R. G. M. Nisbet, Niall Rudd - A Commentary On Horace - Odes Book III (Book 3) - Oxford University Press, USA (2004)

HORACE: ODES
BOOK III
This page intentionally left blank
A CO M M EN TA RY ON
HORACE: ODES
BOOK III
BY
R. G. M. NISBET
AND
NIALL RUDD
1
Great Clarendon Street, Oxford ox2 6dp
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan South Korea Poland Portugal
Singapore Switzerland Thailand Turkey Ukraine Vietnam
Published in the United States
by Oxford University Press Inc., New York
© R. G. M. Nisbet and Niall Rudd 
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
First published 2004
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above.
You must not circulate this book in any other binding or cover
and you must impose this same condition on any acquirer.
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
ISBN 0–19–926314–
3 5 7 9 10 8 6 4
Typeset by Kolam Information Services Pvt. Ltd., Pondicherry, India
Printed in Great Britain on acid-free paper by
Biddles Ltd., King’s Lynn, Norfolk
PR EFAC E
Th i s work follows the same lines as the commentaries by Nisbet and

Hubbard on Books I and II of the Odes (Oxford, 1970 and 1978). It
concentrates on individual poems and problems, and aims to elucidate
the poet’s meaning at the most literal level; it is not another book about
the Odes in general. Yet in view of the lapse of time since the earlier
volumes we have repeated a few facts in the General Introduction, and
at the same time have summarized our approach, particularly on con-
troversial matters.
Recently there has been some discussion about the commentary as
a literary form: see G. W. Most (ed.), Commentaries—Kommentare
(Göttingen, 1999), R. K. Gibson and C. S. Kraus (edd.), The Classical
Commentary (Leiden, 2002). In the case of Horace the size of the
bibliography causes particular difficulty; inevitably our own reading
has been selective. While a commentary should be clear at all costs
and not unreasonably long, these aims would never have been realized if
we had done full justice even to the more important books and articles.
As in the earlier volumes the editors try to support their interpret-
ations by citing parallel passages; these may record an allusion to a
predecessor, exemplify a commonplace, provide the reason for prefer-
ring a textual variant, illustrate a syntactical usage, or give evidence for a
historical or antiquarian point. We use the catch-all ‘cf.’ to introduce
these different types of parallel; it is objected that this obscures import-
ant distinctions and fails to show how the author is using his models,
but the reason for the citation is usually obvious, and where Horace
significantly modifies his predecessor a note is normally supplied. To
avoid clogging the exegesis with lengthy lists, we have often selected the
earliest or most interesting parallels and then added a cross-reference to
TLL, OLD, or a more expansive commentator like Mayor, Pease, or
Bömer. We do not hesitate to cite classical authors later than Horace, as
they may exemplify a standard locution or be derived from a common
source. We have sometimes quoted imitations of Horace in major
English poets; these should not be allowed to determine the interpret-
ation of our text, though of course the reception of Horace is an
important theme in the study of European literature (see for instances
the introduction to 3. 30).
Needless to say, in recording parallels we are not denying Horace’s
originality, as some critics of the first volume supposed. In fact we
regard him as one of the most original of ancient poets for his ability
vi P R EFAC E
to integrate political and philosophical themes in his lyrics, his virtuos-
ity in adapting Greek metres to the heavier Latin language, his use of
traditional forms to present his unique personality, and above all the
range of his style and tone which his imitators have found inimitable.
As our collaboration developed we reached a large measure of agree-
ment. In the few places where we differed, rather than attempt an
unsatisfactory compromise we have used our initials to indicate our
separate positions. As before, the editors owe much to previous com-
mentators, especially Bentley, Orelli–Hirschfelder, and Kiessling–
Heinze, and to the interpretation of the Odes by H. P. Syndikus (edn.
3, Darmstadt, 2001); the attractive short commentary on Book 3 by
David West (Oxford, 2002) appeared too late to be consulted. It
remains only to thank the staff of the Oxford University Press for
bringing the book to completion.
Corpus Christi College, Oxford R. G. M. N.

University of Bristol N. R.
August 2003
CO N TEN TS
bibliography ix
general introduction xix
1. horace’s early life xix
2. the date of Odes i–iii xix
3. the ‘roman odes’ xx
4. horace and augustus xxi
5. maecenas and other addressees xxii
6. horace’s ‘love-poems’ xxiii
7. religion in horace xxiii
8. the meaning of the author xxiv
9. ambiguity xxv
10. person and persona xxvi
11. genre xxvi
12. style xxvii
13. structure xxvii
14. the arrangement of the book xxviii
15. the text xxix
16. the ancient commentators xxix
17. metre xxx
commentary 1
index nominvm 379

index verborvm 383
index rervm 387
B I B LI O G R A P H Y
This bibliography lists books cited in abbreviated form in the commen-

tary; references like ‘Kroll 24’ can be elucidated here. It does not include
either articles or commentaries on other works. In the commentary a
select bibliography is prefixed to each ode; the expression ‘op. cit.’
normally refers to those bibliographies, not to this one. For further
details see W. Kissel, ANRW II. 31. 3 (Berlin, 1981), 141 ff.;
E. Doblhofer, Horaz in der Forschung nach 1957 (Darmstadt, 1992);
W. Kissel in S. Koster (ed.), Horaz-Studien (Erlangen, 1994), 116 ff.
(a) texts and commentaries

For fuller lists see Schanz–Hosius 2. 152 and Kissel (1981) cited above.
Lambinus, D. (1561), Lyons.
Bentley, R. (1711), Cambridge; edn. 3 (1728), Amsterdam (repr. 1869).
Mitscherlich, C. G. (1800), vol. 2, Leipzig.
Peerlkamp, P. Hofman (edn. 2, 1862), Amsterdam.
Schütz, H. (edn. 3, 1881), Berlin.
Orelli, J. C., revised by W. Hirschfelder (edn. 4, 1886), Berlin.
Kiessling, A. (edn. 2, 1890), Berlin.
Page, T. E. (1895), London.
Wickham, E. C. (edn. 3, 1896), Oxford.
Gow, J. (1896), Cambridge.
Keller, O., and Holder, A. (edn. 2, 1899), Leipzig (text and parallels).
Müller, L. (1900), 2 vols., St Petersburg and Leipzig.
Shorey, P., and Laing, G. J. (edn. 2, 1910), Chicago, repr. Pittsburgh, 1960.
Wickham, E. C., revised by H. W. Garrod (edn. 2, 1912), Oxford Classical
Texts.
Darnley Naylor, H. (1922), Cambridge.
Heinze, R. (edn. 7 of Kiessling, 1930; edn. 10, 1960), Berlin.
Campbell, A. Y. (edn. 2, 1953), Liverpool.
Klingner, F. (edn. 3, 1959), Leipzig (text only).
Williams, G. (1969), Oxford (Book 3 only).
Quinn, K. (1980), London.
Borzsák, S. (1984), Leipzig (text only).
Shackleton Bailey, D. R. (1985, revised 2001), Stuttgart (text only).
x B I B LI O G R A P H Y
Syndikus, H. P. (2001), Die Lyrik von Horaz edn. 3, 2 vols., Darmstadt (a literary
commentary with valuable detail).
West, D. (2002), Dulce Periculum, Oxford (Book 3 only).
(b) other books cited

Abbe, E. (1965), The Plants of Virgil ’s Georgics, Ithaca.
Adams, J. N. (1982), The Latin Sexual Vocabulary, London.
—— and Mayer, R. G. (edd.) (1999), Aspects of the Language of Latin Poetry,
Oxford.
Allen, W. S. (1965 and 1978), Vox Latina: A Guide to the Pronunciation of Classical
Latin, Cambridge.
Anderson, J. K. (1961) Ancient Greek Horsemanship, Berkeley.
—— (1985), Hunting in the Ancient World, Berkeley.
André, J. (1949), Études sur les termes de couleur dans la langue latine, Paris.
—— (1967), Les Noms d’oiseaux en latin, Paris.
Appel, G. (1909, repr. 1975), De Romanorum precationibus, Giessen.
Axelson, B. (1945), Unpoetische Wörter: Ein Beitrag zur Kenntnis der lateinischen
Dichtersprache, Lund.
Beard, M., North, J., and Price, S. R. (1998), Religion in Rome, 2 vols., Cam-
bridge.
Bell, A. J. (1923), The Latin Dual and Poetic Diction, London and Toronto.
Binder, G. (1971), Aeneas und Augustus, Interpretationen zum 8. Buch der Aeneis,
Meisenheim am Glan.
Blümner, H. (1875–87, vol. 1, edn. 2, 1912), Technologie und Terminologie der
Gewerbe und Künste bei Griechen und Römern, Leipzig.
—— (1911), Die Römischen Privataltertümer, Munich.
Bo, D. (1960), De Horati poetico eloquio, vol. 3 of Q. Horati Flacci opera (Corpus
Scriptorum Latinorum Paravianum), Turin.
Bolton, J. D. P. (1962), Aristeas of Proconnesus, Oxford.
Bompaire, J. (1958), Lucien écrivain: imitation et création, Paris.
Boucher, J.-P. (1965), Études sur Properce: Problèmes d ’interprétation et d ’art,
Paris.
Bruchmann, C. F. H. (1893), Epitheta deorum quae apud poetas Graecos leguntur,
supplement in Roscher, vol. 7, Leipzig.
Brunt, P. A. (1971) Italian Manpower, Oxford.
—— (1990), Roman Imperial Themes, Oxford.
Burkert, W. (1985), Greek Religion, Archaic and Classical (Oxford), translation of
Griechische Religion der archaischen und klassischen Epoche (1977).
Cairns, F. (1972), Generic Composition in Greek and Roman Poetry, Edinburgh.
Campbell, A. Y. (1924), Horace: A New Interpretation, London.
Capponi, A. (1979), Ornithologia Latina, Genoa.
B I B LI O G R A P H Y xi
Carter, J. B. (1902), Epitheta deorum quae apud poetas Latinos leguntur, supple-
ment in Roscher, vol. 7, Leipzig.
Cavarzere, A. (1996), Sul limitare: Il ‘motto’ e la poesia di Orazio, Bologna.
Christ, F. (1938), Die römische Weltherrschaft in der antiken Dichtung (Tübinger
Beitr. 31), Tübingen.
Collinge, N. E. (1961), The Structure of Horace’s Odes, London.
Commager, S. (1962), The Odes of Horace: A Critical Study, New Haven and
London.
Copley, F. O. (1956), Exclusus Amator: A Study in Latin Love Poetry (Amer.
Philol. Assoc. monograph 17).
Costa, C. D. N. (ed.) (1973), Horace, London and Boston.
Crook, J. A. (1967), Law and Life of Rome, London.
Curtius, E. R. (1953), European Literature and the Latin Middle Ages, London,
translated from the German edition.
Davis, G. (1991), Polyhymnia: The Rhetoric of Horatian Lyric Discourse, Berkeley
and Los Angeles.
Dickey, E. (2002), Latin Forms of Address, Oxford.
Doblhofer, E. (1966), Die Augustuspanegyrik des Horaz in formalhistorischer Sicht,
Heidelberg.
Earl, D. C. (1961), The Political Thought of Sallust, Cambridge.
Ernout, A., and Meillet, A. (edn. 4, 1959), Dictionnaire étymologique de la langue
latine, 2 vols., Paris.
Esser, D. (1976), Untersuchungen zu den Odenschlüssen bei Horaz (Beitr. zur klass.
Philol. 77), Meisenheim am Glan.
Flower, H. I. (1996), Ancestor Masks and Aristocratic Power in Roman Culture,
Oxford.
Fraenkel, E. (1957), Horace, Oxford.
—— (1960), Elementi plautini in Plauto, Florence, translation with addenda of
Plautinisches im Plautus (1922), Berlin.
—— (1964), Kleine Beiträge zur klassischen Philologie, 2 vols., Rome.
Führer, R. (1967), Formproblem—Untersuchungen zu den Reden in der frühgrie-
chischen Lyrik (Zetemata 44), Munich.
Galinsky, K. (1996), Augustan Culture: An Interpretive Introduction, Princeton.
Gatz, B. (1967), Weltalter, goldene Zeit und sinnverwandte Vorstellungen (Spudas-
mata 16), Hildesheim.
Grassmann, V. (1966), Die erotischen Epoden des Horaz (Zetemata 39), Munich.
Griffin, J. (1985), Latin Poets and Roman Life, London.
Gutzwiller, K. J. (1998), Poetic Garlands: Hellenistic Epigrams in Context, Berke-
ley and Los Angeles.
Hand, F. (1829–45), Tursellinus seu de particulis Latinis commentarii, Leipzig,
repr. (1969) Amsterdam.
xii B I B LI O G R A P H Y
Handford, S. A. (1947), The Latin Subjunctive: Its Usage and Development from
Plautus to Tacitus, London.
Hardie, P. (1986), Virgil ’s Aeneid: Cosmos and Imperium, Oxford.
Harrison, S. J. (ed.) (1995), Homage to Horace, Oxford.
—— (ed.) (2001), Texts, Ideas, and the Classics, Oxford.
Hilgers, W. (1969), Lateinische Gefässnamen, Düsseldorf.
Horden, P., and Purcell, N. (2000), The Corrupting Sea: A Study of Mediterra-
nean History, Oxford.
Housman, A. E. (1972), Classical Papers (ed. J. Diggle and F. R. D. Goodyear), 3
vols., Cambridge.
Hubbard, M. (1974), Propertius, London.
Irwin, E. (1974), Colour Terms in Greek Poetry, Toronto.
Kambylis, A. (1965), Die Dichterweihe und ihre Symbolik, Heidelberg.
Keller, O. (1909–20, repr. 1963), Die antike Tierwelt, 2 vols., Leipzig.
Kroll, W. (1924), Studien zum Verständnis der römischen Literatur, Stuttgart, repr.
(1964) Darmstadt.
Lacey, W. K. (1996), Augustus and the Principate: The Evolution of the System (Arca
35), Leeds.
La Penna, A. (1963), Orazio e l ’ideologia del principato, Turin.
—— (1993), Saggi e studi su Orazio, Florence.
Latte, K. (1960), Römische Religionsgeschichte, Munich.
Lattimore, R. (1942), Themes in Greek and Latin Epitaphs (Illinois Studies in
Language and Literature 28), Urbana.
Le Boeuffle, A. (1977), Les Noms latins d ’astres et de constellations, Paris.
Leo, F. (1912), Plautinische Forschungen, Berlin, repr. (1966) Darmstadt.
Lieberg, G. (1982), Poeta Creator: Studien zu einer Figur der antiken Dichtung,
Amsterdam.
Löfstedt, E. (vol. 1, edn. 2, 1942; vol. 2, 1933), Syntactica: Studien und Beiträge zur
historischen Syntax des Lateins, Lund.
Lovejoy, A. O., and Boas, G. (1935, repr. 1997), Primitivism and Related Themes
in Antiquity, Baltimore.
Lowrie, M. (1997), Horace’s Narrative Odes, Oxford.
Lyne, R. O. A. M. (1980), The Latin Love Poets from Catullus to Horace, Oxford.
—— (1987), Further Voices in Vergil ’s Aeneid, Oxford.
—— (1989), Words and the Poet: Characteristic Techniques of Style in Vergil ’s
Aeneid, Oxford.
—— (1995), Horace: Behind the Public Poetry, New Haven and London.
Maltby, R. (1991), A Lexicon of Ancient Latin Etymologies (Arca 25), Leeds.
Marquardt, J., and Mau, A. (edn. 2, 1886), Das Privatleben der Römer, Leipzig.
Meiggs, R. (1982), Trees and Timber in the Ancient Roman World, Oxford.
Millar, F., and Segal, E. (edd.) (1984), Caesar Augustus: Seven Aspects, Oxford.
B I B LI O G R A P H Y xiii
Murray, O., and Tecus˛an, M. (edd.) (1995), In Vino Veritas, London.
Neue, F., and Wagener, C. (edn. 3, 1892–1905), Formenlehre der lateinischen
Sprache, Berlin.
Newman, J. K. (1967), Augustus and the New Poetry (Coll. Latomus 100),
Brussels.
Nicolet, C. (1991), Space, Geography, and Politics in the Early Roman Empire, Ann
Arbor.
Nisbet, R. G. M. (1995), Collected Papers on Latin Literature, Oxford.
Nissen, H. (1883–1902), Italische Landeskunde, 2 vols., Berlin.
Nock, A. D. (1972), Essays on Religion and the Ancient World, 2 vols., Oxford.
Norden, E. (1913), Agnostos Theos: Untersuchungen zur Formengeschichte religiöser
Rede, Leipzig–Berlin, repr. (1956) Darmstadt.
O’Hara, J. J. (1996), True Names: Vergil and the Alexandrian Tradition of Etymo-
logical Wordplay, Ann Arbor.
Oliensis, E. (1998), Horace and the Rhetoric of Authority, Cambridge.
Onians, R. B. (edn. 2, 1954), The Origins of European Thought about the Body, the
Mind, and the Soul . . . , Cambridge.
Oppermann, H. (ed.) (1972), Wege zu Horaz (Wege der Forschung 99), Darm-
stadt.
Otto, A. (1890), Die Sprichwörter und sprichwörtlichen Redensarten der Römer,
Leipzig, repr. (1962) Hildesheim.
Pape, W., and Benseler, G. F. (edn. 3, 1911), Wörterbuch der griechischen Eigenna-
men, Brunswick, repr. (1959) Graz.
Parker, R. (1983), Miasma: Pollution and Purification in Early Greek Religion,
Oxford.
Paschalis, M. (1997), Virgil ’s Aeneid: Semantic Relations and Proper Names,
Oxford.
Pasquali, G. (1920), Orazio lirico, Florence, repr. with addenda by A. La Penna,
1984.
Pichon, R. (1902), De sermone amatorio apud Latinos elegiarum scriptores, Paris ¼
Index verborum amatoriorum, Hildesheim (1966).
Porter, D. H. (1987), Horace’s Poetic Journey: A Reading of Odes 1–3, Princeton.
Pöschl, V. (edn. 2, 1991), Horazische Lyrik, Heidelberg.
Pulleyn, S. (1997), Prayer in Greek Religion, Oxford.
Raaflaub, K. A., and Toher, M. (edd.) (1990), Between Republic and Empire:
Interpretations of Augustus and his Principate, Berkeley, Los Angeles, Oxford.
Rawson, E. (1985), Intellectual Life in the Late Roman Republic, Oxford.
Reitzenstein, R. (1963), Aufsätze zu Horaz, Darmstadt, a collection of reprinted
articles.
Richardson, L., Jr. (1992), A New Topographical Dictionary of Ancient Rome,
Baltimore and London.
xiv B I B LI O G R A P H Y
Rohde, E. (edn. 3, 1914), Der griechische Roman und seine Vorläufer, Leipzig, repr.
(1960) Hildesheim.
Roscher, W. H. (1884–1937), Ausführliches Lexicon der griechischen und römischen
Mythologie, Leipzig.
Ross, D. O., Jr. (1975), Backgrounds to Augustan Poetry: Gallus, Elegy, and Rome,
Cambridge.
Rudd, N. (1966), The Satires of Horace, Cambridge.
—— (ed.) (1993), Horace 2000: A Celebration, London.
Saller, R. P. (1982), Personal Patronage under the Early Empire, Cambridge.
Salmon, E. T. (1967), Samnium and the Samnites, Cambridge.
Santirocco, M. (1986), Unity and Design in Horace’s Odes, Chapel Hill and
London.
Schanz, M., and Hosius, C. (vol. 2, edn. 4, 1927), Geschichte der römischen
Literatur, Munich.
Schmidt, E. A. (2002), Zeit und Form, Heidelberg.
Schönbeck, G. (1962), Der Locus Amoenus von Homer bis Horaz (Diss.), Heidel-
berg.
Scullard, H. H. (1981), Festivals and Ceremonies of the Roman Republic, London.
Shackleton Bailey, D. R. (1982), Profile of Horace, London.
Simon, E. (1990), Die Götter der Römer, Munich.
Sittl, C. (1890), Die Gebärden der Griechen und Römer, Leipzig.
Steinby, E. M. (ed.) (1993–2001), Lexicon topographicum urbis Romae, Rome.
Stroh, W. (1971), Die römische Liebeselegie als werbende Dichtung, Amsterdam.
Suerbaum, W. (1968), Untersuchungen zur Selbstdarstellung älterer römischer Dich-
ter (Spudasmata 19), Hildesheim.
Syme, R. (1939), The Roman Revolution, Oxford.
—— (1978), History in Ovid, Oxford.
—— (1979–91), Roman Papers, 7 vols., Oxford.
—— (1986), The Augustan Aristocracy, Oxford.
Taillardat, J. (1962), Les Images d ’Aristophane: Études de langue et de style, Paris.
Tarán, S. L. (1979), The Art of Variation in the Hellenistic Epigram, Leiden.
Thomas, R. F. (1999), Reading Virgil and his Texts, Ann Arbor.
Thompson, D’A. W. (1936), A Glossary of Greek Birds, Oxford, repr. (1966)
Hildesheim.
Thomson, J. O. (1948), History of Ancient Geography, Cambridge.
Toynbee, J. M. C. (1973), Animals in Roman Life and Art, London.
Treggiari, S. (1991), Roman Marriage, Oxford.
Troxler-Keller, I. (1964), Die Dichterlandschaft des Horaz, Heidelberg.
Wackernagel, J. (edn. 2, 1926–8), Vorlesungen über Syntax, 2 vols., Basel.
Weinstock, S. (1971), Divus Julius, Oxford.
West, M. L. (1992), Ancient Greek Music, Oxford.
B I B LI O G R A P H Y xv
White, K. D. (1970), Roman Farming, London.
White, P. (1993), Promised Verse: Poets in the Society of Augustan Rome, Cam-
bridge, Mass.
Wilkinson, L. P. (edn. 2, 1951), Horace and his Lyric Poetry, Cambridge.
Wille, G. (1967), Musica Romana, Amsterdam.
Williams, G. (1968), Tradition and Originality in Roman Poetry, Oxford.
Wills, J. (1996), Repetition in Latin Poetry, Oxford.
Wimmel, W. (1960), Kallimachos in Rom (Hermes Einzelschriften 16), Wies-
baden.
Wissowa, G. (edn. 2, 1912), Religion und Kultus der Römer, Munich.
Wölfflin, E. (1933), Ausgewählte Schriften, Leipzig.
Woodcock, E. C. (1959), A New Latin Syntax, London.
Woodman, T., and West, D. (edd.) (1984), Poetry and Politics in the Age of
Augustus, Cambridge.
—— and Feeney, D. (edd.) (2002), Traditions and Contexts in the Poetry of
Horace, Cambridge.
Zanker, P. (1988), The Power of Images in the Age of Augustus, Ann Arbor,
translated from (1987), August und die Macht der Bilder, Munich.
(c) concordances
Cooper, Lane (1916, repr. 1961), A Concordance of the Works of Horace, Washing-
ton.
Iso Echegoyen, J.-J. (1990), Concordantia Horatiana, Hildesheim.
(d) abbreviations
For periodicals see L’Année philologique or OCD edn. 3.
ALL Archiv für lateinische Lexikographie und Grammatik (1884–1909).

ANRW Aufstieg und Niedergang der römischen Welt, ed. H. Temporini
and W. Haase (1972– ).
CAH Cambridge Ancient History (edn. 2, 1961– ).
CGL Corpus Glossariorum Latinorum, ed. G. Goetz (1888–1923).
CIL Corpus Inscriptionum Latinarum (1863– ).
CLE Carmina Latina Epigraphica, ed. F. Bücheler and
E. Lommatzsch (1895–1926).
CMA The Oxford Guide to Classical Mythology in the Arts, 1300–1990s,
ed. J. D. Reid, 2 vols. (1993).
Coll. Alex. Collectanea Alexandrina, ed. J. U. Powell (1925).
CRF Comicorum Romanorum Fragmenta, ed. O. Ribbeck (1898).
DNP Der Neue Pauly (1996– ).
D–S C. Daremberg and E. Saglio, Dictionnaire des antiquités grecques
et romaines (1877–1919).
xvi B I B LI O G R A P H Y
Encicl. oraz. Enciclopedia oraziana (1996–8).
Encicl. virg. Enciclopedia virgiliana (1984–91).
FGrH Fragmente der griechischen Historiker, ed. F. Jacoby (1923–58).
FLP The Fragmentary Latin Poets, ed. E. Courtney (1993).
FPL Fragmenta Poetarum Latinorum, ed. W. Morel (1927); edn. 2 ed.
K. Büchner (1982).
GL Grammatici Latini, ed. H. Keil (1855–80).
GLP Greek Literary Papyri: Poetry, ed. D. L. Page (1942).
GV Griechische Vers-Inschriften I: Grab-Epigramme, ed. W. Peek
(1955).
HE Hellenistic Epigrams, ed. A. S. F. Gow and D. L. Page, 2 vols.
(1965).
H–Sz J. B. Hofmann and A. Szantyr, Lateinische Syntax und Stilistik
(Handb. der Alt. 2. 2. 2), 1965.
IG Inscriptiones Graecae (1873– ).
IGRR Inscriptiones Graecae ad res Romanas pertinentes, ed. R. Cagnat
etc. (1901–27).
ILLRP Inscriptiones Latinae Liberae Rei Publicae, ed. A. Degrassi, edn. 2
(1957–63).
ILS Inscriptiones Latinae Selectae, ed. H. Dessau (1892–1916).
K–G R. Kühner and B. Gerth, Ausführliche Grammatik der Grie-
chischen Sprache: Satzlehre (1898–1904).
K–S R. Kühner and C. Stegmann, Ausführliche Grammatik der latei-
nischen Sprache: Satzlehre (1912–14).
LGPN A Lexicon of Greek Personal Names, ed. P. M. Fraser, etc. (1987– ).
LIMC Lexicon Iconographicum Mythologiae Classicae (1981–99).
LSJ Liddell and Scott, Greek–English Lexicon, rev. H. Stuart Jones,
edn. 9 (1925–40).
MRR Magistrates of the Roman Republic, ed. T. R. S. Broughton (1951–
86).
Mus. Lap. Musa Lapidaria, ed. E. Courtney, American Classical Studies 36
(1995).
N–H R. G. M. Nisbet and M. Hubbard, Commentary on Horace, Odes
1 (1970), 11 (1978).
OCD The Oxford Classical Dictionary, edn. 3, ed. S. Hornblower and
A. Spawforth (1996).
OGIS Orientis Graeci Inscriptiones Selectae, ed. W. Dittenberger
(1903–5).
OLD Oxford Latin Dictionary, ed. P. G. W. Glare (1968–82).
PCG Poetae Comici Graeci, ed. R. Kassel and C. Austin (1983– ).
PG Patrologiae Cursus Completus, Series Graeca, ed. J.-P. Migne
(1857–66).
PIR Prosopographia Imperii Romani, edn. 1, (1897–8), edn. 2, (1933– ).
PL Patrologiae Cursus Completus, Series Latina, ed. J.-P. Migne
(1844–64).
B I B LI O G R A P H Y xvii
PLM Poetae Latini Minores, ed. A. Baehrens (1879–81), Leipzig, rev.
F. Vollmer (1911–35).
PMG Poetae Melici Graeci, ed. D. Page (1962).
RAC Reallexikon für Antike und Christentum (1941– ).
RE Real-Encyclopädie der classischen Altertumswissenschaft, ed.
A. Pauly, G. Wissowa, W. Kroll (1893– ).
ROL Remains of Old Latin (Loeb edn.), ed. E. H. Warmington, 4
vols. (1934–53).
SIG Sylloge Inscriptionum Graecarum, ed. W. Dittenberger, edn. 3
(1915–24).
Supp. Hell. Supplementum Hellenisticum, ed. H. Lloyd-Jones and P. Parsons
(1983).
TGF Tragicorum Graecorum Fragmenta, edn. 2, ed. A. Nauck (1889),
suppl. by B. Snell (1964).
TLL Thesaurus Linguae Latinae (1900– ).
TRF Tragicorum Romanorum Fragmenta, ed. O. Ribbeck (1897).
TrGF Tragicorum Graecorum Fragmenta, ed. B. Snell, R. Kannicht, S.
Radt (1971– ).
G EN ER A L I N T RO D U C T I O N
1. Horace’s early life

Horace was born on 8 December 65 bc (3. 21. 1, epist. 1. 20. 27 f., Suet.
vita 71R) at Venusia in Apulia (serm. 2. 1. 34 ff., carm. 3. 4. 9 ff.). His
father had once been a slave, probably as a result of capture in the Social
War (G. Williams ap. Harrison, 1995: 296 ff.); the stigma of servile
origin, however unfair, remained and is made clear at serm. 1. 6. 5 ff.
and 45 f. After his emancipation the father made good as an auctioneer
and provider of credit (serm. 1. 6. 86, Fraenkel 4 f.), and could afford to
educate his son not only at Rome (serm. 1. 6. 76 ff.) but also at Athens
(epist. 2. 2. 43 ff.); Horace exaggerates the humbleness of his origins
(3. 30. 12, epist. 1. 20. 20), but by the standards of his later friends his
background was undoubtedly restricted. In 42 bc he served as a tribunus
militum under Brutus at Philippi (serm. 1. 6. 48, carm. 2. 7. 9 ff., 3. 4. 26),
evidence of energy and organizational ability; but though he lost his
patrimony (epist. 2. 2. 50 f.), within a few years he had made peace with
Octavian’s victorious faction, obtained a high-ranking post in the treas-
ury (serm. 2. 6. 36, Suet. vita 8, Fraenkel 14 f.), and resumed his position
as an eques Romanus (serm. 2. 7. 53, Lyne, 1995: 3 n.). About 37 bc he was
befriended by Maecenas (serm. 2. 6. 40), under whose auspices he wrote
two books of sermones or satires (issued about 35 and 30) and completed
his collection of iambi or epodes (again about 30). In 36 he saw some-
thing of Octavian’s war against Sextus Pompeius (3. 4. 28 n.), and in 31 he
seems to have accompanied Maecenas to Actium (epod. 1 and 9, cf.
perhaps carm. 3. 1. 38–9 n.).1 In the meantime Maecenas had presented
him with a property in the Sabine hills (serm. 2. 6. 1 ff., carm. 1. 17, 3. 1.
47–8 n.) that gave him an income from his tenants’ rents, and the leisure
to write. For further biographical detail see Enciclopedia oraziana 1. 217 ff.
2. The date of Odes I–III

The first three books were issued together (epist. 1. 13. 2 speaks of
volumina), but the poems were not in chronological order. The date
was probably 23 bc in the consulship of Sestius (whose position in 1. 4 is
otherwise hard to explain), before the death of Marcellus later in the
1
See E. Wistrand, Horace’s Ninth Epode, 1958 (¼Opera Selecta, 1972: 293 ff.), R. G. M.
Nisbet ap. Woodman and West (1984), 9 ff. (underlining the need to read huc at epod. 9. 17),
I. M. Le M. Du Quesnay ap. Woodman and Feeney (2002), 17 ff.
xx H O R AC E : O D ES I I I
same year (N–H vol. 1, p. 145 on 1. 12. 45 f.), and before the disgrace of
Murena (the recipient of 2. 10),2 which is put by Dio 54. 3 in 22 bc.
Some of the non-political poems may have been written earlier than
Actium (Encicl. oraz. 1. 220), before the Satires were completed, but
political allusions are the most reliable criterion of date. It is sometimes
said that Horace might have made revisions up to 23; but though an
elegiac or hexameter poet might have done it, this would have been
more difficult with the complex structures of the Odes. G. O. Hutch-
inson now argues that the three books were issued separately (CQ 52,
2002: 517 ff.); though he does not persuade us, he provides some valuable
details.
3. The ‘Roman Odes’

The first six poems of Book III have been called the Roman Odes at
least since Plüss (1882). They share the Alcaic metre in contrast to
Horace’s usual variatio, a substantial length, an absence of individual
addressees, a subject-matter that concentrates on the political and moral
issues which were thought important by the new regime, and an im-
pressive seriousness of style. 3. 4 seems to belong to 29, when Octavian
returned in triumph from the East and demobilized his army (38 n.), 3. 6
looks forward to his repair of the temples in 28 (res gest. 20. 4), a date
that also suits the apparent abandonment of his first attempt at moral
legislation (see the introduction to that poem). In January 27 he as-
sumed the name ‘Augustus’, by which he is described at 3. 3. 11 and 3. 5. 3;
later that year he departed for Gaul, from where he was expected to
invade Britain (cf. 3. 5. 3–4 n. and possibly 3. 3. 56). 3. 2 cannot be dated;
3. 1 serves as an introduction to the series and perhaps to the book as a
whole.
Many have seen in the Roman Odes not just a common form and
purpose but a carefully planned unity of design. Mommsen thought the
series celebrated the new constitution of 27 bc (cf. Reden und Aufsätze,
1905: 168 ff.), but this seems too late for nos. 4 and 6. Domaszewski
found in poems 2–5 the qualities represented on the shield presented to
Augustus in 27, virtus, iustitia, clementia, pietas (RhM 59, 1904: 302 ff.);
but that, apart from being incomplete, is far too schematic a treatment.
Many have claimed to detect various patterns of arrangement and
cross-reference,3 but these are often arbitrary and unconvincing: for
2
The Murena of 2. 10 (a poem that commends the Golden Mean) must be the alleged
conspirator, one of whose associates was the Peripatetic philosopher Athenaeus (Strabo 14.
5. 4, N-H vol. 2, p. 152).
3
See for instance H. Wagenvoort, De Horatii quae dicuntur Odis Romanis, Diss.
Groningen 1911: 18 ff. G. E. Duckworth, TAPA 87, 1956: 299 ff., M. Santirocco (1986),
G EN ER A L I N T RO D U C T I O N xxi
example, the simplicity of life commended in 3. 1 is not the same as the
pauperies that the young soldier must learn to endure (3. 2. 1); 3. 1. 7 ‘clari
Giganteo triumpho’ (of Jupiter himself ) has a different function from 3. 4.
49 ff. (the defeat of the Titans), which is a clear political analogy to the
overthrow of the Antonians. Some have even thought of treating the
series as one long poem,4 but the dates of the odes are different, their
subjects quite distinct, and all have convincing openings and closures.
4. Horace and Augustus

In considering this question we reject two contradictory approaches. A
former generation of scholars, well represented by Fraenkel (1957), was
content to accept Augustan ideology on its own terms, without regard
to the violence and deception that characterized Octavian’s seizure and
retention of power, and to assume that Horace felt the same in every
respect. A contrary and more recent approach has been to exaggerate
the similarities between Augustus and the chief dictators of the twenti-
eth century, and then sometimes to seek hints of subversion in Horace;
this is to ignore the poet’s closeness to the regime, as shown later by
Augustus’ wish to make him his secretary (Suet. vita 18), and also to
disregard the feelings of loyalty that counted for more in Rome than the
political independence valued in modern democracies. It can be debated
how far Horace was sincere in his support of Augustus’ policies, indeed
whether the concept of sincerity is relevant to the public utterances of a
court-poet (see the introduction to 3. 6); but whatever view one takes of
his private commitment, it must be agreed that Horace showed great
skill in selecting illustrations which he knew would have a wide appeal.
Thus Antony is damned indirectly by eloquent Pindaric allegories (3. 4);
Augustus’ moral policy is made more acceptable by vignettes of metro-
politan decadence and rustic simplicity (3. 6); the abandonment of the
prisoners in Parthia is justified by invoking the legendary self-sacrifice
of Regulus (3. 5); references to the ruler-cult in Rome are confined to
the future (3. 3. 11, 3. 5. 2), where they would cause less offence to
traditional attitudes.
Apart from the Roman Odes a few poems in the book are concerned
with Augustus. 3. 24 (like 3. 1) denounces materialism and (like 3. 6) calls
for moral revival; the implication that earlier attempts have failed
(vv. 25 ff.) suggests that it too should be assigned to about 28 bc. In
111 ff. For the independent composition of the six odes see R. Heinze, Vom Geist des
Römertums, edn. 3, 1960: 190 ff., L. Amundsen, SO suppl. 11, 1942: 1 ff. (¼ Oppermann,
1972: 111 ff.).
4
Diomedes (GL 1. 251) regards 3. 7 as the second ode in the book (cf. Porph. on 3. 1. 1);
add S. J. Heyworth in Formative Stages of Classical Traditions (ed. O. Pecere and
M. D. Reeve), 1995: 117 ff., A. Griffiths ap. Woodman and Feeney, 2002: 73 ff.
xxii H O R AC E : O D ES I I I
3. 25, under the inspiration of Bacchus, the poet talks of celebrating the
Princeps; this is often thought to refer to the Roman Odes, but the
celebration may not look beyond the poem itself. In 26 bc Augustus was
absent in Spain fighting the Cantabrians in the north, and in 25 he was
seriously ill at Tarraco on the east coast; see the introduction to 3. 8,
which we assign to the latter year. In 24 bc Horace celebrates the great
man’s return to Rome in an ode that combines his roles as a public and a
private poet (3. 14); here he emphasizes what all reasonable people must
have felt by that date, that the survival of Augustus is at once the
strongest guarantee against the renewal of civil war and the best hope
for the country’s regeneration.
5. Maecenas and other addressees

Three odes in the book are addressed to Maecenas: 3. 8, 3. 16 (beginning
the second half ), and 3. 29 (the last poem before the epilogue); in
addition 3. 1 has some pointers in the same direction (see the introduc-
tion to that poem). Maecenas was not Horace’s patronus (a word not
used in the Augustan period of literary patrons), but rather his amicus—
even if an unusually grand one (Saller, 1982: 8 ff., P. White: 1993, 29 ff.,
280 f.). The poems mentioned above allude to various aspects of his life
and personality—his pride in his Etruscan ancestors, the grandeur of his
life-style, his wide learning, his munificence, and the worries caused by
his political responsibilities (especially in the absence of Augustus). At
the same time Horace is ready to tease him about his eminence (Lyne,
1995: 102 ff.), and even to hint, perhaps, that his wealth has not brought
him so much happiness as the Sabine estate has brought to the poet (see
3. 1 and 3. 16). In Book 4, when Maecenas’ political power seems to have
waned (Lyne, 1995: 136 ff., 189 ff.) Horace still speaks of him warmly (11.
17 ff.), and we are told that Maecenas’ final commendation to Augustus
was ‘Horatii Flacci ut mei memor esto’.
In the period of Odes 1–3 Maecenas was still close to Augustus, and in
spite of his equestrian status he had a deserved reputation for diplomatic
skill (serm. 1. 5. 27 ff., carm. 3. 16. 15 n.). One no longer thinks of him as a
propaganda-minister issuing fiats to poets, but the emphasis is now
sometimes too much the other way. It is not enough to point out that
people absorb their opinions from the prevailing atmosphere, for in the
twenties Augustus was still consolidating his position, and positive
guidance was needed. In the Roman Odes Horace seems to have
followed the official line in every particular (see also 3. 24. 54 ff. for
the criticism of young men’s sports), and Maecenas was the obvious
intermediary between the Princeps and the poet; no one, least of all
Horace himself, would have regarded the gift of the Sabine estate as an
act of wholly disinterested benevolence.
G EN ER A L I N T RO D U C T I O N xxiii
Only three other poems are addressed to identifiable people, namely
Aelius Lamia (17), Murena the augur (19), and the great orator and
statesman Messalla Corvinus (21). As usual Horace includes tactful or
humorous allusions to the personalities and families of the recipients.
Yet, unlike the second book of odes and the first book of epistles, Odes 3
puts rather little emphasis on friendship.
6. Horace’s ‘love-poems’
Horace’s KæøØŒ show little of the emotional involvement found in
Catullus or Propertius. One of his roles is that of the urbane and
experienced consultant. Thus he urges Asterie to ignore her serenading
lover (7), consoles the love-lorn Neobule (12), and warns Pyrrhus not to
compete with a predatory woman for the favours of a good-looking boy
(20). When he professes to speak of his own case, he wittily adapts the
traditional situations of love-poetry, the paraclausithyron in 10, the
renuntiatio amoris in 26, the propempticon in 27; when he reminds
Lyde of the heroically loyal Hypermestra (11) and Galatea of the spec-
tacularly indiscreet Europa (27), his exempla are entertaining rather than
moving. His amusement is often directed wryly at himself: Lydia is
given the last word in her tart exchanges with the poet (26), if Lyce and
Neaera are unresponsive (10 and 14), he will not persist, and though he
pretends (unconvincingly) to have given up his interest in girls, he says
he would like to get his own back on Chloe (26). He admits to many
relationships with both puellae and pueri (epod. 11. 4, serm. 2. 3. 325, carm.
4. 1. 29 ff.), and his references to hetaerae no doubt reflect personal
experience (Griffin, 1985: 20 f.), but that is not to say that the names
and situations are to be taken as historically authentic. He does not lay
claim to lasting affections (4. 1. 30 ff., cf. 1. 13. 17 ff.), whether because of
the ambiguity of his social position or simply his inborn nature. Some-
times he is more brutally sexist than any other Augustan poet (see epod.
8 and 12, serm. 1. 2. 116 ff., carm. 1. 25, 2. 5, 3. 15, 4. 13, epist. 1. 18. 71 ff.); yet
towards the end he seems to have regretted the loneliness which his
bachelor life-style has brought (4. 1. 30 f.). For further discussion see
N–H vol. 1, pp. xvi f., Lyne (1980), 201 ff., B. Arkins ap. Rudd (1993),
106 ff., Encicl. oraz. 1. 527 ff.
7. Religion in Horace
Other people’s religions are often hard to understand. That of ancient
Rome may seem unattractive because of its blood-sacrifices (3. 13. 3 ff.),
its bargaining spirit (3. 18. 5 ff.), its legalistic insistence on verbal accur-
acy (3. 21. 5 n.), the absurdity of its superstitions (3. 27. 11 n. on augury),
its complacency about Rome’s role in the divine purpose (3. 6. 1 ff.). Yet
Horace, like Virgil, conveys some of the deeper feelings that antiquarian
xxiv H O R AC E : O D ES I I I
pedantry and anthropological speculation cannot illuminate: the recur-
ring festivals reflect the age-old rhythms of the agricultural year (3. 13,
3. 18), there is awe at the mystery of woods, caves, and springs (cf. 3. 25.
2 n. and the introduction to 3. 13), the solemn rites convey a sense of
peace and order (3. 1. 2 n., 3. 14. 5 ff., 3. 30. 8 f.), as in the tableaux of the
Ara Pacis. Moreover, Roman religion was unusually tolerant and inclu-
sive, as is shown by the incorporation of Greek cults even in the earliest
times (3. 3. 9 n., 3. 14. 1); it found a place for slaves and freedmen (see
3. 23 on the Penates), women had goddesses to suit their special needs
(3. 22. 2 ff.), and as it was not constricted by any formal creed it could
accommodate even a sceptic like Horace. See further Wissowa (1912)
and Latte (1960) for antiquarian detail; for more modern approaches
add Beard–North–Price (1998), D. Feeney, Literature and Religion at
Rome (1998), especially the summary at 2 ff.
8. The meaning of the author

To establish Horace’s meaning on the most literal level may seem
difficult when one is dealing with a dead language; yet in this respect
Horace is easier than Shakespeare and far easier than many a modern.
According to one theory which has been familiar for over half a century,
the writer’s intention is always unknowable. This dogma exaggerates the
difficulties in the concept,5 and underestimates the amount of common
ground shared by poetry and everyday communication: if even a tenth of
our ordinary discourse were as problematic as poetic discourse is sup-
posed to be, social life would soon become chaotic. So in dealing with
basic questions of language we have followed a long and well-tried
tradition, inviting others to refute our views (or to supplement them)
by evidence and argument.
On broader issues, which Horace may not have consciously con-
sidered, modern theory has more to offer. When he dreams of swim-
ming down the Tiber in pursuit of Ligurinus (4. 1. 40), a Freudian
psychologist might provide analogous case-histories, a social anthro-
pologist could show how the gift of an estate created binding obliga-
tions, types of ancient slavery have been illuminated by Marxists, we
ourselves have no doubt been influenced by feminists in our criticism of
Horace’s treatment of women. On more literary matters the critic can
study ‘interaction in imagery’ (3. 11. 41–2 n.), or the point of view in a
narrative (see the story within a story at 3. 7. 9 ff.), or the poem that
refers to itself (3. 25. 7 n.), or the part of nightfall in closures (3. 28.
16 n.); such things were not discussed in ancient rhetorical theory, which
naturally concentrated on prose, yet they are clearly relevant to the
5
See N. Rudd, Phoenix 18, 1964: 221 ff., M. Heath, Interpreting Classical Texts, 2002:
59 ff.
G EN ER A L I N T RO D U C T I O N xxv
understanding of ancient poems. Where we take issue with post-
moderns is over their dogmatic fervour6 and their esoteric use of
language, which implies that their subject is not for the profani. If such
critics would apply their insights more often to the exegesis of particular
poems, it might be easier to have a constructive debate.
9. Ambiguity
This word may describe various phenomena. In the most obvious kind
of case a double meaning is exploited for an amusing effect, as in the
combination of wine-jar and divinity (3. 21). Often the associations of a
word allow a verbal play not transferable to English (3. 4. 44 ‘fulmine
sustulerit caduco’, 3. 27. 22 ‘orientis Austri’). In aiming at precision the
commentator may make unnecessary distinctions: thus it can be debated
whether at 3. 4. 75 Orcum is the underworld or the god or both at once.
Some would argue that we have given too much weight to precise
syntactical labels that would have meant nothing to users of the lan-
guage; see for instance the note on donec firmaret (3. 5. 45–6). More
importantly, Horace can say things where the superficial meaning is not
the real point: thus at 3. 2. 26 f. ‘betraying the mysteries’ seems to refer
primarily to state secrets; at 3. 14. 27 f., when the poet mentions his hot-
headedness consule Planco, he is referring not so much to his pursuit of
women as to his youthful bravado in joining Brutus’ army.
Such ambiguities present no problem, but it is another matter when
critics tell us that all language is ambiguous and may contain the seeds
of its own contradiction; communication between sensible people usu-
ally works better than that. Of course there may be special problems in
interpreting poets, who sometimes extend normal usage; but though
Virgil’s expressions not uncommonly have a penumbra that is hard to
analyse, Horace is usually more straightforward. Some recent critics
have been too ready to imagine implausible layers of meaning (we quote
a few examples, which could easily be multiplied, at the end of the
introduction to 3. 13). Of course we ourselves may sometimes have made
the wrong choice, as at 3. 15. 4 where we prefer one interpretation of
propior to several others; if we are wrong we can be refuted by lexico-
graphical analysis, but it is no solution to say that incompatible inter-
pretations are equally valid. When the issue is related to the poem’s
central meaning, such uncertainty is more troublesome (as at 3. 26. 11 f.,
3. 27. 69 ff.), but a poet who in principle (ars 448 f.) and practice is
generally clear cannot intend to confuse us; if we have misread the clues,
we are ready to believe that it is our fault rather than his.
6
We take comfort from the reservations of M. Silk (ap. Harrison, 2001: 26 ff.), himself
a modern theorist from whom we are prepared to learn.
xxvi H O R AC E : O D ES I I I
10. Person and persona
Fraenkel thought that Horace never lied (1957: 200 and 260), but many
now go to the opposite extreme and assume that, while the poet plays
various roles, the man remains invisible. In fact discrimination is neces-
sary. It is usually possible to distinguish statements where invention
would have been pointless and in principle detectable from poetical
fantasies (2. 19. 1 ff., 3. 4. 9 ff., 3. 25. 1 ff.); we must not be tempted into
the absurd scepticism of those who regard Ovid’s exile as a poetic fiction.
Moreover, while Horace undoubtedly adopts various personae, for
example that of a simple fellow eager for advice about the law (serm.
2. 1. 1–23) or wine (serm. 2. 4. 1–10), or of one who has much in common
with the country mouse (serm. 2. 6), this device does not prevent us from
drawing valid conclusions about his views and character.7 When he
presents himself in several genres as sceptical and pleasure-loving, and
at the same time hot-tempered (3. 9. 22–3 n.), it is hard to believe that he is
making it all up. We do, however, have to make a sensible allowance for
tactical self-depreciation (‘irony’ in Aristotle’s sense). Was the young man
quite so overwhelmed at his interview with Maecenas (serm. 1. 6. 56 f.)?
And later so utterly ignorant about matters discussed at the dinner-tables
of the great (serm. 2. 6. 51 ff.)? Such passages may lead us to underrate his
position on the fringe of the Augustan court.
11. Genre
We include only points relevant to this commentary. Genres (e.g. lyric)
and their subdivisions (e.g. paean) derive originally from their social
function, and each had its appropriate style and topics; yet even in early
Greece the drinking-songs of Alcaeus and the love-poems of Sappho
need not be exactly what they profess. Hellenistic and Roman poetry
was written primarily for a reading public and was not limited by the
requirements of any particular performance; Horace can allude to the
formulae of hymns (11, 13, 18, 21) or dedications (22, 26) without
following them precisely. Genres in the strict sense should be distin-
guished from the situation-poems analysed by Cairns (1972), which can
cut across generic categories; thus a propempticon or ‘sending-off ’
poem can be found in lyric, bucolic, and elegy, to say nothing of
elements already present in Homer (see the introduction to 27). Cairns
makes use of the rhetorical treatises of Menander Rhetor of about ad
300 (edited by D. A. Russell and N. G. Wilson, 1981); although these
gave prescriptions for ceremonial prose orations (real or artificial) and
employed technical terms that need not all have been familiar to the
Augustans, they can be relevant to our purpose as they draw on topics of
7
See N. Rudd, Lines of Enquiry, 1976, 167 ff.
G EN ER A L I N T RO D U C T I O N xxvii
early Greek poetry. Wilhelm Kroll identified the ‘crossing of the genres’
as a feature of Hellenistic and Roman poetry (1924: 202 ff., cf.
A. Barchiesi ap. Harrison, 2001: 142 ff.); thus 3. 11 includes a hymn, a
poem of courtship, and a mythological narrative, and in 3. 14 the
celebration of Augustus’ adventus is followed by another type of situ-
ation-poem, the preparation for a symposium. Though the Augustans
were obviously influenced by these various conventions, great flexibility
was possible; the poet was always in charge, and we should not judge the
quality of a poem by its correspondence to some preconceived form (cf.
Griffin, 1985: 48 ff.). For further discussion see L. E. Rossi, BICS 18, 1971:
69 ff., G. B. Conte, Genres and Readers, 1994 and OCD 630 f., Lyne
(1995), 59 ff., A. Barchiesi, Encicl. oraz. 2. 35 ff.
12. Style
The style of the Odes (more than that of any other Latin poetry) makes
them the despair of a translator; it is also the hardest feature to character-
ize. A few generalizations are offered in N–H vol. 1, p. xxii, emphasizing
his incisiveness, artificial constructions (including Graecisms), and his
‘unpoetical’ vocabulary (for a clarification of Axelson’s term see 3. 5. 53–
4 n.). In N–H vol. 2 more attention was paid to word-play, partly under
the influence of D. West ap. Costa (1973), 40 ff. In this volume we say
rather more about word-order and the emphasis given by hyperbaton and
position (see further Nisbet ap. Adams–Mayer, 1999: 135 ff.); in these
matters we have derived some benefit from the neglected commentary
of H. Darnley Naylor (1922). We call particular attention to the remarks
on stylistic register made in Adams–Mayer both by the editors (3 ff.) and
by R. G. G. Coleman (21 ff.); they underline the ambiguity of ‘prosaic’
where a term like ‘neutral’ would often be more appropriate. On syntax we
refer not only to the standard works of Kühner and Hofmann but to the
article by Frances Muecke in Encicl. oraz. 2. 755 ff.
13. Structure
Every ode is carefully organized, so that as a rule nothing could be taken
away without impairing the whole. The opening tends to be arresting
(not least in the first poem), and may pose a puzzle that needs to be
resolved (8, 16, 19, 21); apart from the Roman Odes and some other
exceptions most poems are addressed to a real or imaginary person, a
god (4, 18, 22), or a sacred object (the lyre in 11, the spring in 13, the wine
jar in 21), so that they often begin with a question (7, 8, 20, 25) or an
imperative (4, 11, 15, 21). Closures are no less important,8 and may
8
For closure see Barbara H. Smith, Poetic Closure (1968) on English literature, D. Esser
(1976), Deborah H. Roberts, F. M. Dunn, D. Fowler (edd.), Classical Closure (1997).
xxviii H O R AC E : O D ES I I I
contain a trenchant aphorism (2, 6, 16, 24), a ‘breaking-off formula’ (3)
or a diminuendo (5, 28), or a reference to the poet himself (1, 13, 25, 29,
30) which is sometimes self-depreciating (10, 14, 19). The centre of the
poem can act as a pivot; see the introductions to 8 and 14. Sometimes
there is an element of ‘ring-composition’ (13, 25), sometimes the ode
ends far from where it began, describing, in A. Y. Campbell’s phrase, a
parabola rather than a circle (14, 19). Sometimes the second part con-
tains a narrative that may include a speech (3, 11, 27). Again, an ode may
break neatly into pairs of stanzas (1, 9), or it may, for a deliberate effect,
sweep along with few pauses at the end of the line (25). For details
we refer to the introductions to the separate odes; see further N. E.
Collinge (1961), Syndikus ap. Harrison (1995), 17 ff., Y. Nadeau, Coll.
Latomus 266, 2002: 362 f. (with references to earlier articles).
14. The arrangement of the book

Much attention has been paid to this subject, notably by Santirocco
(1986) and Porter (1987). The Roman Odes (1–6) obviously form a
group, though it is less integrated than is sometimes supposed. 3. 1
with its Epicureanism and its possible hints at Maecenas is balanced
to some extent by 3. 29, the great ode to Maecenas that precedes the
personal epilogue. Maecenas is also addressed at the beginning of the
second half of the book (16); for the significance of this position cf. 1. 20
and 4. 8 (which though it has no reference to Maecenas is written in the
same metre as 1. 1 and 3. 30), epod. 9, serm. 1. 6, and G. B. Conte, YCS 29,
1992: 147 ff. When he came to arrange the book Horace no doubt put
some poems together because they invited comparison or contrast: the
light-hearted warning to Asterie (7) comes after the stern denunciation
in 6; 22 and 23 both deal with rustic offerings; it may be significant that
26 (fifth last in the collection) balances 1. 5, just as 30 balances 1. 1. But
those who look for significance in every juxtaposition, and discern
complicated sequences and cycles, forget that such a work would have
been impossible to organize (Nisbet, 1995: 423 f.). We do not believe that
the comparison of the poet to the Adriatic (9. 22 ff.) is linked to the bad
weather of 10. 3 ff., or that the stern father of 11. 45 balances the stern
uncle of 12. 3, or that the boar-hunt of 12. 12 leads to the sacrifice of the
kid in 13. 3. It has been suggested that when he came to prepare the book
for publication Horace made various changes here and there to produce
the kind of relationships envisaged; but, except perhaps for the opening
of 3. 1, there is no evidence to support such a theory, and it is incredible
that he would have tampered with carefully organized poems to achieve
such trivial results.
G EN ER A L I N T RO D U C T I O N xxix
15. The text
Horace’s manuscripts are all medieval, with none earlier than the ninth
century; for some details see Brink, The Ars Poetica, 1971: 1 ff., R. J.
Tarrant in Texts and Transmission (ed. L. D. Reynolds), 1983: 182 ff.,
C. Questa in Encicl. oraz. 1. 319 ff. They divide roughly into two
families, called Æ and in the Oxford text of Wickham and Garrod
(1912), ˛ and in the Teubner texts of Klingner (1959) and Borzsák
(1984); Shackleton Bailey (Teubner, 1985 etc.) presents the evidence
most clearly by keeping the symbol and breaking the ˛ group into
its components. Well-attested readings presumably go back to the
ancient world and so are seldom complete nonsense (contrast the text
of Catullus); this may be a snare for those whose only method is to
follow the manuscripts through thick and thin. When the transmitted
reading causes any doubt, conjectures may be considered; it is irrational
to suppose that a conjecture should not be mentioned unless it is certain.
Of recent editors Borzsák cites insignificant variants but is reluctant to
take conjectures seriously; Shackleton Bailey offers interesting new
proposals but is too ready to print them in his text. In the present
volume the lemmata provide a continuous text, though without an
apparatus or testimonia (for the latter see Klingner or Borzsák). We
have put only two of RN’s conjectures in the lemmata (3. 1. 42 Sidone,
3. 26. 6 lurida), but in the notes have suggested some thirty others with
varying degrees of confidence or diffidence.
16. The ancient commentators

The most important is Pomponius Porphyrio (ed. A. Holder, 1894, repr.
1967), who probably lived in the third century; the commentary of the
so-called ‘pseudo-Acro’ (ed. O. Keller, 1902–4) is a compilation by
various hands that goes back to the fifth century; the ‘commentator
Cruquianus’ is an amalgam printed by Cruquius in his edition of 1611
and contains little of significance that is not found elsewhere. Porphyrio
is much more reliable than ps.-Acro, but even he combines basic infor-
mation (notably on prosopography) with curious misapprehensions. In
textual criticism the scholiasts’ comments deserve more attention than
their lemmata, which usually repeat what is provided elsewhere. For
further detail see N–H vol. 1, pp. xlvii ff., Brink, op. cit. 33f., 38 ff., Encicl.
oraz. 3. 17 ff. and 3. 695 ff. (a complete text of Porph. and ps.-Acro),
S. Diederich, Der Horazcommentar des Porphyrio im Rahmen der kaiser-
zeitlichen Schul- und Bildungstradition, 1999 (reviewed by L. Holford-
Strevens, Gnomon 74, 2002: 501 ff.).
Commentary
on
Horace
Odes, Book III
1 . O D I P RO FA N V M V V LG V S
[K. Barwick, RhM 93, 1950: 259 ff.; F. Cairns, PLLS 8, 1995: 91 ff.; Fraenkel 261 ff.;
W. D. Lebek, ANRW II. 31. 3 (1981), 2065 ff.; Lyne (1995), 158 ff.; G. J. Mader, Acta Classica
(South Africa) 30, 1987: 11 ff.; Pasquali 649 ff., 833 ff.; V. Pöschl, HSCP 63, 1958: 333 ff. ¼
Horazische Lyrik edn. 2, 1991: 144 ff.; E. T. Silk, YClS 13, 1952: 145 ff. and 23, 1973: 131 ff.;
F. Solmsen, AJP 68, 1947: 337 ff. ¼ Oppermann 139 ff.; T. Woodman ap. Woodman and
West (1984), 83 ff.]
1–8. Let the uninitiated depart; I am teaching new chants to a fresh

generation. Know that even dread kings must fear the rule of Jove. 9–16.
Men compete in landed wealth and political advantages, but high and low
alike are subject to Fate. 17–24. The overweening cannot enjoy their luxury
or be lulled to sleep even by music, but sleep does not disdain lowly dwellings
and a shady valley. 25–32. The man of limited desires is undisturbed by the
bad weather that harasses the acquisitive merchant and the dissatisfied
landowner. 33–40. The arrogant encroach on the sea with unnatural con-
structions, but anxieties pursue them even there and cannot be escaped on sea
or land. 41–8. So, since mental pains are not assuaged by exotic luxuries, why
should I rear a grandiose edifice or exchange my Sabine valley for the troubles
of wealth?
The ode opens with an arresting scene: the poet, as priest of the
Muses, bids outsiders depart (1 n.), asks for silence (2 n.), and sings his
new song to the boys and girls who may be thought receptive (4 n.). His
proclamation is earnest and uncompromising: dreaded kings, for all
their earthly power, are subject to divine law (5–8). This thought is
given a particular application in the next two stanzas (9–16), though
with less solemnity: in spite of men’s various ambitions (here described
with some satire), everybody dies in the end. The reader is expected to
keep this idea in mind for the remainder of the poem. There, in a series
of vignettes, the evils of riches are contrasted with the blessings of
simplicity; but except in the first and most extreme example (that of
the tyrant), there is no talk of actual impiety. The emphasis is on private
happiness and how it is threatened by the anxieties of wealth.
When we consider the appropriateness of the ode as an introduction
to a political series, we are confronted with an awkward fact: its prov-
enance is predominantly Epicurean (Pasquali, Pöschl, locc. citt., Lyne,
1995: 162 f.), notwithstanding the different tone and content of the first
two stanzas (Lebek, op. cit. overstresses the non-Epicurean elements).
In particular the poem is strongly influenced by the proem of Lucretius
2 (a passage already imitated by Horace in 2. 16). In Horace as in his
4 H O R AC E : O D ES I I I
exemplar the shortness of life makes nonsense of our strenuous rivalries,
we are often anxious about matters of minor importance, self-
indulgence and ostentation do nothing to assuage our fears, happiness
is to be found in the simple life and the acceptance of what is there.
Many of the details, too, have counterparts in Lucretius, though the
arrangement of the original has been changed (again as in 2. 16): e.g.
competition for status (10 ff., cf. Lucr. 2. 11 ff.), elaborate banquets (17 ff.,
cf. 2. 23 ff.), the sound of music (20, cf. 2. 28), the repose of the poor
(21 ff., cf. 2. 36 ff.), the shady bank (23 f., cf. 2. 29 ff.), mankind’s limited
needs (25, cf. 2. 20 f.), anxiety that cannot be shaken off (40, cf. 2. 48),
useless purple (42, cf. 2. 35, 52), the concluding quodsi (41, cf. 2. 47). The
central doctrine obviously comes from Epicurus himself: P ºØ c B
łı
B ÆæÆ
c Pb c IØ ºª Iªfi A
Ææa h ºF
æ
ø › ªØ hŁ Ææa E ººE Øc ŒÆd æºłØ h
¼ºº Ø H Ææa a IØæı ÆNÆ (sent. Vat. 81, cf. 2. 16. 9 with
N–H). In 3. 29, the corresponding poem before the epilogue of the
whole collection, the Epicurean element is equally clear: there as here
the luxury and anxieties of the city are contrasted with simple meals in
the country ‘sine aulaeis et ostro’ (15).
These Epicurean elements can to some extent be reconciled with
traditional Roman attitudes. Criticisms of avaritia and luxuria had
been popular since Cato with moralizing orators, and had recently
found forceful expression in the monographs of Sallust. Augustan
ideology pointed in the same direction: aristocratic ostentation made
for disharmony, and the rivalries of selfish and ambitious men were no
longer encouraged; contentment and acceptance made for peace and
stability. A simple life-style was commended by the Princeps himself:
one thinks of his Palatine house as described by Suetonius (Aug. 72,
doubtless exaggerating its austerity); and the passages on his clothes,
furniture, and diet are no less relevant (ibid. 73–4). Yet the fact remains
that in our poem Horace rejects luxury because it does not lead to
happiness, not because it is socially and politically unacceptable (Lyne,
1995: 162 f.); contrast 2. 15, which concentrates on ancestral norms rather
than Epicurean precepts. While it would be absurd to suppose that he is
undermining the very system that he professes to support (even Lyne’s
phrase ‘benignly subversive’ goes too far), it is true that by adopting a
predominantly private stance, Horace has written a poem which is less
overtly patriotic than the other Roman Odes. There is some reason in
the conjecture that the portentous opening has been grafted onto a
more personal piece to serve as an introduction to the series.
We have spoken so far of Lucretius, but Horace was also influenced
by the end of the recently issued second Georgic, which in the same
philosophical tradition had drawn a contrast between the happiness of
farmers and the pomp of the rich (2. 461 ff.):
1 . O D I P RO FA N V M V V L G V S 5
si non ingentem foribus domus alta superbis
mane salutantum totis vomit aedibus undam,
nec varios inhiant pulchra testudine postis
inlusasque auro vestis Ephyreiaque aera,
alba neque Assyrio fucatur lana veneno,
nec casia liquidi corrumpitur usus olivi . . .
Here as in the ode we have a domus alta (46 n.) with ornamental doors
(45 n.) and a crowd of clients (13 n.), a mention of exotic dyes (42 n.), and
an instance of the periphrastic usus. Virgil then goes on to speak of
tempe (469, cf. 24 n.), ‘mollesque sub arbore somni’ (470, cf. 21 n.), the
countryman’s lack of envy (498 f.), and the misguided ambitions of
political life (508 ff., cf. 10 f.); for further details see B. Fenik, Hermes
90, 1962: 72 ff. Perhaps Horace took a hint from his friend about how to
conform with the current ideology without compromising his individu-
alistic standpoint. But it is noticeable that whereas Virgil, like Lucre-
tius, tells us little about his own way of life, the ode ends with the
familiar picture of Horace on his Sabine estate.
The mention of Horace’s Sabine valley in the last stanza encourages
us to see an expression of gratitude to Maecenas (see Cairns, 1972: 74 f.
on the eucharisticon); for similar acknowledgements in opening poems,
cf. 1. 1. 35 f., epod. 1. 31 f. By the same token, sublime atrium (46) may
include Maecenas’ Esquiline palace; cf. 3. 29. 10, epod. 9. 3, serm. 2. 6. 102
(the residence of the town mouse). RN detects further hints of Maece-
nas’ life-style (for which see Mayor on Juv. 1. 66), comparing 2. 18,
where again the great man is not actually identified as the householder
(see N–H ad loc., Lyne, 1995: 126 ff.); he would cite the references at v. 9
to extensive estates (cf. 3. 16. 41 n.), at v. 13 to crowds of clients (cf. 2. 18. 8
with N–H), at vv. 20 f. to music as a treatment for insomnia (see note ad
loc.), at vv. 33 ff. to a villa maritima (also perhaps at 2. 18. 20 ff.), at v. 41
to exotic marble (cf. 2. 18. 3), at v. 42 to purple fabrics (cf. N–H on
2. 18. 8), at v. 44 to royal perfumes (cf. perhaps 3. 29. 4). When Care rides
behind the eques (40), RN thinks of the equestrian Maecenas and his
neurotic obsessions; Maecenas may also have owned a ‘private trireme’
like that mentioned in v. 39 (see note). While it would be absurd to
suppose that Horace is sneering, he can suggest that his benefactor has
given him a happier life than his vast wealth has secured for himself.
At 2. 18. 17 NR sees the objectionable tu as quite different from
Maecenas (the potens amicus of v. 12); tall houses and broad acres, like
clients, music and purple, were among the appurtenances of any rich
man and belonged to the tradition of the diatribe (see Lucretius and
Virgil above). There is no proof that Maecenas built over the sea or
owned a private trireme; most important, Horace would never have
suggested that he was greedy or ambitious. So if he is present, he is very
much in the background. What we can say is that, whatever Horace’s
intentions were, and whatever innuendoes were perceived by his en-
emies, such correspondences did not cause serious offence (see epist. 1. 1.
1 ff.), and that they should not be given undue prominence in a poem
which derives its weight from the generality of its truths.
The stanzas of 3. 1 break into groups of two, as in the similar 2. 16
(Barwick, op. cit.); the articulation is underlined by the enjambment at
the end of the odd-numbered stanzas (except the first). Each group
(after the first) implies a different sphere of rivalry and discontent: the
ownership of estates and and the pursuit of political power (3 and 4), the
possession of luxuries (5 and 6), dissatisfaction with available sufficiency
(7 and 8), pretentious and unnatural building (9 and 10), followed by the
personal conclusion ‘if all this restless striving is not satisfied by material
luxuries, why should I leave my Sabine valley?’ (11 and 12). The two
halves of the poem end with pictures of rural serenity: the first (stanza 6)
expressed in general terms, the second (stanza 12) centred on the poet
himself (tempe in v. 24 is picked up by valle in 47). There are no
connections between the pairs, except for the final quodsi at v. 47. The
symmetry of the structure is not easily paralleled in Greek, and Syndi-
kus (2. 12) is right to distinguish the abrupt transitions of Pindar. One
might compare the sententiae of declamation which often repeat the
same point but by ingenious restatement give the appearance of pro-
gression. But here the argument is conducted less by aphorisms than by
a series of representative vignettes, which are mainly drawn from the
contemporary Roman scene—a technique that is very typical of the
Satires. Though the style is dry and rather formal, it is less solemn than
is sometimes implied (cf. 9 n., 33, 35 ff.). And though Horace begins
with a hieratic pronouncement, when he comes back to himself at the
end his tone is human and personal.
Metre: Alcaic.
1. Odi profanum vulgus et arceo: odi (‘I shun’) balances arceo (‘I keep it
at a distance’); here the former verb emphasizes overt rejection, though
emotional dislike is not excluded (cf. 1. 38. 1 ‘Persicos odi, puer, appar-
atus’, epist. 1. 7. 20 ‘spernit et odit’, ars 188, Fraenkel 263). profanum may
mean not ‘in front of the temple’ (Varro, lL 6. 54) but ‘away from the
temple’ (Charis. gramm. p. 305. 20 Barwick ‘porro a fano positus’,
Cairns, op. cit. 94); it is used in religious contexts of the uninitiated
(Catull. 64. 260, Theoc. 26. 13 f. ZæªØÆ BŒ
ø = . . . P
›æØ
ÆºØ) or those not participating in a rite. For the sacral arceo cf.
Pacuv. TRF 304 f. ‘quamquam aetas senet, satis habeam virium ut te ara
arceam’, Ogilvie on Liv. 1. 12. 4. Ancient religion had a strong feeling
for the sacred space; cf. Ar. Ach. 239 ff., ran. 369 f., Eur. Bacch. 68 ff.
›fiH ›fiH; ; = ºŁæØ Œ ø, Æ h- = –Æ
KØŁø (with Dodds), Call. h.2.2 Œa Œa , ‹Ø IºØæ , Virg.
Aen. 6. 258 ‘procul o procul este, profani’, Paul. Fest. 72L ¼ 82M ‘exesto,
extra esto. sic enim lictor in quibusdam sacris clamitabat: hostis, vinctus,
mulier, virgo exesto; scilicet interesse prohibebatur’, O Weinreich, Aus-
gewählte Schriften 2 (1973), 386 f. For the application of the theme to
literature cf. Ar. ran. 354 ff. PE
æc ŒIÆŁÆØ E æØØ
æEØ = ‹Ø ¼Øæ ØH º ªø j ªfi c ŒÆŁÆæØ = j

ªÆø ZæªØÆ MıH r K
æı, Gell. praef. 20 f.
A real Roman priest did not speak for himself (Fraenkel 264) or make
portentous announcements (5 ff.), and already Porph. sees a metaphor
from the mysteries; for the pattern cf. the Orphic testamenta fr. 247.
1 Kern ŁªÆØ x ŁØ K, ŁæÆ KŁŁ, ºØ. Porph.
interprets the metaphor by indoctos and Musarum profanos, and no
doubt H has been influenced by Callimachean manifestos about poetry;
cf. especially ep. 28. 4 ØŒ
Æø Æ a ØÆ, aet. fr. 1. 17 ºº
BÆŒÆ Oºe ª , S. J. Heyworth, MD 33, 1994: 54 ff. But here the
religious metaphor refers primarily to content rather than literary style;
cf. Hippocrates, lex 5 a b ƒæa K Æ æªÆÆ ƒæEØ IŁæØØ
ŒıÆØ, ºØØ b P ŁØ æd j ºŁHØ OæªØØ KØ ,
Plat. symp. 218b. H exhorts the new generation not to be too impressed
by wealth and power (cf. epist. 2. 1. 119–38) and to make time for things
that matter more; similarly in the ensuing odes he emphasizes such
virtues as courage, piety, and chastity. The upper-class girls and boys
who are contrasted with the profanum vulgus are chosen primarily for
their untainted idealism, though NR thinks they may also be expected
eventually to regard Horace’s new poetry with less prejudice than their
elders (epist. 2. 1. 18 ff.).
2. favete linguis: ‘hold your peace’ (for the short opening syllable cf.
tumultuosum in v. 26, N–H vol. 1, p. xl); for sacred silence cf. 2. 13. 29
with N–H, 3. 14. 11 f., 3. 30. 9. The religious formula originally meant
‘make favourable utterance’ (hence the instrumental linguis), but the
safest way of avoiding ill-omened words was to say nothing; cf. Serv.
auct. Aen. 5. 71 ‘praeco magistratu sacrificante dicebat favete linguis,
favete vocibus, hoc est bona omina habete aut tacete’, Pease on Cic.
div. 1. 102, Bömer on Ov. fast. 1. 71, Courtney on Juv. 12. 83. The Greek
PE changed in the same way (F. Williams on Call. h. 2. 17).
The request was made after the exclusion of the profani, as at Eur. Bacch.
68 ff. cited in 1n. above; it preceded a sacred song, as at Ar. Thesm. 39 f.
For the application of the motif to literature cf. Ar. ran. 354, cited above
in 1 n., Prop. 4. 6. 1 ‘sacra facit vates, sint ora faventia sacris’.
2–3. carmina non prius / audita: the plural refers to the Roman Odes as
a whole; in the religious context carmina suggests sacred chants, and the
assonance of carmina . . . canto suits the sacral style (cf. 4. 15. 30 ff.
‘carmine . . . canemus’). Roman poets, like their Greek predecessors
(both early and Hellenistic), often lay claim to originality (3. 25. 7 f.
‘adhuc / indictum ore alio’, 3. 30. 13 n., N–H on 1. 26. 10 and 2. 20. 1);
Porph. should not have restricted the issue to Latin lyric, as there is
nothing like these poems in Alcaeus, and even an early piece like 1. 2
cannot equal the combined authority of the Roman Odes. Newness was
also emphasized in Bacchic and other mysteries, as later in the world
of the New Testament (E. Norden, Die Geburt des Kindes, 1924: 46 f.);
for other literary adaptations of this idea cf. Prop. 3. 1. 3 f. ‘primus
ego ingredior puro de fonte sacerdos / Itala per Graios orgia ferre
choros’.
3. Musarum sacerdos: a poet could be compared to a priest for a variety

of reasons (for some references see Kroll 24 ff., O. Falter, Der Dichter
und sein Gott bei den Griechen und Römern, diss. Würzburg 1934, Kam-
bylis 12 ff., Newman 99 ff., N–H on 1. 31. 2). In early Greece he might be
called the æ or interpreter of the Muses, because like the
Delphic priestess he imposed form on material that seemed to present
itself spontaneously (Pind. fr. 150 Æ, MEÆ, æÆø Kª).
Hellenistic poets and their Roman imitators underlined the arcane
character of their art, their awareness of their high calling, and the
splendour and formality of their procedures. Horace uses vates for the
inspired bard (N–H on 1. 1. 35), sometimes ironically (e.g. epist. 2. 2. 102
‘genus irritabile vatum’); sacerdos on the other hand emphasizes the
authority and dignity of the poet’s pronouncements (for the distinction
cf. Lyne, 1995: 184 f.). At Rome there was an aedes Herculis Musarum, but
the Muses had no independent priesthood (contrast Greece) and
Horace’s reference is entirely literary (N. Horsfall, BICS 23, 1976: 85).
4. virginibus puerisque canto: the collocation of virgines and pueri

seems to suggest that H is training a choir; cf. 1. 21. 1 ff., 4. 6. 31 f.,
carm. saec. 6, epist. 2. 1. 132 ff. ‘castis cum pueris ignara puella mariti /
disceret unde preces vatem ni Musa dedisset?’, Cairns, op. cit. 102 ff.
That in no way implies that the Roman Odes were literally sung; the
motif of a choir is not sustained (as Cairns claims), for it does not suit
the content and manner of the poem or the series (Pasquali 650). H’s
message is directed to a sinless new generation, including future wives
and mothers, because they are still open to moral instruction; cf. Porph.
ad loc. ‘dicit se carmen proditurum quod teneras aetates ad utilia
instituat, quibus ad beatam vitam pervenire possint’, epist. 2. 1. 128 ff.
Later poets echo H’s line simply to suggest an absence of impropriety
(Ov. trist. 2. 370, Mart. 3. 69. 8).
5–6. regum timendorum in proprios greges, / reges in ipsos imperium
est Iovis: the sacerdos now makes his proclamation (ŒæıªÆ). Just as
kings rule their subjects, so Jupiter rules kings: the emphasis lies on the
second clause (cf. 9–16 n.). In the same way the Orphic testamenta, after
dismissing the profane (1 n.), proceed to the praises of the supreme god
(245. 9 f. Kern). H is not just paying the conventional tribute of an
exordium to Jupiter (N–H on 1. 12. 13, Pasquali 654); it is central to his
meaning (here expressed in Stoic terms) that even the rich and powerful
are subject to the laws of the universe.
The kings are Eastern rulers who, because of their total and arbitrary
power, are dreaded by their own subjects (Tarrant on Sen. Ag. 72 f.), but
must dread Jupiter in turn. H is not arguing that the rulers are God’s
representatives on earth (Call. h 1. 79, N–H on 1. 12. 50 ff., cf. 3. 6. 5), and
he would not have regarded Augustus as either a rex or as timendus in
the sense required here (otherwise D. P. Fowler ap. Harrison, 1995: 264);
cf. rather 1. 35. 12 (to Fortune) ‘(te) purpurei metuunt tyranni’, Philemon
31 (PCG 7, p. 244), 4 f. FºØ ÆØºø N, › ÆØºf ŁH, = › Łe
IªŒ , Manil. 4. 93, Sen.Thy. 612 ‘omne sub regno graviore regnum
est’. For H’s use of in cf. 4. 4. 2 ‘regnum in aves’, Plaut. Pers. 343, Ter.
eun. 415. proprios implies ownership, but its main function is to underline
the limits of a king’s imperium; Jupiter on the other hand rules over
everything (cuncta in v. 8). greges is as often disparaging: a good king may
be the shepherd of his people (Hom. Il. 2. 243 etc.), but it is another
matter to call the people the grex of the king.
7. clari Giganteo triumpho: the Giants in their arrogance would not

submit to the moral order of the universe (Cic. sen. 5 ‘quid est aliud
Gigantum modo bellare cum dis nisi naturae repugnare?’), but in spite
of their strength they were subject to the limitations of the earth-born
(Philodemus, de morte 4. 37 A ¼Łæø , Œi N
ıæ æ fi H
ˆØªø, Kæ KØ æe øc ŒÆd ºı). Elsewhere their
rebelliousness is associated with resistance to earthly rulers, notably
Augustus (2. 12. 6 f. with N–H, more explicitly 3. 4. 53 ff.).
8. cuncta supercilio moventis: the phrase represents a variation of an

epic commonplace; cf. Hom. Il. 1. 528 ff. q ŒÆd ŒıÆfi Ø K OæØ F
˚æø = . . . ªÆ KººØ ! ˇºı, Catull. 64. 204 ff. (numine),
Virg. Aen. 9. 106 (‘adnuit et totum nutu tremefecit Olympum’). In our
passage the gesture is not just a sign of assent but indicates Jupiter’s
power to move the cosmos with the minimum effort; cf. Dio Chrys. 12.
25, quoting Homer, loc. cit. F ØÆ Oºªfiø ÆØ H Oæø
e ÆÆ ! ˇºı. For movet cf. Plaut. rud. 1 ‘qui gentes omnes
mariaque et terras movet’, Ov. met. 1. 180, TLL 8. 1544. 53 ff. For other
comments about the use of the eyebrows see Plin. nat. hist. 11. 138,
Quint. 11. 3. 79, Sittl 92.
9–16. est ut viro vir latius ordinet / arbusta sulcis . . . : these two stanzas
have a bipartite structure (Mader, op. cit. 13): (a) men strive for superior-
ity in various fields (9–14, the so-called ‘preamble’or ‘foil’), (b) but all
men are mortal (14–16, the climax or ‘apex’ of the argument); for similar
patterns cf. 1. 1. 3 ff. (with N–H p. 2), 1. 7. 1 ff., 3. 27. 1–2 n. The
conclusion follows naturally from the second stanza, though it is now
expressed in specifically Roman terms: all men, however successful, are
subject to Jupiter or Fate. Woodman maintains that the lives referred
to in the preamble are all meritorious and provide a foil to the impius of
v. 17 (op. cit. 92 f.); on the contrary, they are by implication over-
competitive (Mader, op. cit. 15 f.) and lead up to the criticisms of
greed and ambition that occupy the latter part of the poem.
The periphrastic est ut means ‘it happens that’ (K–S 2. 237); the
formula binds the four clauses together and sets them against the
inevitability of death (Heinze). The polyptoton viro vir (i.e. the repeti-
tion of a word in a different case) recalls sardonically the combats of
epic; cf. Hom. Il. 16. 215, Furius Bibaculus, FLP 10 ‘pressatur pede pes,
mucro mucrone, viro vir’, Virg. Aen. 10. 361 etc., Wills 195 ff. arbusta was
particularly applied to plantations of trees on which vines were trained
(Virg. georg. 2. 416, OLD s.v. 2). Such trees were arranged in neat ranks
(Cic. sen. 59, Varr. rust. 1. 7. 2, Colum. 3. 13. 5 ‘ordinent vineam paribus
intervallis’); ordinent also suits an army on parade (cf. Virg. georg. 2.
279 ff.), and so sustains the heroic tone of viro vir. latius means ‘over a
wider area’ (cf. 2. 2. 9 ‘latius regnes’), not ‘at greater intervals’ (Virg.
georg. 2. 277 ‘indulge ordinibus’); H is critical of latifundia, and viticul-
ture was particularly profitable.
The sulci are the trenches in which the supporting trees were planted
(Colum. 5. 6. 10 ‘sulci . . . qui arbores recipiant praeparandi’); shallower
furrows were dug for the vines (Cato, agr. 49. 2, Virg. georg. 2. 289,
Colum. 3. 13. 5), but it was the former that gave the plantation its
pattern. sulcis is generally taken as a local ablative, but we are inclined
to see it as instrumental (‘arrays plantations with trenches’); cf. Varr.
rust. 3. 5. 11 ‘porticus . . . arbusculis humilibus ordinatae’, Colum. 5. 3. 7
‘vitibus locum . . . ordinare’, Mart. 3. 58. 2 ‘otiosis ordinata myrtetis’ (of a
villa). sulcis belongs to the laborious side of country life, like ‘hedging
and ditching’ (cf. epist. 1. 7. 84 ‘sulcos et vineta crepat mera’); as such it is
set against the pretensions of latius ordinet.
10–11. hic generosior / descendat in Campum petitor: the verb suits a

candidate going down from the hills of Rome, where the rich lived, to
the Campus Martius, where the comitia centuriata elected consuls and
praetors; cf. Val. Max. 7. 5. 1 ‘cum . . . candidatus in Campum descendis-
set’, Vell. 2. 92. 3. But it also suggests ‘entering the fray’ (ŒÆÆÆØ)
whether for athletic (Cic. Tusc. 2. 62) or military honours (Suet. Aug.
96. 2); similarly campum as a common noun suits a battlefield (Liv.
10. 27. 7). petitor is regularly used of a candidate for office, but as it
literally means ‘petitioner’ it undercuts generosior; cf. the effect of sulcis
above. There was still a certain amount of freedom at elections, though
less at consular level (epist. 1. 1. 43, 1. 6. 49 ff., A. H. M. Jones, JRS 45,
1955: 9 ff.).
12–13. moribus hic meliorque fama / contendat: genus and mores are
often contrasted (e.g. Juv. 8. 20 f.), and novi homines opposed their virtus
to the claims of birth (Woodman, op. cit. 86; so in a different context
epist. 1. 20. 22). After petitor the verb contendat implies electoral contests
rather than the more general rivalry mentioned by Lucr. 2. 11 ‘certare
ingenio, contendere nobilitate’. This is the most commendable of the
four types mentioned, but even he exploits his reputation to win polit-
ical advantage.
13–14. illi turba clientium / sit maior: such celebritas was a sign of social
success; cf. Virg. georg. 2. 461 f. quoted in the introduction above, Saller
128 f. Although clients were a practical asset for political candidates
(Cic. Mur. 69–70), there is no need this time to confine the issue to
elections. The four types are arranged in the form of a chiasmus:
landowner (not confined to politics), two types of political candidate,
social celebrity (not confined to politics). turba ‘a mob’ maintains the
pejorative tone.
14–15. aequa lege Necessitas / sortitur insignis et imos: aequa makes a

contrast with the preceding comparatives. The poor do not usually get a
fair deal against the rich (Plaut. cist. 532 f. ‘quando aequa lege pauperi
cum divite / non licet’), but in death they are on an equal footing (see
N–H on 1. 4. 13, and cf. Shirley, The Glories of our Blood and State, 3 ff.
‘Scepter and crown / Must tumble down, / And in the dust be equal
made / With the poor crooked sithe and spade’). Necessitas (#ªŒ)
refers to the laws of the universe and echoes imperium Iovis (6) in more
philosophic terms (Silk, op. cit. 152, Cairns, op. cit. 117 citing Arist. æd
Œ ı 401b 8–9); though the abstraction is more comprehensive than
necessitas leti (1. 3. 32 f.), here too H is concentrating on the inevitability
of death; cf. Philodemus, de morte 4. 38. 12 ff. ¼ø ¼Æ æÆ
ÆŒæa IÆØæ KºÆ e $æ, where e $æ is the phil-
osopher’s substitution for º
Ł %ØÆ in TGF adesp. 127. 10.
Though death is inevitable for all, its timing for each individual is
random, as symbolized by the lot (cf. 2. 3. 26 ff.). The verb, like lege,
sustains the official tone; lot had a limited role in Roman political and
legal life; e.g. in the allocation of public provinces (Dio 53. 13. 2). insignes
et imos is not only a polar expression of the type common in the context
of death (N–H on 1. 4. 13 f. and 2. 14. 11); by means of an ellipse
(¼ ‘insignes et humiles, summos et imos’) it achieves variety and
compression.
16. omne capax movet urna nomen: the urn was shaken; for movere in
this sense cf. serm. 1. 9. 30 (where mota . . . urna is surely right) and Virg.
Aen. 6. 432; eventually a lot jumped out (2. 3. 27 with N–H). omne and
capax both carry weight: there is room for all in the urn (Sen. Hf 191 ), as
in Charon’s boat (ibid. 775) and the underworld itself (ibid. 659). The
line, unusually, has five disyllables, perhaps reflecting the shuffling
motion involved.
17–18. destrictus ensis cui super impia / cervice pendet: according

to the anecdote, Dionysius of Syracuse suspended a sword over the
courtier Damocles to show him that even the most magnificent despots
lived in fear of assassination (Cic. Tusc. 5. 62). Cicero applies the story
to Dionysius the elder (tyrant 405–367), who was notoriously suspi-
cious, though Dionysius II (tyrant 367–357) is mentioned in another
context in conjunction with Damocles (Athen. 6. 250, citing Timaeus).
The sword over Dionysius was a metaphorical one, as opposed to the
real one over Damocles. H has not made a mistake but is generalizing,
as the future elaborabunt shows: ‘Uneasy lies the head that wears
a crown’.
The grandiose ensis is used six times in the odes, but gladius never; the
former is greatly preferred by Virgil and later epic poets, except for
Lucan (Axelson 51, Lyne, 1989: 103 ff.). The strong word impia brings
out the hubris of a tyrant who has defied the laws of God (cf. the Giants
in v. 7); the assassins he fears are the instruments of divine vengeance
(cf. 6). In view of vv. 14 ff. the metaphor is that of impending death
rather than of other anxieties (Pers. 3. 40 ff. refers to a guilty con-
science); the tyrant is the extreme example of the lust for wealth and
power, and his banquets (18 f.) illustrate how, in Epicurean doctrine, the
pleasures of the rich are spoiled by the fear of death; cf. Lucr. 3. 37 ff. ‘et
metus ille foras praeceps Acherontis agendus / funditus humanam qui
vitam turbat ab imo, / omnia suffundens mortis nigrore, neque ullam /
esse voluptatem liquidam puramque relinquit’.
18–19. non Siculae dapes / dulcem elaborabunt saporem: Sicilian

tables (especially those of Syracuse) were proverbial for their luxury;
cf. Ar. PCG 3. 2 fr. 225. 2 &ıæÆŒÆ æÆ, Plat. rep. 3. 404 d, epist.
7. 326b, Cic. Tusc. 5. 100, Athen. 518c, paroem. Gr. 1. 158, Otto 321.
A luxurious banquet was the traditional setting for the story of
Damocles (Cic. Tusc. 5. 62 ‘mensae conquisitissimis epulis exstrueban-
tur’, Amm. Marc. 29. 2. 4, Euseb. praep. evang 8. 14. 29, Sidon. epist. 2.
13. 7); it is relevant that, as noted in the introduction, Augustus encour-
aged frugality. The grandiose dapes can be contrasted with ‘mundae . . .
pauperum / cenae’ (3. 29. 14 f.).
elaborare is rare in the active (cf. Sen. epist. 16. 8 in a similar context),
and would normally be used of the cook rather than the feast; the verb
suits the perfection of an intricate work of art, and is used with irony
of a flavour. For moralists’ criticisms of cookery cf. Plat. Gorg. 462d,
Hor. serm. 2. 2. 19 f., Rudd (1966), 202 ff.
20–1. non avium citharaeque cantus / somnum reducent: avium refers

not to open-air birdsong (as at epod. 2. 26, eleg. in Maec. 1. 36 ‘sederat
argutas garrulus inter aves’) but to the unnatural aviaries of rich men;
cf. Varr. rust. 3. 5. 14 ‘intra retem aves sunt omnigenus, maxime cantrices,
ut lusciniolae ac merulae’, Plin. nat. hist. 10. 141 ‘coepimus carcere
animalia coercere quibus rerum natura caelum adsignaverat’, G. Jenni-
son, Animals for Show and Pleasure in Ancient Rome, 1937: 116 ff. For
music as an aid to sleep cf. epist. 1. 2. 30 f. (of the Phaeacians) ‘cui
pulchrum fuit in medios dormire dies et / ad strepitum citharae
ycessatumy (cessantem cod. det.) ducere somnum’, Sen. dial. 1. 3. 10
‘Maecenatem . . . cui . . . somnus per symphoniarum cantum ex longin-
quo lene resonantium quaeritur’. For insomnia as a literary topic see
further N–H on 2. 11. 8.
21–2. somnus agrestium / lenis virorum: cf. 2. 16. 15 with N–H, epist.
1. 7. 35 ‘nec somnum plebis laudo satur altilium’, Epic. fr. 207 Usener ¼
V48 Bailey ˚æE Ø ŁÆææE Kd Ø ŒÆÆŒØfi j
ÆæŁÆØ
æıB K
fi Œº ŒÆd ºıºB æÆ, Stob. 5. 763
Hense (the rich man) ıŒºøæı
Ø . . . f oı , Ecclesiastes 5:
12 ‘The sleep of a labouring man is sweet, whether he eat little or much:
but the abundance of the rich will not suffer them to sleep.’ For a more
realistic view cf. Theoc. 21. 2 f. Pb ªaæ oØ = IæØ KæªÆÆØØ
ŒÆŒÆd Ææ
Ø æØÆØ.
The function of virorum, which is not necessary with agrestium, is to
underline the toughness of the rustic life; cf. Xen. oec. 5. 4, Cato, agr.
praef. 4 ‘at ex agricolis et viri fortissimi et milites strenuissimi gignun-
tur’, Virg. georg. 2. 167, 472, 531 ff. H is associating himself with the
agrestes viri; cf. serm. 2. 6. 79 ff. (on the Country Mouse), epist. 1. 10. 2.
For lenis cf. Enn. ann. 2 Sk. ‘somno leni placidoque revinctus’, TLL 7. 2.
1144. 51 ff. The contrast between the gentleness of sleep and
the robustness of the countrymen is brought out by the interlacing
word-order.
22–3. non humilis domos / fastidit umbrosamque ripam: non negatives
fastidit, not humilis, as Porph. points out. For the cottages of the just cf.
Aesch. Ag. 774 f., Sen. Thy. 446 ff. with Tarrant. fastidit suggests the
great man’s disdain, which is not shared by sleep. humilis describes literal
lowness as well as modesty; cf. Virg. ecl. 2. 29 ‘humilis habitare casas’.
The countryman can also enjoy a siesta in the open air; cf. epod. 2. 28,
epist. 1. 14. 35, Virg. ecl. 1. 55, georg. 2. 470. H economically suggests
running water and trees; cf. 1. 1. 21 f. with N–H, 3. 29. 22 ff., Sappho 2.
5 ff., Lucr. 2. 30 ‘propter aquae rivum sub ramis arboris altae’. umbrosam
carries associations of leisure (1. 32. 1) and perhaps, in RN’s view, of
obscurity (balancing humilis).
24. non zephyris agitata tempe: the countryman’s valley is disturbed

only by the breeze. The neuter plural tempe, properly the Vale of Peneus
(N–H on 1. 7. 4), came to be used for any wooded valley in later Greek
prose (see LSJ s.v., also Cic. Att. 4. 15. 5); H seems to have been
influenced in particular by Virg. georg. 2. 469 f. ‘frigida tempe / mugi-
tusque boum mollesque sub arbore somni’. The exotic word, here
balancing zephyris, associates the Italian countryside with idealized
Greek landscapes.
25. desiderantem quod satis est: cf. 3. 16. 44, serm. 1. 1. 62, epist. 1. 2. 46
‘quod satis est cui contingit nil amplius optet’, Sen. epist. 119. 7 ‘num-
quam parum est quod satis est’. As H’s countryman is assumed (not
wholly realistically) to have quod satis est, he does not need to worry
about survival. Ancient moralists of various schools preach on this text;
cf. Epic. fr. 473 Usener ¼ V 68 Bailey ˇPb ƒŒÆe fiøffl Oºª e ƒŒÆ ,
G. A. Gerhard, Phoenix von Kolophon, 1909: 56 f., 86 f., Krenkel on
Lucil. 205 ff. (¼ 203 ff. M). desidero, like the English ‘want’, can mean
either ‘desire’ or ‘need’; here it has to mean the former.
25–6. neque / tumultuosum sollicitat mare: though H professes to be

talking of the contented man, he goes on to give two vignettes of the
opposite, as so often in the Satires. The figure of the discontented and
acquisitive merchant is conventionally contrasted with that of the coun-
tryman (e.g. 1. 1. 14 with N–H, epod. 2. 6). The tumult of the sea (cf. 3.
29. 63) is mirrored by the agitation of the merchant’s mind (cf. 2. 16 10,
Stat. silv. 2. 2. 28); Shorey quotes The Merchant of Venice 1. i. 8 ‘Your
mind is tossing on the ocean.’
27–8. nec saevus Arcturi cadentis / impetus: Arcturus, the brightest

star in Bootes, was often associated with bad weather; cf. Perses, anth.
Pal. 7. 539. 1 f. ˇP æØ, ¨ Ø, ŒÆŒc Ø Ø = #æŒæı
(with Gow–Page, HE 2895), Plaut. rud. 70 f. ‘nam Arcturus signum sum
omnium acerrumum: / vehemens sum exoriens, cum occido vehemen-
tior’. H is referring to the so-called ‘evening setting’ at the end of
October (Col. 11. 2. 78 ‘iv Kal. Nov. Arcturus vespere occidit: ventosus
dies’, Plin. nat. hist. 18. 313), i.e. the date when its setting came closest
after sunset (West on Hes. op. p. 379); the merchant’s greed appears
from his readiness to sail at the end of the sailing season. impetus suits
the onslaught of the storm rather than the movements of the star (cf. 31-
2 n.), but its collocation with cadentis makes something of an oxymoron.
See further Le Boeuffle 95 ff.
28. aut orientis Haedi: note the chiasmus ‘Arcturi cadentis . . . orientis
Haedi’. The Haedi are close to Capella (cf. 3. 7. 6); for the singular cf.
Prop. 2. 26. 56, Ov. ars 1. 410, Le Boeuffle 110. The evening rising of the
Haedi about the end of September was associated with bad weather
(Theoc. 7. 53 with Gow); the constellation is combined with Arcturus at
Virg. georg. 1. 204 f.
29. non verberatae grandine vineae: hail, sometimes in the summer, is

a particular menace to vineyards (cf. epist. 1. 8. 4 f. quoted on 30–1 below,
Pease on Cic. nat. deor. 2. 14). For verberare of the elements cf. 3. 27. 24,
OLD s.v. 3b (so plectantur at 1. 28. 27). As the vine-stock (vitis) was an
instrument of punishment (OLD 4), RN thinks there might be a
particular point in the lashing of the vines; vinea is used of the individ-
ual plant at Colum. 4. 10. 2, 4. 22. 5.
30. fundusque mendax: though the earth should repay its debts (3. 16.
30 n.), it sometimes plays false; cf. epist. 1. 7. 87 ‘spem mentita seges’,
Philemon (quoted in next note), Ov. met. 5. 479 f. ‘arvaque iussit / fallere
depositum’ (with Bömer), Petr. 117. 9. Normally such passages refer
to corn-growing, but fundus here is more comprehensive, as arbore
suggests.
30–1. arbore nunc aquas / culpante: the arbor, like the fundus,
is personified; it seems to mean fruit-trees in general, including the
olive. It was a commonplace that a farmer is always grumbling about the
weather; cf. epist. 1. 8. 4 f. ‘haud quia grando / contuderit vitis oleamve
momorderit aestus’, Philemon fr. 92. 10 f. (PCG 7 p. 275) [ ªB] f
Œı IıæŒı Id = æ ÆØ Ø ÆP
e j
IæE,
[Plat.] Axioch. 368c ŒºA ıd b ÆP
, ıd b KæÆ , ıd b
KŒÆıØ, ıd b Kæı, ıd b Łº ¼ŒÆØæ j Œæ (so nunc . . .
nunc in Horace). To blame both rain and drought is a sign of discontent
and a refusal to accept the world as it is.
31–2. nunc torrentia agros / sidera: plants were scorched at the rising of
the Dog-Star in July; cf. 3. 13. 9 ‘flagrantis . . . caniculae’, epod. 3. 15 f., 16.
61 f. ‘nullius astri / gregem aestuosa torret impotentia’, serm. 1. 7. 25 f.,
2. 5. 39 f., Alc. 347. 1 f. e ªaæ ¼æ æØººÆØ, = I þæÆ
ÆºÆ, Æ b łÆØ Pa ŒÆÆ , Plin. nat. hist. 17. 222. This
sideratio was attributed to the baleful influence of the constellation;
cf. Theophr. caus. plant. 5. 9., Plin. nat. hist. 18. 278. Geminus more
sensibly points out that the stars simply mark the wet and stormy
times of year (NÆªøª 17. 10): ‰ ı
æØ ÆæØºø æe
e æªØŒØ A a æd e IæÆ æØØ .
32. nunc hiemes iniquas: hiemes means both ‘winters’ and ‘storms’.
iniquas covers not just ‘harmful’ and ‘unfavourable’ but also ‘unfair’
(continuing the querulous tone of culpante).
33. contracta pisces aequora sentiunt: the extension of villas into the
sea was a topic of the diatribe against luxury (2. 18. 21 with N–H); for
hyperbole in such contexts cf. 2. 15. 1 ff., 3. 24. 1 ff. The contraction of the
fishes’ domain marks a bizarre invasion of a natural element (cf. N–H on
1. 2. 9); for the hubris involved cf. 1. 3. 21 ff. sentiunt implies ‘feel to their
cost’; cf. N–H on 2. 7. 9, epist. 1. 1. 84 f. ‘lacus et mare sentit amorem /
festinantis eri’.
34. iactis in altum molibus: i.e. when moles or piers have been pushed
out into the sea; cf. Caes. ap. Cic. Att. 9. 14. 1 ‘ab utroque portus cornu
moles iacimus’, bc 1. 25. 5, 3. 112. 2 ‘in longitudinem passuum DCCCC in
mare iactis molibus’, Strab. 5. 4. 6, Antiphilus, anth. Pal. 7. 379. 1 f. Ø
N –ºÆ
HÆ = ºÆØ; (on the harbour at Puteoli), Sen.Thy.
459 f. ‘retro mare / iacta fugamus mole’ (cf. ‘jetty’ from the French jetée),
Sidon. carm. 2. 57 f. ‘itur in aequor / molibus et veteres tellus nova
contrahit undas’. Most editors interpret iactis as ‘dropped’; cf. Virg.
Aen. 9. 711 f. ‘saxea pila cadit, magnis quam molibus ante / constructam
pelago iaciunt’, where molibus refers to blocks of masonry. But in our
view the parallels cited above seem to be more convincing illustrations
of what H says; the stages in the building of such piers are described in
Vitruv. 5. 12. 2–3, where the structures are an alternative to aggeres.
34–6. huc frequens / caementa demittit redemptor / cum famulis: huc

(with demittit) means ‘into the deep water’, i.e. where it is enclosed by
the moles. The redemptor is a contractor; locare is ‘to place a contract’,
conducere or redimere is ‘to take it up’ (RN sees a formal contrast between
re- and de-). The prosaic connotations of the word suggest ugly materi-
alism; cf. epist. 2. 2. 72 ‘festinat calidus mulis gerulisque redemptor’, Juv.
3. 38 ‘conducunt foricas, et cur non omnia?’ demittit redemptor carries the
humorous suggestion that the contractor himself wields a shovel. fre-
quens probably means ‘repeatedly’ (Ov. met. 2. 409 ‘dum redit itque
frequens’, Sen. Phaedr. 1076 f. ‘terga nunc torto frequens / verbere
coercet’, OLD s.v. 5); some understand ‘many a contractor’, but one
expects only one master-builder who supervises the work of his famuli.
caementa (another prosaic word) are the broken stones (from caedere)
used as the basis or ‘aggregate’ for concrete; they were bound together by
a mixture of lime, sand, and pulvis Puteolanus (pozzolana, a volcanic
dust that hardens in the water); cf. 3. 24. 3, Strab. 5. 4. 6, Plin. nat. hist.
35. 166, C. D. Curtis, JRS 3, 1913: 197 ff., N. Davey, A History of Building
Materials, 1961: 120 ff. Such constructions could be undertaken as here
by gangs of unskilled labourers.
36–7. dominusque terrae / fastidiosus: the organization of the sentence

puzzled L. Müller and A. Y. Campbell (the former found it as strange as
‘praefectus cum fabris et Caesar’); but by trailing the proprietor after the
contractor and his workmen H seems to suggest that the man is
demeaning himself. Shackleton Bailey (1982: 93) objected that when
dominus and famuli are found together ‘it is natural to think of the latter
as belonging to the former’; he therefore proposed ‘tum famuli dom-
inusque (sc. scandunt)’. But apart from the awkward structure, the
famuli should not be put on the same level as the dominus; they do not
climb to the top of the building to be plagued by Timor et Minae. NR is
attracted by RN’s idea ‘cum famulis dominoque terrae / fastidioso’; this
puts famulis in a pointed relationship with domino (even though they do
not belong to him), and strengthens the innuendo that the latter is no
higher than a manual labourer; for the idiomatic change of case
with dominus (38) cf. ‘somnum . . . somnus’ (21) and 3. 16. 15–16 n.
R. Renehan suggests that cum famulis is to be taken with dominus and
that -que has been postponed (CP 83, 1988: 324 f.); his best parallel is 4. 2.
21 f. ‘flebili sponsae iuvenemve raptum / plorat’, but that passage
is unambiguous whereas here the reader would be misled to no purpose.
dominusque terrae at first sight suggests ‘a lord of the earth’; cf.
‘terrarum dominos’ (1. 1. 6). It then transpires that the genitive depends
on fastidiosus, which echoes fastidit (23); the proprietor’s disdain for dry
land provides a further twist to the satire; cf. ‘parum locuples continente
ripa’ (2. 18. 22). By way of contrast, one thinks of Pliny, happily unaware
of his decadence, fishing from his bedroom window (epist. 9. 7. 4).
37–8. sed Timor et Minae / scandunt eodem quo dominus: Timor et

Minae are personified fears from within and threats from without; they
are represented as malignant companions that cannot be shaken off (cf.
Lucr. 2. 48 ‘curaeque sequaces’, 2. 16. 11 f. ‘curas laqueata circum / tecta
volantis’). For scandunt cf. 2. 16. 21 (cited below on 38–9). In our passage
the verb suggests the height of the new building as opposed to humilis
domos (22); after observing the laying of the foundations H’s camera
sweeps up to the top floor of the completed building. For tall villae
maritimae cf. J. H. D’Arms, Romans in the Bay of Naples, 1970: pls. 13–15b
(wall-decorations from Stabiae and Pompeii).
38–9. neque / decedit aerata triremi: the ‘brazen trireme’ was properly a
warship with three levels of oars like its Greek counterpart ( J. S.
Morrison, J. F. Coates, N. B. Rankov, The Athenian Trireme, 2000:
fig. 45) and a bronze beak for ramming (ibid. fig. 67); cf. Plin. nat. hist.
32. 3 ‘rostra illa aere ferroque ad ictus armata’, Barwick, op. cit. 263, TLL
1. 1059. 23. There is a close parallel at 2. 16. 21 ff. (probably earlier)
‘scandit aeratas vitiosa navis / Cura nec turmas equitum relinquit, / ocior
cervis et agente nimbos / ocior Euro’; the recipient Grosphus may have
been an eques with an interest in cavalry (N–H vol. 2, p. 253). Lucretius
had already said that a military commander finds no peace of mind in
contemplating his forces (2. 40 ff.); for his curae sequaces (2. 48) are not
dispelled by a display of arms. RN sees here a hint of Maecenas, who
might have retained a trireme after service at Actium; cf. epist. 1. 1. 91–3
‘quid pauper? ride: mutat cenacula lectos / balnea tonsores, conducto
navigio aeque / nauseat ac locuples quem ducit priva triremis’. There he
takes pauper as Horace and the owner of the trireme as Maecenas; cf.
94 f. ‘si curatus inaequali tonsore capillos / occurri, rides . . . ’ (where
there is no pronoun to distinguish the subject of occurri from the
pauper).
NR thinks it is going too far to see a covert allusion to Maecenas (see
introduction above), and he interprets epist. 1. 1 differently: in 77 ff. we
hear of men’s failure to find satisfaction; in 83–9 we have the restless
dives—an entirely general figure; the restless pauper in 91–3 is just as
silly, and just as general (the contrasting boats figured in popular
philosophy, as Heinze remarks). In 94 ff. Horace comes to his own
case; he, too, is restless and inconsistent (97–100); but the same is not
said of Maecenas. He laughs at the poet’s haircut and sloppy dress and is
irritated if his nails aren’t trimmed, but he fails to notice his restless
spirit because such a defect is so normal (sollemnia in 101).
39–40. <et> / post equitem sedet atra Cura: et has negligible authority
(cf. Fraenkel 317 n. 5), but the two weak line-endings well convey unre-
mitting pursuit (N–H on 2. 6. 2); Bentley less plausibly proposed postque
equitem. RN, following Barwick, thinks that H is referring to a military
review, as with the parallel turmas equitum at 2. 16. 22 (quoted in the
previous note); as in that passage, the point would be that Care keeps up
with a display of power and speed (for a possible reference to the
equestrian Maecenas see the introduction above). NR prefers to believe
that the motif of 2. 16. 22 has been varied; since the trireme is a private
vessel, the riding is more likely to represent another of the rich man’s
efforts to shake off his Angst (so Heinze). The postponement of Cura
produces not only an elegant chiasmus (balancing Timor et Minae) but a
sinister climax; for atra cf. 3. 14. 13, 4. 11. 35, serm. 2. 7. 115 ‘nam comes atra
premit sequiturque fugacem’, Hom. Il. 4. 117 ºÆØø (æ Oıø.
41. quodsi dolentem nec Phrygius lapis: quodsi (‘Since, then,’) intro-
duces the inescapable conclusion; cf. 1. 1. 35, Lucr. 2. 47, Prop. 1. 1. 37,
Fraenkel 24. dolor was widely used of unhealthy mental and spiritual
states, including anxiety, cf. Varro, Men. 36, Lucr. 4. 1067, Cic. Tusc. 3.
22 ff., 4. 23 ff.
Phrygius lapis was a white marble with purple markings; cf. Strab. 12.
8. 14, Stat. silv. 1. 5. 37 f. ‘(purpura) cavo Phrygiae quam Synnados antro /
ipse cruentavit maculis lucentibus Attis’, Juv. 14. 307 with Courtney,
D. Magie, Roman Rule in Asia Minor, 1950: 1. 50, 2. 815 n. 96, R. Gnoli,
Marmora Romana, 1971: 142 ff. and pl. 126. Phrygius probably made an
alliteration with purpurarum (cf. Allen 26 f.). RN thinks that the
repeated p may have given an opulent effect as at 2. 12. 22 pinguis
Phrygiae.
42–3. nec purpurarum Sidone clarior / delenit usus: the MSS and
Porph. read sidere (cf. 3. 9. 21 sidere pulchrior, Hom. Il. 6. 295 Icæ
S IºÆ of a Sidonian robe, 290 f.); Sidone is RN’s conjecture
(Collected Papers, 144 ff., 343 f.). A place-name is needed to balance
Phrygius, Falerna, and Achaemenium; and Sidone is exactly the word
required (for Sidonian purple cf. epist. 1. 10. 26, Virg. Aen. 4. 137 with
Pease). For the compendious comparison (‘brighter than the purple of
Sidon’) cf. 2. 6. 14 ff. ‘ubi non Hymetto / mella decedunt viridique certat
/ baca Venafro’, Varr. rust. 1. 2. 6 ‘quod oleum (conferam) Venafro?’ (also
of products), K–S 2. 566 f., H–Sz 826. For Sidone with a short o cf. Sil. 8.
436 f. ‘stat fucare colus nec Sidone vilior Ancon / murice nec Libyco’,
Mart. 2. 16. 3, 11. 1. 2; the o is also often short in the adjective; cf. Virg.
Aen. 1. 678 and 4. 75 with Pease. G. Perl regards the conjecture as
methodologically wrong in the absence of earlier parallels for the
short o (Acta Ant. Hung. 39, 1999: 244), but it is unsafe to assume that
Silius must have been the first to use it; F. Cairns objects that no purple
was brighter than Sidonian (Coll. Latomus 266, 2002: 90), but the
hyperbole sharpens Horace’s point.
For usus in the sense of ‘wearing’ cf. Mart. 9. 62. 1 f. ‘tinctis murice
vestibus . . . / utitur’, OLD s.v. utor 2b. The periphrasis is suitably gran-
diose; H seems to have been influenced by Lucr. 2. 52 ‘nec clarum vestis
splendorem purpureai’, but he might also be including hangings and
coverlets (cf. serm. 2. 6. 106 ‘purpurea porrectum in veste locavit’). delenit
suits the quasi-medical meaning of dolentem; cf. epist. 1. 1. 34 f. ‘sunt
verba et voces quibus hunc lenire dolorem / possis’, Phaedr. 3 praef. 44
‘his dolorem delenirem remediis’.
43–4. nec Falerna / vitis Achaemeniumque costum: Falernian, from

the coastal region of north-west Campania, was a wine of the highest
quality: cf. Plin. nat. hist. 14. 62 ‘nec ulli nunc vino maior auctoritas’.
By mentioning the vine rather than the wine H combines the themes of
viticulture (9 ff., 29 ff.) and gastronomic luxury (18 f.); the Italian name
marks a contrast with the exotic Achaemenium. Achaemenes was the
legendary founder of the Persian royal house (2. 12. 21 with N–H); for
the royal association with perfumes cf. epod. 13. 8 ‘Achaemenio . . . nardo’,
Plin. nat. hist. 12. 41, 13. 18 ‘ergo regale unguentum, appellatur quoniam
Persarum regibus ita temperatur, constat myrobalano costo amomo . . . ’,
Diog. Laert. 2. 76 (for another royal perfume add Sappho 94. 18 ff.).
costum (‘putchuk’) probably came from India rather than Persia ( J. I.
Miller, The Spice Trade of the Roman Empire, 1969: 84 ff.), but H was
concerned with literary associations rather than botanical precision (cf.
2. 11. 16 ‘Assyriaque nardo’ with N–H).
45–6. cur invidendis postibus et novo / sublime ritu moliar atrium?:

invidendis postibus is a descriptive ablative depending on atrium and co-
ordinate with the phrase novo sublime ritu. In such contexts postes is used
for the door as a whole (synecdoche), cf. OLD 2b, Blümner (1911), 16 f.,
P. Howell, Philologus 112, 1968: 132 ff. Not just the posts but the panels
could be decorated with ivory, tortoiseshell, etc.; cf. Virg. georg. 2. 463
‘nec varios inhiant pulchra testudine postis’, Lucan 10. 118, Mart. 1. 70.
14. For invidendis (here adjectival like timendorum in v. 5), cf. 2. 10. 7 f.
‘invidenda . . . aula’; such adornments bring not admiration but envy.
After the extravagances mentioned in the previous stanza, H might
have been expected to say ‘why should I seek any of these things?’;
instead he turns in the apodosis to further examples of luxury (for this
characteristic schema cf. N–H on 1. 7. 10, Schmidt 335 ff.).
novo ritu is an oxymoron, as a ritus is normally well established.
The veterum norma (2. 15. 12) had rejected such a scale for private houses;
Augustus avoided excessive grandeur for himself (Suet. Aug. 72. 1) and
deprecated it in others (ibid. 72. 3 ‘ampla et operosa praetoria gravaba-
tur’); but his legislation on the height of city tenements (ibid. 89. 2) is
not relevant here. sublime is a grandiose word, and repeats the idea of
height implied in scandunt (38). moliar describes the effort needed to
rear a pile (cf. 2. 15. 2, 3. 29. 10 ‘molem propinquam nubibus’).
47–8. cur valle permutem Sabina / divitias operosiores: for the modest
personal closure cf. 1. 31. 15 ff., 2. 17. 32, 3. 29. 62 ff., 4. 2. 54 ff., Esser 9 ff.
2 . A N G V S TA M A M I C E 21
(on the ‘Ich-Schluss’). For H’s Sabine valley, cf. 1. 17. 17, epist. 1. 16.
5 ff., 1. 18. 104, G. Highet, Poets in a Landscape, 1957: 137 ff., Encicl.
oraz. 1. 253 ff. As valle suggests both lowness and seclusion it makes a
contrast with sublime. Sabina has a cluster of connotations, e.g. ‘hardy’,
‘simple’, ‘frugal’, and ‘old-fashioned’; cf. 1. 20. 1, 2. 18. 14, epod. 2. 41 f.,
epist. 2. 1. 25.
permutem means here ‘take in exchange’ (cf. 2. 12. 23); the compound’s
commercial nuance suits H’s rejection of materialism. operosiores de-
scribes the over-elaboration of luxury building (cf. Suet. Aug. 72. 3
quoted in the previous note, OLD s.v. 3), but also suggests that wealth
brings nothing but bother; cf. serm. 2. 6. 79 ‘sollicitas . . . opes’, OLD 2a,
Muson. 108. 12 Hense (on grand buildings) æÆªÆÆ
Ø
º.
2. ANGVSTAM AMICE
[L. Amundsen, SO suppl. 11, 1942: 1 ff. ¼ Oppermann 120 ff.; P. J. Connor, Hermes 100,
1972: 241 ff.; G. Davis, Class. Ant. 2, 1983: 9 ff.; V. B. Jameson, TAPA 114, 1984: 219 ff.;
D. Lohmann in Schola Anatolica (Tübingen 1989), 336 ff.; Lyne (1995), 208 ff.; W. J. Oates,
The Influence of Simonides of Ceos upon Horace, 1932 and 1971: 1 ff.; Pasquali 667 ff.;
R. Stoneman in Aischylos und Pindar (ed. E. G. Schmidt, Berlin), 1981: 257 ff.; G. Williams,
Figures of Thought in Roman Poetry, 1980: 184 ff.]
1–16. Let the young soldier learn to accept hardship as well as danger in
fighting the Parthians. Let the enemy princess sigh on the battlements in case
her betrothed provokes him. It is sweet and glorious to die for one’s country,
but the man who runs away is killed dishonourably. 17–24. A man’s virtus
knows nothing of rebuffs, and does not accept or surrender office to suit the
popular whim. Their virtus opens the gate of heaven to those who have
earned immortality. 25–32. Loyal discretion also has its sure reward; I shall
shun those who betray the mysteries (i.e. state secrets). Often Jupiter destroys
the innocent with the guilty; rarely does Nemesis fail to overtake the
criminal.
The first section of the ode (1–16) praises the soldierly virtues of
courage and endurance, which were particularly prized in Rome’s mili-
taristic society (K. Büchner, Studien zur römischen Literatur 3, 1962:
1 ff. ¼ H. Oppermann, Römische Wertbegriffe, 1967: 376 ff.). These
qualities are treated as a training not just for war but for life (1 n.). A
topical note is provided by the emphasis on cavalry (4 n.) and the need
to match the Parthians: Carrhae had not been avenged (1. 2. 51, 3. 5.
4 n.). Eastern wars lead to Homeric scenes, the enemy princess watching
from the battlements (6 n.) and the comparison of the young Roman to
a bloodthirsty lion (11–12 n.); the romantic colouring derives from Hel-
lenistic poetry and tales about Alexander (cf. 1. 29. 5 f. ‘quae tibi virgi-
num / sponso necato barbara serviet?’). This fantasy shows less
humanity than is sometimes suggested, and it ends rather abruptly
with patriotic maxims, drawing again on early Greek poetry, about the
glory of death in battle (13–14 nn.).
The central section (17–24) turns from the virtus or manliness of the
soldier (though the word is not actually used in the previous lines) to the
more superhuman virtus of the great man. Though no individual is
mentioned, Augustus is obviously the example in the poet’s mind; about
27 bc he was awarded a golden shield on which his virtus was recorded
(res gest. 34. 2, A. N. Wallace-Hadrill, Historia 30, 1981: 300 ff., Zanker
95 f., Galinsky 80 ff.). The character is described in Stoic terms: the
great man knows nothing of setbacks because he is superior to the
accidents of fortune (17 n.), and after death he will enjoy an immortal
life in the sky, as memorably expounded in Cicero’s Somnium Scipionis
(21–2 n.); see also 3. 3. 11 n. In all this there is no mention of virtus in
ordinary human relationships, as it was understood by Panaetius and
Cicero in his de officiis, and by Lucilius in his famous lines beginning
(1329 ff.) ‘virtus scire homini rectum utile quid sit, honestum, / quae
bona, quae mala item, quid inutile turpe inhonestum . . . ’ (imitated by
Horace himself at epist. 1. 2. 3 ff.).
In the final section (25–32) Horace does turn to social life, concen-
trating on fidele silentium, ‘loyal discretion’; and here the transition is less
clear. Orelli comments ‘socia virtutis est fides’; but though fides was one
of the virtues, it had no special association with the high-flown virtus of
the previous stanzas. It might be better to see a deliberate decrescendo:
not everyone can show the superhuman qualities of the Stoic hero, but
‘they also serve’ who at least offer loyalty and discretion. Mommsen saw
a reference to the new imperial bureaucracy, and the need to make it
conscious of security (Reden und Aufsätze, 1905: 168 ff.). It is better to
think of the amici principis (cf. A. Wallace-Hadrill, CAH 10, edn. 2, 285
on court societies); in all likelihood Horace himself, in spite of his
modest pretences, had now reached the fringes of the imperial circle,
where caution was advisable. Ironically, even a central figure like
Maecenas was found wanting in reticence (Suet. Aug. 66) when he
betrayed an official secret to his wife; but that was after the publication
of Odes I–III.
The poem is now the least admired of the Roman Odes; in spite of its
trenchant aphorisms and two memorable vignettes (6 ff., 31 f.), the
qualities commended, though in harmony with the Augustan value-
system and intelligible at a time of crisis, are not those that appeal to
more liberal societies. One can also see why Pasquali described it as the
most difficult ode in the cycle (668); for at certain points it seems to
allude to events or situations which, though probably clear enough to
contemporary readers, are puzzling to us. When a great man’s virtus is
said to know nothing of electoral defeat, one wonders whether Horace
has in mind some specific instance (17 n.); the same question may arise
when his virtus is said not to take or leave office to suit the popular
whim (19–20 n.). The vehement denunciation of those who give away
secrets is expressed by analogy with the profanation of the Eleusinian
mysteries (26–7 n.). Some have seen here a reference to the downfall of
Gallus, but though he may have been arrogant and indiscreet (Ov. trist.
2. 446, Suet. Aug. 66. 2, Dio 53. 23. 5), there is no suggestion that he
betrayed confidences. Perhaps in the period before Actium, when the
arcana of the Triumvirate were revealed, they contained some damaging
instances of political duplicity as well as the sexual escapades mentioned
in Suet. Aug. 69 (for Antony’s invective cf. K. Scott, Mem. Amer. Acad.
Rome 11, 1933: 1 ff.); in that case Horace may have been thinking of the
vengeance that followed the battle. Or one might imagine a more
serious instance of the kind of thing described in serm. 2. 6. 51 ff.
(‘numquid de Dacis audisti?’ etc.). But the reference may be entirely
general; see Dio 53. 19. 3 on the cloud of secrecy that descended with the
advent of the Principate.
There is also a literary problem: all three sections contain reminis-
cences of Simonides (see the notes on 14, 21, 25, 27–9), and it has been
conjectured that the ode as a whole reflects a lost poem on civic virtue
(Oates, op. cit. 28 ff., Harrison, cited in introduction to 3. 30). But it is
noteworthy that these imitations all seem to diverge from their model,
and at v. 25 very considerably. What is more, though Horace was adept
at transposing particular phrases to his unique idiom, the construction
of his poems is wholly his own (Syndikus); the themes also arise from
the Augustan situation, not from archaic Greece. Where several imita-
tions have been demonstrated, it is always possible that others are lost;
for instance Simonides might be the source of ‘dulce et decorum est pro
patria mori’ (see 13 n.). But it is not to be supposed that any Greek
prototype would explain the difficulties of the ode.
Metre: Alcaic.
1. Angustam amice pauperiem pati: angustam refers to ‘straitened cir-

cumstances’ (the dead English metaphor may be derived from Juv. 3. 165
‘res angusta domi’); pauperiem as usual describes not indigence but
modest means (cf. 3. 24. 42 n., Apul. apol. 18); the alliteration gives
the line the memorable quality of a motto. The hardship of military
service is a lesson for later life, a sentiment worthy of Augustus himself
(cf. 3. 24. 52 ff.), though unrealistic when applied to the life awaiting a
young cavalry-officer; for similar ideas cf. Cic. Tusc. 2. 37 ff. (it teaches
endurance and fortitude in the face of pain), Sall. Cat. 7. 4–6 and 9. 1–3
(it is connected with the lack of avarice in early Roman society), Jug. 85.
33 (Marius describes his qualifications for office) ‘at illa multo optuma
rei publicae doctus sum: . . . hiemem et aestatem iuxta pati, humi
requiescere, eodem tempore inopiam et laborem tolerare’. This reverses
the usual topic that a frugal upbringing produces good soldiers (Syndi-
kus): see 3. 6. 37–8 n.
amice means ‘in a well-disposed way’; the best parallel is Sen. dial. 7.
6. 2 ‘beatus est praesentibus qualiacumque sunt contentus amicusque
rebus suis’ (cited by J. Delz). Perhaps H is representing some Greek
adverb, e.g. IªÆH ‘contentedly’ (Mitscherlich), ºø (L. Müller),
possibly PH or æH ; cf. Plut. de virtute et vitio 100d ŒÆd Æ
ŒÆd ıªc ŒÆd ªBæÆ KºÆæH ŒÆd æH . . . æıØ. pati with an
adverb means ‘accept’, ‘take’ (OLD 4), like ferre. Bentley argued that the
adverb was superfluous (yet note Nep. Epam. 3. 4 ‘paupertatem . . . facile
perpessus est’); he read the vocative amici (ps.-Acro ‘hanc oden ad
amicos generaliter scribit’), but such an address would not suit the
Roman Odes. Shackleton Bailey, who obelizes, suggests et aeque (the
first two letters of amice could be a dittography after angustam); he
explains ‘aeque et condiscat pauperiem pati et Parthos vexet’.
2–3. robustus acri militia puer / condiscat: acri militia should be taken
with robustus (¼ corroboratus); cf. Cic. Cat. 2. 20 ‘genus exercitatione
robustum’. Here acri, whose root meaning is ‘sharp’, emphasizes the
harshness of the soldier’s life; at 1. 29. 2 ‘acrem militiam paras’ it points
to fierceness against the enemy. For puer of a young soldier cf. epist. 1. 18.
54 f. ‘denique saevam / militiam puer et Cantabrica bella tulisti’ (military
service began at 16 or 17); here the word is pointedly combined with
robustus, which suggests the robur aetatis (‘cum iam robustus est ac per
hoc cum adulescens est’ Porph.). condiscat suits a cadet (cf. Tyrt. 11. 27
æø ZæØÆ æªÆ ØÆŒŁø ºØ); the prefix intensifies (‘let
him learn thoroughly’) as at 4. 11. 34 f. ‘condisce modos amanda / voce
quos reddas’. Yet RN is tempted to consider ‘let him learn at the same
time to put up with poverty’ (¼ addiscat); no parallel presents itself for
such a usage, but for ‘learn together’ in another sense cf. Apul. flor. 18. 42
‘ex iis qui mihi Athenis condidicerunt’.
3–4. et Parthos ferocis / vexet eques metuendus hasta: Augustus at-

tached great importance to equestrian exercises (cf. 1. 8. 5 ff. with N–H,
3. 7. 25, 3. 24. 54 ff., Virg. Aen. 5. 548–602 on the ‘lusus Troiae’, Suet. Aug.
43. 2, Dio 51. 22. 4); for the place of young equites in his organization
of the iuventus cf. Z. Yavetz in Millar–Segal 16 ff. While it seems that
the equites legionis served mainly as mounted messengers and scouts,
the cavalry of the allies was often commanded by Roman officers
( J. Suolahti, The Junior Officers of the Roman Army in the Republican
Period, 1955: 202 f., cf. K. R. Dixon and P. Southern, The Roman Cavalry,
1992: 22); such men would be important in coping with the Parthian
mounted archers. vexet (‘harry’) suits attacks by cavalry; cf. 4. 14. 22 f.
‘impiger hostium / vexare turmas’. For metuendus hasta cf. 1. 12. 23 f.
‘metuende certa, / Phoebe, sagitta’, Stat. Theb. 4. 221 ‘metuendus in
hasta’; the construction is tighter if the ablative is also taken with vexet
(L. Müller, Darnley Naylor).
5–6. vitamque sub divo et trepidis agat / in rebus: divum (or dium) was
an old word for the sky that survived in the phrase sub divo (2. 3. 23) and
sub divum (1. 18. 13). res trepidae was a standard euphemism in official
and historical contexts for alarming situations; cf. Sall. Jug. 91. 5 ‘res
trepidae, metus ingens, malum inprovisum’, Liv. 1. 27. 7 ‘Tullus in re
trepida duodecim vovit Salios fanaque Pallori ac Pavori’, 4. 34. 5, 5. 50. 4,
Tib. 2. 3. 21. rebus is given emphasis by its position at the beginning of
the line before a full stop; this may stress the idea of real military action
as distinct from training; cf. Val. Fl. 2. 380 f. ‘me tecum solus in aequor /
rerum traxit amor’, OLD 7.
6–8. illum ex moenibus hosticis / matrona bellantis tyranni / prospi-

ciens et adulta virgo: from the enemy walls (for the archaic hosticus cf.
Porph. on 2. 1. 1 ‘illi (sc. antiqui) enim civica et hostica, deinde civilia et
hostilia dicebant’, Plaut. capt. 246, serm. 1. 9. 31) a girl of marriageable
age, i.e. only in her early teens (cf. Treggiari 39 ff.) and her mother, the
wife (OLD s.v. matrona 2) of an enemy king (tyrannus implies an
Eastern despot) look out at (prospiciens is transitive) the young Roman
officer who is a threat to the girl’s betrothed (9–11). Unlike most Roman
poets H prefers ex to e before consonants, and indeed never uses the
latter form in the Odes (Axelson 119 f.).
A survey of the battlefield from the city walls (teichoscopia) was a
literary theme since Homer; cf. Il. 3. 161–244 (Helen identifies the
leading Greeks), 22. 463 ff. (Andromache sees Hector’s body being
dragged), Hes. scut. 242 f. Æƒ b ªıÆEŒ Kıø Kd æªø =
ÆºŒø Of ø, Enn. ann. 418 Sk. ‘matronae moeros complent

spectare faventes’, Liv. 31. 24. 13, Tarrant on Sen. Ag. 622 (with further
examples). Here H includes a romantic element derived from Hellenis-
tic poetry (though more usually the girl falls in love with an enemy
warrior); cf. Parthenius, erot. path. 21 (The Story of Pisidice, citing
Apollonius fr. 12 Powell); for further examples see Lightfoot ad loc.,
Prop. 4. 4. 19 ff. (Tarpeia and Tatius, cf. Ogilvie on Liv. 1. 11. 5–9), Ov.
met. 8. 21 ff. and ciris 172 ff. (Scylla and Minos), Val. Fl. 6. 575 ff. (Medea
and Jason).
9–11. suspiret, eheu, ne rudis agminum / sponsus lacessat regius
asperum / tactu leonem: suspiret describes a lover’s sigh, whether of
anxiety or longing; cf. 3. 7. 10, epod. 11. 10, Catull. 64. 98, Call. ep. 43. 1 f.,
Ov. fast. 1. 417, CIL 4. 4342 of a gladiator ‘suspirium puellarum’, ‘the
heart-throb of the girls’. suspiret and sponsus suit only the princess, the
nearer and more important subject; for the zeugma cf. epod. 15. 7 f. ‘dum
pecori lupus et nautis infensus Orion / turbaret hibernum mare’. eheu
represents the girl’s feelings; for this usage cf. epod. 15. 23 f. ‘heu heu,
translatos alio maerebis amores, / ast ego vicissim risero’, Catull. 64. 61
‘prospicit eheu’ (Ariadne on the shore). The ne clause specifies the
anxiety implicit in suspiret. For rudis agminum cf. epist. 2. 2. 47 ‘rudem
belli’ (see OLD 5b for ‘raw recruits’), Cic. Brut. 292 ‘omnium rerum
inscium . . . et rudem’, K–S 1. 437 f., H–Sz 77 f. agmina could be used of
warfare in general and is well suited to troops of cavalry.
The sponsus regius is the prince of an allied community, like Coroebus
(Virg. Aen. 2. 341 ff.) or Turnus (ibid. 7. 56); the adjective could have a
hint of disparagement suggesting Eastern luxury. lacessat suits a chal-
lenge in war; and it was proverbially dangerous to provoke a lion (3. 20.
2 n.). Warriors were compared to lions since Homer (Il. 5. 136, 20.
164 ff.); the metaphor comes more naturally from the patriotic Roman
poet than the foreign princess. In RN’s opinion asperum tactu suggests
not just ferocity at the touch (1. 23. 9 ‘tigris ut aspera’) but tangible
roughness (cf. æØ of bristling animals); the somewhat archaic tactu
is a second supine (cf. serm. 1. 4. 124 ‘inutile factu’, Lucr. 6. 1150 ‘aspera
tactu’, Cic. Tusc. 2. 20 ‘o multa dictu gravia, perpessu aspera’, K–S. 1.
724 f., H–Sz 382 f., Bo 379).
11–12. quem cruenta / per medias rapit ira caedis: for lions’ fury cf. 1. 16.
15 with N–H. For rapit (‘carries along’) cf. epod. 7. 13 ‘furorne caecos an
rapit vis acrior?’, Hom. Il. 20. 170 ff. (of a lion) Pæfi Ð b ºıæ ŒÆd
N
Æ IæøŁ = ÆÆØ, b ÆPe KæØ Æ
ÆŁÆØ, =
ªºÆıŒØ ø NŁf æÆØ Ø, Liv. 10. 41. 1 ‘Romanos ira spes ardor
certaminis . . . in proelium rapit’, OLD 11b–c. All the words in H’s
relative clause contribute to the impression of ferocity and slaughter;
for the transferred epithet cf. Sen. Phaedra 542 f. ‘venit imperii sitis /
cruenta’, Val. Fl. 3. 84 f. ‘clamorque tubaeque / sanguineae’. per medias
caedis again has an epic ring; cf. Hom. Il. 10. 298 i , i ŒıÆ
Ø Æ ŒÆd ºÆ Æx Æ (Diomedes and Odysseus are advancing
like lions).
13. dulce et decorum est pro patria mori: decorum in the sense of ‘noble’
or ‘glorious’ can be paralleled from all periods, OLD 3; in this heroic
context it is enough to cite Tyrt. 10. 1 f. ŁÆØ ªaæ ŒÆºe Kd
æ
ØØ Æ = ¼æ IªÆŁe æd fi ffl ÆæØ Ææ, Callinus
1. 6 ff., C. W. Müller, Gymnasium 96, 1989: 317 ff. But here it has the
additional sense of ‘right and proper’, OLD 4; in philosophy the pleasant
and the honourable are sometimes opposed, but here they are compat-
ible. In view of the patriotic commonplaces in 13 and 14, dulce might be
expected to provide a conventional sentiment, perhaps echoing Simoni-
des (14 n.), yet no exact parallel is available that antedates Horace. It is
not enough to quote Bacchyl. 3. 47 ŁÆE ªºŒØ (for Croesus has
seen his country overrun); the point must be that the thought of well-
earned glory is sweet (the adjective is often used of glory in Pindar,
though not actually of death). Something similar is found in Sextus
Empiricus (contra ethicos 107) lÆØ ŒÆd ªŁÆØ æe f KÆı ,
but the best parallel is provided by Ach. Tat. 3. 22. 1 Iºº bæ ºı Œi
IŁÆE fi , ŒÆºe › Œı , ªºıŒf › ŁÆ (cited by S. J.
Harrison, RhM 136, 1993: 91 ff.). Add Stat. Theb. 4. 230 f. ‘mortis
honorae / dulce sacrum’, Hegesippus (4th cent.) 5. 24. 1 ‘( Josephus) cui
dulce fuerat ante patriam mori et pro patria’ (we owe these last two
passages to Dr Nigel Holmes), Prud. peristeph. 1. 25 ‘hoc genus mortis
decorum’ followed by 1. 51 ‘dulce tunc iustis cremari, dulce ferrum
perpeti’ (see C. H. Gnilka, RhM 138, 1995: 94 ff.).
dulce has offended modern sentiment (cf. Wilfred Owen, no. 144 ed.
Stallworthy ‘The old lie, Dulce et decorum est / pro patria mori’, Ezra
Pound, Ode pour l’élection de son sépulcre 4. 11 f.). Cicero himself mocks
the notion that the wise man under torture says ‘quam hoc suave’ ( fin. 2.
88; cf. Sen. epist. 67. 15). It is unconvincing to argue that H is simply
describing the attitude of the puer (Lohmann, op. cit.), or that the
pleasure is derived by those who hear about the glorious dead
(H. Hommel, RhM 111, 1968: 233 ff. in a useful collection of material).
RN at one time considered ‘dulci decorum est pro patria mori’ (see the
discussion in W. Ludwig (ed.), Horace, Entretiens Hardt 39, 1993: 32 ff.);
he relied on Acts of the Pagan Martyrs 11. 41 ff. Musurillo: Œº KØ
bæ B ªºıŒı ı Ææ ºıBÆØ. This proposal would
produce a more rational and a nobler sentiment; but in view of the
evidence cited above the idea must now be abandoned. It has to
be recognized that the ethos of most societies, including our own,
has often been different from the individualism that now prevails in
the West.
14. mors et fugacem persequitur virum: for the pattern ‘mori / mors’ in
successive lines see Wills 397. The aphorism is modelled on Simonides,
PMG 524 › Æs ŁÆ Œ
ŒÆd e ıª Æ
(the adversative Æs
suggests that this followed a sentence similar to v. 13 above). The
meaning of the two passages does not seem to be the same ( pace
Heinze). H is saying that if a soldier runs away the enemy catches
up on the battlefield (as is shown by the balance of persequitur and
fugacem and by vv. 15 f.); cf. Porph. ad loc. ‘hic ostendit eos maxime
in bello periclitari qui fugiunt’, Xen. Cyr. 3. 3. 45, Sall. Cat. 58. 16–17
‘nam in fuga salutem sperare . . . ea vero dementia est. semper in proeliis
iis maximum est periculum qui maxime timent’, Curt. 4. 14. 25. Si-
monides, however, probably meant that if a man shirks battle altogether
he will die in the end one way or another; cf. Hom. Il. 12. 322 ff.
(Sarpedon’s famous speech on noblesse oblige), Callin. 1. 14 f. ººŒØ
œBÆ ıªg ŒÆd F IŒ ø = æ
ÆØ, K YŒfiø EæÆ Œ

ŁÆı, Aesch. TrGF 3. fr. 362. 3 f., Eur. TGF fr. 10. 1 f., Dem. 18. 97
with Wankel.
15–16. nec parcit imbellis iuventae / poplitibus timidoque tergo: it was

standard practice to slash the hamstrings of a fleeing enemy; cf. Hom. Il.
13. 212, Virg. Aen. 9. 762, Liv. 22. 51. 7 (Cannae), Ov. met. 8. 363 f.
‘trepidantem et terga parantem / vertere succiso liquerunt poplite
nervi’, Veg. mil. 1. 11. 7 ‘poplites et crura succidere’. It was similarly
disgraceful to be struck in the back (Hom. Il. 13. 289, 22. 283, Tyrt. 11.
17 f., Catull. 64. 339 with Ellis; so still Binyon, For the Fallen v. 12 ‘They
fell with their faces to the foe’); the epithet as often is transferred from
the person to the body (1. 37. 1, 3. 5. 22 etc.). Two important ninth-
century MSS (A and B) read timidove, while the bulk of the tradition
has timidoque; both readings can be justified, and both can be paralleled
(Liv. 22. 48. 4 ‘tergaque ferientes ac poplites caedentes’, paneg. Lat. 2. 36.
1 ‘alii poplitibus imminere, alii terga configere’).
17. virtus repulsae nescia sordidae: for the sequence of thought see the
introduction above. H is referring here to the man of virtus, not to the
divine abstraction of 2. 2. 19 (quoted below). He adopts the Stoic view
that the good man knows nothing of humiliating setbacks because his
mind is superior to the accidents of fortune; cf. 2. 2. 17 ff. ‘redditum Cyri
solio Phraaten / dissidens plebi numero beatorum / eximit Virtus’, serm.
2. 7. 88 ‘in quem manca ruit semper Fortuna’, Cic. paradox. Stoic. 6. 43
‘animus oportet tuus se iudicet divitem, non hominum sermo neque
possessiones tuae’, Sen. epist. 57. 3, 118. 4 (see below). Elsewhere in his
private capacity H derides the Stoic paradox that only the wise man is
truly king (serm. 1. 3. 124 ff., epist. 1. 1. 106 f.); cf. also Lucil. 1226 with
Marx, Cic. fin. 3. 75, Lucull. 136 ‘neminem consulem praetorem imper-
atorem, nescio an ne quinquevirum quidem quemquam, nisi sapientem?’
repulsae properly refers to defeat in an election (cf. epist. 1. 1. 43 ‘exiguum
censum turpemque repulsam’, OLD 1), and this motif is continued
below in 19 f. For the same philosophic attitude to the defeat of a
good man cf. Cic. Tusc. 5. 54 (on Laelius’ initial rejection for the
consulship), Val. Max. 7. 5. 6 (de repulsis), Sen. epist. 71. 11, 118. 4, Plin.
nat. hist. praef. 9 (all referring to Cato).
In the early Principate, even if some freedom of elections remained,
there was no question of a major consular candidate being rejected. As
the following stanza on virtus points to Augustus himself (21 n.), the
same may be true here. Augustus, of course, had lost no election, but
repulsae may hint indirectly at some other setback; Williams in his
commentary (p. 37) sees an allusion to a failure to push through his
moral legislation (see below on 3. 6). H invokes philosophy elsewhere as
a consolation for political discomfiture; see 2. 2. 17 ff. (quoted above)
which sympathizes with Tiridates, the defeated Roman candidate for
the Parthian throne, and 4. 9. 30 ff., which vindicates the virtus of
Lollius (‘consulque non unius anni’), who had suffered a military defeat,
and in the literal sense had been consul only once.
18. intaminatis fulget honoribus: the good man who has been denied
honores (magistracies) shines with honores (of a less material sort); cf.
Claud. 17. 1 ff. ‘ipsa quidem Virtus pretium sibi, solaque late / Fortunae
secura nitet. nec fascibus ullis / erigitur, plausuve petit clarescere vulgi’.
intaminatus is not otherwise attested in the classical period; for instances
from Christian Latin cf. TLL 7. 1. 2069. 73 ff. intaminatis fulget makes a
contrast with the immediately preceding sordidae; so Wade’s interminatis
is without merit. For the imagery cf. Cic. Sest. 60 ‘(virtus) lucet in
tenebris . . . splendetque per sese semper neque alienis umquam sordibus
obsolescit’ (of Cato, with heavy irony, but no doubt reflecting things
that were said seriously).
19–20. nec sumit aut ponit securis / arbitrio popularis aurae: cf. epist.
1. 16. 33 f. (of the populace) ‘si / detulerit fascis indigno, detrahet
idem’, Vell. 2. 33. 3 (on Pompey’s magistracies) ‘ut quod cupisset
arbitrio suo sumeret, alieno deponeret’. The fasces were the insignia
imperii (Mommsen, Staatsrecht 1, edn. 3, 373 ff.), and as such were carried
at all stages by the lictors of Augustus; the sentence may contain
a rebuke for anyone who genuinely wished to restore the Republic.
It seems odd, however, that H should refer not to the rods but to the
axes, which were not normally carried within the city; perhaps the
passage reflects discussions about the government of the provinces,
where they were carried. (By the settlement of 27 bc, which may
come after our poem, Augustus kept control of Spain, Gaul, and
Syria, while nominally handing the other provinces back to the people.)
For the fickle breeze of popular favour cf. epist. 1. 19. 37 ‘non ego
ventosae plebis suffragia venor’, Cic. har. resp. 43 ‘popularis aura pro-
vexit’, Liv. 3. 33. 7 with Ogilvie, Virg. Aen. 6. 815 ‘nimium gaudens
popularibus auris’, Sen. Hf 169 ff. with Billerbeck. As arbitrio suggests
a legal adjudication or at least a serious decision, it is paradoxically
combined with aurae.
21–2. virtus recludens immeritis mori / caelum: the repetition of virtus
(17 and 21) recalls the sacral anaphora common with tu and ille; cf. Soph.
Ant. 781 f. ! ¯æø IŒÆ
Æ, = ! ¯æø , n K ŒÆØ Ø ,
Philodemus, anth. Pal. 10. 21 (¼ 8 Sider). Wilkinson (125) objects that
the image is confused and unreal: in 21 ‘Virtus seems to be a goddess
opening the gates of heaven for her devotees’, while in 22 ‘she represents
the man of virtue entering them.’ In fact virtus in both 17 and 21 is the
quality of the man himself, not an external power; for recludere of heroic
achievements cf. Sen. dial. 11. 15. 5 (of Drusus) ‘intima Germaniae
recludentem’, OLD s.v. 2b. For the gates of heaven cf. Hom. Il. 5. 749,
Enn. var. 24 ‘mi soli caeli maxima porta patet’, Virg. georg. 3. 261,
Genesis 28: 17. immeritis mori appears to be the first recorded instance
of immeritus followed by an infinitive.
It was a commonplace that heroes like Hercules, the Dioscuri, and
Romulus attained immortality because of their services to mankind (e.g.
3. 3. 9-10 n., epist. 2. 1. 5 ff., Cic. nat. deor. 2. 62 with Pease, Tusc. 1. 28, 4.
50, off. 3. 25, Sen. HO 1942 f.). Hellenistic and Roman writers also spoke
of similar rewards for great leaders; cf. 3. 3. 11 and epist. 2. 1. 15 ff.
(Augustus) with Brink’s bibliography, Cic. rep. 6. 26 (great men) ‘si
quidem bene meritis de patria quasi limes ad caeli aditum patet’ with
Powell, Virg. Aen. 6. 129 f. ‘pauci quos aequus amavit / Iuppiter aut
ardens evexit ad aethera virtus’, Sen. dial. 6. 25. 1, Luc. 9. 6 ff. (Pompey)
‘quodque patet terras inter lunaeque meatus / semidei manes habitant
quos ignea virtus / innocuos vita patientis aetheris imi / fecit et aeternos
animam collegit in orbes’, Joseph. bJ 6. 5. 47 (on those who die in battle),
Suet. Jul. 88 (Caesar). In particular, H reflects the doctrine of the Stoics,
of whom some (like Cleanthes) maintained that all souls survive till
the next conflagration, while others (like Chrysippus) confined that
privilege to the elect (Diog. Laert. 7. 157); cf. Diog. of Oenoanda,
Smith, fr. 39, III. 13–IV. 12, in particular a b H IıH PŁø
a c ØŒæØØ F Æ ŁæŁÆØ ºªıØ, a b H
ıÆø KØØÆØ IæH. Such ideas have been associated with
the alleged eschatology of Posidonius (Reinhardt, RE 22. 1. 778 ff.),
but there are many uncertainties ( J. F. Dobson, CQ 12, 1918: 182 ff.,
M. Laffranque, Posidonius d’Apamée, 1964: 519 ff.). In H’s case there is no
need to look beyond traditional Stoicism as coloured by the imagination
of Cicero.
H follows the tradition when speaking as the national vates. Else-
where he says that great men survive only in their fame (which in turn
depends on their being celebrated by a major poet); cf. 4. 8. 29 ‘caelo
Musa beat’, 4. 9. 25 ff., Lyne, 1995: 209 ff. This idea was already familiar
from early Greek poetry; cf. for instance Tyrt. 12. 31 ff. P Œº
KŁºe I ººıÆØ P Z ÆPF, = Iºº e ªB æ Kg ªÆØ
IŁÆ , Simonides, anth. Pal. 7. 251. 3 f. ¼ Campbell, epig. ix (probably
on the Greek dead at Plataea) Pb ŁAØ ŁÆ , K Iæc
ŒÆŁæŁ = ŒıÆı IªØ Æ K #ø.
22. negata temptat iter via: temptat (‘ventures’) suggests boldness and is
less tentative than ‘tries’; cf. epist. 1. 17. 34 ‘caelestia temptat’, Virg. georg.
3. 8 f. ‘temptanda via est qua me quoque possim / tollere humo’, Rhianus
1. 15 (Powell, Coll. Alex. ). iter (literally ‘a going’) means the same as ire
(the unmetrical reading of one side of the tradition). negata . . . via
implies ‘by a way denied to ordinary mortals’; cf. 1. 3. 35 ‘pennis non
homini datis’, Pind. Pyth. 10. 27 ›
ºŒ PæÆe h IÆe
ÆPfiH, Sen. Phaedr. 224 ‘solus negatas invenit Theseus vias’, Claud. 26. 69 f.
23–4. coetusque vulgaris et udam / spernit humum fugiente penna: the

wet ground suits grosser natures; cf. serm. 2. 2. 79 ‘atque adfigit humo
divinae particulam aurae’, Ar. nub. 232 f. with Dover, Plat. Phaed. 109b
rÆØ ªaæ ÆÆ
fi Ð æd c ªB ººa ŒEºÆ ŒÆd ÆÆa . . . N L
ıææıŒÆØ oøæ ŒÆd c ›
º ŒÆd e IæÆ (in contrast with
the stars and the aether), Cic. Tusc. 1. 42, nat. deor. 2. 17 with Pease, Virg.
Aen. 6. 730 ff. ‘igneus est ollis vigor et caelestis origo / seminibus,
quantum non noxia corpora tardant / terrenique hebetant artus mor-
ibundaque membra’, Val. Fl. 1. 10; for the fiery nature of the divine see
Pease on Cic. nat. deor. 1. 39, 2. 40. spernit suggests not just departure
(1. 30. 2 ‘sperne dilectam Cyprum’) but disdain (2. 16. 39 f. ‘malignum /
spernere volgus’) and soaring flight (Milton, PL 7. 421 f. ‘Soaring the air
sublime / With clang despised the ground’). fugiente penna recalls the
flight of Justice (Arat. phaen. 1. 133 f.) and H’s metaphorical ascension in
the guise of a swan (2. 20. 3 f. ‘neque in terris morabor / longius’).
25–6. est et fideli tuta silentio / merces: H is imitating the pattern of

Simonides, PMG 582 Ø ŒÆd ØªA IŒı ªæÆ (see especially
Oates, Davis, and Stoneman, opp. citt.); Stoneman takes ªæÆ not as
the reward for silence paid by the laudandus but as the tribute of silence
offered by the laudator. So when Simonides points to the danger of
saying too much, he may simply be using a conventional means
of breaking off an encomium (thus Pind. I. 1. 60 ff. Æ
KØE . . . IÆØæEÆØ æÆ
f æ
ø = o . q a ººŒØ ŒÆd
e ø- = Æ PŁıÆ ø æØ). Pindar, however, says else-
where that ‘the ways of silence are most reliable’ because assertive words
provoke quarrels (fr. 180 Ł ‹ Ø ÆÆØ ØªA › with C. M.
Bowra, Pindar, 1964: 359 f.); that is how Simonides was understood by
Augustus (ap. Plut. regum . . . apoph. 207c), Aelius Aristides (orat. 3. 97
Behr), and Libanius (decl. 15. 4); cf. also Plut. de tuenda san. 125d ŒÆŁæ
› &Øø ºª ÆPfiH ÆºBÆØ ØªÆØ, ŁªÆfiø
b ººŒØ .
By the addition of fideli H gives the aphorism a new direction. In
view of the following sentence on the rites of Ceres (26–7 n.) he must at
one level be referring to the betrayal of religious mysteries (cf. Virg. Aen.
3. 112 ‘fida silentia sacris’, Apul. met. 3. 15. 4 ‘sacris . . . initiatus profecto
nosti sanctam silentii fidem’, Tert. apol. 8); but after the superhuman
political achievements of the previous stanzas this is too peripheral an
issue to be given such prominence. Nor can we argue that by mention-
ing one aspect of piety H is commending piety in general (Heinze,
Syndikus); in the Augustan context he could have found something
more significant than the mysteries, and the formidable sanctions of the
last sentence (29–32) had only a limited application in the tolerant world
of Roman religion. Pasquali (678 ff.) suggests that H is declining to
reveal secrets about the immortality of Augustus, but that was no secret,
and would not have deserved punishment (30–2); the case is quite
different at Pind. N. 5. 16 ff., for there the poet is keeping silent about
an unhappy legend. Williams, loc. cit., sees fidele silentium as an analogy
for assent to Augustan legislation (one might compare Sopater in
Walz, Rhet. Graec. 8. 119 e ØøA ŒÆd æØ a Ææa H ø
æÆ Æ ŒıæE b f ı . . . ); but the observance of reli-
gious secrets seems a very different sort of silence.
It appears more probable that the mysteries here are an analogy for
state secrets; this keeps up the political reference of the previous stanzas.
arcana imperii in Tacitus and elsewhere (Goodyear on ann. 2. 36. 1) is a
religious metaphor; cf. also 1. 28. 9 ‘Iovis arcanis Minos admissus’. In the
same way mysteria is used in secular contexts (TLL 8. 1757. 9 ff.); cf.
Lucil. 651 f. ‘at enim dicis ‘‘clandestino tibi quod commissum foret, /
neu muttires quicquam neu mysteria ecferres foras’’ ’ with Marx, Porph.
on 1. 18. 12 f. ‘non proferam in publicum mysteria tua, inquit, latentia
alioquin ac secreta. sed per allegoriam hoc videtur dicere: ea quae fidei
meae commissa sunt, numquam ego ebrius factus prodam’, [Plut.] de lib.
educ. 10 f ŒÆd Øa F Ø ŒE a ıæØØ ºa ƒ ÆºÆØd
ŒÆØÆ, ¥ K ÆÆØ ØøA KŁØŁ Kd c H IŁæøø
ıæø Ø e Ie H Łø Ææø , de garrul. 505 f.
Augustus notoriously kept people guessing about his intentions (see
Suet. Aug. 50 for the sphinx on his seal-ring), and H repeatedly men-
tions the importance of discretion in both public and private life; cf. 1.
18. 16 ‘arcanique Fides prodiga’, serm. 1. 3. 95, 1. 4. 84 f. ‘commissa tacere /
qui nequit, hic niger est, hunc tu, Romane, caveto’, 2. 6. 46, 2. 6. 57 f.,
epist. 1. 5. 24 f., 1. 18. 37 ‘arcanum neque tu scrutaberis illius umquam’, 1.
18. 70, ars 200.
H’s tuta merces has therefore a different implication from the
IŒı ªæÆ of his model. Simonides seems to have meant that
silence is risk-free: no offence can be caused if you say nothing. On
the other hand, with tuta merces the reward for discretion is assured
(just like the reward for virtus and the punishment for cowardice and
indiscretion). The reward is not specified; it could be position or esteem,
but may simply be the consciousness of duty done; cf. e.g. 2. 2. 21 where
the diadema tutum cannot be taken away because it exists only in
the mind.
26–7. vetabo qui Cereris sacrum / vulgarit arcanae: the mysteries of

Demeter and Persephone at Eleusis in Attica included the enactment of
the myth, which ‘explained’ how corn remained below the earth for part
of the year and above it for the rest (see Burkert 276–8, 285–90); charges
of impiety were brought against those who betrayed the rituals (Aes-
chylus, Andocides, Alcibiades, and Diagoras among others). Following
the precedent of earlier notabilities like Cicero and Atticus (leg. 2. 36),
Octavian was initiated in 31 bc (Dio 51. 4. 1), and observed the duty of
secrecy; cf. Suet. Aug. 93 ‘Athenis initiatus, cum postea Romae . . . de
privilegio sacerdotum Atticae Cereris cognosceret et quaedam secretiora
proponerentur dimisso consilio . . . solus audivit disceptantes’. H, too,
seems to be referring to the Greek Demeter (cf. serm. 2. 8. 13 f. ‘ut Attica
virgo / cum sacris Cereris’). A temple of Ceres was dedicated in Rome
in 493 bc (Dion. Hal. ant. Rom. 6. 17. 2–4, Tac. ann. 2. 49., Scullard
102 f., Richardson 80 f.), but her mysteries were probably confined to
women (Cic. leg. 2. 21, 2. 37, H. Le Bonniec, Le Culte de Cérès à Rome,
1958: 423 ff., B. S. Spaeth, The Roman Goddess Ceres, 1996: 20, 105 f.) and
played a much less important part in religious life than the Eleusinian
mysteries.
vetabo is future in tense because the poet is warning about future
possibilities; for somewhat similar usages cf. epist. 1. 14. 44 ‘censebo’, 1.
19. 8 f. ‘forum putealque Libonis / mandabo siccis, adimam cantare
severis’ (edixi in v. 10 hints at the praetor’s edict), ars 317 ‘iubebo’,
Löfstedt 2. 127 ff., H–Sz 850. The following subjunctive (sit in v. 28) is
more legalistic than the usual infinitive; cf. epist. 2. 1. 239 f. ‘edicto vetuit
ne quis se praeter Apellen / pingeret’, Pers. 1. 112 ‘hic . . . veto quisquam
faxit oletum’. Bentley considered reading arcanum (still an adjective),
and though he judged it unnecessary it should be left on the record: it is
combined elsewhere with sacrum (epod. 5. 52 ‘arcana cum fiunt sacra’, Ov.
her. 12. 79 ‘arcanaque sacra Dianae’, Stat. silv. 3. 3. 65 f.), it makes a
sharper contrast with vulgarit, and it brings out more clearly the analogy
with state secrets. Normally with enallage (transferred epithet) the
genitive adjective is transferred to the noun on which it depends (see
the references collected by Bo 134); for another unusual instance like the
present cf. 3. 24. 44 ‘virtutisque viam deserit arduae’ (with note).
27–9. sub isdem / sit trabibus fragilemque mecum / solvat phaselon:

for the wish not to share a house with the impious cf. Soph. Ant. 372 ff.
Kd ÆæØ = ªØ Y æH = n æØ, Antiphon
5. 11, Plat. rep. 3. 417a, Call. h. 6. 116 f. ˜Ææ, c B Kd º , ‹
Ø I
Ł , = Y › Ø
. For the wish not to sail in the same
boat cf. Aesch. sept. 602 ff., Eur. Elect. 1354 f. oø IØŒE d
Łºø = KØ æŒø Æ ıºø, fr. 852. 5 N edn. 2, Xen. Cyr.
8. 1. 25, Petr. 105. 1. For the combination of both sentiments cf. Greg.
Naz. PG 37. 907. 309 f. (a reference supplied by Mr A. S. Hollis). There
was a story that Simonides was warned in a dream not to sail in a ship
that later sank (PMG 653. 1, Cic. div. 1. 56 with Pease, Val. Max. 1. 7 ext.
3, Oates, op. cit. 4 ff.); and when the hall-roof of the Scopadae fell
down, the poet again miraculously escaped (PMG 510, Call. aet. fr. 64.
11 ff., Cic. de orat. 2. 352–3, Quint. 11. 2. 11 ff., Oates, op. cit. 2 ff., 9 ff.).
The combination of the collapsing roof and the shipwreck may again be
suggested by something in Simonides; but the context must have been
different, for neither the sailors nor the Scopadae were guilty of impiety,
and Simonides was not killed as a result of associating with them.
A phaselos was a boat shaped like a bean-pod; as it had a shallow
draught it was unstable (Ov. am. 2. 10. 9 ‘erro velut ventis discordibus
acta phaselos’); for fragilem cf. 1. 3. 10 f. ‘fragilem . . . ratem’. It seems
originally to have been associated with the Nile (cf. Virg. georg. 4. 289),
yet the name was also used of a sea-going vessel for the conveyance of
passengers; cf. Catull. 4. 1 ff. with Fordyce, Cic. Att. 1. 13. 1 (a voyage to
Epirus), L. Casson, Ships and Seamanship in the Ancient World, 1971:
167 f., RE 19. 2. 1883 f.
29–30. saepe Diespiter / neglectus incesto addidit integrum: Jupiter

punished the impious (1. 3. 40, 1. 12. 59 f.) and in particular oath-
breakers; for neglectus in religious contexts cf. 3. 6. 7. The archaic
Diespiter is used with solemn effect; cf. 1. 34. 5, Paul. Fest. 102L ‘si sciens
fallo, tum me Dispiter salva urbe arceque bonis eiciat, ut ego hunc
lapidem’. The incestus is the man polluted by sacrilege (ars 472, Tib.
1. 2. 81 f. ‘num feror incestus sedes adiisse deorum / sertaque de sanctis
eripuisse focis?’); such miasma was contagious (see Parker 4, and cf. Liv.
39. 9. 1). The integer, on the other hand, was untainted (1. 22. 1 ‘Integer
vitae scelerisque purus’). For the punishment of the innocent with the
guilty cf. Hes. op. 240 with West, Antiph. 5. 82, Virg. Aen. 1. 39 ff.,
Genesis 18: 23–33. addidit is a ‘gnomic’ perfect, applied to things that have
happened in the past and are likely to happen again (K–S 1. 132, H–Sz
318 f.). saepe, like the balancing raro (31), is common in such contexts.
31–2. raro antecedentem scelestum / deseruit pede Poena claudo:

Poena is here personified, like the avenging Fury in Greek; cf. Aeschin.
1. 190 c ªaæ YŁ . . . f MŒ Æ ŒÆŁæ K ÆE æÆªfiøÆØ
—Øa KºÆØ, Tib. 1. 9. 4 ‘sera tamen tacitis Poena venit pedibus’ with
3 . I V S T V M E T T ENAC EM P RO P O S I T I V I RV M 35
Murgatroyd. Divine punishment is conventionally slow but sure (Hom.
Il. 4. 161, Plut. de sera numinum vindicta, Otto 111), and so Justice is
described as slow-footed; cf. Eur. TGF 979. 3 f. EªÆ ŒÆd æÆE
d =
ıÆ æłØ f ŒÆŒf ‹Æ
fi , Val. Max. 1. 1 ext. 3.
26 f. ‘lento enim gradu ad vindictam sui divina procedit ira’, Strato, anth.
Pal. 12. 229. 2 æ ı . . . ˝Ø. The image of limping in such a
context is found in Hom. Il. 9. 502 ff. (the limping Supplications that
follow after Ate), but there is no recorded parallel to H’s limping
Retribution.
deseruit means ‘lets go of ’, ‘cannot keep up with’ (TLL 5. 1. 673. 66 ff.
‘a comite vel duce discedere . . . non sequendo’, 2. 16. 22 ‘nec turmas
equitum relinquit’). For the unremitting pursuit cf. Pease on Virg.
Aen. 4. 471 (the Furies), Coleridge, Ancient Mariner part 6 ‘Because he
knows a frightful fiend / Doth close behind him tread’; in Horace’s raro
deseruit, however, there is an added irony in that the faithless man is
unable to shake off such a faithful companion. The word-order pro-
duces an effect worthy of Hitchcock: first we see a man walking—a
criminal—then behind him the foot—Retribution’s foot—limping.
3 . I V S T V M E T T ENAC EM
PROPOSITI VIRVM
[P. Ceauşescu, Historia 25, 1976: 79 ff.; D. C. Feeney, CQ 34, 1984: 179 ff.; W. Warde Fowler,
Roman Essays and Interpretations, 1920: 216 ff.; Fraenkel 267 ff.; E. L. Harrison ap. Wood-
man and West (1984), 95 ff.; S. Harrison, Entretiens Hardt 39, 1993: 141 ff.; Lyne (1995),
208 ff.; L. P. E. Parker, MH 59, 2002: 101 ff.; Pasquali 681 ff.; H. T. Plüss, Horazstudien,
1882: 211 ff.; Wilkinson 73 f.]
1–8. The just and steadfast man cannot be shaken by any kind of threat.
9–18. Through such qualities divinity was achieved by heroes like Hercules
(among whom Augustus will take his place), and Romulus, when Juno had
given her consent in a speech that proved agreeable to the gods: 18–44.
‘Perfidious Troy was destroyed by Paris and Helen; now that the war is
over I give up my resentment and agree that Romulus should be enrolled
among the gods, provided that Rome and Troy are kept apart; provided that
Troy remains desolate, Roman power may stand firm. 45–68. Let Rome
extend her empire from Spain to Egypt, disregarding mineral wealth, and
from the torrid zone to the rainy north, so long as she refrains from rebuild-
ing Troy; if Troy is rebuilt I shall destroy it again’. 69–72. But cease, Muse,
to diminish such epic themes in the slight mode of lyric.
Firmness of purpose was a virtue particularly prized by the Stoics; in
this they drew on Plato’s account of Socrates, who did not yield to
pressure from either tyrants or the demos (2–3 nn.). Cicero (Tusc. 3. 9)
states that constantia had an important place in the Romans’ traditional
value-system (for examples see Val. Max. 3. 8 and Sen. dial. 2); that
quality is particularly evident in their idealized accounts of the younger
Cato (cf. Parker, loc. cit.). In the first two stanzas there is no direct
reference to Augustus, who was not threatened by tyrants (3–4 n.). But
in the third stanza Pollux and Hercules, and in the fourth Bacchus and
Romulus, are said to have been deified for their steadfastness (hac arte);
so when Augustus is pictured as reclining in their company (11–12), he
was clearly expected to achieve such glory through the same virtue. It is
true that the Princeps’ divinity lies only in the future, and that the poet
concentrates on the apotheosis of Quirinus (i.e. Romulus); but since
Augustus regarded himself as Rome’s second founder and at one time
thought of assuming the name of Romulus (Suet. Aug. 7, Dio 53. 16. 7),
analogies were bound to suggests themselves (cf. Binder 162 ff.). We
shall argue that the steadfastness of Augustus is highly relevant to the
purpose of the ode as a whole.
The mention of Quirinus leads to the heart of the poem (for such a
transition cf. 3. 11. 25). There Juno makes a speech to the council of the
gods in which she agrees to the apotheosis of Romulus (35 f.) on
condition that Troy should not be rebuilt (37 ff., 58 ff.). The apotheosis
itself was guaranteed by Jupiter, as we learn from the Annals of Ennius
(1, fr. xxxiii with Skutsch); for the context see Ov. met. 14. 812 ff. where
Mars reminds Jupiter of his promise: ‘tu mihi concilio quondam prae-
sente deorum / . . . ‘‘unus erit quem tu tolles in caerula caeli’’ / dixisti’; cf.
15–16 n. It is probable that Juno also consented to the apotheosis of her
grandson in this debate (Feeney, op. cit. 181 f.). Such a speech seems to
be the common source behind Horace’s ode and Virg. Aen. 12. 827 f.,
where Juno says ‘sit Romana potens Itala virtute propago’, adding
‘occidit, occideritque sinas cum nomine Troia’; in other words the
Trojans may settle in Italy provided they abandon their distinct identity;
such a limited détente had been foreshadowed by Jupiter at Aen. 10. 8 ff.
But Juno would not be fully reconciled to Rome until she withdrew her
support from Carthage in the Second Punic War (Skutsch on Enn. ann.
8, frr. xv and xvi, Virg. Aen. 1. 179 ff., E. L. Harrison, loc. cit.). In a
different context Horace actually extends her support for Carthage until
46 bc (2. 1. 25–6 with N–H), but in the present poem we are not meant
to remember Carthage at all; otherwise Juno would not have sanctioned
the spread of Roman power from the Straits of Gibraltar to the Nile
(46 ff.).
In the ode Juno insists that the Romans must not rebuild the original
Troy in the Troad (40 ff.); this stipulation cannot have been made in
Ennius (for at the time of Rome’s foundation there was no possibility of
re-establishing contact with Troy), so we must look for another explan-
ation. Suetonius says of Julius Caesar (79. 3): ‘quin etiam varia fama
percrebruit migraturum Alexandream vel Ilium, translatis simul opibus
imperii . . . et procuratione urbis amicis permissa’; cf. also Nicolaus of
Damascus, FGrH 90 F130. 20 (II. 1, p. 404): some said Caesar intended
to make Egypt the centre of empire, ƒ K + ºfiø F ÆÆ ÆPe
ººØ ŒÆŁÆŁÆØ Øa c ÆºÆØa æe e ˜ÆæÆØH ª
ıªªØÆ, Lucan 9. 998 f. (Caesar prays to the dead Trojans) ‘restituam
populos; grata vice moenia reddent / Ausonidae Phrygibus, Romanaque
Pergama surgent’. This rumour, though often in the past thought
significant, is now usually discounted (Fraenkel 268, nn. 1–5), but Caesar
might have thought of establishing a secondary centre of power in the
East once he had conquered Parthia. It is also significant that Livy gives
prominence to a speech by Camillus in which he rejects a proposal to
move the seat of government to Veii (5. 51–4 with Ogilvie); there was
nothing new about this speech (see Enn. ann. 4 fr. v with Skutsch, pp.
314 f. and Plut. Cam. 31–2), but as Livy was writing around the same
time as Horace, it is tempting to look for a contemporary context in
both passages and also in Virg. Aen. 12. 827 f. quoted above. Such fears
were not new; they were foreshadowed by the scare-mongering allega-
tions in Cic. de leg. agr. 1. 18, where the orator attacks in highly coloured
terms a proposal to send a colony to Capua: ‘Capuam deduci colonos
volunt, illam urbem huic urbi rursus opponere, illuc opes suas deferre et
imperi nomen transferre cogitant’, cf. 2. 96.
Obviously Augustus had no intention of abandoning Rome as the
capital of the empire (if he had, Horace would not have denounced the
idea). Yet others may have canvassed less drastic proposals; the richest
and most populous parts of the empire were in Asia, so attractive
economic arguments may have been put forward for a redeveloped
Troy, and in view of its accessiblity from the Via Egnatia the place
may have been thought a promising site for a secondary centre of power.
In 22 Agrippa, Augustus’ loyal general, held a command in Lesbos,
from where he could exercise authority over a wide area and keep a
watch on the sea-routes ( Joseph. aJ 15. 350, Dio 53. 32. 1, J.-M. Roddaz,
Marcus Agrippa, 1984: 425, Lacey 117 ff.); we find him again in the East in
16–13, and his name occurs in an undatable inscription from Ilium
(P. Frisch, Die Inschriften von Ilium, 1975: no. 86). D. Magie’s comment
is significant: ‘The mere creation of this command and the precedent
established for the division of power over the West and the East
between emperor and associate were of far greater importance than
anything that was done by the first incumbent of the office’ (Roman
Rule in Asia Minor, 1950: 1. 468–9). Yet, although the buildings of
various cities were repaired and their economy restored (for Troy, see
C. B. Rose, Studia Troica 2, 1992: 45; and 4, 1994: 75), Augustus never set
up another administrative centre of the kind envisaged. The careers of
Sulla and Pompey had shown how the extension of Roman power could
lead to the emergence of rich and powerful figures with dangerous
ambitions. Antony, now inseparably linked with civil war, had staged
a ceremony at which he handed over large territories in the East to
Cleopatra and her children (Dio 49. 32. 4 f.); worse still, he had
intended, allegedly, to move the capital to Alexandria—a plan which
even his friends viewed with dismay (Dio 50. 4. 1 f.). Augustus’ refusal to
countenance such an idea was vindicated by history; for when Constan-
tine eventually founded a New Rome in Byzantium, the consequences
for the unity of the empire were disastrous; it is worth noting that his
original choice had been Ilium (Zosimus 2. 30. 1). For these matters and
further documentation see Ceauşescu, loc. cit.
Other explanations of the poem may be briefly recorded. Fraenkel
(loc. cit.) refuses to admit a contemporary relevance (just as he does with
3. 5 on Regulus); he sees Juno’s speech as providing just a vivid sketch of
the goddess’s intransigent character, but that does not explain why the
continuing insignificance of Troy should be given such a prominent
place in the Roman Odes. Some see a criticism of Antony’s oriental
policy; but though the ode hints at a correspondence between Helen
and Cleopatra (25 f.), Troy was too different from Alexandria to make an
extended analogy plausible. Others regard Troy, which was notoriously
rich (Enn. trag. x x v i i , 89 ff. J with Jocelyn’s notes on 90 and 91, Virg.
Aen. 2. 763 ff.) and corrupt (21–2 n.) as representing the materialism of
the fallen Republic (cf. Plüss and Wilkinson, locc. citt.); this suits the
denunciation of gold-mining at 49–52, but that single stanza can hardly
be given such importance in the scheme of the ode as a whole. Such a
view, like the supposed allusion to Antony, does nothing to explain the
corresponding passages in Livy and especially Virgil (Aen. 12. 827 f.); and
it is difficult to regard ‘Ilium’ as merely a metaphor in view of Rome’s
continuing interest in the actual site (cf. Rose, loc. cit.).
The theory proposed above offers an explanation of the poem’s length
and gravitas and its position alongside no. 4. It accounts for the import-
ance of Romulus, and allows room for the attack on acquisitiveness
(49–52 n). Above all, Horace’s praise of steadfastness under pressure
makes sense if Augustus is resisting proposals that would impair the
integrity of the empire and his own position as Rome’s second founder
(cf. Warde Fowler, op. cit. 217 ff.).
Metre: Alcaic.
1. Iustum et tenacem propositi virum: iustitia (‘right behaviour’) was

put by Cicero at the centre of morality: cf. off. 1. 20 ‘iustitia in qua
virtutis splendor est maximus ex qua viri boni nominantur’. It involved
not just the avoidance of unfair dealing but a readiness to stand up for
principle and show steadfastness in adversity: cf. Cic. off. 2. 38 ‘nemo
enim iustus esse potest qui mortem, qui dolorem, qui exilium, qui
egestatem timet’. iustitia also occupied a central place in the Romans’
political theory (see the debate in Cic. rep. 3. 8 ff.) and figured promin-
ently in the presentation of their imperial power; thus Caesar claimed
that Roman imperium in Gaul was iustissimum (bG 1. 45. 3) and that he
would win the civil war by iustitia et aequitate (bc 1. 32. 9); similarly the
golden shield presented to Octavian in 27 bc by the senate and people
contained the words ‘virtutis clementiaeque et iustitiae et pietatis caussa’
(res gest. 34. 2 with the introduction to 3. 2 and Weinstock 243 ff.).
propositum can refer not only to a general principle but also to one’s
purpose in a particular situation; hence somebody who holds fast to his
purpose is said tenere propositum (Caes. bc 1. 83. 3, Ov. ars 1. 470);
similarly tenax propositi at Ov. met. 10. 405, Val. Max. 6. 3. 5. For H’s
steadfast man cf. Cic. Lig. 26 ‘eius viri quem de suscepta causa propo-
sitaque sententia nulla contumelia, nulla vis, nullum periculum posset
depellere’.
2. non civium ardor prava iubentium: iubentium refers not just to the
clamour of the mob but to resolutions of an assembly, as in the formulae
velitis iubeatis, Quirites and populus iussit; in RN’s view such a neuter
plural has a Greek tinge that suits philosophical discourse (cf. 1. 29. 16
‘pollicitus meliora’). For criticisms of popular emotion cf. Plat. apol. 32b
(Socrates resists the prosecution of the generals after Arginusae), Cic.
rep. 1. 65–7 (citing Plat. rep. 8. 562 ff.), Sen. dial. 2. 1. 3 (Cato the Younger
set on by the mob); a general exposition of the citizens’ influence is
given by F. Millar, The Crowd in Rome in the Late Republic, 1998. For
things that cannot affect the wise man cf. serm. 2. 7. 83 ff., Sen. Ag. 596 ff.
with Tarrant, Thy. 350 ff.; Syndikus 2. 36 n. 22 comments on the Stoic
colouring of the first two stanzas.
3–4. non vultus instantis tyranni / mente quatit solida: the tyrant no
less than the demos must be resisted, as shown again by Socrates (Plat.
apol. 32c–d, Epictet. 4. 1. 164); for illustrations of the stock situation
cf. epist. 1. 16. 73 ff., M. Winterbottom, index to Seneca the Elder s.v.
‘Tyrants’, Juv. 10. 307 with Mayor, Epictet. 1. 18. 17, 1. 19. 7–8, Philostr.
vit. Apoll. 7. 4. The tyrant’s threatening mien is referred to at Soph. OT
447 f. P e e = Æ æ ø, Cic. off. 1. 112 (on Cato) ‘(cum)
semperque in proposito susceptoque consilio permansisset, moriendum
potius quam tyranni vultus aspiciendus fuit’, Sen. HO 1992 ff. (with
Nisbet, 1995: 211); instantis means ‘louring’, cf. Lucr. 1. 64 f. (on religio)
‘quae caput a caeli regionibus ostendebat / horribili super aspectu
mortalibus instans’. For mente solida cf. Sen. dial. 2. 3. 5 ‘quemadmodum
proiecti quidam in altum scopuli mare frangunt . . . ita sapientis animus
solidus est’. Though parallels are lacking, mente is probably a separa-
tive ablative, as with the compound excutere (cf. [Aesch.] Prom.
360 f. [ŒæÆıe ] ÆPe Kº H łª æø = ŒÆø; Heinze
regards it as an ablative of respect as with animo perturbari, but solida
makes this less likely.
4–5. neque Auster / dux inquieti turbidus Hadriae: after two illustra-
tions from human life come two from the natural world. For the menace
of the south wind in the Adriatic cf. 1. 3. 14 ff. ‘rabiem Noti, / quo non
arbiter Hadriae / maior’, RE 1. 418; but while arbiter implies legal
authority, dux turbidus suggests a turbulent demagogue who stirs up a
restless mob. For the personification cf. also 2. 17. 19 f. ‘tyrannus /
Hesperiae Capricornus undae’; similarly the French mistral is derived
from magistralis. Hadria for the Adriatic is masculine, like #æÆ
in Greek.
6. nec fulminantis magna manus Iovis: the sky-god’s association with

the thunderbolt is shown by the magnificent bronze statue recovered off
Artemisium ( J. Boardman, Greek Sculpture and Painting, 1985: fig. 35);
for the grandiloquent magna manus cf. Hom. Il. 15. 695
Øæd ºÆ
ªºfi . In Rome Jupiter had the standing epithet Tonans (3. 5. 1 n.);
in the same way fulminantis may represent Iæ . (A temple to
Jupiter Tonans was dedicated in 22 bc to celebrate Augustus’ escape
from lightning in the Cantabrian War (L. Richardson 226 f.), but when
this ode was written he was presumably still in Rome.) The Stoic
Seneca commends indifference to the forces of nature; cf. epist. 79. 10,
nat. quaest. 6. 32. 4 ‘securus aspiciet fulminantis caeli trucem atque
horridam faciem, frangatur licet caelum et ignes suos in exitium
omnium, in primis suum, misceat’, Thy. 358 ff. Epictetus says that not
even Zeus can overcome the good man’s moral purpose (1. 1. 23).
7–8. si fractus illabatur orbis / impavidum ferient ruinae: orbis here ¼

caeli orbis (Ov. am. 1. 8. 10, TLL 9. 2. 913. 83 ff.) rather than ‘the universe’
(Ov. met. 1. 6), certainly not ‘the world’, which does not suit illabatur; cf.
Theog. 869 Ø ØÆ Ø ªÆ PæÆe Pæf oæŁ, Ter. heaut.
719 ‘quid si nunc caelum ruat?’, Sen. HO 1385, Otto 61 f.; cf. the saying
‘fiat iustitia ruat caelum’, and Addison’s translation ‘Should the whole
frame of nature round him break, / In ruin and confusion hurled, / He
unconcerned would hear the mighty crack, / And stand secure amidst
a falling world’. The present subjunctive in the protasis indicates a
hypothesis, the future indicative in the apodosis implies certainty;
cf. 2. 17. 14 f., epod. 11. 15 ff., K–S 2. 395. The plural ruinae describes the
debris rather than the crash; cf. Sen. nat. quaest. 2. 59. 3 ‘contemne
mortem, et omnia quae ad mortem ducunt contempta sunt, sive illa
bella sunt . . . sive ruinarum subito lapsu procidentium pondera’.
9. hac arte Pollux et vagus Hercules: arte here means ‘quality’ or ‘virtue’
(OLD 4b, Brink on epist. 2. 1. 13), a usage commoner in the plural; cf.
Cic. imp. Pomp. 36, Cael. 77 ‘civem bonarum artium’, Marvell, Horatian
Ode ad fin. ‘The same arts that did gain / A power must it maintain’.
H refers to iustitia in the wide sense that includes constantia (1 n.).
The Dioscuri and Hercules, like Bacchus and Romulus below, were
added to the gods because of their services on earth; cf. 3. 2. 21 n., epist. 2.
1. 5 ff. ‘Romulus et Liber pater et cum Castore Pollux / post ingentia
facta deorum in templa recepti . . . ’ with Brink, Doblhofer 122 f. Pollux
also includes Castor (3. 29. 64 n.). In the Roman tradition the Dioscuri
fought on horseback against the Latins at Lake Regillus (496 bc), and
later brought news of the victory over Perseus in 168 (Cic. nat. deor. 2. 6
with Pease); for Greek encomia cf. Pind. N. 10. 49 ff., Arist. PMG 842.
9 ff. (the hymn to Arete) F (Œ ŒÆd › E = , ˙æÆŒºB ¸Æ
ŒFæØ = ºº IºÆÆ. See further Apollod. 3. 10. 7 and 3. 11. 2 with
Frazer, Pease on Cic. nat. deor. 2. 6, LIMC 3. 1. 567 f., 3. 2. 456 ff.
Hercules had a special place in such tributes; see Diod. 4. 8 ff. (a list
of his feats), Dio Chrys. 1. 84 ff. (his opposition to tyrants), Apollod. 2.
7. 7 with Frazer (his apotheosis), Cic. nat. deor. 2. 62 with Pease; for a
general study see R. Höistad, Cynic Hero and Cynic King, 1948. For
analogies with Augustus see 3. 14. 1 n.
vagus here refers (obviously without any pejorative nuance) to
Hercules’ expeditions; cf. Pind. I. 4. 55 ff., Eur. Her. 1196 f. PŒ i
N (æ = ºı
Ł æ ºıºÆªŒ æ ŁÆH, Virg.
Aen. 6. 801 ‘nec vero Alcides tantum telluris obivit’, Stat. Theb. 8. 516,
silv. 4. 3. 155. For the use of the adjective see Fordyce on Catull. 64. 271.
10. enisus arces attigit igneas: enisus (‘striving upwards’) is obviously

right against the variant innisus (‘leaning on’); cf. Curt. 7. 11. 10 ‘nihil
tam alte natura constituit quo virtus non possit eniti’, Tac. ann. 1. 70. 4
‘in editiora enisus’. When two singular subjects are combined, the verb
in H is normally singular, even where it refers to people; cf. Wickham
on 1. 3. 10, K–S 1. 45. arces describes the heights of heaven, igneas their
fiery nature as shown particularly by the stars; cf. 3. 2. 21–24 n. and our
‘empyrean’, Ov. met. 1. 26 f. ‘ignea convexi vis et sine pondere caeli /
emicuit summaque locum sibi fecit in arce’.
11–12. quos inter Augustus recumbens / purpureo bibet ore nectar: in

January 27 bc Octavian assumed the name ‘Augustus’ (res gest. 34. 2,
Suet. Aug. 7. 2, Dio 53. 16. 7-8). After death he will participate in gods’
banquets; cf. Hom. Od. 11. 602 (Heracles), Theoc. 17. 22 ff. (Ptolemy
Soter joins Alexander and Heracles), Stat. silv. 3. 1. 26 f. (Hebe gives
Hercules a drink of nectar). Here he is given the central position
regardless of chronology. recumbens suggests the reclining posture of
the Romans (epist. 1. 5. 1); a painting from Herculaneum depicts the
deified Vespasian between Hercules and Theseus (K. Schefold, La
Peinture pompéienne, 1972: 201).
ore with bibet must mean the mouth rather than the face (as at Mart.
8. 65. 4 ‘purpureum fundens Caesar ab ore iubar’); it also rules out the
idea (most recently advocated by J. S. C. Eidinow, CQ 50, 2000: 463 ff.)
that H is referring to the painted face of the triumphator. Parallels to
Augustus’ red lips can be cited from Simon. PMG 585 æıæı Ie
Æ = ƒEÆ øa ÆæŁ , Catull. 45. 12 ‘illo purpureo ore
saviata’, Virg. Aen. 2. 593 (of Venus) ‘roseoque haec insuper addidit
ore’; these examples describe female beauty, but RN thinks H could
be suggesting the vitality of the new god. Some, including NR, believe
that Augustus’ mouth is stained by the red nectar (Hom. Od. 5. 92 f., h.
Hom. 5. 206); so S. Pulleyn, Mnem. 50, 1997: 482 f., but RN thinks this is
too fanciful for the context.
bibet is clearly right as against the variant bibit. To suggest that
Augustus is already drinking in such company is absurd, as his apothe-
osis is still in the future; cf. 1. 2. 45 ‘serus in caelum redeas’, Virg. georg. 1.
24 f., Housman on Manil. 1. 800. As the adopted son of Divus Iulius,
Augustus exploited the popular credulity that supported Caesar’s divine
pretensions (N–H on 1. 2, pp. 19 f.). At the same time, in this matter as
in others, he was astute enough to learn from Caesar’s mistakes, and he
did something to appease traditionalists by suggesting that any claims of
his own to divinity lay in the next world. See further 3. 25. 6 n., L. R.
Taylor, The Divinity of the Roman Emperor, 1931: chapters 6–10, ANRW
II. 16. 2. 845, 878 f., OCD 1338 f. with the bibliography appended, and for
a realistic summary A. Wallace-Hadrill, Augustan Rome, 1993, 79 ff.
13–15. hac te merentem, Bacche pater, tuae / vexere tigres, indocili

iugum / collo trahentes: though Dionysus was originally a god in the
full sense, later writers sometimes included him among the heroes for
his gift to mortals (Pasquali 684 ff.); pater (1. 18. 6) is particularly
associated with his Latin name of Liber (epist. 2. 1. 5), but the picture
here is Hellenistic. Commentators understand caelum with merentem (cf.
Ov. ars 2. 218), but the participle may simply mean ‘deservedly’; perhaps
ad igneas arces should be supplied with vexere (‘conveyed’). Dionysus’ car
was sometimes drawn by panthers (LIMC 3. 1. 461 and 463), sometimes
by tigers, which Bacchus supposedly found in India or Hyrcania south
of the Caspian (Plin. nat. hist. 8. 66, Ov. trist. 5. 3. 21 ff.); the animals
will no doubt have become known as a result of the eastern campaigns
of Alexander (Arr. Ind. 15. 1–3, B. Bosworth, JRS 89, 1999: 2 f.); see Virg.
Aen. 6. 805 with Austin, Ov. met. 3. 668 with Bömer. Tigers were sent by
the Indians to Augustus in Samos in 20 bc (Dio 54. 9. 8) and in 11 he
exhibited a tame specimen in Rome (Plin. nat. hist. 8. 65, Toynbee
70 ff.). indocili: the tamed tigers represent the god’s civilizing power
(cf. ars 393 on Orpheus). For the ascension by chariot cf. Prop. 3. 17. 8
‘lyncibus in caelum vecta Ariadna tuis’ (i.e. Bacchi), Ov. trist. 5. 3. 19 ‘ipse
quoque aetherias meritis invectus es arces’; cf. the legends about Rom-
ulus (15–16 n.), Hercules (Ov. met. 9. 272), and Elijah (2 Kings 2: 11); see
also Weinstock 356 f. In RN’s view, because of the hyperbaton and the
word’s position at the end of the line, tuae is emphatic, underlining the
association of the tigers with Bacchus and balancing Martis below.
15–16. hac Quirinus / Martis equis Acheronta fugit: Quirinus, perhaps

the god of the Quirinal (Latte 113), was identified with the deified
Romulus (Pease on Cic. nat. deor. 2. 62, Bömer on Ov. fast. 2. 475),
probably long before Ennius (Dion. Hal. ant. Rom. 2. 63. 3, Weinstock
176 f.), though Skutsch argues for a later date (ann. p. 246). Romulus
was sometimes said to have suddenly disappeared (Cic. rep. 2. 17, Liv. 1.
16 with Ogilvie, Dion. Hal. ant. Rom. 2. 56. 2); other sources speak of his
ascension by the chariot of Mars; cf. Enn. ann. 1, fr. xxxiii (cited in the
introduction), Ov. fast. 2. 496 ‘rex patriis astra petebat equis’, met. 14.
820 ff. For Acheronta fugit cf. 4. 8. 25 ‘ereptum Stygiis fluctibus Aeacum’,
Pind. fr. 143. 2 f. Ææı Æ = æŁe ıª #
æ , Theoc. 17.
46 f., cons. Liv. 242 ‘[ut Remus et Romulus] effugerent aliqua stagna
profunda via’.
17–18. gratum elocuta consiliantibus / Iunone divis: the council of the

gods had been a feature of epic since Homer; see Il. 1. 533 ff., 4. 1 ff., 8.
1 ff., Cic. nat. deor. 1. 18 with Pease, Virg. Aen. 10. 1 ff. with Harrison,
Ov. met. 1. 167 ff. with Bömer, Otto 109, Milton, PL 2. 1 ff. (the debate in
Satan’s assembly). The event was parodied in the first book of Lucilius
(3, 19, 26 M, C. Cichorius, Untersuchungen zu Lucilius, 1908 and 1964:
219 ff., J. H. Waszink, WS 70, 1957: 322 ff.), Sen. apocol. 9 (see Wein-
reich), Lucian, deorum concilium. The debate about the deification of
Romulus is derived from Ennius, but H gives it a new direction (see
introduction).
For the neuter gratum cf. 3. 25. 7 n. The dative consiliantibus goes with
gratum rather than elocuta; the welcome concession emerges at 33 ff. In
classical Latin deponent verbs are seldom found in the ablative absolute
when combined with an object (Schmalz, ALL 1. 344, K–S 1. 783 f.); for
exceptions see Sall. Jug. 103. 7 ‘Sulla omnia pollicito’, Ov. met. 8. 565.
Mythological speeches at the end of a poem are found in epod. 13. 12–
18, carm. 1. 7. 25–32, 4. 4. 50–72 (in the last case followed as here by a
stanza of comment). The feature is derived from Greek lyric; see N–H
on 1. 7. 24 with p. 93.
18–21. ‘Ilion Ilion / fatalis incestusque iudex / et mulier peregrina

vertit / in pulverem: the feminine Ilios is the form used by Homer,
and by Horace elsewhere in the nominative and accusative (epod. 14. 14,
carm. 4. 9. 18); in the dative and ablative he uses the neuter (e.g. 1. 10. 14
‘Ilio . . . relicto’), like later Greek poets and other Romans (Ov. met. 6.
95, 13. 408, etc.). In such contexts the geminatio is pathetic (Eur. Or. 1381
! +ºØ ! +ºØ, þØ Ø, Tro. 173), which is why L. Müller proposed Ilion
impium; but even if Juno is hostile to Troy she can still profess emotion
at the fall of a mighty city.
iudex refers as often to the ‘Judgement of Paris’, where he awarded the
prize for beauty to Aphrodite rather than Hera or Athena; cf. 1. 15. 1 with
N–H, Apollod. epit. 3. 2 with Frazer, RE 18. 4. 1496 ff., LIMC 7. 1. 176 ff.
fatalis means ‘bringing doom’; cf. Sen. Ag. 730 f. ‘fatalis sedet / inter
potentes arbiter pastor deas’, Hor. carm. 1. 37. 21 ‘fatale monstrum’ (of
Cleopatra), Virg. Aen. 2. 237 ‘scandit fatalis machina muros’. incestus,
‘unchaste’, is combined (shockingly) with iudex because Paris was bribed
by Aphrodite with the promise of an illicit union with Helen; cf. Hom.
Il. 24. 30 c fi
! X ƒ æ Æ
º IºªØ, Stat. Ach. 1. 45, 2.
79. The mulier peregrina is of course Helen of Sparta, but Juno, as an
injured goddess, cannot bring herself to name either her or Paris; a
contemporary would also be reminded of Cleopatra (just as Paris sug-
gests Antony); for the same device in her case cf. epod. 9. 12, carm. 1. 37.
7, and in Antony’s epod. 9. 27. The reduction of Troy to dust recalls
Hom. Il. 9. 593 ºØ Fæ IÆŁØ; cf. Eur. Andr. 111 f. ±Œ
ºØ = ¼ı ŒÆd ŁÆºı ŒÆd Ø K ŒÆØ .
21–2. ex quo destituit deos / mercede pacta Laomedon: ex quo depends

not on vertit but on damnatam (23); for the dating of an inherited curse
cf. epod. 7. 19 f. ‘ut immerentis fluxit in terram Remi / sacer nepotibus
cruor’. destituit means ‘left in the lurch’ (cf. Cic. Sex. Rosc. 117 ‘induxit
decepit destituit . . . fefellit’, OLD 3b); as it does not literally mean
‘cheated’ it can hardly take an ablative like fraudare, so mercede pacta is
presumably absolute with a concessive sense. After Apollo and Poseidon
had built the walls of Troy, Laomedon, Priam’s father, withheld the
promised payment (Hom. Il. 21. 446 ff.); later he also cheated Heracles,
who then captured Troy and killed him (Il. 5. 638 ff., Call. fr. 698, Ov.
met. 11. 215 ff., Apollod. 2. 5. 9 with Frazer). His dishonesty became a
byword; cf. Virg. georg. 1. 502 f. ‘Laomedonteae . . . periuria Troiae’, Aen.
4. 542 with Pease, Ov. met. 11. 199 ff. Juno mentions these events to show
that she is not actuated wholly by pique.
22–4. mihi / castaeque damnatam Minervae / cum populo et duce
fraudulento: Hera and Athena had conspired with Poseidon to put
Zeus in bonds (Hom. Il. 1. 399 ff.), and as a punishment Poseidon was
forced to take service with Laomedon (schol. on Il. 21. 444, Tzetzes on
Lycophron 34); the hostility of the two goddesses to Troy seems to have
begun with this episode. The MSS read damnatum, but we are inclined
to accept damnatam (Glareanus) to agree with the feminine Ilion (18 n.);
as Bentley points out, this also avoids the slight awkwardness caused by
the masculine pulverem. The prominently placed mihi and Minervae are
datives of the agent (‘doomed by me’); cf. Prop. 4. 6. 21 ‘classis . . . Teucro
damnata Quirino’ (as explained by Housman 3. 1237). Some interpret as
addictam, ‘made over to for punishment’ (cf. Virg. Aen. 4. 699 ‘Stygio-
que caput damnaverat Orco’); but Juno and Minerva did not play so
central a role at the time of Laomedon’s perfidy. castae makes a contrast
with incestus above (19 n.); we are again led to suspect that the Judge-
ment of Paris is the real cause of offence; cf. Virg. Aen. 1. 27. The whole
population is associated with the guilt of their ruler (i.e. Laomedon) as
later with that of Paris; cf. Hom. Il. 24. 27 f. I
Ł ! +ºØ ƒæc = ŒÆd
—æÆ ŒÆd ºÆe `ºæı (Œ ¼ .
25–6. iam nec Lacaenae splendet adulterae / famosus hospes: Juno

still refuses to name the guilty pair (cf. 19 f.); for Lacaenae cf. Eur.
Andr. 486, Virg. Aen. 2. 601, 6. 511 with Austin. splendet refers to the
gorgeousness of Paris (cf. 1. 15. 20 with N–H, 4. 9. 13 ff., Hom. Il. 3. 392
Œºº. ºø ŒÆd ¥ÆØ). It is most pointed to take adulterae as
dative, alluding to Helen’s susceptibility (cf. 1. 5. 12 f. ‘quibus / intemp-
tata nites’). Heinze argues that a genitive is not only simpler but
balances Priami below, but splendet on its own is far from balancing
pugnacis . . . refringit. adulterae and famosus reflect Augustan censorious-
ness where Homer was more tolerant (Il. 3. 156 ff.); see I. Opelt,
Die lateinische Schimpfwörter, 1965: 52, Grassmann 141; in particular
the words foreshadow Augustus’ moral legislation, cf. 3. 6. 17 ff., 3. 24.
20, 28 ff. famosus hospes refers to Paris’ offence against ˘f Ø ; cf.
1. 15. 2 ‘perfidus hospitam’, Hom. Il. 13. 623 ff., Alc. 283. 5 L–P, Aesch.
Ag. 399 ff.
26–8. nec Priami domus / periura pugnacis Achivos / Hectoreis opi-

bus refringit: the inherited guilt of Laomedon (21 f.) affects the whole
house of his son Priam; cf. Hom. Il. 4. 31 ff., referring to Hera’s implac-
able anger against Priam and his sons. It was a commonplace that
Hector’s valour had prolonged Troy’s resistance (2. 4. 11 with N–H);
the adjective Hectoreis marks a higher style than the genitive Hectoris
(3. 4. 72 n.) and recalls Homer’s use of , ¯ Œ æ (Il. 2. 416 etc.). opibus in
the sense of ‘help’ (cf. epist. 1. 10. 36 ‘imploravit opes hominis’, 2. 2. 136,
Val. Fl. 3. 713) is less common than ope (1. 6. 15, 4. 2. 2).
29–30. nostrisque ductum seditionibus / bellum resedit: ductum means

‘prolonged’ (OLD 24). By seditionibus Juno suggests that but for the
rebelliousness of other gods the war could have ended sooner. The
assonance of resedit (‘has subsided’) with seditionibus is probably con-
scious, though the latter word was in fact derived from se(d ) and itio (‘a
going apart’).
30–3. protinus et gravis / iras et invisum nepotem / Troica quem

peperit sacerdos / Marti redonabo: ‘from now on I shall give over my
bitter anger and my hated grandson . . . to Mars’. In redonabo, a rare
word attested only here and at 2. 7. 3 in classical Latin, there is a slight
zeugma of two ideas: (a) ‘to renounce something to please somebody’ as
with condonare in Caes. bG 1. 20. 5 ‘uti et rei publicae iniuriam et suum
dolorem eius voluntati ac precibus condonet’, and donare in Cic. fam. 5.
4. 2, (b) ‘to renounce a further claim on somebody’ as with donare in
Petr. 30. 11 ‘dono vobis eum’ of a slave who is forgiven. For Juno’s
vindictiveness cf. Hom. Il. 18. 119 IæªÆº
º (towards Heracles),
Virg. Aen. 1. 11 ‘tantaene animis caelestibus irae?’, 1. 25, 5. 608. As Ares by
some accounts was Hera’s son (Hes. theog. 922 f. with West, Ov. fast. 5.
229 with Bömer), Romulus, the son of Mars and Ilia, was her grandson.
But Ilia, as her name indicates, was a descendant of the hated Aeneas
(Serv. auct. on Aen. 1. 273 says that in Naevius and Ennius she was his
daughter; cf Serv. on 6. 777). Hence Juno maliciously refers to her as
Troica sacerdos (for as a Vestal Virgin she should not have borne a child).
33–6. illum ego lucidas / inire sedes, discere nectaris / sucos, et

adscribi quietis / ordinibus patiar deorum: for the radiance of the
aether cf. Lucr. 1. 1014 ‘caeli lucida templa’, Cic. rep. 6. 16. We print
discere with Porph. and the preponderance of the MS tradition (RN
more confidently than NR); it balances inire and adscribi, not just
because of its prosaic realism but because it represents part of an
initiation (Heinze). Others feel that the picture of a novice at his first
wine-tasting is inappropriate to such a serious occasion and so prefer the
more conventional variant ducere (1. 17. 22, 4. 12. 14, OLD 25b).
The pattern ‘A, B et C’ is found in Plautus, Terence, Varro, and rhet.
Her. , but is unusual in classical Latin (K–S 2. 32); Juno may be adopting
a rather old-fashioned style. Both adscribi and ordinibus have an official
tone; cf. Lucian, deor. conc. 3 IªÆª K e PæÆe ŒÆd ÆæªæÆłÆ.
For quietis cf. Epicur. p. 71, fr. 1 Usener e ÆŒæØ ŒÆd ¼ŁÆæ h
ÆPe æªÆÆ
Ø h ¼ººfiø Ææ
Ø, Lucr. 3. 18 ff. ‘apparet divum
numen sedesque quietae’, Cic. nat. deor. 1. 45 and 1. 52 with Pease, Virg.
Aen. 4. 379 f. again with Pease. In view of the gods’ incessant bickering
there may be a slightly sardonic suggestion that now at last they can be
at peace.
37–8. dum longus inter saeviat Ilion / Romamque pontus: both longus
and saeviat stress the separation, which in Juno’s view must be emo-
tional as well as geographical. See the introduction.
38–9. qualibet exules / in parte regnanto beati: for parte as a region of

the world (55, 3. 24. 37) see Löfstedt (1933), 441 f., Late Latin (1959), 113.
qualibet suggests indifference rather than encouragement, and regnanto
implies ‘they can if they like’; Juno is still grudging. Even in the time of
Romulus she still regards the Trojans as exiles; as often, there is the
suggestion of a sneer in exules, which is combined paradoxically with
regnanto and beati.
40–1. dum Priami Paridisque busto / insultet armentum: according to

Virgil’s account Priam was left unburied on the shore (Aen. 2. 557, where
there is a hint of Pompey), but a tomb is mentioned by Antipater, anth.
Pal. 7. 136. 1 0 ˙æø —æØı ÆØe . Some think that the bustum
is the incinerated city of Troy (epod. 16. 11 f.), but ‘trampling on the
grave’ should be understood more literally; cf. Hom. Il. 4. 177 fiø
KØŁæfiøŒø 1ºı ŒıÆºØ, Eur. Elec. 327, Prop. 2. 8. 20 ‘insul-
tetque rogis calcet et ossa mea’. Though insultet is applied here to
animals, it retains a note of deliberate contempt.
41–2. et catulos ferae / celent inultae: cf. epod. 16. 10 ‘ferisque rursus
occupabitur solum’, 16. 19 f. ‘habitandaque fana / apris reliquit et rapa-
cibus lupis’, Isaiah 13: 21–2 ‘There marmots shall have their lairs,
and porcupines shall overrun her houses; there desert owls shall dwell,
and there he-goats shall gambol; jackals shall occupy her mansions, and
wolves her gorgeous palaces’ (New English Bible), orac. Sib. 8. 41 ŒÆd a
ŁØºÆ ºŒØ ŒÆd IºŒ NŒıØ, E. Fitzgerald, The Rubaiyat of
Omar Khayyam xviii ‘They say the Lion and the Lizard keep / The
courts where Jamshyd gloried and drank deep; / And Bahram, that great
hunter—the Wild Ass / Stamps o’er his Head, but cannot break his
Sleep’. The distribution of material suggests that the commonplace had
an eastern origin (D. Ableitinger-Grünberger, Der junge Horaz und
Politik, 1971: 67).
42–3. stet Capitolium / fulgens: for Juno even the major symbol
of Rome’s permanence (3. 30. 8 n.) is conditional. fulgens points to
the gilded roof of Jupiter’s temple, as well as suggesting more meta-
phorical splendours; cf. Virg. Aen. 8. 347 f. ‘Capitolia . . . / aurea nunc’,
Sen. contr. 1. 6. 4, 2. 1. 1, Plin. nat. hist. 33. 57 ‘cum varie sua aetas
de Catulo existimaverit, quod tegulas aereas Capitoli inaurasset’,
L. Richardson 223.
43–4. triumphatisque possit / Roma ferox dare iura Medis: though

triumphare is intransitive, the perfect participle passive is legitimate; cf.
3. 29. 27, Virg. georg. 3. 33 ‘bisque triumphatas . . . gentis’, Aen. 6. 770 with
Norden, K–S 1. 102. possit again is concessive rather than optative. dare
iura (‘to prescribe laws’) is a sign of sovereignty (Virg. georg. 4. 561 f.,
Aen. 1. 293, OLD s.v. ius 3b, TLL 7. 2. 682. 70 ff.); it is because of her
belligerence ( ferox) that Rome can impose the rule of law (cf. Virg.
Aen. 4. 231 with Pease, H. Fuchs, Augustin und der antike Friedensge-
danke, 1926: 194 n.). For hopes of defeating Parthia see the introduction
to 3. 5.
45–6. horrenda late nomen in ultimas / extendat oras: horrenda goes

with late (cf. 3. 16. 19, 3. 17. 9 n., 4. 4. 23); the adverb is balanced by
extendat (‘wide and far’) and so should not be combined with it. For the
fear of Rome in distant regions cf. 1. 35. 31 f., Ov. fast. 1. 717. For
the extent of the Roman empire cf. 4. 15. 15 ff., Fraenkel 451 n. 4, Austin
on Virg. Aen. 1. 286 ff., Bömer on Ov. fast. 5. 557, Woodman on Vell. 2.
126. 3. For the claim to world dominion see F. Christ (1938), V. Buch-
heit, Der Anspruch des Dichters in Vergils Georgika, 1972: 136 n., Zanker,
figs. 31a, 42, 62b, 64 (coins representing a globe), Nicolet 29 ff.
46–7. qua medius liquor / secernit Europen ab Afro: H is referring

to the Straits of Gibraltar, one of the boundaries of the known world
(B. Bosworth, JRS 89, 1999: 4); Spain is balanced by Egypt (48) at the
other end of the Mediterranean; both were areas of recent expansion.
Since liquor just means ‘water’ and is not used of a sea, there is (in NR’s
view) no significant problem; if there is a momentary hesitation in
anyone’s mind, it is at once removed by v. 48. RN thinks that medius
(‘intervening’) liquor would also be applicable to the Mediterranean and
might therefore cause confusion; L. Müller proposed modicus (cf. Avien.
ora marit. 335 ‘locos utrosque interfluit tenue fretum’), and RN has
considered minimus (which was always liable to corruption because of
the repeated downward strokes). H is presumably imitating Enn. ann.
302 ‘Europam Libyamque rapax ubi dividit unda’, but there the exact
place may have been indicated by the context, while rapax suggests the
turbulence of the straits; in our passage the perfect secrevit might get
round the difficulty (cf. Heinze) as it would allude to the original
separation of the continents (Diod. 4. 18. 5, Virg. Aen. 3. 417 of the
Sicilian straits ‘venit medio vi pontus’). The Greek form Europen is
more grandiloquent that Ennius’ Europam (cf. 3. 4. 76), just as liquor is
more grandiloquent than unda; as the name can refer to the mytho-
logical heroine it can be more readily combined with the personal Afro.
48. qua tumidus rigat arva Nilus: in view of Europen and Afro above,
Nilus too may be personified. In the descriptions of swollen rivers (e.g.
Virg. Aen. 11. 393, Ov. am. 1. 7. 43, OLD tumidus 3), tumidus is usually
destructive; so there is a slight paradox with the benign rigat. If one
accepts a conjecture above like modicus or minimus, RN sees a paradox-
ical contrast between the narrow sea and the swollen river.
49–52. aurum irrepertum et sic melius situm, / cum terra celat, sper-
nere fortior / quam cogere humanos in usus / omne sacrum rapiente
dextra: this stanza should perhaps be combined with its predecessor
(thus Wickham) rather than its successor (as most editors take it); this
makes clearer the parallelism of extendat (46), which refers to west and
east, and tangat (54), which refers to south and north. For the use of
fortior to develop a period cf. 1. 37. 26 ff., Nisbet in Adams–Mayer 150; it
seems less natural to put such an elaboration before the main statement.
It has sometimes been assumed that Juno now introduces a second
stipulation, parallel with the ban on rebuilding Troy, but spernere fortior
is hardly a strong enough expression to indicate such a crucial limita-
tion—contrast dum (37), dum (40), and sed (57). We seem to have here a
more incidental warning against the avarice associated with distant
expeditions (cf. 3. 24. 36 ff.).
The avarice involved in mining was a moral commonplace; cf. Deme-
trius of Phalerum (ap. Posidonius, F240a. 11 f. Kidd ¼ Athenaeus 6.
233e), Ov. met. 1. 138 ff. with Bömer, Manil. 5. 276 ff., Sen. ben. 7. 10. 2,
epist. 94. 57–8, nat. quaest. 5. 15. 3, Plin. nat. hist. 2. 158, Milton, PL.
1. 686 ff. ‘with impious hands / Rifled the bowels of their mother earth /
For treasures better hid’, a clear imitation of our passage; for the
opposite view that resources should be used cf. 2. 2. 2 with N–H, Cic.
nat. deor. 2. 98, 151 with Pease. Mining is also a dangerous process, see
Plin. nat. hist. 33. 70 ‘siduntque rimae subito et opprimunt operarios, ut
iam minus temerarium videatur e profundo maris petere margaritas
atque purpuras’; in our passage H suggests that it takes more courage
to reject mineral wealth than to dig for it ( for a similar paradox cf. 3. 16.
25 n.). The sequence of thought, however, is not obvious. NR suspects
that after the Straits of Gibraltar (46 f.) H thinks naturally of the
Spanish gold-mines ( J. F. Healy, Mining and Metallurgy in the Greek
and Roman World, 1978: 48). RN, however, finds a difficulty in the
intervening reference to Egypt, and thinks it might be relevant that
in 27 bc Augustus was supposed to be planning an invasion of Britain (3.
5. 3–4 n.), a country that was thought to be rich in gold (Tac. Agr. 12. 6
with Ogilvie–Richmond, appendix 4, Healy, op. cit. 48 and 52). On this
hypothesis the stanza would come best after v. 56, which refers to the
bad weather of the north; the search for knowledge (54) provides a more
creditable reason for exploration than greed for gold (a contrast that can
also be felt if a full stop is put after Nilus in v. 48 and a comma after
dextra in v. 52).
Commentators take different views about the articulation of vv.
50b–52. Some interpret ‘braver in scorning gold than in amassing it’
and take the rest of the sentence as an ablative absolute; but though
cogere is used of both gathering crops and collecting money (OLD 5),
without amplification it seems bald (Heinze), while the break after the
third syllable of the line is relatively rare. Others combine (aurum) cogere
with humanos in usus (‘press to human uses’), but if a pause is assumed
after usus that impairs the contrast between humanos and sacrum. Others
again interpret ‘than in pressing to human uses everything sacred with
rapacious hand’; but if rapiente dextra is an instrumental ablative it
straggles awkwardly at the end. It is no use arguing that sacrum is the
object both of cogere and rapiente; the clause either ends at usus or
continues, and in the latter case we do not expect sacrum to be governed
by a second verb. On any assumption v. 52 extends to mining what is
usually said about pillaging shrines; cf. Plin. nat. hist. 33. 2, where after
describing the indignation of our sacrae parentis he says imus in viscera.
53–4. quicumque mundo terminus obstitit / hunc tangat armis: the

terminus is the boundary imposed by the zones referred to in 55 f. (see
also 3. 24. 37 n.); the word is more commonly used of mountains, rivers,
and the ocean (Liv. 21. 44. 5, 37. 54. 23). RN finds the dative difficult, for
(as Peerlkamp saw) the zones are obstacles to adventurers, not to the
world; he therefore prefers to read mundi with Lambinus, though it has
negligible MS support (obstitit might have encouraged a copyist to
expect a dative). Bentley objects that this would require obstiterit, but
we could understand simply ‘has stood in the way’ (i.e. ‘is an obstacle’),
without specifying that a Roman army would be impeded at some time
in the future.
tangat, though another minor variant, seems necessary against the
transmission’s tanget; after extendat we look for a parallel concession.
54–5. visere gestiens / qua parte debacchentur ignes: gestiens implies an

urge that is almost physical; cf. 3. 16. 24 n., Fraenkel 272 n. The Greek
desire for Łøæ was a strong motive for travel (Herod. 1. 30. 2, 3. 139. 1,
Arist. Ath. pol. 11. 1). It affected military expeditions both in Greece and
Rome; Fraenkel cites Thuc. 6. 24. 3, Sall. hist. 1 fr. 103 M ‘more humanae
cupidinis ignara visendi’. This intellectual curiosity was attributed to
Ulysses by Horace (epist. 1. 2. 19 f. ‘multorum providus urbes / et mores
hominum inspexit’, which suggests more purpose than Hom. Od. 1. 3),
by Dante (inf. 26. 120 ‘per seguir virtute e conoscenza’), and by Tenny-
son (Ulysses 69 f. ‘but strong in will / To strive, to seek, to find, and not
to yield’).
parte refers to a region of the world (as at 39), ignes to the heat of the
torrid zone. bacchari is used metaphorically of wild natural forces, most
obviously of winds that rush like Bacchanals (1. 25. 11 with N–H, ciris
480). Words suggesting madness are also applied to intense heat, and
not only with reference to the rabid Dog-Star; cf. 3. 29. 18 f. ‘iam
Procyon furit / et stella vesani Leonis’, epod. 16. 61 f., epist. 1. 10. 16 f.
The prefix de- implies that the heat is raging to its limit (cf. debellare
etc.); see Ter. ad. 184 ‘si sati’ iam debacchatus es’, Plat. rep. 561a (of a
young man) Ka . . . c æÆ KŒÆŒ
ıŁfi , Ð E. Fantham, Comparative
Studies in Republican Latin Imagery, 1972: 49.
56. qua nebulae pluviique rores: cf. 1. 22. 19 ff. ‘quod latus mundi
nebulae malusque / Iuppiter urget’ (followed by the hot and uninhabit-
able regions of 21 f.). Northern countries were often associated with
clouds and rain; cf. Albinovanus Pedo, FLP p. 315 with Courtney, Tac.
Agr. 12. 3 ‘caelum crebris imbribus ac nebulis foedum’ (Britain), ann. 2. 23
(Germany). bacchari does not remotely suit a misty climate, but it could
be used of rain-storms; cf. Val. Fl. 6. 632 ff. ‘velut hiberno proruptus ab
arcu / imber agens scopulos nemorumque operumque ruinas, / donec ab
ingenti bacchatus vertice montis / frangitur’.
57–8. sed bellicosis fata Quiritibus / hac lege dico: fata (from fari), here
with dico, implies that Juno is a fatidica or prophetess. There is a paradox
in the conjunction of bellicosis with Quiritibus, which suggests civilians
(cf. Suet. Jul. 70, where Caesar shames his soldiers by calling them
Quirites). hac lege means ‘with this stipulation’ (OLD 12c).
58–60. ne nimium pii / rebusque fidentes avitae / tecta velint reparare

Troiae: nimium pii is an oxymoron; since Juno was hostile to the old
Troy, she deprecates the Romans’ excessive respect for their ancestors.
nimium should also be taken with fidentes; rebus means ‘circumstances’,
here ‘good fortune’. The emphatic position of avitae imposed by the
long hyperbaton stresses that the city now belongs to the remote past.
61–2. Troiae renascens alite lugubri / fortuna tristi clade iterabitur: for
the repetition of Troiae from the last word of the previous sentence cf. 3.
16. 15 n. ‘If the fortune of Troy comes to life again under a gloomy omen,
it will be re-enacted with all its grim disasters’. alite (literally ‘a bird’)
means an omen (cf. epod. 10. 1, carm. 1. 15. 5, 4. 6. 24); the concern for
augury is a particularly Roman feature (cf. especially Enn. ann. 72 ff. for
the observation of birds at the foundation of Rome).
63–4. ducente victrices catervas / coniuge me Iovis et sorore: Juno
asserts her double claim on Jupiter; cf. Hom. Il. 16. 432, 18. 356, Cic. nat.
deor. 2. 66 with Pease, Virg. Aen. 1. 46 f. ‘ast ego quae divum incedo
regina Iovisque / et soror et coniunx’, Ov. met. 3. 265 f. with Bömer.
65–6. ter si resurgat murus aeneus / auctore Phoebo: three, a signifi-

cant number (3. 19. 11 n.), is sometimes repeated in hyperbolical contexts
(Pease on Virg. Aen. 4. 690). resurgat refers to the rebuilding of Troy in
the Troad, though the word is more often used of the revival of Troy
in Italy (Virg. Aen. 1. 206 ‘illic fas regna resurgere Troiae’, Ov. fast. 1. 523
with Bömer). Apollo built the walls of Troy for Laomedon ‘so that the
city should be unbreakable’ (Hom. Il. 21. 446 f.) and this may suggest
that they were made of bronze (cf. Hes. scut. 243); for bronze structures
in myths and proverbs cf. Hom. Od. 7. 86 etc., 3. 16. 1 n., epist. 1. 1. 60
‘hic murus aeneus esto’. Bronze was a symbol of hardness before iron
(cf. Frazer, on Ov. fast. index under Bronze Age), and D. Wormell,
Hermath. 58, 1941: 116 ff. maintains that in such passages bronze indi-
cates magic as well as indestructibility; how far H was aware of the point
is, of course, unknowable. auctore is this context is supported by Virg.
georg. 3. 36 ‘Troiae Cynthius auctor’; ductore (cod. det.) and structore
(Bentley) are therefore unnecessary.
66–7. ter pereat meis / excisus Argivis: for Juno’s ferocity at the fall of
Troy cf. Virg. Aen. 2. 612 ff. ‘hinc Iuno Scaeas saevissima portas / prima
tenet sociumque furens a navibus agmen / ferro accincta vocat’ (no
doubt with post-Homeric precedents). meis is emphatic; the Heraeum
at Argos was one of Greece’s great cult-centres (OCD 687). The instru-
mental ablative is unambiguous at 1. 15. 6 and 4. 14. 9. excisus (‘extir-
pated’) is a very strong word (OLD 5).
67–8. ter uxor / capta virum puerosque ploret: the misery of war
encapsulated in five words. uxor is usually avoided in the higher genres
of poetry (3. 15. 1 n.).
69–70. non hoc iocosae conveniet lyrae: / quo, Musa, tendis?: iocosae
‘merry’ suggests themes of love and wine (1. 2. 34 with N–H).
70–2. desine pervicax / referre sermones deorum et / magna modis

tenuare parvis: H pretends that his recalcitrant Muse has carried him
away and needs to be reined in. This ‘breaking-off formula’ is sometimes
found in Pindar (cf. 3. 2. 25 n., Fraenkel 239, Thummer on Pind. I. vol. 1,
pp. 122 ff., Syndikus 2. 47); cf. also Pope, Windsor Forest 423 f. ‘Here cease
thy flight, nor with unhallow’d lays / Touch the fair fame of Albion’s
golden days’. More generally, the claim that grand themes are suitable
4 . D ES C EN D E CA ELO 53
only for epic and are diminished by the smaller compass and less exalted
style of lyric ignores Pindar as well as the Roman Odes (to name no
others). It is a characteristically Horatian pose (see last note); yet it is
clear from 1. 1, 2. 20, 3. 30, 4. 3, 4. 9 that the poet would not have been
pleased if others had taken it seriously. For ironic closures as a literary
motif see D. Fowler, Roman Constructions, 2000: 7 ff., 270 ff.
4 . D ES C EN D E CA ELO
[I. Borzsák, Acta Antiqua Hung. 8, 1960: 369 ff.; Davis 98 ff.; A. J. Dunston, Aumla 31, 1969:
9 ff.; Fraenkel 273 ff.; Hardie 86 ff., 98 f.; F. Klingner, Römische Geisteswelt, edn. 4, 1961:
376 ff.; Lowrie 214 ff.; Lyne (1995), 50 ff., 164 ff.; W. Marg, Monumentum Chiloniense
(Festschrift E. Burck), ed. E. Lef èvre, 1975: 385 ff.; J. F. Miller, CQ 48, 1998: 545 ff.; Pasquali
692 ff.; W. Theiler, Schriften der Königsberger Gelehrten Ges., G. Kl. 12.4, 1935: 253 f.
(reviewed by F. Klingner, Gnom. 13, 1937: 36 ff.) ¼ Untersuchungen zur antiken Literatur,
1970: 394 ff.; Williams (1968), 268 ff.]
1–8. Descend from Heaven, Calliope, and sing a lengthy song. Already I
seem to hear you and to be roaming in the sacred wood. 9–20. The doves of
fable protected me as an infant when I was lost on Mt Vultur, to the wonder
of all the places around. 21–36. As your protégé, Muses, I am transported to
the Sabine hills and in the past was rescued from mortal danger; as long as
you are with me, I shall venture beyond the limits of the Roman world.
37–48. You likewise refresh Caesar now that his wars are over, advising
mildness. On the other hand the impious Titans were destroyed by Jove.
49–64. The Giants terrorized heaven, but achieved nothing against Pallas,
Vulcan, Juno, and Apollo. 65–80. Force without wisdom is self-defeating:
witness the destruction of the lustful characters of myth.
In the first stanza the Muse is not just invoked but summoned: that is
a tradition of Greek lyric rather than epic (West on Hes. op. 1 f.). She is
asked for a melos, the Greek word for a song (1–2 n.), to be performed
with a wind instrument or strings (3 n.): these two instruments provide
the accompaniment for a Pindaric ode (4 n.). The climax comes with the
cithara of Apollo, which is by implication Horace’s preferred option:
that recalls the golden lyre of Pindar’s First Pythian, $æıÆ æØª
# ººø ŒÆd NºŒø = ØŒ 1ØA ŒÆ (1 f.). It will
emerge that Pindar has a profound influence on the development of
Horace’s ode; see especially Theiler and Fraenkel, locc. citt.
The poet now hears the response of Calliope (6–7 n.), much as the
Muses spoke to Hesiod on Helicon (theog. 22 n.); this leads him to tell a
myth about his own infancy, when he was lost on Mount Vultur, and
wood-pigeons covered him with bay and myrtle (9–20), as suited a
future poet. In a similar legend doves fed Zeus with ambrosia in the
Cretan cave (Hom. Od. 12. 62 f., Athen. 490–1 citing Moero 1. 3 f.,
Thompson 286; cf. 245 on Semiramis), and when the Babes in the
Wood were murdered, Robin Redbreast covered their bodies with leaves
(Percy’s Reliques ser. 3, book 2, 18). In particular such stories were told
about poets and literary men (Pease on Cic. div. 1. 78): a nightingale
sang on the lips of Stesichorus (Plin. nat. hist. 10. 82, Christodorus,
anth. Pal. 2. 128 ff.), bees fed Pindar with honey (Dio Chrys. 64. 23,
Aelian, var. hist. 12. 45, Antipater, anth. Plan. 305. 3 f.), and the same
thing happened to the mellifluous Plato (Val. Max. 1. 6 ext. 3, Plin. nat.
hist. 11. 55, Olympiodorus, vit. Plat. 382–3). In view of the influence of
Pindar on this ode, it seems significant that he tells how snakes fed
Iamus with honey and covered him with flowers (O. 6. 45 ff., 55 f.). Most
notably of all, Philostratus describes a picture in which the infant Pindar
was laid in bay and myrtle, just like the infant Horace (19), and fed by
bees (imag. 2. 12); for another parallel see 11 n. These tales presumably
originated not from Pindar himself (they do not suit choral lyric) but
from laudatory epigrams; and they were repeated earnestly by biograph-
ers as signs of providential protection and future greatness. See further
Pease on Cic. div. 1. 121, L. Bieler, ¨¯+ˇ& `˝˙2, 1, 1935: 39, M. R.
Lefkowitz, The Lives of the Greek Poets, 1981.
Horace develops his fable with a list of minor Apulian places that
marvelled at the miracle (13–16); he thus equates the surroundings of his
childhood with the landscapes of Greek poetry (cf. Martial 1. 49 on his
Spanish countryside), and at the same time he acknowledges the nar-
rowness of his first horizons. Then he turns to the present with a list of
better-known places where he has found inspiration (21 ff.); as a devotee
of the Muses, here Latinized as Camenae, he is transported to the
Sabine hills, his local equivalent of Helicon (21–2 n.). They have pro-
tected him in the crises of his past life (here enumerated in another list);
without literally believing in his own myths he might have liked to fancy
that he had been preserved for a higher purpose. Finally he turns to the
future: with the Muses’ help he is prepared to journey to the ends of the
earth (29–36); here the list of distant places suggests not just the Muses’
protecting power (cf. N–H on 1. 22. 1), but also the growing extent of his
fame (N–H on 2. 20. 14 ff.), which will reach beyond the present limits
of the Roman empire (see above on 3. 3. 45 ff.).
At v. 37 the ode moves from personal to political themes (perhaps
foreshadowed by the preceding place-names): at the very centre it tells
us that the Muses have a benign influence both on the poet and on the
Princeps himself. This refers not just to the great man’s interest in
literature (39–40 n.), but to the power of poetry to civilize and pacify
(41–2 n). Hesiod again had shown the way. In theog. 81 ff. he had asserted
the sacred link between the Muses and the ruler: they mark him out at
birth and give him eloquent speech; they enable him to make wise
decisions and to settle quarrels among the people, ÆºÆŒEØ
ÆæÆØØ KØ (90), cf. lene consilium (41). Pindar had taken
the idea further in his First Pythian, which speaks of the cosmic power
of poetry: ŒBºÆ b ŒÆd ÆØ ø Łº- = ªØ æÆ I ¸Æ- = Æ
fi Æ ÆŁıŒ ºø 1ØA (12 f.). At 42 ff. Horace turns abruptly to
the Titans and other monsters who were destroyed by the gods; this
contrasting picture was suggested by Pindar’s account of the enemies of
Zeus who were dismayed by the voice of the Muses (P. 1. 13 f. ‹Æ b c
ºŒ ˘ , IÆØ = —Øæø IÆ). Pindar goes on to
describe in graphic terms the sufferings of Typhon as he breathes fire
under Etna (15 ff.); Horace diffuses the effect by listing a whole series of
monsters and gods, but his description of Apollo (61–2 n.) and Etna (76)
confirms that he is still thinking of Pindar.
The monsters of Hesiod’s Theogony had been given an allegorical
significance by later ages; see especially Hardie (op. cit.) for the contrast
between chaos and cosmos, barbarism and civilization. When Pindar
wrote his First Pythian in honour of Hiero and his new foundation of
Aitna, the destruction of Typhoeus symbolized the defeat of the Car-
thaginians at Himera in 480 bc and of the Etruscans at Cumae in 473
(Fraenkel 279). When Callimachus talks of ‘late-born Titans’ (h. 4. 174
OłªØ 3ØB ), he is referring to the Galatian onslaught on Delphi
in 279 bc that was routed by Apollo; and there may have been further
such allusions in Hellenistic epic (Hardie, op. cit. 86 n., citing the
imitations in Nonnus). Public sculpture was even more significant: the
theme had already been used on the Parthenon metopes to allude to
the Persian invasions. The baroque Gigantomachies on the Great Altar
of Zeus at Pergamum symbolized the Attalids’ defeat of the Galatians in
230 and 166 (E. Simon, Pergamon und Hesiod, 1975, R. R. R. Smith,
Hellenistic Sculpture, 1991: 157 ff.); Attalus I had already dedicated a
similar monument on the Athenian Acropolis (Paus. 1. 25. 2, Hardie,
op. cit. 124). Horace would have known the Athenian Gigantomachies
from his days as a student, and he may also have visited Pergamum
when serving under Brutus (Encicl. oraz. 1. 233).
Horace’s ode is the most systematic account of the Gigantomachy
that has survived in Augustan literature, but apart from incidental
references the conflict between order and anarchy is also highly relevant
to the Aeneid, for instance in the episode of Cacus in book 8; it occurs
later in the Silver Age and especially in Claudian (carm. min. 53). The
theme seems to have played less part in contemporary sculpture, perhaps
because its violence was at odds with the Augustan ideal (Hardie,
op. cit. 89). The ivory doors on the Palatine temple of Apollo, dedicated
in 28 bc about the time of our poem, represented the god’s repulse
of the Galatian attack on Delphi (Prop. 2. 31. 13 ‘deiectos Parnasi
vertice Gallos’); but the cult statue of Apollo showed him holding the
lyre and not the bow (N–H on 1. 31. 1, Prop. 2. 31. 16, Zanker 85 and
fig. 186).
At line 65 Horace offers some generalizations (‘vis consili expers mole
ruit sua’); the alternation of aphorisms and exempla is Pindaric, but the
contrast with the legitimate use of force (vim temperatam) is particularly
Augustan. This leads to the mention of four further mythological
sinners, Gyges (69 n.), Orion (71 n.), Tityos (77 n.), and Pirithous
(79 n.); the offence of the last three is explicitly described as sexual,
and something similar may have been said about Gyges (cf. perhaps Ov.
fast. 4. 593 where Ceres says ‘quid gravius victore Gyge captiva tulis-
sem?’). The relevance of all this to Augustan moralizing is evident; in
particular there seems to be a hint at Antony, though it is typical of
Horace’s best political odes that the allusion is oblique.
We have related this ode to its Pindaric forerunner because this
provides a starting-point and a framework for its appreciation; but as
always, Horace’s elaboration is highly original. The autobiographical
first half of the poem, with its blend of realism and fantasy, modesty and
pride (cf. 3. 30), is unparalleled in the First Pythian. The myths of
Apollo’s lyre and the Muses’ protection establish Horace’s position as
a vates, and so enable him to speak with authority on public issues and
(amazingly) to occupy as much space in the poem as the Princeps
himself. Pindar had linked the harmony induced by the lyre with cosmic
and political order, addressing Hiero in the process; Horace gives the
idea an entirely Roman context, employing the Muses to associate his
own poetry with the Augustan settlement. Here we should think not
just of the civilizing power of literature (41 f.) but of something that is
implicit rather than expressed: the poet, like the good ruler, exercises
balance and control, imposing form and structure on raw, often recalci-
trant, material (for further examples of these parallels see NR in The
Camb. Hist. of Class. Lit. ii (1982), 402 f.). In sum, poetic and political
power are derived from the same divine source. This contention had a
lasting influence on the classicizing English ode: it is enough to refer to
Milton’s At a Solemn Musick (‘Blest pair of Sirens, pledges of Heaven’s
joy / Sphere-borne harmonious Sisters, Voice and Verse’), Dryden’s
Song for St Cecilia’s Day (‘From harmony, from heavenly harmony /
This universal frame began’), and Gray’s Progress of Poesy (the tranquil
eagle at vv. 20 ff. comes from Pindar’s First Pythian).
Metre: Alcaic.
1. Descende caelo: for the summons to the Muse cf. Sappho 127, 128
L–P, Stes. PMG 240 Fæ ¼ª, ˚ÆººØ ØÆ ºªØÆ. For descende in such a
kletic hymn cf. 3. 21. 7 (parodic), TLL 5. 1. 642. 46 ff. (Christian
instances), Milton, PL 7. 1 ‘Descend from heaven, Urania’. In early
Greek invocations the Muses are described as living on Olympus
(Hom. Il. 2. 484, Hes. theog. 75 with West), which is interpreted here
as the sky; cf. PMG frag. adesp. 935. 1 ff. ŁÆ = Fæ ºŁ I TæÆH,
Cavarzere 225. Porph. sees the implication ‘come down to earth from the
setting of the previous poem’; but arguably in the final stanza of 3. 3 H
had already come down to earth, and in any case this can have had no
influence on the composition of 3. 4, since the latter was almost certainly
earlier than 3. 3.
1–2. et dic age tibia / regina longum Calliope melos: the tibia was an
instrument with two pipes, blown through the end (not the side like a
flute); see M. L. West, 1992: 81 ff., J. G. Landels, Music in Ancient Greece
and Rome, 1999: 24 ff. Presumably the Muse is being asked to sing and
play on the pipe in alternation. Theiler thinks that she is to play on the
pipe without a song (op. cit. 397); that would make a clearer contrast
with v. 3 (see note), and the instrumental ablative is then rather simpler
(Theiler sees the same usage at 4. 12. 9 f. ‘dicunt . . . / carmina fistula’; cf
also [Theoc.] 20. 29 Œj ÆPºfiH ºÆºø, Œj ÆŒØ, Œj ºÆªØÆºfiø). Yet
some similar passages imply the use of words as well as an instrument (1.
12. 1 f. ‘quem virum aut heroa lyra vel acri / tibia sumis celebrare, Clio?’,
1. 32. 3 f. ‘age dic Latinum, barbite, carmen’, 3. 11. 7 f. ‘dic modos Lyde
quibus obstinatas / applicet auris’); moreover, when the Muse is in-
voked, she should tell the poet what to say and not simply provide
appropriate music. For age with another imperative cf. 1. 32. 3, 2. 11. 22,
OLD 24a; so in Greek invocations Alcman, PMG 14(a) 1 f., 27 (to
Calliope), Stes. 240 (cited above).
Hesiod described Calliope as the most important (ææ) of
the Muses (theog. 79), and this may be enough to explain why she is
singled out here. He goes on to say that she accompanies kings, so H
may associate her with political wisdom (thus Plut. mor. 746d, 801e), yet
the good advice mentioned by Hesiod comes from the Muses as a whole
(theog. 81–93); M. T. Camilloni, Le Muse, 1998, is ready to assign
particular functions to the Muses in Horace, but there are some cases
that do not fit (N–H on 1. 24. 3, P. Murray, CR 50, 2000: 294 f.). The
name Calliope in Greek suggested ‘a beautiful voice’ (Hes. theog. 68 of
the Muses IªÆºº ÆØ Od ŒÆºfi ); Ð the Greek associations are sustained
by melos, which is used in older Latin in the sense of carmen (Enn. ann.
293 ‘tibia Musarum pangit melos’, Naev. trag. 20, Lucr. 2. 412 ‘musaea
mele’), but it is not found elsewhere in the major Augustan poets. For
regina cf. Pind. N. 3. 1 4 ØÆ 1EÆ, Herodas 3. 97, Opp. hal. 1. 78,
anon. anth. Plan. 312. 1 ˚ÆººØ ÆºØÆ; in the sacral style the epithet
is sometimes separated from the proper name (N–H on 2. 19. 8). longum
prepares us for the fact that the ode is the longest in the whole collec-
tion (80 lines) and correspondingly one of the grandest.
3. seu voce nunc mavis acuta: this, the transmitted text, is usually
thought to to refer to singing without instrumental accompaniment;
cf. Varro ap. Non. 77M, 107L ‘et assa voce [i.e. solo voice] et cum
tibicine’; but as voce here is understood with both tibia (1) and cithara
(4), one would expect H to have made it quite clear if the solo voice
were intended. Moreover, in other passages where H mentions pipe and
lyre he says nothing of the solo voice (see 4 n.). Wickham, arguing for
the usual two categories, thinks that H asks first for a melody on the
pipe, then corrects himself and leaves it to Calliope to decide whether it
should be sung voce acuta (and so alternating with the pipe) or voce gravi
(and so accompanied by the lyre). But this gives a very involved con-
struction and has not been taken up by other editors. There is much to
be said for Darnley Naylor’s suggestion that we might read si voce for seu
voce; this would make it clear that there are only two alternatives (cf. 1.
12. 1 f. ‘lyra vel acri / tibia’). acuta indicates that the ancients thought of
the sharp, penetrating quality of the sound (cf. O ) rather than the
height of its pitch; the opposite term was gravis (cf. Ææ ); cf. M. L.
West (1992), 64, n. 73, Landels, op. cit. 54 f.
4. seu fidibus citharaque Phoebi: the MS tradition supports citharave

(so also Victorinus, GL 6. 180), which is defended by Theiler with
Strabo 1. 2. 3 ƒ ıØŒd łººØ ŒÆd ºıæØ ŒÆd ÆPºE ØŒ ,
Ael. Arist. 37. 21, but it is pointless to subdivide the stringed instru-
ments; the corruption would be very easy after seu. fidibus (‘strings’)
makes a hendiadys with cithara (cf. Virg. Aen. 6. 119 f. ‘Orpheus /
Threicia fretus cithara fidibusque canoris’). The lyre of Apollo was
more serene than the pipe of Dionysus, which Plato regarded as too
exciting (rep. 399d–e, W. D. Anderson, Ethos and Education in Greek
Music, 1966: 64 ff., 137 ff., M. L. West, 1992: 105 f.); the two instruments
are cited together by Pindar (O. 3. 8, 7. 12, P. 10. 39, N. 9. 8, I. 5. 27) as by
H himself (1. 1. 32 ff., 1. 12. 1 f., 4. 1. 22 ff., epod. 9. 5 with Mankin). Here
the lyre marks the climax and is presumably preferred; cf. Pind. P. 1. 1 f.
(cited in the introduction). It is significant that the Palatine temple,
which was dedicated on 9 October 28 bc (N–H on 1. 31. 1), must have
been nearing completion when this poem was written; inside stood the
Apollo Citharoedus of Scopas (see the introduction).
5–6. auditis? an me ludit amabilis / insania?: with a dramatic touch

(N–H vol. 1, pp. 310 f.) H asks his audience to confirm that they too
can hear the descending Muse; cf. Ar. nub. 292 fi Łı
! øB ; (can
Strepsiades hear the chorus of the descending clouds?), Call. h. 2. 4 f.
P
›æÆfi Æ ; Kı › ˜ºØ Ø EØ = KÆ (at the epiphany
of Apollo), fr. 227. 1. Ps.-Acro thought the question was addressed to the
Muses collectively; but if the passage is to cohere, auditis must govern
the same object as audire in the next line. insania refers to the ‘divine
madness’ attributed to poets by Democritus, Plato, and others (see the
introduction to 3. 25); madness is usually distressing, but here it is
paradoxically agreeable; cf. 2. 7. 28 ‘dulce . . . furere’, Anacreontea 53. 14
(Campbell)
ÆæØø ÆBÆØ. ludit suggests that the poet may be
imagining things; cf. Plat. Crit. 54d uæ ƒ ŒæıÆØH H ÆPºH
ŒFØ IŒØ. For similar uncertainty cf. Soph. OC. 316 pæ Ø; pæ
PŒ Ø; ª ºÆfi A; Virg. ecl. 8. 108, Keats, Ode to a Nightingale 80
‘Do I wake or sleep?’ (for H’s influence on this poem cf. epod. 14. 1 ff. and
G. Highet, The Classical Tradition, 1949: 637 f.).
6–7. audire et videor pios / errare per lucos: the poet’s request is
fulfilled within the poem; cf. Alcman, PMG 30, Davis 100 f. For videor
of dreams and visions (¼ videor mihi) cf. 2. 1. 21, Virg. ecl. 10. 58 f., Prop.
3. 3. 1 with Fedeli, Pope, Windsor Forest, 267–70. The groves belong to an
ideal poetic landscape, less rugged than the Bacchic countryside of 3. 25.
2; cf. Plat. Ion 534a–b, Troxler-Keller 29 ff. pios ascribes to the groves the
sanctity of the Muses’ devotees, just as Elysium can be called sedes pia
(culex 39, 375); all others are excluded. errare here means not ‘stray’ but
‘roam’ or ‘saunter’; cf. Virg. ecl. 6. 64 ‘errantem Permessi ad flumina
Gallum’, Petr. 27. 1 ‘nos interim vestiti errare coepimus’.
7–8. amoenae / quos et aquae subeunt et aurae: in a hot climate

water and breezes belong to the conventional locus amoenus; cf. 3. 29.
21 ff., N–H, vol. 2, pp. 52 f., with the bibliography there cited. For
amoenus of waters cf. Horsfall on Virg. Aen. 7. 30; here both the
adjective and the verb (‘come up to’) suit aquae better than aurae.
9. me fabulosae Vulture in Apulo: the emphatic me indicates that

the poet is under special protection (cf. 1. 22. 9 ‘namque me’); for
similar legends see the introduction above. fabulosae agrees with
palumbes (12); the long hyperbaton creates suspense and is a mark of
the grand style (cf. 4. 4. 7 ff.). The adjective means ‘told about in legend’
(cf. 1. 22. 7 f. ‘fabulosus / . . . Hydaspes’), and introduces a miraculous
note that is continued in mirum (13); for adjectives ending in -osus in
serious poetry see Brink on ars 4 (vol. 3, p. 579), P. E. Knox, Glotta 64,
1986: 90 ff. Monte Vulture (4,350 ft.) lies only 9 miles to the west of
Venosa, and looms over the whole area (for a fine photograph see Encicl.
oraz. 1. 1. tav. 17); as the name must have suggested a vulture, it would
have had sinister connotations that make a contrast with palumbes
below. Apulo reminds us of H’s pride in his native region (3. 30. 10 n.); cf.
serm. 1. 5. 77 f. ‘incipit ex illo montis Apulia notos / ostentare mihi’.
10. nutricis extra ylimen Apuliaey: the MSS divide between limen
Apuliae and limina Pulliae. The prosodic variation from Apulo (— ^ —)
in the previous line to Apuliae (^ — ^ —) cannot be justified as an
artistic elegance (for such variation see N–H on 1. 32. 11, N. Hopkinson,
Glotta 40, 1982: 162 ff.); it is true that the u of Apulia is always long
(while with Apulus only short u is attested), but the initial A is never
short (except at the corrupt 3. 24. 3). And though the mountain was on
the border with Lucania (cf. serm. 2. 1. 34), extra limen Apuliae not only
contradicts Vulture in Apulo but is quite inappropriate; for in the context
of a straying child limen can only mean the threshold of a house—a
potent symbol that divided the secure from the unknown (cf. Hom.Od.
15. 450 f. ÆEÆ ªaæ Iæe B Kd ªæØ IØººø, = ŒæÆº c
E, –Æ æ
øÆ ŁæÆ). extra limina Pulliae is also unconvincing:
H is thought to be referring to his wet-nurse, but in this grand poem he
would not name so obscure a person (as Bentley forcefully points out),
even if, contrary to common practice, she bore a gentile name (thus
A. Treloar, Antichthon 2,1968:58 ff.). Fraenkel 274 compares the minor
place-names below (14–16), but they were all in the public domain.
The best conjecture is perhaps limina pergulae (Baehrens); this was
proposed independently by Housman (Classical Papers 1. 99 f.), though
he later accepted Pulliae (W. G. Arnott, LCM 11, 1986: 149). For nutricis
as an adjective cf. [Quint.] decl. mai. 13. 4 (p. 269 Håkanson) ‘volui
relinquere avitos lares et conscios natalium parietes et ipsam nutriculam
casam’ (in a context with other Horatian reminiscences). Housman
called attention to ps.-Acro’s paraphrase ‘extra casae limen expositus’
(though on 10 the scholia follow the manuscript readings); note also
Antipater, anth. Pal. 9. 407. 1 f. æ K ºÆf ı =
Eº
ªØ æÆ KŒ ŒÆº (at CGL 2. 337 ŒÆº is glossed by
casa and pergula). It is true that pergula in the sense of a cottage may be
too humble for H’s family (cf. Prop. 4. 5. 70 ‘horruit algenti pergula curta
foco’, Petr. 74. 14, Auson. epist. 13. 6); but perhaps it belonged to H’s
wet-nurse (though in [Quint.], loc.cit. nutriculam casam seems to be the
speaker’s own cottage). In that case H could have been farmed out in a
remote country area: for the practice cf. Tac. dial. 28. 4 ‘non in cellula
emptae nutricis sed gremio ac sinu matris educabatur’, Suet. Aug. 6.
1 ‘nutrimentorum eius ostenditur adhuc locus in avito suburbano iuxta
Velitras permodicus et cellae penuariae instar’, K. R. Bradley in The
Family in Ancient Rome (ed. Beryl Rawson), 1986: 201 f., P. Garnsey in
The Family in Italy (ed. D. I. Kertzer and R. P. Saller), 1991: 86 ff., RAC
1. 381 ff. There can be no objection to the plural limina being applied to a
modest dwelling; cf. Lucil. 1107M.
Wade in his edition of 1731 had proposed villulae; H’s father had
acquired a property near Venusia (cf. serm. 1. 6. 71 agello, epist. 2. 2. 51
fundi). The word is attested at serm. 1. 5. 45 and 2. 3. 10, but the
diminutive may be too colloquial for this grand ode (Housman loc.
cit. contrasts the semi-humorous parmula at 2. 7. 10), and ps.-Acro’s
casae implies something smaller. Other conjectures include sedulae
(Bentley), vilicae (Yonge), devium (Lehrs), dum vagor (Courtney, Phoe-
nix 40, 1986: 319 f.).
11. ludo fatigatumque somno: cf. Hom. Il. 10. 98, Od. 6. 2 ofiø ŒÆd
ŒÆfiø Iæ ; for the same pattern see Enn. ann. 288 ‘vino domiti
somnoque sepulti’ with Skutsch’s parallels. In view of the importance of
Pindar for this ode, it seems significant that similar phrases are used
about him in two separate sources: see Paus. 9. 23. 2 (when he was fed by
bees) Œe ŒÆd o . . . ŒÆºÆ, vita Pindari p. 1 Drachmann
e ººF ŒÆı N o ŒÆ
ŁBÆØ. Horace’s expression is more
striking; it is a paradox that both play and sleep can be tiring. The
artificial position of -que is common enough in Horace (N–H on 1.
30. 6), and here may correspond to a feature of Greek lyric poetry (3. 21.
18, N–H on 2. 19. 27).
12–13. fronde nova puerum palumbes / texere: palumbes were wood-

pigeons; cf. Virg. ecl. 1. 57, Plin. nat. hist. 10. 106 ‘cantus omnibus similis
atque idem trino conficitur versu praeterque in clausula gemitu, hieme
mutis, a vere vocalibus’, Thompson 225 ff., Capponi 375 ff. For flocks of
wood-pigeons on Monte Vulture see C. T. Ramage, The Nooks and
Byways of Italy, 1868: 212. Doves are often associated with Venus (Virg.
ecl. 3. 69 palumbes, Thompson 246), but as they protect infants in
legends where she is not involved, she need not be relevant here; a
connection with the Muses does not seem to be attested (in spite of the
choir of Peleiades at Alcman, PMG 1. 60). fronde nova is fresh green
foliage (cf. 4. 1. 32 ‘novis floribus’), appropriate for a future poet (see
further 19 n.).
13. mirum quod foret omnibus: for the use of quod cf. epod. 2. 27 f.
‘frondesque lymphis obstrepunt manantibus, / somnos quod invitet
levis’. foret, an archaic equivalent of esset, in found at 4. 8. 22, epod. 12.
23, and thirteen times in H’s hexameter poems (especially serm. 1); it was
avoided by Cicero and Caesar, but is found quite frequently in Sallust
and in parts of Tacitus (Woodman–Martin on ann. 3. 14. 4).
14. quicumque celsae nidum Acherontiae: the MSS give Acherontiae or

Acheruntiae (so also Porph., CIL 9. 947 ‘numini Herculis Acheruntini’,
Procop. bG. 3. 23). The spelling Acerentia or Aceruntia is supported by
CIL 9. 417, 10. 482, and the modern name Acerenza; Acherontiae may be
due to popular association with Acheron, and H may have liked the
paradox of combining it with celsae. Acerenza, 13 miles south of Venosa,
was an important strongpoint in the Middle Ages, and its cathedral may
have given its name to the region of Basilicata. It is very high for an
inhabited place (2,700 ft.), and nidum suggests an eagle’s eyrie (as at 4.
4. 6); cf. Cic. de orat. 1. 196 ‘Ithacam illam in asperrimis saxulis tamquam
nidulum affixam’, Macaulay, Lays of Ancient Rome ‘Like an eagle’s nest /
Hangs on the crest / Of purple Apennine’. The genitive defines nidum
(K–S 1. 419, Woodcock 72. 5).
15–16. saltusque Bantinos et arvum / pingue tenent humilis Forenti:

saltus was a woodland clearing used for pasture, usually (as here) in hilly
country. Banzi is 12 miles south-east of Venosa, and as it rises to 1,900 ft.
it is easily visible. Below, the MSS vary between Forenti (supported by
Livy 9. 20. 9, Plin. nat. hist. 3. 105 as well as by the present name) and
Ferenti (supported by Porph. and Diod. 19. 65. 7, perhaps through
confusion with Ferentinum in Latium). humilis points a contrast with
celsae of Acherontia and the upland saltus of Bantia; but the original site
was abandoned (ps.-Acro already comments ‘nunc sine habitatore’); its
situation is disputed (cf. Encicl. oraz. 1. 393 f.). pingue of rich soil (OLD
3) suits ploughed land (arvum) as opposed to the pasture of Bantia; cf.
EÆæ. RN suspects an irrational association with arvina or lard (Virg.
Aen. 7. 627 ‘arvina pingui’).
17–18. ut tuto ab atris corpore viperis / dormirem et ursis: the ut-clause

explains mirum (‘how I slept’); cf. epod. 16. 53 f. ‘pluraque felices mir-
abimur, ut neque largis / aquosus Eurus arva radat imbribus’, Plaut.
merc. 240 f., Tac. hist. 1. 79. 2, Plin. epist. 1. 6. 2, H–Sz 645. It cannot
introduce a purpose clause after texere, for this does not go well with
premerer below; so it is a mistake to talk of syntactical ambiguity as
W. Wimmel does (Glotta 40, 1962: 134 ff.).
Snakes are repeatedly called ‘black’ in Latin poetry (serm. 2. 8. 95,
Virg. georg. 1. 129, Ov. met. 3. 63 with Bömer, Juv. 5. 91); the adjective is
not simply a colour-word but suggests deadliness (Virg. georg. 4. 407)
and in particular black poison (1. 37. 27, Virg. Aen. 2. 130). tuto is
pointedly juxtaposed with atris. For bears in S. Italy cf. Varro, lL 5.
100 (ursus is of Lucanian origin), [Ov.] hal. 58, Mart. spect. 10 (8). 1,
N–H on 1. 22. 14 (the ‘Daunian bear’ tamed by Pythagoras), Keller
1 (1909), 175; they are still sometimes found (Horsfall on Virg. Aen. 7.
17). Heinze comments on the repeated r in these lines; the letter was
‘rolled’ in Latin (Allen 32), and here perhaps suggested a bear’s growl; it
was commonly associated with a dog (cf. N–H on 2. 19. 23, Pers. 1. 109
on the littera canina).
18–19. ut premerer sacra / lauroque collataque myrto: premerer goes
further than texere and suggests that the baby was well protected under a
pile of leaves. The emphatic sacra qualifies both lauro and myrto, which
are combined by -que . . . -que. Likewise collata affects both nouns.
The bay was sacred to Apollo, god of poetry (3. 30. 16 n.), the myrtle
especially to Venus (RE 16. 1. 1180 f.); for the common collocation of
the two shrubs cf. Virg. ecl. 2. 54 with Clausen, 7. 61 f. ‘gratissima . . . /
formosae myrtus Veneri, sua laurea Phoebo’. Yet a pointer to H’s love-
poetry would not be expected in this grand ode to Apollo and the
Muses; cf. rather the legend about the infant Pindar mentioned by
Philostratus, imag. 2. 12 e b ªaæ ÆØ N I ŒØÆØ ŒÆd
ŒºHÆ ıææ (so also Ael. var. hist. 10. 21 gives the infant Plato a bed
of myrtle).
20. non sine dis animosus infans: the litotes is common in Greek
(Hom. Od. 15. 531 h Ø ¼ı ŁF, Pind. P. 5. 76 P ŁH ¼æ etc.);
such phrases mean ‘by Heaven’s help’, and individual gods are not
identified (cf. Nock 1 (1972), 261), but here the Muses must be implied
(for their status as goddesses cf. 1 n., Varr. rust. 3. 16. 7, Ov. ars 3. 348).
animosus means ‘spirited’ (‘quia solitudinem nemoris non expavit’
Porph.); in retrospect, with a touch of pride and humour, H represents
the straying child as courageous. Some commentators see a reference to
H’s future inspiration (cf. Virg. Aen. 6. 11 f. ‘magnam cui mentem
animumque / Delius inspirat vates’); but it is not clear that the word
can bear this extra implication.
21–2. vester, Camenae, vester in arduos / tollor Sabinos: H now

addresses the Camenae, or Latin Muses; cf. Livius Andronicus fr.
1 ‘virum mihi, Camena, insece versutum’) and Naevius frr. 1, 64, N–H
on 2. 16. 38, Suerbaum 303 f., 347 f. The emphatic vester means ‘as your
devotee’ (note the sacral anaphora); in this capacity the poet visits places
where he can write in peace (cf. epist. 2. 2. 77 with Brink).
tollor is not the obvious word for an uphill journey by carriage or on
horseback; when H revisits his beloved estate he seems to be borne
upwards by a power outside himself (cf. ÆYæÆØ, Virg. georg. 3. 8 f.
‘temptanda via est qua me quoque possim / tollere humo’), though he
does not claim to have been wafted in a dream like Callimachus (aet. 4)
or Ennius (Skutsch, pp. 147 ff.).
H’s Sabine estate was over 12 miles beyond Tibur (cf. 3. 1. 47 n); the
villa is 1,300 ft. above sea-level and the neighbouring Lucretilis 3,300 ft.
The word probably means ‘the Sabine countryside’; cf. Liv. 1. 45. 4 ‘bos
in Sabinis nata . . . dicitur’; Housman (Classical Papers 2. 613 and 658)
took Sabinos as referring literally to ‘Sabine people’, but it suits arduos
better if we understand agros. Haupt applied Sabinos more narrowly to
the estate itself (Opuscula 3, 1876: 578); cf 2. 18. 14 ‘satis beatus unicis
Sabinis’ (but see N–H on the text), Plin. epist. 5. 6. 45 ‘habes causas
cur ego Tuscos meos Tusculanis Tiburtinis Praenestinisque praeponam’,
but such precision is not in harmony with the following place-names;
cf. A. Bradshaw, in Studies in Latin Literature and Roman History
(ed. C. Deroux) 5, 1989: 172 ff.
22–3. seu mihi frigidum / Praeneste seu Tibur supinum: Heinze sees a
slight anacoluthon: the sentence develops as if the previous clause had
read ‘vester sum cum tollor’. Praeneste is Palestrina, 23 miles east-south-
east of Rome, famous for the temple of Fortuna Primigenia built by
Sulla (G. Bagnani, The Roman Campagna and its Treasures, 1929: 191 ff.,
RE 22. 2. 1549 ff., Suppl. 8. 1241, F. Coarelli, Studi su Praeneste, 1978);
with a height of 1,500 ft. rising in the arx to 2,460 ft. it provided a cool
refuge from Roman summers; cf. epist. 1. 2. 2, Juv. 3. 190 ‘gelidum
Praeneste’, Suet. Aug. 72. 2 (with Tibur a welcome secessus for Augustus).
Tibur is Tivoli, 15 miles east-north-east of Rome, 750 ft. high (Bagnani
226 ff., Weinstock RE 6A. 1. 816 ff.); it was a favourite retreat for H (2. 6.
5 f., epist. 1. 7. 45, 1. 8. 12), and at some time he acquired a property there
(4. 2. 31, 4. 3. 10, Suet. vita Horati ‘domusque ostenditur circa Tiburni
luculum’, Encicl. oraz. 1. 257 f.). supinum here means ‘sloping backwards’
(Virg. georg. 2. 276 ‘collisque supinos’, 3. 555), though often ‘flat’ like
oØ ; when Juvenal (3. 192) speaks of Tibur as ‘sloping forwards’
(‘proni Tiburis arce’) he is perhaps capping H, by referring to the steeper
slope below the town. RN suspects a decline from the bracing arduos
and frigidum to the laid-back qualities of Tibur and the notorious luxury
of Baiae.
24. seu liquidae placuere Baiae: normally in a hymn seu placuere would
refer to the deity’s choice of residence (N–H on 1. 30. 2), but here the
sacral formula is transferred to the poet. For the fashionable resort of
Baiae see 2. 18. 20 with N–H, epist. 1. 15. 12, where H’s horse has to be
headed elsewhere, J. H. D’Arms, Romans on the Bay of Naples, 1970:
index. liquidae (Encicl. virg. 3. 231 ff.) refers primarily to the pellucid air
(2. 20. 2 ‘per liquidum aethera’ with N–H, epist. 1. 1. 83 ‘nullus in orbe
sinus Bais praelucet amoenis’, Virg. georg. 4. 59, Gray, Ode on Spring 27
‘And float amid the liquid noon’). But it is hard to exclude, as a
subsidiary meaning, the clarity of the waters (ps.-Acro offers both
explanations); cf. Virg. Aen. 7. 760, Ov. met. 6. 400, Shelley, Ode to the
West Wind 33 f. (describing the Bay of Baiae) ‘And saw in sleep old
palaces and towers / Quivering within the wave’s intenser day’.
25. vestris amicum fontibus et choris: amicum here means ‘welcome to’
(cf. 1. 26. 1 ‘Musis amicus’ with N–H, epitaph. Bionis 76 ÆªÆE
ØºØ); because H is loved by the Muses he can rely on their
favour and protection. From early times water was associated with
prophecy (introduction to 3. 13) and the Muses with sacred springs,
Hippocrene on Helicon (Hes. theog. 6 with West), Castalia on Parnas-
sus (61 n.), Pieria near Olympus (West on theog. 53, N–H on 1. 26. 9),
Arethusa at Syracuse (Clausen on Virg. ecl. 10. 1); in Hellenistic and
Roman poets these often symbolized poetry that was fresh and modest
in scale (N–H on 1. 26. 6). The Camenae were originally water-nymphs
(Wissowa 219), with a spring outside the Porta Capena (Steinby 1. 216,
Richardson 63 f.); there they were associated with Egeria, the spring-
goddess of Aricia, who had prophetic gifts and was consulted by Numa
(Ogilvie on Liv. 1. 21. 1, Courtney on Juv. 3. 13). For the dances of the
Muses around springs cf. Hes. theog. 3 f. ŒÆ æd Œæ NØÆ
±ÆºEØ = Oæ
FÆØ (with West); their powers of inspiration and
prophecy are mentioned by Hesiod at theog. 31 f. KıÆ Ø
ÆPc = ŁØ, ¥Æ ŒºØØ K Æ æ K Æ.
26. non me Philippis versa acies retro: the battles of Philippi were
fought in the autumn of 42 bc near the Via Egnatia in eastern Mace-
donia; for the site and H’s participation see Encicl. oraz. 1. 241 ff. The
slaughter was the worst in all the civil wars (Vell. 2. 71. 2). ‘versa acies
retro’ alludes with austere understatement to the rout of the republican
army (cf. 2. 7. 9 f. ‘Philippos et celerem fugam / sensi’). There as here
H ascribes his escape to divine intervention (2. 7. 13 f. ‘sed me per
hostis Mercurius celer / denso paventem sustulit aere’); for the emphatic
me see 9 n.
27. devota non exstinxit arbor: for H’s accident with the tree cf. N–H
on 2. 13. 10 ff., below on 3. 8. 6–7 and 11–12. He makes so much of the
episode (exstinxit is a strong word) that it cannot be fictitious; he may
well have been injured (cf. 3. 8. 7 ‘prope funeratus’). devota means not
just ‘accursed’ as a conventional term of abuse, but ‘made over to the di
inferi’ (L. Watson, Arae, 1991: 209); cf. the amusingly ferocious opening
of 2. 13.
28. nec Sicula Palinurus unda: Capo Palinuro is on the Lucanian coast
12 miles south-east of Velia (Elea); the name in Greek may have implied
‘contrary wind’. Here in 36 bc, in the war against Sextus Pompeius,
Octavian lost many ships in a storm (Vell. 2. 79. 3, Dio 49. 1. 3); we
know that Maecenas was present (App. civ. 5. 99. 414), and our passage
may indicate that H was with him (E. Wistrand, Opera Selecta, 1972:
304 f.); the disaster must have coloured the story of the drowning of
Palinurus at Virg. Aen. 5. 833 ff., 6. 337 ff. Sicula unda (instrumental rather
than local) refers to the mare Siculum; the term is said not to suit anything
so far north, but note Phaedr. 2. 5. 10 of the Mons Misenus in Campania
‘prospectat Siculum et respicit Tuscum mare’. Horace, however, may be
indicating that the waves came from the place where the bellum Siculum
was being fought. The geographical spread, reaching from Macedonia to
Sicily, emphasizes the scope of the Muses’ protective power.
29. utcumque mecum vos eritis: utcumque here means ‘when and
only when’; cf. 1. 17. 10, 2. 17. 11, 4. 4. 35. vos after the main caesura
is emphatic, continuing the sacral anaphora above (vester . . . vester . . .
vestris), and leading to vos . . . vos below (37 and 41).
29–31. libens / insanientem navita Bosphorum / temptabo: for the

danger of these waters cf. 2. 13. 14 and 2. 20. 14; ‘in the centre there is
a rapid current from the Black Sea to the Sea of Marmora, but a
counter-current sets in the opposite direction below the surface and
along the shores’ (Encycl. Brit. , edn. 11, 4. 286); cf. A. E. Zimmern, The
Greek Commonwealth, 1931: 28 f., Walbank on Polyb. 4. 43. 3–10, RE 3.
742 ff. insanire is used like furere of natural forces; cf. 3. 7. 6, Catull. 25. 13
‘vesaniente vento’, Virg. ecl. 9. 43, Sen. nat. quaest. 6. 17. 1. The archaic
form navita is metrically convenient; it is used five times in the Odes and
twice in the Epodes, not always in heroic contexts. For temptare of
dangerous ventures cf. 1. 28. 5, Catull. 11. 13 f. ‘omnia haec . . . / . . . temp-
tare simul parati’, Virg. georg. 1. 207 ‘ (quibus) Pontus et ostriferi fauces
temptantur Abydi’; cf. also Cic. Tusc. 1. 45: ‘aliquid adsequi se putant qui
ostium Ponti viderunt et eas angustias per quas penetravit . . . Argo’.
31–2. et urentis harenas / litoris Assyrii viator: the MSS divide be-
tween urentis and arentis. For the former cf. Sen. nat. quaest. 2. 30.
1 ‘(Aetna) ingentem vim harenae urentis effundit’, 4a. 2. 18 ‘ardens pulvis
nec humani vestigii patiens’, Claud. 8. 598 ‘flagrantis . . . harenas’, Boeth.
cons. phil. 2. 6 carm. 13 ‘ardentes . . . harenas’. For arentis cf. Sil. 6. 140
‘arentis lento pede sulcat harenas’, Sulp. Sev. dial. 1. 3. 4; the assonance
with harenas is no objection (Ov. met. 15. 268 ‘aret harenis’), and the
words were thought to be etymologically connected (Maltby 269). But
the vivid urentis is unlikely to have arisen by corruption, whereas arentis
could easily be an Antizipationsfehler; and while all sand is dry, we look
for an exceptional feature to balance the raging Bosporus. In the same
way navita is picked up by viator: sea and land make a common ‘polar
expression’ (cf. 2. 6. 7 ‘maris et viarum’).
Assyrii properly should refer to the land of Nineveh, in later terms
Parthia or Iraq, but that has no Mediterranean shore; ps.-Acro and
Bentley are surely wrong to apply litoris to anything except the coast.
The adjective was sometimes used for ‘Syrian’, especially where exotic
products were concerned (N–H on 2. 11. 16); but the relatively fertile
coast of Syria was not the obvious place to typify a desert journey (in
spite of Manilius 4. 624 where a mariner gazes at Syriam perustam), and
it was not remote enough to match the other places in the context. Gow
saw a reference to the Persian Gulf and Alexander’s march through the
Makran desert, which was further east in the modern Baluchistan (CAH
6 edn. 2, 835 f.), but unlike the other distant places this does not suit the
Roman perspective. E. Wistrand proposed Austurii (Opera Selecta, 1972:
465 ff.), citing a Moorish people known from the fourth century as
Austuriani and by Corippus as Austures (RE 2. 2592. 40 ff.); this would
provide a southern point of reference as the Syrtes do elsewhere (1. 22. 5,
2. 20. 15 in closely parallel passages), but coming between the Bosporus
and Britain it would be a very obscure reference, and although such lists
sometimes ‘box the compass’ (Theiler, op. cit. 404), complete symmetry
cannot be required. RN has considered limitis Assyrii, i.e. the Roman
frontier with Parthia (OLD limes 2b); cf. Plut. Crass. 22. 5 P Ł b
c #æø ØØ ŒÆd #ıæø ŁæÆ;
33. visam Britannos hospitibus feros: after dangerous places H turns to

savage peoples; visam is more enterprising than the English ‘visit’, cf. 2.
20. 14, 3. 3. 54. Roman interest in Britain is shown again in 3. 5. 3 (see
note); for its remoteness cf. Catull. 11. 12, Cic. nat. deor. 2. 88 with Pease,
Virg. ecl. 1. 66 ‘penitus toto divisos orbe Britannos’. Strabo 4. 5. 2
describes the inhabitants as more barbaric than the Celts; for the
human sacrifices of the Druids cf. Caes. bG 6. 16. 1, Tac. ann. 14. 30. 3
‘nam cruore captivo adolere aras et hominum fibris consulere deos fas
habebant’, Juv. 15. 124.
34. et laetum equino sanguine Concanum: the Concani were a Can-

tabrian tribe in northern Spain (RE 4. 798, Syme 1978: 185), for we hear
of a place there called Concana (Ptol. geog. 2. 6. 51); the area caused
concern even before Augustus’ campaign in 26, and Dio records oper-
ations by Statilius Taurus in 29 (51. 20. 5, cf. Syme, Roman Papers 2. 828).
For their drinking of horses’ blood cf. Sil. 3. 361 ‘cornipedis fusa satiaris,
Concane, vena’ (in a Spanish context); this literal bloodthirstiness puts
them on a par with the Britons. The practice is usually attributed to the
Scythians (Virg. georg. 3. 461 of the Geloni, Lucan 3. 282 ff., Plin. nat.
hist. 18. 100 of the Sarmatians, cf. Mart. spect. 3. 4 ‘epoto Sarmata pastus
equo’, Sen. Oed. 470 of the Massagetae, Stat. Ach. 1. 307 f., Clem. paed. 3.
24. 4); that is why Porph. says ‘Hispaniae gens est vel ut alii dicunt
Scythiae’, and why Silius absurdly describes the Concani as descendants
of the Massagetae (loc. cit.).
35–6. visam pharetratos Gelonos / et Scythicum inviolatus amnem:

two topical areas in the north-east balance two in the west. The Geloni
were a Scythian tribe mentioned several times by Horace and Virgil,
possibly because the Romans associated their name with gelu (2. 9. 23
with N–H, 2. 20. 19, Syme, 1978: 186 n.); for their archery cf. 3. 8. 23 n.
(of the Scythians in general), Virg. Aen. 8. 725 ‘sagittiferosque Gelonos’.
The Scythian river is the Tanais or Don (cf. 3. 10. 1, 3. 29. 28, 4. 15. 24),
which comes out in the Sea of Azov; the periphrasis indicates its
prominence in the Roman mind; cf. 2. 9. 21 ‘Medumque flumen’ and
Call. h. 2. 98 #ıæı ÆE ªÆ Þ (both of the Euphrates).
The Borysthenes or Dnieper had greater commercial importance, but
the Tanais is particularly relevant here because it was regarded as the
boundary of Europe and Asia; cf. 3. 10. 1 n.
inviolatus implies not just ‘unhurt’ but ‘inviolable’, ‘sacrosanct’ (Caes.
bG 6. 23 ‘hospitem violare fas non putant’), even among a cruel people;
cf. 1. 22. 1 ff. (the integer vitae needs no weapons even in barbarous
lands), Prop. 3. 16. 13 f. ‘quisquis amator erit, Scythicis licet ambulet oris,
/ nemo adeo ut noceat barbarus esse volet’.
37–8. vos Caesarem altum, militia simul / fessas cohortis abdidit

oppidis: as usual H avoids addressing the great man directly (the only
exception in Odes 1–3 is 1. 2. 52); but here, quite paradoxically, the Muses
become the mouthpiece of the poet. Octavian, as moderns conveniently
but misleadingly call him to distinguish him from Julius Caesar, as-
sumed the name ‘Augustus’ in January 27 bc; in the immediate after-
math the more grandiloquent title would have been expected, as in
3. 3. 11, 3. 5. 3; but this ode was probably written in the autumn of 29
bc (see below). The temple of Janus had been closed at the beginning of
that year (Dio 51. 20. 4 and the note of Brunt and Moore on res gest. 13).
The honorific altum (cf. Ov. Pont. 2. 3. 63 ‘Caesaris alti’) was to lead to
the medieval ‘Highness’ (cf. the Irish Ard Rı́), but here the word was
well within the heroic tradition; cf. serm. 2. 5. 62 f. ‘ab alto / demissum
genus Aenea’, Virg. Aen. 9. 697 ‘Sarpedonis alti’, OLD 11.
The MSS divide in the main between abdidit and addidit; Bentley
read reddidit from the meaningless redditis of a few MSS (cf. Tac. ann.
1. 17. 6, 1. 30. 3 ‘si . . . suis quisque hibernis redderentur’); E. Courtney has
proposed dididit (Phoenix 40, 1986: 319 ff.). abdidit makes a contrast with
militia, and suggests that the soldiers can now enjoy a quiet life in their
winter quarters (cf. epist. 1. 1. 5 ‘latet abditus agro’); there would be no
repetition of the conflicts of the Republic when the army of a victorious
general (Marius, Sulla, Pompey, Caesar) presented an actual or potential
threat to the political system. addidit would refer to the settlement
of veterans on the land, as undertaken by Octavian in 30 bc (Suet.
Aug. 17. 3, Dio 51. 4. 5–6) and presumably continued after his triumphs in
29; this was sometimes done by adding to existing communities (res gest.
3. 3 ‘deduxi in colonias aut remisi in municipia’, Hygin. grom. p. 177
Lachmann ‘quosdam in veteribus oppidis deduxit’, cf. Tac. ann. 13. 31. 2
‘coloniae Capua atque Nuceria additis veteranis firmatae sunt’). But this
was a lengthy process involving land-surveyors and lawyers; here fessas
suggests that the army had just come back. Finally, though this cannot
be decisive, NR observes that abdidit contains an interesting poetic
metaphor, whereas addidit is a prosaic word.
39–40. finire quaerentem labores / Pierio recreatis antro: the Princeps

can relax now that his troops have been quartered. The passage seems to
refer to the occasion in 29 when he heard Virgil reading the Georgics at
Atella (Suet. vit. Verg. 105 ff. Rostagni, cited by Schütz and independ-
ently by D. A. Malcolm, CR 5, 1955: 242 f.); Maecenas was present,
and so perhaps H himself. The cave is hardly a garden grotto (thus
Malcolm) but belongs to a symbolic landscape (cf. 3. 25. 2–4 nn., Pind.
P. 6. 49 K ı
EØ —Øæø, Mart. 12. 11. 3, Juv. 7. 59 f. ‘sub antro /
Pierio’); for the ‘Pierian spring’ near Olympus cf. 1. 26. 9 with N–H.
recreatis describes the refreshment provided by poetry (cf. 1. 32. 14 f.
‘laborum / dulce lenimen’, Lucr. 6. 94 ‘Calliope, requies hominum’, Cic.
Att. 4. 10. 1 ‘sic litteris sustentor et recreor’); in addition to his exhaus-
tion, Octavian was ill with a throat-infection (Suet. loc. cit.), but
recreatis is tactfully inexplicit. About the same time he rewarded Varius
for his recently staged Thyestes (Schanz–Hosius 2. 163); his serious, if
limited, literary interests are mentioned by Suet. Aug. 84–6.
41–2. vos lene consilium et datis et dato / gaudetis almae: for the
power of the Muses to make the world a gentler place cf. Hes. theog.
80 ff., Pind.P. 1. 10 ff. (see introduction), 5. 66 f. (of Apollo), Plat. Prot.
326b, rep. 441e, Plut. praecepta gerendae reipublicae 801e c ˚ÆººØ
ÆæŒºı m c ÆØºFØ – ÆNØØ OE.
We may assume that the lene consilium (tendered tactfully by ‘the
Muses’, not by Horace in propria persona) coheres with Octavian’s view
of his own policy; cf. carm. saec. 51 f. ‘iacentem / lenis in hostem’, Aug. res
gest. 3. 1 ‘victorque omnibus veniam petentibus civibus peperci’, 34. 2.
More specifically, it is often supposed (and NR agrees) that in 29 bc H
has in mind primarily the idea of clemency to the defeated Antonians;
for though Antony and Cleopatra had been eliminated, there were
many lesser people among Antony’s supporters who could be won
over (like Horace himself after Philippi). If this is right, the rest of
the poem emphasizes that (as in Pythian 1) before peace could come the
enemies of legitimate vis had to be crushed in a terrible war. In that
mythical struggle our attention is focused entirely on the protagonists;
there are no subordinates. RN objects that the last half of the ode says
nothing about mildness or forgiveness but concentrates on the over-
throw and punishment of sinners who rebelled against the divine order;
he thinks that the point about lene consilium is made in more general
terms, and that H is contrasting the humanity and enlightenment of the
Princeps with the barbarism of his opponents (a colon could be printed
after almae). For the same contrast cf. Pind. P. 1. 13 (cited in the
introduction) and 8. 10 ff. (where the address to Hesychia, Goddess of
Peace, is followed by the defeat of the Giants).
The second i of consilium is treated as consonantal, thus lengthening
the previous syllable; for the synizesis cf. 3. 6. 6 ‘hinc omne principium’,
Bo 81 f., Skutsch on Enn. ann. p. 59, S. Timpanaro, Encicl. virg. 4. 877 ff.
The Muses ‘are glad to have given the advice’ when they know it has
been accepted. L. Müller explained dato with the gloss ‘vel a vobis vel ab
aliis’, but ‘ab aliis’ is irrelevant, and his interpretation impairs the
rhetorical pattern by which a verb is picked up by a participle (Virg.
Aen. 1. 736 f. ‘laticis libavit honorem, / primaque libato summo tenus
attigit ore’, Wills 311 ff.). J. Gow explained ‘when gentle advice is given
to you’ (cf. Auson. prof. Burd. 24. 9 f. ‘tam bone dandis / semper consiliis
quam taciturne datis’); but nobody would be so presumptuous as to
advise the Muses.
42–3. scimus ut impios / Titanas immanemque turmam: scimus intro-

duces an exemplum, cf. Aesch. cho. 602 Yø, Soph. El. 837 rÆ. The
Titans were the twelve children of Earth (Gaea) and Sky (Uranus), the
most notable being Cronus and Rhea, the parents of Zeus; see Hom.
Il. 8. 479, Hes. theog. 133 f. with West. Hesiod goes on to describe the
Titanomachy in which Zeus and the other new gods defeated their
predecessors (theog. 617 ff. with West); H seems to conflate this with the
Gigantomachy and other assaults on the gods (below, 53 n.).
One side of the MS tradition has turmam here and turmas in 47; the
other has turbam here and turbas in 47. Here turmam is appropriate as it
describes a military formation, though usually cavalry (cf. 2. 19. 22
‘cohors Gigantum’, Soph. Trach. 1058 f. ªªc = æÆe ˆØªø);
for the same reason turmas in 47 is totally inappropriate. Here immanem
points to the Giants’ monstrous appearance; cf. Apollod. 1. 6. 1
ŒÆŁØØ ÆŁEÆ Œ KŒ ŒÆºB ŒÆd ªø, r
b a Ø
ºÆ æÆŒ ø (they are often portrayed and described as snake-
footed, e.g. Ov. trist. 4. 7. 17 ‘serpentipedesque Gigantas’). H is not using
a hendiadys for ‘the troop of Titans’; to judge from Hesiod and Apol-
lodorus the Titans were not monstrous in appearance, and as the ‘old
gods’ they could hardly be called iuventus in v. 50. If v. 50 refers
exclusively to the ‘Hundred-Handers’ (see note ad loc.), so must turmam
here; but in that case the allusion seems too restricted to be clear.
44. fulmine sustulerit caduco: for the thunderbolt as the weapon of

divine justice cf. 1. 3. 40, 1. 12. 59 f., 3. 3. 6, A. B. Cook, Zeus 2. 11 ff.,
722 ff., Courtney on Juv. 13. 78. caducus means ‘descending’; cf. [Aesch.]
Prom. 360 ˘e ¼ªæı º , = ŒÆÆØ ŒæÆı (directed at
Typhon, for whom see 53 n.), Lycophron 382. The Latin adjective is
normally used of things that are frail and languishing (Prop. 4. 2. 53 ‘tela
caduca’, Lyne, 1989: 165 ff.), but sometimes it has a more sinister note
(e.g. 2. 13. 11 on the falling tree); Bentley’s corusco is not needed. sustulerit
(‘destroyed’) primarily means ‘lifted’; so there is verbal play in combin-
ing it with caduco (D. West, ap. Costa 41).
45–6. qui terram inertem, qui mare temperat / ventosum: Jupiter holds
sway over all the elements; cf. 1. 12. 15 f. ‘qui mare et terras variisque
mundum / temperat horis’, Plaut. rud. 1 ‘qui gentes omnes mariaque et
terras movet’ with Marx, Enn. ann. 556. For ‘terram inertem’ cf. Ov. met.
15. 148, Calp. 4. 109; here the immobility of the earth is contrasted with
the fluidity of the sea (cf. 1. 34. 9 ‘bruta tellus et vaga flumina’). temperat
suits the control of natural forces, regit (48) the government of sentient
beings, including the underworld.
46. et umbras regnaque tristia: umbras (elaborated by regnaque tristia)

is Bentley’s conjecture for urbes; for this common way of referring to the
underworld cf. Pease on Virg. Aen. 4. 25, Sen. HO. 1477 f. ‘emenso freta /
terrasque et umbras’. This last example parallels the trio here of earth,
sea, and underworld, a sequence which is interrupted by urbes; see
Shackleton Bailey 114 f.
47–8. divosque mortalisque turbas / imperio regit unus aequo: turbas

suggests the confusion of human life and the greater numbers of
mortals; the variant turmas (‘squadrons’) has no relevance here (43 n.).
49–50. magnum illa terrorem intulerat Iovi / fidens iuventus horrida

bracchiis: illa iuventus must refer to the immanem turmam of 43; it is an
appropriate description of the Giants (cf. 2. 12. 7 ‘Telluris iuvenes’),
though other monsters are not excluded. Porph. persuasively comments
‘ordo est fidens bracchiis iuventus horrida’; cf. Hom. Il. 12. 135 (of Lapith
warriors)
æØ ØŁ Mb Ø. Some take bracchiis also with
horrida (‘bristling with fore-arms’); on this view H means the ‘Hun-
dred-Handers’, Aegaeon (or Briareus), Cottus, and Gyges (Hom. Il.
1. 401 ff., Hes. theog. 147 with West). According to the usual account
they helped Zeus to resist the Titans (theog. 617 ff. with West, p. 663),
but in another version they joined the attack on Zeus: see Eumelus fr.
2K (Aegaeon an ally of the Titans), Virg. Aen. 10. 565 ff. with Hardie
154 ff. (Aegaeon fights against Jove’s thunderbolts), Bömer on Ov. fast.
3. 805. Yet in our passage, where fidens bracchiis is modelled on Homer, it
is difficult to give horrida bracchiis the required emphasis: one would
expect something like ‘densis iuventus horrida bracchiis’ (suggested by
Housman, Classical Papers 1. 167).
Heinze argued that terrorem described a threat to Jupiter rather than
actual fear; he compared Cic. Mil. 71 ‘declarat se non terrorem inferre
vobis’. Yet see Hom. Il. 1. 406 (on the Hundred-Hander Aegaeon) e
ŒÆd ØÆ ŒÆæ Ł, Ov. met. 5. 322 ‘caelitibus fecisse metum’
(on Typhoeus), Aetna 54 ‘Iuppiter et caelo metuit’ (because of the
Giants), Stat. Ach. 1. 484 f., Sidon. carm. 6. 15, 7. 133 f. Homeric gods
had human emotions.
51–2. fratresque tendentes opaco / Pelion imposuisse Olympo: Otis

and Ephialtes, the gigantic sons of Aloeus, piled up the Thessalian
mountains to make a ramp to heaven; cf. Hom. Od. 11. 315 f. ! ˇÆ
K ˇPºfiø ÆÆ Ł, ÆPaæ K ! ˇfi = —ºØ, Virg. georg. 1.
280 ff. and Thomas 128 ff., Aen. 6. 582 ff., Ov. am. 2. 1. 11 ff. with
McKeown (who records varying arrangements of the mountains), Apol-
lod. 1. 7. 4, F. P. Wilson, Ox. Dict. of English Proverbs, edn. 3, 1970: 600.
tendentes reflects ÆÆ in Homer and ØÆÆ at Hes. theog. 209
(though that refers to the Titans). opaco transfers Homer’s foliage from
Pelion at the top of the heap to Olympus at the bottom (so Virg. georg. 1.
282 ‘frondosum . . . Olympum’). For the tense of imposuisse see 3. 18. 15 n.
53. sed quid Typhoeus et validus Mimas: H now describes the battle of
the Gods and Giants (cf. Apollod. 1. 6. 1–2 with Frazer, F. Vian, La
Guerre des Géants, 1952, with REG 65, 1952: 1 ff., Hardie 85 ff., RE Suppl.
3. 655 ff., LIMC 4. 1. 191 ff.); for its symbolic significance in Pindar and
public sculpture see the introduction. In early Greek mythology it came
later than the battle with the Titans, but in the Hellenistic and Roman
periods these and similar legends were often conflated (Bömer on Ov.
fast. 5. 35, LIMC 4. 1. 193, Lyne, 1995: 51 n.); cf. Milton, PL 1. 196 ff. ‘in
bulk as huge / As whom the fables name of monstrous size, / Titanian
or Earth-born, that warred on Jove, / Briareos or Typhon . . . ’. Instead
of saying ‘what could the Aloadae avail against the Gods?’ H attaches
his question to a parallel illustration; for the schema see 3. 1. 45 n.
Typhoeus, son of Gaea and Tartarus, was half man, half dragon:
Hom. Il. 2. 781 ff., Hes. theog. 820 ff. with West, [Aesch.] Prom. 351 ff.,
Apollod. 1. 6. 3 with Frazer, J. Fontenrose, Python, 1980: 70 ff., LIMC 8.
1. 149 f. In Hesiod he came after the Giants and was still probably
regarded as distinct at Pind. P. 8. 15 ff.; for a later tradition that treated
him as a Giant see Roscher 5. 1440, Fontenrose 80 n. Mimas was a Giant
who hurled the volcanic island of Lemnos against the Gods (Claud.
gigant. 85 ff., Sidon. carm. 15. 25 f.) but was killed by Hephaestus with
masses of red-hot metal (Apollod. 1. 6. 2); there are other accounts of his
death (Hunter on Ap. Rhod. 3. 1225–7), but in view of the reference to
Volcanus (59) this may have been the version that H had in mind. He
is mentioned on the Great Frieze at Pergamum (see further LIMC
4. 1. 193, E. Simon, Pergamon und Hesiod, 1975: 41, and 59 n. below).
The name seems to be derived from the mountain opposite Chios
(Hom. Od. 3. 172, h. Hom. 3. 39, Call. h. 4. 67).
54. aut quid minaci Porphyrion statu: Porphyrion, the king of the
Giants, was overcome by the bow of Apollo (Pind. P. 8. 15 ff. cited
above, RE 22. 1. 272 f., Simon, op. cit. 43–5, taf. 15); in another version
he tried to rape Hera but was destroyed by Zeus and Heracles (Apollod.
1. 6. 2). See also Ar. aves 1252, Naev. fr. 19 ‘inerant signa expressa,
quomodo Titani / bicorpores Gigantes magnique Atlantes / Runcus
ac Purpureus, filii Terras . . . ’; this passage described sculptures on the
temple of Acragas, or perhaps rather the decoration on a shield
(E. Fraenkel, JRS 44, 1954: 14 ff. ¼ Kleine Beiträge 2, 1964: 25 ff.). minaci
statu is a descriptive rather than an instrumental ablative; the epithets
increase in length with each giant in 53–4 (see 67–8 n.). The phrase
suggests the stance of a warrior poised for action (cf. Petr. 95. 8 ‘statum
proeliantis componit’). Some interpret statu as ‘stature’ (OLD s.v. 2); this
avoids the apparent inconsistency with ruentes (58), but the baroque
Gigantomachies of sculpture combine aggressive postures with an
impression of dynamic movement (for this feature of ecphrasis cf.
Kerkhecker on Call. iamb. pp. 177 ff.).
55–6. quid Rhoetus evolsisque truncis / Enceladus iaculator audax:

Rhoetus is a giant at 2. 19. 23, Sidon. carm. 6. 24, but more usually a
centaur; perhaps the name should be emended to Rhoecus everywhere
(Housman, Classical Papers 3. 1103), but for difficulties see N–H on 2. 19.
23. For the giant Enceladus see Batrachomyomachia 283, RE 5. 2578 f.,
LIMC 3. 1. 742 f. His combat with Athena was perhaps portrayed on
the temple of Apollo at Delphi (Eur. Ion 209 ff. ºØ s K
¯ ªŒºfiø = ªæªøe ººıÆ Yı; = ºø —Æºº Ka Ł )
and also at Pergamum (Simon 44); he was buried under Etna according
to Call. aet. 1. 35 f., Virg. Aen. 3. 578 ff., Aetna 71 ff. evolsisque truncis is
best taken as an instrumental ablative with the agent-noun iaculator
(Plaut. Men. 187 ‘uter ibi melior bellator erit inventus cantharo’, H–Sz
128); this binds the sentence more tightly together than an ablative
absolute, and gives Enceladus a long epithet to balance Porphyrion’s
(cf. 54 n.). Fighting with trees was typical of centaurs (Vian on Ap.
Rhod. 1. 64). audax here implies reckless insolence rather than courage
(OLD 2).
57–8. contra sonantem Palladis aegida / possent ruentes?: for Athena’s

part in the Gigantomachy see LIMC 2. 1. 990 ff.; it was shown at
Pergamum (Simon 47–9, taf. 14, 15), and also on the Panathenaic peplos
(Lyne on Ciris 29–34); when her warlike aspects are mentioned, Latin
poets tend to call her Pallas rather than Minerva (cf. 1. 6. 15, 1. 12. 20, 1.
15. 11). The aegis was originally associated with Zeus, but after Homer
more particularly with Athena (OCD 17); sometimes it is a shawl,
fringed with serpents, draped over the goddess’s left arm; sometimes it
is a breastplate with the Gorgon’s head in the middle; sometimes it could
be shaken so as to inspire fear (Hom. Il. 15. 230, 309 f., 17. 595, Virg. Aen.
8. 354); here sonantem points to a breastplate. ruentes suggests a mindless
onslaught (‘inconsiderate venientes’, ps.-Acro); in later writers’ inter-
pretation of the Gigantomachy Athena stood for rationality (Theiler,
op. cit. 415 f., citing Cornutus 39. 13, Ael. Arist. 37. 9 etc.).
58–9. hinc avidus stetit / Volcanus: hinc . . . hinc refers to the different
sides of Athena; for Hephaestus’ part in the Gigantomachy see LIMC
4. 1. 647, Apollod. 1. 6. 2 (53 n.). avidus presumably means ‘eager for the
fray’; for this use, without a genitive, cf. Sall. hist. 2. 67M ‘avidis ita
atque promptis ducibus’, Tac. hist. 1. 45. 1 ‘avidum et minacem militum
animum’, ann. 1. 51. 1 with Goodyear. The adjective could also suggest
fire’s capacity to devour (‘propter ignis aviditatem’ Porph., cf. 75 n.); this
may be how H interpreted Æºæ , the obscure Homeric epithet for fire
(cf. Aesch. cho. 325, Quint. Smyrn. 3. 711 , ˙Æı ÆºæE, 13. 150,
330); for a similar ambiguity cf. 1. 4. 8 ‘Volcanus ardens’. avidus has
sometimes been suspected, for instance by Shackleton Bailey; A. Y.
Campbell at various times considered calidus, rapidus (which he rejected
as unsuitable for a lame god), and validus (which would require a change
at 53 validus Mimas).
59–60. hinc matrona Iuno et / numquam umeris positurus arcum: for

Hera’s presence in the Gigantomachy cf. 54 n., LIMC 4. 1. 702 ff., Simon
20–2, taf. 16. For matrona Tonantis cf. Ov. fast. 6. 33 with Bömer: the
gods support Augustan values against the primitive lust of Porphyrion.
Apollo’s part was more important (LIMC 2. 1. 309 f., Simon, taf. 18),
and is central to H’s ode. The bow, like the lyre, was a standing attribute
of the god; cf. 1. 21. 11 with N–H, h. Hom. 3. 131, Call. h. 2. 19, LIMC 2. 1.
196 ff. Here he does not intend to replace the bow on his shoulder until
the battle is won (thus Kiessling and Gow); most commentators think
that Apollo will not take the bow down from his shoulder, but that is not
the natural interpretation of umeris positurus, and in the heat of battle he
should be represented as actually shooting. At the same time H seems to
have caught a significant moment, as if he were describing a work of art
(cf. Virg. Aen 8. 704 f. ‘Actius haec cernens arcum intendebat Apollo /
desuper’, Keats, Ode on a Grecian Urn 20 ‘For ever shalt thou love and
she be fair’); for Apollo’s bow at Actium cf. also Prop. 4. 6. 55.
61–2. qui rore puro Castaliae lavit / crinis solutos: this stanza reminds
us that the warrior god of the Gigantomachy is also the god of poetry
(cf. 4, Miller, loc. cit.); for the description cf. Hom. Il. 20. 39, epod. 15. 9
‘(dum) intonsosque agitaret Apollinis aura capillos’, Ap. Rhod. 2. 676 f.,
Virg. Aen. 4. 147 f. with Pease.
The hair-washing must also be a traditional motif; cf. 4. 6. 26
‘Phoebe, qui Xantho lavis amne crinis’ (alluding to the Lycian river
and the god’s golden hair), Val. Fl. 1. 449, Stat. Theb. 1. 697 f. rore refers
to a spray of water (OLD 2a); cf. especially the paean of Aristonous
inscribed on stone at Delphi: t —ÆæÆF ªıºø = Pæ ØØ
˚ÆÆºÆ = ÆE e Æ KÆæ- = ø (Powell, Coll. Alex. p.
164, 41 ff.).
62–3. qui Lyciae tenet / dumeta natalemque silvam: Apollo had a

famous oracle at Patara in Lycia (east of Rhodes) near the mouth of
the Xanthus; cf. Herod. 1. 182, Bömer on Ov. met. 1. 516, RE 18. 4. 2555 ff.
Lycia and Delos are mentioned together at Pind. P. 1. 39 (with Castalia),
Virg. Aen. 4. 143 f. ‘qualis ubi hibernam Lyciam Xanthique fluenta /
deserit ac Delum maternam invisit Apollo’ (where see Pease for the
god’s seasonal migrations). For dumeta (thick scrub) cf. 3. 29. 22–4 n.
Apollo was born on Delos (hence natalem) beside a palm or olive (Hom.
Od. 6. 163, h. Hom. 3. 18, Call. h. 4. 210, Paus. 8. 48. 3), and perhaps there
was a sacred grove in his honour; cf. Cic. leg. 1. 2 ‘quam Homericus
Ulixes Deli se proceram et teneram palmam vidisse dixit, hodie mons-
trant eandem’, Ael. var. hist. 5. 4 (the olive and palm are said to flourish
on Delos), Tac. ann. 3. 61. 1 (a rival grove at Ephesus).
64. Delius et Patareus Apollo: the cult-centres are now mentioned

again in reverse order (chiasmus), with the god’s name reserved for
the climax. Though naturally it is not mentioned in the poem, it is
worth recalling that in 42 bc Brutus destroyed Xanthus and looted the
treasure ‘both public and private’ from the neighbouring Patara; H was
on his staff about this time (serm. 1. 7. 18 ff.) and may have witnessed
the outrages; see Plut. Brut. 32 (who tries to excuse Brutus), App. civ.
4. 79–81, Dio 47. 34, Hercher, Epistol. Graec. pp. 182 ff. (letters attributed
to Brutus), Encicl. oraz. 1. 236. Note must have been taken of Xanthus
in the war of words at the time of Philippi; H was not as friendly to the
memory of Brutus as is usually supposed (see N–H on 2. 7. 2).
65. vis consili expers mole ruit sua: mindless force is self-destructive: cf.
Pind. P. 8. 15 Æ b ŒÆd ªºÆı
Æº K
æ fiø, Bacchyl. 15. 62 f.
[0 5æØ ] æØºı = Â ÆEÆ þº ˆªÆÆ , Eur. TGF fr.
732N Þ b IÆŁc ººŒØ ŒØ º, Diod. 5. 71. 5 (on the
Giants’ punishment). consilium here means ‘wise judgment’ (OLD 7), a
quality prized at Rome (H. Fuchs, MH 4, 1947: 166 ff.); the word echoes
consilium (41), though there it refers to good advice. People or cities are
said from time to time to topple either through their own strength
(epod. 16. 2 ‘suis et ipsa Roma viribus ruit’, Liv. 6. 19. 6) or their own bulk
(Liv. praef. 4, again of Rome, ‘ut iam magnitudine laboret sua’, Ov. met.
1. 156 of the Giants ‘obruta mole sua cum corpora dira iacerent’, Sen.
dial. 2. 2. 2); H conflates the two ideas.
66–7. vim temperatam di quoque provehunt / in maius: the phrase vim

temperatam balances vis consili expers in length, metre, and position
(‘vertical responsion’). Force controlled by reason suits the self-
presentation of Augustus (‘hoc ad Augustum vult referri’ Porph.).
67–8. idem odere viris / omne nefas animo moventis: the three cola
on vis increase in size; cf. Fraenkel 351, E. Lindholm, Stilistische Studien
zur Erweiterung der Satzglieder im Lateinischen, 1931. ‘The strength
that plots every kind of wickedness’ is distinguished alike from mind-
less brutality and force controlled by wisdom. As viris in effect
means validos it can be combined with animo moventis (OLD s.v.
movere 19).
69–70. testis mearum centimanus Gyges / sententiarum: Gyges was

one of the Hundred-Handers (for whom see 49–50 n.); the MSS read
gigas here and at 2. 17. 14 (see N–H), but Muretus and Bentley restored
the correct form of the name (West on Hes. theog. 149). testis is often
used of people who illustrate an aphorism; cf. Cic. off. 2. 26 ‘testis est
Phalaris’ (also illustrating an assertion about power), Prop. 2. 13. 53, Ov.
her. 20. 101 ff. By his sententiae H is referring to the aphorisms in 65–7
rather than simply to his personal opinions; mearum placed before its
noun with hyperbaton seems surprisingly emphatic, and Campbell
looked for a reference to divine judgment (ratarum W. M. Edwards);
but the mention of the poet’s standpoint in a grand lyric poem is a
Pindaric feature (cf. O. 4. 17 f., P. 11. 52 f., fr. 169. 3 f.).
70–1. notus et integrae / temptator Orion Dianae: for notus ¼ ‘notori-

ous’ cf. 3. 11. 25 f. ‘notas / virginum poenas’; the word balances testis (69),
and has the same function as scimus (42). This version of the myth is
recorded by Serv. Aen. 1. 535, and seems to be implied by Call. h. 3. 264 f.
In Arat. phaen. 637 ff. Orion was killed by a scorpion which Artemis
sent against him for attempting to rape her. According to Homer, he
was killed by Artemis through jealousy of his love for Eos (Od. 5. 121 ff.);
for various accounts see Apollod. 1. 4. 5 with Frazer, J. Fontenrose,
‘Orion’, Univ. Cal. Class. Stud. 23, 1981, RE 18. 1. 1072 ff. temptare
can refer to attempted rape (cf. Tib. 1. 3. 73 ‘Iunonem temptare Ixionis
ausi’), but the agent-noun is not attested before this and not again till
Tertullian; the word is pointedly juxtaposed to integrae (¼ intactae) as
Orion to Dianae.
72. virginea domitus sagitta: Diana had the title of Virgo (3. 22. 1 n.), as
Artemis that of Parthenos; for the poetic use of the adjective for a
genitive cf. Löfstedt (1942), 107 ff. As Orion was a gigantic hunter
(Hom. Od. 11. 572 ff.) there is a paradox in his subjugation by a young
maiden. domare (‘to tame’) suits the action of a huntress; for its applica-
tion to violent sinners cf. 2. 12. 6 f. ‘domitosque . . . / Telluris iuvenes’,
Pind.P. 8. 17 AŁ, N. 7. 90.
73. iniecta monstris Terra dolet suis: the monsters of lines 42–56 were
all born from Earth (Apollod. 1. 1. 1–3, mainly from Hes., theog.
147 ff., 185, Virg. georg. 1. 278 ff.), as were Orion (Apollod. 1. 4. 3) and
Tityos below (Hom. Od. 11. 576); for iniecta cf. Apollod. 1. 6. 3
ÆPfiH . . . ˘f KææØł `Y Zæ (on Typhoeus). Terra combines
the idea of inanimate earth with that of a sentient goddess; for her
grief cf. Hes. theog. 858
Ø b ªÆEÆ ºæ (on Typhoeus), Quint.
Smyrn. 3. 397 (on Tityos); Ge is shown prostrate under Athena at
Pergamum (Hardie, op. cit. pl. 3, cf. LIMC 4. 1. 172, Simon, taf. 14).
H’s sentence best balances its successor if iniecta is combined with dolet
( ¼ ‘dolet se iniectam esse’); for this Greek construction with verbs of
feeling cf. Virg. georg. 2. 510 ‘gaudent perfusi sanguine fratrum’, K–G 2.
53 f. suis, with its connotations of affection, is paradoxically applied to
the loathsome monstris.
74–5. maeretque partus fulmine luridum / missos ad Orcum: partus

again suggests an affectionate relationship; so, more outrageously, Man-
ilius 3. 6 ‘fulminis et flammis partus in matre sepultos’. Like the Greek
Hades, Orcus can be used both of the god (2. 3. 24, 2. 18. 34, Hom. Il. 1. 3
and usually in Homer) and his kingdom (3. 11. 29 n., Hom. Il. 8. 16,
Pind. P. 5. 96); for NR this is one of the frequent cases where no
distinction can be drawn, but RN thinks that for a poet at least the
construction with ad seems better suited to the place. Similarly luridum
could be applied to the god ([Tib.] 3. 3. 37 f. ‘me vocet . . . luridus Orcus’)
and to the underworld (cf. Prop. 4. 11. 8 lurida porta of its door, and the
commoner use of pallida at Virg. Aen. 8. 244 f. ‘regna . . . pallida’); RN
thinks the latter description is the more arresting.
75–6. nec peredit / impositam celer ignis Aetnen: for the burial of the
monsters under Etna cf. especially Pind.P. 1. 19 f. and [Aesch.] Prom. 365
on Typhoeus (see the introduction), Call. h. 4. 141 ff. on Briareus, aet.
1. 36 on Enceladus (above 56 n.); in Homer’s account Typhoeus was
buried in the country of the Arimi (Il. 2. 782 ff.), and later ‘Inarime’
was identified with Ischia off Campania (Hardie on Virg. Aen. 9. 716,
Dewar on Claudian, de sexto consulatu Honorii, praef. 17 f.). celer ignis is
the fire darting from the monster’s mouth; cf. Pind. P. 1. 25 f. on
Typhoeus ŒE , `ÆØ Œæıf æ = Øı IÆØ,
Ov. met. 5. 353, Aetna 73 on Enceladus. Paradoxically it has not con-
sumed the mountain so as to allow respite or escape (Ov. met. 5. 356 f.);
for peredit cf. 58 n., [Aesch.] Prom. 366 f. on Typhoeus ŒEÆØ . . .
N ÞÆØØ `NÆÆØ o, Petr. 122. 135 f. ‘iamque Aetna voratur
/ ignibus insolitis’.
The Greek form Aetnen is read here by Porph.’s lemma and some
good MSS (cf. Ov. met. 13. 770, 14. 1); it should be accepted here on the
principle of ‘difficilior lectio’ against the better-attested Aetnam. The
reference could have had a particular interest for contemporaries: Etna
had erupted in 44 bc (Virg. georg. 1. 471 ff., Liv. fr. 57), 36 (Sen. epist. 79. 5
citing Cornelius Severus, App. bell. civ. 5. 117), and 32 (Dio 50. 8);
Augustus also wrote a poem on Sicily (Suet. Aug. 85. 2), which may
have referred to Etna, but the date is uncertain.
77–8. incontinentis nec Tityi iecur / reliquit ales: Tityos, another

enormous son of Earth, had his liver gnawed by vultures because of
an assault on Leto (2. 14. 8 with N–H, 4. 6. 2 ‘Tityosque raptor’, Hom.
Od. 11. 576 ff., RE 6A. 2. 1593 ff.); incontinentis refers as often to a lack of
sexual restraint. He was killed by Artemis (Pind. P. 4. 90 ff.) or Apollo
(Ap. Rhod. 1. 759 ff.) or both together (Apollod. 1. 4. 1); Frazer ad loc.
cites works of art for his punishment (LIMC 8. 1. 40), including Poly-
gnotus’ picture of the underworld at Delphi (Paus. 10. 29. 3). The liver
was often regarded as the seat of desire (1. 13. 4 with N–H, TLL 7. 1. 245.
71 ff.), though this has been denied for Homer (Onians 84 ff., West on
Hes. theog. 523 ff.); the myth was rationalized by making Tityos an
example of lust (Lucr. 3. 992 f. ‘sed Tityos nobis hic est, in amore
iacentem / quem volucres lacerant’ with Kenney, [Heraclitus] alleg.
Hom. 18, anth. Lat. 636. 21 ff.). reliquit (balancing peredit) means ‘has
let go of ’, but may also have a hint of ‘abandoned’ (see next note).
Tityos’ liver was eternally renewed (Sen. Ag. 18. ‘fecundum iecur’), as in
the analogous case of Prometheus (Hes. theog. 523 ff.).
78–9. nequitiae additus / custos: nequitia here refers to lust, and does
not have its milder implication of ‘naughty behaviour’ (3. 15. 2 n.); the
abstract noun stands for ‘a lustful creature’. additus means ‘attached to’
or ‘set over’ as a custos or jailer (cf. Plaut. aul. 556 of Argus ‘quem
quondam Ioni Iuno custodem addidit’, mil. 146, OLD s.v. 4); with
reliquit there seems to be a grim suggestion that the guard has not
failed in his duty (cf. the phrase signa relinquere).
5 . CA ELO T O NA N T EM C R ED I D I M V S 79
79–80. amatorem trecentae / Pirithoum cohibent catenae: Pirithous
was a savage Lapith (N–H on 2. 12. 5), son by Zeus of Ixion’s wife
(Hom. Il. 14. 317 f.); he accompanied Theseus to Sparta to abduct the
young Helen (Diod. 4. 63. 2, Hygin. fab. 79), and when he failed to win
her for himself tried with Theseus’ help to abduct Persephone from the
underworld. As a punishment he was fettered eternally, and even Hera-
cles failed to rescue him (except in the novel treatment of Euripides);
see Apollod. 2. 5. 12 with Frazer and especially epit. 1. 23–4, Virg. Aen. 6.
393, RE 19. 1. 115 ff., LIMC 7. 1. 232 ff. trecentae is an indefinite large
number (2. 14. 5 with N–H); the assonance trecentae cohibent catenae may
suggest the clank of chains; in the usual account Pirithous was bound in
the coils of serpents (Apollod. epit. 1. 24).
The emphatic amatorem refers to licentiousness (3. 18. 1, epist. 1. 1. 38
‘vinosus, amator’, Plaut. asin. 921 ‘surge, amator, i domum’, Cic. Tusc. 4. 27
‘aliud est amatorem esse, aliud amantem’). Pirithous was an example not
only of lust but of impiety (Ov. met. 8. 612 f., Diod. 4. 63. 4); he thus
matches Orion and Tityos. Fraenkel (285) suggests that the young lover’s
offence is venial (cf. the sympathetic reference in the much later 4. 7. 27 f.
‘nec Lethaea valet Theseus abrumpere caro / vincula Pirithoo’), but
amatorem tells against an indulgent reading; Lyne adds that Pirithous
was not a monster but a man, just like Antony (op. cit. 167 f.). There is
indeed a hint of Antony, but his infatuation for Cleopatra met with no
understanding from the victors, who represented him as a drunken
womanizer who had betrayed his country to an oriental queen.
5 . CA ELO TO NA N T EM CR ED I D I M V S
IOVEM R EGNARE
[ J. Arieti, TAPA 120, 1990: 209 ff.; Fraenkel 272 f.; H. Haffter, Philologus 93, 1938: 132 ff.;
H. Kornhardt, Hermes 82, 1954: 85 ff.; Lyne (1995), 55 f.; G. Marconi, Rivista di cultura
classica e medioevale 9, 1967: 15 ff.; Pasquali 701 ff.; H. T. Rowell in Studies presented to D. M.
Robinson (ed. G. E. Mylonas and D. Raymond) 2, 1953: 663 ff.; Williams (1968), 438 ff.]
1–4. We believe that Jupiter reigns in the sky because he is the Thunderer;
but Augustus will be held a god on earth when Britain and Parthia have
been conquered. 5–12. The soldiers of Crassus have married Parthian wives
and forgotten their country. 13–18. It was to avoid such a situation that
Regulus repudiated Carthage’s terms, which would have set a damaging
precedent. 18–40. ‘I have seen our standards,’ he said, ‘hanging in Cartha-
ginian shrines, and Romans humiliated as prisoners. A ransomed soldier will
not renew the fight, for he has proved his cowardice by being taken prisoner’.
41–56. Regulus disregarded the pleas of family and friends, and though he
knew the torture that awaited him, he hurried on his way as though he had
settled some legal business and was making for his country estate.
The first stanza looks forward to the conquest of Britain in the West
and Parthia in the East: the former issue was more topical in 27 bc
(3–4 n.), but the latter is more relevant to the development of the poem.
When Crassus was defeated at Carrhae in 53 bc, 20,000 of his men were
killed and 10,000 taken prisoner (Plut. Crass. 31. 7), and it was necessary
for Roman prestige to avenge the defeat and recover the lost standards.
In 36 Antony launched a large expedition against Parthia, but was
forced to withdraw (3. 6. 9–10 n.); in the course of the operation
he tried to negotiate a settlement that would lead to the return of
both the standards and the prisoners (Plut. Ant. 37. 2, 40. 4), though
this may simply have been a ruse to cover his preparations for war (Dio
49. 24. 5). For these and subsequent events see N. C. Debevoise, A
Political History of Parthia, 1938: 108 ff., K.-H. Ziegler, Die Beziehungen
zwischen Rom und dem Partherreich, 1964, A. N. Sherwin-White, Roman
Foreign Policy in the East, 168 bc to ad 1, 1984, chapter 13, C. Pelling,
CAH 10, edn. 2: 30 ff.
Augustus’ approach to the problem was more circumspect. In 26 bc
he supported the rebel Tiridates against Phraates, but this indirect
strategy proved unsuccessful (N–H vol. 1, p. xxxii and on 2. 2. 17). He
must have been reluctant to risk his legions across the Euphrates, yet the
poets talk as if a campaign was imminent (1. 2. 51, 2. 9. 20 f., 3. 3. 43 f.,
Prop. 2. 10. 13 f., 3. 4. 1 ff.); it was doubtless part of official policy to give
this impression (P. A. Brunt, JRS 53, 1963: 174 f. ¼ Imperial Themes,
1990: 104 ff.). In the end diplomatic pressure backed by the threat of
force produced a satisfactory result, and in 20 bc the standards were
restored; cf. 4. 15. 6 ff., epist. 1. 12. 27, 1. 18. 56 ‘sub duce qui templis
Parthorum signa refigit’, Aug. res gest. 29. 2, Vell. 2. 91. 1 with Wood-
man, RE 10. 351, CAH 10, edn. 2: 159 f. It is significant that little is said
about the prisoners; for exceptions cf. Justin 42. 5. 11 ‘itaque a tota
Parthia captivi ex Crassiano sive Antonii exercitu recollecti signaque
cum his militaria Augusto remissa’, Dio 54. 8. 1.
The story of Regulus, which occupies eleven of the ode’s fourteen
stanzas, reflects the official indifference to the prisoners’ fate. According
to the traditional account, he was captured by the Carthaginians in
255 bc, and five years later was allowed to return to Rome on parole
to negotiate an exchange of prisoners; but he persuaded the senate
not to agree to the deal, went back to Carthage in accordance with
his oath, and died there after terrible tortures. The first surviving
mention of this is in a fragment of C. Sempronius Tuditanus, cos. 129
bc (5 Peter ¼ Gell. 7. 4. 1), and the case seems to have been cited by the
eminent jurist Q. Mucius Scaevola, cos. 95 bc, in a technical discussion
of postliminium (the restoration of civic rights); cf. dig. 49. 15. 5. 3. The
demonstration of fides and fortitudo passed into the long series of Roman
exempla (Val. Max. 1. 1. 14, 2. 9. 8); Cicero repeatedly uses the story for
its moral implications (see especially paradox. Stoic. 2. 16, fin. 2. 65, off. 1.
39, 3. 97–115 with A. R. Dyck’s commentary), in which he is followed by
Seneca (dial. 1. 3. 9 etc.).
On the other hand, there is not a word about Regulus’ heroic mission
in Polybius’ account of his African campaigns (1. 25–35 with Walbank);
there he is censured for trying to impose unacceptable terms on the
Carthaginians after his victories—allegedly because he feared that a
successor might have the glory of taking Carthage (1. 31. 4)—and
subsequently for pleading for the mercy that he had denied to his
enemies (1. 35. 3). Walbank with other historians regards the whole
story as fictitious, and thinks Polybius may not even have heard it. Yet
we know that Polybius was following the account of Philinus, who
wrote from the Carthaginian point of view (Walbank 1, p. 93); and
even if he had heard the story he might have felt that Regulus’ self-
rehabilitation would have spoiled the cautionary tale of hubris and its
punishment (repeated later by Diodorus 23. 12–15). So perhaps it is best
to withhold judgment about Regulus’ mission, and also about the
manner of his death (see 49–50 n.). Whatever the truth may be, by the
time of Cicero and Horace the story had won acceptance and could be
used without hesitation for moral instruction. See further Klebs, RE 2.
2086 ff., P. Blättler, Studien zur Regulusgeschichte, 1945, E. R. Mix,
Marcus Atilius Regulus, Exemplum Historicum, 1970, OCD 207.
In the traditional accounts the Carthaginians intended Regulus to
negotiate an exchange of prisoners (Cic. off. 1. 39 ‘de captivis commu-
tandis’, Kornhardt, op. cit. 101 f.) In Horace, however, the issue is one of
ransom. His version echoes the debate that took place after the disaster
at Cannae in 216 bc. On that occasion Hannibal supposedly offered
to accept ransom for the Roman prisoners, and allowed a spokesman
to plead their cause in Rome (Liv. 22. 58. 4–6); he was opposed in
a speech by Torquatus, which is imaginatively reconstructed by Livy
(22. 60. 6–27). Hence Horace’s ironical ‘auro repensus scilicet acrior /
miles redibit’ (25–6) sounds like a reply to the spokesman’s conten-
tion ‘(utemini) nobis etiam promptioribus pro patria, cum beneficio
vestro redempti atque in patriam restituti fuerimus’ (Liv. 22. 59. 11),
and as such it parallels Torquatus’ words ‘pretio redituri estis eo unde
ignavia ac nequitia abistis?’ (22. 60. 16); in the same way Horace’s
‘flagitio additis / damnum’ (26–7) looks like a rejection of the spokes-
man’s general argument ‘neque enim vos pretio pepercisse homines
credent’ (22. 59. 19).
It is unlikely that Livy’s version of Cannae had appeared before
Horace wrote this ode, but some of his material had no doubt
been used by earlier historians (e.g. Acilius, Sempronius Tuditanus,
Claudius Quadrigarius). Ennius, too, may have been a significant
source. When Horace’s Regulus says ‘neque amissos colores / lana refert
medicata fuco’ (27–8), the words have been plausibly connected with
ann. 476 ‘quom illud quo iam semel est imbuta veneno’ (apparently from
the speech of Torquatus after Cannae); see Skutsch, pp. 635 ff., citing
Kornhardt, loc. cit. In her opinion Ennius may also have influenced
Silius’ account of Regulus, 6. 348 f. ‘poscentes [Poeni] vinctam inter
proelia pubem / captivamque manum ductore rependere nostro’ (cf. v.
18 of the ode ‘captiva pubes’) and of the Cannae debate, 10. 650 f.
‘obsecrantes captivum vulgus ut auro / pensarent parvo’ (cf. v. 25 of
the ode ‘auro repensus’). We may note also the famous speech that
Ennius assigns to Pyrrhus over a proposal to ransom prisoners: ‘nec mi
aurum posco nec mi pretium dederitis: / non cauponantes bellum sed
belligerantes / ferro non auro vitam cernamus utrique’ (183–5 with
Skutsch’s parallels); this context may lie behind Horace ‘pacem duello
miscuit’ (38 n.).
The intention of the ode has been disputed. Mommsen thought it
supported Augustus’ policy of renouncing the conquest of Parthia and
resisting clamour for the recovery of the prisoners (Reden und Aufsätze,
1905: 168 ff.); but Horace could not have done this by promising the
Princeps divine status when that conquest had been achieved (2–4).
Fraenkel declined to see a political purpose (loc. cit.), and held that
the magnificent story of Regulus was told for its own sake; but the
analogy with the prisoners of Carrhae is made unmistakable by the
emphatic hoc (13). In fact the poem looks forward to the ultimate defeat
of Parthia, but by arguing against the ransom of prisoners it discourages
any pressure for immediate results. Modern readers are understandably
outraged by the lack of sympathy for the victims of Carrhae (17 f.), but
here too the poet, speaking in his official voice, reflects the stern code of
a militaristic society: cf. Livy 22. 61. 1 ‘exemplum civitatis minime in
captivos iam inde antiquitus indulgentis’.
Metre: Alcaic.
1–2. Caelo Tonantem credidimus Iovem / regnare: ‘we have come to

believe’ and so ‘we believe’; as well as the common novi (‘I know’) cf.
epist. 1. 2. 5 ‘cur ita crediderim . . . audi’, H–Sz 318. caelo is set against
praesentem below (cf. 1. 12. 57 ff.), and so should be taken with regnare; cf.
Lucan 3. 318 ff. ‘sortisque deorum / ignarum mortale genus per fulmina
tantum / sciret adhuc caelo solum regnare Tonantem’. Tonans was
probably a standing attribute of Jupiter (Latte 81 n.), though his
Capitoline temple was not dedicated until 22 bc, and so it does not need
caelo; Martial 7. 60. 2 caps Horace when he says of Jupiter ‘quem salvo
duce credimus Tonantem’.
There is an a fortiori argument (as Porph. recognized): Jupiter wields
his power in the (distant) sky, but Augustus’ achievements will be visible
on earth; cf. Ov. trist. 4. 4. 20 ‘quorum hic aspicitur, creditur ille deus’,
Val. Max. 1. praef. (to Tiberius) ‘cetera divinitas opinione colligitur, tua
praesenti fide paterno avitoque sideri par videtur’, Curt. 8. 10. 1 ‘patrem
Liberum atque Herculem fama cognitos esse, ipsum (Alexandrum)
coram adesse cernique’. The same antithesis occurred in Hermocles’
panegyric of Demetrius Poliorcetes: !`ººØ b j ÆŒæa ªÆæ I
ıØ
Ł, = j PŒ
ıØ tÆ, = j PŒ YØ, j P æ
ıØ E Pb (, =
b b Ææ Ł ›æH, = P ºØ Pb ºŁØ, Iºº IºŁØ (Athen. 6.
253e ¼ Powell, Coll. Alex. p. 174. 15 ff.), but the Greek poet’s scepticism
about traditional gods would not suit the Augustan context; H’s credi-
dimus is not intended to imply any lack of conviction, whatever his
private reservations.
2–3. praesens divus habebitur / Augustus: praesens is a religious word

(KØÆ ) that describes a god’s presence on earth (Brink on epist. 2. 1. 15,
RAC 5. 853 ff.); it also suggests efficacy (Virg. ecl. 1. 41 ‘nec tam praesentis
alibi cognoscere divos’, N–H on 1. 35. 2, Haffter, op. cit. 134 ff., cf. Psalm
46: 1 ‘a very present help in trouble’), as will be evident when Augustus
conquers the Britons and Parthians. For bibliography on his religious
claims see the note on 3. 3. 11 f.
3–4. adiectis Britannis / imperio gravibusque Persis: the poem should

be dated to 27 bc when the Princeps assumed the name ‘Augustus’ (3. 4
and 3. 6 are earlier). When Augustus set out for Gaul in the same year, it
was believed that he intended to conquer Britain, where the agreements
made with Julius Caesar were probably now disregarded (see C. E.
Stevens in Aspects of Archaeology, Essays presented to O. G. S. Crawford,
1951: 332 f.); Dio mentions such plans in 34 (very dubious), 27 and 26 bc
(49. 38. 2, 53. 22. 5, 53. 25. 2). When the poets make similar remarks they
must be reflecting official attitudes (cf. A. Momigliano, JRS 40, 1950:
39 f., P. A. Brunt, JRS 53, 1963: 173 f. ¼ Imperial Themes, 1990: 103 f.);
see especially 1. 35. 29 f. ‘serves iturum Caesarem in ultimos / orbis
Britannos’, which belongs to the same period (Syme, 1978: 50 f.), Prop.
2. 1. 76, 2. 27. 5 ‘seu pedibus Parthos sequimur seu classe Britannos’.
Augustus may have been diverted by the Cantabrian War (see on 3. 14),
but at some stage British embassies were sent to renew submission
(Strabo 4. 5. 3, not dated, Brunt, op. cit. 173 f.). The closure of the
temple of Janus in 25 suggests that the question of annexation was left
hanging in the air.
For the intention of defeating the Parthians see the introduction
above. adiectis refers to the future; but H is not expressing doubt
about either their conquest or that of the Britons. For gravibus (here
‘formidable’) cf. 1. 2. 22 ‘quo graves Persae melius perirent’; H is thinking
of the great Parthian incursion of 40 bc (3. 6. 9–10 n.) as well as the
defeat of Crassus. The ablative absolute at the end of the sentence
suggests the language of officialdom (cf. the Fasti Triumphales); such
effects seem to be more frequent in H than in the other poets (Nisbet in
Adams–Mayer, 1999: 151). In using the names Persae and Medi for the
Parthians, who had occupied the former Persian empire, the Roman
poets perhaps took over some of the associations which those names had
for the Greeks.
5–6. milesne Crassi coniuge barbara / turpis maritus vixit?: the ques-
tion is indignant, and emphasizes the shame of the situation. coniuge
barbara should probably be taken as an ablative absolute explaining
turpis maritus ‘a disgraced husband, seeing that his wife is a barbarian’
(note the chiasmus); some take the ablative with turpis alone (Ov. am. 1.
6. 72 ‘lente nec admisso turpis amante, vale’), but then the phrases would
be less neatly balanced; others combine it with an adjectival maritus
(cf. Ov. her. 4. 134 ‘fratre marita soror’), but then turpis is hard to
accommodate. vixit implies that the soldiers preferred a disgraceful
life to an honourable death; for the shame of ‘going native’ (as it used
to be called) cf. Caes. bc 3. 110. 2 ‘nomen disciplinamque populi Romani
dedidicerant uxoresque (Aegyptias) duxerant’, Virg. Aen. 8. 688 ‘sequi-
turque (nefas) Aegyptia coniunx’ (highly tendentious).
6–8. et hostium /—pro ycuriay inversique mores!—/ consenuit socer-

orum in armis: pro(h) expresses indignation, often in a parenthesis; it
can be combined with the vocative of the god or persons invoked (Plaut.
Poen. 1122 ‘pro supreme Iuppiter!’), or a nominative referring to the
conduct that is deplored (Sen. suas. 7. 11 ‘pro facinus indignum!’, Sen.
dial. 11. 17. 4 ‘pro pudor imperii!’, TLL 10. 2. 1439. 35 ff.). inversi mores
must refer to the degenerate behaviour of the prisoners in marrying
Parthian women (as described in the surrounding sentence). RN obel-
izes curia, which does not combine well with the mores of the prisoners;
the senate could have had no influence on their marriages whether or
not we understand an epithet corresponding to inversi (thus Porph.
explains ‘o mores, inversi estis, et o curia, quam indignas res passa
es!’). Some think that H is blaming the senate for failing to recover or
avenge the prisoners, but apart from the irrelevance of the reproach, it is
not clear when they could have been expected to do this; Carrhae was
a distant memory, Antony had failed, and Parthian policy was now
the reponsibility of Augustus. Cornelissen conjectured saecula, which
suggests the degeneration of the times (3. 6. 17 n., Ter. ad. 304 ‘hoccin
saeclum! o scelera . . . ’); the phrase as a whole recalls Cicero’s ‘o tempora,
o mores!’ (Cat. 1. 2). W. S. Watt has proposed pro gloria, understanding
inversa (Latomus 54, 1995: 608 f.), but the participle does not quite fit.
NR is inclined to accept the view of Orelli, Heinze, and others, that
the senate was supposed to represent the honourable traditions of Rome
and yet had accepted the behaviour described in vv. 5–12. If change is
needed, he would favour Watt’s gloria as the most likely suggestion
so far.
It appears that the Parthians had settled some of the prisoners of
Carrhae at Antiocha Margiane (Merv) 400 miles east of the Caspian
(Plin. nat. hist. 6. 47 ‘in hanc Orodes Romanos Crassiana clade captos
deduxit’). If they were conscripted in the Parthian army, that is why
Propertius envisages Romans fighting Romans by the Caspian (2. 30A.
19 ff.): ‘num tu, dure, paras Phrygias nunc ire per undas, / et petere
Hyrcani litora nota maris, / spargere et alterna communis caede Penates,
/ et ferre ad patrios praemia dira Lares?’ (cited by N.-O. Nilsson, Eranos
45, 1947: 44 ff.); their position on the frontier suggests they were soldiers
or ex-soldiers. According to Justin (based on the Augustan Trogus),
the greater part of the Parthian army consisted of slaves, presumably
prisoners of war (41. 2. 5 ‘exercitum, non ut aliae gentes liberorum, sed
maiorem partem servitiorum habent’). If they were on garrison duty
rather than campaigning they could have married Parthian women
(Curtius 5. 5. 5 describes how Alexander met Greek prisoners who had
married Persians, though in that case they had not been conscripted).
Subsequently a number of Romans appeared at the capital of the Hun
general Chih-Chih 500 miles further east. Perhaps they were mercen-
aries who had managed to escape from the Parthians; 145 are said to
have been taken prisoner by the Chinese (see Homer H. Dubs, AJP 62,
1941: 322 ff.).
These considerations tell in favour of the manuscripts’ armis as
against Faber’s arvis; cf. also Liv. 5. 20. 8 ‘in quo (bello) prope con-
senuerint’, 32. 3. 5 ‘consenuisse sub armis’. Serving in the enemy’s army
(hostium before the parenthesis is emphatic) was more disgraceful than
tilling his fields; and commentators fail to notice that ‘sub rege Medo’
(9) follows much better from armis than from arvis (cf. epist. 1. 18. 55 f.
‘Cantabrica bella tulisti / sub duce . . . ’). Bentley claimed that arvis was
implied by the scholiasts’ notes (‘apud Parthos resedisse’ Porph., ‘velut in
propriis sedibus omnem iam aetatem agerent’ ps.-Acro); but in both his
lemma and his note on 6–8 Porph. has the phrase in armis. It is true,
however, that prisoners of war were at times used as agricultural labour-
ers; cf. epist. 1. 16. 70 of a captive ‘sine pascat durus aretque’, Eur. Rhes.
74 f. ƒ K æ
ØØ ØØ ººØ = 6æıªH IææÆ KŒŁøØ
ªÆE, 176, and in particular 23 f. of the present ode. Bentley also
thought ‘socerorum in armis’ an odd expression, as if the fathers-in-law
were in command, but socerorum simply implies Parthorum; the word
balances coniuge and maritus to emphasize the horror of marrying into
an enemy family.
9. sub rege Medo Marsus et Apulus: any rex was objectionable

to a Roman, a Parthian one doubly so. The alliterative Marsus is
pointedly juxtaposed with Medo; the Marsi from central Italy, whose
very name was associated with Mars, provided stalwart countrymen
for the Roman army (N–H on 1. 2. 39) as did H’s neighbours from
Apulia. The same was true of the Lucanians; according to Pliny it
rained iron in Lucania the year before Crassus was killed by the
Parthians, ‘omnesque cum eo Lucani milites quorum magnus numerus
in exercitu erant’ (nat. hist. 2. 147); it has been suggested that since
his defeat of Spartacus Crassus had particular influence in S. Italy
(P. Moore, CR 23, 1973: 13 f.).
10–11. anciliorum et nominis et togae / oblitus aeternaeque Vestae: the

original ancile fell to earth in the time of Numa; and as the fate of
empire was bound up with it he commanded eleven replicas to be made
so as to frustrate a prospective thief (Ov. fast. 3. 373 ff., Plut. Num. 13,
Dion. Hal. ant. Rom. 2. 71. 1–2, Serv. auct. on Virg. Aen. 7. 188). They
were associated with Mars and as such were carried by his priests, the
Salii (Bömer on Ov. fast. 3. 259, and for illustration see Beard–North–
Price 2. 128). H’s genitive plural is anomalous for ancilium (like vectiga-
liorum); cf. Prob. GL 4. 208. 26, Neue–Wagener, 1. 434.
The Romans’ name was the most powerful symbol of their identity,
transcending local loyalties (Cic. leg. 2. 5); for its vital importance cf.
Cic. Phil. 2. 51 ‘dum genus hominum, dum populi Romani nomen
exstabit (quod quidem erit . . . sempiternum)’, Phil. 11. 36, Verr. 1. 79,
Liv. 26. 41. 13 ‘(Hasdrubal) qui si se cum fratre coniunxisset, nullum iam
nomen esset populi Romani’. In the Aeneid too, after speaking of the
Romans’ name (1. 277), Jupiter goes on to speak of their dress:
‘Romanos, rerum dominos gentemque togatam’ (282).
In RN’s view nominis seems too abstract to combine with anciliorum
and togae, and it is strange to find the word after the less significant
anciliorum. One side of the manuscript tradition omits et before nominis
and the testimonia are also divided; the meaning would be ‘the very
word ancilia’; cf. Cic. Verr. 5. 108 ‘non te . . . hospiti ius atque nomen a
scelere . . . revocare potuit?’, Petr. 26. 3 ‘ne puella quidem . . . expaverat
nuptiarum nomen’, OLD s.v. 8. This reading deserves more attention
than it has received; although the genitive depending on a genitive is
awkward, it is not unexampled (Shackleton Bailey, Propertiana, 1956:
118, 223, H–Sz 65); as well as avoiding the difficulties of the other
reading it produces a list of three nouns where four might be too many.
Our passage may be imitated at Florus 2. 21. 3 (of Antony) ‘patriae
nominis togae fascium oblitus totus in monstrum illud . . . desciverat’,
where nominis presents the same problem.
The assumption of eastern dress could be a matter for reproach
(Thuc. 1. 130. 1 of Pausanias, Flor. loc. cit., Claud. in Rufin. 2. 82 ff.);
yet the toga may have meant as little to Marsian soldiers as the ancilia
(cf. Juv. 3. 171 ‘pars magna Italiae est si verum admitttimus in qua / nemo
togam sumit nisi mortuus’). Vesta was the Roman goddess of the hearth
(Beard–North–Price 1. 51 ff., Lacey 171 f., OCD 1591); her shrine in the
Forum (Steinby 5. 125 ff.) contained the eternal fire that guaranteed
Rome’s survival and was tended by the Vestals (Cic. leg. 2. 20). She is
mentioned together with the ancilia at Liv. 5. 52. 7, 5. 54. 7 (the famous
speech assigned to Camillus) ‘hic Vestae ignes, hic ancilia caelo
demissa’; Livy may have priority (Syme, Roman Papers 1. 416 ff., S. P.
Oakley on Livy 6–10, vol. 1, 109 f.), but the collocation was no doubt
common in patriotic discourse.
12. incolumi Iove et urbe Roma?: Porph. comments ‘mire incolumi Iove,
ut si diceret stante mundo, incolumi rerum natura’, but H is referring
specifically to the temple of Jupiter on the Capitol; cf. Livy 5. 52. 6 and
54. 7 cited in the last note. The phraseology alludes to an old religious
formula that was extended to other contexts: cf. Paul. Fest. 115M ¼ 102L
‘si sciens fallo, tum me Diespiter salva urbe arceque bonis eiciat’ (the
Fetial oath excludes the city from any evil consequences, cf. Polyb. 3. 25.
8, CIL 2. 172. 16–17), Caecil. CRF 146 ¼ Gellius 2. 23. 10 (comically of a
henpecked husband) ‘liber servio salva urbe atque arce’ (with the same
adversative implication as our passage), Cic. Vat. 21, Planc. 71, Juv. 9.
130 f. (parodic); see further Fraenkel (1960), 223 ff., 428, H. J. Rose, CQ
41, 1947: 79 f., Williams (1968), 363 ff.
13–15. hoc caverat mens provida Reguli / dissentientis condicionibus /

foedis: for the story of Regulus see the introduction. Extended mytho-
logical exempla with speeches were a feature of Greek lyric poetry
(cf. 1. 7 with N–H vol. 1, p. 93, epod. 13, Syndikus 2. 74 f.); for instances
that purport to be historical cf. 4. 4. 49 ff. (Hannibal), Bacchyl. 3.
29 ff. (Croesus). The vocabulary of these lines is termed ‘prosaic’
by Axelson (103), as suits the political context (see below, 53–4 n.):
dissentire is paralleled in the major poets only once in Lucretius and
Ovid and twice in H’s hexameters; condicio is seldom found (but cf.
1. 1. 12); pernicies below only at 2. 13. 4, twice in H’s hexameters, three
times in Silver Age poetry. foedis, itself a strong word (‘foul’), is empha-
sized by its position at the beginning of a line before a pause (cf. 1. 5. 12,
4. 9. 26).
15–16 et exemplo ytrahentisy / perniciem veniens in aevum: the MSS
and Porph.read trahentis, which is supposed to mean ‘deducing from a
precedent’, but then ex or ab might have been expected (Cic. div. 2. 121
‘ex insanorum . . . visis innumerabilia coniectura trahi possunt’, OLD s.v.
traho 12b), and veniens in aevum does not combine easily with either
trahentis or perniciem. The Commentator Cruquianus (i.e. the notes
derived by Cruquius from the lost codex Blandinianus) read exemplo
trahenti, which was independently proposed by Canter; trahenti (‘carry-
ing with it’) leads easily to veniens in aevum. NR finds this acceptable
but RN sees some difficulty in combining dissentientis with exemplo;
Regulus could disagree with the terms, but not with the precedent
(unless an awkward zeugma is posited). Bentley proposed exempli tra-
hentis, a genitive of quality with condicionibus (‘of a model carrying with
it destruction’), coordinate with the adjective foedis (cf. Liv. 4. 48. 12
‘largitione pessimi exempli’, Suet. Aug. 32. 1 ‘pleraque pessimi exempli in
perniciem publicam . . . duraverant’, TLL 5. 2. 1335. 15 ff.); but even if we
accept the combination of the single adjective foedis with a phrase of six
words, it is very confusing to have the genitive trahentis after the
rhyming but disparate dissentientis. On the same lines as Bentley, RN
has considered exemplo trahente as an ablative of quality depending on
condicionibus and co-ordinate with foedis; but it must be admitted that in
this context a genitive of quality is easier to parallel than an ablative.
17–18. si non periret immiserabilis / captiva pubes: periret represents

the pereat of direct speech. If the text is sound the third vowel of periret
is long, as in Plautus (Sonnenschein on rudens 390) and Ennius
(Skutsch on ann. pp. 58 f.); for this archaic usage cf. 2. 13. 16 ‘caeca
timet aliunde fata’ with N–H, Nettleship’s appendix to Conington’s
commentary on Virgil, vol. 3, pp. 472 f. A short syllable at this place in
the line would be unexampled in H, though it is an option in Alcaeus.
Conjectures include perirent (Glareanus, to which Bentley added immi-
serabiles), perires (Lachmann, but the apostrophe would be awkward
before the following speech), and the transposition of si non periret
and captiva pubes (W. M. Edwards cited by Campbell). immiserabilis
may be a Horatian coinage (cf. 2. 14. 6 illacrimabilem, 3. 6. 10 n.).
18–21. ‘signa ego Punicis / adfixa delubris et arma / militibus sine

caede’ dixit / ‘derepta vidi . . . : the Romans hung the trophies of battle
on their temple walls (Gallus, FLP 2. 4 f., cf. Virg. Aen. 11. 778 of
Camilla). H. attributes the same custom to the Carthaginians (Punicis
is indignantly emphatic); he may have had a precedent in a lost carmen
of Livius Andronicus (Marconi, op. cit., R. Verdière, Latomus 42, 1983:
383 ff.); see Serv. auct. on Aen. 4. 37 ‘Livius Andronicus refert eos (Afros)
de Romanis saepius triumphasse, suasque porticus Romanis spoliis
adornasse’. The same practice was also attributed (perhaps wrongly) to
the Parthians, who must have been in any Roman’s mind at this time
(4. 15. 7 f., epist. 1. 18. 56 ‘sub duce qui templis Parthorum signa refigit’).
In v. 21 direpta has much better authority than derepta; but the
confusion is common (4. 15. 6, Lucan 1. 240, Sil. 10. 599), and manu-
script evidence counts for little. deripere (‘to tear down’) suits objects on
a temple wall; deripere, as in Ov. am. 1. 5. 13, and diripere (‘to tear away’),
as in Ov. met. 9. 637, are used of clothing; when arms are stripped from a
corpse or captive the issue is not clear-cut, but derepta is supported
by parallels (Virg. Aen. 11. 193 ‘spolia occisis derepta Latinis’, Tac. ann.
2. 45. 3), and here makes a sharper contrast with adfixa delubris.
vidi is used of horrors that one has lived to see; cf. 1. 2. 13, Enn. scen.
97V ¼ 92J ‘haec omnia vidi inflammari’, Lucil. 1324 f. M ¼ 1218 f. W
‘vidimus <vinctum / thomice cann>abina’, referred by Cichorius to the
surrender of Mancinus to the Numantines (Untersuchungen zu Lucilius,
1908: 37 f.), Virg. Aen. 2. 499 with Austin.
21–2. vidi ego civium / retorta tergo bracchia libero: the first word of
the sentence repeats the last word of its predecessor (Wills 392 f.); vidi
ego is a common combination ( J. N. Adams in Adams–Mayer, 1999:
123 f.), in spite of the elision of the long i before the short o. For retorta
cf. epist. 2. 1. 191, Hom. Il. 21. 30 f., Soph. Aj. 71 f. with Jebb, Ov. am. 1. 2.
31 with McKeown’s parallels, (of Roman prisoners) Liv. 9. 10. 7, Vell. 2.
1. 5. Here tergo is probably dative ‘of place whither’ rather than ablative;
cf. Prop. 3. 24. 14 ‘in mea terga’, Plin. nat. hist. 35. 93 ‘ad terga’, Prud.
Symm. 2. 560 ‘in terga’.
It has been thought strange that H should say civium; a man ceased to
be a citizen when captured by the enemy (W. W. Buckland, The Roman
Law of Slavery, 1908: 291 ff.), and this was the attitude of Regulus
himself (42 n.). Rowell (loc. cit.) points out that the captured soldier
did not lose his citizenship until he had been brought within the
enemy’s camp (cf. Pomponius, dig. 49. 15. 5. 1 ‘antequam in praesidia
perducatur hostium, manet civis’); but this is probably an over-subtle
explanation. The use of the word, which balances libero, is meant to be
shocking. For libero applied to a part of the body cf. 1. 37. 1 ‘pede libero’
with N–H; Cornelissen’s livido (referring to the marks of flogging) is
not wanted.
23–4. portasque non clausas, et arva / Marte coli populata nostro:

open gates were a sign of security; cf. ars 199 ‘apertis otia portis’, Sall.
hist. 1 fr. 14M. ‘apertae portae, repleta arva cultoribus’. Here it made
things worse that the fields were being tilled by the Roman soldiers
who had earlier ravaged them (Polyb. 1. 29. 6, 1. 30. 14); for Regulus
had captured Tunis (1. 30. 15) and had he shown more moderation,
could have imposed terms on Carthage (1. 31. 5–8). The instrumental
ablative Marte refers to armed forces (OLD 7) and is combined with
both coli and populata; the word is used with indignation here, as is the
emphatic nostro.
25–7. auro repensus scilicet acrior / miles redibit. flagitio additis /

damnum: the prominently placed auro has the pejorative tone often
found in moralists; cf. especially Enn. ann. 185 Sk ‘ferro non auro’ (cited
in the introduction). repensus, which refers as elsewhere to ransom
(OLD 5b), suggests a literal balancing in the scales; cf. Cic. de orat. 2.
269 ‘pro Gracchi capite erat aurum repensum’ (Septumuleius was given
its weight in gold). The ironic scilicet is combined with acrior (‘keener’),
which is to be taken predicatively with redibit; on the other hand, miles
is the subject (pace Heinze, who wanted to take acrior miles as the
predicate). damnum refers to the financial cost of ransom, not (as Orelli
maintains) to the consequences for future generations; for the combin-
ation of loss and disgrace cf. serm. 1. 2. 52 f., 2. 2. 95 f. ‘grandes rhombi
patinaeque / grande ferunt una cum damno dedecus’, Plaut. merc. 237,
Pseud. 440, H. Usener, RhM 56, 1901: 7 ¼ Kleine Schriften 4, 1913: 362 f.,
Shakespeare, The Tempest iv. i. 210 f. ‘There is not only disgrace and
dishonour in that, monster, but an infinite loss’. The sentiment recalls
the debate after Cannae as recounted by Livy (see the introduction).
27–8. neque amissos colores / lana refert medicata fuco: instead of

saying ‘just as X never happens, so Y never happens’, H says ‘X never
happens, neither does Y’. The first clause is an observable fact of nature,
so the second can be presented as equally incontrovertible. Such para-
tactic comparisons avoid the prosaic complications of subordinate
clauses (cf. Pind. O. 10. 10 ff., Aesch. Ag. 322 ff., cho. 258 ff.); they are
notably common in H’s hexameters, perhaps reproducing a feature of
Greek popular philosophy (see the index to Mayer’s commentary on
Epistles I s.v. ‘parataxis’).
Wool, once dyed, never recovers its original whiteness; cf. Enn. ann.
476 Sk ‘quom illud quo iam semel est imbuta veneno’ (see the introduc-
tion), Quint. inst. 1. 1. 5 ‘nec lanarum colores quibus simplex ille candor
mutatus est elui possunt’, schol. Pers. 3. 37 ‘metaphora a lana quae
corrupta ad pristinum colorem reverti non potest’. Like these writers
H regards the dyeing as bad (cf. medicata, fuco, and 29 f. below); in
Lucretius 6. 1074 ff. it is morally neutral; in Plat. rep. 429–30, where
it refers to education, and Cic. Hortensius 96 Grilli it is good; cf.
K. Berger, ANRW II. 25. 2. 1141 ff. S. J. Harrison (CQ 36, 1986: 502 f.)
points out that in our passage the plural colores suits the various dyed
materials (as in Quintilian) better than the original colour of the wool;
he suggests that H has conflated the two images; alternatively one
might consider amissum . . . colorem, but H may not have wanted a
sequence of -am-, -um, am-, -um, -em all in one line (cf. 50–2 n.
below). medicare and medicamentum, in addition to the beneficial
sense, are used, like venenum and the Greek pharmakon, both of dyes
and poisons; for the latter sense cf. the colloquial ‘to doctor’. fuco, as
often, suggests artificiality and pretence (OLD 2, 4, 5).
29–30. nec vera Virtus, cum semel excidit, / curat reponi deterioribus:
vera virtus is a stock alliterative phrase (epist. 1. 18. 8, Plaut. cist. 198,
H. D. Jocelyn, The Tragedies of Ennius, p. 390); here it corresponds to
the undyed wool. excidit means ‘has been let slip’, balancing amissos; for
the verb’s application to non-material things cf. Ov. fast. 6. 393 ‘spes
excidit’, OLD 9, TLL 5. 2. 1238. 21 ff. curat personifies Virtus, and nec
curat represents a litotes for ‘is unwilling to be restored’. S. J. Harrison
(CQ 36, 1986: 503) suggests that this is a medical metaphor from the
dislocation and setting of limbs (Cels. 8. 11. 4, 8. 20. 4); but at this point
we expect not another analogy but the actual situation that is being
deplored. deterioribus, ‘men who have gone to the bad’ (Nonius 432M ¼
696L ‘peius a malo . . . deterius vero a bono’), suits the conventional
moral explanation of reponi rather than a medical metaphor.
31–3. si pugnat extricata densis / cerva plagis, erit ille fortis / qui
perfidis se credidit hostibus: the syllogism suggests the influence of
rhetorical exercises; Cicero mentions as a subject for debate ‘placeatne a
Karthaginiensibus captivos nostros redditis suis recuperari’ (de orat. 3.
109). A female deer does not show fight at the best of times, still less
when she has been removed from the net and is in the hands of the
hunters; the analogy is more damaging than its predecessor, as it implies
that the soldiers were cowards even before they were captured. tricae are
‘tangles’, though the word is only attested in a metaphorical sense; this is
the first recorded instance of extricare in its literal meaning, but like
irretire it may have been a huntsman’s word. For hunting nets cf. RE 20.
2. 1953 ff., Anderson, 1985: 38 ff., Bömer on Ov. fast. 5. 173.
se credidit means ‘trusted himself to the mercy of ’ (cf. Cic. fam. 4. 7. 3
‘victori sese crediderunt’, Liv. 36. 13. 8). perfidis and se credidit are
juxtaposed to suggest the culpable credulity of the prisoners (cf. 3. 27.
25 f. ‘doloso / credidit tauro latus’); Bentley’s dedidit or tradidit and
Campbell’s tentative perfide lose this point. For the bad faith attributed
the Carthaginians cf. 4. 4. 49 ‘perfidus Hannibal’, Sall. Jug. 108. 3
‘Punica fide’, Liv. 21. 4. 9 ‘perfidia plus quam Punica’ (of Hannibal),
Sil. 6. 482, Otto 291. When H puts a single monosyllable before the
main break of an Alcaic hendecasyllable, it is normally a weak word,
here an unemphatic pronoun (N–H vol. 1, p. xli); for exceptions cf. 3. 21.
10, 4. 14. 33.
34–6. et Marte Poenos proteret altero / qui lora restrictis lacertis /
sensit iners timuitque mortem,: restrictis lacertis repeats the indignity
described in 21–2, and also provides an analogy with the nets of 32. sensit
iners makes something of an oxymoron: the prisoners ‘felt’ their bonds
(N–H on 2. 7. 9) but remained passive. ‘He feared death’ was the
harshest comment that could be made about a Roman soldier; similarly
the Spartan who escaped from Thermopylae was referred to as › æÆ
(Herod. 7. 231).
37–8. yhicy unde vitam sumeret inscius / pacem duello miscuit: NR is

inclined to accept hic in spite of the preceding ille (32); following
Campbell, he points out that ille belongs to a generalization applicable
to all wars (31–3), whereas hic refers specifically to the Roman who has
surrendered in the current war (34–6); it is because he has surrendered
that he should not be ransomed. Shackleton Bailey, like Kiessling, takes
the sentence as interrogative, though he admits the obscurity and
obelizes hic. Courtney has proposed nunc (Phoenix 40, 1986: 320 f.); he
interprets ‘if deer demonstrate pugnacity, the soldier who has shown
himself a coward will become brave, but in reality (F ) he has
confounded peace with war’.
In RN’s view these theories do not meet Housman’s point (Classical
Papers 1. 101) that Regulus is arguing against ransoming the prisoners
(the relevant issue in H’s own day); if a strong pause were accepted after
mortem (36), too much emphasis would be placed on the original
surrender, about which nothing could be done. He proposed (partly
following Bentley) ‘timuitque mortem / hinc unde vitam sumere iustius,
/ pacemque bello miscuit’; like other scholars he explained that the
soldier should seek life from battle rather than surrender (cf. Porph.
ad loc. ‘cum deberet, inquit, vitam virtute atque armis sumere’, Sall. Jug.
39. 1 ‘dedecore potius quam manu salutem quaesiverat’). Housman’s
proposal, however, involves considerable change without achieving
total clarity.
RN thinks that Housman’s difficulty might be met if we put a comma
after mortem (36) and took 37–8 as an asyndetic relative clause (i.e. still
under qui), which turns to the question of ransom. In that case hic has
displaced a word that marks the transition, perhaps Courtney’s nunc,
which RN would understand differently: it would refer to the time of
Regulus’ speech in Rome, which was delivered long after the battle.
It may be objected that pacem does not refer specifically enough to
negotiation, so Campbell’s pactum should be seriously considered; cf.
Sall. Jug. 26. 1, hist. 4 fr. 69. 7 ‘pacto vitam dederant’, Liv. 22. 52. 3 (after
Cannae) ‘pacti ut arma atque equos traderent’ (followed by terms of
ransom). duello is justified by the antique historical context; it is an
archaism that H uses on five other occasions—more often than the
other Augustans (Axelson 26 f., Brink on epist. 2. 1. 254, p. 435). For the
use of miscere cf. Lucan 10. 75 f. ‘et miscuit armis / illicitosque toros et
non ex coniuge partus’.
38–40. o pudor! / o magna Carthago probrosis / altior Italiae ruinis!:

the recurrent words for shame and disgrace (6, 15, 26) lend weight to
Augustus’ intention of confronting Parthia; the exclamations o pudor,
pro pudor, heu pudor are well attested. In the following phrase some
editors wrongly put a comma after Carthago, as if the city were being
invoked; but the meaning is rather ‘to think of the elevation of great
Carthage’ (the ab urbe condita construction). altior refers to the city’s
pride and metaphorically to its enhanced height; ruinis is a causal
ablative (Manil. 4. 6 with Housman, Sil. 3. 585 ‘nobilior sit Roma
malis’). The emphasis on Italy rather than Rome might seem to suit
the Second Punic War rather than the First.
41–3. fertur pudicae coniugis osculum / parvosque natos ut capitis

minor / ab se removisse: fertur need not suggest disbelief; it means
simply that the story is traditional (N–H on 1. 7. 23); such expressions
can have various implications (N. Horsfall, PLLS 6, 1990: 49 ff.), but
here the tone is heroic. The virtuous wife and small children provide a
typical Roman tableau (epod. 2. 39 ff., Lucr. 3. 894 ff., Virg. georg. 2.
523 f.) that makes a contrast with the disgraceful union of the second
stanza; the kisses that Lucretius and Virgil leave to the children are here
allowed to the coniunx. She played some part in the historical tradition:
Cic. off. 3. 99 ‘esse domui suae cum uxore cum liberis’, Eutrop. 2. 25.
1 (from Livy) ‘itaque et uxorem a complexu removit’, Val. Max. 4. 4. 6,
Dio 11 (fr. 43) 27. Silius 6. 403 ff. says she was called Marcia and gives her
a significant role, perhaps derived from earlier accounts (RE 14. 2. 1601).
capitis minor means ‘diminished in respect of citizen rights’ (OLD,
caput 6). The genitive is a legal archaism; cf. 2. 14. 19, Lucil. 783 ‘capitis
dicturum diem’, Caes. bc 3. 110. 2 ‘capitis damnati’, TLL 3. 419. 49 ff., H–
Sz 76. minor recalls ‘capite (de)minuere’, ‘to deprive of citizen rights’; cf.
Liv. 22. 60. 15 (after Cannae) ‘deminuti capite, abalienati iure civium,
servi Carthaginiensium facti’, Fest. 70M ¼ 61L ‘deminutus capite
appellatur . . . qui in hostium potestatem venit’, RE 3. 1523 ff., TLL 3.
420. 53 ff., Kornhardt, op. cit. 85 ff. ab followed by a consonant also has
a somewhat archaic tone (cf. 1. 28. 29, 3. 16. 22, 3. 17. 1) and is therefore
to be preferred to the variant a, which is found in the Odes only at 1. 21.
14, 4. 5. 12.
43–4. et virilem / torvus humi posuisse vultum: virilem completes the

family picture presented by coniugis and natos. torvus (‘grim’ or ‘fierce’)
is also a male word; it is used of Mars (1. 28. 17), the younger Cato (epist.
1. 19. 12), and of wild creatures such as lions, boars, and bulls (OLD 1b).
But here Regulus lowered his gaze in shame, like an embarrassed
woman; cf. Virg. Aen. 3. 320 of Andromache ‘deiecit vultum’, 11. 480
of Lavinia ‘oculos deiecta decoros’; for similar expressions in various
situations see Hom. Il. 3. 217, Ap. Rhod. 3. 22 with Hunter, Virg. Aen. 7.
249 f. with Horsfall, OLD s.v. deicere 5b, and especially F. Muecke, BICS
31, 1984: 108 f.
45–6. donec labantis consilio patres / firmaret auctor numquam alias

dato: the vocabulary is political: consilium is the judicious advice that the
Romans prized so highly (here described as unique because it was so
disinterested); auctor suggests the moral authority behind the recom-
mendation, which was not strictly a formal proposal as Regulus no
longer considered himself a senator (cf. Sil. 6. 459 f.) or even a Roman
(Eutrop. 2. 25. 1, Dio 11. 27). firmaret (¼ confirmaret) balances labantis;
the verb’s mood is usually explained by its position in a subordinate
clause under fertur, and that is supported by properaret (48); yet if we
take the sentence as it comes, it is hard not to see also a nuance of
purpose in firmaret (‘until he should strengthen the wavering senators’).
47–8. interque maerentis amicos / egregius properaret exul: amicos

extends the sequence of wife, children, and senators; they might literally
go into mourning, as Cicero’s supporters did when he was banished
(Cic. Pis. 17 ‘vestis mutatione’). egregius exul produces an oxymoron, for
although the disgrace of exile, with its loss of citizenship and property,
was acutely felt, here the sentence was self-imposed by a man universally
admired; the effect is heightened by assonance of eg- and ex-, though
the quantity of the vowel changes.
49–50. atqui sciebat quae sibi barbarus / tortor pararet: atqui is found
14 times in Horace (including 1. 23. 9, 3. 7. 9), 4 times in Catullus, but
otherwise in the major poets only at Virg. georg. 3. 526 (Axelson 103 f.);
its prosaic quality again suits the tenor of the ode. (For ‘prosaic’ features
of H’s style see below on 53–4.) For the torture of Regulus cf. Tubero ap.
Gell. 7. 4. 3 (where he is put in a dark dungeon and then suddenly forced
to look into strong sunlight), ‘palpebras quoque eius, ne conivere posset,
sursum ac deorsum diductas insuebant’; Cic. Pis. 43 speaks of him being
shut up in a machina, which according to Val. Max. 9. 2. ext. 1 had sharp
spikes all round the inside; Cic. off. 3. 100 says ‘neque vero tum ignor-
abat se ad crudelissimum hostem et ad exquisita supplicia profiscisci’
(surely behind Horace), cf. Val. Max. 1. 1. 14, Sil. 6. 539 ff. (clearly based
on the annalists). The Carthaginians had a reputation for cruelty (Dyck
on Cic. off. 1. 38); in particular they were accused of human sacrifice
(Diod. 20. 14. 4–7, Tert. apol. 9. 2), a belief confirmed by archaeological
discoveries (S. Lancel, Carthage, a History, Eng. trans. 1995: 24); so
barbarus here means not only ‘neither Greek nor Roman’ but ‘barbarous’
in the modern sense (cf. 4. 12. 7). Yet according to Tuditanus (ap. Gell. 7.
4. 4), the sons of Regulus imposed equally dire punishments on some
eminent Carthaginian prisoners, and Diodorus (24. 12) says that his
widow, though attributing his death to neglect, tortured some noble
Carthaginians in revenge.
50–2. non aliter tamen / dimovit obstantis propinquos / et populum

reditus morantem: dimovere is used for clearing a path through a crowd
(cf. Tac. hist. 3. 80. 2 ‘occiditur proximus lictor dimovere turbam ausus’,
OLD 1b); the ob- of obstantis is set against di-. propinquos belongs to the
tradition about Regulus (Cic. off. 1. 39 ‘cum retineretur a propinquis et
ab amicis ad supplicium redire maluit’). populum moves beyond propin-
quos and amicos (47) to a still wider circle of supporters; it too may be
derived from the historical tradition (cf. Sil. 6. 494 f.). For the poetic
plural reditus of a single individual cf. Tib. 1. 3. 13, Stat. Theb. 3. 369, K–S
1. 85; reditum would produce an unwanted homoeoteleuton with populum
and cause confusion with morantem.
53–4. quam si clientum longa negotia / diiudicata lite relinqueret: a

patronus could advise his clientes on their legal problems (cf. epist. 2. 1.
104 ‘clienti promere iura’ with Brink’s note); such business might be
tedious and time-consuming (cf. epist. 1. 5. 30 f. ‘rebus omissis / atria
servantem postico falle clientem’). Here, as shown by lite, Regulus is
pictured as arbitrating in a dispute between clients; this is a very Roman
analogy for Hom. Od. 12. 439 f. q Kd æ Icæ IªæBŁ I
= Œæø ŒÆ ººa ØŒÆø ÆNH (cited by L. V. Hinckley, CB
55, 1978–9, 56 ff.). Some commentators think that Regulus is described
as representing his clients in court (cf. epist. 1. 7. 76 where Philippus goes
to his estate in the legal vacation), but the plural clientum is then less
convincing; the scene is more effective if he is pictured as delivering the
judgments in his own atrium (perhaps an analogy for his decisiveness in
the dispute over his return). Once again the prosaic vocabulary suits the
political tone of the ode, as does the ‘officialese’ ablative absolute; even
negotium, which is found seven times in H’s hexameters and once
elsewhere in the Odes (3. 29. 49), is avoided by the other major poets
before Juvenal (Axelson 107).
It was Axelson’s thesis that in the Odes H persistently used words that
were avoided by the grander Roman poets. This is not always to be
explained by the difference in subject-matter (Williams, 1968: 743 ff.); in
his lyrics H evolved a varied style that in some respects was less
‘dignified’ than that of epic (see P. Watson, CQ 35, 1985: 430 ff., who
refines Axelson’s thesis, taking account of Williams’s objections; cf also
F. Muecke, Encicl. oraz. 2. 776). Catullus and his contemporaries in their
more ambitious poems had favoured a diction that excluded many
common words, and in this they were followed by Virgil in the Aeneid
and the epic poets of the Silver Age (Lucan less than the others).
Horace, who aimed at a more ‘masculine’ style than the neoterics,
followed the more liberal practice of earlier Roman poets; in particular
he had no inhibitions about using words with an official, legal, or even
commercial resonance (1. 3. 5 ff., 4. 7. 13); Shakespeare is an obvious
parallel. Unfortunately Axelson regarded such features as a sign of
inadequacy, and even now it is difficult to use words like ‘prosaic’ and
‘unpoetical’ without seeming to imply a criticism that is not intended.
55–6. tendens Venafranos in agros / aut Lacedaemonium Tarentum:

for the verb cf. Quint. 11. 2. 38 ‘dum rus tendit’ (on the way to his estate
Scaevola remembers where he went wrong in a board-game). Venafrum
was a town on the Via Latina 100 miles south-east of Rome (RE 8A. 1.
668 ff.); it had a fine situation over the upper Volturnus and was famous
for its olive groves (N–H on 2. 6. 15 f. ‘viridi . . . Venafro’, there also
combined with Tarentum). It is implied that Regulus had an estate
there, which might have been acquired in the Samnite Wars; M. Atilius
Regulus Calenus, cos. 335 and A. Atilius Caiatinus, cos. 258 owe their
names to Cales and Caiatia further down the Volturnus valley, and the
family is attested in Campania from the third century bc (M. Freder-
iksen, Campania, 1984: 231), though there seems to be no evidence for a
presence as far north as Venafrum. Our Regulus was said to have had a
modest property in regione Pupiana (Val. Max. 4. 4. 6), but that was on
poor land near Rome. H is imitated by Macaulay, History of England,
ch. 9 (on the revolution of 1688 ) ‘But the iron stoicism of William never
gave way; and he stood among his weeping friends calm and austere, as
if he had been about to leave them only for a short visit to his hunting
grounds at Loo’; imaginative fictions of this kind follow the expansive
traditions of Hellenistic historiography.
Tarentum was a favourite resort in H’s own day (2. 6. 10 with N–H
vol. 2, pp. 94 ff., epist. 1. 7. 45); a historical connection with Regulus must
be an anachronism. It was traditionally founded by the Spartan Pha-
lanthus in the eighth century (N–H on 2. 6. 11, T. J. Dunbabin, The
Western Greeks, 1948: 29 ff.), but was now associated with relaxation and
pleasure. The long epithet Lacedaemonium produces an impressive
closure (cf. the last lines of 3. 1 and 3. 6), but it is not simply a learned
cliché (cf. Virg. Aen. 10. 28 ‘Aetolis . . . Arpis’ with Harrison); for its
connotations recall the unbending character of Regulus. Yet, as in the
case of Tarentum, it too may have an ambivalent function, because for
some readers, like NR, the succession of liquids and nasals (l, m, n, r)
contributes to the quiet ending; others, including RN, remain sceptical.
6 . D ELI C TA MA I O RV M
[B. Fenik, Hermes 90, 1962: 82 ff.; Fraenkel 285 ff.; E. Kraggerud, SO 70, 1995: 54 ff.; Lyne
(1995), 173 ff.; Pasquali 706 ff.; G. Williams, JRS 52, 1962: 28 ff. and (1968), 610 ff.]
1–16. You will continue, Roman, to pay for the omissions of your elders
until you repair the temples; you owe your dominion to your piety, and it is
the neglect of religion that has brought disasters to Italy: the Parthians have
blunted Antony’s attacks, and Cleopatra’s barbarians have threatened the
city. 17–32. The rot began in the home: young girls think of nothing but sex,
and when married take disreputable lovers with the connivance of their
husbands. 33–44. The stalwart rustics who defeated Hannibal and other
enemies did not spring from parents like these, but had stern mothers who
made them collect firewood. 45–8. In recent times each generation has shown
a decline from its predecessor, and the next will be even worse than ours.
The ode begins with a visible sign of national decay: many of Rome’s
temples need cleaning and repair. Fire and flood had played a part
(P. Gros, Aurea Templa, 1976: 18), but decades of neglect were mainly
responsible; cf. Nep. Att. 20. 3 ‘cum aedes Iovis . . . vetustate atque
incuria detecta prolaberetur’, Prop. 2. 6. 35 f. ‘sed nunc immerito velavit
aranea fanum / et male desertos occupat herba deos’, Ov. fast. 2. 58 (on
shrines of Juno) ‘longa procubuere die’. Yet Horace’s criticisms are
exaggerated (cf. J. A. North, JRS 76, 1986: 251 ff.); in 36 Domitius
Calvinus had rebuilt the Regia, and in 33 Sosius may have begun the
restoration of the Aedes Apollinis in Circo (Syme, 1939: 241, though the
dates of some of his examples are uncertain); when the poet complained
‘quare / templa ruunt antiqua deum?’ (serm. 2. 2. 103 f.) he was still
looking for private benefactors. But in 28 bc Octavian, using the wealth
taken from Egypt, produced a programme of his own, which in view of
its extent must have concentrated on repair rather than rebuilding
(though the Palatine temple of Apollo was completed in that same
year); cf. res gestae 20. 4 ‘duo et octoginta templa deum in urbe consul
sextum ex auctoritate senatus refeci, nullo praetermisso quod ex tempore
refici debebat’, Suet. Aug. 30. 2 (see also ch. 29 for later constructions),
Dio 53. 2. 4 (which says that pressure was put on private individuals to
restore temples built by their ancestors).
That must be the context of our poem, though the Princeps is not
directly mentioned; some have argued that its pessimism does not suit a
date so soon after the triumphs of 29 (Fenik and Kraggerud, opp. citt.),
but it would be impossible to say ‘donec templa refeceris’ when the
repairs of 28 were largely complete. In a year or two Livy could describe
Augustus as ‘templorum omnium conditorem ac restitutorem’ (4. 20. 7),
words echoed by Ovid, fast. 2. 63 ‘templorum positor, templorum sancte
repostor’.
Horace goes on to say that Rome had prospered because of her
obedience to divine will (5–6); such claims reflect the orthodoxy of the
Republic (5 n., 6 n.), and they went on being made till they were refuted
by Alaric’s sack of Rome in ad 410. The civil wars are treated as a
punishment for the neglect of religion (7–8) (Cicero had said the same
thing in Caesar’s senate (1 n.), though he at least knew better); Antony’s
failure to deal with the Parthian menace (9–12) is attributed to the
disregard of the auspices (as in the Republic, political rivals were always
eager to detect infringements of ritual requirements). The war of
Actium is represented as a threat to the city from foreign enemies
while she was preoccupied with civil dissension (13–16); again it is
implied that the ultimate cause is the decline of religion. In all this
the poem reflects the programme of the Princeps by re-asserting trad-
itional beliefs; but unlike its three predecessors it says nothing about his
practice of associating himself with the various rituals and observances
of the state religion.
Horace goes on to connect the national decline with sexual immoral-
ity (17–32); as usual his strictures are directed at the adultery of married
women and the connivance of their husbands, not at men’s extra-marital
associations with freedwomen or slaves (for the traditional ‘double
standard’ cf. Treggiari 199 ff., 299 ff.). The need for tighter controls
had already been canvassed during Caesar’s dictatorship (Cic. Marc. 23
‘comprimendae libidines, propaganda suboles’), and it was later to find
expression in Augustus’ social legislation—the lex Iulia de maritandis
ordinibus of 18 bc, the lex Iulia de adulteriis coercendis of about the same
date, and the milder lex Papia Poppaea of ad 9 (Treggiari 60 ff., 277 ff.
and CAH 10, edn. 2: 886 ff.). A primary aim was to restore the birthrate,
at least in the upper classes, as can be seen from the rewards and
penalties of the leges Iuliae (Brunt, 1971: 558 ff., A. Wallace-Hadrill,
PCPS 27, 1981: 58 ff.); even Augustus could not have foreseen the
capacity of the oligarchy to renew itself by incorporating new men.
But beyond this he clearly believed that the health of the nation
depended on a stable family life (K. Galinsky, Philol. 125, 1981: 126 ff.,
less sympathetically L. F. Raditsa, ANRW II. 13. 278 ff., C. Edwards, The
Politics of Immorality in Ancient Rome, 1993: 34 ff.). Orators had de-
nounced metropolitan laxity since the days of Cato (cf. fr. 222 M.),
Metellus Macedonicus (frr. 4–7 de prole augenda), and Scipio Aemilianus
(frr. 17, 30); historians said the same thing from Polybius (31. 25. 4–5) to
Sallust (Cat. 13. 3 etc., Earl 45 ff.), and however over-simplified its
expression, the underlying attitude must have been shared by many
Romans, not least matronae. The topic continued to excite the indigna-
6 . D ELI C TA M A I O RV M 99
tion of moralists, as we see from Sen. ben. 1. 9. 3–4, Juv. 1. 55 f. (with
Mayor) and 6. 21 ff.
Moderns sometimes find it hard to see how a hedonistic bachelor like
Horace could support such doctrines. NR sees no reason to think that
Horace’s view of the national interest was at variance with that of the
Princeps: he could advocate military training without feeling any desire
to take part in operations (he had had quite enough of that); he could
urge the rebuilding of temples and the revival of traditional rituals as a
way of promoting national solidarity, without accepting the concomi-
tant beliefs; and he could support the institution of marriage without
becoming a husband or father. RN would rather not be drawn into
speculations about the poet’s sincerity: it was not his business to formu-
late social policy but to write an effective poem (cf. Ll. Morgan, Patterns
of Redemption in Virgil’s Georgics, 1999: 5 ff.). He proceeds indirectly
(Lyne, 1995: 57), not by praising social legislation (carm. saec. 18 f. ‘decreta
super iugandis / feminis’), but by describing vividly contrasting life-
styles in a way that could evoke a response from his readers.
There is reason to believe that at the time of the ode Octavian had
already tried to cope with these problems; see P. Jörs, Festschrift
T. Mommsen, 1893: 4 ff., G. Williams, op. cit. 28 f. The crucial evidence
comes from Prop. 2. 7. 1 ff., written in 28 or 27 bc: ‘gavisa est certe
sublatam Cynthia legem / qua quondam edicta flemus uterque diu / ni
nos divideret’; the poet goes on to describe how Octavian’s law might
have forced him to marry someone else (8 ff.). There is surely some
exaggeration in all this (F. Cairns, Grazer Beiträge 8, 1979: 188 ff.), but
whatever the exact truth, it is not enough to posit an old tax on
bachelors designed to raise money before Actium (so E. Badian, Philol.
129, 1985: 82 ff.); when Propertius says ‘unde mihi patriis natos praebere
triumphis?’ (13), this implies that the aim of the law was social rather
than fiscal (cf. M. Beck, Philol. 144, 2000: 309). It may be that quondam
just means ‘previously’ (cf. Virg. Aen. 11. 819 ‘purpureus quondam color
ora reliquit’), and sublatam legem may be used imprecisely of a proposal
withdrawn rather than a law rescinded. The theory of failed legislation
is supported by 3. 24. 25 ff.; there Horace, after calling on the Princeps to
curb licentiousness (28 f. ‘indomitam audeat / refrenare licentiam’),
ruefully admits that his efforts will be appreciated only after his death
(note also 33 f. ‘quid tristes querimoniae, / si non supplicio culpa recidi-
tur?’). See further Livy, praef. 9 ‘nec vitia nostra nec remedia pati
possumus’, which in isolation might be entirely general (R. Syme,
HSCP 64, 1959: 42 f. ¼ Roman Papers 1. 416 f., Badian, op. cit. 92),
but in conjunction with Propertius and Horace suggests something
rather specific.
Horace next turns from present-day urban immorality to the rustic
innocence of former times (33–44); family influence continues to be
emphasized, the strict mother being contrasted with the adulterous
wife. This idealization of country life was common in later Greek
moralizing (Dio Chrys. 7, R. Vischer, Das einfache Leben, 1965), and
since the time of Cato had a particular appeal to rich Roman land-
owners. It suited the new emphasis on Italian agriculture when many
veterans were being resettled, and it had recently found powerful ex-
pression in Virgil’s Georgics (37–8 n.). With his Apulian origin and his
Sabine villa Horace could describe forcefully the peasant virtues, and
when he reflects that the modern generation could never have dealt with
Hannibal (33 ff.), that was a thought which might sometimes occur to
any middle-aged patriot. Once again Horace makes his case effective by
a judicious selection of stereotypes. But the rhetoric involves some
exaggeration; for the highly effective Roman army was still recruited
from rural Italy.
The last stanza pictures Rome in continuous decline, with each
generation worse than its predecessor. This was a commonplace of
poetry since the ‘golden ages’ of Hesiod and Aratus (see the index of
Lovejoy and Boas 1935, Gatz 1967). Moral degeneration is also a recur-
rent theme in the historians (Sall. Cat. 5. 9 with Earl 41 ff., Liv. praef. 5),
who give various dates in the second century bc for the start of the rot;
Juvenal capped conventional treatments by saying that future ages could
not get any worse (1. 147f. ‘nil erit ulterius quod nostris moribus addat /
posteritas, eadem facient cupientque minores’). Such unrelieved pessim-
ism, though unexpected in the Roman Odes, may reflect Octavian’s
disappointment at the failure of his attempts at reform. It is harder
to account for the contradiction with the opening stanza (Williams,
op. cit. 32 f.): there it is apparently said that the current generation is
not responsible for its predecessors’ neglect of the temples (immeritus),
a neglect that is held to be typical of the attitudes that led to the civil
wars (7–8). Even if this problem is removed by emendation (1 n.),
some awkwardness remains: the guilt can be expiated by the repair
of the temples (2), and in 28 bc Octavian was actually engaged in
such an operation, yet the last stanza speaks of irreversible decline.
Perhaps in combining the promise of the building programme with
the failure of the social programme Horace has not achieved total
consistency.
Metre: Alcaic.
1. Delicta maiorum immeritus lues: delicta refers to the failure to keep

the temples in repair; the word bears its primary sense of ‘derelictions’,
‘sins of omission’ (cf. Cic. Mur. 61 on the exaggerations of Stoicism
‘omne delictum scelus esse nefarium’, TLL 5. 1. 460. 57 ff.); luere is a
sacral word for ‘atone’. delicta is often thought to describe the guilt of
the civil wars, but here the civil wars, as lines 7–8 show, are the
punishment rather than the offence; cf. Cic. Marc. 18 ‘di immortales . . .
poenas a populo Romano ob aliquod delictum expetiverunt, qui civile
bellum tantum et tam luctuosum excitaverunt’. maiorum (‘fathers’ or
‘elders’) must include the previous generation; to translate ‘ancestors’
would suggest something too remote. For the idea cf. Eur. TGF 980N
a H Œ ø ºÆ N f KŒª ı = ƒ Łd æıØ, Exodus
20: 5 ‘I, thy God, am a jealous God, visiting the iniquity of the fathers
upon the children’, Pease on Cic. nat. deor. 3. 90 (citing many parallels
from Greek tragedy and elsewhere). Sometimes the thought is added
that the punishment is undeserved; cf. 1. 28. 30 f. ‘neglegis immeritis
nocituram / postmodo te natis fraudem committere?’, Solon 13. 31 f.
IÆØØ æªÆ ıØ = j ÆE ø j ª Kø.
But in our passage immeritus causes some difficulty; the current
generation must also have been neglectful of religion. Peerlkamp pro-
posed meritus (which destroys the necessary word-break after the fifth
syllable), Lehrs heu meritus; RN has considered et meritus (‘and what is
more, deservedly’). Normally such a comment would come at the end of
a clause (cf. serm. 1. 6. 22 ‘vel merito’), but a poet might be allowed a
more intricate word-order; for et merito even at the beginning of a
sentence cf. Prop. 1. 17. 1 f. ‘et merito, quoniam potui fugisse puellam, /
nunc ego desertas alloquor halcyonas’ (with Fedeli’s note).
2. Romane, donec templa refeceris: the vocative singular suits an oracle

(which was addressed to a single inquirer, though he might represent a
community) or a quasi-oracular pronouncement; cf. serm. 1. 4. 85 ‘hunc
tu, Romane, caveto’ (parodic), Liv. 5. 16. 9 ‘Romane, aquam Albanam
cave lacu contineri’ (leading to the conclusion ‘sacraque patria, quorum
omissa cura est, instaurata ut adsolet facito’), Virg. Aen. 6. 851 ‘tu regere
imperio populos, Romane, memento’, Ov. fast. 4. 259, met. 15. 637, Val.
Max. 1. 8. 10 ‘nihil . . . ad te hoc, Romane, bellum’, App. civ. 1. 11. 97
Ł Ø, , PøÆE to Sulla, Dickey 209 f.
3. aedesque labentis deorum: an aedes was usually, though not invari-

ably, less important than a templum, which was a sacred space demar-
cated by the augurs (Gell. 14. 7. 7, Wissowa 467 ff., Beard–North–Price
1. 22 f.). For labentis (‘collapsing’) cf. Aug. res gestae 20. 2 ‘rivos aqua-
rum . . . vetustate labentes refeci’, TLL 7. 2. 783. 7 ff. deorum gains sig-
nificance from being placed outside aedes labentis (Darnley Naylor); the
importance of the gods in the argument is again underlined by the
position of dis (5) and di (7).
3–4. et / foeda nigro simulacra fumo: in all but the earliest period
important shrines had cult-images (Latte 150, Beard–North–Price 2.
2 f.); this followed the Greek practice (Burkert 88 ff.). The grime was
produced by the city’s smoke (3. 29. 12) over a long period (ps.-Acro ‘per
nimiam vetustatem sordentia’). Later on, the Christian fathers gloated;
cf. Jerome, epist. 107. 1 ‘auratum squalet Capitolium, fuligine et ara-
nearum telis omnia Romae templa cooperta sunt’, Arnob. 6. 16 ‘non
videtis spirantia haec signa . . . ut nidoribus atque fumo suffita ac deco-
lorata nigrescant?’
5. dis te minorem quod geris imperas: for se gerere with a predicative

adjective (which is not a Ciceronian construction) cf. Sall. hist. 3. 7, Liv.
2. 27. 3 ‘medium se gerendo nec plebis vitavit odium nec apud patres
gratiam iniit’, Sen. epist. 44. 3, Lact. inst. 5. 15 ‘quasi minorem se gesserit’
(a clear reminiscence of our passage), TLL 6. 1947. 59 ff. The thought
was traditional; cf. 1. 12. 57 ‘te minor laetum reget aequus orbem’, Cic.
har. resp. 19 ‘pietate ac religione atque hac una sapientia quod deorum
immortalium numine omnia regi gubernarique perspeximus omnes
gentes superavimus’, nat. deor. 2. 8 with Pease, 3. 5, Weinstock
249 f. For a more objective view cf. Polybius 6. 56. 7 ŒÆ Ø ŒE
e Ææa E ¼ººØ IŁæØ OØØ F ı
Ø a
, PøÆø æªÆÆ, ºªø b c ØØÆØÆ.
6. hinc omne principium, huc refer exitum: it was an old commonplace

that everything begins with the gods (Terpander, PMG 698.
1 ZF ø Iæ
) and is determined by them; for Stoic developments
cf. Norden (1913), 240 ff. (citing Paul, Romans 11: 36, Marc. Aur. 4. 23).
As a corollary, the success of human enterprises depends on divine
favour; cf. Pind. P. 10. 10 `! ºº, ªºıŒf IŁæø º Iæ

Æ Oæ ÆhÆØ, fr. 108a. Thus a poem might start with an
invocation (Terpander above, Arat. phaen. 1 EŒ ˜Øe Iæ
ŁÆ, Gow
on Theoc. 17. 1, Bömer on Ov. fast. 5. 111), a symposium with a libation,
and in Roman public life auspices were taken very seriously; cf. Cic. Vat.
14 ‘omnium rerum magnarum ab dis immortalibus principia ducuntur’,
Liv. 45. 39. 10 ‘maiores vestri omnium magnarum rerum et principia
exorsi a dis sunt et finem statuerunt’, Appel 61. If the outcome was
successful the gods must be duly thanked; if it went wrong, this showed
that they disapproved (cf. inauspicatos in v. 10).
In our passage we should not understand est with principium (for
wicked undertakings do not come from the gods); rather we must
supply from refer an imperative meaning ‘derive’ (cf. Cicero’s ducuntur
above). There is no need to understand omnem with exitum, for the
noun is more incisive without it; cf. Hes. op. 669 K E ªaæ º Kd
›H IªÆŁH ŒÆŒH (with West’s parallels), Ov. her. 20. 44 ‘exitus
in dis est’.
7–8. di multa neglecti dederunt / Hesperiae mala luctuosae: for negle-
gentia deum cf. Sall. Cat. 10. 4, Liv. 3. 20. 5; H’s personal construction
puts more emphasis on the gods’ deliberate action (cf. 3. 2. 29 f. ‘saepe
Diespiter / neglectus incesto addidit integrum’). The ‘many evils’ of civil
war probably include the struggle between Caesar and Pompey (cf.
Pollio’s history, as reflected in 2. 1) in addition to subsequent battles.
multa gains stress through its long separation from mala. For Hesperia as
a poetic name for Italy (originating from a Greek perspective) cf. 1. 28.
26 with N–H, 2. 1. 32 ‘Hesperiae sonitum ruinae’. As an Apulian
H thinks not just of Rome.
9–10. iam bis Monaeses et Pacori manus / inauspicatos contudit

impetus: that is to say, the Romans were defeated once by Monaeses
and once by Pacorus; for this use of bis cf. Virg. georg. 3. 33 ‘bisque
triumphatas utroque ab litore gentes’, which refers to two campaigns,
not four. In 40 bc, as the consequence of an abortive Roman attack on
Palmyra (App. civ. 5. 10. 39), the Parthians under Pacorus, along with
the renegade Q. Labienus, defeated Antony’s legate Decidius Saxa, and
overran Syria and much of Asia Minor; for this major disaster (reflected
in epod. 16. 11 ff.) see Dio 48. 25. 3, C. Pelling in CAH 10, edn. 2: 12 f. In
36 Antony invaded Parthia with perhaps sixteen legions, but after his
legate Oppius Statianus was routed (Plut. Ant. 38. 3), withdrew with
difficulty having suffered serious losses (A. N. Sherwin-White, Roman
Foreign Policy in the East, 1984: 307 ff.); the Parthian Monaeses was
prominent in this episode, and though for some of the time he claimed
to be acting on Antony’s behalf, he may have been playing a double
game (Pelling on Plut. Ant. 37. 1 and CAH 10, edn. 2: 31 f.). Porph. says
that H is referring to the Parthian victory over Crassus (at Carrhae in 53
bc) as well as that over Saxa; it is true that Crassus had neglected
unfavourable omens (Pease on Cic. div. 1. 29), but the victor is elsewhere
called Surenas (admittedly a family title rather than a personal name),
and no Monaeses is mentioned. The following stanza refers to events
much more recent than 53 (H is abusing Antony and Cleopatra in turn);
contrast res gest. 29. 2 ‘Parthos trium exercitum Romanorum spolia et
signa reddere mihi . . . coegi’, which refers to the defeats of 53, 40, and 36.
The MSS divide between inauspicatos (cf. Liv. 7. 6. 11 ‘inauspicatam
legem’) and non auspicatos; the former reading, with its iambic opening,
was legitimate in H’s Alcaics (3. 1. 2 n.), but rare enough to be exposed to
corruption. See Vollmer, ALL 15, 1908: 31 f.
11–12. y nostros y et adiecisse praedam / torquibus exiguis renidet: in

RN’s view nostros is not wanted after inauspicatos impetus, and at the
beginning of the line before a pause seems over-emphatic; Priscian, GL
2. 518 reads nostris (dative), for which Bentley unconvincingly compared
Hirt. bell. Gall. 8. 10. 4 ‘quae res etsi mediocre detrimentum . . . nostris
adferebat’. One would sooner look for an epithet with praedam to make
a contrast with exiguis. Shackleton Bailey has considered nostratem
(which he admits is unpoetical) or Romanam (HSCP 89, 1985: 156);
alternatively he proposed reading nostrorum with praedam, which Bent-
ley had tried with impetus. RN has suggested praeclaram; if this was
corrupted to praedam, then, because of the second praedam, rewriting
was inevitable. NR thinks that nostros can stand; for two epithets are
permitted when one is a possessive (3. 13. 15 f. ‘loquaces / lymphae
desiliunt tuae’), and the emphasis on the word may stress the indignity
of Roman defeats.
Persian kings honoured friends with necklaces (Herod. 3. 20. 1
æ æe æØÆı

Ø, Xen. Cyr. 8. 2. 8, anab. 1. 2. 27), but
the derogatory exiguis here points to something humbler. The Romans
associated torques with Eastern or Gallic peoples (D–S 5. 375 ff., RE 6A.
1800 ff.), though their own soldiers could wear them as booty or reward
(Gow–Page, Garland of Philip 1237). renidet (‘beams’) suggests the naive
pleasures of barbarians (again derogatory); Catullus 39. 2 uses the word
of a silly grin. The unusual infinitive with renidet follows the construc-
tion of gaudet, a typically Horatian development.
13–14. paene occupatam seditionibus / delevit Vrbem Dacus et

Aethiops: paene modifies delevit; for the word-order cf. 2. 13. 20. The
city was ‘engrossed in dissensions’ (OLD occupo 9); there is a touch of
paradox in the phrase, since occupatam implies an action of concen-
tration and seditionibus one of dispersion. For the thought cf. epod. 7.
9 f., 16. 9 ff. (pointing out that the civil wars were Parthia’s opportunity),
[Sall.] epist. 1. 5. 2. For the plural seditiones cf. Juno’s speech in 3. 3. 29,
rhet. Her. 4. 66 ‘nunc vestris seditionibus, o cives, vexor’, Suet. Ner. 26. 2.
Dacian inroads in this period are mentioned in 3. 8. 18, serm. 2. 6. 53
‘numquid de Dacis audisti?’, Virg. georg. 2. 497 ‘coniurato descendens
Dacus ab Histro’; but in view of the conjunction with Aethiops it seems
more relevant that they sided with Antony before Actium (Dio 51. 22. 8)
and figured in the triumphs of 29 (ibid. 22. 6). Aethiops is a derogatory
way of describing the Egyptian forces of Cleopatra (see 15–16 below).
The exaggeration is evident, yet it was believed that Antony intended to
invade Italy (1. 37. 16, cf. Dio 50. 9. 2), and that Cleopatra had
threatened to dispense justice from the Capitol (Dio 50. 5. 4, cf. N–H
on 1. 37. 6).
15–16. hic classe formidatus, ille / missilibus melior sagittis: for Cleo-
patra’s Egyptian fleet cf. Virg. Aen. 8. 705 f. (on Actium) ‘omnis eo
terrore Aegyptus et Indi, / omnis Arabs, omnes vertebant terga Sabaei’.
In reality, according to Plutarch (Ant. 61. 1) Antony had 500 warships in
all, of which only 60 were Cleopatra’s (64. 1, 66. 3). The Dacians, like
other northern peoples (3. 8. 23 n.), were associated with archery (Sil.
1. 324 ff.); missilibus in effect ¼ mittendis. melior makes a comparison in a
way familiar from epic; cf. Hom. Il. 4. 400 x
æØÆ
fi , Iªæfi Ð b
Iø, Virg. Aen. 5. 153 ‘melior remis’.
17–18. fecunda culpae saecula nuptias / primum inquinavere et genus

et domos: in moralizing discourse saeculum often implies criticism
(Löfstedt 2. 470 ff.; cf. M. Winterbottom, The Elder Seneca, vol. 2,
1974: index II, Commonplaces, On the Age); here saecula is not just
plural for singular, as the decline has been going on for a long time.
fecunda suggests that the generations have been prolific in misconduct
(OLD s.v. culpa 3b) rather than children; cf. Sil. 2. 498 ‘fecundum
in fraudes hominum genus’, CIL 6. 31711 ‘omnium virtutum fecundissi-
mae feminae’.
For inquinare cf. epod. 16. 64 ‘ut inquinavit aere tempus aureum’
(which suggests the literal contamination of metals); here the idea of
moral pollution suits the religious tone; cf. 23 ‘incestos amores’, 4. 5. 21
‘nullis polluitur casta domus stupris’. genus refers to purity of lineage
(cf. 4. 5. 23), domos to the stability of the family; cf. Virg. georg. 2. 524
‘casta pudicitiam servat domus’.
19–20. hoc fonte derivata clades / in patriam populumque fluxit:

Peerlkamp proposed labes (cf. Liv. 39. 9. 1 ‘huius mali labes ex Etruria
Romam velut contagione morbi penetravit’, Sil. 14. 596 of a plague); at
first sight this word (or Palmer’s tabes) seems not just to suit inquinavere
but to cohere better with fonte, derivata (which was used primarily of
irrigation), and fluxit. Yet clades can also be used of a spreading plague
(Lucr. 6. 1091, 1125 f. ‘haec igitur subito clades nova pestilitasque / aut in
aquas cadit aut fruges persidit in ipsas’, Cic. fr. de cons. 2. 50 (Traglia),
Ov. trist. 5. 4. 33 ‘subitae contagia cladis’, Amm. Marc. 19. 4. 4 ‘clades
illa . . . paulatim proserpens Atticam occupavit’); the political implica-
tions of the word suit the present context (cf. Sall. Jug. 85. 43 ‘ita
iniustissume luxuria et ignavia . . . rei publicae innoxiae cladi sunt’).
The collocation with fluxit is supported by Sen. Thy. 236 ‘hinc omne
cladis mutuae fluxit malum’.
For patriam populumque cf. Accius, praetext. 6 ‘portenta ut populo
patriae verruncent bene’, Ov. met. 15. 572 ‘patriae . . . populoque Quirini’,
Juv. 14. 70; such alliterative phrases are often found in official contexts
(E. Wölfflin 270 ff.). Bentley proposed ‘inque patres populumque fluxit’;
this gives a commoner collocation (Virg. Aen. 4. 682 with Pease), but
inque is inelegant (in H only at serm. 1. 3. 141), and the senators should
not be singled out in this censorious passage.
21–2. motus doceri gaudet Ionicos / y matura y virgo: motus describes
the gyrations of the dancer, whether those of a rustic (Virg. georg.
1. 350 ‘det motus incompositos’), or a Salian priest (Fest. 334L), or
as here a provocative saltatrix; cf. Ov. am. 2. 4. 30 ‘et tenerum molli
torquet ab arte latus’ with McKeown, copa 2 ‘crispum sub crotalo
docta movere latus’, Sen. epist. 90. 19. In the classical Greek world
Ionians were thought effeminate (Ar. Thesm. 163, Otto 177), and their
dances came to be regarded as indecent; cf. Athen. 14. 629e q Ø ŒÆd
I øØŒc Zæ
Ø ÆæØ , Plaut. Pers. 826, Pseud. 1275, Stich. 769
‘qui Ionicus aut cinaedicust, qui hoc tale facere possiet’ (all referring
to men), L. B. Lawler, TAPA 74, 1943: 60 ff. For moral criticisms of
dancing cf. Scipio Aemilianus, orat. fr. 30M. ‘eunt, inquam, in ludum
saltatorium, inter cinaedos virgines puerique ingenui . . . ’, Sall. Cat.
25. 2 (on Sempronia) ‘psallere saltare elegantius quam necesse est
probae’, Bompaire (1958), 356 f., J. P. V. D. Balsdon, Roman Women,
1962: 274 f.
matura means ‘of marriageable age’ (about 12 or 13); cf. Virg. Aen. 7. 53
‘iam matura viro, iam plenis nubilis annis’. But in this context one
expects the immaturity of the girl to be emphasized (which makes the
immorality worse); the difficulty is compounded by iam nunc below
(‘already now’), which suggests that an adjective expressing immaturity
has been displaced. RN would therefore obelize matura; Peerlkamp
proposed a matre, Lehrs Romana, L. Müller acerba, Shackleton Bailey
innupta (HSCP 89, 1985: 156), Delz nuptura. NR takes matura virgo as ‘a
girl who has just reached puberty’, and sees this as further emphasized
by iam nunc below.
22–3. et fingitur artibus / iam nunc: for fingitur (‘is moulded’) cf. ars
366 f. ‘voce paterna / fingeris ad rectum’ with Brink, Sen. cont. 1. 2. 5
‘docetur blanditias et in omnem corporis motum confingitur’ (an appar-
ent reminiscence of our passage), Colum. 11. 1. 13, Pers. 5. 36 ff. It was a
commonplace that the young were malleable (epist. 1. 2. 64 ff., 2. 2. 7 f.
‘idoneus arti / cuilibet; argilla quidvis imitaberis uda’). Normally they
are moulded by moral training, but here the artes are the various skills of
seduction; the variant artubus, read by Porph., is unintelligible (see
Bentley).
iam nunc is best attached to fingitur, not to the following clause; this
prevents the middle clause of the three from being too short. For the
punctuation after the opening spondee of an Alcaic enneasyllable cf. 3.
4. 79, 3. 5. 27, 3. 17. 7; note also 1. 34. 7 for the adverb plerumque in the
same position and followed by a pause. RN has considered reading
iamdudum to fill the gap between the adolescent girl and the infant of
v. 24 (as he understands that line).
23–4. et incestos amores / de tenero meditatur ungui: in RN’s view the
phrase de tenero ungui refers to the soft nails of babies, hence ‘a prima
infantia’ (Porph.); cf. paroem. Gr. 2. 407. 51a K ±ÆºH O
ø Id F
Ø Ł, Cic. fam. 1. 6. 2 ‘qui mihi a teneris, ut Graeci dicunt, ungui-
culis es cognitus’ (with Shackleton Bailey), Claudian, VI Cons. Honorii
79 f. ‘dilectaeque urbis tenero conceptus ab ungue / tecum crevit amor’
(with Dewar), Alan Cameron, CQ 15, 1965: 80 ff. (For the temporal use
of de cf. serm. 2. 8. 3 ‘de medio potare die’, K–S 1. 498). In view of the
awkward chronological sequence involved by this interpretation NR
follows those who take de tenero ungui as meaning ‘with every fibre of
her being’ (E. W. Fay, AJP 29, 1908: 201 ff., P. Brind’Amour, Latomus 26,
1967: 467 ff., Williams’s commentary pp. 66 f.); cf. Plaut. Stich. 761 ‘ubi
perpruriscamus usque ex unguiculis’, Plut. de lib. educ. 5. 1 Ł ŒÆd e
c ºª K O
ø.
The latter usage is not normally found with ±ÆºH or teneris; yet an
epigram by Automedon begins T c Ie B A Oæ
æÆ, c
ŒÆŒ
Ø =
ÆØ K ±ÆºH ŒØı O
ø, = ÆNø, P
‹Ø
Æ ÆŁÆÆØ, P ‹Ø ººØ = a ±Æºa ±ÆºH z ŒÆd z
æÆ (anth. Pal. 5. 129. 1–4). Here Gow–Page point out (as against
Cameron) that the poet cannot be referring to infancy (Garland of Philip
2. 186 f.); ‘quivering’ indicates ‘with every fibre of her being’. It would be
odd if nails had a different significance in two poems with a clear
relationship. Automedon could have written a little earlier than Horace
(at anth. Pal. 10. 23. 1 he refers to the rhetor Nicetes, whose floruit
is given by Jerome as 31 bc), and here the epigram seems to precede
the ode: Asia becomes the more literary Ionia, and c ŒÆŒ
Ø
. . . O
ø is distributed among three clauses; cf. P. Colaclides and
M. McDonald, Latomus 33, 1974: 382 ff. To meet this difficulty RN
would like to read K ¼Œæø ŒØı O
ø (‘quivering from the tips
of her finger-nails’, cf. Lucian, Tragodopodagra 17
ØæH I ¼Œæø);
perhaps at an early stage ¼Œæø was corrupted to ±ÆºH under the
influence of ±Æºa ±ÆºH below (it seems pointless to apply the
adjective to nails and hands with different implications), and H gave
K ±ÆºH O
ø its usual meaning of ‘from infancy’. NR is uneasy that
the interpretation of two lines should involve two (or perhaps three)
emendations; he also thinks that, apart from the chronological point
noted above, the hyperbole would be so extreme as to undermine the
seriousness of the passage. In RN’s view hyperbole is often extreme; he
also thinks that the emotional interpretation (‘with every fibre of her
being’) does not suit the more deliberate meditatur, which means ‘goes
over in contemplation’. NR thinks the verb can include an emotional
implication, adducing e.g. Aen. 4. 171 (‘nec iam furtivum Dido meditatur
amorem’).
25–6. mox iuniores quaerit adulteros / inter mariti vina: mox moves on
to the stage when the girl is married. iuniores means ‘younger than the
maritus of the next line’; the elderly husband is a stock figure in comedy,
satire, and elsewhere; cf. 3. 19. 24, Catull. 17. The comparative iunior is
used only by Horace among major poets, and is surprisingly rare even in
prose (Axelson 104); aetate minor was the usual expression. For the
fuddled husband cf. Ov. am. 1. 4. 51 ff. ‘vir bibat usque roga’ with
McKeown. vina is often a poetic plural (Löfstedt 1. 48 ff.), but with
inter it suggests a series of drinks; cf. epist. 1. 7. 28, Plaut. Pseud. 947 ‘inter
pocula’, OLD s.v. inter 7.
26–8. neque eligit / cui donet impermissa raptim / gaudia luminibus

remotis: the woman does not pick and choose her lovers (which would
be bad enough) or aim at haste and concealment; by a standard rhet-
orical procedure the negative sentence leads up to her actual behaviour
(29 ff.), which is even more disgraceful (cf. Cic. Verr. 2. 1. 9 ‘non enim
furem sed ereptorem, non adulterum sed expugnatorem pudicitiae . . . ’).
For ‘impermissa raptim gaudia’ cf. Ov. am. 1. 4. 47 ‘properata voluptas’;
impermissa (the only instance recorded) not only suggests the thrill of
forbidden fruit but makes a contrast with the brazen infidelities that
follow. The lamps were sometimes removed in the symposium (Plut.
quaest. conv. 8. 1), clearly as an encouragement to sex; cf. Plaut. asin.
785 f. (an injunction to a lady) ‘si lucerna exstincta sit, ne quid sui /
membri commoveat quicquam in tenebris’, Prop. 2. 15. 4 ‘quantaque
sublato lumine rixa fuit’, Mart. 12. 43. 10, Lucian, symp. 46 (after the
lamp overturned) ººa Kæ
Ł ŒÆd Øa K fiH ŒfiH, Minuc. Fel. 9.
6–7 (alleged Christian orgies when the dog has upset the lamp); for
similar phrases in other contexts cf. Plut. coniug. praecepta 46, paroem.
Gr. 2. 511. 90 º
ı IæŁ ªıc AÆ ÆP. H is partial to the
ablative absolute even at the end of a sentence (3. 5. 3–4 n.).
29–30. sed iussa coram non sine conscio / surgit marito: iussa is
contrasted with eligit (26); the woman comes when called (cf. serm.
1. 2. 122 ‘neque cunctetur cum est iussa venire’). coram ¼ ‘in the presence
of everyone’ (cf. Ov. am. 3. 14. 16 ‘nec pudeat coram verba modesta
loqui’); this makes a better contrast with luminibus remotis (28) than ‘to
her face’ (i.e. not by letter) or ‘face to face’ (it is not specified whether
the adulterer calls in person). The husband does not even feign ignor-
ance, as he does at Juv. 1. 57 ‘doctus et ad calicem vigilanti stertere naso’
(see Mayor and Courtney); note also the joke ‘non omnibus dormio’
(Otto 121), attributed by Lucilius to one Cipius (1223 M) and by Plutarch
to Gabba, where the adulterer is none other than Maecenas (amat.
16. 759 f. PŒ rŁÆ ‹Ø MÆØŒfi Æ ŒÆŁø;). The conniving hus-
band is ridiculed as a leno (Ov. am. 2. 19. 57 with McKeown, Juv. 1. 55);
such persons were penalized by the later Augustan legislation (dig. 48. 5.
2. 2 on lenocinium).
30. seu vocat institor: an adulterous wife is ‘summoned’ at serm. 2. 3. 238

‘unde uxor media currit de nocte vocata’, meretrices at Prop. 4. 8. 33; cf.
N–H on 2. 11. 21. Lucian Müller thought the institor was one of the
guests, but his social position makes that unlikely (see below); and vocat
is naturally interpreted as ‘summons’ rather than ‘attracts the woman’s
attention’. For similar scenes involving only guests cf. Plaut. asin. 774 ff.,
Ov. am. 1. 4, Suet. Aug. 69. 1 ‘feminam consularem e triclinio viro coram
in cubiculum abductam, rursus in convivium rubentibus auriculis
incomptiore capillo reductam’ (an invective by Antony against Octavian
which may have been recalled by Juvenal in his portrait of the meretrix
Augusta at sat. 6. 116 ff.).
An institor was a salesman (RE 9. 1564 f.); he might call on ladies
with his wares, and is sometimes pictured as slick and dissolute; cf.
Prop. 4. 2. 38 ‘mundus demissis institor in tunicis’, Ov. ars 1. 421 ‘institor
ad dominam veniet discinctus emacem’, rem. am. 306, and more sol-
emnly Sen. fr. 52 Haase ‘institores gemmarum sericarumque vestium si
intromiseris, periculum pudicitiae est’. Moral and social scorn merge; cf.
Eliot, The Waste Land, part III, ‘He, the young man carbuncular, arrives,
/ A small house agent’s clerk, with one bold stare, / One of the low on
whom assurance sits / As a silk hat on a Bradford millionaire’.
31–2. seu navis Hispanae magister, / dedecorum pretiosus emptor: the

Spanish sea-captain (OLD s.v. 3) has no settled home in Italy and so can
be thought of as having ‘a girl in every port’ (Plaut. cist. 157 ff., Sen. contr.
2. 7); for the collocation with a travelling salesman cf. epod. 17. 20 ‘amata
nautis multum et institoribus’. As he is engaged in trade, he is despised
as much as the institor, but is presumably richer (especially if he owns his
ship); so v. 32 refers largely, if not exclusively, to him. In such a context
pretiosus normally means ‘expensive’ (for luxury goods from Spain cf.
Catull. 12. 14 ff., Plin. nat. hist. 19. 10 ); here it is applied uniquely to the
big-spending buyer (cf. Gellius 9. 12. 1 for active and passive senses of
adjectives ending in -osus); the adjective also reflects on the woman, who
is greedy as well as lustful. In erotic poetry too the power of money is
often deplored, and the dives amator is a recurrent figure; cf. epod. 11. 11,
15. 19 f., Prop. 1. 8. 37, Tib. 1. 5. 47.
33–4. non his iuventus orta parentibus / infecit aequor sanguine

Punico: husband and wife are both guilty, and their behaviour affects
the next generation (cf. Juv. 14). Here there is a powerful appeal to
patriotic exempla (iuventus and orta are appropriate to the high style);
for a similar contrast cf. Juv. 2. 153 ff. ‘Curius quid sentit et ambo /
Scipiadae, quid Fabricius manesque Camilli, / quid Cremerae legio et
Cannis consumpta iuventus, / tot bellorum animae, quotiens hinc talis
ad illos / umbra venit?’ H. is referring to the Sicilian sea-battles of the
First Punic War, notably at Mylae (260 bc), which was commemorated
in the Forum by the Columna Duilii (Richardson 97), and off the
Aegatian Islands (241). For the blood-stained waters of inshore battles
cf. 2. 1. 34 f. ‘quod mare Dauniae / non decoloravere caedes?’ with N–H,
Juv. 10. 185 ff.; there is a play of words in Punico, which suggested
‘crimson’; cf. 2. 12. 2 f. ‘Siculum mare / Poeno purpureum sanguine’
with N–H.
35–6. Pyrrhumque et ingentem cecidit / Antiochum Hannibalemque

dirum: Pyrrhus, king of Epirus, invaded Italy in 280 in support of
Tarentum; he won ‘Pyrrhic victories’ at Heraclea and Asculum (north
of Venusia), but was defeated in 275 at Beneventum (see Plutarch’s Life,
P. Garoufalias, Pyrrhus, King of Epirus, 1979, CAH edn. 2: 7. 2. 456 ff.);
his devastation is combined with Hannibal’s at Lucan 1. 30 f., Flor. 2. 6.
11, 2. 9. 22. Antiochus the Great, after restoring the Seleucid empire in
the East, invaded Thrace and finally Greece; he was defeated by the
Romans at Thermopylae in 191 and at the Lydian Magnesia in 189
(E. Will, Histoire politique du monde hellénistique, 1982, especially
2. 177 ff.); ingentem not only recalls his usual epithet of Magnum (Cic.
Sest. 58, Deiot. 36) but underlines how mighty an enemy was over-
thrown. cecidit means that the enemy armies were cut down, not neces-
sarily that the general was killed; cf. Cic. ad Brut. 1. 6. 3 ‘Dolabellam
caesum fugatumque esse’. With Hannibal, who invaded Italy from 218
to 203 and threatened Rome itself, the chronological order is disre-
garded to provide a proper rhetorical climax (cf. epod. 16. 3 ff.). The
emphatic dirum, balancing ingentem, was his conventional epithet ‘quo
nihil inveniri possit significantius’ (Quint. 8. 2. 9, cf. N–H on 2. 12. 2,
Juv. 7. 161), and the variant durum in this passage at least can be ruled out.
37–8. sed rusticorum mascula militum / proles: the words rein-

force each other to point a contrast with urban degeneracy; the
alliteration adds to the effect. For the connection of military virtue
with rustic hardihood cf. 1. 12. 41 ff., Cato, agr. praef. 4, Virg. georg. 2.
167 ff. ‘haec genus acre virum, Marsos pubemque Sabellam / . . . extulit’,
Aen. 9. 607 f. with N. Horsfall, Latomus 30, 1971: 1108 ff., Juv. 8. 245 ff.,
Veg. mil. 1. 3 ‘aptior armis rustica plebs quae sub dio et in labore
nutritur’.
38–9. Sabellis docta ligonibus / versare glaebas: a ligo was an imple-

ment like half a pick-axe (but with a cutting edge instead of a point)
used for turning over the clods and removing weeds; cf. epod. 5. 30 f.
‘ligonibus duris humum / exhauriebat’, epist. 1. 14. 27, Juv. 11. 89,
K. D. White, Agricultural Implements of the Roman World, 1967: figs.
16–20, RE 13. 525; for versare glaebas cf. Horsfall on Virg. Aen. 7. 725.
The Sabelli were not Sabines but Samnites and other Oscan-speakers
of central southern Italy (RE 1A. 2. 1571, Salmon 32 ff.) and included
the original inhabitants of Venusia (serm. 2. 1. 36, epist. 1. 16. 49 of
H. himself, cf. G. Williams ap. Harrison, 1995: 301 f.). For the trans-
ferred epithet applied to implements cf. 1. 31. 9 ‘Calena falce’, Catull. 17.
19 ‘Liguri . . . securi’, Virg. Aen. 7. 665 ‘veruque Sabello’. docta here refers
to the unsophisticated skills of the countryman (cf. Prop. 2. 19. 12
‘docta . . . falce’); contrast doceri (21).
39–41. et severae / matris ad arbitrium recisos / portare fustis: the wife is

responsible for looking after the house, including the fire (cf. epod. 2. 43 f.
‘sacrum vetustis exstruat lignis focum / lassi sub adventum viri’, Cato, agr.
143. 2, Virg. Aen. 8. 410). RN at one time considered reading fasces, which
seems more natural than fustes for a load of wood (cf. Liv. 22. 16. 7
‘fascesque virgarum atque aridi sarmenti praeligantur cornibus boum’,
Apul. met. 7. 17. 4 ‘lignorum vero tanto me premebat pondere ut fascium
molem elephanto non asino paratam putares’, TLL 6. 306. 45 ff.); in
conjunction with ‘severae matris ad arbitrium’ the word might suggest
the fasces borne by a lictor at the behest of a stern magistrate. However,
Apuleius continues ‘cum deberet potius gravantis ruinae fustes demere’
(heavier pieces of wood that make the load unbalanced); note also Varro,
lL 5. 137 (of rushes) ‘utuntur in vinea alligando fasces, incisos fustes,
faculas’, CGL 2. 74. 40 fustes: ºÆ. For the collection of wood cf. 3. 17.
13 f. (windfalls), epist. 1. 14. 41 f., Men. dysc. 31 f. ıºæH Œø Id
= H, Apul. met. 7. 24. 4 ‘lignum quod deveherem recidebat’ (lopping
branches, as here); lignum was an important by-product of many rural
economies (Horden–Purcell 183 f.). portare (as opposed to ferre) suggests
sizeable loads; thus a mosaic from Gaul illustrates a labourer shouldering
a bundle of cut willow; see White (1970), pl. 78.
41–2. sol ubi montium / mutaret umbras: the shadows are changed in
both length (Virg. ecl. 1. 83 ‘maioresque cadunt altis de montibus
umbrae’, ecl. 2. 67 ‘sol crescentes decedens duplicat umbras’) and position
(epist. 1. 16. 6 f., Virg. Aen. 1. 607 f. ‘dum montibus umbrae / lustrabunt
convexa’). The change would be most conspicuous in the evening,
particularly in a mountain valley, and this time is made certain by the
cessation of ploughing; cf. epod. 2. 63 f., Virg. ecl. 2. 66, Ov. fast. 5. 497,
and 42–3 n. below.
The subjunctives mutaret and demeret are thought by some to denote
repeated action (so Heinze and NR); this usage is not found elsewhere
in Horace, but is common in Livy (H–Sz 624). Others (e.g. Page and
RN) see the subjunctives as reporting in poetic terms the mother’s
instructions; E. A. Sonnenschein, who perhaps originated this view,
analyses the tense as ‘past prospective’, ‘when the sun should shift the
shadows’ (CR 7, 1893: 7 ff.).
42–3. et iuga demeret / bobus fatigatis: for the unyoking of oxen as a

sign of evening cf. Hom. Il. 16. 779, Od. 9. 58 q MºØ
ıºı , Ar. aves 1500, Ap. Rhod. 3. 1340 ff. with Hunter, Arat.
phaen. 825, Cic. Att. 15. 27. 3 ‘ÆPfi Ð ıºØ cenantibus nobis’. For the
yoking of oxen as a sign of morning add Hes. op. 581, Call. aet. fr. 21. 3
æÆ IØıÆ º e ªæ T Ø.
43–4. amicum / tempus agens abeunte curru: cf. Sapph. 104 ! E æ
Æ æø ZÆ ÆºØ KŒÆ Æhø , Stat. silv. 5. 1. 124 ff. English
makes the distinction between ‘welcome’ as here (cf. Cic. Att. 12. 15 ‘nihil
est mihi amicius solitudine’, TLL. 1. 1904. 45 ff.) and ‘friendly’ where
there is some idea of personification (Virg. Aen. 2. 255 ‘per amica silentia
lunae’, Ov. her. 19. 33 ‘noctis amicior hora’); sometimes amicum is
compared with the use of Pæ for night, but that may be an old
euphemism for a time of danger. For agens cf. Hom. Il. 8. 485 f. K
4ŒÆfiH ºÆæe MºØ, = (ºŒ ŒÆ ºÆØÆ Kd øæ
¼æıæÆ (i.e. like a curtain), Theoc. 25. 85 f.; it is a paradox that the sun
brings happiness by going away.
45. damnosa quid non imminuit dies?: the question, being rhetorical, is
not answered, but is elaborated in what follows; for the thought cf.
Soph. Ai. 714 Ł › ªÆ
æ ÆæÆØ, Lucr. 2. 1173 f. ‘omnia
paulatim tabescere et ire / ad scopulum’, Ov. met. 15. 234 ff., Prud. contra
Symm. 2. 658 f. ‘nam cum mortalia cuncta vetustas / imminuat’;
H transfers the commonplace from physical destruction to moral de-
generation. imminuit is present tense; the generalization is illustrated
from the past, the present, and the future. dies means ‘the lapse of time’
(cf. epist. 2. 1. 34, ars 293, OLD 10); for the metrical convenience of the
feminine gender cf. E. Fraenkel, Glotta 8, 1917: 60 ff. ¼ Kl. Beitr. 1. 63 ff.
46–8. aetas parentum peior avis tulit / nos nequiores, mox daturos /
progeniem vitiosiorem: for the theme of degeneration cf. epod. 16. 64 f.,
Hom. Od. 2. 276 f. ÆFæØ ªæ Ø ÆE ›EØ Ææd ºÆØ, = ƒ
º ŒÆŒı , ÆFæØ Ææe Iæı , Hes. op. 109 ff., and
especially Arat. phaen. 123 f. (H’s immediate model) ¥
æØØ
Ææ ªc Kº =
Øææ, E b ŒÆŒæÆ Ł,
Bömer on Ov. met. 1. 89 ff., Gatz 18 ff.; but in our passage there is no
mention of the Golden Age, and no assertion about human history as a
whole; the deterioration has taken place only since the second century
7 . QV I D F LE S, A S T ER I E ? 113
(36). Though H mentions four generations rather than Aratus’ three,
he includes them in one sentence with remarkable trenchancy and
compression (cf. 4. 7. 9 ff.): peior avis is a compendious comparison
(cf. 3. 1. 42 n.), equivalent to ‘peior aetate avorum’; nos nequiores ¼ ‘nos
qui nequiores sumus’; the future participle daturos (where Cicero would
have said ‘qui daturi sumus’) exemplifies an abbreviated construction
that would develop in the imperial period (H–Sz 390); the last line is
reduced to two words as in 3. 1. 48 ‘divitias operosiores’. The phrase
daturos progeniem seems to reflect an old formula; cf. Catull. 61. 67 f.
‘nulla quit sine te domus / liberos dare’ with Fordyce’s note. Lyne argues
that daturos etc. permits an inexplicit qualification, ‘unless we reform in
the moral-religious way prescribed’ (1995: 174 f.); yet it is hard to see this
ray of hope in view of the bleak generalization in v. 45 and the clear
allusion to the pessimistic passage of Aratus.
7 . QV I D F LES, AS T ER I E ?
[A. Bradshaw, Hermes 106, 1978: 156 ff.; F. Cairns ap. Harrison (1995), 65 ff.; Davis 43 ff.;
S. J. Harrison, CQ 38, 1988: 186 ff.; H. Jacobson, Mnem. 48, 1995: 85; Lowrie 266 ff.; Lyne
(1995), 175 ff.; F.-H. Mutschler, SO 53, 1978: 111 ff.; W. M. Owens, CW 85, 1991: 161 ff.;
Pasquali 463 ff.]
1–8. Why, Asterie, are you weeping for Gyges, who will be restored to you
in the spring with a rich cargo? While he is storm-bound in Epirus he sleeps
alone, shedding tears himself. 9–22. And yet a go-between reports that his
hostess is burning with passion for him; he reminds him of other wives who
have threatened unresponsive young men, all in vain, for Gyges is deaf to
entreaties. 22–32. You in turn must not become too fond of Enipeus, in spite
of his athletic prowess in the Campus Martius. Bolt your door at nightfall,
and ignore his serenades.
In the first two stanzas Asterie is counselled not to weep for Gyges
(for exhortation and advice in Horace’s amatory odes cf. 1. 8, 1. 13, 1. 33,
2. 4, 2. 5). She seems to fear that he has been lost at sea or become
attached to another woman, but the poet reassures her on both counts:
storms have kept him on the wrong side of the Adriatic, and in the
meantime his fidelity is unwavering. It is sometimes assumed that the
man is Asterie’s husband, but there is nothing in the text to justify so
particular an interpretation. It suits the conventions of erotic literature
better if he is regarded as a lover (Syndikus), even if Propertius could
extend those conventions to include a married couple (3. 12).
The central section of the poem (9–22) describes the temptations that
Gyges is supposed to have been resisting. His hostess Chloe has fallen
in love with him, and through a go-between has menacingly reminded
him of certain exempla from myth: Bellerophon and Peleus were nearly
done to death when they refused a lady in similar circumstances. Within
the context of these threats, maturare necem (16) and paene datum Tartaro
(17) stress the extreme danger that those two heroes were in. Gyges is
not meant to reflect that in the end they escaped; if he did, he could not
expect to be rescued by a magic horse or a kindly centaur. In the wider
context of Horace’s address to Asterie, however, the girl is assured that
somehow Gyges will return safe and sound; so we are free to imagine
that she will be gratified to hear him compared to Bellerophon and
Peleus who were models of chastity. Yet the fact remains that the poet’s
melodramatic story about the go-between and his exempla is highly
implausible (even if we allow for authorial omniscience); perhaps it is
best taken as a light-hearted fiction, designed to give point to the poet’s
sly warning at the end—a warning that seems to be amply justified.
A pattern of correspondences can be observed (so Lowrie 268); for
just as the go-between is said to tell stories to seduce Gyges, so the poet
tells his story to deter Asterie; just as Gyges is said to be deaf to the
blandishments of Chloe, so Asterie is urged to be deaf to the serenades
of Enipeus (another name from mythology, but one that suggests
seduction rather than chastity, cf. 22–4 n.). Owens (op. cit.) thinks
that the poet is really encouraging Enipeus, but nothing in the text
warrants such a cynical interpretation (Cairns, op. cit. 67).
The ode has close affinities with elegy (Syndikus 2. 94). The separ-
ation of lovers is a common motif (cf. Tib. 1. 2. 65 ff. with Murgatroyd’s
parallels, and later Prop. 3. 12); the basic situation goes back to Homer,
with the triangle of Odysseus, Calypso, and Penelope (Harrison, op.
cit.), but it must have been repeated often since. The individual myths
have antecedents in Greek tragedy, but the accumulation of such exem-
pla occurs in Hellenistic poetry (e.g. Theoc. 3. 40 ff., Pasquali 464 f.),
and is typical of Propertius (1. 2. 15 ff., 1. 15. 9 ff. etc.). Just as in elegy, we
hear of tears and sighs, constancy and temptation, with a plaintive
serenade at the end; and there are many similarities in the vocabulary
(for a full list see Cairns, op. cit. 69 f.). Yet the over-all tone of the ode is
very different, not at all sentimental, but detached and amused in
Horace’s manner.
The underlying situation is Roman, in spite of the Greek proper
names and Greek exempla: a trader sails to Bithynia to make his fortune
(3 n.), and a young horseman flaunts his prowess in the Campus Martius
(25–6 n.). But Gyges’ fidelity appears, as suggested above, to be a
romantic embellishment; for a Roman merchant, whether married or
single, would not be required to keep chaste when wintering in a distant
port (for the ‘double standard’ see Treggiari 299 ff.), even if he drew the
line at a liaison with his host’s wife. Scholars are concerned about
the relation of the ode to its predecessor with its vivid depiction of a
ship-master’s adultery (3. 6. 29 ff.); some think it reinforces the moral
(Santirocco 125), others that it undermines it (Lyne, 1995: 178). But
socially this ode is in a different dimension: women who were serenaded
were not matrons of good family (Pasquali 463), and the liaisons of an
Asterie would be of little concern to the Augustan regime. Perhaps it is
enough to say that when Horace arranged his collection he followed the
sequence of Roman Odes (including the powerful 3. 6) with a lighter
piece simply for the sake of variety.
Metre: stanzas of two Asclepiads followed by a Pherecratean and a

Glyconic.
1–2. Quid fles, Asterie, quem tibi candidi / primo restituent vere
Favonii?: the woman’s name, from Iæ, implies a star-like beauty
(cf. 3. 9. 21 ‘sidere pulchrior’, Hom. Il. 6. 401, K. Kost, Musaios: Hero und
Leander, 1971: pp. 164 ff., Cairns, op. cit. 76). ‘Asteris’ is used in a similar
way for the bride of Stella by Statius, silv. 1. 2. 197 f., cf. PMG frag.
adesp. 957 (for ‘Aster’ of good-looking young men cf. Plato, anth. Pal. 7.
670. 1, D. L. Page, Further Greek Epigrams, 1981: p. 161).
candidi suggests brightness, balancing Asterie (cf. Colum. 10. 78 ‘can-
didus . . . zephyrus’, Plaut. merc. 876 ‘hic favonius serenust’, mil. 665);
similarly albus of winds implies an absence of storm-clouds (1. 7. 15, 3. 27.
19, cf. Hes. theog. 379 Iæªc ˘ıæ). candidi also has connotations
of ‘propitious’ (Catull. 8. 3 ‘candidi soles’, OLD 7); in the same way
Favonius, the zephyr, was associated with favere, not because it was a
‘favourable wind’ in the modern sense (here a west wind would come
from the wrong direction), but because it had a benign influence on
vegetation. It began to blow in February (Ov. fast. 2. 148, Plin. nat. hist.
2. 122, Colum. 11. 2. 15), and the navigation season was regarded as
beginning then or in early March (N–H on 1. 4. 1–2); the emphatic
primo means that the man will return as soon as possible.
3. Thyna merce beatum: the Thyni occupied the coastal region of

Bithynia (Plin. nat. hist. 5. 150, RE 6A. 1. 734) on the eastern side of
the Sea of Marmora; by H’s time the name is relatively rare (cf. Catull.
31. 5), as ‘Bithynia’ came to be used for the whole area. The province
produced timber, grain, and marble, and its commercial importance was
increased by its access to the Black Sea; cf. 1. 35. 7, epist. 1. 6. 32 f. ‘cave
ne portus occupet alter, / ne Cibyratica, ne Bithyna negotia perdas’. The
riches are mentioned by way of encouragement, for the rewards (like the
risks) of such a voyage could be great (both are illustrated by the career
of Trimalchio as described in Petr. 76). Naturally the merchant’s con-
ventional avarice (1. 1. 16 with N–H) is not mentioned here; yet Asterie
and her lover belong to the classes who cannot be indifferent to profit
(cf. Mutschler, op. cit. on ‘Kaufmannsliebe’).
4–5. constantis iuvenem fidei / Gygen: the virtue of fides was already
prized in the love-poetry of Catullus (Lyne, 1980: 24); cf. further
Boucher 85 ff., Fedeli on Prop. 1. 4. 16 and 1. 12. 8. For the disyllabic
genitive fidei (the reading of the MSS) cf. TLL 6. 1. 662. 43 ff.,
M. Leumann, MH 2, 1946: 254 (citing Plaut. aul. 617, where fidi is
an iambus, as well as plebeive scitum in inscriptions); most editors
emend to the archaic genitive fide, for which see Bömer on Ov. met. 3.
341 (for the genitive die cf. Gell. 9. 14. 4, Virg. georg. 1. 208 with
Mynors). For the emphatic position of Gygen, which frames the first
sentence with Asterie, cf. 2. 11. 21 f. ‘quis devium scortum eliciet domo /
Lyden?’ with N–H; here as there the apposition precedes the proper
name, and to print a comma after fidei may subdivide the sentence
too much.
The proper name has various resonances that may be significant here.
At 2. 5. 20 Gyges is a beautiful young man, also appearing in Ovid
according to Porph. and perhaps reflecting a Hellenistic source. The
seventh-century King Gyges of Lydia was famous for his wealth
(Archil. 19. 1 h Ø a ˆªø F ºı
æı ºØ, Strab. 14. 5. 28),
as is shown by his dedications at Delphi (Herod. 1. 14); this feature
might be humorously applied to a would-be-rich merchant (beatum in
v. 3) who is trading in Asia Minor. What is more, the legend of the
Lydian Gyges and the wife of King Candaules has some points of
contact with the merchant Gyges and the hospita of 9 ff.; according to
Herodotus (1. 8–12), when Gyges saw the queen naked she threatened
him with death if he did not kill Candaules and marry her (for a tragic
fragment on the same lines see Ox. pap. 23. 2382, D. L. Page, A New
Chapter in the History of Greek Tragedy, 1951: 2 ff.). By another version
Gyges tried to seduce the king’s wife but she told Candaules (Nicolaus
of Damascus, FGrH II. 1, 90 F 47. 7–8, H. Diller in Navicula Chilonien-
sis, Studia Jacoby Oblata, 1956: 66 ff.); Cairns (op. cit. 79 ff.) points out
that the similar stories of Bellerophon (13 ff.) and Peleus (17 ff.) also
appeared in Nicolaus’ Historiae (90 F9 and 55).
5. ille Notis actus ad Oricum: Oricus or Oricum (modern Orika) was a

harbour in Epirus (S. Albania) due east of Brundisium, where the
crossing of the Adriatic is shortest; see RE 18. 1. 1059 ff., N. G. L.
Hammond, Epirus, 1967: 125 ff. with pl. 18a. For the haven afforded by
Oricus cf. Prop. 1. 8. 19 f. ‘ut te felici praevecta Ceraunia remo / accipiat
placidis Oricos aequoribus’. It was a dangerous coast (1. 3. 20 ‘infamis
scopulos Acroceraunia’ with N–H), and the Adriatic could be whipped
up by the south wind (3. 3. 4 n.); this is what has prevented Gyges from
continuing to Italy.
6. post insana Caprae sidera: Capra or Capella is the brightest star in

the constellation of Auriga (Pease on Cic. nat. deor. 2. 110, Le Boeuffle
107 ff.); it had its morning rising (cf. 3. 1. 27 n.) about 28 Sept. (Pliny, nat.
hist. 18. 312). For the elliptical use of post cf. 1. 18. 5 ‘post vina’ with N–H.
H uses the plural sidera because he is including the Haedi (Porph.),
whose rising was very close (3. 1. 28 n., Arat. phaen. 157 f.); it is less likely
that the word is a poetic plural (thus Housman, Classical Papers 3. 1227,
citing Ov. met. 14. 172 ‘sidera solis’). insana indicates equinoctial gales
(Catull. 46. 2 ‘caeli furor aequinoctialis’, Ov. met. 3. 594 ‘sidus pluviale
capellae’).
6–8. frigidas / noctes non sine multis / insomnis lacrimis agit: Porph.
comments on frigidas ‘et propter hiemem . . . et propter solitudinem’; cf.
Catull. 68. 28 f. ‘quod hic quisquis de meliore nota est / frigida deserto
tepefactet membra cubili’, Tib. 1. 8. 39 f. with Smith, Prop. 4. 7. 6 ‘lecti
frigida regna mei’, Ov. am. 3. 5. 42. The insomnia of lovers is another
commonplace (Pease on Virg. Aen. 4. 5, McKeown on Ov. am. 1. 2. 1 ff.);
so too are the tears, which answer those of Asterie in v. 1. non sine (PŒ
¼ı) is found ten times in the Odes; it is not only metrically convenient
but here suits H’s taste for dry understatement; cf. further Wackernagel
2. 297 ff.
9. atqui sollicitae nuntius hospitae: the point of the adversative atqui is

that Gyges does not need to sleep alone. The hospita must be the host’s
wife, not just the landlady of the inn (as some have taken it); otherwise
the following exempla are pointless. sollicita in conjunction with hospita
would normally suggest solicitude for the comfort of a guest (OLD 4a);
here the word is ironical, as it can also refer to a lover’s turmoil (Virg.
ecl. 10. 6, Lygdamus ap. [Tib.] 3. 6. 61 ‘sollicitus repetam tota suspiria
nocte’, ciris 340). The nuntius is a trusty go-between, as the matter was
too delicate for a lady to raise face-to-face. The part was played by the
nurse in Euripides’ Hippolytus and in the ciris (see Lyne on 206–385), by
Anna in the Aeneid, and at a less confidential level by the mistress’s maid
in elegy (McKeown in Ov. am. 1. 11); for male intermediaries cf. Joseph.
aJ 2. 252 (on an Ethiopian princess who has seen Moses from the
battlements and fallen in love with him, cited with other material by
Lightfoot on Parthenius xxi. 2).
10–11. suspirare Chloen et miseram tuis / dicens ignibus uri: the name
Chloe suggests that the wife is young (cf. 1. 23. 1, 3. 26. 12 n.); this makes
her a rival to Asterie (Harrison, op. cit. 189), and perhaps implies an
older and unresponsive husband (cf. 3. 19. 24). For suspirare of a lover’s
sigh cf. 3. 2. 9 n. tuis ignibus means ‘with the same passion as yourself ’
(Porph.) or perhaps rather ‘with the passion that properly belongs to
you’ (cf. Tib. 1. 9. 77 ‘blanditiasne meas aliis tu vendere es ausus?’ with
Murgatroyd, Ov. am. 1. 4. 40, her. 20. 145 ‘iste sinus meus est, mea
turpiter oscula sumis’). Some interpret ‘the object of your passion’ (cf.
epod. 14. 13 ‘non pulchrior ignis’, Virg. ecl. 3. 66), but the plural does not
suit this idea. The emphatic tuis, though it represents the purport of the
go-between’s message, really comes from the poet, who is trying to
produce a reaction from Asterie.
12. temptat mille vafer modis: temptare is often used of attempted

seduction; cf. Ov. ars 1. 273, met. 11. 239, Sen. Phaedr. 891. The dispara-
ging vafer (not used elsewhere in the Odes) suggests the crafty slave of
comedy. For the alliterative mille modis cf. Plaut. trin. 264, Bömer on
Ov. met. 5. 596.
13–14. ut Proetum mulier perfida credulum / falsis impulerit crimini-

bus: Sthenoboea (or Anteia), the wife of Proetus, having failed to
seduce Bellerophon, falsely accused him to her husband, who then
sent him to his father-in-law Iobates with a letter ordering his death;
avoiding the odium of murder, but intending the same effect, Iobates
sent Bellerophon to attack the Chimaera, but thanks to the winged
horse Pegasus he prevailed. See Hom. Il. 6. 150 ff., Apollod. 2. 3 with
Frazer, RE 3. 241 ff., LIMC (on Proitos and Sthenoboia) 7. 1. 525 f., 810 f.;
7. 2. 414 ff., 576; Sophocles wrote an Iobates and Euripides a Sthenoboea.
perfida is pointedly juxtaposed to credulum (cf. 3. 5. 33 ‘perfidis se
credidit’), and the line forms a chiasmus.
14–16. nimis / casto Bellerophontae / maturare necem refert: for

Bellerophon as a model of chastity cf. Ov. trist. 2. 397 f. ‘nam quid de
tetrico referam domitore Chimaerae / quem leto fallax hospita paene
dedit?’ (clearly influenced by the whole context in Horace), Plut. de aud.
poet. 32b–c, Juv. 10. 325. The emphatic nimis is paradoxical (cf. 3. 3. 58
‘nimium pii’): Bellerophon, says the nuntius, was too chaste for the
lady’s liking or his own good. Bellerophontae is a ‘dative of disadvantage’
(with maturare necem); for the declension (from Bellerophontes) cf. 3. 12.
8 n. maturare necem means ‘to hasten the killing’, i.e. ‘to kill without
delay’ (cf. Sall. Cat. 32. 2 ‘mandat . . . insidias consuli maturent’, Liv.
26. 14. 5, Gell. 10. 11); some interpret ‘to make him die before his time’
(cf. Cic. Cluent. 171, Apul. met. 6. 31. 3), but necem would not suit a
natural death. The exemplum contains a sinister though unspoken threat
to Gyges; for the pattern of the story cf. 17 n. (Peleus), Genesis 39
( Joseph and Potiphar’s wife), Stith Thompson, Motif Index of Folk
Literature, 1958: K2111, Lightfoot on Parthenius xiv.
17. narrat paene datum Pelea Tartaro: narrat (cf. 3. 19. 3) with the other
verbs of telling suggests that the nuntius talks at some length; cf. 12
above ‘mille vafer modis’, Ov. am. 1. 8. 20 (of the lena) ‘nec tamen
eloquio lingua nocente caret’. Hippolyte, the wife of Acastus who was
king of the Magnetes in north-east Thessaly, tried to seduce Peleus;
when she failed, she falsely accused him to her husband (Pind. N. 4.
54 ff., 5. 26 ff.). In revenge Acastus abandoned him on Mount Pelion,
where he would have been killed by the Centaurs if Chiron had not
rescued him (Apollod. 3. 13 with Frazer). Both Sophocles and Euripides
wrote tragedies about him, and he is a model of chastity at Ar. nub. 1063
and Plat. rep. 391c. See further RE 19. 1. 277 ff.
datum Tartaro is a grandiose euphemism for ‘killed’; cf. 1. 28.
10 f. ‘Orco demissum’ with N–H, 3. 4. 75, Hom. Il. 1. 3 łı
a ` ! œØ
æ.Æł (where Hades is a god), Lucr. 3. 966, Virg. Aen. 2. 398 with
Austin. dare is also combined with leto (Enn. trag. 283 J, where Jocelyn
identifies the usage as sacral), morti (Plaut. merc. 472), and neci (Virg.
georg. 3. 480).
18. Magnessam Hippolyten dum fugit abstinens: Magnessam distin-

guishes this Hippolyte from the homonymous Amazon (ps.-Acro).
H has cleverly noticed that Magnesia gave its name to magnets (Lucr.
6. 908 f., Plin. nat. hist. 36. 128, RE 14. 1. 474 ff.), and magnets were an
analogy for erotic attraction (Nisbet, PCPS suppl. 15, 1989: 87 ¼ Collected
Papers 261 f.); cf. Ach. Tat. 1. 17. 2 Kæfi A ªF ÆªÆ ºŁ F Øæı
Œi Yfi ŒÆd Łªfi, æe Æc ¥ºŒı, uæ KæøØŒ Ø

ıÆ, anon. anth. Pal. 12. 152, Claud. carm. min. 29. 38 f. ‘ferrumque
maritat / aura tenax’, Nonn. 32. 24, RE 14. 1. 482, W. S. Gilbert, Patience
II. 1 ‘A magnet hung in a hardware shop . . . ’. Magnets could also repel
(D. West cites Lucr. 6. 1042 f.); see further Austin–Bastiniani on Posi-
dippus, epig. 17.
19–20. et peccare docentis / fallax historias movet: peccare is used of

sexual misdemeanours with varying degrees of condemnation; cf. 1. 27.
17 with N–H, 3. 24. 24, Pichon 227. historiae here means ‘stories’; cf. TLL
6. 2840. 5 ff. As Porph. points out, we must understand ‘other stories’ (cf.
1. 10. 5 ‘Iovis et deorum’ with N–H); the case of Phaedra and Hippolytus
was the most obvious, but as it ended in tragedy for both parties it
would hardly have been cited by the nuntius. There is a paradox in
peccare docentis, as such exempla do not usually ‘teach to sin’.
We have preferred movet to monet, though the latter has somewhat
stronger manuscript support. movet means ‘adduces’; cf. Ov. ars 3. 651
‘quid iuvat ambages praeceptaque parva movere’ (where monere of the
MSS does not suit ambages), Pont. 2. 2. 55 f. ‘num . . . excuses . . . factum, /
an nihil expediat tale movere, vide’, Sen. suas. 5. 6 ‘locum movit non
inutiliter’, OLD s.v. 18; here with fallax it makes the go-between sound
like a dishonest advocate. monet is rare with an ‘accusative of thing’,
except in the case of pronouns like hoc or neuter plural adjectives (cf. Ov.
am. 1. 8. 21 f. of the lena ‘illa monebat / talia’); different again are Plaut.
Stich. 58 ‘qui manet ut moneatur semper servos homo officium suum’
(where the person warned is the subject), Cic. cons. fr. 2. 27 ‘terribiles
formae bellum motusque monebat’ (where the accusative describes a
thing warned against). But the construction monet historias in the sense
of ‘tells cautionary tales’ is highly doubtful, and monet is not needed to
assist the paradox ‘peccare docentis’.
21–2. frustra: nam scopulis surdior Icari / voces audit adhuc integer:
for the placing of frustra (also followed by nam) cf. 3. 13. 6, serm. 2. 7. 115,
Catull. 21. 7. The insensibility of rocks is usually a paradigm for a less
laudable intransigence; cf. epod. 17. 54 ‘non saxa nudis surdiora navitis’,
Hom. Il. 16. 34 f., Eur. Med. 28 f. ‰ b æ j ŁÆºØ = Œºø
IŒØ ıŁı ºø, Virg. Aen. 4. 365 f. with Pease, 6. 470 f., Otto
313 f. Icarus or Icaria (still Ikaria) was a rocky island between Samos and
Myconos (Strabo 10. 5. 13, RE 9. 977 f.); the Icarian Sea, which
according to tradition received its name from Icarus (4. 2. 3 f.), was
associated from early times with storms (1. 1. 15, Hom. Il. 2. 145, Ov. fast.
4. 283 with Bömer, her. 18. 50). The specific place-name not only adds
vividness in H’s manner but also suits the context of the eastern Aegean
implied in Gyges’ voyage. integer combines the ideas of ‘heart-whole’
(2. 4. 22) and ‘with virtue unimpaired’ (1. 22. 1). adhuc means ‘still’,
emphasizing the constancy of Gyges; ps.-Acro interprets ‘so far’ (im-
plying doubt about the future) which weakens both the description of
Gyges and H’s appeal to Asterie.
22–4. at tibi / ne vicinus Enipeus / plus iusto placeat cave: at intro-

duces a threat (cf. epod. 3. 19 with Mankin); with the emphatic tibi H
comes to the nub of the poem: he has been commending the fidelity of
Gyges primarily to persuade Asterie to behave likewise. When meetings
between the sexes are restricted, neighbours (not necessarily next door
to each other) have a particular importance (cf. 3. 19. 24, Theoc. 14. 24,
Ov. met. 4. 57 on Pyramus and Thisbe, Pers. 3. 109 f. ‘sive / candida vicini
subrisit molle puella’). This has a special relevance here, when Gyges is
so far away. Enipeus is attested as a personal name at IG 14. 841
(Puteoli), CIL 2. 3583 (Spain), and here recalls the river in Thessaly;
for river-names applied to fictitious characters see note on 3. 12. 6 Hebri
(both Enipeus and Hebrus swim in the Tiber). Homer tells how Tyro
fell in love with Enipeus ‘by far the fairest of rivers’ (Alcaeus 45. 1 says
the same of Hebrus) and was seduced by Poseidon in his guise (Od. 11.
235 ff., Prop. 1. 13. 21 with Fedeli, Apollod. 1. 9. 8 with Frazer, RE 7A. 2.
1869 ff.); Sophocles wrote two plays on the subject (TrGF 648 ff.), in
which Enipeus, in appearance at least, will have figured as a seducer.
Tyro and Enipeus were the grandparents of Acastus (17 n.), Tyro and
Cretheus were the parents of Hippolyte (18); the collocation of these
stories may be due to some mythological handbook (Cairns, op. cit.
91 f.) or to a Hellenistic writer. The poet’s warning is conveyed gently
by the euphemistic ‘plus iusto placeat’; for the erotic sense of placere
cf. OLD 1d.
25–6. quamvis non alius flectere equum sciens / aeque conspicitur

gramine Martio: non alius, like non alter, is commonly used to indicate
an exceptional quality; cf. Hom. Il. 8. 483 Kd P Œæ ¼ºº,
Virg. georg. 4. 372, Aen. 6. 164, 9. 179 f. with Hardie. For flectere equum cf.
Virg. Aen. 9. 606, Ov. Pont. 2. 9. 58, Sen. Phaedr. 811; it took particular
skill to turn a horse in narrow circles (Ov. ars 3. 384 ‘in gyros ire coactus
equus’, Coleman on Stat. silv. 4. 7. 3 f.); see Anderson (1961), 98 ff. The
emphatic aeque is to be taken not with conspicitur (ps.-Acro) but with
sciens, balancing citus aeque at the end of v. 27. conspicitur means ‘attracts
attention’; cf. epist. 1. 15. 46, Plaut. merc. 406 f. ‘quando incedat per vias /
contemplent conspiciant omnes’, Liv. 21. 4. 8, TLL. 4. 497. 1 ff.; quamvis
with the indicative, rare in Republican prose, is well attested in Horace
(1. 28. 11 ff., 3. 10. 13 ff. etc., K–S 2. 443). For exercises in the Campus
Martius cf. 1. 8. 6 f., 3. 12. 8 (there also impressing a young woman),
Tib. 1. 4. 11 (followed as here by a reference to swimming). For gramine
cf. 4. 1. 39 f., ars 162 ‘gaudet equis canibusque et aprico gramine Campi’;
gramine Martio is an unusual phrase for gramine Campi Martii; RN
thinks H may have been influenced by the belief that grass was sacred
to Mars (Onians 142 compares Fest. 97M ¼ 86L (Gradivus), Serv.
Aen. 12. 119).
27–8. nec quisquam citus aeque / Tusco denatat alveo: after working
out in the Campus Martius athletes swam in the Tiber (3. 12. 7 n.); it is
grandiloquently called the Tuscan stream because it flows south from
Etruria (serm. 2. 2. 32 f., Virg. georg. 1. 499, Aen. 8. 473, Ov. ars 3. 386).
denatare is not attested elsewhere till Augustine. It is not clear whether
in this stanza we are meant to think of formal competitions.
29–30. prima nocte domum claude, neque in vias / sub cantu querulae
despice tibiae: cf. Ov. am. 2. 19. 38 ‘incipe iam prima claudere nocte
forem’. prima nocte formally balances primo . . . vere (2). Asterie’s bed-
room is, as usual, upstairs; cf. 1. 25. 2 with N–H’s references, and
add Babrius 116. 5 ff., W. Fauth, Abh. Akad. Mainz 6, 1966: 331 ff.,
A. J. Graham, JHS 118, 1998: 22 ff., Shakespeare, The Merchant of Venice
i i .v. 28 ff. ‘Hear you me Jessica, / Lock up my doors; and when you hear
the drum / And the vile squeaking of the wry-necked fife, / Clamber
not you up to your casements then, / Nor thrust your head into the
public street’ (the parallel with Horace was noted by Malone, but need
not be an imitation). No evidence has so far been produced for vias as a
‘plural for singular’; perhaps it generalizes, suggesting ‘into the public
domain’.
For the paraclausithyron or serenade see below on 3. 10; here, contrary
to the usual pattern, the poet is urging resistance to its blandishments;
cf. Cairns (1972), 209. sub cantu means ‘at the music’ (cf. copa 2 ‘crispum
sub crotalo docta movere latus’, Ov. met. 4. 523 f. ‘Bacchi sub nomine
Iuno / risit’ with Bömer, OLD s.v. sub 14); the variant sub cantum is also
possible (Ov. fast. 3. 342 ‘sub verbum querulas impulit aura fores’, OLD
24). The music of the pipe is plaintive (Lucr. 4. 584 f. ‘dulcisque querelas
/ tibia quas fundit’) as suits the lover (2. 9. 18, 2. 13. 24); for its use in the
serenade cf. Aristaenetus 1. 14. 1 h ÆPºe ÆæÆ r ææØ h
ºæfi Æ Ø KºŒÆØ æÆ Iæªıæı
øæ . Voice and instrument alter-
nate (cf. Virg. ecl. 5. 14 ‘alterna notavi’, 8. 17 ff.).
31–2. et te saepe vocanti / duram difficilis mane: saepe underlines the

persistence of Enipeus. duram represents the viewpoint of the exclusus
amator (cf. Tib. 1. 8. 50 with Murgatroyd, Prop. 1. 16. 30, Ov. ars 2. 527
‘postibus et durae supplex blandire puellae’, OLD 5b, Pichon 136); it
corresponds to the unfeeling rocks of v. 21. difficilis, on the other hand, is
the poet’s own term of praise (3. 10. 11 ‘Penelopen difficilem procis’,
Pichon 130), the opposite of facilis (Mart. 1. 57. 2 ‘nolo nimis facilem
difficilemque nimis’).
8 . M A RT I I S CA ELEB S QV I D
AG A M K A LEN DI S
[Fraenkel 221 ff.; K. Kumaniecki, Eos 50, 1959–60: 147 ff.; Lyne (1995), 109 ff.; Ernst
A. Schmidt, Antike und Abendland 26, 1980: 26 f. ¼ Zeit und Form, 2002: 258 ff.; Williams
(1968), 103 ff.]
1–12. If you are surprised in spite of your learning that I am celebrating on

Matrons’ Day, it is because I made a vow to Bacchus on my escape from the
falling tree; today will uncork a jar laid down in the year of that event. 13–16
Drink, Maecenas, in honour of your friend’s deliverance, and away with all
8 . M A RT I I S CA ELEB S QV I D AG A M K A LEN D I S 123
hubbub. 17–24. You can forget your worries about home affairs; our enemies
have been defeated in Dacia and Parthia, Spain and Scythia. 25–8. Do not
be over-anxious about the people of Rome, and enjoy the moment.
This ode begins with a parody of aetiology. The Homeric hymns and
the tragedians already allude to the origins of ceremonies (cf. DNP 1.
369 ff.); such interests were formalized by the learning of the Hellenistic
age, particularly among local historians (P. M. Fraser, Ptolemaic Alexan-
dria 1, 1972: 511 ff., 775 ff.). Hence the part played by aetiology in the
poetry of the period, most famously in the Aetia and other writings of
Callimachus (Fraser 721 ff.), but also in Theocritus (18. 39 ff.) and Apol-
lonius (Fraser 627 ff.), Eratosthenes (Erigone) and Euphorion (cf. Virg.
ecl. 6. 72). From the time of Cato such studies attracted the Romans, as
suited their respect for the calendar and their love of factual particulars;
they reached their culmination in the works of Varro; cf. Rawson 233 ff.
In the Augustan age Virgil shows a concern for aetiology, most notably
in Aeneid 8 (E. V. George, Aeneid VIII and the Aitia of Callimachus,
Mnem. suppl. 27, 1974); afterwards came the major experiments in the
genre by Propertius in his fourth book (Hubbard 118 ff.), and by Ovid in
his Fasti (A. Barchiesi, The Poet and the Prince, 1997, especially 214 ff.).
In the spirit of these researches Maecenas is supposed to have asked
Horace why he is holding a celebration on the day of the Matronalia
(1 March). For similar curiosity in aetiological poetry cf. Call. aet. frr. 3,
7. 19 ff., 178. 21 f. (on the ceremonies of Icos) ‹Æ KE Ł æÆ
Łıe IŒFÆØ = N
ÆØ Ø º IØæfiø, Virg. Aen. 8. 311 f.
‘singula laetus / exquiritque auditque’, Ov. fast. 3. 169 f. (also on 1 March)
‘cum sis officiis, Gradive, virilibus aptus, / dic mihi matronae cur tua
festa colant’; for Maecenas’ puzzlement (3) cf. Ov. fast. 1. 165 f. ‘post ea
mirabar cur non sine litibus esset / prima dies: ‘‘causam percipe’’ Ianus
ait’, Prop. 4. 2. 1 f. ‘qui mirare meas tot in uno corpore formas, / accipe
Vertumni signa paterna dei’. Horace explains that he had made a vow
after his escape from the falling tree . The ritual is an annual one
(9 ‘anno redeunte’, Call. aet. fr. 178. 3 +ŒÆæı ŒÆd ÆØe ¼ªø
KØ ±ªØ, Virg. Aen. 8 268 f. ‘ex illo celebratus honos laetique
minores / servavere diem’); consule Tullo seems to indicate the date of its
inauguration (12 n.).
The first three stanzas deal with the preliminary religious ceremonies,
which are held in the open air, presumably by daylight. As for Maece-
nas’ ignorance, we need not insist on complete realism; the great man
would not drift in without being asked (especially if the ode is set at the
Sabinum), and if he were asked he would probably know the purpose of
the celebration. At v. 13 (‘sume . . . cyathos’) the party has actually begun
(for such a progression cf. 3. 19. 9); this fourth stanza serves as a bridge
between vv. 1–12, which concentrate on Horace, and 17–28, which
contain an exhortation to his guest (for the importance of Maecenas’
name in this central stanza see L. A. Moritz, CQ 18, 1968: 116 ff.,
M. Marcovich, ICS 5, 1980: 72 ff.).
As usual the ode is designed to suit the temperament of the recipient
(N–H vol. 2, p. 2). Horace mentions Maecenas’ eclectic learning (5),
which made possible the aetiological opening; elsewhere he hints at an
interest in astrology (N–H vol. 2, p. 273), while the elder Pliny in his
table of contents (book 1) cites him as a source on aquatic creatures (for
books 9 and 23) and on precious stones (for book 37). The milieu of the
drinking-party reminds us that Maecenas had himself written a Sympo-
sium (see below on 3. 21), which by the conventions of the genre could
have ranged from the etiquette of such occasions (13 ff.) to erudite
conversations. In 13 f. (‘amici sospitis’) Horace recalls Maecenas’ affec-
tionate concern over the episode of the tree (for all we know, the poet
may actually have been injured), and though he says nothing about his
friend’s simultaneous recovery from illness (2. 17. 17 ff.), ‘consule Tullo’
may be a discreet reminder of that happy event (12 n.). The final stanza
contains tactful allusions to other facets of Maecenas’ career: his unrest-
ing vigilance (Vell. 2. 88. 2 ‘vir ubi res vigiliam exigeret sane insomnis’
with Woodman’s note, eleg. in Maec. 1. 14), his political involvement
from a private station (26 n.), his love of pleasure (27), cf. Mayor on
Juv. 1. 66, and perhaps his tendency to morbid anxiety (28); cf. N–H
vol. 2, pp. 273 f. By recalling the collapse of dangerous enemies Horace
encourages him to indulge his hedonism without compromising his
patriotism; for this combination of public duty and private relaxation
cf. 3. 14 and 3. 29.
The date of the ode can be established from its references to foreign
affairs (18–24). The victory over the Dacian Cotiso is compatible with
M. Crassus’ operations in 29–28 bc (Dio 51. 23. 2, Liv. per. 134, see 18 n.
below), for which he held a triumph in 27; obviously the memory of
these events would remain fresh for several years. The allusion to the
Parthian civil war points to the rebellion of Tiridates in the spring of 26,
which was over by about August 25, as appears from variations in the
coinage (N–H on 2. 2. 17, N. C. Debevoise, A Political History of Parthia,
1938: 137 f., D. Timpe, Würz. Jahrbücher für die Altertumswissenschaft 1,
1975: 155 ff.); this date is supported by Justin’s remark that Tiridates took
refuge with Augustus in Spain (42. 5. 6), i.e. in 25. The defeat of the
Cantabrians (21 f.) refers to Augustus’ own inconclusive campaign in 26
(for this series of operations see especially R. Syme, Roman Papers 2,
1979: 825 ff., also 3. 14 introduction); he then withdrew to Tarraco on the
east coast because of ill health, leaving his legates to continue the war
against the more westerly Astures in 25. The Scythians’ weakness (23 f.)
explains their embassy to Augustus (res gestae 31. 2); Orosius says it
arrived at Tarraco (6. 21. 19), an event which points again to late 26 or 25.
The combination of these last three items suggests that the purported
date of Horace’s symposium is 1 March 25 bc (Kumaniecki, op. cit.),
before the failure of Tiridates’ rebellion, after Augustus’ campaign
against the Cantabrians, and after the events that gave rise to the
Scythian embassy. In the spring or early summer of 24, when Augustus
returned to Rome, the picture had changed, and more emphasis would
have been placed on the Princeps himself.
Metre: Sapphic.
1. Martiis caelebs quid agam Kalendis: on 1 March the Matronalia

were celebrated in honour of Juno Lucina, goddess of childbirth, who
was the mother of Mars; cf. Ov. fast. 3. 167 ff. with Bömer, RE 14.
2306 ff., J. Gagé, Matronalia, 1963: 66 ff., Scullard 86 f. Husbands prayed
for the survival of their wives (ps.-Acro on our passage), and wives for
the glory of their husbands (Ausonius, ecl. 16. 7 f.). Women dressed up
for the occasion ([Tib.] 4. 2. 1), and received presents from their
husbands and others; cf. Plaut. mil. 690 ‘da, mi vir, calendis meam qui
matrem moenerem’, Ov. ars 1. 405 ff., dig. 24. 1. 31. 8 ‘si vir uxori munus
immodicum Kalendis Martiis aut natali die dedisset, donatio est’ (i.e. is
against the law). Pomponius wrote an Atellan farce called Kalendae
Martiae where somebody disguises himself as a woman (CRF 57 f.
‘vocem deducas oportet ut videantur mulieris verba’); this suggests that
men were excluded from some of the ceremonies, as from those of the
Bona Dea (cf. Scullard 247). caelebs is paradoxical when juxtaposed to
Martiis; Juv. 9. 53 speaks of ‘femineis . . . Kalendis’.
2. quid velint flores: ‘what mean the flowers?’; for velle with non-
personal subjects cf. Cic. Verr. 2. 150 ‘quid ergo illae sibi statuae equestres
inauratae volunt?’, OLD s.v. 17. Flowers suited the time of year; cf. Ov.
fast. 3. 253 f. ‘ferte deae flores: gaudet florentibus herbis / haec dea: de
tenero cingite flore caput’.
2–3. et acerra turis / plena miraris: incense was scattered from the
acerra on the fire (carbo in v. 3), but unlike the turibulum or censer
(Hilgers 294 f.) the acerra had no flame itself. plena is given some
emphasis by its position.
miraris means cum admiratione quaeris (Ter. ad. 642, Prop. 1. 5. 21 f.,
TLL 8. 1064. 50 ff.), and here expresses more surprise than the English
‘wonder’. Ancient poems sometimes profess to be sparked off by some
observation from an interlocutor (N–H on 2. 17. 1); it helps the illusion
of a conversation if the verb is taken as interrogative. There is a
progressive build-up of curiosity from quid agam to quid velint to
miraris.
3–4. positusque carbo in / caespite vivo: one side of the textual trad-
ition attaches in to the fourth line, the other omits it; the preposition
seems genuine in view of 4. 6. 11 f. ‘posuitque collum in / pulvere
Teucro’. For carbo or charcoal see R. J. Forbes, Studies in Ancient
Technology 6, 1958: 16 ff.; there is an implicit colour-contrast, underlined
by the alliteration, with the fresh green turf. Temporary arae caespiticiae
were natural in simple country sacrifices (N–H on 1. 19. 13); the clods
had to be vivi (Calp. 5. 25, TLL 3. 113. 69 ff.), i.e. with the grass
still growing.
5. docte sermones utriusque linguae?: Horace teasingly withholds the

name of his addressee; for the independent vocative participle cf. N–H
on 2. 7. 1. sermones are ‘discourses’, ‘treatises’, particularly those in
dialogue form (3. 21. 9 f. ‘Socraticis . . . sermonibus’, Cic. orat. 151, OLD
3b); it may be relevant that Maecenas himself wrote dialogi (Charis. GL
1. 146). Though doctus has the implications of the adjective (‘learned’),
the following accusative shows it is a participle (cf. 3. 9. 10 ‘dulcis docta
modos’, Liv. 25. 37. 3); for various types of ‘retained accusative’ with
passive verbs see S. J. Harrison on Virg. Aen. 10, pp. 290 ff. For
Maecenas’ wide-ranging learning see the introduction above and epist.
1. 19. 1 ‘Maecenas docte’, where Horace is reminding him of a saying
of Cratinus.
utriusque linguae refers to Greek and Latin, the only languages that
mattered to a Roman; for this common locution cf. serm. 1. 10. 23 ‘sermo
lingua concinnus utraque’, Cic. off. 1. 1, Plut. Luc. 1. 3 XŒ ŒÆd ºªØ
ƒŒÆH ŒÆæÆ ªºHÆ, CIL 8. 8500 ‘litterarum studiis utriusque
linguae perfecte eruditus’, M. Dubuisson, L’Antiquité classique 50, 1981:
274 ff. For other allusive uses of uterque cf. N–H on 2. 2. 11, Fordyce on
Catull. 31. 3 ‘uterque Neptunus’, Horsfall on Virg. Aen. 7. 100 f. linguae
(‘tongue’) follows well from sermonibus (‘talks’).
6–7. voveram dulcis epulas et album / Libero caprum: dulcis, ‘deli-

cious’, need not be confined to sweet cakes and fruit; cf. 3. 1. 19, serm. 2.
2. 73 ff. ‘at simul assis / miscueris elixa, simul conchylia turdis, / dulcia se
in bilem vertent’, Enn. var. 44 ‘dulces quoque echini’. album . . . caprum
specifies the content of the feast (hendiadys); white victims suit the
upper gods (ps.-Acro ad loc., cf. Virg. Aen. 4. 61 with Pease, RE suppl. 5.
244 f.). Goats were regularly sacrificed to Bacchus as enemies of the
vine; cf. Varr. rust. 1. 2. 19 ‘sic factum ut Libero patri, repertori vitis, hirci
immolarentur, proinde ut capite darent poenas’, Virg. georg. 2. 380 f.,
Euenus, anth. Pal. 9. 75, Ov. fast. 1. 353 ff. with Bömer, RE suppl. 5. 251.
Elsewhere H attributes his escape to Faunus (2. 17. 28) or the Muses
(3. 4. 27). Liber combines both associations: he was a rustic god, and
like Bacchus (N–H vol. 2, pp. 316 f.) he could be represented as a
patron of poets (epist. 1. 19. 3 ‘ut male sanos / adscripsit Liber Satyris
Faunisque poetas’). Here the very name suits the god who delivered H
from danger, though like ‘Lyaeus’ it usually suggests psychological
liberation.
7–8. prope funeratus / arboris ictu: the verb means ‘to give a funeral to’
(Sen. contr. 8. 4 ‘eos qui vivi uruntur, poena funerat’), not simply ‘to kill’;
there is a piquant contrast between the sombreness of a funeral and the
dies festus that H is celebrating, perhaps also at a verbal level with caespite
vivo. H’s humour is taken further by Petr. 129. 1 ‘funerata est illa pars
corporis qua quondam Achilles eram’. Since ictus normally implies a
deliberate blow, as at 2. 17. 28 f. ‘nisi Faunus ictum / dextra levasset’, the
tree is made to appear malevolent.
9. hic dies anno redeunte festus: festus suggests the Roman calendar;
this holiday is H’s substitute for the Matronalia. The ritual is to recur
annually (cf. 3. 22. 6 ‘per exactos . . . annos’, Virg. ecl. 1. 42 etc. ‘quotan-
nis’). Such anniversaries suit aetiological writing (see introduction
above).
10. corticem adstrictum pice dimovebit: for the use of cork cf. Cato,
agr. 120 ‘in amphoram mustum indito et corticem oppicato’, Plin. nat.
hist. 16. 34 ‘usus eius (suberis) . . . cadorum obturamentis’. The stopper
was smeared with pitch to make it tight (cf. 1. 20. 3 with N–H, Theoc. 7.
147); pice balances corticem as it came from the picea or spruce (Meiggs
422, 467 ff.). dimovebit means ‘will separate from the jar’ (the prefix is the
counterpart of ad-), and in the absence of corkscrews implies some
effort; cf. Archil. 4. 7 W Œ.ºø Æ ¼ºŒ Œø. To have dies as
the agent is not unusual; cf. 2. 17. 8, 3. 14. 13, Headlam on Herodas 5. 22,
Murgatroyd on Tib. 1. 7. 3 f.
11–12. amphorae fumum bibere institutae / consule Tullo: amphorae

seems to be a kind of ‘dativus incommodi’, like that found with verbs of
separating and depriving (cf. 3. 29. 5 ‘eripe te morae’); see K–S 1. 331,
H–Sz 95, Woodcock 61. Wine was stored in an apotheca (bodega) in the
roof, where it was thought to be improved by the smoke (Tib. 2. 1. 27,
Colum. 1. 6. 20, Juv. 5. 34 f. with Courtney); a more sceptical view is
found in Plin. nat. hist. 23. 39, Galen 11. 663K. For the use of bibere (in
the sense of ‘absorb’) cf. Cato, agr. 100 ‘metreta oleum non bibet’, Mart.
13. 32, Sidon. carm. 5. 303 f. institutae means literally ‘set in the first
place’; the formal word suggests the inauguration of a rite, the sort of
thing that interested aetiological writers. Here there is an amusing
paradox in a wine-jar being taught to drink. For the following infinitive
cf. Virg. ecl. 2. 32 f., georg. 1. 147 f. ‘prima Ceres ferro mortalis vertere
terras / instituit’, Juv. 1. 71 f.; there as here the verb implies not only
teaching but the initiation of a continuing activity.
When they were laid down, wines were regularly labelled by the
names of the consuls. H is probably referring to L. Volcacius Tullus,
cos. 33 bc (RE Suppl. 9. 1838 f.) rather than his homonymous father, the
consul of 66; the reader would think first of the more recent and familiar
Tullus (E. Ensor, CR 16, 1902: 210). It suits the parody of aetiology if H
is dating the episode of the tree and the institution of his rite (see N–H
vol. 1, p. 244, E. A. Schmidt, op. cit. 26 f.); ex hypothesi this must have
occurred soon after his acquisition of the Sabinum. At first sight his
description might suggest a more venerable vintage, but 66 bc has no
relevance to the occasion, and there is humour in applying such lan-
guage to a recent tradition. H showed the same care when he laid down
a cheap Sabine wine on Maecenas’ recovery from illness (1. 20. 2 f.
‘Graeca quod ego ipse testa / conditum levi’); as this happened about
the same time as H’s escape (2. 17. 21 f.), he may mean us to regard the
two wines as commemorating both events, perhaps even to be identical
(N–H vol. 1, p. 244).
13–14. sume, Maecenas, cyathos amici / sospitis centum: the cyathus

was the ladle that transferred the wine from the mixing-bowl to the cup
(D–S 1. 1675 f., Hilgers 56 f., 166 f.); it contained a twelfth of a sextarius
or pint. In toasts the number of such measures could be specified (3. 19.
11 n.); for the hyperbolic centum (note the emphatic hyperbaton) cf.
Antiphanes 81. 2 f. (PCG 2, p. 354) ª
Ø . . . ŒıŁı ŁH ŒÆd
ŁÆØH ıæı . There was a hope that the number of cyathi somebody
drank would be equalled by the remaining years of his life (Ov. fast. 3.
531 f. ‘annosque precantur / quot sumant cyathos’ with Bömer); such an
implication coheres with sospitis.
amici sospitis means ‘in honour of your friend’s deliverance’; the
emphasis is on sospitis, as in the ab urbe condita construction, but the
Latin is less abstract than the English translation (cf. N–H on 2. 4. 10).
The toasts belong to the person to whom they are offered (‘this drink
is Heliodora’s’); the genitive reflects a usage of the Greek symposium
(3. 19. 10 n., K–G 1. 376, G. H. Macurdy, AJP 53, 1932: 168 ff.). For poems
celebrating an escape (soteria) cf. N–H vol. 2, pp. 272 f., citing Cairns
(1972), 73 f., 222 f.
14–15. et vigiles lucernas / perfer in lucem: lamps are a symbol of the

symposium (1. 27. 5, Cic.Cael. 67); here they are personified, as some-
times in erotic verse (Meleager, anth. Pal. 5. 197. 3 f., anth. Lat. 711. 4 f.
‘ludite: sed vigiles nolite extinguere lychnos, / omnia nocte vident,
nil cras meminere lucernae’). For the prolongation of the symposium
cf. 3. 21. 24 n., epist. 1. 5. 10 f., Plat. symp. 223c, Cic. Pis. 67; perfero is used
in OLD sense 3. The play on lucem and lucernas is not affected by the
change of quantity.
15–16. procul omnis esto / clamor et ira: it was a convention (not

always observed by H) that the symposium should be orderly; cf.
Hom.Od. 1. 369 f. b f = ø, carm. 1. 27. 5 ff. with N–H.
Sacred days, like the one parodied here, should also be peaceful (Cic.
div. 1. 102 with Pease), and both procul and esto imitate religious lan-
guage (cf. Virg. Aen. 6. 258 f. ‘procul o procul este profani’). Note also
Paul (?), Ephesians. 4: 31 AÆ ØŒæÆ ŒÆd Łıe ŒÆd Oæªc ŒÆd ŒæÆıªc ŒÆd
ºÆÆ IæŁø I H, cited by E. Skard, SO 30, 1953: 101 ff.; that
passage, however, refers not to a religious occasion but to the regular
behaviour expected of the devout. The warning may seem superfluous at
a gentlemanly celebration, but H is also hinting at the conflicts of the
wider world; he thus leads into the military successes of the following
stanzas.
17. mitte civilis super Vrbe curas: it was conventional in sympotic verse
to say ‘do not trouble about serious matters’ (below on 3.19), and in
addressing his important friends H applies the commonplace to na-
tional issues (cf. 1. 26. 3 with N–H, 2. 11. 1 ff., 3. 29. 25 ff.). civilis seems to
mean ‘civilian’ as opposed to ‘military’ (cf. Liv. 9. 3. 5 ‘non militaribus
solum sed civilibus quoque abscesserat muneribus’, Vell. 2. 97. 2); even
though he held no official position (26 n.), Maecenas was concerned
with home affairs (‘ne qua populus laboret’ in v. 25). It is objected that
this interpretation does not lead well to the campaigns of vv. 18–24; but
the argument may be ‘domestic problems have become unimportant in
view of our victories abroad’. If civilis is taken to mean no more than
‘political’ (cf. epist. 1. 1. 16 ‘mersor civilibus undis’), the specific super Vrbe
loses its force.
super has an official tone that suits the political context; cf. 4. 2. 42 f.
‘super impetrato / fortis Augusti reditu’, carm. saec. 18 f. ‘decreta super
iugandis / feminis’, J. N. Adams, CQ 22, 1972: 358 f. The archaic prepos-
ition finally prevailed over de in the spoken language (cf. H–Sz 281);
hence the French sur.
18. occidit Daci Cotisonis agmen: a series of short clauses describing

the state of the nation is also found in epist. 1. 12. 26 ‘Cantaber Agrippae,
Claudi virtute Neronis / Armenius cecidit’ and carm. 4. 5. 17 ff. For
similar brevity in inscriptional elogia cf. ILLRP 309 (¼ ROL 4, p. 2 )
‘Taurasia Cisauna Samnio cepit, / subigit omne Loucanam opsidesque
abdoucit’, E. Fraenkel, Glotta 33, 1954: 158 (citing Virg. Aen. 4. 655 f.); the
res gestae of Augustus are written in a similar style. Military commu-
niqués show the same arrogant brevity, carried to excess in Caesar’s ‘veni
vidi vici’ (Suet. Jul. 37); cf. Cic. Att. 5. 20. 3 (discussed by E. Fraenkel,
Kleine Beiträge 2, 1964: 69 ff.). Similar proclamations appear on board-
games (tabulae lusoriae): ‘Parthi occisi / Britto victus / ludite Romani’;
‘virtus imperi / hostes vincti / ludant Romani’; ‘hostes victos / Italia
gaudet/ ludite Romani’ (cited by Ch. Huelsen, MDAI(R) 19, 1904: 142).
For such board-games see RE 13. 2008 ff., R. G. Austin, G&R 4, 1934:
31 ff., R. K. Gibson on Ov. ars 3. 363 f.
The Dacian Cotiso (RE 4. 1676, 1960 f.) appears first before Actium:
Antony pretended in an invective that Octavian wished to marry his
daughter (Suet. Aug. 63, who calls him Getarum regi). In 29 bc
M. Crassus defeated the Bastarnae near the Ciabrus river (Cibrica) in
north-west Bulgaria (Dio 51. 24, RE 13. 1. 276 f., CAH 10 edn. 2: 550);
after further campaigns in 28 extending towards the Danube delta
(which might have included Dacians) he held a triumph in 27 ‘ex
Thracia et Geteis’, but nothing is said in Dio’s account about Cotiso.
A. Mócsy (Historia 15, 1966: 511 ff.) posits a victory of Crassus over the
Dacians in 30 or early 29 (Dio 51. 22. 6 mentions Dacian prisoners at
Octavian’s triumph in 29); this date is earlier than those referred to in vv.
19–24, and after the victories of 29–8 might seem less relevant. Florus
mentions Cotiso in his account of the Dacian war two chapters after his
account of Crassus: 2. 28 ¼ 4. 12. 18 ‘Daci montibus inhaerent, inde
Cotisonis regis imperio, quotiens concretus gelu Danuvius iunxerat
ripas, decurrere solebant et vicina populari’. Perhaps the campaigns of
Crassus were continued by a successor in 27–26; when Florus speaks in
the following sentences about the campaigns of Lentulus (placed about
10–6 bc by Syme (1986), 289 ff.), we should perhaps allow for the
compression of an epitome.
19–20. Medus infestus sibi luctuosis / dissidet armis: for the Parthian
rebellion see the introduction above; the collective singular is not
uncommon in military contexts, cf. 1. 19. 12 with N–H. The emphatic
sibi should be combined with luctuosis, which needs a supplement; the
Parthians now give grief to themselves rather than to the Romans
(contrast 3. 6. 8 ‘Hesperiae mala luctuosae’, where H goes on to speak
of Parthian victories). In the phrase Medus infestus the adjective needs
no dative; if it is combined with sibi the line breaks at a less natural
place, and it becomes harder to give the pronoun enough emphasis
when taking it with luctuosis. dissidet stands on its own, as in Phaedrus
1. 30. 1 ‘humiles laborant ubi potentes dissident’; if one tries to take it
with sibi, the ablative luctuosis armis gets in the way.
21–2. servit Hispanae vetus hostis orae / Cantaber sera domitus

catena: the date is discussed in the introduction above. Hispanae orae
probably refers to the furthest edge of northern Spain (3. 14. 3–4 n.);
‘the Spanish region’ (OLD 3) would be too vague an expression.
The genitive with hostis is normally objective, but here specifies the
enemy’s habitat.
sera contrasts with vetus; it implies ‘later than desirable’ (cf. Prop. 4. 6.
79 ‘hic referat sero confessum foedere Parthum’, Quint. inst. 6. 3. 4,
Augustine, conf. 10. 27 ‘sero te amavi’). The time taken to complete the
conquest was notorious; Livy 28. 12. 12 says ‘prima Romanis inita pro-
vinciarum quae quidem continentis sunt, postrema omnium nostra
demum aetate ductu auspicioque Augusti Caesaris perdomita est’ (in
fact the achievement of Agrippa in 19 bc, acting on Augustus’ behalf ).
domare is regularly applied to the conquest of Rome’s enemies (1. 12. 54,
4. 8. 18); the word suggests the breaking-in of animals (cf. 2. 6. 2
‘indoctum iuga ferre nostra’); according to Strabo (3. 3. 8), the Cantab-
rians and their neighbours have been subdued by Augustus Caesar, but
where the Roman presence is less felt they are
ÆºæØ ŒÆd
ŁæØøæØ. catena suits unruly slaves, cf. servit (21).
23–4. iam Scythae laxo meditantur arcu / cedere campis: the Scythians’
embassy to Augustus is mentioned above in the introduction; it looks as
if they had suffered some reverse, but it is not known where or how this
happened. For laxo arcu, which implies that they are no longer a threat,
cf. Antip. Thess. anth. Pal. 9. 297. 3 Ø b fiø Œ
ÆºÆÆ Æ (of
the Parthians). The Scythians were expert archers (R. Rolle, The World
of the Scythians, Eng. trans. 1989: figs. 42, 45, 55, 80, 95); perhaps like the
Parthians they were thought to shoot while retreating (N–H on 1. 19.
11). meditantur means ‘are planning’; in other circumstances the verb
would suggest a sinister plot (cf. 2. 11. 1 f. ‘quid bellicosus Cantaber et
Scythes . . . cogitet’, 3. 29. 28 ‘quid Seres et . . . Bactra parent Tanaisque
discors’). The Scythians controlled the vast area of the steppes; now
they are said to be about to retreat from them—a hyperbole that goes
even further than 2. 9. 22 f. ‘intraque praescriptum Gelonos / exiguis
equitare campis’.
25. neglegens ne qua populus laboret: by a characteristic syntactical

innovation neglegens is given a ne-clause of the kind that might follow
noli curare; Porph. compares securus (e.g. 1. 26. 5 f. ‘quid Tiridaten terreat
unice / securus’). Some editors attach ne laboret to cavere (26), but then
neglegens is left isolated, and the symmetry of the clauses is disturbed.
qua means ‘in any respect’ (e.g. the price of corn); H is talking only of
the city (17 n. above).
26. parce privatus nimium cavere: it could seem curious to the Romans
that a private citizen should worry about political problems; cf. Plaut.
Pers. 75 f. ‘sed sumne ego stultus qui rem curo publicam / ubi sint
magistratus quos curare oporteat?’, Cic. rep. 2. 46. This passage tells
against the suggestion that Maecenas held the praefectura Vrbis in 25 bc
(even if this was not technically a magistracy). He may, however, have
exercised a general supervision over home affairs, as he had in 31 (Tac.
ann. 6. 11. 2, Dio 51. 3. 5); this did not prevent H from teasing him about
his avoidance of public offices.
There is a touch of paradox in combining parce with cavere, which
suggests ‘caution’. nimium is also pointed, as parum cavere would nor-
mally be a matter for criticism. After cavere (here used in an absolute
sense) the manuscript tradition reads et, but the stanza articulates more
evenly if it breaks into two pairs of two lines. Bentley transposed the
conjunction to the end of v. 27 (where see note).
27–8. dona praesentis cape laetus horae, / linque severa: for the Epicur-
ean thought cf. epist. 1. 11. 22 f. ‘tu, quamcumque deus tibi fortunaverit
horam, / grata sume manu neu dulcia differ in annum’. The variant rape
would suit contexts emphasizing urgency (1. 11. 8 with N–H, Cic. de orat.
3. 162), but with gifts that are readily available cape or sume is quite
adequate (Bentley); indeed the more violent word is less appropriate as
an antithesis to linque. Bentley’s transposition of et (cf. 26 n.) would make
it clear that 27 combines more closely with 28 than with 26; for its position
at the end of a Sapphic line cf. N–H vol. 1, p. xliv; on the other hand, ae is
elided in the Odes only at 3. 4. 78 (Leo, Plautinische Forschungen, 1912:
357 f.). ac is read by one significant manuscript (actually at the beginning
of 28 ), but H never puts it at the end of a line in either lyrics or
hexameters.
9 . D O N E C G R AT V S ER A M T I B I
[M. Owen Lee, Word, Sound, and Image in the Odes of Horace, 1969: 103 ff.; Lyne (1980),
222 ff.; R. M. Nielsen, Ramus 6, 1977: 132 ff.; Pasquali 408 ff.; M. C. J. Putnam in Ancient
and Modern: Essays in Honor of G. F. Else, 1977: 139 ff. ¼ Essays on Latin Lyric, Elegy, and
Epic, 1982: 107 ff.; R. J. Tarrant ap. Harrison (1995), 46 ff.; D. West, ibid. 100 ff.]
1–4. ‘As long as I found favour with you, I was more blessed than the
King of the Persians. ’ 5–8. ‘As long as Lydia was preferred to Chloe, I was
more famous than the mother of Romulus. ’ 9–12. ‘I am now ruled by Chloe,
and in her place I shall not fear to die. ’ 13–16. ‘I am fired with passion
for Calais and he for me, and in his place I am prepared to die twice over. ’
17–20. ‘What if our former love is restored, if Chloe is thrown over and we
get together again?’ 21–4. ‘Though he is more beautiful than a star and you
more unstable than cork, with you I should happily live and die. ’
9 . D O N E C G R AT V S ER A M T I B I 133
The poem consists of a dialogue between Lydia (6 n.), who has been
supplanted by Chloe (6 n.), and a man who is presumably Horace
himself. The odes regularly profess to be the poet’s own utterance, the
only apparent exceptions being 1. 15 (the narrative on Paris) and 1. 28
(where a dead sailor’s spirit addresses Archytas). The fact that the man
(unlike the woman) is unnamed points in the same direction; for
otherwise there would be a lack of symmetry in a very symmetrical
poem. When Lydia compares her fame to Ilia’s (7–8), it can only be
because she has been celebrated by Horace. When the man’s rival is
described as a puer (16 n.), that coheres with Horace’s comments on his
own middle age (2. 4. 23 f. etc.); and the accusations of fickleness (22)
and hot temper (23, if the text is sound) suit his rueful self-presentation
elsewhere (22–3 n.).
In form the poem is a carmen amoebaeum (an ‘answer-poem’), in which
alternate stanzas are spoken by the poet and Lydia (cf. 3. 28. 9 ff.). The
structure had its origin in folk-song; cf. PMG 852, epist. 2. 1. 146 ‘versibus
alternis opprobria rustica fudit’ with Brink, Liv. 7. 2. 5. It is particularly
associated with bucolic poetry, where rival shepherds cap one another’s
snatches: see, for instance, Theoc. 4, 5. 80 ff. (with Gow, pp. 92 f.), 8
(especially 33–40), 27 (see below), Virg. ecl. 3 and 7. It is also found in
religious ceremonies; for choruses of girls and boys see N–H on 1. 23 (vol.
1, pp. 253 ff.). Similarly in Catullus 62 (the Graecizing epithalamium)
choirs of young men and girls debate the desirability of marriage; there
may well have been a prototype in Sappho (cf. 105 L–P). For the pattern
cf. also the exchange between Lorenzo and Jessica (‘in such a night as
this’) at the beginning of Act 5 of The Merchant of Venice; for verbal
skirmishes between men and women in European literature see further
W. Parks, Verbal Dueling in Heroic Narrative, 1990: 12 f.
Our poem has affinities with no. 27 in the Theocritean corpus
(probably not by Theocritus himself ), where a shepherd and a girl
converse in alternate lines. The idyll like the ode describes a developing
situation, in that case the stages of a seduction; the man counters each of
the girl’s objections with some verbal responsion in the bucolic manner
(1 f., 3 f., 15 f., 20 f.), but the poem is diffuse compared with Horace’s and
lacks his elegant patterning. At one point the two parties boast of their
parentage (42 ff.), just as Lydia claims that her new partner (unlike
Horace) comes of good family (14). Though the idyll does not nor-
mally figure in discussions of our ode, it is more relevant than some
passages that are cited, e.g. the abusive back-chat in Aristophanes, eccles.
892 ff. (cf. Pasquali 415), the interchange between a man and a young
woman in Powell’s Collectanea Alexandrina, p. 184, or the dialogue
between a prostitute and her client in Philodemus, anth. Pal. 5. 46 (¼
20 Sider), where there is no symmetry and no capping. The poetic
conversations between a man and a woman in Sappho 137 (see Page,
Sappho and Alcaeus, 1955: 108) and Archilochus, 196a W. edn. 2 (the
Cologne papyrus) offer nothing directly comparable.
The hendecasyllables of Catullus on Septimius and Acme (45) are
illuminating for purposes of comparison and contrast (Putnam, op. cit.
1982: 117 ff.). That poem, like Horace’s, has a symmetrical structure: the
declaration of love by Septimius (1–7) is matched by that of Acme (10–
16), each being rounded off by a refrain about Cupid’s sneezes; there
follows an eight-line benediction by the poet, consisting of four couplets
arranged ABBA. Here, as in Horace, the attitudes of the man and
woman are differentiated: the forthright speech of Septimius (‘ni te
perdite amo’) with its heroic pledge (‘solus in Libya Indiaque tosta /
caesio veniam obviam leoni’) is capped by Acme’s more emotional
response (‘ut multo mihi maior acriorque / ignis mollibus ardet in
medullis’, cf. 21 ff.). But there is no parallel in the ode for the sensuous
details that appeal to the eye: ‘tenens in gremio’ (2), ‘leviter caput
reflectens’ (10), ‘ebrios ocellos’ (11), ‘purpureo ore saviata’ (12). Catullus’
poem seems to reflect a genuine love-affair; it is perverse to look, as
some have done, for cynical undercurrents.
Horace’s ode is very different: the poet is not an observer but a
participant, there is no scene-setting or visual detail, and we must draw
our conclusions only from what the parties say. The three pairs of stanzas
deal in turn with the past, present, and future; instead of Catullus’ static
tableau with its unchanging emotion, the ode enacts a miniature drama.
In each pair Lydia mimics and at the same time scores off the poet (cf.
Collinge 58 f.); differences between the male and female temperaments
are subtly suggested (Lyne, op.cit. 224 ff., cf. below on 6, 13, 15, 21 f.), and
even at the end, when Horace tentatively suggests an amoris integratio,
Lydia has to reproach him once again with his personal defects before
agreeing. The exchanges are more interesting for their verbal cut and
thrust than for their emotional intensity, and when in the last line Lydia
professes undying loyalty to the poet (24), we recall the similar protest-
ations at the end of stanzas 3 and 4, and suspect that this is just another
stopping-place in love’s merry-go-round (cf. B. Arkins ap. Rudd, 1993:
117). Such a conclusion is not disturbed by the poet’s presence; for Horace
is well able to view his own behaviour with ironical detachment.
The word-order of the poem is simple, and reflects in the main the
language of conversation (Syndikus 2. 106 f.), even if there are some
inflated allusions (8, 14). Apart from the correspondence between one
stanza and another, there is great regularity in each stanza (Lee, loc.
cit.); nouns and adjectives neatly balance, there is much alliteration, and
the phrase usually ends with the line. For the way in which the third
pair of stanzas picks up and develops the first see Tarrant, loc. cit. Those
who believe that the odes could have been set to music, as the Carmen
Saeculare undoubtedly was, might find the case here less implausible
than usual; for this issue see N. A. Bonavia-Hunt, Horace the Minstrel,
1969, M. von Albrecht, Atti del convegno di Venosa, 1993: 75 ff.,
E. Doblhofer, 1992: 79 ff., and on the sceptical side L. E. Rossi, Seminari
romani di cultura greca 1. 1, 1998: 163 ff. (with bibliography).
The ode’s witty treatment and structure have encouraged translations
and imitations, including poems by Jonson, Herrick, and Prior (see
Shorey’s commentary).
Metre: alternating Glyconics and Asclepiads.
1. Donec gratus eram tibi: gratus, ‘ agreeable’, is an understated word

when applied to a lover (cf. Prop. 1. 12. 7 ‘olim gratus eram’, Ov. am. 2.
19. 30 with McKeown); H does not presume to say that Lydia felt
passionately about him. tibi in this position is not enclitic (Nisbet in
Adams–Mayer, 1999: 143 ff.), but has no exceptional stress; the contrast
is between the speaker and quisquam.
2–3. nec quisquam potior bracchia candidae / cervici iuvenis dabat: for
potior of a preferred rival, a stock figure in love-poetry (N–H on 1. 33. 3,
R. D. Brown on Lucr. 4. 1139), cf. epod. 15. 13 ‘non feret assiduas potiori te
dare noctes’ (with Grassmann 154), serm. 2. 5. 76, Tib. 1. 5. 69. quisquam
receives emphasis from the hyperbaton (Fraenkel 152 n. 1), and iuvenis
hints that the rival is a younger man (cf. puero in v. 16 and 1. 5. 1, iunior in
1. 33. 3). candidae, in a prominent position, is a tactful compliment (cf.
Eur. Med. 30 j æłÆÆ ººıŒ æ, Hipp. 70 f., Virg.
georg. 4. 337); ladies of leisure in the Mediterranean world were not
admired for their tan (cf. R. D. Brown on Lucr. 4. 1160). For dare bracchia
of embraces cf. Prop. 4. 3. 12 ‘cum rudis urgenti bracchia victa dedi’; for
other uses of the phrase see N–H on 2. 12. 18. H implies that Lydia is
responsible for the rupture, but by making the rival the subject of the
sentence he lets her off relatively lightly; contrast the direct arsisti (6).
4. Persarum vigui rege beatior: Persian kings were proverbially rich

(a common implication of beatus), cf. 2. 12. 21, Juv. 14. 328 with Courtney,
Otto 273. They were therefore thought ‘blessed’ in the proper sense of
the word (Chariton 3. 1. 8 ÆŒÆæØæ F ªºı ÆØºø ); for the
protests of moralists see 2. 2. 17 ff. with N–H, Cic. Tusc. 5. 35 ‘tu igitur ne
de Persarum quidem rege magno potes dicere beatusne sit?’, Dio Chrys.
6. 7, Athen. 545. For similar comparisons in erotic contexts cf. Prop. 1. 14.
13 ‘tum mihi cessuros spondent mea gaudia reges’, Tib. 1. 8. 34 with K. F.
Smith, Donne, The Sunne Rising 21 ‘She is all States, and all Princes I’.
5–6. ‘donec non alia magis / arsisti: Lydia mimics H by using two
donec-clauses linked by nec. The meaning is ‘as long as you did not fall in
love with another (more than you were in love with me)’; the perfect
tense describes not a state (ardebas) but an action, as at 2. 4. 7, 4. 9. 13,
epod. 14. 9. As arsisti is stronger than gratus eram it makes Horace appear
more to blame for the break. For the ablative alia cf. 1. 4. 19, 2. 4. 8, epod.
14. 9 f. ‘non aliter dicunt arsisse Bathyllo / Anacreonta Teium’, H–Sz 133;
there is no advantage in reading aliam, a poorly attested variant that has
been produced by the following m-.
6. neque erat Lydia post Chloen: Roman meretrices often had Greek
names in life as well as in literature; though not necessarily born in
Greece, they might assume professional pseudonyms. For the exotic
associations of ‘Lydia’ cf. N–H on 1. 8. 1. The use of the proper name
rather than the pronoun ‘I’ not only identifies the speaker but expresses
hurt and indignation; elsewhere it has a pathetic effect (Catull. 8. 12 ‘vale
puella, iam Catullus obdurat’, K. F. Smith on [Tib.] 4. 8. 2, Mankin
on epod. 15. 12). ‘Chloe’ suggests greenness and immaturity (N–H on
1. 23. 1); by naming her rival, Lydia shows her resentment more explicitly
than H, who says simply ‘nec quisquam potior’. For the use of post cf.
4. 13. 21 ‘post Cinaram’, OLD 5b.
7–8. multi Lydia nominis / Romana vigui clarior Ilia’: multi nominis,
‘of much renown’, is a genitive of quality representing the Greek
compound adjective ºıı (Hes. theog. 785 etc.); for the word-
order cf. 1. 36. 13 ‘multi Damalis meri’, 4. 1. 15. Lydia is renowned
because of H’s poetry (cf. Prop. 2. 5. 5 f., Ov. am. 2. 17. 28 ‘et multae
per me nomen habere volunt’ with McKeown); she thus repays the
compliment about her white neck, and by appealing to the poet’s vanity
edges a little towards a reconciliation.
Romana goes one better than Persarum; both words are emphasized
by hyperbaton. Ilia was not strictly a Roman (Rome had not yet been
founded), but as the mother of Romulus by Mars, she could be regarded
as a national heroine (cf. Ov. fast. 3. 9 ‘Romana sacerdos’); moreover, she
was celebrated near the beginning of Ennius’ Annales; cf. Ov. trist.
2. 259 f. ‘sumpserit Annales—nihil est hirsutius illis— / facta sit unde
parens Ilia, nempe leget’. The analogy is extravagant for a hetaera;
some commentators suggest an imitation of Asclepiades, anth. Pal. 9.
63. 1 f. ¸ıc ŒÆd ª Nd ŒÆd hÆ H Ie ˚ æı = æ
ÆH Nd Ø #Æ
(‘because of Antimachus I am honoured more
than all the ladies of Colophon who claim descent from Codrus’).
9–10. me nunc Thressa Chloe regit / dulcis docta modos et citharae

sciens: H now acknowledges that he is under Chloe’s rule (instead of
being like a rex himself ); regit implies autocratic control (cf. 4. 1. 3 f. ‘non
sum qualis eram bonae / sub regno Cinarae’), like the common domina
(N–H on 2. 12. 13). Thressa suggests not just servile origin (thus making
a paradox with regit), but perhaps a wild and passionate temperament;
by emphasizing ‘Thracian’ (note the word-order) the poet may be
contradicting the immaturity implied by ‘Chloe’.
modos is a retained accusative after docta (cf. 3. 6. 21), sciens a participle
that has taken the construction of an adjective (cf. 1. 15. 24 f. ‘sciens /
pugnae’, Hom. Il. 5. 549 ˚æŁø ˇæº
,
KV N ,
Alexander Aetolus 4. 2 Powell ŒØŁæ YÆ ŒÆd ºø, Cic. de orat. 1.
214, K–S 1. 437). H hints that Chloe is more accomplished than Lydia;
for the musical talents of hetaerae see the introduction to 3. 28. In the
same way Terence’s parasite, Gnatho, recommends the praise of a rival
to provoke a loved one’s jealousy: ‘ubi nominabit Phaedriam, tu Pam-
philam / continuo; si quando illa dicet ‘‘Phaedriam / intro mittamus
comissatum,’’ Pamphilam / cantatum provocemus’ (eun. 440 ff.).
11–12. pro qua non metuam mori / si parcent animae fata superstiti:
metuam is future indicative (as parcent shows) and makes a stronger
statement than a subjunctive (cf. 3. 14. 15). animae means ‘my soul ’ in the
extended sense of ‘my dearest’; cf. Plaut. Bacch. 194 ‘animast amica
amanti’, Catull. 45. 13 ‘mea vita, Septimille’, Prop. 1. 2. 1 with Fedeli,
Juv. 6. 195 øc ŒÆd łı
, and for similar expressions N–H on 1. 3. 8.
superstiti is proleptic (‘so that she survives’); the phrase gives a new and
completely different meaning to ‘the survival of one’s soul’. The senti-
ment is found in a variety of situations (A. Setaioli, Prometheus 3, 1977:
75 ff.); see especially Eur. Alc. 280 ff. (in the context of marriage), Plaut.
asin. 609 f. (to a hetaera), CLE 995 ¼ Mus. Lap. 180. 13 ff. (of contuber-
nales) ‘si pensare animas sinerent crudelia fata / et posset redimi morte
aliena salus / quantulacumque meae debentur tempora vitae / pensassem
pro te, cara Homonoea, libens’.
13–14. ‘me torret face mutua / Thurini Calais filius Ornyti: for torrere of
sexual passion (¼ OA) cf. 1. 33. 6 with N–H., 3. 19. 28, 4. 1. 12, Catull.
68. 52. For the metaphorical use of face cf. Prop. 1. 13. 26 ‘tibi non tepidas
subdidit illa faces’ (with Fedeli), Thomas (1999), 60 f.; it is already
implicit in Valerius Aedituus fr. 2. 1 f. ‘quid faculam praefers, Phileros,
qua est nil opus nobis?’ The poet, who has just described himself as the
subject of Chloe, without saying anything about her feelings for him, is
now outdone by Lydia, who claims that she and her partner share an
equal passion; cf. 2. 12. 15 f. ‘bene mutuis / fidum pectus amoribus’ with
N–H, 4. 1. 30, epod. 15. 10, Catull. 45. 20 (Septimius and Acme) ‘mutuis
animis amant amantur’, [Tib.] 4. 5. 13 f. with K. F. Smith.
Lydia’s new lover is given a name with exotic and implausible associ-
ations (see I. Düring, Eranos 50, 1952: 91 ff.); perhaps we are meant to
think she is romancing. ‘Calais’ recalls the homonymous Argonaut
(RE 3. 724), the winged son of Boreas and Oreithyia, whose name is said
to be derived from ŒÆº (schol. Pind. P. 4. 182 ¼ 324 ˚ºÆœ x ŒÆºH
¼Æ); as he was born in Thrace (Ap. Rhod. 1. 213) and by one account
was the beloved of Orpheus (Phanocles, Coll. Alex. p. 106 Powell), Lydia
might be claiming to outdo Chloe’s Thracian connections. It is less
likely to be significant that cal(l)ais means ‘turquoise’ (Plin. nat. hist. 37.
151, Putnam, op. cit. 111 f.). Calais’ father is mentioned as if the family
was important, in order to put Chloe’s in the shade. ‘Ornytus’ seems to
be derived from ZæıØ and to suggest a rushing wind (cf. the horseman
at Virg. Aen. 11. 677 f. ‘procul Ornytus armis / ignotis et equo venator
Iapyge fertur’, where Iapyge recalls the wind of that name); H may have
remembered the Ornytus who is mentioned just four lines earlier than
Calais at Ap. Rhod. 1. 207. Thurii was a Greek city built near the site of
Sybaris on the Tarentine Gulf, but there is no evidence that it had the
same luxurious associations (cf. the glamorous young Sybaris of 1. 8. 2);
more probably the name again suggests ‘rushing’ (ŁæØ ); there was a
cult of Boreas at Thurii (Ael. var. hist. 12. 61), though this might not
have been generally known. Oddly, Augustus originally bore the cogno-
men ‘Thurinus’ (Suet. Aug. 7. 1, 2. 3, RE 6A. 646); Antony called him
this by way of insult, but Octavian refused to regard it as a matter for
reproach.
15–16. pro quo bis patiar mori / si parcent puero fata superstiti’: Lydia’s
hyperbolic bis patiar goes one better than non metuam, but patiar is less
bold and more passive, as suits the feminine stereotype; for multiple
deaths cf. Eur. Orest. 1117 d ŁÆE P
–ÆØ, Plat. apol. 30c ººŒØ
ŁÆØ, Peek, GV 1010. 7 f. ø i = d ŁÆØ ÆPe H
Kb ºØ . puero, which replaces animae, not only expresses
womanly solicitude but rubs in the fact that Calais is younger than
the middle-aged Horace.
17–18. quid si prisca redit Venus / diductosque iugo cogit aeneo?: quid
si is followed by a lively present to introduce a bright idea, the tentative
suggestion that the couple might be reconciled. prisca is more solemn
than pristina, as suits the mention of Venus. For diductos of lovers’
separation cf. Prop. 2. 7. 3 f. ‘diducere amantes / non queat invitos
Iuppiter’, OLD 1c; cogit balances the participle and means ‘brings to-
gether’ (¼ co-agit). The image of the yoke is used for pairs in both
marriage (coniugium) and love-affairs (Fedeli on Prop. 1. 5. 2, Murga-
troyd on Tib. 1. 4. 16); the brazen yokes that do not break belong only to
legend (1. 33. 11 ‘sub iuga aenea’ with N–H, Ap. Rhod. 3. 1284, 1308). The
use of the word here is a persuasive ploy on the part of the ‘Horace
figure’ in the poem; the poet himself no doubt regarded the hyperbole as
unconvincing.
19. si flava excutitur Chloe: excutitur may be translated as ‘shaken off ’
(RN) or ‘thrown out’ (NR); it is a vigorous and even brutal word (cf.
Lyne, 1987: 52, 58 f.). As ‘Chloe’ in Greek describes green vegetation,
there seems to be an oxymoron with flava, which suits ripe corn.
It may also be relevant that Chloe is described as a Thracian (9),
for the Romans admired the fair hair of northern women (Pease on
Virg. Aen. 4. 590, Murgatroyd on Tib. 1. 5. 43 f., McKeown on Ov. am. 1.
14. 45–50).
20. reiectaeque patet ianua Lydiae: if the text is sound, Lydiae must be
dative; as Heinze says, patet is set against reiectae, and the woman who
previously was rejected is offered an open door (thus also Bentley on
3. 15. 8). By this interpretation H could for the first time be acknow-
ledging his foolish error in discarding her, or he could be accepting her
version of what had happened (even if only for tactical reasons). For
passages where the woman visits the man cf. 1. 17. 17 ff., 2. 11. 21 ff. with
N–H, 3. 28, 4. 11, serm. 1. 2. 122, Prop. 4. 8. 29 ff.; for the exclusa amatrix
cf. 3. 15. 8 f., Asclepiades, anth. Pal. 5. 164. 3 f., W. J. Henderson, Acta
Classica 16, 1973: 61 ff.
It is argued on the other side that the door in erotic poetry normally
belongs to the woman (see 1. 25, 3. 10, W. Wimmel, Glotta 40, 1962:
124 ff.). reiectae derives emphasis from its position, and it might seem
tactless to stress that Lydia had once been scorned; cf. Shackleton Bailey
(1982), 94. Peerlkamp took Lydiae as genitive and proposed reiectoque
(accepted by Campbell and Shackleton Bailey); this conjecture not only
spoils the balance of the line but seems to emphasize too sharply that
Lydia was to blame. RN has considered praelataeque (‘and the door of
the preferred Lydia lies open’); the participle would provide a contrast
with excutitur and make a diplomatic concession to Lydia. He would
argue that the mention of the door without a possessive adjective makes
it natural to take Lydiae as genitive. Whatever view one adopts, it is
surely impossible to follow Wimmel, loc. cit., who saw a deliberate
ambiguity in the transmitted reading.
21–2. ‘quamquam sidere pulchrior / ille est: pulchrior at once caps flava
(19) and implies that H is less good-looking than Calais. For comparisons
with stars cf. Hom. Il. 6. 401 (on the infant Astyanax), Virg. Aen. 8. 589 ff.
(on Pallas) with Lyne (1989), 85 ff. Here the image is erotic (cf. 3. 19. 26),
as pulchrior shows. Hyperboles of the type melle dulcior seem to have had a
semi-proverbial origin (cf. Virg. ecl. 7. 37 f., Löfstedt 1, 1942: 307 ff., H–Sz
107 ff.); they appear repeatedly in this poem (4, 8, 22, 23).
22–3. tu levior cortice et improbo / iracundior Hadria: for the asyn-

deton of contrasting elements in a subordinate clause cf. 3. 18. 7 n. Cork
was used for the belts of young swimmers (serm. 1. 4. 120 ‘nabis sine
cortice’) and for the floats of fishing-nets (Pind. P. 2. 80, Ov. trist. 3. 4.
11); the comparison seems to have been proverbial, cf. paroem. Gr. 2. 482
Œı æ ººF, and for a double hyperbole Strabo 1. 2. 30
Œı æ ººF ŒØA . Given the privilege of the last word, Lydia
gets in two further digs before agreeing to peace.
For the storms of the Adriatic see 3. 3. 5 n.; they are compared with
human tantrums also in 1. 33. 15 ‘fretis acrior Hadriae’ (of a bad-
tempered freedwoman). improbus (‘unconscionable’) was used of people
who had no regard for decent limits; for the word’s extension to natural
phenomena cf. Ov. trist. 1. 11. 41 (hiems), Sen. nat. quaest. 4. 4 (imber),
Stat. silv. 1. 5. 46 (sol). iratus is used of the sea (epod. 2. 6 ‘neque horret
iratum mare’, OLD d), and though iracundus is an unusual personifica-
tion we may compare 1. 31. 8 ‘taciturnus amnis’. H admits elsewhere that
he was prone to anger (1. 16. 22 ff., serm. 2. 3. 323, epist. 1. 20. 25 ‘irasci
celerem tamen ut placabilis essem’); NR thinks it natural that in a poem
which refers to the couple’s past experience Lydia should comment on
this fault. RN, however, finds it surprising; he maintains that within
the context of the poem the only reason for H’s bad temper would
be her infidelity, and she would not want to raise that topic. He
therefore has considered inconstantior (PCPS suppl. 15, 1989: 88),
which combines well with levior (Cic. Sull. 10 ‘inconstans ac levis’,
Ov. am. 2. 9. 49 ‘tu levis es multoque tuis ventosior alis’); for the general
idea cf. 1. 5. 5–8; similarly Demosthenes compared the fickleness
of the populace with the changeability of the sea (19. 130); cf. Cic.
Mur. 35, Planc. 15, Liv. 28. 27. 11. levior . . . inconstantior would make a
contrast with Lydia’s double declaration of constancy in the following
lines (amem . . . obeam); iracundior on the other hand introduces an
extraneous notion.
24. tecum vivere amem, tecum obeam libens’: even while blaming
H for his fickleness Lydia has abandoned her own declared loyalty
to Calais (15 f.). Instead of being ready to die for someone else, Lydia
is now prepared to live and die with Horace. For this idea cf. 2. 17.
8 f., Prop. 2. 20. 18 ‘ambos una fides auferet, una dies’, 2. 28. 39 ff., Virg.
Aen. 9. 444 with Hardie, Ov. met. 8. 708 ff. (Philemon and Baucis),
Xen. Ephes. 1. 9. 3 e KæÆc
Ø ¼æÆ Ł y B ŒÆd IŁÆE
æÆØ ªıÆØŒd æØ, Longus 2. 39. 2, F. Olivier, Essais, 1963:
156 ff. (ıÆŁŒø in a variety of texts, including theology and
historiography). obire is a solemn word which significantly is common
on tombstones; cf. 2. 17. 3, 2. 20. 7, Lyne (1989), 108 ff. There is a slight
but perceptible distinction between amem and libens: Lydia would love
to live with the poet and willingly die with him.
10. EX TREMVM TANAIN
[Cairns (1972), s.v. komos; Copley 62 ff.; G. Pascucci, SIFC 54, 1982: 29 ff.; Pasquali 419 ff.;
Pöschl 375 ff.]
1–4. If you were a harsh Scythian, Lyce, you would weep to see me lying on
your doorstep exposed to the northern blasts. 5–12. The trees are groaning in
the wind, and the snow is freezing. Stop being so haughty: don’t strain my
patience too far. Your father was an Etruscan, so you are no Penelope. 13–20.
If nothing else persuades you, I throw myself on your mercy. You are as hard
as oak and as cruel as a snake: I shan’t put up with the hard ground or the
pouring rain for ever.
This poem is a serenade or paraclausithyron (for the name cf. Plut.

amat. 753b), the lament sung by an excluded lover in front of the
woman’s closed door. The type is attested as early as Alcaeus 374 L–P
ÆØ ŒøÆ, ÆØ, ºÆ ºÆØ (which suggests the
repetitions of folk-song). Comedy exploits the humour of the situation,
e.g. Ar. eccl. 952 ff.; cf. Plaut. Curc. 147 ff. ‘pessuli heus pessuli, vos saluto
libens / vos amo, vos volo, vos peto atque obsecro . . . ’. Hellenistic
epigrammatists provide variations on the theme (anth. Pal. 5. 23, 103,
189, 191). We may distinguish the more boisterous ŁıæŒØŒ , the
poem that describes riotous assaults on the premises (cf. 3. 26. 6 ff.,
Headlam on Herodas 2. 34 f.). Roman poets use the paraclausithyron to
express various attitudes, with particular emphasis on the front door
(Copley 28 ff.): Lucretius mocks the lover’s behaviour (4. 1177 ff.), Ca-
tullus uses the door to make scandalous allegations (67), in Propertius
the door complains of its treatment and quotes the lover’s lament (1. 16),
Tibullus presents an elaborate love-poem, allegedly recited on a door-
step (1. 2 with Murgatroyd), Ovid addresses the concierge with witty
epigrams (am. 1. 6), Horace’s ode is a deflating parody (for other
allusions to the form see epod. 11. 19 ff., serm. 2. 3. 259 ff., carm. 1. 25, 3.
7. 29 ff.).
It is clear that the serenade persisted in real life; Lucretius did not
waste his ammunition on purely literary targets, and the last stanza of 3.
7 has no force if it describes something that could not happen. There is
disagreement, however, about how far the Augustan poets engaged in
the kind of behaviour which they describe (contrast Griffin, 1985: 54 f.
with Nisbet, 1995: 214). For example, while RN regards epod. 11. 19 ff. as
fiction, NR thinks it might reflect an experience of the poet’s youth; but
however that may be, it would be fanciful to imagine that the present
ode was ever recited outside a woman’s door. See further J. C. Yardley,
Eranos 76, 1978: 19 ff., Fedeli on Prop. loc. cit., DNP 9. 317 f.; for the
related topic of the sleepless lover add Thomas (1999), 33 ff. (¼ HSCP
83, 1979: 179 ff.).
Horace gives his serenade a Roman setting at the pulchra tecta (6 n.) of
a high-class courtesan; he calls her Lyce, the Greek for a she-wolf (1 n.),
but her father is Etruscan (12) and therefore presumed to be a voluptu-
ary (Maecenas would have enjoyed this). She is contrasted with a
Scythian woman, not only because she lives in the fashionable metro-
polis, but also because she is not married to a brutally repressive
husband. More light is shed on the second point in v. 15, where we
learn that she has a paelex, a word which implies that Lyce has the prior
rights of a wife (15–16 n.). But the existence of a rival suggests that the
husband is not over-possessive (see 15 n.), which in turn lends force to
the poet’s appeal. When the poet professes to court a married woman
he is not trying to undermine Augustan strictures against adultery
as expressed in 3. 6. 25 ff., 3. 24. 20 ff.; our poem shows rather how
little such condemnation mattered except where the highest classes
were concerned. The paraclausithyron conventionally emphasized the
hard-heartedness of the woman (3. 7. 32, Call. anth. Pal. 5. 23. 4 f.,
Prop. 1. 16. 17 ff. etc.); Horace adopts the theme with comic over-
statement (1–4, 17–18). It was a tradition that there should be bad
weather to match the lover’s misery (Asclepiades, anth. Pal. 5. 167. 1 f.
, Y e q ŒÆd , ŒÆd e æ ¼ºª æøØ, = r ŒÆd æ
łı
æ , Kªg b with Gutzwiller (1998), 140 ff., 5. 189. 2, Prop. 1.
16. 24 ‘frigidaque Eoo me dolet aura gelu’ with Fedeli, Ov. am. 2. 19. 22);
Horace gives us all at once wind, snow, ice, and rain, and we are not
expected to ask how far these weather-conditions are consistent with
one another.
Other traditional elements of the form are significantly absent: there
are no maudlin tears and kisses (Lucr. 4. 1177 ff.), no dedication of
the drunken reveller’s garland (Syndikus 2. 112). It was not uncommon
to say to the woman ‘one day you’ll be old and then you’ll be sorry’
(1. 25. 9 ff. with N–H, Asclepiades, anth. Pal. 5. 164. 3 f., Call. ibid. 5. 23.
1 ff.), but even in v. 10 Horace does not go so far as this. The excluded
lover sometimes looks for sympathy by saying he is likely to die (Ar. eccl.
963, Theoc. 3. 53 ŒØFÆØ b , ŒÆd d ºŒØ z ÆØ, Prop.
2. 9. 39 f.), or he even threatens to hang himself ([Theoc.] 23. 21, 36 f.);
but Horace is more robust (Syndikus 2. 114 n. refuting Pöschl 376 f.).
After throwing himself on Lyce’s mercy (16 f.), he executes a typical
volte-face: he will not put up with her cruelty for ever. For a similar
closure cf. his instructions in 3. 14. 23 f. ‘si per invisum mora ianitorem /
fiet, abito’.
Metre: quatrains of three Asclepiads followed by a Glyconic.

1 0 . E X T R EM V M TA NA I N 143
1. Extremum Tanain si biberes, Lyce: i.e. ‘even if you and I were both
in Scythia’. As usual, the ends of the earth are thought to be uncivilized
(cf. 1. 22. 17 ff., 2. 20. 18, Catull. 11. 11 f., Virg. ecl. 8. 44); so extremum
balances saevo and asperas below. The Tanais or Don (3. 4. 36 n., RE 4A.
2. 2162 ff.) was regarded as the boundary of Europe (Manil. 4. 677 with
Housman, Lucan 3. 273, Thomson 254). For its bleak associations cf.
Virg. georg. 4. 517 ‘Tanaimque nivalem’ (reflecting the grief of Orpheus,
though it was nowhere near Thrace), Pope, Dunciad 3. 87 ff. ‘Lo! where
Maeotis sleeps and hardly flows / The freezing Tanais thro’ a waste of
snows . . . ’. The inhabitants of a country are often identified in the poets
by the river they drink (Hom. Il. 2. 825, N–H on 2. 20. 20, Pascucci, op.
cit. 33 ff.); here there is a suggestion that the cold water makes the
women cold (contrast Nilotic Memphis in 3. 26. 10).
Lyce (‘she-wolf ’), like similar Greek names, suits a hetaera (Antip.
Thess. anth. Pal. 11. 327. 1, Lucian, dial. meretr. 12. 1); animal-words
could suggests wildness, rapacity, or as here cruelty, but Lyce does not
have the sordid associations of lupa and lupanar. H uses the name again
in 4. 13 when he says of Lyce ‘felix post Cinaram notaque et artium /
gratarum facies’; the circumstantial detail looks plausible, especially as
Cinara is a recurring character (epist. 1. 7. 28, 1. 14. 33, carm. 4. 1. 4, 4. 13.
21). So H may have had a particular woman in mind, who need not have
been called ‘Lyce’; but even if the ode expresses personal experience of a
lover’s frustrations, it is too frivolous to be a description of a current
situation.
2. saevo nupta viro: the phrase recalls not just the savagery imputed to
the Scythians (Cic. Verr. 5. 150, Prop. 3. 16. 13, Juv. 15. 115 with Mayor)
but also the strict sexual morality imposed by the husbands (cf. 3. 24. 24
‘et peccare nefas, aut pretium est mori’).
2–4. me tamen asperas / porrectum ante fores obicere incolis / plor-

ares Aquilonibus: me in the emphatic position is indignant. porrectus,
‘stretched out’, is used elsewhere of dead soldiers (Prop. 2. 8. 34, Virg.
Aen. 9. 589) or washed-up corpses (epod. 10. 22), but here describes the
undignified posture of the excluded lover; Bentley’s proiectum (‘pros-
trate’) would be acceptable in itself (McKeown on Ov. am. 2. 19. 21 f.),
but not when followed by obicere. asperas refers primarily to the harsh-
ness of Scythian women (cf. Ov. ars 2. 185 ‘quid fuit asperius Nonacrina
Atalanta?’), but Kiessling thinks it may also suggest the roughness of
Scythian carpentry; contrast Catull. 61. 161 ‘rasilemque subi forem’,
where the adjective seems to represent the Greek (see also
K. Latte, Glotta 32, 1953: 35 f.). durus too has a double meaning at epod.
11. 22, Prop. 1. 16. 18 ‘quid mihi tam duris clausa taces foribus?’ with
Fedeli, Tib. 1. 8. 76.
For plorare with the infinitive (‘weep to do something’) cf. Plaut.aul.
308 of an old skinflint ‘aquam hercle plorat, quom lavat, profundere’; for
the rarity of the verb in most Latin poets cf. 3. 27. 38 n., Brink on epist. 2.
1. 9. On the north wind’s place of origin cf. Bolton (1962), 44 f. (Aquilo’s
cave), 22 f. (the Hyperboreans); for incolis as an adjective cf. Ov. fast. 3.
582 ‘incola turba’ with Bömer. The Aquilo was thought of as a swooping
eagle; so obicere might suggest exposure to a bird of prey (OLD 1b).
The idea that it is uncivilized for a wife to reject a lover, which is
more in keeping with the insouciance of Ovid, brings out the contrast
between the ethos of this poem and that of 3. 24.
5–7. audis quo strepitu ianua, quo nemus / inter pulchra satum tecta
remugiat / ventis?: the doors of antiquity rattled and creaked easily, as is
clear from comedy; cf. also Ov. am. 1. 6. 51 ‘impulsa est animoso ianua
vento’ (another paraclausithyron). Lyce’s fine house has trees in its
peristyle (so Porph.); cf. epist. 1. 10. 22 ‘nempe inter varias nutritur
silva columnas’, Alfenus Varus, dig. 8. 5. 17 (referring to a house)
‘(cum) Seius in eo (loco) silvam sevisset’, [Tib.] 3. 3. 15, Sen. contr.
2. 1. 13, Mayor on Juv. 4. 6. Some editors think that a grove is too
luxurious even for an expensive courtesan and must belong to rich
mansions in the same area. But our attention should be concentrated
on Lyce; it is her pulchra tecta that are contrasted with Scythian huts
(and the poet’s discomfort). satum is a much more vivid word than the
variant situm and is supported by ps.-Acro ‘inter tecta satum nemus
viridarium dicit’.
ventis (dat.) means ‘in answer to the wind’; cf. epod. 10. 19 f. ‘Ionius
udo cum remugiens sinus / Noto carinam ruperit’, Lucr. 2. 28, Stat. silv.
5. 1. 153. Some commentators take it as ablative (cf. 3. 29. 57 f. ‘si mugiat
Africis / malus procellis’), but this is incompatible with quo strepitu. In
this position and followed by a pause the word may seem too emphatic,
but it is set against the corresponding nives.
7–8. et positas ut glaciet nives / puro numine Iuppiter?: we must

understand sentis vel sim. from the preceding audis; for the zeugma cf.
1. 14. 3 ff., h. Hom. 3. 264 f., [Aesch.] Prom. 21 f., Virg. Aen. 4. 490. For
lying snow cf. Prop. 1. 8. 7 ‘positas fulcire pruinas’, Ov. fast. 2. 72. This is
the first sure instance of the transitive glacio; in Q. Cic. carm. 1. 12
‘bruma gelu glacians’ (Courtney, FLP p. 179) the word may be intransi-
tive; cf. conglaciaret at Cic. nat. deor. 2. 26. The snow is frozen because
there is no cloud cover; for purus of a clear sky cf. 1. 34. 7 f. ‘per purum
tonantis / egit equos’, OLD 6.
Iuppiter is the old weather-god, who is sometimes identified with the
sky (Pease on Cic. nat. deor. 2. 65); cf. 1. 1. 25 ‘sub Iove frigido’, 1. 22. 20
with N–H, CIL 6. 431 ‘Iovi sereno’. puro is an unusual attribute of
numine, but is perhaps no more surprising than sub Iove puro would be.
Bentley read duro numine, but whereas puro suits the god of the sky, duro
points rather to the ground (cf. 3. 24. 39 ‘durataeque solo nives’); Scaliger
preferred puro lumine, but the scene takes place at night.
9. ingratam Veneri pone superbiam: a refusal of sexual favours could be

represented as arrogance; cf. 4. 10. 2, Prop. 1. 18. 5, 3. 25. 15 ‘exclusa inque
vicem fastus patiare superbos’, Ov. fast. 1. 419 ‘fastus inest pulchris,
sequiturque superbia formam’, Pichon 271. By extension it could be
claimed that such arrogance was an affront to the gods (Tib. 1. 8. 69
‘oderunt, Pholoe, moneo, fastidia divi’) and in particular to Venus; cf. 3.
10. 9, Theoc. 27. 15 F F, A —ÆÆ
º – ŒÆd ª, ŒæÆ,
Agathias, anth. Pal. 5. 280. 7 f. £ Ø ŒÆºe = ººÆ
, K
ŁÆæØ a
ÆæıÆ , Rohde 157 f. The chastity of Hippolytus was an extreme
provocation; cf. Eur. Hipp. 6 (Aphrodite) ººø ‹Ø æFØ N
A ªÆ.
10. ne currente retro funis eat rota: a drum with a rope around it is
wound to lift a heavy weight; for such cranes cf. epist. 2. 2. 73 ‘torquet
nunc lapidem, nunc ingens machina tignum’, Lucr. 4. 905 f. ‘multaque
per trocleas et tympana pondere magno / commovet atque levi sustollit
machina nisu’, RE IA. 1151. In our passage rota is the drum (Vitruv.
10. 2. 5 ‘tympanum amplum quod nonnulli rotam appellant’); for currere
of rapid rotation cf. ars 22; for retro cf. epod. 17. 7 ‘citumque retro solve,
solve turbinem’. Commentators combine retro with eat as well as with
currente, cf. Aristid. Panath. 118 KFŁ X Æ, uæ Œºı
ÞÆª , K
æ Oø, but the word-order is unconvincing; for
absolute eat cf. Lucr. 6. 564 ‘trabes impendent ire paratae’. H’s warning
recalls Lucian, dial. meretr. 3. 3 ‹æÆ c ŒÆa c ÆæØÆ Iææø
ı ıÆØ e ŒÆºfiø Ø, Aristaenetus, 2. 1. 35 (combined as in H
with a warning against pride), paroem.Gr. 2. 298; that is to say, if the
hetaera resists her suitor too strongly, his patience will snap and she will
lose him. In our passage the proverb is modernized, perhaps to suit the
Roman interest in hoisting loads by a windlass: the rope does not break
as in the Greek passages, but goes into reverse along with the drum
when Horace lets go of the handle. The threat anticipates in metaphor-
ical form what the poet says in 19–20, that he will simply give up the
struggle and dump Lyce; there is no suggestion that one day the
positions will be reversed and she will be the suppliant (thus Pasquali
435 f.).
11–12. non te Penelopen difficilem procis / Tyrrhenus genuit parens:

for the pattern cf. Virg. Aen. 3. 42 f. ‘non me tibi Troia / externum tulit’,
Stat. silv. 5. 2. 15 ff.; H is cynically reversing the commonplace that fine
people produce fine children (4. 4. 29 ‘fortes creantur fortibus et bonis’,
Hom. Od. 4. 63 f., 611, Virg. Aen. 1. 606, N–H on 2. 4. 20). Penelope was
the paradigm of wifely fidelity; cf. Theog. 1126 ff., Catull. 61. 221 ff.,
Prop. 3. 12. 38 ‘vincit Penelopes Aelia Galla fidem’, Lucian, dial. meretr.
12. 1, Otto 272, RE 19. 1. 483 f. Here difficilem is the opposite of facilem
(‘compliant’); cf. 3. 7. 32 n. In procis Penelope’s suitors foreshadow Lyce’s
hopeful lovers. As Etruscans were supposed to be licentious (Athen.
12. 517d–518b, citing Timaeus and Theopompus, N–H on 2. 18. 8, RE
6. 754 f.), and Lyce’s father was an Etruscan, she could be represented as
licentious too. Tyrrhenus, the Greek for ‘Etruscan’, has a grandiose effect
(cf. 3. 29. 1); the word balances Penelopen just as parens balances procis.
13–14. o quamvis neque te munera nec preces / nec tinctus viola pallor
amantium: o leads up to the wish at parcas (17); see 3. 24. 25 n., where
again the interjection is a long way from the verb. For lovers’ gifts cf.
Prop. 1. 16. 36, Tib. 1. 8. 29, 2. 4 with Murgatroyd, Ov. ars 2. 261 ff.; for
their supplications cf. Prop. 1. 1. 16 with Fedeli, Ov. am. 1. 6. 61 with
McKeown.
Frustrated lovers are conventionally portrayed as ‘palely loitering’;
such pallor does not come just from a sudden emotion (Sappho 31. 14
L–P) but is a continuing feature; cf. Prop. 1. 1. 22 with Enk, Ov. ars 1. 729
‘palleat omnis amans: hic est color aptus amanti’, Rohde 167 n. The
colour described is most probably a pale yellow; cf. Theoc. 2. 88 with
Gow, Tib. 1. 8. 52 ‘nimius luto corpora tingit amor’; pallor induced by
other causes is described as luteus by Horace (epod. 10. 16), and is
sometimes compared to gold (Catull. 64. 100, 81. 4, Ov. met. 11. 145,
Stat. silv. 4. 7. 15 f., Sil. 1. 233). The lover in Nemesianus (2. 41) combines
pallor with the violet (‘pallidior buxo violaeque simillimus erro’), so we
must think of a violet with a corresponding colour; Plin. nat. hist. 21. 27,
in addition to purpureae and albae, speaks of luteae; that is probably what
Virgil had in mind at ecl. 2. 47 ‘pallentis violas’, where Servius comments
‘amantum tinctas colore’ and quotes our passage.
Some refer viola to pitiable signs of stress additional to the pallor,
‘hectic red splashes on the lover’s cheeks or the dark lines under his eyes’
( J. Gow); for such colour contrasts (usually indicating beauty) cf. Ap.
Rhod. 3. 297 f., Virg. Aen. 12. 67 ff., Stat. Achill. 1. 307 ff., Ach. Tat. 3. 7. 3.
Pasquali (437) suggests that the exclusus amator is blue with the cold, and
indeed tinctus seems better suited to an overall tinge than Gow’s varie-
gated blotches (for which distinctus might be more appropriate); cf.
perhaps Serv. auct. georg. 1. 236 ‘caeruleae frigore scilicet, quia ipse
color convenit frigori’, M. Marcovich, Mnem. suppl. 103, 1988: 93.
15–16. nec vir Pieria paelice saucius / curvat: ‘nor does it sway you that
your husband has been smitten by a Macedonian mistress’; for the
ab urbe condita construction cf. 1. 37. 12 f. ‘sed minuit furorem / vix una
sospes navis ab ignibus’ with N–H, 2. 4. 10, 2. 12. 4. paelex (ÆººÆŒ)
properly describes a rival woman in relation to a wronged wife (3. 27. 66,
epod. 3. 13, Cic. Cluent. 199 ‘filiae paelex’); she is often socially inferior
and sometimes, as here, has the stigma of being foreign. Pieria was the
district of Macedonia north of Mt. Olympus. Perhaps Lyce’s husband
has been ensnared when travelling abroad (like Gyges at 3. 7. 5); that
would explain his absence from the scene, and suit the contrast between
Lyce and Penelope. But as Pieria had no particular associations with
foreign trade, she may simply be a Macedonian woman living in Rome;
one would like to find some special significance in the name, for
instance a suggestion of musical accomplishments (the Pierides are the
Muses at 4. 3. 18, 4. 8. 20), but in the absence of parallels the question
remains open.
For saucius (æøŁ ) cf. Enn. trag. 216J (¼254V) ‘Medea animo aegro
amore saevo saucia’, Lucr. 4. 1048 with R. D. Brown, Virg. Aen. 4. 1 with
Pease, Pichon 259; the strength of the word is designed to weaken Lyce’s
scruples. As the intention of the woman is not emphasized, the instru-
mental ablative without a is quite natural. curvat (¼ ‘bends’, flectit) is
unattested elsewhere in this sense, but is found in a compound form at
Persius 1. 91 ‘qui me volet incurvasse querella’. For the indicative after
quamvis cf. 3. 7. 25–6 n.
16–17. supplicibus tuis / parcas: H has appealed in vain to various emo-

tions that might have moved Lyce: greed (munera), soft-heartedness
(preces), pity (pallor), jealousy (paelice); now he can only throw himself
on her mercy. For supplex in an erotic context cf. Pichon 271. The plural
here has a generalizing effect (‘show mercy to those who beg for mercy’),
as with pallor amantium above; see Löfstedt 1 (1942), 38 f., who quotes
such passages as Cic. Att. 1. 17. 3 ‘vereor ne dum defendam meos [my
brother] non parcam tuis [your sister]’. H has no interest in the welfare
of rival lovers.
17–18. nec rigida mollior aesculo, / nec Mauris animum mitior angu-
ibus: the vocative attributes without a noun give a grandiloquent effect
(cf. epod. 17. 20, 3. 8. 5, N–H on 2. 7. 1); this is deflated in vv. 19–20. The
aesculus, which was sacred to Jupiter (Plin. nat. hist. 12. 3, 16. 11), and is
celebrated for its size in Virg. georg. 2. 15 f. and 291 f., had nobler
associations than the quercus and is identified by some with the Valonia
oak (Mynors on Virg. georg. 2. 14–16; see also Abbe 84 f.). For the
proverbial hardness of robur cf. N–H on 1. 3. 9, Ov. met. 13. 799 (the
Cyclops to Galatea) ‘durior annosa quercu’. rigida, set against mollior,
continues the image of ‘nec . . . curvat’, and nec mollior aesculo is balanced
by nec mitior anguibus (‘vertical correspondence’).
North Africa was notorious for its snakes (serm. 2. 8. 95 ‘peior
serpentibus Afris’, Lucan 9. 619 ff.) and Mauris adds a hint of barbarian
menace (at 1. 22. 2 ‘Mauris iaculis’ suggests poison); the hyperbole shows
that H is not serious. For mitis and immitis cf. 1. 33. 2, Tib. 1. 4. 53, Ov.
met. 13. 804 ‘calcato immitior hydro’, TLL 8. 1154. 50 ff. animum is a
Graecizing accusative of respect with mitior, a construction found else-
where in H only with cetera (4. 2. 60, epist. 1. 10. 3,50); cf. Virg. Aen. 1.
320 ‘nuda genu’ with Austin, K–S 1. 286, H–Sz 37 f., R. G. Mayer in
Adams–Mayer 163 ff. The variant animo (the normal prose construction)
should be rejected in view of the other ablatives.
19–20. non hoc semper erit liminis aut aquae / caelestis patiens latus:
hoc is given emphasis by the long hyperbaton and implies ‘whatever
others may do’. non semper means not just that the poet will soon go
away but that he will never come back (Lambinus); cf. Theoc. 7. 122
ŒØ Ø æıæø Kd æŁæØØ. limen includes the area in front of
the threshold; for its hardness cf. epod. 11. 22, Ov. met. 14. 709 f. ‘posuit-
que in limine duro / molle latus’. A participle like patiens can take a
genitive even in prose when used as an adjective (K–S 1. 450). aqua
caelestis for rain is not a poeticism; cf. Vitruv. 8 praef. 1 ‘caelestium
imbrium’, OLD s.v.caelestis 1a.
1 1 . M ERCV R I , NA M TE D O C I LI S
MAGISTRO
[A. Bradshaw, Hermes 106, 1978: 156 ff.; F. Cairns, G&R 22, 1975: 129 ff.; R. W. Carrubba,
CJ 84, 1988–9: 113 ff.; Lowrie 275 ff.; Pasquali 144 f.]
1–12. Mercury and lyre of tortoise-shell, play a song to win the ear of
Lyde, who now like a skittish filly shies away from a mate. 13–24. You can
make tigers follow you and bring rivers to a halt; Cerberus yielded to your
blandishments, and you brought respite to sinners like the Danaids. 25–32.
Let Lyde note the punishment of those virgins who cruelly killed their
bridegrooms. 33–52. One was magnificently deceitful and urged her husband
to get up, saying ‘Even if my father imprisons or banishes me, go with the
blessing of Venus, and carve the sad story on my tombstone’.
Horace opens by invoking Mercury: just as, at the building of

Thebes, the god taught Amphion to move stones by his music, so
may he help the poet to win the unyielding Lyde. Mercury is coupled
with his lyre of tortoise-shell (3 n.), which takes over from its inventor as
1 1 . M ERC V R I , NA M T E D O C I LI S M AG I S T RO 149
the recipient of a formal hymn (for the religious formulae see N–H vol.
1, p. 127, below on 3. 21). Its attributes are described in a series of
appositions (5 n.), and an aretalogia follows on its power to cajole wild
nature and soothe the torments of the damned; other sacral features are
the conventional potes (13–14 n.) and the anaphora of tu and tibi (15–
16 n.). Yet for all its solemnity, the invocation derives a whimsical note
from the tortoise-shell’s humble origins.
This light-hearted tone is repeated in the description of Lyde, who is
compared to a frolicsome filly in the erotic language of Anacreon (9 n.).
The secularization of hymns for private purposes can already be seen
in Sappho’s appeal to Aphrodite to restore the affections of a beloved
(1. 15 ff.); closer in mood is Anacreon, PMG 357. 9 f. ˚ººfiø IªÆŁe
ª = ıº , e K ª æø- = , t ˜ ı,
ŁÆØ (cf. Cairns,
op. cit. 137). The blandishments of music and poetry in courtship were
also a topic in Augustan elegy; cf. Prop. 1. 7. 5 f., 3. 2. 1 ff. (like Horace,
mentioning Orpheus and Amphion), Tib. 2. 4. 19 ‘ad dominam faciles
aditus per carmina quaero’ (with Murgatroyd), W. Stroh (1971).
In the central lines (25–8) Horace turns to the Danaids, whose crime
takes up the six final stanzas (for this type of conclusion cf. N–H on 1. 7.
21): when coerced by Danaus to marry their cousins, the sons of
Aegyptus, they killed them on the wedding-night; of the fifty sisters
only Hypermestra spared her bridegroom, Lynceus. The tale had been
told in a trilogy by Aeschylus, in which the surviving Supplices must
have been less relevant than the lost Danaids; for theories about the
controversial ending see A. F. Garvie, Aeschylus’ Supplices (1969) and the
commentary by H. F. Johansen and E. W. Whittle (1980), especially vol.
1, pp. 29–55. There was also a legend about unnamed water-carriers in
the underworld, who were compelled to fill a leaky vessel because they
themselves were incomplete (IºE ), i.e. uninitiated; cf. Plato, Gorg.
492d (with Dodds), Paus. 10. 31. 9 (on the painting by Polygnotus at
Delphi), E. Keuls, The Water Carriers in Hades, 1974. In the Hellenistic
age the water-carriers were regularly identified with the Danaids
([Plat.] Axioch. 371e); perhaps they were ‘incomplete’ because their
marriage had not been consummated (for various views cf. E. Rohde,
Psyche, appendix 3, Garvie 234 f., Keuls 53). See further Lucr. 3. 1008 ff.,
Apollod. 2. 1. 4–5, RE 4. 2087 ff., LIMC 3. 1. 337 ff., 3. 2. 250 ff.
The myth as presented by Horace is ingeniously linked to the
opening hymn. The power of the lyre is illustrated by various examples,
culminating in the story of how Orpheus in the underworld brought
respite to sinners like the Danaids. Having reached the notorious man-
haters in this apparently casual way, Horace uses them as a warning to
Lyde (25–6 n.); at the same time he glorifies Hypermestra, who was
merciful to her man, even at the risk of her own life. This turns out to be
the song by which he hoped to win Lyde’s ear (7 f.). In addition to the
contrast between wild and tame (tigers and lionesses as against filly and
calves) there is a persistent tension between hardness and gentleness:
thus lapides (2), obstinatas (7), duro ferro (31 f.), saevis catenis (45) pull
against amica (6), blandienti (15), mulces (24), mollior (43), clemens (46),
Venus (50). This is enough to show that, although the myth is the most
memorable feature of the poem, the opening section is not just an
introduction prefixed to the story and then forgotten (Fraenkel 190;
cf. F. Klingner, JRS 48, 1958: 175), but rather an integral part of the ode,
explaining why the myth is recounted (Syndikus 2. 117 f., Bradshaw, op.
cit. 164). If this explanation is disregarded, the nature of the ode as a
whole is liable to be misconstrued (see the end of the introduction).
One contemporary representation of the story was to be seen in the
portico in front of Apollo’s Palatine temple: see Prop. 2. 31. 1 ff. ‘quaeris
cur veniam tibi tardior? aurea Phoebi / porticus a magno Caesare aperta
fuit. / tota erat in speciem Poenis digesta columnis / inter quas Danai
femina turba senis’, Ov. am. 2. 2. 3 f., ars 1. 73 f. ‘quaque parare necem
miseris patruelibus ausae / Belides et stricto stat ferus ense pater’, trist. 3.
1. 61 f. ‘signa peregrinis ubi sunt alterna columnis / Belides et stricto
barbarus ense pater’, schol. Pers. 2. 56 ‘Acron tradit quod in porticu
quadam Apollinis Palatini fuerunt L Danaidum effigies, et contra eas
sub divo totidem equestres filiorum Aegypti’ (the presence of the sons is
doubted by many).Three herms of black marble depicting women are
now believed to have come from this site: see L. Balensiefen, MDAI(R)
102, 1995: 189 ff. (with illustrations and bibliography). The monument is
likely to have influenced Horace’s ode, but throws little light on its date.
Propertius 2. 31 implies that the opening of the portico was distinct from
the dedication of the temple in October 28, and the latter part of his
second book is assigned to 26 or 25; cf. Hubbard (1974), 43 f.; but even if
the formal opening was delayed till Augustus’ return from Spain in 24,
the construction could have been completed earlier.
The symbolism is not obvious. One looks for some connection with
Apollo, as in the case of the Gauls’ defeat at Delphi and the destruction
of Niobe and her children—motifs depicted on the temple doors (Prop.
2. 31. 13 f.)—or with Augustus’ victories; for if he did not actually
conceive the design, Augustus certainly sanctioned it (res gest. 19. 1).
According to one view, Danaus is present as an ancestor of Antony’s,
and the murderous Danaids, who came from Egypt, represent the
much-hated forces of Cleopatra as well as symbolizing a hostile attitude
towards marriage; see B. Kellum in The Age of Augustus (ed. R. Winkes),
1985: 173 ff., and especially S. Harrison in Vergil’s Aeneid: Augustan Epic
and Political Context (ed. H.-P. Stahl), 1998: 223 ff. Others think that
they stand for the triumph of civilization in the same conflict (E.Simon,
Augustus. Kunst und Leben in Rom um die Zeitwende, 1986: 19 ff.,
E. Lefèvre, Das Bild-Programm des Apollo-Tempels auf den Palatin,
1989: 14 ff.); but this hardly suits the hostile reaction of the Augustan
poets (cf. Virg. Aen. 10. 497. ‘impressumque nefas’). P. Zanker sees a
suggestion of atonement and reconciliation after the fratricidal war
(Città e architettura nella Roma imperiale, Analecta Romana Instituti
Danici, suppl. 10, 1983: 27 ff.); but that is not in line with Ovid’s descrip-
tion of Danaus or the triumphalism of the temple as a whole.
K. Galinsky, while in general accepting the more favourable view of
the Danaids, speaks of ‘multilayered inspirations’ and ‘alternative inter-
pretations’ (1996: 222); yet it is easier to see or imagine such ambiguities
in poetry than in a large and prominent work of public sculpture. If the
first hypothesis is preferred (according to Apollodorus 2. 1. 5 and
Hyginus, fab. 170. 3 one of the Danaids was actually called Cleopatra),
a problem still remains: in such monuments victorious killers are usually
celebrated rather than reviled.
In style the poem has some affinities with Greek lyric (Cairns,
op. cit.), for instance in the concluding speech (37 ff.); but it lacks the
amplitude of Pindar and seems closer to the narratives of Sappho and
Ibycus (N–H vol. 1, p. 189, Kroll 239 f.). Unlike Ovid, her. 14, it does not
tell a continuous story, but alludes to what is assumed to be known. The
search for a new way of handling a familiar theme, and the light-hearted
attitude to the myth, are more reminiscent of Callimachus than of early
Greek poetry. The didactic impact of the story would be blunted by the
psychological conflicts that we find in Ovid’s imitation (for which see
H. Jacobson, Ovid’s Heroides, 1974: 124 ff.). The combination of brevity
and sensationalism is paralleled in the Europa-ode (3. 27); perhaps the
influence of declamation, as seen in the elder Seneca, is already at work.
This brings us to the poet’s intention. It has been pointed out that
here, in contrast to 1. 23, Horace gives no overt sign of any personal
interest in the girl; hence Bradshaw suggests that Lyde is a Roman
maiden, and Horace is giving her some kindly avuncular advice about
fidelity in marriage (op. cit. 156, 164, 172). Yet the name Lyde does not
indicate the context of Roman marriage any more than it does in 2. 11. 22
or 3. 28. 3. So, whatever the significance of the monument mentioned
above, here a less serious interpretation is called for, and this is sup-
ported at several points by the poet’s tone. In the first section the
magical achievement of Amphion in founding Thebes, and that of
Orpheus in charming the hound of hell, are put on the same level as
the poet’s attempt to seduce an indifferent young girl. When in the
central stanza (25 ff.) Horace refers to the punishment of the wicked
Danaids (‘scelus atque notas / virginum poenas’) he is hinting that their
virginity was part of their offence. In the second part there is an
amusing incongruity in recommending a heroine with a noble literary
pedigree in Aeschylus and Pindar as a role-model to a girl who, like her
predecessor in Anacreon, is behaving like a skittish filly. And when the
poet emphasizes the soft-heartedness of Hypermestra he is suggesting
that Lyde should show the same quality to him (see note on viro in
v. 46). Fraenkel finds in Hypermestra’s speech ‘a dignity of thought and
expression worthy of tragedy’ (197), but it seems nearer the mark to
think of melodrama. And when the story is set in the context of an
attempted seduction, the discrepancy would appear to justify Pasquali’s
verdict on the ode: ‘frivolo scherzo’ (144).
Metre: Sapphic.
1. Mercuri: Hermes was the inventor of the tortoise-shell lyre (1. 10. 6
with N–H, epod. 13. 9, h. Herm. 25 ff., Frazer on Apollod. bibl. 3. 10. 2);
so he could be regarded as a patron of poetry, though less exalted than
Apollo. As the god of persuasion (N–H on 1. 10. 1 facunde), he was a
suitable intermediary in the poet’s courtship of Lyde; cf. N–H on 1. 30. 8,
citing Cornutus, nat. deor. 24 where he is associated with Aphrodite. H
regards him as his protecting deity (N–H vol. 1, pp. 127 f.), who had
rescued him at Philippi (2. 7. 13 f.) even before he became a significant
poet; there was also apparently an astrological connection (2. 17. 29 f.
with N–H). For the god’s name at the very beginning of a hymn cf. 1. 10,
1. 30, Cairns, op. cit. 138 n. 2.
1–2. —nam te docilis magistro / movit Amphion lapides canendo—:

in hymns a god’s past assistance to the suppliant, or as here to others, is
commonly used as a reason for requesting his help (1. 32. 1 with N–H,
Pulleyn 17, 35 f.). In both sacred and secular contexts the reason for the
address can be given in a parenthetic clause following the vocative;
cf. epod. 17. 45, serm. 2. 6. 51 f., Hom. Il. 24. 334 ff. , ¯ æÆ, d ªæ
ºØ ª ºÆ KØ = Iæd ÆØæÆØ, Alcaeus 308. 1 f., Virg.
Aen. 1. 65 ‘Aeole, namque tibi’ (with Austin), Ov. met. 1. 2 (with Bömer),
Milton, PL 3. 654 ff. ‘Uriel! for thou . . . ’, J. D. Denniston, The Greek
Particles, 1954: 69, H–Sz 472. For Hermes’ tuition of Amphion cf. Paus.
9. 5. 8 (mentioning a hexameter poem on Europa), Apollod. 3. 5. 5 with
Frazer, Philostr. imag. 1. 10. 1; similarly, Bacchus appears as a music-
teacher at 2. 19. 1 f. ‘Bacchum . . . vidi docentem’. docilis, pointedly juxta-
posed with magistro, indicates that Amphion was an apt pupil (cf. 4. 6.
43 f. ‘docilis modorum / vatis Horati’); for a word’s insertion in the
middle of an ablative absolute cf. 3. 29. 1–2 n.
Amphion with his brother Zethus built the walls of Thebes (Hom.
Od. 11. 262 f., who gives no details); in other accounts he attracted the
stones by the music of Hermes’ lyre, notably in the Antiope of Euripides;
there Hermes says to him (łÆØ Ø = æÆØ KæıÆd ıØŒfi Ð
ŒºÆØ = æ æe KŒºØ Ł ºØÆ ( J. Diggle, Trag. Graec.
Frag. Selecta, 1998, p. 92, 92 ff. ¼ GLP, p. 68, 86 ff.). The artistic
Amphion was celebrated in Hellenistic poetry (Ap. Rhod. 1. 740 f.);
a similar source may lie behind Virg. ecl. 2. 24 ‘Amphion Dircaeus
in Actaeo Aracyntho’ (Clausen ad loc. suggests Parthenius); cf. Prop.
3. 15. 42. In epist. 1. 18. 41 f. H takes an unsentimental line, favouring the
practical Zethus; elsewhere he rationalizes Amphion’s song to suggest
the civilizing power of literature (ars 394 ff. with Brink). For the famous
Antiopa of Pacuvius (cf. 5 n.), which was based on Euripides (Cic. fin.
1. 4), cf. M. Valsa, Marcus Pacuvius poète tragique, 1957: 10 ff. See further
Sen. Phoen. 566 ff., RE 1. 1944 ff., Pease on Cic. div. 2. 133, Lieberg 37 ff.,
LIMC 1. 854 ff.
Propertius cites the exemplum of Amphion to show the uses of poetry
in courting a woman (3. 2. 5 f. ‘saxa Cithaeronis Thebis agitata per artem
/ sponte sua in muri membra coisse ferunt’); so there may have been a
Hellenistic prototype. H’s movit lapides suggests the persuasion of stony
hearts; cf. Ov. am. 3. 7. 58 ‘surdaque blanditiis saxa movere suis’ and
Fedeli on Prop. 1. 9. 31. canendo could include both music and words
(ars 395 cited above); in the myth of Amphion the former was more
important, with H the latter.
3–4. tuque, testudo, resonare septem / callida nervis: for the bowl-
shaped lyre, here of tortoise-shell, cf. M. L. West (1992), 56 f.: it was ‘the
ordinary instrument of the non-professional’, and hence appropriate to
the courtship of Lyde. For addresses to the testudo or
ºı cf. 1. 32. 14,
Sappho 118, Alcaeus 359 L–P. tu marks the change of addressee (cf. Liv.
1. 32. 10 ‘audi, Iuppiter, et tu, Iane Quirine’); from now on the lyre is
invoked rather than Mercury.
For the seven strings cf. Pind. P. 2. 70, N. 5. 24, Eur. Alc. 446 f., West,
op. cit. 62; the substitution of the heptachord for the tetrachord was
usually attributed to Terpander (D. A. Campbell, Greek Lyric 2, 1988:
testimonia 1 and 16, fr. 6), but sometimes to Amphion (Paus. 9. 5. 7, cf.
Philostr. imag. 1. 10. 5. e b E
ıº, ‹Ø B ºæÆ ƒ Ø).
callida transfers to the lyre the trickiness of Mercury himself (1. 10. 7);
for the epexegetic infinitive cf. N–H on 1. 1. 8. nervis is probably dative
(cf. serm. 1. 4. 76 ‘suave locus voci resonat conclusus’); the tortoise-shell
acted as a sound-box for the strings.
5. nec loquax olim neque grata: the apposition alludes to the deity’s
origin; cf. N–H on 1. 10. 1 citing Norden (1913), 148. Since Latin does
not distinguish the tortoise from its shell, H was able to conflate the
two; cf. (in a different context) Juv. 11. 94 f. In life the tortoise makes
no significant sound (Arist. hist. anim. 536a7, Pacuvius, Antiopa 2 ff.
R. ‘quadrupes tardigrada agrestis humilis aspera / capite brevi, cervice
anguina, aspectu truci, / eviscerata inanima cum animali sono’), but now
it can talk; for the same antithesis cf. h. Herm. 38 j b Łfi Œ
ºÆ ŒÆºe IØ , Soph. TrGF 4. F314. 300 (¼ GLP, pp. 46 ff., vv.
240 ff.), Nicand. alex. 560 f., Manil. 5. 324 ff. ‘nunc surgente Lyra testu-
dinis enatat undis / forma per heredem tantum post fata sonantis . . . ’.
loquax, like ºº , suggests fluent chatter rather than eloquence. grata
points to a contrast between the prized shell and its ugly inhabitant; so
also Pacuvius, loc. cit.
5–6. nunc et / divitum mensis et amica templis: the contrast between

‘formerly’ and ‘now’ was a common device in ancient literature (cf.
3. 26. 1 n.., Cairns, op. cit. 129 f., 138 n. 3). The most relevant parallels
here are the transformations of one object by craftsmanship into an-
other; cf. serm. 1. 8. 1 ff. ‘olim truncus eram ficulnus, inutile lignum’ with
Fraenkel 121 f., Philippus, anth. Pal. 6. 99, Simmias, 6. 113 (a bow made
from antlers), Crinagoras 6. 229, Fedeli on Prop. 1. 16. 1–5.
The lyre is called ‘comrade of the feast’ (ÆØd . . . Ææ) at Hom.
Od. 17. 271, h. Herm. 31; this suits mensis, but in conjunction with temples
amica need mean no more than ‘welcome friend’. For music at Roman
banquets cf. Blümner (1911), 411, Wille 143 ff.; H is not just thinking of
the ‘lays of ancient Rome’ mentioned at 4. 15. 29 ff., Cic. Tusc. 4. 3. For
the lyre at sacrifices cf. Porph. ‘fidicines hodieque Romae sacrificiis
adhiberi videmus’, 1. 36. 1 with N–H, Wille 29 ff. templis does not
quite balance divitum mensis, and deorum must be understood; for a
similar combination of men and gods cf. Varro, rust. 3. 16. 5 (of honey)
‘et deis et hominibus est acceptum’, Virg. georg. 2. 101 (of a grape).
Baehrens proposed caelitum for divitum (cf. 1. 32. 13 f. ‘dapibus supremi
/ grata testudo Iovis’); but then the first et ought to have come before
mensis.
7–8. dic modos Lyde quibus obstinatas / applicet auris: dic modos suits
both verses and music (3. 4. 1–2 n.). The name Lyde, here in a prominent
position before quibus, has exotic associations (N–H on 2. 11. 22). For
‘bending the ears’ cf. carm. saec. 71 f. (of Diana) ‘votis puerorum amicas /
applicat auris’, Symm. epist. 3. 6. 1. obstinatus sometimes has a good
sense (Liv. 1. 58. 5 ‘obstinatam pudicitiam’), but here Lyde’s obduracy is
not regarded as a merit; as the word implies stiffness, a contrast with
applicet (which suggests bending) is obvious.
9. quae velut latis equa trima campis: in Greek lyric poetry a stanza is
sometimes connected with its predecessor by a relative pronoun; Cairns
(op. cit. 130) cites exempli gratia Alc. 34. 5, Pind.O. 5. 4, 6. 29, 8. 67, 13. 63.
For the comparison of girls with fillies cf. Anacr. PMG 417. 1 Hº
¨æfi Œ, Ar. Lys. 1308, Lucil. 1041M ‘anne ego te vacuam atque ani-
mosam, / Tessalam ut indomitam, frenis subigamque domemque?’,
V. Buchheit, Studien zum Corpus Priapeorum, 1962: 104 n. Horses bred
for speed need space to exercise (Colum. 6. 27. 2 ‘spatiosa . . . pascua’),
and here the girl has freedom to roam (cf. N–H on 2. 5. 5 f. ‘circa virentis
est animus tuae / campos iuvencae’); contrast the walled garden of the
secluded maiden (Catull. 62. 39 ff., Virg. ecl. 8. 37, Ov. met. 14. 635 f.),
and the medieval figure of the hortus conclusus. Mares were mated at the
age of 2 (Colum. 6. 28. 1) or better 3 (Arist. hist. anim. 6. 575b24).
10. ludit exsultim metuitque tangi: H is echoing Anacreon 417. 5

ºØH ŒÆØ = ŒF ŒØæHÆ ÆØ (see last note); ludit
implies high-spirited capers but not sexual activity (contrast 2. 5. 8,
3. 15. 12). The drily archaic exsultim is attested only here, but cf. Suet.
Aug. 83 subsultim, Gell. 9. 4. 9 saltatim, which is given as a gloss by
ps.-Acro on our passage; for exsultare of horses cf. Cic. off. 1. 90
‘ferocitate exsultantes’, Nep. Eum. 5. 5, TLL 5. 2. 1948. 16 ff. metuit
indicates not just timidity but actual shying away; there is a sexual
innuendo in tangi (cf. serm. 1. 2. 54 ‘matronam nullam ego tango’, and
the concept of virgo intacta).
11–12. nuptiarum expers et adhuc protervo / cruda marito: expers

means ‘having no part in’; cf. Virg. Aen. 4. 550 with Pease, Stat. Theb.
7. 298 f. ‘expertem thalami crudumque maritis / ignibus’, Musaeus 31
with Kost, TLL 5. 2. 1689. 30 ff. nuptiarum is ambiguous, as it can refer
to sexual experience as well as marriage; cf. Plaut. cist. 43 ‘haec . . . cotidie
viro nubit’, auct. ad Her. 4. 45 ‘cuius mater cottidianis nuptiis delectetur’,
Petr. 26. 3. cruda means ‘unripe’, the opposite of ‘iam matura viro’ (Virg.
Aen. 7. 53); for the combination of imagery from animals and fruit cf.
2. 5. 10, Theoc. 11. 21, Catull. 17. 15 f. protervo suggests aggressive
masculinity; the word can be used of animals (2. 5. 15 implying a piquant
reversal of roles, Ov. met. 14. 63); marito can be used of animal mates
(1. 17. 7 with N–H, 2. 5. 16). The marital connotations of nuptiarum and
marito provide a link with the story of Hypermestra.
13–14. tu potes tigris comitesque silvas / ducere et rivos celeris morari:

these are the accomplishments regularly attributed to Orpheus (indeed
vv. 15 ff. must allude to him rather than Amphion); cf. 1. 12. 7 ff. (a later
poem) with the parallels cited by N–H. In particular H is alluding to
Virg. georg. 4. 510 ‘mulcentem tigris et agentem carmine quercus’; there
may be a humorous suggestion that savage tigers are led by the slow and
lowly tortoise. In the rationalizing version in ars 391 ff. Orpheus is again
combined with Amphion. For the rivers see N–H on 1. 12. 9 (citing
Ap. Rhod. 1. 26 f.), Clausen on Virg. ecl. 8. 4. The lyre checks their
progress by holding their attention (OLD morari 3); the verb points a
contrast with both celeris and ducere. tu potes suggests the hymnal style
(N–H on 1. 28. 28, Fedeli on Prop. 1. 14. 17); the lyre can do these things,
because it has already done them; cf. 2. 19. 17 ‘tu flectis amnis’ with
N–H. comites is to be taken with both tigris and silvas; the word-order
is characteristic of Greek lyric poetry, particularly hymns (N–H on
2. 19. 27, Cairns, op. cit. 138 n. 11).
15–16. cessit inmanis tibi blandienti / ianitor aulae: for descents to

the underworld see N–H vol. 2, pp. 203 ff., Mynors on Virg. georg.
4. 453 ff., Bömer on Ov. met. 10. 1 ff., R. J. Clark, Catabasis: Vergil
and the Wisdom Tradition, 1979: 79 ff. (Heracles), 95 ff. (Orpheus).
For the need to neutralize Cerberus by music or other means cf. 2. 13.
33 ‘illis carminibus stupens’ with N–H, 2. 19. 29 ff. (Bacchus gets
past), Virg. georg. 4. 483 ‘tenuitque inhians tria Cerberus ora’, Synesius,
hymni 8. 19 ff. on Christ’s descent æ ªæø = #Æ ›
ÆºÆØª , = ŒÆd ºÆ æ Œø = . . . › ÆæıŁc I
Æ ºF
(cf. cessit). Terzaghi ad loc. sees a common source behind Synesius,
Virgil, and our passage.
Is inmanis genitive with aulae or nominative with ianitor? In favour of
the first is the fact that aulae needs qualification (cf. Sil. 2. 552 ‘lacrimo-
sae ianitor aulae’) and can be supported by phrases like ‘spelunca imma-
nis’ (Virg. Aen. 6. 237) and ‘immane barathrum’ (ibid. 8. 245); this is the
preference of most editors, including NR, who thinks it may be imitated
in Matthew Arnold’s ‘the vasty hall of death’ (Requiescat 16). RN thinks
that after cessit the reader would naturally take inmanis as a contrasting
nominative and apply it to the monstrous Cerberus (cf. Virg. Aen. 6. 400
‘ingens ianitor’, 6. 417 f. ‘Cerberus adverso recubans immanis in antro’);
the misunderstanding would not be removed until aulae (16). He has
considered immensae (cf. Sen. Tro. 178 ‘immensos specus’), but that
might not be enough to identify the aula. He has also thought of
reading in manis for inmanis of many MSS (cf. Hom. Od. 11. 563 f.
of Ajax B b ¼ººÆ = łı
a N ! ¯ æ Œø ŒÆÆŁø,
Synes. loc. cit. 17 f.); but Ajax can join the shades more appropriately
than Cerberus. He has also tried emending aulae to Orci; if ianitor Orci
was compressed by haplography into ianitorci, then aulae might have
been supplied to fill the gap.
tibi after tu (13) exemplies the ‘Du-Stil’ common in hymns (3. 21. 13 n.);
the pronoun is emphatic, not enclitic (Nisbet ap. Adams–Mayer, 1999:
149). blandienti, which suits the charms of Orpheus (1. 12. 11, 1. 24. 13)
hints at the blandishments of the poet’s own courtship (cf. Stroh 115 ff.).
For aula as ‘the hall of death’ cf. 2. 18. 31 with N–H, Prop. 4. 11. 5
‘fuscae . . . aulae’; for the entrance to the underworld cf. Hom. Il. 5. 646,
H. Usener, Kleine Schriften 4, 1913: 226 ff., Bömer on Ov. met. 4. 453.
A door-keeper was not a friendly character in the ancient world; for
Cerberus in this capacity cf. Eur. Her. 1277, Virg. Aen. 8. 296 ‘ianitor
Orci’, anon. anth. Pal. 7. 319. 1 f., Roscher 3. 2. 3331, TLL 7. 1. 132. 66 ff.
17–20. Cerberus . . . : this stanza, which was known to Porph., has been
deleted by several scholars, including Buttmann, Naeke (Opuscula Phi-
lologica 1, 1842: pp. 73 ff.), Peerlkamp, L. Müller, Heinze; it has been
defended by Jahn (ap. Orelli), Campbell (edn. 2), Williams, Syndikus
(2. 121 f.), Bradshaw (RhM 118, 1975: 311 ff.), Cairns (op. cit. 130 f.). In our
view there is nothing problematic about it except eius atque (18 n.). The
lines are alleged to disrupt the high-flown sequence of 13–16 and 21–4,
but they too underline the power of poetry; the three stanzas on
Orpheus (13–24) balance their three predecessors. The appositional
Cerberus has been thought an interpolated explanation of ianitor that
was built up into a stanza (Naeke, op. cit., Fraenkel on Aesch. Ag.
1224–6), but such appositions are well attested; cf. for instance Hom.
Il. 6. 394 f. Ł ¼º
ºøæ KÆ qºŁ ŁıÆ, = #æ
,
Virg. Aen. 7. 761 ‘ibat et Hippolyti proles pulcherrima bello, / Virbius’.
17–18. quamvis Furiale centum / muniant angues caput: for the

appended concessive clause cf. 2. 19. 25 ff.; the subjunctive has been
doubted, but is normal with quamvis in Republican Latin (K–S
2. 442 f.), and is used by H in the same sense at 4. 6. 6 f. ‘filius quamvis
Thetidis marinae / Dardanas turris quateret’. Cerberus had snakes in his
hair, like the Furies (2. 13. 35 f. ‘intorti capillis / Eumenidum . . . angues’,
Virg. Aen. 6. 419, where they are seen as a kind of mane, Billerbeck on
Sen. Hf 785–7, LIMC 6. 1. 31); for muniant cf. Lucr. 5. 27 ‘hydra
venenatis . . . vallata colubris’. He was usually given three heads (Cic.
Tusc. 1. 10, Virg. Aen. 6. 417, Ov. met. 4. 450 f.), though sometimes fifty
or a hundred (2. 13. 34, Hes. theog. 311 f., Billerbeck on Sen. Hf 784 ); in
art for obvious reasons the number of heads varies from one to three
(LIMC 6. 1. 24 ff., 6. 2. 12 ff.). In our passage the singular caput has been
impugned by those who delete the stanza, but furiale caput is not
inconsistent with a plurality of heads; cf. Sen. Hf 784 f. where Cerberus’
three heads (trina capita) are followed by the singular caput: ‘sordidum
tabo caput / lambunt colubrae’; see further on v. 20.
18–19. yeius atquey / spiritus taeter: the first two words present three
problems. (1) the genitive eius is very rare in high poetry (Axelson 72); it
suits the somewhat old-fashioned style of Lucretius (35 instances) and is
attested in elegy (Tib. 1. 6. 25, Prop. 4. 2. 35, 4. 6. 67, Ov. trist. 3. 4. 27,
Pont. 4. 15. 6 and also at met. 8. 16), but though found in H’s Sermones
(2. 1. 70, 2. 6. 76) it appears elsewhere in the Odes only in the suspect
4. 8. 18. (2) Though atque ends a line at 2. 10. 21, the sequence of two
inert words is uncharacteristically clumsy. (3) Though parallels can
be found for a verb suiting the second of its two subjects better than
the first, spiritus does not combine well with manet (19), which implies
some kind of liquid. Taken together, these points have led us to obelize
the phrase (though NR has doubts), and to look for a verb that has been
displaced by eius atque.
Bentley proposed exeatque (cf. Ov. met 3. 75 f. ‘halitus exit / ore niger
Stygio’), though he conceded that elsewhere exit spiritus refers to dying
breath (Ov. ars 3. 745, trist. 4. 3. 41); but one would have preferred
something less comprehensive and more clearly differentiated from
manet below. Gesner (1752) proposed effluatque, which occurred inde-
pendently to Housman (Classical Papers 1. 3); the verb could be applied
to breath (Cic. nat. deor. 2. 101 ‘aer effluens’, TLL 5. 193. 56 ff.), but is also
perhaps too similar to manet. Cunningham (1721) proposed aestuatque
(making the three verbs in the stanza indicative), and Williams con-
sidered aestuetque (while retaining eius atque); the word would refer to a
hot dry blast as at Lucr. 3. 1012 ‘Tartarus horriferos eructans faucibus
aestus’, Virg. Aen. 8. 258 ‘nebulaque ingens specus aestuat atra’, Sil. 6. 219
‘(serpens) Stygios aestus fumanti exsibilat ore’. aestuetque seems the
best solution; unlike exeatque, it has the advantage of standing alone
and not combining with ore trilingui. The archaic taeter is avoided by
most Latin poets; but as it is found in Virgil (Aen. 3. 228, 10. 727) and
Seneca, Heinze was wrong to count it one of the suspicious features in
the stanza.
19–20. saniesque manet / ore trilingui: for Cerberus’ poisonous dis-

charge cf. Ov. met. 4. 501 ‘oris Cerberei spumas’ (carried by the Fury),
Plin. nat. hist. 27. 4 (aconite is produced by the froth). Porph. suggests
that the blood comes from human bodies, and Cairns (op. cit. 131) sees
a reference to Œæ æ ‘meat-eating’, the supposed etymology of
Cerberus (Maltby 121); but the sanies seems to be produced by the dog
himself (cf. Ov. met. 4. 494 of the Fury’s snakes ‘saniemque vomunt’,
Plin. nat. hist. 27. 50 ‘si [aures] manent sanie’). For the ablative of source
with manare cf. 2. 9. 1, Ov. met. 3. 85 with Bömer, TLL 8. 320. 14. ore
trilingui seems to be a condensed way of saying ‘from his three muzzles,
each with its tongue’; cf. 2. 19. 31 f. with N–H’s note: ‘Cerberus had
usually three heads, and therefore three mouths, three tongues, and
three barks’.
21–2. quin et Ixion Tityosque vultu / risit invito: quin introduces the
climax of the underworld scene, as at 2. 13. 37; both passages must be
influenced by Virg. georg. 4. 481 f. ‘quin ipsae stupuere domus atque
intima leti / Tartara’. Ixion attempted to seduce Hera, and was punished
by being tied to a revolving wheel (Pind. P. 2. 21 ff. with schol., Prop.
1. 9. 20 with Fedeli, RE 10. 2. 1373 ff., LIMC 5. 1. 857 ff., 2. 555 ff.). For
Tityos cf. 3. 4. 77 n. The respite from the torments of the damned (Virg.
georg. 4. 484, Prop. 4. 11. 23 ff., Ov. met. 10. 41 ff., Sen. HO 1068 ff.)
presumably goes back to a lost catabasis of Orpheus.
On vultu invito Porph. comments ‘intellegas tantam fuisse gratiam
cantus, ut in tormentis poenae constitutis extorserit tamen risus’; i.e.
they couldn’t help smiling in spite of their pain. In other contexts the
phrase might more naturally have suggested a deliberately forced smile;
cf. Aesch. Ag. 794 with Fraenkel, Stat. Ach. 1. 194 ‘ficto risit Thetis anxia
vultu’. RN has considered insueto; cf. Tib. 1. 4. 48 with Murgatroyd,
Lucan 5. 163 ‘insueto concepit pectore numen’, where invito is a variant.
22–4. stetit urna paulum / sicca, dum grato Danai puellas / carmine
mulces: for the myth of the daughters of Danaus see the introduction.
urna refers to the pitcher of each individual Danaid (cf. Plato, Gorg.
493b quoted in 26–7 below); the girls are so entranced that instead of
refilling their pitchers, they set them down on the ground. H is
following the usual account by which the holes were in the dolium,
not the urnae (27 n.). mulcere (lit. ‘to stroke’) is used regularly of the
charms of music (Lucr. 5. 1390, Ov. fast. 2. 116, notably in Virgil’s
description of Orpheus in georg. 4. 510). H hopes that his song will
have an equally agreeable effect on Lyde.
25–6. audiat Lyde scelus atque notas / virginum poenas: for the subtle
transition see the introduction. Porph. points out the mock-serious
moral: ‘audiat Lyde qua poena damnatae sint quae crudeles amatoribus
fuerint’; cf. Tib. 1. 3. 79 ff. ‘et Danai proles, Veneris quod numina laesit, /
in cava Lethaeas dolia portat aquas. / illic sit quicumque meos violavit
amores’. H cunningly gives the impression that virginity was part of
their crime; he clearly ignores the version by which they were forced to
have intercourse (Apollod. 2. 1. 5), as that would have diminished their
guilt and confirmed Lyde’s suspicions of men.
26–7. et inane lymphae / dolium fundo pereuntis imo: et adds an

explanation of poenas; cf. OLD s. v. 11. For the genitive with inane cf.
Cic. de orat. 1. 37, K–S 1. 441. pereuntis means ‘going to waste’; the verb is
used of leaks at Frontin. aqu. 2. 88 ‘ne pereuntes quidem aquae otiosae
sunt’, Mart. 12. 50. 6 ‘et pereuntis aquae fluctus ubique sonat’. The large
dolium, which held 10–15 amphorae, was big enough for Diogenes;
cf. Hilgers 171 ff. fundo means ‘through the bottom’; imo is pleonastic,
like summus with vertex.
H puts the leaks in the dolium to which the water is being carried
(cf. Tibullus quoted in the last note, Phaedrus, app. 7. 10 ‘urnis scelestae
Danaides portant aquas, / pertusa nec complere possunt dolia’); this
is the predominant version in the iconographic tradition (LIMC
3. 2. 250 f.). Leaky pitchers are assigned to the Danaids by Sen. Med.
748, but this was not Seneca’s innovation (pace Costa ad loc.): both
forms of the legend are applied to nameless sinners at Plat. Gorg. 493b
æE N e æ Ł oøæ æfiø Øfiø æfiø
ŒŒfiø. H has the same account as Bion fr. 28 Kindstrand (¼ Diog.
Laert. 4. 50), Paus. 10. 31. 9.
28–9. seraque fata / quae manent culpas etiam sub Orco: sera means
‘long-postponed’ (cf. Virg. Aen. 6. 569 ‘distulit in seram commissa
piacula mortem’); for the slowness of retribution cf. 3. 2. 32 n. For the
sinister manent cf. epod. 13. 13, N–H on 1. 28. 15, 2. 18. 31. sub Orco means
‘in Hades below’, not ‘under the rule of Orcus’; it is sometimes hard to
distinguish the two meanings (3. 4. 74–5 n.), but in the context of the
underworld sub is naturally taken as local; moreover, etiam would lack
point if Orcus were personal, for one expects the ruler of the dead to
punish people.
30. impiae (nam quid potuere maius?): impiae is emphasized by both

its position and its isolation (see further on 31); the violation of marital
loyalties is central to H’s version of the story, as later in Ovid, her. 14.
quid maius? would normally mean ‘what greater deed?’, but here the
context requires ‘greater in the scale of evil’, i.e. ‘more monstrous’; cf.
Prop. 4. 7. 67 ‘narrat Hypermestre magnum ausas esse sorores’. Heinze
suggests that the parenthesis points forward to 31, which specifies
murders of surpassing wickedness, rather than back to impiae; he there-
fore interprets nam quid not as causal but as the equivalent of quidnam?
(epist. 1. 1. 76, K–S 2. 117). Some editors begin the parenthesis with quid;
but though there are parallels in H for nam in the second place (4. 14. 9,
epod. 17. 45) and even for the break after the fourth syllable in the
Sapphic line (3. 11. 50), the combination of both irregularities is quite
unconvincing. RN sees attractions in numquid, a variant without au-
thority recorded by Bentley.
31–2. impiae sponsos potuere duro / perdere ferro: impiae is further

emphasized by its repetition after the parenthesis; cf. Quint. inst. 9. 3. 29
‘similis geminationis post aliquam interiectionem repetitio est, sed paulo
etiam vehementior’, Wills 66 ff. sponsos recalls the mutual pledges of
betrothal, and thus accentuates the disloyalty; note the repeated refer-
ences to marriage (33 face nuptiali, 37 marito). The reiterated potuere now
takes on the nuance of ‘could bear to’ (ºÆ); cf. epod. 9. 13 f. ‘(miles)
spadonibus / servire rugosis potest’. duro refers to both literal hardness
and cruelty (Porph.).
33–4. una de multis face nuptiali / digna: una, namely Hypermestra

(this is the correct form of the name); the asyndeton presents a sharp
contrast between evil and good. For Hypermestra’s uniqueness among
her sisters cf. [Aesch.] Prom. 865 (see on 35–6 below), Pind. N. 10. 6 P
, 5ææÆ Ææºª
Ł, - = łÆ K ŒºfiH ŒÆÆ
EÆ ;
for the common antithesis of one and many see Fraenkel on Aesch. Ag.
1455, Headlam on Herodas 6. 35.
34–5. periurum fuit in parentem / splendide mendax: the father’s

treachery justified the daughter’s deception (Danaus had pledged his
daughters in marriage to his nephews). For splendide mendax cf. Aesch.
TrGF fr. 301 I ØŒÆÆ , Ov. met. 9. 711, Tac. hist. 4. 50. 2 ‘egregio
mendacio’; oxymoron suits H’s succinct and pointed style (see
F. Muecke, Encicl. oraz. 2. 781). The assonance of the phrase was probably
thought attractive; cf. 1. 15. 19 f., 2. 1. 35 f., Catull. 11. 4 ‘tunditur unda’,
Virg. Aen. 6. 223 ‘triste ministerium’, 6. 314 ‘ripae ulterioris amore’.
35–6. et in omne virgo / nobilis aevum: for in omne aevum cf. Ov. met.
1. 663 ‘aeternum . . . in aevum’, TLL 7. 1. 751. nobilis balances splendide
and means both ‘famous’ (nosco) and ‘glorious’; for similar praise of
Hypsipyle’s rescue of her father cf. Val. Fl. 2. 243 ff. By one account
Lynceus left Hypermestra a virgin (Apollod. 2. 1. 5 ¸ıªŒÆ Øø
ÆæŁ ÆPc ıºÆÆ), by another he did not ([Aesch.] Prom. 865
¥æ ŁºØ e c = ŒEÆØ ı, schol. on Eur. Hec. 886
, 5 ææÆ KÆ F ¸ıªŒø , Ie B ø ØŁØ K
ŒıEÆ
æe ÆP . H calls her virgo because that is how the Danaids were
described (otherwise Cairns, loc. cit.).
37. ‘surge’ quae dixit iuveni marito: for exhortations to rise up in

various contexts cf. Hom. Il. 18. 170 (Iris to Achilles) Zæ, Virg. Aen.
3. 169, 8. 59, Ov. her. 14. 73 (Hypermestra to Lynceus) ‘surge age, Belide’,
Val. Fl. 2. 249 f. (Hypsipyle speaks) ‘fuge protinus urbem, / surge, pater’,
Aus. ephem. 1. 17 ff. (an amusing adaptation of our passage). Such
imperatives are commonly found in the ‘Aubade’, where the woman
urges her lover to leave at dawn; cf. anon. PMG 853. 2 Iø, A. T.
Hatto, Eos, An Enquiry into the Theme of Lovers’ Meetings and Partings
at Dawn in Poetry, 1965, Cairns (1972), 84 f. (on the ‘diegertikon’) and op.
cit. 135 f. iuveni marito echoes the pathos of virgo; Lynceus, too, is
unnamed. marito suggests ‘bridegroom’; the murder was committed
within hours of the wedding.
38–9. ‘surge, ne longus tibi somnus unde / non times detur: longus is
euphemistic for aeternus (N–H on 2. 14. 19) and somnus for mors (N–H
on 1. 24. 5). ‘From where you least expect it’ is vaguely sinister; the
source of the danger is then specified in 39 f.
39–40. socerum et scelestas / falle sorores: the alliterative sibilants bind

the phrase and suggest a furtive whisper. Family affection should have
been expected from a socer (a more emotive word than ‘father-in-law’);
cf. Catull. 72. 3 f. ‘dilexi tum te . . . pater ut gnatos diligit et generos’; so
scelestum should probably be understood, though the emphasis is on
sorores. sorores can include ‘sisters-in-law’; cf. the corresponding use of
frater (OLD 2); after socerum (tuum) it would be awkward to understand
meas. falle, like ºŁ, means ‘give the slip to’; cf. 1. 10. 16, epist. 1. 5. 31
‘postico falle clientem’, Bömer on Ov. met. 4. 85.
41–2. quae velut nactae vitulos leaenae / singulos eheu lacerant: the
simile of lion and cattle comes from high poetry; cf. Hom. Il. 5. 161, Eur.
IT 296 f. (on Orestes) › b
æd Æ , =
ı OæÆ N
Æ ºø ‹ø , = ÆØ Øæfiø ºÆª Æ . vitulos suits both animals and
young men (cf. 2. 8. 21 iuvencis); for nactae of finding prey cf. epist.
1. 15. 38, Ter. hec. 65 ‘spolies mutiles laceres quemque nactus sis’. After
the collective hunt, singulos turns to the separate murders; cf. [Aesch.]
Prom. 862 (on the same story) ªıc ªaæ ¼æ (ŒÆ ÆNH æE.
lacerant (‘rend’), though possible of daggers, is particularly suited to
lions’ claws; for the transference of elements from the simile to the thing
illustrated see M. S. Silk, Interaction in Poetic Imagery, 1974, passim,
Davies on Soph. Trach. 31 ff. The horror of lacerant is emphasized by
the interjection eheu and the delay imposed by the parenthesis; for
Hypermestra’s imaginative sympathy cf. Ov. her. 14. 35 ‘circum me
gemitus morientum audire videbar’.
42–4. ego illis / mollior nec te feriam, neque intra / claustra tenebo:
mollior makes a contrast with duro (31). claustra can mean either ‘bolts’
(which suits a bedroom) or ‘an enclosed chamber’ (which suits intra); as
the word is sometimes used of an animal’s cage, RN thinks it may be
relevant that Lynceus suggests ‘a lynx-man’. Hypermestra refuses even
to detain her husband till Danaus’ arrival.
45–6. me pater saevis oneret catenis: Hypermestra is willing to be

imprisoned as long as Lynceus escapes. Her forebodings were justified
in the usual form of the legend (otherwise Cairns, op. cit. 136): see
Apollod. 2. 1. 5 ŒÆŁæÆ ÆPc ˜ÆÆe KææØ, Ov. her. 14. 3 ‘clausa
domo teneor gravibusque coercita vinclis’. For the use of oneret cf. TLL
9. 631. 52 ff.
46. quod viro clemens misero peperci: viro, set against pater (45),
illustrates the conflict of family loyalties; but in the general sense of
‘man’ it can refer to Horace, hinting that Lyde should similarly spare
the poet. clemens combines naturally with misero; for the juxtaposition
of complementary words cf. 1. 3. 10 ‘fragilem truci’, 3. 7. 13 ‘perfida
credulum’.
47–8. me vel extremos Numidarum in agros / classe releget: Danaus
was king of Libya as well as of Argos. Numidarum suggests savagery; cf.
2. 6. 3 ‘barbaras Syrtis’, Eur. Hel. 404, Pease on Virg. Aen. 4. 41. extremos
means ‘at the end of the earth’, not ‘the most remote parts of Numidia’;
cf. 3. 10. 1.
relegare is used by the poets in a non-technical sense (‘to banish’); cf.
Cic. poet. frag. e Sophocle, 70. 16. T ¼ Cic. Tusc. 2. 20 (vers.) 16, Virg.
georg. 3. 212, Aen. 7. 775, OLD 2. classe seems to have the general meaning
‘by ship’ (OLD 3b); cf. the use of Æı at Pind. P. 10. 29, where it is
contrasted with . RN thinks H may be alluding to Danaus’
connection with the beginning of navigation (Plin. nat. hist. 7. 206
‘nave primus in Graeciam ex Aegypto Danaus advenit’, Hygin. fab.
277. 5); note the references to ships in Aeschylus’ Suppliants (177 fiH
ÆıŒºæfiø Ææ, 713 ff., 764 ff.).
49. i pedes quo te rapiunt et aurae: Lynceus is to go quickly by land

and sea wherever chance takes him (cf. epod. 16. 21 f. ‘ire pedes quocum-
que ferent, quocumque per undas / Notus vocabit aut protervus Afri-
cus’); the present rapiunt is livelier than the more literal future. When
the feet are said to carry someone, that may simply describe walking
(Hom. Il. 18. 148); but usually the expression suggests non-deliberate
movement, either because the journey is familiar (Plin. epist. 7. 5. 1) or
because the direction is left to chance (Theoc. 13. 70 fi Æffl pª K
æØ
with Gow, paroem. Gr. 1. 404, Otto 275 f.). aurae carries the same
connotation; cf. Plat. rep. 394 d ‹ i › º ª uæ FÆ æfi .
50. dum favet Nox et Venus: for the warning cf. Ovid’s imitation, her.
14. 77 ‘dum nox sinit, ‘‘effuge,’’ dixi’. Apart from hints in 35 f., the
combination of Nox and Venus (cf. 3. 28. 13 ff.) is the first clear indication
of a love interest to balance the poet’s courtship of Lyde; this feature is
lacking in Ovid’s version, where Hypermestra concentrates on pietas and
duty (H. Jacobson, Ovid’s Heroides, 1974: 125 f.). Yet in Aeschylus’
Danaids Aphrodite brings about her acquittal by an Argive court; the
famous lines on the marriage of Earth and Heaven were spoken by the
goddess on that occasion (TrGF fr. 44); and when the case was over,
Hypermestra dedicated a statue to her (Paus. 2. 19. 6). A glass vessel of
the third century ad shows Lynceus fleeing in the presence of a Pothos
or Cupid (W. H. Friedrich, A & A 12, 1966: 6 with Abb. 2 ¼ LIMC 5. 1.
589 f.) For the rejection of sex by the other Danaids see F. Zeitlin,
Playing the Other, 1996: 123 ff.
50–2. i secundo / omine et nostri memorem sepulcro / scalpe quere-

lam’: these lines contain standard features of the propempticon: the
departing friend is sent on his way with a hope for good omens (contrast
3. 27. 1 ff.), and urged not to forget the speaker (cf. 3. 27. 14 ‘et memor
nostri, Galatea, vivas’ with the note ad loc.). The tomb belongs to
Hypermestra herself, who with growing pessimism realizes that the
parting is permanent; cf. Porph.’s gloss ‘monumento meo’ and Ov. her.
14. 128 ff. ‘sculptaque sint titulo nostra sepulcra brevi’. It cannot be that of
Lynceus (thus L. Müller, Williams) after he has been sent away secundo
omine; rather he is to write an inscription in the manner of the laudatio
Turiae (OCD 822). Cairns (op. cit. 136) refers to the tradition that the pair
were buried together (Paus. 2. 21. 2): ‘by mentioning the tomb, which
everybody was familiar with, Horace is delicately implying this happy
ending without destroying the pathos which he needs to conquer Lyde’.
But in Ovid’s expanded imitation Hypermestra’s epitaph is to read ‘exul
Hypermestra, pretium pietatis iniquum, / quam mortem fratri depulit,
ipsa tulit’ (her. 14. 129 f.), and Horace’s closure seems equally sombre
(even though such heroism will not be expected of Lyde); some editors
think of a cenotaph, but H leaves the question of location open.
sculpe is a variant without authority or merit; the older form scalpe
could mean ‘sculpt’ or ‘carve’ as well as ‘scratch’, though in the later
Empire the words were differentiated (Ernout–Meillet, Dictionnaire
étymologique 2. 1055 f.). sculpo is used in an imitation of our passage
carved by a Roman lady on the pyramid of Gizah during a visit to
Egypt in the second century ad: ‘Vidi pyramidas sine te, dulcissime
frater, / et tibi, quod potui, lacrimas hic moesta profudi, / et nostri
memorem luctus hanc sculpo querelam.’ For the whole inscription and
comment see Mus. Lap. no. 74 ¼ CLE 270.
1 2 . M I S ER A RV M ES T
[F. Cairns, QUCC 24, 1977: 138 ff.; Cavarzere 229 ff.; R. W. Fortuin, Der Sport im augus-
teischen Rom, Palingenesia 57, 1996: 196 ff.; W. Kissel, WS 14, 1980: 125 ff.; R. M. Nielsen in
Studies in Latin Literature and Roman History (ed. C. Deroux) 2, 1980: 237 ff.; Pasquali
86 ff.; Pöschl 324 ff.; Williams 203 ff.]
1–3. Wretched are the girls who can neither enjoy love nor drown their
sorrows in wine without incurring an uncle’s wrath. 4–9. You are distracted
from your household tasks, Neobule, by the winged Cupid and the gleaming
shoulders of Hebrus when he has swum in the Tiber—a better horseman than
Bellerophon, undefeated at boxing and running, (10–12) skilled too at
shooting the fleeing deer and waylaying the lurking boar.
The opening of this ode imitates in theme and metre the opening of
a fragmentary poem by Alcaeus: ºÆ, ÆÆ ŒÆŒø
1 2 . M I S ER A RV M ES T 165

ØÆ (10 L–P, cf. Page, Sappho and Alcaeus, 1955: 291 ff.); for such
‘mottoes’ see Cavarzere, op. cit. There the poet speaks unusually in the
character of a woman (for such role-playing cf. Anacr. PMG 385, Ascle-
piades, anth. Pal. 12. 153 with Gow–Page, HE 898 ff., Theoc. 2, Führer,
1967: 5 ff., H. Jacobson, Ovid’s Heroides, 1974: 343 ff., Fortuin, op. cit.
196 n.); she is clearly lamenting an unhappy love-affair (cf. 380 L–P
˚ıæªÆ ÆºÆØØ, which may well come from the same
poem). Because of this parallel most commentators think that Horace too
has assumed the character of a woman, though almost always he speaks in
propria persona (for exceptions cf. introduction to 3. 9); on the other hand,
the ancient commentators thought the poet himself was the speaker, in
which they are followed by Müller, Syndikus, Cairns, and Pöschl.
According to the former view, when the speaker says tibi (4), the girl is
addressing herself (for this practice cf. 2. 5, Catull. 8 etc., W. Schadewaldt,
Monolog und Selbstgespräch, 1926: 35 ff., Williams 461 ff.); but in this
context Horace might have been expected to follow his source and
write mihi for the sake of clarity. It is alleged that the emphasis on the
young athlete’s looks and accomplishments is more natural in the mouth
of a woman, but this is not a necessary assumption (cf. 1. 8, 3. 20. 13 ff., 4. 1.
38 ff.); the jaunty references to sex and drink (1–2) suit a confident male
better than a carefully supervised young girl. Nor need the mention of the
stern uncle (3) and wool-making (4) come from the girl: it is a stock
situation, and individualizing touches are not to be expected.
The ode also has affinities with a poem by Sappho (102 L–P):
ªºŒÆ Aæ, hØ ÆÆØ ŒæŒ e Y = Łfiø ØÆ ÆE
æÆÆ Ø A æÆ. Here as in Alcaeus the woman is describing
her own unfulfilled love, and as in Horace this is distracting her
from her domestic tasks (4 n.). The theme, which may belong to trad-
itional folk-song, is repeated in Goethe’s Gretchen am Spinnrad (Faust,
part 1, 3374 ff.) and Landor’s ‘Mother I cannot mind my wheel’ (for
singing at work see further Fortuin, op. cit. 202 n.). The despair of love-
sickness, particularly in women, is a typically Hellenistic motif (with
some encouragement from Euripides); cf. Theoc. 2, the ‘fragmentum
Grenfellianum’ (Collectanea Alexandrina, ed. Powell, 177 ff.), Rohde
(1914), 173 ff., S.Trenkner, The Greek Novella in the Classical Period,
1958: 66 (adding references to medical writers), Jacobson, loc. cit. No
such sentimentality is present in Horace’s poem, and none of the
emotional intensity of Alcaeus: the grievous hurt (Aæ IÆ) and
the frenzy (ÆØ ) have disappeared, and the tone is one of slightly
ironic detachment. Instead of showing the empathy with women that
we find in Virgil and sometimes in Ovid, Horace concentrates on the
vigour of the male athlete, which he describes with appropriate verve. It
is quite characteristic for him to take a ‘motto’ from a Greek poet and
then to give his ode an original direction; cf. 1. 9, 1. 18, 1. 37.
As usual with poems of this type, Horace gives the characters Greek
names (5–6 nn.), and his use of Greek mythology heightens the stylistic
level (4 n., 8 n.). At the same time, he sets the scene in the Campus
Martius, so the characters are felt to be Roman: a young man shows off
his prowess in swimming, riding, boxing, and running, sports encour-
aged by Augustus (N–H vol. 1, pp. 108 f., Fortuin, op. cit.). In a Greek
situation such activities attracted male admirers (though note Theoc. 2.
77 ff., Aristaenetus 1. 8, where the athlete is æØ Ł ! A øØ ÆE
ÆæÆØ ). In Horace it is a girl who is attracted, and her ambiguous
status perhaps reflects social realities: she is freer than most Greek
women to watch and admire, yet she has an old-fashioned guardian
(3 n.) who denies her wine (1–2 n.). At the end the poet turns to the very
Roman sport of boar-hunting (11–12 n.), but though such activities
might be found near the city (epist. 1. 6. 57), they would lie outside a
young girl’s experience; examples of women hunting are fictional, like
Atalanta in Ov. Met. 8. 317; see Anderson (1985), 89 ff.
The poem is written in ionics a minore (^^ — —), like the prototype
in Alcaeus (10 L–P); cf. M. L. West, Greek Metre, 1982: 124 ff. Synaphea
prevails throughout, i.e. there is metrical continuity between the lines
and no variation of prosodic rules at the end of a line; thus in v. 4 telam
(or vellus) would be impossible before the opening vowel of operosae.
The endings of the metra coincide with word-endings to an unusual
extent (Kissel, op. cit.); for exceptions note ex-animari (2), ver-bera (3),
operosae-que (5), la-vit (7), Bel-lerophonte (8), ex-cipere (12). Kissel sug-
gests that the monotony of the metre reflects the sound of Neobule’s
loom, for which see Horsfall on Virg. Aen. 7. 14; but the metre comes
from Alcaeus (where there is nothing about weaving in the surviving
fragment), and in Horace’s ode Neobule has given up weaving. RN
thinks the unremitting onset of Horace’s lines may suit the verve of
Hebrus, the true subject of the poem.
Following Bentley and most editors we have posited four three-line
stanzas with lines of 4 þ 4 þ 2 ionic metra. Alternatively one could
consider stanzas with lines of 4 þ 3 þ 3 metra (Cavarzere op. cit. 229 n.),
which suits best the limited evidence from Alcaeus (R. Führer, NAG phil.-
hist. Kl. 6, 1976: 253 ff.); this also avoids word-breaks at line-endings, but if
we begin new lines at metuentis (2) and Neobule (5), that impairs the typical
pattern of patruae verbera linguae and Liparaei nitor Hebri, where adjective
and noun frame the line. K. E. Bohnenkamp, following the Latin metrical
writers, posits lines of 3 þ 3 þ 4 metra. (Die Horazische Strophe, 1972:
88 ff.); this produces word-breaks at line-endings (ex-animari, la-vit,
Bel-lerophonte). Heinze posits a single four-line stanza, in which each
line has 10 metra; this suits the principle (observed by Meineke and
Lachmann) that the norm for a Horatian stanza is four lines, not three,
though the Lesbian poets sometimes used three-line stanzas.
1. Miserarum est neque amori dare ludum: ‘wretched are the girls whose
lot it is . . . ’; for a similar movement from general to particular cf. Alpheus,
anth. Pal. 12. 18. 1 ff. T º , x IæÆ ı h ªaæ æÆØ =
PÆæ , h NE K Ø Ø Łø, followed by ŒÆd ªaæ Kªg F Nd
º æÆ . Some editors see a statement about the female sex as a whole
(cf. Agathias, anth. Pal. 5. 297 H œŁØ PŒ Ø , › E
= ÆE IÆºł
Ø
æÆ ŁºıæÆØ , Ov. her. 19. 9 ff., pointing a contrast
with the male pursuits of hunting, wrestling, riding, and drinking); but it
is not the case that in Augustan Rome all women were as restricted as the
girl here, or that they all had censorious uncles. dare ludum means ‘to allow
enjoyment to’, i.e. ‘to indulge’; cf. Plaut. Bacch. 1083 ‘nimi’ nolo desidiae ei
dare ludum’, Cic. Cael. 28 ‘datur enim concessu omnium huic aliqui ludus
aetati’, TLL 7. 2. 1791. 45 ff.
1–2. neque dulci / mala vino lavere: lavere (third conjugation) is an old
form (‘antiqua declinatione’ Porph.) that is particularly used in meta-
phorical contexts (2. 3. 18 ‘villa . . . quam Tiberis lavit’, cf. Fronto p. 58 van
den Hout). Though it is supported here by Porph.’s comment and by
several testimonia (GL 6. 129, 169, 303, 387), the use of the simple verb for
‘to wash away troubles’ is surprising, especially as lavit is found in a literal
sense in v. 7; the word is more natural at Ter. Phorm. 973 ‘venias nunc
precibus lautum peccatum tuum?’ (where the supine lautum means ‘to
cleanse’), but there are no close parallels to our present case till late Latin
(cf. TLL 7. 2. 1052. 74 ff. for the meaning ‘abluendo tollere’). As early
corruption cannot be excluded, one might consider Withof ’s conjecture
eluere (cf. 4. 12. 19 f. ‘(cadus) amaraque / curarum eluere efficax’, Sen. dial.
9. 17. 8); another possibility is abluere (cf. Cic. Tusc. 4. 60 ‘omnis . . .
perturbatio animi placatione abluatur’). As the passage was of interest to
writers on metre, they might have got rid of the elision in the interest of
simplicity; for this kind of corruption cf. N–H on 1. 8. 2.
The lover’s sorrows were traditionally drowned in wine; cf. epod. 11.
11 ff., Tib. 1. 5. 37 f. with Murgatroyd, Prop. 3. 17. 6 ‘tu vitium ex animo
dilue, Bacche, meo’. In earlier days Roman women were not supposed
to drink wine, though there was no such restriction in Greece (Pasquali
92 f.), and according to Cato their male relatives kissed them to detect
breaches of the rule (Plin. nat. hist. 14. 90); Augustus followed the
ancient custom by denying wine to the banished Julia (Suet. Aug. 65. 3);
see further Polyb. 6. 11a. 4 with Walbank, Plut. quaest. Rom. 6, Gell.
10. 23. 1, Blümner (1911), 365. dulce like is a conventional epithet
referring to the pleasurable quality of wine; elsewhere it distinguishes
sweet from dry wines (Plin. nat. hist. 14. 63).
2–3. aut exanimari metuentis / patruae verbera linguae: aut is elliptical

and implies ‘or else’ (cf. 3. 24. 24, Plaut. Pseud. 995 ‘necesse hodie Sicyoni
me esse aut cras mortem exsequi’, OLD s. v. 7); the construction is also
found in Greek (cf. Plut. Sol. 21. 1 HÆ b ŒÆŒH ºªØ KŒºı . . . j
æE æÆ
a . . . IØ, LSJ X A3). exanimari perhaps refers not just
to fainting (Ter. Andr. 251 ‘me miseram exanimavit metu’) but hyperbol-
ically to dying (cf. 2. 17. 1). The paternal uncle, unlike the avunculus or
maternal uncle, was proverbially censorious; cf. serm. 2. 2. 97, 2. 3. 88 ‘ne
sis patruus mihi’, Catull. 74. 1, Cic. Cael. 11 ‘fuit in hac causa pertristis
quidam patruus censor magister’, Pers. 1. 11 with Kissel. In Rome agnate
relatives (i.e. on the male side) had an interest in preserving the family
fortune, and so were regularly appointed tutores or guardians (Crook
113 ff.); the position was different in Greece (Pasquali 89 f.), though we
are told that Philemon wrote about a patruus obiurgator (Apul. flor. 16a).
patruus as an adjective is very rare when textual corruptions are excluded
(TLL 10. 1. 794. 33 ff.). For ‘tongue-lashings’ cf. Cic. rep. 1. 9 ‘neque liberi
(esse) . . . contumeliarum verbera subire’, OLD 3b; verberare is often so
used, sometimes in combination with verbis (OLD 1b).
4. tibi qualum Cythereae puer ales, tibi telas: the qualum or quasillum
was a work-basket (calathus) that held wool for spinning (D–S 1. 812
with an illustration, Hilgers 42 ff., 128 f.); for the virtuousness of such
activities cf. 3. 15. 13–14 n. Cythera is the island off the south coast of the
Peloponnese where Aphrodite came ashore after her birth. The grandi-
ose periphrasis for Cupid (‘The winged son of the Cytherean’) is faintly
ironical (at Prop. 2. 30B. 31 he is simply called Ales); for similar Greek
epithets cf. Bruchmann 115. The presentation of Eros as a mischievous
boy is Hellenistic; for a lost wall-painting from Pompeii in which he
steals a basket from Leda see W. Helbig, Wandgemälde der von Vesuv
verschütteten Städte Campaniens, 1868: 43, and for a similar theft K.Sche-
fold, Die Wände Pompeiis, 1957: 209. Cythereae is contrasted with Miner-
vae below, and so is placed before puer ales rather than in the middle; for
this common opposition cf. anon. anth. Pal. 6. 48, 6. 283, Tarán (1979),
115 ff. For love as a distraction from domestic tasks cf. Sappho 102 (cited
in the introduction), Sen. Phaedr. 103 ‘Palladis telae vacant’; for the
neglect of other activities cf. Theoc. 11. 72 ff., Virg. ecl. 2. 70 ff., Aen.
4. 86 ff., ciris 177 ff., Longus 1. 13. 6.
We combine tibi telas with tibi qualum (aufert Cupido); most editors
combine it with operosaeque Minervae studium (aufert Hebrus). The
former interpretation suits the line-division suggested in the introduc-
tion, and also keeps the clauses more evenly balanced. It may be
objected that Cupid would find it difficult to steal the web from the
loom, but such thefts could be implausibly outrageous (1. 10. 11). aufert
with studium is not referring to theft but simply to distraction; it
therefore readily admits a second subject. que (5) need not join telas
and studium; it can link the two clauses just as easily.
4–5. operosaeque Minervae studium aufert, Neobule: Minerva was an
old Italian deity who acquired some of the characteristics of Athena,
including the patronage of both spinning and weaving (Bömer on Ov.
met. 4. 33 f., Simon 168 ff.); her temple on the Aventine was a centre for
artisans (L. Richardson 254). operosa represents Kæª or Kæª ,
epithets of Athena (Bruchmann 8); the objective genitive with studium
(OLD 2) is made easier because minervae can be used as a common noun
for spinning and weaving (Murgatroyd on Tib. 2. 1. 65). The only
Neobule otherwise attested is the woman derided by Archilochus (fr.
171 with West’s testimonia), but nothing in the surviving fragments
points to an allusion here. The name suggests that she has ‘new designs’.
6. Liparaei nitor Hebri: the form of the phrase mirrors Cythereae puer
ales: the power of Cupid is manifested in Hebrus. The focus now moves
from the girl to the young man, who is described in a series of clauses
that continue to the end of the poem. nitor refers to shining beauty (cf.
1. 19. 5 ‘urit me Glycerae nitor’), here enhanced because athletes rubbed
themselves with olive oil (‘unctos . . . umeros’ in v. 7); cf. Cic. div. 1. 22
‘nitidoque Lyceo’ with Pease, Ov. fast. 5. 667 ‘nitida . . . palaestra’ with
Bömer, Theoc. 2. 79 with Gow (Simaetha admires the ŁÆ ºÆ
of Delphis as he leaves the wrestling-school), Call. Hec. 71. 3 with
Hollis. Hebrus is said to come from Lipari (off the north coast of Sicily)
because ºØÆæ is used of the sheen of oil (for such verbal play see
O’Hara (1996), Paschalis (1997)); cf. Hom. Od. 15. 332, Ar. nub. 1002
ºØÆæ ª ŒÆd PÆŁc K ªıÆØ ØÆæłØ , Theoc. 2. 51
ºØÆæA . . . ÆºÆæÆ with Gow.
‘Hebrus’ is properly the great Thracian river, and as such balances
Tiberinis below; it is attested as a man’s name in a few inscriptions
(LGPN 2. 138, 3A. 137), as an early Virgilian scholar (RE 7. 2589 f.), and
in epic (Virg. Aen. 10. 696, Val. Fl. 3. 149, 6. 618); fictitious characters
are sometimes named after rivers (3. 7. 23 ‘Enipeus’, also a swimmer
in the Tiber (ibid. 27 f.), C. Saunders, TAPA 71, 1940: 544 f., Dewar on
Stat. Theb. 9. 152). One would like to find a specific explanation here:
thus M. Treu (WJA 4, 1949–50: 224 f.) follows H. Fränkel in seeing an
allusion to Alcaeus 45 L–P; there the Hebrus is described as the fairest of
rivers, and its water, where the girls bathed, is compared to unguent (7 f.
e e ‰ ¼ºØÆ = Łœ høæ with Page, Sappho and Alcaeus, 286 ff.).
One might also point to the notorious chilliness of the river (epist. 1. 3. 3,
1. 16. 13, RE 7. 2588 f.); that might suggest that the young man is cold and
unresponsive; cf. Sithonia nive in 3. 26. 10 (also Thracian).
7. simul unctos Tiberinis umeros lavit in undis: for swimming in the

Tiber cf. 1. 8. 8 with N–H, 3. 7. 27 f., Cic. Cael. 36, Griffin (1985), 89 ff., RE
Suppl. 5. 847 ff.; for other passages where an athlete impresses women
cf. 3. 7. 25 ff., Tib. 1. 4. 11 f. (again riding and swimming), Ov. her. 18. 95 f.
(Leander) ‘nunc etiam nando dominae placuisse laboro, / atque oculis
iacto bracchia nostra tuis’. The bathing follows the anointing, which in its
turn follows the riding, boxing, and running, olive oil being the equiva-
lent of soap; for this sequence of events cf. serm. 2. 1. 7 f. ‘ter uncti /
transnanto Tiberim’, Ov. trist. 3. 12. 21 f., Veg. mil. 1. 3. 4, 1. 10. 3 ‘Romani
veteres . . . Campum Martium vicinum Tiberi delegerunt in quo iuventus
post exercitium armorum sudorem pulveremque dilueret ac lassitudinem
cursus natandi labore deponeret’. In our passage the events are not
mentioned in chronological order, any more than in 1. 8. 3 ff.
8. eques ipso melior Bellerophonte: eques is in apposition to Hebrus,

the subject (understood) of lavit. With Athena’s help Bellerophon
succeeded in breaking in and riding the winged horse Pegasus; see 4.
11. 27 f., Pind.O. 13. 84 ff., I. 7. 44 ff., LIMC 7. 1. 221 ff., 7. 2. 152 ff.
Bellerophonte is the ablative of Bellerophontes (not Bellerophon); metre
requires that the final vowel should be long. The accusative is probably
Bellerophonten at 4. 11. 28 (Bentley for Bellerophontem), the dative Beller-
ophontae at 3. 7. 15. For the declension of such names see Housman 2. 829
and 833, M. Leumann, MH 2, 1945: 237 ff.
8–9. neque pugno / neque segni pede victus: vincere is sometimes

combined with an instrumental ablative (Ov. met. 1. 448 f. ‘quicumque
manu pedibusve rotave / vicerat’). The phrase here would have to mean
‘defeated because of slowness of fist or foot’ (ibid. 1. 544 ‘victa labore
fugae’), but unfortunately it might more naturally be taken as ‘surpassed
in slowness’ (cf. Plaut. Poen. 532 ‘vicistis cochleam tarditudine’). RN
(PCPS suppl. 15, 1989: 89 ¼ Collected Papers 263 f.) has considered
emending segni to a proper name like Cycni (in late Latin the spelling
cigni would be close, particularly if c was already pronounced like s); the
genitive would then be Ie ŒØF with pugno and pede. Cycnus is
mentioned along with Bellerophon as a participant in the games for
Pelias (Hyg. fab. 273. 11 ‘Cycnus Martis filius armis occidit Pilum
Diodoti filium, Bellerophontes vicit equo’, perhaps from Stesichorus);
but nothing is said there about boxing or running.
10–11. catus idem per apertum fugientis agitato / grege cervos iaculari:
Porph. comments ‘catus acutus et per hoc doctus’ (cf. Varro, lL 7. 46);
the word refers to practical skills at 1. 10. 3, Cic. Arat. 304 ‘tornare cate
contortos . . . orbis’, Auson. epigramm. 104. 1 ‘doctus Hylas caestu, Phe-
geus catus arte palaestrae’; it is somewhat archaic in tone, and is avoided
by Virgil, Ovid, Lucan, Valerius, Silius, and Statius; for the following
infinitive cf. 1. 29. 9 ‘doctus sagittas tendere Sericas’, N–H on 1. 1. 8, Bo
268. Bentley considered transposing catus with celer (11), and though he
1 3 . O F O N S BA N DV S I A E 171
dropped the idea it is worth noting: shooting at a moving target requires
speed (cf. Virg. Aen. 9. 178 ‘iaculo celerem levibusque sagittis’), whereas
tracking a hidden boar calls for cunning (see, however, 11–12 n.). For
agitare of driving animals in the hunt cf. 2. 13. 40.
11–12. et celer arto latitantem / fruticeto excipere aprum: there is

substantial manuscript support for the omission of et, but the asyndeton
is unconvincing; see above on 1–2 for a possible attempt by writers on
metre to get rid of elision in this poem. celer excipere balances catus
iaculari; for Bentley’s doubts see above. celer seems better suited to an
infinitive like exigere ‘to drive out’, but Horace may be telescoping
events for the sake of vividness: the boar lurks in the thicket, and
when it bursts out a quick reaction is needed.
latitantem makes a contrast with fugientis, and arto fruticeto with per
apertum; for the haunts of boars cf. Xen. cyn. 10. 5, Ov. met. 8. 334 ff. arto
goes well with both fruticeto and latitantem (cf. Hom. Il. 11. 118, Od. 19.
439 K º
ıŒØfi Ð ŒÆŒØ ªÆ F , Ov. met. 1. 122 ‘densi frutices’);
it should be preferred to the variant alto, a commoner but here less
precise word (cf. Hom. Il. 11. 415 ÆŁ KŒ ıº
Ø, Ov. her. 4. 170,
Sen. Ag. 892). excipere is the vox propria for receiving an animal’s onset;
cf. Virg. ecl. 3. 18, Sen. dial. 3. 11. 2 ‘(venator) venientis excipit et fugientis
persequitur’, OLD 13. aprum makes an effective climax after latitantem
has aroused our curiosity; for boar-hunting cf. 1. 1. 28 with N–H, epod. 2.
32, epist. 1. 18. 49 ff., Anderson (1985), index.
1 3 . O F O N S BAN DV S I A E
[ L. and P. Brind’Amour, Phoenix 27, 1973: 276 ff.; F. Cairns, AC 46, 1977: 523 ff.; Commager
322 ff.; Fraenkel 202 ff.; R. Hexter in Homo Viator (ed. M. Whitby et al.), 1987: 131 ff.; M. R.
Lefkowitz, CJ 58, 1962: 63 ff.; G. Nussbaum, Phoenix 25, 1971: 151 ff. and ANRW 31. 3.
2133 ff.; Oliensis 98 ff.; Pasquali 553 ff.; K. Quinn, Latin Explorations, 1963: 75 ff.; Ernst
A. Schmidt, Antike und Abendland 23, 1977: 105 ff. ¼ Zeit und Form, 2002: 131 ff.; D. R.
Smith, Latomus 35, 1976: 822 ff.; D. W. T. Vessey in Studies in Latin Literature and Roman
History (ed.C. Deroux) 4, 1986: 383 ff.; R. Westman, Classica et Mediaevalia, Diss. 9, 1973:
301 ff.; Williams 673 ff.; J. R. Wilson, CJ 63, 1967–8: 289 ff.]
1–8. Spring of Bandusia, more glittering than glass, tomorrow you will
receive the sacrifice of a kid. 9–12. You provide welcome coolness for oxen and
goats. 13–16. You will be counted among the famous springs because of my
poem.
Water was a mysterious substance in the Graeco-Roman world, and

could never be taken for granted. Hippocrates discussed its healing
properties and its effects on various localities (Airs, Waters, and
Places, Loeb edn. vol. 1); Roman agricultural writers stressed the neces-
sity of irrigation (cf. Horden–Purcell 238 ff.); Vitruvius (8. 3), Strabo (3.
5. 7), Seneca (nat. quaest. 3), and Pliny (nat. hist. 2. 227–32) recorded
miscellaneous lore about unusual sources; Frontinus savoured the qual-
ities of the different aqueducts that meant so much for urban civiliza-
tion (aqu. 1. 13–15, 2. 89–92). In a hot climate where water was scarce and
most rain fell in the winter the perennial spring had a particular appeal,
and figured regularly in descriptions of beauty-spots; cf. serm. 2. 6. 2
(quoted below), Hom. Od. 17. 205 ff., Sappho 2 L–P, Plat. Phaedr.
230b, Schönbeck (1962), 21 ff., N–H on 2. 3. 12 and vol. 2, pp. 52 ff.,
below 14–15 n.
In these circumstances springs played an important part in cult:
cf. Serv. Aen. 7. 84 ‘nullus enim fons non sacer’, N–H on 1. 1. 22, Prop.
1. 18. 27 ‘divini fontes’ with Fedeli, Horden–Purcell 412 f. Water was
connected with prophecy and inspiration (M. Ninck, Die Bedeutung des
Wassers im Kult und Leben der Alten, Philol. suppl. 14, 1921; repr. 1960,
Kambylis 23 ff.); the Muses were associated with Hippocrene and the
rest (T. R. Glover, Springs of Hellas, 1945, ch. 1), and the Latin Camenae
were originally water-nymphs (3. 4. 25 n.). At Rome there was an
ancient deity called variously Fons, Fontus, Fontanus (ILS 3882–9,
Wissowa 221 f., RE 6. 2838 ff.); he had an altar on the right bank of
the Tiber (Cic. leg. 2. 56) and a temple founded in 231 bc (Steinby 2.
256 f., Pease on Cic. nat. deor. 3. 52). This feeling of reverence persisted
long after the period of classical antiquity; churches were built by
ancient wells, and even today people still throw coins into fountains—
an unconscious survival of primitive animism.
The whereabouts of Bandusia has been disputed. Since the name is
too unusual to be a fiction (contrast Greek epigrams, which do not
specify their springs), it is reasonable to assume that it was situated near
Horace’s Sabine villa (thus ps.-Acro on 3. 13. 1 and Porph. on epist. 1. 16.
12); substantial support is provided by serm. 2. 6. 2 ‘tecto vicinus iugis
aquae fons’ and epist. 1. 16. 12 ff. ‘fons etiam rivo dare nomen idoneus, ut
nec / frigidior Thracam nec purior ambiat Hebrus’ (for the ‘Fonte
Oratina’ see the map in Encicl. oraz. 1. 254). It is true that a papal bull
of 1103 mentions ‘ecclesia sanctorum martyrum Gervasii and Protasii in
Bandusino fonte apud Venusiam’ together with a neighbouring ‘castel-
lum Bandusii’ (P. Jaffè, Regesta Pontificum, edn. 2, 1881–5: no. 5945); for
modern attempts to identify the site at San Gervasio, 7 miles south of
Venosa, see B. Capmartin de Chaupy, Découverte de la maison de cam-
pagne d’Horace 3, 1769: 536 ff., E. T. Ramage, The Nooks and Byways of
Italy, 1868: 210, N. Douglas, Old Calabria, 1915: ch. 7. ‘Bandusia’ looks
like a south Italian corruption of the Greek Pandosia, ‘giver of every-
thing’ (cf. Skutsch on Enn. ann. 166, Burrus for Pyrrhus); Pandosia was
the name of several places in S. Italy (RE 18. 3. 549 ff.), and is paralleled
by similar formations (Venusia, Canusium, Brundisium), whereas in
Latium a Greek name would be unexpected, and the intervocalic
s should have changed to r (see further F. Ribezzo, Rivista Indo-
Greco-Italica 21, 1937: 63 f., 93 f.). The best way of accounting for this
evidence is to assume that Horace gave to a Sabine spring the name of a
famous landmark near his birthplace; for this practice cf. Cicero’s
‘Academia’ and ‘Lyceum’ at Tusculum (Tusc. 2. 9, div. 1. 8), Augustus’
‘Syracuse’ in his Palatine palace (Suet. Aug. 72. 2), Andromache’s ‘falsi
Simoentis’ in Epirus (Virg. Aen. 3. 302), W. Görler, ‘Sentimentale
Namengebung’ in Pratum Saraviense (ed. W. Görler and S. Koster),
Palingenesia 30, 1990: 169 ff.
The professed occasion of the ode is uncertain. When Horace says
that a kid will be offered to the spring ‘tomorrow’ (3), he seems to be
thinking of a particular festival. As the poem is addressed to a fons, most
commentators refer to the Fontinalia on 13 October (Latte 77). Varro
says garlands were offered to Fons on that day (lL 6. 22: ‘(Fontinalibus)
in fontes coronas iaciunt et puteos coronant’), but in mid-October
coronae were probably made of leaves and other greenery rather than
flowers (2 n.); again, the libation of wine need not be connected with the
October vintage. For blood-sacrifices to Fons cf. Ov. fast. 3. 300 (Numa
offers a sheep on 1 March), Acta Fratrum Arvalium (ed. G. Henzen,
1874, p. 146), ‘Fonti verbeces II’ (on 8 Feb. and 7 Nov.); but a sacrifice on
the Fontinalia is not attested. More importantly, this theory does not
allow the most direct and immediate interpretation of the reflected
sunlight (1), the ‘burning Dog-Star’, which rose on 18 July (9 n.), and
the shade provided for ploughing-oxen (10–12 n.). Cairns, however, sees
these as representing the most striking features of the spring, regardless
of date.
There is also a further complication. The usual pattern of goat-
breeding was for the animals to mate in the autumn and for the kids
to be born in the following spring (Varr. rust. 2. 3. 8 of the nanny-goat
‘quae concepit [desistente autumno], post quartum mensem reddit
tempore verno’, Plin. nat. hist. 8. 200 ‘concipiunt Novembri mense ut
Martio pariant’, Colum. 7. 6. 6). As a kid’s horns begin to bulge a month
after birth, and have normally broken through after two, Horace’s
kid (4–5) could not have been at the stage described much after the
middle of May—five months too soon for the Fontinalia. So one has
to assume a later birth. K. D. White (1970), 314 says kids could be
born twice a year; cf. D. Mackenzie, Goat Husbandry edn. 3, 1978: 218,
who refers to cases in Africa, and Virg. georg. 2. 150 who cites ‘bis
gravidae pecudes’ as one of the marvels of Italy, though the animals
are not specified. If Horace’s kid was born in early September, it would
be at the right stage for the Fontinalia in October. In any case,
some variation in the pattern must be assumed in order to account for
the sacrifice of a tener haedus on 5 December (3. 18. 5) and a possibly
older haedus on 13 February (1. 4. 11 f.); for information of this topic
we are indebted to Dr Catharine Bazeley of the Bristol University
Veterinary School.
Schmidt (op. cit. 109) sets Horace’s sacrifice at the festival of the
Camenae on 13 August; but when Horace refers to his own poetry in the
last stanza, he gives no hint that the Camenae have inspired him.
Others set Horace’s sacrifice at the Neptunalia on 23 July (thus the
Brind’Amours, op.cit.); this festival was originally concerned not with
the sea but with springs, whose continuing existence in midsummer
occasioned popular rejoicing (see introduction to 3. 28). The date fits
the references to sunlight, heat, and leafy shelter mentioned above; it
also suits the ploughing-oxen (10–12 n.). The sacrifice to Neptune
was regularly a bull (RE 16. 2. 2520), but so grand an offering was
appropriate only to public occasions. As with the Fontinalia, the date
finds no support in the statements of Varro, Pliny, and Columella;
one notes, however, that Hesiod (op. 592) enjoys the flesh of young
kids (æøª ø Kæø) ‘when Sirius parches head and knees’ (587).
It is true that if Horace was thinking of the Neptunalia one might
have expected him to mention it, as he does in 3. 28, but he may have
thought that the poetic details referred to above gave a sufficient
indication of the season.To sum up: if one is prepared to rely on
Hesiod to fill the gap in our knowledge of Italian goat-breeding, the
Neptunalia meets the other criteria and seems on balance the most
attractive hypothesis.
The description of the sacrifice has provoked disagreement of a
different kind. When Horace contrasts the glittering spring with the
stain of the victim’s blood (inficiet in v. 6), Campbell (1924: 2) comments
‘who wants a drink after that?’ But Horace says nothing about drinking,
and the blood would soon be washed away. Fraenkel sees only sympathy
for the kid ‘warm, pretty, and amusing in its youthful pranks, which
only too soon will come to an end’ (203, comparing 4. 2. 54 ff.). David
West speaks of ‘a complex stimulus, the life spurting from the animal’s
jugular, an ancient religious observance of your race, the promise of a
good supper, good wine and good company, perhaps with some music
and love’ (Reading Horace, 1967: 130). True, an agricultural society was
not squeamish about slaughtering animals (for the clash of attitudes cf.
Hardy’s Jude the Obscure part 1, ch. 10); and in the ancient world sacrifice
had a potency not now easily understood (see W. Burkert, Homo Necans,
English edn. 1983: 40 for the excitement generated on such occasions).
None of this primitive emotion is apparent in Horace’s poem, but some
moderns find a disconcerting element in the aesthetic contrast of red
blood and cool water (cf. J. Griffin, JRS 87, 1997: 63); this impression is
not removed by observing that the sacrifice lies in the future (Smith and
Vessey, opp. citt.). It is perhaps worth adding that in presenting the
natural scene Horace, unlike a romantic poet, says nothing directly
about his feelings.
Formally the ode has features of a hymn, even if such a hymn could
not figure in the ceremony envisaged: here we may note the opening
address (1 n., 2 n.), the praise of the spring’s virtues (9–12), the anaphora
of tu . . . tu . . . tuae; cf. Norden (1913), 149 ff., N–H on 1. 10. 9. But in
practice it owes more to Greek epigrams (Pasquali, op. cit.). Sometimes
these are descriptive, emphazing cool water and leafy shade (see e.g.
Anyte, anth. Pal. 9. 313, anth. Plan. 228, Gutzwiller 68 ff.); rocks and a
solitary tree may also be mentioned (14–15 n.). Dedicatory epigrams are
another influence (Cairns, op. cit.); here we may compare Leonidas,
anth. Pal. 6. 334 and 9. 326, Crinagoras, ibid. 6. 253 (all illustrating
natural features). Yet it would be wrong to characterize Horace’s very
original lyric too schematically; the offering is promised more indirectly
than is usual in an epigram (3 n.), the conventional themes are attached
to a real place in the Italian countryside and are associated with a real
Italian ritual, while the final stanza displays a different and deeper
purpose.
Although the concluding vignette matches the opening line, Horace
now reveals that he is offering more than a kid: because of this very ode
Bandusia will be counted with the springs celebrated by Greek poets.
For such claims cf. Theog. 237, Ibycus, PMG 282. 47 f. ŒÆd
, —ºŒæÆ , Œº ¼ŁØ (Ø = ‰ ŒÆ IØa ŒÆd Ke Œº ,
Prop. 3. 2. 17 f., Ov. trist. 5. 14. 4 ‘tu tamen ingenio clara ferere meo’,
Shakespeare, sonnet 18. 13 f. ‘So long as men can breathe and eyes can see
/ So long lives this and this gives life to thee’ (with J. B. Leishman,
Themes and Variations in Shakespeare’s Sonnets, 1961: ch. 1). In spite of
Horace’s hymnal style, the boast could not have been made in address-
ing a deity, or for that matter a ruler (though note 4. 9. 25 ff. to Lollius);
for a similar claim with reference to a natural feature cf. Cic. leg. 1. 1 (the
oak at Arpinum described in Cicero’s Marius) ‘manet . . . et semper
manebit; sata est enim ingenio’.
Most critics see that the ode is more than a pretty nature-poem, but
some of their efforts to formulate a more profound interpretation only
muddy the waters. One says that ‘as [the] warm blood mingles with the
lucid water it is easy to sense a suggestion of the transformation of life
into art’; another, noting that frons and cornua were parts of a book-roll,
thinks that ‘the death of the kid . . . would be no less the death of poetic
overreaching in epic vein, a kind of recusatio under the veil of metaphor’;
another believes that ‘the sacrifice of the kid represents the sacrifice of
the living individual that poetry itself demands’; another, commenting
on me dicente (14) writes ‘Horace’s perfectly self-reflexive participial
construction thus harbors its own semantic turgidity, like the kid’s brow
and like the spring itself, forever brimming over into the future’. Such
notions remind us that to be open and receptive is not the critic’s only
business.
Metre: two Asclepiads followed by a Pherecratean and a Glyconic.
1. O fons Bandusiae, splendidior vitro: the formal o is not generally

used with Latin vocatives but here suits the hymnal style. The genitive
gives the name of the spring (for which see the introduction), not of a
nymph, who would have been addressed directly, nor yet of a district
(thus ps.-Acro, but that would be too colourless); for the construction,
where Cicero would have said fons Bandusia, cf. Virg. Aen. 1. 247 ‘urbem
Patavi’, H–Sz 62 f. splendidior means ‘more glittering’, and refers to the
reflection of the sunlight; cf. Call. Hec. fr. 18. 2 Hollis (¼ 238. 16 ff.)
æÆ ºØ Ææ PæÆe qł, Ov. met. 13. 791 9 (to
Galatea) ‘splendidior vitro’. More usually glass suggests translucence,
sometimes with a hint of green (N–H on 1. 17. 20); cf. 1. 18. 16 ‘perlu-
cidior vitro’, 4. 2. 3 f. ‘vitreo . . . ponto’, Virg. Aen. 7. 759 ‘vitrea . . . unda’
(with Horsfall), Plin. epist. 8. 8. 2 ‘purus et vitreus’ (on the source of the
Clitumnus); see further M. L. Trowbridge, ‘Philological Studies in
Ancient Glass’ (Univ. of Illinois Stud. in Lang. and Lit. 13, 1928: 59 ff.).
2. dulci digne mero non sine floribus: ‘unmixed wine’ was used in
libations and other ritual offerings; ‘sweet’ is a conventional epithet
applied to wine in general, but here it may emphasize literal sweetness
(3. 12. 1–2 n.). digne suits the religious language (3. 21. 6); the alliterative
dulci digne (balanced by donaberis below) helps to suggest an archaic rite.
For the offering of wine to springs cf. Longus 4. 32. 3 (of Chloe) KŒæÆ
b ŒÆd c ªc Yfiø, Schmidt, op. cit. 108 n. For the litotes non sine cf.
3. 7. 7 n. Flowers suit summer better than autumn, unless we are
supposed to see a contrast between the usual offerings and tomorrow’s
special sacrifice.
3. cras donaberis haedo: a real dedication would normally use a ‘per-

formative’ present tense (‘I hereby give’), unless where the offering
depended on a reciprocal benefit (Virg. ecl. 7. 31 f., Petr. 85. 6); here a
promise is made on the eve of the celebration (cf. 3. 17. 14, epist. 1. 5. 9 f.).
Similarly the passive verb detaches the lyric from the usual dedicatory
formulae and puts the emphasis on the spring rather than on the poet,
whose most important offering turns out to be, not the kid, but the ode
itself. For blood-sacrifices at springs see the introduction; add Hom. Od.
17. 240 ff. (Odysseus offers sheep and kids to the nymphs), Mart. 6. 47. 5,
P. Stengel, Die griechischen Kultusaltertümer, edn. 3, 1920: 135 ff.
4–5. cui frons turgida cornibus / primis et Venerem et proelia destinat:
the emphatic primis means ‘earliest’ (cf. Sen. Tro. 538 ‘primisque non-
dum cornibus findens cutem’), balancing destinat, not ‘tips of the
horns’ (Quinn, Williams) which are not yet evident; for the picture
cf. Lucr. 5. 1034 f. (of a bull calf ) ‘cornua nata prius vitulo quam
frontibus extent, / illis iratus petit atque infestus inurget’, Sen. Hf 142
with Billerbeck, Mart. 6. 38. 8, Juv. 12. 9, Galen, de usu partium 1. 3. 6
(p. 4 Helmreich).
For ‘Venus’ applied to animals cf. 2. 5. 3 (with N–H vol. 2, p. 78), Virg.
georg. 2. 329 ‘et Venerem certis repetunt armenta diebus’, 3. 210, OLD 4b;
horns are symbols of aggressive masculinity, for they are used to repel
sexual rivals (Onians 238 ff.). proelia is smilingly grandiloquent in this
context; cf. Virg. georg. 3. 219 f. ‘pascitur in magna Sila formosa iuvenca:
/ illi alternantes multa vi proelia miscent’ and 4. 5.
6–7. frustra; nam gelidos inficiet tibi / rubro sanguine rivos: for the
detached frustra followed by nam cf. 3. 7. 21 n., E. Wölfflin, ALL 2. 10 f.
inficiet (‘discolour’) is set against the purity of splendidior (1); for a similar
prediction cf. Theoc. epig. 1. 5 f. (¼ anth. Pal. 6. 336. 5 f.) øe ,
ÆƒØ ŒæÆe æª y › Æº , = æŁı æªø
Æ
IŒæ Æ (as in our poem there is a contrast with the living animal).
gelidos implies that the blood is warm, rubro that the spring is clear
(Kiessling ad loc., Bell 341, Schmidt 344 ff.). RN thinks that tibi has
some emphasis as part of the sacral anaphora at 9 ff., NR prefers to take
it as enclitic with inficiet. For the plural rivos (ÞŁæÆ) cf. Tib. 1. 1. 28.
8. lascivi suboles gregis: lascivi picks up Venerem; as applied to goats it

means not only ‘frisky’ but ‘lustful’ (OLD 4). suboles [sub þ alo] is used of
an animal that replenishes the stock; according to Cicero (de orat. 3. 153),
the word has an archaic tinge and suits poetry (Tusc. 2. 23 ‘Titanum
suboles’) rather than oratory (yet cf. Marc. 23 ‘propaganda suboles’).
9–10. te flagrantis atrox hora Caniculae / nescit tangere: for the em-
phatic te see the introduction. Canicula is Sirius, the Dog-Star (Hom.Il.
22. 26 ff., Hes. op. 417 ff. with West, 587, N–H on 1. 17. 17, serm. 1. 7. 25 f.,
anon. anth. Pal. 10. 12. 7 Ø b ıª OøæØF Œıe pŁÆ, Le
Boeuffle 134 ff.); its ‘rising’ on 18 July (Plin. nat. hist. 2. 123), i.e. the date
when this came closest before sunrise (3. 1. 27 n.), coincided with the
hottest time of the year, and its invisible presence was supposed to have
caused the heat (3. 1. 32 n., Plin. nat. hist. 18. 269 f.); for the ‘Dog-Days’ see
further B. Blackburn and L. Holford-Strevens, The Oxford Companion to
the Year, 1999: 595 f. hora here means ‘season’, a Greek usage not attested in
Latin before Horace; cf. 1. 12. 16 with N–H, ars 302 ‘sub verni temporis
horam’, TLL 6. 3. 2964. 1 ff. (for the Greek cf. Alc. 347. 2 L–P I þæÆ
ÆºÆ, Opp. hal. 3. 47 f. łØ ‰æ = &Øæı). atrox, a strong word

sometimes used of weather (OLD 1b), also suits a savage dog (OLD 5); cf.
epist. 1. 10. 16 ‘rabiem Canis’, Manil. 1. 396 f., Pers. 3. 5. nescit, although
like atrox grammatically attached to hora, suggests a sentient creature that
is at a loss how to hurt. The euphemistic tangere suits both climatic
afflictions (OLD 4) and dog-bites.
10–12. tu frigus amabile / fessis vomere tauris / praebes et pecori vago:

for the picture cf. 3. 29. 21 ff., Virg. ecl. 2. 8 ‘nunc etiam pecudes umbras
et frigora captant’. frigus is set against flagrantis, and amabilis against
atrox; as the noun often has unpleasant connotations, there is an
element of paradox here. The agricultural writers distinguished vomer
‘ploughshare’ from aratrum ‘plough’, e.g. Cato, agr. 10. 2 ‘aratra cum
vomeribus’, Varro, rust. 1. 29. 3, but the poets often used vomer for
‘plough’ by synecdoche; cf. carina, ‘keel’, for ‘ship’. Ploughing is com-
patible with both summer and autumn; for the former cf. Xen. oec. 16.
14, Theophr. caus. plant. 3. 20. 7, Theoc. 25. 25 (with Gow), Varro, rust. 1.
32. 1, Colum. 11. 2. 54; this ploughing was supposed to be completed
between the solstice and the Dog-Star. The normal word for ploughing
oxen was boves (epod. 2. 3 ‘bobus exercet suis’), the uncastrated taurus
being used for breeding, but poets sometimes ignored the distinction; cf.
Virg. georg. 1. 45 (with Mynors), Ov. fast. 1. 698. pecori vago here refers
not to sheep (OLD 1b) but to goats (corresponding to gregis in v. 8),
which unlike the yoked oxen can range freely (cf. 1. 17. 6 f. ‘deviae /
olentis uxores mariti’, culex 104 f. ‘vagae . . . capellae’). The adjective vago
balances fessae in a chiastic pattern.
13. fies nobilium tu quoque fontium: H is thinking of such celebrated

springs as Arethusa, Castalia, and Hippocrene; but whereas they in-
spired poets and made them famous, here it is the poet who confers the
fame (see the introduction). For this type of partitive genitive cf. epist. 1.
9. 13 ‘scribe tui gregis hunc’, Plaut. mil. 1016 ‘cedo signum si harunc
Baccharum es’, Strab. 8. 6. 12 , ¯æØ Kd H PŒ Iø ºø,
K–S 1. 453.
14–15. me dicente cavis impositam ilicem / saxis: me is emphatic,

answering tu (13). dicente as often refers to celebration in song or poetry
(OLD 7b). For the combination of rock and water cf. Hom. Il. 9. 14 f.
Œæ ºıæ , = l ŒÆ ÆNªºØ æ æe
Ø oøæ,
Leonidas, anth. Pal. 9. 326. 1, Theoc. 1. 8, 7. 136 f., Ap. Rhod. 3. 227,
Ov. fast. 3. 295 ff. (also with an ilex), T. B. L. Webster, Hellenistic Poetry
and Art, 1964: 164. The rocks are described as hollow because a grotto
was regarded as picturesque (Bömer on Ov. fast. 2. 315), perhaps
also to suggest an echo. The solitary tree makes a conspicuous vertical
1 4 . H ERC V LI S R I T V 179
feature (as at 3. 22. 5); for its position by the spring cf. Plat. Phaedr. 230b,
Leonidas, anth. Pal. 6. 334. 2, anth. Plan. 230. 4 ff. aæ Œfi Æ Øfi Æ
ıœ = æØ ŒºÆæ KßŒæı Øa æ = AÆ, ´æØÆ
łı
æ æ Ø , Theoc. 1. 1 f. The ilex or holm-oak (instead of the
conventional plane or pine) is an Italian touch (RE 5. 2058 ff., Abbe 87 f.,
Meiggs 45, 218); it may be significant that the tree is evergreen and long-
lived (Lefkowitz, op. cit., Plin. nat. hist. 16. 237 ‘vetustior autem urbe in
Vaticano ilex’). Rocks, trees, water, and goats are regular components of
the so-called ‘sacral-idyllic landscape’; see E. W. Leach, Virgil’s Eclogues:
Landscapes of Experience, 1974: 86 ff., R. Ling, Roman Painting, 1991: 39
and fig. 38, 55 and fig. 55, 142 f. with fig. 153.
15–16. unde loquaces / lymphae desiliunt tuae: loquaces is predicative

with desiliunt; the adjective is not attested for water before H. For
leaping water cf. OLD salio 3, Catull. 68. 58 ‘prosilit’, epod. 16. 47 f.
‘mella cava manant ex ilice, montibus altis / levis crepante lympha desilit
pede’—a passage very like this in its sound and in its personification of
lympha, which was in fact connected with nympha (OLD 1, serm. 1. 5. 97,
RE 17. 2. 1581 f.). The alliteration of l, qu, c, l, found also in ilicem above
and in the lines from epod. 16, has an onomatopoeic appeal; cf. Virg.
georg. 1. 108 ff. ‘ecce supercilio clivosi tramitis undam / elicit; illa cadens
raucum per levia murmur / saxa ciet’, elaborating Hom. Il. 21. 260 ff.,
L. P. Wilkinson, Golden Latin Artistry, 1963: 57. tuae in the final position
picks up tu (13) and balances me (14), bringing out the close, quasi-
personal, relationship between the poet and the object addressed. The
ode ends, as it began, with an aesthetic response to the spring’s beauty.
14. HERC VLIS RITV

[Cairns 179 ff.; M.Dyson, G&R 20, 1973: 169 ff.; Fraenkel 288 ff.; T. S. Johnson, Philologus
141, 1997: 323 ff.; D.Kienast, Chiron 1, 1971: 239 ff.; Lyne (1995), 169 ff.; R. G. M. Nisbet,
PLLS 4, 1983: 105 ff.; Pasquali 195 ff.; M. C. J. Putnam, Zeitgenosse Horaz (ed. H. Krasser
and E. A. Schmidt), 1996: 442 ff.; U. W. Scholz, WS 5, 1971: 123 ff.; E. Wistrand, Miscella-
nea Propertiana, 1977: 20 ff.]
1–12. Augustus, who was thought to be dying, is returning victorious like

Hercules from Spain. Let his wife come forth with other matrons in solemn
thanksgiving, while the children keep ritual silence. 13–16. I shall not fear
internal strife while Augustus rules the world. 17–28. Bring me old wine for
my private celebration, if a jar has escaped the Italian wars; ask Neaera to
join me, but if the doorman stops her, come away. I am less hot-blooded than
I was in the year of Philippi.
This ode celebrates the return of Augustus from Spain in the early
summer of 24 bc; he had left for Gaul in 27 and had not been back since.
In 26 he commanded in person against the Cantabrians in the moun-
tains of northern Spain (see the introduction to 3. 8 citing Syme, and the
note on 3. 8. 21). Later he fell ill and retired to Tarraco on the east coast
(Dio 53. 25. 7); the campaign of 25 against the Astures was fought by his
legates, and at the end of the year his nephew Marcellus came home
without him to marry his daughter Julia (9 n.). When Augustus finally
reached Rome, he declined a triumph (Flor. 2. 33. 53), but was voted
other honours to celebrate his recovery and return (Dio 53. 28. 3).
Though the temple of Janus had been closed at the end of 25 (Dio 53.
26. 5), the war had taken longer than expected, and the illness of
Augustus had caused grave anxiety.
The ode contains features of a conventional type of panegyric, the
celebration of a great man’s arrival (adventus): apart from descriptions in
Greek (Polyb. 16. 25, Posidonius fr. 253. 35 f. Kidd, Matthew 21: 8–11, etc.,
Pasquali, loc cit.), cf. two other odes (4. 2. 33 ff. and 4. 5), Cicero’s
account of his return from exile (Att. 4. 1. 4–5, Sest. 131, Pis. 51–2), a
couple of epigrams by Martial (7. 8, 10. 6), the panegyrici of Pliny (22–3)
and his fourth-century followers, the prescriptions of Menander Rhetor
377–88 on the epibaterios logos (speech of welcome on arrival), and the
poems by Dryden and Cowley on Charles II’s restoration. Typical
commonplaces are the comparison with Hercules (1 n.), the participa-
tion of all classes (1 n.), the hint of an IØ or ‘coming to meet’
(6 n.), the religious ceremonies (6 n.), the presence of children (10 n.),
the brightness of the great day (13–14 n.). Some features are also found
in private addresses of welcome (N–H vol. 2, p. 107, Cairns 21 ff.): the
dangers past (‘morte venalem’ in 2), the distant lands visited (3 f.), the
return to the Penates (3), the celebratory symposium (17 ff.). For adven-
tus see further A. Alföldi, Die monarchische Repräsentation im römischen
Kaiserreiche, 1970: 88 ff., T. E. V. Pearce, CQ 20, 1970: 313 ff., S. Mac-
Cormack, Historia 21, 1972: 721 ff., I. M. Le M. Du Quesnay in S. J.
Harrison, 1995: 135 ff. (with bibliography).
The ode also describes a supplicatio (cf. ‘supplice vitta’ in v. 8); this was
a national day either of prayer for the future or thanksgiving for the past
(Wissowa 423 ff., RE 4A. 1. 942 ff.), and as here (5 ff.) women played a
central part in the rituals (Liv. 27. 51. 9 ‘cum matronae amplissima veste
cum liberis . . . omni solutae metu deis immortalibus grates agerent’).
Such supplicationes were often voted on news of a victory, sometimes as a
preliminary to a triumph (L. Halkin, La Supplication d’action de grâces
chez les Romains, 1953, G. Freyburger, ANRW 16. 2. 1418 ff.), but here the
situation is different as the victory lies in the past. repetit (3) shows that
Augustus is now on his way back, iustis divis (6) implies that the gods
have delivered him safe from both war and illness (so one assumes he is
nearing home), dies festus (13) is the day of his arrival, but the nature of
the preparations suggests that by now he is known to have refused a
triumph. In determining the occasion of the poem we are not obliged to
choose between an adventus (thus e.g. Syndikus 2. 137) and a supplicatio
(Heinze, Kienast, Wistrand).
The first three stanzas celebrate the public occasion with patriotic and
religious language, e.g. Herculis ritu (1), morte venalem (2), victor (4),
operata divis (6), supplice vitta (8), sospitum (10), male nominatis (11).
Here Horace first addresses the people in a way unrealistic for a Roman
(plebs 1 n.); then, following a Callimachean technique (Pasquali 196 cites
hymns 2, 5, and 6), he announces what everybody is expected to do.
Even in this formal part of the poem there is a significant emphasis on
the private sphere: Augustus’ domestic Penates (3), Livia’s joy in her
husband (5), Octavia’s affection for her brother (7 n.), the grace and
reverence of the matrons (7 f.), the stilling of the children’s chatter
(10 ff.). The rites evoke the traditional piety of the Roman people, in
contrast to 4. 5 (the return of Augustus in 13 bc), where the libations
are poured to Augustus himself and there are already hints of the
imperial cult.
The fourth stanza occupies the centre of the poem and links the
very different sections that come before and after; for the significance
of this position see introduction to 3. 8. Horace now turns from the
public ceremonies to his personal reactions (for the same movement
cf. 1. 31. 15 ff., 4. 2. 45 ff.). The transition is marked by mihi (13) and ego
(14), which point forward, balancing non ego at the end of the poem.
At the same time Horace looks back to the patriotic themes of the
earlier part: Caesar (3) is picked up by Caesare (16) and morte (2) by mori
(15). But now the emphasis is not on Spanish victories but on the
freedom from civil strife that Augustus’ return will surely guarantee.
This reminds us that, in spite of the air of confidence conveyed by the
poets, this was a time of crisis for the Principate; cf Syme (1939), 333,
Kienast, op.cit.
In the last three stanzas Horace prepares for his private symposium in
a way that recalls the home-coming of Numida (1. 36) and Pompeius
(2. 7). The brisk instructions to a slave (i in 17, dic in 21) are set against
the hushed procession in the first part of the poem; Neaera and her
perfumed hair belong to a different world from that of the pious
matrons of 7 f. ‘decorae supplice vitta’. Scholars are worried by the
dichotomy, but Horace is contrasting the formal rites of the imperial
family with the spontaneous rejoicing of an ordinary citizen, which in
the poet’s case takes the form of a symposium; in the same way Proper-
tius hopes to watch a triumph ‘in sinu . . . nixus . . . puellae’ (3. 4. 15), and
commemorates Actium with an all-night carousal (4. 6. 85 f.). Far from
complaining about a lack of unity, we should be fascinated by the
ingenious juxtaposition of diverse elements (cf. the introduction to 3. 11
and N–H vol. 1, p. xix). What is more, Horace’s symposium contains a
serious message; for just as the public stanzas take account of family
affections, so the private stanzas contain allusions to the nation’s
troubled past: Horace sends for a wine that recalls the Italian conflicts
of 91–87 bc and has eluded the marauding bands of Spartacus (73–71 bc);
and if the glamorous Neaera is kept from coming to his party, so be it;
there are to be no rows. This may sound like no more than a wistful
statement about his changed attitude to sex, but when he says he is now
less hot-headed and pugnacious than in the year of Philippi (27–8 n.), he
is also distancing himself (not without a twinge of nostalgia) from the
political fervour of his youth; the way to a peaceful future lies in
supporting Augustus.
Metre: Sapphic.
1. Herculis ritu: the phrase should be combined with repetit (3), as ps.-
Acro saw, not with petiisse (2), as is usually done; the victorious return
from Spain is the point of the comparison (Nisbet, op. cit. 106 f.). ritu
‘in the manner of ’ keeps its original suggestion of a solemn ceremony
(cf. Suet. Vit. 10. 2 ‘ritu triumphantium’); elsewhere it is applied to ways
of life (Cic. amic. 32 ‘pecudum’), natural phenomena (3. 29. 33 ‘fluminis’),
or old precedents (serm. 2. 1. 29 ‘Lucili’).
Augustus was often compared with Hercules, the civilizer of the
world (3. 3. 9 n., 4. 5. 36, Virg. Aen. 6. 801 ff., Dio 56. 36. 4, the funeral
oration ascribed to Tiberius). On his return from the East in 29 bc he
had entered the city on 13 August, the day after the ceremonies devoted
to Hercules at the Ara Maxima; an association is presumed by Virgil in
Aen. 8. 102–305 (Binder 42 ff., 145 ff.). Similarly on this occasion orators
must have suggested an analogy between Augustus’ return from the
Cantabrian wars and Hercules’ visit to the site of Rome after his defeat
of the Spanish giant Geryon (Aen. 8. 201 ff., Apollod. 2. 5. 10 with
Frazer). See further E. Norden, RhM 54, 1899: 466 ff. ¼ Kl. Schr.
(1966), 422 ff. (on a commonplace originating with Alexander), A. R.
Anderson, HSCP 39, 1928: 44 ff., Christ (1938), 129 ff., R. Schilling, RPh
16, 1942: 31 ff., E. Doblhofer, RhM 107, 1964: 327 ff., Scholz, op. cit.
127 ff., K. Galinsky, The Herakles Theme, 1972: 136 ff.
1–2. modo dictus, o plebs, / morte venalem petiisse laurum: ‘who was
reported but lately. . . to have sought a bay-crown at the cost of his life’.
The content of the rumours is given by the emphatic morte venalem, not
by petiisse laurum (as translators often imply); dictus would not be
applied to anything so obvious as Augustus’ quest for glory, and modo
suits the illness of 25 bc better than the departure of 27. For morte
venalem cf. Pind. P. 6. 39 æÆ b ŁÆØ ŒØa Ææ , Xen.
Cyr. 3. 1. 36, Isoc. 6. 109, Virg. Aen. 5. 230, 9. 205 f. ‘istum / qui vita bene
credat emi, quo tendis, honorem’, Quint. inst. 9. 3. 71.
The vocative plebs is paralleled perhaps only at Ov. Ib. 79, and populus
and vulgus are very rare as vocatives; see J. Svennung, Anredeformen,
1958: 237 ff., 284 ff., Fraenkel 289 n. 1, Dickey 295, H–Sz 24. Roman poets
followed their Greek predecessors in addressing the citizen body,
though in real life they had no such right; cf. 3. 24. 45 ff., epod. 7. 1 ff.,
16. 15 ff. (with Fraenkel 42 ff.). Descriptions of an adventus often em-
phasize the participation of all classes; cf. Cic. Pis. 52 ‘omnes viri ac
mulieres omnis fortunae ac loci’, Ov. fast. 4. 293, trist. 4. 2. 15 ‘plebs
pia cumque pia laetetur plebe senatus’, Stat. silv. 1. 6. 44, Dio
51. 19. 2 (30 bc). It is relevant that in 24 Augustus made donations to
the urban plebs (Dio 53. 28. 1); his assumption of the tribunicia potestas in
23 represented another populist gesture (CAH edn. 2, 10. 86, otherwise
Lacey 154 ff.).
3–4. Caesar Hispana repetit Penatis / victor ab ora: Hispana gains

emphasis from the long hyperbaton and underlines the analogy with
Hercules; its juxtaposition with Caesar associates the man and the
country as in the triumphal Fasti. repetit makes a formal contrast with
the similarly placed petiisse (2).
Victor, like Invictus, was sometimes a title of Hercules (Carter 43 f.),
and he had two temples at Rome under this name (L. Richardson 188 ff.
and Steinby 3. 22 ff.). ora probably refers to the northern coast of Spain,
where Augustus had won his victories (cf. 3. 8. 21 ‘servit Hispanae vetus
hostis orae’); the word, which primarily means ‘edge’, here suggests
‘distant shores’ (epist. 1. 3. 1 ‘quibus terrarum militet oris’). It is true
that Augustus had returned from Tarraco on the east coast, where he
had been ill for much of 25 bc, but if that were the reference, ora would
strike a less heroic note.
5. unico gaudens mulier marito: the formidable Livia, a daughter of

the Livii Drusi, had married Octavian in 39, and their marriage lasted
over fifty years. unico means ‘one and only’, i.e. ‘incomparable’
(E. Dutoit, Latomus 15, 1956: 481 ff.); for its application to husbands
and wives cf. Plaut. merc. 768, Tac. ann. 15. 63 ‘unice dilectam’, Suet. Aug.
62 ‘dilexitque et probavit (Liviam) unice ac perseveranter’. unicus can
also be used of military and political leaders; cf. Catull. 29. 11 ‘imperator
unice’ (there ironic), Liv. 6. 6. 17 ‘unico imperatore’ (Camillus),
A. Cameron, Hermes 104, 1976: 156 ff. (who suggests that the phrase
belonged to imperatorial acclamations). In this passage we are perhaps
led to expect a political compliment, then by the delayed marito the
force of the adjective is shifted to domestic affection. It is wrong to
point to the Roman esteem for the univira (the woman who had
married only once); Livia had been married previously to the father
of Tiberius, who had lost her to Octavian in scandalous circumstances,
and it would be embarrassing to suggest that she was a kind of
honorary univira.
6. prodeat: ‘let her come forth’ (i.e. from the imperial palace), not ‘let
her advance in procession’; the verb suggests a formal appearance as at
Catull. 61. 96 ‘prodeas, nova nupta’. Nisbet (op. cit. 108 f.) at one time
compared obviam prodire (to go out of the city gates to meet someone,
especially at a formal adventus), but here there is no obviam and no
indication of place, and Livia’s immediate purpose is to offer sacrifice
(see next note). Even so, ‘prodeat (domo)’ balances ‘repetit Penatis’; so a
meeting in the near future is implied.
iustis operata divis: operata refers to a sacrifice; such a rite suited any
home-coming (1. 36. 1 f., 2. 7. 17 with N–H), in particular a public
adventus, and it played a central part in a supplicatio. Before the elder
Pliny the word is attested only in the -atus form, which seems originally
to have been an adjective like feriatus (see J. P. Postgate, JPh 26, 1899:
314 ff., TLL 9. 2. 690. 7 ff.); so here it means not ‘having sacrificed’
(where it would have to refer to the household gods), but ‘performing
due ritual’ without any indication of priority (cf. Porph. ‘pro prodeat et
operetur divis, id est sacrificet’).
The gods are just because they have repaid the nation’s prayers and
sacrifices; cf. Ov. her. 6. 151 f. ‘quod si quid ab alto / iustus adest votis
Iuppiter ipse meis’, trist. 4. 2. 11 f. ‘cumque bonis nuribus pro sospite
Livia nato / munera det meritis saepe datura deis’. The variant iustis . . .
sacris is equally well attested (though the scholiasts support divis); the
case would still be dative rather than ablative (cf. Liv. 1. 31. 8 ‘operatum
his sacris’, Quint. 10. 3. 13 ‘scholae adhuc operatum’, OLD s.v. operatus
2). The phrase would naturally be taken as ‘conventionally appropriate
rites’ (not ‘rites that Augustus deserved’); and this is less pointed than
iustis . . . divis. The latter reflects the ancient feeling that gods can be
held to bargains.
7. et soror cari ducis: Octavia; she had additional importance at this

time as the mother of Marcellus. One side of the manuscript tradition
reads clari, which would be more usual; cf. Cic. Marc. 30, Sen. Tro. 255
‘clari ducis’, TLL 5. 1. 2327. 48 ff. Yet the affectionate cari provides a
better parallel to unico . . . marito; for the unexpected collocation with
ducis cf. 1. 1. 2 ‘dulce decus meum’, which also combines the public and
private spheres.
7–8. et decorae / supplice vitta: decorae (‘seemly’) suggests grace as well
as propriety. The vitta was a headband reserved for matrons (Plaut. mil.
792, Tib. 1. 6. 67, Blümner 273); woollen vittae were used on religious
occasions (Bömer on Ov. fast. 3. 29 f., Headlam on Herodas 8. 11).
supplice refers not to a petition but to a thanksgiving (Sen. Hf 875 f.
‘Thebis laeta dies adest: / aras tangite, supplices’); for the application of
the adjective to things cf. OLD s.v. 2. Soph. OT 3 ƒŒæØ ŒºØØ
KØ.
9–10. virginum matres iuvenumque nuper / sospitum: iuvenum refers

to officers who have recently returned from Spain; the virgines are their
fiancées whose marriage has been delayed by the war. These categories
include Octavia’s son Marcellus and Augustus’ daughter Julia, who are
indirectly alluded to here. Marcellus was still in Spain at the end of the
campaign of 25 (Dio 53. 26. 1), but then returned to Italy after his
depositio barbae (cf. Crinagoras, anth. Pal. 6. 161. 3 f. ÆŁc æH
ŒØæ ªØÆ _ º Ææd = oø , ŒÆd łÆØ ÆEÆ ŒÆd ¼æÆ
ºÆE); this was a preliminary to his marriage to Julia, which took place
still in 25, before Augustus had returned (Dio 53. 27. 5). Livia’s son
Tiberius had also served in the war (Suet. Tib. 9. 1, Dio 53. 26. 1); so she
as well as Octavia is associated with the matres.
sospitum is more solemn than salvorum and here suggests sacral
language (cf. 1. 36. 4 ‘Hesperia sospes ab ultima’, Ov. trist. 4. 2. 11 quoted
above on v. 6). The adjective must be taken with virginum as well as
with iuvenum, otherwise it is not clear who these young women are;
their fate depends on the survival of their loved ones (cf. Plaut. Epid. 556
‘salva sum quia te esse salvom sentio’).
10–11. vos, o pueri et puellae / yiam virum expertaey: vos should be

preceded by a semi-colon; the shift to the vocative not only adds
liveliness (note o), but detaches the pueri and puellae from the official
procession of matrons. If the transmitted text is right (and it is sup-
ported by ps.-Acro), one has to assume that the girls who were virgines
when their fiancés returned at the end of 25 were now married women
(like Julia) when Augustus arrived in 24; thus W. H. Alexander, CW 36,
1943: 162 f. But pueri does not suit victorious officers (when Marcellus is
addressed as ‘miserande puer’ at Aen. 6. 882, that intentionally evokes
pathos), and the collocation with puellae seems too sentimental for
the context.
In the description of a ceremony it is more natural to take pueri et
puellae as children, who might well be reminded not to chatter. For their
presence on religious occasions cf. 1. 21. 1 f., 4. 6. 31 f., carm. saec. 6
‘virgines lectas puerosque castos’, epist. 2. 1. 132 ‘castis cum pueris ignara
puella mariti’ with Brink, Catull. 34. 2 ‘puellae et pueri integri’; note
particularly the mention of children in the context of a supplicatio (Cic.
Cat. 3. 23, Liv. 27. 51. 9 etc.) as well as their appearance on the Ara Pacis.
As iam virum expertae is now impossible, Bentley proposed non virum
expertae, which NR is inclined to accept, even though it is rather close to
virginum above, who represent a different category. J. Delz has suggested
coniugi expertes (MH 30, 1973: 53 f.) on the lines of 3. 11. 11 f. ‘nuptiarum
expers et adhuc protervo / cruda marito’; again the expression seems too
close to virginum. RN has conjectured labis expertes (the spelling lavis
could have originated the corruption); similar expressions are used of
moral purity (e.g. Ov. trist. 1. 9. 43 ‘vitae labe carentis’, Sen. epist. 4. 1
‘mentis ab omni labe purae’) and of victims in ritual contexts (Ov. fast.
4. 335 ‘sine labe iuvencam’, met. 15. 130 ‘victima labe carens’). The phrase is
applicable to the boys as well as the girls and is more comprehensive than
casti: children taking part in religious ceremonies had to be not just
chaste, but also free-born with both parents living (cf. Wissowa 496);
note especially Liv. 37. 3. 6 (in the context of a supplicatio) ‘decem ingenui
decem virgines, patrimi omnes matrimique ad id sacrificium adhibiti’,
Macr. sat. 1. 6. 14. Other interpretations, some very implausible, are
recorded by P.Tremoli, GIF 7, 1954: 159 ff., Nisbet, op. cit. 112.
11–12. male nominatis / parcite verbis: male nominatis is not normal

Latin for ‘ill-omened’, but can be defended as a literal translation of the
Greek ıı ; for such ‘calques sémantiques’ cf. N–H on 1. 27. 9 and
2. 19. 29. ıı can mean not only ‘bearing an unlucky name’
(Soph. Aj. 914) or ‘unmentionable’ (West on Hes. theog. 171) but ‘of ill
omen’; cf. the euphemistic use of Pı to mean ‘inauspicious’. To
apply male nominatis to verbis is a further oddity, but H might be
attracted by the play on nomen and verbum; cf. the archaic nuncupare
verba of ritual formulae, as J. G. F. Powell suggests to us (cf. Liv. 9. 9. 5
‘verba legitima dedentium urbes nuncupare’, Val. Max. 5. 10. 1 ‘nuncu-
pationem sollemnium verborum’).
There is some manuscript support for male ominatis, but the hiatus is
surely impossible; the alleged parallels cited by Williams (1969) 94 ff. are
all different. Bentley proposed male inominatis, in which the informal
male hardly suits the solemn context; the adverb is found with negative
adjectives like insulsus and ineptus, but these words admit intensification
more naturally than inominatis, as Williams points out. Delz (MH 30,
1973: 54) proposes ab inominatis (after expertes); for this rare construction
he compares Liv. 25. 25. 6 ‘precantes ut a caedibus et ab incendiis
parceretur’.
13. hic dies vere mihi festus: the day will be festus for H in more than
the technical sense; for the same sort of point cf. Claud. de sext. cons.
Hon. 603 ff. ‘hinc te iam patriis laribus Via nomine vero / Sacra refert’.
13–14. atras / exiget curas: ater is associated not only with worries
(3. 1. 40 n., 4. 11. 35), but with death (serm. 2. 1. 58, epist. 1. 7. 6, André,
1949: 51). The curae refer both to Augustus’ health and to the violence
mentioned in the following words.
The important codex Bernensis has exiget; Priscian, GL 3. 189 has
exigit, but the day is only beginning, and it is desirable to have a parallel
to metuam (15). For exigere cf. Gratt. cyn. 475 f. ‘Liber tenuis e pectore
curas / exigit’, epist. 1. 15. 19 ‘quod curas abigat’; it combines well with hic
dies and atras (Sil. 15. 251 ‘Aurora . . . terris exegerat umbras’), and suits
the imagery of an adventus (4. 5. 5 ‘lucem redde tuae, dux bone, patriae’
with Doblhofer 86 ff., Claud. de sext. cons. Hon. 537 ff. ‘solitoque decen-
tior aer / . . . principis et solis radiis detersa removit / nubila’ with
Dewar’s parallels). The preponderance of the MSS have eximet, which
is found elsewhere of the removal of anxieties (epist. 1. 5. 18 ‘sollicitis
animis onus eximit’, Cic. Tusc. 2. 29), but the verb is less forceful (cf.
A. Minarini, Boll. d. stud. lat. 9, 1997: 42 ff.), and contributes nothing to
the imagery.
14–15. ego nec tumultum / nec mori per vim metuam: the emphatic ego
balances Caesare below and marks the beginning of the transition
described in the introduction above. tumultus was a traditional euphem-
ism for insurrection in Italy or Gallia Cisalpina; cf. Cic. Phil. 8. 3
‘maiores nostri tumultum Italicum, quod erat domesticus, tumultum
Gallicum, quod erat Italiae finitimus, praeterea nullum nominabant’,
T. N. Habinek, The Politics of Latin Literature, 1998: 78 f. Similarly vim
refers not to foreign battles but to violent disorder at home (Mommsen,
Strafrecht, 1899: 652 ff.); the jurists associated vis with metus or terror
(Ulp. dig. 4. 2. 1 ‘quodcumque vi atroci fit, id metu quoque fieri videtur’).
For the combination with civil war cf. 4. 15. 17 f. ‘custode rerum Caesare,
non furor / civilis aut vis exiget otium’ with Heinze’s note.
The infinitive mori with nec metuam would normally mean ‘I shall not
hesitate to die’ (cf. 3. 9. 11 ‘pro qua non metuam mori’); it is argued that
exiget curas and the parallelism of tumultum with mori prevent misun-
derstanding, but RN shares the doubts of some editors (Peerlkamp,
L.Müller, Campbell).
15–16. tenente Caesare terras: for the Roman claim to mastery of the
world cf. 3. 3. 45–6 n. The protection of Augustus is a guarantee of
internal peace; cf. 4. 15. 17 ff. (quoted in the last note), A. Alföldi, op. cit.
(introduction above), 86 ff. on the theme ‘te salvo salvi sumus’.
17. i, pete unguentum, puer, et coronas: for i, puer cf. serm. 1. 10. 92,
Prop. 3. 23. 23; for shopping-lists in comedy and sympotic poetry cf.
N–H vol. 1, pp. 421 ff.
18. et cadum Marsi memorem duelli: the wine-jar recalls the ‘Social
War’ of 91–87 bc, when Venusia had joined the revolt of Rome’s Italian
allies (App. civ. 1. 52, Diod. 37. 2. 10); this may have caused the
enslavement of H’s father (G. Williams ap. S. J. Harrison, 1995:
300 ff.). At this period the war was usually called bellum Marsicum
(cf. epod. 16. 3) because the Marsi were the first to take up arms
(Diod. 37. 2. 1) or bellum Italicum; the archaic form duelli, with its
suggestion of ‘old unhappy far-off things’ makes a contrast with the
delights of the symposium. For the associations of old wines cf. Juv. 5. 31
‘calcatumque tenet bellis socialibus uvam’ (hyperbolic), B. Baldwin,
AJP 88, 1967: 173 ff. (on Opimianum). Here H’s jar can tell a cautionary
tale about the tumultus of sixty-five years before, and though wine is
by nature ‘oblivious’ (N–H on 2. 7. 21) this one remembers, cf.
Oliensis 148.
19–20. Spartacum si qua potuit vagantem / fallere testa: the terrible

slave-revolt of 73–71, which was led by Spartacus (RE 3A. 1528 ff.,
R. Seager, CAH edn. 2, 9. 221 ff.), is described by Caesar as a tumultus
(bG 1. 40. 5). The insurrection started at Capua in the wine-growing
country and spread over southern Italy (Flor. 2. 8. 5 ‘totamque perva-
gantur Campaniam’, Plut. Crass. 9. 6 K æŁı KØæı Ø c
+ÆºÆ); though the slaves were well enough organized to defeat
consular armies, H’s picture of a marauding band (vagantem) looting
wine-stores doubtless contained an element of truth. Charisius, who
quotes the line for another purpose, reads vagacem (GL 1. 66), and this
might be right (Brink on ars, pp. 34 f.); although the word (‘ram-
pageous’) is not attested elsewhere, the drily humorous coinage suits
H’s tone (cf. epist. 2. 1. 70 f. ‘plagosum . . . Orbilium’), and a grammarian
is unlikely to have used it by a mere slip of the pen.
si qua means either ‘if anywhere’ or ‘if by any means’. qua cannot be
nominative, as some translations imply, as the syllable must be long; si
quae (considered by Lambinus) is implausible, as H does not use this
form for the indefinite. testa (presumably nominative) seems repetitious
after cadum; perhaps the colloquial word implies that a smaller pot
would do. Bentley saw the difficulty when he took ‘si qua testa’ together
as ablatives (cf. N–H on 2. 3. 12 for places where the ablative comes close
to repeating the subject); he compared Phaedr. 3. 1. 1 ff., ‘amphoram /
. . . testa nobili / odorem quae iucundum late spargeret’, but there testa
suggests the material that diffused the fragrance. One might have
expected an ablative agreeing with qua to describe the place where the
jar was hidden; RN has tried cista, ‘hamper’ (Blümner, 1911: 131), perhaps
suggesting the cista where the emblems of Bacchus were concealed
(N–H on 1. 18. 12).
21. dic et argutae properet Neaerae: the position of properet is artificial,
but it adds emphasis and urgency; for the request cf. 2. 11. 22 ff. ‘eburna
dic age cum lyra / maturet’ with N–H (for further resemblances in that
passage see below, 22 n.). argutae means ‘clear-sounding’ (ºªØÆ); here it
implies that the girl is a singer (ps.-Acro comments ‘citharistae sive
cantrici’). Neaera was a name borne by hetaerae, including heroines of
comedies (see B. L. Ullman, CQ 9, 1915: 29); as the word means ‘young’,
it supplies a contrast to albescens (25).
22. murreum nodo cohibente crinem: murreum means ‘scented with

myrrh’ (ps.-Acro); cf. Archil. 48. 5, Meleager, anth. Pal. 5. 175. 2, Prop.
1. 2. 3 ‘crines perfundere murra’, Virg. Aen. 12. 100. Editors believe
Porph. when he mentions a colour ‘inter flavum et nigrum’, but what-
ever the usage may have been in his day (hodie), that does not certify the
adjective murreus as meaning ‘yellow’ here, in spite of Ov. met. 15. 399
‘fulva . . . murra’. Some cite as parallels for the alleged colour Prop. 3. 10.
22 ‘et crocino nares murreus ungat onyx’, [Tib.] 3. 4. 28 ‘stillabat Syrio
myrrhea rore coma’ (myrtea codd.), but in both these passages the
adjective must refer to a scent.
We interpret nodo here as a simple band tied by a knot; cf. 2. 19. 19 f.
‘nodo coerces viperino / . . . crinis’, Sen. Tro. 100 f. ‘coma demissa est /
libera nodo’, Phaedr. 401 f. ‘et nodo comas / coegit emisitque’, especially
HO 1546 ‘comas nullo cohibente nodo’, OLD s.v. 6a. In some other
passages it is better taken as a bun (Ov. met. 3. 169 f., 8. 319 ‘crinis erat
simplex nodum collectus in unum’); but here cohibente suits a band
better, and it takes less time to put on a band than to arrange a bun.
At 2. 11. 23 f., where there are difficult textual problems, N–H read
‘maturet, incomptum Lacaenae / more comae religata nodum’ (taking
nodum as a bun); there Bentley and others prefer ‘incomptam Lacaenae /
more comam religata nodo’ (taking nodo as a band).
cohibente is Muretus’ conjecture (endorsed by Bentley) for the cohibere
of the manuscripts and apparently the scholiasts. H’s chief point is that
Neaera should come quickly, not that she should make haste to put her
hair in a band; cf. Sen. HO 1546 f. quoted above. With cohibere the pace
of the sentence is checked, for we have to revise our view of the syntax.
Moreover, mora (23) should answer to properet; the janitor may try to
obstruct Neaera’s departure, but he is not going to interfere with her
hair-style.
Our passage may have influenced Milton at Lycidas 67 ff. ‘Were it not
better done as others use / To sport with Amaryllis in the shade / Or
with the tangles of Neaera’s hair?’ There the second ‘with’, which seems
inert as a preposition, has been taken by some as the verb ‘withe’. For
the source of these lines see J. B. Leishman, Milton’s Minor Poems, 1969:
280 ff. (citing also George Buchanan and Johannes Secundus).
23–4. si per invisum mora ianitorem / fiet—abito: the concierge is
conventionally a surly slave; cf. Sen. dial. 2. 14. 2 ‘sapiens non accedet
ad fores quas durus ianitor obsidet’, 5. 37. 3, Blümner (1911), 29. Some-
times he was employed as a custos by the woman’s vir (e.g. Ov. am. 3. 4.
1 f.), and as such was an unpopular figure in erotic poetry; see Antip.
Thess. anth. Pal. 5. 30. 3 f., Prop. 2. 23. 9 (‘vultum custodis amari’), Tib.
1. 2. 5, Ov. am. 1. 6, ars 3. 587 f. Usually, however, the door-keeper
prevents the man’s entry rather than the girl’s exit. invisum is amusingly
overplayed and makes a contrast with Neaera’s glamour.
abito is a ‘future imperative’; the form is used in instructions that are
not to be immediately fulfilled, and is therefore regularly found in legal
texts and general precepts (K–S 1. 198, H–Sz 340 f., Woodcock 126).
Here the word means ‘come away’ and balances i (17). The unusual
break in the adonius underlines the surprise (we had half expected the
slave to be ordered to use force).
25. lenit albescens animos capillus: for H’s grey hair see the parallels
cited by N–H on 2. 11. 15; capillus is more prosaic than crinem (22), cf. the
tables in TLL 3. 314. For the sentiment cf. Philodemus, anth. Pal. 5. 112.
3 f. ºØc ªaæ KªÆØ Id ºÆ = Łæd X, ıB ¼ªªº
ºØŒ (with reference to sex); for the opposite view cf. Herodas 1.
67 f. ˆıºº, a ºıŒa H æØ
H IÆºØ = e F. The plural
animos means ‘fiery spirit’ (cf. epist. 1. 19. 24 f. ‘animos . . . Archilochi’,
Virg. Aen. 1. 57 ‘mollitque animos ac temperat iras’, OLD s.v. 11); for H’s
hot temper cf. 3. 9. 23 n. For the mollifying effect of age cf. epist. 2. 2. 211
‘lenior et melior fis accedente senecta?’
26. litium et rixae cupidos protervae: litium refers to verbal wrangling;

cf. Juv. 6. 268 f. ‘semper habet lites alternaque iurgia lectus / in qua nupta
iacet’ with Courtney. rixae goes further and suggests physical brawling
(cf. Prop. 3. 8. 1 on a row with Cynthia); particularly relevant is the
violence of the exclusus amator (3. 26. 7 f., Ov. ars 3. 71, Copley 148 n. 26).
protervae also indicates the young man’s aggressive pursuit of sex (1. 25.
2, 2. 5. 15).
27–8. non ego hoc ferrem calidus iuventa / consule Planco: non ferrem
can mean the same as non tulissem in early Latin, and has often this
sense even in Cicero; cf. Handford 123 ff.
For the fervour of youth cf 1. 16. 23 ff., ars 115 f. For similar reflections
cf. Homer Il. 23. 626 ff. (Nestor), Ar. Ach. 211 ff. PŒ i K KB ª
- = , ‹ ªg æø = IŁæŒø æ = MŒºŁı 6Æ7ººfiø
æ
ø, = z Æºø i › æ y K- = F
ØøŒ = Kıª, Herodas 2. 71 ff. with Headlam, Virg. Aen. 5.
398 ff. H’s adaptation of the motif provides a typically self-depreciating
1 5 . V XO R PAV P ER I S I B YC I 191
closure (cf. 1. 33. 13 ff., 2. 4. 22 ff., 3. 19. 28, epod. 14. 15 f.). For L. Munatius
Plancus see RE 16. 1. 545 ff., N–H vol. 1, pp. 90 ff. This eminent man was
still alive at the time of writing (censor in 22 bc); he had been consul in
42 bc, when Horace was 22–3, and no contemporary was likely to forget
that this was the year of Philippi.
15. VXOR PAVPERIS IBYCI

[C. C. Esler in Old Age in Greek and Latin Literature (ed. T. M. Falkner and J. de Luce),
1989: 174 ff.; Pasquali 449 ff.; M. C. J. Putnam, CP 71, 1976: 90 ff. ¼ Essays on Latin Lyric,
Elegy, and Epic, 1982: 126 ff.]
1–6. Wife of Ibycus, set a limit to your misbehaviour, and now that you are
growing old, stop casting a cloud over the frolics of the girls. 7–10. It is more
defensible for your daughter Pholoe to storm young men’s houses, but it is
inappropriate, Chloris, for you. 11–16. It is all right for her to frolic like a
frisky young animal, but wool-working is suitable for you, not lyres or roses,
or wine-casks drunk to the dregs.
Mockery of ageing women had a long literary tradition behind it.

Archilochus’ gibes at Neobule already show the pattern: see 188. 1 f.
W PŒŁ ›H ŁººØ ±Æºe
æ Æ, Œº., 196a (edn. 2). 26 ff. ÆNÆE,
ØæÆ, d (‘over-ripe, twice your age’) / ¼Ł IææŒ
ÆæŁœ = ŒÆd
æØ m æd KB, A. P. Burnett, Three Archaic
Poets, 1983: 77 ff. (for Hipponax and Semonides see Grassmann 12 f.).
Comedy depicts old women as grotesquely devoted to sex and drink; see
H. G. Oeri, Der Typ der komischen Alten in der griechischen Komödie,
Diss. Basel, 1948, J. Henderson, TAPA 117, 1987: 105 ff. For Greek
epigram see Rufinus, anth. Pal. 5. 21, 27, 28, 76; in Roman elegy
(Prop. 3. 25. 11 ff.) an unresponsive woman is told ‘one day you will
be old and ugly, and then you will be sorry’ (cf. N–H vol. 1, pp. 289 ff.).
In satire Lucilius speaks of a ‘vetulam atque virosam / uxorem’
(282 f.).The impulse for such invective can sometimes be explained
by jealousy or a wish for revenge; for apparently impersonal abuse
for its own sake see for instance Mart. 3. 93, 11. 29 (with Kay), Juv. 6.
119 ff. See further K. Cokayne, Experiencing Old Age in Ancient Rome,
2003: 134 ff.
Horace wrote a handful of poems within this tradition. In his epodes
he reviles older women with the obscene physical detail that went with
the iambic genre (Grassmann 5 ff.): 8. 1 ff. ‘rogare longo putidam te
saeculo / viris quid enervet meas, / cum sit tibi dens ater . . . ’, 12. 7 ff.
‘qui sudor vietis et quam malus undique membris / crescit odor, cum
pene soluto / indomitam properat rabiem sedare’. In carm. 1. 25 he
predicts Lydia’s old age; in a later poem he exults that Lyce’s old age
is now arriving (4. 13. 2 ff. ‘fis anus et tamen / vis formosa videri /
ludisque et bibis impudens . . . ’).
Our poem contains some of the traditional sneers: Chloris has one
foot in the grave (4 n.), but shows an unseemly addiction to sex (12 n.)
and drink (16 n.). There is none of the overt obscenity associated with
iambi, and no display of the personal bitterness found in epigram and
elegy; the censure is conveyed with some degree of moderation (e.g.
laboribus in v. 3, ludere in v. 12), and the criticism is ostensibly based on
decorum (decet in v. 8, decent in v. 14). Nevertheless, the tone throughout
is derisive rather than genuinely moralistic; and though Horace’s
manner is more restrained than in most of the prototypes, a good deal
of their inhumanity remains. One may contrast the rueful realism
with which he contemplates his own advancing years (cf. J. N. Bremmer
in Sexual Asymmetry (ed. J. Blok and P. Mason), 1987: 191 ff., Esler,
op. cit. 177 ff ).
1. Vxor pauperis Ibyci: the form of address does not reflect Roman
custom, whereby married women retained their gentile name, but
indicates the wife’s domestic responsibilities. If the scene is some Apu-
lian town, as one might infer from Luceriam (14), then an Apulian wife
ought to have behaved better; cf. epod. 2. 41 ff. ‘Sabina qualis aut perusta
solibus / pernicis uxor Apuli . . . ’ (where RN has proposed parcentis)—a
passage imitated at Stat. silv. 5. 1. 122 ff. ‘velut Apula coniunx / agricolae
parci vel sole infecta Sabina . . . ’. uxor is the everyday word for ‘wife’,
avoided in epic but found seven times in the Odes; cf. Axelson 57,
P. Watson, CQ 35, 1985: 431 f., Lyne (1989), 43 ff., 60 ff. pauperis suggests
respectable frugality rather than destitution (3. 2. 1 n.), and makes a
contrast with the wife’s flightiness.
The name Ibycus is thought to recall the sixth-century poet from
Rhegium; but his hedonism and reputed lechery (Cic. Tusc. 4. 71) are
not in point here. Commentators invoke the proverbs Iæ
ÆØ æ (or
Iæ ) +Œı ( paroem. Gr. 1. 207, 1. 251): the paroemiographer
explains that Ibycus was called old-fashioned (or silly) because he
declined the position of tyrant, but the analogy to the hapless husband
is hardly close enough. It might be more relevant that an early Pythag-
orean (presumably also from the South of Italy) was called Ibycus
(Athen. 2. 69e cites him on lettuce); as the frugality of the sect was
notorious (Mayor on Juv. 15. 173 f., Kerkhecker on Call. iamb. p. 40), that
would provide a possible association with pauperis. ‘Antiquated’ would
suit the Pythagorean better than the poet, and the same could be true of
the offer of the dictatorship, for the Pythagorean Archytas did become
strategos at Tarentum.
2. tandem nequitiae fige modum tuae: the emphatic tandem suggests

impatience; cf. 1. 23. 11 f. ‘tandem desine matrem / tempestiva sequi viro’.
fige modum ‘fix a limit’ is stronger than the variant pone; cf. Cic. parad. 25
‘si quidem rerum modum figere non possumus, animorum modum
tenere possumus’, Arnob. nat. 1. 10. When H says ‘limit’ rather than
‘end’, he is being sarcastically polite (cf. N–H on 1. 16. 2), as if moder-
ation was the best that could be hoped for.
nequitia is the quality of the nequam or ‘good-for-nothing’, and
is used with various degrees of censoriousness; the following line
shows that here it refers to sexual misconduct; cf. Gallus, fr. 2. 1 ‘tristia
nequitia . . . Lycori, tua’, Prop. 1. 15. 38, Ov. ars 2. 392, fast. 1. 414.
The postponed possessive tuae may indicate a contrast with the
husband’s frugality (‘that profligacy of yours’); possessive adjectives in
this position sometimes have more emphasis than is usually realized;
cf. 1. 3. 8, 2. 13. 10, 2. 17. 1, 3. 13. 16, 3. 16. 30, 3. 19. 28, Nisbet in
Adams–Mayer 143.
3. famosisque laboribus: ‘scandalous exertions’; famosis makes some-

thing of an oxymoron with the laboribus expected of an Apulian house-
wife.
4. maturo propior desine funeri: in this kind of context we expect

maturus to refer to a ripe old age (OLD s.v. 6), so funeri comes as a
surprise; cf. Ar. vesp. 1365 (of a randy old man) ŁE KæA ØŒÆ
‰æÆÆ æF, where ‘coffin’ is an unexpected substitute for ‘girl’. propior
means ‘quite close’, not ‘daily nearer’ (Page), or ‘closer to death than to
the girls’ (Williams), or ‘nearer than the virgines of v. 5’ (Quinn); for a
similar point cf. 2. 18. 18, where the greedy old man continues to build
mansions ‘sub ipsum funus’.
Derision of the elderly sometimes suggests they have ‘one foot in the
grave’; cf. epod. 8. 11 f., Ar. eccl. 905, 926, 996, 1073 with L. K. Taaffe,
Aristophanes and Women, 1993: 123 ff., Plut. 1008, 1033, Nicarchus, anth.
Pal. 11. 71. 4, Plaut. merc. 290 f. ‘Acherunticus / senex vetus decrepitus’
with Enk’s note, Mart. 3. 32. 2 ‘sed tu mortua non vetula es’, 3. 93. 19
‘virumque demens cineribus tuis quaeris’, Rufinus, anth. Pal. 5. 21. 6.
Sometimes the old person is actually described as a tomb; cf. Eur. Med.
1209 f. e ªæÆ OæÆe Ł = ŁØ; with Page,
Ar. Lys. 372 t with Taillardat (1962), 53, anon. anth. Pal. 11.
425. 2, Plaut.Pseud. 412 ‘ex hoc sepulchro vetere viginti minas / effodiam
ego hodie’, Priap. 57. 1 ‘cornix et caries vetusque bustum’, Grassmann
16 f., 21.
5. inter ludere virgines: for the position of inter cf. 3. 3. 37, 3. 27. 51, epist.
1. 3. 4, K–S 1. 588, TLL 7. 2147. 52 ff.; the word-order does not accord
with normal prose usage, but we doubt whether it is meant to reflect the
woman’s position in the ring of dancers (contrast the passage from PLM
quoted below). ludere, ‘frolic’, here suggests choral dancing; cf. 2. 12. 18 f.
‘nec dare bracchia / ludentem nitidis virginibus’, PLM 5. 77. 59 f. (a post-
classical imitation) ‘tuque puellarum dum ludis in agmine princeps /
inter virgineos lucida stella choros’. The Suda 1. 539. 30 f. cites as a
proverb (cf. 10 n.) ˆæÆF
æØ Kd H Ææ uæÆ Ø ØÆæÆø
ÆE ªaæ ÆØ æØ e
æØ (cf. decet in v. 8). The impropriety is
aesthetic rather than sexual (as is shown by v. 6), though ludere is
repeated with a sexual implication in v. 12.
6. et stellis nebulam spargere candidis: epic describes how heavenly

bodies are obscured by cloud (Hom. Il. 11. 62 f., Od. 9. 144 f., Ap. Rhod.
4. 1696 ff., Virg. Aen. 3. 586). It is a convention of panegyric to compare
the superiority of the laudandus over his company to that of one
heavenly body over others (1. 12. 48 with N–H); cf. also Ap. Rhod. 1.
239 f. ƒ b ÆØd = Iæ S Ø æ. In our passage this
latter image is reversed: the old woman casts a shadow over the radiant
girls. candidis suits both bright stars (Plaut. rud. 3) and beautiful girls;
they are perhaps thought of as wearing their whitest dresses for the
dance (cf. 2. 12. 19 nitidis, cited in 5 n.).
7. non, si quid Pholoen satis: non negates decet (8); for the hyperbaton
see K–S 1. 819. As Chloris (8) is the uxor (1), so Pholoe seems to be her
filia (8); otherwise her appearance is unmotivated and the balance of the
passage is upset. The two names are combined also at 2. 5. 17 f. ‘dilecta
quantum non Pholoe fugax, / non Chloris albo . . . umero nitens’. The
parallel is hard to explain, for here Chloris is not beautiful; perhaps
there was a Greek prototype which H has developed in two different
ways. ‘Pholoe’ is a name persistently used of a girl who is hard to get; see
1. 33. 6 f. ‘Cyrus in asperam / declinat Pholoen’, 2. 5. 17 ‘Pholoe fugax’
with N–H, Tib. 1. 8. 27 ‘nec tu difficilis puero tamen esse memento’, 1. 8.
69 ‘oderunt, Pholoe, moneo, fastidia divi’, Stat. silv. 2. 3. 10 (a nymph
pursued by Pan). One assumes in our passage that she is one of the
dancing virgines (5).
8. et te, Chlori, decet: the adjective

ºøæ often suggests a pale yellow,
the colour of withered grass (Sappho 31. 14 f.
ºøææÆ b Æ = Ø)
or of frightened people; cf. Irwin (1974), 62 ff., A. Lorenzoni, Eikasmos 5,
1994: 139 ff. The name ‘Chloris’ (Pape–Benseler 2. 1687) was given to a
daughter of Niobe who grew pale when her brothers and sisters were
killed (Paus. 2. 21. 9). More relevant, perhaps, is the fact that a Chloris
was the mother of Nestor (Hom. Od. 11. 281 ff., Hyginus 97); for similar
sneers at vetulae cf. Lollius Bassus, anth. Pal. 11. 72. 2 ªæÆEÆ Ø m ˝øæ
PŒØ æÆ , Myrinus, ibid. 11. 67. 2 f., Priap. 57. 3 f. ‘quae forsan
potuisset esse nutrix / Tithoni Priamique Nestorisque’ (cf. 12. 1 f.), Mart.
3. 76. 4, 10. 67. 1. Admittedly
ºøæ , like
ºæ , sometimes implies
greenness and the sap of youth (Onians 177 n. 9), the name Chloris could
mean ‘greenfinch’, and at 2. 5. 18 it suggests a beautiful paleness which
can have nothing to do with Nestor’s mother; but it is hard to see an
ironical suggestion of youthfulness in our present passage.
Ancient moralists sensibly emphasize that what is decorum (æ)
varies according to one’s temperament, status, and notably age (Dyck on
Cic. off. 1. 93–107); cf. also off. 1. 110 ff., especially 122 ‘et quoniam officia
non eadem disparibus aetatibus tribuuntur aliaque sunt iuvenum, alia
seniorum . . . ’, 123 ‘luxuria vero cum omni aetati turpis, senectuti
foedissima est . . . ’; see further epod. 13. 5, epist. 1. 14. 36 ‘nec lusisse
pudet sed non incidere ludum’ (echoing Philodemus, anth. Pal. 5. 112.
5 f. ŒÆd ÆØ ‹ ŒÆØæ , KÆÆ ŒÆ ŒÆd F = PŒØ, ºøœæ
æ ±ł ŁÆ), 2. 2. 216 ‘lasciva decentius aetas’, Cic. Cael. 41–2,
Otto 106.
8–9. filia rectius / expugnat iuvenum domos: ‘it is more appropriate for
your daughter . . . ’; for this use of an adverb cf. ars 129 ‘rectius Iliacum
carmen deducis in actus’, 3. 16. 39–40 n. expugnat, ‘takes by storm’,
belongs to the militia amoris; cf. 3. 26. 2 n., Lyne (1989), 34 f. Here
there is a hyperbolical reversal of roles, as conventionally it is the
roistering male who breaks down doors (3. 26. 6–8 nn.). For such
aggressive behaviour by a woman cf. Plaut. mil. 1249 f. (a meretrix
converses with her maid) ‘durare nequeo / quin eam intro.—occlusae
sunt fores.—ecfringam.—sana non es’, Sen. nat. quaest. 4a, praef. 6
‘Crispus Passienus saepe dicebat adulationi nos non cludere ostium
sed operire, et quidem sic quemadmodum opponi amicae solet, quae si
impulit grata est, gratior si effregit’. These are extreme instances of the
mulier proterva who takes the sexual initiative; cf. 1. 33. 13, 2. 5. 13 with
N–H, Ar. eccl. 877–1111.
Yet if the filia is indeed Pholoe, this behaviour is strangely out of line
with the normal implications of her name (7 n.). To meet this difficulty
we suggest reading expugnet (‘it would be more appropriate for your
daughter’), without implying that she is actually behaving in this way; in
fact it is the elderly Chloris who is supposed to storm young men’s
houses and to behave like a Bacchanal (10 n.).This would require read-
ing L. Müller’s cogat in v. 11 (see note).
10. pulso Thyias uti concita tympano: a Thyias was a Bacchanal (N–H
on 2. 19. 9, below on 3. 25); the word, which is a trochee, is supposed to
be derived from ŁØ, ‘to rush’. Similes comparing emotional women
to Bacchanals suited the male perception of female psychology, and
were a long-standing poetic motif; cf. Hom. Il. 22. 460 ªæØ
Øı ÆØØ Y, h. Dem. 386 with Richardson, Virg. Aen. 4. 301 f.
‘qualis commotis excita sacris / Thyias’ with Pease, Prop. 3. 18. 14 with
Fedeli. The old woman playing the Bacchanal became a proverb; see
paroem.Gr. 1. 57 ˆæÆF ÆŒ
Ø, 1. 228 (so above, 5 n.). The tympanum or
tambourine (1. 18. 14 with N–H, Wille 53 f.) was particularly associated
with Bacchus and Cybele; cf. Ov. fast. 4. 183 ‘ibunt semimares et inania
tympana tundent’; the ancients were well aware of the unnatural excite-
ment produced by a throbbing drum-beat.
11. illam cogit amor Nothi: L. Müller proposed cogat (accepted by SB),
taking the clause as concessive. If expugnet is read in v. 9 (see note), that
would import a different nuance: ‘let love of Nothus drive her (not the
mother, as is in fact the case)’. This interpretation allows the mother to
remain on centre stage.
The name Nothus (the Greek for ‘bastard’) is attested in S. Italy
(LGPN 3A. 329) and several times in the index to CIL 6 (Rome).
Though ‘bastard’ was not a regular term of abuse in antiquity, it may
hint at a background which is not quite respectable; cf. Putnam, op. cit.
129 n. 11.
12. lascivae similem ludere capreae: a caprea is normally a small roe-

deer, the feminine of capreolus (Varr. lL 5. 101 ‘caprea a similitudine
quadam caprae’) and conventionally in poetry a victim of wild animals
(1. 33. 8, epod. 12. 26); see RE 1A. 512 f. There is a difficulty here, for
elsewhere it is shy young girls who are compared with deer (1. 23. 1 ‘vitas
inuleo me similis, Chloe’ with N–H). Ps.-Acro comments on our
passage ‘genus pecudis libidini et lasciviae aptius’; but pecudis suggests
that he has made a simple confusion with the domestic goat (capra or
capella). It would make sense here if the caprea could be a kind of wild
goat or ibex (RE 3A. 2. 2238 ff.); possible parallels, none of them
decisive, are Virg. Aen. 10. 724 f. ‘si forte fugacem / conspexit capream
aut surgentem in cornua cervum’ (imitating Hom. Il. 3. 24 j ºÆ
ŒæÆe j ¼ªæØ Ær ªÆ), Ov. fast. 2. 491 ‘est locus, antiqui Capreae dixere
paludem’ (cf. Liv. 1. 16. 1 ‘Caprae paludem’, which is a variant in Ovid),
Mart. 13. 98. 1 ‘pendentem summa capream de rupe videbis’ (imitating
Virg. ecl. 1. 76 of goats ‘dumosa pendere procul de rupe videbis’), Juv. 10.
93 f. ‘principis angusta Caprearum in rupe sedentis / cum grege Chal-
daeo’ (the island of Capri is suited to goats rather than deer, a point
underlined by grege).
ludere echoes ‘inter ludere virgines’ (5), but this time it is a sexual
euphemism, like ÆØ; cf. 4. 13. 4, Pub. Syr. 30 ‘anus cum ludit, morti
delicias facit’, Adams (1982), 162 f. For an older woman’s sexual interests
see the introduction above with Oeri, op. cit., Grassmann 14, 17 f., 26,
F. J. Brecht, Philologus suppl. 22. 2, 1930: 62 f.
13–14. te lanae prope nobilem / tonsae Luceriam: the plural is natural

with materials in prose and verse, cf. 2. 16. 37, K–S 1. 75; so more often in
Greek (K–G 1. 15). The virtuous Roman housewife is conventionally
portrayed as either spinning herself or superintending others (but not as
shopping or cooking). See especially Liv. 1. 57. 9 (Lucretia) ‘nocte sera
deditam lanae inter lucubrantes ancillas in medio aedium sedentem
inveniunt’ (with Ogilvie), Suet. Aug. 64. 2, 73. 1 (the Augustan ideal),
Treggiari, 1991: 243 f.; so in inscriptions ILS 8403. 8 ‘domum servavit
lanam fecit’, 8394, 8402. 2 ‘lanifica pia pudica frugi casta domiseda’, laud.
Turiae 1. 30, Lattimore 297. Such labours were particularly important for
the poor and the elderly; cf. Ter. Andr. 75 f. ‘primo haec pudice vitam
parce et duriter / agebat lana ac tela victum quaeritans’, heaut. 285,
perhaps from New Comedy (F. Leo, Plautinische Forschungen, 144),
Phaedr. 4. 5. 4 f. ‘unam formosam et oculis venantem viros, / at alteram
lanificam et frugi rusticam’. The Greek attitude was very much the
same, see e.g. Xen. rep. Lac. 1. 3, Plat. Lys. 208d, Musonius p. 18
Hense ¼ p. 48 Lutz; for some parallels emphasizing poverty see Hom.
Il. 12. 433 ff., Ap. Rhod. 3. 291 ff., 4. 1062 ff. (imitated at Aen. 8. 409 ff.),
anon. anth. Pal. 6. 283 (a hetaera reduced to weaving), and S. Tarán, The
Art of Variation in the Hellenistic Epigram, 1979: 115 ff.
Luceria was one of the leading towns in Apulia, though its import-
ance had diminished (Strabo 6. 3. 9); cf. RE 13. 1565 f., E. T. Salmon,
Samnium and the Samnites, 1967 (index). For Apulian wool see Strabo
loc. cit. KæÆ ÆºÆŒøæÆ b B 3ÆæÆ K, ºÆæa b w,
Plin.nat. hist. 8. 190, Mart. 14. 155. 1 f. ‘velleribus primis Apulia, Parma
secundis / nobilis’, CIL 9. 826 (a lanarius from Luceria), Salmon, op. cit.
67 ff., Brunt (1971), 366 ff. Luceria boasted a temple of Minerva (cf. 3. 12.
5 ‘operosaeque Minervae’) with a statue reputedly brought by Diomedes
from Troy (Strabo 6. 1. 14, 6. 3. 9); this association must have been
familiar to the Apulian Horace. By nobilem H is adding dignity to the
place as if it were a famous Greek city, cf. 1. 7. 1 ‘claram Rhodon’; some
interpret the word as ‘famous for its wool’ (as in Mart. loc. cit.), but here
there is no ablative.
14. non citharae decent: to provide music at a symposium was not the
business of an elderly woman; cf. 3. 28. 11 n. decent echoes decet (8).
15. nec flos purpureus rosae: roses were a regular feature of the sympo-
sium, where their petals might be scattered (see e.g. 3. 19. 22 ‘sparge
rosas’, Prop. 4. 8. 40 of a castanet-girl showered with roses, Ov. fast. 5.
360), or their blooms might be woven into wreaths or garlands (e.g. 2. 11.
14 ‘rosa / canos odorati capillos’, 3. 29. 3, Cic. Tusc. 3. 43 ‘sertis redimiri
iubebis et rosa?’), but their beauty soon withered; cf. 2. 3. 13 f. ‘nimium
brevis . . . rosae’. purpureus is conventionally used of roses but may also
hint at the rosy complexion of youth; cf. 2. 5. 12 with N–H, 4. 10. 4 ‘qui
color est puniceae flore prior rosae’, Phrynichus fr. 13 (TrGF 1 p. 77)
ºØ Kd æıæÆØ Ææfi Ø Ð H æø . In that case there would be
a contrast with the pallor of Chloris (8 n.).
16. nec poti vetulam faece tenus cadi: the bibulous old woman, as
remarked in the introduction, is a stereotype of comedy and epigram;
cf. 4. 13. 4 f. ‘ludisque et bibis impudens’, Ar. nub. 555 æŁd . . . ªæÆF
Ł F Œ æÆŒ o
with Dover, Athen. 10. 440d–442a (citing
Middle Comedy), anth. Pal. 7. 353, 455, 456, 11. 73, 297–8, Plaut. Curc.
96–140. For abundant further material see Brecht, op. cit. (12 n.) 66,
Oeri, op. cit. (in the introduction above), 13 ff., Grassmann 21,
P. Zanker, Die Trunkene Alte, 1989 (with illustrations, including the
famous sculpture from Munich).
For ‘drinking to the lees’ cf. 1. 35. 26 f. ‘cadis / cum faece siccatis’ with
N–H, Theoc. 7. 70, Gaetulicus, anth. Pal. 11. 409. Here citharae, flos
purpureus rosae, and poti faece tenus cadi represent, in an increasing
tricolon, pleasures which are suited to attractive young women. The
isolated vetulam (read by the scholia where the bulk of the MS tradition
offers vetula) underlines the embarrassing incongruity of Chloris’s be-
haviour. Pasquali (453) (not followed by NR) suggests that v. 16 has an
ironic dimension that distinguishes it from its predecessors: in one sense
of decet music and roses do not ‘grace’ an old woman as they do the
young; but the verb can also mean ‘suit’ (cf. epist. 1. 7. 44 ‘parvum parva
decent’), and in this sense heavy drinking does suit the stereotype of
the vetula (a derogatory word). Taking a hint from Pasquali, RN has
considered the further insult that she herself has now reached the dregs;
cf. Ar. plut. 1084 ff. XP. KØc ŒÆd e r Mı = Ø, ıŒ
K Ø ŒÆd c æªÆ. NE. Iºº Ø ŒØfi Ð æf ÆºÆØa ŒÆd Ææ—
‘old and fusty dregs’. (For another metaphor from the symposium
directed at an older woman see 4. 13. 26 ff. ‘possent ut iuvenes visere
fervidi / . . . dilapsam in cineres facem’.)
With the above interpretation one might think of accepting vetula;
by applying to the wine-lees the adjective often used of old women
Horace would be underlining the analogy. But though vetulus is used
affectionately of old wine (Catull. 27. 1 ‘minister vetuli puer Falerni’,
Mart. 1. 18. 1 with Citroni, cf. Eubulus fr. 121 ¼ PCG 5, p. 263 ¸Ø
ªæÆ ŒÆæÆªB), it is not attested in a disparaging sense of lees;
the nearest equivalent would be Aristophanes, loc. cit.
16. INCLVSAM DA NAEN
[R. W. Carrubba, MH 47, 1990: 139 ff.; Fraenkel 229 ff.; Lyne (1995), 122 ff.; R. J. Schork,
TAPA 102, 1971: 515 ff.; D. West ap. Costa 36 f.; Williams 600 ff.]
1–8. When Danae was confined in a brazen tower, Jupiter gained access
by turning himself into a bribe. 9–16 Gold breaks through all barriers, as is
shown by instances from mythology and history. 17–20. Yet as money brings
worry and grander ambitions, I have been right, Maecenas, to keep my head
low. 21–32. The more one denies oneself, the more the gods will bestow. I gain
more glory from the possessions I reject than if I hoarded all the grain in
Apulia, and more happiness from my modest estate than the owners of
provinces. 33–44. Though I derive no wealth from agriculture, I am not
poor, and you would not refuse me if I asked for more; I shall expand my
revenues by limiting my desires rather than by ruling the kingdoms of the
East. Those who seek much lack much; he is well-off who is granted enough.
The ode begins with an arresting exemplum, whose full implications

are not immediately apparent. Danae and the shower of gold show the
irresistible power of money; in this Horace follows not the traditional
myth (1 n.) but the later rationalization (8 n.). The story is sketched
light-heartedly, as suits the cynical intention; cf. adulteris (which in-
cludes Jove), custodem pavidum (of the anxious Acrisius), and the meta-
morphosis converso in pretium deo. It is pointless to complain with
Fraenkel that the ode ‘has no wings’; Horace knew many ways of writing
a poem (cf. Schork, op. cit.), and this amusing opening is highly
effective. In the third and fourth stanzas he rubs in the moral with
instances ranging from Greek myth to contemporary Roman history.
With exaggerated sententiousness he rings the changes on words for
money (9 n.), and exemplifies its power with a series of forceful verbs
that are varied to suit the context, such as perrumpere (10), demersa and
diffidit (13), subruit (14), inlaqueant (16).
In the fifth stanza (17–20) there is an unexpected change of direction:
instead of emphasizing the power that money confers on the donor, he
points to the trouble it causes the recipient. He gives his maxim a
personal application by referring to his own modest aims (18 f.); then
in v. 20 he at last mentions Maecenas, the addressee and in part the
subject of the poem. This stanza provides the pivot of the ode, as it
looks back to the exempla of vv. 1–16 and forward to the personal
affirmations of vv. 21–44. In the sixth, seventh, and eighth stanzas
Horace continues the account of his own life-style. With a parody of
the military imagery used by philosophers he presents himself as a
deserter from the camp, not of virtue, but of riches (23 n.). Great estates
are rejected with the hyperboles of diatribe and the paradoxes of
Stoicism (25 n., 28 n.), which on Horace’s lips could only be half-serious.
The description of the Sabinum (29–32) gives a personal and original
note to his moralizing; by describing alike its blessings and its simplicity,
the poet shows himself both grateful and ungrasping in his relations
with his benefactor.
The final three stanzas repeat the themes of their three predecessors,
though in a different order; the sequence is roughly ABC: BCA.
Horace turns once more from the latifundia that he rejects (33–6 answer
25–8) to the sufficiency that he owes to his patron (37–8 answer 29–32);
he then winds up by reverting to more general sententiae (39–44), first
stated in a personal, then in an impersonal, form (the reverse order
from 21–4). The balance of the two sections is underlined by several
significant repetitions (cf. Syndikus, 2. 157): quam si (41) corresponds to
quam si (26), multa petentibus (42) to nil cupientium (22), multa . . . multa
(42 f.) to plura . . . plura (21 f.), deus (43) to ab dis (22). Though Horace is
too discreet to press the point, the god who has given such blessings
recalls the poet’s human patron (cf. serm. 2. 6. 5 and 41, Virg. ecl. 1. 7
‘namque erit ille mihi semper deus’). The beneficent deity who knows
where to stop also allows a contrast with the Jupiter of the opening
parable (8).
So the poem is not just about ostentation versus simplicity, but also
about the right relation to one’s benefactors. ‘How to receive’, was a
standard theme of the ancient literature de beneficiis (Sen. ben. 6); if the
topic is unwelcome to modern readers, that is because we feel uncom-
fortable with a system that accepted inequalities of wealth and status,
where liberality was a duty owed, not to suffering strangers, but to
grateful relatives, friends, and clients. If the situation of poet and patron
is related to stanzas 1 and 2, we have an implied contrast: Maecenas is
quite unlike the scheming Jupiter; and Horace, unlike Danae’s guards, is
not open to corruption. (In the same way, the anecdote about Philippus
in epist. 1. 7 shows the wrong way to be a benefactor.) Any possible
offence is dissipated by the humour, which shows an easy friendship
without sycophancy in spite of the difference in status. Horace’s refusal
to make himself conspicuous includes a sally at Maecenas himself
(20 n.); the allusion to Lydian latifundia is a friendly gibe at the des-
cendant of Etruscan kings (41–2 n.); the final reference to the god’s
measured generosity is a deft compliment to Maecenas’ practice. If the
ode is seen as a comment on a personal relationship within a vanished
social framework, it is not so trite or so smug as is sometimes supposed.
For the expression of thanks to a benefactor (eucharisticon) see further
2.18 (with N–H, p. 290), Cairns 74 f., I. M. Le M. DuQuesnay, PLLS 3,
1981: 97 ff.
1 6 . I N C LV S A M DA NA EN 201
The ode begins the second half of the book (see G. B. Conte on
‘delayed proems’ in YCS 29, 1992: 147 ff.). The same central position is
occupied by other poems to Maecenas (epod. 9, serm. 1. 6, carm. 1. 20); in
all these books Maecenas is also addressed in the opening poem. In the
present ode the praise of a quiet life on the Sabine estate (29 ff.) recalls
the conclusion of 3.1 ‘cur invidendis postibus et novo / sublime ritu
moliar atrium? / cur valle permutem Sabina / divitias operosiores?’
Metre: quatrains of three Asclepiads followed by a Glyconic.
1. Inclusam Danaen: when Acrisius was told by an oracle that he would

be killed by his grandson, he imprisoned his daughter Danae in a bronze
chamber; but Zeus descended to her in a shower of gold, and she gave
birth to Perseus. See Apollod. bibl. 2. 4. 1 with Frazer, RE 4. 2084,
LIMC 3. 1. 327 ff., 3. 2. 243 ff., CMA 1. 319 ff. (remarking that in medieval
interpretations Danae became a symbol of chastity and her story prefig-
ured the annunciation). Originally the myth probably suggested the
divine semen (see the illustrations in LIMC); for later rationalizations
see 8 n.
For inclusam cf. Men. Sam. 591 ŒÆŁØæª (also of Danae); for non-
mythical women cf. Call. fr. 401. 1 ÆE ŒÆŒºØ , Prop. 3. 14. 23
‘clausae tutela puellae’, Ov. am. 1. 4. 61 ‘nocte vir includet’. The epithets
in the first three lines are all placed next to their nouns, contrary to H’s
usual practice; they reinforce one another to emphasize the security of
Danae’s imprisonment (West, op. cit.).
turris aenea: for Danae’s bronze prison cf. Soph. Ant. 945 K
ÆºŒØ ÆPºÆE , Apollod. bibl. 2. 4. 1, Pausanias (2. 23. 7), who claims
to have inspected the site at Argos. The tower seems to appear for the
first time here, but is unlikely to have been H’s invention; it may be
relevant that æª could be used for the women’s part of the house.
For brazen structures cf. 3. 3. 65 ‘murus aeneus’ with note, 3. 30. 1 n.,
Prop. 2. 32. 59 ‘aerato . . . muro’ (again of Danae).
2. robustaeque fores: the doors are not just ‘strong’ but more literally
‘oaken’; the adjective balances aenea (1); cf. 1. 3. 9 ‘robur et aes triplex’
with N–H. The noun robur was applied to the Tullianum at Rome; cf.
2. 13. 18 f. with N–H, Plaut. Curc. 692 ‘in robusto carcere’.
2–3. et vigilum canum / tristes excubiae: dogs are regularly mentioned

as an obstacle to the lover; cf. epod. 5. 57 f. ‘senem quod omnes rideant
adulterum / latrant Suburanae canes’, Ar. Thesm. 416 f. 1ºØŒf =
æıØ æºıŒEÆ E Ø
E ŒÆ , Ov. am. 2. 19. 40 with
McKeown. The power of money could surmount such barriers; cf.
Tib. 2. 4. 33 f. ‘nam pretium si grande feras, custodia victa est / nec
prohibent claves et tacet ipse canis’, Antip. Thess. anth. Pal. 5. 30. 3 f. j
b ªaæ e
æÆªÆ æfi , º , h Łıæøæe = K , h Œø K
æŁæØ ÆØ. excubiae, like vigilum, is properly used of a human
picket (Ov. am. 1. 6. 7); a contrast can be heard with the prefix of inclusam.
The word tristes, which is not normally applied to dogs, suggests the grim
aspect of a human concierge (3. 14. 23 n.). For watch-dogs see further
Toynbee, 1973: 107 ff. with pl. 43 (the familiar mosaic from Pompeii).
3. munierant satis: the indicative, like the off-hand satis, underlines

Acrisius’ assurance; what appears to be a statement of fact is then
falsified by si non risissent. In such conditional sentences the apodosis
regularly precedes the protasis (2. 17. 28 with N–H). For the theme cf.
Apollod. Caryst. 6. 1 ff. (PCG 2, p. 491) ŒÆd ŒºŁ ŁæÆ
ºE . (B)
Iºº Pb x = Œø O
ıæa oø K ŁæÆ, = Ø w ªÆºB ŒÆd
Ø
e PŒ Næ
ÆØ.
4. nocturnis ab adulteris: adulter, like moechus, does not always imply

a married accomplice; cf. 1. 33. 9, 1. 36. 19, 1. 25. 9. Here the word
points forward to Jupiter himself; for the cynical treatment cf. Plautus’
Amphitruo.
5–6. si non Acrisium virginis abditae / custodem pavidum: the ‘jailer’

is a derisive way of describing Acrisius himself; pavidum makes a
paradox, as normally the prisoner is afraid of the custos. Traditionally
an oracle is said to account for Acrisius’ fear (see n. 1 above), but here the
adjective may simply describe the fussy father of the fille mal gardée; the
very name ‘Acrisius’ suggests indecision and lack of judgment.
6–7. Iuppiter et Venus / risissent: Venus’ traditional smile (Hom. Il.

3. 424 etc., Pease on Virg. Aen. 4. 128) often turns to the mockery of
lovers’ ways, while Jupiter laughs at lovers’ perjuries (for both deities see
N–H on 2. 8. 13); but in our passage the gods’ cynicism is diverted from
mortals’ inconstancy to their corruptibility. Bentley proposed risisset, as
H normally uses a singular verb when two singular subjects are com-
bined (cf. 2. 11. 2, 2. 13. 38, 3. 3. 10 n., 3. 11. 22, C. O. Brink PCPS 28, 1982:
30 ff.), and his conjecture is supported by the scholium on Stat. Theb. 6.
265 (287), which reads the corrupt misisset in quoting our passage.
Exceptions to the rule seem to have particular explanations, and the
singular may well be right.The humorous climax is helped by the
position of the verb and the pause which follows it.
7. fore enim tutum iter et patens: tutum points back to the fierce dogs
and patens to the stout doors, thus forming a chiasmus. A verb of
thinking is understood with fore; though the construction is commoner
in the historians than in the poets, it is found e.g. at Virg. Aen. 1. 444 f.,
Ov. met. 1. 250.
8. converso in pretium deo: dative after tutum and patens rather than
ablative absolute. The amusingly cynical in pretium goes beyond the
expected in aurum; cf. Ovid’s imitation ‘Iuppiter admonitus nihil esse
potentius auro / corruptae pretium virginis ipse fuit’ (am. 3. 8. 29 f.).
The rationalization of the golden shower goes back at least to the
Danae of Euripides (TGF fr. 324 t
æı, øÆ ŒººØ æE ).
Treatments of the tale in New Comedy are found in Men. Sam. 590 f.
‰ ª
æıe › ˘f Kææ = Øa ªı ŒÆŁØæª ÆE
K
ı , Ter. eun. 584 ff. with Donatus. A Menippean narrative
lies behind Lucian, gall. 13 (Zeus) PŒ
ø . . . ‹ø i ØÆŁæØ F
#ŒæØı c æıæ---IŒØ ı ‰
æı Kª (cf. R. Helm,
Lucian und Menipp, 1906: p. 325). The Greek epigrammatists blame
Danae’s own avarice; e.g. Antip. Thess. anth. Pal. 5. 31. 5 f. Œø ‹Ø
ŒÆd ˜Æfi ˘f = P
æı ,
æıF qºŁ æø ŒÆ , Parmenio,
ibid. 5. 33 and 34, Strato 12. 239. 2. H implies that guards were bribed (as
in Lucian) rather than Danae herself; the former follows better from
‘tutum iter et patens’ and leads better to ‘per medios ire satellites’.
9. aurum per medios ire satellites: though gold is softer than bronze, it
turns out to be stronger; cf. Asclepiades, anth. Pal. 5. 64. 6 Øa
ÆºŒø
æıe ı ŁÆºø, Paul. Sil. ibid. 5. 217, Apul. met. 9. 18. 2. For
similar aphorisms on the power of money cf. Soph. Ant. 296 f. F ŒÆd
ºØ = æŁE, ¼æÆ KÆØ ø, TGF adesp. 129, Cic.
Verr. 1. 2. 4 ‘nihil esse tam sanctum quod non violari, nihil tam munitum
quod non pecunia expugnari possit’, Otto 50. satellites describes the
henchmen at a despot’s court.
10–11. et perrumpere amat saxa potentius / ictu fulmineo: in a world

ignorant of electricity, and without lightning-conductors (which were
not invented until the eighteenth century), lightning was a terrifying
thing; this fear can be seen in the practice of railing off the structure
which had been hit and declaring it a consecrated site (see ars 471 ‘triste
bidental’ with Porph.’s note). It was a truism that lightning could cleave
rocks (Lucr. 6. 229 ‘transit per saxa’), or as here man-made fortifications
(ibid. 240 ‘potest ictu discludere turris’). amat is often translated ‘is
accustomed’ (ØºE), but it seems to include a hint of capriciousness:
perhaps ‘has a way of ’. For potentius cf. Diphilus, fr. 103 (PCG 5, p. 111)
N
ıæ æ Œæø e
æØ º = a Æ fiø ÆØ ŒÆd
æÆØ, Ov. am. 3. 8. 29 (cited 8 n.). fulmineo recalls Jupiter’s thun-
derbolt, and so follows well from 8; the use of the adjective for the
genitive is poetical (Lucr. 2. 382, cf. 3. 1. 7, 3. 4. 72 n.).
11–12. concidit auguris / Argivi domus: here H varies his exempla
by describing the victim of bribery. Amphiaraus joined the ‘Seven
against Thebes’ under the persuasion of his wife Eriphyle, who had
been bribed by Polynices with a golden necklace. The episode was
mentioned in passing by Homer (Od. 11. 326 f., 15. 244 ff.) and was the
subject of a poem by Stesichorus and plays by Sophocles and Accius. It
was often used as a moral object-lesson (Plato, rep. 590a, Cic. Verr. 4. 39,
Prop. 2. 16. 29, 3. 13. 57 f., Stat. Theb. 4. 193 ff., Hense on Musonius 4
p. 14). See further Diod. 4. 65 f., Apollod. bibl. 3. 6. 2 with Frazer,
RE 6. 460 ff., LIMC (Eriphyle) 3. 1. 843 ff., 3. 2. 606 ff. The crime led
to the deaths of Amphiaraus (see next note), Eriphyle, and ultimately
their son, Alcmaeon, who had taken revenge on his mother. concidit
domus refers to the whole family (as at 1. 6. 8 ‘saevam Pelopis domum’),
but the image also suggests the collapse of a building; in this meta-
phorical sense it balances perrumpere (10), diffidit (13), subruit (14).
Amphiaraus was an augur or Ø (cf. 1. 2. 32 ‘augur Apollo’) who
on this occasion foresaw his own death (cf. Cic. fam. 6. 6. 6, Stat.
Theb. 3. 546 ff.).
12–13. ob lucrum / demersa exitio: ob lucrum should be taken at least

partly with demersa exitio, which otherwise seems a weak appendage.
Though (de)mergere is used elsewhere for ‘to ruin’, it recalls here how
Amphiaraus and his chariot were swallowed up by the earth; cf. Apol-
lod. 3. 6. 8, Pind. O. 6. 13 ff., N. 9. 24 ff. › #ØÆæE
ŒæÆıfiH
Æfi Æ = ˘f a ÆŁæ
Ł Æ, Œæł – ¥ Ø (for the
thunderbolt cf. ictu fulmineo above), Stat. Theb. 7. 794 ff., 9. 646 f.
‘demersi funera . . . / auguris’, Apollod. 3. 6. 8 with Frazer, RE 1. 1891,
LIMC 1. 1. 698 ff., 1. 2. 560 ff. exitio is ablative; cf. Virg. Aen. 6. 429
‘funere mersit acerbo’, Housman, Classical Papers, 3. 915 f.
13–14. diffidit urbium / portas vir Macedo: the delayed muneribus (15)
shows that diffidit is metaphorical. H is referring to Philip II, father of
Alexander the Great; the periphrasis vir Macedo, which balances auguris
Argivi (11–12), alludes to the hostile use of the phrase 1ÆŒg Iæ in
Dem. Phil. 1. 10. His capture of the Thracian coastal cities, and espe-
cially Olynthus (348 bc), was attributed in part to bribery; cf. Dem. fals.
leg. 265–8, Cic. Att. 1. 16. 12 ‘Philippus omnia castella expugnari posse
dicebat in quae modo asellus onustus auro posset ascendere’, Diod. 16.
53. 2, Val. Max. 7. 2. ext. 10, Plut. Aem. Paul. 12. 6, Juv. 12. 47 ‘callidus
emptor Olynthi’ with Mayor, paroem. Gr. 1. 209 (an oracle to Philip)
IæªıæÆE º ª
ÆØ
ı, ŒÆd ø ŒæÆØ . Of course Philip’s mas-
tery of siege-warfare was the most important factor; cf. G. Cawkwell,
Philip of Macedon, 1978: 160 ff., N. G. L. Hammond and G. T. Griffith,
A History of Macedonia 2, 1979: 444 ff.
14–15. et subruit aemulos / reges muneribus: subruit, ‘undermined’,
keeps up the military imagery; tunnels played an important part in
ancient siege-warfare. The particular reference is uncertain, but aemulos
reges probably refers to rival Macedonian princes (OLD s.v. rex) rather
than external enemies. For instance, about 359 bc Philip bribed
the Paeonians to thwart Argaeus and Pausanias; cf. Diod. 16. 3. 4,
Hammond and Griffith, op. cit. 2. 210 f. It may also be relevant
that by one account he destroyed Olynthus so as to capture his rival
half-brothers Arrhidaeus and Menelaus ( Justin 8. 3. 10–12, CAH edn.
2, 6. 748 ff.); H’s muneribus connects the subversion of princes with the
destruction of cities.
15–16. munera navium / saevos inlaqueant duces: H turns from land to

sea to show the universal sway of money; for the collocation of muner-
ibus and munera in successive clauses cf. 3. 5. 21 n., Wills 271 ff. He must
be referring to a particular case, like those of Amphiaraus and Philip
above; in view of the present inlaqueant and the omission of the name,
this must be a fairly recent instance. Editors plausibly refer to the
freedman Menas, later called Menodorus, who in the naval war of 38
deserted Sex. Pompeius and was rewarded by Octavian with gold rings
and equestrian status (App. civ. 5. 80, Dio 48. 45. 7). Conceivably
Maecenas played a part in this transaction, for his reputation as a fixer
was high at this time. To give the reference more precision RN has
considered reading servos duces (a sneer at Menas’ lowly origin); but
though this would produce an effective oxymoron, saevos combines well
with inlaqueant to suggest the trapping of a wild beast.
17. crescentem sequitur cura pecuniam: as in 3. 1. 48 ‘divitias opero-

siores’, H. refers to the complications that wealth brings; he follows
Epicurus in regarding such anxieties as unnecessary (KD XV, frr. 468–79
Usener, Lucr. 5. 1118 f.; so already Democritus B283–4 ). pecunia is a
prosaic word (Axelson 108); here it has a derogatory sense that suits the
moralizing tone of the passage (see 3. 24. 61, 4. 9. 38).
18. maiorumque fames: for the neuter plural in gnomic utterances

cf. 2. 16. 26, epist. 1. 10. 32 ‘fuge magna’, Alc. 38. 4 L–P, Theoc. 16. 65,
Paul, Romans 12: 16. As a man becomes richer he longs for greater
things, a thought that leads to ‘tollere verticem’ (19). This is a develop-
ment of the commonplace that the love of money grows with what it
feeds on; cf. 2. 2. 13 ‘crescit indulgens sibi dirus hydrops’ with N–H,
epist. 2. 2. 146 f., Sen. ben. 2. 27. 3, Juv. 14. 139 with Mayor. For the
metaphorical fames see epist. 1. 18. 23, Virg. Aen. 3. 57 ‘auri sacra fames’
(so also ØA, ‘to crave for’); sitis is more common (2. 2. 14, serm.
1. 1. 68 etc.).
18–19. iure perhorrui / late conspicuum tollere verticem: H’s horror is
humorously exaggerated; yet it is true that at some stage he declined to
become Augustus’ private secretary (Suet. vita). The high head is a sign
of both eminence and pride; cf. 1. 1. 36 ‘sublimi feriam sidera vertice’
(also addressed to Maecenas), 1. 18. 15 with N–H. conspicuum is proleptic;
here it is combined with late (cf. 3. 3. 45 n.), which makes a contrast of
dimensions with tollere.
20. Maecenas, equitum decus: a discreet reminder that Maecenas

claimed to keep a low profile; cf. serm. 1. 6. 98 ‘iudicio . . . sanus fortasse
tuo’, ps.-Acro ‘Maecenas enim eques Romanus fuit nec voluit transire in
ordinem senatorium’, eleg. in Maec. 1. 31 ff., Vell. 2. 88. 2, Dio 55. 7. 4; he
preferred to exercise influence from a private station rather than face the
hazards of elections. There is humour in using so rich and magnificent a
person as an example of moderation; H is imitated by Prop. 3. 9. 2 ‘intra
fortunam qui cupis esse tuam’, 22 ‘cogor et exemplis te superare tuis’. At
the same time decus is a term of panegyric that H applies elsewhere
to Maecenas (1. 1. 2, 2. 17. 4); it was tactful to note that he was no
ordinary equestrian.
21–2. quanto quisque sibi plura negaverit / ab dis plura feret: H now
turns from the anxieties brought by cupidity to the positive advantages
of self-denial. It was a philosophical paradox that to limit one’s desires is
in itself to increase one’s wealth; cf. 2. 2. 9 ff. ‘latius regnes avidum
domando / spiritum quam si Libyam remotis / Gadibus iungas’ (see
N–H for Epicurean parallels), Cic. parad. 51 ‘non esse cupidum pecunia
est, non esse emacem vectigal est; contentum vero suis rebus esse
maximae sunt certissimaeque divitiae’, Sen. epist. 14. 17. The preposition
ab is often combined with dis even in Cicero, perhaps giving a slightly
archaic effect ( J. C. Rolfe, ALL 10, 1898: 468).
22–3. nil cupientium / nudus castra peto: a particular consequence of

the preceding maxim; H sometimes dispenses with conjunctions to
connect his argument. For the philosophic nil cupere cf. Sen. Thy. 389
‘rex est qui cupiet nihil’, TLL 4. 1430. 48 ff. In the terms of the metaphor
nudus means ‘without weapons’ (see next note), but at the literal level it
refers to the renunciation of superfluous possessions.
23–4. et transfuga divitum / partis linquere gestio: a humorous

reversal of the philosophers’ metaphor by which ‘to desert the camp of
virtue’ is disgraceful; cf. epist. 1. 16. 67 ‘perdidit arma, locum virtutis
deseruit’, Cic. fam. 9. 20. 1 ‘in Epicuri nos, adversarii nostri, castra
coiecimus’, Sen. epist. 2. 5, dial. 12. 5. 2. To amuse his benefactor
H suggests that he himself belongs to the camp of the rich; vis-à-vis
the pauperiorum turba, this is not unrealistic, but in comparison with
Maecenas and the very rich H was no more than comfortably off. gestio
suggests physical excitement (3. 3. 54, Austin on Cic. Cael. 67); it is
applied paradoxically to self-denial, which is presented as a fulfilment
of desire.
25. contemptae dominus splendidior rei: ‘a more glorious master of the

wealth that I reject’; cf. Cowley’s imitation: ‘Slaves to the things we too
much prize, / We masters grow of all that we despise’. dominus is used
in two senses: he is more in control of the wealth that he rejects than
if he were an owner of large tracts of land; cf. serm. 2. 7. 75, 93, Cic.
Stoic. parad. 5. 33, Min. Fel. 36. 5 ‘omnia si non concupiscimus, possi-
demus’, Maxim. eleg. 1. 54 ‘et rerum dominus nil cupiendo fui’. For the
use of contemnere cf. Virg. Aen. 8. 364 ‘aude, hospes, contemnere
opes’, Sen. epist. 62. 3 ‘brevissima ad divitias per contemptum divitiarum
via est’. Most editors understand contemptae rei as ‘the property despised
by others’; that is a less pointed interpretation of this condensed
passage, it fails to take proper account of the topos cited above, and
such a description of the Sabine estate might well be discourteous
to Maecenas.
26. quam si quidquid arat impiger Apulus: arat here means ‘produces
by tilling’, i.e. ‘harvests’, as occultare horreis shows (cf. Cic. Verr. 3. 113 ‘ut
plus quam decem medimna ex iugero ararent’, Sen. clem. 1. 6. 1, TLL
3. 627. 24 ff.); in other places it indicates the extent of a landowner’s
property, e.g. Virg. georg. 2. 224–5, Persius 4. 26 ‘dives arat Curibus
quantum non miluus errat’. For the lengthening of -at before a vowel
see 2. 13. 16 with N–H, 3. 5. 17 n. The hyperbolical quidquid suits
descriptions of latifundia (cf. below 31 n.). The fertility of Apulia is
mentioned by Varro, rust. 1. 2. 6 ‘quod triticum Apulo (conferam)?’,
1. 57. 3, Strab. 6. 3. 9, Brunt (1971), 368 f. impiger Apulus is a tribute to H’s
hard-working fellow-countrymen; cf. 3. 15. 1 n., Lucan 5. 403 ‘piger
Apulus’ (a paradoxical result of war).
27. occultare meis dicerer horreis: the large landowner is a hoarder who
does not understand that wealth is for use (serm. 1. 1. 44 ‘quid habet
pulchri constructus acervus?’, epist. 2. 2. 177 ff.). By a paradox his hidden
riches can only be a matter for report (dicerer), while the man who
rejects wealth enjoys a brighter lustre (splendidior in 25). meis in this
emphatic position brings out the pride of possession.
28. magnas inter opes inops: the rich man is ‘without resources’ be-
cause he does not use his wealth; the epigram is the counterpart of v. 25.
For similar paradoxes cf. serm. 2. 3. 142 ‘pauper Opimius (suggesting
opimus) argenti positi intus et auri’, Sen. epist. 74. 4 ‘in divitiis inopes’,
dial. 12. 11. 4, Hf 168, Otto 51.
29–30. purae rivus aquae silvaque iugerum / paucorum: for similar

descriptions of H’s estate see serm. 2. 6. 1 ff. ‘modus agri non ita magnus,
/ hortus ubi et tecto vicinus iugis aquae fons, / et paulum silvae super his
foret’, epist. 1. 16. 9 ff., 1. 18. 104 ‘gelidus Digentia rivus’. Ps.-Acro well
comments on our passage ‘laus mediocritatis sub agelli descriptione’; the
stream of pure water belongs to the traditional blessings of simplicity
(serm. 1. 1. 54 ff., 2. 11. 20 with N–H, Lucr. 2. 30), while ‘iugerum
paucorum’ suggests antique frugality (2. 15. 1, Virg. georg. 4. 127 f. ‘cui
pauca relicti / iugera certa soli’).
30. et segetis certa fides meae: cf. Cic. fam. 16. 17. 1 ‘ager etiam fidelis dici
potest’. In RN’s view certa presents a problem: most farming was unreli-
able (cf. 3. 1. 30 ‘fundusque mendax’ with note), and one does not expect
exaggerated claims for the fertility of H’s estate; cf. epist. 1. 14. 26 ff. and
(of Mena’s Sabine property) 1. 7. 87 ff. ‘spem mentita seges’; on the other
hand, a reference here to spiritual benefits impairs the contrast with
beatior below. RN has therefore considered curta fides, ‘defective grati-
tude’ (cf. Juv. 14. 166 f. ‘ingratae / curta fides patriae’); this takes one step
further the unassuming tone of rivus aquae and iugerum paucorum; the
phrase might be defended as a tease of Maecenas that is heavily out-
weighed by the following eulogy. In NR’s view harsh comments about the
land (3. 1. 30 etc.) and friendly ones (serm. 2. 6. 1 ff. etc.) are a matter of
mood (contrast Virg. georg. 1. 199 ff. with 2. 458 ff.); here meae, which
seems to stress the pride of ownership (cf. 27), suits a complimentary
adjective like certa, and fallit (32) makes a typical verbal contrast with fides.
31. fulgentem imperio fertilis Africae: this means the imaginary owner
of impossibly large estates in the province of Africa, the modern Tunisia
(cf. OCD 34); the fertility of the area is referred to in serm. 2. 3. 87
‘frumenti quantum metit Africa’, Juv. 8. 117 f. ‘messoribus illis / qui
saturant Vrbem’ with Mayor. He is described hyperbolically as ruling
over the country; cf. 2. 2. 9 ff. (quoted on 21–2 above). H does not have
in mind a real contemporary king (for Africa was a Roman province),
nor yet a Roman proconsul (for the genitive Africae suggests a personal
domain, and the fertility of the land would not enhance a proconsul’s
magnificence). fulgentem suits the imaginary monarch and balances
splendidior (25); for the ablative imperio cf. 3. 2. 18 ‘(virtus) intaminatis
fulget honoribus’, Sil. 13. 605, Juv. 8. 42. Bentley proposed fulgente
imperio, looking for an ablative of comparison with beatior; but this
can easily be understood, and it is confusing to combine beatior with two
ablatives of different categories.
32. fallit sorte beatior: the poet’s modest property ‘escapes the notice’ of
the big landowner ‘as being a more blessed allocation’. In Greek one
could say ºÆŁø OºØæ þ ‘I escape your notice being more
fortunate’, but esse lacks a present participle, and so the construction is
less clear; cf. Virg. Aen. 12. 634 ‘nequiquam fallis dea’, TLL 6. 1. 189. 79 ff.,
S. Eklund, The Periphrastic, Completive and Finite Use of the Present
Participle in Latin, 1970: 89 ff. (who underestimates the Greek influ-
ence). beatus can be used of things that bring blessedness, as in 1. 29. 1 f.
‘beatis . . . / gazis’. sorte is an ablative of respect, common with words like
felix; by an odd extension it is used here where the subject is not the
possessor of the lot but the actual blessing itself.
33. quamquam nec Calabrae mella ferunt apes: in ancient times

Calabria was in south-east Italy; speaking of Tarentum, H praises its
honey, olives, and vines (2. 6. 14 ff.); Virgil mentions the abundance of
flowers from which the old bee-keeper profited (georg. 4. 109 ff.);
Matine bees (4. 2. 27) belonged to the same district (N–H on 1. 28. 3);
see also Varro ap. Macrob. sat. 3. 16. 12. For the plural applied to
substances cf. K–S 1. 73 ff., Wackernagel 1. 96 f.; mella is metrically
convenient and is found in the satires (2. 2. 15, 2. 4. 24) as well as the
odes. Similar lists of dispensable luxuries occur in 2. 18. 1 ff. and 3. 1. 41 ff.
34–5. nec Laestrygonia Bacchus in amphora / languescit: the Homeric

Laestrygones (Od. 10. 80 ff.) were traditionally located at Formiae, on
the coast of South Latium; cf. 3. 17. 6 n. The adjective is humorously
grandiose: the Laestrygones were cannibals. It may be relevant that
Maecenas’ wife, Terentia, seems to have had roots in the area; a
Murena, perhaps her brother, is attested there at serm. 1. 5. 38 (cf.
below 3. 19. 7 n.). Formian wine is mentioned in 1. 20. 11, Athen. 26e;
for the transferred epithet cf. 1. 9. 7 f. ‘Sabina / . . . diota’, 1. 31. 9 ‘Calena
falce’; there were no vines on H’s estate (epist. 1. 14. 23). languescit refers
to mellowing with age (cf. 3. 21. 8).
35–6. mihi nec pinguia Gallicis / crescunt vellera pascuis: mihi is

emphatic (¼ K) as in 2. 16. 37 ‘mihi parva rura’. So it comes at the
beginning of the clause (preceded by a comma), and should be taken
primarily with crescunt, though it has to be understood also with the two
previous verbs. If it were combined primarily with languescit, it would be
unemphatic and would be oddly placed at the end of its clause.
For the wool of Cisalpine Gaul cf. Strabo 5. 1. 12 (Padua), Plin. nat.
hist. 8. 190 ‘alba (lana) circumpadanis nulla praefertur’, Colum. 7. 2. 3,
Blümner (1911), 238 f. pinguis is used of thick fleeces in Claud. 20. 384;
the word is also applied, sometimes pejoratively, to thick clothes (Mart.
6. 11. 7 ‘te Cadmea Tyros, me pinguis Gallia vestit’, Juv. 9. 28, Suet. Aug.
82 ‘cum pingui toga’) and luxuriant hair (Suet. Ner. 20). crescunt implies
physical growth (cf. 3. 23. 11, Sen. dial. 10. 12. 3 of a barber ‘dum
decerpitur si quid proxima nocte succrevit’) with the hint of a capital
increment, as in ‘crescentem . . . pecuniam’ (17).
37. importuna tamen pauperies abest: pauperies is a relative term; it is

often used by H to mean ‘modest circumstances’ in contrast to the great
wealth possessed by many of his acquaintances (e.g. 2. 18. 10, 3. 29. 56),
but here it describes actual distress (cf. epist. 2. 2. 199 ‘pauperies
immunda’). importuna refers to the insistent demands of the body, as
in Plaut. St. 387 ‘importunam exigere ex utero famem’, Lucr. 2. 17 ‘nil
aliud sibi naturam latrare’; it leads to the idea of nagging one’s patron
(cf. epist. 1. 17. 46 ff.) that is mentioned in the following line (deneges). In
the context of possessions abest normally refers to material deficiencies
(3. 24. 64, Lucr. 3. 957 ‘sed quia semper aves quod abest, praesentia
temnis’), but here it is used paradoxically: what H lacks is poverty.
38. nec si plura velim, tu dare deneges: H declines further enrichment

in a way that shows gratitude for past benefits; cf. 2. 18. 12 ff., epod. 1. 31 f.,
serm. 2. 6. 4.
39–40. contracto melius parva cupidine / vectigalia porrigam: H re-

sumes the theme of 21 f. (see note); add Epicur. fr. 135 Usener ¼
C28 Bailey N ºØ ºØ —ıŁŒºÆ ØBÆØ, c
æø
æŁØ, B b KØŁıÆ IÆæØ (translated by Sen. epist. 21. 7).
melius qualifies the whole clause (‘I shall do better to . . . ’); cf. 1. 2. 22
‘melius perirent’ with N–H, 3. 15. 8 (rectius), K–S 1. 795, H–Sz 827.
cupidine is ‘avarice’ (for the metrically impossible cupiditate); the word is
always masculine in H.
vectigalia refers to private rents from the Sabine estate (epist. 1. 14. 2 f.);
cf. Cic. Att. 12. 19. 1 ‘nihil egeo vectigalibus et parvo contentus esse
possum’, rep. 4. 7, parad. 49 ‘non intellegunt homines quam magnum
vectigal sit parsimonia . . . ex meo tenui vectigali, detractis sumptibus
cupiditatis, aliquid etiam redundabit’. Of course the word normally
refers to public revenues (Porph. ad loc.), particularly from Asia
Minor; so it makes a good contrast with the kingdom of Alyattes
below (cf. serm. 2. 2. 100 f. ‘ego vectigalia magna / divitiasque habeo
tribus amplas regibus’). porrigam (‘extend’) forms a contrast with con-
tracto (the two words frame the sentence); it also suits the expansion of
property (cf. Sen. epist. 89. 20 ‘quo usque arationes vestras porrigetis?’)
and so leads well to 41 ff.
41–2. quam si Mygdoniis regnum Alyattei / campis continuem:

Mygdon was a legendary king of Phrygia (Hom. Il. 3. 182 ff.);
1 7 . A ELI V E T V S T O N O B I LI S A B L A M O 211
cf. 2. 12. 22 ‘aut pinguis Phrygiae Mygdonias opes’. Alyattes was a
sixth-century king of Lydia, the father of Croesus; for the genitive
Alyattei (a five-syllable name), restored by Faber for aliat(t)hii or
halyattici, cf. 1. 6. 7 Vlixei, 1. 15. 34 Achillei. It is significant that Maecenas
could claim a Lydian origin (serm. 1. 6. 1; cf. N–H on 2. 18. 6). For
continuare cf. Cic. leg. agr. 2. 70 ‘Sullanus ager . . . latissime continua-
tus’, 2. 78, 3. 14, Sall. Cat. 20. 11; the whole expression continues the
hyperbole of 26 and 31; cf. epist. 2. 2. 178. For similar rejections of wealth
cf. N–H on 2. 12. 21, Lucil. 671 f. with Marx, Alphaeus, anth. Pal. 9.
110. 1 f. ˇP æªø ÆŁıº.ı IææÆ , = PŒ Zº º
æı, x Æ
ˆª . . . Phrygia was associated with Midas, and Lydia with silver
coinage and gold-bearing rivers, but here H is thinking of agricul-
tural wealth.
42–3. multa petentibus / desunt multa: the chiastic repetition gives

epigrammatic form to the Stoic commonplace; cf. 25 n., Cic. paradox.
44 ‘qui . . . innumerabiles cupiditates habet . . . hunc quando appellabo
divitem, cum ipse egere se sentiat?’, Sen. epist. 90. 38, 108. 9 quoting
Publil. 1.7 (¼ Min. Lat. Poets, Loeb edn. 275) ‘desunt inopiae multa,
avaritiae omnia’.
43–4. bene est cui deus obtulit / parca quod satis est manu: bene est is
an expression of contentment; cf. serm. 2. 6. 4 ‘bene est; nil amplius oro’,
OLD 8 b. The omission of ei suits a somewhat archaic aphorism; cf. 2. 16.
13 ‘vivitur parvo bene cui etc.’, K–S 2. 281 ff.; the singular cui, in contrast
to the plural petentibus, allows the sententia to point to H himself. parca
manu would normally suggest a lack of generosity, but here paradoxic-
ally is a blessing. After bene est the word satis expresses positive satisfac-
tion; cf. 2. 18. 14 ‘satis beatus unicis Sabinis’, epod. 1. 31 f. ‘satis superque
me benignitas tua / ditavit’.
1 7 . A ELI V E T V S T O N O B I LI S A B L A M O
[L. Mondin in Temi Augustei (ed. G. C. Marrone), 1998: 37 ff; Syme (1986), 394 f.;
S. Treggiari, Phoenix 27, 1973: 246 ff.; T. P. Wiseman, G&R 21, 1974: 153 ff. ¼ Roman
Studies, 1987: 207 ff.]
1–9. Aelius, illustrious descendant of ancient Lamus (for all the Lamiae
derive from the founder of Formiae who held sway as far as Minturnae)—
9–14. Tomorrow there will be a rainstorm, if the prophetic crow can be relied
on; so gather dry firewood while you can. 14–16. Tomorrow you will enjoy a
simple meal with your household, released from work by the weather.
The recipient of this ode, L. Aelius Lamia, belonged to a family of
growing importance; for his place in it see N–H on 1. 26, Treggiari, op.
cit., Syme, loc. cit. His homonymous father, a man of far-reaching
business interests (Syme, 1939: 82 n. 1), became plebeian aedile in 45 bc
and praetor probably in 42; as a leading eques he had rallied support
for Cicero at the time of his exile (Sest. 29, fam. 11. 16. 2), and when
Cicero was killed near Formiae at the end of 43, he used his position as a
local magnate (6 n.) to give his maimed body a funeral (anth. Lat. 2. 608,
611, 614 Riese, H. H. Davis, Phoenix 12, 1958: 174 ff.). Horace’s friend
may be one of the people who acquired patrician status about 29 bc
(Syme, 1939: 382), and at some stage he became a xvvir sacris faciundis
(AE, 1948: n. 93); this suits the antiquarian concerns suggested by vv.
1–8. When Augustus returned from Spain in 24 he left Lamia in
Hispania Tarraconensis as legatus pro praetore (Dio 53. 29. 1, PIR edn.
2, A199, Syme, Roman Papers 2, 1979: 829, 848 f.). Thereafter the man
surprisingly disappears from view, and Syme (1986: 395) suggests that he
might be the Lamia who is mourned by his brother at epist. 1. 14. 6 ff.
Like the other ode to Lamia (1. 26), which is dated to 26–25 by the
reference to Tiridates, our poem must have been composed before the
Spanish appointment.
A son of the legate, another L. Aelius Lamia, rose to be consul in ad
3 and prefect of the city in 32 (Tac. ann. 6. 27. 2, PIR edn. 2, A200), and
it used to be thought that he was the recipient of Horace’s poems. N–H
(loc.cit.), while contesting this view, identified the future consul with
the Lamia of 1. 36. 7–9 (otherwise Treggiari, op. cit.), and also with the
Lamia who mourns for his brother; these suggestions are rejected by
Syme (1986), 394, who thinks that the future consul was born about 32
bc, i.e. too late for him to figure in Horace. The family continued to be
important in the early empire; cf. Juv. 4.154 of Domitian ‘Lamiarum
caede madenti’ with Courtney.
Though his friend was not a nobilis (1 n.), Horace teasingly celebrates
his descent from Lamus, the ruler of Homer’s Laestrygones (1 n.).
Aristocratic Roman families were always very aware of their ancestors,
with whom they identified (Griffin 188 ff.); we have only to think of the
stemmata in the atrium (Courtney on Juv. 8. 1–9) and the masks worn at
great men’s funerals (Flower, 1996). M. Messalla Rufus (cos. 53 bc)
wrote a book de familiis (Plin. nat. hist. 35. 8), Atticus investigated
several genealogies (Nepos, Att. 18. 2–4 with Horsfall), and Brutus had
the resulting tree painted on a wall of his house (Cic. Att. 13. 40. 1); see
Rawson, 1985: 231 f. Apart from this genuine antiquarianism, families
traced their origin to mythical Greek heroes (see especially Wiseman,
op. cit.)—a practice encouraged by the works of Varro and Hyginus
de familiis Troianis (P. Toohey, Arethusa 17, 1984: 1 ff.); thus the Julii
claimed descent from Aeneas, the Antonii from Anto, a son of
Hercules, the Memmii from Mnestheus, the Licinii apparently from
the kings of Argos (see below on 3. 19). For a parallel to our passage
Mondin cites Suet. Vitell. 1. 2–3 ‘extat Q. Elogii ad Quintum Vitellium
Divi Augusti quaestorem libellus, quo continetur Vitellios Fauno Abor-
iginum rege et Vitellia quae multis locis pro numine coleretur ortos,
toto Latio imperasse . . . ’. But needless to say, when Horace connects
Lamia with the man-eating Laestrygones, he does not expect his pon-
derous excursus to be taken seriously (1 n.).
After the disquisition on genealogy (1–9) the ode looks forward to a
modest meal on the following day (cf. 3. 13. 3), not to celebrate a festival
or a birthday (14–15 n.), but because the weather is likely to be wet; in
poems of this kind storm outside makes a contrast with the snugness
within, and sometimes gives a reminder of life’s vicissitudes (cf. 1. 9. 1 ff.
with N–H 116 ff., epod. 13. 1 ff.). The scene is set near the coast, presum-
ably at Lamia’s estate near Formiae (Porph. on 13–16); the wine is not
specified, though the area was famous for its vintages (1. 20. 11, 3. 16. 34).
The main course is to be a sucking piglet (an exception to the tendency
for people to eat in the Satires and drink in the Odes), and the household
slaves participate (16 n.), again a sign of wholesome simplicity. Campbell
exclaimed, ‘From the ancestry of Aelius Lamia to dry faggots and a
sucking pig! What is the point? And where is the poetry?’ (1924: 5); but as
he himself seems to realize (113 f.), the point and the poetry lie precisely
in the contrast between pretentious fantasies about remote ancestors
and the actual pleasure provided by a simple meal consisting of a
recently born piglet.
Aelius Lamia was no doubt as ambitious as the rest of his family, but
he must have had some literary understanding: otherwise the tribute in
1. 26 could not have been addressed to the Muse. Ps.-Acro (on ars 288)
mentions a Lamia who wrote togatae and praetextae (for a mutilated
reference cf. also Fest. 181 M ¼ 192 L); and though this man’s date is
uncertain, he appears in a list of dramatists that also includes Maecenas’
secretary Melissus (Schanz–Hosius 2. 176 f., 292 f.). If the two are
identical, then Lamia has something in common with Fuscus, who
figured in serm. 1. 9. and according to Porph. on epist. 1. 10. 1 also
wrote comedies. However that may be, Horace’s friend seems to have
been the sort of person who could accept friendly banter and perhaps
even a tactful admonition (13–14 n.).
Metre: Alcaic.
1. Aeli, vetusto nobilis ab Lamo: in Roman political life a nobilis was

somebody directly descended in the male line from a consul; cf. D. R.
Shackleton Bailey, AJP 107, 1986: 255 ff. ¼ Selected Classical Papers, 1997:
309 ff., Syme (1986), 50 ff. Aelius Lamia could not claim this distinction,
so H gives him nobilitas in a non-technical sense through his descent
from Lamus; this mythical figure appears in Homer as ruler of the
Laestrygones (Od. 10. 81 f.), who were fancifully located at Formiae,
where the Aelii Lamiae had their estate (for similar cases see 6–9 n.
below). Lamia was a hideous bogy in Greek mythology ( J. Fontenrose,
Python, 1959 and 1980: 100 ff., DNP 6. 1079 f., LIMC 6. 1. 189, 6. 2. 90 f.),
and she is described in one source as queen of the Laestrygones (schol.
Theoc. 15. 40). The cognomen may in fact go back only to the grand-
father of H’s friend, yet another L. Aelius Lamia; he suffered from some
disfigurement, and when L. Crassus the orator sneered ‘audiamus pul-
chellum puerum’, he retorted ‘non potui mihi formam ipse fingere,
ingenium potui’ (Cic. de orat. 2. 262). As the Romans were insensitive
about such things, the combination of disfigurement with residence at
Formiae would have been enough to produce the cognomen; it is not
clear that the family itself claimed descent from Lamia the baby-killer
(ars 340) and Lamus the cannibal king.
Lineage was affirmed on coins and inscriptions and was a standard
topic of encomium; for such compliments at the beginning of a poem cf.
1. 1. 1, 3. 29. 1, Prop. 3. 9. 1, Pind. P. 10. 2 etc. vetustus is relatively rare in
Cicero, but is found in Virgil and Ovid, particularly in solemn contexts,
but also in more mundane passages (epod. 2. 43 ‘vetustis . . . lignis’,
Colum. 3. 11. 4 ‘stercore vetusto’). It may have an archaic ring, like
ante-consonantal ab (3. 5. 43, 3. 16. 22 n.).
2–3. (quando et priores hinc Lamias ferunt / denominatos: in Latin, as

in Greek, a parenthesis sometimes follows a vocative, either to give
reasons for the address (3. 11. 1 n.) or to justify a statement that it
contains; for some poetical uses of the construction see R. J. Tarrant
in Style and Tradition (ed. P. Knox and C. Foss), 1998: 141 ff. Here the
parenthesis continues for half the poem, and is meant to give the
illusion of a precise argument; cf. 4. 4. 18 ff., Catull. 44. 1 ff. ‘o funde
noster seu Sabine seu Tiburs (nam te esse Tiburtem autumant . . . ’).
quando in the sense of ‘since’ contributes to the didactic tone; apart from
instances in Lucretius cf. serm. 2. 5. 9, 2. 6. 93. priores refers to the earliest
Lamiae before there were written records; ferunt suggests an oral trad-
ition, as does dicitur below (6). The prosaic denominatos again suits
the pretence of a serious historical disquisition; cf. Hygin. fab. 127
‘Italus qui Italiam ex suo nomine denominavit’, TLL 5. 1. 535. 41 ff.
hinc means ‘from Lamus’; cf. Virg. Aen. 1. 21 f. ‘hinc populum . . . /
venturum’, 1. 234 f.
3–5. et nepotum / per memores genus omne fastus / auctore ab illo

ducit originem: nepotum refers to the remoter descendants of Lamus,
as opposed to the priores Lamias (a different view is mentioned at the
end of this note). For memores cf. 4. 14. 4 ‘per titulos memoresque
fastus’ (of Augustus); the fasti have an accurate memory in contrast
to the vague tradition implied by ferunt (2). In our passage fastus
(TLL 6. 1. 326. 36 ff.) is as well attested as fastos, and in 4. 14. 4 has
better support (Priscian, GL 2. 256 says he found both forms in MSS
of Horace); it should probably be preferred as an appropriate archaism
and the less obvious reading. The word cannot refer here to the
consular fasti, as no Lamia so far had achieved the consulship, but it
is sometimes used generally for ‘records’; cf. 4. 13. 14 ff. ‘tempora quae
semel / notis condita fastis / inclusit volucris dies’, serm. 1. 3. 112, TLL
6. 1. 328. 57 ff.
auctore refers to the ancestor of the family (i.e. Lamus); cf. 1. 2. 36,
OLD 15. ducit is the conjecture of D. Heinsius, supported by Bentley, for
ducis of the MSS; ‘derives its origin from’ goes well with nepotum genus
omne, which is now the subject. ducis is explained in two different ways,
both implausible. (a) If the parenthesis is closed after fastus (4), the
remarks on Lamia’s ancestry are pointlessly divided; what is worse,
the transition at cras (9) is intolerably abrupt, for there is no connec-
tion between the legendary Lamus and the coming storm. (b) If
the parenthesis is closed after tyrannus (9), quando no longer explains
v. 1 (as one expects), but simply the following clause; this produces the
banal truism ‘since the Lamiae are all descended from Lamus, you are
descended from Lamus’. Moreover, if ducis is retained, nepotum genus
omne has to be an accusative coordinate with priores Lamias; but ferunt
and denominatos suit the earlier members of the family better than
the nepotes.
Shackleton Bailey (1982: 95) has proposed ducet; in his view nepotum
refers to the descendants of H’s addressee. He finds this more natural
than ‘the descendants of the earlier Lamiae’, but probably the meaning
is ‘the descendants of Lamus’; there is a progression from vetusto to
priores and nepotum (all three words are given emphasis by their pos-
ition). H has allowed his friend a kind of nobilitas because of his
ancestors, and it seems an irrelevance to introduce future generations.
SB argues that his conjecture would allow fastus to bear the usual
meaning of ‘consular fasti’, but these would not record a mythical
ancestor; and it blunts the compliment to suggest that Lamia would
have descendants more eminent than himself.
6–9. qui Formiarum moenia dicitur / princeps et innantem Maricae /

litoribus tenuisse Lirim, / late tyrannus): Formiae, a coastal town in the
south of Latium, some 80 miles south-east of Rome, was associated
with Homer’s Laestrygones (1 n., 3. 16. 34, Cic. Att. 2. 13. 2 ‘si vero in
hanc 3ºıº veneris ¸ÆØæıª, Formias dico’, Ov. fast. 4. 69
with Bömer, met. 14. 233 f., Plin. nat. hist. 3. 59, RE 6. 2857). The
grandiose reference to the city walls suits foundation legends; cf. Virg.
Aen. 3. 132 etc., Horsfall 21 in the article cited below. princeps means
‘first’, a word naturally applied to founders as to other originators (3. 30.
13 n.); it does not mean ‘ruler’ here, as that notion is supplied by tyrannus
(9). Many Italian towns were similarly linked with legendary Greeks,
Tibur with Catilus (1. 18. 2), Lanuvium and others with Diomedes
(3. 19. 1 n.), Tusculum with Telegonus (3. 29. 4), Patavium with Antenor
(Virg. Aen. 1. 242 etc.). See further T. P. Wiseman, op. cit. 209 f., N.
Horsfall, Vergilius 35, 1989: 3 ff.; for ŒØ in the Greek world
add F. Cairns, Tibullus, 1979: 68 ff., T. J. Cornell, RAC 12, 1983:
1108 ff., C. Dougherty, The Poetics of Colonization, 1993.
The Liris (now called Garigliano in its lower course) enters the sea
near Minturnae, nine miles east of Formiae. There were extensive
marshes at the mouth (epist. 1. 5. 4 f., ‘palustris / . . . Minturnas’), where
Marius hid from his enemies (Vell. Pat. 2. 19. 2 ‘nudus ac limo obrutus,
oculis tantummodo ac naribus eminentibus, extractus harundineto circa
paludem Maricae’, Plut. Mar. 39. 4); the whole coast in antiquity was
swampier than today (M. Frederiksen, Campania, 1984: 17 ff.). innantem
refers to the inundation of the flood-plain; cf. Manil. 4. 757 (as emended
by Housman) ‘ultimus et sola vos tranans colit Indica Ganges’, Plin. nat.
hist. 5. 54 (of the Nile) ‘fecundus innatat terrae’, Plin. epist. 8. 17. 2 (for a
similar use of
ÆØ see Gow on Theoc. 21. 17 f.). Between Minturnae
and the sea were the shrine and grove of the goddess Marica which
seems still to have existed in Porph.’s day (see his note on 7–8); cf. Liv. 27.
37. 2, Strab. 5. 3. 6, Horsfall on Virg. Aen. 7. 47, Latte 192, DNP 7. 894 f.
The stately late tyrannus is meant to recall Pæf Œæø (Homer, Il.
1. 102 of Agamemnon) and Pæıø (Pind. O. 8. 31 of Poseidon); for
the attachment of the adverb to the noun (here influenced by the Greek
parallels) cf. Virg. Aen. 1. 21 ‘populum late regem’, Plin. epist. 3. 5. 4
‘Germaniae latissime victor’. tyrannus here has a nuance of friendly
irony; but it is used straightforwardly by Virgil of Latinus (Aen. 7. 342)
and by Latinus of Aeneas (ibid. 266). When H extends the sway of
Lamus beyond Formiae, this may suggest that Lamia had his villa near
Minturnae (Heinze).
9–12. cras foliis nemus / multis et alga litus inutili / demissa tempestas
ab Euro / sternet: by nemus H can hardly have in mind the sacred
grove of Marica, where ‘cutting wood’ and ‘carting away anything that
belongs to the grove’ (cf. CIL 1. 2. 366 ¼ ROL 4, p.154) are likely to
have been prohibited. multis in this emphatic position may seem
colourless to moderns; yet cf. 4. 2. 29 f. ‘per laborem / plurimum
circa nemus’ (where Bentley took the adjective with nemus), 4. 11. 4 f.
‘est hederae vis / multa’ (again in an emphatic position); for much
seaweed cf. Hom. Il. 9. 7 (ŒFÆ) ººe b Ææb –ºÆ FŒ
ı.
Wet leaves (and by implication wet firewood) were useless for burning;
seaweed was proverbially useless (for vilior alga cf. serm. 2. 5. 8, Virg.
ecl. 7. 42).
12–13. aquae nisi fallit augur / annosa cornix: a seemly caution is

commonly found with prophecies; cf. epist. 1. 20. 9 ‘quod si non odio
peccantis desipit augur’, Soph. OT 1086, Fraenkel 358 n. 3, J. J. O’Hara,
Death and the Optimistic Prophecy in Vergil’s Aeneid, 1990: 55 f.; how
much actual uncertainty is intended is often hard to decide. For augur
in the sense of ‘prophet’ cf. 3. 16. 11, serm. 2. 5. 22, OLD 2; it may be
relevant that Lamia was a priest (see introduction), and certainly the
crow was important in Roman augury (3. 27. 16 n.). In particular it was
believed to forecast rain by its caw; cf. Euphorion 89 Powell (¼ 93 van
Groningen) ÆØ ‹ ŒæØ Œæ, Theoph. sign. 6. 3. 39,
Arat. phaen. 949 ff., 1022, Lucr. 5. 1084 ff., Virg. georg. 1. 388, Thompson
171 f. aqua of rain is commoner in the plural, but H’s phrase represents
ÆØ ; the word’s prominent position makes a contrast with
aridum below.
For the legend of the crow’s longevity cf. 4. 13. 25, Hes. fr. 304 M–W
(it lives for nine generations of men), Ar. aves 609 with Dunbar (who
gives five to ten years as a typical life-span), Juv. 10. 247 with Courtney,
Thompson 169. Because it is ‘full of years’, it knows a thing or two;
hence in Callimachus it can describe what happened many generations
earlier (Hecale fr. 260. 42 ff. ¼ 73. 13 Hollis). Davis suggests a contrast
with the brevity of human life (1991: 154). The termination of annosus
does not prove that the word is colloquial and prosaic (cf. Brink on epist.
2. 1. 26 with p. 579); in fact it is found mainly in poetry, perhaps
representing ºı .
13–14. dum potes, aridum / compone lignum: the archaic potis has
some MS support and is read by Bentley (the rare word was exposed
to corruption); for potis without est cf. Enn. ann. 164 ‘quis potis ingentes
oras evolvere belli?’ with Skutsch, Virg. Aen. 3. 671, TLL 10. 2. 336. 63 ff.,
but its dignity is perhaps too great for this context. compone means
‘gather’ (OLD 2a), not ‘heap on the fire’ (Lambinus, Orelli); the coming
rainstorm will soak the wood that needs to be dry. The collection of
firewood was part of country life (3. 6. 39–40 n.); we are not expected to
dwell on the fact that an important man like Lamia would have had
ample stores of wood, or that others would have collected it. Here we
have a variation of the carpe diem theme (cf. 2. 11. 16 ‘dum licet’ with
N–H, 4. 12. 36); for those concerned are to take advantage of the day
(cf. epod. 13. 3 f, though in our passage it is the next day); when the storm
is over there will be no more leisure; other variations are found in 1. 7,
1. 9, and 1. 11. Another theme is also seen to be present, namely ‘uti
compositis’ (serm. 2. 3. 109 f., 1. 1. 37 ff., epist. 1. 1. 12 ). Such combinations
are typical of Horace.
14–15. cras Genium mero / curabis et porco bimestri: the Genius was a
man’s guardian spirit (see epist. 2. 2. 187 ff. with Brink, appendix 19); he
can be given offerings as an external being (‘unmixed wine’ suggests a
libation as at Pers. 2. 3 ‘funde merum Genio’), but where the enjoyment
of life is concerned he is almost identified with the individual himself;
cf. epist. 2. 1. 144 ‘(piabant) floribus et vino Genium memorem brevis
aevi’, where RN is tempted to read memores (Porph. ad loc. ‘qui indul-
gent genio suo memores sunt vitam humanam non esse diuturnam’), ars
210, Pers. 5. 151 ‘indulge Genio, carpamus dulcia’, Serv. georg. 1. 302
‘quotiens voluptati operam damus, indulgere dicimur genio’, Mondin,
op. cit. 47. Offerings to the Genius were particularly associated with
birthdays, which one’s Genius was thought to share (Tib. 2. 2. 5,
Pers. 2. 3, Censorinus, de die nat. 2. 2, RE 7. 1143), but that cannot be
the case here (as Heinze points out); for the repetition of cras shows that
the festivity is caused by the bad weather.
curare corpus ‘to look after onself ’ is a common phrase (OLD s.v. curo
1 b), but curare by itself can mean ‘to refresh’ (Enn. ann. 367 ‘vino
curatus’, Liv. 34. 16. 5); the future is not so much a command as a tactful
prediction (‘tomorrow you will be refreshing . . . ’). The Genius was not
normally offered blood-sacrifices (Censorinus, de die nat. 2. 2 citing
Varro ‘ut . . . manum a caede ac sanguine abstinerent, ne die qua ipsi
lucem accepissent aliis demerent’); our passage has been thought to
contradict this (I. M. Le M. Du Quesnay, PLLS 3, 1981: 108 f.), but
perhaps the rule applied only to birthdays. Pigs were not weaned under
two months (Varro, rust. 2. 1. 20, 2. 4. 13) and sucking pigs were regarded
as a great delicacy; cf. Apic. 8. 7. 6 ‘porcellum lacte pastum’. The poem is
set in the autumn when the storm brings down the leaves; a two-
month-old piglet might have been born in July (Colum. 7. 9. 3) or
soon after.
16. cum famulis operum solutis: in vignettes of country life the slaves
appear as members of the familia (cf. epod. 2. 65, serm. 2. 6. 66, Mart.
3. 58. 22); for their part in household cults in both Greece and Rome see
Fraenkel on Aesch. Ag. 1036–8. In the case of regular holidays Cicero
says (leg. 2. 29) ‘feriarum festorumque dierum ratio in liberis requietem
habet litium et iurgiorum, in servis operum et laborum’. Bad weather
also could allow leisure: when Ofellus is operum vacuus, a friend comes
for a meal through the rain (serm. 2. 2. 119 f.), but strict masters might
see such conditions as an opportunity for indoor occupations (Cato, agr.
2. 3, 39. 2 ‘per imbrem in villa quaerito quid fieri possit . . . cogitato, si
nihil fiet, nihilo minus sumptum futurum’, Virg. georg. 1. 259 ff.). solutus
1 8 . FAV N E , N Y M P H A RV M F VG I EN T V M A M AT O R 219
usually takes the ablative (4. 2. 12, epod. 2. 4, serm. 1. 6. 129 ), but
adjectives expressing lack or separation can take the genitive (cf. operum
vacuus cited above, ars 212 ‘liberque laborum’, H–Sz 78); the same is true
of verbs in early Latin (Plaut. rud. 247 ‘ut me omnium iam laborum
levas’, H–Sz 83), but in our passage the construction of ºø may have
had some influence (Löfstedt 2.417).
1 8 . FAV N E , N Y M P H A RV M
F VG I EN T V M A M AT O R
[W. Warde Fowler, The Roman Festivals of the Period of the Republic, 1899: 256 ff.; Pasquali
559 ff.]
1–8. Walk gently through my lands, Faunus, and be kind to the new-born
animals, seeing that a kid is sacrificed to you at the end of every year. 9–16.
At your December festival beasts and men keep holiday in your honour; the
lambs do not fear the wolf, and the peasants dance.
Fauni were spirits of the woods who disturbed the countryside with
their ‘noctivago strepitu ludoque sonanti’ (Lucr. 4. 582); their mysterious
voices were thought prophetic (Skutsch on Enn. ann. 207, Pease on Cic.
nat. deor. 2. 6). Their name was probably derived from favere (Serv.
georg. 1. 10, RE 6. 2057), not fari (Varr. lL 7. 36, cf. Maltby 226); it may
have been used euphemistically of potentially sinister beings (cf. 3–4 n.).
At some stage an individualized Faunus was identified with the pastoral
god Pan, whom the Greeks, with their more visual imagination, por-
trayed in the likeness of a goat; many authorities think that the Fauni
were derived from Faunus on the lines of the Greek Panes, but in the
context of Roman religion it would make sense if the more nebulous
beings came first (note ps-Acro on v. 10 ‘Nonis enim Decembribus
Faunalia, <quae> et Faunorum culta dicebantur’). With their genius
for regularizing superstition the Romans gave Faunus a temple on the
Insula Tiberina (Ov. fast. 2. 193 with Bömer, Steinby 2. 242); the
dedication day was 13 February 194 bc, only two days before the festival
of Lupercus, with whom Pan was sometimes identified (Ov. fast. 2. 267
with Bömer, Plut. Rom. 21. 7). Horace’s poem suggests that the rural
festival on 5 December was much more significant, when Faunus was
honoured for his protection of flocks throughout the past year; he is
given this same function in 1. 17. 12 when he leaves Pan’s habitat in
Arcadia to visit the Sabine estate. See further Wissowa 208 ff., Otto,
RE 6. 2054 ff., Latte 83 ff.
Horace’s ode has some of the formal characteristics of a hymn. The
opening vocative is followed by a phrase in apposition describing an
attribute of the god (1 n.). The summons to Faunus categorizes the
hymn as ‘kletic’ (cf. 1. 30 with N–H, Menander Rhetor 334. 25 ff.,
Murgatroyd on Tib. 2. 5, pp. 164 ff.); words for ‘propitious’ like lenis
(3 n.) and aequus (4) are regular in such hymns. The poet promises an
appropriate sacrifice in return for the god’s protection; for such bargain-
ing see 5 n. It is not specified when the promise is made; the sacrifice to
Faunus at 1. 4. 11 should probably be associated with 13 February, the
date of the urban Faunalia (N–H are too cautious), but in our poem
aprica rura (2) suggests something later in the year. Of the animals
mentioned in the first stanza lambs were born from mid-October to
mid-December and weaned in March (K. D. White, 1970: 305); kids
were normally born in March, so the tener haedus of v. 5 seems to belong
to a later brood (see the introduction to 3. 13).
In the second half of the poem Horace vividly describes the Faunalia
on 5 December, when the god receives his annual reward; the regard for
the precise date is typically Roman, though the rural festival is not
recorded in the official calendars. The stressed tibi (cf. ‘Thy kingdom
come’) keeps up the sacral tone (10 n., 14 n.), but a more realistic cult
hymn would not describe an occasion that was due to occur at some
distance in the future. V. Bartoletti (SIFC 15, 1938: 75 ff.) compares
Sappho 2, a hymn in Sapphics to Aphrodite with a tricolon similar to
Horace’s; there too we meet altars, incense, leaves, and a meadow; but
rather than a specific imitation we should recognize a natural structure
and shared commonplaces (a god summoned to a locus amoenus). Greek
epigrammatists describe dedications to Pan (Leonidas, anth. Pal. 6. 13,
6. 35, etc.), sometimes in return for protection from wolves (Philippus,
ibid. 6. 99), and Theocritean shepherds invoke him (1. 123 ff, 7. 103 ff.);
Longus’ Daphnis and Chloe is described as an offering to Pan and the
Nymphs (praef. 3), and his important part in the love-story may some-
times reflect the influence of Philitas (I. M. Le M. Du Quesnay in
Creative Imitation and Latin Literature, ed. D. West and T. Woodman,
1979: 60). But instead of including such frivolities Horace draws on the
time-honoured rituals of a real fiesta in the Italian countryside; Warde
Fowler, loc. cit., looks beyond antiquarian details and anthropological
speculations to catch the underlying spirit of such occasions. For paral-
lels to Horace’s vignette see Lucr. 5. 1379 ff., Virg. georg. 1. 338 ff. and
2. 380 ff. with Mynors, Tib. 2. 1 with Murgatroyd, Ov. fast. 1. 657 ff. with
Bömer, Calp. Sic. 4. 122 ff., H. Kier, De Laudibus Vitae Rusticae, Diss.
Marburg 1933: 67 ff.
Heinze comments on the elegant balance of the poem, which is very
different from the archaic formulae of genuine Roman cult. The first
pair of stanzas summons Faunus and promises a sacrifice; the second
pair describes the fulfilment of the vow. All four stanzas break into three
clauses: in the first pair these clauses occupy 1 þ 1 21 þ 1 21 , in the more
end-stopped second pair 1 þ 1 þ 2 lines. The word-patterns at the end
of each stanza also correspond: 3 f. ‘parvis / aequus alumnis’ is picked up
by 7 f. ‘multo / fumat odore’, and 12 ‘cum bove pagus’ by 16 ‘ter pede
terram’. In the third stanza two lines on animals are followed by two
lines on people—a pattern repeated in the fourth stanza. We may also
note the rhyming long o’s: pleno . . . anno (5), multo (7), herboso . . . campo
(9), otioso (11). So like many other ‘simple’ poems, this proves to be
carefully crafted.
Metre: Sapphic.
1. Faune, nympharum fugientum amator: in hymns and prayers the

vocative name is often followed by a phrase in apposition that gives a
standing attribute of the god; cf. 1. 10. 1 ‘Mercuri, facunde nepos
Atlantis’ with N–H, 3. 22. 1 (with the order reversed) ‘Montium custos
nemorumque Virgo’, PMG 887, carm. conv. 1 f. t —a #æŒÆÆ ø
ŒºA , = Oæ
a æÆØ OÆb ˝ÆØ , frag. adesp. 936. 1 f. (also
to Pan), Norden (1913), 148. Here Faunus, taking over the goat-like
qualities of Pan, is presented as both nimble (cf. 1. 17. 1 velox) and lustful
(for the tone of amator see 3. 4. 79 n.); cf. Eur. Hel. 187 ff., Ov. met.
1. 691 ff. (Pan and Syrinx) with Bömer, her. 5. 137 f., Ach. Tat. 8. 6. 7 ›
—a s KøŒ ÆPc æ KæøØŒ , Longus 2. 39. 3 ÆÆØ b
P ˜æıØ K
ºH ŒÆd ¯ ØºØ ˝ÆØ Ææ
ø æªÆÆ.
Naturally H’s humorous formulation would be impossible in genuine
cult, but belongs rather to the sensual world of Graeco-Roman litera-
ture and art; moderns deplore the sexual harassment (L. C. Curran,
‘Rape and rape victims in the Metamorphoses’, in Women in the Ancient
World, ed. J. Peradotto and J. P. Sullivan, 1984: 263 ff.), but typically
of the ancient world Porph. emphasizes the lascivia of the nymphs
(he cites Virg. ecl. 3. 65 ‘et fugit ad salices et se cupit ante videri’).
fugientum is an archaic form, metrically convenient for fugientium, the
standard genitive plural of the present participle; cf. also 3. 24. 21 n.
2–3. per meos finis et aprica rura / lenis incedas: meos preceding its
noun is in its stronger position: in a prayer it is natural to underline the
reciprocal relationship of worshipper and god. The possessive should
also be understood with rura, the land that H owns (cf. epod. 2. 3
‘paterna rura bobus exercet suis’). aprica presents an idyllic picture of
the sunny Sabinum in spring or summer (cf. epist. 1. 14. 30, 1. 16. 6 f.).
Kletic hymns naturally contain a word for ‘come’ (1. 2. 30 venias with
N–H, Pulleyn 136 ff., 219), and here incedas is often taken to mean no
more (cf. TLL 7. 1. 856. 31 ff.), but with lenis (¼ leniter) it means ‘walk
gently’, suggesting a contrast with the boisterous pursuit of the
Nymphs; in ancient prayers a god’s manner of approach is sometimes
specified (Fraenkel 204 n., citing such passages as the prayer to Dionysus
in PMG, carm. pop. 871 KºŁE læø ˜Ø ı = . . . fiH fiø d Łıø,
Ar. ran. 326 ff., Catull. 61. 9 f. ‘huc veni, niveo gerens / luteum pede
soccum’). lenis also suggests that Faunus should be gentle to the flocks;
for such adjectives in kletic hymns cf. 1. 19. 16 ‘mactata veniet lenior
hostia’ with N–H, Anacr. PMG 357. 6 f. f Pc = ºŁ ,
Eur. Hec. 538 æıc E ªF, Ar. Thesm. 1148 lŒ
hæ , YºÆØ, Virg. georg. 1. 18 (to Pan) ‘adsis, o Tegeaee, favens’,
Ov. am. 2. 13. 21 (to Ilithyia) ‘lenis ades’, Pulleyn 145.
3–4. abeasque parvis / aequus alumnis: by saying ‘come and go favour-

ably’ H refers to the whole time of Faunus’ sojourn; so ‘te veniente die,
te decedente canebat’ (Virg. georg. 4. 466) implies that Orpheus
mourned Eurydice all day. For similar ‘polar expressions’ (some more
difficult than this) cf. Aesch. Ag. 634 f. H ªaæ ºªØ
ØHÆ ÆıØŒfiH
æÆfiH = KºŁE ºıBÆ ÆØ ø Œ fiø; Eur. Phoen. 533 f. (on
Ambition) ººf K YŒı ŒÆd ºØ PÆÆ = KBºŁ ŒIBºŁ
K OºŁæfiø H
æøø, Rhes. 811 f., Xen. Hell. 5. 2. 39 (cited among
other passages by E. Kemmer, Die polare Ausdrucksweise in der grie-
chischen Literatur, 1903: pp. 225 ff.). So also Psalm 121: 8 ‘The Lord shall
preserve thy going out and thy coming in’; comprehensiveness is a
natural feature of religious language.
Some scholars have supposed that Faunus is being urged to leave
before he does any harm (cf. Pasquali 563, A. W. J. Holleman, AC 41,
1972: 563 ff.). It is true that in some contexts he had a dangerous aspect
(Porph. on v. 1 ‘pestilentem deum’, RE 6. 2060 ff.), just like Pan (Gow
on Theoc. 1. 16); but here he is a protector of flocks. The reading adeas
appears in the ninth-century codex Bernensis, and that verb is occasion-
ally used in similar contexts (Cic. har. resp. 62, Virg. Aen. 8. 302 ‘et nos et
tua dexter adi pede sacra secundo’). But abeas (supported by Servius on
Aen. 7. 91) gives a good contrast to incedas and leads more smoothly to
the next stanza, which implies that Faunus’ visit will last till his De-
cember festival; so also 1. 17. 2 ff. ‘igneam / defendit aestatem capellis /
usque meis pluviosque ventos’.
The alumni are lambs and kids (not slave-boys as ps.-Acro oddly
supposes); parvis gains emphasis from the hyperbaton and stresses their
need for protection. Most editors understand meis from meos above
(cf. 1. 17. 3 f. ‘capellis . . . meis’) suggesting H’s pride of ownership;
Heinze understood tuis, but that takes the god’s concern too much for
granted. If meis is regarded as too sentimental, one might think rather of
the unweaned nurslings (ŁæÆÆ) of the flock itself; the same alterna-
tive is available at 3. 23. 7. aequus means ‘favourable’ or ‘kindly’ (cf. serm.
2. 3. 164 f. ‘immolet aequis / hic porcum Laribus’, carm. saec. 65); in the
ancient world Faunus was already connected by some with favere (see
the introduction above).
5. si tener pleno cadit haedus anno: the conditional clause gives a

justification for the god’s help (N–H on 1. 32. 1, Pulleyn 16 ff.). Normally
the tense is either past or future (Hom. Il. 1. 40 f. N Ø ŒÆa
Æ æ ŒÆ = Ææø M ÆNªH, Ø Œæ Kºøæ, Virg.
Aen. 1. 334 ‘multa tibi ante aras nostra cadet hostia dextra’ with Austin);
but here the present tense underlines that the sacrifice recurs every year.
The euphemistic cadere of a sacrificial victim (TLL 3. 25. 9 ff.) is a
vox propria rather than a poeticism. As Faunus was assimilated to the
goat-like Pan, he is offered a male kid (less appropriately a ewe at
Ov. fast. 4. 653); at 1. 4. 12 he is given the choice. pleno . . . anno means
‘when the calendar year is complete’ and refers to the Faunalia on 5 Dec.
(an idea developed in the next two stanzas); distinguish 3. 22. 6 ‘per
exactos . . . annos’, which refers to anniversaries.
6–7. larga nec desunt Veneris sodali / vina creterrae: the MSS vary
between creterrae and craterae; the former was naturalized in Latin at an
early stage, but being less familiar in late antiquity was often corrupted
(W. Clausen, CQ 13, 1963: 85 ff.). The mixing-bowl was needed in the
first instance for a libation at the sacrifice (see RE 15. 2039, Hilgers
157 ff.); cf. especially Leonidas, anth. Pal. 9. 99. 6, where the goat is
sprinkled with wine from the vine that he has nibbled. But H envisages
an uninhibited rustic festival that goes beyond the thank-offering to
Faunus; hence the emphatic larga (which balances pleno in a purely
formal way) and the litotes nec desunt (cf. 1. 36. 15 ‘nec desint epulis
rosae’). Hence, too, the mention of Venus, which would not belong to
the authentic cult; for the association of wine and sex cf. Ar. PCG 3. 2 fr.
613 ¼ Athen. 10. 444d Ø r #æ ªºÆ, Eur. Bacch.
773 Yı b Œ Z PŒ Ø ˚æØ , Ter. eun. 732 ‘sine Cerere
et Libero friget Venus’ (with Pease on Cic. nat. deor. 2. 60), Ov. ars 1. 244
‘Venus in vinis, ignis in igne fuit’, Otto 366. For the mixing-bowl as
‘boon-companion of Venus’ cf. 1. 25. 19 f. ‘hiemis sodali / . . . Euro’ with
N–H, Hom. Od. 17. 271 (where the lyre is the companion of the feast), h.
Hom. 4. 31 with Allen and Halliday, anon. anth. Pal. 5. 135. 3 ´Œ
ı ŒÆd
1ıø ƒºÆæc ºæØ ŒÆd ˚ıŁæ (to a wine-jar).
7–8. vetus ara multo / fumat odore: Peerlkamp inserted et before vetus,
unnecessarily; for the asyndetic combination of two parallel subordinate
clauses cf. Lucr. 3. 957 ‘sed quia semper aves quod abest, praesentia
temnis . . . ’, Leo (1912), 272 n. 4. For the hallowed associations of the
altar of Faunus cf. Men. dysc. 1 ff. (set at Pan’s famous altar at Phyle),
E. W. Leach, Vergil’s Eclogues: Landscapes of Experience, 1974: ch. 3
(illustrating ‘sacral-idyllic’ scenes from poetry and wall-painting). vetus
evokes the age-old rhythms of the agricultural year, and reinforces the
impression that cadit (5) refers to an annual event. multo . . . odore refers
to incense, as Porph. says (cf. 1. 30. 3. to Venus ‘ture te multo’, 3. 23. 3 to
the Lares); this was used even in simple sacrifices, perhaps to exclude
less agreeable smells (see S. Lilja, The Treatment of Odours in the Poetry
of Antiquity, 1972: 31 ff.). For smoking incense cf. Eur. Andr. 1026, Bacch.
144 ff. &ıæÆ ‰ ºØı ŒÆ- = e › ´ÆŒ
f I
ø = ıæ º ªÆ
ŒÆ , Ov. met. 10. 273 ‘turaque fumabant’, Sil. 7. 457 ‘Paphos centum
mihi fumet in aris’; fumare is more often applied to the burnt offerings
themselves (Ov. fast. 2. 193 ‘Idibus agrestis fumant altaria Fauni’), but
after the sacrifice (5) H turns to its concomitants, wine and incense.
9. ludit herboso pecus omne campo: this is no longer part of the

conditional clause, but a new sentence describing the winter Faunalia
(as Bentley saw). For ludere of animals cf. 3. 11. 10 (a frisking horse),
Lucr. 1. 261, 2. 320 ‘et satiati agni ludunt blandeque coruscant’, Ov. fast.
1. 156, anth. Lat. 238a. 1 ‘adludunt pavidi tremulis conatibus agni’, TLL 7.
2. 1771, 16 ff.; the prominent opening verb sets the tone for the human
merrymaking that follows. herboso suggests that there was still grass on
the lower ground when the flocks left the hills for the winter; for
‘transhumance’ cf. Varro, rust. 2. 2. 9, K. D. White (1970), 507 n. 130,
Mynors on Virg. georg. 3. 146, Horden–Purcell (2000), 549 ff.
10. cum tibi Nonae redeunt Decembres: for the festival on 5 Dec. see
the introduction above. tibi (repeated in 14) is the emphatic pronoun
often found in hymns and prayers (Norden, 1913: 149 ff., N–H on 1. 10. 9).
For redire of an anniversary cf. 3. 8. 9.
11–12. festus in pratis vacat otioso / cum bove pagus: vacat and otioso
emphasize that the festival is a day of rest for the farm-workers and the
ploughing-oxen; cf. 3. 17. 16 ‘cum famulis operum solutis’, Tib. 2. 1. 5 ff.
‘luce sacra requiescat humus, requiescat arator, / et grave suspenso
vomere cesset opus . . . ’ with K. F. Smith. Sacral law forbade work
on such a day, but there were various reasonable exceptions (Virg.
georg. 1. 268 f. with Mynors, Nock, 1972: 2. 738). A pagus was a scattered
rural community that had some administrative and religious responsi-
bilities (RE 18. 2. 2318 ff., Bömer on Ov. fast. 1. 669, OCD 1092). H’s
pagus included Mandela and apparently Varia, now Vicovaro; cf. 2. 13. 4,
epist. 1. 18. 104 f. ‘gelidus Digentia rivus, / quem Mandela bibit, rugosus
frigore pagus’, epist. 1. 14. 3 (for the topography of the area see S. Q.
Gigli in Encicl. oraz. 1. 253 ff.).
pardus is read in a number of significant MSS (sometimes as a
variant), though pagus is supported by ps.-Acro’s comment and the
preponderance of the tradition (cf. also Ov. fast. 1. 669 ‘pagus agat
festum, pagum lustrate coloni.’). As Bentley pointed out with amuse-
ment, the former reading is due to a reminiscence of Isaiah 11: 6
‘habitabit lupus cum agno et pardus cum haedo accubabit’; for monastic
corruptions cf. Petr. 43. 1 ‘abbas secrevit’ (for ‘ab asse crevit’), R. M.
Ogilvie, G & R 18, 1971: 32 ff., J. Willis, Latin Textual Criticism, 1972:
100 ff.
13. inter audacis lupus errat agnos: the wolf is proverbially the enemy
of the flock (cf. epod. 4. 1, 12. 25 f., 15. 7, TLL 7. 2. 1855. 31 ff.), but here the
lambs are unnaturally bold (audacis is emphasized by the hyperbaton). It
was a proverbial impossibility (‘adynaton’) for the laws of animal nature
to be reversed; cf. 1. 33. 8 with N–H, Ar. pax 1076 æ Œ ºŒ r
ÆØE, Virg. ecl. 8. 52 ‘nunc et ovis ultro fugiat lupus’, Otto 198, TLL 7.
2. 1853. 8 ff. In descriptions of an ideal world this impossibility is
portrayed as an actuality; here a major influence is Virg. ecl. 5. 60 ‘nec
lupus insidias pecori . . . ’, 4. 22 ‘nec magnos metuent armenta leones’
(probably drawing on the paraphrase of Isaiah 11: 6 at orac. Sib. 3. 791 f.).
In our passage H’s idyllic fantasy goes beyond the parallel at 1. 17. 8 f.
‘nec viridis metuunt colubras / nec Martiales haediliae lupos’; for a
further elaboration cf. Prud. cath. 3. 153 ff., who in addition to fearless
flocks speaks of an ineffectual serpent (153), obedient lions (162), and an
aggressive dove (164 f.).
14. spargit agrestis tibi silva frondes: the wood scatters leaves in
honour of Faunus; the repeated tibi suits the hymnal style (10 n.), and
the position in the line is emphatic (Nisbet, ap. Adams–Mayer 144 f.).
For the scattering of flowers or leaves as a mark of honour (IŁºE or
ıºººE) cf. Pind. P. 9. 123 f. ººa b ŒEØ Œ = ºº Ø ŒÆd
ı , Virg. ecl. 5. 40 ‘spargite humum foliis’, Ov. trist. 4. 2. 50,
Matthew 21: 8 ¼ººØ b Œ Œºı Ie H æø ŒÆd
Kæı K fi Ð ›fiH (cf. Mark 11: 8), Apul. met. 11. 9. 2 with Griffiths,
RE 20. 1025, Browning, The Patriot 1 f. ‘It was roses, roses, all the way /
With myrtle mixed in my path like mad’. For such tributes by inanimate
Nature cf. Lucr. 1. 7 f. (the hymn to Venus) ‘tibi suavis daedala tellus /
summittit flores’, Virg. ecl. 4. 18 ff. agrestis (accusative) shows that the
foliage belongs to the wild woods rather than being made up by some
coronarius; cf. Virg. ecl. 10. 24 ‘agresti capitis Silvanus honore’, Tib. 2. 5.
117 ‘lauro devinctus agresti’, TLL 1. 1418. 11 ff. Faunus himself is called
agrestis (Ov. fast. 2. 193, cf. 3. 315), just as Pan is Iªæ Æ (Leonidas, anth.
Pal. 6. 13. 2) or IªæÆºfiø (Archias, ibid. 6. 179. 1).
Cornelissen proposed arentes (Mnem. 16, 1888: 310), but withered
leaves would be a poor compliment; he himself cites 1. 25. 19 f. ‘aridas
frondes . . . / dedicet Euro’ (a pejorative passage that does not help
his case). To avoid this difficulty Lucian Müller combined arentes
with ubi (for tibi); and this has been accepted by Shackleton Bailey.
But though leaves in Italy might fall as late as December (epod.
11. 5 f. ‘hic tertius December . . . / silvis honorem decutit’), H could not
say ‘when the leaves fall’ as a way of pointing to the Faunalia on
5 December.
15–16. gaudet invisam pepulisse fossor / ter pede terram: fossor is

sometimes used contemptuously for a clodhopper (Catull. 22. 10 ‘capri-
mulgus aut fossor’, Pers. 5. 122), but here it has the more genial tone of
Virg. georg. 2. 264 ‘labefacta movens robustus iugera fossor’. ‘To beat the
ground with one’s feet’ describes the vigorous dancing of antiquity (cf. 1.
4. 7 ‘alterno terram quatiunt pede’ with N–H, 1. 37. 2. ‘pulsanda tellus’);
for rustic dances cf. epist. 1. 14. 24 f. (to the vilicus) ‘meretrix tibicina
cuius / ad strepitum salias terrae gravis’, Lucr. 5. 1402 ‘duriter et duro
terram pede pellere matrem’, Nonn. 47. 37. But stamping with the feet
could also be a sign of anger; cf. Cic. de orat. 1. 230 with Wilkins, Sen.
de ira 1. 1. 4 ‘complosae saepius manus et pulsata humus pedibus’. Here
invisam suggests that the fossor gets his own back by kicking the earth
that has cost him so much trouble; cf. Porph. ‘naturale est enim omnibus
odisse laboris sui materiam’. A dance with triple beat (tripudium) was
associated with the Salii (1. 36. 12 with N–H, 4. 1. 28) and the Arval
Brethren (ILLRP 4, ROL 4, p. 250, E. Norden, Aus altrömischen Priest-
erbüchern, 1939: 238 f.). The movements described by Plut. Num. 13. 4 f.
must have been much more sophisticated than the rustic dances referred
to here and at Ov. fast. 6. 330 ‘et viridem celeri ter pede pulsat humum’,
Calp. Sic. 4. 128 f. Although ps.-Acro’s comment is not clearly expressed
(‘ter vero ad rhythmum rettulit sonos’) it is plain that H’s ter pede terram
has an onomatopoeic effect.
pepulisse here means ‘to strike’, not ‘to have struck’; the poets find the
perfect infinitive metrically convenient, and often there is no clear
distinction between it and the present (cf. K. F. Smith on Tib. 1. 1.
29–32, Bo 270). But in our passage as sometimes elsewhere H seems to
emphasize an instantaneous action as opposed to a continuing state; cf.
1. 34. 16 ‘posuisse gaudet’, 3. 4. 51 f. ‘fratresque tendentes opaco / Pelion
imposuisse Olympo’, serm. 1. 2. 28, 2. 3. 187. We may compare the ‘aspect’
of the Greek aorist infinitive, where OæªØŁBÆØ means ‘to get angry’ as
opposed to the present OæªŁÆØ ‘to be angry’ (cf. Arist. eth. Nic. 1173b
1 f., W. W. Goodwin, Syntax of Moods and Tenses of the Greek Verb, edn.
2, 1889: 28 ff.). So in Latin we find S. C. de Bacchanalibus 2 (ROL 4,
p. 256) ‘neiquis eorum Bacanal habuise velet’, Cato, agr. 5. 4, tabulae
1 9 . QVA N T V M D I S T E T A B I NAC H O 227
Vindolandenses 2. 505 (A. K. Bowman and J. D. Thomas, Britannia 27,
1996: 324) ‘cras quid velis nos fecisse rogo, domine, praecipias’; these are
clearly not literary affectations. See further K–S 1. 133 ff., H–Sz 351 f.,
A. Ernout and F. Thomas, Syntaxe latine, 1953: 259 f., Adams–Mayer 8
(with R. G. G. Coleman, ibid. 83).
1 9 . QVA N T V M D I S T E T A B I NAC H O
[E. Bignone, RFIC n s . 7, 1929: 457 ff.; J. F. G. Gornall, G&R 18, 1971: 188 ff.; H. U.
Instinsky, Hermes 82, 1954: 124 ff.; J. Rüpke, MH 53, 1996: 217 ff.; H. Tränkle, MH 35, 1978:
48 ff.; L. Wickert, RhM 97, 1954: 376 f.; Williams 115 ff.]
1–8. You keep talking of remote chronologies, but say nothing about a
symposium this cold night. 9–17. Pour a toast to the new month, the new day,
and the augurate of Murena: the bard shall have nine measures and the
decorous only three. 18–28. Let us have wild music and a profusion of roses, so
that old Lycus next door may be envious. Rhode makes advances to handsome
young Telephus; I am smouldering with love for my Glycera.
This ode celebrates the installation as augur of a certain Murena,

though the situation does not emerge until vv. 10 ff. He is identified by
some with the Licinius (Murena) of 2. 10, (the brother of Maecenas’
wife Terentia), who was accused of conspiracy probably in 22 bc (Dio
54. 3. 4 f.) and killed ‘while intending to escape’; that ode’s commen-
dation of the Golden Mean suits the Peripatetic interests of ‘the con-
spirator’, as shown by N–H vol. 2, pp. 152 f.. However, the Murena of
our poem was probably not the same person but perhaps his brother,
perhaps the A. Terentius Varro Murena who was elected consul for 23
bc but was for some reason replaced; for the difficult prosopography
of this family see especially G. V. Sumner, HSCP 82, 1978: 187 ff.,
Syme (1986), 387 ff., J. S. Arkenberg, Historia 42, 1993: 326 ff., 471 ff.
A relationship to Maecenas would give a point of contact with Horace;
at the same time the ode’s indirect approach suits a nobilis who was
not a close intimate, just as 3. 21 is addressed not to Messalla but to
a wine-jar.
The poem begins abruptly with a protest: somebody is going on
about chronological questions that all seem very remote (1–4). The
natural scene for such a discourse is not a casual encounter but a dinner:
one recalls Callimachus’ aetiological questions to the Ician (aet. fr. 178),
as well as such treatises as the Quaestiones Conviviales of Plutarch, the
Deipnosophistae of Athenaeus, the Noctes Atticae of Aulus Gellius,
the Saturnalia of Macrobius ( J. Martin, Symposion, 1931). A suitable
occasion would be the cena aditialis at Murena’s installation as augur,
where boring antiquarians were likely to be found. For such banquets cf.
Varr. rust. 3. 6. 6 ‘primus (pavones) Q. Hortensius augurali cena posuisse
dicitur’, Sen. epist. 95. 41 ‘et deciens tamen sestertio aditiales cenae
frugalissimis viris constituerunt’, Wissowa 491, RE 2. 2319; for pontifical
dinners in general cf. also 2. 14. 28, Macrob. Sat. 3. 13. 10–12 (with a long
menu). If Horace’s comments sound too discourteous for a formal
dinner of this kind, we can always understand them as an interior
monologue.
In the second stanza it becomes clear that Horace would sooner hear
about arrangements for the coming symposium (5–8); when in v. 8 he
speaks of escaping from the cold, we must assume that it is a chilly night
and that he has in mind the walk to the venue of the drinking-party. It
was a poetic convention to say that guests should concentrate on
enjoyment rather than serious preoccupations. Sometimes these were
questions of war and politics; cf. 1. 26. 3 ff., 2. 11. 1 ff., 3. 8. 17 ff., Theog.
763 f., Anacr. eleg. 2. 1 ff., Xenophanes, eleg. 1. 21 ff. For other rejected
topics cf. 1. 11. 1 ff. (astrology), Anacreontea 50. 1 ff. Bergk ¼ 52 (a)
Campbell (schemes), Alexis (?) PCG 2 fr. 25. 1 ff. ¼ pp. 822 ff. in W. G.
Arnott’s commentary (philosophy): ÆFÆ ºæE , ºÆH ¼ø
Œø = ¸ŒØ, #ŒÆØÆ, 4Øı ºÆ , = ºæı ØH; Pb
£ ø ŒÆº = ø, Kø . . . Particularly relevant are some
myuric or mouse-tailed hexameters of the Roman period ( J. U. Powell,
Coll. Alex. p. 199, no. 37. 7 ff. ¼ Page, GLP no. 125. 8 ff. or N. Hopkinson,
A Hellenistic Anthology, 1988: pp. 80 f.): c ŒÆ E Ł lºØ j
Ł oøæ, = Iººa Ł e æ ŒÆd f ı Iªæfi . Here,
as in our ode, the rejection of serious themes is followed by questions
about purchases for the symposium (6 n.); the Greek poem next men-
tions measures of various liquids (3 of honey, 5 of milk, 10 of wine, 12 of
myrrh, and 2 of spring water), and goes on to speak of a girl, a lyre, and a
Phrygian pipe. The resemblances suggest shared antecedents in the
drinking-songs of the Greek symposium if not a direct borrowing by
Horace (thus Bignone, op. cit.).
The last five stanzas describe the ensuing symposium; it was a
common practice for diners to go elsewhere for their drinking (Plaut.
most. 315 ff. ‘nam illi ubi fui inde ecfugi foras, / ita me ibi male convivi
sermonisque taesumst. / nunc comissatum ibo ad Philolachetem’, Liv.
40. 7. 5 ‘quin comissatum ad fratrem imus?’, Blümner, 1911: 400); for a
similar progression cf. 3. 14. 17 ff. (again moving from an official occasion
to an informal party). In our poem, however, the change of scene is
disconcertingly abrupt. Some think that Horace interrupts a boring
dinner with a call for wine, but that is too rude; it is also too late to
discuss the price (5 n.), and nonsensical to ask about the place and time
of the festivities (see 5–6 n. and 7 n.). Some commentators have regarded
the last five stanzas as a scene imagined by the diner, in anticipation of
the coming symposium; but that loses vividness and raises questions at
v. 23 Lycus and perhaps v. 27 petit (see the notes below). In view of
the very real difficulty it might be worth considering that after v. 8
a transitional stanza has fallen out (as once suggested to RN by
W. A. Camps).
Preparations for the drinking-party were a standard poetic theme
(5–6 n.); here Horace envisages an ‘eranos’ on the Greek model, in
which one man provides the accommodation (‘praebente domum’ in
v. 7) and another prepares the hot drinks (6 n.). Not only does he
propose toasts (9 n.), which any guest could do, but as self-appointed
magister bibendi (N–H on 1. 4. 18) he prescribes the quantities to be
drunk (11 ff.). Then, as commentator on a developing scene, he keeps up
a flow of instructions and descriptions till the end of the poem; for this
mime-like technique cf. 1. 27 with N–H, E. Stemplinger, Philologus 75,
1919: 466 ff. The typical concomitants of a drinking party are duly
recorded (N–H vol. 1, p. 402, vol. 2, p. 168, RE 4. 618, Blümner, 1911:
410 ff.): the Chian wine with its possibly significant origin (5 n.), the
wild music that annoys the neighbours (18 ff., 22 ff.), the roses that are
scattered with hyperbolical abandon (22 n.), the on-coming girl (27 n.).
The wintry evening is also traditional (8 n.) and is set against a very
Roman use of hot wine (6 n.).
The wildness of the party is reflected in the style (Syndikus 2. 171 f.)
with its rapid choriambic metre (cf. 1. 11, 1. 36, and for different reasons
3. 25), its increasingly short sentences (contrast the opening period with
its leisurely A et B et C), its mixture of crisp imperatives and impatient
questions, its sudden breaks in unexpected places (11, 15, 17, 22). In all
this there is none of the mellow wisdom that usually characterizes
Horace’s sympotic odes; cf. A. La Penna in O. Murray and M. Tecus˛an
(1995), 275. The poet mentions the restraining influence of the Graces
(16 n.), but he himself chooses the bigger drinks (14 f. with 11 n.), and the
less sober role (18 insanire). But though the fleeting pleasures of youth
make an arresting contrast with the antiquated concerns of pedants,
Horace seems no longer quite at home even with the former: though he
is not jealous like the misanthrope next door (22–3), he cannot match
young Telephus’ lustrous hair (25 n.) and sexual magnetism (28 n.). Here
he not only paints a vivid and economical picture, but communicates an
emotion. Once again he has shown how to celebrate an official occasion
without abandoning his lyric persona (cf. 3. 14. 17 ff. and in its different
way 1. 31. 17 ff.); contrast the formality of Tibullus 2. 5. 1 (on the elevation
of Messalinus to a priesthood) ‘Phoebe fave; novus ingreditur tua
templa sacerdos’.

1. Quantum distet ab Inacho: Inachus was the legendary first king of
Argos; for his proverbial antiquity cf. N–H on 2. 3. 21. He became a
point of reference for ancient chronographers (E. Schwartz, AGG 40. 2,
1895: 11 ff.); cf. Ocellus Lucanus 42, pp. 218 ff. Harder E ºªıØ c
B , ¯ ººØŒB ƒæÆ Iæ
c Ie +
ı r ÆØ F #æªı, Castor of
Rhodes (1st cent. bc) FGrH 250 F3, cf. RE 9. 2. 1218 f., 10. 2352 (Inachus
begins the list of Argive kings), Cens. 21. 2 ‘a priore scilicet cataclysmo
ad Inachi regnum anni sunt circiter quadringenti’, Apollodorus, 2. 1. 1
(Inachus begins the book), Clem. strom. 1. 102. 1, Aug. civ. Dei 18. 3
(from Varro). But though the king was so remote, the pedant in the
ode must fix him on a time-scale. For the place of Inachus in early
Greek genealogies add M. L. West, The Hesiodic Catalogue of Women,
1985: 177, R. L. Fowler, PCPS 44, 1998: 7 ff.; for later calculations see
A. A. Mosshammer, The Chronicle of Eusebius and Greek Chronographic
Tradition, 1979.
One is tempted to try a further speculation about the function of the
Argive king in the ode. The Licinii Murenae came from Lanuvium
(Cic. Mur. 86), and must have been its most distinguished family; the
foundation of the place was attributed to the Argive Diomedes (App.
civ. 2. 20), and its famous cult of Juno Sospita (Cic. Mur. 90, Wissowa
188) was associated with Argive Hera (Ael. nat. an. 11. 16); a sculpture
of Alexander in the precinct and fragmentary inscriptions have been
connected with the victories of L. Murena (cos. 62) and Lucullus (also a
Licinius) over Mithridates (F. Coarelli, Coll. de l’école fr. de Rome 55,
1981: 251 ff.). The name Licin(n)ius could have been associated with
Licymnius (RE Suppl. 8. 259 ff.), the eponymous hero of Licymna
(the acropolis at Tiryns), whose grave was still shown at Argos (Paus.
2. 22. 8, Plut. Pyrrh. 34. 2), and who in Apollodorus comes twelfth in
line from Inachus (2. 1. 1–2. 4. 5); at 2. 12. 13 ps.-Acro identifies Licymnia
with Terentia, the sister of Licinius the alleged conspirator (see N–H ad
loc. and vol. 2, pp. 180 ff.). There was a vogue at the time for imaginative
family history (cf. the introduction to 3. 17); so Murena may have traced
his roots to the kings of Argos (cf. 2. 3. 21 to Dellius who might have
claimed to be ‘prisco natus ab Inacho’). The pedant could then have
been shown as flattering Murena by referring to his ancestors.
2. Codrus pro patria non timidus mori: the legendary last king of
Attica is significantly combined with the first king of Argos; his death
marked a stage in chronological systems (cf. Vell. 1. 2. 1, RE 11. 986 ff.,
E. Schwartz, op. cit (1 n.), 16). The name was proverbial for old-
fashioned ways; cf. paroem. Gr. 2. 148 etc. Iæ
ÆØ æ ˚ æı, RE
11. 993 f. When the Spartans invaded Attica, and the Delphic oracle
promised victory to the side that lost its general, Codrus disguised
himself, picked a quarrel with some Spartan soldiers, and was duly
killed. The story flowed from Greek patriotic discourse to Roman
exempla (Hellanicus, FGrH 4 F125, Lycurgus, Leocr. 84–6, Cic. nat.
deor. 3. 49 with Pease). With non timidus mori cf. 4. 9. 51 f. ‘non ille
pro caris amicis / aut patria timidus perire’; for H’s use of the infinitive
see N–H on 1. 1. 8.
3. narras: emphatically placed, to balance taces (8). narrare has various

idiomatic nuances (Lejay on serm. 1. 9. 52) and here suggests ‘a long
story’; cf. Petr. 44. 1 ‘narratis quod nec ad caelum nec ad terram pertinet,
cum interim nemo curat quid annona mordet’ (the same sort of contrast
as here), Mart. 4. 61. 15 f. For tedious talk at a convivial occasion cf. Plut.
quaest. conv. 614e, Gell. 13. 11. 4.
et genus Aeaci: the line consisted of Aeacus, Peleus, Achilles,
Neoptolemus; the last was pre-eminent in the capture of Troy, where
Eratosthenes began his chronography (FGrH 241 F1, RE 6. 382).
4. et pugnata sacro bella sub Ilio: the adjective is Homeric (Il. 4. 46

! +ºØ ƒæ) and the preposition is also traditional (Il. 2. 216 e ! +ºØ
qºŁ Virg. Aen. 5. 261 ‘sub Ilio alto’). The plural bella seems to include a
reference to Troy’s earlier capture by Telamon, who was himself a son
of Aeacus. For the irrelevance of Homeric scholarship Bignone op. cit.
cites Diog. Laert. 6. 27 (on Diogenes) ªæÆÆØŒf KŁÆÆ
a b F ˇıø ŒÆŒa IÆFÆ , a YØÆ IªFÆ , Sen.
epist. 88. 6–7 ‘quid, inquam, annos Patrocli et Achillis inquirere ad rem
existimas pertinere?’
RN speculates that there may also be an indirect allusion to the augur
Murena’s great-grandfather (by adoption), L. Licinius Murena. When
Sulla rescued Ilium in 85 after its sack by Fimbria, Murena as one of his
principal subordinates may have played a significant part; these sensa-
tional events were naturally compared with the Trojan War (Strab.
13. 1. 27, App. Mith. 53, Aug. civ. Dei 3. 7), and were commemorated
in the Augustan sanctuary (A. Barchiesi in Style and Tradition, ed.
P. Knox and C. Foss, 1998: 139). So they would have made an appropri-
ate topic of conversation at the installation of the augur (see also the
following note).
5–6. quo Chium pretio cadum / mercemur: for Chian wine cf.
Hermippus ap. Athen. 1. 29e (Dionysus calls it the best of all), Plaut.
Poen. 699, Lucil. 1131M, Varro, rust. 2, praef. 3, Tib. 2. 1. 28, Plin. nat.
hist. 14. 73, 96, RE 3. 2291; H refers to it at serm. 1. 10. 24, 2. 3. 115 f., 2. 8.
15, epod. 9. 34 (in connection with Actium), but only here in the Odes,
where it suits the Greek type of symposium. His wines sometimes have
a particular significance, so there may again be a reference to the
achievements of Murena’s great-grandfather. In 86 bc Mithradates
deported the Chians (App. Mith. 47), but the island was recaptured by
Lucullus in 85 (Plut. Luc. 3. 3), and in the peace treaty of Dardanus in
that year Sulla insisted on the restoration of the inhabitants (App. Mith.
55, D. Magie, Roman Rule in Asia Minor 1, 1950: 224 ff.). On Sulla’s
return to Rome, Murena was left as governor of Asia for 84; in view of
the other honours which he received from Greek cities (SIG edn. 3, 745,
RE 13. 445), he may have won some of the credit for restoring
the Chians, and perhaps even established a hereditary connection.
When Lucullus returned to Rome from the East in 80 he was said
to have distributed more than 100,000 jars of Greek wine (Plin. nat.
hist. 14. 96).
mercemur is an indirect deliberative subjunctive; for similar prepar-
ations cf. 1. 38 with N–H vol. 1, pp. 421 ff., 3. 14. 17 ff., and especially
Page, GLP 125. 8 ff. (cited in the introduction above). The plural verb
implies that the cost is being shared (cf. the Greek æÆ ). It has been
argued that the price is not money, but a song that begins at v. 9 (Gow),
or clever conversation (Wickert, op. cit.), or toasts to Murena (Tränkle
op. cit.); for such a quid pro quo at a symposium cf. 4. 12. 16 ‘nardo vina
merebere’, Catull. 13, Gell. 7. 13. 2. But as literal purchases are natural in
a poem of this kind, and as the following clause ‘quis aquam temperet
ignibus?’ admits no metaphorical explanation, the obvious interpret-
ation should be accepted.
6. quis aquam temperet ignibus: temperare is used of modifying an

extreme (1. 20. 11 with N–H), here of warming cold water. Horace is
not giving orders to slaves (as at 2. 11. 18 ff.) but allocating functions
among the contributors to the symposium (as at 2. 7. 23 ff.); this is
shown by the parallel quo praebente domum. Ps.-Acro says that the hot
water is for mixing with the wine; for this Roman custom note the
thermipolium already in Plautus (cf. Fraenkel, 1960: 149 n. 2, 415 f.), and
see further Marquardt–Mau 332 f., Blümner (1911), 401, Mayor on Juv.
5. 63, T. Kleberg, Hôtels, restaurants et cabarets dans l’antiquité romaine,
1957: 24 f. On this interpretation hot water follows easily after the
mention of wine, and explains the escape from the cold in v. 8 (frigoribus
balances ignibus). It is objected that the owner of the house could easily
have supplied the amount of hot water that was needed, but the Greek
type of symposium may have had a tradition of self-service, with the
help of an authepsa or portable stove (Blümner, 1911: 402, fig. 59) and
personal slaves; anything to do with the mixing of the wine played a
central part in the rituals. Lambinus and others have seen a reference
to water for baths; normally these were taken before dinner (cf.
R. Ginouvès, Balaneutikè, BEFAR 200, 1962: 157), and though they
were sometimes available afterwards (Petr. 72. 3, Mayor on Juv. 1. 143),
they were not central to the symposium.
7. quo praebente domum: the phrase is co-ordinate with quota
(see next note); that is to say it goes with caream (8), not with temperet
(6) as Orelli supposes. praebere domum applies to one who simply
makes his house available (here as his contribution to the symposium);
cf. serm. 1. 5. 37 f. ‘in Mamurrarum lassi deinde urbe manemus / Murena
praebente domum, Capitone culinam’ (though on that occasion
Murena apparently provided accommodation too), Plin. epist. 8. 12. 2
(for a poetry recital), Tac. ann. 11. 4. 1 (for a sexual assignation; cf. Ca-
tull. 68. 68). Some think that our Murena, who may be the man
mentioned in the first parallel above, is actually the host (Instinsky,
op. cit. 125 f.); but a great man would hardly lend his house for a
celebration of his augurate without offering further hospitality; and
if he were present, our attention would surely be focused on him,
Horace would not be acting as the magister bibendi, and we would not
expect to hear about the poet’s sexual aspirations, much less those of
a meretrix.
et quota: ‘and at what (hour)?’; quota corresponds to an ordinal
numeral. The time is naturally mentioned in invitations (epist. 1. 7. 71
‘post nonam venies’, 1. 5. 3 ‘supremo te sole domi, Torquate, manebo’,
Ter. eun. 541 ‘locus, tempus constitutum est’, Varro, Men. 335 ¼ Gell.
13. 11. 3). Heinze (followed by Gornall, loc. cit.) thought that domum
referred to the guests’ lodging after the symposium and quota to the
time when the party was scheduled to break up. But as the scene is
presumably Rome (8 n.), the guests would not be looking for a bed; and
an interest in when the party is going to finish accords badly with the
wild mood of the latter part of the poem. It is also wrong to suppose
that the ‘house’ is the scene of the function already taking place (the
information would then be superfluous), and that quota refers to the
time of the toasts (cf. 10 noctis mediae): the guests cannot be expected to
wait around till midnight before they get a hot drink.
8. Paelignis caream frigoribus taces: the district of the Paeligni, i.e. the
valley of Corfinium and Sulmo in the central Apennines, is almost
surrounded by mountains (the Maiella is over 9,000 ft. high), and the
cold was commemorated by its greatest son: see Ov. am. 2. 16. 36, fast.
4. 81, trist. 4. 10. 3, also Sil. 8. 510, Nissen, 2. 1 (1902), 445 ff., RE 18. 2.
2230. The sophisticated symposium with its winter roses and familiar
hetaerae is surely not set in this remote area (thus Shackleton Bailey,
1982: 96 f.); rather the proper name suggests chill East winds and
corresponds to literary epithets like ‘Sithonia nive’ (3. 26. 10) or
Ø
˚ºØ (Philodemus, anth. Pal. 10. 21. 4). Wintry weather often sup-
plies a background to symposia, making a contrast with the snugness
within (cf. 1. 9 with N–H, vol. 1, p. 117). caream is deliberative (like
mercemur and temperet) and means ‘escape from’ (cf. N–H on 2. 14. 13),
particularly by drinking hot wine; cf. Petr. 42. 2 ‘sed cum mulsi pultar-
ium obduxi, frigori laecasin dico’.
9–10. da lunae propere novae, / da noctis mediae: for the difficult

change of place see the introduction above. On our interpretation all the
questions asked in the second stanza have now been answered. As the
wine was ladled in the cyathus from mixing-bowl to cup, a toast could
be named with each ladleful; cf. Porph. ‘veteres singulos vini cyathos
sub singulorum deorum aut carissimorum sibi nominibus solebant in
poculum defundere’, Alexis, fr. 116. 1 ff. (below, 11 n.), Call. ep. 29. 1,
Meleager, anth. Pal. 5. 136. 1 f., Argentarius, ibid. 5. 110. 1 f., RE 4. 614 f.
The Greek character of the party is underlined by the use of the genitive
for the toast; see 3. 8. 13 n., W. G. Arnott on Alexis, PCG 2, frr. 59 and
116 (composite toasts to a deity and a person).
lunae novae refers not to the new moon (which would be unlikely to
coincide with the right date) but the new month when Murena assumed
his augurate; cf. the Greek ıÆ for the beginning of the month. As
the time is winter (8) one is tempted to think of the first of January, but
then one would expect the new year to be mentioned; perhaps the first
of February is meant, which might have been agreeable to Murena as
the day of Juno Sospita in Rome (Ov. fast. 2. 55 f. with Bömer) and
perhaps also Lanuvium (cf. 1 n.). Midnight marks the beginning of
Murena’s augurate.
10–11. da, puer, auguris / Murenae: puer, like ÆE, is a normal address to
a slave (cf. 1. 38. 1), notably in contexts that refer to pouring wine (epod.
9. 33, Alexis fr. 116 cited 11 n., Stat. silv. 1. 5. 10); so it is unlikely to refer
to a symposiast (like Thaliarchus in 1. 9), though in the Greek sympo-
sium free-born young men could act as wine-pourers ( J. Bremmer in
Sympotica, ed. O. Murray, 1990: 139 ff.). In the phrase auguris Murenae
the emphasis is on auguris; cf 3. 8. 13 f. ‘sume, Maecenas, cyathos amici /
sospitis centum’ with note. The repeated rhymes lunae, novae, mediae,
Murenae enforce the impression of a ritual (cf. Virg. ecl. 4. 4–10, 8. 80),
just like the magic numbers that follow.
11. tribus aut novem: the reference is to the numbers of ladlefuls

poured; cf. 3. 8. 13 f. ‘cyathos . . . centum’, Alexis, PCG 2 fr. 116. 1 f.
with Arnott’s commentary pp. 324 ff. ÆE, c ªº ,
Æ =
ØºÆ ŒıŁı <f > H Ææ ø ÆæÆ , = f æE ! ¯æø
æÆØ oæ= ( #Øª ı F ÆØºø Œ ŒÆºH = ŒÆd
F ÆŒı ŒÆŁ ˜æı: = æ e æ . . . 6ºÆ #æ ,
Plaut. Stich. 706 f. ‘vide quot cyathos bibimus?—tot quot digiti sunt tibi
in manu. / cantio Graeca est: j j æÆ E j c ÆæÆ’ (in the
Greek saying ‘five’ and ‘three’ are total quantities, as is clear from Plut.
quaest. conv. 657c), Auson. XV (Green) griphus ternarii numeri 1 ff. ‘ter
bibe vel totiens ternos; sic mystica lex est . . . ’ (explicitly imitating our
passage), Marquardt–Mau 334 ff., Blümner (1911), 403. For three and
nine as significant numbers cf. 3. 22. 3 n., E. Wölfflin ALL 9, 1894–6:
333 ff., W. Deonna and M. Renard, ‘Croyances et superstitions de table
dans la Rome antique’ (Collection Latomus 46, 1961: 67 ff.). Bentley
commended aut tribus aut novem, which he found in a minor manu-
script; but the elision of (Muren)ae is unattractive, particularly at a major
pause (3. 8. 27–8 n.).
Some commentators think that H is referring to the proportions of
wine and water: as 12 cyathi make one sextarius or pint, one man drinks 3
parts of wine to 9 of water, and another 9 parts of wine to 3 of water (for
this and similar proportions see Hes. op. 596 with West, RE 4. 613). A
common mixing-bowl was generally used at the symposium, though at
the cena the drinker could follow his own taste (Marquardt–Mau 333,
Blümner, 1911: 402); the distinction is doubted by K. M. D. Dunbabin
in Murray–Tecus˛an 261, but that does not affect the issue here. Where
a toast was offered with each cyathus (cf. Call. ep. 29. 1 ª
Ø ŒÆd ºØ
Nb ,˜ØŒº ), the wine must have been already mixed in proportions
laid down by the magister bibendi. Note also T. E. Page’s comment: ‘how
wine that was palatable with the addition of three times its own bulk of
water could be drinkable when mixed with a third of its own bulk
of water I cannot conceive’.
12. y miscentur y cyathis pocula commodis: the number of ladlefuls

has to be consonant with the subjects of the toasts in vv. 9–11 (Heinze),
i.e. to be three or a multiple of three; cf. Auson. XV griphus ternarii
numeri ‘Flacci ecloga in qua propter mediam noctem et novam lunam et
Murenae auguratum ternos ter cyathos attonitus petit vates’. Others
connect commodis with the three Graces and the nine Muses of the
following lines (thus ps.-Acro); but the reference would not yet be clear.
Some again translate ‘full ladlefuls’ (cf. Plaut. rud. 1318 ‘talentum argenti
commodum magnum’); but even if H is glancing at this usage, he surely
means something more.
Horace should be expressing his wishes as the arbiter bibendi, so one
expects a verb in the present subjunctive or future indicative; he is not
laying down a rule that applies to all symposia, as the number of
ladlefuls is chosen to suit the particular occasion. Rutgers proposed
miscentor (approved by Heinsius, Bentley and Shackleton Bailey); this
‘future imperative passive’ is attested only in grammarians (Prisc. GL
3. 456 docentor, Tränkle, op. cit. 57 f.), and seems impossibly archaic for
this context. It is an even more serious difficulty that misceri seems the
wrong verb (11 n.). RN once proposed mitescent or mitescant (cf. 4. 7. 9
‘frigora mitescunt zephyris’); but it can be objected that this is a tepid
way of referring to a hot drink. He now tentatively proposes umescent or
uvescent (or the corresponding subjunctives); no exact parallel presents
itself, but cf. Lygdamus (ap. Tib. 3. 6. 5) ‘care puer, madeant generoso
pocula Baccho’.
13. qui Musas amat imparis: a humorous periphrasis for Horace him-
self; for a poet’s love of the Muses and vice versa see N–H on 1. 26. 1.
imparis avoids the obvious novem. Porph. compares Virg. ecl. 8. 75
‘numero deus impare gaudet’, and in our passage too there is a hint of
a magic number (cf. 14 n.).
14–15. ternos ter cyathos attonitus petet / vates: H humorously pro-

claims that he himself will take the larger total to match the number of
the Muses. ternos ter again suggests a magic number (cf. Virg. ecl. 8. 77
‘necte tribus nodis ternos, Amarylli, colores’); that is why Ausonius
speaks of a mystica lex (11 n.). attonitus, ‘thunderstruck’ represents the
Greek Kæ ; cf. also Archil. 120. 2W, Yfiø ıªŒæÆıøŁd
æÆ . vates, as often, has a mock-grandiloquent note; cf. 2. 6. 24
‘vatis amici’, epist. 1. 7. 11 ‘vates tuus’.
15–16. tris prohibet supra / rixarum metuens tangere Gratia: tris supra
means ‘more than three’ (OLD supra 5a); in vv. 11–15 the numbers ‘three’
and ‘nine’ form a chiasmus. rixarum metuens equals quae rixas metuit (cf.
K–S 1. 450 f.); for the avoidance of brawls cf. 1. 27. 5 with N–H. tangere
has a suggestion of ‘touching forbidden things’ and so goes well with
prohibet. The Graces were sometimes connected with Dionysus and the
symposium (e.g. Pind. O. 14. 8), where their presence had a restraining
effect; cf. 3. 21. 22 n., Panyassis fr. 13 Kinkel. For the play on the nine
Muses and the three Graces cf. Varro, Men. 333 (¼ Aul. Gell. 13. 11. 1)
‘dicit convivarum numerum incipere oportere a Gratiarum numero et
progredi ad Musarum’, though this refers to a cena, not a symposium.
17. nudis iuncta sororibus: an elegant way of referring to the three

Graces; cf. 4. 7. 5 f. ‘Gratia cum Nymphis geminisque sororibus audet /
ducere nuda choros’. iuncta describes the characteristic grouping in
art, which survives in Botticelli’s Primavera and Canova’s sculpture; cf.
E. Wind, Pagan Mysteries in the Renaissance, 1958: ch. 7, LIMC 3. 1.
203 ff., 3. 2. 151 ff., CMA 474 ff. nudis implies unadorned beauty, as at
4. 7. 6; ps.-Acro adds ‘ideo Gratiae nudae pinguntur veluti sine dolo,
sine fraude, sine aliquo velamine’. In Hellenistic and Roman art the
Graces were often portrayed naked.
18. insanire iuvat: there may be something of an oxymoron here as at

2. 7. 28 ‘dulce . . . furere’; see N–H.
18–19. cur Berecyntiae / cessant flamina tibiae?: Mt. Berecyntus in
Phrygia was one of the cult centres of the great goddess Cybele,
whose wild processions could still sometimes be observed in Rome in
spite of official discouragement (Wiseman ap. Woodman and West,
1984: 117 ff.). The Phrygian tibiae were two pipes held outwards from the
mouth and played with a band around the cheeks to keep the instru-
ment steady. One of the pipes was curved at the end and emitted a deep
note (1. 18. 13 f. with N–H, Soph. TrGF 4, fr. 513 ´æŒıÆ æ with
Pearson, Ov. fast. 4. 181 f. ‘inflexo Berecyntia tibia cornu / flabit’ with
Bömer, The New Oxford History of Music 1.404 ff. with plates 11a and 12).
For blasts of music cf. Eur. Bacch. 127 f. ŒæÆÆ ±ı fi Æ 6æıªø =
ÆPºH ÆØ. cessant suits the metaphor of ‘winds’ (TLL 3. 961.
33 ff.), though here it suggests, not a lull, but a failure to begin.
20. cur pendet tacita fistula cum lyra: the fistula is the syrinx or pan-
pipe, in which reeds of different length were laid side by side; cf. Virg.
ecl. 2. 32, D–S 4.1596 ff., RE 4A. 2. 1779, Wille 112 ff. The instruments are
pictured as hanging on a peg; cf. Hom. Od. 8. 67, Pind. O. 1. 17, Bacchyl.
20b1. As the verb implies a lack of use it is balanced by tacita (for which
cf. Call. h. 2. 12 Øøºc ŒŁÆæØ). For the combination of pipe and lyre
cf. 3. 4. 4 n.
21–2. parcentis ego dexteras / odi: dexteras implies more vigour than
manus; there must be no holding back. For odi of overt rejection see
3. 1. 1 n.
22. sparge rosas: for the scattering of flowers at a symposium cf. 1. 36. 15
with N–H, epist. 1. 5. 14 f. ‘potare et spargere flores / incipiam, patiarque
vel inconsultus haberi’. Winter roses were a luxury (N–H on 1. 38. 4), so
there is something extravagant about sparge.
22–3. audiat invidus / dementem strepitum Lycus: with the attitude

typical of a party-goer H implies that the old man will object to the row
only because he is jealous; as invidus implies ‘looking in a marked
manner’, it seems to be pointedly combined with audiat. Lycus is a
genuine Greek name (Pape–Benseler 2. 824 f., LGPN 2. 287 f. etc.),
which may here suggest a misanthropic wolf; for lone wolves cf.
R. Buxton in Interpretations of Greek Mythology (ed. J. Bremmer), 1988:
63 ff. If the latter part of the ode were a scene in the mind of the diner
and the location were still unsettled, the specific mention of Lycus and
his wife would be out of place.
dementem,which personifies the strepitum, picks up insanire (18); for
noisy parties cf. Ion of Chios 27. 7 W, Cic. Sex. Rosc. 134, Verr. 5. 31,
Prop. 3. 10. 26 ‘publica vicinae perstrepat aura viae’.
24. et vicina seni non habilis Lyco: ‘the woman next door’ (understand
nobis) is presumably the wife of Lycus rather than a mistress (who would
have had more freedom of action). non habilis means ‘incompatible’
(OLD 2), perhaps also ‘hard to handle’ (OLD 1); as she is much younger
than her husband (seni), she hears the party with lively interest. For the
repetition Lycus . . . Lyco (polyptoton in successive clauses) cf. Wills
272 ff.; it may underline the man’s disagreeable character.
25. spissa te nitidum coma: Telephus has the thick glossy hair of youth;
nitidum can also suggest hair-oil (cf. 2. 7. 7 f. ‘nitentis / malobathro Syrio
capillos’ and 1. 4. 9 with N–H). In view of the balance of te and me (26,
28), there is probably an implied contrast with H’s hair, which was going
thin and grey; cf. 3. 14. 25 n., epist. 1. 7. 25 f. ‘reddes / . . . nigros angusta
fronte capillos’.
26. puro te similem, Telephe, Vespero: in view of Vespero (the evening

star and the planet Venus), the name Telephus may have suggested
‘shining far’ (Bº and H ); for encomiastic comparison with stars
cf. 3. 9. 21, N–H on 1. 12. 47. This glittering young man cannot be
the boring antiquarian of the opening stanza, as Porph. and some
commentators suppose, nor is there anything to identify him with the
wine-pourer of v. 10. puro suggests a cloudless sky and unimpaired
beauty (2. 5. 19 with N–H); as the word can be used of a clear space
(epist. 2. 2. 71 ‘purae sunt plateae’), RN sees a formal contrast with
spissa (both adjectives are given an emphatic position). Like the
repeated Telephi at 1. 13. 1 (see N–H), the repeated te emphasizes the
girl’s fixation.
27. tempestiva petit Rhode: for the name (¼ ‘rose-bush’) cf. Men. fr.
210. 6K, Acts 12: 13, Pape–Benseler 2.1310, LGPN 2. 391. She has been
identified, without good reason, with Lycus’ wife in v. 24 (Rüpke, op.
cit.); with that name she is more likely to be a hetaera, like Glycera (28).
tempestiva means ‘the right age for you’; there is a contrast with non
habilis (24). For petere of sexual initiative (paradoxically ascribed to a
woman) cf. N–H on 1. 33. 13 and 2. 5. 16. Hints of erotic activity at the
end of a sympotic ode also occur in 1. 36. 17 ff., 3. 28. 13 ff.; for comments
on a guest’s love-affair cf. 1. 27. 10 ff., epod. 11. 9 f., Asclepiades, anth. Pal.
12. 135, Call. ep. 43. If the latter part of the poem were just imagined,
petit would have to be seen as a vivid ‘present for future’; it would then
be in a different category from torret below, which in view of lentus
should refer to a genuinely present situation.
28. me lentus Glycerae torret amor meae: Porph. says ‘lentus ignis
dicitur qui latens ac sine flamma urit’; cf. 1. 13. 8 ‘quam lentis penitus
2 0 . N O N V I D ES QVA N T O M OV E A S P ER I C LO 239
macerer ignibus’ (where N–H cite evidence for
ø and ø), Call.
ep. 44. 1 f. Ø Ø Æd e —AÆ ŒŒæı, Ø Ø Æfi = Æd a
˜Øı Fæ e fi Ð Øfi , Ð Tib. 1. 4. 81, Ov. ars 3. 573 with R. K.
Gibson. A Glycera also appears as a love of Tibullus at 1. 33. 2, who is
said to be less sweet than her Greek name implies (‘immitis Glycerae’).
H’s poems sometimes end with a depreciating contrast between himself
and somebody more successful (2. 16, 2. 17, 4. 2, Esser 9 ff., 32 f.); cf.
especially epod. 14. 15 f. ‘gaude sorte tua: me libertina nec uno / contenta
Phryne macerat’. Here he occupies an intermediate position between
the jealous and disapproving old Lycus and the glamorous and sought-
after young Telephus.
2 0 . N O N V I D ES QVA N T O
MOVEAS PERICLO
[Lyne (1980), 230 ff.; E. Oliensis ap. Woodman and Feeney (2002), 100 ff.]
1–8. Do not disturb the lioness’s cubs, Pyrrhus; you will be worsted when
she comes to reclaim Nearchus. 9–12. While you and she prepare for battle, the
boy shows his indifference, while posing exquisitely like Nireus or Ganymede.
According to traditional Roman mores homosexuality was reprehen-

sible. This attitude was beginning to change, at least in certain sections
of society, by the second century bc, when, as the result of Rome’s
conquests, an increasing influx of Greeks entered the city, bringing with
them the palaestra, the symposium, and a much freer view of sexual
relations. Polybius said that in the period after Pydna (167 bc) many
young Romans abandoned themselves to boy-friends and courtesans
(31. 25. 3 f.). The new attitudes did not go unopposed, as is clear from
remarks in the speeches of Scipio Aemilianus (Gell. 6. 12. 5, Macrob. sat.
3. 14. 7) and Cato the Censor (e.g. Liv. 39. 42. 8–12, Gell. 9. 12. 7);
Lucilius, too, had some harsh things to say (e.g. 74, 418 ff., 1058).
In assessing the evidence from the closing decades of the Republic one
has to take account of the speaker, the occasion, and the social milieu.
Cicero’s invectives against the practice (e.g. Verr. 2. 1. 32 ff., Cat. 1. 13,
Phil. 2. 44) must have been acceptable to the majority of his audience
(cf.R. MacMullen, Historia, 31, 1982: 484 ff.). But among the poets a
more indulgent attitude was common, as in Catullus’ Iuventius poems,
Virgil’s Eclogues, and the Marathus elegies of Tibullus; on the other
hand, Propertius refers only once to pederasty (2. 4. 18 ff.), and for Ovid
it had little attraction (ars 2. 683 ff.). With many versifiers it is not easy to
tell the conventional and sentimental from the realistic; cf. Shackleton
Bailey (1982), 67 ff. One has to bear in mind that the active role attracted
much less censure than the passive; and whereas stuprum with ingenui
was forbidden by the lex Scantinia de infanda venere, which was still in
force in 50 bc (Rotondi, Leges publicae populi Romani, 293), a dominus
had complete power over his own slave-boys. During his schooling
Horace was carefully protected by his father (serm. 1. 6. 81 ff., cf. Pliny,
epist. 3. 3. 3 f. and Quint. 2. 2. 15); yet as a young man he could advise the
reader to avoid adultery by exploiting a girl or boy slave (serm. 1. 2. 116 ff.;
cf. epist. 1. 18. 72 ff.). See further S. Lilja, Homosexuality in Republican and
Augustan Rome, 1983; E. Cantarella, Secondo natura: La bisessualità nel
mondo antico, 1988 (English translation, 2002); J. P. Sullivan, Martial,
The Unexpected Classic, 1991: 186 ff., 207 ff.; D. M. Halperin, OCD 720 ff.;
T. K. Hubbard, Homosexuality in Greece and Rome, 2003.
In the present ode Horace warns Pyrrhus not to appropriate the boy
Nearchus, who is in the clutches of a possessive woman; for such an
ambivalent stage of adolescence cf. 1. 4. 19 f. ‘nec tenerum Lycidan
mirabere, quo calet iuventus / nunc omnis et mox virgines tepebunt’
(a more common sequence of events), 2. 5. 21 ff., Petr. 113. 7, Mart. 11. 22.
9 f. (quoted below on 7–8). He begins with a tease by presenting Pyrrhus
as a hunter who is disturbing the lioness’s cubs; the Homeric imagery
(2 n.) is sustained throughout the poem. In the second stanza the
metaphor is clarified: the ‘lioness’ (i.e. a predatory woman) will charge
through the bands of huntsmen (i.e. rival suitors) and reclaim the
handsome boy, who is attractive to both sexes (7–8 n.). The third stanza
begins with more mock-heroic diction (9–10 nn.), but the reader’s inter-
est is then transferred to Nearchus. By standing on the victor’s palm
(11–12 n.) he shows his contempt for the whole business; cf. Theoc. 1.
33 ff. (a scene engraved on a cup) where a woman is untouched by the
wrangles of rival suitors. The final cameo has the static quality found in
descriptions of works of art (cf. 13–14 n. for a parallel on Narcissus in the
Imagines of Philostratus); and though the poem concludes with more
Homeric illustrations, they come through the medium of Hellenistic
erotic writing. Lyne (loc. cit.) thinks Pyrrhus is being warned that
Nearchus is not worth having; but the description of the boy seems
too sensuous for that. The ode really has no serious purpose, but simply
shows an amusement which is perhaps not wholly detached.
Metre: Sapphic.
1. Non vides quanto moveas periclo . . . ?: a question is livelier than a

statement; elsewhere H says nonne vides? (1. 14. 3), videsne? (epod. 4. 7), or
plain vides? (so probably 1. 9. 1 ff., 3. 27. 17 f.). moveas means ‘disturb’
(ŒØE); cf. paroem. Gr. 1. 277 c ŒØE ŒÆŒe s Œ (with parallels).
Heinze less naturally interprets as ‘amoveas’ (cf. ps.-Acro ‘coneris auferre’,
1. 10. 10 ‘per dolum amotas’ with N–H); but H is leading only gradually to
the more drastic raptor (4). quanto periclo is an ‘ablative of attendant
circumstances’ (K–S 1. 410 f., Woodcock 47); cf. the common periculo suo.
2. Pyrrhe: the Greek name as elsewhere dissociates the poem from any
actual situation, even if the pattern is true to life. Pyrrhus implies red
hair (ıææ , burrus), and may suggest a fiery temperament; cf. R.
Förster, Scriptores Physiognomici Graeci et Latini 1. 250 ‘saepenumero
capillum vidi illo etiam rubicundiorem et simul mores eorum ferarum
similes et in ipsis impudicitiam et concupiscendi amorem’. RN thinks
we are perhaps encouraged to remember Pyrrhus (Neoptolemus), the
aggressive and cowardly son of Achilles (cf. Virg. Aen. 2. 469 ff.), who
abducted Andromache at the sack of Troy.
Gaetulae catulos leaenae: the lion that protects its cubs, or looks for
them when robbed, was a motif of poetic similes from the time of
Homer; cf. Il. 17. 133 ff., 18. 318 ff. (of Achilles mourning Patroclus)
ıŒa ºÆ
ø u ºd MßªØ , = fiøffl Þ Ł e Œı
KºÆ º ±æfi Icæ = oº KŒ ıŒØB . Some writers refer explicitly
to the female of the species (a lioness or later a tigress); see Theoc. 26.
20 f. æ b ŒÆºa ıŒÆ ÆØe ºEÆ, = ‹ æ Œ
ºŁØ ŒÆ ºÆÆ , RE 13. 988. The plural catulos suits a real lioness
rather than the woman of the ode, in which only one boy is being stolen;
it can be justified because of the semi-proverbial nature of the expres-
sion, just as we could say ‘let sleeping dogs lie’ even if only one person
was involved. ‘Gaetulian’ (from north-west Africa) was a conventional
epithet of lions (cf. 1. 23. 10, Virg. Aen. 5. 351); it has associations of
savagery (Pease on Virg. Aen. 4. 40). The identification with a lioness
suits a dominating woman; it may be worth noting that the name was
borne by courtesans (Pape–Benseler 2. 779, LGPN 2. 280).
3–4. dura post paulo fugies inaudax / proelia raptor: for ‘hard-fought
battles’ cf. 2. 13. 28 ‘dura fugae mala, dura belli’, Prop. 3. 5. 2 ‘sat
mihi cum domina proelia dura mea’, Liv. 40. 16. 8, Sil. 17. 571. In our
passage the adjective, emphasized by hyperbaton, seems less to the
point, as the contest would be a short one (post paulo). RN has con-
sidered dira, ‘terrifying’; cf. periclo (1), inaudax (2), timendos (10);
so Lucan 3. 312 f. ‘at si funestas acies, si dira paratis / proelia’, 7. 689,
Sil. 8. 300, 17. 397. For similar textual problems cf. 2. 12. 2 ‘durum
Hannibalem’ (where the convincing dirum has negligible manuscript
support), serm. 1. 2. 40 ‘dura inter saepe pericla’ (where RN proposed
dira in CR 16, 1966: 327).
paulo post is the normal word-order in Cicero, but for post paulo cf.
serm. 1. 2. 120, epist. 1. 6. 43, 1. 18. 33, TLL 10. 1. 833. 8 ff. (citing Caesar,
Sallust, Livy); in H’s Sapphics the fifth syllable is never a monosyllable
unless it is preceded by another monosyllable as in dum tu (9). inaudax
(‘unbold’), not found elsewhere, is presumably a Horatian coinage
(¼ ¼º ); this is the earliest occurrence of an adjective in -ax with
a negative prefix (Wackernagel 2. 290). It makes an amusing oxymoron
with raptor, which sometimes refers to predatory animals (Virg. Aen. 2.
356 ‘lupi . . . raptores’), but here is applied to a hunter (so Mart. 8. 26. 2
where he is also frightened). In classical Latin agent-nouns in -tor
normally refer to repeated actions (cf. 3. 4. 79 amatorem), but in verse
(as in early Latin) only a single action may be represented (cf. Plaut.
Men. 65 of a river ‘rapidus raptori pueri subduxit pedes’). For the change
from bluster to cowardice cf. epod. 6. 1 f. ‘quid immerentis hospites vexas
canis / ignavus adversus lupos?’
5–6. cum per obstantis iuvenum catervas / ibit: metaphorically catervas

refers to parties of hunters (the ŁÆºæd ÆN of Hom. Il. 3. 26), but
literally to the numerous male admirers of Nearchus (cf. 1. 4. 19, 2. 8. 7 f.
‘iuvenumque prodis / publica cura’). ibit suggests purposeful movement
(OLD s.v. 7).
6. insignem repetens Nearchum: after the mock-heroic teasing the

situation now becomes explicit. The proper name is a genuine one in
Greek (LGPN 2. 328) and here suggests the literal meaning ‘leader of the
youth’ (cf. 1. 9. 4 ‘Thaliarche’). insignem ‘outstanding’ goes well with
the proper name; normally the adjective would be given further specifi-
cation, as at 1. 33. 5 ‘insignem tenui fronte Lycorida’, but here it is clear
that striking looks are the relevant point; cf. Virg. Aen. 7. 762 ‘Virbius
insignem quem mater Aricia misit’. repetens (IÆØH) suggests a claim
for what is due, and is used in legal and similar contexts; cf. 1. 15. 6 ‘quam
multo repetet Graecia milite’ with N–H.
7–8. grande certamen, tibi praeda cedat / maior an illi: if the text
is sound, then in Pöschl’s view (270 n. 8) the victor will get more
satisfaction from the sexually ambivalent Nearchus (cf. Mart. 11. 22.
9 f. ‘divisit natura marem: pars una puellis, / una viris genita est’), and
so perhaps will fight harder (RN). Some interpret maior as ‘relatively
big’ (cf. A. Ker, PCPS 10, 1964: 45 f.); the comparative is similarly used
with hostia (TLL 6. 3. 3064. 72), but after certamen and before an one
expects a genuine comparison, and the sense produced is feeble. Most
modern editors accept Peerlkamp’s conjecture ‘tibi praeda cedat, maior
an illa (sit)’ (cf. epist. 1. 10. 35 ‘minor in certamine longo’), but it is
unnatural to separate maior from praeda (especially in the third and
fourth lines of a Sapphic stanza), and the ellipse of the subjunctive sit
is unparalleled in the Odes; see E. Kraggerud (SO 57, 1982: 101 ff.).
Shackleton Bailey has considered ‘tibi praeda cedat, cedat an illi’. The
question remains open.
The construction of grande certamen is generally thought to be the
so-called ‘accusative in apposition to the sentence’, a kind of internal
accusative that extends and explains the action of the previous clause; cf.
Sall. hist. 4. 69. 8 ‘Eumenen . . . prodidere Antiocho, pacis mercedem’,
Virg. Aen. 6. 223 with Austin, Tac. ann. 1. 27. 1 with Goodyear. But the
intervention of the woman cannot be described as either a contest or a
point of dispute, as one might expect on a strict interpretation of
the appositional accusative (the usage in Tacitus becomes more free).
Accordingly we are inclined to take the phrase as a loosely attached
nominative; cf. Cic. Tusc. 1. 65 ‘nec Homerum audio qui Ganymeden ab
dis raptum ait propter formam . . . : non iusta causa cur Laomedonti tanta
fieret iniuria’ with Dougan, K–S 1. 248, H–Sz 430; when the noun is neuter,
the case may be difficult to determine (see Fordyce on Virg. Aen. 8. 683).
grandis occurs in the Odes also at 1. 6. 9, 2. 1. 11, 2. 17. 4 but 10 times in
the hexameters; in Virgil only twice in the Aeneid, but 8 times in the
Eclogues and Georgics; the word encroaches at the expense of magnus in
late Latin and especially in the Romance languages. praeda suits the
hunt as ‘prey’, but the contest as ‘prize’; for its application to a person cf.
Vell. 2. 86. 3 ‘ero praeda victoris’ (from Pollio), Petr. 80. 1, Sen. HO 511.
cedat ‘fall to’ is common in legal contexts; cf. serm. 2. 2. 134 f. ‘cedet in
usum / nunc mihi nunc alii’, epist. 2. 2. 174 ‘cedat in altera iura’.
9–10. interim dum tu celeris sagittas / promis: interim refers to the

run-up to the fight. Shooting at animals continues the Homeric illus-
tration (cf. Il. 11. 475 f. I ºÆ ŒæÆe º, ‹ Æº Icæ =
NfiH Ie ıæB ).
10. haec dentes acuit timendos: another dum-clause parallel to its

predecessor (the teeth balance the arrows); for such asyndeton in subor-
dinate clauses cf. 3. 18. 7 n. dentes acuit is again mock-Homeric; cf. Il.
11. 416 Łªø ºıŒe O Æ, 13. 474 f., Virg. georg. 3. 255, [Tib.] 3. 9. 3
(all of boars). For lions frendere is the usual term (Cic. Tusc. 2. 22
‘vestrone pressu quondam Nemeaeus leo / frendens efflavit graviter
extremum halitum’), but cf. rhet. Her. 4. 51 ‘statim, sicut e cavea leo
emissus aut alia aliqua taeterrima belua soluta ex catenis, volitabit et
vagabitur . . . acuens dentes’.
11. arbiter pugnae: it now appears that the result depends not on the
contestants but on Nearchus, who is supremely indifferent to the
squabble (see next note) and will bestow his favours simply according
to his own caprice. In Roman law an arbiter was a single official
appointed to settle a dispute, sometimes outside the normal judicial
process (Crook, 1967: 78 ff.); for a less formal use cf. Ov. her. 16. 69 of
Paris ‘arbiter es formae: certamina siste dearum’; that is the sense of the
present phrase. In other contexts the word may simply mean ‘witness’;
cf. Plaut. Cas. 90 ‘circumspicite ne quis adsit arbiter’ and OLD (1);
for the idea of ‘onlooker’ cf. Soph. Trach. 523 ff. (Deianira stands
by while Heracles fights Achelous) ± PHØ ±æa = ºÆıªE Ææ
Z
Łfiø = w, e n æı IŒÆ, Theoc. 1. 33 ff., Ov. am. 2. 12.
25 f. ‘vidi ego pro nivea pugnantes coniuge tauros: / spectatrix animos
ipsa iuvenca dabat’, Virg. Aen. 12. 718 f.
11–12. posuisse nudo / sub pede palmam: the palm was a symbol of
victory from the fourth century (not in early Greek poetry); cf. 1. 1. 5
‘palmaque nobilis’ with N–H, RE 20. 401 ff. For its metaphorical use in
erotic contexts cf. Tib. 1. 9. 81 f. with Murgatroyd, Prop. 4. 1. 140 ‘eludet
palmas una puella tuas’. Here, instead of handing the palm to the victor
(as Paris handed the apple to Aphrodite), the boy stands on it; for
similar acts of disdain cf. Lucr. 1. 78 ‘religio pedibus subiecta’, Cic. Pis.
61 (on Piso’s supposed contempt for military glory) ‘ut ad portam
Esquilinam Macedonicam lauream conculcarim’, Prop. 1. 1. 4 ‘et caput
impositis pressit Amor pedibus’, Meleager, anth. Pal. 12. 101. 4, Sittl 107.
13. fertur: the verb sustains the mock-heroic tone, as if Nearchus were a
legendary figure. At the same time it distances H from the vignette, as if
to imply that he disclaims any personal erotic interest; for Horace’s own
bisexuality cf. serm. 2. 3. 325 ‘mille puellarum, puerorum mille furores’,
epod. 11. 27 f., carm. 4. 1. 29.
13–14. et leni recreare vento / sparsum odoratis umerum capillis: a

typical picture of the puer delicatus whose long hair flutters in the breeze;
cf. 4. 10. 3 ‘quae nunc umeris involitant comae’, Chaeremon fr. 1. 5 ff.
(TrGF 1 edn. 2, p. 216) Œ ÆØ b Œæ
æø ‰ IªºÆ = ÆPEØ
æ
ØØ KŒºÆı = ıŁEØ IØ Kæø æÆØ,
Theoc. 5. 91 ºØÆæa b Ææ ÆP
Æ ŁØæÆ, Philostratus, imag. 1. 23.
5 (Narcissus) ıæÆØ ªaæ ÆPB [B Œ ] Æƒ ŒØØ K fiH æ fiø ŒÆd
Aºº KØa e Iı Ø ı ªÆØ, Rufinus, anth. Pal. 5.
28. 3 f. recreare means ‘to refresh’ in the sense of ‘cool’; cf. 1. 22. 17 f.
‘ubi nulla . . . / arbor aestiva recreatur aura’. The perfumed hair suits the
romantic picture; see N–H on 1. 5. 2, and Ov. fast. 2. 309 (of a woman)
‘ibat odoratis umeros perfusa capillis’.
15. qualis aut Nireus fuit: the MSS generally read Nereus, but Nireus was
a common exemplum of good looks, as Porph. recognizes; cf. epod. 15. 22
‘formaque vincas Nirea’ (where again Nerea has far better manuscript
support), Hom. Il. 2. 673 f. ˝Øæ , n ŒººØ Icæ e ! +ºØ qºŁ,
2 1 . O NATA M E C V M 245
Prop. 3. 18. 27, Ov. ars 2. 109, Quint. Smyrn. 7. 7 ff., Otto 243 f., RE 17.
1. 708, Burton, Anatomy of Melancholy 3. 2. 5. 3 ‘Beautiful Nireus, by that
Homer so much admired, once dead, is more despised than Thersites’.
15–16. aut aquosa / raptus ab Ida: as often in heroic poetry two similes are
combined. Ganymede was described by Homer as the fairest of mankind
(ŒººØ ŁH IŁæø), and he was snatched away to be the cup-
bearer of Zeus (Il. 20. 232 ff.); the erotic element of the story was already
made explicit by Ibycus (PMG 289) and Theognis (1345 f.), and is well
attested in Strato’s Musa Puerilis (anth. Pal. 12). In particular the abduc-
tion by an eagle was a favourite theme in art and descriptions of art; cf.
Theoc. 15. 123 f., Plaut. Men. 143 f. ‘dic mi, enumquam tu vidisti tabulam
pictam in pariete / ubi aquila Catamitum raperet?’ (significantly catamitus
is derived from Ganymedes through Etruscan), Virg. Aen. 5. 252 ff., Petr.
83. 3, Plin. nat. hist. 34. 79 (the sculpture by Leochares), J. Carcopino, La
Basilique pythagoricienne de la porte majeure, 1946: 111 ff., LIMC 4. 1. 159 ff.,
4. 2. 82 ff., CMA 1. 452 ff. For further references see Pease on Cic. nat. deor.
1. 112, RE 7. 741 ff., and cf. Tennyson, The Palace of Art 122 f. ‘Or else flush’d
Ganymede, his rosy thigh / Half-buried in the Eagle’s down’.
aquosa recalls ºıEÆ and ØÆ, the Homeric epithets for
Mt. Ida in the Troad (Il. 8. 47, 11. 183); hence Tennyson’s refrain in Oenone:
‘O mother Ida, many-fountained Ida’. That was the traditional scene of
Ganymede’s abduction; cf. RE 7. 741 ff. raptus ¼ ‘he who was snatched’
(› ±æÆŁ ), a usage particularly rare with the nominative singular; cf.
2. 16. 2 ‘in patenti / prensus Aegaeo’, Ov. am. 1. 10. 1 f. ‘qualis ab Eurota
Phrygiis avecta carinis / coniugibus belli causa duobus erat’, K–S 1. 223 f.,
H–Sz 156. The omission of the name bears witness to the fame of the story,
cf. Theoc. loc. cit., Petr. loc. cit. (Idaeum), Lucan 9. 972, Stat. Theb. 1. 548.
2 1 . O NATA M E C V M
[I. Borzsák, Acta Classica (Debrecen) 12, 1976: 47 ff.; J. N. Grant, CJ 73, 1977–8: 22 ff.;
R. G. M. Nisbet ap. Woodman and Feeney (2002), 80 ff.; Norden 143 ff.; Pasquali 613 ff.]
1–8. O born with me in the consulship of Manlius, whether you carry

reproaches or fun or easy sleep, kindly jar, under whatever title you preserve
the Massic wine, descend at Messalla’s bidding. 9–12. Though he is steeped in
Socratic dialogues, he will not disregard you; even the austerely virtuous Cato
is said to have grown hot with unmixed wine. 13–20. You apply gentle
torture to minds that are usually stiff. You give hope to the anxious and
confidence to the poor. 21–4 You will proceed with Venus and the Graces till
Phoebus returns to rout the stars.
The opening of this ode keeps the reader guessing about its purport.
First we think that Horace is addressing a woman (Grant, op. cit. with
some exaggeration); this implication can still be felt in vv. 2 and 3. At the
same time he uses the language of a hymn—a feature that continues
throughout the poem (cf. especially Norden, loc. cit., with N–H on 1. 10
and 2. 19); for details see the notes on nata (1), seu . . . sive (2), pia (4),
quocumque . . . nomine (5), moveri and bono die (6), descende (7),
tu . . . tu . . . te (13 ff.), the divine retinue (21 ff.), producent (23). At v. 4
it transpires that the poet is really addressing a wine-jar; such personifi-
cations are found in comedy (Taillardat 139 ff.) and more relevantly
in epigrams by Posidippus (3rd cent. bc) and others, e.g. anth. Pal. 5.
134, 135 (quoted 21 n.); see particularly Marcus Argentarius, ibid. 9.
229 #æ
Æ Ø, ŒÆºØŒa æÆ ØºFÆ, = hºÆº, æ7ªºø ,
h, ÆŒææı, = ÆNb KB æÆ
ıº Ø, ºªı,
= qºŁ ‹ø Kc
Eæ
æ Ø . = ÆYŁ Zº ŒÆd ¼ØŒ
Iı Ææ , = ¼Łæ ‰ Œæ æe Ø Kæ
. Here,
as in Horace, the flagon is personified as a woman, given a series of
attributes, and characterized with a number of ingenious puns (thus
Ø means both ‘drink’ and ‘husband’); the epigram may have been
influenced by the ode (it was probably written a little later); certainly
both poems drew on a common tradition.
In v. 7 we meet at last the real recipient of the ode, M. Valerius
Messalla Corvinus, the most versatile aristocrat in Augustan Rome
(see especially Syme, 1986: 200 ff., Nisbet, op. cit.). He fought for the
Republic at Philippi and for Octavian at Actium; in 27 bc he celebrated
a triumph for the conquest of Aquitania. He became consul in 31 and
prefect of the city in 26 (though he soon resigned); he was made an
augur when still under 30 and later an Arval Brother. He was the most
eloquent and charming orator of his day, and as a young man he had
written bucolic poetry in Greek (for his identification with the Codrus
of Virg. ecl. 7. 21 ff. see A. Rostagni, Virgilio minore, edn. 2, 1961: 405 ff.,
Nisbet, op. cit. 90 ff.). Among the poets befriended by him were Virgil,
Horace, Tibullus, his own niece Sulpicia, and later Ovid; the poems
written for him included the Panegyricus Messallae, Tibullus 1. 3 and 1. 7,
catalepton 9, and later the ciris.
Horace says nothing of Messalla’s achievements in war and politics,
but alludes obliquely to certain aspects of his life and personality (for
detailed suggestions see Nisbet, op. cit.). He was born in 64 bc, as is
now generally believed (Syme, 1986: 201), or possibly in 65 (Nisbet,
op. cit.), which would make him an exact coeval of Horace and the
wine-jar. The liturgical language would please a man at the heart of
Rome’s religious establishment. The Massic wine may be chosen to
recall the campaigns of his ancestor, the first Valerius Corvus (5 n.).
The allusion to Socratic dialogues (9 f.) would appeal to the Attic purist
2 1 . O NATA M E C V M 247
who translated Hyperides’ defence of Phryne (Quint. inst. 10. 5. 2) and
had studied in Athens at the same time as Horace. The reference to
Cato’s drinking habits might have an additional resonance if there was a
family connection with the younger Cato (11 n.). In the last stanza the
Graces (22 n.) are relevant to an elegant stylist in both prose and poetry.
In particular the praises of wine in the last three stanzas link up with a
fragment of Maecenas’ Symposium where in the company of Virgil and
Horace a certain Messalla commends the virtues of wine (Serv. auct.
Aen. 8. 310 ‘idem umor ministrat faciles oculos, pulchriora reddit omnia,
et dulcis iuventae reducit bona’). The identification with Messalla
Corvinus was disputed by Cichorius (Römische Studien, 1922: 233 ff.),
but his reasons were inadequate (Nisbet, op. cit. 87 f.); the nostalgia for
‘sweet youth’ in no way rules out the middle-aged Corvinus, and as he
had an estate at Arezzo (Sulpicia ap. Tib. 3. 14), which was Maecenas’
home town, Messalla made a suitable interlocutor. We may infer from
the common theme of the dialogue and the ode that Messalla was a
connoisseur of wine; it is also possible that some of the topics found in
the ode (see notes on 13–20) had figured in the dialogue. Certainly we
can be sure that Horace’s suavity and humour were calculated to suit
Messalla himself.
Metre: Alcaic.
1. O nata mecum consule Manlio: o nata suggests an address to a

woman, as at 1. 16. 1 ‘O matre pulchra filia pulchrior’; mecum recalls
poems to coevals (1. 36. 9) and comrades (2. 7. 1 with N–H). When we
see that H is using religious language, the grandiloquent O nata is
appropriate, for hymns sometimes mention a god’s birth (1. 10. 1 with
N–H). When at v. 4 we find that H is addressing a wine-jar, it becomes
relevant that natum can be used of wine; cf. serm. 2. 8. 47 ‘vino . . . citra
mare nato’, Cic. Brut. 287, ILS 8580 (quoted below); as with the English
‘bottle’, the container can stand for the contents.
The formal dating exemplifies the Roman organization of chron-
ology (Greek poetry had nothing similar), and with its prosaic associ-
ations makes a stylistic contrast with the hymnal invocation. L. Manlius
Torquatus (RE 14. 1. 1199 ff.) was consul in 65 bc, the year of H’s
birth; cf. epod. 13. 6 ‘tu vina Torquato move consule pressa meo’.
Wine-jars were sometimes dated by the year when the wine was
transferred (diffusum) from the large dolium to the amphora (epist.
1. 5. 4 ‘iterum Tauro diffusa’, ILS 8583–4, Marquardt–Mau 462 f.,
Mayor on Juv. 5. 30), sometimes by the year of the vintage; cf. ILS
8580 ‘Ti. Claudio P. Quinctilio cos. (13 bc) a.d. XIII k. Iun. vinum
diffusum quod natum est duobus Lentulis cos (18 bc) autocr(atum)’.
A wine could be known familiarly by the name of only one of the
two consuls (cf. Plin. nat. hist. 14. 94 on Opimianum), and a poet had
no room for both; though Manlius Torquatus came second to L. Cotta
on the fasti, he was the grandfather of the friend addressed in 4. 7
and epist. 1. 5, and his family had an old association with Messalla’s
(below, 5 ad fin.).
2. seu tu querelas sive geris iocos: this line is compatible with the
suggestion that a woman is being addressed: querelae suits a mistress’s
reproaches, and ioci (‘fun’) is often used in erotic contexts (1. 2. 34, epist.
1. 6. 66 etc.). The use of seu . . . sive to mark alternative attributes or
functions is a feature of the sacral style (see below, 5 n.); the mention
of two extremes ascribes comprehensiveness to the god (cf. 2. 19. 27
with N–H). geris of a person means ‘bring’, of a wine-jar ‘contain’;
cf. Mart. 4. 88. 6 ‘nec quae (testa) cottana parva gerit’, 7. 94. 1, TLL 6.
1935. 78 ff..
3. seu rixam et insanos amores: the line goes a stage further than its
predecessor: the querelae have developed into a brawl (cf. 1. 13. 10 f.
‘immodicae mero / rixae’ of lovers), the light-hearted ioci into obsessive
passion. For the conventional insanus amor cf. Virg. ecl. 10. 44, serm. 1. 4.
49, Prop. 2. 14. 18 with Enk, Otto 18 (amens amans), TLL 7. 1. 1833. 79 ff.
4. seu facilem, pia testa, somnum: H’s climax corresponds to a possible

sequence of events at a drinking-party (cf. Plat. symp. 223d). For the
soporific qualities of wine cf. Eur. Bacch. 282 with Dodds, Euenus,
anth. Pal. 11. 49. 6, Sil. 7. 205, ‘Somnus, Bacche, tibi comes additus’,
Plut. Mar. 45. 3. For facilem of sleep which comes when asked cf. 2. 11. 8
with N–H.
The separation of the vocative noun from nata (1) suits the hymnal
style (N–H on 2. 19. 8) and delays the identification of the addressee
(cf. 3. 8. 5). The everyday testa (1. 20. 2, 3. 14. 20, Hilgers 286 f.) is
humorously incongruous with the religious pia. The adjective here
means, not ‘devout’ (which does not suit a deity), but ‘kindly’, ‘loving’,
‘caring’ (qualities that reciprocate human devotion). For this rare
usage cf. Virg. Aen. 2. 536 ‘di, si qua est caelo pietas quae talia
curet’ with Austin’s note, 4. 382 with Pease, 5. 688 f. (addressed to
Jupiter) ‘si quid pietas antiqua labores / respicit humanos’, Martial 11.
3. 9 ‘cum pia reddiderint Augustum numina terris’, Weinstock 249
(citing Oscan material), RE 20. 1. 1180, Encicl. virg. 4. 95. This is
the meaning of pie Jesu, and pietas in this sense has given rise to the
English ‘pity’.
5–6. quocumque y lectum y nomine Massicum / servas: in sacral

language it was prudent to address a god by alternative names or
2 1 . O NATA M E C V M 249
epithets; cf. serm. 2. 6. 20 ‘matutine Pater, seu Iane libentius audis’,
carm. saec. 14 ff. To avoid any possibility of error, these might be
followed by the kind of blanket clause imitated here; see Catull. 34.
21 f. ‘sis quocumque tibi placet / sancta nomine’, Lucian, Tim. 1 ŒÆd Y
¼ºº ƒ Kæ Ø ØÆd ŒÆºFØ, Apul. met. 11. 2. 1–3, CIL 11. 1823
‘hunc ego aput vostrum numen demando devoveo desacrifico uti vos,
Aquae ferventes, sive vos Nimfas sive quo alio nomine voltis adpellari,
uti vos eum interemates interficiates intra annum’, Burns, To the
Deil ‘Oh thou, whatever title suit thee, Auld Hornie, Satan, Nick
or Clootie’. For further examples see Appel 76 ff., Norden, op. cit.
144 ff., Fraenkel on Aesch. Ag. 160 ff., S. Pulleyn, CQ 44, 1994: 19 ff.
H humorously suggests that different jars have different titles or func-
tions (‘bringer of sleep’ etc.), and here he covers himself against
any omissions.
lectum is usually translated as ‘choice’ (OLD 2), but this makes no
sense when the word is taken with quocumque nomine, as the order
naturally suggests. RN has therefore proposed laetum (‘in whatever
form of address the Massic that you preserve rejoices’); for similar
developments of the religious formula cf., in addition to the passages
just quoted, Plat. Crat. 400e K ÆE P
ÆE Kd E
h
ŁÆØ, ¥ Ø ŒÆd › Ł
ÆæıØ OÆ Ø, ÆFÆ ŒÆd
A ÆPf ŒÆºE, Prot. 358a, Cic. Tim. 4 ‘omne igitur caelum sive
mundus sive quo alio vocabulo gaudet . . . ’. In NR’s view the nomen
should belong not to the Massic but to the testa addressed in geris,
servas, and descende; so he would prefer laetans.
The Mons Massicus, on the north-western edge of Campania,
produced one of the finest Italian wines; cf. 1. 1. 19, 2. 7. 21, RE 14. 2.
2153. There is perhaps a personal compliment here: about 340 bc
Messalla’s ancestor Valerius Corvus may have played a part in the
conquest of the area in conjunction with the first Manlius Torquatus
(see Nisbet, op. cit. 84). servas suits both the storing of wine (2. 14. 26,
Cato, agr. 114. 2, OLD 8) and protection by a deity (1. 35. 29).
6. moveri digna bono die: the wine-jar is to be moved from its resting-
place in the apotheca; cf. epod. 13. 6 (quoted in 1 n.). As the jar is holy, the
verb may also hint at the bringing out of sacred emblems at a festival; cf.
Pease on Virg. Aen. 4. 301 ‘commotis excita sacris’ (though there the
compound suggests violent brandishing), Serv. auct. ad loc. ‘moveri
enim sacra dicebantur cum sollemnibus diebus aperiebantur templa
instaurandi sacrificii causa’; cf. epod. 11. 14 ‘arcana promorat loco’.
bono die is used of religious festivals (cf. Plaut. Poen. 497 ‘die bono
Aphrodisiis’, Catull. 14. 15 ‘Saturnalibus, optimo dierum’, Ov. fast. 1. 72,
TLL 2. 2092. 66 ff.); the day is also auspicious because of the happy
symposium.
7. descende: i.e. from the loft, where according to Columella 1. 6. 20
wines should be stored (see 3. 8. 11 n.).The verb suits not only the
personified wine-jar but also the divinity invoked in a kletic hymn
(Norden, op. cit. 148, compares KºŁ, ÆE, ƒŒF, º); cf. 3. 4. 1 ‘Des-
cende caelo’, TLL 5. 1. 642. 46 ff. (Christian instances). In a context like
this such humour would not have been thought blasphemous; for other
instances see H. Kleinknecht, Die Gebetsparodie in der Antike, 1937.
7–8. Corvino iubente / promere languidiora vina: H’s use of Messalla’s

agnomen may be intended to associate him with his ancestor’s cam-
paigns near the Mons Massicus (5 n.). The scene takes place in the
poet’s own house (otherwise descende would be discourteously impa-
tient); for similar requests by guests cf. 2. 3. 14 ‘ferre iube’, 2. 11. 18 ‘quis
puer ocius . . . ?’ promere is used of bringing out wine (3. 28. 2 n.), but
might also suit the production of sacred emblems from their cista.
languidiora means ‘vetustate lenita’ (Porph.); so 3. 16. 35 languescit of
wine. The comparative means ‘more mellow than on ordinary occa-
sions’; cf. epod. 9. 33 ‘capaciores adfer huc, puer, scyphos’. RN thinks that
the wine may be meant to suit Messalla’s relaxed demeanour.
9–10. non ille, quamquam Socratibus madet / sermonibus: Messalla is

steeped in the Socratic dialogues, especially those of Plato and Xeno-
phon; cf. ars 310 ‘Socraticae . . . chartae’, Lucil. 709M, Cic. Tusc. 3. 43,
Prop. 2. 34. 27. H is referring here to their ethical content, as quamquam
shows; in spite of his serious reading-matter Messalla enjoyed his wine.
Messalla must also have been attracted by their informal elegance of
style; cf. Cic. off. 1. 134 ‘sermo in quo Socratici maxime excellebant lenis
erat minimeque pertinax’, Plin. epist. 3. 12. 1 ‘sit expedita, sit parca
(cena), Socraticis tantum sermonibus abundet’; for his Atticizing tastes
cf. the introduction above. For madet of the drinker cf. Plaut. Cas. 245
‘ubi bibisti? mades mecastor’, OLD 3; for metaphorical instances cf. Sen.
anth. Lat. 405. 8 ‘cuius Cecropio pectora melle madent’, Mart. 1. 39. 3 f.
‘si quis Cecropiae madidus Latiaeque Minervae / artibus’ with Citroni,
TLL 8. 34. 78 ff.
10. te negleget horridus: te is emphasized by its position as first word in

its clause and as a single monosyllable before the caesura (N–H vol. 1,
p. xli). RN thinks H is suggesting that ‘a Corvinus at any rate will not
neglect Massic’ (see the observation about his ancestor in 5 n. and 7–8 n.
above). negleget suits religious dereliction (3. 6. 7 ‘di . . . neglecti’); the
future should be preferred to the variant neglegit as the drinking has not
yet begun. horridus suggests the uncouth and forbidding appearance of
the ascetic philosophers with whom Socrates was associated (Ar. aves
1554 ¼ºı , cf. nub. 835 ff. z e B ØøºÆ = IŒæÆ Pd
2 1 . O NATA M E C V M 251
P MºłÆ = P K ÆºÆE qºŁ ºı ); as it primar-
ily describes a dry unkemptness, RN sees a verbal contrast with madet
above. Of course Messalla was a model of nitor alike in his demeanour
and his writings (Quint. inst. 1. 7. 35, 10. 1. 113).
11–12. narratur et prisci Catonis / saepe mero caluisse virtus: narratur

appeals to tradition without implying that the story is false (3. 5. 41 n.);
the verb, reinforced by prisci, points to the elder Cato, censor in 184 bc
(thus ps.-Acro); cf. epist. 2. 2. 117 ‘priscis memorata Catonibus atque
Cethegis’, Sulpicia minor, sat. 48 ‘prisci sententia dia Catonis’. Plut. Cato
maior 1. 2 says he was originally called ‘Priscus’, but the name was perhaps
used to differentiate him from his great-grandson, Cato Uticensis (A. E.
Astin, Cato the Censor, 1978: 1 n. 1), and NR thinks it must have that
function here. The Censor is often mentioned as an example of old-
fashioned austerity (2. 15. 11 ‘intonsi Catonis’, Astin, 8), but he was said to
enjoy convivial occasions in the country (Plut. ibid. 25. 3); cf. Cato, agr.
156. 1 ‘si voles in convivio multum bibere cenareque libenter . . . bibesque
quantum voles’, Cic. sen. 46 ‘non intellego ne in istis quidem ipsis
voluptatibus (sc. potionis et cibi) carere sensu senectutem. . . . et pocula
(delectant), sicut in Symposio Xenophontis est, minuta et rorantia’.
RN suspects there may also be a secondary hint at the younger Cato
(praetor 52); in his case prisci would be limited to ‘old-fashioned’, and
narratur would be used humorously of a recent tradition. The younger
Cato’s heavy drinking is attested much more explicitly than the
Censor’s, and figured in Caesar’s Anticato (see Plin. epist. 3. 12. 3); cf.
Mart. 2. 89. 1 f. ‘quod nimio gaudes noctem producere vino / ignosco:
vitium, Gaure, Catonis habes’, Plut. Cato minor 6. 1. æ.Ø b fiH
æ fiø ºØÆ æ e Ø, u ººŒØ K Yfiø ØªØ N

ZæŁæ. It is possible that Messalla’s wife Calpurnia was the step-
daughter of Cato’s daughter (Syme, Roman Papers 6, 1991: 198 ff.,
Nisbet, op. cit. 2002: 86); in that case H could be implying that Messalla
had authority within his own family for drinking.
virtus Catonis, a phrase which resembles serm. 2. 1. 72 ‘virtus Scipiadae’
(cf. Il. 5. 781, ˜Ø ), refers to the tough old Roman manliness
of the Censor. If there is also a hint of the younger Cato in this stanza,
virtus Catonis could include a philosophical dimension, picking up
Socraticis sermonibus.
13–14. tu lene tormentum ingenio admoves / plerumque duro: the

repeated tu, listing a god’s functions and achievements, is a mark of
the hymnal style (Norden, op. cit. 149 ff., N–H on 1. 10. 9); for poetical
catalogues of wine’s virtues cf. 1. 18. 3 (with N–H vol. 1, p. 228), epist. 1. 5.
16 ff.; there was also a large prose literature on the symposium, notably
in the first two books of Plato’s Laws. Wine is said to apply torture
because it makes people talk; cf. epist. 1. 18. 38 ‘commissumque teges et
vino tortus et ira’, ars 434 f. For the oxymoron with lene cf. Bacchyl. fr.
20B. 6 ªºıŒE IªŒÆ (of wine). admoves (ææØ ) ‘bring to bear’
suits an instrument of torture; cf. Cic. Verr. 5. 163, TLL 1. 770. 65 ff.,
OLD s.v. 6b.
duro, ‘stiff ’, makes a contrast with tormentum. In NR’s view H is
thinking here specifically of a rack; cf. OLD, torqueo 3 (the rack was of
course operated by a twisting mechanism), Sil. 1. 176 f. ‘extenti, quantum
tormenta iubebant, / creverunt artus’; see also LSJ under æº ø II. 2.
Taciturn natures are loosened by wine; cf. epist. 1. 5. 19 ‘fecundi calices
quem non fecere disertum?’, Plut. quaest. conv. 715e ‘Some people have
an inventive nature that as long as they are sober remains unadventurous
and stiff (Ø . . . IºæÆ ŒÆd ªıEÆ), but when they
start drinking they give out exhalations like incense from the heat’.
Bentley thought that H was referring to himself, but that does not suit
sapientium curas below.
14–16. tu sapientium / curas et arcanum iocoso / consilium retegis

Lyaeo: for wine as a reliever of cares cf. 1. 18. 4 with N–H, 2. 11. 17 f.
Here ‘the wise’ are not philosophers but serious people whose anxieties
could be relaxed at a symposium (cf. 4. 12. 28 ‘dulce est desipere in loco’).
It was also a commonplace that wine reveals a person’s character and
intentions (in vino veritas); cf. Alc. 366, Theoc. 29. 1 with Gow, paroem.
Gr. 1. 85, Otto 372, W. Rösler in Murray–Tecus˛an 106 ff. arcanum
consilium seems to refer to the drinker’s own inmost thoughts; cf. epod.
11. 13 f. H is much less tolerant about the betrayal of others’ secrets (1. 18.
16 with N–H, 3. 2. 25 f., epist. 1. 18. 38, ars 434 ff., Sen. epist. 83. 9–15).
retegis makes a contrast with arcanum, which suggests something inside
an arca or strong-box; the verb is reinforced by Lyaeo ‘the Loosener’, the
cult-title of Bacchus (epod. 9. 37 f. ‘curam metumque . . . iuvat dulci
Lyaeo solvere’, N–H on 1. 7. 22, Plut. quaest. conv. 613c).
17. tu spem reducis mentibus anxiis: for the hopes of the drinker cf. 4.
12. 19 ‘spes donare novas largus amaraque / curarum eluere efficax’, epist.
1. 5. 17 ‘spes iubet esse ratas’, 1. 15. 19, Arist. eth. Nic. 1117a14 ƒ
ŁıŒ Ø . . . PºØ . . . ªÆØ. The theme was imaginatively
developed by Pind. fr. 124 and Bacchyl. fr. 20B; there as here hope is
contrasted with anxiety.
18. viresque et addis cornua pauperi: the artificial word-order (instead

of ‘viresque addis et cornua’) suggests a schema of Greek lyric, perhaps
particularly common in hymns (3. 4. 11 n.). The bull’s horns, indicating
courage and pugnacity, are often attributed to Bacchus (2. 19. 29 with
N–H); for the use here cf. paroem. Gr. 1. 302 ŒæÆÆ
Ø, Ov. ars 1. 239
2 1 . O NATA M E C V M 253
‘tunc pauper cornua sumit’, Paul. Fest. 33L ¼ 37M ‘cornua Liberi patris
simulacro adiciuntur . . . eo quod homines nimio vino truces fiunt’, TLL
4. 973. 21 ff. Wine’s relief of the poor man is part of the commonplace in
both Greek and Latin; cf. 1. 18. 5 ‘quis post vina gravem militiam
aut pauperiem crepat’ with N–H, epist. 1. 5. 20, Macedonius, anth. Pal.
11. 63. 2, together with the passages from Pindar and Bacchylides
mentioned above (17 n.).
19–20. post te neque iratos timenti / regum apices neque militum

arma: for the brachylogy post te (¼ ‘after drinking wine’) cf. 1. 18. 5
‘post vina’ with N–H. . apex refers here to the mitres of eastern kings
(Call. h. 4. 166), more usually to the conical caps of some Roman priests,
a relic of the days when the king had also been the chief priest (D–S 2.
1167 ff.); cf. 1. 34. 14 ff. with N–H (alluding both to eastern kings and to
the Roman Tarquinius Priscus). iratos can count as a ‘transferred epithet’
(cf. 1. 3. 40 ‘iracunda Iovem ponere fulmina’); yet the construction
emphasizes the intimidating character of the towering apex. For courage
given by drinking cf. 1. 18. 5 (quoted in 18 n.), epist. 1. 5. 17 ‘ad proelia
trudit inertem’, Diphilus fr. 86. 3 ff. (PCG 5 p. 104), Macedonius, anth.
Pal. 11. 63. 5 ff.
21. te Liber et si laeta aderit Venus: ‘Liber’, like the Greek ‘Lyaeus’
above (16), emphasizes the god’s liberating power. Here the wine-god
himself is a member of the wine-jar’s thiasos (for such retinues in
hymnal contexts cf. 1. 30. 5 ff., 1. 35. 17 ff.); for a more normal precedence
cf. anon. anth. Pal. 5. 135. 3 (to a flagon) ´Œ
ı ŒÆd 1ıø ƒºÆæc
ºæØ ŒÆd ˚ıŁæ . Venus is often associated with Bacchus (1. 18. 6, 1.
19. 2 with N–H, 1. 32. 9); H is discreetly referring to sexual activity at the
symposium (cf. Griffin 15 ff.), but Venus’ favour cannot be taken for
granted, while Bacchus’ presence is inevitable. laetus is regularly used of
gods’ visitations (1. 2. 46, Catull. 61. 8 f. ‘laetus huc / huc veni’, TLL 7.
888. 49 ff.), but it is particularly appropriate for Venus.
22. segnesque nodum solvere Gratiae: the three Graces or Charites

were named by Hesiod (theog. 907–9) as Aglaea (Radiance), Euphrosyne
( Joy), and Thalia (Bloom). For their presence at a symposium cf. 3. 19.
16 f., Pind. O. 14. 8. They suit the personality and literary style of
Messalla (cf. his association with Aglaie at catalepton 9. 60); see
Quint. inst. 1. 10. 21 for their connection with educated people, and
10. 1. 82 for their influence on style. The Graces were regularly
depicted arm-in-arm or hand-in-hand as a symbol of the bond of love
and friendship; according to Sen. ben. 1. 3. 4, their ring of three is ruined
if the circle is broken and is ‘pulcherrimus si cohaeret et vices servat’;
cf. Serv. Aen. 1. 720 ‘ideo connexae, quia insolubiles esse gratias decet’.
NR follows the scholiasts and most editors in referring nodum to this
traditional ring (for the metaphorical nodum cf. Cic. amic. 51 ‘amabil-
issimus nodus amicitiae’). RN prefers to interpret nodum as ‘belt’ (cf.
Mart. 9. 101. 5 ‘Scythico discinxit Amazona nodo’, OLD 2a); at 1. 30. 5
the Graces are invited to come to a hetaera’s house ‘solutis . . . zonis’, and
it is economical to take solvere in the same sense in both passages. For
this view cf. D. West, Reading Horace, 1967: 147 n. 55 ‘the point here
may be that the full delights of beauty, joy, and conviviality cannot
be rushed’.
23. vivaeque producent lucernae: for lamps as symbols of the sympo-

sium cf. 3. 8. 14–15 n.; for vivae cf. OLD 5, Eur. Bacch. 8 Ø HÆ º ªÆ
(contrast mori of dying fires, OLD 4b). producent suits a religious or
other escort; cf. Stat. Theb. 2. 313 f. ‘una soror producere tristis / exsulis
ausa vias’. It also carries the sense of prolongation (serm. 1. 5. 70 ‘cenam
produximus illam’, Mart. 2. 89. 1 f. ‘quod nimio gaudes noctem produ-
cere vino / ignosco’, OLD s.v. 10.); perhaps ‘will proceed with you’ would
bring out the two implications.
24. dum rediens fugat astra Phoebus: when dum meant ‘until’, and
there was no idea of purpose or anticipation, the indicative was normal
(as here) in republican Latin, but from the time of Livy the subjunctive
tended to encroach; see K–S 2. 378 ff., Tib. 1. 9. 61 f. ‘convivia ducere
Baccho / dum rota Luciferi provocet orta diem’, Prop. 4. 6. 85 f. ‘sic
noctem patera, sic ducam carmine, donec / iniciat radios in mea vina
dies’. For such prolonged parties cf. 3. 8. 15 n. For fugare and fugere of
heavenly bodies cf. N–H on 2. 9. 12, Fitzgerald, The Rubaiyat of Omar
Khayyam, 2 ‘has flung the stone that put the stars to flight’. Phoebus
balances not only lucernae (23) but also Liber and Venus (21).
22. MONTIVM CVSTOS

[F. Cairns, Philologus 126, 1982: 227 ff.; Fraenkel 201 f.; J. Henderson, Ramus 24, 1995: 103 ff.
¼ Writing down Rome, 1998: 114 ff.; Reitzenstein 7 ff.; W. H. D. Rouse, Greek Votive
Offerings, 1902.]
1–4. Diana, guardian of woods and helper of women in childbirth, (5–8)

let the pine that overhangs my villa be dedicated to thee, so that on each
anniversary I may gratefully offer it a male piglet.
The first stanza of this poem takes the form of a hymn. The goddess
is addressed, not by name, but by her common appellation Virgo (1 n.);
22. MONTIVM CVSTOS 255
her attributes are recorded in hymnal style by appositions and a relative
clause; cf. Norden (1913), 168. Her last title, diva triformis, is separated
from the main vocative, as often in Greek and Latin hymns (3. 21. 4 n.);
diva is a common way of addressing goddesses, and triformis bears
witness to her triple nature (4 nn.). For other hymns to Artemis or
Diana see h. Hom. 9 and 27, Alcaeus (?) 304 L–P (expounded with
supplements by Page, Sappho and Alcaeus, 1955: 261 ff.), Anacreon, PMG
348, Eur. Hipp. 61 ff. (so Sen. Phaedr. 54 ff.), Call. h. 3, not to mention
Horace’s own 1. 21 and carm. saec. (where she is addressed alongside
Apollo). An illuminating comparison may be drawn with Catull. 34.
9 ff. ‘montium domina ut fores / silvarumque virentium / saltuumque
reconditorum / amniumque sonantum. / tu Lucina dolentibus / Iuno
dicta puerperis, / tu potens Trivia et notho’s / dicta lumine Luna’; here
as in the ode we have Diana’s dominion over mountains and woods, her
assistance to women in childbirth, and her identification with other
goddesses. But Catullus is more comprehensive than Horace: he also
mentions Diana’s birth (1 ff.), her nourishment of crops (17 ff.), and
her protection of the Roman state (21 ff.).
In the ode, on the other hand, the hymnal aspect is not developed;
by the ‘crossing of the genres’ familiar in Horace (Kroll 209 ff.) the
second stanza has more of the character of a dedicatory epigram
(for dedications see below). In the article cited above, Cairns points
to such standard elements as the donor (as the Odes are written in
propria persona, Horace does not name himself ), the recipient, here
Diana, with her attributes (1–4), the gift with a short description
(5 ‘imminens villae pinus’), the formula of dedication (for some
abnormalities see 5 n.), the function of the gift (the tree is to serve
as a rustic shrine for annual sacrifices), the donor’s gratitude (6 laetus)
and the modesty of his offering (below; Cairns, op. cit. 236). It is
also natural to give a reason for the offering: thus in a Greek epigram
that begins like Horace’s ode a farmer promises Artemis a she-goat
and some lambs if she shoots robbers (Theodoridas, anth. Pal. 6. 157):
@æØ , ˆ æªØ ºÆ Œø ŒÆd IªæF, = fiø b ŒºHÆ
ºº, ı b ºı = ŒÆ Ø KØææØ ˆ æª
ØæØ Ø =
Æx Æ ŒÆd ‰æÆı ¼æÆ Kd æŁæØ . But in Horace’s ode the justifi-
cation for the tree and the sacrifice is left inexplicit; some theories will
be mentioned below.
For the dedication of trees cf. Catull. fr. 1. 1 ‘hunc lucum tibi dedico
consecroque, Priape’, Virg. Aen. 7. 62, Thyillus, anth. Pal. 6. 170. 1 ff. `ƒ
ºÆØ fiH —Æ, ŒÆd Æƒ ÆıŒ ÆyÆØ = NÆØ, l Ł ƒæa ŒIØºÆc
ºÆ = . . . ¼ªŒØÆØ, Plin. nat.hist. 12. 3 ‘prisco ritu simplicia rura
etiamnunc deo praecellentem arborem dicant’, CLE 19. 10 ff. ¼ Mus.
Lap. 149. 8 ff. (to Silvanus) ‘tu me meosque reduces Romam sistito, /
daque Itala rura te colamus praeside: / ego iam dicabo mile magnas
arbores’, Rouse, op. cit. 39 ff., Cairns, op. cit. 228 ff. Trees could also serve
as informal shrines for other dedications (Rouse, op. cit. 50 f.); thus
Callimachus told how a hunter dedicated a boar’s head to himself rather
than Artemis, only to be killed when it fell on him from the poplar tree
where it hung (aetia fr. 96 ff.; cf. Ov. Ibis 505 f.). Leonidas describes how
the antlers of a stag were fixed to a tall pine (anth. Pal. 6. 110. 3 f.); in
Erucius a steer’s horns are dedicated to Pan on a plane tree (ibid 6. 96).
For similar passages in the Roman poets cf. Virg. ecl. 7. 29 f. (to Diana),
Prop. 2. 19. 17 ff., Ov. met. 12. 266 f., Stat. Theb. 9. 585 ff. with Dewar.
Celebrations of the hunt are well attested in real life ( J. Aymard,
Essai sur les chasses romaines, 1951: 582 ff.). See for instance the inscrip-
tions on a shrine to Diana at León in Spain, belonging probably to the
second century ad (CLE 1526 ¼ Mus. Lap. 141); here a legionary
commander offered the goddess various trophies of the chase: ‘aequora
conclusit campi divisque dicavit / et templum statuit tibi, Delia virgo
triformis, / Tullius e Libya rector legionis Hiberae / ut quiret volucris
capreas, ut figere cervos, / saetigeros ut apros . . . ’, cf. Mus. Lap. 139, 140
‘umbrarum ac nemorum incolam, / ferarum domitricem, / Dianam
deam virginem . . . ’, 142 ‘Latonia sancta virago’. A relief from Sorrento
of the first century ad shows a boar being brought to Diana near a pine-
tree (Simon 57 f. with pl. 73).
Horace, however, is not offering a wild boar that has already been
killed, or even promising one in the future. Rather he is undertaking to
sacrifice an animal at present living on his estate, and to do this every
anniversary (6 n.). Therefore the verres that he promises is not a wild
boar at all (which would be called an aper), but a domestic pig, one of
the commonest sacrificial victims (3. 23. 3–4 n.); when he speaks of the
animal’s sidelong blow, just as if it were a wild boar (7–8 n), he is
whimsically comparing himself to the mighty hunters of literature and
life. Normally the victim would be of the same sex as the deity (Arnob.
7. 19 ‘feminas feminis, mares autem hostias dis maribus immolari sacri-
ficiorum iura praescribunt’), but exceptions occur (1. 4. 11 f., 2. 17. 32,
Latte 380); a wild boar is offered to Artemis at anth. Pal. 6. 240. 6, and
here the sacrifice of a male pig keeps up the parallel with hunting.
For the occasion of the ode Cairns suggests 13 August (op. cit. 237), the
foundation date of the temple of Diana on the Aventine (2. 12. 20 ‘Dianae
celebris die’); this anniversary might have been celebrated even in local
cults. For annual offerings in return for success in hunting cf. ILS 3257
‘Dianae deae, nemorum comiti, victrici ferarum, annua vota dedi’; but
after the first line Horace gives no hint of this activity, and (as noted
above) the sacrificial pig is a domestic animal. Unlike Catullus 34, the
whole emphasis of the poem is on Diana as a goddess of childbirth
(2 n.). In Greek epigrams women dedicate gifts to Artemis or Ilithyia
in return for successful childbirth; cf. anth. Pal. 6. 146, 202 (quoted in
2 n.), 270, 272, Rouse, op. cit. 251 ff.; in the Roman world the goddess’s
grove at Nemi was hung with ribbons as thank-offerings (Frazer and
Bömer on Ov. fast. 3. 267, ILS 3234 ‘pro Cn. filiod’); but in the present ode,
as some editors have noticed, such concern seems strange for a bachelor
like Horace.
Darnley Naylor thought that a friend’s wife had come safely through
a confinement; Verrall talked wildly of the infant son of Julia; Dacier
and his French followers mention the women of the Odes (Glycera and
the rest); R. S. Conway, more intriguingly, suggested that a girl on the
estate bore a child ‘in which Horace had the best of reasons to be
interested’ (PCPS 136, 1927: 37), and Quinn, after talking of Horace’s
position as paterfamilias to the slaves on the estate (serm. 2. 6. 65 ff.),
cites Mart. 1. 84. 4 ff. ‘futuit ancillas / domumque et agros implet
equitibus vernis. / pater familias verus est Quirinalis’. The hypothesis
cannot be proved; but it is not refuted by the social policy of Augustus,
which was concerned with adultery in the governing classes, not with
irregular unions in Sabine villages; and it does give a precise and
significant point to the last three lines, for the happy anniversaries will
then be the child’s birthdays.
There can be little doubt that the hills and woods which Diana is
asked to protect are those of the Sabine countryside, and that the setting
of the ode is the poet’s own estate, though he does not describe it
elsewhere as a villa. One critic has detected a pessimistic tone in the
mention of menacing pine, suffering woman, and slaughtered animal
(C. J. Carter, Proc. Class. Ass. 70, 1973: 39, even suggesting an anti-
Augustan element). Fraenkel’s sentimental approach is more appealing:
‘the simple words imminens villae tua pinus esto cause a surge of delight
and nostalgia in the heart of everyone who is as fond of the Italian
countryside as he is fond of Horace’ (op. cit. 202). By using his imagin-
ation an urban rationalist can still recapture something of the feeling for
rural cults, the communion with the spirits of the wild, a sense of the
sanctity of trees, and the significance of an annual blood-sacrifice in
gratitude for the renewal of life.
Metre: Sapphic.
1. Montium custos nemorumque Virgo: ‘Virgin that guardest the

mountains and the woods’. The whole line makes a single unit; editors
print a comma after nemorumque, but that encourages the modern
reader to pause, perhaps also to take Virgo primarily with the relative
clause (which impairs the balance). For Virgo as an epithet of Diana see
Carter 31 (with Bruchmann 49 for —ÆæŁ of Artemis); in such
contexts the word suggests fierce independence (Sen. Phaedr. 54 ‘diva
virago’) rather than maidenly modesty. By modern conventions it
should be spelt with a capital; like (Athena) Parthenos and the Virgin
(Mary) the goddess does not need to be named.
As a huntress Artemis belonged to the wilds (L. R. Farnell, The Cults
of the Greek States 2, 1896: 434, Burkert 149 ff.); cf. Hom. Il. 5. 51 f. Æ
ªaæ ` ! æØ ÆPc = ººØ ¼ªæØÆ Æ, æØ hæØ oº,
Alcaeus (?) 304. 6 L–P (see introduction above), Call. h. 3. 10. In
Roman life and poetry Diana was particularly associated with woods
(for the cult of Diana Nemorensis at the Lago di Nemi see Wissowa
247 ff.); cf. 1. 21. 5 ff., carm. saec. 1, Catull. 34. 9 ff., Virg. Aen. 11. 557, Sen.
Phaedr. 406 ff., ILS 3257 (introduction above), 3258. In our passage the
word-order highlights both genitives; cf. 3. 25. 14 f. ‘Naiadum potens /
Baccharumque’. custos is found in religious contexts for a divine pro-
tector (see serm. 2. 6. 15, Virg. Aen. 9. 405, and Carter 116); similarly
ºÆ is applied to Artemis at anth. Pal. 6. 157. 1 (introduction above).
2. quae laborantis utero puellas: Artemis was invoked by Greek

women in childbirth, an agonizing experience before the invention of
chloroform; cf. Eur. Med. 250 f. ‰ æd i Ææ IÆ = BÆØ ŁºØ
i Aºº j ŒE –Æ, Hipp. 166 f., Leonidas, anth. Pal. 6. 202
¯PŁÆ Ø ›F ŒÆd ŒÆØ = #Łd ÆæŁø
ŁBŒ oæŁ ŁıæH, = KŒ Œı, t ¸ø., Ææı ‹ f (¼
utero) / øe I Tø ºÆ B æ (with Gow–Page, HE
1955 ff.), Call. h. 3. 21 f. In Rome this function normally belonged to
Juno Lucina; cf. Plaut. aul. 691 f. ‘perii, mea nutrix, obsecro te, uterum
dolet. / Iuno Lucina, tuam fidem’, Ter. Andr. 473 ‘Iuno Lucina, fer
opem, serva me, obsecro’ (Donat. ad loc. ‘obstetriciam hanc potestatem
Iunoni attribuit quamquam illam Menander Dianam appellet’), Cic.
nat. deor. 2. 68 with Pease, Dion. Hal. ant. Rom. 4. 5. 5, Appel 86. But
sometimes the role is given to Diana, usually with other epithets; cf.
carm. saec. 13 ff. ‘rite maturos aperire partus / lenis Ilithyia, tuere matres,
/ sive tu Lucina probas vocari / seu Genitalis’, Catull. 34. 13 f. (cited
above), CIL 10. 1555 (Puteoli), ‘Dianae Loch(iae)’, Tert. anim. 39. 2
‘itaque omnes idololatria obstetrice nascuntur . . . dum in partu Lucinae
et Dianae eiulatur’. Clearly Diana suits the rustic setting better than the
matronly Juno, especially if an extra-marital relationship is involved.
For puellae in the context of childbirth cf. Ov. am. 2. 13. 19 (to Ilithyia)
‘tuque laborantis utero miserata puellas’ (imitating Horace), fast. 2. 451,
and so even in prose: Sen. epist. 24. 14, Tac. ann. 14. 64. 1 (of Octavia),
Gell. 12. 1. 4; see P. Watson, Glotta 61, 1983: 135 ff. Here the word seems
to suggest a contrast with Virgo at the same place in the previous line (an
instance of ‘vertical responsion’). It was a familiar paradox that the
virgin goddess was invoked in childbirth; cf. Plat. Theaet. 149b, Call.
aet. fr. 79 with Pfeiffer, h. Orph. 36. 4 (to Artemis), Serv. on Aen. 3. 73
‘cum Diana sit virgo tamen a parturientibus invocatur’.
3. ter vocata audis: cf. Call. h. 6. 138 ¥ ºÆŁ Ø æººØ ªÆ ŒæØÆ
Łø, ILLRP 4 ¼ ROL 4 p. 250 ‘enos Lases iuvate’ (repeated three
times in all), Virg. Aen. 6. 506 ‘magna manis ter voce vocavi’. The
number three was significant in many forms of ritual; cf. 1. 28. 36, 3. 19.
11 n., serm. 2. 1. 7, epist. 1. 1. 37, Theoc. 2. 43 with Gow, Virg. ecl. 8. 75,
Aen. 4. 510 with Pease, M. Alexiou, The Ritual Lament in Greek Trad-
ition, 1974: index s.v. ‘repetition’. audis implies that the goddess not only
hears but listens; cf. 1. 2. 27, Antip. Thess. anth. Pal. 9. 46. 5 . . . `
! æØ
KŒ (on the birth of a child), Addaeus, ibid. 9. 303, IG 14. 964 (¼
IGRR 1. 35) c ŒıæÆ ŒÆd PæªØ ŁÆ KŒ ÆæŁ [!`æ]Ø
¯ [ØÆ, O. Weinreich, Ausgewählte Schriften 1, 1969: 138 f.
adimisque leto: for the ‘dative of disadvantage’ cf. 3. 8. 11–12 n. leto is
more grandiose than morti, and leto dare was a phrase used in sacral
language (3. 7. 17 n.); the word may be derived from the name of an
Etruscan deity (cf. Lucr. 1. 852 ‘leti sub dentibus ipsis’, J. H. Waszink,
Mnem. ser. 4, 19, 1966: 249 ff. ¼ Opuscula 260 ff., RE 12. 2148). Cairns,
op. cit. 235, suggests a play on Leto, the Greek equivalent of Diana’s
mother, Latona, but the point of such a play is not obvious. It is perhaps
more significant that the phrase leads into diva triformis (4), which
recalls Hecate, Diana’s infernal aspect (see next note).
4. diva triformis: goddesses are often addressed as dea or diva, whereas

the vocative of divus is seldom used, and that of deus never in
pre-Christian Latin; cf. Löfstedt (1942), 92 ff. tri- balances ter above,
reinforcing the suggestion of ritual; cf. Virg. Aen. 4. 510 ‘ter centum
tonat ore deos Erebumque Chaosque / tergeminamque Hecaten, tria
Virginis ora Dianae’. A triple aspect is common in mythological
and religious beings; one thinks of three-headed Cerberus, the three-
bodied Geryon, the three Graiae, Fates, and Graces, the Capitoline
triad ( Jupiter, Juno, Minerva), the Holy Trinity (cf. H. Usener,
‘Dreiheit’, RhM 58, 1903: 1 ff., 161 ff., 321 ff. ¼ Vorträge und Aufsätze,
1907: 159 ff.).
In her capacity as a goddess of witchcraft and magic, Hecate was
associated with the moon (RE 7. 2778 f.); on earth she haunted junctions
where three roads met (Soph. TrGF 535, Burkert 171); and she was also a
goddess of the underworld. So a triple form was assigned to her at least
as early as the fifth century (Pease on Virg. Aen. 4. 511, Bömer on Ov.
fast. 1. 141), and she is often represented thus in Greek art (Roscher
1. 1903 ff., LIMC 6. 1. 1014 ff., 6. 2. 661 ff.). Since Artemis/Diana was, for
different reasons, a moon goddess (LIMC 2. 1. 689 f., 2. 2. 512 ff., Pease
on Cic. nat. deor. 2. 68), she also, by a typical process of syncretism,
became associated with Hecate as a goddess of crossroads and of the
underworld (Roscher 1. 1896 f., RE 7. 2770 f., LIMC 2. 1. 686 ff.). Hence
triformis came to refer to Diana’s power in heaven, earth, and the
underworld; cf. a coin of 43 bc illustrated by Beard–North–Price 2. 15,
and the three statues portrayed on a wall-painting in the ‘House of
Livia’ on the Palatine (Simon 57 with pl. 69).
5. imminens villae tua pinus esto: ‘let the pine that overhangs the
villa be thine’. tua is predicative and emphatic, which suits its position
after the main caesura of the Sapphic hendecasyllable. The possessive
has a semi-legal nuance; Cairns, op. cit. 231, cites Virg. ecl. 9. 4 ‘haec
mea sunt: veteres migrate coloni’, Ov. am. 1. 4. 40 ‘et dicam ‘‘mea
sunt’’ iniciamque manum’, and the formula of mancipatio ‘hunc ego
hominem ex iure Quiritium meum esse aio, isque mihi emptus esto
hoc aere aeneaque libra’ (note the formal esto, as in our passage).
Horace’s language is unusual for a dedication, which normally uses a
deictic pronoun (e.g. hanc) and a verb like do, dono, or dedico; in the
same way the subjunctive donem (8) avoids the formulae of genuine
inscriptions.
As well as being in harmony with the intimate setting of Horace’s
villa (cf. 2. 3. 9, 2. 11. 14, Henderson, op. cit. 110 ff.), the umbrella pine
fits Diana as ‘montium custos nemorumque Virgo’ (just as elsewhere it
is associated with Pan). Editors compare the ‘sacral-idyllic’ landscapes
of Campanian wall-painting, see 3. 13. 14–15 n., but Horace’s poem is
typically less cluttered and concentrates on a single picturesque pine.
6. quam per exactos ego laetus annos: ego balances tua (5) and quam
balances quae (2). per exactos annos means ‘as each year is completed’;
cf. 2. 3. 6 f. ‘per dies / festos’ with N–H, 2. 14. 15 ‘per autumnos’. The
phrase shows the Roman feeling for the calendar and anniversaries; cf. 3.
8. 9 ‘anno redeunte’, Virg. Aen. 5. 46 ‘annuus exactis completur mensibus
orbis’, 8. 268 f., ILS 3257 ‘annua vota’ (introduction above). laetus, which
commonly occurs in such rituals, shows the joy of the worshipper as he
pays his return for services rendered; cf. the abbreviation in inscriptions
VSLLM ¼ ‘votum solvit laetus libens merito’ (ILS 3. 2, p. 795).
7–8. verris obliquum meditantis ictum / sanguine donem: a sacrifice to

the goddess’s tree is a sacrifice to the goddess herself. The piglet is
already practising with its tusks (cf. Virg. georg. 3. 255 ‘dentesque Sabel-
licus exacuit sus’); as with the kid in 3. 13. 6 f., the animal’s promise will
not be fulfilled. meditatur suits poetic descriptions of a wild boar about
to charge; cf. Hes. scut. 387 Œæ
ÆıºØ ø æØ ŁıfiH Æ
ÆŁÆØ,
Pind. fr. 234. 3 f. Œæfiø b ıºØ ŒÆ
æ = ºŁı
KıæE. In the same way obliquum alludes to the sideways cut familiar
from descriptions of boar-hunts: Hom. Il. 12. 148
ø I.,
Od. 19. 451 ºØŒæØd I.Æ , Hes. scut. 389
øŁ , Ov. met. 8. 344 (of
the Calydonian boar) ‘obliquo latrantis dissipat ictu’.
2 3 . CA ELO S V P I NA S S I T V LER I S M A N V S
[A.Barigazzi, Rivista di cultura classica e medioevale 18, 1976: 71 ff.; F. Cairns, AC 46, 1977:
535 ff.; A. Dubourdieu, Les Origines et le développement du culte des Penates à Rome, École
française de Rome, 1989: esp. 89–91; Pasquali 603 ff.; M. J. C. Putnam in Rome and
her Monuments (ed. S. K. Dickison and J. P. Hallett), 2000: 521 ff.; F. A. Sullivan, CP 55,
1960: 109 ff.]
1–8. If you sacrifice to the Lares at the new moon, Phidyle, the crops and
animals will not suffer in the unhealthy autumn. 9–16. The pontiffs of the
state cults may slaughter cattle, but for you a sweet-smelling garland is a
sufficient offering. 17–20. The hand that has brought no contribution can
placate the Penates with grain and salt.
The Lares had no Greek counterpart, and although Festus p. 108L

says ‘Lares . . . animae putabantur esse hominum redactae in numerum
deorum’, their origin is still unclear (OCD 815 f.). Their images stood by
the hearth, sometimes in a cupboard, where offerings were regularly
made to them (3–4 n., 15–16 n.). Their protection included the country-
side (as in Horace’s poem), and the Lares compitales were worshipped
at simple roadside shrines. They were particularly important for the
underprivileged and excluded, that is to say for women, freedmen, and
slaves (cf. epod. 2. 65 f., serm. 1. 5. 66). When Cato told his vilica that
religion was his business, he made an exception for the Lares (agr.
143. 2): ‘Kalendis, Idibus, Nonis, festus dies cum erit, coronam in
focum indat, per eosdemque dies Lari familiari pro copia supplicet’;
here pro copia means ‘as resources permit’, which also suits the tenor
of our poem. See further Wissowa 166 ff., RE 12. 814 f., Latte 90 ff.,
D. G. Orr, ANRW 2. 16. 2. 1563 ff., Simon 119 ff. (with lively illustra-
tions), T. P. Wiseman, JRS 91, 2001: 185 ff.
When Horace instructs Phidyle on her offerings to the Lares, he does
not imply criticism: good advice (paraenesis) is a regular feature of his
lyrics, which give the illusion of speech, and is a way of suggesting the
ethos of the addressee (N–H vol. 2, pp. 3 f.). He gives the woman a
significant Greek name (2 n.), and so avoids the banality of mentioning
an individual Italian countrywoman. The place is not identified either,
though the reference to the Alban Hills (9 ff.) would suit the general
vicinity of Rome; we cannot specify the poet’s Sabine estate, as the
fruitful vine of v. 6 did not grow there (epist. 1. 14. 23). But though not as
particularized as a modern might expect, the ode reflects the sympathy
for ordinary country people that Horace shows elsewhere (cf. serm. 2. 2
on Ofellus, 2.6 on the country mouse). At the same time its idealization
of rustic piety is a truly Augustan theme (cf. Virg. georg. 2. 493 ff. etc.).
Horace is unlikely to have believed that the Lares could protect the
crops; but the traditional cults required no formal creed, and rituals were
performed because they always had been. It is important to distinguish
emotional from intellectual sincerity; elsewhere Horace repeatedly
demonstrates his love of the countryside, and religion formed an inte-
gral part of that life (cf. 3. 13 on Bandusia, 3. 18 on the Faunalia, 3.22 on
the sacrifice to Diana). The recurring ceremonies followed the rhythms
of the natural world, linked present to past, and gave the participant a
sense of belonging. This may have been important for the poet as well as
for a peasant.
The poem claims that the widow’s mite is as acceptable to the gods
as the most grandiose sacrifice; this is a commonplace, sometimes
with the added thought, implicit here, that the piety of the donor is
what matters. See Hes. op. 336 Œa ÆØ æØ ¥ æ IŁÆØØ ŁEØ
= ±ªH ŒÆd ŒÆŁÆæH (quoted with approval by Socrates according to
Xen. mem. 1. 3. 3), Eur. TGF 327. 4 ff. (no doubt reflecting sophistic
enlightenment) Kªg b ººŒØ øæı = Æ ¼æÆ NæH
H ºıø = ŒÆd <f > ŁEØ ØŒæa ŁÆ º = H ıŁıø
ZÆ Pæı , 946, [Plat.] Alc. 2. 149e ŒÆd ªaæ i Øe Y N æe
a HæÆ ŒÆd a ŁıÆ IºıØ H ƒ Łd Iººa c æe c
łı
, ¼ Ø ‹Ø ŒÆd ŒÆØ J ıª
fi . Horace’s ode has a particular
affinity with various epigrams of the Greek Anthology; cf. Archias, 10. 7.
5 ff. (hecatombs unnecessary), Antip. Thess. 9. 93. 4 (incense enough), as
well as the dedications cited by Cairns, op. cit. 537 (though he goes too far
when he assigns our ode to the category of ‘anathematikon’). For in-
stances of the topic in Latin cf. Cic. nat. deor. 2. 71 with Pease, leg. 2. 24
‘probitatem gratam esse deo, sumptum esse removendum’, paneg. Mess.
14 f. (cited below, 19–20 n.), Ov. trist. 2. 75 f. ‘sed tamen ut fuso taurorum
sanguine centum, / sic capitur minimo turis honore deus’ (with Luck’s
parallels), Pont. 4. 8. 39 f., Sen. ben. 1. 6. 3 ‘itaque boni etiam farre ac fitilla
religiosi sunt’, Pers. 2. 75 ‘haec [spiritual and moral qualities] cedo ut
admoveam templis, et farre litabo’, Plin. nat. hist. praef. 11 ‘verum dis lacte
rustici et mola tantum salsa litant qui non habent tura’. In the Judaeo-
Christian tradition cf. 1 Samuel 15: 22 ‘to obey is better than sacrifice, and
to hearken than the fat of rams’.
As always the ode is carefully structured (R. E. Deutsch, CW 36,
1942–3: 165 ff., with some overstatement). The first two stanzas list the
offerings that Phidyle is to make and the blights that will be avoided:
fruge (4) is picked up by seges (6), and porca (4) by alumni (7); the
correspondence between ture (3) and vitis (6) is less clear, though wine
and incense are combined at 3. 18. 7 f. The third stanza describes the
cattle sacrificed by the pontiffs in two closely parallel clauses; the fourth
mentions the two types of simple garland offered by Phidyle. The fifth
stanza, with its ethical focus and token offerings, forms the culmination
2 3 . CA ELO S V P I NA S S I T V LER I S M A N V S 263
of the ode; Horace has already proceeded from incense and a pig to
garlands of rosemary and myrtle, now he says that even emmer and salt
on their own are sufficient. The final stanza, though a general statement
rather than an address to Phidyle, balances the first: ‘si tuleris manus’ (1)
is answered by ‘tetigit manus’ (17), ‘placaris . . . Lares’ (3 f.) by ‘mollivit
Penates’ (19), ‘fruge . . . avidaque porca’ (4) by ‘farre et saliente mica’
(20) (the latter phrase also balances ‘marino rore . . . fragilique myrto’
in 15 ff.).
The date of the ode is uncertain, as is usually the case in non-political
poems. Syndikus suggests that the diction points to an early date
(2. 193 n.): sumptuosa and nihil attinet are prosaic, while pomifer is an
unusual compound for Horace’s middle period. These arguments carry
little weight, as variations of style depend on the nature of the poem
rather than on the exact time of writing; yet the conclusion may be
correct. The last line of the ode is closely similar to paneg. Mess. 14
(see below, 19 n.); it is unlikely that Horace would imitate so inferior a
poem, so if the panegyric is correctly assigned to 31, when Messalla was
consul (Syme, 1986: 200 and 203), the ode may belong to the triumviral
period, when the Georgics were being written. The metre is unusual at
v. 7 ‘robiginem aut dulces alumni’, where the third line of an Alcaic
begins with an elided quadrisyllable; yet cf. 3. 5. 7 and 3. 29. 59.
Metre: Alcaic.
1. Caelo supinas si tuleris manus: when praying, the ancients regularly

stood with hands uplifted; cf. D. Aubriot-Sévin, Prière et conceptions
religieuses en grèce ancienne, 1992: 127 ff., Pulleyn 188 ff. The palms were
upturned (supinas), usually ‘angled outward at about forty-five degrees’
(Pulleyn 189); he cites Boëdas’s statue of ‘The Boy at Prayer’ (for an
illustration see Enciclopedia dell’arte antica 2, 1959: 123); cf. [Aesch.]
Prom. 1005 ªıÆØŒØ ØÆØ
æH, Virg. Aen. 3. 176 f. ‘ten-
doque supinas / ad caelum cum voce manus’, 4. 205 with Pease, Liv.
26. 9. 7, Plut. comp. Philop. et Flam. 2. The upturned hands could
indicate submission, as it was used by suppliants in other contexts; cf.
Liv. 3. 50. 5 ‘supinas . . . tendens manus’, Petr. 114. 4, Sittl 148 f., 174 n. 2;
but the posture could also indicate an expectation of bounty; cf. Ar.
eccl. 781 ff. ‹Æ ªaæ P
ŁÆ Ø ÆØ IªÆŁ, = (Œ KŒÆ c
Eæ Æ = P
u Ø Iºº ‹ø Ø ºłÆØ, Suet. Vit. 7. 3
‘advenientem . . . exercitus . . . libens et supinis manibus excepit, velut
dono deum oblatum’.
For the dative caelo cf. Virg. Aen. 2. 688 ‘caelo palmas cum voce
tetendit’, 5. 457 ‘it clamor caelo’. According to Löfstedt (1942), 180 f.,
the construction is made easier because caelo implies deis; he compares
Pind. I. 6. 41 › IÆÆØ PæÆfiH
EæÆ I
ı with Hom. Il. 3. 318
ŁEØ b
EæÆ I
. tuleris stands for sustuleris; cf. Virg. Aen. 2. 153
‘sustulit . . . ad sidera palmas’. The usage may be an archaism, stereo-
typed in a religious context; cf. TRF incert. 70 ‘tetulit senilis Poeas ad
caelum manus’, where the reduplicated form of the perfect is certainly
archaic.
2. nascente Luna, rustica Phidyle: cf. Cato, agr. 143 cited above (offer-
ings to the Lares on the Kalends among other dates), Tib. 1. 3. 34
‘reddereque antiquo menstrua tura Lari’, more cynically Prop. 4. 3. 53 f.
‘raris adsueta kalendis / vix aperit clausos una puella Lares’. Phidyle
makes her offering at the new moon, which originally marked the
Kalends (Liv. 1. 19. 6, Macrob. Sat. 1. 15. 20, Wissowa 186 f.); it was
still an illiterate peasant’s best way of observing the calendar (cf. Virg.
georg. 1. 276 ff. with Mynors). There is nothing in Horace’s words to
confine the offering to the first of May, when sacrifices were made for
the crops (thus Cairns, op. cit. 538). nascente Luna is not simply a
poeticism; cf. Cic. acad. 2, fr. 2 ‘(lunae) nascentis et insenescentis’,
Hor. serm. 2. 4. 30 ‘lubrica nascentes implent conchylia lunae’.
Phidyle is a Greek name that means ‘sparing’; it is attested in an Attic
inscription, LGPN 2. 444, where the diminutive 6ıººÆ is also cited
(cf. also 3A. 445); so Bentley’s Phidyli can be forgotten. Here rustica does
not mean just ‘living in the country’, but suggests simplicity and inno-
cence, without any disparagement.
3–4. si ture placaris et horna / fruge Lares avidaque porca: tus was an
aromatic resin imported largely from S. Arabia (RE Suppl. 15. 700 ff.,
N. Groom, Frankincense and Myrrh, 1981, OCD 752 f.); its granules were
burnt in religious ceremonies to evoke reverent emotions (M. Detienne,
Les Jardins d’Adonis, 2nd edn. 1989) and to exclude less agreeable smells
(cf. 3. 18. 7 f.). It was widely enough distributed to be regarded as a
simple offering (so Prop. 2. 10. 24 metaphorically of his poetry ‘pauper-
ibus sacris vilia tura damus’), and was given to the Lares at all periods;
cf. Plaut. aul. 23 f. (spoken by the Lar) ‘ea mihi cottidie / aut ture aut
vino aut aliqui semper supplicat’, Tib. 1. 3. 34, Ov. fast. 2. 631, Juv. 9. 137,
Prud. perist. 10. 261 ‘fuliginosi ture placantur Lares’. In early Latin the
final syllable of words like placaris is short in the future-perfect though
long in the perfect subjunctive (Neue–Wagener, 3. 428 ff.). In Horace’s
odes two other instances in the future-perfect are unambiguously long
(4. 7. 20, 21) and one short (4. 10. 6); in his hexameters two are
unambiguously long (serm. 2. 2. 74, 2. 5. 101) and nine short (Bo 86).
horna ‘of this year’ is applied to rustic things, perhaps reflecting rustic
speech (so epod. 2. 47 of wine, serm. 2. 6. 88 of straw, Prop. 4. 3. 61 of
lambs); here the adjective is emphasized, as the grain must be of the
most recent harvest (Serv. on Aen. 2. 133, citing our passage). Pork was
the basic meat in Italy, as appears already from Plautus ( J. André,
L’Alimentation et la cuisine à Rome, 1961: 139 ff., E. Gowers, The Loaded
Table, 1993: 69 ff.). Pigs were the most common sacrificial victims, at
least in private cult (Wissowa 411 f., Bömer on Ov. fast. 1. 349), and are
offered to the Lares at serm. 2. 3. 164 f., Plaut. rud. 1208, Tib. 1. 10. 26
‘hostiaque e plena rustica porcus hara’; for an illustration from Pompeii
of a man carrying the pig over his shoulders cf. ANRW 2. 16. 2. 1583. The
victim is a sow perhaps because it was supposed to be more efficacious
(Serv. on Aen. 8. 641 ‘quia in omnibus sacris feminini generis plus valent
victimae’); it is not clear whether Phidyle’s sex is relevant here. Pigs are
conventionally greedy (ps.-Acro ‘avida edaci’) and so by implication fat;
the humorous touch would not occur in actual sacral language, but as
Horace’s phrase is repeated by Ovid at fast. 1. 349 there is no need to
think it corrupt.
5–6. nec pestilentem sentiet Africum / fecunda vitis: in the same way
Cato’s farmer prays that Mars will ward off ‘morbos . . . calamitates
intemperiasque’ from his animals and crops as well as from his shep-
herds and household (agr. 141. 2). The hot dry Sirocco from the Sahara
(called Atabulus at serm. 1. 5. 78) was noxious both to vegetation and to
living creatures (2. 14. 15 f. ‘nocentem / corporibus metuemus Austrum’
with N–H, RE 17. 1118, and see further on grave tempus below). For
sentire of unpleasant experiences cf. N–H on 2. 7. 9; for its application to
inanimate objects cf. ars 65 f. ‘palus . . . sentit aratrum’ with Brink.
6–7. nec sterilem seges / robiginem: robigo is mildew (Virg. georg.

1. 151), known in English sometimes as ‘rust’, and recognizable in its
early stages by its reddish spores. At the Robigalia on 25 April sacrifices
were sometimes made to Robigo or Robigus (Ov. fast. 4. 907 with Bömer,
Wissowa 196, Latte 67), to the derision of the Christian fathers. sterilem
by a common type of transference is applied to the agent of sterility
(Lucan 4. 108 ‘sterili frigore’, OLD 2c); the word balances pestilentem (5)
and grave (8), as well as making a contrast with fecunda (6).
7–8. aut dulces alumni / pomifero grave tempus anno: the alumni
(cf. 3. 18. 4) are the unweaned lambs and kids. dulces is readily applied
to children, as at Lucr. 3. 895, and here suggests the engaging behaviour
of young animals, perhaps from Phidyle’s point of view; cf. Virg. georg.
3. 178 ‘sed tota in dulcis consument ubera natos’ (though that is pre-
sented as the cow’s attitude). tempus refers to climatic conditions (Lucr.
5. 231 ‘nec varias quaerunt vestis pro tempore caeli’, OLD 11b). grave
combines the ideas of ‘oppressive’ and ‘dangerous’; cf. serm. 2. 6. 18 f.
‘plumbeus Auster / autumnusque gravis, Libitinae quaestus acerbae’,
Cic. fam. 5. 16. 4 ‘hoc gravissimo et pestilentissimo anno’, Sen. ben.
6. 38. 3 ‘medicis gravis annus in quaestu est’. pomifero anno means
‘pomifera parte anni’ (Porph.); for this use of annus with an adjective
cf. epod. 2. 29 ‘annus hibernus’, Virg. Aen. 6. 311 ‘frigidus annus’ (for
some similar locutions see Gow on Theoc. 7. 85, where he takes
uæØ to mean the spring). Autumn was a bad time for malaria, which
comes from a parasite injected into the bloodstream by the female
mosquito; in Italy mosquitoes hibernate, and attacks are confined to
late summer and early autumn. The cause of malaria (‘bad air’) was not
known until the 1890s, and in antiquity it was associated with the
oppressive wind (which was simply an accidental concomitant). But it
was recognized that swampy districts were especially dangerous; Silius
(8. 379) speaks of ‘pestifera Pomptini uligine campi’, and earlier Varro
had advised against building a villa in such a place; he actually refers to
‘animalia quaedam minuta, quae . . . per aera intus in corpus per os et
nares perveniunt’ (rust. 1. 12. 2). For other references to unhealthy
autumn see N–H on 2. 14. 15, Courtney on Juv. 4. 56.
9–10. nam quae nivali pascitur Algido / devota: nam is to be taken

with ‘te nihil attinet’ (13); the intervening stanza, though co-ordinate in
syntax with the fourth, is subordinate in sense; a countrywoman would
be more interested in the rearing of the animal than in the official
sacrifice in the city. Mons Algidus (RE 1. 1476) is a high ridge to the
east of the Mons Albanus (Monte Cavo), about 15 miles south-east of
Rome; nivali plays on the name (cf. 1. 21. 6 gelido, and Stat. silv. 4. 4. 16
where horrens ¼ ‘shivering’); for etymological puns on proper names see
O’Hara (1996), Paschalis (1997). The cool saltus provided excellent
summer pasture (Varr. rust. 2. 5. 11 ‘aestu abigantur in montes frondo-
sos’), though obviously not when there was snow on the ground as in
epist. 1. 7. 10 ‘quodsi bruma nives Albanis illinet agris’. Particular areas
were designated for the perfect cattle needed at public sacrifices (for the
Clitumnus cf. Virg. georg. 2. 146 ff., Juv. 12. 13 ff.; for the ager Faliscus cf.
Ov. fast. 1. 84 with Bömer); such an area at Alba was said to have been
reserved from the regal period (Dion. Hal. ant. Rom. 3. 29. 6). pascitur
suggests that while the animal is peacefully grazing it is doomed (devota
carries some emphasis); cf. epist. 1. 3. 36 ‘pascitur in vestrum reditum
votiva iuvenca’ and 11 n. below.
10. quercus inter et ilices: quercus is the common oak, illustrated in

Abbe (1965), 82; Liv. 3. 25. 7–8 depicts such a tree on Algidus. The ilex is
the evergreen holm-oak (Abbe p. 86); J. Sargeaunt describes a line of
them at Albano (The Trees, Shrubs, and Plants of Virgil, 1920: 62). The
leaves of the holm-oak are dark (cf. 4. 4. 57 f. ‘duris ut ilex tunsa
bipennibus / nigrae feraci frondis in Algido’); so there is a colour-
contrast with nivali. For the sandwiched position of inter, found also
in prose, cf. epist. 1. 4. 12 ‘timores inter et iras’, K–S 1. 587, TLL 7. 1. 2147.
28 ff.; the stylistic function here and in most other cases is not to
represent the idea of ‘coming between’ (an overworked modern notion),
but simply to separate two parallel nouns by a word of less importance.
11–12. aut crescit Albanis in herbis / victima: closely parallel to ‘nivali

pascitur Algido / devota’; as Albanis suggests ‘white’, there seems again
to be a colour-contrast with the green grass. crescit is used of fattening
for sacrifice; cf. the witty metaphor at 2. 8. 17 ‘adde quod pubes tibi
crescit omnis’ with N–H; for a similar use of iuvenescit cf. 4. 2. 55.
victima (a larger animal than hostia) belongs to the relative clause; cf.
2. 3. 24, where even without devota the word has the sense of ‘destined
victim’ (see N–H’s note).
12–13. pontificum securis / cervice tinguet: by its prominent position

pontificum makes a contrast with te (13); varying sacrifices reflect a
difference of status (N–H on 2. 17. 32, Headlam on Herodas 4. 16).
A future like tinguet is sometimes called ‘concessive’, but the prediction
is more confident than with etsi; cf. 1. 7. 1 ‘laudabunt alii’ with N–H.
cervice is a gruesome brachylogy for cervicis sanguine; cf. Prop. 4. 1. 111 f.
‘idem Agamemnoniae ferrum cervice puellae / tinxit’. The actual blow
would be delivered, not by the pontifex, but by the attendant (popa); for
the procedure cf. Latte 388. For sacrificing cattle an axe was sometimes
used rather than a knife; cf. Virg. Aen. 2. 223 f. ‘fugit cum saucius aram /
taurus et incertam excussit cervice securim’, Ov. met. 12. 248 f., fast.
3. 805, trist. 4. 2. 5 f. (there is an illustration in R. M. Ogilvie, The
Romans and their Gods, 1974, facing p. 41). The singular securim (adopted
by Bentley, Kiessling, and Heinze) has a little manuscript support, and
would be more precise after victima (contrast Prud. apoth. 461 f. ‘ponti-
ficum festis ferienda securibus illic / agmina vaccarum steterant’); but
the animal represents a class, and plural pontificum suits plural securis
(unless indeed several priests are thought of as acting together). It is true
that the homoeoteleuton with herbis (11) is unusual in the third and
fourth lines of an Alcaic stanza unless the words agree (as at 3. 3. 59 f.
avitae . . . Troiae) or correspond (as at 3. 3. 55 f. ignes . . . rores); yet there
are exceptions, as in 2. 1. 39 f. antro . . . plectro. Bentley’s objection to
successive lines ending in -s (ilices, herbis, securis) is hypercritical; cf. 1. 2.
13 ff., 1. 12. 10 ff., 1. 15. 14 ff.
13–14. te nihil attinet / temptare multa caede bidentium: for the

unpoetical nihil attinet (‘it is not for you’) cf. 1. 19. 12 ‘quae nihil attinent’
with N–H (citing Axelson 101). temptare means ‘to attempt to influence’;
cf. ars 404 f. ‘gratia regum / Pieriis temptata modis’, Virg. Aen. 4. 113
‘tu coniunx, tibi fas animum temptare precando’ (with Pease), OLD s. v.
6. Peerlkamp thought the word unsuitable for religious supplications,
but it seems no more objectionable than 1. 35. 5 ‘ambit sollicita prece’
(of prayers to Fortuna). More defensibly, he disliked taking deos (15) as
the object of temptare as well as coronantem; to put it another way, the
clause naturally ends with a pause after bidentium, and this would
be easier if the infinitive were intransitive. He proposed certare (not
too far from tentare), citing 4. 1. 31 ‘nec certare iuvat mero’, Lucr. 2. 11
‘certare ingenio’, Virg. ecl. 2. 57 ‘si muneribus certes’ (where the idea of
competition is more explicit). This suggestion has disappeared from
view but is favoured by RN; there was no question of offering sheep to
the Lares.
bidens ‘two-toothed’ is a word for a sacrificial sheep. In its second year
(the right time for sacrifice) a sheep has two second-phase teeth that are
prominent compared to the six surviving primary teeth; see the lucid
account by Henry on Aen. 4. 57, cited and supplemented by Pease.
caedere is used in sacral contexts for ‘to sacrifice’, but the noun at this
date may have had a hint of excessive slaughter, especially when accen-
tuated by multa. Instead of saying ‘the pontiffs can sacrifice cattle and
sheep’, H by a characteristic distribution puts the sheep into a later
clause (for parallels see 3. 1. 45–6 n.).
15–16. parvos coronantem marino / rore deos fragilique myrto: the

images of the Lares (for dei in this sense cf. 2. 18. 26 f. with N–H) are
literally small (Ov. fast. 5. 130); the adjective suggests less directly that
Phidyle has a humble home (cf. Juv. 9. 137 f. ‘o parvi nostrique Lares
quos ture minuto / aut farre et tenui soleo exornare corona’ with
Courtney); and its emphatic position underlines the contrast with the
public grandeur of what goes before. coronantem is not negatived by nihil
attinet, and implies ‘provided that you wreathe . . . ’; for offerings of
garlands to the Lares cf. Plaut. aul. 23 ff., 385 f. ‘nunc tusculum emi
hoc et coronas floreas: / haec imponentur in foco nostro Lari’, Cato, agr.
143. 2, Juv. 12. 87 with Mayor, D. G. Orr, ANRW 16. 2. 1583 (nails at
Pompeii to hold garlands). marino rore (rosemary, Rosmarinus officinalis),
‘gets its name from its liking for sea coasts and spray’ (Sargeaunt, op. cit.
(10 n.), 110); see also RE 1A. 1128 f., Mynors on Virg. georg. 2. 213, Bömer
on Ov. fast. 4. 741, Abbe 176 with illustration; according to [Apul.] herb.
79, ‘antequam tus sciretur hac herba deos homines placabant’ (for such
plants see N–H on 1. 19. 14 verbenas). Myrtle (Sargeaunt 82 f., Abbe 146
with woodcut) is common on the coasts of S. Italy; its liking for the
shore may account for its dedication to Venus, who was worshipped at
Paphos on the coast of Cyprus; hence Virgil’s Paphiae . . . myrtus (georg.
2. 64). It is combined with rosemary as another aromatic shrub (so
Ov. ars 3. 690), and it is offered to the Lares at Tib. 1. 10. 28. fragili
represents the Greek ŁæÆı (used of another shrub by Theophrastus,
hist. plant. 5. 3. 6); a sprig is ‘easily broken off ’, and this accords with the
simplicity of Phidyle’s offering.
17. immunis aram si tetigit manus: immunis bears its normal sense of
‘not making a contribution’ (here ‘not bringing an offering’); cf. 4. 12.
22 f. ‘non ego te meis / immunem meditor tingere poculis’, epist. 1. 14. 33,
TLL 7. 1. 505. 47 ff. If someone is unable to contribute even a garland
(immunis is stressed by hyperbaton), then a symbolic offering of grain
and salt is acceptable (see also below on 19); some commentators object
that grain and salt are also a contribution, but these have no intrinsic
value. Porph. glosses ‘understand scelerum’ (cf. Vell. Pat. 2. 7. 2 ‘immu-
nisque delictorum paternorum’) and Barigazzi, op. cit., explains ‘immu-
nis caedis’ (cf. Ov. her. 14. 8), but to produce such a meaning an explicit
genitive is required; in any case H is contrasting the lavishness of
the offerings, not the morality of the donors. For touching the altar
in sacrifices and other rituals cf. Plaut. rud. 1333 ‘tange aram hanc
Veneris’, Macrob. sat. 3. 2. 7, Pease on Virg. Aen. 4. 219 (with copious
illustrations).
18. non sumptuosa blandior hostia: this is the beginning of the

apodosis, not the end of the protasis: H is saying that a nominal offering
is as acceptable as a blood-sacrifice, not that no offering at all need
be made. The construction is made clearer if in translation we postpone
v. 18: ‘if a hand touches an altar without a contribution, it appeases
the offended Penates with reverent emmer and sputtering salt; it is
not made more appealing by a costly victim’. Cf. Ov. her. 5. 35 f.
‘sumptisque decentior armis / venit in arbitrium nuda Minerva tuum’,
i.e. at the Judgment of Paris the naked Minerva would have been
more seemly if she had been wearing armour; see also Lucan 5. 166
with Housman.
19–20. mollivit aversos Penatis: because no offering has been brought

(immunis), the gods are unpropitious; but they can be mollified by a
token gift ( J. P. Postgate, CQ 4, 1910: 109 ff., Sullivan, op. cit. 111);
aversos on this interpretation has a particular point, but it is used
more generally at epod. 10. 18, Prop. 4. 1. 73, Ov. trist. 1. 3. 45 ‘multaque
in aversos effudit verba Penatis’. mollibit is more widely attested than
mollivit (the common confusion in late Latin of v and b), but the archaic
future of the fourth conjugation is paralleled in Augustan poetry only at
Prop. 3. 21. 32 lenibunt; the imperfect mollibat at Ov. met. 6. 21 gives little
support, as molliebat is impossible in hexameters. mollivit is an instance
of the ‘gnomic perfect’, where something can be treated as a general rule
because it has happened often in the past (3. 2. 30 n.); there are decisive
parallels to the present passage at paneg. Mess. 14 f. ‘parvaque caelestes
placavit mica’, Stat. silv. 1. 4. 130 f. ‘sed saepe deis hos inter honores /
caespes et exiguo placuerunt farra salino’.
20. farre pio et saliente mica: far is emmer (RE 3A. 1600 ff., L. A.
Moritz, Grain-Mills and Flour in Classical Antiquity, 1958: xxii ff.), the
standard grain of the early Romans (Ov. fast. 2. 519 f., Plin. nat. hist. 18.
7), but because it had to be roasted and pounded in order to remove the
husks, it was superseded by wheat, which only needed threshing before
grinding. For pio cf. Virg. Aen. 5. 745 ‘farre pio’, [Tib.] 3. 4. 10 (quoted
below), OLD 2c (of ritual offerings). Even a small offering can fulfil
religious obligations; cf. the passages cited in the introduction above.
mola salsa was a mixture of grain and salt (Frazer, fast. vol. 4, 175 f., Pease
on Virg. Aen. 4. 517, Bömer on Ov. fast. 1. 128, K. Freudenburg, TAPA
125, 1995: 214 and nn. 24 and 25; cf. Greek Pº
ÆØ) that was sprinkled
on the victim (hence immolare) or as here on the altar. mica is a grain of
salt, but there is no implication that the ingredients are thrown separ-
ately. saliente describes the sputtering of the salt in the flame (cf. [Tib.]
3. 4. 10 ‘farre pio placant et saliente sale’, Ov. fast. 4. 409 f. ‘farre deae
micaeque licet salientis honorem / detis’); the pun on sale is probably
implicit in our passage (cf. Isid. orig. 16. 2. 3 ‘sal quidam dictum putant
quod in igne exsiliat’). The sputtering was a sign that the gift was
acceptable (cf. Virg. ecl. 8. 105 f. ‘aspice, corripuit tremulis altaria flammis
/ sponte sua . . . cinis ipse’); for divination by a so-called IºØ ÆØ ,
see A. Bouché-Leclercq, Histoire de la divination dans l’antiquité 1,
1879: 178 ff., W. R. Halliday, Greek Divination, 1913: 184 ff.
2 4 . I N TAC T I S O P V LEN T I O R
[Fraenkel 240 ff.; Lovejoy–Boas (1935, repr. 1997), especially 315 ff.; Pasquali 428 ff.;
Zs. Ritoók, Acta Classica (Debrecen) 29, 1993: 163 ff.; Treggiari, 1991: 291 ff.; G. Williams,
JRS 52, 1962: 29 ff.]
1–8. Though you build extravagant villas, you will not escape the fear and
actuality of death. 9–24. The Scythians and Getae live better; though their
homes are only caravans, their women are brought up to be chaste. 25–44.
Whoever intends to stop civil strife must pass laws to curb permissiveness; but
laws are vain when entrepreneurs do anything to make money. 45–50. If we
truly repent of the civil wars we must offer our wealth to the gods or else
dump it in the sea. 51–62. Self-indulgence must be rubbed out, and the well-
born boy must be toughened by manly pursuits; as things are he idles and
gambles, while his father commits fraud only to leave him money. 63–4.
Wealth grows remorselessly, yet something is always lacking.
2 4 . I N TAC T I S O P V LEN T I O R 271
This poem has much in common with the so-called Roman Odes
(3. 1–6). It handles national issues at considerable length, but without
any intrusion of the first person: there are none of the autobiographical
touches we find in 3. 1 and 3. 4, or the vatic proclamations that introduce
3. 1 and 3. 6. There is no addressee, though the Princeps (25 ff.) and his
programme of regeneration are the central theme. In particular the ode
contains an attack on avarice and extravagance, here represented by
luxury building (as in 3. 1); it upholds the importance of virtus (as in
3. 5), especially in the case of young men (as in 3. 2); it calls for a curb on
violations of marriage, which are seen as part of the national malaise
(as in 3. 6). But in addition to these common elements, this poem has
distinctive features in terms of both subject-matter and tone.
In the first part of the poem (1–24) Horace contrasts the ostentatious
extravagance of Rome with the superior morality of the Scythians and
Getae. The Scythians had a particular fascination for Greek and Roman
ethnographers: there was an extensive and unsentimental excursus by
Herodotus (4. 1–81), some interesting observations by Hippocrates
(de aere aquis locis 17–22), a less objective treatment by Ephorus as
cited by Strabo 7. 3. 9 (¼ FGrH 70 F42, C. van Paassen, The Classical
Tradition of Geography, 1957: 256 ff.), a lost account by Aristotle in his
—æd ø ÆæÆæØŒH, and in Horace’s own time or soon after,
discussions by Sallust (hist. 3. 61–80 M., R. Syme, Sallust, 1964: 195),
Pompeius Trogus (ap. Justin 2. 1–2), and Strabo (7. 3. 6–9); see also DNP
11. 644 ff. In addition there was much fictitious Greek lore (Rohde,
215 ff.), extending from the fantastic Arimaspea attributed to Aristeas
(Bolton, 1962) to the moralizing commonplaces of Dio Chrysostom
(especially orat. 36) and Lucian (Bompaire 228 ff.); no doubt some of
this comes from the Cynics (Lovejoy–Boas, op. cit. 117 ff.), as is shown
for instance by the imaginary letters of Anacharsis (Hercher, Epistolo-
graphi Graeci, 1873: 102 ff., Lovejoy–Boas 329 f.). As a people living at the
margins of the familiar world the Scythians were sometimes idealized
(cf. Kroll 303 f., 320 ff., Lovejoy–Boas 315 ff.), much as prehistoric soci-
eties could be misrepresented by fables about the Golden Age (cf.
Gatz 189 ff.). The Romans were particularly ready to praise the hardi-
hood of northern barbarians (cf. Tacitus’ Germania), as opposed to the
luxury and softness of the East (cf. vv. 1–2); Rousseau may have envied
the noble savage for his ‘natural’ existence, free from the tyranny of
convention, but Horace professes to admire the Scythians for their self-
discipline and social cohesion. In our own day attempts to understand
undeveloped societies on their own terms are sometimes combined with
denunciations of Western civilization; in antiquity some Cynics no
doubt talked this way, but Horace’s hyperboles should be seen as part
of the rhetorical context rather than a full reflection of his personal
views or of Augustan ideology.
Ancient ethnographical writing often ignored distinctions of time
and place, and by Horace’s day the Scythians were often confused with
the Getae and the Sarmatians; but as a poet he could select the
commonplaces that best suited his purpose. Northern peoples had
long been regarded as paragons of justice; cf. Hom. Il. 13. 5 f. IªÆıH
, +ºªH = ªºÆŒªø, #ø ØŒÆØø IŁæø with
F. Buffière, Les Mythes d’Homère et la pensée grecque, 1956: 362 ff., Bolton,
op. cit. on the Hyperboreans. The Scythians were similarly admired for
their fair dealing (Aesch. TrGF 198, Ephorus, loc. cit., ps.-Scymnus, orb.
descript. 852 f. and 859, Pomp. Trog. ap. Just. 2. 2, Curt. 7. 8. 29,
Dio Chrysost. 69. 6); the same attitude is found in our poem (16 aequali,
18 innocens); in other contexts the Scythians are associated with a
savagery appropriate to their climate (3. 10. 1 n., Herod. 4. 62–5,
Ephorus, loc. cit.), or alternatively physical degeneracy (Hippocrates,
loc. cit.) or even decadent luxury (Clearchus of Soli ap. Athen. 524c,
Strabo 7. 3. 7). Horace, with many other writers, represents them as
living in caravans (below, 10 n.), but he also sees them as practising
agriculture; Herodotus had already spoken of some as ‘the farming
Scythians’ (9 n.), and modern excavations of their settlements have
revealed a rich variety of artefacts (E. H. Minns, Scythians and Greeks,
1913, T. Talbot Rice, The Scythians, 1957, M. I. Artimonov, Treasures from
Scythian Tombs, 1969, DNP 11. 644 ff., more generally R. Rolle, The
World of the Scythians, 1989). Horace also describes them as holding their
land in common (12–13 n.); here by the process of ‘transference’ common
in ethnography he may be introducing an element associated with the
Germans, as he does in connection with annual migration and the
rotation of duties (see the notes on 12–16). According to some writers
the Scythians also shared their women and children (ps.-Scymnus loc.
cit. 857, Pomp. Trog. loc. cit., Strabo 7. 3. 7 a ªıÆEŒÆ —ºÆøØŒH

Æ ŒØa ŒÆd ŒÆ, cf. 7. 3. 4 on the Getae, and later Tert. adv.
Marc. 1. 1. 3 ‘promiscua libido’). This tradition, however, was not
universal (Lucian, Scyth. 4).
In the central section of the poem (25–44) Horace seems to come
closer than usual to a concrete political manifesto. He deplores the vices
of the times, in which he includes both avarice (largely male, 1–8, 36–50,
59–64) and sexual immorality (largely female; the licentia of 29 is
contrasted with the alleged castitas of Scythian women in 20–4, just as
the villas of 1 ff. are contrasted with the Scythian plaustra); and he
connects both aspects of Rome’s degeneration with the scelus of civil
war (26 and 50, cf. A. Wallace-Hadrill, Past and Present 95, 1982: 22 ff.).
What is more, he calls for penal legislation (28 ff., 33 f.), which in view of
the preceding stanzas must be directed against sexual immorality. The
whole passage (25–36), with its assertion that great men are appreciated
only when dead (30 ff.), seems to reflect the failure of the Princeps’s
original attempt at marriage legislation. For this important and contro-
versial issue see the introduction to 3. 6.
The final section of the poem (45–64) resumes the attack on materi-
alism. It begins with a call to the Romans to surrender their treasures to
the gods on the Capitol or else to cast them into the sea: such an address
to the citizen body recalls the imitation of early Greek poetry in the
Epodes (45 n.). The content of the exhortation is equally unrealistic,
being drawn from Greek popular moralizing, sometimes imprecisely
called diatribe; see especially 45 n. and 47 n. for similar stories about
Pythagoras and Crates the Cynic. Such writing had some influence on
Hellenistic iambographers like Phoenix (ed. G. A. Gerhard, 1909) and
Cercidas (Pasquali 210 ff.), and Horace frequently adapts it to the
criticism of his own society, not only in the Satires but sometimes also
in the Odes (see N–H vol. 2, index s. v. diatribe, Kroll 243 and WS 37,
1915: 223 ff.). On the other hand, when he turns to the training of the
young, his comments about horse-riding and playing with hoops are so
specifically Roman that they seem to reflect conservative opinion, per-
haps even some observations of the Princeps himself. One thinks of the
frequent performances of the Lusus Troiae which were staged at his
command (Suet. Aug. 43. 2), the exercitatio campestris, and the various
privileges that were granted to the well-born youth (see the introduc-
tion to 3. 2 and Z. Yavetz in Millar–Segal, 1984: 16 ff. with his refer-
ences).
The ode has not been greatly admired in modern times, though
it inspired Du Bellay’s ‘Conte des Avaritieux’ (ed. Charmand III,
108–19). It lacks the imaginative range of the best Roman Odes
(where the symbols of Greek myth and early Roman history are brought
to bear powerfully on present issues), and the Asclepiad metre with its
frequent end-stopped couplets cannot rival the sweep of Alcaics. It is
best seen as a versified diatribe in the manner of the Greek iambogra-
phers and Horace’s own 2. 15 and 2. 18. The clipped and bleak querimo-
niae suggest affinities with Sallust, whose influence is also apparent in
the Epodes (see G. Schörner, Sallust und Horaz über den Sittenverfall und
die sittliche Erneuerung Roms, Diss. Erlangen, 1934: 74 ff.); and the
somewhat prosaic vocabulary suits down-to-earth preaching: cf. 3 cae-
mentis, 10 plaustra, 16 vicarius, 18 privignus, 19 dotata, 28 subscribi, 30
quatenus, 57 trocho, 60 consortem, 61 pecuniam, 64 curtae. These features
are combined here and there with a certain banality (as in 51–8) which
recalls the weaker stanzas of the Carmen Saeculare (in particular 17–20
and 45–8). In his public poems Horace is at his best when he treats
political matters more indirectly; cf. N–H vol. 1, p. xix, Lyne (1995),
49 ff., 201.

1–2. Intactis opulentior / thesauris Arabum: H’s compression is
notable; conventional Latin prose would have used another concessive
clause (licet sis opulentior). Brevity is also helped by the compendious
comparison, where symmetry would have required ‘richer than the
Arabs who enjoy untouched treasure-chambers’ (cf. 2. 6. 14 f. ‘ubi non
Hymetto / mella decedunt’ with N–H’s note). The accumulation of four
words suggesting wealth increases the impression of opulence; for a
similar effect cf. 1. 17. 14 ff. thesauris evokes the proverbial riches of
Arabia Felix (in the south-west corner) soon to be made topical by
Aelius Gallus’ expedition about 26 bc; cf. 1. 29. 1 f. ‘Icci, beatis nunc
Arabum invides / gazis . . . ?’ with N–H, 2. 12. 24, S. Jameson, JRS 58,
1968: 71 ff., G. Bowersock, Roman Arabia, 1983: 46 ff. The strong-rooms
are inviolate, not just because they were locked and sealed (as in the tale
of Aladdin), but because Arabia had hitherto been immune from inva-
sion; cf. 1. 29. 3 f. ‘non ante devictis Sabaeae / regibus’ (with N–H), Prop.
2. 10. 15 f. ‘India quin, Auguste, tuo dat colla triumpho / et domus
intactae te tremit Arabiae’.
et divitis Indiae: the wealth of India was also proverbial, cf. epist.
1. 6. 6, Prop. 3. 4. 1 ‘arma deus Caesar dites meditatur ad Indos’, Otto
174. Herodotus already says
æıe ¼º ÆP ŁØ K (3. 106. 2); as he
wistfully points out, the finest products seem to come from the most
distant countries. The definition of India was imprecise, and could
sometimes include the Persian and Arabian Gulfs; cf. A. Dihle, PCPS
10, 1964: 15 ff., Murgatroyd on Tib. 2. 2. 15 f. For western trade with
India see E. H. Warmington, The Commerce between the Roman Empire
and India, 1974; M. G. Raschke, ANRW II. 9. 2. 604 ff.; V. Begley and
R. de Puma (edd.), Rome and India, 1992; OCD 754.
3. caementis licet occupes: for the building of villas over the sea see
3. 1. 33 n. For caementa (broken stone) see 3. 1. 35 n.; other prosaic words
in this austere poem are mentioned in the introduction above. occupare is
usual in legal contexts for ‘staking a claim’; it is hubristic to do this in
the sea.
4. Tyrrhenum omne tuis et mare Punicum: we have accepted the

thinly attested Punicum as an old equivalent of mare Libycum; cf. Flor.
1. 18. 17 ‘nec defuerant qui ipso Punici maris nomine ac terrore deficer-
ent’. Horace is hyperbolically suggesting that the rich man’s palace
might extend across the Mediterranean (note omne); for similar fantasies
cf. 2. 2. 10 ff. ‘si Libyam remotis / Gadibus iungas et uterque Poenus /
serviat uni’ with N–H.
The MSS for the most part read Apulicum or Ponticum, while
ps.-Acro’s comment (‘omnibus patens’) and the vetus Blandinianus sup-
port the variant publicum. Though the u of Apulia may be either long or
short, the first vowel is never short (Housman, Classical Papers, 1. 99 f.);
for at 3. 4. 10 limen Apuliae makes no sense (see note ad loc.). It is true
that the Apulian sea (to the east of Italy) is often combined with the
Tyrrhenian to the west (see RE 14. 2. 1673 ff. for mare superum and mare
inferum), but it was not a place for luxury building. This consideration
also rules out Ponticum, as the Black Sea was far too remote.
publicum might seem to allude to the common belief and legal
doctrine that the sea is common to all; cf. Soph. TrGF 4. 673.
ºŒØ #ØæÆ, Plaut. rud. 975 ‘mare quidem commune certost
omnibus’ with Marx, Ov. met. 6. 349 ‘usus communis aquarum est’ with
Bömer, dig. 43. 8. 3 ‘maris communem usum omnibus hominibus ut
aeris, iactasque in id pilas eius esse qui iecerit’ (in our passage the legal
implication would suit occupet). But then balance is only achieved by
reading terrenum (Lachmann) for Tyrrhenum; for the same confusion cf.
Lucan 6. 401. Porph. ad loc. comments ‘aedificiis novis non terram
tantum verum etiam mare occupantem’, but this is a possible comment
even if H made no mention of the land. Heinze explains terrenum as soil
suitable for agriculture (Liv. 23. 19. 14 ‘quidquid herbidi terreni extra
murum est’, Colum. 3. 11. 8), but this seems too restrictive to balance the
comprehensive mare. Ritoók, op. cit. cites Lucan 10. 204 ‘luna suis
vicibus Tethyn terrenaque miscet’, but the pl. of the adj. (¼ terrestria)
is easier in poetry at least than terrenum as a noun. If Punicum is
accepted, the four place names (two eastern, two western) provide a
typically Horatian pattern (so 1. 31. 4 ff., 3. 1. 42 ff.); and one may agree
with Gow that it is absurd to alter the transmitted and characteristically
specific Tyrrhenum in order to support the variant publicum.
Palmer proposed ‘caementis licet occupes / Tyrrhenum omne tuis et
mare sublicis’ (CR 5, 1891: 140 f.); and this is read by Shackleton Bailey.
Here et is intended to join caementis and sublicis (wooden piles); but in
that case the conjunction would be followed immediately by sublicis
rather than by mare (as explained in A. Y. Campbell’s 2nd edn. 116).
5 –7. si figit adamantinos / summis verticibus dira Necessitas / clavos:

‘if indeed dread Necessity fixes adamantine bolts on the rooftops’. The
picture has veered from the literal to the metaphorical: death is the
culmination of all our strivings (cf. 2. 18. 29 ff., also in terms of a domus).
So in Greek ŁæØªŒ or ‘coping-stone’ is used for the final stage or
climax (Aesch. Ag. 1283 ŒØØ ¼Æ ŁæØªŒø ºØ , Eur. Her.
1280 HÆ ŁæØªŒHÆØ ŒÆŒE ).
Necessitas is the Gk. #ªŒ (Roscher 3. 1. 70 ff.), the inescapable
Doom of Death; cf. 1. 35. 17 ff. ‘te semper anteit saeva Necessitas / clavos
trabalis et cuneos manu / gestans aena’ (aenos Campbell) with N–H ad
loc., 3. 1. 14 n. For the association of Necessity with words for fixing
( figo, destino) and binding (ø, necto) see Onians 322 f., 363. For clavi
(sometimes half a yard long) as emblems of fixity see N–H on 1. 35. 18;
the word derives emphasis from its position at the end of the clause and
the beginning of the line. In Etruscan funerary art Charun is apparently
equipped with a hammer (Latte 156), and Arthrpa (¼ Atropos) appears
on the famous mirror from Perugia with a hammer and a nail, the latter
pointing at the head of Adonis (LIMC 3. 1. 1 f.). It may be relevant that
nails were fixed annually in Nortia’s temple at Volsinii, as also in
Jupiter’s Capitoline temple; Liv. 7. 3. 7–8 regards these ceremonies as a
tally of the years, but they may rather have fixed the fates for the coming
year, or declared the immutability of what had already happened
(RE 17. 1. 1049). Adamant was a legendary metal of exceptional hard-
ness (1. 6. 13), mentioned first in Hesiod (West on Hes. theog. 161,
[Aesch.] Prom. 6 etc., RE 5. 322 ff.); it was associated with Necessity
(orac. ap. Herod. 7. 141. 3, Pind. P. 4. 71 f. b Œı ŒæÆæE
IÆ = B –ºØ ; Plat. rep. 10. 616c) and with death (Gow on
Theoc. 2. 33 f., Prop. 4. 11. 4 ‘non exorato stant adamante viae’).
This sentence poses three distinct problems. (a) If si is genuine, it
must mean, as often, si quidem, ‘if indeed’, not indicating scepticism but
introducing an agreed truth (as in 3. 1. 41). But even so it is less assertive
than translations like ‘whereas’ and ‘seeing that’. Here it seemed too
ambiguous to Bentley, who commented ‘quid si non figat clavos Neces-
sitas, nonne morietur nihilo minus?’; he therefore proposed sic, putting
the clause in a parenthesis: ‘(such is the decree of Fate) you will not
escape death’. For sic in similar contexts cf. e.g. 2. 17. 15 f. ‘sic potenti /
Iustitiae placitumque Parcis’, epod. 9. 3 ‘tecum sub alta (sic Iovi gratum)
domo’, serm. 2. 6. 22 (‘sic dis placitum’), Virg. Aen. 3. 375 ‘sic fata deum
rex / sortitur’, 4. 614 ‘sic fata Iovis poscunt, hic terminus haeret’. So we
see considerable attraction in the decisiveness of sic rather than the
circumspection of si. Against this one must weigh the fact that such
parentheses usually refer back, and that here a certain degree of compli-
cation is introduced.
(b) The prosody of ‘si figit adamantinos’ is unusual though not
certainly wrong. At 1. 3. 36 ‘perrupit Acheronta Herculeus labor’ a
long vowel in the perfect -it can be paralleled in comedy and by inscrip-
tions in -eit (N–H ad loc.); elsewhere H lengthens the final vowel of
verbs in the present tense ending in -et (particularly before a main
caesura); this can again be justified as an archaic usage (3. 5. 17 n.). But
present figı́t cannot be similarly defended; the instances cited by Lindsay
(on Plaut. capt. p. 13) and Skutsch (on Enn. ann. p. 58) are of the fourth
conjugation or otherwise different. Also different are passages where
short syllables are lengthened before the principal caesura of a hexam-
eter (serm. 1. 4. 82 ‘defendit alio’, 2. 3. 260 ‘agit ubi’). In our passage Axt,
cited by Orelli, proposed ‘si summis adamantinos / figit verticibus’,
a conjecture worth recording in an apparatus criticus; summis would
gain emphasis from the hyperbaton and provide a contrast with the
rubble of the foundations.
(c) summis verticibus would have to refer to rooftops, as in Mart. 8. 36.
11 f. ‘haec, Auguste, tamen quae vertice sidera pulsat / par domus est
caelo’. In the present context it is less easy to understand villarum, but as
caementis refers to the foundations, that may make the point clearer.
Because of the difficulty RN has considered imbricibus, i. e. the topmost
roof-tiles (PCPS suppl. 15, 1989: 91). For the fixing of tegulae by bolts cf.
the lex Puteolana, CIL 1. 2. 698 (¼ ROL 4, p. 276) ‘tegulas primores
omnes . . . ferro figito’, G. Brodribb, Roman Brick and Tile, Gloucester
1987: 10 ff.; this was not the norm with imbrices (Prof. Trevor Hodge
informs us), though for some evidence see Brodribb 26. But if imbricibus
is read, there is no point in transposing figit, as the prosody remains
difficult.
summis verticibus is explained by some as literally the plutocrat’s own
head (cf. Theog. 1012 of old age ŒÆºB –ÆØ IŒæ , Rhianus,
Coll. Alex. 1. 18 (Powell), ¼Œæfi K ŒÆºfi Ø);
Ð but the plural verticibus
and the absence of tuis make this hard to accept, and caput (8) is
confusingly anticipated. Whatever view one takes, the villas must be
in some way associated with the life of the builder; just as the nails
driven into the rooftops show that the houses are finished, so the nails of
Necessity will inevitably finish the house of the rich man’s life. Bentley
took summis verticibus as meaning that the nails were driven up to their
heads; but then the construction is impossibly awkward, and we are not
told where the nails are fixed.
7–8. non animum metu, / non mortis laqueis expedies caput: animum
balances caput, and mortis should be supplied with metu. By Epicurean
doctrine avarice and ostentation do nothing to assuage the fear of death,
but manifest a vain attempt to run away from it (3. 1. 17 ff., Lucr. 3. 63 f.
‘haec vulnera vitae / non minimam partem mortis formidine aluntur’).
A laqueus was a noose used to catch birds or animals (epod. 2. 35, Virg.
georg. 1. 139); for H’s metaphor cf. Apul. met. 10. 24. 2 ‘crudelissimis
laqueis mortis insidiari’; for comparable metaphors see Aesch. Ag. 1115,
Stat. silv. 5. 1. 155 ff. As expedire originally meant ‘to disentangle one’s
foot’, there may be a verbal point in the juxtaposition with caput (for the
contrast cf. Otto 74); so perhaps 1. 4. 9 ‘caput impedire myrto’. The lines
entail the assumption that, unlike the builder in 2. 18. 18 f. who is sepulcri
immemor, this man is well aware of the finality of death.
9. campestres melius Scythae: it is a paradox that even a proverbially

savage people (Strabo 7. 3. 9, Mela 2. 1. 11–15) has a better way of life
than over-civilized Romans; for unrealistic admiration of northern
tribes see the introduction. The steppes of Russia are mentioned also
at 2. 20. 16 ‘Hyperboreosque campos’, 3. 8. 23 f. ‘Scythae . . . meditan-
tur . . . cedere campis’, cf. Herod. 4. 47. 1 l ªaæ ªB KFÆ Øa Æo
Ø ŒÆd hıæ KØ. Herodotus contrasts the Scythian nomads
(the Greek word simply means ‘pastoral’) with the ªøæª or tillers of
the soil (4. 18. 1). The prosaic campestres does not itself represent
but rather a geographical term like ØØØ (cf. Liv. 39. 53. 13); here the
emphatic word suggests wide open spaces (cf. Otto 315 for Scytharum
solitudines) that mark a contrast with the overcrowding and cluttered
buildings of the civilized world.
10. quorum plaustra vagas rite trahunt domos: plaustra, the word for
ordinary farm-carts, describes the nomads’ unglamorous caravans. The
best account of these vehicles is in Hippocrates, de aere aquis locis 18: Æƒ
b ±ÆÆ NØ Æƒ b Kº
ØÆØ æŒıŒºØ Æƒ b ŒıŒºØ . . . Nd b
ŒÆd
ÆÆØ uæ NŒÆÆ, a b ØºA a b æØºA . . . a b
±Æ (ºŒıØ ªÆ a b a b æÆ H . . . K Æfi Ø b
s fi Ø
Ð ±fi Ø Æƒ ªıÆEŒ ØÆØFÆØ, ÆPd K ¥ ø O
FÆØ
ƒ ¼æ ; see further Hes. fr. 151 M–W ˆºÆŒªø K ªÆEÆ IÆ
NŒ K
ø, Pind. fr. 105. 4 ff. Ø ªaæ K &ŒŁÆØ IºAÆØ
æÆH, = n IÆ æ r Œ P ÆÆØ, = IŒºc Æ (the
disgrace of not possessing a caravan), [Aesch.] Prom. 709 f. &ŒŁÆ
Ifi Æ Q ºŒa ªÆ = æØØ Æı K PŒŒºØ Z
Ø ,
Herod. 4. 46. 3 NŒÆ Ø fi Kd ıªø, Sall. hist. 3. 76 ‘Scythae
nomades . . . quibus plaustra sedes sunt’, Strab. 7. 3. 17 H b ˝ø
Æƒ ŒÆd ØºøÆd ªÆØ Kd ÆE ±ÆØ K Æx ØÆØHÆØ, Tac.
Germ. 46. 2 (of the Sarmatians) ‘in plaustro equoque viventibus’.
vagas with domos is a paradox, as a house should be stabilita; there is a
contrast with the firm foundations of the plutocrat’s villa (3). rite suits
ethnography (like ritus), and makes another oxymoron with vagas; for
the nomads a refusal to settle is a principle of life. In view of the
parallels one would have expected vehunt rather than trahunt. We
might even consider ‘quorum rite vagas plaustra vehunt domos’; that
would give more point to the combination of rite and vagas, and the
final -tra of plaustra could have helped to produce trahunt.
11. vivunt et rigidi Getae: the Getae were a people of the lower Danube
(cf. 4. 15. 22), who were later confused with the Goths (see H. Wolfram,
History of the Goths, 1988: 19). rigidi refers to the severity of their
discipline (epist. 1. 1. 17 ‘virtutis verae custos rigidusque satelles’, 2. 1. 25
‘rigidis . . . Sabinis’) and the grimness of their demeanour (Ov. trist.
5. 1. 46 ‘barbariam rigidos effugiamque Getas’). Porph. comments
‘propter frigida regionum earum’ (cf. Sen. dial. 4. 15. 5), and there is
certainly a hint of the cold climate (cf. 4. 5. 25 ‘gelidum Scythen’, Virg.
georg. 3. 354 ff.), but when the adjective is applied to people, this nuance
cannot be primary. vivunt should be taken in the first instance with
Getae (cf. 38 n.), but also with Scythae.
12–13. immetata quibus iugera liberas / fruges et Cererem ferunt:

Orelli refers this clause only to the Getae, to balance the clause about
the Scythians in line 10. But the Scythians cannot be ignored; because of
their life-style, they held their land in common. The same sort of system
was used by the Suebi (Caes. bG 4. 1. 7) and by the Germans in general
(6. 22. 2). Such collectivism recalled the Golden Age; cf. Virg. georg.
1. 126 f. ‘ne signare quidem aut partiri limite campum / fas erat’, Tib.
1. 3. 43 f. with Murgatroyd, Ov. met. 1. 135 with Bömer, orac. Sib. 2.
319 ff., Gatz 229 (absentia rerum privatarum). H’s account of Getic land-
tenure may not be in accordance with the evidence (B. Gerov, Eirene
14, 1976: 31 ff.).
immetata (lit. ‘unmeasured’) occurs only here. Like mensor in Ov. met.
1. 135 f. (‘communemque prius, ceu lumina solis et auras, / cautus
humum longo signavit limite mensor’) it suggests the Roman agrimen-
sores, and so makes an oxymoron with iugera, a word notably inappropri-
ate to the open country of the nomads. (For a similar point cf. 2. 15. 14 ff.
‘nulla decempedis / metata privatis . . . / porticus’.) fruges et Cererem,
which Kiessling wanted to distinguish, is a hendiadys for ‘crops of
grain’, the general being followed by the particular. liberas means that
nobody claimed private rights.
14. nec cultura placet longior annua: for Scythian agriculture see 9 n.
For such annual migrations cf. Caes. bG 4. 1. 7 (about the Suebi) ‘neque
longius anno remanere uno in loco colendi causa licet’, 6. 22. 2 ‘anno
post alio transire cogunt’, Tac. Germ. 26. 2 (about the Germans) ‘arva
per annos mutant, et superest ager’.
15–16. defunctumque laboribus / aequali recreat sorte vicarius: Caes.

bG 4. 1. 4 describes how, when some of the Suebi went to war, ‘reliqui
qui domi manserunt se atque illos alunt: hi rursus in vicem anno post in
armis sunt, illi domi remanent’. Caesar may have influenced Horace, as
he combines the same three points: the holding of land in common, the
annual migration, and the simultaneous rotation of duties. But in
Caesar work on the land alternates with military service, whereas in
H the weary labourer seems to get a period off.
recreat implies ‘a new lease of life’, whereas ‘relieves’ in English has
lost most of its original force. aequali sorte means ‘by an impartial
allotment of duty’ (cf. Sil. 9. 92 f. ‘dum sorte vicissim / alternat portae
excubias’). The emphatic aequali underlines the fairness of the lot;
Roman land-tenure could be notorious for its injustice. A vicarius
(RE 8A. 2. 2015 ff.) was a proxy in many spheres of Roman life
(cf. serm. 2. 7. 79 f. ‘sive vicarius est qui servo paret’); a vicar in the
Church was originally a deputy. Here, as there is no subordination,
vicarius is best translated by ‘successor’; cf Cic. Verr. 4. 81 ‘succedam ego
vicarius tuo muneri’.
17–18. illic matre carentibus / privignis mulier temperat innocens:

the stepmother, unlike her Roman counterpart, does no injury to
her motherless stepchildren (innocens has the strong sense of innoxia):
contrast Eur. Alc. 305 ff. ŒÆd c Øªfi E æıØa ŒØ , = lØ
ŒÆŒø s KF ªıc Ł fiø = E EØ ŒIE ÆØd
EæÆ æÆºE.
The stepmother was proverbially unkind, cf. epod. 5. 9 ‘quid ut noverca
me intueris?’, Hes. op. 825 ¼ºº æıØc ºØ æ, ¼ºº æ
with West, Juv. 6. 627 with Courtney, Otto 245 f., S. Dixon, The Roman
Mother, 1988: 155 ff. The stereotype was not invariable; cf. Prop. 4. 11. 87 f.
(the dead Cornelia speaks) ‘coniugium pueri laudate et ferte paternum: /
capta dabit vestris moribus illa manus’. See further Treggiari, index s.v.
‘stepmothers’, P. A. Watson, Stepmothers: Myth, Misogyny and Reality,
Mnem. suppl. 143, 1995.
19–20. nec dotata regit virum / coniunx: a Roman bride of any standing
brought a dowry to her husband, but if the marriage broke down she
could usually claim it back, cf. Crook 104 ff., Treggiari 323 ff. So a woman
in this position (dotata is emphatic) might dominate her husband in a
way that to men at least seemed unnatural; for regit, ‘tyrannizes over’, cf.
regnare at Juv. 6. 149. For the stock situation see Eur. Phaethon 158 f.
KºŁæ J Fº KØ F º
ı , = æÆ e HÆ B æB

ø (with Diggle’s note), Arist. eth. Nic. 1161a1 f. K b ¼æ
ıØ Æƒ
ªıÆEŒ KŒºæØ sÆØ, Plaut. asin. 87 ‘argentum accepi, dote imper-
ium vendidi’, aul. 534 f. ‘nam quae indotata est, ea in potestate est viri: /
dotatae mactant et malo et damno viros’, Men. 766 f., Titin. CRF 70, Juv.
6. 136 ff. (with Courtney’s note), F. Wilhelm, RhM 70, 1915: 177. The
absence of a dowry may be part of the ethnographic tradition; cf. Tac.
Germ. 18. 1 ‘dotem non uxor marito sed uxori maritus offert’. By some
accounts some northern tribes gave women a privileged position; cf.
Herod. 4. 26. 3 of the Issedones NŒæÆ b ›ø Æƒ ªıÆEŒ EØ
IæØ, Bolton 79 (with a quotation from Afghanistan), E. Hall,
Inventing the Barbarian, 1989: 202 ff. (noting fictions about Amazons).
20. nec nitido fidit adultero: cf. Tacitus’ tribute to the Germans ‘pau-
cissima in tam numerosa gente adulteria’ (Germ. 19. 1). Here nitido
describes a smart but ungentlemanly sleekness (epist. 1. 7. 83 of the
auctioneer Vulteius Mena, Juv. 3. 157 of an auctioneer’s son), and has a
hint of hair-oil (1. 4. 9 with N–H, 3. 19. 25); barbarians, on the other
hand, were shaggy and unkempt (horridi). Shackleton Bailey objected to
fidit: ‘do we think of Lesbia and her like as trusting their lovers?’ (Philol.
184, 1990: 225). He therefore proposed and printed laedit, a word used of
disloyalty to a partner (TLL 2. 868. 64 ff.). Yet it does not seem unrea-
sonable to translate ‘nor does she place her trust in some sleek adulterer
(rather than in her husband)’.
21–2. dos est magna parentium / virtus: ‘their big dowry is their
parents’ worth’ (picking up dotata in v. 19). For the inheritance of
moral qualities cf. 3. 10. 11–12 n. For the commonplace cf. Hippothoon
6, TGF p. 828 Æo ªaæ æd NŒÆ fiø Ø , Soph. TrGF 4.
201d, Plaut. Amph. 839 ff. ‘non ego illam mi dotem duco esse quae dos
dicitur, / sed pudicitiam et pudorem et sedatum cupidinem’, Antip.
Thess. anth. Pal. 9. 96. 5 f. (¼ Gow–Page, Garland of Philip, 195 f.) j
¥ Œfi N ÆØ, #
ÆØ. XŁÆ æe =
æa ºÆ, Ø
æEŒÆ ÆØ, Plut. apophthegmata Lac. 242b —ÆæŁ Ø
æa
KæøŁEÆ Æ øØ fiH ªÆFØ æEŒÆ, , c æØ
, øæ . For ‘my face is my fortune’ cf. Afran. CRF 156 ‘formosa
virgo est: dotis dimidium vocant . . . ’, Otto 121; for a cynical application
add Ov. ars 2. 155 ‘dos est uxoria lites’. parens was by origin a present
participle with genitive plural in -ium, but when it was treated as a noun
the form in -um was normally used. H has both forms (as ps.-Acro
notes), -um at 1. 2. 23, 2. 20. 6, 3. 6. 46; -ium only here. As with other
genitive plurals, the longer form does not suit hexameters, but is
sometimes convenient in lyrics; see further Bo 224 f.
22–3. et metuens alterius viri / certo foedere castitas: certo foedere, a

common collocation, is ablative of quality with castitas. For the very
Roman idea of the foedus of marriage see Catull. 64.373 (the first attested
use, but not necessarily an innovation), Virg. Aen. 4. 339, R. Reitzen-
stein, SHAW 1912, 12. Abh., 9 ff. (on the word’s extension to less formal
relationships), TLL 6. 1. 1004 f., Lyne (1980), 34 ff. metuens suggests
timidity rather than terror (cf. 3. 11. I0 ‘metuitque tangi’); for the gen. cf.
3. 19. 16 ‘rixarum metuens’ (shunning brawls). ‘alterius viri’ means
‘another man’ (the gen. of alius is very rare), as at Prop. 1. 16. 33 ‘nunc
iacet alterius felici nixa lacerto’; distinguish alter of a specific ‘other man’
(Prop. 2. 23. 3).
24. et peccare nefas—aut pretium est mori: peccare, ‘to err’, originally
meant ‘to stumble’ (a sense played on by H in epist. 1. 13. 4); it is used
euphemistically of sexual lapses (1. 33. 9, 3. 7. 19). Here its relative
mildness, reflecting Roman mores, makes a pointed contrast with
nefas, ‘an abomination’. et joins ‘their dowry is morality’ with this
obverse formulation of the same point; it should be preceded by a
semicolon after castitas, for there is a climax here.
aut is elliptical and implies ‘or else’; cf 3. 12. 2 n. pretium is used, as
often, for a penalty (Ov. met. 10. 572 ‘mors pretium tardis’, Paul, Romans
6: 23 a ªaæ OłØÆ B ±ÆæÆ ŁÆ , Sen. HO 1336 ‘mors erit
pretium omnium’). For est mori some editors accept the thinly attested
emori, which can have various nuances (‘to fade away’, ‘to die once and
for all’, etc.); but in describing the objective penalty the simple verb is
much better. For barbarian punishments of women’s adultery see Tac.
Germ. 19. 2.
25–6. o quisquis volet impias / caedis et rabiem tollere civicam: any-

body who is going to eradicate civil strife must curb ‘permissiveness’.
H connects the civil wars with sexual licence (3. 6. 17 ff.) or materialism
(below, 47 ff.) or the decline of religion (3. 6. 1 ff.), but he naturally says
nothing about great men’s lust for power. Here with quisquis he is
obviously thinking of the Princeps, but he makes his point discreetly,
in general terms. In the same way the future volet avoids the suggestion
that the issue is a current one (see the introduction). impias points to the
sinfulness of civil as opposed to foreign wars; cf. 1. 2. 29 f. ‘cui dabit
partis scelus expiandi / Iuppiter?’, 2. 1. 30 ‘impia proelia’. rabiem is a
strong word that suggests animal savagery (epist. 2. 2. 75, ‘hac rabiosa
fugit canis’, ars 393 ‘lenire tigres rabidosque leones’); so epod. 7. 11 ff.
(of the civil wars) ‘neque hic lupis mos nec fuit leonibus . . . ’. civicam
is emphatic, chiastically balancing impias; here it suggests civil war (cf.
2. 1. 1 ‘motum . . . civicum’) as more commonly civilis does (cf. 4. 15. 17 f.
‘Furor/ civilis’). civicus has an archaic ring (Porph. on 2. 1. 1 says ‘antiqua
figura’ and compares hosticus).
Bentley read ‘o quis quis volet’, with a question-mark after civicam; he
is supported by the preponderance of one side of the tradition (though
in a matter of word-division MS testimony has no weight), and appar-
ently by Diomedes, GL 1. 329. 36 ff. He argued that H cannot say
‘everybody who’, as if many people were in a position to reform the
state; but here quisquis means ‘anybody who’. He also maintained that
o suits the excitement of a question, shown also by the repeated quis
(cf. epod. 7. 1. ‘quo quo scelesti ruitis?’), but it can equally well lead up to
the wish audeat (28). For this use of o (¼ o si or utinam) cf. 1. 35. 38 f.,
3. 10. 13 ff. ‘o . . . supplicibus tuis / parcas’ (with intervening clauses as
here), serm. 2. 6. 8, Virg. ecl. 2. 28 ‘o tantum libeat’, Tib. 1. 1. 51 ‘o
quantum est auri pereat’, Fraenkel 242 n., TLL 9. 2. 6. 71 ff. Bentley’s
punctuation makes it hard to see a connection between line 26 and what
follows; it also impairs the argument, for H is not saying ‘who will put
an end to civil strife?’, a question irrelevant to the poem (contrast 1. 2.
29 f. cited above); indeed the repeated quis might suggest that the
saviour is hard to identify.
27–8. si quaeret pater urbium / subscribi statuis: urbium should be
taken only with pater, not with statuis (Kiessling) nor even with both
nouns (Heinze). H is grandly describing the municipia of Italy (4. 4. 42),
which were more usually called oppida; the great cities of the East (as at
1. 35. 10) would not be concerned with a programme of moral regener-
ation. For the use of pater cf. 1. 2. 50 with N–H; Augustus was unoffi-
cially pater patriae (Dio 55. 10. 10) long before he was voted the title in 2
bc. For such honours outside Rome cf. Cic. 1. 1. 31 ‘parentem Asiae’ (of
Q. Cicero), CIL 11. 720 (Bononia) ‘divus Aug. parens dedit’, 3. 13264
(Iadera in Illyricum) ‘imp. Caes. divi f. Augustus parens coloniae’, inscr.
numm. Cohen I p. 176 n. 9 (Gades) ‘municipii parens M. Agrippa’.
Porph. comments on our passage ‘ut . . . in titulo earum pater urbium
ascribatur’; but in fact each inscription would name only the city
concerned (pater urbis).
Under the Republic statues were a recognized form of honour in Italy
as well as in Rome, and from Caesar’s dictatorship they were used, as in
Hellenistic kingdoms, to assert the authority of the ruler and the
goodwill of those who erected them. They were more conspicuous
than coinage, and more accessible to most people than literature.
Already at the beginning of 43 Octavian was voted an equestrian statue,
and the practice spread through Italy (as our passage indicates) and the
whole empire. See further G. Lahusen, Untersuchungen zur Ehrenstatue
in Rom, 1983, W. Eck in Millar–Segal (1984), 142 ff., Zanker (1987), 46 ff.
¼ (1988), 37 ff., and for the typology of Augustus’ portraits U. Haus-
mann, ANRW II. 12. 2 (1981), 513 ff. subscribere is the proper term for
putting an inscription on the base of a statue (Cic. Cluent. 101 etc.); its
prosaic tone suits the political context (so too the use of statuis rather
than the more grandiloquent signis). On the other hand, the predicative
use of pater is an artificial Horatian brachylogy; cf. 1. 6. 1 f. ‘scriberis
Vario fortis et hostium / victor’ with N–H.
A. Y. Campbell (edn. 2) argued that quaeret (the reading of the MSS
and Porph.) suggests too blatant ambition; he noted also the repetition
at 32 quaerimus (in a different sense). He therefore suggested gaudet,
comparing 1. 2. 50 ‘hic ames dici pater’; but the corruption is unlikely,
and the present indicative is out of keeping with volet and audeat. If
any change is needed (which is doubtful), RN has considered curet
(a subjunctive here would not be impossible); cf. Virg. georg. 1. 503 f.
‘iam pridem nobis caeli te regia, Caesar, / invidet atque hominum
queritur curare triumphos’. For the Princeps’s modesty, or caution, in
this respect cf. res gestae 24. 2 (he removes 80 statues of himself ).
28–9. indomitam audeat / refrenare licentiam: the need for moral

reform had previously been canvassed in Caesar’s dictatorship (Cic.
Marc. 23 ‘comprimendae libidines, propaganda suboles, omnia quae
dilapsa iam diffluxerunt severis legibus vincienda sunt’). audeat under-
lines the risks involved. refrenare, ‘to curb’, is used by Lucretius and
Ovid, but is more prosaic than the simple frenare; indomitam suits the
metaphor, as it suggests the taming of wild animals (cf. v. 26 rabiem).
The lex Julia de maritandis ordinibus was not passed until 18 bc and the
lex de adulteriis coercendis about the same time; at carm. saec. 17 ff.
H. prays for the success of this legislation, and later implausibly claims
it has succeeded (4. 5. 21, 4. 15. 9 f.). Such assertions were a commonplace
of panegyric (Men. Rhet. 376.4 ff.).
30. carus postgenitis: clarus is the reading of the MSS and is supported
by ps.-Acro’s comment ‘posteris admirandus’. carus is a variant without
authority, commended by Lambinus but generally rejected. Neverthe-
less, it provides a contrast with odimus (31) and is strikingly supported by
epist. 2. 1. 14 ‘exstinctus amabitur idem’, from the parallel passage on the
envy directed at Augustus (see below on 31 f.); for carus of political
popularity cf. Cic. off. 3. 80 ‘nemo umquam multitudini fuit carior’,
TLL 3. 504. 36 ff. It is paradoxical that the Princeps should be loved
largely by those who could not know him, but necessary legislation
sometimes incurs unpopularity at the time; on the other hand, after
Actium even enemies would concede that he was now clarus. postgenitis
is found only here; it seems to be a Horatian coinage for a word like
KØª Ø .
quatenus—heu nefas—: quatenus ‘insofar as’ in the sense of ‘since’ is
again a prosaic word; it is used three times in the sermones and four
times by Lucretius, but not by Virgil or in Silver Age epic (apart from
Sil. 17. 373). nefas (cf. 24) suggests an unspeakable outrage; the word was
connected by the Romans with fari (Maltby 407). For the indignant
parenthesis cf. 4. 6. 17 ‘heu nefas, heu’, epod. 16. 14 ‘nefas videre’, Catull.
68. 89, Virg. Aen. 7. 73, 8. 688 ‘sequiturque—nefas—Aegyptia coniunx’,
10. 673; cf. clauses beginning with pro (3. 5. 7) and indignum (epist. 1. 6. 22,
McKeown on Ov. am. 1. 6. 1); see further Horsfall on Virg. Aen. 7. 64.
Such parentheses are best marked by dashes rather than brackets, which
tend to obscure the heavy emphasis.
31–2. virtutem incolumem odimus, / sublatam ex oculis quaerimus

invidi: it was a commonplace that one resented the successful in their
lifetime and only appreciated them when they were dead. See 2. 20. 4
with N–H, epist. 2. 1. 12 ‘comperit invidiam supremo fine domari’ with
Brink, Pind. pae. 2. 55 X Ł Y
ÆØ = H ºÆØ æŁÆ ø,
Mimnermus (?), TGF p. 829 Ød ªaæ Iæd Kb PŒºE =
HØ ŁBÆØ, ŒÆŁÆ Æ ÆNÆØ, Arist. rhet. 1388a (ØºØEÆØ) . . .
æe f Kı j ŁHÆ P , ps.-Sall. epist. 2. 13. 7 ‘nam vivos
interdum fortuna, saepe invidia fatigat; ubi anima naturae cessit,
demptis optrectatoribus, ipsa se virtus magis magisque extollit’, Ov.
am. 1. 15. 39 ‘pascitur in vivis Livor, post fata quiescit (with McKeown),
Vell. 2. 92. 5 ‘praesentia invidia, praeterita veneratione prosequimur’,
R. Häussler, Tacitus und das historische Bewusstsein, 1965: 233. Plutarch,
on the other hand, thought that we do not envy truly great men like
Alexander (de invidia 538a). For a typical quip cf. Mart. 5. 10. 12: ‘si post
fata venit gloria, non propero’.
odimus implies a public rejection (‘we spurn’) as much as an internal
emotion (3. 1. 1 n.). incolumem, ‘as yet unharmed’, means in effect
viventium, but brings out more sharply the malice of human nature.
quaerimus implies a vain longing (OLD s.v. 2a), like requirimus or
desideramus.
invidi (the reading of Porph. as well as the MSS) has to be taken
with both clauses and translated in some such way as ‘jealous creatures
that we are’. The second sentence poses something of a problem in
view of the widespread commonplace (illustrated above) that the dead
are no longer envied. One approach (favoured by RN) is to look for
a convincing emendation. Crusius and O. de Rooy proposed invidis;
but the stock phrase ex oculis ‘out of sight’ (OLD s.v. 4) does not lead
us to expect an adjective. Cornelissen (Mnem. 16, 1888: 310) proposed
irriti (which could be spelt inriti); for the application to people
(¼ ‘disappointed’, ‘to no purpose’) cf. Tib. 2. 3. 22 ‘venit et a templis
irrita turba’, OLD s.v. 4, TLL 7. 2. 434.7 ff. This loses a specific
reference to invidia; yet the lack of such a reference might encourage
corruption.
33. quid tristes querimoniae . . . ?: the querimoniae are laments de sae-

culo; cf. Sall. Cat. 52. 7 ‘saepe de luxuria atque avaritia nostrorum civium
questus sum’, Sen. the Elder, Loeb edn. (Winterbottom), vol. 2, p. 635,
index under ‘commonplaces’, ‘on the age’, and ‘on the good old days’.
Dactylic verse uses querelae. The verb does not emerge until proficiunt in
v. 36 (so Porph.), but the second quid makes the drift clear; Palmer’s quo
is unnecessary and rhetorically inferior.
34. si non supplicio culpa reciditur: culpa is ‘wrong-doing’, here as

often of a sexual nature (OLD 3b). reciditur is a metaphor, not from
surgery as in Ov. met. 1. 190 f., but from pruning; cf. serm. 1. 3. 122 f. ‘et
magnis parva mineris / falce recisurum simili te’, Sen. dial. 4. 18. 2
‘difficulter reciduntur vitia quae nobiscum creverunt’. In the same way
luxuria is used both of rank vegetation and profligate self-indulgence.
35–6. quid leges sine moribus / vanae proficiunt?: vanae goes closely
with sine moribus: ‘what use are laws, vain as they are without morals?’
At the same time it marks a contrast with proficiunt (a word suited to
moralizing).
The superiority of mores over leges was particularly illustrated from
undeveloped societies; cf. Sall. Cat. 9. 1 ‘igitur domi militiaeque boni
mores colebantur; concordia maxuma, minuma avaritia erat; ius bonum-
que apud eos non legibus magis quam natura valebat’, Ov. met. 1. 90 ff.,
Pomp. Trog. fr. 35 ¼ Justin 2. 2 (on the Scythians) ‘iustitia gentis
ingeniis culta non legibus’, Tac. Germ. 19. 5 ‘plus . . . ibi boni mores valent
quam alibi bonae leges’, Woodman–Martin on ann. 3. 26. 1 ‘vetustissimi
mortalium, nulla adhuc mala libidine, sine probro scelere eoque sine
poena aut coercitionibus agebant’. For a different and defensible view-
point cf. Sen. epist. 94. 37 ‘leges quoque proficiunt ad bonos mores’.
Augustus claimed in the end to have taken charge of both mores and
leges; cf. 4. 5. 22 ‘mos et lex maculosum edomuit nefas’, epist. 2. 1. 2 f.
‘moribus ornes, / legibus emendes’ (with Brink), res gestae 6.1 ‘curator
legum et morum’. Tennyson no doubt had mores and leges in mind when
he wrote ‘Ring in the nobler modes of life, / With sweeter manners,
purer laws’ (In Memoriam cvi, 15 f.).
36–8: si neque fervidis / pars inclusa caloribus / mundi: for pars as a

region of the earth cf. 3. 3. 38–9 n.; for the collocation with mundi
cf. paneg. Mess. 150 ‘teque <manet> interiecto mundi pars altera sole’,
Lucan 4. 106 f. ‘sic mundi pars ima iacet quam zona nivalis / perpetuae-
que premunt hiemes’. H. is referring to the doctrine of the zones, which
was known to the pre-Socratics and Aristotle before Eratosthenes
(Powell, Coll. Alex. fr. 16): the temperate part of the earth comes
between regions that are uninhabitable from heat or cold; cf. Lucr.
5. 204 f. ‘inde duas porro prope partis fervidus ardor / assiduusque geli
casus mortalibus aufert’, Virg. georg. 1. 233 ff. (of the corresponding
zones of the sky) ‘quinque tenent caelum zonae: quarum una corusco /
semper sole rubens et torrida semper ab igni’, Ov. met. 1. 49 ff., Plin. nat.
hist. 2. 172 ‘media terrarum . . . exusta flammis et cremata comminus
vapore torretur’, Thomson (1948), index, Pease on Cic. nat. deor. 1. 24,
and especially K. Abel, RE Suppl. 14. 989 ff.
inclusa seems at first to suggest that the zone in question is sur-
rounded by two barriers of heat. Housman in the astronomical appendix
of his Lucan, pp. 330 ff., pointed to the idea that the middle strip of the
torrid zone was in fact habitable; cf. Strabo 2. 3. 2-3, citing Eratosthenes
and Polybius (34. 1. 1), and probably using Posidonius (fr. 49 with Kidd,
cf. fr. 210); see also Thomson, op. cit. 163, 209 f., and RE Suppl. 14.
1049 f., 1063 f. Housman gave the same interpretation to Lucan 5. 23 ff.
‘nam vel Hyperboreae plaustrum glaciale sub Vrsae / vel plaga qua
torrens claususque vaporibus axis / nec patitur noctes nec iniquos cres-
cere soles’; there he compared our passage. He took the same view at
Lucan 9. 538 f. ‘at tibi, quaecumque es Libyco gens igne dirempta, / in
Noton umbra cadit, quae nobis exit in Arcton’; he admitted that at first
sight these lines suggest the temperate zone to the south of the tropic of
Capricorn, but he argued from the following lines that Lucan was
referring to a temperate band round the equator.
Nevertheless, the theory is a surprising complication in our passage,
where we expect a simple contrast between the frozen Arctic and the
torrid equatorial zone; cf. 1. 22. 17 ff. ‘pone me pigris ubi nulla campis /
arbor aestiva recreatur aura, / quod latus mundi nebulae malusque /
Iuppiter urget; / pone sub curru nimium propinqui / solis in terra
domibus negata . . . ’, Virg. Aen. 7. 225 ff. ‘et si quem tellus extrema refuso
/ submovet Oceano, et si quem extenta plagarum / quattuor in medio
dirimit plaga solis iniqui’. In NR’s opinion a region ‘enclosed in heat’
need not have a temperate centre (any more than an area enclosed in
mist or darkness need have a bright centre); in Lucan 5. 24 (quoted
above) he takes the ‘clausus vaporibus axis’ to be the same as the ‘plaga
torrens’. If that seems too difficult, RN tentatively suggests exclusa
caloribus (‘shut off by the heat’); for the instrumental ablative cf. Caesar,
bG 5. 23. 5 ‘ne anni tempore a navigatione excluderentur’, Virg. georg.
4. 147 ‘spatiis exclusus iniquis’, perhaps also Lucan 5. 24 ‘claususque
vaporibus axis’, where some interpret ‘a region barred by the heat’.
Copyists sometimes change prefixes; and if exclusa caloribus was wrongly
interpreted as ‘shut off from heat’, that could have caused the corrup-
tion. The merchant of v. 40 defies Nature’s boundaries, and this point is
made more explicitly by exclusa than by inclusa; cf. 1. 3. 21 ff. and Sen.
Med. 335 ff.
38. nec Boreae finitimum latus: for latus of an outlying region of the
earth cf. 1. 22. 19 ‘quod latus mundi’ (cited in the previous note), Pers.
6. 76 ‘omne latus mundi’, TLL 7. 2. 1028. 50 ff., OLD 7a. In our passage
mundi is perhaps to be understood from the end of the previous clause;
hence its unusually emphatic position, when we might have expected a
pause after caloribus. But as pars needs no genitive (3. 3. 38 cited above)
perhaps mundi should be taken primarily with latus (with a comma after
caloribus); for the word-order cf. 11 n.
39. durataeque solo nives: if this is right it must mean ‘snows hardened
on the ground’ (see OLD 1b for this meaning of duro); but without in or
an adjective solo is oddly attached (even though it appears in a sand-
wiched position). Bentley proposed gelu, but this is remote; it cannot be
defended by ps.-Acro’s note ‘gelu, nimietate frigoris solidatae’, for there
gelu is part of the explanation, not part of the lemma. A lesser change
would be S. Wyngaarden’s polo, ‘hardened by the arctic sky’; cf. Strab.
2. 3. 1 a b æe fiH ºfiø (IŒ KØ) Øa łF
. RN has tentatively
considered durataeque solum nives: ‘the fact that hardened snow makes
up the ground does not drive away the merchant’; for the mannered
expression see below on 42.
40. mercatorem abigunt: poets and moralists regarded the trader as

reckless because of his avarice; see N–H on 1. 1. 16 and 1. 3 (p. 43), West
on Hes. op. 686 ff., Murgatroyd on Tib. 1. 3. 39–40, Fedeli on Prop. 1. 17.
13 f., T. Heydenreich, Tadel und Lob der Seefahrt, 1970 (including modern
imitations).
40–1. horrida callidi / vincunt aequora navitae: not ordinary seamen

but merchant ship-owners; for the form navitae see 3. 4. 30 n. callidi
describes the cleverness that comes from experience (cf. Soph. Ant. 348
æØæÆc Iæ in the celebration of human ingenuity); here the word
has a pejorative tone, hinting that these men are too clever by half.
horrida implies both ‘rough’ and ‘frightening’, as at epod. 10. 3; it makes
an oxymoron with aequora, which properly describes a flat surface (cf. 1.
5. 5 f. ‘aspera / . . . aequora’). For the use of vincunt cf. Tac. Agr. 25. 1
‘(cum) hinc terra et hostis, hinc victus Oceanus militari iactantia com-
pararentur’; it corresponds to nec . . . abigunt better than Cornelissen’s
unnecessary findunt.
42. magnum Pauperies opprobrium iubet: the clause is still introduced

by si (36). Kiessling and others put a question-mark at the end of 41, but
then the transition is too abrupt; and the relevance of moribus (35) is
clearer if it is picked up in the same sentence by virtutis (44). For
the sandwiched apposition cf. 1. 20. 5 with N–H, 4. 8. 31 ‘clarum
Tyndaridae sidus’, epist. 1. 18. 104 ‘gelidus Digentia rivus’, and especially
J. B. Solodow, HSCP 90, 1986: 129 ff. He points out that the outside
element is normally more important, with the inside element added in
apposition (Virg. ecl. 1. 57 ‘raucae tua cura palumbes’). But this tendency
does not apply when the inside element is a single word.
Pauperies describes straitened circumstances rather than destitution
(for the distinction cf. Ar. plut. 552 ff.), and is not a disgrace; cf. 3. 2. 1
‘angustam amice pauperiem pati’, serm. 2. 3. 91 ff. ‘credidit ingens /
pauperiem vitium . . . ’, Sall. Cat. 12. 1 ‘pauperies probro haberi’, Sen.
epist. 115. 11, Mayor on Juv. 3. 152. Like Æ, the word is easily personi-
fied in popular moralizing; cf. 3. 29. 55 f. with note. In our passage
magnum opprobrium is an integral part of the subject: the avaricious are
impelled not so much by poverty (thus Theog. 649 ff., Lucian, cited in
note on 43) as by the disgrace of poverty (Heinze). This point is obscured
in translation if Poverty is made the subject; on the other hand, if we say
‘the reproach of poverty’, that destroys the personification; for a similar
dilemma cf. 2. 4. 10 ‘ademptus Hector’ with N–H’s note.
43. quidvis et facere et pati: the seaman combines arrogant recklessness
with an unnatural tolerance of suffering; cf. 1. 3. 25 ‘audax omnia perpeti’
(in a similar context). To stop at nothing is a mark of wickedness (cf. the
Greek ÆFæª ), but by ancient standards to submit to every indignity
is just as deplorable; cf. Arist. eth. Nic. 2. 1108a and 4. 1126a. For the
collocation of facere and pati (like ØE and
Ø cf. Cic. Pis. 11 ‘qui
nihil sibi umquam nec facere nec pati turpe duxit’ (with a sexual
innuendo), Vell. 2. 100. 3 with Woodman, TLL 6. 1. 124. 74 ff.; for a
more favourable use cf. Liv. 2. 12. 9 ‘et facere et pati Romanum est’. For
similar expressions about poverty in Greek cf. Lucian, apol. 10 c
Æ Æ ØE ŒÆd
Ø IÆŁıÆ ‰ KŒªØ Ø ÆP,
less closely Timocles, PCG 7, p. 776, fr. 30 ººf ªaæ KŁ Æ
ØÆØ = IØ ÆH æªÆ Ææa Ø ØE, cf. fr. 37.
44. Virtutisque viam deserere arduae: the steep path to #æ goes
back to Hes. op. 287 ff. (see West), especially 289 ff. B IæB
ƒæHÆ Łd ææØŁ ŁŒÆ = IŁÆØ ÆŒæe b ŒÆd ZæŁØ
r K ÆPc = ŒÆd æ
f e æH Kc N ¼Œæ ¥ ŒÆØ, =
ÞØ c ØÆ ºØ,
Æº æ KFÆ (though Hesiod is thinking of
the path to glory rather than moral excellence), Simonides, PMG 579,
Pearson on Soph. fr. 397. The image was given an ethical application
in popular philosophy, as was the kindred parable of Hercules at the
crossroads; cf. Xen. mem. 2. 1. 21–34 ¼ Prodicus B2, Pers. 3. 56 f., 5. 34 f.,
Lact. inst. 6. 3, Otto 36, J. Alpers, Hercules in Bivio, Diss. Göttingen
1912, E. Panofsky, ‘Hercules am Scheidewege’, Studien der Bibliothek
Warburg 18, 1930 (with Renaissance and modern illustrations),
O. Becker, ‘Das Bild des Weges’, Hermes Einzelschriften 4, 1937: 57 ff.,
Bompaire 258 ff.
deserere is Bentley’s conjecture for deserit of the MSS (also in Porph.’s
lemma); for the elision cf. 1. 3. 12 ‘praecipitem Africum’. H is talking
not about what Poverty does but about what the reproach of poverty
makes people do, cf. iubet in v. 42. For the abandonment of Virtue cf.
serm. 2. 3. 13 ‘invidiam placare paras virtute relicta?’, epist. 1. 16. 67 ‘locum
virtutis deseruit’, Lact. inst. 7. 1 ‘et virtutis viam deserunt cuius acerbitate
offenduntur’ (apparently influenced by our passage).
arduae includes the ideas of both ‘lofty’ and ‘difficult’, as at 1. 3. 37 f.
where ‘nil mortalibus ardui est’ is followed immediately by ‘caelum
ipsum petimus’; for the combination with virtus cf. Stat. silv. 5. 2. 98 f.
‘ardua virtus / affectata tibi’, Theb. 10. 845 f. ‘hac me iubet ardua virtus /
ire’, Lucan 4. 576 (in a somewhat different sense). Lambinus conjectured
arduam, which is closer to Hesiod; for the ardua via to Virtus cf. also
[Sall.] epist. 2. 7. 9, Corn. Sev. fr. 2 (Courtney, FLP p. 321), Sil. 2. 578,
15. 102. Lambinus did not print his conjecture because of the absence of
manuscript support, but RN views it with some favour. The merchant
abandons the arduous path to Virtue (which bears the emphasis) in
order to pursue a no less arduous course of his own (43 ‘quidvis et facere
et pati’); in epist. 1. 1. 45 ff. H denies that morality is more difficult than
the merchant’s hyperactivity, and similarly here the epithet ‘arduous’
should perhaps not be applied to Virtue exclusively.
45. vel nos in Capitolium . . . : with an emphatic and repeated nos

Horace exhorts his fellow-citizens in the manner of early Greek poetry;
cf. 3. 14. 1–2 n. First he urges them to dedicate their valuables inside
Jupiter’s Capitoline temple (hence in rather than ad). There was a
legend that Pythagoras persuaded the women of Croton to dedicate
their finery to Hera ( Just. 20. 4. 11, Iamblichus, vit. Pyth. 11); in 217 bc,
after Hannibal’s victory at the Trebia, Roman matrons donated money
to Juno on the Aventine (Liv. 22. 1. 18); in 210 all classes made generous
contributions of gold and silver for the war effort (26. 36), a practical
purpose not envisaged in Horace’s poem. It is relevant that Octavian
had acquired treasures from Cleopatra (Dio 51. 22. 3) which he dedicated
on the Capitol, presumably at or soon after his triumph in 29 bc; cf. res
gestae 21. 2 ‘dona ex manibiis in Capitolio . . . consecravi’, Suet. Aug. 30. 2
‘utqui in cellam Capitolini Iovis sedecim milia pondo auri gemmasque
ac margaritas quingenties sestertium una donatione contulerit’.
Horace keeps us in suspense about what he has in mind: the valuables
appear only in v. 48 and the verb not till mittamus (50), where it means
‘let fall’. Commentators suggest that we should recognize a zeugma and
infer something like feramus here. But the difficulty is lessened if we first
understand mittamus in the different sense of ‘send’ or even ‘bring’;
for the latter usage (¼ the Greek Ø) Lucian Müller cites Virg.
ecl. 9. 6 ‘hos illi . . . mittimus haedos’, Aen. 6. 380 ‘tumulo sollemnia
mittent’ (as here of a sacred offering). ‘Consign’ in English would
keep the ambiguity.
46. quo clamor vocat et turba faventium: clamor is here applause (OLD
1 c), like acclamatio. faventium goes with clamor as well as with turba (cf.
Liv. 1. 25. 9 ‘clamore . . . faventium’); for such demonstrations of approval
cf. Liv. 1. 12. 10, Virg. Aen. 5. 148 f. ‘tum plausu fremituque virum
studiisque faventum / consonat omne nemus’, TLL 6. 1. 377. 22 ff.
Horace is describing a completely imaginary procession of citizens
offering up their valuables; it would be tactless to suggest that they
should tag on at the end of Octavian’s triumph.
47. vel nos in mare proximum . . . : by an even more extravagant fantasy

Horace then urges the Romans to dump their treasures in the sea.
proximum implies ‘most readily available’; there is to be no shilly-
shallying. The proposal belongs to popular philosophy; cf. Lucian,
Timon 56 N ªæ Ø ŁØ, ºØÆ b ‹º K c ŁºÆÆ KÆºE
ÆPe (e ºF), Pb IÆªŒÆE Iæd IªÆŁfiH ZÆ, vitarum auctio
9, Philostratus, vit. Apoll. 1. 13, where Crates the Cynic is alleged
actually to have done this; cf. Jerome, adv. Iov. 2. 9. 338 ‘Crates ille
Thebanus, proiecto in mare non parvo auri pondere, ‘‘abite’’ inquit
‘‘pessum, malae cupiditates; ego vos mergam ne ipse mergar a vobis’’ ’.
In the Duomo at Siena there is an intarsio by Pinturicchio of about 1505
which portrays Sapientia (or rather Virtus?) between Socrates and
Crates, who is throwing jewels over a cliff (Pöschl 381 ff. with illustra-
tion); an inscription above reads ‘huc properate, viri, salebrosum scan-
dere montem’ (cf. v. 44 above on the path to Virtus). H may also have
been influenced by parables about jettisoning treasure to save a ship
(3. 29. 61 n.).
48. gemmas et lapides aurum et inutile: gemmae and lapides are not
consistently distinguished, but where the two words are combined the
latter refers to pearls; cf. Tib. 1. 8. 39, Manil. 4. 398 f., where Housman
compares Cic. Verr. 4. 1 ‘ullam gemmam aut margaritam’ among other
passages; add Suet. Aug. 30. 2 cited above on 45, TLL 7. 2. 951. 56 ff. This
explanation is supported by Ov. med. fac. 20 ff. ‘conspicuum gemmis
vultis habere manum: / induitis collo lapides Oriente petitos / et
quantos onus est aure tulisse duos’, Mart. 11. 49 (50). 4 ‘gemma vel
a digito vel cadit aure lapis’. Though men might wear ostentatious
rings, jewellery was naturally associated with women (for details see
J. P. V. D. Balsdon, Roman Women, 1962: 262 ff., with bibliography
on 336); for typical denunciations cf. Sen. ben. 7. 9. 4, Plin. nat. hist.
9. 104–5.
inutile is not a euphemism for perniciosum, as some suggest, but
‘useless’ in a moralist’s sense (supervacuum); cf. Lucian, Timon 56
(cited on 47) , anth. Lat. 649. 2 ‘caecus inutilium quo ruit ardor
opum?’
49. summi materiem mali: the parallels suggest that the ‘supreme evil’
here is avarice, not luxury or the decline of morals; cf. Cato p. 82, fr.
1 Jordan (¼ Gell. 11. 2. 2) ‘avaritiam omnia vitia habere putabant’, Paul, 1
Timothy 6: 10 ÞÆ ªaæ ø H ŒÆŒH KØ ØºÆæªıæÆ, Diog.
Laert. 6. 50 (on Diogenes) c ØºÆæªıæÆ r æ ºØ ø H
ŒÆŒH, Otto 51. H must have been influenced particularly by Sall. Cat.
10. 3 ‘primo pecuniae, deinde imperi cupido crevit; ea quasi materies
omnium malorum fuere’. But in Sallust avarice is the raw material from
which other evils originate; in our passage gold is the substance that
fuels avarice itself (cf. TLL 8. 463. 30 ff.). Justin combines both ideas
when he says ‘auri argentique usum velut omnium scelerum materiam
sustulit’ (3. 2. 12).
50. mittamus, scelerum si bene paenitet: mittamus combines the ideas
of ‘drop’ (OLD 8) and ‘get rid of ’.The conditional clause is to be taken
with mittamus, not with 51 ff.; for the very strange and extreme recom-
mendations need the further elaboration provided by the si-clause,
whereas the general assertion which follows does not; moreover, it is
more usual to round off the sentence at the end of the line. H is
referring to the crime of repeated civil wars (for scelerum cf. 1. 2. 29
with N–H, 1. 35. 33); for these he gives an unconvincing economic
interpretation (so Lucan 1. 158 ff.).
51–2. eradenda cupidinis / pravi sunt elementa: cupidinis (masculine

in H as in early Latin) is generally taken here as ‘avarice’ (so in 2. 16. 15,
3. 16. 39); this coheres alike with the previous lines and the end of the
poem. Mitscherlich, however, interpreted it as the self-indulgence trad-
itionally associated with young men; this suits the remedy prescribed in
54 f. ‘asperioribus . . . studiis’. In fact both views are true: cupidinis in the
general sense of ‘desire’ suits the transition from one phase of the
argument to the next; other instances of words with a ‘cardinal’ function
are serm. 1. 10. 39 (theatris), 2. 1. 82 (mala carmina), epist. 2. 2. 57 (quid
faciam vis?). For the importance of early training cf. epist. 1. 2. 64–70,
Pers. 5. 36 ff., [Plut.] de lib. educ. 4, Juv. 14. 123 f. ‘sunt quaedam vitiorum
elementa, his protinus illos iuvenes / (pater) imbuit et cogit minimas
ediscere sordes’ (much of the satire is relevant). As elementa often has
the specific sense of ‘letters of the alphabet’ (Greek Ø
EÆ), it suits
eradenda, which suggests erasures on a wax tablet; cf. also Sen. epist.
104. 20 ‘omnem ex animo erade nequitiam’.
52–4. et tenerae nimis / mentes asperioribus / firmandae studiis:

firmandae is Bentley’s conjecture for formandae of the MSS. He recog-
nized that the latter is the vox propria for moulding the characters of the
young; cf. epist. 2. 1. 128 ‘mox etiam pectus praeceptis format amicis’ with
Brink, TLL 6. 1. 1104. 6 ff.; if tenerae implies ‘young and malleable’ (cf.
Pers. 5. 36 ff. ‘teneros tu suscipis annos . . . / artificemque tuo [animus]
ducit sub pollice vultum’), formandae is particularly appropriate (Bentley
cites Sen. epist. 34. 1 ‘qui ingenia educaverunt et quae tenera formaver-
unt, adulta subito vident’, Stat. Ach. 1. 478 ‘teneros formaverit annos’).
But then he subtly observes that nimis is a fatal objection: how can
minds be reproached for being too young and malleable? Therefore
tenerae must mean rather ‘soft and effeminate’; such characters need to
be toughened and stiffened by asperioribus studiis, i.e. firmandae sunt ; cf.
3. 5. 46, Quint. 10. 1. 131 on Seneca ‘iam robustis et severiore genere satis
firmatis legendus’, TLL 6. 1. 810. 76 ff. This argument appears to be
unanswerable.
54–5. nescit equo rudi / haerere ingenuus puer: for haerere, ‘to hold
one’s seat’, cf. Cic. Deiot. 28 ‘itaque Deiotarum cum plures in equum
sustulissent, quod haerere in eo senex posset, admirari solebamus’, Ov.
met. 4. 27 (on Silenus) ‘pando non fortiter haeret asello’, Sidon. carm.
2. 264 ‘equo ceu fixus adhaeret’; this was not easy in a world without
stirrups. Riding was important in Augustan youth-movements (cf. N–
H on 1. 8, p. 108), and every young officer had to know how to ride a
horse, though the cavalry were mainly supplied by the socii. The drill
was believed to be beneficial in itself, and anybody with duties in
difficult country would need a horse.
rudi is Cornelissen’s conjecture for rudis of the MSS and Porph.’s
lemma (Mnem. 16, 1888: 311); it describes untrained animals, e.g. oxen
(Varr. rust. 1. 20. 1, Ov. fast. 1. 83 ‘rudes operum’) and elephants (bell.
Afr. 27. 2). Cornelissen argued that any boy of good family must have
been able to keep his seat on a horse, provided it had been broken in;
on the other hand, riding an untamed animal was evidence of energy,
and is significantly combined with hunting at serm. 2. 2. 9 f. ‘leporem
sectatus equove / lassus ab indomito’ (see further, 56–7 n.). The conjec-
ture also provides a satisfactory balance between equo rudi and ingenuus
puer, whereas the asyndetic combination of rudis and ingenuus is un-
usual in Latin poetry (for limited exceptions see Virg. Aen. 5. 24 with
Heyne–Wagner, 10. 391 with Harrison, H–Sz 160 f.). In favour of rudis
it might be contended that it balances doctior (56), which comes at
the same place in the line, but the natural connection of doctior is
with the foregoing nescit. Hiatus between lines is usually rare in an
Asclepiad system (N–H vol. 1, p. xl), but it occurs in this poem at 11 f.,
24 f., and 61 f.
56. venarique timet: ancient hunting was done for the most part on
foot (cf. Ov. met. 8. 331 ff. and N–H on 1. 1. 25, 1. 37. 18), and the danger
arose not from the riding but from the quarry, particularly wild boars
(see Pease on Virg. Aen. 4. 158 f.); as a result the sport was regarded as
evidence of masculinity and as a training for war ( J. Henderson, PCPS
47, 2001: 19; cf. serm. 2. 2. 10 f., epist. 1. 18. 49 f. ‘Romanis sollemne viris
opus, utile famae / vitaeque et membris’). In giving moral precepts to
the well-born young Lollius, H uses as examples the training of horses
and hounds (epist. 1. 2. 64 ff.). For Roman hunting see J. Aymard, Les
Chasses romaines, 1951, Anderson 1985.
56–7. ludere doctior / seu Graeco iubeas trocho: the trochus was a metal
hoop which was bowled along (hence the Greek æ
from æ
ø); cf.
Prop. 3. 14. 6 ‘versi clavis adunca trochi’, Ov. ars 3. 383 with Brandt,
Mart. 11. 21. 2 ‘arguto qui sonat aere trochus’ with Kay. H presents it
here as paradoxical that mere play should be regarded as an art, but he
admits elsewhere that the hoop needed skill: ars 379 f. ‘ludere qui nescit
campestribus abstinet armis, / indoctusque pilae discive trochive quies-
cit’. For the Greek character of such sports cf. serm. 2. 2. 10 ff. ‘vel si
Romana fatigat / militia [i.e.hunting, and riding an unbroken horse]
assuetum graecari, seu pila velox / molliter austerum studio fallente
laborem, / seu te discus agit, pete cedentem aera disco’; the Greek
word trochus could itself convey contempt (cf. other Greek examples at
Juv. 3. 67 f.) and H naturally uses the everyday Graeco rather than the
poetical Graio. Yet in other contexts H and his friends take part in ball-
games (serm. 1. 5. 48 ff., 1. 6. 126), and Strabo talks of the Campus
Martius being full of people playing with balls and hoops (5. 3. 8). For
the hoop see further H. A. Harris, Sport in Greece and Rome, 1972: 133 ff.
with plates 52 and 54, RAC 10. 855.
58. seu malis vetita legibus alea: for this use of malle in presenting an
alternative cf. 1. 4. 12, 3. 4. 3. Because dicing often involved ruinous
gambling (cf. epist. 1. 18. 21 ‘quem praeceps alea nudat’), it was already
prohibited in the time of Plautus (mil. 164 f.); the Digest mentions three
laws and a senatus consultum (11. 5. 2–3), but the rules were relaxed at the
Saturnalia (Mart. 5. 84 etc.). We hear of penalties of four times the
wager (ps.-Asc. on Cic. Caecil. 24, p. 194 Stangl) and even of exile (Cic.
Phil. 2. 56), but they had little effect; even emperors played, including
Augustus himself, and Claudius wrote a treatise on the subject (Suet.
Aug. 71. 1, Calig. 41–2, Claud. 33. 2, Dom. 21). For the procedures of
dicing cf. J. P. V. D. Balsdon, Life and Leisure in Ancient Rome, 1969:
155 ff., RE 13. 1933 ff., RAC 10. 849 f.; for denunciations and penalties cf.
Mayor on Juv. 11. 176, Owen on Ov. trist. 2. 472, RE 1. 1358 f.
59–60. cum periura patris fides / consortem <et> socium fallat et

hospitem: it is not surprising that the son breaks the law in his self-
indulgence seeing that the father shows no scruple in his acquisitiveness.
Early editors (including Lambinus) joined the cum clause to what
follows (62 ff.), not to what precedes; but then the connection at 59 is
very abrupt. In the remarks on the son we should not lose sight of the
father, whose materialism is the main object of H’s criticism.
Every word underlines the father’s villainy: his promises are perjured
(periura fides being an oxymoron), and he is guilty of deceiving people to
whom he owes a special duty. For cheating a business partner cf. epist.
2. 1. 122 f. ‘non fraudem socio . . . incogitat ullam’ with Brink’s note,
Cic. Sex. Rosc. 116, Rosc. Com. 16 ‘aeque enim perfidiosum et nefarium
est fidem frangere . . . et socium fallere qui se in negotio coniunxit’. The
MSS read consortem socium, which implies that the partner shared the
capital, but the attribute adds little; in view of Porph’s comment
(cum omnes scelere grassentur in hospitem in socium in consortem) Bentley
plausibly inserted et, taking consortem to mean ‘co-heir’. For treachery to
one’s guest-friend (perhaps the climax of the man’s bad faith) cf. 2. 13. 8
with N–H; to maintain a proper balance, hospitem seems preferable
to hospites, even though the latter has rather greater manuscript
support and the former produces a hiatus before indignoque in 61 (see
above on 54).
61–2. indignoque pecuniam / heredi properet: the acquisitive man

derives no enjoyment from his money, which will pass to his heir; cf.
2. 3. 19 f. ‘exstructis in altum / divitiis potietur heres’; this heir must be
the spendthrift of the preceding lines, not a remote connection as might
often be the case (N–H on 2. 14. 25). pecunia is an appropriately prosaic
word (Axelson 108); cf. 3. 16. 17, 4. 9. 38. For the greedy man in a hurry
cf. Juv. 14. 178 ‘properantis avari’, G. A. Gerhard, Phoinix von Kolophon
(1909), 99 f.; for the transitive use of properare cf. N–H on 2. 7. 24
(deproperare), epod. 12. 22, Virg. georg. 4. 171.
62–3. scilicet improbae / crescunt divitiae: the emphatic improbae is

predicative (‘wealth accumulates shamelessly’) and suggests an inability
to draw the line (Mynors on Virg. georg. 1. 146); it does not refer merely
to the ill-gotten gains of 59 f. As divitiae is not accompanied by a
genitive, H is presumably stating a generalization; for this feature at
the end of a poem cf. 1. 24. 19 f., 4. 12. 28, Esser 62 ff. That suits scilicet
which means ‘in truth’ and is not here a concessive particle (OLD 2e)
balancing tamen.
63–4. tamen / curtae nescioquid semper abest rei: curtus means ‘having
something broken off ’, and hence ‘defective’, cf. mancus. For the com-
monplace that greed is never satisfied cf. 2. 2. 13 ‘crescit indulgens sibi
dirus hydrops’ with N–H, 3. 16. 42 f., epist. 1. 2. 56 ‘semper avarus eget’,
Otto 51; the observation comes not just from Horace, the objective
commentator, but also represents the mentality of the greedy man
himself, who is never content; cf. serm. 1. 1. 62 ‘nil satis est’ and the
prayers of the fool in serm. 2. 6. 8 ff. curtae must be taken as an instance
of prolepsis (for which see H–Sz 413 f.), i.e. ‘something is always missing
from our possessions, so that they are defective’. In the belief that this
was over-complicated, Cornelissen proposed partae, ‘the wealth already
acquired’ (Mnem. 16, 1888: 311 f.); this gives fair, if rather bland, sense,
but the change is not easy to explain. Campbell (edn. 2) proposed
structae; the word suits the piling up of wealth (serm. 1. 1. 34 f. ‘addit
acervo / quem struit’, 1. 1. 44, carm. 2. 3. 19, Pers. 2. 44). So the conjecture
should be taken seriously: we look for a contrast with nescioquid semper
abest, not an anticipation (Syndikus 2. 207 n.).
2 5 . QVO ME , BACCH E , R A P I S ?
[Commager 344 ff.; P. J. Connor, AJP 92, 1971: 266 ff.; Davis 111 ff.; Fraenkel 257ff; Pöschl
164 ff. (¼ H. Oppermann, Wege zu Horaz, 1972: 258 ff.); Troxler-Keller 47 ff.]
1–6. Where are you rushing me, Bacchus, possessed by your divinity? In
what cave shall I rehearse my song on Augustus’ consecration among the
stars? 7–14. I shall utter something striking and original. As the wakeful
Maenad gasps at the mountains she has traversed, so I leave the beaten track
and marvel at the uninhabited landscape. 14–20. O god that has power over
the potent Bacchae, I shall say nothing mundane. It is delightful as well as
dangerous to follow you.
Maenadism left a lasting impression on the ancient imagination, but

its exact nature is a subject for debate. In his commentary on Euripides’
Bacchae and the third chapter of The Greeks and the Irrational (1951)
E. R. Dodds took Dionysiac ecstasy seriously as a manifestation of
paranormal psychology; he rightly dissociated it from the later picture
of ‘Jolly Bacchus’ and literal intoxication. More recently Albert Hen-
richs has pointed to the confusion of myth, with its fantasies about
snakes and raw flesh, and ritual as it was actually practised (see especially
HSCP 82, 1978: 121 ff.); largely relying on Hellenistic inscriptions, he
postulates an organized celebration on the mountains that was physic-
ally exhausting rather than psychologically disturbing. J. N. Bremmer
adopts an intermediate position (ZPE 55, 1984: 267 ff.); while acknow-
ledging the exaggerations in Euripides, he recognizes that repetitive
music and cold night air could produce strange states of mind.
R. Osborne insists that late inscriptions should not outweigh the exten-
sive evidence for Maenadism in literature and art (Aeschylus as well as
Euripides, and a very large number of vases from 530 onwards); parallels
in other cultures cannot be brushed aside, and a genuine quest for
communion with the divine ought to be recognized (in Greek Tragedy
and the Historian, ed. C. Pelling, 1997: 187 ff.). C. Segal among others
has emphasized Dionysus as the ‘outsider’ and the Maenads’ impulse to
escape from the literal and metaphorical bounds of the city (Dionysiac
Poetics and Euripides’ Bacchae, 2nd edn. 1997: 350 ff.). See further
A. Henrichs, OCD, 479 ff. (with a valuable bibliography), Horsfall on
Virg. Aen. 7. 373 ff.
Plato with some irony associated the poet’s fine frenzy with
Maenadism (see Ion, passim, apol. 22a–c, Men. 99c–e, Phaedr. 245,
Russell on Longinus 13. 2); this should be distinguished from the
inspiration provided by the Muses since the earliest Greek poetry
2 5 . QVO M E , BACC H E , R A P I S ? 297
(Hes. theog. 31 f., P. Murray, JHS 101, 1981: 87 ff.). Bacchus was the patron
of dithyramb and drama, and his status as a poetical god was enhanced
in the Hellenistic and Roman periods (N–H on 2. 19, 316 f.). The
Roman dramatists continued to write Bacchic plays (e.g. the Lycurgus
of Naevius, the Pentheus of Pacuvius, and the Bacchae of Accius), and
passages in Plautus mock their conventions (cf. 1 n.). There was an
outburst of the cult in 186 bc which was controlled by the surviving
‘Senatusconsultum de Bacchanalibus’ (ILLRP 511), and described by
Livy with some sensationalism (39. 8–18, J. A. North, PCPS 25, 1979:
85 ff., P. G. Walsh, G&R 43, 1996: 188 ff.). But by the Augustan period
Maenadism in Italy seems no more than a literary topic that appealed to
poets’ taste for the exotic and bizarre (as did the continuing rites of
Cybele); cf. especially Catull. 64. 251, Virg. Aen. 4. 300 ff. (where Pease
supplies many later parallels), 7. 373 ff. (Amata). Maenads continued to
figure in art (A. Bruhl, Liber Pater, 1953: 145 ff.), and when Messalina put
on a Bacchic performance (Tac. ann. 11. 31. 2–3) that, needless to say, had
no religious significance.
Horace in his hexameter poems treats Bacchic possession with satir-
ical humour. His own poetry, according to Damasippus, is one of the
symptoms of his insanity (serm. 2. 3. 321.). Ever since Liber (i.e. Bacchus)
enrolled poets in his thiasos, the Muses have smelt of drink (epist. 1. 19.
3 ff.). Ever since Democritus pronounced poets mad, they have sought
solitude and avoided the baths (ars 296 ff.), and sane people shun them
in case they recite them to death (455 ff.). Horace recognizes the place of
ingenium (natural ability) as well as ars (craftsmanship) in poetry
(408 ff.), but in that passage he does not consider the notion that a
poet may be ‘carried away’.
The ethos of the Odes is very different: for Bacchus’ various roles see
T. Oksala, Religion und Mythologie bei Horaz, 1973: 43 ff., H. Krasser,
‘Horazische Denkfiguren: Theophilie und Theophanie als Medium der
poetischen Selbstdarstellung des Odendichters’, Hypomnemata 106,
1995: 131 ff. He appears often enough as the god of wine (for his wilder
aspects in this capacity cf. 2. 7. 26 f. ‘non ego sanius / bacchabor Edonis’,
3. 19. 14 ff.); in particular 1. 18 shares some of the rapidity and excitement
of our poem, though it associates the god with mysteries (12 ff.) rather
than Maenads. Elsewhere he appears as a god of poetry (note the
nymphs and satyrs in the programmatic overture at 1. 1. 31); Bacchic
enthusiasm is a more violent force than the inspiration attributed to
Apollo and the Muses (1 n.), and the symbolic terrain of our poem is
more rugged than the ideal landscape where Horace hears Calliope
(3. 4. 5 ff.). The closest analogy is in 2. 19 (‘Bacchum in remotis carmina
rupibus / vidi docentem’), which describes the poet’s confused emotions
when possessed by the god (cf. below, 18 n.); but there the Maenads are
mentioned only in a line (‘pervicacis . . . Thyiadas’ in v. 9). They occupy
a more central position in our poem, where their ecstasy is a metaphor
for the poet’s frenzy. For this condition cf. Theseus’ famous lines in
A Midsummer Night’s Dream, beginning ‘The lunatic, the lover and
the poet / Are of imagination all compact . . . / The poet’s eye in a fine
frenzy rolling, / Doth glance from heaven to earth, from earth to
heaven’ (v . i. 7 ff.), and Coleridge’s picture of the strange visionary
‘And all should cry Beware! Beware! / His flashing eyes, his floating
hair! / Weave a circle round him thrice, / And close your eyes with holy
dread, / For he on honey-dew hath fed, / And drunk the milk of
Paradise’ (Kubla Khan 49 ff.).
After his ecstatic outburst Horace makes a surprising announcement:
in his Bacchic cave he will rehearse the apotheosis of Augustus (6 n.),
and the result will be something remarkable, fresh, and never said before
(7 f.). What work does he have in mind? By the usual account (e.g.
Fraenkel 259) he is referring to the ‘Roman Odes’ and proclaiming their
originality (cf. 3. 1. 2 f. ‘carmina non prius audita’); but that does not
account for the poem’s late position in the book. The most likely
explanation is that the future tenses in 3. 25. 7 and 18 represent what
Horace is actually doing; in the same way at 1. 12. 21 ‘neque te silebo’ he
is already praising Liber (cf. 4. 9. 30 ff. in honour of Lollius), and in 1. 21. 2,
when his choir sing ‘dicite Cynthium’, they are already telling of Apollo.
When Pindar announces what he will sing, he is referring to the
immediate occasion, not to some future poem (E. Bundy, ‘Studia
Pindarica I’, University of California Publications in Classical Philology
18, 1962: 21 f.). In the same way Horace’s intention is fulfilled in the
ode itself.
The claim to originality can be seen in the same way. Horace does not
here mean the combination of Roman political themes and Greek poetic
forms, for he had written other such poems from the time of the Epodes.
He is thinking rather of the spirit that informs the ode and the manner in
which it is expressed: here as nowhere else he has conveyed the strange
feeling of Dionysiac possession and the excitement associated with the
dithyramb (for which see B. Zimmermann, OCD 487). The exuberance
and rapidity of that kind of poem is exemplified in Pratinas, PMG 708
(below, 14 n.) and Pindar fr. 75. Horace cannot produce this effect by an
accumulation of short syllables (as Catullus does in no. 63), but by
persistent enjambment he reinforces the choriambs to suggest a move-
ment that is out of control (Pöschl 171); cf. the imitation of Pindar’s
‘torrential’ lyrics in 4. 2. 5 ff., where the verse sweeps now across caesuras,
now over line-endings, and even over strophic divisions. When Horace,
‘velox mente nova’, imitates the revelling Maenads and follows the god in
ecstasy, he creates a poem of unusual rapidity.
That is not to deny that the ode, like some others at the end of
the book, has a retrospective element. It looks back to the theme
of Augustus’ greatness that had found its most memorable expression in
3. 2–6. Yet 3. 25 is unlike 3. 4, where the inspired poet and the divinely
guided Princeps are equally prominent. Here, although the celebration
of Augustus is the avowed purpose of the ode, that is not its most
striking feature; indeed a Maenad’s ecstasy makes an odd analogy for
political commitment, however fervid. What one takes away, rather, is
some sense of the mysterious phenomenon of poetic possession.
Metre: alternating Glyconics and Asclepiads, as in 3. 24, though the use

of enjambment produces a notably different effect.
1–2. Quo me, Bacche, rapis tui / plenum?: H professes to be swept

along like a Maenad possessed by Bacchus; the religious and literary
background are indicated in the introduction. For the use of rapere in
Bacchic contexts cf. Ov. fast. 4. 457 f., Sen. Ag. 722, Lucan 1. 676, Stat.
Theb. 5. 93, Herrick ‘Whither dost thou whorry me, / Bacchus, being
full of thee?’ In such contexts the Maenad does not know where she is
going (cf. Plaut. Men. 835 ‘Bromie, quo me in silvam venatum vocas?’,
clearly parodying the topic as it occurred in Greek and Roman tragedy);
the question does not expect an answer any more than when Cassandra
cries ‘where have you brought me?’ and the Chorus uncomprehendingly
explains ‘to the house of the Atridae’ (Aesch. Ag. 1087 f.). For plenum dei
of divine possession (¼ Ł ) cf. 2. 19. 6 ‘plenoque Bacchi pectore’ with
N–H; note especially Virg. Aen. 6. 77–80 with Norden, Sen. suas. 3. 5–7,
Luke 1: 67 ˘Æ
ÆæÆ . . . KºŁ —Æ ,`ªı ŒÆd æı,
E. R. Dodds, The Greeks and the Irrational, 1951: 64 ff., F. Pfister, RAC 4.
944 ff. As the Maenads in Euripides’ Bacchae, contrary to Pentheus’
allegation in vv. 221 f., are not drunk (see 686 f.), so Horace is not
referring to intoxication; distinguish the humorous 3. 19. 14 ff. where
the attonitus vates calls for wine.
Porph. comments ‘videtur allegoricos significare non sufficere spir-
itum suum laudibus Augusti [4–5], nisi Liberi munere (nam et ipse
musicus deus est) adiuvetur’. For Bacchic frenzy as a metaphor for the
poetic impulse see the introduction. H’s rapis recalls also the impetus of
the orator, for which see M. Winterbottom in D. Innes et al., Ethics and
Rhetoric, 1995: 313 ff.; for poetic instances cf. Ov. fast. 1. 23 with Bömer,
Pont. 4. 2. 25 f. ‘impetus ille sacer qui vatum pectora nutrit’; see also the
note on velox (3).
2. quae nemora aut quos agor in specus: one significant MS and two
citations read et instead of aut, but in such cases Latin sometimes uses
the disjunctive even when no real alternatives are offered (K–S 2. 102,
H–Sz 499 f.). For agor of Bacchic possession cf. Virg. Aen. 7. 384 (of
Amata) ‘per medias urbes agitur’ (the passive is significant). in must be
taken with nemora as well as with specus, a construction which occurs not
uncommonly in high poetry (e.g. Virg. Aen. 6. 692 f. ‘quas ego te terras
et quanta per aequora vectum / accipio’), but is not confined to that
level; cf. epist. 2. 1. 25 ‘vel Gabiis vel cum rigidis aequata Sabinis’, Catull.
33. 5 f. ‘cur non exsilium malasque in oras / itis?’; see Leo, Analecta
Plautina 1, 1891: 42 f. ¼ Kleine Schriften 1. 117 f., K–S 2. 561 f., H–Sz
835. For a similar word-order in Greek poetry cf. K–G 1. 550, Wilamo-
witz on Eur. Her. 237, G. Kiefner, Die Versparung (1964), 27 ff.
Woods could convey to the Romans some of the mysterious feelings
associated with animism; cf. Ov. fast. 3. 295 f. ‘lucus . . . quo posses viso
dicere ‘‘numen inest’’ ’, met. 3. 28 f. with Bömer, Sen. epist. 41. 3 ‘illa
proceritas silvae et secretum loci et admiratio umbrae in aperto tam
densae atque continuae fidem tibi numinis faciet’. They are sometimes
regarded as appropriate for poetic composition; cf. 1. 1. 30 with N–H,
4. 3. 10 ff., epist. 2. 2. 77 ‘scriptorum chorus omnis amat nemus et
fugit urbem’ with Brink, Juv. 7. 56 f., Tac. dial. 9. 6 with Peterson and
Gudeman. But in our passage the scenery is not localized, suggesting
rather an ideal landscape; cf. the rather gentler picture in 1. 26. 6 ff., 3. 4.
6, Prop. 3. 1. 1 f. ‘Callimachi manes et Coi sacra Philetae, / in vestrum
quaeso me sinite ire nemus’ (with Boucher 216 f.), Troxler-Keller 40 ff.,
92 ff., Kambylis 178 f.
Caves were also awe-inspiring places; cf. Sen. loc. cit. ‘si quis specus
saxis penitus exesis montem suspenderit, non manu factus sed natur-
alibus causis in tantam laxitatem excavatus, animum tuum quadam
religionis suspicione percutiet’. They were particularly associated with
the cult of Dionysus (Dodds on Eur. Bacch. 120); sometimes these were
artificial (Athen. 4. 148b, M. P. Nilsson, The Dionysiac Mysteries of the
Hellenistic and Roman Age, 1957: 61 ff.), but obviously not in the wild
terrain described here. Like the woods, they are part of the poet’s
imaginary world, cf. 2. 1. 39 with N–H, 3. 4. 40, Prop. 3. 1. 5 f. (quoted
below on vv. 3–5); Horace did not write in real caves any more than
Homer or Euripides (for the fictions about them see Paus. 7. 5. 12,
Eur. vita 62–5).
3. velox mente nova: the three words should be taken closely together
(‘sped on by a strange state of mind’). velox after agor suggests literal
fleetness of foot (as at 1. 17. 1 of Faunus), a regular attribute of Maenads
(Eur. Bacch. 169 ŒHº ¼ªØ Æ
ı ŒØæÆØ Œ
Æ, 748), which is
why they were called Thyiades; metaphorically the adjective refers to
the impetus of composition and the dithyrambic rapidity of the ode
itself. mente nova suits divine possession (cf. Sen. Ag. 720 ff. ‘quid me
furoris incitum stimulis novi / . . . rapitis?’, Lucan 5. 167 f. ‘mentemque
priorem / expulit’); in other religious contexts novus is applied to
spiritual regeneration (Paul, Romans 12: 2 ÆæFŁ fi Ð IÆŒÆØØ
F e H, Paul. Nol. carm. 10. 142 f. ‘mens nova mi, fateor, mens
non mea, non mea quondam / sed mea nunc auctore Deo’). Here the
metaphor indicates a new poetic impulse; cf. the quasi-Bacchic passage
of Lucretius (1. 922 ff.) paraphrased below (12 n.).
3–5. quibus / antris egregii Caesaris audiar / aeternum meditans

decus . . . : the three questions, increasing in length (as well as particu-
larity), exemplify a common rhetorical principle (3. 4. 67–8 n.), and here
suggest that H’s dithyramb is gaining momentum.
antris means ‘caverns’, not ‘rocky glens’ (as at Prop. 1. 1. 11, 4. 4. 3,
Manil. 5. 311); see n. 2 above on specus, especially 3. 4. 40 and Prop. 3. 1. 5 f.
‘dicite, quo pariter carmen tenuastis in antro, / quove pede ingressi?’ The
case is a local ablative, not a dative of the agent (as some have taken it);
to be heard only ‘by hollow caverns’ would imply a degree of isolation
that is not appropriate here. audiar is a confident future indicative, like
dicam (7), not a tentative present subjunctive (‘am I to be heard?’). H is
referring here only to the ode itself, not to some future composition; see
the introduction.
egregii Caesaris (i.e. Augustus) gains emphasis from its position before
aeternum . . . decus (where aeternum itself is stressed). decus implies that
the glory of Augustus will add lustre to the skies; cf. 2. 19. 13 f. of
Ariadne ‘additum / stellis honorem’ with N–H, carm. saec. 2 ‘lucidum
caeli decus’ of Phoebus and Diana, Virg. Aen. 8. 301 ‘decus addite divis’
of Hercules. meditari implies not just ‘to think of doing something’ but
actually to work it out; the word is used of practising a tune (Virg. ecl. 1. 2
‘silvestrem tenui meditaris harundine Musam’) or composing a poem, as
in serm. 1. 9. 2 ‘nescio quid meditans nugarum’, Suet. Aug. 85. 2 ‘exstat
alter epigrammatum (liber) quae fere tempore balinei meditabatur’.
Here meditans is followed by inserere (6), which governs decus, so the
meaning will be ‘practising placing etc.’; cf. Ter. ad. 896 ‘meditor esse
adfabilis’, Sen. epist. 121. 8 ‘infans qui stare meditatur’. This is preferable
to taking inserere with audiar, as ps.-Acro does.
6. stellis inserere et consilio Iovis?: the divinity of the stars was affirmed
in Platonic and later philosophy (Pease on Cic. nat. deor. 1. 30), and even
Cicero found a place there for the illustrious dead (rep. 6. 16); at a more
popular level a belief in catasterism had spread to Rome from the East
(see the summary in F. Cumont, After-Life in Roman Paganism, repr.
1959: ch. 3). After Julius Caesar’s assassination the appearance of a comet
was hailed as evidence of his apotheosis (Virg. ecl. 9. 46 ff., Suet. Jul. 88,
Weinstock 370 ff.); in the same way Virgil predicted a celestial destiny
for Octavian ( georg. 1. 32 ff. ‘anne novum tardis sidus te mensibus
addas . . . ’), in which he was imitated in later panegyrics on emperors
(Ov. met. 15. 846 ‘animam caelestibus intulit astris’, Manil. 1. 385 f.,
Lucan 1. 45 ff. with Getty’s parallels). The fantasy owes much to
Hellenistic literary convention (2. 19. 13 f. with N–H on Ariadne, Call.
aet. fr. 110 on Berenice, imitated in Catull. 66), but the underlying idea
was soon to be embodied in genuine cult: Augustus’ apotheosis, as
distinct from the worship of his genius, has to await his death (3. 3.
11–12 n.), but the event itself is already ordained, as aeternum decus
implies. inserere suggests not just inclusion (cf. 1. 1. 35 ‘quod si me lyricis
vatibus inseres’ with N–H, Tac. dial. 10. 3), but the physical insertion of
a new arrival in the starry circle (cf. 2. 5. 21 ‘quem si puellarum insereres
choro’); Virg. loc. cit. had talked of the constellations making room for
Octavian, and so also Lucan loc. cit. of Nero. H suggests elsewhere that
poetry can confer immortality (4. 8. 29 ‘caelo Musa beat’), but though
that claim is no doubt implied here (Pöschl 167), one thinks primarily of
the figure by which a poet is said to do himself what he descibes as being
done; cf. ecl. 6. 62 f. ‘tum Phaethontiadas musco circumdat amaro /
corticis’ (where Silenus does not perform the metamorphosis but simply
narrates it), Lieberg (1982), 46 ff.
consilio describes the informal group of advisers that played such a
part in Roman public and private life (see J. A. Crook, Consilium
Principis, 1955: 4 ff.). The variant concilio (also well attested) represents
the Homeric ‘council of the gods’, which was often described in later
epic or satirical writing; cf. 3. 3. 17 n. H must have remembered Virg.
georg. 1. 24 f. ‘tuque adeo quem mox quae sint habitura deorum / concilia
incertum est’, but in our passage Iovis (rather than deorum) suits consilio
better (in spite of Rutil. Namat. 1. 18 ‘concilium summi . . . dei’); it
coheres with Augustan ideology that Jupiter should take wise advice
from his amici rather than permit an oligarchic debate.
7–8. dicam insigne recens adhuc / indictum ore alio: H is describing

his own poem (as is shown particularly by recens), not the theme of
apotheosis. For neuter singular adjectives treated as nouns cf. K–S 1. 228,
H–Sz 153 f.; one would have expected the addition of aliquid in prose.
insigne means ‘something that stands out’; cf. 1. 12. 39 ‘insigni referam
Camena’ with N–H. recens means that Horace’s poem is something
new and unfamiliar (cf. PMG 851 (b) 1 ff. , ´Œ
, 1FÆ
IªºÆ. = . . . ŒÆØa IÆæŁı, h Ø ÆE æ = Œ
æÆ
fi ÆEØ . . . ), but it may also hint at the poet’s mystical experience; cf. 2.
T
19. 5 ‘euhoe recenti mens trepidat metu’. indictum ore alio makes a
stronger assertion of originality, that need not be limited to Roman
poetry (contrast epist. 1. 19. 32 f. ‘hunc (Alcaeum) ego non alio dictum
prius ore Latinus / volgavi fidicen’). For claims to originality see 1. 26. 10
with N–H, 3. 1. 2–3 n., 3. 30. 13–14 n., 4. 9. 3 f.
8–9. non secus in iugis / exsomnis stupet Euhias: Euhias is a feminine
noun constructed from the bacchic cry ‘Euhoe!’ exsomnis means ‘sleep-
less’ (‘vigilans per furorem’ ps.-Acro); cf. Virg. Aen. 6. 556 ‘vestibulum
exsomnis servat’ (v. l. insomnis), Sil. 9. 4 f. ‘consul traducere noctem /
exsomnis’, TLL 5. 2. 1880. 56 ff. Maenads notoriously roamed all night
without sleep; cf. 2. 19. 9 ‘pervicacis . . . Thyiadas’ with N–H, Eur. Ion
1077 K
Ø ¼ı of Bacchus, Nonn. 12. 397 ¼ªæı æ, 24. 348
IŒØØ
æ .
Bentley objected to exsomnis that at night the Maenad would not see
the view described below, while the next morning she is regularly
portrayed as asleep; cf. Eur. Bacch. 683, Prop. 1. 3. 5 ‘assiduis Edonis
fessa choreis’ with Fedeli, Ov. am. 1. 14. 21 f. ‘ut Threcia Bacche / cum
temere in viridi gramine lassa iacet’, culex 113, Th. Birt, RhM 50, 1895:
60 ff. It is no use arguing that exsomnis means ı ‘awakened from
sleep’ (a usage not attested elsewhere, though note CGL 2. 66. 27
exsomniat KıØ); for such an interpretation is incompatible with
the Maenad’s sleepless wanderings. In spite of the passages cited above,
the only way out is to say that at dawn the Maenad is still awake.
Bentley also maintained that iugis is feeble without an epithet (yet cf.
Virg. georg. 3. 292 ‘iuvat ire iugis . . . ’, Aen. 3. 125 ‘bacchatamque iugis
Naxon’). He therefore proposed Edonis (ablative), referring to the
Thracian tribe of Lycurgus; he cited 2. 7. 26 f. ‘non ego sanius / baccha-
bor Edonis’, Ov. trist. 4. 1. 42 ‘dum stupet Edonis exululata iugis’ (but
there the right reading is ‘Idaeis . . . modis’), Lucan 1. 674 f. ‘nam qualis
vertice Pindi / Edonis Ogygio decurrit plena Lyaeo’, Sil. 4. 776 f.
‘Edonis ut Pangaea super trieteride mota / it iuga’. Yet exsomnis coheres
so well with the common representation of the Maenads’ nocturnal
revel that it is unlikely to have arisen from palaeographical accident.
10–11. Hebrum prospiciens et nive candidam / Thracen: the Hebrus

(Maritza, Evros) is the great river of eastern Thrace (for the geography
of the country see S. Casson, Macedonia, Thrace, and Illyria, 1926: 3 ff.);
for its wintry associations cf. epist. 1. 3. 3 ‘Hebrusque nivali compede
vinctus’ (similarly Flaccus, anth. Pal. 7. 542, 1 f. 0 ¯ æı
ØæØ
. . . ŒæıEØ Ł ), 1. 16. 13, Theoc. 7. 112, Virg. ecl. 10. 65. It appears
in the story of Orpheus, who was torn to pieces by Maenads; cf. Virg.
georg. 4. 524, Ov. met. 11. 50 and 55, Milton, Lycidas 62 f. ‘His gory visage
down the stream was sent, / Down the swift Hebrus to the Lesbian
shore’. For prospiciens cf. Catull. 64. 61 (on Ariadne) ‘saxea ut effigies
bacchantis prospicit’, Ov. her. 10. 49 (again Ariadne) ‘aut mare prospi-
ciens in saxo frigida sedi / quamque lapis sedes tam lapis ipsa fui’ (for
such representations of Ariadne cf. LIMC 3. 1. 1058–60, 3. 2. 731–2); it
looks as if H knew a similar depiction of a Bacchanal surveying the
scene (but cf. P. Hardie in Rudd, 1993: 122). In these imagined vistas we
must not look for geographical precision, though in fact H knew
southern Thrace from the Philippi campaign. Lucan gives an anti-
Caesarian twist to our passage when his Bacchanal exclaims ‘quo feror,
o Paean? qua me super aethera raptam / constituis terra? video Pangaea
nivosis / cana iugis latosque Haemi sub rupe Philippos’ (1. 678 ff.).
Thrace had particular associations with Dionysus (M. P. Nilsson,
Geschichte der griechischen Religion 1, 3rd edn., 1967: 564 ff.), notably in
the legend of Lycurgus king of the Edoni (2. 19. 16 with N–H); but it is
now thought unlikely that the cult entered Greece from that quarter
(A. Henrichs, OCD 479 f., citing the appearance of the god’s name in
Linear B tablets from Pylos and Crete). The area was conventionally
covered with snow (Hom. Il. 14. 227 ¨æfi ŒH ZæÆ Ø Æ, Eur. Hec. 81,
Andr. 215), which was an eerie feature in Bacchic landscapes, cf. Dodds
on Eur. Bacch. 661 f. The wild terrain makes a particular appeal to the
Romantic imagination (note the imitation by Novalis cited by Fraenkel
and Pöschl); but the ancients’ wonder at such scenes included a large
element of dread (RE 16. 2. 1817, 1831 f., 1853, 1859 f.).
11–12. et pede barbaro / lustratam Rhodopen: et was read by Gesner’s

edition of 1772 for ac of the MSS. The sequence ‘A et B ac (or atque) C’ is
unusual, except where B and C are closely linked (Hand 2. 526) or atque
is metrically convenient, but this point alone would not justify emend-
ation; if, however, ac is read below in v. 12 (see note), it becomes
unnecessarily confusing here.
Rhodope was the great range of Thrace, rising to 7,000 ft.; for its
association with wildness and cold cf. Theoc. 7. 77, Virg. ecl. 8. 44, georg.
4. 461. By a false etymology the name may have suggested roses, and so
offered a colour contrast with nive candidam (10). lustrare primarily
meant ‘to purify’ (probably connected with luo rather than lavo), and
so was applied to religious processions (W. Warde Fowler, The Religious
Experience of the Roman People, 1911: 209 ff., Austin on Virg. Aen. 4. 6,
Fordyce on Aen. 7. 391); hence it came to mean ‘traverse’. H is referring,
not to the inhabitants of Thrace in general (as some editors suppose),
but to the Maenad and her fellow-worshippers; this is shown by the
ritual associations of lustro and the Maenads’ reputation for traversing
mountains (cf. Virg. georg. 2. 487 f. ‘virginibus bacchata Lacaenis /
Taygeta’). barbaro is not just a transferred epithet but emphasizes
the outlandish nature of the dance; the Maenads were pictured in
literature and on vases as bare-footed, to underline their escape from
the proprieties of the city (Eur. Bacch. 664 ff. with Dodds, Nonn. 14. 367
¼ºº Ææa æıÆ ŒÆ hæÆ ªıa ºø, 14. 382 IÆº,
46. 147 IØº ).
12. ac mihi devio: ac was accepted by Bentley with negligible MS
support; the MSS in general read ut ; Porph.’s lemma has quam. By the
norms of Latin syntax non secus should be followed by ac or quam (ars
148 f. ‘in medias res / non secus ac notas auditorem rapit’, K–S 2. 18 ff.).
Editors defend ut by citing 1. 16. 7 f. ‘non Liber aeque, non acuta / sic
geminant Corybantes aera / tristes ut irae’ (but there ut picks up sic, not
aeque), Prop. 1. 15. 7 f. ‘nec minus . . . ut’ (but there minus points back-
wards, and Enk glosses nec minus with atque etiam); positive sentences
involving pariter ut and perinde ut give no support to non secus ut.
Housman 1. 133 f. defends ut by comparing Ov. met. 15. 179 f. ‘assiduo
labuntur tempora motu / non secus ut flumen’ (but though a good MS
has ut, the bulk of the tradition has ac); he cites Virg. georg. 2. 277 ‘nec
setius . . . ut’ where setius in fact points backwards (see Mynors); at carm.
2. 3. 1 f., where the MSS have ‘aequam memento rebus in arduis / servare
mentem, non secus in bonis’, he reads ut bonis rather than the obvious ac
bonis which is found in two of the deteriores (see N–H). In the present
passage ut should probably be regarded as an interpolation by somebody
unfamiliar with this use of ac. If ac is restored here, the ac in v. 11
becomes not just inelegant but confusing; so et should probably be read
there (see note).
The Maenad is presented as literally in a remote place, having
abandoned the city for the mountains, but Horace has ‘left the beaten
track’ only in imagination, and in doing so he claims originality for his
poem; Porph. comments correctly ‘diximus autem haec allegoricos dici,
quia per ea intellegi vult se inusitatum Romanis carmen tractare’. If we
consider the four themes: Bacchic possession, poetic originality, moun-
tain landscape, and remoteness (devio), we may be put in mind of the
following passages (the words quoted in the original show a direct
resemblance to the ode):
Virgil (georg. 3. 287 ff.) is about to deal with sheep and goats; it is a
large task to add glory (addere honorem) to such a lowly subject, but a
sweet desire whirls him (raptat) over the lonely steeps (deserta per ardua)
of Parnassus; he is eager to tread on mountain ridges (iugis) where no
predecessor’s wheel-track winds (nulla priorum devertitur orbita); he will
speak in a resonant and elevated style (magno ore sonandum).
Lucretius (1. 922 ff.) is coping with an obscure subject, but the hope of
renown has struck his heart with a sharp wand (thyrso), inspiring him
with a love of the Muses; so he wanders through the pathless haunts of
the Pierides which no one has trodden before (avia Pieridum pera-
gro . . . loca nullius ante / trita solo).
Callimachus (aetia fr. l) shuns long works about kings and heroes,
advocating small neatly made poems (cf. h. 2. 108 ff.). Apollo bids him
avoid the highway (see Wimmel, 1960: 106 ff.). It is clear that the points
of contact become fewer the further back one goes. The debt to Virgil
seems plain enough; the echo of Lucretius is fainter; and although we
may still want to see a remnant of the Callimachean passage in the ode,
it is wiser (in view of 17 f. ‘nil parvum aut humili modo . . . loquar’) to
regard it as a transformation rather than an imitation.
13–14. rupis et vacuum nemus / mirari libet: rupis is Muretus’ conjec-

ture for the ripas of the MSS; just as nemora and specus were combined
in v. 2, so rupis and nemus are joined here (cf. also Virg. ecl. 10. 58 ‘per
rupis . . . lucosque sonantis’). For crags in the wild Bacchic landscape
cf. 2. 19. 1 f. ‘Bacchum in remotis carmina rupibus / vidi docentem’,
Eur. Bacch. 1094 IªH, 1051 ¼ªŒ IŒæ, Theoc. 26. 10
(Pentheus spies on the Bacchanals from a rock), Sil. 2. 73 f. (a pastiche
of our passage) ‘quales Threiciae Rhodopen Pangaeaque lustrant /
saxosis nemora alta iugis’, Nonn. 14. 382 f. ł ŁØ æ = æ
Æºfiø
æHØ . . . Kæ (and so often on vases). In Virgil’s ‘deserta per
ardua’ (quoted above, 12 n.) deserta corresponds to Horace’s vacuum,
which can be applied to rupis as well as nemus.
Editors generally accept ripas, which balances Hebrum (10); indeed
the word is often used where we should expect ‘rivers’ (L. Håkanson,
Statius’ Silvae, 1969: 68). ripae are part of a locus amoenus at 3. 1. 23, 3. 29.
24, but they normally suggest the gentle and peaceful aspects of nature
(flowers, trees, birds); the word is combined with nemus at 4. 2. 31 f. ‘circa
nemus uvidique / Tiburis ripas’, but that is not the kind of landscape
visualized here. Bentley proposed rivos, which again suits peaceful
contexts (3. 16. 29, 3. 29. 22, epist. 1. 10. 7, 1. 18. 104, Lucr. 2. 30); they
are often combined with groves and caves (Bömer on Ov. fast. 2. 315),
but in a symbolic landscape the word might too easily suggest the small
streams of the unpretentious poet.
In a literal sense H’s imagined grove might be called ‘empty’ because
it is lonely and remote (cf. Aetna 22 ‘desertam vacuo Minoida litore’); at
the same time H’s metaphorical grove is ‘unoccupied’ because his poetry
is original; cf. epist. 1. 19. 21 f. ‘libera per vacuum posui vestigia princeps,
/ non aliena meo pressi pede’ (the legal tinge of vacuum can also be
detected in our passage). mirari (balancing stupet in v. 9) suits awe at the
more impressive aspects of nature; cf. epist. 1. 6. 1 ff. (where admirari is
associated with wonder at the heavens), Virg. georg. 4. 363 (Aristaeus
under the river). libet means not just ‘I am willing’ but ‘I am delighted’;
cf. epod. 2. 23 ‘libet iacere modo sub antiqua ilice’.
14. o Naiadum potens: the nymphs regularly belong to Bacchus’ thiasos

(1. 1. 31, 2. 19. 3, Anacreon, PMG 357. 1 ff.), and are often represented as
Naiads; cf. Pratinas, PMG 708. 3 f. Ke Ke › ´æ Ø , Kb E
ŒºÆE, Kb E ÆÆªE = I ZæÆ a ˝Æœø. potens
(like o) suits religious language (cf. ŒæÆø etc.), with a genitive to
mark the god’s sphere of influence; cf. 1. 3. 1 ‘sic te diva potens Cypri’
with N–H. Here H is not making a petition but acknowledging the
source of his inspiration, cf. 1. 26. 9 f. ‘nil sine te mei / possunt honores’,
Pulleyn 39 ff.
15–16. Baccharumque valentium / proceras manibus vertere fraxinos:

the Bacchanals, when possessed by the god, had unnatural strength, as
shown also by their tireless dances over the mountains. valentium takes
further the idea of power suggested by potens; it is because of this divine
power that H writes nil mortale (18). vertere means ‘overturn’ (OLD
s.v. 5), like evertere (Virg. georg. 1. 256 ‘silvis evertere pinum’); for the
Bacchanals’ ability to bend and uproot trees cf. Eur. Bacch. 1064 ff. and
1109 ff. (hence Cornelissen’s vellere, for which cf. Virg. Aen. 3. 27 f.
‘ruptis radicibus arbos / vellitur’, OLD s. v. 2a). procerus is a grandiose,
perhaps somewhat archaic, word for ‘lofty’ (or ‘long’), particularly used
of trees; cf. epod. 15. 5, Enn. ann. 178 ‘proceras pinus pervortunt’ (this,
like Virg. georg. 1. 256, tells in favour of vertere here), Catull. 64. 289,
Cic. leg. 1. 15, Virg. ecl. 6. 63.
17–18. nil parvum aut humili modo, / nil mortale loquar: H is referring
to his own style, and hence by implication to his subject-matter. parvum
means ‘trivial’, the opposite of magnum as found in ars 280 (of Aes-
chylus) ‘et docuit magnumque loqui nitique cothurno’; for ªŁ in
literary criticism (involving value as well as size) see D. A. Russell and
M. Winterbottom, Ancient Literary Criticism, 1972: index under ‘Grand-
eur’. humili refers to banality and vulgarity, like ÆØ (see Russell
and Winterbottom, index under ‘Low words’). mortale means primarily
‘quod mortali conveniat’ (ps.-Acro), as at Virg. Aen. 6. 50 ‘nec mortale
sonans’ (of the inspired Sibyl); there is also an unmistakable suggestion
that H’s poetry will not die; cf. 3. 30. 6 n., 4. 9. 1, Pind. I. 4. 58 F ªaæ
IŁÆ øA (æØ. loquar balances audiar (4) and dicam (7) and
need refer to nothing outside the present poem. It may seem to come
late in the poem for a statement of this kind, but cf. Pind. I. 4. 90b.
18. dulce periculum est: Bacchic ecstasy is both thrilling and dangerous,
not because of snowfields and precipices (Fraenkel), but because meet-
ing a god and submitting one’s mind to him is a terrifying experience;
cf. 2. 19. 5 ‘euhoe, recenti mens trepidat metu’ with N–H, Lucr. 3. 28 f.
‘his ibi me rebus quaedam divina voluptas / percipit atque horror’ (see
Bailey on this quasi-religious sensation). The oxymoron particularly
suits the double nature of Dionysus, who was both genial and pitiless
(N–H on 2. 19. 27, C. Segal, Dionysiac Poetics and Euripides’ Bacchae,
2nd edn., 1997: 350 f.), cf. 2. 19. 6 f. ‘turbidum / laetatur’ with N–H,
Eur. Bacch. 66 f = ŒÆ PŒÆ, Antip. Thess. anth.
Pal. 9. 186. 3 f. (of Aristophanes) M ‹ ˜Ø ı
Ø º , x Æ b
FŁØ = M
FØ, æH ºŁ Ø
Ææø, anth. Plan. 290. 3
æe (of a pantomimus playing Bacchus). In the same way the
ambitious poet experiences dangers as well as delights; cf. 4. 2. 1 ff. (the
imitator of Pindar risks the fate of Icarus), epist. 2. 1. 210 f. (of the
dramatic poet) ‘ille per extentum funem mihi posse videtur / ire
poeta’, ars 10 (on poetic audacia) with Brink, Stat. silv. 4. 5. 25 (repeating
dulce periculum), Kroll 42 n. 43.
19. o Lenaee, sequi deum: ¸BÆØ was a word for Bacchanals (see Gow’s
preface to Theoc. 26), and the Lenaea was Dionysus’ winter festival at
Athens (A. Pickard-Cambridge, The Dramatic Festivals of Athens, edn. 2,
revised by J. P. A. Gould and D. M. Lewis, 1968: 25 ff.). The cult-title
¸ÆE is rare in extant Greek (Alcaeus of Messene, anth. Pal. 9. 519. 1,
Orph. hym. 50. 5, 52. 2, Bömer on Ov. met. 4. 14), but is common in
Augustan poetry. As it was thought, probably wrongly, to be derived
from º a wine vat (Diod. 4. 5. 1 etc.), it suits contexts that deal with
the vine and drinking (Alcaeus of Messene, loc. cit., Virg. georg. 2. 4 ff.
‘huc, pater o Lenaee: tuis hic omnia plena / muneribus, tibi pampineo
gravidus autumno / floret ager’). In the same way it coheres in H with
pampino (20).
Bentley objected that the sentence would amount to no more than
‘dulce est, o Bacche, sequi Bacchum’; he therefore suggested, but did not
print, ‘te, Lenaee, sequi ducem’. But the transmitted reading is less
tautological than he suggests. deum implies the superhuman power of
Bacchus: ‘it is a dulce periculum,’ says the poet, ‘to follow your divinity’.
20. cingentem viridi tempora pampino: after some hesitation we have

concluded that cingentem refers not to the god but the poet (thus
Heinze); it is on him that our attention should be fixed, particularly at
the end of the ode (cf. 3. 30. 15 f.). For the accusative and infinitive
cf. Plaut. mil. 68 ‘nimiast miseria nimi’ pulchrum esse hominem’, K–S
1. 695, H–Sz 358 f. As he joins the thiasus Horace wreathes himself as a
sign of his new commitment; normally the vine-shoots were worn by
the god rather than his votaries, but cf. Ov. met. 6. 592 ‘vite caput
tegitur’ (of Procne). As the colometry of the passage suggests a break
at the end of v. 19, the dissociation of cingentem from deum seems
acceptable.
Some commentators think that cingentem must be taken with deum. It
cannot be a standing epithet (‘the god who wreathes his temples’); that
is possible at 1. 32. 9 f. ‘Veneremque et illi / semper haerentem puerum’,
but if it were the construction in our passage, the perfect cinctum would
be required (cf. 4. 8. 33 of Liber ‘ornatus viridi tempora pampino’). The
meaning would have to be ‘while he wreathes his temples’; cf. Sen. Med.
70 ‘praecingens roseo tempora vinculo’, where Hymen is preparing for
2 6 . V I X I P V ELLI S 309
a particular wedding. But here it would be odd for the god to do the
wreathing during the revel-rout.
Syndikus (2. 214 n.) now thinks that the god is wreathing the poet’s
brow (Porph. had already suggested that he is wreathing Horace as well
as himself ); this view was firmly rejected by Bentley, for then mea or
mihi would have to be expressed; cf. 3. 30. 15 f. ‘et mihi Delphica / lauro
cinge volens, Melpomene, comam’. It may also be doubted whether the
revel-rout is the right occasion for the formal coronation of the poet.
26. VIXI PVELLIS

[C. P. Jones, HSCP 75, 1971: 81 ff.; Pasquali 498 ff.; T. C. W. Stinton, Phoenix 31, 1977:
164 ff.; Williams 206 ff.]
1–8. I was until lately successful in the warfare of love, but now I am
dedicating my lyre and weapons; hang them on the wall that protects Venus’
left side. 9–12. Goddess of Cyprus and Memphis, let the disdainful Chloe feel
a stroke of your whip.
In this short ode Horace combines and refabricates themes from

several types of epigram, adapting them all to his own purported
situation. By opening with ‘vixi puellis nuper idoneus’ he is parodying
the language of sepulchral epitaphs; see CLE 106. 1 ‘vixi beatus dis
amicis literis’, 381. 1 ‘vixi viro cara custosque fidelis’, 1869. 15 ‘Felix
vocatus, felix vixit cum suis’. The following militavi would suit a
soldier’s gravestone; cf. ILS 2030 ‘vixit ann. XVIII, militavit ann. II’
(cited with other instances by R. Merkelbach, ZPE 17, 1975: 140). When
Horace goes on to speak of his lyre (3 f.), he is alluding to the lover’s
serenade (paraclausithyron), a stock motif of amatory epigram as of
other genres (see 3. 7. 29 f. and the introduction to 3. 10).
Horace also draws on dedicatory epigram, a common type in life and
literature; see anth. Pal. book 6, W. H. D. Rouse Greek Votive Offerings,
1902, H. Kühn, Topica epigrammatum dedicatoriorum Graecorum, Diss.
Vratislavia 1906, H. Roth, Untersuchungen über die lateinischen Weihge-
dichte auf Stein, Diss. Giessen 1935. A victorious soldier might dedicate
his own weapons as well as his enemy’s (anth. Pal. 6. 127–32 etc., Rouse
111 ff., Roth 38 ff., cf. epist. 1. 1. 4 f. of a gladiator); Horace exploits this
motif through the traditional topic of the militia amoris (2 n.). Again,
dedications on a craftsman’s retirement, attested sometimes from real
life (Rouse 366), were a favourite topic in the Anthology (Gutzwiller 92),
where minor implements are catalogued with pedantic realism; thus an
ageing courtesan might dedicate her unwanted mirror to Aphrodite
(Plato, anth. Pal. 6. 1. 3, cf. 6. 210, 211). Here Horace applies a similar
motif to himself with a blend of humour and wistfulness that seems
entirely his own: he pretends that he used to break into his girl-friends’
apartments like the aggressive rakes of comedy (7–8 n.), but now he is
feeling his age and is hanging up his crowbars.
Such a renuntiatio amoris was another familiar topic of Greek epi-
grammatists and their Latin successors (anth. Pal. 5. 179, Tib. 1. 9. 83 f.,
Ov. am. 3. 11. 32 ‘non ego sum stultus ut ante fui’; see further the
catalogue in Cairns 79 ff.); but instead of voicing bitter reproaches
Horace shows a wry amusement at his own situation. For his philo-
sophical attitude to changed circumstances cf. Philodemus, anth. Pal.
5. 112. 5 f. (no. 5 in D. Sider’s edition) ŒÆd ÆØ ‹ ŒÆØæ , KØÆ:
ŒÆ ŒÆd F = PŒØ, ºøœæ æ ±ł ŁÆ (imitated at epist.
1. 14. 32 ff.). It is worth noting that in his renunciation of love Propertius
dedicates himself to Good Sense (3. 24. 19 f.): ‘Mens Bona, si qua dea es,
tua me in sacraria dono; / exciderant surdo tot mea vota Iovi’. RN thinks
that in the same way Horace may be making his dedication not to
Venus, as is generally assumed, but on the neighbouring wall of the
shrine of Good Sense (see below, 5 n.).
Horace’s humour is shown again in the grandiloquent prayer to
Venus in the last stanza. We expect that his address will lead to a
request for future peace; in fact by a witty surprise he asks for the
haughty Chloe to be struck with the goddess’s whip. Venus’s whip is
intended both to punish Chloe for her past disdain and (like Cupid’s
arrows) to make her fall in love; for love’s blend of pain and pleasure cf.
Mart. 6. 21. 9 f. (on the marriage of Stella after some delinquencies):
‘dixit (Venus) et arcano percussit pectora loro. / plaga iuvat: sed tu iam,
dea, caede duos’; that is to say, the stroke is not just a punishment but a
stimulus, which is why the goddess is invited to strike the bride as well.
According to the usual interpretation Horace hopes that Chloe will
after all be smitten with love for himself (see especially Stinton, loc.
cit.). He thus acknowledges that his renuntiatio amoris is a sham; for
such mockery of his own backsliding cf. 2. 4. 22 ff. ‘fuge suspicari / cuius
octavum trepidavit aetas / claudere lustrum’ (where the suspicion is
meant to be well founded), and especially 4. 1. 33 f. (after he has diverted
Venus to a more suitable target) ‘sed cur, heu Ligurine, cur / manat
rara meas lacrima per genas?’ For a final volte-face in other contexts add
2. 1. 37 ff., 3. 3. 69 ff.
There is, however, a different interpretation, to which we now in-
cline: Horace is begging Venus to give Chloe a taste of her own
medicine by making her love, not himself (for that is not specified)
but some superciliously indifferent man (thus Jones, loc. cit.). For the
thought cf. Theoc. 7. 118 (to the Erotes) ‘Strike with your arrows the
lovely Philinus, for he has no pity for my friend’ (i.e. make Philinus
2 6 . V I X I P V ELLI S 311
the victim of an unrequited passion for somebody else, as explained by
Gow), Ov. met. 3. 405 (addressed to Nemesis by a lover who has been
spurned by Narcissus) ‘sic amet ipse licet, sic non potiatur amato’. The
view that this is a prayer for revenge is supported by arrogantem in the
emphatic last position and by the severe punishment implied by sublimi
flagello (11–12 n.); for if Chloe fell in love with the poet, he would surely
respond and there would be no punishment at all. Stinton objects
(166 f.) that prayers for revenge should, as in Ovid’s passage, be ad-
dressed to Nemesis or Dike, whereas Venus should be asked for love.
But in a real sense Venus is being asked to give love (unrequited love
to Chloe). That, and not a miserable old age (as in 1. 25) constitutes
the revenge.
Here another complication must be considered. When he dedicates
his barbitos towards the end of the collection, Horace is renouncing
not only love but also love poetry (cf. Porph. ad loc. ‘hac T fi fi Ð videtur
iam renuntiare carminibus’, Heinze 361). In doing so he repeats themes
that appeared in the first book: e.g. the dedication on Venus’ temple
wall (1. 5. 13 ff.), the slaves carrying offerings to Venus (1. 19. 13 ff.
quoted below on v. 6), the unresponsive Chloe (1. 23), the assault on
the hetaera’s house (1. 25. 1 f.), the grandiose prayer to the Queen
of Cyprus (1. 30. 1). If the traditional view of our ode is correct, that
ought to involve a volte-face in the area of love-poetry as well as in love.
One would have to argue that such a return takes place in the final
stanza: in still hoping for Chloe’s love he turns out to have written
another love-poem. On the other hand, with Jones’s interpretation the
renunciation is not revoked within the poem, but we do not have to wait
until 4. 1 to see this happen. For, whatever the chronology of their
composition, with the odes in their present order we shortly come to
3. 28, which in part at least can be regarded as a love-poem (thus
Syndikus 2. 219).
Metre: Alcaic (in itself a sign that Horace has created something more
complicated than the epigrams on which he draws).
1. Vixi puellis nuper idoneus: ‘I lived of late fit for the girls’; for the
parody of epitaphs see the introduction above. The predicative adjective
describes a life-style; cf. Sen. epist. 88. 37 ‘(quaeritur) libidinosior Anac-
reon an ebriosior vixerit’, OLD s.v. 9; distinguish the places where vivere
means no more than esse (Catull. 10. 33 f. ‘sed tu insulsa male ac molesta
vivis’, OLD s.v. 2). Horace is not saying ‘I have truly lived’ (as at 3. 29. 43),
nor yet ‘I have completed my life’ (as at Virg. Aen. 4. 653), for such
usages are incompatible with an adjective; and if one puts a comma after
vixi (thus Campbell, followed by Shackleton Bailey), the isolation of
the verb spoils the contrast with 3 ‘nunc . . . habebit’. vixi is an aorist,
not a true perfect (which would require adhuc rather than nuper);
cf. W. Peek, Griechische Vers-Inschriften 1113a (appendix), 6 f. I
ø
(or - ) KøÆ a ºø Œb ıªH.
For puellis C. Franke proposed duellis, which found favour with
Housman (reported by L. P. Wilkinson, JRS 60, 1970: 256). They
understood by it not, of course, real wars but the militia amoris (2 n.),
in which case H’s erotic meaning would first emerge at barbiton (4) and
Veneris (5). But duellis is too archaic to combine well with idoneus, and
the variation of form at bello (3) seems impossibly awkward. idoneus is a
prosaic word (Axelson 105 f.), often used in military contexts (2. 19. 26,
serm. 2. 2. 111, Prop. 1. 6. 29 ‘non ego sum laudi, non natus idoneus
armis’); it is therefore appropriate to the militia amoris and should be
translated as ‘fit’ (with a suggestion of sexual vigour). The rare personal
dative, not attested before our passage (cf. Quint. inst. 2. 3. 1 ‘etiam cum
idoneos rhetori pueros putaverunt’, TLL 7. 1. 231. 23 ff.), suggests that
puellis had a surprise effect.
Contrasts between ‘then’ and ‘now’ are common in many types of
epigram (3. 11. 5–6 n.). In particular, sepulchral epitaphs contrast death
with recent activity (see Lattimore 172 ff.); nuper, ‘only the other day’,
suits the unexpectedly deceased.
2. et militavi non sine gloria: the theme of love as a campaign, occa-

sionally found in earlier Greek poetry, became a commonplace of
comedy, epigram, and novel. The militaristic Romans developed the
parallel with gusto; cf. Lucil. 1323M. ‘vicimus, o socii, et magnam
pugnavimus pugnam’ (‘pugnam pro stupro’ Donatus), Tib. 1. 10. 53
with Murgatroyd, Prop. 2. 7 with M. R. Gale, JRS 87, 1997: 77 ff.; see
especially Ovid’s extended double-entendre at am. 1. 9, beginning ‘mili-
tat omnis amans’ (‘every lover is on active service’) and continuing with
references to hardships, wounds, sleeplessness, stratagems, night-
attacks, close combat, triumph, and spoils; see further McKeown ad
loc., A. Spies, Militat omnis amans, Diss. Tübingen 1930, P. Murgatroyd,
Latomus 34, 1975: 59 ff., Lyne (1980), 71 ff. H gives new life to the
metaphor by enumerating the actual implements of his siege-warfare
(7). For gloria of sexual success cf. Ov. am 2. 12. 11 f. with McKeown; for
non sine cf. 3. 7. 7–8 n.
3–4. nunc arma defunctumque bello / barbiton hic paries habebit: for
habere of a dedication cf. Virg. Aen. 10. 423 ‘haec arma exuviasque viri
tua quercus habebit’. arma at first seems metaphorical, then turns out to
refer to literal implements; it is convenient for H’s joke that the word
arma can comprise other than military equipment (OLD s.v. 10).
defuncta bello must be understood with arma; for the word-order cf.
1. 5. 5 f. ‘fidem mutatosque deos’.
2 6 . V I X I P V ELLI S 313
The Greek barbitos is a poetical word for ‘lyre’ (N–H on 1. 32. 4); so
here it is the instrument not just of the serenading lover but of the
Graecizing love-poet. defunctum suits the military metaphor as it sug-
gests a discharged soldier; as it can also mean ‘dead’, it may allow a
contrast with vixi (1). As the lyre was conventionally imbellis (1. 6. 10
with N–H), there is point in the collocation with bello.
5–6. laevum marinae qui Veneris latus / custodit: laevum is given

emphasis by the long hyperbaton, but it has not been explained why
H is so particular about the placing of his offerings; in real life dedica-
tions were made all over a shrine and its precincts (D–S 2. 378 f.), and
literary epigrams do not specify a position with any exactitude (anth.
Pal. 6. 52. 1, 6. 114. 2, 6. 124. 2, Kühn, op. cit. in introduction, 22).
Headlam compared our passage at Herodas 4. 19 f. KŒ ØB e
ÆŒÆ, ˚ŒŒº, B = B , Y ªØ , but there the tablet is placed
on the goddess’s right because Asclepius is on her left (4. 4 f.).
As was indicated in the introduction, RN would like to suggest that
the wall belongs to the aedes of Mens, which (as every Roman would
have known) stood on the Capitol next to Venus Erycina; the two
shrines were vowed (217 bc) and dedicated (215 bc) at the same time
(R. Schilling, La Religion romaine de Vénus, 1954: 250 ff.), and were
separated only by a gutter (Liv. 23. 31. 9 ‘utraque in Capitolio est canali
una discreta’, Steinby 3. 240 f.). latus custodit suggests the action of a
faithful adherent who protects his superior’s left, or unguarded, side
(serm. 2. 5. 18 ‘utne tegam spurco Damae latus?’); the shrine of Venus was
the more important of the two and was the responsibility of the consul
Fabius (Liv. loc. cit.). Philodemus in his similar renunciation of love
(anth. Pal. 5. 112. 6) hoped for ‘a better mind’; Propertius actually
dedicated himself in the sacraria of Mens Bona (3. 24. 19); and Ovid
(am. 1. 2. 31) implies that Amor cannot expect to hold a triumph until
Mens Bona has been manacled.
NR thinks that laevum is sufficiently justified by the conceit that the
wall, like a shield, protects Venus’ exposed flank, thus continuing the
military imagery; as Heinze and Plessis point out, there are other cases
where H uses a precise word largely for the sake of immediacy, e.g. 1. 11.
6, 16. 4, 33. 7. Here Horace’s concluding prayer to Venus (9 ff.) would
seem strange if the dedication had been made to Mens Bona, which is
not mentioned. But see further 9–10 n.
marina suits Venus as a goddess of the sea, born from the foam; for
the familiar association cf. N–H on 1. 5. 16, Hes. theog. 195 ff., LIMC
2. 1. 114 ff., 2. 2. 117 ff., CMA 1. 144 ff. (for the period since the Renais-
sance). marina is not attested as an epithet of the Capitoline Erycina,
but the goddess of the Sicilian Eryx on the north-west coast may have
been worshipped as Aphrodite Pelagia (cf. A. Schulten, Klio 23, 1929:
424, K. Galinsky, Aeneas, Sicily, and Rome, 1969: 74 ff.). Some authorities
believe that the Aphrodite Anadyomene of the Ludovisi throne (where
she is portrayed rising from the sea) belonged to the Venus Erycina of
the Colline Gate, near which the sculpture was found (Schilling, op. cit.
256 ff., Galinsky, op. cit. 243 ff.). H’s dedication on the sacred wall
balances the one at the end of the Pyrrha ode (1. 5. 13 ff. ‘me tabula
sacer / votiva paries indicat uvida / suspendisse potenti / vestimenta
maris deo’), where we both take the divinity as Venus, but RN now
retains deo (1995: 426 f.); that poem was fifth from the beginning of the
three-book collection, this is fifth from the end.
6–7. hic, hic ponite lurida / funalia: Horace has attendants to fetch and
carry, and presents economically a scene reminiscent of comedy (cf.
Plaut. Curc. 1 ff., Ter. eun. 774). One recalls his very different description
of a sacrifice to Venus at 1. 19. 13 f.: ‘hic vivum mihi caespitem, hic /
verbenas, pueri, ponite, turaque’. For the geminatio of adverbs of place
cf. Wills 110 f.; H sounds an urgent note now that he has found the right
spot for his offerings. For ponere of dedications (ØŁÆØ) cf. OLD s.v. 8c.
funalia were torches made of rope coated with wax (Val. Max. 3. 6. 4
‘funalem cereum’, Blümner, Techn. 2. 224 ff.). As the Romans had no
street lamps, such lights were associated with late-night roisterers, but
H was thinking of more than that (as vectes shows): torches could be
used for burning down a woman’s door, or perhaps destroying the lintel
in which the pivot was set (Headlam on Herodas 2. 65); cf. Theoc.
2. 127 f. N ¼ººfi Æ TŁE ŒÆd ± ŁæÆ Y

ºfiH, = ø ŒÆ
ºŒØ ŒÆd ºÆ qŁ K Æ (with Gow’s note); add Ar. Lys.
249, Men. dysc. 60, Plaut. Pers. 569, Turpil. CRF 200, Ov. am. 1. 6. 57 f.
with McKeown.
lurida, ‘yellowing’, is RN’s conjecture for lucida of the MSS; this
makes the torches parallel to the implements and the lyre which are
now no longer useful (3 f.). lucida is regarded as a conventional epithet
that continues to be used when no longer applicable (cf. 1. 35. 20
‘liquidumque plumbum’, Hom. Od. 6. 74 KŁ~ Æ ÆØ of clothes for
the wash, Kroll 277), but it is particularly inappropriate in this passage
with its contrast between past and present. lurida (a negligible alter-
ation) suits the wax of old torches that are now discoloured with age and
smoke; cf. 4. 13. 10 f. ‘luridi / dentes’, epod. 17. 22 ‘pelle . . . lurida’, Stat.
silv. 4. 9. 25 ‘vel mantelia luridaeve mappae’, Fulg. myth. 1 praef. p. 6. 9
‘fumo lurida parietibus aratra pendebant’, André (1949), 137 f. The
relatively rare lurida was exposed to corruption (TLL 7. 2. 1861. 80 ff.);
the same was true a fortiori of luror and lurere; cf. epod. 17. 33 ‘virens in
Aetna flamma’ (lurens Onians), and Ov. met. 1. 239 (of a homicide
turned into a wolf ) ‘idem oculi lucent, eadem feritatis imago est’
2 6 . V I X I P V ELLI S 315
(where Housman, Classical Papers 1. 162, proposes lurent and provides
parallels for the corruption).
7–8. et vectes yet arcusy / oppositis foribus minacis: vectes were crow-
bars for splitting open the folding doors ( fores); cf. Ter. eun. 774 ‘in
medium huc agmen cum vecti, Donax’, Lucil. 839M cited below, Fest.
519L ¼ 378M ‘vecticularia vita dicitur eorum qui vectibus parietes
alienos perfodiunt furandi gratia’. For such assaults on houses in the
comissatio cf. 1. 25. 2, Headlam on Herodas 2. 34 f., K. F. Smith on Tib.
1. 1. 73, McKeown on Ov. am. 1. 6. 57 ff., Bompaire 325, Copley 40 ff.,
57 f., 148 n. 26.
et arcus seems to be corrupt; bows and arrows are quite unconvincing,
even if they are supposed to be aimed at the doorkeeper. G. P. Bidder
( J. Phil. 35, 1920: 113 ff.) suggested that the word means ‘bowdrills’
(Blümner, Techn. 2. 224 and 226); he cited the Italian archetto and the
French archet, though there is no parallel in Latin. Iæ , the Greek
equivalent, occurs in dedicatory epigrams describing carpenters’ tools
(Leonidas, anth. Pal. 6. 205. 5 with Gow–Page, HE 1996, Philippus,
ibid. 6. 103. 1 f.); but such an implement is unattested in the present kind
of context, and seems altogether too mechanical for the ardent lover.
Some have thought that arcus refers symbolically to Cupid’s bow; but
such an object would be out of place with funalia and vectes, the plural
would be awkward, and Cupid could not be mentioned in such a
condensed and casual way.
Bentley’s conjecture securesque, which he did not print, may well be
right (cf. Theoc. 2. 128 ºŒØ ŒÆd ºÆ , Plaut. Bacch. 1119 ‘nisi
mavoltis fores et postes comminui securibus’, Lucil. 839M. ‘vecte atque
ancipiti ferro effringam cardines’, Athen. 585a); the elision of the hyper-
metric syllable is paralleled (2. 3. 27, 3. 29. 35), but if a copyist was
puzzled enough to omit it, rewriting would inevitably follow. Housman
objected that after vectes another implement was unnecessary (Classical
Papers 1. 3 f.), but an accumulation of accusatives is common in Greek
dedicatory epigrams and is found in the corresponding sacrifice to
Venus at 1. 19. 13 ff. (quoted above on 6); his own tentative sacrate is
much more otiose. Another approach would be to give vectes an epithet;
Giangrande proposed aduncos (Eranos 64, 1966: 82 ff.), and RN has
considered retusos to balance his lurida as well as defunctum (3); cf.
1. 35. 39 f. ‘retusum . . . ferrum’. vectes can have two asyndetic epithets
if they are in separate clauses (cf. 3. 1. 7 f. ‘clari Giganteo triumpho, /
cuncta supercilio moventis’).
9. o quae beatam, diva, tenes Cyprum: dedicatory inscriptions fre-

quently end with prayers (see anth. Pal. 6 passim). Marks of the sacral
style are o, the relative clause (N–H vol. 1, p. 127), diva (3. 22. 4 n.), which
should be enclosed in commas as a vocative, tenes (¼
Ø , cf. 3. 4. 62),
the alternative cult-centres (N–H vol. 1, p. 343). For Cyprum, which
suggests the familiar ˚æØ , cf. 1. 3. 1 and 1. 30. 1 with N–H; beatam
alludes to ‘Macaria’, an old name for the island (Plin. nat. hist. 5. 129).
The inflated periphrasis builds up suspense before the surprise ending;
for similar grandiloquence in a mock-supplication to Venus cf. Catull.
36. 11 ff. ‘nunc o caeruleo creata ponto, / quae sanctum Idalium Vriosque
apertos, / quaeque Ancona Cnidumque harundinosam / colis . . . ’. In
view of our argument with regard to Mens, it is worth remarking that,
while Catullus addresses his prayer to Venus, his dedicatory offering is
actually made to Vulcan; both deities, however, are explicitly present.
9–10. et / Memphin carentem Sithonia nive: for the temple of

Aphrodite at Memphis (near the modern Cairo) cf. Herod. 2. 112. 2
(but that may refer to the Phoenician Astarte), more relevantly Strab.
17. 1. 31. The heat of Memphis makes a symbolical contrast with the
Thracian snow; cf. Byron, Don Juan c. 63 ‘What men call gallantry, and
gods adultery, / Is much more common where the climate’s sultry’.
R. Reitzenstein suggested a reference to Isis (NJA 21, 1908: 93 f. ¼
Aufsätze zu Horaz, 1963: 13 f.), but she had no special connection with
Cyprus; though the two goddesses were sometimes identified, the cults
at Memphis were distinct (RE 15. 682), and there is no need to invoke
such syncretism here.
Egypt conspicuously lacks snow (Herod. 2. 22. 3, Sen. nat. 4. 2. 18, RE
1. 987). H’s inflated phrase imitates expressions in Greek poetry where
the most irrelevant qualities are sometimes negated; cf. in particular
Bacchyl. fr. 30. 1. a I
Æ 1Ø (for the use of carentem to
represent an alpha privative cf. N–H on 2. 8. 11 f.). Sithonia is the middle
of the three peninsulas of Chalcidice; Sithonian snow, presumably a
Hellenistic epithet (cf. Virg. ecl. 10. 66, Ov. am. 3. 7. 8), represents
Chloe’s frigidity; cf. Philodemus, anth. Pal. 10. 21. 4 (of his loveless
state) e
Ø Ø łı
c ˚ºØ Ø . Some editors compare 3. 9. 9
‘me nunc Thressa Chloe regit’; but women in H’s odes do not always
have constant characteristics, and it is not certain that the earlier passage
is meant to be remembered here.
11–12. regina, sublimi flagello / tange Chloen semel arrogantem: regina

applied to Venus is an appropriate poeticism (1. 30. 1, Prop. 4. 5. 65) but
not a regular cult-title (Bömer on Ov. fast. 6. 37). The second vocative
after diva is a mark of the hymnal style; cf. 2. 19. 8 with N–H,
Philodemus, anth. Pal. 11. 41. 7 f. ŒæøÆ ªæłÆ, 1FÆØ, = Æ
æ , Ø , Æ . A flagellum was a severe instrument (serm.
1. 3. 119 ‘horribili . . . flagello’), and sublimi underlines the seriousness of
2 7 . I M P I O S PA R R A E R E C I N EN T I S O M EN 317
the threat; after that, tange semel suggests by way of litotes that a single
stroke will be enough (tange does not necessarily imply a light blow as at
Stat. silv. 5. 4. 18 ‘extremo me tange cacumine virgae’). The name Chloe
hints at greenness and immaturity (1. 23. 1 with N–H), and arrogantem
sounds like a jaundiced complaint against a young girl who had rejected
an older man’s advances. For the interpretation of v. 12 see the discussion
in the introduction.
2 7 . I M P I O S PA R R A E R E C I N EN T I S
OMEN
[T. Berres, Hermes 102, 1974: 58 ff.; A. Bradshaw, Hermes 106, 1978: 165 ff.; K. Büchner,
Gnomon 14, 1938: 636 ff.; Cairns 189 ff.; J. S. Clay, CJ 88, 1992–3: 167 ff.; Fraenkel 192 ff.;
W.-H. Friedrich, NGG 1959: 5, 81 ff.; S. J. Harrison, Hermes 116, 1988: 427 ff. and Fond.
Hardt, Entretiens 39, 1993: 148 ff.; R. S. Kilpatrick, Grazer Beiträge 3, 1975: 191 ff.; C. W.
Macleod, CQ 24, 1974: 88 ff. ¼ Collected Essays, 1983: 165 ff.; K. Quinn, Latin Explorations,
1963: 253 ff.]
1–16. Let evil omens attend the journeys of the impious: I shall seek
favourable auguries for my friend. May you be happy, Galatea, wherever
you prefer, and may there be no ill omens to keep you from going. 17–24. But a
storm is rising, and I know the dangers of the Adriatic. I should wish such
anxiety only on my enemies’ families. 25–32. Europa did not think of these
dangers when she trusted the treacherous bull. Soon, instead of picking
flowers, she saw nothing except stars and waves. 33–44. On reaching
Crete, she exclaimed ‘Alas for the father I have deserted in my madness!
Am I really awake, or is it all a bad dream? 45–56. I should like to break the
horns of the monster I loved. For my shameful behaviour may I wander
naked among lions, and provide a meal for tigers. 57–66. ‘‘Wretched Europa,’’
I imagine my father saying, ‘‘Why don’t you hang yourself—unless you prefer
the indignities of slavery?’’ ’ 66–76. Venus appeared and said ‘Cease your
angry complaints. Behave like Jove’s consort. Learn to accept a great destiny:
your name will be borne by a continent.’
The first third of the ode (1–24) is a propempticon, or sending-off

poem, addressed to a woman who is given the name ‘Galatea’ (14 n.).
The stock themes are found in Greek poetry of all periods, usually as
incidental elements in larger compositions: cf. Hom. Od. 5. 204 f. (13 n.),
Sapph. 94. 7 f. L–P (14 n.), Call. fr. 400, Theoc. 7. 52 ff., Lightfoot on
Parthenius p. 40 n. 108. Perhaps Parthenius was the first to write a self-
contained verse propempticon (see Lightfoot on frr. 26 and 36); for his
Roman successors cf. Cinna, FLP 1–4 with Courtney, Prop. 1. 8 with
Quinn, op. cit. 242 ff., Hor. epod. 10 (a parody), carm. 1. 3 with N–H,
Ov. am. 2. 11 with McKeown, Stat. silv. 3. 2, Paulinus of Nola, carm.
17. These developments must have been influenced by the prose
propempticon, a standard form in epideictic oratory and later in rhet-
orical exercises, that was often marked by sentimentality and flowery
language (cf. Men. Rhet. 2. 395–9 with Russell). See further Cairns
50 ff., 115 ff., 190 ff., A. Kerkhecker on Call. iamb. 6, pp. 171 ff., DNP 10.
414 f.
Horace’s version shows some of the conventional features. It was
usual to hope for favourable conditions for the journey (1. 3. 1 ff.,
Theog. 691 f., Prop. 1. 8. 17 ff., Ov. am. 2. 11. 34); here Horace promises
to conjure up desirable omens (7b–12). Like Prop. 1. 8, the poem is
addressed to a woman, so a sentimental interest should be understood;
but as no return is in prospect the natural assumption is that she is
leaving the poet for another man. When Horace hopes that Galatea will
be happy and remember him, his acquiescence in her departure reflects a
pattern that goes back to Sappho (13–14 nn.). At the same time he hints
at a reluctance to see her go, implied by ‘ubicumque mavis’ in v. 13; such
regret also belongs to the conventional pattern, even if here the
reproaches (
ºØÆ ) are less explicit than normal (Cairns 190).
But instead of praying for calm seas and mild Zephyrs, and looking
forward to Galatea’s safe return, Horace shows his continuing disquiet
by drawing attention to the storm that is supposed to be raging (17–24);
these menacing lines lead into the ordeal of Europa as described in the
longer section of the ode (25–76).
According to the myth Europa was carried from Phoenicia to Crete
by Zeus in the form of a bull (cf. Apollod. bibl. 3. 1. 1 with Frazer, Pease
on Cic. nat. deor. 1. 78, DNP 4. 293 f.). The story suits Crete very well
(cf. the bulls of Minoan art as well as the Minotaur), but it apparently
had a Phoenician origin; cf. Lucian, de Syria dea 4 (coins of Sidon show
Europa riding the bull) and Ach. Tat. 1. 1 (a realistic painting at Sidon).
It was told by Hesiod (frr. 140–1 M–W), Stesichorus (PMG 195),
Bacchylides (fr. 10), and Aeschylus (TrGF 3. 99 Radt, Loeb edn. vol.
2. 415 f., 601 f.). In Herodotus (1. 2) Europa was carried off by Greeks—
an event leading to the rivalry of Europe and Asia that culminated in
the Persian Wars (RE 6. 1298 ff.); this rationalization was repeated in
Lycophron, Alex. 1296 ff., where the vessel was shaped like a bull. In the
second century bc Moschus wrote a narrative ‘epyllion’ on the subject
(ed. W. Bühler, 1960, and M. Campbell, 1991), which must have been
known to Horace; for later accounts see Ovid, met. 2. 836ff., fast. 5.
603 ff., Lucian, dial. mar. 15. As suited a picturesque story, there were
numerous artistic representations in antiquity; see LIMC 4. 1. 76 ff. (over
200 items) and the illustrations in 4. 2. 32 ff.; W. Bühler, RAC 6, 1986:
964 ff.; Steinby 4. 121 f. and Richardson 313 (the porticus Europae in the
Campus Martius as described by Martial 2. 14. 3 etc.); for more modern
examples see CMA 1. 421–9.
The epyllion of Moschus, mentioned above, will serve as a reference-
point for the ode. Europa dreams that Asia and another figure are
struggling to possess her (1–27). She goes to gather flowers (28–36).
Her golden basket is described, with its significant representation of
Zeus and Io (37–62). The narrative resumes with the girls picking
flowers (63–71). Zeus changes into a bull (72–88). He courts Europa
and carries her off (89–114). The sea is calm and sunny (115–30). Europa
speaks to the bull, aware that he is a god (131–51). Zeus gives a comfort-
ing reply (152–61). In Crete he resumes his shape, and intercourse takes
place, with the assurance of royal sons (162–6). Here then is a straight-
forward narrative in a predominantly happy mood, where the main
stages of the action are given more or less equal attention. In the ode,
however, there is no dream; the naming of the continent is switched to
the end; the flower-picking is drastically reduced (29–30, 43–4). The
ecphrasis on the basket, characteristic of an epyllion, is omitted; the
same is true of the god’s metamorphosis and courtship. The voyage,
over a rough sea, takes place at night, without an escort of marine
deities; it is described in five lines and is over before the poem reaches
half way. The emphasis falls heavily on the heroine’s distress and
remorse, which continue over eight stanzas and include her father’s
imagined condemnation. At the end it is Venus and not the god himself
who reveals the identity of the bull.
Here we must confront a controversial issue. According to some
commentators, Jupiter does not have intercourse with Europa until he
has resumed his proper form; cf. Mosch. 2. 163 f., Ov. fast. 5. 615 ff.,
Lucian, dial. mar. 15. 4 (where the bull disappears and Zeus leads his
blushing bride to the Dictaean cave), Nonn. 1. 344 ff.; on this hypothesis
the consummation must lie outside the poem, for in v. 73 Europa still
does not know who the bull is. This scenario is favoured by Orelli,
L. Müller, vol. 1, p. 246, Berres, op. cit. 63 ff. According to others, the
bull mates with Europa on reaching Crete (33 f.), though in Horace’s
compressed narrative this crucial event is omitted; perhaps in pursuing
the analogy with Galatea the poet is skating over the grossly absurd
details. We prefer this second view, as it provides a more plausible
explanation for Europa’s extreme remorse. For some consequences of
the two theories see the notes on 33 f., 37 f., 50, 58 ff., 66 ff., 71 f.
On the question of the poem’s overall coherence, Fraenkel (loc. cit.)
and others have regarded the address to Galatea as simply a loosely
attached prelude to the myth of Europa; he showed a similar misun-
derstanding of 3. 11 (Hypermestra). In fact, as many have seen, the ode is
cleverly integrated, the second part serving as a warning to Galatea. The
dangers of the sea that we meet at the end of the first part (17–24) are
repeated at the beginning of the second (26–8, 31–2); the treachery of
the bull (25) hints that Galatea’s new lover may also prove unreliable;
when Europa’s love turns to hate (45–8) that is a clear warning to
Galatea; when Europa gives way to wild self-reproaches, they seem
meant to waken some feelings of guilt in Galatea. Nevertheless, Horace
manages to express the hope that the girl will be happy (13–14), and
the story of Europa ends with the assurance that a continent will be
named after her. So in spite of everything Galatea’s adventure may
turn out to have some compensations; yet the extravagance of the
conclusion suggests that here too the poet does not expect to be taken
too seriously.
The tone of the poem is highly original. The opening lines on
impious travellers and evil omens (1–7a) are far removed from the
friendly spirit of the usual propempticon (see 1 n.); the pedantic details
of favourable and unfavourable omens (1–12) sound like a parody of
augural lore; and to compare Galatea’s departure with Europa’s ride on
the bull is an amusing flight of fancy. Speeches by abandoned heroines
were a traditional motif, especially in tragedy and epic (e.g. Eur. Med.
160 ff., 214 ff., Ap. Rhod. 4. 355 ff., Catull. 64. 132 ff., Virg. Aen. 4. 534 ff.),
but Europa’s hysterical outburst is (at least on the first theory mentioned
above) grotesquely hyperbolic: when she threatens to break the bull’s
horns (47 f.), or hopes to provide juicy prey for tigers (53–6), or contem-
plates hanging herself by her girdle from a tree (58–60), the melodrama
is calculated to entertain rather than move the reader. In 1. 3 Horace
wrote an introductory propempticon in a spirit of warm affection:
here the tone is teasing and at places even cynical. In 3. 11, where he
professes to have hopes of winning Lyde, Horace used Hypermestra as a
splendid example of loyalty: here Galatea is supposed to be leaving him,
so the exemplary figure is a rather silly Europa. The dryness of the style
suits the situation, deliberately avoiding the sentiment of the conven-
tional propempticon and the glamour normally attached to the legend.
In this respect Horace is more like Callimachus than Moschus or
Propertius.
Metre: Sapphic.
1–2. Impios parrae recinentis omen / ducat: this is an instance of the

‘priamel’ (praeambulum), where the matter in hand is preceded by a ‘foil’
by way of contrast or analogy; cf. 1. 1. 3 ff. with N–H pp. 2 f., 1. 7. 1 ff.
‘laudabunt alii . . . ’, Sappho 16 L–P, A. Henrichs, HSCP 83, 1979:
207 ff., W. H. Race, The Classical Priamel from Homer to Boethius
(Mnem. suppl. 74, 1982). At first sight this seems to reverse the usual
benediction of the propempticon (Cairns 56 f., 130 f.); cf. epod. 10 on
Mevius, especially v. 1 ‘mala soluta navis exit alite’. In 7 ff. H excludes
Galatea from this category and shows that he has come to terms with
her rejection, but this is not yet clear. Faithless women are sometimes
called impiae (2. 8. 19, Pichon s.v.); Propertius in his propempticon
addresses Cynthia as periura (1. 8. 17), though he too is ready to forgive.
The parra was an ill-omened night-bird whose identity is uncertain.
Many think it was a type of owl (Capponi, 1979: 381 f.), sometimes citing
the Venetian dialect word ‘parruzza’; cf. Paul. Nol. carm. 11. 36 ‘si confers
fulicas cycnis et aedona parrae’ (or picae), where a contrast between owl
and nightingale is paralleled at Theoc. 1. 136 and Calp. Sic. 6. 8. J. André
(1967), 118 calls it a night-jar or caprimulgus; its cry is mentioned at
PLM. 5. 61. 9 f. ‘parrus enim quamquam per noctem tinnitet omnem /
sed sua vox nulli iure placere potest’, where the onomatopoeic tinnitet
suits a night-jar better than an owl; H’s recinentis suggests a sinister
repetition, as characteristic of a night-jar as of an owl. When the
glossaries identify it with an ÆNªŁÆº or titmouse (CGL 7. 50) there
is presumably a confusion with the ÆNªŁºÆ or caprimulgus (Thomp-
son 24, 59, Capponi 126 ff.). Others, less convincingly, have identified it
with the lapwing (Keller 2. 178), the wheatear or oenanthe (compare Plin.
nat. hist. 10. 87 with 18. 292), and the jay ( J. W. Poultney, Studies
Presented to D. M. Robinson, 2, 1953: 469 ff.); but these are not primarily
nocturnal birds.
ducat (‘escort’, ‘speed on their way’) would be used more naturally
with a favourable omen; so it sounds sardonic here. For unfavourable
omens cf. for instance Aesch. Ag. 104 ff., Cic. div. 1. 29 and 2. 84 with
Pease, Tib. 1. 3. 17, Ov. met. 10. 452 f., Posidippus, epig. 21, 28, 29 (ed.
Austin and Bastianini), RAC 5. 423 ff.
2. et praegnans canis: for this and the following portents cf. Paul. Fest.
244M ¼ 287L ‘pedestria auspicia nominabant quae dabantur a vulpe
lupo serpente equo ceterisque animalibus quadrupedibus’; similarly a cat
crossing the road is a bad sign (Ar. eccl. 792, Theophr. char. 16. 2), and a
crocodile crawling from right to left prevents a journey (Heliod. 6. 1).
For ill-omened bitches cf. Plaut. Cas. 973 ‘caninam scaevam spero
meliorem fore’, Virg. georg. 1. 470 ‘obscaenaeque canes importunaeque
volucres’. praegnans refers to the time when birth was imminent, for to
encounter a beast or bird at this stage was inauspicious; cf. Plin. nat. hist.
10. 30 ‘(cornix) inauspicatissima fetus tempore’. Pollution was associated
with birth, though less emphasized than in the Jewish–Christian
tradition; see Parker 48 ff.
2–3. aut ab agro / rava decurrens lupa Lanuvino: Lanuvium stands on a

hill 20 miles south-east of Rome, a little to the west of the Appian Way;
it was therefore to the right of anybody heading south to Brindisi in
order to sail for Greece (cf. 19 f.). For wolves in Italy cf. N–H on 1. 22. 9,
Bömer on Ov. fast. 4. 766, T. Ashby, Some Italian Scenes and Festivals
(1929), 112 n. 1; of their sinister character Pliny (nat. hist. 8. 80) says: ‘in
Italia quoque creditur luporum visus esse noxius vocemque homini
quem priores contemplantur adimere ad praesens’ (thus Virg. ecl. 9.
53 f.). ravus is a blend of niger and fulvus (ps.-Acro); cf. Encycl. Brit.,
edn. 11, 28. 772 ‘The ordinary colour of the wolf is yellowish or fulvous
grey’. The adjective is commonly applied to eyes (André, 1949: 70), but
here the eyes seem too small a detail; the word must surely refer to the
animal’s skin, as in epod. 16. 33 ‘ravos . . . leones’.
4. fetaque vulpes: for foxes in Greece and Italy (where they were less
significant than wolves) cf. Keller 1. 88 ff., RE 7. 189 ff. If feta means
‘pregnant’ (OLD 2a), the adjective repeats the sense of praegnans
(ps.-Acr. glosses it with gravida), and several commentators, including
Bentley, have found this objectionable. If it means ‘having recently
given birth’ (OLD 1), we have to picture the vixen as lying with her
cubs within sight of the road; i.e. we have to dissociate her from
decurrens, though she and the wolf seem closely linked by -que. On
either view the arrangement of the omens is less patterned than one
expects in Horace: feta, instead of balancing rava, corresponds to praeg-
nans in v. 2. RN would have looked for a colour-adjective such as fulva.
5. rumpat et serpens iter institutum: for portents involving snakes

cf. Theophr. char. 16. 4, Ter. Phorm. 707 ‘anguis in impluvium decidit
de tegulis’, Pease on Cic. div. 1. 36, 72, 79. Bentley favoured the variant
rumpit, arguing that it was in the interests of the impii to break off their
journey before they came to any harm. But the point is simply that their
hope of a successful journey should be completely frustrated: rumpat
goes further than ducat. The present indicative rumpit would produce a
banal truism applicable to all journeys and not simply to those of the
impii; that would spoil the contrast with what follows.
6–7. si per obliquum similis sagittae / terruit mannos: per obliquum

suits an unexpressed verb of motion; the snake is pictured as darting
across the road (cf. Virg. Aen. 5. 273 f. ‘serpens / aerea quem obliquum
rota transiit’). The simile of the arrow suggests speed (Eur. TGF fr.
1063. 13 ŁA b NF ŒÆd æF, Virg. Aen. 5. 242, 10. 248, 12. 856);
conversely Apollo’s arrow is described as a snake (Aesch. Eum. 181 f.
e Iæªc ZØ, =
æıºı ŁØªª Kæ). The
shape of the head may assist the comparison. manni (probably an
Illyrian word) were fast ponies used for drawing carriages; cf. epod. 4.
14 ‘Appiam mannis terit’ (where Porph. regards them as a sign of
luxury), epist. 1. 7. 77, Lucr. 3. 1063, Ov. am. 2. 16. 49 with McKeown,
Prop. 4. 8. 15 ff. (of a trip to Lanuvium), E. Wölfflin, ALL 7, 1890: 318 ff.
7–8. ego cui timebo / providus auspex: an auspex (¼ avi-spex) took
omens from the observation of birds; the word is used in connection
with auspicia privata, especially at a wedding, whereas augur is the right
word at official ceremonies. The poet plays this role at epist. 1. 20. 9
‘quodsi non odio peccantis desipit augur’, Call. iamb. 5. 31 Kªg ´ŒØ Ø
ŒÆd &ıººÆ with Kerkhecker p. 140. providus here means not just
‘prophetic’ (which would be tautologous with auspex) but also indicates
concern for the future (cf. 3. 5. 13 ‘mens provida Reguli’). cui ¼ ei cui
(for the absence of the antecedent cf. 3. 16. 43 n.); ei is dative with
providus, as it can be with providere (OLD 4b), and not just with the
remote suscitabo (11); cui is dative with timebo (cf. Ov. am. 2. 11. 9 ‘quam
tibi, me miserum, Zephyros Eurosque timebo’ with McKeown, OLD
1b). Heinze thinks the future refers to a principle that H will always
follow (cf. 3. 2. 26), but the following stanza must describe a particular
occasion; the tense presumably indicates that Galatea’s journey has
not yet started (cf. the passage of Ovid cited above, also from a
propempticon).
9–10. antequam stantis repetat paludes / imbrium divina avis immi-

nentum: when birds of sea or marsh splash in the water, it is a sign of
storm; see Aratus, phaen. 942 ff., Cic. div. 1. 14, Varro Atacinus, FLP 14.
1 f. ‘tum liceat pelagi volucres tardaeque paludis / cernere inexpletas
studio certare lavandi’ (with tardae paludis cf. H’s stantis, i.e. stagnantis,
paludes), Virg. georg. 1. 384 ff. ‘dulcibus in stagnis’, Lucan 5. 555 f. The
subjunctive repetat indicates that H is trying to forestall the bird’s return
to the marshes (K–S 2. 368, Wackernagel 1. 246 f.), where it may signal
bad weather. divina means ‘prophetic’; cf. serm. 1. 9. 30 ‘divina anus’
(with Cruquius and Bentley), Cic. div. 1. 120 ‘in avibus divina mens’
with Pease, contradicted by Virg. georg. 1. 415 f. in Lucretian vein: ‘haud
equidem credo quia sit divinitus illis / ingenium aut rerum fato
prudentia maior’. For the genitive cf. ars 218 ‘divina futuri’, Milton, PL
9. 845 f. ‘Yet oft his heart, divine of something ill, / Misgave him’.
11–12. oscinem corvum prece suscitabo / solis ab ortu: the corvus

(which refers back to avis) or Œ æÆ was properly a raven (Thompson
168 f., André 61 ff., Capponi 196 ff., RE 1A. 1. 19 f.), while the cornix or
Œæ (below, 16) was a crow (Thompson 168 ff., Capponi 190 f.); both
birds were associated with augury and weather-signs. oscen is an augural
word (from obs- and cano), referring to a bird that gives omens by its cry
rather than its flight; Fest. 197M ¼ 214L, citing Appius Claudius,
describes both the corvus and the cornix as oscines, and the repeated
c’s in H’s v. 11 suggest the raven’s croak; but our passage implies that
their flight mattered as well. See further Cic. fam. 6. 6. 7 ‘non igitur
ex alitis involatu nec e cantu sinistro oscinis, ut in nostra disciplina
est, tibi auguror’, div. 1. 120 with Pease, Prud. Symm. 2. 571 ‘cornice vel
oscine parra’.
In taking omens a Roman faced south (whereas a Greek faced north),
and flight from the left was generally favourable (Enn. ann. 87 ‘laeva
volavit avis’ with Skutsch); but in Cicero’s Marius an eagle flying from
the west was thought to be a good omen (div. 1. 106). We are also told
that a corvus (unlike a cornix) was favourable when it appeared on the
right (Plaut. asin. 260 ‘picus et cornix a laeva, corvus parra a dextra
consuadent’, aul. 624, Cic. div. 1. 12 with Pease, 1. 85); this is inconsistent
with our passage but is supported by Jerome (PL 26. 219 f.) ‘cum
a sinistro oscinem corvum audiam crocitantem’ (a disparaging reference
to St Ambrose). The Pomptine Marshes, which H’s corvus is to be
prevented from reaching, lay to the west, i.e. to the right, of someone
going south on the Appian Way. But here there seems to be a conflation
of augury (where direction was all-important) and weather-signs (where
it was not).
H of course is not taking augural lore seriously (cf. his treatment of
astrology at 2. 17. 17 ff.). For the nonsensical pedantry of the system see
especially the lengthy inscription in the Umbrian dialect from Gubbio
(ed. J. W. Poultney, The Bronze Tablets of Iguvium, 1959). As a specimen
it is enough to cite ‘ef aserio parfa dersua curaco dersua peico mersto
peica mersta . . . ’ (VIa, pp. 228 ff.); Poultney translates ‘there observe a
parra in the west, a crow in the west, a woodpecker in the east, a magpie
in the east’.
13. sis licet felix ubicumque mavis: if we take the words as they come,
this would naturally mean ‘as far as I am concerned, you can be happy
wherever you prefer’, implying indifference; but this is incompatible
with ‘et memor . . . vivas’ (14), which must be an independent wish, not
dependent on licet. Some translate ‘may you be happy’ comparing Ov.
met. 3. 405 ‘sic amet ipse licet, sic non patiatur amato’ (as if licet denoted
a wish); but that passage means ‘it could be that he himself will fall in
love’; cf. Shackleton Bailey (1982), 97. In our passage Shackleton Bailey
follows T. E. Page, who took licet as parenthetic.
RN thinks it jerky to take the second word in the sentence as a
parenthesis, and he has therefore considered ilicet, ‘without more ado’
(PCPS suppl. 15, 1989: 92). This archaic word was originally a formula of
dismissal (¼ ire licet); Charisius (261B) says it means extemplo as early as
Afran. 215R; and though it is not found elsewhere in Horace it is used
five times by Virgil in this sense (Austin on Aen. 2. 424, S. Timpanaro,
Contributi di Filologia, 1978: 17 ff.). It is combined with a subjunctive at
Sil. 9. 28 f. ‘tradant immo mihi revocatos ilicet enses, / tradant arma
iube’. The meaning ‘without more ado’ would suit the context of the
propempticon; cf. Hom. Od. 5. 205 (Calypso to Odysseus) ÆPŒÆ F
KŁºØ NÆØ; This conjecture would allow us to supply vivere (from 14)
with mavis, which seems more natural than to supply esse felix, as one
tends to do if sis is retained; for the combination of vivere with felix and
memor cf. Virg. Aen. 3. 493 ‘vivite felices’, [Tib.] 3. 5. 31 ‘vivite felices,
memores et vivite nostri’. This idea also leads well to the next sentence,
where it is taken up by ‘nec vetet ire etc.’
14. et memor nostri, Galatea, vivas: ‘remember me’ is a natural com-

monplace at partings; cf. Sappho 94. 7 f. L–P
ÆæØ æ
Œ¼Ł =
ÆØ , Tib. 1. 3. 2 (to Messalla), Ov. am. 2. 11. 37 ‘vade memor nostri’,
[Tib.] 3. 5. 31 (quoted in 13 n.), Juv. 3. 318 ‘ergo vale nostri memor’, Men.
Rhet. 398. 27 f. (prescription for prose propemptica) IØØ ÆPe
BŁÆØ B ºÆØ ıŁÆ , B PÆ , B ØºÆ , Paul. Nol.
carm. 17. 9 ‘i memor nostri’. For similar formulae in letters cf. Lucian,
dial. meretr. 10. 3, Chariton 8. 4. 6, Sid. Apoll. epist. 6. 11. 12, H.Kos-
kenniemi, Studien zur Idee und Phraseologie des griechischen Briefes, 1956:
123 ff., 146 ff. (including papyri).
Galatea is a hetaera’s name at Athen. 1. 6 (the mistress of Dionysius
seduced by Philoxenus), Paul. Sil. anth. Pal. 5. 244. 1, 256. 1; see also
LGPN 3B. 90. H is probably thinking of the nymph of Theoc. 6 and
especially 11 (cf. also Philoxenus, PMG 815–21, Virg. ecl. 9. 39, Ov. met.
13. 738–817, Philostr. imag. 2. 18, LIMC 8. 1, suppl. 1016 ff., 8. 2. 673 f.);
she was wooed unsuccessfully by Polyphemus, who like H is finally
reconciled to failure, and she took a perverse pleasure in the sea (note
also Nonnus 1. 58 on Europa, ‘you would think she was Thetis or
Galatea’). The name is used differently in other propemptica, where
as a Nereid Galatea is asked to assist the voyager; see Prop. 1. 8. 17 f. ‘sed
quocumque modo de me, periura, mereris, / sit Galatea tuae non aliena
viae’, Ov. am. 2. 11. 34 with McKeown.
15–16. teque nec laevus vetet ire picus / nec vaga cornix: the emphatic
te implies ‘whatever bad omens others may experience’. A picus is a
woodpecker (André 127 f., Capponi 414 ff., T. S. Mackay, AJP 96, 1975:
272 ff.); for its importance in augury cf. Thompson 249 f. In Plautus a
picus or cornix (crow) on the left was a good sign (11–12 n.); on the other
hand, cf. Virg. ecl. 9. 15 ‘nisi ante sinistra cava monuisset ab ilice cornix’.
vaga means that the crow is flying randomly and so failing to provide an
omen (cf. Sen. nat. 2. 32. 4 ‘fortuita et sine ratione vaga divinationem
non recipiunt’); it is not here a conventional epithet applicable to all
birds (as at 4. 4. 2, Ov. met. 1. 308) as that would not balance laevus.
17–18. sed vides quanto trepidet tumultu / pronus Orion?: a question

rather than a statement suits the liveliness of the context (cf. 3. 20. 1 n.).
The imagined scene apparently moves on to the Adriatic. The
constellation Orion (Le Boeuffle 129 ff.) ‘sets’ in the technical sense in
November, i.e. its morning setting comes closest before sunrise (3. 1.
27–8 n.). For pronus ‘moving down to its setting’ cf. 1. 28. 21 ‘devexi . . .
Orionis’, Theoc. 7. 53 f.
þÆ K æØ ¯ æØ ªæa ØŒfi =
ŒÆÆ,
Tæø ‹ K TŒÆfiH Æ Y
Ø (also from a propempticon)
with Gow. For the association of Orion’s setting with storms see epod.
10. 10, 15. 7, Hes. op. 619 ff. with West, Virg. Aen. 4. 52 with Pease.
trepidet (‘is in commotion’) refers to the storms that Orion is supposed
to cause, and not just to accompany (3. 1. 31–2 n.).
18–19. ego quid sit ater / Hadriae novi sinus: the Adriatic, seen as a
Œ º or gulf of the Ionian Sea (epod. 10. 19 ‘Ionius . . . sinus’), is
notoriously stormy (3. 3. 5, 3. 9. 23 n.). H claims to know what it is like
from bitter experience (ego is emphatic) because he has sailed to Greece;
for the position of novi cf. Soph. OT 1251
þø b KŒ H PŒ r
I ººıÆØ, Call. aet. fr. 6 with Pfeiffer, Theoc. 16. 16 f., 29. 3 with Gow.
For ater of the darkened sea cf. serm. 2. 2. 16 f. ‘atrum / hiemat mare’,
OLD 2b; it is not just a conventional epithet as in Homer’s ºÆØ
fiø (Il. 24. 79), and it is more sinister than niger (23 below, N–H on 1.
5. 7). quid sit must mean ‘what it is like’ (cf. Virg. ecl. 8. 43 ‘nunc scio quid
sit amor’); in isolation this would cause no doubt, but RN thinks that sit
may be too weak to balance the following peccet (PCPS suppl. 15, 1989:
92). He has considered quo sit ‘to what purpose the sea is black’ (i.e. it is
brewing a storm).
19–20. et quid albus / peccet Iapyx: Iapyx was the wind that blew west-
north-west from Iapygia (the ‘heel’ of Italy) to Greece; cf. 1. 3. 4, Virg.
Aen. 8. 710, Gell. 2. 22. 21, RE 8A. 2. 2299 ff. It is called ‘white’ because it
cleared the clouds away, like the Greek Argestes (cf. 1. 7. 15 f. ‘albus ut
obscuro deterget nubila caelo / saepe Notus’ with N–H); the word is
placed so as to produce a colour contrast with ater immediately above.
For the misbehaviour of the winds cf. Virg. Aen. 1. 136, where Aeolus
says ‘post mihi non simili poena commissa luetis’, Tac. ann. 14. 3. 3
‘quem adeo iniquum ut sceleri adsignet quod venti et fluctus deliquer-
int?’ As peccare properly describes a stumble, RN (loc. cit.) has suggested
a play on the horses of Iapygia; cf. Virg. Aen. 3. 537 f. (the white horses of
Castrum Minervae), 7. 691 ‘at Messapus equum domitor, Neptunia
proles’ with Horsfall (Messapia was roughly the same as Iapygia though
in the Aeneid Messapus is oddly connected with south Etruria), 11. 678
‘et equo venator Iapyge fertur’, Sil. 4. 555 f. ‘Iapyge . . . / . . . equo’.
21–2. hostium uxores puerique caecos / sentiant motus orientis Austri:

for the diversion of trouble to others (aversio) cf. 1. 21. 13 ff. ‘hic bellum
lacrimosum, hic miseram famem / . . . in / Persas atque Britannos / vestra
motus aget prece’ with N–H’s parallels. Enemies are conventionally
mentioned in such contexts; cf. Nic. ther. 186 K
ŁæH ı æÆ ŒEÆ
ŒÆæÆØ KºØ, Virg. georg. 3. 513, Ov. ars 3. 247 f. ‘hostibus eveniat
tam foedi causa pudoris / inque nurus Parthas dedecus illud eat’; but
here H has in mind private enemies (OLD 3), cf. Prop. 3. 8. 20 ‘hostibus
eveniat lenta puella meis’, Ov. am. 3. 11. 16, her. 16. 219; Williams
compares our own saying ‘I wouldn’t wish it on my worst enemy’. For
variations on the theme cf. Hollis on Call. Hecale, fr. 47. 10–11.
H is describing the first indications of the gathering storm (thus
ps.-Acro); cf. Hom. Il. 14. 16 ff., Arat. phaen. 909 ff., Virg. georg. 1.
356 ff. caecos motus refers to ‘waves without foam or spray, groundswell’
(OLD 7b); cf. Sisenna, hist. 24 ‘caecos . . . fluctus in se provolvere leniter
occepit (mare)’. sentiant makes something of an oxymoron with caecos
motus, though not with the phenomena that follow. For the dangerous
south winds in the Adriatic cf. 1. 3. 15, 3. 3. 4 f. Though orientis is used
here of a rising storm, the other sense of ‘eastern’ suggests to RN a
typical Horatian word-play with Austri.
22–4. et / aequoris nigri fremitum et trementis / verbere ripas: the

gathering of the storm is marked by the movement from caecos to orientis
to fremitum to trementis verbere; note the repetition of r, which the
Romans rolled (3. 4. 17–18 n.). For fremitum, a premonitory rumble
rather than a roar, cf. Virg. Aen. 1. 56 ‘circum claustra fremunt’, where
the winds are like growling animals. trementis verbere means that the
shore shudders, cf. Sen. Phaedr. 1013 ‘saxa cum fluctu tremunt’; but it
also includes the notion of fear. verbere in such contexts normally refers
to the wind (cf. Meleager, anth. Pal. 5. 180. 5 f. Iø ØØ ŁºÆÆ
= æÆ
f fi A, Lucr. 5. 957 ‘verbere ventorum’, Musaeus 297 f., Nonn. 13.
390, OLD 4); but the verb is also used of waves (Curt. 4. 3. 6). ripae
refers to the sea-shore as at 2. 18. 22, and may imply quite a steep
coastline; cf. Mela 3. 89 ‘vasta omnia vastis praecisa montibus ripae
potius sunt quam litora’.
It is objected that voyagers are more aware of their own troubles than
of the shaking of the sea-shore: hence gementes (Bentley, cf. 2. 20. 14),
rupes (Cunningham, cf. Soph. Ant. 592), costas (Shackleton Bailey, of the
ship’s ribs, cf. epod. 10. 3). RN suggests that the ‘wives and children’ are
watching from dry land, like H; the hostes themselves and not their
children would be the right people to be exposed to danger (cf. 1 ff.). NR
takes the usual view that the wives and children are thought of as
experiencing the perils of the sea, like Galatea (and Europa in the
following myth); sentiant motus suits voyagers, and to those caught by
a storm in a sailing ship the reverberation of waves on a coast would be
terrifying. But the shuddering would more truly be felt by people
standing on the shore.
25–6. sic et Europe niveum doloso / credidit tauro latus: for sic intro-
ducing a mythological analogy cf. epist. 1. 18. 41, Fraenkel on Ag. 718; for
et cf. Soph. Ant. 944 ºÆ ŒÆd ˜ÆÆ PæØ H = IººÆØ Æ .
niveum is a picturesque touch, though the description of girl on bull is
much less elaborate in H than in some other writers; cf. Mosch. 2. 125 ff.,
Ov. fast. 5. 605 ff., met. 2. 874 f., Ach. Tat. 1. 1. 10 ff., Nonn. 1. 65 ff.
Depictions of wind-swept and half-naked heroines are a feature of
Hellenistic and later poets (Catull. 64. 63 ff., Griffin 103 ff.) and owe
much to works of art (see LIMC cited in the introduction).
credidit is set against doloso; the bull invited Europa to sit on his back
(Mosch. 2. 100, 108), and in literature and art she sits sideways (Ach.
Tat. loc. cit. P æØ, Iººa ŒÆa ºıæ), grasping one of
his horns. credidit latus may also echo a stock phrase about ‘trusting
one’s flank to somebody’; cf. Plin. pan. 23. 2 ‘quod primo statim die
latus crederes omnibus’, Sen. nat. 4a praef. 3. 3 ‘ne adulatoribus
latus praebeas’.
26–7. et scatentem / beluis pontum: the expression is based on Hom.

Od. 3. 158 ªÆŒÆ , though that refers to the size of the
monsters rather than their abundance (see Gow on Theoc. 17. 98); cf.
also 1. 3. 18 (‘monstra natantia’), 4. 14. 47, Virg. Aen. 6. 729, Albinovanus
Pedo 5 ff. (FLP p. 315 with Courtney). scatere is an archaic word (avoided
by Cicero and in general by the Augustan poets) that means ‘to gush’ of
liquids (Enn. [?] scaen. 155V ‘fontes scatere’ quoted at Cic. Tusc. 1. 69,
Lucr. 5. 952) and ‘to teem’ of animals; cf. Lucr. 5. 39 f. ‘ita ad satiatem
terra ferarum / nunc etiam scatit’, Plin. nat. hist. 31. 18 ‘lacum scatentem
albis serpentibus’. In Moschus the creatures of the deep provide a joyful
escort (2. 116 ff. with Bühler); here they are frightening.
27–8. mediasque fraudes / palluit audax: fraudes here means ‘dangers’

(OLD 1), but as it often has the sense of ‘treachery’ (OLD 5), it recalls
doloso (25) in the same position. medias means ‘intervening’, ‘in her
path’; cf. Varius, FLP 4. 5 ‘non amnes illam medii, non ardua tardant’,
Virg. Aen. 9. 142 ‘medii fiducia valli’, TLL 8. 583. 34 ff., Xen. Cyr. 5. 2. 26
K fiø Kd F ıÆØ; (what is to hinder the union
of forces?)
palluit is best analysed as the perfect of pallescere, ‘to grow pale’ (cf.
epist. 1. 3. 10 ‘Pindarici fontis qui non expalluit haustus’) rather than of
pallere (Pers. 5. 184 ‘recutitaque sabbata palles’); inceptive verbs are
sometimes transitive (H–Sz 298), whereas the construction in Persius
is more unusual. palluit audax is a typical Horatian oxymoron, describ-
ing Europa’s understandable pallor combined with her exceptional
boldness in facing the sea. audax cannot describe the recklessness of
the whole enterprise, as she had not intended to be abducted.
29–30. nuper in pratis studiosa florum et / debitae Nymphis opifex
coronae: the picking of flowers was part of the tradition from Hesiod
(fr. 140 M–W) through Bacchylides (fr. 10 Snell) to Moschus (2. 63 ff.).
The uncut meadow is a symbol of virginity (Eur. Hipp. 73 f. d
ºŒe Æ K IŒæı = ºØH . . . ŒÆ æø), like the
enclosed garden (Catull. 62. 39–45), and the picking foreshadows deflor-
ation; cf. h. Dem. 5 ff. with Richardson, Moschus, loc. cit. with Bühler,
Ov. met. 5. 392 ff., Claud. rapt. Pros. 2. 128 ff., Lyne (1989), 151 ff., Milton,
PL 4. 268 ff. ‘Not that fair field / Of Enna, where Proserpine gathering
flowers / Herself a fairer flower by gloomy Dis / Was gathered’.
debitae means ‘vowed’ (1. 36. 2 ); garlands for the Nymphs are men-
tioned by Men. dysc. 51 and Longus 1. 9. 2. The Nymphs are already
present in Hesiod (loc. cit.), but their place is taken by ordinary compan-
ions in Moschus. opifex ‘artificer’ is a dry way of describing Europa’s
activity, as if she were a professional maker of garlands (coronarius).
31–2. nocte sublustri nihil astra praeter / vidit et undas: sublustri means
‘slightly lit’, hence ‘glimmering’, just as ‘me subpudet’ means ‘I am
slightly ashamed’; for the combination with nocte (perhaps from early
poetry) cf. Liv. 5. 47. 2 (when the Gauls try to ascend the Capitol), Virg.
Aen. 9. 373 ‘sublustri noctis in umbra’ (with Serv. and the schol. Ver.).
There is a contrast with the flowery meadows, which by implication
were sunlit (1. 26. 7 ‘apricos necte flores’, epist. 1. 14. 30 ‘aprico . . . prato’);
compare the suppressed adjectives at 3. 13. 6 f. For the gloomy seascape
cf. Hom. Od. 14. 301 f., Mosch. 2. 132 f. Æ h IŒ Ø ±ºææŁ
h Zæ IØ, = Iºº Icæ b oæŁ æŁ b Iæø (with
Bühler on 131–4), Lucr. 4. 434 ‘ubi nil aliud nisi aquam caelumque
tuentur’, Virg. Aen. 3. 193, Ov. trist. 1. 2. 23. For the anastrophe of
disyllabic praeter see H–Sz 216.
33–4. quae simul centum tetigit potentem / oppidis Creten: in the

traditional story, as represented by Moschus 2. 162 ff., the god carried
Europa to Crete, where he was united with her. simul tetigit appears
to leave no time for this; for an explanation see the introduction.
centum . . . oppidis alludes to the Homeric epithet ŒÆ ºØ (cf.
Virg. Aen. 3. 106). The word ‘syncretism’ was associated with the
federation of these cities (Wackernagel 1. 301), and potentem probably
refers to the resulting power; for the power of Europa’s son Minos see
Thuc. 1. 4 with Hornblower. The Greek form Creten suits H’s lyrics (cf.
Europe in vv. 25 and 57, 3. 3. 47); contrast epod. 9. 29 ‘centum nobilem
Cretam urbibus’.
34–6. ‘pater o relictum / filiae nomen pietasque’ dixit / ‘victa furore!’:

Europa’s father was called Phoenix in Homer (Il. 14. 321) and Moschus
(2. 7 with Bühler), but Agenor in Herodotus (4. 147. 4) and much of the
later tradition. pater is an exclamatory nominative (K–S 1. 274); it is tied
by the alliteration and by -que to pietas, which could not be a vocative
(cf. Virg. Aen. 6. 878, ‘heu pietas, heu prisca fides!’). relictum filiae nomen
is in apposition to pater, and is balanced by victa furore; there should be
no comma after pater as the words form a single unit (cf. Virg. ecl. 3. 3
‘infelix o semper oves pecus’, though there the word-order is more
artificial). filiae is dative of agent: by leaving her home (49) Europa
has relinquished the right to speak of a father. Some editors take it as a
genitive, regarding pater as a vocative, but then pater becomes impos-
sibly isolated and its correspondence to pietas is destroyed. For nomen
applied to a family relationship cf. Ov. her. 10. 69 f. ‘pater et tellus . . . /
prodita sunt facto, nomina cara, meo’, met. 8. 508 ‘nunc animum pietas
maternaque nomina frangunt’.
37. unde quo veni?: in Greek two questions are sometimes combined in
one clause (cf. Hom. Od. 1. 170 Ł N IæH; K–G 2. 521 f.); in
Latin cf. Cic. Verr. 3. 191 ‘quo ex loco in quem locum?’, H–Sz 459 f.
‘Where am I?’ is a cry of the disoriented; cf. Aesch. Ag. 1086 ff., Virg.
Aen. 4. 595 ‘quid loquor? aut ubi sum?’, 10. 670 ‘quo feror? unde abii?’ H
has given the conventional idiom an additional sense; for as the se-
quence shows (i.e. 34–6 and 37b–40), Europa is also thinking of the
change in her moral situation.
37–8. levis una mors est / virginum culpae: for the commonplace
(established by the plural virginum) cf. Eur. Heraclidae 959, Dem.19. 110
with MacDowell, Prop. 4. 4. 17 ‘et satis una malae potuit mors esse
puellae . . . ?’ with Fedeli. culpa is often used of women’s sexual miscon-
duct (3. 6. 17, Pease on Virg. Aen. 4. 19, OLD 3b). Those who think that
Europa was still a virgin argue that in her overwrought state she is
exaggerating a thoughtless indiscretion: in Moschus she fondled and
kissed the bull (2. 95 f.) and finally sat on his back (108), and in the ode
she admits to having fallen in love with him (multum amati in 47). They
can compare this to the kissing imputed to less innocent herdsmen
([Theoc.] 27. 7 ŒÆº Ø ÆºÆ ØºØ, PŒ ¼ıªÆ ŒæÆ), so it
could not be regarded as an innocent game; on the other hand,
H might not wish to suggest to Galatea that Europa had actually
mated with a bull. Those who think otherwise argue that virginum
culpae, like turpe commissum (39), is too serious an expression to indicate
anything less than the loss of virginity; such feelings of guilt, however
unjustified, are frequently experienced by victims of rape. As for
Galatea, the myth (even without an explicit mention of the mating)
serves as a sardonic warning to a girl who is ill-advised enough to go
overseas with a dubious lover.
38–9. vigilansne ploro / turpe commissum: the emphasis lies on
vigilansne, which introduces a chiasmus: is she awake and guilty or
innocent and dreaming? plorare (whence the French pleurer) is avoided
by Virgil and later epic, and for the most part in classical prose; but it is
found in elegy and 12 times in Horace, including 5 instances in the Odes
(Axelson 28 f., cf. 3. 10. 4 n.).
39–42. an vitiis carentem / ludit imago / vana quae porta fugiens

eburna / somnium ducit?: false dreams issue from the gate of ivory,
true dreams from the gate of horn; cf. Hom. Od. 19. 564 f. H [ºø]
Q Œ ºŁøØ Øa æØF KºÆ , = ¥ Þ KºÆæÆØ,
IŒæÆÆ æ , Plat. Charm. 173a, Virg. Aen. 6. 895 f. (presumably
later than our passage) ‘altera candenti perfecta nitens elephanto, / sed
falsa ad caelum mittunt insomnia manes’ (Austin gives further parallels).
Heinze explains that the imago is the objective simulacrum that emerges
from the underworld and the somnium the subjective dream that it
produces; in that case vana should probably be taken with the following
relative clause (so Bennett in the Loeb edition); it would then make a
frame with eburna, reflecting Homer’s play on KºÆ ‘ivory’ and
KºÆæÆØ ‘deceive’. But note should be taken of Sanadon’s conjecture
quam; this requires taking vana with the previous line (for the uncom-
mon pause after the first word of the stanza cf. 4. 11. 5); imago vana
would then correspond to Homer’s ‘unfulfilled words’ and somnium ducit
to his (ZØæØ) IŒæÆÆ æ . For vana of illusory images cf. ars
7 ‘velut aegri somnia, vanae / fingentur species’, Sen. dial. 2. 11. 1.
42–4. meliusne fluctus / ire per longos fuit an recentes / carpere

flores?: Europa abandons the idea that her experience is just a bad
dream, and with ‘self-taunting irony’ (Shorey) puts another pair of
alternatives. longos refers not to the length of the waves but to their
remoteness from her home, ‘faraway waves’ (OLD 6); elsewhere the
word describes a long, or wide, stretch of sea (OLD 4a, 3. 3. 37 f.
‘longus . . . pontus’, Moschus 2. 133 Iæø) and that could be
the sense here, but in RN’s view that would suit pontus better than
fluctus. Alliteration sets fluctus against flores; RN senses an antithesis
between the far-off waves and the fresh flowers, even if one adjective
refers to space and the other to time (for contrasts between the distant
and the ready-to-hand cf. N–H on 2. 16. 25).
45–6. si quis infamem mihi nunc iuvencum / dedat iratae: infamem is a

strong word (cf. 2. 14. 18 f. ‘Danai genus / infame’); infamia was a legal
concept that could entail serious sanctions (Crook 83 ff.). There may be
an implication here that the bull was guilty of rape (Williams). For the
threat (‘if I could only get my hands on him’) cf. Ter. ad. 311 ff. ‘nil est
quod malim quam illam totam familiam dari mi obviam, / . . . vah quibus
illum (Syrum) lacerarem modis’.
46–8. lacerare ferro et / frangere enitar modo multum amati / cornua

monstri: Campbell proposed lacerare frontem because it was not obvious
where Europa would obtain a sword or dagger, and RN was once
interested in this idea; but we are not invited to ask such questions of
a hysterical girl. The transmitted reading is supported when Venus says
‘laceranda reddet / cornua taurus’ (71 f.). lacerare can be used of shatter-
ing hard objects like ships and bridges (OLD 2); Hercules broke the
horns of Achelous with his bare hands (Ov. met. 9. 85 f.) but Theseus
used his club on the horns of the Cretan bull (Hollis on Call. Hecale,
fr. 69. 1).
49. impudens liqui patrios penates: ‘shameless’ is too strong a word for
Europa’s actual behaviour, just as ‘abandoning home’ is a very unfair way
to speak of her abduction; this language is part of her exaggerated self-
excoriation. penates can be applied to foreign homes (Virg. Aen. 1. 527,
4. 21 etc.), but the Roman touch has particular resonance in an address
to Galatea. For the collocation with patrios cf. serm. 2. 5. 4, Tib. 1. 3. 33.
Here the adjective means ‘paternal’ (cf. 34, 57) rather than the more
general ‘ancestral’; cf. Eur. Med. 166 t æ, Mosch. 2. 146 f. HÆ =
Ææe IæºØFÆ, Catull. 64. 180.
50. impudens Orcum moror: the anaphora links Europa’s offence to

her proposed expiation by suicide. If she is still a virgin, she regards her
indiscretion at Sidon as serious enough to deserve death. If she
has already been raped (a thing she is reluctant to mention explicitly)
she regards this as the consequence of her original shamelessness.
Orcus is another very Roman touch (N–H on 2. 18. 30), more appro-
priate to Galatea than Europa. Death is ‘kept waiting’ in various
contexts: cf. Eur. Alc. 255, Ov. her. 10. 82, Stat. Theb. 7. 364, OLD 1b.
50–2. o deorum / si quis haec audis, utinam inter errem / nuda

leones: for the invocation cf. epod. 5. 1 f. ‘At o deorum quidquid in
caelo regit / terras et humanum genus’; si quis reflects the blanket clause
common in prayers (2. 1. 25 with N–H, 3. 21. 5 n.). For errare of the
distraught cf. Virg. ecl. 6. 52 (Pasiphae), Prop. 1. 1. 11 (Milanion), Ov. am.
3. 6. 50 (Ilia). For the position of nuda between inter and leones cf.
3. 3. 37, 3. 15. 5, K–S 1. 588; here it may reflect Europa’s position amid
the lions.
Elsewhere deserted heroines fear lions (Catull. 64. 152 f., Ov. her.
10. 85), and at Octavia 636 ff. Agrippina wishes she had been torn
by beasts; for a closer parallel cf. Nonn. 48. 736 f. (perhaps from a
Hellenistic prototype) where the distraught Aura visits a lioness’s den in
the hope of meeting her death. One thinks of Helen’s agonies of guilt
in All’s Well That Ends Well i i i . ii. 116 ff.: ‘Better ’twere / I met the ravin
lion where he roar’d / With sharp constraint of hunger’. nuda empha-
sizes Europa’s vulnerability, but there is also a hint of sexual titillation
(cf. Griffin 106). This element, which is more characteristic of Ovid,
should be associated with fictional fantasies rather than the arena;
malefactors could indeed be thrown to the beasts even in the Republican
and Augustan periods (Cic. Pis. 89, Strab. 6. 2. 6, Mommsen, Römisches
Strafrecht, 1899: 925 ff.), but we do not seem to hear of women being
punished in this way till the Christian martyrdoms (see H. Musurillo,
Acts of the Christian Martyrs, 1972: 106 ff. for that curious psychological
document, the Passio Sanctarum Perpetuae et Felicitatis).
53–4. antequam turpis macies decentis / occupet malas: elsewhere such

things are said by way of a compliment when somebody dies young
(Soph. Ant. 819 h ŁØØ ºÆªEÆ Ø [ŒÆÆØ], Stat. silv.
2. 1. 154, Chariton 1. 5. 7.). For occupet cf. [Tib.] 4. 4. 5 ‘effice ne macies
pallentes occupet artus’.
54–5. teneraeque sucus / defluat praedae: tenerae describes the softness

of youth; cf. 1. 1. 26 ‘tenerae coniugis immemor’, 4. 1. 26. For sucus of
people cf. Ter. eun. 318 (of a girl of 16 ) ‘color verus, corpus solidum et
suci plenum’, Ov. met. 3. 397 f., Apul. apol. 4, OLD 3b, Paul. Sil. anth.
Pal. 5. 258. 1 —æ ŒæØ KØ, 6ºØÆ, c Þıd j Oe l = ,
Suckling, ‘A Ballad upon a Wedding’, stanza 6, ‘No grape . . . could be /
So . . . soft as she, / Nor half so full of juice’. Human ‘sap’ was thought to
rise and fall as in a tree; cf. epod. 13. 4 ‘dumque virent genua’, Aesch. Ag.
76 f. ‹ Ææe ıºe æø = Ke I fi Æø
Æ with Fraenkel, Plin.
nat. hist. 2. 189, Onians 177 f., 192, 223 f., E. Irwin 35 ff.
55–6. speciosa quaero / pascere tigris: speciosa is emphatic ‘while still

good-looking’. The succession of s’s and c’s gives a harsh effect as in
v. 11. tigris balances leones at the same place in the previous stanza.
Tigers were known to Romans at this date only from literature (3. 3.
13–15 n.).
57–8. ‘vilis Europe,’ pater urget absens, / ‘quid mori cessas? . . . : urget
absens produces something of an oxymoron, since urges primarily
suggests physical pressure; the situation resembles Virg. Aen. 4. 83
‘illum absens absentem auditque videtque’. For the question (which
echoes the statement in v. 50) cf. Catull. 52. 1 ‘quid moraris emori?’,
Ov. her. 9. 146 ‘impia quid dubitas, Deianira mori?’ cessas implies not just
delay but neglect of duty.
58–60. potes hac ab orno / pendulum zona bene te secuta / laedere
collum: the lively hac continues the idea that the parent can see her from
afar; for the word’s deictic function cf. 2. 11. 13 ‘cur non sub alta vel
platano vel hac / pinu iacentes . . . ?’ with N–H. The manna-ash was
often barren (Virg. georg. 2. 111 ‘steriles orni’) and so could be character-
ized as an infelix arbor, suitable for executions; cf. Liv. 1. 26. 6 ‘infelici
arbori reste suspendito’ with Ogilvie, N–H on 2. 13. 11. Hanging was
an informe letum (Virg. Aen. 12. 603) and the usual form of suicide in
Greek poetry for self-reproaching women, e.g. Jocasta and Phaedra (cf.
N. Loraux, Façons tragiques de tuer une femme, 1985: 38 ff.).
The girdle was a symbol of virginity (Hom. Od. 11. 245, h. Hom. 5. 164,
Catull. 2. 14); cf. especially Mosch. 2. 73, 164 (of Zeus) ºF ƒ æ
with Bühler. For its employment as a noose cf. Aesch. suppl. 457,
Aristodicus, anth. Pal. 7. 473. 3 f., Parthenius, fr. 33. 5 Lightfoot, where
Byblis hangs herself from an oak. bene te secuta means ‘that you fortu-
nately have brought with you’. sequi can mean ‘accompany’ as well as
‘follow’; for this use, even of things, cf. Catull. 68. 36 ‘capsula me
sequitur’, Varr. rust. 1. 55. 4 (of olive oil) ‘dominum et in balneas et
gymnasium sequitur’, Fronto p. 62. 6 van den Hout. On one view this
implies that Europa is still a virgin, even if she does not deserve to be;
but the point seems rather to be that, since it no longer has any symbolic
function, the girdle may as well be used for suicide.
laedere is a grim euphemism such as sometimes occurs in mentions of
death; cf. Soph. Ant. 54 ºŒÆEØ IæÆØØ ºøAÆØ . e-/lidere is
found in some insignificant MSS (for the division of a word cf. 1. 2. 19, 1.
25. 11, 2. 16. 7); it is paralleled at Prud. peristeph. 10. 1108 ‘elidit illic fune
collum martyris’, and is here the vox propria. Shackleton Bailey prints it
and NR is inclined to accept it, even though this impairs the character-
istic rhyme of zona and secuta, at least to some extent; we do not know
the exact sound produced by the synaloephe of a and e; cf. Allen 81.
61–2. sive te rupes et acuta leto / saxa delectant: the three standard
methods of suicide were the sword, the noose, and the precipice (with
poison sometimes substituted for the sword); cf. schol. f on Pind. O. 1.
97 , Iª
, Œæ , paroem. Gr. 1. 164, 2. 662, epod. 17. 70 ff., Sen.
Phaedr. 258 ff., Plin. nat. hist. 2. 156, E. Fraenkel, Philologus 87, 1932:
470 ff. ¼ Kleine Beiträge, 1964: 1. 465 ff. For the jump cf. for instance
[Aesch.] Prom. 748, Ar. ran. 129 ff., Theoc. 3. 25, Virg. ecl. 8. 59 f., and
especially Sappho’s legendary leap from the Leucadian rock (Turpilius,
CRF pp. 113 ff., Ov. her. 15. 171 ff. with Palmer, fast. 5. 630 with Bömer,
Strab. 10. 2. 9). See further Y. Grisé, Le Suicide dans la Rome antique,
1982: 117, A. J. L. van Hooff, From Autothanasia to Suicide, 1990: 73 ff.
Commentators in general think that acuta leto saxa means ‘rocks
sharp for death’; acutus as an adjective is elsewhere applied to rocks
(Virg. Aen. 1. 45, Sen. Ag. 571), but here it would make the construction
easier if it retained some of its force as a participle (for the dative cf.
epist. 1. 3. 23 ‘seu linguam causis acuis’, Colum. 10. 105 ‘quaeque viros
acuunt armantque puellis’). RN has some doubts, but it would be at least
as awkward to take leto as an ablative with acuta. Mitscherlich con-
sidered the possibility that leto is to be combined with delectant; this
would be easier if we could give the verb its archaic sense of ‘to entice’:
cf. Enn. trag. 303 J ‘set me Apollo ipse delectat ductat Delphicus’
(Non. 97. 29 ‘delectare inlicere adtrahere’), Quadrigarius 46P ‘Fabius de
nocte coepit hostibus castra simulare obpugnare, eum hostem delectare’,
TLL 5. 1. 422. 43 ff. delectare is generally avoided at the grander levels of
Roman poetry (Axelson 106), but not in the Odes. In the present case it
carries a sneer.
62–3. age te procellae / crede veloci: Europa goes further than the self-
reproachful Helen, who wishes that at birth she had been blown to
distant seas (Hom. Il. 6. 345 ff.). For the ironic use of se credere cf. Virg.
Aen. 6. 15 (of Daedalus) ‘ausus se credere caelo’, Sen. Med. 301 ff.;
committere is used similarly at Juv. 12. 57. Philoctetes hoped for a similar
fate at Acc. TRF 562 f. ‘heu! qui salsis fluctibus mandet / me ex sublimo
vertice saxi?’ (¼ ROL 2. 564 f.).
63–6. nisi erile mavis / carpere pensum / regius sanguis dominaeque

tradi / barbarae paelex: for the fate of captive women as domestic slaves
cf. Hom. Il. 6. 456 ŒÆ Œ K ` ! æªØ KFÆ æe ¼ºº ƒe ÆØ ,
Eur. Hec. 359 ff., etc. pensum was the amount of wool (already scoured,
washed, and dyed) weighed out for a woman to spin (Virg. georg.
4. 348 f., Prop. 3. 15. 15). Before spinning could begin it was first teased
into a fluff, a process described by carpere or vellere; cf. Virg. georg. 1. 390
‘carpentes pensa’, 4. 334 f., Blümner 1 (1912), 109. erilis is frequent in
comedy in expressions like erilis filius, but later the use of the adjective
instead of the genitive is a mark of a grander style (Virg. Aen. 7. 490,
8. 462, Löfstedt 1, 1942: 116 ff.). The degradation was greater because
Europa was a king’s daughter (cf. Creusa’s words at Aen. 2. 785 ff.); regius
is set against erile above and dominae below (all three words are in an
emphatic position). Such a servant might become a concubine of her
master and a hated rival (paelex) of her mistress (cf. Aesch. Ag. 1441 of
Cassandra); for parallel situations with Roman ancillae cf. Courtney
on Juv. 2. 57. Though Europa was from Phoenicia, her family was
Greek; she can think of barbarians because she does not know she is
in Crete.
66–8. aderat querenti / perfidum ridens Venus et remisso / filius arcu:

aderat expresses the suddenness of the epiphany. The conventional
smiles of Venus (Hom. Od. 8. 362, Sapph. 1. 14) were often accompanied
by deceit (cf. Hes. theog. 205), and were at times hard to interpret
(Theoc. 1. 95 f. with Gow); here she has amused herself at Europa’s
expense and is only now about to reveal the happy ending in store
for her. For the Greek adverbial accusative cf. 1. 22. 23 ‘dulce ridentem’,
2. 12. 14 with N–H, Apul. met. 6. 13. 2 ‘subridens amarum’ with Kenney,
F. Muecke, Encicl. oraz. 2. 764.
remisso . . . arcu: the slackened bow implies that Cupid has now shot
his bolt; cf. 3. 8. 23 f. ‘iam Scythae laxo meditantur arcu / cedere campis’.
Ps.-Acro comments ‘merito remisso arcu Cupido, tamquam qui Euro-
pae iam debellasset pudorem’ (presumably on arrival at Crete); if Europa
is still a virgin one has to imagine that Cupid fires again as she finishes
her complaint, but that can hardly be extracted from the text. In NR’s
view the phrase suggests that Jupiter’s passion has also run its course (cf.
Moschus 75 f.); but RN thinks we should focus entirely on Europa.
69–70. mox ubi lusit satis ‘abstineto’ / dixit ‘irarum calidaeque rixae:
with cum, the transmitted reading in 71, abstineto is a true future impera-
tive by which the order becomes operative at a later stage; cf. 3. 14. 23 f.
‘si per invisum mora ianitorem / fiet, abito’ with note, epod. 13. 17 ‘illic
omne malum vino cantuque levato’. The form is often combined with
subordinate clauses referring to the future, and in RN’s view its presence
is a significant reason for retaining cum in v. 71. For the Graecizing
genitive (as with I
ŁÆØ) cf. 4. 9. 37 f. ‘abstinens / . . . pecuniae’,
2. 9. 17 f. ‘desine mollium / tandem querellarum’. The plural irarum
refers to fits of rage (cf. 1. 16. 9).
71–2. cum tibi invisus laceranda reddet / cornua taurus: ps.-Acro

comments ‘ironicos dictum’, following Porph.’s explanation ‘nec laceres
eum . . . cum se tibi rursus obtulerit’. On this interpretation Europa is to
stop whingeing when Jupiter reappears in human form and the pair are
united; Venus expresses this by echoing Europa’s words in 46 f. Bentley
saw a metaphorical hint of passionate love-making here and RN thinks
this is a possibility; cf. Claud. 14.5 ff on the marriage of Honorius ‘ne
cessa, iuvenis, comminus adgredi, / inpacata licet saeviat unguibus’. RN
thinks there may be a conflation of the two versions: (a) rape by the bull,
(b) the return of the god in his true form. On the other hand, in all the
amours of Zeus/Jupiter sex takes place at the earliest opportunity and is
not repeated; so NR does not believe that the god is expected to return,
especially in view of remisso arcu (67 f.); as there is some awkwardness in
taking the cum clause as the complete opposite of the truth after we have
been told that Venus is no longer teasing (69), he favours Fabricius’ non,
though this involves taking abstineto (69) as the equivalent of an ordin-
ary imperative (K–S 1. 198e, H–Sz 340 f.).
73. uxor invicti Iovis esse nescis: of the two interpretations given by
Porph. we prefer ‘you do not know how to play the part of Jupiter’s wife’.
For this use of nescire cf. Val. Max. 2. 6. 6 ‘abi igitur et esto servus
quoniam liber esse nescisti’, Aesch. suppl. 917 b r ÆØ æH PŒ
KÆÆØ. This compresses two ideas: ‘You are now the wife of Jupiter’
and ‘You must behave as such’. The injunction is then elaborated by bene
ferre . . . disce (74 f.). In Jupiter’s case, of course, uxor is a euphemism.
The alternative is ‘you do not realize that you are the wife of Jove’
(¼ ‘uxorem te esse nescis’); in that case bene ferre and disce are not mere
variations of nescis esse, but mark a new consequential injunction. H
would then be using a Greek construction; cf. epist. 1. 7. 22, Catull. 4. 2
‘ait fuisse navium celerrimus’ with Fordyce, K–S 1. 702, H–Sz 363 f.
However, as Postgate points out (CR 30, 1919: 190 f.), there are no
Latin precedents for this construction with scire or nescire; in Greek
such verbs require a participle rather than an infinitive (K–G 2. 51);
hence Statius, Theb. 7. 792 ‘scit peritura ratis’. It is not a satisfactory
answer to say that in Latin the verb ‘to be’ has no present participle; for
how was the reader to know that here H had departed from the norm?
For invictus as an epithet of Jupiter cf. Ov. fast. 6. 650 ‘invicto sunt data
templa Iovi’, though it is not clear that there was a separate cult of
‘Jupiter Invictus’ (L. Richardson 227).
74–5. mitte singultus, bene ferre magnam / disce fortunam: this pas-
sage has some of the character of an annunciation, which often begins
with words of reassurance; cf. schol. on Od. 11. 322 (Aphrodite to
Ariadne) ŁÆææE ÆPfi Ð ÆæÆØE ˜Øı ªaæ ŁÆØ ªıÆEŒÆ ŒÆd
PŒºB ªŁÆØ, h. Hom. 5. 193 ff. (Aphrodite to Anchises), Theoc. 24.
73 ff. (Teiresias to Alcmena), Mosch. 2. 154 f. (Zeus to Europa) ‘ŁæØ
ÆæŁØŒ c ØŁØ Ø r Æ: = ÆP Ø ˘ NØ’ (with Bühler),
Dion. Hal. ant. Rom. 1. 77. 2 (Mars to Ilia), Ov. ars 1. 556 (Bacchus to
Ariadne), Luke 1: 30–1 (Gabriel to Mary) ‘1c F, 1ÆæØ yæ
ªaæ
æØ Ææa fiH ¨fiH’. bene ferre means ‘to sustain with proper
dignity’, cf. McKeown on Ov. am. 1. 2. 10, TLL 2. 2119. 36 ff.
75–6. tua sectus orbis / nomina ducet’: sectus orbis is a division of the
world, even though sectus is a participle, not a noun; perhaps the
implication is ‘half the world’, just as at Stat. Theb. 8. 64 sectum . . .
annum means ‘half the year’. H is representing a use of the Greek ø
(Isoc. 4. 179 B ªB ±
Æ , ŒÆd B b #Æ B
¯Pæ ŒÆºı , Aristides 2, p. 60D e BÆ Œæ). Some-
times the ancients spoke of two continents (Varro, lL 5. 31, Plin. nat.
hist. 3. 5), sometimes of three (Sall. Jug. 17. 3); cf. Hardie (1986), 311 ff.
Those who add Africa include Pind. P. 9. 8, Cic. nat. deor. 2. 165 (see
Pease), Ov. fast. 5. 618 (on Europa) ‘parsque tuum terrae tertia nomen
habet’. ‘Europe’ as the name of the continent was known to Hecataeus
(cf. Herod. 4. 36. 2) and Aeschylus (Pers. 799, cf. 176 ff.), but its origin is
uncertain (Herod. 4. 45, RE 6. 1298 ff.).
nomina is a ‘poetic’ plural; cf. 4. 2. 3 ‘daturus nomina ponto’, Ov. trist.
1. 1. 90, Manil. 4. 609, H–Sz 16 ff. The word-order emphasizes tua. For
ducet cf. serm. 2. 1. 66 ‘duxit ab oppressa Carthagine nomen’.
2 8 . F ES T O QV I D P O T I V S D I E ?
[Th. Birt, Horaz’ Lieder, 1926: 27 ff.; A. Bradshaw ap. Woodman and Feeney (2002), 7 ff.;
J. Griffin, JRS 87, 1997: 58 ff.; Pöschl 180 ff.; E. A. Schmidt, Zeit und Form, 2002: 223 ff.;
D. West ap. Costa (1973), 43 f.; Williams 120 f.]
1–4. It is Neptune’s day, so bring out the Caecuban, Lyde. 5–8. There is no
time to lose. 9–12. We two shall sing in turn of Neptune and the Nereids, you
of Latona and Diana. 13–16. The final song will celebrate Venus and Night.
Following Hellenistic precedents (e.g. Theoc. 15), the poem is set at a

festival, presumably the Neptunalia on 23 July (Wissowa 225, RE 16.
2521 ff.). Though here he is accompanied by Hellenizing Nereids (10 n.),
the Roman Neptune was not originally a god of the sea; he was
associated rather with springing waters, hence his day belonged to the
hottest and driest time of the year. This was a suitable season for
drinking (cf. Alcaeus 347. 1 ff.), and as the day was a holiday in the
Roman calendar, the celebrations could begin soon after midday (5 n.).
Festus mentions ‘leafy huts’ in connection with the Neptunalia
(377M¼519L ‘umbrae vocabantur Neptunalibus casae frondeae pro
tabernaculis’); we may compare the festival of Anna Perenna when the
populace took to the Campus Martius and spent the day drinking and
love-making under similar screens (Ov. fast. 3. 527 ff.). But though the
Neptunalia provides the background for our poem, Horace’s celebration
is typically private and discreet.
Neptune received more solemn honours in Augustan Rome, because
he had assisted Octavian in his naval war against Antony; cf. Virg. Aen.
8. 699, Binder 243 ff., J. Pollini ap. Raaflaub–Toher (1990), 346 ff. In 25
bc Agrippa completed the Porticus Neptuni (Dio 53. 27. 1, Zanker 143),
where a famous fresco depicted the voyage of the Argonauts (cf. below
10 n. on the Nereids); the Basilica Neptuni may or may not have been a
separate structure (cf. Steinby 1. 182 f., 3. 341). This would have provided
a suitable occasion for the ode, but Horace says nothing about it
(contrast 1. 31, where the dedication of Apollo’s temple provides the
event to which Horace makes his private response). It has been thought
2 8 . F E S T O QV I D P O T I V S D I E ? 339
significant that according to the Athenian calendar Poseidon’s day was 8
December, Horace’s birthday (Bradshaw, loc. cit. 7–10), but it is unlikely
that the coincidence of dates would be generally known, and festo die
points to a Roman festival. Horace’s escape from drowning seems a
more cogent reason for his address to the god (cf. 3. 4. 28 n.), though
again nothing is said about it (contrast 3. 8, where he celebrates the
Matronalia as the anniversary of his escape from a falling tree).
We turn now to the poem’s mise-en-scène, which is not to imply that
it is an exact description of a particular occasion. The psaltria Lyde
seems to have been invited to Horace’s town-house, much as the Lyde
of 2. 11. 21 f. (there also a psaltria) is invited to Quinctius’ suburban horti
(‘quis devium scortum eliciet domo?’ with N–H); for musical girls of the
type cf. also 3. 9. 10, 4. 11. 34 f., RE 8. 2. 1341, Marquardt–Mau 151,
Griffin (1985), 26 ff. Some commentators picture the poet as dropping in
at Lyde’s own apartment; but in that case he could hardly have insisted
on a vintage Caecuban which he knows is well stowed away (2 f.). More
strangely, some regard Lyde as a hard-working housekeeper at the
Sabine villa (strenua in v. 3), disregarding the girl’s exotic name (see
below, 3 n.) and her elegant accomplishments. Birt (loc. cit.) thinks of
her as the hostess of a tavern (which might suit the wine-store in v. 6);
but Lyde has none of the robust vulgarity conventionally associated with
such women (cf. the copa Surisca of the Appendix Vergiliana), and a
tavern would be an odd place for the consummation implied at the end.
The ode is carefully articulated, and each stanza is self-contained.
The first two deal with the wine; in the other two Horace gives
directions about the songs (as in 1. 21). First he himself will sing about
Neptune (as befits the occasion), Lyde about the Nereids (9 n.). Then
Lyde will sing with musical accompaniment about Diana and her
mother Latona (11–12); these are appropriate subjects for a woman,
and balance Neptune and the Nereids (the huntress Diana is set in
opposition to the sea). The theme, however, seems surprisingly chaste
for a psaltria, and the implication may be that the sexual conclusion of
the entertainment is not to be rushed (for the reticence of the poem see
Griffin loc. cit.); RN is also interested in Pöschl’s idea (op. cit. 190 ff.)
that the song to the virgin Diana suggests Lyde’s initial hesitation, just
as the following song implies that she has given way. That song, in
honour of Venus and Night, shows that the ode’s intention is not
religious but erotic; the impersonal dicetur allows us to suppose that
both parties are now participating.
Metre: alternate Glyconics and Asclepiads.
1–2. Festo quid potius die / Neptuni faciam?: the ode begins abruptly
in the middle of a developing situation (as in 1. 27). festo and Neptuni are
both stressed: ‘What better can I do on a feast-day, and that the feast-
day of Neptune?’ (Darnley Naylor).
2–3. prome reconditum, / Lyde, strenua Caecubum: having aroused

the reader’s curiosity, H at once provides the appropriate answer:
cf. N–H on 2. 3. 6 f. ‘per dies / festos’, Tib. 2. 1. 29 f. ‘non festa
luce madere / est rubor, errantes et male ferre pedes’, Liv. 40. 14. 1 ‘non
est res qua erubescam, pater, si die festo inter aequales largiore vino
sum usus’.
Caecuban was one of the best Italian wines; cf. 1. 37. 5, 2. 14. 25 ff.,
epod. 9. 1 and 36, Plin. nat. hist. 14. 61, Athen. 1. 27a. reconditum suggests
that it was only brought out on special occasions (cf. 2. 3. 8 ‘interiore
nota Falerni’). The participle is typically balanced by the contrasting
prome, a natural word for bringing out wine (1. 36. 11, 3. 21. 8). D. West
(loc. cit. 43) sees military metaphors in these lines (see also 6 n. and 7 n.
below): prome reconditum could suggest the drawing of a sheathed sword
(the prefix meaning ‘out of the way’ as in re-motum rather than ‘again’),
and the wine would then be a weapon in the assault on the ramparts of
wisdom (4).
Lyde, like Lydia (1. 8. 1), is a Greek name, suitable for a meretrix.
strenua (‘energetically’) is predicative with prome; the adjective is gener-
ally avoided by the poets (Axelson 106), except for Horace’s hexameters
(4 times) and Ovid (7 times); it suits military behaviour, but is here used
amusingly in a hedonistic context. The suggestion that there is no time
to lose is common in sympotic verse; cf. 3. 29. 5 ‘eripe te morae’, 4. 12. 25
‘verum pone moras’.
4. munitaeque adhibe vim sapientiae: for the fortress of wisdom cf. 2.

6. 21 with N–H, especially Lucr. 2. 7 f. ‘munita tenere / edita doctrina
sapientum templa serena’. Porph. comments ‘Graeci aiunt vim sapien-
tiae adhiberi per vinum’; his precise authority is unknown, but cf.
Eubulus, PCG 5, fr. 133 e r H fiH æE KØŒE, Plin.
nat. hist. 23. 41 ‘in proverbium cessit sapientiam vino obumbrari’, Otto
372. For the thought that conventional good sense may be folly in a
symposium cf. 3. 19. 18 n. H’s exhortation applies to Lyde as well as to
himself; there is humour in ascribing sapientia to a girl of this type, and
some commentators, including RN, think there may be a suggestion
that her defences need to be broken down.
5–6. inclinare meridiem / sentis: strictly speaking, it is the sun that is

sinking, having passed its zenith (cf. Liv. 9. 32. 6 ‘sol meridie se
inclinavit’); elsewhere the day is the subject (Cic. de orat. 3. 17 ‘inclinato
iam in postmeridianum tempus die’, Tac. ann. 12. 39. 2), which justifies
here the use of meridies (medius þ dies). On working days the sympo-
sium did not normally begin till evening or near enough (N–H on 2. 7. 6);
so there is humour in saying inclinare meridiem rather than solem.
6. et veluti stet volucris dies: et has sometimes an adversative implica-

tion (‘and yet’), cf. TLL 5. 2. 893. 4 ff.; this nuance may be less natural
with the variant ac, though atque sometimes has such a sense. For ‘the
winged day’ cf. 4. 13. 14 ff. ‘quae semel / notis condita fastis / inclusit
volucris dies’, Thymocles, anth. Pal. 12. 32. 3 uæ P ›
Ø K
ÆNŁæØ ÆæŁØ ZæØ (so Marvell speaks of ‘Time’s winged chariot’).
The poet is vividly aware of the shortness of life and love (cf. Pöschl
183 f.); cf. 1. 11. 7 f. ‘dum loquimur, fugerit invida / aetas’, epod. 13. 3 f.
‘rapiamus, amici, / occasionem de die’, Catull. 5. 5. stet makes a charac-
teristic contrast with the juxtaposed volucris, and also balances inclinare;
for that contrast cf. Cic. de orat. 2. 187 ‘stantem inclinare’ (eloquence can
topple the man who is standing his ground).
West sees a further military metaphor; cf. 2. 7. 6 f. ‘cum quo mor-
antem saepe diem mero / fregi’ (which he does not quote). Admittedly
the image has changed from storming a citadel to attacking a battle-
line, and volucris in particular does not fit; but RN accepts West’s
defence that H is playing with a series of military terms rather than
aiming at a sustained image.
7. parcis deripere horreo: many editors take this as a statement, but a

question goes better with the urgency of the context. parcere with the
infinitive means ‘to refrain’; cf. 3. 8. 26 ‘parce . . . nimium cavere’, Milton,
Sonnet to Mr Lawrence, 13 f. ‘He who of those delights can judge, and
spare / To interpose them oft, is not unwise’; normally such restraint
would be commendable, but not on this occasion; cf. 3. 19. 21 f. ‘parcentis
ego dexteras / odi. sparge rosas.’ deripere continues the image of impa-
tient violence; Caecuban was usually treated with more circumspection.
West sees the amphora as an ally (which does not suit deripere), RN
rather as a weapon (cf. Virg. Aen. 10. 475 ‘vagina deripit ensem’).
A horreum is usually a barn for grain, but the word is sometimes used
of a wine-store; cf. Sen. epist. 114. 26 ‘aspice veteraria nostra et plena
multorum saeculorum vindemiis horrea’, TLL 6. 3. 2987. 23 ff. The word
suggests a large store-room rather than an apotheca in the roof (3. 8.
11 n.); deis said to imply the latter (cf. 3. 21. 7 descende, Colum. 1. 6. 20),
but need mean no more than ‘from’. Birt thought that the horreum was a
warehouse in the city where wine could be obtained (cf. 4. 12. 17 f.
‘cadum / qui nunc Sulpiciis accubat horreis’); but deripere does not suit
that idea.
8. cessantem Bibuli consulis amphoram?: M. Calpurnius Bibulus was

consul with Julius Caesar in 59 bc; so the wine is over thirty years old
and particularly fine. Bibulus was a notoriously ineffective consul (anon.
ap. Suet. Jul. 20. 2 ‘nam Bibulo fieri consule nil memini’), but a reference
to his obstruction of Caesar would now be remote; H may just be
playing on the man’s name. cessantem would suit any slacker (epist. 2.
2. 14), including a soldier who is hanging back; cf. 1. 35. 15 f. ‘cessantis ad
arma / concitet’, Plaut. Stich. 705 ‘sed interim, stratege noster, cur hic
cessat cantharus?’ It makes a contrast with strenua (3); cf. epist. 1. 2. 70
‘quod si cessas aut strenuus anteis’.
9–10. nos cantabimus invicem / Neptunum et viridis Nereidum

comas: nos means ‘we two’, as in vicem implies; the pronoun is emphatic
(cf. 11 n. on tu). H will sing of Neptune, Lyde of the Nereids; for the
distribution of themes between the sexes note especially 1. 21. 1 f. ‘Dia-
nam tenerae dicite virgines, / intonsum pueri dicite Cynthium’. Some
editors think that nos refers to Horace alone (cf. 1. 6. 5 and 17); this is
supposed to make a clearer contrast with ‘tu recines’ in v. 11, which again
has two objects. In that case invicem would be attached to the first of
two parallel clauses (cf. Ov. her. 17. 180 ‘inque vicem tua me, te mea
forma capit’); but it would be impossibly confusing to combine
this word-order with the singular use of nos (‘I, on the one hand,’).
Wickham interprets ‘We two shall sing in turn, I of Neptune and the
Nereids, you of Latona and Diana’, but such a change of direction is
hard to justify. The variant in vices is first securely attested in Ovid
(in the form ‘inque vices’); invicem is paralleled at 1. 25. 9, serm. 1. 3. 141
(‘inque vicem’).
The Nereids were often associated with the Greek Poseidon (cf. J. M.
Barringer, Divine Escorts, 1995: passim). In the Roman world they
belong to literature (Catull. 64. 15 ff.) and art rather than cult; they
were included in the group by Scopas in Domitius’ temple of Neptune
(Plin. nat. hist. 36. 26 ‘Nereides supra delphinos et cete aut hippocampos
sedentes’), and they surely found a place in the frescoes of Agrippa’s
Porticus Argonautarum. viridis is occasionally applied to the sea (Ov. ars
2. 92, 3. 130, ciris 461) and hence to the Nereids (Ov. her. 5. 57) and other
sea deities (Ov. trist. 1. 2. 59); cf. Gow on Theoc. 7. 59, H. Blümner, Die
Farbenbezeichnungen bei den römischen Dichtern, 1892: 215). For its use
here of the Nereids’ hair cf. Ov. met. 2. 12 ‘viridis siccare capillos’ (with
Bömer’s note). H may have in mind the colour not just of the sea but of
seaweed (cf. Sid. Apoll. carm. 2. 333 f. ‘harundinis altae / concolor in
viridi fluitabat silva capillo’).
11. tu curva recines lyra: Lachmann’s tum is supposed to get round the
awkwardness of contrasting nos and tu; but after the emphatic nos an
emphatic tu is positively desirable. recines must refer to a combination of
playing on the lyre and singing the themes described in v. 12; the
instrumental ablative can be defended by 3. 4. 1 ‘dic age tibia’ (see note ad
loc.). On the other hand, it is strange that Lyde should ‘answer’ an
amoebaean song in which she herself has been participating. Heinze
saw the difficulty when he interpreted ‘voce canes, lyra recines’ (i.e.
Lyde’s instrument answers her voice); but H’s words do not seem
explicit enough to bear this interpretation. RN has considered reading
curvae . . . lyrae; i.e. Lyde’s voice answers her instrument (a different sort
of alternation from that in v. 9). For the conventional curva of the lyre
(originally a hollow tortoise-shell) see N–H on 1. 10. 6.
12. Latonam et celeris spicula Cynthiae: Diana is called ‘Cynthia’ from

Cynthos, a small hill on Delos; this is the first attested use of the epithet
(Bömer on Ov. met. 2. 465), which is modelled on ‘Cynthius’, a title of
Apollo given currency in Hellenistic poetry (Call. h. 4. 10 etc.) and
popularized by Virgil (ecl. 6. 3). ‘Cynthia’ was not a girl’s name in Latin,
but was chosen by Propertius to indicate his mistress’s devotion to the
arts of Apollo; in this he was following Gallus, who had called his
mistress ‘Lycoris’ because of ‘Lycoreus’, another title of Apollo.
Diana is associated, as often, with her mother Latona (N–H on 1. 21.
1 n.); their statues stood on either side of Apollo in his Palatine temple
(N–H on 1. 31. 1, Zanker 240 f. with fig. 186). Her huntress’s arrows (1.
12. 22, Hom. Od. 6. 102 ` ! æØ . . . N
ÆØæÆ) make a formal contrast
here with the curved lyre (11). For such accoutrements in a hymn cf. 1. 21.
11, Eur. Hec. 464 f. #æØ ŁA =
æıÆ ¼ıŒÆ Æ Pºªø,
Call. h. 2. 19 (Apollo’s lyre and bow). Diana’s epithet celeris is also
attested at Ov. met. 4. 304, ciris 297; as ancient hunting was usually
done on foot (N–H on 1. 37. 18), she was conventionally dressed in a
short chiton for speed (LIMC 2. 1. 640 ff., 2. 2. 461 ff., Simon pls. 70
and 71).
13–14. summo carmine quae Cnidon / fulgentisque tenet Cycladas: the

verb is left unspecified until the impersonal dicetur (16). summo means
‘last’, as at epist. 1. 1. 1 ‘summa dicende Camena’. Cnidos in Caria, on the
tip of a peninsula south of Cos, had three shrines of Aphrodite, repre-
sented the goddess on its coinage, and was the home of the famous
statue by Praxiteles (LIMC 2. 1. 49 ff., 2. 2. 36 ff.); cf. 1. 30. 1 ‘regina
Cnidi Paphique’ with N–H. For the goddess’s association with the
Cyclades and especially Paros cf. RE 1. 2749 f., 18. 4. 1846 f.; fulgentis,
like nitentis in 1. 14. 19, describes the effect of the sun shining on Parian
marble, as it appeared to the poet’s imagination.
14–15. et Paphon / iunctis visit oloribus: after Cnidon the Greek form
Paphon is preferable to the even better attested Paphum; in spite of
Bentley the rhyme is unobjectionable (cf. 3. 25. 11 f. etc.). Paphos in
south-west Cyprus was where Aphrodite landed from the sea, after first
going to Cythera (Hes. theog. 192 f., h. Hom. 6. 3 ff.); the omission of her
name and the list of her cult-sites are ritual features, here used in a
light-hearted way (cf. Catull. 36. 11 ff.). Ancient deities were not omni-
present, and visere is regularly used for their visitations; cf. 1. 4. 8 with
N–H, Posidippus, anth. Pal. 12. 131. 1 ff. 8 ` ˚æ, – ˚ŁæÆ, ŒÆd L
1º KØ
E = . . . ºŁØ ¥ ºÆ ˚ÆººØfiø.
olor is the proper Latin word for a swan, though most poets found
the Greek cycnus more poetical and more tractable metrically in the
oblique cases. For Venus’ team of swans cf. 4. 1. 10 ‘purpureis ales
oloribus’, Prop. 3. 3. 39, Ov. met. 10. 708 with Bömer, Thompson
184 f.; evidence from Greek and Etruscan art is provided by LIMC
2. 1. 96 ff., 2. 2. 89 ff.
16. dicetur merita Nox quoque nenia: dicere is often used of singing
(OLD 7b); one assumes that H participates in the amorous conclusion.
Nox was often personified by the poets, but there was no official cult
(as opposed to magic); cf. epod. 5. 51 with Mankin, Virg. Aen. 6. 249 f.,
7. 138, Ov. fast. 1. 455 with Bömer ad loc., RE 17. 1229, 1663 ff.
The reference to Night balances meridiem (5), and suits the closure of
the poem (cf. Virg. ecl. 1, 3, 7, 9, 10, Curtius 90 f., Barbara H. Smith,
Poetic Closure, 1968: 176). For similar erotic endings cf. Prop. 2. 28. 62
‘votivas noctes et mihi solve decem’, 3. 10. 31 f. ‘annua solvamus thalamo
sollemnia nostro / natalisque tui sic peragamus iter’. Pöschl 185
following ps.-Acro sees a reference to the night of death, but this is
incompatible with the mention of Venus and quite contrary to the spirit
of the poem.
nenia suggests primarily a repetitious chant (pace J. L. Heller, TAPA
74, 1943: 215 ff.); it can refer to a dirge (N–H on 2. 1. 38 and 2. 20. 21),
an incantation (epod. 17. 29 ‘caputque Marsa dissilire nenia’), a jingle
of children at play (epist. 1. 1. 63), a fabulist’s ditties (Phaedr. 3 pr. 10,
4. 2. 3). In our passage it seems to mean a lullaby; cf. Arnobius nat. 7. 32
‘lenes audiendae sunt neniae’. It is objected that Horace is not thinking
of sleep, but one can argue that after the love-making (represented
by the invocation of Venus in 13 ff.) Night is honoured by a lullaby
in gratitude for her services; cf. Tib. 2. 1. 87 ff. (after an invitation
to Amor) ‘ludite: iam Nox iungit equos . . . / postque venit tacitus
furvis circumdatus alis / Somnus’, Virg. Aen. 8. 405 f. Heinze thought
that in our passage nenia meant the coda of a song (thus Paul. Fest.
163M ¼ 155L, and Diomedes, GL 1. 485. 6 ff.); but this does not explain
the other usages of the word. The misunderstanding probably arose
from such passages as Plaut. Pseud. 1278 ‘id fuit nenia ludo’—a
humorous expression which means simply ‘that was the death-dirge of
the show’.
2 9 . T Y R R H ENA R E G V M P RO G EN I ES
[ J.-M. André, Mécène, 1967; G. Davis 172 ff.; Fraenkel 223 ff.; Lyne (1995), 111 ff., 132 ff.;
Pasquali 329 ff., 635 ff.; Pöschl 198 ff.; N. Rudd in Horace 2000, 1993: 64 ff.; E. Zinn ap.
Oppermann, 1972: 225 ff.]
1–16. I have everything ready, Maecenas, for your enjoyment; tear yourself
away from the business that detains you, and do not simply gaze from afar at
the countryside. Leave your towering mansion and the bustle of Rome: a
simple meal in a modest home relieves worry. 17–28. Now is the hottest time
of year and all the countryside is at rest; but you keep worrying about Rome’s
constitution and the intentions of far-off peoples. 29–48. The future is
unknowable, so calmly set at rest the matter of the moment; everything else
sweeps past as unpredictably as the Tiber. Treat each day as the last; even if
tomorrow brings misfortune, nothing can annul past blessings. 49–64. If
Fortune flies away, I renounce what she has given and woo Poverty. If a
storm rages, unlike the rich merchant who bargains with the gods in the hope
of saving his possessions, I shall make land in my lifeboat with the help of the
Dioscuri.
This is the penultimate poem of Odes I–III before the personal

declaration or sphragis (3. 30), and as such it has a particular significance.
Like 1. 1 it is addressed to Maecenas, thus framing the collection (see 1 n.
for a parallel reference to his Etruscan ancestry); but Horace now writes
with greater confidence and authority. There is a similar arrangement in
the first book of Epistles, where the first and nineteenth poems are
addressed to Maecenas, and the twentieth poem is again a sphragis; in
the first book of Odes the penultimate poem is the important ode on
Cleopatra, and it is followed only by a personal tail-piece. Our ode is
likely to have been written not long before the collection was issued in
23 bc; the reference to Maecenas’ anxiety for the city suggests a time
before Augustus’ return from Spain in 24 (26 n.). In 1. 1 Horace had
already acknowledged Maecenas’ support and protection, and looked
forward eagerly to gaining his approval for a new poetic venture; here he
turns once more to address him, using this as an occasion for the most
memorable statement of his thoughts about friendship and life.
The ode has some of the characteristics of an invitation-poem,
though it develops into something much deeper and more complex;
for the type see N–H on 1. 20, Courtney on Juv. 11, I. M. Le M. Du
Quesnay, PLLS 3, 1981: 90 ff. with n. 345. Horace offers his guest mellow
wine (2 n.), summer roses (3 n.), and exotic perfume (4 n.); this coheres
with what is known about Maecenas’ luxurious tastes, particularly from
Seneca’s unsympathetic account (epist. 114, André, op. cit. 19 ff.).
A teasing contrast is drawn between his towering mansion on the
Esquiline (10 n.) and simple dwellings where there are no elaborate
meals or purple furnishings (14–15 n.); Horace is thinking of his own
Sabine villa (as is confirmed by the pastoral references in 21 ff.), and is
making the comparison between grandeur and modesty that was con-
ventional in such poems. The reference to hot weather is appropriate in
sympotic verse (20 n.), and the astronomical detail may have struck a
chord with Maecenas (17–18 n.). The exhortation to stop thinking about
remote matters (26 ff.) suits the type (3. 8. 17 ff., 3. 19. 1 ff.); but again it is
especially relevant both to Maecenas’ political responsibilities and to his
personality (for his persistent worrying see 16 n.).
What distinguishes the ode from most invitation-poems is its sus-
tained ethical dimension. Horace derived his wisdom from various
sources (Rudd, op. cit.); thus in our poem potens sui (41) has a Stoic
ring, and some of the illustrations draw on popular moralizing associ-
ated with Cynics and others (see 54 n. on the cloak of Virtue, 55 n. on the
wooing of a personified Poverty, 62 n. on the metaphor of big and little
ships). But the tone is predominantly Epicurean (Pasquali 639 f.), even
if some of the ideas were shared by other schools (see 41–2 n. and 45–6 n.
for the Stoic Seneca). Here we may mention the desirability of occa-
sional relaxation (5 n., 13 n.), the inner peace (IÆæÆÆ) afforded by
simplicity (14 n.), the pointlessness of worrying about what may never
happen (27–8 n.), the impossibility of foretelling the future (29–30 n.),
the absurdity of prayer in storms at sea (58–9 n.). Above all, the moral
exhortation (paraenesis) at the centre of the poem expounds the attitude
to time that lies at the heart of Epicurean ethics: the way to achieve
happiness and inner peace is to concentrate on the present (32–3 n.),
regarding life as complete as each day ends (41–2 n.); no matter what the
future may bring, nothing can take away the good things that have
happened in the past (45–8 nn.).
Some have supposed that Maecenas was himself an Epicurean
(André, op. cit. 15 ff., rejected by Lyne, op. cit. 114 ff.). It is no objection
to this view that he was very rich, or that he took part in politics;
Epicurus sanctioned such participation in times of crisis (Cic. rep. 1. 10,
Sen. dial. 8. 3. 2), and several Roman Epicureans were public figures,
like Piso (cos. 58), Memmius the patron of Lucretius, and Cassius the
tyrannicide (A. Momigliano, JRS 31, 1941: 157 ff.). But it is hard to
believe that the extravagant and neurotic voluptuary described by later
writers (3–5 n., Sen. epist. 114) was in any significant sense an adherent of
the school; at the most he may have justified his own hedonism by
misappropriating a few aphorisms (that would be enough to suit the
principle that Horace’s advice is likely to be acceptable to the recipient).
Though Horace is grateful to his friend for the Sabine estate and all that
it implies in the way of personal freedom, he knows him well enough to
2 9 . T Y R R H ENA R E G V M P RO G EN I ES 347
draw a contrast between the happiness that his own simplicity has
brought him and the anxieties that Maecenas’ luxurious life-style has
failed to alleviate (14–15 n.). When he proclaims his self-sufficiency at
the end of the poem (53–64) some see this as a subtle assertion of
independence from his benefactor (54 n., Lyne, op. cit. 115); but such
an innuendo would hardly be appropriate to an ode of friendly support
in which Horace invites Maecenas to join him at the Sabine villa. At the
end H is claiming for himself the privilege of every wise man (cf. epist. 1.
16. 79); in the last resort even Maecenas was free to say
ø PŒ
ÆØ
The context of epist. 1. 7, written a few years later, is quite different.
The movement of the ode crosses schematic divisions. Thus the
generalization in 13–16, with its contrast of wealth and simplicity, is
related to the same contrast in 53–61. A violent element (heat) in 17–20 is
balanced by a violent element (water) in 36–41; the peaceful scene with
its trees and animals in 21–4 is contrasted with the terrible scene in 36–41
where the once calm river sweeps away trees and animals. The central
advice in 32 extends into the next stanza (‘quod adest memento /
componere aequus’); the flying hour in 48, which betokens rapid change,
foreshadows winged Fortune in 53–4; and indulgent but capricious
Fortune (53–4) is succeeded by the reliability of Poverty (54–7); the
poet’s indifference to wealth (54–6) leads smoothly into the image of
his self-sufficiency (62–4). A series of antitheses runs through the poem:
business/leisure, wealth/poverty, heat/coolness, violence/peace, anxiety/
calm, danger/safety. Several words are used to describe nature which are
also appropriate to human beings: furit and vesani of stars (18 f.),
taciturna of a bank, vagis of breezes (24), pace of a river (35), clamore of
mountains and vicinae of woods (39), fera and irritat of a flood, quietos
of streams (40 f.), avaro of the sea (61), tumultus of the Aegean (63).
Throughout we find the Greek past alive in the Roman present, an
inimitable blend of grandeur and intimacy, solemnity and humour; a sad
awareness of transience and insecurity combined with a tough-minded
intention to survive. Only 3. 4 can match this ode for the breadth and
majesty of its scope.
Metre: Alcaic.
1. Tyrrhena regum progenies: Maecenas claimed descent on his

mother’s side from the Cilnii, who had once been the ruling house in
Arretium (Arezzo), his home town in eastern Etruria; cf. 1. 1. 1 ‘Maece-
nas atavis edite regibus’ (where ‘kings’ may be an overstatement), serm.
1. 6. 1 ff., Prop. 3. 9. 1 ‘Maecenas, eques Etrusco de sanguine regum’, Liv.
10. 3. 1 ‘Cilnium genus praepotens’, Sil. 7. 29, Syme (1939), 129 n. 4,
Lyne, op. cit. 132 f. Such claims might mean much to some (cf. Pers. 3.28
‘stemmate quod Tusco ramum millesime ducis’, Flower 212, 348 f.), but
counted for little in the official hierarchies of Rome; here with a typical
blend of flattery and teasing H magnifies the ancestry of his friend. His
phrase is even more grandiloquent than the first line of Book 1 which it
balances: the Greek Tyrrhena is used for the normal Latin Etrusca;
progenies (‘scion’) is rather stately when applied to an individual (4. 15.
32, Catullus 34. 6, Val. Max. 4. 3. 4 ‘Anci regis clara progenies’); the
transferred epithet is another mark of the high style (cf. Virg. Aen. 8. 526
‘Tyrrhenusque tubae clangor’, Prop. loc. cit., H–Sz 160).
1–2. tibi / non ante verso lene merum cado: the emphatic tibi, begin-
ning a new clause, means ‘all for you’ (cf. 2. 7. 20 ‘cadis tibi destinatis’).
Maecenas listed the qualities of wine in his Symposium (see introduction
to 3. 21) and was doubtless a connoisseur (1. 20. 9 f., Plin. nat. hist. 14.67).
A cadus (Œ ) was a large storage vessel (Blümner, 1911: 151, Hilgers
125 ff.); the everyday word was common in comedy and inscriptions and
occurs ten times in the Odes, but it is also attested in Virgil (Aen. 1. 195,
6. 228). Wine was drawn by tilting the jar; cf. serm. 2. 8. 39, Plaut. Stich.
721 ‘quamvis desubito vel cadus vorti potest’, Enn. ann. 532 ‘vortunt
crateras aenos’. The insertion of the subject between the two parts of the
ablative absolute is not a Ciceronian construction but is common in
Livy (H–Sz 402, Nisbet in Adams–Mayer 150 f.).
It has been suggested that the wine as yet unused is a metaphor for
the philosophical content of H’s verse epistles (P. L. Bowditch, Horace
and the Gift Economy of Patronage, 2001: 178 ff.). Such interpretations,
now fashionable, are occasionally convincing (at 1.26 the garland is the
poem), but here the rose-blossoms and hair-oil suit Maecenas’ style in
life and poetry rather than Horace’s; see further Nisbet (1995), 416 ff.
3. cum flore, Maecenas, rosarum: rose-blossoms were worn in garlands

(2. 3. 14, 3. 15. 15, Simon. PMG 506.2), and their petals were sometimes
scattered at the symposium (3. 19. 22); roses were spring flowers in Italy
(N–H on 1. 38. 4), and though they would last longer in the Sabine hills,
in the summer their presence implies luxury. Here the roses are first and
foremost roses; they do not ‘stand for’ anything else; but as so often they
carry sweet associations of transient beauty and pleasure; cf. 2. 3. 13 ff.
and M. Ferber, A Dictionary of Literary Symbols, 1999: 172–7. The
separation of Maecenas from a previous attribute is again a mark of
high style; cf. 3. 21. 1 ff. in a pseudo-hymn, Fraenkel 222. The insertion
of the vocative between flore and rosarum may serve to recall his extrava-
gant tastes; cf. ‘dulci, Tyndari, fistula’ with Fraenkel 206, A. La Penna,
Mnemosynum, Studi in onore di A. Ghiselli, 1989: 335 ff.
3–5. et / pressa tuis balanus capillis / iamdudum apud me est: balanus

was the fruit of an Egyptian palm from which a perfumed oil was
crushed; cf. Plin. nat. hist. 12.100–2, Scrib. Larg. 129, 131. For hair-oil at
symposia cf. 1. 4. 9 with N–H, 2. 7. 7 f.; Cicero regarded it as effeminate
(Pis. 25 ‘madentes cincinnorum fimbriae’). When Augustus teased
Maecenas for his myrobrechis cincinnos (Suet. Aug. 86. 2), he was prob-
ably alluding to more than just his literary style; for his luxurious habits
see Mayor on Juv. 1. 66, RE 14. 1. 214, and for his alleged work de cultu
suo see Sen. epist. 114. 4, Schanz–Hosius 2. 20 f. iamdudum adds
urgency to the appeal; cf. epist. 1. 5. 7 ‘iamdudum splendet focus et tibi
munda supellex’.
5. eripe te morae: for se eripere cf. Sen. epist. 19. 1 ‘subduc te istis
occupationibus, si minus eripe’, Sil. 2. 564. morae means ‘business that
causes delay’; cf. 4. 12. 25 (also an invitation), ‘pone moras’, OLD 8. The
dative is regular with verbs of ‘taking away’ (3. 8. 11–12 n.), and is
extended from persons to things by the poets, Livy, and later prose-
writers. There seems to be a hint of the Epicurean doctrine that life is
frittered away by the postponement of enjoyment; cf. sent. Vat. 14 › b
ººfiH ÆæÆ ººıÆØ, Sen. dial. 10. 9. 1, Pöschl, op. cit. 207 n.
6–7. nec semper udum Tibur et Aefulae / declive contempleris arvum:

the present subjunctive in a prohibition is common in early Latin but
otherwise rare (Handford 47, H–Sz 336); yet in Horace cf. 1. 33. 1 ‘Albi,
ne doleas’, which is sometimes mistakenly regarded as a purpose clause,
2. 11. 4 f. ‘nec trepides in usum / poscentis aevi pauca’, serm. 2. 3. 88 ‘ne sis
patruus mihi’; the construction suits a continuing state of affairs (note
semper) rather than a single occasion (for which ne with the perfect
subjunctive would be appropriate). Syndikus points out (2. 236 n. 13)
that along with eripe a prohibition is wanted to balance the two impera-
tives in 9 and 11 (desere, omitte); in particular ne contempleris (a word
derived from augury) suits serious contemplation and so balances omitte
mirari (11). The MSS divide between nec and ne; but as the latter might
be thought to introduce a purpose clause we have preferred nec, even
though this involves losing a parallel with the asyndeton at vv. 10–11.
From the eastern side of the Esquiline, where Maecenas had his
mansion, one could look across to Tibur, Praeneste, and Tusculum; cf.
Strabo 5. 3. 11–12. For the attraction of vistas cf. epist. 1. 10. 23 ‘lauda-
turque domus longos quae prospicit agros’, 1. 11. 26, Sen. epist. 89. 21,
Stat. silv. 2. 2. 72 ff., Mart. 4. 64. 9 ff., RE 16. 2. 1858 f. For Tibur see 3. 4.
23 n.; udum alludes not just to the spectacular waterfalls but also to the
well-irrigated orchards (cf. 1. 7. 13 f. ‘uda / mobilibus pomaria rivis’, 4. 2.
30 f., 4. 3. 10, Ov. fast. 4. 71). Since the text does not assign any special
prominence to Tibur (if there were any difference, Tusculum would
carry the most emphasis), there is no justification for assuming that the
invitation comes from there. It is not known when Horace acquired the
house mentioned by Suetonius (vita); on the other hand, parvo sub lare
(14) is in line with H’s other references to his Sabine property (e.g. 2. 16.
37, 2. 18. 14, serm. 2. 6. 1 ff.), and vv. 21–4 indicate a rural setting.
Maecenas may have had properties at the three places mentioned, but
that has nothing to do with Horace’s invitation.
Aefulae is the reading of H’s MSS and is supported by CIL 14. 3530
‘sub monte Aeflano’ (see further E. Hübner, Hermes 1, 1866: 426); old
editions read Aesulae, which appears in MSS of Velleius and Pliny. The
height of the place is confirmed by Liv. 26.9.9 (on the approach of
Hannibal) ‘praesidia etiam in monte Albano atque arce Aefulana
ponuntur’. It seems to have been situated south of Tibur and east of
Hadrian’s villa; to the bibliography provided by Briscoe on Liv. 32. 29. 2
add Encicl. oraz. 1. 491 f.
7–8. et / Telegoni iuga parricidae: Tusculum, about 15 miles south-east

of Rome, was situated above Frascati at a height rising to 2,200 ft.; many
notables had mansions there, including Lucullus, Cicero, and Gabinius
(RE 7A. 2. 1487 f.). The place claimed as its founder Telegonus, son of
Ulysses and Circe; cf. epod. 1. 30 ‘Circaea . . . moenia’ (also addresssed to
Maecenas, so possibly he too had property there), Liv. 1. 49. 9 with
Ogilvie, Prop. 2. 32. 4, Dion. Hal. ant. Rom. 4. 45. 1, CIL 14. 2649 (the
base of a statue of Telegonus in the theatre); for similar legends see 3. 17.
6 n. According to the story as told in the Niptra of Sophocles and
Pacuvius, Telegonus unwittingly killed his father; cf. Opp. hal. 2.
497 ff., Frazer on Apollod. epit. 7. 36, J. L. Lightfoot, Parthenius of
Nicaea 386, RE 5A1. 314 ff. Eustathius similarly characterizes Telegonus
as › Łæıº Ææ (Od. p. 1660. 5); in our passage there
may be a hint of mischievous disrespect to counter the standard eulogies
of Tusculum.
9. fastidiosam desere copiam: for the paradox that abundance brings

distaste cf. Plaut. trin. 671, Liv. 3. 1. 7 ‘fecit statim ut fit fastidium copia’,
Plin. nat. hist. 12. 81. fastidiosus is transferred from the person who feels
disgust to the thing that causes it; cf. 2. 7. 21 ‘oblivioso . . . Massico’, Sen.
contr. 9. 2. 4 ‘fastidiosos ob ebrietatem cibos’. Such hypallage is very
common in Greek and Latin poetry; see Bell 315 ff. (especially 324 ff. for
the transfer from persons to things), H–Sz 160.
9–10. et / molem propinquam nubibus arduis: Maecenas’ skyscraper is

also mentioned in Porph.’s note on v. 6 ‘turrim Maecenas dicitur in
hortis suis extruxisse’, epod. 9. 3 ‘sub alta . . . domo’ (already in 31 bc),
Suet. Ner. 38. 2 ‘hoc incendium e turre Maecenatiana prospectans’; cf.
Y. Perrin, Latomus 55, 1996: 399 ff., RE 14. 215 f., Steinby 3. 71. Elsewhere
H is more critical of regiae moles (2. 15. 1 f.); even here he is no doubt
making a smiling contrast with his own modest villa (3. 1. 45 f.), but it
goes too far to see an allusion, however humorous, to the perils of
greatness (thus Lyne, 1995: 112 citing 2. 10. 10 f. ‘celsae graviore casu /
decidunt turres’). For the hyperbole of propinquam nubibus cf. Virg. Aen.
4. 89, 8. 99 f. ‘tecta vident quae nunc Romana potentia caelo / aequavit’,
Fraenkel 225 n. 3, Sen. Ag. 92 with Tarrant, Hardie 271 ff., Shakespeare,
The Tempest i v . i. 152 ‘The cloud-capped towers, the gorgeous palaces’.
With arduis the clouds are perhaps visualized as heights that even
Maecenas cannot quite reach.
11–12. omitte mirari beatae / fumum et opes strepitumque Romae: for

omitte (‘cease’) cf. epist. 1.18.79 ‘quem sua culpa premet, deceptus omitte
tueri’, Plaut. Pers. 431 ‘iam omitte iratus esse’, OLD 5b. The first syllable
of the third line in an Alcaic stanza is short 11 times in Horace (to the
list in Bo 37 add 2. 15. 15). mirari means ‘to look with awe upon’ (cf. 3. 25.
14), an inappropriate attitude to material objects; cf. epist. 1. 6. 1 ‘nil
admirari’ (b ŁÆıØ).
beatae gains force from the hyperbaton; Rome was obviously ‘rich’
and in the popular view (not wholly shared by the poet) ‘blessed’; for
such ambiguities cf. N–H on 2. 2. 18. The city was smoky even in
summer because of the need to cook (cf. Juv. 8. 8 with Courtney,
Rutil. Namat. 1. 193 f.). H interposes opes, which most people thought
desirable, between fumum and strepitum, which were clearly objection-
able; for the strategy cf. Johnson, The Vanity of Human Wishes 160 ‘Toil,
envy, want, the patron, and the jail’. Tennyson avoids any ambiguity
about wealth when he writes ‘Far from noise and smoke of town’ (To
the Rev. F. D. Maurice 14); for his imitation of the ode see Rudd,
Hermathena 150, 1991: 5 ff.
13. plerumque gratae divitibus vices: understand sunt, not explicuere

(16): H proceeds from a general aphorism to its particular application
(Heinze). plerumque suits such a maxim; the word is avoided by most
Roman poets (except significantly Lucretius), but is found also at 1. 34.
7, 3. 21. 14, Virg. georg. 1. 300 (Axelson 106). vices can refer not just to the
process of change but also its result. The word is neutral and derives its
colour from the context; cf. 1. 4. 1 ‘grata vice veris et Favoni’, culex 211
‘acerbas cogor adire vices’. The converse of H’s point is found in
Epicurus, epist. 3. 131: luxuries should be approached KŒ ØÆºØø
(‘after intervals’).
14–15. mundaeque parvo sub lare pauperum / cenae: mundus describes

a simple elegance; cf. 1. 5. 5 ‘simplex munditiis’ with N–H. For mundae
cenae cf. Milton, To Mr Lawrence 9 ‘neat repast’. Here the idea of
simplicity predominates, as the word reinforces parvo and pauperum as
against aulaeis et ostro (15); for the same emphasis cf. serm. 2. 2. 65
‘mundus erit qua non offendat sordibus’ (simple without being mean).
The adjective is avoided in the higher styles of poetry (Axelson 106 f.);
so is cena (ibid.), which is also found at 2. 14. 28. pauperum as usual
describes a lack of riches rather than destitution (3. 2. 1 n.); for such
modesty in an invitation poem cf. 1. 20. 1 f., 4. 12. 24, epist. 1. 5. 1 f.,
Philodemus, anth. Pal. 11. 44. 1 f. (see further Hollis on Call. Hecale, pp.
349 f.). Here as in 3. 16. 39 ff., without envy or reproach, H explicitly
contrasts his life-style with that of Maecenas; an implicit contrast
emerges in 2. 18 and in RN’s view at 3. 1. 33–48.
15. sine aulaeis et ostro: aulaea usually refers to curtains hung vertically,
whether in the theatre, on the walls of a room, or on rails going round
the dining-couches (Curt. 8. 5. 21 ‘cum post aulaea quae lectis obduxerat
staret’, 9. 7. 15 ‘lectis circumdederat aulaea purpura auroque fulgentia’);
but at the cena Nasidieni they were canopies above the table (serm. 2. 8.
54 ff. with Porph., Blümner, 1911: 145). aulaeis et ostro is probably a
hendiadys for ‘purple hangings’ (cf. Virg. georg. 2. 192 ‘pateris libamus
et auro’). ostrum was the dye (of various shades from purple through
crimson to scarlet) produced from shell-fish, notably in Tyre and Sidon
(3. 1. 42 n.). It is relevant that Maecenas the dandy seems to have had a
liking for the colour (cf. Juv. 12. 38 f. ‘vestem / purpuream teneris quoque
Maecenatibus aptam’); such luxury suited the Etruscan stereotype (N–
H on 2. 18. 8).
16. sollicitam explicuere frontem: cf. serm. 2. 2. 125 ‘explicuit vino

contractae seria frontis’; explicuere is a gnomic perfect (3. 23. 19 n.). The
lines lead up to the cares of state (25–8); Maecenas was an inveterate
worrier (cf. 2. 17. 1 ff. with N–H, vol. 2, pp. 273 f., André, op. cit. 32 ff.). It
was a central Epicurean doctrine that anxiety is not banished by riches,
and that happiness may be found in simplicity; see 2. 16. 9 ff. with N–H,
3. 1. 17 ff. with the introduction to that poem, and Lucr. 2.1–61.
17–18. iam clarus occultum Andromedae pater / ostendit ignem:

Cepheus, king of the Ethiopians, was the father of Andromeda and
husband of Cassiepia; he offered his daughter up to a sea-monster from
which she was rescued by Perseus (cf. Frazer on Apollodorus 2. 4. 3,
Bömer on Ov. met. 4. 663 ff., LIMC 1. 774 ff., 2. 622 ff., RE 11. 1. 223 f.).
For Cepheus’ constellation cf. Roscher 6. 884 ff., Cic. nat. deor. 2. 111
with Pease, Le Boeuffle 125 ff. Its ‘evening rising’ was placed by Colu-
mella on 9 July (11. 2. 51); but such calculations must have been based on
Alexandrian astronomy, for in the latitude of Rome it never completely
set (Housman on Manil. 5. 449, D. R. Dicks, Hermes 91, 1963: 63 n. 4).
It may be conjectured from the astrological references in 2. 17 that
Maecenas had some interest in astronomy, but there as here H makes
grandiose references without caring for precision. Pöschl (op. cit. 212 n.
18) rightly rejects the view that pater refers to Jupiter and ignem to the
constellation of Andromeda (thus W. Riedel, Philologus 95, 1943: 308 f.).
ignem describes the fiery material of Cepheus, which was thought to
produce the summer heat; occultum refers to its previous supposed
disappearance, and makes a contrast with clarus and ostendit; for the
latter cf. Catull. 62. 7 ‘nimirum Oetaeos ostendit Noctifer ignes’, Virg.
georg. 4. 232, which confirms that the subject of ostendit in our passage is
the constellation itself rather than Jupiter.
18. iam Procyon furit: Procyon was the brightest star in Canis Minor
(Pease on Cic. nat. deor. 2. 116, Roscher 6.1002 ff., RE 23. 1. 613 ff.);
the Greek name, which means ‘the one that comes before the Dog’, is
translated by Cicero as Antecanis (Arat. 222). Its ‘morning rising’ is given
by Columella as 15 July (11. 2. 52), by Pliny as 4 July for Egypt, 17 July for
Assyria (nat. hist. 18. 268–9). furit, like vesani, suits a rabid animal.
19. et stella vesani Leonis: not the constellation of Leo as a whole,

but its brightest star, the so-called stella regia (Plin. nat. hist. 18. 235,
Housman 3. 907, Le Boeuffle 164); Columella dates its rising to 29 July
(11. 2. 53 ‘Leonis in pectore clara stella exoritur’). The sun enters Leo on
20 July (Colum. 11. 2. 52) heralding the hottest season of the year; so the
constellation is often given epithets like furibundus and rabidus (for a list
see RE 12. 1981 f.). The ancients thought of constellations as gigantic
figures in the night sky; cf. the charts in Goold’s Loeb edition of
Manilius and, for general information, RE 3A2. 2412–39. The insepar-
able prefix ve- can either negate or intensify (Gell. 16. 5. 5, H–Sz 257); cf.
vecors, vegrandis.
20. sole dies referente siccos: the prominently placed siccos includes
human thirst, a point that suits H’s invitation; cf. 4.12.13 ‘adduxere sitim
tempora, Vergili’ (there of the spring), Hes. op. 588 f. (of the Dog-Days),
Alc. 347. 1 f. ªª ºÆ Yfiø, e ªaæ ¼æ æØººÆØ, = . . .
Æ b łÆØ Pa ŒÆÆ (the star is Sirius), Theog. 1039 f. ¼æ
¼ŁæøØ ŒÆd ØØ, YØ r = c ı ¼æı ŒÆd Œıe
Iæ
ı.
21–2. iam pastor umbras cum grege languido / rivumque fessus

quaerit: for the anaphora of iam in identifying seasons or times of day
cf. 1. 4. 5 with N–H, Sen. apocol. 2. 1, Juv. 4. 56 f. (mock-heroic); here the
word adds to the urgency of iamdudum (above, 5). Descriptions of
summer heat naturally mention water, shade, and repose: see for in-
stance epod. 2. 23 ff., Theoc.1. 15 ff. with Gow, Virg. ecl. 2. 8, georg.
3. 327 ff., culex 104 ff., anth. Plan. 227. 5 ff. At the same time the heat
of the day can be thought of as nature’s parallel to emotional strain,
while the shade corresponds to Epicurus’ IÆæÆÆ; cf. OLD umbra 5,
P. Smith, Phoenix 19, 1965: 298 ff., T. Rosenmayer, The Green Cabinet,
1969: ch. 4.
22–4. et horridi / dumeta Silvani caretque / ripa vagis taciturna ventis:

dumeta is the subject of carent understood, not the object of quaerit;
for the word-order cf. 3.4. 11 n. In the first part of the stanza (framed
by pastor . . . quaerit) umbras and rivum are joined by -que; in the second
part dumeta and ripa are joined by another -que; dumeta answers to
umbras, and ripa to rivum. The prickly scrub would not have provided
a comfortable resting-place for the pastor, but it was enough for his
animals; cf. culex 154 ff. ‘at circa passim fessae cubuere capellae / excelsis
subter dumis, quos leniter adflans / aura susurrantis poscit confun-
dere venti’.
dumeta denotes the thick scrub (macchia, maquis) that is such a feature
of the south European landscape (Horden–Purcell 182 ff.). Silvanus was
an old Italian deity of the countryside, more benign than Pan, with
whom he was sometimes associated (Wissowa 213 ff., P. F. Dorcey, The
Cult of Silvanus, 1992, Simon 200, LIMC 7. 1. 763 ff.). horridi alludes to
both his appearance (Martial 10. 92. 6) and his habitat (Virg. Aen. 8. 348
‘silvestribus horrida dumis’). taciturnus, properly applied to people, is
transferred to the river-bank (cf. 1. 31. 8 ‘taciturnus amnis’ with N–H);
there is not even the usual wandering breeze to rustle the vegetation
(contrast culex 156 above).
25–6. tu civitatem quis deceat status / curas: tu emphasizes that

Maecenas is out of step with the natural order; curas balances times
and here implies anxiety as well as attention (cf. 3. 8. 17 ‘mitte civilis
super Vrbe curas’). It looks as if Maecenas was taking an important part
in the deliberations that led to the settlement of 23 bc, by which
Augustus based his rule on imperium proconsulare and tribunicia potestas
(Lacey 110 ff., 154 ff., J. A. Crook, CAH 10 edn. 2. 86 ff., 117 ff.); as the
following clause implies that Augustus had not yet returned from Spain
(24 bc), the same must be true of this. It may be relevant that in the
context of 29 bc Dio assigns a long speech to Maecenas advocating
monarchical government (52. 14–40); the details are largely anachronis-
tic (F. Millar, A Study of Cassius Dio, 1964: 102 ff.), but there may be a
substratum of truth (André, op. cit. 78 f.). Our passage at any rate
implies that Maecenas had views on constitutional questions.
26. et Vrbi sollicitus times: Maecenas had previously been left in

charge of Rome and Italy after Actium (Dio 51. 3. 5); he must also
have had exceptional influence during Augustus’ absence in the West
(27–24 BC), though Porph. is presumably wrong when he says he was
prefect of the city. sollicitus picks up sollicitam (16); for the vigilance of
Maecenas cf. the introduction to 3. 8. The dative should be taken
certainly with times, perhaps also with sollicitus, but not with the plots
of the Chinese mentioned below.
27–8. quid Seres et regnata Cyro / Bactra parent Tanaisque discors: as

Seres is a Greek plural, the second e is short. The threat from the
Chinese was a hyperbolical fantasy, as was talk of conquering them
(1. 12. 56 with N–H); for knowledge of China see Thomson 32 f., RE
2A. 1678 ff., OCD 1392 f. Bactra (neut. pl.) denotes both the country of
Bactria (roughly northern Afghanistan) and its capital (Prop. 4. 3. 63,
where it is part of the Parthian empire; cf. 3. 1. 16 ff.). To make it seem
more remote and irrelevant H describes it as ‘ruled by Cyrus’ (for the
grandiloquent phrase cf. 2. 6. 11 ff. with N–H). The sixth-century Cyrus
the Great was the subject of much hagiography, notably in Xenophon’s
Cyropaedia (N–H on 2. 2. 17); according to that work he conquered
Bactria (1. 1. 4), but most that was known about the country came as a
result of Alexander’s expedition; see further W. W. Tarn, The Greeks in
Bactria and India, 1951, DNP 2. 414 ff., OCD 231.
The Tanais or Don (3. 10. 1) here stands for the Scythians; cf. 2. 9. 21
‘Medumque flumen gentibus additum’ with N–H, Virg. georg. 1. 509
‘hinc movet Euphrates, illinc Germania bellum’. H is claiming that their
remoteness makes any threat illusory (2. 11. 1 ff. ‘quid . . . Scythes / . . .
cogitet Hadria / divisus obiecto remittas / quaerere’); at this time they
were in fact making overtures to Augustus at Tarraco (3. 8. 23 f. with
the introduction to that poem). discors refers to internal dissensions,
which as in the case of the Parthians (3. 8. 19 f.) made their supposed
threat even less serious; cf. 1. 26. 3 ff. ‘quis sub Arcto / rex gelidus
metuatur orae’ (where gelidae orae is probably dative and refers to the
Scythians themselves). Porph. (as one among other theories) took discors
to mean that the Tanais separated Europe from Asia, and so was too
remote to worry about; that is clearly impossible, and even Bentley’s
dissors (‘of divided lot’, i. e. belonging to both continents) lacks a
convincing parallel.
29–30. prudens futuri temporis exitum / caliginosa nocte premit deus:

it was an old maxim that no one could predict the future (Theog. 1075 f.
æªÆ IæŒı
ÆºÆ KØ ºıc = ªHÆØ, Pind. P. 3. 21 f.
etc.); many writers, pagan and Christian, add that it would be undesir-
able to do so (Cic. div. 2. 22 with Pease, 2. 105 ‘magnus Dicaearchi liber
est nescire ea melius quam scire’, Shorey on the present passage, N–H
on 1. 11. 3). prudens suggests the wise foresight of divine Providence
(cf. 1. 3. 21 ff. ‘nequiquam deus abscidit / prudens Oceano dissociabili /
terras’); the formulation here is Stoic rather than Epicurean, though the
Stoics actually believed in prophecy and the Epicureans did not. For the
murk that conceals the future cf. Theog. 1077 Zæ ªaæ ÆÆØ, Pind.
O. 12. 9, Juv. 6. 556 ‘at genus humanum damnat caligo futuri’, Milton, PL
7. 122 f. ‘Things not revealed, which the invisible king, / Only omnisci-
ent, hath suppressed in night’ (a clear imitation of our passage). deus is
more philosophical than the personal Iuppiter (cf. 1. 3. 21, 1. 34. 13);
premit means ‘conceals’ (OLD 17).
31–2. ridetque si mortalis ultra / fas trepidat: for the scornful amuse-
ment of the gods at human folly cf. 2. 8. 13 with N–H, Ov. ars 1. 87; for
similar derision on the part of a human cf. epist. 2. 1. 194 ‘rideret
Democritus’, Juv. 10. 51. mortalis is contrasted with deus and underlines
human presumption; cf. 1. 11. 1 f. ‘tu ne quaesieris, scire nefas, quem mihi
quem tibi / finem di dederint’. trepidat describes a fussy agitation; cf. 2.
11. 4 f. (quoted on 6–7).
32–3. quod adest memento / componere aequus: quod adest makes a

contrast with futuri temporis exitum; cf. Cic. fin. 1. 55 ‘nam corpore nihil
nisi praesens et quod adest sentire possumus’, off. 1. 11. memento is
regularly used in didactic contexts both serious (1. 7. 17, 2. 3. 1) and
parodic (serm. 2. 4. 12, 2. 5. 52). aequus, ‘calm in mind’ (2. 3. 1, OLD 7), is
contrasted with trepidat (32). To correspond with aequus, componere
seems to mean ‘to set at rest’ as well as ‘to set in order’.
33–4. cetera fluminis / ritu feruntur: cetera describes everything outside

the immediate present; cf. epod. 13. 7 ‘cetera mitte loqui’, carm. 1. 9. 9,
Eur. Alc. 788 f. e ŒÆŁ æÆ = ºªı , a ¼ººÆ B
.
For the age-old image of time as a stream see 2. 14. 2 with N–H,
Heraclitus (as summarized by Diog. Laert. 9. 1. 8) ÞE a ‹ºÆ
ÆF Œ (cf. Plat. Crat. 402a), Marc. Aur. 4. 43 Æ Ø KŒ
H ªØø ŒÆd ÞFÆ ÆØ › ÆN, ‘Time like an ever rolling
stream / bears all its sons away’ (the metrical version of Psalm. 90: 5
‘Thou carriest them away as with a flood’); here the emphasis is not only
on the inevitability of change but also on its unpredictability. feruntur,
suggesting a lack of control, suits the image of a river; cf. Sen. epist. 23. 8
‘ceteri eorum more quae fluminibus innatant non eunt sed feruntur’,
OLD 5b.
34–6. nunc medio alveo / cum pace delabentis Etruscum / in mare:

medio alveo means ‘in the middle of its channel’; cf. 3. 7. 28 ‘Tusco
denatat alveo’, Virg. Aen. 7. 303 ‘Thybridis alveo’. medio aequore is read
by one side of the MS tradition (perhaps influenced by aequus above); it
would mean ‘in the middle of the plain’, which does not provide a
proper contrast to ‘in flood’. For cum pace (‘placidly’) cf. Cicero Tusc. 5.
83 ‘nos quidem illud cum pace agemus’.
The sea to the west of Italy was normally called mare Tyrrhenum or
Tuscum; in this context Etruscum is more recherché (cf. Liv. 4. 52. 5). The
epithet reminds us that the ‘Tuscan Tiber’ rose near Arezzo, Maecenas’
home town (1. 20. 5 f. ‘paterni / fluminis’, Plin. nat. hist. 3. 53 ‘finibus
Arretinorum profluit’). Like other Italian rivers the Tiber was relatively
low in summer (Liv. 2. 5. 3 ‘Tiberim tenui fluentem aqua ut mediis
caloribus solet’), but was subject to violent flash-floods when the snows
in the mountains melted ( J. Le Gall, Le Tibre, 1953: 29 ff., N–H on 1. 2.
13). The hypermetric line with its elision seems to represent a gliding
motion, for H is not yet talking of flooding; contrast 4. 2. 22 and 23
which represent the Pindaric river in spate (5 ff.).
36–8. nunc lapides adesos / stirpisque raptas et pecus et domos /

volventis una: descriptions of rivers in flood were a traditional set-
piece; see e. g. Hom. Il. 5. 87 ff., 11. 492 ff., Lucr. 1. 281 ff., Virg. georg.
1. 324 ff., Ov. met. 1. 285 ff., Plin. epist. 8. 17 (on the Tiber). For the
boulders cf. Hom. Il. 13. 137 f., Lucr. 1. 289, Virg. Aen. 10. 362. adesos
ought to mean that boulders which have already been eroded are now
swept downstream, but that impairs the balance with raptas; Heinze
implausibly interprets lapides as the rocky river-banks (cf. Sen. Phoen. 72
‘partesque lapsi montis exesas rotat’, Sil. 3. 470 ‘adesi fragmina montis’);
RN has considered abesos, though abedo is attested only in grammarians
(Diom. GL 1. 362. 21, Charis. 345. 16 Barwick). For the tree-trunks cf.
Hom. Il. 11. 494, Ar. equ. 528, Lucr. 1. 284. For the cattle and houses
cf. Ov. met. 1. 286 f. ‘pecudesque virosque / tectaque’, Stat. Theb. 1. 366 f.,
Plin. epist. 8. 17. 4 ‘boves aratra rectores . . . arborum truncos aut villarum
trabes’.
una means ‘all together’ (not ‘along with it’) and suggests total
confusion; cf. Plaut. most. 277 ‘cum una multa iura confudit cocus’.
The polysyndeton allows us to visualize each horror in turn, and leads
to a climax with domos; cf. Pope’s fine imitation of Horace, epist. 2. 2 (To
Colonel —— 262 f.) ‘Inexorable Death shall level all, / And trees, and
stones, and farms, and farmer fall’.
38–9. non sine montium / clamore vicinaeque silvae: at several points

in the next four lines the natural world is described in human terms (the
‘pathetic fallacy’). For clamor applied to natural objects cf. OLD 5 (so
Hes. op. 511 AÆ fi A æØ oº). The roar of the torrent is a
regular feature of such descriptions; cf. Hom. Il. 4. 455, Lucr. 1. 288,
Virg. georg. 1. 327, Aen. 2. 307 f. ‘stupet inscius alto / accipiens sonitum
saxi de vertice pastor’, Sil. 3. 471, 4. 521.
40–1. cum fera diluvies quietos / irritat amnis: clause follows clause to
imitate the impetuous flood (cf. 4. 2. 5 ff., where Pindar’s style is
compared to a torrent); the scene of chaos is given more prominence
than that of quiet order, just as the convulsions of human life make the
greatest impact. For the archaic diluvies (instead of the more normal
diluvium) cf. 4. 14. 28, Lucr. 5. 255, 6. 292; the form is metrically
convenient as it can be combined with an adjective ending in a short
a. fera suggests a wild creature ‘untamed and intractable’ (Eliot, The Dry
Salvages 2); cf. N–H on 2. 14. 4. The adjective makes a contrast with
quietos, as does irritat; the river is like a sleeping animal that is danger-
ous when stirred (cf. 3. 20. 1 f., paroem. Gr. 1. 277). amnis tends to be a
grandiose word; cf. TLL 1. 1943. 5 ff., Encicl. virg. 1. 140 f. Here the plural
means ‘waters’; cf. Virg. Aen. 7. 464 f. ‘aquai / . . . amnis’ (of water in a
cauldron), Val. Fl. 4. 345 ‘qua rigidos eructat Bosporus amnis’, TLL. 1.
1942. 86 ff. H seems to be influenced by Hom. Il. 21. 312 (Scamander to
Simois) Æ Oæ Łı KÆºı (cited by Fraenkel). Some com-
mentators think H is including the Tiber’s tributaries, which are men-
tioned in Pliny’s vivid description (epist. 8. 17. 2), but whereas Pliny is
including the whole area affected, including the Anio, H concentrates
on the Tiber itself in contrasting moods.
41–3. ille potens sui / laetusque deget, cui licet in diem / dixisse ‘vixi’:
self-mastery (KªŒæØÆ) often refers to the control of one’s appetites,
but here it describes rather the conquest of hope and fear; cf. Sen. epist.
12. 9 ‘ille beatissimus est et securus sui possessor qui crastinum sine
sollicitudine expectat’, dial. 10. 10. 2–4. There, in spite of some Stoic
colouring, Seneca seems to depend on Epicurus (cf. epist. 12. 11); it was a
central doctrine of Epicureanism that one should take each day as it
comes without worrying about the future; see fr. 490 Usener (¼78B)
› B ÆhæØ lŒØÆ lØÆ æ ØØ æe c ÆhæØ, Hor.
carm. 1. 11. 8 f. with N–H, 2. 16. 25 f. ‘laetus in praesens animus quod
ultra est / oderit curare’, epist. 1. 4. 12 f. ‘omnem crede diem tibi diluxisse
supremum: / grata superveniet quae non sperabitur hora’.
in diem ¼ ‘from day to day’ (cf. serm. 2. 6. 47 f. ‘subiectior in diem et
horam / invidiae noster’), though the plural is more usual (Liv. 26. 12. 9
‘patriae occasum in dies exspectabant’, K–S 1. 566). The phrase should
be taken with dixisse, not with vixi in the sense of ‘I have lived for the
day’ (cf. Cic. de orat. 2. 169 ‘barbarorum est in diem vivere’); that would
impair the seriousness of what follows. At the end of each day we
should be able to regard our life as complete; cf. Sen. epist. 12. 8–9 ‘sic
ordinandus est dies omnis, tamquam cogat agmen et consummet atque
expleat vitam . . . quisquis dixit ‘‘vixi’’ cotidie ad lucrum surgit’ (cf. Hor.
carm. 1. 9. 13 ff.), epist. 101. 7–8 ‘cotidie cum vita paria faciamus . . . qui
cotidie vitae suae summam manum inposuit, non indiget tempore’. We
have taken vixi on its own as the terse and sufficient pronouncement of
the wise man; this is supported by Seneca’s story of Pacuvius (epist. 12.
8 f.) who had a funeral feast every day and was carried out to chants of
øÆØ but of course Pacuvius’ conception of the good life was very
different from Seneca’s.
43–5. cras vel atra / nube polum Pater occupato / vel sole puro: clouds,
as often, symbolize trouble (cf. 1. 7. 15), just as sunshine symbolizes good
times (Catull. 8. 3 ‘fulsere quondam candidi tibi soles’). puro means
‘unclouded’ as at 2. 5. 19 f. ‘pura luna’, 1. 34. 7 ‘per purum’ (of the sky).
Pater (i. e. Jupiter) suits the metaphor of the weather (3. 10. 8 n., Theoc.
4. 43
T ˘f ¼ººŒÆ b ºØ ÆYŁæØ , ¼ººŒÆ oØ); the idea is
traditional and poetic, and does not imply any religious commitment
on H’s part (Pöschl, op. cit. 226 f.). occupato is a third-person imperative;
Shackleton Bailey takes it as second-person, with Pater as vocative, but
that does not suit efficiet below. The alliteration of b and p gives the
solemn effect of early Roman poetry, and may hint at thunder (though
that does not suit puro); cf. Ar. nub. 391 ŒØfi Ð æfi A ÆÆÆ.
The emphasis lies on cloud rather than sunshine, but Heinze goes too
far when he talks of a zeugma with illustrato suppressed.
45–6. non tamen irritum / quodcumque retro est efficiet: irritum has a
legal tinge, ‘null and void’ (OLD 1). For the thought cf. Epicur. sent. Vat.
55 PŒ Ø ¼æÆŒ ØBÆØ e ªª . This idea was quite acceptable
to the Stoic Seneca, see ben. 3. 4. 1 (citing Epicur. fr. 435 Usener) ‘cum
certior sit nulla voluptas quam quae iam eripi non potest’, cf. dial. 10. 7.
9 ‘de cetero fors fortuna, ut volet, ordinet; vita iam in tuto est. huic adici
potest, detrahi nihil’. For efficere with a predicative adjective in the sense
of ‘render’ cf. Plaut. trin. 669 (of Amor) ‘mores hominum moros et
morosos efficit’.
46–7. neque / diffinget infectumque reddet: diffingere means ‘to

reshape’, hence ‘to alter’; elsewhere the verb is attested only at 1.
35. 38 ff. ‘o utinam nova / incude diffingas retusum / . . . ferrum’ and
serm. 1. 10. 37 ‘diffingit [v. l. diffindit] Rheni luteum caput’, perhaps
mocking a phrase of Bibaculus; cf. Nisbet (1995), 395. ‘infectumque
reddet’ goes further than both diffinget and irritum efficiet, for it
means ‘cause it not to have been made’; for this piece of traditional
wisdom cf. Theog. 583 f., Simon. PMG 603, Pind. O. 2. 17, Agathon
TrGF 1. fr. 5 ı ªaæ ÆPF ŒÆd Łe æŒÆØ, = IªÆ ØE –
i fi æÆªÆ, Epicur. sent. Vat. 55 (quoted on 45–6 above), Plin. nat.
hist. 2. 27, Otto 129 f., H. Fuchs, Fest. Tschudi, 1954: 48 ff., Milton, PL 9.
926 f. ‘But past who can recall, or done undo? / Not God omnipotent,
nor fate’.
48. quod fugiens semel hora vexit: commentators disagree about the
sense of vexit, which in itself simply means ‘has carried’. Some (followed
by RN) interpret ‘has brought’ (cf. Varro, Men. 333 ‘nescis quid vesper
serus vehat’, Lucr. 3. 1083 ‘posteraque in dubio est fortunam quam vehat
aetas’, Virg. georg. 1. 461); see 2. 16. 31 f. ‘et mihi forsan tibi quod negarit /
porriget hora’, and for the Epicurean sentiment anth. Lat. 474. 7 f. R
‘pervixi, neque enim fortuna malignior umquam / eripiet nobis quod
prior hora dedit’. There would be a paradox in the combination of
fugiens and vexit; cf. 3. 6. 43 f. ‘amicum / tempus agens abeunte curru’,
Bell 391. Others (including NR) think that in view of ‘quodcumque
retro est’ (46) and especially fugiens, which more naturally refers to
‘movement away from’, vexit means ‘has carried away’; cf. Sen. dial. 1.
5. 8 ‘irrevocabilis humana pariter ac divina cursus vehit’. On the second
view semel means ‘once an experience is over it cannot be altered’; RN
thinks this does less than justice to the Epicurean doctrine that once a
blessing has been received it cannot be annulled.
49–50. Fortuna saevo laeta negotio et / ludum insolentem ludere

pertinax: Fortune is characterized by two paradoxes: she delights in
her business (which is not serious but cruel), and perseveres in her game
(which is not pleasant but arrogant). For her cruelty cf. serm. 2. 2. 126, 2.
8. 61, Catull. 64. 169, Sall. Cat. 10. 1, Apul. met. 5. 5. 2 with Kenney,
Claud. in Eutrop. 1. 24 f. ‘quaenam ista iocandi / saevitia?’ For her sport
cf. 2. 1. 3 with N–H, Plaut. merc. 225, Sen. dial. 9. 11. 5 ‘ludos sibi facit’,
epig. 24. 65 with Prato, Lucian, Nigr. 20, Juv. 3. 40, 6. 608; as she
personifies the accidents that govern life, she is frequently described in
this way; the idea is less common and more shocking when ascribed to
the gods, e. g. by Shakespeare, King Lear 4. 1. 37 f. ‘As flies to wanton
boys are we to the gods; / They kill us for their sport’, Hardy, Tess of the
D’Urbervilles, ch. 59 ‘the President of the Immortals . . . had ended his
sport with Tess’. ludum ludere is a so-called figura etymologica (K–S 1. 275,
Wills 243 ff.), common in Greek (K–G 2. 582 f.) and early Latin (cf.
Porph. ‘veteribus usitatum elocutionis genus’); though vigorously em-
phatic it is unexpected in the economical style of the Odes, and at one
time RN considered ducere (‘protract’), but in a passage that owes
something to Horace, Boethius (cons. phil. 2. 2. 9) has Fortuna say
‘hunc continuum ludum ludimus’. RN now thinks that lusum . . . ludere
would be better than ludum . . . ludere (cf. Petr. 13.1 ‘o lusum fortunae
mirabilem’).
51–2. transmutat incertos honores, / nunc mihi nunc alii benigna: for
the fickleness of Fortune see for instance Soph. Ant. 1158 f., Men. dysc.
803 f. with Sandbach, Cebes, tab. 31. 5, Cic. nat. deor. 2. 43 with Pease,
Sen. suas. 1. 9 ‘dixit deinde locum de varietate Fortunae’, Sen. epig. 24. 64
with Prato, Plin. nat. hist. 2. 22 ‘volubilis, a plerisque vero et caeca
existimata, vaga inconstans incerta varia indignorumque fautrix’, Otto
142 (the wheel of Fortune). H makes the same point in three different
ways: by the verb, by an adjective with the object, and by an adjective
with the subject. With mihi he moves from the general to the particular,
and so continues to the end of the poem; for personal closures cf. 2. 16.
37 ff., 3. 1. 45 ff., 3. 3. 69 ff., Esser 9 ff.
53. laudo manentem: Fortuna could not be expected to stay for long
(Boethius, loc. cit. 2. 1. 13 ‘nec manendi fida’, 17), but people naturally
hoped she would; this idea seems to be expressed by the legend on a coin
from the reign of Commodus f o r t u n a e m a n e n t i , ‘to enduring
Fortune’ (LIMC 8. 1. 132, no. 113, 8. 2. 101). According to a flattering
speech of Plutarch’s (fort. Rom. 318a) she came to Rome with the
intention of staying (‰ FÆ), but this was because she had joined
forces with Virtue (316e).
53–4. si celeris quatit / pennas, resigno quae dedit: Fortuna takes off
with a beating of wings; for winged Fortune cf. anon. PMG 1019. 5, Val.
Max. 7. 1. 1 ‘volubilis’, Plut. loc. cit. (she took off her wings on arriving at
Rome ), Lucian, Timon 20, Fronto, pp. 150 f. van den Hout ‘omnes ibi
Fortunas Antiatis Praenestinas Respicientis balnearum etiam Fortunas
cum pennis . . . reperias’, LIMC 8. 1. 137, nos. 191 (a wall-painting from
Pompeii), 194–6 (gems of 1st cent. ad), 8. 2. 109.
resigno means rescribo (Fest. 281M ¼ 352L), i. e. ‘repay a debt’ (serm. 2.
3. 76, Ter. Phorm. 922); cf. epist. 1. 7. 34 ‘hac ego si compellor imagine,
cuncta resigno’ (a rather disingenuous protestation to Maecenas). For
the thought cf. Sen. dial. 7. 21. 2, 9. 11. 2–3 ‘quandoque autem reddere
iubebitur, non queretur cum Fortuna, sed dicet ‘‘gratias ago pro eo quod
possedi habuique . . . sed quia ita imperas, do cedo gratus libensque’’ ’,
Epictet. 2. 16. 28.
54–5. et mea / virtute me involvo: as opposed to the external posses-

sions that he renounces H’s virtus is his own (hence the emphatic mea);
cf. Acc. TRF 619 f. ‘nam si a me regnum Fortuna atque opes / eripere
quivit, at virtutem non quiit’. Although basically serious, H uses a self-
depreciating image drawn from popular philosophy: a rough cloak or
æø was the badge of the Stoic-Cynic preacher (Diog. Laert. 6. 13, 6.
22, Kindstrand on Bion pp. 161 ff.) and his one protection against the
elements. For various moral illustrations from clothing see Aristippus
30–4 Mannebach (cf. Hor. epist. 1. 17. 23 ff.), Plat. rep. 457c, Varro, Men.
571 ‘non quaerenda est homini qui habet virtutem paenula in imbri’,
Favorinus, de exsilio 20B, Athen. 7. 281 (about a convert to Epicurean-
ism) n ¼ØŒæı If e B IæB
ØHÆ IŁØa ØÆ.
55–6. probamque / Pauperiem sine dote quaero: for quaero cf. epist. 1. 2.
44 f. ‘quaeritur argentum puerisque beata creandis / uxor’ (where the
cynical beata means ‘rich’). A personified Greek Penia figured in the
Plutus of Aristophanes and most relevantly in popular philosophy; cf.
Bion F17. 15 ff. Kindstrand, Petr. 84. 4 ‘Bonae Mentis soror est Pauper-
tas’, Otto 268 f., RE 19. 1. 495 ff.
57–8. non est meum si mugiat Africis / malus procellis: non est meum
means ‘it is not a characteristic of mine’, analogous to genitives like
miserarum est (3. 12. 1); cf. Plat. Alc. 1. 106b e K , Plaut. asin. 190, Ter.
heaut. 549 ‘non est mentiri meum’, Cic. Pis. 75, K–S 1. 454, OLD 6. For
mugiat of inanimate things cf. epod. 10. 19 f. ‘Ionius udo cum remugiens
sinus / Noto carinam ruperit’, Hes. op. 508 ıŒ b ªÆEÆ ŒÆd oº with
West, Plin. nat. hist. 18. 360 ‘montium sonitus nemorumque mugitus’.
The wind is specified to add vividness; for the stormy south-wester cf. 1.
14. 5 f. ‘et malus celeri saucius Africo / antemnaeque gemant’, Virg. Aen.
1. 85 f. ‘creberque procellis / Africus’ with Austin.
58–9. ad miseras preces / decurrere et votis pacisci: decurrere means ‘to

have recourse to’, ‘take refuge in’ (cf. serm. 2. 1. 32, OLD 9); it suits the
seaman who runs for port in a storm (bell. Afr. 3. 5 ‘portum . . . quo naves
decurrerent’, OLD 4). Epicureans thought it irrational to pray when in
peril on the sea, for the gods were indifferent and the storms had
scientific causes; cf. Lucr. 5. 1229 ff. ‘(induperator) non divum pacem
votis adit ac prece quaesit / ventorum pavidus paces animasque secundas
/ nequiquam . . . ?’, N–H on 2. 16. 1. For miseras (‘abject’) cf. epod. 10. 17 f.
‘et illa non virilis eiulatio / preces et aversum ad Iovem’.
Seamen in storms promised offerings to the gods if they and their
cargoes were saved; cf. 1. 5. 13 ff. with N–H, Pease on Cic. nat. deor. 3. 89.
pacisci is a legal and commercial verb that suggests a discreditable
bargain; cf. Plat. Euthyphr. 14e KæØŒc . . .
, Sen. nat. quaest. 4. 7
‘negant posse fieri ut cum grandine aliquis paciscatur et tempestates
munusculis redimat’, Pers. 2. 29 f. ‘aut quidnam est qua tu mercede
deorum / emeris auriculas? pulmone et lactibus unctis?’
60–1. ne Cypriae Tyriaeque merces / addant avaro divitias mari:

Cyprus produced copper (cyprum) and timber (N–H on 1. 1. 13, RE 12.
70 ff.); Tyre exported purple textiles (serm. 2. 4. 84, epod. 12. 21) as well as
spices and perfumes from further east (RE 7A. 2. 1902 f.). In a crisis it
was customary to jettison part or all of the cargo (Greek KŒº); cf.
Dem. 35. 11, Acts 27: 38 KŒØ e ºE, KŒÆºº Ø e E N
c ŁºÆÆ, Juv. 12. 33 f. ‘decidere iactu / coepit cum ventis’ with
Courtney, Ach. Tat. 3. 2. 9, dig. 14. 2 de lege Rhodia de iactu, J. Rougé,
Recherches sur l’organisation du commerce maritime en Méditerranée sous
l’empire romain, 1966: 397 ff. Moral analogies were pointed out already
in Aeschylus: see sept. 769 ff. æ æıÆ KŒºa æØ = IæH
IºA = Zº ¼ªÆ Æ
ıŁ , Ag. 1008 ff. with Fraenkel. As Davis
(op. cit. 180) points out, H does not identify himself with the acquisitive
mercator.
addant underlines that shipwrecks were common. Elsewhere H
describes the sea as greedy (1. 28. 18 ‘exitio est avidum mare nautis’);
here avaro (‘miserly’) goes further, suggesting that it never gives up
its hoards.
62. tunc me biremis praesidio scaphae: a scapha was a lifeboat normally

towed behind a ship but sometimes kept on deck; cf. Plaut. rud. 74 f.,
162 ff., 173 ff., summed up at 366 ff. ‘de navi timidae ambae in scapham
insiluimus, quia videmus / ad saxa navim ferrier; properans exsolvi
restim, / dum illi timent; nos cum scapha tempestas dextrovorsum /
differt ab illis . . . / vix hodie ad litus pertulit nos ventus exanimatas’,
Petr. 102. 1 ‘ ‘‘quin potius’’ inquam ego ‘‘ad temeritatem confugimus et
per funem lapsi descendimus in scapham, praecisoque vinculo reliqua
fortunae committimus?’’ ’, 102. 4–5, 114. 7, Acts 27: 30
ÆºÆø c
Œ N c ŁºÆÆ, Ach. Tat. 3. 3. 4. For further evidence see
Zinn, op. cit. 234 f., L. Casson, Ships and Seamanship in the Ancient
World, 1995: index s. ‘ship’s boat’.
The MSS divide between tunc and tum; the former avoids the
assonance with tutum and tumultus (though modern instincts about
euphony are not infallible). biremis here means ‘with two oars’ (so
Lucan 8. 562, 10. 56); it usually means ‘with two banks of oars’
(Œæ ). praesidium is found four times in the Odes, but is generally
avoided in the higher genres of poetry except for Seneca (Axelson 98,
N–H on 1. 1. 2); there is a paradox in applying the word to a small
and fragile craft. For the instrumental ablative cf. Lucr. 5. 873 f.
‘quare pateremur eorum / praesidio nostro pasci genus esseque tutum’,
Caes. bG 6. 14. 4; in our passage it need not be combined with
tutum, which seems more effective if it begins the next clause (the
assonance provides a meaningful contrast with tumultus). For parables
about big and small boats in popular philosophy cf. epist. 1. 1. 92 f. with
Heinze, epist. 2. 2. 199 f., Plut. tranq. animi 466b; note also Antip.
Thess. anth. Pal. 9. 107 (where the small vessel trusts in the gods),
Stat. silv. 1. 4. 120 f. (where both ships suffer equally). We are told by
Seneca (dial. 9. 14. 3) that when Zeno heard of a shipwreck in which all
his possessions had been lost he said ‘iubet me Fortuna expeditius
philosophari’. Similarly, according to Vitruvius 6. 1 the shipwrecked
Aristippus said ‘eiusmodi possessiones et viatica liberis oportere parari,
quae etiam e naufragio una possent enatare’ (cf. Mannebach nos. 9A,
99B, 9D, 9E).
63–4. tutum per Aegaeos tumultus / aura feret geminusque Pollux: the
Aegean was notorious for its squalls (2. 16. 1 f. ‘in patenti / prensus
Aegaeo’), and its name may have suggested ÆNª ‘a storm-wind’. tumul-
tus suits the upheavals of life as well as those of the sea. aura is
contrasted with procellis (58); the Dioscuri could miraculously still the
waves (1. 12. 29 f. ‘defluit saxis agitatus umor, / concidunt venti’, Catull.
68. 63 ff.). Bentley read ferat, which has negligible MS support, because
of the hypothetical mugiat (58); but as tunc is the equivalent of ‘if that
does happen’, a confident future is more effective.
Ps.-Acro rightly comments ‘Pollux cum Castore intellegendus est’;
for the Dioscuri as joint protectors of seamen see 1. 3. 2 ‘lucida sidera’,
N–H on 1. 12. 27. H’s brachylogy may be a coinage of his own (imitated
at Ov. ars 1. 746 ‘geminus . . . Castor’), but it reflects the established
usage by which the Dioscuri could be called ‘Castores’ or less commonly
‘Polluces’; see Plin. nat. hist. 10. 121 ‘supra Castorum aedem’, Tac. hist. 2.
24. 2 ‘locus Castorum vocatur’, paneg. Lat. 2 (12). 39. 4 ‘Castoras
geminos’, Serv. georg. 3. 89 ‘ambo libenter et Polluces et Castores
vocantur’, TLL onomast. 2. 244. 55 ff. For similar expressions cf. Stat.
silv. 4. 6. 15 f. with Coleman, Juv. 11. 105 (of Romulus and Remus) with
Courtney. In Homer the dual `YÆ occasionally refers to Ajax and his
brother Teucer (Il. 4. 285, 13. 197, etc.); cf. Wackernagel, Kl. Schrift., 1955:
1. 144 f. citing Sanskrit parallels, Bell 64 f., D. L. Page, History and
the Homeric Iliad, 1959: 235 ff., 272 f. See further Löfstedt 1 edn. 2,
67 ff., H–Sz 19, Bell 1 ff.
3 0 . EX E G I M O N V M EN T V M
[M. von Albrecht, Antike und Abendland 18, 1973: 64 ff.; I. Borzsák, Acta Ant. Hung. 12,
1964: 137 ff.; E. Doblhofer, Horaz in der Forschung nach 1957 (1992), 111 ff.; Fraenkel 302 ff.;
A. Hardie, Stud. Clas. 21, 1983: 49 ff.; D. Korzeniewski, Gymnasium 79, 1972: 380 ff., and 81,
1974: 201 ff.; Lowrie 71 ff.; Pöschl 246 ff. (¼ GIF 20, 1967: 261 ff.); S. R. Slings in Ultima
Aetas (ed. C. Kroon and D. den Hengst), 2000: 5 ff.; Suerbaum, 1968; T. Woodman in
Quality and Pleasure in Latin Poetry (ed. T. Woodman and D. West), 1974: 115 ff.]
1–5. I have finished a monument more everlasting than bronze and more
conspicuous than the pyramids, beyond the reach of time and the elements.
6–9. A large part of me will survive, and my fame will grow as long as the
pontifex and vestal ascend the Capitol. 10–14. I shall be proclaimed in Apulia
as the first to have brought Aeolian verse-forms to Italian poetry. 14–16. Take
a justifiable pride, Melpomene, and crown me with Apollo’s bays.
At the end of a book ancient poets sometimes said something about

themselves and their expectation of future fame; for the so-called
3 0 . EX E G I M O N V M EN T V M 365
sphragis or ‘seal’ cf. W. Kranz, RhM 104, 1961: 3 ff., 97 ff., N–H on 2. 20,
McKeown on Ov. am. 1. 15. Here Horace does not name himself
(contrast 4. 6. 44), and mentions his Apulian birthplace only obliquely;
for his characteristic boast of lowly origin cf. the note on v. 12 below. In
particular the ode balances 1. 1, the programmatic poem of the collec-
tion; they are both written in an unvaried sequence of Asclepiads, a
metre found elsewhere in the Odes only at 4. 8 (the central poem of the
book, making a similar claim to fame). But Horace’s tone is more
assertive now that his task is completed: in the opening piece the
poet’s occupation is one among many, here he concentrates entirely on
himself; there he speaks cautiously in a future tense (1. 1. 29 ‘quod si me
lyricis vatibus inseres’), here confidently in a perfect tense (‘exegi mon-
umentum’); there he hopes for the Muses’ support in two modest
conditional clauses (1. 1. 32 ff. ‘si neque tibias / Euterpe cohibet
nec Polyhymnia / Lesboum refugit tendere barbiton’), here he asks
Melpomene to take pride in the achievement; there he is crowned
with Bacchus’ ivy (1. 1. 29), here more originally with the bay of Apollo
(15–16 n., cf. Suerbaum, op. cit. 310 f.).
Pindar had compared his songs to glittering buildings: O. 6. 1 ff.
$æıÆ Æ P- = Ø
E æŁæfiø ŁÆºı = ŒÆ ‰ ‹
ŁÆe ªÆæ = , fr. 194 (C. M. Bowra, Pindar, 1964: 20 ff.,
323); he is a particular source for this ode (3 ff.) when he claims that
neither rain nor wind will sweep away his treasure-house (P. 6. 10 ff. e
h
ØæØ Zæ , KÆŒe KºŁ = . . . h ¼ K ı
= ±ºe
¼ØØ Æ æfiø
æØ = ı ). Aristophanes uses metaphors
from masonry to describe the verses of Aeschylus (ran. 1004, cf. Tail-
lardat 438 f.); a scrap of Callimachus (aetia fr. 118) seems to describe two
temples in terms suggestive of poetry (R. F. Thomas, 1999: 76 ¼ CQ 33,
1983: 97 ff.). Such imagery suited the Roman admiration for solidly
constructed buildings and poems; thus at georg. 3. 13 ff. Virgil describes
his future epic as a marble temple. But whereas Pindar had used his
architectural image to glorify the laudandus, the Roman poets empha-
size their own achievement; cf. Suerbaum, op. cit. 151 ff. (Ennius), 165 ff.
(Horace), 172 ff. (Virgil), 194 ff. (Propertius), A. Kerkhecker on Call.
iamb., pp. 11 ff.
The poem also contains imitations of Simonides (for his influence on
Horace see Oates, cited on 3. 2, and the articles by Barchiesi and
Harrison in The New Simonides, ed. D. Boedeker and D. Sider, 2001:
255 ff., 261 ff.). See especially the dirge on the dead at Thermopylae
(PMG 531. 4 f., D. Steiner, CQ 49, 1999: 383 ff.), where he says that a
tribute such as his own, unlike material offerings, will be dimmed
neither by decay nor by time: KØ b ØF h Pæg = hŁ ›
ÆÆøæ IÆıæØ
æ (see further 2 n.). Elsewhere he speaks of
‘sharp-toothed time’ (eleg. W. edn. 2, 2. 88. 1) $æ Of O Æ ,
which may have influenced Horace’s imber edax (3 n.). A further
connection is suggested by an epitaph attributed to Cleobulus
(Korzeniewski, op. cit. (1974), 202 f., Slings, op. cit. 7 ff.); this describes
how a bronze statue on the tomb of Midas will last ‘as long as water
flows and trees are green’ (anth. Pal. 7. 153 i oøæ fi , ŒÆd
æÆ ŒæÆ Łºfi ), a boast which Simonides censures (PMG 581. 5).
Horace’s comparison with the pyramids (2) has no recorded parallel in
Greek poetry. An analogy has been sought in an Egyptian papyrus of
about 1200 bc; see H. Fuchs, #øæ Edgar Salin, 1962: 149 ff.,
Borzsák, op. cit. 138 ff., M. L. West, HSCP 73, 1969: 132 f. This says in
praise of ancient scribes that ‘they did not make themselves pyramids of
bronze with iron plaques . . . their teachings are their pyramids; . . . their
monuments are covered with earth . . . but their names are mentioned
because of their books’. If this line of thought had any influence on
Horace’s ode, it could only have been done through a Hellenistic
intermediary; Egyptian influences on Alexandrian poetry cannot be
entirely ruled out (cf. M. L. West, loc. cit.), but they can never
be tracked down with any precision. It has been suggested that a link
with Horace is provided by an epigram of Posidippus for the tomb of
Doricha at Naucratis (no. 17 Gow–Page, HE 1, p. 171 ¼ 122 Austin–
Bastianini, I. Trencsényi–Waldapfel, Acta Ant. Hung. 12, 1964: 149 ff.).
This shares certain commonplaces with Horace’s poem (see notes on
6–7 and 8–9); but though Doricha was wrongly connected with a
pyramid (cf. Herod. 2. 134, Page, Sappho and Alcaeus, 48 ff.), nothing is
said of that in the epigram. The theory of Egyptian influence here is
therefore too speculative (E. Maróti in Festschrift Altheim 1, 1969: 452 ff.).
In spite of Greek influences, Horace’s ode is very Roman. The
reference to the pyramids is highly topical after the annexation of
Egypt in 30 bc. The remark about ‘escaping Libitina’ (6–7 n.) not only
gives local colour but has a sardonic tone that would not suit Pindar.
The picture of the pontiff and vestal on the Capitol evokes a solemn
ceremony at the heart of Rome’s religious life (8–9 n.). When Horace
turns to his other patria in Apulia (10 ff.) he shows a feeling for his rural
background that has already appeared in the Roman Odes (3. 4. 9 ff., 3.
6. 37 ff.). But when he asks for a garland it is that of a victor at the
Pythian games, not that of a triumphator (15–16 n.).
The poem has a monumental solidity that matches its subject and is
again truly Roman. It consists of seven proud declarations, not seriously
modified by a tribute to the Muse (14–16), and each of these sections can
be broken into two balancing subdivisions; thus 1 ‘exegi . . . perennius’ is
answered by 2 ‘regali . . . altius’, 3 ‘quod non imber . . . ’ by 3–4 ‘non
Aquilo . . . diruere’, and so on. There is a pervasive series of contrasts:
death against growth, transience against permanence, temporal against
literary power; the interaction of these opposites gives unity to the poem
and generates much of its energy. The imagery moves between culture
and nature: monuments and their decay (1–5), the violence of the
elements (3 f. and 10), the rituals of the Roman state (8 f.), the evergreen
vitality of the poet’s creation (8 crescam, 16 lauro, with perhaps a hint at
1 perennius). Lucretius had described the power of poetry in terms of
florescence as well as masonry: ‘quo tot facta virum totiens cecidere
neque umquam / aeternis famae monimentis insita florent’ (5. 328 f.).
Horace gives something of the same impression here.
The ode has had a fruitful Fortleben. Propertius uses the image of the
pyramids to make a contrast with his power to immortalize (3. 2. 19, see
below, 2 n.); for specific echoes see below on altius (2), quaesitam meritis
(15). Ovid draws on our poem at the end of am. 1 to predict the fame of
his book (1. 15. 7 f. ‘mihi fama perennis / quaeritur’); see especially the
notes below on multaque pars mei (6) and dum (8). Above all, at the end
of met. 15 he skilfully refabricates Horatian motifs (15. 871 ff.): ‘iamque
opus exegi quod nec Iovis ira nec ignis / nec poterit ferrum nec edax
abolere vetustas . . . / parte tamen meliore mei super alta perennis / astra
ferar, nomenque erit indelebile nostrum, / quaque patet domitis
Romana potentia terris / ore legar populi’. The topics were taken up
by Du Bellay and Ronsard (‘Plus dur que fer j’ai fini mon ouvrage’),
Spenser (Epilogue to the Shepeardes Calender), Herrick (Pillar of Fame),
Milton (On Shakespeare), Klopstock and Pushkin (von Albrecht, op.
cit.). The following parallels in Shakespeare’s sonnets are worth noting:
‘Not marble nor the gilded monuments / Of princes shall outlive this
powerful rhyme’ (55. 1 f.), ‘And thou in this shalt find thy monument, /
When tyrants’ crests and tombs of brass are spent’ (107. 13 f.), ‘No! Time,
thou shalt not boast that I do change: / Thy pyramids built up with
newer might / To me are nothing novel, nothing strange . . . / For thy
records and what we see doth lie, / Made more or less by thy continuous
haste’ (123. 1–3, 11 f.); see further J. B. Leishman, Themes and Variations
in Shakespeare’s Sonnets, 1961: 27 ff. and his index under ‘Horace’;
C. Burrow, The Oxford Shakespeare: Complete Sonnets and Poems, 2002
(index under ‘Horace’). For Horace’s more general influence see
E. Stemplinger Das Fortleben des Horazischen Lyrik seit der Renaissance,
1906, Orazio nella Letteratura Mondiale, 1936, H. Krasser and
E. A. Schmidt (edd.), Zeitgenosse Horaz, 1996.
Metre: Asclepiad.
1. Exegi monumentum: exegi means not just ‘I have completed’

but ‘I have perfected’ (‘finii perfeci’ ps.-Acro); cf. Ov. rem. 811, fast.
1. 723, met. 15. 871 f. (cited above). The former usage is found with
actions (Virg. Aen. 6. 637 ‘his demum exactis’) or periods of time
(OLD s.v. 5); for a similar declaration on a real monument cf. CLE 89. 3
‘monumentum absolvi’. But when exigere is applied to artefacts it sug-
gests ‘finished’ in the sense of ‘perfected’, as is most familiar from the
participle exactus; cf. epist. 2. 1. 71 f. (of early Roman poetry) ‘sed emen-
data videri / pulchraque et exactis minimum distantia miror’, Prop. 3. 1. 8
‘exactus tenui pumice versus erat’, 3. 21. 30, Ov. met. 1. 406 with Bömer.
For the much rarer use with other forms of the verb cf. Cic. Verr. 2. 1. 133
‘ad perpendiculum columnas exigere’, Brink on epist. 2, appendix 5
(who suggests it is a stonemason’s word); though in our passage there
is no clarifying phrase (as in Cicero, loc. cit.), the implication suits
monumentum; cf. R. F. Thomas loc. cit (see introduction), n. 36.
monumentum is commonly used of works of literature that preserve an
author’s memory; cf. Lucil. 1084M, Catull. 95. 9, Cic. off. 1. 156 ‘hoc
idem etiam post mortem monumentis litterarum persequuntur’, Dion.
Hal. ant. Rom. 1. 1. 2 EÆ . . . ŒÆÆºØE, L c ıÆÆØŁÆØ E
ÆØ ÆPH e F
æ ı. Horace has given new life to the word
by describing his poetry as a sepulchral monument (as is shown by the
comparison with the pyramids in v. 2 and the reference to death in v. 6);
cf. Prop. 3. 2. 18, Ov. trist. 3. 3. 77 f., Suerbaum, op. cit. 327 f., H. Häusle,
‘Das Denkmal als Garant des Nachruhms’, Zetemata 75, 1980: 37 n.,
Lowrie 294 f. For the grandiosity of such edifices see S. Hornblower,
Mausolus, 1982: 224 ff., Korzeniewski, op. cit. 1972 (citing with illustra-
tions the tomb of Caecilia Metella on the Appian Way and of Munatius
Plancus at Gaeta).
aere perennius: aere is at first naturally understood as the metal,
which is associated by the poet with hardness and durability (3. 16.
1 n.); for similar comparisons cf. Ov. am. 3. 8. 3 ‘pretiosius auro’, trist.
1. 5. 53 ‘firmius aere’. But the mention of the pyramids (2) makes us focus
specifically on bronze statues (called aera at epist. 1. 6. 17, 2. 1. 240, for the
singular aes see Mart. 9. 43. 2, Val. Max. 2. 10, ext. 1. 25), which Horace
contrasts elsewhere with the artefacts of poets; cf. 4. 2. 19 f. ‘centum
potiore signis / munere donat’, 4. 8. 2 ff., following Pind. N. 5. 1. Poems
and speeches are compared elsewhere with monuments of marble or
bronze; cf. Pind. N. 4. 81 ºÆ Ł —Ææı ºŁı ºıŒæÆ, Isoc.
antid. 7 E . . . ºf ŒººØ H
ÆºŒH IÆŁø, Liban. orat.
12. 10. The distinction between ‘bronze’ and ‘a bronze’ is much clearer in
English than in Latin; ps.-Acro interprets ‘durabilius metallo’ and
‘durabilius statuis’ without suggesting that these are alternatives.
perenne can mean both ‘lasting through the years’ (as here) and
‘lasting all year’ (used particularly of evergreen foliage and water that
never runs dry). The former sense is applied to poetry by Catullus at 1. 10
‘plus uno maneat perenne saeclo’ (going further than Call. fr. 7. 14 ¥ Æ
Ø ıºf øØ ), and both by Lucretius at 1. 118 (of Ennius)
‘detulit ex Helicone perenni fronde coronam’. Even in our passage there
might be a latent sense of ‘ever-renewed’ (cf. Cleobulus and Simonides
cited in the introduction); this is then made explicit in the metaphor at
v. 8. perenne does not properly admit a comparative, and the form here is
unparalleled.
2. regalique situ pyramidum altius: the noun situs is regularly used of

the situation of a building (OLD s.v. 2); but the site of the pyramids was
not high, and H could hardly have thought otherwise. Heinze inter-
prets as ‘resting-place’ or ‘grave’ (an unattested meaning), yet sinere can
have the sense ‘to leave lying’ (Ernout–Meillet s. v.), the participle situs
often means ‘buried’ (as in ‘hic situs est’), and the fourth declension
noun positus is used for burial-places in Val. Fl. 6. 109 f.; similarly ŁŒ
from ŁØ is often used of a tomb (see further below). Others under-
stand the word to mean ‘edifice’ (Korzeniewski, op. cit. 1972: 383 f.);
though this also is an unattested use of the noun, the participle situs can
mean ‘set up’, as if it came from sisto; cf. Tac. ann. 1. 74. 3 ‘statuam
Marcelli altius quam Caesarum sitam’, 2. 7. 2 ‘veterem aram Druso sitam
disiecerant’ (with Goodyear’s parallels). This meaning too would follow
well from monumentum (for the grandiose periphrasis cf. Prop. 3. 2. 19
‘pyramidum sumptus ad sidera ducti’), and it might seem to suit altius
better than the former theory; cf. Stat. silv. 2. 7. 71 f. (to Lucan) ‘Pharo
cruenta / Pompeio dabis altius sepulcrum’, Greg. Naz. PG 35. 720 Æo
Ø Ææ H º H , ˙æÆŒºø ºH łºæÆ. Yet if situ
means ‘resting-place’ the idea of height may be sufficiently conveyed
by pyramidum.
Others, especially Pöschl, op. cit., give situs its common meaning of
‘decay’. Martial seems to have recalled our passage at 8. 3. 5 ‘et cum rupta
situ Messalae saxa iacebunt’, but if so he has altered it; for here ‘regal
decay’ would strike a satirical note at odds with the gravity of the
context, and it would be rhetorically inept to compare the height of
H’s monument to something that was crumbling; in any case Diodorus,
who had visited Egypt not long before, emphasized that the pyramids
were in an excellent state of repair: ØÆıØ
æØ F F ƒ ºŁØ c
K Iæ
B ŁØ ŒÆd c ‹º ŒÆÆŒıc ¼ ØÆıº
(1. 63. 5). Nevertheless, RN thinks it relevant to cite Simonides, PMG
531 (quoted in the introduction) where Pæ , the corruption of the
grave, is contrasted with the enduring qualities of poetry. situs in its
sense of ‘decay’ seems to be derived from sinere ‘to leave lying’ (Ernout–
Meillet, cited above); so if situ here can mean ‘grave’, it may at a
secondary level have associations with ‘decay’.
The pyramids were known to everybody as a wonder of the world
(Herod. 2. 134, Diod. 1. 63–4, Strab. 17. 1. 33–4, RE 23. 2. 2271 f.). It is
unnecessary to see a reference to Cleopatra’s mausoleum or Gallus’
boastful inscriptions (thus E. Maróti, Fest. Altheim 1, 1969: 452 ff., B. J.
Gibson, CQ 47, 1997: 312 ff.). For their unparalleled height (up to 500 ft.)
cf. Prop. 3. 2. 19 (quoted above), anth. Lat. 417. 3 (¼ Sen. epig. 26. 3
Prato) ‘Pyramidesque ausas vicinum attingere caelum’ (on a theme
similar to H’s), Tac. ann. 2. 61. 1 ‘instar montium eductae pyramides
certamine et opibus regum’. The Romans’ awe at their scale was
combined with disapproval of their uselessness; cf. Plin. nat. hist. 36.
75 ‘regum pecuniae otiosa ac stulta ostentatio’, Frontinus, aqu. 1. 16 ‘tot
aquarum tam multis necessariis molibus pyramidas videlicet otiosas
compares’.
3–4. quod non imber edax, non Aquilo impotens / possit diruere:
H is recalling Pindar’s storm at P. 6. 10 (see introduction), which suits
Delphi, but edax is more appropriate to long-term erosion. For the idea
that ‘constant dripping wears away a stone’ cf. Lucr. 4. 1286 f. ‘nonne
vides etiam guttas in saxa cadentis / umoris longo in spatio pertundere
saxa?’ with R. D. Brown, Tib. 1. 4. 18 with K. F. Smith, Otto 156 f.
Words for nibbling were sometimes applied to erosion by rivers; cf.
1. 31. 8 ‘mordet aqua taciturnus amnis’ with N–H (citing Call. ep. 44. 4
æªø among other passages), 3. 29. 36 ‘lapides adesos’. The more
forceful edax was used of parasites by Plautus (Pers. 421) and by H in
epist. 2. 1. 173, but was given metaphorical applications by others; cf.
Virg. Aen. 2. 758 ‘ignis edax’ with Lyne, 1989: 51 ff., Prato on Sen. epig.
1. 1. The parallel with Simonides, eleg. W. edn. 2, 2. 88. 1 has been noted
in the introduction; both authors are combined at Ov. met. 15. 234 ff.
‘tempus edax rerum, tuque invidiosa vetustas, / omnia destruitis, vitia-
taque dentibus aevi / paulatim lenta consumitis omnia morte’. For
the impermanence of monuments see further D. Fowler, Roman
Constructions, 2000: 193 ff.
impotens, like IŒæÆ , means ‘without self-control’, cf. 1. 37. 10 f.
(of Cleopatra) ‘quidlibet impotens / sperare’; it is applied to natural
forces at Catull. 4. 18 ‘tot per impotentia freta’, Mart. 1. 49. 19 ‘bruma
impotens’ (so impotentia at epod. 16. 62). Elsewhere the word means
‘powerless’ (cf. 2. 1. 26); here the two senses combine to suggest an
impotent fury (cf. ‘non . . . possit diruere’). The personification of Aquilo
(cf. 1. 3. 12 ff.) recalls the unruliness of winds in epic (Virg. Aen. 1. 81 ff.);
see also Stat. silv. 1. 1. 91 ff. (a clear reminiscence of our passage) ‘non hoc
imbriferas hiemes opus [Domitian’s statue] aut Iovis ignem / tergemi-
num, Aeolii non agmina carceris horret / annorumve moras [a reversal
of H’s fuga temporum]; stabit dum terra polusque, / dum Romana dies’.
4–5. aut innumerabilis / annorum series et fuga temporum: the kings

of Egypt, like Shelley’s Ozymandias, no doubt boasted that their
monuments were indestructible; cf. OGIS 383. 37 ff. (the inscription
of Antiochus of Commagene in the ‘Asiatic’ style) Kd b ƒæŁı
F ŒæEÆ I æŁ
æ ı ºÆØ PæÆø ¼ª
ØÆ Łæ ø
ŒÆÆÆŁÆØ æŁ. Literary men were more sceptical; cf.
above 3n., Enn. ann. 406 with Skutsch, Lucr. 5. 306 ff., Cic. Marc. 11
‘nihil est enim opere et manu factum quod non conficiat et consumat
vetustas’, Phil. 9. 4, Ov. fast. 5. 131 f., met. 15. 871 f. annorum series reminds
us of the Romans’ ability to organize chronology (D. Feeney ap. Rudd
1993: 58), but even they found the date of the pyramids incalculable. The
transferred epithet innumerabilis avoids a cumbrous genitive. Time,
though infinite, is also fugitive (2. 14. 1 with N–H); it is a paradox
that it can destroy by running away.
6–7. non omnis moriar multaque pars mei / vitabit Libitinam: the
Romans were cautious when they expressed a belief in survival after
death; cf. Prop. 4. 7. 1 ‘sunt aliquid manes, letum non omnia finit’, Sen.
Tro. 382 f. ‘an toti morimur nullaque pars manet / nostri?’, Stat. silv. 3. 3.
195 ‘non totus rapiere tamen’. Horace gives this platitude a different
meaning: that he will live on in his poems, which are part of what he is.
For the commonplace that a great poet’s writings survive cf. 1. 32. 3 with
N–H, 4. 9. 1 ff., Plat. symp. 209d, Call. ep. 2. 5 Æƒ b Æd ıØ I
(where ‘Nightingales’ was probably the title of Heraclitus’ poetry-book),
Posidippus 17. 5 (Gow–Page, HE 3146 f.) &ÆfiHÆØ b ıØ º Ø
ŒÆd ıØ = T fi B Æƒ ºıŒÆd Łªª ÆØ º , A. O. Hulton,
Latomus 31, 1972: 499 ff. Sometimes it is the poet himself or herself
who is said to survive; cf. Antip. Sid. anth. Pal. 7. 713. 3 f. (on Erinna)
ØªæØ PŒ Xæ, Pb ºÆ = ıŒe e ŒØæfi Ð
ŒøºÆØ æıªØ, Tullius Laurea, ibid. 7. 17. 7 (where Sappho speaks)
ªÆØ ‰ #.ø Œ Œıª. For claims by poets themselves cf.
the paraphrase at Sappho fr. 193 ‰ P IŁÆ ÆØ ºŁ, Enn.
var. 18V ‘volito vivos per ora virum’; Catull. 1. 10. Horace’s multa (for the
more usual magna) is imitated by Ovid am. 1. 15. 42 ‘parsque mei multa
superstes erit’ (McKeown gives other parallels to the idea; cf. also met.
15. 875, cited in introduction above).
Libitina was the goddess of funerals (Wissowa 245, Blümner, 1911:
489 f., RE 13. 1. 113); she kept a register of deaths (Dion. Hal. ant. Rom.
4. 15. 5, Suet. Nero 39. 1) for which she received a tax (serm. 2. 6. 19
‘Libitinae quaestus acerbae’, Phaedrus 4. 21. 26, ILS 6726 from Bergamo
‘lucar Libitinae’); her grove at Rome, presumably on the Esquiline
(Steinby 3. 189 f.), was the headquarters of the undertakers, who kept
their equipment there (ps.-Acro on serm. 2. 6. 19, Ascon. Mil. 29KS). In
the arena the gate through which the corpses were dragged was called
the ‘Porta Libitinensis’. H’s reference is wry and down-to-earth; cf. epist.
2. 1. 49 ‘miraturque nihil nisi quod Libitina sacravit’ with Brink, Juv. 12.
122 ‘si Libitinam evaserit aeger’.
7–8. usque ego postera / crescam laude recens: this climactic sentence
expands on its predecessor and therefore should be preceded by a colon.
usque is combined with crescam, a verb which implies progression; it
cannot be taken even partly with an adjective like recens. For a similar
sentiment cf. Ap. Rhod. 4. 1773 ff. (the end of the poem), expressed in
less personal terms, as befits an epic: Æ¥ IØÆd = N K
ªºıŒææÆØ r IØ = IŁæØ . crescam would normally suggest
growing fame in life (Virg. ecl. 7. 25 ‘crescentem ornate poetam’, Nepos,
Cato 2. 4 ‘quoad vixit virtutum laude crevit’); here it is extended to
posthumous reputation (cf. 1. 12. 45 f. ‘crescit occulto velut arbor aevo /
fama Marcelli’, Enn. ann. 365 ‘ergo postque magisque viri nunc gloria
claret’, Sil. 6. 63, laudatio Turiae 2. 60 ‘crescere tui memoriam’). postera
laude is to be taken with both crescam and recens; the ablatives mean ‘by
reason of ’, cf. 4. 4. 45 f. ‘post hoc secundis usque laboribus / Romana
pubes crevit’; Prop. 3. 1. 33 f. ‘Homerus / posteritate suum crescere sensit
opus’ echoes our passage in a condensed form. recens paradoxically
suggests that even in later ages Horace will remain fresh; cf. epist 2. 1.
53 f. ‘Naevius in manibus non est et mentibus haeret / paene recens?’
(there ironically). H’s image suggests the freshness of vegetation, as if
his reputation were watered by praise; cf. 1. 12. 45 (cited above), Pind.
N. 8. 40 ff. IØ Iæ,
ºøæÆE KæÆØ = ‰ ‹ æ <- ->,
Virg. ecl. 10. 73 f., Pöschl, op. cit. 249 n. It was suggested in 1 n. above
that it might be legitimate here to take up the vegetative connotations
of perennius which were only latent in the opening line.
8–9. dum Capitolium / scandet cum tacita virgine pontifex: just as 2.

20 describes the extent in space of H’s fame, so this poem describes the
extent in time; and just as the pyramids are contrasted with H’s work, so
the Capitol is seen as a parallel to it. For the hyperbole cf. epod. 15. 7 ff.
‘dum pecori lupus et nautis infestus Orion . . . ’ (following Call. fr. 202.
69 f.), Theog. 1. 252 Zæ i ªB ŒÆd MºØ , Cleobulus (introduction
above), Tib. 1. 4. 65 with K. F. Smith, Virg. Aen. 1. 607 ff. with Austin,
Shakespeare, sonnet 18 ‘So long as men can breathe and eyes can see / So
long lives this, and this gives life to thee’; the formula is the positive
counterpart of an adynaton (1. 33. 7 with N–H). Such dum clauses often
give an illustration that suits the matter in hand; cf. Critias 8. 6 ff. Diehl
(Anacreon will live as long as the rituals of the symposium), Posidippus,
epig. 17. 7 (Gow–Page, HE 3148 f.) hÆ e ÆŒÆæØ , n ˝ÆŒæÆØ
z ıºØ = i Yfi ˝ºı ÆF K ±ºe ºª, Ov. am. 1. 15. 9
‘vivet Maeonides Tenedos dum stabit et Ide’, and especially 25 f. ‘Aeneia-
que arma legentur / Roma triumphati dum caput orbis erit’ (see
McKeown on both passages). In the same way H is here proclaiming
himself the poet of Roman institutions; in the opening stanza of the
book he had called himself ‘Musarum sacerdos’ (3. 1. 3), so here there
might be the hint of an analogy between himself and the pontifex
(scandet balances crescam).
The Capitol was the symbol of Roman imperium; cf. Binder 124 ff.
The city’s survival was thought to depend on its preservation (3. 3. 42
‘stet Capitolium’, Amm. Marc. 2. 26. 12); for a similar attitude to the
Colosseum in later times cf. Bede, PL 94. 543B ‘quamdiu stat Colisaeus,
stat et Roma; quando cadet Colisaeus, cadet et Roma; quando cadet
Roma, cadet et mundus’, Gibbon vol. 7, ch. 71, p. 330 (Bury), Byron,
Childe Harold’s Pilgrimage 4. 145 ‘While stands the Coliseum, Rome
shall stand’. For the immortality of the res publica, assumed in the rites
of Vesta (3. 5. 11 n.), cf. already Cic. Att. 9. 10. 3, Marc. 22; Rome first
appears as ‘The Eternal City’ at Tib. 2. 5. 23 f. ‘Romulus aeternae
nondum formaverat urbis / moenia’ (see K. F. Smith), Ov. fast. 3. 72
(see Bömer). For the theme in Augustan ideology cf. Liv. 4. 4. 4. with
Ogilvie, Virg. Aen. 1. 278 with Austin; for Tiberius cf. Vell. 2. 103. 4 with
Woodman; see further Melinno, Supp. Hell. 541. 13 ff., C. Koch, Religio,
1960: 142 ff. A more sober view is seen in Polybius 38. 22 on Scipio’s
forebodings about Rome, Cic. rep. 6. 21 ff. on Scipio’s perception of
Rome’s insignificance, Cic. fam. 4. 5. 4 giving Sulpicius’ sombre reflec-
tions on the transience of power, Hor. ars 63–9 ‘debemur morti nos
nostraque . . . mortalia facta peribunt’; see further P. Hardie in A. Powell
(ed.), Roman Poetry and Propaganda in the Age of Augustus, 1992: 59 ff.
More generally, the end of the world, whether by decay or conflagration,
was a familiar element in Epicurean and Stoic thought.
The Pontifex Maximus had a particular responsibility for the Vestals
(Beard–North–Price 1. 57 f.), but here H is less specific; at this point
the office was still held by Lepidus, who was living in retirement at
Circeii (Suet. Aug. 16. 4), and it was not assumed by Augustus till 12 BC
(G. W. Bowersock in Raaflaub–Toher 380 ff.). For rites similar to the
one described cf. Varro, lL. 5. 47 (on the Via Sacra) ‘qua sacra quotquot
mensibus feruntur in Arcem’, Joh. Lydus, de mens. 4. 49 ¼ 36 (on a
sacrifice to Jupiter on the Ides of March) ªıı F Iæ
Øæø ŒÆd
H ŒÆ æø B 1æ (i. e. the chief vestal), Wissowa 517 (who,
however, associates this ceremony with the Flamen dialis). The Vestal is
silent because it is a solemn occasion (cf. ‘favete linguis’ in 3. 1. 2); a
similar situation is reflected at Virg. Aen. 7. 71 f. ‘castis adolet dum altaria
taedis / et iuxta genitorem astat Lavinia virgo’. H pictures the steep
climb from the Forum up the Clivus Capitolinus (T. P. Wiseman ap.
Steinby 1. 280 f.); victorious generals ‘climbed the Capitol’ (Liv. 42. 49. 6)
by the same route, but that does not mean that we are invited to see H
himself as a triumphator (see n. on vv. 15–16 below).
H’s lines bear some resemblance to Virg. Aen. 9. 447 ff. (on Nisus and
Euryalus) ‘si quid mea carmina possunt, / nulla dies umquam memori
vos eximet aevo, / dum domus Aeneae Capitoli immobile saxum /
accolet imperiumque pater Romanus habebit’ (the two passages are
recalled at Mart. 9. 1. 5 ff.). The similarity is best explained by postulat-
ing a common source in Ennius’ proem to book 16 of the Annals; cf.
Suerbaum, op. cit. 159 and Skutsch’s commentary, pp. 568 f.
10. dicar qua violens obstrepit Aufidus: ‘some day someone will say’
was a poetic formula going back to Homer (Il. 6. 459 etc., Ov. ars 3. 341);
for the proud dicar cf. Prop. 4. 11. 36 ‘in lapide hoc uni nupta fuisse
legar’, Ov. met. 15. 877 f. cited in the introduction (capping Horace’s
boast), trist. 3. 7. 50 ff. Here H himself looks for fame in his native
Apulia; for such local patriotism Fraenkel 304 f. cites Cic. Planc. 19–22,
leg. 2. 3–6, Prop. 4. 1. 63 ff. ‘ut nostris tumefacta superbiat Vmbria
libris . . . ’, Ov. am. 3. 15. 8 ff. ‘Paelignae dicar gloria gentis ego . . . ’,
Mart. 1. 61. 12; for a less optimistic assessment cf. Matthew 13: 57 PŒ
Ø æ ¼Ø N c K fi Ð ÆæØ ÆPF.
Many commentators take the qua clause not with dicar but with
deduxisse (14); for the word-order cf. Virg. ecl. 9. 7 ff. But H had left
Apulia before he became a poet. Regions were often characterized by
their rivers (3. 29. 27–8 n.), and the Aufidus (Ofanto) was the chief river
of Apulia (Encicl. oraz. 1. 398). H’s birthplace at Venusia was only ten
miles away, and he mentions the river’s boisterousness elsewhere (4. 9. 2
‘longe sonantem natus ad Aufidum’, 4. 14. 25 ff., serm. 1. 1. 58); nearer the
Adriatic it is less impressive, particularly in the summer (climatic
changes and human intervention may also have reduced its volume).
The turbulence of the river marks a contrast with the quiet procession
on the Capitol (9 tacita) and with the general dryness of the area. RN
thinks that the prefix of obstrepit may suggest that H’s praises can be
heard above the noise of the elements, which damage only material
monuments (3 ff.); cf. Woodman, op. cit. 123 f.
11–12. et qua pauper aquae Daunus agrestium / regnavit populorum:

Daunus was the legendary king of Daunia, the northern part of Apulia
(N–H on 1. 22. 14, Strab. 6. 3. 8–9, Bömer on Ov. fast. 4. 76); hence 4. 6.
27 ‘Dauniae defende decus Camenae’ (grandiloquent and defiant, just
like our passage). It is a paradox that a king (proverbially regarded as
rich) should lack something as cheap as water (serm. 1. 5. 88 ‘vilissima
rerum’); for the genitive, which as with egens indicates ‘lack’, see K–S
1. 441 f. For drought in the area cf. epod. 3. 16 ‘siticulosae Apuliae’, serm.
1. 5. 91 (even at Canusium near the Aufidus until Herodes Atticus built
an aqueduct), Ov. met. 14. 510, K. D. White 73 f. The primitive Apulian
tribes are called agrestes because the first towns were not yet established
in Daunus’ time (they were attributed to his son-in-law Diomedes; see
serm. 1. 5. 92 with Lejay’s references); at the same time, H wishes to
emphasize the rusticity of his origins. regnavit with a genitive is a
Graecism (cf. N–H on 2. 9. 17, Serv. Aen. 11. 126).
12. ex humili potens: here humili is masculine (cf. Cic. part. or. 57 ‘ex
beato miser’, Soph. OT 454 ıºe KŒ æŒ , K–S 1. 18, 505); at
Prop. 2. 10. 11 ‘ex humili’ the adjective is neuter (see Enk). A reference to
H’s origins suits a concluding poem; cf. 2. 20. 5 f. ‘pauperum / sanguis
parentum’, epist. 1. 20. 20 ‘me libertino natum patre et in tenui re’, where
he adds (22) ‘ut quantum generi demas virtutibus addas’. potens describes
the poet’s power to influence people’s thinking (cf. 4. 8. 26 f. ‘lingua
potentium / vatum’); RN stresses the analogy with King Daunus, who
was powerful in a political sense. Bentley applied potens to Daunus
himself, but though this suits him well enough, ex humili does not;
elsewhere he is described as ‘Illyricae gentis claro viro’ (Fest. 69M ¼ 60L),
one of the many sons of the notable Lycaon (Ant. Lib. met. 31).
13–14. princeps Aeolium carmen ad Italos / deduxisse modos: the

Aeolians settled in Lesbos near the end of the second millennium
bringing with them the dialect later used by Sappho and Alcaeus.
Here H is referring to their metres and (in very general terms) their
manner and matter; cf. 4. 3. 12 ‘fingent Aeolio carmine nobilem’. H feels
at liberty to ignore the Sapphics of Catullus 11 and 51 because those were
the only two examples and he is thinking primarily of Alcaeus (cf. 2. 13.
26 f., epist. 2. 2. 99). Italos modos means, not ‘Italian metres’ (for the
metres were taken from Greece), but ‘Italian melodies’, so described
because they are played on the Roman lyre, i. e. they are in Latin. H says
Italos rather than Latinos to stress his own place of origin; cf. Catull. 1. 5
on Nepos, Prop. 3. 1. 4.
Roman poets sometimes claim to be the first to have imitated Greek
models; cf. epist. 1. 19. 23 f. ‘Parios ego primus iambos / ostendi Latio’,
32 f. ‘hunc (Alcaeum) ego, non alio dictum prius ore, Latinus / vulgavi
fidicen’, Lucr. 1. 117 ff. ‘Ennius ut cecinit noster qui primus amoeno /
detulit ex Helicone perenni fronde coronam / per gentis Italas hominum
quae clara clueret’ (perhaps echoing Ennius himself ), Prop. 3. 1. 3 f. with
Fedeli, Wimmel, 1960: s.v. ‘primus-Motiv’, Kambylis 155 ff. In our
passage princeps suggests leadership as well as priority (cf. epist. 1. 19.
21 ‘libera per vacuum posui vestigia princeps’, followed by ‘qui sibi fidit /
dux reget examen’); as with potens above, RN thinks it may suggest an
analogy with Daunus, a political leader; see further on 15–16 below.
deduxisse literally means ‘to have brought X to Y’; cf. ars 244, Virg.
Aen. 2. 800, Ov. fast. 3. 151 (of Numa) ‘primus oliviferis Romam deductus
ab arvis’, Calp. 4. 161. That is the primary sense here: H has brought
Aeolian verse to Latin poetry; the terms of the metaphor are varied in
epist. 1. 3. 12 f. ‘fidibusne Latinis / Thebanos aptare modos studet auspice
Musa?’, and reversed in carm. 1. 32. 3 f. ‘dic Latinum, barbite, carmen’,
where the carmen is Latin and the lyre Greek. In the present passage
scholars have suggested a number of extra nuances. One is that of
founding a colony (OLD s.v. 9, E. Maróti, Acta Ant. Hung. 13, 1965:
101 ff.); though not accepting all the connotations of this idea, RN
sees a continuation of the analogy with Daunus (see on potens and
princeps above).
Others see a reference to a triumphant Roman general (Borzsák, op.
cit. 145 f., Woodman, op. cit. 124, A. Hardie, op. cit. 52); cf. 1. 37. 31
‘privata deduci triumpho’, Liv. 28. 32. 7 ‘quos secum in patriam ad
meritum triumphum deducere velit’, and especially Virg. georg. 3. 10 f.
‘primus ego in patriam mecum . . . / Aonio rediens deducam vertice
Musas’, where the imagery is supported by 9 and 17 ‘victor’ and 12
‘primus Idumaeas referam tibi, Mantua, palmas’ (see Mynors). We
accept that in a general sense H was influenced by his friend’s deducam;
nevertheless Virgil’s repeated victor is abandoned, the object of deduxisse
is no longer the Muses but Aeolian verse-forms; and those forms are
brought not to Horace’s patria, still less to the Capitol, but to ‘Italian
melodies’ (i. e. Latin poetry). In the closing lines H emphasizes that his
garland is very different from that of the triumphator (see 15–16 n.).
Finally, it is held that deducere evokes a metaphor from spinning (Ross
134 f.); the verb is used literally of ‘drawing down threads’ (Catull. 64.
312, OLD 4a) and metaphorically of ‘spinning verses’ (serm. 2. 1. 4, epist.
2. 1. 225 ‘tenui deducta poemata filo’, OLD 4b); see further W. Eisenhut,
Gedenkschrift für G. Rohde, 1961: 91 ff. (reprinted in his Properz, 1975:
247 ff.), Pöschl 258 f. This interpretation does not cohere well with ‘ad
Italos modos’; it is true that ad can be used of a musical accompaniment
(epist. 1. 2. 31 ‘ad strepitum citharae’), but in this geographical context
the reader would expect deducere ad to indicate some sort of destination.
This argument tells a fortiori against connecting deduxisse with the
1FÆ ºÆº of Callimachus. Moreover, a reference to Callima-
chean poetics goes against the tenor of the ode, which begins with a
metaphor from monumental masonry and ends with an acknowledge-
ment of the classical lyric poets.
14–15. sume superbiam / quaesitam meritis: H is modifying the pre-

sumption of his previous claims by attributing his success to the Muse;
cf. 1. 26. 9 f. ‘nil sine te mei / possunt honores’, 4. 3. 21 ff. ‘totum muneris
hoc tui est, / quod monstror digito praetereuntium / Romanae fidicen
lyrae; / quod spiro et placeo, si placeo, tuum est’, Pind. pae. 7b. 15 ff.,
Call. fr. 228. 1. But the terms in which he does so are not entirely clear.
Most editors are probably right in taking the whole expression to mean
‘assume a well-earned pride’; cf. Caes. bG 1. 33. 5 ‘Ariovistus tantos sibi
spiritus . . . sumpserat’, Prop. 2. 25. 21 ‘qui pleno fastus assumis amore’,
Ov. am. 2. 17. 9 with McKeown (and contrast 3. 10. 9 ‘pone superbiam’).
As no possessive adjective is expressed with meritis, one is naturally
inclined to understand tuis from the imperative sume; for gods’ ‘deserts’
cf. Plaut. capt. 922 ‘Iovi disque ago gratias merito magnas’, ILS 3834
‘Aescolapio donom dat lubens merito’ (the abbreviation L. M. is
common in inscriptions), TLL 8. 823. 53 ff. For superbia as justifiable
pride (rather than the more usual arrogance) cf. Prop. 4. 1. 63 ‘ut nostris
tumefacta superbiat Vmbria libris’.
Some argue that superbiam refers to H’s own pride, and meritis to his
deserts; this follows well from the preceding account, and receives some
support from Prop. 3. 2. 25 ‘ingenio quaesitum nomen’ (a passage that
draws on the ode, but without any divine address). sume then means
‘take over’; thus Porph. interprets ‘adroga, inquit, tibi gloriam ubertate
ingenii quaesitam’. sume suits a dedication to a god (cf. CLE 868. 5 ‘sume
libens simulacra’), and cinge (16) the reciprocal boon requested on such
occasions. There is some inconsistency, however, in handing over his
pride to the Muse while at the same time claiming the credit for
himself.
D. Korzeniewski understood superbiam as ‘a thing to take pride in’
(like the English ‘pride and joy’), i. e. the book of poems (loc. cit., 1974:
204 ff.); cf. the concrete sense of words like gloria, honor, and gaudium.
He saw a particular analogy in ¼ªÆºÆ, a word sometimes applied to an
offering (it was explained as ‘a thing in which one delights’ as well as ‘a
thing that glorifies’). Although this explanation would suit sume and the
reciprocal cinge, it can hardly be reconciled with quaesitam meritis; the
sense assigned to superbiam is very strange, and even as a calque it does
not represent ¼ªÆºÆ satisfactorily.
15–16. et mihi Delphica / lauro cinge volens, Melpomene, comam:

H does not as a rule differentiate the functions of the Muses (at 3. 4.
1–2 n.), but he may have associated Melpomene with lyric (cf. 4. 3. 1,
N–H on 1. 24. 1). volens (often combined with propitius) is found in
prayers for a god’s goodwill, particularly in granting a request that could
be refused (‘of thy grace’); cf. Cato, agr. 141. 2 ‘Mars pater, te precor
quaesoque uti sies volens propitius’, Liv. 1. 16. 3 with Ogilvie, 24. 38. 8 ‘ut
nobis volentes propitii adsitis’, Appel 122 f. Łºø is similarly used in
Greek; cf. Aesch. Ag. 664 with Fraenkel, cho. 19 ªF b Æ

Łºø K, Pind. I. 6. 42 f., Pulleyn 144 f.
Laurus is the ‘sweet-bay, L. nobilis . . . not the cherry-laurel of English
shrubberies’ (Mynors on Virg. georg. 2. 18); cf. Abbe 94 f. with illustra-
tion; for general information see M. B. Ogle, AJP 31, 1910: 287 ff. The
plant was particularly associated with prophecy and the Delphic
cult of Apollo (Lucr. 6. 154 ‘Phoebi Delphica laurus’, F. Williams on
Call. h. 2. 1); the Muses on Helicon gave Hesiod a branch of bay (theog.
30), and a garland of bay was awarded to victors at the Pythian games,
including poets (Call. fr. 194. 33, Paus. 10. 7. 22–4); H requests it here
because it suits the inspired vates (Kambylis 18 ff.), cf. 4. 2. 9 (of Pindar)
‘laurea donandus Apollinari’. In Roman life the laurea corona was also
associated with the triumph, which provided a variety of images for
Augustan poets (K. Galinsky, WS 82, 1969: 75 ff.): the Muse of Proper-
tius holds a triumph (3. 1. 9 ff.), Ovid is a prisoner in Cupid’s triumph
(am. 1. 2. 19 with McKeown). But Delphica (emphatic in this position)
shows that Horace is asking for the bays of a poet rather than those of a
general; a general dedicated his bays to Jupiter (RE 7A. 1. 510), but
Horace is to receive his from the Muse. See also the important parallel
at 4. 3. 6 ff. (on the poet blessed by Melpomene): ‘neque res bellica
Deliis / ornatum foliis ducem / . . . ostendet Capitolio’, followed by a
contrast with the foliage of Tibur, which will make him famous as
a poet.
In this symbolic context we should not picture the poet’s hair as it was
in real life (3. 14. 25 n., 3. 19. 25 n.) but rather as the idealized locks of the
poetic victor; similarly the last ode of Pindar’s Olympians (14. 24) ends
Kø . . .
ÆÆ (Oliensis, 1998: 104 n. 3). For crowning as a
closural motif see also Pind. P. 9. 124 f., I. 7. 49 ff., Athen. 15. 674 f citing
Aristotle e b Ø ºæø ØÆ ÆØ (see I. Rutherford in
Classical Closure, ed. D. H. Roberts et al., 1997: 49 ff.).
I N D EX N O M I N V M
(Brackets denote indirect references)
Achaemenium costum, 20 Bactra, 355

Acheron, 43 Baiae, 64
Acherontia, 61–2 Bandusia, 172–3, 176
Achivi, 45 Bantini saltus, 62
Acrisius, 202 Bellerophontes, 118, 170
Aeacus, 231 Berecyntia tibia, 237
Aefula, 349 Bibulus, 341
Aegaei tumultus, 364 Boreas, 287
Aelius, see Lamia Britannus, 67, 83
Aeolium carmen, 375
Aethiops (¼ Aegyptius), 104 Caecubum, 340
Aetne, 77 Caesar (Augustus), 68, 183, 187, 301; see also
Afer (African), 48 Augustus
Africa, 208 Calabrae apes, 209
Africus (ventus), 265, 362 Calais, 137
(Agenor), 329, 333 Callimachus, 7, 305, 376
Albanae herbae, 267 Calliope, 57
Alcaeus, 164 Camenae, 63
Algidus, 266 Campus Martius, 10, 121
(Aloidae), 72 Canicula, 177
Alyattes, 210 Cantaber, 124, 130
(Amphiaraus), 204 Capitolium, 47, 290, 372–3
Amphion, 152 Capra, 117
Andromeda, 352 Carthago, 80–2, 93
Antiochus, 120 Castalia, 75
Antony, 38, 44, 79, 80, 103 (Castor) 364
Apollo, 74, 75; see also Cynthius; Phoebus Cato (censor), 251
Apulus, 86, 207 Cato (pr. 52), 36, (?) 251
Aquilo, 143, 370 Catullus, 134, 255
Arabes, 274 (Cepheus), 352
Aratus, 112 Cerberus, 156–8
Arcturus, 14 Ceres, 33, 279
Argivi, 52 Chius cadus, 231
Assyrium litus, 66 Chloe, 117, 136, 139
Asterie, 115 Chloris, 194
Aufidus, 374 Cleopatra, 44, 104
Augustus, xxi–ii, 22–3, 36, 41, 80, 83, 98–9, Cnidos, 343
180, 298; see also Caesar Codrus, 230
Auster, 40, 326 Concanus, 67
Corvinus, 250; see also Messalla
Bacchae, 307 Cotiso, 129; see also Dacus
Bacchus, 42, 209, 296, 299, (306); see also Crassus, 80, 84
Lenaeus; Liber Crete, 329
380 I N D EX N O M I N V M
Cupido, (168), (335) Hannibal, 110
Cyclades, 343 Hebrus (flumen), 303
Cycnus, (?) 170 Hebrus (iuvenis), 169
Cynthia (Diana), 343 Hectoreae opes, 45
Cynthius, 343; see also Apollo; Phoebus (Helen), 44, 45
Cypriae merces, 362 Hercules, 41, 182
Cyrus, 355 Hesiod, 53, 57, 70, 77, 289
Cytherea (Venus), 168 Hesperia, 103
Hippolyte, 119
Dacus, 104; see also Cotiso Hispana (navis), 31; (ora), 130, 183
(Damocles), 12 (Hypermestra), 160–4
Danae, 201, 203
Danaides puellae, 149–51, 159 Iapyx, 326
(Danaus), 162–3 Ibycus, 192
Daunus, 374 Icarus (insula), 120
Delius Apollo, 75 Ida (mons), 245
Delphica laurus, 377 Ilia, (46), 136
(Diana), 77, 254–60 Ilios, 44, 47; see also Troy
Diespiter ( Jupiter), 34 Ilium, 231
Inachus, 230
Enceladus, 23 India, 274
Enipeus, 120 Ionici motus, 106
Ennius, 36, 374 Italia, 93
Etruscum mare, 356 Italici modi, 375
Euhias (Baccha), 303 Iuno, 36, 43, 52, 74
Europe (mulier), 318–19, 328, 333 Iuppiter, 9, 40, 71, 82, 87, 144, 202, 301, 337
Europe (pars orbis), 48, 337–8 Ixion, 158
Eurus, 216
Keats, 59
Falerna vitis, 20
Faunus, 219–20, 221 Lacaena adultera (Helen), 45
Favonii, 115 Lacedaemonium Tarentum, 96
Fontinalia, 173 Laestrygones, 209, 215
Forentum, 62 Lamia, Aelius, 212–13
Formiae, 209, 215 Lamiae, 214
Fortuna, 360, 361 Lamus, 213
Lanuvium, 230, 321
Gaetula leaena, 241 Laomedon, 45
Galatea, 325 Lar(es), 261–2, 264
Gallica pascua, 209 Latona, 343
(Ganymede), 245 Lenaeus (Bacchus), 308
Geloni, 67 Leo, 353
Getae, 278 Liber (Bacchus), 126, 253
Giganteus triumphus, 9 Libitina, 371
Gigantomachia, 9, 55–6, 72 (Licinius), see Murena
Glycera, 238 (Livia), 183
Graecus trochus, 293 Luceria, 197
Gratia(e), 236, 253 Lucretius, 3–4, 18, 305
Gyges centimanus, 76 Lyaeus (Bacchus), 252
Gyges (mercator), 116 Lyce, 143
Lycia, 75
Hadria, 40, 139, 326 Lycus, 237, 238
Haedus (sidus), 15 Lyde, 154, 159, 340
I N D EX N O M I N V M 381
Lydia, 136, 139 Orion, 76, 325
(Lynceus), 161 Ornytus, 137
(Orpheus), 155–8
Macedo vir (Philip), 204–5 Ovid, 163–4, 347
Maecenas, xxii, 5–6, 123–4, 128, 199–201, Owen, Wilfred, 27
206, 247, 345–8
Magnessa Hippolyte, 119 Pacorus, 103
Manlius, 247 Paeligna frigora, 233
(Marcellus), 185 Palinurus, 65
Maricae litora, 215 Pallas, 73
Mars, 43, 46, 89, 92 Paphos, 343
Marsum duellum, 188 Paris, (44), (45), 47
Marsus, 86 Parthi, 24, 80, 103; see also Medus; Persae
Martiae Kalendae, 125 Patareus Apollo, 75
Martium gramen, 121 Pater (Iuppiter), 359
Massicum vinum, 248 Pelion, 72
Mauri angues, 147 Penates, 183, 269, 332
Matronalia, 125 Penelope, 145
Medus, 48, 130 Persae, 83, 135
Medus rex, 86 Phidyle, 264
Melpomene, 377 (Philip), 204
(Menas), 205 Philippi, 65
(Mens?), 313 Phoebus (Apollo), 52, 58, 254
Mercurius, 152 Pholoe, 194, 195
Messalla, 246–7; see also Corvinus Phrygius lapis, 19
Milton, 31, 49, 189, 323, 329, 341, Pierius, 69, 146
351, 356 Pindar, 54–5, 58, 70, 365
Mimas (Gigas), 72 Pirithous, 79
Minerva, 45, 169 Plancus, 191
Moschus, 319 Poena, 34
Murena, 227, 234 Poeni, 92
Musa(e), 8, 52, 57, 236 Pollux, 41, 364
Mygdonii campi, 210 Porphyrion, 73
Praeneste, 64
Naiades, 306 Priamus, 45, 47
Neaera, 189 Procyon, 353
Nearchus, 242 Proetus, 118
Necessitas, 11, 275 Pullia (?), 60
Neobule, 169 Punicus, 88, 109, 274; see also Poeni
Neptunalia, 174, 338 Pyrrhus (Epiri rex), 110
Neptune, 338, 339, 342 Pyrrhus (iuvenis), 241
Nereides, 342
Nilus, 49 Quirinus, 36, 43
Nireus, 244 Quirites, 51
Nothus, 196
Notus, 116 Regulus, 80–1, 87, 96
Nox, 344 Rhode, 238
Numidae, 163 Rhodope, 304
Rhoetus, 73
(Octavia), 184 Roma, 47, 48, 351
Olympus, 72 Romana Ilia, 136
Orcus, 77, 160, 332 Romane, 101
Oricus, 116 (Romulus), 36, 46; see also Quirinus
382 I N D EX N O M I N V M
Sabelli ligones, 110 Thyna merx, 117
Sabina vallis, 20 Tiberinus, 169
Sabinos (agros), 63 (Tiberis), 169, 356–7
Sappho, 165 Tibur, 64, 349
Scythes(ae), 132, 271–2, 277 Titanes, 70
-icus amnis, 67 Tityos, 78, 158
Seres, 355 Troicus, 46
Shakespeare, 122, 133, 333, 367 Troy, 36–8, 51; see also Ilios
Siculus, 12, 65 Tullus, 127
Sidon, 19 (Tusculum), 350
Silvanus, 354 Tuscus alveus, 121
Simonides, 23, 365–6 Typhoeus, 76
Socraticus, 250 Tyriae merces, 362
Spartacus, 188 Tyrrhenus, 145, 274, 347
(Sthenoboea), 118
Venafrani agri, 96
Tanais, 68, 143, 355 Venus, 145, 163, 202, 223, 253, 313, 335,
Tarentum, 96 343–4
Tartarus, 119 ¼ amor, 138, 177
Telegonus, 350 Vesper, 238
Telephus, 238 Vesta, 86
Tennyson, 351 Virgil, 4–5, 36, 158, 305, 373–4
Thrace, 303 Virgo (Diana), 257
Thurinus, 137 Vulcanus, 74
Thyias (Baccha), 195 Vultur, 59
IN DEX V ER B O RV M
ab, 93, 206, 214 civicus, 282

abesus (?), 357 civilis, 129
acutus, 58, 334–5 clades, 105
additus (custos), 78 classe, 163
aestuare, 158 claustra, 162
altus (Caesar), 68 clavi, 275–6
alumni ( pecoris), 222, 265 cogere, 50, 138
amator, 79, 221 commodus, 235
amice pati, 24 componere, 356
amnes (‘waters’), 358 condiscere, 24
anciliorum, 86 consilium, 69–70, 75–6, 94, 302
anima (amatae), 137 consors, 294
annus, 266 coram, 108
arbiter, 243–4 cornix, 217, 323–4, 325
arcanus, 33, 252 cornua, 177, 252–3, 332, 336
arcere, 6–7 corvus, 323–4
arcus (?), 315 credidimus, 82
arduus, 63, 289, 351 crescere, 267, 372
argutus, 189 creterra, 223
arte (¼ virtute), 41 curtus, 208 (?), 295
asper, 26, 143
ater, 62, 187, 326 debacchari, 51
atqui, 94 decet, decorum, 26–7, 195
attonitus, 236 deducere, 375–6
augur, 204, 217 delectare, 335
aulaea, 352 delicta, 100
aut, 167–8, 281 deripere, 89, 341
avidus (Volcanus), 74 descendere, 10–11, 56–7, 250
deserere, 35
balanus, 348–9 dicere, 57, 154
barbitos, 312 difficilis, 122, 146
beatus, 47, 135, 209, 316, 351 diffingere, 359
bidens (ovis), 268 dirus, 110, 241 (?)
bono die, 249 diva, 259, 316
divinus, 323
caducus, 71 dominus, 17, 207
caecus, 327 duellum, 92, 188
caementa, 17, 274 dulce (mori), 27
capitis minor, 93 dulcis, 126, 167, 176
caprea (¼ ibex?), 196 dumeta, 75, 354
carere, 233–4, 316
carus, 284 edax, 370
catus, 170 efficere irritum, 359
cessare, 237, 333, 342 eheu, 26
cista (?), 188 eius (?), 157
384 I N D EX V ER B O RV M
elementa, 292 in diem, 358
ensis, 12 innare, 216
erilis, 335 innocens, 280
exigere, 187, 367–8 institor, 109
exsomnis, 303 instituere, 127
exsultim, 155 insultare, 47
intaminatus, 29
faece tenus, 198 invicem, 342
fallit beatior, 209 invito vultu, 159
far, 270 iocosus, iocus, 52, 252
fastus (¼fasti), 215 ire (of rope), 145
favete linguis, 7 irritus, 285, 359
fax, 137 iura dare, 48
fertur, 93, 244 iusti . . . dei, 184
feta, 322 iustitia, 38–9
fidei, 116
firmare, 292 labis (?) expertes, 186
foedus, 281 lacerare, 162, 332, 336
foret, 61 laedere collum, 334
fremitus, 327 laetus, 249, 260
frequens, 17 lapides (‘pearls’), 291
funeratus, 127 laqueus (mortis), 277
fustes, 111 latus, 148, 287, 313, 328
laurus, 63, 377–8
geminus Pollux, 364 lavere (¼eluere), 167
gerere, 102, 248 lenis (incedas), 221–2
gestire, 50, 207 lentus (amor), 238–9
glaciare, 144 limen, 60, 148
grandis, 243 limitis (?) Assyrii, 67
liquidus, 64
haerere equo, 293 liquor, 48
hora, 177–8 ludere, 59, 155, 196–7, 224, 360
hornus, 264 ludum dare, 167
horridus, 250–1, 288, 354 luridus, 77, 314–15
hosticus, 25 lustrare, 304
lusum (?) ludere, 360
iactis molibus, 16
idoneus (puellis), 312 madere, 250
iecur, 78 male nominatis (?) . . . verbis, 186
ilex, 178, 266 manare (þ abl.), 158
ilicet (?), 324–5 manni, 322
imbrex (?), 277 materies mali, 291
immanis, 156 maturare necem, 118
immetatus, 279 maturus, 106 (?), 193
immunis, 269 meditari, 107, 131, 260, 301
importunus, 210 medius, 48, 328
impotens, 370 melior sagittis, 105
improbus, 140, 295 melius (adv. iudicantis), 210
inaudax, 242 memor nostri, 325
incedere, 221 merita (of Muse), 376–7
inclinare, 340, 341 metuere, 187, 281
incolumis, 87, 285 miscentur (?) pocula, 235
inconstantior (?) Hadria, 140 mitto, 290, 292
I N D EX V ER B O RV M 385
monumentum, 368 praebere domum, 233
mores (and leges), 286 praesens, 83
movere historias, 119–20 princeps, 216, 375
movere sacra, 249 pro(h), 84
mundae cenae, 351–2 procerus, 307
murreus, 189 prodire, 184
mutuus, 137 profanus, 6–7
promere, 243, 250, 340
narrare, 119, 231 protervus, 155, 190
natum (of wine), 247 publicum (?) mare, 275
neglegere (deos), 34, 103, 250 puella, 185, 258
nenia, 344 purpureus, 42, 198
nequitia, 78, 193 purus, 144–5, 238, 359
nescire (þ inf.), 337
nitidus, 238, 280 quamvis (þ ind.), 121, 147
nitor, 169
nobilis, 178, 197, 213–14 ravus, 322
nodus, 189, 254 recens, 302, 331, 372
nomen, 86–7, 248–9, 330, 337 recreare, 69, 244, 279
non secus ac, 304–5 redemptor, 16
non sine, 63, 176, 312, 357 redonare, 46
nuptiae, 155 regnare, 355, 375
relinquere, 78
o (¼ 0 si), 146, 282 renidere, 104
odi, 6, 237, 285 resignare, 361
olor, 344 retusus (?), 315
operatus, 184 ripae, 306 (?), 327
opibus (¼auxilio), 45–6 rite, ritu, 20, 182, 278, 356
ora (ultima), 130–1, 183 rixa, 190, 236, 248
ordinare sulcis, 10 robigo, 265
oscen, 323–4 ros marinus, 268
-osus, 59, 109, 217, 350 rudis, 26, 293
paelex, 147, 335 saeculum, 84 (?), 105

parcere (þ inf ), 132, 341 saliente mica, 270
parra, 321 saucius, 146–7
pars (mundi), 47, 51, 286 scalpere, 164
pater (urbis), 283 scapha, 363
pauperies, 23, 210, 288, 362 scatere, 328
peccare, 119, 281, 326 sectus orbis, 337
pedes rapiunt, 163 sentire (malam rem), 16, 92, 265
pergula (?), 60 sequi, 334
phaselos, 34 sermones, 126, 250
picus, 325 situs, 369
pius, 59, 248, 270 sospes, 128, 185
plebs (voc.), 183 strenuus, 340
plenus (dei), 299 structae (?) . . . rei, 295
plorare, 144, 331 sublustris, 329
poples, 28 suboles, 177
post, 117, 253 sucus (puellae), 333
postes, 20 super (=de), 129
postgeniti, 284 superbia, 145, 377
potens, 306, 358, 375 supercilium, 9–10
386 I N D EX V ER B O RV M
supinus, 64, 263 ungui, de tenero, 107
supplice vitta, 185 unicus, 183–4
surge, 161 utriusque linguae, 126
suspirare, 26, 118 uxor, 52, 192, 337
tempe, 14 vacuus, 306

temperare, 71, 76, 232, 280 vagax (?), 188
temptare, 31, 66, 76–7, 118, 267–8 (?) vagus, 41, 178, 278, 325
tibia, 57–8, 122, 237 vetulus, 198
torvus, 93–4 vetustus, 214
trepidus, -are, 25, 326, 356 vicarius, 279–80
triformis (diva), 259–60 viola, 146
trochus, 293–4 virtus, 22, 28, 30, 251, 289
tuleris (¼sustuleris), 264 visere, 50, 67, 344
tumultus, 187 vixi, 309, 311, 358
turma, 70, 71 vultus, 39, 94, 159
tuta merces, 31–2
zona, 334
umescere (?) pocula, 236
I N D EX RERV M
ablative absolute, 43, 84, 108, 152, 348 country life idealised, 13–14, 100, 220, 261
ab urbe condita construction, 93, 147, 234
accusative of respect 148; retained, 126, 137 dancing, 106 (bad), 194 (innocent), 226
adjective for genitire, 77, 203, 335 (of peasant)
adventus, 180 degeneration, 100, 112–13
adynaton, 155, 225 diatribe, 273, 289, 361–2
aetiology, 123 dicing, 294
agent-nouns, 73, 79, 242 discontent, 14, 15
alphitomantis, 270 Dog-days, 16, 177, 353
amoebaeum carmen, 133, 342 dowry, 280–1, 362
animals desecrate grave, 47 dreams, false, 331
anniversaries, 123, 127, 224, 260 dye (in metaphor), 90–1
annunciation, 337
apotheosis, 30, 41–3, 46 elections in Principate, 10–11, 29
apposition encloses subject, 288 envy of the living, 284–5
aretalogy, 155–6, 251–3 Epicureanism, 3–4, 132, 346, 358–60
ascension by chariot, 43 epigram, Greek, 175, 246, 255–6, 309–10
assonance, 46, 161 equites, 18, 24–5
asyndeton, 92 (?), 139, 223, 243 ethnography, 271–2
athletes, 166, 169–70 eucharisticon, 5, 200
augury, 217, 323–4, 325 exploration, 50–1
aversio of trouble, 326–7
flower-picking, symbolic, 329
benefactors, 200 founders, legendary, 216, 350, 374
board-games, 130
breaking-off formula, 31, 52 genealogies, 212–13
bronze, 52, 138, 201, 368 genius, 218
bulletins, military, 129–30 gigantomachy, 9, 55, 70–8
goats, 173–4, 177, 223
calque sémantique, 186 go-between, 117
caravans, Scythian, 278
catabasis, 156 hendiadys, 58, 126, 279, 352
catasterism, 301–2 hetaerae, musical, 137, 147 (?), 189, 339
caves, poetic, 69, 300 holidays, 218, 224, 339–40
cena aditialis, 227–8 homosexuality, xxiii, 239–40
childbirth, 258 horsemanship, 121, 170, 293
children in ceremonial, 8, 185–6 hunting, 91, 171, 256, 293
chronographers, 230–1 hymnal style, 152–9, 175, 220, 246–50, 254–5,
civil wars, 101, 104, 187, 282 257–60, 315–16
closure, 20–1, 52–3, 96, 239, 344, 378 hyperbole, 16, 19, 52, 107, 138, 207, 211, 274
colour-contrast, 126, 177, 266, 304, 326
compendious comparison, 19, 113, 274 imperative, future in -to, 190, 336 (?)
concilium deorum, 43, 302 incense, 125, 224, 264
constitutional reform, 354 insomnia, 13, 117
continents, 337–8 interaction in imagery, 162
cookery criticised, 13 invitation-poem, 127–9, 213, 250, 345–6
388 I N D EX R ERV M
kerygma, 9 polar expressions, 12, 66, 222
kings, Eastern, 9, 86, 135, 253 pollution, 34, 321
prayer, posture in, 263
lamps, 108, 128, 254 priamel, 10, 320–1
landscape, imaginary, 59, 69, 300, 306 prolepsis, 206, 295 (?)
latifundia, 10, 207, 210–11 propempticon, 163–4, 317–18, 324–5
lifeboats, 363 proper names, significant, 116, 137–8, 169 (?)
lightning, 40, 203 prosody, irregular, 88, 207, 276–7
lions, 26, 162, 241, 332–3 purple, 19, 110, 352
locus amoenus, 178–9, 353–4 pyramids, 366, 369
lot, 12
lovers: who would die for partner, 137, 140; ransom, 81, 90
pale, 146; neglect duties, 165, 168; rivers: represent country, 68, 143, 355, 374;
compared to soldiers, 312; have rivals, provide personal names, 120, 169;
135, 137; give up the struggle, 148, 190, compared to time, 356; in flood, 357
310 Roman empire worldwide, 48, 187
roses, 197, 237, 348
maenadism, 296–8, 303–4
magnets, erotic, 119 sacrifice, 174, 176, 223, 256–7, 260, 263–7
merchants, 14, 109, 115–16, 288–9, 362 secrets, 32–3, 252
mining, 49 sexism, xxiii, 191–2, 221
monastic corruptions, 225 shopping-lists, 187, 228
moral legislation, 98–9, 283–4 silence, religious, 7, 185–6, 372
mysteries, 7, 33 singular verb with multiple subjects, 41, 202
mythical exempla, 87, 118–19, 149–50, 152–3, smoke of Rome, 102, 351
182, 201, 318–19 sphragis, 364–5
springs, 64–5, 172
nominative ‘in apposition to sentence,’ 243; stars: beautiful, 115, 139, 194, 238; rising and
exclamatory, 330 setting, 15, 117, 177, 326, 353; influence
of, 16
omens, 320–2 stepmothers, 280
onomatopoeia, 179 Stoicism, 28, 30, 36, 358
oracles, 101 storm, signs of, 323, 326–7
oxymoron, 59, 161, 242, 278, 307–8, 328 subjunctive, 94, 111, 349
sufficiency, 14, 211
paraclausithyron, 121–2, 141–2, 309 suicide, 334–5
paraenesis, 113, 131–2, 151, 261, supplicatio, 180–1
340, 346 swimming, 121, 169
parataxis, 90 symposium: preparations for, 125–7, 229,
parenthesis, 84, 152, 160, 214 348; orderly, 129; seriousness avoided,
perfect, gnomic, 34, 269–70, 352 131–2, 228; toasts, 128, 234–6;
perfect infinitive, 226–7 prolonged, 128, 254
perfume, 20, 189, 244, 348–9
personification, 17, 18, 288, 362 teichoscopia, 25
phyllobolia, 225 temples to be repaired, 97, 101–2
pigs, 218, 256, 260, 264–5 thermipolium, 232
place-names, sentimental, 173 three, 52, 234–6, 259
plural for singular, 95 (reditus), 122 (vias), tigers, 42–3, 155–6, 333
338 (nomina) tortoise-shell lyre, 153–4
poet: infancy of 53–4, 61–2; inspired, 59, torture, 94, 251–2
299–307; original, 8, 298, 302, 375; as transferred epithet, 28, 111, 209, 253, 348,
priest of Muses, 7; crowned, 308, 378; 350
survival of, 371 transhumance, 224
I N D EX R ERV M 389
trees in peristyle, 144; for dedications, windlass, 145
255–6; arbor infelix, 334 wine: dated, 128, 247–8, 341; historical
trireme, private, 18 associations, 188, 231–2, 249; imbibes
smoke, 127; ‘languid’, 209, 250; mixed
uncles, censorious, 168 with hot water, 232, 235; its praises,
‘unpoetical words’, 16–17, 87–8, 95–6 247, 251–3
wolves, 143, 225, 321–2
vertical responsion, 76, 147, 183, 258, 326 women: work wool, 168, 197, 335;
villa maritima, 16–17, 274 responsible for fire 111; denied wine,
vistas, 349 167; differentiated from men, 134;
reversal of roles, 195, 238; vetula
wealth rejected, 3–4, 21, 49, 207; jettisoned, derided, 191–2, 193, 198
290–1, 362–3 woods, awe of, 300
wet-nurses, 60
widow’s mite, 262 zones, 50–1, 286–7

R. G. M. Nisbet, Niall Rudd - A Commentary On Horace - Odes Book III (Book 3) - Oxford University Press, USA (2004)

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

R. G. M. Nisbet, Niall Rudd - A Commentary On Horace - Odes Book III (Book 3) - Oxford University Press, USA (2004)

Enviado por

Direitos autorais:

Formatos disponíveis

HORACE: ODES

Th i s work follows the same lines as the commentaries by Nisbet and

Corpus Christi College, Oxford R. G. M. N.

index nominvm 379

This bibliography lists books cited in abbreviated form in the commen-

(a) texts and commentaries

(b) other books cited

ALL Archiv für lateinische Lexikographie und Grammatik (1884–1909).

1. Horace’s early life

2. The date of Odes I–III

3. The ‘Roman Odes’

4. Horace and Augustus

5. Maecenas and other addressees

8. The meaning of the author

14. The arrangement of the book

16. The ancient commentators

1–8. Let the uninitiated depart; I am teaching new chants to a fresh

æEØ = ‹Ø ¼Øæ ØH  º ªø j ªﬁ  c ŒÆŁÆæØ = j

3. Musarum sacerdos: a poet could be compared to a priest for a variety

4. virginibus puerisque canto: the collocation of virgines and pueri

7. clari Giganteo triumpho: the Giants in their arrogance would not

8. cuncta supercilio moventis: the phrase represents a variation of an

10–11. hic generosior / descendat in Campum petitor: the verb suits a

14–15. aequa lege Necessitas / sortitur insignis et imos: aequa makes a

17–18. destrictus ensis cui super impia / cervice pendet: according

18–19. non Siculae dapes / dulcem elaborabunt saporem: Sicilian

20–1. non avium citharaeque cantus / somnum reducent: avium refers

24. non zephyris agitata tempe: the countryman’s valley is disturbed

25–6. neque / tumultuosum sollicitat mare: though H professes to be

27–8. nec saevus Arcturi cadentis / impetus: Arcturus, the brightest

29. non verberatae grandine vineae: hail, sometimes in the summer, is

34–6. huc frequens / caementa demittit redemptor / cum famulis: huc

36–7. dominusque terrae / fastidiosus: the organization of the sentence

37–8. sed Timor et Minae / scandunt eodem quo dominus: Timor et

43–4. nec Falerna / vitis Achaemeniumque costum: Falernian, from

45–6. cur invidendis postibus et novo / sublime ritu moliar atrium?:

1. Angustam amice pauperiem pati: angustam refers to ‘straitened cir-

3–4. et Parthos ferocis / vexet eques metuendus hasta: Augustus at-

6–8. illum ex moenibus hosticis / matrona bellantis tyranni / prospi-

ÆºŒø O f ø, Enn. ann. 418 Sk. ‘matronae moeros complent

15–16. nec parcit imbellis iuventae / poplitibus timidoque tergo: it was

23–4. coetusque vulgaris et udam / spernit humum fugiente penna: the

25–6. est et ﬁdeli tuta silentio / merces: H is imitating the pattern of

26–7. vetabo qui Cereris sacrum / vulgarit arcanae: the mysteries of

27–9. sub isdem / sit trabibus fragilemque mecum / solvat phaselon:

29–30. saepe Diespiter / neglectus incesto addidit integrum: Jupiter

31–2. raro antecedentem scelestum / deseruit pede Poena claudo:

1. Iustum et tenacem propositi virum: iustitia (‘right behaviour’) was

6. nec fulminantis magna manus Iovis: the sky-god’s association with

7–8. si fractus illabatur orbis / impavidum ferient ruinae: orbis here ¼

10. enisus arces attigit igneas: enisus (‘striving upwards’) is obviously

11–12. quos inter Augustus recumbens / purpureo bibet ore nectar: in

13–15. hac te merentem, Bacche pater, tuae / vexere tigres, indocili

15–16. hac Quirinus / Martis equis Acheronta fugit: Quirinus, perhaps

17–18. gratum elocuta consiliantibus / Iunone divis: the council of the

18–21. ‘Ilion Ilion / fatalis incestusque iudex / et mulier peregrina

21–2. ex quo destituit deos / mercede pacta Laomedon: ex quo depends

25–6. iam nec Lacaenae splendet adulterae / famosus hospes: Juno

26–8. nec Priami domus / periura pugnacis Achivos / Hectoreis opi-

29–30. nostrisque ductum seditionibus / bellum resedit: ductum means

30–3. protinus et gravis / iras et invisum nepotem / Troica quem

33–6. illum ego lucidas / inire sedes, discere nectaris / sucos, et

38–9. qualibet exules / in parte regnanto beati: for parte as a region of

æEØ = ‹Ø ¼Øæ ØH º ªø j ªﬁ c ŒÆŁÆæØ = j

ÆºŒø Of ø, Enn. ann. 418 Sk. ‘matronae moeros complent

æ æe æØÆı