Professional Documents
Culture Documents
SERIES ENTOMOLOGICA
EDITOR
K.A. SPENCER
VOLUME45
Host Specialization
in the World Agromyzidae (Diptera)
by
KENNETH A. SPENCER
With illustrations by Ann Spencer
Spencer. Kenneth A.
Host specialization in the world Agromyzidae (Diptera) I by
Kenneth A. Spencer; with illustration by Ann Spencer.
p. cm. -- (Series entomlogica ; 45)
Includes bibliographical references.
ISBN 0-7923-0402-0
1. Agromyzidae--Host plants. 2. Insect-plan~ relationships.
I. Title. II. Series: Series entomologica ; v. 45.
QL537.A4S643 1990
595.77'4--dc20 89-15569
CIP
ISBN-I J:97S-9.t-0 I 0-7.Bti-J e-ISBN- B:97S-9.t-009-IS7.t-0
DOl: 10.1007/97S-9.t-009-1 S7.t-0
Preface IX
Introduction Xl
1. Division BRYOPHYTA 1
2. Divisions EQUISETOPHYTA and POLYPODIOPHYTA 4
Class EQUISETATAE 6
Class FILICOPSIDA 8
3. Division PINOPHYTA 15
4. Division MAGNOLIOPHYTA (ANGIOSPERMS) 17
A. Class MAGNOLIOPSIDA (DICOTYLEDONS) 17
Subclass MAGNOLIIDAE 17
Subclass HAMAMELIDAE 52
Subclass CARYOPHYLLIDAE 62
Subclass DILLENIIDAE 71
Subclass ROSIDAE 95
Subclass ASTERIDAE 179
Order ASTERALES 247
B. Class LILIOPSIDA (MONOCOTYLEDONS) 322
Superorder LILIIFLORAE 323
Superorder ARIFLORAE 337
Superorder ALISMATIFLORAE 337
Superorder BROMELIIFLORAE 340
Superorder ZINGIBIFLORAE 340
Superorder COMMELINIFLORAE 342
Order CYPERALES 344
Order POALES 352
5. Polyphagous species 381
6. Concluding remarks: phylogeny, coevolution, colonization 388
Taxonomic Appendix 394
Acknowledgements 407
References 408
Index 417
Preface
Phytophagous insects represent a very particular not really belong to their host plant range. This may
group of organisms. Firstly, their number amounts lead to mistaken conclusions especially in regions
to more than one quarter of all recent species (ex- where only few observations were possible, as well
cluding fungi, algae and microbes) and together with as in the case of uncommon insect species. Fourthly,
the green plants on which they feed they form al- the great majority (99.4%) of the agromyzid species
most one half of all living species described so far. studied show a high degree of host specialization
Secondly, their overwhelming majority shows very which makes these insects especially suitable for
narrow host plant specialization, that is they feed taxonomic-phylogenetic considerations.
only on one or a few, mostly closely related plant With such an enormous amount of data, it may
species, a characteristic that led J. H. Fabre to elab- have been tempting to draw far-reaching conclu-
orate the notion of the 'insects' botanical instinct' a sions. However, the author has been very careful in
century ago. doing this. One of the great scientific values of this
Considering the above exceptional features it is book is the moderateness of phylogenetic assump-
not surprising that the relations between phyto- tions which are firmly supported by the available
phagous insects and their host plants have been the data, but the author also always clearly indicates
subject of many evolutionary considerations and those questions which remain to be answered. One
discussions. The publications in this domain became can fully agree with the author's reasoning, namely,
especially numerous after the appearance of Ehrlich that the host plant relations found in agromyzid flies
and Raven's (1964) paper expounding the coevolu- do not support Ehrlich and Raven's (1964) coevolu-
tionary theory based on the host plant relations of tionary theory; instead it is concluded that 'the
certain families of Lepidoptera. Unfortunately, the sequential colonization of ever more new hosts
authors of most related publications tried to prove resulting in the present diversity of agromyzid spe-
their assumptions by a few, arbitrarily chosen exam- cies has been substantially facilitated by the chemi-
ples. Therefore, their reasoning was often uncon- cal similarity of new, related hosts but random colo-
vincing and the conclusions drawn had limited nization of unrelated hosts also forms a significant
validity. part of the complex pattern of host specialization
It has always been the general feeling of special- apparent in the Agromyzidae today.' This means
ists working in this field that only a world-wide sur- that in some cases the host plant data of closely
vey of larger groups of phytophagous insects, and related agromyzid species may help botanists in
especially a great number of well documented data clarifying phylogenetic relations of higher plant taxa;
on the host plants, are the prerequisite for suitably however, this method should be used with care.
founded evolutionary considerations on insect-plant The correct phylogenetic reasoning that con-
relationships. A survey of the relevant literature sistently characterizes the whole book is its principal
shows unequivocally that no other publication has but not its only value. This is the first modern mono-
been issued so far that contains such a wealth of graph on Agromyzidae that, besides a vast number
reliable data as Kenneth Spencer's work. Its out- of host plant data, also critically summarizes all
standing value is due to the following points: main results of the taxonomic research on this group
Firstly, it covers more than one half of the known in which the authors take the absolute leadership in
agromyzid species, i.e. the sample is, statistically the entomological community of the world.
speaking, very representative. Secondly, it com- Of considerable value are the figures which,
prises data from all geographical zones and from a together with the host plant data, facilitate the iden-
great diversity of plant communities. Thirdly, since tification of agromyzid species. This is extremely
agromyzids are mining insects, they permit identifi- important not only for those who are interested in
cation of the host plant species more reliably than further research on the taxonomy and ecology of
insects that feed on the surface of the plants, which these insects, but also for applied entomologists.
may occasionally be found on plant species that do Agromyzid species, mainly as introduced pests, are
x Preface
of increasing concern in many regions of the world. of outstanding importance for a better understand-
This volume is therefore an indispensable standard ing of the evolution of insect-plant relationship in
handbook for practical purposes, too. general, and for further taxonomic, ecological and
In conclusion, Kenneth Spencer's monograph is applied research on agromyzids in particular.
Agromyzidae are exclusively plant feeders, popular- esting but not normally directly relevant to this sur-
ly and best known as leaf miners but the larvae may vey. Cronquist's division of the dicots into six sub-
form external stem-mines, bore internally in stems classes is perhaps somewhat arbitrary and differs, in
of herbaceous plants or in the cambium of trees, or some cases significantly, from the systems proposed
feed in roots or flower-heads. They are virtually ubi- by other workers, particularly Thorne, but at least it
quitous and are known from northern Greenland to provides a clear framework within which one can
Tierra del Fuego in South America and subantarc- work. The host specialization of the Agromyzidae in
tic islands south of New Zealand. They successfully most cases provides no evidence conflicting with
colonize hosts on the seashore and on mountains up Cronquist's arrangement but in a few cases it does
to the snowline, and with special adaptations are (such as the inclusion of the Proteaceae and Elaea-
even found in deserts. gnaceae in a single order Proteales). Thorne's pro-
A wealth of information is available on the host posals also in some cases accord better with the
specialization of the Agromyzidae but this is scat- relationships of the Agromyzidae on particular
tered throughout countless publications in many families. These cases are discussed individually in
languages. Basic facts, such as the number of plant the main text below.
families colonized, the proportion of species with Where recent major taxonomic revisions are
known hosts and particularly the evolutionary rela- available, such as for the Leguminosae (Polhill and
tionship, if any, between the agromyzids and their Raven, 1981), for the Monocotyledons (Dahlgren et
hosts, have hitherto been the subject of mere specu- al., 1985) and for the grasses (Clayton and Ren-
lation. Here I present a factual account of the rela- voize, 1986), it has seemed appropriate to follow
tionship between all species whose hosts are known these but with comments as appear desirable where
and their host families. In some cases species whose they depart significantly from Cronquist.
hosts can be deduced with certainty, primarily from With Agromyzidae known on liverworts, horse-
their male genitalia, have also been discussed. I treat tails, ferns and gymnosperms, I found that there is
the species in the evolutionary progression of their considerable divergence among botanists on the
hosts, from primitive Hepaticae (liverworts) to nomenclature of higher taxa at the level of Divisions
Equisetaceae (horsetails), Polypodiophyta (ferns), and Classes. Cronquist, Takhtajan and Zimmerman
Pinatae and Gnetatae (Gymnosperms) through the (1966) attempted to rectify this unsatisfactory situa-
dicotyledons and monocotyledons. This work has tion. They wrote: 'There is now a bewildering
shown that 160 plant families in the dicots and plethora of systems and partial systems, each of
mono cots have been colonized out of a total of which may lay some claim to being the best repre-
some 380 which are currently accepted, and ap- sentation of the similarities, differences, and evolu-
proximately 50% of Agromyzidae have known hosts tionary relationships within all or some part of the
(see Section 6) out of the world total of 2500 plant kingdom.' They proposed what they hoped
(Henshaw and Howse, in press). might become a universally accepted, standard sys-
At the beginning of this study I had to decide tem of classification for the whole plant kingdom.
which of several available classifications of the Their proposals have found only limited acceptance
angiosperms I would follow. Differing systems have but I have nevertheless found it useful to follow their
been proposed by Takhtajan (1969), Cronquist classification in the discussion below of the four
(1981), Thorne (1983) and Goldberg (1986). For non-angiosperm groups on which there is limited
an entomologist Cronquist's system is by far the colonization by agromyzids.
most useful, treating every accepted family of dicots It quickly became clear that it was inadequate
and monocots in a logical evolutionary sequence merely to associate agromyzid species with indi-
and providing exactly the background information vidual host families. The families needed to be sub-
required on fossil history, relationships and distribu- divided into subfamilies, tribes and genera and only
tion. The detailed floristic data also given are inter- then was the host specialization of the agromyzids
xii Introduction
fully revealed. The clearest way of presenting this studied since their original description early in work
information, apart from the more detailed discus- on the family in the 1920s and 1930s, which are
sion within the main text, is the provision of Tables now mainly in the Museums in Berlin and Vienna.
for each plant family, including the subfamily and Consequently some revisionary taxonomic work has
tribal divisions and listing all genera with their as- been inevitable and this has resulted in the discovery
sociated agromyzids. These Tables effectively repre- of a number of new species, new synonyms and new
sent an identification guide to species throughout combinations. For convenience, all such new taxo-
the world once the host has been accurately identi- nomic information and formal changes in nomencla-
fied. Within the text, for most species I have also ture are presented together in an Appendix.
given one or more illustrations of the male genitalia Studies in phytochemistry have made great ad-
which will permit a positive identification to be vances in recent years and a complete account of the
made of species initially identified from their host chemistry of all plant families is provided by Frohne
association. This practical aspect did not form part u. Jensen (1985). Unfortunately time has not per-
of my original plan for this work but I now realise it mitted the detailed correlation between agromyzid
is an important by-product, which will be of real host specialization and the chemistry of their hosts
value to agricultural entomologists who, for the first which is obviously desirable. It is to be hoped that
time, will now have available in a single volume a this aspect will be investigated further by future
tool for instant identification of any leaf miner found workers.
in any part of the world. Experience has shown that the male genitalia of
The host specialization of individual species is it- the Agromyzidae give a clearer indication of evolu-
self striking but equally remarkable is the variation tionary relationships than any other complex of
in host selection by the different agromyzid genera characters and therefore in the limited space avail-
which has hitherto not been fully appreciated. This able I have provided illustrations of the genitalia of
is discussed further in Section 6 (Table 93) but is most species. Even when species appear identical on
already indicated in a further set of Tables given at external morphology, the genitalia can immediately
the beginning of the discussion on all the larger host reveal their distinctness. One striking example of
families, in which the genera present on the family this is the totally different genitalia of two of the leaf
are listed with the number of species in each. To cite miners on Heracleum (Apiaceae) - Phytomyza
one example here of the disparity noted on families spondylii (Fig. 653) and Phytomyza heracleana (Fig.
of comparable size, Phytomyza which is the largest 659).
agromyzid genus has just 6 species on the Legu- I have completed this study in 2 years. Ideally
minosae but 64 on the Apiaceae, the two families more time should have been devoted to it. Inevitably
being in the same Subclass Rosidae. there will be some omissions and some mistakes will
In contrast to the high proportion of species have been made. I can only apologise for these. With
which are restricted in their feeding to a single no funding for this work apart from the limited
family or genus, only 16 species (0.6% of the total) resources of the Spencer Foundation, I have been
are truly polyphagous, feeding on a number of un- under pressure to complete the work as quickly as
related families. A few feed on just one or two un- possible. Much new detail has been provided, an
related hosts. This in some cases may be construed overall picture of the world host specialization of
as accidental or in others, where such feeding is not the Agromyzidae is presented and initial conclu-
abnormal but is encountered regularly, it may well sions are drawn.
represent the first stages of incipient polyphagy. All As Professor Hering said to me on a number of
these species are discussed in Section 5. occasions, our knowledge of the Agromyzidae can
During the preparation of this study I have con- only advance 'schrittweise' (step by step). This study
sulted virtually all the literature with details of agro- I feel represents one further useful step forward and
myzid hosts and have reassessed my own world- I hope will encourage further study of this complex
wide collection of leaf mines (and in part also and fascinating family. I also hope the data provided
Prof. Hering's mine collection now in the British will be of value in general studies of insect-plant
Museum). It has also been necessary to examine relationships.
types of a number of species which have not been
1. Division BRYOPHYTA
Agromyzidae are only known on the classes Antho- pears like a typical Liriomyza, with the head largely
cerotae and Hepaticae (liverworts), none on the yellow but the third antennal segment slightly
Musci (mosses). darkened apically, the mesonotum 5-banded and
Stewart (1983) states that an unequivocal fossil the scutellum entirely yellow. The description was
bryophyte is known from compressions in rocks of based on a single male which, when it was recently
Upper Devonian age (350 million years before re-examined by Ing. M. Martinez at INRA, Ver-
present). There is apparently an unusually high sailles, was in poor condition. Examination of the
degree of similarity between the extant genus Palla- male genitalia showed the species to belong in the
vicinia and Pallavicinites devonicus. Further 'beauti- genus Phytoliriomyza, to which it was transferred by
fully preserved specimens' of liverworts are known Spencer and Martinez (1987).
from the Carboniferous. Additional fossils are I briefly examined the genitalia of the holotype
known in the Triassic, and yet others from Oligo- during a visit to Versailles in 1987 and later ar-
cene Baltic amber are assigned to modern genera. ranged for it to be sent to me in Cornwall for illus-
Agromyzidae are known on the two orders tration. Most regrettably it failed to arrive and was
Marchantiales and Anthocerotales. Fossil Hepati- lost by either the French or British postal authori-
cites cyathoides from Middle Triassic shows unequi- ties. Sellier (1945) re-discovered Phlir. mesnili in
vocal relationships with the Marchantiales (agro- Brittany on a non-aquatic liverwort, Riccia beyri-
myzids known from Mexico). Among a 'fine fossil chiana and (1947) gave an excellent and detailed
flora of liverworts' from Sweden Ricciopsis bears a description of its biology, recording that the larva
striking resemblance to modern Riccia (agromyzid pupates internally in the thallus where it has been
known from France). This all follows 'a predictable feeding. From Sellier's many illustrations I repro-
pattern where many genera of Tertiary floras bear duce here his figure of the male genitalia (Fig. 1),
marked resemblance to extant ones.' which, although lacking detail, at least shows the
Today the liverworts are known to occur form of the aedeagus.
throughout the world, with some 8000 species in ca. Although Riccia natans has recently been trans-
n families, classified in eight orders. The known ferred to Ricciocarpus it is worth commenting that
species of Agromyzidae on the liverworts are listed Phlir. mesnili finds no difference between these two
in Table 1. In view of the large number of possible genera and appears to feed equally happily on both.
hosts, it can be expected that many more agromyzid It will be interesting to discover in due course with
species will be discovered in the future. further collecting whether the species has an even
Although only a single species of Agromyzidae wider host range.
has been described from Hepaticae, felicitous co- Hering (1957c) recorded mines in the thallus of
operation between Prof. Hering in Berlin and Prof. Megaceros spp. from the West Indies and New Zea-
J. Proskauer in the Department of Botany, Univer- land, and gave a detailed description of a dead larva
sity of California, Berkeley resulted in the descrip- found in Megaceros sp., Juan Fernandez Is., Masa-
tion by Hering (1957c, 1966) of larval and pupal fuera, Loberia South, Chorro de las Cabras,
remains from herbarium specimens discovered by 2l.ii.1955, referred to by Hering (1966) as 'sp.a.'
Proskauer at Berkeley. This work shows that Agro- Striking features of this larva are the form of the
myzidae are present on liverworts probably cephalo-pharyngeal skeleton (Fig. 2) and the un-
throughout the world, with records from the West usually large right-hand 'tooth' of the mouth-parts
Indies, Mexico, Peru, Juan Fernandez Islands, New (Fig. 3).
Zealand, and of course France where adults have Hering (1966) described a puparium in the thal-
been reared. lus of Megaceros vincentianus from Peru as sp.B.
Aguilar (1944) described as Liriomyza mesnili a The most interesting feature is the extreme elonga-
species reared from Ricciocarpos (as Riccia) natans tion of the anterior spiracles (Fig. 4) which penetrate
from Versailles, France. Externally, this species ap- through the epidermis of the thallus. Sp. c. was de-
2 Division BRYOPHYTA
Table I
Class ANTHOCEROT AE
Anthocerotaceae
Megaceros sp. a. (Juan Fernandez Islands),
sp. b. (Peru), sp. (West Indies), sp. (New Zealand)
Class HEPATICAE
Order MARCHANTIALES
Marchantiaceae
Dumortiera sp. (Mexico)
Monociea sp. c. (Mexico)
Ricciaceae
Riccia Phytoliriomyza mesnili
Ricciocarpos Phytoiiriomyza mesnili
scribed and illustrated from shallow, upper surface species [sic] feeding on Hepaticae can be considered
blotch mines in Monoclea sp., Mexico, Oaxaca as living primarily on liverworts, and that they have
State, 1300-2100 m. Mines in Dumortiera hirsuta, not transferred secondarily to liverworts from
also from Mexico, were recorded but these were angiosperms on which the majority of the other
entirely empty. numerous species in the genus occur.'
Commenting on the double arms of the main In view of the dominant occurrence also on ferns
ventral process of the cephalopharyngeal skeleton in of species in Phytoliriomyza, I feel it cannot be a
sp.a. (Fig. 2), Hering (1957c: 50) writes (in Ger- coincidence that this genus, which is now accepted
man): 'One might be inclined to see in this Lirio- as primitive, feeds on both Bryophyta and Pterido-
myza [sic] a particularly primitive form at the border phyta. It thus seems very probable that the coloniza-
of the Agromyzinae and Phytomyzinae, and the oc- tion of these two classes is not secondary. I have
currence of the species on the Hepaticae not as detected no angiosperm hosts from which Phyto-
secondary host selection.' Later, referring to Phlir. liriomyza might have transferred to these more
mesnili and sp. a., Hering comments: 'The structure primitive groups and I am therefore inclined to sup-
of the mouth parts of the larva of the two species, port Hering's tentative suggestion that the presence
combined with their feeding on liverworts, perhaps of Phytoliriomyza on liverworts represents an early
justifies the conclusion that these two Liriomyza primary colonization.
2. Divisions EQUISETOPHYTA and POLYPODIOPHYTA
Here I am considering the two classes SPHENO- Willis (1985) usefully includes a list of the 48
PSIDA (Equisetatae, horsetails) and FILICO- family names which have been proposed for extant
PSIDA (ferns) which provide hosts for Agromyzi- Pteridophyta prepared by Pichi-Sermolli (1970) but
dae. only some 40 families are now accepted, grouped in
Horsetails and ferns formed a dominant part of ten orders. Of these, only ten in six orders are
the flora during the Mesozoic long before the advent known to support Agromyzidae. The number of
of Agromyzidae. For instance Lovis (1977) quoting modem species is estimated at between ten and
an earlier author writes: 'Ferns are so abundant in 12000.
the Yorkshire Jurassic flora that they may well have The Agromyzidae known on horsetails and ferns
formed the dominant herbs on land.' It is generally are shown in Tables 2 and 3.
true that ferns are prominent in the Mesozoic until
the Lower Cretaceous, whereafter they diminish in Table 2
importance, no doubt in direct consequence of the
diversification and increasing abundance of angio- Agromyzid genera on SPHENOPSIDA (horsetails)
and FILICOPSIDA (ferns)
sperms in the latter half of the Cretaceous period
(Lovis, 1977: 236). Restricted Polyphagous
From the Tertiary a great many fragmentary fern to class
fossils are known, a high proportion of which have
been attributed to modem genera but Lovis (1977: SPHENOPSIDA
250) rightly urges caution in such identifications Liriomyza 3
' ... when one considers the difficulties that botanists FILICOPSIDA
can experience in identifying with certainty some
Tropicomyia 2 (undescribed)
modem ferns even when supplied with good fertile Phytoliriomyza 10 (2 undescribed)
specimens, it makes no sense to give a specific name Chromatomyia 4
to a minute fragment of a sterile imprint. The attri-
bution of sterile debris to living genera such as 19
Asplenium, Dennstaedtia and Dryopteris can only
cause confusion.' It appears that the situation With suitable hosts of horsetails, ferns and dicots
changed radically early in the Tertiary, when ac- available for the emerging Agromyzidae from the
cepted records of a number of extant families Late Cretaceous on to the Tertiary, it is puzzling
became available. that so few species succeeded in colonizing horse-
'There is at present no evidence to refute the con- tails and ferns. However, it is well known that few
clusion that the bulk of modem families... are of insects are found feeding on ferns. This is almost
remarkably recent origin, their main radiation certainly associated with their specific chemistry.
materializing in the fossil record even later than the Although a great deal of work has already been
appearance of modem flowering plant families.' done, Swain and Cooper-Driver (1973) write: 'Bio-
(Lovis, 1977: 313.) chemical studies on ferns are only in their in-
The classification of the ferns is still in a state of fancy .... Our knowledge of biochemical pathways is
flux. The oversimplified arrangement proposed by almost negligible and even that of secondary metab-
Engler & Prantl has been followed by six systems of olites fragmentary.'
classification since 1938 but no one system has
been adopted universally. Some of the existing dis-
parities are commented on under the relevant fami-
lies below. The arrangement I follow here is basical-
ly that proposed by Crabbe, Jermy and Mickel
(1975).
Class SPHENOPSIDA (EQUISETATAE) horsetails 5
Table 3
Order EQUISETALES
Equisetaceae
Equisetum Liriomyza equiseti, L. occipitalis, L. virgo
Order MARATTIALES
Angiopteridaceae
Angiopteris Tropicomyia sp. n. ( India)
Marattiaceae
Marattia ? Phytoliriomyza sp. n. ( Hawaii)
Order POLYPODIALES
Polypodiaceae
Microsorium Phlif. flavopleura ( N. Z. )
Pleopeltis Tropicomyia sp. ( Cameroun)
Polypodium Chromatomyia scolopendri ( Europe)
Grammitidaceae
Grammitis Phlif. flavopleura ( N. Z. )
Order PTERIDALES
Adiantaceae
Cheilanthes Chf. cheilanthus ( India)
Dennstaedtiaceae
Pteridium Phlif. clara ( U.S.A. ), Phlif. hilarella, Phlif. pteridii ( Europe)
Histiopteris Phlif. kuscheli ( luan Fernandez Is.)
Cyatheaceae
Cyathea Phlif. cyatheae ( N. Z. ), Phlif. sp. n. ( Argentina, Guadeloupe)
Phlif. tearohensis
6 Divisions EQUISETOPHYTA and POLYPODIOPHYTA
Order ASPIDIALES
Aspleniaceae
Asplenium Tropicomyia sp. ( Cameroun ); Phlir. flavopleura ( N. Z. ), Phlir. felti
( U.S.A. ); Chr. dryoptericola, Chr. scolopendri
Camptosorus Phlir. felti
Ceterach Chr. dorsata
Phyllitis Chr. scolopendri
Dryopteridaceae
Diplazium Phlir. diplazii ( Malaysia)
Dryopteris Chr. dryoptericola ( Japan)
Davalliaceae
Nephrolepis T. polyphaga ( India)
8 9
Figs. 8, 9. Liriomyza occipitalis: 8, posterior spiracles of larva; 9, aedeagus.
10
Fig. 10. Liriomyza nordica: aedeagus. Fig. 11. Liriomyza freyella: aedeagus.
Hering (1962: 42) suggested that there might be a Marattiaceae (7 genera, 100 + species)
fourth species feeding on Equisetum and this could
possibly be L. jreyella which was caught together In Ha~aii an undescribed species, known only from
with L. occipitalis on Equisetum in Finland leaf mmes on Marattia sp., was recorded by Swezey
(Spencer, 1976a: 250). It will be of the greatest (1954: 90). Hardy and Delfinado (1983: 211) sug-
interest in due course to discover the hosts of gested that the species was probably Liriomyza coc-
L. jreyella, virgula and nordica, as this information culi. ~Frick) which feeds on a wide range of dicot
could provide valuable evidence of relationships of famihes but I consider this is improbable and the
the Equisetum feeders with species feeding on species is almost certainly an undescribed
angiosperms. Phytoliriomyza sp. The genus Tropicomyia which
The known colonization of Equisetum by three has species on ferns in Mrica and India does not
Liriomyza species remains puzzling and their exact occur in Hawaii.
affiliations must remain a matter for further study
and speculation. Order POLYPODIALES
Interestingly, few Liriomyza species are known
on early angiosperms, for example five on Ranun- The group of seven families on which agromyzids
culaceae, two on Hamamelidae, four on Caryophyl- are known are here treated within the single order
lidae, four on Dilleniidae, none on Rosaceae, but 75 Polypodiales, although a relatively recent system
on Asteraceae. This proliferation on the Asteraceae (Willis, 1985) treats this group as a subclass Filici-
reflects the many new hosts which evolved in this dae, divided into 14 orders; Lovis (1977) accepts 13
modern family during the later Tertiary. orders, while Jones (1987) increases the number to
20. This adequately illustrates the lack of consensus
among pteridologists on the higher classification of
Class FILICOPSIDA/POLYPODIATAE this group.
(ferns)
Polypodiaceae (50 genera)
Order MARATTIALES (4 families) It is now agreed that the early delimitation of Poly-
This is an old group with fossils identified in the podiaceae (sensu Engler) does not represent a
Triassic rocks in Greenland. Leaves have been natural group and the 'polypodiaceous' ferns are of
assigned to Marattiopsis which have been recorded polyphyletic origin. Early, Mesozoic records of such
as differing from extant Marattia 'only in age.' From fern~ are ~ot now accepted as representing Poly-
the Triassic to the end of the Jurassic the Marattiales podlaceae m the modern sense (Lovis, 1977). It is
reached their climax. clear that the limits of the family and particularly of
Agromyzidae are known on the two families the genus Polypodium need further study.
Angiopteridaceae and Marattiaceae. Three agromyzids are known on Polypodiaceae.
In Europe Chromatomyia scolopendri occurs on
Polypodiu"! vulgare but perhaps more commonly
Angiopteridaceae (1 genus, 100 species) on Asplenzum and Phyllitis (see below). In New
The Angiopteridaceae (sole genus Angiopteris) were Zealand Phytoliriomyza Jlavopleura has been re-
earlier included within the Marattiaceae but are corded on Microsorium (as Phymatodes) scandens
more recently treated as a separate family, with the (Polypodiaceae) but also on Grammitis (Gramrniti-
main species A. evecta split into as many as 100 d~ceae) and Asplenium, and both these species are
microspecies. A. evecta s.l. occurs in Madagascar, dlscussed further under Aspleniaceae below.
tropical Asia and Polynesia and is a characteristic De Meijere (1940: 170) records a mine on an
tree fern, with a massive stem up to 2 m high, with epiphytic fern, Pleopeltis sp. (as Pleiopeltis), on Mt.
fronds as long as 3 m. Came~oun, Cameroun (as Melanagromyza spec.)
A long, linear leaf mine of an obviously unde- and thls clearly represents a Tropicomyia. It is prob-
scribed Tropicomyia sp. (Fig. 13) was discovered on ably the same species recorded near the same local-
an immature plant in forest in Kerala when I visited ity on Asplenium (see below). It seems likely that
southern India in March, 1985. This host was identi- these records refer to an undescribed species.
fied by A. M. Paul (B.M., Nat. Hist.) as 'either
Angiopteris evecta or Marattia jraxinea, probably Grammitidaceae (11 genera, 500 species)
the former.' This is the fourth record of a Tropico-
myia on a fern and probably represents a switch This family was formerly included in Polypodium
from an ancestor of Tr. polyphaga, which has been and consists of small epiphytes. Phytoliriomyza
recorded on Nephrolepis in northern India (see Jlavopleura has been recorded in New Zealand on
Davalliaceae below).
ClassFILlCOPSIDA/POLYPODIATAE 9
Grammitis billardieri but is discussed further under Phytoliriomyza hi/are/la (Zetterstedt, cf. Spencer,
Asplenium below. 1976a) forms short leaf mines on bracken and is
w.idespread throughout Europe. Surprisingly, it was
dIscovered that a second species, Phlir. pteridii
Order PTERIDALES Spencer (1973b) forms virtually identical mines on
bracken, also in Europe. Distribution of the two
Adiantaceae species overlaps but I am not aware of these two
species actually occurring on the same plant.
Phytomyza cheilanthus Garg (1971) was reared
Although they can only be separated with difficulty
from Cheilanthes virga-aurea in northern India. The
on external morphological characters, the male geni-
species pupates in the leaf and the illustration of the
talia of P. hi/are/la (Fig. 14) and P. pteridii (Figs. 15,
genitalia given by Garg (1971: Fig. 3j) is not incon-
16) are entirely distinct.
sistent with Chromatomyia, to which the species is
In the United States Phlir. clara (Melander, cf.
formally transferred in the Appendix below (see
Spencer and Steyskal, 1986) also feeds on bracken
also discussion of Chromatomyia scoLopendri under
and the genitalia (Fig. 17) show that it is closely re-
Aspleniaceae).
lated to P. hilare/La and appears to have diverged in
North America from a common ancestor.
Order DENNSTAEDTIALES Bracken is virtually cosmopolitan and it will be of
great int~rest to discover with further collecting
Dennstaedtiaceae (24 genera, 410 species) whether It supports additional species of Agromy-
zidae elsewhere in the world.
Also formerly included in Polypodiaceae, this family
Phytoliriomyza kuscheli (Spencer, 1964b, as Li-
was erected recently (Pichi-Sermolli, 1970) and now
riomyza) (transferred to Phytoliriomyza by Spencer,
includes Pteridium, the common bracken, on which
1982) was described from 'Grutas' (caves) on Masa-
3 agromyzids are known. One further host-specific
fuera, Juan Fernandez Islands, 17.ii.Sl (leg.
species is known on Histiopteris in the Juan Fernan-
Kuschel). When describing this species I had no
dez Islands.
evidence of the host. However, during a fortuitous
14
Fig. 13. Tropicornyia sp.n.: leaf mine on Angiopteris evecta Fig. 14. Phytoliriornyza hilarella: aedeagus.
15
Figs. 15, 16. Phytoliriornyza pteridii: 15, aedeagus, side view; 16, distiphallus, ventral view.
10 Divisions EQUISETOPHYTA and POLYPODIOPHYTA
~
''' ...
.~ : ..'
1/'17
~,
20
Figs. 20, 21. Phytoliriomyza cyatheae: 20, aedeagus, side view; 21, same, ventral view.
22
Figs. 22-24. Phytoliriomyza sp.n. (Argentina): 22, anterior spiracles of puparium; 23, aedeagus, side view; 24, same, ventral view.
Class FILICOPSIDA/POLYPODIATAE (ferns) 11
meeting with Dr. Kuschel at the British Museum of the two (Figs. 23, 24 and 20, 21), it seems prob-
(Natural History) on 10 June, 1987, he mentioned able that they are directly related. Cyathea was cer-
that the type specimens were caught on the fern His- tainly present on Gondwanaland before its breakup
tiopteris and were certainly breeding on the plants. and the present populations of Cyathea in South
Josephine Camus at the Herbarium, BM (NH) kind- America and the Caribbean, and in New Zealand
ly identified from the literature the only Histiopteris are clearly derived from ancestral Gondwanan spe-
present on the Juan Fernandez Is. as H. incisa. The cies. The presence of related Phytoliriomyza species
male genitalia (Fig. 18) and epandrium (Fig. 19) on Cyathea in these two areas suggests an early ori-
confirm the generic position of this species in Phyto- gin when direct dispersal between West Africa and
liriomyza but give no indication of its nearest rela- South America was still possible.
tive. During a recent visit to Kenya I searched on the
local Cyathea, C. manniana, at three localities for
leaf mines, but unfortunately unsuccessfully. I still
Order CYATHEALES
suspect that an intermediate population between
that in South America and New Zealand is present
Cyatheaceae (2 genera, 625 species)
in Africa. If and when this can be discovered, it will
Stems of arborescent ferns, such as Cyathocaulis be exciting evidence suporting my suggestion of the
and others, are known from late Jurassic and early early dispersal of the ancestral Cyathea leaf miner
Cretaceous in India and Japan. These tree ferns from Gondwanaland both to the East (south-east)
have been split into two orders, the Dicksoniales and West.
and Cyatheales, and Pichi-Sermolli (1970) recog- Also in New Zealand, I caught a single male Phy-
nised a whole series of families of tree ferns. More toliriomyza, described as Phlir. tearohensis Spencer
recently these have been united in the single family (1976b), on Cyathea dealbata on North Island, and
Cyatheaceae. Lovis (1977: 294) writes: 'Overall, the with no other suitable plants in the vicinity, I am
cytological evidence favours the unification of the satisfied that this was the host. The male genitalia
tree-ferns into a single family.' (Fig. 25) appear to indicate closer relationship with
The family is widespread in the Old and New Phlir. pteridii in Europe (Fig. 15) than with any pos-
World tropics, extending to Guadeloupe in the sibly related species in New Zealand or Australia.
Caribbean and South Africa and New Zealand in The exact affiliations of this species remain prob-
the southern hemisphere. lematical.
Phytoliriomyza cyatheae Spencer (1976b) was
described from leaf mines on Cyathea dealbata and
Order ASPIDIALES
C. smithii on both North and South Islands, New
Zealand. This is a relatively large species, with wing
Aspleniaceae (78 genera, 2700 species)
length from 2.5-2.9 mm. The aedeagus (Figs. 20,
21) is distinctive, but the epandrium and surstyli Leaf remains suggesting affinity with Asplenium
(Spencer, 1976b: fig. 84) are characteristic of the have been reported from the Cretaceous but the
genus, with a group of strong bristles at the hind genus is not well represented until the Tertiary (Ste-
comer of the epandrium and a 'comb' of even longer wart, 1983: 208).
bristles on the inner margin of the surstyli. There is still no consensus on the taxonomic
When collecting in Guadeloupe in Jan./Feb. limits of this family. For example, Willis (1985)
1987 I discovered many mines on Cyathea arborea treats Dryopteris and Diplazium as in the Aspi-
but unfortunately all were empty but two contained diaceae (= Dryopteridaceae), while Mabberley
dead larvae. The anterior spiracles are unusual, each (1987) splits the Aspidiaceae into six subfamilies,
having numerous minute pores on two separate pro- including Asplenioideae, Athyrioideae and Dryop-
jections. I have also seen adults reared from an un- terioideae. Here, I accept the Aspleniaceae and
identified fern at Tucuman, Argentina by P. Wygod- Dryopteridaceae as separate families, each with
zinsky, 20.x.53. The anterior pupal spiracles (Fig. their own distinct agromyzids (see Table 3).
22) are of the same form as in the larvae from Gua- In New Zealand Phytoliriomyza flavopleura
deloupe and it is possible that the same species is (Watt, cf. Spencer, 1976b) is not uncommon on As-
present in the two areas. The species represents an plenium !alcatum, A. flaccidum and A. lucidum and
undescribed Phytoliriomyza sp., which will be de- has also been recorded on Grammitis billardieri
scribed in due course by Dr. G. Valladares, with its (Grammitidaceae) and Microsorium (as Phyma-
name formed from 'helecho,' Spanish for fern. The todes) scandens (Polypodiaceae). The male genitalia
adults of this species are not obviously distinguish- (Figs. 26, 27) indicate considerable divergence from
able from P. cyatheae in New Zealand and, although P. cyatheae on Cyathea (Figs. 20, 21) but the two
there is significant difference between the genitalia species could be directly related.
12 Divisions EQUISETOPHYTA and POLYPODIOPHYTA
In North America (Illinois and New York) Phyto- sis in New Zealand (Fig. 25). Frost (1924, Plate V:
liriomyza felti (Malloch, d. Spencer and Steyskal, Fig. 3) illustrates the leaf mine on Camptosorus,
1986) feeds on hosts in two genera, Asplenium pin- which appears to be a linear-blotch. The larva is
natifidum and Camptosorus rhizophyllus. The geni- recorded as pupating in the mine.
talia (Fig. 28) suggest relationship both with Phlir. In Europe a species known hitherto as Phytomyza
pteridii in Europe (Fig. 15) and with Phlir. tearohen- scolopendri R.-D. (cf. Spencer, 1976a) feeds com-
;-~ .....
~::: ".;:--
"'-- ;~;.-
26
, ....
Fig. 28. Phytoliriomyza!elti: aedeagus.
29
30
Fig. 29. Chromatomyia dorsata: aedeagus. Fig. 30. Chromatomyia scolopendri: aedeagus.
31
Fig. 31. Phytoliriomyza diplazii: aedeagus.
Class FILICOPSIDA/POLYPODIATAE (ferns) 13
monly on Asplenium ruta-muraria and Phyllitis sco- sider that this is probably the same species which he
lopendrium (also on Polypodium vulgare). I found found on Pleopeltis (Polypodiaceae, see above).
mines of this species on Phyllitis in August, 1988 in
woodland in North Devon, England and re-exami-
Dryopteridaceae
nation of the puparia in the mines strongly suggests
that this species should correctly be included in the Dryopteris first appears in the fossil record in the
genus Chromatomyia. The 'slipper-shaped' puparia Tertiary. Although this family is here accepted as
and the habit of the larva to pupate upside down in distinct, the main genus Dryopteris was included in
the leaf, with its anterior spiracles projecting the Aspidiaceae by Willis (1985) and in the Asple-
through the epidermis are characteristic of this niaceae by Mabberley (1987), while Diplazium was
genus. This new combination is formally established included in the Athyriaceae by Willis, and in the As-
in the Appendix below. The male genitalia (Fig. 30) pleniaceae by Mabberley. This further illustrates the
are somewhat atypical of Chromatomyia, as also is problems fern taxonomists have in the classification
the yellow scutellum but this is known in one North of this difficult group.
American species, Chr. compta Spencer (in Spencer In Japan (Mt. Hiko, Kyushu) Sasakawa (1961)
and Steyskal, 1986), the host of which has not been described Phytomyza dryoptericola from Dryopteris
discovered. lacera, the larva forming a narrow mine following
A second European species, closely related to the leaf segments and making a 'pupal blister' at the
Chr. scolopendri, is only known on Ceterach offici- end of the mine. This is a largely yellow species but
narum in Dalmatia - Chr. dorsata (cf. Hendel, 1935 the mesonotum has '3 brown stripes.' Both in gener-
and the Appendix below). The male genitalia of al colour and particularly in the type of pupation, it
these two species (Figs. 29, 30) confirm beyond seems clear that this species is related to Chr. scolo-
doubt that they are sister-species and it is thus cer- pendri and it is formally transferred to Chroma-
tain that the group of their fern hosts were originally tomyia in the Appendix below. Although the holo-
colonized by a single ancestor. The population type is a male, the genitalia have not been examined
adapted to Ceterach appears to be confined to this and the affiliations of the species cannot be further
genus and this supports the classification of clarified. The same species is also recorded as feed-
Ceterach as a genus distinct from Asplenium, with ing on Asplenium incisum (Sasakawa, 1961: 448).
which it has been synonymised by some botanical Recently Sasakawa (1988) described Phytolirio-
authorities. There must be some distinctive chemical myza diplazii from Diplazium sp. in Malaysia. The
differentiation between the two genera, inhibiting genitalia (Fig. 31) are of a form not previously found
P. scolopendri from feeding on Ceterach and simi- in species on ferns, with the aedeagus coiled, as in
larly inhibiting P. dorsata from transferring to Phlir. arctica (Fig. 978) feeding on Sonchus and
Asplenium or Scolopendrium (Aspleniaceae) or probably other composites, and also as in other
Polypodium (Polypodiaceae). species known in Australia and India, the host of
Externally Chr. dorsata and Chr. scolopendri are which is unknown. This suggests a relatively recent
morphologically similar, with the conspicuously yel- switch of ancestral species between ferns and As-
low head, including the entire hind-margin of the teraceae and is of particular interest.
eye, and only the third antennal segment black. The
mesonotum is shining black in both species and the
Davalliaceae (13 genera, 220 species)
femora are bright yellow. An essential difference
between the two is that the scutellum is yellow in This family is widespread in tropical and subtropical
Chr. sc%pendri, black in Chr. dorsata. areas mainly in the Oriental Region. Nephrolepis is
The close relationship of these two species has included here, although some authorities include it
not previously been recognised, owing to their ap- in a separate family Nephrolepidaceae.
pearing at different places in Hendel's (1935) key to A Tropicomyia reared from Nephrolepis in
Palaearctic Phytomyza species, with P. scolopendri northern India (Sehgal) surprisingly proved to
running to couplet 5, while with the scutellum black represent the widespread polyphagous species
P. dorsata continues to couplet 36. Tr. polyphaga. The genitalia (cf. Fig. 667) are readily
It now seems highly probable that Phytomyza recognisable and there seems little doubt about the
dryoptericola Sasakawa (1961) in Japan belongs to identification. This appears to represent an unusual
this same complex. The holotype was rearedjrom case of transfer to a fern from a dicot.
Dryopteris but the same species is also recorded on There is no obvious explanation for the presence
Asplenium incisum (see Dryopteridaceae below). of only four agromyzid genera on the Sphenopsida
De Meijere (1940: 170) recorded a Tropicomyia and Filicopsida - Liriomyza on the Sphenopsida
sp. (as Melanagromyza spec.) mining Asplenium (Equisetum), and Tropicomyia, Phytoliriomyza and
auriculatum on Mt. Cameroun, Cameroun. I con- Chromatomyia on the Filicopsida. The total number
14 Divisions EQUISETOPHYTA and POLYPODIOPHYTA
of species - 19 - is also small, with 3 on Equisetum species known today in New Zealand and Guade-
and 16 on the Filicopsida (see Table 3). All but one loupe. I find no other acceptable explanation for the
of these 19 species (Tr. polyphaga) are restricted to dispersal of these related species to these two widely
these two classes and show a high degree of host separated areas.
specificity (see Table 3). It could be that this assem- I am here postulating a single original coloniza-
blage of 19 species is ancient, representing primitive tion of ferns by Phytoliriomyza, a single colonization
genera which evolved when horsetails and ferns still by Tropicomyia, now known in India and West Afri-
formed a dominant part of the flora among the ca (with possibly a secondary, recent colonization
emergent angiosperms. This may be part of the ex- by Tr. polyphaga), and a single colonization by
planation. Phytoliriomyza which is the dominant Chromatomyia, with two species in Europe, one in
genus on ferns and is the exclusive genus on Hepa- India and one in Japan.
ticae is now accepted as primitive, associated with Of the 19 species on ferns, 12 are restricted to a
Metopomyza on monocots (see this section below). single genus, P. felti is known on two genera in the
With colonization of Hepaticae in Europe, Mexico, Aspleniaceae, P. dryoptericola on separate genera in
the Juan Fernandez Islands and New Zealand, and the two families Aspleniaceae and Dryopteridaceae,
of ferns in Europe, Japan, North America, the Chr. scolopendri on one genus in the Polypodiaceae
Caribbean, Argentina, the Juan Fernandez Islands, and two genera in the Aspleniaceae, while P. flavo-
Malaysia and New Zealand, this clearly attests to an pleura occurs on three genera and three separate
early origin leading to subsequent worldwide dis- families: the Aspleniaceae, Gramittidaceae and
persal. The presence of Phytoliriomyza cyatheae on Polypodiaceae. All these families are in closely relat-
two Cyathea species in New Zealand and an unde- ed orders, with only two species on separate families
scribed Phytoliriomyza on C. arborea on Guade- in the Marattiales. There are thus many orders and
loupe and a related species in Argentina on an un- families which appear to be completely avoided by
identified fern seems particularly significant. The Agromyzidae but with so little specific collecting
most logical explanation is that ancestral species having been done on ferns, the significance of these
were present in Gondwanaland, perhaps late, short- gaps cannot at this stage be properly evaluated.
ly before its breakup, evolving into the modern
3. Division PINOPHYTA (Gymnospermae)
for colonization by agromyzid leaf miners. these will prove to represent primitive species or
Gymnosperms have hitherto been totally ignored modern species which for reasons known only to
by collectors of Agromyzidae but now that species themselves have switched from angiosperms to
in two genera are known on this Division, hopefully gymnosperms as in Tropicomyia atomella remains
more species will be discovered in future. Whether to be established.
4. Division MAGNOLIOPHYTA (Angiosperms)
The angiosperms, which provide the hosts of the less effective. Under this concept the chemical arse-
majority of Agromyzidae, 'enter the fossil record nal of the Magnoliidae may consist largely of weap-
in ... the Lower Cretaceous, some 130 million years ons that have been superseded in more advanced
ago .... The combination of pollen and leaf charac- groups.'
ters is compatible with only one sub-class of modern In addition to Phytomyza, the primitive genus
angiosperms, the Magnoliidae. Phytobia whose larvae bore in the cambium of trees
' ... The earliest significant dichotomy in the evo- is well represented on the subclass, with records of
lutionary diversification of the angiosperms is that larval feeding on six families (see below) but no
between the Magnoliopsida (dicots) and Liliopsida adults are yet known. Only six polyphagous species
(monocots). The two classes are now sharply dis- are known on the subclass (Tables 4, 5).
tinct. .. .' (Cronquist, 1981)
Order MAGNOLIALES (10 families,
3000 species)
A. Class MAGNOLIOPSIDA
(Dicotyledons) Larval feeding of Phytobia has been recorded on
two families in this order (Suss and Muller-Stoll,
1980 and Suss (pers. comm.):
Subclass MAGNOLIIDAE
Magnoliaceae (12 genera, 220 species)
This subclass, as defined by Cronquist (1981) con-
sists of eight orders, 39 families and 12 000 species. Magnolia acuminata L. - USA
Agromyzid leaf miners are known on only ten of the
more primitive families but with an extraordinary
Annonaceae (130 genera, 2300 species)
proliferation on the Ranunculaceae, with a total of
116 species restricted to the family, of which 91 Hexabolus crispiflorus A. Rich. - West Africa
belong to the single genus Phytomyza (cf. Table 5). Polyalthia subtevillei var. hedinii - ? locality
This dominance of Phytomyza is approached in the Polyalthia sp. - West Africa
Apiaceae (Umbelliferae: subclass Rosidae), where Saccopetalum tomentosum Hook. f. et Thoms. -
64 Phytomyza species are known out of the total of eastern India.
87 on the family. The four agromyzid species known
on the ten families in five orders other than the
Myristicaceae (15 genera, 300 species)
Ranunculales are shown in Table 4. It is thus
obvious that this subclass is generally unattractive to This family consists mainly of trees and is wide-
and not easily colonized by Agromyzidae. This is spread in the tropics in both the Old and New
doubtless largely explained by the specific chemistry Worlds.
of the family. An unidentified Liriomyza sp. has been recorded
'The Magnoliidae have their own set of chemical on Myristica wallichii in India (Singh and Ipe, 1973:
defences. A great many of them have isoquinoline 267) but it is felt that this requires confirmation.
alkaloids and related compounds, especially benzyl-
isoquinoline and aporphine types. These are rare
Order LAURALES (8 families, 2500 species)
(though not unknown) in other groups of angio-
sperms. The various other sorts of alkaloids found in This order is unattractive to leaf miners but in con-
other angiosperms are largely or wholly wanting trast there are many records of larval feeding by
from the Magnoliidae' (Cronquist, 1981). Cronquist Phytobia.
considers that 'chemical defences of plants tend to
evolve in successive waves as old weapons become
18 Class MAGNOLIOPSIDA (Dicotyledons)
Table 4
Order MAGNOLIALES
Magnoliaceae
Magnolia Phytobia sp.
Annonaceae
Hexabolus Phytobia sp.
Polyalthia Phytobia sp.
Saccopetalum Phytobia sp.
Myristicaceae
Myristica Liriomyza sp.
Order LAURALES
Monimiaceae
Doryophora Tropicomyia polyphyta
Mollinedia Phytobia sp.
Siparuna Phytobia sp.
Calycanthaceae
Calycanthus Phytobia sp.
Lauraceae
Actinodaphne Phytobia sp.
Apollonias Phytobia sp.
Cinnamomum Phytobia sp.
Dehaasia Phytobia sp.
Iteadaphne Phytobia sp.
Laurus Phytobia sp.
Lindera Phytobia sp.
( = Parabenzoin )
Litsea Phytobia sp.
Nectandra Phytobia sp.
Neolitsea Phytobia sp.
Notaphoebe Phytobia sp.
Ocotea Phytobia sp.
Phoebe Phytobia sp.
Order PIPERALES
Piperaceae
Peperomyia Liriomyza cocculi
Piper Tropicomyia piperi, Trop. sp. ( West Africa)
Order ARISTOLOCHIALES
Aristolochiaceae
Aristolochia Liriomyza schmidti
Subclass MAGNOLIIDAE 19
Order ILLICIALES
Illiciaceae
Illicium Phytomyza sp.
Schisandraceae
Kadsura Phytobia sp.
Liriomyza cocculi Frick, a polyphagous species nated from the Magnoliales through something like
originally described from Cocculus (Menisperma- the Illiciales, which differ from the Magnoliales most
ceae: order Ranunculales), has been reared from notably in having more advanced, triaperturate pol-
Peperomyia sp. in Hawaii (Hardy and Delfinado, len' (Cronquist, 1981).
1983: 211). Four of eight families in the order serve as hosts
for Agromyzidae. While only six species (three high-
ly polyphagous) are known on the small families
Order ARISTOLOCHIALES (1 family)
Lardizabalaceae, Menispermaceae and Papavera-
ceae, in contrast, on the Ranunculaceae, there is a
Aristolochiaceae (8-10 genera, 600 species)
large assemblage of 116 species restricted to the
It is now generally accepted that this family is direct- family in seven genera (Table 5). There are in fact
ly derived from the Magnoliales. It is clearly equally more species on the Ranunculaceae than any other
unacceptable to Agromyzidae as are the families in family apart from the Asteraceae with 295. As dis-
that order. cussed below I consider that this colonization of the
Only the highly polyphagous species, Liriomyza Ranunculaceae must be primary, not secondary, and
schmidti has been recorded on Aristolochia sp. in relatives of host-specific species on Ranunculaceae
Florida. It is noteworthy that one of the most poly- are clearly detectable on a number of later families,
phagous species of all, L. trifolii, which is common particularly Aquifoliaceae, Apiaceae, Crassulaceae
in Florida, has not colonized this family. (Rosidae) and Lamiaceae (Asteridae).
Table 5
Order ILLICIALES (2 families, 90 species)
Agromyzid genera on Ranunculaceae
It is generally agreed that the Illiciales are related to
and derived from the Magnoliales. Family-specific Polyphagous
Melanagromyza
IIliciaceae (1 genus, 40 species) Ophiomyia 5 (1 undescribed)
Liriomyza 2 3
The single agromyzid known on this family, Phyto- Napomyza 14
myza jucunda Frost and Sasakawa (ct. Sasakawa, Chromatomyia 2
1961), normally feeds on !lex (Aquifoliaceae: sub- Phytomyza 91
class Rosidae), but has been recorded on Illicium Ptochomyza
religiosum in Japan. If this identification is correct, it
Total 116 3
would represent a most unusual switch but the illus-
tration of the mine on Illicium given (ct. Sasakawa, All species on the order are shown in Table 6
1961: fig. 130,0 2) differs substantially from that of following the botanical arrangement of their hosts.
typical mines on flex and it seems probable that a
separate species is involved.
Ranunculaceae (50 genera, 2000 species)
It is agreed among botanists that this family is close-
Schisandraceae (2 genera, 50 species)
ly related to and probably derived from the Mag-
This is an ancient family known from the Upper noliales and in the opinion of Cronquist 'there is no
Cretaceous. It supports no leaf miners but Suss (loc. obvious reason why the Ranunculaceae might not be
cit.) records larval feeding of Phytobia on Kadsura ancestral to all the other families of the order.' Al-
grandifolia Wall. - India to Japan and Greater though the earliest pollen record is Lower Miocene
Sunda Islands. (Muller, 1981: 15), Blanc-Louvel (1984) has dis-
cussed in detail and illustrated four fossils from the
Berriasian, Early Cretaceous from Province Lerida,
Order NYMPHAEALES (5 families, 65 species)
Spain of a species ascribed to Ranunculus and ap-
This primitive, aquatic order, which includes such pears close to R. aquatilis. However, more detailed
well-known plants as Nuphar (Nymphaeaceae), the study by other workers refutes the affiliations of
yellow water-lily in Europe, appears to be entirely these fossils to the Ranunculaceae and suggest that
repellent to Agromyzidae. they merely represent an unidentifiable non-angio-
spermous seed-plant.
The classification of the Ranunculaceae cited by
Order RANUNCULALES (8 families,
Willis (1985) is based on an early study by Hutchin-
3200 species)
son (1923). It quickly became apparent to me that
'It is widely believed that the Ranunculales origi- this in some respects could not be a natural system,
Subclass MAGNOLIIDAE 21
Table 6
Order RANUNCULALES
Ranunculaceae
Subfamily Helleboroideae
Tribe Trollieae
Trollius Phytomyza subrostrata, P. trollii, P. trolliophila, P. trolliovora
Caltha P. calthae, P. calthivora, P. calthophila, P. soenderupi
Tribe Cimicifugeae
Actaea Melanagromyza actaeae; P. actaeae, P. kaltenbachi (= P. philactaeae)
Cimicifuga P. actaeae
Tribe Helleboreae
Helleborus P. hellebori
Nigella Phytomyza sp.
Tribe Delphineae
Aconitum P. aconitella, P. aconiti, P. aconitophila
Delphinium Ophiomyia delphinii; P. aconiti, P. delphinivora
Subfamily Ranunculoideae
Tribe Anemoneae
Subtribe Anemoninae
Anemone P. abdominalis. P. albimargo, P. anemonantheae, P. anemones,
P. anemonivora, P. buhri, P. f1avofemoralis, P. hendeli, P. multifidae,
P. narcissif1orae, P. nigricoxa. P. nigripennis, P. pulsatillae, P. pulsatillicola,
P. socia, P. soenderupiella, P. yasumatsui
Knowltonia P. knowltoniae
Subtribe Clematidinae
Clematis Napomyza ciematidicolla, N. clematidis; Chromatomyia clemativora,
Chr. cIematoides; P. actaeae, P. buhri, P. cirrhosae, P. cIematadi,
P. cIematidella, P. cIematidicaulis, P. clematidiphoeta, P. clematiphaga,
P. ciematisana, P. cIematisella, P. cIematisi, P. clematoides, P. dalmatiensis,
P. esakii, P. fulgens, P. kaltenbachi, P. ligusticifoliae, P. loewii,
P. novitzkyi, P. orientalis, P. pampeana, P. paniculatae, P. philoclematidis,
P. rectae, P. varii, P. vitalbae, P. \'italbella, P. williamsoni. P. yasumatsui,
P. anemones ssp. cIematobia, Phytomyza sp. n. ( Montana);
Ptochomyza mayeri
Tribe Ranunculeae
Subtribe Adonidinae
Adonis Liriomyza xanthocera; Chr. horticola
Callianthemum p, callianthemi
Subtribe Ranunculinae
Ranunculus O. ranunculicaulis; L. ranunculoides, L. trifolii, L. xanthocera;
Napomyza albipennis, N. costata, N. evanescens, N. eximia, N. improvisa,
N. lyalli, N. marginalis, N. nigritula, N. ranunculella, N. ranunculicaulis,
N. renovata, N. subeximia; P. aquilegiana, P. aquilonia, P. buhriana,
P. caulinaris, P. cortusifolii, P. davisii, P. fallaciosa, P. f1avofemoralis,
22 Class MAGNOLIOPSIDA (Dicotyledons)
Subfamily Isopyroideae
Tribe Isopyreae
Subtribe Aquilegiinae
Aquilegia O. aquilegiana; P. aquilegiae, P. aquilegiana, P. aquilegioides,
P. aquilegiophaga, P. aquilegivora, P. columbinae, P. krygeri, P. minuscula,
P. sonorensis
Tribe Coptideae
Coptis P. tamui
Subfamily Thalictroideae
Tribe Thalictreae
Thalictrum O. aquilegiana, O. thalictricaulis, Ophiomyia sp.n. (Argentina);
L. cordobensis nom. nud, L. thalictri; P. albifrons, P. aquilegiae,
P. aquilegioides, P. columbinae, P. minuscuia, P. piumiseta, P. thaiictreIla,
P.-thalictri, P. thaiictricoia, P. thalictrivora
Lardizabalaceae
Akebia P. akebiae
Menispermaceae
Cocculus L. cocculi
Stephania Tropicomyia polyphyta
with Aqui/egia and Thalictrum placed far apart in These two specialists thus produce somewhat dif-
separate subfamilies, although five agromyzids (Phy- ferent systems. The agromyzid host selection sup-
tomyza aquilegiae, P. minuscula, P. aquilegioides, ports Tamura who has Caltha and Trollius in the
P. columbinae; and Ophiomyia aquilegiana) feed on same tribe Trollieae of the Helleboroideae but he
these two genera and on no others. Jensen (1968: does separate Aquilegia and Thalictrum in separate,
292), on the basis of serological tests, was able to although adjoining, subfamilies. On the other hand
postulate the close affinity of Aqui/egia and Thalic- Jensen has Caltha and Trollius in separate subfami-
trum and formally erected the tribe Thalictreae to lies but Aquilegia and Thalictrum are in the same
embrace these two genera, although placing them in tribe Thalictreae, although in different subtribes.
separate subtribes. Both agree in including Anemone, Hepatica and
Tamura (1966,1967,1968), in detailed morpho- Pulsatilla in the same subtribe of the Anemoneae.
logical studies, divided the family into six subfami- Here, I basically follow Tamura.
lies (now five), further subdivided into ten tribes The remarkable dominance of the genus Phyto-
with eight subtribes. Jensen (1968) only accepted myza on the Ranunculaceae is apparent from Table
two subfamilies, the Hydrastoideae and Ranunculoi- 5 above. The Lauraceae and Piperaceae each with a
deae, the latter divided into six tribes which are virtually identical number of species (2000) support
further subdivided into 13 subtribes. no species in Phytomyza, the essential difference
Subclass MAGNOLIIDAE 23
between the families being the largely tropical or Treatment of the NAPOMYZA,
subtropical nature of the former, while the Ranun- CHROMATOMYIA, PHYTOMYZA, and
culaceae are essentially temperate, occurring in the PTOCHOMYZA species in the botanical
tropics only at high altitudes, thus in areas where sequence proposed by Tamura
Phytomyza can flourish. (1966, 1967, 1968)
I discuss below first the eight family-specific spe-
cies in Melanagromyza, Ophiomyia and Liriomyza, Subfamily Helleboroideae
but with such a large number of Phytomyza species
involved, it seems desirable to discuss these sequen- Tribe Trollieae
tially under the subfamilies, tribes and genera of the Trollius. Four Phytomyza species are monophagous
family to illustrate more clearly the monophagy and on this genus, three leaf miners and one seed-feeder.
occasional oligophagy of the species concerned. It The two leaf miners, P. trolliophila Hering (Figs. 34,
will be seen that while most species closely follow 35) and P. trolliovora Hering (Figs. 36, 37) are more
the classification adopted by botanists, others cross closely related to P. subrostrata Frey (Figs. 38, 39),
subfamily (Aquilegia and Thalictrum) and tribal which feeds in the flower-head, than to the third leaf
boundaries (Actaea/Clematis, Ranunculus/Clematis miner, P. trollii Hering (Figs. 40, 41). All these spe-
and Anemone/Clematis). In view of the close rela- cies are discussed by Spencer (1976a). I have also
tionship of Napomyza, Chromatomyia and Ptocho- seen a photograph of a linear mine on Trollius yun-
myza to Phytomyza, the 108 species in these four nanensis at 13,000 ft. in Tibet (preserved in the
genera are discussed together below. New York Botanical Garden Herbarium) taken and
The only Melanagromyza known on the family, kindly sent to me by Dr. Olle Pellmyr, New York. It
M. actaeae Sehgal (1971b) feeds in stems of Actaea would be of great interest to know whether this
in Alberta, Canada but unfortunately only females miner represents one or other of those known in
are known and the affiliations of the species cannot Europe or is yet another species.
be established. Caltha. Four host-specific species are known on
Of the four Ophiomyia species in Europe, three Caltha, all in Europe, three being leaf miners, one
0. aquilegiana (Lundquist (= thalictrina Griffiths)), an internal feeder. There are conspicuous differen-
0. ranunculicaulis Hering and 0. thalictricaulis ces in their leaf mines, P. calthae Hering (cf. Hendel,
Hering (cf. Spencer, 1964c for discussion of all 1935) forming a large blotch, calthophila Hering (cf.
three) - form external stem-mines and there is Spencer, 1976a) a long, narrow mine (Fig. 45) and
nothing in the genitalia to suggest that they have not calthivora Hering (cf. Spencer, 1976a) a shorter,
evolved from a single ancestral colonization. How- broader mine (Fig. 47). The genitalia show that all
ever, the larval posterior spiracles in O. ranunculi- three are closely related: P. calthae (Fig. 42),
caulis (Fig. 32) are unique, with numerous spiracu- calthophila (Figs. 43, 44) and calthivora (Fig. 46).
lar pores on three arms and Hering in his descrip- Particularly in P. calthivora the posterior spiracles
tion (1949c: 22) remarks: 'It is surprising that with a of the larva (and puparium) are greatly enlarged
larva of such striking form there are not greater (Fig. 48), as found in other species in an aquatic en-
differences in the imago.' vironment (cf. Liriomyza equiseti and P. cicutae: Fig.
0. delphinii Hendel is a strictly monophagous 620). P. soenderupi Hering (cf. Spencer, 1976a)
leaf miner on Delphinium staphysagria, known only feeds and pupates internally in the petiole and the
in Dalmatia, Yugoslavia. I found it to be abundant genitalia (Figs. 49, 50) confirm relationship with the
on Zlarin Is. in the Adriatic in June, 1962. The host three leaf miners.
is widespread in the Mediterranean area and ex-
tends into Asia Minor, and 0. delphinii may well Tribe Cimicifugeae
prove to be more widespread. Actaea. P. actaeae Hendel (cf. Hendel, 1935) forms
Only females are known of the two described a secondary blotch mine, mainly on A. spicata. The
Liriomyza species, L. thalictri Hering (1932) from genitalia (Fig. 51) indicate a relationship with
Germany and L. ranunculoides Spencer (1969b) P. aconitophila Hendel (see below). Beiger (1980b)
from Ontario, Canada. The distinctive mesonotum discovered P. actaeae also feeding on Clematis al-
of the latter is shown in Fig. 33. A further species pina in Poland.
has been reared from Thalictrum in Argentina (G. Hering (1932) described P. philactaeae on
Valladares) but has not yet been described. Of the Actaea in Switzerland as a var. of P. kaltenbachi
polyphagous Liriomyza species, L. xanthocera is Hendel which feeds on Clematis. Only females are
known on Adonis and L. trifolii on Ranunculus. known of philactaeae but this form is now accepted
as synonymous with P. kaltenbachi. The discovery
of a species feeding on both Actaea and Clematis
24 Class MAGNOLIOPSIDA (Dicotyledons)
34
32
<':~~
, i
36 35
,:'
38
3..77 I:'i.:
/:~.,;
pJ
40
Fig. 32. Ophiomyia ranunculicaulis: posterior spiracles of larva (Hering).
Fig. 33. Liriomyza ranunculoides: mesonotum.
Figs. 34, 35. Phytomyza trolliophila: 34, aedeagus, side view; 35, same, ventral view.
Figs. 36, 37. Phytomyza trolliovora: 36, aedeagus, side view; 37, same, ventral view.
Figs. 38, 39. Phytomyza subrostrata: 38, aedeagus, side view; 39, same, ventral view.
Figs. 40, 41. Phytomyza trollii: 40, aedeagus, side view; 41, same, ventral view.
Subclass MAGNOLIIDAE 25
43
44
45
49
48
50
---.:.::.----
51
53
55 56
)
.,
t
.-.'.-
58
which are in separate subfamilies is unexpected. ly as in P. fallaciosa (Figs. 166, 167) on Ranunculus
Cimicifuga. The only species recorded here is and adults of the two species cannot be reliably dis-
P. actaeae. tinguished. However, with P. fallaciosa being multi-
voltine, this represents a consistent biological dis-
Tribe Helleboreae tinction from anemones. Although there appear to
Helleborus. P. hellebori Kaltenbach (cf. Spencer, be no detectable differences in the genitalia between
1976a) forms linear-blotch mines on many species P. anemones and its ssp. buhri, in the latter the
of this genus. The genitalia (Fig. 52) indicate rela- femora are wholly yellow, not largely black and
tionship with species on both Anemone (P. anem- narrowly yellow only at the end (tip), the host is
ones) and Ranunculus (P. fallaciosa). P. hellebori restricted to Anemone hortensis and there are at
subsp. buhri Hering (in Buhr, 1930) has now been least two generations, not one as in P. anemones. On
upgraded to a full species on Anemone. this basis ssp. buhri was raised to specific status by
Nigella. Mines considered to represent an unidenti- Spencer and Martinez (1987: 256). A further sub-
fied Phytomyza sp. have been found on this genus in species of P. anemones on Clematis, P. (a.) clemato-
Germany (Hering, 1957b). bia, is discussed below.
P. abdominalis Zett. (cf. Spencer, 1976a) is a
Tribe Delphinieae largely black species but with the abdomen strik-
Aconitum, Delphinium. Of the four leaf miners on ingly yellow in both sexes. It forms blotch mines and
this tribe, two, P. aconitella Hendel (1934) (Figs. its primary host is A. hepatica. The genitalia (Figs.
53,54) and P. aconitophila Hendel (1927) (Fig. 55) 64, 65) indicate close relationship with P. socia
in Europe, are restricted to Aconitum, P. delphini- Brischke (Figs. 66, 67) which forms similar mines
vora Spencer (Figs. 56, 57) in Alberta and Alaska on the same host but in this species the abdomen is
(Spencer, 1969b) is restricted to Delphinium, while black (cf. Spencer, 1976a). P. campanariae Nowa-
P. aconiti Hendel (Figs. 58, 59) occurs on both kowski (1958) is now accepted as synonymous with
Aconitum and Delphinium and is holarctic, being P. socia (see Appendix below). Two somewhat
common in Europe and is also present in the eastern similar dark species occur on A. nemorosa. P. albi-
U.S. and in Quebec (cf. Spencer, 1976a). margo Hering (Spencer, 1976a) forms blotch mines
P. aconiti and P. aconitophila are related, and (Fig. 68) and can occur together with P. hendeli
also related to P. actaeae (Fig. 51), but P. delphini- Hering (Spencer, 1976a) which forms linear mines
vora is an isolated species representing a separate (Fig. 71). The genitalia of P. albimargo are un-
colonization of Delphinium. known. In P. hendeli the genitalia (Figs. 69, 70)
show that this belongs to a separate group and is
closely related to and even doubtfully distinct from
Subfamily Ranunculoideae P. pulsatillae Hering (1924b) and P. rectae on
Clematis (see below). Adults, genitalia (Figs. 72, 73)
Tribe Anemoneae and leaf mines of P. pulsatillae are identical to
P. rectae but there are slight larval differences, with
Subtribe Anemoninae
16-18 pores on the posterior spiracles of P. pulsa-
Seventeen species are known on this subtribe in tillae, while consistently only 12 in P. rectae (He-
Europe, one in Japan, and 2 in Canada, all on ring, 1968: 321). A further species, P. pulsatillicola
Anemone (including Pulsatilla), and one on Knowl- Hering (1962), was described from blotch mines on
tonia in South Africa. Pulsatilla pratensis (= Anemone pratensis) from the
Anemone. Species on this genus all show generic Leutra Tal, Jena, East Germany which Hering con-
monophagy apart from two species in Japan - trasted with P. campanariae Now. but citing the
P. yasumatsui recorded on Clematis and P. flavo- same larval characters as for P. pulsatillae. The male
femoralis on Ranunculus. genitalia of the holotype (Figs. 74, 75) confirm that
P. nigripennis Fallen (Spencer, 1976a) is proba- this is a good species. Pulsatilla is now treated as a
bly the most primitive species on Anemone. It is synonym of Anemone (Mabberley, 1987).
exceptionally large, with wing length up to 3.5 mm, In P. multifidae Sehgal (1971b) which is known
and feeds internally in the stem of A. nemorosa on Anemone multifida in Alberta the genitalia (Fig.
(long suspected but only recently reared by M. von 76) are little differentiated from those of the other
Tschirnhaus). The genitalia (Figs. 60, 61) show a species in this group. This either represents a holarc-
clear relationship with the species on Caltha and tic species or if it is treated as distinct, it can be ac-
Trollius. cepted as the sister-species of one of the group
P. anemones Hering (Spencer, 1976a) is known known in Europe (P. hendeli, rectae, pulsatillae). As
only on A. nemorosa, with a single early spring mentioned under Ranunculus, P. ranunculivora be-
generation. The genitalia (Figs. 62, 63) are essential- longs to this same group.
28 Class MAGNOLIOPSIDA (Dicotyledons)
60
: 64
61
,.-,--
~
66 ...
. - - - ...
'
6S
67
Subclass MAGNOLIIDAE 29
P. anemonantheae Spencer (1969a) is only (Figs. 84, 85) are of the characteristic form pre-
known on A. silvestris in Germany and the genitalia viously noted in species in Africa, Australia and
(Figs. 77, 78) show that this is an isolated species. New Zealand on Ranunculus. N. ranunculella
P. anemonivora Spencer (1969b) which has once Spencer (1974), known only in Israel is doubtfully
been reared from 'head of Anemone sp.' at Ottawa, distinct from N. clematidis but with no other species
Canada is a striking species with the abdomen bright known to feed on both Clematis and Ranunculus, I
yellow as in P. abdominalis but the genitalia (Fig. propose to treat the two as distinct until more de-
79) show that the two species are not closely related. tailed revisionary studies can be undertaken. In
In Europe P. nigricoxa Hendel (cf. Spencer, 1976a) Australia the leaf miner clematidicolla (cf. Spencer,
also feeds in seed-heads of A. pratensis and the 1977 c) was described in Phytomyza but the genitalia
genitalia (Fig. 80) suggest that this may be related to (Figs. 86, 87) clearly associate it with N. clematidis
P. anemonivora. P. soenderupiella Spencer (1976a) and it is now transferred to Napomyza (cf. Appen-
is an entirely black species (which was misidentified dix below). Significantly in this species pupation
earlier as nigritella Zett.) and feeds on the same host takes place in the mine. It seems clear that N. clema-
in Europe in the upper stem and receptacle (Fig. tidicolla is derived from a stem-feeding species and
81). after developing into a leaf miner, it has retained the
P. narcissiflorae Hering (cf. Hendel, 1935) was habit of pupating internally.
described from a single female reared from A. nar- Two species of Chromatomyia are known on Cle-
cissiflora in Switzerland. It generally resembles matis. Chr. clemativora (Frost, cf. Spencer and
P. abdominalis but in the absence of males its status Steyskal, 1986) was described from Texas (genitalia:
cannot be further clarified. Fig. 88) and Chr. clematoides (Spencer and Steys-
Two species are known in Japan on Anemone. kal, 1986) was discovered in Washington State,
P. yasamatsui (cf. Sasakawa, 1961) has been record- U.S.A. mining C. ligusticifolia (genitalia: Figs. 89,
ed on Anemone but its primary hosts are Clematis 90). In both species the genitalia are typical of the
spp. (see below). P. flavofemoralis Sasakawa (cf. genus (see also Appendix below).
1961) feeds primarily on Anemone but was later re- With 30 Phytomyza species known on Clematis,
corded also on Ranunculus. The genitalia (Fig. 82) this is the largest assemblage in a single genus
show that this species belongs to the anemones- known on any host genus in the Ranunculaceae.
group but its exact relationship with anemones and Two Phytomyza species are known as stem
fallaciosa on Ranunculus requires more detailed in- feeders on Clematis recta in Austria, both pupating
vestigation. in the stem. Both are largely black, with strongly
Knowltonia. P. knowltoniae Hering (1957a) is projecting orbits. The genitalia of P. clematidicaulis
known only in South Africa on K. capensis. The Hering (Fig. 91) and of P. novitzkyi Hering (Fig. 92)
genitalia (Fig. 83) associate the species closely with are strongly differentiated and the two species may
further South African species on Ranunculus, P. na- not be closely related (Hering, 1958).
talensis and P. ranunculina (see below). Together with P. clematidicaulis and P. novitzkyi,
Novitzky discovered a third species in stems of Cle-
Subtribe Clematidinae matis recta near Vienna which has most recently
Only the single genus Clematis is included here. It been treated as Phytomyza mayeri (Spencer)
has a virtually worldwide distribution and Napomy- (L. Suss, 1977-78) but which I have now dis-
za and Phytomyza species have succeeded in fol- covered properly belongs to the small genus Ptocho-
lowing it from Europe to North America and Japan, myza (see Appendix below and Fig. 93). The only
to Central and South America as far south as other host known of Ptochomyza species is Aspara-
Argentina, to East and South Africa, Australia and gus (Liliaceae). A logical explanation for a host
New Zealand. With 35 species and one subspecies transfer between Clematis and Asparagus is difficult
recorded on Clematis, this is more than on any other to conceive. I consider there must have been a pure-
genus in the family. ly fortuitous switch, probably as a result of a mistake
Napomyza clematidis (Kaltenbach) (= Phytomyza in oviposition by some ancestral female. I have col-
mallorcensis Spencer, cf. Tschirnhaus, 1981: 32) lected Pt. asparagivora in Kenya on Asparagus
feeds in flower-heads and stems and the genitalia africana. This is a climber and was growing in such
Figs. 60, 61. Phytomyza nigripennis: 60, aedeagus, side view; 61, same, ventral view.
Figs. 62, 63. Phytomyza anemones: 62, aedeagus, side view; 63, distiphallus, ventral view.
Figs. 64, 65. Phytomyza abdominalis: 64, aedeagus, side view; 65, same, ventral view.
Figs. 66, 67. Phytomyza socia: 66, aedeagus, side view; 67, same, ventral view.
Fig. 68. Phytomyza albimargo: leaf mine on Anemone nemorosa (Hering).
Figs. 69-71. Phytomyza hendeli: 69, aedeagus, side view; 70, same, ventral view; 71, leaf mine on Anemone nemorosa (Hering).
30 Class MAGNOLIOPSIDA (Dicotyledons)
72
74
76 77
Subclass MAGNOLIIDAE 31
~
?<.J ~ .
..
82
84
86
85
87 89
Figs. 72, 73. Phytomyza pulsatillae: 72, aedeagus, side view; 73, same, ventral view.
Figs. 74, 75. Phytomyza pulsatillicola: 74, aedeagus, side view; 75, same, ventral view.
Fig. 76. Phytomyza multifidae: aedeagus (Sehgal).
Figs. 77, 78. Phytomyza anemonantheae: 77, aedeagus . side view; 78, same, ventral view.
Fig. 79. Phytomyza anemonivora: aedeagus.
Fig. 80. Phytomyza nigricoxa: aedeagus.
Fig. 81. Phytomyza soenderupiella: aedeagus.
Fig. 82. Phytomyza fiavofemoralis: aedeagus (Sasakawa).
Fig. 83. Phytomyza knowltoniae: aedeagus.
Figs. 84, 85. Napomyza clematidis: 84, aedeagus, side view; 85, same, ventral view.
Figs. 86, 87. Napomyza clematidicolla: 86, aedeagus, side view; 87, same, ventral view.
Fig. 88. Chromatomyia clemativora: aedeagus.
Figs. 89, 90. Chromatomyia clematoides: 89, aedeagus, side view; 90, same, ventral view.
32 Class MAGNOLIOPSIDA (Dicotyledons)
92
;,:/
91
93
95
94
96
97
Fig. 91. Phytomyza clematidicaulis: aedeagus.
Fig. 92. Phytomyza novitzkyi: aedeagus.
Fig. 93. Ptochomyza mayeri: aedeagus (L. Suss).
Fig. 94. Phytomyza vitalbae: aedeagus.
Fig. 95. Phytomyza dalmatiensis: aedeagus.
Figs. 96, 97. Phytomyza kaltenbachi: 96, aedeagus, side view; 97, distiphallus, ventral view.
Figs. 98, 99. Phytomyza cirrhosae: 98, aedeagus, side view; 99, same, ventral view.
Subclass MAGNOLIIDAE 33
close association with numerous other plants that it detailed studies clematobia is retained as a valid
was not easy to be certain to which plant a particular subspecies, while P. buhri has been raised to full
stem belonged. Under such conditions a switch from specific status (see above).
Clematis to Asparagus could have occurred inad- P. fulgens (d. Hendel, 1935) in Europe has
vertently, the non-host Asparagus was found accept- entirely different genitalia (Figs. 100, 101) which
able and resulted in the evolution of two new species associate it closely with P. loewii Hendel (Figs. 102,
on Asparagus. 103) (= clematidis Loew) (cf. Spencer and Steyskal,
Five leaf miners and one subspecies have been 1986) which is widespread in North America and
described on Clematis in Europe, belonging to three has extended its range south to Costa Rica (where it
different groups. P. vitalbae Kalt. (d. Spencer, was described as P. centralis Frost), Cuba and
1976a) and Phytomyza dalmatiensis (Spencer, Colombia.
1961f) (described in Napomyza, with outer cross- Six further species have been described in North
vein present) both have the aedeagus coiled (vitalbae: America. P. clematisella Spencer (Spencer and
Fig. 94; dalmatiensis: Fig. 95) and can be associated Steyskal, 1986) (Figs. 104, 105) was reared from
with P. ranunculi (Fig. 156). The former pupates stems of Clematis sp. in Arkansas, U.S.A. P. clema-
externally, the latter, which feeds basically in the tidiphoeta Spencer (1969b) (Figs. 106, 107) was
midrib and veins, pupates internally, a relic of ances- caught on 'wild Clematis' in Ontario, Canada and
tral internal feeding. In Western Australia, south of the genitalia of both these species show a clear as-
Perth I found leaf mines of P. vitalbae occurring sociation with P. clematidicaulis in Europe. I caught
together, even on the same plant (c. pubescens), one male (Fig. 108) and eight females of P. clema-
with N. clematidicolla (see above) but the mines are tisana Spencer (1981) individually on foliage of
readily distinguishable, with N. clematidicolla pu- C. ligusticifolia in southern California and I am
pating in the mine, P. vitalbae externally. In P. kal- satisfied that this is an internal feeder in either the
tenbachi Hendel (cf. Spencer, 1976a) the genitalia stem or seeds.
(Figs. 96, 97) show that this is related to species on P. clematiphaga Spencer (1969b) is known in
Caltha and Trollius (see above). Pupation in this Alberta, Canada and Montana, U.S.A. (Figs. 109,
species takes place in the mine. P. kaltenbachi is 110). A further undescribed species was noted in
unusual in feeding not only on Clematis but also on the series from Montana reared from C. colum-
Actaea in the Helleboroideae. P. actaeae also feeds biana (Fig. 111 ) but this cannot be further clarified
on these 2 hosts, having been recently recorded on at the present time. No males are known of P. ligus-
Clematis in Poland (Beiger, 1980b). P. cirrhosae ticifoliae Spencer (1981) which forms distinctive
Spencer (1969a) is only known on Clematis cirrhosa blotch mines (Fig. 112) on C. ligusticifolia in south-
in Morocco, the larva forming a whitish blotch mine ern California but with the differing form of mine
and the genitalia, with the distiphallus strongly re- from the other species on Clematis which form
curved (Figs. 98, 99) suggest that this may be related linear mines, it is probable that this belongs to a
to the hendeli group on Anemone. further group.
P. rectae Hendel (cf. Hendel, 1935; Hering, Three host-specific species are known on Clem-
1967) feeds on Clematis recta and C. flammula, and atis in Japan, all forming linear mines. P. yasumatsui
the genitalia cannot be distinguished from those of (cf. Sasakawa, 1961), described in Napomyza owing
P. hendeli and P. pulsatillae on Anemone, although to the presence of the outer cross-vein, is a typical
there are small larval differences (see above under Phytomyza (Fig. 113). P. esakii (Sasakawa, 1961)
Anemone). (Fig. 114) and P. paniculatae (cf. Sasakawa, 1961)
Hering (1967) recorded a species from Clematis (Fig. 115) have genitalia suggesting they are closely
flammula on Hvar, Yugoslay.ia which he originally related.
intended to describe as P. clematobia. Examination Of the five species known on Clematis in East
of the genitalia (Hering, 1967: fig. 30) showed these and/ or South Africa, one, P. clematidella Spencer
to be identical to those of P. anemones (Figs. 62, (1959) known only in Kenya, is a gall-causer (or
63). The femora and mesopleura are more exten- possibly a stem feeder). Unfortunately only a single
sively yellow, similar to P. buhri on Anemone hor- reared female is known but the distinctive anterior
tensis, which has similar genitalia but in P. clema- larval spiracles are shown in Fig. 116. The four
tobia emergence time was later, from the last ten other species are leaf miners all belonging to a group
days of May to the first ten days of June, while in with the aedeagus strongly recurved. P. clematisi
buhri the adults emerged during the last ten days of Spencer (1964a) (Figs. 117, 118) is a blotch miner,
April until the first ten days of May. In view of these known in Ethiopia and Kenya, while the other three
differences Hering decided to treat clematobia as a form linear mines. P. philoclematidis Hering
subspecies of anemones, differentiated from buhri (1957a) was described from South Africa and mines
solely on this biological difference. Pending more I found near Nairobi, Kenya almost certainly repre-
34 Class MAGNOLIOPSIDA (Dicotyledons)
100 .'
102
":<:"'\:i>'3:\\ :$. ,
.,'
101 ":/.~;~.<»~./
... - .~ . 103
104 106
\
105 107
108
Subclass MAGNOLIIDAE 35
111
110
~, T---
--- --- --- .'\. ..
----~J
-' ~
.'
116
118
Figs. 109, 110. Phytomyza clematiphaga: 109, aedeagus, side view; 110, distiphallus, ventral view.
Fig. Ill. Phytomyza sp.n. near clematiphaga: distiphallus, ventral view.
Fig. 112. Phytomyza ligusticifoliae: leaf mine on Clematis ligusticifolia.
Fig. 113. Phytomyza yasumatsui: aedeagus, side view (Sasakawa).
Fig. 114. Phytomyza esakii: aedeagus, side view (Sasakawa).
Fig. 115. Phytomyza paniculatae: aedeagus, side view (Sasakawa).
Fig. 116. Phytomyza clematidella: anterior spiracles of puparium.
Figs. 117, 118. Phytomyza clematisi: 117, aedeagus, side view; 118, same, ventral view.
Subclass MAGNOLIIDAE 37
\\
~
.......
119
121
122
.'
...........,......
............
120
123
124 125
Fig. 126. Balloon travel from Christchurch to South America at an altitude of 12000 m.
Several of the Napomyza species discussed below the stem and flower-heads of Ranunculus in Israel,
were described in Phytomyza before their correct the genitalia (Figs. 131, 132) have the distiphallus
generic position was appreciated. fully divided and the species is closely related to
N. evanescens Hendel (cf. Spencer, 1976a) (Fig. N. clematidis (see above). Ancestral relatives of
128) feeds internally in the stem of Ranunculus spp. N. ranunculella have dispersed to the southern
This is a holarctic species, being common in Europe hemisphere and are present in East and South Afri-
and is present in western Canada and California. ca, Australia and New Zealand, all feeding either
N. albipennis (Fallen, cf. Spencer, 1976a), a distinc- internally or as leaf miners on Ranunculus. N. reno-
tive species with silvery-white wings, is known only vata (Spencer, 1960a) was described from South
in Europe and, although not yet reared, the genitalia Africa and later many specimens were caught in
(Fig. 127) clearly associate it with N. evanescens. A Kenya on R. multifidus. The genitalia (Figs. 133,
second holarctic species feeding in both stems and 134) indicate close relationship with N. ranunculel-
receptacle, N. nigritula (Zetterstedt) (= cineracea lao N. subeximia Spencer (1985a) (Figs. 135, 136)
Hendel, cf. Spencer, 1976a), has somewhat modi- was caught on the same host also in Kenya, and is
fied genitalia (Fig. 129). N. marginalis (Frost) (d. accepted as an internal feeder. A further species in
Spencer and Steyskal, 1986) is close to N. nigritula South Africa, N. eximia (Spencer, 1964a) was de-
but the small genitalia differences (Fig. 130) suggest scribed in Phytomyza but is now accepted in Napo-
that the species are distinct. myza and (d. Appendix below) has genitalia (Figs.
In N. ranunculella Spencer (1974) which feeds in 137, 138) indicating that it is closely related to
127 128
129 130
~,
131 132
133
).- 136
135 .~
40 Class MAGNOLIOPSIDA (Dicotyledons)
137
139
. ;dI
141 143
144
142
.~
~..
146
Figs. 137, 138. Napomyza eximia: 137, aedeagus, side view; 138, distiphallus, ventral view.
Figs. 139, 140. Napomyza ranunculicaulis: 139, aedeagus, side view; 140, same ventral view.
Figs. 141-143. Napomyza lyalli: 141, aedeagus, side view; 142, same, ventral view; 143, leaf mine on Ranunculus lyallii.
Figs. 144, 145. Napomyza coslata: 144, aedeagus, side view; 145, same, ventral view.
Figs. 146, 147. Napomyza improvisa: 146, aedeagus, side view; 147, same, ventral view.
Subclass MAGNOLIIDAE 41
N. renovata and N. subeximia, and it can be as- nunculi (Schrank, ct. Spencer 1976a) (Fig. 156)
sumed with a high degree of probability that the host which is present in Europe, California and Japan, to
is Ranunculus. In Australia south of Perth I found the extreme form in the boreal-alpine holarctic
N. ranunculicaulis in stems of R. colonorum. I de- species P. aquilonia Frey (d. Spencer, 1976a)
scribed the species in Phytomyza (Spencer, 1977c) (= vibeana Griffiths, 1966) (Fig. 157) which was de-
but both the genitalia (Figs. 139, 140) and the slen- scribed from Finland, is known in the Tatra Moun-
der brown puparia confirm that the species can be tains in Poland, and in Greenland and Alaska. Two
associated with the other Napomyza species in this species on Clematis also belong to this group,
group. Two further species in Australia, P. placita P. dalmatiensis (Spencer) and P. orientalis Spencer
Spencer (1977c: figs. 364, 365) and P. pulchella (see under Clematis above).
Spencer (1977c: figs. 368, 369) are virtually cer- At Abisko in northern Sweden I caught a new
tainly feeders on Ranunculus but are not formally species, which I was proposing to describe, but von
associated with this host here. Tschirnhaus (pers. comm., 1976) informed me that
In New Zealand two species reared from and one he had reared a long series in the mountains in
caught on Ranunculus were described in Phytomyza southern Norway from Ranunculus glacialis. He
but are now accepted in Napomyza. N. lyalli considers that this host should be referred to the
Spencer (1976b) (Figs. 141, 142) feeds primarily in genus Beckwithia, rather than Ranunculus, and re-
the stem of the large plant R. lyallii, which can attain quested that he should be allowed to describe it -
a height of 150 cm, and in one plant I counted 50 which he proposed to do in the near future. I there-
puparia. However, the larva can also move into fore provided an informal description (Spencer,
leaves and a true leaf mine (Fig. 143) can be formed 1976a: 390) as 'sp. (Beckwithia).' Von Tschirnhaus
(see Appendix below). N. costata (Harrison, ct. has unfortunately still not described the species 13
Spencer, 1976b), as far as is known, is only a leaf years later and it thus remains without a name. The
miner, the puparium remaining in the leaf. The geni- genus Beckwithia was described in California for the
talia (Figs. 144, 145) clearly associate this species single species B. austinae. Later, B. austinae was
with the other Napomyza species in this group (see synonymised with Ranunculus andersonii and
Appendix below). I caught a third species, with Beckwithia was thus sunk in Ranunculus. Tamura
many specimens on Ranunculus, N. improvisa (pers. comm., 16.i.88) wrote: 'About Beckwithia, I
(Spencer, 1976b), and the genitalia (Figs. 146, 147) don't think it is necessary to accept the independent
clearly indicate that this species also belongs in genus for Ranunculus glacialis L. I want to follow
Napomyza. Benson (1948) who treated R. glacialis L. as the
I consider that the habit of pupating within the representative of Ranunculus subgen. Crymodes A.
leaf is a modification of pupating internally in the Gray.' Benson treated Beckwithia as synonymous
stem or flower-head, retained after the actual larval with Crymodes. The genitalia of P. sp. (Beckwithia)
feeding has advanced to leaf mining. An exactly (Fig. 149) suggest that it could be most closely relat-
comparable case can be seen in N. clematidicolla ed to P. callianthemi (Fig. 150).
(Spencer) in Australia (see above). Two black species in Europe belong in different
Among the 19 Phytomyza species on Ranunculus groups - P. ranunculicola Hering (ct. Spencer,
two distinct groups are present, with a number of 1976a) (Figs. 158, 159) appears to be related to the
more isolated species. The notata-group, with the species on Trollius (see above), while P. ranunculi-
distiphallus largely fused or divided only towards vora Hering (cf. Spencer, 1976a) (Fig. 160) with the
the end, consists of eight species present in Europe, distiphallus broad and recurved can be associated
Japan, North America, East Africa and Papua New with P. hendeli on Anemone and P. pulsatillae on
Guinea. The distiphallus is short in P. notata Pulsatilla, and also with P. multifidae in Alberta (see
Meigen (d. Spencer, 1976a) (Fig. 148) (= pseudo- under Anemoneae above). P. ranunculoides (in
notata Hering) showing a progressive evolutionary Spencer and Steyskal, 1986) (Figs. 161, 162) which
extension, from Phytomyza sp. (Beckwithia) (Fig. is only known from Minnesota, U.S.A., is clearly re-
149) in Norway to P. callianthemi Hering (Fig. 150) lated to P. loewii on Clematis (see above).
in Germany, to P. cortusifolii Spencer (1965) (Fig. In East and South Africa P. natalensis Spencer
151) in the Canary Islands to P. humilis Spencer (Fig. 163) and P. ranunculina Spencer (Fig. 164)
(1969b) (Fig. 152) present only in California and have the distiphallus strongly recurved and are pos-
Alberta. As the distiphallus lengthens further, it be- sibly associated with P. ranunculivora.
comes coiled, incipient in P. infelix Spencer (1969b) Two further species in Europe cannot be fully
(Fig. 153) (host unknown, Canada but probably clarified in the absence of males. P. buhriana Hering
Ranunculaceae) to P. caulinaris Hering (Fig. 154) (1949c: 23) is known from a single female reared
with two coils, known only in Germany, to P. multi- from R. sceleratus in Mecklenburg, northern Ger-
fidi Spencer (1985a) in Kenya (Fig. 155), to P. ra- many. Hering was unable fully to clarify the mine
42 Class MAGNOLIOPSIDA (Dicotyledons)
,I ./
148 .. /- ....
. ~.;:,.(.:. . <:.
I","
/
151
150
152
/:
.':,',
.'
153
Fig. 148. Phytomyza notata: aedeagus, Fig. 152. Phytomyza humilis: aedeagus,
Fig. 149. Phytomyza sp, (Ranunculus glacialis): aedeagus, Fig. 153. Phytomyza infelix: aedeagus,
Fig. 150. Phytomyza callianthemi: aedeagus, Fig. 154. Phytomyza caulinaris: aedeagus,
Fig. 151. Phytomyza cortusifolii: aedeagus.
Subclass MAGNOLIIDAE 43
but with the outer cross-vein present the species can with Aquilegia and subtribe Thalictrinae with Thal-
only be compared with P. rydeni from which it dif- ictrum.
fers both in colour and chaetotaxy. P. stolonigena
Hering (1949b) is known from two females also Tribe Isopyreae
from Mecklenburg. The adults are not satisfactorily
distinguishable from the variable P. ranunculi but Subtribe Aquilegiinae
the mines (Fig. 165) are distinctive, running along Aquilegia. This genus has been colonized by eight
the petiole, with offshoots into the leaf blade. The Phytomyza species, with clear radiation, particularly
species is not uncommon in Germany on various in North America in what may be referred to as the
Ranunculus spp. and I have found mines on aquilegiae-group. The only species in Europe is
R. repens at Hampstead, London. P. aquilegiae Hardy (cf. Spencer, 1976a) (Figs. 171,
Four further species feed on Ranunculus in 172). Its relatives in North America are P. aquilegia-
Europe in a second group. P. fallaciosa Brischke (cf. na Frost (Figs. 173, 174) (cf. Spencer and Steyskal,
Spencer, 1976a) (Figs. 166, 167) is closely related 1986), P. aquilegioides Sehgal (cf. Spencer and
to P. anemones (see above) but biological differ- Steyskal, 1986) (Figs. 175, 176), P. sonorensis
ences justify treating it as distinct. P. rydeni Hering Spencer (1981) (Figs. 177, 178) and P. columbinae
(cf. Spencer, 1976a) is an unusual species with the (cf. Spencer and Steyskal, 1986) (Figs. 179, 180).
outer cross-vein present and is only known in All are leaf miners. The differing biology of two
Scandinavia and Scotland. Although Hering de- species is clearly reflected in the genitalia. P. krygeri
scribed the species in Napomyza (at the time treated Hering (cf. Spencer, 1976a) in Europe feeds in the
as a subgenus of Phytomyza), the genitalia (Fig. 168) seed capsules (Figs. 181, 182) and in Canada
show that it correctly belongs in Phytomyza. P. aquilegiophaga Spencer (1969b) has once been
P. davisii Walton (cf. Spencer and Steyskal, 1986) recorded as from the root but probably feeds mainly
is a closely related species which is widespread in in the stem (Figs. 183, 184). All these species have
the eastern U.S., also having the second cross-vein the frons yellow and the mesonotum grayish-black,
and forming similar primary blotch mines (Fig. 169). with the distal tubules of the aedeagus symmetrically
The genitalia (Fig. 170) are also similar. This species paired. Two species in a different group are largely
pair represents a further case of dispersal via Be- black, with the mesonotum shining, and the dis-
ringia. tiphallus consists of a single tubule. P. minuscula
Myosurus. Phytomyza ranunculi has once been Goureau (cf. Spencer, 1976a), which is also com-
recorded on this host. mon on Thalictrum, is the only truly holarctic
species on Aquilegia. It is common in Europe and
known in the western U.S., also in the Himalayas in
Subfamily Isopyroideae
northern India. The genitalia (Fig. 185) associate it
Tamura (1968) insists on maintaining this subfami- with P. aquilegivora Spencer (cf. Spencer and Steys-
ly, with A quilegia , as distinct from the Thalictroi- kal, 1986) which is widespread in the United States
deae, with Thalictrum. However, he concedes (pers. (Figs. 186, 187). These two species are related to
comm., Jan., 1988) 'Of course I think the close P. thalictrivora Spencer (see below).
relationship between Thalictroideae and Isopyroi-
deae is apparent, and Thalictrum was derived from Tribe Coptideae
the Isopyroideae. But I think, at the same time, the Tamura (pers. comm.) has now decided to abandon
distinction between follicles or multiovular carpels his sixth subfamily Coptidoideae and Coptis is
and achenes or uniovular carpels cannot be under- included in a new tribe within the Isopyroideae.
valued.' It has to be stressed though that five agro- Coptis. The only species known on this genus is
myzids (three in Europe: Ophiomyia aquilegiana, P. tamui Sasakawa (cf. Sasakawa, 1961) in Japan on
P. aquilegiae and P. minuscula; and two in North C. quinquefolia. The genitalia (Fig. 188) suggest
America: P. aquilegioides and P. columbinae) feed relationship with the three Caltha feeders.
on both Aquilegia and Thalictrum and only on these
two genera. They thus find no distinction between
the genera. Subfamily Thalictroideae
As a result of his serological tests, Jensen (1968:
283) confirmed the close relationship between Tribe Thalictreae
Aquilegia and Thalictrum but found Aquilegia to be Thalictrum. Six Phytomyza species are restricted to
closer to Leptopyrum (on which no agromyzids are this genus, three in Europe and three in North
known) than to Thalictrum. He formalized this with America, and in addition P. aquilegiae and
a tribe Thalictreae, containing subtribe Isopyrinae P. minuscula which were discussed above under
44 Class MAGNOLIOPSIDA (Dicotyledons)
155
156
157
160
159
161 162
Subclass MAGNOLIIDAE 45
163 164
165 166
:
168 ;
' ~ - ' - . ".,
,. ::.:..
/'
169
Fig. 155. Phytomyza multifidi: aedeagus.
Fig. 156. Phytomyza ranunculi: aedeagus.
Fig. 157. Phytomyza aquilonia: aedeagus.
Figs. 158, 159. Phytomyza ranunculicola: 158, aedeagus, side view; 159, distiphallus, ventral view.
Fig. 160. Phytomyza ranunculivora: aedeagus.
Figs. 161, 162. Phytomyza ranunculoides: 161, aedeagus, side view; 162, same, ventral view.
Fig. 163. Phytomyza natalensis: aedeagus.
Fig. 164. Phytornyza ranunculina: aedeagus.
Fig. 165. Phytornyza stolonigena: leaf mine on Ranunculus repens.
Figs. 166, 167. Phytornyza!allaciosa: 166, aedeagus, side view; 167, distiphallus, ventral view.
Fig. 168. Phytornyza rydeni: aedeagus.
Figs. 169, 170. Phytornyza davisii: 169, leaf mine on Ranunculus abortivus; 170, aedeagus.
46 Class MAGNOLIOPSIDA (Dicotyledons)
171
~
r~ _
~<~~
:->"
.-. <-:---~
. . . ~ .. ,
:,:;.....
172
174
< 177
178
Subclass MAGNOLIIDAE 47
Aquilegia are equally common on Thalictrum in Menispermaceae (70 genera, 400 species)
Europe and in North America. P. aquilegioides and
Liriomyza cocculi (Frick, 1953) (described as Phy-
P. columbinae, described from Aquilegia, also occur
on Thalictrum. In Argentina Liriomyza cordobensis tobia (Praspedomyza) in view of its dark scutellum)
was recorded as a leaf miner on Cocculus sp. No
nom. nud. has been found on Thalictrum decipiens
details of the leaf mine are known. Leaf mines have
(~alladares, 1986) and is also known on Legu-
been referred to this species from Piperaceae
mmosae and Tropaeolaceae (see below).
(:eperomyia), Chenopodiaceae in the Caryophyl-
. P. thalictri Escher-Kundig in Rougemont (1912)
hdae, Fabaceae and Thymeleaceae in the Rosidae
IS only known in Switzerland where it has been
and in the Asteridae Apocynaceae, Lobeliaceae and
found feeding in seeds of T. aquilegifolium. The
Myoporaceae. L. cocculi on the basis of these
smooth, strikingly shining orange-brown puparium
records has been treated as a polyphagous species.
suggests that this species is related to P. minuscula
However, with no adults known from any of these
which has similar puparia. However, the genitalia
hosts the identification must be treated as suspect or
(Figs. 189, 190) associate the species with the seed-
at least requiring confirmation, particularly in view
feeder on Aquilegia, P. krygeri (d. Figs. 181, 182).
of the presence in Hawaii of two other polyphagous
A lectotype of P. thalictri is designated in the
species, L. huidobrensis and L. sativae.
Appendix below. P. thalictricola Hendel (d. Hen-
The dark scutellum suggests that L. cocculi may
del, 1935) (Fig. 191) forms blotch mines and
well be a primitive species and its presence on
P. albifrons Groschke (1957) (Fig. 192) which
Piperaceae can probably be accepted as correct.
closely resembles it, forms narrow linear mines. The
The male genitalia (Figs. 197, 198) show a close
genitalia indicate that the two are closely related.
relationship with L. huidobrensis. It is possible that
In North America P. plumiseta Frost (d. Spencer
L. cocculi represents an early immigrant into Ha-
and Steyskal, 1986) forms large, white blotch mines
waii from ancestral stock of L. huidobrensis which it
on Thalictrum polygamum and other T. spp. and the
is believed is related to the L. bryoniae/strigata com-
genitalia (Figs. 193, 194) clearly associate this
plex in Europe. On this assumption L. huidobrensis
species with P. aquilegiana which forms blotch
speciated in North America after crossing Beringia
mines on Aquilegia. P. thalictrella Spencer (1981) is
and L. cocculi could have speciated in isolation in
only known from a high elevation in California with
the single host T. fendleri. Details of the leaf'mine Hawaii. It will thus be of great interest to obtain
adults from further hosts in Hawaii to clarify this
were not recorded but the genitalia (Fig. 195) give
problem further.
no clear indication of its relationships. P. thalictri-
vora Spencer (1969b) was described from Alberta
In Australia the polyphagous Trop. polyphyta has
forming short linear mines. The genitalia (Fig. 196) been recorded on Stephania.
indicate relationship with P. aquilegivora (Fig. 186).
In addition to the five species mentioned above Order PAPAVERALES (2 families,
feeding on both Aquilegia and Thalictrum, it can 600 species)
now be seen that there are also two pairs of sister-
species on the two genera: P. aquilegivoraiP. thalic- Papaveraceae (25 genera, 200 species)
trivora; P. aquilegianaiP. plumiseta. This further
confirms the close relationship between Aquilegia ~nly a single ~amily-specific species, Phytomyza par-
vl~ella (~oquIllett, d. Spencer, 1969b), is known on
and Thalictrum.
thIS famIly, on Papaver radicatum. It was described
from Alaska and has subsequently been found in the
Lardizabalaceae (8 genera, 30 species) ~ountains in Poland. It is thus a truly holarctic spe-
CIes. The outer cross-vein has been retained result-
In Japan Phytomyza akebiae (Sasakawa, 1954a) is a
ing in its original description in Napomyza ~d the
leaf miner on Akebia quinata. Only females are
genitalia (Figs. 199, 200) indicate that it is an an-
known but with the outer cross-vein present, it is
cient, isolated species.
probably a primitive species.
F~gs. 171, 172. Phytomyza aquilegiae: 171, aedeagus, side view; 172, distiphallus, ventral view.
F~gs. 173, 174. Phytomyza aqu~legia~a: 173, aedeagus, side view; 174, distiphallus, ventral view.
F~gs. 175, 176. Phytomyza aqUlleglOldes: 175, aedeagus, side view; 176, distiphallus, ventral view.
F~gs. 177, 178. Phytomyza sonorensis: 177, aedeagus, side view; 178, distiphallus, ventral view.
FIgs. 179, 180. Phytomyza columbinae: 179, aedeagus, side view; 180, distiphallus, ventral view.
48 Class MAGNOLIOPSIDA (Dicotyledons)
-_.. --
d
'.'
181 183
182
185
187
Figs. 181, 182. Phytomyza krygeri: 181, aedeagus, side view; 182, same, ventral view.
Figs. 183, 184. Phytomyza aquilegiophaga: 183, aedeagus, side view; 184, same, ventral view.
Fig. 185. Phytomyza minuscula: aedeagus.
Figs. 186, 187. Phytomyza aquilegivora: 186, aedeagus, side view; 187, same, ventral view.
Fig. 188. Phytomyza tamui: aedeagus (Sasakawa).
Subclass MAGNOLIIDAE 49
:.....
-'.
,.,
190
189
192
,c~,~
/! 195
193 ,...~
/'.~
Figs. 189, 190. Phytomyza thalictri: 189, aedeagus, side view; 190, distiphallus, ventral view.
Fig. 191. Phytomyza thalictricola: distiphallus.
Fig. 192. Phytomyza albifrons: aedeagus.
Figs. 193, 194. Phytomyza plumiseta: 193, aedeagus, side view; 194, distiphallus, ventral view.
Fig. 195. Phytomyza thalictrella: aedeagus.
Fig. 196. Phytomyza thalictrivora: aedeagus.
50 Class MAGNOLIOPSIDA (Dicotyledons)
Some botanists have associated the Papaverales sphere (South Africa, Australia, New Zealand), no
with the Capparales but chemically these orders Phytomyza is known on the Asteraceae but all feed
have nothing in common, while many of the second- on Ranunculaceae. It thus appears that the early
ary compounds present in Papaverales clearly asso- species which reached the southern hemisphere by
ciate the order with the Ranunculales. It seems long-distance 'mountain hopping' from the north
probable therefore that the ancestor of P. parvicella were all Ranunculaceae feeders. When collecting in
will be in the Ranunculales, although its affilitations South Africa, it was striking to find composites
are not immediately apparent. completely ignored by leaf miners, even when grow-
Two polyphagous species, Liriomyza strigata and ing in the immediate proximity of species feeding on
Chromatomyia horticola have succeeded in coloniz- ranunculaceous genera.
ing Glaucium and Papaver in Europe. The differences in larval feeding, which frequent-
ly accompany mutations involving external mor-
The dominance of Agromyzidae on the Ranun- phology or male genitalia, can be illustrated by Phy-
culaceae is very apparent from the above analysis, tomyza ranunculi which forms an irregular linear
this family containing all but three of the family- mine on any part of the leaf and P. stolonigena
specific species on the subclass. Nevertheless, the Hering which feeds initially along the midrib, with
Ranunculaceae are a family which is not generally offshoots along lateral veins (Fig. 165).
acceptable to Agromyzidae. This is indicated by its Within the Ranunculaceae the agromyzids show a
colonization by only seven agromyzid genera (Table remarkable degree of generic specificity (Table 6).
5), the complete absence of Agromyza, the presence All 14 Napomyza species are host-specific on either
of only two family-specific Liriomyza species, and Clematis (2) or Ranunculus (12). The two Chroma-
its general unacceptability to polyphagous species, tomyia and one Ptochomyza species are only known
with only isolated records of Liriomyza xanthocera on Clematis. In Phytomyza 83 species are restricted
(= L. crucifericola), L. trifolii, L. cordobensis and to a single genus, two species occur on two genera in
Chromatomyia horticola. Many genera in the family the same tribe (P. aconiti, P. buhri), four on both
are toxic and obviously repellent, and only 18 of the Aquilegia and Thalictrum (P. aquilegiae, P. minus-
available 50 genera have been colonized (Table 6). cula; P. aquilegioides, P. columbinae; and also
Despite this, the family supports the remarkably Ophiomyia aquilegiana). The presence of P. kalten-
high total of 116 family-specific species, of which 14 bachi primarily on Clematis but also on Actaea (as
are in Napomyza and 91 in Phytomyza. P. philactaeae) and of P. actaeae primarily on Actaea
After carefully considering the species present on but also on Clematis appears to be exceptional.
all other families of dicots, I have reached the in- It is thus clear that transfer between genera is
escapable conclusion that the colonization by Napo- extremely rare. The evolutionary radiation which
myza and Phytomyza is primary, not secondary fol- has produced 14 Napomyza and 91 Phytomyza
lowing switches from other families. Napomyza is species has obviously been a slow process and, in
the more primitive genus, and its long association view of the difficulties in colonization even of re-
with the Ranunculaceae is evident from the pres- lated genera, this might possibly represent a case
ence of holarctic species on Ranunculus (N. evanes- mainly of co-evolution of the agromyzids together
cens, N. nigritula) and particularly the spread of the with their host genera.
genus on Ranunculus and Clematis to South Africa, The difficulty for species in switching between
Australia and New Zealand. The initial evolution of genera even when adjusted to the unusual chemistry
Phytomyza may well have been from ancestral of the family is clear evidence for the even greater
Napomyza species on the Ranunculaceae. Napo- difficulty species on other families have in coloniz-
myza species are primarily internal feeders in the ing the Ranunculaceae. The presence of only two
stem or receptacle but significantly I found N. lyalli family-specific Liriomyza species (L. thalictri and
in New Zealand with many puparia in the stem but L. ranunculoides) and only three polyphagous spe-
some larvae already feeding as leaf miners. From the cies fully supports this.
genitalia and larval characters I accept several spe- Conversely, I have identified four cases of switch-
cies in Napomyza which are exclusively leaf miners ing from the Ranunculaceae to more advanced fami-
in South Africa (N. eximia, N. renovata), Australia lies. The species on Aquifoliaceae (P. ilicis group,
(N. clematidicolla) and in New Zealand (N. costata). Figs. 541, 542) appear to be derived from ancestors
Significantly, in these cases pupation takes place in of the P. fallaciosa group (Ranunculus). P. sedicola
the leaf, indicating that the species have evolved (Crassulaceae) (Figs. 350, 351) is clearly related to
from more typical, ancestral Napomyza species P. hendeli or relatives (Anemone) - Figs. 69, 70 -
which would have pupated internally in the stem. or P. ranunculivora - Fig. 160. The P. angelicae
The favourite host family of Phytomyza is the group on Apiaceae also show unmistakable rela-
Asteraceae, yet remarkably, in the southern hemi- tionship with the fallaciosa group and are certainly
Subclass MAGNOLIIDAE 51
Oi5~;jt;7~
Ej Jfi[j§~--p<uc:
198
197
Figs. 197, 198. Liriomyza cocculi: 197, aedeagus, side view; 198, same, ventral view.
Figs. 199,200. Phytomyza parvicella: 199, aedeagus, side view; 200, distiphallus, antero-ventral view.
more closely related to this group than to the other present on other primitive families in this Subclass
main group on Apiaceae - P. spondylii group; final- probably reflect relatively recent colonization by a
ly P. anthocercidis in Australia (Solanaceae) clearly backwards switch from a more modem family, with
represents a switch from an ancestor of P. orienta lis inadequate time for any subsequent radiation on
(Clematis) or P. ranunculi (Ranunculus) - Fig. 156. relatives of the newly acquired but unfamiliar host.
The three cases in which only single species are
Subclass HAMAMELIDAE (11 orders, 24 families, 3400 species)
It is agreed among botanists that this is an ancient group representing 'a series of broken-up fragments
group which can be traced back to the Lower Creta- - the main orders being relictual survivors of an
ceous, some 100 million years before present. early phase of angiosperm evolution and remnants
Cronquist (1981: 152) suggests that while the plata- of a broad transitional group between the Magnoli-
noid line was differentiating from the magnoliids, a idae and the Rosidae-Dilleniidae.'
sister-group was differentiating as the probable It was accepted at the Reading meeting that the
ancestors of the Rosidae. He also accepts specula- Urticales 'do not have close relationships with other
tion that the origin and early expansion of the Hamamelidae but should be placed elsewhere.' It
Hamamelidae took place in Laurasia rather than was suggested that the Urticales have a sister-group
Gondwanaland. Interestingly, there is a striking relationship with the Malvales and, more remotely,
relationship between agromyzids on one hand on with the Violales. This was in fact already formally
the Ulmaceae and Betulaceae and on the other with proposed by Thome (1983).
species on the Rosaceae. Relationships between the '[he new proposals regarding the relationships
Hamamelidae and Rosidae are supported by similar within the Hamamelidae and of the Subclass with
chemistry. Polyphenols are common to both groups, other Subclasses are still to some extent conjectural
particularly prodelphinidin, myricetin and certain and I am inclined here to adopt the established
tannins (ellagic acid) (Frohne and Jensen, 1985). framework provided by Cronquist (1981), although
Here, it may be mentioned that Thome (1983: 106) it is apparent that this will certainly be modified in
in his realignment of angiosperms, places his Super- the future.
order Hamamelidifiorae adjacent to the Rosifiorae, Only four of the ten orders in the Hamamelidae
and his order Fagales immediately adjacent to the are known to support Agromyzidae with 30 species
Rosales. This system better explains the many cases in 13 genera recorded below. Of these, three are
of related insects, in addition to Agromyzidae, cambium borers, six stem-borers, one (known only
which feed on Betulaceae and Rosaceae. The close from larvae) is a gall-causer and 19 are leaf miners.
relationship of the 'Amentiferae' and Rosaceae was In addition, Suss and Muller-Stoll (1980) and Suss
proposed as early as 1925 by Mez on the basis of (pers. comm.) have detected larval feeding of Phy-
serum diagnosis, and Hering (1951a), who largely tobia species in trees in five previously unrecorded
followed Mez in his discussion of plant relation- families (see below); three common polyphqgous
ships, prophetically wrote: 'We are convinced that species are also known here. ,
future investigations, possibly with more subtle The agromyzid genera known on Hamamelidae
methods of serum diagnosis, will confirm the rela- are listed in Table 7 and the host families and genera
tionship between these two groups of plants.' are shown in Table 8.
Reference must also be made here to the recent
Symposium on the Hamamelidae held at Reading
Order HAMAMELIDALES (5 families,
University, England, 22-25 March, 1988. In a brief
110 species)
discussion of this Symposium, Humphries (1988)
comments on the 'remarkable progress made in No extant Agromyzidae are known on this order but
palaeobotany over the past decade' which has re- one fossil species has been described from Platan-
sulted in a clearer understanding of the origins and aceae and larval feeding of Phytobia has been
subsequent evolution of the Hamamelidae. The detected in four genera of Hamamelidaceae.
early divergence of the Hamamelidae and Rosidae
in the mid-Cretaceous is supported both by pollen
Platanaceae (1 genus, 6-7 species)
and megafossils, with 'a wealth of wholly fossil taxa,
and it appears that modem genera made their first One species ancestral to modem Phytobia was
real appearances in the Tertiary.' Humphries states described by Suss and Muller-Stoll (197 5) as
that it is now considered that the Hamamelidae are a Palaeophytobia platani, based on the distinctive
SubclassHAMAMELIDAE 53
Table 8
Order HAMAMELIDALES
Platanaceae
Platanoxylon Palaeophytobia platani
Hamamelidaceae
Altingia Phytobia sp.(I)
Distylium Phytobia sp.(I)
Hamamelis Phytobia sp.(I)
Liquidambar Phytobia sp.O)
Order URTICALES
Ulmaceae
Celtis Phytobia sp.; Cecidomyiaceltis deserta; Agromyza trebinjensis, A. varifrons
Ulmus Phytobia sp. (I, Europe, W. Africa), (2, N.America); A. aristata
Cannabaceae
Cannabis Liriomyza cannabis, L. strigata; Chromatomyia horticola
Humulus A. fJaviceps, A. igniceps, A. pseudoreptans (?); Chr. horticola
Moraceae
Broussonetia Phytobia sp.(I)
Cudrania Tropicomyia (?) theae
Ficus Phytobia sp.O); Ophiomyia fici
Moquilea Phytobia sp.(2)
Morus Phytobia sp.(3); A. morivora
Olmediopsis Phytobia sp.(?)
Sorocea Phytobia sp.(4)
54 Class MAGNOLIOPSIDA (Dicotyledons)
Urticaceae
Tribe Urticeae
Laportea A. pseudoreptans
Urera Calycomyza sp.
Urtica Melanagromyza aenea, M. martini, M. urticella, M. urticivora;
A. anthracina, A. hiemalis, A. pseudoreptans, A. rep tans; L. urticae;
Napomyza sp. n. (Japan); Phytomyza flavicornis
Tribe Boehmerieae
Boehmeria Melanagromyza boehmeriae ( China)
Tribe Parietarieae
Parietaria A. anthracina, A. pseudoreptans
Phenax Calycomyza sp. n.
Order JUGLANDALES
Juglandaceae
Pterocarya Phytobia sp.
Order FAGALES
Fagaceae
Castanea Japanagromyza viridula
Quercus Phytobia sp. ( 1, U.S.A.), ( 2, Europe ), ( 3, Mediterranean); J. quercus,
J. viridula
Betulaceae
Alnus Phytobia betulae, Phb. betulivora, Phytobia sp.; A. alnivora
Betula Phb. betulae, Phb. betulivora, Phytobia sp. (I, Manchuria); A. alnibetulae, A. betulae
Carpinus Phb. betulae
Corylus Phb. betulae
Ostrya Phytobia sp. (2, N. America)
Hamamelis mollis Olivo - China Suss (loc. cit.) has detected larval feeding of
Liquidambar styraciflua L. - USA, Central America Phytobia on two genera of Ulmaceae:
These records probably refer to at least 3 distinct Celtis sp. - West Africa
species. Ulmus carpinifolia Gleditsch - Europe, North
Africa, Western Asia
U. glabra Uds. - North and central Europe, Western
Order URTICALES (6 families, 2200 species) Asia
U.laevis Pall. - North America
There is some disagreement among botanists con-
cerning the correct affiliation of this order. Thorne These records probably represent at least three
(1983: 105) may be mentioned as one who con- distinct species.
siders that the Urticales are allied to the Malvales. A species of considerable interest is known in the
There is no relationship between agromyzids on United States on Celtis occidentalis with the unfor-
Urticales and Malvales to support this and here I tunate name Cecidomyiaceltis deserta Patton (1897).
therefore follow Cronquist. Patton, a specialist on Cecidomyiidae, discovered
galls on twigs of Celtis occidentalis at Orange, Conn.
and believed the species was a cecidomyiid in a new
Ulmaceae (18 genera, 150 species)
genus and hence the name. Dr. R. Gagne, of the Sys-
This is acepted as an ancient family, with pollen of tematic Entomology Laboratory, U.SDA., Wash-
Ulmus type present in deposits of Maestrichtian ington, rediscovered this species on the same host in
(late Upper Cretaceous). Montgomery Co., Maryland in 1983 and, although
SubclassHAMAMELIDAE 55
no adults were obtained, it is clear from the dead Cannabaceae (2 genera, 3 species)
larva kindly sent to me by Dr. Gagne that the species
Two Agromyza leaf miners are known on this fami-
is an agromyzid. The larval characters show th~t the
ly, both in Europe on Humulus, but it appears. that
species is close to Melanagromyza but, .untIl the
the two species are not directly related. A. f!avlceps
male genitalia of an adult can be exanuned, the
Fallen (cf. Spencer, 1976a) forms a conspicuous
generic name Cecidomyiaceltis is retained (Spencer,
blotch mine and the genitalia (Fig. 203) suggest that
1988). . . this is an isolated species. A. igniceps Hendel (d.
This will certainly prove to be an Isolated specIes
Spencer, 1976a) forms irregu~ar ~inear-blo~ch mines
and it will be of great interest in due course to estab-
and the genitalia (Fig. 204) IndIcate pOSSIble rela-
lish its affiliations.
tionship with species on Leguminosae. It see.ms .that
Three Agromyza leaf miners are known on the
these two species represent separate colomzatIons
Ulmaceae, one in Europe, two in North America.
of Humulus and a long period of isolation on this
Agromyza trebinjensis Strobl (d. Hendel, 1?31) was
host has now obscured their affiliations.
described from Yugoslavia and NowakowskI (1960)
Buhr (1937) records Agromyza reptans (probably
discovered the host to be Celtis australis, giving a
A. pseudoreptans) on Humulus in the B?tanical
detailed description of the species as Agromyza
Gardens Rostock but this represents an Isolated
celtidis sp.n. The synonymy of A. celtidis with
transfer ~nd almost certainly does not occur in the
trebinjensis was established by Spencer (19.66e).
wild.
This species is closely related to A. varif!0~s Liriomyza cannabis Hendel (1931) is the only
Coquillett (cf. Spencer and Steyskal, 1986) WhICh IS
leaf miner known on Cannabis (apart from two po-
common from Ontario and Quebec to the southern
lyphagous species). It was originally identi~ed. by
United States on Celtis laevigata and C. occidentalis.
Hering (1921) as L. eupatorii (Kaltenbach) In VIew
There are significant differences in the external
of the similarity of the leaf mine which begins with a
morphology of the two species but their sister-
conspicuous spiral before continuing with. a sh.ort
species relationship can be deduced from the com-
linear section. The genitalia of L. cannabls (FIgs.
mon host and the close similarity of the male gen-
205,206) differ only slightly from those of L. .eupa-
italia - A. trebinjensis (Fig. 201) and A. vari-
torii (Figs. 1099, 1100) but the difference In the
frons (Fig. 202). This represents a clear case of
distiphallus in ventral view justifies retention. of
speciation on either side of Beringia fro~ the single
L. cannabis as a distinct species. There seems httle
ancestral popUlation which was present In the Arc-
doubt that this represents a relatively recent switch
to-Tertiary forest which extended without break
from Eupatorium, with still only slight differences in
from Northern Europe to Siberia and across Be-
external morphology, genitalia and leaf mines be-
ringia into North America. Nowakowski (1960: Fig.
tween the two species (cf. Spencer, 1976a: 236). It is
6e) illustrates the larval posterior spiracles, .each
possibly significant in this connection that L. eupa-
process having three pores (contrast A. anstata
torii is found regularly on Galeopsis tetrahit (Lami-
below).
aceae) in addition to its normal host Eupatorium
The genitalia of these two species indicate rel~
(see below under Compositae). This may represent
tionship with several species on Rosaceae and It
an even more recent switch, with no sign yet of
seems possible that they may have been precursors
of the radiation which is known on Rosaceae today speciation on the two hosts. Such. ca~e~ may t~us
develop into incipient polyphagy or InCIpIent specia-
(see below).
tion but they are rare.
Agromyza aristata Malloch (cf. Spencer and
Steyskal, 1986) (= A. ulmi Frost) is widespread on
Ulmus americana from Ontario throughout the east- Moraceae (40 genera, 1000 species)
ern United States to Arkansas. Although Ulmus is a
This family is widespread in tropical and subtropical
well-known component of the Arcto-Tertiary flora,
regions and by far the largest genus is Ficus, with
no leaf miner is known on this genus in Europe.
some 500 species. Fossil pollen has been recorded
A. aristata is possibly related to the Celtis feeders
at least to early Eocene. Only two described agro-
but is significantly differentiated.
myzids are known as host-specific on the Moraceae.
In contrast to A. trebinjensis in which the larval
Suss (loc. cit.) has detected larval feeding of Phy-
posterior spiracular processes have the primitive
tobia in six genera.
arrangement of three pores, when collecting in
Arkansas in 1979, I found that A. aristata has 8-10 Broussonetia papyrifera Vent. - S.E. Asia
pores. This might well indicate that A. aristata. is the Ficus benjamina L. - Java
more derived species, if indeed they are dIrectly F. hispida L. - Java
related. F. pilosa Reinw. ex Blume - Java
Moquilea tomentosa Benth. - Brazil
56 Class MAGNOLIOPSIDA (D"Icotyledons)
201
:~ ~i·~gp
202
203
206
207 208
211
Subclass HAMAMELIDAE 57
Morns nigra L. - Caspian Sea area, Caucasus fusion in rearing records. This synonymy is estab-
Olmediopsis obliqua Krst. - (no locality recorded) lished in the Appendix below.
Sorocea affinis Hemsl. - Panama Of the four Agromyza leaf miners on Urtica in
These records must represent at least 4 species. Europe, A. reptans Fall. (Fig. 214) and A. hiemalis
Becker (cf. Spencer, 1965a) (Fig. 215) are directly
Ophiomyia fici Spencer and Hill (1976~ is a uni~ue
related, with A. reptans widespread in central and
species only known in Hong Kong, feedmg on Flcus
northern Europe, while A. hiemalis Becker is
macrocarpa. The female lays two eggs in the leaf
restricted to the Mediterranean area. A. anthracina
each side of the midrib near its base and the two lar-
Meigen (cf. Spencer, 1976a) belongs to the same
vae form raised linear mine-galls towards the apex group but is more differentiated both in external
of the leaf (cf. Spencer and Hill, 1976: fig. 3). With
morphology and male genitalia (Fig. 216). The
the puparium remaining in the leaf, the posterior
fourth species, A. pseudoreptans Nowakowski (orig-
spiracles are characteristic ?f. 0p'hiomyia ~nd the inally as A. urticae) which was confused with A. rep-
male genitalia, although dIstmctIve, confIrm .t~e tans until the examination of genitalia (Fig. 217) by
generic position of the species i~ Ophi~myia: !~s IS Nowakowski (1964), is more closely related to
an interesting and isolated species and ItS afflhatIOns species feeding on Compo sitae (A. rufipes Meigen)
may be clarified when more collecting is undertaken
and five species restricted to Boraginaceae (see
in S.E. Asia and particularly southern China. Subclass Asteridae below).
Agromyza morivora Sasakawa and F~kuhara
Nowakowski (1964: 209) suggested a common
(1965) is known only in Japan as a.l.ea~ mmer o.n origin on Urtica of A. reptans and A. pseudoreptans,
Morus (cultivated mulberry). The affihatIons of thIS
and thus direct relationship in view of their common
species are not clear (Figs. 207, 208).
host. Logically this should be correct. However, the
obviously closer relationship between A. pseudorep-
Urticaceae (45 genera, 700 species) tans and both A. rufipes (Artemisia) and the five
species on Boraginaceae than between it. and the
This family is widely distributed in tropical and sub-
other species on Urtica can better be explamed by a
tropical areas. It has been divided into five. tribes,
switch from Urtica to the Asteridae (Boraginaceae
with 12 species of Agromyzidae known on SIX gen-
feeders + Artemisia).
era in three tribes (see Table 8).
Liriomyza urticae (Watt, cf. Spencer, 1976b) is
Four Melanagromyza species have been found in
the only known Liriomyza feeding on Urtica, the
stems of Urtica, M. aenea (Meigen, cf. Spencer,
host being U. ferox in New Zealand. The male gen-
1976a) in Europe, M. martini Spencer (1969b) in
italia (Figs. 218, 219) are typical of the genus and
Canada, M. urticella Spencer (1981) in California
the species may have switched from Hebe (Scroph-
and M. urticivora Spencer (1966a) in Pakistan.
ulariaceae) which is the commonest host of N.z.
There is no positive evidence that these species are
Liriomyza species or possibly from Melicytus (Viol-
directly related but equally it is not impossible that
aceae), the host of L. Jlavolateralis (Watt), the gen-
they are derived from a single colonization of Urtica.
italia of which closely resemble those of L. urticae
The male genitalia of M. aenea (Figs. 209, 210) and
(Figs. 287, 288). Humphries (1988) suggested a
the larval posterior spiracles (Fig. 211) are typical of
possible relationship between the Urticales and
many groups of Melanagromyza, while the most
Viol ales and, if this is confirmed, it could explain
differentiated of the four species is M. urticella
this switch between Urtica and Melicytus.
known only in California (Figs. 212, 213).
A leaf miner on Phenax sonneratii was originally
Wenn (1985) described M. boehmeriae from
discovered at the edge of tropical forest in Venezu-
stems of Boehmeria nivea in Zhejiang Prov., China,
ela (Spencer, 1973c: 80) but I failed to rear adults,
comparing it with M. urticella. .
recording the species as either a Liriomyza or Caly-
I am satisfied that M. radicicola Steyskal (m
comyza. Recently Dr. Valladares obtained a Calyco-
Spencer and Steyskal, 1986) which was presumed to
myza sp.n. on Phenax in Argentina which I accept
feed in the roots of Urtica, in fact represents
as identical with the species in Venezuela. This
M. minimoides and was described as a result of con-
212
to 213
215
Figs. 212, 213. Melanagromyza urticella: 212, aedeagus, side view; 213, same, ventral view.
Fig. 214. Agromyza reptans: aedeagus.
Fig. 215. Agromyza hiemalis: aedeagus.
Fig. 216. Agromyza anthracina: aedeagus.
Fig. 217. Agromyza pseudoreptans: aedeagus.
Figs. 218, 219. Liriomyza urticae: 218, aedeagus, side view; 219, same, ventral view.
clearly represents a switch from one of the many Phytomyza flavicornis Fallen (cf. Spencer,
hosts on which this genus is known in the Neotropi- 1976a), one of the few known stem-boring Phyto-
cal Region but in the absence of males, the affilia- myza species, is strictly monophagous on Urtica.
tions of the species cannot be determined. The posterior spiracular processes of the puparium
Napomyza lateralis Fall., a species restricted to are in the form of two plates (as in Melanagromyza)
Compo sitae, was recorded in stems of Urtica japo- each bearing some 20 irregular pores. In the male
nica L. in Japan (Sasakawa, 1961: 424) but this genitalia the postgonites are uniquely extended
almost certainly represents a misidentification and anteriorly and the aedeagus (Fig. 220) confirms the
an undescribed species. isolated position of the species. I am satisfied
Subclass HAMAMELIDAE 59
that P. luteiceps Sehgal (1971b), caught in Alberta Gulf of Mexico, while 1. quercus (cf. Sasakawa,
but with the host suggested as Urtica, is identical 1961) is known only in Japan, on Quercus glauca.
with P. f/avicornis and this synonymy is formally Both form blotch mines and the male genitalia are
established in the Appendix below. P. f/avicornis, generally similar (J. viridula, Fig. 221). The larvae
with a holarctic distribution, strongly projecting provide further confirmation that the two species
orbits and very deep jowls, unique larval spiracles are in the same group, with the posterior spiracles
and genitalia, is clearly a primitive species and may each having three main processes and these are
well be closely associated with Napomyza. further subdivided, with 12-14 individual spiracular
pores in 1. quercus (Fig. 223, larva) and irregularly
three or four in 1. viridula (Fig. 222, puparium).
Order JUGLANDALES (2 families, 60 species)
With Quercus being a distinctly unacceptable host
This is certainly an ancient order, with pollen recog- for Agromyzidae, I suggest that the two known leaf
nisable in the Upper Cretaceous. There is an on- miners, both in the same genus, represent a single
going difference of opinion among botanists on the original colonization, with speciation each side of
affiliations of this order. No adult Agromyzidae are Beringia. 1. viridula has also been found on the
known on the order but larval feeding of a Phytobia introduced Castanea mollissima (Chinese Chestnut)
has been detected in Pterocarya caucasica C. A. in Pennsylvania, U.S.A. It was originally thought
Mey. in Transcaucasian countries and Japan (Suss, that the species concerned was distinct from 1. viri-
loco cit.). dula but I am satisfied that, with no possibility of an
endemic U.S. species on this introduced tree, 1. viri-
dula has successfully transferred to Castanea which
Order FAGALES (3 families, 900 species)
is closely related to Quercus.
This is clearly an ancient group, with pollen trace- Remarkably, however, 1. viridula is clearly more
able back to the Campanian of the Upper Creta- closely related to 1. salicifolii (Fig. 298), a leaf miner
ceous. Wood considered to represent the tribe on Populus and Salix in southern Europe and
Coryleae of the Betulaceae has been found in upper- extending eastwards through Israel and Turkey to
most Campanian deposits. Both leaves and pollen of Kirghizia, U.S.S.R. (see below under Salicaceae).
Alnus appear in the Maestrichtian (Cronquist, There is a more obvious difference between the
1981). adults, with the mesonotum being conspicuously
The two well-known families Fagaceae and Betul- grey in 1. viridula, more shining black in 1. salicifolii,
aceae each have their own agromyzids but in all only than in the male genitalia which show only the smal-
seven described species are known on the order. lest difference. The switch from Quercus to Salica-
ceae has thus probably occurred late in the Tertiary
in the Arcto-Tertiary forest, with speciation each
Fagaceae (6-8 genera, 800 species)
side of Beringia but with little detectable difference
Agromyzidae are only known on Castanea and in the genitalia.
Quercus, not on Fagus or Nothofagus (see below). Givulescu (1984) described as Phytomyzites quer-
Larval feeding of unknown Phytobia species has ci a mine on a fossil Quercus leaf found at Chiuz-
been recorded on Quercus rubra in Pennsylvania, baia, Romania from the Pliocene. However, in view
U.S.A. by Greene (1914) and in northern France on of the total lack of Agromyzidae mining Quercus in
Quercus robur (Martinez, pers. comm.). Also Suss Europe today and with over 70 mining micro-Iepi-
(loc. cit.) records larval feeding of Phytobia on doptera on this host (Hering, 195 7b), the likelihood
Quercus ilex L. in the Mediterranean area, on of this fossil mine representing an agromyzid must
Q. laurina Humb. et Bonpl. at Hidalgo, Michoacan, be considered as remote and this species is not
Mexico and on Q. pedunculata Ehrh. - Europe, accepted.
North Africa, western Asia. Knigge and Bonnemann (1969) recorded larval
I have seen a female of an undescribed Phytobia feeding of an undescribed Phytobia sp. in Notho-
sp. caught on Mt. Meron in northern Israel (leg. fagus dombeyi in southern Chile. It was with some
Freidberg) in an area where the dominant tree was excitement therefore, during a visit to Chile in 1978,
Quercus ilex and it now seems certain that this was that I made a point of visiting the University at Val-
the host. More material is awaited before this spe- divia where this 'agromyzid' had been discovered,
cies is formally described. only to find that it had subsequently been estab-
Two leaf miners are known on Quercus, both in lished that the species concerned was not an agro-
the same genus, lapanagromyza. 1. viridula (Coquil- myzid but an undescribed species of Opostegidae
lett, cf. Spencer and Steyskal, 1986) is not uncom- (Micro-Lepidoptera).
mon on Quercus rubra from Nova Scotia and
Ontario throughout the eastern United States to the
60 Class MAGNOLIOPSIDA (Dicotyledons)
.·u·--~)~--~.:
. .~~1
. :.(".'C; :.).~FT....
~.P-
221
220
" I
226
227
228
225
Subclass HAMAMELIDAE 61
Betulaceae (6 genera, 120 species) and A. alnibetulae Hendel (1931) to Betula. The
latter name is unfortunate, resulting from Hendel's
Five described species of Agromyzidae are known
belief when describing it that the species on Betula
on this family, in the genera Alnus, Betula, Carpinus,
was identical to that on Alnus. The two species are
or Corylus. In addition larval feeding of Phytobia
generally similar but there are distinctive differences
has been recorded in Ostrya virginiana in North
(Spencer, 1976a: 93) and these are confirmed as of
America (Suss, loco cit.).
specific significance by the male genitalia - A. alni-
Two Phytobia species have been described on
vora (Fig. 226) and A. alnibetulae (Fig. 227). There
this family, Pb. betulae Hendel (1931) which has
is a further species on Betula in Japan, A. betulae
been reared from Alnus, Betula, Carpinus and
Sasakawa (1961), which the genitalia confirm as dis-
Corylus in Europe and P. betulivora Spencer
tinct (Fig. 228). Sasakawa (1961: 314) suggests this
(1969b) in Canada. The male genitalia show that
species 'is very closely allied to A. spiraeae Kalten-
these two species belong to separate groups (Fig.
bach' (now a junior synonym of A. potentillae
225) (d. Pb. betulae (Fig. 224) and Pb. betulivora
(Kaltenbach» and the genitalia of A. potentillae do
which may represent early speciation and differen-
indicate a probable relationship (Fig. 366).
tiation on the same host genus either side of Berin-
The inacceptability of families in the Hamameli-
gia. The genitalia of Pb. betulae are strikingly similar
dae to polyphagous species is noteworthy, with only
to those of Pb. cambii on Salicaceae (Dilleniidae)
Liriomyza strigata and Chromatomyia horticola
(Fig. 293) and there must have been an early switch
recorded on Cannabis (Cannabaceae), and Tropico-
from Betula to either Salix or Populus resulting in
myia theae (probably) on Cudrania (Moraceae).
speciation on these genera.
Suss (loc. cit.) has detected larval feeding of
This is a relatively small Subclass and, as might be
Phytobia on the following hosts:
expected, the number of agromyzid species re-
Alnus acuminata HBK. - southern New Mexico, stricted to it is also small (30). This assemblage is of
Arizona, Mexico, Central America to Peru particular interest, containing species which are
A. glutinosa (L.) Gaerth. - Europe, Siberia, North holarctic (Agromyza pseudoreptans, A. reptans; Phy-
Africa (probably all P. betulae) tomyza flavicornis) and species pairs which have
A. incana (L.) Moench - Europe, North America diversified each side of Beringia (Japanagromyza
A. viridis (Chaix) DC. - Europe (Alps, Carpathians, viridulal l. quercus; Agromyza trebinjensisl A. vari-
Balkans) frons). P. flavicornis is clearly a primitive species
Betula davurica Pall. - Dahurien, Manchuria and is possibly an annectant species with Napomyza.
B. papyrifera Marsh. - North America The closer relationship of lap. viridula on Quer-
B. populifolia Marlsh. - North America cus with lap. salicifolii on Salix than with the
B. pubescens Ehrh. - North and Central Europe, second species on Quercus, lap. quercus, is evidence
Siberia of a puzzling switch between Quercus and Salix.
B. verrucosa Ehrh. - Europe, Asia Minor Two isolated species with unusual biology - Cecido-
Carpinus betulus L. - Europe, Asia Minor myiaceltis deserta on Celtis and Ophiomyia fici on
Corylus avellana L. - Europe, Asia Minor Ficus - indicate their early evolutionary isolation,
Ostrya virginiana (Mill.) Koch - North America and Liriomyza urticae on Urtica ferox in New Zea-
land is an isolated species of uncertain affinity.
With Pb. betulae known to feed on both Alnus
The obvious affinity of the Agromyza species on
and Betula, it has now been established that its host
Alnus and Betula (Betulaceae) with species on
range also includes Carpinus (confirmed from larva)
Rosaceae provides support for the recent proposal
and Corylus. Larval feeding was recently found on
by Thorne (1983) to include these families in ad-
these hosts in northern France and adults have now
jacent Superorders. On the other hand agromyzids
been reared (Martinez, pers. comm.).
provide no evidence of a closer relationship be-
Alnus and Betula each have their own host-spe-
tween the Urticales and Malvales cited by Hum-
cific Agromyza leaf miners. In the West Palaearctic
phries (1988).
A. alnivora Spencer (1969a) is restricted to Alnus
4 (1)
Table 10
Order CARYOPHYLLALES
Phytolaccaceae
Rivinia Liriomyza schmidti
Nyctaginaceae
Bougainvillea Tropicomyia flacourtiae; L. schmidti
Cactaceae
Opuntia gen., sp. indet.
Chenopodiaceae
Atriplex Ophiomyia atriplicis; Amauromyza luteiceps; L. bryoniae:
Chromatomyia horticola
Subclass CARYOPHYLLIDAE 63
Amaran thaceae
Achyranthes Pseudonapomyza alternantherae
Alternanthera M. marellii; O. alternantherae, O. buscki; Ps. alternantherae
Amaranthus M. amaranthi; Am. abnormalis; L. bryoniae. L. trifolii; Hap. minutus;
Chr. horticola
Celosia L. bryoniae. L. tri folii
Gomphrena Hap. gomphrenae nom. nud.
Pfaffia O. pfaffiae
Philoxerus Hap. philoxeri
Pupalia L. trifolii
gen. indet. L. flagellae nom. nud.
Portulacaceae
Portulaca Hap. palliatus
Basellaceae
Anredera L. schmidti
Basella L. trifolii
Caryophyllaceae
Agrostemma Am. flavifrons
Cerastium Am. flavifrons; L. betae; Chr. horticola
Dianthus Am. flavifrons; L. dianthicola, L. trifolii
Gypsophila Am. flavifrons; L. gypsophilae, L huidobrensis. L. trifolii; Chr. horticola
Lychnis Am. flavifrons; L. strigata
Melandrium O. melandricaulis, O. melandryi, Ophiomyia sp. n. ( Japan );
Am. flavifrons
Moehringia O. melandricaulis; Am. flavifrons
Saponaria Am. flavifrons; L bryoniae. L. strigata, L xanthocera
Silene Am. flavifrons; L. betae
Stellaria O. melandricaulis; Am. flavifrons; L. betae, L. strigata; Chr. horticola
Vaccaria Am. flavifrons; L. strigata
Order POLYGONALES
Polygonaceae
Polygon urn Japanagromyza polygoni; Agromyza pittodes, A. polygoni;
Chr. horticola
Rumex Am. monfalconensis
Ruprechtia Phytobia sp. ( larval feeding only)
64 Class MAGNOLIOPSIDA (Dicotyledons)
Order CARYOPHYLLALES (12 families, firm that this species is entirely typical of others in
10 000 species) the genus known in the northern hemisphere. It is
closely related to M. amaranthi (see below) recently
Eight of the 12 families in this order have been
discovered in Venezuela. The Chenopodiaceae are
colonized by Agromyzidae, the great majority, 20
believed to be directly ancestral to the Amaran-
species, on the two families Amaranthaceae and
thaceae and it is logical to assume that the switch of
Chenopodiaceae.
the ancestral species took place from Chenopodium
to Amaranthus.
Phytolaccaceae (18 genera, 125 species) Ophiomyia atriplicis Spencer and Hawkins
(1984) is an isolated species occurring on Atriplex
The Phytolaccaceae are generally agreed to be the
polycarpa, a plant adapted to the desert climate in
most archaic family in the order and it is widespread
eastern California near Palm Springs, on which the
in tropical and sUbtropical regions,especially in the
larvae form bud-galls and, with this specialised
New World.
adaptation to xerophytic conditions, are able to
Only the single polyphagous species, Liriomyza
survive the long, dry, hot summers.
schmidti, is known on Rivinia sp. in Argentina.
Amauromyza chenopodivora Spencer (1971b),
known only in Europe, was originally misidentified
Nyctaginaceae (30 genera, 300 species) as the North American Am. abnormalis which feeds
on Amaranthus (see below), until the genitalia (Fig.
It is now generally agreed that this family is closely
230) were examined. Am. chenopodivora was de-
allied to and probably derived from the Phytolacca-
scribed when the only male identified as Am. abnor-
ceae. Pollen referable to the Nyctaginaceae is known
malis was a specimen from Quebec and the genitalia
from early Eocene. of these two are obviously different. It can now be
Bougainvillea, native to South America, was
seen that the European species (Fig. 231) closely
recorded as a host of the polyphagous species Tropi- resembles a male discovered later in California. This
comyia flacourtiae (Seguy cf. Spencer, 1985a) in
complex represents a clear example of speciation
Madagascar and I have seen it frequently mined by
either side of Beringia.
the same species in the Nairobi area, Kenya. Lirio-
Am. flavifrons (Meigen, cf. Spencer, 1976a) is
myza schmidti (Aldrich) a highly polyphagous
occasionally found on Beta and Spinacia but its
species in South America has been recorded on
normal hosts are genera in the Caryophyllaceae (see
Bougainvillea spectabilis in Florida.
Table 10). It is apparently able to transfer without
difficulty from the Caryophyllaceae to the Cheno-
podiaceae.
Cactaceae (30-200 genera,
Amauromyza luteiceps (Hendel, cf. Spencer,
1000-2000 species)
197 6a) occurs in Europe as a stem borer in Atriplex.
The relationships of this family are still debated but The genitalia (Fig. 232) confirm its close relation-
'it is now generally believed that the Cactaceae take ship with Am. abnormalis on Amaranthus in North
their origin in or near the Phytolaccaceae' (Cron- America and Am. chenopodivora on Chenopodium
quist, 1981). in Europe.
No agromyzid has been described from this fami- Two Liriomyza species are known on Beta in
ly but I saw mines on Opuntia sp. with dead larvae Australia, L. caulophaga (Kleinschmidt, 1961, cf.
clearly of Agromyzidae near the entrance to the Spencer, 1977c) (Fig. 233) and the species hitherto
Zoological Institute at Manguifios, near Rio de known as L. chenopodii (Watt, cf. Spencer, 1976b
Janeiro, Brazil, 23.viii.1957. The species causing and 1977c), which also occurs in New Zealand. It
these mines was of uncertain genus but it will be of has recently been discovered that L. chenopodii
great interest to establish its identity and affiliations. represents a synonym of L. betae (Coquillett, 1900)
and this new synonymy is established in the Appen-
dix below. In both L. caulophaga and L. betae the
Chenopodiaceae (100 genera, 1500 species)
outer cross-vein is lacking.
This family is of cosmopolitan distribution but is Watt (1924) recorded Chenopodium album,
especially abundant in desert and semi-desert condi- white and silver beet, and in the Caryophyllaceae
tions. Many species are now widespread weeds. Cerastium vulgatum and Stellaria media as hosts of
Fifteen species of Agromyzidae are known on the L. betae. These are all introduced plants in New
family (see Table 10). Zealand and it is accepted that the fly has been
Melanagromyza chenopodii Spencer (1963c) was introduced from Australia where it is widespread. In
described from Chile, feeding in the stem of Cheno- the Perth Botanical Gardens I caught one fly and
podium ambrosioides. The genitalia (Fig. 229) con- found many mines on Silene gallica. The male gen-
Subclass CARYOPHYLLIDAE 65
/ t~'!:~
.....
.
·.. ···.V··
230
231
233 235
'. "1;;-
... -'-
_
///
, !
..
Fig. 230 . M
Fig. 229 A ean
I agromyza chenG
Fig. 231' A mauromyza
. . mauromy
abnor:a~dll: ..
aedeagus.
IS: aedea
F.Ig. 232 . A mauro za chenopod'Ivora' a d gus.
F Igs.233 234 L .myza luteiceps' aed . e eagus.
Fig. 235 'LI'" IrlOmyza caulop'h eagus.
rlOmyz a betae: aedeagus'
F"Ig. 236.• Urio aga' 233 , aedeagus; 234 .
myza flagellae ( ' ' epandnum
nom . nud). : aedeagus. .
# . "'- .
66 Class MAGNOLIOPSIDA (Dicotyledons)
~ ~
"
/. ~!<~:-==-
....----.-~ ..
.. :
237 238
'~~~~';=. ~
(
241
239
('.
___ _ r-'~: ....
"
-"~ ..
~~~'
245 243
244
~~\, \
\
246 247
Figs. 237, 238. Xeniomyza ilicitensis: 237, aedeagus, side view; 238, same, ventral view,
Figs. 239, 240. Melanagromyza amaranthi: 239, aedeagus, side view; 240, same, ventral view,
Figs. 241-243. Ophiomyia alternantherae: 241, aedeagus, side view; 242, same, ventral view; 243, posterior spiracles ofpuparium.
Figs. 244, 245. Melanagromyza marellii: 244, aedeagus; 245, galls on Alternanthera philoxeroides.
Figs. 246, 247. Ophiomyia buscki: 246, aedeagus; 247, posterior spiracles of puparium,
Subclass CARYOPHYLLIDAE 67
italia (Fig. 235) are entirely typical of Liriomyza and Amaranthaceae (65 genera, 900 species)
it must be assumed that the original host is an
This family is obviously allied to the Chen~po~i
endemic species of either Chenopodiaceae or
aceae and the two families stand side-by-sIde In
Caryophyllaceae.
virtually all systems of classification. ,
The only known host of L. caulophaga is. Be~a
Melanagromyza amaranthi Spencer and Havra-
vulgaris var. cicla, the larva feeding and pupatIng In
nek (1989) was found in stems of Amaranthus dub ius
the midrib or petiole. The male genitalia show this to
which grows commonly as a weed where tomatoes
be an isolated species of extreme interest. Klein-
are cultivated in Tachira Province, Venezuela. The
schmidt (1961: fig. 10) gave a general photographic
genitalia (Figs. 239, 240) show that the species is
illustration of the genitalia of the male allotype
closely related to M. chenopodii in Chile (see
reared from Beta, and more detailed drawings of the
above).
aedeagus are shown in Fig. 233. The long, paired
Two unusual species, accepted in Melanagromyza
process within the epan~rium. is visibl~ in. Klein-
by Spencer (1963c), are known from Argentina a?d
schmidt's illustration and IS agaIn shown In FIg. 234.
Brazil on Alternanthera philoxeroides. The specIes
This structure varies in form and length but is pre-
described as M. alternantherae forms leaf mines and
sent in all Liriomyza species and is probably
is a small black species, with ill-defined male gen-
homologous with the 'rod-like scler~tes' (Lan~fo~
italia. It cannot now properly be accommodated in
satze of Nowakowski, 1973) present In all specIes In
Melanagromyza and is provisionally transferred to
all subgenera of Cerodontha, a genus restricted to
Ophiomyia (see Appendix below) pending further
the mono cots (see below).
revision. The aberrant larval posterior spiracles are
Liriomyza flagellae (nomen nudum) has been
shown in Fig. 243 and the genitalia in Figs. 241,
recorded by Valladares in Argentina in an unpub-
242). M. marellii (Brethes, cf. Spencer, 1963c)
lished report (1986) as forming leaf mines on un-
forms internodal stem galls (Fig. 245) on the same
identified species of Chenopodiaceae and Amaran-
host, with very different genitalia (Fig. 244) a~d
thaceae. I have seen specimens, and with the unusual
larval spiracles. This species is temporarily left In
male genitalia (Fig. 236), the affiliations of this
Melanagromyza but is clearly an isolated species,
species will merit further study when it is formally
possibly deserving separate generic status but fur-
described.
ther revision is called for.
Valladares (loc. cit.) records Haplopeodes suae-
A further species on the same host, Ophiomyia
dae as a new species (nomen nudum) on Suaeda
buscki (Frost) forms external stem mines and is a
divaricata in Argentina. The formal description will
more typical Ophiomyia. It was described from
follow in due course.
Panama and has since been recorded in Argentina
Xeniomyza ilicitensis Meijere (cf. Hering, 1936)
and Brazil (Spencer, 1963c). The male genitalia and
was discovered in Spain by Hering mining the fine
larval posterior spiracles are shown in Figs. 246,
leaves of Suaeda jruticosa and has also been re-
247. These three species are not related and repre-
corded on Salicornia in the south of France. This is
sent three separate colonizations of Alternanthera.
a minute species, with reduced chaetotaxy (post-
Ophiomyia pfaffiae (Spencer, 1963c) has only
ocellar bristles lacking) and without the outer cross-
once been found in Venezuela, forming leaf mines
vein. The erection of a mono typic genus to accom-
and pupating in the leaf of Pfaffia iresinoides. The
modate this species was largely based on these
genitalia and leaf mine showing the puparium at the
reductions. Adaptation to such extreme halophytic
end of the mine are shown in Figs. 248, 249.
genera as Salicornia and Suaeda represents an inter-
In Argentina Liriomyza flagellae has been record-
esting switch from a host growing in a more normal
ed by Valladares on an unidentified species of
habitat and this has doubtless been in part respons-
Amaranthaceae and also on Chenopodiaceae (see
ible for the speciation which has occurred on the
above).
new, unusual host. Zlobin (1979) described a
Pseudonapomyza alternantherae (Seguy, 1951) is
second species in this genus as X. intermedia from
a further leaf miner on Alternanthera, which was
an inland area of the USSR and the unknown host
described from Madagascar on A. sessilis and has
was probably also a halophyte in a salt-marsh area.
subsequently been recorded in India and Thailand.
The genitalia of X. ilicitensis (Figs. 237, 238) indi-
Incipient speciation is detectable from the male
cate its close relationship with Liriomyza.
genitalia between the populations in Madagascar
Three polyphagous Liriomyza species and also
(Figs. 250, 251) on A lternanthera , and in India/
Chromatomyia horticola attack a number of genera,
Thailand (Figs. 252, 253) on the related genus
as shown in Table 10.
Achyranthes. This species was recently discussed by
Spencer (1986c).
Leaf miners in the genus Haplopeodes have also
68 Class MAGNOLIOPSIDA (Dicotyledons)
249
248
252
250
253
251
255
254
257 256
Figs. 248, 249. Ophiomyia pfaffiae: 248, aedeagus; 249, leaf mine on Pfaffia iresinoides.
Figs. 250-253. Pseudonapomyza alternantherae: 250, aedeagus, side view; 251, same, ventral view (holotype, Madagascar); 252, aedea-
gus, side view; 253, same, ventral view (Thailand).
Fig. 254. Ophiomyia melandryi: aedeagus.
Figs. 255, 256. Agromyza pittodes: 255, aedeagus; 256, leaf mine on Polygonum viviparum (Hering).
Fig. 257. Amauromyza monfalconensis: aedeagus.
Figs. 258, 259. Japanagromyza polygoni: 258, aedeagus, side view; 259, same, ventral view.
Subclass CARYOPHYLLIDAE 69
colonized Amaranthus, Chenopodium, Gomphrena stem mines. The genitalia of both species have a
and Philoxerus. H. minutus (Frost, cf. Spencer and short sclerotized process above the distiphallus
Steyskal, 1986) feeds on Amaranthus and Cheno- complex and it seems clear that there has been a
podium in Canada and the United States. single colonization, with speciation occurring prob-
H. gomphrenae Valladares, nomen nudum (1982, ably on different host genera and each with its own
unpublished Doctoral Thesis) was reared from ecological niche.
Gomphrena sp. at Cordoba, Argentina and I dis- Sasakawa (1961: 359) recorded 0. melandryi 'in
covered H. philoxeri on Philoxerus vermicularis at the pith of stem of Melandryum sp.' in Japan. How-
the edge of the Everglades in Florida (Spencer and ever, his illustration of the genitalia (1961: fig. 39d)
Stegmaier, 1973). Empty leaf mines I found on shows that this represents a further species (which
Amaranthus retroflexus on Guadeloupe in Febru- remains to be described), which has evolved in isola-
ary, 1987 probably represent H. minutus. tion in Japan from O. melandryi or its immediate
Three polyphagous species are known on the ancestor.
Amaranthaceae - Liriomyza bryoniae in Europe Amauromyza flavifrons (Meigen, cf. Spencer,
and L. trifolii in both Europe and Kenya, and Chro- 1976a) has been recorded on all genera in this fami-
matomyia horticola in Europe. ly known to be attacked by Agromyzidae, the larva
Amauromyza abnormalis (Malloch, cf. Spencer forming a conspicuous white blotch mine. Larval
and Steyskal, 1986) was reared from Amaranthus in feeding has also been noted not infrequently on
U.S.A. and is compared above with its sister-species Beta and Spinacia (Chenopodiaceae) and these
in Europe, Am. chenopodivora from Chenopodium. alternate hosts are obviously also fully acceptable.
In Ontario, eastern Canada Am. flavifrons has
been recorded on Dianthus and Saponaria but with
Portulacaceae (20 genera, 500 species)
these cultivated, introduced plants as its only known
This small family is cosmopolitan but is most devel- hosts, it seems probable that it has itself been intro-
oped in western North America and the Andes. duced relatively recently.
Only Haplopeodes palliatus (Coquillett, cf. Two Liriomyza species are known in Europe on
Spencer and Steyskal, 1986) is known as a leaf Dianthus and Gypsophila. L. dianthicola (Venturi,
miner on Portulaca sp. It is certainly directly related cf. Spencer, 1973a) has only been found on cultiva-
to the three species recorded above on Amarantha- ted Dianthus. It is an anomalous species, with the
ceae. third antennal segment slightly angulate and the cos-
ta ending shortly after the termination of vein r 4 + 5.
This has led to confusion as to its true generic posi-
Basellaceae (4 genera, 15-20 species)
tion. It was originally described in Pseudonapomyza,
The Basellaceae are generally considered to be later in Phytomyza by Seguy, transferred to Phyt-
related to but more advanced than the Portulaca- agromyza by Hering, to Paraphytomyza by Spencer
ceae. The family occurs mainly in tropical or sub- and finally to Liriomyza by von Tschirnhaus (1981:
tropical areas in the New World. 319) on the basis of the presence of the 'stridulating
Only two highly polyphagous species have been organ.' Von Tschirnhaus (loc.cit.) suggests that
found on this family, Liriomyza schmidti on Anre- L. dianthicola is related to a group of 35 Liriomyza
dera cordifolia in Argentina (Valladares) and L. tri- species feeding on a range of different hosts. How-
folii on Basella alba in Mauritius. ever, this group consists of a number of certainly un-
related species (mesnili now known to be a Phyto-
liriomyza) and its inclusion in this 'group' needs
Caryophyllaceae (75 genera, 2000 species)
reappraisal.
This is one of the largest families in the order but L. gypsophilae Beiger (1972b) is only known
has only five host-specific agromyzids, with two from the Tatra Mtns., Poland where it was reared
overlapping on the Chenopodiaceae. It is also at- from Gypsophila repens. It also has the third anten-
tacked by five polyphagous species. The Agromyzi- nal segment angulate and belongs to the group with
dae in four genera, which are known on 11 genera of eight pores on the posterior spiracular processes (as
the Caryophyllaceae, are shown in Table 10. in L. dianthicola). With these synapomorphies and
Two typical Ophiomyia species have colonized despite other differences, it seems probable that the
this family in Europe. 0. melandryi (Meijere, cf. two are monophyletic, representing a single original
Spencer, 1964c) (Fig. 254) feeds in the hollow stem colonization of the family. Cultivated Gypsophila is
of Melandrium, while its sister-species, 0. melandri- a favourite host of the two polyphagous pest species
caulis Hering (cf. Spencer, 1964c) has a wider host L. huidobrensis and L. trifolii in California and
range, feeding not only on Melandrium but on Colombia.
Moehringia and Stellaria, the larva forming external
70 Class MAGNOLIOPSIDA (Dicotyledons)
Order POL YGONALES Rumex spp., the larva feeding internally in the stem.
The genitalia (Fig. 257) show that this species is
This order consists of the single family Polygona-
closely related to the three species feeding in stems
ceae.
of Amaran~hus, Atriplex and Chenopodium (see
above) and It can be accepted that the four species
Polygonaceae (30 genera, 1000 species) are derived from a single ancestor on the Caryo-
phyllales. Other related species are known as stem-
This family occurs chiefly in North Temperate
borers, including Am. madrilena Spencer on
regions and pollen attributed to it dates from the
Phlomis (Lamiaceae, see below).
Palaeocene (Cronquist, 1981).
The Agromyzidae on the Polygonaceae show no
Only four described species of Agromyzidae are
relationship with species on the other families in-
known as host-specific on this family. The only poly-
cluded by Cronquist in the Caryophyllidae.
phagous species known on Polygonum is Chroma-
tomyia hortieola.
In general, each family colonized in the two orders
Phytobia is accepted as one of the most primitive
on which Agromyzidae are known, the Caryophyl-
genera in the Agromyzidae and larval feeding has
l~les and Polygonales, has its own host-specific spe-
been detected in the stem of the shrub, Rupreehtia
CIes. Two paIrs of related species are present on the
eorylifolia in Argentina (Suss and Muller-Stoll,
Chenopodiaceae and Amaranthaceae - Melanagro-
1980).
myza ehenopodii (Chile) and M. amaranthi (Vene-
Japanagromyza polygoni Spencer (1973c) was
zuela), and Amauromyza ehenopodivora (Europe)
discovered in Venezuela forming blotch mines on
and Am. abnormalis (North America). However,
Polygonum in the Andes near Merida. The male
thre~ species cross .family boundaries. Amauromyza
genitalia (Figs. 258, 259) and other characters of the
flavifrons w~ose pnmary hosts are in the Caryophyl-
epandrium suggest that this species is closely related
l~ceae also fmds Beta and Spinaeia in the Chenopo-
to the species feeding on Fabaceae (see below).
dlaceae acceptable - but interestingly is never found
Two Agromyza leaf miners are known on Poly-
on Amaranthaceae which is more closely related to
gonum: A. pittodes Hendel (d. Spencer, 1976a)
th~ . Chenopodiaceae than to the Caryophyllaceae.
feeds on P. viviparum and it appears to be a boreal-
Lznomyza betae whose primary hosts are in the
alpine species, occurring now in Norway and north-
Chenopodiaceae, also feeds on three genera in the
ern Sweden, with isolated populations in the Aus-
Caryophy!laceae but also avoids the Amarantha-
trian Alps and Tatra Mountains in Poland. The male
ceae. However, L. flagellae (nom. nud.) feeds on
genitalia are shown in Fig. 255, and the leaf mine,
Chenopodiaceae and Amaranthaceae but not on the
which is largely confined to the midrib area in Fig.
Caryophyllaceae.
256. A. polygoni Hering (1941) feeds on P. bistorta
On Polygonum, Japanagromyza polygoni is closely
and is only known from Mecklenburg, East Ger-
r~lated to species on Fabaceae, and Agromyza
many, forming an irregular linear-blotch mine, not
patodes and A. polygoni are related to species on
associated with the midrib (d. Hering, 1957b: fig.
Rosaceae. The four Amauromyza species are re-
483). The male genitalia are only slightly differenti-
lated to species on Lamiaceae and Scrophulariaceae
ated from those of A. pittodes. A close relationship
(see below). It is thus clear that in the past there has
with the A. potentillae group on Rosaceae is evident
been complex switching between hosts in different
from the genitalia of these two species.
subc~asses to produce the interesting assemblage of
Amauromyza monfalconensis (Strobl, ct. Hendel,
speCIes now known on the Caryophyllidae.
1931) occurs widely throughout Europe and its host
has recently been discovered (von Tschirnhaus) as
Subclass DILLENIIDAE
This subclass as defined by Cronquist (1981) con- Salicaceae, and these are recorded below. Only the
sists of 13 orders, 78 families and nearly 25,000 subclass Rosidae on present evidence has more
species. Cronquist writes: 'The orders that make up Phytobia species (see below).
the Dilleniidae evidently hang together as a natural It should be noted that the Salicaceae are in-
group .... It is perfectly clear that the Dilleniidae cluded here, a catkin-bearing family which was
take their origin in the Magnoliales .... The Theales included by earlier authors with amentiferous fami-
are the central group of Dilleniidae.... The Mal- lies discussed above in the Hamamelidae. There is
vales, Lecythidales, Violales, Capparales and Ne- still no agreed consensus among botanists on the
penthales all appear to have arisen from a common phylogenetic position of the Salicaceae and the
complex in the Theales .... The Salicales are an Agromyzidae feeding on this family provide little
amentiferous offshoot from the Violales .... Pollen indication of its affiliations, although the close rela-
that appears to represent the Dilleniidae dates from tionship of Agromyza species with those on Rosa-
about the beginning of the Upper Cretaceous.' ceae is detectable.
Sixty-one species of Agromyzidae have suc- The 14 agromyzid genera known on the subclass
ceeded in colonizing 26 families, most of which are are shown in Table 11 and the host genera with their
in only four orders. The subclass is thus generally agromyzid species are listed in Tables 14 and 15.
not highly acceptable. However, possibly as many as
22 further tree-feeding Phytobia species on 20
Order DILLENIALES (2 families, 400 species)
genera in 12 families have been identified from
larval 'pith-flecks' by Suss and Muller-Stoll (1980)
Paeoniaceae (1 genus, 300 species)
and Suss (pers. comm.), including a fossil species on
No Agromyzidae are known on this family but this
Table 11
is worth noting, since it was earlier included with the
Ranunculaceae (cf. Willis, 1957) and it was always a
Agromyzid genera on Dilleniidae puzzle to Prof. Hering that it had no leaf miners. He
suspected correctly, early in the 1960s when I dis-
Species restricted Polyphagous cussed this question with him, that Paeonia was
to subclass
probably included incorrectly in the Ranunculaceae
Palaeophytobia and this is now fully confirmed. Cronquist (1981:
Phytobia I' 301) writes: 'Paeonia has traditionally been referred
Hexomyza 6 to the Ranunculaceae, where it is sadly misplaced.'
Melanagromyza 3
Ophiomyia 4
Tropicomyia 2 5 Order THEALES (18 families, 3500 species)
Japanagromyza 1
Agromyza 5 Only two of the 18 families in this order have been
Liriomyza 8 6 colonized by Agromyzidae.
Galiomyza 3
Calycomyza 3
Paraphytomyza 7 Theaceae (40 genera, 600 species)
Chromatomyia 2
Phytomyza 3 Larval feeding of Phytobia has been noted in 2
genera by Suss and Muller-Stoll (1980) and Suss
49 12 (pers. comm.):
* 1. As many as 22 further species are known only from larval Camellia japonica L. - Japan, China
feeding. Cleyera japonica Thunb. - eastern Asia
2. The Liriomyza sp.n. on Marah in California is included.
3. L. brassicae occasionally transfers to Pisum (Fabaceae). Only a single leaf miner, Tropicomyia theae
72 Class MAGNOLIOPSIDA (Dicotyledons)
(Cotes, ct. Spencer, 1973a) is known on this family Suss (loc. cit.) has detected larval feeding of Phy-
and it is considered as a minor pest of tea in Sri tobia on 3 genera of this family:
Lanka. Like other species in the genus, the larvae
form shallow, silvery epidermal mines and positive Luehea divaricata Mart. et Zucco - Brazil, Argentina
identification of adults is only possible from the L. grandiflora Mart. - Brazil, Paraguay, Argentina
male genitalia (Figs. 260, 261). Sloanea sigun K. Schum. - Java
Tilia americana L. - North America
I am now satisfied that this is a polyphagous
species and represents a senior synonym of Tr. cof- These records must represent 3 distinct species.
feae (see under Rubiaceae below) and this new
synonymy is established in the Appendix below. It is Only the single agromyzid, Ophiomyia tiliae (Cou-
now clear that the genus Tropicomyia needs further den, ct. Spencer and Steyskal, 1986) has been
revision. described on this family, where it forms character-
istic twig galls on Tilia americana (Fig. 264). This
Guttiferae (50 genera, 1200 species) led to the species being earlier included in the gall-
causing genus Hexomyza. However, the male gen-
Suss (loc. cit.) has detected larval feeding of Phyto- italia (Figs. 262, 263) associate this species with
bia in 2 genera: 0. abutilivora (see below) and it appears that the
Endodesmia callophylloides Benth. - West Africa differing reaction of the host produces gall-like stem
Symphonia globulifera L.f. - West Africa and Brazil mines on Abutilon but true, oval galls on Tilia.
260 261
262
264
~.d~~~~~~
r.
0;
265
-_." .---
<&st;fJj ;;0
- - ---
267 \'; 268
Figs. 260, 261. Tropicomyia theae: 260, aedeagus, side view; 261, same, ventral view.
Figs. 262-264. Ophiomyia tiliae: 262, aedeagus, side view; 263, same ventral view; 264, galls on Tilia americana.
Figs. 265, 266. Hexomyza abutilonicaulis: 265, aedeagus, side view; 266, same, ventral view.
Fig. 267. Melanagromyza hicksi: aedeagus.
Fig. 268. Melanagromyza hibisci: aedeagus.
74 Class MAGNOLIOPSIDA (Dicotyledons)
270 269
271
(~.
·\~;:0,,;".~~~$~.~:~; ~::y
272 273
274
276
~~.
" .
277
Subclass DILLENIIDAE 75
similarity (M. hicksi: Fig. 267 and M. hibisci: Fig. C. sidae Spencer (in Spencer and Stegmaier, 1973).
268) but the two cannot be directly related. The genitalia of these two species are very similar
Ophiomyia abutilivora Spencer (in Spencer and but the leaf mines are distinctively different. C. mal-
Steyskal, 1986), which is widespread in eastern vae forms a long, narrow linear mine (Fig. 274, on
U.S.A. and has also been recorded in Chile (Spen- Althaea rosea, Toronto), while C. sidae forms a
cer, 1982), forms unusual stem-mines which are short linear-blotch mine (Fig. 275). A further spe-
raised to form a 'welt' and thus show incipient gall cies, C. longicauda (Blanchard, d. Spencer, 196 3c),
formation. The posterior spiracles of the puparium is known only in Argentina (Buenos Aires, Tucu-
(Fig. 269) are typical of Ophiomyia but the male man), forming blotch mines on several Abutilon
genitalia (Figs. 270, 271) indicate relationship with a species. The distinctive aedeagus (Fig. 276) gives no
small group of species of which the host is unknown indication of the affiliations of this species.
(Spencer and Steyskal, 1986: 251), and also with The Malvaceae are clearly highly acceptable to
0. tiliae discussed above under Tiliaceae. polyphagous species. Tropicomyia flacourtiae is
Only a single Tropicomyia is known on the known on Gossypium in Nigeria; of 5 Liriomyza
family, reared from Gossypium in Nigeria. This was species, L. strigata and L. xanthocera (= cruciferi-
originally treated as a distinct species T gossypii cola) (Europe) are each known on Althaea, Lava-
Spencer (1959) but was later synonymised with the tera and Malva; L. sativae has been recorded on
polyphagous T flacourtiae. Abelmoschus, Malva and Sida in the U.S.; and
Two unrelated Agromyza species have colonized L. trifolii is known on Gossypium, Hibiscus and
Abutilon and Malva. A unique species, Agromyza Malva. Chromatomyia horticola is known on seven
brunnicosa Becker (cf. Spencer, 1965a), forms leaf genera (cf. Table 15).
mines on Malva neglecta in the Canary Islands
where it is apparently endemic. The male genitalia
Order LECYTHIDALES
give no indication of its affiliations and its long
isolation obscures its relationships. A. abutilonis This order consists of the single family Lecythida-
Spencer (1959) is known in Kenya and Uganda, ceae.
forming linear-blotch mines on Abutilon and Hibis-
cus. The male genitalia of the allotype from Uganda
Lecythidaceae (20 genera, 400 species)
are shown in Figs. 272, 273. These are puzzling, as
the distinctive form is characteristic of many species The family is confined to tropical regions and is best
feeding on Poaceae (cf. Figs. 1339, 1341). There is developed in rain forests.
no doubt about the host of this species. I found Suss (loc. cit.) has recorded larval feeding of
mines on Abutilon in Kenya and recently many Phytobia in Brazil on Lecythis ollaria L.
specimens have been reared at Nairobi (B. Ten- The only leaf miners known here are two of the
gecho). It appears that a unique switch has occurred highly polyphagous Tropicomyia species - T poly-
from Poaceae to Abutilon which is difficult to phaga Spencer (1961b) on Barringtonia acutangula
explain apart from highly random colonization. (included by some authors in a separate family
Agromyza malvaceivora Seguy (1951) was de- Barringtoniaceae) which I found at Bombay in
scribed from Madagascar from adults reared from January, 1960, and T polyphyta (Kleinschmidt)
an unidentified species of Malvaceae. Only three known in Queensland, Australia on B. gracilis.
females are available and they appear virtually
identical to A. abutilonis and the mines of the two
Order VIOLALES (24 families, 5000 species)
species are similar. The linear-blotch mines of
A. malvaceivora were illustrated by Paulian (1953). Cronquist (1981: 378) considers that 'the bulk of
I suspect that the two species may be identical but a the order hangs together as a natural group with the
male from Madagascar will be required before this Flacourtiaceae at or near its evolutionary base.' An
can be confirmed. important difference in Thorne's (1983) treatment is
In North America and the Neotropical Region that the Salicaceae are included within the order
two leaf miners occur commonly on Malva and Violales, not as a separate order as by Cronquist.
Sida, Calycomyza malvae (Burgess) (= C. althaeae The assemblage of Agromyzidae on these two
Spencer (1969b) described from Canada) and groups is totally distinct, with no indication of close
Figs.269-271. Ophiomyia abutilivora: 269, posterior spiracles of puparium; 270, aedeagus, side view; 271, same, ventral view.
Figs. 272, 273. Agromyza abutilonis: 272, aedeagus, side view; 273, same, ventral view.
Fig. 274. Calycomyza malvae: leaf mine on Althaea rosea.
Fig. 275. Calcycomyza sidae: leaf mine on Sida sp.
Fig. 276. Calycomyza longicauda: aedeagus.
Figs. 277, 278. Tropicomyiaflacourtiae: 278, aedeagus, side view; 279, same, ventral view.
76 Class MAGNOLIOPSIDA (Dicotyledons)
279 282
281 284
',",
. \'\
~"....-...----':>
285 286
relationship and this evidence, however slight, sup- (named after his daughter Takako). A third species
ports Cronquist's treatment. was discovered in the United States in Maryland,
Of the 24 families included here eight have been Mississippi and Pennsylvania, and was described as
colonized by 18 species of Agromyzidae in four Galiomyza violivora (Spencer and Steyskal, 1986:
genera (see Table 15). These represent a mixed 298).
assortment, with the only possible radiation on the An isolated species on Galium, of uncertain
order apparent in the three species of Tropicomyia. generic affiliation, morio (Brischke, 1880) was
accepted in Liriomyza until von Tschirnhaus (1971:
563) recorded that this species lacks the stridulating
Flacourtiaceae (85 genera, 800 species)
organ which has been accepted as a definitive gen-
Suss (loc. cit.) has detected larval feeding of Phy- eric character of Liriomyza. Von Tschirnhaus re-
tobia in three species of Case aria as follows: moved morio from Liriomyza but avoided the dif-
ficult problem of its generic affiliation, simply leav-
Casearia glomerata Roxb. et Boupe. - Hidalgo,
ing it unplaced generically. After discovering further
Michoacan, Mexico
species on Galium in Canada and California, I
C. greviaefolia Vent. - Java
erected for this group the new genus Galiomyza
C. rhopalaefolia Krst. - (locality?)
(Spencer, 1981: 288). The new species on Viola in
Tropicomyiaflacourtiae (Seguy, 1951, as Melana- America could not be accommodated in Liriomyza
gromyza) was described from Madagascar on Fla- and generally resembled the group of Galium
courtia ramontchi and is known on Flacourtia sp. in feeders. To avoid undesirable splitting, I therefore
East Africa. I also found mines on Dovyalis caffra at included both violivora and the European violi-
Nairobi in November, 1961. The male genitalia of phaga in Galiomyza.
T. flacourtiae, type of the genus Tropicomyia, are From the excellent illustrations of Liriomyza
shown in Figs. 277, 278. This is a highly poly- takakoae given by Sasakawa (1954c and 1961), I
phagous species occurring on 21 families in four am now satisfied that this Viola miner is mono-
subclasses (see Table 92). phyletic with violiphaga and violivora and is formal-
At Krishnagar, India I found mines on Flacourtia ly transferred to Galiomyza in the Appendix below.
sp. in January, 1960 and this species I accept as The aedeagi of G. violiphaga, violivora and taka-
T. polyphaga (Spencer, 1961a). koae are shown in Figs. 279, 280, 282. All three
species lack discrete surstyli and the epandrium of
takakoae is shown in Fig. 283. Also, all three
Cistaceae (8 genera, 200 species)
species form whitish blotch mines - G. violivora
No species have been described from this family but (Fig. 281) and G. takakoae (Fig. 284).
Suss (loc. cit.) has detected larval feeding of Phy- I suggest that this monophyletic group on Viola is
tobia on 4 species of Cistus as follows: derived from a single ancestor possibly in eastern
Asia, with speciation in the three isolated popula-
Cistus crispus L. - western Mediterranean
tions in Western Europe, Japan and North America.
C. ladaniferus L. - southern France, Spain, Por-
Galiomyza can now be seen as an inappropriate
tugal, Algeria, Morocco
generic name for the Viola feeders and with further
C. monspeliensis L. - Mediterranean area
more detailed revisionary studies with emphasis on
C. salviaefolius L. - Mediterranean area
the host association, it is possible that in due course
the Galium and Viola feeders will be treated as each
Violaceae (16 genera, 800 species) deserving separate generic status.
Liriomyza violicaulis Hering (1962) is known
This family is generally resistant to attack by
only in Germany feeding on Viola tricolor arvensis,
phytophagous insects and it has been colonized by
initially as a leaf miner but mainly in the stem.
only family-specific agromyzids and also the poly-
Examination of the holotype shows that this is not
phagous Chromatomyia horticola. Hering (1926:
related to the three leaf miners on Viola discussed
467) discovered the first leaf miner, on Viola silva-
above and is a true Liriomyza. The male genitalia
tica in the Swiss Alps, which was later described by
are shown in Figs. 285, 286. L. flavalateralis Spen-
Hendel (1932) as Liriomyza violiphaga. For reasons
cer (1976b) is another typical Liriomyza in New
which are not clear Hering (1957b: 1116) without
Zealand (Figs. 287, 288), the larva forming a broad,
comment placed this species in Metopomyza and
white linear mine on the tree Melicytus ramiflarus
this generic transfer was accepted by Nowakowski
(whitey-wood) which can reach a height of 10m.
(1962: 153) and by Spencer (1971b: 163). Sasa-
kawa (1954c: 116) described a second species in
Japan on Viola nipponica as Liriomyza takakoae
78 Class MAGNOLIOPSIDA (Dicotyledons)
Turneraceae (8 genera, 120 species) is n?t reflected. in the host selection of the agro-
myzIds. Only mne described species are known on
The highly polyphagous species, Liriomyza trifolii,
the family, all but one polyphagous, and three each
has been recorded in Florida on Piriqueta carolini-
cross three tribal boundaries. These agromyzids are
ana.
thus not aware of these tribal divisions. Here there-
fore I find it adequate and simpler to treat the 12
Passifloraceae (16 genera, 650 species) genera known to be colonized in alphabetical order.
Despite the small number of host-specific agro-
Passiflora spp. are frequent hosts of Tropicomyia
myzids on the Cucurbitaceae all the cultivated
spp. in Africa, India, Papua New Guinea and
genera are frequently attacked in pest proportions
Australia, and there is a recent record from Fiji.
by polyphagous Liriomyza species.
T passiflo~ae (Hering, 1957a, as Melanagromyza)
Epidermal-mining Tropicomyia species are
was descnbed from South Africa on Passiflora
known in India, Papua New Guinea and Samoa.
coerulea; this species was synonymised with T fla-
T momordicae (Sasakawa, 1963b) is the one de-
court!ae (Seguy) by Spencer (1973a). It may be
scribed species in this genus, reared from mines 'in
mentIOned here that the Passifloraceae 'are widely
leaf of spiny melon vine, Momordica sp.' in Western
regarded as being related to and derived from the
Samoa. The illustration of the male genitalia (Sasa-
Flacourtiaceae' (Cronquist, 1981). The synonymy of
kawa, 1963b: fig. 3) confirms that this is a distinct
.T passiflorae with T flacourtiae is thus not surpris-
species.
mg.
In Papua New Guinea I found mines of a Tropico-
In India 'Melanagromyza atomella' (Malloch, d.
m~ia. on Melothria sp. in the Eastern Highlands
Spencer, 1973a) has been recorded on Passiflora
DIstnct, 2350 m. The mines were mainly charac-
odorata, while in Australia a number of Passiflora
teristically epidermal but in places substantially
spp. are attacked by T polyphyta. Mines on Passi-
deeper and it is probable that this represents an
flora foetida I found in Papua New Guinea, identi-
undescribed species. T atomella was recorded in
fied as T theae, now accepted as a polyphagous spe-
India from Coccinia (as Cephalandra indica) but the
cies in eastern Asia and the western Pacific. Speci-
true identity of this species is uncertain; the record
mens recently seen from Fiji from Passiflora sp. are
of Melanagromyza centrosematis de Meij. on the
tentatively identified as T momordicae (Sasakawa,
same host, also in India (Singh and Ipe, 1973: 272)
1963b), based on the similarity of the male genitalia.
~epresents a certain misidentification, as this species
IS now accepted as an Ophiomyia feeding exclusive-
Achariaceae (3 monotypic genera) ly as a stem miner on Fabaceae.
Lirior:zyza . bryoniae (Kaltenbach, d. Spencer,
Tropicomyia ceratiosicyi (Hering, 1957a), a leaf
1973a) IS a highly polyphagous species and occurs
miner on Ceratiosicyos ecklonii in South Africa is
commonly as a significant pest on several genera of
the only species recorded on this family. It is kno~n
~ultiv~t~d cucurbits in Europe but is rarely found on
from a single female and its affiliations cannot be
ItS ongmal host, Bryonia. The male genitalia are
further clarified. However, in view of the close rela-
shown in Figs. 289, 290, and the larval posterior
tionship between this small family and the Passi-
spiracles in Fig. 291. With about ten spiracular
floraceae it seems likely that T ceratiosicyi may
pores on each process and the generally similar
merely represent T flacourtiae (see Passifloraceae
genitalia, L. bryoniae is closely related to L. strigata,
above).
another polyphagous species which occurs most
commonly on Asteraceae (see below). However the
Cucurbitaceae (90 genera, 700 species) le~f mines of ~he two species are entirely diffe;ent,
WIth L. bryomae forming an irregular linear mine on
Fossil leaves are known in the Palaeocene but pollen
any part of the leaf (Fig. 292), while the mine of
has . no~ been found earlier than Oligocene. The
L. strigata always follows the midrib, with short
fanuly IS largely tropical but is well known in tem-
offshoots into the leaf blade. L. cocculi, another
perate areas from the widespread cultivation of such
polYl?hagous species described from Cocculus (see
genera as Citrullus, Cucurbita and Cucumis.
Memspermaceae, Ranunculales above) in Hawaii
There is some dispute about the affiliations of the
and L. hui~obrensis (see under Fabaceae below)
family but I see no reason not to retain it in the
~elo~g to t~IS same group. It seems clear that specia-
Violales, following Cronquist (1981). From an
tIon m thIS group has occurred in the ancestral
earlier ?ivision into five tribes, more recently the
popUlations following isolation each side of Ber-
CucurbItaceae have been further revised and are
ingia.
now split i~to n~n~ tribes (Mabberley, 1987, citing
L. strigata and the 3 polyphagous N ew World
Jeffrey). ThIS sphttmg, however justified botanically,
species L. trifolii, L. sativae and L. huidobrensis,
Subclass DILLENIIDAE 79
~
-........
289
• ',4
290
293
297
295 296
299
Figs. 289-292. Liriomyza bryoniae: 289, aedeagus, side view; 290, distiphallus, ventral view; 291, posterior spiracles of larva; 292, leaf
mine on Cucumis melo.
Figs. 293, 294. Phytobia cambii: 293, aedeagus; 294, larva.
Fig. 295. Hexomyza schineri: galls on Populus tremuloides.
Fig. 296. Hexomyza simplicoides: anterior end of puparium.
Fig. 297. Hexomyza winnemanae: posterior spiracles ofpuparium.
Figs. 298, 299. Japanagromyza salidfolii: 298, aedeagus, 299, epandrium, side view.
Subclass DILLENIIDAE 81
S. hookeriana Baratt - North America 298) and the epandrium with the conspicuously long
S. laevigata Bebb - California and narrow surstyli (Fig. 299) closely resemble
S. nigra Marsh. - North America those of 1. viridula. The two species are obviously
S. nuttallii Sargent - North America related and there must at some time have been a
S. purpurea L. - Europe, North Africa, central Asia, switch from Fagaceae to Salicaceae, with subse-
Japan quent speciation, and further speciation of the
S. rigida Miihlenb. - North America Quercus-feeders producing 1. quercus in Japan.
S. triandra L. - Europe, Siberia, China, Japan Three Agromyza leaf miners are found on the
S. viminalis L. - central Europe, Caucasus, Asia, Salicaceae, two in Europe and one in North Ameri-
Siberia, Himalayas, China, Japan ca. A. albitarsis Meigen (cf. Spencer, 1976a) occurs
commo~ly in Europe on Populus, and it was origi-
These records must represent one, possibly two
nally belIeved that the species on Salix, described as
further species.
A. lygophaga Hering (1937) was distinct but this
The gall-causing genus Hexomyza is represented
was synonymised with albitarsis by Spencer
by five species, with only H. schineri (Giraud, cf.
(1969a). A. albitarsis has also been recorded in
Spencer, 1976a) on Populus. This species was de-
British Columbia. A. isolata Malloch (cf. Spencer
scribed in western Europe but is also widespread in
and Ste~skal, 1986) (= A. populoides Spencer,
Canada from New Brunswick to British Columbia
1969b) IS a related species occurring on Populus
on P. tremuloides and in the United States from
and Sa.lix. in Canada and U.SA. The two species
Washington State south to New Mexico on other
form sImIlar upper surface blotch mines and the
Populus spp. The typical oval twig galls are shown in
posterior spiracles of the larvae have the normal
Fig. 295. The distinctive posterior spiracles of the
three pores. The male genitalia (A. albitarsis, Fig.
puparium are on short, raised stalks bearing three
300; A. isolata, Fig. 301) confirm their close rela-
spiracular pores (differentiating the genus from
tionship, and also suggest relationship with the
Melanagromyza, cf. Fig. 211). It seems clear that this
A. po~e~tillae grou~ on Rosaceae (cf. Fig. 366».
species was present in the Arcto-Tertiary forest
A. sallcma Hendel IS only known in northern Eu-
extending from Europe across Beringia to North
rope on Salix and is differentiated in various charac-
America. Two of the species on Salix occur in
ters fr:>m A. albitarsis. The leaf mine is produced
Europe and two in North America. Of the former
exclusIvely on the lower leaf surface (Fig. 303) and
H. simplicoides (Hendel, 1931) is distinctly larger
the larval posterior spiracles each have 8-10 elon-
than H. cecidogena (Hering, 1927c) and the two
gateyo:es. The male genitalia (Fig. 302) neverthe-
have generally similar genitalia; however, the larva
less mdicate general relationship with A. albitarsis.
of simplicoides is strongly differentiated, with the
.In E~rope a closely related group of five leaf
front segments reddish-brown and conspicuously
mmers m the genus Paraphytomyza feed on Salica-
sloping ventrally (Fig. 296). In North America
ce~e which has been referred to as the populi group.
!f. salicis (Malloch, cf. Spencer and Steyskal, 1986) It IS generally agreed that these species are not
IS possibly the sister-species of H. cecidogena, with
mon.ophylet!c with other species known in the genus
slight but distinctive differences in genitalia. H. win-
feedmg mamly on eight families in the Subclass
nemanae (Malloch, cf. Spencer and Steyskal, 1986)
Asteridae. There ~as been hesitation about splitting
(= H. albicula Spencer, 1969b), described from
the genus to provIde separate generic status for the
~arylan? and also known in Ontario, is most clearly
p~puli group until the affiliations of more species
dIfferentIated by the pale, whitish wings, pale squa-
wIth known hosts have been further clarified (Spen-
mae and in the larva (puparium) by the posterior
cer, 1987). Three species feed on Populus, P. populi
spiracular processes being contiguous (Fig. 297).
(Kaltenbach), P. populicola (Walker) and P. tremu-
The genitalia nevertheless indicate its close rela-
lae (He~ng), and on Salix, P. fulvicornis (Hendel)
tionship with the other Salix feeders.
and P. tndentata(Loew) - all discussed by Spencer
lapanagromyza salicifolii (Collin, 1911) occurs
(1976a). The most closely related pair of species are
on both Salix and Populus. It was described from
probably P. populicola and P. tremulae, having in
Egypt and has since been found in the Canary
co~on a striking synapomorphy with all body
Islands and is present throughout the Mediter-
bnstles yellow. The male genitalia of P. populicola
ranean area from Portugal to Israel and Turkey,
are shown in Fig. 304 and P. tremulae in Fig. 306.
with its range extending eastwards to Kirghizia in
The larval leaf mines of P. tremulae are unique in
the U.S.S.R. The majority of species in this genus
the group in occurring exclusively in the spongy
feed on herbaceous plants, particularly in the Faba-
parenchyma on the lower leaf surface of Populus
ceae, but two species on Quercus in U.S.A. and
tremula (Fig. 307), while in populicola the larva
Japan were discussed earlier (Subclass Hamameli-
fo~ms upper surface blotch mines (Fig. 305). P. pop-
dae). Both the male genitalia of 1. salicifolii (Fig.
ullcola appeared for the first time in Canada in the
82 Class MAGNOLIOPSIDA (Dicotyledons)
300 301
~.:-~~.;:.~ ........ ,
~c:;;>' .•'''';'::::::;7'~
302
303
"C~~~UJ
309
312 313
314 315
317
319
Figs. 308, 309. Melanagromyza cleomae: 308, aedeagus, side view; 309, same, ventral view.
Figs. 310, 311. Ophiomyia alliariae: 310, head; 311, aedeagus.
Figs. 312, 313. Ophiomyia texana: 312, head; 313, aedeagus.
Figs. 314, 315. Liriomyza alyssi: 314, aedeagus, side view; 315, distiphallus, ventral view.
Figs. 316, 317. Liriomyza pseudopygmina: 316, aedeagus, side view; 317, distiphallus, ventral view.
Figs. 318, 319. Liriomyza xanthocera: 318, aedeagus, side view; same, ventral view.
Subclass DILLENIIDAE 85
Table 14
Capparaceae
Cleome Melanagromyza c1eomae; Liriomyza brassicae, L. bryoniae L. strigata;
Chromatomyia horticola
Gynandropsis L. brassicae, L. bryoniae
Brassicaceae ( Cruciferae )
Aethionema L. xanthocera '
Alliaria Ophiomyia alliariae; L. strigata, L. xanthocera; Chr. horticola;
Phytomyza rufipes
Alyssum L. alyssi, L. strigata; Chr. horticola; P. alyssi
Arabis L. xanthocera; Chr. horticola
Arabidopsis L. strigata; Chr. horticola
Armoracia L. xanthocera
Barbarea Ophiomyia sp.; L. brassicae, L. strigata, L. xanthocera
Bertorea L. xanthocera; Chr. horticola
Biscutella L. xanthocera; Chr. horticola
Brassica Mel. c1eomae; L. brassicae, L. xanthocera; Paraphytomyza mamonowi;
Chr. horticola; P. rufipes
Brassicella Ophiomyia sp. ( N. France ); L. strigata; Chr. horticola
( = Rhynchosinapis )
Bunias L. strigata, L. xanthocera; Chr. horticola
Cakile L. brassicae; Chr. horticola
Callepina L. strigata, L. xanthocera; Chr. horticola
Carnelina L. xanthocera
Capsella Ophiomyia sp. ( East Germany); L. strigata, L. xanthocera; Chr. horticola
Cardamine o. alliariae; L. cardamines, L. strigata; Chr. horticola
Cardaria L. strigata; Chr. horticola
Cheiranthus L. brassicae, L. strigata, L. xanthocera; Chr. horticola
Cochlearia L. strigata, L. xanthocera; Chr. horticola
Conringia L. strigata, L. xanthocera; Chr. horticola;P. rufipes
Coronopus L. xanthocera; Chr. horticola
Cram be L. strigata, L. xanthocera; Chr. horticola
Descurainia O. texana; Chr. horticola
Diplotaxis L. strigata, L. xanthocera; Chr. horticola; P. rufipes
Eruca L. xanthocera
Erucastrum L. xanthocera; Chr. horticola
Erysimum L. strigata, L. xanthocera; Chr. horticola
Hesperis L. strigata, L. xanthocera; Chr. horticola
Hirschfeldia L. brassicae, L. strigata; Chr. horticola
Iberis L. strigata, L. xanthocera; Chr. horticola
Isatis L. brassicae, L. strigata, L. xanthocera; Chr. horticola
Lepidium Ophiomyia sp. ( Corsica ); L. brassicae, L. lepidii, L. strigata, L. xanthocera;
Chr. horticola
Malcolmia L. strigata
Matthiola L. brassicae, L. strigata; Chr. horticola
Moricandia L. brassicae, L. strigata; Chr. horticola; P. rufipes
Myagrum L. strigata, L. xanthocera; Chr. horticola
Neslia L. xanthocera
Peltaria L. strigata, L. xanthocera; Chr. horticola; P. rufipes
Raphanus Ophiomyia sp. ( East. Germany); L. brassicae, L. strigata, L. xanthocera;
Chr. horticola; P. rufipes
Rorippa O. texana; Chr. horticola; P. rufipes
Sinapis Ophiomyia sp.; L. brassicae, L. strigata, L. xanthocera;
Paraphytomyza mamonowi; Chr. horticola; P. rufipes
Sisymbrium L. brassicae, L. bryoniae, L. pseudopygmina, L. strigata, L. xanthocera;
Para. sisymbrii; Chr. horticola; P. rufipes
Thelypodium Tropicomyia flacourtiae ( ? )
Thlaspi L. strigata, L. xanthocera; Chr. horticola
Turritis L. strigata, Chr. horti cola
Moringaceae
Moringa L. sativae
Resedaceae
Reseda L. brassicae, L. strigata, L. xanthocera; Chr. horticola
pygmina (Hering, 1933) (= L. sisymbricaulis L. xanthocera. The third species in this complex,
Hering, 1962) (Figs. 316, 317) and L. xanthocera Liriomyza alyssi (Hering, 1960) was described from
(Czerny = crucifericola Hering, 19 51 b) which is Alyssum in Thuringia, East Germany. The genitalia
polyphagous), are a monophyletic group, unusual in of the holotype confirm that this species is close to
Liriomyza in having the scutellum black. L. xantho- but distinct from L. pseudopygmina. The larva in the
cera (as crucifericola) was described from Isatis only population known forms initial mines in small
tinctoria at the Botanical Gardens, Rostock, East upper leaves but then continues feeding in the stem.
Germany. The genitalia have not previously been In view of the close relationship of L. xanthocera,
illustrated and their examination from a paratype L. pseudopygmina and L. alyssi it seems certain that
surprisingly showed that they are identical to those they are derived from a single original colonization.
of L. xanthocera, a widespread species with most With the scutellum black and larval feeding in the
records from Spain and the Mediterranean. The stem (pseudopygmina), in the leaf and stem (alyssi)
aedeagus of L. cisti (Spencer) which I caught on and in the midrib (xanthocera) these are clearly
Cistus at Barcelona and which was synonymised relatively primitive species. With acrostichals pre-
with L. xanthocen (Spencer, 1977b) is shown in sent in L. alyssi and pseudopygmina but entirely
Figs. 318, 319. lacking in L. xanthocera, this suggests that the latter
L. xanthocera (as crucifericola) has been found may be the derived species.
on 31 genera, mainly in the Rostock Botanical In Japan Liriomyza cardamines Sasakawa (cf.
Gardens. It is highly polyphagous and has been Sasakawa, 1961) has been recorded on Cardamine
recorded on 12 families (see Table 92 and Buhr, scutata. The male genitalia show that this species is
1953). Surprisingly, a particularly frequent host is not related to any of the other five feeding primarily
Vida faba (Fabaceae). Quite unexpectedly this on the family. The leaf mine (Fig. 320) is strictly
species has also been found on two genera, Adonis associated with the midrib but without the frequent
and Ranunculus, in the Ranunculaceae. I briefly lateral offshoots found in L. strigata.
discussed its presence on these two genera under In New Zealand Liriomyza lepidii Harrison (cf.
Ranunculaceae (Magnoliidae). The larva feeds with- Spencer, 1976b) has been reared from Lepidium
in the midrib, with offshoots into the leaf blade, oleraceum. The male genitalia (Figs. 321, 322) indi-
which is probably a more primitive type of feeding cate relationship with the group of feeders on Hebe
than true leaf mining. (Scrophulariaceae), particularly with L. f!avocentra-
Phytobia (Cephalomyza) sisymbricaulis (Hering, lis (Watt) (cf. Figs. 287, 288). L. lepidii is unique
1962), which forms external stem mines on Sisym- among the Liriomyza species known in New Zea-
brium was described from Berlin. It correctly be- land in having the mesonotum partially yellow, with
longs in Liriomyza despite the dark scutellum. This distinctive banding (Fig. 323). This is a striking
species was synonymised with L. pseudopygmina by example of the subordinate significance of colour as
Nowakowski (1967: 652) and the male genitalia an indicator of relationship, as the genitalia clearly
(Figs. 316, 317) show the close relationship with confirm that this species is closely related to the
Subclass DILLENIIDAE 87
321
322
~/,
324 \\ / 323
325 ,....................,..
'.
....
other Hebe feeders, in which the mesonotum is host is given as Sinapis and the same species was re-
uniformly black. ported as occurring on Chenopodium but the mines
Liriomyza brassicae (Riley, ct. Spencer and Steys- on this host sent to Hering by Mamonow were of an
kal, 1986) was described in the United States on unidentified anthomyiid. Only two females are
Brassica (cabbage) and is now known on 15 further known and the species cannot be further clarified.
genera. It is semi-cosmopolitan but in Europe is Hering (195 7b) cites Brassica as a further host.
only known in Botanical Gardens. It is present Only two Phytomyza species are known on this
throughout the Old World tropics, particularly on family. P. alyssi Nowakowski (1975) has been found
Capparaceae and Tropaeolum (Tropaeolaceae), a only in Poland, feeding on Alyssum saxatile. The
favourite host, on which it occurs commonly in complex genitalia (Fig. 327) give no indication of its
Australia, including Tasmania. It occurs occasional- relationship. The larvae form an upper surface
lyon Resedaceae and an unusual switch is repre- blotch mine. P. rufipes Meigen (ct. Spencer, 1976a)
sented by its occasional occurrence on Pisum, on is a common pest of Brassica spp. in Europe and has
which it has been found on several occasions in been recorded on eight further genera. I recently
Kenya and India. The male genitalia (Figs. 324, 325) discovered a well-established population of P. rufi-
suggest possible relationship with L. baccharidis pes in the agricultural area around Bogota, Colom-
(Asteraceae) (Figs. 1059, 1060). Although L. bras- bia, where it has apparently been introduced with
sicae cannot be reliably distinguished from the cabbages imported from Europe. The larva feeds
important pest L. sativae on external characters, the initially in the petiole or stem and with its large size
genitalia are entirely distinct (ct. Figs. 499, 500). and internal feeding is almost certainly a primitive
Tavormina (1982) conducted interesting host species. It is not directly related to P. alyssi.
. selection experiments with L. brassicae in the
United States on the three hosts Barbarea vulgaris,
Order ERICALES (8 families, 4000 species)
Brassica nigra and Thlaspi arvense. He found that
'Insects collected from a particular host produced a No species of Agromyzidae have been described on
greater proportion of their mines on that host than this order but larval feeding of Phytobia has been
did insects that were not collected from that host, detected on the Ericaceae.
and their offspring had a shorter developmental
time on that host than did offspring whose parents
Ericaceae (125 genera, 3500 species)
were not collected from that host.' It was established
that 'these observed differences are due to genetic All species in this family are shrubs or small trees.
divergence within the wild population.' The confir- Fossil seeds have been found in Palaeocene deposits
mation that sympatric divergence was occurring in in England and pollen possibly representing Erica-
the population of L. brassicae which was studied is ceae is known from the Maestrichtian (Upper Cre-
significant in view of the unusual host range of this taceous) in the northern hemisphere and from the
species, with larval feeding known on Capparaceae, Eocene in Australia (Cronquist, 1981).
Resedaceae and Tropaeolaceae, and even on Pisum. The family is customarily divided into four sub-
Obviously such sympatric divergence can lead ulti- families and Suss (loc. cit.) has detected larval
mately to polyphagy or even speciation. feeding of Phytobia in three genera in two of the
I consider it must be significant that among the six subfamilies (see Table 15).
Liriomyza species whose primary hosts are in Bras-
Calluna vulgaris (L.) Hull- Europe, Asia Minor
sicaceae, larval feeding in four and also in the poly-
Erica arborea L. - southern Europe, North Africa,
phagous L. strigata is associated with the stem,
Caucasus
petiole or midrib and they are not thus true leaf
Vaccinium laurifolium Miq. - Java
miners feeding exclusively in the leaf blade. This
V. myrtillus L. - Central Europe
suggests a primitive origin, with the ancestral species
V. vitis idaea L. - northern Europe to Siberia and
being stem-feeders, with their modern descendants
Japan, arctic North America
forming secondary leaf mines.
Two Paraphytomyza species (as Phytagromyza) These records appear to represent two or pos-
have been recorded on Brassicaceae. Pa. sisymbrii sibly three distinct species.
(Hering, 1962) forms exclusively stem mines and is
known only from one locality near Berlin. The male
Order EBENALES (5 families, 1750 species)
genitalia are unique, with the distiphallus greatly
elongated (Fig. 326), giving no indication of the One host-specific and two polyphagous species are
affiliations of this isolated species. Pa. mamonowi known on this order and in addition larval feeding
(Hering, 1930) was described from Rostov, U.S.S.R. of Phytobia is known on two families.
but there is some doubt about its larval feeding. The
Subclass DILLENIIDAE 89
Sapotaceae (70 genera, 800 species) Upper Cretaceous. Botanists largely agree that the
Symplocaceae and Styracaceae are closely related.
This largely tropical family is known in both the Old
A paratype of Trop. polyphaga Spencer (1961b)
and New Worlds. Fossil pollen is known in Maes-
was reared from Symploeos spieata in Sri Lanka.
trichtian and more recently. The genitalia of this species, which has been tenta-
Suss (loc. cit.) has noted larval feeding of two
tively identified on 12 families, have not previous-
certainly distinct Phytobia species on this family:
ly been illustrated and those of the holotype (from
Chrysophyllum roxburghii G. Don - Java Tylophora flava, Asclepiadaceae) are shown in Fig.
Gluema ivorensis Aubr. et Pellegr. - Ivory Coast 667. It will be seen that there are small but distinct
differences from T. styrieieola (Fig. 328) and
T. theae (Fig. 260).
Ebenaceae (5 genera, 450 species)
Largely tropical and subtropical, this family is also
Order PRIMULALES (3 families, 1900 species)
old. Leaves thought to represent Diospyros (ebony)
have been found in Upper Cretaceous deposits but The relationships of this order are still obscure but
these records require confirmation. Pollen attri- Cronquist suggests that it is probably related to the
buted to the group dates from Lower Eocene and Ebenales. The few agromyzids known on the order
wood is known from the Oligocene (Cronquist, are isolated and show no relationships with those on
1981). any other orders in the Dilleniidae.
Suss (loc. cit.) has detected larval feeding of Phy-
tobia of two distinct species on this family:
Myrsinaceae (30 genera, 1000 species)
Maba marguensis Hiern. var. blumei Koord. et
This family, consisting mainly of evergreen trees or
Valet. - Java
shrubs, primarily in the tropics and subtropics, has
Royena lucida L. - South Africa (Natal), Brazil
been colonized by two cambium-boring Phytobia
species and by a single Tropieomyia.
Styracaceae (10 genera, 150 species) Suss (loc. cit.) has detected larval feeding of Phy-
tobia in five species in two genera:
This family appears to be younger than the two
preceding ones with no fossils known earlier than Ardisia diversifolia Koord. et Valet. - Java
the Tertiary. A. marginata Blume - Java
Tropieomyia styrieieola (Sasakawa, 1954b; 1961) Myrsine arvensis A. DC. var. venosa Koord. et
was described from Japan on Styrax japoniea, the Valet. - Java
larva forming silvery, epidermal mines characteristic M. melanopleos R. Br. - South Africa
of this genus. The male genitalia are shown in Figs. M. tenuifolia Koord. et Valet - Java
328, 329 and the serrated mouth-hooks of the
The only leaf miner known on the family is Tropi-
larva, also characteristic of the genus in Fig. 330.
eomyia flaeourtiae, on Maesa in Madagascar and the
This species shows considerable polyphagy and has
leaf mine was figured by Paulian (1953).
been recorded on 12 families (see Table 92). There
is no indication of its primary host. It is noteworthy
that the genitalia of T. styricieola differ significantly Primulaceae (30 genera, 1000 species)
from those of T. theae (Figs. 260, 261), which is
The Primulaceae occur mainly in temperate and
known from Sri Lanka to Papua New Guinea. It is
cold regions of the northern hemisphere, with some
impossible to assess at this time the degree of rela-
species on tropical mountains.
tionship between these two species.
Only three species of Agromyzidae are known on
Phytomyza kisakai Sasakawa (1954c; 1961) is
the family. Liriomyza eyclaminis Suss (1987) ap-
known only in Japan, also feeding on Styrax japoni-
pears to be not uncommon in central and southern
ea, forming normal upper surface leaf mines. This is
Italy and leaf mines were earlier found on the Adri-
an entirely black species known only from the
atic island Hvar by Hering (1967). The description
female holotype and its affiliations are unclear. The
is based on adults reared from Cyclamen hederi-
posterior spiracles of the larva are unique, with up
folium and the mines on Hvar were found on
to 49 pores arranged in ten groups (Fig. 331).
C. repandum (as C. verna Ie ). It seems probable that
L. eyclaminis is a primitive species, with the orbits
Symplocaceae (1 genus, 300-400 species) greatly projecting and the orbital setulae upright
(not reclinate as in most species), as illustrated by
This family occurs in moist tropical and subtropical
L. Suss (1987: fig. 1). In the male genitalia the
regions and is old, with pollen dating from the
90 Class MAGNOLIOPSIDA (Dicotyledons)
328
~~
333 335 334
Figs. 328-330. Tropicomyia styricicola: 328, aedeagus, side view; 329, same, ventral view; 330, larval mouth parts (Sasakawa).
Fig. 331. Phytomyza kisakai: larval posterior spiracles (Sasakawa).
Figs. 332, 333. Liriomyza cyclaminis: 332, aedeagus, side view; 333, distiphallus, ventral view (L. Siiss).
Fig. 334. Chromatomyia primulae: leaf mine on Primula vulgaris (Hering).
Fig. 335. Chromatomyia soldanellae: leaf mine on Soldanella alpina (Hering).
aedeagus ends in two elongate tubules (Figs. 332, of 'the same general form.' However, the mines of
333), which appears to associate the species with the two species differ significantly. That of C. pri-
L. equiseti on Equisetaceae (Fig. 7), L. lutea on mulae is a long, narrow linear mine (Fig. 334), while
Apiaceae, L. buhri on Campanulaceae (Figs. 868, in C. soldanellae the mine is linear only in the first
869) and several others of which the host is not instar and then develops into a large, roundish
known. blotch (Fig. 335 and Stary, 1950: fig. 2).
Two Chromatomyia species are known on Pri- It seems probable that Chr. primulae and
mula and Soldanella, respectively. Chr. primulae Chr. soldanellae had a common ancestor, which
(Robineau-Desvoidy, cf. Spencer, 1976a) is com- may well have existed prior to the evolution of Pri-
mon on Primula vulgaris in Western Europe, and in mula and Soldanella into separate genera. As the
Poland has also been reared from P. elatior (Beiger, hosts diverged, so did the leaf miners. This pair of
197 5b). She mentions that 'In the mountains it species might thus well represent a case of co-evolu-
reaches the alpine layer.' Chr. soldanellae was de- tion. Both are strictly monophagous, and Chr. solda-
scribed by Stary (1950) from mines on Soldanella nellae is now isolated with its host at high altitudes
carpatica in Slovakia, 1353 m. Beiger (1975b; in the mountains of central Europe.
1980b) illustrated the male genitalia of Chr. primu- The Primulaceae are clearly largely unacceptable
lae and Chr. soldanellae and mentions that these are to Agromyzidae and for many years the lack of leaf
Subclass DILLENIIDAE 91
mines on such apparently suitable hosts as Anagallis The relationship of the three Agromyza species
and Lysimachia has caused surprise among special- on Salicaceae with species on Rosaceae is note-
ists on the family. worthy, comparable to a similar relationship of
species on the Ulmaceae and Betulaceae (Hama-
There are a number of points of interest in the melidae). Agromyza abutilonis is of special interest,
generic composition of agromyzids on the Dilleni- representing the only known case of a switch of a
idae: grass-feeding species to a dicot. A. brunnicosa on
Malva in the Canary Islands is a unique, isolated
1. The large number, approximately 22, of Phyto-
species.
bia species detected from larval feeding.
Liriomyza lepidii in New Zealand on Brassica-
2. The continuing small number of Melanagromyza
ceae and L. violicaulis in Europe on Violaceae
species (three), fewer than in the Hamamelidae
represent isolated colonizations from unrelated
and Caryophyllidae, with five each.
ancestral hosts.
3. The four Ophiomyia species, 0. tiliae (Tiliaceae),
Considering its size (78 families, 25 000 species),
0. abutilivora (Malvaceae), 0. alliariae and
the Subclass Dilleniidae supports relatively few
0. texana (Brassicaceae) appear to represent
agromyzids (61), though, as might be expected,
four separate colonizations.
more than the smaller Hamamelidae and Caryo-
4. The first appearance of Paraphytomyza (seven
phyllidae.
species) and of family-specific Chromatomyia on
Primulaceae (two species)
5. The unique group on Viola included in Galio-
myza
6. The high number of polyphagous species in both
Tropicomyia (five) and Liriomyza (six).
Table 15
Order THEALES
Theaceae
Camellia Phytobia sp.; Tropicomyia theae
Cleyera Phytobia sp.
Guttiferae
Subfamily Calophylloideae
Endodesmia Phytobia sp. ( West Africa)
Subfamily Clusioideae
Symphonia Phytobia sp. (West Africa, Brazil)
Actinidiaceae
Saurauia Phytobia sp.
Order MALVALES
Tiliaceae
Luehea Phytobia sp. ( I, South America)
Sioanea Phytobia sp. ( 2, Java)
Tilia Phytobia sp. ( 3, North America ); Ophiomyia tiliae
92 Class MAGNOLIOPSIDA (Dicotyledons)
Sterculiaceae
Brachychiton T. polyphyta
Guazuma Phytobia sp. ( South America)
Melochia Phytobia sp. ( Java)
Malvaceae
Tribe Malopeae
Kitaibelia Chromatomyia horticola
Tribe Malveae
Abelmoschus Liriomyza sativae
Abutilon Hexomyza abutilonicaulis; O. abutilivora; Agromyza abutilonis;
Calycomyza longicauda
Althaea Melanagromyza hicksi; L. strigata, L. xanthocera; Cal. malvae;
Chr. horticola
Anoda Chr. horticola
Malva A. abutilonis, A. brunnicosa; L. strigata,
L. trifolii, L. xanthocera; Cal. malvae,
Cal. sidae; Chr. horticola
Palava Chr. horticola
Sida O. abutilivora; Cal. malvae, Cal. sidae
Tribe Hibisceae
Gossypium T. flacourtiae; L. tri f olii
Hibiscus Mel. hibisci; A. abutilonis; L. trifolii;
Chr. horticola
Gen. indet. A. malvaceivora
Order LECYTHIDALES
l.ecythidaceae
Barringtonia T. polyphaga, T. polyphyta
l.ecythis Phytobia sp. (Brazil)
Order VIOLALES
Flacourtiaceae
Casearia Phytobia sp.
Dovyalis T. flacourtiae
Flacourtia T. flacourtiae, T. polyphaga
Cistaceae
Cistus Phytobia sp.
Violaceae
Melicytus L. flavolateralis
Viola L. violicaulis; Galiomyza takakoae,
G. violiphaga, G. violivora; Chr. horticola
Turneraceae
Piriqueta L. trifolii
Subclass DILLENIIDAE 93
Passi f10raceae
Passiflora T. atomella, T. f1acouniae,
? T. momordicae; L. schmidti
Achariaceae
Ceratiosicyos T. ceratiosicyi
Cucurbitaceae
Bryonia L. bryoniae, L. strigata
Ceratosanthes L. sativae
Citrullus L. trifolii
Coccinia T. atomella
Cucumis L. bryoniae, L. huidobrensis, L. sativae,
L. trifolii; Ch ... horticola
Cucurbita L. bryoniae, L. sativae, L. strigata,
L. trifolii; Chr. horticola
Ecballium L. strigata
Luffa L. criptica nom. nud. (Argentina),
L. strigata
Marah Liriomyza sp.n. (California)
Melothria Tropicomyia sp.n. (Papua New Guinea)
Momordica T. momordicae
Rytidostylis ? Liriomyza sp.
Loasaceae
Caiophora L. bryoniae
Loasa A.? pseudoreptans; L. strigata;
Chr. horticola
Mentzelia L. strigata
Gen. indet. ? Calycomyza sp. (Argentina)
Order SALICALES
Salicaceae
Populoxylon Palaeophytobia salicaria
Populus Pb. cambii; Hexomyza schineri;
Japanagromyza salicifolii; A. albitarsis,
A. isolata; Paraphytomyza populi,
Par. populi co la, Par. tremulae
Salix H. cecidogen~, H. salicis, H. simplicoides,
H. winnemanae; 1. salicifolii; A. albitarsis,
A. salicina; Par. fulvicornis, Par. tridentata
Order ERICALES
Ericaceae
subfamily Ericoideae
Calluna Phytobia sp: (I)
Erica Phytobia sp. (I)
subfamily Vaccinioideae
Vaccinium Phytobia sp. (I or 2 ?), (3, Java)
94 Class MAGNOLIOPSIDA (Dicotyledons)
Order EBENALES
Sapotaceae
Chrysophyllum Phytobia sp. (Java)
Gluema Phytobia sp. (Ivory Coast)
Ebenaceae
Maba Phytobia sp. (Java)
Royena Phytobia sp. (South Africa. Brazil)
Styracaceae
Styrax T. styricicola; Phytomyza kisakai
Symplocaceae
Symplocos T. polyphaga
Order PRIMULALES
Myrsinaceae
Ardisia Phytobia sp. (1. Java)
Maesa T. flacourtiae
Myrsine Phytobia sp. (1, Java).
Phytobia sp. (2, South Africa)
Primulaceae
Cyclamen L. cyclaminis
Primula L. trifo1ii; Chr. primulae
Soldanella Chr. soldanellae
Subclass ROSIDAE
This large Subclass as proposed by Cronquist by the random nature of the colonization and radia-
(1981) consists of 18 orders, 114 families and about tion which has occurred.
58 000 species. Forty of the families have been
colonized by Agromyzidae in 15 genera and with
Order ROSALES (24 families, 6600 species)
250 species restricted to the Subclass. This is slight-
ly more than half those known on the Asteridae, Although the 24 families in this order appear to be
although the two Subclasses contain approximately interrelated, only six have been colonized by Agro-
the same number of species. myzidae.
Although Cronquist considers that the Rosidae The largest family, the Rosaceae (3000 species),
form a natural group, there is no overlap in the has been colonized by only marginally more species
agromyzid species on the various orders. Most of (19) than the other five families in the order (to-
the orders are not acceptable to polyphagous spe- gether some 2300 species) on which 17 host-specif-
cies, but the Euphorbiales represent a minor excep- ic agromyzids are known.
tion. The Rosales are considered to be the most
primitive order, the other orders all being derived Table 16
directly or indirectly from them. Cronquist never-
Agromyzid genera on order Rosales
theless concedes that the Rosales are morphologi- (excluding Rosaceae, see below)
cally diffuse and difficult to define. He writes: 'The
Rosidae are evidently derived from the Magnoli- Host-specific OJigophagous Polyphagous
idae, as indicated by the fossil records as well as by species
comparative morphology. The fossil record suggests
Phytobia
that the Rosidae began to diverge from the ancestral Melanagromyza
Magnoliidae near the Albian stage of the Lower Tropicomyia 4
Cretaceous.' Agromyza
Thus, with very similar male genitalia, there is no Phytoliriomyza 2
inherent reason why such groups as the feeders on Liriomyza 5 (? + 2)
Chromatomyia 4
flex (Aquifoliaceae) known from fossil pollen in Phytomyza 4
early Upper Cretaceous, and the Phytomyza angeli-
cae group on Apiaceae should not be direct de- 17 5
scendents of one or other of the ancestral groups on
Ranunculaceae. Relationships of species on the
Pittosporaceae (9 genera, 200 species)
Rosidae are discussed further at the end of this
section. Some authors, notably Frohne and Jensen (1985)
Radiation of closely related agromyzid species on place this family in close relationship with the
individual families is becoming increasingly appar- Araliales primarily on chemical affinity. However,
ent in this Subclass, particularly in Agromyza on the this is not accepted by Cronquist who states that the
Rosaceae (seven species), Phytomyza on !lex, Aqui- Pittosporaceae 'must be considered to be an early
foliaceae (eight species) and most strikingly on the simple-leaved offshoot that independently devel-
Apiaceae (64 Phytomyza species). oped the secretory system and some of the special
Theoretically, two families in the same subclass chemicals of some of the more advanced Rosidae.'
might be expected to show comparable colonization Three host-specific agromyzids are known on Pit-
by Agromyzidae. However, this is not so, as is tosporaceae, appearing to be isolated and not
immediately apparent in the two families Legumino- directly related to species on other families in the
sae and Apiaceae (Tables 21, 26). In the one genus Order. Two have been reared from Pittosporum in
Phytomyza there are six species on the Leguminosae N.S.W., Australia.
but 64 on the Apiaceae. This can only be explained Phytoliriomyza pittosporocaulis Hering (cf. Spen-
96 Class MAGNOLIOPSIDA (Dicotyledons)
\':.. ,,;~
-v" ____- - _
336
339
340
341
Figs. 336, 337. Phytoliriomyza pittosporocaulis: 336, aedeagus; 337, epandrium.
Figs. 338, 339. Phytoliriomyza pittosporophylli: 338, aedeagus; 339, epandrium.
Figs. 340, 341. Liriomyza obscurata: 340, aedeagus, side view; 341, same, ventral view.
Figs. 342, 343. Liriomyza philadelphivora: 342, aedeagus, side view; 343, distiphallus, ventral view.
Subclass ROSIDAE 97
cer, 1977c) forms stem mines and P. pittosporophyl- in Figs. 342, 343. Unfortunately the genitalia of
Ii Hering (cf. Spencer, 1977c) produces unique leaf L. philadelphi have not been illustrated but it seems
galls. The aedeagus of these two species is remark- highly probable that the two species are monophy-
ably similar (pittosporocaulis: Fig. 336 and pitto- letic. Also in Japan Phytomyza deutziae Sasakawa
sporophylli: Fig. 338) but the form of the epandrium (cf. Sasakawa, 1961) forms leaf mines on Deutzia
is strikingly different. In P. pittosporocaulis the sur- crenata (pupating internally) and P. hydrangeae
stylus is discrete, without spines, but there is a Sasakawa (d. Sasakawa, 1961) feeds on Hydrangea
strong bunch of these at the hind-corner of the paniculata (pupating externally). Only females are
epandrium (Fig. 337), while in pittosporophylli the known of P. deutziae and the genitalia of the male
surstylus is fully fused with the epandrium, which holotype of P. hydrangiae have not been illustrated.
bears a row of spines along its entire inner margin It seems possible or even probable that these species
(Fig. 339). With the differing biology, this appears to are monophyletic but this cannot be confirmed on
be a case where sympatric speciation would have available information.
been possible but it is more likely that speciation
occurred in isolated populations which developed
Grossulariaceae (25 genera, 350 species)
different larval feeding habits, and the two species
now overlap in distribution. Opinions of botanists on the status of this family
Liriomyza obscurata Spencer (cf. Spencer, have varied from the one extreme in the Englerian
1977c) is believed to be the species responsible for system in which all genera were included within the
the leaf mines which are not uncommon on Billar- Saxifragaceae to the other extreme when eight
diera scandens in the Sydney area and I caught flies separate families were recognised by Takhtajan.
on this host. The male genitalia (Figs. 340, 341) are Here I follow Cronquist in accepting the Grossu-
remarkably similar to those of L. antipoda Harrison lariaceae, with only two of the genera, Escallonia
(cf. Spencer, 1976b) known only on Antipodes and Ribes, recorded as attacked by larvae of Phyto-
Island, New Zealand (cf. Spencer, 1976b: figs. 24, bia.
25). The two species may well be synonymous or in Grossenbacher (1910) records 'pith flecks' in
any case must be closely related and thus can only Ribes in the United States but these represent not
represent a case of long distance aerial dispersal Phytobia but a microlepidopteron.
from Australia. Suss (pers. comm.) has noted larval feeding of
I found mines on Bursaria spinosa in the Botani- Phytobia in two genera:
cal Gardens, Sydney and also on Sollya heterophylla
Escallonia iveyi hort. - South America
on Mt. Chudalup, W.A. probably caused by Lirio-
Ribes nigrum L. - Europe, Central Asia, Himalayas
myza spp. but unfortunately no adults were ob-
tained. Two polyphagous Tropicomyia spp. are These records must represent two distinct spe-
known on Pittosporum - T flacourtiae in Africa cies.
and T styricicola in Japan. Also, in Argentina, Griffiths discovered Agromyza potentillae (which
Liriomyza schmidti is known on this genus. normally feeds on Rosaceae) on Ribes glandulosum
in Alberta representing an interesting transfer from
its normal hosts in the Rosaceae.
Hydrangeaceae (17 genera, c. 170 species)
There is disagreement among botanists regarding
Crassulaceae (25 genera, 900 species)
the taxonomic position of the genera Deutzia, Hy-
drangea and Philadelphus on which leaf miners are This family is generally accepted as being closely
known. Willis (1957) included these genera in Saxi- related to the Saxifragaceae but the six agromyzids
fragaceae and this is accepted by Farr et al. (1979). known on each family show no evidence of being
Cronquist (1981) uses the family name Hydrangea- directly related.
ceae, while Willis (1985) and Goldberg (1986) use In South Africa Melanagromyza cotyledon us
both Hydrangeaceae and Philadelphaceae. Here I Spencer (1960a) feeds in the leaves of Cotyledon
accept Hydrangeaceae for all three genera. orbiculata. Melanagromyza species are not leaf
Two polyphagous Tropicomyia species are miners and I suspect that the larvae feed either in
known on Hydrangea. I found mines of T polyphyta the midrib or in the fleshy leaves which may in some
on H. macrophylla at the Botanical Gardens, Syd- physiological characters be comparable to stems.
ney and T polyphaga has been recorded in India on The posterior spiracles of the puparium (Fig. 346)
Hydrangea sp. Two Liriomyza species are known on confirm beyond question that the species is a true
Philadelphus, L. philadelphivora Spencer (1969b) Melanagromyza. The male genitalia (Figs. 344, 345)
in Canada and L. philadelphi Sasakawa (1961) in are entirely typical of the genus.
Japan. The aedeagus of L. philadelphivora is shown Tropicomyia kalanchoes Spencer (1985a) forms
98 Class MAGNOLIOPSIDA (Dicotyledons)
~gf7~
344
:,\
~-3LJ 346
345
347 ci~
,-,-,'
~.·~:~.1?
- .. - - '
349
350 351
/'~--- ~I
------' -- -
"'i I
352 ,
I 'I:
, \
" I
Subclass ROSIDAE 99
deep mines in the palisade parenchyma (not epider- from Saxifraga carpatica in the Tatra Mtns., Poland.
mal as is normal in the genus) in the succulent leaves The male genitalia (Fig. 355) give no indication of
of Kalanchoe densiflorum and is only known from the affiliations of this species but it does belong to
the Aberdare Mountains, Kenya. Externally there is the group with an ellipse of eight pores on the
nothing to distinguish this species from others in the posterior larval spiracles.
genus but the form of leaf mine and the male gen- Five species described in Phytomyza but sub-
italia (Fig. 347) (which were possibly damaged in sequently transferred to Chromatomyia by Griffiths
the only available male) both suggest that this is an (1972a) are known as leaf miners on this family, 3
isolated species. on Saxifraga and one each on Tiarella and Mitella.
The single Liriomyza on Umbilicus, L. umbilici The male genitalia of all five are generally similar (d.
Hering (1927a) is only known from Tenerife, Ca- Chr. deirdreae: Fig. 356) but there is greater dif-
nary Islands and the Azores. The species is a typical ferentiation in the form of the larval posterior spira-
Liriomyza and the aedeagus (Figs. 348, 349) is close cles (Fig. 357).
to that of L. bryoniae (d. Figs. 289, 290). Chr. aizoon (Hering, d. Griffiths, 1972a) is
Three Phytomyza species are known on the known only in the mountains of Central Europe but
family, two on Sedum and one on Rhodiola. P. sedi- is doubtless more widespread with its host Saxifraga
cola Hering (d. Hendel, 1935), present only in paniculata (= aizoon Jacq.) in mountains further
Europe, has genitalia (Figs. 350, 351) clearly as- east to the Caucasus. The posterior spiracles have
sociating it with Ranunculaceae feeders on Ane- the 10 pores arranged in a circle (Fig. 358). Chr.
mone, Clematis and Ranunculus (d. Figs. 69, 160) saxifragae (Hering, d. Griffiths, 1972a), known on
and it seems clear that there was an ancestral switch Saxifraga rotundifolia, is also a high elevation spe-
to Sedum from a species on one of these genera. Of cies, in Central Europe and the Balkans, and the
the 2 other species, P. sedi Kaltenbach (d. Hendel, posterior spiracles have 22-25 pores in a wide
1936) is a European species (male genitalia: Figs. ellipse (Fig. 359). Chr. mitellae (Griffiths, 1972a) is
352, 353), while P. rhodiolae Griffiths (1976a) is known only in Alberta, Canada on Mitella nuda and
only known in the Yukon Territory, Canada (male the posterior spiracles are generally similar but with
genitalia: Fig. 354). Griffiths suggests that these are fewer pores (14-17). Chr. tiarellae (Griffiths,
sister-species and if they are monophyletic with 1972a) was described from Alaska on Tiarella tri-
P. sedicola, there has been considerable divergence. foliata and has subsequently been recorded on
The Crassulaceae may thus have been colonized on Heuchera and Tellima in California; Tolmiea men-
two separate occasions. ziesii is a further probable host in Alaska. Here the
posterior spiracles have a similar arrangement with
about 20 pores. These three species are clearly
Saxifragaceae (40 genera, 700 species)
monophyletic and apparently restricted to different
There has been considerable divergence of opinion genera in the family. Chr. deirdreae (Griffiths,
among botanists on the limits of this family which 1972a) has a wider distribution and, as far as is
has been variably classified by different specialists, known, is restricted to the single genus Saxifraga. It
including Engler, Hutchinson and Takhtajan. Here, was described from Alaska, is present in Alberta
I follow Cronquist (1981) who restricts the family to and British Columbia, and occurs also in Japan
herbaceous genera, with the largely woody genera where it was misidentified by Sasakawa (1961) as
segregated into the Hydrangeaceae and Grossularia- Phytomyza saxifragae Hering. The larval posterior
ceae. spiracles are of the same basic form as in Chr. mite/-
Eocene fossils have been referred to this family lae, saxifragae and tiarellae but become enlarged
but Savile (1975) considers that 'it is difficult to and differentiated into two 'arms,' with up to 40-45
believe that the Saxifragaceae originated before the pores (Fig. 357). I feel it can be accepted that all five
Oligocene period.' species have evolved from a single ancestral coloni-
Of the six agromyzids known on the Saxifraga- zation of the family.
ceae, one is in Liriomyza, five in Chromatomyia.
Liriomyza clarae Beiger (1972b) has been reared
Figs. 344-346. Melanagromyza cotyledonus: 344, aedeagus, side view; 345, same, ventral view; 346, posterior spiracles of puparium.
Fig. 347. Tropicomyia kalanchoes: aedeagus.
Figs. 348, 349. Liriomyza umbilici: 348, aedeagus, side view; 349, same, ventral view.
Figs. 350, 351. Phytomyza sedicola: 350, aedeagus, side view; 351, same, ventral view.
Fig. 352, 353. Phytomyza sedi: 352, aedeagus, side view; 353, distiphallus, ventral view (Griffiths).
Fig. 354. Phytomyza rhodiolae: aedeagus (Griffiths).
Fig. 355. Liriomyza darae: aedeagus (Beiger).
100 Class MAGNOLIOPSIDA (Dicotyledons)
357
356 .... ""-
358 359
Figs. 356, 357. Chromatomyia deirdreae: 356, aedeagus; 357, posterior spiracles of puparium (Griffiths).
Fig. 358. Chromatomyia aizoon: posterior spiracles of puparium (Griffiths).
Fig. 359. Chromatomyia saxifragae: posterior spiracles of puparium (Griffiths).
Table 17
Phytobia 6
Melanagromyza 1
J apanagromyza
Agromyza 7 (1 undescribed)
Liriomyza 1
Paraphytomyza 1
Phytomyza 2
19
Table 18
Pittosporaceae
Billardiera Liriomyza obscurata
Bursaria Phytobia sp.; Liriomyza sp.(?)
Pittosporum Tropicomyia flacourtiae, T. styricicola; Phytoliriomyza pittosporocaulis,
Phlr. pittosporophylli; L. schmidti
Sollya Liriomyza sp.(?)
Hydrangeaceae
Deutzia Phytomyoza deutziae
Hydrangea T. polyphaga, T. polyphyta; P. hydrangeae
Philadelphus L. philadelphi, L. philadelphivora
Grossulariaceae
Escallonia Phytobia sp.
Ribes Phytobia sp.; Agromyza potentillae
Crassulaceae
Cotyledon Melanagromyza cotyledonus; P. sedicola
Kalanchoe T. kalanchoes
Rhodiola P. rhodiolae
Sedum P. rhodiolae, P. sedi, P. sedicola
Umbilicus L. umbilici
Saxi fragaceae
Heuchera Chromatomyia tiarellae
Mitella Chr. mitellae
Saxifraga L. clarae; Chr. aizoon, Chr. deirdreae, Chr. saxifragae
Tellima Chr. tiarellae
Tiarella Chr. tiarellae
Tolmiea Chr. tiarellae
Rosaceae
Subfamily Spiraeoideae
Aruncus A. spiraeoidearum,
Spiraea A. spiraeoidearum, A. valdorensis
102 Class MAGNOLIOPSIDA (Dicotyledons)
Subfamily Rosoideae
Agrimonia A. potentillae; L. trifolii
Alchemilla A. potentillae
Aremonia A. potentillae
Comarum A. potentillae
Duchesnea Japanagromyza duchesneae
Filipendula A. filipendulae, A. potentillae, A. sulfuriceps
Fragaria A. fragariae, A. potentillae, A. sulfuriceps
Geum A. potentillae
Potentilla Mel. sagehenensis; A. filipendulae, A. potentillae, A. sulfuriceps
Poterium A. potentillae
Rosa Phytobia sp.; A. potentillae
Rubus A. potentillae, A. sulfuriceps, Agromyza sp. n.
Sanguisorba A. potentillae, A. sulfuriceps
Sibbaldia A. potentillae
Subfamily Prunoideae
Parinarium Phytobia sp.
Pruninium Palaeophytobia prunorum ( U.S.A. )
Prunus Phb. cerasiferae, Phb. pruni, Phb. prunivora, Phytobia spp.
( Java, Mongolia ); P. persicae
Subfamily Maloideae
Amelanchier Phb. amelanchieris
Chaenomeles Phytobia sp.
Cotoneaster Phytobia sp.
Crataegus Phb. carbonaria, Phb. pruni; L. trifolii
Cydonia Paraphytomyza cydoniae; Phytobia sp.
Malus Phb. carbonaria, Phytobia sp. ( North America ); P. heringiana
Mespilus Phytobia sp.
Pyracantha Phytobia sp.
Pyrus Phytobia sp.
Sorbus Phb. aucupariae
361
365
364
366
368
and was described from two localities in Japan. It species Phytobia pruni (Grossenbacher, cf. Spencer
has since been recorded in Papua New Guinea and and Steyskal, 1986) was described from New York
on Vanuatu in the Pacific. The male genitalia are on Prunus and Crataegus. Unfortunately no speci-
shown in Fig. 364 and the posterior spiracles of the mens can now be traced and the affiliations of the
larva in Fig. 365. species cannot be established. It is surprising that
Of the four Agromyza species occurring on this this species should feed on both Prunus and Cratae-
subfamily, the commonest is A. potentillae (Kalten- gus which is in the further subfamily Maloideae.
bach, cf. Spencer, 1976a) (known in earlier litera- Pb. prunivora Spencer (1981) is known in Califor-
ture as A. spiraeae Kalt.) which is widespread in nia feeding on Prunus and the male genitalia (Fig.
Europe and North America. It is highly oligopha- 373) do not suggest close relationship with other
gous, occurring on Fragaria, Geum, Potentilla and known species. Pb. cerasiferae (Kangas, 1955) was
Rubus (Potentilleae), and crosses tribal boundaries described from Prunus cerasifera and other Prunus
to colonize Agrimonia, Alchemilla, Aremonia, spp. in England and has since been found in Ger-
Comarum, Poterium and Sanguisorba (Sanguisor- many. The genitalia (Fig. 374) suggest possible rela-
beae), Filipendula (Ulmarieae) and rarely Rosa tionship with Pb. amelanchieris on Amelanchier
(Roseae). Interestingly, while it attacks Rubus (subfamily Maloideae, see below).
idaeus (raspberry), it never feeds on R. fruticosus A fossil species, Palaeophytobia prunorum, has
(blackberry). The male genitalia are shown in Fig. also been described from Eocene wood described as
366 and the linear-blotch mine in Fig. 367. Pruninium gummosum from the Yellowstone Na-
A. sulfuriceps Strobl is present in both Europe tional Park, Wyoming, U.S.A. (Suss and Muller-
and North America, with the only confirmed hosts Stoll,1980).
being Filipendula, Potentilla, Rubus and Sanguisor- Suss (pers. comm.) has detected larval feeding of
ba. This species is differentiated from others in the Phytobia spp. in 20 modern species of Prunus, as
group by having the frons and antennae bright follows:
yellow (dark in the others) but the male genitalia
P. allegheniensis Porter - North America
(Fig. 368) differ only slightly from A. potentillae.
P. americana Marsh. - North America
Here also this species feeds on three tribes - Poten-
P. amygdalus Matsch - Western Asia, North Africa
tilleae, Sanguisorbeae and Ulmarieae.
P. armeniaca L. - Transcaucasia to western Hima-
A. Jilipendulae Spencer (1976a) forms exclusive-
laya, northern China and Mongolia
ly linear mines (Fig. 370) only on Filipendula in
P. avium L. - Europe to western Siberia, northern
Europe and was earlier confused with A. potentillae
Iran and southern Turkestan
until recently, although the differences in the leaf
P. cerasus L. - Asia Minor
mines had been recorded by Hering (1957b: No.
P. domestic a L. - Central Europe to Asia
2204). There is little differentiation between the two
P. emarginata (Dougl.) Walp. - North America
in external characters or in the male genitalia (Fig.
P. hortulana Bailey - North America
369) and this is an interesting case where the evolu-
P. ilicifolia (Nutt.) Walp. - North America
tion is more apparent in the larval feeding instinct
P. isititia L. - Europe, Asia Minor, Himalayas.
and the different host selection (although A. poten-
P. javanica Miq. - Java
tillae does occur on Filipendula) than in external
P. lusitanica L. - Spain, Canary Islands
morphology.
P. mahalab L. - Europe, Asia Minor, Caucasus,
A. fragariae Malloch (cf. Spencer and Steyskal,
Turkestan, Syria
1986) which is known only in California, Alberta
P. pennsylvanica L. - North America
and Quebec was originally considered identical with
P. serotina Ehrh. - North America
A. potentillae but the male genitalia (Fig. 371)
P. spinosa L. - Europe, North Africa, western Asia
confirm its distinctness and the only host appears to
P. ssiori F. Schmidt - Japan
be Fragaria.
P. virginiana L. - North America
I have found what certainly represents an unde-
scribed Agromyza sp. to be locally abundant on Some of these records certainly refer to Phytobia
Rubus odoratus both in Canada and Virginia and on cerasiferae and others to Pb. pruni or Pb. prunivora
R. strigosus near Minneapolis, Minnesota which but it seems probable that in other areas further
forms a long, linear mine (Fig. 372). It is to be hoped undescribed species are represented.
that this species can be reared in due course. Larval feeding of Phytobia spp. has also been
detected in Parinarium campestre in tropical Africa
and in P. curatelaefolium in West Africa (Suss, pers.
Subfamily Prunoideae
comm.).
Two genera, Parinarium and Prunus, have been Phytomyza persicae Frick (cf. Spencer and Steys-
colonized by Agromyzidae. The cambium-boring kal, 1986) is the only leaf miner known on Prunus,
Subclass ROSIDAE 105
369
371 370
.....
372
374
376
Figs. 369, 370. Agromyza jilipendulae: 369, aedeagus; 370, leaf mine on Filipendula ulmaria (Hering).
Fig. 371. Agromyza fragariae: aedeagus.
Fig. 372. Agromyza sp.n.: leaf mine on Rubus odoratus.
Fig. 373. Phytobia prunivora: aedeagus.
Fig. 374. Phytobia cerasiferae: aedeagus.
Fig. 375. Phytomyza persicae: aedeagus.
Fig. 376. Phytomyza heringiana: aedeagus.
106 Class MAGNOLIOPSIDA (Dicotyledons)
on P. persica in the United States. The male gen- (Fig. 3?8). Also in Europe Pb. aucupariae (Kangas,
italia (Fig. 375) closely associate this species with 1949) IS known on Sorbus aucuparia but no males
Phytomyza heringiana (cf. Hendel, 1935) (Fig. 376) are known.
which feeds on Malus sylvestris in Europe (subfami- Suss (pers. comm., 1987) has detected larval
ly Maloideae). Although the adults differ only slight- feeding of Phytobia spp. on the following species in
ly, the form of leaf mine is distinct in the two. In both ten genera:
species pupation takes place in the mine. It seems
Amelanchier alnifolia Nutt. - Alaska to southern
clear that the two are sister-species.
border of California, Nebraska, Colorado, New
The distribution of these two species is puzzling.
Mexico
Cultivated peach is the only known host of P. per-
Chaenomeles japonica (Thunb.) Lindl. - Japan
sicae in the United States and although peach has
Cotoneaster acuminata Lindl. - eastern India from
escaped from cultivation in some areas, it was intro-
Beas to Sikkim
duced from Europe, being a native of China. If the
C. bacillaris Wall. - eastern India
speciation now detectable has occurred following
C. integerrima Med. - Europe, SE Mediterranean
natural dispersal via Beringia, the original host of
area, isolated records in Siberia
P. persicae was presumably one of the 30 or so
C. microphylla Wall. - Himalayas from Kashmir to
native Prunus species present in the United States,
Bhutan and Sikkim
on which it may well still be present. It is significant
C. tomentosa (Ait.) Lindl. - mountains of southern
that no leaf miner has been found on this host in
Europe, Balkans
Europe and thus P. heringiana appears to be fully
Crataegus monogyna Jacq. - Europe, Caucasus
isolated on Malus.
Syria, North Africa, Siberia, Himalayas '
C. x degeni Zsak - Hungary
Subfamily Maloideae (Pomoideae) C. nigra W.K. - Hungary, Romania, Yugoslavia (Ser-
bia)
Agromyzidae are known on ten genera in this
C. oxycantha L. - Europe, W. Asia, West Africa
subfamily. Only two are leaf miners, the other
C. punctata Jacq. - U.S.A.
records all being of Phytobia species.
Cydonia vulgaris Pers. - Southern Europe western
Three Phytobia species have been reared. Pb. Asia '
amela~chieris (Greene, cf. Spencer and Steyskal,
Malus angustifolia Michx. - North America
1986) IS known on Amelanchier canadensis in Cana-
Mespilus (? = Crataegus) germanic a L. - southern
da and the United States and the male genitalia (Fig.
and central Europe, also Asia (cultivated)
377) show beyond doubt that this species is more
Pyracantha coccinea Roem. - Italy, southern Eu-
closely related to Pb. betulivora (Fig. 225) known
rope, Caucasus, Asia
on Betula nigra, also in Canada and the United
Pyrus communis L. - Europe, Asia Minor (also
States, than to the other species with hosts in the
recently found in Pyrus in France, M. Martinez,
Rosaceae. Pb. carbonaria (Zetterstedt, cf. Spencer,
pers. comm.)
1976a) feeds on Crataegus and Malus in Europe
377
378
Fig. 377. Phytobia arnelanchieris: aedeagus, side view.
Fig.378. Phytobia carbonaria: aedeagus, ventral view.
Subclass ROSIDAE 107
P. piraster Borkh - Central Europe undertaken Polhill and Raven (1981), based on the
Sorbus americana Marsh - North America International Legume Conference at Kew in 1978,
S. aria (L.) Cr. - Europe, Caucasus, Siberia, Hima- write: 'The recognition of one family or three fami-
layas lies within the Leguminales may still be regarded as
S. aucuparia L. - Europe, northern Asia, NW Hima- a matter of opinion .... In general the evidence of
layas this volume would seem to support the concept of a
S. chaemaemespilus Crantz - Europe single family ... .' Later, it is added that: 'In practice it
S. domestica L. - southern Europe is immaterial whether Leguminosae is [sic] regarded
S. intermedia (Ehrh.) Pers. - northern Europe as one family or three.' Here it is appropriate to
S. sambucifolia Roem. - southern Greenland, in point out that the Agromyzidae detect a very sharp
U.S.A., New Mexico, central California division between the three groups, which could
S. torminalis (L.) Cr. - Europe, Asia Minor, North justify giving them family rank. Of the 96 family-
Africa. restricted species, there are just two on the Caesal-
In this subfamily two leaf miners are known, of pinioideae, one on the Mimosoideae and 93 on the
which Phytomyza heringiana on Malus was dis- Papilionoideae. This indicates a striking chemical
cussed above under Prunus. In addition, Paraphyto- difference between the three subfamilies. Neverthe-
myza cydoniae Hendel (1936) is known on Cydonia less, as my study follows the classification outlined
oblonga in Morocco. The species is largely yellow in Polhill and Raven, it will obviously be less con-
and, in view of the close relationship detectable in a fusing to follow their treatment of the single family
number of genera of species feeding on Salicaceae Leguminosae, with three subfamilies Caesalpinioi-
and Rosaceae, it may possibly be related to the deae, Mimosoideae and Papilionoideae.
populi group on Salix (see Dilleniidae above). It will Cronquist (1981) cites the number of species in
be of great interest to confirm this when a male his order Fabales as 'about 14000.' In Polhill and
becomes available for study. Raven the figures given are 650 genera and 18000
species, making the Leguminosae 'the largest family
Points of particular interest about the Agromyzidae of flowering plants after the Compositae and Orchi-
on Rosaceae are: daceae.'
The time of origin of the Leguminosae has not
1. The small number of genera (7) represented
been established with any certainty but fossil woods
2. As in the Hamamelidae, the high proportion of
and pollen indicate the presence of both Mimoso-
Agromyza species (7)
ideae and Caesalpinioideae in the Upper Creta-
3. The large number of genera (14) on which larval
ceous (Maestrichtian, 65-70 m.y.!BP). The family
feeding of Phytobia has been detected and also
may be somewhat older but it does not appear to
the relatively large number of species (6) which
have become large and diverse until the Eocene
have been reared and described
(38-54 m.y.!BP). The Papilionoideae had certainly
4. The further confirmation of the close rela-
also originated by mid-Eocene. 'There were direct
tionship between species on Betulaceae and Ro-
overland connections between Europe and North
saceae, with Phytobia amelanchieris clearly more
America until 49 m.y. BP and between Asia and
closely related to Pb. betulivora (Figs. 377, 225)
North America via Beringia, with some interrup-
than to any of the other 5 species on Rosaceae.
tions through much of the Miocene. There was un-
5. The apparently isolated position of Rubus fruti-
doubtedly an early interchange of legumes between
cosus within the genus, with Agromyza potentil-
Africa and South America. Some rodents and even
lae feeding on R. idaeus and 11 further genera
primates seem to have reached South America by
(Table 18) but never on R.fruticosus.
this route even as late as the Early Oligocene' (36
6. The presence of only a single polyphagous
m.y. BP) (from Polhill & Raven, 1981: 27-28). With
species on the family, Liriomyza trifolii.
the well-known ability of the agromyzids to cross
substantial water gaps (Spencer and Stegmaier,
1973; Spencer, 1976b), I can see no reason why an-
Order FABALES
cestral agromyzids should not have similarly dis-
Cronquist (1981) places the Fabales as a separate persed between Europe and North America, and
order within the subclass Rosidae, with three sepa- Africa and South America. Interestingly, in mid-
rate families Mimosaceae, Caesalpiniaceae and Fa- October, 1988 considerable numbers of the Desert
baceae, contrary to his earlier (1968) treatment Locust, Schistocera gregaria, crossed the Atlantic
where the Leguminosae were considered as one of from North Africa and were positively identified on
17 families in the order Rosales, with the three several islands in the Caribbean (Kevan, 1989).
universally agreed divisions treated as subfamilies. It has been generally accepted that the Legumi-
In the most exhaustive study of this group yet nosae are derived from the Rosales but more
108 Class MAGNOLIOPSIDA (Dicotyledons)
recently Thorne (1983) and Dahlgren (1986) have tic, four in the N eotropical, none in the Afrotropical
come to the conclusion that the Leguminosae are or Oriental Regions and just one in New Zealand.
more closely related to the Sapindaceae. Other Phytomyza, the largest world genus has adapted
authors suggest relationship with still further fami- poorly to the Leguminosae, with only two species in
lies. 'Since it is almost certain that no extant dicoty- the Palaearctic, four in the N earctic and none in any
ledonous family evolved directly from any other other region.
extant family, the phylogenetic origin of the legumes To date the Leguminosae are known to be colo-
is obscure' (Polhill & Raven, 1981: 39-50). nized by 106 species of agromyzids. Of these, six
I have detected no obvious affiliations between occur on the Caesalpinioideae and only one on the
family-specific agromyzids on Legumes and species Mimosoideae (plus L. schmidti), while 99 species
on other families suggested as directly ancestral to are present on the Papilionoideae. In contrast, larval
the Leguminosae but as more material from further feeding of the primitive tree-boring genus Phytobia
collecting becomes available, this is an aspect which has been detected by Suss and Muller-Stoll (1980)
deserves further study. The information which has and Suss (pers. comm.) on seven genera of Caesal-
come to light during the course of this study is pinioideae, eight genera of Mimosoideae, but on
briefly reviewed at the end of this section. only five genera of Papilionoideae. It will be of the
Fifteen genera of agromyzids (out of the world greatest interest in due course to discover the iden-
total of 30) are present on the Leguminosae and tity of these species, in particular to ascertain the
include tree-borers, gall-causers, stem-borers, stem- possible relationships between the species in Africa
miners, seed- and root-feeders, one species feeding and South America (see below) and the degree of
in young green thorns, with the majority being leaf relationship between the species on the three sub-
miners. These genera are shown in Table 19. families.
Of the 106 species of Agromyzidae on the Legu-
Table 19 minosae, 80 are restricted to a single tribe (75+%),
14 (13+%) are oligophagous (crossing tribal boun-
Agromyzid genera on Leguminosae (Fabaceae) daries) and 12 (11 +%) are polyphagous. These
figures, broken down by the 26 tribes which have
Species restricted Polyphagous
to family been colonized, are shown in Table 20. The most
highly oligophagous species in the family is Lirio-
Phytobia - (larval feeding only) myza congesta in Europe, which is known on 27
Hexomyza 2 genera in nine tribes.
Melanagromyza 23 (1 Mimosoideae)
The disparity both in numbers of species in
Ophiomyia 11 (1 Caesalpinioideae)
Tropicomyia 4 2 differing agromyzid genera on the Leguminosae (see
Kleinschmidtimyia 2 below) and of the subfamilies and tribes which they
J apanagromyza 7 have colonized appears to present a picture of
Agromyza 16 random host selection, as well as the degree of
Phytoliriomyza 3
radiation (evolution) which has occurred on indi-
Amauromyza 3
Liriomyza 14 9 vidual tribes. If the hosts as here indicated could be
Calycomyza 2 (1 Caesalp.) accurately correlated with the chemical structure of
Paraphytomyza 1 the individual genera, it is not impossible that some
Chromatomyia logical explanation of the agromyzid host selection
Phytomyza 6 could be discovered but it is doubtful whether any
94 12 such detailed analysis will ever be undertaken and
the interpretation of randomness seems to be the
only answer.
Twelve polyphagous species (out of 16 known) For ease of reference to the tribes accepted by
have colonized the family - four in Europe, six in Polhill & Raven, I have followed their numbering
North or South America, one each in Africa and throughout the discussion below. In view of the
Australia, and are known on many genera. These small numbers of species on the Caesalpinioideae
are listed together with the host genera of all other and Mimosoideae, these are discussed under the
species in Table 21. two subfamilies. The 100 species on the Papiliono-
There is remarkable disparity in the representa- ideae are better treated in the sequence of the agro-
tion of different genera in different regions. For myzid genera, and the phylogenetic arrangement of
example, Agromyza has 17 species in the Palae- the host genera is shown in Table 21.
arctic, none in the Nearctic, Neotropical, Oriental or
Australia/Pacific Regions. By contrast, Liriomyza
has eight species in the Palaearctic, ten in the Nearc-
Subclass ROSIDAE 109
Table 20
Subfamily Caesalpinioideae
1. Caesalpinieae 1
2. Cassieae 1 (? + 1) 3
3. Cercideae 1
4. Detarieae (Phytobia, larval feeding only)
Subfamily Mimosoideae
Subfamily Papilionoideae
106
Table 21
Agromyzidae on Leguminosae
Subfamily Caesalpinioideae
Tribe 10 Caesalpinieae
Caesalpinia Phytobia spo; Liriomyza caesalpiniae
Tribe 20 Cassieae
Cassia Ophiomyia cassiae (1); Tropicomyia flacourtiae, To polyphyta; L. trifolii;
Calycomyza cassiae
Tribe 30 Cercideae
Bauhinia L. schmidti
Tribe 40 Detarieae
Chidlowia Phytobia spo
Copaifera Phytobia spo
Cynometra Phytobia spo
Eperua Phytobia spo
Oxystigma Phytobia spo
Schotia Phytobia spo
Subfamily Mimosoideae
Tribe 10 Parkieae
Parkia Phytobia spo
Tribe 30 Mimoseae
AmblygonocarpusPhytobia spo
Calpocalyx Phytobia spo
Piptadenia Phytobia spo
Tetrapleura Phytobia spo
Tribe 40 Acacieae
Acacia Phytobia spo ( I, Africa ), spo ( 2; USA ), spo ( 3, Chile ), spo ( 4, Asia );
Melanagromyza acaciae
Tribe 50 Ingeae
Albizia Phytobia spo ( I, Philippines ); L. schmidti
Enterolobium Phytobia spo ( Brazil )
Subfamily Papilionoideae
Tribe 20 Sophoreae
Bowdichia Phytobia spo
Tribe 40 Dalbergieae
Dalbergia Phytobia spo
Tribe 60 Tephrosieae
Mundulea Agromyza munduleae
Tephrosia Japanagromyza tephrosiae
Wisteria Hexomyza websteri; Ao wistariae; Ao demeijerei (1)
Subclass ROSIDAE 111
Tribe 7. Robinieae
Gliricidia L. schmidti
Poissonia L. criptica nom. nud. ( Argentina), L. sativae
Robinia Phytoliriomyza robiniae; L. congesta
Tribe 8. Indigofereae
Indigofera Phytobia sp.; Mel. sojae; Kleinschmidtimyia indigoferae; L. sativae
Tribe 9. Desmodieae
Alysicarpus o. alysicarpi
Desmodium Mel. albisquama, Mel. floridensis; J. desmodivora
Subtribe DiocIeinae
Calopogonium Mel. dolichostigma, Mel. koizumii; O. centrosematis
Canavalia o. phaseoli
Subtribe Glycininae
Glycine Mel. koizumii, Mel. pseudograta, Mel. sojae, Mel. vignalis; T. vigneae;
O. centrosematis, O. phaseoli, O. shibatsuji; J. tristella; L. congesta,
L. trifolii; Chromatomyia horticola
Pueraria Mel. koizumii; O. centrosematis, O. phaseoli, O. puerarivora; J. tristella
Subtribe Clitoriinae
Centrosema O. centrosematis; J. centrosemae
Subtribe Phaseolinae
Dolichos Mel. bonavistae; O. phaseoli; T. labi, T. vigneae
Lablab Mel. bonavistae; o. spencerella; T. vigneae
Phaseolus Mel. koizumii, Mel. obtusa; Mel. phaseolivora; O. centrosematis,
O. phaseoli, o. phaseoloides, O. spencerella; T. vigneae; J. phaseoli;
L. bryoniae, L. congesta, L. strigata, L. trifolii
Vigna Mel. bonavistae, Mel. candidipennis, Mel. chaIcosoma, Mel. obtusa,
Mel. vignalis; O. centrosematis, O. phaseoli, O. spencerella, O. vignivora;
T. vigneae, T. polyphyta; J. aequalis, J. inaequalis; L. strigata, L. trifolii;
Chr. horticola
Subtribe Cajaninae
Cajanus Mel. chalcosoma, Mel. obtusa, Mel. sojae; O. phaseoli; L. sativae
Flemingia Mel. chalcosoma, Mel. sojae
Rhyncosia Mel. koizumii
Subtribe Astragalinae
Astragalus Mel. adsurgenis, Mel. astragali; Phlr. variegata; L. eongesta, L. strigata;
Chr. hortieola
Caragana Amauromyza obseura; Paraphytomyza earaganae; L. eongesta
Oxytropis Phlr. variegata; L. eongesta; Phytomyza oxytropidis; Chr. hortieola
Subtribe Genistinae
Cytisus Hex. sarothamni; A. johannae; P. cytisi
Genista Phytobia sp.; A. johannae, A. pulla
Laburnum A. demeijerei; P. cytisi
Petteria P. cytisi
Spartium A. johannae, A. luteifrons, A. pulla
Ulex A. johannae
380
379
383
381
385
384
\1 \\ '
, ~
\
" !
:
'I\~
! i
\ i (
" 390
"'~,:/
/
389 ,\,
Subclass ROSIDAE 115
"'\.~''''.'
'~""-",:,,.~ &;;w:..€io
~:;;i~'~ :;
394
391
~::t~ :::;,,···:'":··',,··4DB';'··
395
392
~'~"~~5'~~;w~-
396
~:~y,g 393
397
401
398
Subclass ROSIDAE 117
A. munduleae (Seguy, 1951) was reared from usual, being exclusively epidermal, and can easily be
leaf mines on Mundulea suberosa in Madagascar. overlooked as true mines. The male genitalia (Figs.
This species is unusual in having only a very small 391,392) indicate that this is an isolated species.
third dorso-central bristle close to the second and
the halteres are darkened, pale brownish. Unfor- Tribe 8. Indigofereae (4 genera)
tunately only the unique female holotype is known Suss (loc. cit.) has detected larval feeding of Phyto-
and the species cannot be further clarified until a bia in Indigofera argentea L. in India.
male becomes available. Kleinschmidtimyia indigoferae (ct. Spencer,
lapanagromyza tephrosiae (de Meijere, ct. Spen- 1986a) is known only in Australia on Indigofera
cer, 1961b) was described from Tephrosia sp. in suffruticosa, forming narrow, white mines (Fig. 393).
Java and has subsequently been reared from T can- The genitalia (Figs. 394, 395) confirm that this is
dida in northern India. It is now clear that it has an closely related to a second species in this small
even more extensive range and I am satisfied that genus feeding on Leguminosae, K. pisi which is
1. badia Spencer (1977c) described from N.S.W., discussed below in Tribe 21 (Vicieae).
Australia represents 1. tephrosiae and this new The genus Kleinschmidtimyia was erected for the
synonymy is formally established in the Appendix small group of minute black leaf miners which
below. The aedeagus, with exceptionally long, generally resemble Melanagromyza, Tropicomyia
paired distal tubules is shown in Fig. 386. and some Ophiomyia species but which must be
excluded from these genera on characters of the
Tribe 7. Robinieae (21 genera) genitalia, larvae or biology. New genera are needed
Only a single described species is known on Robi- for comparable groups in Africa and the Caribbean
nia, described by Valley (1982) as Liriomyza robi- area but more material needs to be studied before
niae. This is only known in the eastern United these can be formally erected. Relevant species are
States, the larvae forming blotch mines (Fig. 389) in discussed under Tropicomyia in the Appendix be-
young leaflets on the tree R. pseudoacacia. Careful low.
examination of the male genitalia satisfies me that Melanagromyza sojae which is highly oligophag-
this species does not correctly belong in Liriomyza ous but whose primary hosts are in the Phaseoleae
and it is now provisionally placed in Phytoliriomyza. (Tribe 10) has been recorded in Java on Indigofera
The unusual male genitalia are illustrated in Figs. and Swainsona. In U.S.A. the polyphagous Lirio-
387, 388 and the species is discussed further in the myza sativae has been found on Indigofera.
Appendix below.
Robinia has been widely introduced in Europe Tribe 9. Desmodieae (27 genera)
and I found short linear mines (Fig. 390), all unfor- Of the 22 Melanagromyza species known on the
tunately empty, in the leaves of young seedlings Leguminosae, 21 are found on the Papilionoideae,
growing below mature trees in Budapest, Hungary with two on Desmodium, both feeding internally on
in August, 1986. These are not referable to any the seeds in the pods. M. albisquama (Malloch, ct.
known European species and the identity must Spencer, 1973a) is known from Africa, across the
remain in doubt until adults can be obtained. Oriental Region to northern Australia and Fiji.
Two polyphagous species are known on this tribe Although it has only once been reared from Des-
in Argentina. Liriomyza sativae has been found on modium, I suspect that it may be present on other
Poissonia hypoleuca and on the same host also genera. M. floridensis Spencer (ct. Spencer and
L. criptica Valladares (nomen nudum) (see under Stegmaier, 1973) was described from D. tortuosum
Nicotiana, Solanaceae below). in Florida, the larva feeding in a single seed within
The polyphagous Liriomyza schmidti was de- the pod. It occurs widely throughout the Caribbean
scribed from Gliricidia in Costa Rica, and has been area to Guyana and Venezuela and I recently found
reared from Albizia and Bauhinia in Florida. In the it on Guadeloupe. The genitalia are somewhat atypi-
United States it has also been found on Bougainvil- cal of Melanagromyza and it is by no means certain
lea (Nyctaginaceae), Passiflora (Passifloraceae) and that these two widely-separated Desmodium feeders
Smilax sp. (Smilacaceae). A list of host families is are closely related (M. albisquama: Fig. 396; M. flo-
given in Section 5 below. The leaf mines are un- ridensis: Fig. 397).
Figs. 391, 392. Liriomyza schmidti: 391, aedeagus, side view; 392, same, ventral view.
Figs. 393-395. Kleinschmidtimyia indigoferae: 393, leaf mines on Indigofera suffruticosa; 394, aedeagus, side view; 395, same, ventral
view.
Fig. 396. Melanagromyza albisquama: aedeagus.
Fig. 397. Melanagromyzafloridensis: aedeagus.
Fig. 398. Japanagromyza desmodivora: aedeagus.
Figs.399-401. Ophiomyia alysicarpi: 399, aedeagus, side view; 400, same, ventral view; 401, leaf mine on Alysicarpus vagina lis.
118 Class MAGNOLIOPSIDA (Dicotyledons)
402
405
407
406
409
Figs. 402-404. Melanagromyza erythrinae: 402, gallon Erythrina tricocarpa; 403, aedeagus, side view; 404, same, ventral view.
Fig. 405. Melanagromyza sojae: aedeagus.
Figs. 406-408. Melanagromyza dolichostigma: 406, aedeagus; 407, anterior spiracles of larva; 408, posterior spiracles.
Fig. 409. Melanagromyza pseudograta: aedeagus.
Figs. 410, 411. Ophiomyia shibatsuji: 410, aedeagus; 411, posterior spiracles of larva.
Subclass ROSIDAE 119
Seven Japanagromyza species feed as blotch recorded on Cana valia , also in Java (see below
miners on the Papilionoideae, of which only J. des- under Phaseolinae).
modivora Spencer (Spencer and Stegmaier, 1973) is
known on D. tortuosum in Florida. This species Subtribe Glycininae
appears to be closely related to J. aequalis on Vigna Two genera, Glycine and Pueraria, provide hosts for
(Tribe 10) (male genitalia: Fig. 398). , nine species in the genera Melanagromyza, Ophio-
Ophiomyia alysicarpi (Bezzi, d. Spencer, 1977c) myia and Japanagromyza.
is an isolated species, the larva forming distinctive The primary host of Melanagromyza sojae
mines associated with the midrib (Fig. 401). It is (Zehntner, d. Spencer 1973a) appears to be Gly-
present both in tropical northern Australia and in cine but this species is highly oligophagous, re-
Papua New Guinea. The only known host is Alysi- corded also on Indigojera, Flemingia, Cajanus,
carpus vaginalis. The male genitalia (Figs. 399, 400) Aeschynomene, Medicago, Melilotus and Swain-
suggest relationship with 0. conspicua Spencer, a sona. M. sojae feeds as an internal stem-borer with
stem-feeder on Asteraceae (see below, Fig. 1203). wide host selection on five different tribes. The gen-
italia are distinctive in having the ventral bladder
Tribe 10. Phaseoleae (84 genera) (which is present in all Melanagromyza species) con-
This tribe has the largest number of genera in the spicuously bending to the right in ventral view
family, of which only 14 are known to support (Fig. 405). M. sojae is essentially an Oriental species
Agromyzidae. In all, 33 species of Agromyzidae in but it reaches Egypt and Israel. It is clearly the
the six genera Melanagromyza (11 species), Ophio- sister-species of M. cunctans, with similar external
myia (7), Tropicomyia (3), Japanagromyza (5), morphology and very similar genitalia but M. cunc-
Liriomyza (6) and Chromatomyia (1) are recorded tans is a gall-causer on Lotus (see below).
on the Phaseoleae. Of these, 19 are tribe-specific, six In Japan Melanagromyza koizumii Kato (d.
oligophagous and eight polyphagous. Spencer, 1973a) is an internal stem-borer known on
Polhill & Raven (1981) divide the tribe into eight cultivated Glycine max and the weed Rhyncosia
subtribes, of which six have been colonized by acuminatijolia (subtribe Cajaninae). It appears to be
Agromyzidae. Most of the subtribal boundaries are closely related to M. dolichostigma de Meijere (d.
crossed by a number of agromyzid species (see Spencer, 1973a), which is one of the most oligoph-
Table 21), indicating that the subtribes are weakly agous species in this group, known in Java and
differentiated, at least in relation to their chemistry, Taiwan. Main feeding takes place in the stem but
thus ensuring wide acceptability as hosts. oviposition is in the leaf. Hosts recorded include
Ca/opogonium, Glycine, Pueraria, Phaseolus and
Subtribe Erythrininae Crotalaria in tribe 29 (Crotalarieae). The male gen-
Suss (loc. cit.) has detected larval feeding of Phy- italia (Fig. 406) are characteristic of Melanagromyza
tobia in Erythrina lithospermum Blume in Java. but both anterior and posterior spiracles are modi-
Only a single host-specific species is known on fied, the anterior ones being exceptionally long (Fig.
Erythrina, Melanagromyza erythrinae (de Meijere, 407), while the posterior spiracles (Fig. 408) each
1910), also in Java. This is a species with unusual have an ellipse of some 12 pores, with no trace of
biology, the larva forming leaf galls on E. tricocarpa the central horn normal in the genus. In Australia
(Fig. 402). From the male genitalia (Figs. 403, 404) only a single species of Melanagromyza, M. pseudo-
it might be suggested that this species is related to grata Spencer (1977c), is known on the Papilionoi-
M. obtusa which is a pod-feeder primarily on Ca- deae, recorded from Glycine clandestina, the larva
janus (see below and Fig. 434). A lectotype of feeding in the pod. It closely resembles M. albis-
M. erythrinae is designated in the Appendix below quama, a pod-feeder on Desmodium, but the male
where the species is discussed further. genitalia of the two show that they are not closely
Two polyphagous species have colonized Ery- related (M. albisquama: Fig. 396; M. pseudograta:
thrina, Tropicomyia polyphyta in Australia and Fig. 409).
Liriomyza schmidti on Guadeloupe. L. schmidti has Ophiomyia shibatsuji (Kato, d. Spencer, 1973a)
also been recorded on Mucuna (as Stizolobium) in is known only in Japan and feeds exclusively in the
Florida. stem and root of Glycine (male genitalia: Fig. 410
and posterior spiracles: Fig. 411). 0. puerarivora
Subtribe Diocleinae Sasakawa (19 81 b) also in Japan feeds in the pods of
Two oligophagous species, Melanagromyza dolicho- Pueraria lobata. Sasakawa gives many illustrations
stigma and Ophiomyia centrosematis (discussed but it is not possible to decide on the affiliations of
further below) are known on Calopogonium in Java, this species. 0. phaseoli, a major pest on beans
and M. koizumii attacks this host in Japan. Another throughout the Old World tropics, with its range
highly oligophagous species, 0. phaseoli, has been extending to Egypt and Israel and tropical northern
120 Class MAGNOLIOPSIDA (Dicotyledons)
...,;,~,'>..
ICf/l~f"~ .~
.::. ,/ ""./:'
412 413
' .. . ,
414 416
. '-.-.--- . ~- -"
417
415
,:
419 .
"\:'';'
418
Subclass ROSIDAE 121
Australia, occurs commonly on Glycine but is highly M. bonavistae Greathead (1971) is only known in
oligophagous and is discussed further below under East Africa on Dolichos, Lablab and Vigna. Sur-
Phaseolus. prisingly it appears to be the sister-species of M. vig-
Japanagromyza tristella (Thomson, d. Spencer, naZis Spencer (d. Spencer, 1973a) which occurs
1973a) forms large blotch mines on both Glycine widely in Africa from Sudan to South and West
and Pueraria and is restricted to this subtribe. It is Africa and is only known as a pod-feeder on Glycine
widely distributed from India to Japan and Indone- and Vigna. The male genitalia (M. bonavistae: Fig.
sia. The male genitalia (Fig. 412) are distinctive and 418 and M. vigna lis: Fig. 419) confirm their close
the larval posterior spiracles are of interest, with the relationship.
primitive three pores each further subdividing into M. chalcosoma Spencer (d. Spencer, 1973a)
16-18 secondary pores (Fig. 413). This arrange- occurs widely in Africa, attacking the pods of Vigna,
ment is found in other species in the genus, such as Cajanus and Flemingia, while M. obtusa Malloch
1. quercus and J. viridula (d. Figs. 223, 222) (Fa- (d. Spencer, 1973a) has similar biology, with Vigna,
gales, Hamamelidae). Cajanus and Phaseolus as hosts in India and else-
where in the Oriental Region. The main host of both
Subtribe Clitoriinae species appears to be Cajanus (see subtribe Caja-
The oligophagous Ophiomyia centrosematis (de ninae, below).
Meijere, d. Spencer, 1973a), which was described On Phaseolus M. phaseolivora (Spencer, 1973a)
from Centrosema pubescens in Java, forms external is only known as a pod-feeder on P. vulgaris in
stem-mines and has been recorded in addition on Ecuador and Colombia. The male genitalia (Figs.
Calopogonium, Glycine, Phaseolus, Vigna and Cro- 420, 421) show that this species is not closely
taZaria, being most common on the latter genus. The related to M. floridensis, a pod-feeder on Desmo-
distinctive male genitalia are shown in Figs. 414, dium.
415 and the posterior pupal spiracles in Fig. 416. Two serious Ophiomyia pest species, 0. phaseoli
Japanagromyza centrosemae (Frost, d. Spencer, (Tryon, d. Spencer, 1973a) and 0. spencerella
1973a) is known on Centrosema pubescens in Pana- Greathead (d. Spencer, 1973a) are oligophagous
ma. Recent examination of the unique holotype, (see Table 21) but occur most commonly on Pha-
which was described as a male, has shown it to be a seolus. The larvae of both feed mainly in the lower
female and the affiliations of the species cannot be stem and root. The two species are remarkably
further clarified. similar on external characters, both even having the
unusually long, narrow and brilliantly shining ocellar
Subtribe Phaseolinae triangle but the male genitalia of the two are entirely
This is by far the most acceptable subtribe to agro- distinct (0. phaseoli: Fig. 422 and 0. spencerella:
myzids, supporting 17 species in the genera Melana- Fig. 423). Also the puparium is whitish in 0. pha-
gromyza, Ophiomyia, Tropicomyia and Japanagro- seoli, black in 0. spencerella. Both the geographical
myza. Surprisingly only four of 23 genera in the and host range of the two species also differ con-
subtribe are known to be colonized - Dolichos, siderably. 0. spencerella is only known in East and
Lablab, Phaseolus and Vigna. Of the 11 species re- West Africa and the only hosts known are Phaseo-
stricted to the subtribe, four are host-generic on Ius, Lablab and Vigna. On the other hand 0. pha-
Vigna, three on Phaseolus and one on Dolichos. seoli occurs throughout the Old World tropics,
Of the seven Melanagromyza species, two feed in reaching Egypt and Israel, and extending eastwards
stems and five in pods. M. candidipennis (Lamb, d. across Asia to Pacific islands and northern Aus-
Spencer, 1985a) was described from the Seychelles tralia. Larval feeding is known on Glycine, Cajanus,
and is widespread in Africa, but uncommon. The Canavalia and Crota laria , apart from Dolichos,
host has been discovered in Nigeria, with the larva Phaseolus and Vigna. The present direct competi-
feeding and pupating in the stem of Vigna unguicu- tion in East Africa on the same hosts, with the larvae
lata (cow-pea). It is an aberrant species, the adult feeding in essentially the same way, was earlier
having largely colourless wings and veins and the (Spencer, 1973a: 66) explained by the original isola-
male genitalia (Fig. 417) confirm its isolated posi- tion of 0. spencerella in East Africa, followed by re-
tion.
Figs. 412, 413. Japanagromyza tristella: 412, aedeagus; 413, posterior spiracles of larva.
Figs. 414-416. Ophiomyia centrosematis: 414, aedeagus, side view; 415, same, ventral view; 416, posterior spiracles of puparium.
Fig. 417. Melanagromyza candidipennis: aedeagus.
Fig. 418. Melanagromyza bonavistae: aedeagus.
Fig. 419. Melanagromyza vignalis: aedeagus.
Figs. 420, 421. Melanagromyza phaseolivora: 420, aedeagus, side view; 421, same, ventral view.
122 Class MAGNOLIOPSIDA (Dicotyledons)
;1jY ."
422
, ,
Jj
423
424 425
~...........•..........•......
'L.(JA'~~~:5 /. ~
. . ·. . ~·A
~<~
........ : ..
426
~":,", ..
427
,',
428
Fig. 422. Ophiomyia phaseoli: aedeagus.
Fig. 423. Ophiomyia spencerella: aedeagus.
Figs. 424, 425. Ophiomyia phaseoloides: 424, aedeagus; 425, posterior spiracles of puparium.
Figs. 426, 427. Ophiomyia vignivora: 426, aedeagus; 427, same, ventral view.
Fig. 428. Japanagromyza inaequalis: aedeagus.
Fig. 429. Japanagromyza aequalis: aedeagus.
Subclass ROSIDAE 123
colonization of Africa from Asia by the now distinct 330), the larval posterior spiracles (Fig. 432) have
0. phaseoli. more pores than the usual three in Tropicomyia and
An interesting new leaf miner described as even the male genitalia do not agree exactly with the
0. phaseoloides Martinez (in Spencer et aI., 1990) form typical of Tropicomyia (Figs. 278, 279). How-
has recently been discovered on Guadeloupe, form- ever, for the time being this species is best left in
ing long, narrow mines on Phaseolus vulgaris, with Tropicomyia (see Appendix below).
up to ten mines in a leaf. The larva pupates on the
lower leaf surface, the puparium remaining in the Subtribe Cajaninae
mine. The complex genitalia (Fig. 424) and the There are no host-specific species on genera on this
distinctive posterior puparial spiracles (Fig. 425) subtribe but two important pod-feeders, M. chal-
give no clear indications of the affiliations of this cosoma in Africa and M. obtusa in the Oriental
species. It is one of a complex of diminutive black Region feed primarily on Cajanus, both also on
leaf miners in the Caribbean area which probably Vigna, the former also on Flemingia and the latter
deserve separate generic status when more informa- also on Phaseolus. The genitalia (M. chalcosoma:
tion becomes available (see Appendix under Tropi- Fig. 433 and M. obtusa: Fig. 434) show their close
comyia). Attention is drawn to the striking similarity relationship. In addition, M. koizumii (main host
of the posterior spiracles between 0. phaseoloides Glycine) has been recorded in stems of Rhyncosia
and Tropicomyia crotalariae (Fig. 508), a leaf miner acuminatifolia in Japan and the internal stem-borer,
on Crotalaria in South Africa. This can only be M. sojae, which feeds primarily on the Phaseoleae
interpreted as a remarkable case of convergence. (Tribe 10), has been recorded on Cajanus indicus
Only a single Ophiomyia, 0. vignivora Spencer and Flemingia in Java.
(1973a) from Pakistan, is either a leaf miner or
external stem-miner on Vigna unguiculata. It be- Tribe 14. Aeschynomeneae (25 genera)
longs to a small group which occurs mainly in the Two species are known to have colonized this tribe
southern hemisphere with the squamal fringe white, - Melanagromyza sojae, whose primary host is
not black as is normal in species in this genus. The Glycine, but which has been found in stems of
male genitalia are shown in Figs. 426, 427. Aeschynomene indica in Java and Liriomyza trifolii
Three lapanagromyza species are known on this which has been found on Arachis in South America.
subtribe, all forming blotch mines. 1. inaequalis
(Malloch, cf. Spencer, 1973a) and 1. aequalis Spen- Tribe 16. Galegeae (20 genera)
cer (cf. Spencer, 1973a) are not uncommon on Six of the 20 genera in this tribe are known to
Vigna luteola in the Caribbean area and reach support Agromyzidae, with a total of 12 species in
southern Florida. These two are not closely related. eight genera. Of these, five are tribe-specific, five
The genitalia of 1. inaequalis (Fig. 428) show that it oligophagous, two polyphagous.
is most closely related to 1. phaseoli (see below),
while 1. aequalis (Fig. 429) may well be the sister- Subtribe Coluteinae
species of 1. desmodivora (see above). In Europe Phytoliriomyza variegata (Meigen, cf.
1. phaseoli Spencer (1983 and sp. (Peru), 1973c) Spencer, 1976a) occurs most frequently on Colutea
is widespread from Costa Rica to Peru feeding on arborescens but also commonly on Astragalus in
Phaseolus vulgaris, forming large blotch mines. The subtribe Astragalinae, less often on Coronilla (Tribe
genitalia (Fig. 430) suggest close relationship with 20) and Lathyrus (Tribe 21). The hosts of few Phy-
1. inaequalis on Vigna. toliriomyza species are known and the affiliations of
Tropicomyia labi (Singh and Ipe, 1971) has been the species are not clear, even from the male gen-
reared from Dolichos lablab at Agra, India but no italia (Figs. 435, 436). In New Zealand Liriomyza
genitalia illustrations are available and the species clianthi Watt, cf. Spencer, 1976b) (Figs. 437, 438) is
cannot be further clarified. It is tentatively accepted commonly found on the cultivated shrub Clianthus
as distinct from T. polyphaga which occurs on many puniceus but see also Carmichaelia (Tribe 17).
families in India but is not known on the Papilionoi- Empty leaf mines I found on Swainsona galegifolia
deae. Thus T. labi may well be host-specific on at the Botanical Gardens, Canberra were referred to
Dolichos. Kleinschmidtimyia indigoferae (see Tribe 8) but this
T. vigneae (Seguy, cf. Spencer, 1973a) is an identification requires confirmation.
oligophagous species known from the Cape Verde
Islands to Indonesia feeding on Dolichos, Glycine, Subtribe Astragalinae
Lablab, Phaseolus and Vigna. It is a somewhat Astragalus is considered to be the largest genus in
aberrant species and larval mining is only in part the angiosperms with 2000 species but it supports
epidermal, the mouth-parts (Fig. 431) are not whol- only six agromyzids, three of which are polyphag-
ly serrated which is normal in Tropicomyia (Fig. ous. Melanagromyza astragali Spencer (1976a) is
124 Class MAGNOLIOPSIDA (Dicotyledons)
431
/"
/'
«,
~"\'
433
434
437
~,~:t::t ~
436 438
Subclass ROSIDAE 125
one of only five species in this genus known on the Tribe 18. Hedysareae (7 genera)
Papilionoideae in the Palaearctic Region, present in Agromyza nana is oligophagous, feeding occa-
Sweden as a stem-borer in Astragalus glyciphyllus. sionally on Onobrychis viciifolia, although its fa-
The male genitalia are typical of many species and vourite hosts are in the Trifolieae (see below, Tribe
give no indication of its affiliations. M. adsurgenis 23). Liriomyza congesta is also known here.
Wenn (1985) was reared from stems of Astragalus
adsurgens in Hanan Province, China and the gen- Tribe 19. Loteae (4 genera)
italia (Figs. 439, 440) show that it is well differen- Only one host-specific species is known on this
tiated from M. astragali but there may well have tribe. Melanagromyza cunctans (Meigen, cf. Spen-
been only a single colonization of this unusual host. cer, 1966b) forms oval galls on the upper stem of
Phytoliriomyza variegata crosses the subtribal Lotus corniculatus in Europe (Fig. 444) and the
boundary and is frequently present on Astragalus genitalia (Figs. 445, 446) are unusual in having the
and Oxytropis, as well as on Colutea (see above). ventral bladder conspicuously bending to the right
Caragana has two host-specific species. Amauro- in ventral view as in M. sojae on Glycine (see Fig.
myza obscura (Rohdendorf-Holmanova, d. Spencer 405). The two are clearly sister-species with a wide
and Martinez, 1987: 256) is host-specific on C. pu- jump in hosts between tribes, and the differing
bescens. This little black species can scarcely be biology represents a significant evolutionary differ-
distinguished on external characters from Am. cha- entiation.
maebalani Hering which feeds on Lathyrus and Agromyza nana, which feeds most commonly on
Pisum (see below) but its genitalia are entirely Trifolieae (Tribe 23), has also been found on An-
distinct (Fig. 441) and strongly differentiated from thyllis.
those of Am. chamaebalani (Fig. 471); on the same Liriomyza cicerina (Rondani, ct. Spencer, 1973a)
host the only Paraphytomyza known on the Legu- which normally feeds on Cicer (see Tribe 22) and
minosae, Pa. caraganae Rohdendorf-Holmanova also on Ononis (Tribe 23) has once been found on
(1959b) forms blotch mines and was described from Hymenocarpus circinnatus in Yugoslavia.
Khirgizia, USSR, in central Asia. I have also seen Phytomyza brischkei Hendel (cf. Hendel, 1935)
specimens from Irkutzk, Siberia (colI. Buhr) and the which is one of two species in this genus in Europe
genitalia of a male of this series are shown in Fig. feeding on Leguminosae (also P. cytisi on Labur-
442. Pa. caraganae closely resembles species in the num) was described from Trifolium. It has now been
populi-group (feeders on Salicaceae) but the gen- discovered that two further species described on
italia are unique within the genus, providing no Anthyllis, P. anthyllidis Hering (1954) (valid larval
evidence for direct relationship with this group. It description antedating description of adult by
is also generally similar to Pa. cydoniae Hendel Groschke (1957), ct. Henshaw and Howse, 1989), a
(1936) (Rosaceae) but with only females known of leaf miner, and P. vulnerariae Spencer (1957c),
this species, the possible relationship of the two found only mining sepals on Anthyllis vulneraria,
remains wholly speculative. represent synonyms of brischkei. Hering (1957b)
A single Phytomyza is known on Oxytropis, was misled into accepting these species as distinct
P. oxytropidis Sehgal (1971b) in Alberta, Canada. owing to the different host genus from brischkei and
The species forms linear mines, unusually in this the fact that vulnerariae is known only from sepals.
group pupating in the mine. The spinules on the These two new synonymies are formally established
basal sclerites of the aedeagus (Fig. 443) associate in the Appendix below. The distinctive genitalia of
this species with the spondylii-group on Apiaceae brischkei (Fig. 447) show that this species is clearly
and the albiceps-group on Asteraceae (see below). related to P. ranunculivora (Fig. 160) and also to
P. sedicola (Fig. 350). There have thus been signifi-
Tribe 17. Carmichaelieae cant switches in hosts on three families by this
Liriomyza clianthi, whose primary host is Clianthus group.
(Tribe 16) has also been found on North Island,
N.z. on Carmichaelia williamsii and on South Island Tribe 20. Coronilleae (6 genera)
on C. grandiflora. No host-specific species are known on this tribe but
442
443 444
,?,:.;-%~~r~,
/
447
446 .;/i~'!;~
.~~"\'
~
450
448
451
Fig. 441. Amauromyza obscura: aedeagus.
Fig. 442. Paraphytomyza caraganae: aedeagus.
Fig. 443. Phytomyza oxytropidis: aedeagus (Sehgal).
Figs. 444-446. Melanagromyza cunctans: 444, gall on Lotus corniculatus; 445, aedeagus, side view; 446, same, ventral view.
Fig. 447. Phytomyza brischkei: aedeagus.
Figs. 448, 449. Melanagromyza azawaii: 448, aedeagus, side view; 449, same, ventral view.
Figs. 450, 451. Melanagromyzafabae: 450, aedeagus, side view; 451, same, ventral view.
Subclass ROSIDAE 127
both Phytoliriomyza variegata and Liriomyza con- I think it can be accepted that the most primitive
gesta in Europe and L. trifoliearum in the U.S.A. species is A. erythrocephala Hendel (d. Spencer,
have been recorded on Coronilla, with L. congesta 1976a) which forms galls in the stem of Vicia cracca
also known on Hippocrepis, Ornithopus and Scor- and other Vicia spp. This is the largest species in the
piurus. group, with wing length up to 4 mm and the orbits
are strongly projecting above the eye (Fig. 459).
Tribe 21. Vicieae (5 genera) However the genitalia (Fig. 460) are characteristic
This is a highly acceptable tribe, colonized by the of the group. A. marionae Griffiths (1963a), known
following nine genera of Agromyzidae, with a only in England, forms external stem mines on Vicia
total of 33 species (number of species indicated cracca but only a single female has been reared.
in brackets): Melanagromyza (3), Ophiomyia (1), Four host-specific Agromyza leaf miners on Vicia
Kleinschmidtimyia (1), Agromyza (10), Amauro- all form blotch mines but in A. bicophaga Hering
myza (2), Phytoliriomyza (2), Liriomyza (12), Chro- (1925a) and A. vicifoliae Hering (1932) the first
matomyia (1), Phytomyza (1). Four of the five host instar larva forms a short linear mine which may
genera support Agromyzidae, only the mono typic become largely invisible within the later blotch but
genus Varilovia, present from Lebanon and Turkey can frequently be detected by the arrangement of
to Caucasia, is not known to be attacked by agro- frass.
myzids. The favourite genus is Vicia (see Table 21). A. bicophaga was reared from Vicia tetrasperma
Twenty of the 33 species are restricted to the tribe, near Berlin. In this species the genitalia are well
six are oligophagous and seven polyphagous. differentiated (Figs. 461, 462) without the separate
Melanagromyza azawii Spencer (1973a) is only areas of pigmentation in the distiphallus present in
known in Iraq, feeding primarily in the stems of other members of the group. The genitalia of A. vici-
Vicia faba. It has also been reported at the same foliae (Fig. 463) are characteristic of the group.
locality near Baghdad in stems of Sesamum indicum A. felleri Hering (1941) is found most frequently
(Pedaliaceae, Scrophulariales, Subclass 5) at the on Vicia sepium. It most closely resembles A. de-
same locality. If these records are correct, this repre- meijerei on Laburnum (Tribe 32, Genisteae) and
sents an unusual case of host transference. The male the genitalia (Fig. 464) suggest that the two may be
genitalia are shown in Figs. 448, 449. M. fabae sister-species. A. viciae Kaltenbach (cf. Hendel,
Spencer (1973a) is known only on Vicia faba in 1931) forms a primary blotch mine. I have not been
England where the larva feeds and pupates within able to locate a male in any museum and the affilia-
the root. The male genitalia (Figs. 450, 451) are tions of the species cannot be established.
distinctive but typical of the genus. In China, Gansu A. nana Meigen (cf. Spencer, 1976a) has a wide
Province, M. viciae Wenn (1985) was also reared host range, occurring only rarely on Vicia and is
from stems of Vicia faba. The genitalia (Figs. 452, most common on Medicago (Tribe 23, Trifolieae).
453) suggest close relationship with M. fabae. Lathyrus has three host-specific Agromyza spe-
The only Ophiomyia known on the tribe, 0. orbi- cies, one of which may occur also on Pisum, apart
culata (Hendel, cf. Spencer, 1973a) appears to be from oligo- and polyphagous species. In Europe
strictly host-specific on Pisum, the larva feeding in Agromyza lathyri Hendel (cf. Spencer, 1976a) forms
the stem, pupating in the lower stem or root. The a large, whitish blotch mine on both Lathyrus and
genitalia (Fig. 454) suggest close relationship with Pisum (Fig. 465). The posterior spiracles of the
the Old World tropical species 0. centrosematis larva are highly differentiated, with up to 40 spira-
(Fig. 414) but the pupal spiracles (Fig. 455) with cular pores on each process (Fig. 466) but the gen-
17-22 bulbs suggest possible relationship with italia (Fig. 467) are typical of the group. A. varicor-
o. heracleivora on Heracleum (Fig. 594). nis Strobl (cf. Hendel, 1931) (= A. watersi Spencer,
Kleinschmidtimyia pisi (Kleinschmidt, cf. Spen- 1957b) feeds on Lathyrus sylvestris and L. latifolius,
cer, 1986a) occurs widely in Australia on Pisum mining initially in the leaf but mainly in the winged
sativum and Trifolium repens (both introduced stem. This species closely resembles A. lathyri. The
plants and its endemic host has not yet been dis- genitalia (Fig. 468) are of similar form but the larvae
covered). The male genitalia are shown in Figs. 456, are differentiated by having only the basic three
457 and the larval mouth-parts in Fig. 458, which posterior spiracular pores. A. orobi Hendel (cf.
are normal, that is without the serration present in Spencer, 1976a) is strictly monophagous on Lathy-
the other related leaf-mining genus Tropicomyia rus vern us, forming distinctive blotch mines and the
(Fig. 330). K. pisi is closely related to K. indigoferae larval spiracles are as in A. varicornis. The genitalia
discussed under Tribe 8. (Fig. 469) are typical of the group.
Agromyza is the dominant genus on the Vicieae, Two Phytoliriomyza species, which are not di-
with ten species, seven on Vicia and three on rectly related have colonized Vicia. PI. variegata
Lathyrus. occurs most commonly on the Galegeae (Tribe 16)
128 Class MAGNOLIOPSIDA (Dicotyledons)
'~;.")-::J
452 453
455
454
456
458
457
~/
459 460
Figs. 452, 453. Melanagrornyza viciae: 452, aedeagus, side view; 453, same, ventral view (Wenn).
Figs. 454, 455. Ophiornyia orbiculata: 454, aedeagus; 455, posterior spiracles of puparium.
Figs. 456-458. Kleinschrnidtirnyia pisi: 456, aedeagus; 457, same, ventral view; 458, mouth parts and cephalo-pharyngeal skeleton.
Figs. 459, 460. Agrornyza erythrocephala: 459, head; 460, aedeagus.
Subclass ROSIDAE 129
F~;~': -~';r'
~--.
.
. ~; . ::.""'"
. - . ~ , (""--"-~(--
"
461
463
464
465
467
~ -.:
469
Figs. 461, 462. Agromyza bicophaga: 461, aedeagus, side view; 462, same, ventral view,
Fig. 463. Agromyza vicifoliae: aedeagus.
Fig. 464. Agromyza feUeri: aedeagus.
Figs. 465-467. Agromyza lathyri: 465, leaf mine on Pisum (Hering); 466, posterior spiracles of larva (Hering); 467, aedeagus, ventral
view.
Fig. 468. Agromyza varicornis: aedeagus, side view.
Fig. 469. Agromyza orobi: aedeagus.
130 Class MAGNOLIOPSIDA (Dicotyledons)
471
470
472
~
. . . . . . . . :.<i: ... ' _________
.\"~':'>. "' . . . . . . . . . . . ~=;::-
. : .t·· '.".; ,,"
473 476
~"~"')u
478 477
479
Fig. 470. Phytoliriomyza viciae: aedeagus.
Fig. 471. Amauromyza chamaebalani: aedeagus.
Figs. 472, 473. Amauromyza lathyroides: 472, aedeagus, side view; 473, same, ventral view.
Figs. 474, 475. Liriomyza persica: 474, aedeagus, side view; 475, same, ventral view.
Figs. 476, 477. Liriomyza lathyri: 476, aedeagus, side view; 477, same, ventral view (Sehgal).
Fig. 478. Liriomyza pisivora: aedeagus.
Figs. 479, 480. Liriomyza congesta: 479, aedeagus, side view; 480, distiphallus, ventral view.
Subclass ROSIDAE 131
but has also been found on Vicia. However, Phlir. related to L. congesta in Europe than to L. trifolii in
viciae Spencer (1969b), known only in Canada on North America. Its hosts include Lathyrus and Vicia
V. cracca, represents a remarkable case, with ob- (and in the Trifolieae Medicago and Trifolium). Its
vious relationships with the two fern feeders, Phlir. distribution is limited, being only known in Califor-
clara (North America) and Phlir. hi/are/la (Europe), nia, Washington and more recently Wisconsin.
both feeding exclusively on Pteridium aquilinum L. trifoliearum has been recorded on Pisum (also
(male genitalia: Fig. 470). This suggests a most on Coronilla, Tribe 20) but its primary hosts are be-
unusual ancestral switch from Pteridium to Vicia. lieved to be Medicago and Trifolium (see below).
Two Amauromyza species are known on the Of the nine polyphagous species attacking the
Vicieae, one each in Europe and California. Vicieae (see Table 21), Liriomyza xanthocera is not
Am. chamaebalani (Hering) (= soosi Zlobin, d. uncommon in W. Germany on Vicia faba and
Spencer and Martinez, 1987: 268) forms blotch L. huidobrensis, which occurs commonly on Aste-
mines on Lathyrus tuberosus and Pisum sativum raceae (see Tribe Anthemideae), was a particularly
and has been confused with Am. obscura (Rohden- important pest on Pisum in California in the 1950s
dorf-Holmanova, 1959b), whose host is Caragana and was given the official American name 'Pea Leaf
pubescens (see Tribe 16). The genitalia of the two Miner.' For reasons which are not clear, it is now
are entirely different (Am. chamaebalani: Fig. 471 considered to be of little economic importance on
and Am. obscura: Fig. 441) showing that they are this host. L. cordobensis Valladares (nomen nudum)
not closely related despite their superficial resem- is a recently discovered species in Argentina, feed-
blance. Am. lathyroides Spencer (1981), known ing on Thalictrum and Tropaeolum, in addition to
only on Lathyrus vestitus in central California, is an Lens and Pisum. In northern India Chromatomyia
isolated species (male genitalia: Figs. 472, 473) horticola is a serious pest on Pisum.
possibly related to the oligophagous Asteraceae- Surprisingly, the Leguminosae are unattractive to
feeder Am. maculosa (d. Fig. 1045), with an early Phytomyza, the largest world genus, with only six
switch from Leguminosae to Asteraceae. species known on the family, and a single species
The genus Liriomyza has successfully colonized known on the Vicieae. P. subtilis Spencer (1969b)
the Leguminosae, with seven species in the Palae- forms blotch mines on Lathyrus ochroleucus, L. ves-
arctic Region and ten in North America. Two host- titus and Vicia americana in Alaska, Alberta and
specific species are known on the Vicieae. L. persi- California. The male genitalia (Fig. 483) suggest
ca Griffiths (1964a) is known only in northern Iran possible relationship with species on Ranuncula-
on Vida cracca and is clearly an isolated species, ceae, such as P. aquilegiana (Figs. 173, 174) on
with the third antennal segment black and the outer Aquilegia or P. plumiseta (Figs. 193, 194) on Tha-
cross-vein lacking. New illustrations of the genitalia lictrum, and it appears there must have been an
are shown in Figs. 474, 475. L. lathyri Sehgal early ancestral switch between the Ranunculaceae
(1971b) is only known from blotch mines on and Leguminosae.
Lathyrus ochroleucus at Edmonton, Canada. The
genitalia (Figs. 476, 477) suggest closer relationship Tribe 22. Cicereae (1 genus)
with several species on Asteraceae than on Legu- Four species have been recorded on Cicer. Ophio-
minosae. myia cicerivora Spencer (1961b) is only known in
In Europe Liriomyza pisivora Hering (1957b) Pakistan, forming stem mines on Cicer sp. (gram).
forms largely lower surface mines on Pisum sativum The head (Fig. 484), with the well-developed vibris-
(also on other Pisum spp. in Botanical Gardens) and sal fasciculus, is characteristic of the genus, as are
on Lathyrus spp. Externally this species cannot be both genitalia (Fig. 485) and the pupal posterior
reliably distinguished from L. congesta but the male spiracles (Fig. 486). Liriomyza cicerina (Rondani, d.
genitalia (Fig. 478) are distinctively different and in Spencer, 1973a) feeds mainly on Cicer arietinum
the larva there are eight pores on each posterior (chick pea) in the Mediterranean area, on which it
process, three only in L. congesta. can be a serious pest. The genitalia (Figs. 487, 488)
Three oligophagous Liriomyza species occur show that it is related to species on Asteraceae, for
commonly on the Vicieae. In Europe L. congesta example L. sonchi Hendel (d. Figs. 990, 991). It
(Becker, d. Hendel, 1931) is common on Lathyrus, also occurs on Ononis spp. (Tribe 28) and has once
Pisum and Vicia, and has been recorded on Lens; it been found on Hymenocarpus (Tribe 19).
is also equally common on the Trifolieae (Tribe 23). The oligophagous species L. congesta and the
The male genitalia (Figs. 479, 480) show that it is common polyphagous species Chromatomyia hor-
closely related to L. fricki Spencer (d. Spencer and ticola are also found here.
Steyskal, 1986) (which was earlier confused with
L. trifolii, see Tribe 23). The genitalia (Figs. 481, Tribe 23. Trifolieae (7 genera)
482) clearly suggest that L. fricki is more closely Seven genera of Agromyzidae - Melanagromyza
132 Class MAGNOLIOPSIDA (Dicotyledons)
,~~:~- .....
482
481
483 484
485
,{ ,".
488
487
489 490
-..-~-.-.:: ;::,.:.::.~
Figs. 481, 482. Liriomyza fricki: 481, adeagus, side view; 482, distiphallus, ventral view.
Fig. 483. Phytomyza subtilis: aedeagus.
Figs. 484-486. Ophiomyia cicerivora: 484, head; 485, aedeagus; 486, pupal spiracles.
Figs. 487, 488. Liriomyza cicerina: 487, aedeagus, side view; 488, distiphallus, ventral view.
Fig. 489. Melanagromyza gibsoni: aedeagus.
Figs. 490, 491. Ophiomyia ononidis: 490, aedeagus; 491, posterior spiracles of puparium.
Subclass ROSIDAE 133
(2), Ophiomyia (1), K[einsehmidtimyia (1), Agromy- found commonly on Medicago, Melilotus and Tri-
za (2), Liriomyza (9), Chromatomyia (1) and Phyto- folium, less frequently on Trigonella, with the whit-
myza (1), with in all 17 species are represented on ish blotch mine always associated with the midrib.
this tribe, of which only three are tribe-specific. Five A. nana is oligophagous, found also on tribes 18
of the seven genera assigned to the tribe have been (Hedysareae), 19 (Loteae) and 21 (Vicieae). The
colonized: Medieago, Melilotus, Ononis, Trifolium male genitalia (Fig. 494) closely resemble those of
and Trigonella. A. Jrontella. Interestingly, A. nana has managed to
Melanagromyza gibsoni (Malloch, cf. Spencer and extend its range through or round the Himalayas to
Steyskal, 1986) is an internal stem-borer in Medi- Delhi and Agra in India where it has been found
eago sativa. It is widespread in the southern United feeding on Melilotus.
States and is also present in Chile. Externally this Eight Liriomyza species have colonized this tribe,
species is virtually indistinguishable from M. floris mainly on the genera Medieago, Melilotus and Trifo-
Spencer, an oligophagous seed-feeder on Astera- lium, all being oligo- or polyphagous.
ceae (see below) but the male genitalia (M. gibsoni: L. congesta (Becker, cf. Spencer, 1976a) is wide-
Fig. 489 and M. floris: Fig. 1218) show that the two spread in Europe, feeding commonly on Medieago
are not directly related. Ophiomyia ononidis (Spen- but its main hosts appear to be in the Vicieae (see
cer, 1966b) feeds exclusively on Ononis spp., form- above). L. fricki is known in U.S.A. on Medicago
ing an external stem-mine. The adult closely re- and Trifolium and is discussed above under the
sembles 0. curvipalpis (Zett.), an oligophagous Vicieae, where it attacks Lathyrus and Vida. Al-
feeder on Asteraceae (see below), and the two though the European species L. deerina feeds pri-
cannot be separated on external characters. How- marilyon Cicer (see Tribe 22), it is not infrequently
ever, the differences between the two species are also found on Ononis. Medicago is the favourite
clearly reflected in the genitalia (Fig. 490) and also host of L. trifoliearum Spencer (cf. Spencer and
in larval characters, with the posterior spiracles in Stegmaier, 1973), which also feeds on Trifolium,
0. ononidis having 6-7 pores on each process (Fig. and on Tribe 20 (Coronilla) and Tribe 21 (Pisum).
491), while 0. curvipalpis retains the plesiomor- It occurs widely in U.S.A. and also in Canada. The
phous state of three. genitalia (Figs. 495, 496) suggest that it is an isolated
The highly oligophagous Melanagromyza sojae, species.
who1>e primary hosts are in the Phaseoleae (Tribe Two most important pest species, which are both
10), has been recorded in Egypt on Medicago and in highly polyphagous, were described from Medieago
India on Melilotus. In Australia Kleinsehmidtimyia sativa (L. sativae) and Trifolium repens (L. trifolii).
pisi has been found on Trifolium repens, in addition L. trifolii Burgess (cf. Spencer, 1973a) was de-
to its most frequent host, Pisum. scribed from Washington, D.C. and is now known as
AgromyzaJrontella (Rondani, cf. Spencer, 1976a) one of the most polyphagous species in the Agromy-
is tribe-specific, feeding primarily on Medicago but zidae, with records from 25 families (see Section
also on Melilotus. The genitalia (Figs. 492, 493) are 5). It has also developed a greater degree of resist-
of similar form to species on the Vicieae and Genis- ance to insecticides than any other species and has
teae (cf. Figs. 463,522) and the linear-blotch mine is been introduced widely to many countries, primarily
comparable to that of A. nana and A. vieifoliae. A with infested chrysanthemum cuttings. L. sativae
new synonym of this species has been discovered Blanchard (cf. Spencer, 1973a) was described from
during the course of this work. Hering (1930) de- Medicago sativa in Argentina and is widespread in
scribed A. drepanura from two specimens reared much of North and South America. Although these
from Medicago in the southern U.S.S.R. Unfortu- two species differ in colour, the genitalia (trifolii:
nately the male has been mislaid but I have exam- Figs. 497, 498 and sativae: Figs. 499,500) show that
ined the female syntype and its identity is not in they are closely related. In Europe L. strigata and
doubt. This synonymy is established in the Appen- L. xanthocera occur on a number of genera, as also
dixbelow. Chromatomyia hortieola (see Table 21).
In his description of A. bicaudata, Hendel (1920) Phytomyza brischkei was described from Trifo-
suggested that this species might represent a syno- lium and it has in the past been considered that this
nym of A. frontella. I have recently examined ten of was its only host. However, it has now been dis-
the original 12 type specimens and it is clear that covered that it also feeds on Anthyllis (Tribe 19)
this is a distinct species. The form and colour of the where it is discussed further above.
head does resemble A. Jrontella but the cerci (Fig. It should be mentioned here that Phytomyza
1334) are exceptionally long and the genitalia (Fig. medieaginis Hering (1925b) was described as a
1333) show that this must be a grass feeder. result of confusion in breeding records (Spencer,
The primary hosts of A. nana Meigen (cf. Spen- 1977b: 251) and is in no way associated with Medi-
cer, 1976a) appear to be in the Trifolieae where it is eago but represents the senior synonym of P. sym-
134 Class MAGNOLIOPSIDA (Dicotyledons)
.~.-~
492 493
494 495
496
497 499
498 500
Figs. 492, 493. Agromyza frontella: 492, aedeagus, side view; 493, same, ventral view.
Fig. 494. Agromyza nana: aedeagus.
Figs. 495, 496. Liriomyza trifoliearum: 495, aedeagus; 496, distiphallus, ventral view.
Figs. 497, 498. Liriomyza trifolii: 497, aedeagus; 498, same, ventral view.
Figs. 499, 500. Liriomyza sativae: 499, aedeagus; 500, same, ventral view.
Subclass ROSIDAE 135
phyti Hendel (1935). Acceptance of this would 508. T. laburnifoliae (Spencer, 1964a) forms a short
represent an undesirable change of name for the linear mine on Crotalaria laburnifolia in Ethiopia.
Symphytum feeder (Lamiaceae) and a Submission No males are known and its exact affiliations remain
to the International Commission on Zoological to be established.
Nomenclature is therefore in preparation for the The genus Calycomyza, which has proliferated in
suppression of the name Phytomyza medicaginis. North and South America and has colonized 13
The 17 species, including the six polyphagous plant families mainly in the Asteridae, has just one
species occurring on this tribe are shown in Table species on Leguminosae, C. crotalarivora Spencer
21. (1963c). This forms a linear-blotch mine on Crota-
laria anagyroides in Venezuela. The genitalia (Figs.
Tribe 29. Crotalarieae (16 genera) 509,510) suggest relationship with species on Mal-
Only the single genus Crotalaria is known to be vaceae or Verbenaceae.
colonized by Agromyzidae, with six host-specific, Of the oligophagous species, Melanagromyza
four oligophagous and one polyphagous species. dolichostigma and Ophiomyia phaseoli have been
Melanagromyza crotalariana Spencer (1959) is recorded on Crotalaria juncea in Java, and 0. cen-
known from Uganda from stems of Crotalaria sp. trosematis on C. mucronata in East Africa. Great-
and is generally similar to M. sojae but the male head (1969) suggests that this host is preferred to
genitalia (Fig. 501) are very different and the species Centrosema, Phaseolus and Vigna.
appears to be isolated on Crotalaria. At Nairobi,
Kenya a new Ophiomyia sp., described in the Tribe 31. Thermopsideae (6 genera)
Appendix below as 0. crotalariella was recently Only the genus Baptisia has been colonized by
discovered forming external mines in young, ter- Liriomyza baptisiae (Frost, cf. Spencer and Steyskal,
minal twigs of the small tree, Crotalaria agatiflora. 1986) in Pennsylvania, U.S.A. forming linear-blotch
Its presence is revealed by leaf drop and withering mines on Baptisia tinctoria. It was later found in
of affected twigs. The black puparium remains Montana on Lupinus laxiflorus. The genitalia (Figs.
beneath the epidermis at the end of the mine, as is 511, 512) show that L. baptisiae is not closely
normal in this genus. The male genitalia (Figs. 502, related to the three leaf miners which are host-spe-
503) are not typical of the genus, suggesting long cific on Lupinus (see Tribe 32).
isolation of this species in its unique niche.
lapanagromyza parvula Spencer (1961a) was Tribe 32. Genisteae (20 -genera)
described from a single female from Kakoa, 1200 m Seven of the genera assigned to this tribe have been
near Mt. Kilimanjaro, Tanzania. Surprisingly, a colonized by 17 species of Agromyzidae in the five
population forming blotch mines on Crotalaria genera Hexomyza (1), Agromyza (4), Liriomyza (8),
agatiflora, growing in the grounds of the Herbarium, Phytomyza (3) and Chromatomyia (1). All but six
National Museums of Kenya, Nairobi proved to are restricted to the tribe, with only Liriomyza bap-
represent this species. 1. parvula is small, black, with tisiae occurring also on Baptisia (Tribe 31).
white halteres and distinctive genitalia, with the
aedeagus ending in two unusually long, paired Subtribe Lupininae
tubules (Fig. 504). The hypandrium (Fig. 505) has Lupinus is now placed in this mono typic subtribe
an extended apodeme and the sperm pump (Fig. and agromyzid host selection tends to support the
506) is exceptionally long. The larval posterior segregation of this genus. Lupinus has three host-
spiracles have the normal arrangement of three well- specific Liriomyza and two Phytomyza species and,
separated, oval pores. A male from South Island, in addition L. baptisiae indicating the close rela-
Aldabra (leg. B. Cogan) is externally similar but the tionship between Baptisia and Lupinus. Only Agro-
genitalia show that this represents a further unde- myza johannae de Meijere in Europe which is highly
scribed species, possibly related to 1. parvula and oligophagous, recorded on three further genera (see
feeding on the only Crotalaria present on Aldabra, below), crosses the subtribal boundary.
C. laburnoides. However this is an assumption and The three Liriomyza species associated with
the host is unknown. Lupinus were all described from high altitudes in
Two leaf miners on Leguminosae in Africa, de- California. Liriomyza lupinella Spencer (1981) and
scribed in Melanagromyza are now temporarily L. lupiniphaga Spencer (1981) both form mines in
retained in Tropicomyia (see Appendix below). part associated with the midrib. L. lupini Spencer
T. crotalariae (Hering, 1957a) forms normal linear (1981) feeds more generally in the leaf blade and is
mines (not epidermal) on Crotalaria capensis in thus probably a more advanced species. It has
Natal (illustrated in Hering (1957a: fig. 12)). The successfully followed its host into Central America
distinctive genitalia are shown in Fig. 507 and the where I recently found it in Costa Rica at 2820 m on
equally aberrant posterior larval spiracles in Fig. Lupinus ashenbonnii, north of San Jose. The male
136 ClassMAGNOLIOPSIDA (Dicotyledons)
501 502
504 503
505 506
\ ...
f .~
'.~'
507 508
509 510
Subclass ROSIDAE 137
511 513
512 514
~~~~~Q: )!E:~>
515 ,) ~(»" 516
517
fll;f'!;"l
.!~-.'
518
519
520 521
522
524 525
, /
Figs. 519-521. Hexomyza sarothamni: 519, distiphallus, ventral view (holotype); 520, aedeagus (ex Genista aetnensis, Sicily); 521, gall
on Genista aetnensis,
Fig. 522. Agromyza demeijerei: aedeagus,
Fig. 523. Agromyzajohannae: aedeagus,
Fig. 524. Phytomyza cytisi: aedeagus.
Fig. 525. Agromyza pulla (= genistae): aedeagus.
Subclass ROSIDAE 139
genitalia associate L. lupiniphaga (Figs. 513, 514) also once been found on Petteria. Agromyza pulla
with species on Asteraceae (Figs. 1049, 1127); Meigen (cf. Hendel, 1931, as genistae Hendel) (Fig.
L. lupinella has genitalia (Fig. 515) suggesting rela- 525) feeds normally on Genista tinctoria but has
tionship with L. lupini (Fig. 516). also been recorded on Spartium (Hering). A. lutei-
Phytomyza lupini Sehgal (cf. Spencer, 1969b) frons Strobl (cf. Hendel, 1931), as far as is known, is
appears to be a primitive species and is one of the host-specific on Spartium junceum and on this host
few in this genus feeding and pupating internally in at Tibidabo, above Barcelona, Spain, I have caught
the stem. The larval posterior spiracles resemble adults and also reared A. johannae from mines on
those of stem-boring Melanagromyza species, hav- the bushes found in April. Although the adults of
ing the posterior spiracles on two strongly sclero- these three species are readily distinguishable, the
tized plates each with 12 pores around a central male genitalia show their close relationships.
horn. This spiracular arrangement has presumably
arisen independently in the two unrelated genera Briefly summarising and reviewing the host selec-
and is a modification which has evolved with the tion of the 108 species of Agromyzidae on the
spiracular horn assisting locomotion within the Leguminosae, the following facts emerge:
stem. The genitalia (Figs. 517, 518) suggest that this 1. The small number of species (8) and the host-
is an isolated species. Phytomyza lupinivora Sehgal specificity of 4 on the Caesalpinioideae and
(1971b) is known only from a single female reared Mimosoideae point to the wide separation of
from a leaf mine on Lupinus sericeus in the moun- these two groups from the Papilionoideae and
tains in Alberta. It is a largely black species and it would support their treatment as separate fami-
will be interesting to discover its affiliations when a lies if botanists are so inclined.
male becomes available. 2. Among the species restricted to the Papilionoi-
deae, every degree of host specialization is
Subtribe Genistinae apparent, from strict monophagy on a single spe-
Suss (loc. cit.) has detected larval feeding of Phy- cies or genus, limited oligophagy on two or more
tobia in Genista lucida Cambess in the Balkans and genera in the same tribe or, less frequently, on
Syria. two or more tribes, to extensive oligophagy on
A gall-causer, Hexomyza sarothamni (Hendel, cf. up to 27 genera in nine tribes (Liriomyza con-
Spencer, 1966b) forms twig-galls on Cytisus (Saro- gesta).
thamnus) scoparius in western Europe. This species 3. Examples of strict monophagy are Melanagro-
possibly also occurs on Genista aetnensis in Sicily myza acaciae on Acacia (Mimosoideae), lapana-
from which I reared a long series in April, 1964. gromyza tephrosiae on Tephrosia (Tribe 6), Phy-
The aedeagus of the holotype from Austria (Fig. toliriomyza robiniae on Robinia (Tribe 7), Mel.
519) is not significantly different from that of the floridensis and lap. desmodivora on Desmodium
population from Sicily (Fig. 520). Specimens from (9), Amauromyza obscura and Paraphytomyza
Cytisus scoparius from the Loire Valley, France caraganae on Caragana (16), two Melanagromy-
show greater differentiation in the genitalia but with za and five Agromyza species on Vicia (21),
variation detectable between individual specimens Melanagromyza crotalariana, Tropicomyia crota-
from Genista, it is not proposed to split this complex lariae, T. laburnifoliae, lap. parvula on Crota-
further at this time. The gall on Genista from Sicily laria (29).
is shown in Fig. 521. H. websteri which forms galls 4. The Leguminosae are highly acceptable to
on Wisteria (Tribe 6, Tephrosieae) in Japan has very polyphagous species, particularly Liriomyza
different genitalia (Fig. 385) but it seems neverthe- (nine species) and Chromatomyia horticola.
less probable that the two species are directly
related. In the genus lapanagromyza, both genitalia and
Agromyza demeijerei Hendel (cf. Spencer, 1976a) larval characters indicate strong differentiation be-
is monophagous on Laburnum spp., forming char- tween species in the Neotropical and Oriental
acteristic blotch mines. Interestingly this species is Regions. The proliferation of spiracular pores from
only known on Laburnum cultivated in gardens in the plesiomorphous three as in l. tristella (Fig. 413)
Western Europe but is presumably present where is lacking in the Neotropical species but surprisingly
the host occurs naturally in S.E. Europe. The gen- is present in incipient form in l. viridula in the
italia (Fig. 522) indicate close relationship with United States (Quercus) and is fully advanced in
A. johannae (Fig. 523) whose normal host is Cytisus l. quercus in Japan (Fig. 223). The distinctive
but which has also been found on Genista, Spartium forked dorsal process in the aedeagus of l. tristella
and very young leaves of Ulex. Phytomyza cytisi (Fig. 412) is also present in l. duchesneae (Rosa-
Brischke (cf. Spencer, 1976a) (Fig. 524) was re- ceae) but in this species the larval posterior spiracles
corded on Laburnum (as Cytisus laburnum) and has have proliferated differently, to form a more or less
140 Class MAGNOLIOPSIDA (Dicotyledons)
regular circle of ten pores (Fig. 365). This is an diverse assemblage of agromyzids on the Legumi-
interesting case of differential evolution of different nosae has immensely complex relationships with
parts of the insect - the genitalia diverging more in ancestral species on other families and detailed
some closely-related species but the divergence is reconstruction of the switches which have obviously
greater in larval characters in others. Japanagromyza occurred, even on unrelated families, is no longer
tristella thus shows relationship both with species on possible.
Quercus (Fagaceae) and Duchesnea (Rosaceae).
Although traditionally the Leguminosae are con- The agromyzid genera on 36 families in 12 orders of
sidered to have evolved from species of Rosalean the ROSIDAE discussed below are shown in Table
ancestry, this is the only case where any relationship 22 and the individual species are listed in Table 24.
is detectable between species on Leguminosae and
Rosaceae. Table 22
An isolated and puzzling relationship is apparent
Species
between Phytoliriomyza viciae (Vicia) and Phlir.
clara in the U.S.A. on Pteridium (Figs. 470, 17). I Family-specific Polyphagous
can only suggest that there has been a chance switch
Phytobia
from an ancestral fern feeder to Vicia. Hexomyza 1
The distinctive genitalia (Fig. 447) of Phytomyza Melanagromyza 2 (1 undescribed)
brischkei (Anthyllis and Trifolium) appear to in- Ophiomyia 3
dicate definite relationship with both P. sedicola Kleinschmidtimyia 1
Tropicomyia 2 4
(Sedum, Crassulaceae) (Fig. 350) and to the P. ra-
Japanagromyza 3
nunculivora group on Ranunculus and Anemone Agromyza 2
(Figs. 69, 160). I postulate that the Ranunculaceae Phytoliriomyza 1
feeders are the more primitive group, from which Amauromyza 2
both Sedum, and Anthyllis and Trifolium (P. brisch- Liriomyza 3 7
Calycomyza 1 (undescribed)
kei) have been colonized. Phytomyza subtilis in
Napomyza
western U.S.A., Alaska and Canada (Lathyrus and Chromatomyia 3
Vicia), which is in a different group, also has gen- Phytomyza 4
italia suggesting possible relationship with species
on Ranunculaceae (Aquilegia, Thalictrum). 29 13
Calycomyza cassiae from Panama (Cassia, Cae-
salpinioideae) is clearly related to the two feeders on
Malvaceae in the Nearctic/Neotropical Regions, Order PROTEALES (2 families, 1050 species)
Cal. malvae and Cal. sidae. However, I feel that too Cronquist (1981) includes in this order the Protea-
much significance should not be attached to this, ceae and Elaeagnaceae. The Proteaceae are an old
pending confirmation of this early record on Cassia. family, with fossil pollen known from mid-Upper
Two species suggest relationship with species on Cretaceous deposits in Australia (some 82 million
the Apiaceae. Phytomyza oxytropidis (Fig. 443) has years ago) and from somewhat younger deposits
genitalia associating it with either the P. spondylii- from the latest Cretaceous in South America. The
group (Fig. 653) on many genera of Apiaceae or Elaeagnaceae appear to be considerably younger,
with the P. albiceps-group which is even more wide- with the earliest known pollen dating from the
spread on the Asteraceae. Ophiomyia orbiculata on Palaeocene in early Tertiary.
Pisum has larval characters (Fig. 455) strongly sug- Frohne and Jensen (1985) place the Proteaceae
gesting relationship with 0. heracleivora (Fig. 594) and Elaeagnaceae in widely separated monotypic
on Heracleum. orders, and Goldberg (1986) also places the Protea-
Relationships with Asteraceae feeders are ceae in a monotypic order, while the Elaeagnaceae
detectable in three genera. Oph. alysicarpi in are left isolated under 'Incertae Sedis.'
northern Australia (Fig. 399) appears to be related No agromyzids are known on the Proteaceae
to 0. conspicua feeding on Eidens and almost cer- (1000 species) but seven species have been re-
tainly other genera. Amauromyza lathyroides in corded on the Elaeagnaceae (50 species). The
California known only on Lathyrus is possibly re- Elaeagnaceae are thus a relatively acceptable family,
lated to the common oligophagous Asteraceae-feed- while the Proteaceae are totally unacceptable.
er, Am. maculosa. Finally, three Liriomyza species, Cronquist concedes that 'it is not at all certain that
L. lathyri (Lathyrus), L. cicerina (Cicer) and L. lupi- the two groups are closely related.' The striking
niphaga (Lupinus) all have genitalia suggesting rela- contrast in their acceptability to Agromyzidae
tionship with species on Asteraceae. strongly suggests that inclusion of the two families in
It is clear from the foregoing that the large and a single order is not well founded.
Subclass ROSIDAE 141
Elaeagnaceae (3 genera, 50 species) larval feeding of Phytobia has been detected on ten
genera (cf. Table 24). .. .
The Elaeagnaceae are a taxonomically isola.ted
It is considered that tannins, includmg ellagIc aCId
family, present mainly in temperate and subt~opIcal
and often proanthocyanins, play a major role in
areas of the northern hemisphere. The famIly has
discouraging predators. This possibly explains the
been colonized by seven species of Agromyzidae in
absence of leaf miners on the Onagraceae, with such
the four genera Ophiomyia, Japanagromyza, Amau-
apparently suitable hosts as Circaea, Epilobium,
romyza and Chromatomyia (cf. Table 24). In J~pan
Fuchsia and Oenothera. The family appears to be
Ophiomyia sasakawai was reared from lon~ ~mear
almost wholly unacceptable to Agromyzidae, with
mines on Elaeagnus pungens and was ongmally
only a single record of the most highly polyphagous
described in a monotypic genus as Carinagromyza
species, Chromatomyia horticola, on Oenothera.
heringi Sasakawa (1954d). The ~pecies is slig~tly
aberrant in having a narrow, raIsed keel runmng
down the centre of the frons but is otherwise not Combretaceae (20 genera, 400 species)
atypical of leaf miners accepted in Ophiomyia.
Only polyphagous Tropicomyia species are known
Larval characters fully support the inclusion of the
here both on Combretum and Quisqualis, with
species in Ophiomyia and the generic synonymy was
T. p~lyphaga recorded in India and T. polyphyta in
established by Spencer and Martinez (19~7: .256). Australia (Table 24).
The male genitalia conform to other specIes m the
genus (Fig. 526). . .
Also in Japan and in the Bomn Islands (MICro- Myrtaceae (140 genera, 3000 species)
nesia) Japanagromyza elaeagni Sasakawa (1954b)
This family occurs in tropical and subtropical are~s
occurs as a leaf miner on Elaeagnus and the gen-
throughout the world and is also well developed m
italia (Fig. 527) indicate its isolated position with the
temperate Australia.
distiphallus being entirely membranous.
The only leaf mining agromyzid known on Euca-
Two related Amauromyza species are known on
lyptus, Japanagromyza eucalypti . (cf. Spencer,
the family, Am. elaeagni (Rohdendorf-Holmanova,
1977c), was described from AustralIa and Java on
1959a) in Europe on Elaeagnus and Am. shepher-
E. camaldulensis, and is probably present on other
diae Sehgal (1971b) in western Canada on Shepher-
Eucalyptus species. The male genitalia (Fig. 533)
dia. The male genitalia show considerable differen-
clearly indicate relationship with J. salicifolii in
tiation but are of the same general form (elaeagni:
Europe (Fig. 298) and J. viridula in the eastern
Figs. 528, 529; shepherdiae: Figs. 530, 531). .
United States (Fig. 221), although there is differen-
Griffiths (197 6b) discusses three Chromatomyza
tiation in the epandrium and surstyli which lack the
species on Shepherdia, which he writ~s ~re '~eakly
prominent spines present in the other two species
differentiated.' Chr. shepherdiana Gnffiths IS pre-
(Fig. 534). It must be assumed that there has been
sent in the Yukon Territory, Chr. leptargyreae Grif-
dispersal from the northern hemisphere via Java to
fiths in Alberta and Chr. merula (Spencer, 1969b)
Australia, presumably with additional unknown
in Alberta and the Yukon. The genitalia of Chr.
intermediate hosts.
merula (Fig. 532) are generally similar to tho.se of
Tropicomyia polyphaga is known on Careya in
the other two species which can be more satIsfac-
India and T. polyphyta on Angophora in Australia.
torily differentiated on minor morphological diff~r
ences. There has thus been relatively recent specIa-
tion on the same host from a single original coloni- Thymeleaceae (50 genera, 500 species)
zation.
Three leaf miners in the unrelated genera Klein-
The relationships of the Agromyzidae on the
schmidtimyia and Liriomyza are known on this
family are not obviously apparent and. provid~ .no
family.
indication regarding the correct botamcal pOSItIon
K. wikstroemiae (Kleinschmidt, cf. Spencer,
of the Elaeagnaceae.
1977c and 1986c) is known in Australia on Pimelea
and Wikstroemia (Fig. 536), belonging to the group
Order MYRTALES (13 families, 9000 species) of small black leaf miners which have been referred
to as the pis i-group (see under Fabaceae). The male
Seven families in this large order support Agromyzi-
genitalia are characterised by the greatly elongated
dae, with only three family-specific species in t?~
hypandrial apodeme (Fig. 535).
genera Kleinschmidtimyia, Japanagromyza and Lm-
In Europe Liriomyza approximata (Hend~l,
omyza. The order is generally unacceptable to p~l~
1920) is monophagous on Daphne mezereum. ,!hIS
phagous Liriomyza speci~s, althou~h L. cocculz IS
is an isolated species, with the scutellum essentIally
known on Wikstroemia m Hawall. By contrast,
black and the larvae having 18 minute bulbs on each
142 Class MAGNOLIOPSIDA (Dicotyledons)
,--
(fi
.,::../~/
... ~., .' .. .' . :,:
'
526 527
529
528
~
. 'J2;;;;;;~
~
:!
- - ..
. --:-,.. ..
-,:~: \ ....
531
530
532 534
533
535
Subclass ROSIDAE 143
posterior spiracular process. The genitalia, however uncommon on the shrub Comus sanguinea in Eu-
(Fig. 537) are not atypical of Liriomyza. rope and on the low herbaceous C. canadensis in
Suss and Muller-Stoll (1980) and Suss (pers. Canada and the U.S.A. This is thus a true holarctic
comm.) record larval feeding of Phytobia on six species which has dispersed across Beringia. P. cor-
families in the Myrtales on ten different genera as ni Kaltenbach is a doubtful species described in
follows: Germany and it has not been rediscovered since its
description in 1859. In Japan P. nishijimai (cf. Sasa-
Combretaceae: Combretum sokodense Engl. et
kawa, 1961) is found on Comus kousa and the gen-
Diels - Togo, Nigeria; C. splendens Engl. et Diels
italia (Fig. 539) show that this is not related to
- Uganda
P. agromyzina (Fig. 538).
Lythraceae: Lagerstroemia lanceolata Wall. - India;
Suss and Muller-Stoll (1980) and Suss (pers.
L. ovalifolia Teysm. et Binn. - Java; L. parvifolia
comm.) have reported larval feeding by Phytobia in
Roxb. - eastern India; L. speciosa Pers. - Java
Comus mas L. and C. sanguinea L. in Europe.
Melastomataceae: Melastoma asperum Blume -
Java; M.lanuginosum Blume - Java
Myrtaceae: Acmena floribunda DC: - Central Order SANTALALES (10 families,
America; Aphanomyrtus tetraquetra Valet. - 2000 species)
Java; Eucalyptus globulus Lab. - Australia;
This is an old order but 'interpretation of the fossil
E.lanceolatus - Australia; E. macrorrynca F.
record is in a state of flux' (Cronquist, 1981). Ofthe
Muell. - Australia; Eugenia discophora Koord. et
two families on which agromyzids are known, San-
Valet. - Java; E. jambolana Lam. - Australia;
talaceae and Loranthaceae, the former has been
Psidium pyriferum L. - America
considered to enter the fossil record in the late
Punicaceae: Punica granatum L. - Balkans to Hima-
Palaeocene and the Loranthaceae in lower Eocene.
layas
Thymeleaceae: Schoenobiblus peruvianus StandI. -
Peru Santalaceae (35 genera, 400 species)
These records could refer to at least six unde- Liriomyza thesii Hering (cf. Hendel, 1931) is a leaf
scribed species. miner on Thesium spp. in mountains in southern
Germany. The male genitalia (Fig. 540) clearly show
this to be an isolated species differentiated on its
Order RHIZOPHORALES (1 family,
semi-parasitic host.
14 genera, 100 species)
Cronquist (1981) with some hesitation raises the Loranthaceae (60-70 genera, 700 species)
Rhizophoraceae to ordinal rank. No agromyzids
have been described from the family but Suss and Japanagromyza loranthi Spencer (1966a) has been
Muller-Stoll (1980) and Suss (pers. comm.) record recorded once as a leaf miner on Loranthus sp. in
larval feeding of Phytobia on Gonotroches axillaris India. Only females are known and the affiliations of
Blume in Java. the species cannot be determined.
#:
C)il~~~)~~ ..
537
~/~<~ " ./
539
~-r~~-_/
/'
/~~:-:
541 542
544
only on the genus flex, with host-specific species istic of the genus and the genitalia are shown in Fig.
known in Europe, Japan and the United States. 548. With only a female known of T cassin is, it
'Fossil pollen attributed to flex first appears in cannot be further clarified but I suspect that it is in
Turonian (early Upper Cretaceous) deposits in fact identical to T gymnosporiae, with the two 'spe-
Australia' (Cronquist, 1981). cies' feeding on related genera in the Celastraceae.
The genus Phytomyza is represented on flex with The formal synonymy is best delayed until more
eight species, of which four are strictly monophag- material is available, particularly a male from Cas-
ous on individual species. The most primitive spe- sine.
cies is possibly P. ilicis Curtis (cf. Hendel, 1935), I found old mines, also near Cape Town, on
which is common in Europe on I. aquifolium. Ovi- Elaeodendron capense which probably represent
position takes place at the base of the petiole in a T gymnosporiae.
young leaf and after feeding initially in the midrib, a Suss and Muller-Stoll (1980) and Suss (pers.
conspicuous blotch mine is formed in the leaf. The comm.) record larval feeding of Phytobia on Celas-
genitalia (Figs. 541, 542) are of the same form as in trus luteolus Del. in Ethiopia and on the Zanzibar
the other seven species confirming their monophyly. coast.
P. jucunda Frost and Sasakawa (cf. Sasakawa,
1961) is known only in Japan where it has been
Order EUPHORBIALES (4 families,
found on four local flex species. The genitalia (Fig.
7560 species)
543) differ from all other species in having the
distiphallus strongly recurved. There has been re-
Pandaceae (3 genera, 15 species)
markable radiation in the U.S.A. where the six
further species are found, mainly in the S.E. States. This small family distributed in the Old World
P. glabricola Kulp (1968) is host specific on I. glab- tropics was earlier included within the Euphorbia-
ra, P. verticillatae Kulp (1968) on I. verticillata (Fig. ceae but is now generally accepted as distinct.
544), P. vomitoriae Kulp (1968) on I. vomitoria Larval feeding of Phytobia has been detected by
(Fig. 545). P. ditmanni Kulp (1968) appears to be Suss (loc. cit.) in Microdesmis puberula Hook. f. in
restricted to I. decidua and I. serrata, while P. opa- the Ivory Coast. It should be noted that similar
cae Kulp (1968) and P. ilicicola Loew (cf. Spencer larval feeding has been recorded on Macaranga and
and Steyskal, 1986) have each been recorded on Sapium in the Euphorbiaceae, also in the Ivory
three !lex species. Although differences in the male Coast. It seems probable that the same Phytobia sp.
genitalia are small, there are also slight differences in is involved, which would suggest that the Pandaceae
external morphology which can be associated with are weakly differentiated from the Euphorbiaceae.
significant differences in biology.
In Japan the polyphagous Tropicomyia styricicola
Euphorbiaceae (300 genera, 7500 species)
has been found on flex and larval feeding of Phyto-
bia has been observed in the stems of flex in the This family is cosmopolitan but best developed in
U.S.A. (Greene, 1914). tropical and subtropical regions. It has been colon-
ized by 20 species of Agromyzidae in seven genera
(Table 23) and they and their 16 host genera are
Celastraceae (50 genera, 800 species)
listed in Table 24. Larval feeding of Phytobia has
A gall-causer, Hexomyza gymnosporivora (Spencer, been detected in seven genera (see below). Cron-
1963b) is present in South Africa and Kenya on quist (1981) points out that with the wide range of
Maytenus buxifolius (formerly Gymnosporia). The vegetative and chemical features and in pollen
distinctive galls are shown in Fig. 546 and the morphology the Euphorbiaceae could be com-
genitalia in Fig. 547. The genitalia differ significantly pared to several different families and orders, par-
from those of H. sarothamni, type of the genus (Fig. ticularly to Malvaceae (Dilleniidae). Frohne and
519), but in view of the similar biology, I feel it is Jensen (1985) do consider the family closer to the
justified to include gymnosporivora in Hexomyza. Malvales and Urticales but point out that homolog-
Also in South Africa, an unidentified Melanagromy- ous lectines in the Fabaceae and Euphorbiaceae and
za sp. has been recorded forming leaf galls on Celas- the occurrence of certain aphids (Aphis and Acyr-
trus sp. (Spencer, 1960a) and in Australia T poly- thosiphon) indicate the possible inclusion of the
phyta has been recorded on Celastrus subspicatus. family within the Rosidae. The diverse assemblage
Two Tropicomyia species have been described of Agromyzidae on the Euphorbiaceae provide no
from Cape Town, South Africa by Hering (1957a), evidence for any association with the Dilleniidae but
T cassinis on Cassine sphaerophylla and T gymnos- nor of direct relationship with other species on the
poriae on Gymnosporia buxifolia (now Maytenus). Rosidae. Here I follow the classification proposed
T gymnosporiae forms epidermal mines character- by Cronquist.
146 Class MAGNOLIOPSIDA (Dicotyledons)
546 548
--.:.-;.
549
551 552
553
Figs. 546, 547. Hexomyza gymnosporivora: 546, galls on Maytenus buxifolius; 547, aedeagus.
Fig. 548. Tropicomyia gymnosporiae: aedeagus.
Figs. 549, 550. Melanagromyza kenyensis: 549, aedeagus, side view; 550, same, ventral view.
Fig. 551. Japanagromyza kalshoveni: aedeagus.
Figs. 552-555. Tropicomyia clutiae: 552, larval mouth parts; 553, leaf mine on Clutia pulchella; 554, aedeagus, side view; 555, same,
ventral view.
Subclass ROSIDAE 147
Euphorbia sp., has apparently been colonized In Australia the polyphagous Tropicomyia polyphy-
relatively recently, with little differentiation in the ta has been recorded on Alphitonia.
genitalia. Suss and Muller-Stoll (1980) and Suss (pers.
Three of the European polyphagous species comm.) have detected larval feeding of Phytobia on
attack the Euphorbiaceae (Lir. bryoniae, L. strigata nine genera of Rhamnaceae:
and Chr. horticola) and in South America L. sativae
Alphitonia philippinensis Braid - Philippines
occurs commonly on Ricinus communis. In Kenya
Ampeloziziphus amazonicus Ducke - Brazil
the introduced L. trifolii has been found on the
Ceanothus coeruleus Lag. - Mexico, U.S.A.
same host.
C. cuneatus (Hook) Nutt. - U.S.A.
In Florida leaf mines referable to L. schmidti
C. incanus Torr. et Gray - U.S.A.
have been found on Acalypha and Breynia (Steg-
C. palmeri Trel. - U.S.A.
maier, 1966), and in Argentina L. schmidti has been
C. thyrsiflorus Eschw. - U.S.A.
found on Manihot glazionii.
Colubrina arborescens (Mill.) Sarg. - Cuba
Suss and Muller-Stoll (1980) and Suss (pers.
C. ferruginosa (Lam.) Brongn. - Puerto Rico
comm.) record larval feeding of Phytobia on seven
C. rufa Reiss - Brazil
genera of Euphorbiaceae:
Discaria discolor (Hook.) Benth. et Hook. - Argen-
Aporosa frutescens Blume - Java tina, Chile
A. microcalyx Hassk. - Java Lasiodiscus articulatus Capuron - Madagascar
Baccaureajavanica Muell.-Arg. - Java L. chevalieri Hutch. - West Africa
B. racemosa Muell.-Arg. - Java Reynosia mucronata Gris. - Cuba (Lorna de la Pita)
Bridelia minutiflora Hook. f. - Java Rhamnus cathartica L. - Europe to Transcaucasia,
Macaranga capensis Sim. - southern Africa West Siberia, Algeria
M. hendelotii Baill. - Ivory Coast Rh.lanceolata Pursh - U.S.A.
Manihot palmata Muell.-Arg. - Arizona to Argen- Scutia commersonii Brongn. - Africa
tina
Sapium aubrevillei Leandri - Ivory Coast These records could refer to at least seven dif-
Spirostachys africana Sond. - South and East Africa ferent species.
558
556
JW
557 559 ~~\ . 560
~
561 562
... ~.
564
563
565 566
Figs. 556, 557. Tropicomyia crotonella: 556, aedeagus; 557, inner margin of epandrium.
Fig.558. Ophiomyia cabanae: aedeagus (Valladares).
Figs. 559, 560. Ophiomyia ricini: 559, anterior larval spiracles; 560, posterior spiracles (de Meijere).
Fig. 561. Liriomyza pascuum: aedeagus.
Fig. 562. Liriomyza balcanica: aedeagus.
Fig. 563. Liriomyza myrsinitae: aedeagus.
Fig. 564. Liriomyza heringi: aedeagus (Nowakowski).
Fig. 565. Phytomyza ceanothi: aedeagus.
Fig. 566. Liriomyza polygalae: aedeagus.
150 Class MAGNOLIOPSIDA (Dicotyledons)
Linaceae (6 genera, 220 species) Phytobia, with only one species described, others
are known from larval feeding only.
Napomyza lateralis (Fallen), a species which is
primarily oligophagous on Asteraceae (see below),
has been recorded in stems of Linum in Greece. Aceraceae (2 genera, 112 species)
The species concerned is very close to the true
The dominant genus is Acer which is widespread in
N. lateralis but more detailed studies, including the
both temperate and some tropical regions. 'Maples
examination of more material, might be profitable to
are well represented in the fossil record from the
confirm whether this represents an unusual case of
base of the Miocene to the present, and pollen
host transfer or whether the species in Linum is
considered to belong here dates from the Oligocene'
possibly host-specific and distinct.
(Cronquist, 1981).
In Europe Liriomyza strigata, L. xanthocera (as
Phytobia setosa (Loew), redescribed later as
crucifericola) and Chromatomyia horticola have
Agromyza aceris Greene (d. Spencer and Steyskal,
been recorded on Linum and in the U.S.A. L. hui-
1986) was reared from young trees of Acer rubrum
dobrensis is known on the same host.
in the eastern U.S. and is also not uncommon in
Ontario and Quebec, Canada. The genitalia (Fig.
Order POL YGALES (7 families, 567) clearly indicate relationship with Pb. amelan-
2300 species) chieris (Greene) (cf. Fig. 377) feeding on Amelan-
chier (Rosaceae). Further species of Acer in the
The Polygales and Linales are regarded by Cron-
United States in which larval feeding of Phytobia
quist (1981) as a pair of closely related offshoots
has been detected by Suss (loc. cit.) are A. macro-
from the Rosales. Only three families have been
phyllum Pursh (western U.S.A.), A. pennsylvanicum
colonized by agromyzids with a single host-specific
L., A. saccharinum L. and A. saccharum Marsh.
species on the Polygalaceae.
Other species have been attacked in other regions as
follows:
Malpighiaceae (60 genera, 1200 species)
Acer campbellii Hook. f. et Thoms. - Himalayas,
Only the polyphagous Tropicomyia polyphyta has Sikkim
been recorded on this family in Australia, on Heter- A. campestre L. - Europe, western Asia
opteris nitida. A. platanoides L. - Europe, Caucasus
A. pseudo-platanus L. - Europe, western Asia
Polygalaceae (12 genera, 750 species) The Indian and European records almost certain-
ly refer to two distinct species.
This nearly cosmopolitan family is obviously old,
with pollen known from the Palaeocene and fruits
known from Eocene deposits in Egypt and the Anacardiaceae (60-80 genera, 600 species)
Rocky Mountains. This widespread family has a good fossil record,
Liriomyza polygalae Hering (cf. Hendel, 1931) is with wood from Patagonia known in the Palaeocene,
widespread in Europe with its host Polygala vulgaris. and is known from Eocene deposits in England.
The male genitalia (Fig. 566) show this to be an Fossil pollen similar to that of Rhus comes from the
isolated species. Palaeocene in Europe (Cronquist, 1981).
In Argentina an unidentified Calycomyza sp. has The only leaf miner known on the family is the
been found on Noumina sp., (Valladares, unpub- polyphagous Chromatomyia horticola on Rhus and
lished report, 1986). Cotinus in Europe. However, Suss (loc. cit.) has
recorded larval feeding of Phytobia on four genera:
Vochysiaceae (7 genera, 200 species) Cotinus coggygria Scop. (= Rhus cotinus L.) -
Species in this family are mostly native to South southern Europe, Caucasus and S.W. Siberia
America. Suss and MUller-Stoll (1980) and Suss Haematostaphis barteri Hook. f. - West Africa
(pers. comm.) have detected larval feeding of Phyto- (Ivory Coast, Sierra Leone, Upper Guinea)
bia on Qualea albiflora Warm. in Guayana. Lannea welwitschii Engl. - West Africa
Semecarpus albescens Kurz - Java
Order SAPINDALES (15 families, 5400 species) These records appear to represent three distinct
species.
Eight families in this large order are known to
support Agromyzidae but, apart from six polyphag-
ous leaf miners, all are in the cambium-boring genus
Subclass ROSIDAE 151
'.'
, ,
~~~~~~;~~~li
567 ~~
569
568
570 571
Fig. 567. Phytobia setosa: aedeagus.
Figs. 568, 569. Agromyza nigrescens: 568, aedeagus, side view; 569, same (as oycoviensis Beiger).
Fig. 570. sp. (Oxalis): leaf mine on Oxalis acetosella.
Fig. 571. Phytoliriomyza melampyga: aedeagus.
Burseraceae (16-20 genera, 600 species) regions. The earliest fossil record is pollen referred
to the Meliaceae from the Oligocene.
This pantropical family 'probably originated no later
The only leaf miner known on the family is Tropi-
than the Eocene epoch. Palaeogene fruits from
comyia polyphyta on Melia azederach in Australia.
England are attributed to the Burseraceae, as are
Suss (loc. cit.) records larval feeding of Phytobia
some Eocene macrofossils from western U.S.A., and
on four genera:
it has been suggested that the amber in the Eocene
London Clay is of burseroid origin' (Cronquist, Cabralea cangerana Mart. - Brazil
1981). Chisocheton divergens Blume - Java
No leaf miners are known on the Burseraceae but Dysoxylum densiflorum Miq. - Java
Suss (loc. cit.) records larval feeding of Phytobia on D. densiflorum Miq. var. minor Koord. et Valet. -
2 genera: Java
D. nutans Miq. - Java
Canarium schweinfurthii Engl. - West Africa
Trichilia heudelotii Planch. - Senegal/Gambia, Gui-
Zanha galungensis Hiern - West Africa (Dahomey)
nea, Cameroun, Spanish Guinea, Gabon, Nigeria,
Ghana, Liberia
Meliaceae (51 genera, 550 species)
These records must refer to 3 distinct species.
This family is widespread in tropical and SUbtropical
152 Class MAGNOLIOPSIDA (Dicotyledons)
Rutaceae (150 genera, 1500 species) quist discusses possible other orders for the
No host-specific leaf miners are known on this
two families Balsaminaceae and Tropaeolaceae.
family but it has been colonized by three polyphag-
Three host-specific agromyzids are known on
ous species. Tropicomyia flacourtiae has been found
the order, Phytoliriomyza melampyga on the
Balsaminaceae and two Agromyza species on
on Citrus in Nigeria, and in Australia T. polyphyta
the Geraniaceae. Only the Tropaeolaceae are
has been reared from Murraya paniculata and mines
generally acceptable to polyphagous species.
have been found on Citrus aurantifolia. Chromato-
myia horticola is known on Ruta in Europe.
Suss (loc. cit.) has recorded larval feeding of Phy- Geraniaceae (11 genera, 700 species)
tobia on six genera:
The largest genus, Geranium (300 species), is pre-
Acronychia laurifolia Blume - Java sent in temperate areas and the montane tropics,
A. laurifolia var. densiflora Blume - Java while Pelargonium (250) is mainly South African
A. trifoliata Zollo et Mor. - Java but is widely cultivated in Europe, as Geranium.
Citrus aurantium L. - Italy, Sardinia, Spain Agromyza nigrescens (cf. Hendel, 1931) is a
Esenbeckia fasciculata R. Barban - Brazil common leaf miner on Geranium in Europe, feed-
Limonia warneckei Engl. - West African coast from ing on a number of species, particularly G. pratense.
Liberia to Nigeria Beiger (1960a) described A. oycoviensis in Poland,
Murraya exotica L. var. sumatrana Hook. - Java the only known host being Geranium phaeum. She
Phellodendron amurense Rupr. - USSR (Sakhalin), conceded that this species was very close to A. nig-
Amur and Ussuri regions, Korea. rescens and no differential diagnosis was given, per-
mitting the two species to be separated. Interest-
These records represent at least 4 species.
ingly, A. nigrescens has never been found on Pelar-
gonium.
Sapindaceae (140 genera, 1500 species) In Japan Sasakawa (as Tsujita before his marri-
age) described A. japonica (1951) as a subspecies of
Suss and Muller-Stoll (1980) and Suss (pers. comm.)
nigrescens on Geranium nepalense. The genitalia of
record larval feeding of Phytobia on three genera:
a specimen from Spain (Fig. 568) differ primarily in
Arytera littoralis Blume - Java
the degree of pigmentation from A. oycoviensis (Fig.
Koelreuteria paniculata Laxm. - northern China
569) and ssp. japonica and there is virtually no
Pometia tomentosa Teysm. and Binn. - Java
difference between these two. I consider that in fact,
In Venezuela I found empty linear-blotch mines
with no differences between adults or in larval
on Serjania, near S. cordiophyllum which cannot
characters in the three taxa, only a single species is
definitely be placed to genus but probably represent
involved and ssp. japonica and A. oycoviensis are
a Calycomyza.
synonymised with nigrescens in the Appendix
below. A. nigrescens is closely related to species
Staphyleaceae (5 genera, 50 species) feeding on Rosaceae, as rightly noted by Sasakawa
(d. A. potentillae, Fig. 366). Agromyza erodii
This family is 'irregularly distributed in the Ameri-
Hering (1927a) has only been recorded as a leaf
cas and in Eurasia and extending also into the Malay
miner on Erodium laciniatum in Tenerife, Canary
Archipelago' (Cronquist, 1981). No leaf mines are
Islands. Only the unique female holotype is known
known but Suss (loc. cit.) gives a single record of
and the affiliations of this interesting species (which
larval feeding of Phytobia on Staphylea pinnata L. in
is not closely related to A. nigrescens) cannot be
central Europe, Asia Minor and the Caucasus.
further clarified.
The Geraniaceae are obviously not generally
Zygophyllaceae (30 genera, 250 species) acceptable to Agromyzidae and colonization of the
genus Geranium must have occurred relatively
This largely tropical and subtropical family has been
early, to explain the distribution of A. nigrescens in
colonized only by two polyphagous species: Lirio-
both Europe and Japan.
myza schmidti on Bulnesia in Venezuela, and L. tri-
folii on Kallstroemia in Florida and on Guadeloupe,
and on Tribulus terrestris in Florida. Oxalidaceae (7-8 genera, 900 species)
Although this family is primarily tropical or sub-
Order GERANIALES (5 families, tropical, several species are common in Europe,
2600 species) particularly as garden weeds. The only agromyzid
record known in Europe is given by Hering (1957b)
This is a somewhat ill-defined order and Cron-
of a linear miner on 0. acetosella (Fig. 570) in
Subclass ROSIDAE 153
northern Germany but the adult is unknown. and Chromatomyia horticola, are not uncommon on
Liriomyza huidobrensis, a highly polyphagous Tropaeolum.
species in South America and California, has recent- A recently discovered polyphagous species, Li-
ly been recorded on Oxalis sp. in Argentina (Valla- riomyza cordobensis Valladares (nomen nudum) has
dares, unpublished report) but the family is un- been reared from Tropaeolum majus at Cordoba,
acceptable to most polyphagous species. Argentina, and also in Argentina from Thalictrum
(Ranunculaceae), Lens and Pisum (Leguminosae).
Tropaeolaceae (3 genera, 90 species) Frohne and Jensen (1985: 178) suggest that on
the basis of the chemical characteristics of Tropae-
Species in this family occur naturally only in the olum the family should be placed near the Brassica-
New World. Tropaeolum, the dominant genus, is ceae in the Dilleniidae: Capparales. The host selec-
found from Mexico to Chile, but is widely cultivated tion of L. brassicae would support this.
throughout the world (as 'nasturtium').
No host-specific species are known on the family,
Balsaminaceae (2 genera, 450 species)
but Tropaeolum is a favourite host of Liriomyza
brassicae. In Australia and Malaysia it was remark- Impatiens, the main genus, is native mainly in trop-
able to find even isolated plants of Tropaeolum in ical Asia and Africa but with a few species in tem-
gardens and parks, in some cases 100 km or more perate regions of both the Old and the New World.
from their nearest neighbours, invariably with a Two host-specific agromyzids are known on the
population of L. brassicae. The presence of mus- family, one holarctic, the other restricted to Japan.
tard-oil glucosides in Tropaeolum would account Phytoliriomyza melampyga (Loew, d. Spencer,
for its high degree of acceptability to L. brassicae, 1976a) is a leaf miner on Impatiens, described in
whose normal hosts are in the Brassicaceae. U.S.A. and common in Canada but also common in
Liriomyza emiliae Seguy (1951), known only in Europe with its host. The long, coiled tubules of the
Madagascar, feeds primarily on Emilia (Senecio- aedeagus (Fig. 571) suggest possible relationship
neae) and other unidentified composites. Among the with PI. arctica (Fig. 978), a stem miner on Astera-
type series there are specimens reared from Tro- ceae and there has clearly been a switch of the
paeolum which I have confirmed from the male ancestral species between these families, presum-
genitalia (Figs. 1129, 1130) also represent L. em i- ably from the Balsaminaceae to Asteraceae.
liae. This obviously represents a switch from the In Japan Phytomyza jimbriata (d. Sasakawa,
composite-hosts to Tropaeolum, which is present in 1961), a species with yellow frons, has been reared
Madagascar only as an introduced ornamental. from Impatiens noli-tangere but only the female
In Europe the three common polyphagous spe- holotype is known and the affiliations of this inter-
cies, L. xanthocera (= L. crucifericola), L. strigata esting species cannot be further clarified.
Table 24
Order PROTEALES
Elaeagnaceae
Elaeagnus Ophiomyia sasakawai; Japanagromyza elaeagni; Amauromyza elaeagni
Shepherdia Am. shepherdiae; Chromatomyia leptargyreae, Chr. merula,
Chr. shepherdiana
Order MYRTALES
Combretaceae
Combretum Phytobia sp.; Tropicomyia polyphaga, T. polyphyta
QuisquaJis T. polyphaga, T. polyphyta
Lythraceae
Lagerstroemia Phytobia sp.
154 Class MAGNOLIOPSIDA (Dicotyledons)
Onagraceae
Oenothera Chr. horticola
Melastomataceae
Melastoma Phytobia sp.
Myrtaceae
Acmena Phytobia sp.
Angophora T. polyphyta
Aphanomyrtus Phytobia sp.
Careya T. polyphaga
Eucalyptus Phytobia sp.; J. eucalypti
Eugenia Phytobia sp.
Psidium Phytobia sp.
Punicaceae
Punica Phytobia sp.
Thymeieaceae
Daphne Liriomyza approximata
Pimelea Kieinschmidtimiyia wikstroemiae
Schoenobi blus Phytobia sp.
Wikstroemia K. wikstroemiae; L. cocculi
Order RHIZOPHORALES
Rhizophoraceae
Gonotroches Phytobia sp.
Order CORNALES
Alangiaceae
Alangium T. atomella
Cornaceae
Comus Phytobia sp.; Phytomyza agromyzina, P. nishijimai
Loranthaceae
Loranthus J.loranthi
SantaJaceae
Thesium L. thesii
Order CELASTRALES
Aquifoliaceae
Hex Phytobia sp.; T. styricicola; P. ditmanni, P. glabricola, P. i1icicola, P. iii cis,
P. jucunda, P. opacae, P. verticillatae, P. vomitoriae
Celastraceae
Cassine T. cassinis
Celastrus Phytobia sp.; Melanagromyza sp.; T. polyphyta
Subclass ROSIDAE 155
Order EUPHORBIALES
Pandaceae
Microdesmis Phytobia sp.
Euphorbiaceae
Acalypha Agromyza sp. n.; L. schmidti; ? gen. indet. ( Argentina)
Antidesma J. kalshoveni
Aporosa Phytobia sp.
Baccaurea Phytobia sp.
Breynia L. schmidti
Bridelia Phytobia sp.
Clutia T. clutiae
Croton Mel. kenyensis; O. cabanae; T. crotonella, T. philocroton
Euphorbia L. ba\canica, L. heringi, L. myrsinitae, L. pascuum, L. strigata;
Chr. horticola
Macaranga Phytobia sp.
Mallotus J. yanoi
Manihot Phytobia sp.; L. schmidti
Mercurialis L. bryoniae, L. strigata
Ricinus O. ricini; L. bryoniae, L. sativae, L. strigata, L. trifolii
Sapium Phytobia sp.
Spirostachys Phytobia sp.
Order RHAMNALES
Rhamnaceae
Alphitonia Phytobia sp.; T. polyphyta
Ampeloziziphus Phytobia sp.
Ceanothus Phytobia sp.; P. ceanothi
Colubrina Phytobia sp.
Discaria Phytobia sp.
Lasiodiscus Phytobia sp.
Reynosia Phytobia sp.
Rhamnus Phytobia sp.
Scutia Phytobia sp.
Vitaceae
Cissus gen., sp. indet. ( Argentina)
Order LlNALES
Erythroxylaceae
Erythroxylon Phytobia sp. ( I, Tropical America ), ( 2, West Africa)
Linaceae
Linum L. huidobrensis, L. strigata, L. xanthocera; Napomyza lateralis (?);
Chr. horticola
156 Class MAGNOLIOPSIDA (Dicotyledons)
Order POLYGALES
Malpighiaceae
Heteropteris T. polyphyta
Polygalaceae
Noumina Calycomyza sp.
Polygala L. polygalae
Vochysiaceae
Qualea Phytobia sp.
Order SAPINDALES
Aceraceae
Acer Phytobia setosa, Phytobia sp.
Anacardiaceae
Cotinus Phytobia sp.; Chr. horticola
Haematostaphis Phytobia sp.
Lannea Phytobia sp.
Rhus Chr. horticola
Semecarpus Phytobia sp.
Burseraceae
Canarium Phytobia sp.
Zanha Phytobia sp.
Meliaceae
Cabralea Phytobia sp. (I)
Chisocheton Phytobia sp. (2)
Dysoxylum Phytobia sp. (2)
Melia T. polyphyta
Trichilia Phytobia sp. (3)
Rutaceae
Acronychia Phytobia sp. (I)
Citrus Phytobia sp. (2); T. flacourtiae, T. polyphyta
Esenbeckia Phytobia sp. (3)
Limonia Phytobia sp. (4)
Murraya Phytobia sp.; T. polyphyta
Phellodendron Phytobia sp.
Ruta Chr. horticola
Sapindaceae
Arytera Phytobia sp.
Koelreuteria Phytobia sp.
Pometia Phytobia sp.
Serjania gen., sp. ( ? Calycomyza )
Staphyleaceae
Staphylea Phytobia sp.
Subclass ROSIDAE 157
Zygophyllaceae
Bulnesia L. schmidti
Kallstroemia L. trifolii
Tribulus L. trifolii
Order GERANIALES
Geraniaceae
Erodium A. erodii
Geranium A. nigrescens
Balsaminaceae
Impatiens Phytoliriomyza melampyga; P. fimbriata
Oxalidaceae
Oxalis L. huidobrensis; gen. indet. ( mine only, Germany)
Tropaeolaceae
Tropaeolum L. brassicae, L. cordobensis nom. nud., L. emiliae, L. strigata,
L. xanthocera; Chr. horticola
Subclass F: Rosidae but in his order 60: Cornales, entire absence of Agromyza and of Liriomyza
closely associated with the Cornaceae and 'probably (apart from polyphagous species) is striking. The
derived from the Saxifragales.' Thorne (1983) in- assemblage is in direct contrast with that on the
cludes the Apiaceae in his Superorder Corniflorae, Leguminosae (in the same subclass), where Agro-
well removed from the Rosiflorae, and takes the myza (17) and Liriomyza (13) are well represented
radical step of submerging the Apiaceae as one of but Phytomyza (6) poorly. No logical explanation
four subfamilies (Aralioideae, Hydrocotyloideae, can be suggested for such disparity, apart from the
Saniculoideae and Apioideae) in the family Aralia- random nature of the colonization of differing fami-
ceae. On the basis of detailed discussion and analy- lies to which attention has already been drawn. The
sis of the chemical composition of the Apiaceae and host genera of the 87 species known on the family
Asteraceae, Frohne and Jensen (1985: 191) come to are listed in Table 26.
the conclusion that the relationship between these Here, I accept the subfamily and tribal classifica-
two families is not due to convergence but 'Bio- tion of the Umbelliferae adopted by Heywood
chemical discoveries have shown that close rela- (1971).
tionships can be accepted.' As will be seen from the
discussion at the end of this section there are no
Subfamily Hydrocotyloideae
direct relationships between the Agromyzidae on
the Apiaceae and other families in Cronquist's This subfamily supports no host-specific species but
Rosidae but there are striking relationships between only the three polyphagous species on Hydrocotyle,
the largest Phytomyza group on the Apiaceae (spon- Liriomyza bryoniae and L. strigata in Europe, and
dylii + obscurella groups) with the albiceps group L. trifolii in the New World.
on the Asteraceae.
Agromyzidae are known on 53 genera of the
Subfamily Saniculoideae
available 300. Doubtless this number will be in-
creased considerably with further collecting. Iden- Two species have been described from Sanicula,
tification of Apiaceae presents difficulties for Phytomyza brunnipes Brischke (d. Spencer, 1976a)
entomologists and the all important fruit is frequent- in Europe and P. saniculae Spencer (1981) in Cali-
ly not available on the young plants on which mines fornia. Both genitalia and leaf mines differ consider-
are found. Furthermore, delimitation of genera pre- ably (brunnipes: Figs. 577, 578; saniculae: Figs. 579,
sents difficulties for botanists, although species are 580) indicating a long period of isolation of the two
generally well characterized. However, the classifi- populations which it can be accepted are derived
cation of the family in the three subfamilies estab- from a common ancestor. Both belong to the spon-
lished by Engler is generally accepted and is fol- dylii group. The single species, P. astrantiae Hendel
lowed here. (d. Hendel, 1935) is known on Astrantia in Europe
and the genitalia (Fig. 581) show this to be related to
Table 25 the two species on Sanicula.
Sasakawa (1961) recorded Phytomyza polycladae
Agromyzid genera on Apiaceae (which was described from Angelica) as also feeding
Species restricted Polyphagous
on Sanicula. The host combination Angelica/Sanicu-
to family species la, with the two genera being in different subfami-
lies, seemed improbable and was so treated by
Melanagromyza 10 1 (M. splendida: Griffiths (1973a). I have now been able to examine
xenophagy) the genitalia (Fig. 648) of a male reared from Angeli-
Ophiomyia
5
ca at Yazu-gum, Tottori Province (M. Kuroda) and
Liriomyza
Calycomyza 1 it will be seen that this species is very close to
Napomyza 3 P. saniculae in California (Fig. 579). From this it can
Chromatomyia 2 (Chr. syngenesiae: be deduced either that in Japan the same species
xenophagy) does indeed feed on Angelica and Sanicula or that
Phytomyza 64 (3 undescribed)
two separate species are now involved but with only
79 8 slight differentiation. In either case there has been
Total 87 an unusual switch between the two host genera.
Sasakawa (1961: fig. 141d) illustrated the genitalia
of P. polycladae but omitted to record whether the
The colonization of this family by agromyzids is specimen was from Angelica or Sanicula, and in any
remarkable, with 15 family-specific species in four case it appears that the genitalia were damaged
genera, together with an extraordinary radiation of during preparation, with the end of the distiphallus
Phytomyza with 64 species (see Table 25). The missing. It would be useful to examine additional
Subclass ROSIDAE 159
Table 26
Araliaceae
Acanthopanax Phytomyza acanthopanacis
Aralia P. araliae, P. aralivora
Cussonia Melanagromyza cussoniae
Panax Mel. panacis
Apiaceae ( Umbelliferae )
Subfamily Hydrocotyloideae
Hydrocotyle Liriomyza bryoniae, L. strigata, L. trifolii
Subfamily Saniculoideae
Astrantia P. astran ti ae
Eryngium L. strigata; Calycomyza chilena; Chromatomyia horticola
Sanicula P. brunnipes, P. saniculae
Subfamily Apioideae
Tribe 2. Scandiceae
Anthriscus Napomyza glabra; P. chaerophylli )
Chaerophyllum Mel. chaerophylli; P. aurei, P. biseta, P. chaerophylli. P. chaerophylliana
Myrrhis Phytomyza sp. ( prob. P. chaerophylli )
Orlaya P. chaerophylli
Osmorhiza P. osmorhizae
Tribe 3. Coriandreae
Bifora P. chaerophylli
Coriandrum L. bryoniae, L. strigata; Chr. horticola
Tribe 4. Smyrnieae
Conium P. aurei, P. chaerophyUi, P. conii, P. coniopais; Chr. horticola
Pleurospermum Phytomyza sp.
Scaligeria P. ferulae ssp. scaligeriae
Smyrnium P. smyrnii
Tribe 6. Peucedaneae
Angelica Mel. angelicae, Mel. angeJiciphaga; P. angeJicae, P. angelicastri,
( = Archangelica )P. angelicivora, P. archangelicae, P. arnaudi, P. heracleana (1),
P. kibunensis, P. polycladae
Conioselinum P. angelicae ( 1 ), P. conioselini, P. sitchensis, Phytomyza sp.n. ( 1 )
Ferula Mel. ferulae; P. ferulae, P. ferulivora
Herac1eum * Mel. angeJiciphaga, Mel. !imata; Ophiomyia heracleivora; L. strigata;
P. angelicae, P. heracleana, P. lanati, P. pastinacae, P. spondylii,
P. spondylii ssp. heracleiphaga, P. sphondyliivora, P. tiingitica
Laser P. heracleana
Levisticum L. bryoniae, L. xanthocera; Chr. horticola; P. spondylii
Lomatium Mel. lomatii
Myrrhidendron Phytomyza sp.n.
Pastinaca * Mel. angeJiciphaga, Mel. sativae; P. heracleana, P. pastinacae, P. spondyJii
Peucedanum Chr. horticola; P. heracleana, P. pauliloewii, P. peucedani, P. thysseJini,
P. thysselinivora
Sphenosciadium Phytomyza sp.n.
TordyJium L. strigata; Chr. horticola; P. chaerophylli
Tribe 7. Laserpitieae
Laserpitium P. angelicae, P. heracleana (1), P. latifolii, P. zarzyckii
Tribe 8. Dauceae
Caucalis Phytomyza sp. ( prob. chaerophylli ), P. heracleana ( ? )
Daucus N. carotae; P. chaerophylli, P. ferulae, P. mylini; Chr. syngenesiae
Torilis Mel. sativae; P. chaerophylli
material to further clarify this interesting species. known to support Agromyzidae. The largest tribe,
Only a single species, Calycomyza chilena Spen- the Apieae has the largest number of Agromyzidae,
cer (1982) is known on Eryngium. This was de- 40, of which 31 are tribe-specific (see Table 26).
scribed from caught specimens in Chile but was Of 11 Melanagromyza species known on genera
subsequently reared from Eryngium agavifolium in in the Apioideae, seven are known in Europe, three
Argentina (Valladares, unpublished report, 1986). in North America and one in Australia. Coloniza-
The genitalia are typical of a number of species and tion of the tribes appears to have been random, with
it is not possible to determine from which family this one species each on the Scandiceae and Dauceae,
species has switched to Eryngium. five on the Apieae and six on the Peucedaneae.
M. angelicae (Frost, ct. Spencer and Steyskal,
1986). Hosts Angelica spp., known only in U.S.A.
Subfamily Apioideae
Male genitalia (Fig. 582) showing close relationship
Eight tribes have been recognized in this subfamily, with M. angeliciphaga in Europe. Posterior spiracu-
with only the small Echinophoreae which is re- lar plates of puparium (Fig. 583) widely separated.
stricted to the Mediterranean and the Near East not M. angeliciphaga Spencer (1969a). Previously
Subclass ROSIDAE 161
573
575
574
576
577 578
580
Fig. 572. Melanagromyza cussoniae: aedeagus.
Fig. 573. Melanagromyza panacis: aedeagus.
Fig. 574. Phytomyza acanthopanacis: aedeagus (Sasakawa).
Fig. 575. Phytomyza araliae: aedeagus (Sasakawa).
Fig. 576. Phytomyza aralivora: aedeagus.
Figs. 577, 578. Phytomyza brunnipes: 577, aedeagus; 578, leaf mine on Sanicula europaea (Hering).
Figs. 579, 580. Phytomyza saniculae: 579, aedeagus; 580, leaf mine on Sanicula sp.
162 Class MAGNOLIOPSIDA (Dicotyledons)
582
581
583
585
584 586
587 588
"qfJ~~,
589 590
Fig. 581. Phytomyza astrantiae: aedeagus.
Figs. 582, 583. Melanagromyza angelicae: 582, aedeagus; 583, posterior spiracular plates of puparium.
Figs. 584, 585. Melanagromyza angeliciphaga: 584, aedeagus; 585, posterior spiracular plates of puparium.
Fig. 586. Melanagromyza apii: aedeagus.
Fig. 587. Melanagromyza chaerophylli: aedeagus.
Fig. 588. Melanagromyza ferulae: inner margin of epandrium.
Fig. 589. Melanagromyzafoeniculi: aedeagus.
Fig. 590. Melanagromyza limata: aedeagus.
Subclass ROSIDAE 163
confused with M. angelicae until examination of There is nothing in the genitalia of these Melanagro-
the genitalia (Fig. 584) revealed that the Nearctic myza species to suggest that the seven in Europe are
and Palaearctic species on Angelica are distinct. not monophyletic. They are closely related to spe-
Normal host Angelica sylvestris, occasionally found cies on both Asteraceae and Leguminosae and it is
on Heracleum and Pastinaca. Posterior spiracular impossible to reconstruct accurately the evolution-
plates of puparium (Fig. 585) adjoining. ary progression on these families.
M. apii Hering (cf. Spencer, 1977c). Only known Only the single Ophiomyia, 0. heracleivora on
in Australia on cultivated celery, Apium graveolens, Heracleum (d. Spencer, 1964c), is known on the
which has probably been colonized from a wide- Apiaceae, in Europe. The genitalia (Fig. 593) are
spread local species, such as A. australe. The gen- simplified and atypical of the genus; the posterior
italia (Fig. 586) are typical of the genus but with 27 larval spiracles (Fig. 594) agree closely with those of
Melanagromyza species known in Australia it is not 0. orbiculata on Pisum (Fig. 455).
possible to postulate its exact affiliations. Three Napomyza species are known on Anthris-
M. chaerophyl/i Spencer (1969a) is only known cus, Bupleurum and Daucus in Europe. N. glabra
from East Germany, reared from stems of Chaero- (Hendel, 1935) feeds internally in Anthriscus stems.
phyllum bulbosum. Male genitalia: Fig. 587. The genitalia (Fig. 595) are somewhat atypical of
M. ferulae Spencer (1966b). Known from a single Napomyza and indicate an early switch to Anthris-
male caught on Ferula communis on Mt. Etna, cus. N. carotae Spencer (1966d) attacks Daucus
Sicily. The male genitalia are typical but the cerci are carota, the larva sometimes forming a short leaf
unusual, bearing some eight strong spines on the mine, but rapidly moves to the stem via the petiole
inner margin (Fig. 588). and not infrequently then down into the carrot root
M. foeniculi Spencer (d. Spencer, 1966b). Only (Fig. 596). The genitalia (Fig. 597) are entirely
known from the Barcelona area of Spain, feeding in typical of the genus, showing little differentiation
stems of Foeniculum vulgare. Male genitalia: Fig. from species feeding on Asteraceae and there has
589. thus been a relatively recent switch between these
M. limata Spencer (1971b). Described from families. Hering (1957b, hand-written note in his
single specimen caught in North Wales, (Fig. 590), personal copy) records Petroselinum as a further
subsequently reared from Heracleum sphondylium host of this species. I have not been able to confirm
in South Wales by S. Derji, a PhD. student at Car- this record but during larval transplantation experi-
diff University. ments in Switzerland (Wiesmann, 1961) found that
M. lomatii Steyskal (d. Spencer and Steyskal, larvae developed normally in Apium, Heracleum,
1986). Known only from Oregon, U.S.A., reared Pastinaca and Petroselinum. The natural host range
from Lomatium nudicaule. Lomatium is the largest of N. carotae thus needs further investigation.
genus in the family in U.S.A., with 80 species. Many Phytomyza bupleuri Hering (1963) on Bupleu-
were originally included in Peucedanum, and Loma- rum falcatum has been found correctly to belong in
tium is also closely related to Angelica and Hera- Napomyza to which it is formally transferred in the
cleum. The genitalia (Fig. 591), with the wide gap Appendix below. The head with the enlarged third
between the basal sclerites and the distiphallus antennal segment (Fig. 598) and the male genitalia
complex suggest relationship with M. angelicae (Fig. (Figs. 599, 600) are characteristic of Napomyza.
582). This might represent a case of coevolution but Nowakowski (1962: 105) was the first to attempt
there is no detailed evidence to confirm this. to indicate relationships among the Phytomyza leaf
M. nibletti Spencer (d. Spencer, 1966b). Only miners on Apiaceae. His work was followed by
known in southern England ex stems of Silaum Griffiths (1973a) in an exemplary study of the
silaus. The genitalia (Fig. 592) are entirely typical of species on Angelica, Heracleum, Laserpitium and
the genus. Pastinaca in Europe and North America.
M. sativae Spencer (d. Spencer, 1966b). Reared On the basis of my study of the 64 Phytomyza
originally from Pastinaca sativa and Tori/is japonica species now known on Apiaceae, I feel it is not
at the same locality in southern England, and sub- wholly appropriate to apply the term 'albiceps'-
sequently found in northern Iran in stems of group to the large number of species on Apiaceae
Pimpinella (Griffiths, 1964a). Closely resembling which can be associated with P. albiceps on Astera-
M. nibletti. ceae. In the sense used by Nowakowski and Griffiths
Melanagromyza splendida Frick (1953) was de- albiceps was treated as the common miner on Ar-
scribed from Apium graveolens in Hawaii but as far temisia. However, examination of the type in Vienna
as I know has not subsequently been found on this showed that this is in fact the exceptionally large
host and is now well known as an oligophagous species later described by Hering (1949d) as P. ry-
feeder on Asteraceae. Its occurrence on Apium can deniana (Fig. 958) feeding on Cirsium heterophyl-
best be treated as xenophagy (see Section 5). fum (see under Asteraceae) and a new name was
164 Class MAGNOLIOPSIDA (Dicotyledons)
592
594
595
~
.. '"
~
.. 4~
.. '
····················•·····
..... ,"
:.......
596
·· ,.:.::: ...
A.. f t · ·
I .. .
Ij '".
597
~.'~(~
~'~?;~~
599 .<:)
602
601
604
~.:~;::J!t>
\.,
"
605
607
608
Subclass ROSIDAE 167
face exit slit and falling from the leaf on pupation, Athamanta. P. athamantae Hering (1943) is only
while in P. conii the larva makes an exit slit through known from Austria. Athamanta is essentially a
the upper leaf surface and the puparium remains Mediterranean genus where 15 species are known.
glued to the leaf beyond the end of the mine. The genitalia (holotype, Fig. 614) are typical of the
P. coniopais Hering (1931b) was originally dis- angelicae group.
covered at the same locality as P. chaerophylliana Bupleurum. Three species have been described
near the River Oder on the German-Polish border from this genus (P. bupleuri now transferred to
and these two very similar yellow-frons species are Napomyza, p. 163), while a fourth is known only
each restricted to their own host and with very dif- from its distinctive mine (Hering, 1957b: III, 208),
ferent genitalia - Chaerophyllum (Fig. 603) and which begins with a long, narrow channel on the
Conium (Figs. 606, 607). P. coniophila Hering lower surface of the leaf. P. elsae Hendel (cf. Hen-
(1931a), with the frons black, represents P. chaero- del, 1935) is unusual in having the third antennal
phylli. segment greatly enlarged but the genitalia (Figs.
Scaligeria. A largely black form but with genitalia 615, 616) suggest it is affiliated to the angelicae-
as in P. ferulae ssp. ferulae was reared from Scali- group. P. facialis Kaltenbach (cf. Hendel, 1935) is
geria cretica on the Adriatic island Hvar, Yugoslavia the species with the longest distal tubules of the
and described by Hering (1967) as P. ferulae ssp. aedeagus in the whole angelicae-group (Figs. 617,
scaligeriae. As this subspecies is only known in 618). Examination of the male lectotype of P. inuli-
association with its host in this tribe, I accept it as cola Hering (1937) has shown that this represents a
valid, although the genitalia are as in the typical synonym of P. facialis (cf. Appendix). Hering mis-
subspecies. takenly believed that the type series, which was
Smyrnium. P. smyrnii Spencer (1954b) is known caught on Inula hirta, were feeding on this host,
from a single female reared from S. olusatrum in whereas they were merely resting on the plants or
Portugal. Its affiliations cannot be further clarified otherwise attracted to them, possibly due to their
but the wing of the holotype is shown in Fig. 608 being slightly aromatic - which is very pronounced
and the leaf mines on Smyrnium in Fig. 609. in Inula conyza.
Carum. The mines which are not uncommon on
Tribe 5. Apieae (formerly Ammineae) this genus represent P. chaerophylli (see Tribe 2)
This is the largest tribe and supports 25 Phytomyza but Hering (1957b) also records P. mylini (see Seli-
species which are tribe-specific, the majority feeding num below).
on a single genus. Cicuta. Three yellow-frons species are known on
Aegopodium. Three host-specific European spe- this genus, two in Europe, one in California. P. cicu-
cies are known here and also occasionally P. angeli- tae Hendel (cf. Spencer, 1976a) and P. cicutivora
cae (Tribe 6). P. obscurella Fallen (cf. Spencer, Hering (cf. Hendel, 1935) both feed on Cicuta viro-
1976a) is common on A. podagra ria and, although sa, the larvae forming linear mines. They are not
with the frons dark, the genitalia (Fig. 610) show directly related, P. cicutae belonging to the spon-
that it is closely related to the yellow-frons species dy/ii-group (Fig. 619), P. cicutivora to the angelicae-
such as spondylii (Fig. 653). P. podagrariae Hering group (Fig. 621). P. cicutella Spencer (1981) on
(1957b) is only known from the Botanical Gardens, C. douglasii in California is related to P. cicutivora
Berlin on the same host and Hering correctly diag- in Europe and is only slightly differentiated from
nosed it as distinct from P. obscurella on the basis of P. angelicae (cf. Figs. 641, 642) (genitalia: Figs. 622,
differences in the leaf mines and with the puparium 623). Both form blotch mines.
being brown, not black. The genitalia (Fig. 611) The larval posterior spiracles in P. cicutae are
show that it is closely related to P. obscurella. The highly modified (Fig. 620), with numerous spiracu-
third species, P. pubicornis Hendel (1920) which is lar pores which are arranged on two large 'arms.'
widespread in western Europe appears like a typical The host plant grows in an aquatic environment and
member of the obscurella group but the genitalia this modification is primarily to facilitate respira-
(Figs. 612, 613) indicate that it is an isolated species, tion, with the enlarged structure interpreted by
not closely related to either of the other two main Hering, followed by Nowakowski (1962) as a 'cling-
groups. Its affiliations are unclear. ing organ' to prevent the puparium falling from the
\~~~
; ,I,
612
610
613
615
614 616
617 618
Fig.610. Phytomyza obscurella: aedeagus.
Fig.611. Phytomyza podagrariae: aedeagus.
Figs. 612, 613. Phytomyza pubicornis: 612, aedeagus, side view; 613, distiphallus, ventral view.
Fig. 614. Phytomyza athamantae: aedeagus.
Fig. 615, 616. Phytomyza elsae: 615, aedeagus, side view; 616, distiphallus, ventral view.
Figs. 617, 618. Phytomyza!acialis: 617, aedeagus, side view; 618, distiphallus, ventral view.
Subclass ROSIDAE 169
mine into the water below. It is interesting that this and Chromatomyia hortieola have been found here.
represents a comparable, parallel development to Pimpinella. Of the four main species known on
that which has evolved in the two feeders on Caltha this genus, all in Europe, two, P. adjuneta Hering
(Ranunculaceae), P. ealthivora (Fig. 48) and P. eal- and P. melana Hering have the frons dark and
thophila whose hosts also grow in or near water. belong to the obseurella group, while in P. pauliloe-
Cnidium. P. enidii Griffiths (1973a), known only wii Hendel and P. pimpinellae Hendel (d. Spencer,
in north-west Canada, is a typical member of the 1976a for the four species) it is yellow and they
obseurella group. belong to the angelieae group.
Conopodium. Only P. ehaerophylli is known on P. adjuneta and P. melana are generally similar
this genus (see Tribe 2). and can occur together on the same plant but the
Foeniculum. P. femlae Hering (1927a) is com- mines of adjuneta are normally longer and rigidly
mon in the Mediterranean area, described from adjoin the leaf margin. The genitalia show consider-
Femia (Tribe 6) but also present on Daueus (Tribe able divergence (adjuneta: Fig. 627; melana: Fig.
8). Empty mines found by Buhr in Corsica can with 628). P. pauliloewii (Fig. 629) and P. pimpinellae
confidence be ascribed to P. ferulae. (Figs. 630, 631) are obviously closely related, with
Ligusticum. In Europe P. mylini (main host Seli- the male genitalia of the same form but differing
num) and Chr. hortieola are known on this genus. In considerably in detail. The differing larval feeding
California I reared 2 females from mines on L. grayi instinct is reflected in the very different mines of the
(cf. Spencer and Steyskal, 1986: 323) which I accept two species. P. pauliloewii forms a compact blotch,
as representing an undescribed species. The larva while P. pimpinellae forms a short, rather broad
forms narrow mines following the serrations of the linear mine (Fig. 632), which is readily distinguish-
leaf and the species belongs to the obseurella group. able from the longer, narrower mines of adjuneta
Meum (= Mutellina). Beiger (1961) described and melana.
P. mutellinae from the Tatra Mountains, Poland, P. heracleana has been recorded on Pimpinella
recorded as reared from Mutellina purpurea. Similar (Griffiths, 1973b) but this is probably not a normal
empty mines were recorded by Hering (1957b) host.
from the Rila Mountains, Bulgaria. The male gen- Selinum (= Mylinum). The two yellow-frons
italia (Fig. 624) show that this species belongs to the species on Selinum were described from Germany,
spondy/ii group. P. selini Hering (d. Hendel, 1935) from the former
Oenanthe. Four species have been described on German-Polish border near the River Oder and
this genus, two in Europe (P. oenanthiea Hering P. mylini Hering (1957b) from the Botanical Gar-
(1949c) and P. phellandrii Hering, 1957b), one in dens, Berlin (named after Mylinum when this was
Japan (P. oenanthes Sasakawa (cf. Sasakawa, 1961» considered as a genus distinct from Selinum).
and one in California (P. oenanthoides Spencer, Hering correctly diagnosed the two species as dis-
1981). The frons is dark in P. phellandrii, yellow in tinct, with minor differences in the mines and larval
the other three. The genitalia of P. oenanthes (Fig. characters. The genitalia (mylini: Figs. 633, 634 and
625) and P. oenanthoides (Fig. 626) are of the same selini: Figs. 635, 636) show the two species to be
general form and indicate close relationship, both closely related, both belonging to the angelieae
in the spondylii group. The relationship between group. Presumably they speciated in isolation and it
p. oenanthes in Japan and P. cieutae in Europe (Fig. is not known whether their distributions now over-
619) appears even closer, with the genitalia being lap. Hering (1957b) records P. mylini as also feed-
virtually identical but the larval posterior spiracles ing on Ligustieum and Daueus and (1957b: 979)
are less enlarged in P. oenanthes. It seems clear that possibly also on Seseli.
there has been a relatively recent switch between Seseli (= Libanotis). Two host-specific yellow-
these hosts which has resulted in two only slightly frons species are known on Seseli, both in Germany.
differentiated species on Cieuta (Europe) and P. libanotidis Hering (d. Spencer, 1976a) (de-
Oenanthe (Japan). scribed after Libanotis when this was considered a
Only females are known of P. oenanthiea and genus distinct from Seseli) forms slightly different
P. phellandrii and their exact status cannot be estab- linear mines from those of P. seseleos Hering
lished. All, however, have the enlarged posterior (1957b) and there are also larval differences. The
spiracles in larva (and puparium), present also in genitalia of P. libanotidis (Fig. 637) show this to be
P. cieutae (see above). closely related to P. heracleana on Heracleum (Fig.
Petroselinum. No Phytomyza is known on this 659) but only females are known of P. seseleos and
genus but young mines of Napomyza earotae in the its affinities cannot be established.
leaves could be mistaken for a Phytomyza (see The oligophagous P. heracleana and P. mylini
under N. earotae above). have been recorded on Seseli by Hering (1957b).
The two polyphagous species Liriomyza bryoniae Silaum. Only P. silai Hering (1935) is known in
170 Class MAGNOLIOPSIDA (Dicotyledons)
620
622
623
.. ..
~
624
626
//
/.::" ......
627,.
Figs. 619, 620. Phytomyza cicutae: 619, aedeagus; 620, posterior spiracles ofpuparium.
Fig.621. Phytomyza cicutivora: aedeagus.
Figs. 622, 623. Phytomyza cicutella: 622, aedeagus, side view; 623, distiphallus, ventral view,
Fig. 624. Phytomyza mutellinae: aedeagus,
Fig.625. Phytomyza oenanthes: aedeagus (Sasakawa).
Fig. 626. Phytomyza oenanthoides: aedeagus,
Fig. 627. Phytomyza adjuncta: aedeagus,
Fig. 628. Phytomyza melana: aedeagus,
Subclass ROSIDAE 171
630
629
)
634
636
\t .c·,···· ~~.~
.. :,~t-...,'._~-.-.):-=.;;
''I(.,~{
/'
... --,/
638 _ .. -'
640
639 641
642
643 644
.~.~.-.-~: ".--".". ----~.::::::-.--
645 646
Figs. 638, 639. Phytomyza silai: 638, aedeagus; 639, front end of larva.
Fig. 640. Phytomyza sii: aedeagus.
Figs. 641, 642. Phytomyza angelicae: 641, aedeagus, side view; 642, distiphallus, ventral view.
Fig. 643. Phytomyza angelicivora: aedeagus (Griffiths),
Fig. 644. Phytomyza kibunensis: aedeagus (Sasakawa).
Fig. 645. Phytomyza angelicastri: aedeagus.
Fig. 646. Phytomyza archangelicae: aedeagus.
Subclass ROSIDAE 173
western Europe as host-specific on this genus. The (Fig. 644) show slight but distinct differences from
genitalia (Fig. 638) are extremely close to P. selini P. angelicae.
(Figs. 635, 636), with greater differentiation in the Hering (1957b) records P. heracleana on Angeli-
adults, the side of the thorax being more yellow in ca but as such records are based on mines alone, it
silai. The larvae of the two species are also not is felt that this requires confirmation from reared
obviously distinguishable, with similar posterior adults, as the blotch mines of the two species might
spiracles and each have the conspicuous 'frontal well be confused.
projection' (Fig. 639) above the mouth-parts. Larval P. angelicastri Hering (cf. Spencer, 1976a; Grif-
characters were discussed by Allen (1957). Hering fiths, 1973b) is common in western and northern
(1957b) also records here P. mylini. P. silaicomes Europe, forming an irregular linear mine. The gen-
Hering (1960) was described from Silaum in Eng- italia (Fig. 645) are characteristic of the spondylii
land but this species was found to be synonymous group. In P. archangelicae Hering (cf. Spencer,
with P. adjuncta (Spencer, 1971b). 1976a; Griffiths, 1973b) the genitalia (Fig. 646) are
Sison. P. sison is Hering (1943) was described somewhat modified but the species can also be
from S. amonum in northern France (leg. Buhr). assigned to the spondylii group. It is present in
Only the female holotype is known but I am satisfied northern Europe, the Faroe Islands and Alaska.
that this species merely represents P. chaerophylli P. arnaudi Sasakawa (cf. Sasakawa, 1961; Griffiths,
and this synonymy is formally established in the 1973b) in Japan is closely related to P. archangeli-
Appendix below. cae (Fig. 647). P. polycladae Sasakawa was de-
Sium (= Berula). P. sii Hering (d. Hendel, 1935) scribed from Angelica polyclada in Japan, with only
was described together with P. berulae Hering in the the unique female holotype. Later, Sasakawa (1961)
same paper, both from Mecklenburg in northern illustrated the genitalia but added the additional
Germany. Hering admitted that there were no sig- host Sanicula for the species, but he did not make
nificant differences between the adults on the two clear whether the illustration refers to a male from
hosts but considered the differences in the mines Angelica or Sanicula. I illustrate here the genitalia of
indicated distinct species. This was at a time when a male (Fig. 648) reared from Angelica decursiva at
he considered that each genus in the Apiaceae had Yazu-gum, Tottori Pref. (leg. Kuroda). This differs
its own species of miner. De Meijere (1937: 230) slightly from the illustration given by Sasakawa
found no differences between larvae from Sium and (1961: fig. 141d) but closely resembles that of
Berula and therefore synonymised P. berulae with P. saniculae from Sanicula in California. The spe-
sii. It is now accepted that Berula is in fact synony- cies in Japan on Angelica and Sanicula may thus
mous with Sium. The genitalia of P. sii (Fig. 640) indeed be identical as recorded by Sasakawa but
are typical of the spondylii group. equally well they may represent distinct but very
similar species. In either case there must have been a
Tribe 6. Peucedaneae switch at some time of an ancestral species between
Sixteen tribe-specific Phytomyza species are known these hosts in separate subfamilies.
here, in addition there are four which cross tribal Conioselinum. Two further species in the obscu-
boundaries. rella group, P. conioselini Griffiths (1973a) (Fig.
Angelica (= Archangelica). This genus has been 649) and P. sitchensis Griffiths (1973a) (Fig. 650)
colonized by six species which are monophagous, were described from Alaska, and P. sitchensis has
three in Europe and three in Japan, and also since been discovered in the mountains in southern
P. angelicae Kaltenbach (cf. Spencer, 1976a; Grif- Germany (von Tschirnhaus) and in England (Hen-
fiths, 1973b) which also feeds on Heracleum and shaw and Irwin). These two species are clearly
Laserpitium, and is known both in Europe and N.W. closely related, although the genitalia are well dif-
America. Its genitalia (Figs. 641, 642) are character- ferentiated. A third species on Conioselinum dis-
istic of the 19 species referred to as the angelicae covered by Griffiths in Alaska forms blotch mines
group. P. angelicivora Hering (cf. Griffiths, 1973b) and possibly represents P. angelicae (Griffiths,
closely resembles P. angelicae but forms a more 1973a: fig. 14). Empty mines were found by Buhr
linear mine rather than the large blotch of angelicae near Leningrad which Hering (19 57b) considered to
and is of more limited distribution, confirmed only represent an undescribed species.
at Berlin, in eastern Germany and western U.S.S.R. Ferula. Two closely related species are known on
The genitalia (Fig. 643) show its close relationship this genus, both occurring primarily in the Mediter-
also with P. selini and P. silai. P. kibunensis Sasa- ranean area. P. ferulae ssp. ferulae Hering (1927a)
kawa (cf. Griffiths, 1973b) was described as a sub- (Fig. 651) was described from Gran Canaria, Cana-
species of angelicae in Japan but was raised to ry Islands and has since been found on most Medi-
specific status by Griffiths (1973b) and the genitalia terranean islands and in Israel. A secondary host, on
174 Class MAGNOLIOPSIDA (Dicotyledons)
647 648
649 650
, .
'., \,
, .
653
Fig.647. Phytomyza arnaudi: aedeagus (Sasakawa).
Fig.648. Phytomyza polycladae: aedeagus.
Fig. 649. Phytomyza conioselini: aedeagus (Griffiths).
Fig.650. Phytomyza sitchensis: aedeagus (Griffiths).
Fig.651. Phytomyzaferulae: aedeagus.
Fig.652. Phytomyza ferulivora: aedeagus.
Fig.653. Phytomyza spondylii, subspecies spondylii: aedeagus (Griffiths).
Fig.654. Phytomyza spondylii subspecies heracleiphaga: aedeagus (Griffiths).
Subclass ROSIDAE 175
which it occurs with equal frequency to Ferula, is 1973b) (Fig. 658) forms yellow, interparenchymal
Daucus carota (= P. dauci Spencer (1957d) de- mines somewhat similar to those of P. tlingitica and
scribed from southern Spain). A dark form treated it seems probable that these two are sister-species
as ferulae ssp. scaligeriae is accepted as valid, partic- each side of the Bering Straits. The genitalia provide
ularly in view of the different tribal position of the no evidence against this assumption. P. heracleana
host Scaligeria (see Tribe 4). P. ferulivora Griffiths Hering (cf. Griffiths, 1973b) is the only species on
(1956) (Fig. 652) was described from Rome and I Heracleum belonging to the angelicae group (gen-
have also found it on Mt. Etna, Sicily. The genitalia italia: Fig. 659) and is widespread throughout much
of the two species differ significantly and there is of Europe. The mines are distinctive, greenish, inter-
some colour difference between the adults of the parenchymal and frequently irregularly oval, con-
two species. fined in the angle between two veins. There are also
Heracleum, Pastinaca. Six species and one sub- records of this species on Angelica, Laser, Pastina-
species occur on Heracleum, two confined to ca, Peucedanum (Tribe 6), Pimpinella and Seseli
Europe, three to North America, while two have a (Tribe 5) and Caucalis (Tribe 8) but no reared
holarctic distribution. The close relationship of adults are available from these hosts to provide
Heracleum and Pastinaca is indicated by three spe- confirmation of the identification. Most are prob-
cies, P. spondylii, P. pastinacae (spondylii group) ably accurate.
and P. heracleana (angelicae group) feeding on Thus, while P. sphondyllivora, spondylii ssp. hera-
both, but equally their differences are emphasised cleiphaga, lanati and tlingitica are strictly mono-
by P. sphondyliivora in Europe and P. lanati in phagous on Heracleum, P. spondylii spondylii and
California which feed only on Heracleum. pastinacae show limited oligophagy, and the exact
The common P. spondylii Robineau-Desvoidy host range of P. heracleana requires further clarifi-
(cf. Spencer, 1976a; Griffiths, 1973b) forms white cation.
linear mines on Heracleum sphondylium in Europe. Laser. Hering (1957b) records P. heracleana
A similar species known in California and found by here and also empty mines possibly referable to two
Griffiths in Alaska is treated as subspecies heraclei- further Phytomyza spp.
phaga. The European subspecies has fewer spinules Levisticum. P. spondylii recorded by Hering
above the right basal sclerite or they may be entirely (1957b).
lacking (Fig. 653) and is smaller than ssp. heraclei- Myrrhidendron. Empty mines I found on M. don-
phaga Spencer (1969b) (Fig. 654) which was re- nell-smithii near the rim of Vulcan Po as (2700 m.),
duced to subspecific status by Griffiths (1973b). In Costa Rica (30 km north of San Jose) certainly
Europe spondylii also mines Pastinaca, from which represent an undescribed species. The stems of this
P. pastinacae Hendel (cf. Griffiths, 1973b) was host show some aspects of woodiness, with 'the
described but this species also occurs on Heracleum. structure of the xylem that of normal dicotyledon
With no external difference between the adults, the wood' (Heywood, 1971: 74). The Phytomyza occur-
two species can only be separated by the small but ring here could be of considerable interest and
consistent differences in the genitalia. The large hopefully adults will be obtained in due course.
group of spinules between the basal sclerites in Pastinaca. For P. heracleana, pastinacae and
P. pastinacae (Fig. 655) is always lacking in P. spon- P. spondylii see above under Heracleum.
dylii, as also the two sclerites A and B in spondylii Peucedanum. Three host-specific species are
(Fig. 653) which is present on both hosts in Europe. known on this genus. Two were described from
P. pastinacae is found on Heracleum lanatum in P. palustre (when this host was included in Thys-
Alberta and on Angelica atropurpurea in New York. selinum), P. thysselini Hendel (cf. Spencer, 1976a),
P. lanati Spencer (1966c; Griffiths, 1973b) feeds on a typical member of the obscurella-group (Fig. 660)
H. lanatum in California and Alberta and, although and P. thysselinivora Hering (cf. Spencer, 1976a), a
superficially resembling P. pastinacae and spondylii slightly aberrant member of the angelicae-group
ssp. heracleiphaga, the genitalia (Figs. 656, 657) are (Fig. 661). P. peucedani Ryden (cf. Spencer, 1976a)
significantly different. P. tlingitica Griffiths (1973b) forms blotch mines on P. cervaria (and was mis-
was described from Alaska on H. lanatum and sub- identified by Hendel (1935: 426) as marginella
sequently found in California. Although both adults Fallen). The genitalia (Fig. 662) indicate close rela-
and genitalia closely resemble the other species tionship with P. thysselinivora.
known in North America on Heracleum, the leaf Sphenosciadium. Empty mines on S. capitellatum
mines differ substantially and are greenish and inter- found at the Luther Pass, California (leg. K. Frick)
parenchymal, not white and on the upper surface as almost certainly represent an undescribed species.
in the other species. The regular blotch at the apex of a leaf segment (cf.
Two further species on Heracleum are confined Spencer and Steyskal, 1986: fig. 1361) is distinctive
to Europe. P. sphondyliivora Spencer (cf. Griffiths, and cannot be ascribed to P. angelicae which forms
176 Class MAGNOLIOPSIDA (Dicotyledons)
655
658 659
660 661
- - - . t.
- - - - --
~
,
663
Subclass ROSIDAE 177
much larger blotches in the centre of the leaf. The Agromyzidae on the Apiaceae are remarkable
Tordylium. P. tordylii Hendel (1927) is now ac- by the unique imbalance of their generic composi-
cepted as a synonym of P. chaerophylli. The two tion, with 64 species (or 81% of the total of 79
polyphagous species, L. strigata and Chr. horticola which are family-specific) in the one genus Phyto-
have also been recorded here. myza. This is combined with the complete absence
of Agromyza, and Liriomyza represented only by
Tribe 7. Laserpitieae four polyphagous species. This is in striking contrast
Laserpitium. Two species are known only on L. lati- with the larger assemblage on the Leguminosae,
folium with their distribution overlapping in Poland. where there are 94 species, with only six in Phyto-
P. latifolii Groschke (cf. Griffiths, 1973b) is not myza but 16 in Agromyza and 14 in Liriomyza
uncommon on Laserpitium in mountains in central (Table 19).
Europe, extending eastwards to Poland. It can be The degree of generic specificity is high. In Mel-
included in the angelicae group in a broad sense but anagromyza, of the 11 species, six feed on a single
the aedeagus (Fig. 663), with the distal tubules genus; M. angeliciphaga whose main host is Ange-
recurved, closely resembles the form present in lica occasionally transfers to Heracleum and Pasti-
P. hendeli (Fig. 69) (Anemone) and P. ranunculi- naca (in the same tribe), but M. sativae has been
vora (Fig. 160) (Ranunculus). Griffiths (1973b: found on Pastinaca, Pimpinella and Torilis con-
233) suggested the possible relationship of the sidered to be in three different tribes. The Ophio-
angelicae group with some groups of Ranuncula- myia and Calycomyza known here both feed on a
ceae feeders. I have discussed this earlier under single genus, while Napomyza glabra is known on
Ranunculaceae, and P. latifolii strongly suggests Anthriscus and Chaerophyllum (closely related
such a relationship, with, I suggest, a switch from genera), N. bupleuri only on Bupleurum and N. ca-
Ranunculaceae to Apiaceae. rotae only on Daucus (probably).
P. zarzyckii Nowakowski (1975) is apparently Of the 64 Phytomyza species, 49 (76%) are
widespread in mountains in Poland and closely restricted to a single genus, six are found on two or
resembles P. pimpinellae, both externally and in the three genera, while three - P. chaerophylli, P. herac-
form of the genitalia. It can occur together with leana and P. mylini have a substantially wider host
P. latifolii. range. The species showing the widest oligophagy is
The blotch miner which is more common on P. chaerophylli feeding on 12 genera in six tribes.
Laserpitum represents P. angelicae. Hering (1957b) Only two are known to cross subfamily boundaries:
also records P. heracleana based on mines only on P. spondylii whose normal hosts are Heracleum and
Caucalis and P. mylini on Daucus (Tribe 8, Dau- Pastinaca but which I found on Astrantia sp. at the
ceae) but it is felt that such records require confir- Royal Botanical Gardens, Kew - with so many
mation from reared adults. possible hosts growing in close proximity in Botan-
ical Gardens, such unusual transfers are not uncom-
Tribe 8. Dauceae
mon, with many comparable records discovered by
Caucalis and Tori/is have traditionally been in- Buhr in the Rostock Botanical Gardens, East Ger-
cluded in the Scandiceae but following recent stu- many; and in Japan P. polycladae which has been
dies, they are now placed in the Dauceae (Heywood, identified on Sanicula and Angelica.
1971: 37). No host specific agromyzids are known The dominance of Phytomyza in the Palaearctic
on the tribe. Region is indicated by the presence of 50 species on
P. chaerophylli has been found on Caucalis, Apiaceae, as against only 11 in North America
Daucus and Tori/is in north-west Europe (once (including one in Costa Rica). Just three species -
described as P. anthrisci daucivora Hering, 1957b), P. angelicae, P. pastinacae and P. sitchensis - are
and in the Mediterranean area P. ferulae is the most truly holarctic. P. angelicae which is common on
frequent species found on Daucus carota (described Angelica in Europe was found by Griffiths (1973b)
in southern Spain as P. dauci Spencer, 1957c). on both Angelica and Heracleum in Alaska and Al-
berta. On Heracleum and Pastinaca, P. pastinacae is
also common in Europe and is now known on Ange- The genitalia of Napomyza carotae (Fig. 597)
lica in New York and on Heracleum in Alberta. suggest relationship with N. latera lis on Asteraceae.
P. sitchensis, feeding on Conioselinum and de- I also draw attention to the possible relationship
scribed from Alaska, was later found in the Euro- between Ophiomyia heracleivora and 0. orbiculata
pean Alps and England. These three are presumably on Pisum (Leguminosae). The larval spiracles of
relatively recent immigrants into North America. these two are strikingly similar (0. heracleivora: Fig.
P. spondylii, whose normal host in Europe is Herac- 594, 0. orbiculata: Fig. 455). Morphological char-
leum sphondylium has been in North America long acters and the genitalia of these two species differ
enough to have evolved as a distinct subspecies, considerably and it therefore cannot be excluded
heracleiphaga, on H. lanatum. Close relationship that the similarity in the larvae merely represents an
but with greater differentiation is shown by P. brun- unusual case of convergence.
nipes (Europe) and P. saniculae (California), both
on Sanicula, and by P. sphondyliivora (Europe) and
P. tlingitica (Alaska), both on Heracleum. These Relationships of Agromyzidae on the Rosidae are
probably represent sister-species. A comparable detectable both with species on more primitive sub-
relationship is apparent in the two Melanagromyza classes, on different families within the Rosidae
species on Angelica, M. angelicae (North America) and also with species on the more advanced Astera-
and M. angeliciphaga (Europe). ceae.
Many other species, particularly on the same host The certain relationship between Agromyza po-
genus in Europe and North America, in which tentillae (which has a wide host range on the Rosa-
direct relationship is not so apparent, are doubtless ceae), and A. trebinjensis (Europe) and A. varifrons
derived from even earlier migrations of ancestral (U.S.A.), both feeding on Ulmaceae (Hamamelidae),
species across Beringia. and with A. alnivora, A. alnibetulae (Europe) and
The acceptability of the Apiaceae to polyphagous A. betulae (Japan), all on Betulaceae, was already
species is limited. Records of the five Liriomyza noted under the Hamamelidae. There is a similar
species are scattered and relatively uncommon, relationship between A. potentillae and three Agro-
while Chromatomyia horticola, which is the most myza species on the Salicaceae (Dilleniidae) -
polyphagous of all species in the family is known on A. albitarsis, A. isolata and A. salicina, and also
12 genera in both the Saniculoideae and Apioideae. with A. nigrescens on Geranium (Geraniaceae) and
The Agromyzidae on Apiaceae show clear rela- the johannae group on Leguminosae.
tionships with species on just two families, primarily Within the Rosidae there is obvious relationship
the Asteraceae but also the Ranunculaceae. This is between Phytobia amelanchieris on Amelanchier
most obvious with the Phytomyza species. The (Rosaceae) and Phytobia setosa (= Pb. aceris) on
species in the spondylii + obscurella groups, most Acer (Sapindaceae).
with the characteristic spinules in the area of the In the Apiales the Phytomyza angelicae group
basal sclerites of the aedeagus (Figs. 640, 645) can appears to be directly related to and derived from
be incontrovertibly associated with those in the ancestral species on the Ranunculacae (d. P. falla-
albiceps group on the Asteraceae (Figs. 948, 958). ciosa), while the P. spondy/ii group is related to the
This can only be explained by a switch of an ances- P. albiceps group on the Asteraceae.
tral species, almost certainly from Apiaceae to Thus with its intermediate evolutionary position
Asteraceae, with subsequent radiation on both the Rosidae have species derived from earlier sub-
families. classes and others which appear to be ancestral to
The genitalia of the Phytomyza angelicae group species on the more advanced subclass Asteridae.
(Figs. 641, 642) are strikingly similar to a number of There is evidence of some switching between fami-
species on the Ranunculaceae (Figs. 166, 167) and lies within the subclass but this is limited and there
it seems improbable that this can represent con- appears to be virtually no relationship between
vergence. A switch of an ancestral species between species on the two largest families, the Leguminosae
these two families is difficult to explain but seems to and Apiaceae.
have occurred. If my contention that the Phytomyza This pattern of host specialization indicates some
species on Ranunculaceae represent a primitive degree of logical colonization between related hosts,
assemblage which evolved from Napomyza species however distant the relationship, but there is much
is correct, then the switch must have been from the evidence of random host selection in many families.
Ranunculaceae to the Asteraceae, with perhaps an
intermediate step via the Aquifoliaceae (Figs. 541,
542).
Subclass ASTERIDAE
This subclass, as defined by Cronquist (1981), is of loids which are probably effective repellents to most
comparable size to the Rosidae, with about 60,000 predators.
species in 11 orders and 49 families. It has been Four of the six families have been colonized by
colonized by far the largest number of agromyzids Agromyzidae but only on the Gentianaceae has
of any subclass, with some 530 species. There has there been any significant radiation, with ten species
been substantial radiation on many families, most present in the single genus Chromatomyia. How-
notably on the Asteraceae. This large family makes ever, throughout the Old World Tropics from Afri-
up about one third of the subclass and supports ca, across Asia to northern Australia the two fami-
many more agromyzids than any family of either lies Apocynaceae and Asclepiadaceae appear to be
dicots or monocots, with 295 species (Table 52). widely acceptable to species in the genus Tropico-
Botanically, the families of Asteridae hang to- myia but only in this genus.
gether by their sympetalous flowers and are the The agromyzid genera known on the Gentianales
most advanced subclass of dicotyledons and 'pos- are shown in Table 27 and the host genera which
sibly the most recently evolved' (Cronquist, 1981). have been colonized in Table 28.
Cronquist suggests that the Asteridae are derived
from ancestral Rosidae, and adds: 'There is no
reason to suppose that the group originated before
the beginning of the Tertiary period.' Although
fossil pollen referred to the Apocynaceae dates
from the Palaeocene, 'The Asteridae began to playa Table 27
prominent role in the flora of the world only during
the Oligocene epoch.' Agromyzid genera on the families Apocynaceae,
Asclepiadaceae, Gentianaceae and Loganiaceae
As will be seen from the colonization by Agro-
myzidae of the various orders and families discussed Genus Species
below, in general each family has its own species and
it is the exception to find any obvious relationship Restricted Polyphagous
between the Agromyzidae on the different orders. to family
Colonization has thus taken place sufficiently early Apocynaceae
for the evolution of the distinct family assemblages
Tropicomyia 4
detectable today. Liriomyza 2
Asclepiadaceae
Order GENTIANALES (6 families,
5500 species) Tropicomyia 4
Liriomyza 2
This is a relatively recent order which on fossil Paraphytomyza 1
evidence does not antedate the Tertiary. Pollen
referable to the Apocynaceae has been recorded in Gentianaceae
the Palaeocene but macrofossils of the Gentiana- Tropicomyia 1
ceae are not known before the Miocene. It should be Ophiomyia 1
Liriomyza 1
added though that this is one of the few orders of the Chromatomyia 10
Asteridae which can be traced back beyond the Phytomyza 1
Eocene.
The most pronounced chemical characteristics of Loganiaceae
the order are the general occurrence of iridoid Tropicomyia 1
compounds, glycosides and certain groups of alka- Liriomyza I? 1
180 Class MAGNOLIOPSIDA (Dicotyledons)
Table 28
Order GENTIANALES
Agromyzidae on Apocynaceae, Asclepiadaceae, Gentianaceae and Loganiaceae
Apocynaceae
Allamanda Tropicomyia theae, T. polyphaga
Pachypodium T. flacourtiae
Plumeria T. polyphyta
Pteralyxia Liriomyza cocculi (?)
Thevetia T. flacourtiae (?), T. polyphaga
Vinca L. schmidti
Asciepiadaceae
Asclepias L. asciepiadis, L. subasclepiadis
Cryptolepis T. polyphaga (?)
Cryptostegia T. flacourtiae
Cynanchum Paraphytomyza metaplecicola
Hoya T. polyphyta, T. theae ?
Marsdenia T. polyp~yta
Gentianaceae
Blackstonia Chromatomyia blackstoniae sp. n.
Centaurium Chr. centaurii sp. n.
Enicostema Ophiomyia akbari
Gentiana L. amarellae, L. bryoniae; Chr. crawfurdiae,
( = Crawfurdia) Chr. gentianae, Chr. gentianella, Chr. gentii, Chr. hoppiella sp. n.,
Chr. pseudogentii, Chr. skuratowiczi; Phytomyza vernalis
Swertia Chr. swertiae
Loganiaceae
Dregia T. polyphaga
Labordia L. cocculi
Spigelia Liriomyza sp. ( ? blechi )
:. ...:.:.:.:...
665
664
666
669
• <,'
671
670 .~'. -.
673 672
Fig. 664. Liriomyza asclepiadis: aedeagus.
Fig. 665. Liriomyza subasclepiadis: aedeagus.
Fig. 666. Paraphytomyza metaplecicola: aedeagus (Sasakawa).
Figs. 667, 668. Tropicomyia polyphaga: 667, aedeagus, side view (holotype ex Tylophora); 668, same, side view (ex Cryptolepis).
Figs. 669, 670. Chromatomyia blackstoniae: 669, aedeagus, side view; 670, same, ventral view.
Figs. 671, 672. Chromatomyia centaurii: 671, aedeagus, side view; 672, supporting sclerites, dorsal view.
Fig. 673. Chromatomyia crawfurdiae: aedeagus (Sasakawa).
182 Class MAGNOLIOPSIDA (Dicotyledons)
675
676
678
\ ...
677
679
Figs. 674-676. Chromatomyia gentianae: 674, aedeagus, side view; 675, supporting sclerites, dorsal view; 676, leaf mines on Gentiana
(Hering).
Figs. 677-679. Chromatomyia gentianella: 677, aedeagus; 678, supporting sclerites; 679, leaf mines on Gentiana acaulis (Hering).
Figs. 680-682. Chromatomyia gentii: 680, aedeagus; 681, supporting sclerites; 682, leaf mine on Gentiana lutea (Hering).
Subclass ASTERIDAE 183
Also in U.S.A., known only in Washington State on from their leaf mines, a few on external characters
A. ovaliiolia, L. subasclepiadis Spencer (in Spencer but a positive identification may only be possible
and Steyskal, 1986) (Fig. 665) forms a linear-blotch from the male genitalia. Essential data on these ten
mine. The genitalia of the two species are generally Chromatomyia species are as follows:
similar, suggesting relatively recent speciation.
C. blackstoniae Spencer, sp.n. (see description in
In Japan Paraphytomyza metaplecicola Sasakawa
Appendix). Host Blackstonia perfoliata, adults
(1961) forms a small blotch mine on Metap,lex,is
known only in western Ireland, mines also in
japonica and Cynanchum dickinsii. The gemta,1Ia
England and northern France. Male genitalia
(Fig. 666) differ substantially from all other speCIes
(Figs. 669, 670) show that this species is strongly
in the genus and the affiliations of the species are not
differentiated and isolated from the species on
detectable.
Centaurium and Gentiana.
Polyphagous species attacking this family are
C. centaurii Spencer, sp.n. (see description in Ap-
known only in the essentially tropical genus Tropi-
pendix). Hosts Centaurium umbellatum, western
comyia. In southern Japan T styricicola is known on
Europe. Mine normally at base of leaf adjoining
Tylophora aristolochioides and in East Africa T fla-
stem, linear but developing into a blotch. Male
courtiae has been found on Cryptostegia. In Sri
genitalia: Figs. 671, 672.
Lanka T polyphaga (Spencer) was described from
C. crawfurdiae (Sasakawa, 1954a, 1961). Host
Tylophora flava (Fig. 667, 668) and in Australia
Gentiana (as Crawfurdia) japonica, Japan. Larva
T polyphyta was reared on T barbata. Also in
forming irregular linear mine. Male genitalia: Fig.
Australia T polyphyta has been recorded on Mars-
673. Vein tp present.
denia rostrata. C. gentianae (Hendel, 1920; 1935). Hosts Gentiana
In Thailand I found mines on Cryptolepis elegans spp., common in mountains in Europe, also pre-
probably referable to T polyphaga and in Papua sent in Japan. Mine somewhat stellate initially,
New Guinea at Bulolo mines on Roya sp. probably then developing into a large blotch (Fig. 676).
represent T theae. Genitalia (Figs. 674, 675).
C. gentianella (Hendel, 1932). Hosts Gentiana spp.,
Gentianaceae (75 genera, 1000 species) widespread in mountains in southern Europe,
also in Poland. Male genitalia (Figs. 677, 678)
This family is cosmopolitan but most common in
close to but obviously distinct from C. gentianae
temperate and subtropical regions and on tropical
and C. hoppiella sp.n., wing with vein tp present.
mountains.
Mine initially linear, then developing into large
The bitter taste of species in this family, resulting blotch (Fig. 679).
from terpenoids and iridoid compounds, acts as an
C. gentii (Hendel, 1920). Host probably Gentiana
effective deterrent to most phytophagous insects.
lutea, central Europe. Mine at apex of leaf, with
Nevertheless 11 agromyzids in Europe, one in India
several short channels running down the leaf blade
and one in Japan are restricted to the Gentianaceae,
which becomes deformed (Fig. 682). Genitalia
present mainly on Gentiana but also o~ Black~tonia, (Figs. 680, 681), with supporting sclerites largely
Centaurium, Enicostema and Swertza, ten m the
membranous. This species differs from all others
single genus Chromatomyia (see Tables 27 and 28):
in the group in having the frons pale yellow,
The genitalia in eight of the ten Chromatomyza
although the orbits on either side are black. Vein
species are generally similar, characteristic of other
tp present.
species in the genus on other families but they differ
C. hoppiella Spencer, sp.n. (see Appendix). Host
significantly in C. gentii and C. blackstoniae sp.n.
Gentiana excisa, known only in Switzerland.
On wing venation the species fall into two groups,
Originally confused with both C. gentii (Hering,
four with the outer cross-vein (tp) present, six with
1928) and C. gentianella (Hendel, 1932) but
this vein lacking. The presence of this vein can be
male genitalia (Figs. 683, 684) distinctive.
accepted as the retention of a plesiomorph~us
C. pseudogentii (Beiger, 1972a). Host Gentiana
character in four species but these are not otherwIse
punctata, Poland. Mine similar to that of C. gentii
clearly differentiated externally from the six species
(Fig. 682). Genitalia: Fig. 685. Puparium black
which have lost this vein.
(pale in all other species) but pupating in leaf.
The majority of these species are known in the Vein tp present.
mountains of Central Europe, with one in Japan and
C. skuratowiczi (Beiger, 1972b). Host Gentiana
two at low altitudes in western Europe. Three new verna. Poland. Larva forming large blotch mines
species have been discovered in this review, one on
in one or more leaves of the basal rosette (Beiger,
Gentiana in Switzerland and two which appear to be
1972b: fig. 8). Genitalia: Fig. 686. In this species
widespread in Western Europe on Blackstonia and
the wing lacks the outer cross-vein.
Centaurium. Some species can be reliably identified
184 Class MAGNOLIOPSIDA (Dicotyledons)
684
683
685 686
688
687
689 691
Figs. 683, 684. Chromatiomyia hoppiella: 683, aedeagus; 684, supporting sclerites.
Fig. 685. Chromatomyia pseudogentii: aedeagus (Beiger).
Fig. 686. Chromatomyia skuratowiczi: aedeagus (Beiger).
Figs. 687, 688. Chromatomyia swertiae: 687, aedeagus, side view; 688, supporting sclerites.
Figs. 689, 690. Liriomyza amarellae: 689, aedeagus, side view; 690, same, ventral view.
Figs. 691, 692. Phytomyza vernalis: 691, aedeagus, side view; 692, same, ventral view.
Subclass ASTERIDAE 185
C. swertiae (Hering, 1937). Host Swertia perennis. (included by some authors in a separate family
Mecklenburg, northern Germany, also in Poland Spigeliaceae) but it is possible that the species con-
in Alpine zone of Tatra Mountains (Beiger, cerned may be distinct from L. blechi. Its identity
Nowakowski). Mine linear at beginning, then cannot be further clarified without additional mate-
developing into a blotch. Genitalia: Figs. 687, rial.
688, similar to species mining Gentiana. Buddleja has often been included in this family
(and is retained in the Loganiaceae by Mabberley
Liriomyza amarellae Hering (1963: 230) was
(1987)) but is here accepted with family status in the
discovered in the mountains in southern Germany
Scrophulariales. This is strongly supported by agro-
mining Gentiana germanica and G. ciliata. The
myzid host selection (see Scrophulariaceae below).
mines are unusual, commencing in the stem, the
larva then following the midrib, before forming
short offshoots into the leaf blade. This may well Order Solanales (8 families, 5000 species)
therefore be a primitive species derived from a
Five of the eight families in this order have been
stem-mining ancestor. The genitalia (Figs. 689, 690)
colonized by Agromyzidae, the majority of the 34
appear to resemble some species on Aster~ceae but
species restricted to the order being on the .largest
the exact affiliation of this species remams to be
family, the Solanaceae. The Solanales are eVIdently
established.
related to the Gentianales but no relationships are
Phytomyza vernalis Groschke (1957) is. only
detectable between the agromyzids feeding on the
known on Gentiana verna in the mountams of
two orders. The Agromyzidae on the Solanaceae are
southern Germany. The larva forms blotch mines in
several basal leaves, moving via the stalks from one shown in Table 30 and on the Convolvulaceae, Cus-
cutaceae, Polemoniaceae and Hydrophyllaceae in
leaf to another, pupating externally (in contrast to
Table 32. The fossil record suggests that the order
the 10 Chromatomyia species). The male genitalia
does not antedate the Tertiary, the oldest macrofos-
(Figs. 691, 692) are aberrant and indicate long isola-
sils attributable to the Solanaceae occurring in the
tion on Gentiana, and its nearest relative cannot be
Eocene. It thus 'appears that most of the diversifica-
determined.
tion of the order into families is post-Eocene'
In India a leaf miner on Enicostema verticillatum
(Cronquist, 1981).
was described as Melanagromyza akbari Singh and
Ipe. It was transferred to Ophiomyia by Sasakawa
(1977). Solanaceae (85 genera, 2800 species)
Among European polyphagous species, only This family is virtually cosmopolitan but is best
Liriomyza bryoniae has been recorded on Gentiana.
developed in tropical South America. This is re-
It is noteworthy that only two species in Lirio- flected in the 21 described agromyzids restricted to
myza and one in Phytomyza have succeeded in
the Solanaceae, of which 16 are in South or Central
colonizing the Gentianaceae and in view of the America, two in Africa, one in Papua New Guinea
obvious general inacceptability of the family to leaf
and two in Australia. Only eight agromyzid genera
miners, it seems probable that the present group in are represented on the family, with significant radia-
Chromatomyia has radiated from a single ancestral
tion only in Haplopeodes, with eight species. In
colonization. How early this occurred and whether addition, there are seven polyphagous species, with
before the family split into the currently known
three in Europe, and four in North and South
genera with subsequent coevolution can only remain America. The breakdown by genera is shown in
a matter of speCUlation.
Table 29.
The Loganiaceae are largely tropical and have no Agromyzid genera on Solanaceae
family-specific agromyzids. Three species only are
known on the family. Tropicomyia polyphaga has Family-specific Polyphagous
species species
been recorded on Dregia volubi/is in Sri Lanka, and
in Hawaii leaf mines on Labordia sp. were recorded Phytobia 1 (larva only)
by Swezey (1913) which have been tentatively re- Melanagromyza 3
ferred to the polyphagous Liriomyza cocculi (see Ophiomyia 3
under Menispermaceae, Magnoliidae). Phytoliriomyza
Liriomyza 5 6
In Florida Liriomyza blechi Spencer (in Spencer Haplopeodes 8
and Stegmaier, 1973) (normally on Acanthaceae, Chromatomyia
see below) has once been recorded on Spigelia sp. Phytomyza
186 Class MAGNOLIOPSIDA (Dicotyledons)
695
693 __ 5 .• "
696
694
698
697
700
699
701
Figs. 693, 694. Melanagromyza tomaterae: 693, aedeagus, side view; 694, same, ventral view.
Figs. 695, 696. Melanagromyza solanidis: 695, aedeagus, side view; 696, same, ventral view.
Figs. 697-699. Ophiomyia solanicola: 697, aedeagus, side view; 698, same, ventral view; 699, leaf mine on Solanum prinophyllum.
Figs. 700, 70l. Ophiomyia ferox: 700, aedeagus, side view; 701, same, ventral view.
Subclass ASTERIDAE 187
I here follow the revision of the Solanaceae by genitalia (Figs. 702, 703) give no clear indication of
Hunziker (1979) who divides the family into two its nearest relatives. Liriomyza braziliensis Frost (cf.
subfamilies, the Solanoideae and Cestroideae, and Spencer, 1963c) occurs widely from Argentina to
11 tribes. All but two of the 21 host-specific species Colombia on Solanum tuberosum and S. andige-
are present on the Solanoideae and only one, Lirio- num feeding in the lower stem or potato tuber (Fig.
myza quadrata, is also found on Cestroideae. 704). This is the largest known Liriomyza but the
genitalia are typical of the genus and suggest rela-
tionship with L. quadrata (see below). However, the
Subfamily Solanoideae
surstyli (Fig. 705) are unique, indicating probable
Solaneae. Agromyzidae are known on six genera in evolution from Phytoliriomyza (see remarks below).
this tribe. Three stem-feeding Melanagromyza spe- Three further leaf mining Liriomyza species are
cies are known, two in South America and one in restricted to the Solanaceae, all in South America.
Africa. M. tomaterae Steyskal (cf. Spencer, 1985b) The species with the widest host range is L. quadra-
feeds in the stem of Lycopersicon esculentum and is ta Malloch (cf. Spencer, 1963c) which forms large,
known in Colombia, Ecuador and Venezuela. white linear-blotch mines on Lycopersieon, Physalis
M. caucensis Steyskal (cf. Spencer, 1985b) has only and Solanum (Fig. 706) and is widespread from
once been found on the same host in Colombia. The Colombia to Argentina. L. solanita Spencer
genitalia of M. tomaterae (Figs. 693, 694) are typical (1963c) is also not uncommon at high altitudes in
of many species in the genus and those of M. cau- Colombia and Venezuela on Datura arborea, Physa-
censis are little differentiated. Tomato is also severe- lis pubescens and has been found on Solanum mar-
ly attacked by the polyphagous species Liriomyza gina tum (Bogota) which is an Old World species, in-
sativae and L. trifolii in Florida and California, and troduced to South America by seeds in Persian car-
in Europe by L. trifolii and Chromatomyia horti- pets. The irregular linear mine (Fig. 707) of L. sola-
cola. nita is immediately distinguishable from that of
In Uganda M. solanidis Spencer (1959, 1965b) is L. quadrata and the genitalia suggest that they are
reported as causing galls in the stems of Solanum not closely related. In Venezuela L. solanivora
melongena. Whether or not these are true galls or Spencer (1973c) forms long, narrow mines on S. a-
merely swellings caused by internal feeding in the merieanum, S. bieolor, S. hirtum and S. rugosum
stem remains to be established. The genitalia (Figs. and the genitalia (Figs. 708, 709) confirm that this
695, 696) are distinctive in the large gap between species belongs to a different group to L. quadrata.
the short basal sclerites and the distiphallus com- There has been considerable radiation in the
plex. genus Haplopeodes on the Solanoideae in South
Three leaf-mining Ophiomyia species are known America with seven species known on the Solaneae
on Solanum, all in the Old World tropics from and two on the Lycieae (one undescribed). I found
Ethiopia, Kenya, South Africa and Madagascar to two species, H. bullati (Spencer, 1963c) and H. ver-
Papua New Guinea and N.E. Australia. 0. solani- bascifolii (Spencer, 1963c) on the tree solanums,
vora Spencer (1961f) is typical of the genus, feeding S. bullatum and S. erianthum near the Botanical
on S. indicum in Madagascar and South Africa Gardens at Sao Paulo, Brazil. The latter host, after
(Transvaal) and on S. incanum in Ethiopia and which H. verbascifolii was named, was unfortunately
Kenya. 0. ferox Spencer (1977a) was described misidentified locally in Brazil as S. verbascifolii and
from Papua New Guinea from leaf mines on Sola- has now been accurately identified as S. erianthum
num identified locally as S. ferox. However, the (S. Knapp, BMNH). Although more black than
hosts have now been accurately identified as S. tor- other species on Solanum and Amaranthaceae
voideum and S. torvium (det. S. Knapp, BMNH). which are pale grey or yellow, the reduced, simpli-
Empty mines found on S. schefferi at the same local- fied male genitalia (H. bullati: Fig. 710; H. verbas-
ity in P.N.G. as those of 0. ferox are also believed to cifolii: Fig. 711) confirm the close relationship of
represent this species. 0. ferox is probably related these two species with others in the genus. Hap-
to O. solanicola Spencer (1977c) which I obtained lopeodes eurhabdus Steyskal (1980) was described
at Clyde Mt., N.S.W., Australia forming mines (Fig. from Argentina 'ex Solanum elaeagnifolium.' It is
699) on S. prinophyllum. The genitalia (Figs. 697, accepted that the species was actually reared from
698) are of the same general form as in 0. ferox Solanum, although such labels can sometimes mean
(Figs. 700, 701). that the specimen was merely caught on the plant,
Two internal feeders are known on potato in and this does not necessarily imply that this is its
South America. Phytoliriomyza papae Spencer host. H. lopesi Oliveira and Silva (cf. Steyskal,
(1985b) has only been found in the stalks of Sola- 1980) was reared from S. argenteum at Rio de
num tuberosum in Bolivia. This is one of the few Janeiro, Brazil. H. kefi Steyskal (1980) was de-
species in the genus whose host is known and the scribed from S. umbelliferarum in California.
188 Class MAGNOLIOPSIDA (Dicotyledons)
.
("'"
"/~:\
,/~
~-
.' _IN
702 703
704 705
706 707
709
Figs. 702, 703. Phytoliriomyza papae: 702, aedeagus; 703, epandrium.
Figs. 704, 705. Liriomyza braziliensis: 704, mines on potato; 705 surstylus.
Fig. 706. Liriomyza quadrata: leaf mine on tomato.
Fig. 707. Liriomyza so/anita: leaf mines on Physalis peru viana.
Figs. 708, 709. Liriomyza solanivora: 708, aedeagus, side view; 709, distiphallus, ventral view.
Subclass ASTERIDAE 189
710 711
712 713
714 716
715
I possess a larva of an undescribed Phytobia sp. polyphagous species, Liriomyza cnptlca (nomen
obtained from the small tree, Solanum quitoensis, in nudum) has been identified by Valladares (unpub-
Colombia. This plant is well known and a sweet lished report, 1986) on Nicotiana glauca, the larvae
drink called locally 'lulo' is obtained from the fruits. forming greenish linear mines. In the same area the
Local entomologists are aware of this Phytobia sp. species also attacks Luffa (Cucurbitaceae) and Pois-
and hopefully adults will be obtained in due course. sonia (Leguminosae). L. cestri Spencer (1982)
Two further Haplopeodes species have been forms similar greenish mines on Cestrum parqui in
described from Argentina, H. capsici Spencer Chile but the genitalia of L. criptica (Fig. 714) sug-
(1973a) and H. flavinotus Valladares (unpublished gest closer relationship with L. sativae or L. bac-
thesis, 1982) both forming leaf mines on Capsicum charidis. Petunia serves as host for L. bryoniae and
annuum. Differences between the two are slight and Chr. horticola in Europe, and for Lir. huidobrensis
H. flavinotus is probably doubtfully distinct from and L. trifolii in Florida and South America.
H. capsici. Salpiglossideae. Three genera in this tribe have
Liriomyza bryoniae has been recorded on Atropa been colonized, the most remarkable case being
belladonna and, with its wide host range on 16 Anthocercis littorea in Western Australia, on which
families, this is the only species which has succeeded I found numerous leaf mines of a new species,
in adapting to the various toxic chemicals present in Phytomyza anthocercidis Spencer (1977 c), on this
this plant. yellow-flowering bush at the coast south of Perth.
Datureae. Liriomyza solanita has been found on This represents a striking switch from Ranuncula-
Datura arborea near Bogota and in addition three ceae, the adults closely resembling P. ranunculi
polyphagous Liriomyza species are known on Datu- which is not known in Australia, and the genitalia
ra, L. huidobrensis and L. sativae in South America (Fig. 715) with the coiled aedeagus clearly associ-
and L. bryoniae in Botanical Gardens in Europe. ated with both P. ranunculi and P. vitalbae (Fig. 94),
L. bryoniae has also been found on Scopolia carnio- the latter being present in Australia on Clematis
lica in Europe. pubescens.
Jaboroseae. An unidentified Haplopeodes sp. has I found numerous empty mines (Fig. 716) on
been found on Salpichroa in Argentina. L. bryoniae BrowalZia demissa in Costa Rica in and near San
has once been recorded on Withania aristata on the Jose. These almost certainly represent an unde-
Canary Islands. This plant is only known on the scribed Liriomyza sp. In Argentina the polyphagous
main islands of the Canaries group. L. schmidti has colonized Brunfelsia sp.
Lycieae. Haplopeodes lycivora Valladares (1982,
Relationships with species in a number of other
unpublished thesis) was described from Lycium
families can be detected in the Agromyzidae re-
cestroides in Argentina. This species appears to be
stricted to the Solanaceae, which can be briefly
close to H. minutus Frost which feeds on Amaran-
summarised as follows:
thus and Chenopodium (see Amaranthaceae). In
1. Phytomyza anthocercidis on Anthocercis in
Europe Lycium is attacked by L. bryoniae and
south-west Australia shows direct relationship
Chromatomyia horticola. An unidentified Haplope-
with species on Ranunculaceae. The male gen-
odes sp. has been found in Argentina by Valladares
italia (Fig. 715) reveal an unmistakable switch
(loc. cit.) on Grabowskia obtusa.
from species ancestral to either P. ranunculi
Nicandreae. Nicandra physalodes (a native of
(Ranunculus) or P. vitalbae (Clematis). Coloni-
Peru) has been colonised in Europe by L. bryoniae.
zation of Anthocercis may well have been forced
This is a highly poisonous plant (see remarks under
in the absence of the normal host of the ancestral
Atropa above).
species on Ranunculaceae. Initial attraction may
have been to the yellow flowers and presumably
Subfamily Cestroideae the female when ovipositing and the larva when
Cestreae. Only a single host-specific species has feeding found the new host acceptable, and a
colonized this tribe, with in addition five polyphag- new popUlation and species became established
ous species. In Chile I found Cestrum parqui to be on Anthocercis. It may be noted that no other
heavily attacked by Liriomyza cestri Spencer species in Phytomyza, the largest world genus, is
present on the Solanaceae.
(1982), the larvae forming short, irregular linear
2. Three generally similar leaf-mining Ophiomyia
mines. The genitalia (Figs. 712, 713) suggest prob-
able relationship with L. brassicae (Figs. 324, 325). species on Solanum in South Africa (and Mada-
gascar), Kenya, Papua New Guinea and Austra-
L. sativae and L. trifolii are also found on Cestrum.
L. quadrata is known on Nicotiana tabacum and in lia probably do not represent separate coloniza-
Europe this host is colonized by Chr. horticola. In tions. If they are strictly monophyletic, the area
of origin of the ancestral species and its subse-
Argentina (Tucuman, leg. Wygodzynski) a new
....
Subclass ASTERIDAE 191
---
;,.,"
719
717
720
718
.'~~'~ r"~'"
"\::"'""'."
~~ " ,
C"",
. -'. ',,~.,/
721 722
724
Figs. 717, 718. Melanagromyza albocilia: 717, aedeagus, side view; 718, same, ventral view,
Figs. 719, 720. Melanagromyza convolvuli: 719, aedeagus, side view; 720, same, ventral view.
Figs. 721, 722. Melanagromyza caerulea: 721, aedeagus, side view; 720, same, ventral view,
Figs. 723, 724. Melanagromyza ipomoeavora: 723, aedeagus (holotype); 724, same, (paratype).
quent dispersal to give rise to the three now 4. Liriomyza braziliensis, with its large size and
widely separated populations can only be specu- internal feeding habit, is a primitive species and
lative. this is supported by the structure of the surstyli
3. It is not certain that the Liriomyza species on (Fig. 705). The presence of four strong bristles in
Solanaceae are a monophyletic group. The only addition to the group of long hairs is unique
probable relationship is of L. cestri with L. bras- within the genus and is more characteristic of
sicae (Brassicaceae). Phytoliriomya than Liriomyza. However, the
192 Class MAGNOLIOPSIDA (Dicotyledons)
Table 30
Agromyzidae on Solanaceae
Classification after Hunziker (1979)
Subfamily Solanoideae
Solaneae
Atropa Liriomyza bryoniae
Capsicum L. bryoniae, L. huidobrensis, L. sativae; Haplopeodes capsici,
Hap. flavinotus
Hyoscyamus L. bryoniae, L. strigata; Chromatomyia horticola
Physalis L. quadrata, L. sativae, L. solanita, L. trifolii; Hap. kefi
Lycopersicon Melanagromyza caucensis, Mel. tomaterae; L. bryoniae, L. quadrata,
L. sativae, L. trifolii; Chr. horticola
Solanum Phytobia sp. ( larva only); Mel. solanidis; Ophiomyia ferox, O. solanicola,
O. solanivora; Phytoliriomyza papae; L. braziliensis, L. bryoniae,
L. quad rata, L. sativae, L. solanita, L. solanivora, L. strigata,
L. trifoliearum (?), L. trifolii; Hap. bullati, Hap. eurhabdus, Hap. kefi,
Hap. lopesi, Hap. verbascifolii; Chr. horticola
Datureae
Datura L. solanita, L. bryoniae, L. huidobrensis, L. sativae
Scopolia L. bryoniae
laboroseae
Saipichroa Hapiopeodes sp.
Withania L. bryoniae; Chr. horticoia
Lycieae
Grabowskia Hapiopeodes sp.
Lycium L. bryoniae; Hap. lycivora nom. nUd.; Chr. horticola
Nicandreae
Nicandra L. bryoniae
Subfamily Cestroideae
Cestreae
Cestrum L. cestri, L. sativae, L. tri folii
Nicotiana L. criptica nom. nud. ( Argentina ), L. quadrata; Chr. horticola
Petunia L. bryoniae, L. huidobrensis, L. trifolii; Chr. horticola
Salpiglossideae
Anthocercis Phytomyza anthocercidis
Browallia Liriomyza sp. n. ? ( Costa Rica)
Brunfelsia L. schmidti
Subclass ASTERIDAE 193
Convolvulaceae (100 genera, 1500 species) layensis Sasakawa (1963a: fig. 9), described from
Selangor, Malaysia, clearly belongs to the same
This family has been divided into eight tribes (Willis, group but differs more from the two specimens of
1985), of which only two have been colonized by M. ipomoeavora than they do from each other and
eight host-specific agromyzids and six polyphagous this can be accepted as a distinct species, with no
species (Tables 31, 32). indication of the host.
It is generally accepted that the Convolvulaceae Four Calycomyza species have been described
are closely related to the Solanaceae and this is from Ipomoea in South America, the Caribbean
supported by chemical characteristics, such as the area and southern Florida. C. ipomaeae (Frost, cf.
presence of tropan alkaloids and the absence of Spencer, 1963c) is the most widespread species,
iridoid compounds. However, there appears to be forming an irregular linear mine and with distinctive
no direct association between the Agromyzids on genitalia (Fig. 725). C. stegmaieri Spencer (in Spen-
the two families. cer and Stegmaier, 1973) has very different genitalia
and forms a circular blotch mine (Fig. 726), with the
Table 31 frass deposited in a more or less regular spiral. This
Agromyzid genera on Convolvulaceae
was described from Florida and is also present on
Guadeloupe. In Argentina C. brewerae (nom.
Species restricted Polyphagous emend. from breweri) Valladares (1981) (Fig. 727)
to family species is only known on 1. cairica, the larva forming a
linear-blotch mine. A new species, C. ipomoeapha-
Melanagromyza 4
Tropicomyia 2
ga Martinez (in Spencer et al., 1990) was recently
Liriomyza 3 discovered on Guadeloupe (genitalia: Fig. 728), the
Calycomyza 4 larvae forming a distinctive narrow linear mine.
Chromatomyia 1 Six polyphagous species are known on the family
- two Tropicomyia, three Liriomyza and Chromato-
8 6
myia horticola (see Table 32). In Cameroun, W. Af-
rica Buhr found mines (de Meijere, 1940: 169) on
Convolvuleae. Two closely-related internal stem- an unidentified Ipomoea sp. This possibly repre-
borers are known on Convolvulus arvensis. Melana- sents Tropicomyia flacourtiae. In Papua New Guinea
gromyza albocilia Hendel (cf. Spencer, 1966b) was I reared a species from Ipomoea alba which I accept
described from Austria and has since been recorded as the polyphagous T. theae.
in Egypt and Israel, and M. convolvuli Spencer (cf. Apart from some radiation of Calycomyza in
Spencer, 1973a) was described from Pakistan. South America, the Convolvulaceae have only been
These species are only separable by their distinct colonized by isolated species and are thus not gener-
male genitalia (M. albocilia: Figs. 717, 718 and ally acceptable to leaf miners. It is significant that the
M. convolvuli: Figs. 719, 720). dominant genus in South America, Liriomyza, has
Ipomoeeae. Two unrelated Melanagromyza spe- not succeeded in establishing any populations on
cies are known on Ipomoea. M. caerulea Malloch Convolvulaceae (apart from the polyphagous
(cf. Spencer, 1973a) occurs widely in Florida and L. schmidti).
the Gulf States, U.S.A. and throughout the Carib-
bean area, feeding in one or more of the four seeds
Cuscutaceae (1 genus, 150 species)
produced by each flower of 1. batatas and other
Ipomoea species. The distinctive genitalia are shown This family is included by some authors as a sub-
in Figs. 721, 722. family of the Convolvulaceae but I follow Cronquist
M. ipomoeavora Spencer (1963f) was described (1981) in treating it as distinct. Cronquist writes:
from Truk (Caroline Is.), with paratypes from 'The differences [between the families] seem as sig-
Guam (Mariana Is.). The holotype is labelled 'Ex nificant, however, as those separating other recog-
Ipomoea' and I have in the past assumed that this nized families in the order.'
indicated that the species had been reared from Only a single species, Melanagromyza cuscutae
Ipomoea. I now feel it is possible that the specimen Hering (cf. Spencer, 1966b), is known on Cuscuta
was caught resting on the plant and this mayor may europaea in Europe and on C. reflexa in northern
not be an indication of its host. A new illustration is India and Pakistan, the larvae feeding either in the
now given of the genitalia of the holotype (Fig. 723) seeds or stem. The genitalia (Fig. 729) are entirely
and also of a paratype from Guam (Fig. 724). typical of the genus and cannot be directly associ-
Although these are not identical, the complex struc- ated with the two species feeding on Convolvulus
ture is generally similar and pending further study I (see above).
accept them as representing a single species. M. ma-
194 Class MAGNOLIOPSIDA (Dicotyledons)
Table 32
Convolvulaceae
Convolvuleae
Convolvulus Melanagromyza albocilia, Mel convolvuli; Liriomyza schmidti, L. strigata;
Chromatomyia horticola
Ipomoeeae
Ipomoea Mel. caerulea, Mel. ipomoeavora; Tropicomyia flacourtiae ? ( Cameroun)
( = Calonyction ) T. theae ( PNG ); Calycomyza brewerae ( as breweri), Cal. ipomaeae,
Cal. ipomoeaphaga, Cal. stegmaieri; L. schmidti, L. trifolii
Cuscutaceae
Cuscuta Mel. cuscutae
Polemoniaceae
Phlox L. sativae, L. trifolii; Chr. horticola
Polemonium Mel. polemonii; L. bryoniae, L. strigata, L. xanthocera ( as crucifericola )
Hydrophyllaceae
Eriodictyon Chr. eriodictyi
Nemophila Chr. mimuli
Phacelia L. strigata; Chr. horticola, Chr. mimuIi; Phytomyza phaceliae
725
726
.. F~~t~"l:.:.{.'.~.··.
:~! -' .
. : .-'
.
728 729
730
'!'~-'---.-:-,~,~~-
<:::J"",o~~;'~"~;"jJi ·.t)ri?illd
735 736
738
737
739 740
--
--......,.~-'-:~.;.....: :..---.::::' ':!
742
741
737, 738) show close relationships with A. myosoti- The male genitalia are essentially similar, differing
dis Kaltenbach (cf. Spencer, 1976a) which feeds on only in detail, but there is also differentiation in the
Cynoglossum and also Borago, Symphytum and larvae, primarily in the form of the posterior spir-
Myosotis (see below), and also some degree of rela- acles (Nowakowski, 1964; Griffiths, 1975). Five
tionship with A. pseudoreptans (Fig. 217) on Urtica species are known in the Palaearctic Region and
in the Urticaceae. The close relationship between three in Canada and Alaska. P. symphyti Hendel
A. abiens and A. myosotidis is interesting in view of (1935) forms a linear-blotch mine on Symphytum
the striking colour difference between the two, spp. The genitalia (Figs. 746, 747) are characteristic
A. abiens being yellow and grey, A. myosotidis of the whole group. Of the other species in Europe,
largely black. P. nowakowskiana Beiger (1975a) is known only in
A. canadensis Spencer (1969b) is believed to Poland and P. rhodopaea Beiger (1975a) only in
feed on Cynoglossum officinale (mines found in Bulgaria, both feeding on Symphytum spp. P. pul-
Ontario) and it has also been recorded in California. monariae Nowakowski (1959) was described pri-
This is a pale species, with distinctive genitalia (Figs. marily from Pulmonaria but additional hosts re-
739, 740). Remarkably, the genitalia are extremely corded were Cynoglossum (Cynoglosseae), both in
close to those of A. pseudorufipes Nowakowski Poland, and also Symphytum in Bavaria.
(1964), a dark species, feeding on Myosotis in
Europe and Trigonotis in Japan (as A. rufipes,
Sasakawa, 1961). This pair of species clearly repre- Lithospermeae
sents dispersal via Beringia. Two further Phytomyza species on Lithospermum
Calycomyza cynoglossi Frick (1956) is the only and Myosotis are known in Germany and Poland
species in this genus found on the Boraginaceae and respectively, P. lithospermi Nowakowski (1959)
was reared from Cynoglossum in Indiana, U.SA. It and P. myosotica Nowakowski (1959), both hither-
belongs to a group which is weakly differentiated to considered as host-specific. However, a popula-
and only identifiable by the genitalia (Figs. 741, tion on Omphalodes (Cynoglosseae) in East Ger-
742). Calycomyza species are found on 12 other many appears to represent P. myosotica. Of the four
families and it is not possible to determine its N earctic species, P. ovalis Griffiths (1975) (Figs.
nearest relative. 748, 749) is known in NW Canada (Alberta, British
P. pulmonariae Nowakowski (1959) in Europe Columbia, Yukon) and feeds on Mertensia panicu-
on Cynoglossum, Pulmonaria and Myosotis and lata and Myosotis alpestris, and also on Hackelia
P.ovalis Griffiths (1975) (Figs. 748, 749) in NW (Cynoglosseae). The other three, P. beringiana and
Canada on Hackelia, Mertensia and Myosotis both P. petiolaris (both Griffiths, 1975) and P. merten-
cross tribal boundaries (Table 34). siae Sehgal (1971b) feed exclusively on Mertensia
paniculata in NW Canada and Alaska. P. petiolaris
Boragineae differs from the other species by the larva feeding
not exclusively as a leaf miner but also in the midrib
A single Melanagromyza, M. symphyti Griffiths (cf. and petiole.
Spencer, 1966b) is known on Symphytum officinale Griffiths (1975) gives useful keys to male gen-
in England, the larvae feeding in the stem and thick italia and leaf mines of all species in this group
leaf stalks. The genitalia (Fig. 743) indicate that this known both in Europe and the Nearctic, conceding
species may well be directly related to M. miranda that it is not possible to identify adults on external
on Lithospermum in North America (see below). characters alone. The speciation in this group can-
A. abiens occurs on Anchusa, Lycopsis, Nonea, not have resulted exclusively from specializing on
Pulmonaria and Symphytum. A. myosotidis, a large- individual genera as postulated by Nowakowski
ly black species, has been recorded in this tribe on (1959), with four species now known to feed on
Anchusa, Borago and Symphytum. A third species, Symphytum in Europe and four on Mertensia in
A. ferruginosa Wulp (cf. Hendel, 1931) occurs on North America. Also there is greater oligophagy
Pulmonaria and Symphytum. This is a large, strik- than was appreciated by Nowakowski, with P. myo-
ing, largely yellow species with genitalia (Figs. 744, sotica on Omphalodes and Myosotis, and P. pul-
745) associating it with A. abiens but it has consi- monariae on Cynoglossum, Pulmonaria, Symphy-
derably diverged, although Symphytum is a host in tum and Myosotis. The slight differentiation be-
common. tween many of these species suggests that speciation
A monophyletic group of largely host-specific, has been recent. I also feel that the diagnoses pre-
small, black Phytomyza species, which has been re- sented by Nowakowski and Griffiths should if pos-
ferred to as the symphyti group, feeds predominantly sible be tested by examining material from further
on this tribe but two species cross tribal boundaries localities and on further host genera.
and colonize genera in two and even three tribes. Melanagromyza miranda Spencer (1969b) was
Subclass ASTERIDAE 199
originally described from British Columbia and the tinct, A. canadensis being pale, yellow and grey, and
host, Lithospermum arvense, was later discovered in A. psudorujipes largely black.
Montana, U.S.A. The genitalia suggest that this spe-
cies may well be directly related to M. symphyti
Echieae
feeding on Symphytum in Europe.
Agromyza lithospermi Spencer (1963b) is known Echium is a favourite host of A. abiens which
from England and Romania on Lithospermum offi- occurs commonly on genera in the five tribes which
cinale and is closely related to A. abiens. A. pseudo- have been colonized.
rujipes Nowakowski (1964) is known on Myosotis
in Europe and Trigonotis in Japan. Surprisingly, the It is interesting that the Phytomyza nepetae group on
genitalia are scarcely differentiated from those of Lamiaceae is far closer to the symphyti group on
A. canadensis on Cynoglossum in Canada (Figs. Boraginaceae than to the obscura group on Lamia-
739, 740), although the two species are entirely dis- ceae (see below).
Table 34
Agromyzidae on Boraginaceae
Classi fication after Mabberley ( 1987 )
Subfamily Cordioideae
Cordia Pseudoliriomyza cordiae
Subfamily Heliotropioideae
Heliotropium Melanagromyza heliotropii; Liriomyza blechi ( ? ), L. strigata;
Chromatomyia horticola
Subfamily Boraginoideae
Cynoglosseae
Asperugo Agromyza abiens; Chr. horticola
Caccinia Chr. horticola
Cynoglossum A. abiens, (?) A. canadensis, A. myosotidis; Calycomyza cynoglossi;
Chr. horticola; Phytomyza pulmonariae
Hackelia P.ovalis
Omphalodes A. abiens; Chr. horticola; P. myosotica
Trichodesma Chr. horticola
Boragineae
Anchusa A. abiens, A. myosotidis; Chr. horticola
( = Pentaglottis )
Borago A. myosotidis; Chr. horticola
Lycopsis A. abiens; Chr. horticola
Nonea A. abiens; Chr. horticola
Pulmonaria A. abiens, A. ferruginosa; Chr. horticola; P. pulmonariae
Symphytum Mel. symphyti; A. abiens, A. ferruginosa, A. myosotidis; Chr. horticola;
P. nowakowskiana, P. pulmonariae, P. rhodopaea, P. symphyti
Eritricheae
Lappula A. abiens; Chr. horticola
Lithospermeae
Cerinthe A. abiens; Chr. horticola
Lithospermum Mel. miranda; A. lithospermi; Chr. horticola; P. lithospermi
Mertensia A. canadeHsis; P. beringiana, P. mertensiae, P. ovalis, P. petiolaris
Myosotis A. abiens, A. myosotidis, A. pseudorufipes; Chr. horticola; P. myosotica,
P. oval is, P. pulmonariae
Podonosma A. abiens, A. pseudorufipes; Chr. horticola
Trigonotis A. pseudorufipes
Echieae
Echium A. abiens; Chr. horticola; P. symphyti
200 Class MAGNOLIOPSIDA (Dicotyledons)
'.',,~~
i~~~'
744 745
746 747
... _........",.
:'
748(\, 749
-'-l ... - ~.-
Figs. 744, 745. Agromyzaferruginosa: 744, aedeagus, side view; 745, distiphallus, ventral view.
Figs. 746, 747. Phytomyza symphyti: 746, aedeagus, side view; 747, same, ventral view (Nowakowski).
Figs. 748, 749. Phytomyza ovalis: 748, aedeagus, side view; 749, same, ventral view (Griffiths).
Table 35
Agromyzidae on Verbenaceae
Classification following Briquet in Willis (1957)
Verbeneae
Duranta Phytobia sp. (larval feeding); Tropicomyia flacourtiae;
Calycomyza durantae, Cal. lantanae
Lantana Ophiomyia camarae, O. lantanae, O. legitima; Cal. lantanae
Lippia Melanagromyza riparella; O. Jippiae; Cal. lantanae, Cal. verbenivora ? ;
Liriomyza nr. blechi (? sp. n.)
Stachytarpheta Liriomyza spp. (I, Argentina),(2, Costa Rica)
Verbena Mel. verbenae; O. verbenivora; L. bryoniae, L. strigata; Cal. lantanae,
Cal. verbenae, Cal. verbenivora; L. trifolii; Chromatomyia horticola
Viticeae
Holmskioldia L. ? sativae, L. trifolii
Vitex Phytobia sp. (larval feeding)
Nyctantheae
Nyctanthes Tropicomyia sp.
Avicennieae
Avicennia Mel. avicenniae
val feeding of Phytobia on Duranta plumieri Jacq. in plant is placed in Phyla by some authors but this is
'tropical America.' generally accepted as a small segregate within
Lantana has been colonized by four species. Lippia). The male genitalia are typical of the genus
Ophiomyia lantanae (Froggatt) (cf. Spencer and but give no indication of the correct relationship of
Stegmaier, 1973) feeds in the fleshy receptacle of the species. Ophiomyia lippiae Spencer (cf. Spencer
several Lantana spp. It occurs commonly in the and Stegmaier, 1973) is known only in Florida and
Caribbean area and much of tropical South America the Bahamas. Although recorded as a seed-feeder, I
and has been widely introduced for biocontrol of suggest the larva forms external stem mines, pos-
Lantana, but with little success, in Africa, tropical sibly pupating in the upper stem. The posterior
Asia and northern Australia. 0. legitima Spencer (in spiracles (Fig. 755) are typical of species feeding in
Spencer and Stegmaier, 1973) which was described this way. In addition, a species tentatively identified
from Dominica has recently been discovered in as Liriomyza blechi (cf. Blechum, Acanthaceae) has
Guadeloupe feeding in the fruits of Lantana camara. been reared from Lippia in Florida. Valladares
Although the adults of these two species, with (unpublished report) has found mines of a Lirio-
apparently identical biology and the same host, are myza sp. on Stachytarpheta sp. in Argentina and I
not distinguishable externally, the genitalia confirm found mines on this host west of San Jose, Costa
that they are distinct. 0. camarae Spencer (1963c) Rica, which are narrow and linear, not associated
is known from Trinidad to Argentina, forming with the midrib and these probably also represent a
distinctive leaf mines on Lantana spp., the larva Liriomyza sp.n.
feeding along main veins, with short offshoots into Four host-specific species are known on Verbena.
the leaf blade (Fig. 752). The asymmetrical genitalia Melanagromyza verbenae Spencer (1982) is an
are characteristic of the genus (Figs. 753, 754). Ca- internal stem-borer in V. litoralis. I found many
lycomyza lantanae (Frick) (cf. Spencer and Steg- specimens on this host in central Chile. A related or
maier, 1973) forms large blotch mines primarily on possibly identical species is present in Florida on
Lantana. I have also reared it from Verbena litoralis V. scabra. The genitalia show that M. verbenae is
in Venezuela (Caracas) and it is known on Lippia in not directly related to M. riparella on Lippia and
Florida and Argentina. the two species appear to represent separate coloni-
Two species appear to be restricted to Lippia. zations of the family. Ophiomyia verbenivora Valla-
Melanagromyza riparella (Hendel) (cf. Spencer and dares (1986a) was described from Verbena bona-
Steyskal, 1986) is known only in the Eastern U.S.A., riensis and V. litoralis at Cordoba, Argentina.
feeding internally in stems of Lippia nodiflora (this Th~ species is in all ways typical of the genus, but
202 Class MAGNOLIOPSIDA (Dicotyledons)
751
750
753 755
756 c, 757
Subclass ASTERIDAE 203
the leaf mine is entirely different from that of L. trifolii and possibly L. sativae on Holmskioldia in
0. camarae on Lantana (Fig. 752), here forming a Jamaica.
normal linear mine at any point on the leaf blade, The only probable direct relationship of species
the puparium remaining in the leaf. on Verbenaceae with any on other families is detect-
Two Calycomyza species which are not readily able in Calycomyza verbenivora and C. hyptidis on
distinguishable externally feed on Verbena, both Lamiaceae. The genitalia of the two species are
forming blotch mines. C. verbenae (Hering) (cf. remarkably similar (cf. Figs. 757 and 784) and the
Spencer, 1963a) is present in New Mexico and greatest difference is apparent in the larval feeding,
Florida (Fig. 756) and C. verbenivora Spencer with a linear-blotch mine formed by C. verbenivora
(1963c) is known in Venezuela and Argentina (Fig. and a star-shaped blotch by C. hyptidis (Fig. 785
757). If these two species represent a single original below).
colonization of Verbena, they have now diverged
significantly in their genitalia. However, it seems
Lamiaceae (200 genera, 3200 species)
more probable that they are not directly related and
that C. verbenivora is related to C. hyptidis on The Lamiaceae were until recently divided into
Hyptis (Lamiaceae, see below). Valladares (1982, seven subfamilies but the most recent revision by
unpublished thesis) tentatively records Lippia tur- Cantina and Sanders (1986) based on pollen type,
binata as host of C. verbenivora in Argentina. seed type and chemistry proposed merely two major
divisions: the Lamioideae and Nepetoideae.
The 31 species of Agromyzidae restricted to the
Viticeae
family fit well into these two subfamilies, the domi-
I found linear mines (Spencer, 1963c: fig. 103n) of nant genus Phytomyza with 11 species (Plus one
an unidentified species on Holmskioldia sanguinea Chromatomyia) being predominantly on the Nepe-
in a Botanical Garden in Kingston, Jamaica. I now toideae, while the next largest group, Amauromyza,
feel that this probably represents Liriomyza sativae, feeds mainly on the Lamioideae. The few transfers
although this would be the only record on the Ver- between the subfamilies are detailed below. Of the
benaceae of this highly polyphagous species. strictly host-generic species, two occur on the
Larval feeding of Phytobia has been recorded in Lamioideae, eight on the Nepetoideae. All species
Java on Vitex incisa Lam. (Suss, lac. cit.). known on the family are shown in Table 37.
Cronquist (1981) suggests that the Lamiaceae are
Avicennieae
most closely related to the Verbenaceae, but the
host selection of the Agromyzidae could suggest an
Melanagromyza avicenniae Spencer (1977c) was equal degree of relationship with the Boraginaceae.
found in large numbers on the grey mangrove, The Phytomyza nepetae group on Lamiaceae (two
A vicennia marina, north of Sydney, Australia. Al- species) is clearly closely related to, and scarcely
though no specimens were reared, no other possible distinguishable from, the symphyti group on Boragi-
host was present in the vicinity of the mangroves naceae (three species). A further group of Phyto-
and this large species, with wing length up to 4 mm myza species which can be referred to as the petoei
is certainly an internal borer in the stems. As might group (five species) shows unmistakable relation-
be expected on this unusual host, the species is ship with some species on the Ranunculaceae (see
clearly isolated and the genitalia (Figs. 758, 759) above).
give no indication of its direct relationships. This large family has been colonized by ten
The small number of agromyzid genera and genera of Agromyzidae, with a total of 39 species, as
species feeding on the Verbenaceae shows that this indicated in Table 36.
family is not generally acceptable to the Agromy-
zidae. This is further indicated by the few records of
Subfamily Lamioideae
polyphagous species which have been found on the
family - Liriomyza bryoniae, L. strigata and Chro- Fourteen genera in this subfamily have been
matomyia horticola on Verbena in Europe, and colonized by the Agromyzidae, with seven species
Figs. 750, 751. Calycomyza durantae: 750, aedeagus, side view; 751, distiphallus, ventral view.
Figs. 752-754. Ophiomyia camarae: 752, leaf mine on Lantana camara; 753, adeagus, side view; 754, distiphallus, ventral view.
Fig. 755. Ophiomyia lippiae: posterior spiracles of puparium.
Fig. 756. Calycomyza verbenae: aedeagus.
Fig. 757. Calycomyza verbenivora: aedeagus.
Figs. 758, 759. Melanagromyza avicenniae: 758, aedeagus, side view; 759, same, ventral view.
204 Class MAGNOLIOPSIDA (Dicotyledons)
Table 37
Agromyzidae on Lamiaeeae
Classification after Cantino & Sanders (1986)
Subfamily Lamioideae
Ajuga ~~auromyza labiatarum; Liriomyza bryoniae, L. strigata, L. trifolii;
't1fromatomyia hortieola; Phytomyza abdita
BaIlota
...Am. labiatarum, Am. lamii, Am. morionella; L. strigata; Chr. hortieola;
P. saIviae (as P. ballotae), Phytomyza sp. (Istanbul)
Galeopsis Am. labiatarum, Am. lamii; Ophiomyia labiatarum; L. bryoniae,
L. eupatorii, L. strigata, L. xanthoeera (as L. erueiferieola); Chr. hortieola
Lamium Am. labiatarum, Am. lamii; O. labiatarum; Agromyza flavipennis;
L. bryoniae, L. strigata; Chr. hortieola
Leonotis Melanagromyza leonotidis; Agromyza sp. (mines only, Kenya)
Leonurus Am. leon uri; L. bryoniae, L. strigata; Chr. hortieola
Marrubium Am. labiatarum, Am. lamii, Am. morionella
Melittis Am. labiatarum
Molueella L. strigata, L.trifolii; Chr. hortieola
Phlomis Am. balcaniea, Am. madrilena
Prostanthera Phytobia sp. (larval feeding)
Scutellaria Am. labiatarum, Am. lamii
Staehys O. labiatarum; Am. labiatarum, Am. lamii;
(= Betoniea) L. staehyos, L. strigata; Napomyza sp. n.?; Chr. hortieola
Teuerium Am. labiatarum; L. strigata; Chr. hortieola
Subfamily Nepetoideae
Bystropogon Chr. hortieola
Clinopodium P. petoei
Coleus O. eolei
Draeoeephalum L. strigata; P. nepetae
Gleehoma Am. labiatarum, Am. lamii; L. strigata; Napomyza nigrieeps; P. gleehomae
Hyptis Calyeomyza hyptidis, Cal. hyptisieola, Cal. menthae
Lallemantia Am. labiatarum, Am. lamii; L. bryoniae, L. xanthoeera, L. strigata;
Chr. hortieola
Lavandula Chr. hortieola
Lepeehinia Chr. platensis
Lyeopus Am. labiatarum; Chr. hortieola, Chr. platensis; P. Iyeopi
Majorana Chr. hortieola
Melissa P. petoei
Mentha Am. labiatarum, Am. lamii; Cal. menthae; Chr. hortieola, Chr. platensis;
P. petoei, P. tetrastieha
Monarda Cal. menthae
Monardella Cal. menthae
Nepeta O. labiatarum; L. strigata; Chr. hortieola; P. nepetae
Ocimum O. oeimi, O. oeimivora; Cal. menthae; L. strigata, L. trifolii; Chr. platensis
Origanum Chr. hortieola; P. origani
Pleetranthus Am. pleetranthi, Amauromyza sp. (mines only)
Prunella O. labiatarum; Am. labiatarum; Chr. hortieola; ? P. abdita
Salvia L. strigata, L. trifolii; Chr. hortieola, Chr. platensis; P. salviae, P. seotina
Satureja O. labiatarum, Am. labiatarum; Chr. hortieola; P. obseura
Thymus P. thymi
? genus O. aberrans
206 Class MAGNOLIOPSIDA (Dicotyledons)
.(~~~
,......
760
f······.7~2
~.:/ )j
/,
... j
765
767
768) .' ,
769
Subclass ASTERIDAE 207
This is a primitive biological character and reflects also an accompanying differentiation in larval char-
long association with and isolation on Stachys. acters, particularly in the number of pores on the
Allen (1956) cites Stachys germanica as one of posterior spiracles. It seems clear that speciation in
many hosts of Napomyza latera lis (Fallen, 1823, cf. this group has thus been relatively recent.
Spencer, 1976a). However, this is a misidentifica- Phytomyza obscura Hendel (cf. Nowakowski,
tion. N. lateralis feeds exclusively on Asteraceae, 1959) is host-specific on Satureja. In the male gen-
but the record is of interest indicating that an italia the distal tubules of the aedeagus are greatly
undescribed species probably feeds on Stachys. enlarged (Figs. 770, 771). P. origani Hering (cf.
Phytomyza abdita Hering (cf. Spencer, 1977b) is Nowakowski, 1959) is found only on Origanum
an isolated species known only from Switzerland, vulgare, and P. tetrasticha Hendel (cf. Nowakowski,
where it was reared from Ajuga, according to the 1959) on Mentha species. In both species the gen-
original description. However, Hering (1927b: 51) italia are of the same form as in P. obscura but differ
states that there is doubt about the host, which may considerably in detail. It is interesting that the
have been Prunella. The genitalia (Fig. 768) give no species on Thymus, P. thymi, belongs to a different
indication of its affiliations. Hering (1924a) de- group.
scribed P. ballotae from Ballota nigra, from Rostov- P. nepetae Hendel (cf. Nowakowski, 1959),
on-Don, southern U.S.S.R. I have now established whose primary host is Nepeta cataria but which has
that this species represents a junior synonym of also been found on Dracocephalum, has significant-
P. salviae (see Appendix). It is possible, though ly different genitalia (Figs. 772, 773) and the sperm
unlikely, that this species feeds on both Salvia pump is smaller than in the other groups. Nepeta has
(Nepetoideae) and Ballota (Lamioideae). I consider been introduced to North America and somehow
that it is more probable that there was confusion in P. nepetae appears to have been similarly intro-
the identity of the host and therefore Ballota as a duced and it is known in Ontario and Quebec and
host of P. salviae requires confirmation. also in Washington state, U.S.A. P. lycopi Nowa-
Larval feeding of a Phytobia sp. in the tree Pros- kowski (1959) is known only on Lycopus europaeus
tanthera lasianthus Lam. has been recorded in and has generally similar genitalia but very different
Australia by Suss and Muller-Stoll (1980) and Suss larval spiracles. The genitalia of these two species
(pers. comm.). Smaller plants in this genus are show clear association with the P. symphyti group
popularly referred to as 'mint-bush.' on Boraginaceae.
The Phytomyza petoei group consists of five
species in Europe - P. petoei (d. Spencer, 1976a),
Subfamily Nepetoideae
P. glechomae Kaltenbach (cf. Spencer, 1976a),
A single Napomyza, N. nigriceps Wulp (cf. Spencer, P. salviae Hering (1924a), P. scotina Hendel (1920)
1976a), has recently been discovered to feed in the and P. thymi Hering (1928). P. petoei feeds pri-
stems of Glechoma hederacea (von Tschirnhaus, marily on Mentha and also on Melissa. The genitalia
pers. comm.) in western Europe. The genitalia (Fig. (Figs. 774, 775), with the exceptionally large sperm
769) indicate relationship with the feeders in the pump, immediately differentiate it and others in the
stem and receptacle of Ranunculus, N. evanescens group from the nepetae group. The genitalia are in
(Fig. 128) and N. nigritula (Fig. 129). fact remarkably similar to P. calthophila (Figs. 43,
Nowakowski (1959) discussed in detail five of the 44) (Caltha, Ranunculaceae). This cannot represent
small black Phytomyza species feeding on Lamia- convergence and, I suggest, indicates a switch from
ceae in Europe, placing three in the obscura group Ranunculaceae to Lamiaceae. Subsequent radiation
and two in the nepetae group. All are restricted to has accounted for the speciation on Glechoma, Sal-
the Nepetoideae. A further group of five species via and Thymus of the four other species in the
now referred to as the Phytomyza petoei group is group. P. glechomae and P. salviae have retained
also largely restricted to this subfamily. With few the outer cross-vein (tp) but the genitalia of P. sal-
exceptions species in this group are reliably iden- viae (Figs. 776, 777) closely resemble those of
tifiable only by the male genitalia and even here P. petoei; in P. glechomae there is an additional pair
differences are frequently slight. However, there is of short sclerites behind the distiphallus. A second
770 771
773
774 775
.
~
>~c@_~+- ~~.
776 777
778 779
Subclass ASTERIDAE 209
species on Salvia, P. scotina, lacks this cross-vein been collected in Sri Lanka and at Bangalore,
but is otherwise similar to P. salviae both on ex- southern India. It seems possible that these species,
ternal characters and in the genitalia (Fig. 778). On as with the leaf miners on Solanum in South Africa,
Thymus, P. thymi was described from Switzerland Madagascar, Papua New Guinea and Australia
and is known in Germany; it is probably widespread (Clyde Mountain, south of Sydney) (cf. Solanaceae)
with its host. It is the smallest species in the group reflect a Gondwanan distribution.
but the genitalia (Fig. 779) indicate its affiliations. Three Calycomyza species are known on Hyptis,
In South America Chromatomyia platensis of which two appear to be restricted to this genus.
Brethes (cf. Spencer, 1982) (Fig. 780) appears to be Cal. hyptidis (cf. Spencer and Stegmaier, 1973) was
widespread mainly in more temperate areas associ- described from Florida on Hyptis pectinata and is
ated with Lamiaceae. It was described in Argentina present in South America in Venezuela and Brazil. I
on Salvia splendens at Buenos Aires and has re- also recently found it on Guadeloupe on H. atroru-
cently been recorded at Cordoba on this and bens. The larva forms a distinctive star-shaped mine
other Salvia species and also on Lepechinia, Mentha (Fig. 785), returning to the centre of the mine when
and Ocimum (Valladares, unpublished thesis, not feeding and there depositing its frass. The
1982). Steyskal (1976) redescribed this species as genitalia (Fig. 784), with the elongate tubules of the
Phytomyza winderi on Salvia farinacea (introduced distiphallus, indicate close relationship with Cal.
from southern U.S.A.) at Curitiba, Brazil. The syno- verbenivora (Fig. 757). The second species, Cal.
nomy of P. winderi with Chr. platensis is established hyptisicola Spencer (1973c), is only known from
in the Appendix. At Valdivia in south-central Chile I Venezuela on Hyptis sp., the larva forming a regular
reared this species from Lycopus europaeus, intro- blotch mine. The genitalia (Fig. 786) indicate that
duced from Europe, growing on the university the two species are not directly related. Cal. hyptisi-
campus. Adults and mines were also present on cola appears to be most close to Cal. lantanae
Mentha arvensis at Concep<;ion, and I have also seen (Verbenaceae). Cal. menthae Spencer (1969b) was
flies from four further localities in Chile north to La described from Mentha in Ontario, Canada and has
Serena. I think it can be accepted that Salvia is the since been recorded on Hyptis, Monarda and Oci-
original host and it is of interest that introduced mum in Florida and on Monardella in California.
Lycopus and Mentha have also been colonized. The genitalia (Figs. 787, 788) suggest possible
Six species in Africa, North and South America in relationshp with Cal. hyptisicola.
the genera Ophiomyia and Calycomyza are re- In Japan Amauromyza plectranthi Sasakawa
stricted to the Nepetoideae, occurring on the genera (1961) is known as a leaf miner on Plectra nth us sp.
Coleus, Hyptis, Mentha and Ocimum. In South (no details of mine recorded). No males are known
Africa O. colei Spencer (1965b) forms long, wind- and the affiliations of the species cannot be further
ing mines on Coleus, probably C. hybridus hort. clarified. I found empty mines on Pl. laxiflorus at
(Spencer, 1960a: fig. 21a) and the genitalia (Fig. two localities in Natal and mines have also been
781) are asymmetric and typical of the genus. found on Pl. flaccidus in Kenya. These certainly
0. ocimi Spencer (1965b) (Fig. 782) is known as a represent an Amauromyza sp. and it is highly prob-
leaf miner on O. gratissimum in Guinea, West able that the species is distinct from Am. plectranthi
Africa (leaf mines not preserved). 0. ocimivora in Japan.
Spencer (1985a) (Fig. 783) is probably an external One exceptional case of host transference has
stem miner on 0. lamiifolium in Kenya. The gen- been frequently noted in Europe, with Liriomyza
italia of the three species do not indicate that the eupatorii (normal host: Eupatorium, Asteraceae)
species are closely related. If they are monophyletic, forming its characteristic leaf mine on Galeopsis
there has been remarkable differentiation in the tetrahit. It can only be speculated that there is some
genitalia. attractive substance in Galeopsis which makes it
A unique species in Ophiomyia, O. aberrans acceptable to this feeder on Asteraceae.
(Spencer, 1960a), has the halteres snow white. This The Lamiaceae are widely acceptable to Euro-
species was described from Zaire and subsequently pean polyphagous species. The two subfamilies
reared in South Africa (Pretoria) from flower-heads appear to be almost equally acceptable and one or
of an unidentified Labiate. A related species, host more of the four Liriomyza species, L. bryoniae,
unknown, 0. subaberrans (Spencer, 1986b), has L. xanthocera, L. strigata or L. trifolii and also
Figs. 770, 771. Phytomyza obseura: 770, aedeagus, side view; 771, distiphallus, ventral view (NOWakowski).
Figs. 772, 773. Phytomyza nepetae: 772, aedeagus, side view; 773, distiphallus, ventral view.
Figs. 774, 775. Phytomyza petoei: 774, aedeagus, side view; 775, distiphallus, ventral view.
Figs. 776, 777. Phytomyza sa/viae: 776, aedeagus, side view; 777, distiphallus, ventral view.
Fig. 778. Phytomyza seotina: distiphallus, ventral view.
Fig. 779. Phytomyza thymi: distiphallus, ventral view.
210 Class MAGNOLIOPSIDA (Dicotyledons)
780
.. ; ,', ,.~ .
./
".
782 783
784
...... ,.
,:: .. j'
\ 785
.~H;;'~;~
787 788
Subclass ASTERIDAE 211
--"\
\
1)
789 792
790 791
Agromyzidae. The Scrophulariaceae are the largest no related groups of Agromyzidae common to the
family, with by far the largest number of Agromy- two orders.
zidae. Cronquist considers that most of the other Species on the eight families Bignoniaceae, Bud-
families in the order 'may logically be considered to dlejaceae, Gesneriaceae, Globulariaceae, Myopora-
be specialized derivatives of the Scrophulariaceae.' ceae, Oleaceae, Orobanchaceae and Pedalhlceae are
Although the Scrophulariales are considered to be listed in Table 39, those on the Acanthaceae in
closely related to the Solanales, there appear to be Table 38 and on the Scrophulariaceae in Table 41.
Table 38
Agromyzidae on Acanthaceae
Thunbergioideae
Thunbergia Melanagromyza thunbergiae; Tropicomyia thunbergivora
Acanthoideae
Contortae
Blechum Mel. ruelliae; Liriomyza blechi
Hygrophila Mel. asteracanthae
( = Asteracantha )
Ruellia Mel. ruelliae
Strobilanthes Mel. ruelliae
Imbricatae
Asystasia Pseudonapomyza asystasiae
Blepharis Ophiomyia blepharidis
Chaetoth y lax Two unidentified spp. ( Argentina)
Dicliptera Ps. matopi; unidentified sp. ( Argentina)
Hypoestes Ps. hypoestis, Ps. hypoestivora, Ps. urundensis
Justicia Mel. diantherae; Ps. justiciae sp. n., Ps. urundensis;
( = Dianthera ) unidentified sp. ( Argentina)
Peristrophe Pseudonapomyza sp. ( South Africa)
'? genus Ps. acanthacearum
Table 39
Bignoniaceae
Bignonieae
Clytostoma Liriomyza schmidti
Tecomeae
Catalpa Phytobia sp. ( larval feeding ); Amauromyza pleural is
Dolichandrone Phytobia sp. ( larval feeding)
Jacaranda Phytobia sp. ( larval feeding ); Phytoliriomyza jacarandae
Markhamia Phytobia sp. ( larval feeding)
Newbouldia Phytobia sp. ( larval feeding)
Stereospermum Phytobia sp. ( larval feeding)
Crescentieae
Kigelia Phytobia sp. ( larval feeding ); Tropicomyia polyphyta
214 Class MAGNOLIOPSIDA (Dicotyledons)
Buddlejaceae
Buddleja Am. verbasci; Liriomyza sp. n. ( Argentina)
Gesneriaceae
Cyrtandra Tropicomyia sp. n. ( leaf mines only)
Globulariaceae
Globularia L. globulariae; Phytomyza globulariae
Myoporaceae
Myoporum L. citreifemorata, L. cocculi
Oleaceae
Oleoideae
Fraxineae
Fraxinus Phytobia sp. ( larval feeding ); Paraphytomyza fraxini, Para. fraxinivora,
Para. heringi
Oleeae
Chionanthus Phytobia sp. ( larval feeding)
Ligustrum Phytobia sp. ( larval feeding)
Olea Phytobia sp. ( larval feeding)
Phillyrea P. phillyreae
Syringa Phytobia sp. ( larval feeding)
lasminoideae
lasmineae
lasminum Tropicomyia sp.; Para. jasmini
Forsythieae
Forsythia Phytobia sp. ( larval feeding)
Orobanchaceae
Orobanche P. orobanchia
Pedaliaceae
Sesamum Melanagromyza azawii
bergia natalensis in South Africa and the genitalia species in this group are restricted in their host to a
are characteristic of this genus. single genus and only Ps. urundensis has been re-
corded on two genera. In six species the outer cross-
vein is present, forming a small discal cell but this
Subfamily Acanthoideae
vein is lacking in Ps. hypoestivora in Madagascar. In
Contortae. Two Melanagromyza and one Liriomyza contrast to the related group feeding on Poaceae in
species are known on this tribe. which the third antennal segment is distinctly angu-
Blechum, Ruellia, Strobilanthes. Melanagromyza late (Ps. asiatica, Fig. 1416), here the third antennal
ruelliae Spencer (in Spencer and Stegmaier, 1973) segment is round. An important larval character
was originally discovered in Florida on Blechum, linking the two groups is the presence of con-
Ruellia and Strobilanthes, feeding in the seeds. It spicuous spinules around the segment boundaries
was subsequently found in the Bahamas. The gen- (cf. Ps. atra, Poaceae, Fig. 1420).
italia (Fig. 796) show that this species is not related The general similarity in male genitalia and also in
to M. diantherae (Fig. 800, see below). larval characters, with the primitive arrangement of
Liriomyza blechi Spencer (1973, loco cit.) was three pores on the posterior spiracles, suggests that
described from Florida, forming leaf mines on the species on Acanthaceae are monophyletic, re-
B. pyramidatum, invariably associated with the mid- presenting a single original colonization of the
rib (Fig. 797). It belongs to what is probably a group family. Ps. asystasiae Spencer (1965b) was de-
of closely related species which have been reared scribed from South Africa and leaf mines referable
from Heliotropium (Boraginaceae), Lippia (Ver- to the species were also found by the author at
benaceae), Plantago (Plantaginaceae) and Spigelia Victoria Falls, Zimbabwe and have recently been
(Loganiaceae). The male genitalia (Fig. 798) and found near Mombasa, Kenya. The normal host
particularly the anterior spiracles (Fig. 799) are appears to be Asystasia gangetica and the distinctive
distinctive and indicate close relationship with genitalia are shown in Fig. 804. Ps. matopi Spencer
L. marginalis which feeds exclusively on a number (1965b) (Fig. 805) has been found only once in the
of genera of Poaceae (see below). When this switch Matopos Hills, south of Bulawayo on Dicliptera
occurred and how it was motivated can only be a monroi. Three species are known on Hypoestes. Ps.
matter of speculation. hypoestis Hering (1957a) was described from Natal
Hygrophila (= Asteracantha). M. asteracanthae from leaf mines on Hypoestes aristata. This species
Spencer (1966a) was described from Bangalore, (and also Ps. justiciae which I recently discovered in
southern India, reported as a leaf miner on Astera- Kenya) differs from others in the group in having the
cantha sp. The genitalia and larval spiracles confirm halteres dark brown (not white or yellow) but the
that this is a true Melanagromyza and I suspect it genitalia (Fig. 806) do not suggest any significant
feeds in the midrib and is not a true leaf miner. difference from related species. Ps. urundensis
Imbricatae. One Melanagromyza, one Ophiomyia Spencer (1959) was described from a single female
and seven Pseudonapomyza species (Plus one un- from Zaire (Belgian Congo) and was subsequently
described) are known on genera in this tribe. found in South Africa and ultimately reared from
M. diantherae (Malloch) (cf. Spencer and Steys- Hypoestes triflora at Addis Ababa, Ethiopia. The
kal, 1986) is known in the eastern United States, genitalia of a male from Natal are shown in Fig. 807,
feeding internally in the stems of lusticia americana. and the pupal spiracles in Fig. 808. With the strong
The genitalia (Fig. 800) indicate that this species is pigmentation, the aedeagus of this species can be
not closely related to the other Melanagromyza sp. more closely associated with grass-feeders (cf. Figs.
known on Acanthaceae in North America, M. ruel- 1419, 1427) than with the others known on Acan-
liae (see above). thaceae. Ps. hypoestivora Seguy (1951) was de-
A species feeding in the seeds of Blepharis sp. in scribed from Madagascar from Hypoestes betsilien-
South Africa was described as Melanagromyza sis. The genitalia of this species closely resemble
blepharidis Spencer (1960a). The genitalia of the those of Ps. hypoestis but the two species are entire-
unique holotype have now been examined (Figs. ly distinct, with the halteres brown in hypoestis, white
801, 802) and show that the species correctly in hypoestivora. Also Ps. hypoestivora lacks the
belongs in Ophiomyia (see Appendix below). This is outer cross-vein and the wing venation is thus as is
supported by the form of the posterior larval spir- usual in the group on Poaceae. Ps. acanthacearum
acles, with just three elongate pores (Fig. 803) on Spencer (1959) was reared on Mt. Cameroun from
each process. a 'white-flowered Acanthaceae.' The genitalia are
The eight Pseudonapomyza species have a range distinctive and confirm the close association with
extending from Cameroun in the west to Ethiopia, others in the group. During a recent visit to Kenya
Kenya, South Africa and Madagascar. On the (February, 1988) I found a single immature mine on
limited information available, it appears that most lusticia diclipteroides near Nairobi and caught a
216 Class MAGNOLIOPSIDA (Dicotyledons)
796
795
.-.' . :', \~~ .
"\
. .. ' . t:,'
.:.:;",' ...
.:, ... .~, .. :.. :
,~d') "--_/
799
797 798
801
802
800
803 804
Fig. 795. Melanagromyza thunbergiae: posterior spiracles of puparium.
Fig. 796. Melanagromyza ruelliae: aedeagus.
Figs. 797-799. Liriomyza blechi: 797, leaf mine on Blechum pyramidatum; 798, aedeagus; 799, anterior spiracles of puparium.
Fig. 800. Melanagromyza diantherae: aedeagus.
Figs.801-803. Ophiomyia blepharidis: 801, aedeagus, side view; 802, distiphallus, antero-ventral view; 803, posterior spiracles of
puparium.
Fig. 804. Pseudonapomyza asystasiae: aedeagus.
Subclass ASTERIDAE 217
806
807 808
809 810
Fig. 805. Pseudonapomyza matopi: aedeagus.
Fig. 806. Pseudonapomyza hypoestis: aedeagus. . . .
Figs. 807, 808. Pseudonapomyza urundensis: 807, aedeagu~; 80~, postenor spiracles of ~upanum.
Figs. 809, 810. Pseudonapomyza justiciae: 809, aedeagus, side view; 810, same, ventral view.
single male sweeping the plants. This is described i.n Bignoniaceae (100 genera, 800 species)
the Appendix below as Ps. justiciae. The male gem-
This mainly tropical family is best represented in
talia (Figs. 809, 810) suggest that this species is most
closely related to Ps. hypoestis and this is supported tropical America but the two host-specific Agro~y
zidae are known in Canada, U.S.A. and Argentma.
by the halteres being dark in both species. . .
The family is relatively old, with fossil pollen known
I found empty mines on Peristrophe natalensls m
from middle Eocene and 'well preserved flowers
Natal but identification of the Pseudonapomyza
and fruits attributed to Catalpa have been found in
species concerned is not possible. .
the Eocene London Clay' (Cronquist, 1981).
In Argentina (Valladares, unpubbshed report)
The family was until recently divided into four
records two different types of mine, one linear, one
tribes (Willis, 1957, after Schumann) but has more
a blotch, on Chaetothylax sp. at Tucuman (leg.
recently been split into seven tribes by Gentry (cf.
Wygodzinsky). One probably represents a Lirio"!y-
Cronquist, 1981). The four species of Agromyzidae
za and the other a Calycomyza, although no specIes
known on the family are found on three tribes. In
in these genera have hitherto been recorded on
addition, larval feeding of Phytobia species has been
Acanthaceae. Further unidentified mines have been
detected on seven genera by Suss and Muller-Stoll
found at Tucuman on Dicliptera tweediana and on
(1980) and Suss (pers. comm.), as shown in Table
lusticia sp.
39.
The close relationship of the Pseudonapomyza
species on this family with t~ose on Poac~ae,.which I
accept as representing a smgle genus, mdIcates a Bignonieae
remarkable switch of an ancestral species, probably
The polyphagous Liriomyza schmidti has been
from Acanthaceae to Poaceae and at this stage I can
found on Clytostoma callistegioides in Argentina
offer no explanation for this other than a fortuito.us
(Valladares).
accident, resulting in considerable subsequent radIa-
tion on Poaceae.
218 Class MAGNOLIOPSIDA (Dicotyledons)
Buddlejaceae (10 genera, 150 species) Myoporaceae (3/4 genera, 125 species)
There still appears to be no agreement among This small family is best represented in Australia but
botanists on the taxonomic position of this family. also extends to Pacific islands and eastern Asia. The
Willis (1957) and Mabberley (1987) treat it as two agromyzids known on the family have been
synonymous with Loganiaceae (order Gentianales). found in New Zealand and Hawaii.
However, Cronquist (1981) and Willis (1985) Myoporum. The only host-specific species on this
consider it to represent a separate family close to the genus is Liriomyza citreifemorata Watt (d. Spencer,
Scrophulariaceae. This is supported by the host 1976b), the larva forming a narrow linear mine on
selection of the agromyzid Amauromyza verbasci M. faeta in New Zealand. The genitalia (Figs. 816,
which feeds normally on Verbascum (see Scrophu- 817) suggest that this genus has been colonized by
lariaceae below) but not infrequently also on Budd- an ancestral composite-feeding species (d. L. cras-
leja. This species is not uncommon in Europe and I pediae on Craspedia: Fig. 1031), rather than any
have also seen empty leaf mines on Buddleja panicu- species on the more closely related host, Hebe (d.
lata from Simla in the Himalayas, N.W. India (leg. Figs. 840, 842). In Hawaii L. cocculi which shows
Sehgal) which almost certainly represent the same limited polyphagy, is known on Myoporum sp., in
species. addition to genera in eight further unrelated fami-
Valladares (unpublished report) records an un- lies. It is not related to L. citreifemorata.
described Liriomyza sp. from Buddleja sp. from
Tucuman, N.W. Argentina (leg. Wygodzinsky).
Oleaceae (30 genera, 600 species)
Gesneriaceae (120 genera, 2500 species) This family has in the past been associated with the
Gentianales but here, following Cronquist (1981)
It is generally agreed that this family is closely re- and Johnson (1957), I include it in the Scrophulari-
Subclass ASTERIDAE 219
//~
•. . '-.....c ..- ·
813
':-,
817
819
821
Figs. 811, 812. Amauromyza pleuralis: 811, aedeagus, side view; 812, distiphallus, ventral view.
Figs. 813, 814. Phytoliriomyza jacarandae: 813, leaf mines on Jacaranda mimosifolia; 814, epandrium.
Fig. 815. Tropicomyia sp.: leaf mines on Cyrtandra sp.
Figs. 816, 817. Liriomyza citreifemorata: 816, aedeagus, side view; 817, same, ventral view.
Figs. 818-820. Paraphytomyza heringi: 818, wing; 819, aedeagus, side view; 820, same, ventral view.
Figs. 821, 822. Paraphytomyza fraxini: 821, aedeagus, side view; 822, same, ventral view (Beiger).
220 Class MAGNOLIOPSIDA (Dicotyledons)
ales. It appears to be a relatively young family, with tinctive, with the outer cross-vein present (Fig. 818).
fossil pollen not known before Upper Miocene. The male genitalia (Figs. 819, 820) suggest possible
The five species of Agromyidae restricted to the relationship with the populi group on Salicaceae.
family all appear to be isolated, with no obvious Pa. fraxini Beiger (1980a) is known only in Bul-
indications of their affiliations. Larval feeding of garia, forming leaf mines on F. oxycarpa. The gen-
Phytobia species has been detected on six genera italia (Figs. 821, 822) show that this species is not
by Suss and Muller-Stoll (1980) and Suss (pers. closely related to Pa. heringi. In Japan P. fraxinivora
comm.), as indicated in Table 39. Sasakawa (1961) is known on two local Fraxinus
species. The genitalia are not available for study.
Subfamily Oleoideae The wing venation is as in Pa. fraxini but there are
significant differences in larval characters, with the
Fraxineae posterior spiracles having 13-16 minute pores in
Fraxinus. Three Paraphytomyza species are known Pa. fraxinivora, three larger, elongate pores in Pa.
on Fraxinus, two in Europe, one in Japan. Pa. fraxini.
heringi Hendel (1920) (cf. Spencer, 1976a) is widely
Oleeae
distributed in association with its host, F. excelsior,
from western Europe to the Soviet Asian Republics Phillyrea. Phytomyza phillyreae Hering (in Buhr,
of Khirgizia, Kazakstan and Uzbekistan. There is 1930, ct. Hendel, 1935) is an isolated, largely yellow
apparently only a single generation between Sep- species forming mines on the small trees, P. angilsti-
tember and November. The wing venation is dis- folia and P. media, pupating deep within the leaves,
''''~
.. c.":-::>'
... -.~ .
823
824
825
Figs. 823, 824. Phytornyza phillyreae: 823, aedeagus, side view; 824, same, ventral view.
Figs. 825, 826. Paraphytornyza jasrnini: 825, aedeagus, side view; 826, same ventral view.
Fig. 827. Phytornyza orobanchia: aedeagus.
Subclass ASTERIDAE 221
an adaptation for surviving the dry, hot summers in Scrophulariaceae (190 genera, 4000 species)
the Mediterranean area where it is only known to
This family is of cosmopolitan distribution but is
occur. The larvae are unusual for this genus, having
best represented in temperate regions. The majority
five elongate pores on the posterior spiracles (nor-
of the agromyzids known on the family are present
mally more in Phytomyza). The male genitalia (Figs.
in Europe and North America, but with five species
823, 824) are unusually elongate, presumably a
restricted to New Zealand, one in East Africa and
parallel development with the female genitalia which
one (at a high altitude) in Venezuela and also in
are modified for inserting the eggs deep within the
California. Only six genera of the Agromyzidae are
leaf tissue.
represented, the great majority (18 of the total of
33) in Phytomyza (see Table 40).
Subfamily Jasminoideae The Scrophulariaceae are now divided into three
subfamilies, each with a number of tribes. The Agro-
Jasmineae myzidae show a high degree of host specificity, with
Jasminum. The only described species known on only Amauromyza verbasci present on two of the
this genus is Paraphytomyzajasmini Spencer (1987) subfamilies, all other species, apart from Phytomyza
which has been reared from J. pubescens near rostrata, being restricted to a single genus. Ten
Bangalore in southern India. The distinctive male Phytomyza and one Chromatomyia species are
genitalia (Figs. 825, 826) give no firm indications of restricted to the hemiparasitic genera in the tribe
its affiliations. Epidermal mines of a Tropicomyia Rhinantheae (Rhinanthoideae), mostly in Europe,
sp. (possibly the polyphagous T flacourtiae) were but these are closely related to two internal feeders
recently found in the Kieni Forest, Aberdare Moun- in the Scrophularieae (Scrophularioideae).
tains, Kenya on J. abyssinicum. The distribution of all species recorded on the
family is shown in Table 41.
Orobanchaceae (17 genera, 150 species)
Subfamily Verbascoideae
This parasitic family is considered to be closely
related to the Scrophulariaceae, or at least to those Verbascum. Two species in Europe are known on
genera which are hemiparasitic, such as Euphrasia, this genus. Melanagromyza verbasci Spencer (cf.
Melampyrum, Odontites and Pedicularis. The single Spencer, 1966b) was discovered feeding in stems of
agromyzid known on Orobanche, Phytomyza oro- V phlomoides in western Germany. The male gen-
banchia Kaltenbach (cf. Spencer, 1973a) supports italia are characteristic of many species in the genus
this, as the genitalia (Fig. 827) indicate close rela- and the posterior spiracles of the puparium are also
tionship with species feeding on these genera, such typical of the genus but distinctive (Fig. 828).
as P. ten ella (Fig. 859). Amauromyza verbasci Bouche (cf. Spencer, 1976a)
P. orobanchia is widespread in Europe and forms large blotch mines primarily on Verbascum
Western Asia in association with its host and has but also on Scrophularia, and interestingly also on
reached Ethiopia. It has been successfully used for Buddleja (see Buddlejaceae). This clearly indicates
biological control of Orobanche which in some the close relationship between Buddleja and the
countries in eastern Europe causes significant dam- Scrophulariaceae, which is now generally accepted
age to tomatoes, beans, tobacco and other crops by botanists, although Buddleja was earlier included
which it parasitizes. The species is unusual, as the in the Loganiaceae (Order Gentianales). The gen-
larva feeds not only in the seeds but also in the
upper stem.
Table 40
Table 41
Agromyzidae on Scrophulariaceae
Verbascoideae
Verbascum Melanagromyza verbasci; Amauromyza verbasci; Liriomyza bryoniae,
L. strigata; Chromatomyia horticola
Scrophularioideae
Antirrhineae
Antirrhinum L. bryoniae, L. strigata, L. trifolii; Chr. horticola
Linaria L. bryoniae, L. strigata, L. trifolii; Chr. horticola
Maurandia Chr. horticola
Scrophularieae
Chelone Phytomyza chelonei
Collinsia Chr. horticola
Cordylanthus P. trivittata
Penstemon P. penstemonella, P. penstemonis
Scrophularia Mel. scrophulariae; Am. verbasci; Chr. horticola
Gratioleae
Mimulus Chr. horticola, Chr. mimuli
Rhinanthoideae
Digi tali deae
Digitalis P. digitalis
Hebe L. flavocentralis, L. hebae, L. umbrina, L. umbrinella, L. umbrosa
Veronica L. strigata; Chr. horticola; P. aphyllae, P. crassiseta, P. "eronicicola
Gerardieae
Striga Ophiomyia strigalis
Rhinantheae
Castilleja Chr. castillejae, Chr. castillejae nordica
Euphrasia P. affinis, P. rostrata
Melampyrum Chr. horticola; P. flavofemorata, P. rostrata
Odontites P. isais, P. rostrata
Pedicularis P. diversicornis, P. pedicularicaulis, P. pedicularidis. P. pedicularifolii,
P. tenella
Rhinanthus Chr. horticola; P. rostrata, P. \aripes
828 830
829
832 833
834 835
836 837
Fig. 828. Melanagromyza verbasci: posterior spiracles of puparium.
Fig. 829. Amauromyza verbasci: aedeagus.
Fig. 830. Phytomyza chelonei: aedeagus.
Fig. 831. Phytomyza trivittata: aedeagus.
Figs. 832, 833. Phytomyza penstemonis: 832, aedeagus, side view; 833, same, ventral view.
Figs. 834, 835. Chromatomyia mimuli: 834, aedeagus, side view; 835, distiphallus, ventral view.
Figs. 836, 837. Phytomyza digitalis: 836, aedeagus, side view; 837, distiphallus, ventral view.
224 Class MAGNOLIOPSIDA (Dicotyledons)
California, Colorado and Nevada and is believed to which five are in Liriomyza in New Zealand and
feed in the seeds of C. nevinii. The male genitalia are four in Phytomyza in Europe.
shown in Fig. 831. This species is also related to the Digitalis. Only the single species, Phytomyza digi-
group on the Rhinantheae. talis Hering (192 5b) is known as a leaf miner on this
Penstemon. Two leaf miners are known on this genus. The genitalia (Figs. 836, 837) indicate its
genus, P. penstemonis Spencer (1969b) which was close relationship with the two species known on
described from Alberta and is also known in Missis- Veronica, P. crassiseta and P. veronicicola.
sippi, and Washington State, and P. penstemonella Hebe. Although not all botanists agree that it is
Spencer (1981), known only in California. The gen- justified to treat Hebe as distinct from Veronica, this
italia of P. penstemonis (Figs. 832, 833) show that is accepted here, as five Liriomyza species are
this species is not related to the two internal feeders known in New Zealand on Hebe where its distinct-
on this tribe. The genitalia of P. penstemonella are ness from Veronica is not questioned. Hebe is clear-
not known but it is probable that the two species are ly an old element in the New Zealand flora, with
closely related. P. penstemonella pupates internally, some 100 species known which are mostly endemic
P. penstemonis externally, the former thus probably but two are shared with South America and one ex-
being the more primitive of the two. tends to the Falkland Islands (Allan, 1961).
Scrophularia. Melanagromayza scrophulariae Four Liriomyza leaf miners on Hebe were de-
Spencer (1981) is only known from California, scribed by Watt (cf. Spencer, 1976b). The male
feeding as an internal stem-borer in S. californica. genitalia show that they are closely related and have
The male genitalia are typical of the genus. Amauro- radiated, possibly in isolation or on different Hebe
myza verbasci which normally feeds on Verbascum species. The genitalia are distinctive in each species
(see above) has also been found on Scrophularia in - L. Jlavocentralis (Fig. 838), L. umbrina (Fig. 839),
Europe. L. umbrinella (Fig. 840) and L. umbrosa (Fig. 841).
I discovered a fifth species, L. hebae Spencer
Gratioleae (1976b) which was present both in Auckland and
Only Mimulus in this tribe has been colonized by Wellington (Fig. 842).
the single agromyzid, Chromatomyia mimuli Spen- All these species except L. Jlavocentralis are
cer (1981) (Figs. 834, 835), known in California on remarkable in being almost totally black, including
M. aurantiacus. I also found it to be particularly the scutellum, the yellow colour of which is normally
common on Phacelia spp. (Hydrophyllaceae) and its a diagnostic character of Liriomyza. However, their
close relationship with Chr. eriodictyi Spencer generic position has been confirmed by the presence
(1981) on the same family (cf. Figs. 732, 733) of the stridulating organ, which is present in all
suggests that Phacelia is possibly its primary host, males in Liriomyza. The exact host range on Hebe
with subsequent transfer to Mimulus. It is agreed of these five species remains to be established and,
that the two families Hydrophyllaceae and Scrophu- in view of the large number of Hebe spp. known in
lariaceae, although in different orders, are fairly New Zealand, with considerable hybridism, careful
closely related. and detailed collecting will be necessary before this
can be clarified.
Veronica. P. crassiseta Zetterstedt (cf. Hendel,
Subfamily Rhinanthoideae
1935) occurs commonly on this genus in Europe,
This subfamily embraces three tribes, of which and is also present in the United States and Chile,
plants in the Digitalideae have a normal root system, almost certainly as an introduction. Pupation taking
while in the Gerardieae and Rhinantheae most are place in the leaf would clearly facilitate chance
hemiparasitic. This division is fully reflected in the introduction. The male genitalia (Figs. 843, 844)
host selection of the agromyzids, with no overlap show the close relationship with P. veronicicola
between hosts in the three tribes. Hering (1925) (Figs. 846, 847) which is known only
in Europe. In P. crassiseta the arista is conspicuous-
Digitalideae ly thickened (Fig. 845) but is normal in P. veronici-
The three genera on which agromyzids are known cola. However, in P. digitalis the arista has the same
have been colonized by nine host-specific species, of apomorphic form as in P. crassiseta but the genitalia
~tkF"':~
838 839
840 841
842
845
, "
\. ~- -. --.' .....
846
226 Class MAGNOLIOPSIDA (Dicotyledons)
848
851 850
852 853
"
".
,'.'
<:".;';"
indicate that these two species are less closely some variation in length and curvature in different
related than the two feeders on Veronica. populations. Six species, including P. melampyri
P. aphyllae Beiger (1964) is only known on Ve- Hering, were therefore synonymised with P. flavo-
ronica aphylla above 1600 m in the Tatra Moun- femorata after examining the genitalia of a syntype
tains in Poland. The genitalia (Fig. 848) show that of Strobl's species from Celje, northern Yugoslavia,
this species is not directly related to the two other . and of the holotypes of the other six (Spencer,
species on Veronica. 1976a: 421). Dr. v. Tschirnhaus (pers. cornm.) has
recently suggested that two further species feed in
Gerardieae the seeds of Melampyrum. It has not been possible
The hemiparasitic genus Striga has been colonized to examine any specimens of these species and, in
by a single species, Ophiomyia strigalis Spencer view of the variation both in colour and in the male
(1963c), the larva feeding in the lower stem and root genitalia in this group which I have detected, I here
of S. hermonthica in East Africa. This species maintain my earlier diagnosis.
belongs to the group lacking the vibrissal fasciculus P. rostrata Hering (cf. Spencer, 1976a) was de-
in the male but the genitalia (Spencer, 1963c: fig. scribed from specimens caught on Melampyrum
1 7b) are characteristic of the genus. With the squa- nemorosum and was later reared by Buhr from this
mae and fringe white, it is probably related to other genus and also from Euphrasia odontites and Rhi-
species with this character. nanthus. It is thus the only species in this group
feeding on more than one genus. Buhr, quoted by
Rhinantheae Hering (1960: 143) gives a detailed account of larval
Six genera in this tribe of hemiparasitic species each feeding, which takes place primarily in the pith of
have their own specific agromyzids, ten in Phyto- the stem but always commences with a short leaf
myza in Europe and northern Canada and one in mine. The genitalia (Fig. 855) suggest that this
Chromatomyia known in California and Venezuela. species may be most closely related either to the
Only one species, P. rostrata, has a more extensive seed-feeder on Odontites, P. isais, or the leaf-miner
host range, recorded on four genera. The Phytomy- on Pedicularis, P. pedicularifolii.
za species are closely related and it is accepted that Odontites. P. isais Hering (cf. Spencer, 1976a)
they are monophyletic but three groups are detect- feeds only in seeds of O. rubra. The genitalia (Fig.
able, two feeding in seeds, one in stems. The one 856) indicate the relationship with P. flavofemorata.
exclusive leaf miner, P. pedicularifolii, as might be The only other species known on this genus is the
expected, is somewhat isolated. stem-feeder, P. rostrata (see above).
Castilleja. Chromatomyia castillejae (Spencer, Pedicularis. Five species feed only on this genus,
1973c) was described from Venezuela in the moun- two in seeds and two in stems, one in each group
tains north of Caracas (1000 m) on C. fissifolia, isolated geographically between Europe and North
forming short leaf mines and pupating in the leaf. America (northern Canada). The fifth species is
The genitalia (Figs. 849, 850) distinctively differ exclusively a leaf miner.
from those on C. latifolia in California (Figs. 851, P. diversicornis Hendel (cf. Spencer, 1976a) (Fig.
852), which I treated as a subspecies, Chr. castillejae 857) feeds internally in the pith of the stems of
nordica (Spencer, 1981). It might be justified to Pedicularis palustris in Europe. P. pedicularicaulis
treat these two populations as representing distinct Spencer (1969b) (Fig. 858) is known only in Que-
species. However, until the populations are known bec, feeding in the stems of Ped. canadensis. The
which are certainly present in Mexico and Central male genitalia clearly show that the two are sister-
America, I prefer to treat this as a case of incipient species, with dispersal of the ancestral population
speciation in the two now isolated populations. from Europe either via Beringia or across the North
Euphrasia. P. affinis Fallen (cf. Spencer, 1976a) Atlantic via Iceland and Greenland.
is host-specific on this genus, feeding in the seeds. In Europe P. tene/la Meigen (cf. Spencer, 1976a)
The genitalia (Fig. 853) indicate close relationship (Fig. 859) is common in seeds of Ped. palustris and I
with P. tenella on Pedicularis. P. rostrata has been think it can be accepted that this is the sister-species
found in stems of this genus and it is discussed of P. pedicularidis Spencer (1969b), known only in
further below under Melampyrum. Quebec and Labrador where it has been found in
Melampyrum. There has been confusion about seeds of Ped. groenlandica. The genitalia (Fig. 860)
the species commonly found on this genus but confirm their close relationship.
following examination of all available material at the The only species feeding exclusively as a leaf
time, I decided that the prior name for the seed- miner in this group is P. pedicularifolii Hering
feeding species is P. flavofemorata Strobl (cf. Spen- (1960) which forms blotch mines on Ped. palustris
cer, 1976a). The large, paired, black processes of in Europe. The genitalia (Figs. 861, 862) show that
the distiphallus (Fig. 854) are distinctive but show this is an isolated species, possibly related to
228 Class MAGNOLIOPSIDA (Dicotyledons)
I. _. . . .
"':'-y"
.....
,.: ....
........
:~-r
857
"",
856
::"/
....
.I!!IIJ.; ,- " ,
-----.
859
858
860 861
.::-::,-~-:-'
--:,
.-" . "". '"\>
"\'
/~~~;t~;:t/·i:-~~~~~ ',::, .
863 862
Fig. 856. Phytomyza isais: aedeagus.
Fig. 857. Phytomyza diversicornis: aedeagus.
Fig. 858. Phytomyza pedicularicaulis: aedeagus.
Fig. 859. Phytomyza tenella: aedeagus.
Fig. 860. Phytomyza pedicularidis: aedeagus.
Figs. 861, 862. Phytomyza pedicularifolii: 861, aeadeagus, side view; 862, same, ventral view.
Fig. 863. Phytomyza varipes: aedeagus.
Subclass ASTERIDAE 229
.~:"'''i'.'
864
865
867
866
870
868
869
871
Spencer, 1964c) (Fig. 866) feeds on C. persicifolia, In New Zealand L. wahlenbergiae Spencer
Jasione and Phyteuma, and also on ten genera of (1976b) was caught on Wahlenbergia albomarginata
Asteraceae, mainly in the tribe Lactuceae (Table 60) and this is accepted as the host. The genitalia (Fig.
and also on Matricaria (Anthemideae). This is one 870) show close relationship with L. craspediae
of two species on Campanulaceae which are also Spencer (1976b), a leaf miner on Craspedia uniflora
found on Asteraceae. In view of the oligophagy on (Asteraceae), also in New Zealand (cf. Fig. 1031). In
both families of 0. heringi, it is unclear which is its North America no described species are known on
primary host family. Campanula but I found a mine (Fig. 871) of a Lirio-
Amauromyza gyrans (Fallen, d. Spencer, 1976a) myza on C. americana at Pelee, southern Ontario
feeds primarily on Campanula but also not uncom- near the U.S. border.
monly on Phyteuma spicata and more rarely on A single Napomyza is known on the family in
Legousia, and in Japan on Adenophora, forming a Europe, N. hirticornis Hendel (cf. Spencer, 1976a),
white linear-blotch mine. The genitalia (Fig. 867) the original host being Jasione montana. It was later
associate this species with others in the genus, par- found on Centaurea in Ireland (Griffiths, 1968). The
ticularly Am. flavifrons (Caryophyllaceae), Am. genitalia (Fig. 872) indicate close relationship with
labiatarum (Lamiaceae) and Am. verbasci (Scro- the common N. lateralis feeding on Asteraceae (cf.
phulariaceae). Fig. 1167).
In Europe Liriomyza buhri Hering (d. Spencer, Two Phytomyza species feed on Campanula.
1976a) is unusual in forming external stem-mines, P. campanulae Hendel (cf. Spencer, 1976a) is wide-
mainly on a number of Campanula species but also spread in Europe on Campanula spp., forming an
on Jasione and Phyteuma. The genitalia (Figs. 868, irregular, whitish linear mine. The genitalia (Fig.
869), with the elongate tubules of the distiphallus, 873) associate it with species of the albiceps group
could possibly associate this species with L. lutea on Asteraceae (see below under Asteraceae). A
(feeding in seeds of Heracleum and Pastinaca, related species, though with well differentiated gen-
Apiaceae) and with others whose host is unknown, italia (Figs. 874, 875), P. rapunculi Hendel (d.
such as L. canescens Spencer (1976a: fig. 405). It Hendel, 1935) occurs on Campanula spp. and
cannot be excluded that there is a direct, although Phyteuma spp., with a disjunct distribution in the
remote relationship between L. buhri and L. lutea, Bavarian Alps, the Polish Tatra Mountains (as P. ta-
comparable to the Phytomyza albiceps group on trica Beiger) and in central Sweden (as P. campanu-
Apiaceae and on Asteraceae. Transfers between livora Spencer). This also belongs to the albiceps
Apiaceae, Campanulaceae and Asteraceae could group. P. rapunculi thus represents the senior syno-
clearly have been facilitated by the presence of nym of both P. campanulivora Spencer (1971a) and
polyacetylenes in all three families. P. tatrica Beiger (1972a) (see Appendix below).
Table 43
Campanulaceae
Campanuloideae
Adenophora Amauromyza gyrans
Campanula Ophiomyia campanularum, O. eucodonus, O. heringi; Am. gyrans;
Liriomyza buhri, L. strigata, Liriomyza sp. ( Canada );
Phytomyza campanulae, P. rapunculi
Canarina L. strigata
Jasione O. heringi, L. buhri, L. strigata; Napomyza hirticornis;
Chromatomyia horticola
Legousia Am. gyrans
Phyteuma O. heringi; Am gyrans; L. buhri, L. strigata; Chr. horticola; P. rapunculi
Symphandra L. strigata
Wahlenbergia L. wahlenbergiae
Goodeniaceae
Goodenia O. cornuta
Scaevola O. cornuta; L. scaevolae
232 Class MAGNOLIOPSIDA (Dicotyledons)
l~--
H- ":»'5>.
874
873 875
877
878 876
Fig. 873. Phytomyza campanulae: aedeagus.
Figs. 874, 875. Phytomyza rapunculi: 874, aedeagus, side view; 875, distiphallus, antero-ventral view.
Figs. 876, 877. Ophiomyia comuta: 876, leaf mine on Goodenia ovata; 877, aedeagus.
Fig.878. Liriomyza scaevolae: aedeagus.
Only two polyphagous species are known on the family) and the Asterales, with just the Asteraceae.
family, Liriomyza strigata and Chromatomyia horti- There are two cases of species feeding on both
cola (see Table 43). Campanulaceae and Asteraceae. Ophiomyia heringi
The primary relationship of species on Campanu- feeds not only on Campanulaceae but also on ten
laceae is with those on Asteraceae and this supports genera of Asteraceae, and Napomyza hirticornis has
the view of Stebbins (1977) who writes that the once been found on Centaurea in Ireland (Griffiths,
Asteraceae and Campanulaceae may have had an 1968). In other cases a clear relationship is detect-
'immediate or not very distant common ancestor,' able between species on Campanulaceae and Aster-
and also the taxonomic arrangement proposed by aceae, in particular Liriomyza wahlenbergiae and
Frohne and Jensen (1985) in which their Subclass L. craspediae, and Phytomyza campanuiaelP. rapun-
Asteridae contains just two orders, the Campanula- culi and the P. albiceps group.
les (in which Lobeliaceae are treated as a separate Other relationships indicate switches for which
Subclass ASTERIDAE 233
there is no obvious explanation, such as Ophiomyia family into three subfamilies with a total of 29 tribes,
campanularum - 0. melandricaulis (Caryophylla- based on the study of 11 characters, particularly
ceae), Amauromyza gyrans - A. flavifrons (Caryo- hairs and pollen. The Agromyzidae known on the
phyllaceae), A. labiatarum (Lamiaceae), A. verbasci family are shown in Table 45.
(Scrophulariaceae), and Liriomyza buhri - L. lutea The Rubiaceae are primarily of tropical or sub-
(Apiaceae). tropical distribution but the only genus on which
there has been significant radiation is Galium, in the
northern hemisphere. Only the surprisingly small
Goodeniaceae (14 genera, 300 species)
number of 16 species have so far been recorded on
This small family is primarily Australian but extends this large family, of which 13 are restricted to the
to New Zealand, the Pacific and with a few species family. It is noteworthy that no Melanagromyza,
more widely distributed. Only Goodenia and Scae- Chromatomyia or Phytomyza species have colo-
vola have been colonized by Agromyzidae. nized the family and the number of polyphagous
Ophiomyia cornuta de Meijere (1910, d. Spen- species is limited to four, with Liriomyza strigata
cer, 1977 c) feeds primarily on Scaevola but I found and Chromatomyia horticola notable absentees. The
mines also on Goodenia ovata, north of Sydney, breakdown of species by genera is shown in Table
Australia. The larva forms a narrow, winding mine 44.
(Fig. 876) and the genitalia (Fig. 877) show that this
is an isolated species, although typical of the genus Table 44
on external characters.
The other species known on the family, Liriomyza Agromyzid genera on Rubiaceae
scaevolae Spencer (1977c), I found to be abundant Genus Species restricted Polyphagous
in association with its host, the bush Scaevola nitida, to family species
on the SW coast of Western Australia. The genitalia
(Fig. 878) are not atypical and its nearest relatives Hexomyza 2
cannot be determined. The species has certainly Ophiomyia 1
Tropicomyia 1 (undescribed) 3
long been isolated on Scaevola.
Liriomyza 1
Galiomyza 1
Paraphytomyza 6
Order RUBIALES (2 families, 6500 species)
Gymnophytomyza 1
This order consists essentially of the large family
13 4
Rubiaceae, with three species in Theligonum given
family status as Theligonaceae, closely associated
with the Rubiaceae.
Subfamily Rubioideae
In New Zealand Hexomyza coprosmae Spencer
Rubiaceae (450 genera, 6500 species)
(1976b) forms galls (Fig. 880) on the bush Copros-
The systematic position of the Rubiaceae is still in ma propinqua. The genitalia (Fig. 879) show that
dispute. Takhtajan (1969), following earlier studies this is an isolated species, not closely related to
by other workers, and also Frohne and Jensen Hexomyza species known in Europe. However, in
(1985) include the family in the Gentianales. Cron- Japan Sasakawa (1954b) described Melanagromyza
quist (1981), whom I follow here, places the family paederiae which forms similar galls (Fig. 881) on the
in the separate order Rubiales between the Cam- twigs of Paederia chinensis. I consider that this
panulales and Dipsacales. He comments: 'The Ru- species should be treated as a Hexomyza and this
biaceae form a connecting link between the Gentia- combination is established in the Appendix below.
nales and the Dipsacales and would be an aberrant This is justified both by the gall-forming habit of the
element in either order.' There are no relationships larvae and the presence on the posterior spiracles of
between the Agromyzidae on Gentianaceae and only three spiracular pores (always more in Melana-
Rubiaceae, but the one genus Paraphytomyza has six gromyza). The male genitalia (Fig. 882) could indi-
species on Rubiaceae, and in the Dipsacales nine on cate that this species is monophyletic with H. cop-
Caprifoliaceae and one on Dipsacaceae. This sup- rosmae and intermediate forms probably await
ports a close relationship between these three fami- discovery in SE Asia or Australia.
lies. Galium is the favourite host genus, with nine
There is a similar lack of agreement among agromyzids in four genera feeding on it either in the
botanists on the internal classification of the Rubia- stem or as leaf miners.
ceae and two systems are outlined by Willis (1985). Ophiomyia galii Hering (ct. Spencer, 1964c) is
Verdcourt (1958), whom I follow here, divided the known in Europe on Asperula and Galium mol/ugo,
234 Class MAGNOLIOPSIDA (Dicotyledons)
882
879
884 885
887
Figs. 879, 880. Hexomyza coprosmae: 879, aedeagus; 880, gall on Coprosma propinqua.
Figs. 881, 882. Hexomyza paederiae: 881, gall on Paederia chinensis; 882, aedeagus (Sasakawa).
Fig. 883. Liriomyza galiivora: aedeagus.
Fig. 884. Galiomyza morio: aedeagus.
Fig.885. Paraphytomyza anteposita: aedeagus.
Fig.886. Paraphytomyza buhri: aedeagus.
Fig. 887. Paraphytomyza galii: aedeagus.
Subclass ASTERIDAE 235
forming external stem mines. The male genitalia are for the isolated species G. heteroneura in Europe,
characteristic of the genus but give no indication of feeding in individual seeds of Galium aparine and
its nearest relative. Liriomyza galiivora (Spencer, pupating externally. The species is probably related
1969b, originally described in Praspedomyza) was to the group of Paraphytomyza on Galium but the
first found in Alberta and Quebec, later in Min- simplified genitalia (Fig. 889) show it to be well
nesota, U.S.A. and subsequently identified in Eu- differentiated.
rope from Corsica and Berlin. Von Tschirnhaus
considers that the species belongs in Liriomyza but
Subfamily Cinchonoideae
in view of its morphological characters and partic-
ularly its genitalia (Fig. 883) which clearly associate No host-specific species have been described on this
it with G. morio, I feel the relationship of these two subfamily but the conspicuously green mines I
species requires further study. found in northern Thailand on the small bush,
Galiomyza morio (Brischke, 1880, cf. Spencer, Canthium horridum, appear to represent a new
1976a) is an unusual species forming leaf mines on Tropicomyia sp. (Spencer, 1986c).
Asperula, Galium, Rubia and Sherardia. With the Tropicomyia theae (Cotes, cf. Spencer, 1973a =
head and scutellum black, its generic position was T. coffeae Koningsberger, see Appendix below) is
unclear until it was finally included in the small now accepted as polyphagous and in Papua New
genus Galiomyza (Spencer and Steyskal, 1986), Guinea it was particularly common on Coffea (Figs.
together with two species in the United States, one 890,891). I also found mines I ascribe to it in Papua
on Galium, one on Viola. The male genitalia (Fig. New Guinea on Apocynaceae, Passifloraceae and
884) indicate its isolated position. Sapindaceae, and also on the monocot family Dios-
Six monophyletic Paraphytomyza species are coreaceae. The genitalia (Figs. 260, 261) ex tea are
known on Galium, of which five are present in well differentiated from other species in this genus,
Europe, one in California. including T. flacourtiae (Figs. 278, 279) (Flacourtia-
Pa. anteposita (Strobl, cf. Spencer, 1966e). Host: ceae, Dilleniidae), which occurs commonly on CoF
G. aparine, internal stem-feeder. Male genitalia: Fig. fea in East Africa where it was redescribed as
885. Melanagromyza coffeae (Hering, 1940). Five further
Pa. buhri (de Meijere, cf. Spencer, 1976a). Host: genera in Africa, India and Australia have been
G. mollugo and Asperula, external stem-miner. colonized by polyphagous Tropicomyia spp. (only
Male genitalia: Fig. 886. Morinda in the Rubioideae). In the United States the
Pa. galii (Groschke, 1957). Host: G. rotundifo- highly polyphagous Liriomyza schmidti has been
lium, only found once in Bavaria. Oviposition takes recorded on Chiococca and Hamelia (see Table 45).
place in the stem but the larva quickly moves to a It will be seen from Table 45 that larval feeding of
leaf where a large blotch mine is formed. It may feed Phytobia has been detected in the single genus
on via the petiole and stem to a second leaf. The Urophyllum in the Rubioideae (Java) and in ten
genitalia (Fig. 887) are somewhat differentiated genera in the Cinchonoideae in Africa, South Amer-
from those of the stem-feeders, reflecting the more ica and Java. It is certain that a number of different
isolated position of this species as the only leaf mi- species are involved and it will be of great interest to
ner in the group. examine adults when these can in due course be
Pa. lucens (de Meijere, cf. von Tschirnhaus, obtained in association with their specific hosts.
1969). Host: G. aparine (on which specimens The Agromyzidae on the Rubiaceae are of inter-
caught but none reared). The male genitalia associ- est in the almost total absence of the large genera
ate the species with the other Galium feeders. Phytomyza and Liriomyza Gust one species) and the
Pa. orphana (Hendel, 1920, cf. Spencer, 1976a). general inacceptability to polyphagous species. In
Hosts: G. aparine, G. palustris. Male genitalia: Fig. contrast, there is striking radiation on Galium of the
888, of similar form to Pa. buhri but more strongly small genus Paraphytomyza. The detection of larval
pigmented. This is more common than Pa. lucens feeding of Phytobia on 11 genera in tropical areas
and with the same primary host, it seems probable of both the Old and New World is also noteworthy
that there is some difference in the larval feeding of and it will be of great interest to discover the affilia-
the two species but this remains to be clarified. tions of these species in due course.
Pa. praecox Spencer (1981). This is the only
species in the group known outside Europe, in
Order DIPSACALES (4 families,
central California. From its external characters, it is
1000 species)
assumed that this is an internal feeder in Galium but
no specimens have been reared and no male is 'The Caprifolaceae are obviously the most archaic
known. family in the Dipsacales. The Valerianaceae appear
Gymnophytomyza was erected by Hendel (1936) to take their origin directly from the Caprifoliaceae.
236 Class MAGNOLIOPSIDA (Dicotyledons)
Table 45
Agromyzidae on Rubiaceae
After Verdcourt (1958) and pers. comm.
Subfamily Rubioideae
Anthospermeae
Coprosma Hexomyza coprosmae
Hamelieae
Hamelia Liriomyza schmidti
Morindeae
Morinda Tropicomyia polyphyta (Australia)
Paederieae
Paederia Hex. paederiae
Rubieae
Asperula Ophiomyia gaJii; Paraphytomyza buhri;
Galiomyza morio
Galium O. galii; L. galiivora; Par. anteposita, Par. buhri,
Par. galii, Par. lucens, Par. orphana,
Par. praecox ( ? ); G. morio;
Gymnophytomyza heteroneura
Rubia G. morio
Sherardia G. morio
Urophylleae
Urophyllum glabrum Wall. Phytobia sp. (larval feeding, Java)
Subfamily Cinchonoideae
Chiococceae
Chiococca L. schmidti
Cinchoneae
Corynanthe Phytobia sp. (larval feeding, West Africa)
(= Pseudocinchona)
Mitragyna macrophylla Hiern. Phytobia sp. (larval feeding)
(tropical Africa)
Mitragyna sp. (West Africa) Phytobia sp. (larval feeding)
Gardenieae
Diplospora javanica Koord. et Valet. Phytobia sp. ( larval feeding) - Java
Gardenia T. polyphyta ( Australia)
Heinsieae
Bertiera montana Hiern. Phytobia sp. ( larval feeding) - Java
Isertieae
Mussaenda T. flacourtiae ( Africa)
Muss. frondosa L. Phytobia sp. ( larval feeding) - Java
Subclass ASTERIDAE 237
Naucleeae
Nauclea lanceolata Blume Phytobia sp. ( larval feeding) - Java
Sarcocephalus sp. Phytobia sp. ( larval feeding) - West Africa
Pavetteae
Pavetta T. flacourtiae
Tarenna fragrans Koord. et Valet. Phytobia sp. ( larval feeding) - Java
Vanguerieae
Canthium horridum Tropicomyia sp. n. ( Thailand)
VangueriJIa vanguerioides ( Hiern.) Verd. Phytobia sp. ( larval feeding) - Togo
The Dipsacaceae are a little more removed, but tianaceae. The few agromyzids restricted to the
evidently of the same ancestry. On the basis of fossil Valerianaceae show no relationship to those on the
pollen, it appears that the Caprifoliaceae can be other families in this order.
traced back to the middle Eocene, and the Dipsaca-
ceae to the middle Miocene' (Cronquist, 1981).
Caprifoliaceae (15 genera, 400 species)
There is no general agreement among botanists
on the systematic position of this order. Frohne and Seven genera of this family have been colonized by
Jensen (1985), for example, place the order in a Agromyzidae and a notable exception is Weigela, of
subclass Cornales near the Gentianales. However, eastern Asian origin. No transfers to Weigela from
there is nothing in the relationship of the Agromy- Lonicera or Symphoricarpos, the favourite host
zidae to support this view. On the other hand, the genera, have ever been recorded, although all three
one genus Paraphytomyza has ten species on the plants are not infrequently found in close proximity
Caprifoliaceae, one on Dipsacaceae and six on in parks or gardens in western Europe.
Rubiaceae, but none on either Cornaceae or Gen- There has been considerable radiation in the two
889
888
891
Fig. 888. Paraphytomyza orphana: aedeagus.
Fig.889. Gymnophytomyza heteroneura: aedeagus.
Figs. 890, 891. Tropicomyia theae (ex Coffea): 890, aedeagus, side view; 891, same, ventral view.
238 Class MAGNOLIOPSIDA (Dicotyledons)
Table 47
Agromyzidae on Caprifoliaceae
892 893
894
895
.'.:.:.
.. ~-'. / '
896 897
899
cer, 1976a) is a pale species, also present on both inaccurate and his illustration of the genitalia (Fig.
host genera but the mine is distinctive, being filled 903) shows that an undescribed species is involved.
centrally with greenish frass. The genitalia of Pa. The two European species Chr. lonicerae and
atlantidis Spencer (1967) (Fig. 898), Pa. cornigera Chr. aprilina (Goureau, ct. Griffiths, 1974a) have
(Fig. 896) and Pa. luteoscutellata (Fig. 897) show retained the outer cross-vein which is accepted as a
their close relationship, with the first two appearing primitive character. Chr. lonicerae (Fig. 904) (better
to be sister-species. Pa. hendeliana (Fig. 899) is known as 'Napomyza xylostei' Kaltenbach) is wide-
somewhat isolated from the other three. spread in Europe on both Lonicera and Symphori-
Pa. cornigera and Pa. luteoscutellata are both also carpos, the larva fonning a distinctive mine which
present in North America and it seems probable, in commences with a small stellate area (Fig. 905). In
view of their wide distribution, that this represents Europe Chr. aprilina is restricted to western areas
natural dispersal rather than recent introduction. and has also been found at Tangiers, Morocco. In
In North America Pa. orbitalis (Melander, cf. this species the larva forms a long white linear mine,
Spencer and Steyskal, 1986) is a common species exclusively on Lonicera, and transforms into a
on Lonicera and the genitalia (Fig. 900) confirm its distinctive green puparium. The head of the adult is
close relationship with Pa. cornigera. Pa. lonicerina largely yellow (black in all other species in the
Spencer (1981) forms distinctively green mines on group). These morphological differences are re-
Lon. involucrata in Canada and California. Its gen- flected in the male genitalia (Fig. 906) which differ
italia (Fig. 901) suggest it is a more isolated species. considerably from all others in the group, confirm-
Sehgal (1971b) described Pa. spenceri from Alber- ing its isolated position.
ta, a leaf miner on both Lonicera and Symphoricar- Chr. periclymeni (de Meijere, cf. Griffiths, 1974a)
pos and his illustration of the genitalia (1971b: fig. is common throughout Europe on both Lonicera
76) shows that the species is well differentiated from and Symphoricarpos, forming an irregular blotch
the others known in North America. In the same mine. This, with the following ten species, have been
paper he recorded for the first time the host of Pa. referred to by Griffiths (1974a) as the periclymeni
plagiata as Lon. involucrata in Alberta. The species group. The male genitalia of Chr. periclymeni (Fig.
was previously known in Idaho and Montana in the 907) are characteristic of all others in the group,
western U.S.A. The holotype is a female but, which differ only in minor detail. The greatest dif-
although Sehgal caught a male the genitalia are not ferentiation between these species is in the larval
available for illustration. feeding instinct and thus in the form of the mine they
Pa. loniceroides Spencer (1987) is only known produce.
from females and forms leaf-mines on Lonicera Chr. nervi (Groschke, 1957, ct. Griffiths, 1974a)
quinquelocularis in the low Himalayas, 1500 m, in is only known from five females reared from Lon.
Pakistan. The wing, with the outer cross-vein lacking alpigena in Bavaria. The larva forms an irregular
is shown in Fig. 902. blotch mine on the midrib, and the affiliations of the
Griffiths (1974a) clarified in detail the 13 Chro- species seem clear.
matomyia species on Linnaea, Lonicera and Sym- In North America there are four generally similar
phoricarpos and illustrated their genitalia. Of these, species on Lonicera. Chr. gregaria (Frick, ct. Grif-
five occur in Europe, seven in North America, with fiths, 1974a) forms a complex blotch mine along the
only a single species having a boreal, hoi arctic dis- midrib, with many larvae feeding together (Fig.
tribution. Lonicera is again the favourite host genus 908). The only known host is L. involucrata and the
with nine species, two occur on both Lonicera and species has only been found in Alberta, British
Symphoricarpos, three only on Symphoricarpos. Columbia and California. On the same host adults
Sasakawa (1961) identified as Napomyza xylostei of Chr. involucratae Spencer (1969b) have been
Kaltenbach a species on Lonicera gracilipes and caught in Alberta and British Columbia, and the
L. japonica in Japan. This name represents a syno- genitalia clearly indicate that this is the host. Chr.
nym of Chr. lonicerae (Robineau-Desvoidy, cf. nigrilineata Griffiths (1974a) forms an irregular
Griffiths, 1974a) but Sasakawa's identification is blotch mine on L. dioica in Alberta and this species
902 903
906 907
"~'--------""''''''''''
('
----
.,,
,
, I
I
I
I
,I
,, ~'
,'.
....... _------
,,
".,.. ..... -----_ ....
910
242 Class MAGNOLIOPSIDA (Dicotyledons)
is distinctive, having a black line centrally along the ceae but the relationship of the two groups is
ventral surface of the white puparium. evident. The nearest earlier family on which Chro-
The larvae of only two species in Chromatomyia matomyia species have radiated is the Gentianaceae
leave their mines to pupate on the ground - Chr. and it can be postulated that the ancestral switch to
alpigenae (Groschke, 1957) in Europe and Chr. Caprifoliaceae occurred from this family. In the
chamaemetabola Griffiths (1974a) in Alberta. The species on both families the outer cross-vein is in
genitalia indicate that these are sister-species. Chr. process of being lost. However, this primitive vein
alpigenae appears to be restricted to mountains in has been retained in Pa. comigera, fallax, hendelia-
central Europe and the most frequent hosts are na, lonicerina and plagiata but is absent in the other
Lonicera alpigena and L. nigra. Many larvae may specIes.
feed together, forming a large blotch mine based on
the midrib. The only known host of Chr. chamae-
Dipsacaceae (10 genera, 270 species)
metabola is L. involucrata on which a similar com-
munal mine to that of Chr. alpigenae is formed (Fig. This small family is native to Eurasia and Africa and
909). The mine somewhat resembles that of Chr. is only known to have been colonized by Agromy-
gregaria but here the puparia remain in the leaf (Fig. zidae in Europe, with a total of ten species in five
90S). genera (Table 4S). Five of the available ten genera in
The only holarctic species in this group is Chr. the family are known to support Agromyzidae, as
linnaeae Griffiths (1974a), which was described shown in Table 49.
from Linnaea borealis in Alberta and the Yukon. I
found an immature mine on the same host in Table 48
northern Norway near Narvik on 3 July, 1972. The
Agromyzid genera on Dipsacaceae
male genitalia (Fig. 910) show that this species is
most closely related to the three restricted to Sym- Restricted to family Polyphagous
phoricarpos in North America, Chr. caprifoliae
Spencer (1969b), Chr. fricki Griffiths (1974a) and Ophiomyia
Chr. symphoricarpi Griffiths (1974a). Agromyza 2
Liriomyza
The only Phytomyza known on the Caprifolia- Paraphytomyza
ceae is P. abeliae Sasakawa (1961) in Japan, which Chromatomyia 4
is a small species forming short mines on Abelia
serrata. Sasakawa emphasises that he is unable to 8 2
identify any related Japanese species and the male
genitalia (Fig. 911) suggest long isolation on its host
in Japan. Ophiomyia longilingua (Hendel, = 0. knautiae
It is noteworthy that polyphagous species largely Hering, cf. Spencer, 1976a) is an aberrant species
avoid this family both in Europe and North Amer- which is restricted to Knautia, forming reddish stem
ica. The only exception is Tropicomyia polyphyta mines. The genitalia (Fig. 912) confirm its isolated
which has been found on Lonicera periclymenum in position.
Queensland, Australia. Of the two Agromyza species on this family,
There is no obvious chemical explanation for the A. dipsaci Hendel (cf. Hendel, 1931) is restricted to
limited colonization of the Caprifoliaceae by Agro- Dipsacus, while its sister-species A. woerzi Grosch-
myzidae. It is interesting to note that no species in ke (1957) is only known on Knautia. Both form
the large genera Melanagromyza, Ophiomyia or funnel-shaped mines always along the leaf margin
Agromyza have succeeded in establishing them- and the complex male genitalia (A. dipsaci: Figs.
selves on the family. The two Liriomyza species and 913, 914 and A. woerzi: Figs. 915, 916) are of the
the single Phytomyza represent isolated transfers same unique form but differing considerably in
and it cannot be established from which families detail. Although the species are monophyletic, there
these transfers took place. In both Paraphytomyza has been significant divergence in A. dipsaci in the
(ten species) and Chromatomyia (14) I consider larva, in which the primitive arrangement of three
that, given the general unacceptability of the family, posterior spiracular pores has evolved into a group
there was a single original colonization with remark- of about 40, while in A. woerzi which retained the
able subsequent radiation in both genera. It seems three pores, in the wing there has been a reduction
certain that in Paraphytomyza the switch to Capri- in the costa which extends only to vein R4+ 5, not to
foliaceae came from the Rubiaceae on which more M1 +2 which is certainly the ancestral condition still
primitive stem-feeding species are present. The present in A. dipsaci. A mine on Cephalaria in
genitalia of the leaf miners on Caprifoliaceae show Bessarabia (Hering, 195 7b) almost certainly repre-
considerable differentiation from those on Rubia- sents either A. dipsaci or A. woerzi.
Subclass ASTERIDAE 243
912 ~
915
~~~~/:e,~,
L~~~~/~ 917
916
~
~~
920
918
922
Fig. 912. Ophiomyia longilingua: aedeagus.
Figs. 913, 914. Agromyza dipsaci: 913, aedeagus, side view; 914, same, ventral view.
Figs. 915, 916. Agromyza woerzi: 915, aedeagus, side view; 916, same, ventral view.
Fig. 917. Paraphytomyza similis: aedeagus.
Fig. 918. Chromatomyia scabiosarum: aedeagus.
Figs. 919, 920. Chromatomyia succisae: 919, aedeagus, side view; 920, distiphallus, (dorsal view).
Figs. 921, 922. Chromatomyia scabiosae: 921, aedeagus, side view; 922, same, (dorsal view).
244 Class MAGNOLIOPSIDA (Dicotyledons)
Table 49
Agromyzidae on Dipsacaceae
Paraphytomyza similis (Brischke, d. Spencer, known. Agromyza dipsaci and A. woerzi are prob-
1976a) forms whitish linear-blotch mines on Knau- ably related to species on Boraginaceae, which is far
tia and Succisa. This is an isolated species, although removed in the Lamiales placed here early in the
the genitalia (Fig. 917) suggest direct relationship Asteridae. The three pairs of ventral lobes which are
with the species on Galium (Rubiaceae, see above). highly developed in A. woerzi (Fig. 915) are unique
Two Chromatomyia species are restricted to within the genus Agromyza. Although Paraphyto-
Scabiosa. Chr. scabiosae (Hendel, 1935) and Chr. myza similis can be accepted as related to the
scabiosarum (de Meijere, 1934) are not directly species on Rubiaceae and Caprifoliaceae in this
related, the latter being the sister-species of Chr. order, its genitalia have now significantly evolved
succisae, which occurs on all four genera, Dipsacus, since this family was colonized by any ancestor
Knautia, Scabiosa and Succisa throughout much of which can be postulated. The two groups of Chro-
Europe, while Chr. scabiosarum is restricted to the matomyia, Chr. scabiosarum and Chr. succisae, and
Mediterranean area. The larva of Chr. succisae Chr. scabiosae, are both more isolated than any
forms a narrow linear mine which can be largely other groups in the genus. Finally, Chromatomyia
obscured by strong purple discoloration of the leaf ramosa has highly complex genitalia which appear
around the mined area. The mine of Chr. scabio- to have evolved in isolation on the Dipsacaceae.
sarum is very similar to that of Chr. ramosa (see Fig. A strange aspect of evolution is apparent in
924), with short offshoots into the leaf blade from Agromyza dipsaci and A. woerzi. While the two
the midrib. The genitalia (Chr. scabiosarum: Fig. Agromyza species are very similar externally, evolu-
918) and Chr. succisae: Figs. 919, 920) indicate that tionary differentiation is apparent in A. woerzi in
this species pair represent an isolated group within the wing with reduction of the costa and in A. dip-
Chromatomyia. The genitalia of the species I accept saci in the larva with the remarkable proliferation of
as Chr. scabiosae (Figs. 921, 922) are typical of the pores on the posterior spiracles. It is interesting
Chromatomyia, although this is also obviously an to note that the larval feeding is virtually identical in
isolated species. Beiger (197 5b) identified as Chr. Chromatomyia scabiosarum and Chr. ramosa (Fig.
scabiosae a specimen from Scabiosa sp. in Poland 924). Such mines, with lateral offshoots into the leaf
but her illustration of the genitalia suggests that this blade from the midrib, are rare but not unique but are
is a further species which remains to be clarified. characteristic, for example, of Ophiomyia cunctata
Chromatomyia ramosa (Hendel, d. Spencer, 1976a) (Asteraceae: Lactuceae) and of the two polyphagous
feeds most commonly on Dipsacus but is also found species, Liriomyza huidobrensis and L. strigata.
on Knautia and Succisa. The mine on Dipsacus is
characteristic, with short offshoots from the midrib
Valerianaceae (13 genera, 300 species)
(Fig. 924), while on Knautia the lateral feeding
channels are much longer (Fig. 923). Examination 'It is generally agreed that the Valerianaceae are
of genitalia (Fig. 925) of specimens reared from allied on the one hand to the Caprifoliaceae, and on
both hosts confirms that the same species is in- the other to the Dipsacaceae, being more advanced
volved. than the former group and more archaic than the
The eight species restricted to Dipsacaceae are a latter' (Cronquist, 1981).
virtually unique assemblage in the high degree of Only three agromyzid genera are known on this
differentiation evident in all four genera repre- small family, and only four species are restricted to
sented. Ophiomyia longilingua can only be associ- it (Table 50). Four of 13 genera in the family have
ated with 0. rostrata (Hendel, d. Spencer, 1976a) been colonized by Agromyzidae, as shown in Table
but unfortunately the host of this species is not 51.
Subclass ASTERIDAE 245
925
,'.}
-.'
923 924
926
927
928
Figs. 923-925. Chromatomyia ramosa: 923, leaf mine on Knautia; 924, leaf mine on Dipsacus (both Hering); 925, aedeagus.
Fig. 926. Liriomyza valerianae: aedeagus.
Fig. 927. Liriomyza valerianellae: aedeagus.
Fig. 928. Chromatomyia kluanensis: aedeagus (Griffiths).
246 Class MAGNOLIOPSIDA (Dicotyledons)
antennal segment being partially blackish in L. vale- The host selection of polyphagous species (Table
rianellae. The male genitalia of the two species 51) and also Liriomyza valerianae which are all
(L. valerianae: Fig. 926 and L. valerianellae: (Fig. found on the three genera Centranthus, Valeriana
927) suggest that they are not directly related, repre- and Valerianella shows that these agromyzids find
senting separate colonizations of Valeriana. Their no essential difference between these genera. This
nearest relatives are not immediately apparent. family is thus more acceptable to polyphagous
Chromatomyia kluanensis Griffiths (1974c) is species than the Caprifoliaceae and Dipsacaceae but
only known in mountains and tundra in Yukon appears to have been colonized twice by ancestral
Territory and British Columbia, Canada forming a Liriomyzas. It is also noteworthy that no species in
long, narrow linear mine on Valeriana capitata and the largest leaf mining genus, Phytomyza, have
V sitchensis. This species is not directly related to succeeded in colonizing the family.
the 13 on Caprifoliaceae but belongs to the Chr.
syngenesiae group (d. Griffiths, 1967a). The male
genitalia (Fig. 928) show that it is an isolated spe-
Table 51
cies, obviously with a long association with Valeria-
na in NW Canada. Agromyzidae on Valerianaceae
Sasakawa (1972) recorded Calycomyza artemi-
siae on Patrinia in Taiwan, although this species is Centranthus Liriomyza strigata, L. valerianae, L. xanthocera
Patrinia Calycomyza sp.
only known as a feeder on Asteraceae. I hesitate to
Valeriana L. strigata, L. valerianae, L. xanthocera;
accept this identification and believe that it is more Chromatomyia horticola, Chr. kluanensis
probable that this represents an undescribed species Valerianella L. strigata, L. valerianae, L. valerianellae,
(see Artemisia, Anthemideae below). L. xanthocera; Chr. horticola
Order ASTERALES
This order consists of the single family Asteraceae. published by Heywood, Harborne and Turner,
Thorne (1983) proposes an identical arrangement 1977) and following detailed discussion of all tribes,
within his Superorder Asteriflorae. the family was left without any formal subfamily
division. This decision was followed by Bremer
(1987) who has produced the most detailed recent
Asteraceae (1300 genera, 21 000 species, after
analysis of tribal relationships in the Asteraceae. He
Mabberley,1987)
takes a strictly cladistic approach and proposes 20
This is by far the largest family of dicots and it is tribal groups which he considers can be accepted as
perhaps not suprising that, with 295 species, it sup- monophyletic. For practical purposes it seems ap-
ports the largest number of agromyzids. propriate to follow this work here in analysing the
There has been substantial disagreement about host selection of the 295 agromyzids known on the
the date of origin of the Asteraceae. Raven and family.
Axelrod (1974) suggested that the family was not There has also been wide lack of agreement about
older than mid-Oligocene, basing this view on the which tribes are primitive within the family. Cyna-
lack of fossil pollen prior to uppermost Oligocene. reae, Heliantheae and Cardueae have each been
However, a much earlier origin is now generally postulated as the most primitive tribes, but in Jef-
postulated, with the family present in the Oligocene frey's arrangement the Lactuceae are treated as the
with similar tribal groupings present as today, indi- most primitive,followed by the Mutisieae.Here, I es-
cating a much earlier origin for the family as a sentially follow the evolutionary progression pro-
whole. A Cretaceous origin now seems to be gen- posed by Bremer, with the Mutisieae-Barnadesiinae
erally agreed. The dramatic increase in fossil pollen treated as the most primitive group and the Coreop-
from the Miocene onwards indicates substantial sideae as the most advanced. The agromyzids on
specific radiation in the later Tertiary and this is each of Bremer's accepted tribes, taxa of near equiv-
reflected by the numerous family-specific leaf alent rank or subtribes are listed in Tables 54 to 80.
miners in the two largest world genera, Phytomyza There is continuing debate about the nearest
with 87 species and Liriomyza with 67. living relatives of the Asteraceae. Stebbins (1977)
Traditionally the Asteraceae have been divided concluded: 'the basic organization of the flower
into two subfamilies, the Asteroideae (Tubuliflorae) favors the same kinds of changes in Campanulaceae
and Cichorioideae (Liguliflorae), with 11 tribes in as have been predominant in Compositae, and
the former, 1 in the latter. Many of the genera are might therefore be taken as evidence for an im-
not clearly defined and the tribal relationships have mediate or not very distant common ancestor of the
been and still are the subject of much debate. two families.' Jeffrey (in Heywood et at., 1977)
Various proposals have been made for a modified suggests that 'one line of evolution led eventually to
subfamily division and this was formalised by Jeffrey the Gentianales-Dipsacales, a second to the Campa-
(1978) with two subfamilies: nulaceae, Calyceraceae and Asteraceae.' On the
other hand, Bremer (1987) stated: 'The Lobelia-
Cichorioideae: Lactuceae, Mutisieae, Vernonieae,
ceae, excluding the Campanulaceae, are considered
Liabeae, Cardueae, Arctotideae
here as the sister group of the Asteraceae.'
Asteroideae: Anthemideae, Astereae, Calenduleae,
With no Agromyzidae known on Lobeliaceae or
Inuleae, Heliantheae, Eupatorieae and Senecio-
the small South American family Calyceraceae, they
neae
can provide no evidence about the relationship of
Although Jeffrey's proposal has not been general- these two families. However, there is strong evid-
ly accepted, essentially the same subfamily division ence, admittedly limited, supporting the close rela-
was adopted by Thorne (1983). Earlier, during the tionship of Campanulaceae and Asteraceae (specifi-
five-day Symposium on the Asteraceae held at cally the Lactuceae), with one species, Ophiomyia
Reading University, England in 1975 (Proceedings heringi feeding on three genera of Campanulaceae
248 Class MAGNOLIOPSIDA (Dicotyledons)
and eight genera of the Lactuceae. Few other cases tionships between the agromyzids on the Asteraceae
are known, apart from polyphagous species, of spe- and on other families are briefly reviewed at the end
cies feeding on two separate families, and when this of the discussion of individual tribes.
does occur it may provide some evidence of the
relationship of the families concerned. The most
Cynareae s. lat.
striking fact emerging from my analysis of the 295
agromyzid species known on the Asteraceae is the Dittrich (in Heywood et a!., 1977) proposed split-
remarkably high degree of tribal specificity, with an ting the traditional tribe Cynareae and this was ac-
average of 81 % for the whole assemblage. This is cepted by Bremer (1987) using the names Carli-
broken down for the 15 genera known on the family neae, Echinopsideae and Cardueae. He cautiously
in the consolidated Table 52 and this also shows the wrote: 'It seems conceivable that all three cynaroid
43 inter-tribal and 11 polyphagous species. Rela- tribes represent monophyletic taxa' and in his treat-
Table 52
Agromyzidae on the Asteraceae
Tribe-specific, inter-tribal and polyphagous species
Tribe-specific • 31 21 2 2 58 17 1. 8 ··i
13 84 241
Inter-tribal 19 2 9 I5 •3 43
Polyphagous 1 l 8 11
Two species of Agromyzidae, on information cur- Echinops Chr. 'atricornis'; Phytomyza bipunctata
rently available, are restricted to this tribe. Melana-
gromyza pseudometallica Spencer (1966a) was de-
Cardueae (66 genera, 2400 species)
scribed from caught specimens on New Ireland.
Later the same species was identified from Burma. This large tribe has been colonized by 11 genera of
During the course of this study I have re-examined Agromyzidae, with 24 tribe-specific species, seven
specimens reared from stems of Carthamus tinctor- which are inter-tribal and four polyphagous. The
ius at New Delhi and the distinctive genitalia are breakdown by genera is shown in Table 55 and all
250 Class MAGNOLIOPSIDA (Dicotyledons)
930
... ; . .
~~
931
929
933
932
/~~,f~7'>
"'-'.;/
934
935 937
Subclass ASTERIDAE 251
Table 56
. -3
~ ...~
..•.....
==~~;....--~ ... .
938 939
940 942
941 943
944 945
947
Figs. 938, 939. Melanagromyza dettmeri: 938, aedeagus, side view; 939, distiphallus, ventral view.
Fig. 940. Melanagromyza lappae: aedeagus.
Fig. 941. Melanagromyza marinensis: aedeagus.
Fig. 942. Melanagromyza cirsiophila: aedeagus.
Fig. 943. Agromyza apfelbecki: aedeagus.
Figs. 944, 945. Liriomyza arctii: 944, aedeagus, side view; 945, same, ventral view.
Fig. 946. Liriomyza centaureae: distiphallus.
Fig. 947. Liriomyza sorar: distiphallus.
Subclass ASTERIDAE 253
sent a synonym of M. aeneoventris. This is a com- cies, with wing length up to 4.2 mm, and mines along
mon species throughout Europe, particularly on the midrib with short offshoots into the leaf blade on
Cirsium. A second species on Cirsium, M. eriolepi- Cirsium, Carduus and Cynara. It occurs primarily in
dis Spencer (d. 1966b) is known only on C. erio- the Mediterranean area where it is invariably pre-
phorum in West Germany. The genitalia of sent on cultivated Cynara scolymus (globe arti-
M. aeneoventris are shown in Fig. 936 and the char- choke), and has also been introduced to Chile with
acteristic posterior spiracles of the puparium in Fig. this plant.
937. It seems probable that M. aeneoventris may This is clearly an isolated and, I suggest, a primi-
also feed on Senecio jacobaea where it was growing tive species, as indicated by its large size, the type of
together with Cirsium vulgare, at Scratch Wood, larval feeding and the unusual genitalia (Fig. 943)
north of London (mistakenly identified as M. dett- which do not directly associate it with any other
meri by Spencer, 1966b). groups in the genus.
M. dettmeri Hering (d. Spencer, 1966b) was de- A single Paraphytomyza, Pa. jaceicaulis (Hering,
scribed from Centaurea jacea in the Netherlands 1960) was reared by Buhr in East Germany from
and the genitalia (Figs. 938, 939) can be seen to be stem-mines on Centaurea jacea. Unfortunately only
distinctively different from those of M. aeneoventris. a single female is known and the affiliations of the
Records of this species given by Spencer (1966b) species cannot be clarified.
from Achillea, Artemisia, Hieracium and Crepis Calycomyza Jlavinotum Frick (d. Spencer and
were inaccurate and have now been identified as Steyskal, 1986) forms large blotch mines on Arc-
hosts of a new species, M. oligophaga sp.n. de- tium in eastern U.S.A. but probably feeds primarily
scribed in the Appendix below. The male genitalia on Eupatorium (see Eupatorieae below).
(Figs. 967, 968) are quite distinct from those of Three Liriomyza species are restricted to this
M. dettmeri, although the adults agree closely on ex- tribe. In Ontario and Minnesota L. arctii Spencer
ternal characters (see Lactuceae below). (1969b) (Figs. 944, 945) forms narrow linear mines
M. lappae (Loew, d. Spencer, 1966b) is a large on Arctium. No host-specific Liriomyza is known
species, restricted to Arctium. The genitalia (Fig. on this genus elsewhere. Two species are known in
940) clearly show the general similarity but also dis- Europe. L. centaureae Hering (d. Spencer, 1976a)
tinctive differences between species in this groups. forms short linear mines on several Centaurea spe-
The species described in Japan as M. lappivora Koi- cies, including C. nigra and C. scabiosa and is com-
zumi on Arctium lappa was found to represent an mon in much of Europe. This species belongs to a
Ophiomyia (see below). small group with long pubescence on the third an-
In the United States M. marinensis Spencer tennal segment and in the larva the posterior spira-
(1981) is only known from Cirsium in California. cles each have an ellipse of eight to nine pores. The
This is one of the largest species in the genus, with genitalia (Fig. 946) suggest that this is a somewhat
wing length in the female up to 4.25 mm. M. cirsio- isolated species. L. soror Hendel (1927) belongs to
phila Spencer (1981) feeds on Cirsium and Car- a different group, with normal antennae, the primi-
duus, it is smaller, the squamal fringe is pale (dark in tive arrangement of three spiracular pores and
marinensis) and the genitalia of the two species forms short linear mines on Cirsium, most common-
show that they are not closely related (M. marinen- lyon C. oleraceum. The male genitalia (Fig. 947) are
sis: Fig. 941, M. cirsiophila: Fig. 942). Neither ap- distinctive but give no indication of its nearest rela-
pears to be directly related to the two Cirsium feed- tives.
ers in Europe. Liriomyza sabaziae Spencer (1973c) is a highly
Ophiomyia curvipalpis (Zetterstedt, cf. Spencer, oligophagous species which was described from
1976a) is an oligophagous stem-miner occurring Sabazia in Venezuela (Heliantheae, see below) and
commonly in Europe and recorded on Centaurea is also known on genera in the Astereae, Helenieae,
but feeding mainly on the Anthemideae and also on Heliantheae and Inuleae. On the Cardueae it has
Aster (see below). In Japan Ophiomyia lappivora been found in California on Carduus, Cirsium and
(Koizumi, cf. Spencer, 1973a) was described in Silybum. The polyphagous L. trifolii is also known
Melanagromyza but the genitalia, larval characters on Centaurea and Cirsium.
and biology indicate its inclusion in Ophiomyia. Its Specimens reared from stems of Centaurea have
host is Arctium lappa, of which the roots are a com- been identified as Napomyza lateralis (d. Spencer,
mercial crop in Japan, the larva feeding from the 1976a). This is a highly oligophagous species re-
base of a leaf stalk, down the stem into the root. The ported on ten genera in seven different tribes. Per-
affiliations of this species are not clear. haps the favourite host is Matricaria (Anthemideae)
Surprisingly, only three Agromyza species are where the species is discussed in greater detail. Grif-
known on the Asteraceae. A. apfelbecki Strobl (d. fiths (1968) has also recorded N. hirticornis (Hen-
Spencer, 1973a) is one of the largest Agromyza spe- del, d. Spencer, 1976a) from Centaurea in Ireland
254 Class MAGNOLIOPSIDA (Dicotyledons)
··-"""rr, •.
~:.J;."~." •..
948 950
949 951
952
,
//
f
954
953
" 955'··
Fig. 948. Phytomyza lappae: aedeagus.
Fig. 949. Phytomyza lappivora: aedeagus.
Figs. 950, 951. Phytomyza cinerea: 950, aedeagus; 951, epandrium.
Figs. 952, 953. Phytomyza spoliata: 952, aedeagus, side view; 953, same, ventral view.
Fig. 954. Phytomyza montana: aedeagus.
Figs. 955, 956. Phytomyza spinaciae: 955, aedeagus, side view; 956, same, ventral view.
Subclass ASTERIDAE 255
and probably near London (see under iasione, firm that this is a typical member of the P. albiceps
Campanulaceae). group.
No Chromatomyia species are known which are In earlier literature a common species forming
restricted to the tribe but most, if not all genera, are distinctive greenish linear mines primarily on Cir-
attacked by either Chr. horticola or Chr. syngene- sium spp. but also on Carduus, Cnicus and Serra-
siae. In many cases only mines are known which are tula was known as P. affinis Fallen, until examina-
here referred to Chr. 'atricornis', the name in wide tion of Fallen's type showed affinis to represent the
use before the species was split by Griffiths (1967a). seed-feeder in Euphrasia (Scrophulariaceae). This
Ten Phytomyza species are apparently restricted Cirsium feeder is now known as P. spinaciae
to this tribe, all in Europe. Two form linear mines on Hendel (cf. Spencer, 1976a), an unfortunate name
Arctium species, most commonly on A. Zappa. but based on examination of the type. Hendel mis-
P. lappae Goureau (cf. Spencer, 1976a) forms un- takenly believed this had been reared from Spinacia
usually long, white mines, frequently with several or but it must have been merely resting on the plant
many in a leaf and even a large leaf can be complete- and in no way otherwise associated with it. The geni-
ly eaten out. The genitalia (Fig. 948) show this spe- talia (Figs. 955, 956) suggest that this species, which
cies to belong to the albiceps group. P. lappivora is very close to P. autumnalis, may also correctly be-
Hendel (1927) has only been found once, near long in Chromatomyia and more detailed studies
Vienna, Austria, on Arctium lappa. Details of the will be necessary to clarify its exact status. The
mine have not been recorded. The genitalia of the puparium remains in the mine, as in P. autumnalis,
only known specimen (Fig. 949) suggest that it is not but it is white, not black as in autumnalis.
closely related to P. lappae and it should be noted Three further species feel on Cirsium. Phytomyza
that, while P. lappivora is largely black, P. lappae is cirsii Hendel (cf. Spencer, 1976a) forms white linear
an unusually pale, yellowish species. mines on several Cirsium spp., pupating externally.
Four species feed on Centaurea. P. cinerea The genitalia (Fig. 957) show that this is a typical
Hendel (cf. Hendel, 1935) forms a communal mine member of the albiceps group and is not directly re-
containing several larvae, always at the apex of the lated to the more common species on Cirsium,
leaf, normally on C. scabiosa, with the puparia re- P. spinaciae. Cirsium heterophyllum, which in
maining loose in the mine. I believe this can be con- southern and central Europe is restricted to moun-
sidered as a primitive species, the dark colour, pro- tains, is more widespread is Scandinavia, and has its
jecting frons, deep jowls and internal pupation sup- own host-specific species. This was earlier known as
port this view. The male genitalia (Fig. 950) confirm P. rydeniana Hering, until this name was found to
that it is an isolated species, and the epandrium represent a synonym of P. albiceps Meigen (cf.
which has two well-defined sclerites lying internally Spencer, 1976a). The albiceps group has been used
(Fig. 951) is unique within the genus. P. spoliata by various authors to denote species related to the
Strobl (cf. Spencer, 1976a) is another isolated spe- feeder on Artemisia, which was originally misidenti-
cies, with the only known host Centaurea calcitrapa fied as P. albiceps but later described by Spencer
in Spain. The larva forms a broad linear mine which (1971b) as P. artemisivora (see below under Anthe-
may develop into an irregular blotch. The male geni- mideae). Fortunately, both species belong to the sa-
talia (Figs. 952, 953) confirm the isolated position me group and the group name can continue to be
of this species. used without any inconsistency. The genitalia of
Griffiths (1959) described Phytomyza autumnalis P. albiceps (Fig. 958), with the characteristic spinu-
from Centaurea nigra, with Onopordum as a less les between the basal sclerites and the paired distal
frequent host. In England the deep, inter-paren- tubules, confirm the close relationship of this spe-
chymal mines are only produced in October/ cies with P. cirsii (Fig. 957)), P. artemisivora (Fig.
November and the shining black puparium remains 1170) and many other species on Asteraceae and
in the mine. Although this species has been treated also on Apiaceae (see below).
as a Phytomyza, the illustration of the genitalia given Mines have been found on Saussurea of an un-
by Griffiths (1959: fig. 1) suggests that it may cor- described species almost certainly belonging to the
rectly belong in Chromatomyia, despite the black albiceps group but unfortunately no adults have yet
puparium. Here it may be mentioned that in been reared (cf. Spencer, 1976a: 346).
Chr. pseudogentii the puparium is also black (see P. continua Hendel (cf. Spencer, 1976a) (= car-
Gentianaceae). dui Hering, 1943) is the only species on the Car-
Groschke (1957) described Phytomyza montana dueae belonging to the robustella group (Griffiths,
from Centaurea montana in Bavaria and it is record- 1964b) which has eight species on the Lactuceae,
ed as common around Stuttgart. The larva forms an five on the Senecioneae and one on the Anthemi-
unusually long mine. The genitalia (Fig. 954) con- deae. Species in this group are large, probably prim-
256 Class MAGNOLIOPSIDA (Dicotyledons)
958
itive, the larvae forming gall-like swellings associa- Bremer (1987) accepts the Barnadesiinae as the
ted with feeding at the base of the midrib, where most primitive group within the family and in his
pupation also takes place. The genitalia are essen- dendogram the Mutisieae s.s. are separated from it
tially similar in this group, having very long basal by the Arctoteae and Cynareae s.1. With only a few
sclerites with no spinules as in the a/biceps group agromyzids involved, it seems convenient here to
and paired tubules of the distiphallus of varying treat the two taxa together.
shape. The genitalia of P. continua (Figs. 959, 960) The agromyzid genera on the Mutisieae are
are characteristic of the group, although being an shown in Table 57 and the species on their host
isolated species on the Cardueae. genera in Table 58.
Table 57
Mutisieae
Agromyzid genera on Mutisieae
Bremer (1987) concluded that the Mutisieae are
paraphyletic, and writes: 'Different parts of the Genus Species
Mutisieae are probably more closely related to other Tribe-specific Inter- Poly-
tribes than to other Mutisieae.' He therefore decided tribal phagous
that it is necessary to divide the tribe into smaller
taxa and accepts a division into four subtribes, Melanagromyza 3
pending the outcome of detailed generic studies cur- Liriomyza 1 (? + 2) 2
Calycomyza (?1)
rently being undertaken. Only two of the subtribes Phytomyza (?1) (undescribed)
are considered here, as no agromyzids are known on
the Nassauviinae or Gochnatiinae. 4 (? + 4) 2
Subclass ASTERIDAE 257
961
963 964
965 966
Figs. 961, 962. Melanagromyza chaptaliae: 961, aedeagus; 962, posterior spiracles of puparium.
Figs. 963, 964. Melanagromyza gerberae: 963, aedeagus, side view; 964, same, ventral view (holotype).
Figs. 965, 966. Melanagromyza gerberivora: 965, aedeagus, side view; 966, same, ventral view (paratype).
Table 58
Petasites and Ligularia are also recorded as hosts of origin of the Lactuceae. The fossil record of the tribe
P. lappae and Sasakawa illustrates the genitalia (fig. is not extensive. Pollen first appears in the Miocene
132d). This appears to represent an undescribed and an early Miocene origin of the tribe has been
species but as the host of the specimens concerned postulated. However, Tomb (in Heywood et al.,
is not indicated proper identification is not possible. 1977) writes: 'The Lactuceae probably arose in the
It should be mentioned that Adenocaulon has early Tertiary in Asia .. .'.
been variously included in the Inuleae and Senecio- It is interesting to note that with only nine agro-
neae but following recent studies has now been re- myzid genera on the Lactuceae, the number of spe-
ferred to the Mutisieae (Mabberley, 1987). cies - 56 - is only surpassed by the Astereae with
Mines of undescribed species are known on Ono- 60. The breakdown by genera is shown in Table 59,
seris and Trixis in Argentina, and in India a Phyto- and the species on the 23 host genera are listed in
myza sp. is recorded as having been reared from Table 60.
Ainsliaea aptera in the Simla Hills, Himachal Pra- It appears that the latex present in the stems has
desh, 2852 m by Beri (1971). This species was mis- acted as an effective deterrent against feeding by the
identified as P. pheUandrii Hering which feeds ex- internal stem-boring species of Melanagromyza. In
clusivelyon Oenanthe (Apiaceae). Europe only a single species (which I have recently
Among polyphagous species only Liriomyza stri- differentiated from M. dettmeri, see Cardueae), is
gata is known on Perezia in Europe and L. trifolii known on the tribe, with hosts in Crepis, Hieracium
on cultivated Gerbera in North and South America, and Picris. Surprisingly, the same species has been
and in Europe. Larvae have recently been found in reared from Artemisia (Anthemideae) at two locali-
Italy mining in the petals (L. Suss, 1988). ties in southern England and I accept Achillea as a
further host. With hosts in two tribes which are not
closely related, this species may well be present on
Lactuceae (70 genera, 2300 species)
other tribes and I therefore describe it as M. oligo-
Traditionally this group has been treated as a sep- phaga sp.n. in the Appendix below (Figs. 967, 968).
arate subfamily Cichorioideae (Liguliflorae), distinct
from the Asteroideae (Tubuliflorae) embracing all Table 59
other tribes. However the most recent studies, par- Agrornyzid genera on Lactuceae
ticularly by Bremer (1987), have shown that the As-
teroideae in this sense are not monophyletic and the Genus Number of species Inter- Poly-
proposal by Jeffrey (1978) to expand the normal restricted to tribe tribal phagous
concept of the Lactuceae to a subfamily Lactucoi-
Melanagromyza 2
deae to embrace six further tribes in addition to the Ophiomyia 9 2
Lactuceae is also not accepted. The Lactuceae as Arnauromyza 1
here discussed are therefore treated as a single Phytoliriomyza 1
monophyletic tribe, characterised by the presence of Liriornyza 13 1 3
ligulate florets, latex and distinctive pollen. How- Caiycomyza 1
Napomyza 1
ever, the latex present in most genera is not confined Chromatomyia 5
to this tribe but is also present in the Arctoteae Phytornyza 13
( Gundelia). The tribe has a mainly north-temperate
distribution. 42 10 4
Botanists are not agreed on the probable date of Total 56
Subclass ASTERIDAE 259
Table 60
968
)
,
967
970 971
972 973
974 975
976 977
Figs. 967, 968. Melanagromyza o/igophaga: 967, aedeagus, side view (ex Crepis); 968, distiphallus, ventral view (ex Artemisia).
Fig. 969. Ophiomyia cunctata: leaf mine on Taraxacum.
Figs. 970, 971. Ophiomyia cichorii: 970, aedeagus, side view; 971, distiphallus, ventral view.
Figs. 972, 973. Ophiomyia chondrillae: 972, aedeagus, side view; 973, same, ventral view.
Fig.974. Ophiomyia beckeri: aedeagus.
Fig.975. Ophiomyia pulicaria: aedeagus.
Fig. 976. Ophiomyia nasuta: aedeagus.
Fig. 977. Ophiomyia pinguis: aedeagus.
Subclass ASTERIDAE 261
In addition, there is a single record of the highly Reichardia and Tragopogon. O. curvipalpis has been
oligophagous species M. splendida Frick on Lactua recorded on Reichardia but occurs most commonly
in Hawaii - it occurs most commonly on genera in on three genera of Anthemideae (see below).
the Heliantheae. It is not clear whether the two species with a
In contrast to Melanagromyza, 11 species of the holarctic distribution - 0. nasuta and pulicaria -
related genus Ophiomyia have between them colo- have dispersed naturally or whether they have been
nized 17 genera, of which seven occur only in introduced to North America. The fact that the
Europe, two only in North America, while three are puparia in each species remain in the leaf at the base
holarctic. The Ophiomyia species can be divided in- of the stalk could clearly facilitate introduction with
to two distinct groups, based on biology, rather than dried plants in hay. However, with both species also
male genitalia. I suggest that group 1 is the more being present in Japan (0. nasuta as madizina
primitive, consisting of 0. beckeri (Hendel, d. Spen- Hendel), 0. nasuta being particularly common in
cer, 1976a), O. cunctata (Hendel, d. Spencer, northern Scandinavia and Finland and the range of
1976a) and 0. pulicaria (Meigen, d. Spencer, 0. pulicaria known to extend from western Europe
1976a), the larvae of these species feeding within to Siberia, natural dispersal seems a distinct possi-
the midrib, with short offshoots into the leaf blade bility. 0. beckeri (Fig. 974) is particularly common
(0. cunctata: Fig. 969). 0. nasuta (Melander, cf. in the Cape area, South Africa on Sonchus, and I re-
Spencer and Steyskal, 1986) on Taraxacum and cently found 0. pulicaria (Fig. 975) on the same
O. pinguis (Fallen, d. Spencer, 1976a) on Cichor- host at Nairobi, Kenya. It seems likely that these two
ium and Leontodon can be closely associated with European species have been introduced relatively
this group, with larval feeding taking place in the recently.
midrib and stalks (0. nasuta) and primarily in the The relationship between o. cymbonoti on Cym-
midrib in 0. pinguis. It is puzzling that the three bonotus in Australia, and 0. nasuta and pinguis is
species beckeri, cunctata and pulicaria appear to discussed above under the Arctoteae. In view of the
have identical larval feeding with many hosts in considerable differentiation in genitalia (0. cym-
common. More detailed biological investigations bonoti: Fig. 930; 0. nasuta: Fig. 976 and 0. pinguis:
may possibly reveal differences which have hitherto Fig. 977), this could not have occurred in the short
been undetected. The six species in group 2 all form time since the arrival of Europeans in Australia, and
the typical shallow mines in the stem immediately a species ancestral to the group must have reached
beneath the epidermis which is characteristic of Australia from the north at an early date. This sup-
most species in the genus. In Europe four species ports the possibility that O. nasuta and 0. pulicaria
belong to this group: 0. curvipalpis (Zetterstedt, d. are also sufficiently old to have reached North
Spencer, 1976a), 0. cichorii Hering (d. Spencer, America via Beringia.
1964c), 0. heringi Stary (d. Spencer, 1964c) and Amauromyza maculosa has been recorded on
0. hieracii Spencer (1964c); and in North America Sonchus and Taraxacum in North and South Amer-
two species: 0. chondrillae Spencer (in Spencer and ica. This species is common in these two Regions
Steyskal, 1986) and 0. coniceps (Malloch, d. and has colonized genera in nine further tribes. It
Spencer, 1969b). appears to favour the Astereae (see below).
Nine of these 11 species occur only on the Lactu- Phytoliriomyza arctica (Lundbeck, d. Spencer,
ceae but most of these are found on many genera 1976a) has been recorded as a stem miner on
(see Table 60). Four species are more restricted in Crepis, Lapsana and Sonchus (Hering, 1957b) but I
their hosts. O. cichorii is only known on Cichorium caught a long series on Solidago at Ottawa, Canada
intybus in eastern France, 0. chondrillae only on and in view of the almost cosmopolitan distribution
Chondrilla juncea in Washington State, U.S.A. The of the species, it seems certain that it has other hosts.
genitalia suggest that these could be sister species PI. arctica was described from Greenland, I caught a
(0. cichorii, Figs. 970, 971 and 0. chondrillae: Figs. specimen just below the snowline of the volcano
972, 973). 0. hieracii is only known on Hieracium Llaime in southern Chile, it has dispersed through-
and Tragopogon in the German Democratic Repub- out Europe and to Sri Lanka and Taiwan, and has
lic, while O. coniceps, which is widespread in Cana- close relatives in Australia and New Zealand. Such a
da and the United States, has only been reared from distribution clearly indicates an ancient species and
Sonchus asper but it seems probable that it may this supports my view that Phytoliriomyza is a primi-
have other hosts. tive genus. Whether Pl. arctica is truly congeneric
0. heringi (Fig. 866) is one of the few species with species present on Hepaticae and ferns (see
which is known on two separate families. It was de- above) will require more detailed studies; at least it
scribed from Campanula in Czechoslovakia, and on represents a separate group within the genus. The
the Lactuceae has been recorded on Crepis, Hier- long, coiled aedeagus (Fig. 978) associates this spe-
acium, Lapsana, Leontodon, Mycelis, Prenanthes, cies with Pl. pilosella which has been found breed-
262 Class MAGNOLIOPSIDA (Dicotyledons)
., ,
,
.. .
-
~ ~
A -::: - - -- -"~,!i... .
978
979 980
981 982
" .~
983 985
ing on Cotula (Anthemideae) on Gomera, Canary The host-specificity of Liriomyza sonchi has been
Islands (von Tschirnhaus). tested experimentally in Canada (Peschken and
Thirteen tribe-specific Liriomyza species are Derby, 1988) for possible introduction of the spe-
known here, of which eight are restricted to a single cies for use in biological control of the weed Son-
genus, L. endiviae Hering (1955) has been reared chus arvensis, which causes crop losses in rapeseed
on Lactuca and Sonchus, L. puella (Meigen, cf. of up to 6 million dollars annually in Alberta, Mani-
Spencer, 1976a) on Lapsana and Prenanthes, toba and Sasakatchewan. Before introduction of
L. scorzonerae on Scorzonera and Lactuca (see L. sonchi from Europe could be approved, its host
below) and L. taraxaci Hering on Arnoseris but specificity on Sonchus had to be established. Adult
most commonly on Taraxacum. flies were caged on a number of hosts in the Lactu-
L. eupatorii (Kaltenbach, cf. Spencer, 1976a) be- ceae and on others in other tribes and in other fami-
longs to a different group and is more widely oligo- lies. The results proved conclusively that L. sonchi
phagous, recorded on Lapsana (but I consider this is restricted to Sonchus. Even Lactuca and Taraxa-
record as doubtful - no mines on Lapsana are pre- cum were virtually resistant to attack and not a
served in Hering's mine collection, 13.vi.88) - and single specimen could be reared from Gerbera,
known on Aster, Eupatorium and Helianthus, and Achillea, Chrysanthemum, Helianthus, Heliopsis,
commonly (according to Hering, 1957b) on Gale- Rudbeckia, Aster, Bellis, Callistephus, Erigeron,
opsis (Lamiaceae). Despite this, I consider that its Solidago, Arctotis, Gaillardia, Helenium, Calendula,
primary hosts are in the Asteraceae rather than the Dimorphotheca, Senecio, Helichrysum, Leontopo-
Lamiaceae in view of the wide occurrence through- dium, Ageratum, Arctium, Carduus, Carlina, Cen-
out the Oriental Region of a closely related species, taurea, Cynara, Stokesia, or from non-composite
L. pusilla, on Crassocephalum (Senecioneae) and hosts tested in the Chenopodiaceae, Cruciferae, Le-
on Tithonia and Xanthium (Heliantheae), see be- guminosae, Solanaceae and Poaceae.
low. This represents the most exhaustive tests of host-
The genitalia of the species known on the Lactu- specificity undertaken with any species of Agromy-
ceae are remarkably similar but differ in the detail of zidae and strikingly illustrates the rigid association
the distiphallus which is most apparent in ventral with a single host-genus in L. sonchi.
view as illustrated below. In some species there are distinctive morphologi-
cal differences and also minor differences in larval
L. andryalae Hering (1927a), host Andryala, Figs. characters, particularly the number and arrange-
979,980 ment of the pores on the posterior spiracles. All spe-
L. aposeridis Beiger (1972c), host Aposeris, Fig. cies (apart from L. eupatorii) form blotch mines of
981 varying shape, normally roundish or slightly more
L. bulgarica Beiger (1979), host Sonchus (Beiger, elongate as in L. sonchi (Fig. 992) but the two spe-
1979: fig. 2) cies on Tragopogon form very different mines. In
L. endiviae Hering (1955), hosts Lactuca, Sonchus, L. groschkei the blotch is confined to the apex of the
?Cichorium, Fig. 982 leaf (Fig. 984), while in L. tragopogonis the larva
L. eupatorii, hosts Lapsana (?), Aster, Helianthus, forms an elongate blotch along the midrib (Fig.
Galeopsis, Figs. 1099, 1100 995).
L. groschkei Spencer (1956), host Tragopogon, I have recently discovered that L. scariolae on
Figs. 983, 984. Lactuca is identical with L. scorzonerae on Scor-
L. hieracii (Kaltenbach, cf. Spencer, 1976a), host zonera and this new synonymy is established in the
Hieracium, Fig. 985 Appendix below. Earlier, L. mycelis Beiger
L. hieracivora Spencer (1971a), host Hieracium, (1960b), described from Mycelis, was found to be
Fig. 986 synonymous with L. puella, known on Lapsana and
L. puella (Meigen, cf. Spencer, 1976a), hosts Lap- Prenanthes (Beiger, 1981). The possibility that
sana, Mycelis, Prenanthes, Fig. 987) others of the seven species now considered to be
L. serriolae Hering (1955) host Lactuca, no male monophagous on a single genus may in fact have a
known wider host range within the tribe cannot be exclud-
L. scorzonerae Ryden (cf. Spencer, 1976a), hosts ed. More collecting will be necessary to clarify this
Scorzonera, Lactuca, Figs. 988, 989 further. Also the status of L. serriolae on Lactuca,
L. sonchi Hendel (cf. Spencer, 1976a), host Son- of which only a single female is known, requires
chus, Figs. 990, 991 further study when males become available.
L. taraxaci Hering (cf. Spencer, 1976a), hosts Ta- The three polyphagous species L. huidobrensis,
raxacum, Arnoseris, Sonchus, Fig. 993 L. strigata and L. trifolii are known on a number of
L. tragopogonis de Meijere (cf. Spencer, 1976a) genera (see Table 60). The Lactuceae are the fa-
(= pusio Mg.), host Tragopogon, Fig. 994 vourite tribe of L. strigata with records on 16 genera
264 Class MAGNOLIOPSIDA (Dicotyledons)
989
993 994
~'.l:ffi~;
~ ,4 ....... .
''® ,', .. ~\
996
997
999 995
Figs. 988, 989. Liriomyza scorzonerae: 988, aedeagus, side view; 989, distiphallus, ventral view.
Figs. 990-992. Liriomyza sonchi: 990, aedeagus, side view; 991, distiphallus, ventral view; 992, leaf mines on Sonchus.
Fig. 993. Liriomyza taraxaci: distiphallus.
Figs. 994, 995. Liriomyza tragopogonis: 994, distiphallus; 995, leaf mine on Tragopogon pratensis.
Figs. 996-998. Liriomyza strigata: 996, aedeagus, side view; 997, distiphallus, ventral view; 998, leaf mine on Pisum.
Fig.999. Calycomyza sonchi: aedeagus.
Subclass ASTERIDAE 265
and Sonchus is possibly the favourite host genus. dimorphic species, with long pubescence on the
The genitalia (Figs. 996, 997) indicate close rela- third antennal segment, which is greatly enlarged in
tionship with L. bryoniae and the North American the female (Figs. 1001, 1002) and the unusually
L. huidobrensis but the larval feeding (Fig. 998), long leaf mines are normally on the lower leaf sur-
with the mine closely associated with the midrib and face. The genitalia (Fig. 1003) confirm that the spe-
lateral veins, suggests unmistakably a sister-group cies belongs to the syngenesiae group. In Poland
relationship with L. huidobrensis, with the common Chr. griffithsiana Beiger (1977) was described from
ancestor crossing Beringia presumably in the late Cicerbita (as Mulgedium) alpina. The genitalia are
Tertiary. typical of Chromatomyia but the adult is distinctive
A single Calycomyza, Cal. sonchi Spencer in the long pubescence on the third antennal seg-
(1969b), is known on this tribe, on both Sonchus ment and the larva forms a long, linear mine on the
and Taraxacum and only recorded in Alberta and lower leaf surface. These two characters are shared
Manitoba, Canada. The genitalia (Fig. 999) indicate with Chr. lactuca in North America but Chr. lactuca
possible relationship with Cal. humeralis (oligo- differs in having the third antennal segment greatly
phagous) or Cal. promissa (Astereae) but in both enlarged in the female (Fig. 1002). These two are
these species the third antennal segment is angulate, clearly sister-species, with dispersal during the Ter-
while in Cal. sonchi it is round. Its exact affiliations tiary between Europe and North America via
are thus uncertain but its presence on the Lactuceae Beringia.
represents a significant switch from a host in an- Two widespread Chromatomyia species, both
other tribe. with an extensive host range occur commonly on the
Two Napomyza species are known on the Lactu- Lactuceae. These were jointly known as Phytomyza
ceae, N. cichorii Spencer (1966a) and N. lateralis atricornis Meigen until Griffiths (1967a) discovered
(Fallen, cf. Spencer, 1976a). N. cichorii Spencer (cf. that two species were represented, with differing
Spencer, 1973a) was described as a serious pest on genitalia and a different host and geographical
cultivated chicory in Belgium and the Netherlands range. The two species cannot be distinguished on
and is also known in Switzerland and England. Ex- external characters or on the form of the mine and
ternally and in the male genitalia (Fig. 1000) there thus in the many cases in which mines have been
are small but distinctive differences between this recorded but without a positive identification from
species and N. lateralis which has a wide host range adults, I refer to these as Chr. 'atricornis'.
in the Asteraceae, occurring on seven different Chr. syngenesiae Hardy (cf. Griffiths, 1967a)
tribes, and recorded in the Lactuceae on Crepis, feeds almost exclusively on Asteraceae and has been
Hypochoeris and Lactuca. It is discussed further be- identified on five tribes in addition to the Lactuceae.
low under Matricaria (Anthemideae) which is per- It has only been positively identified on two hosts in
haps its commonest host. other families - Daucus (Apiaceae) and Pisum (Fa-
Five tribe-specific Chromatomyia species are baceae). It is widespread in Europe but less com-
known here. Chr. aragonensis Griffiths (1967a) is mon in the south, and has been introduced to Aus-
only known in Spain, feeding on Lactuca tenerrima. tralia and New Zealand, and also to both eastern
The genitalia (Griffiths, 1967a; fig. 5) are somewhat and western U.S.A. from where there has only been
similar to Chr. syngenesiae (Fig. 1004) but the adult limited dispersal. The larva forms a whitish linear
is distinguished by exceptionally long pubescence mine and pupates in the leaf. The male genitalia (Fig.
on the third antennal segment. Chr. farfarella 1004) are distinctive of the genus in having a pair of
Hendel (ct. Griffiths, 1967a) feeds on Leontodon 'supporting sclerites' dorsally, above the actual disti-
and Taraxacum and is also generally similar to Chr. phallus. On the Lactuceae this species has been con-
syngenesiae. firmed on Arnoseris, Crepis, Lactuca, Picris, Son-
Chr. ixeridopsis Griffiths (1977) is known only at chus and Taraxacum. The genitalia of Chr. horti-
a high altitude in Alberta and in the Yukon Terri- cola (Goureau, cf. Griffiths, 1967a) are essentially
tory, Canada, feeding on Crepis elegans and similar but the supporting sclerites are significantly
C. nana. The species is well distinguished morpho- expanded apically (Fig. 1005). This species appears
logically from Chr. syngenesiae but belongs to the to have dispersed naturally across temperate areas
same group. The name ixeridopsis is derived from of Africa to South Africa, it is common in parts of
the section Ixeridopsis of Crepis which is apparently India and has reached eastern Asia but is entirely
not monophyletic with Crepis but has not yet been absent from Australia, New Zealand and the U.S.A.
raised to generic rank. Chr. lactuca (Frost, cf. Grif- It is equally, or even more widespread than Chr. syn-
fiths, 1977 and Spencer and Steyskal, 1986) occurs genesiae on the Asteraceae and has been recorded
widely in the northern United States and western on 30 genera in 13 tribes (see below) (Griffiths,
Canada and has been recorded on Crepis, Lactuca, 1967a). On the Lactuceae it is known on Andryala,
Sonchus and Taraxacum. It is a distinctive, sexually- Lactuca, Launaea, Sonchus and Taraxacum. It is
266 Class MAGNOLIOPSIDA (Dicotyledons)
1000
1001
1004
1003
,~ ....
:,"_._;:::;·t~::------ -;>.:
1005 1006
1007 1008
-~
-,-~
~
....• ./ ..
1010
Subclass ASTERIDAE 267
the most polyphagous species in the Agromyzidae, P. japonica is therefore accepted with Hieracium as
comparable only to Liriomyza strigata and L. tri- host with some reserve.
folii, and is known on 35 families in all subclasses, Species in the P. robustella group all feed in the
including also Allium in the monocots. petiole, sometimes causing a gall-like swelling, the
Fourteen Phytomyza species are known on this whitish puparium remaining at the base of the mid-
tribe, six in the albiceps group (P. marginella, rib. Some of the species are very large and the inter-
columbiana, aposeridis, japonica, monori and ta- nal feeding and pupation suggests that this group is
raxaci), and eight in the robustella group. relatively old. Griffiths (1964b) discusses, with illus-
In Europe P. marginella Fallen (d. Spencer, trations of the genitalia, the following six of the eight
1976a and Griffiths, 1977) is known to feed on the species known in this group on the Lactuceae:
following 12 genera: Cicerbita, Crepis, Hieracium,
P. robustella Hendel (1936) (= P. crepidocecis
Lactuca (syn.: Mulgedium), Lapsana, Leontodon,
Hering (1949)). Host Crepis biennis
Mycelis, Picris, Prenanthes, Reichardia, Sonchus and
P. araciocecis Hering (d. Griffiths, 1964b). Host
Taraxacum. Hering, considering in his early work
Crepis paludosa
that each genus had its own species, described
P. cecidonomia Hering (d. Griffiths, 1964b). Host
P. marginella under various names which were
Hypochoeris radicata (Figs. 1008, 1009)
synonymised with P. sonchi R.-D. by Hendel (1935)
P. penicilla Hendel (1936, d. Griffiths, 1964b).
but sonchi was itself later found to represent a syno-
Host Lactuca serriola, genitalia similar to P. ara-
nym of P. marginella. Griffiths (1977) described
ciocecis but adults distinguishable by the long
P. columbiana in Alaska and Alberta with two con-
pubescence of the third antennal segment in this
firmed hosts, Agoseris and Hieracium and probably
specIes
also Taraxacum and considered P. marginella and
P. picridocecis Hering (1957b, d. Griffiths, 1964b).
P. columbiana to be 'geographically vicariant sister-
Host Picris hieracioides (Fig. 1010)
species.' However, the genitalia of the two species
P. wahlgreni Ryden (d. Griffiths, 1964b as P. ta-
(marginella: Fig. 1006 and columbiana: Fig. 1007)
raxacocecis). Host Taraxacum (Figs. 1011, 1012)
are now well differentiated.
Groschke (1957) reared two species from The two further species restricted to this tribe are
Aposeris foetida in the mountains in Bavaria. Both P. crepidis Spencer (1981) in California, host Cre-
are known only from single females but the two spe- pis, male genitalia (Figs. 1013, 1014), and P. rufes-
cies are believed to belong in the albiceps group. In cens Roser (= P. hieracii Hendel, d. Spencer,
P. aposeridis the mine is upper surface and the larva 1976a), host Hieracium, male genitalia (Fig. 1015).
pupates externally, while in P. monori the mine is al- An empty mine on Agoseris glauca in California re-
most exclusively on the lower surface and the pupar- mains unidentified.
ium remains in the mine. Beiger (1973) considered Five species in this group restricted to the Sene-
mines found in S.E. Poland on the same host to rep- cioneae (see below) are closely related to those on
resent P. aposeridis. Hendel (1927, d. 1935) de- the Lactuceae.
scribed P. taraxaci as a largely black species forming
long mines, largely on the lower leaf surface in The most significant fact about the host selection of
Pomerania (now Poland). The unique female holo- the agromyzids on the Lactuceae is the large num-
type has not been seen by any later specialist and ber of species (42) which are restricted to the tribe.
there must be some doubt about the validity of this This is more than on any other tribe. The presence
species. In Japan Sasakawa (1953 and 1961) de- of Melanagromyza oligophaga on both the Lactu-
scribed and discussed P. japonica, with the type ceae and Anthemideae is surprising and may indi-
series bred from Aster (see below under Astereae). cate that this species has an even wider host range.
Later the same species was recorded from genera in
the Anthemideae and Heliantheae, and also from
Vernonieae (70 genera, 100 species)
Hieracium. If these records all do represent the
same species, it has a uniquely wide host range and This large tribe is pantropical in distribution, with
,
1012
.",
\
1011
1014
\. ,I-··'~::-.
ii '<~
1013 ~~, 1015
Figs. 1011, 1012. Phytomyza wahlgreni: 1011, aedeagus, side view; 1012, distiphallus, ventral view.
Figs. 1013, 1014. Phytomyza crepidis: 1013, aedeagus, side view; 1014, same, dorsal view.
Fig. 1015. Phytomyza rufescens: aedeagus.
1016 1018
.
,~ .. /
1017 1019
,'. 1021
?/'::':"
( .
:~:.
"
'
",".
~.
1020 1022
Fig. 1016. Melanagromyza vernoniae: aedeagus.
Fig. 1017. Melanagromyza vernoniana: aedeagus.
Fies. 1018, 1019. Ophiomyia atralis: 1018, aedeagus, side view; 1019, same, ventral view.
Fig. 1020. Calcycomyza dominicensis: aedeagus.
Figs. 1021, 1022. Pseudonapomyza vernoniae: 1021, aedeagus, side view; 1022, same, ventral view.
ly been found to have a wide range from South Afri- The host of Calycomyza dominicensis Spencer (in
ca to Calcutta, Bangkok, New Caledonia, Micro- Spencer and Stegmaier, 1973), which was originally
nesia and Queensland, Australia, presumably in as- described from caught specimens from Dominica,
sociation with its only known host, Vernonia has recently been established on Guadeloupe as
cinerea, on which the larva forms a long, narrow Elephantopus pilosus. The large blotch mines of this
stem mine. The species belongs to the group which species were earlier found empty on E. elatus in
is widespread in the southern hemisphere Old Florida. The male genitalia (Fig. 1020) are very dis-
World tropics, in which the squamae and fringe are tinctive.
white. The male genitalia (Figs. 1018, 1019) are Seguy (1951) described Pseudonapomyza ver-
typical of the genus. noniae (as Phytomyza) from Vernonia appendicu-
270 Class MAGNOLIOPSIDA (Dicotyledons)
1. salicina which Hering considered might be refer- the Inuleae which are normally found on other
able to 0. heringi (feeding normally on Campanula- tribes. L. pusilla (as compositella Spencer, 1961b)
ceae and Asteraceae: Lactuceae). The identity of was recorded on Epaltes in Taiwan (Sasakawa,
this species remains to be established. 1972) but appears to be most frequent on Helian-
Amauromyza maculosa, a highly oligophagous theae (Tithonia, Xanthium) and is also known on
species feeding on nine tribes but favouring genera Senecioneae ( Gynura). It is discussed further below.
in the Astereae (see below), has been recorded on L. sabaziae Spencer (1963c) has been found on
Gnaphalium peregrinum in Florida. In Argentina an Gnaphalium in California but occurs widely on four
Amauromyza sp.n. has been found forming blotch other tribes (see Heliantheae below). In addition, the
mines on Pterocaulon rugosum and has been de- three polyphagous species, L. strigata (Europe),
scribed by Valladares and named after the host L. trifolii (U.S A.) and L. huidobrensis (Brazil) have
genus (unpublished thesis, 1982). The genitalia indi- been found on this tribe (Table 64).
cate close relationship with Am. maculosa but are At the edge of the Ololua Forest, west of Nairobi,
significantly different, with the distal tubules conspi- Kenya I found a mine on Sphaeranthus suaveolens
cuously curving dorsally. which may possibly represent L. trifolii but is more
Nemorimyza posticata, another oligophagous probably of an undescribed species.
species, has been recorded tentatively on Buphthal- Calycomyza platyptera (Thomson) has a wide
mum in Europe (Hering, 1957b) but occurs most host range on six tribes, occurring most commonly
commonly on genera in the Astereae where it is dis- on the Heliantheae (eight genera, see below) and on
cussed further below. the Inuleae has been recorded on Gnaphalium in
Of the Liriomyza species known on the Inuleae, Florida.
five are tribe-specific. L. anaphalidis Sasakawa Chromatomyia is only represented in Europe, by
(1961) is known only in Japan on Anaphalis yedoen- Chr. horticola and Chr. syngenesiae which have
sis, the larva forming an irregular linear-blotch mine been identified on seven genera (see Table 64). In
and pupating in the leaf. No male is known. Australia Chr. syngenesiae has been recorded on
L. helichrysi Spencer (d. Spencer, 1977c) was de- Helichrysum rutidolepis.
scribed from New South Wales and I later dis- Three Phytomyza species have been described on
covered it in Western Australia. The larva forms ir- the Inuleae. P. conyzae Hendel (cf. Spencer, 1976a)
regular linear mines, pupating in the leaf, which is is a common species in the albiceps group (genitalia:
unusual in this genus. The posterior spiracles have Fig. 1032) occurring most frequently on Inula but it
three small, circular pores (puparium: Fig. 1029) has also been recorded on Anaphalis, Buphthal-
which differ from the normal arrangement, where mum, Dittrichia, Pallenis, Pulicaria and Telekia.
the pores are more elongate and one is invariably There are also records in northern Germany and
longer than the other two. L. cassiniae Spencer Denmark on Arnica montana and the species on
(1977c) forms short mines on the bush Cassinia this host was described by Hering (1931b) as P. ar-
quinquefolia, also pupating in the mine. The form of nicophila (see Helenieae below).
pupation, the posterior spiracles of the larva and the Hering described two species with apparent dif-
male genitalia (L. cassiniae: Fig. 1030 and L. heli- ferences in their leaf mines - P. kyffhusana (d.
chrysi: Figs. 1027, 1028) suggest that these are Hering, 1963) on Inula hirta and P. gnaphalii
sister-species. In Natal, South Africa, L. helichrysi- Hering (1963) on Antennaria, Buphthalmum and
vora Spencer (1965b) is only known from the Gnaphalium. These two species are indistinguish-
unique female halo type. In this species pupation able on external characters and after careful con-
takes place externally and there is no evidence that sideration I have decided that only a single species,
the two species on Helichrysum are directly related. P. kyffhusana is involved, with a wide host range on
I found mines of probably the same species on Antennaria, Buphthalmum, Filago, Helichrysum,
H. odoratissimum on the eastern edge of the Inula, Leontopodium and Pulicaria. The formal
Ruwenzori Range, Uganda and on H. foetidum on synonymy of P. gnaphalii with P. kyffhusana is es-
the Aberdares, Kenya. tablished in the Appendix below. The male genitalia
L. craspediae Spencer (1976b) was described (Fig. 1033) show this species to belong to the
from New Zealand on Craspedia uniflora. The geni- P. spondylii group and it is closely related to the pair
talia (Fig. 1031) suggest close relationship with of sister-species P. erigerophila and P. solidaginis
L. wahlenbergiae in New Zealand (see Campanula- (d. Fig. 1084, Astereae).
ceae above) (Fig. 870). The larva of L. craspediae is Leaf mines (Fig. 1034) of a Phytomyza sp. prob-
unusual in having an elongate ellipse of some 20 ably belonging to the P. robustella group were found
pores on each of the two posterior spiracular proc- on Antennaria plantaginifolia in Arkansas, U.SA.
esses. but no adults were reared.
Two Liriomyza species have been recorded on
272 Class MAGNOLIOPSIDA (Dicotyledons)
1023
1025 1026
1027 1028
.<fp1\'jf~~f"
.' :~-:.
1030 1029
:-~'''\
1031 i;'
( 1032
1033 1034
Subclass ASTERIDAE 273
Table 64
Table 66
Agromyzidae on Astereae
acceptable to leaf miners. For example, I carefully (Loew) normally feeds on Eupatorium (see below)
searched on Celmisia on South Island, New Zealand but has been recorded from Heterotheca as M. he-
but found nothing. On the other hand I did find spe- terothecae Spencer (cf. Spencer and Stegmaier,
cies on Olea ria in both New Zealand and Western 1973) in Florida.
Australia. Five described Ophiomyia species are known as
A detailed list by host genera of all species known leaf or stem miners on this tribe. 0. maura (Meigen,
on the Astereae is given in Table 66. cf. Spencer, 1976a) is well known in Europe, form-
Only four Melanagromyza species are known on ing conspicuously long mines only on Solidago (Fig.
the Astereae. The only tribe-specific species is 1037). The genitalia of a specimen from Como, Italy
M. erigeron is Spencer (1963c) reared from stems of are shown in Figs. 1038, 1039. The species has been
Erigeron bonariensis at Sao Paulo, Brazil, The gen- recorded in Japan (Sasakawa, 1961) and the geni-
italia (Figs. 1035, 1036) are distinctive in their talia illustrated by Sasakawa suggest that the J apa-
strong pigmentation. M. gnaphalii was discussed nese and European populations are identical. How-
above (Inuleae) but has also been recorded from ever, in Japan the host range is different, with Aster
Conyza in California. M. minimoides is an oligo- and Erigeron, and also Eupatorium recorded as
phagous seed-feeder occuring most commonly on hosts. Until confirmation is provided from the male
the Heliantheae (see below) but has once been genitalia of specimens from Eupatorium, I feel this
reared from Aster simmondsii in Florida. M. virens host record must be considered questionable and a
Subclass ASTERIDAE 275
separate species may be involved. The name maura on Solidago in Europe and in North America also
has been widely applied to populations in North on Aster and Baccharis, and on Erechtites (Senecio-
America. However, my recent studies (Spencer and neae). This species also forms blotch mines but with
Steyskal, 1986) have shown that the true maura conspicuous feeding lines. The affiliations of Nem.
may be restricted to eastern Canada, where Griffiths posticata have never been fully clarified and it is cer-
has reared specimens from Solidago with genitalia tainly an isolated species. It cannot be directly asso-
apparently identical to those of European maura. ciated with any other genus from the complex geni-
The genitalia of a Canadian specimen (Spencer, talia (Fig. 1047) but a number of characters indicate
1969b: fig. 142) as O. maura (locality either Mani- its relationship with both Phytobia (cambium
toba or Quebec) suggest that this represents a dis- borers) and Amauromyza (stem-borers and leaf-
tinct species. A further species which I have con- miners). Characters common to the group include
firmed from Ontario and from Aster in Arkansas strongly-haired discrete surstyli, large black post-
and Solidago at Ithaca, N.Y. forms shorter, though gonites and strong pre-scutellar bristles. Nem. posti-
similar mines (Fig. 1040) to those of 0. maura and cata has in common with Phytobia the large size and
the genitalia (Figs. 1041, 1042) are quite distinct. silvery lunule. A synapomorphy with Amauromyza
This species was described as 0. quinta Spencer is the bowl-shaped base of the sperm pump (Fig.
(1969b) and see also Spencer and Steyskal (1986: 1048). Unique characters in Nem. posticata are the
257). bristle on the fore-tibia, the loss of the pre-sutural
It thus appears that the host range of 0. maura dorso-central bristle and the yellow colour of the
has expanded from Solidago also to Aster and hind-tergites. I feel it is unlikely that Nem. posticata
Erigeron in the population which has reached has evolved directly from tree-boring Phytobia spe-
Japan, and in North America speciation has oc- cies and more logically has acquired the leaf mining
curred with 0. quinta also feeding on Aster and habit from the intermediate step of stem-boring in
Solidago. More detailed studies on populations herbaceous hosts, which is normal in several groups
from further localities in North America will be of Amauromyza. Its present distribution in Europe,
necessary fully to clarify this complex. Japan and North America, with dispersal south-
A further leaf miner, 0. asterovora Spencer wards to Costa Rica and Colombia, indicates that it
(1969b), is only known on Aster sagittifolius, in is an ancient species. It has retained a number of
southern Ontario, in which the genitalia (Figs. 1043, primitive characters of Phytobia, shares the leaf
1044) are entirely distinct from those of the maura mining habit with some groups in Amauromyza but
group. has developed its own unique characters and male
In Brazil 0. spicatae Spencer (196 3c) forms stem genitalia.
mines on Baccharis spicata. No male is known but A long series of Phytoliriomyza arctica (see Lac-
this belongs to a different group, in which the tuceae above) was caught on Solidago canadensis
squamae and fringe are white. near Ottawa, Canada and I am satisfied that this
In Europe 0. curvipalpis, an oligophagous spe- population was breeding on this host.
cies feeding most commonly on three genera of Sixteen Liriomyza species are known on this
Anthemideae (see below) and also on Reichardia tribe, of which six are tribe-specific, five oligo-
(Lactuceae) has been recorded on Aster. phagous on the Asteraceae and 5 are polyphagous.
Valladares (unpublished report, 1986) records Four species are known on Aster. L. asteris
stem mines of an Ophiomyia sp., probably unde- Hering (1928) has only once been found, in the
scribed, in Argentina on both Conyza bonariensis Harz Mountains, Germany, forming linear mines on
and Hysterionica bakeri. A. amelus. Hering (1957b) distinguished the mine
Amauromyza maculosa (Malloch, cf. Spencer by having a long initial section on the lower leaf sur-
and Steyskal, 1986) which occurs widely from the face. The genitalia (Figs. 1049, 1050), which have
United States throughout South America to Chile is not been previously examined, show that this is an
known on eight tribes but appears to favour the isolated species, with a black sclerite lying within the
Astereae, with records on Baccharis, Conyza, Eri- epandrium on each side near the outer wall.
geron and Solidago. It is a large, distinctive species, L. eupatorii (normally on Eupatorium) is also occa-
with the hal teres largely white but partially black sionally found on Aster. L. pusilla (Meigen, ct.
above. Several larvae frequently feed together to Spencer, 1971b) (= fasciola Meigen) (Figs. 1051,
form a very large blotch mine. In the genitalia the 1052) is best known on Bellis but is recorded by
aedeagus is divided, with two strong tubules (Fig. Hering (195 7b) also on Aster and Solidago.
1045) and the sperm pump is shown in Fig. 1046. L. compositella Spencer (1961b) has been identi-
Nemorimyza posticata (Meigen, ct. Spencer, fied widely in the Oriental Region from India, Sri
1976a and Spencer and Steyskal, 1986) partially Lanka, Thailand and Taiwan, on Callistephus and
overlaps in host range with Am. maculosa feeding Solidago in this tribe but also on seven further gen-
276 Class MAGNOLIOPSIDA (Dicotyledons)
1035 1036
1038
1041
1044
Figs. 1035, 1036. Melanagromyza erigeronis: 1035, aedeagus, side view; 1036, same, ventral view.
Figs. 1037-1039. Ophiomyia maura: 1037, leaf mine on Solidago canadensis (Hering); 1038, aedeagus, side view; 1039, same, ventral
view.
Figs. 1040-1042. Ophiomyia quinta: 1040, leaf mine on Aster; 1041, aedeagus, side view; 1042, same, ventral view.
Figs. 1043, 1044. Ophiomyia asterovora: 1043, aedeagus, side view; 1044, distiphallus, ventral view.
Subclass ASTERIDAE 277
era in five tribes: Bidens (Coreopsideae), Synedrella, from the examination of Meigen's holotype and dis-
Tithonia and Xanthium (Heliantheae), Epaltes covered that it represented the species known at the
(Inuleae), Crassocephalum (Senecioneae) and Ver- time as L. fasciola on Bellis.
nonia (Vernonieae). It should be mentioned here that L. cannabis
In the past differences primarily in the leaf mines Hendel (1935), which was originally considered to
and larvae, differing hosts and different geographi- represent L. eupatorii (Hering, 1921), with Canna-
cal distribution have been the basis for treating these bis sativa as its sole host, also belongs to the pusilla
two species as distinct. However, I am now satisfied complex (see under Urticaceae above).
that a single species is involved representing the true Five New World species have colonized this tribe.
pusilla and this new synonymy is established in the L. togata (Melander, ct. Spencer and Steyskal,
Appendix below. 1986) was established as senior synonym of
In a further species in this complex, the larva of L. douglasii Spencer (1981) which was reared from
L. eupatorii commences the mine with a conspi- Baccharis douglasii in California. The genitalia
cuous spiral of 4/5 coils (Fig. 1101), while in pusilla (Figs. 1055, 1056) clearly place this species in Lirio-
the mine is irregularly linear from the beginning. myza, although it had earlier been included in
L. asterivora (ct. Sasakawa, 1961) appears to be Haplomyza (now Haplopeodes). L. zinniae Spencer
marginally distinct, with the mine beginning with (1981), which is known only in California and
2/3 coils and in the larva the posterior spiracles Washington, feeds primarily on Zinnia (see Helian-
have only 5 pores, as against an ellipse of 8/10 in theae below) and has also been recorded on Hetero-
L. pusilla and up to 12 in L. eupatorii. The genitalia theca.
of L. asterivora (ct. Sasakawa, 1961: fig. 71 d) ap- L. microglossae Spencer (1963c) was described
pear to be close to L. pusilla. In specimens identi- as a leaf miner on Solidago microglossa, Sao Paulo,
fied as L. compositella in India the mine is normally Brazil and has since been reared in Colombia, and
irregularly linear (Fig. 1054) but I have seen a ten- in Argentina was reared from Baccharis salicifolia
dency to a spiral beginning (Fig. 1053) in a popula- (Valladares, unpublished thesis). The male genitalia
tion on Crassocephalum (Senecioneae) in Kerala, (Figs. 1057, 1058), with the distiphallus divided
India. Conversely, the spirals may be less pro- apically, show this to be an isolated species.
nounced in L. eupatorii in Canada, where it is L. baccharidis Spencer (1963c) was described from
known on Aster from Ontario to Alberta. I found Colombia, near Bogota feeding on Baccharis flori-
mines on Eupatorium sp. at Sao Paulo, Brazil in bunda and has been found in Venezuela on the
August, 1957 in which the spirals are very conspi- same host. I found it to be common in central and
cuous but no adults were obtained to confirm the southern California on Baccharis and Conyza, and
identity of this species. also on Artemisia (Anthemideae). L. sabaziae Spen-
It is clear that L. asterivora, L. eupatorii and cer (1963c) was described from Sabazia (Helian-
L. pusilla represent a monophyletic complex and, theae, see below) in Venezuela and has been re-
with its wide distribution in Europe, Asia, Japan and corded on five tribes. Here it is known on Baccharis
North America, it must be of early origin. The ten- in California. It is a paler species than L. bacchari-
dency to a spiral beginning of the leaf mine, some- dis. The genitalia (L. baccharidis: Figs. 1059, 1060)
times detectable in L. pusilla in part of its range suggest that the two species are closely related. An
(India) and also in L. asterivora in Japan but fully empty mine found on Baccharis genistelloides (Fig.
and consistently developed in L. eupatorii is pre- 1061) near Sao Paulo may represent one or other of
sumably a relatively recent evolutionary modifica- these species.
tion. Two species are known as leaf miners on Olearia.
The presence of L. pusilla on Lord Howe Island, In New Zealand L. oleariae Spencer (1976b) was
Australia (Spencer, 1977c) is difficult to explain, described from 0. hectori and 0. virgata, and in
particularly as there are no records on mainland Western Australia I found mines at several localities
Australia. Recent introduction cannot be excluded. on 0. axillaris and described L. oleariana Spencer
A female identified by Sasakawa as L. compositella (1977c). The genitalia show that the two species are
on Biak Island, N.W. New Guinea may possibly rep- not closely related (L. oleariae: Fig. 1062 and
resent a misidentification of L. brassicae. L. oleariana: Fig. 1063). Also L. oleariae pupates
It is important to appreciate the restricted sense externally and there are eight pores on the posterior
of the L. pusilla group as here used, as the name spiracles of the larva, while L. oleariana pupates in-
pusilla was applied in earlier literature to a far wider ternally and the larva has the more primitive ar-
range of species, until initially Hendel (1935) re- rangement of three pores.
stricted the name (erroneously) to the species on The polyphagous species L. bryoniae and L. stri-
Hieracium (now L. hieracii Spencer, see under Lac- gata in Europe and L. huidobrensis, L. sativae and
tuceae). Later I finally clarified the identity of pusilia L. trifolii in the New World have been found on a
278 Class MAGNOLIOPSIDA (Dicotyledons)
( ..,
~.~ 1047
1045
1046 1048
1050
1049
Figs. 1045, 1046. Arnaurornyza rnaculosa: 1045, distiphallus; 1046, sperm pump.
Figs. 1047, 1048. Nernorirnyza posticata: 1047, aedeagus, side view; 1048, sperm pump.
Figs. 1049, 1050. Liriornyza asteris: 1049, aedeagus, side view; 1050, distiphallus, ventral view.
Figs. 1051-1054. Liriornyza pusilla: 1051, aedeagus, side view; 1052, distiphallus, ventral view (both holotype) 1053, leaf mines on
Crassocephalurn crepidioides (India); 1054, leaf mine on Solidago canadensis (India).
Subclass ASTERIDAE 279
1055 1057
1056 1058
1060
1059
1062
1063 1061
Figs. 1055, 1056. Liriomyza togata: 1055, aedeagus, side view; 1056, distiphallus, ventral view.
Figs. 1057, 1058. Liriomyza microglossae: 1057, aedeagus, side view; 1058, distiphallus, ventral view (after Valladares).
Figs. 1059, 1060. Liriomyza baccharidis: 1059, aedeagus, side view; 1060, same, ventral view.
Fig. 1061. Liriomyza sp.: leaf mine on Baccharis genistelloides.
Fig. 1062. Liriomyza oleariae: aedeagus, ventral view.
Fig. 1063. Liriomyza oleariana: aedeagus, ventral view.
280 Class MAGNOLIOPSIDA (Dicotyledons)
/1
/.:/
:/i
Y'
y
'j
1065
1064
1068
1067 1069
-_ .. -
1071 1070
<'~\
•····<,i'~'
1072 j //i';> 1073
Subclass ASTERIDAE 281
number of genera in this tribe (see Table 66). Papua New Guinea and Australia, including Lord
Of the six Calycomyza species known here, four Howe Island.
are tribe-specific. All form blotch mines and pupate C. solidaginis (Kaltenbach, d. Spencer and Steys-
internally in the mine. C. promissa Frick (d. kal, 1986) feeds only on Solidago, both in Europe
Spencer and Steyskal, 1986) feeds on Aster and is and North America. With similar larval characters
widespread from southern Canada to Florida and to C. humeralis, it can be accepted that the two are
California. The genitalia (Figs. 1064, 1065) are typi- sister-species but C. humeralis is distinguishable by
cal of many species in this group but pupation is the angulate third antennal segment. The genitalia
unusual with the puparium attached to a central pile are of the same general form in the two species but
of black frass at an angle of approx. 45°. The only differ in detail (c. solidaginis: Fig. 1071).
known host of C. minor Spencer (in Spencer and Napomyza tripolii Spencer (1966d) was caught at
Stegmaier, 1973) is Conyza canadensis and the spe- the coast in southern England on Aster tripolium
cies is known in Florida and South Carolina, and re- and has been caught on the same plant in Schleswig-
cently also on Guadeloupe (Spencer et al., 1990). It Holstein, northern Germany. It seems reasonably
is the smallest species known in the genus, with wing certain that this is the host, the larva feeding inter-
length of 1.1-1.25 mm. C. exquisita (Spencer, nally in the stem.
1963c) is only known from Sao Paulo, Brazil and Two species are known on Bellis. The oligo-
was reared from a mine formed centrally along the phagous Napomyza lateralis feeds primarily in the
midrib (Fig. 1066), on a bush believed to be Bac- stem and root and has been recorded from B. peren-
charis sp. nis in northern France. Its commonest host is prob-
C. platyptera (Thomson, d. Spencer and Steyskal, ably Matricaria (see Anthemideae below). N. be/li-
1986) has the widest host range of any species in the dis Griffiths (1967b) has only been found in north
genus and is known on seven tribes but most com- London on Bellis, and larval feeding is restricted to
monly on Astereae, where six genera are attacked: the leaves of the basal rosette, mining in the midrib
Aster, Baccharis, Conyza, Grindelia, Heterotheca but with some channels formed in the leaf-blade. Al-
and Solidago. The larva forms a distinctive white though adults in this group are very similar, there
blotch mine at the base of a leaf, with short offshoots are small but distinctive differences in the male geni-
into the leaf blade (Fig. 1067). The genitalia indicate talia (N. tripolii: Fig. 1072 and N. bel/idis: Fig.
that c. platyptera is closely related to C. humeralis. 1073).
Until recently C. platyptera was confused with C. ju- Three Chromatomyia species are only known on
cunda and all records of C. jucunda (apart from the this tribe, on Aster and Erigeron. Chr. asteris
holotype from Wisconsin, host unknown) prior to (Hendel, d. Spencer, 1976a) is host-specific on the
Spencer (1981) refer to c. platyptera, which is a salt-marsh plant, Aster tripolium and occurs widely
more southern species, present in southern states of in Europe in association with its host. The genitalia
the U.S.A. and extending to the Caribbean area. (Fig. 1074) confirm that it belongs to the syngene-
C. humeralis (Roser, d. Spencer and Steyskal, siae group (see under Lactuceae). Two species are
1986) is most common on the Astereae, where it has restricted to Erigeron and were discussed in detail
been recorded on ten genera (Table 66) and is also by Griffiths (1976c). Chr. erigerontophaga (Spencer,
known on the Heliantheae and Madieae. A distinc- 1969b) (Fig. 1075) was described from northern
tive character of the species is the angulate third an- Greenland and is also known on Baffin and Elles-
tennal segment (Fig. 1068) and the larval posterior mere Islands, Canada, Alaska and Labrador and
spiracles are unusual, with up to ten pores, of which mines ascribed to it have been found at high eleva-
four are elongate, finger-like, and the others shorter, tions in the Canadian Rockies on E. debilis. Chr.
more curving. The genitalia are of a form present in thermarum Griffiths (1976c) (Fig. 1076) is known
a number of species found mainly on Asteraceae but only in Alberta and British Columbia on E. phila-
also on Lamiaceae (Figs. 1069, 1070). C. humeralis delphicus. Griffiths refers to these two species as the
has a wide, almost cosmopolitan distribution. Al- erigerontophaga group and the genitalia indicate
though this species was described in Europe, it oc- their close relationship.
curs throughout the New World, in Africa, India to The polyphagous Chr. horticola and also Chr.
Figs. 1064, 1065. Calycomyza promissa: 1064, aedeagus, side view; 1065, distiphallus, ventral view.
Fig. 1066. Calycomyza exquisita: leaf mine on Baccharis sp.
Fig. 1067. Calycomyza platyptera: leaf mine on Bidens sp.
Figs. 1068-1070. Calycomyza humeralis: 1068, third antennal segment; 1069, aedeagus, side view; 1070, same, ventral view.
Fig. 1071. Calycomyza solidaginis: distiphallus, ventral view.
Fig. 1072. Napomyza tripolii: aedeagus.
Fig. 1073. Napomyza bellidis: aedeagus.
282 Class MAGNOLIOPSIDA (Dicotyledons)
,~ ......................•.
1075
1074
~ ...
1077
1076
1079 1081
~ "''''''
." Y
..
1082 1083
Subclass ASTERIDAE 283
syngenesiae have been recorded on a number of single large blotch mine. Male genitalia (Fig.
genera (Table 66) but never on Solidago, either on 1082) little differentiated from P. astotinensis
the native European S. virga urea or the introduced (Solidago), P. peregrini (Erigeron), P. scopulina
North American S. canadensis. (Erigeron and Solidago), the four treated by Grif-
The 18 Phytomyza species known on the As- fiths as the P. phalangites superspecies.
tereae represent a remarkable case of radiation, all P. pieninica Nowakowski (1963). Host Aster alpi-
but one being restricted to one or other of the four nus, found once in Carpathian Mountains (747
genera, Aster, Bellis, Erigeron or Solidago, and all m), Poland. Larva forms linear blotch mine,
belonging to the albiceps group (see above under pupating internally. Nowakowski gives detailed
Cardueae). comparison of this species with P. conyzae (hosts
exclusively in Inuleae) but, despite superficial re-
Species on Aster semblance of the two, I suggest they are not close-
ly related. Nowakowski further suggests that
P. anserimontis Griffiths (1976c). Host Aster
P. pieninica appears to be the sister-species of
modestus, Alberta. The male genitalia (Fig. 1077)
P. asteris. This is fallacious, as asteris is a Chro-
with the very short distiphallus suggest close rela-
matomyia (see above) with entirely different geni-
tionship with P. solidaginophaga on Solidago.
talia, despite the superficial resemblance of the
P. asterophaga Spencer (1969b). Only known host
two on external characters, and is a striking exam-
Aster conspicuus, Alberta, North Carolina. Male
ple of the significance of male genitalia when as-
genitalia (Fig. 1078) characteristic of albiceps
sessing relationships. The male genitalia, with 2
group.
rows of spinules between the basal sclerites, con-
P. ciliolati Spencer (1969b). Aster spp. (not conspi-
firm that this species belongs to the albiceps
cuus), Alberta. Mine (Fig. 1079) readily distin-
group.
guishable from that of P. asterophaga by the more
P. solidaginivora Spencer (1969b). Host consid-
conspicuous line of black frass. Male genitalia il-
ered by Griffiths (1976c) to be Aster modestus,
lustrated by Griffiths (1976c).
not Solidago sp. as originally identified, Alberta.
P. despinosa Griffiths (1976c). Only known host
Male genitalia (Fig. 1083), completely spinules
A. sibiricus, Alberta, Yukon Territory. Male gen-
but belonging to the P. albiceps group.
italia without spinules (Fig. 1080).
P. erigerophila Hering (as P. asteribia Hering) (cf.
Species on Bellis
Spencer, 1976a). Host Aster amellus, but normal-
ly Erigeron (see below). P. bellidina Hering (cf. Griffiths, 1976c). Host Bel-
P. hoppi Hering (cf. Griffiths, 1976c). Only known lis, not uncommon in Mediterranean area, iden-
host A. bellidiastrum, mountains of central and tity of empty mines in Germany and England re-
eastern Europe. Male genitalia without spinules. mains uncertain. Genitalia of albiceps type but
Mine linear, widening at end (Fig. 1081). entirely without spinules.
P. japonica Sasakawa (1953 and 1961). Hosts Aster P. japonica Sasakawa - see above under Aster.
(original description), later Artemisia, Bellis, P. solidaginis Hendel- see below under Solidago.
Chrysanthemum, Hieracium, Rudbeckia, only P. virgaureae Hering - see below under Solidago.
known in Japan. Larva forming linear mine,
greatly widening towards end. The genitalia are Species on Erigeron
typical of the albiceps group, though without
P. erigerophila Hering (cf. Spencer, 1976a and Grif-
spinules. If the hosts cited by Sasakawa are ac-
fiths, 1976c). Host Erigeron acris (Europe, Al-
curate, this species has an unusually wide host
berta), E. glabellus (Alberta). Adult with third an-
range and confirmation from male genitalia
tennal segment enlarged, male genitalia (Fig.
seems desirable.
1084) with left basal sclerite greatly reduced,
P. phalangites Griffiths (1976c). Host Aster, Alber-
spinules strong, distiphallus membranous (not
ta. Larvae frequently feed communally to form
seen in Spencer, 1976a: fig. 721). The genitalia
Fig. 1074. Chromatomyia asteris: aedeagus.
Fig. 1075. Chromatomyia erigerontophaga: aedeagus (Griffiths).
Fig. 1076. Chromatomyia thermarum: aedeagus (Griffiths).
Fig. 1077. Phytomyza anserimontis: aedeagus (Griffiths).
Fig. 1078. Phytomyza asterophaga: aedeagus (Griffiths).
Fig. 1079. Phytomyza ciliolati: leaf mine on Aster sp.
Fig. 1080. Phytomyza despinosa: aedeagus (Griffiths).
Fig. 1081. Phytomyza hoppi: leaf mine on Aster bellidiastrum (Hering).
Fig. 1082. Phytomyza phalangites: aedeagus (Griffiths).
Fig. 1083. Phytomyza solidaginivora: aedeagus.
284 Class MAGNOLIOPSIDA (Dicotyledons)
show that this is the sister species of P. solidaginis spp., Alberta, larva forming linear mine. A typical
on Solidago. member of the P. albiceps group, male genitalia
P. ovimontis Griffiths (1976c). Host E. caespitosus, (Fig. 1085).
Yukon Territory (only record). This species P. solidaginis Hendel (cf. Spencer, 1976a and Grif-
agrees in all essential characters (adult, genitalia, fiths, 1976c). Host Solidago virgaurea, Europe, in
larva) with P. demissa Spencer on Artemisia (An- Poland also Bellis. Larva forming linear mine.
themideae, see below) and is possibly synony- Male genitalia (Fig. 1086) indicate close rela-
mous, despite the hosts in different tribes. tionship with P. erigerophila (Fig. 1084).
P. peregrini Griffiths (1976c). Host E. peregrinus, P. virgaureae Hering (cf. Spencer, 1976a and Grif-
Alberta, possibly also Alaska. Belonging to fiths, 1976c). Hosts Solidago virgaurea and Bellis
P. phalangites superspecies (see above), differ- perennis, widespread in Europe. Larva forming
ences from P. phalangites and P. scopulina slight. linear mine, male genitalia (Fig. 1087) character-
P. scopulina Griffiths (1976c). Hosts Erigeron, Soli- istics of a/biceps group.
dago, Alberta, British Columbia, Yukon Terri-
With ten agromyzid genera represented on the
tory. A member of the P. phalangites super-
Astereae, the number of species, 60, is the largest on
species.
any tribe, followed closely by 56 on the Lactuceae.
Of the 36 tribe-specific species all but three are
Species on Solidago
present in the northern hemisphere, with one each
P. astotinensis Griffiths (1976c). Host Solidago spp., in South America, Australia and New Zealand. The
Alberta. Larva forming narrow linear mine. Male striking radiation in Phytomyza with 17 tribe-specif-
genitalia essentially as in P. phalangites (Fig. ic species has occurred mainly in North America
1082). and among these just three, P. erigerophila, solida-
P. scopulina Griffiths (1976c). Hosts Solidago and ginis and virgaureae have a holarctic distribution.
Erigeron, Alberta, British Columbia, Yukon The discovery that Liriomyza compositella, with
Territory. Larva forming linear-blotch mine. A its wide range from India to Thailand, Hong Kong
member of the P. phalangites superspecies. and Taiwan and south to Sri Lanka, Papua New
P. solidaginophaga Sehgal (1971b). Host Solidago Guinea and Lord Howe Island, is identical with the
/\)"
I ••
, .
.", .,
',' I
,. ,
,,
'.
1084
I
\
I
-,...
I ,
I I ~ \
I I \ \
I \: I
,I '
"
,\
'
true L. pusilla hitherto known only in Europe, is in- lection with Robinson's 19 groups. Surprisingly,
teresting and surprising. However, this range is very only eight of the 160 genera available as hosts are
comparable to that of Chromatomyia hortieola, known to be colonized by Agromyzidae (Table 68).
known originally as a temperate species in Europe Six Melanagromyza species are known here, two
but now known to be widespread in Africa and being host-specific on Eupatorium, four inter-tribal.
across India and Asia to Japan. M. eupatoriella Spencer (in Spencer and Stegmaier,
Although 36 species are restricted to the As- 1973) has the mesonotum black and abdomen green
tereae, the 16 further species which are oligo- and is only known from stems of Eupatorium sp. in
phagous, occurring on one or more tribes, indicates Trinidad. M. malloehi (Hendel, ct. Spencer and
that the tribal boundaries can be crossed with rela- Stegmaier, 1973) is very similar and is only known
tive ease. from Puerto Rico from stems of E. odoratum but
the genitalia of the two species are entirely distinct
(M. eupatoriella: Fig. 1088; M. malloehi: Fig. 1089).
Eupatorieae (160 genera, 2000 species) In Europe M. eupatorii Spencer (ct. Spencer,
Bremer (1987) writes that the 'circumscription of 1966b) was described from E. eannabinum and was
the Eupatorieae presents few problems.' Mabberley subsequently reared from Inula eonyza (Inuleae),
(1987) however places the tribe in the subfamily Leueanthemum vulgare (Anthemideae) and Senecio
Lactucoideae (after Jeffrey) but adds 'referred to spp. The male genitalia are little differentiated from
Asteroideae by some authors.' This lack of agree- related species but the posterior spiracles of the lar-
ment on the position within the family of a major, va (Fig. 1090) with an ellipse of 12 small pores and
well-defined tribe illustrates the difficulties botanists the small central horn (contrast M. aeneoventris,
are having in reaching a consensus on the classifica- Fig. 937) fully distinguish the species. M. virens
tion of the Asteraceae. The tribe is best represented (Loew, ct. Spencer and Stegmaier, 1973) has been
in the New World. widely recorded in North American literature but
The Eupatorieae have been divided into four sub- the only hosts I accept are Eupatorium and Heter-
tribes but, following recent and ongoing studies, otheca (Astereae). M. virens may be confused with
such a division is considered as 'highly artificial' and other greenish, stem-boring species. However, the
has been replaced 'with a series of 19 groups many orbits are unusually broad and in the male genitalia
of which are potential subtribes .... ' (Robinson and the surstyli bear three or more distinctive spines
King in Heywood et aI., 1977). (Fig. 1091). Reared specimens are only known in
A relatively late derivation of the Eupatorieae Florida. M. minima (Malloch, ct. Spencer and
from the Astereae in the Middle Miocene has been Steyskal, 1986) appears to have as its main host
suggested, while other workers considered that the Wedelia spp. (Heliantheae, see below) but Eupa-
Eupatorieae were closest to the Vernonieae. Bremer torium odoratum is certainly an additional host, on
(1987) supports the former relationship. which I found adults on the flowers in Jamaica.
In view of the small number of Agromyzidae so M. metalliea (Thomson, ct. Spencer, 1985a) is a
far recorded on the tribe - eight genera, with 27 spe- widespread species in the Old World tropics from
cies, of which 13 are tribe-specific (Table 67), it has Africa, across Asia to northern Australia, with Ager-
seemed pointless to attempt to correlate the host se- atum and Bidens (Coreopsideae) as confirmed
hosts but there are almost certainly others in further
tribes. It is one of a large group of shining greenish
Table 67
species which can only be identified by the male
genitalia (Figs. 1092, 1093). The posterior spiracles
Agromyzid genera on Eupatorieae of the larva (and puparium) are typical of the genus
but the central horn present in most species (cf.
Genus Species M. angeliciphaga: Fig. 585) is here atrophied to a
Tribe-specific Inter- Poly- mere scar.
tribal phagous Frost (1924: 42) records as Agromyza eurvipalpis
var. texana Malloch a species reared from Eupator-
Melanagromyza 2 4
Ophiomyia 1 (?)
ium urtieaefolium, which Hendel (1931) treats as
Tropicomyia 2 Ophiomyia maura. I consider that it is highly im-
Amauromyza probable that the species represents 0. maura
Liriomyza 4 (1 undescribed) 1 2 which feeds on Aster and Solidago and its identity
Calycomyza 5 (1 undescribed) 3 must remain in doubt.
Chromatomyia
Phytomyza
Two polyphagous Tropicomyia spp. are known
on this tribe, both on Eupatorium. I found empty
13 9 5 mines of T. polyphaga on E. odoratum in Sri Lanka
286 Class MAGNOLIOPSIDA (Dicotyledons)
"~2~~~~
1089
/
..... ~'. ~ ... ,,- ., "
"
'":~:~«(
1092
1093
1096
1097 1098
Subclass ASTERIDAE 287
and Sasakawa (1988) recorded T. atomella on the folii are known on Ageratum and Eupatorium
same host in Malaysia. (Table 68).
The highly oligophagous Amauromyza maculosa Eight Calycomyza species are known on the
which favours genera in the Astereae (see above) Eupatorieae, which is more than on any other tribe.
has been found in Florida in the Miami area mining Of these five are tribe-specific. Species in this genus
E. odoratum. cannot normally be identified on external characters
Four Liriomyza species are restricted to the but the host will frequently give a good indication of
Eupatorieae, of which three feed on Mikania. L. mi- the species. The form of the mine may occasionally
kaniovora Spencer (1973c) was described from Ve- have some specific characters but a positive identifi-
nezuela on M. cordifolia, the larva forming a long, cation will normally only be possible from the male
narrow mine (Fig. 1094). No male was known when g~nitalia, which are more differentiated than in spe-
I first discovered the species but it has subsequently C1es on the Astereae or the Lamiaceae.
been found in Argentina (Valladares) and the male
genitalia will be illustrated in due course. Remark-
C. artemisiae (Kaltenbach, ct. Spencer, 1976a). This
species was originally described from Artemisia
ably similar mines were found in Kenya near N ai-
in Europe (see under Anthemideae below). In
robi on M. cordata (Fig. 1095) but no adults are
Europe mines are also regularly found on
known and it will be of great interest to discover the
Eupatorium cannabinum, while in the United
affiliations of this species when a male can be ob-
States known hosts include Ageratina altissima,
tained. A certainly unrelated species, L. prostrata
Eupatorium odoratum and Artemisia douglasia-
(Sasakawa, 1963c) (= L. mikaniae Spencer (1977a),
na. The male genitalia are highly asymmetric and
synonymy established by Spencer, 1989) is known
thus with every 10° of inclination can appear very
in Papua New Guinea, also on M. cordata, the larva
different. With records from China, Colombia
feeding in a spiral and producing a large secondary
and Japan and in view of the variation detectable
blotch mine (Fig. 1098). The species is unusual in
in these populations, I now feel it is not certain
having the scutellum entirely black (normally yellow
whether this represents incipient or possibly even
in this genus). The male genitalia (Figs. 1096, 1097)
full speciation. More detailed studies are necessa-
and also the larval feeding conclusively indicate the
ry to clarify this interesting species.
relationship of this species with L. eupatorii (Kal-
C. eupatorivora Spencer (in Spencer and Stegmaier,
tenbach, ct. Spencer, 1976a) (Figs. 1099, 1100)
1973). Hosts Eupatorium, Alomia. Jamaica,
which feeds primarily on Eupatorium in both
Guadeloupe, Venezuela, Brazil. This species was
Europe and North America, and also on Galeopsis
originally misidentified as C. flavinotum, until it
(Lamiaceae) in Europe. The leaf mine starts with a
was found from the distinctive genitalia (Figs.
spira~ but then the larval feeding instinct changes
and 1t forms an irregular linear mine (Fig. 1101). 1104, 1105) that three species are represented on
Eupatorium.
L. eupatoriana Spencer (1954a) is known only on
C. flavinotum .(Frick, ct. Spencer and Steyskal,
Eupatorium cannabinum in Europe and here the
1986). Conf1rmed hosts Eupatorium maculatum,
mine is always largely confined to the apex of the
leaf, being essentially linear but in the confined area E. rugosum; Eupatoriadelphus maculatus; prob-
which is mined forming a secondary blotch (Fig. ably Ageratina; also Arctium lappa, A. minus
(Cardueae), not Viburnum pubescens (Caprifolia-
1102). The genitalia (Fig. 1103) show that this spe-
cies is not directly related to L. eupatorii. In Argen- ceae), as recorded by Frick (1956). Not uncom-
tina an undescribed Liriomyza has been found mon from New York to Minnesota, also Ontario.
forming linear mines on Stevia achalensis similar to Although this species cannot be reliably distin-
those of L. brassicae or L. sativae but pupating in guished from C. eupatorivora, its genitalia (Fig.
the mine. The genitalia (to be illustrated by Valla- 1106) are quite distinct. In fact, in side view the
dares in due course) confirm that this represents a aedeagus closely resembles that of C. frickiana
(Fig. 1225), suggesting that these two species are
distinct species.
The polyphagous Liriomyza strigata and L. tri- closely related, although in ventral view the
aedeagus is highly differentiated (Figs. 1106
1226). '
Fig. 1088. Melanagromyza eupatoriella: aedeagus.
Fig. 1089. Melanagromyza mallochi aedeagus.
Fig. 1090. M elanagromyza eupatorii: posterior spiracles of larva.
Fig. 1091. Melanagromyza virens: surstylus.
F~gs. 1092, 1093. Melanagromyza metallica: 1092, aedeagus, side view; 1093, same, ventral view.
FIg. 1094. Liriomyza mikaniovora: leaf mine on Mikania cordifolia.
Fig. 1095. Liriomyza sp.: leaf mine on Mikania cordata (Kenya).
Figs. 1096-1098. Liriomyza prostrata: 1096, aedeagus, side view; 1097, same, ventral view; 1098, leaf mine on Mikania cordata.
288 Class MAGNOLIOPSIDA (Dicotyledons)
Ii:Ij~
1100
1099
1104
1105
1106 1107
1108 1109
Figs. 1099-1101. Liriomyza eupatorii: 1099, aedeagus, side view; 1100, distiphallus, ventral view; 1101, leaf mine on Eupatorium
cannabinum.
Figs. 1102, 1103. Liriomyza eupatoriana: 1102, leaf mine on Eupatorium cannabinum; 1103, distiphallus.
Figs. 1104, 1105. Calycomyza eupatorivora: 1104, aedeagus, side view; 1105, distiphallus, ventral view.
Fig. 1106. Calycomyzaf!avinotum: distiphallus, ventral view.
Fig. 1107. Calycomyza mikaniae: aedeagus.
Figs. 1108, 1109. Calycomyza servilis: 1108, aedeagus, side view; 1109, distiphallus, ventral view.
Subclass ASTERIDAE 289
Table 68
e. mikaniae Spencer (in Spencer and Stegmaier, It is puzzling that so few species are known on the
1973). Host Mikania spp. Described from Flor- Eupatorieae and the explanation is probably in part
ida, subsequently found in J amaica, Venezuela, due to the limited collecting in the New World
Guadeloupe. Male genitalia (Fig. 1107). where the tribe is best represented. Two aspects of
e. platyptera (Thomson). Highly oligophagous, the host selection of the Agromyzidae are particu-
known here only on Eupatorium, most common- larly noteworthy - the radiation of Calycomyza spe-
lyon Heliantheae (see below). cies, with five specific to the tribe and three further
e. servilis Spencer (in Spencer and Stegmaier, oligophagous species, and the presence of only a
1973). Host Eupatorium odoratum, Trinidad, single Phytomyza, the largest world genus which is
Dominica, Argentina (Cordoba). Larva forms represented on the Astereae with 17 species and on
small blotch mine. Male genitalia (Figs. 1108, the Senecioneae with 15. There appears to be no ob-
1109) suggesting relationship with C. artemisiae vious reason for the unattractiveness of the Eupa-
group. torieae to Phytomyza and the partially random na-
e. steviae Spencer (1973c). Host Stevia elatior, Ve- ture of host selection is presumably the explanation.
nezuela. The genitalia (Figs. 1110, 1111) confirm
that this species belongs to the artemisiae group.
Calenduleae (7 genera, 110 species)
Valladares (unpublished report, 1986) has found
a species on S. effusa in Argentina which she pro- This small tribe is represented primarily in the Me-
poses to describe as distinct from C. steviae. diterranean area and South Africa. It is generally
Mines I found on Pique ria trinervia in Costa Rica agreed to be closely related to the Senecioneae. On-
probably represent C. steviae. ly five agromyzid genera, with ten species, have
In Chromatomyia only Chr. horticola has been
positively identified, on Ageratum in India. Records Table 69
on Eupatorium in Europe can only be treated as
Agromyzid genera on Calenduleae
Chr. 'atricornis'.
Surprisingly only a single Phytomyza has been Genus Species
described on the Eupatorieae, P. eupatorii Hendel Inter-tribal Polyphagous
(cf. Hendel, 1935) which is strictly host-specific on
Melanagromyza 2
Eupatorium cannabinum in Europe and on Eupa-
Amauromyza 1
torium sp. in Japan. The larva forms an irregular Liriomyza 2 3
linear mine and the genitalia (Fig. 1112) show that Napomyza 1
this is a typical member of the albiceps group. A Chromatomyia 1 (or 2)
mine on Ageratina altissima (Fig. 1113) in Cali-
fornia probably represents an undescribed Phyto- 6 4 (or 5)
myza sp. TotallO (or 11)
290 Class MAGNOLIOPSIDA (Dicotyledons)
1110 1111
- - - - - . --
..
",'
fo,',
---....
"''',\ ~",
'\;, "---:'-
"\ -........
~'-'.," "
1112 1113
1114
1115
Figs. 1110, 1111. Calycomyza steviae: 1110, aedeagus, side view; 1111, distiphallus, ventral view.
Fig. 1112. Phytomyza eupatorii: aedeagus.
Fig. 1113. Phytomyza sp.: leaf mine on Ageratina altissima.
Figs. 1114, 1115. Liriomyza spencerella: 1114, aedeagus, side view; 1115, same, ventral view.
Table 70
Agromyzidae on Calenduleae
colonized the tribe (Table 69) and Agromyzidae are Senecioneae (100 genera, 2000 species)
only known on three host genera (Table 70). Sur-
Botanists are still working on the classification of
prisingly, no species are known which are restricted
this tribe but here I follow Bremer (1987). Arnica,
to the tribe. previously included in the Senecioneae, has now
Two Melanagromyza species, which are present
been assigned to the Helenieae (Nordenstam in
in Florida and/or California and are widespread in
Heywood et al., 1977). None of the problematic
the Caribbean area, are known on Calendula.
genera in the tribe are known to be colonized by
M. floris Spencer (1963f) feeds in seeds of Calendu-
Agromyzidae and need no special comment here.
la but the main host is believed to be Bidens (Co-
The tribe is cosmopolitan and tribe-specific agro-
reopsideae, see below), while M. viridis (Frost, cf.
myzids are known from Europe, North America,
Spencer and Steyskal, 1986) has also been recorded
Africa, Australia and New Zealand.
in flower-heads of Calendula but the main hosts are Nine genera of agromyzids are known on the
in four genera of the Heliantheae (see below). Senecioneae, as shown in Table 71 and only 16 gen-
Five Liriomyza species are known on Calendula. era of the tribe support leaf-miners or stem borers
L. zinniae Spencer (1981), known only in California (Table 72).
and Washington is oligophagous on Calendula offi- Of 11 Melanagromyza species known to feed in-
cinalis and Heterotheca (Astereae) and Zinnia ternally in stems or flower-heads of genera in this
(Heliantheae, see below). An interesting oligo- tribe, Europe, Australia and New Zealand each has
phagous species, L. spencerella Valladares (1984), one three are known in North America, four in
was described in Argentina. The designated holo- Africa and one in Madagascar. I once reared speci-
type was reared from Calendula and it therefore mens from stems of Senecio jacobaea at Scratch
seems appropriate to discuss the species here but Wood, north of London which were originally iden-
other hosts known are Bidens (Coreopsideae), tified as M. dettmeri (Spencer, 1966b) but I am now
Chrysanthemum (Anthemideae) and Helianthus satisfied that the species was M. aeneoventris which
(Heliantheae). The genitalia (Figs. 1114, 1115) normally feeds on Cirsium (Cardueae, see above).
clearly show that this species belongs to the L. pusil- In Florida M. wedeliae was described from stems
la group. It is the only species in this group, apart and flower-heads of Wedelia paludosa (see under
from the holarctic L. eupatorii, which is known in Heliantheae) and also Senecio confusus. It is impos-
the New World. It is also of interest, being the only sible at this stage to say which is the primary host
species in the group in which the larva pupates inter- but as the holotype was from Wedelia, the species is
nally, at the end of the mine on the lower surface. discussed under this genus in the Heliantheae.
The oligophagous Amauromyza maculosa (see Five Melanagromyza species feed on Senecio in
Astereae above) has been recorded in the United Africa. M. seneciocaulis Spencer (1960a) was
States on Calendula. reared from stems of S. juniperinus and S. ruderalis
Three polyphagous Liriomyza species are also in South Africa (East London and Pretoria). The
known on Calendula, L. strigata in Europe, L. hui-
dobrensis in Chile and L. trifolii in Europe and
North America. Table 71
Napomyza lateralis (Fallen) can occur in pest pro-
portions in stems of cultivated Calendula an~ is als.o Agromyzid genera on Senecioneae
known to attack Dimorphotheca and SeneclO. ThIS
Genus Species
highly oligophagous species is discussed in more
detail in the Anthemideae below, which include Restricted Inter-tribal Polyphagous
to tribe
what is probably its favourite host, Matricaria.
The Chromatomyia sp. on Calendula in Europe Melanagromyza 7 4
has not yet been positively identified and can at the Ophiomyia 1
moment only be referred to as Chr. 'atricornis'. In Amauromyza
Kenya Chr. horticola is known on Chrysanthemo- Nemorimyza 1
Liriomyza 5 2 2
ides. Calycomyza 2
Calendula seems fully acceptable to oligophagous Napomyza 1
species known on other tribes and also to poly- Chromatomyia 3 1
phagous species. It is interesting though that no Phy- Phytomyza 16
tomyza species has colonized it, although the genus
32 12 3
is so well represented on neighbouring tribes (As-
tereae, 17 species; Senecioneae, 15). Total 47
292 Class MAGNOLIOPSIDA (Dicotyledons)
-. ~
1118
1116
~:J.~~~;:;6. ,,:.,:::.~
_1IIiI!"!."...,_.•:....
----- .. ~.:::-
1119
1117
.......-.. _-;.'.'
//
r:.";~"':'.ifl;~; . .... "
1120
1121
1122 1123
1124
Figs. 1116, 1117. Melanagromyza compositana: 1116, aedeagus, side view; 1117, same, ventral view.
Fig. 1118. Melanagromyza gynurae: aedeagus.
Figs. 1119, 1120. Melanagromyza heatoni: 1119, aedeagus, side view; 1120, same, ventral view.
Fig. 1121. Melanagromyza ruandae: aedeagus.
Fig. 1122. Melanagromyza senecionella: aedeagus.
Fig. 1123. Melanagromyza seneciophila: aedeagus.
Figs. 1124, 1125. Ophiomyia senecionina: 1124, aedeagus, side view; 1125, distiphallus, ventral view.
Subclass ASTERIDAE 293
Table 72
Agromyzidae on Senecioneae
male genitalia are characteristic of the genus Compo sitae confirm their close relationship. It will
(Spencer, 1961e: fig. 11). M. compositana Spencer be interesting to discover when material from Cras-
(1959) was recorded from various countries in Afri- socephalum in Africa can be examined whether the
ca and the genitalia of the holotype 'ex Composite l' speciation now detectable between the population
from Naivasha, Kenya show that this is not closely on Crassocephalum in Madagascar and on Senecio
related to M. seneciocaulis. One series was also in Kenya has resulted from isolation of the popula-
reared from a host referred to as 'Senecio 532', and tion in Madagascar or whether it is associated with
the genitalia (Figs. 1116, 1117) are identical to the different hosts, Crassocephalum and Senecio.
those of the holotype. This species is very close to Also in Kenya I recently obtained a series of a
M. gynurae Spencer (1959, nom. nov. for viridis new greenish Melanagromyza (described as M. hea-
Seguy, 1951, preoccupied) in Madagascar. The host toni sp.n. in the Appendix below) which was feeding
is Crassocephalum sarcobasis (originally in Gynura) as an internal stem-borer in S. moorei at 3000 m in
and the species was recorded as a leaf miner. This I the Aberdares National Park, west of Nyeri. The
do not accept. I have recently re-examined the a genitalia show that this species is related to M. sene-
and 2 QQ in the type series and the species is typical ciocaulis but it is substantially larger and is the larg-
of the greenish group feeding in stems or flower- est species in the genus known in Africa, with wing
heads. (Seguy, 1951, made a further mistake in re- length up to 3.7 mm. A second species caught on
cording the host as Bougainvillea spectabilis, al- Senecio moorei at the same locality as M. heatoni
though the data labels indicate the host as Gynura proved to represent M. ruandae Spencer (1959), of
and this was confirmed by Pauli an (1953: 21) who which a long series is known from Addis Ababa
was working in Madagascar at that time). The geni- (Spencer, 1961 d). This is a black species and is en-
talia of M. gynurae (Fig. 1118) and of M. composita- tirely different from M. heatoni and the genitalia of
na (Fig. 1116) from Senecio and possibly other the two show that they are not directly related
294 Class MAGNOLIOPSIDA (Dicotyledons)
~ ..
1126 1127
1129
1130
1134
1135 1133
Fig. 1126. Liriomyza latigenis: aedeagus, ventral view.
Fig. 1127. Liriomyza erucifolii: aedeagus.
Fig. 1128. Liriomyza kleiniae: aedeagus.
Figs. 1129, 1130. Liriomyza emiliae: 1129, aedeagus, side view; 1130, distiphallus, ventral view.
Fig. 1131. Calycomyza majuscula: aedeagus.
Figs. 1132, 1133. Chromatomyia seneciophila: 1132, aedeagus, side view; 1133, same, dorsal view (both holotype).
Figs. 1134, 1135. Chromatomyia seneciovora: 1134, aedeagus (holotype); 1135, aedeagus (Senecio moorei).
Subclass ASTERIDAE 295
(M. heatoni: Figs. 1119, 1120, M. ruandae: Fig. show no significant differentiation, despite the isola-
1121). I suspect that M. ruandae may feed in the tion of the species geographically and on its semi-
flower-heads of Senecio moorei and probably other succulent host.
Senecio species. In Alberta, Canada L. seneciovora Sehgal
A single Melanagromyza is known in New Zea- (1971b) was described from S. pauciflorus, and near
land, M. senecio nella Spencer (1976b), which is not Sao Paulo, Brazil I discovered L. seneciella Spencer
uncommon in stems of the introduced Senecio jaco- (1963c) forming linear mines on S. braziliensis. The
baea but I also found it on native S. minima. The male genitalia give no firm indication of the affilia-
genitalia (Fig. 1122) show clear relationship with tions of these two species.
M. seneciophila Spencer (1977c) (Fig. 1123) which Seguy (1951) described L. emiliae in Mada-
is widespread in southern Australia from N.S.W. to gascar, forming linear mines on Emilia citrina. I
WA. The species was reared from S. vagus in have examined the type series of 14 specimens and I
N.S.W. and I obtained many specimens on S. hypo- am satisfied that all are conspecific. Eight were
leucus and S. quadridentatus near Adelaide; there reared from Emilia, four from unidentified 'compo-
was also a large population on S. ramosissimus on sees' and two from Tropaeolum. The male genitalia
the Porungurups, south of Perth. Specimens were (Figs. 1129, 1130) suggest relationship with L. pu-
also seen on Chrysanthemum, Gnaphalium and He- silla but this is puzzling as L. pusilla, although wide-
lichrysum but it remains to be established whether spread in India and elsewhere in the Oriental
the species was breeding on these hosts. It seems Region on Crassocephalum, is not known in Africa.
certain that the New Zealand species is derived L. emiliae is treated as oligophagous in Table 71.
from an Australian ancestor which succeeded in The two polyphagous Liriomyza species, L. stri-
crossing the Tasman Sea. gata and L. trifolii are found not uncommonly on
In Florida M. erechtitidis Spencer (cf. Spencer various Senecio spp., the former in Europe, the
and Stegmaier, 1973) feeds in the flower-heads of latter in both Europe and North America.
Erechtites hieracifolia, and M. splendida, which is Calycomyza majuscula (Frick, d. Spencer and
highly oligophagous and found most commonly on Steyskal, 1986) was described from California on
genera of the Heliantheae, also feeds in the stems of Senecio triangularis and has subsequently been
the same host. found on S. lugens and Petasites sp. in Alaska. The
Only a single Ophiomyia has been recorded on genitalia (Fig. 1131) show that this species belongs
Senecio, 0. senecionina Hering (cf. Spencer, to the C. humeralis group. The highly oligophagous
1964c), known only in England and northern C. platyptera, which has been recorded on seven
France, forming external stem-mines on S. jacobaea tribes but primarily on the Heliantheae, is also
and S. erucifolius. The male genitalia (Figs. 1124, known on Senecio sp.
1125) are highly asymmetric but characteristic of Napomyza lateralis, which is highly oligophagous
the genus. and is discussed below under Anthemideae, was
In South America Amauromyza maculosa which found to transfer readily to wild Senecio vulgaris
is found most frequently on genera of Astereae (see growing near a large population infesting cultivated
above) is common on Emilia. Calendula in southern England (Morgan, 1954).
Five Liriomyza species are known which are re- Two Chromatomyia species are known in Africa,
stricted to the Senecioneae. In Europe L. latigenis associated with Senecio s.1. Chr. seneciovora (Spen-
(Hendel, cf. Spencer, 1966e) has been reared from cer, 1959; cf. Spencer, 1985a) was found on the tree
Senecio jacobaea near Paris, France. The species is Senecio, now Solanecio mannii (cf. Jeffrey, 1986)
unusual in lacking the outer cross-vein (Hendel thus on Mt. Cameroun, West Africa and Chr. senecio-
described it in Haplomyza) but the genitalia (Fig. phila Spencer (1985a) was reared from another tree
1126) although distinctive, are not atypical of the Senecio, now S. johnstonii ssp. battiscombei var.
genus. L. erucifolii de Meijere cf. Spencer, 1976a) battiscombei (cf. Jeffrey, 1986) at Cave Falls, Aber-
forms linear mines on S. erucifolius and S. jacobaea. dare Mountains, Kenya. During a recent visit to
The species was discovered by Hering who formally Kenya in February, 1988 I obtained specimens
described it as L. senecifolii but de Meijere's de- clearly feeding on S. moorei, a herbaceous species,
scription based on larvae sent to him by Hering near the Fishing Lodge, close to the locality of Chr.
antedated Hering's description by two weeks. The seneciophila. This species is closer to Chr. senecio-
genitalia (Fig. 1127) suggest relationship with species vora than to Chr. seneciophila and I am treating it as
in the L. sonchi group (cf. Figs. 990, 991). L. klei- Chr. seneciovora, despite the different host and the
niae Hering (1927a) is known only in the Canary Is- distance separating the two populations. The male
lands on Fuerteventura and Tenerife, the larva genitalia of Chr. seneciophila (Figs. 1132, 1133) are
forming irregular linear blotch mines on Kleinia ne- obviously distinct but the aedeagus of Chr. senecio-
riifolia. The genitalia of the holotype (Fig. 1128) vora (Fig. 1134) is very close to that of the Kenyan
296 Class MAGNOLIOPSIDA (Dicotyledons)
1136 1137
1138
1139
1142 1143
Subclass ASTERIDAE 297
population (Fig. 1135) but the differences suggest P. aronici Nowakowski (d. Griffiths, 1974b). Host:
incipient speciation. A specimen reared from Sene- Doronicum c/usii, Tatra Mountains, Poland. In
cio sp. on Mt. Mem, Tanzania is close to but not his description Nowakowski treated this species
identical with Chr. seneciophila. as closely related to P. doronici (see below) but
It seems clear that active speciation is occurring Griffiths (1974b) considered it to be more close
in populations of this group isolated at 2500-3000 to P. alpina and distinguishable only by minor
m on the mountains in both East and West Africa details of the genitalia. With P. alpina now known
and it will be of great interest to study additional to feed on Petasites in addition to its originally re-
material, particularly from Mt. Kenya and Mt. corded host, Senecio, the possibility that it may
Kilimanjaro. Spencer (1985a) suggested that Chr. have Doronicum as a further host cannot be ex-
seneciovora may also feed on Helichrysum, Coreop- cluded and I consider that the status of P. aronici
sis, Gynura and a large-leaved species in the tribe should be reconsidered when additional material
Inuleae (after de Meijere, 1940). No adults were becomes available.
reared from these hosts and the host selection of P. burchardi Hering (1927a). A single female reared
Chr. seneciovora requires further study. from Senecio papyraceus on Tenerife, Canary Is-
In Alberta and Alaska Chr. senecio nella (Sehgal, lands. The larva forms a rather broad linear mine
1971b) feeds probably primarily on Senecio spp. (Hering, 1927a: fig. 86). Hering refers the species
but has also been found on Petasites jrigidus. The to the albiceps group and, pending discovery of a
male genitalia (Sehgal, 1971b: fig. 123 and Griffiths, male, this is accepted.
1972b: fig. 13) have the 'supporting sclerites' slend- P. californica Griffiths (1974b, cf. Spencer and
er as in Fig. 1135 but the upcurved distiphallus is Steyskal, 1986). Host: Senecio triangularis, only
strongly developed and it is improbable that there is known locality summit of Luther Pass (7740 ft),
any direct relationship between the two species. central California. Male genitalia: Fig. 1137, typi-
The polyphagous species Chr. horticola, and also cal of albiceps group.
Chr. syngenesiae both occur commonly on Senecio P. doronici Hering (d. Griffiths, 1974b). Host:
species and the long white mines, normally of Chr. Doronicum spp., mountains of Central Europe.
syngenesiae, are also invariably present on culti- No male is known and the exact affiliations of the
vated Cineraria cruentus in glasshouses. species remain to be established.
There has been remarkable radiation of Phyto- P. homogyneae Hendel (d. Griffiths, 197 4b). Host:
myza species on the six genera Adenostyles, Doroni- Homogyne alpina, mountains of Central Europe.
cum, Homogyne, Petasites, Senecio and Tussilago, Male genitalia (Fig. 1138), with strong patches of
with a total of 15 species and three subspecies, all in spinules at front end of basal sclerites. Sasakawa
Europe, Japan or North America. Griffiths (1972b, (1961) recorded this species in Japan, with both
197 4b) has discussed these species in detail and it Homogyne and Aster as hosts. His illustration of
will suffice here to recapitulate essential informa- the genitalia (1961: fig. 127d) shows that the
tion. It should be noted that Griffiths included Arni- strong spinules of P. homogyneae are entirely
ca within Senecioneae, but this genus is now accept- lacking and it seems probable that a further spe-
ed in the tribe Helenieae (see below). cies is involved. Whether the host of the specimen
illustrated was Aster or Homogyne was not indi-
Phytomyza albiceps group (10 species, 3 sub- cated.
species): P. integerrimi Griffiths (1974b). Host: Senecio inte-
gerrimus, Oregon, Colorado. Male genitalia (Fig.
P. adenostylis Hering (d. Griffiths, 1974b). Host:
1139), without spinules normally present in the
Adenostyles spp., mountains of central Europe.
albiceps group.
No male is known.
P. senecionis Kaltenbach (d. Griffiths, 1972b).
P. alpina Groschke (1957). Hosts: in Europe Sene-
Hosts: Senecio spp., particularly S. fuchsii and
cio, in Alaska, Yukon and British Columbia also
S. nemorensis, widespread in central and eastern
Petasites. A typical member of the albiceps
Europe, reaching Denmark but not Britain. Male
group, with two patches of spinules adjoining the
genitalia (Fig. 1140) with two patches of strong
basal sclerites (Fig. 1136).
spinules. Sasakawa (1955b, cf. 1961) described ulations are also present in northern Scandinavia
ssp. ravasternopleuralis in Japan on Senecio pal- or Siberia.
matus. Griffiths (1974b) raised this subspecies to
specific rank, primarily on the basis of apparent Phytomyza robustella group (6 species): six species
differences in the male genitalia between Sasa- in this group are known on Petasites, Senecio and
kawa's illustration and a specimen from Germany Tussilago, all pupating internally.
which he had illustrated (Griffiths, 1972b: fig.
P. buhriella Spencer (1969a, cf. Griffiths, 1972b).
11). However, this was a damaged specimen with
Hosts: Petasites, Tussilago, Europe. Larvae mine
the distiphallus missing and a replacement draw-
mainly in petioles of basal leaves. Male genitalia
ing was published later (Griffiths, 1974b: fig. 14).
(Fig. 1143), with distiphallus strongly sclerotized.
There is thus no essential difference in the geni-
P. farfarae Hendel (cf. Griffiths, 1974b). Hosts: Pe-
talia between Sasakawa's ssp. ravasternopleuralis
tasites, Tussilago, mountains of Central and Eas-
and typical senecionis. The Japanese population
tern Europe. Mines deep, interparenchymal, nor-
is somewhat darker than typical senecionis, and
mally beside veins in leaf blade. Genitalia: Fig.
the leaf mine is more irregular, while in Europe
1144.
the mines are conspicuously long and frequently
P. hasegawai Sasakawa (1981b). Host: Peta~ites ja-
adjoining the midrib or lateral veins (Fig. 1141).
ponicus, Japan (Hokkaido). Larvae feed III both
It is therefore possibly justified to treat ravaster-
the stem and flower-head. Male genitalia (Fig.
nopleuralis as a subspecies but not as a full spe-
1145). This is the largest species in the group on
cies and this new status is formally established in
Petasites, with wing length up to 3.5 mm.
the Appendix below.
P. hyperborea Griffiths (1972b). Host: Petasites fri-
P. tussilaginis Hendel (cf. Griffiths, 1972b). Hosts:
gidus, Alaska. Male genitalia generally similar to
Tussilago, Petasites, widespread in Europe. Male
P. faifarae but adult distinguishable by long
genitalia (Fig. 1142), typical of albiceps group. I
pubescence on third antennal segment.. ..
described a particularly pale form on Petasites in
P. hypophylla Griffiths (1972b). Host: Petasltes frlgl-
Alberta as P. petasiti Spencer (1969b). Griffiths
dus, Alaska, Alberta. A darker species th~n
(1972b) reduced this to a ssp. of tussilaginis and
P. hyperborea, male genitalia generally as III
recorded it also in British Columbia, Yukon Ter-
P. faifarae but differing in detail. Leaf mine ge~
ritory and Alaska. He also (loc. cit.) described a
erally on lower leaf surface (on upper surface III
somewhat melanic form with slightly differing
P. hyperborea on same host), distribution in the
wing venation as ssp. kevani, present in the
two species probably overlapping.
Yukon and Northwest Territories. With no ob-
P.lugentis Griffiths (1972b). Hosts: Senec~o sp~.,
vious differences in the genitalia between the
British Columbia. A somewhat dark speCIes as III
three subspecies, differentiation must be relative-
P. hypophylla, but male genitalia distinctive. Lar-
ly recent and it will be of interest to discover
vae mine on either upper or lower leaf surface.
whether this incipient speciation has occurred on-
ly in the Canadian Arctic or whether similar pop-
r. \'
-----.1 t
1144 \) 1145
Fig. 1144. Phytomyza farfarae: aedeagus (Griffiths).
Fig. 1145. Phytomyza hasegawai: aedeagus (Sasakawa).
Subclass ASTERIDAE 299
1146 1148
1149
1147
1152
1150 1153
'-'.::::..::"
/' ti /'
~'I
Table 74
that this was the host family of the ancestor of in Japan and Argentina. It seems clear that the
this group with a later isolated coloniz~tion of Arte- radiation has resulted from multiple colonization.
misia. In southern Canada A. ambroslvora Spencer The favourite host is Artemisia with seven species
(1969b) was discovered on Ambrosia artemisiifolia and in addition three polyphagous species have
(see Heliantheae below). Adults were cau~ht ~d colonized several genera (see Table 74).
empty mines found on Artemisia douglaszana m
Liriomyza analis (Rondani, cf. Hendel, 1936). Host:
southern California. This is an interesting and un-
Chrysanthemum coronarium, Madeira (according
usual host combination.
to handwritten note in Hering's personal copy of
Amauromyza maculosa which is found most
his 'Keys' (1957b)). The species was described
commonly on genera of Astereae (see above) feeds
from northern Italy and its rediscovery on C. co-
regularly on Chrysanthemum. .
ronarium in Madeira is not implausible but needs
Phytoliriomyza scotica Spencer (1962), descnbed
confirmation.
from caught specimens in Scotland, has recently
L. artemisiae Spencer (1981). Host: Artemisia
been found in large numbers on Gomera, Canary Is-
douglasiana, southern California. The larva forms
lands, certainly feeding on Cotula austra~i~ (von
a normal short linear mine but is unusual in
Tschirnhaus). This plant is of Australian ongm and
pupating in the leaf. Male genitalia (Figs. 1148,
is now established on several islands in the Canaries
1149) not suggesting direct relationship with
group, on Madeira and in Portugal. Phlir. scotica
L. artemisicola in Europe.
must have other native composite hosts and the
L. artemisicola de Meijere (cf. Spencer, 1976a).
most likely is the widespread genus Matricaria
Host: Artemisia vulgaris, larva forming short mine
(same tribe as Cotula). It is now bel~eved that Phlir.
normally limited to single leaf segment (Fig.
pilosella (in Spencer and Stegmmer, 1~73), ~e
1151). Widespread in much of Europe. Male gen-
scribed from Florida, may represent Phlzr. scotlca
italia (Fig. 1150).
but more detailed study will be necessary before this
L. demeijerei Hering (cf. Spencer, 1976a). Host: Ar-
possible synonymy can be formally establish~d.
temisia vulgaris, mine longer than in previous
Eleven Liriomyza species are known as tnbe-spe-
species, frequently following leaf margin or veins.
cific on the Anthemideae, mainly in Europe but also
Western Europe. Male genitalia: Fig. 1152.
302 Class MAGNOLIOPSIDA (Dicotyledons)
1159 1160
''4'a~?r;r>
.... ':'
1161 1162
1163 1164
Subclass ASTERIDAE 303
L. dracunculi Hering (d. Spencer, 1976a). Hosts: from Europe to China and Japan, and in Canada,
Artemisia absinthium, A. campestris, larva mining U.SA. and Colombia. In Canada in addition to a
in the very fine leaves of these hosts. Germany, population identified as artemisiae s.s., I described
Scandinavia. Male genitalia: Fig. 1153. C. marcida Spencer (196 9b), the host later being es-
L. flavopicta Hendel (d. Spencer, 1976a). Host: tablished in California as Artemisia douglasiana,
Achillea millefolium, larva forming external stem- and I therefore reduced marcida to subspecific
mine immediately below flower-bud. Western status. The slight but distinctive differences in the
Europe, uncommon. Male genitalia: Fig. 1154. genitalia are apparent from the following illustra-
L. gudmanni Hering (d. Spencer, 1976a; von tions: Europe (Figs. 1159, 1160), southern Cali-
Tschirnhaus, 1981). Hosts: Artemisia maritima, fornia (Figs. 1161, 1162), ssp. marcida (Figs. 1163,
A. rupestris. Germany, Denmark, Poland. Male 1164). Sasakawa (1961: fig. 61d) illustrated the gen-
genitalia: Fig. 1155. italia of a species identified as artemisiae in Japan
L. katoi Sasakawa (1961). Host: Artemisia vulgaris, but this is clearly distinct from typical artemisiae
Japan. The larva forms an upper surface leaf mine from Europe. A record given by Sasakawa (1972) of
and the male genitalia are typical of this group a male of artemisiae reared from Patrinia villosa
(Sasakawa, 1961: fig. 80). (Valerianaceae) in Taiwan must be treated as doubt-
L. matricariae was recorded in Schleswig-Holstein, fully correct without confirmation from the male
north Germany as a nomen nudum by von genitalia. A closely related species I found in
Tschirnhaus (1981), with Matricaria matricarioi- Colombia and Venezuela reared from Stevia elatior
des as host. (Eupatorieae) was described as C. steviae Spencer
L. ptarmicae de Meijere (d. Spencer, 1976a) (1973c)(Figs. 1110, 1111).
(= L. millefolii Hering, L. chrysanthemi Hering). With no differences between adults from these
Hosts: Achillea, Leucanthemum, common in various populations so far detected, this species-
western Europe, also present in Alberta. Adult complex, with its remarkably wide distribution, pro-
distinguished by long pubescence on third anten- vides clear evidence of incipient speciation and it
nal segment (Fig. 1157), male genitalia (Fig. must remain a matter of subjective judgement
1156) characteristic of group. whether individual populations are considered to
L. spencerella Valladares (1984). Hosts: Chrysanth- represent mere variation, or are given subspecific
emum morifolium but oligophagous, also on Ca- status or are treated as full species. A detailed revi-
lendula, Bidens and Helianthus. Only known in sion of the complex, with the study of even more
Argentina. Genitalia: Figs. 1114, 1115. material, could prove of great interest.
L. tanaceti de Meijere (cf. Spencer, 1976a). Host: Five Napomyza species are known here, of which
Tanacetum, widespread in much of Europe. On four appear to be monophagous. In Europe N. an-
larval characters, with the posterior spiracles each nulipes (Meigen, cf. Hendel, 1935) forms root galls
having 8-10 pores (instead of the normal 3), this on Artemisia campestris and A. scoparia. This is a
species can be associated with L. strigata but the large, uncommon species but the genitalia (Fig.
male genitalia (Fig. 1158) are not atypical of the 1165) are characteristic of the genus. N. plumea
pusilla group. Spencer (1969b) is a distinctive species, with the
L. togata (Melander, d. Spencer and Steyskal, third antennal segment pubescent and, in the male
1986). Host of one female in California given as genitalia, the distiphallus long and pigmented (Fig.
Artemisia douglasiana but discussed above under 1166). The boreal distribution is known in Scandi-
Baccharis (Astereae). Genitalia: Figs. 1055, navia north of the Arctic Circle, Alaska, northern
1056. Canada and extending south to northern California.
The species has only once been reared, from the
Calycomyza artemisiae (Kaltenbach, d. Spencer,
stem of Achillea sp. at San Francisco. N. maritima
1976a) occurs commonly in Europe on Artemisia
Tschirnhaus (1981) has only been reared from Arte-
vulgaris, and less frequently on Eupatorium canna-
misia maritima on the coast of Schleswig-Holstein,
binum, forming conspicuous white blotch mines. In
north Germany but specimens treated as paratypes
the male genitalia the distiphallus is asymmetrical
were recorded from Istria, Yugoslavia and the south
and can appear very different with every 10° of rota-
of France, with other white-flowering Artemisia spp.
tion. The species has a remarkably wide distribution,
Figs. 1159-1162. Calycornyza arternisiae: 1159, aedeagus, side view; 1160, distiphallus, ventral view (both Europe); 1161, aedeagus,
side view; 1162, same, ventral view (both southern California).
Figs. 1163, 1164. Calycornyza arternisiae ssp. rnarcida: 1163, aedeagus, side view; 1164, same, ventral view (both California).
Fig. 1165. Napornyza annulipes: aedeagus.
Fig. 1166. Napornyza plurnea: aedeagus.
Fig. 1167. Napornyza lateralis: aedeagus.
304 Class MAGNOLIOPSIDA (Dicotyledons)
1168
1171
1172
,~ ...... ~
.,
1177
Subclass ASTERIDAE 305
as hosts. The genitalia are generally as in N. lateralis in the robustella group. Seven of the species have
(see below) but have a characteristic sharp bend to been recorded in Europe, three in Japan and four in
the rear of the distiphallus (more rounded in N. late- NW Canada and Alaska. Griffiths (1974d) made a
ralis). Also in Schleswig-Holstein von Tschirnhaus detailed study of Phytomyza species on Artemisia
recorded N. achilleanella as a nomen nudum feed- but other hosts now recorded include Achillea
ing on Achillea millefolium. Chrysanthemum and Leucanthemum (see Tabl~
Napomyza lateralis (Fallen, cf. Spencer, 1976a) is 74). Essential details concerning these species are as
highly oligophagous and crosses tribal boundaries follows:
but with numerous misidentifications in the past and
the close similarity of a group of species on Astera- Europe and Japan:
ceae and others on other families, revisionary work
will be necessary accurately to establish the hosts of Phytomyza achilleae Hering (cf. Hendel, 1935).
the true latera lis , described originally from a single Host: Achillea millefolium, only reared once near
Berlin. Male genitalia (Fig. 1169), with disti-
caught specimen in Sweden. I accept records on the
four tribes Cardueae (Carduus, Centaurea, Sily- phallus ~arge, membranous, undivided, generally
rese~bhng some species on Apiaceae (see dis-
bum), Inuleae (Helichrysum, Inula), Calenduleae
CUSSIon below). I found empty mines on Achillea
( Calendula) and Coreopsideae (Bidens), in addition
on Mt. Vesuvius, Italy in April, 1960 which I
to the Anthemideae (Anthemis, Matricaria).
believe represent this uncommon species (pupa-
Records on Lactuca (Lactuceae) and also on Linum
rium black).
(Linaceae) require confirmation.
P. artemisivora Spencer (1971b). Host: Artemisia
Matricaria inodora is a favourite host of this spe-
vul~aris, widespread in Europe (misidentified by
cies and I have recently re-examined the male geni-
earher workers as P. albiceps Meigen (cf. Car-
talia of specimens reared from this host in the
dueae». Larva forming long, white linear mine
Netherlands and Cornwall, England and they agree
genitalia (Fig. 1170) remarkably similar to P. cor:
virtually exactly with those of the holotype from
vimontana but adults clearly distinguishable.
Sweden (Fig. 1167). Although oviposition takes
P. corvimontana Hering (cf. Hendel, 1935). Host:
place in a leaf, the larva feeds primarily in the stem
Achillea ptarmica, local in western Europe. Des-
and capitulum where it pupates. N. lateralis is com-
pite very similar genitalia (Fig. 1171) to those of
mon in Europe and I have seen specimens from the
~. artem~sivora, the two species are entirely dis-
Northwest Territories, Canada 'Assoc. with Matri-
caria inodora.' The species thus appears to have a tmct, WIth the scutellum in P. corvimontana
bright yellow (as P. conyzae in this group, cf.
natural hoi arctic distribution.
Inuleae).
Two closely related Chromatomyia species are
known in Europe on Leucanthemum vulgare. Chr. P. hiemalis Griffiths (1974a). Host: Artemisia
ciliata (Hendel, cf. Spencer, 1976a) (described in keiskiana, only known from Botanical Gardens
Phytomyza but transferred to Chromatomyia in the Tokyo. During a brief visit to Tokyo in January:
Appendix below) is a boreal-alpine species which I 1960, I made an obligatory visit to the Botanical
reared in Switzerland. The genitalia (Fig. 1168) are Gardens despite snow and intense cold and was
close to Chr. paraciliata Godfray (1985) which was delighted to find larvae feeding in the basal leaves
found on the same host in southern England. In ad- of Artemisia beneath the snow. The resulting
dition the common polyphagous Chr. horticola and a?ults proved to represent an undescribed spe-
the oligophagous Chr. syngenesiae (treated as 'atri- CIes which Griffiths diagnosed 14 years later. The
male genitalia (Fig. 1172) suggest relationship
cornis' when known only from leaf mines and not
with P. japonica.
positively identified) have been recorded on many
genera (Table 74). P. jap?nica Sasakawa (cf. Griffiths, 1974d). Type
Fourteen tribe-specific Phytomyza species are senes reared from Aster indicus (see under As-
known here, 13 in the albiceps group and only one tereae) but additional hosts recorded later as Ar-
~,-.,-.-
,\ ',
,
1179
Chrysanthemum in Japan (and also reported on Asteraceae the following numbers of feeding punc-
Lactuceae, Astereae and Heliantheae) has not suc- tures and eggs laid were recorded:
ceeded in colonizing any host in the Senecioneae.
The favourite host genus both in Europe and NW Host Feeding punctures Eggs
Canada is Artemisia, with a total of 27 species and
this approaches the number on Senecio (30). Achillea sibirica 567 28
Artemisia sp. 227 10
Despite the many cases of closely related species Aster ciliolatus 10 0
on either side of the Bering Straits, there are only Chrysantheum sp. (cultivated) 467 34
three - Napomyza lateralis, Liriomyza ptarmicae Helianthus annuus 292 19
and Phytomyza pullula (= matricariae) - which are Matricaria matricarioides 441 30
known both in Europe and NW Canada. With their Senecio vulgaris 54 0
Solidago sp. 12 0
hosts being weeds - Achillea, Artemisia and Matri- Sonchus uliginosus 4 0
caria - it seems possible that there has been fortui- Tanacetum vulgare 443 35
tous introduction to Canada, although I accept Taraxacum officinale 124 4
N. lateralis as a natural holarctic species. Zinnia sp. (cultivated) 44 3
Phytomyza achilleae has proved to be a most in-
teresting species. It has only been recorded once It will be seen that most eggs were laid on Tanace-
from a single locality north of Berlin and the geni- tum vulgare, the host from which the females had
talia have not been previously examined. These been reared, with a Chrysanthemum sp. having a
show that the species is closely related to P. spon- comparable number, followed by two further genera
dylii (primary host Heracleum, Apiaceae), both be- in the same tribe, Achillea and Matricaria. The few
longing to the alhiceps group. However, normally eggs laid on Artemisia is consistent with the rare oc-
the distiphallus is divided, with some degree of pig- currence of P. pullula on this host in nature, while
mentation, while in these two species it is almost ful- the higher number on Helianthus is perhaps surpris-
ly fused into a single large, entirely membranous ing, as P. pullula has never been found on any host
tubule, and another significant character in common in the Heliantheae. It is interesting that Aster, Sene-
is the presence of a weakly pigmented semicircular cio and Solidago were totally rejected, while the two
sclerite near the end of the left basal sclerite (Figs. species in the Lactuceae had only four eggs (Taraxa-
1169,653). It thus seems that there has been a rela- cum) and none (Sonchus). These results illustrate
tively recent switch (from Heracleum to Achillea) of the high degree of discrimination between hosts in
P. achilleae or its immediate ancestor. the different tribes of the Asteraceae. Obviously it
Von Tschirnhaus (1981) recorded and named would be of great interest to carry out comparable
three new species from the coast of Schleswig-Hol- experiments with other species on the Asteraceae,
stein in north Germany (with suggested hosts indi- and beyond this an infinite amount of experimental
cated in brackets): Liriomyza matricariae (Matri- work on many other families would be valuable in
caria matricarioides), Napomyza achilleanella confirming the results obtained by Sehgal.
(Achillea millefolium) and Phytomyza achilleana
(Achillea ptarmica). These names all represent
Helenieae (58 genera, 400+ species)
nomina nuda but von Tschirnhaus is reasonably cer-
tain of their hosts and the species are included here. The taxonomic position of this assemblage, which is
largely of North American distribution, has received
Experiments testing the host preference of Phyto- much attention by botanists but there is still no gen-
myza pullula (as matricariae) carried out by Sehgal eral consensus. For example, Willis (Sixth Edition,
(1971a) at Edmonton, Alberta showed conclusively 1957) treated the Helenieae as distinct from the
the preference of this species to hosts in the tribe Heliantheae but Mabberley (1987), largely follow-
Anthemideae. Twenty-seven plant species were test- ing Cronquist, includes both the Helenieae and Ta-
ed in the Polypodiaceae (1 species), Ranunculaceae geteae within the Heliantheae sensu lato. The tribe is
(3), Papaveraceae (1), Chenopodiaceae (1), Cucur- discussed in detail in Heywood et al. (1977) and
bitaceae (1), Brassicaceae (2), Rosaceae (2), Faba- given separate status and this is accepted by Bremer
ceae (5), Tropaeolaceae (1), Solanaceae (3), Scro- (1987). This splitting finds little support from the
phulariaceae (1), Lamiaceae (1), Campanulaceae host selection of the few agromyzids on the tribe,
(1), and in the moncots Liliaceae (2), Poaceae (1) with only Arnica (7 species) and Helenium (1)
and Typhaceae (1). In these, although a few feeding having species restricted to the tribe. Conversely, 12
punctures were recorded (most on Pisum), in no inter-tribal or polyphagous species find the tribe ac-
cases were any eggs deposited. This provides an ex- ceptable.
cellent illustration of the ability of females to detect Until recently Arnica was included in the Sene-
and reject unsuitable hosts. In contrast, on the cioneae but the genus was discussed by a number of
308 Class MAGNOLIOPSIDA (Dicotyledons)
authors in Heywood et al. (1977) and it was general- have also been recorded on Gaillardia and
ly agreed, primarily on chemical grounds, that Arni- Helenium respectively.
ca cannot be logically accommodated in that tribe. No described Ophiomyia is known on this tribe
Nordenstam, in a detailed discussion of the genus but a mine referable to this genus was found on
(Heywood et al., p. 822) gives 16 reasons for ex- Arnica in Denmark. This may represent an unde-
cluding Arnica from the Senecioneae. However, scribed species but may possibly be of 0. curvi-
there is no clear agreement where the genus proper- palpis which feeds mainly on genera in the Anthe-
ly belongs. It is even suggested that a new tribe 'Ar- mideae but also on Reichardia (Lactuceae) and
niceae' might be created for the Arnica assemblage Aster.
but 'Thorough studies are required before such a The highly oligophagous Amauromyza maculosa,
step is formally taken.' Turner and Powell (in Hey- which has been recorded on nine tribes (see As-
wood et al., 1977) proposed a new subtribe Arni- tereae), has been found feeding on Dahlia, Gail-
cinae within the Senecioneae but they write further: lardia and Helenium.
'of course, the point we make is that Arnica and the The only tribe-specific Liriomyza found here is
genera centering about it are not typical senecioid L. helenii Spencer (1981) found in the mountains in
elements; we believe them to be ancestral elements California on Helenium hoopesii. The male genitalia
with phyletic connections back to the asteraceous (Fig. 1181) are distinctive but give no definite indi-
complex which gave rise to both the Senecioneae cation of its affiliations. I found L. venegasiae
and Heliantheae.' The general agreement is that Ar- Spencer (1981) to be abundant at several localities
nica belongs in or near the Heliantheae or in southern California north of Los Angeles on
Helenieae and here I follow Bremer (1987) who in- Venegasia carpesioides and also reared the species
cludes it in the Helenieae. from Eidens pilosa at the University Campus at
The generic breakdown of the Agromyzidae Riverside. It somewhat resembles L. sabaziae re-
known on the Helenieae is shown in Table 75 and corded here on Dahlia but feeding mainly on Saba-
the host genera with their agromyzid species are list- zia and Galinsoga (Heliantheae). The genitalia (Fig.
ed in Table 76. 1182) suggest that the two species may be closely
related.
Table 75 Of the polyphagous Liriomyza species, L. bryo-
niae, L. strigata and L. trifolii have been found on
Agromyzid genera on Helenieae
Dahlia, and the two latter also on Gaillardia. Both
Genus Species Chromatomyia horticola and Chr. syngenesiae have
also been confirmed on the same hosts.
Restricted to tribe Inter- Poly-
tribal phagous In December, 1958 I found an empty mine (Fig.
1183) on Villanova anemonifolia at the foot of the
Melanagromyza 2 Teleferico, Bogota, Colombia. This probably repre-
Ophiomyia (1) (undescribed, sents a Liriomyza sp. but its identity remains to be
status uncertain)
confirmed when adults can be obtained. I also found
Amauromyza 1
Liriomyza 1 (+2, mines only) 2 3
mines of two species on Porophyllum ruderale at
Calycomyza Caracas, Venezuela, possibly representing Am. ma-
Chromatomyia culosa and a Liriomyza sp. (Fig: 1184A, B).
Phytomyza 7 The oligophagous Calycomyza platyptera has
been recorded on Helenium in the United States.
9 (+2?) 8 4
There has been remarkable radiation of Phyto-
Total 21 (? + 2) myza on Arnica, with four species in the albiceps
group and three in the robustella group, restricted to
The only Melanagromyza restricted to this tribe is this genus. P. arnicae and P. arnicicola were first re-
M. arnicarum Hering (cf. Spencer, 1976a) which corded from Sweden but have now been discovered
feeds in stems of Arnica montana and is known only in NW Canada, the latter extending its range south-
in Denmark and Germany. The distinctive male gen- wards to mountains in California and Colorado.
italia (Fig. 1180) show striking similarity to M. pole- Griffiths (1974b) discussed these species in detail
monii (Fig. 730) (Polemonium, Solanales). I con- (treating Arnica within the Senecioneae) and it will
sider that it is improbable that this represents con- suffice to give here the following brief details:
vergence and suggest there was a switch between
Arnica and Polemonium of the species ancestral to Phytomyza albiceps group
M. arnicarum and polemonii.
In the United States the two inter-tribal species, P. arnicae Hering (cf. Griffiths, 1974b; Spencer,
M. splendida and M. minimoides (see Heliantheae) 1976a). Mountains of Central Europe, Sweden;
Subclass ASTERlDAE 309
Alberta. Larvae forming blotch mines. Male geni- in NW Canada. Mine similar to that of P. cam-
talia (Fig. 1185) with numerous strong spinules pestris. Male genitalia (Griffiths, 1974b: fig. 15).
between basal sclerites, distiphallus appearing
The small size of this tribe is reflected in the small
membranous.
number of agromyzids which have colonized it. It is
P. arnicicola Lundquist (Griffiths, 1974b; Spencer,
fully acceptable to inter-tribal and polyphagous spe-
1976a). Northern Sweden; Alaska, California,
cies (Table 76) and it seems surprising that it sup-
Colorado. Larvae forming linear mines. Male
ports only single tribe-specific Melanagromyza and
genitalia: Fig. 1186.
Liriomyza species. In contrast to this, the presence
P. conyzae Hendel (= P. arnicophila Hering) (cf.
of seven host-specific Phytomyza species on Arnica,
Griffiths, 1974b; Spencer, 1976a). Recorded as
two of holarctic distribution, is remarkable. Differ-
P. arnicophila but synonymy with conyzae estab-
ences between these species are slight but are fully
lished by Hendel (1935) and confirmed by Grif-
reflected in the male genitalia and in the larval feed-
fiths and Spencer. Normal host Inula (see Inu-
ing instinct, resulting in differentiation in the leaf
leae), found once in Denmark on Arnica montana.
mines in most species. Although speciation has pre-
P. saximontana Griffiths (1974b). Alberta, Cali-
sumably been relatively recent, the presence of
fornia. Larvae forming linear mines which may
P. arnicae and arnicicola with a natural holarctic
greatly widen towards end. Male genitalia (Fig.
distribution suggests that this boreal dispersal must
1187) with few spinules and distiphallus pigment-
have taken place at least during the late Tertiary.
ed in apical bifid section.
P. tundrensis Spencer (196 9b, d. Griffiths, 197 4b).
Described from Alaska, widespread in NW Cana- Madieae (15 genera, 125 species)
da. Larvae forming large blotch mine. Male geni-
This small tribe is monophyletic and predominantly
talia (Fig. 1188) generally resembling P. arnici-
American. Only the single genus Madia in this as-
cola (Fig. 1186) but spinules fewer and weaker.
semblage has been colonized by the oligophagous
Calycomyza humeralis (see Astereae) in the United
Phytomyza robustella group (larvae pupating in
States, and the evidence for the validity of this tribe
leaf)
provided by agromyzid host selection is thus entire-
P. arnicivora Sehgal (1971b, d. Griffiths, 1974b). ly negative.
Alberta. Larvae forming narrow linear mines,
partly or mostly on lower leaf surface. Male geni-
Heliantheae sensu stricto (155 genera,
talia (Fig. 1189) with paired distal tubules short,
700 species)
broad, characteristic of this group.
P. campestris Griffiths (1974b). Alberta, Alaska. No consensus has yet been reached by botanists on
Larvae forming narrow linear mine. Male geni- the composition and limits of this tribe. Jeffrey, cited
talia (Fig. 1190) differing only in detail from by Mabberley (1987), included the four genera
P. arnicivora. Helenium, Tagetes, Bidens and Coreopsis within the
P.oreas Griffiths (1974b). Alaska and widespread Heliantheae. However, Bremer (1987) has attempt-
Table 76
0"1.".
. • · ·. . .:. . ·.~.·~~-v?:~
l.-__ '--. ..' _.'
.?
/~ . / / /
,~ ... ~//~ ....
\A~/'
1180 1181
1182 1183
! .
/
1186
1185
1187 1188
1189 1190
Subclass ASTERIDAE 311
ed to subdivide the tribe and in an arrangement genitalia (Fig. 1191) show that the two are not di-
which he concedes is provisional treats the Helian- rectly related. M. verbesinae Spencer (in Spencer
theae sensu stricto and Helenieae sensu stricto as and Steyskal, 1986) is a stem-borer in Verbesina
separate tribes, and has further removed 'four alternifolia and is only known in the eastern U.S.A.
monophyletic groups as separate taxa' - Madieae, The male genitalia (Fig. 1192) suggest that this is an
Heliantheae - Flaveriinae sensu lato, Tageteae and isolated species. M. spilanthis Spencer (1985b)
Coreopsideae (see below). (Figs. 1193, 1194) is only known as a seed-feeder in
I am in no position to judge the validity of such Spilanthes americana at Bogota, Colombia. It is
treatment but it is based on the most up-to-date, closely related to M. wedeliae (see below). M. biden-
detailed study of the family and it seems appropriate tis Spencer has also been recorded on Verbesina in
to follow it here. Florida (see Coreopsideae below).
It should be added that Bremer agrees that the Of the 18 oligophagous species, M. cunctanoides
Heliantheae sensu stricto are 'not clearly mono- Blanchard (1954) is only known in Argentina
phyletic.' It is a large, cosmopolitan, though mainly boring in the stems of Helianthus annuus and Ver-
American group. Nine genera of Agromyzidae have besina encelioides, and also in Anthemis (Anthemi-
colonized this tribe, with a total of 39 species, as in- deae), Gnaphalium (Inuleae) and Tagetes (Tage-
dicated in Table 77. The Agromyzidae on 31 host teae). Valladares (1982, unpublished thesis) has dis-
genera in the tribe are shown in Table 78. cussed this species in some detail and illustrated the
The sole record of larval feeding of Phytobia male genitalia and early stages. M. minima Malloch
on the Asteraceae is on the bush Oyedaea verbesi- (cf. Spencer and Stegmaier, 1973) has only been
no ides in 'tropical America, northwards to Mexico' reared from seeds of Wedelia paludosa and W trilo-
(Suss and Muller-Stoll, 1980 and Suss, pers. comm.). bata in Florida and I found a large population ob-
Eleven Melanagromyza species are known on this viously feeding on W calycina in Guadeloupe. Also
tribe, of which only three are tribe-specific. In Vene- I found a female on flowers of Eupatorium odora-
zuela M. polymniae Spencer (1973c) feeds in stems tum in Jamaica and I am satisfied that this wide-
of the tall plant, Polymnia riparia. Although this spread species in the Caribbean area also feeds on
species superficially resembles M. caerulea which this host. This species closely resembles M. mini-
feeds in seeds of Ipomoea spp. (cf. Fig. 721), the moides Spencer (cf. Spencer and Stegmaier, 1973)
which overlaps in distribution with M. minima and
Table 77 is a seed-feeder which favours the Heliantheae. It is
known on six genera - Borrichia, Heliopsis, Melan-
Agromyzid genera on Heliantheae
thera, Rudbeckia, Verbesina and Viguiera, and also
Genus Species on Aster (Astereae) and Helenium (Helenieae). The
two species are readily distinguishable both by the
Tribe-specific Inter- Poly-
tribal phagous male genitalia and the differing length of the female
ovipositor (M. minima: Figs. 1195, 1196 and
Melanagromyza 3 8 M. minimoides: Figs. 1197, 1198). It is interesting
Ophiomyia 1 that, although both species are seed-feeders, the
Agromyza male genitalia and ovipositor are substantially
Phytobia 1 (larval feeding only) -
Amauromyza 1
longer in M. minima and this must be an adaptation
Liriomyza 6 (1 undescribed) 4 4 to the differing structure of the flower buds where
Calycomyza 4 3 oviposition takes place, necessitating greater pene-
Chromatomyia 1 tration into the tissue in M. minima. The aedeagus
Phytomyza has evolved in length to match the female terminalia.
16 18 5 M. splendida Frick (cf. Spencer and Steyskal,
1986) is a greenish stem-borer, possibly favouring
Total 39
1191
(
-
.
-
-.
.
--~=--I
tt:~. /
-. -I
1195
~'~
.:. . . .'~'
".'·.1'.·.~
~ ~ _%?
.~
.'. .~.'.~. .'; :~". ".:;"'.)'; :. '."
c•':.'•..• . •.,-...• '...,.
' .•". '. '..• ... ' .,. ,•.. . . . ·!.. ',• .• .. . .:. .:. .·.·.;:.
~.~~
1197
~~·~··ii····
1200 1201
\\
i/ ,~
\~::.~::.~~::/'
1203
Subclass ASTERIDAE 313
hosts in the Heliantheae, reared on Ambrosia, misia douglasiana. This is an unusual host combina-
Helianthus and Parthenium, but has a remarkably tion and is the only species known on both Helian-
wide host range and is known on seven further theae and Anthemideae (apart from polyphagous
tribes: Lactuceae (Lactuca), Coreopsideae (Bidens), species and the highly oligophagous Amauromyza
Helenieae ( Gaillardia), Heliantheae-Flaveriinae maculosa). The distinctive genitalia of A. ambrosi-
(Flavena), Inuleae (Gnaphalium), Senecioneae vora (Fig. 1204) suggest relationship with A. anthra-
(Erechtites) and Tageteae (Tagetes). Also it shows cina which is host-specific on Urticaceae (Fig. 216).
limited polyphagy, having been reared from Apium This is further supported by the general similarity of
(Apiaceae) in Hawaii. The male genitalia (Fig. 1199) external characters. If this relationship is correct, it
are distinctive but characteristic of the genus. indicates a puzzling switch of an ancestral species
M. viridis (Frost, ct. Spencer and Steyskal, 1986) between Urticaceae and Asteraceae.
(= M. helianthi Spencer, 1963b) feeds in seed- The highly oligophagous Amauromyza maculosa
heads, primarily in genera of the Heliantheae, re- has been recorded here on Helianthus, Melanthera
corded hosts being Encelia, Helianthus, Tithonia and Zinnia but occurs most commonly on genera of
and Zinnia. It has also been found on Calendula Astereae (see above).
(Calenduleae). It occurs widely from California and Fourteen Liriomyza species are known on the
Florida to Chile. The genitalia (Fig. 1200) suggest Heliantheae, of which five plus one undescribed are
probable relationship with M. neotropica whose on- tribe-specific:
ly known host is Eidens.
L. assimilis (Malloch, ct. Spencer and Steyskal,
M. wedeliae Spencer (in Spencer and Stegmaier,
1986), a leaf miner on Helianthus in Illinois and
1973) feeds primarily in flower-heads of Wedelia
Pennsylvania, was later discovered in Texas feed-
paludosa in Florida. This species resembles most
ing on Lindheimera texana. In the genitalia the
closely M. spilanthis which is only known on
extended distiphallus (Fig. 1205) is distinctive.
W americana in Colombia but there are small but
L. debilis Sasakawa (ct. Sasakawa, 1961) is only
distinctive differences in the male genitalia (M. we-
known in Japan, forming linear-blotch mines on
deliae: Figs. 1201, 1202). Also recorded on Senecio
Helianthus debilis. No males have been found.
and Bidens.
L. helianthi Spencer (1981 and Spencer and Steys-
On Guadeloupe in February, 1987 I caught a
kal, 1986) is a leaf miner on Helianthus annuus
series of M. wedeliaphoeta Martinez (in Spencer et
and Xanthium strumarium (no details of leaf
al., 1990) mixed with M. minima on the bush Wede-
mine recorded), known only in California and
lia calycina on which the flies were clearly breeding.
Washington. The male genitalia (Fig. 1206) show
This species is immediately distinguishable from
that this species belongs to the pusilla group.
M. minima by having the squamal fringe white (not
L. melantherae Spencer (1959). Host Melanthera
black as in M. minima). The male genitalia are of the
?brownei, Mt. Cameroun. Larva forming linear
same general form as in M. minima but differ in de-
mine, reported as pupating in leaf. Male genitalia
tail.
(Figs. 1207, 1208) differentiated but adult not
Only a single Ophiomyia is known here, O. con-
distinctive.
spicua Spencer (1961b) which feeds on Eclipta and
L. synedrellae Martinez (Spencer et al., 1990). Host
Siegesbeckia in eastern Asia and north-east Aus-
SynedreUa nodiflora, known only in Guadeloupe
tralia. This entirely black species does not have the
and Venezuela, larva forming characteristic leaf
external characters of the genus (vibrissal fasciculus
mine along midrib and lateral veins (Fig. 1209).
in the male, antennae divided by a raised facial keel)
but the male genitalia (Fig. 1203) indicate its correct In Kenya a leaf miner (Fig. 1210) on Aspi/ia mos-
generic position in Ophiomyia. sambicensis appears to represent an undescribed
One of only 3 Agromyza species known on the species. The mines were not uncommon at two
Asteraceae, A. ambrosivora Spencer (1969b) was widely separated localities at Ruiru north of Nairobi
described from Canada (southern Ontario and Otta- and at Kibwesi west of Mombasa. I have also found
wa), forming leaf mines on Ambrosia artemisiifolia the mines on A. africana near Kampala, Uganda.
and was subsequently found in California on Arte- Of four inter-tribal species, L. zinniae Spencer
Fig. 1191. Melanagromyza polymniae: aedeagus.
Fig. 1192. Melanagromyza verbesinae: aedeagus.
Figs. 1193, 1194. Melanagromyza spilanthis: 1193, aedeagus, side view; 1194, same, ventral view.
Figs. 1195, 1196. Melanagromyza minima: 1195, aedeagus; 1196, female ovipositor.
Figs. 1197, 1198. Melanagromyza minimoides: 1197, aedeagus; 1198, ovipositor.
Fig. 1199. Melanagromyza splendida: aedeagus.
Fig. 1200. Melanagromyza viridis: aedeagus.
Figs. 1201, 1202. Melanagromyza wedeliae: 1201, aedeagus, side view; 1202, same, ventral view.
Fig. 1203. Ophiomyia conspicua: aedeagus.
314 Class MAGNOLIOPSIDA (Dicotyledons)
>~
1204 1205
1207
1208
1212
1213
Subclass ASTERIDAE 315
(1981), whose primary host is probably Zinnia ele- talia of these two species suggest that they are di-
gans, has also been reared from Calendula (Calen- rectly related (c. ecliptae: Fig. 1215, C. verbesinae:
duleae) and Heterotheca (Astereae). The male geni- Fig. 1216).
talia (Fig. 1211) suggest that this species belongs to Of the three inter-tribal species, C. frickiana has a
the pusilla group but the aedeagus is unusual in restricted host range, known in the Heliantheae on
having a particularly strong ventral sclerite. L. pusil- Helianthus and Rudbeckia but the primary host ap-
la has been found on Synedrella in India (see As- pears to be Bidens (see Coreopsideae below). The
tereae) and L. spencerella on Helianthus annuus in species belongs to the artemisiae group. The highly
Argentina (see Calenduleae). L. sabaziae Spencer oligophagous species, C. humeralis and C. platyp-
(1963c) was described from Sabazia urticaefolia in tera favour genera in the Astereae (see above) but
Brazil, the larva forming an irregular linear mine, both are found here on a number of genera (see
and I recently found it to be very common on Galin- Table 78).
soga caracasana in Colombia around Bogota. It is Cal. enceliae Spencer (1981) was caught on
also widespread in California and Washington and Encelia caliJornica in southern California in March,
has been recorded on four further tribes - Astereae 1977 but no mines were found confirming this as its
(Baccharis), Cardueae (Carduus, Cirsium, Silybum), host.
Helenieae (Dahlia) and Inuleae (Gnaphalium). The The polyphagous Chromatomyia horticola and
genitalia (Fig. 1212) show that this species belongs probably Chr. syngenesiae (as 'atricornis') occur on
to the pusilla group. several genera of the Heliantheae (Table 78).
The four polyphagous species, Liriomyza huido- The only Phytomyza known on the Heliantheae is
brensis, sativae, triJolii in the New World and L. stri- P. helianthi Sasakawa (1955d and 1961) in Japan,
gata in Europe have been found on a number of the larva forming mines on Helianthus tuberosus (in-
genera (Table 78). troduced from North America). The genitalia are
Seven Calycomyza species have colonized the typical of species in the albiceps group. Presumably
Heliantheae, of which four are tribe-specific. C. am- the species also feeds on one or other local Japanese
brosiae Frick (cf. Spencer and Stegmaier, 1973) hosts from which it has colonized Helianthus tuber-
forms blotch mines on Ambrosia artemisiijolia and osus.
is only known in southern states of U.SA. from The following unidentified leaf mines on the
Florida to California. Although the adults are simi- Heliantheae should be recorded:
lar to many species in the malvae group (Lamia-
1. Delilia martyni. Linear mines, possibly repre-
ceae), the male genitalia (Fig. 1213) are distinctive.
senting Liriomyza huidobrensis, north of San
Differences in the leaf mines between this species
Jose, Costa Rica.
and Agromyza ambrosivora have not been clarified.
2. Oyedaea verbesinoides, linear mine. Caracas,
C. melantherae Spencer (in Spencer and Stegmaier,
Venezuela. Liriomyza sp. (?).
1973) is a southern species known only in Florida
3. Smallanthus uvedalia, blotch mine. D.C., U.SA.
and the Bahamas, forming bloth mines on Melan-
Calycomyza sp.
thera deltoidea. The male genitalia (Fig. 1214) are
4. Viguiera sylvatica. Linear and blotch mines, Cos-
more differentiated than any others in the artemisiae
ta Rica - Liriomyza and Calycomyza (?).
group, to which this species appears to belong. Two
species are known only in South America. C. eclip- The relatively large number of Liriomyza and
tae (Spencer, 1963c) was described from the Botan- Calycomyza species on the Heliantheae contrasts
ical Gardens at Rio de Janeiro, Brazil and subse- with the single Phytomyza, known only in Japan.
quently found in Venezuela, the larvae forming ir- This generic colonization differs strongly from the
regular blotch mines on Eclipta alba. C. verbesinae Astereae (17 Phytomyza species), the Senecioneae
(Spencer, 1963c) is known only from near Sao Pau- (16) and the Helenieae (7) and is probably ex-
lo, Brazil, with Verbesina sp. as host. The male geni- plained by the largely New World, tropical distribu-
tion of the tribe. However, many genera occur in veria in this small assemblage has been colonized by
North America and it is puzzling that not a single Agromyzidae. In Florida the oligophagous Melana-
Phytomyza is known there on the Heliantheae. gromyza splendida (see Heliantheae) has been found
on Flaveria linearis and the polyphagous Liriomyza
trifolii on F. trinervia. Valladares (unpublished re-
Heliantheae - Flaveriinae sensu lata
port, 1986) found mines in Argentina on F. bidentis
(7 genera, 32 species)
which possibly represent L. trifolii.
Bremer (1987) recognises the above as a distinct Thus on Flaveria the agromyzids provide no con-
taxon within the Heliantheae sensu lato. Only Fla- firmation for the separate taxonomic status of this
Table 78
group. However, it would clearly be of interest to (Sehgal, 1971b: fig. 27) very similar to M splen-
discover leaf miners on the further American genera dida (which also feeds on Bidens but further
Clappia, Haploesthes, Jaumea, Pseudoclappia, Sart- south in U.S.A.).
wellia and Varilla which are included here. M. bidentis Spencer (cf. Spencer and Stegmaier,
1973). Host Bidens pilosa, also Verbesina
(Heliantheae), seed feeder, Florida, Venezuela.
Tageteae (16-18 genera, 240 species)
The long aedeagus (Fig. 1217) is characteristic of
The genera in this tribe are largely American but seed feeders but here the length is achieved by the
Tagetes has been widely introduced and is the only extension of the distiphallus, in contrast to
genus in the tribe known to support Agromyzidae. M. minimoides where comparable length results
In Argentina Melanagromyza cunctanoides has from a long membranous gap between the basal
been reared from stems of Tagetes sp., the primary sclerites and the distiphallus complex (Fig. 1197).
host probably being Helianthus annuus. Also in Ar- M bidentis is distinguishable from M floris
gentina the oligophagous Amauromyza maculosa which also feeds on seeds of Bidens by the
and polyphagous Liriomyza huidobrensis, and in squamal fringe being dark, silvery-white in
U.S.A. L. tri/olii have also been found attacking Mfloris.
Tagetes. M. floris Spencer (1963c). Hosts: Bidens, Calen-
dula, seed feeder widespread in Caribbean area,
including Mexico, Venezuela and Guadeloupe.
Coreopsideae (26 genera, 500 species)
Male genitalia compact (Fig. 1218), contrast
This tribe is cosmopolitan but predorninantlly M bidentis .
American. Only the two genera Coreopsis and M. neotropica Spencer (1963c). Host Bidens, seed
Bidens support Agromyzidae. Coreopsis is native in feeder, widespread throughout much of South
North America but no agromyzids are known on it America, from Argentina and Chile to Vene-
there. It is widely cultivated and is attacked by the zuela. This is one of the large complex of greenish
polyphagous species, Liriomyza strigata and also species with the squamal fringe pale, which can
Chromatomyia syngenesiae in Europe, by Chr. horti- only be positively identified by the male genitalia
cola in India and Chr. syngenesiae in Australia. (Fig. 1219). It is possibly related to M viridis (see
Bidens appears to be particularly attractive to Heliantheae above, Fig. 1200).
Agromyzidae and supports six species which are M. splendida Frick (cf. Spencer and Stegmaier,
host-specific and 12 others which are oligophagous 1973). Hosts Bidens and other genera in the Heli-
or polyphagous in five genera as shown in Table 79. antheae (see above) and its subordinate tribes,
The 18 species on the tribe are listed in Table 80. stem-borer. Widespread from Mexico and the
Caribbean area to N ew York and Michigan. Male
Table 79 genitalia (Fig. 1199) distinctive, very similar to
M. bidenticola and the two species appear to
Agromyzid genera on Bidens have a partially overlapping distribution.
Host-specific Inter- Poly- M. wedeliae Spencer (in Spencer and Stegmaier,
tribal phagous 1973). Normal host Wedelia (see Heliantheae),
also found on Senecio, reared once from Bidens
Melanagromyza 2 5 pilosa in Colombia (Spencer, 1985b).
Amauromyza 1
Liriomyza 3 (1 undescribed) 2 The highly oligophagous Amauromyza maculosa
Calycomyza 1 2 is not uncommon on Bidens but the blotch mines
Chromatomyia may be mistaken for those of Calycomyza allecta, at
6 10 2 least when old.
Seven Liriomyza species attack Bidens, of which
Seven Melanagromyza species are known to feed three are host-specific on this genus.
on Bidens, either in the flower-heads or stems. All
Liriomyza archboldi Frost (cf. Spencer and Steg-
but one are restricted to the New World.
maier, 1973). Host Bidens, larva forming blotch-
M. metallic a (cf. Spencer, 1985a). Widespread in mine. Described from Florida, also now known in
the Old World tropics and the only known hosts the Bahamas and Costa Rica. Male genitalia (Fig.
are Ageratum and Bidens (cf. Eupatorieae, Figs. 1220), distinctive, suggesting that this is an isolat-
1092,1093). It is particularly common on Bidens ed species.
in Papua New Guinea. L. insignis Spencer (1963c). Host Bidens pilosa, lar-
M. bidenticola Sehgal (1971b). Host: Bidens cernua, vae feeding in seeds in Costa Rica (Spencer,
Edmonton, Alberta. Stem-borer. Male genitalia 1985b). Also known in Brazil. This is an unusual
318 Class MAGNOLIOPSIDA (Dicotyledons)
1216
1215
y·::·~'····"c"Ai.~<t:fy#'
1217 1218
(\.~. • ·. .c:(C
JJ
1219
1220
1223
1225
1224
1226 .• .....
Subclass ASTERIDAE 319
species, with the ovipositor in the female greatly a more northern species, with no records south of
elongated (for oviposition within the young seed- Virginia and Pennsylvania.
heads), the mesonotum yellow centrally. The C. platyptera (Thomson). Highly oligophagous, oc-
male aedeagus (Figs. 1221, 1222) is somewhat curring most commonly on genera of the Helian-
extended, as would be expected with such a long theae (see above).
female ovipositor. Chromatomyia horticola is known on Bidens in the
Canary Islands, Kenya and India.
Only three other Liriomyza species are known to
Phytoliriomyza bidensiphoeta Spencer (1977a) was
feed in buds or seeds, L. urophorina in Lilium mar-
described from a single male I caught on Bidens
tagon, L. wachtli in Veratrum and L. lutea in Hera-
pilosa in Papua New Guinea. Although this speci-
cleum and Pastinaca (Apiaceae), all in Europe.
men may have found Bidens attractive (hence the
L. pusilla (Meigen). Recorded by Sasakawa (1972)
name), this cannot be taken as firm evidence that
in Taiwan on Bidens biternata. Highly oligo-
Bidens was the host.
phagous but most frequent on Astereae (see
above).
Species requiring clarification:
L. spencerella Valladares (1984). Host Bidens, also
Calendula (Calenduleae), Chrysanthemum (An- 1. Melanagromyza kenyensis (Spencer, 1959). This
themideae) and Helianthus (Heliantheae). Larva species was described from a series from Kenya
forming narrow linear mine (discussed above (Naivasha, Nairobi) 'ex Composite 7' (van
under Calendula). Someren, H.J. Turner) and a further series 'ex
L. trifolii (Burgess). Common on Bidens, see above Croton' (Euphorbiaceae, see above). It is im-
under Fabaceae. probable that a species would have hosts in the
L. venegasiae Spencer (1981). Recorded on Bidens Euphorbiaceae and Asteraceae but Van
in California but normal host Venegasia (see Someren is remembered as being meticulous in
above under Helenieae). the accuracy of his data, which can only be ac-
Liriomyza sp.n. Host Bidens near pilosa, Tucuman, cepted as correct. The genitalia of M. kenyensis
Argentina (Valladares, unpublished report, (Figs. 549, 550) are atypical of other species
1986). The description will be published by Val- known to feed on Asteraceae in Africa but as the
ladares in due course. species is black rather than green (as in most
stem-boring species), the unusual genitalia are
Of three Calycomyza species known on Bidens
not surprising. Further clarification of the hosts
only one appears to be host-specific and t\yo are
of this species is obviously desirable.
oligophagous.
2. Calycomyza compositana Spencer (1973c) was
C. allecta Melander (cf. Spencer, 1963c). Host described from an unidentified composite in Ve-
Bidens, described from St. Vincent, also known in nezuela. The mined leaf (Fig. 1229) gives no in-
Brazil and Guadeloupe. The white blotch mines dication of the tribal affiliation but the male geni-
were particularly common in Guadeloupe on talia (Figs. 1227, 1228) suggest that the species
B. pilosa in Jan/Feb, 1987. The male genitalia are belongs to the artemisiae group.
distinctive (Figs. 1223, 1224).
Time has not permitted a detailed analysis of the
C. frickiana Spencer (in Spencer and Steyskal,
relationships of the 295 species discussed here on
1986). Hosts: Bidens, Helianthus, Rudbeckia.
the Asteraceae but some facts are worthy of note:
The primary host appears to be Bidens, the larva
forming a large blotch mine. C. frickiana was con- 1. The close relationship between species in the
fused with C. allecta until the genitalia were Phytomyza spondylii group on Apiaceae with
examined, showing the two species to be entirely those of the albiceps group on the Asteraceae
distinct (c. allecta: Figs. 1223, 1224; C. fricki- was pointed out earlier. The albiceps group oc-
ana: Figs. 1225, 1226). C. allecta is a Caribbean curs widely on the family, with representatives
and South American species, while C. frickiana is on eight tribes and with particular proliferation
1227 1228
1229
Figs. 1227-1229. Calycornyza cornpositana: 1227, aedeagus, side view; 1228, distiphallus, ventral view; 1229, leaf mine on Compo sitae
sp.
Table 80
Heliantheae - Flaveriinae
Flaveria Melanagromyza splendida; Liriomyza trifolii
Tageteae
Tagetes Mel. cunctanoides; Amauromyza maculosa; L. huidobrensis, L. trifolii
Coreopsideae
Bidens Mel. bidenticola, Mel. bidentis, Mel. floris, Mel. metallica, Mel. neotropica,
Mel. splendida, Mel. wedeliae; Am. maculosa; L. archboldi, L. insignis,
L. pusilla, L. spencerella, L. trifolii, L. venegasiae, Liriomyza sp. n.
(Argentina); Calycomyza allecta, Cal. frickiana, Cal. platyptera;
Chromatomyia horticola
Coreopsis L. strigata; ChI. horticola, ChI. syngenesiae
on the Astereae, Senecioneae and Anthemideae. florae. Whatever the degree of relationship between
Interestingly, the isolated tribe Echinopsideae the Apiaceae and Asteraceae, it is clear that the Phy-
has its own host-specific species in this group, tomyza species in the spondylii and albiceps groups
P. bipunctata known only on Echinops. have evolved from a switch of an ancestral species
from the Apiaceae to the Asteraceae and this was
Although the Asteraceae are chemically very dis-
tinct, there are striking similarities with the Apia- clearly facilitated by the chemical similarity (in some
respects) between the two families.
ceae and this leads Hegnauer in Heywood et al.
(1977) to postulate an evolutionary line Magnoliales 2. The similarity between Melanagromyza arni-
-> Rutales -> Umbellales -> Asterales. Com- carum (Arnica) and M. polemonii (Pole-
menting on the chemical relationship between the monium, Boraginaceae) may well not be due
Asteraceae and Apiaceae, Cronquist (1981) writes: merely to convergence. Hegnauer (loc. cit.) sug-
'The family that may be chemically most nearly like gests that a relationship between Polemoniales
the Asteraceae is the Apiaceae, in the subclass Rosi- and Asterales seems possible on chemical
dae, but the morphological differences make any at- grounds.
tempt to link these two families patently absurd.' 3. Hexomyza centaureae (gall-causer on Centaurea)
Here it may be mentioned that Thorne (1983) is clearly related to H. sarothamni (gall-causer
places the Apiaceae in his Superorder Corniflorae on Cytisus, Leguminosae).
close to the Asteraceae in his Superorder Asteri- 4. Amauromyza carlinae is related to Am. lamiil
Subclass ASTERIDAE 321
Am. leonuri on Lamiaceae and probably repre- tribes appear to the agromyzids almost as well dif-
sents a switch of an ancestral species from La- ferentiated as families in other orders. Nevertheless
miaceae to Asteraceae. the 43 species which cross tribal boundaries (15%)
5. The close relationship between the Campanula- is not inconsiderable. It could be of interest to
ceae and Asteraceae is indicated by one of the analyse the subfamily specificity (in the two sub-
few cases of an agromyzid feeding on two fami- families as discussed earlier in the introduction). If a
lies. Ophiomyia heringi occurs on a number of higher degree of specificity was discovered than in
genera both in the Campanulaceae and particu- individual tribes, this would provide some evidence
larly in the Lactuceae. supporting the division of the Asteraceae into two
subfamilies.
Here I again emphasise the inacceptability of the
It will be seen from Table 52 that the number of
Asteraceae to the relatively large genus Agromyza
tribe-specific species is comparable in the Lactuceae
which is represented merely by three species (Table
(42), Astereae (36) and Anthemideae (36) and is
52). In contrast, the radiation of both Liriomyza (67
closely followed by the Senecioneae (32). However,
species) and Phytomyza (87) is striking. It is perhaps
the most acceptable tribe, including also inter-tribal
surprising that the tribe in which the majority of gen-
and polyphagous species, is the Astereae, with 60
era have latex (Lactuceae) should support the larg-
species. The number of species colonizing a tribe is
est number of tribe-specifric Agromyzidae (42),
not merely a function of the size of the tribe. Admit-
plus a further 14 which are inter-tribal or poly-
tedly, the Astereae with 2500 species is the largest
phagous. This indicates that the latex in no way acts
tribe and does support more agromyzids than any
as a deterrent to leaf miners, although the virtual ab-
other tribe. However, surprisingly the Lactuceae
sence of the internal-boring Melanagromyza species
with the largest number of tribe-specific species
is probably due to the discomfiture or even risk of
comprises only some 1000 plant species and is only
drowning by feeding in stems and rupturing the
the eighth in size. Subtle factors, which we do not
latex-bearing vessels.
begin to understand but which are doubtless related
The remarkably high degree of tribal specificity,
to their chemistry, have clearly influenced the agro-
with 241 of the 295 species known on the family
myzids in their choice of host in the past and the
rigidly restricted to a single tribe is indicated in the
degree to which subsequent radiation has occurred,
consolidated Table 52. This confirms in general
although admittedly randomness has also certainly
terms the taxonomic judgment of botanists in their
been a significant factor.
circumscription of the tribes within the family. The
B. Class LlLIOPSIDA (Cronquist, 1981)
MONOCOTYLEDONS (Dahlgren et al., 1985)
Traditionally the monocots have been treated as an ales, while the Poaceae are placed in a separate
evolutionary group distinguished from the dicots order Poales. In both proposals these families are in-
by the presence of only a single cotyledon. Dahl- cluded in the same higher category - Commelini-
gren et al. (1985) add a second condition of the florae of Dahlgren and Commelinidae of Cronquist.
mono cots - the presence of sieve tube plastids and It seems logical here to follow Dahlgren et al.
state: 'Thus with these two character conditions the (1985), as this is the most recent, specialist work on
monocotyledons can be unequivocally defined.' the monocotyledons as a whole, and the very high
They then list 14 characters by which 'most or all degree of host specificity apparent in the monocot-
monocotyledons differ from all or most of the dico- feeding agromyzids fits well into their arrangement.
tyledons.' However, they go on to add: 'We shall, Although it has been argued that the monocotyle-
however show that certain dicotyledonous fami- dons are more ancient than the dicots, it is now
lies ... possess a considerable number of the mono- widely accepted that the mono cots evolved from
cotyledon attributes [enumerated above]. Likewise, primitive dicots. Fossil pollen suggests that the
certain monocotyledons ... each resemble various monocots diversified from the dicots in early to
different families that are assigned to the dicotyle- mid-Cretaceous. The first pollen which can almost
dons. In some cases similarities between two taxa on certainly be ascribed to the monocotyledons has
either side of the moncotyledonldicotyledon border been described as Liliacidites from the Albian
can be so strong that we are still in serious doubt as (some 115 million years ago). By Upper Cretaceous
to the distinctness of the monocotyledons. We have time fossils referable to mono cots appear increas-
also avoided classifying monocotyledons and dico- ingly but still cannot be referred unequivocably to
tyledons as units of definite rank, such as class, in any modern groups. It is thus certain that by the time
this presentation.' Agromyzidae had evolved and were beginning to
There are now significantly differing views among diversify in late Cretaceous or Upper Tertiary
botanists on the most logical system of classification various groups of monocots were already in exis-
of the monocots. Cronquist (1981) divides his Liliop- tence and available for colonization.
sida class into five subclasses, treating the Alisma-
tidae as the most primitive, although conceding that Agromyzidae have colonized six of Dahlgren's ten
they are not on the main line of evolution of the superorders, with a total of 195 mono cot-specific
class. He follows with the Arecidae, including the species with known hosts (plus about 155 whose
Arecales (palms, on which no agromyzids are hosts can be deduced as feeding on Poaceae from
known) and the Araceae, continuing with the Com- their male genitalia) and also eight polyphagous spe-
melinidae and ending with the Liliidae. Dahlgren's cies known on dicots. The 12 genera represented
more recent treatment results in the acceptance of are indicated in the Tables below under each
more, smaller, more homogeneous and monophyle- Superorder. Eleven of the genera are present on the
tic families arranged according to their presumed re- Liliiflorae but the number of species on this
lationships. Their scheme claims to 'avoid the con- Superorder is small (34 + 8 polyphagous species
straints of mere convention.' Dahlgren divides the known also on dicots). The greatest proliferation has
Monocotyledons into ten Superorders, treating the occurred on the Commeliniflorae, with 56 species
Liliiflorae as the most primitive, then largely follow- on the Cyperales and 107 (+ 155) on the Poales. The
ing the same sequence as Cronquist, but placing the large number on the Poaceae (approx. 262) in part
Areciflorae (palms) at the very end. An important reflects the large size of the family, with about
difference between the two schemes which is rele- 10,000 species (Clayton and Renvoize, 1986).
vant to agromyzid host selection is the inclusion by However, the number of available hosts is certainly
Cronquist of the Cyperaceae and Poaceae in the not the only criterion for the radiation of agromy-
same order Cyperales, while Dahlgren includes the zids, as is clear from the mere seven species known
Juncaceae with the Cyperaceae in his order Cyper- on the orchids which are now estimated to have
MONOCOTYLEDONS 323
some 20000 species (see below). The discussion of the monocot-feeding agromy-
Consideration of the agromyzid genera repre- zids below does not answer all the questions and
sented on the mono cots is to some extent revealing solve all the problems relating to this intriguing
and does support the traditional division of the group, but it does show the complex pattern of the
angiosperms into two classes. In the discussion be- host relationship of these 195 species and I hope
low of the individual agromyzid genera on their dif- will provide a useful basis for further investigations.
ferent plant hosts an attempt has been made to
postulate their ancestral dicot-feeding sister-groups.
In some cases this is apparent but in other groups, Superorder LlLIIFLORAE
with obviously early colonization of the monocots, This complex, as defined by Dahlgren et al. (1985),
this relationship has been obscured by time. Three comprises five orders. Dahlgren's treatment of his
genera occur only on monocots. Cerodontha, with Liliiflorae differs substantially from Cronquist'S
seven subgenera and some 82 reared species is vir- Subclass Liliidae. In particular the Asparagales are
tually cosmopolitan and is restricted to the mono- treated as an enlarged order of 31 families, many of
cots on the four families Poaceae, Cyperaceae, Iri- which are included in the Liliales by Cronquist. This
daceae and Juncaceae. It represents an early assem- rearrangement is briefly discussed below under the
blage, closely related to Liriomyza, which has a prim- orders Discoreales, Asparagales and Liliales (which
itive species, L. commelinae restricted to the Com- include the orchids).
melinaceae in the southern U.S. and South America. Dahlgren's Liliiflorae include the 'extant forms
This has retained several important characters of among the monocotyledons that combine the major-
Cerodontha (see below). Penetagromyza aloes in ity of the features here considered to be ancestral. A
South Africa with its only host Aloe petricola is a number of these features are also found in Magnolii-
particularly interesting species, having characters of florae, and we postulate that the monocotyledons
both Melanagromyza and Ophiomyia and it can on- and the Magnoliiflorae are sister-groups in the evo-
ly be construed as a relict from an ancestor before lutionary sense. Most of the presumably ancestral
the division into the two genera Melanagromyza and attributes under this assumption are concentrated in
Ophiomyia. Metopomyza, with the single species Dioscoreales.'
M. ornata on Butomus, and others on Cyperaceae, The Liliiflorae as here circumscribed have been
Juncaceae and Poaceae, I consider to represent a colonized by a mere 31 described species of Agro-
diversification on monocots from an ancestral fern- myzidae (plus eight polyphagous species), while the
feeding species in Phytoliriomyza. Superorder (with the orchids) has some 30,000
Agromyza which feeds on 18 dicot families is well plant taxa. Eleven genera of agromyzids are repre-
represented on Poaceae. It has two groups, with sented as shown in Table 81. The host genera in the
three species which have switched almost certainly 11 families attacked with their agromyzid species
from Rosaceae, while the majority, with remarkably are listed in Table 84.
uniform male genitalia, are possibly related to
modern species on Urticaceae. A unique case of the
single species Agromyza abutilonis on Malvaceae in
East Africa has genitalia indistinguishable from the
monocot-ieeding type and appears to represent an Table 81
isolated backwards switch from mono cots to dicots. Agromyzid genera on Liliiflorae
Species in the other eight genera on mono cots are
present in the main in small numbers and represent Genus Species restricted Polyphagous
isolated switches, which have resulted in little subse- to the superorder species
quent radiation. The most successful genus in this
Penetagromyza 1
group is Chromatomyia which now has 23 species Melanagromyza 6
on Poaceae. The situation with the orchids is puz- Ophiomyia 3
zling. This is by far the largest family of monocots Kleinschmidtimyia 1
and has merely one host specific Tropicomyia on Tropicomyia 4 (+ 1 undescribed) 4
Amauromyza 1
Eulophia in Kenya (probably related to Tr. haeman-
Cerodontha sg.
thi on Iridaceae in South Africa) and six further spe- Dizygomyza 5
cies, all in Melanagromyza. One in Florida repre- Liriomyza 7 (+ 1 undescribed) 3
sents a switch from one or other of the large groups Napomyza 1
of greenish species known on a number of dicot Chromatomyia
families, while the other five all appear to be closely Ptochomyza 2
gen. indet. - (+ 1 undescribed)
related and belong to a single group, with species
known in Brazil, Kenya, Japan and Australia. 31 (+3 undescribed) 8
324 Class LILIOPSIDA
Agavaceae (8 genera, 300 species) into the root where it pupates. The large size (wing
length up to 3.5 mm) and the strongly projecting
This family is widely distributed in the Old and New
frons (Fig. 1233) suggest that this is a primitive spe-
Worlds and pollen is known from the Eocene. The
cies and the genitalia (Fig. 1234) are not atypical of
Agavaceae occur mainly in arid regions where
Napomyza. I consider that this is a relict, possibly
Agromyzidae would not be expected. .
with the transfer to Allium from an ancestor on the
In Papua New Guinea mines of a large .TroP1C~
Ranunculaceae, on which 14 Napomyza species are
myia sp. were common wherever Cordylme frU:1-
known.
cosa was cultivated. This plant is frequently found m
Chromatomyia horticola occurs on Allium in
gardens and also aroun~ .burial. gr?~nds, as it ap-
England and leaf mines on this host found in India
pears to have some relIgIOus sIgmfIcance. Unf~r
and Japan are also probably referable to this spe-
tunately I only obtained two females but I am satIs-
cies.
fied that the species is undescribed.
1230
1232
"~'\'~ .'.':.'
.'".'.,./,.',/'...
..•.'. ,,),
1235
1237
~~";~)f~~~
Q(~~I~~~·_j;~"~~~~o
MONOCOTYLEDONS 327
is usual in the genus. I cannot suggest the most close- ily which have been colonized by Agromyzidae.
ly related dicot-feeder. This treatment differs from that of Cronquist (1981)
0. asparagi Spencer (1964c) is a smaller species who includes Aloe in a separate family Aloeaceae
known only in the Mediterranean area, forming and Asphodelus within the Liliaceae. This is of no
stem mines on wild Asparagus acutifolius in addition significance for interpreting the host selection of the
to cultivated A. officinalis. In this species the male Agromyzidae, as each genus has its own agromyzid
has the vibrissal fasciculus typical of most species in species and how these genera are classified is not
the genus (Fig. 1238) and the genitalia (Figs. 1239, relevant to this study. I nevertheless follow Dahl-
1240) are highly asymmetrical. It is unlikely that As- gren's arrangement.
paragus has been colonized twice and I suggest that Asphodelus is a small genus of 12 species occur-
0. asparagi is the more primitive species. ring from the Mediterranean to the Himalayas. In
Two species in the small genus Ptochomyza feed southern Spain near Algeciras I found leaf mines on
in the leaves and stems of Asparagus species. Pt. as- A. microcarpus on 24.iv.1955 and obtained only a
paragi Hering (cf. Spencer, 1973a) is only known on single female described as Liriomyza asphodeli
A. officinalis the larvae feeding in the fine leaves and Spencer (1957d). In the absence of males, its affilia-
more rarely in the stem, pupating internally. It has tions cannot be clarified.
been found in large numbers in northern and eas- A species reared by Dr. K. Munro from Aloe
tern Germany and was also found at Peking (Buhr, petricola in his garden in Pretoria, South Africa
1960) in leaves of asparagus among flowers present- from plants whose natural habitat was eastern
ed to a visiting East German Delegation. The head Transvaal was described in the monotypic genus
(Fig. 1241) has the orbits strongly projecting and the Penetagromyza as P. aloes Spencer (1959). This is
wing venation (Fig. 1242), with the second costal unique in the family, with the post-vertical bristles
section unusually short, is characteristic of the entirely lacking. It also has a unique combination of
genus. The genitalia (Fig. 1243) are greatly simpli- characters: colour largely black, including halteres,
fied. The second species, Pt. asparagivora Spencer but frons reddish and protruding, wing with costa
(1964a; 1985) feeds on wild, climbing Asparagus ending at vein R4+ 5 and outer cross-vein lacking.
species, incl. A. acutifolius, A. ajricanus, A. falcatus The male genitalia, with the aedeagus membranous
and A. stipularis and has a wide distribution from and asymmetrical (Figs. 1244, 1245), appear char-
the Mediterranean to Ethiopia (Addis Ababa), acteristic of Ophiomyia. However, larval characters,
Kenya (Nairobi) and Pakistan (Rawalpindi). It with the posterior spiracles with an ellipse of eight
seems probable that Asparagus has been colonized pores around a well-marked central scar, each on a
once, with subsequent allopatric speciation. Surpris- well-defined protuberance, are entirely typical of
ingly, it was discovered during the course of this Melanagromyza. Also larval feeding deep within the
study that the stem miner on Clematis in Europe thick leaf and internal pupation immediately below
described as Phytagromyza mayeri also belongs in the epidermis is very similar to that of Melanagro-
Ptochomyza (see above under Ranunculaceae). It is myza cyrtorchidis on the succulent epiphytic orchid
difficult to find any logical explanation for the switch Cyrtorchis arcuata in Kenya (see below).
which has occurred between Ranunculaceae and Hering (1968: 267) considered the structure of
Asparagaceae but this must have occurred early and the larval cephalo-pharyngeal skeleton to be highly
with the wide-ranging distribution of the two Pto- primitive, indicating possible relationship with
chomyza species on Asparagus, it is suggested that Ophiomyia. I therefore postulate that Pen. aloes
these are not of recent origin. The exact phylogen- represents a relict which has retained plesio-
etic position of Ptochomyza is an intriguing problem morphous characters in isolation on its unusual
which requires further study. host, the ancestral population having colonized
Aloe from an ancestor of the Melanagromyza-
Ophiomyia complex before the differentiation into
Asphodelaceae (18 genera, 750-800 species)
the two separate genera we recognise today. It may
Dahlgren et al. (1985) divide this family into two be added that in a number of modern species the
subfamilies, the Asphodeloideae and Alooideae; division between these two genera is not entirely
Asphodelus and Aloe are the two genera in the fam- clear, but their generic assignment is frequently pos-
1241
1243
1244 1245
1246 1247
1248 1249
......../
Figs. 1241-1243. Ptochomyza asparagi: 1241, head; 1242, wing; 1243, aedeagus.
Figs. 1244, 1245. Penetagromyza aloes: 1244, aedeagus, side view; 1245, same, ventral view.
Figs. 1246, 1247. Ophiomyia kwansonis: 1246, head; 1247, aedeagus (Sasakawa).
Figs. 1248, 1249. Kleinschmidtimyia dianellae: 1248, aedeagus, side view; 1249, same, ventral view.
Fig. 1250. Liriomyza sp.: leaf mines on Bomarea sp.
MONOCOTYLEDONS 329
1251
1253
1255
1256
1258
1261
MONOCOTYLEDONS 331
lin. The larvae may feed singly but up to eight have tionally been divided into three subfamilies, al-
been found in a single bud; pupation takes place on though the classification has been split further by
the ground. The female has an extended ovipositor Dahlgren et al. (1985) into five subfamilies. Of the
(Fig. 1253) for penetrating into the bud tissue and six agromyzids known on the family, five are found
the male genitalia (Fig. 1254) are comparably elon- only on the lridoideae (primarily Iris but also
gate, with the additional length resulting from an ex- Belamcanda), while Cerodontha (Dizygomyza) ireos
tension of the basal sclerites rather than of the disti- has a wider host range, having been found, probably
phallus or as a gap between the basal sclerites and exceptionally, on the lxioideae ( Gladiolus). A
distiphallus, as is found in some other species, such record of this species on Typha in Finland is reject-
as Melanagromyza minimoides (Asteraceae). Al- ed as a misidentification (Hering, 1957b) and a
though the majority of Liriomyza species are leaf record of Cer. (D.) thompsoni on Typha in U.S.A. is
miners, these are certainly congeneric with L. uro- similarly the result of a misidentification (see below).
phorina which is the type-species of the genus. Amauromyza (?subg. Cephalomyza) belamcan-
L. wachtli Hendel (cf. Spencer, 1976a) was stated dae Sasakawa (1981b) forms linear mines on
to have been bred from seed-capsules of Veratrum Belamcanda chinensis in Honshu, Japan, the larva
album and V nigrum in Austria. I have also seen pupating externally. The species is distinctive in hav-
specimens from Finland and Archangel in northern ing the third antennal segment angulate and the
U.S.S.R. (possibly introduced with cultivated Vera- outer cross-vein lacking. While the male genitalia
trum). Surprisingly L. veratri Groschke (1955), de- (Figs. 1262, 1263) possibly indicate inclusion ofthis
scribed as a leaf miner on V album in southern species in the subgenus Cephalomyza, as Sasakawa
Germany, proved to be identical with L. wachtli and writes: 'this is clearly an isolated species and prob-
the synonymy was established by Spencer (1976a). ably deserves separate subgeneric status.'
It is puzzling that the same species should have lar- Five species in Cerodontha, subgenus Dizygo-
vae capable of feeding in seeds and also as a normal myza, feed on Iris.
leaf miner (Fig. 1257), and further clarification of
C. (D.) ireos (Goureau, cf. Spencer, 1976a). Host
the biology of this species is desirable.
primarily Iris pseudacorus in Europe, in Japan
In Canada and northern California L. smilacinae
also local Iris species, and in addition Belamcan-
Spencer (1969b) feeds on Smilacina stellata in
da chinensis (Sasakawa, 1961). The larvae feed
Alberta forming leaf mines (Fig. 1260) which are
singly, forming a white mine down the leaf (Fig.
very similar to those of L. wachtli. This is not fortui-
1265), pupating vertically (contrast C. (D.) iridis).
tous, because the genitalia of L. smilacinae (Figs.
The male genitalia (Fig. 1264) indicate closer re-
1258, 1259) differ only slightly from those of
lationship with C. (D.) iridophora and C. (D.)
L. wachtli (Figs. 1255, 1256) and the two are clear-
thompsoni in the U.S. and with several Carex-
ly sister-species, representing a species pair each
feeders in Europe (cf. Figs. 1327, 1329) than with
side of the Bering Straits. Pollen considered to rep-
the second Iris feeder in Europe, C. (D.) iridis.
resent the Liliaceae is known as early as Upper
The record of C. (D.) ireos on Gladiolus in
Eocene and thus the early speciation of L. wachtli
Romania appears to represent a unique attempt
and L. smilacinae from a common ancestor in
by an adventurous female to colonize this host. I
northern Eurasia is entirely plausible.
have seen the only mine known, which is incom-
A mine of an unidentified species of Agromyzi-
plete and the larva has died early in its life, ob-
dae has been found on Maianthemum trifolium near
viously finding Gladiolus unacceptable and
Moscow (Hering, 1957b). I have seen this mine
lethal.
which clearly represents an undescribed species
C. (D.) iridicola (Koizumi, cf. Sasakawa, 1961). Host
(Fig. 1261).
Iris ensata, known only in Japan. No male is
known but the species is probably related to C.
Iridaceae (70 genera, 1400 species) (D.) ireos.
C. (D.) iridis (Hendel, cf. Hendel, 1931). Host Iris
Although this family is best represented in the
foetidissima, known only in Europe, severallar-
southern hemisphere, particularly in South Africa,
vae feeding together, pupating laterally side by
Agromyzidae have colonized it only in Europe,
side across the leaf (Fig. 1267). The male genitalia
Japan and North America. The Iridaceae have tradi-
Figs. 1251, 1252. Tropicomyia gloriosae: 1251, aedeagus, side view; 1252, same, ventral view.
Figs. 1253, 1254. Liriomyza urophorina: 1253, female ovipositor; 1254, aedeagus.
Figs. 1255-1257. Liriomyza wachtli: 1255, aedeagus, side view; 1256, same, ventral view; 1257, leaf mine on Veratrum album.
Figs. 1258-1260. Liriomyza smilacinae: 1258, aedeagus, side view; 1259, same, ventral view; 1260, leaf mine on Smilacina stellata.
Fig. 1261. sp. indet.: leaf mine on Maianthemum (Moscow).
332 Class LILIOPSIDA
1262 1263
1264
1265
1268
Figs. 1262, 1263. Amauromyza belamcandae: 1262, aedeagus, side view; 1263, same, ventral view (Sasakawa).
Figs. 1264, 1265. Cerodontha (Diz.) ireos: 1264, aedeagus; 1265, leaf mines on Iris pseudacorus (Hering).
Figs. 1266, 1267. Cerodontha (Diz.) iridis: 1266, aedeagus; 1267, leaf mine on Iris sp. (Nowakowski).
Figs. 1268, 1269. Cerodontha (Diz.) iridophora: 1268, aedeagus; 1269, posterior spiracles of puparium.
MONOCOTYLEDONS 333
(Fig. 1266) suggest close relationship with Carex the world orchids and treat the group as a single
feeders, such as C. (D.) luctuosa (Fig. 1304) and family. Here they are included as one of ten families
C. (D.) hirtae. in the order Liliales.
C. (D.) iridophora Spencer (in Spencer and Steg- The Agromyzidae have only succeeded in colon-
maier, 1973). Host Iris sp., Florida. The larva izing nine genera, with seven species, in two of the
mines towards the base of the leaf where it four subfamilies (tribes of Cronquist) into which the
pupates. The posterior spiracles are very distinc- family is divided. Six of the species are in Melana-
tive (Fig. 1269) with the three pores elongate and gromyza, a genus not represented elsewhere on the
free from each other. The male genitalia (Fig. monocots, with one Tropicomyia which is probably
1268) indicate close relationship with C. (D.) related to Tr. haemanthi (Iridaceae, see above).
thompsoni.
Melanagromyza cyrtorchidis Spencer (1985a).
C. (D.) thompsoni (Frick, d. Spencer and Steyskal,
Hosts Cyrtorchis arcuata and Rangaeris amanien-
1986). Host Iris versicolor, eastern U.S., also On-
sis (Kenya), also present in Uganda, the larvae
tario and Quebec. The plant reacts to larval feed-
feeding deep within the thick, succulent leaves,
ing by producing a gall around the pupation site.
pupating near the lower surface. The male geni-
Claassen (1918) gives a detailed description of
talia (Fig. 1270) are not atypical of Melanagro-
the biology of this species with 20 illustrations.
myza. No other Melanagromyza is known to feed
It has recently been confirmed that a record of C. exclusively in leaves but the larval feeding in the
(D.) thompsoni on Typha in the U.S.A. is a result of succulent leaves of the host approximates to the
a misidentification of the host. Dr. R.B. Faden, normal stem-boring in this genus. Also the habit
Department of Botany, U.S.N.M., Smithsonian Insti- of the larva to eat out a channel to the lower epi-
tution, Washington, D.C. made a cytological exam- dermis of the leaf which is left intact and then re-
ination of a leaf fragment preserved with specimens treat some way back into the leaf for transforma-
in the U.S.N.M. collection and he established that tion is exactly comparable to the practice of stem-
the host did not represent Typha. However, there is borers. It thus seems certain that this leaf-mining
a record in Europe of Liriomyza trifolii having been species has switched to its orchid hosts from an
found on Typha (see below). ancestral stem-borer.
It seems certain that there has been direct inter- M. dipodii Spencer (1977c). Host Dipodium punc-
change of ancestral species on Iris between Eurasia tatum, known only at three localities near Syd-
and North America, with subsequent speciation of ney, New South Wales. I have seen larvae feeding
the five modern species now known. There has also in the flower-heads. This is a typical black Mela-
been a switch in species in this group between nagromyza but no other species is known in Aus-
Cyperaceae and Iridaceae - and also Poaceae (see tralia with the combination of black mesonotum
below). and abdomen and white squamal fringe. It is thus
an isolated species (Fig. 1271), with undetectable
affiliations.
Orchidaceae (800 genera, 18-20 000 species)
M. galeolae Sasakawa (1961). Host Galeola septen-
This large family is of cosmopolitan distribution but trionalis, known only in Japan. The larvae form
with the majority of species in the tropics. There is external mines in the fruits of the host, with be-
no agreement among specialists on the family re- tween 5-35 larvae in a single fruit. The colour of
garding either its exact size, or its taxonomic posi- the adult is as in the two preceding species but the
tion. male genitalia are highly differentiated (Fig.
Cronquist (1981) writes: 'The Orchidales are evi- 1272).
dently derived from the Liliales.' He places the Or- M. miamensis Spencer (in Spencer and Stegmaier,
chidaceae and Liliaceae in separate, though adjoin- 1973). Host 'orchid', Miami, Florida, 'larva des-
ing orders. Rasmussen (in Dahlgren et al., 1985) troys bloom'. This is a typical greenish species
also considers that the orchids evolved within the and the genitalia (Figs. 1273, 1274) are charac-
order Liliales but he split them into three taxa of teristic of many species in the genus. The eye is
family rank. As there is no available intermediate stated to be bare and this suggests that the ances-
rank in discussion of the mono cots between order tral relationship was not with an Asteraceae-feed-
and family, the orchids are thus left in his treatment er, in which the eye, at least in the male, is nor-
without formal rank and they are merely referred to mally conspicuously pilose.
as 'the orchids'. However, I prefer to follow the tra- M. orchidearum (Costa Lima, 1950). Hosts Cattleya
ditional treatment adopted by Cronquist (1981) and and Laelia, known only from Santos, Brazil, lar-
particularly by Dressler (1981) in his monograph of vae feeding in the roots. The species is black. A
334 Class LILIOPSIDA
.-:":;
'/~j,~
... -,-
1270 1271
1272 1275
1273 1274
~_' :-:.~.~.~
........_.-:=""Jc.)_).... .,:
. ) .. ..
1276 1277
Fig. 1270. Melanagromyza cyrtorchidis: aedeagus.
Fig. 1271. Melanagromyza dipodii: aedeagus.
Fig. 1272. Melanagromyza galeolae: aedeagus (Sasakawa).
Figs. 1273, 1274. Melanagromyza miamensis: 1273, aedeagus, side view; 1274, same, ventral view.
Fig. 1275. Melanagromyza tokunagai: aedeagus (Sasakawa).
Figs. 1276, 1277. Tropicomyia eulophiae: 1276, aedeagus, side view; 1277, same, ventral view.
male was present in the type series but the geni- A single Tropicomyia is known on the Orchida-
talia remain unknown and the species cannot be ceae. Mines I found at Kapsabet, Western Highlands,
further clarified. Kenya on Eulophia porphyroglossa produced adults
M. tokunagai Sasakawa (cf. Sasakawa, 1961). Host originally identified as the polyphagous T. flacour-
Cymbidium virescens, only known in Japan. The tiae. However, the mines are greenish, linear, with
egg is laid in the flower-bud and the young larva conspicuous frass and re-examination of male now
mines towards the stem, pupating externally. This satisfies me that this represents a new species,
is another black species and the complex genitalia Tr. eulophiae, which is described in the Appendix
(Fig. 1275) give no indication of its affiliations. below. The genitalia (Figs. 1276, 1277) are entirely
External pupation is virtually unknown in Mela- typical of Tropicomyia and it is not possible to be
nagromyza. certain of its exact affiliations.
MONOCOTYLEDONS 335
Table 84
Order DIOSCOREALES
Dioscoreaceae
Dioscorea Tropicomyia theae, T. flacourtiae; Liriomyza schmidti
Smilacaceae
Eustrephus T. polyphyta
Smilax T. flacourtiae, T. polyphaga, T. polyphyta; L. schmidti
Order ASPARAGALES
Agavaceae
Cordyline Tropicomyia sp. n.
Alliaceae
Allioideae
Allium L. cepae, L. chinensis, L. huidobrensis, L. nietzkei, L. trifolii;
Napomyza gymnostoma; Chromatomyia horticola
Amaryllidaceae
Haemantheae
Cyrtanthus T. cyrtanthi
Haemanthus T. haemanthi
Asparagaceae
Asparagus Ophiqmyia asparagi, O. simplex; Ptochomyza asparagi, Pt. asparagivora
Asphodelaceae
Asphodeloideae
Asphodelus L. asphodeli
Alooideae
Aloe Penetagromyza aloes
Hemerocalii daceae
Hemerocallis O. kwans0nis
Phormiaceae
Dianella Kleinschmidtimyia dianellae
Alstroemeriaceae
Alstroemeria L. huidobrensis, L. trifolii
Bomarea Liriomyza (probably sp. n.)
Liliaceae
Gloriosa T. gloriosae, T. ? polyphaga
Lilium L. urophorina
Maianthemum Gen. indet. sp. n.
336 Class LILIOPSIDA
Smilacina L. smilacinae
Veratrum L. wachtli
Iridaceae
Iridoideae
Belamcanda Amauromyza belamcandae; Cerodontha ( Dizygomyza ) ireos
Iris Cerodontha ( Dizygomyza ) ireos, Cer. ( Diz.) iridicola,
Cer. ( Diz.) iridis, Cer. ( Diz.) iridophora, Cer. ( Diz.) thompsoni
Ixioideae
Gladiolus Cerodontha ( Dizygomyza ) ireos; L. trifolii
Orchidaceae
Epidendroideae
Vanilleae
Galeola Melanagromyza galeolae
Vandoideae
Cymbideae
Cattleya M. orchidearum
Cymbidium M. tokunagai
Dipodium M. dipodii
Laelia M. orchidearum
Vandeae
Cyrtorchis M. cyrtorchidis
Rangaeris M. cyrtorchidis
Eulophia T. eulophiae
( syn.: Lissochilus )
Subfamily, genus indet. M. miamensis ( Florida)
This is probably the same species referred to by rind of the fruit and unusual leaf-mining. The geni-
de Meijere (1940: 169) as Melanagromyza spec., talia of the five black species do not indicate direct
based on mines with larvae found by Buhr on Mt. relationship and it seems therefore that there have
Cameroun in 1938 on Lissochilus horsfalli. Lisso- been several colonizations, but surprisingly all by
chilus is now accepted as synonymous with Eulo- species in Melanagromyza and with similar external
phia. characters. This is a puzzling assemblage deserving
In their biology the six Melanagromyza species further study.
known on Orchidaceae represent a remarkably Tropicomyia eulophiae in Kenya may well be
varied assemblage. Although five are generally simi- related to T. haemanthi in South Africa (see Amar-
lar on external characters, the biology varies from yllidaceae above).
feeding in roots, stems and flowers, to mining in the
Superorders ARIFLORAE, ALiSMATIFLORAE, BROMELlIFLORAE
and ZINGIBIFLORAE
Only eight species of Agromyzidae in the three gen- spread polyphagous species T. polyphaga and again
era Tropicomyia, Metopomyza and Liriomyza are very probably represents a further undescribed spe-
known on these four Superorders. They are shown cies. Detailed studies of the male genitalia will be ne-
with their host genera in Table 85. cessary to establish the identity of these two species.
In Argentina linear-blotch mines were found on
Zantedeschia aethiopica by Valladares but no adults
Superorder ARIFLORAE were obtained. This plant is introduced and of South
African origin and thus the agromyzid concerned
In the treatment by Dahlgren et al. (1985) this cannot be a local, host-specific species but I suspect
superorder consists of the single order Arales, com- an abnormal mine of one of the polyphagous Lirio-
prising the two families Araceae and Lemnaceae. myza species.
Although of such very dissimilar appearance, there
is no dispute about their close association in the
same order. Agromyzidae are only known on Ara- Superorder ALiSMATIFLORAE
ceae.
There are significant similarities between the Ari- As defined by Dahlgren et al. (1985) this superorder
florae and the Alismatiflorae, and there are also comprises two orders, the Alismatales and Naja-
some similarities between the Ariflorae and the dales. Cronquist considers that this group represents
Dioscoreales, the most primitive order in the Lilii- a 'near-basal, side branch, a relictual group that has
florae. Dahlgren suggests 'a likely hypothesis is that retained a number of primitive characters ... within
the Ariflorae share their ancestors with those of the the Magnoliidae the aquatic order Nymphaeales
more primitive Liliiflorae. The Alismatiflorae may presents the closest approach to the Alismati-
represent a sister-group.' dae .... The fossil record carries the Alismatidae
It is thus probably not coincidental that the only back to the Paleocene epoch, but records from the
agromyzid genus known on the Dioscoreaceae, Tro- Upper Cretaceous are doubtful or incorrect.'
picomyia, has also three times been found on genera However, Cronquist concedes that the Alismati-
of the Araceae. dae can scarcely be on the main branch of evolution
of the class.
Dahlgren considers that some features found in
Araceae (110 genera, 2450 species)
members of the Ranunculaceae and Alismataceae
This family is of worldwide distribution but is are 'without any doubt the result of convergent evo-
centred in the Old World tropics. lution' but he agrees that there are 'numerous and
Three Tropicomyia species are known on dif- far-reaching similarities between various Alisma-
ferent genera, two at least probably representing un- tales and various Nymphaeales which cannot ... be
described species. In Cameroun Buhr found mines dismissed.'
on Arisaema sp. (identified as Melanagromyza by de
Meijere (1940: 172)). This certainly represents a
Order ALiSMATALES
Tropicomyia, possibly Tr. flacourtiae. Sasakawa
(1972) recorded as Melanagromyza theae two fe- Dahlgren (loc. cit.) places five families in this order,
males reared from Colocasia antiguorum in Taiwan. while Cronquist restricts it to three. Only the family
The species with virtual certainty represents a Tropi- Butomaceae has been colonized by the single agro-
comyia sp.n. In January, 1960 I found mines on myzid, Metopomyza ornata.
Lasia spinosa at Bangkok, Thailand (Spencer,
1986b). The epidermal mines are typical of Tropico-
Butomaceae
myia but the empty puparia are unusually large and
I believe the species cannot be treated as the wide- This family consists of one species, Butomus umbel-
338 Class LILIOPSIDA
latus, which is native to temperate Eurasia but is length from 2.6 mm in male to 3.5 mm in female.
naturalized in parts of North America. 'Seeds Spencer (1976a) transferred ornata to Metopo-
thought to represent Butomus occur in Oligocene myza, suggesting that the genus might be treated as a
and younger rocks. Pollen similar to that of Buto- subgenus of Phytoliriomyza. The close relationship
mus dates back at least to the Miocene' (Cronquist, between Metopomyza and Phytoliriomyza is clear.
1981). In view of its worldwide distribution, particularly the
Metopomyza ornata (Meigen, d. Spencer, 1976a) group of fern-feeders, Phytoliriomyza can be ac-
feeds in the stems of Butomus. The generic position cepted as the older, more primitive genus and I now
of this interesting species has been unclear to earlier consider that the small group of Metopomyza spe-
workers. After its description in Agromyza, it was cies feeding on the monocot families Butomaceae,
transferred to Liriomyza by Hendel (1920) by vir- Juncaceae, Cyperaceae and Poaceae represent an
tue of its yellow scutellum and later to Cerodontha early offshoot from Phytoliriomyza, restricted on
by von Tschirnhaus (1971). However, the absence the basis of present knowledge to the Palaearctic
of the paired sclerites within the epandrium charac- Region (Europe and Siberia).
teristic of all species in Cerodontha, precludes its in- Selachops Jlavocincta Wahlberg (d. Spencer,
clusion in that genus. The aedeagus itself (Figs. 197 6a) is an intriguing and obviously primitive spe-
1278, 1279) closely resembles that of grass and cies. The head (Fig. 1281) has enormously project-
Carex feeders (d. Figs. 1310, 1312) but in Metopo- ing orbits, the genitalia (Figs. 1282, 1283) and par-
myza the strong bristles on the surstyli and along the ticularly the epandrium (Fig. 1284) all suggest prob-
inner margin of the epandrium (Fig. 1280) are able relationship with Metopomyza ornata (see fur-
exactly as in some species of Phytoliriomyza (par- ther discussion in Section 6). Although the host is
ticularly fern-feeders such as Phlir. hilarelia, cya- unknown, it seems desirable to mention this species
theae and kuscheli (Fig. 19). As in many primitive here.
species, Met. ornata is exceptionally large with wing
Superorder ALISMATIFLORAE 339
1278 -"
1279 1280
1282
(
~ I :
.;,: " ~, i I
::.'.
...~
: "
: ",
1285 1287
1286 1288
Figs. 1278-1280. Metopomyza ornata: 1278, aedeagus, side view; 1279, same, ventral view; 1280, epandrium.
Figs. 1281-1284. Selachops flavocincta: 1281, head; 1282, aedeagus, side view; 1283, distiphallus, ventral view; 1284, epandrium.
Figs. 1285, 1286. Liriomyza angulicornis: 1285, aedeagus, side view; 1286 same, ventral view.
Figs. 1287, 1288. Liriomyza latipalpis: 1287, aedeagus, side view; 1288, same, ventral view.
340 Superorders ARIFLORIAE, ALISMATIFLORAE, BROMELIIFLORAE, ZINGIBIFLORAE
Table 85
ARIFLORAE
Araceae
Arisaema Tropicomyia sp. ( possibly flacourtiae )
Colocasia Tropicomyia probably sp. n.
Lasia Tropicomyia probably sp. n.
Zantedeschia ? Liriomyza sp.
ALISMATIFLORAE
Butomaceae
Butomus Metopomyza ornata
BROMELIIFLORAE
J uncaginaceae
Triglochin L. angulicornis, L. latipalpis
Typhaceae
Typha L. trifolii
ZINGIBlFLORAE
Zingi beraceae
? genus Agromyzid sp. ( ? genus)
Superorder COMMELINIFLORAE
As defined by Dahlgren et al. (1985) this superorder regions and Agromyzidae have been found in the
embraces four orders, of which the Commelinales, southern U.S.A., Cuba and South America south to
Cyperales and Poales support Agromyzidae. Argentina on the two genera Commelina and
Dahlgren states that the Commelinales and pos- Tradescantia. Mines of an undescribed species have
sibly the Cyperales have a South American concen- also been found near Nairobi, Kenya.
tration and probably a West-Gondwana origin, I have" seen an undescribed Amauromyza sp.
while the Poales are concentrated in the Old World. reared from Commelina elegans in Cuba. The geni-
There was a rather late, explosive diversification of talia (Fig. 1289) clearly indicate the generic position
both Poaceae and Cyperaceae during the Tertiary. of this species and I believe the accuracy of the host
The closest relationships of the Commeliniflorae record can be accepted. Dr. E. Rohdendorf who dis-
among the monocotyledons are probably with the covered this species proposed to describe it in 1969
Zingibiflorae, on which only a single agromyzid is but this was not done and now that she has been
known on the Zingiberaceae. transferred to other work in Moscow regrettably the
It will be seen from Table 86 that the number of description is unlikely to be completed. This is a
Agromyzidae known on the orders of the Com- most unusual host for an Amauromyza and can only
meliniflorae varies significantly between those on represent an 'accidental' switch. A number of
the Commelinales and the Poales, certainly to some Amauromyza species have been described from
extent reflecting the size of the orders and thus the southern California of which the hosts are unknown
number of potential hosts. and it cannot be postulated from which family the
switch may have occurred.
Table 86 Liriomyza commelinae Frost (d. Spencer and
Steyskal, 1986) is known on both Commelina and
Agromyzidae on orders of COMMELINIFLORAE Tradescantia. It was described from St. Vincent and
Order
has since been found widely from Florida, through-
Commelinales out the Caribbean and in Venezuela, Brazil and
Commelinaceae 2 (+2 undescribed) Argentina. It is a remarkable and significant species,
Cyperales aberrant in a number of respects. Externally L. com-
Juncaceae 14 melinae has the aspect of a Liriomyza, with a dis-
Cyperaceae 26 (+2 undescribed)
Poales
tinctive black and yellow pattern on the mesonotum
Poaceae 107 (+"155, deduced from male genitalia) (Fig. 1290) but in the male the third antennal seg-
ment (Fig. 1291) is greatly enlarged. This and char-
acters of the larva, puparium and leaf mine are
Order COMMELINALES
atypical of Liriomyza. The puparium is black, elon-
As defined by Dahlgren et al. (1985) this order gate, somewhat flattened and remains in the mine.
comprises 5 families, with agromyzids only known Each posterior spiracular process retains the primi-
on the Commelinaceae. This family is widely dis- tive arrangement of three pores but here one is
tributed in the tropics and extends into temperate greatly enlarged, hook-like, with two minute ones at
regions of both the northern and southern hemi- its base (Fig. 1292), generally resembling the form
spheres, while the other families are concentrated in found in many Butomomyza and Dizygomyza spe-
South America. Pollen assumed to represent the cies (d. Fig. 1308). The leaf mine is long, narrow,
Commelinales first appears in the Senonian, about with frass deposited in widely-spaced black lumps,
the middle of the Upper Cretaceous. whereas in Liriomyza it is invariably in narrow
strips.
The male genitalia (Fig. 1293) give no clear indi-
Commelinaceae (50 genera, 700 species)
cation of the relationship of L. commelinae. How-
This family is widespread in tropical and subtropical ever, the stridulating organ is present and the spe-
Order COMMELINALES 343
1289
;;?(F~
j
1290 1292 )1293
,~.~r· ., '.
1 2 9 6 ' 1297 \1" 1299
\.
cies is thus retained in Liriomyza. With other char- writes 'On the basis of all the available information,
acters of both adults and larva (puparium) associat- including chemical and anatomical as well as gross
ing the species with Dizygomyza, I suggest that this is morphological features, I believe that the Cypera-
a relict, which has retained primitive characters of ceae and Poaceae are closely related lines diverging
ancestral species before the diversification of Cero- at an angle from a common source .... The two fami-
dontha, subgenus Dizygomyza. lies are, all considered, too similar to be dissociated;
Liriomyza robustae Spencer (1985b) was dis- they must stand side by side in the system.'
covered in the Andes in central Colombia, feeding Treating the Cyperaceae and Poaceae as in separ-
on Commelina robusta. This species differs signifi- ate orders following Dahlgren accords better with
cantly from L. commelinae but I consider the two the host specialization of the Agromyzidae. There is
are sister-species, but with considerable diversifica- a very different generic composition in the agro-
tion. In L. robustae the third antennal segment is en- myzids on the two families (see Tables 88 and 90)
larged but the mesonotal pattern (Fig. 1294) is more and proportionately far fewer species on the Cyper-
extensively yellow, the male genitalia (Figs. 1295, aceae. Clayton (1986: 19) even states: 'The similar-
1296) are well differentiated and the leaf mine (Fig. ity between Gramineae and Cyperaceae is a matter
1297) is entirely different from that of L. com- of convergence rather than kinship.' The most
melinae, the larva forming a blotch on the midrib, striking difference between the Agromyzidae on
with irregular offshoots into the leaf blade. The Cyperaceae and Poaceae is the complete absence on
puparium is yellowish, with the anterior spiracles the former of species in Agromyza, the subgenus
two-horned (Fig. 1298) and the arrangement of the Cerodontha and Pseudonapomyza, and only a single
posterior spiracles essentially as in L. commelinae. species of Poemyza.
The only known host of L. robustae, Commelina It is thus clear that the agromyzids are very aware
robusta, is believed to be restricted to the Andes of the differences between the two families.
and presumably L. robustae has evolved in isolation
both on this host and geographically.
Juncaceae (8 genera, 300 species)
I recently found leaf mines on Commelina imber-
his at Langata, near Nairobi, Kenya which are an ir- In the past botanists have not been wholly in agree-
regular blotch (Fig. 1299), with external pupation. I ment on the derivation of the Juncaceae and it has
assume these are referable to an undescribed Lirio- been customary to consider the family as florally
myza species and it will be of great interest to dis- reduced descendants of the Liliaceae. However,
cover the relationship of this species to the two on Cronquist (1981) convincingly argues that the fam-
Commelina known in South America. It seems un- ily resembles typical Commelinidae rather than typi-
likely that Commelina would have been colonized cal Liliidae. This is fully supported by the host selec-
twice independently in Africa and South America tion of the Agromyzidae, with 14 species found on
and the two populations possibly indicate direct dis- either funcus or Luzula in the genera Cerodontha
persal between Africa and South America. and Chromatomyia, all closely related to species
feeding on Cyperaceae or Poaceae, and with no re-
Order CYPERALES lationship with any of the species on Liliidae (see
below).
As defined by Dahlgren et al. (1985) this order con- Only three genera of Agromyzidae - Metopo-
sists of the Juncaceae, Cyperaceae and the small myza, Cerodontha (subgenera Icteromyza and Dizy-
family Thurniaceae on which no agromyzids are gomyza) and Chromatomyia have colonized the
known. Dahlgren considers that there is strong evi- Juncaceae, with in all 14 species, as shown in Tables
dence that the Cyperales are a monophyletic unit 87 and 89.
and that there is close relationship between the three In the small genus Metopomyza (consisting of 16
families. species according to von Tschirnhaus, 1981: 332)
Traditionally the Cyperaceae and Poaceae have only M. junci von Tschirnhaus (1981) has been con-
been closely associated in systems of classification firmed as feeding on funcus. Hosts known of other
but more recent authors, while recognising their af- species are Carex and grasses (see below) and it
finities have placed them in separate orders. Takh- seems certain that all species feed on genera in one
tajan (1969) placed the two groups not only in or other of the three families Cyperaceae, Juncaceae
separate orders but these well removed from each orPoaceae.
other, the Cyperaceae in his Order 84, the Poaceae M. funci was reared from funcus gerardii on the
in Order 89. Frohne and Jensen (1985) include the coast of Schleswig-Holstein near the Danish border.
two families in separate, although adjacent, orders. The larva forms a linear-blotch mine, pupating ex-
After considering the similarities and differences be- ternally, the male genitalia not significantly differing
tween the two families Cronquist (1981: 1135) from M. scutellata (see under Carex below).
Order CYPERALES 345
1300 1301
1302 1303
1306
1304
./ 1307
1305 1308
Cyperaceae (110 genera, c. 4000 species, after several larvae feed together and pupate at the base
Kew and BMNH.) of the leaf of Eriophorum latifolium, in Europe. The
genitalia are shown in Fig. 1313. Carex pseudo-
This family is of cosmopolitan distribution but is
cyperus has been recorded as a further host of this
best represented in temperate regions. Fossils
species but this requires confirmation (Nowa-
ascribable to the Cyperaceae date from the Eocene
kowski, 1973). In South Africa C. (let.) stuckenber-
in several areas.
giella Spencer (1977b) is indistinguishable ~ro~
Only three agromyzid genera (Cer~dontha .",:ith
C. geniculata on external characters but the gemtaha
four of its subgenera) have succeeded m colomzmg
are significantly different (Fig. 1314). One species of
the Cyperaceae (Table 88). The family is now ?ivid-
Eriophorum is known in South Africa in the Trans-
ed into five subfamilies of which the CyperOldeae,
vaal but Carex, which is more widely distributed
Rhyncosporoideae and Caricoideae support Agr?-
with 16 species, is a more probable host. C. (let.)
myzidae but the majority of species are .o~ Carex m
piliseta (Becker, cf. Spencer, 1985a) is a wide-
the Caricoideae (see Table 89). Surpnsmgly, only
ranging species in the Old World tropics and is not
seven of the available 110 genera are known to sup-
uncommon in the Lae area, Papua New Guinea. I
port agromyzids but with further collecting, particu-
found a mine in the Botanical Gardens, Lae on Fim-
larly outside Europe, it is likely this figure will great-
bristylis dichotoma and the larval characters defi-
ly increase.
nitely indicate that the species is a Cerodontha,
There is obviously considerable disagreement
either subgenus leteromyza or Dizygomyza. Fim-
among botanists on the taxonomy of the Cyper-
bristylis is widely distributed in the Old World trop-
aceae. The number of genera and species as cited by
ics with 250 species and I think it is very probable
Dahlgren et al. (1985) - 96: 9300 differs substan-
that this is the host of C. piliseta. I caught a speci-
tially from those in Cronquist (1981) - 70: 4000
men at Embu, Kenya in an area where Fimbristylis is
and in Mabberley (1987) - 115: 3600. Dahlgren's
likely to occur. The distinctive genitalia of C. piliseta
figures are an obvious error (see his own text) and
are shown in Fig. 1315.
here I accept the figures agreed between Kew and
Only a single Poemyza is known on Cyperaceae,
BMNH.
C. (Po.) macminni Spencer (Spencer and Steyskal,
Interestingly, this family is totally avoided by
1986) feeding on Carex crus-corvi in Mississippi,
polyphagous species.
U.S.A. The larvae form a long, brownish blotch,
4-16 larvae feeding together. Externally and in
Table 88
characters of the larva and male genitalia (Fig. 1316)
Agromyzid genera on Cyperaceae this species closely resembles a number of grass-
feeding Poemyza species. During his 12-year study
Metopomyza 2 of this genus in Europe Nowakowski did not dis-
Cerodontha cover a single Poemyza on Carex and erected the
Icteromyza 2
new subgenus Butomomyza 'for the Cyperaceae-
Poemyza 1
Butomomyza 11 (+ 1 undescribed) mining Poemyza-group.' The discovery of Po. mac-
Dizygomyza 9 (+ 1 undescribed) minni on Carex thus represents a unique switch
Chromatomyia 1 from the normal hosts of the subgenus on grasses.
The dominant subgenus on Carex is Butomo-
26 (+2)
myza. The subgenus is virtually cosmopolitan, with
a total of 20 described species, of which ten have
Two species in the small, primitive genus Metopo- been reared from Carex. Only But. rohdendorfi is
myza are known to feed on Carex. M. scutellata known on Poaceae (see below). Nowakowski (1973)
(Fallen, cf. Spencer, 1976a) wa~ reared by Gr?sc~ke has discussed the European species in great detail,
(1954) from a narrow linear mme on Carex sllvatlca and it will suffice to record here only essential refer-
in southern Germany (misidentified as M. xantha- ences and other facts of particular interest:
spis, see Spencer, 1976a). The aedeagus and epa~
C. (But.) angulata (Loew, cf. Spencer, 1976a). Host
drium are shown in Figs. 1310, 1311. M. xanthaspls
Carex, rarely Scirpus sylvaticus, widespread and
(Loew, cf. Spencer, 1976a) forms leaf mines on
common in Europe, Canada, U.S A., Guyana
Carex spp. and the aedeagus (Fig. 1312) is c~ear~y
(South America). Male genitalia: Fig. 1317.
distinct from that of M. scutellata. Other speCIes m
C. (But.) caricivora (Groschke, 1954). Host Carex
this genus are known to feed on grasses (see below).
sylvatica, C. pilosa, central and northern Europe.
Cerodontha is the main genus feeding on Cyper-
Male genitalia: Fig. 1318.
aceae, with the four subgenera leteromyza, Poemyza,
C. (But.) eucaricis Nowakowski (cf. Nowakowski,
Butomomyza and Dizygomyza represented.
1973). Hosts Carex spp. (not sylvatica or pilosa).
In C. (lct.) geniculata (Fallen, cf. Spencer, 1976a)
348 Superorder COMMELINIFLORAE
1310
1311
~,";-{-
...
1314
1315
1316
1317
1318 1319
Order CYPERALES 349
Not uncommon in Europe, widespread in Cana- pupal case on Cyperus ajax in the Aberdare Moun-
da, closely related to C. caricivora. Male genitalia: tains west of Nyeri, Kenya which certainly repre-
Fig. 1319. Block (1974), following the study of sents an undescribed species, probably in Buto-
larval chromosomes, suggests that two semispe- momyza.
cies may exist among populations in southern I recently discovered mines of C. (But.) scleriae
Sweden: Martinez (in Spencer et al., 1990) on Scleria pterota
C. (But.) mellita Spencer (1971b, cf. Nowakowski, on Guadeloupe and adults were later reared by the
1973). Hosts Carex, central and northern Eu- local entomologist, Ing. J. Etienne. The genitalia
rope, incl. England but uncommon. Male geni- (Fig. 1326) give no real indication of its nearest rela-
talia: Fig. 1320, closely related to C. angulata. tive.
C. (But.) pseuderrans (Hendel, d. Loew, 1976a).
C. (But.) cornigera (de Meijere, 1934) was de-
Host Carex hirta, central Europe, also England,
scribed from Java, reared from a species of
uncommon. Male genitalia: Fig. 1321, indicating
Cyperaceae, and was later reared from Carex
close relationship with C. eucaricis.
chinensis in Taiwan (Sasakawa, 1972). Sasakawa
C. (But.) scirpi (Karl, cf. Spencer, 1976a). Host Scir-
(1963a) illustrated the genitalia of a male from
pus sylvaticus, larva forming long, narrow mine,
Malaysia and this confirms that the species cor-
normally pupating deep in leaf sheath. Wide-
rectly belongs in Butomomyza, not in Poemyza as
spread in Europe, also in Canada. In the past con-
recorded by de Meijere and Sasakawa.
fused with C. (But.) scutellaris (Roser, cf.
Spencer, 1976a), the host of which is uncon- Ten further species are known in Butomomyza
firmed but probably Carex. Male genitalia: Fig. with no firm indication of the host, although almost
1322, indicating that C. scirpi is a more isolated certainly Carex: C. pubicata (Spencer, 1959) in
species, although associated by Nowakowski with Zaire, C. poemyzina in Australia (cf. Spencer,
C. pseuderrans. 1977c), six in the United States (Spencer and Steys-
C. (But.) staryi (Stary, cf. Nowakowski, 1973). Host kal, 1986) and one each in Argentina and Chile
Carex, central Europe. (Hering (1931c) formally (Spencer, 1982).
described this species but Starfs paper discussing
it as staryi Hering appeared earlier in 1930 and Subgenus Dizygomyza is widely distributed, with
the name staryi must therefore be attributed to some 60 species, of which 40 are known in Europe,
Stary). Male genitalia illustrated by Nowakowski 16 in the United States and two each in South
(1973: fig. 149) suggest relationship with C. scirpi America and Africa. The main host genus is Carex
but Nowakowski treats it as an isolated species. but a few species are known on lridaceae and Poa-
The larval posterior spiracles retain the primitive ceae. There have been numerous misidentifications
arrangement of three pores but these are highly in the literature of species in this subgenus but the
differentiated, with one greatly elongated (Fig. male genitalia are in all cases distinctive (see below).
1323). C. (Diz.) caricicola (Hering, cf. Spencer, 1976a).
C. (But.) vigneae Nowakowski (cf. Spencer, 1976a). Hosts Carex, widespread in Europe. Male geni-
Hosts Carex, uncommon, known only in eastern talia: Fig. 1327, with distal tubules exceptionally
Europe and southern Sweden. Male genitalia: Fig. long (very close to Diz. chaixiana on Poa, see be-
1324, indicating close relationship with C. pseu- low).
derrans. C. (Diz.) eriophori Nowakowski (cf. Nowakowski,
1973). Host Eriophorum latifolium, only known
In Africa C. (But.) cariciphaga (Spencer, 1963b) in Poland. Male genitalia (cf. Nowakowski, 1973:
(originally misidentified as Phytobia (Poemyza) cari- fig. 160) with aedeagus short and distal end en-
civora) was reared by Buhr from Carex sp. in Came- larged.
roun. This species is possibly related to C. angulata C. (Diz.) gallic a Nowakowski (cf. Nowakowski,
but the genitalia are well differentiated (Fig. 1325). 1973). Hosts Carex, widespread in Europe but
In October, 1970 I found a mine with an empty uncommon. Male genitalia (cf. Nowakowski,
~
-.;,
.~'.
1320
1321
1322
:::':,'"
~
1324
1325
1326
1327
1329 1328
Order CYPERALES 351
Table 89
Order COMMELINALES
Commelinaceae
Commelina Amauromyza sp. n. ( Cuba ); Liriomyza commelinae, L. robustae,
Liriomyza sp. n. ( Kenya)
Tradescantia L. commelinae
Order CYPERALES
Juncaceae
Juncus Metopomyza junci; Cerodontha ( Icteromyza ) capitata,
Cer. (Iet.) longipennis, Cer. (Ict.) duplicata, Cer. (Ict.) triplicata;
Cer. ( Oizygomyza ) luctuosa
Luzula Cer. (Oiz.) bimaculata, Cer. (Oiz.) carpatica, Cer. ( Oiz.) handlirschi,
Cer. (Oiz.) luzulae, Cer. (Oiz.) silvatica; Chromatomyia beigerae,
Chr. cygnicollina, Chr. luzulae
Cyperaceae
Subfamily Cyperoideae
Cyperus ? Cer. ( Butomomyza ) sp. n. ( Kenya ), Cer. (But.) sp. ( South Africa );
Cer. (Oiz.) sp. n. ( Venezuela)
Eriophorum Cer. (Iet.) geniculata; Cer. (Oiz.) eriophori
Fimbristylis Cer. (lct.) piliseta ( ? )
Kyllinga Oetails cannot be traced
Scirpus Cer. (But.) angulata, Cer. (But.) scirpi;
Cer. (Oiz.) scirpivora
Subfamily Rhyncosporoideae
Scleria Cer. (But.) scleriae
Subfamily Caricoideae
Carex Met. scutellata, Met. xanthaspis; Cer. (But.) angulata,
Cer. (But.) cariciphaga, Cer. (But.) caricivora, Cer. (But.) cornigera,
Cer. (But.) eucaricis, Cer. (But.) mellita, Cer. (But.) pseuderrans,
Cer. (But.) scirpi, Cer. (But.) staryi, Cer. (But.) vigneae;
Cer. (Oiz.) caricicola, Cer. (Oiz.) gallica, Cer. (Oiz.) hirtae,
Cer. (Oiz.) magnicornis ( probably Carex ), Cer. (Oiz.) morasa,
Cer. (Oiz.) spinata, Cer. (Oiz.) suturalis; Cer. ( Poemyza ) macminni;
Chr. pseudomilii
1973: fig. 173) with aedeagus of similar form to Germany virtually identical specimens treated as
C. caricicola but much shorter. paratypes were reared from Deschampsia. This spe-
C. (Diz.) hirtae Nowakowski (ct. Nowakowski, cies is discussed further below under Poaceae.
1973). Host Carex, common in Europe, also re- It may be mentioned here that this is the only spe-
corded in Japan (as luctuosa) and California. cies in any genus known to have hosts in both
Male genitalia: Fig. 1328, closely resembling Cyperaceae and Poaceae.
C. morosa in side view (ct. Fig. 1329) but distal
tubules parallel, not diverging as in luctuosa (ct.
Order POALES
Fig. 1305).
C. (Diz.) morosa (Meigen, ct. Spencer, 1976a). As circumscribed by Dahlgren et al. (1985) this
Hosts Carex, widespread in Europe. Male geni- order consists of five very small families and the
talia: Fig. 1329, indicating close relationship with Restionaceae with 40 genera and 400 species, none
C. caricicola, C. gallica and C. chaixiana on Poa- of which are known to support Agromyzidae, and
ceae (see below). These species are treated by the large family Poaceae. As explained under
Nowakowski as the morosa superspecies. Cyperales above, I here treat the Poaceae and
C. (Diz.) scirpivora (Spencer, 1969b). Host Scirpus, Cyperaceae as in separate orders, rather than in a
only known in Ontario, Canada. Male genitalia: single order as proposed by Cronquist (1981).
Fig. 1330. Up to five larvae feed together, pupat- Pollen referable to Restionaceae, a southern
ing side by side laterally across the leaf (as in Diz. hemisphere family centred on South Africa and
iridis on Iris) (Fig. 1267). Australia, has been reported from the late Creta-
C. (Diz.) spinata (Groschke, 1954). Hosts Carex ceous, whereas typical grass pollen is younger. 'The
pilosa, C. sylvatica, central Europe and Poland. grasses obviously began their expansion about the
Male genitalia (Nowakowski, 1973: fig. 167) beginning of Tertiary, and their cosmopolitan dis-
closely resembling C. caricicola (Fig. 1327) but tribution is thus relatively late in time' (Dahlgren et
the species is associated by Nowakowski with al., 1985).
C. bimaculata (Luzula) and C. luctuosa (funcus) ,
in which the aedeagus is very different (ct. Figs.
Poaceae (651 genera, 10 000 species)
1304, 1307).
C. (Diz.) suturalis (Hendel, ct. Spencer, 1976a). I have followed Clayton and Renvoize (1986) in
Hosts Carex hirta, ? Scirpus maritimus (1 Q only), their division of the Poaceae into six subfamilies
widespread in Europe but not common. Male (rather than seven as proposed by Dahlgren et al.
genitalia (Fig. 1331), indicating close relationship (1985», who treat the Stipa group (including
with C. (Diz.) magnicornis (Loew, ct. Spencer and Milium) as a seventh subfamily Stipoideae. The six
Steyskal, 1986, male genitalia: fig. 595) which is subfamilies, referred to by numbers below are I
widespread in U.S.A., also present in Canada. Bambusoideae; II Pooideae; III Centothecoideae;
The host is unknown but can be assumed to be IV Arundinoideae; V Chloridoideae; VI Panicoi-
Carex. deae.
Clayton considers that the origin of the grasses is
At Caracas, Venezuela I found a mine on Cyperus
obscure, 'for whatever link existed with the rest of
tenuis in December, 1958, in which larvae were
the Monocotyledons is now extinct.' Three evolutio-
feeding together, and pupated laterally across the
nary lines can be detected on taxonomic evidence:
leaf. The puparium (Fig. 1332), with the conspi-
cuous prothoracic horns, is characteristic of a num- Bambusoideae - tropical, specializing in forest and
ber of species in Dizygomyza. This species is certain- aquatic habitats
ly undescribed. Arundinoideae - Chloridoideae - Panicoideae -
Only a single Chromatomyia is known on Carex, mainly tropical or southern hemisphere
Chr. pseudomilii Griffiths (1980), which was reared Pooideae - temperate.
at the same locality in Alaska as further specimens
There is some support for this division in the host
from what is probably the main host at that area,
selection of the Agromyzidae, with only ten species
Agrostis. Two specimens were reared by Griffiths
present on the Bambusoideae, of which five are re-
from Carex, both on the same day, one from the
stricted to this subfamily. Nine species are host-
locality where the main type series was found on
specific on the Arundinoideae and two on the
Agrostis, the other on a mountain not far away.
Chloridoideae, while by far the largest number are
Whether Carex can be considered as a regular alter-
found on the temperate Pooideae. The agromyzid
nate host or whether it merely represents a local ab-
genera on the family are shown in Table 90 and the
normality of females under oviposition pressure re-
detailed distribution on the five subfamilies which
mains to be established by further collecting. In East
support agromyzids is shown in Table 91.
"41
" " :·
.'.
',': .
. .
..
.~
Order POALES 353
1330
( I
!
1332
Fig. 1330. Cerodontha (Diz.) scirpivora: aedeagus.
Fig. 1331. Cerodontha (Diz.) suturalis: aedeagus.
Fig. 1332. Cerodontha (Diz.) sp.: puparium (Venezuela).
Table 91
I Subfamily Bambusoideae
Bambuseae
Arundinaria Cerodontha (Poemyza) bisetiorbita
Bambusa Agromyza ceylonensis; Cer. (Po.) bambusae
Sasa Cer. (Po.) sasae
Ehrharteae
Ehrharta Cer. (Po.) muscina
Oryzeae
Oryza A. oryzae; Cer. (Po.) orbitona, Cer. (Po.) oryzivora;
Pseudonapomyza asiatica
Zizania A. oryzae; Cer. (Po.) incisa
II Subfamily Pooideae
Aveneae
Agrostis A. albipennis, A. nigripes; Cer. (Po.) atra, Cer. (Po.) beigerae,
Cer. (Po.) incisa, Cer. (Po.) pygmaea; Cer. (Phytagromyza) flavocingulata;
Chromatomyia milii, Chr. nigra, Chr. pseudomilii
Aira Chr. nigra
Alopecurus A. yanonis; Cer. (Cer.) denticornis; Cer. (Po.) calamagrostidis,
Cer. (Po.) incisa; Cel'. (Phytag.) flavocingulata; Chl'. alopecuri,
Chr. fuscula, Chr. nigra
Ammophila Cer. (Po.) incisa, Cel'. (Po.) pygmaea, Cer. (Po.) superciliosa;
? Pseudonapomyza sp. n.; Chr. nigra
Anthoxanthum Cer. (Po.) incisa; Chr. nigra
Apera Cer. (Po.) pygmaea, Cer. (Po.) superciliosa; Ps. atra; Chr. nigra
Arrhenatherum A. albipennis, A. rondensis; Cer. (Cer.) flavifrons; Cer. (Po.) pygmaea;
Chr. fuscula, Chr. nigra
Avena A. ambigua, A. nigrella; Cer. (Cer.) dorsalis; Cer. (Po.) incisa,
Cer. (Po.) pygmaea, Cer. (Po.) superciliosa; Liriomyza cruciata, L. orbona,
L. trifolii; Ps. atra; Chr. fuscula, Chr. nigra
Beckmannia Chr. fuscula
Calamagrostis A. rondensis; Cer. (Cer.) denticornis, Cer. (Cer.) hennigi; Cer. (Po.) atra,
Cer. (Po.) beigerae, Cel'. (Po.) calamagrostidis, Cer. (Po.) incisa,
Cer. (Po.) lapplandica, Cer. (Po.) melicae, Cer. (Po.) zoerneri;
Cer. (Phytag.) flavocingulata; Chr. nigra, Chr. norwegica
Chaetopogon Chr. nigra
(= Chaeturus)
Cinna Chr. cinnae, Chr. fuscula, Chr. norwegica
Deschampsia A. albipennis, A. lucida, A. nigripes; Cer. (Po.) imbuta, Cer. (Po.) pygmaea;
( = Vahlodea) L. orbona, L. pedestris; Chr. alopecuri, Chr. fuscula, Chr. nigra,
Chr. norwegica, Chr. pseudomilii, Chr. torrentium
Gaudinia Chr. nigra
Helictrotrichon Chr. fuscula, Chr. nigra
Hierochloe Cer. (Po.) muscina; Chr. ? milii, Chr. nigra
Holcus A. nigrella, A. nigripes; Cer. (Cer.) angustipennis, Cer. (Cer.) denticornis,
Cer. (Cer.) flavifrons; Cer. (Po.) incisa, Cer. (Po.) muscina,
Cer. (Po.) pygmaea; Cer. (Phytag.) flavocingulata; L. flaveola; Chr. milii
Lagurus Chr. nigra
Order POALES 355
Phalaris A. albipennis; Cer. (Cer.) denticornis, Cer. (Cer.) dorsalis; Cer. (Po.) incisa,
Cer. (Po.) phalaridis, Cer. (Po.) zoerneri; Ps. atra; Chr. fuscula,
Chr. nigra
Phleum A. albipennis, A. nigrella; Cer. (Cer.) denticornis, Cer. (Cer.) dorsalis;
Cer. (Po.) pygmaea, Cer. (Po.) ? superciliosa; Cer. (Phytag.) flavocingulata;
Chr. fuscula, Chr. nigra
Polypogon Cer. (Cer.) flavifrons; Chr. nigra
Trisetum A. nigrella; Cer. (Po.) alpina, Cer. (Po.) incisa; Chr. fuscula, Chr. nigra,
Chr. opacella
Bromeae
Bromus A. bromi, A. intermittens, A. rondensis; Cer. (Cer.) dorsalis;
Cer. (Po.) incisa, Cer. (Po.) muscina, Cer. (Po.) ? pygmaea;
Cer. (Phytag.) flavocingulata; L. flaveola, L. septentrional is; Chr. fuscula,
Chr. nigra
Hainardieae
Pholiurus Chr. nigra
Meliceae
Glyceria A. alunulata, A. lucida, A. nigrella, A. nigripes; Cer. (Po.) pygmaea
Melica Cer. (Po.) melicae; Chr. fuscula, Chr. nigra
Nardeae
Nardus Chr. nigra
Poeae
Arctagrostis Chr. arctagrostidis, Chr. norwegica
Brachypodium Cer. (Po.) pygmaea; Chr. nigra
Briza Cer. (Po.) incisa; Chr. nigra
Catapodium Chr. nigra
( = Scleropoa)
Cynosurus Chr. nigra
Desmazeria Chr. nigra
( = Brizopyrum)
Dactylis A. cinerascens; A. nigrella, A. rondensis; Cer. (Cer.) australis,
Cer. (Cer.) denticornis, Cer. (Cer.) dorsalis; Cer. (Po.) incisa,
Cer. (Po.) muscina, Cer. (Po.) pygmaea, Cer. (Po.) superciliosa;
Cer. (Phytag.) flavocingulata; Cer. (Oiz.) crassiseta; L. flaveola, L. phryne;
Chr. fuscula, Chr. nigra
Echinaria A.lucida
Festuca A. albipennis, A. nigrella; Cer. (Cer.) denticornis; Cer. (Po.) incisa;
Cer. (Po.) lapplandica, Cer. (Po.) muscina, Cer. (Po.) pygmaea;
Cer. (Butomomyza) rohdendorfi; Cer. (Phytag.) flavocingulata; Chr. nigra
Lamarkia Chr. nigra
Lolium A. nigrella; Cer. (Cer.) denticornis, Cer. (Cer.) dorsalis; Cer. (Po.) incisa,
Cer. (Po.) muscina, Cer. (Po.) pygmaea; L. lolii, L. septentrionalis;
Ps. atra; Chr. fuscula, Chr. nigra
Poa A. albipennis, A. nigrella, A. ron dens is; Cer. (Cer.) australis,
Cer. (Cer.) denticornis, Cer. (Cer.) dorsalis, Cer. (Cer.) fulvipes;
Cer. (Po.) alpina, Cer. (Po.) incisa, Cer. (Po.) melicae, Cer. (Po.) muscina,
Cer. (Po.) pygmaea, Cer. (Po.) superciliosa; Cer. (But.) rohdendorfi;
Cer. (Phytag.) flavocingulata; Cer. (Oiz.) chaixiana, Cer. (Diz.) fasciata;
L. flaveola, L. graminivora, L. orbona; Ps. atra; Chr. fuscula, Chr. milii,
Chr. nigra, Chr. poae, Chr. torrentium
356 Superorder COMMELINIFLORAE
Stipeae
Milium Cer. (Po.) incisa, Cer. (Po.) muscina; L. flaveola; Chr. fuscula, Chr. milii,
Chr. nigra
Triticeae
Aegilops Chr. nigra
Agropyron A. ambigua, A. nigrociliata; Cer. (Cer.) denticornis; Cer. (Po.) incisa,
? Cer. (Po.) inconspicua, Cer. (Po.) lateralis, Cer. (Po.) muscina,
Cer. (Po.) pygmaea, Cer. (Po.) superciliosa, Cer. (Po.) zuskai;
Cer. (Phytag.) flavocingulata; L. phryne; Chr. nigra
Elymus Cer. (Po.) incisa, Cer. (Po.) inconspicua, Cer. (Po.) superciliosa
Hordeum A. albipennis, A. ambigua, A. hordei, A. intermittens, A. luteitarsis,
A. megalopsis, A. nigrella, A. nigrociliata, A. rondensis, A. yanonis;
Cer. (Cer.) australis, Cer. (Cer.) denticornis, Cer. (Cer.) dorsalis;
Cer. (Po.) incisa, Cer. (Po.) lateralis, Cer. (Po.) muscina,
Cer. (Po.) superciliosa; L. graminivora, L. septentrionalis, L. trifolii;
Ps. atra, Ps. spinosa; Chr. fuscula, Chr. nigra
Secale A. albipennis, A. ambigua, A. cinerascens, A. intermittens, A. luteitarsis,
A. megalopsis (?), A. nigrella, A. nigrociliata, A. ron dens is;
Cer. (Cer.) denticornis, Cer. (Cer.) dorsalis; Cer. (Po.) incisa,
Cer. (Po.) lateralis, Cer. (Po.) superciliosa; Ps. atra; Chr. fuscula, Chr. nigra
Triticum A. albipennis, A. ambigua, A. mobilis, A. nigrella, A. nigrociliata,
A. prespana, A. rondensis, A. yanonis; Cer. (Cer.) australis,
Cer. (Cer.) denticornis, Cer. (Cer.) dorsalis, Cer. (Po.) incisa,
Cer. (Po.) lateralis, Cer. (Po.) superciliosa; Ps. atra, Ps. spinosa, Ps. spicata;
Chr. fuscula, Chr. nigra
III Centothecoideae
No records
IV Arundinoideae
Arundineae
Arundo Cer. (Cer.) phragmitophila
Molinia Cer. (Po.) melicae, Chr. nigra
Phragmites A. baetica, A. graminicola, A. hen deli, A. phragmitidis, A. spenceri;
Cer. (Cer.) phragmitophila; Cer. (Po.) incisa, Cer. (Po.) phragmitidis
V Chloridoideae
Cynodonteae
Cynodon Ps. asiatica, Ps. dilatata, Ps. malayensis
Eragrostideae
Eleusine Cer. (Cer.) dorsalis; L. marginalis; Ps. dilatata, Ps. spinosa; Chr. nigra
Eragrostis Cer. (Cer.) dorsalis; Ps. asiatica
Sporobolus Agromyza sp. (Kenya)
Order POALES 357
Leptureae
Lepturus Chr. nigra
VI Panicoideae
Andropogoneae
Coix Cer. (Po.) incisa; Ps. dilatata
Miscanthus Ps. dilatata
(Ischaemum)
Rottboellia Ps. philippinensis
Saccharum Ps. dilatata, Ps. spicata
Sorghum A. pallidifrons; Cer. (Cer.) dorsalis; Ps. hispanica
Zea A. parvicornis; Cer. (Cer.) dorsalis; Cer. (Po.) incisa, Cer. (Po.) orbitona,
Cer. (Po.) superciliosa; L. marginalis; Ps. asiatica, Ps. dilatata, Ps. gujaratica,
Ps. spicata, Ps. zeae; Chr. nigra
Paniceae
Brachiaria Agromyza sp. (Argentina); Ps. spinosa; Chromatomyia sp. n. (Costa Rica)
Digitaria A. penniseti; Cer. (Po.) incisa; L. marginalis
Echinochloa A. parvicornis, A. proxima; Cer. (Po.) incisa, Cer. (Po.) muscina, Cer. (Po.) superciliosa
Oplismenus Agromyza sp. (Australia)
Panicum A. panici, A. parvicornis, A. proxima, Agromyza sp. n. (Kenya),
Agromyza sp. n. (Venezuela); Cer. (Cer.) dorsalis; Cer. (Po.) incisa,
Cer. (Po.) javana; L. marginalis; Ps. spicata
Paspalum L. marginalis
Pennisetum A. penniseti, A. pennisetivora
Setaria A. albipennis, A. catherinae, A. nigrella, A. papuensis; Cer. (Cer.) ipei;
Cer. (Po.) aristella, Cer. (Po.) incisa, Cer. (Po.) sasakawai, Cer. (Po.) setariae;
Chr. nigra
Urochloa A. pallidifrons, A. susannae
(Fig. 1333) are entirely characteristic of grass-feeders essential references are given for each species, with
but on external characters, with the frons reddish and a numbered reference to the host subfamily, and
3 + 1 strong dorso-central bristles, the species ap- tribes are also included. The genus is normally only
pears like a typical feeder on Fabaceae and Hendel cited in Table 91. It should be added that the classi-
in his original description (1920: 123) even suggest- fication of the Poaceae has been somewhat modified
ed it might be synonymous with Agromyza Jrontella. since that cited by Griffiths (1963b) and I have fol-
However, this has entirely different genitalia. A dis- lowed Clayton and Renvoize (1986). I have not in-
tinctive character is the exceptional length of the cluded subtribes.
cerci (Fig. 1334). I have examined ten of the 12 orig- The agromyzid species are treated in alphabetical
inal specimens and the species is known from south- order for ease of reference in each zoogeographical
ern U.S.S.R., Hungary and Austria. It will be of great region.
interest to confirm its host.
Just three species found only in Europe and
Palaearctic Region
Japan form the cinerascens-group, with an entirely
different type of aedeagus, which strongly suggests
Agromyza ambigua-nigripes group
relationship with species on Rosaceae and Fabaceae
and I suggest that in this group colonization of the Agromyza albipennis Meigen (cf. Griffiths, 1963b;
Poaceae was by an ancestral feeder on one of these Spencer, 1976a). Hosts: II, Poeae, Aveneae, Triti-
two families. ceae; VI, Paniceae. The species has a wide host
There is only limited overlap in the host speciali- range, primarily on the Pooideae, with only a
zation of species between the six subfamilies of the single record on Setaria in the Panicoideae. Re-
Poaceae. In the discussion below brief details with cords on Phragmites are inaccurate. Very com-
358 Superorder COMMELINIFLORAE
c1fr6j;)~~
1333
1335 1336
1334
~ - ~--
1337 1338
\
\
1341 1342
1340
1343
1344
Order POALES 359
mon in western Europe, also present in Japan. Japan but Sasakawa's illustration of the genitalia
Holarctic, widespread in Canada but no speci- (1961) suggests that a closely related species may
mens known to have been reared. In America on- be involved. Closely related to A. nigrociliata
ly known host Triticum. (Fig. 1345).
A alunulata Hendel (1931, cf. Spencer, 1976a) A nigrella (Rondani, cf. Spencer, 1976a). Hosts: II,
(= A. distorta Griffiths, 1955). Host: II, Meliceae, A veneae, Poeae, Triticeae; VI, Paniceae. With re-
only known host Glyceria. Local in western Eu- cords on 13 genera, this species has one of the
rope. Genitalia (Fig. 1335) typical of the group widest host ranges in this group. A. nigrella was
but the anterior larval spiracles enlarged, with up earlier confused with A. ambigua. Common in
to 50 spiracular pores (Fig. 1336) modified for Europe, also recorded in the U.S.A (Figs. 1343,
the aquatic environment of the host. A similar 1344).
modification found also in A. oryzae (see below). A nigripes Meigen (cf. Griffiths, 1963b; Spencer,
These two are thus certainly sister-species but in- 1976a). Hosts: II, Aveneae, most commonly on
terestingly feeding on different subfamilies. Glyceria maxima; some other records in the liter-
A ambigua Fallen (= A. niveipennis Zett.) (Spencer, ature are certainly inaccurate, others require con-
1976a). Hosts: II, Aveneae, Triticeae, common in firmation from bred material. Older records on
Europe. Holarctic, also present in Canada and Phragmites refer to A. hendeli. The record of
U.S. A. nigripes in the U.S.A is based on a misidentifi-
A baetica Griffiths (1963b). Host: IV, Arundineae, cation of A. albipennis; however, the species is
only known host Phragmites communis from two present in Canada, known from four specimens I
localities in southern Spain. collected in Alberta (Spencer, 196 9b).
A bromi Spencer (1966e). Host: II, Bromeae, only A nigrociliata Hendel (1931, cf. Spencer, 1976a).
known host Bromus, found once in Berlin Botan- Hosts: II. Triticeae. Closely related to A. mobilis
ical Gardens. The male genitalia (Fig. 1337) are and nigrella, male genitalia (Fig. 1346) only
well differentiated. In the larva (puparium) the slightly differentiated.
posterior spiracles are widely separated (Fig. A oryzae (Munakata, cf. Sasakawa, 1961; Spencer,
1338)(contrast A. megalopsis, Fig. 1342). 1973a). Hosts: I, Oryzeae, described from Japan,
A graminicola Hendel (1931; Spencer, 1976a). also present in eastern Siberia, Taiwan and Thai-
Host: IV, Arundineae, only host Phragmites com- land. This pest of rice has been discussed in detail
munis. Not uncommon in eastern Europe, but by Sasakawa (1961). Male genitalia (Fig. 1346)
not known in the west. Male genitalia (Fig. 1339), distinctive, larva with proliferation of pores on
leaf mine on Phragmites (Fig. 1340). anterior spiracles (Fig. 1347), associating the spe-
A hendeli Griffiths (1963b). Host: IV, Arundineae, cies with A. alunulata (Fig. 1336).
only host Phragmites communis. Widespread in A phragmitidis Hendel (cf. Griffiths, 1963b; Spen-
Europe, previously confused with A. lucida. cer, 1976a). Hosts: IV, Arundineae, only host
Alucida Hendel (cf. 1931; Griffiths, 1963b) Phragmites communis. Widespread in Europe,
(= A. airae Karl). Hosts: II, Aveneae, Meliceae, also present in Japan.
Poeae. Records of Agropyron, Bromus and Hol- A prespana Spencer (cf. Griffiths, 1963b). Host: II,
cus given by Hering (1957b) considered by Grif- Triticeae, only known host Triticum. Described
fiths as doubtful. Widespread in Europe, also re- from Yugoslavia, subsequently found to be wide-
corded in Ontario, Canada from Glyceria grandis spread in Europe north to Sweden but local.
(K.AS.). A rondensis Strobl (cf. Spencer, 1976a) (= A. oc-
A megalopsis Hering (cf. Spencer, 1973a). Hosts: cellaris Hendel, A. nigrifemur Hering, A. veris
II, Triticeae, normally Hordeum, possibly Secale. Hering). Hosts: II, A veneae, Bromeae, Poeae,
Widespread in Europe, can be a serious pest on Triticeae, widespread from Spain to Sweden.
barley. Male genitalia (Fig. 1341), posterior spira- Male genitalia (Fig. 1348) showing close rela-
cles ofpuparium (Fig. 1342). tionship with A. conjuncta Spencer (1966e)
A mobilis Meigen (cf. Spencer, 1976a). Host: II, which is most common in southern Europe but
Triticeae, possibly with wider host range. Wide- has also been recorded in Scotland; its hosts are
spread and common in Europe, recorded in not known but it obviously feeds on Poaceae.
1345 1346
1348 .:,.
::\',
/
.-/-/
~1i!(;.
\
1349 1347
1350 1351
.~9Y
1352
Order POALES 361
A spenceri Griffiths (1963b). Host: IV, Arun- cies is a well-known pest of Zea mays but popula-
dineae, only known host Phragmites communis, tion build-up is normally prevented by a complex
found only once in southern France. This species of 18 recorded hymenopterous parasites.
is well differentiated from other Phragmites feed- A proxima Spencer (in Spencer and Stegmaier,
ers both on external characters and in the male 1973). Hosts: IV, Paniceae, only known in Flori-
genitalia (cf. Griffiths, 1963b). da and New York. Although generally similar to
A yanonis (Matsumura, cf. Sasakawa, 1961). Hosts: A. parvicornis, the genitalia of this species (Fig.
II, A veneae, Triticeae. Known only in Japan, 1354) are longer and narrower.
Sasakawa compares this species with A. nigrella
(as ambigua), from which it is well differentiated,
Neotropical Region
particularly by the larvae.
Surprisingly, only a single species, A. parvicornis,
has been reared anywhere in South America from a
Agromyza cinerascens group
known host. This species was present in small num-
A cinerascens Macquart (cf. Spencer, 1976a). bers on Zea mays in Guadeloupe when I was there
Hosts: II, Poeae, Triticeae, widespread in Europe in February, 1987.
but local. Male genitalia (Fig. 1349) showing Mines have been found in Venezuela near Mara-
close relationship with the two following species cay on Panicum sp. and on Brachiaria sp. in Argen-
but also with some species on the Fabaceae and tina (both VI, Paniceae). Also, I reared four speci-
Rosaceae. mens at Sao Paulo, Brazil from an unidentified grass
A intermittens (Becker, cf. Spencer, 1976a). Hosts: which are very close to A. venezolana Spencer
II, Bromeae, Triticeae, widespread in Europe, (1963c) described from a male caught in the Botani-
also in North Africa (Tunis). Male genitalia (Fig. cal Gardens, Caracas, Venezuela.
1350). This species has been confused with In all, nine Agromyza species have been de-
A. luteitarsis but Aguilar et al. (1976), who inves- scribed in South America which from their genitalia
tigated this species and others attacking cereals in are clearly grass feeders. The most interesting is
the Paris area, discovered a difference in the A. fusca Spencer (1963a) known from Nova Teu-
colour of the puparia in two species - in A. inter- tonia, Brazil, in which the halteres are dark brown
mittens it is uniformly brown, while in A. lutei- (normally yellowish-white in this genus) and the
tarsis the front segments are blackish. wing (Fig. 1355) is strongly infuscated.
Aluteitarsis (Rondani, cf. Spencer, 1976a). Hosts:
II, Triticeae, widespread in Europe but local.
Afrotropical Region
Male genitalia (Fig. 1351).
Agromyza species on Poaceae are well represented
throughout Africa. Hosts are known of the four spe-
Nearctic Region
cies discussed below, and in addition four described
In addition to the four holarctic species, A. albi- species are certainly grass-feeders. Mines on three
pennis, A. ambigua, A. lucida and A. nigrella, two further hosts are also recorded.
species have been reared and 15 further species
A catherinae Spencer (1959). Host: VI, Paniceae,
have been described with unknown hosts in the Poa-
probably Setaria. This species was described
ceae, six in the U.S. and nine in Canada which are
from Natal and later I found numerous mines on
not discussed here.
S. chevalieri at Pietermaritzburg and Pretoria,
A parvicornis Loew (cf. Spencer, 1973a). Hosts: and also at Victoria Falls, Zimbabwe.
VI, Andropogoneae, Paniceae, widespread in A pallidifrons Spencer (1959). Hosts: VI, Paniceae.
U.S., extending north to British Columbia and Described from caught specimens in Zimbabwe
Ontario in Canada and south to the Caribbean and Zaire, later identified from Bombay, India ex
(Guadeloupe). Male genitalia (Figs. 1352, 1353): Sorghum and Calcutta (caught) and from Pretor-
distiphallus unusually short and broad. This spe- ia, South Africa reared from Urochloa panico-
"
'''~\,\\\\,\,,~\\\,,~\,~\'''''''''''''''''~~~'~
I", II"" II, '" 'II ,\\, 11""\\\\\,,
1356
1355
1358
1361 1359
1362
,,;~--- -
1363 1364
Fig. 1355. Agrornyzafusca: wing.
Fig. 1356. Agrornyza pallidifrons: aedeagus, ventral view.
Fig. 1357. Agrornyza penniseti: aedeagus.
Fig. 1358. Agrornyza ceylonensis: aedeagus.
Figs. 1359, 1360. Agrornyza panici: 1359, aedeagus, side view; 1360, same, ventral view.
Fig. 1361. Agrornyza pennisetivora: aedeagus.
Fig. 1362. Agrornyza rnellita: aedeagus.
Fig. 1363. Agrornyza venusta: aedeagus.
Fig. 1364. Agrornyza sp.: leaf mines on Oplisrnenus cornpositus.
Order POALES 363
ides. Male genitalia: Fig. 1356. ceae, reared from Pennisetum typhoides, Delhi,
A. penniseti Spencer (1959) (= A. graminivora India. Male genitalia (Fig. 1361) well differentia-
Spencer). Hosts: VI, Paniceae, reared from Penni- ted from those of A. penniseti in Africa (Fig.
setum purpureum ('Elephant grass'), Cameroun 1357).
and Uganda and from P. typhoides ('millet'),
In addition, four species which are certainly
Nigeria, also from Digitaria adscendens, Senegal.
grass-feeders but with the hosts unknown have been
The puparium is unusual with the anal segments
described in the Oriental Region (Burma, Taiwan).
curving downwards and the posterior spiracles,
Four further grass-feeders have been described in
each with three pores, lying ventrally. Male geni-
India by Indian workers (cf. Singh and Ipe, 1973)
talia: Fig. 1357.
but unfortunately the identifications are so inac-
A. susannae Spencer (1959). Host: VI, Paniceae,
curate that, without revision, the records cited are of
described from caught female in Zambia. Subse-
little value. For example the host of A. nigripes was
quently identified (tentatively) from Urochloa pa-
given as Potentilla (Rosaceae) and of A. phragmi-
nicoides, Pretoria, South Africa.
tidis as Oplismenus. Here it is at least worth noting
Mines on three further hosts (colI. K.A.S.) almost that mines are known on Oplismenus and Setaria.
certainly represent Agromyza spp.:
(i) Sporobolus filipes (V, Eragrostideae), Game Australia/Pacific Region
Park, Nairobi, Kenya, 3.ii.64. Sasakawa (1963b)
The genus Agromyza is greatly reduced in the south-
described Melanagromyza sporoboli from
ern hemisphere and this is particularly apparent in
American Samoa and Fiji from specimens
Australia with three species, and New Zealand with
caught on Sporobolus. The name is misleading,
none.
because Sporobolus is quite certainly not the
A. papuensis Sasakawa (1963c) was described
host of this species (now transferred to Ophio-
from SE Papua New Guinea. I found mines of an
myia).
Agromyza sp. at a number of localities and in PNG
(ii) Panicum maximum (VI. Paniceae), Victoria
on Setaria palmifolia (VI, Paniceae), in one case
Falls, Zimbabwe, 1.ii.64.
with two larvae, and I am satisfied that this is the
(iii) P. calvum, Aberdare Mts, Kenya, 13.x.71. This
host of A. papuensis.
mine is very distinctive, centrally filled with
Of the three species known in Australia, two are
greenish-black frass, and certainly represents a
grass-feeders. A. mellita Spencer (1977c) and A. ve-
sp.n.
nusta Spencer (1977c) were both described from
Queensland and the male genitalia (A. mel/ita: Fig.
Oriental Region 1362, A. venusta: Fig. 1363) are characteristic of all
grass-feeders elsewhere. Mines of an Agromyza sp.
Six Agromyza species have been described from
were found at Clyde Mt., NSW on Oplismenus com-
known hosts and six further species have been re-
positus (Fig. 1364) and these may well represent
corded, which are certainly grass-feeders.
one or other of these two species.
A. ceylonensis Spencer (1961b). Host: I, Bam-
Two species have been reared from unidentified
buseae, described from a caught specimen in Sri
grasses:
Lanka, later reared from Bambusa vulgaris at
Namkum, India (Sehgal). Male genitalia: Fig. 1. A. venezolana Spencer (1963c), described from a
male from Venezuela, later reared at Sao Paulo,
1358.
Brazil (K.A.S.).
A. hordei Spencer (1961b). II, Triticeae, reared
2. A. sahyadriae Ipe (cf. Singh and Ipe, 1973), de-
from Hordeum at Delhi, India. Only single female
scribed from a single female reared 'ex grass', Ke-
known.
rala, India.
A. oryzae (see above in Palaearctic Region) has
been recorded in Thailand and Taiwan. In all, 76 Agromyza species are known as feeders
A. pallidifrons Spencer (1959). VI, Andropo- on Poaceae, of which the hosts are known of 33. It
goneae, reared from Sorghum at Bombay, also can be confidently assumed that the further 43 spe-
caught at Calcutta; known in Africa (Fig. 1356). cies accepted as grass feeders are all restricted to the
A. panici de Meijere (cf. Spencer, 1986c). Host: VI, Poaceae. None are known on Cyperaceae. Within
Paniceae, reared from Panicum palmifolium in the family, the two subfamilies Pooideae and Pani-
J ava, later caught in northern Thailand. Male gen- coideae support the majority of species, with only
italia (lectotype, designated by Spencer, 1986c): A. ceylonensis (Bambusa) and A. oryzae (Oryza,
Figs. 1359, 1360. Zizania) known on the Bambusoideae, and with five
A. pennisetivora Spencer (1961b). Host: VI, Pani- species known on the Arundinoideae (Phragmites).
364 Superorder COMMELINIFLORAE
It is noteworthy that in Africa the hosts of the four species on Poaceae. This is the only known case of a
species known are all in the Panicoideae but in species of obvious poaceous origin feeding on any
Kenya two unidentified species are known, on the family of dicots.
Chloridoideae (Sporobolus) and Panicoideae (Pani-
Cerodontha is now split into seven subgenera,
cum).
following Nowakowski's detailed studies (1973). /c-
With the high degree of host-specificity on sub-
teromyza, Butomomyza and Dizygomyza are known
families, tribes and even subtribes, it is perhaps sur-
primarily as feeders on Cyperaceae and in smaller
prising that two obvious sister-species, A. oryzae in
numbers on Iridaceae, Juncaceae and Poaceae,
Japan and eastern Asia and A. alunulata in Europe
while the subgenera Cerodontha, Xenophytomyza,
should have hosts in different subfamilies. Three
Poemyza and Phytagromyza feed almost exclusively
species have a particularly wide host range, mainly
on Poaceae. Only a single Poemyza in the U.S.A. is
on the Pooideae or in two cases even crossing sub-
known on Cyperaceae (see above) and only one
family boundaries. A. albipennis is known on nine
Butomomyza and four Dizygomyzas are known on
genera in three tribes in the Pooideae and on one
Poaceae (see below).
genus (Setaria) in the Paniceae in the Panicoideae;
Here, the subgenera are treated in the phylogene-
A. nigrella is known on 12 genera in three tribes in
tic sequence proposed by Nowakowski but the indi-
the Pooideae and also on Setaria, while A. rondensis
vidual species, as with Agromyza, for ease of refer-
has been recorded on eight genera in four tribes in
ence are discussed in alphabetical order by Regions.
the Pooideae.
For the sake of brevity host genera are not normally
The distribution of the grass-feeding Agromyzas
included in the discussion below but all are included
is predominantly in the northern hemisphere, Eu-
in Table 91.
rope and Canada having most species, with a
Cerodontha s.s. is the largest subgenus, with 54
marked reduction in South America, southern Afri-
known species but of these the hosts are only re-
ca and Australia, and none in New Zealand. This
corded of nine. Adults may be caught in large num-
could suggest a relatively recent origin in the north-
bers but larval feeding is difficult to detect, taking
ern hemisphere, with insufficient time for any con-
place within the leaf sheath and thus normally not
siderable dispersal and subsequent radiation in the
clearly visible. The subgenus has a virtually cosmo-
south. Nevertheless, with four holarctic species -
politan distribution in temperate areas, from Swe-
A. albipennis, A. ambigua, A. lucida and A. nigrella
den and Alaska to southern Chile, large populations
- this group is sufficiently old to have been able to
are present in the mountains in Africa and India,
disperse naturally across Beringia.
and the subgenus is well represented in Australia
The homogeneity of the main ambigua-nigripes
and reaches New Zealand.
group is remarkable, with essentially similar struc-
ture in the male genitalia of species in the north (ct.
Figs. 1358, 1361) and those in Australia (d. Figs. Palaearctic Region
1362, 1363). The sister-group of these species re-
Hosts are known of four of the 21 species known in
mains a puzzle. The switch of an ancestral species
the Region.
from the dicots must have occurred early for the
grass-feeders to have proliferated and dispersed the Cer. (Cer.) denticornis (Panzer, ct. Spencer, 1976a).
way they have. I suggest that this must have taken Hosts: II, 13 genera in the tribes A veneae, Poeae
place at the latest in mid-Tertiary and no obvious and Triticeae. Very common in Europe, extend-
relative on the dicots is now detectable. A possible ing to North Africa, also present in Japan. Third
candidate might be a species on the Urticaceae. Spe- antennal segment with characteristic spine (Fig.
cies on this family are certainly old, with representa- 1365) present in most species in the subgenus but
tives known in Europe and North America. There is replaced by an angular projection in others (Fig.
considerable similarity in external characters but the 1373). Male genitalia (Fig. 1366) characteristic of
genitalia are now highly differentiated but this most species in the subgenus.
should not be unexpected over such a time span. Cer. (Cer.) fulvipes (Meigen, ct. Spencer, 1976a).
The switch from Fabaceae or Rosaceae of an ances- Host: II. Poeae, only recorded host Poa. Wide-
tor of the cinerascens-group might well be more re- spread and common in Europe, also present in
cent. The group is only known in Europe and Japan Japan. Closely related to C. denticornis.
and the genitalia of some species on these two fami- Cer. (Cer.) hennigi Nowakowski (ct. Nowakowski,
lies show obvious similarity with those of the ciner- 1973). Host: II, Aveneae, only recorded host
ascens group. Calamagrostis epigeios. Widespread in Europe,
A. abutilonis, known on Abutilon (Malvaceae) in but uncommon. An isolated species, confirmed
Kenya, has genitalia (Figs. 272, 273) which are in- by the male genitalia (Nowakowski, 1973: fig.
distinguishable from those of the many Agromyza 110).
Order POALES 365
Cer. (Cer.) phragmitophila Hering (cf. Nowakowski, C. denticornis but the genitalia are distinctively dif-
1973). Hosts: IV, Arundineae, primarily on ferent, with the paired distiphallus greatly enlarged
Arundo donax, also on Phragmites. Southern Eu- (Fig. 1369). Other species in the Himalayan area are
rope, related to C. denticornis. even more closely related to C. denticornis.
Further species are known in Europe, the Azores,
eastern USSR and Mongolia, with no indication of Australian/Pacific Region
their hosts but these can be assumed to be grasses.
Eight species are known in Australia, two in the
denticornis-group, six in what may be termed the
Nearctic Region robusta-group.
Cer. (Cer.) australis Malloch (cf. Spencer, 1976b;
Three Cerodontha species are known in the United
1977c) is only known in Australia in the Sydney
States and NW Canada, and three further species
area but is widespread virtually throughout New
were described from mountains in Mexico (Spencer,
Zealand. No hosts are known in Australia but in
1977b). Of these, the hosts are known only of Cer.
New Zealand it feeds commonly on the introduced
(Cer.) dorsalis (Loew, cf. Spencer and Ste~skal,
grasses Dactylis (II, Poeae), Hordeum and Triticum
1986) which has a wide host range on the POOldeae
(II, Triticeae) and I found it also on one native spe-
and h~s also been recorded on the Chloridoideae
cies, Poa anceps. In Australia C. australis was wide-
(V) and Panicoideae (VI). It is unusual in having a
ly misidentified, and most specimens I examined
pale eastern form and a d~rk wester~ form which
proved to represent Cer. (C~r.) milleri Spenc~r
meet in Colorado, and thIS colour dIfference has
(1973a). No hosts of this speCIes are know? and It
not been satisfactorily explained. The species has a
will be of interest to discover to what extent ItS hosts
wide range from Canada, throughout the. U:S. to
overlap with those of C. australis. The genitalia of
Colombia Brazil and Chile. The male gemtaha are
the two species indicate their close relationship but
distinctiv~ (Fig. 1367) and the posterior larval spir- are quite distinct (c. australis: Fig. 1370; C. milleri:
acles show an interesting proliferation, each with up
Fig. 1371).
to 15 minute pores.
Although no hosts are known of the six species in
the robusta-group (cf. Spencer, 1977c), I assume
Neotropical Region that these will also be grasses. These species are
larger and stouter than those in the denticornis-
Cerodontha is well represented from Colombia to
group and represent an isolated complex in Aus-
southern Chile, with nine species but only Cer.
tralia, with very different genitalia (c. robusta: Fig.
(Cer.) flavifrons (Philippi, cf. Spencer,. 1982! has
1372). Also the spine on the third antennal segment
known hosts, three genera in the subfamIly POOldeae
is here replaced by a stout projection (Fig. 1373).
(Aveneae). Widespread and common in Chile from
Two further species are known in New Zealand,
north of Santiago to Tierra del Fuego and recently
one in the denticornis-group, the other Cer. (Cer.)
identified in the Falkland Islands. The genitalia (Fig.
angustipennis Harrison (cf. Spencer, 1976b) with
1368) show some similarity with C. dorsalis which is
Holcus as its only recorded host, is entirely different
present in northern Chile.
both in colour, chaetotaxy and in the male genitalia
(Fig. 1374). This species is clearly derived fro~ an
Afrotropical Region early immigrant and I have speculated on possIble
relationship with species known in high mountains
Five species are known in Africa from W est Afri~a,
in Mexico (Spencer, 1977b).
the Abyssinian Highlands, to East and South Afnca
(Spencer, 1986a). Surprisingly, non.e ha.ve. been
The wide distribution of this subgenus throughout
reared but relationships with C. dentlcorms III Eu-
the world indicates early colonization of the Poa-
rope are obvious and it is clear that in the past an
ceae but dispersal has doubtless been facilitated by
ancestral species has dispersed to the s~uth and by
the virtual ubiquity of suitable hosts. Nevertheless,
mountain hopping reached southern Afnca.
there is considerable discrimination between the six
host subfamilies and also between tribes and even
Oriental Region genera. C. dorsalis has the widest host range, feed-
ing on 15 genera in the subfamilies Pooideae, Chlor-
Seven species are known in India and Thailand
idoideae and Panicoideae. C. denticornis in Europe
(Singh and Ipe, 1973; Spencer, 1986c), of which on-
and Japan in contrast is restricted to the Pooideae,
ly Cer. (Cer.) ipei Garg (cf. Spencer, 1986~) has
feeding on 13 genera in three tribes. Several species
been reared, from Setaria glauca (VI, Pamceae).
are known only on a single tribe or genus but it re-
This species superficially resembles the European
mains to be established whether this accurately re-
366 Superorder COMMELINIFLORAE
1365
1367
1368
1369
\,:ffi, ........
1371
1374
1373
1372
Order POALES 367
flects their host-specificity or is merely due to Bambuseae, only known in Japan on Arun-
undercollecting. It seems probable that C. hennigi dinaria. The genitalia illustrated by Sasakawa
on Calamagrostis, C. phragmitophila on Arundo (1961: fig. 49d) confirm that this species on the
and Phragmites, and probably C. ipei on Setaria do Bambusoideae is highly differentiated from spe-
have a restricted host range on these genera. cies on the Pooideae.
C. (Po.) calamagrostidis (cf. Nowakowski, 1973).
Xenophytomyza is a small subgenus, with three
Hosts: II, A veneae, Europe. Male genitalia with
species known in Europe and one in North Ameri-
long, curving aedeagus resembling that of Po.
ca. There are no published records of the hosts of
phragmitidis (Fig. 1377, as Po. spenceri). Nowa-
these species but it has been clear that they must be
kowski also described Po. spenceri on Calama-
grass-feeders, C. (Xen.) atronitens (Hendel) always
grostis and Po. tschirnhausi on Alopecurus with
being common in meadows in early spring. This spe-
generally similar genitalia but these were both
cies, C. (Xen.) biseta (Hendel) and C. (Xen.) venturii
synonymised with Po. calamagrostidis by Zlobin
Nowakowski are all discussed by Spencer (1976a).
(1984a).
The genitalia (Fig. 1375) show general resemblance
C. (Po.) imbuta (Meigen, cf. Spencer, 1976a, as des-
to those of species in sg. Cerodontha.
champsiae Spencer, 1957c): Host: II, Aveneae,
Recently a colleague has informed me that he has
only known host Deschampsia. Male genitalia
discovered the hosts of the three European species
(Fig. 1378) showing close relationship with Po.
but at this stage he is unwilling to divulge the genera
atra and beigerae.
of grasses serving as hosts, proposing to publish this
C. (Po.) incisa (Meigen, cf. Spencer, 1976a). Hosts:
himself later.
I, II (5 tribes), IV and VI, known on 28 genera
Subgenus Poemyza is only slightly smaller than and thus with the widest host range of any species
sg. Cerodontha, with 45 known species, of which 26 in the genus. Common in Europe, present in
have been reared, 20 being in Europe or Japan. Poe- Japan (as okazakii, cf. Sasakawa, 1961) and
myza is also widely represented outside Europe but North America. Male genitalia: Fig. 1379, asso-
significantly declines in numbers in the southern ciating the species with Po. pygmaea.
hemisphere. Although two species are known in C. (Po.) inconspicua (Malloch, cf. Spencer, 1976a).
Papua New Guinea, the subgenus is not known in Host: II, Triticeae, uncommon in Europe (Fin-
Australia or New Zealand. In the Neotropical land, Leningrad area, Poland) but widespread in
Region two species have recently been discovered in USA and Canada. Host Agropyron repens in
Guadeloupe (see below and Spencer et al., 1990). Poland, once reared from Agropyron in Colora-
do. Male genitalia very similar to Po. atra (Fig.
1380).
Palaearctic Region
C. (Po.) lapplandica (Ryden, cf. Spencer, 1976a)
C. (Po.) alpina Nowakowski (cf. Nowakowski, (= tatrica Nowakowski). Hosts: II, Aveneae,
1973). Hosts II, Aveneae, Poeae, only known in Poeae, only confirmed on Calamagrostis and Fes-
Poland. Closely related to the common Po. mus- tuca. Closely resembling Po. pygmaea, male geni-
cina. talia (Fig. 1381).
C. (Po.) atra (Meigen, cf. Spencer, 1976a). Hosts: II, C. (Po.) lateralis (Macquart, cf. Spencer, 1976a).
Aveneae, common in Europe, primarily on Cala- Hosts: II, Triticeae, widespread in Europe, also
magrostis, also on Agrostis but records on other present in Japan. Closely resembling Po. super-
hosts represent misidentifications. Male genitalia: ciliosa, also in genitalia (cf. Fig. 1387) but distin-
Fig. 1376, showing close relationship with Po. in- guishable by the conical anal projection (Fig.
conspicua. 1382, contrast superciliosa: Fig. 1388).
C. (Po.) beigerae (cf. Nowakowski, 1973). Hosts II, C. (Po.) melicae (cf. Nowakowski, 1973). Hosts: II,
A veneae, central Europe. Closely resembling Aveneae, Meliceae, Poeae; IV, Arundineae.
Po. atra on the same hosts. Widespread in Europe, earlier confused with Po.
C. (Po.) bisetiorbita (cf. Sasakawa, 1961). Host: I. pygmaea which it closely resembles.
Figs. 1365, 1366. Cerodontha (Cer.) denticornis: 1365, third antennal segment; 1366, aedeagus.
Fig. 1367. Cerodontha (Cer.) dorsalis: aedeagus.
Fig. 1368. Cerodontha (Cer.) flavifrons: aedeagus.
Fig. 1369. Cerodontha (Cer.) ipei: aedeagus (Garg).
Fig. 1370. Cerodontha (Cer.) australis: aedeagus.
Fig. 1371. Cerodontha (Cer.) milleri: aedeagus.
Figs. 1372, 1373. Cerodontha (Cer.) robusta: 1372, aedeagus; 1373, third antennal segment.
Fig. 1374. Cerodontha (Cer.) angustipennis: aedeagus.
Fig. 1375. Cerodontha (Xen.) atronitens: aedeagus.
368 Superorder COMMELINIFLORAE
1376
.:' -.
1378
1380
1382
;;fir;
~.~-- . ". .."
'
1384
Order POALES 369
1386
1387
1389
1390
1392
Order POALES 371
isolation, while Po. orbitona has probably diversi- the male genitalia, larvae and biology of Poemyza,
fied on its primary host Oryza. Dizygomyza and Phytag. flavocingulata, Nowa-
kowski (1972) concuded that this species is most
closely related to Poemyza. The male genitalia of
Oriental Region
Phytag. flavocingulata (Fig. 1396) differ consider-
Seven Poemyza species are known in India, Malay- ably from those of P. frankensis (Fig. 1397) in North
sia, Sri Lanka, the Philippines and Taiwan. The America (no host known) but it can be accepted that
hosts are known of two: the two are sister-species.
e. (Po.) javana (de Meijere, cf. Spencer, 1986c).
Although subgenus Butomomyza is dominant on
Host: VI, Paniceae, reared from Panicum in Java.
Cyperaceae, it is represented on the Poaceae by the
Male genitalia (Fig. 1393) exceptionally long.
single species, C. (But.) rohdendorfi (cf. Nowa-
e. (Po.) oryzivora Spencer (1973a) (= Po. flavior- kowski, 1973) feeding on Festuca and Poa in the
bitalis Spencer, 1975 from Sri Lanka). Host: I,
Poeae; it is only known in Finland and Poland. This
Oryzeae, once reared from rice in Malaysia. The
species is considered to be closely related to
male genitalia (Figs. 1394, 1395) show this to be
C. (But.) angulata feeding on Carex but the male
an isolated species, although it superficially
genitalia (Fig. 1398) are well differentiated.
resembles others in the subgenus.
The subgenus Dizygomyza is likewise found pre-
It is now clear from the genitalia illustrations
ponderantly on Cyperaceae but with four species on
(Spencer, 1975: figs. 6, 7) that Cer. (Po.) flavior-
Iridaceae and six on Juncaceae. Four species are
bitalis Spencer is identical with Po. oryzivora and
known to feed on Poaceae:
this new synonymy is formally established in the
Appendix below.
e. (Diz.) crassiseta (Strobl, cf. Nowakowski, 1973)
(= poae Hering). Hosts: II, Poeae, widespread in
Poemyza is poorly represented in the Pacific/
Europe but uncommon. Distinctive species with
Australian Region, with no species known in Aus-
short, thickened arista, male genitalia with
tralia or New Zealand and just two, with no indica-
aedeagus unusually short (Fig. 1399).
tion of their host, recorded in Papua New Guinea.
With 26 described Poemyza species all host-
e. (Diz.) chaixiana (Hering, cf. Spencer, 1978).
Host: II, only known on Poa chaixii, Germany
specific on Poaceae, it is accepted that the 19 fur-
and Poland. There has been confusion with this
ther species known in the subgenus all feed on
species, as in the long type series reared by
grasses. The only known exception is Poemyza mac-
Groschke near Stuttgart, two species were repre-
minni in Mississippi which has transferred to Carex
sented, chaixiana and jasciata, with very different
as host (see under Cyperaceae above).
genitalia. The genitalia of Diz. chaixiana (Nowa-
Subgenus Phytagromyza consists of only two spe- kowski, 1973: fig. 171) appear virtually identical
cies, C. (Ph.) flavocingulata in Europe and eastern with those of Diz. caricicola (Fig. 1327), while in
Canada, and C. (Ph.) frankensis in western Canada, Diz. fasciata (Fig. 1400) the aedeagus is shorter
California and Wyoming. This subgenus is unique in and the species are obviously distinct.
Cerodontha in having the costa terminating at vein Nowakowski (1972) suggests that Diz. chaixiana
R4 + 5. Hendel (1920) erected Phytagromyza with may merely represent a 'bionomic subspecies' of
flavocingulata as type-species and included in the caricicola but makes no reference to this in his
genus other species with hosts in the Caprifoliaceae, later Monograph (1973). If this was accepted, it
Dipsacaceae, Rubiaceae and Salicaceae now placed would represent one of two cases of a species
in Paraphytomyza (see these families above). Nowa- feeding on both Cyperaceae and Poaceae (also
kowski first noted that the genitalia of flavocingu- known only in Chromatomyia pseudomilii on
lata, a grass-feeder, associated it with Cerodontha, Carex and Agrostis in Alaska, and on Deschamp-
particularly Poemyza. sia in East Germany (see under Carex above and
C. (Ph.) flavocingulata is restricted to the Pooi- under Chromatomyia below).
deae, with hosts in the A veneae, Poeae and Triticeae e. (Diz.) elbergi (cf. Nowakowski, 1973). Host: II,
(see Table 91). After an analysis of 17 characters of
Fig. 1386. Cerodontha (Po. ) pygmaea: aedeagus.
Figs. 1387, 1388. Cerodontha (Po.) superciliosa: 1387, aedeagus; 1388, anal plate.
Fig. 1389. Cerodontha (Po.) bambusae: aedeagus (Martinez).
Fig. 1390. Cerodontha (Po.) aristella: aedeagus.
Fig. 1391. Cerodontha (Po.) orbitona: aedeagus.
Fig. 1392. Cerodontha (Po.) setariae: aedeagus.
Fig. 1393. Cerodontha (Po. ) javana: aedeagus.
Figs. 1394, 1395. Cerodontha (Po.) oryzivora: 1394, aedeagus, side view; 1395, distiphallus, ventral view.
372 Superorder COMMELINIFLORAE
1396
1398
............ f
....... _....•'/
1397 1399
1400 1401
1403
1404
Order POALES 373
only known on Sesleria tatrae in Poland, also re- L. pedestris Hendel (cf. Spencer, 1976a). II, only re-
corded in U.S.S.R. and Mongolia. Closely related corded host Deschampsia. Widespread in Europe
to C. (Diz.) crassiseta. but commoner in south. Male genitalia (Fig.
C. (Diz.) fasciata (Strobl, cf. Spencer, 1976a; 1978): 1406) well differentiated.
Host: II, Poa chaixii. Reared by Groschke to- L. phryne Hendel (cf. Spencer, 1976a). Hosts: II,
gether with Diz. chaixiana near Stuttgart. Male only known on Poeae and Triticeae, widespread
genitalia (Fig. 1400) close to and I consider iden- but uncommon. Male genitalia (Fig. 1407) only
tical with the population in Canada and U.S.A. slightly differentiated from L. orbona.
(Fig. 1401) originally misidentified as chaixiana
It is interesting that these five species all occur
(Spencer, 1969b) (cf. Spencer and Steyskal,
only on the Pooideae. Their genitalia confirm their
1986).
very close relationship, L. flaveola being the most
Diz. crassiseta and elbergi belong to one group, differentiated species.
Diz. chaixiana and fasciata to two further groups, all The polyphagous L. trifolii has recently been
three with close relatives on the Cyperaceae. Thus found mining young barley plants (II, Triticieae)
on at least three occasions ancestral species on growing in a glasshouse where it was present in pest
Cyperaceae appear to have transferred to Poaceae. proportions on cultivated Gerbera in Italy (L. Suss,
1988). This is the second case of this species attack-
The large genus Liriomyza is not represented by
ing a grass - it has been recorded on A vena in Flori-
many species on Poaceae but some of those present
da (see below).
can occur in large numbers and are widespread and
In North America six Liriomyza species can be
common. Ten species have been reared from
accepted as grass-feeders but only two have been
grasses, and from their distinctive genitalia 10 more
reared:
can be deduced as grass-feeders. Two common spe-
cies in this group are L. infuscata Hering (cf. L. septentrionalis Sehgal (cf. Spencer and Steyskal,
Spencer, 1976a) and L. richteri Hering (cf. Spencer, 1986). Hosts: II, on three genera in the three
1976a). tribes Bromeae, Poeae and Triticeae but probably
Relationships with species on dicot families are others also, common in western Canada, Cali-
not normally obvious but one clear case is L. margi- fornia and Colorado. Male genitalia (Fig. 1408)
nalis in the United States and South America which showing close relationship with L. flaveola in Eu-
is closely related to L. blechi on Acanthaceae in the rope.
United States and to L. andina on Plantago in Chile L. marginalis Malloch (cf. Spencer and Steyskal,
(see below). 1986). Hosts: V, VI, primary host appears to be
Eight species are certainly grass-feeders in Eu- Paspalum. Not uncommon in southern parts of
rope, of which five have been reared and are dis- the U.S.A., extending south to Caribbean and
cussed below: widespread in South America. Male genitalia
(Fig. 1409) and epandrium (Fig. 1410), showing
L. flaveola (Meigen, cf. Spencer, 1976a). Hosts: II,
this to belong to a different group and closely re-
on five genera in four tribes. One of the com-
lated to species on Acanthaceae and Plantagin-
monest species on grasses. Male genitalia (Fig.
aceae (see above). The anterior spiracles of the
1402) with curvature of the distiphallus charac-
larva (Fig. 1411) are also unique to this group.
teristic of the group.
L. trifolii (Burgess). This highly polyphagous species
L. graminivora Hering (1949c). Hosts: II, only
on 25 dicot families has been recorded in Florida
known on Hordeum and Poa, in Berlin. Male
on A vena. It is not clear whether this is a regular
genitalia (Figs. 1403, 1404) show close relation-
alternate host or represents an isolated exception.
ship with the more widespread species L. orbona
and L. pedestris. Two Liriomyza species have been reared on
L. orbona (Meigen, cf. Spencer, 1976a). Hosts: II, grasses in South America, either in Argentina or
Aveneae, Poeae, common in spring. Male geni- Chile and four other species have genitalia charac-
talia (Fig. 1405) indicating close relationship with teristic of grass-feeders.
L. phryne.
Fig. 1396. Cerodontha (Phytag.) Jlavocingulata: aedeagus.
Fig. 1397. Cerodontha (Phytag.) frankensis: aedeagus.
Fig. 1398. Cerodontha (But.) rohdendorfi: aedeagus.
Fig. 1399. Cerodontha (Diz.) crassiseta: aedeagus (Nowakowski).
Fig. 1400, 1401. Cerodontha (Diz.) fasciata: 1400, aedeagus (Europe); 1401, same (Canada).
Fig. 1402. Liriomyzaflaveola: aedeagus.
Figs. 1403, 1404. Liriomyza graminivora: 1403, aedeagus, side view; 1404, distiphallus, ventral view.
Fig. 1405. Liriomyza orbona: aedeagus.
374 Superorder COMM
~~_ ELmwL~E
~*~"
1406
1409
1411
~-.";;,~~,,
",.~ ..
'.~:~--'''\'' :
:}-."
::-"
---::,:...
1413
1414
Order POALES 375
L. cruciata (Blanchard, cf. Spencer, 1982). II, only Ps. asiatica Spencer (1961b). Hosts: I, V, VI, wide-
known host A vena in Argentina, three specimens spread in Afrotropical and Oriental Regions.
also known in Chile (but no males). The third an- Third antennal segment (Fig. 1416) angulate, as
tennal segment is conspicuously angulate (Fig. in all species in this group, genitalia (Fig. 1417),
1412) and this led to the species being treated as leaf mines on Cynodon dactylon (Fig. 1418).
a Cerodontha (cf. Fig. 1365). Ps. atra (Meigen, cf. Spencer, 1976a). II, confirmed
L.lolii Spencer (1982). II, only known host Lolium on eight genera in the A veneae, Triticieae, wide-
perenne, widespread in central Chile. Male geni- spread in Europe, also present in U.S.A. Male
talia (Fig. 1413) characteristic of grass-feeders. genitalia (Fig. 1419), puparium (Fig. 1420) with
spinules characteristic of many species in this
Four further species in Chile can be accepted as
group.
grass-feeders. L. peullae (Malloch, cf. Spencer,
Ps. dilatata Sasakawa (cf. Spencer, 1973a). V, VI,
1982) is common in central Chile and is unusual in
known on six genera, only recorded in Fiji and
having the scutellum entirely black. The genitalia
Samoa. Male genitalia: Fig. 1421.
(Fig. 1414) indicate close relationship with L. lolii.
Ps. gujaratica Shah (1982). VI, only known host Zea
L. schwabei Spencer (1963c) is widespread in cen-
mays in India. Male genitalia (Figs. 1422, 1423)
tral Chile. The genitalia (Fig. 1415) confirm that this
distinctly different from Ps. asiatica with which
is a grass-feeder but suggest that it belongs to a sep-
the species was originally confused.
arate group.
Ps. hispanica Spencer (1973a). VI, only suspected
In the Afrotropical Region no Liriomyza species
have been reared from Poaceae but from their geni- host Sorghum. Described from Spain and Israel,
subsequently discovered in East Africa. Not ac-
talia two described species can be referred here,
tually reared but caught on Sorghum in Israel.
while in the Oriental and Australian/Pacific Region
Ps. malayensis Spencer (1973a). Host; V, only
none are known as grass-feeders.
known on Cynodon dactylon in Malaysia. Male
From the details recorded above, it is apparent
genitalia: Fig. 1424.
that Liriomyza on grasses are more numerous in the
Ps. philippinensis Spencer (1961b). Host: VI, only
northern than in the southern hemisphere, although
known on Rottboelia exaltata in the Philippines.
the species in Chile indicate some proliferation in
This species is unusual, with a number of larvae
temperate areas in South America. Their total ab-
feeding together to form a large, brownish blotch
sence in Australia and New Zealand is striking.
mine (contrast the mines with single larvae of Ps.
The genus Pseudonapomyza colonizes a restrict- asiatica (Fig. 1418)). Male genitalia characteristic
ed number of host families, including Acanthaceae of the group.
(five genera), Amaranthaceae (Achyranthes, Alter- Ps. spicata Malloch (cf. Spencer, 1973a). Hosts: II,
nanthera) and Asteraceae (Vernonia) in the dicots VI, known only on Triticum, Panicum, Sac-
but the greatest radiation has occurred on the Po a- charum, Zea but also on '14 wild grasses' in Ha-
ceae. On this family hosts have been confirmed of waii (details not recorded), widespread from
ten species, in Europe, Africa, the Oriental Region Egypt and Israel across the Old World Tropics
and Australia but 26 more are known only from and Pacific, apparently not in tropical Africa or
caught specimens. Speciation has occurred most Australia, confused in the past with Ps. spinosa.
rapidly and probably relatively recently in East Afri- Exact host range remains to be established. Male
ca, with the adults little differentiated on external genitalia (Figs. 1425, 1426) only certain way of
characters but with striking differences in the male distinguishing this from related species which can
genitalia. I recently discovered six new species in have hosts in common with it.
Kenya (Spencer, 1985a) and five further species Ps. spinosa Spencer (1973a). Hosts: II, V, records
from Tanzania await description (von Tschirnhaus). on Hordeum, Triticum, Eleusine and probably
Three species have been described in India from fe- Brachiaria. Range in part overlapping with Ps.
males alone (Singh and Ipe, 1973) and with the male spicata but present in West, East and South Afri-
genitalia unknown, positive re-identification of these ca across Oriental Region to northern Australia
species will be impossible. (Sydney) and Pacific, absent from Hawaii. Male
1416 1417
1419 1418
1420 1421
1423
1425
1424
1427 1426
1429
OrderPOALES 377
genitalia (Figs. 1427, 1428) distinctly different lae superspecies are all assumed to feed on Luzula
from Ps. spicata. and have been discussed above under Juncaceae.
Ps. zeae Spencer (1973a). Host: VI, only known on Griffiths concedes that his superspecies comprise
Zea mays, Ghana. Male genitalia: Fig. 1429. 'relatively weakly differentiated species,' identifiable
in many cases only by minor differences in the male
In Schleswig-Holstein, N. Germany an unde-
genitalia or by their host genus if this is known. He
scribed species has been caught with the host be-
suggests that this complex is the result of recent evo-
lieved to be Ammophila arenaria (II, Aveneae) (von
lution.
Tschirnhaus, 1981).
In the discussion below I follow the arrangement
With only two species known in North America
proposed by Griffiths (1980).
(both holarctic) and one on Guadeloupe, it appears
that the genus crossed Beringia relatively recently, A. Chr. milii superspecies (six of nine species with
with insufficient time for any radiation in North known hosts)
America and very limited dispersal and speciation in
Chr. milii (Kaltenbach, cf. Spencer, 1976a; Griffith.s,
the Caribbean and South America. In contrast there
1980). Hosts primarily in Pooideae (Agrostls,
has been extensive radiation in Africa, the Oriental
Holcus, Poa; also Milium). The genitalia of this
Region and the Pacific. Interestingly, the two wide
species, as in all in the group and also in the Chr.
ranging species Ps. spicata and spinosa appear to
opacella superspecies, have a pair of 'wedge-
have important host differences, with spicata pre-
shaped sclerites' between the supporting sclerites
sent on Panicum and Saccharum in the Panicoi-
of the dorsal lobe and the sclerites immediately
deae while spinosa is known on Eleusine in the
below the terminal section of the ejaCUlatory duct
Chlo'ridoideae and only on Brachiaria in the Pani-
(Fig. 1431). The anterior larval spiracles are
coideae. The two species overlap on Triticum in the
unique in this group, projecting upwards thro~gh
Pooideae. It must be stressed though that more de-
the leaf epidermis in the puparium on a COnIcal
tailed collecting and accurate identification of hosts
projection divided into two 'horns' with a total of
will be necessary for a full understanding of the host
some 25 spiracular pores (Fig. 1432). Wide-
range of this group of species.
spread in Europe from northern Scandinavia to
the Mediterranean (including Morocco).
The genus Chromatomyia is well represented on
Chr. norwegica (Ryden, cf' Griffiths, 1980). Hosts:
Poaceae, primarily in the northern hemisphere, with
Pooideae, tribes A veneae (Calamagrostis, Cinna,
19 species described mainly in Europe a~d NW Ca-
Deschampsia (as Vahlodea) and Poeae (Arcta-
nada and Alaska. Griffiths (1980) provIded a de-
grostis). Paler than Chr. milii, with upper ~alf?f
tailed study of this group, which need only be briefly
frons clear yellow. Distribution boreal-alpme m
summarized here. Of the 12 species with known
Europe, most common in the Arctic but prese~t
hosts which are discussed below, most are selective
in mountains of central Europe, and also m
in the choice of hosts, feeding on from one to four
Alberta, Canada.
genera, while the two common species Chr. fuscula
Chr. arctagrostidis Griffiths (1980). Host: only
and Chr. nigra have a far wider host range, recorded
known on Arctagrostis, Pooideae. Closely re-
on 20 and 46 genera respectively (see Table 91).
sembling Chr. milii in all characters but 'puparia
After the appearance of Griffiths' paper, I have
golden-yellow to dark red-brown'. Only known in
extended the known range of Chromatomyia on
British Columbia.
Poaceae, by discovering a mine on Brachiaria s~.
Chr. pseudomilii Griffiths (1980). Hosts: Agrostis
(Fig. 1430) (Panicoideae) at sea l~vel on the Atl~~t1c
(Alaska), Deschampsia (East Germany), also
coast of Costa Rica, near Limon, and descnbmg
Carex (Cyperaceae, see above). Male genitalia
Chr. nigrissima and Chr. subnigra from a high alti-
(Fig. 1433) with significant difference from Chr.
tude on Mt. Kenya, Kenya (see below).
milii. It remains to be established whether the lar-
Griffiths (1980) divides Chromatomyia on mono-
val feeding on both Agrostis and Carex is normal,
cots into five superspecies, each comprising from
or whether it represents a temporary, local switch
one to nine species. The six species in the Chr. luzu-
in one area of Alaska.
Figs. 1416-1418. Pseudonapomyza asiatica: 1416, third antennal seg~ent; 1417, aedeagus; 1418, leaf mines on Cynodon dactylon.
Figs. 1419, 1420. Pseudonapomyza atra: 1419, aedeagus; 1420, pup anum.
Fig. 1421. Pseudonapomyza dilatata: aedeagus. . . .
Figs. 1422, 1423. Pseudonapomyza gujaratica: 1422, aedeagus; Side view; 1423, same ventral view.
Fig. 1424. Pseudonapomyza malayensis: aedeagus. .. .
Figs. 1425, 1426. Pseudonapomyza spicata: 1425, aedeagus, ~Ide ~Iew; 1426, same ventral view.
Figs. 1427. 1428. Pseudonapomyza spinosa: 1427, aedeagus, Side view; 1428, same, ventral VieW.
Fig. 1429. Pseudonapomyza zeae: aedeagus.
378 Superorder COMMELINIFLORAE
'C~5_>~
1431
.":;
1432
('
,, '
,
,
, ,'
,!
--
"
\\,I
/
" ", ,
" " ,,
1433 1434
............
'~---~""""
(
. "." · ; .. : ; n ·
1435 ",
1430
I
,
I
I
I
I
I
/
-'
"
.
;.-- -'
/
1436 1437
\.;.&.,., - '-',-
1438 1439
379
Chr. cinnae Griffiths (1980). Only known host Cin- Alberta. Male genitalia (Fig. 1437) characteristic
na (Aveneae), Pooideae, in Alberta. Male geni- of the superspecies.
talia distinctive in the very small wedge-shaped Chr. alopecuri Griffiths (1980). Hosts Alopecurus,
sclerites (Fig. 1434). Deschampsia (Aveneae). Only known in Alberta.
Chr. torrentium Griffiths (1980). Hosts: Deschamp-
sia, Poa (Pooideae). Male genitalia differing in D. Chr. nigra superspecies
detail from others in the group but associating the Griffiths (loc. cit.) treats this as a single species but
species with Chr. cinnae. with five forms diverging from the typical West
European form - in Madeira, the Baltic, Kamchatka
There are three other species in this group but
(Siberia), Japan and North America (Pacific Coast).
with no hosts known.
These six forms may differ slightly in colour, size
B. Chr. opacella superspecies and the structure of the male genitalia but cannot be
associated with particular hosts.
Griffiths includes two species here, of which only
Chr. nigra has the widest host range of any Chro-
Chr. opaceUa has known hosts. It was described by
matomyia species on the Poaceae or indeed in the
Hendel from mountains in northern Italy and is well
genus, with records on 46 genera in 11 tribes, main-
known in Arctic Scandinavia, and also in Iceland
ly in the Pooideae but also on five genera in the
and Greenland. More recently specimens were
Arundinoideae, Chloridoideae and Panicoideae
reared in the Tatra Mountains, Poland from Trise-
(Table 91).
tum (Aveneae) and Sesleria (Poeae). The presence
The male genitalia (Fig. 1438) are uniquely com-
of wedge-shaped sclerites in the aedeagus (Fig.
plex and indicate an obscure origin for Chr. nigra.
1435) associate this group with the milii super-
Griffiths suggests the species may have evolved
species but with a small difference in the structure of
from the fuscula superspecies or even from the syn-
the genitalia, and the third antennal segment bare
genesiae superspecies which includes the highly
here and more pubescent in Chr. milii. Griffiths
polyphagous Chr. horticola. A diagnostic character
(1980) treats the two groups as separate super-
of this superspecies is the thickly pilose eyes in both
species.
sexes.
C. Chr. fuscula superspecies Two further species in this complex, Chr. nigrissi-
ma and subnigra Spencer (1985a) (Fig. 1439) were
Four species are included here, all with known hosts
described from a high altitude on Mt. Kenya, north
and characterised by the reduced chaetotaxy of the
of Nairobi, Kenya. The genitalia immediately con-
mesonotum and limited pigmentation of the
firm relationship with Chr. nigra. Both also have
aedeagus.
pilose eyes. Although no hosts are known, it is cer-
Chr. fuscula (Zetterstedt, d. Spencer, 1976a; Grif- tain that these are local grasses.
fiths, 1980). Hosts: restricted to the Pooideae but
with a wide range on 20 genera in six tribes (see E. Chr. syngenesiae superspecies
Table 91). Male genitalia (Fig. 1436) with re- Chr. horticola is the most highly polyphagous spe-
duced sclerotization of the distal sections. Wide- cies in the family, known on 35 families but occur-
spread in Europe, particularly in Scandinavia but ring most commonly on Fabaceae, Asteraceae (Fig.
absent from Britain; also present in Alberta, 1105) and Brassicaceae (see above under Faba-
Canada. ceae). It also regularly attacks Allium but is not
Chr. puccinelliae (Spencer, 1969b; Griffiths, 1980). known on the Poaceae.
Only known host Puccinellia (Poeae), a northern Griffiths (1980: 7) suggests that: 'We must postu-
New World species only known in NW Canada late that transference to monocotyledonous hosts
and Greenland. Male genitalia close to Chr. poae. has occurred two or three times' in the complex of
Chr. poae Griffiths (1980). Only known from Poa, Chromatomyia species on Poaceae and Juncaceae.
Fig. 1430. Chromatomyia sp.: leaf mine on Brachiaria sp. (Costa Rica).
Figs. 1431, 1432. Chromatomyia milii: 1431, aedeagus; 1432, anterior spiracles oflarva (Hering).
Fig. 1433. Chromatomyia pseudomilii: distal section of aedeagus (Griffiths).
Fig. 1434. Chromatomyia cinnae: distal section of aedeagus (Griffiths).
Fig. 1435. Chromatomyia opacella: aedeagus.
Fig. 1436. Chromatomyiafuscula: aedeagus.
Fig. 1437. Chromatomyia poae: distal section of aedeagus (Griffiths).
Fig. 1438. Chromatomyia nigra: aedeagus.
Fig. 1439. Chromatomyia subnigra: aedeagus.
380
Although Griffiths has clarified differences between transfer from Poaceae to Carex was fortuitous.
his superspecies on these two families, they are rela- There are just two cases where single species have
tively slight and I feel are more similar and thus switched, on the one hand from Poaceae to Cyper-
more probably monophyletic than that there is a aceae and on the other from Cyperaceae to Poaceae.
sister-group relationship between any single group C. (Poemyza) macminni in Mississippi is a typical
of the monocot feeders and any group of dicot feed- member of the large group of grass-feeding Poe-
ers. The switch of an ancestral dicot feeder to mono- myza species which is only known on Carex, and
cots would not have been easy and I can see no good conversely C. (But.) rohdendorfi in Europe is a typi-
reason for more than a single transfer. With Chro- cal member of the Carex-feeding group of Butomo-
matomyia species known on seven families of dicots myza species which is only known on Poaceae. The
(Asteraceae, Caprifoliaceae, Dipsacaceae, Gen- majority of Dizygomyza species feed on Carex but a
tianaceae, Hydrophyllaceae, Primulaceae and Saxi- few which are closely related also on Poaceae (and
fragaceae) and also on ferns, apart from Chr. horti- on lridaceae).
cola, I agree with Griffiths that the ancestor of the That the association between Cerodontha, par-
monocot feeders seems on present evidence most ticularly the subgenera Butomomyza and Dizygo-
likely to have transferred from the Saxifragaceae. myza, with Carex is ancient is indicated by the pres-
The host specialization of the Agromyzidae on ence of four holarctic species on this genus - those
the Cyperaceae provides some evidence of the simi- in Butomomyza (B. angulata, eucaricis and scirpi)
larity of the family to the Poaceae but even more of and Dizygomyza hirtae.
the distinctness of the two families. It is certainly sig- There are other sister-species each side of the
nificant that Chromatomyia pseudomilii is the sole Bering Straits, such as Diz. suturalis in Europe and
species known to feed on both families and, as this Diz. magnicornis in North America representing an
occurred at the same locality in Alaska where the even earlier dispersal across Beringia.
hosts were growing together, it may well be that the
5. Polyphagous species
In contrast to the great majority of agromyzid spe- cies can be characterised as pre-polyphagous. It will
cies which show a striking degree of host specializa- be interesting for entomologists 100 000 years
tion, a mere 16 species (0.6%) are truly poly- hence to test whether in this relatively short time
phagous. such species have developed more extensive,
The term polyphagy is used in a different sense by genuine polyphagy.
different authors. Here I treat as polyphagous spe- Time does not permit any detailed comparison of
cies which feed regularly on more than two hosts in the degree of polyphagy in the Agromyzidae with
families in different orders or subclasses. Polyphagy that found in other groups of insects but just a few
is not a static condition but is the reflection of an on- examples may be cited. Hollis (1987) mentions that
going dynamic process. I consider it has evolved in the Psylloidea host specialization is normally on a
slowly, at differing rates in individual species, also family to host-family basis. The genus Diaphorina
with differing times of origin and with differing (Aphalaridae) is exceptional, with 45 of 71 species
causes. While the original feeding of primitive spe- feeding on 18 families in ten orders in the Sub-
cies ancestral to Agromyzidae may have been poly- classes Dilleniidae, Rosidae and Asteridae. Hollo-
phagous, possibly saprophagous, early host speciali- way and Hebert (1979) discussed the host selection
zation developed, favoured by the countless new of 571 species of moths in Canada and found that
niches which became available as the angiosperms 39% are polyphagous, defining polyphagy as feed-
radiated at an increasing rate from early Tertiary on- ing on three or more families. Ehrlich and Raven
wards. The polyphagy which we known today is thus (1964) in their comprehensive analysis of the food
secondary and in the Agromyzidae is extremely plants of butterflies found a close association be-
limited. The host range can be extensive embracing tween most tribes and related host families. How-
35 families in Chromatomyia horticola, 28 in Tropi- ever, there is little general polyphagy - two excep-
comyia polyphyta and 29 in Liriomyza strigata. tional species are mentioned - Euptoieta claudia
It is clearly not easy for a species to extend its (Argynninae) feeding on ten families and Precis lavi-
host range. If it were, the rigid specialization which nia on eight (both Nymphalidae). Ackery (1988)
is dominant throughout the family in all genera, over made a detailed study of the hosts of the Nymphali-
the millions of years during which many species dae, in particular attempting to clarify problems of
have existed, would have become eroded and a classification within the 17 recognised subfamilies
reversal to ancestral polyphagy would have oc- by analysing their host plants. There is considerable
curred. host specialization and on the dicots only limited
There are many examples where females for a polyphagy. He cites Cynthia cardui as the most
variety of reasons have attempted to colonize a non- polyphagous species, feeding on nine families, par-
host. In such cases the larvae most frequently die, or ticularly on 35 genera of the Asteraceae. In the
very rarely may produce adults which successfully Coleoptera Johnson (1981) discusses the host spe-
adapt to the non-host. With adequate time this can cificity of the Bruchidae. There is a high degree of
be the basis for full speciation. In early stages, how- host specificity, 84% of species being restricted to
ever, the populations on the normal and abnormal the Leguminosae. The genus Amblycerus feeds on
hosts can be recognised as the same species. Having ten families in addition to the Leguminosae, with
adapted once to a non-host, it is probably easier to just one undescribed species feeding on four fam-
continue the process, and so genuine polyphagy may ilies. The Chrysomelidae have a very comparable
develop. I refer below to the few species known to host range to the Agromyzidae and have been much
feed regularly on just one abnormal host. When this studied. Hosts are known of some 10 000 of the
is an isolated or accidental occurrence, the term 35000 described species (28%). Jolivet (1988) con-
xenophagy has been used (Voigt, 1932), but when siders that the primitive polyphagy in this family
the two unrelated hosts are used regularly, such spe- evolved into oligophagy and monophagy, with later
382 Polyphagous species
Table 92
Hosts of polyphagous species
Liriomyza Chromatomyia
«I
"8
...
'';::;
o
Host ..c
Families ...:
..c
....i ....i ....i ....i ....i....i U
GYMNOSPERMAE
_____ .Gnetaceae __________ ---J.C----______ c____ r ___
---~--- -- ---- -
-,," --~ ---...:..-----.... ---------.:.-------- .... -- ........ -.-- ------------------_ ----.
....
POLYPODIOPHYT A . .
_____ Davalliaceae. _____________ , _____ • _-X. ____ :______________________ J____ L ___ c _____ • ____ • ____ • _____________________________ - ___ _
MAGNOliOPHYTA
MAGNOLIIDAE
.. Aristolochiaceae ... x
Menispermace.ae x X
Monimiaceae . X
.. . .... Papaveraceae X
Piperaceae x
Ranunculaceae X X X x
HAMAMEUDAE
Cannabaceae . X X
_____ Mru:aceae.__ ________ _ _____ ____ •____ c__ x. __ X _____________ . _________ ; ____ . _________ •_______________________________________ _
CA RYOPHYLLID AE
Amar.anthaceae X X X
Basellaceae ... X
CaryophyUaceae X X X X X X
.. Chenopodlaceae. X X X X X X
Nyctaginaceae x
Phytolaccaceae X
-. ____P.oLygonaceae__ -. ---~-- .-- --r--·---c ---.• --. -<-.--t------.------.----. --------------.--------------
DILLENIIDAE
Actinidiace.ae
.. Brassicaceae ... X X X .X
Capparac.eae X X X
Cucurbitaceae ... X XXX X X ..X
.. Flacourtiaceae.
Lecythirla.ceae
. . .... Loasaceae ... X X x
Malvaceae X X X X
Moringaceae X
.Myrsinaceae ... x
Passifloraceae X· x X X
Primulaceae. X
Rese.daceae.... X X
Salicaceae
Sterculiaceae.
Styracaceae
Symplocaceae
Theaceae..
Turneraceae. X
----- Ylolaceae_ .--------- i-" ----. _________ _ .-. ___ ... ________ .. _, ..X _______________ 0 _._______ __ __ _______ X-_________________ .
Continued
Polyphagous species 383
Table 92 (continued)
Hosts of polyphagous species
Host
Families
ROSIDAE.
.......... A1angiaceae ..... .
.Anacardiaceae X
. Apiaceae .. x x.xx X
...... Aquifoiiaceae ..... . ............. : .. . . .... -:....
Celastraceae X X
....... Combretaceae ..... . X
Euphorbiaceae X x X X X
Hydrangeaceae X
....... Leguminosae ... X x X X X X X X X X X X
.Linaceae X X X X
· ....... Malpighiaceae
· ........ Meliaceae... .
Myrtaceae
...Onagr.aceae ....
Oxalidaceae
Pittosporaceae
..... Rhamnaceae ...
Rosaceae
....... Rutaceae..... X ..
... Sapindaceae........ .
..... .Saxifragaceae
...... Thymeiaeaceae. X
Tropaeoiaceae X • X X
.____ LygopbJdlaceae ____ ~--- ___ ;.----~------.---.;- ___ __________________ 1____ ~ _______ x.. ____ c--X-----
ASTERIDAE
Apocynaceae x x X
....... Asclepiadaceae X
Asteraceae xt X X X X X
Bignoniaceae X X
......Boraginaceae . X X
Campanulaceae X X
Caprifoliaceae
Convoivulaceae X X X X X X
Dipsacaceae X
........ .Gentianaceae .. x
... Hydrophyllaceae .. X .. X
Lamiaceae X X X X X
........ LabeIiaceae... x.
Laganiaceae X
Myoporaceae X
.Oleaceae ... x
Piantaginaceae X X X
....... Poiemoniaceae . X X X X X
........ Rubiaceae ...... . x X
Scrophulariaceae X X X X
· ....... SOianaceae .. . ........... L .. X ........ X X X X X X X X X
Valerianaceae X· X X
......V.erhenaceae...............·.. :X .......... : ...... x ................... ·...x ........ x .. x ............... x ....... J:ontioued
384 Polyphagous species
Table 92 (continued)
-ci
Tropicomyia Liriomyza '2 Chromatomyia
"
t::
'"t::
'r;j cd
'"
cd cd C cdO
....
"',- u 0
cd cd
::::0
'P
.... ~
..c: '"cd ~
C '"
.... +=l
...
cd u .... u
'P
';:: ;;
~
.D -0
t::
~ O~
..c:,_ ....
'E
;:I
"'u 0.
" 0 cd
S'u 0 ;.
... .... ~ "'u
;>, 0 .0 0
Host 0 -0
..c:
bIl
..c:
°5 "5
0
0,- ~ 00,
;>,
.... 0
"E '5 '.;J t::;:1
... c 0 cd 0 cd ....
Families cd;>, .D 0 c() ..c: '" '" '" ><:0 ...;
..c:
~~ ~ ..j ..j .J ..j ..j ..j ..j~ U
MONOCOTYLEDONS
.. Alliaceae ... X X
Alstroemeriaceae X
Araceae ..
Dioscoreaceae..
lridaceae X
Poaceae.... X
Smilacaceae
Typhaceae X
Total Families 18 29 25 12 3S
a reversion in some species to secondary polyphagy, The relationship between T polyphaga (Oriental
However, no statistics appear to be available on the Region) and T atomella needs further investigation.
degree of such polyphagy, T polyphyta is restricted to Australia, Papua New
There is thus an infinite range of host specializa- Guinea and Pacific islands, while T flacourtiae oc-
tion in phytophagous insects, varying from strict curs in Africa and Madagascar. It will be seen that
monophagy to varying degrees of polyphagy, as in several of the Tropicomyia species have family hosts
the Agromyzidae, No detailed analysis of the host .in common in the different Regions in which they
range of the polyphagous Agromyzidae has hitherto occur.
been undertaken and the most significant fact now The recent discovery by Sasakawa and Pong
emerging is just how limited the polyphagy is, even (1988) of T atomeUa on Gnetum, representing the
in the species with the widest host range. Many only record of any leaf miner on the Gymnosper-
widespread families are totally avoided. Also only mae, is of considerable interest. I have checked the
species in the three leaf-mining genera Tropicomyia, single specimen of T polyphaga known on the fern
Liriomyza and Chromatomyia are polyphagous. Nephrolepis (Davalliaceae) in northern India and
Phytomyza, the largest genus, shows no indication am satisfied that this is identical with the types
even of pre-polyphagy. Similarly, in the primarily reared from Symplocaceae and Asclepiadaceae in
stem-feeding genera Melanagromyza and Ophio- Sri Lanka. Between them these five Tropicomyia
myia there is no polyphagy, with only isolated cases species have colonized a wide scatter of families but
of extended oligophagy and xenophagy. it is remarkable that only T atomella and T poly-
The host ranges of the 16 polyphagous species phaga are known on Asteraceae (Eupatorium) at
are shown in Table 92. The host genera of these spe- two localities in Sri Lanka and Malaysia.
cies are all given in the family Tables above but Of the ten Liriomyza species, L. bryoniae, strigata
some of the commoner species are not mentioned and xanthocera (= crucifericola) are restricted to Eu-
under the individual families in the main text. rope; L. huidobrensis, L. sativae and L. trifolii are
The five Tropicomyia species are virtually indis- native to North and South America (but L. trifolii
tinguishable on external characters and are differen- has become semi-cosmopolitan during the last 12
tiated primarily by their male genitalia. Even here years by commercial introduction); L. schmidti is
differences are slight. Sasakawa (1977) synonym- restricted to South America and Florida, L. cordo-
ised both T atomella and T styricicola with bensis and L. criptica to Argentina and L. cocculi to
T theae. I accept the synonymy of the first two Hawaii.
(Oriental Region, southern Japan) but consider that L. trifolii is a remarkable species, with its original
T theae should be treated as distinct, representing hosts apparently in the Leguminosae in the eastern
the senior synonym of T coffeae (see Appendix). U.S.A. and quite recently it has extended its host
Polyphagous species 385
range dramatically. It is the only polyphagous spe- further paratypes reared 'from lettuce leaves'
cies to have colonized the Rosaceae (Agrimonia, (presumably the stalks). Frick also recorded ad-
Crataegus in Europe). In parallel with its polyphagy, ditional material from Cucurbita pepo ('squash
indicating its ability to adjust to the many chemicals leaf') and 'ex Ipomoea batatas' (sweet potato). I
present in its wide range of hosts, it has been able to am not convinced that the specimens from
develop resistance to the many insecticides used Cucurbita and Ipomoea are conspecific with
against it, primarily in the U.S.A., more rapidly than M. splendida.
any other species. Parrella, Keil and Morse (1984) This species has subsequently been found to
report that most classes of insecticides (chlorinated occur widely in the southern United States as an
hydrocarbons, organophosphates, carbamates, pyr- internal stem-borer in Asteraceae, primarily in
ethroids) were used intensively against L. trifolii in the tribe Heliantheae (see above). I know of no
the late 1970s and 'the average effective field-life of further records of the species feeding on Apium
an insecticide used against this species in Florida has and it is an open question whether the host
been less than three years.' In California permethrin Apium represents a case of pre-polyphagy or
and Penncap (methyl parathion) also had a short xenophagy,
useful life. 2. Melanagromyza azawii spencer (1973a) was de-
Buhr (1954) recorded a further eight families scribed from Baghdad, Iraq, feeding in stems of
from which L. bryoniae had been bred, mainly from Vicia jaba and also in Sesamum indicum (Peda-
the Botanical Gardens, Rostock and still more with liaceae) growing in nearby plots at the Agricul-
mines probably referable to the species. However, tural Experiment Station, at Abu Ghraib. This
Hering (1957b) does not include these families in probably represents a case of xenophagy. No
his 'Keys' and I have located no reared specimens in other records of this species are known, and I
Hering's collection in the BMNH. The close proxi- suggest this represents a chance transfer to an
mity of many species in Botanical Gardens provides alternate host which was growing in the imme-
an unnatural environment in which polyphagous diate vicinity of Vicia jaba.
species may flourish and I therefore do not include 3. Ophiomyia heringi Stary feeds regularly on
Buhr's records here. For the same reason I record Campanulaceae and Asteraceae (primarily Lac-
here only Hering's (1957b) records of L. xanthoce- tuceae). In view of the accepted close relations-
ra (= crucijericola), not those given by Buhr (1954). hip of these two families, I feel this can best be
Valladares (1986) in an unpublished Report re- treated as extended oligophagy, rather than pre-
cords two species in Argentina, both as nomina polyphagy.
nuda, which can be accepted as polyphagous. Lirio- 4. Amauromyza flavifrons (Meigen) with normal
myza cordobensis feeds on Thalictrum (Ranuncula- hosts in the Caryophyllaceae has regularly been
ceae), Lens and Pisum (Leguminosae) and Tropae- found also on Beta and Spinacia (Chenopodia-
olum (Tropaeolaceae), while the hosts of L. criptica ceae), both families in the same order Caryo-
are Luffa (Cucurbitaceae), Poissonia (Leguminosae) phyllales, their feeding can best be considered as
and Nicotiana (Solanaceae). The relationships of extended oligophagy.
these two species need further study. 5. Amauromyza verbasci (Bouche) with normal
The only polyphagous species in Chromatomyia, host Verbascum (Scrophulariaceae) is also not
C. horticola, has the widest host range of any spe- infrequently found on Buddleja. The taxonomic
cies (35 families in all subclasses except Hamameli- position of Buddleja has been unclear and it was
dae), most records known in Europe, though some included in the Loganiaceae (Asteridae: Gentia-
in India. nales). It is now accepted in a separate family
Table 92 clearly indicates the random nature of Buddlejaceae, close to the Scrophulariaceae.
the host selection by the polyphagous species and This accords well with the host specialization of
there seems no logical explanation for the varying Am. verbasci and again can be construed as ex-
choice of hosts by the different species. tended oligophagy, rather than pre-polyphagy.
It seems surprising that a more evident transition 6. Liriomyza brassicae (Riley) is a native Nearctic
in host selection between the wide host range of tru- species but is also semi-cosmopolitan, known in
ly polyphagous species and the narrow specializa- East Africa, the Oriental Region and Australia,
tion of the vast majority of species is not detectable. and in Europe occurring commonly in Botanical
The four pre-polyphagous species and cases of Gardens. Its primary hosts are in the Brassica-
xenophagy are briefly discussed below. ceae and I have found it frequently also on Cap-
paraceae (Cleo me, Gynandropsis) in Singapore,
1. Melanagromyza splendida Frick (1953) was de-
the Philippines, Sri Lanka and Papua New Gui-
scribed from Hawaii, with the holotype and pa-
nea. A further favourite host is Tropaeolum, on
ratypes reared from Apium graveolens and nine
which I found it at several localities in Australia,
386 Polyphagous species
Malaysia and California. Hering also cites a re- ternate host. I suspect that in due course this
cord on Resedaceae. In Kenya and in northern species will successfully colonize further hosts
India L. brassicae has also been found attacking and I therefore treat this species as pre-poly-
Pisum (Leguminosae). The surprising switch phagous.
here is to Pisum. I feel this species must be trea- 9. Napomyza latera lis (Fallen) feeds on hosts in the
ted as pre-polyphagous and the discovery of fur- Asteraceae, primarily in the tribe Anthemideae.
ther hosts can be expected. It is closely related to species on a number of
7. Liriomyza blechi Spencer feeds primarily on other families and the overall similarity both in
Blechum (Acanthaceae) in the southern United external characters and in the male genitalia sug-
States and I recently found it on Guadeloupe. gest recent speciation. The species present on
Isolated specimens which cannot be satisfactor- Unum (Linaceae) is accepted as representing
ily distinguished from L. blechi have also been N. lateralis. However, more detailed studies
reared in Florida from Heliotropium (Boragina- may show that this population also represents a
ceae), Lippia (Verbenaceae), Plantago (Planta- distinct species. At the present time, however, I
ginaceae) and Spigelia (Loganiaceae). I original- am inclined to treat N. lateralis as pre-poly-
ly inclined to the view that the species on these phagous.
hosts represented a complex of closely-related 10. Chromatomyia mimuli Spencer (1981) was
species. Now however I feel it is not impossible found to be common in California, with large
that all represent L. blechi. It should be noted populations feeding on both Mimulus (Scrophu-
that all these further hosts are in orders of the lariaceae) and Phacelia (Hydrophyllaceae).
Asteridae. These two genera are in different orders in the
L. blechi is a distinctive species, with unusual Asteridae and it is impossible to determine
larval spiracles (Fig. 799) and characteristic ma- which represents the primary host. More infor-
le genitalia (Fig. 798). There has been one cer- mation is needed on the distribution and host
tain switch to Plantago in South America where range of this species and at the moment I can
I discovered this to be the host of L. andina only treat the presence of C. mimuli on Mimu-
(Malloch) during a visit to Chile in 1978. Also, Ius and Phacelia as extended oligophagy, rather
even more striking has been a switch to Poaceae than pre-polyphagy.
(primarily Paspalum) where L. marginalis (Figs. 11. Chromatomyia syngenesiae (Hardy) is very com-
1409, 1411) is certainly closely related to mon in Europe and has been introduced to both
L. blechi. the East and West coasts of the U.S.A., and to
It might be justified to treat L. blechi as a Australia and New Zealand. Griffiths (1967a)
polyphagous species. With certain switches be- lists 18 genera of Compositae as hosts con-
tween Acanthaceae, Plantaginaceae and Poa- firmed from male genitalia and also two isolated
ceae which have resulted in the evolution of occurrences on Daucus (Apiaceae) in Denmark
three separate species, the hosts in four further and Pisum (Leguminosae) in California. Grif-
families indicate the host-switching potential of fiths treats the two records on non-composite
L. blechi. The status of the populations on hosts as xenophagy and this I accept. There are
Heliotropium, Lippia, Plantago (in Florida) and further early records of 'Phytomyza atricornis'
Spigelia are not yet fully understood and at this (before this species was split into C. horticola
stage L. blechi is best treated as pre-poly- and C. syngenesiae) in North America on Lami-
phagous. If hosts can be discovered outside the aceae, Malvaceae and Leguminosae which al-
Asteridae, the species could justifiably be trea- most certainly represent misidentifications. Ear-
ted as polyphagous. ly records from Australia and New Zealand are
8. Uriomyza eupatorii (Kaltenbach) feeds com- also not accepted. Thus Chr. syngenesiae which
monly in Europe on Eupatorium and some has been widely treated as polyphagous is in fact
other genera of Asteraceae but also regularly on restricted to the Asteraceae, apart from two
Galeopsis tetrahit (Lamiaceae). The spiral begin- confirmed cases of xenophagy.
ning of the leaf mine (Fig. 1101) is very distinc- 12. Chromatomyia pseudomilii Griffiths. Recorded
tive. The male genitalia (Figs. 1099, 1100) show in Alaska on Milium and Carex, and in Europe
that L. eupatorii is closely related to L. pusilla on Deschampsia. These records suggest that this
which is widespread in Europe and is now ac- species is basically a Poaceae feeder and Carex
cepted as having an extensive distribution across as an alternate host probably represents xeno-
Asia, with hosts exclusively in the Asteraceae. It phagy.
can thus be accepted that this family provides
In his detailed discussion of the mechanism of host
the primary hosts of L. eupatorii. Its occurrence
on Galeopsis is frequent and this is a regular al- selection in phytophagous insects, Dethier (1954)
Polyphagous species 387
comments on the advantages of polyphagy, primar- er are in the process of developing it. I cannot here
ily 'the lack of complete dependence upon the fluc- discuss the possible reasons for this but restrict my-
tuating fortunes of a single plant species.' In view of self to the outline above of the actual situation exist-
this it may seem surprising that so few Agromyzidae ing in the Agromyzidae today.
have developed secondary polyphagy and even few-
6. Concluding remarks: phylogeny, coevolution, colonization
The phylogeny of the Agromyzidae must be inex- the host specialization of the Agromyzidae and the
tricably linked to the processes of host specializa- evolutionary development of their hosts. It is now
tion which have evolved in the family probably since clear, having studied all species whose hosts are
the end of the Cretaceous and at least from early known, that this is an oversimplification of the pro-
Tertiary. blem. Relationships are detectable between related
The Agromyzidae are an isolated family and agromyzids and related hosts as discussed in many
there has been no general agreement on their closest cases in the main text above. However, the overall
relatives. Hendel (1920) united them with the Odi- picture becomes very apparent when considering
niidae in two subfamilies, considered to be related the host associations of each agromyzid genus as
to the Milichiidae. Later (1931) he related them to discussed below (see Table 93). There is really little
the Tephritidae and Ortalidae with a group name logic in the hosts which have been colonized by the
Ortalidiformes proposed by Frey. Hennig (1958) majority of genera. There have been countless
discusses the relationships of the Agromyzidae in switches between unrelated families, which may
some detail but finally has to conclude that he can- then result in a varying degree of radiation, such as
not decide on their sister-group. Later (1965) he of eight Phytomyza species on the Aquifoliaceae, 64
postulates a sister-group relationship with the Odi- Phytomyza species on the Apiaceae. In most cases
niidae. Griffiths (1972c), in his detailed discussion the colonization of an unrelated host leads to the
of the relationships of the Cyclorrhapha, proposes evolution of a single isolated species on this host,
the Clusiidae as the nearest relatives of the Agromy- with the nearest relative now totally obscured by
zidae, uniting the two families in the Prefamily time. Thus the only explanation for the diverse hosts
Agromyzoinea. both at the specific and generic level is the remark-
The lack of· agreement among specialists on the able degree of randomness which has now been re-
Acalyptrates and the obscure relationships of the vealed.
Agromyzidae can only be the result of extinctions The extensive literature on insect/plant rela-
which have left wide gaps between modem families tionships during the past 25 years has advanced
and between genera of the Agromyzidae themselves. conflicting theories. In my review of the host special-
Although the Agromyzidae are a relatively large ization of the Agromyzidae, it has been possible to
family, they are remarkably homogenous, with only consider only a few of the more important papers
30 currently accepted modem genera (contrast the with a direct bearing on the Agromyzidae. The spe-
Chloropidae, a family of comparable size with 150 cies I have considered with known hosts number
genera). Nevertheless the internal classification of 1190 (47% of the world total of 2500) but it is not
the Agromyzidae has remained unclear. The present unreasonable to include the 155 species accepted as
study, with emphasis on host plants, but with full feeding on Poaceae from their male genitalia, al-
consideration of patterns of distribution, would I though they have not actually been reared (see order
had hoped clarify the phylogeny of the family. This Poales above). This increases the total to 1345, rep-
is only partially so. resenting 54%. This substantial figure forms my
Hering (1951a) envisaged the evolution of the basis for testing the views expressed by other work-
Agromyzidae as occurring in parallel with that of ers in this field.
their hosts. He never actually used the term coevolu- The main debate has centred round coevolution
tion as far as I know but discussed at considerable versus colonization. Coevolution has been referred
length the evolution of the agromyzids following the to by different authors in differing terms and with
system of Mez (1925) based on serum diagnosis. differing meanings. Ehrlich and Raven (1964) in
Working closely with Hering from 1950 until his their study primarily of Papilionidae implied by co-
death in 1968 I was inevitably considerably influ- evolution reciprocal evolutionary changes between
enced by his ideas and when starting this study I ex- parasite and host. Jermy (1984) referred to this con-
pected that I might discover a neat parallel between cept as 'classic coevolutionary theory'. Mitter and
Phylogeny, coevolution, colonization 389
'lapU! snuaD - - - - ~
- -
- -
-eZAWOlj3Old N
o:!
~o. llZAWod-euopnasd <'l ~
-- -- -
e::l
02 <'l 0 ..., 00
llZAwolAljd ' <'l
'- -
<'l
~oo
0 Vl r--
--
Z -e!AWOlllWOllj:J
~ ~ 00 M
~
llZAwodllN .
<'l M -~ 00
sapoadoldllH
-eZAWOlAljdouWAD
-:• •
-~
..., -- ,
~
-
o:!
-ezAwolAljdllllld
-
<'l
N
<'l \D
\D
--
0
«'J
- -
~
N
>. -eZAW03Alll:J
~
- -
~ Eo.
-
w ·cO::l0 llZAWO!UaX
Z
w .- ...
0...... 00
llZAWO!lllD
...,
N
0
- -
~
-
llljlUopola:J
-- -
>.
..c::
-
0 \D Vl r--
0... llZAWO!ln ~ <'l \D N
~
N <'l
>- -
-
-ezAwop!I0pnasd
::E N M
0 llZAwodolaw
«'J
a- ..., ...,
'"
~
0 llZAWOP!l°lAljd
M r--
-
\D
:0
o:!
~
u.. - -- a-
- -- - - -'" - -
Vl
--
V) «'J ~
E- llZAWOll!Y •
O
rn -eZAWOll!llU-edllf a-
W
o:! ..., N ~ <'l \D
'"
::l 1l!AW03!dol.L
~
-
N Vl
>.
..., -
~ § §"
- -
<'l M
1l!AW!W!wlj3SU!al)l
~ ... 0 Vl
u..
-
M ~ <'l
-
~~ ll!AWO!ljdO <'l
E- .....
'"
- ..., - -
rn t:: SPPOll!AWOP!3a:J
0
X '"
0... llZAWOlll-eUllPW
,- N r-- \D
<'l
N <'l N \D
llzAwoxaH
llZAwOJll-elaUad -
- -
- --
llzAwpowaN
.;':! 0.
--
~ N \D ~
-
.D::l llZAWOJnllWY
........
00
:\0 :\0
-
N \D 00 0
..2 00 ll!qOlAljd <'l
'"
0... -
- --
ll!qolAljdolQ.ld
«'J
ll!qolAljdoa-ellld
:
rn
rn
~ Z
I
0
...'"o:!
W
0... 0
rn '" o:! W
...0'"o:! ...'"0o:! ....'"o:! c0 E'"
til
'" :3'" =:
0 -0 ..J
~ o:!
G .....>. .....'" ~ o:!
~ >- ro ,C C 0
.,
,-0 '0
'" '"
o:! o:!
E
'" '"
o:! E-
Q)
E o '2 o:! .;:: 0 0 ...
Z ...... :0
o:! ..2-0 .... o:! ...... C Q) o:!
~ ..c:: :D
0. .-
~
0 o:! "0
'" 0.0>. ~ o:!
u C E
... '" 'm0 2til Z0 .- c.;:: :.~ E....0 'bbt:: E0
-0
-
~
0. 0
I 0 .~
~
t::
E >. E
... W0"0
>.
Z III
::l
0...
t::
i:i:
bl)
o:! 'o:! o:! :;::
::Ex U 0 ~ ~ 0 ;J ~ III N U :< ~
0
Z ::E ~
Brooks (1983) used the term 'association by des- ed examples of parallel cladogenesis between insects
cent' for coevolution but without postulating evolu- and plant are known.' In the Agromyzidae I con-
tionary pressure on the host by the action of the par- sider that their feeding in no way influences the evo-
asite. The alternative hypothesis they called simply lution of the hosts. I have cited a few examples
'colonization'. Miller (1987) posed the question: where coevolution, implying simply speciation of
'parallel cladogenesis or colonization?' He reviewed the agromyzids in parallel with the diversification of
the available literature on insect/plant relationships the hosts, appears possible and it is an attractive
and was able to write: 'I conclude that no document- theory. For instance, in the United States the two
390 Concluding remarks
stem-boring species Melanagromyza angelicae (Fig. to subfamily status (within the Geomyzidae) (Tull-
582) and M. lomatii (Fig. 591), feeding on Angelica gren and Wahlgren, 1920-1922: 546) as Selacho-
and Lomatium, are closely related as shown by their pinae, and as Eucelocerinae (corruption of Encelo-
male genitalia. Angelica and Lomatium are also cerinae) by Hackman and Vaisanen (1985: 200) but
considered to be closely related. The 13 Liriomyza these taxa have not received further recognition.
leaf miners on genera in the Tribe Lactuceae (Aster- I now consider that four main phyletic lines can
aceae) are a monophyletic group and it is not incon- be recognised as indicated in Table 93 but these
ceivable that they have speciated on isolated popu- groups cannot be wholly linked to any association of
lations of the diverging host genera. The two Chro- related hosts. The host families colonized by each
matomyia leaf miners on Primulaceae, C. primulae 'line' represent an apparently random spread, al-
on Primula and C. soldanellae on Soldanella might though clear relationships are detectable in many in-
well have speciated together with their hosts. It must dividual cases. There has been much radiation on in-
be emphasised though that in these three cases dividual families, with generic colonization obvious-
where coevolution seems plausible, there is no firm ly facilitated by their close relationship. I briefly
evidence that this actually occurred. Botanists can comment below on some of the genera.
give no dates when generic speciation has taken
place and equally dates of speciation of individual
Phytobia group
species in the Agromyzidae remain pure conjecture.
No cladograms exist for the majority of plant fami- Phytobia. Nowakowski (1964: 212) expressed the
lies and none for the Agromyzidae, and any possible firm conviction that this genus is the most primitive
congruence in the relative cladograms is thus impos- in the family and this view has not been disputed by
sible to confirm. If such cladistic investigations can later workers. The biology - internal boring in the
be carried out in the future, the results could be of cambium of trees or bushes, the large size of many
great interest. In the meantime it has to be accepted species, with wing length of 6.5 mm in Pb. gigas
that the high probability is that the speciation in the Spencer (1966a) in Burma, the worldwide distribu-
Agromyzidae has been the result of sequential colo- tion of the genus, with many undescribed species
nization. Clearly this will have been greatly facilitat- discovered in rain forest in Peru by von Tschirnhaus
ed when a switch has been between related hosts. and six species known on Lord Howe Island (Spen-
Many obviously random switches will have also oc- cer, 1977 c), all indicate an ancient origin. This is
curred but this will not have been easy. The agro- supported by the discovery of feeding tracks in fossil
myzid female takes great care when ovipositing that rosaceous wood of Eocene age in North America.
the host she is selecting is the normal, correct one or Modern species are only known on the Betula-
at least compatible. The mechanism adopted for en- ceae and Fagaceae (Hamamelidae), Salicaceae (Dil-
suring the suitability of the host is threefold: (1) leniidae), Aceraceae (Rosidae), Oleaceae, Rosaceae
smell; (2) taste by the formation of numerous feed- and Solanaceae (Asteridae). However, as will be
ing punctures before an egg is laid; and (3) the fail- seen from Table 93, 53 further families are now re-
ure of the ovaries to produce viable eggs unless the corded as hosts based on larval feeding detected by
appropriate chemicals from an acceptable host have Suss and Muller-Stoll (1980).
first been imbibed. The formation of feeding punc- Amauromyza. Many species are internal stem
tures appears to be universal throughout the family borers in herbaceous plants and this represents a
but I know of only one series of experiments testing logical progression from boring in young saplings.
the viability of ova which were carried out on Napo- The leaf mining habit of some species represents a
myza carotae by Wiesman (1961) in Switzerland. more advanced evolutionary development from
The high degree of host specialization is clear from stem boring.
the data I have provided and has been fully con- Nemorimyza. N. posticata, a leaf miner on
firmed experimentally on Liriomyza sonchi on Son- Asteraceae, is an isolated species in this monotypic
chus (see above and Peschken and Derby, 1988) genus, with characters associating it with Phytobia
and on Phytomyza pullula (as matricariae) (see but more closely with Amauromyza.
above and Sehgal, 1971a).
Traditionally, the Agromyzidae have been divid-
Penetagromyza group
ed into two subfamilies, the Agromyzinae and Phy-
tomyzinae. Von Tschirnhaus (1971) examined the Penetagromyza. The monotypic P. aloes, known on-
relationship between these groups and concluded ly in South Africa feeding internally in the thick
that there is no sufficiently clearcut division between leaves of Aloe petricola, I now consider to be ances-
them to justify the division of the existing 30 mod- tral to both Melanagromyza and Ophiomyia (see
ern genera into two subfamilies. Selachops flavocinc- above under Agavaceae).
ta Wahlberg (cf. Hendel, 1936) has also been raised Hexomyza. This small genus, with the larvae
Phylogeny, coevolution, colonization 391
forming twig galls, is close to Melanagromyza and is genus is best represented in the northern hemi-
known only on six families in the Dilleniidae, Rosi- sphere, with no species known in Argentina, Chile
dae and Asteridae. or New Zealand, and only a few on grasses in Aus-
Melanagromyza. This is a predominantly tropical tralia.
genus, becoming progressively more numerous from
the Magnoliidae to the Asteridae and has also colo-
Phytoliriomyza group
nized the Orchidaceae, with six species known from
Brazil to New Zealand. The genus declines north- Considerable diversification has occurred in this
wards in the Palaearctic and Nearctic Regions. The group, with the evolution of 11 genera from the
primary feeding habit is internal stem boring, a few primitive Phytoliriomyza to the most advanced
species also feeding in flower-heads or pods of Haplopeodes.
Leguminosae. Phytoliriomyza. It was a surprise to discover
Ophiomyia. A widespread genus with many spe- during the course of this study that this is the genus
cies, it is more advanced than Melanagromyza, the present both on Hepaticae (not Liriomyza as had
larvae mainly forming external stem mines but some been believed) and ferns. I consider that the associa-
have developed into leaf miners. Relatively few tion with these two primitive groups must be an-
hosts are known, owing to the difficulty of detecting cient, reflected in the wide distribution of species
the inconspicuous stem mines. Species are known from Europe and North America through South
on two families of monocots, the Asparagaceae and America to the Juan Fernandez Islands and to Aus-
Hemerocallidaceae. tralia and New Zealand. The presence of related
Kleinschmidtimyia. This small Australian leaf species on the tree fern genus Cyathea in both New
mining genus is known only on the Leguminosae Zealand and South America (also Guadeloupe)
and Thymeleaceae (Rosidae), and also on the Phor- even suggests possible dispersal following the break-
miaceae in the monocots. The minute black adults up of Gondwanaland in association with Cyathea.
appear like little Melanagromyzas but the male geni- Relatively few hosts are known, with a wide jump
talia differ significantly. from the ferns to three families in the Rosidae (Pit-
I believe that comparable leaf mining genera have tosperaceae, Leguminosae, Balsaminaceae) and
evolved independently from Melanagromyza both in three in the Asteridae (Solanaceae, Bignoniaceae
Africa and South America but the few species con- and Asteraceae). The internal feeding habit of the
cerned are inadequately known and are here includ- larvae in many species makes discovery of the hosts
ed in Tropicomyia (Africa) or in Ophiomyia (South difficult. For instance, 25 species are known in Aus-
America). I am convinced that new genera will be tralia but the host is known of only two - both on
erected for these species in due course. Pittosporaceae. Phlir. papae, feeding in the stalks of
Tropicomyia. This is a purely tropical genus feed- potato in Bolivia, is certainly a primitive species.
ing on a wide range of hosts in all subclasses of the Metopomyza. M. ornata is an isolated relict on
dicots, on two superorders of the mono cots and also Butomus (monocots), which with other species in
on ferns and Coniferae (Gnetum). The larval leaf the genus feeding on Juncaceae, Cyperaceae and
mines are distinctive, being shallow and purely epi- probably Poaceae, represents an early switch to
dermal, the puparium remaining in the leaf. The monocots in the northern hemisphere of an ances-
genus may have evolved early, independently of tral species associated with Phytoliriomyza.
Melanagromyza, from an ancestor near Penetagro- Selachops flavocincta Wahlberg (d. Spencer,
myza. It is known only in Africa, Asia, southern 1976a) is an obviously primitive species whose host
Japan, northern Australia and the Pacific. More is unknown. I agree with Zlobin (1984b) who sug-
families have been colonized (55) than in any other gested a relationship with Metopomyza. This is par-
genus in this group. ticularly apparent in the genitalia of S. flavocincta
Japanagromyza. This small, tropical, leaf mining (Figs. 1282, 1283) and M. ornata (Figs. 1278,
genus has limited hosts and is best represented on 1279). I now predict on this basis that S. flavocincta
the Leguminosae. It appears to be intermediate be- feeds on a mono cot and probably in the root, in
tween Melanagromyza and Agromyza. view of its large size. The head shape (Fig. 1281) is
Agromyza. This genus is represented on a wide unique within the Agromyzidae, with the enormous-
range of hosts, with some radiation on the Hama- ly wide and projecting orbits.
melidae, Rosidae and Asteridae, though only three Pseudoliriomyza. Only P. cordiae is known in this
species are known on the Asteraceae. In the mono- monotypic genus, a leaf miner on the tree, Cordia
cots it has a large group of grass-feeding species, in ovalis (Malvaceae) in southern Africa and Java. I
which the male genitalia are distinctive and suffi- consider this represents an early offshoot from Phy-
ciently differentiated for any certain affiliation with toliriomyza.
a sister group on the dicots to be obscured. The Liriomyza. This large, worldwide genus has the
392 Concluding remarks
widest host range within the family, known on 76 Galium, is certainly closely related to the Galium
families which include the Equisetaceae, all sub- feeders in Paraphytomyza.
classes of the dicots and 11 families of monocots. Haplopeodes. This small New World genus is
Interpretation of the three species on Equisetaceae known on only four families - Chenopodiaceae,
is difficult. Colonization of this family obviously oc- Amaranthaceae and Portulacaceae (Caryophylli-
curred early, with both L. equiseti and L. virgo dae) with five species, and Solanaceae (Asteridae)
known in Europe and North America, having dis- with eight species. It appears to be derived from
persed across Beringia probably early in the Ter- Liriomyza.
tiary. Other primitive species are known, such as
L. antiquaria Spencer (1977b) in Mexico but not
Napomyza group
their hosts. It is thus possible but by no means cer-
tain the Equisetaceae feeders represent primary Napomyza is clearly the most primitive genus in this
colonization of this primitive family. group, directly ancestral to Phytomyza, with Chro-
Few Liriomyza species are known on early matomyia, Pseudonapomyza and Ptochomyza
angiosperms - only five on Ranunculaceae but 75 evolving in association with their specific host fami-
on Asteraceae. Several separate switches to the lies.
mono cots have taken place. There has been some Napomyza. I consider that the 14 species known
radiation of closely related species on Poaceae and on Ranunculaceae (most being internal feeders in
three are known on Allium (Alliaceae). Four further the stem or flower-head), including the hoi arctic
isolated species are known on the monocots: species, N. evanescens, are descendants of the origi-
L. asphodeli (Asphodelaceae), L. smilacinae and nal ancestor of the genus. The genus has dispersed
L. wachtli, a species-pair on Liliaceae, and an inter- to southern Africa, Australia and New Zealand
nal, bud-feeding species on Lilium, L. urophorina. where the hosts are exclusively in the Ranuncula-
Also two species have colonized Triglochin (Junca- ceae. N. gymnostoma feeding on Allium porrum is
ginaceae). L. commelinae (Commelinaceae) is a the most primitive extant species which must have
particularly interesting, primitive species, represent- colonized this monocot host early. It remains as a
ing a link with Cerodontha. This genus is restricted relict, with no relatives on the monocots.
to the monocot families Iridaceae, Juncaceae, Cype- Napomyza has later colonized further dicot fami-
raceae and Poaceae. It is of worldwide distribution lies, and new species have evolved on the Urticaceae
and the switch to mono cots must have occurred ear- (Hamamelidae), Apiaceae (Rosidae), Lamiaceae
ly. It is generally agreed that Cerodontha is closely and Campanulaceae (Asterideae), with considerable
related to Liriomyza and L. commelinae (see Com- radiation on the Asteraceae (nine species).
melinaceae) appears to be an annectant species. Phytomyza. Hendel (1931) treated Napomyza
The remaining six genera in this group (Table 93) and Phytomyza as subgenera in the single genus
represent a varied assemblage and their exact rela- Phytomyza, but the two groups are now fully accept-
tionships remain unclear. Galiomyza is a small ed as distinct genera. The extraordinary prolifera-
genus of five species restricted to the Violaceae and tion of Phytomyza on Ranunculaceae with 91 spe-
Rubiaceae, related to Liriomyza. Xeniomyza has cies suggests that the genus may have evolved on
only two species, the hosts being known only of this family, in direct descent from Napomyza, with a
X. ilicitensis in Spain - Salicornia and Suaeda (Che- biological change to leaf-mining from the more
nopodiaceae). Xeniomyza appears to represent an primitive internal feeding of Napomyza. A number
isolated offshoot from Liriomyza. Calycomyza is of leaf mining species on Clematis and Ranunculus
largely restricted to North and South America and in Australia and New Zealand pupate in the leaf,
has colonized a variety of host families from the Ur- which can be construed as a relictual biological
ticaceae (Hamamelidae), Malvaceae (Dilleniidae), character of the ancestral stem-boring habit in
Leguminosae, Polygalaceae, Sapindaceae and Apia- which the larvae all pupate internally.
ceae (Rosidae) to the Convolvulaceae, Boragina- No Phytomyza species are known on non-angio-
ceae, Lamiaceae, Valerianaceae and Asteraceae sperms and few are known on other families of Mag-
(Asteridae). The main radiation has been on the As- noliidae, Hamamelidae and Dilleniidae (none on the
teraceae. Caryophyllidae), with only seven species on five
Paraphytomyza. The five species known on Sali- families in these subclasses. The genus becomes
caceae (Hamamelidae) in Europe form a compact more numerous on the Rosidae, with 22 species on
group which are probably not congeneric with the six families but then follows a striking radiation on
main monophyletic group of 17 species on Rubia- Apiaceae producing 64 species, with a high degree
ceae, Caprifoliaceae and Dipsacaceae (Asteridae). of monophagy on individual genera. However, the
Gymnophytomyza. This monotypic genus with the full development of Phytomyza occurs on the Aster-
single species G. heteroneura, feeding in seeds of idae with 13 families colonized with some 40 spe-
Phylogeny, coevolution, colonization 393
cies (clear radiation on the Lamiaceae (13) and Typ.e B includes. oligophagous, closely related spe-
~crop~ulariaceae (17)) but another remarkable pro- CIes on a restncted group of plant species. There
hfe:atIOn on the Asteraceae, with 87 species, of are also many examples of this type in the Agro-
W~IC~ 84 are tribe-specific (see Table 52). Although myzidae.
thIS IS the largest world genus, the number of fami- Type <:= embraces monophagous, closely related
lies colonized (28) is less than half the number in speCIes on closely related plant species.
Liriomyza (76) which has far fewer species. Type D. Polyphagous species.
Chromatomyia. This is either descended directly
All four of Jermy's feeding types are evident in
from Napomyza or represents an early offshoot
the Agromyzidae. Striking examples of Type A are
from Phytomyza. Four species feed on ferns and it
the presence of Napomyza species on Ranuncula-
seems possible that the genus evolved in association
ceae and the primitive N. gymnostoma on Alliaceae,
wi~h .t~eir. different hosts, the species retaining the
and Ptochomyza mayeri on Clematis and P. aspara-
p.nmItIve Internal pupation of Napomyza. Two spe-
gi and P. asparagivora on Asparagus. On a wide ge-
CIes are known on Ranunculaceae and there is then
neric level, it will be seen from Table 93 that there
a j~mp to the Primulaceae (Dilleniidae) and the
are countless further examples of Type A. Examples
SaxIfragaceae and Eleagnaceae (Rosidae). The main
of Types Band C need not be cited here but are
development of the genus occurs on the Asteridae
with. significant radiation on the Gentianaceae (10 enumerated throughout the main text above. Just 16
polyphagous species fall into Type D (see Section 5
specIes), the Caprifoliaceae (12) and the Asteraceae
above).
(13). The lar.ge number of families colonized (47)
In his discussion of coevolution Jermy (1984)
reflects the WIde polyphagy of Chr. horticola.
concludes that there is no firm evidence to support
Chromatomyia has successfully switched to the
wha~ he terms the 'classic coevolutionary theory' of
mono cots, with a single species, C. pseudomilii on
Ehrh~h and Raven (1964), in which reciprocal
~yperaceae (though the primary hosts are probably
selective pressure between insects and their host
In the Poaceae), the Juncaceae (six species) and the
plants is postulated. No evidence for this is apparent
Poaceae (23).
in the Agromyzidae. Nor are any examples known
Pseudonapomyza. This genus can only represent
of ?irect interspecific competition exerting any evo-
an offshoot from Phytomyza. It consists of two well-
lutI?nary pressure on species within the family. Oc-
d~fined ?roups, the first in the Old World tropics
caSIOnally two species are found mining the same
WIth a SIngle species on Amaranthaceae, eight on
plant or even leaf but this is so rare that it can be of
Acanthaceae (Asteridae) and one on Vernonia (As-
no evolutionary significance.
teraceae). This is closely associated with a far larger
Jermy further suggests that 'recent insect/host
group which has switched to the monocots, known
plant systems have emerged and are evolving in a
on the Poaceae in temperate areas of the northern
'sequential' way: the insects following the evolution
and southern hemispheres and extensively also in
of plants.' I consider that this is exactly what has
the Old World tropics, with a single species on Gua-
happened in the Agromyzidae, certainly in the great
deloupe. Hosts are known of ten species but 26
majority of species. His suggestion that 'there is no
more which are clearly referable to the genus both
correlation between the evolutionary age of plant
on external characters and male genitalia are only
groups and that of insect species living on them' is
kn~wn f:om caught specimens. Four new species
not wholly supported by the host specialization of
WhICh wIll be described shortly have just been dis-
the Agromyzidae. I consider that the worldwide
~overed in Czechoslovakia (M. Cerny). Ptochomyza
presence of Phytoliriomyza on both ferns and liver-
IS a very small genus, with a single species known on
worts (Hepaticae), of Metopomyza ornata on Buto-
Clematis (Ranunculaceae) and two on Asparagus
mus and of Napomyza on Ranunculaceae and Al-
(monocots: Asparagales). This represents a remark-
lium (Liliaceae) does indicate a primary association
able switch (see under Asparagaceae).
between early agromyzids and primitive hosts. The
sequential colonization of ever more new hosts
J~rmy. (1984) considered that insect/plant rela-
r~sulting in the present diversity of agromyzid spe-
tIOnshIps can be classified into four types:
CIes has been substantially facilitated by the chemi-
Type A includes related insect species on distantly cal similarity of new, related hosts but random colo-
related plant species and it was suggested that this nization of unrelated hosts also forms a significant
i~ common in the Agromyzidae. Many examples part of the complex pattern of host specialization
CIted above fall into this category. apparent in the Agromyzidae today.
Taxonomic Appendix
Inevitably during the course of this study which in- 3 paratypes in National Museums of Kenya, Nairo-
volved the revisionary examination of some hun- bi, further paratypes in author's collection.
dreds of species, new species, new synonyms and Remarks. This is the largest Melanagromyza known
new combinations have been discovered. In addi- in Africa, with M. seneciocaulis Spencer (1960a),
tion, lectotypes have been designated, an incorrect a stem-feeder in S. ruderalis in South Africa
original spelling has been emended and the status of being marginally smaller. M. compositana Spencer
species in three genera has been discussed. It seems (1959), which is widespread in the Nairobi area is
appropriate to include all this taxonomic informa- generally similar but also smaller, and more green-
tion here, rather than scattered through the main ish. M. heatoni doubtless also feeds on other high
text. altitude Senecio species among the 85 known in
East Africa.
I have pleasure in dedicating this fine species to
1. New species my friend Tom Heaton, who organised and provid-
ed transport for the safari to the Aberdares.
1a. Two new Melanagromyza species
On re-checking European specimens identified as
During a visit to Kenya in February, 1988, I ob- M. dettmeri Hering (Spencer, 1966b) reared from a
tained a series of a large Melanagromyza on Senecio number of different hosts, particularly those from
moorei in the Aberdares National Park, which Crepis and Hieracium (tribe Lactuceae), I dis-
proved to be undescribed. covered that these and others represent an unde-
scribed species, close to but clearly distinct from
Melanagromyza heatoni sp.n. M. dettmeri.
(Figs. 1119,1120)
Melanagromyza oligophaga sp.n.
Exceptionally large, coppery-greenish species.
(Figs. 967, 968)
Head. Frons not or at most narrowly projecting
above eye; orbits pronounced, with 2 upper orbital Adult virtually indistinguishable from M. dettmeri,
bristles (ors), normally 2 lower (ori), orbital setulae with following essential characters: frons not pro-
in 2 rows, those nearest eye margin reclinate, inner jecting above eye, orbital setulae in 2 rows, those
proclinate; frontal triangle narrow, only moderately nearest eye margin reclinate, inner row proclinate;
shining; eye in male with patch of strong hairs at eye with distinct pilosity in both male and female;
level of ors, in female eye distinctly but more sparse- mesonotum normally faintly greenish; squamae
ly pilose; arista long, bare. pale, margin pale brown, fringe whitish to ochrous;
Wing. Length in male 3.1, in female 3.6 mm, inner wing length normally about 3 mm. Male genitalia as
cross-vein near centre of discal cell. in Figs. 967, 968.
Colour. Head black, mesonotum and abdomen pre- Hosts. Crepis biennis, Hieracium umbellatum, Picris
dominantly coppery but sometimes faintly green- hieracioides; Achillea millefolium, Artemisia vul-
ish; squamae whitish, margin pale brown, fringe garis. Larval posterior spiracles on 2 adjoining
silvery. plates, with a strong central horn surrounded by an
Male genitalia. Aedeagus as in Figs. 1119, 1120. ellipse of 15 pale pores.
Host. Senecio moorei, larva boring in stem (empty Holotype 0, German Democratic Republic,
puparium found).
Holotype 0, Kenya, Aberdares N.P., near Fishing
Lodge, W. of Nyeri, roadside from Kinaini Bridge to
1932; ex Crepis biennis; paratypes: 11 °
Miihlhausen, Thuringia, emerged (forced) Feb.
and Q,
same locality; 1 Q, near Meissen, spring, 1966, ex
Elephant Bridge, 2100-2500 m, 11.ii.88; paratypes: Hieracium umbellatum (all leg. H. Buhr); Thuringia,
9 Q, same data, all on Senecio moorei. Holotype and Miihlhausen,l Q, 1929, ex Picris hieracioides; Eng-
New species 395
land, Surrey, 1 d, 1 Q, ex same host, 1950; 1 d, ex Holotype d, Kenya, Nairobi, at Westlands Shop-
Achillea millefolium, 23 v. 50; 1 d, ex Artemisia vul- ping Centre, 15.i.87; paratypes: 4 d, 1 Q, same data,
garis, 1950 (all M. Niblett); Betchworth, Surrey, 1 all reared ex Crotalaria agatiflora (J.M. Ritchie).
d, 4 Q, ex same host, emerged spring 1958 from Holotype and 2 paratypes in National Museums of
puparia collected Sept., 1957 (K.A. Spencer). Holo- Kenya, 3 paratypes in author's collection.
type and paratypes in B.M. (Nat. Hist.), further para- Remarks. Ten species are known in Africa with the
types in author's collection. squamae and fringe pale but none have known
Remarks. M. oligophaga runs to M. dettmeri in the hosts. 0. crotalariella is readily distinguishable from
author's (1966b) key to European Melanagromyza others in the group by the lack of a vibrissal fasci-
species and most of the type specimens were dis- culus and the very narrow facial keel. This is a par-
cussed as M. dettmeri in that paper. At that time the ticularly interesting species, being the only one
significance of minor differences in the genitalia known to feed in twigs of a tree.
were not fully appreciated but the variation was
commented on. The differences in the genitalia be-
1c. Two new Tropicomyia species from Kenya
tween M. dettmeri (Fig. 938) and M. oligophaga
(Fig. 967) are particularly striking in side view. Tropicomyia eulophiae sp.n.
The host of M. dettmeri must now be restricted to (Figs. 1276, 1277)
Centaurea jacea and the species is only known in
Adult. Generally resembling T. flacourtiae (type of
Denmark. With hosts of M. oligophaga known in the
the genus), with following essential characters: en-
tribes Lactuceae and Anthemideae, it is possible
tirely black, small but somewhat larger than T. fla-
and even probable that the species has an even
courtiae, wing length from 1.9-2.25 mm; costa
wider host range and it is hoped that this will be es-
ending at vein R4 + 5, inner cross-vein well beyond
tablished in due course by further collecting.
centre of discal cell; mesonotum deep black, moder-
A female from East Germany, Jena, Leutra Tal,
ately shining from rear.
spring, 1966 ex Inula sp. (leg. Buhr) possibly repre-
Male genitalia. Aedeagus as in Figs. 1276, 1277,
sents M. oligophaga but this specimen is not treated
with distinct differences from T. flacourtiae.
as a paratype.
Host/Biology. Eulophia porphyroglossa (Orchida-
ceae), larva forming long, slightly irregular linear
1 b. A new Ophiomyia species from Kenya mine, up to 12 cm in length, distinctly greenish, with
finely scattered black frass; the mine appears to be
Dr. J. Mark Ritchie, Head of the Department of En-
deeper than the epidermal mines normal in this
tomology at the National Museums of Kenya, Nai-
genus; pupation in leaf, with stalked anterior spir-
robi noticed that terminal twigs of the small tree
acles projecting through the epidermis, posterior
Crotalaria agatiflora were withering and dying and
spiracles on a short, stout central projection, with 3
discovered that this was due to mining by agromyzid
pores at each side.
larvae. Six specimens (unfortunately now in poor
Holotype d, Kenya, Western Highlands, Kapsa-
condition) were reared and I have identified the spe-
bet, 19 .x. 71 ex leaf mine on Eulophia porphyro-
cies as an undescribed Ophiomyia belonging to the
glossa coIl. 18.x.71; paratypes: 4 Q, 3 same data,
small group with the squamae and fringe white.
1 emerged 2.xi.71 (all K.A.S.), in BMNH.
Remarks. I originally misidentified this species as
Ophiomyia crotalariella sp.n.
T. flacourtiae despite differences in the mine and
(Figs. 502, 503)
small differences in the male genitalia. It is now clear
Head. Frons not projecting above eye; ocellar that T. eulophiae is distinct, as might be expected on
triangle broad above, half length of frons; jowls nar- its unusual host.
row, vibrissa strong but fasciculus lacking, antennae Records of mines on Lissochilus horsfalli on Mt.
divided by low, narrow keel; third antennal segment Cameroun, West Africa can be accepted as repre-
small, round. senting this species (Lissochilus now accepted as
Mesonotum. 2 dorso-central bristles. synonym of Eulophia).
Wing. Length from 1.6-1.75 mm, costa extending
weakly to vein M1 + 2, last and penultimate sections Tropicomyia gloriosae sp.n.
of vein M3 + 4 equal. (Figs. 1251, 1252)
Colour. Entirely black, apart from silvery-white
Adult. Essentially as in T. eulophiae but smaller,
squamae and fringe.
wing length 1.75 mm. Male genitalia. Aedeagus as in
Male genitalia. Aedeagus as in Figs. 502, 503.
Figs. 1251, 1252, unusually small (same scale as
Host. Crotalaria agatiflora, a single larva forming an
T. eulophiae, Figs. 1276, 1277).
external mine in a terminal twig, pupating beneath
the epidermis, puparium black.
396 Taxonomic Appendix
Host/Biology. Gloriosa superba (Liliaceae), details Male genitalia. Aedeagus as in Figs. 683, 684.
of leaf mine not recorded. Host. Gentiana excisa, larva initially forming linear
Holotype d, Kenya, Nairobi, Arboretum, ex mine which then develops into a large blotch. Pupa-
Gloriosa sp., 28.iii.84 (B. Tengecho), in National tion in mine.
Museums of Kenya, Nairobi but temporarily re- Holotype d, Switzerland, 'bei Lenzerheide',
tained in BMNH. 22.vii.27 (No. 3226a), ex Gentiana excisa (leg. W.
Remarks. A mine with larva of this species was also Hopp) in BMNH.
found on Gloriosa superba in the Shimba Hills Re- Remarks. The background of this species is dis-
serve at the coast, 2 km from Kivunoni gate, 6.viii.84 cussed above. Hering's confusion over the status of
(M. Ritchie and B. Tengecho). this species and its misidentification as 'Napomyza
Gloriosa superba is native in East Africa and it is gentii' is in part explained by the fact that a second
not surprising to find that it has its own specific leaf specimen, a female, reared by Hopp does have the
miner. As an introduction in northern India, the spe- outer cross-vein, is larger and obviously represents
cies attacking Gloriosa proved to be the poly- the species described by Hendel as 'Napomyza gen-
phagous Tropicomyia atomella. tianella'. No mines collected by Hopp have been
preserved in Hering's mine collection.
The genitalia of Chr. hoppiella (Figs. 683, 684)
1d. Three new Chromatomyia species
indicate close relationship with Chr. gentianae (Figs.
on Gentianaceae
674,675) but the adults are readily distinguishable
There has been some confusion over species reared by the more yellow abdomen in Chr. gentianae.
from Gentiana, Blackstonia and Centaurium.
Hendel (1920: 150) described one species as Napo- Chromatomyia blackstoniae sp.n.
myza gentii (now in Chromatomyia). This species is (Figs. 669, 670)
distinctive, with the frons bright yellow and the
Adult. Closely resembling G. hoppiella but smaller,
outer cross-vein present. The genitalia are atypical
wing length 2.25 mm in male, costal sections 1,3, 4
of the group (Figs. 680, 681) in being largely mem-
in ratio 45 : 10 : 12; acrostichals less numerous, in 4
branous. Hering (1928: 553) misidentified this spe-
rows; fore-femora distinctly yellow at apex.
cies from 2 or more specimens reared from Gen-
Male genitalia. Aedeagus as in Figs. 669, 670.
tiana excisa near Lenzerheide, Switzerland by W.
Host. Blackstonia perfoliata, larvae feeding either in
Hopp. Hering at this stage was primarily interested
basal rosette or in stem-leaves, particularly those
in the differences in the mines of this species from
immediately below the flower; mine initially linear,
the most widespread species he referred to as Phyto-
later developing into irregular blotch, whitish pupar-
myza gentianae. I have 2 specimens of this series
ium remains in the mine.
before me. The female represents the species de-
Holotype d, Ireland, Co. Clare, Murrough,
scribed later by Hendel (1932: 310) as Napomyza
28.vi.65 ex mines colI. 16.vi.65; paratype Q, same
gentianella. This has the outer cross-vein present as
data (leg. G.CD. Griffiths), both in colI. Griffiths,
in Chr. gentii and Hendel rightly added 'Syn.
Canada.
Napom. gentii Hering (1928).' This is a widespread
Remarks. This species was misidentified by Griffiths
species in the mountains of central Europe.
(1968) as Phytomyza gentianae. The two species are
The second of the 2 specimens obtained by Hopp
distinguishable by the more yellow abdomen in Chr.
is a male and this lacks the outer cross-vein. The
gentianae. Surprisingly, the male genitalia are of en-
male genitalia show this to represent an undescribed
tirely different form from Chr. gentianae and other
species.
related species and do not conform to Griffiths'
Specimens reared from Blackstonia and Cen-
(1974a) re-definition of the genus Chromatomyia
taurium and incorrectly identified as Chromatomyia
based on male genitalia. However, the adult, type of
gentianae have proved to represent 2 further species
mine and puparium clearly associate this species
and are described below.
with the other Gentianaceae feeders and an expand-
ed definition of Chromatomyia is discussed below,
Chromatomyia hoppiella sp.n.
to accommodate this species and the fern feeders
(Figs. 683, 684)
previously included in Phytomyza.
Adult. Closely resembling Chr. gentianae Hendel I found mines on Blackstonia in the West of Eng-
(1920; 1932), with following essential characters: land at Kilcot, Glos., l1.viii.58 and I have also seen
frons uniformly brownish, orbits slightly paler; mines from northern France (leg. H. Buhr). Chr.
acrostichals in some 6 rows; legs entirely black; ab- blackstoniae is thus probably widely distributed
domen uniformly black; wing length 2.7 mm in male, with its host plant.
costal sections 2, 3, 4 in ratio 5.5: 1.4: 1.5.
New speci~ 397
Sweeping plants of lusticia diclipteroides at Langata, Melanagromyza cleomae was described from Cle-
near Nairobi on which I found an immature mine I ome (Capparaceae) and is also known on Brassica.
obtained a single male Pseudonapomyza which 'is M. ~ustriraphani was also described from Bangkok,
described below. ThaIland from Brassica ('red radish, cabbage?').
Sasakawa's illustrations of the genitalia of M. austrir-
Pseudonapomyza justiciae sp.n. aphani confirm that this is identical with M. cleo-
(Figs. 809, 810) mae (Figs. 308, 309).
Head. Orbits with 2 strong, equal reclinate upper or- Melanagromyza minimoides Spencer (1966f)
bitals, 2 weaker, inclined lower orbitals; orbital setu-
lae sparse, reclinate; ocellar triangle not differenti- Melanagromyza radicicola Steyskal, 1981: 40,
ated beyond foremost ocellus; jowls slightly ex- syn. nov.
tended at rear, in centre one-sixth vertical height of The unique holotype of M. radicicola was presumed
eye; third antennal segment round, with fringe of to have been reared from roots of Urtica sp. and was
short pubescence; arista long, equal to height of eye found dead in a rearing container. No puparium
distinctly pubescent. ' from which it might have emerged could be located.
Mesonotum. Damaged, chaetotaxy not detectable, Earlier, specimens of M. minimoides had been
398 Taxonomic Appendix
reared in the same container from Heliopsis and Cerodontha (Poemyza) oryzivora (Spencer, 1967c)
Rudbeckia. This problematic species was discussed
Cerodontha (Poemyza) flaviorbitalis Spencer,
by Spencer and Steyskal (1986: 246) and it was my
1975: 212, syn. nov.
opinion at the time that M. radicicola merely repre-
sented M. minimoides and some confusion in rear- Cerodontha (Poemyza) oryzivora was described from
ing records led to the assumption that the host was Malaysia and flaviorbitalis from Sri Lanka. The ge-
Urtica. I am now convinced that M. radicicola is nitalia of the two species are identical (d. Figs.
identical with M. minimoides and this new synony- 1394, 1395).
my is now formally established.
Liriomyza betae (Coquillett, 1900)
Tropicomyia theae (Cotes, 1896)
Haplomyza chenopodii Watt (1924: 683),
Tropicomyia coffeae (Koningsberger, 1897: 25), syn. nov.
syn. nov. Haplomyza imitans Malloch, 1934: 1, syn. nov.
(synonymy of imitans with chenopodii established
There has been confusion about the status of these
by Spencer (1963d: 333)
two minute, black species and the only significant
difference between these and other closely related Phytomyza betae Coquillett was treated as a homo-
species lies in the male genitalia. I am now satisfied nym of Phytomyza betae Macquart (1855: 258) by
that T theae is a wide-ranging, polyphagous species, Spencer (1977c). However, it has now been accept-
described from Sri Lanka, but with its range extend- ed that Macquart's description represents a nomen
ing to Indonesia and Papua New Guinea where it is nudum (Spencer and Martinez, 1987) and thus
particularly common. The most recent illustrations Coquillett's name betae is valid and has priority over
of the male genitalia of T theae (Sri Lanka: Fig. chenopodii Watt. This new synonymy is therefore
260) and coffeae (Papua New Guinea: Fig. 890) now formally established.
satisfy me that only a single species is involved.
Sasakawa (1977: 252) treated T styricicola Liriomyza pusilla (Meigen, 1830)
(Sasakawa, 1954b), with its main distribution in
Liriomyza Jasciola (Meigen, 1838: 204)
southern Japan, as synonymous with T theae. How-
(synonymy established by Spencer (1971b: 173))
ever, I consider the genitalia of T styricicola (Figs.
Liriomyza compositella Spencer (1961b: 87),
328, 329) are sufficiently differentiated to be ac-
syn. nov.
cepted as distinct from T theae (see under Styraca-
ceae, Dilleniidae above). The relationship between This species is one of a large group and was con-
T atomella (Malloch, 1914), T polyphaga (Spen- sistently misidentified by early workers, including
cer, 1961b), T styricicola and T theae needs Hendel (1935) who restricted the name pusilla to
further clarification. the species on Hieracium (Lactuceae) which was
subsequently described as Liriomyza hieracii
J apanagromyza tephrosiae (de Meijere, 1917) Spencer (1971b). The name pusilla was also widely
used in the United States but it is now clear that it
Japanagromyza badia Spencer (1977c: 111),
does not occur there.
syn. nov.
The species which is found commonly on Bellis in
J. tephrosiae was described from Java and is also Europe was identified by Hendel (1935) as Lirio-
known in northern India. The distinctive genitalia myza Jasciola Meigen but I established (Spencer,
(Fig. 386) are identical with those of 1. badia de- 1971b) that this represents the true pusilla, with
scribed from N.S.W., Australia. which Jasciola is synonymous. Additional hosts in
Europe are Aster and Solidago (surprisingly though
Agromyza nigrescens Hendel (1920) only the introduced S. canadensis is attacked, not
the native S. virga urea ). From 1971 therefore the
Agromyza nigrescens japonica Tsujita (1951: 53)
name pusilla has been restricted to the leaf miner on
(= Sasakawa), syn. nov.
Aster, Bellis and Solidago in Europe.
Agromyza oycoviensis Beiger (1960: 89),
Malloch (1914: 314) identified 9 caught speci-
syn. nov.
mens in Taiwan as Agromyza pusilla. However,
I have carefully considered the relationship between Hennig (1941: 173), after comparing the genitalia of
the subspecies japonica in Japan and A. oycoviensis a male of this series with a male from Europe (his
in Poland with A. nigrescens which is widespread in figs. 21, 22) rightly concluded that the species in
western Europe (see above under Geraniaceae) and Taiwan was not identical with the species he had
have decided that these two taxa cannot be ade- examined from Europe. Unfortunately it is clear
quately separated from typical nigrescens. from his fig. 22 that the European specimen was
New synonymies 399
something quite different and one of the complex of Liriomyza xanthocera (Czerny, 1909: 263)
grass-feeders. Hennig therefore referred to the pop-
Liriomyza crucifericola Hering (1951b: 31),
ulation from Taiwan as Liriomyza pusilla Malloch
syn. nov.
(nec Meigen). In view of this homonymy with Lirio-
Liriomyza cisti Spencer (1960c: 377),
myza pusilla (Meigen), I renamed the species
synonymy with xanthocera established by Spencer
L. compositella Spencer (1961b).
(1977b: 250)
During subsequent collecting in India (New Delhi
and Bombay), Sri Lanka, Hong Kong and Thailand, Examination of the male genitalia of L. crucifericola
I found this same species on Callistephus, Crasso- for the first time has shown that this species is syn-
cephalum, Solidago, Tithonia and Xanthium (not onymous with L. xanthocera (see above under Bras-
Gynura, a misidentification of Crassocephalum in sicaceae).
Sri Lanka). Sasakawa (1972) recorded L. composi-
tella on Bidens and Epaltes in Taiwan and also Chromatomyia platensis (Brethes, 1923)
identified caught specimens from Biak Island, Papua
Phytomyza winderi Steyskal (1976: 820),
New Guinea. I identified caught specimens on Lord
syn. nov.
Howe Island, Australia (Spencer, 1977c).
Hering (1957b) recorded mines of a Liriomyza Chromatomyia platensis is a widespread leaf miner
sp. on Xanthium strumarium in Bessarabia, Roma- on several genera of Lamiaceae in Brazil, Argentina
nia. I found a very large population on this host at and Chile. It is clear from Steyskal's description and
New Delhi in November, 1962, recorded as L. com- genitalia illustrations of Phytomyza winderi from
positella (Spencer, 1986b). Xanthium is an intro- Curitiba, Brazil that the species is identical with
duced plant of almost cosmopolitan distribution but Chromatomyia platensis.
probably of South American origin and it is logical
to assume that its leaf miner is a species with an orig- Phytomyza brischkei Hendel (1922)
inal population on a native host, such as Bellis in
Phytomyza anthyllidis Hering (1954: 128) and
Europe.
Groschke (1957: 124), syn. nov.
Re-examination of the genitalia of specimens
Phytomyza vulnerariae Spencer (1957a: 25),
reared from Bellis and Solidago in Europe and
syn. nov.
comparison with those from Crassocephalum and
Xanthium in India has convinced me that a single Hering (195 7b), after examining P. anthyllidis and
species is involved, with a remarkably wide range, P. vulnerariae, included them both in his key to
comparable to but not identical with that of Chro- mines on Anthyllis, considering that P. vulnerariae,
matomyia horticola The synonymy of Liriomyza only known mining sepals, must be distinct from the
compositella with L. pusilla is therefore formally es- leaf miner, P. anthyllidis. He also misinterpreted
tablished herewith. P. anthyllidis, referring to the third antennal seg-
ment as round, whereas it is distinctly enlarged,
Liriomyza scorzonerae Ryden (1951) slightly concave above and cut away below. I have
examined a male of P. brischkei ex Trifolium repens,
Liriomyza scariolae Hering (1955: 206),
from Berlin and the holotypes of P. anthyllidis and
syn. nov.
P. vulnerariae, and the distinctive genitalia (Fig.
Examination of the male genitalia of the holotype of 447) are identical in the three species. P. anthyllidis
L. scariolae from West Germany has shown that this and P. vulnerariae are therefore formally synonym-
is identical with L. scorzonerae (Figs. 988, 989) ised with P. brischkei herewith.
from S. Sweden and this new synonymy is estab-
lished herewith. Phytomyza chaerophylli Kaltenbach (1856)
External characters of L. scariolae, with 1 upper
Phytomyza sisonis Hering (1943: 58),
orbital bristle and 2 lower and the long pubescence
syn. nov.
of the third antennal segment, also the similar larval
mines, suggested that the species was probably iden- Although only a single female is known from Sison,
tical with L. scorzonerae even before examination of there is little doubt that P. sisonis merely represents
the genitalia. Hering was misled into believing the the common P. chaerophylli and this new synonymy
species from the different hosts Scorzonera and is formally established herewith.
Lactuca must be distinct, an error he made on
several occasions by attaching too much importance
to the assumed monophagy of agromyzids.
400 Taxonomic Appendix
I only accept Ballota as host of P. salviae with con- 3b. Ophiomyia blepharidis (Spencer, 1960a),
siderable reserve. comb. nov.
In the grounds of the Old Citadel at Istanbul I did
Melanagromyza blepharidis (Spencer, 1960a: 18).
find BaUota nigra with many mines in January, 1960
Holotype a from South Africa in Natal Museum,
but these only produced a single female Phytomyza
Pietermaritzburg
(without the outer cross-vein) and the identity of this
species cannot be established. I tentatively identified The genitalia (Figs. 801, 802) and the pupal spir-
it as P. abdita Hering (1927b, cf. Spencer, 1977b: acles (Fig. 803) of the unique holotype show that
251) (host: Ajuga) but I now feel this identification this species correctly belongs in Ophiomyia.
is doubtful.
3c. Phytoliriomyza robiniae (Valley, 1982),
Phytomyza socia Brischke (1880)
comb. nov.
Phytomyza campanariae Nowakowski (1958: 13),
Liriomyza robiniae Valley (1982: 781).
syn. nov.
Holotype a from Pennsylvania in U.S.N.M.,
The genitalia of these two species, P. campanariae Washington, D.C.
and P. socia (Figs. 66, 67) appear identical. In his
Valley described Liriomyza robiniae, a species
description Nowakowski considered that it was im-
forming blotch mines on Robinia pseudoacacia in
probable that Anemone and Pulsatilla would have
Pennsylvania, New York, Virginia and West Vir-
any agromyzid in common and, while comparing
ginia. In the description it was stated that 'the spe-
P. campanariae with a number of other species, he
cies is anomalous in the genus Liriomyza'. Examina-
did not consider P. socia which was not recognised
tion of the male genitalia of the holotype satisfies me
as distinct until later (Spencer, 1976a). Now that bo-
that this species cannot be retained in Liriomyza. At
tanists tend to include Pulsatilla within Anemone
this stage I am unwilling to erect a monotypic new
and as Nowakowski was unaware of P. socia when
genus to accommodate it and it is provisionally
describing P. campanariae the two can now by syn-
transferred to Phytoliriomyza.
onymised.
This isolated species can be briefly characterised
as follows:
3. New combinations Externally resembling Liriomyza, differing in the
lack of a fully developed pre-sutural dorso-central
bristle (dc); the first and second dc (nearest the scu-
Eighteen new combinations are established below,
tellum) are normal but these are followed in line by
including 7 species transferred from Phytomyza to
up to 6 bristles decreasing in size, with normally 3
Napomyza in South Africa, Australia and New Zea-
post-sutural and 3 pre-sutural; head with orbits dis-
land.
tinctly projecting above eye in profile; conspicuous
epistoma above mouth-margin; palps expanded dis-
3a. Ophiomyia alternantherae (Spencer, 1963c), tally, paddle-shaped; mesonotum uniformly mat, sil-
comb. nov. very-grey; scutellum black at sides but distinctly yel-
low centrally; stridulating organ lacking; male geni-
Melanagromyza alternantherae Spencer
talia: aedeagus, with distiphallus (Figs. 387, 388)
(1963c: 307). Holotype a from Argentina
weakly pigmented, divided distally; sperm pump
in U.S.N.M.
minute, black; epandrium with conspicuous rod-like
This minute black species, with wing length of 1.75 sclerite attached to wall at midpoint and an addi-
mm, is a leaf miner on Alternanthera philoxeroides. tional hooked sclerite at rear; surstyli rounded, with-
Although appearing like a small Melanagromyza, it out bristles, at most with a few weak hairs; cerci ap-
cannot be retained in this genus, in which the larvae parently lacking.
feed in stems or seeds. The genitalia (Figs. 241, 242) The description by Valley includes comments by
show that the species is isolated and this is further G. Steyskal referring to 'abundant stridulatory
confirmed by the larva, which has 20 or so finger- spicules on the lateral abdominal membrane of the
like posterior spiracular pores (Fig. 243). Until this male' and on this evidence the species was assigned
group of neotropical leaf miners can be further re- to Liriomyza. However, I cannot confirm the pres-
vised and a new genus erected for them, the species ence of the 'file' on the abdomen and the 'scraper' on
is provisonally transferred to Ophiomyia. the hind-femora is also lacking. The aedeagus itself
Externally 0. alternantherae cannot be distin- is also not consistent with inclusion in Liriomyza.
guished from 0. pfaffiae, a leaf miner on Pfaffia ire- I consider that robiniae is a relict and its evolu-
sino ides known in Venezuela. tionary affiliations are now obscured. It can best be
402 Taxonomic Appendix
interpreted as an offshoot from the primitive genus genitalia or larval feeding (see under Ranunculaceae
Phytoliriomyza and therefore pending further study above).
and possibly the discovery of further closely related
species, it is best placed in Phytoliriomyza.
South Africa
3d. Galiomyza takakoae (Sasakawa, 1954c),
comb. nov.
3f. Napomyza eximia (Spencer), comb. nov.
Liriomyza takakoae Sasakawa (1954c: 116)
from Kyoto, Japan. Holotype d in Entomological Phytomyza eximia Spencer (1964a: 35). Holotype d
Laboratory, Saikyo University in Natal Museum, Pietermaritzburg.
This leaf miner on Viola is certainly monophyletic
with Galiomyza violiphaga and G. violivora (see 3g. Napomyza renovata (Spencer), comb. nov.
Violaceae: Dilleniidae) and is now formally trans-
Phytomyza renovata Spencer (1960a: 35). Holotype
ferred to Galiomyza.
<:;? in Natal Museum, Pietermaritzburg.
type d in California Academy of Sciences, San 3s. Ptochomyza mayeri (Spencer), comb. nov.
Francisco.
Phytagromyza mayeri Spencer, 1966e: 304.
This species was mistakenly described in Phyto-
Phytomyza mayeri, L. Suss, 1977: 160.
myza, the dorsal sclerites erroneously being con-
As Phytomyza czernyi Strobl, 1909 (Hering, 1958:
sidered as the distiphallus. The genitalia (Figs. 89,
77).
90) are generally similar to those of Chr. clemati-
A third species discovered in stems of Clematis
vora (Coquillett, cf. Spencer and Steyskal, 1986).
recta together with Phytomyza clematidicaulis and
Chr. clematoides is one of the small number of spe-
P. novitzkyi by Novitzky near Vienna (Hering,
cies in the genus which pupates on the ground, not
1958) has surprisingly proved to belong correctly in
in the mine as is normal.
the genus Ptochomyza.
This puzzling species was originally misidentified
3n. Chromatomyia ciliata (Hendel), comb. nov. by Hering (1958) as Phytomyza czernyi Strobl, then
described as Phytagromyza mayeri Spencer (1966e),
Phytomyza ciliata Hendel (1935: 375). Holotype d
following comparison with the male type of Phyto-
from Austria in Naturhistorisches Museum, Vienna.
myza czernyi, and finally transferred to Phytomyza
The male genitalia of this species (Fig. 1168), the
by L. Suss (1977).
internal pupation and the white, 'slipper-shaped'
Hering (1967: 32) rediscovered this species on
puparium indicate inclusion of this species in Chro-
Clematis flammula on the island Hvar, Yugoslavia
matomyia.
and illustrated the larval spiracles. With only 3 pores
The following 4 feeders on ferns are transferred on each posterior spiracular process, the species
to Chromatomyia on the basis of the form of their clearly does not belong in Phytomyza and the larval
puparia and method of pupation (see discussion characters associate the species with the two Pto-
under Chromatomyia below). chomyza species which have been reared from
Asparagus, Pt. asparagi Hering and Pt. asparagivora
Spencer. The head of Pt. mayeri (Spencer, 1966e:
30. Chromatomyia cheilanthus (Garg),
fig. 34) and the wing (L. Suss, 1977: fig. 19) are also
comb. nov.
typical of Ptochomyza. Suss (1977: fig. 23) also il-
Phytomyza cheilanthus Garg (1971: 251) ex Chei/- lustrated the genitalia of a specimen he discovered
anthes. Holotype d in Insect Survey of India, Cal- in Italy and a further illustration is given here (Fig.
cutta. 93).
Phytomyza mayeri is therefore now formally
transferred to Ptochomyza herewith.
3p. Chromatomyia dorsata (Hendel),
comb. nov.
Phytomyza dorsata Hendel (1920: 157) ex Ceterach. 4. New status
Holotype d from Ragusa, Yugoslavia in Naturhis-
torisches Museum, Vienna. Phytomyza senecionis ssp. ravasternopleuralis
Sasakawa (1955b: 19), new status.
Griffiths (1972b) raised ssp. ravasternopleuralis
3q. Chromatomyia dryoptericola (Sasakawa),
to specific rank following the apparent difference
comb. nov.
between the aedeagus of typical P. senecionis in Eu-
Phytomyza dryoptericola Sasakawa (1961: 446) ex rope and ssp. ravasternopleuralis in Japan. However,
Dryopteris. Holotype d formally deposited in Ento- the specimen of P. ravasternopleuralis illustrated by
mology Laboratory, Kyoto Prefectural University Griffiths was damaged, thus accounting for the dif-
but now mislaid (Sasakawa, pers. comm.). ference he thought he detected. There is in fact no
significant difference in the genitalia between the
two populations and P. ravasternopleuralis in Japan
3r. Chromatomyia scolopendri
is again now treated merely as a subspecies of
(Robineau-Desvoidy), comb. nov.
P. senecionis (see text above under Senecioneae).
Phytomyza scolopendri Robineau-Desvoidy (1851:
402) ex 'Ie scolopendre', types lost.
A species transferred to Ptochomyza Hering
(1942).
404 Taxonomic Appendix
s.s. in 2 or more characters, all but T. clutiae with In view of the intermediate status of T. clutiae
the costa extending to vein M1 + 2. I had proposed and T. vigneae, no host known of T. capeneri, only
to erect a new genus to accommodate these but on females known of T. dicksoni, T. philocroton and
reviewing the species I now feel it is more appro- T. laburnifoliae and the apparently isolated position
priate, at least temporarily, to 'lump' them within of T. crotalariae and T. crotonella, it is obviously
Tropicomyia. . undesirable to erect a new genus for species which
Re-examination of T. clutiae (Spencer, 1963b) are probably not monophyletic. It seems better at
has necessitated a slight expansion of the original this stage to leave all these species in an expanded
concept of Tropicomyia. While the wing venation, concept of Tropicomyia, until more information be-
larval spiracles and genitalia (Figs. 554, 555) are comes available permitting more rational splitting
typical of the genus, the larval mouth-parts (Fig. into one or more further genera.
552) each have only 2 large teeth and the leaf mine
is upper surface, not epidermal.
7c. Chromatomyia Hardy (1849)
T. vigneae (Seguy, 1951) has been accepted in
Tropicomyia, although the costa extends to vein Hardy (1949) named as Chromatomyia a group of
M1 + 2, the larval mouth-parts (Fig. 431) are slightly species he separated from Phytomyza 'on account of
atypical, there is a proliferation of pores on the pos- differences in the pupa state, accompanied by a cor-
terior spiracles (Fig. 432), the leaf mine is not ex- responding variation in habit,' with 'slipper-shaped
clusively epidermal and the male genitalia diverge pupae, whose transformations take place entirely
from the form characteristic of Tropicomyia s.s. within the leaf.' Griffiths (1974a) restricted the con-
T. capeneri Hering (1957a) was reared from an un- cept of Chromatomyia on the basis of genitalia stu-
identified host in South Africa and was originally dies to 'all those species of Phytomyza s.l. in whose
confused and synonymised with T. vigneae (Spen- males the distal section of the ejaculatory duct is
cer, 1964: 29) from specimens reared from Vigna in simple (not bifid) and lies below a lobe on the "dor-
Senegal. The genitalia of this population are close to sal" (in anteriorly directed rest position) side of the
but not identical with those of T. vigneae (Spencer, aedeagus.' Eighty species in 10 families of dicots
1973a: figs. 272, 273) and this synonymy was then (excluding the polyphagous Chr. horticola) and 2
withdrawn. With no host known, T. capeneri need families of monocots fully conform to this definition
not be further discussed here, but it is accepted as (cf. Chr. gentianae: Fig. 674; Chr. milii: Fig. 1431).
distinct from T. vigneae, particularly in view of the In all but 2 species (Chr. alpigenae and Chr. cha-
distinctive leaf mine (Hering, 1957a: fig. 9). maemetabola, both on Lonicera, Caprifoliaceae) the
T. ceratiosicyi Hering (1957a), feeding on Cera- puparium remains within the mine (according with
tiosicyos in South Africa (Achariaceae: Dilleniidae), Hardy's definition of Chromatomyia), the anterior
is unusual in the form of the larval posterior spir- spiracles bending downwards and projecting
acles, with each process having an ellipse of 20-25 through the epidermis.
pores. With the study of more material than originally
Five further little black leaf miners are known in considered by Griffiths, it is apparent that there are
Africa which were all described in Melanagromyza problems in the delimitation of Chromatomyia. Two
but transferred to Tropicomyia by Spencer (1973a). species which must be included in this genus on the
All have the costa extending to vein M1 + 2 but 3 basis of their obvious affinities with other species on
are known only from females. T. dicksoni Hering the same host family (Gentianaceae) - Chr. gentii
(1957a) was reared from an unknown host in South (Fig. 680) and Chr. blackstoniae (Figs. 669, 670)
Africa (possibly Bosquiea sp., Moraceae) and have genitalia which cannot be encompassed within
Hering considers the mine to be possibly epidermal. Griffiths' restricted definition. Also I have now
In T. philocroton (Hering, 1957a) no mine is pre- transferred the 4 fern feeders, cheilanthus, dorsata,
served and no further comment is possible. T. labur- dryoptericola and scolopendri from Phytomyza to
nifoliae (Spencer, 1964a) was described from Cro- Chromatomyia (see section 3 above). The identical
talaria laburnifolia in Ethiopia. The mine is not epi- form of puparium and method of pupation I saw re-
dermal but no larvae are known and the exact status cently in fresh specimens of scolopendri convinced
of this species remains uncertain. me that this group should logically be considered as
Two further species diverge in most characters monophyletic with the main group of Chromato-
from Tropicomyia - T. crotalariae (Hering, 1957a) myia species, despite the differing genitalia (Chr.
and T. crotonella (Spencer, 1964a), both from scolopendri, Fig. 30).
Ethiopia. It seems clear that these two species are Griffiths (1974a) mentions that 3 other groups -
not monophyletic with Tropicomyia s.s. but prob- the Phytomyza robustella group, the P. anemones
ably also not even with each other (cf. T. crotalariae, group and the P. ilicis group transform in the same
Figs. 507, 508; T. crotonella: Figs. 556, 557). manner as typical Chromatomyia species. Depend-
406 Taxonomic Appendix
ing on the weight given in evolutionary terms to However, it should be mentioned here that Chro-
male genitalia on the one hand and form of pupar- matomyia Walker (1849, date of publication 8 De-
ium and method of pupation on the other, differing cember) formally antedates Chromatomyia Hardy
classifications of this whole group of species will be (1849, date of publication December (no date) and
arrived at. This must be a matter of more detailed therefore taken as 31 December). However, Walker
research in the future and I cannot discuss the pro- himself in his original publication of Chromatomyia
blem further here. I can only emphasise that I con- on p. 1162 in a section 'Errata and Addenda' for-
sider the generic limits of Chromatomyia are not as mally synonymised Chromatomyia with Lampro-
clear-cut as originally defined by Griffiths and my gaster Macquart. The name Chromatomyia Walker
inclination now is to revert to Hardy's original con- thus has no validity.
cept.
Acknowledgements
This study has in large part been based on adult and supplied much unpublished information on his
agromyzids and larval leaf mines I personally col- research on the larval feeding of this genus on
lected in many parts of the world during the past 25 numerous previously unknown hosts.
years. However, I could not have carried out the Assistance with botanical problems was willingly
work without frequently consulting the collections given by the staff at the Herbarium, British Museum
in the Department of Entomology at the British (Nat. Hist.). It is impossible to thank individually all
Museum (Natural History) and I would like to thank the botanists I consulted but I mention particularly
the Trustees and staff of the Diptera Section for the Mr. Arthur Chater and Dr. Chris Humphries and
facilities provided and the willing help I received. also, at the Royal Botanic Gardens, Kew, myoid
Type specimens were also borrowed from the fol- friend Jan Gillett. I would also like to thank Prof.
lowing Institutions: J.B. Harborne, (Department of Botany, University
of Reading) for an interesting discussion on plant
Zoologisches Museum, Berlin, DD.R.
chemistry.
(Dr. H. Schumann)
The librarians of both the Entomology and Bo-
Naturhistorisches Museum, Wien
tany Libraries at the BM gave me invaluable assis-
(Dr. R. Contreras-Lichtenberg)
tance during my many visits to London and special
Museum National d'Histoire Naturelle
thanks are due for their willing help.
(Dr. L. Matile)
Zoologisch Museum, Amsterdam (B. Brugge) I must also thank the Trustees of the Nuffield
Istituto di Entomologia Agraria, Milan Foundation for their support from the 'Small Grants
Scheme for Research in Science' primarily to cover
(Prof. L. Suss)
travel expenses from Cornwall to the Natural His-
Natal Museum, Pietermaritzburg (Dr. J.G.H. Londt)
tory Museum, and also the British Entomological
Institute of Systematic Zoology, Poznan, Poland
Society for a grant from the Professor Hering Me-
(Dr. M. Beiger)
morial Research Fund towards typing expenses.
Entomologisches Institut, Zurich
Special thanks are due to David and Diane
(Prof. Dr. W. Sauter)
Henshaw for checking much of my manuscript and
I express my appreciation and thanks to the for preparing on their computer the many Host
above colleagues for their efficient cooperation in Tables which form an important part of this book.
supplying all loans I requested. Our many discussions in London and on the phone
I must also thank Dr. M. von Tschirnhaus (Biele- materially assisted me in completing my work in the
feld) and Ing. M. Martinez (INRA, Versailles) for allotted time.
advice given in many telephone conversations, and Finally I must thank my wife Ann for finding the
particularly Dr. T. Jermy (Budapest) for the encour- time to prepare most of the drawings of genitalia,
agement given in many letters and phone calls. which are so significant for proper understanding of
Prof. H. Suss (Berlin, DD.R.) kindly met me in Ber- the Agromyzidae.
lin for discussions on the important genus Phytobia
References
Ackery, P. R., 1988. Hostplants and classification: a review of and systematics of some Phytomyza Fallen species (Diptera,
nymphalid butterflies. Bioi. 1. Linn. Soc. 33: 95-203. Agromyzidae). Bull. Soc. Amis Sci. et lettres Poznan. Ser. D.
Aguilar, J. d', 1945. Description d'un Liriomyza nouveau vivant 20: 135-143.
sur Riccia natans L. Bull. Soc. ent. Fr. 84: 143-6. __ 1981. Studies on mining insects of the Tatry National Park.
Aguilar, J. d', Chambon, J.-P. et Touber, F., 1976. Les Agromyza 12. (in Polish). Poznanskie towarzystwo przyjacial nauk 57:
mineurs de feuilles de cereales (Dipteres, Agromyzidae) dans 3-68.
la region parisienne. Ann. Zool.-Ecol. animo 8(4): 579-593. Benson, L., 1948. A Treatise on the North American Ranunculi.
Allan, H. H., 1961. Flora of New Zealand. Vol. 1. Wellington, Amer. Mid!. Nat. 40(1): 1-261.
N.Z.,pp.l-l085. Beri, S. K, 1971. Immature stages of Agromyzidae (Diptera)
Allen, P., 1956. Observations on the biology of some Agromyzi- from India. VIII. Taxonomy and Biology of Twentyone spe-
dae(Diptera). Proc. R. ent. Soc. Lond. A31 (7/9): 117-131. cies of Phytomyza Fallen. Orient. Ins. Suppl. 1: 119-164.
__ 1957. Larval morphology of some species of Phytomyza Blanc-Louvel, C., 1984. Le genre 'Ranunculus L.' dans Ie berria-
Fallen (Diptera: Agromyzidae). Proc. R. ent. Soc. London A sien (Cretace Inf.) de la province de Lerida (Espagne). Ilerda
32: 171-181. 45: 81-92.
Beiger, M., 1960a. Owady minujace Ojcowskiego Parku Naro- Blanchard, E., 1938. Descripciones y Anotaciones de Dipteros
dowego. Pro Kom. bioi., PTPN, 23(2): 1-154 (87-240). argentinos. AGROMYZIDAE. An. Soc. cient. Argentina 126:
__ 1960b. Liriomyza mycelis Beiger - a new species of the 352-359.
family Agromyzidae (Diptera) which mines leaves of Mycelis Blanchard, E., 1954. Sinopsis de los Agromyzidos argentinos.
muratis (L.). Bull. Soc. Amis Sci. Lett. Poznan, (D) 1: 161- Min. Agr. Ganaderia (A) 56: 1-50.
167. Block, K, 1974. Chromosomal variation in Agromyzidae (Dip-
__ 1961. A new Phytomyza species from Poland (Dipt. Agro- tera) III. Cerodontha (Butomomyza) eucaricis Nowakowski -
myzidae). Bull. Soc. Amis Sci. Lett. Poznan, (D) 2: 37-42. two semispecies or sibling species. Hereditas 78: 125-140.
__ 1964. Studies on mining insects of the Tatra National Park. Bremer, K, 1987. Tribal Interrelationships of the Asteraceae.
2. Phytomyza aphyllae n. sp. (Diptera, Agromyzidae). Bull. Cladistics 3(3): 210-253.
Soc. Amis Sci. Lett. Poznan, (D) 4: 55-60. Bremer, K and Humphries, C. J., in press. Generic Monograph
__ 1972a. Studies on mining insects of the Tatra National of the Asteraceae-Anthemideae. Bull. Brit. Mus. (Nat. Hist.).
Park. 3. Phytomyza tatrica sp.n. and Phytomyza skuratowiczi Brethes, J., 1923. Sur un Diptere mineur des feuilles de Salvia
sp.n. (Diptera, Agromyzidae). Polskie Pismo ent. 42(1): 109- splendens et deux Hymenopteres, ses parasites. Rev. Zoo!.
120. Agric., Bordeaux 22(6): 153-6.
__ 1972b. Studies on mining insects of the Tatra National Brischke, C. G. A., 1880. Die Blattminirer in Danzig's Umge-
Park. 4. Liriomyza gypsophilae sp.n. and Liriomyza clarae bung. Danzig, pp. 1-58.
sp.n. (Diptera, Agromyzidae). Polskie Pismo ent. 42(1): 121- Buhr, H., 1930. Einige Blattminen und Gallen von der Insel
130. Lesina (Hvar) in Dalmatien. Sitzber. und Abh. naturf Ges.
__ 1972c. Studies on mining insects of the Tatra National Rostock 3(2): 128.
Park. 6. Phytomyza pseudogentii sp.n. (Diptera, Agromyzi- __ 1937. Parasitenbefall und Pflanzenverwandschaft. Bot. Jb.
dae). Bull. Acad. pol. Sci., Bioi. 20(3): 183-186. 68: 142-198.
__ 1972d. Materials concerning the knowledge of the fauna of __ 1953. Die Frassbilder und weiteren Nahrungspflanzen der
mining insects in South-Eastern Poland. II. Liriomyza aposeri- an Vicia faba L. lebenden Minierinsekten. Beitr. En!. 3(3):
dis sp.n. (Diptera, Agromyzidae). Bull. Acad. pol. Sci., Bio!. 258-278.
20(4): 233-236. __ 1954. Mecklenburgische Minen. Archiv Freunde Natur-
__ 1975. Beitrag zur Kenntnis der minierenden Insekten geschichte Mecklenb. NF I: 238-288.
Siidostpolens. IV. Eine neue Art aus der Gruppe der Phyto- __ 1960. Bemerkenswerte oder neue Gallen und Minen aus
myza symphyti Hd. (Diptera, Agromyzidae). Polskie Pismo Thiiringen. Mitt. thur. bot. Ges. 2(1): 56-150.
ent. 45(1): 133-139. Cantino, P. D. and Sanders, R. W., 1986. Subfamilial Classifica-
__ 1977. Phytomyza (Chromatomyia) griffithsiana sp.n. (Dip- tion of Labiatae. Syst. Bot. 11(1): 163-185.
tera, Agromyzidae), a new mining fly on Mulgedium alpinum Claassen, P. W., 1918. Observations on the life history and biol-
Less. Bull. Acad. pol. Sci., BioI. 25(5): 317-321. ogy of Agromyza laterella Zetterstedt. (Diptera). Ann. Ent.
__ 1979. Liriomyza bulgarica sp.n. and Phytomyza rhodopaea Soc. Am. 11: 9-18.
sp.n. (Diptera, Agromyzidae), two new mining flies from Bul- Clayton, W. D. and Renvoize, S. A., 1986. Genera Graminum
garia. Bull. Acad. pol. Sci., Bioi. 26(1978) (10): 687-693. Grasses of the World. Kew Bull. Add. Ser. 13, pp. 389.
__ 1980a. Paraphytomyza [raxini sp.n. (Diptera, Agromyzi- Collin, J. E., 1911. A new fly of the genus Agromyza from Egypt.
dae), a new mining fly from Eastern Bulgaria. Bull. Soc. Amis Bull. Ent. Res. 2(1): 79-80.
Sci. Lett. Poznan (D) 20: 145-148. Coquillett, D. W., 1900. Descriptions of two new species of Dip-
__ 1980b. Studies on mining insects of the Tatry National tera from Western Australia. Proc. Linn. Soc. N.S. W. 25:
Park. 11. Materials concerning the morphology, taxonomy 389-901.
References 409
Costa Lima, A. da, 1950. Agromyzideos (Diptera) do tomateiro the Families of Dicotyledons. Smithsonian Contribution to
e de orquideas. Arq. Inst. Bioi. 20(3): 35-40. Botany 58, Washington, 314 pp.
Cotes, E. c., 1896. An account of the insects and mites which Greathead, D. J., 1969. A study in East Africa of the bean flies
attack the tea plants in India. Indian Mus. Notes 3(4): 1-74. (Dipt., Agromyzidae) affecting Phaseolus vulgaris and of their
Crabbe, J. A., Jermy, A. C. and Mickel, J. T., 1975. A new natural enemies, with the description of a new species of
generic sequence for the Pteridophyte Herbarium. Fern Gaz. Melanagromyza Hend. Bull. ent. Res. (1968) 59: 541-561.
11: 141-162. __ 1971. A new species of Melanagromyza Hendel (Dipt.,
Cronquist, A., 1968. The Evolution and Classification of Flower- Agromyzidae) from pods of the bonavist bean (Dolichos lab-
ing Plants. Nelson, London, pp. 396. lab) in Uganda. Bull. ent. Res. 60: 463-465.
__ 1981. An Integrated System of Classification of Flowering Greene, C. G., 1914. The cambium miner in river birch. J. agric.
Plants. Colombia Univ. Press, pp. 1262. Res. 1(6):471-474.
Cronquist, A., Takhtajan, A. and Zimmermann, W., 1966. On Greguss, P., 1970. Ein Callitris iihnliches Holz aus dem Tertiar
the higher taxa of Embryobionta. Taxon 15(4): 129-134. in Limburg (Niederlande). Senk. leth. 51: 265-275.
Czerny, L. and Strobl, G. Spanische Dipteren III. Verh. zool.- Griffiths, G. C. D., 1955. A new British species of Agromyzidae
bot. Ges. Wien 59 (6): 121-301. (Diptera). Entomologist's Gaz. 6: 10-12.
Dahlgren, R. M. T., 1980. A revised system of classification of __ 1956. A note on some Agromyzidae (Dipt.) from Italy, in-
the angiosperms. Bot. J. Linn. Soc. 80: 91-124. cluding a species new to science. Entomologist's mono Mag.
Dahlgren, R. M. T., Clifford, H. T. and Yeo, P. F., 1985. The 42: 145-6.
Families of the Monocotyledons. Springer Verlag, 520 pp. __ 1959. A new species of Phytomyza (Dipt. Agromyzidae)
Darvas, B., Szarukan I. and Papp, L., 1988. Napomyza gym- near Phytomyza affinis Fallen. Ibid. 10: 101-6.
nostoma (Loew) (Dipt., Agromyzidae). An agromyzid pest on __ 1963a. The Agromyzidae (Diptera) of Woodwalton Fen.
leek in Hungary (in Hungarian). NdvenyvMelem 24. Evfolyam Entomologist's mon. Mag. 98(1962): 125-58.
(10): 450-454. __ 1963b. A revision of the palaearctic species of the nigripes
Dethier, V. G., 1954. Evolution of feeding preferences in group of the genus Agromyza Fallen (Diptera, Agromyzidae).
phytophagous insects. Evolution, N. Y. 8: 33-54. Tijdschr. Ent. 106(2): 113-168.
Dressler, R. L., 1981. The orchids. Harvard University Press, pp. __ 1964a. A note on some Agromyzidae (Diptera) from
332. northern Persia, including a new Liriomyza species. Entomol-
Ehrlich, P. R. and Raven, P. H., 1964. Butterflies and Plants: a ogist's mon. Mag. 99(1963): 75-79.
Study in Coevolution. Evolution N. Y. 18: 586-608. - - 1964b. The Agromyzid Fauna of Iceland and the Faroes,
Enderlein, G., 1936a. 22. Ordnung: Zweifliigler. Diptera. Tier- with Appendices on the Phytomyza milii and robustella
welt Mitteleur. 6(2) (= Insekten 3. Teil): 1-259. Groups (Diptera, Agromyzidae). Ent. Medd. 32: 393-450.
__ 1936b. Notizen zur Klassifikation der Agromyziden __ 1966. The Agromyzidae (Diptera) of Greenland. Meddr
(Dipt.). Mitt. dt. ent. Ges. 7: 42. Gronland 170(4): 1-32.
Fallen, C. F., 1823. Phytomyzides et Ochtidiae Suecia, pp. 1-10. __ 1967a. Revision of Phytomyza syngenesiae group (Diptera:
Lund. Agromyzidae), including species hitherto known as 'Phyto-
Farr, E. R., Leussink, J. A. and Stafleu, F. A., 1979. Index Nomi- myza atricornis Meigen'. Stuttg. Beitr. Naturk. 177: 1-28.
num Genericorum. Vols 1-3, pp. 1896. Bohn, Scheltema and __ 1967b. Notes on the genus Napomyza Westwood (Diptera:
Holkema, Utrecht; Dr. W. Junk, The Hague. Agromyzidae). Proc. R. ent. Soc. Lond. (B) 36: 128-130.
Frick, K. E., 1953. Further studies on Hawaian Agromyzidae - - 1968. Agromyzidae (Diptera) from Ireland. Proc. R. Irish
(Diptera) with Descriptions of Four New Species. Wash. State Acad. 67(B, 2): 37-61.
agr. Exp. Sta. Bull. 1182: 207-215. __ 1972a. Studies on boreal Agromyzidae (Diptera). I. Phyto-
__ 1955. Nearctic species in the Liriomyza pusilla complex, myza mines on Saxifragaceae. Quaest. ent. 8: 67-80.
No.3. L. alliovora, new name for the Iowa onion miner. J. __ 1972b. Studies on Boreal Agromyzidae (Diptera). II. Phy-
Kans. ent. Soc. 28(3): 88-92. tomyza miners on Senecio, Petasites and Tussilago (Composi-
__ 1956. Revision of the North American Calycomyza species tae, Senecioneae). Quaest. ent. 8: 377-405.
North of Mexico (Phytobia: Agromyzidae, Diptera). Ann. ent. - - 1972c. The phylogenetic classification of Diptera Cyclorrha-
Soc. Am. 49: 284-300. pha, with special reference to the structures of the male post-
Frohne, D., Jensen, u., 1985. Systematik des P[lanzenreichs. abdomen. Ser. ent. 11 Dr. W. Junk, The Hague, pp. 1-340.
Gustav Fischer Verlag, Stuttgart, pp. 355. __ 1973a. Studies on boreal Agromyzidae (Diptera). III. Phy-
Frost, S. W., 1924. A study of the leaf-mining Diptera of North tomyza miners on Cnidium and Conioselinum (Umbelliferae).
America. Mem. Cornell Univ. Agric. Exp. Stn. 78: 1-228. Quaest. ent. 9: 3-11.
Frost, S. W. and Sasakawa, K., 1954. New Agromyzidae from - - 1973b. Studies on boreal Agromyzidae (Diptera). IV. Phy-
Japan. II. Mushi 27(7): 49-52. tomyza miners on Angelica, Heracleum, Laserpitium and Pas-
Garg, P. K., 1971. Studies on Agromyzidae (Diptera) from the tinaca (Umbelliferae). ibid. 9: 219-253.
Gangetic Basin. Part V. Descriptions of three new species of __ 1973b. Description of an unnamed Paraphytomyza (Dip-
Phytomyza Fallen. Orient. Ins. Suppl. 1: 247-256. tera, Agromyzidae). Entomologist's Gaz. 24: 307-310.
Geissert, F., Niitzold T. & Siiss, H., 1981. Pflanzenfossilien und __ 1974a. Studies on boreal Agromyzidae (Diptera). V. On
Palaeophytobia salicornia Siiss, eine neue fossile Minierfliege the genus Chromatomyia Hardy, with revision of Caprifolia-
(Agromyzidae, Diptera) aus dem Plioziin des Elsass. Mitt. ceae-mining species. Quaest. ent. 10: 35-69.
bad. Landesver. Naturk. Naturschutz (N.F.) 12: 221-231. - - 1974b. Studies on boreal Agromyzidae (Diptera). VI. Fur-
Givulescu, R., 1984. Pathological elements on fossil leaves from ther Phytomyza miners on Senecioneae (Compositae). ibid.
Chiuzbaia (galls, mines and other insect traces). Dari Seama 10: 103-129.
Sed. Inst. Geol. Rom. 68(3): 123-133 (1981). __ 1974c. Studies on boreal Agromyzidae (Diptera). VII. A
Godfray, H. C. J., 1985. Chromatomyia paraciliata sp.n. (Dip- new Chromatomyia miner on Valeriana. ibid. 10: 217-222.
tera: Agromyzidae, a leaf-miner of Leucanthemum vulgare - - 1974d. Studies on boreal Agromyzidae (Diptera). VIII.
from southern England. Entomologist's Gaz. 36: 47-50. Phytomyza miners on Artemisia (Compositae). ibid. 10: 295-
Goldberg, A., 1986. Classification, Evolution and Phylogeny of 314.
410 References
__ 1975. Studies on boreal Agromyzidae (Diptera). IX. Phyto- tik blattminierender Insekten (Minenstudien VIII). Z. angew.
myza miners on Boraginaceae in North America. ibid. 11: Ent. 13: 156-198.
129-142. __ 1927c. Eine neue ga1lenerzeugende Agromyzide (Dipt.).
__ 1976a. Studies on boreal Agromyzidae (Diptera). X. Phy- (Melanagromyza cecidogena sp.n.). Z. wiss. InsektBiol. 22:
tomyza miners on Crassulaceae. ibid. 12: 203-210. 314-316.
__ 1976b. Studies on borea1 Agromyzidae (Diptera). XI. __ 1928. Beitrage zur Kenntnis der Oekologie und Systematik
Chromatomyia miners on Elaeagnaceae. ibid. 12: 211-216. blattminierender Insekten (Minenstudien IX.). Zool. Jb. Syst.
__ 1976c. Studies on boreal Agromyzidae (Diptera). XII. Phy- 55: 535-588.
tomyza and Chromatomyia miners on Astereae (Compositae). __ 1930. Beitrage zur Kenntnis der Oekologie und Systematik
ibid. 12: 239-278. blattminierender Insekten (Minenstudien X). Z. angew. Ent.
__ 1977. Studies on boroeal Agromyzidae (Diptera). XIII. 17: 431-471.
some Phytomyza and Chromatomyia miners on Cichoreae - - 1931a. Minenstudien XI. Z. wiss.InsektBiol. 26: 93-108,
(Compositae). ibid. 13: 327-345. 157-182.
__ 1980. Studies on boreal Agromyzidae (Diptera). XN. __ 1931b. Minenstudien XII. Z. PflKrankh. 42: 529-551.
Chromatomyia miners on Monocotyledones. Entomologica __ 1931c. Eine neue Agromyzide aus Carex-Minen aus der
Scand. Supp!. 13: 1-61. Tschechoslowakei (Dipt) Dizygomyza (Poemyza) staryi sp.n.
Groschke, F., 1954. Miszellen iiber Blattminen und Blattmi- Mitt. dt. ent. Ges. 2(6): 93-96.
nierer. 1. Dt. ent. Z. (N.F.) 1: 138-156. __ 1932. Minenstudien XI. Z. wiss. InsektBiol. 26(7/10):
__ 1955. Miszellen iiber Blattminen und -minierer II. ibid 2: 157-182.
171-181. __ 1933. Neue Agromyziden (Dipt.). Konowia 12: 33-40.
__ 1957. Miszellen iiber Blattminen und -minierer III. ibid. 4: __ 1935. Die Blattminen Mittel- und Nordeuropas, Lief. 1: pp.
113-134. 1-112. Feller, Neubrandenburg.
Grossenbacher, J. G., 1910. Medullary spots: a contribution to __ 1936. Blattminen von Spanien, Eos, Madr.11: 331-384.
the life history of some cambium miners. Tech. Bull. N. Y. __ 1937. Agromyziden-Nachlese (Dipt.) Neue Liriomyza-
Agric. Exp. Stn 15: 49-65. Geneva, N.Y. und Phytomyza Arten I. Dt. ent. Zt. 1936: 73-80.
Hackman, W. and Viiisanen, R, 1985. The evolution and phylo- Hering, E. M., 1940, Melanagromyza coffeae sp.n., a leaf-miner
genetic significance of costal chaetotaxy in the Diptera. Ann. ofthe coffee-plant. Ann. Mag. nat. Hist. 5(11): 471-472.
Zool. Fennici 22: 169-203. __ 1941. Minenstudien 16. Dt. ent. Z. 1941: 10-23.
Hardy, D. E. and Delfinado, M. D., 1983. Family Agromyzidae, __ 1942. Ein neuer Spargelfeind. Z. Pfl. Krankheit. Pfl. Schutz
in: Insects of Hawaii 13: 190-222. Honolulu. 52: 529-533.
Hardy, J., 1849. On the primrose leaf-miner; with notice of a __ 1943. Neue pa1aearktische Agromyzidae (Dipt.). Mit
proposed new genus, and characters of three species of Dip- einem Anhang: Agromyziden-Funde in Spanien. Eos, Madr.
tera. Ann. Mag. nat. Hist. (2)4 (24): 385-392. 19: 51-62.
Harrison, R A, 1959, Acalypterate Diptera of New Zealand. __ 1944.Minenstudien 18. Mitt. dt. ent. Ges.13: 116-119.
N.z. D.S.I.R. Bul!. 128. 382 pp. __ 1949a. Die Phytomyza-Arten von Pedicularis (Dipt. Agro-
Hendel, F., 1920. Die palaarktischen Agromyziden (Prodromus myz.). Entomon 1(8): 182-3.
einer Monographie): Arch. Naturgesch. A 84(7): 109-174. __ 1949b. Biologische Unterarten bei Phytomyza ranunculi
__ 1927. Beitrage zur Systematik der Agromyziden. Zool. Schrk. (Dipt. Agromyz.). ibid. 1(9): 207-210.
Anz. 69: 248-271. __ 1949c. Neue palaearktische Agromyziden. Notul. ent. 29:
__ 1931-6. Agromyzidae: in Lindner. FUeg. Pal. Reg. 59: 18-32.
1-570. __ 1949d. Eine neue minierende Phytomyza aus Schweden
Hennig, W., 1941. Verzeichnis der Dipteren von Formosa. Ent. (Dipt. Agromyz.). Opusc. ent. 14: 159-161.
Beihefte 8: 1-239. __ 1951a. Biology of the Leaf Miners. Dr. W. Junk, 's-Graven-
__ 1958. Die Familien der Diptera Schizophora und ihre phy- hage. 420 pages.
logenetischen Verwandschaftsbeziehungen. Beitr. Ent. 8: __ 1951b. Neue pa1aarktische und nearktische Agromyziden
505-688. (Dipt.). Notul. ent. 31: 31-45.
__ 1965. Die Aca1yptratae des Baltischen Bernsteins. Stuttg. __ 1954. Die Larven der Agromyziden (Diptera). 1. Tijdschr.
Beitr. Naturk. 145: 1-215. Ent. 97: 115-136.
Henshaw, D. and Howse, D., 1989. Revised Status of Eight Spe- __ 1955. Die Liriomyza-Arten von Lactuca und Sonchus
cies of European Agromyzidae (Diptera) described both as (Dipt. Agromyz.). Dt. ent. Z. (N.F.) 2: 204-209.
Larvae and Adults, and Four New Names for Homonyms. __ 1956. Die Minierfliegen der Oberlausitz (Dipt. Agromyz.).
Entomologist's mon. Mag. 125: 83-84. 1. Nachtrag. Abh. Ber. Naturk. Mus. Forschstelle Garlitz 35:
Henshaw, D. J. de Courcy and Howse, D. in collaboration with 111-121.
M. von Tschirnhaus, in press. A List of Taxa of the Agromyzi- __ 1957a. Zur Blattminenkunde von Siidafrika. Ann. Trans.
dae (Dipt.) of the World. Harley Press, Colchester, England. Mus. 23: 59-80.
Hering, M., 1921. Minenstudien II. Dt. ent. Z. 1921: 123-147. __ 1957b. Bestimmungstabellen der Blattminen von Europa.
__ 1924a. Zur Kenntnis der Blattminenfauna des Banats. Z. Vols. 1-2, 1-1185, Vo!' 3: 1-221. Dr. W. Junk, 's-Graven-
wiss.InsektBiol. 19: 1-15; 31-41. hage.
__ 1924b. Minenstudien IV. Ztschr. Morph. Okol. Tiere 2(1/ __ 1957c. Minierfliegen in Lebermoosen (Dipt. Agromyzi-
2): 217-250. dae). Mitt. dt. ent. Ges. 16: 48-51.
__ 1925a. Minenstudien V. Z. wiss. InsektBiol. 20: 125-136; __ 1958. Stengelbewohnende Larven an Clematis recta (Dipt.
161-174. Agromyzidae). Dt. ent. Z. (N.F.) 5: 72-78.
__ 1925b. Minenstudienn VI. Ztschr. Morph. Okof. Tiere 4: __ 1960. Neue Blattminen-Studien. Dt. ent. Z. (N.F.) 7: 119-
502-539. 145.
__ 1926. Minenstudien VII. ibid. 5: 447-488. __ 1962. Neue Blattminen-Studien II. Dt. ent. Z. (N.F.) 9: 30-
__ 1927a. Die Minenfauna der Canarischen Inseln. Zool. Jb. 65.
Syst. 53: 405-486. __ 1963. Neue Blattminen-Studien III. Dt. ent. Z. (N.F.) 10:
__ 1927b. Beitriige zur Kenntnis der Oekologie und Systema- 221-250.
References 411
__ 1966. Minierfliegen in Lebermoosen II (Dipt., Agromyzi- Lovis, J., 1977. Evolutionary Patterns and Processes in Ferns.
dae). Dt. ent. Z. 13: 231-5. pp. 230-415, in Preston, R. D. and Woolhouse, H. W. Ad-
__ 1967. Blattminen der Insel Hvar. Dt. ent. Z. (N.F.) 14: vances in Botanical Research. 4. Academic Press, pp. 440.
1-80. Mabberley, D. J., 1987. The Plant-Book. Cambridge Univ.
__ 1968. Briefe iiber Blattminierer Letters on Leaf Miners. Ed. Press, 706 pp.
K. A. Spencer. Dr. W. Junk, 450 pages. Macquart, J., 1855. Plantes herbacees d'Europe et leurs insectes.
Heywood, V. H., 1971, Ed. The Biology and Chemistry of the 3. Partie. Mem. Soc. imp. Sci. Agric. Arts, Lille (1854): 157-
Umbelliferae. Academic Press, London, 1-438. 330.
Heywood, V. H., Harborne, J. B. and Turner, B. L., 1977. Eds. Malloch, J. R., 1914. Formosan Agromyzidae. Ann. Hist.-Nat.
The Biology and Chemistry of the Compositae. Two Vols. Aca- Mus. Hung. 12: 306-336.
demic Press, London, 1189 pp. __ 1934. Notes on Australian Dipteren 8. Proc. Linn. Soc.
Hollis, D., 1987. A new citrus-feeding psyllid from the Comoro N.S. W. 59: 1-8.
Islands, with a review of the Diaphorina amoena species a
Martinez, M., 1972. Contribution l'etude des Agromyzidae de
group (Homoptera). Syst. Ent. 12: 47-61. France, Bull. Soc. ent. Fr. 87: 302-308.
Holloway, J. D. and Hebert, P. D. N., 1979. Ecological and taxo- Martinez, M., Gumez J.-L. et Munier, P., 1985. Un ravageur mal
nomic trends in macrolepidoptera host plant selection. Bioi. 1. connu: la mouche mineuse du cambium des peupliers. Phy-
Linn. Soc. 11: 229-251. toma 372: 51-53.
Humphries, c., 1988. Patterns before process. Nature, Lond. Mason, B. J., 1971. Global atmospheric research programme.
333: 300-1. Nature, Lond. 233: 382-8.
Hunziger, A. T., 1979. South American Solanaceae: a synoptic Meigen, J. W., 1830. Systematische Beschreibung der bekannten
survey, pp. 49-85, in The Biology and Taxonomy of the Sola- europiiischen zweiJliigeligen lnsekten. 6: 1-40, Hamm.
naceae, Eds. Hawkes, J. G., Lester, R. N., Ske1ding, A. K. __ 1838. Systematische Beschreibung der bekannten euro-
Linn. Soc. Symposium 7. piiischen zweiJliigeligen lnsekten. 7: 1-434. Hamm.
Hutchinson, J., 1923. Contributions towards a phylogenetic Meijere, J. C. H. de, 1910. Studien tiber stidostasiatische Dip-
classification of flowering plants. I. Kew Bull. 2: 65-89. teren. IV. Die neue Dipterenfauna von Krakatau. Tijdschr.
Janchen, E., 1942. Das System der Cruciferen. Ost. bot. Z. 91: Ent. 53: 58-194.
1-28. __ 1917. Studien tiber stidostasiatische Dipteren XIII. ibid.
Jeffrey, C., 1978. Compositae, pp. 263-8, in Flowering Plants of 60: 249-251.
the World, Ed. Heywood, V. H., Oxford University Press, 335 __ 1922. Zur Kenntnis javanischer Agromyzinen. Bijd. Dierk.
pages. 22: 17-24.
__ 1986. The Senecioneae in East Tropical Africa. Kew Bull. __ 1934. Die Larven der Agromyzinen 9. Nachtrag. Tijdschr.
41: 873-943. Ent. 77: 244-290.
Jensen, U., 1968. Serologische Beitrage zur Systematik der __ 1937. Die Larven der Agromyzinen. Dritter Nachtrag.
Ranunculaceae. Bot. lb. 88: 205-310. Tijdschr. Ent. 80: 167-243.
Jermy, T., 1984. Evolution of Insect/Host Plant Relationships. __ 1940. Die Larven der Agromyzinen. Ftinfter Nachtrag.
Am. Nat. 124: 609-630. Tijdschr. Ent. 83: 160-188.
Johnson, C. D., 1981. Seed Beetle Host Specificity and the Mez, c., 1925. Der sero-diagnostische Stammbaum des Pflan-
Systematics of the Leguminosae, pp. 995-1022, in Polhill and zenreichs. In: Drei Vortrage tiber die Stammesgeschichte der
Raven (Eds.) (op. cit.). Pflanzenwelt. Naturwirtschaft u. Landwirtschaft 4(192): 1-44.
Johnson, L. A. S., 1957. A review of the family Oleaceae. Contr. Miller, J. E., 1987. Host-plant Relationships in the Papilionidae
N.s. W. Natn. Herb. 2: 395-418. (Lepidoptera): parallel cladogenesis or colonization? Cladis-
Jolivet, P., Petitpierre, E. and Hsiao, T. H., 1988. Eds. Biology of tics 3(2): 105-120.
Chrysomelidae. Series Ent. 42, 615 pp. Kluwer Academic Mitter, C. and Brooks, D. R., 1983. Phylogenetic aspects of co-
Publishers, Dordrecht, Boston, London. evolution, pp. 65-98. In Futuyama, D. J. and Slatkin, M., eds.
Jones, D. L., 1987. Encyclopaedia of Ferns. Brit. Mus. (Nat. Coevolution. Sinauer, Sunderland, Mass.
Hist.), 433 pp. Morgan, H. G., 1954. Napomyza lateralis (Fall.): a new pest of
Kaltenbach, J. H., 1856. Die deutschen Phytophagen aus der Calendulas. PI. Path. 3: 85-6.
Klasse der Insekten. Verh. naturk. Ver. Preuss. Rheinl. 13: Muller, J., 1981. Fossil pollen records of extant angiosperms.
165-256. Bot. Rev. 47: 1-142.
Kangas, E., 1935. Die Braunfleckigkeit des Birkenholz und ihre Nowakowski, J. T., 1958. Studies on mining flies (Diptera, Agro-
Urheber Dendromyza (Dizygomyza) betulae n.sp. Comm. Inst. myzidae. 1. A new species of Phytomyza Fall. in Pulsatilla
Forest. Fenn. 22: 1-31. Mill. Annis zool., Warsz. 17: 11-28.
__ 1949. Die Dendromyza-Arten von Eberesche, Espe und __ 1959. Studien tiber Minierfliegen 3. Revision der in Labia-
Salweide und ihre Biologie. ibid. 14: 106-116. ten und Boraginaceen minierenden Arten aus der Gruppe der
__ 1955. Dendromyza cerasiferae n.sp. (Dipt. Agromyzidae) Phytomyza obscura Hend. Dt. ent. Z. (N.F.) 6: 185-229.
from England. Ann. Ent. Soc. Fenn. 21(4): 165-170. __ 1960. Studien tiber Minierfliegen. 5. Eine neue Art von
Kevan, D. K. McE., 1949. Transatlantic travellers. Antenna Agromyza Fall. aus Jugoslawien. Polskie Pismo ent. 30: 367-
13(1): 12-15. 379.
Kleinschmidt, R. P., 1961. New Species of Agromyzidae from __ 1961. Studien tiber Minierfliegen (Diptera, Agromyzidae).
Queensland. Ql. l. Agric. Animal Sci. 27(1960): 341-384. 6. Eine neue Art von Liriomyza Mik nebst einem Beitrag zur
Knigge, W. and Bonneman, A., 1969. Die Markfleckigkeit des Kenntnis des miinnlichen Genitalapparats bei dieser Gattung.
Coigue-Holzes (Nothofagus dombeyi (Mirb.) (Oerst.). Holz Dt. ent. Z. (N.F.) 8: 140-146.
Roh-Werkstoff27: 224-232. __ 1962. Introduction to a Systematic Revision of the Family
Koningsberger, J. c., 1897. De dierlijke vijanden van de koffie- Agromyzidae (Diptera) with some Remarks on Host Plant
cultuur op Java. Meded. Lds. PITuin, Batavia 20: 25-26. Selection by these Flies. Annis. Zool. Warsz. 20(8): 67-183.
Kulp, L. A., 1968. The taxonomic status of the dipterous Holly __ 1963. Studien tiber Minierfliegen (Diptera, Agromyzidae).
Leaf Miners (Diptera: Agromyzidae). Univ. Maryland agric. 8. Eine neue Art von Phytomyza Fall. aus dem Pieninen-
Exp. Stn. Bull. A-ISS: 1-42. Gebirge. Annis. zool., Warsz., 21: 43-52.
412 References
__ 1964. Studien iiber Minierfliegen (Dipt. Agromyzidae) 9. __ 1955b. New Agromyzidae from Japan VIII. Kontyu 23(1/
Revision der Artengruppe Agromyza reptans Fall. - A. rUfipes 2): 16-21.
Meig. Dt. ent. Z. (N.F.) 11: 175-213. __ 1955c. A new species of Melanagromyza from Shikoku,
__ 1967. Vorlaufige Mitteilung zu einer Monographie der Japan (Diptera, Agromyzidae). Trans. Shikoku Ent. Soc. 4(5/
europiiischen Arten der Gattung Cerodontha Rond. (Diptera, 6): 87-89.
Agromyzidae). Polskie Pismo ent. 37: 633-661. __ 1955d. Six new species of the genus Phytomyza. Scient.
__ 1972. Zweite vorlaufige Mitteilung zu einer Monographie Rep. Saikyo Univ., Agric. 7: 26-34.
der europaischen Arten der Gattung Cerodontha Rond. (Dip- - - 1955e. New Agromyzidae from Japan (Diptera). XI. Two
tera, Agromyzidae). Polskie Pismo ent. 42(4): 735-765. new Leaf-miners of Angelica. Akitu 4(4): 93-97.
__ 1973. Monographie der europaischen Arten der Gattung - - 1956. A new species related to Phytomyza saxifragae
Cerodontha Rond. (Diptera, Agromyzidae). AnnIs. zool., Hering. Insecta matsum. 19(3/4): 105-8.
Warsz. 31: 1-327. __ 1957. Descriptions of Two New Phytomyza-species from
__ 1975. Drei neue Arten von Agromyzidae (Diptera) aus Mt. N achi, Japan (Diptera, Agromyzidae). Akitu 6 (4): 90-93.
dem Pieninengebirge Nationalpark. AnnIs. zool., Warsz. __ 1961. A study of the Japanese Agromyzidae (Diptera), 2.
32(18): 407-429. Pacif Insects 3: 307-472.
Papp, L., 1984. Family Agromyzidae, pp. 263-343 in Catalogue __ 1963a. Oriental Agromyzidae (Diptera) in Bishop
of Palaearctic Diptera Vol. 9. Budapest. Museum, Part 1. Pacif Insects 5: 23-50.
Parrella, M. P., Keil, C. B. and Morse, J. G., 1984. Insecticide __ 1963b. A Revision of Polynesian Agromyzidae (Diptera).
resistance in Liriomyza trifolii. Calif Agric. Jan-Feb: 22-23. Pacif Insects 5: 489-506.
Patton, W. H., 1897. A principle to observe in naming galls: two - - 1963c. Papuan Agromyzidae (Diptera). Pacif Insects 5:
new gall-making Diptera. Can. Ent. 29: 247-8. 797-835.
Paulian, R., 1953. Recherches sur les insectes d'importance bio- __ 1972. Formosan Agromyzidae (Diptera). Scient. Rep. Kyo-
logique it Madagascar. XII-XX. Mem. Inst. Sci. Madagascar to Pref Univ., Agric. 24: 43-82.
(E) 3: 1-27. - - 1977. Family Agromyzidae, pp. 243-269 in A Catalog of
Peschken, D. P. and Derby, J.-A. L., 1988. Host specificity of the Diptera of the Oriental Region Vol. III, Delfinado, M. D.
Liriomyza sonchi Hendel (Diptera: Agromyzidae), a potential and Hardy, D. E., Eds. Honolulu.
biological agent for the control of weedy sow-thistles, - - 1981a. Agromyzidae of Thailand (Diptera). Akitu n. ser.
Sonchus spp., in Canada. Can. Ent. 120: 593-600. 37:1-6.
Pichi-Sermolli, R. E. G., 1970. A provisional catalogue of the __ 1981b. Descriptions of Three New Leaf-Mining Pests
family names of living Pteridophytes. Webbia 25: 219-97. (Diptera: Agromyzidae). appl. Ent. Zool. 16(2): 149-155.
Polhi1l, R. M. and Raven, P. H., Eds., 1981. Advances in Legume Sasakawa, M. and Fukuhara, N., 1965. A new Mulberry leaf-
Systematics, Royal Botanic Gardens, Kew. Parts 1 and 2, miner in Japan (Dipt.: Agromyzidae). Jap. 1. appl. Ent. Zool. 9
1049 pp. (20): 121-124.
Raven, P. H. and Axelrod, D. I., 1974. Angiosperm Biogeogra- Sasakawa, M. and Pong, T. Y., 1988. Agromyzidae (Diptera) of
phy and past Continental Movements. AnnIs. Miss. Bot. Malaysia (Part 1). Kontyu 56(4): 853-866.
Garden 61: 539-673. Savile, D. B. D., 1975. Evolution and biogeography of Saxifraga-
Raven, P. H. and Raven, T. E., 1976. The genus Epilobium in ceae with guidance from their rust parasites. Ann. Missouri
Australasia: a systematic and evolutionary study. D.5.I.R. Res. Bot. Gard. 62: 354-361.
Bull., Auckland 216,321 pp. Seguy, E., 1951. Dipteres mineurs de Madagascar. Mem. Inst.
Riley, C. v., 1884. The Cabbage Oscinis. Ann. Rep. U.S. Dep. sci. Madagascar (A) 5: 309-21.
Agric. 1884: 322. Sehgal, V. K., 1963. A new stem mining Melanagromyza species
Robineau-Desvoidy, M., 1851. Descriptions d'Agromyzes et de from India (Diptera: Agromyzidae). Ann. Mag. nat. Hist.
Phytomyzes ecloses chez M. Ie Col. Goureau. Rev. Mag. Zool. 13(6): 541-3.
(2)3: 391-405. - - 1971a. Biology and Host-plant relationships of an oligo-
Rohdendorf-Holmanova, E. B., 1959a. Neue Minierfliegen phagous leaf miner Phytomyza matricariae Hendel (Diptera:
(Diptera, Agromyzidae) aus der USSR. Ent. Obozr. 37: 380- Agromyzidae). Quaest. ent. 7: 255-280.
391. - - 1971 b. A taxonomic survey of the Agromyzidae (Diptera)
__ 1959b. Zwei neue Agromyziden-Arten (Diptera) aus der of Alberta, Canada, with observations on host-plant relation-
Caragana arborescens Lam. Cas. Esl. Spol. ent. 52(2): 150- ships. ibid. 7: 291-405.
154. Sellier, R., 1945. Notes de faunistique armoricaine. Bull. Soc.
Rougemont, F. de, 1912. Details biologiques sur la Phytomyza Sci. Bretagne. 20: 3-4.
du Thalictrum. Mitt. schweiz. ent. Ges. 12: 82-7. __ 1947. Contribution it I'etude de Liriomyza mesnili d'Agui-
Ryden, N., 1951. Zur Kenntnis der schwedischen Blattminierer lar. Ann. des Sci. Nat. Zool., 11. ser., IX: 27-38.
XIV. Liriomyza scorzonerae nov. sp. Opusc. ent. 16: 188-190. Singh, S. and Ipe, M. I., 1971. Description of two new species of
Sasakawa, M., 1953. Descriptions and Records of Dipterous Melanagromyza Hendel (Agromyzidae: Diptera) from India.
Leaf-miners from Japan (Agromyzidae) 1. Scient. Rep. Saikyo Oriental Ins. 5(2): 223-228.
Univ., Agric. 4: 9-22. Spencer, K. A., 1954a. Notes on the British Agromyzidae (Dipt.)
__ 1954a. New and unrecorded Agromyzidae (Diptera) from - III. Entomologist's mon. Mag. 90: 217-218.
Japan IV. Kontyu 20(3/4): 55-61. - - 1954b. Agromyzidae (Dipt.) in Portugal, including a de-
__ 1954b. Neue Agromyzidae aus Japan. Trans. Shikoku Ent. scription of a new species. ibid. 90: 219-221.
Soc. 4: 35-49. - - 1956. A new species of Agromyzidae (Dipt.) from Ger-
__ 1954c. New Agromyzidae from Japan, with record of many. ibid. 92: 66-67.
several little-known species VII. Scient. Rep. Saikyu Univ., - - 1957a. Observations on the British Agromyzidae - III. En-
Agric. 6: 106-130. tomologist's Gaz. 8(1): 21-28.
__ 1954d. Neue Agromyzidae aus Japan (Diptera). III. Kon- - - 1957b. Agromyzidae (Dipt.) in the Canary Isles. Comm.
tyu 22(1/2): 23-28. BioI. 16(3): 1-3.
__ 1955a. New Agromyzidae from Japan VI. Kontyu 22(3/4): - - 1957c. A new species of Agromyzidae (Diptera) from Bri-
59-64. tain. Proc. R. ent. Soc. Lond. (B) 26(9-10): 162-4.
References 413
__ 1957d. Records of Agromyzidae in Spain including 3 spe- __ 1969a. Notes on European Agromyzidae (Diptera) - 2.
cies new to science. Eos, Madr. 33(1-4): 113-122. Beitr. Ent. 19: 5-26.
__ 1959. A synopsis of the Ethiopian Agromyzidae (Diptera). __ 1969b. The Agromyzidae of Canada and Alaska (Diptera).
Trans. R. ent. Soc. Lond. 111(10): 237-329. Mem. Ent. Soc. Canada 64: 1-311.
__ 1960a. Records of further Ethiopian Agromyzidae (Dip- __ 1971a. Notes on European Agromyzidae (Dipt.) - 3. Beitr.
tera), mainly from South Africa, including eighteen species Ent. 21: 249-65.
new to science. ibid. 112(2): 15-36. __ 1971b. Notes on a Revision of the British Agromyzidae
__ 1960b. A new species of Agromyzidae from South Africa. (Diptera). Entomologist's Gaz. 22: 141-195.
I. ent. Soc. SthAfr. 23(2): 314-16. __ 1973a. Agromyzidae (Diptera) of economic importance.
__ 1960c. Seven new species of Agromyzidae from Spain, to- Series Ent. 9. Dr. W. Junk, The Hague, pp. 1-405.
gether with other new and interesting records (Diptera). Eos, __ 1973b. A new species of Agromyzidae mining bracken
Madr. 36(3): 375-386. (Pteridium aquilinum L.). Entomologist's Gaz. 24: 315-7.
__ 1961a. Notes on the African Agromyzidae (Diptera) - 1. __ 1973c. The Agromyzidae of Venezuela. Revista Fac.
Stuttg. Beitr. Naturk. (A)46: 1-5. Agron., Maracay 7(2): 5-108.
__ 1961b. Some interesting Agromyzidae from Yugoslavia, __ 1974. Some Agromyzidae (Diptera) from Israel. Israel 1.
including the description of two new species. Entomologist Ent. 9: 141-157.
1961: 53-58. __ 1975. Agromyzidae from Ceylon. Ent. scand. Suppl. 4:
__ 1961 c. A synopsis of the Oriental Agromyzidae (Diptera). 209-220.
Trans. R. ent. Soc. Lond. 113(4): 55-100. __ 1976a. The Agromyzidae (Diptera) of Fennoscandia and
__ 1961d. Notes on the African Agromyzidae - 3 (Diptera). Denmark. Fauna ent. scand. Vol. 5, Part 1: 1-304; Part 2:
I. Ent. Soc. Sth. Afr. 24(2): 322-344. 305-606.
__ 1961e. Notes on the African Agromyzidae - 2. (Diptera). __ 1976b. The Agromyzidae of New Zealand (Insecta: Dip-
Dt. ent. Z. (N.F.) 8(5): 415-430. tera). I. Roy. Soc. N.z. 6(2): 151-211.
__ 1961f. The Agromyzidae of Madagascar (Diptera). Mem. __ 1977a. Agromyzidae in New Guinea. Paci! Insects 17 (2-
Inst. sci. Mad. (E) 12: 269-287. 3): 339-369.
__ 1962. Observations on the British Agromyzidae - 6. A __ 1977b. Notes on world Agromyzidae (Diptera), with the
new Phytoliriomyza sp. from Scotland. Entomologist's Gaz. description of 16 new species. Beitr. Ent. 27(2): 233-254.
13: 18-20. __ 1977c. A Revision of the Australian Agromyzidae (Dip-
__ 1963a. A new Phytomyza species on Plantago media L. tera). Spec. PubIs. w.A. Museum 8: 1-255. Perth.
Stuttg. Beitr. Naturk. 103: 1-5. __ 1978. Clarification of the status of Cerodontha (Dizygo-
__ 1963b. The Agromyza rufipes Mg. group of leaf-miners on myza) chaixiana (Hering, 1956). Stuttg. Beitr. Naturk. (A) 306:
Boraginaceae (Diptera). ibid. 115: 1-6. 1-3.
__ 1963c. Notes on the African Agromyzidae (Diptera) - 4. __ 1981. A Revisionary Study of the Leaf-mining Flies (Agro-
I. ent Soc. Sth. Afr. 26(1): 94-124. myzidae) of California. Univ. Calif., Div. Agric. Sciences, Spec.
__ 1963d. The Australian Agromyzidae (Diptera, Insecta). Publ. 3273: 1-489.
Rec. Aust. Mus. 25(15): 305-354. __ 1982. Agromyzidae (Diptera) in Chile. Stuttg. Beitr.
__ 1963e. Diptera: Agromyzidae in Insects of Micronesia Naturk. (A) 357: 1-55.
14(5): 135-162. Bishop Museum, Honolulu. __ 1983. Leaf Mining Agromyzidae (Diptera) in Costa Rica.
__ 1963f. A Synopsis of the Neotropical Agromyzidae. Trans. Revta BioI. Trop. 31(1): 41-67.
R. ent. Soc. Lond. 115(12): 291-389. __ 1985a. East African Agromyzidae (Diptera): further de-
__ 1964a. Notes on the African Agromyzidae (Diptera) - 5. scriptions, revisionary notes and new records. 1. nat. Hist. 19:
Dt. ent. Z. (N.F.) 11(1/2): 15-41. 969-1027.
__ 1964b. Some Agromyzidae (Diptera) from the Juan Fer- __ 1985b. The Agromyzidae (Diptera) of Colombia, including
nandez Islands. Paci! Insects 6(2): 251-255. a new species attacking potato in Bolivia. Rev. col. Ent.l0(1-
__ 1964c. A revision of the palaearctic species of the genus 2): 3-33.
Ophiomyia Braschnikov (Diptera: Agromyzidae). Beitr. Ent. __ 1986a. New records of Agromyzidae (Diptera) from
14(7/8): 773-822. southern India, with revisionary notes on other Indian species.
__ 1965a. Notes on the Agromyzidae (Diptera) of the Canary Colemania 3: 1-12.
Islands. Stuttg. Beitr. Naturk. 140: 1-5. __ 1986b. Agromyzidae (Diptera) from Thailand: new spe-
__ 1965b. Notes on the African Agromyzidae - 6. 1. ent. Soc. cies, revisionary notes and new records. Proc. Ind. Acad. Sci
Sth. Afr. 28: 233-276. (Animal Sciences) 95(5): 487-507.
__ 1966a. Notes on the Oriental Agromyzidae - 4. Stuttg. __ 1986c. A new genus of Agromyzidae (Diptera) from Aus-
Beitr. Naturk. 147: 1-15. tralia and Papua New Guinea. Entomologist's man. Mag. 122:
__ 1966b. A Revision of European species of the genera Me- 249-252.
lanagromyza Hendel and Hexomyza Enderlein, with a supple- __ 1987. The agromyzid genus Paraphytomyza Enderlein new
ment on the genus Ophiomyia Braschnikov. Beitr. Ent. 16: to the Oriental Region (Diptera). I. Nat. Hist. 21: 557-562.
3-60. __ 1988. Cecidomyiaceltis Patton, an overlooked genus in the
__ 1966c. A new Phytomyza species from California. Pan-Pa- Agromyzidae (Diptera). Entomologist's mon. Mag. 124: 63-
cific Ent. 42: 108-110. 64.
__ 1966d. A clarification of the genus Napomyza Westwood __ 1989. Family Agromyzidae, pp. 538-47. In Evenhuis,
(Diptera, Agromyzidae). Proc. R. ent. Soc. Lond. (B) 35: 29- N. L., Ed., Catalog of the Diptera of the Australasian and
40. Oceanian Regions. Bishop Museum Press, Honolulu, E. J.
__ 1966e. Notes on European Agromyzidae - 1. Beitr. Ent. Brill, Leiden. 1156 pp.
16: 285-309. Spencer, K. A. and Hawkins, B., 1984. An interesting new gall-
__ 1966f. New and Interesting Agromyzidae from Florida. forming Ophiomyia species (Diptera: Agromyzidae) on Atri-
Stuttg. Beitr. Naturk. 158: 1-20. plex (Chenopodiaceae) in southern California. Proc. Ent. Soc.
__ 1967. Some Agromyzidae (Diptera) from Morocco. Ento- Wash. 83: 664-668.
mologist's mon. Mag. 103: 126-130. Spencer, K. A. and Hill, D. S., 1976. A new species of Ophio-
414 References
myia (Diptera: Agromyzidae) causing leaf-galls on Ficus Swezey, O. H., 1913. Leaf-Miners of the Hawaiian Islands. Proc.
microcarpa in Hong Kong. Bull. Dep. Agric. Hong Kong 1 (5): Haw. Ent. Soc. 11(5): 221-7.
419-423. __ 1954. Forest Entomology in Hawaii. Spec. Pubis. Bernice
Spencer, K. A & Martinez, M., 1987. Additions and corrections P. Bishop. Mus. 44 (Agromyzidae pp. 27, 90,117,120,176).
to the Agromyzidae Section of the Catalogue of Palaearctic Takhtajan, A, 1969. Flowering Plants - Origin and Dispersal.
Diptera (Papp, 1984). Ann. Soc. ent. Fr. n.s. 23(3): 253-271. Oliver & Boyd: Edinburgh, 310 pp.
Spencer, K. A and Stegmaier, C. E., 1973. The Agromyzidae of Tamura, M., 1966. Morphology, Ecology and Phylogeny of the
Florida with a Supplement on Species from the Caribbean. Ranunculaceae VI. Science Reports, Osaka Univ. 15(1): 13-
Arthropods of Florida 7: 1-205, Gainesville. 35.
Spencer, K. A & Steyskal, G. c., 1986. Manual of the Agromyzi- __ 1967. Morphology, Ecology and Phylogeny of the Ranun-
dae (Diptera) of the United States. U.S. Dep. Agric. Agricul- culaceae VII. ibid. 16(2): 21-43.
tureHandbook No. 638. 478 pp. __ 1968. Morphology, Ecology and Phylogeny of the Ranun-
Spencer, K. A, Martinez, M. et Etienne, J., 1990. Les Agromy- culaceae VIII. ibid. 17(1): 41-56.
zides (Diptera) de Guadeloupe. Ann. Soc. ent. Fr. 126. Tavares, J. da Silva, 1901. As Zoocecidias Portugezas. Ann. Sci.
Spencer, K. A and Havranek, D., 1989. A new species of Agro- Nat. Porto VII: 75.
myzidae (Diptera) from Venezuela. Fla. Ent. 72(3): 441-4. Tavormina, S. J., 1982. Sympatric genetic divergence in the leaf-
Stary, B., 1950. Phytomyza soldanellae, n.sp., une nouvelle mining insect Liriomyza brassicae (Diptera: Agromyzidae).
espece de la famille des Agromyzidae (Dipt.). Acta Soc. ent. Evolution, N. Y. 36(3): 523-534.
Cechosl. 97: 162-5. Thorne, R F., 1983. Proposed new realignments in the angio-
Stebbins, G. L., 1977. Developmental and Comparative ana- sperms. Nord. 1. Bot. 3(1): 85-117.
tomy of the Compositae. In Heywood, V. H. and Harborne, Tschirnhaus, M. von, 1969. Zur Kenntnis der Variabilitat, Eido-
J. B., eds. The Biology and Chemistry of the Compositae. Aca- nomie und Verwandschaft bemerkenswerter Agromyzidae
demic Press, London. (Diptera). Senckenberg. bioI. 50 (3/4): 143-157.
Stegmaier, C. E., 1966. Host plants and parasites of Liriomyza __ 1971. Unbekannte Stridulationsorgane bei Dipteren und
schmidti in Florida (Diptera: Agromyzidae). Fla Ent. 49(2): ihre Bedeutung fur Taxonomie und Phylogenetik der Agro-
119-122. myziden. Beitr. Ent. 21: 551-579.
Stewart, W. N., 1983. Paleobotany and the evolution of plants. __ 1981. Die Halm- und Minierfliegen im Grenzbereich
Cambridge University Press, pp. 405. Land-Meer der Nordsee. Spixiana, Supp!. 6, pp. 1-405.
Steyskal, G. C., 1972. Two New Species of Melanagromyza Tsujita, M., 1951. A new subspecies of Agromyza nigrescens
Hendel (Diptera, Agromyzidae) that bore in Tomato Stalks in Hendel (Diptera, Agromyzidae). Sci. Bull., Fac. Agric., Kyu-
Colombia and Ecuador. l. Wash. Acad. Sci. 62(3): 265-267. shu Univ. 13,53-7.
__ 1976. A New Species of Leafmining Fly on Mealycup Sage Tullgren, A and Wahlgren, E., 1920-1922. En orienterande
(Diptera: Agromyzidae). U.S.D.A. Coop. Plant Pest Rep. handbok vid studiet om vart hands Insektfaune. Stockholm,
1(43): 820-2. pp.812.
__ 1980. Haplopeodes, a New Genus for Haplomyza of Valladares, G., 1981. Contribucion al conocimiento de las espe-
Authors (Diptera, Agromyzidae). Proc. Ent. Soc. Wash. cias de Calycomyza Hendel (Diptera, Agromyzidae), mina-
82(1): 140-151. doras de hojas en la Republica Argentina. Revta Soc. Ent.
Strobl, G., 1909. Neue i:isterreichische Muscidae Acalypterae II. Arg. 40(1-4): 221-9.
Wien. ent. Ztg 28: 283-301. __ 1982. Agromyzidae (Diptera) de la Provincia de Cordoba
Suss, H., 1979. Durch Protophytobia cupressorum gen. nov., sp. (Argentina). Doctoral Thesis, pp. 1-318. Universidad nacio-
nov. (Agromyzidae, Diptera) verursachte Markflecke in nal de Cordoba, Argentina (unpublished).
einem Holz von luniperoxylon aus dem Tertiar von Sud-Lim- __ 1984. Una nueva especie de Liriomyza Mik (Diptera,
burg (Niederlande) und der Nachweis von Markflecken in Agromyzidae) de la Argentina. Neotropica 30(84): 245-9.
einer rezenten Callitris-Art. Feddes Repertorium, Berlin __ 1986a. The genus Ophiomyia Braschnikov (Dipt., Agro-
90(3): 165-172. myzidae) in Argentina. Entomologist's mon. Mag. 122: 111-
Suss, H. and Muller-Stoll, W. R, 1975. Durch Palaeophytobia 115.
platani n.g., n.sp. (Agromyzidae, Diptera) verursachte Mark- __ 1986. Agromyzidae (Diptera) Neotropicales. Especies del
flecken im Holz fossiler Platanen aus dem ungarischen Mio- N.O. Argentina, pp. 1-34 in Annual Report for CONICET,
zan: Wiss. Ztschr. Humboldt Univ. Berlin, Math. - Nat. R. Argentina (unpublished).
24(4): 515-519. Valley, K. R, 1982. A new Liriomyza mining leaflets of Black
__ 1980. Das fossile Holz Pruninium gummosum Platen Locust (Diptera: Agromyzidae). Proc. Ent. Soc. Wash. 84(4):
emend. Suss u. Muller-Stoll aus dem Yellowstone National- 781-5.
park und sein Parasit Palaeophytobia prunorum sp. nov. nebst Vaughan, J. G., McCleod, A J. and Jones, B. M. G., 1976. Eds.
Bemerkungen uber Markflecke. In Vent, W., 100 lahre Arbor- The Biology and Chemistry of the Cruciferae. Academic Press,
etum (1879-1979): 343-364. London, 355 pp.
Suss, L., 1977-78. Osservazioni su alcuni Agromyzidi poco noti Verdcourt, B., 1958. Remarks on the Classification of the Rubi-
o nuovi per l'entomofauna italiana. Boll. Zool. agr. Bachic. aceae. Bull. lard. Bot. Brux. 28: 209-90.
Ser.II, 14: 145-164. Walker, F., 1849. List of the specimens of dipterous insects in the
__ 1986-87. Liriomyza cyclaminis, nuova specie fillomina- collection of the British Museum (part). 4. pp. 689-1172.
trice dei Cyclamen. Boll. Zool. agr. Bachic. Ser. II. 19: 23-29. British Museum, London.
__ 1988. Foglie di orzo e petali di gerbera nuovi substrati ali- Watt, M., 1924. The Leaf-mining Insects of New Zealand: con-
mentari di Liriomyza trifolii (Burgess). 1nformatore Fitopato- tinued, and Gracilaria selonitis Meyr. (Lepidoptera). Trans.
logico 6: 45-47. Proc. N.Z Inst. 55: 674-687.
Swain, T. and Cooper-Driver, G., 1973. Biochemical systematics Wenn, c., 1985. Notes on three new stem miners. Entomotaxo-
in the Filicopsida, pp. 111-134, in The Phylogeny and classifi- nomia 7(2): 109-113.
cation of the ferns, Ed. Jermy, Crabbe and Thomas. Bot. l. Wiesman, R, 1961. Phytomyza lateralis Fal!., ein weinig beach-
Linn. Soc. 67, Supp!. 1. teter Mi:ihren- und Karottenschadling. Mitt. ent. Ges. Basel
11: 39-62.
References 415
Willis, J. c., 1957. A Dictionary of the Flowering Plants and (Diptera, Agromyzidae). II. Subgenus Poemyza muscina
Ferns. Sixth Edition. Cambridge University Press. group (in Russian). in: Dvukryl fauny SSSR i ikh roe vecosiste-
__ 1985. A Dictionary of the Flowering Plants and Ferns. makh: 45-52, Leningrad.
Eighth Edition. Cambridge University Press. __ 1984b. New Species of Selachops Wahlberg and Metopo-
Zlobin, V. V., 1979. A new species of mining fly Xeniomyza in- myza Enderlein (Diptera: Agromyzidae) from the Far East (in
termedia sp.n. (Diptera, Agromyzid) from Kyzykum (in Rus- Russian). In: Diptera (Insecta) their Systematics, Geographic
sian). Trydy vses. ent. obshch. 61: 204-6. Distribution and Ecology: 45-52 (1983). Zoological Institute,
__ 1984a. Mining flies of the genus Cerodontha Rondani Leningrad.
INDEX
Page numbers with host tables are shown in bold, those with illustrations by an asterisk.
For genera, oligophagous, polyphagous and widespread species with multiple entries, the page
number with the main discussion is underlined. Junior synonyms, nomina nuda and
misidentifications are indented. Host plants are indexed to genera only. Place names of
particular interest or relevant to dispersal only are indexed.
A 293
abdita, Phytomyza 205, 206*, 207, 401 aequalis, Japanagromyza Ill, 122*, 123
abdominalis, Phytomyza 21, 27, 28* Aeschynomene 111,119,123
Abelia 238, 242 Aeschynomeneae Ill, 123
abeliae, Phytomyza 238, 241 *, 242 Acthionema 85
Abelmoschus 75, 92 Aelhusa 159
aberrans, Ophiomyia 205, 209 afTinis, Phytomyza 222, 226*, 227
abiens, Agromyza 196, 197*, 198, 199,400 affinis, Phytomyza (misidentification) 255
abiens group, Agromyza 299 Afrotropical Region 108, 361, 365, 369
Abies 15 Agastache 204
abnormalis, Amauromyza 63, 64, 65*, 70, 204 Ageratina 287, 289, 290*
abutilivora,Ophiomyia 74*,75,91, 92 Ageratum 263, 285, 287, 289
Abutilon 75, 92 Agoseris 259, 267
abutilonicaulis, Hexomyza 72, 73*, 92 Agrimonia 102, 104, 385
abutilonis, Agromyza 74*, 75, 91, 92, 323, 364 Agromyza 50, 91, 95, 100, 108, 127, 158, 177, 196, 229,
Acacia 110,113,115,139 242, 321, 323, 344, 353, 363, 389, ill
acaciae, Melanagromyza 110, 114*, 115, 139 Agrol11yza sp. 102, 104, 105*, 147, 155, 204, 205, 356,
Acacieae 110, 115 357, 363
Acalypha 147, 148, 155 Agrol11yzid sp. 340, 341
acanthacearum, Pseudonapomyza 213,215 agrol11yzina, Phytomyza 143, 144*, 154
Acanthoic1eae 213, 215 Agromyzinae 390
acanthopanacis, Phytomyza 157, 159, 161* Agromyzoinea 388
Acanthopanax 157, 159 Agropyron 356, 367
Acer 150, 156, 178 Agrostel11ma 63
aceris, Phytobia (= Phb. setosa) 150, 178 Agrostis 352, 354, 377
Achillea 253, 258, 263, 299, 301, 303, 306, 307, 394 Ainsliaca 258, 258
achilleana, Melanagromyza 299, 301 Aira 354
achilleae, Ophiomyia (= O. curvipalpis) 299 airae, Agromyza (= A. lucida) 359
achilleae, Phytomyza 301, 304*, 305, 307 aizoon, Chromatomyia 99, 100*, 101
achille ana , Phytomyza 301, 307 Ajuga 204, 205, 207, 211, 401
achilleanella, Napomyza nom.nud 301, 303, 307 akbari, Ophiomyia ISO, 185
Achyranthes 63,67 Akebia 22, 47
Acmena 143, 154 akebiae, Phytol11yza 22,47
aconitella, Phytomyza 21,26*, 27 Alangium 143, 154
aconiti, Phytomyza 21, 26*, 27, 50 alaskan<l, Phytol11yza 301, 304*, 306
aconitophila, Phytomyza 21, 26*, 27 alhiceps group, Phytomyza 125, 140, 158, ill, 165,
Aconitum 21, 27 178, 231, 249,255, 267, 271, 283, 284, 297, 305, 307,
Acronychia 152, 156 308,315,319
Actaea 21,23 albiceps, Phytol11yza 163, 251, 255, 256*, 305
actaeae, Me1anagromyza 21 albicula, Hexomyza (= Hex. winnemanae) 81
actaeae, Phytomyza 21, 23, 26* albifrons, Phytomyza 47, 49*
Actinodaphnc 18, 19 alhimargo, Phytomyza 21, 27, 28*
Acyrthosiphon 145 alhipenllis. Agromyza 354, 355, 356, 357, 361, 364
Adenocaulon 257, 258 albipennis, Napomyza 21, 38, 39*
Adenophora 231 albisquama, Me1anagromyza lIl, 116*,117
Adenostyles 293, 297 albitarsis, Agromyza 81, 82*, 93, 100, 178
adcnostylis, Phytomyza 293, 297 Albizia 110, 115, 117
adjuncta, Phytomyza 160, 169, 170* albocilia, Melanagromyza 191 *, 193, 194
Adonidinae 21, 35 Alchemilla 102, 104
Adonis 21,23, 35, 86 Aldabra 135
adsurgenis, Melanagromyza 112, 124*, 125 Alismatales 337
Aegilops 356 Alismatidae 322, 337
Aegopodium 159, 167 Alismatiflorae 337, 341, 389
aenea, Me1anagromyza 54, 56*, 57 Allamanda ISO, 180
aencoventris, Melanagromyza 250*,251, 251, 253,29;, allecta, Calycol11yza 318*,319,320
418
costata, Napomyza 21,40*,41,50, 402 cymbonoti, Ophiomyia 249, 249, 250*, 261
Cotinus ISO, 156 Cymbonotus 249, 249, 261
Cotoneaster 102, 106 Cynanchum 180, 183
Cotu1a 263, 301 Cynara 251, 253, 263
Cotyledon 97, 101 Cynareae 247, 248
cotyledonus, Melanagromyza 97, 98*, 101 Cynodon 356, 375, 376*
Crambe 85 Cynodonteae 356
Craspedia 271, 273 Cynoglosseae 196, 199
craspediae, Liriomyza 232, 271. 272*, 273 cynoglossi, Calycomyza 197*, 198, 199
crassiseta, Cerodontha (Dizygomyza) 355, 371, 372*, Cynoglossum 198, 199
373 Cynometra 110, 113
crassiseta, Phytomyza 222, 224, 225* Cynosurus 355
Crassocephalum 277, 278*, 293, 399 cyparissiae, Liriomyza (= L. balcanica) 147
Crataegus 102, 104, 106, 385 Cyperales 322, 342, 344, 351
Crawfurdia (= Gentiana) 180, 183 Cyperoideae 347, 351
crawfurdiae, Chroma tomyia 180, 181 *, 183 Cyperus 349, 351, 352
crepidis, Phytomyza 259, 267, 268* Cyphioideae 229
Crepis 253,258,259,261,265,267,394 Cyrtandra 214,218,219*
Crescen tieae 213, 218 Cyrtandroideae 218
criptica, Liriomyza (nom. nud.) 79, 93, Ill, 117, cyrtanthi, Tropicomyia 325, 335
189*, 190, 192,382,383,384, 384, 385 Cyrtanthus 325, 335
Cronquist ix cyrtorchidis, Me1anagromyza 327, 333, 334*, 336
Crotalaria 113, 119, 121, 135, 139, 395,405 Cyrtorchis 327, 333, 336
crotalariae, Tropicomyia 113, 123, 135, 136*; 139,405 cytisi, Phytomyza II3, 138*, 139
crotalariana, Melanagromyza 113, 135, 136*, 139 Cytisus 113, 139,251. 404
Crotalarieae 113, 135 czernyi, Phytomyza (misident.) 403
crotalariella,Ophiomyia 113, 135, 136*, 395
crotalarivora, Calycomyza 113, 135, 136* D
Croton 147, 155 Dactylis 355, 365, 369
crotonella, Tropicomyia 147, 149*, 155,405 Dahlia 308, 309
cruciata, Liriomyza 354, 374*, 375 Da1bergia 110, 115
Crucifera (nom. nud.) 83 Dalbergieae 110, liS
crucifericola, Liriomyza (= L. xanthocera) 75, 85, Dahlgren 322
86, 194, 205, 382, 383, 384, 384, 399 dalmatiensis, Phytomyza 21, 32*, 33
Crymodes 41 Daphne 141, 154
Cryptolepis 180, 183 Datura 187, 190, 192
Cryptostegia 180, 183 Datureae 190, 192
Cryptostemma 249 Dauceae 160, 177
Cucumis 78, 93 dauci, Phytomyza (= P. ferulae) 175,177
Cucurbita 78, 93 Daucus 160, 163, 165, 169, 175, 177, 265
Cudrania 53, 61 davisii, Phytomyza 21, 43, 45*
cunctanoides, Melanagromyza 270, 273, 299, 301, 311, debilis, Liriomyza 313, 316
316,317,320 Dehaasia 18, 19
cunctans, Me1anagromyza 112, 119, 125, 126* deirdreae, Chromatomyia 99, 100*, 101
cunctata, Ophiomyia 244, 259, 260*, 261 Delilia 315, 316
Cunninghamia 15 Delphineae 21, 27
cupressorum, Protophytobia 15 delphinii, Ophiomyia 21, 23
Cupressus IS Delphinium 21, 23, 27
curvipalpis, Ophiomyia 133, 251, 253, 259, 261, 274, delphinivora, Phytomyza 21, 26*, 27
299, 301 demeijerei, Agromyza II 3, 127, 138*, 139
curvipalpis var. texana, Ophiomyia 285 demeijerci, Liriomyza 300*, 301, 301
Cuscuta 193, 194 demissa, Phytomyza 284, 301, 306*
cuscutae, Melanagromyza 193, 194, j 95* Dennstaedtiales 5, 9
Cussonia 157, 159 denticornis, Cerodontha (Cerodontha) 354, 355, 356,
cussoniae, Melanagromyza 157, 159, 161 * 364, 365, 366*
Cyathea 5, 11, 391 denticornis group, Cerodontha 365
cyatheae, Phytoliriomyza 5, 10*, 11, 14, 338 Deschampsia 352, 354, 367, 369, 373, 377, 379, 386
Cyatheales 5, 11 deschampsiae, Cerodontha (Poemyza) (= Cer. (Po.)
Cyathocaulis 11 imbuta) 367
Cyclamen 89, 94 Descurainia 83, 85
cyclaminis, Liriomyza 89, 90*, 94 deserta, Cecidomyiaceltis 53, 54
Cydonia 102, 106, 107 Desmazeria 355
cydoniae, Paraphytomyza 102, 107, 125 Desmodieae Ill, 117
cygnicollina, Chromatomyia 345, 351 Desmodium Ill, 117, 139
Cymbideae 336 desmodivora, Japanagromyza 111,116*,119,123,139
Cymbidium 334, 336 despinosa, Phytomyza 274, 282*, 283
425
Endodesmia 72, 91 Eupatorium 270, 274, 285, 287, 289, 303, 311, 384
Enicostema 180, 185 eupatorivora, Calycomyza 287, 288*, 289
Enterolobium 1I0, liS Euphorbia 147, 148, 155
Epaltes 271, 273, 277, 399 Euphorbiales 95, 145, 155
Eperua llO, 113 Euphrasia 221,222,227
Ephedra 15 eurhabdus, Haplopeodes 187, 192
Epidendroideae 336 Eustrephus 324, 335
Epilobium 141 evanescens, Napomyza 21, 38, 39*, 50, 402
Equisetales 5, 6 Evax 273
Equisetatae (= Sphenopsida) 5, 389 eximia, Napomyza 21, 38, 40*, 50, 402
equiseti, Liriomyza 5, 6, 7* exquisita, Calycomyza 274, 280*, 281
Equisetites 6
Equisetophyta 4, 5 F
Equisetum 5, 6 fabae, Melanagromyza 112, 126*, 127
Eragrostideae 356 Fabales 107
Eragrostis 356 facialis, Phytomyza 159, 167, 168*
Erechtites 275, 293, 295, 313 Fagales 52, 54, 59
erechtitidis, Melanagromyza 293, 295 Fagus 59
Erica 88, 93 Falkland Islands 224, 365
Ericales 88, 93 fallaciosa, Phytomyza 21, 27, 43, 45*, 178
Ericoideae 93 fallaciosa group, Phytomyza 50
Erigeron 263, 274, 281, 283 fall ax, Paraphytomyza 238, 242
erigeronis, Me1anagromyza 274, 276* farfarae, Phytomyza 293, 298*, 299, 306
erigerontophaga, Chromatomyia 274, 281, 282* farfarella, Chromatomyia 259, 265
erigerontophaga group, Chromatomyia 281 fasciata, Cerodontha (Dizygomyza) 355, 371, 372*, 373
erigerophila, Phytomyza 271, 274, 283, 284* fasciola, Liriomyza (= L. pusilla) 274, 275, 277, 398
eriodictyi, Chromatomyia 194, 195*, 224 felleri, Agromyza 1I2, 127, 129*
Eriodictyon 194 felti, Phytoliriomyza 6, 12*
eriolepidis, Me1anagromyza 251, 251, 253 ferns 4,8, 13,261, 380, 393
eriophori, Cerodontha (Dizygomyza) 349, 351 ferox,Ophiomyia 186*, 187, 192
Eriophorum 345, 347, 349, 351 ferruginosa, Agromyza 198, 199, 200*
Eritricheae 199 Ferula 160, 163, 173
erodii, Agromyza 152, 157 ferulae, Melanagromyza 160, 162*, 163
Erodium 152, 157 ferulae, Phytomyza 159, 160, 169, 173, 174*
Eruca 85 ferulae ssp. scaligeriae, Phytomyza, 159, 173, 175
Erucastrum 85 ferulivora, Phytomyza 160, 174*, 175
erucifolii, Liriomyza 293, 294*, 295 Festuca 355, 371
Eryngium 159, 160 lici, Ophiomyia 53, 57
Erysimum 85 Ficus 53, 57
Erythrina 111, 119, 404 Fiji 78
erythrinae, Melanagromyza Ill, 118*, 119, 404 Filago 271, 273, 400
Erythrininae Ill, 119 Filicopsida 4, 5
erythrocephala, Agromyza 112, 127, 128* Filipendula 102, 104
Erythroxylon 148, 155 fllipendulae, Agromyza 102, 104, 105*
esakii, Phytomyza 21, 33, 36* f11llbriata, Phytomyza 153, 157
Escallonia 97, 101 Fimbristylis 347, 351
Esenbeckia 152, 156 Flacourtia 77, 92
esulae, Liriomyza (= L. balcanica) 147 flacourtiae, Tropicomyia 62, 64, 74*, 75, 77, 78, 83, 86,
eucalypti, Japanagromyza 141, 142*, 154 89, 92, 93, 94, 97, 101, 1I0, 113, 152, 156, 180, 183,
Eucalyptus 141, 143, 154 193, 194,200,201,221, 235, 236, 324, 335, 337, 341,
eucaricis, Cerodontha (Butomomyza) 347, 348*, 351, 382, 383, 384, 384, 395
380 flagellae, Liriomyza (nom. nud.) 63, 65*, 67, 70
Euce10cerinae (= Encelocerinae) 390 flaveola, Liriomyza 354, 355, 356, 372, *, 373
eucodonus, Ophiomyia 229, 230*, 231 Flaveria 313, 316, 320
Eugenia 143, 154 Flaveriinae, see Heliantheae - Flaveriinae
Eulophia 323, 334, 336, 395 flaviceps, Agromyza 53, 55, 56*
eulophiae, Tropicomyia 334*, 336, 336, 395 flavicornis, Phytomyza 54, 58, 60*
Euonymus 155 flavifrons, Amauromyza 63,69,70,204,231,233, 385
Eupatoriade1phus 287, 289 flavifrons, Cerodontha (Cerodontha) 354, 355, 365,
eupatoriana, Liriomyza 287, 288*, 289 366*
Eupatorieae 248, 285, 289 flavinotum, Calycomyza 251, 253, 287, 288*, 289
eupatoriella, Melanagromyza 285, 286*, 289 flavinotus, Haplopeodes 190, 192
eupatorii, Liriomyza 205, 209, 259, 263, 274, 275, 277, flaviorbitalis, Cerodontha (Poemyza) (= Cer. (Po.)
287, 288*, 289, 291, 386 oryzivora) 371, 398
eupatorii, Melanagromyza 270, 273, 285, 286*, 289 flavipennis, Agromyza 204, 205, 206*
eupatorii, Phytomyza 289, 290* flavocentralis, Liriomyza 86, 222, 224, 225*
427
Juniperoxylon 15 Laportea 54
Juniperus 15 lappae, Me1anagromyza 251,251,252*,253
Jurinea 251 lappae, Phytomyza 251, 254*, 255, 257, 258, 293
Justicia 213, 215, 397 lappivora, Ophiomyia 251, 253
justiciae, Pseudonapomyza 213,217*, 217, 397 lappivora, Phytomyza 251, 254*, 255
lapplandica, Cerodontha (Poemyza) 354, 355, 367,
K 368*
Kadsura 19, 20 Lappula 199
Kalanchoe 99, 101 Lapsana 259,261,263,267
kalanchoes, Tropicomyia 97, 98*, 101 larissa, Liriomyza (= L. virgula) 7
Kallstroemia 152, 157 larval chromosomes 349
kalshoveni, Japanagromyza 146*, 147, 155 Laser 160, 175
kaltenbachi, Phytomyza 21.23, 32*, 33, 50 Laserpitieae 160, 177
katoi, Liriomyza 301, 303 Laserpitium 160, 163, 173, 177
kefi, Haplopeodes 187, 192 Lasia 337, 341
kenti, Liriomyza (= L. equiseti) 6 Lasiodiscus 148, 155
kenyensis, Melanagromyza 146*, 147, 155,319 lateralis, Cerodontha (Poemyza) 356, 367, 368*
kevani, Phytomyza (ssp. of P. tussilaginis) 293, 298, lateralis, Napomyza 58, 150, 155,207,231,251, 253,
299 259, 265, 274, 281, 290, 291, 293.295, 301, 302*,
kibunensis, Phytomyza 160, 172*, 173 305, 307, 386
kiefferi, Agromyza (= Hexomyza sp.) 404 lateralis group, Napomyza 35
Kigelia 213,218 lathyri, Agromyza 112, 127, 129*
kisakai, Phytomyza 89, 90*, 94 lathyri, Liriomyza 112, 130*, 131, 140
Kitaibelia 92 lathyroides, Amauromyza 112, 130*, 131, 140
Kleinia 293, 295 Lathyrus 112, 123, 127, 131, 140
kleiniae, Liriomyza 293, 294*, 295 latifolii, Phytomyza 160, 176*, 177
Kleinschmidtimyia 117, 329, 389, 391 latigenis, Liriomyza 293, 294*, 295
kluanensis, Chromatomyia 245*, 246 latipalpis, Liriomyza 339*, 340, 341
Knautia 242, 244 Launaea 259, 265
knautiae, Ophiomyia (= O. longilingua) 242 Laurales 17, 18
Knowltonia 21.29 Laurasia 52
knowltoniae, Phytomyza 21.29, 31 * Laurus 18, 19
Koelreuteria 152, 156 Lavandula 205, 211
koizumii, Me1anagromyza Ill, 119, 123 Lavatera 75, 92
krygeri, Phytomyza 22. 43, 48* Lecythidales 71,75,92
kuscheli, Phytoliriomyza 5, 10*, 338 Lecythis 75, 92
kwansonis, Ophiomyia 328*, 329, 335 legitima, Ophiomyia 201
kyfThusana, Phytomyza 271, 272*, 273, 400 Legousia 231
Kyllinga 351 Leguminales 107
Lens 112, 131, 385
L leonotidis, Melanagromyza 204, 205, 206*
labi, Tropicomyia Ill, 123 Leonotis 204, 205
labiatarum, Amauromyza 204, 205, 206*, 211, 222, 231 Leontodon 259, 261, 265, 267
labiatarum, Ophiomyia 204, 205, Leontopodium 263, 271, 273,400
Lablab Ill, 121, 123 leon uri, Amauromyza 204, 205, 206*, 211, 249, 321
Labordia ISO, 185 Leonurus 204,205,211
laburnifoliae, Tropicomyia 113, 135, 139, 405 Lepechinia 205, 209
Laburnum 113, 127, 139 lepidii, Liriomyza 85, 86, 87*, 91
Lactuca 259, 263, 265, 267, 313, 399 Lepidium 83, 85, 86
lactuca, Chromatomyia 259, 265 266* Lepidoptera vii
Lactuceae 247,248,258,259, 321, 390, 395 butterflies 381
Lactucoideae 258, 285 cardui, Cynthia 381
LaeJia 333, 336 claudia, Euptoieta 381
Lagerstroemia 143, 153 lavinia, Precis 381
Lagurus 354 moths 381
Lallemantia 204, 205, 211 Nymphalidae 381
Lamarkia 355 Papilionidae 388
Lamiales 194 leptargyreae, Chromatomyia 141, 153
lamii, Amauromyza 204,205, 249, 320 Leptureae 357
Lamioideae 203,205 Lepturus 357
Lamium 204, 205, 211 leucanthemi, Phytomyza 301, 304*, 306
lanati, Phytomyza 160, 175, 176* Leucanthemum 270, 285, 301, 305, 306
Lannea 150, 156 Levisticum 160, 175
Lantana 201, 202* Liabeae 247
lantanae, Calycomyza 201, 209 libanotidis, Phytomyza 160, 169, 171 *
lantanae, Ophiomyia 201 Libanotis (= Seseli) 160, 169
431
Ligularia 258, 293 lupiniphaga, Liriomyza 113, 135, 137*, 139, 140
ligusticifoliae, Phytomyza 21, 33, 36* lupinivora, Phytomyza 113, 139
Ligusticum 169 Lupinus 113, 135, 140
Ligustrum 214 lutea, Liriomyza 160, 231, 233
Liliacidites 322 luteiceps, Amauromyza 62, 64, 204
Lilirues 323, 324, 329, 335 luteiceps, Phytomyza (= P. flavicomis) 59, 400
Liliidae 322, 323, 344 luteifrons, Agromyza 113, 139
Liliiflorae 322, 323, 335, 389 1uteitarsis, Agromyza 356, 360*, 361
Liliopsida 322 1uteoscutellata, Paraphytomyza 238,239*, 240
Lilium 329, 335 Luzula 344, 345, 351
limata, Melanagromyza 160, 162*, 163 1uzu1ae, Cerodontha (Dizygomyza) 345, 346*, 351
Limonia 152, 156 1uzu1ae, Chromatomyia 345, 346*, 351
Linrues 148, 155 1uzulae superspecies, Chromatomyia 345, 377
Linaria 222 1yalli, Napomyza 21, 40*, 41, 402
Lindera 18, 19 Lychnis 63
Lindheimera 313, 316 Lycieae 187, 190, 192
Linnaea 238, 242 Lycium 190, 192
linnaeae, Chromatomyia 238, 241 *, 242 lycivora, Hap1opeodes (nom. nud.) 190, 192
Linum 150, 155 Lycopersicon 187, 192
Lipochaeta 316 lycopi, Phytomyza 205, 207
Lippia 201, 215 Lycopsis 198, 199
lippiae, Ophiomyia 201, 202* Lycopus 204,205,207,209,211
Liquidambar 53, 54 1ygophaga, Agromyza (= A. rubitarsis) 81
Liriomyza 4, 8, 17, 18, 20, 50, 77, 108, 139, 158, 177, Lysimachia 91
235,247,299,321, 323, 342, 373, 384, 389, 390, ~
Liriomyza sp. 93,101,190, 192,201,214,217,218, M
229, 230*, 231, 257, 258, 273, 279*, 287, 289, 308, Maba 89,94
309,310*,313,315,316,319,320,329,335,337, Macaranga 145, 148, 155
341, 343*, 344, 351 macminni, Cerodontha (Poemyza) 347, 348*, 351, 371,
Lissochilus (= Eu1ophia) 336, 395 380
Lithospermeae 198, 199 macu1osa, Amauromyza 131, 140, 251, 259, 261, 271,
lithospemli, Agromyza 199, 199 273, 274,275, 278*, 287, 289, 290, 291, 293,
lithospermi, Phytomyza 198, 199 295,301, 308, 309, 313, 316, 317, 320
Lithospermum 196, 199, 199 Madia 309
Litsea 18, 19 Madieae 248, 309
Loasa 79,93 madizina, Ophiomyia (= O. nasuta) 261
Lobelia 229 madrilena, Amauromyza 204, 205, 206*, 211
Lobelioideae 229 Maesa 89, 94
loewii, Phytomyza 21, 33, 34* magnicornis, Cerodontha (Dizygomyza) 351, 352, 380
lolii, Liriomyza 355, 374*, 375 Magnolia 17, 18
Lolium 355, 375 Magno1iales 17, 18, 71
lomatii, Melanagromyza 160, 163, 164*, 390 Magnoliidae 17, 18, 52, 95, 382, 389
Lomatium 160, 163, 390 Magnoliillorae 323
longicauda, Calycomyza 74*, 75, 92 Magnoliophyta 382
longilingua, Ophiomyia 242, 243*, 244 Maianthemum 330*, 331, 335
longipennis, Cerodontha (lcteromyza) 345, 346*, 351 Majorana 205, 211
Lonicera 237, 238, 238 majuscula, Calycomyza 293, 294*, 295
lonicerae, Chromatomyia 238,240,241* malayensis, Melanagromyza 193
lonicerina, Paraphytomyza 238, 239*, 240, 242 ma1ayensis, Pseudonapomyza 356, 375, 376*
loniceroides, Paraphytomyza 238, 240, 241 * Ma1colmia 85
lopesi, Haplopeodes 187, 192 male genitalia x, 251
loranthi, Japanagromyza 143, 154 mallochi, Melanagromyza 285, 286*, 289
Loranthus 143, 154 mallorcensis, Phytomyza (= Napomyza c1ematidis)
Lord Howe Island 277, 281, 390, 399 29
Loteae 112, 125 Mallotus 147, 155
Lotus 112, 125 Ma10ideae 100, 102, 104, 106
lucens, Paraphytomyza 235, 236 Malopeae 92
lucida, Agromyza 354, 355, 359, 361, 364 Malus 102, 106
luctuosa, Cerodontha (Dizygomyza) 333, 345, 346*, 351 Malva 75,92
Luehea 72, 91 ma1vaceivora, Agromyza 75, 92
Luffa 79,93, 190, 385 ma1vae, Calycomyza 74*, 75, 92, 113, 140
lugentis, Phytomyza 293, 298, 299 malvae group, Calycomyza 315
lupinella, Liriomyza 113, 135, 137*, 139 Malva1es 52, 71, 72, 91, 145
lupini, Liriomyza 113, 135, 137* Malveae 92
lupini, Phytomyza 113, 137*, 139 mamonowi, Paraphytomyza 85, 86, 88
Lupininae 113, 135 Maniliot 148, 155
432
oligophaga, Melanagromyza 251, 253, 258, 259, 260*, Palaeophytobia 79, 389
267, 299, 301, 394 Palava 92
oligophagy 198, 381, 384, 386 Pallavicinia 1
01mediopsis 53 Pallavicinites 1
Omphalodes 198, 199 Pallenis 271, 273
Onobrychis 112, 125 palliatus, Haplopeodes 63
ononidis, Ophiomyia 112, 132*, 133 pallidifrons, Agromyza 357, 361, 362*, 363
Ononis 112, 131, 133 pampeana, Phytomyza 21, 35, 37*
Onopordum 251, 255 panacis, Melanagromyza 157, 159, 161 *
Onoseris 258 Panax 157, 159
opacae, Phytomyza 145, 154 Paniceae 357
opacella, Chromatomyia 355, 356, 378*, 379 panici, Agromyza 357, 362*, 363
opacella superspecies, Chromatomyia 379 Panicoideae 352, 357, 364
Ophiomyia 20, 190, 242, 323, 327, 384, 389, 390, 121. paniculatae, Phytomyza 33, 36*
Ophiomyia group 1 261 Panicum 357, 361, 363, 375
Ophiomyia group 2 261 papae, Phytoliriomyza 187, 188*, 192
Ophiomyia sp. 22,63,69, 85,86,271, 274, 275, Papaver 22, 47, 50
289, 308, 309 Papaverales 22,47, 50
Oplismenus 357, 362*, 363 Papilionoideae 107, 110, 139
Opostegidae 59 papuensis, Agromyza 357, 363
Opuntia 62, 64 Parabenzoin (= Lindera) 18
orbiculata, Ophiomyia 112, 127, 128*, 140, 163, 178 paraciliata, Chromatomyia 301, 305
orbitalis, Paraphytomyza 238, 239*, 240 Paraphytomyza 81, 91, 100,233,235, 237, 238, 242,
orbitona, Cerodontha (Poemyza) 354, 357, 369, 370* 389, 392
orbona, Liriomyza 354,355, 372*, 373 Parietaria 54
Orchid ales 333 Parietarieae 54
orchidearum, Me1anagromyza 333, 336 Parinarium 102, 104
oreas, Phytomyza 309, 309 Parkia 110, 115
orientalis, Phytomyza 21, 35, 37*, 51 Parkieae 110, 115
Oriental Region 108, 363, 365, 371 Parthenium 313, 316
origani, Phytomyza 205, 207 parvicella, Phytomyza 22, 47, 51 *
Origanum 205, 207, 211 parvicornis, Agromyza 357, 360*, 361
Orlaya 159, 165 parvula, Japanagromyza 113, 135, 136*, 139
ornata, Metopomyza 323, 338, 339*, 341, 393 pascuum, Liriomyza 147, 149*, 155
Ornithopus 112, 127 Paspalum 357, 373
Orobanche 214, 221 Passiflora 78, 93, 117
orobanchia, Phytomyza 214, 220*,221 passiflorae, Melanagromyza (= Tropicomyia
orobi, Agromyza 112, 127, 129* flacourtiae) 78
orphana, Paraphytomyza 235, 236, 237* Pastinaca 160, 163, 175
Ortalidae 388 pastinacae, Phytomyza 160, 175, 176*, 177
Oryza 354, 363, 369 Patrinia 246, 303
oryzae, Agromyza 354, 359, 360*, 363, 364 pauliloewii, Phytomyza 160,169,171*
Oryzeae 354, 359, 371 Pavetta 237
oryzivora, Cerodontha (Poemyza) 354, 369, 370*, 371, Pavetteae 237
398 pedestris, Liriomyza 354, 373, 374*
Osmorhiza 159, 165, pedicularicaulis, Phytomyza 222, 227, 228*
osmorhizae, Phytomyza 159, 165, 166* pedicularidis, Phytomyza 222, 227, 228*
Ostrya 54, 61 pedicularifolii, Phytomyza 222, 227, 228*
Othonna 293 Pedicularis 221, 222, 227
ovalis, Phytomyza 198, 199, 200* Pelargonium 152
ovimontis, Phytomyza 274, 284 Pe1taria 86
oviposition 390 Penetagromyza 389, 390
Oxalis 151*,153, 157 Penetagromyza group 389, 390
Oxystigma 110, 113 penicilla, Phytomyza 259, 267
oxytropidis, Phytomyza 112, 125, 126*, 140 penniseti, Agromyza 357, 362*, 363
Oxytropis 112, 125 pennisetivora, Agromyza 357, 362*, 363
oycoviensis, Agromyza (= A. nigrescens) 152, 398 Pennisetum 357, 363
Oyedaea 311, 315, 316 Penstemon 222, 224
penstemonella, Phytomyza 222, 224
p penstemonis, Phytomyza 222,223*, 224
Pachypodium ISO, 180 Pentaglottis (= Anchusa) 199
Paederia 233, 236 Peperomyia 18, 20
paederiae, Hexomyza 233,234*, 236 peregrini, Phytomyza 274, 283, 284
Paederieae 236 Perezia 257, 258
Paeonia 71 pereziae, Liriomyza 257, 258
Palaearctic Region 108, 357, 364, 367 periclymeni, Chromatomyia 238, 240, 241 *
435
peric1ymeni group, Chromatomyia 240 143, 145, 148, ISO, lSI, 152, 153, 190, 192,200,201,
Peristrophe 213, 217 205,207,213,214,217,220,235,236,237,238,275,
persica, Liriomyza Il2, 130*, 131 311, 316, 389, 390
persicae, Phytomyza 102, 104, 105* Phytobia group 389, 390
Petasites 258, 293, 295, 297, 298, 299 phytochemistry
petasiti, Phytomyza (ssp. of P. tussilaginis) 293, 298, alkaloids 17, 179, 229
299 anthocyanines 62
petiolaris, Phytomyza 198, 199 Apiaceae 158, 320
petoei group, Phytomyza 203, 207, 211 aporphine alkaloids 17
petoei, Phytomyza 205, 207, 208* Arnica 308
Petroselinum 160, 163 Asteraceae 158, 320
Petteria 113, 139 Asterales 320
Petunia 190, 192 benzylisoquinoline 17
Peucedaneae 160, 173 betalains 62
peucedani, Phytomyza 160, 175, 176* carbamates 385
Peucedanum 160, 163, 175 Caryophyllidae 62
peullae, Liriomyza 374*, 375 chlorinated hydrocarbons 385
Pfaffia 63, 67, 68* Cooper-Driver 4
pfaffia group, Ophiomyia 147 Cyperaceae 344
pfaffiae,Ophiomyia 63, 67, 68*, 401 ellagic acid 141
Phacelia 194, 195*, 224 Frohne and Jensen x, 52, 62, 95, 153, 158, 233
phaceliae, Phytomyza 194, 195 * Gentianaceae 183
phalangites, Phytomyza 194, 195*, 274, 282*, 283 Gentianales 179
phalangites superspecies, Phytomyza 283, 284 glycosides 153, 179
phalaridis, Cerodontha (Poemyza) 355, 368*, 369 Hamamelidae 52
Phalaris 355, 369 iridoid compounds 179, 183, 193
Phaseoleae HI, 119 isoquinoline alkaloids 17
phaseoli, Japanagromyza HI, 123, 124* latex 258, 321
phaseoli,Ophiomyia Ill, 119, 121, 122*, 135 lectines 145
Phaseolinae Ill, 121 - Lobelia 229
phaseolivora, Melanagromyza Ill, 120 *, 121 Magnoliidae 17
phaseoloides,Ophiomyia Ill, 122*, 123 myricetin 52
Phaseolus Ill, 119, 121, 123, 135 myrosin 82
phellandrii, Phytomyza 159, 169, 258 Onagraceae 141
Phellodendron 152, 156 organophosphates 385
Phenax 54, 57 Papaverales 50
philactaeae, Phytomyza (= P. kaltenbachi) 21,23 Penncap 385
phil adelphi, Liriomyza 97, 101 permethrin 385
philadelphivora, Liriomyza 96*, 97, 101 Pit tosporaceae 95
Philade1phus 97, 101 Poaceae 344
philippinensis, Pseudonapomyza 357, 375 Polemoniales 320
Phillyrea 214, 220 polyacetylenes 231
phillyreae, Phytomyza 214, 220* polyphenols 52
philoc1ematidis, Phytomyza 21, 33 proanthocyanins 141
philocroton, Tropicomyia 147, 155, 405 prode1phinidin 52
philoxeri, Haplopeodes 63, 69 pyrethroids 385
Philoxerus 63, 69 Rosidae 52
Phleum 355 serum diagnosis 43, 52, 388
Phlomis 204, 205 tannins 52, 141
Phlox 194, 194 terpenoids 183
Phoebe 18, 19 triterpine-saponins 62
Pholiurus 355 Tropaeolum 153
Phragmites 356, 357, 358*, 359, 361, 365, 369 tropan-alkaloids 193
phragmitidis, Agromyza 356, 359 Phytoliriomyza 3,8,10, 11,14,117,192,196,261,
phragmitidis, Cerodontha (Poemyza) 356, 368*, 369 323, 338, 389, 12..!., 393
phragmitophila, Cerodontha (Cerodontha) 356, 365, Phytoliriomyza sp. a 2*
367 Phytoliriomyza sp. b 2*
phryne, Liriomyza 355, 356, 373, 374* Phytoliriomyza group 389, 391
Phyla (= Lippia) 201 Phytomyza 17, 20, 23, 50, 95, 100, 131, 145, 158, 163,
Phylli tis 6, 13 177, 190, 196,221,235,246, 247, 28~ 289, 291, 29~
Physalis 187, 192 299,306,308,315,321,384,388,389, 392,405
Phytagromyza, sg., Cerodontha 371 Phytomyza sp. 20, 21,22, 33, 36*, 160, 175, 205, 251,
Phyteuma 231 258, 259, 271, 273, 289, 293
Phytobia 15,17, 18,19,52,53,54,59,61,70,71,72, Phytomyzinae 390
75,77,79,88,89,91,91, 92,93,94,97, 100, 101, Phytomyzites 59
102, 104, 107, 108, 109, 110, Ill, 113, lIS, 139, 141, Picea IS
436
Poa 355, 365, 371, 377, 379 primulae, Chromatomyia 90, 90*, 94, 390
poae, Cerodontha (Oizygomyza) (= Cer. (Oiz.) Primulales 89, 94
crassiseta) 371 promissa, Calycomyza 265, 274, 280*, 281
poae, Chromatomyia 355, 378*, 379 Prostanthera 205, 207
Poales 322, 342, 352 prostrata, Liriomyza 286*, 287, 289
podagrariae, Phytomyza 159, 167, 168* Protea1es 140, 153
Podonosma 199 Protophytobia 389
Poeae 355 proxima, Agromyza 357, 360*, 361
Poemyza 347, 364, 367, 371 Prunella 204,205,207,211
poemyzina, Cerodontha (Butomomyza) 349 pruni, Phytobia 102, 104
Poissonia 111, 117, 190, 385 Pruninium 100, 102, 104
Po1emonia1es 196, 320 prunivora, Phytobia 102, 104, 105*
polemonii, Melanagromyza 194, 195*,308, 320 Prunoideae 100, 102, 104
Polemonium 194, 308 prunorum, Palaeophytobia 102, 104
pollen 52,54,55,59,62,64,70,71,72,78,79,88,89, Prunus 100, 102, 104
95, 107, 145, 150, 151, 157, 179, 196,203,217,220, pseuderrans, Cerodontha (Butomomyza) 349, 350*,
233, 237, 247, 258, 322, 331, 338, 342, 352, 402 351
Polya1thia 17, 18 Pseudocinchona (= Corynanthe) 236
polydadae, Phytomyza 158, 160, 173, 174*, 177 Pseudodappia 317
Polygala 150, 156 pseudogentii, Chromatomyia ISO, 183, 184*
polygalae, Liriomyza 149*, 150, 156 pseudograta, Me1anagromyza 111, 118*, 119
Polygales 150, 156 Pseudoliriomyza 389, ~
Polygonales 63, 70 pseudometallica, Me1anagromyza 249, 251 *, 257
polygoni, Agromyza 63, 70 pseudomilii, Chromatomyia 351, 352, 354, 371, 377,
polygoni. Japanagromyza 63, 68*, 70 378*, 380, 386
Polygonum 63, 70 Pseudonapomyza 214, 217, 344, 375, 377, 389, 392, 393
Polymnia 311, 316 Pseudonapomyza sp. 213, 354, 377
polymniae, Melanagromyza 311,312*,316 pseudonotata, Phytomyza (= P. notata) 41, 400
polyphaga, Tropicomyia 6, 13, 75, 77, 89, 92,94,97, pseudopygmina, Liriomyza 83, 84*, 86, 86
101,123, 141, 153,154, ISO, 181*, ill, 185,285, pseudoreptans, Agromyza 54, 57, 58*, 93, 198, 299
289, 324, 329, 335, 337, 382, 383, 384, 384 pseudorufipes, Agromyza 198, 199
polyphagy 55, 381, 384, 387, 398 Psidium 143, 154
polyphyta, Tropicomyia 18, 19,22,47,72,75,78,92, Psylloidea 381
97, 101, 110, 111, 113, 119, 141, 145, 148, ISO, 152, ptarmicae, Liriomyza 300*, 301, 303, 307
153, 154, 155, 156, ISO, ill, 213, 218, 236, 238, 242, ptarmicae, Phytomyza 301, 304*, 306
324, 335, 382, 383, 384, 384 Pteralyxia ISO, 180
Polypodiales 5, 8 Pteridales 5, 9
Polypodiatae (= Filicopsida) 5, 389 pteridii, Phytoliriomyza 5, 9*
Polypodiophyta 4, 5, 382 Pteridium 5, 131, 140
Polypodium 5, 8, 13 Pteridophyta 4
Polypogon 355 Pterocarya 54, 59
Pometia 152, 156 Pterocaulon 271, 273
Pomoideae (Maloideae) 106 Ptochomyza 23, 29, 327, 389, 392, 393, 403
Pooideae 352, 354 pubicata, Cerodontha (Butomomyza) 349
populi, Paraphytomyza 81,93 pubicornis, Phytomyza 159, 165, 167, 168*
populi group, Paraphytomyza 81, 82, 107, 125, 220 Puccinelli a 356, 379
populicola, Paraphytomyza 81, 93 puccinelliae, Chromatomyia 356, 379
populoides, Agromyza (= A. isolata) 81 puella, Liriomyza 259, 262*, 263
Populoxylon 79, 93 Pueraria Ill, 119, 121
Populus 79, 82, 93 puerarivora, Ophiomyia 111, 119
Porophyllum 308, 309, 310* pulchella, Phytomyza 41
Portulaca 63 Pu1icaria 271, 273, 400
posticata, Nemorimyza 271, 273, 274, 275, 278*, 293, pulicaria, Ophiomyia 259, 260*, 261
390 pulla, Agromyza 113,138*, 139
Potentilla 102, 104 pullula, Phytomyza 301, 304*, 306, 307, 390
potentillae, Agromyza 61, 97, 100, 102, 103*, 104, 152, Pulmonaria 198, 199
178 pulmonariae, Phytomyza 198, 199
potentillae group, Agromyza 70, 81 Pulsatilla (= Anemone) 27,401
Potentilleae 102, 104 pulsatillae, Phytomyza 21, 27, 30*
Poterium 102, 104 pulsatillico1a, Phytomyza 21, 27, 30*
praecox, Paraphytomyza 235, 236 Punica 143, 154
Praspedomyza (= Liriomyza) 235 Pupalia 63
pre-polyphagy 381, 384, 386 pusilla group, Liriomyza 147, 291, 313, 315
Prenanthes 259, 261, 263 pusilla, Liriomyza 263, 270, 271, 273,274,275, 277,
prespana, Agromyza 356, 359 278*, 285, 295, 315, 316, 319, 320, 386, 398
Primula 90, 94 pusio, Liriomyza (= L. tragopogonis) 263
439
pygmaea, Cerodontha (Poemyza) 354, 355, 356, 369, Ricinus 147, 148, 155
370* riparella, Me1anagromyza 201
Pyracantha 102, 106 Rivinia 62
Pyrus 102, 106 Robinia Ill, 114*, 117, 139, 401
robiniae, Phytoliriomyza 111,114*, 117, 139, 401
Q Robinieae Ill, 117
quadrata, Liriomyza 187, 188*, 190, 192 robusta, Cerodontha (Cerodontha) 365, 366*
Qualea 150, 156 robusta group, Cerodontha 365
querci, Phytomyzites (doubtful species) 59 robustae, Liriomyza 343*, 344, 351
Quercus 54, 59, 81, 139 robustella group, Phytomyza 255,267, 271, 298, 305,
quercus, Japanagromyza 54,59,60*,81, 121, 139 306, 308, 309, 405
quinta, Ophiomyia 274, 275, 276* robustella, Phytomyza 259, 267
Quisqualis 141, 153 rohdendorfi, Cerodontha (Butomomyza) 347, 355, 371,
372*, 380
R rondensis, Agromyza 354, 355, 356, 359, 360*, 364
radicicola, Melanagromyza (= Mel. minimoides) 57, Rorippa 83, 86
397 Rosa 100, 102, 104
ramosa, Chromatomyia 244, 245* Rosales 52, 95, 101, 148
Rangaeris 333, 336 Roseae 102, 104
Ranunculales 20, 21 Rosidae 52, 95, 107, 140, 153, 157, 179, 383, 389
Ranunculeae 21, 35 Rosiflorae 52, 158
ranunculella, Napomyza 21, 29, 38, 39* Rosoideae 100, 102, 102
ranunculella group, Napomyza 35 rostrata, Ophiomyia 244
ranunculi, Phytomyza 22,41,43,44*,50,51, 190 rostrata, Phytomyza 222, 226*, 227, 229
ranunculicaulis, Napomyza 21,40*,41,402 Rottboellia 357, 375
ranunculicaulis, Phytomyza 402 - Royena 89, 94
ranunculicaulis, Ophiomyia 21, 23, 24* ruandae, Melanagromyza 292*, 293
ranunculicola, Phytomyza 22, 41, 44* Rubia 235, 236
ranunculina, Phytomyza 22,41,45* Rubiales 233
Ranunculinae 21, 35 Rubieae 236
ranunculivora, Phytomyza 22, 41, 44*, 50, 125, 140, Rubioideae 233, 235, 236
177 Rubus 102, 104, 105*, 107
ranunculivora group, Phytomyza 140 Rudbeckia 263, 283, 311, 315, 316, 319, 398
Ranunculoideae 21, 27 Ruellia 213,215
ranunculoides, Liriomyza 21, 23, 24*, 50 ruelliae, Melanagromyza 213, 215, 216*
ranunculoides, Phytomyza 22,41,44* rufescens, Phytomyza 259, 267, 268
Ranunculus 20, 21, 23, 35, 41, 50, 86, 140 rufipes, Agromyza 57, 299, 300*, 301
Raphanus 83, 86 rufipes, Phytomyza 85, 86, 88
rapunculi, Phytomyza 231, 232*, 400 Rumex 63,70
ravasternopleuralis, Phytomyza (ssp. of P. senecionis) Ruprechtia 63, 70
293, 298, 403 Ruta 152, 156
rectae, Phytomyza 21, 27, 33 Rutales 320
Reichardia 259, 261 rydeni, Phytomyza 22, 43, 45*
renovata, Napomyza 21,38, 39*, 50, 402 rydeniana, Phytomyza (= P. albiceps) 163, 251, 255
reptans, Agromyza 54, 55, 57, 58*, 79 Rytidostylis 79, 93
Reseda 86
Reynosia 148, 155 S
Rhamnales 148, 155 Sabazia 315, 316
Rhamnus 148, 155 sabaziae, Liriomyza 251,253, 271, 273, 274, 277, 308,
Rhinantheae 222, 224, 227 309, 314*, 315, 316
Rhinanthoideae 222, 224 Saccharum 357, 375
Rhinanthus 222, 227, 229 Saccopetalum 17, 18
Rhizophorales 143, 154 sagehenensis, Me1anagromyza 102*
Rhodiola (= Sedum) 99, 101 sahyadriae, Agromyza 363
rhodiolae, Phytomyza 98*, 99 Salicales 79, 93
rhodopaea, Phytomyza 198, 199 salicaria, Palaeophytobia 79, 93
Rhus 150, 156 salicifolii, Japanagromyza 59, 80*, 81, 93, 141
Rhynchosinapis (= Brassicella) 85 salicin a , Agromyza 81, 82*, 93, 178
Rhyncosia 111,119,123 salicis, Hexomyza 81, 93
Rhyncosporoideae 347, 351 Salicornia 63, 67
Ribes 97, 101 Salix 79, 93
Riccia 1,2 Salpichroa 190, 192
Ricciocarpos 2 Salpiglossideae 190, 192
Ricciopsis 1 Salvia 205,209, 211
richteri, Liriomyza 373 salviae, Phytomyza 205,208*, 209, 400
ricini, Ophiomyia 147, 149*, 155 Sambucus 238
440
Trifolium 112, 127, 131, 133, 140, 399 va1dorensis, Agromyza 101, 102, 103*
Triglochin 340, 341, 392 Valeriana 245, 246
triglochinae, Uriomyza (= L. angulicornis) 340 valerianae, Liriomyza 245*,246
Trigonella 112, 133 Valerianella 245, 246
Trigonotis 198, 199 valerianellae, Uriomyza 245*, 246
triplicata, Cerodontha (Icteromyza) 345, 346*, 351 Vandeae 336
tripolii, Napomyza 274, 280*, 281 Vandoideae 336
Trisetum 355, 379 Vanguerieae 237
tristella, Japanagromyza 111,120*, 121, 139 Vanguerilla 237
Triticeae 356 Vanilleae 336
Triticum 356, 359, 365, 375 varicornis, Agromyza 112, 127, 129*
Triumfetta 72 variegata, Phytoliriomyza 111. 112, 123, 124*, 125,
trivittata, Phytomyza 222, 223*, 229 127
Trixis 258 varifrons, Agromyza 53, 55, 56*, 178
Trollieae 21. 23 varii, Phytomyza 21, 35
trollii, Phytomyza 21.23, 24* Varilla 317
trolliophila, Phytomyza 21. 23, 24* Varilovia 127
trolliovora, Phytomyza 21. 23, 24* varipes, Phytomyza 222, 228*, 229
Trollius 21. 22, 23 Venegasia 308, 309
Tropaeolum 88, 131, 153, 157, 295, 385 venegasiae, Liriomyza 308, 309, 310*, 319, 320
Tropicomyia 4, 8, 117, 123, 179, 323, 324, 337, 384, venezo1ana, Agromyza 361, 363
389, 391, 404 venturii, Cerodontha (Xenophytomyza) 367
Tropicomyia sp. 5,8,9*,13,78,93,201,219*,235, venusta, Agromyza 362*, 363
237, 325, 335, 337, 341 veratri, Uriomyza (= L. wachtli) 331
tschirnhausi, Cerodontha (Poemyza) (= Cer. (Po.) Veratrum 330*, 331, 336
calamagrostidis) 367 verbasci, Amauromyza 204,214, 218, 221, 222, 223*,
tundrensis, Phytomyza 309, 309, 310* 224,229,231,233, 385
Turritis 86 verbasci, Me1anagromyza 221, 222, 223*
tussilaginis, Phytomyza 293,296*, 298 verbascifolii, Haplopeodes 187, 189*, 192, 192
tussilaginis ssp. kevani, Phytomyza 293, 298 Verbascoideae 221, 222
tussilaginis ssp. petasiti, Phytomyza 293, 298 Verbascum 221, 222
Tussilago 293, 296*, 298, 299 Verbena 201, 203
Tylophora 89, lBO, 183 verbenae, Ca1ycomyza 201, 202*, 203
Typha 331, 333, 340, 341 verbenae, Me1anagromyza 201
Typha1es 340 Verbeneae 200, 201
verbenivora, Ca1ycomyza 201, 202*, 203
U verbenivora, Ophiomyia 201
Ulex 113, 139 Verbesina 311, 315, 316
Ulmarieae 104 verbesinae, Ca1ycomyza 315, 316, 318*
ulmi, Agromyza (= A. aristata) 55 verbesinae, Me1anagromyza 311,312*,316
Ulmus 53,54 veris, Agromyza (= A. rondensis) 359
Umbella1es 320 vernalis, Phytomyza lBO, 184*, 185
umbilici, Uriomyza 98*, 99, 101 Vernonia 268, 270, 277
Umbilicus 99, 101 vernoniae, Me1anagromyza 268, 269*, 270
umbrina, Uriomyza 222, 224, 225* vernoniae, Pseudonapomyza 269*, 270
umbrinella, Uriomyza 222, 224, 225* vernoniana, Me1anagromyza 268, 269*, 270
umbrosa, Uriomyza 222, 224, 225* Vernonieae 248, 267, 270, 285
Urera 54 Veronica 222, 224
Urochloa 357, 361, 363 veronicicola, Phytomyza 222, 224, 225*
urophorina, Uriomyza 329, 330*, 335, 392 verticillatae, Phytomyza 144*, 145, 154
Urophylleae 236 vibeana, Phytomyza (= P. aquilonia) 41
Urophyllum 235, 236 Viburnum 238, 287
Urtica 54, 57, 397, 400 Vicia 112, 127. 131, 139, 140
urticae, Agromyza (= A. pseudoreptans) 57 viciae, Agromyza 112
urticae, Uriomyza 54, 57, 58* viciae, Me1anagromyza 112, 127, 128*
Urticales 52, 53, 54, 145 viciae, Phytoliriomyza 112, 130*, 131, 140
Urticeae 54 Vicieae 112, 127
urticella, Melanagromyza 54, 57, 58* vicifoliae, Agromyza 112, 127, 129*
urticivora, Me1anagromyza 54, 57 Vigna 111, 121, 123, 135,405
urundensis, Pseudonapomyza 213,215,217* vigna1is, Melanagromyza Ill, 120*, 121
vigneae, Cerodontha (Butomomyza) 349, 350*, 351
V vigneae, Tropicomyia 111, 123, 124*, 405
Vaccaria 63 vignivora, Ophiomyia 111, 122*, 123
Vaccinioideae 93 Viguiera 311, 315, 316
Vaccinium 88, 93 Villanova 308, 309, 310*
Vah10dea (= Deschampsia) 354, 377 Vinca lBO, 180
444
W
wachtli, Liriomyza 330*, 331, 336, 392
Wahlenbergia 229, 231
wahlenbergiae, Liriomyza 230*,231, 232, 271
wahlgreni, Phytomyza 259, 267, 268*
watersi, Agromyza (= A. varicornis) 127
websteri, Hexomyza 110, 114*, 115, 139
Wedelia 291,311,313,316
wedeliae, Melanagromyza 291,293,311, 312*, 313,
316,317,320
wedeliaphoeta, Melanagromyza 313, 316
Weigela 237
We1witschia 15
Wikstroemia 141, 154, 329
wikstroemiae, Kleinschmidtimiyia 141, 142*, 154
Willemetia (= Chondrilla) 259
williamsoni, Phytomyza 21, 35, 37*, 109
winderi, Phytomyza (= Chr. platensis) 209, 399
winnemanae, Hexomyza 80*, 81, 93
wistariae, Agromyza 110, 115
Wisteria 110, 115
Withania 190, 192
woerzi, Agromyza 242, 243*, 244
X
xanthaspis, Metopomyza 347, 348*, 351
Xanthium 277, 313, 316, 399
xanthocera, Liriomyza 21, 23, 35, 50, 63, 75, 83, 84*,
85,86, 86, 92,112, 113, 131, 133, 150, 153, 155,157,
160, 194,205, 211, 246,251, 382, 383, 384, 384, 385,
399
Xeniomyza 389, 392
xenophagy 163, 381, 384, 386, 387
Xenophytomyza, sg. Cerodontha 364, 367
Xeranthemum 251
y
yanoi, Japanagromyza 147, 155
yanonis, Agromyza 354, 356, 361
yasumatsui, Phytomyza 21, 29, 33, 36*
Yellowstone National Park 19, 100, 104
Z
Zanha 151, 156
Zantedeschia 337, 341