Art. XLII.—The Genus Cordyceps in New Zealand. By G. H. Cunningham. With Special Entomological Notes on the Hosts, by J. G. Myers. [Read before the Wellington Philosophical Society, 27th October, 1920; received by Editor, 31st December, 1920; issued separately, 8th August, 1921.] Plates LIX–LXII. In the genus Cordyceps are included those fungi which produce the so-called “vegetable caterpillars,” “vegetable wasps,” &c., which are insects that have been attacked by fungi and their tissues replaced by the vegetative portion of the attacking fungus. In writings of about a century ago the various species of Cordyceps were supposed to be insects changing into plants. To quote one example, an author (25) in 1763, describing Cordyceps sobolifera Tul., which he called the “vegetable fly,” states, “In the month of May it buries itself in the earth and begins to vegetate. By the latter end of July the tree arrives at its full growth and resembles a coral branch, and is about three inches high, and bears several little pods which, dropping off, become worms, and from thence flies, like the English caterpillar.” Naturally, the earlier systematists had some difficulty in placing such peculiar fungi. Species of this genus were first included under Clavaria (12), a Basidiomycete; they were then transferred to Sphaeria (19), a genus which at that time covered all the genera now included in the Pyrenomycetes; thence to Cordyceps by Link (11); from this to Torrubia* The genus Cordyceps of Link was by Tulasne (22) divided into two genera: (1) Torrubia, because of the presence of two spore-forms in the life-cycle; and (2) Cordylia, embracing all forms growing on subterranean fungi. by Tulasne; and, as this latter genus was not tenable, back to Cordyceps. For the most part, the species of Cordyceps grow on insects, but two—C. capitata (Holmsk.) Link, and C. ophioglossoides (Ehr.) Link—grow on subterranean fungi, Elaphomyces spp. C. ophioglossoides has recently been recorded growing on a locust in Japan (15). Distribution. The genus Cordyceps is widely distributed, being found in Britain, Europe, North and South America, China, Ceylon, Japan, Australia, and New Zealand. Many species are extremely limited in their distribution, while others again are more or less cosmopolitan: e.g., Cordyceps gracilis Grev. has been recorded from Britain, Europe, North America, Algeria, Australia, and, doubtfully, from New Zealand (as Cordyceps entomorrhiza (Dicks.) Link). Biology. Little is known of the life-history of Cordyceps. Tulasne (21) and de Bary (1) have worked out the life-history of the common European species, C. militaris (L.) Link. Their investigations tend to show that a spore, on coming in contact with a host, germinates and produces a germtube which penetrates the cuticle and body-wall. Inside the body-cavity this germ-tube branches, forming hyphae, which penetrate to all parts of the body. In the blood gemmae are produced: these are cells asexually produced from the ends of hyphae. They are exceedingly small, and are

rapidly carried in the blood-stream to different parts of the body, where they in turn give rise to hyphae. In this manner the fungus rapidly spreads and quickly kills the host. Infection of the host may occur from the germ-tube from an ascospore, or from hyphae developed from conidia borne by the Isarial form of Cordyceps. A conidium may germinate, and the subsequent hyphae live saprophytically on decaying wood or other organic matter for some considerable time. These hyphae on coming in contact with a host are eapable of entering the host-tissues. In the decaying wood from which Cordyceps Aemonae Lloyd was taken, mycelial development was so pronounced as to be visible to the naked eye. The writer carried out some rough experiments to ascertain whether this mycelium was capable of attacking the larvae of Aemona hirta Fabr., the host of C. Aemonae. Healthy host larvae (quiescent) were obtained from rotting logs in which no sign of Cordyceps was found, and were buried in pots filled with sterilized sawdust in which were mixed fragments of infected wood taken from the centre of the log that contained C. Aemonae. The pots were kept moist and covered with bell jars. In two months' time these larvae were exhumed, and were all found to be dead and surrounded by hyphae. They were replaced, and in three months stromata bearing the Isarial stage of C. Aemonae appeared above the surface of the sawdust. Unfortunately this experiment was not carried further to determine whether the perithecial stage could be obtained; but at the time of the first experiment Isarial forms of C. Aemonae were brought into the laboratory from logs in the forest in which they were found, and were buried in sawdust with the stromata alone showing. The pots were kept moist and covered; in three months immature perithecia had appeared on one or two of the stromata. (Plate LIX, fig. 1, b). The sawdust used in these experiments was obtained by sawing up dead, sound, dry logs of mahoe (Melicytus ramiflorus Forst.). In the host the hyphae continue to develop until finally the whole of the internal tissues are replaced by the mycelium of the fungus, when it forms a hard, compact mass, the cuticle and sometimes portions of the alimentary system alone remaining unaltered. (Plate LIXI, fig. 2.) This mycelial mass is known as a sclerotium; from it, usually after a period of rest, the stromata bearing the fructifications of the fungus arise. The stromata vary considerably in shape, size, and number, according to the nature and habitat of the host. If the host is subterrancan, then the stromata will necessarily have to be long enough to rise to the surface of the ground, so the length would be governed by the depth of the host. Again, if the host is exposed, as in the case of Cordyceps clavulata (Schw.) Ellis & Ev.,* This species occurs on various species of Lecanium. the stromata would necessarily be short. In some species there are two kinds of fructification: the first is known as the Isarial form, and bears conidia; the second form, which appears after the Isarial (when the latter is present), bears the ascospores. Conidia are simple, short-lived spores, and are abjointed in immense numbers from the ends of hyphae. They may be borne on a stroma, in which case they are abjointed from the terminals of the hyphae forming the stroma, or may occur on the terminals of hyphae which form a loose covering over the external surface of the host. The relationship between this Isarial and the later (or Cordyceps) stage is known in a few species only, and in the majority of cases is assumed merely on account of the occurrence of both forms from the same host. As mentioned above, Isaria is capable of

living as a saprophyte; Isaria-like forms also occur as the conidial stages in the life-cycle of Xylaria, the species of which are saprophytes, occurring on dead logs, grass, &c. The ascospores of Cordyceps are filiform, multicellular bodies borne in asci (cylindrical sacs), which in turn are enclosed in perithecia (variously shaped receptacles bearing asci on their inner walls). The perithecia are, as a rule, densely packed on the surface of or embedded in the substance of the stroma. Each is provided with a definite opening (ostiole) through which the spores escape at maturity. Each ascus beais a small cap on its distal end, pierced by a minute pore. The ascospores are filiform, and lie closely packed in parallel fascicles, eight in each ascus; they are at first continuous, but when mature are divided by many transverse septa—a hundred or more. Eventually they break up at these septa into secondary spores. Each secondary spore is capable of germinating and infecting a host. From this it is obvious in what enormous numbers these spores are produced. Assuming a stroma to bear 100 perithecia, each perithecium to contain 100 asci, and each ascospore to break up into 100 secondary spores, the number of ascospores produced would total 8,000,000 —and this is, of course, a very modest estimate of the actual contents of each peritheciuni, ascus, &c.: for example, a large specimen of Cordyceps Robertsii Hook. contains many thousands of perithecia. Distribution of Spores. Conidia are light, minute bodies borne on the ends of hyphae, and are thus admirably adapted for wind distribution. Ascospores, being enclosed in perithecia, are primarily dependent on other means of distribution. If a mature perithecium be placed in water, in an hour or so enormous numbers of asci are seen to be collected outside the ostiolum. They have been forced out of the perithecium by the swelling of certain hyphal tissue at the base of the asci. No doubt in nature a similar condition exists: here the spores are forced out and remain on the exterior of the perithecia, or are washed on to the ground, leaves, logs, &c., and when dry may be carried by wind, insects, or other agency to some distance from their source. Technical Description of the Species. Although a large number of species have been described, five only are definitely known to occur in New Zealand; of these, four are endemic, and one occurs also in Australia and Tasmania. Cordyceps* The name Cordyceps was first used by Fries as the name of a tribe of the Pyrenomycetes, including the genera Cordyceps and Xylaria. Torrubia was first used by Léveillé in manuscript in the Paris Museum Herbarium, and was later adopted by Tulasne (l.c.). (Fries) Link, Handbk., vol. 3, p. 347, 1833 (emended). Sphaeria § Cordyceps Fries, Syst. Myc., vol. 2, p. 323, 1823; Torrubia Lév. Tulasne in Fung. Carp., vol. 3, p. 5, 1865. Stromata† Stromata: This term is used very loosely by the various mycologists who have worked on this genus; as here used it includes both fertile and sterile portions of the clubs. arising from a sclerotium composed of mycelial tissue within the bodies of insects (rarely in other fungi), simple or branched; sterile below, fertile on upper portion. Perithecia immersed or superficial, seated on or in fertile portion of stroma; spherical, oval, flask-shaped, &c.; ostiolate. Asci cylindric, 8-spored, hyaline, distal end capitate; paraphyses absent.

Spores hyaline, filiform, multiseptate, arranged in the asci in parallel fascicles, or interwoven; breaking up in the asci into secondary spores, or remaining entire. Isarial stage when present forming an effused downy weft or an erect simple or variously branched stroma, consisting of hyphae bearing the hyaline continuous conidia on their apices. 1. Cordyceps Sinclairii Berk., Fl. N.Z., vol. 2, p. 338, 1855. (Plate LXII, fig. 2.) Sphaeria Basili Taylor, N.Z. and its Inhabitants, p. 424, 1844. Torrubia caespitosa Tulasne, Select. Fung. Carp., vol. 3, p. 11, 1865. Cordyceps caespitosa Sacc., Syll., vol. 2, p. 565, 1883. Isarial Stage: Stromata growing from head of host, yellowish, from 18 mm. to 25 mm. high; stems cylindrical, slender, simple or forked, sometimes confluent, 8 mm. or more high, divided above into numerous more or less cylindrical simple or slightly-lobed heads, which are sometimes disposed into a flabelliform mass clothed with innumerable oblong conidia 7–8 μ long. (Berkeley.) Perithecia unknown. Hosts.—Melampsalta cingulata Fabr.; M. cruentata Fabr. (Plate LXII, fig. 3.) Type Locality.—Tauranga, Poverty Bay, in loose gravelly soil in garden of Bishop Williams; “growing from larva of some orthopterous [sic] insect.” Distribution.—Tauranga (Colenso); Farewell Spit, Nelson (Benham) (2); Weraroa (E. H. Atkinson)! Hokitika (unknown collector)! There is a fine specimen in the Canterbury Museum collection! (Plate LXII, fig. 2, c). No. 79, Biol. Lab. Herb. (Crypt.), Wellington. This form should really have been named as an Isaria, as only he conidial form is known. It is possible that this may be the conidial stage of Cordyceps sobolifera Tul., as this species occurs on cicada in Japan. As all the other species occurring in New Zealand are endemic, with the exception of Cordyceps Robertsii, which is found only in Australia and New Zealand, it is, however, more likely that Cordyceps (?) Sinclairii is also endemic. It is a very variable form, and assumes many different shapes. The colour of the stroma ranges from white in the most immature specimens, through yellow (colour mentioned by Berkeley), light brown, in more mature forms becoming pink, deepening in colour with age. Although specimens are fairly plentiful in the New Zealand museums, none are known in any of the mycological collections abroad (14). A most interesting account of this species (with plate) is given by Benham (2). Notes on the Hosts (by J. G. Myers).—Of the four specimens available for study, only two are in at all a good state from an entomological point of view; but it is significant that all four hosts are nymphs of the final instar, with wing-pads well developed and the whole appearance suggestive of almost immediate emergence. This is of interest in that it is an indication that the nymphs are, of course, full-grown—a fact which enables an estimate of their species to be made with greater accuracy than would otherwise be possible. The two large specimens can be assigned almost certainly to Melampsalta cingulata Fabr. (6), while the two others, both of the same size and smaller than the other two, belong to one of the smaller cicadas, most probably to Melampsalta cruentata Fabr. (6).

2. Cordyceps Craigii Lloyd, Myc. Notes, p. 527, f. 718, 1911 (emended). (Plate LX, fig, 3; and text-figs. 1, 2.) Isarial stage unknown. Stroma solitary, 5–7 cm. long; growing from head of host; stem 3–4 mm. thick, 3–4 cm. long; fertile portion brown when fresh, blackening with age, flattened, falcate, 2–3 cm. long, 8–10 mm. wide, 3–4 mm. thick; surface smooth, or punctate with ostioles of perithecia. Perithecia completely immersed, densely packed in stroma, flask-shaped, with long slender slightly curved necks; up to 1,500 μ long, 300–500 μ wide; walls 35 μ thick. Asci hyaline, narrowly cylindrical, tapering slightly towards distal end, markedly towards proximal end, terminating in a long slender pedicel, not constricted below cap; 250–330 × 6–7 μ. Spores in parallel fascicles in asci, same thickness throughout, ends bluntly pointed, 180–260 × 2 μ; secondary spores 3–4 × 2 μ; readily separable in asci. Fig. 1.—Transverse section through fertile portion of stroma. Fig. 2.—A. Perithecia (enlarged). B. Capitate apex of ascus. C. Base of ascus. D. Secondary spores, 3–4 × 2 μ. Host.—Porina enysii Butl.; growing from head. (Plate LX, fig. 2.) Type Locality.—Old and abandoned kumara (Ipomoea batatas Poir) beds, Auckland. Distribution.—Auckland (E. Craig); Wellington, in ground under a karaka (Corynocarpus laevigata Forst.), in forest, vicinity of Wireless Hill (unknown collector)! No. 192, Biol. Lab. Herb. (Crypt.), Wellington. “Mr. Craig also sends two specimens collected in the bush which are very similar and probably the same species. I could not say positively, however, from the specimens, as they are both immature.” (Lloyd.) Specimen 192 was given me by Mr. H. Hamilton, of the Dominion Museum. He obtained it from a man who dug it up in the forest under a karaka. Note on the Host (by J. G. Myers).—As this species has so far been recorded only from the North Island, the host, taking its size into considertion, is almost certainly Porina enysii Butl. (1), the larva of which, in the North Island, is the victim also of Cordyceps Robertsii Hook.

Fig. 1.—Cordyceps Aemonae Lloyd. × 6. a, c. Showing characteristic fasciculate growth of stromata. b Isarial form. Arrow points to developing perithecia. Fig. 2.—Larva of Aemona hirta Broun, the host of Cordyceps Aemonae. Lloyd. × 2½. Fig. 3.—Imago of Aemona hirta. ×2½.

Fig. 1.—Cordyceps consumpta n. sp. Natural size. Fig. 2.—Porina enysii Butl., the larva of which is the host of C. Ciaigii and C Robertsii. Natural size. Fig. 3.—Cordyceps Craigii Lloyd. Natural size

Fig. 1.—Cordyceps Robertsii Hook. Reduced ½. Fig. 2.—Tiansverse section through sclerotium; taken from anal portion of specimen marked × in fig. 1. × 10.

Fig. 1.—Porina dinodes Meyr., the larva of which, in the South Island, is the host of C. Robertsii, Natural size. Fig. 2.—Cordyceps Sinclarii Berk. Specimens from a to d show gradual development of the stroma; the colour ranges from white in a, through yellow in b, brown in c, to pink in d. In a, b, and c the conidia are borne on the tufted apices; in d they form a packed mass round the central axis. Reduced ⅔. Fig. 3.—a, Melampsalta cruentata Fabr., the nymps of which are the hosts of a and d. b. Melampsalta cingulata Fabr., the nymphs of which are the hosts of b and c, Reduced ⅔.

3. Cordyceps consumpta, n. sp. (Plate LX, fig. 1; and text-figs. 3, 4.) Isarial stage unknown. Stromata gregarious, two springing from head; 2–3 cm. long; fertile portion cylindrical, curved, apex obtuse, black, 8–10 mm. long, 2–3 mm. thick; rough with projecting necks of the perithecia; sterile portion slender, cylindrical, straight or curved, glabrous, black, 20 mm. long, 1–5 mm. thick. Perithecia completely immersed, flask-shaped, or more frequently very irregular and distorted; not crowded in the stroma, each perithecium being separated by stromal hyphae; necks protruding; 1,000–1,200 μ, long, 200–500 μ wide; necks short; walls 30 μ thick. Asci hyaline, narrowly cylindrical, tapering slightly towards distal end, markedly towards proximal end, not constricted below capitate apex; 250 × 7 μ. Spores in parallel fascicles in asci, same thickness throughout, ends bluntly pointed, 180–220 μ; secondary spores 4–5 × 1–1.5 μ, readily separable in asci. Fig. 3.—Transverse section through fertile portion of stroma. Fig. 4.—A. Perithecia (enlarged: note distortion). B. Capitate apex of ascus. C. Base of ascus. D. Secondary spores, 4–5 × 1.5 μ. Host.—Porina sp. (see note); growing from head. Type Locality.—Rotorua, N.Z., growing from larva buried in soil (A Lush)! Distribution.—Known only from type locality. No. 230, Canterbury Museum collection. (Type.) In macroscopic characters this species resembles Cordyceps falcata Berk., but differs in having the perithecia completely immersed; in C. falcata they are perfectly superficial. In microscopic characters there is a strong resemblance to Cordyceps Craigii Lloyd; but the difference in perithecial characters, together with the difference in all macroscopic characters, indicates that this is a valid species. It bears a closer resemblance to C. falcata and C. Craigii than to any other described species. This specimen, together with many others, was kindly forwarded for examination by Mr. G. Archey, of the Canterbury Museum. It was collected by Mr. A. Lush at Rotorua in June, 1920. Unfortunately, no particulars as to exact locality were appended. Note on the Host (by J. G. Myers).—The larva infected, unless it be immature, must in this case be that of one of the three smaller common

species of the genus Porina—namely, P. cervinata Walk. (23), P. signata Walk. (24), or P. umbraculata Guen. (8). At present we have no ascertained constant structural characters by which to distinguish these larvae. The insect is far too small for a larva of Porina enysii Butl. (5)—at any rate, for a full-grown one. 4. Cordyceps Robertsii Hook., Fl. N.Z.;, vol. 2, p. 202, 1855 (emended). (Plate LXI, figs. 1, 2; and text-figs. 5, 6.) ? Sphaeria larvarum Westw., Proc. Ent. Soc. Lond., vol. 2, p. 6, 1836. Sphaeria Robertsii Hook., Icon. Pl., vol. 1, t. 11, 1837, S. Hugelii Corda, Icon., vol. 4, 44 f, p. 129, 1840. S. Forbesii Berk. in Lond. Jour. Bot., vol. 7, p. 578, 1848. Torrubia Robertsii Tul., Sel. Fung. Carp., vol. 3, p. 6, t. 1, 1865. Cordyceps Selkirkii Olliff in Ag. Gaz. N.S.W., vol. 6. p. 411, 1895. ? C. Coxii Olliff, l.c. C. larvarum (Westw.) Olliff, l.c. Isarial stage unknown. Stroma slender, 10–38 cm. long; fertile portion 6–12 cm. long, 3–4 mm. thick, acute, densely covered with superficial perithecia, which reach to apex of stem; brown, becoming black with age; sterile portion slender, 5–15 cm. long, 2–3 mm. thick, same colour as fertile portion. Fig. 5.—Transverse section through fertile portion of stroma. Fig. 6.—A. Perithecium (enlarged). B. Capitate apex of ascus (note constriction below cap). C. Base of ascus. D. Secondary spores, 5–6 × 3 μ. Perithecia superficial, small, elongate-obovate or elliptical, densely packed around central axis, easily separable; dark brown, composed of coarse hyphal threads; 600–880 × 300–400 μ; wall thick, 30–50 μ. Asci hyaline, narrowly cylindrical, tapering slightly towards distal end, markedly towards proximal end, terminating in a long slender pedicel; slightly constricted below capitate apex; 280–400 × 9–10 μ. Spores in parallel fascicles, filiform, equally thick throughout, bluntly pointed, multiseptate, 280 × 3 μ; secondary spores 5–6 × 3 μ, not readily separable in asci. Hosts.—Porina enysii Butl.; P. dinodes Meyr. Growing usually from head, rarely from anal region. (Plate LX, fig. 2; Plate LXII, fig 1.) Type Locality.—Given in Icones Plantarum as “N.Z.” Distribution.—More or less general throughout the North Island. Specimens have been recorded from the following localities: Rotorua

(H. Hill) (10), Petane (A. Hamilton) (9), Auckland (E. Craig) (13), Waikanae (H. C. Field) (7), Raurimu (E. H. Atkinson)! In the South Island this species appears to be less common. Specimens have been recorded from the following localities: Catlin's and Tokonui Range, near Gore (teste Benham); Riverton (W. G. Howes)! Lloyd (16) states that it has been collected in Australia by Cheel. Rodway (20) records its incidence in Tasmania. Australia, N.S.W. (E. Cheel). Tasmania (L. Rodway), gully at foot of Mount Wellington. No. 191, Biol. Lab. Herb. (Crypt.), Wellington. Many improbable tales of the life-history of this species may be found in the earlier articles on the subject. One assertion that has gained credence is that these fungi are found only under the rata (Metrosideros spp.). It is true that they are often found under the rata, but they occur as frequently in areas in which no rata is found growing; for example, Hamilton (9) records its occurrence under Coprosma grandifolia Hook. f. Frequent mention is made of its being used as food by the Maori. I am informed by Mr. Elsdon Best, that, although in times of famine the Maori undoubtedly made use of certain terrestrial and arboreal fungi as articles of food, they certainly did not eat the awheto, as this fungus is called by them. The sclerotium of Cordyceps Robertsii was, however, made use of in tattooing, Mr. Best stating that the vegetative portion of the fungus, or awheto—the living grub being known as ngutara—was burnt and pulverized, and the powder so obtained mixed with water to form a black paste. The pattern of the tattoo having been marked out on the limbs and body (the pigment was not used on the face, as it did not give a deep enough black), the edge of the uhi whakatataramoa was placed on a line of the pattern and the back struck with the take rarauhe,* The take rarauhe was made from a piece of fern-stalk (Pteridium esculentum Cockayne), lashed round at the end to prevent its splitting. Fuller information on tattooing as practised by the Maori may be obtained in the Journal of the Polynesian Society, vol. 13, p. 166, 1904. causing the skin to be severed. A second implement, the uhi puru (which had a serrated edge), was then dipped in the pigment, applied to the cut made by the uhi first used, and struck with the take rarauhe, the pigment remaining on removal of the uhi. Cordyceps Robertsii is an extremely variable form. (Plate LXI, fig. 1.) I have specimens with a single stroma; with stromata occurring in pairs from the head; with a single stroma bifurcate about half-way between apex and base; and with stromata growing from both head and anal regions. Specimens are most plentiful in the summer months. Hill (10) states that the mature stage is most common in October, November, and December, but he has seen Maori children offering them for sale along the Rotorua railway-line as late as March. Note on the Hosts (by J. G. Myers).—Practically all the earlier naturalists accepted without question the current belief that the host of this species was Hepialus virescens Dbld. Hamilton, Field, Maskell, and other writers in the early volumes of the Transactions of the New Zealand Institute considered this to be the only larva large enough to coincide with the “vegetable caterpillar” in size. This was, however, merely a conjecture. G. V. Hudson was the first to point out the improbability of the arboreal Hepialus as a host, seeing that the infected larvae were invariably found underground. He suggested Porina mairi Buller as the host; but only one specimen of this moth has been taken, and the supposition rests only on its large size,

its metamorphosis being unknown. The argument that its rarity may be due to the heavy larval mortality consequent on the attacks of Cordyceps Robertsii is obviously inadmissible. Porina mairi may be a host, but that it is the usual host is highly improbable. In 1895 Olliff (Ag. Gaz. N.S.W., vol. 6, p. 407) supported the hypothesis that the victim was the larva of a species (not necessarily P. mairi) of Pielus (syn. Porina), a view which subsequent evidence has justified. The first experimental indications of the host's specific identity were published in 1903, when A. Philpott registered his opinion that the larva of Porina dinodes Meyr. “is the vegetable caterpillar. No other moth in this district [Southland] known to me is large enough to warrant the assumption that its larva may be the host of the fungus. I have several times found the fungus-attacked larvae here, and, so far as a comparison between these and the living larvae of P. dinodes can be trusted, I think it bears out my opinion.” W. G. Howes gives similar evidence (1910) regarding Cordyceps found plentifully at Riverton. He found “along with the fungi … an apparently healthy larvae of Porina dinodes, and, so far as I can see, all the vegetable caterpillars there were those of this moth. The largest specimen I took was 5 in., but I have never seen a living dinodes larva of this length, and suppose that the fungus growth distends the skin of the host.” As Porina dinodes is confined to the South Island, definite evidence of the North Island host was lacking. This evidence was however, forthcoming in 1905, when G. V. Hudson, at Karori, received “two Hepialid larvae, one very recently dead and infested with Sphaeria [Cordyceps] fungus … the other an identical larva unaffected by the fungus—alive and very healthy. Both the larvae were found in the earth, close together, amongst the roots of some native shrubs.” The healthy larvae was successfully reared, and proved to be that of Porina enysii Butl. As distinguished from the smaller and commoner species of the genus, this frequents the bush, and is by no means rare in the localities where Cordyceps abounds. It is interesting to note that the abundance of the imago, like that of many other Lepidoptera, is somewhat periodic. One season may produce large numbers in a locality where the moth was at other times rare. It seems probable that the usual host of Cordyceps Robertsii is Porina dinodes Meyr. in the South Island, and P. enysii Butl. in the North. For the suggestion that the larva of Sphinx convolvuli L. may sometimes be the victim no grounds of evidence exist. Moreover, an affected caterpillar of this species would be immediately recognizable by Sphingid characters which no fungous attack would completely obscure. 5. Cordyceps Aemonae Lloyd, Myc. Notes, p. 932, fig. 1695, 1920. (Plate LIX, fig. 1; and text-figs. 7, 8.) Isarial stage preceding the perithecial on the same stroma; at first white and pruinose with conidia, becoming light brown; conidia hyaline, subglobose, 4–6 μ. Stromata fasciculate, 3–5; stipitate, short, 2–3 mm. long; tipped with sterile apices, light brown; growing from head of host. Perithecia subsuperficial, irregularly globose, obtuse, contiguous; light brown, becoming dark with age; 300–500 μ in diameter; wall thick, up to 80 μ. Asci hyaline, narrowly cylindrical, tapering slightly towards distal end, markedly towards proximal end, terminating in a long slender pedicel; not constricted below capitate apex; 180–220 × 5–6 μ.

Spores in parallel fascicles in asci, filiform, same thickness throughout, ends bluntly pointed, multiseptate, 100–120 × 2–2.5 μ; secondary spores easily separable in ascus, 3–4 × 2–2.5 μ. Host.—Larva of Aemona hirta Fabr.; growing from head. (Plate LIX, figs. 2, 3.) Type Locality.—Weraroa (G. H. C.), in rotting logs of mahoe (Melicytus ramiflorus Forst.). Distribution.—Known only from type locality. No. 78, Biol. Lab. Herb. (Crypt.), Wellington. (Co-type.) These specimens were collected by the author and sent to C. G. Lloyd. They were obtained from rotting logs of mahoe. This is a brittle, soft wood, soon decaying on contact with the ground. The parasitized larvae were all found with their heads towards the surface of the log. It is apparently necessary for the stromata to make their way through about 5 mm. of solid wood before coming to the surface; frequently, however, they follow the old larval tunnels until they come to an opening at the exterior. Fig. 7.—Transverse section through fertile portion of stroma. (Enlarged.) Fig. 8.—A Perithecium (enlarged), showing coalescence of walls. B. Capitate apex of ascus. C. Base of ascus, showing hyaline pedicel. D. Secondary spores, 3–4 × 2–2.5 μ. Although the perithecia are superficial, in section they appear to be immersed. That this is due to the coalescence of the perithecial walls with the formation of a pseudo-stromal tissue, careful microscopic examination shows. (Text-fig. 2, A.) Note on the Host (by J. G. Myers).—Aemona hirta Fabr., Broun in Man. N.Z. Coleopt., p. 570, 1275: This fairly common Cerambycid beetle was reared by David Miller, Government Entomologist, from healthy larvae taken with specimens undoubtedly of the same species, infested with the mycelium of Cordyceps Aemonae Lloyd. This beetle is variable in size, colour, and relative quantity of pubescence one variety formerly ranking as a distinct species under the name of Aemona humilis Newman. The latter species, commonly known as the “flat-headed lemon-tree borer,” falls, according to Broun (3), into synonymy with Aemona hirta Fabr. under the name of A. humilis. The species has been recorded as a pest of lemon-trees in the Auckland District (4). The beetle passes its larval and pupal stages in manuka (Leptospermum scoparium Forst.) (3), mahoe, and a variety of other trees and shrubs (teste W. W. Smith, Taranaki).

Doubtful Species. The following have been recorded as occurring in New Zealand, but as no specimens have to my knowledge been collected, and as none from this biological area exist in any of the mycological herbariums abroad, they are here recorded as doubtful. Cordyceps gracilis Grev. Host.—“Larvae of insects.” Massee (17) states that this species—determined by him as Cordyceps entomorrhiza (Dicks.) Link—was collected by Colenso and sent to Kew. There are no specimens of this fungus from New Zealand in the Kew Herbarium (14). Cordyceps Gunnii Berk. Olliff (l.c.) doubtfully records this species for New Zealand without mention as to who collected it, where it was collected, or where the original reference was obtained. It is probable that a specimen of Cordyceps Craigii Lloyd has been mistaken for it. Frequent mention is made of a species of Cordyceps attacking Hepialus virescens Dbld. Unfortunately, I have not seen any specimens, and so cannot do more than record this animal as a host. The following particulars have been supplied by Mr. Myers :— In the Entomologist, London, vol. 31, p. 128, 1898, W. G. Howes records the discovery of “vegetable caterpillars” in the trunks of trees buried at a considerable depth and exposed by mining operations at Orepuki, Southland. The situation of these infected larvae was taken as indubitable proof that Hepialus was the host [of C. Robertsii]. The matter, however, must remain extremely uncertain, since Hepialus virescens is confined to the North Island. The second case occurred about 1903, when G. V. Hudson was “shown a specimen of a vegetable caterpillar in the trunk of a tree.…On examination I at once recognized the insect as a larva of Hepialus virescens, and the portion of the tree-trunk with the burrow in which this larva was situated precisely agreed with the usual habitat of that species.” Literature cited. 1. Bary, A. de, Comp. Morph. & Biol. of Fungi, &c., 1887. 2. Benham, W. B., Trans. N.Z. Inst., vol. 32, p. 4, 1900. 3. Broun, T., N.Z. Dept. Ag Rep., p. 157, 1896. 4. —–, N.Z. Dept. Ag Rep., p. 151, 1897. 5. Butler, A. G., Proc. Zool. Soc. Lond., p. 381, 1877. 6. Fabricius, J. C., Syst. Ent., p. 680, 1775. 7. Field, H. C., Trans. N.Z. Inst., vol. 28, p. 623, 1896. 8. Guenée, Ent. Mo. Mag., vol. 5, p. 1, 1868. 9. Hamilton, A., Trans. N.Z. Inst., vol. 18, p. 210, 1886. 10. Hill, H., Trans. N.Z. Inst., vol. 34, p. 401, 1902. 11. Link, H. F., Cordyceps, Handbook, vol. 3, p. 347, 1833. 12. Linnaeus, C., Clavaria, Species Plantarum, p. 1182, 1753. 13. Lloyd, C. G., Myc. Notes, p. 528, 1911. 14. —–, Synopsis Cordyceps Australasia, p. 11, 1915. 15. —–, Myc. Notes, p. 809, fig. 1260, 1918. 16. —–, Myc. Notes, p. 912, 1920. 17. Massee, G., Annals of Botany, vol. 9, p. 26, 1895. 18. Meyrick, E., Trans. N.Z. Inst., vol. 22, p. 206, 1890. 19. Persoon, C. H., Sphaeria, Syn., p. 90, 1801. 20. Rodway, L., Proc. Roy. Soc. Tas., p. 116, 1919. 21. Tulasne, L. R., Ann. Sci. Nat., 4, ser. 8, pp. 34–43, 1857. 22. —–, Torrubia Lév., Sel. Fung. Carp., vol. 3, p. 4, 1865. 23. Walker, F., Cat. Lep. Brit. Mus. Suppl., vol. 2, p. 595, 1865. 24. —–, Cat. Brit. Mus., vol. 7, p. 1563, 1856. 25. Watson, J., Clavaria sobolifera in Phil. Trans. Lond., vol. 53. pp. 271–74, 1763.