380 Transactions. 



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 main may be a host, but that 

 it is the usual host is highly improbable. 



In 1895 Olliff {Ay. 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 appa- 

 rently healthy larvae of Porina dinodes, and, so far as T 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 cater- 

 pillar 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 /*. 



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 /x in diameter ; wall thick, up 

 to 80 /x. 



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 X 5-6 \x. 



