Jan. IS. 1920 Further Studies of Sorosporella wvella 407 



certain of the smuts, such as Urocystis and Tubercinia, a resemblance 

 noted by Sorokin (77) in 1888. Furthermore, early stages in the ger- 

 mination of the resting spores in hanging drops of water (PI. 51, C) are 

 similar to analogous processes of promycelium and sporidium formation 

 in some of the Ustilaginales ; but although this resemblance is rather 

 striking, it is purely superficial and of no phylogenetic significance. 



The resting-spore aggregations are more naturally comparable to that 

 class of propagative bodies known as bulbils, which, as Lyman {13), 

 Hotson {10), and others have shown, occur in the life histories of certain 

 Ascomycetes and Basidiomycetes. Their method of development, how,- 

 ever, is quite different. 



So far as can be determined from his imperfect description, something 

 similar to the conidiophore production apparently was observed by 

 Krassilstschik {11) ; and this phase of development is in itself sufficient 

 to show that the organism can not possibly be included in the Entomoph- 

 thorales, since the coenocytic hyphae of the latter, together with the 

 peculiar method by which their conidia are formed and discharged, are 

 radically different and in no way comparable to corresponding char- 

 acteristics of the form under consideration. 



The germinations thus obtained from unbroken larvae treated as 

 described above were found to be confined to those spores only which 

 lay near the surface, immediately beneath the integument. When, 

 however, the infected larvae were torn open and the red spore masses 

 freely exposed to the air and light, a somewhat different result was 

 obtained in several instances. Under these conditions the germination 

 involved practically all the spores composing the mass, and a luxuriant 

 growth developed quite unlike that obtained in the first instance. The 

 differences observ^ed were due to the fact that germ tubes from adjacent 

 chlamydospores cohered in such a way that Isaria-like fascicles of co- 

 nidiophores were produced (PI. 51, G). Conidia were formed in the same 

 manner and are apparently identical in every way with those described 

 above. This type of growth does not invariably appear when infected 

 larvae are treated as described, but it is not unusual. The larv^ae died 

 but a few days before the tests were performed, so the resting spores 

 were comparatively young and fresh. 



The Isaria-like grouping of the conidiiferous hyphae is somewhat 

 similar to that which occurs in certain species of the genus Cordyceps, 

 a condition that may be permanently conidiiferous, or represent a stage 

 preliminary to the perfect or acigerous phase of the development. Such 

 a condition, which was found to be parasitic upon leafhoppers (Perkinsi- 

 ella) and other insects in Hawaii, and which is similar in many respects 

 to Isaria saussurei that attacks wasps, has been described and figured 

 by the writer (19) under the name "sterile Cordyceps." Furthermore, 

 it has been suggested by Cooke (4) that Isaria saussurei is the imperfect 



