202 BOTANICAL GAZETTE [MARCH 
the inner region of the hold-fast. The latter presents clearly the char- 
acter of the typical rhizoid; thick, frequently yellowish walls, contain- 
ing but a thin layer of protoplasm, which bounds externally the 
large vacuolar cavity that almost fills the hypha. The walls of the 
ordinary vegetative hyphae, on the other hand, are thin; those of the 
rhizoids in this case appear to have undergone a gelatinous or slimy 
modification, and the contents seems to be undergoing degeneration. 
The growth of the conidiophores, in the case of E. sciarae, proceeds 
more slowly than that of the rhizoidal hyphae, a phenomenon which is 
probably due, in part at least, to the slower absorption of water by 
those hyphae destined to bear conidia. At any rate, the vacuoles 
which are formed at this time in the cells increase slowly in size, and a 
conidiophore arises from near the end of each cell and grows out radi- 
ally, boring its way through the tissues of the host (figs. 12, 14, 15)- 
As THAXTER and others have noted, the conidiophores of certain 
species remain simple and unbranched, as in the case of E. muscae 
(fig. 40); or, in other species, they may become normally septate and 
branched, as shown in E. sciarae (fig. 16), and Empusa sp. (fig. 23)-° 
From this fact arise most interesting cytological variations in the vari- 
ous conidia of these species. THAXTER has brought out clearly the 
variation in size, shape, etc., of the conidia of many species, but 
CAVARA (’9Q9) was the first to contrast the multinucleate condition 
of the conidia of Empusa muscae with that of the uninucleate conidia 
of Entomophthora Delpiniana. Of the six species studied by me, 
four have uninucleate conidia (E. sciarae, fig. 27; E. americana, 
figs. 36, 37; E. aphidis, figs. 42, 43; and Empusa sp., figs. 22, 26): 
The conidia of E. culicis are normally two-, rarely three-nucleate 
(figs. 31, 32); while those of E. muscae have a more or less 
indefinite number, frequently about 15-18 (figs. 38, 39).2. The more 
common uninucleate conidia arise primarily from the septation of the 
conidiophores into uninucleate segments; whereas, on the other hand, 
the simple conidiophore of E. muscae does not usually become septate 
except at the conidium. In the last case, therefore, all of the many 
nuclei of the last-formed vegetative cell, which forms the origin of the 
2 While convinced of the value of CAVARA’s suggestion as to the use of nuclear 
characters in the classification of the Entomophthoreae I do not think that this one 
character alone would justify the separation of Entomophthora from Empusa. 
