152 ROLAND THAXTER ON THE 



veiiieiit receptacle for specimens; and the latter should be kept separate. In the labora- 

 tory each specimen should be inverted over a slide or cover in a moist chanil)er, until a 

 sufficient number of conidia have been discharged, when it may be dried for the herba- 

 rium. Conidia, obtained in this way, may usually be kept for reference for an indefinite 

 period; and, since they allow the comparison of very large numbers of spores side by 

 side, are most convenient for study. 



The artificial propagation of Empusae, by the infection of fresh hosts, I have found a 

 much more difficult matter than one would suppose, even when the host infected was of 

 the same species as that from which the spores were obtained for this purpose. Infec- 

 tion between dissimilar hosts I have found still more difficult; although, in two instances, 

 I have been successful in infecting caterpillars with E. Gri/Ui developed on grass- 

 hoppers, as well as in transferring E. sphaerosperma from leaf hojipei's ( Tyjjhlocyhct) 

 to a Pieris larva. The method which I have adopted for infection consists in the use of 

 a tightly-covered jelly tumbler in which the upper portion is separated from the lower by 

 a round piece of wire netting. By placing the hosts to be infected in the loAver of the 

 two chamljers thus formed and fastening a specimen of JEinpusa m the upper one, the 

 living hosts below can hardly escape the spores discharged through the netting. 



The period which ensues after the infection of a host until its death varies to some ex- 

 tent. In the larger hosts, such as Hies or catei-pillars, death may not take i)lace for twelve 

 days; although the usual period is from five to eight days. In minute hosts this period 

 must be considerably shortened, owing to the ephemeral character of many forms known 

 to be subject to the attack of Empusae. The first visible symptom of the disease is a 

 general restlessness of the host. In caterpillars, for instance, the insect leaves its food 

 plant and wanders restlessly about; usually endeavoring to climb upwards before death, 

 which is apparently quite sudden and iniaccompanied by contortions of the body. The 

 host insect thus remains clinging to the object on which it rests or is fastened to it by 

 rhizoids. Certain insects are fixed by the insertion of their probosces into the substratum 

 on which they rest, as is the case with aphides. The house-fiy is, I believe, always fas- 

 tened by its proboscis which adheres firmly to the substratum. "Where rhizoids ai-e de- 

 veloped, they often appear before the death of the host, and I have seen a geometrid 

 moth, which was thus firmly attached to a pine needle, fluttering violently in its at- 

 tempts to escape. 



In the account of the separate species of Empusae which follows I have used this 

 generic name for all the forms, employing the names EntomojMhova and Trijdosporium 

 in brackets as groups of subgeneric value. In Empusa p)roper I have inchuled forms 

 in which the branching of the conidiophores is of the simple type, and the formation of 

 resting spores presumably asexual, taking E. Muscae as the type. Under Triplosporium, 

 I have included the two forms E. Fresenii and E. lageniformis, the jjositiou of which in 

 Empusa is only provisional. The group is characterized I)_y the production of conidia 

 having a smoky tint, thick-walled, with evenly granular contents and producing peculiar 

 almond-shaped secondary conidia on capillai-y conidiophores; while its zygospores are 

 elliptical, each originating as a bud which rises upward from the point of conjugation 

 of gametes arising from two hyphal bodies. Further study of E. lagenifor^nis may 



