432 Journal of Agricultural Research Voi.xvin, no.s 



To examine the various parts of the intestine and the body wall, how- 

 ever, it was necessary to cut serial sections of the body. The insects 

 were killed and fixed in Carnoy's solution in periods varying from one 

 to six days after inoculation, after which they were prepared by the 

 usual methods and stained in Erlich's haematoxylin and eosin, such stains 

 having been found of value in differentiating blastocysts within the 

 blood cells. 



Although hundreds of sections from a large number of insects were 

 examined, it has been impossible to observe the conidial germ tubes 

 penetrating either the body wall or the intestine. When insects were 

 fed with leaves smeared with conidia, it was possible actually to observe 

 the fungus entering the mouth ; yet sections cut of such individuals did 

 not show conidia even in the lumen of the alimentary tract, in spite of 

 the fact that such insects were killed and fixed before the conidia-bearing 

 leaf fragments were voided. In certain slides, however, blastocysts 

 were observed within the folds of a tissue connecting the longitudinal 

 and transverse muscles of the intestine; but it could not be determined 

 whether this position was attained by entrance of the blastocysts from 

 the body cavity or the intestine (PI. 52, K). Similarly no conidia were 

 observed externally penetrating the body wall of the inoculated insects, 

 though special care was taken to examine the regions noted above and 

 in spite of the fact that in many cases the conidial spore paste was applied 

 in definite regions to facilitate examination. 



That the fungus does gain entrance into the body cavity is obvious, 

 however; and as the following tests will show, it seems probable that both 

 the body wall and the intestine may be penetrated by the conidial germ 

 tubes. Inability to observe them is perhaps due to improper time of 

 killing or faulty technic. 



To test the parasitism of the fungus and to determine a method of 

 infection that would be adapted to the inoculation of insects artificially 

 on a large scale many tests were conducted of a more general nature 

 than those noted above. A variety of insects, representative of nearly 

 all the larger orders, were inoculated by one or sometimes by all of the 

 methods described below. The hosts used were as follows : 



DIptera — larvse of Musca domestica L. 



Coleoptera — larvse of Lachnostema, larvse of Elateridae. 



Orthoptera — nymphs and adults of grasshoppers. 



Isoptera — workers of Reticulitermes sp. 



Lepidoptera — larvae of Feltia jaculifera Guen., Fcltia subgothica Haw., Peridroma 

 saucia Hiibn., Agrotis ypsilon Rott., Noctua c-nigrum L., Chloridea obsoleta Fab., 

 Hyphaniria texior Harr., Leucania unipuncta Haw., Phlegethontius sexta Joh., Bombyx 

 mori L., and several other members of the family Noctuidae. 



Although the fungus has been collected in the field in this country 

 upon two species of Euxoa only, all the Noctuidae used in the labora- 

 tory tests, including the corn earworm and army worm, were shown 



