30 Annals Entomological Society of America [Vol. XI, 



hour bouillon culture of B. poncei. October 13th the animals 

 were all alive. I pulled out one metathoracic leg from each 

 animal and permitted a drop of blood from each to flow into a 

 nutrient bouillon tube. Three tubes were kept at room tem- 

 perature and three were incubated, yet all six remained per- 

 fectly sterile. Stained smears of some of the blood also failed 

 to reveal any micro-organisms. Sooner or later, I think, the 

 gut would have ruptured liberating the intestinal flora into the 

 body cavity, so I thought it best to make the tests the second 

 day. 



I further conclude that passage infections performed by 

 using the alimentary tract are hopeless on account of the 

 extensive flora. Blood passages, with B. poncei, were Hkewise 

 useless, in most cases, for the reason that the gut ruptured after 

 a short time. Passages by means of the blood are possible with 

 other bacteria, however, as I will show later. 



Experiments with Cultures ''Souche Cham'' and 

 "Souche Sidi.'' 



The infection experiments with "Souche Cham" and 

 "Souche Sidi" were much more satisfactory than those with 

 B. poncei. In regard to "Souche Cham," I successfully per- 

 formed two passages, but curiously enough, as can be seen 

 from Tables XV-XVII (1 and 2), obtained no increase in 

 virulence. Perhaps if I had measured the time between inoc- 

 ulation and death in hours instead of in days, I might have 

 noticed something, but many deaths unfortunately occurred 

 during the night. However, measurement of time in days is 

 sufficient and if a marked increase in virulence had manifested 

 itself, I surely would have noticed it. 



The gut of M. atlanis never ruptured, so the blood or muscle 

 tissue could readily be used as a basis for further inoculations. 

 In no case, however, can extracts from the stomach or the 

 intestines be used for further passages. A series of examinations 

 conclusively proved that these are invariably contaminated 

 even in perfectly normal looking animals. 



Tables XVIII and XIX represent experiments deahng with 

 food infections. M. atlanis was also the subject for these tests. 

 In general the time between infection and death is somewhat 

 extended which is to be expected in this mode of experiment, 



