GASTRIC C/ECA OF THE HETEROPTERA. 143 



parasite, although the salivary glands often appeared almost 

 completely filled with them. 



Some time later, upon examining specimens of Podisus mac- 

 uliventris, an insect of the same family, but in which the caeca 

 are wholly wanting, a similar infection was observed. This 

 time, however, the alimentary canal was the seat of the infection; 

 and although a very careful examination was made of the salivary 

 glands in specimens showing a very heavy intestinal infection, 

 these organs were found to be invariably free from any infection 

 with the flagellate. In Podisus the infection appears to be re- 

 markably common, fully fifty per cent, of the specimens from 

 some localities containing this parasite. The third stomach 

 seems to be the place of greatest multiplication of the flagellates 

 in this insect, and not only the Hcrpetomonas but also various 

 forms of foreign bacteria were observed in this region. This is 

 in marked contrast to the condition found in the midgut of those 

 forms which are provided with caeca; for of the hundreds of 

 typical pentatomids examined, flagellates were found in the 

 alimentary canal of but two specimens of Ccenus delius, which 

 were apparently on the point of dying when dissected, and in 

 these the parasites were confined chiefly to the first stomach 

 instead of reaching their greatest development in the third 

 stomach as in Podisus. 



For a long time after the singular infection of the salivary 

 glands in Peribalus had been observed, no satisfactory explana- 

 tion could be offered for the strict localization of the typical 

 intestinal flagellates in these organs. It was not realized at first 

 that there might be some possible connection between this 

 phenomenon and the bacteria normally infecting the caeca of the 

 insect, but later when this case was compared with similar infec- 

 tions in such forms as Podisus, in which the caeca were wholly 

 wanting, it seemed that the caecal bacteria must in some way be 

 responsible, and this view was later confirmed by direct culture 

 experiments. 



We may assume that these Herpetomonas parasites, upon being 

 continually introduced into the alimentary canal of Peribalus, 

 and being unable to develop in the midgut in the presence of the 

 antagonistic caecal bacteria, gradually become adapted to a life 



