HONEY ANTS. 



373 



lower is undoubtedly the opening of the epinotal chamber and leads 

 directly into one of the large inflated cavities. As all of my specimens 

 are dry and carded, I am unable to ascertain the histological structure 

 of these organs. I am convinced, however, that they represent, as 

 [anet supposed, an enormous development of the organs found in the 

 coresponding portion of the epinotum of our common ants. This is 

 also indicated by an examination of specimens of C. diffonnis from 

 Perak and of rnncronata from Sumatra. In these the openings of the 



FIG. 220. Repletes of Myrmecocystus horti-deorum hanging from roof of honey 



chamber. (McCook.) 



epinotal chambers are more ventral and more slit-shaped than in inflata, 

 and may therefore be described as intermediate between those of inflata 

 and our northern species of Cremastogastcr. 



As the function of the epinotal glands, even in our common ants, is 

 still unknown, we can hardly expect to form a satisfactory conception 

 of the hypertrophied homologues of these organs in a few Indomalayan 

 ants that have hardly been studied in a living condition. That these 

 organs should secrete a sweet liquid to be fed to the ants or their young 

 is surprising at first thought and suggests the nursing habits of the 

 Mammalia, but when we stop to consider that ants are in the habit of 

 feeding their young and one another with a secretion of the labial, or 

 salivary glands, we can see no reason why, in certain species, the 

 thoracic glands might not be developed for a similar purpose. It will 

 be very interesting, nevertheless, if future investigation proves that cer- 

 tain species of Cremastogastcr, a genus whose members are so con- 



