26 Bulletin American Museum of Natural History. [V'ol. XXIII, 



that the latter is greatly reduced in size. This interpretation, which derives 

 its plausibility from the soldier-like structure of the thorax in the parasitized 

 Pheidole ahswda, is equally applicable to P//; covimutata. Emery is, how- 

 ever, unable to account satisfactorily for the appearance of the ocelli and 

 larger eyes. The case of Pheidole is certainly complicated by the existence 

 of a soldier caste. Both in the Ponerine ants studied by Emery (Odonto- 

 machus, Neoponera, Paraponera, etc.), and in the species of the Myrmicine 

 Cremastogaster above described, this caste does not exist, and hence the 

 mermithergates may be only excess developments of the worker, or defective 

 developments of the female. Although in these cases I believe that the 

 former alternative is the more probable, at least in some of the species, the 

 question can hardly be settled without much additional material. 



Emery takes it for granted that the gaster is an excess development, 

 because it is so greatly enlarged in the adult mermithergates. But I have 

 shown that in the pupffi in Ph. com,mutata, at least, the gaster is proportion- 

 ally no more developed than any other part of the body. Hence we can- 

 not conclude that the head is small because the gaster is large. We must 

 rather suppose that the whole body exhibits excessive and uniform growth, 

 and this must, of course, mean a uniform distribution of the metabolized 

 nutriment in the larva. ^ It would seem, therefore, that the mermithergates, 

 if they really arise from soldier larvte, retain small heads because the para- 

 sites and not the abdominal tissues appropriate the substances which in the 

 normal soldier go to form the cephalic region. This is indicated by the 

 fact that the head of the soldier does not develop till the semi-pupal stage, 

 when the pellucid cephalic substances may be seen rapidly accumulating in 

 the previously slender anterior region. This explanation, which resolves 

 itself into a very obvious struggle between the larva and its parasite and not 

 between the parts of the larva, will also account for the smaller heads in the 

 Ponerine mermithergates. In all cases the pupa tends to become a well- 

 proportioned whole. It is evident that this tendency is very different from 

 an unequal struggle of the various body regions with one another, since it 

 implies a uniform distribution of the available nutriment. It implies, more- 

 over, defect developments in the head, in the case of the soldier larvae. 

 The failure to develop the huge head of the soldiers in Pheidole mermither- 

 gates is in all probability due to the suppression of a very recently acquired 

 character by the parasite, since there is every reason to suppose that the 

 soldier caste among these and other ants is of much less phylogenetic anti- 

 quity than the worker. This seems to be clearly indicated in the ontogene- 



1 Dr. T. H. Montgomery informs me that crickets (Gryllus ahbreviatiis) infested with Mermis 

 or with Paragordius varius are apt to be larger than uninfested individuals. The enlargement 

 is not confined to the abdomen but also affects the other regions of the body as in mermithergates. 



