1907.] Wheeler, The Polymorphism of Ants. 27 



tic development of colonies of Pheidole, C amponotus , etc., for in these only 

 small-headed workers are at first produced by the queens. Here scarcity 

 of food produces almost the same results as parasitism — namely microce- 

 phaly. At the same time that the tendency to produce a well-proportioned 

 pupa suppresses the recently acquired macrocephalic characters of the 

 species, it may be conceived to encourage the development of certain ancient 

 phylogenetic characters like the ocelli and larger eyes. This is especially 

 the case in the Pheidole larva for two reasons : first, because the mermither- 

 gates of these ants probably approach more nearly the ancestral stature of 

 the workers, since it is very probable that the Pheidole workers have under- 

 gone a reduction in size during the phylogeny of the genus, and second, 

 because the larva, notwithstanding its parasites, nevertheless succeeds in 

 accumulating more formative substance than is necessary for the production 

 of a normal worker.^ 



On the basis of the above interpretation, which I would substitute for 

 Professor Emery's, an interesting comparison may be instituted between 

 the mermithergates of Pheidole commntata and ahsiirda on the one hand, 

 and the phthisergates of Ph. instahilis on the other. In the former cases 

 the parasitic ilffr??z;'.s-, situated in the abdomen of the larval ant, appropriates 

 to its own growth the substances that will later be required to increase the 

 volume of the pupal and imaginal head. In the latter case, the ectoparasitic 

 Orasema larva, situated just behind the head of the semipupal ant, inter- 

 cepts and extracts the head-producing substances as they flow forward 

 from the abdominal region info the anterior end of the body. In both cases 

 the result is a pathological microcephaly, extreme and peculiar in the mori- 

 bund pupal phthisergates of Ph. instahilis, much less pronounced but per- 

 sisting into the imaginal stage in the less seriously parasitized Ph. absurda 

 and commutata. 



Emery has called attention to the interesting fact that all the known 

 mermithergates occur among American ants. They occur moreover in 

 species peculiar to the tropical or subtropical portions of the New World. 

 This would seem to indicate that the single or several species of Mermis 

 which produce this anomaly in ants are of rather circumscribed distribution. 

 The genus Mermis, however, seems to be cosmopolitan, judging from von 

 Linstow's recent revision of the group (/. c). He records species from 

 Europe, Turkestan, INIadagascar, Southwest Africa, New Zealand, United 

 States, Costa Rica, and Brazil. Some of the species, like the European M. 



1 The effects of the parasitic Mermis on the host are not confined to mere mectianical dis- 

 tortion and the withdrawal of nutriment. There i.s much evidence to .show tiiat ijotli Nematodes 

 and Ce.stodes secrete toxic substances that iiave a positive and far reacliiiiK effect on the tissues 

 of their liosts. In this connection see: von Ftirth, Vergleichende Chemisclie Physiologic der 

 Niederen Tiere, Jena, 1903, pp. 308-310, and Faust, Die Tierischen Gifte, Braunschweig. 1906. 

 pp. 223-228. 



