54 Saw-flies, Gall-flies, Ichneumons, 



such a division of the group can be fairly made. These three families can 

 be distinguished by the following key: 



Basal peduncle of the abdomen composed of a single segment (the first) (Fig. 743). 



Abdomen not constricted between the second and third segments (Fig. 743, r ). 



CAMPONOTID.E. 



Abdomen constricted between the second and third segments (Fig. 743, 2 ) . PONERID.E. 

 Basal peduncle of the abdomen composed of two segments (Fig. 743, 3 ) . . 



Of these families that of the Poneridae is the smallest in number of species, 

 and includes the least specialized (as regards sharply marked division of 

 labor, differentiation into castes, and complexity of 

 the communal life) of all the ants. In the following 

 brief accounts of a few of the better known American 

 ants the family relationship of each of the species 

 referred to is indicated. 



Of the Poneridae only about 25 species are so far 

 known in this country; all are stingers, although 

 not very strong ones, and but a few species are at 

 all common. Little was known of their habits 

 and life-history before the recent studies of Profes- 

 sor Wheeler on three species occurring in Texas, 

 FIG. 743. _ Diagrams of namely, Odontomachus hamalodes, Pachycondyla 



lateral aspect of abdo- harpax, and Leptogenys elongata. The nests, made 



men of representatives i . ... 



of the three families of under stones or lo S s > are pnmitive structures, corn- 



ants: i, Camponotidse; posed of a few simple and irregular burrows or gal- 



2 Poneridae; 3, Myrmi- j rf SQme f which mn a] th surface of th 



cidse. a, thorax; b, nrst 



abdominal segment; c, soil immediately beneath the stone or log, while 



second abdominal seg- ot h er s extend obliquely or vertically downwards 



ment; a, third abdom- .... . * _,, . 



inal segment. for from 8 to 10 inches. Ihere are no widened 



chambers. The nests of L. elongata comprise ten 



to fifty individuals, those of P. harpax fifteen to one hundred, and those of 

 O. hamatodes one hundred to two hundred. Ergatoid (worker-like) females, 

 no larger than and almost exactly like the true workers, existed in all the 

 nests; the workers of none of the species fed each other or the males and 

 females, and the larvae were fed simply by giving them pieces of freshly killed 

 insects, which they chewed and devoured by means of their unusually well- 

 developed mandibles. This method of larval feeding is more primitive 

 (demands less care and manipulation on the part of the workers) than in 

 the case of any other ants, indeed of any other social insects, for even the 

 wasps, which also feed their young pieces of insects, masticate these insect 

 morsels thoroughly before turning them over to the tender larvae. The 

 feeding of the Ponerine larvae is also very irregular and capricious both as 



