WHEELER— ANT LARV^. 301 



the surface of the egg-packet. I have repeatedly seen the workers thus feed- 

 ing the very young larvae, a single globule of regurgitated food serving for 

 a meal of which four or five larvse successively partook. 



Undoubtedly the majority of Myrmicinse, Dolichoderinse and Cam- 

 ponotinse, the three most highly specialized subfamilies of ants, 

 feed the brood throughout its larval stages with regurgitated liquids. 

 Concerning larval feeding in the Dorylinse nothing is known. 



I come now to a consideration of some of the ant larvse collected 

 by Mr. Lang in the Belgian Congo. Four of these, all belonging to 

 the subfamily Myrmicinse, are of unusual interest. One of the 

 species is a new Pccdalgus which I shall describe elsewhere as P. 

 termifolesfes sp. nov., the third of the genus to come to light, as 

 only one Indian and one other West African species were previously 

 known. The workers of termitolcstes are minute brownish yellow 

 ants which live in the masonry of large termite hills and undoubtedly 

 prey on their inhabitants. Their habits therefore resemble those of 

 the well-known thief-ants, Solcnopsis iiiolcsta of Xorth America and 

 6". fngax of Europe. The larva (Fig. 5) has a singular shape, 

 being almost spherical, with a short neck, small head and minute, 

 bidenticulate mandibles. The delicate integument is studded with 

 very short, stifif hairs, each of which has two recurved branches. 

 The larvje, which are held together in compact masses by the inter- 

 locking of these hooked hairs, are fed with liquid food by regurgi- 

 tation as is evident from the contents of their large spherical 

 stomachs and the very feeble development of their mouthparts. 

 Although, like other Myrmicinse, they do not spin cocoons but form 

 naked pupae, they nevertheless possess huge salivary glands. Even 

 in the very young larva (Fig. 5^) the salivary receptacle on each 

 side is full of a clear liquid secreted by the large cells of the two 

 branches of the gland. In the nearly full-grown female larva (Fig. 

 5-B) the glands are very voluminous and have their lumen and that 

 of the receptacle fitll of secretion shown as dark, compact masses in 

 the figure, which was, of course, drawn from a specimen hardened 

 and dehydrated in alcohol. As such an amount of saliva would 

 hardly be necessary for digestive purposes and as it is not used in 

 the form of silk by the full-grown larva, it probably serves as a 

 store of food for the nurses. The Pccdalgus larvae, therefore, would 



