WHEELER— ANT LARV^. 323 



tivities slow down so that the larva may hatch before all the prey 

 is brought in (" approvisionnement massif ralenti"). This condi- 

 tion leads naturally to a feeding of the larva from day to day with 

 living but paralyzed caterpillars ("education surveillee indirecte") 

 and eventually to a stage essentially like that of the social wasps in 

 which the caterpillar is chewed up and placed as a pellet in the mouth 

 of the larva ("education surveillee directe"). Synagris cornuta has 

 reached this last stage. The mother insect, while malaxating the 

 caterpillar, herself imbibes its juices. 



The internal liquids having partly disappeared during this process of 

 malaxation, the prey is no longer, as it was in the beginning, soft and juicy 

 and full of nutriment for the larva. It is possible, in fact, to observe that the 

 caterpillar patee provided by the Synagris cornuta is a coarse paste which 

 has partly lost its liquid constituents. There is no exaggeration in stating 

 that such food would induce in larvse thus nourished an increase of the 

 salivary secretion in order to compensate for the absence of the liquid in 

 the prey and facilitate its digestion. 



It is here that the further development to the condition seen in 

 Bclonogastcr and other social wasps sets in. The mother wasp 

 finds the saliva of the larva agreeable and a trophallactic relation- 

 ship is established. As Roubaud says, 



the nursing instinct having evolved in the manner here described in the 

 Eumenids, the wasps acquire contact with the buccal secretion of the larva, 

 become acquainted with it and seek to provoke it. Thence naturally follows 

 a tendency to increase the number of larvse to be reared simultaneously in 

 order at the same time to satisfy the urgency of oviposition and to profit by 

 the greater abundance of the secretion of the larvse. 



Although considerable evidence thus points to trophallaxis as the 

 source of the social habit in wasps, ants and termites, it must be ad- 

 mitted that the phenomenon has not been observed in the social 

 bees. That the latter may have passed through a phylogenetic stage 

 like that of Synagris seems to be indicated by the solitary bees of 

 the genus Allodape to which I have already referred (p. 318). 

 Brauns' observations, though meager, show nevertheless that Allo- 

 dape has reached Roubaud's fourth stage, that of direct feeding of 

 the larvae from day to day, and if I am right in supposing that the 

 peculiar appendages of the larva: are exudate organs, there would 

 be grounds for assuming that trophallaxis occurs in this case. On 



