328 WHEELER— ANT LARV^. 



3. The various parasitic ants, of which a number of species have 

 come to Hght within recent years and have been described by Was- 

 mann, Donisthorpe, Emery, myself and others, can be shown to 

 have established trophallactic relations with their host species. One 

 of the most instructive is Leptothorax emersoni which lives with 

 Myrmica canadensis. I have described its habits and those of one 

 of its subspecies in three of my former papers (1901, 1903, 1907). 



4. The relations of ants to plant-lice and other Homoptera and 

 to the larvae of Lepidoptera outside the nest are, as I have said, 

 mcompletely trophallactic, since these insects are not fed, though 

 they may be defended by the ants. The Homoptera are not fed 

 probably for the simple reason that their mouthparts are so pecu- 

 liarly speciaHzed for piercing plant-tissues and sucking their juices, 

 and the Lepidopteron larvse have, as a rule, no occasion to abandon 

 their leaf diet. There are, however, several cases in which both 

 caterpillars and Homoptera have entered into more intimate asso- 

 ciation with the ants. Many of the root aphids and coccids and 

 their eggs are collected and kept by the ants in their nests, at least 

 ■during certain seasons of the year. Two of the caterpillars that 

 have acquired closer relations with the ants are so instructive, as 

 illustrating one of the ways in which the myrmecophilous habit has 

 been developed, that they merit more detailed description. 



F. P. Dodd (191 2) found that the first-stage larva of a small 

 gray Queensland moth, Cyclotorna monoccntra, is ectoparasitic on 

 a Jassid Homopteron which feeds on certain trees and is attended 

 and " milked " by an ant of the genus Iridomyrmcx. The ant car- 

 ries the parasite but not the Jassid into its nest. There the former 

 spins a temporary cocoon and later emerges from it as a peculiar, 

 flat, bright red (symphilic color), second stage larva, with two 

 long tails. In this stage it subsists " solely on the ant grubs by 

 sucking out their juices," but as in the case of Lomechnsa in the 

 nests of the European Formica sanguinca, the ant is partially re- 

 compensed for the loss of its brood. Dodd says : 



Reference has been made to the caterpillars raising their terminal seg- 

 ments, even the small ones from the cocoons doing so. This was quite suffi- 

 cient to warrant investigation. Consequently at various times I have placed 

 them with ants and grubs under glass, in order that they could be seen to 

 advantage and without risk of disturbance. When the anal parts are pro- 



