796 THE BUTTERFLIES OF NEW ENGLAOT). 



The caterpillars are known by their resemblance to wood-lice, whence 

 the term "onisciforni" which has been frequently applied to them; they 

 are naked or pilose ; their legs often extremely short, so that they resemble 

 slugs in their sliding movement. The generally minute head, almost 

 always wholly retractile within the greatly enlarged first thoracic segment, 

 the coalesced condition of some terminal joints of the body, and the atro- 

 phied legs — features shared in part by the Lemoniinae, — are here devel- 

 oped to such an unusual extent as to distinguish these caterpillars from 

 those of all other groups. 



They feed upon various exogenous plants, some groups showing a pre- 

 ference for trees and shrubs, others for annuals ; a single species will often 

 have a great range of food. Indeed not a few of them are not only 

 polyphagous, but in stress at least will even devour their neighbors 

 whether of their own or another species. Thus Thwaites remarks (Moore, 

 Lep. Ceylon, i : 70) : 



It is difficult to realize that the larvae of some species of these lovely Lycaenidae, 

 such as Amplypodia, etc., are carnivorous or even cannibal in their haljits, and do not 

 hesitate to eat their ovpn brethren of the same brood, when any of the latter are com- 

 mencing their change into the inactive chrysalis state, with their consequent inability 

 to protect themselves from their voracious kindred, who devour them with avidity. 



The same carnivorous tendency has been observed in several of our 

 own species ; one, indeed, feeds exclusively upon plant lice, and it is 

 not improbable that the habit may be shared by others in other parts 

 of the world, as Dr. Holland has su'jg'ested. This habit seems the more 

 curious, as both plant lice and caterpillars of at least one group of Lycae- 

 ninae are often accompanied by ants, who seek them and protect them for 

 the same purpose, namely to lap the sweet exudations which flow from 

 special organs at the hinder extremity, of which further details will be 

 given under the Lycaenidi. Thwaites, in the passage aboA'e quoted, goes 

 80 far as to say that the ants actually protect them from their cannibalistic 

 brethren. Nature, says he, 



finds a protection for these said helpless individuals, in the instinct of a species of 

 ant (Formica smaragdina Fabr.) which, finding a substance most palatable to it 

 secreted naturally from a glandular defined spot upon the bodies of these helpless lar- 

 vae, talces possession of them as "cows," surrounding each separate one and the leaf 

 on which it had been feeding with a'few silken strands of its web, protecting tliem 

 jealously and attacking most fiercely any living thing intruding upon them. 



Whatever the value of this statement, we must confess to a doubt 

 whether the ants spin the web seen. The secretions which are thus 

 attractive to ants are exuded from an evaginable vescicle in the middle of 

 the dorsum of the seventh abdominal segment, which appears to be found 

 in the caterpillars of all Lycaenidi, but only in some of the other tribes, 

 though only a few of them (and these all Lycaenidi) are known to be 

 attended by ants. Some of the caterpillars which possess this vescicle, 



