20 IMPERFECT SOCIETIES OF INSECTS. 



ness of the ground, is sometimes deserted by them : it 

 is however attempted by others with success, unless it 

 happens to roll into some deep hollow chink, where they 

 are constrained to leave it; but they continue their 

 work by rolling off the next ball that comes in their 

 way. Non'e of them seem to know their own balls, but 

 an equal care for the whole appears to affect all the 

 community 11 ." 



Many larvae also of fyepidoptera associate with this 

 view, some of which are social only during part of their 

 existence, and others during the whole of it. The first 

 of these continue together while their united labours are 

 beneficial to them ; but when they reach a certain period 

 of their life, they disperse and become solitary. Of this 

 kind are the caterpillars of a little butterfly (Melitcea 

 Cinxia) which devour the narrow-leaved plantain. The 

 families of these, usually amounting to about a hundred, 

 unite to form a pyramidal silken tent, containing several 

 apartments, which is pitched over some of the plants 

 that constitute their food, and shelters them both from 

 the sun and the rain. When they have consumed the 

 provision which it covers, they construct a new one over 

 other roots of this plant ; and sometimes four or five of 

 these encampments may be seen within a foot or two of 

 each other. Against winter they weave and erect a 

 stronger habitation of a rounder form, not divided by 

 any partitions, in which they lie heaped one upon an- 

 other, each being rolled up. About April they separate, 

 and continue solitary till they assume the pupa. 



Reaumur, to whom I am indebted for this account, has 

 also given us an interesting history of another insect, the 



* Catesby's Carolina, ii. 111. See above, VOL. I. 51. 



