738 THE BUTTERFLIES OF NEW ENGLAND. 



side ; if extended horizontally, to the lower surface. If, however, it has 

 been bom late in the season on a flower pedicel it attacks the flowers them- 

 selves and eats down into the ovaries at first, not attacking the leaves until 

 later. The caterpillar eats voraciously, and ordinarily matures rapidly. 

 Dr. Harris says it lives but a fortnight in the larval state, casting its skin 

 three times ; the latter statement is corroborated by Eiley, but contra- 

 dicted by Edwards, whose observations entirely agree with mine, that 

 there are four moults. Sometimes, however, it takes it three or more 

 weeks to attain its growth. Looked for toward evening it will ordinarily 

 be found quiet, apparently resting for the night, planted on the under sur- 

 face of the midrib of a leaf, half way betAveen the base and apex, its head 

 outward : from this it might appear that it fed only by day, but some 

 caged specimens certainly did eat at night, and I have found it resting early 

 in the morning on the top of the leaves on a cloudy day ; it is almost 

 always found near the top of a plant, and when disturbed, so as to be 

 knocked oif a leaf, the caterpillar coils like a galley worm. Dewitz, 

 writing of the larva in Venezuela, says it spins a thread on being seized, 

 but I cannot understand the statement ; it spins less thread than almost 

 any caterpillar known to me. When walking, the anterior filaments are 

 alternately moved forward, so far that they nearly touch the ground. 

 While feeding they are nervously twitched backward and forward ; and 

 when the caterpillar is alarmed, the movement, though no more rapid, is 

 of nmch greater violence, the filaments nearly touching the body. 



My attention was attracted one morning to one of these caterpillars 

 while moulting its skin : it had been stationary at least twenty-four hours ; 

 and noAv first began swaying its body from side to side, falling over so far 

 that the thoracic filament of the upper side became perpendicular, and 

 then drawing itself forcibly back to an opposite position ; the muscular 

 eflPort caused a considerable indentation along the falling side of the sway- 

 ing larva at the point where the white band widens, and at which muscles 

 are attached. The motion was repeated about once in three seconds and 

 continued for nearly three-quarters of an hour ; now and then the larva 

 would violently shake its filaments or strain forward the front of the tho- 

 racic segments, thus gradually detaching the old skin from the new ; at 

 last, after remaining quiet, as if to gather strength for a final eflPort, it 

 began to make violent contortions, especially about the thoracic regions, 

 which at first seemed ineffectual, but suddenly the integument parted 

 between the head and body, and, by the movements of the larva, passed 

 backward over the new skin, slipping over the whole body at once and 

 leaving a little empty pellicle at the hinder extremity. The skin was Avith 

 difficulty removed from the filaments, especially from one whose tip had 

 been bent in the former stage, and which only parted after strong exer- 

 tions ; the fresh filaments lay limp along the back until they were 



