NYMPIIALINAE : ARGYNNIS CYBELE. 561 



iug many full-grown caterpillars. The period of the chrysalis is from 

 fourteen to sixteen days according to Saunders, sixteen to twenty accord- 

 ing to Edwards, and he gives one instance where it was twenty-four days. 



Life-history. Tlie insect is single-brooded in New England, passing 

 till' winter in the larval state. The caterpillars become full-grown in 

 eTiuie, and the earliest butterflies appear in the latter part of June, some- 

 times as early as the 16th in the latitude of Boston, usually not much 

 before the 21st, become common by the first of July when the female first 

 emerges, continue to escape from the chrysalis until at least the middle of 

 July, and fly until the middle of September and occasionally later. The 

 butterflies generally pair at the end of July, but the eggs do not begin to 

 assume their proper size until about the middle of August, and are not 

 laid until the last of August or first of September. Miss Soule obtained 

 eggs in Stow, Vt., on August 20, which is the earliest New England date 

 known to me. "It is evident," writes Mr. Edwards, speaking from ex- 

 perience, "that these butterflies take their time for laying their eggs, 

 instead of doing the work all at once as moths do, who die from exhaus- 

 tion afterwards." 



Mr. W. H. Edwards now regards the species as probably digoneutic in 

 the south. There the sexes may be found pairing late in June or early 

 in July, and the early brood of the butterflies, to quote his own words. 



Appears in great force here [W. Va.] by 1st June, on the clover blossoms, first the 

 males, and in a few days the females. After the 15th to 20th June they disappear, 

 and in July I scarcely ever see an example. By 15th August fresh males appear again, 

 and soon after fresh females, and I can always obtain eggs between 1st and 20th Sep- 

 tember. ... I sliould not have doubted there being two broods were it not for the 

 fact that the several stages of the larvae which feed in spring are so remarkably pro- 

 longed that it seemed unlikely that between 15th June and 18th August the several 

 stages of egg, larva and chrysalis could be passed; and furthermore, that I had repeat- 

 edly dissected females ... in June, and when I could obtain them, in July and first half 

 of August, and never yet found the least appearance of a formed egg. . . . But suddenly, 

 about the middle of August, the eggs begin to take shape, and in a week or ten days 

 are ready to be laid. But the hot weather of July and August, the mercury constantly 

 running between 80° and 95° in this region, and the nights (in which tliese larvae prin- 

 cipally feed) being warm, may accelerate all tlie preparatory stages." (Can. ent., ix : 

 35-3G.') 



This phenomenon has been discussed under the genus. 



The eggs hatch, as has been said, in about fifteen days, but the cater- 

 pillars from them go immediately into hibernation without eating anything 

 more than their eerg shells. 



Mr. Edwards was able on one occasion to carry some of the larvae 

 through the winter in a cool room in West Virginia, and theu- growth 

 during the early spring months diflered curiously from that of others 

 which had been kept over winter (five months) in an ice house, as shown 

 bv the following: table : 



