56 BKITISH BUTTERFLIES. 



triple-brooded in the southern ; the latter are continously-brooded in the 

 warm southern localities in which they live, and, spreading from 

 these centres, are triple- or double-brooded in the more northern parts 

 they reach, being annihilated in the last, i.e., winter, brood, and 

 relying on fresh incursions to keep any position in these areas. 

 The hybernating larvae of Colias philodice in their more southern 

 habitats in North America feed up rapidly and give imagines by 

 May, just as is the case along the Mediterranean littoral with C. hyale; 

 the larval stage of the next brood does not last more than about 

 eighteen days in the favourable lowlands of West Virginia to almost 

 double the time at a moderate elevation in the Catskill Mountains; 

 the larvae of the next brood leave the eggs in August and September 

 and hybernate when about half-grown, although some of these 

 attempt, under very favourable conditions, to complete the cycle as 

 imagines in November; in the more northern parts of its range it is only 

 double- brooded, exactly as is C. hyale in its more northern permanent 

 haunts in central Europe. The larval habits of Colias eurytheme are 

 almost parallel with those of Colias edusa. In its more southern haunts, 

 in the lowlands of California and Texas, under the most favourable 

 climatic and geographical conditions, the species produces " forward " 

 larvae very rapidly. The winter-larvae during the early part of their 

 hybernation appear merely to be in a sort of restless sleep, and, though 

 they may remain motionless for days, will, if breathed upon, start as if 

 alarmed (Fletcher); they commence feeding in the earliest spring, and 

 feed, as do the larvae of C. edusa, in its southern European habitats, at 

 different rates, so that the earliest imagines are out in Texas in March 

 and April, whilst slow-feeders of the same brood are somewhat later. 

 The species then goes on by a continuous series of broods, until the late 

 autumn, when the last lot of larvae divide into two sections : those 

 that get beyond the third moult passing on to the final imaginal stage 

 in November, tbe slower ones of the same brood going, as already 

 described, into partial hybernation, and producing the earliest spring- 

 imagines, so that, in North America, at different latitudes, we find the 

 species two-brooded in North Carolina, three-brooded in Illinois and 

 Nebraska, four-brooded in California and Texas, whilst Edwards' 

 remark that C. eurytheme is, in North Carolina, killed off in the 

 autumn, suggests that its conditions there are exactly those of 

 C. edusa when it penetrates into central Europe in the spring and gets 

 a summer footing in a suitable spot in a favourable season. 



The food-habit appears to be a very fixed one, almost all the 

 species choosing leguminous plants ; and the growth of these crops 

 for fodder, both in the Palaearctic and Nearctic regions, probably 

 explains the occasional autumnal abundance of certain species when 

 immigration takes place. Among our European species, Colias 

 palaeno offers an exception, being said to feed on Vaccinium uliyinosum, 

 whilst Colias scudderii also feeds on Vaccinium (Bruce), and willow, 

 the larvae refusing clover (Edwards) [Can. Ent., 1892, p. 54) ; 

 C. nastes is also said to live on willow, but in confinement feeds well 

 on Hedymrum, (Bean, Can. /Cut., 1892, p. 54) ; the larva of another 

 species, Colias bc/irii, feeds on a species of ground huckleberry 

 (Vaccinium) in the mountain meadows of the high Sierras of 

 California (Dyar, Can. Ent., 1893, p. 158), and yet, again, the 

 larva of Colias interior feeds upon Vaccinium (Lyman, Can. Ent., 



