78 BRITISH BUTTERFLIES. 



depressed portion of the crown are simple, being without raised knobs 

 at the junctures ; those surrounding the side are prominent but 

 diminished on nearing base ; the cells between the reticulations have 

 a fine granular surface (Frohawk). Described from photographs in 

 several positions x 20, the width is found to be 0'42mm., the height 

 0'2mm. The sides are very upright, making the width of top or base 

 comparatively little less than the maximum width at the equator. 

 The arrangement of the cells of the white surface coat is largely 

 " engine-turned," and, hence, many of them are quadrangular; they 

 are about O03mm. in diameter at the margin and sides, but smaller 

 towards the middle of the top, ceasing at the margin of the micropylar 

 depression, which is about 0-07mm. across, and a little sunk, owing to 

 the want of the adventitious coat, but apparently with network of cells 

 somewhat continuous. At the angles of the cells are low eminences 

 rather than knobs or pillars ; they hardly exist on top except at the 

 margins. On emerging the young larva eats out a good section of the 

 upper surface (Chapman). [See also Scudder (antea, p. 72).] 

 Edwards describes the egg of comyntas, Can. Ent., viii., p. 208. 



Habits of larva. — The habits of the summer-feeding larvae have 

 been described by Frohawk, who states that the young larva (hatched 

 July 30th, 1904) makes its exit by eating the micropylar area as well 

 as a portion of the side of the eggshell. The larvae are very active when 

 first hatched, crawling rapidly for such small creatures and feeding on 

 various parts of the plant ; one was found half buried in the end of a 

 stem which-had been cut off ; they also feed greedily on the seeds of 

 Medicayo lupuliiia, eating through the capsule and devouring the con- 

 tents, and appearing to feed mostly at night ; later, after the third 

 moult, they appear to feed as much by day as by night, but are much 

 slower between the moulting — the first moult occurring on August 3rd, 

 the larvae four days old, the second August 7th, the third August 12th, 

 whilst the fourth and last did not take place till August 23rd. One 

 larva, in its last stage, completely ate up the whole of the petals of 

 a red clover blossom in two days; it appeared to be feeding continually, 

 day and night, the whole time. During the earlier stages the larvae 

 are distinctly cannibalistic. One larva, after its first moult, was 

 observed feeding on another,* but cannibalism was not noticed in the 

 last two stages. These larvae all pupated in due course, and produced 

 imagines in September (Frohawk). The habits of the larvae were 

 described in much greater detail by Zeller, who observes that the 

 newly-hatched larvae, in confinement, ate the upper cuticle and paren- 

 chyma of the young leaves of }dedica<jo falcata and Lotus corni- 

 culatuH, leaving the lower cuticle untouched, rarely eating through 

 the thickness of the leaf, the spaces thus eaten being small and 

 several occurring on one leaf ; after the second moult, which occurred 

 in from four to five days ('? after the first), the little larvae readily 

 took to the leaves of Trifolium repens, which they also ate from the 

 upperside and somewhat skeletonised, one caterpillar resting peace- 



* Frobavvk's details of this (Ent., xxxvii., p. 245 et seq.) must be taken with cau- 

 tion. On p. 246, he states that "all the eggs hatched on the same day, July 30th, 

 1904," and that " the first moult took place on August 3rd." On p. 248 he stated that 

 he " noticed one larva, after the first moult, feeding on a newly-hatched larva, 

 which it seized as it emerged from the egg," which is very wonderful, if both 

 statements be carefully studied together. 



