28 [July, 
larva moulted, then assumed a scattered appearance, this change in its 
arrangement arising from the larva jerking its posterior segments about 
each time it deposited a pellet of " frass." 
The larva continued feeding for ten days longer, by vi^hich time the 
dorsal vessel lost its green appearance, but could be traced by its faint 
brownish outlines. 
As the larva lay in the clear space which it makes in the mine 
previous to quitting it, it appeared, judging from the movement of the 
jaws, to be eating something, but though I watched it very closely with 
a lens, I failed to detect the exact nature of the substance it was ap- 
parently swallowing. The larva then being quite full fed bit a hole in 
the leaf, which hole was exactly of the same shape and form as the front 
and sides of the head of the larva, and gradually drew its body out of 
the mine. The larva was then exactly four lines in length, and rather 
more than half-a-line in width. 
(The larva does not always spin up on a leaf or twig, nor even 
amongst the leaves on the ground, several instances having come under 
my notice in which the larva has penetrated the damp sand in the 
breeding jar to the depth of more than an inch, and there spun its 
cocoon.) 
The larva under observation, directly it had crawled on to the sur- 
face of the leaf, twisted and tossed itself about, as if not knowing what 
to make of its new position ; these gambles occupied some little time, 
after which it set about the more serious work of fabricating its cocoon, 
and began carpeting with silk the part of the leaf on which its body 
lay curled up, twisting and twining itself about in all kinds of ways, 
whilst constructing the flooring of its cocoon ; when this was done, the 
larva still keeping its body in the same cramped position and moving its 
anterior segments backwards and forwards, gradually threw a number 
of silken filaments over its body, fastening them to the sides of the 
cocoon flooring, thus imparting a slightly convex form to the cocoon ; 
this formed only the frame-work of the cocoon, but the larva set 
vigorously to work and continued spinning additional threads till, after 
nearly nine hours of assiduous labour, it had entirely completed its 
cocoon. 
I find that a larva, if ejected from its cocoon at this period of its 
existence, is not only unable to spin another cocoon, but does not even 
possess the power of entering the pupa state, and, after lingering a few 
days, ultimately dies. 
The body of the larva, curled round as it was, occupied nearly the 
whole of the interior of the cocoon ; a fortnight later its body had so 
