LIMENITIS I. 
stitute the base of the perch. This is bound and lengthened with frass and 
serves every purpose. 
Both these species of larvae have a habit of accumulating little scraps of leaf 
at the base and under side of the perch till quite a packet is formed, and this is 
rolled back as the substance of the leaf is eaten so as to be close to the cut edsre 
O 
of the leaf. This edge, in willow, is kept nearly square, a section being eaten 
from one lobe and then a corresponding one from the other. In beginning on a 
fresh section, the larva lies diagonally across one corner, the anal legs clasping 
the base of the perch, and its head will strike the side of the leaf about two 
tenths inch above the corner. It eats a canal nearly perpendicular to the side 
and towards the midrib ; not all at once, by any means, for this is the result of 
several meals, in the intervals always returning to the perch. When the rib is 
reached, the larva then begins to feed on the lower side of the canal next the 
rib, and to keep the slender and 'unsteady bit of leaf in position it spins guys 
from the end and edge to the solid leaf opposite and to the rib. As the feeding 
proceeds and a considerable triangle is held only by a narrow strip, which dimin¬ 
ishes at each mouthful, more guys are put out, and, at last, when the triangle 
falls, it is held by the threads and swings to the base of the perch. If not, it is 
soon brought there by fixing one thread after another from it to the rib and leaf 
till it is pulled to its place. Here it is bound loosely. As other bits are added, 
there comes to be an open packet, held together by simple threads, and of about 
one tenth inch diameter. In the two younger stages this is moved along as the 
larva feeds, and is always kept close to the leaf, partly by pushing, what is gained 
at each effort being secured by threads, or it is rolled by attaching successive 
threads from the farther side to the leaf and rib till the mass is turned over. 
After the second stage the packet is left behind, and no additions are made to it. 
(Fig. Ji). I was at first puzzled to account for this construction ; but happening to 
see one of the caterpillars back down the perch and drop its excrement directly 
into the packet, it occurred to me that really this was the magazine whence the 
larva drew its materials for lengthening the perch. On pulling some of the pack¬ 
ets apart a few grains were always found in them. This I believe to be the use 
of the packets, and without some contrivance to catch the frass, it is difficult to 
see how the larva obtains the materials it uses. Apparently it drops just about 
enough into the packet for the object in view, for it is certain that the grains are 
usually expelled wherever the larva happens to be, and fall to the ground. After 
the end of the perch is sufficiently strengthened and there is no further need of 
the grains, the packet is dropped behind and neglected. 
The larvae of Arthemis hatch in from seven to nine days, undergo two moults, 
and construct, each for itself, cases or hybernacula of leaves in which to pass the 
