8 
just before the transformation to chrysalis. At the time of the trans- 
formation, however, a bluish-white color is assumed. Although the 
size of the younger larve is not greater than that of the Bark-lice which 
they attack, they spin no silk at first and the spinnerets are rudimen- 
tary. Assoon as the contents of one Bark-louse have been devoured 
the caterpillar abandons it and seeks a new prey,and is sometimes 
found exposed upon the twigs. Apparently realizing the necessity for 
protection, as well as additional food, it hastens to bury itself in a new 
victim, which usually is a matter of but some minutes. After having 
entered a Bark-louse it seems to be effectually protected from parasites. 
After ten days, and after having passed through several molts, the 
exact number of which has not been observed, the larva begins to con- 
struct a silken covering spun from the spinnerets, which make their 
appearance only at this late stage. This covering enlarges the case 
formed by the coccid shell and to it are attached excrement and the 
débris of the Coccide. This artificial case the larva now carries about 
with it, during its final stage. It has left an opening for the head 
and feet, and in fact for the larger part of its ventral surface and crawls 
about, devouring several Coccide every day. The covering is held 
in place by the anal prolegs, two pairs of abdominal prolegs being func- 
tional. As the larva continues its destructive work it is with difficulty 
perceived. Hidden in this case, which it never leaves, and from the 
anterior of which only the head, which is coriaceous, small, and blackish, 
issues, it is very difficult to distinguish it upon a surface covered with 
smut fungus, as twigs infested by black scale usually are. The effi- 
cacy of this protection is indicated by the fact that, although several 
hundreds of cocoons of Erastria were obtained by M. Rouzaud, upon 
only one occasion did parasites emerge instead of the moth. In this 
instance 50 small Chalcidids were reared, and so far as he is able to 
judge, the species is identical with one which he reared from a Coc- 
cinellid (Hxochomus quadripustulatus). In this statement, however, 
we are inclined to think that M. Rouzaud is mistaken. 
Upon arriving at full growth the larval case has become enlarged far 
beyond the dimensions of the original Bark-louse, which in fact now 
occupies a superior position on the anterior end of the case. The case 
has been built out with silk, excrement, and the fragments of eaten bark- 
lice to a distance of perhaps four times the length of the full grown 
Lecanium and nearly three times its height. The larva then searches 
for a favorable position in which to fasten its case, prepare its cocoon, and 
transform to pupa. It chooses a leaf very rarely and is more often found 
in the angle formed by two branches, or in large crevices in the bark of 
the trunk. Perhaps most frequently it is at the collar of the trunk, 
where the bark is roughest, that the cocoon is found. The cocoon is 
formed by increasing the silk layer and completing it beneath the larva, 
first clearing the surface upon which it is to be fastened by means of 
the mandibles. Five or six cocoons are sometimes found together on the 
. 
