110 Annals Entomological Society of America  [Vol. VII, 
a good deal in some cases, which made them more difficult to 
see in a dim light, but brought them out in contrast with the 
fat. For the work with the thorax it is important to open 
up the caterpillar and pin it out before hardening, as other- 
wise many of the muscles will be broken. If specimens are to 
be preserved whole for dissection purposes they should be 
killed in hot water, or injected with a stronger solution of 
preservative to prevent decay. The prothorax and last seg- 
ment are particularly hard to get ina satisfactory condition, 
because of their peculiar shape, and the close connection 
of the former to the head, which should be split vertically 
when opening the caterpillar. It 1s most convenient to open 
the caterpillar near, but not quite on the middorsal line. 
The viscera and loose fat of the body cavity are removed 
as a preliminary, as well as the wings of the heart, leaving the 
heart, trunk trachea and nervous system as long as possible 
for landmarks. The dissection can follow the order given by 
Lyonet, to advantage, but it 1s often unnecessary to open any 
specimens by the venter, as the dorsal musculature is com- 
paratively simple, and is often uninjured on one side. Plate 
XIX follows this order fairly closely in its six stages, but in 
other plates less stages are shown for economy. 
It soon appeared that the muscles represented by Lyonet, 
for the Goat moth caterpillar, could be found in such 
widely divergent forms as a Sphinx, a Noctuid and a Lasio- 
campid, so his lettering is used in the figures. Lubbock’s dis- 
section was hardly as perfect, while Berlese’s figures are wholly 
diagrammatic and useless for the study of homologies; in fact 
several of his comparisons with the muscles of other insects are 
invalid. We will only be able to come to a true knowledge of 
the homologies when the nerve-muscle relations are fully worked 
out; and in the Lepidoptera the matter is much confused by the 
anastomoses between all three pairs of nerves, and even between 
successive segments. 
THORAX. 
To go on to details; I have not made any satisfactory dis- 
section of the prothorax. Evidently however, Lyonet represents 
the state of affairs much more accurately than Berlese. The 
dissection figured on Plates 1 and 2 may be taken as accurate so 
aa 4 
