ac Fe 
1914] A Structural Study of Caterpillars. nee 
from those resulting directly from the slender body, are the 
setension of F to the acrosternite, under a and ¢, the extension 
of d well into the following segment outside the other ventral 
muscles and the simplified leg-muscles of A8, correlated with 
the much reduced proleg; the most noticeable point in this is 
that while not forming a longitudinal series of transverse 
fibres as they do in the seventh segment, muscles m and q are 
represented by a series of several evidently homologous fibres. 
The nervous system also shows the highest specialization I 
have seen in a caterpillar; as the first abdominal ganglion 
is moved forward well into the thorax (Pl. 1, Fig. 1) and the 
fused last ganglia have also moved forward into the preceding 
segment and nearly fused with the sixth abdominal. Their 
nerves, however, are normal, except for the oblique direction. 
Lubbock has notes on a few other species in his paper, 
among them a Pieris. 
SUMMARY. 
1. Corrosive sublimate produces the most perfect material 
for dissection of the muscles; four per cent. formalin is more 
generally satisfactory. 
2. The caterpillars should be opened out before hardening. 
3. Lyonet’s work has proved fully satisfactory, and 
Lubbock’s sufficiently so for almost perfect correlation of 
the two. 
4. The- meso- and metathorax are alike, the first eight 
segments of the abdomen much alike, and capable of general 
correlation with the thorax. 
5. The ninth abdominal segment is much reduced, but an 
unquestionable segment. It is less reduced in the most gen- 
eralized form. 
6. The first and last segments are of a different type, and 
probably compound in nature. 
7. The first spiracle has- moved forward, the second 
has become rudimentary in situ and the other eight have 
moved back, but the first rather less than the others. All 
have the same innervation, in the abdomen largely from the 
preceding segment. 
8. The meso- and metathorax alone have muscles traversing 
the nerve-cord. 
