366 
Moreover it is desirable to apply this same method of projection to 
the dorsal and ventral body-wall, in such a sense that the circum- 
ference of the figure no longer corresponds to that of the sideview 
of the animal, but roughly forms a trapezium, the upper- and under- 
side of which represent the median dorsal and ventral lines, or better 
still, stretch a little over them. This can also be expressed in such 
a way, that the body becomes compressed from left to right, thereby 
erowing higher in the dorsoventral direction, a condition so often 
shown in reality by squeezed specimens. As to the position of the 
wing-sexangles in regard to each other, I am of opinion that 
they should “be placed in one and the same horizontal body-axis, 
the one behind the other, instead of the anterior margin of the 
hindwing passing beneath the posterior one of the forewing. In size 
the two wings may be represented alike. The arguments for both these 
assumptions can be found in the wings of Hepialids. 
Consecutively the following points should be attended to in the 
investigation : 
a. comparison of the markings of the thoracal with those of the 
abdominal rings in one and the same individual, therefore in the 
caterpillar with its different instars, in the pupa, and in the imago. 
b. comparison of the colour-pattern of all these stages, to each 
other in the same species. 
c. comparison of kindred species in their successive stages with 
each other. 
For each of these comparisons an example may be given. 
Of the few Sphingid-caterpillars which were at my disposition, I 
choose the fullgrown larva of Protoparce convolvuli (scil. the brown 
variety) as a fit object for comparison of thoracal with abdominal 
markings. For in this species the relation of the one to the other 
can be very clearly observed, and in doing so, we are impressed 
by the fact that the first seem to bear a more prisnitive character 
than the second. For on the thorax the obliquely ascending lateral 
bars, which are so characteristic of the abdominal segments of these 
as of so many other Sphingid caterpillars, are absent. The pattern 
is restricted to longitudinal light and dark stripes, which in their 
turn are evidently composed of rows of spots, whose number cor- 
responds to that of the annuli or secondary rings, which enter into 
the composition of each body-segment of the caterpillar. The number 
of these annuli is constant, eight for each segment of the abdomen, 
except the posterior two; it is likewise diminished in the thoracal 
ones, probably an effect of reduction. On each annulus one row of 
light spots on a dark ground is seen. The relative size of the single 
