62 
may be clearly observed. Even in Smerinthus they can be easily dis¬ 
tinguished under a lin. objective, owing to their overtopping the 
smaller secondary hairs, which in Smerinthus alone are present on the 
general body area in the 1st instar. The arrangement of the dorsal 
and lateral tubercles, while agreeing in its general aspect with that of 
other lepidopterous larva?, presents some special features. Briefly, the 
arrangement is: i and ii in trapezoidal position, iii supra-spiracular, 
iv beneath the spiracle, and v normally shifted up in front of spiracle 
on the anterior margin of segment. In one or two sj>ecies, however, 
v on the 1st abdominal segment is clearly situated on the lateral 
flange, only slightly above the plane of iv, and decidedly sub- 
spiracular, while in Hi/les euphorbiae tubercles iv and v of 1st ab¬ 
dominal segment are conjoined, the two setae being situated on a small 
oval plate beneath the spiracle, so that there is no doubt of the 
identity of the prespiracular of the Sphinges with v of the larva? of 
other families. On the 2nd and 3rd thoracic segments, iv is wanting, 
iii bears two hairs on a small oval plate, and v is on anterior margin 
of the segment, as on the abdominals. This arrangement, so far as 
my present rather slender knowledge goes, holds good throughout the 
group, but probably only the position of v and the absence of iv on 
meso- and meta-thorax is likely to be characteristic of the Sphingids. 
When, however, we come to the dorsal tubercles on the last-named 
segments, we get the distinct line of cleavage within the group that 
has been referred to above. In h'uinorpha ( Choerocampa ), and also in 
Phryxids ( Deilephila ), the dorsals i and ii on the meso- and meta-thoracic 
segments are situated on separate sub-segments, and form the corners 
of an oblong or low trapezoid when viewed from above; while in all 
the other species I have seen of Sonia ( Macrorjlossa ), He mans (Sphinx), 
and Amorpha ( Smerinthus ), they are arranged as twin tubercles with 
conjoined bases (i + ii) on the same sub-segment. 
\\ ith regard to the tubercles of lepidopterous larva? in general, I 
am as yet very far from a sufficiently complete or exhaustive know¬ 
ledge to allow of my mapping out their full significance or phyiogeny, 
but I feel tolerably sure that they were more numerous on the ancestral 
larvae than with the existing forms, and that specialisation has largely 
been by way of reduction in number, as a general tendency, but often 
secondarily by way of an increase in size of base and multiplication of 
secondary setie on or around the individual primary tubercles. The 
more or less complete coats of minute spicules and secondary hairs 
springing from the general skin surface I am not here referring to, as 
they in all probability form a problem apart. The two processes 
above-mentioned of the evolution or devolution of primary tubercles 
may be in progress at one and the same time, as, for instance, in 
Saturniids and Lachneids, or Liparids and Arctiids; the former in¬ 
creasing the size of i, while reducing or losing ii, and the latter showing 
a waning tendency or want of equally vigorous growth on the part o'f 
i compared with ii. Beneath the spiracle, the tubercles not infre¬ 
quently show tendency to combine. Above it, on the alxlominal s <v/- 
ments, the tendency is not for combination, but for the atrophy of one 
or the other pair, while on the thoracic segments it seems to be 
towards approximation in a single transverse band, if not to actual 
combination. 
Now as to the value, for classificatory purposes, of the position of 
