Notes in 1887 upon lepidopterous larve, éc. 569 
of these characters in the early stages, and their subse- 
quent diminution or complete disappearance, in both 
Aglia and Sphingide, is strongly opposed to such an 
interpretation. 
I have examined a large number of figures of the 
larve of tropical Sphingide, and of genera allied to 
Aglia, in order to see if further light can be thrown 
upon the subject. 
The results of this comparison lead to the following 
conclusions. The tubercular covering of Saturnian 
larvee is frequently developed into a spinous covering, 
especially in the genus Attacus, the change being 
doubtless protective. There is a tendency towards the 
special development of a single median dorsal spine 
on the 8th abdominal segment, and of several spines 
upon the thoracic segments. As these spines increase 
in size, they usurp the protective functions of those 
upon the rest of the body, which begin to lessen in 
size. The most complete transition between the extremes 
can be made out. As the thoracic spines become more 
important, further specialisation takes place, and the dor- 
sal pair upon the 2nd and 8rd thoracic segments become 
far more important than the others. Finally the speciali- 
sation reaches a pitch equal to that attained by Aglia, 
with four large thoracic spines, and the large caudal 
horn, all the other elements being reduced to a minimum. 
There is, however, this difference, that the anterior pair 
of thoracic spines is upon the prothorax of Aglia, and 
upon the mesothorax of the species of Ceratocampa 
(Attacus). Furthermore, the five chief spines of certain 
species of Attacus, do not possess the forked terminations 
and certain other details which unite Aglia so closely to 
the Sphingide. In one species (Rhescynthis erythine), 
evidently closely allied to Ceratocampa, and placed 
immediately after it by Burmeister (‘ Atlas of the Lepid- 
optera of the Argentine Republic.’ Plate XXI., figs. 1 
and 1 a), the young larva (fig. 1 4) has a distinct caudal 
horn, and a smaller horn behind it (probably upon the 
9th abdominal, like the anterior of the bifid tubercles 
alluded to in Aglia), while the four spines upon the 2nd 
and 5rd thoracic segments are immensely large. All the 
six spines are thorny. Noother spines are shown in the 
figure, and if present, they must be quite rudimentary. 
The older larva (fig. 1) has lost all the spines, just as Aglia 
