61 
example of the former, and Geometra papilionana of the latter; and 
it is remarkable that with butterfly larva*, as Dr. Dyar has pointed 
out, and also with G. papilionaria, the shagreen or mammillary based 
hairs usually occur as a secondary growth, making their first appear¬ 
ance after the first moult, as is the case with the Sphingids other than 
Smerinthus. Bifid or forked hairs are also to be met with among the 
butterfly larvae, Leucophasia sinapis, Neincobius Incina, and Chryso- 
phaniis phlaeas being examples, as well as among the Sphingid larvae, 
and this character is also paralleled among the Geometrids by Geometra 
papilionaria, a further illustration that similarity of structure goes 
with a similarity of habits, and presumably of needs. 
These, among many other examples, had so influenced my views 
since writing on the relationship of Dimorpha versicolora to the 
Sphingids, that when the problem of the origin of the last-named 
group was raised by Mr. Tutt, in view of the approaching publication 
of the third volume of his work on The British Lepidoptera, it was by 
no means clear to me whether the Sphingids were to be considered as 
a homogeneous group arising from a common Sphingid ancestor, or if 
they might not be the result of parallel development such as we find 
between the imaginal stages of Si/n tom is phetjea and the Zygamids, or 
between Arsilonche alhorenosa and the Calamias, Leucanias, Nonagrias, 
and Tapinostolas, which make up the group called “ Wainscots ”—a 
parallel development, or perhaps convergence, that is, of different 
stirpes related to each other, and at some very distant base, of course, 
springing from a common ancestor. But the common ancestor was 
in no sense a Sphingid, and gave rise to groups as widely divergent, 
say, as an Ay lata tail, Dimorpha ( Kndromis ) versicolora, Bombi/x mori, 
and Tli/perchiria i<>, which subsequently converged and developed on 
parallel lines to form the apparently homogeneous group of moths we 
call “ Hawks.” 
Evidence of breaks or faults in the larval characters of the 
Sphingids is not wanting, and one at least occurs with regard to the 
minuter larval characters of the 1st instar, characters that, to my 
thinking, are less likely to be the results of convergence than the more 
striking and obviously useful ones. Some discussion and correspond¬ 
ence that 1 had with Dr. T. A. Chapman, however, convinced me that 
there was in reality no doubt of the Sphingids being a truly homo¬ 
geneous family (see Dr. Chapman’s remarks on the antenna? of the 
Sphingids, in Tutt’s British Lepidoptera, iii., pp. 377-378) with a 
common descent. But the evidence of a rift within the group 
remains, and to my mind points to the sub-division of the European 
species of the superfamily into two main branches. Whether exotic 
material would show others of equal value, or if all the species would 
fall into one or other of my main divisions, remains to be seen. 
Although the Sphingid larva presents a high degree of specialisa¬ 
tion in the form, coloration, markings, and in some points of structure 
— such as the conjunction and raising of the bases of tubercle i on 
the 8th abdominal segment into a more or less developed horn bearing 
the two tubercles and their hairs on its summit, the shape of the head 
in some species, and in the tendency to have either a uniform coat 
of generally distributed fine short hairs, or to become completely 
naked—it still retains in all groups other than Smerinthus a more or 
less primitive 1st skin in which the single-haired primitive tubercles 
