of the Amazon Valley. 179 
characters at all constant for the formation of groups. There seem 
indeed to be two distinct types of larve, namely, that of the Apatura 
group, distinguished by the smooth skin and bifid head, which will 
connect together Apatura, Charaxes, Prepona*, and a few other 
genera; and that of Argynnis, Vanessa, and Limenitis, in which the 
body is studded with branched spines. This latter or Vanessa type 
seems again to be divisible into two, in one of which the head is 
bifid and surmounted by two long branched spines (Epicalia, Cal- 
lithea, Diadema, Ergolis, &e.), and in the other rounded and spine- 
less (Argynnis, Vanessa, &c.). 
We are now acquainted with the transformations of thirty-eight 
genera of Nymphaline, and are thus enabled to test the value of 
larval structure as a systematic character. The result is that the 
value is very small; for cephalic spines reappear in members of the 
Apatura group (Siderone, Paphia) and of the Argynnis group 
(Colenis, Agraulis), thus weakening the importance of this charac- 
ter; and again, one genus, Limenitis, shows a great diversity of larval 
form between allied species, as may be seen on comparing figures of 
the larvee of LZ. populi and L. Sibylla. Moreover the larve of some 
genera are of quite an aberrant character, supplying no clue to the 
affinities of the imagos; such are those of Timetes (7. Petreus figured 
by Stoll), and Adolias Acontea (figured by Horsfield). 
If we turn to the perfect insects, we find the intricacy of relation- 
ships almost as great as that revealed by the larvee; for whether we 
take the form and style of colouring of the wings, the neuration, 
shape of antenne, palpi, or legs, we cannot discover amidst the great 
diversity which exists in all these parts, any constancy of form 
within any group of genera that may serve as points of distinction 
from other groups of genera. Thus, although Apatura and Prepona 
agree pretty well as to their larvee and the form of wings of the 
imagos, they differ greatly in neuration, and to some extent also in 
the antenne. Stderone, Protogonius, and other genera, again, belong 
to the Apatura group in their larval condition, but differ much both 
from Apatura and Prepona in their neuration. At the opposite end 
of the subfamily we find a large number of genera allied to Argynnis 
which agree in the shape and clothing of the palpi and in style of 
coloration, and we seem to have here a natural group; but a 
feature exists in some of the members of it (Colenis, Agraulis, 
Clothilda), namely, the closure of the fore wing-cell by a perfect 
* The transformations of Prepona have not yet been made known; I have, 
however, bred P. Amphimachus and possess drawings of the larva and pupa. 
