SEASONAL POLYMORPHISM. isl 
The ordinals in Roman figures represent the various grades in the 
extent of the orange lunules on upperside, which cannot be indicated with 
more accuracy, by stating the actual number of lunules present, because 
their size varies more than their number and has more importance: 
in the look of the insect. In this, as in the following characters, it 
- should be understood that the grade indicated is that of the medium 
individual variation, but that the latter is, as a rvle, so 
extensive as to include various grades. The different variations of the 
underside of the wings, which constitute the leading characters of the 
various races, are represented by letters and the different grades 
of each by arabic figures, beginning by O where the character is 
absent: A—size of black dots; B=brightness of fulvous tinge of 
underside in the single generation of artaxerwes and of salmacis, and 
in the second and third generations of the other races; the gray 
colour, with which it is generally mixed, may be more or less dark, or 
even absent in the various individual variations ; C=medium of the 
different individual variations in the intensity of the gray of under- 
side; the total absence of gray (C,) as well as of fulvous (B,) does not 
figure in the table, because this combination only occurs in extreme 
individual forms, exactly as the extreme nymotypical form of 
infracandida does not occur as a race, because the medium intensity of 
the race corresponds to grade infralbens (C,); in the case of race 
sarmatis and in that of race infracandida the grade of the first brood 
figures in brackets in the Table because the following broods are 
unknown to me, and in the case of race pallidefulva that of the Il. and 
Ill. brood is also in brackets, because in this race seasonal 
dimorphism is more marked than in the others, in this particular 
character (C), and the first brood has a different (C,) grade of 
gray, so that the grade indicated does not refer to all the broods, as in 
the case of the other races. 
This Table shows clearly that the vast majority of the more 
widespread races consist in gradations of only one series of variations, 
from medon to infracandida, similar to those of A. thetis, of P. icarus, 
etc., and that only a few local races stand apart. Experimental 
breeding would probably show, as in the case of R. phlaeas, that the 
series only consists of ontogenetic races produced by the direct action 
of surroundings on individual elasticity, or perhaps that a slight 
degree of hereditary differentiation exists between the extremes of the 
series; a positive fact is that no sign of a ‘transverse scission ”’ 
is perceptible at any point of the series. On the contrary scissions are 
clearly discernible, which might, by a material image, be called 
“longitudinal ’’ in respect of the direction of variation just mentioned, 
and which cause a division of the two extremes of the series in 
“eollateral’’ divergent branches. That artaverves is not simply 
produced by a further continuation of the variation which leads up to 
medon seems to me clearly shown by the fact that it is never found as 
extreme individual variation even in localities where form medon is 
most abundant as compared to the other forms, and by the fact that 
the fulvous colour is often more accentuated than in the latter both by 
the extent of the lunules and by the tinge of underside, which can 
reach the grade of agestis and even of aestiva; at this level in 
consequence one suspects that the branch of semiallows and medon, and 
that of salmacis and artaweraes have separated. It must be noted that 
the disappearance of the black dots below cannot be entirely due to a 
