SEASONAL POLYMORPHISM. 135 



presume this is very near the extreme northern limit at which a second 

 brood is produced. Compared with its own first brood differences do 

 exist : for instance, the male constantly has a large apical crescent and 

 a spot on disc, whereas these are usually absent in the first ; the under- 

 side streaks are greenish-gray, whereas they are much darker and 

 sharper in the first ; one male of my series has the slightly broader 

 wings and more convex outer margin, characteristic of summer broods ; 

 this, no doubt, is the race which produces the yellow form siilphtoea, 

 Schoyen, described from the extremely rare male, but more frequent 

 in the female sex. Its first brood seems to answer the designation of 

 race arctica trans, ad napi, Vrty.-L., but I have not enough material 

 at hand to make certain of it. 



Grade la conveys, to my mind, as well as possible, the relationship 

 of the following little group of races to the others. The latter consti- 

 tute, more or less, one series of grades along the same line of variation, 

 and their differential characters consist in the extent of the pattern, 

 and are thus purely quantitative on an average, but the races in 

 question in this paragraph are puzzling as to their position, because 

 they could not be placed either before or after Grade I. By both 

 surfaces of the male sex and of most females, as well, and by the 

 underside of all the females, they would most positively fall into Grade 

 III., but a considerable percentage of females exhibit on upperside a 

 remarkably greater extent of pattern than is normally found even in 

 Grade I., and a bright yellow ground colour, which is very clearly the 

 equivalent of form bnjoiuae, O., of the first generation. Here, as in 

 that case, the brijuniae-like individuals and those of the form stand- 

 ing opposite to it, by its white ground-colour and by its limited dark 

 pattern, constitute two groups pointing to dimorphism, whilst inter- 

 mediate forms are quite scarce. In the case of the first brood, however, 

 the dark streaks of underside are, as a rule, proportionately as extensive 

 as those of upperside in the bnjo)iiae forms, and these fall in, quite 

 naturally, as the culminating grades in the extent of the pattern. In 

 the cases of the second and third brood, on the contrary, the bnjoniae 

 features decidedly give an impression of abnormality, such as of atavism 

 of some female individuals, making its appearance in races which 

 ■would otherwise have a very natural position in Grade III., and would, 

 in fact, be identical with the races I will describe in that paragraph. 

 That is why I consider the races I am dealing with in this paragraph 

 as a collateral variation to the main line, or in other words, as a sort 

 of dimorphism amongst races. The doubt I have is whether it should 

 not be called grade IIIc/, bat the analogy to bnjoniae and the consider- 

 able average extent of dark pattern in the female sex on upperside, 

 owing to the bniuniae-Wka forms, have decided me to call it grade la, 

 because it must evidently be more closely connected with the first 

 generation than are the races described as grade III. 



'RoiCe fiavescens, Wagner, Verh. zuol.-bot. Ges., liii., p. 174, pi. I., 

 fig. 1 (1903), is, I think, the name which should designate the one of 

 Modling, near Vienna, whose first generation I have discussed under 

 the name of interjecta. Rober. The name of fiavescens is the oldest 

 given to any individual form from that locality, and designates the 

 one which constitutes the characteristic of that race in the second 

 generation. I have figured it in Il/top. Fal., pi. xxxii.,figs. 46 and 47 ; 



