SEASONAL POLYMOKPUISM. 



71 



be a proof of the existence of at least two generations. As I had never 

 conceived the possibility of any butterfly living nine or ten months 

 at the imago stage in a climate such as that of peninsular Italy, I had 

 concluded that an autumn emergence must take place also in Florence, 

 though the insects did not show themselves on the wing at that time 

 of year. Last year we have actually had a proof of the longevity of 

 rhatimi: Querci in September found a female still alive in a paper, 

 where it had been placed in July. When it was taken out it made no 

 attempt to fly, but remained motionless till February, notwithstanding 

 the warmth of the room, when on a sunny day, it was suddenly heard 

 knocking vigorously against the window-panes. 



C'olias liijale, L., and C. croceiis, Fourcr. — The first generation of these 

 two species I have figured and described at length in Hhopalocera 

 Palaearctica (April, 1909), under the name of vernalifi. Stauder, not 

 being aware of it, gave that of cmceiis the name of iiu'diterranea in 

 1914. The features of vernaUs in both species are : great individual 

 variability ; small size in most individuals ; paleness of colour ; 

 increased extent of yellow powdering; of streaks on neuration ; and of 

 spaces, according to the sex and the species, in the marginal dark 

 band ; this is reduced in extent ; so is the discocellular spot ; more or 

 less thick and extensive green powdering on underside; prominent 

 pre-marginal row of spots on this surface ; costa and fringes con- 

 spicuously pink. In ('. ixjale the two summer broods of June and of the 

 end of August and September are quite similar to each other, as far as 

 I have been able to make out from my series of Central Italy ; they 

 belong to the form I have called calida in the Ent. Eec. for May, 1916, 

 p. 99, taking as typical a Tuscan couple figured in PJwp. Pal., pi. xl., 

 figs. 31 and 36 ; they stand opposite to rernalis by all the characters 

 mentioned above. The fourth emergence, which begins towards the 

 10th of October in Florence, usually consists only in a few sporadic 

 individuals : in exceptionally favourable autumns, such as this last 

 one of 1921, quite a fair number make their appearance (over 200 

 were collected in one locality in 1921), but whether they reproduce 

 and their oft'springs join the first generation in the following spring, 

 so that one can really speak of the October emergence as of a fourth 

 extraordinary generation in some years, probably depends entirely on 

 the time when the first severe frosts take place ; when they are not 

 too precocious they may allow the larvae to reach the very definite 

 fourth stadium in which, only, it seems that hybernation can safely be 

 accomplished. Anyhow the October individuals afl'ord striking 

 features of their own: the chief one consists in the extremely pale 

 tinge, more Avhitish in male than rernalis ; the marginal dark band is 

 as developed as in calida, or rather more so, on an average ; the under- 

 side is always conspicuously dusted with green. We thus have a 

 mixture of some characters of calida with some of rernalis. I have 

 pointed out in 1916 (I.e.) that this form quite resembles the summer 

 generation of the more northern parts of the habitat of the species, 

 and thus answers to the English insect of Petiver's figure, which is 

 the first quoted by Linnfeus ; besides, as the two other seasonal forms 

 have been named, this, also by exclusion, would remain the nymo- 

 typical one of the species. 



As to C. croceus, there is quite a difference between the 



