(1) 
1907, completing the series; no further larve or pup re- 
maining. These four include three pale (hyerana) forms and one 
dark (marginata). The hyerana are a dark 9, darker than 
anything preceding, which happens also to be the only ¢ 
specimen at all referable to var. alpha, Mill., another very 
dark ?, anda male that met with some unaccountable accident 
in emerging, and whose precise form is therefore indetermin- 
able; it was further remarkable in the pupal stage possessing 
mandibles of a larval structure, although the imago appears to 
have no peculiarities in thisregion. ‘The fourth January speci- 
men is a female, the very darkest of the marginata form I have 
seen, Summarising, we have out of a total of fifteen specimens 
that emerged in November, December (1906), and January 
(1907), thirteen that are darker than any that appeared in the 
normal season of August, September and October, although the 
latter were over 400 innumber. We should like to know why 
these were so late in emerging, and why they are so dark ; 
and what is the co-relation between these two facts. Why 
they are so late, Ido not know. They were kept along with the 
others that emerged at the normal dates. Three months is a 
long period for a normal emergence to last over. One must 
suppose, therefore, that it has an advantage in securing that 
some specimens shall appear when weather or other circum- 
stances are favourable to them, either in early, late, or mid- 
season, and if early specimens are selected one year and late 
another, and so on, the race will acquire a habit of prolonged 
emergence and a variability in date of emergence that will 
include extreme specimens, beginning, as I have shown, in June 
and extending to the following March. 
“This does not involve any explanation of the darkness, and, 
as a matter of fact, late specimens in preceding years did not 
present any special darkening to attract my attention. It 
happens that this season we have had various spells of cold 
weather, and the room in which I kept the larvee and pup 
was only intermittently warmed and often fell in November 
to as low as 50° and even 47°. This would of course make 
still later those specimens that were late enough to fall under 
its influence. But it had the further effect of delaying 
emergence after the pupal state was assumed, so that whilst 
