66 



in a restricted district, for it is well known that a species with 

 a fairly constant form in one locality often branches off 

 and produces distinct local races in other districts, each 

 local race being very fairly constant in its general facies 

 within its own area. The case of MelitcEa atirinia, of Gnophos 

 obscurata, &c., will at once occur to all naturalists. 



It is well known that whole broods of Lepidoptera reared 

 in confinement often give a much greater range of variation 

 than the same number of examples taken by chance in the 

 habitat of the parents, and that have been selected by 

 nature before reaching maturity. This is probably due to 

 the facts that the whole brood, or a great part of it, is reared 

 by the experimenter without much loss, and because, having 

 been reared under artificial conditions, no selection has 

 taken place. This, too, probably explains why, in the years 

 of comparative abundance of a species, there is frequently a 

 wider range of variation than when the species is compara- 

 tively scarce, for the struggle for existence is more or less 

 intermittent in its most severe forms, and under a combina- 

 tion of favourable circumstances that may occur from lime 

 to time a larger number of individuals come to maturity. 

 Nor must it be overlooked that the diversity and complexity 

 of conditions forming the environment tend, in different 

 years, to the selection of different characters : e.g. extreme 

 cold will select those that can bear, and kill off those that 

 cannot bear, a very low temperature ; extreme drought will 

 kill off those weak in other characters, whilst extreme wet 

 will tend to the preservation of those that can stand ex- 

 treme moist conditions, and the destruction of those that 

 cannot. 



It is probably this complexity of characters, resulting 

 from a complexity of environmental conditions, that permits 

 of such a great range of variation as we often see in some 

 species, although in every district (and I make this proviso 

 so as to include distinct local forms, which often in their 

 essence are almost species in themselves) that variation 

 plays about a mean which may be looked upon as repre- 

 senting the type of the species. That this is so is distinctly 

 proved by the difference that v/e often find existing in the 

 types of a species from different districts, in which the 

 species is subjected to different environmental conditions, 

 including among many other factors heat, cold, drought, 

 and moisture. The typical form for a given district repre- 

 sents, then, that form which is able to survive the various 

 changes to which the species is subjected in different 



