54- Studies in the Theory of Descent. 



equally erroneous to infer that the cold of winter 

 had no influence in determining the type. In this 

 case also the determining causes must have been in 

 operation during innumerable generations. After 

 the winter form of the species has become esta- 

 blished throughout such a long period, it remains 

 constant, even when the external influence which 

 produced it (cold) is occasionally withdrawn. 



Experiments cannot farther assist us here, since 

 we cannot observe throughout long periods of time ; 

 but there are certain observations, which to me 

 appear decisive. When, both in Germany and 

 Italy, we see Polyommatus Pkl&as appearing in 

 two generations, of which both the German ones 

 are alike, whilst in Italy the summer brood is 

 black, we cannot ascribe this fact to the influence 

 of a shorter period of development, because this 

 period is the same both in Germany and Italy 

 (two annual generations), so that it can only be 

 attributed to the higher temperature of summer. 



Many similar cases might be adduced, but the 

 one given suffices for proof. I am therefore of 

 opinion that it is not the duration of the period of 

 development which is the cause of change in the 

 formation of climatic varieties of butterflies, but 

 only the temperature to which the species is ex- 

 posed during its pupal existence. In what manner, 

 then, are we to conceive that warmth acts on the 

 marking and colouring of a butterfly ? This is a 

 question which could only be completely answered 



