in Temperature- Variation. 445 
between which at that time the struggle was being 
waged.* If, on the other hand, there is no such influ- 
ence, but the effect is entirely a direct one and modifies 
the individual determinant, then all the as yet unex- 
hausted determinants that are capable of reacting to this 
particular disturbance would be affected in some degree ; 
though no doubt, as Weismann supposes, to a greater 
extent if they had reached the point of disintegration than 
otherwise. 
Again, it seems to me to be of great importance to 
ascertain if possible which of these modifications are 
transmissible to descendants. If all the modifications, 
including those which I have supposed to be produced 
in the first way, can be shown to be hereditarily trans- 
missible, this would amount to a demonstration that the 
second explanation is adequate ; and the first may then 
be abandoned as unnecessary. Should only some be 
inheritable, the presumption would be in favour of the 
co-existence of both modes of action; moreover, the 
greater the number of non-transmissible variations that 
can be produced, the more will the case be strengthened 
against pangenesis, and in favour of the “ centrifugal” 
theory. 
I am anxious to see, if possible, the results of breeding 
experiments on specimens like these for yet another 
reason. It seems to me that by comparative experi- 
ments, with and without artificial selection, on such 
variations as may be transmissible, a measure might be 
obtained of the relative importance of selection and the 
mere action of external influences in the transformation 
of a species. I think, too, that no better group for such 
experiments as these of Mr. Merrifield’s could be selected 
than the Vanessas. For, in the first place, it is only 
among poikilothermic animals that the direct effects of 
temperature can be fully studied ; then among these the 
Lepidoptera are pre-eminent for the extremely delicate 
register of variation afforded by their wings ; and, lastly, 
* The fact that in V. polychloros forcing invariably tends to pro- 
duce a certain effect, whether preceded by warming or cooling ; 
while refrigeration brings about another definite effect, whether 
followed or not by forcing, seems so far favourable to this hypo- 
thesis. See above, p. 452. 
