146 
Variation and Inhenta7ice in Aphis 
the ratio the environment might be expected to have some effect, but the greatest 
effect would naturally be on the fraternal correlation. Theoretically the mothers 
and all the broods ought to have lived on exactly the same sort of leaf, or failing 
this the mothers could have lived on one kind of leaf and all the broods on some 
other kind. Unfortunately this is not wholly possible in practice. 
(12) Summary. 
The present investigation well illustrates the exceeding difficulty of breeding 
organisms for statistical purposes, and it emphasizes the necessity for the greatest 
caution in drawing conclusions from the results. 
The reaction of the organism to its environment is so prompt that it is practi- 
cally impossible to breed two generations under such conditions as to ensure the 
same mean and the same variability in the two generations. The differences 
which occur are due to two sets of causes: (1) the unavoidable removal of certain 
selective agencies which would act under purely natural conditions, (2) the direct 
effect of the environment both on the parents and on the offspring. The resultant 
effect of numerous influences under either or both of these headings might increase 
or diminish the apparent intensity of inheritance. I will now summarize in sepa- 
rate paragraphs the facts and conclusions deduced from the present investigation. 
(1) The mean of the ratio rises very greatly with the growth of the indi- 
vidual. 
(2) An individual was said to be "mature" or "adult" after the period of 
rapid growth had passed. 
(3) Under unfavourable conditions the individual grows slowly, after a time 
offspring may be produced, and ultimately growth nearly ceases and the indi- 
vidual is " mature " but permanently small. 
(4) The unfavourable conditions of the environment had a marked effect on 
the absolute dimensions of the adult third generation, but they had little effect on 
the mean of the ratio. These conditions did not induce any great formation of 
winged individuals. 
(5) The number of offspring in a brood may be taken as a rough index of the 
fertility of the parent. There was but little correlation between fertility and 
the dimensions measured on the parents. 
(6) In the second generation there was a death-rate of 12-8 per cent, among 
the immature animals. The larger individuals tended to have healthier families. 
Offspring of fertile mothers were somewhat stronger in constitution than those 
of less fertile mothers. But those families in which the majority of the adult 
members were widely divergent from the filial mean exhibited little difference in 
their death-rate. 
(7) The frontal breadth and the length of antenna are closely correlated 
in the adult individual, the coefficient being '80. The frontal breadth (taken 
