ALCOHOL AND HEREDITY 127 



of this extreme, a certain number will fall in Class <z, . 

 while the rest will be in Class b, all showing injury. 

 Still diminishing the dose in proportion to the power of 

 resistance, Class c will begin to appear, and become 

 larger as the amount of poison used decreases. In 

 birds, the high temperature and rapid metabolism 

 doubtless favor the rapid elimination of alcohol ; thus 

 the dose, though apparently identical with that of the 

 guinea pigs, is in effect less. It may well happen in 

 some cases that when Class b is small and Class c large, 

 the statistical results will show an actual improvement 

 over the normal population, in spite of the fact that a 

 certain number suffer injury. 



10. If, as appears certain, alcohol thus discriminates Outstanding 

 against the weaker gametes in the fowl, what will be the pr 

 effect on future generations ? It all depends on the 



source of the relative weakness. Is it a matter of 

 hereditary composition, or of differences of nourish- 

 ment, dependent possibly on position ? In the latter 

 case there will be no permanent effect ; in the former, 

 the average of the later generations should at least in 

 some degree maintain the observed superiority. Thus, 

 by an extraordinary paradox, it would be possible to 

 improve a breed of fowls by administering alcohol to 

 one or more generations. Experiments are now in 

 progress, designed to settle this question. 



If the germinal difference is hereditary, we should 

 expect a strongly heterozygous or cross-bred type to 

 show the effects more distinctly than a homozygous one. 

 In such a mixed type there might be many different 

 sorts of gametes, which might respond differently to 

 environmental influences. 



11. We now return once -more to our original ques- 

 tion. There is no reason to suppose that alcoholism, as 



