PARENTAL ALCOHOLISM AND THE PROGENY 289 



A' >A, because a loses to yi' as result of the agent's action 



(a' + 6') <a, for the same reason. 

 The relation of a' to b' will vary for different species, and even 

 probably individuals, depending upon the relative resistance of 

 the cells at intensity P. As P for any really effectively harmful 

 agent increases, a' will become, smaller, approaching as a limit, 

 when no normal offspring will be produced. This brings us to 

 an interesting paradox, viz., that so long as a' >0, that is so long 

 as any normal offspring are produced at all, the quality of these 

 normal offspring will be higher the greater the value of P, that 

 is the more intense the action of the deleterious agent. This is 

 obvious on biological grounds, because the more intense the 

 action of the harmful agent, the more intense the selection: 

 hence only the very 'best' germ cells will survive and make 

 zygotes.^ 



For any given organism, deleterious agent, and intensity of 

 dosage or appHcation, there should be a stable relation between 

 a', h', A' and P. 



Let us now return to our experimental data in the light of 

 these theoretical considerations. The essential point to Stock- 

 ard's results, as I understand them, is that the value of P for 

 alcohol on guinea pigs is relatively so high that practically 

 speaking a' = and the fundamental gametic equation has 

 become 



X'=A'-{-h'. 



If a' has any value it is very small. One gathers that after a 

 sufficiently high degree of alcoholization of the guinea pigs a 



^ The symbolic proof of the point is simple. Let Ma denote the mean vigor of 

 the a germ cells, and Ma- the mean vigor of the a' germ cell. Then we have 



Ma' > Ma 

 because 



nMa = y^ ^ aZ 



^~^ ai 



and n'Ma' = > a'Z' 



where S denotes smnmation between the indicated limits, n and n' are the number 

 of cases, Z and Z' are frequencies, and h is any subscript greater than 1 and ^ m. 



THE JOCRNAL OF EXPERIMENTAL ZOOLOGY, VOL. 22, XO. 2 



