16 (208) SURPASSING UPWARD AND DOWNWARD OF THE INDICES 
minus-variations of the hereditary variations 79 or 80, then the indices 
of children arrange themselves as non-hereditary variations round __ 
those of parents. Then there would be no segregation in this family; the 
parents being homozygotes and having the same hereditary consti- 
tution. 
For cases in which the indices of children differ much from those of — 
both parents, or, in which the indices of both parents differ much, 
dominance may be accepted. For famliy 115c we may accept that the 
parent with the index 78 is a minus-varaition of f. i. the hereditary 
variation 80 or 81; the children have the indices 80-84 ; so this case 
is an example of dominance of the high index over the low one. This 
explanation is not sufficient for family 240d and other ones. Here we 
see that four of eight children have importantly higher indices than Ki: 
both parents. | 
So, though there are surely among the cases of the tables II and 
III (p. 13) some which by means of accepting a large non-hereditary 
variability, can be brought into the frame of the simple mendelian 
heredity, yet, in these cases we must admit the possibility of still 
another working, namely that of selection or prepotence of hereditary 
factors. 
With a view to the cases of the tables II and III (p. 13), besides 
segregation, dominance and non-hereditary variability, we also accept, 
prepotence of hereditary factors. In table II there is prepotence for 
the high index in table III for the low one. 
It is not yet clear how we have to understand this. CASTLE with his 
mendelian experiments has always had in view deviations in the results 
of the mendelian rules (1905—1919). For all the cases CASTLE denies 
the homozygoteness of the extracted recessives (experiments with long- 
haired and short-haired, 1906) and accepts variations of the heredity- 
factors. What is explained by others by accepting multiple factors, is 
explained by CASTLE by accepting a contamination of the gametes, 1. e. 
that in the heterozygotes the gametes influence each other in this way 
that their difference becomes smaller. If in such a case F, is intermed- 
late, then in these circumstances, the varlability of F, can nevertheless 
be larger than that of F,; moreover contamination of all gametes does 
not take place. 
An extensive series of experiments (1907—1919, 35000 rats) has been 
done by CASTLE and his cooperators in order to find out the value of 

