108 



The Journal of Heredity 



In the studies of distribution as actu- 

 ally occurring in the Rufer family, the 

 germinal constitution was necessarily 

 inferred from the character of the in- 

 dividual. That is, low ability was 

 taken to indicate the germinal consti- 

 tution cc, or all germ cells with relative 

 absence of determiners for calculating 

 ability. Cc medium ability, with half 

 the germ cells having determiners and 

 half not having sufficient determiners 

 for this ability; while high ability indi- 

 cates the germinal constitution CC, or 

 all germ cells with sufficient determiners 

 for this ability. 



Let us examine first the distribution 

 of the grades of this ability in a group of 

 the Rufer family where the level of the 

 ability is distinctly lower than that of 

 the general population. Here class I, 

 indicated by L, includes all those having 

 no ability to handle quantities beyond 5 ; 

 class 2, indicated by M, includes all 

 those able to perform simple multi- 

 plication and division, and make small 

 change; class 3, indicated by H, includes 

 all those having ability beyond this, but 

 still not above the average for the 

 general population. The progeny of 48 

 matings, including 177 offspring were 

 graded and their distribution in the 

 three classes compared with the theoret- 

 ical expectation as given above. 



Comparison of percentages realized 

 with percentages expected: 



The realized percentages support the 

 following conclusions: (1) When both 

 parents show a low grade of this ability, 

 all the children are similarly of low 

 grade. (2) When both parents show high 

 ability, practically all the children are 

 similarly endowed. (3) When one par- 

 ent shows a low grade and the other a 

 medium or high grade the increase in 

 the percentage of medium and high is 

 proportionate to the grade of the abler 

 parent. Although a small proportion of 

 low grades are produced by the mating 



of a medium with a medium of high 

 grade, here again the proportion of low 

 grades shows greater decrease when one 

 consort is of high grade than when he 

 is of medium grade. 



A similar study of the distribution of 

 grades of calculating ability was made in 

 the Riel group. The distribution of 

 thirty-two families including 120 indi- 

 viduals was determined, only here, 

 different values were assigned to the 

 three classes. L included all those who 

 could perform simple multiplications 

 and divisions and make small change; 

 M included all those with average abil- 

 ity, while H included all those with 

 ability above the average. The com- 

 parison of percentages realized with 

 percentages expected follows: 



The realized percentages in this group 

 support all the conclusions drawn from 

 study of the Rufer group, with the 

 exception of the second conclusion, the 

 mating of high ability with high not 

 having occurred. 



If we adopt the hypothesis of a unit- 

 like series of determiners in the germ 

 plasm to account for the development of 

 this ability, we may conceive the number 

 which means presence in a low degree 

 in the poor strains, to be replaced by 

 a larger number of determiners and 

 consequent good ability in the bet- 

 ter strains. Should this replacement, 

 through "good matings" occur for a 

 number of geilerations, a point may be 

 reached where fair and good ability 

 would still appear in the offspring, 

 even though there be mating with low 

 ability. That is, relative absence in the 

 germ plasm of the better consort would 

 behave as relative presence to the lack 

 in the weaker consort; the resulting 

 simplex individual would show and 

 transmit fair ability. The serious danger 

 would arise from matings in two consec- 

 utive generations with low or no 

 abilitv. In such cases there would be 



