HETER0ZYGO1 8 M STINGS. 107 



If one superimposes in the same drawing the distribution polygon! 

 of the parental and F, generations tables 23, 24, 25 and 30 I Bg L0 

 the modes of the fleshy parents and the 1 generation stand at fleshy, 

 though the mean of the F x generation is intermediate between tin- 

 means of the parental stocks. 



Class IV. Mating of Prouably Heterozygous Parents mn I 



Generation I 



We have next to consider the distribution of build in the progi 

 of parents both of whom are certainly or probably heterozygous and. 

 as such, probably carry gametes for slenderness, as well as fleshin 

 This is the mating in which the geneticist finds evidence for segrega- 

 tion, in that the parental types reappear nearly or quite as they went 

 into the combination. At any rate, this generation is characterized 

 by a greater variability than the F 2 generation. However, this greater 

 variability is the less marked the greater the Dumber of factors in- 

 volved in the prevailing dominant trait — in this case fleshiness. 



Two classes of matings will be considered — that of parents, whether 

 medium or fleshy, who probably carry gametes for slenderness, and 

 that of other parents who are both of medium build. 



Mating 1. Both Parents Either Medium or Fleshy and Probablt Cabbtow <'.\ metes 



for Slenderness as well as I*Yi-iiin. 



This mating includes certain parents selected, usually on the 

 ground of dissimilar grandparents on each side, as those which are 

 probably of the Fi generation. The families used are summarized in 

 table 31, which is made up of the appropriate families taken from 

 tables X. XI, and XIII. 



The range of variation of the F,> generation includes both the slender 

 and the fleshy types which are found in the grandparents. The 

 distribution of the progeny is more in accord with the hypothesis of 

 the 6-zygotic factors than that of 4-zygotic factors. 



The mean variability as measured by the standard deviation is 

 6.78 ± 0.26, as contrasted with that of 5.02 ± 0.30 of table 27 (the 

 Fi generation). The difference is 1.76, which is more than four times 

 the probable difference between the means. Thus, the variability of 

 the F- generation is nearly a third greater than of the I'\ generation. 

 That the difference is not greater is, of course, due in lame pari to the 

 presence of multiple factors. 



The family histories upon which table 31 depend- arc given below. 



Bea-17 Family. (Fig. 41. t 



A family reared in the Central States. Father ill 5) a contractor i i 

 farmer, who died of pneumonia at 70 years, had the formula 170 71 ; build 

 2.4 (341; medium. Three brothers who were killed in the Civil War at 17 

 19, and 21 years respectively had the formula 183 72 build 2.6 



