86 FAMILY STUDIES IN BUILD. 



variability are 2.41-^-28.47 and 5.35-^37.56, or 8.46 per cent and 

 14.25 per cent respectively. That is, the coefficient of variability of 

 build of the offspring of the F X F mating is 68 per cent greater than 

 that of the S X S mating. The ranges of the logarithms of the indices 

 of the S and the F groups is as near as possible the same, and the 

 arithmetic range of the F is only 40 per cent greater than of the S 

 group. Hence the greater variability of the offspring of the F X F 

 as compared with the S X S matings can not be accounted for on a 

 difference in the range of index values of the S and F classes. The 

 most reasonable conclusion seems to be that the gametes of the fleshy 

 parents are somewhat more variable than the gametes of the slender 

 and the very slender parents. 



Class III. Mating of Slender with Fleshy Stock — The F l 



Generation. 



After having considered the inheritance of build in the progeny of 

 slender and of fleshy stock respectively, we have next to consider the 

 distribution of build in hybrids between such stocks. This class of 

 matings produces the familiar Fi generation. Incidentally it may be 

 pointed out that where multiple factors are concerned the Fi genera- 

 tion has not that uniformity that it has when only one factor is con- 

 cerned. This is partly because there are several kinds of Fi offspring. 

 For example, the progeny of very fleshy X slender, fleshy X very 

 slender, fleshy X slender, and even subdivisions of some of these 

 categories. If we analyze the progeny of all these matings together, 

 a considerable range of variation in the Fj generation is to be expected. 

 If we deal with each type of mating separately, the numbers are insuffi- 

 cient to be significant. Compromises of various sorts will be necessary 

 between too small numbers of offspring, on the one hand, and too het- 

 erogeneous parents on the other. Further complications arise from the 

 fact that on the hypothesis that fleshy persons carry factors that domi- 

 nate over slender, it is to be expected that the children shall be, on the 

 average, of above medium build; and, again, from the fact that the 

 children are younger than the parents and, for the most part, do not as 

 yet show their full adult build. This is a matter of especial importance 

 in this mating, since heterozygous individuals are slow in developing 

 their potential build. Three kinds of matings will be considered. 



Mating 1. One Parent Slender, of Slender Stock, and the Other Fleshy, of Fleshy 



Stock. (Table VIII in Part.) 



If mainly those fleshy parents are selected whose parents and sibs 

 (so far as known) are fleshy, then the progeny will correspond as 

 nearly as possible to a Mendelian Fi generation, and should be char- 

 acterized by relatively low variability. Such selected matings are 

 given in table 27, extracted from table VIII, Appendix. 



