HETEHOZYcnrs MATIN) 119 



SI \1M\IIY \\|) DISCI SSION. 



The matings selected because most obviously adapted to gh i 

 F* progeny produced an offspring about one-third more variable than 

 the Fi progeny. The wide range of variability of the I 

 more in accordance with the hypothesis of 6 zygotic factors for build 

 than that of 4 zygotic factors. 



The mating of two "medium-build'* parents gives a progeny that 

 is not very variable in build; not more variable than the progeny of 

 the S X F mating. The result is like thai obtained in Btudies in hered- 

 ity of stature, where the progeny of the M v - M was not extremely 

 variable. This result, in the case of build, that the medium build is 

 relatively invariable, suggests that some persons of medium build 

 really belong to a biotype in which fleshy build is due to 4-zygotic 

 factors, and that such parents are zygotically of the AAbb type, and 

 produce offspring of whom 100 per cent are again of that type. Ex- 

 amples would seem to be found in the following families with 4 of 

 more children: Bat-2, Bea-7, Kel-4a, Old-2, Rig-1, War- 3! Tl 

 is, of course, genetical justification for the assumption that the same 

 character may in certain strains be due to 3 gametic factors, in oth 

 to only 2. This is especially well shown in susceptibility to tumor- 

 growth of mice. 



Class V. The "Back-Cross 



The mating of heterozygous individuals with the "pure" recessives 

 of the parental stock, as well as with "pure" dominants, is an impor- 

 tant procedure in genetics, in order to determine the number of I 

 involved in the trait. This mating is of especial importance in detect- 

 ing "crossing over," but our data are insufficient to permit use for 

 this purpose. Though the total of our data is considerable, yet the 

 nature of traits due to multiple factors is such that the table has to be 

 broken up into a large number of small ones, of which the "constai 

 are subject to large probable errors. It seems best to distinguish 

 two kinds, of matings: (1) heterozygous medium or fleshy • slender, 

 and (2) heterozygous fleshy X fleshy (or very fleshy) of fleshy Btock. 

 Mating 1. Hbiebozyooub Mtoium oh Plebhi 9 nm. 



The data for this analysis are in table 32, which contains the appar- 

 ently most significant matings from Appendix tables [II, IV, and \ 1 1 1. 



The totals of table 32, b + c, give the following proportions: 3 \ S 

 26 S, 36 M, 20 F, 4 VF. 



The matings of 4-factor and 2-factor parents respectively gii 

 distribution of two types. Uniting these (on the assumption that 

 they are equallv numerous), we L r et proportions as follov 3 VE 

 25 S, 44 M, 25 F, 3VF. 



This theoretical series is of the Bame general type as that a by 

 the data, and the agreement (considering all of the unknown fa 



