HUMAN GENETICS 333 



method of searching for such genes in man is as follows for the simplest 

 case, that of a dominant. If a man receives an incompletely sex-linked 

 dominant from his mother, his constitution will be AX.aY and on 

 mating with a normal female aX.aX, he produces mainly affected 

 daughters and normal sons, with a few affected sons and unaffected 

 daughters due to crossing-over. But if he received this gene from his 

 father, it will be in the Y and will go mainly to his sons. Thus in both 

 cases the majority of affected children of affected fathers should be of 

 the same sex as their grandparents. The conditions for incompletely 

 sex-linked recessives are more complex. Haldane^has presented evidence 

 that five or six genes are incompletely sex-linked in man, and the 

 degree of linkage made it possible to draw a provisional Unkage map of 

 the homologous parts of the X and Y chromosomes (Fig. 138). 



5. Continuous Variation 



The greater part of the variation in normal human populations 

 appears to be continuous ; we can at best draw arbitrary Hues between 

 the tall and the short, the clever and the stupid, etc. The methods of 

 spotting genes which we have hitherto discussed are quite inapplicable 

 to such a situation, and other methods of attack have to be found. 



We may expect that the differences between human beings are 

 partly due to genetic differences and partly due to differences in the 

 environments in which the genotypes have developed. The problem 

 which is often posed is to determine the relative importance of heredity 

 and environment in producing the differences which occur. But this is, 

 strictly speaking, a meaningless question. We have seen (p. 189) that if 

 two different genotypes are reared in two different environments, the 

 difference between the resulting phenotypes is a function of all four 

 variables, the two genotypes and the two environments; it is therefore 

 impossible to get any true insight into what is happening by means of 

 observations on only two objects, namely, the two phenotypes. The 

 problem must be analysed into two separate questions. Let us consider 

 two environments and let the same genotype develop in each of them; 

 are the environments such that the phenotypes differ, and if so by how 

 much? Again, what differences are produced when two different geno- 

 types develop in the same environment? 



The technique which is employed in the study of continuous varia- 

 tion is that of statistics.^ The two main notions involved in the dis- 

 cussion which follows are variance and correlation. Suppose we select 



^ Haldane 1936a. 2 ^f. Fisher 1928a, Pearl 1930, Yule 1929. 



