VARIATION IN DOMESTIC ANIMALS 459 



Modifiability and Correlation. Low correlation does not necessarily 

 indicate the relative degree and importance of modifiability in a popula- 

 tion, for it is possible to have low correlation between parent and off- 

 spring as a consequence of genetic variability. This matter may be 

 pointed our very clearly by considering a few abstract cases. Thus a 

 population consisting largely of different, but homozygous, forms would 

 exhibit a high degree of correlation between parent and offspring; 

 whereas one containing more heterozygous individuals would display 

 a lower correlation coefficient, simply because of the segregation which 

 would take place in such a population. This, of course, is simply a de- 

 velopment of the general case that an individual of genotype AA would 

 produce only A individuals, provided A were completely dominant, 

 whereas the genotype Aa would produce individuals of its own type and 

 also of type a. On the other hand, modifiability may be the factor 

 determining the value of the coefficient of correlation, even when the 

 degree of conformance to a given standard is very high. Thus if a 

 population be homogeneous genetically, as is the case in pure lines, then 

 the correlation coefficient within the population would be nil, but when 

 a number of different pure lines are mixed together and the correlation 

 coefficient is determined for the mixed population, the value would be 

 very high, although each pure line within itself would exhibit zero 

 correlation. 



This matter requires specific attention because some little confusion 

 has grown up from the use of the coefficient of correlation as a measure of 

 the intensity of inheritance, a practice of doubtful scientific propriety 

 and one which might well favor misleading conclusions. The lack of 

 correlation, for instance, in egg production of hens and their daughters 

 theoretically does not indicate that attempts to increase egg production 

 will be absolutely futile. It does indicate that some method of 

 breeding must be adopted that will discount at their proper value the 

 influence both of modifiability and genetic variability consequent upon 

 segregation. The theoretical interpretation in this case is borne out 

 strikingly by the practical results of the application of the method of 

 genotypic selection, for as is shown graphically in Fig. 185, there has 

 been a striking increase in egg production from the year in which the 

 method of breeding was changed. 



Variation by Recombination. Unquestionably the greatest possi- 

 bility for improvement in animal breeding as well as in plant breeding 

 lies in the isolation of recombinations of germinal elements which are 

 better adapted to specific purposes. There is every reason to believe, 

 as we shall point out in the next following chapter, that the architecture 

 of the germ-plasm, if we may use such a term, in domestic animals is 

 similar to that in Drosophila, that is, that the elements of the germ-plasm 



