578 GENETICS IN RELATION TO AGRICULTURE 



that eventually a given result may be produced, his time is limited and 

 he, therefore, desires, and rightly, to achieve a given result in the shortest 

 possible time. A case in point is that which we have already discussed 

 in some detail, modifiability in relation to selection for high egg produc- 

 tion in the domestic fowl. Here Pearl found that modifiability was so 

 great that simple phenotypic selection of the highest producers for breed- 

 ing stock resulted in no improvement whatever in laying capacity. On 

 the other hand, the application of a method of breeding which fully 

 allowed for this effect of modifiability and which further took into account 

 the germinal constitutions of the fowls selected for breeding purposes 

 immediately resulted in gratifying improvement. With most characters 

 the influence of modifiability is not so great as in fecundity of fowls. 

 Often in fact modifiability may actually be utilized to advantage by the 

 breeder in determining relative excellence. Thus any system of develop- 

 ment which tends to call forth the highest possible expression of the 

 capabilities of individuals tends to widen the differences between superior 

 and inferior individuals. Both good and poor dairy cows tend to give 

 increased milk yield when fed richly, but the increase is often more marked 

 in the good cows. On the other hand, with horses in general it is possible 

 by training to increase speed, but it is a question whether the increase in 

 such a case is more marked in good or poor horses. We may say with 

 confidence, however, that here training by developing the full capa- 

 bilities of the animal tends to bring its speed up to such a standard 

 that when compared with breed records, the superior excellence of the 

 individual is definitely established. Modifiability, therefore, is on the 

 one hand a factor which tends to decrease the possible effectiveness of the 

 method of breeding from the best; on the other hand, if properly utilized 

 it is a powerful aid in the accurate selection of those individuals which 

 possess the highest inborn capabilities. 



When we come to consider the influence of germinal diversity in 

 phenotypic selection, we approach more nearly the problem of the real 

 limitations under which the method of phenotypic selection labors. Here 

 we may distinguish different ways in which germinal diversity may hinder 

 phenotypic selection. 



Phenotygic Selection Does Not Distinguish Between Homozygous and 

 Heterozygous Individuals. To the student of Mendelism this diffi- 

 culty requires no further comment. It may be pointed out, however, 

 that the difficulty increases as the number of factors for which selection 

 is being practised increases. As with modifiability, however, this diffi- 

 culty tends to be obliterated by long-continued selection, for such 

 selection inevitably increases the proportion of homozygous individuals 

 within a given phenotype or standard of selection. Roughly it may be 

 said that the rate of increase of the proportion of homozygous individuals 



