CHAPTER XXXVII 

 METHODS OF BREEDING 



Like modes of research, methods of breeding are the means by which 

 certain results are attained. It is necessary to emphasize this fact, 

 because even yet there is much confusion in the minds of breeders as to 

 the relation which a particular method of breeding bears to results which 

 have been produced by its employment. Not infrequently statements 

 are made to the effect that a certain method of breeding is the cause of 

 the excellence of one race or strain or the inferiority of another. There 

 is a wide difference between the method of producing a given result, and 

 the cause of its attainment. For the sake of clarity of thought we shall 

 endeavor to emphasize this distinction, so far as is possible in the present 

 state of our knowledge, in the discussions which follow. 



Phenotypic Selection. The oldest method of breeding was simply 

 that of mating together the most excellent individuals. In popular 

 phraseology this is the method of breeding from the best its funda- 

 mental postulate is expressed in the old statement, like produces like. 

 We have called it the method of phenotypic selection in order to empha- 

 size the fact that the basis of choice for breeding in this method is the 

 sum total of expressed characters of the individual. 



It is not necessary to recount here at any great length the sort of 

 improvement which has been effected in modern breeds of domestic 

 animals by the application of this method of breeding. Let it be sufficient 

 to state that much of the excellence of modern breeds is an earnest of the 

 efficiency of phenotypic selection as a mode of breed amelioration. It 

 may, also, be stated justly that all later methods of breeding; out- 

 breeding, line-breeding, inbreeding, and genotypic selection; are simply 

 refined methods of breeding from the best they are methods of pheno- 

 typic selection plus something else; the something else usually ill-defined, 

 but sometimes, as in genotypic selection, more definitely conceived. 



The limitations of the cruder form of phenotypic selection depend 

 upon two primary causes, somatic modifiability of characters and geno- 

 typic differences among like phenotypic individuals. Since differences 

 which are due to modifiability tend in the long run to group themselves 

 around a mean in the form of a normal variability curve, it may be stated 

 dogmatically that long-continued phenotypic selection should tend to 

 obliterate them. But it is not enough for the practical breeder to know 

 37 577 



