312 AN INTRODUCTION TO MODERN GENETICS 



times as many females as he could serve naturally. The sperm can, 

 moreover, be shipped considerable distances before use.^ 



The best method of selecting males for use in grading up has been 

 the subject of considerable investigation. It may be possible to select a 

 male who himself shows the characters desired, as would be the case, 

 for instance, with bulls of beef-producing types. But it is clear that 

 what we really wish to select is a certain genotype, and this may be 

 very imperfectly expressed in the phenotype of its bearer; in par- 

 ticular, bulls for breeding dairy cattle can in the nature of the case 

 show no direct sign of their genotypic capacity for milk production. 

 Attempts to correlate external characters with the nature of the hidden 

 genotype have been made empirically by breeders frequenting the sale- 

 ring, and scientifically by correlation studies, but in neither case have 

 led to reliable or easily communicable results, though some breeders 

 undoubtedly develop a flair for judging good breeding stock. 



It should be possible to tell something about an animal's genotype 

 from an inspection of his pedigree; and the recording of all pure-bred 

 animals in herd books, etc., is an attempt to provide a basis for pre- 

 diction of this kind. In the past, however, too much attention has been 

 paid to the presence of a famous animal somewhere, perhaps quite far 

 back, in an animal's ancestry, and it has been forgotten that the other 

 animals involved, particularly those in recent generations, have made 

 important contributions to the genotype under investigation. Again, 

 indices of inbreeding,^ designed to give a hint as to the degree to which 

 an animal is likely to be homozygous, have usually been too largely 

 influenced by inbreeding some generations back, whereas it is the 

 recent generations which have most effect on homozygosis. 



The only satisfactory way, in fact, of determining what an animal's 

 genotype is like is to test it. We must allow the male to sire offspring 

 and determine from the offspring themselves whether their father is 

 capable of transmitting useful characteristics. Various technical methods 

 have been proposed for disentangling the contribution of the male 

 from those of the females in such tests, but we shall not here'^ go into 

 these different methods of arriving at a "bull-index"; the essential 

 point is to use the bull with several females and study the resemblances 

 between his daughters. The effects of selection guided by such progeny 

 performance tests, as they have been called, have been very strikingly 

 shown in, for example, an experiment on egg-production at the Maine 

 Agricultural Experiment Station. From 1899 to 1907 selection was 



^ Landauer 1933^5 Walton 1933. - Cf. Wright 1922. ^ Cf. Rice 1934. 



