HALDANE ON SELECTION 



M9 



This result was previously given by 

 Punnett (8). 



case e: selection of an alternatively 

 dominant character 



Certain factors, such as those determin- 

 ing the presence or absence of horns in 

 Dorset and Suffolk sheep according to 

 Wood (9), are dominant in one sex and 

 recessive in the other. With complete 

 selection, all members of the type domi- 

 nant in the female sex are weeded out. 

 With k = 1, the type disappears in the 

 male sex and is halved in successive 

 female generations. If k — —00, the 

 recessive type in the female sex disappears 

 in that sex and is halved in successive 

 male generations. For small values of k, 

 that is, for a low rate of selection, we 

 have kn = 2.log. e u n , so that selection 

 occurs on the whole more rapidly than in 

 the case of a simple Mendelian character. 



case f: bisexual selection of a sex- 

 linked CHARACTER 



The case of selection in a population 

 whose members differ with regard to a 

 sex-linked factor leads to the conclusion 

 that for complete selection, no dominants 

 survive to breed and the selection is 

 complete in one generation. If k = 1, 

 and no recessives survive to breed, there 

 are no recessive females produced and the 

 number of recessive males is halved in 

 each generation. Selection is therefore 

 vastly more effective than on an auto- 

 somal character. If colorblind or hemo- 

 philic people were prevented from breed- 

 ing, these conditions would be almost 

 abolished in a few generations, which 

 would not be the case with feeble-minded- 

 ness. 



In the case of slow selection, if genera- 

 tions are reckoned from a standard 

 population when u Q = 1, and 50 per cent 

 of the males and ±5 per cent of the females 



are recessives, we find that within an 

 error of the order of k, we have 



and 



m 



kn = log e a„ -+- 2. log 



y n = (i - O" 2 

 & = (1 + 0~ l 



when y n = per cent of recessives of the 

 homozygous sex and in = per cent of 

 recessives of the heterozygous sex. Thus 

 selection acts more rapidly on a sex- 

 linked character in the homozygous sex 

 than on an autosomal character and at 

 about the same rate in the heterozygous 

 sex as on an autosomal character. 



CASE G: BISEXUAL FAMILY SELECTION OF A 

 SEX-LINKED CHARACTER 



Considering the case where the family 

 within which selection occurs has both 

 parents in common and letting the domi- 

 nants have an advantage of 1 to 1 — k over 

 the recessives in the mixed families, the 

 author finds that with complete selection 

 for k = 1 the recessives disappear at a very 

 fast rate. When k is small, the selection 

 proceeds more as in the case of racial 

 selection, but at from one-half to one- 

 third of the rate. 



CASE H: SELECTION OF A SEX-LINKED 



CHARACTER IN THE HOMOZYGOUS 



SEX ONLY 



Several sex-linked factors are known 

 which have a much more marked effect on 

 the homozygous than the heterozygous 

 sex; thus in Drosophila melanogaster ' 'fused" 

 females are sterile, males are fertile, while 

 the character "dot" occurs in 8 per cent 

 of the genetically recessive females, but 

 only 0.8 per cent of the males (Morgan 

 and Bridges, 10). Under these conditions, 

 with complete selection for k "= — °° , the 

 dominants disappear in two generations. 



