628 



THE AM ERIC AX XATUBAL1ST 



[Vol. XLIX 



by introducing sex-factors into the discussion at all, when 

 all actual individuals have both stamens and pistils, or 

 else neither. When we assume — as Goldschmidt does — 

 that the factor S is so modified in the double-throwing 

 races as to insure the sterility of pollen-grains receiving 

 it, the known facts must follow; it seems wholly superflu- 

 ous to refer the sterility to linkage of S t with a sex-factor. 

 The hypothesis seems quite unnecessarily complex; there 

 is no real evidence here for the existence of a distinctly 

 heritable femaleness factor, or for any elimination of sex- 

 factors in pollen-formation, or for the occurrence of non- 

 functional pollen in ordinary pure single (SS) races. 



In a reply to Goldschmidt, Miss Saunders (1913) gives 

 a very clear presentation of both her formulation and 

 that of Professor Goldschmidt, urging most of the ob- 

 jections to the latter scheme which are stated above, but 

 especially emphasizing its failure to explain the excess 

 of doubles over 50 per cent. She also objects to the as- 

 sumption of the existence of non-functional pollen, but I 

 can not agree with her on this point. 



I have sectioned anthers prepared for cytological 

 study, and have frequently observed stages subsequent to 

 the reduction divisions. The spore-tetrads appear nor- 

 mal, and there seems to be no early and conspicuous evi- 

 dence of later degeneration. The "single" pollen, how- 

 ever, might even germinate and yet be strictly non-func- 

 tional because of weak growth; and, as is shown below, 

 the singles are actually inferior to the doubles in vigor. 

 Selective partial sterility seems to be a rather common 

 phenomenon, and it very probably occurs here. 



The only other recourse seems to be the hypothesis of 

 somatic segregation mentioned above, and somatic segre- 

 gation, except as a rare accident of abnormal cell-divi- 

 sion, has no decisive evidence in its favor 5 and an over- 

 whelming convergence of probabilities against it. Bel- 

 ling (1915) calls attention to decisive evidence against 



