MULTIPLE FACTORS IN MENDELIAN INHERITANCE 179 



it would be expected that they would have intermediate ear 

 lengths. The numbers obtained by Castle were small enough 

 not to show the new combinations expected on this theory. 



On account of the many external conditions that may influence 

 the development of the hereditary size factors, one could not ex- 

 pect to find the various grades in F2 clearly enough separated 

 to discover the number of factors that might be involved. How- 

 ever, evidence favoring a multiple factor interpretation of size 

 would be found in the appearance in F2 of extremes not found in 

 Fi or the increased variability of F2 over Fi, especially if ex- 

 tremes beyond the parental races appear. 



In plants numerous experiments favoring a multiple factor 

 hypothesis have been described. Since definite ratios may be 

 obtained for qualitative differences, such characters afford the 

 most convincing evidence. Supporting Nilsson-Ehle's results, 

 East ('10 and '11), and East and Hayes ('11) in endosperm and 

 pericarp colors of maize, found ratios which are apparently to 

 be explained only by multiple factors. Tammes ('11) crossed 

 two different sorts of blue Linum; an intermediate first gen- 

 eration was followed by a second generation wdth a wide range 

 of variability, the intermediate shades having the greatest 

 frequencies. 



From quantitative differences in color one is led to quantitative 

 differences in shapes and forms. The work of Kajanus ('11 and 

 '12) on the roots of Brassica and Beta affords evidence to sup- 

 port the hypothesis that various form types of these roots are 

 due to combinations of multiple factors for length, or length 

 and roundness. Emerson ('10) reports increase in the variabil- 

 ity of the shapes of the second generation from certain crosses 

 with squashes and gourds. These cases of form differences do 

 not show dominance. Shull ('11 b) described a case involving a 

 double gene for the same character, that is dominant. He found 

 that the flat and triangular seed capsule of Bursa bursa-pastoris 

 is dominant to the oval and, in section, round seed capsule of its 

 recent mutant heegeri. The ratios of the second and third gen- 

 erations strongly indicate that there are two genes involved, 

 each of which can produce the pastoris type of capsule. 



