F. L. Engledow 121 



(fi) Similarly it contains no component exactly like 



(7) The F.2 distribution is unimodal but (a) and (y8) suggest that 

 'shift" has occurred. Possibly therefore the distribution consists of 

 three constituents which lie very close on account of shift and which, by 

 marked overlap, produce unimodality of the whole. 



If the grain be one generation ahead of the plant which bears it, the 

 ratio in F^ should be 3 : 2 : 3 ; if not a 1:2:1 is expected. The ratio is 

 investigated in § VI below. 



It is to be observed that the F2 distribution is based entirely on 

 measurements and consequently the phenomena of "shift," "unimodality," 

 etc. are unaffected by the fact that in sorting it will be assumed that 

 K (or / or P) grains always accompany K (or / or P) glumes only. 



The measurements and numbers of plants con-espond with those for 

 glume-length in § III above. Owing to the sterility of a few F3 plants 

 the distribution totals differ in some cases from those for the corre- 

 sponding glumes. 



VI. The Measurable Characteristics of the K, /, and P Grain 

 Types in F^, F^, and F^. 



It is necessary to recall that " K grain" (or / or P) implies the grain 

 borne by plants whose glumes have been classed by eye as " iT " (or / 

 or P). That is to say, glume-length and grain-length are assumed to be 

 "genetic inseparable," to be governed by the same factor. Eye-impres- 

 sions amply support the assumption, and its validity may now be tested. 

 Glume-length showed segregation in Fz, as was made clear by examina- 

 tion of the ^3 progenies of the F2 pla,nts. The vindication of the 

 assumption necessitates a demonstration of the segregation of grain- 

 length in F2 similar to that given for glume-length. This is afforded by 

 the measurable characteristics of the K, I, and P grain types in ^0. -^2. 

 and F^, which were as follows : 



K. 



Meastirable Characteristics of the Kubanka Grain Type in F^„ F^, and F^. 



