118 Glume-Length and Grain-Length iyi Wheat 



Interesting evidence of the occurrence of "shift" accompanying an 

 otherwise apparently single segregation of one factor is to be found in 

 the results of a number of investigations. Caporn [(3), p. 259] has 

 remarked upon it and it is clearly shown in some of Backhouse's results 

 [e.g. (2), p. 130; Fig. la]. In both of these cases Polish was one parent 

 of the cross. The importance of a recognition of "gametic impurity" 

 or "contamination" has been emphasised by Ruggles-Gates (9) and in a 

 variety of crosf^es the occurrence of the phenomenon is exemplified. 

 Among these are Punnett and Bailey (10) and (11), Castle and Phillips 

 (12), Davenport (13), Martin Leake [(14) — leaf factor in Cotton], East 

 (15),Balls [(17), p. 69 on seed-weight], and possibly some of the intangible 

 facts discovered in Tomato investigations by Groth (7) might be 

 explained as results of " shift." It is interesting to observe that the 

 results recorded by Belling (16) constitute one of the closest studies of 

 length inheritance and that they show complete absence of "shift." 



/. 



Measurable Characters of the Heterozygous Glume Type in F^, F^, and F.^. 



From Tables XII and XIII it appears that the frequency distribu- 

 tion of glume-lengths for the plants of F2 and Fs which are neither P 

 nor K (and are therefore classed as /) is unimodal, and this is in 

 accordance with the one-factor explanation which has been adopted. 

 With regard to the relation of the mean glume- length of the hetero- 

 zygote to the means of the parents, the following values are of 

 interest. 



(Fo .Mp+Fo. Mk)/^ = 21-27 mm. (F^ .MpxF,. Mj^f = 19-Omm. 



{F^ .Mp+F^. Mk)/2 = 17-30mm. {F^ .MpxF^. Mk)^ = 16-27 mm. 



F2.Mi=ie-4>lmm. 



It will be observed that a modified form of the "golden mean" 

 theory of Groth (7) and (8) appears to fit the facts of this case. 



The complete ^3 progeny of the Fz . I plants should, on the hypo- 

 thesis adopted, present a glume-length distribution similar to the 

 distribution of the complete F^. That this is the case is shown by 

 Table XIV. There appear to be three constituents with modes 

 approximately at : 



ll-5mm. I7'0mm. 24-5mm. 



