F. L. Engledow 117 



K. 



Afeasurable Characteristics of the Kubanka Glume Type in F^, F^, and F^. 



The F^.K population was small (114 plants) and the 1913-14 

 seasonal difference was doubtless responsible to a considerable extent 

 for differences in growth. From the above values it is concluded that 

 the K type maintained its measurable characters practically constant in 

 ^0, F2, and ^3. 



P. 



Measurable Chara/iteristics of the Polish Glume Type in Fq, F^, and F^. 



Table of Frecyiency 

 Year Generation M a A' Distribution 



1'J14 F^.V 30-84 ±0-104 3-78 GOO I 



1913 Fi.F 23-18 ±0-186 2-75 101 X 



1914 Fa.Pexi^a-^ 24-66 ±0078 3 78 1054 XI 



As in the case of Kubanka, the mean of {F^ 1914) is slightly greater 

 than that of F^ (.1913) and the difference is small enough to be attri- 

 butable to season. 



The differences 



F,.Mp-F^.Mp = 7-66 mm. = 24-83 7, of F^ . Mp 

 and F^.Mp-F^.Mp = Q-l%mm. = 2004 % of F, . Mp 



are so great in comparison with the relevant probable errors that there 

 is full justification for stating that in the P x. K cross, the parental P 

 does not re-appear in F^. In its place are found P plants which closely 

 resemble F^.P in general appearance but whose mean glume-length is 

 more than 20°/„ lower than that of Fq . P. This "reduced" form of P in 

 F2 produces, when selfed, P plants of the same "reduced" glume-length 

 in J^3. The possibility that "minor multiplying factors "are responsible 

 in some way for this phenomenon of "reduction" or "shift" demands, 

 perhaps, some form of reservation. It may be that the true ^0 • P type 

 is reproduced in F2 but in extremely small proportion. Apart from the 

 value of a safeguarding reservation there is nothing to recommend any 

 theory of multiplying factors in this case. On the grounds of simplicity 

 and agreement with observation, the most acceptable theory is that of 

 the segregation of one factor with the attendant phenomenon of "shift." 



