DWARFS IN GLUYAS EARLY (WHEAT) HYBRIDS. 545 



point of view, unfortunate quality. These selections were in 

 some respects even very fortunate, as theoretically two-thirds of 

 the total should have given rise to dwarf progeny. The reason 

 for this fortunate result remains difficult to explain, and many 

 factors may have a bearing on the case. 



The populations taken as examples, though embracing only 

 io6 individuals in the case of 36a X 41A, and 41 in the case of 

 57& X 410, are nevertheless representative for illustrating some 

 phases which may be of interest, and may perhaps throw some 

 light on this matter. 



Quantitative characteristics, unlike certain qualitative charac- 

 teristics, still remain unsolved when considered from a definite 

 Mendelian point of view, but here may be an instance, even 

 though the object be practical breeding, in which the segregation 

 of such a quantitative characteristic may perhaps be more nearly 

 solved, the more so as at least one recessive is evident. Knowing 

 that dwarfness is recessive, wuU it be correct to assume that 



Fig. 5. — Frequency Curve for Ear Length (cm.). 57^ X 4i«. 



normal height is dominant, and that height and dwarfness are 

 an allelomorph? 



This need not be, for the two ])arent forms already differ in 

 their normal height, and it is not incorrect to predict that strains 

 may be established breeding true to type as far as height is con- 

 cerned, such strains varying probably from a height greater than 

 that of the tallest parent to a height much less than that of the 

 shortest parent; therefore height and dwarfness cannot be taken 

 as one allelomorphic pair of characters in the present case. 



It is not even necessary that in the case of 36a X 41 A the 

 characteristic height may embody the same number of allelo- 

 morphic pairs of characters as in the case of 57/; X 41^?. "or need 

 this be expected. If, however, there is an indication that it is 

 possible to divide these quantitative characteristics into various 

 classes, each class being either homogeneous or heterogeneous in 



