348 Journal of Aqriculiural Research voi. xxn.No. 7 



So far as the ratios of speltoid to wheat forms are concerned, these 

 crosses segregated in the simple mendeUan fashion. In series 13228a 

 (Turkey X Gatineau) there were 53 speltoids and 15 wheats, showing 

 a deviation from the theoretical numbers of 2.0 with a probable error of 

 ±2.4 In series 13229a (Seneca Chief X Gatineau) there were 46 speltoids 

 and 17 wheats, the deviation here being 1.2 with a probable error of ±2.3. 

 In both cases the approximations of the figures obtained to those 

 expected are within the range of their probable erros. Therefore, it can 

 safely be concluded that the ratio is 3 to i and that there very likely is 

 but one spelt factor difference. 



The question of interest in the inheritance of this speltoid form 

 (Gatineau) is not so much in its 3 to i ratio as in the way it differs in 

 details from the spelt X wheat crosses first discussed, which segregated in 

 the ratio of 3 to i. 



In order to compare the Fj curves of these two groups of spelt X wheat 

 crosses, the Fj frequencies in Table VII (series 13228a and 13229a) and 

 the first two series in Table I (13255a and 13260a) are represented graphi- 

 cally (fig. 2). 



The comparative characteristics of the curves of these two sets of 

 crosses are as follows : 



The curves of the true spelt X wheat crosses (13255a and 13260a) begin 

 at class I , where they have their highest spelt frequencies. They gradu- 

 ally drop until they reach classes 5 to 9, inclusive, where there seems to 

 be an indefinite fluctuation of frequencies. Then the curves suddenly 

 rise again at class 10, which contains the spelt-free populations. 



As to the curves of the wheat X speltoid crosses, the Fj spelt popula- 

 tions begin at about class 3, where but a very few individuals are found. 

 Beginning at class 5, the curves steadily rise until they reach their maxi- 

 mum height at class 10. The curves produced by the true spelt crosses, 

 it will be recalled, continuously dropped instead of ascending. 



When these two sets of crosses are compared it will be observed that, 

 although there is but one spelt factor difference in each, the spelt factor 

 present in crosses 13255a and 13260a is entirely different from the spelt 

 factor present in crosses 13228a and 13229a. The wheat parent has 

 had no influence is producing this variation in distribution, the same 

 wheat parent plant, Turkey, C. I. No. 3375, selection 13389, having been 

 used in crosses 13260a and 13228a. 



GENERAL DISCUSSION OF THE GROUPINGS WITHIN THE SPELT AND 



SPELTOID CLASSES 



Having discussed the question of ratios, let us turn our attention to 

 the analyses of the details of the variations within the spelt and speltoid 

 classes. 



The possibility of distinguishing the homozygous from the heterozyg- 

 ous spelts of the Fj generation is of primary interest. Table VIII has 



