34^ Journal of Agricultural Research voi. xxn, no. 7 



though it is borne in mind that the values of the probable error are likely 

 to be too high in data of this kind. 



Summing up the types of groups C and D, the former yielded 67 spelts 

 to 5 wheats, and the latter group yielded 56 spelts to 17 wheats, the 

 deviation being 0.5 and 1.2 and the probable errors ±1.38 and ±2.50, 

 respectively. 



Of the 2 7 plants tested (Tables V and VI) , 5 (group C) showed an ap- 

 proximation to the 15 to I ratio and 8 (group D) to the 3 to i ratio. The 

 theoretical number of plants belonging to each of the two unstable groups 

 was 6.7 — that is, one-fourth of the total Fj plants tested. Comparing the 

 results obtained with those expected, it will be noted that there were in 

 the F, generation two (1.7 actual) SjSiSaSg individuals less and one (1.3 

 actual) SjSiSjS, or s^SiSaS, more than expected (Table VI) . 



In considering the 15 spelt to i wheat segregation, it should be borne 

 in mind that if the progeny of a heterozygous Fj plant is less than about 

 10 individuals, the chances are that the wheat form, which is expected 

 to appear but once in a population of 16 individuals, will not be obtained. 

 Such heterozygous plants producing only spelts and no wheats, on account 

 of their small F3 population, would be classified under group B. Had Fj 

 plants 13126a — 12, — 20, and — 25 produced more than 4 or 5 individuals, 

 I or 2 of them might have produced a wheat form which would have placed 

 them in group C. The experimental ratios then would almost coincide 

 with the theoretical. 



Considering the closeness of agreement even in the details of the 

 analysis, with such small numbers, there seems to be no question that we 

 are dealing here with two spelt factors and that the ratio observed is the 

 ratio of 15 to I. 



In the absence of more experimental evidence, the simplest hypotheses 

 were given to account for the 3 to i and 15 to i segregations. Notwith- 

 standing the surprisingly close agreements between the experimental and 

 theoretical ratios, however, the real explanation concerning the produc- 

 tion of the spelt character is still a matter of speculation. 



The same spelt parent plant was used in crosses 13124a, 13125a, and 

 13126a. The wheat parents were of different varieties. No satisfactory 

 explanation can be offered as to why the same spelt parent should pro- 

 duce a 15 to I ratio in cross 13126a and a ratio of 3 to i in crosses 13124a 

 and 13125a. Three possibilities, however, may be mentioned. 



1. The spelt plant used in the above-mentioned three crosses might 

 have been heterozygous for one of the spelt factors. Such a heterozygous 

 SjSiSjSj plant bears gametes producing 3 to i and 15 to i ratios in the 

 F2 generation when crossed with a double recessive (SjSjSjSj) wheat. 



2. The spelt parent may be assumed to have carried two spelt factors 

 Sj and S2 and some of the wheats might have carried an inhibiting factor 

 I. If the wheat carried the I factor the ratio of spelt to nonspelt would 

 be about 3 to i and if it did not the ratio would be 15 to i. 



