[THOMPSON 



RIPENING PERIODS IN WHEAT 85 



see whether the expected result is the same as we obtained in a direct 

 cross of the eariiest with the latest parents. 



Since we are dealing with typical blending inheritance we must 

 make the postulate that the difference between each pair of parents is 

 more than one pair of determiners. Otherwise there would be segrega- 

 tion and no constant intermediate race. This is true whether the 

 difference be small or great. Where the difference between the paren- 

 tal means is small the F2 variation curve might be considered to in- 

 dicate merely a lack of dominance along with segregation of one pair 

 of determiners. But if this were the case there would be no true breed- 

 ing intermediate race but segregation in each generation. Other 

 experiments show that true breeding intermediate races do result. 



Now crosses involving a difference of only five or six days were 

 made at different regions in the whole range of variation. In each 

 case the result was typical blending with the F2 variation covering the 

 parental range. Consequently for each five days' difference there 

 must be at least two pairs of hereditary determiners. Therefore in 

 crosses involving a parental difference of twenty or twenty-one days 

 there should be an hereditary difference of at least eight pairs of 

 factors. But in order to recover the extreme conditions in crosses 

 involving eight pairs of factors it would be necessary to raise about 

 65,000 plants (Mendel's 4" where n equals 8) and yet we find that in 

 the cross represented in table VIII the parental means are recovered 

 in fewer than 120 plants and one of the parental extremes is exceeded. 

 In the cross represented in table XI, in which the parents differ by 

 about 17 days the parental means are exceeded among fewer than 

 150 plants of the second hybrid generation. Assuming the -minimum 

 number of determiners (6 pairs) we should expect to recover these 

 among not less than 4,096 plants. Considering the crosses of the 

 earliest parent only, we have cases involving parental difference of 

 6, 12, 15, 17 and 21 days. Nevertheless F2 plants more extreme than 

 the average of either parent are recovered among fewer than 200 

 individuals in each case, even though each successive parent must 

 differ from the preceding by at least two factors. 



It is evident from the tables that it is somewhat more difficult 

 to recover the extremes in cases where the parents are very different 

 than in cases where they are more alike. But apparently it is not 

 nearly so difficult as it should be if the multiple determiner hypothesis 

 is correct. 



It should also be stated that there is good evidence (not yet in 

 a condition to be published) that crosses involving differences of less 

 than five days give similar results to those reported. If this is the 

 case it is necessary to assume that more than one pair of factors are 



