412 



Journal of Agricultural Research 



Vol. V, No. 10 



Dividing the second -generation variates into groups on each side of 

 the means, we have : 



In each case there are fewer variates above than below the mean. 

 This agrees with the hypothesis that the factors act as multipHers. 



The second-generation means, including both short and long, were 

 85.9 and 86.9 mm. These two determinations average 86.4 mm. If E 

 is completely dominant and the minor factors act symmetrically, the 

 second-generation mean will be X(62.8 -t- 3 X 94.5) = 86.6. This is sensibly 

 the same as the actual average, 86.4. 



If factor E is a multiplier and completely dominant, we may find its 

 multiplying value in several ways : 



Parents — 



(Two plants each.) 

 (Including samples.) 



1910. .Lyon bean-^Florida velvet bean=92. 7-4-62. 9=1. 47. (Lyon bean value is too 



low.) 

 1912. .Lyon bean-r-Florida velvet bean=94.5-j-62.8=i.5o. 

 1912. .Lyon bean^- Florida velvet bean=95. 6-5-63. 2=1. 51. 



Second generation — 



1910. .Long-^short=94. 2-4-62.7 =1.50. 

 1912 . .Long-4-sliort=94. 7-4-62. 7 = 1. 51. 



This gives 1.50 to 1.51 for the multiplying value of Ee or E^, compared 

 with €2- 



The extremes of the two crosses were: 



1910. 

 1912. 



The results in the third and fourth generations show that these extreme 

 values are inherited. The values of 1912 are probably the more reliable. 

 If E is completely dominant and the factors are multipliers, the multi- 

 plying value of E is given by : 



Shortest long pod-J-shortest short pod= 79-^-53 = 1.49 

 Longest long pod-4-longest short pod =113-4-7 5 =1.51 



If E had shown incomplete dominance, the second value should have 

 been markedly greater than the first. The average multiplying value of 

 Ee or £"2 is here 1.50. 



