436 BROOKLYN BOTANIC GARDEN MEMOIRS 



tion to the expected ratios. In families No. 15405 and 15406, which 

 were given the most careful and detailed study, there was little evi- 

 dence of the suppression of dominant characters, except that in a 

 few specimens it was a little difficult to be quite sure whether the 

 A lobe was present or not and it is not improbable therefore that a 

 few heteris plants have been erroneously included in the rhomboidea 

 group, but this does not affect the ratios relative to the presence or 

 absence of the B factor. On the whole, the three families derived 

 from Fi plants of pedigree 14378, showed the least marked tendency 

 to the suppression of the dominant lobing, and these families show a 

 close approximation to the expected ratio 45 AB : 15 aB : 2> Ab : i ab. 

 The close agreement with this ratio in these families, indicates not 

 only the duplication of the B factor but also the independence of the 

 two B factors from the A factor. 



While 15 : i ratios in the Fo give evidence of duplication, it is 

 highly important to carry the analysis forward at least into the F3 

 generation in order to secure more convincing proof that the B factor 

 was really duplicated in the dominant parent of the original cross. 

 Until now the only families beyond the F2 which have been grown from 

 material in which the B factor is duplicated, have been derivatives 

 from the earlier Tucson cultures and, as before, these families were 

 grown primarily for the study of the capsules, and only incidental 

 attention was given to the rosettes. The ratios in these families are 

 also defective, therefore, but they give, nevertheless, strong support 

 to the hypothesis that the B factor is duplicated in the Tucson plants. 

 These F3 families are brought together in Table 5. 



This table has been arranged into the three groups which are 

 expected in the F3 of a cross involving duplication of determiners. 

 In the first section are the families which bred true to the B lobing; 

 in the second section are those which split in the ratio 15 : i, and in 

 the third section are those which split into 3:1. The results may be 

 summarized as follows: 19 F3 families contained neither tenuis nor 

 simplex individuals, seeming to indicate that the 19 parents of these 

 had at least one of the B factors homozygous; 3 families showed 

 ratios which may be assumed to represent the 15 : i class, showing 

 that the 3 parents of these had both B and B' present in the hetero- 



as probably having a duplication ot the B lobe. As stated in footnote to Table I, 

 15:1 and 3 : i ratios might both occur in the F2 families grown from plants in a 

 single Fi family, if the wild form used in the cross happened to be heterozygous for 

 one or the other of the B factors. Thus 



BBB'b' X bbb'b' = BbB'b', yielding 15 : i, and Bhb'b' yielding 3:1. 

 This situation appears to have been realized in two cases, involving Fj families 

 14357 from Wales and 14368 from Berlin. These two crosses are included in both 

 Tables i and a. 



