1913.] PUBLIC DOCUMENT — No. 31. 195 



111 the third and fourth generations from the cross it appears 

 commonly among the progeny of light pink and crimson parents, 

 and in two cases from white-flowered parent plants, and in one 

 case from a pink-flowered parent. It is probable that these 

 three cases shonld be regarded as exceptional, and it is possible 

 that one or two may arise from erroneous observations. When 

 self-fertilized it commonly breeds true, or it may yield light 

 pink or white progeny. In one case it yields one crimson- 

 flowered plant. There have been several cases where snch plants 

 have appeared most unexpectedly from parent plants of lighter- 

 colored flowers, a thing that we are unable to explain unless they 

 be regarded as mutants. 



Light pink flowers appear in all crosses of Table 5, except the 

 one with Davis, and pink fl.owers in all of them. 



A study of the ratios of these several blossom colors leads only 

 to confusion. It is probable that the theoretical ratios dift'er 

 in the difterent crosses, and, as in many other cases, are corre^ 

 lated with the seed-coat pigmentation. 



Table 6 gives the results of crossing pink and light pink- 

 flowered varieties. Only the parent types appear, as a rule ; 

 the pink is uniformly dominant, and the proportions, while 

 departing considerably in some cases from the expected ratio 

 of 3:1, are perhaps not further than might be expected. 



