592 



THE AMERICAN NATURALIST [Vol. XLVIII 



wrinkled surface. This explanation is rendered less prob- 

 able, however, by the unusual behavior of the aleurone 

 color in this same ear. In the previous discussion the 

 aleurone color has not been considered. To treat of the 

 aleurone color would naturally lead to the question of 

 correlation between that character and endosperm texture, 

 a subject which in these crosses is very complicated and 

 for the treatment of which the results thus far obtained 

 are inadequate. It may be said, however, that with the 

 exception of Ped. 1965 the proportions of colored to white 

 seeds in all the ears bear out the assumption that the in- 

 heritance of the aleurone color is governed by two factors, 

 both of which must be present to produce color. In Ped. 

 1965, however, which was grown from a colored seed, ocly 

 23 of the 249 seeds were white. The colored and white 

 seeds are beautifully distinct with no intermediate or 

 doubtful seeds. The ratio of 9.2 per cent, white might be 

 explained as an approximation to the dihybrid ratio of 

 6.25 per cent, but we must then admit that instead of both 

 factors being necessary for the development of color 

 either factor alone may produce color. 



The 19 ears from horny seeds that showed all three 

 classes are assumed to have the same gametic composi- 

 tion as the original second xenia generation, previously 

 reported. The numbers are given in Table IV. The last 

 column of the table gives the odds in 1,000 that deviations 

 equal to those observed are not chance deviations from the 

 expected proportions, as calculated by Pearson's formula. 

 Thus in Pedigree 1953 the odds are 809 to 191, or practi- 

 cally 4 to 1, that the deviation is not the result of chance. 



As in the original ears, the approximation is sufficiently 

 close to render futile any attempt to predicate a different 

 arrangement of factors, but many of the deviations are 

 too large to be ascribed to chance. In the totals the sweet 

 class is too low and the waxy too high, in fact there is no 

 significant difference between the totals for these two 

 classes. The deviation from the expected is, however, 



