2<>S 



THE AMERICAN NATURALIST [Vol. XLIII 



IX, a, the occurrence of red is not in accord with the 

 hypothesis. Certainly there is no blending in an " all N " 

 ancestry of red and yellow as in the Earn, Ric-C and Wol 

 families of Table VII, b. The results clearly do not 

 accord with the law of blending inheritance. 



It remains to consider if inheritance of hair color fol- 

 lows Galton 's law which states that the two parents deter- 

 mine 50 per cent, of the ancestry, the four grandparents 

 together 25 per cent, and earlier ancestry altogether form 

 the remaining 25 per cent. 



Taking the 12 families of Tables X, c, whose grand- 

 parents are all given, we have in the parentage a total of 

 red in 24 parents ; and of 5 red in 48 grandparents, or 

 about 10 per cent. Assuming the same proportion of 

 red in the unknown earlier ancestry, we have the total 

 expected proportion of red in the offspring given by 

 the sum 



Actually, there are about 40 per cent, of the red type, 

 and, making every allowance, at least 18 per cent, are to 

 be expected. Five per cent, certainly fits the facts very 

 badly. We conclude, therefore, that Galton 's law does 

 not fit the facts as well as Mendel 's law and that heredity 

 of hair color is alternative. 



E. The Non-transgressibility of the Upper Limit 



While the application of the law of alternative inherit- 

 ance to human hair color lacks something in ideal clear- 

 ness and precision, one general rule stands out promi- 

 nently. It is that in the midst of the varying degrees of 

 intensity of the melanic pigmentation the intensity of the 

 melanic pigmentation of the offspring never exceeds that 

 of the more intense parent. 



The general intensity relations of melanic hair pigment 

 in parents and offspring are brought out in Table XI. 



An inspection of this table shows that, in general, in 



24 x 50 : 

 .104 X 25 : 



X oil = 11 



