CASTLE. — LAWS OF HEREDITY. 225 



color follows closely neither tbe law of Galtou nor the series suggested 

 by Pearson. 



The lack of at^reement in this case with Gallon's law has been pointed 

 out by Davenport for certain of the color categories. lie concludes that 

 in the case of gray alone does the color inheritance among von Guaita's 

 mice conform closely with Gallon's law. But in reality, even in the 

 <-:ise of gray, close agreement does not occur ; Davenport's conclusion 

 lliat it does occur results from the inclusion by him in a single 

 color category of two sorts of mice which are clearly quite distinct, 

 namely, (1) mice gray all over like the wild house-mice, and (2) giay 

 mice witii white markings. Even when these two categories are com- 

 bined, Davenport's figures show close agreement between the observed 

 and calculated numbers in two only of the five filial generations with 

 which he deals, namely, in the third and sixth generations (of von Guaita's 

 nomenclature), in which he finds observed and calculated to agree per- 

 fectly. But in the three remaining generations he finds observed and 

 calculated percentages to be related as follows : — 



Generation II. IV. V. 



Observe.! 100% 58% 48% 



Calculated 0% 48% 60% 



Davenport, moreover, has excluded from the category of ''albinos" 

 white mice which possess the dancing character. But this is manifestly 

 an error, for the dancing character has nothing to do with coat-color, 

 and is inherited quite independently of it. Davenport's classification, 

 accordingly, makes the category of albinos appear smaller than it really 

 is. If we include all albinos (whether dancers or not) in one category, 

 and make separate classes for gray, gray-white, black, and black-white 

 mice, the relations between the observed and calculated numbers in each 

 generation are found to be as indicated in Table I, Davenport's method 

 of calculation being followed. 



An examination of this table shows no close agreement between cal- 

 culated and observed conditions throughout any single category or any 

 single generation, althougli the totals turn out better than the predic- 

 tions for the generations consiilered separately. In Generation II the 

 discrepancies are glaring. In the column, white, the grand totals alone 

 agree closely, yet this agreement is clearly without significance; it is a 

 chance agreement in the totals of two series divergent throughout. 



The observed numbers, it is evident, agree no better with one of 

 Pearson's series than with that of Galton. The discrepancies noted 



VOL. .XXXIX. — 15 



