SELECTION IN MAN. 389 



has directed attention ; but if we assume that the con- 

 clusions which result from them are at all approximately 

 Correct, it follows that the blonds in America have less 

 chance than the brunets of contributing their due propor- 

 tion to the next generation. Under these conditions the 

 blonds ought to diminish relatively, and the brunets to 

 increase ; and accordingly we find that of accepted soldiers 

 there were among the white natives of the United States 

 about (per cent.) 



66 light and 34 dark complexioned, but 

 among the English 70 ,, 30 ,, 



Irish 70 „ 30 



„ Germans 69 ,, 29 „ 



Thus the men of American birth yielded a larger pro- 

 portion of brunets than those of any of the nations that 

 had most largely contributed to their ancestry, which is 

 nearly equivalent to saying that the Americans are more 

 generally dark complexioned than their ancestors were. 

 Gould (quoted by Ripley) found that the natives of the 

 eastern states were also darker than those of the west. 

 But whether this last fact is occasioned by the parentage 

 of the western men being more directly European, or 

 whether it is connected with the more migratory character 

 of the blond type, must be left for the present undeter- 

 mined. 



Of European evidence on the relation of complexion 

 and disease there is, so far as I am aware, no great amount. 

 My own observations have shown that it is a mistake to 

 suppose, as many do, that light-haired persons are in 

 England more liable to phthisis than others. I have also 

 pointed out that cancer is more common in persons of dark 

 complexion, and in this I am supported by the observations 

 of Dr. Roger Williams. This last fact has, however, very 

 little bearing on the subject in hand, for as cancerous disease 

 usually attacks persons who are beyond the child-producing 

 age it can have very little effect on the proportions of the 

 different complexions of the next generation. 



As we possess for France not only elaborate recruiting 

 statistics, with numerical lists for the principal disqualifying 



