PREPOTENCY 571 



but upon the theoretical ground that males lead a more active 

 life and take the lead in sexual selection. The data just cited 

 seem to substantiate this assertion. For the species and charac- 

 ters involved it appears that male offspring follow more closely 

 the parental type than do the female, and (which amounts to 

 the same thing) female offspring, or sisters, are more nearly alike 

 than are male offspring, or brothers, tending to the conclusion 

 that males are more variable than females. 



Pearson, 1 however, records data of an exceedingly exhaustive 

 series of investigations of variability in men and women, not 

 absolutely but relatively, as expressed in the coefficient of 

 variability. 2 While he finds men more variable at certain ages 

 and in certain characters, yet he does not find pronounced and 

 decided differences, nor are these the same for different races 

 of men. He concludes, on the whole, that for all races studied, 

 ancient and modern, and for all characters covered by the 

 studies, there is " no evidence of greater male variability, but 

 rather of a slightly greater female variability." 



He finds, for example, that English men are slightly more 

 variable as to height than are English women (4.07 : 4.03), but 

 that, among Germans, women are considerably more variable 

 than men (4.26:4.02), as they are also among the French 

 (4.35:3.88). 



In the long bones sometimes one is more variable, sometimes 

 the other ; thus, as to the femur, men are more variable : Libyan 

 (5.05:4.46), French (5.05:5.04), Aino (4.65:4.18), and Neo- 

 lithic man (4.73 :4-5i) ; but with the ancient inhabitants of the 

 Canary Islands the reverse is true (men, 4.64; women, 4.71). 

 In all cases examined, except the French (men, 4.975 ; women, 

 5.365), the tibia is more variable in men; but in most cases the 

 humerus and radius are more variable in women. 



1 Pearson, Chances of Death, I, 256-377. 



2 Manifestly the coefficient of variability is the only correct estimate of com- 

 parative variability, because in its calculation each instance is compared with its 

 own type as a base. This is necessary, because the stature of women, for example, 

 is different from that of men ; hence the two cannot be compared on any common 

 basis. This is the only way, for instance, as Pearson points out, in which we can 

 compare the variability of man with that of the elephant ; in any other way the 

 elephant would appear more variable, because he is bigger. 



