EVOLUTION 437 



different groups of husbands and wives and for two very- 

 distinct characters, stature and eye-colour, we have found 

 quite sensible measures of homogamy. We cannot doubt 

 in the face of this that like actually tends to mate with 

 like in the case of man. Whether this arises directly 

 from choice in stature or in eye-colour, or from choice of 

 correlated organs it is not possible to determine. If from 

 the latter source, then it is very probable that the measure 

 of homogamy would be somewhat larger, if we could hit 

 upon the directly selected organ. 



The considerations, however, which it seems necessary 

 to emphasise are these : — 



{a) The possibility of permanent differentiation in a 

 form of life largely depends upon its habit or mode of 

 reproduction. Is the mating pangamic, or is there any 

 form of sexual selection such as autogamy, endogamy, 

 apolegamy, or homogamy, using these terms in their 

 broadest senses ? 



(b) If such exist, we must not merely state opinions as 

 to their possible effectiveness, but following Darwin in his 

 treatment of heterogamy collect statistics and obtain 

 quantitative measurements, which will determine definitely 

 whether they are or are not verce causes of evolution. 



(c) As illustration we have taken a special case. We 

 have asked whether sexual selection exists in man. We 

 have found that preferential mating does take place in 

 eye-colour, and assortative mating in both eye-colour and 

 stature. The assortative mating is indeed so great that 

 husband and wife have for the average of these two 

 characters, a sensibly greater degree of resemblance than 

 that assumed on theoretical grounds to hold for first 

 cousins,^ and approaching the degree of resemblance 

 found for uncle and niece. 



I 4. — Genetic (^Reproductive) Selection 

 In discussing in the previous sections how far various 



^ We shall return to the degree of resemblance between cousins later 

 (p. 481). I say theoretical grounds, for I know at present of no actual statistics 

 on this point. 



