TRANSACTIONS OF SECTION D. 



895 



crabs will be produced by that quarter of the present generation in which the 

 antero-lateral margin is longest. And as the oiFspring will inherit a large per- 

 centage of the parental character, the mean of the race may be sensibly raised 

 in a single generation. 



This view of the possible effect of selection seems to have escaped the notice of 

 those who consider that favourable variations are of necessity rare, and likely to 

 be swamped by intercrossing when they do occur. You see that in this case there 

 are a few individuals considerably different from the mean in either direction, and 

 a very large number which differ from the mean a little in either direction. If 

 such deviation be associated vnth some advantage to the crabs, so that crabs 



Fig. 4. 



400 



300 



200 



100 



400 



300 



200 



100 



123 45 6789 IC 11 12 



Diagram showing the number of Miiller's glands in each of 2,000 female swine. 



which possess such abnormality are mora fertile than those which do not, it is a 

 certainty that the mean character of the next generation will change, if only a 

 little, in the direction advantageous to the race ; and the opportunity for selective 

 modification of this kind to occur in either direction is very nearly the same. 



In the next case, this is not true. 



The diagram (fig. 4) represents the number of female swine, out of a batch of 

 two thousand examined in Chicago, which have a given number of MiiUerian 

 glands in the right fore-leg. The distribution is much more skew than in the case 

 of the crabs, and you see again the very beautiful way in which Professor 

 Pearson's curve expresses it. You see that the range of variation is much greater 



