CERTAIN LAWS OF VARIATION. 211 



VT r 2 .* For instance, we have seen that in the case 

 of man and likewise also in other organisms the co- 

 efficient of correlation between parent and offspring 

 is .3, and between mid-parent and offspring .424. 

 The variability of all the children of parents of any 

 given stature will therefore be Vl-ri)90 .9539, or 95 

 per cent, of that of the race, and the variability of the 

 children of mid-parents of any given stature will be 

 90.55 per cent, of that of the race. Again, Professor 

 Pearson and his co-workers t have found that in the 

 series of measurements upon plants examined by them, 

 there is no relation whatever between the variability of 

 the species and the intensity of homotyposis. Thus 

 on splitting up all the 22 series of measurements into 

 two groups, the mean variability of the first group (as 

 expressed by the coefficient of variation) was 22.95 per 

 cent., and of the second, 14.28 per cent., or not two- 

 thirds as much. Nevertheless the mean homotypic 

 correlation (or degree of similarity between the undif- 

 ferentiated like organs of an individual, as compared 

 with organs of other individuals of the same race) was 

 .456 in the first group, and .458 in the second; i. e., 

 was practically the same. " Hence there seems," says 

 Professor Pearson, " so far as our researches go, no 

 ground for asserting that increased intensity of heredity 

 means decreased intensity of variation, and vice versa." 

 To return to Mr. Sedgwick's views, if rigid artificial 

 selection can only reduce variability by 11 per cent., 

 then obviously Natural Selection can scarcely have any 



*L. c., p. 397 and p. 472. 

 fPhil. Trans. A. p. 275, 1901. 



