viii.] DISCUSSION OF THE DATA OF EYE COLOUR. 153 



seen that the calculations in Class III. are by far the 

 most precise. In more than one-half of those calcula- 

 tions the error does not exceed 0*5, whereas in more than 

 three-quarters of those in I. and II. the error is at least of 

 that amount. Only one-quarter of Class III., but some- 

 where about the half of Classes I. and II., are more than 

 1*1 in error. In comparing I. with II. , we find I. to 

 be slightly but I think distinctly the superior estimate. 

 The relative accuracy of III. as compared with I. and 

 II., is what we should have expected, supposing the 

 basis of the calculations to be true, because the addi- 

 tional knowledge utilised in III., over what is turned 

 to account in I. and II., must be an advantage. 



My returns are insufficiently numerous and too 

 subject to uncertainty of observation, to make it worth 

 while to submit them to a more rigorous analysis, but 

 the broad conclusion to which the present results 

 irresistibly lead, is that the same , peculiar hereditary 

 relation that was shown to subsist between a man and 

 each of his ancestors in respect to the quality of 

 Stature, also subsists in respect to that of Eye-colour. 



