200 SCIENCE PROGRESS. 



operating in the direction of an exaggeration of the vertical 

 height and probably also a diminution of the breadth of the 

 head. Thus, if we compare, in the adjoining provinces of 

 the north-west and of Behar, castes identical in name, and, 

 we may presume, nearly identical in blood, we get the 



Here the correspondence of the figures in the first and 

 fifth columns is so close that the wide discrepancies in 

 the second, third, and fourth columns become almost in- 

 credible ; and as they invariably run in the same directions 

 of less breadth and greater height, for the castes of Oudh, 

 etc., one is almost irresistibly led to attribute them, at least 

 in part, to the personal equation of the observer. It is not 

 difficult to conceive the habitual error in the horizontation 

 of the head, which has led to the errors in the third and 

 fourth columns : in the second the fault, if fault there was, 

 lay probably in measuring at the spots where the maximum 

 breadth might be expected to be, instead of groping about 

 to find the maximum. 1 I see no internal evidence to lead 

 one to distrust the other measurements, but of course all 

 the indices for Oudh and the N.W. Province, into which 

 the vertical height enters as an element, must have come 

 out too small. 



The very extensive distribution of some of the castes 

 supplies an argument against the ethnological explanation 



1 Or it may be that the Oudh operator put a little more pressure on 

 the callipers than did his fellow-workers. The frontal and bizygomatic 

 measurements for Oudh, etc., are smaller than those for Behar, but not 

 beyond the limits of likelihood. 



