INTRODUCTION 33 



tion than twice this amount would in a table such as the femur length, 

 where the lengths involved run between four and five hundred millimeters. 

 In order, then, to directly compare deviations from different tables, we 

 need some sort of index which expresses the relation of the actual devia- 

 tion to the actual mean. Such an index is the Coefficient of Variation, or 

 "Coefficient of Dispersion," as it is also called, which uses the mean as 

 the standard ( = 100), and compares with it the deviation, as follows: 



~ „ . , __ . ,. Deviation X 100 



Coefficient of Variation = ^ — — . 



Mean 



In the example used here, where the average or Mean of the Cristal 

 Breadths of 100 students is 271.20 mm., and the average deviation of the 



items is 13.26, the Coefficient of Variation is jyrTi™ or 4.9 % of the mean. 



This figure gives thus the proportionate amount of variation exhibited 

 by a certain measurement, as taken in a number of individuals, and may 

 be compared directly with a similar figure taken from a different list, 

 involving a totally different measurement. By such methods the amount 

 of variation (proportionately) in the radius length could be compared 

 with the amount of variation in the femur length, and the conclusion 

 definitely taken as to which is relatively the more variable, in spite of 

 he actual differences in the lengths of the two bones. 



