BIOMETRY 59 



measurements would occur as frequently as either of the 

 others, so that the middle value would be a fair representa- 

 tive of the class and could be used in statistical computations 

 as the class value with entire propriety and accuracy. Weight 

 classes are also formed of three kilos extent in classifying the 

 weights. The numbers of individuals found in each height 

 class are shown in the totals at the bottom of Table 1. The 

 largest number of individuals is found in the class, 173-175 

 cm., viz., 188. On either side of this class the numbers of 

 individuals (called frequencies) fall off steadily reaching a 

 frequency of four in the shortest class and of one in the tallest 

 class. In Fig. 9 the relative frequencies of the height classes 

 are shown graphically, each column of the figure being pro- 

 portional in altitude to the frequency of the class which it 

 represents. This method of representing variation is called 

 the " method of loaded ordinates." By joining the tops of 

 the several columns of the figure, as in the dotted line, a so- 

 called variation curve is obtained. 



The class with greatest frequency in a group of variates is 

 called the mode, i. e., the fashionable class. It has, of course, 

 the tallest ordinate in the variation figure (class 174, Fig. 9). 



A classification of the same one thousand students as re- 

 gards weight is given in the totals at the right of Table 1, 

 and a graphic presentation of the same data in Fig. 10. The 

 modal class is that which has as its middle value sixty-three 

 kilos. This has a frequency of one hundred and fifty-four 

 with the two adjacent classes almost as large and more remote 

 classes diminishing in frequency to minima in classes forty- 

 five and one hundred and five. The falling off is more rapid 

 to the left than to the right of the modal class, so that in all 

 there are only six classes below the mode but there are four- 

 teen classes in the range of variation above the mode. This 

 results in a '* skew " or asymmetrical curve obtained by 

 joining the tops of the ordinates (dotted line. Fig. 10). The 

 variation curve for the height measurements (Fig. 9) was 

 also slightly skew, but its skewness was much less than that 

 of the curve for weight. 



