II] VARIATION 11 



frequent heights were 67 and 69 inches, and that the 

 more the stature differed from 68 inches in either 

 direction, the fewer would be the men having that 

 measurement. This could be represented graphically 

 by arranging vertical lines representing the heights 

 of every man in order of their height; a line joining 

 their tops would then rise rapidly at the lower end, 

 would be nearly flat as it passed over the men having 

 heights near the 'modal value' of 68 inches, and 

 would rise again steeply to the exceptionally tall 

 men at the upper end of the row (fig. 1). [13]^ 



A more instructive method of graphically repre- 

 senting the distribution of variation is to take a base- 

 line and divide it into equal parts, each representing 

 an equal increment in the structure measured. From 

 each division of the base-line a vertical line is drawn 

 representing by its length the number of individuals 

 having that measurement. 



In the imaginary case taken above, the base-line 

 would have 17 divisions representing successive 

 heights of from 60 to 7Q inches; at each division 

 a vertical line is drawn which by its length repre- 

 sents the percentage of the population which have 

 that height (fig. 2). By joining the tops of the per- 

 pendiculars (ordinates) a curve, or more strictly a 

 polygon, is obtained which graphically represents the 



1 For references see the end of the Volume. 



