1886.] Family Likeness in Stature. 47 



Fig. 3. 



from its axis) by the name of " deviates," and to describe either of 

 those two symmetrical deviates of the normal ogive that stand at 

 + 25° by the name of " quartile deviate," or, more briefly, " quartile." 

 I also give this name to the mean length of the upper and lower 

 quartile, in those ogives which are drawn from observed data, and 

 which are not strictly symmetrical. The numerical value of the 

 quartile is identical with that of the well-known but here inappro- 

 priate term of " probable error." 



Construction of Stature- Schemes and of Ogives from Observations. — 

 The method of drawing an ogive from observations of stature is as 

 follows. The observations (see Tables III, IV, and V, and compare with 

 VI and VII) are sorted into grades, such as " . . . cases of 60 inches 

 and under 61," " . . . cases of 61 inches and under 62," &c. If we are 

 constructing a stature-scheme, or desire to obtain the median value of 

 the series, we have to consider these values of inches, but in con- 

 structing no more than an ogive, which is only the upper boundary of 

 a stature-scheme, it suffices to consider them as successive grades 

 of 1 inch each, and I reckon the first grade not as 0, but as 1. 

 This has been done in column A, Table VI, for the sake of 

 treating different groups on a uniform plan. The number of 

 cases in these grades are then summed from the beginning, and 

 the sum, up to each grade inclusive, is written down, as shown 



