This table represents the measurement of 294,893 recruits. Unfortu- 

 nately and in obvious disregard of scientific requirements the results for 

 the measurements for ages 25 and over are given as a group. The average 

 for such a group including all ages up to the military limit must neces- 

 sarily be misleading and utterly inconclusive, since this group includes 

 125,323 measurements, or 42.5 per cent, of the total measurements, 

 at all ages. The scientific error of returning the results for ages 25 and 

 over in the aggregate is of sufficient magnitude to seriously impair the 

 value of the returns as a whole. For the purpose of ascertaining the 

 frequency distribution of stature in relation to age, the returns should 

 have been given by single years of life, certainly up to age 40; and if 

 this had been done, the results would have been much more valuable 

 in their application to practical uses than as published annually in their 

 present form in the Surgeon General's report. 



Fitting the above data to frequency curves, Mr. Fisher has 

 obtained for me the following values for the various statistical 

 parameters : 



FREQUENCY DISTRIBUTION OF STATURE, UNITED STATES 

 ARMY RECRUITS 



1906-1916 

 Statistical Parameters 



Dispersion Skewness Bxcess 



1.855 inches -0.0283 -0.0269 



1.898 ' -0.0235 -0.0275 



1.936 ' -0.0018 -0.0290 



1.929 ' 0.0009 -0.0303 



1.943 ' 0.0024 -0.0310 



1.939 ' 0.0004 -0.0325 



1.948 ' 0.0027 -0.0321 



1.948 ' 0.0011 -0.0317 



The theoretical frequency distribution per 1,000 recruits at single 

 years of life under age 25 and at ages 25 and over, considered as a 

 group, is as follows : 



FREQUENCY DISTRIBUTION OF STATURE, UNITED STATES 



37 



