90 STANDARD OF COMPARISON. 



that our comparisons were between reproductions a century apart, 

 it will be seen that there is another reason for thinking that our 

 standard is, if anything, abnormally high. To still further test 

 the matter, I compared the average ages of the marriages from 

 which our standard of births are taken with the ages recorded in 

 marriage licenses issued in Chicago at different times during 1900, 

 and I find that they are almost identical, although there is a differ- 

 ence of more than a century of time between them. 



Though there are so many reasons for considering the adopted 

 standard as being high, I have still decided to retain it because it is 

 a definite and known standard of known accuracy, and because, if 

 the men measured by it are found to be high in comparison to it, 

 it will be known that they are absolutely high, and will be relatively 

 high as compared to any standard that may be made from the mass 

 of human beings. 



INTERPRETATION OF MEASUREMENTS BY THE SCALE. 



Having adopted a standard of birth-ranks, and having divided 

 this standard so that it becomes a scale of equal divisions, the law 

 of probabilities declares that if we take any miscellaneous group of 

 men and find their birth-ranks, it will be found that they are pretty 

 evenly distributed along the length of the scale. A deduction from 

 this law is that if we take a selected group of men and compare 

 their birth-ranks with a standard scale of birth-ranks, then if we 

 find that there is an unusual accumulation at a certain part of the 

 scale or an unusual absence of cases at some other part of the scale, 

 this accumulation or this absence must be in some way connected 

 with the manner in which that group of men was selected. This de- 

 duction is very old and well known. Aristotle recognized it when he 

 held that anything which occurs regularly cannot be the result of 



