80 American Statistical Association. 



17.6 centimeters [6.93 in.] for males, and 19.1 centimeters 

 [7.52 in.] for females) by the statement tliat his observations 

 were limited to individuals ''reguliercment construits," and 

 that the number of persons subjected to measurement was 

 "pen considerable."* In the introductory portion of the 

 work he describes as followsf his method of ascertaining the 

 proportions of the human bod}' : " I contented m3^self, there- 

 fore, with measuring carefully ten individuals of each age, of 

 the male as well as of the female sex, but choosing them in 

 general of a form which could be regarded as regular. The 

 averages of the different groups gave me the condition of 

 development of man from year to year." J It seems, there- 

 fore, evident that Quetelet's observations were made on a 

 comparatively small number of individuals, selected on 

 account of their more or less close conformity to what was 

 regarded as a normal type. No measurements seem to have 

 been taken, except on persons having a " regular form." 

 This determination of the normal type in advance of the 

 measurements must, of course, have been largely a matter of 

 conjecture, and might well have led to the rejection of per- 

 fectly healthy and normal individuals whose dimensions did 

 not correspond to a preconceived idea of the typical man or 

 woman. It is therefore probable that if Quetelet's observa- 

 tions had been more numerous and less selected it would 

 have been found that the curves of growth of the two sexes 

 in Belgium intersect each other much in the same way as in 

 England and in this community. 



This view derives confirmation IVom the admission of 

 Quetelet,§ that the growth of any one individual is far from 

 being as regular as that indicated by the table of averages. 



* Anthropomotrie, p. 182. 



t Aiitliropom6trie, p. 24. 



X Asa reason for being satisfied with so small a number as ten observations he states that, 

 on dividing the measurements made on thii'ty individuals into three groups of ten each, so 

 chosen that the average hcu/Jits for all three groups were about alike, he found that the 

 other average measurements of these three groups differed from each other less than 

 might have been expected in three successive measurements made on the same individual. 



§ Ajithropometrie, p. 183. 



