70 American Statistical Association. 



nationalities wliicli would give at each age a sufficient num- 

 ber of observations to justify the calculation of an average. 

 The selection was of course at the outset to a great extent 

 conjectural and tentative, and could be definitely made only 

 as the work progressed. It was finally decided to limit the 

 tabulation by nationality of parents to the following groups: 

 I. Parents, both American. 

 II. Parents, both Irish. 



III. Parents, one American and one Irish. 



IV. Parents, both German. 



V. Parents, one or both English. 



In the last three of these groups the observations were not 

 sufficiently numerous to establish the rate of growth with 

 very great precision ; but the results have a certain value as 

 approximations to the truth. 



The observations thus tabulated were placed in the hands 

 of professional accountants, by whom the average heights 

 and weio-hts for the different ao-es and nationalities were cal- 

 culated both in the English and in the French systems of 

 weights and measures. The results are given in Tables 

 Nos. 1 and 2, at the end of this article. 



It will be noticed that it has been assumed in this investi- 

 gation that the rate of growth of children may be ascertained 

 by computing at any one time the average height and weight 

 of children of different ages, as well as b}' determining the 

 average height and weight of a given set of children in suc- 

 cessive years. This assumption is doubtless perfectly justi- 

 fiable, though certain theoretical objections may be urged 

 against it. It may be said, for instance, to involve the further 

 assumption of the prevalence at any given time of equally 

 favorable conditions for the growth of children of all ages. 

 It is, however, conceivable that at a certain time particularly 

 favorable or unfavorable conditions for the growth of young 

 children may prevail, while the growth of older children may 

 be less affected. The rate of growth determined by observa- 

 tions taken at this time will therefore show a deviation from 



