356 ANNALS NEW YORK ACADEMY OF SCIENCES 



to perpetuate this record. On the other hand, the Indians themselves 

 maintain a winter-count, or calendar, by which they keep track of their 

 ages and other more important events. With this calendar, Dr. Walker 

 familiarized himself so that parents could be carefully questioned. All 

 this gives the age of the child as to year ; but in the determination of the 

 birth month and hence the nearest birthday, there is necessarily greater 

 error than in the case of similar measurements for white children. The 

 mixed blood ages, on the other hand, are almost as accurate as those for 

 whites. ISTow we may waive all this and pass to the data themselves. It 

 seems scarcely necessary at this late day to state that, as errors of age 

 are likely to be due to guesses and memory lapses and hence accidental, 

 they will have little effect upon the average measurements, but will in- 

 crease the calculated variability to an appreciable extent. Thus, a com- 

 parison of variabilities for Indian and white children may serve as a 

 check on accuracy in age. For the white standard, we may take values 

 calculated for a publication by Professor Franz Boas and the writer, 

 from which Table 4 has been compiled.^ A comparison with Table 1 

 shows at a glance that the variabilities for stature and weight are, with 

 few exceptions, less for Indian children; for the mixed blood the same 

 tendency is clear, but of less magnitude. As these differences cannot be 

 accounted for by absolute differences in stature and weight, we are forced 

 to the alternative that either these Indians show an almost abnormal 

 lack of variability in the type, or that the ages and measurements as 

 taken are about as accurate as those made on white children. The 

 former seems unworthy of serious consideration. Of course, it is ap- 

 ]iarcnt that these Indian children are less variable than white, but their 

 ages have been given with sufficient accuracy as not to obscure that differ- 

 ence. Further, in giving the age of white children by year, it should not 

 be overlooked that in the group for any year there are children who differ 

 in age almost a full year and that no matter how accurately the ages are 

 known this great difference will tend to conceal differences and disturb 

 the variability. Even in cases where Indians did not give the part of 

 the year of birth correctly, the assignment of the case to the age shown 

 by the year-count would, it is true, give a greater error than where such 

 assignment was made according to the nearest birthday ; but the differ- 

 ence would be of degree rather than kind and affect chiefly the variability 

 in the group, as some children would be somewhat too 3^oung and some 

 too old. 



•Report of the TJ. S. Commissioner of Education for 1004, p. 25. 1005. 



