142 BUREAU OF AMERICAN ETHNOLOGY [Bull. 197 



The first of these characteristics to note is the median age, which 

 was calculated for each of the 31 distributions, by sex. Comparing the 

 male medians with the female medians for each of the Navaho age 

 distributions, we find that, in general, the female medians are slightly 

 liigher than the corresponding male medians. The same feature can 

 be noted in the medians derived from each of the 10 model age distri- 

 butions given in table 36 (p. 178). One might conclude that there is 

 no apparent sex bias in the reporting of the Navaho population (or in 

 its enumeration). Slightly higher female medians can be viewed as 

 a reflection of slightly higher female life expectancy, which is a com- 

 mon characteristic of the populations whose mortality experience is 

 typified in the model life tables used to develop the 10 model age 

 distributions. 



Significant underenumeration of adult Navaho males would, how- 

 ever, produce the same eli'ect. Actually, only 1 of the 25 Navaho 

 age distributions, that of the Kamah coimnunity, 1880 to 1898, pro- 

 duces a median age for males which is more than 1 year below the 

 corresponding female median. It should be noted that the same popu- 

 lation, from 1920 to 1948, displays a reversal of this relationship, with 

 a male median that is over 1 year above the corresponding female 

 median. Definitive conclusions are unwarranted by these findings. 

 The variation observed in the Eamah population may well reflect 

 actual changes m the sex-age composition of this small community; 

 in any case, it can be attributed also to the wide range of probable 

 error associated with small frequencies.^" 



For the period prior to 1920, we have six sets of Navaho median 

 ages. The first two of these (1880-98; 1900-1918) pertain to the 

 Ramah community. Both of these sets of medians are lower than any 

 of the others given in table 28. It is possible to dismiss these low 

 values as a function of the chance variation associated with the very 

 small frequencies in this population. Such an interpretation is not 

 refuted by the observations which follow, but these observations do 

 lend a certain credibility to the Eamah statistics. 



The first observation to be noted pertains to the extremely careful 

 procedures whereby the Ramah age data were obtained. The 

 anthropological researchers who collected these data were attempting 

 to reconstruct the Ramah population, family by family, back to its 

 origin around 1870. In this reconstruction, detailed genealogical 

 records were prepared of every family in the Ramah commimity, 

 including carefully authenticated and verified information on the 

 dates of birth and death of each of its members. The basic data that 



80 As Is explained In table 31, footnote 3, p. 166, the age distributions of the Ramah 

 Navaho are synthetic figures obtained by summing the age-speciflc frequencies reported 

 annually during the period specified. Since each of these totals represents the sum of 16 

 annual figures, the average size of the actual Ramah population during any of the specified 

 periods would approximate one-sixteenth of the values shown in tables 28 and 31. 



