Johnston] NAVAHO POPULATION 81 



Thirdly, the listing procedures were themselves mappropriate in 

 that no recognition was given to the niatrilineal system under which 

 the typical Navaho family is organized. Instead, each family group 

 was listed according to the surname of the male head ; i.e., patrilin- 

 ealiy. In theory, such a procedure would not necessarily create in- 

 superable diliiculties in the actual preparation of the roll, but it could 

 and did greatly complicate the task of maintaining the roll in later 

 years when it became necessary to record additional family members 

 according to their father's name when the person in question was 

 likely to be known and recorded elsewhere according to his mother's 

 name. An incidental feature of the listing procedure added still fur- 

 ther complexity to the problem of identifying individual enrollees: 

 Navaho names were crudely "translated" by the recording clerks into 

 some sort of English equivalents, but the result was a bewildering 

 variety of spellings. In not a few instances, the clerks apparently 

 abandoned any attempt to record the Navaho name given them, and 

 merely "assigned" a common English name to the person in question. 

 Still further evidence of the artificiality with which the Navaho 

 population was classified in the familiar categories appropriate to 

 American social structure is to be found in the total absence of 

 polygynous marriages on the roll. It is possible that plural wives may 

 have been listed as daughters in a few cases, but the impression of the 

 reported ages does not permit any definite conclusions in this regard. 



A final source of error on the 1885 roll is to be found in the record- 

 ing of ages. The roll was prepared on a large ledger, similar in de- 

 sign to the conventional enumeration schedules whereon all informa- 

 tion for a given individual can be entered on a single data line. The 

 male head of a given family was listed first, followed by his wife and 

 their children. In the 1885 roll, however, this procedure resulted in 

 an interesting bias. Most Navaho families appear to have reported 

 their children by sex rather than by age, givmg (in most cases) the 

 names of their female children first. Since the age of each person 

 listed had to be recorded, and the Navaho seldom knew or could 

 communicate their own or their children's ages effectively, the record- 

 ing clerk apparently adopted the practice of merely totaling the 

 number of children reported for a given family, assigning some plaus- 

 ible age to the first child listed, and then assigning ages to all the 

 succeeding children at 2- or 3-year intervals. 



As a result of this intriguing innovation, only 5 of the 70 families 

 contained in a 5-percent sample of the 1885 roll have children of 

 both sexes with overlapping ages. Of the remaining 65 families, 

 49 have all their female children recorded as being older than all of 

 their male children, and 16 have all of their male children recorded 

 as being older than all of their female children. This listing bias 



