176 BUREAU OF AMERICAN ETHNOLOGY [Bull. 197 



matehed births." This refers to the percentage of infants under 3 

 months of age (in 1950) enumerated by the census takers whose infant 

 cards were successfully matched with birth certificates on file at the 

 appropriate State registrar's office.^^ 



The results of these tests can be summarized briefly. In 1940, 

 approximately 40 percent of all births recorded by the census enumer- 

 ators as occurring among the Indian populations of Arizona and New 

 Mexico during the 4 months prior to the census date were matched 

 with birth certificates on file at the respective State registrar's offices. 

 In 1950, the corresponding percentage of matched births rose to about 

 60 (Hadley, 1952 a, table 1). 



A closer analysis of the results of the 1940 test, focusing upon the 

 six counties in Arizona and New Mexico whose non-White population 

 was predominantly Navaho, indicated that only about one-third of 

 known biiths occurring among the non- White population of this area 

 (as recorded by the census enumerators) could be matched with birth 

 certificates at this time. In 1947, a further effort was made to match 

 Navaho births and deaths that were known to agency officials against 

 birth and death certificates on file at that time. This test indicated 

 that 43.7 percent of known births, and 77.9 percent of known deaths, 

 were supported by appropriate certificates in 1947 (Hadley, 1951, 

 p. 5). 



The results of these several tests are admittedly inconclusive. In 

 his discussion of the findings of the 1947 study, Hadley made pointed 

 reference to the difficulties in matching that are associated with the 

 peculiarities of Navaho nomenclature. He states that some of the 

 birth certificates that were duly filed could not be matched solely be- 

 cause the name of the child recorded thereon differed from the name 

 as reported to agency officials. Although the registration of deaths 

 was evidently far superior to the registration of births at this time, 

 the Navaho predilection for plural names undoubtedly impaired the 

 effectiveness of the attempt to match known deaths against death 

 certificates as well.^'' 



^^ It should be noted that the results of such a test cannot be entirely conclusive, espe- 

 cially in areas experiencing severe underregistration of births, since the probability that 

 a birth will not be registered is related to the probability that the same infant will not be 

 enumerated. 



1" The results of the 1940 matching test, as they pertain to American Indians throughout 

 the United States, are summarized in Hadley, 1950. 



The inconclusiveness of these tests, as applied to the Navaho, can be illustrated by 

 adjusting the birth and death rates derived from the reported Navaho births and deaths 

 in 1947 according to the percentages of birth and death certificates that were matched. 

 Such an adjustment would increase the crude birth i-ate from 22.5 to 51.4 per 1,000, and 

 the crude death rate from 11.1 to 14.2 per 1,000, giving a crude rate of natural increase 

 of 3.72 percent per year at this time. 



It should be noted, in this connection, that the underregistration (or delayed registra- 

 tion) of births ma.y produce a disproportionate reduction in the derived death rate. This 

 effect is produced because the majority of infant deaths occur soon after birth. Hence 



