One of the principal objections to the use of census survival ratios for estimating net migration is the error that must arise from geographic variations in enumeration error and in mortality rates around the national averages. The possibility of reducing this type of error emerged with the tabulation, in two successive censuses, of birth-residence statistics for the native population of the United States by age, sex and color. By treating each group of persons having a common area of birth as a closed population, one can derive census survival ratios for the decade 1950-1960 that are specific for area of birth as well as for age, sex and color.