Using a computerized genealogical database, inbreeding coefficients were calculated for a sample of 435777 Utah Mormons. The population was divided into ten ten-year birth cohorts (1846-1945) and 22 geographic subdivisions in order to assess temporal and spatial variation in inbreeding. The average inbreeding coefficient for this population is 0.000 106. The average within-groups random kinship coefficient is 0.000 312, reflecting consanguinity avoidance. Random kinship matrices were formed by estimating the average kinship within each spatial subdivision and between all pairs of subdivisions. These matrices were compared statistically with kinship matrices previously estimated using migration matrices and isonymy data. The isonymy approach consistently overestimates random and total inbreeding as well as Wright's Fst. This can be attributed primarily to the assumption of monophyletic origin of surnames. The migration matrix method underestimates random inbreeding and Fst. This is due mainly to the assumption that outside immigrants are derived from a genetically homogeneous population. While the absolute values of the kinship coefficients estimated by each method differ substantially, the patterns of between-groups kinship coefficients given by each method are highly congruent. Logistic and linear regression analyses of 85,235 marriages demonstrate that consanguinity is significantly dependent upon year of marriage, geographic distance between husband's and wife's birthplaces, and the population size of husband's and wife's birthplaces.