The aim of this study was to estimate inbreeding, kinship, genetic distances, and population structure of the island of Hvar using surnames of the population from 10 villages. Total inbreeding on the island, estimated from the frequency of isonymous marriages, amounted to 0.0228. Total kinship (unreduced variance) of the entire island amounted to 0.0411, a value indicative of very isolated populations with accumulated kinship over time. The analysis of genetic relationships based on surnames among the villages did not reveal the division of the island's villages into two clusters, "eastern" and "western", as was the case in previous studies when clustering was based on biological, biocultural, and sociocultural traits. In addition, kinship and genetic distances determined from surnames revealed a poor fit to an isolation by distance model (for kinship, R2 = 14.6%, a = 0.1101, b = 0.3573; for genetic distances, R2 = 0.0%, a = 0.5686, b = 121.52), which is in contrast to the results of previous analyses based on other traits. In order to estimate how closely parameters obtained from surnames represented real interpopulational distances, matrices of kinship and genetic distances between 10 villages (determined from surnames) were correlated to corresponding distance matrices based on migrational, sociocultural, anthropometric, physiological, dermatoglyphic, and genetic traits that were previously reported in the literature on the very same samples. The kinship matrix correlated poorly with other distance matrices, while the genetic distance matrix determined from surnames was in a good correlation with distance matrices based on linguistics (r = 0.65 for basic and r = 0.63 for cultural vocabulary), metacarpal bone dimensions (r = 0.54) and genetics (determined from systems of erythrocyte antigens and isoenzymes, r = 0.45). Am. J. Hum. Biol. 9:595-607, 1997. © 1997 Wiley-Liss, Inc.