The Victorian Family Heart Study was established to address the causes of familial patterns in cardiovascular risk factors. From 1990 to 1996, a representative population sample of 783 adult families (2,959 individuals), each comprising both parents (40-70 years) and at least one natural adult offspring (18-30 years), was recruited in Melbourne, Australia. Included in both generations were 461 monozygotic and dizygotic twins as pairs or singletons. A multivariate normal model was used for pedigree analysis of height, weight, body mass index, diastolic and systolic blood pressure, pulse rate, and total and high density lipoprotein cholesterol. All traits showed evidence for additive genetic variation, explaining from 55% (height) to 26% (pulse) of age- and sex-adjusted variance. An effect persisting into adulthood of shared family environment during cohabitation explained from 39% (body mass index) to 13% (systolic blood pressure) of variance (not nominally significant for diastolic blood pressure). These shared environmental effects were strongest within twin pairs, less so for sibling pairs, and least for parent-offspring pairs (in which an effect was not observed for weight, diastolic and systolic blood pressure, and total cholesterol). On a background of genetic influences, there are periods in early life during which the family environment cements long-term correlations between adult relatives in cardiovascular risk factors.