Familial correlation and segregation analyses of forced vital capacity (FVC) were performed on data from 309 nuclear families with 1,045 individuals in the town of Humboldt, Saskatchewan, in 1993. FVC was preadjusted for age, height, and weight in four separate groups (mothers, fathers, daughters, and sons). Residual FVC was standardized within the four groups. Class D regressive model was first used to examine the familial resemblance of FVC without a major gene. While mother-father correlation was not significantly different from zero and mother-sibling and father-sibling correlations were not significantly different from each other, sibling-sibling correlation was greater than parent-sibling correlation. Segregation analysis for all 309 families indicated that both the Mendelian and no-parent-offspring-transmission models fitted the data as did the general model with arbitrary transmission probabilities. Likelihoods under the Mendelian model (LMendelian) and the environmental model (Lenvironmental) were calculated. Based on the value of In(LMendelian/Lenvironmental), 309 families were divided into two groups: 196 families with the value of In(LMendelian/Lenvironmental) greater than zero (group I) and 113 families with the value In(LMendelian/Lenvironmental) less than zero (group II). The Mendelian transmission model without familial correlations was the most parsimonious model for the families in group I. For group II, there were two best models of choice: 1) no-parent-offspring-transmission model with possible heterogeneity plus familial correlations [Akaike's information criterion (AIC) = 1,213.76] and 2) Mendelian transmission plus sibling-sibling correlation model (AIC = 1,202.36). The results suggest there are major genetic mechanisms in FVC with possible heterogeneity.