The familial etiology of the response in total fat mass (FM) and abdominal visceral fat (AVF) to 20 weeks of exercise training was investigated in families participating in the HERITAGE Family Study. AVF (measured by computed tomographic scanning) add FM (measured by underwater weighing techniques) were assessed at baseline (in a sedentary state) and after 20 weeks of exercise training. The response AVF (AVF(Δ)) and response FM (FM(Δ)) were computed as the simple delta values (posttraining - baseline) and adjusted for the effects of sex, generation, and a polynomial in age using multiple regression analysis. To index the AVF response independently of the response in FM and the initial level of visceral fat, the AVF(Δ) was also adjusted for age and baseline AVF (AVF(B)) and FM(Δ). Familial correlation analysis was used to investigate the multifactorial familial effects (polygenic and/or familial environmental), and segregation analysis was used to search for major gene effects. For the age-adjusted AVF(Δ), a putative recessive locus accounting for 18% of the variance (q2 = 1%) was detected. Adjusting AVF(Δ) for AVF(B) and FM(Δ) slightly increased the percentage of variance accounted for (to 26%, q2 = 3%) but did not radically alter the pattern of the parameter estimates. For FM(Δ), a putative dominant locus accounting for 31% of the variance (q2 = 49%) was noted. In conclusion, the results were consistent across methods in suggesting that there is little evidence of a multifactorial heritability for either AVF(Δ) or FM(Δ). Rather, the familial etiology of the response to exercise training appears to be primarily due to putative major genes (a recessive locus for AVF(Δ) and a dominant locus for FM(Δ)). In addition, a pleiotropic/oligogenic system urderlying these variables was inferred. That is, the putative loci for FM(Δ) and/or AVF(B) also may impact the AVF(Δ), with an additional independent major locus effect on AVF(Δ) after the former influences have been removed.