The aim of this study was to investigate whether there are familial influences on cortisol levels at baseline and in response to endurance exercise training and, if so, whether there is evidence for a major gene effect. There were 476 white individuals in 99 nuclear families and 247 black individuals in 105 families with valid cortisol data in the HERITAGE Family Study. Data adjustments were carried out separately in each of 8 sex by generation by race groups, using stepwise multiple regression procedures. The familial factors underling the variability in baseline cortisol (log-transformed and adjusted for age and baseline body mass index [BMI]) and its training response (post-training minus baseline, adjusted for age, baseline BMI, and the baseline cortisol value) were assessed by estimating familial correlations and carrying out segregation analysis. In the white sample, significant familial resemblance was detected for both baseline cortisol and the training response, with maximal heritabilities of 38% and 32%, respectively. However, significant familial correlations were not detected for either cortisol phenotype in the black sample, perhaps owing, in part, to the much smaller family sizes. Results of segregation analysis of the white sample provided evidence for Mendelian additive genes influencing baseline cortisol and its training response. The major genes accounted for 33% and 31% of the variance for baseline cortisol and the training response with 48% and 5% of the sample homozygous for the genotype leading to high values, respectively. In conclusion, we found significant familial effects influencing levels of baseline cortisol and its training response in the white sample. The putative major gene effects appear to explain most of the observed familial resemblance, this will motivate further linkage and association studies.