The maintenance over time (4 months) and generalizability (to a work setting) of laboratory-based heart rate reactivity control training was investigated. Forty-three subjects received 5 weeks of biofeedback, imagery, and breathing training and 41 subjects acted as controls. Although there were no significant differences between the reactivities of the control and those of the treatment subjects at any of three pretraining tests, significant differences were noted immediately following training, 7 weeks, and 4 months later, showing maintenance effects. Data collected during work activities on the frequency of reactivity of a magnitude similar to that in the laboratory also showed that treatment subjects had significantly fewer incidents of reactivity than control subjects after training. Followup interviews with treatment subjects indicated that lifestyles were also enhanced in a variety of areas, suggesting that the training had wider benefits. Issues for further research are discussed.