This study investigated transfer of improvements in stability recovery performance to novel perturbations. Thirty adults (20-53 yr) were assigned equally to three treadmill walking groups: groups exposed to eight trip perturbations of either low or high magnitude and a third control group that walked unperturbed. Following treadmill walking, participants were exposed to stability loss from a forward-inclined position (lean-and-release) and an overground trip. Lower limb joint kinematics for the swing phase of recovery steps was compared for the three tasks using statistical parametric mapping and recovery performance was analysed by margin of stability and base of support. The perturbation groups improved stability (greater margin of stability) over the eight gait perturbations. There was no group effect for stability recovery in lean-and-release. For the overground trip, both perturbation groups showed similar enhanced stability recovery (margin of stability and base of support) compared to controls. Differences in joint angle kinematics between treadmill-perturbation and lean-and-release were more prolonged and greater than between the two gait perturbation tasks. This study indicates that: (i) practising stability control enhances human resilience to novel perturbations; (ii) enhancement is not necessarily dependent on perturbation magnitude; (iii) differences in motor response patterns between tasks may limit transfer.