Ventricular myocardium was obtained from guinea-pig, ferret and human hearts. In each case small strips were mounted isometrically in a tissue bath and superfused with a physiological saline at 37 degrees C. The preparations were stimulated at 1 Hz and ectopic stimuli of different preceding intervals were given. The relationship between the force produced by the ectopic contraction and the duration of the preceding interval was plotted to form mechanical restitution curves (MRC). In the guinea-pig the MRC is described by a rapid recovery phase with an exponential time constant of 220 +/- 22.7 ms (mean +/- S.E.M.) followed by a decay (28.5 +/- 8.4 s). In ferret and man the rising phase is described by two exponentials (192.5 +/- 43.2 ms and 4.4 +/- 1.5 s in the ferret; 259.7 +/- 45.2 ms and 3.0 +/- 1.0 s in man). The decay phase is slower in ferret (22 +/- 156 s, P less than 0.02) and man (177 +/- 70 s, P less than 0.002) than in guinea-pig. There was no significant difference between the time constants of the rapid recovery phase of mechanical restitution in each species. The time constant of the rapid recovery phase (tau 1) was abbreviated by ryanodine, ouabain and adrenaline in human myocardium and by ryanodine alone in guinea-pig. Verapamil increased tau 1 in both species. The decay time constant (tau 3) was prolonged by ouabain, verapamil and by increasing extracellular [Ca2+] in human myocardium and by ouabain and verapamil in guinea-pig. The recovery of the second inward current in human myocardium was not correlated to the recovery of mechanical function. It is suggested that tau 1 is dependent on the recycling of Ca2+ within the cell as well as the reactivation of the second inward current. The decay phase, tau 3, is dependent on the rate of Ca2+ efflux from the cell, possibly via a Na+/Ca2+ exchange mechanism. The mechanisms underlying the slow recovery time constant, tau 2, are unclear but it is important to calculate tau 2 for the proper evaluation of tau 1.