We have provoked sustained mechanical alternans (MA), by a combination of low temperature (27 ± 0. PC) and rapid pacing (3-4.5 Hz) in isometric guinea-pig papillary muscles (0.32-0.89 mm maximum diameter, n = 8) superfused with Tris (20 mM) buffered Tyrode (pH 7.3-7.4) aerated with 100% 02.Mechanical restitution curves were plotted using data obtained by introducing test stimuli at various intervals following trains of steady-state stimuli. Figure 1 shows the curves for test stimuli applied after the small and large contractions during sustained MA for one preparation. The mean half-times for restitution following the small and large contractions in MA are 140 ± 40 ms (mean ± S.E.M., n = 7) and 150 ± 41 ms (mean ± S.E.M., n = 8) respectively. The difference between the means is not statistically significant. These results do not support the hypothesis (Mahler & Rogel, 1970) that MA results from dependence of the restitution time course upon the myocardial force of a given contraction. REFERENCE Mahler, H. & Rogel, S. (1970). Clin. Sci. 39, 625-639. J. Physiol. (1991) Force in relation to membrane potential was studied in voltage-clamped ferret papillary muscles at 37TC using the single sucrose gap technique. This method allows accurate force measurements during voltage clamp. The preparation was stimulated with rectangular voltage clamp pulses at a rate of 1.0 Hz. The holding membrane potantial was -40mV and the preparation was depolarized to + 20 mV for 200 ms to initiate force production (clamps 'O'in Fig. 1). A test clamp denoted '1' was introduced at selected times. It consisted of an initial depolarizing step of 20 ms duration to +20 mV in order to initiate the first part of the second inward current (Ica,f, Arlock & Noble, 1985). This was followed by a second voltage step to different potentials between -20 and +20 mV at a varied duration (0.2-3.0 s). Test clamp 1 was followed by a second and a third test clamp pulse (2, 3 in Fig. 1) with the same amplitude and duration as in the control period. The variations in test clamp 1 produced different degrees of contractile potentiation of test contractions 2 and 3. There was a straight relation between peak force of contraction 3 (F3) and peak force of contraction 2 (F2) (Fig. 1). When the holding potential between clamps 2 and 3 was -70 mV, the curve shifted downwards in a more or less parallel manner compared to the curve obtained at -4OmV (Fig. 1). The straight relation between F3 and F2 could be explained by a recirculation of calcium within the cardiac cell. The shift of the curve would follow from an increased outflow of calcium via sodium-calcium exchange caused by a greater driving force. The unchanged slope suggests that recirculation mainly takes place during depolarisation. 3PT ,B-Adrenergic agonists can restore action potentials of ventricular myocytes rendered inexcitable by hyperkalaemia (Paterson et al. 1990). Hence we have argued that catecholamines may provide protection from the high (7-8 mM) plasma K+ levels found in exercise-induced h...