COMPELLING evidence has existed for more than a decade for a sodium/calcium (Na-Ca) exchange mechanism in the surface membrane of mammalian heart muscle cells1-3 which exchanges about three sodium ions for each calcium ion4-6. Although it is known that cardiac muscle contraction is regulated by a transient increase in intracellular calcium ([Ca2+]i) triggered by the action potential7, the contribution of the Na-Ca exchanger to the [Ca2+]i transient and to calcium extrusion during rest is unclear. To clarify these questions, changes in [Ca2+]i were measured with indo-1 in single cardiac myocytes which were voltage clamped and dialysed with a physiological level of sodium. We find that Ca entry through the Na-Ca exchanger is too slow to affect markedly the rate of rise of the normal [Ca2+]i transient. On repolarization, Ca extrusion by the exchanger causes [Ca2+]i to decline with a time constant of 0.5 s at -80 mV. The rate of decline can be slowed e-fold with a 77-mV depolarization. Calcium extrusion by the exchanger can account for about 15% of the rate of decline of the [Ca2+]i transient (the remainder being calcium resequestration by the sarcoplasmic reticulum (SR)). The ability of the cell toextrude calcium was greatly reduced on inhibiting the exchanger by removing external sodium, which itself led to an increase in resting [Ca2+]i. This finding is in contrast to the suggestion that calcium extrusion at rest is mediated mainly by a sarcolemmal Ca-ATPase8. Near equilibrium, the properties of the Na-Ca exchanger were consistent with a stoichiometry of 3Na:1 Ca, and since the exchanger can operate in the absence of potassium, the cardiac Na-Ca exchanger must be different from that of rod outer segments9.