1. A method has been developed for calculating the flux of Ca2+ across the sarcoplasmic reticulum (SR) during excitation‐contraction coupling in mammalian heart cells. FSR will symbolize the net rate of movement of Ca2+, per litre of accessible cytoplasm, into or out of the sarcoplasmic reticulum. FSR has the units MS‐1. 2. A theory of the cytoplasmic [Ca2+]i transient in mammalian heart cells is presented in which the [Ca2+]i transient results from the various cellular processes that tend to increase or decrease cytoplasmic [Ca2+]i. According to the theory, FSR can be calculated if all cellular processes that contribute to the [Ca2+]i transient (other than Ca2+ fluxes across the SR) are either eliminated or are known quantitatively. 3. To obtain the measurements required to apply this theory, [Ca2+]i transients and membrane currents were recorded in guinea‐pig single ventricular myocytes subjected to whole‐cell voltage clamp and internal perfusion. [Ca2+]i transients were recorded through the use of the Ca2+ indicator, Fura‐2 (pentapotassium salt). 4. Ca2+ fluxes through the sodium‐calcium exchanger were eliminated in all experiments, by perfusing the cells, internally and externally with Na(+)‐free solutions. Ca2+ flux through the sarcolemmal L‐type Ca2+ channel was measured as the verapamil‐sensitive current. Influx of Ca2+ through all other voltage‐dependent pathways was found to be negligible for the calculation of FSR over the time course of a single [Ca2+]i transient. 5. In the combined absence of Ca2+ current, Na(+)‐Ca2+ exchange and fluxes across the SR (10 mM‐caffeine), the net rate of removal of Ca2+ from the cytoplasm, which includes presumed contributions from sarcolemmal Ca(2+)‐ATPase and mitochondrial Ca2+ transport, was found to be a negligible quantity in the calculation of FSR, over the time course of a single [Ca2+]i transient. 6. Calculation of FSR requires that the Ca(2+)‐binding capacity of cytoplasm be known. [Ca2+]i transients recorded during measurable total Ca2+ influx into the cytoplasm (verapamil‐sensitive current in the absence of fluxes across the SR) were compared with theoretical Ca2+ transients computed on the assumption that the entering Ca2+ could bind only to intracellular ligands (values for ligands taken from literature) and to Fura‐2 (30 microM). The slope of the regression line relating calculated total change in [Ca2+]i to the measured total Ca2+ influx was 0.99, not different from the perfect theoretical slope of 1.0 (correlation coefficient, 0.81; standard deviation of slope, 0.14; n = 7).4+ the SR and FSR had a similar time course to that on depolarization. 10. The unidirectional efflux of Ca2+ from the SR, symbolized FSR, rel was calculated utilizing assumed characteristics of the Ca2+ pump of the SR. The value of FSR, rel was not affected by repolarization from voltage‐clamp pulses greater than 150 ms in duration.(ABSTRACT TRUNCATED AT 400 WORDS)