Puzzled by recent reports of differences in specific ligand binding to muscle Ca2+ channels, we quantitatively compared the flux of Ca2+ release from the sarcoplasmic reticulum (SR) in skeletal muscle fibers of an amphibian (frog) and a mammal (rat), voltage clamped in a double Vaseline gap chamber. The determinations of release flux were carried out by the "removal" method and by measuring the rate of Ca2+ binding to dyes in large excess over other Ca2+ buffers. To have a more meaningful comparison, the effects of stretching the fibers, of rapid changes in temperature, and of changes in the Ca2+ content of the SR were studied in both species. In both frogs and rats, the release flux had an early peak followed by fast relaxation to a lower sustained release. The peak and steady values of release flux, Rp and Rs, were influenced little by stretching. Rp in frogs was 31 mM/s (SEM = 4, n = 24) and in rats 7 +/- 2 mM/s (n = 12). Rs was 9 +/- 1 and 3 +/- 0.7 mM/s in frogs and rats, respectively. Transverse (T) tubule area, estimated from capacitance measurements and normalized to fiber volume, was greater in rats (0.61 +/- 0.04 microns-1) than in frogs (0.48 +/- 0.04 micron-1), as expected from the greater density of T tubuli. Total Ca in the SR was estimated as 3.4 +/- 0.6 and 1.9 +/- 0.3 mmol/liter myoplasmic water in frogs and rats. With the above figures, the steady release flux per unit area of T tubule was found to be fourfold greater in the frog, and the steady permeability of the junctional SR was about threefold greater. The ratio Rp/Rs was approximately 2 in rats at all voltages, whereas it was greater and steeply voltage dependent in frogs, going through a maximum of 6 at -40 mV, then decaying to approximately 3.5 at high voltage. Both Rp and Rs depended strongly on the temperature, but their ratio, and its voltage dependence, did not. Assuming that the peak of Ca2+ release is contributed by release channels not in contact with voltage sensors, or not under their direct control, the greater ratio in frogs may correspond to the relative excess of Ca2+ release channels over voltage sensors apparent in binding measurements. From the marked differences in voltage dependence of the ratio, as well as consideration of Ca(2+)-induced release models, we derive indications of fundamental differences in control mechanisms between mammalian and amphibian muscle.