Cut muscle fibers from Rana teraporaria (sarcomere length, 3.4-4.2 p.m) were mounted in a double and equilibrated with end-pool solutions that contained 20 mM EGTA and 1.76 mM Ca. Sarcoplasmic reticulum (SR) Ca release was estimated from changes in pH (Pape, P. C., D.-S.Jong, and W. I~ Chandler. 1995.Journal of GeneralPhysiology. 106:000-000). Although the amplitude and duration of the [Ca] transient, as well as its spatial spread from the release sites, are reduced by EGTA, SR Ca release elicited by either depolarizing voltage-clamp pulses or action potentials behaved in a manner consistent with Ca inactivation of Ca release. After a step depolarization to -20 or 10 mV, the rate of SR Ca release, corrected for SR Ca depletion, reached a peak value within 5-15 ms and then rapidly decreased to a quasi-steady level that was about half the peak value; the time constant of the last half of the decrease was usually 2--4 ms. Immediately after an action potential or a 10-15 ms prepulse to -20 mV, the peak rate of SR Ca release elicited by a second stimulation, as well as the fractional amount of release, were substantially decreased. The rising phase of the rate of release was also reduced, suggesting that at least 0.9 of the ability of the SR to release Ca had been inactivated by the first stimulation. There was little change in intramembranous charge movement, suggesting that the changes in SR Ca release were not caused by changes in its voltage activation. These effects of a first stimulation on the rate of SR Ca release elicited by a second stimulation recovered during repolarization to -90 mV; the time constant of recovery was ~25 ms in the action-potential experiments and ~50 ms in the voltage-clamp experiments. Fura-2, which is able to bind Ca more rapidly than EGTA and hence reduce the amplitude of the [Ca] transient and its spatial spread from release sites by a greater amount, did not prevent Ca inactivation of Ca release, even at con-