1. Using a fast flow, computer-controlled, two-Vaseline-gap chamber, single muscle fibres were subjected to 'pulses' of caffeine at Ca2+ releasing concentrations, combined with voltageclamp depolarizations, while monitoring intracellular [Ca2+].2. Ca2+ release flux elicited by caffeine reached 2-5 mm s-, or less, after 3 s of exposure, then decayed to zero. The caffeine-releasable pool of sarcoplasmic reticulum (SR) Ca2+ was 2-9 + 0A4 mM (mean+ S.E.M., n = 10).3. In parallel with release induced by caffeine, release induced by voltage pulses applied during a caffeine exposure increased in the first second of exposure, then decreased, to abolition after 5 s. and, upon repolarization, termination of the caffeine-induced release. This is similar to repolarization-induced stop of caffeine contracture (RISC) in embryonic murine myoballs. The permeability elicited by caffeine (ratio of flux to calcium in the releasable pool) was not affected by depolarizing pulses. Therefore, the mechanism of the RISC-like effect was Ca2P depletion. 6. Caffeine-induced release did not depend on the holding potential. 7. Whether caffeine was present or not, release activated by voltage remained always under voltage control, ending rapidly upon repolarization. A depolarizing pulse induced a release permeability with an early peak, followed by decay to a steady level. Caffeine (10 mM) shifted the mid-activation voltage of both peak and steady components by -15 mV and increased the steepness of their voltage dependence by 15%. The maximum permeability increased by 30 % for the peak and 25% for the steady component (n = 5). These results neither support nor disprove the hypothesis that the peak of Ca2P release is activated by Ca2+.8. The similar potentiation by caffeine of both components of release, the continued ability of voltage to control release in the presence of caffeine, and its failure to alter caffeine-induced permeability indicate that caffeine and the voltage sensor enhance independently the channel's tendency to open.