SUMMARY1. In single striated muscle fibres from the barnacle Balanus nubilu8 tension development following axial injection of caffeine (50 mM in 150 mM-KCl, pH 7.1) was used as an index of releasable Ca. It was shown that fibres incubated in 0 Ca (Na replaced) salines for up to 400 min gave ca. 15 % of the control tension response. Inclusion of 1 mM-La in the 0 Ca (Na) saline significantly reversed this decline.2. Estimates of the total fibre calcium assayed under the same experimental conditions indicated a 67 % loss of fibre Ca in 0 Ca, and only a 37 % loss in La-0 Ca media. No correction was made for the loss of calcium from the extracellular space.3. Experiments with 45Ca indicated that the efflux of calcium from this preparation was inhibited by 1 mM-La externally and that this effect was still significant even in the presence of 0 Ca (Na) salines. The caffeinestimulated efflux of Ca was also reduced by ca. 70 % in the presence of 1 mm La saline externally.4. The influx and efflux of 14C caffeine were shown to be rapid, and apparently passive. The diffusion coefficient for caffeine following intracellular injection was 2-4 + 0-2 x 10-6 cm2 sec-' at 18-22°C.5. There was no significant loss of 140La over a period of 2 hr following axial micro-injection into the fibres.6. In 140La uptake experiments there was a progressive increase in the La space over 10-5 hr, in contrast to the results with [3H]inulin, whose uptake saturated in ca. 1-5 hr. The probability of surface binding and the precipitation of La salts in the extensive extracellular space was suggested as an explanation.7. It is concluded that internal Ca within the sarcoplasmic reticulum, and not cleft or extracellular Ca is the most significant source for these caffeine-induced contractions. Fluxes across the surface membrane can however alter the internal Ca stores over longer periods of time.