SUMMARY1. Intact rod outer segments (r.o.s.) isolated from bovine retinas were used to measure net Ca2+ fluxes using the optical Ca2+ indicator Arsenazo III. Ca2+ fluxes were observed, which could change the internal Ca2+ content of isolated r.o.s. by as much as 0 5 mM s-5.2. The Ca2+ content of isolated intact r.o.s. was strongly dependent on the Na/Ca ratio in the isolation medium, and could be made less than 01 mol Ca2+ mold1 rhodopsin (zero Ca2+ in isolation medium) or up to 7 mol Ca2+ mold1 rhodopsin (zero Na+ in isolation medium).3. Ca2+ efflux from r.o.s. rich in Ca2+ was observed only when Na+ was added to the external medium (as opposed to any other alkali cation); in Ca2+-depleted r.o.s. Ca2+ uptake required the presence of internal Na+ and was inhibited selectively by external Na+. These results suggest that Na-Ca exchange across the plasma membrane operated freely in both directions and controlled the internal Ca2+ concentration in r.o.s.4. Na+-stimulated Ca2+ efflux depended on the external Na+ concentration in a sigmoidal way. This suggests that the simultaneous binding of two Na ions is rate limiting for transport.5. In Ca2+-depleted r.o.s. and in the absence of external Na+, 1 mol Ca2+ mold1 rhodopsin (or 3 mM-total Ca2+) could be taken up within 1 min by intact r.o.s. at a free external Ca2+ concentration of about 1 /tM. 6. Only part of the internal Ca2+ was available for Na-Ca exchange. The external Na+ and K+ concentration as well as the temperature were factors controlling the accessibility of internal Ca2+ to participate in Na-Ca exchange.7. Ca2+ fluxes in r.o.s. with a permeabilized plasma membrane but intact disk membranes were very similar to those observed in intact r.o.s.; Na-Ca exchange could operate in both directions across the disk membrane.8. In addition to Na-Ca exchange, leaky r.o.s. also showed a guanosine 3', 5'-cyclic monophosphate (cyclic GMP)-induced Ca2+ release that was about I of the rate of Present address: