Membrane current was recorded from a single primate rod with a suction pipette while the cell was bath perfused with solutions maintained at a temperature of ~ 38°C. A transient inward current was observed at the onset of bright illumination after briefly exposing the outer segment in darkness to Ringer's (Locke) solution containing 3-isobutyl-l-methylxanthine (IBMX), an inhibitor of cGMP phosphodiesterase. After briefly removing external Na ÷ from around the outer segment in darkness, a similar current was observed upon Na + restoration in bright light. By analogy to amphibian rods, this inward current was interpreted to represent the activity of an electrogenic Na÷-dependent Ca 2+ efl]ux, which under physiological conditions in the light is expected to reduce the free Ca 2+ in the outer segment and provide negative feedback (the "Ca 2+ feedback") to the phototransduction process. The exchange current had a saturated amplitude of up to ~ 5 pA and a decline time course that appeared to have more than one exponential component. In the absence of the Ca 2+ feedback, made possible by removing the Ca 2÷ influx and efflux at the outer segment using a 0 Na+-0 Ca 2+ external solution, the response of a rod to a dim flash was two to three times larger and had a longer time to peak than in physiological solution. These changes can be approximately accounted for by a simple model describing the Ca 2÷ feedback in primate rods. The dark hydrolytic rate for cGMP was estimated to be 1.2 s -~. The incremental hydrolytic rate, [3*(t), activated by one photoisomerization was ~0.09 s -~ at its peak, with a time-integrated activity, ff3*(t)dt, of ~ 0.033, both numbers being derived assuming spatial homogeneity in the outer segment. Finally, we have found that primate rods adapt to light in much the same way as amphibian and other mammalian rods, such as showing a Weber-Fechner relation between flash sensitivity and background light. The Ca 2÷ feedback model we have constructed can also explain this feature reasonably well.