1. Limulus ventral photoreceptor cells were voltage-clamped with two intracellular micro-electrodes. The light-induced membrane current was recorded for brief stimuli. From observation of discrete waves (quantum bumps) at low stimulus energies and the early receptor potential at high energies, the stimulus energy was related to the number of rhodopsin molecules photosiomerized. 2. In the dark-adapted cell the log (peak light-induced current) reached almost its maximum value when about 10(3) of the 10(9) rhodopsin molecules in the cell were photoisomerized. 3. The magnitude of the maximum light-induced current was not significantly altered after iontophoresis of EGTA into the cell. This treatment is known to counteract the Ca2+-mediated reduction in sensitivity to light. 4. Current pulses were injected into the unclamped cell during the receptor potential. The form of the voltage deflexion (a step followed by a curve) suggested that the effective electrical equivalent of the cell was a membrane capacitance in parallel with a light-dependent membrane resistance, Rm, and in series with another, light-invariant, resistance, Rs. Rs ranged from 7 to 24 k omega (five cells). 5. During a receptor potential the ratio Rm/Rs was never observed to fall below 1.7 no matter how intense the light flash. Hence, it is concluded that the light-induced current saturated essentially because Rm fell to a minimum value. 6. Charging curves gave a value for the capacitance, and hence the area, of the surface membrane. From this it was estimated that there were 10(5)-10(6) microvilli on each cell. 7. These results show that the light-induced increase in membrane conductance in a dark-adapted cell comes close to its maximum value when the number of photoisomerizations is about 1/1000 the total number of microvilli. We suggest that absorption of a photon by a rhodopsin molecule in a microvillus causes an increase in membrane conductance on parts of the surface membrane beyond that microvillus. 8. In the presence of moderate background illumination the sensitivity to non-saturating superimposed flashes was greatly decreased (e.g. by 10(3) while the saturating light-induced current was only slightly decreased (e.g. by 15%). At higher background intensities the saturating light-induced current was further decreased (e.g., with a background that photoisomerized 10(6.25) molecules per second the saturating light-induced current was reduced by 47%).