In cell-attached recordings from rods in the intact lizard retina, light decreased a standing inward membrane current with a reversal potential approximately 60 mV more positive than the resting potential. The peak amplitude of saturating responses depended upon the area of recorded membrane and varied from cell to cell over approximately 100-fold range. Small patches of membrane gave variable responses to identical moderately intense flashes. Whole-cell voltage-clamp recordings were obtained on isolated frog rods with intact ellipsoids. Peak whole-cell photocurrent was related to flash intensity by a Michaelis equation with saturating response amplitudes ranging up to 30 pA in 0.1 mM-Ca2+ Ringer solution. In darkness the steady-state current-voltage relation, determined with whole-cell voltage clamp, showed outward rectification. Photocurrent had nearly constant amplitude between -80 and -10 mV, a mean reversal potential of +8 mV and recovered from flashes more slowly at positive holding potentials. Although it was not possible to resolve light-sensitive single-channel current events, power spectral analysis revealed both low- and high-frequency components of the light-sensitive noise in both cell-attached and whole-cell recordings. The low-frequency component was described by the product of two Lorentzians using time constants derived from the kinetics of the dim flash response. The high-frequency component of the light-sensitive noise was described by a single Lorentzian with a half-power frequency of 62 Hz in lizard and 212 Hz in frog. The half-power frequency was not appreciably affected by steady illumination. The Lorentzian nature of the noise suggests that the light-sensitive channel is a pore rather than a shuttle-type carrier. In cell-attached recordings the high-frequency component declined monotonically with increasing light intensity, suggesting that less than one-half of the channels are open in darkness. Furthermore, the ratio of the variance of the high-frequency noise to the mean photocurrent was independent of light intensity. Changing external Ca2+ from 0.1 to 0.5 mM reduced the ratio from 19.7 to 9.0 fA without a significant effect on the cut-off frequency of the noise. The results support the conclusion that the light-sensitive pore is opened by an internal transmitter that acts as an agonist and that both open and closed states of the pore may be blocked by external Ca2+. The conductance of the light-sensitive pore in the absence of external Ca2+ is estimated to be 1.25-2 pS.(ABSTRACT TRUNCATED AT 400 WORDS)