We investigated the kinetics and sensitivity of photocurrent responses of salamander rods, both in darkness and during adaptation to steady backgrounds producing 20-3,000 photoisomerizations per second, using suction pipet recordings. The most intense backgrounds suppressed 80% of the circulating dark current and decreased the flash sensitivity approximately 30-fold. To investigate the underlying transduction mechanism, we expressed the responses as a fraction of the steady level of cGMP-activated current recorded in the background. The fractional responses to flashes of any fixed intensity began rising along a common trajectory, regardless of background intensity. We interpret these invariant initial trajectories to indicate that, at these background intensities, light adaptation does not alter the gain of any of the amplifying steps of phototransduction. For subsaturating flashes of fixed intensity, the fractional responses obtained on backgrounds of different intensity were found to "peel off" from their common initial trajectory in a background-dependent manner: the more intense the background, the earlier the time of peeling off. This behavior is consistent with a background-induced reduction in the effective lifetime of at least one of the three major integrating steps in phototransduction; i.e., an acceleration of one or more of the following: (1) the inactivation of activated rhodopsin (R*); (2) the inactivation of activated phosphodiesterase (E*, representing the complex G(alpha)-PDE of phosphodiesterase with the transducin alpha-subunit); or (3) the hydrolysis of cGMP, with rate constant beta. Our measurements show that, over the range of background intensities we used, beta increased on average to approximately 20 times its dark-adapted value; and our theoretical analysis indicates that this increase in beta is the primary mechanism underlying the measured shortening of time-to-peak of the dim-flash response and the decrease in sensitivity of the fractional response.