To further characterize the properties of retinal horizontal cell electrical synapses, we have studied the gating characteristics of gap junctions between cone-driven horizontal cells from the hybrid striped bass retina using double whole-cell voltage-clamp techniques. In a total of 105 cell pairs, the macroscopic conductance ranged from 0.4-100 nS with most cell pairs exhibiting junctional conductances between 10 and 30 nS. The junctional current-voltage relationship showed that peak or instantaneous currents (Iinst) were linear within the Vj range of +/- 100 mV and that steady-state junctional currents (Iss) exhibited rectification with increasing voltage beginning around +/- 30-40 mV Vj. The normalized junctional current-voltage relationship was well fit by a two-state Boltzmann distribution, with an effective gating charge of 1.9 charges/channel, a half-maximal voltage of approximately +/- 55 mV, and a normalized residual conductance of 0.28. The decay of junctional current followed a single exponential time course with the time constant decreasing with increasing Vj. Recovery of junctional current from voltage-dependent inactivation takes about 1 s following a pulse of 80 mV, and is about five times slower than the inactivation time course at the same Vj. Single-channel analysis showed that the unitary conductance of junctional channels is 50-70 pS. The overall open probability decreased in a voltage-dependent manner. Both the mean channel open time and the frequency of channel opening decreased, while the channel closure time increased. The ratio of closed time/total recording time significantly increased as Vj increased. Increased Vj reduced the number of events at all levels and shifted the unitary conductance to a lower level. Kinetic analysis of channel open duration showed that the distribution of channel open times was best fit by two exponentials and increased Vj significantly reduced the slower time constant. These results indicate that bass retina horizontal cells exhibit voltage-dependent inactivation of macroscopic junctional current. The inactivation occurs at the single-channel level mainly by increasing the rate of closure of voltage-sensitive channels.