Responses of R/G-type horizontal cells recorded intracellularly in eye cup preparations with normalcirculation of crucian carp under dark-adapted conditions were studied and photoreceptor inputs convergingonto them were identified. Spectral sensitivities of the hyperpolarizing component of responses, determinedat different phases, suggested that it was produced by input from green-sensitive cones. This input andinput from long wavelength sensitive cones, causing depolarization of cells, were almost linearly sum-mated. Based on the linear summation assumption, an algebraic procedure was developed to isolate thedepclarizing component from responses of R/G-type cells to flashes at wavelengths longer than 540 nm.Spectral sensitivities of the isolated depolarizing component were determined and agreed reasonably wellwith those of red-sensitive cones. Therefore, it is unlikely that depolarization of R/G-type cells maybe due to input from some presumptive deep-red sensitive cones. Responses of R / G-type horizontal cells recorded intracellularly in eye cup preparations with normalcirculation of crucian carp under dark-adaptedconditions studied and photoreceptor inputs convergingonto them were identified. Spectral sensitivities of the hyperpolarizing component of responses, determinedat different phases, suggested that it was produced by input from green-sensitive cones. This input and input from long wavelength sensitive cones, causing depolarization of cells, were almost linearly sum-mated. Based on the linear summation assumption, an algebraic procedure was developed to isolate the breakdown components from responses of R / G-type cells to flashes at wavelengths longer than 540 nm. Spectral sensitivities of the isolated depolarizing component were determined and agreed reasonably well with those of red-sensitive cones. Thus, it is unlikely that depolarization of R / G-type cells maybe due to input from some presumptive deep-red sensitive cones.