Electrical reponses of luminosity horizontal cells (L cells) to monochromatic stimuli were analyzed by intracellular recordings in the retinas of the freshwater turtle (Pseudemys scripta elegans) and of the sea turtle (Chelonia mydas mydas). Light intensity, duration, and wavelength were varied to assess temporal effects. For a given intensity of monochromatic light, response amplitude increased with stimulus duration until maximum amplitude occurred at a specific duration. This suprathreshold metric of temporal integration is called here summation time, and it is wavelength-dependent.

L cells always had some level of red-sensitive cone input, although in some cells inputs from green- and blue-sensitive cones were also observed. For these latter cells, summation times were shorter for 640-nm than for 540-nm or 450-nm lights. These results were most evident in cells that received dominant inputs from blue- or green-sensitive cones.

Responses of some other L cells were almost completely dominated by inputs from red-sensitive cones. Summation times of these cells were not wavelength-dependent. But when these inputs also included green-sensitive cones, shorter summation times were obtained to 640-nm light than to 540-nm light, even though dominant inputs were still from red-sensitive cones. These results, obtained from both retinal and 3,4-dehydroretinal photopigment systems, are consistent with reported observations in Pseudemys scripta elegans that show linear responses of red-sensitive cones to have shorter integration times and times-to-peak than green-sensitive cones.

Responses from horizontal cells dominated by blue-sensitive cone inputs were the most sensitive of all; they also had the longest summation times. These results support the hypothesis that a gain in sensitivity occurs from the integration of absorbed photons over longer periods of time.

These intracellular responses are of particular importance because behavioral critical durations in turtle, as defined by Bloch&'s law, are similarly wavelength-dependent.

We characterized the chromatic and temporal properties of a sample of 177 red–green parvocellular neurons in the LGN of Macaca nemestrina, using large-field stimuli modulated along different directions through a white point in color space. We examined differences among the properties of the four subclasses of red–green P-cells (on- and off-center, red and green center). The responses of off-center cells lag the stimulus more than do those of on-center cells. At low temporal frequencies, this causes the phase difference between responses of the two kinds of cells to be considerably less than 180 deg. For isoluminant modulations the phases of on- and off-responses were more nearly 180 deg apart. A cell's temporal characteristics did not depend on the class of cone driving its center. Red center and green center cells have characteristically different chromatic properties, expressed either as preferred elevations in color space, or as weights with which cells combine inputs from L- and M-cones. Red center cells are relatively more responsive to achromatic modulation, and attach relatively more weight to input from the cones driving the center. Off-center cells also attach relatively more weight than do on-center cells to input from the class of cone driving the center.

We studied the interaction between the chromatic and temporal properties of parvocellular (P) neurons in the lateral geniculate nucleus (LGN) of macaque monkeys. We measured the amplitudes and phases of responses to stimulation by spatially uniform fields modulated sinusoidally about a white point in a three-dimensional color space, at a range of temporal frequencies between 1 and 25 Hz. Below about 4 Hz, temporal frequency had relatively little effect on chromatic tuning. At higher frequencies chromatic opponency was weakened in almost all cells. The complex interactions between temporal and chromatic properties are represented by a linear filter model that describes response amplitude and phase as a function of temporal frequency and direction in color space along which stimuli are modulated. The model stipulates the cone inputs to center and surround, their temporal properties, and the linear combination of center and surround signals. It predicts the amplitudes and phases of responses of P-cells, and the change of chromatic properties with temporal frequency. We used the model to investigate whether or not the chromatic signature of the surround in a red–green cell could be estimated from the change in the cell's chromatic properties with temporal frequency. Our findings could be equally well described by mixed cone surrounds as by pure cone surrounds, and we conclude that, with regard to temporal properties, there is no benefit to be gained by segregating cone classes in center and surround.