1. The reversal potential of the surround response in cones was measured by inactivating the outer segment, polarizing the cone to different potential levels with extrinsic current, and flashing a full-field stimulus. This elicited a 3- to 4-mV depolarizing response at the normal -40-mV dark level. Its reversal potential was typically near -60 mV. 2. Cones were also polarized by a steady center light while flashing an annulus. The annular response was suppressed when the cone was polarized to near -60 mV. No reversal was measured, presumably because the maximum center light response could not exceed -60 mV. 3. Items 1 and 2 provide a method for measuring surround responses in cells post-synaptic to the cones in the absence of the surround response in the cones themselves; cone surround responses are suppressed although some horizontal cell surround responses are enhanced (item 6) at high center intensities that polarize the cones to near -60 mV. 4. Bipolar cells of both types (depolarizing and hyperpolarizing) generate a surround response that opposes the center response. In both bipolar cell types, the surround response is completely suppressed near center intensities that suppress the surround response in the cones. 5. Narrow-field horizontal cells behave like cones and bipolar cells; the depolarizing surround response is suppressed at high center intensities bright enough to suppress the cone surround response. At higher center intensities the surround response becomes hyperpolarizing, presumably due to direct coupling to peripheral horizontal cells. 6. Broad-field horizontal cells always hyperpolarize in response to surround illumination. The hyperpolarizing surround response is augmented at high center intensities, presumably because the depolarizing component, initiated in the cones, is suppressed, while the direct coupling to peripheral horizontal cells remains intact.