When mapped with a small spot of light, the central receptive fields of bipolar cells in the salamander retina are much larger than the extent of bipolar cell dendrites. Furthermore responses of bipolar cells to distant spots of light are considerably delayed relative to proximal spots. Using quantitative modelling, electrical coupling between bipolar cells is examined and rejected as a sufficient explanation of the data. An active process appears to shape signal waveform as signals spread laterally in the bipolar cell layer. Chemical synaptic coupling between bipolar cells is considered and shown to be inconsistent with the data. It is suggested that local, transient negative feedback from amacrine cells is involved in shaping bipolar cell signals.