In the lizard, Ctenophorus ornatus, anatomical studies have revealed that optic axons regenerate to visual centers within 2 months of nerve crush but that, from the outset, the regenerated projections lack topographic order (Beazley et al., 1997; Dunlop et al., 1997b). Here we assess the functional topography of the regenerated retinotectal projections by electrophysiological recording of extracellular multiunit responses to visual stimulation and by observing the lizards' ability to capture live prey. At the completion of the electrophysiology, DiI was applied locally to the retina and the topography of the tectal projection later assessed. Electrophysiology revealed that, at 2–4.2 months, responses were weak and habituated readily; no retinotopic order was detected. Between 4.5–6 months, responses were more reliable and the majority of lizards displayed a crude retinotopic order, especially in the ventro-temporal to dorso-nasal retinal axis. Although responses were variable between 6–9 months, they tended to be more reliable again thereafter. However, from 6–18 months, the projection consistently lacked topography with many retinal regions projecting to each tectal locus. Lizards, including those with electrophysiological evidence of crude retinotopy, were consistently unable to capture live prey using the experimental eye. Labelling with DiI confirmed the absence of anatomical retinotopy throughout. Taken together, the electrophysiological and anatomical data indicate that retinotopically appropriate axon terminals (or parts thereof) are transiently active whilst inappropriately located ones are silent. Presumably in lizard map-making cues fade with time and/or the mechanisms are lacking to stabilize and refine the ephemeral map. Moreover, the transient retinotopy is insufficient for useful visual function.