In 1992, Goodale and Milner proposed a division of labour in the visual pathways of the primate cerebral cortex between a dorsal stream specialised for the visual control of action and a ventral stream dedicated to the perception of the visual world. In the years since this original proposal, support for the perception-action hypothesis has come from neuroimaging experiments, human neuropsychology, monkey neurophysiology, and human psychophysical experiments. Indeed, some of the strongest support for this hypothesis has come from behavioural experiments showing that visually guided actions are largely refractory to perceptual illusions. Although controversial, the findings from this literature both support the original hypothesis and suggest important modifications. The ongoing challenge for neurobiologists is to map these behavioural findings onto their corresponding neural substrates.