Recent work has shown that pictorial illusions have a greater effect on perceptual judgements than they do on the visual control of actions, such as object-directed grasping. This dissociation between vision for perception and vision for action is thought to reflect the operation of two separate streams of visual processing in the brain. Glover and Dixon claim, however, that perceptual illusions can influence the control of grasping but that these effects are evident only at early stages of the movement. By the time the action nears its completion any effect of illusions disappears. Glover and Dixon suggest that these results are consistent with what they call a 'planning and control' model of action, in which actions are planned using a context-dependent visual representation but are monitored and corrected online using a context-independent representation. We reanalysed data from an earlier experiment on grasping in the Ebbinghaus illusion in which we showed that maximum grip aperture was unaffected by this size-contrast illusion. When we looked at these data more closely, we found no evidence for an effect of the illusion even at the earliest stages of the movement. These findings support the suggestion that the initial planning of a simple object-directed grasping movement in this illusory context is indeed refractory to the effects of the illusion. This is not to suggest that more deliberate and/or complex movements could not be influenced by contextual information.