The problem of how the brain creates a spatially accurate representation of the world for movement and perception despite the unreliability of the retinal signal has been of interest to neuroscientists since the 19th century. Helmholtz saw a patient with one blind eye who developed a paralysis of the lateral rectus muscle in the good eye.  The patient said that when he tried to look towards his ear the visual world jumped in the direction opposite the attempted eye movement even though the eye did not actually move. Helmholtz postulated that the brain compensated for eye movements by feeding the motor command back to the visual system.  Sherrington, who was interested in proprioception, suggested that because the human extraocular muscles had many muscle spindles, the brain compensated for eye moves by measuring the position of the eye in the orbit and then calculating the position of objects inn the world in suprarenal coordinates.  I will present evidence for both of these strategies: a corollary discharge remaps the visual receptive fields of parietal neurons before a saccade, and the representation of eye position in monkey somatosensory cortex Area 3a is necessary for establishment of spatially accurate long term visual memory.