MORE than 30 years after Hubel and Wiesel¹ first described orientation selectivity in the mammalian visual cortex, the mechanism that gives rise to this property is still controversial. Hubel and Wiesel¹ proposed a simple model for the origin of orientation tuning, in which the circularly symmetrical receptive fields of neurons in the lateral geniculate nucleus that excite a cortical simple cell are arranged in rows. Since this model was proposed, several experiments^2–6 and neuronal simulations^7,8 have suggested that the connectivity between the lateral geniculate nucleus and the cortex is not well organized in an orientation-specific fashion, and that orientation tuning arises instead from extensive interactions within the cortex. To test these models we have recorded visually evoked synaptic potentials in simple cells while cooling the cortex⁹, which largely inactivates the cortical network, but leaves geniculate synaptic input functional. We report that the orientation tuning of these potentials is almost unaffected by cooling the cortex, in agreement with Hubel and Wiesel's original proposal¹.