Neurons in the primary visual cortex of higher mammals are arranged in columns, and the neurons in each column respond best to light-dark borders of particular orientations. The basis of cortical cell orientation selectivity is not known. One possible mechanism would be for cortical cells to receive input from several lateral geniculate nucleus (LGN) neurons with receptive fields that are aligned in the visual field (Hubel and Wiesel, 1962). We have investigated the relationship between the arrangement of the receptive fields of geniculocortical afferents and the orientation preferences of cortical cells in the orientation columns to which the afferents provide visual input. Radial microelectrode penetrations were made into primary visual cortex of anesthetized adult sable ferrets. Cortical cells were recorded throughout the depth of the cortex, and their orientation preferences were determined. Cortical cell responses were then eliminated by superfusion of the cortex with either kainic acid (Zahs and Stryker, 1988) or muscimol. After the drug treatment, responses from many single units with distinct receptive fields were recorded. These responses were presumed to be those of geniculocortical afferents, because they had the response properties characteristic of LGN neurons, and because they could be recorded only in cortical layers that receive geniculate input. In 16 of 18 cases, the afferent receptive fields recorded in a single penetration covered an elongated region of visual space. In these penetrations, the best-fit line through the centers of the afferent receptive fields generally paralleled the preferred orientation of cortical cells recorded at the same site in cortex. These results are consistent with the Hubel and Wiesel (1962) model for the construction of oriented visual cortical receptive fields from geniculate inputs with aligned receptive fields.