Using a combination of anatomical and physiological techniques we have studied some of the neural connections subserving binocular vision in two species of artiodactyl ungulates (the sheep, Ovis sp., and the goat, Capra hircus). After monocular injections of tritiated proline, transsynaptic transport was observed bilaterally in layers 4 and 6 of visual cortical areas V1 and V2, but there were no sharply defined ocular dominance columns of the kind seen in cats and rhesus monkeys. In coronal sections there was a discontinuity in density of labelling between areas V1 and V2 corresponding to a point in the visuotopic map about azimuth - 15 degrees in the ipsilateral visual field. This discontinuity was most pronounced in the hemisphere ipsilateral to the injected eye. We conclude, therefore, that while the cortical representation of ipsilateral visual space can be explained by the retino-geniculo-cortical input pathway from the contralateral eye, the physiologically demonstrated cortical contribution to ipsilateral visual space from the ipsilateral eye cannot be explained in this way. This conclusion was reinforced by experiments using retrograde transport of horseradish peroxidase from the lateral geniculate nucleus (LGN) and medial interlaminar nucleus (MIN) to retinal ganglion cells in flattened whole mounts. These experiments revealed a sharp nasotemporal decussation in the ipsilateral retina, which could not thereby subserve any significant representation of the ipsilateral visual field. In contrast the contralateral nasotemporal decussation was smeared, with many labelled ganglion cells in the temporal retina which could subserve visual input from the ipsilateral hemifield. When we estimated the projection of the nasotemporal decussation line into visual space, we found that it was tilted from vertical by about 5 degrees in each eye, in a similar way to that already reported in the cat. Neurophysiological recordings from binocular neurons in area V1 with different vertical eccentricities also showed that the vertical horopter (the midsagittal reference plane for binocular vision) would be tilted in life when the cyclotorsional position of the eyes was taken into account. Thus both anatomical and physiological methods concur in the prediction that ungulates have a tilted vertical horopter like that described for two other terrestrial species, the burrowing owl and the cat. Anatomical experiments reveal other similarities between the organisation of the ungulate's visual pathways and that of the cat. For example, after tritiated proline injections in V1, we found visuotopic labelling in the calustrum, dorsal LGN, cortical area V2, and the superior col