The aim of our study was to analyse the components of visual lateralization in pigeon homing, a large-scale spatial task. In a series of 13 releases, birds were tested as binocular controls or monocularly with the right or left-eye covered. Occlusion of either eye had a significant effect on initial orientation and homing performance. Vanishing bearings were deflected to the side of the open eye, vanishing intervals were longer, and homing speed was reduced. These parameters were affected to a different degree. Initial orientation was markedly lateralized, with birds using their right-eye deviating less from the mean of control birds and showing significantly less variance. One minute after release, the deviation and variance were similarly large in both monocular groups. However, while the right-eyed birds improved their performance until leaving the release site, the left-eyed birds failed to do so. Vanishing intervals were similar in both monocular groups, but homing speed was reduced to a lesser extent in pigeons using the right-eye. The degree of lateralization varied across different releases, but superiority of the right-eye/left hemisphere prevailed. Lateralization did not depend on familiarity with the release site. This suggests that the crucial processes involved the eyes, but did not depend on visual memory of landscape features at the release site. Results reveal, for the first time, asymmetries of directional orientation as an essential component of lateralized homing performance. As likely mechanisms we suggest hemispheric differences in magnetic compass orientation and in the adjustment to optic flow.