In order to test the importance of vision in homing pigeons, their vision was impaired by frosted contact lenses. Pigeons wearing such lenses seemed unable to recognize artificial landmarks at 6-m distance. Nevertheless, most birds homed from distances of 15 km, and some even from 130-kin distance. This result indicates that, contrary to common expectation, vision need not play an essential role in homing.Attempts to explain homing in pigeons have usually focused on vision. A key role has been attributed to the measurement of astronomical data, as well as to the recognition of landmarks (1-4). Recently, a considerable shift away from an emphasis on visual cues has taken place (5, 6), although there have been earlier attempts to explain homing by nonvisual cues (7). However, experiments entailing interference with vision have not been common. We have now succeeded in fitting contact lenses to pigeon eyes (8), and have extended our preliminary homing experiments that used colored, partially painted, and frosted lenses (8). We report here data from flights of experienced birds at distances of 15 and 130 km from their homes.
MATERIALS AND METHODSAll birds were fitted with clear contact lenses several days before each release. Habituation to wearing lenses considerably enhanced the willingness of the birds to fly upon release. On the evening before the release, the clear lenses were replaced with frosted lenses for the experimental birds, while clear lenses were left in the eyes of the control birds. Both groups were transported to the release site immediately before release, and were released in the usual manner (10). Initial orientation and homing performance were used as criteria for homing performance. The presence of lenses was checked upon arrival of each bird. Data from experimental birds that had lost one or both frosted lenses, or the frosted layer of one or both lenses, were discarded. RESULTS Fig. 1 gives initial orientation and homing performance of four releases from 15-km west, north, and south of the home loft under sunny conditions, including those published in ref 8. Fig. 2 gives initial orientation and homing performance of three releases from 130-km south, west, and north under sunny conditions. Initial orientation of one release (130-km north) could not be included in Fig. 2, because the distribution of the control birds was random. The original data of this release are, therefore, given in Fig. 3. In this instance initial orientation of experimental birds with frosted lenses was superior to that of controls.The preliminary finding (8, 9) has been confirmed: frosted lenses had no negative effect on the accuracy of initial orientation of experienced birds, though homing performance clearly was affected. Although many experimental birds did not reach home, a considerable number did so, some even at normal speeds. The number of experimental birds lost is somewhat misleading, inasmuch as several birds got close to the loft, but then missed it by some small distance. Such birds have been found ...