10 experiments were conducted in which a group of experimental pigeons were allowed to smell natural ambient air for a period of 3 h at a site which was 25–55 km away from the loft (‘false site’) and then subsequently released at another site (‘ true release site ’) situated in a roughly diametrical position from home. Controls were directly transported to the true release site, and during the same 3 h period they were allowed to smell natural air there. Except for those 3 h, both controls and experimentals were prevented from smelling atmospheric odours by filtration of the air flowing through the containers in which the birds were held. Air filtration was never interrupted in a third group of birds, which were transported together with the experimentals. All the birds were released under local anaesthesia of their olfactory mucosae. Experiments were performed in pairs: Within each pair of experiments, the site used as ‘false site’ in the first experiment was the ‘true release site ’ in the second and vice versa.
On average, initial bearings of the controls were homeward oriented. Bearings of experimentals, which smelled natural air at the ‘false site’, were oriented, with similar accuracy, towards a compass direction pointing from this ‘false site’ to home, and hence these birds actually flew predominantly away from home (Fig. 4). The third group of pigeons, which never could smell natural air, differed from both of the others in their initial orientation.
Faster homing speeds were significantly more frequent in the controls than in the other two treatment types. Experimentals that had smelled ‘falsesite ’ odours were slowest.
Thus, pigeons not only can be prevented from orienting their courses homeward by elimination of certain substances from the breathing air, but they also can be systematically misguided by false olfactory information.
There was no indication that site-specific stimuli independent of olfaction might additionally be necessary for site localization in pigeons.