Previous experiments showed that pigeons allowed to smell olfactory cues at a 'false' release site, and subsequently transported to and released from another unfamiliar locality, oriented according to the home direction at the false site but eventually homed despite their wrong initial orientation. In the above 'site simulation' experiments, data collection was restricted to the initial bearings and homing performance, and no information was obtained on the actual route flown by pigeons. Route correction in site simulation tests is now investigated by releasing pigeons equipped with bird-borne direction recorders to track their route. Our results show that the experimental birds can actually fly for a long time in wrong directions, related to the home direction at the false release site, before correcting their path to the true homeward direction. This correction occurs 2 h or more after release, when the birds are supposed to have recovered from the anaesthesia of their olfactory membranes which they had been subjected to just prior to release. This result confirms the basic role of the olfactory information, collected during the outward journey, in the pigeon's homing process.