THE ROLE OF VISUAL FAMILIARITY WITH THE LANDSCAPE IN PIGEON HOMING

In an introductory discussion it is argued why a topographical 'patter n map' based on a familiar visual landscape should come into conflict with an angular shift of the sun compass, whereas a 'point map' based on atmospheric odours is expected to be freely rotatable by compass sh ifts. To test whether these predictions are met, releases were conduct ed in a familiar area with pigeons whose circadian clock was shifted 6 h forward. Half of the pigeons were prevented from smelling natural a irborne odours by air filtration before release and nasal anaesthesia upon release. Analysis of initial bearings, also including earlier rel eases in unfamiliar areas, led to the following results. (1) Unfamilia r area, olfaction unimpaired (only assumed point map applicable): full deflection of bearings according to the amount of sun-compass shift. (2) Familiar area, olfaction unimpaired (both point map and pattern ma p applicable): average deflection of bearings somewhat smaller than co mpass shift. (3) Familiar area, no olfactory access to natural air (on ly assumed pattern map applicable): average deflection of bearings onl y about half compass shift; increased variability. Corresponding non-s hifted controls were always homeward orientated. Results support the h ypothesis that pigeons in a familiar area, ii relying exclusively on n onolfactory spatial information, utilize the visible landscape to find the way home. Certainly they also do so under normal conditions when additionally available airborne odours co-determine behaviour, so that the investigator is unable to separate the roles of vision and olfact ion. This most plausible and parsimonious hypothesis should be falsifi ed before consideration of alternative hypotheses suggesting there are some mysterious unknown factors at work. How pigeons may view and eva luate the landscape for orientation over longer distances is briefly d iscussed.