The Effect of a "Permanent" Clock-Shift on the Orientation of Experienced Homing Pigeons: I. Experiments in Ithaca, New York, USA

A group of experienced homing pigeons was subjected to a 6 h slow shift of their internal clock and kept under these conditions for more than 2 months. During the overlap time between the natural and artificial photoperiods they were released for training flights to familiarize them with an area while living in a "permanent shift". Tested outside the permanent shift training range, the experimentals always deviated about 30° clockwise from the mean of their controls, markedly less than in a regular 6 h slow shift. Inside the permanent shift training range, however, they oriented like the controls (Fig. 2). When their internal clock was returned to normal, the birds showed a larger counterclockwise deflection on their first flight, which was roughly comparable to the effect of a regular 6 h fast shift (Fig. 3). On later flights after normalization, this large shift was no longer found; instead we observed a roughly 30° counterclockwise deflection when they were released inside the permanent shift training range in the morning. This deflection did not seem to occur in the afternoon or outside the permanent shift training range (Figs. 4, 5), and it disappeared when the birds were repeatedly released from the same site (Fig. 6). The occurrence or non-occurrence of the deflection was independent of the duration of the shift or the time passed after normalization; it seemed to depend solely on whether the birds had become familiar with a given site in the situation of the "permanent" shift. This argues against an effect based on the sun compass. We tend to assume that the still unknown navigational "map" is involved. In this case, however, as the deflection is independent of the home direction and the type of release site bias, the factors in question would act very differently from the gradients on which the traditional concepts of the navigational "map" are based. The processes establishing and updating the "map" and their possible differences are discussed.