When homing pigeons are subjected to a shift of their internal clock, their vanishing bearings are deflected with respect to untreated controls,indicating the use of the sun compass. However, the induced deflection is often smaller than the difference in sun azimuth between the birds' subjective time and the real time of day (e.g. Schmidt-Koenig, 1961). This phenomenon is most pronounced in old, experienced pigeons, which, on average,show only about 50-60% of the expected deflection(Wiltschko et al., 1994; Wiltschko and Wiltschko, 2001). Wallraff et al. (1999), as well as Kamil and Cheng (2001),hypothesized that the reduced deflection was caused by these pigeons' greater experience, in particular their familiarity with local landmarks. Gagliardo et al. (2005) recently published a paper claiming that familiarity with the release site is indeed one of the factors reducing the deflection induced by clock-shifting.
This claim, however, is not justified. An analysis of the response to clock-shifting must be based on the behavior of normal, otherwise untreated pigeons, i.e. the data of the control group C in Series I and the data of Series II (tables 3 and 4 in Gagliardo et al., 2005). Unfortunately, the pigeons familiar and unfamiliar with the release sites were from different groups, with possible differences in pre-experience, and were released in different years. More importantly,however, the pigeons familiar with the release sites were released as clock-shifted birds at all three sites used, whereas the pigeons unfamiliar with the sites were released only once under clock-shifted conditions. This is a serious flaw in test design, making any comparison problematic. The familiar pigeons undertook their second and third flights as clock-shifted birds after they had just experienced that their sun compass provided them with false information and that they had to ignore this false information to return home. This affects the sun compass: repeated releases alter the sun compass readings and eventually lead to a recalibration, as has been convincingly demonstrated in two experimental series transferring pigeons to a `permanent' clock-shift(Wiltschko et al., 1976, 1984). The test design of Gagliardo et al. (2005), by repeatedly releasing the familiar birds under clock-shifted conditions, thus tends to reduce the deflection of this group.
The authors claim that this is not the case, referring to Foà and Albonetti (1980), who released clock-shifted birds several times and observed larger shifts at two unfamiliar sites. However, deflections induced by clock-shifting always show considerable variation. Foà and Albonetti(1980) failed to do the critical test indicating whether the sun compass was still intact, namely to record the orientation of the formerly clock-shifted birds immediately after their sun compass was set back to normal. This was done in the two studies by Wiltschko et al. (1976, 1984), with a pronounced deflection in the reverse direction indicating a recalibration of the sun compass.
The familiar birds of Gagliardo et al.(2005) indeed showed their largest deviation, 98°, on their first clock-shift release, a deflection that lies clearly within the range of the deflections observed in the unfamiliar birds - 127°, 116° and 68°. The later deflections of the familiar birds, 78° and 31°, are smaller and can no longer be assumed to reflect an intact sun compass.
Yet even so, the claim that familiarity with the release site reduces the deflection is not justified. A statistic based on matched pairs of data cannot be applied on just three releases, and a test such as the t-test does not show a significant difference (t=1.334, P>0.05). Hence, although the test design leads to a decrease in deflections of the familiar birds, one might say that, at best, the case is open.
To answer the question requires more data obtained under equal conditions for the familiar and unfamiliar birds. We have just completed a series of six such tests, with birds of equal pre-experience released on the same days at sites that were familiar to some of the birds but unfamiliar to the others,with controls and 6 h fast-shifted birds of both groups. Our data clearly contradict the claim of Gagliardo et al.(2005): the deflections vary considerably, but those of the familiar birds range from 39° to 169°(median, 81°) and those of the birds unfamiliar with the site range from 27° to 124° (median, 68°)(Wiltschko et al., 2005). That is, the size of the deflection induced by clock-shifting is not affected by familiarity with the release site.
To answer the question of what factor reduces the clock-shift-induced deflection, we may point out that clock-shifted pigeons show the expected deviation when their magnetic compass is temporarily disrupted by magnets(Wiltschko and Wiltschko,2001): old pigeons seem to fly a compromise between the directions indicated by their sun compass and their magnetic compass.
