We will not respond in detail to the lengthy and speculative arguments by which Muheim et al. try to maintain the general validity of their hypothesis, although we are surprised by their negative attitude toward adaptations of birds to specific situations of their migration, in particular concerning rapid changes in declination. Instead, we will focus on some crucial points:

(1) There is agreement that birds can recalibrate celestial cues—stars and the polarisation pattern—by use of the magnetic field, if they can see the central part of the sky. Muheim et al. (2006a, b) claim, however, that birds recalibrate the magnetic course by the pattern of polarisation at sunrise and sunset, if they can see the area just above the horizon. Assigning particular importance to the area just above horizon makes little sense, when considering the degree of polarisation of the natural sky—it has its maximum, about 70%, at the zenith, but is substantially less near the horizon (Wehner 1982), with the amount depending strongly on the haze in the atmosphere. Also, as we have already pointed out, assuming a reversal of the calibration processes depending on whether or not the birds can see the sky just above horizon is counter-intuitive and makes no sense.

(2) In their argument in support of the crucial importance of the area just above horizon for the calibration of the magnetic compass, Muheim et al. (2006a, b) rely heavily on two papers, those by Able and Able (1995a) and Cochran et al. (2004). The latter, as we detailed in a footnote in our paper (Wiltschko et al. 2008), might involve an entirely different phenomenon. Using the paper by Able and Able (1995a) to support their argument is problematic, as it is not clear at all whether these birds could see the sky down to the horizon—it does not say so in the paper. Muheim et al. seem to base their hypothesis on a mere assumption. Another paper (Able and Able 1995b) describes how handraised savannah sparrows recalibrated their magnetic compass to polarized light when exposed in Emlen funnels. Hence the studies by Able and Able (1995a, b) do not support the assumption that the area just above horizon is crucially important for compass calibration.

(3) Selecting birds that happen by chance to head in the migratory semicircle from a randomly oriented sample does not mean that these birds are in migratory mood. Muheim et al. quote a high fat score and restlessness as evidence for the migratory state of their birds, but this is not convincing as long as the crucial criterion, namely orientation in a migratory direction, is lacking.

(4) Faced with the differing results of their study (Muheim et al. 2006b) and ours (Wiltschko et al. 2008) Muheim et al. now present another highly speculative explanation based on mere assumptions, namely that our birds may have completed compass calibration prior to the experimental exposure, e.g., when they were held in the outdoor aviary. Yet our birds could not see the horizon or the experimental site from the aviary, as it was partly shielded against the sun and wind, surrounded by buildings and trees, and was lower on the hill.

More importantly, however, these arguments make no sense at all, because any calibration in the aviary would have reinforced the natural calibration from the wild before capture. The same applies to the birds of the Muheim et al. (2006b) study—had these birds already migrated, they would already have made the natural calibration of their compass systems and, according to the present comment by Muheim et al., would not be misled by one short exposure to conflicting cues. Here Muheim et al. (2006a, b and comment) show an inconsistency in their argument—they originally argue that the magnetic compass would be regularly recalibrated by the polarisation pattern, but, faced with the discrepancy between findings, now change their argumentation in favour of existing stable calibrations.

(5) We agree with Muheim et al. that the different findings may partly be caused by the exposure protocol. If birds taken from a random sample are exposed to an artificial stimulus (a stimulus shown to produce artefacts before, see Helbig and Wiltschko 1989); it is perhaps not surprising that we could not replicate these results with birds preferring their migratory direction, exposed to the natural polarisation pattern.