Although most recent hypotheses of pigeon homing have assigned an essential role to the sun, there has been some evidence suggesting that the sun is not essential. Two series of releases were designed to examine the question more carefully. Birds whose internal clocks had been shifted 6 hours were used in the critical tests. Under sun, the vanishing bearings of the clock-shifted birds were deflected in the direction predicted by a hypothesis of use of the sun as a simple compass. By contrast, under total overcast the bearings of both the clock-shifted and the control birds were homeward oriented and there was no difference between them, even at a release site the birds could never have seen previously. Therefore it is concluded that the sun is used as a compass when it is available, but that the pigeon navigation system contains sufficient redundancy to make accurate orientation possible in the absence of both the sun and familiar landmarks; the orientational cues used under such conditions do not require time compensation. This conclusion is in complete disagreement with the Matthews sun-arc hypothesis of pigeon navigation, and it makes necessary a major reformulation (at the very least) of the other principal hypothesis, that of Kramer.