Fig. 1: Magnetic secular variation and philopatry.

a–d A schematic demonstrating how slight magnetic secular variation might cause changes in distribution even in animals with low-resolution magnetoreceptors. In Case a, the magnetic field has shifted left slightly, causing a drop in recovery probability at the starred location even though most animals return to the natal site (owing to the tight distribution caused by low magnetic sensitivity). Case b is intermediate, whilst Case c shows the opposite phenomenon; the field has shifted right, and a response can be recorded at the starred site even though the sensitivity of any given animal is very low. e A simulation showing how even noisy sensors can achieve remarkably accurate philopatry (<0.05 degrees) through repeated sampling of the magnetic field. Each point represents 100 simulations where the magnetic field is sampled n times with both sensor error and local secular variation. Whiskers represent standard deviation. Magnetic data is derived from the Hartland Magnetic Observatory (UK).