Fig. 2

Simulated autumn migration routes of Willow warblers using alternative compass mechanisms and aiming to the wintering ground (first winter stop) through: (a) direct route or (b) an intermediate goal area (stopover region). In panel (A) only the sun compass route (b) crosses the identified stopover region (filled black triangles) and neither of the two possible solutions for the magnetoclinic route (d and d*) could reach the wintering area (see text). In panel (B) all routes have a solution that bring birds to the stopover region before switching the compass direction toward the wintering ground (black circle 1). However, for the magnetoclinic route the switch must occur earlier and at higher latitudes (between black circle 2 and 3) than for the other compass mechanisms (black circle 1). c A sun compass can produce alternative routes when the bird’s internal clock partially adapts to the local time (i.e. partial longitudinal time-shift), but an intermediate goal area would still be necessary to cross the stopover region. d Alternative magnetoclinic routes can be generated from locations with alternative geomagnetic inclination values (i.e. different starting points) for which a change in compass on route must always occur not earlier than 85—95 °E longitude. For all the panels, departure location (filled black circle) is the breeding area where birds where tagged. GLS data for the three birds tracked in this study are also reported. Maps are in Mercator projection with 15° grid and all simulated routes are 12,000 km except the northerly magnetoclinic route (d) in panel (A) that is 7750 km