Pigeon homing in relation to geomagnetic, gravitational, topographical, and meteorological conditions

Abstract

Experienced homing pigeons were released at sites unfamiliar to them and with magnetic and gravity anomalies as well as in areas with rather normal fields throughout the FRG (41 releases when sunny, 14 when overcast; Figs. 1–3). The second-order release data were subjected to both univariate and multivariate statistical analysis (stepwise regression, factor analysis). The magnetic field strength and its gradients within the 1-km circle around the release site were determined from aeromagnetic maps of the anomalies of magnetic total intensity. Analogous variables were derived from gravity anomaly maps. It was tested whether the pigeons fly along that gradient to minimize the difference between the magnetic or gravity field at the release site and the loft at maximum rate. Further independent variables described magnetic K index, day-to-day variations of the magnetic components, topography, meteorological conditions, the number of the releases the pigeons had done, and the distance. Over magnetic anomalies widely varying in strength (departing − 250 nT to 300 nT from normal 600 m above ground), extent, and distance from loft, the pigeons vanished with less deviation from the homeward direction and faster than they did in areas with less irregular fields under sunny conditions; this is in contrast to other studies on magnetic anomalies, except one. At sites of gravity anomalies (15–49 mgal), the pigeons were significantly less homeward oriented and homed slower than at less anomalous sites (−9 to 14 mgal). Variables related to gravity were best predictors in 8 and and second predictors in 3 out of 15 regression analyses of the navigational parameters for the releases under sun. Six times the (absolute) amount of the gravity difference between the release site and the loft was selected first (Figs. 7B, C, 8B, C). The results suggest gravity to be involved in navigation as the pigeons' distance measure. Homeward directedness declined with increasing amount of the gravity gradient in the first 12 releases under sunny skies as well as when overcast (Fig. 8A). A preferred compass direction towards north-northeast was determined, being closest to the grand mean vector of the ascending gravity gradient (Fig. 6). The analyses failed to show directional preferences as assumed by the hypotheses tested. Temperature and degree of cloud cover provided some information for predicting mean vector lengths and mean vanishing times, respectively. The surface wind component in the homeward direction was correlated with median homing performances.