A subterranean mammal uses the magnetic compass for path integration

Abstract

Path integration allows animals to navigate without landmarks by continuously processing signals generated through locomotion. Insects such as bees and ants have evolved an accurate path integration system, assessing and coding rotations with the help of a general directional reference, the sun azimuth. In mammals, by contrast, this process can take place through purely idiothetic (mainly proprioceptive and vestibular) signals. However, without any stable external reference for measuring direction, path integration is highly affected by cumulative errors and thus has been considered so far as valid only for short-distance navigation. Here we show through two path integration experiments (homing and shortcut finding) that the blind mole rat assesses direction both through internal signals and by estimating its heading in relation to the earth's magnetic field. Further, it is shown that the greater the circumvolution and length of the traveled path, the more the animal relies on the geomagnetic field. This path integration system strongly reduces the accumulation of errors due to inaccuracies in the estimation of rotations and thus allows the mole rat to navigate efficiently in darkness, without the help of any landmark, over both short and long distances.

Fig. 1.

Apparatus, Experiment 1: wheel-shaped maze used to test homing behavior. (A) The (guided) outward journey leads from the nest to a peripheral junction (X) and then to a central baited goal (cG). There the animal chooses one radial arm to return to the nest. (B) Enlargement of a peripheral junction, which consists of three doors to control the animal's movement and a removable plastic cap to seal the peripheral end of the arm.

Fig. 3.

Experiment 2. (A and B) First shortcut route chosen by each individual trained to follow the long predefined route (dashed line on A) leading from the nest (N) to a feeding goal (G) and tested with all maze routes open. The tests took place either under the natural geomagnetic field (A) or under the altered magnetic field (B). (C and D) Average (±SE) route length (C) and number of turns (D) to reach the goal in 12 trials by the group tested under the natural geomagnetic field (filled squares) and the group tested under the altered magnetic field (open circles). The horizontal dashed lines in C and D represent (i) the route length and number of turns of the training path (upper line) and (ii) the route length and number of turns of the theoretical shortest path to the goal (lower line). Asterisks indicate that the results obtained for the natural and the altered magnetic fields differed significantly (P < 0.01, Mann–Whitney U test).

Fig. 2.

Experiment 1: homing direction of six subjects, tested under the natural (Left) or altered (Right) magnetic field and after a short (less than one full round, Top) medium (more than one but less than two full rounds, Middle) or long (more than three but less than four full rounds, Bottom) peripheral outward journey. The histograms around the circles indicate the percentage of trials in which a particular radial arm was chosen by the subjects. Arrows with small arrowheads represent the mean homing direction of each subject; dashed arrows with a large arrowhead represent the mean direction of the whole experimental group. Statistical data are specified in Table 1.