Résumé
An orientation component leads to directionally biased paths, with major consequences in animal population redistribution. Classical orientation analyses, which focus on the overall direction of motion, are useless for detecting such a component when the preferred direction is not common to the whole population, but differs from one path to another. In-depth path analyses are required in this case. They consist of determining whether paths are more suitably represented as biased or unbiased random walks. The answer is not easy because most animals' paths show some forward persistence propensity that acts as a purely local directional bias and, hence, blurs the possible occurrence of an additional, consistent bias in a preferred direction. I highlight the key differences between biased and unbiased random walks and the different ways orientation mechanisms can generate a consistent directional bias. I then examine the strength and weakness of the available methods likely to detect it. Finally, I introduce a new procedure based on the backward evolution of the beeline distance, from the end of the path, which might correspond to a goal toward which the animal orients itself, to each of the animal's preceding locations. This new procedure proves to be very efficient, as it requires only a small sample of short paths for detecting a possible orientation component.