Abstract
Diaspore (e.g., seed and spore) dispersal is recognized as a key mechanism in plant dynamics, including endozoochory, which can be a risky journey for diaspores. Endozoochory is achieved when diaspores are consumed and may germinate after the mastication, the gut and fecal matrix passage, all representing filters for diaspores. Nevertheless, endozoochory is a highly studied mechanism through numerous methods, notably based on the observation of frugivorous behavior, diaspores retrieved in feces or germination experiments. However, most of those methods consider partially the endozoochorous filters (ingestion, mastication, gut, feces). Hence, the combined effect of the methods and filters consideration may lead to biased inference of endozoochory. In this study, we used a collection of 52 brown bear (Ursus arctos) feces to highlight five methods inferring endozoochory. Two methods consider the ingestion filter and used metabarcoding of fecal eDNA to identify fleshy fruits (i) or plants during fruiting periods (ii). The third method (iii) was based on the intact propagules retrieved in feces, considering ingestion and mastication filters. Another method (iv) was based on the germination from disaggregated feces, considering up to the gut passage filter. The last method (v) was based on the germination from aggregated feces, considering the four filters. We compared the number of taxa, the community and the plant life forms inferred among methods. We inferred the largest number of taxa in method (iii), but the germination-based methods inferred the most diverse plant life forms. We identify few shared taxa among methods. The metabarcoding-based methods might be an interesting tool as a first approximation of endozoochory while detailing the diet. The method (v) appeared as the most reliable. Overall, we invite future studies to mitigate their interpretations according to the filters of endozoochory considered and plant detectability related to the method used.