In tropical regions, most primary ecosystems have been replaced by mosaic landscapes in which species must cope with a large shift in the distribution of their habitat and associated food resources. Primates are particularly vulnerable to habitat modifications. Most species persist in small fragments surrounded by complex human-mediated matrices whose structure and connectivity may strongly influence their dispersal and feeding behavior. Behavioral plasticity appears to be a crucial parameter governing the ability of organisms to exploit the resources offered by new matrix habitats and thus to persist in fragmented habitats. In this study, we were interested in the dietary plasticity of the golden-crowned sifaka (Propithecus tattersalli), an endangered species of lemur, found only in the Daraina region in north-eastern Madagascar. We used a DNA-based approach combining the barcoding concept and Illumina next-generation sequencing to (i) describe the species diet across its entire range and (ii) evaluate the influence of landscape heterogeneity on diet diversity and composition. Faeces from 96 individuals were sampled across the entire species range and their contents were analyzed using the trnL metabarcoding approach. In parallel, we built a large DNA reference database based on a checklist of the plant species of the Daraina region. Our results suggest that golden-crowned sifakas exhibit remarkable dietary diversity with at least 130 plant species belonging to 80 genera and 49 different families. We highlighted an influence of both habitat type and openness on diet composition suggesting a high flexibility of foraging strategies. Moreover, we observed the presence of numerous cultivated and naturalized plants in the faeces of groups living in forest edge areas. Overall, our findings support our initial expectation that P. tattersalli is able to cope with the current level of alteration of the landscape and confirm our previous results on the distribution and the dispersal ability of this species.

Characterizing the trophic relationships between large herbivores and the outstanding plant diversity in rainforest is a major challenge because of their elusiveness. This is crucial to understand the role of these herbivores in the functioning of the rainforest ecosystems. We tested a non-invasive approach based on the high-throughput sequencing of environmental samples using small plant plastid sequences (the trnL P6 loop) and ribosomal ITS1 primers, referred to as DNA metabarcoding, to investigate the diet of the largest neotropical herbivore, the lowland tapir. Sequencing was performed on plant DNA extracted from tapir faeces collected at the Nouragues station, a protected area of French Guiana. In spite of a limited sampling, our approach reliably provided information about the lowland tapir's diet at this site. Indeed, 95.1% and 74.4% of the plant families and genera identified thanks to the trnL P6 loop, respectively, matched with taxa already known to be consumed by tapirs. With this approach we were able to show that two families and eight new genera are also consumed by the lowland tapir. The taxonomic resolution of this method is limited to the plant family and genera. Complementary barcodes, such as a small portion of ITS1, can be used to efficiently narrow identifications down to the species in some problematic families. We will discuss the remaining limitations of this approach and how useful it is at this stage to unravel the diet of elusive rainforest herbivores and better understand their role as engineers of the ecosystem.

While the populations of large herbivores are being depleted in many tropical rainforests, the importance of their trophic role in the ecological functioning and biodiversity of these ecosystems is still not well evaluated. This is due to the outstanding plant diversity that they feed upon and the inherent difficulties involved in observing their elusive behaviour. Classically, the diet of elusive tropical herbivores is studied through the observation of browsing signs and macroscopic analysis of faeces or stomach contents. In this study, we illustrate that the original coupling of classic methods with genetic and ethnobotanical approaches yields information both about the diet diversity, the foraging modalities and the potential impact on vegetation of the largest terrestrial mammal of Amazonia, the lowland tapir. The study was conducted in the Guianan shield, where the ecology of tapirs has been less investigated. We identified 92 new species, 51 new genera and 13 new families of plants eaten by tapirs. We discuss the relative contribution of our different approaches, notably the contribution of genetic barcoding, used for the first time to investigate the diet of a large tropical mammal, and how local traditional ecological knowledge is accredited and valuable for research on the ecology of elusive animals.

Traditional rangelands in many developing countries are currently being encroached by cultivation, driving some herders to illegally use protected areas for grazing their cattle. Since cattle are an exotic species in these ecosystems, they might have an impact on the local wild herbivore communities, notably through competition. We used point pattern statistics to characterise the spatial relationships between wild ungulate species and cattle herds within a protected area in west Africa undergoing seasonal intrusions by cattle. We predicted that the wild ungulate species that are ecologically and morphologically similar to cattle, in terms of body mass and diet, would be more sensitive to grass depletion by cattle and would be separated from cattle to a larger extent. The spatial distribution of browsing and mixed-feeding antelopes did not seem to be affected much by cattle presence, whereas most grazing species showed spatial separation from cattle. Interestingly, elephants also showed significant separation from cattle herds. We discuss the likely processes that may have contributed to the observed spatial patterns. The spatial displacement of certain wild species, including megaherbivores, affects the whole community structure and, thus, other components of the ecosystem