Large terrestrial herbivores play crucial roles in shaping ecosystem structure and function through their foraging activities. Still, the dietary ecology of elusive tropical species remains poorly understood. We investigated the diet composition of lowland tapirs (Tapirus terrestris), the largest terrestrial herbivore in the Neotropics, using DNA metabarcoding of fecal samples from 31 latrines in Carlos Botelho State Park, Brazil. We characterized local plant communities through vegetation plots and analyzed five leaf economic spectrum (LES) traits from both consumed and surrounding vegetation to assess selective feeding patterns. Lowland tapirs consumed 61 plant species from 69 genera and 46 families, predominantly those from the Melastomataceae, Asteraceae, and Myrtaceae families. Beta diversity analysis revealed high compositional turnover among latrines, with a high dissimilarity index, indicating that the samples being compared are distinct in terms of species composition. The plant composition in tapir diets differed significantly from that of the surrounding vegetation, suggesting that this species forages on distinct plant species across its extensive home range rather than consuming locally abundant species. Finally, the functional trait analysis revealed no significant differences between the dietary species and the surrounding vegetation in terms of LES traits. Tapirs consumed plants that spanned both ”fast” (high specific leaf area and high nitrogen content) and ”slow” (high leaf dry matter content and thick leaves) strategies, indicating a broad dietary tolerance rather than trait-based selectivity. This suggests that tapirs can adapt to diverse plant textures and nutritional profiles, browsing from tough to softer, more digestible leaves. Our findings demonstrate that lowland tapirs exhibit generalist feeding strategies, which promote high plant species turnover, potentially contributing to the maintenance of tropical forest diversity, as observed in the Atlantic forest. Given the critical threats facing this endangered megafauna, understanding their generalist diet is essential for developing effective conservation strategies.

Fruit-eating birds play a vital role in seed dispersal across fragmented and regenerating landscapes, shaping forest recovery. Simultaneously, human disturbance influences early successional forest communities by favoring habitat-generalist birds and facilitating the spread of exotic plants. To evaluate quantitative seed dispersal effectiveness (SDE) and identify bird-consumed plants in these novel communities, we collected bird droppings at ten early successional sites in semi-deciduous Atlantic Forest in São Paulo, Brazil. Monthly mist netting over one year (2763 net hours) yielded fecal samples from captured birds. Droppings were inspected for seeds to assess the quantitative component of SDE among bird species. To broaden our diet analysis beyond seeds, we also conducted DNA metabarcoding of 74 droppings from the most effective seed dispersers using a universal plant mini-barcode. Only 25% of droppings from frugivores contained seeds. We identified Chiroxiphia galeata (Pipridae) as the most quantitatively effective seed disperser, followed by Turdus leucomelas (Turdidae) and Ramphocelus carbo (Thraupidae). Quantitative SDE was associated with species’ degree of frugivory but not body mass. Using both seeds and DNA metabarcoding, 84 plant taxa were detected across 42 families, including exotic, weedy, and potentially invasive species. Native Casearia (Salicaceae) and Miconia (Melastomataceae) were the most frequently consumed plants. Despite limited habitat quality and low fruit availability, a forest-dependent species (C. galeata) proved to be the most effective disperser. In contrast, the diets of generalist species reflected resource use in the matrix (e.g., pastures, residential areas), suggesting these birds may facilitate both forest connectivity and the spread of exotic plants. Our study highlights the complementary roles of generalist and forest-dependent bird species, as well as the potential consequences of matrix foraging on the future community composition in recovering neotropical forests.

Environmental DNA (eDNA) and invertebrate-derived DNA (iDNA) have been increasingly recognized as a powerful tool for biodiversity assessment and conservation management. However, because of uncertainties on the efficiency of eDNA/iDNA approach in comparison to conventional methods, its use to assess the vertebrate diversity is still rare. Here we assessed the efficiency of eDNA/iDNA in comparison to conventional methods to survey vertebrate diversity across several type of samplers, vertebrate groups, and location (tropical vs temperate zones), as well as its efficiency to be used as a proxy for relative abundance or biomass across different molecular methods (qPCR and metabarcoding) and types of experiments (in the lab or in the field). The metanalysis showed that, in general, there is no difference in the number of species detected or number of sites that a target species was detected when using eDNA/iDNA or conventional methods, suggesting that eDNA/iDNA and conventional methods were equally efficient in characterizing the biodiversity. However, for water sampler and fish, separately, the risk of not finding a species was greater using conventional method than eDNA, suggesting that eDNA/iDNA was more efficient in finding the target species. Abundance and biomass showed similar correlation patterns, and there was a positive correlation between eDNA/iDNA and abundance/biomass data, suggesting that eDNA/iDNA can be used as a proxy for abundance and biomass. Therefore, eDNA/iDNA has proved to be an efficient tool to assess vertebrate diversity in terms of both diversity of species and abundance or biomass.

Ingested-derived DNA (iDNA) from insects can represent a powerful tool for assessing vertebrate diversity because insects are easy to sample, have a varied diet, and are widely distributed. Despite these advantages, the use of iDNA for mammalian detection is still little explored, especially in the neotropical region. Here we aimed to compare the effectiveness of mosquitoes and flies to detect mammals in a semi-controlled area of a Zoo that houses native and non-native species. We evaluated the number of mammal species detected by the iDNA samplers and verified the distance range of each sampler group for detecting the mammal species. To capture mosquitoes and flies we used CDC (Center for Disease Control) and fish-baited plastic bottle traps, respectively, distributed in eight sampling points during five days. Using two mini-barcodes (12SrRNA and 16SrRNA) and the metabarcoding approach, we identified 45 Operational Taxonomic Units from 10 orders. There was no difference between the number of species recovered per individual insect, although the number of flies captured was higher, resulting in more mammal species recovered by this insect group. Eight species were recorded exclusively by mosquitoes and 20 by flies, suggesting that using both samplers allowed a more comprehensive screening of the biodiversity. The maximum distance recorded was 337 m for flies and 289 m for mosquitoes, but the average range distance did not differ between insect groups. Our essay proved to be quite efficient for the mammal detection, considering the high number of species detected with a reduced sampling effort. Thus, combining iDNA from different samplers and metabarcoding can be a powerful tool for mammal survey and monitoring in the neotropics.