Abstract
Mutualistic plant-animal interactions for seed dispersal are crucial for vegetation dynamics, benefiting over half of the world's plant species. Beyond the tropics, the Mediterranean biome harbors the highest proportion of species adapted to endozoochory, yet major gaps remain in quantifying interaction diversity in these biodiversity-rich areas and their links to ecosystem functioning. High-resolution, quantitative interaction data are essential not only to fill these gaps but also to enable large-scale ecological modeling of species interactions across biomes. Here, we present the FRUGivory INTegration (FRUGINT) dataset - an extensive collection of quantitative frugivory interactions and associated species traits from a Mediterranean biodiversity hotspot in southwestern Spain. By integrating six complementary sampling methods (camera trapping, continuous-monitoring cameras, DNA-barcoding, mist-netting, direct observation and track records) across multiple years, the dataset overcomes limitations of sampling biases, variable effort and spatio-temporal heterogeneity, providing a comprehensive picture of plant-frugivore interactions across the region. Based on a total of 37,923 interaction records and 481 unique pairwise interactions, involving 26 fleshy-fruited plant species present in Doñana and 78 frugivorous vertebrate species, FRUGINT yields estimates of regional-scale plant-frugivore networks based on pairwise interaction probabilities. The dataset encompasses both common and numerous rare interactions, offering a valuable resource for advancing research on plant-animal interactions, network ecology, and biodiversity conservation.