Invasive non-native species are major drivers of biodiversity loss, particularly on island ecosystems where endemicity rates are high and food webs simplified compared to the mainland. Herbivores, specifically, are among the most destructive invaders worldwide, as they dramatically damage native flora and alter ecosystem functioning. Plant reproductive performance can be directly impacted by herbivore consumption of flowers and indirectly by resource allocation and disruption of plant-animal interactions. However, the understanding of these indirect effects remains limited. Moreover, plant-herbivore interactions can switch from positive to negative outcomes depending on the identity and density of the interacting species, making the situation more complex when multiple herbivore species affect the same plant. It is crucial thus to comprehensively understand the full range of ecological impacts exerted by non-native herbivores in order to effectively preserve native biodiversity and associated ecosystem services. This thesis aims to enhance our knowledge of how contrasting invasive non-native herbivores influence plant reproduction and plant-animal interactions from a multi-specific and density-dependent perspective. To this end, I used the model system composed by the Mediterranean dwarf palm Chamaerops humilis L. and its two main invasive herbivores in Mallorca (Balearic Islands, Spain), the stem-boring moth Paysandisia archon Burmeister and the feral goat Capra hircus L. Direct and indirect effects were evaluated throughout several stages of the palm’s life cycle by combining intensive observational field work and spatially explicit modelling. Spatial point pattern analyses revealed that the probability and intensity of moth herbivory and goat herbivory were determined by the density of dwarf palms. The sign and strength of the spatial patterns were highly variable as they changed across distance (distance- and density-dependent effects), types of herbivory (moth herbivory, goat florivory, goat folivory), study sites, and even between palm sexes. At the flower stage, moth herbivory positively affected pollinator abundance and fruit initiation, while goat herbivory had negative effects, mainly though intense florivory. Interestingly, only the nitidulid pollinator Meligethinus pallidulus Erichson was impacted by herbivory, positively by moth herbivory while negatively by goat herbivory. Furthermore, both herbivore species exerted nonadditive effects on palm reproduction as, for instance, the palms that experienced simultaneous attacks showed remarkably lower rates of fruit initiation. At the fruit stage, moth-attacked palms showed higher rates of pre-dispersal insect-seed predation, likely due to the observed decreased pulp content in fruits. Surprisingly, legitimate seed dispersers consumed fruits more often from attacked palms, suggesting that herbivore-mediated changes might boost frugivore attraction or, alternatively, that both mutualistic and antagonistic animals select the same palm traits. Lastly, a comparative spatially explicit analysis among three study plots with varying levels of goat activity revealed that the net outcome of the palm-goat interaction shifted from mutualism to antagonism depending on goat density. Low goat activity led to an aggregated seed rain and, in turn, higher rates of insect-seed predation and lower rates of germination. High goat activity almost completely canceled palm recruitment, leading to the population reproductive collapse. By contrast, an intermediate goat activity maintained a balance between mutualistic and antagonistic effects. Overall, this thesis provides valuable insights into the diverse and multiple pathways through which contrasting non-native herbivores can impact plant reproductive performance, highlighting the importance of considering intraspecific variability and the context-dependence nature of plant-animal interactions. Given the ongoing increase in non-native species worldwide, it becomes essential to implement adaptive and cooperative management strategies among the different social stakeholders to efficiently safeguard native biodiversity.