Are the assemblages of tree pollination modes being recovered by tropical forest restoration?

Questions Do the assemblages of pollination modes in restored (tree plantings) and secondary (naturally regenerated) forests change in comparison to primary forests, and how do these assemblages relate to species turnover at regional scale? Location Southeast region of Brazil. Methods We classified tree species found in a total of 40 forest sites (18 primary, 11 restored, 11 secondary) according to pollination mode, based on the literature. We calculated and compared functional dissimilarity distances, amounts of species and accumulated abundance of pollination modes, and functional indices of richness and evenness between forest types. Results Functional dissimilarity distances were much smaller than species dissimilarity distances within forest types (mean <20%, >80%, respectively), indicating a small variation in pollination modes between sites. Functional indices of richness and evenness did not differ between forest types. However, significant changes were found in the species and abundance proportions of several pollination modes. Primary forests were characterized by the predominance of generalized insect‐pollinated species, followed by secondary proportions of bee, wind and moth pollination; other pollination modes were underrepresented. In restored forests, reductions were found in generalized insect, moth, wind, fly, pollen‐consuming insect and very‐small insect pollination, whereas the species pollinated by bees and bats more than doubled. Smaller changes were found among secondary forests, including reductions in moth, fly and fig‐wasp pollination, whereas there were incremental changes in bee, beetle, big animal and small insect pollination. Conclusions Our results indicate a rather stable assemblage of pollination modes and also high ecological redundancy among trees regardless of the species replacement at the regional scale. Major changes among restored forests are probably in response to larger disturbance effects and/or restoration practices conducted in these sites. In contrast, smaller changes among secondary forests could be in response to smaller disturbance effects and natural selection processes, and also seem to suggest that highly resilient degraded areas are more likely to recuperate their functional diversity through natural regeneration alone. In both cases, however, efforts to recover such patterns should be encouraged to avoid possible negative effects in plant–pollinator interactions.