Plant communities are often dispersal-limited and zoochory can be an efficient mechanism for plants to colonize new patches of potentially suitable habitat. We predicted that seed dispersal by ungulates acts as an ecological filter -which differentially affects individuals according to their characteristics and shapes species assemblages -and that the filter varies according to the dispersal mechanism (endozoochory, fur-epizoochory and hoof-epizoochory). We conducted two-step individual participant data meta-analyses of 52 studies on plant dispersal by ungulates in fragmented landscapes, comparing eight plant traits and two habitat indicators between dispersed and non-dispersed plants. We found that ungulates dispersed at least 44% of the available plant species. Moreover, some plant traits and habitat indicators increased the likelihood for plant of being dispersed. Persistent or nitrophilous plant species from open habitats or bearing dry or elongated diaspores were more likely to be dispersed by ungulates, whatever the dispersal mechanism. In addition, endozoochory was more likely for diaspores bearing elongated appendages whereas epizoochory was more likely for diaspores released relatively high in vegetation. Hoof-epizoochory was more likely for light diaspores without hooked appendages. Furepizoochory was more likely for diaspores with appendages, particularly elongated or hooked ones. We thus observed a gradient of filtering effect among the three dispersal mechanisms. Endozoochory had an effect of rather weak intensity (impacting six plant characteristics with variations between ungulate-dispersed and non-dispersed plant species mostly below 25%), whereas hoof-epizoochory had a stronger effect (eight characteristics included five ones with above 75% variation), and fur-epizoochory an even stronger one (nine characteristics included six ones with above 75% variation). Our results demonstrate that seed dispersal by ungulates is an ecological filter whose intensity varies according to the dispersal mechanism considered. Ungulates can thus play a key role in plant community dynamics and have implications for plant spatial distribution patterns at multiple scales.Plant communities are often dispersal-limited and zoochory can be an efficient mechanism for plants to colonize new patches of potentially suitable habitat. Our analysis is the first synthesis of ungulate seed dispersal that compares characteristics from both non-dispersed and dispersed diaspores, distinguishing the three zoochory mechanisms ungulates are involved in: endozoochory, hoof-epizoochory and fur-epizoochory. We confirmed that seed dispersal by ungulates is an ecological filter whose intensity increases from endozoochory, then hoof-epizoochory to finally fur-epizoochory. By filtering seed traits through dispersal, ungulates can thus play a key role in plant community dynamics and have implications for plant spatial distribution patterns at multiple scales. Synthesis

Plant dispersal syndromes are allocated based on diaspore morphology and used to predict the dominant mechanisms of dispersal. Many authors assume that only angiosperms with endozoochory, epizoochory or anemochory syndromes have a longdistance dispersal (LDD) mechanism. Too much faith is often placed in classical syndromes to explain historical dispersal events and to predict future ones. What is usually recorded as the 'endozoochory syndrome' is in reality a 'frugivory syndrome' and this has often diverted attention from endozoochory by non-frugivores (e.g. waterbirds and large herbivores) that disperse a broad range of angiosperms, for which they likely provide the maximum dispersal distances. Neither the endozoochory nor the epizoochory syndromes provide helpful predictions of which plants non-frugivores disperse, or by which mechanism. We combined data from previous studies to show that only 4% of European plant species dispersed by ungulate endozoochory belong to the corresponding syndrome, compared to 36% for ungulate epizoochory and 8% for endozoochory by migratory ducks. In contrast, the proportions of these species that are assigned to an 'unassisted syndrome' are 37, 31 and 28%, respectively. Since allocated syndromes do not adequately account for zoochory, empirical studies often fail to find the expected relationship between syndromes and LDD events such as those underlying the colonization of islands or latitudinal migration rates. We need full incorporation of existing zoochory data into dispersal databases, and more empirical research into the relationship between plant traits and the frequency and effectiveness of different dispersal mechanisms (paying attention to unexpected vectors). Acknowledging the broad role of non-frugivores in facilitating LDD is crucial to improve predictions of the consequences of global change, such as how plant distributions respond to climate change, and how alien plants spread. Networks of dispersal interactions between these vertebrates and plants are a vital but understudied part of the Web of Life.

The intentional removal or addition of species or specific human impacts on ecosystems trigger changes that can help us understand species interactions. In many temperate forests, deer populations are increasing and so is the need to understand how they influence ecosystems. We took advantage of the introduction of Sitka black-tailed deer (Odocoileus hemionus sitchensis Merriam) to the Queen Charlotte Islands (Haida Gwaii), British Columbia, Canada, to study how hunting pressure affects the impact of deer on tree regeneration after logging. We show that although the regeneration of western redcedar (Thuja plicata Donn ex D. Don) is drastically reduced in presence of deer, regeneration is better and browsing stress lower, in areas where deer are more exposed to hunting. Similar effects of accessibility for hunters are observed on browsing stress of Sitka spruce (Picea sitchensis (Bong.) Carrière). Western hemlock (Tsuga heterophylla (Raf.) Sarg.) is not significantly affected, and its regeneration is not correlated to hunting. We suggest that the effect of hunting on tree regeneration could be explained by the incidence of hunting on deer behaviour rather than by the actual number of deer killed by hunters. These results suggest that the future occurrence of redcedar stands in second-growth forests on this archipelago may depend on the amount and distribution of deer hunting.

Large wild ungulates are a major biotic factor shaping plant communities. They influence species abundance and occurrence directly by herbivory and plant dispersal, or indirectly by modifying plant-plant interactions and through soil disturbance. In forest ecosystems, researchers' attention has been mainly focused on deer overabundance. Far less is known about the effects on understory plant dynamics and diversity of wild ungulates where their abundance is maintained at lower levels to mitigate impacts on tree regeneration. We used vegetation data collected over 10 years on 82 pairs of exclosure (excluding ungulates) and control plots located in a nation-wide forest monitoring network (Renecofor). We report the effects of ungulate exclusion on (i) plant species richness and ecological characteristics, (ii) and cover percentage of herbaceous and shrub layers. We also analyzed the response of these variables along gradients of ungulate abundance, based on hunting statistics, for wild boar (Sus scrofa), red deer (Cervus elaphus) and roe deer (Capreolus capreolus). Outside the exclosures, forest ungulates maintained higher species richness in the herbaceous layer (+15%), while the shrub layer was 17% less rich, and the plant communities became more light-demanding. Inside the exclosures, shrub cover increased, often to the benefit of bramble (Rubus fruticosus agg.). Ungulates tend to favour ruderal, hemerobic, epizoochorous and non-forest species. Among plots, the magnitude of vegetation changes was proportional to deer abundance. We conclude that ungulates, through the control of the shrub layer, indirectly increase herbaceous plant species richness by increasing light reaching the ground. However, this increase is detrimental to the peculiarity of forest plant communities and contributes to a landscape-level biotic homogenization. Even at population density levels considered to be harmless for overall plant species richness, ungulates remain a conservation issue for plant community composition.