Information about the relative importance of competitive or facilitative pollinator-mediated interactions in a multi-species context is limited. We studied interspecific pollen transfer (IPT) networks to evaluate quantity and quality effects of pollinator sharing among plant species on three high-Andean communities at 1600, 1800 and 2000 m a.s.l. To estimate the sign of the effects (positive, neutral or negative), the relation between conspecific and heterospecific pollen deposited on stigmas was analysed with GLMMs. Network analyses showed that communities were characterised by the presence of pollen hub-donors and receptors. We inferred that facilitative and neutral pollinator-mediated interactions among plants prevailed over competition. Thus, the benefits from pollinator sharing seem to outweigh the costs (i.e. heterospecific deposition and conspecific pollen loss). The largest proportion of facilitated species was found at the highest elevation community, suggesting that under unfavourable conditions for the pollination service and at lower plant densities facilitation can be more common.

Summary1. Most plant-pollinator network studies are conducted at species level, whereas little is known about network patterns at the individual level. In fact, nodes in traditional speciesbased interaction networks are aggregates of individuals establishing the actual links observed in nature. Thus, emergent properties of interaction networks might be the result of mechanisms acting at the individual level. 2. Pollen loads carried by insect flower visitors from two mountain communities were studied to construct pollen-transport networks. For the first time, these community-wide pollentransport networks were downscaled from species-species (sp-sp) to individuals-species (i-sp) in order to explore specialization, network patterns and niche variation at both interacting levels. We used a null model approach to account for network size differences inherent to the downscaling process. Specifically, our objectives were (i) to investigate whether network structure changes with downscaling, (ii) to evaluate the incidence and magnitude of individual specialization in pollen use and (iii) to identify potential ecological factors influencing the observed degree of individual specialization. 3. Network downscaling revealed a high specialization of pollinator individuals, which was masked and unexplored in sp-sp networks. The average number of interactions per node, connectance, interaction diversity and degree of nestedness decreased in i-sp networks, because generalized pollinator species were composed of specialized and idiosyncratic conspecific individuals. An analysis with 21 pollinator species representative of two communities showed that mean individual pollen resource niche was only c. 46% of the total species niche. 4. The degree of individual specialization was associated with inter-and intraspecific overlap in pollen use, and it was higher for abundant than for rare species. Such niche heterogeneity depends on individual differences in foraging behaviour and likely has implications for community dynamics and species stability. 5. Our findings highlight the importance of taking interindividual variation into account when studying higher-order structures such as interaction networks. We argue that exploring individual-based networks will improve our understanding of species-based networks and will enhance the link between network analysis, foraging theory and evolutionary biology.

Aim Interaction networks are being increasingly used to evaluate macroecological patterns. We explored a global dataset to identify differences in the structure of pollination networks from islands (of oceanic and continental origin) and mainlands. For oceanic islands, we further evaluated the effects of key island traits on network structural parameters.Location Fifty-two quantitative plant-pollinator networks from continental islands (n = 23), oceanic islands (n = 18) and mainlands (n = 11) located world-wide. MethodsThe effect of geographical origin upon network structure was explored by means of generalized mixed models, accounting for biogeographical region, sampling intensity, latitude and network size. For oceanic island networks, the influence of area, age, elevation and isolation was also evaluated. ResultsThe structure of pollination networks was fairly consistent between mainland and continental islands and only a few differences were noted. Oceanic island networks, however, were smaller and topologically simplified, showing a lower interaction diversity, and higher plant niche overlap than mainland and continental island networks. Isolation and elevational range of oceanic islands influenced the total number of species and interactions. Networks from higherelevation oceanic islands were less nested and those located towards the equator exhibited higher interaction richness. Island area showed no significant effect on any of the network metrics studied here.Main conclusions Pollination networks appear structurally similar regardless of their geographical origin. However, networks from continental islands are more similar to their mainland counterparts than to those from oceanic islands, probably due to the geological nature of continental islands, which are fragments of the mainland to which they were once connected. Oceanic island networks are the least species-and link-rich, and exhibit the lowest interaction diversity and the highest plant niche overlap, possibly due to lower pollinator richness. The most isolated and low-elevation islands show the simplest networks, and are thus probably the most vulnerable to pollination disruptions.

Coextinction models are useful to understand community robustness to species loss and resilience to disturbances. We simulated pollinator extinctions in pollination networks by using a hybrid model that combined a recently developed stochastic coextinction model (SCM) for plant extinctions and a topological model (TCM) for animal extinctions. Our model accounted for variation in interaction strengths and included empirical estimates of plant dependence on pollinators to set seeds. The stochastic nature of such model allowed us determining plant survival to single (and multiple) extinction events, and identifying which pollinators (keystone species) were more likely to trigger secondary extinctions. Consistently across three different pollinator removal sequences, plant robustness was lower than in a pure TCM, and plant survival was more determined by dependence on the mutualism than by interaction strength. As expected, highly connected and dependent plants were the most sensitive to pollinator loss and collapsed faster in extinction cascades. We predict that the relationship between dependence and plant connectivity is crucial to determine network robustness to interaction loss. Finally, we showed that honeybees and several beetles were keystone species in our communities. This information is of great value to foresee consequences of pollinator losses facing current global change and to identify target species for effective conservation.

Studies on pollination networks have provided valuable information on the number, frequency, distribution and identity of interactions between plants and pollinators. However, little is still known on the functional effect of these interactions on plant reproductive success. Information on the extent to which plants depend on such interactions will help to make more realistic predictions of the potential impacts of disturbances on plant-pollinator networks. Plant functional dependence on pollinators (all interactions pooled) can be estimated by comparing seed set with and without pollinators (i.e. bagging flowers to exclude them). Our main goal in this study was thus to determine whether plant dependence on current insect interactions is related to plant specialization in a pollination network. We studied two networks from different communities, one in a coastal dune and one in a mountain. For ca. 30% of plant species in each community, we obtained the following specialization measures: (i) linkage level (number of interactions), (ii) diversity of interactions, and (iii) closeness centrality (a measure of how much a species is connected to other plants via shared pollinators). Phylogenetically controlled regression analyses revealed that, for the largest and most diverse coastal community, plants highly dependent on pollinators were the most generalists showing the highest number and diversity of interactions as well as occupying central positions in the network. The mountain community, by contrast, did not show such functional relationship, what might be attributable to their lower flower-resource heterogeneity and diversity of interactions. We conclude that plants with a wide array of pollinator interactions tend to be those that are more strongly dependent upon them for seed production and thus might be those more functionally vulnerable to the loss of network interaction, although these outcomes might be context-dependent.