New Results

Niche complementarity among pollinators increases community-level plant reproductive success

Ainhoa Magrach, Francisco P. Molina, Ignasi Bartomeus
doi: https://doi.org/10.1101/629931

Abstract

Our understanding of how the structure of species interactions shapes natural communities has increased, particularly regarding plant-pollinator interactions. However, research linking pollinator diversity to reproductive success has focused on pairwise plant-pollinator interactions, largely overlooking community-level dynamics. Here, we present one of the first empirical studies linking pollinator visitation to plant reproduction from a community-wide perspective. We use a well-replicated dataset encompassing 16 plant-pollinator networks and data on reproductive success for 19 plant species from Mediterranean shrub ecosystems. We find that statistical models including simple visitation metrics are sufficient to explain the variability observed. However, a mechanistic understanding of how pollinator diversity affects reproductive success requires additional information on network structure. Specifically, we find positive effects of increasing complementarity in the plant species visited by different pollinators on plant reproductive success. Hence, maintaining communities with a diversity of species but also of functions is paramount to preserving plant diversity.

Introduction

Pollinators provide key services to plants by facilitating pollen flow (Garibaldi et al 2013). Declining trends for some pollinator species in some regions (Potts et al. 2010, Bartomeus et al. 2019) have led researchers to focus on the functional impacts of these changes in pollinator diversity, especially for plant reproductive success (Biesmeijer et al. 2006).

Many studies have evaluated reproductive success on individual plant species (Albrecht et al. 2012; Thomson 2019), and used relatively simple visitation metrics (e.g., the number of pollinator species visiting a plant or the number of visits they perform) to explain the differences observed (e.g., Bommarco et al 2012). Contrastingly, community-level analyses remain scarce (Bennett et al. 2018). Yet plants and pollinators do not interact in isolation but are embedded within larger networks of interactions encompassing other plant and pollinator species (Memmot et al 2004). We are thus missing an important part of the picture, including direct interactions between the whole ensemble of plants and pollinators, but also indirect ones between species within one guild (e.g., plants) through their shared resources (Pauw 2013; Carvalheiro et al. 2014; Lázaro et al. 2014; Mayfield & Stouffer 2017; Johnson & Bronstein 2019). Understanding how changes in pollinator diversity and community structure affect ecosystem functioning is thus a major challenge that requires attention.

The few studies that have analyzed the effects of pollinator diversity on reproductive success at the community level have mainly used experimental setups. As an example, a study that experimentally recreated a plant community with 9 plant species and differing levels of pollinator diversity, found a positive effect of pollinator species diversity on seed set, but also an important effect of niche complementarity between pollinators, a measure of community structure (Fründ et al. 2013). These findings show that not only the diversity of species present, but also the diversity of roles they play and thus the way in which a community is structured are determinant factors of ecosystem functions.

Indeed, theoretical research has long suggested that the structure of multitrophic communities has an effect for ecosystem functioning (reviewed in Thompson et al. 2012). This line of research, rooted in niche theory and revamped by food-web studies (MacArthur & Levins 1967; May & MacArthur 1972, Tilman 1982, Godoy et al 2018), has greatly advanced theory, but the relationship between structure and function has seldom been tested using empirical data (but see Poisot et al. 2013, Kaiser-Bunbury et al 2017, Lazaro et al 2019). Specifically, a major knowledge gap resides in understanding which aspects of structure determine which aspects of function (Thompson et al. 2012). This is because although a network perspective has promised to encapsulate complex ecological mechanisms occurring at the community level – such as indirect interactions (Holt 1977, Abrams et al 1998) or niche overlap (Woodward & Hildrew 2002)- less attention has been given to the ways in which these mechanisms relate to observed ecosystem processes (Blüthgen 2010). We are now at a point where we understand some of the emergent patterns characterizing mutualistic interaction networks at the community level, especially in the case of pollination (Bascompte & Jordano 2007). Amongst them is the prevalence of nested structures, i.e., arrangements where specialist species interact with a subset of the species that generalists interact with (Bascompte et al 2003). Further, plant-pollinator interaction networks seem to exhibit a relatively high extent of complementary specialization at the community scale, which may be directly related to key ecosystem functions (Blüthgen & Klein 2011). However, the mechanisms by which these attributes affect plant reproduction remain to be understood (Winfree 2013). The time is thus ripe to explore the relationship between community structure and ecosystem functioning empirically, with special emphasis on the underlying ecological mechanisms that drive these relationships.

Here, we present an empirical study linking pollinator visitation and plant reproductive success at the community level. We use a well-replicated dataset encompassing plant-pollinator interaction networks collected at 16 sites coupled with data on the reproductive success of 19 plant species recorded in Mediterranean shrub ecosystems. Our study focuses on understanding whether adding information on selected interaction network structure indices to previously used simple visitation metrics (e.g., the number and diversity of pollinator species visiting a plant species) aids in better explaining the differences observed in communitywide reproductive success. In doing so, we conducted our analyses focusing on reproductive success at two different levels: (i) at the species level by considering the association between the position of a focal species within the larger network and its link to individual reproductive success, and (ii) at the site level, by evaluating how attributes that describe the whole site might affect average values of reproductive success for all species measured within one particular site. Specifically, our study focuses on how the interplay between the complementarity in plant species visited by different pollinators, and the redundancy in this function relate to reproductive success. Plant reproductive success requires the delivery of conspecific pollen and thus of a certain degree of niche complementarity (Blüthgen & Klein 2011). Yet, greater values of redundancy in species functions (e.g., that provided by nested structures), are thought to promote species diversity (Bastolla et al. 2009) and stability (Thébault & Fontaine 2010) within plant-pollinator networks. At present, we do not know how either of these network characteristics affects the functions performed by pollinators.

Our results suggest that models including information on simple visitation metrics alone are able to explain differences in reproductive success. However, a mechanistic understanding requires additional information on network structure, notably information on the complementarity between the niches occupied by different pollinator species. Specifically, we find a positive effect of increasing niche complementarity between pollinators on plant reproductive success.

Methods

Plant pollinator interactions

Our study was conducted in SW Spain within the area of influence of Doñana National Park (Fig. S1). Sites were located within similar elevations (ranging from 50 to 150 m a.s.l.), similar habitat and soil types, and presented similar plant composition (plant mean Sørensen beta-diversity among sites = 0.41), reducing potential confounding factors. We surveyed 16 Mediterranean woodland patches with an average distance of 7 km between them (min= 3 km, max= 46.5 km). Each site was surveyed every two weeks for a total of 7 times during the flowering season of 2015 (from February to May) following a 100-m x 2 m transect for 30 mins. Along each transect, we identified all plant species and recorded all the floral visitors that landed on their flowers and touched the plant’s reproductive parts. Only floral visitors (from now on referred to as pollinators) that could not be identified in the field were captured, stored and identified in the laboratory by FPM and another expert entomologist (see acknowledgements). All surveys were done under similar weather conditions, avoiding windy or rainy days, during mornings and afternoons with the sampling order being established randomly. Within each transect every 10 m we surveyed a 2×2 m quadrant where the number of flowers per species were counted, i.e., 10 quadrats per transect which makes 40m2 of area surveyed overall.

Plant reproductive success

Within each site, we marked 3-12 individuals (mean ± SD: 6.49 ± 2.37) belonging to 1-6 plant species (mean ± SD: 4.06 ± 1.69, Table S2). For each individual, at the end of the season, we recorded fruit set (i.e. the proportion of flowers that set fruit), the average number of seeds per fruit and the average fruit and seed weight per fruit (1-36 fruits subsampled; mean ± SD: 11.17 ± 6.85, Table S3). These last two variables show a strong correlation (Pearson correlation= 0.89), and thus we only present results on fruit weight. Our survey included a total of 19 different totally or partially self-incompatible plant species that depend on pollinators to maximize their reproduction (Table S4) across our 16 sites. All plant species were common and widespread shrubs. Individuals were selected depending on the presence of flowers during the sampling events. We also calculated the average reproductive success at the site level by averaging values of reproductive success obtained for each species.

Data analyses

To evaluate the sampling completeness, we estimated the asymptotic number of species of plants, pollinators and interactions present (Chao et al. 2009), a non-parametric estimator of species richness for abundance data. This estimator includes non-detected species and allowed us to calculate the proportion detected with our original data. We used Chao 1 asymptotic species richness estimators (Chao et al. 2009) and estimated the richness of pollinators, plants and plant–pollinator links accumulated as sampling effort increased up to 100% sampling coverage using package iNEXT (Hsieh et al. 2016) within the R environment (R Development Core Team 2011). We then extracted the values covered by our sampling.

To evaluate differences in network structure between communities, we constructed plant-pollinator interaction networks by pooling the data for the 7 rounds of sampling. We thus obtained one interaction network per site, representing the number of individuals of different pollinator species recorded visiting each different plant species. For each network, we extracted a series of relevant network metrics at the species and site levels.

Additionally, we checked for spatial autocorrelation in our data using Mantel correlograms. Autocorrelation values were non-significant for all variables, except for pollinator richness where we have a small but significant effect at small spatial scales (Fig. S2). Hence, we treat each site as independent in our analysis.

Species-level network analysis

At the species level, we focused on attributes defining the position of a focal plant species within the larger community. As such, we considered two metrics providing complementary non-redundant information: (i) average niche overlap in terms of pollinators between a focal plant species and each of the other plant species in the community, and (ii) the contribution to nestedness of each individual plant species. Niche overlap estimates the potential indirect interactions between plant species through shared resources (in this case pollinators) and the potential for increased heterospecific pollen deposition (Arceo-Gómez et al 2019). We calculated it as the average overlap in pollinator species visiting a focal plant and each of the other plants in the community using the Morisita overlap index, a measure of similarity between two sets of data (Zhang 2016). A plant species’ contribution to nestedness is calculated by comparing the nestedness observed in a given community to that generated by randomizing the interactions in which a focal species is involved. Species that show important contributions to overall nestedness will have values >0, while species that do not contribute to overall nestedness wil show values <0 (Saavedra et al 2011).

Site-level network analysis

At the site level, we followed the same logic as the one presented at the species level. We also calculated two network metrics providing complementary non-redundant information. In this case, we focused on nestedness, a measure of the redundancy in the plants visited by different pollinators, and pollinator niche complementarity, a measure of the complementarity in plant species visited by different pollinator species.

Nestedness is the property by which specialists interact with a subset of the species that generalists interact with (Bascompte et al. 2003). Although there is an ongoing debate in the literature (e.g., James et al 2012), some theoretical studies have found that nested networks are more stable and resilient to perturbations because nestedness promotes a greater diversity by minimizing competition among species in a community (Bastolla et al. 2009). However, many network attributes vary with network size and complexity (Blüthgen et al. 2006). In the case of nestedness, we know it can be affected by network size and connectance (Song et al. 2017). An approach that is often used to correct for this are null models, comparing null-model corrected nestedness values across different networks. However, this approach presents the same issues, as z-scores also change with network size and connectance (Song et al. 2017). We thus used a normalized value of the widely used nestedness metric NODF based on binary matrices (Almeida-Neto & Ulrich 2011), NODFC (Song et al. 2017). This normalized value is calculated as N0DFC = NOD Fn/(C * log(S)), where C is connectance and S is network size, calculated as Embedded Image is calculated as NODF/max(NODF), which is independent of network size and thus comparable across different networks (Song et al 2017). To calculate max(NODF) we used a corrected version of the algorithm (Simmons et al 2019) whenever possible. Results did not change qualitatively when using the uncorrected version of the algorithm for all sites as both are highly correlated (Spearman correlation = 0.94).

To calculate niche complementarity, we used a community-level measure defined as the total branch length of a dendrogram based on qualitative differences in visitor assemblages between plants (Petchey & Gaston 2007; Devoto et al. 2012). All network metrics were calculated using package bipartite (Dormann et al. 2009).

Statistical analyses

To evaluate whether adding information on network structure improves our ability to explain differences in reproductive success - both at the species and the site level - we used generalized linear (GLMs) and generalized linear mixed models (GLMMs) respectively. In both cases we fit three types of models: (i) model 0, a null model with no explanatory variables, (ii) model 1, that only included simple visitation metrics and (iii) model 2 that additionally included information on network structure. These models are meant to be additive, so that the network metrics included are intended to complement rather than substitute the simple metrics traditionally used.

At the species level, response variables included fruit set analyzed using a binomial distribution and the average number of seeds per fruit, and the average fruit weight fitted using normal distributions. The number of seeds per fruit was centered and scaled (i.e., we subtracted column means and divided by standard deviation) to allow meaningful comparisons across species with contrasting life histories. As explanatory variables, model 1 included the number of pollinator species observed, and the visitation rate received by each plant species. Visitation rate was calculated as the total number of visits received by a plant species divided by the average number of flowers of that species found in the 10 2×2 m quadrats per transect. In turn, model 2 added the two network attributes calculated at the species level: average plant niche overlap and contribution to nestedness. For both models, we included plant species identity nested within site and site as random effects to account for multiple individuals of the same plant species measured at each site.

At the site level, response variables were the average reproductive success of all plants surveyed within a site (i.e., average fruit set analyzed using a binomial distribution, average number of seeds per fruit and average fruit weight using a normal distribution). We thus had a single value per site and no random effects are needed. Here, model 1 included total pollinator richness and total pollinator abundance (i.e. number of visits received by all plants within the community) as explanatory variables. Model 2, in turn, added information on network structure by including nestedness and pollinator niche complementarity.

Average values of reproductive success at the site level can be driven by a single plant species. Yet, what will determine the persistence of a diverse plant community, is the presence of some sort of “equity” or evenness in reproductive success across the whole community. We therefore calculated the proportion of species with normalized (between 0 and 1) average fruit set values that were above the 50th percentile as a measure of equity. As any selected threshold is arbitrary, we repeated this using the 25th and 75th percentile thresholds (Byrnes et al 2014). We then used the same framework as that used for species and site-level analyses and fit the same models 0, 1 and 2 using equity in reproductive success as response variable and fitting a binomial distribution.

In all cases, we used variance inflation factors to check for collinearity between explanatory variables. Additionally, we ran residual diagnostics to check if model assumptions were met and used the Akaike Information Criterion (AIC) to compare model performance and complexity. Whenever the difference between the AIC of the models was < 2 (ΔAIC < 2), we considered all models equally good (Burnham et al. 2011). In the case of mixed models, for comparison, models were fitted by maximum likelihood and then the best model was refitted using restricted maximum likelihood. All predictor variables were standardized prior to analysis. For every model we also calculate the R2 value, using the approximation suggested for GLMMs when necessary (Nakagawa et al 2017).

Finally, we tested whether the importance of network structure in explaining differences in equity in reproductive success increases with the number of plant species being considered. We expect that when only one plant species is considered the importance of network structure will be negligible, while we expect it to increase as more plant species are considered (up to a maximum number of 6 species which is the maximum we have measured in our study at a particular site).

To test this, we ran a simple simulation in which the number of species considered increased at each step and for each step we re-calculated equity in reproductive success. Instead of drawing plant species randomly for each step, we tested all possible combinations for each plant number level and network, as the number of combinations is small (e.g. for n = 3 plants selected out of 6 there are only 20 possible combinations). Then, we tested if the relationship between equity in reproductive success and niche complementarity (given its importance in determining differences in reproductive success, see Results section) changes as a function of the number of plants considered within our simulated communities. To this end, for each level of species number considered, we randomly selected one of the generated equity values across each of the 16 communities and regressed these 16 values against our network level predictor and extracted the model slope estimates. We repeated this process 1,000 times and averaged all slope estimates. We expect that the more plants considered, the larger the resulting average estimates will be. Note that we only interpret the mean effects, as the variance among different plant number of species considered depends on the initial number of possible combinations.

Results

Within our sampling we recorded 655 plant-pollinator interactions involving 162 pollinator species and 46 plant species (Table S1). Within the pollinator community the distribution of individuals in different orders was: 92.18% Hymenoptera, 5.69% Diptera, 1.29% Coleoptera and 0.63% Lepidoptera.

Our sampling completeness analyses revealed that our survey was able to capture 17-54% of pollinator species (average = 35%), 43-100% of plant species (average = 80%) and 9-32% of plant-pollinator links (average = 20%; Fig. S3). Our values of sampling completeness were slightly smaller in the case of pollinators, probably as a consequence of the great diversity found in the Mediterranean region and within our study area in particular, a hotspot of insect diversity (Nieto et al. 2014).

Species-level analyses

At the species level, in the case of fruit set, our results showed that model 2 had the best fit to our data (lowest AIC value), and fixed effects explained 9% of the variability observed (conditional R2 = 17%). We found a positive relationship between fruit set, pollinator species richness, and a network structure metric, the contribution to nestedness of a focal plant within the overall network (Table 1, Fig. 1, Fig. S4).

View this table:
Table 1.

Results of GLMM showing the association between simple visitation and network structure metrics and species-level fruit. Bold letters indicate variables with large effects (see Figure S4 for estimate confidence intervals).

Figure 1.
Figure 1.

Partial residual plots showing the effect of A) pollinator species richness and B) the contribution to nestedness of each plant species on fruit set. Dots represent each of the individuals sampled for each species within each site.

For the average number of seeds per fruit at the species level as well as for fruit weight, our results showed that none of the models fitted were better than the null model explaining differences across plant species.

Site-level analyses

At the site level, in the case of fruit set and the number of seeds per fruit, we found that both model 1 and 2 were equally good when penalizing for model complexity (i.e.,ΔAIC < 2; Burnham et al 2011). This suggests model 2 was a good model despite its added complexity, and actually showed a substantially better predictive ability than model 1 (R2 = 0.46 for model 2 versus 0.27 for model 1 in the case of fruit set and R2 = 0.49 for model 2 versus 0.35 for model 1 in the case of the number of seeds per fruit) and therefore we will comment results for this model only. Specifically, we found that both fruit set and the number of seeds per fruit were positively related to niche complementarity between pollinators (Table 2, Fig. 2, Fig. S5). Additionally, we found a negative association between site-level pollinator richness and average fruit set (Table 2A, Fig. 2, Fig. S5).

View this table:
Table 2.

Results of GLM showing associations between simple visitation and network structure metrics and A) site-level average fruit set and B) site-level average number of seeds per fruit based on best model selected. Bold letters indicate variables with large effects.

Figure 2.
Figure 2.

Partial residual plots showing the effect of the single predictor which best explains the variability in site-level reproductive success. A) Shows the effect of pollinator richness, and B) of niche complementarity among pollinator species on site-level average fruit set. C) Shows the effect of niche complementarity among pollinator species on the average number of seeds per fruit at the site level. Dots represent average values of fruit set at the level of the community for all plant species considered (N=16 sites).

In the case of fruit weight, we found that both the null model and model 1 were equally good (i.e.,ΔAIC < 2; Burnham et al 2011). Model 1, i.e., that only including simple visitation metrics, showed an R2 of 0.23. In this case, we found a positive link with site-level pollinator richness (Table S5A, Figs. S5-S6). This association was maintained even after removing a site that has a particularly large pollinator richness value (Table S5B, Fig. S7, Fig. S5).

Equity in fruitset

When evaluating the relationship between community composition and network structure on equity in reproductive success across the different species within a community, we found that using the 50th percentile all models were equally good (i.e.,ΔAIC < 2; Burnham et al 2011), but none of the variables considered showed any strong associations (Table S6). In the case of the other two thresholds considered (25th and 75th percentiles) model 0, the null model, was the best model.

Within our simulation evaluating the relationship between niche complementarity and equity in reproductive success at increasing number of plant species considered, we found that the link to complementarity became more important as more species were considered (Fig. 3). This importance seemed to reach a plateau. However, this should be further evaluated, as this was the maximum number of species simultaneously observed in a community for our study, which precludes us from simulating further numbers of species.

Figure 3.
Figure 3.

Results of simulation evaluating the importance of niche complementarity in determining differences in equity in reproductive across communities harboring from one to six species. Points represent average values across 1,000 simulated combinations.

Discussion

The existence of relationships between interaction network structure and ecosystem function have been long hypothesized, yet, the specific mechanisms underlying this relationship remain elusive (Thompson et al. 2012). Our results suggest that different aspects of network structure affect different dimensions of ecosystem functioning. Specifically, we find that the contribution to nestedness of a plant species within a community has a positive association with its fruit set. From a plant’s perspective, this indicates that being connected to other plant species via shared pollinators has a positive outcome (e.g. by ensuring a stable pollinator supply through time) rather than a negative one (e.g. via heterospecific pollen transport). At the site level, we find that greater values of niche complementarity between pollinators result in larger average values of reproductive success.

Most of our analyses reveal that model 1 and 2 were equally good, which suggests that the added complexity of measuring the full network of interactions may not pay off for rapid assessments. Hence, simple visitation metrics, such as pollinator richness, might be enough to describe general patterns (Garibaldi et al. 2013, 2015). Yet, adding network level information may inform us of the potential ecological mechanisms underlying the processes driving the observed patterns.

Consistent with previous experimental (Fontaine et al. 2005; Fründ et al. 2013), theoretical (Pauw 2013), and empirical studies (Valdovinos et al. 2016, Poisot et al. 2013), we find that niche complementarity is key in determining differences in reproductive outputs at the community level, both greater fruit set and larger numbers of seeds per fruit. These results show that reproductive success in plants requires of a certain degree of specialization amongst pollinator species on a particular plant resource in order to avoid the negative effects of inter-specific pollen deposition (e.g., pollen loss, Flanagan et al. 2009) or interference with conspecific pollen (Morales & Traveset 2008). However, we also find that some level of redundancy in these functions is needed as revealed by the positive effect of plant niche overlap on the number of seeds per fruit at the species level.

We did not find a relationship between nestedness and any of the reproductive success measures. This metric, widely used across network analysis, does not seem to play a direct role on plant reproductive success. However, our study is limited to a maximum of six common plant species per community, and including more species, especially rare species, might reveal different patterns. Further, although we sampled each site seven times in a randomized order in an attempt to better represent interactions through time, our surveys were able to capture ~20% of interactions given the great diversity of our study system. This could be explaining part of the low effect sizes we find at the species level, where a stronger contribution of pollinator visits is expected given their obligate dependence (Garibaldi et al 2013). In addition, it is important to note that plant reproductive success is affected by other environmental variables which we do not attempt to measure in this study and that could explain a large portion of the variability observed.

Our measure of reproductive success at the site level using average values represents an important part of the functions delivered by pollinators to plants. However, average values might mask a great deal of variability amongst plant species, and thus a nuanced view of the effect of pollinators on whole-plant ensembles is needed. This can be captured by the effect of pollinators on equity in reproductive success across plant species. This aspect ensures that reproductive success is equally distributed amongst a larger number of species. Indeed, we know that plant species diversity within a community is largely driven by different types of direct and indirect interactions including those amongst plant species (e.g., resource competition, Goldberg & Barton 1992, or facilitation, Bruno et al. 2003), as well as those defining antagonistic (e.g., involving pathogens, Bagchi et al. 2010), or mutualistic interactions (e.g, pollinators, Benadi et al. 2013; Lanuza et al. 2018). However, equitability in reproductive success across species is seldom taken into account, despite increasing theoretical and empirical support to the idea that minimizing fitness differences among species is an important mechanism of species coexistence (Godoy et al 2014). In our case, we did not find a strong effect of either simple visitation or network structure metrics on reproductive equity. However, the results of our simulation, shows us that the effect of network structure increases when more than four plant species are considered. This implies that if we were able to measure reproductive success for all the plant species in all the communities (which is not feasible given constraints on sampling effort), we might find that the effects of network structure on equity might be more prevalent.

One of the unexpected results of our analyses is the strong negative relationship between total pollinator richness and fruit set at the site level. One possible explanation for this is that greater richness means greater transfer of heterospecific pollen (Arceo-Gómez et al 2019). Another possible explanation to this might be the fact that pollinator richness includes all the pollinators recorded during our sampling efforts, i.e., it includes species that do not pollinate some of the species whose reproductive success was measured. More complex communities with more pollinators, but also with more plant species (Pearson correlation between plant and pollinator richness = 0.42 in our case) may require stabilizing mechanisms that reduce the competition exerted by the dominant plant species. A way to reduce the competition exerted by these dominant species, which are precisely those evaluated in this study, is by reducing their reproductive success (Lanuza et al 2018, Stavert et al 2019). These ideas open the door to exploring the positive or negative effects of the complete pollinator community on full plant species coexistence, which may be determined by density-dependence effects (Benadi & Pauw 2018). In our case, while fruit set at the site level is negatively related to pollinator richness, it is important to note that fruit set at the species level and fruit weight show the opposite relationship, indicating that this density-dependent effect might only be limiting fruit quantity and not fruit quality. Thus, taking into account the densities of co-flowering plant species may be the next step (Vanbergen et al. 2014).

Our study illustrates the challenges of measuring and linking network structure to ecosystem function empirically. There is an ongoing debate as to which network metrics better reflect classic ecological mechanisms, such as niche partitioning or competition (Delmas et al 2018). Here, we focus on testing two specific hypotheses, but other structural properties can be explored in the future. Furthermore, the structure of plant-pollinators networks is dynamic due to ecological and evolutionary reasons, but so far, we are only able to characterize it for single snap-shots. Moreover, different aspects of functioning may be important, such as the need to consider the functioning of both trophic levels (Godoy et al 2018). In terms of plant reproductive success and the functions performed by pollinators we can measure different aspects, ranging from pollen deposition (the direct pollinator function), to its final effects on plant fitness. Here, we focus on an intermediate stage including fruit quantity and quality, which is of clear ecological importance.

In summary, our findings show that the analysis of natural communities using network analysis represents an ideal way of visualizing the complexity present within these communities, but also represents a manner of mechanistically representing the differences observed across the reproductive success of individuals and/or species while linking them to potential ecological mechanisms. Our findings represent a step forward in our understanding of how community structure affects function, yet they also show that more studies with better resolved communities are needed, with a special focus being placed in evaluating reproductive success of a larger array of plant species.

Data accessibility

All the data used are available at: https://zenodo.org/account/settings/github/repository/ibartomeus/BeeFunData and the code used to generate all results can be found at: https://doi.org/10.5281/zenodo.3364037

Conflict of interest disclosure

The authors of this preprint declare that they have no financial conflict of interest with the content of this article. AM and IB are recommenders for PCI Ecology.

Statement of authorship

AM led data analysis, FPM collected data, IB obtained funding and supervised data collection and analysis. All authors contributed to writing

Acknowledgements

The authors would like to thank Oscar Aguado for identifying pollinator species. We would also like to acknowledge the efforts of the editor Cédric Gaucherel, as well as those of Michael Lattorff, Nicolas Deguines, Alberto Pascual, Nico Blüthgen, J. M. Bennett and six anonymous reviewers who provided very helpful comments and suggestions in a previous version of this manuscript. AM received funding from a Juan de la Cierva (IJCI-2014-22558) and Ikerbasque fellowships. IB acknowledges funding from MSC-PCIG14-GA-2013-631653 BeeFun Project. We thank Doñana’s Singular Scientific-Technical Infrastructure (ICTS-RBD) for access to the park. Research also supported by the Spanish State Research Agency through María de Maeztu Excellence Unit accreditation (MDM-2017-0714) and the Basque Government BERC Programme.

Version 5 of this preprint has been peer-reviewed and recommended by Peer Community In Ecology (10.1101/629931).

Footnotes

  • Cite as: Ainhoa Magrach, Francisco P. Molina, Ignasi Bartomeus. Niche complementarity among pollinators increases community-level plant reproductive success (2020), bioRxiv, 629931, ver. 7 peer-reviewed by Peer Community in Ecology. 10.1101/629931

  • This article has been peer-reviewed and recommended by Peer Community in Ecology doi: 10.24072/pci.ecology.100037

  • Reviewers: Nicolas Deguines, Michael Lattorff and 3 anonymous reviewers

  • -include conflict of interest disclosure -link to PCI Ecology recommendation

References

  1. ↵
    Abrams, P.A., Holt, R.D. & Roth, J.D. (1998). Apparent Competition or Apparent Mutualism? Shared Predation When Populations Cycle. Ecology, 79, 201–212.
    OpenUrlCrossRefWeb of Science
  2. ↵
    Albrecht, M., Schmid, B., Hautier, Y. & Müller, C.B. (2012). Diverse pollinator communities enhance plant reproductive success. Proceedings of the Royal Society B: Biological Sciences, 279, 4845 LP–4852.
    OpenUrl
  3. ↵
    Almeida-Neto, M. & Ulrich, W. (2011). A straightforward computational approach for measuring nestedness using quantitative matrices. Environ Model Softw, 26, 173–178.
    OpenUrlCrossRef
  4. ↵
    Arceo-Gómez, G., Schroeder, A., Albor, C., Ashman, T.-L., Knight, T.M. & Bennett, J.M. et al. (2019). Global geographic patterns of heterospecific pollen receipt help uncover potential ecological and evolutionary impacts across plant communities worldwide. Scientific Reports, 9, 8086.
    OpenUrl
  5. ↵
    Bagchi, R., Swinfield, T., Gallery, R.E., Lewis, O.T., Gripenberg, S. & Narayan, L. et al. (2010). Testing the Janzen-Connell mechanism: pathogens cause overcompensating density dependence in a tropical tree. Ecology Letters, 13, 1262–1269.
    OpenUrlCrossRefPubMed
  6. ↵
    Bartomeus, I., Stavert, J., Ward, D. & Aguado, O. (2019). Historical collections as a tool for assessing the global pollination crisis. Philosophical Transactions of the Royal Society B: Biological Sciences, 374, 20170389.
    OpenUrlCrossRefPubMed
  7. ↵
    Bascompte, J. & Jordano, P. (2007). Plant-Animal Mutualistic Networks: The Architecture of Biodiversity. Annu Rev Ecol Evol S, 38, 567–593.
    OpenUrl
  8. ↵
    Bascompte, J., Jordano, P., Melián, C.J. & Olesen, J.M. (2003). The nested assembly of plant–animal mutualistic networks. P. Natl. Acad. Sci. USA, 100, 9383–9387.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    Bastolla, U., Fortuna, M.A., Pascual-Garcia, A., Ferrera, A., Luque, B. & Bascompte, J. (2009). The architecture of mutualistic networks minimizes competition and increases biodiversity. Nature, 458, 1018–1020.
    OpenUrlCrossRefPubMedWeb of Science
  10. ↵
    Benadi, G. & Pauw, A. (2018). Frequency dependence of pollinator visitation rates suggests that pollination niches can allow plant species coexistence. Journal of Ecology, 106, 1892–1901.
    OpenUrl
  11. ↵
    Benadi, G., Blüthgen, N., Hovestadt, T. & Poethke, H.-J. (2013). When can plant–pollinator interactions promote plant diversity? Am Nat, 182, 131–146.
    OpenUrlCrossRefPubMed
  12. ↵
    Bennett, J.M., Thompson, A., Goia, I., Feldmann, R., Ştefan, V. & Bogdan, A. et al. (2018). A review of European studies on pollination networks and pollen limitation, and a case study designed to fill in a gap. AoB PLANTS, 10, ply068–ply068.
    OpenUrl
  13. ↵
    Biesmeijer, J.C., Roberts, S.P., Reemer, M., Ohlemuller, R., Edwards, M. & Peeters, T. et al. (2006). Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science, 313, 351–354.
    OpenUrlAbstract/FREE Full Text
  14. ↵
    Blüthgen, N. (2010). Why network analysis is often disconnected from community ecology: A critique and an ecologist’s guide. Basic and Applied Ecology, 11, 185–195.
    OpenUrl
  15. ↵
    Blüthgen, N. & Klein, A.-M. (2011). Functional complementarity and specialisation: The role of biodiversity in plant-pollinator interactions. Basic and Applied Ecology, 12, 282–291.
    OpenUrl
  16. ↵
    Blüthgen, N.N., Menzel, F. & Blüthgen, N.N. (2006). Measuring specialization in species interaction networks. BMC Ecology, 6, 1–12.
    OpenUrl
  17. ↵
    Bommarco, R., Marini, L. & Vaissière, B.E. (2012). Insect pollination enhances seed yield, quality, and market value in oilseed rape. Oecologia, 169, 1025/32.
    OpenUrl
  18. ↵
    Bruno, J.F., Stachowicz, J.J. & Bertness, M.D. (2003). Inclusion of facilitation into ecological theory. Trends in Ecology & Evolution, 18, 119–125.
    OpenUrl
  19. Burnham, K.P. & Anderson, D.L. (2002). Model selection and multi-model inference: A practical information-theoretic approach. Springer, New York, USA.
  20. ↵
    Burnham, K.P., Anderson, D. & Huyvaert, K.P. (2011). AIC model selection and multimodel inference in behavioral ecology: some background, observations, and comparisons. Behav Ecol Sociobiol, 65, 23–35.
    OpenUrlCrossRef
  21. ↵
    Byrnes, J.E.K., Gamfeldt, L., Isbell, F., Lefcheck, J.S., Griffin, J.N. & Hector, A. et al. (2014). Investigating the relationship between biodiversity and ecosystem multifunctionality: challenges and solutions. Methods in Ecology and Evolution, 5, 111–124.
    OpenUrl
  22. ↵
    Carvalheiro, L.G., Biesmeijer, J.C., Benadi, G., Fründ, J., Stang, M. & Bartomeus, I. et al. (2014). The potential for indirect effects between co-flowering plants via shared pollinators depends on resource abundance, accessibility and relatedness. Ecology Letters, 17, 1389–1399.
    OpenUrlCrossRefPubMed
  23. Chacoff, N.P., Vázquez, D.P., Lomáscolo, S.B., Stevani, E.L., Dorado, J. & Padrón, B. (2012). Evaluating sampling completeness in a desert plant-pollinator network. Journal of Animal Ecology, 81, 190–200.
    OpenUrl
  24. ↵
    Chao, A., Colwell, R.K., Lin, C. & Gotelli, N.J. (2009). Sufficient sampling for asymptotic minimum species richness estimators. Ecology, 90, 1125–1133.
    OpenUrlCrossRefPubMedWeb of Science
  25. Delmas, E., Besson, M., Brice, M.-H., Burkle, L.A., Dalla Riva, G.V. & Fortin, M.-J. et al. (2019). Analysing ecological networks of species interactions. Biological Reviews, 94, 16–36.
    OpenUrl
  26. ↵
    Devoto, M., Bailey, S., Craze, P. & Memmott, J. (2012). Understanding and planning ecological restoration of plant–pollinator networks. Ecology Letters, 15, 319–328.
    OpenUrlCrossRefPubMed
  27. ↵
    Dormann, C.F., Frund, J., Bluthgen, N. & Gruber, B. (2009). Indices, graphs and null models: analyzing bipartite ecological networks. The Open Ecology Journal, 2, 7–24.
    OpenUrlCrossRef
  28. ↵
    Flanagan, R.J., Mitchell, R.J., Knutowski, D. & Karron, J.D. (2009). Interspecific pollinator movements reduce pollen deposition and seed production in Mimulus ringens (Phrymaceae). American Journal of Botany, 96, 809–815.
    OpenUrlAbstract/FREE Full Text
  29. ↵
    Fontaine, C., Dajoz, I., Meriguet, J. & Loreau, M. (2005). Functional Diversity of Plant–Pollinator Interaction Webs Enhances the Persistence of Plant Communities. PLOS Biology, 4, e1.
    OpenUrl
  30. ↵
    Fründ, J., Dormann, C.F., Holzschuh, A. & Tscharntke, T. (2013). Bee diversity effects on pollination depend on functional complementarity and niche shifts. Ecology, 94, 2042–2054.
    OpenUrlCrossRefPubMedWeb of Science
  31. Garibaldi, L.A., Carvalheiro, L.G., Leonhardt, S.D., Aizen, M.A., Blaauw, B.R. & Isaacs, R. et al. (2015). From research to action: enhancing crop yield through wild pollinators. Front Ecol Environ, 12, 439–447.
    OpenUrl
  32. ↵
    Garibaldi, L.A., Steffan-Dewenter, I., Winfree, R., Aizen, M.A., Bommarco, R. & Cunningham, S.A. et al. (2013). Wild Pollinators Enhance Fruit Set of Crops Regardless of Honey Bee Abundance. Science, 339, 1608–1611.
    OpenUrlAbstract/FREE Full Text
  33. ↵
    Godoy, O., Bartomeus, I., Rohr, R.P. & Saavedra, S. (2018). Towards the Integration of Niche and Network Theories. Trends in Ecology & Evolution, 33, 287–300.
    OpenUrl
  34. ↵
    Godoy, O., Kraft, N.J.B. & Levine, J.M. (2014). Phylogenetic relatedness and the determinants of competitive outcomes. Ecology Letters, 17, 836–844.
    OpenUrlCrossRefPubMed
  35. ↵
    Goldberg, D.E. & Barton, A.M. (1992). Patterns and Consequences of Interspecific Competition in Natural Communities: A Review of Field Experiments with Plants. The American Naturalist, 139, 771–801.
    OpenUrlCrossRefWeb of Science
  36. ↵
    Holt, R.D. (1977). Predation, apparent competition, and the structure of prey communities. Theoretical Population Biology, 12, 197–229.
    OpenUrlCrossRefPubMedWeb of Science
  37. ↵
    Hsieh, T.C., Ma, K.H. & Chao, A. (2016). iNEXT: iNterpolation and EXTrapolation for species diversity. R package version 2.0.8.
  38. ↵
    James, A., Pitchford, J.W. & Plank, M.J. (2012). Disentangling nestedness from models of ecological complexity. Nature, 487, 227–230.
    OpenUrlCrossRefPubMedWeb of Science
  39. ↵
    Johnson, C.A. & Bronstein, J.L. (2019). Coexistence and competitive exclusion in mutualism. Ecology, 100, e02708.
    OpenUrl
  40. ↵
    Kaiser-Bunbury, C.N., Mougal, J., Whittington, A.E., Valentin, T., Gabriel, R. & Olesen, J.M. et al. (2017). Ecosystem restoration strengthens pollination network resilience and function. Nature, 542, 223.
    OpenUrlCrossRefPubMed
  41. ↵
    Lanuza, J.B., Bartomeus, I. & Godoy, O. (2018). Opposing effects of floral visitors and soil conditions on the determinants of competitive outcomes maintain species diversity in heterogeneous landscapes. Ecology Letters, 21, 865–874.
    OpenUrl
  42. ↵
    Lázaro, A., Gómez-Martínez, C., Alomar, D., González-Estévez, M.A. & Traveset, A. (2019). Linking species-level network metrics to flower traits and plant fitness. Journal of Ecology, n/a.
  43. ↵
    Lázaro, A., Lundgren, R. & Totland, Ø. (2014). Experimental reduction of pollinator visitation modifies plantplant interactions for pollination. Oikos, 123, 1037–1048.
    OpenUrl
  44. ↵
    Macarthur, R. & Levins, R. (1967). The Limiting Similarity, Convergence, and Divergence of Coexisting Species. The American Naturalist, 101, 377–385.
    OpenUrlCrossRefWeb of Science
  45. ↵
    May, R.M. & MacArthur, R.H. (1972). Niche Overlap as a Function of Environmental Variability. Proceedings of the National Academy of Sciences, 69, 1109 LP–1113.
    OpenUrl
  46. ↵
    Mayfield, M.M. & Stouffer, D.B. (2017). Higher-order interactions capture unexplained complexity in diverse communities. Nature Ecology & Evolution, 1, 62.
    OpenUrl
  47. Memmott, J., Waser, N.M. & Price, M.V. (2004). Tolerance of pollination networks to species extinctions. Proc. R. Soc. Lond. B, 271, 2605–2611.
    OpenUrlCrossRefPubMedWeb of Science
  48. ↵
    Morales, C.L. & Traveset, A. (2008). Interspecific Pollen Transfer: Magnitude, Prevalence and Consequences for Plant Fitness. Critical Reviews in Plant Sciences, 27, 221–238.
    OpenUrlCrossRefWeb of Science
  49. ↵
    Nakagawa, S., Johnson, P.C.D. & Schielzeth, H. (2017). The coefficient of determination R2 and intra-class correlation coefficient from generalized linear mixed-effects models revisited and expanded. Journal of The Royal Society Interface, 14, 20170213.
    OpenUrl
  50. ↵
    Nieto, A., Roberts, S., Kemp, J., Rasmont, P., Kuhlmann, M. & García Criado, M. et al. (2014). European Red List of Bees. Publication Office of the European Union, Luxembourg.
  51. ↵
    Pauw, A. (2013). Can pollination niches facilitate plant coexistence? Trends in Ecology & Evolution, 28, 30–37.
    OpenUrl
  52. ↵
    Petchey, O.L. & Gaston, K.J. (2007). Dendrograms and measuring functional diversity. Oikos, 116, 1422–1426.
    OpenUrlCrossRefWeb of Science
  53. ↵
    Poisot, T., Mouquet, N. & Gravel, D. (2013). Trophic complementarity drives the biodiversity–ecosystem functioning relationship in food webs. Ecology Letters, 16, 853–861.
    OpenUrlCrossRefPubMed
  54. ↵
    Potts, S.G., Biesmeijer, J.C., Kremen, C., Neumann, P., Schweiger, O. & Kunin, W.E. (2010). Global pollinator declines: trends, impacts and drivers. Trends Ecol Evol, 25, 345–353.
    OpenUrlCrossRefPubMedWeb of Science
  55. ↵
    R Development Core Team, R. (2011). R: A Language and Environment for Statistical Computing. R foundation for statistical computing.
  56. ↵
    Saavedra, S., Stouffer, D.B., Uzzi, B., and Bascompte, J. (2011) Strong contributors to network persistence are the most vulnerable to extinction. Nature 478, 233–235.
    OpenUrlCrossRefPubMedWeb of Science
  57. ↵
    Simmons, B.I., Hoeppke, C. & Sutherland, W.J. (2019). Beware greedy algorithms. Journal of Animal Ecology, 88, 804–807.
    OpenUrl
  58. ↵
    Song, C., Rohr, R.P. & Saavedra, S. (2017). Why are some plant–pollinator networks more nested than others? Journal of Animal Ecology, 86, 1417–1424.
    OpenUrl
  59. ↵
    Stavert, J.R., Bartomeus, I., Beggs, J.R., Gaskett, A.C. & Pattemore, D.E. (2019). Plant species dominance increases pollination complementarity and plant reproductive function. Ecology, 0, e02749.
    OpenUrl
  60. ↵
    Thébault, E. & Fontaine, C. (2010). Stability of Ecological Communities and the Architecture of Mutualistic and Trophic Networks. Science, 329, 853–856.
    OpenUrlAbstract/FREE Full Text
  61. ↵
    Thompson, R.M., Brose, U., Dunne, J.A., Hall Jr., R.O., Hladyz, S. & Kitching, R.L. et al. (2012). Food webs: reconciling the structure and function of biodiversity. Trends in Ecology & Evolution, 27, 689–697.
    OpenUrl
  62. Thomson, D.M. (n.d.). Effects of long-term variation in pollinator abundance and diversity on reproduction of a generalist plant. Journal of Ecology, 0.
  63. ↵
    Tilman, D. (1982). Resource competition and community structure. Princeton University Press.
  64. ↵
    Valdovinos, F.S., Brosi, B.J., Briggs, H.M., Moisset de Espanés, P., Ramos-Jiliberto, R. & Martinez, N.D. (2016). Niche partitioning due to adaptive foraging reverses effects of nestedness and connectance on pollination network stability. Ecology Letters, 19, 1277–1286.
    OpenUrlCrossRef
  65. ↵
    Vanbergen, A.J., Woodcock, B.A., Gray, A., Grant, F., Telford, A. & Lambdon, P. et al. (2014). Grazing alters insect visitation networks and plant mating systems. Functional Ecology, 28, 178–189.
    OpenUrl
  66. ↵
    Winfree, R. (2013). Global change, biodiversity, and ecosystem services: What can we learn from studies of pollination? Basic and Applied Ecology, 14, 453–460.
    OpenUrl
  67. ↵
    Woodward, G. & Hildrew, A.G. (2002). Body-size determinants of niche overlap and intraguild predation within a complex food web. Journal of Animal Ecology, 71, 1063–1074.
    OpenUrl
  68. ↵
    Zhang, J. (2016). Package spaa: SPecies Association Analysis.
Back to top
Posted June 08, 2020.
Download PDF

Supplementary Material

Data/Code
Email
Niche complementarity among pollinators increases community-level plant reproductive success
Ainhoa Magrach, Francisco P. Molina, Ignasi Bartomeus
bioRxiv 629931; doi: https://doi.org/10.1101/629931
Digg logo Reddit logo Twitter logo Facebook logo Google logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area