Wright et al. (2017) The Overlooked Role of Facilitation in Biodiversity Experiments PDF

Review The Overlooked Role of Facilitation in Biodiversity Experiments [238_TD$IF]Alexandra J. Wright,1,* David A. Wardle,2,3 Ragan Callaway,4 and Aurora Gaxiola5,6,7 Past research has demonstrated that decreased biodiversity often reduces...

Review The Overlooked Role of Facilitation in Biodiversity Experiments [238_TD$IF]Alexandra J. Wright,1,* David A. Wardle,2,3 Ragan Callaway,4 and Aurora Gaxiola5,6,7 Past research has demonstrated that decreased biodiversity often reduces Trends ecosystem productivity, but variation in the shape of biodiversity–ecosystem Understanding the functional role of function (BEF) relationships begets the need for a deeper mechanistic under- biodiversity in an ecosystem is an standing of what drives these patterns. While mechanisms involving competi- essential component of predicting the consequences of biodiversity loss. tion are often invoked, the role of facilitation is overlooked, or lumped within Experimental studies have consistently several less explicitly defined processes (e.g., complementarity effects). Here, shown that the loss of biodiversity can we explore recent advances in understanding how facilitation affects BEF lead to a loss in ecosystem functioning (BEF relationships). relationships and identify three categories of facilitative mechanisms that can drive variation in those relationships. Species interactions underlying Our ability to predict the conse- quences of biodiversity loss in under- BEF relationships are complex, but the framework we present provides a step studied ecosystems, and in a global toward understanding this complexity and predicting how facilitation contrib- change context, requires a deeper utes to the ecosystem role of biodiversity in a rapidly changing environment. mechanistic understanding of BEF relationships. BEF Experiments and the [24_TD$IF]Consequences of Biodiversity Loss Here, we highlight three categories of facilitation that can be important dri- Current and projected rates of global species loss emphasize the need to more precisely vers of BEF relationships: indirect bio- understand how biodiversity promotes ecosystem function in different types of ecosystem. tic interactions due to pathogens and Over the past 20 years, ecologists have tackled this need via controlled experimental manip- mutualists; abiotic interactions due to ulations of the richness of species, genotypes, and functional groups. These BEF experiments nutrient enrichment; and abiotic inter- actions due to microclimate have been used to assess the role of biodiversity in many different ecological contexts. We amelioration. now know that, when species diversity decreases, ecosystem responses, such as net primary productivity (NPP), stability of NPP, and resistance to species invasion, often also decrease. We demonstrate how increased envir- onmental severity, abundance of spe- cialist pathogens, and biological More recently, work has focused on understanding BEF relationships across systems and nitrogen fixation rates likely drive scales at a mechanistic level [3,4]. This research has primarily examined the roles of so-called increased facilitation and, thus, the niche complementarity and selection effects (see Glossary) to explain BEF relationships [5–7]. strength of the BEF relationship, across ecosystems. Niche complementarity is usually examined as the way in which coexisting species differ in their resource needs and acquisition strategies. Greater overall species diversity leads to increased occupation of niche space (up until some point of saturation), more comprehensive resource use, and increased community-level biomass production (but see [8,9] for examples of other mechanisms that are explored within the calculation of ‘complementarity effects’ as defined by 1 FIT – Science & Mathematics, 227 W ). By contrast, selection effects can occur when higher-diversity communities are more 27th St, New York, NY 10001, USA 2 productive than lower-diversity communities due to the increased probability of including a Swedish University of Agricultural Sciences, Department of Forest particularly productive species that dominates in a mixture [7,10]. Ecology and Management, 901 83 Umea, Sweden 3 Despite the commonly reported positive effect of species diversity on ecosystem functioning, Asian School of the Environment, Nanyang Technological University, 50 there is a great range in the magnitude and shape of the BEF relationship that is not easily Nanyang Avenue, 639798, Singapore explained by resource complementarity or selection effects [11–13]. In several cases, 4 University of Montana, Division of Trends in Ecology & Evolution, May 2017, Vol. 32, No. 5 http://dx.doi.org/10.1016/j.tree.2017.02.011 383 © 2017 Elsevier Ltd. All rights reserved. facilitation has been suggested as a mechanism that drives variation in BEF relationships [14– Biological Sciences,32 Campus Drive, Missoula, MT 59812, USA 19]. Here, we propose that a better understanding of these facilitative mechanisms can help 5 Pontificia Universidad Catolica de explain variation in the shape and magnitude of BEF relationships in different ecological Chile, Department of Ecology, Casilla contexts [20,21]. We develop a conceptual framework for how different types of facilitative 114-D, Santiago, Chile 6 Instituto de Ecología y Biodiversidad, mechanism affect BEF relationships. This framework should also help us to predict how Las Palmeras 3427, Santiago, Chile biodiversity could influence ecosystem functioning in systems where it may be impractical 7 Laboratorio Internacional [241_TD$IF]en Cambio to establish large-scale BEF experiments. Global (LINCGlobal, CSIC-PUC), Pontificia Universidad Católica de Chile, Alameda 340, Santiago, Chile Community and Species-Specific Mechanisms Responsible for BEF BEF research has often used the terms ‘niche complementarity’ and ‘overyielding’ interchange- *Correspondence: ably [22–25]. To reduce confusion, we suggest distinguishing between community overyielding [email protected] (A.J. Wright). and species-specific overyielding. Here, we restrict our discussion to the mechanistic underpinnings of species-specific overyielding. As such, species-specific overyielding is the case where a species grows more in mixture than it does in monoculture, after accounting for differences in proportion of seed planted. Using this definition, there are at least three groups of facilitative mechanism that can explain species-specific overyielding in BEF experi- ments: (i) indirect biotic facilitation; (ii) abiotic facilitation via nutrient enrichment; and (iii) abiotic facilitation via microclimate amelioration. Past work focused strongly on the role of resource partitioning (stronger intraspecific than interspecific competition or interference com- petition) for driving species-specific overyielding and, thus, we direct the reader to that work for a more comprehensive discussion of these processes and their role in driving BEF relationships [9,13,28]. Indirect Biotic Facilitation When species grow in dense conspecific clusters or in conspecific soils, species-specific pathogen loads can increase, which can lead to the decreased success of conspecifics. In the context of BEF experiments, negative density dependence due to species-specific patho- gens is a clear demonstration of facilitation that could explain species-specific overyielding , and the BEF relationship in general [14,29–31]. Specifically, diversity can confer a facilitative effect by diluting the effects of pathogens in higher diversity communities (Figure 1A). In addition, higher diversity communities will accumulate a greater diversity of specialist patho- gens, driving the absolute abundance of any individual specialist pathogens to lower levels, thereby potentially resulting in plant species overyielding in mixtures. Indirect biotic facilitation in BEF experiments can also occur via positive effects of belowground mycorrhizal fungi and rhizobacteria [31,32]. Wagg [243_TD$IF]et al. proposed that higher diversity plant communities may be better at harboring more diverse AMF communities. These higher diversity AMF communities might then help expand the total niche space utilized by the plant commu- nity. This could lead to increased performance of individual species (species-specific over- yielding) in higher diversity mixtures. However, empirical support for this proposed mechanism is still lacking. Indirect biotic facilitation can also be the result of indirect competitive interactions in higher diversity systems. When more than two species interact in a plant community, there is the potential for complex indirect interaction networks. For example, species a might limit species b. If species b is usually a strong competitor and limits the success of species c, we might see an indirect positive interaction between species a and species c. In the context of BEF experiments, this should theoretically be more likely with increasing species diversity (because the probability of indirect interactions increases with an increasing number of species), although the probability of negative interactions could also increase with an increasing number of species. These complex interaction networks can also be amplified by the higher diversity soil biota found in higher diversity plant communities. 384 Trends in Ecology & Evolution, May 2017, Vol. 32, No. 5  Species a Species b Glossary monoculture monoculture Mixture Abiotic facilitation: facilitation that (A) Indirect bioc is mediated through changes in the facilitaon abiotic environment (e.g., vapor pressure deficit, soil porosity, soil moisture, or nutrient enrichment). Biotic facilitation: facilitation that results from the activity of a higher order trophic interaction (e.g., bacterial, rhizobial, or arbuscular (B) Abioc facilitaon– mycorrhizal fungal communities). nutrient enrichment Facilitation: occurs when an increase in the density of species b increases the performance of species a. Resource complementarity: occurs when species have unique (C) Abioc facilitaon– and complementary resource amelioraon requirements that can allow some species to stably coexist; these groups of species can be more productive and capture available resources more comprehensively than any species in monoculture. Selection effects: occurs when Figure 1. Facilitative Mechanisms that Explain Species-Specific Overyielding in Biodiversity–Ecosystem higher diversity mixtures have a Function (BEF) Experiments. There are at least three facilitative mechanisms that can explain species-specific higher statistical probability of overyielding. First, indirect biotic facilitation can occur via diversity effects on species-specific pathogen loads (A), or including particularly productive through indirect competitive interactions increasing the productivity of a species growing in mixture. Here, all pathogens species. When those species that (fungal, bacterial, viral, etc.) are indicated with a drawing of an insect. When a single plant species grows alone in are more productive in monoculture monoculture, it can accumulate species-specific pathogens over time. When these same species grow together in are also better competitors in mixture, the species-specific pathogen load is reduced, and plants can grow more due to overall release from pathogenic mixture, higher diversity communities attack. Second, facilitation of neighbors can result from abiotic effects on nutrient availability. In particular, the legume– can be more productive than lower rhizobia symbiosis can directly increase nitrogen availability for neighboring plants (B). In (B) if species a is leguminous, it diversity communities. can have a positive effect on the growth of species b due to nitrogen inputs into the soil. Third, facilitation can be mediated Species-specific overyielding: the through abiotic effects on microclimate conditions. For example, if species a is sensitive to irradiance or high temperatures, case where an individual species the microclimate effect provided by species b can improve the performance of species a in mixture of those species (C). grows more in mixture than it does in monoculture, after accounting for differences in the proportion of seed planted. For example, corn seed in monoculture might be planted at Abiotic Facilitation: Nutrient Enrichment 100%, while corn seed in a two- species mixture might be planted at By far the most well-discussed form of facilitation in the BEF literature is the direct positive 50%. If corn grows 100 g per unit effects that certain species (e.g., legumes) can have on neighbors due to species effects on area in monoculture, but greater than nutrient availability [36–39]. Importantly, legumes likely contribute to both competitive and 50 g per unit area in a two-species facilitative interactions. Legumes can increase resource partitioning when they uniquely have mixture, this is considered species- specific overyielding. direct access to atmospheric nitrogen, a source of nitrogen that is otherwise not accessible to the plant community (e.g., ). However, here we focus on instances where nitrogen inputs facilitated by legumes increase resource availability for nonlegume neighbors (Figure 1B) a clear indication of interspecific facilitation. We extend the well-documented positive effects that legumes have on nitrogen availability and cascading consequences for nonlegume neighbors [18,36,41] to include several other species interactions that may be common and that should result in similar predictions for species-specific overyielding. Biological nitrogen fixation is a widespread phenomenon that occurs in diverse hosts (e.g., legumes, feather mosses, and woody actinorhizal species) and symbiont taxa (e.g., Rhizobia, Frankia, and cyanobacteria). Past BEF experiments have shown that nonleguminous plants can overyield by up to twofold when growing in the presence of legumes [18,42,43]. Similarly, actinorhizal species (e.g., Alnus spp.) can have positive effects on overyielding of neighbors via nitrogen fixation with Frankia. Feather mosses in boreal ecosystems can also positively Trends in Ecology & Evolution, May 2017, Vol. 32, No. 5 385 affect neighbors via nitrogen fixation with the cyanobacteria that they host. Both legumes (e.g., Lupinus) and nonlegumes (e.g., Buddleja davidii) can also have positive effects on neighbors via enhanced phosphorus mobilization (due to the production of phosphate-mobi- lizing root exudates), which enhances the growth of neighboring species [39,45,46]. These mechanisms should all theoretically increase species-specific overyielding through facilitation and increase the strength of the BEF relationship. Abiotic Facilitation: Microclimate Amelioration Neighboring plants can also benefit each other through amelioration of adverse microclimatic conditions. Plants growing in severe climates are often more limited by physiological strain than by competition with neighbors. In these instances, physiological strain and microclimate amelioration in higher diversity communities can increase overyielding and affect the shape of the BEF relationship. Research using BEF experiments has demonstrated the importance of microclimatic ame- lioration in higher diversity plant communities (Figure 1C). In ecosystems that experience periodic drought stress, increased aboveground biomass in higher diversity experimental plots increases shade, which, in turn, reduces surface drying and increases surface soil moisture. Furthermore, increased shade decreases temperature, increases relative humidity, and decreases vapor pressure deficit around the leaves, particularly on unusually hot and dry days. While this effect is likely partially driven by aboveground biomass effects on shade, it may also be related to a type of sampling effect. Higher diversity communities are more likely to include species that have greater drought tolerance and that can maintain higher stomatal conductance during drought. These species are likely to facilitate others via the cooling effects of evapotranspiration. A strong microclimatic amelioration effect on hot, dry days can lead to reduced water stress for neighboring plants that are less drought tolerant [15,48,49]. While potential temperature amelioration effects in Arctic or alpine systems have yet to be shown in the context of biodiversity experiments, they are likely to operate in similar ways (e.g., the buffering of low temperature extremes in higher diversity or higher biomass communities ). Implications for BEF Relationships Understanding the underlying mechanisms behind BEF relationships will allow us to extend our knowledge of BEF patterns to untested systems. Below, we discuss each of the three types of facilitation in detail, together with conceptual predictions about how these might alter BEF relationships in different types of ecosystem. Indirect Biotic Facilitation across Systems We predict that indirect biotic facilitation (Figure 2A–C) should increase with increasing spe- cialist pathogen load and increasing abundance of species-specific mutualistic associations (although for opposite reasons). Specialist pathogen load can vary with latitude , although the directionality of this response is debated [52,53]; it can also shift with elevation and with different agricultural practices. High specialist pathogen loads should drive most species to perform poorly in monoculture. At higher levels of diversity, dilution effects should universally decrease pathogen pressures and all species should overyield. Conversely, beneficial micro- organism diversity should increase niche space available for stable species coexistence. A low diversity of beneficial microorganisms would lead to reduced diversity of niche space, resulting in low overall species coexistence, and experimental additions of species should not result in large increases in productivity. As the diversity of beneficial microorganisms increases, higher diversity plant communities should be increasingly capable of stable coexistence, which should lead to increased productivity and a steeper BEF relationship (Figure 2A). 386 Trends in Ecology & Evolution, May 2017, Vol. 32, No. 5  Increasing abundance of specialist pathogens and beneficial microorganisms (A) (B) (C) Producvity Producvity Producvity Diversity Diversity Diversity Increasing proporon of species that increase resource availability (e.g. nitrogen-fixers) (D) (E) (F) Pmax Producvity Producvity Producvity Pmax Pmax D50 D50 D50 Diversity Diversity Diversity Increasing environmental severity (G) (H) (I) Producvity Producvity Producvity Diversity Diversity Diversity Figure 2. Illustration of Potential Diversity–Productivity Curves [i.e., Biodiversity–Ecosystem Function (BEF)] across Ecosystems. The figure demonstrates how plant diversity (or species richness) on the X-axis can affect productivity (or biomass production) on the Y-axis. (A–C) demonstrate that an increasing abundance of specialist pathogens (solid line) or beneficial microorganisms (dotted line) could increase the strength of the BEF. These biotic facilitative interactions drive stronger BEF relationships for two different reasons. The solid line shows that, as specialist pathogen load increases, monocultures become more suppressed. This is consistent with the findings of Hendriks et al. , that monoculture suppression occurred in the presence of pathogens but not when pathogens were absent. Conversely, the [237_TD$IF]broken line shows that, as specialist beneficial microorganisms increase, niche space can theoretically increase , and productivity of mixtures might be enhanced. (D–F) demonstrate that an increasing proportion of nitrogen-fixers (or other nutrient-enhancing species) can also increase the magnitude of the BEF relationship (abiotic facilitation via nutrients). More nitrogen-fixers can increase the total size of the nitrogen pool and increase the maximum potential productivity (Pmax) of the system. A greater proportion of nitrogen-fixers would also increase the probability of including a nitrogen-fixer at lower diversity and, therefore, increase the slope of the BEF [e.g., smaller D50 in (F)]. These predictions follow patterns observed due to experimental nutrient enrichments in a recent meta-analysis. (G–I) demonstrate how increasing environmental severity can increase the strength of the BEF, consistent with the stress gradient hypothesis (abiotic facilitation via microclimate). When environmental severity is low, diversity–productivity curves likely saturate as a function of niche space (in line with most past BEF experimental evidence ). As environmental severity increases, average monoculture productivity should be suppressed because an increasingly large number of species cannot survive in monoculture (i.e., classic nurse plant effects in deserts ). As diversity increases, there should be a higher likelihood of including a particularly well-adapted nurse plant species that ameliorates the environment for other species and makes it possible for them to persist (sampling effects). The cumulative effects of higher species richness on microclimate can also improve microclimatic conditions and enhance species-specific overyielding at higher levels of diversity (e.g., ). In the special case where environmental severity reduces the overall resource pool, Pmax could also decrease with increasing environmental severity (not shown here). Abiotic Facilitation via Nutrient Enrichment across Systems When there is a greater abundance of plant species that serve as nutrient enrichers in the species pool, there will be greater abiotic facilitation via nutrient enrichment, which should also increase the strength and slope of the BEF relationship (Figure 2B). This is because these nutrient enrichers should increase the size of the available resource pool, leading to greater maximum productivity of the system (Pmax in Figure 2B). Increased abundance of legumes Trends in Ecology & Evolution, May 2017, Vol. 32, No. 5 387 and other nutrient enrichers in an ecosystem should also increase the probability of adding a Outstanding Questions nutrient enricher through sampling effects. As the proportion of nitrogen-fixers in the species What is the relative importance of facil- pool increases, the probability of a low-diversity system including nitrogen-fixers should also itative interactions versus competitive interactions in driving species-specific increase and, thus, the slope and half-saturation constant of the saturation curve should overyielding and BEF relationships? increase (D50 in Figure 2B). Thus, systems that do not have a high proportion of nitrogen-fixers in their species pools can demonstrate weaker BEF relationships than those that do. What are the mechanisms by which facilitative interactions contribute to biodiversity–productivity relationships Abiotic Facilitation via Microclimate Amelioration across Systems in severe environments (e.g., arid, How environmental severity gradients shape BEF relationships has been explored explicitly in a nutrient depleted, or cold) and how handful of experiments involving bryophytes or algae, and among ecosystems that vary in does this affect species redundancy? disturbance regimes. The results from those experiments have been idiosyncratic. Mulder How might this be important for biodi- versity conservation in severe [243_TD$IF]et al. found that bryophyte richness had more positive effects on productivity when environments? bryophytes were subjected to drought, because bryophytes reduced desiccation by increasing microclimatic humidity. Steudel [243_TD$IF]et al. found that heat and salinity stress had stronger How strongly is the role of facilitation in suppressive effects for lower diversity communities, because species growing in monoculture biodiversity–productivity relationships are more vulnerable to temperature and salinity extremes. driven by indirect biotic interactions? Would BEF relationships be weaker in the absence of certain groups of biota, We hypothesize that the effects of direct abiotic facilitation should increase with increasing such as mutualists or pathogens? environmental severity, in line with the stress gradient hypothesis ([24_TD$IF]Figure 2G–I). Furthermore, results from BEF experiments suggest that the shape of the BEF relationship changes with How strongly do nonlegume nutrient environmental stress. In relatively severe climates where facilitation is theoretically important, enrichers drive biodiversity–productiv- ity relationships in different many species are likely to grow poorly in monoculture [245_TD$IF](Figure 2I). However, when more species ecosystems? are present, the effect of diversity on microclimate should be strong enough to reduce physiological strain and promote the growth of sensitive species (e.g., ). At this level of How do facilitative interactions contrib- microclimate amelioration, there might be an inflection point whereby most species overyield ute to ecosystem resistance and resil- due to either incremental whole-community habitat amelioration (e.g., ), or an increased ience in a global change context? How does facilitation help buffer higher probability of including key facilitator species. In more benign environmental conditions, diversity communities against the most competitive interactions and niche complementarity might be more important than facilitation negative effects of climate change? and, therefore, the BEF might follow a similar saturating relationship, but with less monoculture suppression due to environmental severity [246_TD$IF](Figure 2G). Do facilitative interactions affect decomposition rates, elemental fluxes, and ecosystem stability in contrasting Concluding Remarks and Future Directions ways depending on biodiversity con- The past 20 years of BEF research have illuminated our understanding of the role of biodiversity text? In particular, are there instances in [247_TD$IF]ecosystems worldwide, but many types of ecological system remain poorly tested. Here, where facilitation improves productivity in higher diversity communities while we have outlined the mechanisms responsible for these relationships and the overlooked decreasing some other ecosystem importance of three types of facilitation that can drive these patterns (Figure 1). We have also functions (e.g., decomposition)? introduced a conceptual framework for how and why these three types of facilitation can drive changes in the shape of the BEF relationship across different types of system (Figure 2). This framework can be used to predict how biodiversity might affect ecosystem functioning in systems that have been studied less intensively in the past (see Outstanding Questions). Beyond this, there is a need for future work to focus on how facilitative mechanisms can affect BEF relationships for ecosystem functions other than biomass production. In particular, the relationships between facilitation and decomposition, elemental fluxes, and ecosystem stability are likely to be complex. For example, while indirect biotic facilitation might improve the productivity of higher diversity mixtures, the increased interaction complexity resulting from this facilitation might decrease ecosystem stability, depending on the interaction strength in particular. Finally, understanding the mechanisms that drive BEF relationships will be essential to pre- dicting how biodiversity will be affected by global change phenomena. Global change factors will likely independently affect both competitive and facilitative interactions: drought might, for 388 Trends in Ecology & Evolution, May 2017, Vol. 32, No. 5 example, increase competition for water, but aridity might simultaneously increase the impor- tance of facilitation. While these underlying species interactions are complex, our framework presents a first step toward [248_TD$IF]teasing out this complexity. Acknowledgments A.G. acknowledges support from the [249_TD$IF]Grants ICM-MINECON, P05-002 IEB and CONICYT PFB-0023 (Chile). D.A.W. acknowledges support from a Wallenberg Scholars Award. A.G., D.A.W., and R.C. thank the Universidad Internacional de Andalucía, Baeza-Spain. R.M.C. acknowledges support from the NSF EPSCoR Track-1 EPS-1101342 (INSTEP 3). We thank three anonymous reviewers for helpful comments that served to improve the manuscript. References 1. Hooper, D.U. et al. (2012) A global synthesis reveals biodiversity 21. Steudel, B. et al. (2012) Biodiversity effects on ecosystem func- loss as a major driver of ecosystem change. Nature 486, 105–108 tioning change along environmental stress gradients. Ecol. Lett. 2. Cardinale, B.J. et al. (2009) Effects of biodiversity on the func- 15, 1397–1405 tioning of ecosystems: a summary of 164 experimental manip- 22. Tilman, D. (1999) Diversity and production in European grass- ulations of species richness. Ecology 90, 854–854 lands. Science 286, 1099–1100 3. Wright, A. et al. (2014) Local-scale changes in plant diversity: 23. Roscher, C. et al. (2005) Overyielding in experimental grassland reassessments and implications for biodiversity–ecosystem func- communities – irrespective of species pool or spatial scale. Ecol. tion experiments. ProcPoS 1, e6 Lett. 8, 419–429 4. Ebeling, A. et al. (2014) A trait-based experimental approach to 24. Marquard, E. et al. (2009) Plant species richness and functional understand the mechanisms underlying biodiversity–ecosystem composition drive overyielding in a six-year grassland experiment. functioning relationships. Basic Appl. Ecol. 15, 229–240 Ecology 90, 3290–3302 5. Loreau, M. and Hector, A. (2001) Partitioning selection and com- 25. Ravenek, J.M. et al. (2014) Long-term study of root biomass in a plementarity in biodiversity experiments. Nature 412, 72–76 biodiversity experiment reveals shifts in diversity effects over time. 6. Mueller, K.E. et al. (2013) Root depth distribution and the diver- Oikos 123, 1528–1536 sity-productivity relationship in a long-term grassland experiment. 26. Vandermeer, J. (1981) The interference production principle: an Ecology 94, 787–793 ecological theory for agriculture. Bioscience 31, 361–364 7. Hector, A. et al. (2010) General stabilizing effects of plant diversity 27. DeWit, C.T. (1960) On Competition, Landbouwpublikaties on grassland productivity through population asynchrony and 28. Turnbull, L.A. et al. (2012) Coexistence, niches and biodiversity overyielding. Ecology 91, 2213–2220 effects on ecosystem functioning. Ecol. Lett. 16, 116–127 8. Felten von, S. et al. (2009) Belowground nitrogen partitioning in 29. Hendriks, M. et al. (2013) Independent variations of plant and soil experimental grassland plant communities of varying species mixtures reveal soil feedback effects on plant community over- richness. Ecology 90, 1389–1399 yielding. J. Ecol. 101, 287–297 9. Loreau, M. et al. (2012) Niche and fitness differences relate the 30. Maron, J.L. et al. (2011) Soil fungal pathogens and the relation- maintenance of diversity to ecosystem function: comment. Ecol- ship between plant diversity and productivity. Ecol. Lett. 14, 36– ogy 93, 1482–1487 41 10. Huston, M.A. (1997) Hidden treatments in ecological experi- 31. Eisenhauer, N. et al. (2012) Increasing plant diversity effects on ments: re-evaluating the ecosystem function of biodiversity. productivity with time due to delayed soil biota effects on plants. Oecologia 110, 449–460 Basic Appl. Ecol. 13, 571–578 11. Balvanera, P. et al. (2006) Quantifying the evidence for biodiver- 32. Wagg, C. et al. (2011) Mycorrhizal fungal identity and diversity sity effects on ecosystem functioning and services. Ecol. Lett. 9, relaxes plant–plant competition. Ecology 92, 1303–1313 1146–1156 33. Wagg, C. et al. (2015) Complementarity in both plant and mycor- 12. Marquard, E. et al. (2009) Positive biodiversity-productivity rela- rhizal fungal communities are not necessarily increased by diver- tionship due to increased plant density. J. Ecol. 97, 696–704 sity in the other. J. Ecol. 103, 1233–1244 13. Cardinale, B.J. et al. (2006) Effects of biodiversity on the func- 34. Aschehoug, E.T. and Callaway, R.M. (2015) Diversity increases tioning of trophic groups and ecosystems. Nature 443, 989–992 indirect interactions, attenuates the intensity of competition, and 14. Schnitzer, S. et al. (2011) Soil microbes drive the classic plant promotes coexistence. Am. Nat. 186, 452–459 diversity-productivity pattern. Ecology 92, 296–303 35. Lankau, R.A. et al. (2010) Plant-soil feedbacks contribute to an 15. Caldeira, M.C. et al. (2001) Mechanisms of positive biodiversity- intransitive competitive network that promotes both genetic and production relationships: insights provided by delta13C analysis species diversity. J. Ecol. 99, 176–185 in experimental Mediterranean grassland plots. Ecol. Lett. 4, 439– 36. Spehn, E.M. et al. (2002) The role of legumes as a component of 443 biodiversity in a cross-European study of grassland biomass 16. Bessler, H. et al. (2012) Nitrogen uptake by grassland communi- nitrogen. Oikos 98, 205–218 ties: contribution of N2 fixation, facilitation, complementarity, and 37. Vitousek, P.M. et al. (2013) Biological nitrogen fixation: rates, species dominance. Plant Soil 358, 301–322 patterns and ecological controls in terrestrial ecosystems. Philos. 17. Cardinale, B.J. et al. (2002) Species diversity enhances ecosys- Trans. R. Soc. Lond. B: Biol. Sci. 368, 20130119 tem functioning through interspecific facilitation. Nature 415, 38. DeLuca, T.H. et al. (2008) Ecosystem feedbacks and nitrogen 426–429 fixation in boreal forests. Science 320, 1181–1181 18. Hille Ris Lambers, J. (2004) Mechanisms responsible for the 39. Lambers, H. et al. (2013) How a phosphorus-acquisition strategy positive diversity-productivity relationship in Minnesota grass- based on carboxylate exudation powers the success and agro- lands. Ecol. Lett. 7, 661–668 nomic potential of lupines (Lupinus, Fabaceae). 100, 263–288. 19. Brooker, R.W. et al. (2016) Facilitation and sustainable agricul- 40. Fargione, J. et al. (2007) From selection to complementarity: shifts ture: a mechanistic approach to reconciling crop production and in the causes of biodiversity–productivity relationships in a long- conservation. Funct. Ecol. 30, 98–107 term biodiversity experiment. Proc. R. Soc. B: Biol. Sci. 274, 871– 20. Mulder, C. et al. (2001) Physical stress and diversity-productivity 876 relationships: the role of positive interactions. Proc. Natl. Acad. Sci. U. S. A. 98, 6704 Trends in Ecology & Evolution, May 2017, Vol. 32, No. 5 389 41. Schmidtke, A. et al. (2010) Plant community diversity and com- 51. Abdala-Roberts, L. (2016) Test of biotic and abiotic correlates of position affect individual plant performance. Oecologia 164, 665– latitudinal variation in defences in the perennial herb Ruellia nudi- 677 flora. J. Ecol. 104, 580–590 42. Craine, J. et al. (2003) The role of plant species in biomass 52. Schemske, D.W. et al. (2009) Is there a latitudinal gradient in the production and response to elevated CO2 and N. Ecol. Lett. importance of biotic interactions? Annu. Rev. Ecol. Evol. Syst. 40, 6, 623–625 245–269 43. Tilman, D. et al. (2001) Diversity and productivity in a long-term 53. Moles, A.T. and Ollerton, J. (2016) Is the notion that species grassland experiment. Science 294, 843–845 interactions are stronger and more specialized in the tropics a 44. Wall, L.G. (2000) The actinorhizal symbiosis. J. Plant Growth zombie idea? Biotropica 48, 141–145 Regul. 19, 167 54. Xu, M. and Yu, S. (2014) Elevational variation in density depen- 45. Li, L. et al. (2007) Diversity enhances agricultural productivity via dence in a subtropical forest. Ecol. Evol. 4, 2823–2833 rhizosphere phosphorus facilitation on phosphorus-deficient 55. Lin, B.B. (2011) Resilience in agriculture through crop diversifica- soils. Proc. Natl. Acad. Sci. U. S. A. 104, 11192–11196 tion: adaptive management for environmental change. Biosci- 46. Bellingham, P.J. et al. (2005) Contrasting impacts of a native and ence 61, 183–193 an invasive exotic shrub on flood-plain succession. J. Veg. Sci. 56. Fridley, J. (2002) Resource availability dominates and alters the 16, 135–142 relationship between species diversity and ecosystem productiv- 47. Bertness, M. and Callaway, R. (1994) Positive interactions in ity in experimental plant communities. Oecologia 132, 271–277 communities. Trends Ecol. Evol. 9, 191–193 57. Wardle, D.A. and Zackrisson, O. (2005) Effects of species and 48. Wright, A. et al. (2015) Daily environmental conditions determine functional group loss on island ecosystem properties. Nature the competition–facilitation balance for plant water status. J. Ecol. 435, 806–810 103, 648–656 58. Bruno, J. et al. (2003) Inclusion of facilitation into ecological 49. Wright, A. et al. (2014) Living close to your neighbors-the impor- theory. Trends Ecol. Evol. 18, 119–125 tance of both competition and facilitation in plant communities. 59. Rooney, N. and McCann, K.S. (2012) Integrating food web diver- Ecology 95, 2213–2223 sity, structure and stability. Trends Ecol. Evol. 27, 40–46 50. Cavieres, L.A. et al. (2007) Microclimatic modifications of cushion 60. Craven, D. et al. (2016) Plant diversity effects on grassland plants and their consequences for seedling survival of native and productivity are robust to both nutrient enrichment and drought. non-native herbaceous species in the high Andes of central Chile. Philos. Trans R. Soc. Lond. B Biol. Sci. 371, 20150277 Arct. Antarct. Alpine Res. 39, 229–236 390 Trends in Ecology & Evolution, May 2017, Vol. 32, No. 5