Aboveground–belowground interactions as a source of complementarity effects in biodiversity experiments
- First Online:
- Cite this article as:
- Eisenhauer, N. Plant Soil (2012) 351: 1. doi:10.1007/s11104-011-1027-0
- 1.4k Views
The positive relationship between biodiversity and ecosystem functioning (BEF) is due mainly to complementarity between species. Most BEF studies primarily focused on plant interactions; however, plants are embedded in a dense network of multitrophic interactions above and below the ground, which are likely to play a crucial role in BEF relationships.
In the present review I point out the relevance of aboveground–belowground interactions as a source of complementarity effects in grassland biodiversity experiments. A review of the current knowledge on the role of decomposers, arbuscular mycorrhizal fungi, rhizobia, plant growth promoting rhizobacteria, invertebrate ecosystem engineers, herbivores, pathogens and predators in biodiversity experiments, indicates that soil biota can drive both positive and negative complementarity between plant species via a multitude of mechanisms.
I pose four main processes by which aboveground–belowground interactions determine positive complementarity effects: enlarging biotope space, mediating legume effects, increasing plant community resistance, and maintaining plant diversity. By contrast, soil biota may also reinforce negative complementarity effects by competing with plants for nutrients or by exerting herbivore or pathogen pressure, thereby reducing community productivity. Thus, considering aboveground–belowground interactions as well as interactions between antagonistic and mutualistic consumers may improve the mechanistic understanding of complementarity effects in plant diversity–ecosystem functioning experiments and should inspire future research.
KeywordsBiodiversity–ecosystem functioningDecomposersPlant diversity–productivity relationshipSoil feedbackSoil mutualistsSoil pathogens
The rapid loss of species due to human activities and its important implications for ecosystem functioning, services and human well-being have prompted biodiversity research to grow into one of the most active fields in ecological research during the last 20 years (Loreau et al. 2001, 2002; Hooper et al. 2005). Recent meta-analyses showed that biodiversity and ecosystem functioning are predominantly positively linked (Balvanera et al. 2006; Cardinale et al. 2007, 2011). While the phenomenological understanding of the positive relationship between diversity and ecosystem functioning is well established (but see e.g. Huston et al. 2000), current research focuses on the underlying mechanisms (Hector and Bagchi 2007; Fornara and Tilman 2009; Ashton et al. 2010; Maron et al. 2011; Schnitzer et al. 2011). One of the most prominent models explaining why diverse communities perform better than less diverse ones is that different species complement or facilitate each other in chemical, spatial and temporal resource use (Loreau and Hector 2001; McKane et al. 2002; Fox 2005). I briefly outline this concept of species complementarity, its implications for ecosystem functioning and some prominent examples below. A contrasting explanatory approach which I will not discuss in detail in this paper is the selection effect (Huston 1997; Loreau and Hector 2001; Fox 2005). In essence, the performance of high diverse communities may increase due an elevated probability of containing and becoming dominated by one or a few species (Huston 1997; Loreau and Hector 2001). In other words, few dominant species alone can provide the same level of function as more diverse communities. I will focus on complementarity effects due to the recent findings that they may be more important than selection effects in the long term (Cardinale et al. 2007; Fargione et al. 2007; Isbell et al. 2009a; Marquard et al. 2009a; Van Ruijven and Berendse 2009).
The investigation of the BEF relationship and the scientific appreciation of the relevance of species complementarity in this context have increased exponentially in the last 15 years (Balvanera et al. 2006; Fig. 1). Grassland studies have played a dominant role (Fig. 1), primarily highlighting the significance of positive species interactions. The present paper focuses on the findings and mechanisms reported in studies in terrestrial ecosystems (mostly grasslands), in which plant diversity represents the manipulated variable influencing above–belowground interactions. Yet, complementarity between taxa has also been shown in aquatic and bacterial systems (e.g., Bell et al. 2005; Bruno et al. 2005; Stachowicz et al. 2008; Langenheder et al. 2010; Jousset et al. 2011).
List of processes driven by soil biota playing a role in positive plant complementarity effects (see also Fig. 2). Given are varying functional groups of soil biota (in alphabetical order) and the mechanisms they govern. This table is not exhaustive, though intends to give an overview of the multitude of mechanisms representing the source of complementarity effects. See text for details and references. Most mechanisms have not been investigated in plant diversity experiments so far and may stimulate future research
Soil functional group
Enlargement of biotope space
Allocation of varying N forms
Allocation of varying P forms
Exploitation of soil profile by roots
Allocation of varying P forms
Exploitation of soil profile by roots
Fixation and allocation of N2
Mediating legume effects
Litter breakdown and nutrient mineralization
Transport of N in hyphae
Leakage of N into the soil
Plant community resistance
Increase of plant defence compounds
Increase of plant defence compounds
Plant growth promoting rhizobacteria
Reduction of pathogen pressure
Reduction of herbivory
Increase of plant defence compounds
Maintaining plant diversity
Increase of spatial heterogeneity
Promotion of sub-dominant species
Promotion of sub-dominant species
The concept of complementarity
The mathematical background for the differentiation of selection and complementarity effects is given in Loreau and Hector (2001). Considering their major importance and increasing relevance with experimental duration (Cardinale et al. 2007; Marquard et al. 2009a), I will only examine complementarity effects, encompassing resource use complementarity and facilitation. Yet, aboveground–belowground interactions may also promote positive and negative selection effects by favouring only single plant species (Mulder et al. 1999; Bever et al. 2010). In contrast to complementarity effects, selection effects however tend to decrease over time in biodiversity studies (Cardinale et al. 2007; Isbell et al. 2009b).
Aboveground and belowground compartments of terrestrial ecosystems have traditionally been studied in isolation from one another (Wardle et al. 2004), hampering a holistic understanding of ecosystem functioning. In the last decade, soil ecologists in particular worked against this limitation by pointing out the prime relevance of aboveground–belowground interactions for ecosystem functioning (Scheu 2001; Van der Putten et al. 2001; Wardle 2002; Bardgett and Wardle 2003, 2010; Wardle et al. 2004; Bardgett 2005). One view to emerge from this work is that both plants and generalist predators function as essential integrators between the above- and belowground subsystems (Scheu 2001). Plants provide organic carbon inputs to the decomposer subsystem, as well as the food source of belowground herbivores, pathogens and mutualists (Scheu 2001; Wardle et al. 2004). In turn, decomposers affect plant performance via their essential role in organic matter breakdown and nutrient recycling (indirect pathway; Wardle et al. 2004). This action has been shown to cascade to herbivores and higher aboveground trophic levels (Scheu et al. 1999; Wurst and Jones 2003; Eisenhauer and Scheu 2008a; Eisenhauer et al. 2010a; Wurst 2010). Soil biota exert also direct effects on plants by feeding on roots and forming antagonistic (herbivores, pathogens) or mutualistic (mycorrhizal fungi, plant growth promoting rhizobacteria, rhizobia) associations with plants (direct pathway; Wardle et al. 2004). In addition to plants, generalist invertebrate predators, such as spiders, staphylinid and carabid beetles, link the aboveground and belowground subsystems by feeding on both aboveground and belowground prey (Scheu 2001).
Both direct and indirect pathways, as well as interactions between them, influence plant community dynamics and composition (Van der Putten et al. 1993; Klironomos 2002; Wurst et al. 2008; Bardgett and Wardle 2010). In the following, I will shortly summarize how plant diversity affects soil biota, as well as how diversity of soil biota feeds back to plant performance. Both topics need further attention, as plant diversity effects on soil biota are likely to have been underrated due to the predominance of non-representative short-term studies (Eisenhauer et al. 2010b, 2011a), and since soil biota diversity feedback effects to plants are mainly based on laboratory experiments with only a few taxa and ecosystem functions investigated (Wolters 2001). I focus on soil biota since most of the examples given below refer to soil feedback effects on plants.
Plant diversity effects on soil biota
The relationship between plant diversity and the diversity of soil organisms remains controversial. Bardgett and Wardle (2010) conclude that plant diversity exerts weak or non-existent effects on decomposers considering the results of >30 studies, whereas recent long-term studies found a significant positive relationship between plant diversity and soil herbivore (Viketoft et al. 2009; Scherber et al. 2010a; Eisenhauer et al. 2011a, b) and decomposer density and diversity (Scherber et al. 2010a; Eisenhauer et al. 2011a; Sabais et al. 2011). Given the close relationship between increasing plant species richness effects on plant productivity and significant plant diversity effects on many groups of soil biota, though delayed (Eisenhauer et al. 2011a), I will focus on positive plant diversity effects on soil biota. The underlying mechanisms are likely to encompass microhabitat diversity and the quantity, quality and diversity of litter materials and root-derived resources (Hooper et al. 2000; Wardle and Van der Putten 2002; Wardle et al. 2004; De Deyn and Van der Putten 2005). Moreover, plant diversity has been suggested to stabilize the populations of soil organisms by increasing the reliability and consistency of plant derived belowground inputs (Milcu et al. 2010).
Hooper et al. (2000) suggested a step-by-step hypothesis how the diversity of primary producers results in higher belowground diversity assuming strong bottom-up control of biodiversity in soil communities. Briefly, increased diversity of plant derived resources increases the diversity of decomposer microorganisms, detritivores and herbivores in soil, which in turn promotes the diversity of other components of the soil food web. Indeed, Scherber et al. (2010a) recently showed that plant diversity effects cascade from belowground decomposers and herbivores to predators supporting the bottom-up perspective of plant diversity effects on multitrophic interactions.
Soil biota diversity effects on plant performance
One of the most influential, albeit criticized (Wardle 1999), studies highlighting the relevance of soil organism diversity for aboveground functioning is the work by van der Heijden et al. (1998). These authors showed that the diversity of arbuscular mycorrhizal fungi (AMF) drive the diversity, productivity and variability of plants (more details below). Similarly, Smith et al. (2000) showed that different AMF species complement each other by acquiring P close to roots (Scutellospora calospora) and far from roots (Glomus caledonium), suggesting functional complementarity among AMF species, thereby increasing overall P availability to and growth of plants (Koide 2000). In line with this finding, Wagg et al. (2011a) recently reported that AMF species richness significantly increased plant productivity and coexistence in two varying soils. The authors suggested that belowground diversity may act as insurance for maintaining plant productivity under different environmental conditions.
However, there are also less supportive studies (e.g., Vogelsang et al. 2006), and the relevance of soil organism diversity, in particular that of decomposers, for plant performance is disputed. There is evidence that decomposer diversity is crucial for decomposition processes and plant N availability (Bardgett and Cook 1998; Bardgett and Shine 1999; Mikola et al. 2002; Heemsbergen et al. 2004; Tiunov and Scheu 2005), although effects may saturate at low levels of diversity (Laakso and Setälä 1999; Cragg and Bardgett 2001; Bardgett and Wardle 2010). Ecosystem functioning may be mainly affected by the functional differences between decomposers. For instance, Heemsbergen et al. (2004) showed that litter decomposition was driven by the functional dissimilarity between decomposer invertebrates rather than by the number of species per se. Moreover, Eisenhauer et al. (2010a) found non-additive effects of microbial decomposers and invertebrate detritivores synergistically increasing plant and herbivore performance.
Bradford et al. (2002) concluded that positive and negative faunal-mediated effects in soil communities cancel out each other resulting in plant productivity remaining unaffected. Similarly, Wurst et al. (2008) reported that earthworms counterbalanced the negative effects of microorganisms and plant-feeding nematodes on plant community evenness and productivity. By contrast, Wolters (2001) highlighted the relevance of belowground biodiversity for ecosystem functioning by proposing that the number of soil biota species needed to maintain ecosystem functioning may depend on the number of functions investigated. This is in accordance with recent findings on the relevance of plant diversity for ecosystem multifunctionality (Hector and Bagchi 2007; Gamfeldt et al. 2008; Zavaleta et al. 2010; Isbell et al. 2011). Indeed, the latter studies found that the more ecosystem processes were considered, the more species were found to affect overall functioning.
To my knowledge, however, no study has so far experimentally investigated the effect of soil biodiversity on ecosystem multifunctionality. In addition, our current knowledge on the importance of belowground diversity effects on plant performance is almost exclusively based on short-term laboratory experiments, complicating the generalization of weak or missing effects (Cardinale et al. 2007; Eisenhauer et al. 2010b). The increasing number of long-term studies reporting significant soil biota effects on plant productivity suggests that the functional importance of belowground biodiversity may have been underestimated so far. Thus, more research on this topic is needed.
Enlargement of habitat space
Dimitrakopoulos and Schmid (2004) showed that plant diversity effects on primary productivity increase linearly with biotope space (sensu Hutchinson 1978). They defined biotope space as the soil volume which is associated with niche dimension of nutrient acquisition, i.e., species-rich plant communities profited from a large soil volume, allowing for complementary resource use. Recently, biotope space has also been used to express the complexity of resources permitting species complementarity (Jousset et al. 2011). Biotope space can thus be multi-dimensional, i.e., in the case of resource availability it can have a chemical, spatial and temporal dimension (McKane et al. 2002). Noteworthy in this context, multitrophic interactions above and below the ground can influence nutrient cycling and the exploitation of the soil profile, the prerequisites for complementarity between plants.
Decomposer identity and diversity may enlarge biotope space for plant roots by driving organic matter decomposition and nutrient cycling, i.e., the availability of resources in all three dimensions mentioned above (McKane et al. 2002), and by stimulating the exploitation of deeper soil layers and/or resources in a more efficient way.
In a review of approximately 30 papers on litter decomposition, Gartner and Cardon (2004) summarized that litter mixtures mostly experience higher and non-additive decomposition rates than litter of a single species. Recently, Vos et al. (2011) showed that litter diversity effects on decomposition (potentially representing nutrient cycling and availability for plants) are mediated by macro-detritivores. This study investigated litter diversity effects on litter mass loss in the presence/absence of three detritivore species and found that changes due to leaf litter mixing were caused predominantly by complementarity effects, but only in the presence of detritivores. Hättenschwiler and Gasser (2005) and Vos et al. (2011) concluded that the crucial role of macro-detritivores may explain inconsistent findings of previous litter diversity experiments which had been largely ignored. Indeed, macrofauna detritivores have also been shown to play a key role in litter decomposition in grassland (Milcu et al. 2008) and may thus facilitate complementary nutrient use by plants. Moreover, a recent synthesis paper on the influence of detritivore diversity effects on carbon cycling reported significant detritivore species richness effects on decomposition in 100% (richness ≤ 10 species) and 64% (richness > 10 species) of the studies (Nielsen et al. 2011).
Despite the common assumption of functional redundancy of belowground animals and the dominance of generalist feeders in the decomposer community, recent studies point to more specific relationships between plants and decomposers (Bardgett and Wardle 2010). For instance, Bezemer et al. (2010) showed that plant species identity is more important than the surrounding plant community in shaping the composition and structure of decomposer communities of individual plants. Plant species may select for a decomposer community that enhances the decomposition of their own litter (Wardle 2002; Ayres et al. 2009). For instance, the composition and diversity of specific litter compounds, such as polyphenols, can strongly influence soil processes and ecosystem functioning (Hättenschwiler and Vitousek 2000; Hättenschwiler et al. 2005). These findings point to specific associations between single plants and decomposers as well as distinct decomposer feedback effects on plant competition, co-existence and community productivity.
Herbivores and pathogens
Herbivores are assumed to accelerate nutrient cycling and thereby nutrient availability for plant communities at high fertility (McNaughton et al. 1997; Belovsky and Slade 2000; Bardgett and Wardle 2003). Moreover, aboveground and belowground herbivores increase root exudation in grassland plants, which in turn stimulates soil microbial biomass and activity (Holland et al. 1996; Hamilton and Frank 2001) as well as the density of microbivore fauna (Mikola et al. 2001). Increased microbial and faunal activity may result in positive feedback effects to plant productivity due to increased N mobilization (Hamilton and Frank 2001; Bardgett and Wardle 2003). The crucial role of root exudates in stimulating soil microbial activity has been highlighted recently by De Graaff et al. (2010). This study showed that the amount of root exudates entering the rhizosphere is essential for the balance between the stimulation of decomposer microorganisms and decomposition, and compositional shifts towards bacterial-dominated opportunistic communities. The addition of low concentrations of synthetic root exudates increased soil respiration and plant residue decomposition, whereas intermediate concentrations had no effect, and high concentrations even decreased decomposition, with the latter being most likely due to changes in microbial community composition (de Graaff et al. 2010). Since the density and diversity of herbivores often vary with plant diversity (Haddad et al. 2009; Scherber et al. 2010a), herbivory may modulate positive and negative complementarity effects either by stimulating the activity of soil biota or by promoting nutrient immobilization. However, experimental evidence supporting this view is scarce. Previous studies indicate that aboveground herbivores alter plant community composition and attenuate the positive relationship between plant diversity and productivity (Mulder et al. 1999; Scherber et al. 2006, 2010b).
In addition to components of traditional niche theory, soil-borne pathogens may create one essential niche in a multi-dimensional niche space (Hutchinson 1978; Petermann et al. 2008). Petermann et al. (2008) suggested that “pathogen niches”, i.e., pathogen-free space, determined plant coexistence and can explain, at least in part, the positive relationship between plant diversity and productivity (the poor performance of species-poor communities). This means that varying susceptibility of different plant species to soil-borne pathogens as well as “dilution” of host plant individuals in species-rich plant mixtures could play a critical role in the BEF relationship. The role of pathogens will be further developed in the section “Plant community resistance”. A recent meta-analysis suggests that plant–soil feedback effects have medium to large negative effects on plant growth (Kulmatiski et al. 2008), thereby altering plant community composition and probably the balance between positive and negative complementarity effects. However, it should be noted that most knowledge is based on laboratory experiments (Kulmatiski et al. 2008; Bardgett and Wardle 2010); thus, the relevance and interplay of positive and negative soil feedback effects in natural communities needs future attention.
AMF are the most common form of mycorrhizal association in grassland ecosystems, as 80% of herbaceous plants are colonized by AMF (Wang and Qiu 2006). These fungal symbionts build hyphal networks extending the plant root system, enhancing plant nutrient uptake and growth (Smith and Read 1997). Thus, AMF influence productivity and diversity of plant communities, as well as plant competition and aboveground multitrophic interactions (Van der Heijden et al. 1998; Hartnett and Wilson 1999; Klironomos et al. 2000; Wagg et al. 2011a,b). Hyphal length has been shown to significantly increase with the number of AMF taxa, resulting in elevated plant phosphorus (P) content and reduced soil P concentration (Van der Heijden et al. 1998; but see Vogelsang et al. 2006). Thus, increasing AMF diversity resulted in more efficient exploitation of soil P and better use of ecosystem resources. This implies that AMF control P use and thus plant performance in a complementary way, facilitating P use by plant species and enlarging habitat space (Figs. 5, and 6d).
This idea is supported by findings by Smith et al. (2000) showing that different AMF species complement each other in acquiring P, and suggesting functional complementarity among AMF species (Koide 2000). Similarly, Maherali and Klironomos (2007) documented functional complementarity between coexisting AMF taxa, thereby enhancing plant community productivity. Although the study by Klironomos et al. (2000) suggested that AMF increased the functional redundancy of different plant species by decreasing plant species-specific constraints in resource uptake, AMF often seem to enlarge habitat space for plants, thereby promoting plant coexistence and complementarity (Van der Heijden et al. 1998, Wagg et al. 2011a,b). Biomass of AMF often increases with plant species richness (Hedlund et al. 2003; Scherber et al. 2010a), indicating that plants may incrementally promote different AMF species to better access soil P in more complex plant communities.
Predator–prey interactions in the rhizosphere modify nutrient availability, plant productivity and aboveground community dynamics (Moore et al. 2003, Bardgett and Wardle 2010). One prominent example is the microbial loop in soil (Clarholm 1985; Bonkowski 2004). Plants secrete a significant proportion of their assimilates into the rhizosphere via root exudation (Lynch and Whipps 1990) to feed a wide spectrum of mutualists, including N2-fixing microorganisms, AMF and plant growth promoting rhizobacteria (Weller et al. 2002; Dennis et al. 2010). As discussed below, soil microorganisms may represent strong competitors for plants for nutrients in soil. However, nutrients become only temporarily locked up in bacterial biomass since microbivores, such as protozoa and nematodes, successively liberate them by grazing on bacteria with major consequences for nutrient cycling and plant nutrition (Bonkowski 2004). Scherber et al. (2010a) recently showed that the abundance of amoebae strongly increased with plant diversity, suggesting that microfaunal grazing is more pronounced and nutrient cycling accelerated in diverse plant communities. The same relationship has been reported for microbivore nematodes (Eisenhauer et al. 2011c).
Mediating legume effects
Nitrogen is one of the most limiting elements in terrestrial ecosystems (LeBauer and Treseder 2008). Thus, legumes and their association with N fixing bacteria (rhizobia; Fig. 6c) play a crucial role for plant performance and productivity in many terrestrial ecosystems (Mulder et al. 2002; Spehn et al. 2002; Temperton et al. 2007) and for soil biota (Milcu et al. 2008, Viketoft et al. 2009). Fornara and Tilman (2009) suggested that increased plant productivity in species-rich communities relies on seasonal capture of soil nitrate and the accumulation of root N pools in N-limited grasslands, and they proposed seven mechanisms explaining the positive relationship between plant diversity and productivity; six of them are associated with N dynamics. Although Hooper and Dukes (2004) also highlighted the importance of the presence of N-fixing legumes facilitating other non-legume plants, they argued that this was not the only mechanism explaining overyielding of diverse communities. Nevertheless, one of the most prominent examples for facilitative interactions between plant species relies on the symbiotic association between plants and rhizobia. However, legume effects on plant community productivity are also mediated by the action of decomposers, AMF and pathogens, which will be discussed in the following.
Legumes mostly increase the density and diversity of soil biota (Spehn et al.2000; Stephan et al. 2000; Gastine et al. 2003; Milcu et al. 2008; Eisenhauer et al. 2009a; Viketoft et al. 2009). Eisenhauer et al. (2009b) concluded that legumes and earthworms benefit from each other’s presence and form a loose mutualistic relationship. N-rich legume litter and rhizodeposits beneficially affect earthworm biomass (Milcu et al. 2006a, 2008, Eisenhauer et al. 2009a), while legumes benefit from the presence of earthworms through increased decomposition of plant residues and the associated increase in soil N availability (Eisenhauer and Scheu 2008a). Although legumes are able to satisfy most of their N supply through N2 fixation of associated root-nodule bacteria, they also depend on mineralized N in soil (Lee et al. 2003; Eisenhauer and Scheu 2008a). Moreover, N in legume litter becomes available for neighbouring plants only after its recycling by the decomposer community.
In addition to the crucial role in plant P uptake, N transfer from legumes to non-leguminous neighbouring plants has been shown to take place through mycelia of AMF (Haystead et al. 1988), a process potentially facilitating the performance of neighbouring plants and increasing community productivity.
Dromph et al. (2006) showed that the transfer of legume N to neighbouring grass species depended on the density of plant-parasitic nematodes in the rhizosphere. They concluded that increasing infestation of legume roots by plant-parasitic nematodes increases the leakage of N into the rhizosphere and the transfer to neighbouring plants, thereby stimulating their growth. Although primarily discussed in context of plant community succession, the mechanism may also contribute to the often observed key role of legumes for the positive plant diversity–productivity relationship (Mulder et al. 2002; Fornara and Tilman 2009).
Plant community resistance
The coexistence, performance and complementarity of plant species is not only driven by the resource niche, but also by the pathogen niche. Recent studies suggest that species-poor plant communities experience higher pathogen pressure than species-rich ones (see below; Petermann et al. 2008; Maron et al. 2011; Schnitzer et al. 2011). Thus, by attracting different pathogens, diluting host plant species or by being more resistant to pathogen and herbivore pressure, species-rich plant communities may be more productive than species-poor ones.
Decomposers are increasingly recognized to improve plant resistance to herbivores. A recent review on earthworm effects on above- and belowground herbivores reported that influences range from positive to negative (Wurst 2010). Indirect plant-mediated effects are based on altered resource uptake and shifts in microbial community composition, whereas direct earthworm effects on soil nematodes likely are due to the ingestion by earthworms and changes in soil structure. Moreover, Wurst (2010) summarized that earthworms can affect the production of primary and secondary plant metabolites and induce the activity of stress genes. Lohmann et al. (2008) showed that detritivores alter systemic plant responses, i.e., the concentrations of plant defense compounds were significantly affected by the presence of earthworms. Thus, improved nutrient availability due to the action of decomposers may be crucial for the resistance of the plant community against pathogens, herbivory and plant community performance.
Herbivores and pathogens
Herbivores and plant pathogens represent important drivers of plant community dynamics, diversity and composition (van der Putten et al. 1993; Klironomos 2002; De Deyn et al. 2003; Schädler et al. 2004; Weisser and Siemann 2004; Bezemer and van Dam 2005; Allan et al. 2010). Recently, Schnitzer et al. (2011) reported reduced productivity in plant monocultures due to pathogens, whereas productivity in plant mixtures was little affected. Similarly, Maron et al. (2011) showed that the positive relationship between plant diversity and productivity disappeared when the soil was treated with fungicide. The findings that fungicide in particular increased plant production at low plant diversity and little at high plant diversity suggest that soil pathogens may be important drivers of diversity–productivity relationships. Although these findings are very promising and provide a mechanistic explanation for the diversity–productivity relationship, the literature on pathogens and herbivores in BEF studies by no means is consistent. For instance, Bezemer et al. (2004) showed that the number and proportion of plant parasitic nematodes on Cirsium arvense was significantly higher in species-rich than in species-poor plant communities. Nevertheless, community shoot biomass tended to be higher in high diversity than in low diversity plots. Future “home-and-away” experiments, studying plant performance of species growing on their own soils (“home”) and on soils from other plant species (“away”), as well as investigating the soil organisms driving soil feedback effects, may be one promising tool to uncover the relevance of pathogens in plant diversity experiments and the balance between positive and negative soil feedback effects to plant performance (Petermann et al. 2008; Bever et al. 2010).
AMF have been shown to play a crucial role in plant resistance against herbivory (Gehring and Bennett 2009; Koricheva et al. 2009). Recently, Kempel et al. (2010) confirmed earlier work that mycorrhizal colonization increased plant and aboveground herbivore (caterpillar) performance; but after inducing resistance due to short-term herbivory, the beneficial effect of mycorrhizal colonization on plant and herbivore performance disappeared. The authors concluded that plant defence against herbivory depends on AMF due to the allocation of resources into defence compounds, and pointed to the relevance of AMF-induced plant resistance. Similar results were obtained by Sikes et al. (2009) and Vanette and Hunter (2011) for AMF effects on the protection of plants against soil pathogens. Thus, AMF may increase plant community performance due to elevated resistance against herbivores and pathogens above and below the ground. However, the review by Koricheva et al. (2009) revealed that the magnitude and direction of mycorrhizal fungi effects on herbivore performance depends on the feeding mode and diet breadth of the herbivore as well as on the identity of the fungi, complicating the relevance of AMF-induced plant resistance in plant diversity experiments.
Plant growth promoting rhizobacteria
Plant growth promoting rhizobacteria (Fig. 6e) play an essential role for plant performance by inhibiting soil-borne pathogens (Weller et al. 2002; Van der Heijden et al. 2008). Plant diversity beneficially affects the abundance and activity of soil microorganisms (Stephan et al. 2000; Zak et al. 2003; Eisenhauer et al. 2010b), including plant growth promoting rhizobacteria (E. Latz et al., unpublished data; but see Bakker et al. (2010) for inconsistent findings). As typical, the increase in plant growth promoting rhizobacteria caused a pronounced feedback effect to plants by elevating the suppressiveness of the soil against fungal plant pathogens in a biocontrol assay (E. Latz et al., unpublished data). Thus, there may be an indirect positive soil feedback in which species-rich plant communities foster the abundance of soil bacterial populations with biocontrol potential. Given the significance of pathogen effects on the productivity and composition of grassland plant communities (Petermann et al. 2008; Allan et al. 2010; Maron et al. 2011; Schnitzer et al. 2011), this likely prompts plants to cultivate plant growth promoting rhizobacteria. E. Latz et al. (unpublished data) present the increased protection against soil-borne pathogens in species-rich plant communities as novel mechanism contributing to the positive plant diversity–productivity relationship.
Generalist predators perform a crucial ecosystem service as regulators of agricultural pests (Bell et al. 2008). Predation may reduce above- and belowground herbivory, thereby improving plant community performance. Recent studies in arable fields indicate that prey from the belowground system (mostly Collembola) forms a substantial food source for generalist predators, which in turn may suppress plant infestation by aboveground herbivores (Bell et al. 2008; Von Berg et al. 2009, 2010). The authors concluded that biological control by generalist predators can be strengthened by engineering the decomposer subsystem via detrital subsidies. These findings can also have important implications for grassland biodiversity experiments: elevated densities of decomposers in species-rich plant communities (Scherber et al. 2010a; Eisenhauer et al. 2011a,b; Sabais et al. 2011) may increase plant community performance by promoting generalist predators, which in turn reduce herbivore infestation rates. However, the relevance of trophic cascade effects has, to my knowledge, rarely been investigated in plant diversity experiments so far. Haddad et al. (2009) found evidence for the enemies hypothesis that predators control herbivore abundances in plant polycultures. They assumed that this top-down control of herbivores by predators contributes to overyielding of primary productivity in biodiversity experiments. According to Scherber et al. (2010a), however, bottom-up effects seem to be more important than top-down forces.
In addition to affecting plant growth, the association with rhizobia may increase plant resistance against herbivores (Mathesius 2003; Dean et al. 2009). The recent study by Kempel et al. (2009) showed that N provided by rhizobia may be used for the production of N-based plant defence compounds. The authors used nodulating and non-nodulating as well as cyanogenic and acyanoganic strains of Trifolium repens and inoculated these plants with Rhizobium persicae. They showed that mostly positive effects of rhizobia on caterpillar performance did not occur in a cyanogenic legume strain, suggesting that the symbiosis between legumes and rhizobia may be an important driver of herbivore infestation, legume performance and thus plant community productivity. However, and similar to the uncertainty mentioned for AMF effects, the level of plant resistance to herbivores may differ between rhizobia strains (Dean et al. 2009).
Maintaining plant diversity
Theory predicts that diversity is maintained by species interactions that promote overyielding (Vandermeer 1981; Loreau 2004; Isbell et al. 2009a). That means that species overyield when interspecific interactions are less detrimental or more facilitative than intraspecific interactions. As a consequence, there may be less competitive interactions and more complementary interactions in plant mixtures than in monocultures (Isbell et al. 2009a). Thus, biological processes that promote species coexistence and diversity, e.g., that of ecosystem engineers, pathogens and herbivores, and mycorrhizal fungi, may also favour species complementarity.
Ecosystem engineers modify their environment by changing the distribution of materials and energy via non-trophic interactions with abiotic and biotic components of the respective ecosystem (Jones et al. 1994, 1997). These processes affect the distribution, establishment, abundance and diversity of species via the modification, maintenance, formation or destruction of habitats (Jones et al. 1997; Wright et al. 2004). Soil invertebrates, such as ants and earthworms, constitute major ecosystem engineers in many temperate ecosystems (Wright and Jones 2006; Eisenhauer 2010).
Simultaneously manipulating belowground ecosystem engineers and plant diversity, Eisenhauer et al. (2008) found the number and diversity of weed and total plant species to increase with earthworm numbers. This indicates that earthworm middens, i.e., mounts at the soil surface built by anecic earthworms and consisting of litter materials collected on the soil surface and casts, increase the spatial heterogeneity of grassland plant communities (Milcu et al. 2006b; Eisenhauer and Scheu 2008b). Since middens and casts represent nutrient-rich micro-sites with reduced competition between seedlings and the resident plant community (Milcu et al. 2006b) earthworms likely facilitate seedling establishment. Supporting this assumption, Grant (1983) found 70% of the seedlings in temperate grasslands to germinate from earthworm casts, although casts only covered about 25% of the soil surface. Thus, particularly anecic earthworms may increase plant density by stimulating germination and by providing regeneration niches (Grubb 1977). In a grassland biodiversity experiment, Marquard et al. (2009b) found that increased density of overyielding species in mixtures to be the main driver of the positive biodiversity–productivity relationship—a mechanism that may be driven in part by earthworms. In the same experiment, earthworms enhanced plant community productivity (Eisenhauer et al. 2009b). However, experimental evidence for the significance of belowground ecosystem engineers in grassland biodiversity experiments is scarce, coming solely from the investigation of earthworm effects in the Jena Experiment. Although the regeneration niche concept was developed more than 30 years ago (Grubb 1977), more studies are needed to explore the validity of the above-mentioned findings and to investigate the significance of other ecosystem engineers, such as ants and voles.
Herbivores and pathogens
Herbivores and pathogens are likely to change competitive interactions between plants and plant community composition. Herbivores (Collins et al. 1998; Carson and Root 1999; Schädler et al. 2004) and pathogens (van der Putten et al. 1993; Kardol et al. 2005; Petermann et al. 2008) have been reported to decrease the dominance of some plant species, while favouring others. Thereby, they can promote plant diversity (Allan et al. 2010). This is line with the resource concentration hypothesis by Root (1973), assuming that plants experience higher herbivore pressure in species-poor communities, potentially leading to enhanced plant diversity and biomass productivity. However, experimental evidence is scarce and inconclusive. The study by Allan et al. (2010) rather indicates that, although increasing plant diversity, pathogens decrease plant biomass. Mulder et al. (1999) found that the reduction of insect herbivores resulted in greater evenness of relative plant abundances and in a strong positive relationship between plant species richness and shoot biomass. The authors of the latter study assumed that herbivores may rather favour sampling effects and weaken the relationship between diversity and stability of plant productivity.
AMF have been shown to promote plant coexistence (Maherali and Klironomos 2007; Wagg et al. 2011a), relax plant competition (Wagg et al. 2011b), and to maintain plant diversity (van der Heijden et al. 1998) by providing varying resource niches as well as by increasing the use of limiting resources (Klironomos et al. 2000). Specific and complex associations between plants and AMF may thus essentially influence plant community composition and diversity (Kiers et al. 2000; van der Heijden et al. 2008). For instance, AMF diversity has been shown to increase the relative performance of sub-dominant species, thereby increasing overall plant diversity (van der Heijden et al. 1998).
Negative complementarity effects
In contrast to the discussion above, there are also hints that aboveground–belowground interactions can cause negative complementarity effects under certain circumstances. For instance, decomposers may reinforce interspecific plant competition more than intraspecific competition and reduce habitat space by competing with plants for nutrients (Kaye and Hart 1997; Partsch et al. 2006).
Similar to decomposers, effects of AMF on plants form a continuum spanning from beneficial (mututalism) to detrimental (parasitism; Klironomos 2003; Hoeksema et al. 2010). Using meta-analysis Hoeksema et al. (2010) showed mostly positive plant responses to mycorrhizal colonization in P-limited rather than in N-limited systems. In fact, these responses might be interrelated; Wurst et al. (2004) and Eisenhauer et al. (2009c) reported that increased plant P uptake by AMF was associated with reduced N uptake, suggesting competition between AMF and plants for soil N. This assumption was supported by the study by Hodge and Fitter (2010) revealing that AMF take up substantial amounts of soil N from decomposing organic materials and show enhanced productivity in the presence of organic matter patches.
As generally true for mutualistic interactions, the preponderance of beneficial and detrimental effects is context-dependent and varies with nutrient availability, and AMF and plant species (Hoeksema et al. 2010). Thus, interactions between plants and AMF are likely to range from positive to negative complementarity effects. More studies under varying environmental conditions are needed to unravel the context-dependency of positive and negative decomposer and AMF effects on ecosystem functioning and plant species complementarity.
Synthesis, perspectives and conclusions
Positive plant diversity effects have almost exclusively been ascribed to facilitative and complementary interactions between plants so far (Hooper and Dukes 2004; Van Ruijven and Berendse 2005; Marquard et al. 2009a). However, there is increasing awareness that the ecology of plants and the functioning of plant communities can only be understood when the complex network of multitrophic interactions above and below the ground is considered (Klironomos et al. 2000; Worm and Duffy 2003; Bardgett et al. 2008; Bever et al. 2010; Miki et al. 2010). Future research thus should investigate the role of antagonists (herbivores and pathogens) and mutualists (in the broadest sense, including decomposers, AMF, plant growth promoting rhizobacteria, rhizobia) in shaping the relationship between plant diversity and productivity (Fig. 7). Based on the present compilation of results, it is likely that negative effects of antagonists are more pronounced in species-poor plant communities, decreasing their functioning, whereas positive effects of mutualists may dominate in species-rich plant communities and increase their functioning (N. Eisenhauer et al., unpublished data). In particular, there is a lack of information on the balance between antagonists and mutualists in the diversity–functioning relationship. In addition to increasing the phenomenological understanding, future research has to identify the actual multitrophic drivers, e.g., by shedding more light into the soil black box. This information is crucial to realistically predict the consequences of biodiversity loss, since the mostly unnoticed loss of soil biodiversity and/or key belowground species may have stronger impacts on ecosystem functioning and plant complementarity than the loss of single plant species.
The intimate interconnection between the above- and belowground subsystems led Bonkowski (2004) to conclude that “soil, fauna, flora, root, shoot, herbivores and predators in many ways act like a single connected organism, with rhizosphere processes being virtually the basis for understanding plant ecology”. The present review paper presents well-established and hypothetical mechanisms that explain how aboveground–belowground interactions drive or mediate complementarity effects of plants, and may thus inspire future research. The multiple interactions between plants and other biotic ecosystem components are key elements for the mechanistic understanding of complementarity effects in biodiversity experiments (Fig. 2).
I thank Wolfgang W. Weisser for inviting the present review paper to the ESA meeting in Austin, Texas (2011) and for helpful comments on an earlier version of this paper. Comments by Stefan Scheu, Christiane Roscher, Kevin Mueller, Forest Isbell and Cindy Buschena helped to improve this paper. I moreover thank Lois Chaplin, Alexandre Jousset, Stephan König and Claudio Valverde for providing pictures, and Richard Bardgett for help with the illustration of figures. I thank two anonymous reviewers for constructive comments that improved the paper considerably. Further, I gratefully acknowledge funding by the Deutsche Forschungsgemeinschaft (German Research Foundation; Ei 862/1-1).