Positive species interactions in nutrient-poor soils
We observed enhanced growth only when the focal plant Melaleuca grew beside an EMF host (E. marginata) and a nutrient-mining plant with cluster roots (B. menziesii) in a mesh microcosm where root intermingling was not possible in pots with nutrient-poor soils. The enhanced growth found for Melaleuca coincided with a higher P, Mn, Ca, Fe and B content of the shoots. These results suggest a positive fungal mediated interaction mechanism, where enhanced nutrient acquisition is likely and depended on the combination and complementarity of distinct nutrient-acquiring strategies of neighbouring plants. This kind of positive root interaction amongst plants is perhaps an under-appreciated mechanism partly responsible for plant coexistence in plant communities growing in severely nutrient-impoverished soils. Mycorrhizal networks, as an example of a fungal mediated interaction mechanism, have been demonstrated to function differently with different plant assemblies, which corroborates our findings (Walder et al. 2012). Higher P, Mn, Ca, Fe, and B content did not coincide with increased concentrations of these nutrients, most likely because the concentrations of some of these nutrients were severely limiting plant growth, thus any increased uptake would lead to increased growth only (de Groot et al. 2001, 2003).
Positive interactions are typically observed under harsh environmental conditions, and it is referred to as ‘context dependent’ (Brooker et al. 2007; Dickie et al. 2007; van der Heijden & Horton 2009). Here, for the first time, we show that probable growth promotion is also dependent on neighbouring plants with different nutrient-acquisition strategies including a non-mycorrhizal strategy. We also demonstrate that this probable growth-promoting effect is dependent on soil nutrient levels, where positive or negative plant interactions are observed only when nutrients are limiting plant growth. In crop plants, analogous improved mineral nutrition has been shown when one of the two interacting plants fixes N2 when grown in mixtures (Hauggaard-Nielsen & Jensen 2005; Gunes et al. 2007; Li et al. 2007). The possibility therefore exists that as nutrients levels in soils decline, there is an increase in the ability to ‘pack’ more species into a given area via nutrient-sharing arrangements, rather than the principle of ‘winner takes all’ that is a phenomenon of eutrophic systems (Hautier, Niklaus & Hector 2009).
A few root-mediated mechanisms may have been responsible for the enhanced nutrient uptake that we observed. There was likely greater fungal scavenging capability of Melaleuca due to higher and preferential EMF colonization brought upon by the neighbouring EMF host plant Eucalyptus. Melaleuca formed both types of mycorrhizas with either AMF or EMF or both simultaneously, but in the presence of Eucalyptus, Melaleuca was preferentially colonized by EMF, perhaps due to suppressive effects of AMF when grown in the presence of Verticordia (Janos et al. 2013). This more extensively colonized root system presumably allowed Melaleuca to more effectively scavenge the nutrients mobilized by cluster-rooted Banksia (Muler et al. 2014). However, the release of limiting nutrients into soil by neighbouring plants could also have resulted from rhizodeposition or phosphatase activity (Walker et al. 2003; Lynch & Brown 2008), but transfer of nutrients back to the focal plant Melaleuca was ultimately mediated by EMF.
The greater root biomass of Melaleuca when grown with Eucalyptus and Banksia in the nutrient-poor soils further supports a facilitative capacity of the root interaction. Specifically, the mechanism entailed the mobilization of limiting macronutrient P and limiting micronutrient Mn (Muler et al. 2014) by cluster roots of Banksia in the presence of nearby EMF-scavenging hyphae connected to Melaleuca. Similarly, a recent study using Banksia attenuata showed that cluster-root activity promoted the acquisition of nutrients by neighbouring AMF host Scholtzia involucrata (Muler et al. 2014). Our study extends our understanding of enhanced nutrient uptake to plant communities with EMF and suggests a greater positive effect of EM on growth and nutrient uptake compared with AM plants. Thus, we provide evidence to support the claims that EM plants can withstand poorer soils compared with AM plants (Lambers et al. 2008b).
Mycorrhizal networks may also have enhanced nutrient uptake, since the possible growth-promoting effects were preferentially seen when the plants were separated by a root-restricting mesh barrier. In the no-mesh treatment, positive growth effects due to mycorrhizal networks could still have operated, but their magnitudes were likely reduced by root competition. Furthermore, the root intermingling and scavenging capability of non-mycorrhizal Banksia was more prominent in the no-mesh treatment, thereby potentially reducing the occurrence of mycorrhizal networks as a result of more non-mycorrhizal cluster roots dominating the intermingling space. However, hyphal links that could form between Melaleuca and Eucalyptus may have transferred nutrients bidirectionally when grown with Banksia, the non-mycorrhizal species, as the third partner. In this case, the EMF on the Eucalyptus host plant may have scavenged nutrients in the vicinity of Banksia cluster roots, and then connected the roots of the Melaleuca focal plant via mycorrhizal hyphae or rhizomorphs. Transfer of nutrients from Eucalyptus to Melaleuca could then have been promoted by a relatively large sink for nutrients in a larger faster-growing Melaleuca focal plant (data not shown).
When grown with Eucalyptus and Banksia, Melaleuca showed greater shoot P, Mn, Ca, Fe and B contents, possibly explaining the positive growth effect. Enhanced uptake of key limiting nutrients such as P and Mn (Gardner, Barber & Parbery 1983; Gardner & Boundy 1983) may not be the only mechanism explaining the enhanced growth, since shoot concentrations were less consistent than the shoot content data. However, dilution of acquired nutrients certainly occurred, since nutrient levels were only measured on whole shoots that included woody stems, which lowered the concentration values. Furthermore, lack of coincident P and Mn concentrations does not dismiss the evidence against ‘promoted uptake’ since plant biomass gains are not invariably correlated with the concentrations of limiting nutrients (de Groot et al. 2001, 2003). Larger roots of neighbouring plants correlated with larger Melaleuca shoot biomass gains. Had the neighbouring plants been competing for nutrients, the expected effect on Melaleuca would have been the opposite, that is, a decreased nutrient content. This result strengthens the evidence for enhanced nutrient uptake, made possible by the combination of efficient nutrient-mining as well as nutrient-scavenging neighbours in nutrient-poor soils. In accordance, fertilization removed the positive effect of neighbours, as soil nutrient levels were no longer limiting.
Ectomycorrhizal fungi and cluster–root interactions
Mycorrhizal colonization levels and the type of mycorrhiza formed on Melaleuca were influenced by the neighbouring plants. The presence of the obligate AMF host V. nitens favoured the dominance of AMF on the focal plant Melaleuca. Similarly, in the presence of EMF host Eucalyptus, Melaleuca formed more ectomycorrhizas. These results show that neighbouring plants also promote greater colonization levels, likely via mycorrhizal networks (Simard et al. 2012), and influence the type of mycorrhiza that predominates in plant communities as reported in other ecosystems (van der Heijden & Horton 2009). We note that AMF may have suppressive effects on EMF formation as suggested in recent studies with Australian plants (Janos et al. 2013).
Arbuscular mycorrhizal fungi are well known for enhancing P uptake in plants, up to 90% in some cases (van der Heijden, Bardgett & van Straalen 2008). This is one mechanism by which AMF can enhance plant productivity, especially when P is limiting such as in nutrient-poor soils of south-western Australia. However, our results point to enhanced plant productivity mediated by EMF in the presence of a P-mining cluster-rooted plant, most likely because the soils are more severely P impoverished than the nutrient-poor soils where AMF tend to have a positive effect (Lambers & Teste 2013; Lambers et al. In press).
Ectomycorrhizal fungi have traditionally been viewed as key symbionts in maintaining plant productivity in boreal and temperate forests (Smith & Read 2008). These ecosystems are usually poor in N where EMF can scavenge vast amounts of organic compounds in soil using extracellular enzymes (Read & Perez-Moreno 2003). Our study points to the underestimated importance of EMF in P acquisition, as suggested before by others (Cairney 2011). The importance of EMF in uptake of P is noticeable when P is limiting plant growth (Plassard & Dell 2010). We propose that EMF scavenges released P in the vicinity of cluster roots where large amounts of P can be mobilized (Lambers et al. 2012). Thus, our findings provide support for the expectations that possible facilitative root interactions prevail in nutrient-poor soils (Hauggaard-Nielsen & Jensen 2005). However, confirmation through rigorous experiments using tracers and in vivo functional visualization techniques is warranted (Courty et al. 2010).
Root intermingling and competition
Root competition has traditionally been thought to commonly occur when plants grow together in close proximity in soils with nutrient-rich patches (de Kroon, Mommer & Nishiwaki 2003). In nutrient-poor soils, root competition should also be common; however, here a plant's ability to intermingle with roots of neighbouring plants can have net positive effects on growth (Wilson 1989; Lambers & Teste 2013). The greater growth of Melaleuca growing beside Eucalyptus and Banksia coincided with the greatest amount of root intermingling compared only with one of the other plant assemblies only under nutrient-poor soils. We suggest that, in nutrient-poor soils, hotspots of nutrient mobilization and mineralization are found around certain neighbouring plant roots and thus promote root proliferation analogous to nutrient patches. In fertilized soils, root intermingling declined and showed no pattern with neighbour identity. With fertilization, plant diversity and coexistence decreased in plant mixtures where only root competition occurred (Rajaniemi, Allison & Goldberg 2003).
Root competition is predicted to be critical to plant productivity in mixtures especially when light is abundant but nutrients are limiting (Wilson 1988). Positive plant–plant interactions can occur in crop plant assemblies, where one plant species takes up more nutrients than the other two neighbouring plants (Wilson 1989). In this study, light and water were not limiting; thus, it appeared advantageous to explore larger volumes of soil laterally despite the presence of neighbouring plants' root systems. Net positive effects of this nature have been reported in dry relatively nutrient-poor forests where greater access to mycorrhizal networks and roots promoted survival of forest seedlings (Teste et al. 2009).
The observation that the greatest stimulation of Melaleuca growth was found in the presence of mesh barriers between neighbouring plants suggests a positive effect of spatial separation of roots. However, Melaleuca shoot growth was greater, rather than smaller when larger amounts of neighbouring plant roots intermingled (Fig. 7), indicating that sharing rhizospheres led to a positive net benefit. Root proliferation in plant mixtures should follow an evolutionarily stable strategy, where the overall fitness and productivity of the plant is maximized. This strategy entails the proliferation of roots in unoccupied soil first, then intermingling with interspecific neighbours, before exploring soil already occupied by a plant's own roots (Gersani et al. 2001). Root proliferation is typically affected by local nutrient enrichment (Fransen et al. 1999), and our results suggest local enrichment is likely in interspecific rhizosphere soil in nutrient-poor soils. Thus, our study points to simultaneous proliferation of roots in unoccupied soil and in soil occupied by different species.