Community ecology of ectomycorrhizal fungi: an advancing interdisciplinary field


  • Anders Dahlberg

    Corresponding author
    1. Department of Forest Mycology and Pathology, PO Box 7026, Swedish University of Agricultural Sciences, S-750 07 Uppsala, SWEDEN
    Search for more papers by this author

Author for correspondence: Tel: +46 (0) 18 67 27 94 Fax: +46 (0) 18 30 92 45


A long-term goal of community ecology is to identify spatial and temporal factors that underlie observed community structures. Ultimately, ecologists seek to relate community patterns to ecosystem processes and functions. Since the mid 1990s, ectomycorrhizal (ECM) research has been equipped with tools to identify and fully quantify the taxonomic diversity in below-ground ECM fungal communities in detail and address such questions. Many of the most important functions of terrestrial ecosystems, as well as interactions, between plants take place below ground and mycorrhizal fungi are among the key players in soil ecology. Here the rapidly increasing knowledge of ECM fungal community ecology is reviewed and the prospects discussed for elucidating processes that structure ECM fungal communities and the way in which such knowledge might be integrated with, and advance, the understanding of plant ecology and ecosystem processes.


Mutualistic associations between organisms, which reciprocally increase the fitness of both partners, have been a significant driving force in the course of evolution. The importance of endosymbiotic associations between different procaryotes in the evolution of eucaryotic cells is now firmly established and there is also increasing recognition that symbiosis at the level of more complex organisms is the rule, rather than the exception (Smith & Douglas, 1987). An ancient and ecologically important example is the mycorrhizal symbiosis between vascular plants and soil fungi dating back to the first land plants (Blackwell, 2000). However, despite major advances in our understanding of mycorrhizas, the extent and significance of mycorrhizal root symbioses remains greatly underrated in plant ecology. In fact, vascular plants can be regarded as successful entrepreneurs that have benefited from symbiotic association with procaryotes and later fungi to form energy and nutrient transferring biological membrane interfaces, within chloroplasts mitochondria and root-fungal organs, the mycorrhizas. Effective membrane surface area is significantly increased through invagination and external enlargement, and in the case of the mycorrhizal interface, this has critical implications for efficient root uptake of water and nutrients. Through these symbioses plants have succeeded in colonizing all terrestrial biomes (Blackwell, 2000) and consequently, plant-communities are the template for these terrestrial communities of microorganisms. In a strikingly similar manner to the well-being of humans and animals that is intimately linked to community composition of gut bacteria, so also is the functioning of plants closely dependent upon the composition and activities of their below-ground mycorrhizal associates.

World-wide patterns and regulation of plant species richness and their impacts on ecosystem processes have long been of interest, but less is known about the diversity, its regulation and functional importance of mycorrhizal fungi (Molina et al., 1992). While a relatively small number of plants, c. 8000, form ectomycorrhiza (ECM), their global importance is greatly increased by their disproportionate occupancy of terrestrial ecosystems. Many trees, such as those in the Betulaceae, Pinaceae, Fagaceae, and Dipterocaraceae are important and obligate ECM plants, forming vast boreal, temperate and subtropical forests (Smith & Read, 1997). A conservative estimate of the number of ECM fungal species in the world is approaching 6000 (Molina et al., 1992). Beside the diversity of known taxa, there are other ECM fungi which do not produce conspicuous reproductive structures, a feature that precludes their inclusion in known listings, and still further, a huge number of fungal taxa remain to be discovered.

Until the middle of the 1990s, knowledge of ECM fungal community structure, and the degree to which species assemblages show predictable colonization patterns in space and time, was almost exclusively based on sporocarp surveys. A major advantage of such surveys is the potential to record and identify all fruiting ECM fungi within an area with relatively little work load. However, because of the importance of exogenous environmental factors in triggering fruiting, and the consequent irregularity of reproduction, such characterization requires long-term monitoring. Moreover, even after decades of surveying, additional fungal species may still appear. Even more alarmingly, recent demonstrations of the poor correspondence between the species composition of sporocarps and fungi colonizing roots raises some concern about the validity of sporocarp survey data. It was, for example, observed that whereas the most common sporocarps in a jack pine stand were those of Suillus tomentosum, the mycorrhizas identified by distinct morphology rarely constituted more than 5% of the total mycorrhizas (Danielson, 1984).

During the past decade major advances in molecular biology and morphotyping have made it possible to identify mycobionts from single mm-sized mycorrhizas, enabling monitoring and spatiotemporal analysis of below-ground ECM fungal communities (Horton & Bruns, 2001). Thereby, many diversity and functional properties of ECM communities be integrated and their key contributions related to the overall diversity and ecosystem processes, an area of growing interest in terrestrial ecology (Copley, 2000). Examples of such recent break-through achievements are demonstrations of significant net transfer of nutrients and carbon between different plant taxa through common mycorrhizal mycelia (Simard et al., 1997) and the implication that richness and composition of arbuscular mycorrhizal fungi affect diversity and growth of plants (van der Heijden et al., 1998).

Here, I briefly discuss the current potential for terrestrial ecologists to identify and quantify ECM fungal communities. In doing this I provide: an overview of mycorrhizal distributions in soils and methods to identify mycobionts in single mycorrhizas; a description of recent findings in spatiotemporal characteristics of ECM fungal communities; and a discussion of how such an appreciation may advance the understanding of plant ecology and ecosystem processes.

Standing mass and identification of mycorrhizas

Within days of emergence, almost every fine root of an ectomycorrhizal plant is colonized by mycorrhizal fungi (Smith & Read, 1997; Taylor et al., 2000). The number of the mycorrhizal tips, commonly a few mm in length, is often very high; in boreal forests the numbers typically range between 104 and 5 × 105 m−2 forest floor, a figure that in Norway spruce forests roughly corresponds to the number of green needles (Dahlberg et al., 1997). Mycorrhizas are short-lived structures, being regenerated on a yearly basis. Regeneration may occur throughout the growing season, but peaks at certain periods. Mycorrhizas are vertically distributed to depths occupied by host tree roots (i.e. down to several meters). However, the overwhelming majority are located in surface soil layers (e.g. at the interface of organic matter and mineral soil) where mineralization processes are most active. In dry forest ecosystems, mycorrhizas are located deeper in the mineral soil than their mesic counterparts (Visser, 1995; Taylor & Bruns, 1999). Typically up to 90% of the mycorrhizas are located within the upper 10 cm of the soil profile in boreal soils, mainly associated with the thick organic humus or mor layer. Fungal tissue in mycorrhizal roots have fungal biomass that is 10–100 times higher than that in the sporocarps, even without considering the often extensive extramatrical mycelia (Dahlberg et al., 1997).

One well-developed approach to identify ECM fungi colonizing roots involves detailed morphological and structural descriptions of mycorrhizas – morphotyping (Agerer, 1987–1999). In certain cases, morphology of mycorrhizas is species-specific and allows unambiguous identification of mycorrhizas collected in the field (e.g. Smith et al., 2000). More generally, this approach only allows grouping into mycorrhizal morphotypes, which do not represent single species. However, the most significant methodological advance has been the application of high-resolution molecular tools that allow identification of individual mycorrhizas (Gardes & Bruns, 1993; Horton & Bruns, 2001). These methods offer three advantages over morphological methods: they can be directed at any taxonomic level or to detect genotypes; they require less time to learn, and are clearly described; and they are fully reproducible in time and space. Fungal, or plant DNA can be selectively amplified from mixtures of plant and fungal DNA from single mycorrhizas and, with subsequent analysis, species identified (Gardes & Bruns, 1993; Horton & Bruns, 1998). The internal transcribed spacer (ITS) region in the nuclear ribosomal unit is frequently sequenced, digested with restriction enzymes and resulting restriction fragment length polymorphism analysed (Gardes & Bruns, 1996; Kårén et al., 1997; Peter et al., 2000). ITS nucleotide databases are also being constructed for certain ECM fungal groups for identification and phylogenetic purposes allowing for direct sequence comparison; particularly informative is the 5′ end of the LSU gene (Kõljalg et al., 2000; Moncalvo et al., 2000). Sequence data in the ML5/ML6 region of the mitochondrial LrDNA representing all major ECM basidiomycete genera is established (Bruns et al., 1998). Thus, any single mycorrhiza can potentially be identified to species either by ITS-RFLP comparison with reference material or by sequence comparison narrowed down to smaller subsets of putative species affinity that ultimately, with supplementation of appropriate references, will be identified.

Slightly more than 50 such ECM community studies have been published in the period between 1996 and late 2000, few before (cf. Horton & Bruns, 2001; cf. Table 1). Two thirds of these have been based on morphotyping and one third on molecular methods, sometimes with morphotyping-based stratification as a powerful approach (Jonsson et al., 1999a).

Table 1.  A compilation of the three most abundant or frequent ectomycorrhizal (ECM fungal taxa, as recorded from mycorrhizas, in 49 ECM fungal community studies published so far. Only studies where the taxonomic identity of at least one of the three most abundant taxa was certified are included. Figures indicate at how many studies certain taxa dominated. The data from Canada includes four assessments of ECM communities on seedlings. Besides 21 papers referred to in the present review paper, 28 additional studies are included
Forest typeCountryNumber of studies with dominance of the following taxa
Total no of studiesSporocarps lacking or nonconspiceousConspicous sporocarpsSporocarps lacking or nonconspiceous
Cenococcum geophilumCorticea- ceaeThelephor- aceaeSuilloid (incl Rhizopogon)RussulalesCortinaria- ceaeMRA E-strainRisseia
Coniferous forest
BorealCanada1510 7 3 4 41 9 
 Sweden12 612 3 2 21 1 
TemperateSwitzerland & Germany 3 1 3 3  1   
 USA15 4  7 8 7   
Deciduous forests
 USA 1 1       
 Germany 1   1     
 Netherlands 1     1  
 Indonesia 1       1
In total: 4922221714143101

Characteristics of ECM communities

Studies from North America and Europe unanimously confirm that fungi that do not form obvious fruiting structures form the major mycorrhizal abundance. Fruiting species merely constitute 20–30% of the mycorrhizas (e.g. Gardes & Bruns, 1996; Jonsson et al., 1999a; Fig. 1). Apparently, carbon budget or resource allocation to production of sporocarps vs mycorrhizas and mycelial growth vary among species, as discussed in Gardes & Bruns (1996) seminal community paper. Typically, certain species that commonly form mycorrhizas are poorly or nonrepresented in the above-ground fruiting record, and commonly fruiting species may conversely have only a limited representation in below-ground mycorrhizal networks. Thelephoraceae, Russulaceae and, at least in post fire settings, also suilloid fungi repeatedly emerge as the most frequent or abundant mycorrhizas in forests of western USA, whereas Corticeaceous fungi and Cenococcum geophilum, followed by Russulaceae and Thelephoraceae appear as the most common in boreal forests (Table 1).

Figure 1.

(opposite) Late successional Scots pine stands, in northern Sweden, where both ectomycorrhizal fungal community structure from sporocarps and mycorrhizas, and fungal population structure of Suillus variegatus have been analysed. The last major disturbance was a forest fire in 1647. The relative proportion of mass in sporocarps (open red bars, 34 taxa) and of mycorrhizal abundance (closed blue bars, 24 taxa, distinguished by ITS-RFLP) is shown (Jonsson et al., 1999b). At the same site the spatial distribution Suillus variegatus genets was monitored from sporocarps analysed by somatic incompatibility reactions (from Dahlberg (1997), with permission).

The general rule is that a few fungal ECM taxa account for the most, (> 50%), of the mycorrhizal abundance and are widely spread whereas the majority of species are only rarely encountered (Gehring et al., 1998; Jonsson et al., 1999a,b; Peter et al., 2000;Stendell et al., 1999; Grogan et al., 2000). In two stands of Douglas-fir, 69 morphotypes were recognized from 17 500 assessed mycorrhizas; most types were rare, 19 types colonized > 1% of the mycorrhizas and 9 types colonized together 67% (Goodman & Trofymow, 1998a). This pattern of dominance of a few species and rarity of most others is a common community pattern of any taxonomic group of organisms.

Major determinants of the composition of ECM fungal community are host plant composition ( Molina et al., 1992) and edaphic factors (Gehring et al., 1998; Kernaghan & Harper, 2001). Within local forests, the spatial variation of ECM fungi is very high and most species show aggregated distributions (Gardes & Bruns, 1996; Kranabetter & Wylie, 1998; Bidartondo et al., 2000). A study of pinyon trees showed one or a few ECM fungal taxa to dominate single trees, and the dominant fungi varied between trees (Gehring et al., 1998). Spatial variation was concluded to have a larger effect on ECM fungal species composition, than that caused by low intensity wildfire, the major natural disturbance in Swedish boreal forests (Jonsson et al., 1999a). One exception is reported in a study of 1000-yr-old bristlecone pine stands at high elevation in California; here one ECM Pyronemataceae sp. was consistently present as inoculum in soil at all forest sites (Bidartondo et al., 2001). Commonly, small volumes of soil contain several ECM species, for example single four year old seedlings were colonized by up to 11 ECM fungi (Kranabetter & Wylie, 1998). Certain taxa also appear to grow better in either the organic or mineral layer (cf. Goodman & Trofymow, 1998).

A significant step forward in understanding ECM fungal community structure is to establish studies that allow estimation of the relative importance of different mechanisms that create this diversity, and take into consideration spatial and temporal scales. Hypotheses to explore can be placed in two categories: those based on niche differentiation between coexisting species; and, paradoxically, those based on similarities in competitive ability, that is nonequilibrium models. An example of the former is the intermediate disturbance hypothesis stating that a disturbance leads to a predictable sequence of species replacing each other. An intermediate range of disturbance would allow a range of species traits, early to late successional species, to coexist and lead to a high species richness. The latter assumes elements of stochasticity which enhance coexistence: so called lottery models state that due to poor dispersal ability, low local abundance and chance events, species may be absent in a neighbourhood and their abundance recruitment may thus be limited. Therefore the best competitor that happens to colonize a particular local site may well be inferior on a larger scale. Potentially, this can lead to unlimited species richness.

Temporal dynamics of ECM communities

Successional changes in primary vegetation successions, and thus of primary ECM fungal colonization, are described from glacial retreats and farmlands. Such questions have been studied in the course of natural secondary forest succession in a few cases. It must be remembered that forest management procedures, such as clearcutting, impose novel selection pressure upon ECM fungi. Marked changes in ECM community composition occur after intense, stand replacing fire that totally consumes the organic layer, a natural occurrence in certain forest ecosystems (Visser, 1995; Grogan et al., 2000). However, in low intensity fires, where part of the organic layer remains unburnt and a large proportion of trees survive the fire, the legacy of ECM fungi appears large (Jonsson et al., 1999a). Short-term variation in abundance of certain taxa of mycorrhizas may occur, but as yet has been poorly investigated, as weather conditions affect photosynthesis and thereby allocation of assimilates to the roots, and the mycorrhizas, and the temperature and humidity of the soil. This certainly affects the patterns of fruiting. However, a detailed molecular study of mycorrhizas in a Swiss Norway spruce forest conducted over three years showed a striking similarity in abundance of the 10 most abundant taxa (Peter et al., 2001).

Processes in ECM fungal populations underlie ECM fungal community patterns and thus, a knowledge of the ability of individual ECM fungal species to bridge distances in time and space is central to the understanding of ECM community structure. However, little is known about the relative importance of mycelial spread and longevity vs novel colonization from spores for any ECM fungal species. From the nearly 20 studies of ECM fungal population structures conducted so far (cf Redecker et al., 2001), it is apparent that genets in some species are numerous, short-lived (1–2 yr) and small (< few m, i.e. Hebeloma cylindrosporum, Laccaria amethystina, Amanita franchetii, Russula cremoricolor). Other species may have few and due to persistence (decades to centuries) and mycelial spread, relatively large genets (20–40 m in diameter, i.e. Suillus spp. and Cortinarius. Small genets may suggest colonization from recent meiospores, whereas fewer, larger genets indicate old mycelial structures, that have grown from a point source over decades. The latter are often found in old forests (Fig. 1). Some studies suggest mycelial colonization as the primary avenue of ECM fungal colonization in mature undisturbed forests, i.e. that clonal traits are of higher fitness than sexual traits under undisturbed conditions (e.g. Jonsson et al., 1999b; Kranabetter et al., 1999). Resistant ECM fungal propagules also persist, probably in the mineral soil, after major disturbances such as stand replacing wildfires (Baar et al., 1999). Exceptions do occur, for example L. amethystina appear to be short-lived in old beech forests (Gherbi et al., 1999). Demographic patterns in ECM fungi, or any soil dwelling fungus, have still not been explored.

There is a need to launch studies of regional scale dynamics of ECM fungal populations, and communities, to elucidate important population processes. There is also a need to determine which models are most applicable for different ECM fungi: theories of remnant population dynamics with discontinuous, nonequalibrium conditions and a pervasive impact of ‘accidents of history’ or, on the other hand, meta-population models with continuous, equilibrium conditions and without historic dimensions (cf. Eriksson, 1996).

Interactions with plant communities

Most ECM fungal species typically have broad, but differing, plant host ranges. They may form mycorrhizas with any ECM angiosperm or gymnosperm or only with angiosperms or gymnosperms, or a single family of hosts. This potentially extensive, but importantly sometimes also limited, linking of ECM plants via intra- and interspecific ECM mycelia, within and between species, may have major consequences for our understanding of plant community characteristics, plant competitive interactions and performance (Molina et al., 1992). These linkages potentially allow for nutrient and carbon transfer between tree species, but their significance in nature remains to be settled. Evidence for the potential of such mycelial linkages is accumulating. Field and laboratory soil bioassays showed Douglas fir to share ECM fungi and potentially form hyphal links with several other plant species (Massicotte et al., 1999). Jones et al. (1997) found that 91% of paper birch and 56% of Douglas fir intermingled mycorrhizal roots examined were colonized by the same mycorrhizal fungi. In these tree species, a 4–7% net transfer of isotopically labelled carbon from birch to Douglas fir has been demonstrated, which may be ecologically significant (Simard et al., 1997). If net transfer is predominantly from pioneer plant species to late successional species, ECM fungi may be an important contributing factor determining plant community development.

In a mixed stand of bishop pine and Douglas fir the majority of identified mycobionts were associated with both hosts, even within small soil volumes, suggesting that individual fungal genotypes linked the putatively competing tree hosts (Horton & Bruns, 1998). One study in the Canadian Rockies recorded all present ECM fungi to form mycorrhizas with both Dryas and Salix (Kernaghan & Harper, 2001). In a comparison between seedlings of lodgepole pine, white spruce and subalpine fir seedlings, 43 of 74 mycorrhizal morphotypes were shared by all conifers, constituting 52% of the total mycorrhizal abundance (Kranabetter et al., 1999). However, it is also important to appreciate certain ECM fungi, for example suilloid fungi, as rigidly restricted to certain plant taxa and thus only potentially linking intraspecific plants.

Ericoid and ectomycorrhizal plants often coexist in boreal and Mediterranean forests and are normally considered as separate mycorrhizal systems with mycobionts of different taxonomic affinity. However, recently an ericoid mycorrhizal fungus was shown to associate with Quercus roots (Bergero et al., 2000). The possibility that some sharing may occur between ericaceous plants and ECM trees has been discussed, but questions concerning whether it occurs and what possible ecological significance it may have still need to be answered in appropriate, controlled studies. The results of such experiments may have important implications for our understanding of interactions between trees and the understorey.

Effects on ecosystem function

To date, ECM fungal community studies have been mainly of a descriptive or correlative nature and few experimental studies have been performed. Causal mechanisms underlying the dynamics, and implications for ecological functions, have rarely been addressed under controlled experimental conditions (Lilleskov & Bruns, 2001). It is likely that many ECM fungi may fulfil similar ecological functions and that a degree of functional redundancy exists in ECM fungal communities. In an experiment on the effect of plant litter composition, there was no evidence that the composition of the present ECM fungi affected the performance of seedlings (Nilsson et al., 1999). Given their taxonomic diversity, communities are still likely to retain vast amounts of functional heterogeneity. Yet it is difficult to generalize about the ecological functionality of ECM communities because relatively few investigations of ECM fungal physiology have been conducted with more than a few isolates, and those typically include only a handful of taxa that typically occur in low abundance (Cairney, 1999). The potential relationship between ECM fungal species richness and ecosystem productivity has once been explored without finding any correlation (Gehring et al., 1998). As debated at length, there is no mechanistic relationship between species richness, or any diversity measure, and ecosystem function (Bengtsson, 1998). Probably, it is more vital to understand linkages between the actions of key species or functional groups and ecological functions, as diversity is an abstract aggregated property of species in the context of communities or ecosystems (Fig. 2). The functioning of terrestrial ecosystems seems to depend heavily on soil biodiversity. We now have the potential to address and experimentally test questions of ecological significance, as abundant ECM fungi, potentially being key species, can be identified. One approach could be to use A set of ECM fungi from selected environmental settings, and to investigate whether the abilities of these ECM fungal communities differ for various functions, such as carbon cost, rate of nutrient-cycling or response to perturbations.

Figure 2.

Diversity and ecosystem function have no direct relationships to each other, but are both functions of the presence and activities of individual species or functional groups and their interactions. The environment affects species, and the outcome of their interactions, but not diversity or ecosystem processes directly. The question mark indicates the question as to whether diversity and ecosystem processes are related (redrawn from Bengtsson, 1998).

One promising approach is to explore ECM fungal community functioning under simulated forest conditions using microcosm systems (Tammi et al., 2001). In these systems, intact mycorrhizal root systems comprising individual species, for example abundant taxa likely to be functionally important, or natural communities, can be manipulated and analysed to determine carbon and nitrogen relations and host root-fungal metabolic activities that contribute to plant growth or plant community productivity. The application of molecular methods is already providing detailed insights into the complexity of ECM fungal communities and offers exciting prospects to elucidate processes structuring ECM fungal communities and advance our understanding of plant ecology, such as plant interactions, and ecosystem processes.


I thank Robin Sen, Marie-Charlotte Nilsson, Tom Bruns, Åke Olsson, Roger Finlay and Jan Stenlid for critical reading of earlier drafts of this paper. Major support over the years for this research came from the Swedish Council for Forestry and Agricultural Research.