Host preference, niches and fungal diversity


Ectomycorrhizal fungi occur in remarkably species-rich assemblages. One of the prevailing hypotheses to explain this diversity is niche differentiation; by occupying distinct ecological niches within a site, multiple fungal species are able to co-occur (Bruns, 1995). In this issue of New Phytologist (pp. 430–440), Ishida and colleagues make a significant contribution to our understanding of niche differentiation by showing that co-occurring host species have distinct mycorrhizal communities, reflecting both host taxonomy and, arguably, successional status. Although host specificity is a well-known phenomenon (Molina & Trappe, 1982), it has not previously been clear to what extent co-occurring species of plants support different species of ectomycorrhizal fungi. Using individual root collections from co-occurring plants, Ishida and colleagues have effectively demonstrated that host specificity (or, more accurately, host preference) is an important factor in local diversity. Regrettably, statistical power issues prevent a robust determination of whether host preference is more common at the family than at the genus level. Nonetheless, there are strong indications that both host family and successional status are important in determining plant–fungal associations.

‘… they show unequivocal evidence that host preference is an important component of the correlation of ectomycorrhizal fungal diversity with plant diversity’

The estimate of over 300 fungal species in Ishida and colleagues’ study represents the highest ectomycorrhizal fungal species richness yet described. For comparison, I used data from other recent papers where species richness has been calculated using the same estimator of total species richness and my own unpublished data. Although based on a small data set, a remarkably clear pattern emerges: estimated fungal richness is a linear function of the number of ectomycorrhizal host species (n = 11, P < 0.001, r2 = 0.95; Fig. 1, Table 1). Thus, while the extremely high diversity found by Ishida and colleagues is indeed remarkable, it falls exactly in line with previous data from systems with fewer ectomycorrhizal plant species. As the rapid development of molecular tools permits ever larger and more comprehensive surveys of fungal communities, it will be interesting to see if, and at what level, the increase in fungal diversity reaches an asymptote.

Figure 1.

Estimated total ectomycorrhizal fungal species richness as a function of the number of ectomycorrhizal plant species; data from published reports of below-ground fungal diversity where total richness has been estimated (fungal richness = 2.4 + 49.1 × plant richness; P < 0.001; adjusted r2 = 0.95). Circles and the regression line are based on second-order jackknife estimates of species richness. Data from Ishida et al. (2007) are indicated by a closed circle and included in the regression. For comparison, additional points have been added from reports using other richness estimators (first-order jackknife (crosses) or Chao2 (triangle)), but are not included in the analysis. The outlier Chao2 estimate with three plant species of only 37 fungal species is from an early successional community on Mt Fuji (Nara, 2006). See Table 1 for data.

Table 1.  Literature values for estimated richness as a function of number of host speciesa
CitationHost speciesObserved richnessBootstrapChao1Chao2Jackknife 1Jackknife 2
  • a

    Values in bold represent the data points shown in Fig. 1. Data were obtained by searching Google Scholar using the terms ‘ectomycorrhiza and diversity and (Chao1 OR Chao2 OR Jackknife OR Bootstrap)’ with all papers including a jackknife estimate of diversity based on molecular identification included. Host species was the total number of hosts present for studies using soil cores, or the number of species sampled for studies using bioassay seedlings or direct root identification. Where more than one estimate was provided (e.g. for different treatments) the highest estimate was used.

  • b Richness estimates; personal communication from A. Izzo, based on data in Izzo et al. (2005).

  • c Data from Cedar Creek Long-term ecological research (LTER) site, MN, USA. Host species are Quercus ellipsoidalis, Quercus macrocarpa, Corylus americana, and Helianthemum bicknellii. Data were collected by Dickie, Avis, Dentiger, McLaughlin et al.

  • d

    Mature trees.

  • e

    Seedlings near mature trees.

Ishida et al. (2007)8205362315387
Tedersoo et al. (2006)6172322329
Izzo et al. (2005)b4101230163207
Dickie et al. (unpublished data)c4125148167175194
Luoma et al. (2006)4101136145
Toljander et al. (2006)366112149
Nara (2006)33639.937.3
Walker et al. (2005)275116143
Kjøller (2006)13143.348.1
Cline et al. (2005)d14353.357.156.863.2
Cline et al. (2005)e12027.162.730.735.5
Koide et al. (2005)12730.836.042.0
Korkama et al. (2006)13446.5
Saari et al. (2005)11619

Host preference is only one explanation of increased fungal diversity with increasing number of plant species. Increased plant diversity is also likely to create more heterogeneous litter inputs, which may create opportunities for niche differentiation by ectomycorrhizal hyphae (Conn & Dighton, 2000; Wardle, 2006). Alternatively, species richness of ectomycorrhizal plants may be correlated with site conditions that independently favor high species richness of ectomycorrhizal fungi. This is where the detailed work of Ishida and colleagues is invaluable; by independently sampling roots of eight plant species they provide unequivocal evidence that host preference is an important component of the correlation of ectomycorrhizal fungal diversity with plant diversity. Causality and directionality remain, of course, unproven.

Mechanisms of host preference

While there are genetic and physiological barriers to certain plant–fungus associations (Molina & Trappe, 1982), host specificity of ectomycorrhizal fungi does not appear to be absolute. It has been noted that plant–fungus associations that form under laboratory conditions are not always indicative of host specificity under natural conditions, a phenomenon sometimes termed ‘ecological specificity’. The observation of ecological specificity implies that environmental factors have a direct role in determining host specificity. Thus, host preference of mycorrhizal fungi reflects a realized, rather than fundamental, niche.

Restricted realized niches generally result from competition. Ectomycorrhizal fungi compete for roots (Wu et al., 1999), and we know that small differences in the rate of initial stages of mycelial growth onto roots can have longer term impacts on competitive outcomes, through priority effects (Kennedy et al., 2006). Nonetheless, while competitive interactions are generally important in soil fungal communities, our understanding of ectomycorrhizal competition and the influences that plant hosts may have on this competition remains limited (Wardle, 2006).

An alternative hypothesis to strict competition would be direct plant selection of one fungal associate over another. It may be that plants, in the presence of multiple potential symbiotic partners, are able to selectively allocate resources to ‘preferred’ mycorrhizal associates. If this occurs, a hypothetical species ‘A’ might be able to form mycorrhiza with a plant host under laboratory conditions, but be excluded in the presence of a hypothetical species ‘B’ under field conditions. Nonetheless, evidence for preferential plant allocation of resources to one fungal partner over another is limited. It is also interesting that in arbuscular mycorrhiza, at least, an opposite pattern has emerged: the mycorrhizal community developing under particular plant species can be inferior in terms of increasing plant growth (Bever, 2002). This may suggest that plant selection for ‘preferred’ symbionts is either nonexistent or ineffective at optimizing fungal community composition.

The n-dimensional hypervolume of mycorrhizal niche space

Hutchinson (1957) defined a niche as ‘an n-dimensional hypervolume … every point in which corresponds to a state of the environment which would permit the species to exist indefinitely’. The work by Ishida et al. confirms the importance of host preference as one environmental dimension (or niche axis) upon which fungal niche differentiation can occur. Other known ectomycorrhizal niche axes include soil depth (Dickie et al., 2002; Genney et al., 2006), seasonality (Koide et al., 2007), and distance from trees (Dickie & Reich, 2005). Factors such as stand age (Gebhardt et al., in press) or soil type (Lekberg et al., 2007) are also important in structuring mycorrhizal communities; however, these larger scale factors would generally increase between-site (or β) diversity, rather than within-site (or α) diversity.

Both at the plant interface of the ectomycorrhizal root-tip and in the soil as hyphae, ectomycorrhizal fungi encounter a highly variable environment with myriad possible niche dimensions. Many of these niche dimensions are relatively narrow in breadth. Nonetheless, dimension breadth is relatively unimportant compared with dimension numbers (n), as available niche space in a community, i.e. the ‘n-dimensional hypervolume’, increases multiplicatively with niche breadth but exponentially with increasing dimension numbers. Given this, it is perhaps not surprising to find that ectomycorrhizal fungi occur in such species-rich communities. Other factors, such as dispersal limitation (Lekberg et al., 2007), trophic interactions (Wardle, 2006) and soil disturbance, are likely to further contribute to this fungal diversity.

Coda: the jack-of-all-trades

There is at least one notable exception to the rule of niche differentiation: the ectomycorrhizal fungus Cenococcum geophilum. It comes as no surprise that C. geophilum was found on every host tree species studied by Ishida and colleagues. The same species has been found across soil profiles (Dickie et al., 2002), at all stages of stand development (Gebhardt et al., in press), at every distance from forest edges (Dickie & Reich, 2005), and at every season of the year (Koide et al., 2007). Even accepting that C. geophilum may be a closely related species complex, such a wide distribution of a genus is still remarkable, particularly given that C. geophilum has no known long-distance dispersal mechanism. The invocation of niche differentiation as an explanation for fungal diversity has to be tempered by the recognition that some fungi, such as C. geophilum, have yet to show any real evidence of niche restriction.


R. T. Koide, R. G. FitzJohn, P. G. Kennedy and P. G. Avis provided helpful comments and discussion. The author is supported by research funds from the Foundation for Research, Science and Technology of New Zealand.