General latitudinal gradient of biodiversity is reversed in ectomycorrhizal fungi


(Author for correspondence: tel +372 56 654 986;

Tropical rainforests support a tremendous biodiversity of animals and plants, but because of difficulties in observation and identification, little is known about the richness patterns of microbes, including fungi. The general latitudinal gradient of diversity (LGD) demonstrates that nearly all terrestrial and marine macroorganisms studied so far have peak richness at low latitudes (Hillebrand, 2004). Indeed, recent research in the Americas revealed that neotropical rainforests comprise the highest diversity of endophytic fungi. Plant taxa harbor different endophyte communities and therefore fungal diversity increases with host diversity (reviewed in Arnold, 2008). Similar patterns may enhance the diversity of ectomycorrhizal fungi (EcMF) in the temperate zone (Bruns et al., 2002; Ishida et al., 2007; Tedersoo et al., 2008), but the relevance of this and other potential factors remain unknown in tropical ecosystems because of a lack of published studies. In this issue of New Phytologist, Peay et al. (pp. 529–542) address the biotic and abiotic factors driving the composition of an EcMF community in a Bornean rainforest that comprises one of the highest plant diversities in the world.

‘… many of the proposed causal mechanisms that have been developed to explain the general LGD pattern may be inapplicable to EcMF or need substantial modification.’

Do tropical and temperate EcMF communities differ? Extensive fruit-body surveys have revealed that tropical and temperate forests share many common EcMF lineages, such as the /boletus, /cantharellus, /clavulina, /russula–lactarius, /tomentella–thelephora, etc. (reviewed in Tedersoo et al., 2010a). However, EcMF fruit-bodies do not provide useful data for quantitative comparisons of the local EcMF diversity among studies, because of their strongly seasonal and yearly unpredictable production and ephemeral habit. Several common EcMF lineages, such as the /tomentella–thelephora, /piloderma and /hysterangium, produce either resupinate fruit-bodies on the underside of debris or hypogeous sporocarps that remain unnoticed unless specifically searched for. Moreover, many ascomycete lineages and some basidiomycete lineages probably lack fruit-bodies. The development of molecular techniques has facilitated the identification of EcMF from root tips and mycelia that can be found in soil throughout the year. Rapidly advancing sequencing and microarray technologies enable quantitative analyses of EcMF diversity and ecology based on numerous independent soil or root samples. Peay et al. collected 101 soil samples from seven 400-m2 plots at Lambir Hills, Borneo, and recovered 105 EcMF species, most of which belong to the /russula–lactarius, /boletus and /tomentella–thelephora lineages. These data are valuable for shedding light on EcMF diversity and community composition in tropical habitats, but do not represent all tropical forests that differ in historical biogeography and composition of ectomycorrhizal host lineages. To address differences in EcMF diversity and community structure between temperate and tropical rainforests, we supplemented the data from Lambir Hills with three data sets from different tropical regions (Africa, South America and South East Asia) and compared them with the results of temperate studies that employed generally similar methods for sampling and identification (Table 1).

Table 1.   Site descriptions of four temperate and four tropical forests used
SiteGeocodeVegetation typeSite size (ha)Host lineagesReference
  1. 1It remains unknown whether the root tips of these taxa were actually sampled.

Subtropical montane cloud forest3FagalesMorris et al. (2009)
Mt Field, Tasmania42°41′S
Wet sclerophyll forest20Leptospermoideae, Fagales, PomaderreaeTedersoo et al. (2008)
Sierra Foothill, California, USA39°17′N
Mediterranean woodland50Fagales, PinaceaeMorris et al. (2008); Smith et al. (2009)
Tagamõisa, Estonia58°45′N
Temperate woodland10Fagales, Pinaceae, Salicaceae, TiliaTedersoo et al. (2006)
Bukit Bangkirai, Indonesia01°01′S
Tropical rainforest30Dipterocarpaceae
K. Nara et al. (unpublished)
Lambir Hills, Malaysia04°20′N
Tropical rainforest18Dipterocarpaceae, Fagales, Leptospermoideae1, Detariae1Peay et al. (2010) (this issue)
Monts de Cristal, Gabon00°37′N
Tropical rainforest20Amhersteae, Dipterocarpaceae, UapacaL. Tedersoo et al. (unpublished)
Yasuni, Ecuador00°37′S
Tropical rainforest30Pisoniae, CoccolobaeTedersoo et al. (2010b)

Across these studies, the most species-rich lineages were /thelephora–tomentella and /russula–lactarius, followed by /cortinarius, /sebacina, /clavulina, /boletus and /inocybe (Supporting Information Fig. S1). Surprisingly, the relative richness of only two EcMF lineages differed significantly between tropical and temperate habitats –/russula–lactarius (t-test: = 18.8; = 0.005) and /inocybe (= 8.4; = 0.027), which were relatively more diverse in tropical and temperate forests, respectively. However, the statistical significance collapsed when adjusted using the Bonferroni correction for multiple tests.

In contrast to the compositional similarity of EcMF, the diversity of EcMF in tropical sites was lower than that in temperate sites, as revealed by the rarefied species richness (Fig. 1a; t-test: = 61.3; < 0.001) and minimal richness estimators – Jackknife 2, Chao2 and ICE (Fig. S2). This pattern in EcMF strongly contradicts the negative LGD (Hillebrand, 2004). Thus, many of the proposed causal mechanisms, such as the mid-domain, geographical area and climatic stability effects that have been developed to explain the general LGD pattern (Lomolino et al., 2006), may be inapplicable to EcMF or need substantial modification.

Figure 1.

 Rarefied accumulation curves of (a) species and (b) lineages of ectomycorrhizal fungi (EcMF) in four temperate and four tropical sites. Dotted lines indicate 95% CI. Brown, Tagamõisa, Estonia; green, Mt Field, Australia; blue, Huizteco, Mexico; black, UC Sierra Foothill, CA, USA; purple, Lambir Hills, Malaysia; red, Monts de Cristal, Gabon; orange, Bukit Bangkirai, Indonesia; pink, Yasuni, Ecuador. The minimal richness estimates were calculated based on original data sets by use of a computer program estimate s (Colwell, 2006; estimate s: statistical estimation of species richness and shared species from samples. Version 8. Persistent URL <>) and 1000 replicates.

What, then, causes the lower EcMF richness in tropical ecosystems? Although the information available is scant, we propose three possible mechanisms. First, historical and biogeographical effects may partly explain the differences in EcMF richness pattern observed between tropical and temperate ecosystems. Many EcMF lineages occur in both tropical and temperate biomes. While several temperate lineages are notably absent from tropical forests, no strictly tropical lineages are known (Tedersoo et al., 2010a). The EcMF community data support this observation: temperate forests harbor more EcMF lineages than tropical forests (Fig. 1b; t-test: = 345.4; < 0.001). The strictly temperate EcMF lineages probably evolved at higher latitudes with the Pinaceae hosts, but may be inferior competitors in tropical conditions. However, this can only partly explain the observed pattern, because these temperate lineages are relatively species-poor and usually form a minor component in temperate forests (except for the /amphinema–tylospora and strictly Pinaceae-specific /suillus lineages).

Second, when both soil and roots are regarded as habitats for EcMF, the lower diversity and abundance of these habitats may account for the lower EcMF diversity in the tropics. The roots of host plants are an obligatory energy source for all EcMF. Therefore, the co-existence of different hosts enhances habitat diversity. With certain exceptions (Morris et al., 2008), host preference is rarely evident at the host species level, but is probably more important at higher taxonomic levels, that is, from genus to phylum (Ishida et al., 2007). While tropical forests are often dominated by a single ectomycorrhizal host lineage, such as the Dipterocarpaceae or certain groups of Fabaceae, temperate habitats are often composed of multiple codominant host lineages, for example, Fagales, Pinaceae and Salicaceae. Although many different dipterocarp or leguminous host species co-exist in a tropical site, their phylogenetic habitat difference is lower compared with temperate forests. Similarly, the soil habitat is less diverse in the tropics than in temperate forests. In tropical forests, the soil profile is usually poorly differentiated and has thin organic and litter layers as a result of the rapid consumption of organic matter by mesofauna and the fast decomposition rates that occur in constantly warm and humid conditions. Many temperate EcMF species display niche differentiation by soil horizons (e.g. Lindahl et al., 2007) and EcMF communities accumulate more species in better developed soils (Nara et al., 2003).

Third, resource availability and fragmentation may explain the observed pattern of EcMF richness. Many temperate forests are exclusively dominated by suitable host trees that may account for 100% in basal area or the number of stems. Conversely, EcMF hosts usually contribute < 75% to the basal area in South East Asia and in monodominant patches of other tropical regions. Tropical EcM hosts are most often distributed sparsely, forming small isolated host islands in non-EcM vegetation. Fragmentation and availability of fewer EcM roots may reduce the population size of EcMF and eventually result in fewer co-existing EcMF species (Peay et al., 2007; Tedersoo et al., 2010b)

The comparison of temperate and tropical forests demonstrates that the EcMF richness pattern is an exception to the general LGD. The underlying mechanisms of the EcMF richness pattern may include historical–biogeographical factors, habitat diversity and resource availability–fragmentation. Accumulating EcMF data from seasonal tropical and subarctic ecosystems will probably improve our understanding of these causal mechanisms and the entire shape of the latitudinal gradient of EcMF diversity. In addition to high-quality DNA sequence data, collection of various environmental and biological data for each study site is essential to develop a general model of EcMF diversity and community composition on a global scale (Lilleskov & Parrent, 2007).