Pioneer dwarf willow may facilitate tree succession by providing late colonizers with compatible ectomycorrhizal fungi in a primary successional volcanic desert


Author for correspondence: Kazuhide Nara Tel: +81 424 655601 Fax: +81 424 655601 Email:


  • • To advance understanding of the contribution of ectomycorrhizal (ECM) fungi to tree successional processes, natural establishment patterns of secondary colonizing hosts and their ECM fungal communities were investigated with special reference to pioneer hosts.
  • • In the volcanic desert on Mount Fuji, Japan, vegetation is sparsely distributed, resembling islands in a sea of scoria. Of 509 vegetation islands in the research area, 161 contained Salix reinii (Salix), the first colonizing ECM host species. The spatial coincidence between secondary colonizing timber species and Salix was analysed, and ECM fungal communities were studied using molecular identification methods.
  • • I found 39 and 26 individuals of Betula ermanii and Larix kaempferi, respectively. Without exception, these individuals were all accompanied by Salix. The ECM fungal communities of these timber species showed high similarity to that of Salix and were dominated by generalists that were compatible with two or more plant families.
  • • In this desert, available ECM propagules are limited. Pioneer Salix may contribute to tree succession by providing adjacent late colonizers with compatible ECM fungal symbionts.


Ectomycorrhizal (ECM) fungi are symbiotic microorganisms that colonize the roots of many tree species. These fungi usually dominate forest soils and effectively scavenge soil nutrients that are subsequently transferred to the host plants. This nutrient supply from ECM fungi appears to be a requisite for normal growth in many host plants (Smith & Read, 1997). In single forest stands, tens to hundreds of ECM fungal species coexist. Many of these microorganisms are generalists that are compatible with many different tree species (Molina et al., 1992). Because generalists usually dominate ECM communities, their spores and living mycelia are ubiquitous in forest soil. Hence regenerating seedlings of host species can be colonized rapidly by these ECM fungi (Jonsson et al., 1999; Matsuda & Hijii, 2004).

Disturbance can reduce or change the availability of ECM sources and alter the ECM communities on regenerating hosts (Perry et al., 1987 and references therein), but usually not to the extent that ECM association is impossible. For example, after a stand-replacing forest fire, Rhizopogon showed increased dominance on Pinus muricata seedlings relative to the prefire forest (Horton et al., 1998; Baar et al., 1999). Even in a gradually receding glacier area (Helm et al., 1999), a sand dune (Ashkannejhad & Horton, 2006), and a mining district (Ingleby et al., 1985), seedlings were readily associated with some ECM fungi. Thus host growth is not suppressed by the complete absence of ECM associations in most potential forest areas. It is therefore difficult to elucidate the contribution of ECM associations to ecological processes because of the difficulty in preparing nonmycorrhizal controls under natural conditions in most habitats.

Successful ECM colonization is problematic in some primary successional sites, especially in severely devastated volcanic deserts (Allen et al., 1992). If nonmycorrhizal controls are available in natural settings, studies in these sites can provide important and fundamental knowledge concerning the role of ECM symbioses in ecological processes such as succession. In the volcanic desert on Mount Fuji, Japan, the first ECM plant to colonize is an alpine dwarf willow, Salix reinii (hereafter Salix unless otherwise specified; Nara et al., 2003a). Although all established Salix shrubs are associated with ECM fungi (Nara et al., 2003b), nonmycorrhizal seedlings that are transplanted into most parts of this area remain uncolonized by any ECM fungi until the end of the first growing season because of limited propagule availability (Nara & Hogetsu, 2004). Current-year Salix seedlings can easily develop ECM associations only beside established Salix shrubs via the spread of ECM mycelia from the established shrubs (Nara & Hogetsu, 2004; Nara, 2006). The growth of Salix seedlings is significantly improved once the seedlings are connected to the established ECM systems of most fungal species (Nara, 2006). In this way, early-established Salix shrubs facilitate the subsequent establishment of conspecific seedlings by providing ECM fungal symbionts.

ECM fungi on established Salix may also aid in the establishment of secondary colonizing tree species, thereby facilitating tree succession in the volcanic desert. A previous transplanting experiment showed improved ECM colonization on seedlings of late-colonizing tree species beside established Salix, but did not detect any positive effects of ECM associations on the growth of the secondary colonizers (Nara & Hogetsu, 2004), probably because of nutrient competition with the early-established Salix. However, this does not necessarily mean that ECM fungi do not contribute to primary tree succession. Because any manipulative approach in the field must be restricted in time and space, such an approach may not be suitable for evaluating primary tree succession, which requires at least several centuries to proceed. Because of the longevity of tree species, the rare establishment of single trees could have large long-term effects during primary succession. To understand the contribution of ECM fungi to such rare events, studies of the natural establishment patterns of late colonizers may overcome the shortcomings of manipulative approaches (Nara & Hogetsu, 2004) because rare events within the whole research area are temporally integrated.

Another important aspect of ECM symbioses in primary succession of trees is host–fungus compatibility. Because ECM propagules are limited in this volcanic desert, few ECM sources other than that of symbionts on early-established hosts can be used by late-colonizing plants (Nara & Hogetsu, 2004; Nara, 2006). If ECM fungi on early-established hosts are not compatible with late colonizers, the late colonizers must wait for the stochastic arrival of compatible ECM propagules. ECM communities on naturally established saplings of late colonizers, with reference to host ranges of individual fungal species, should provide valuable information on the contribution of ECM fungi to primary succession of trees.

To advance our knowledge on the role of ECM fungi in tree successional processes, the spatial patterns were analysed of naturally established saplings of secondary colonizing timber species 300 yr after the last volcanic eruption, with special reference to pioneer ECM shrubs in the volcanic desert on Mount Fuji, Japan. The ECM communities of two timber species were compared with those of early-established shrubs. The host ranges of individual fungal species were also evaluated. Finally, the importance of ECM fungi in primary succession is discussed.

Materials and Methods

Research site

Mount Fuji, the highest and best known mountain in Japan, erupted in 1707, and its south-eastern area was completely covered with scoria (tephra, typically 2–30 mm in diameter), up to 10 m deep. The existing vegetation was completely destroyed, and is now recovering. Whereas the treeline is generally located at 2500 m asl on the other sides of the mountain, it is located at c. 1300 m on the south-eastern slope, and continues to increase in altitude after 300 yr of vegetation recovery. The study area was located above the treeline at 1450–1600 m asl.

At the study site vegetation is distributed patchily, forming isolated vegetation islands in a sea of volcanic desert. Each vegetation patch is normally initiated by the perennial herb Polygonum cuspidatum. This plant species is usually nonmycorrhizal, but on rare occasions it is colonized by arbuscular mycorrhizal fungi at low levels if accompanied by other arbuscular mycorrhizal hosts in the volcanic desert (Wu et al., 2004). Individuals of P. cuspidatum spread vegetatively to 10 m in diameter, providing stable habitats for subsequent plant species on the unstable scoria desert (Adachi et al., 1996). Many plant species are able to invade this stable habitat (see list of species and their frequencies in Nara et al., 2003a). Salix reinii usually establishes beside a P. cuspidatum patch, where water and light conditions are favourable. This dwarf Salix is not significantly different from early-established P. cuspidatum in terms of plant height (usually <1 m) and leaf mass. Salix is the first pioneer ECM host species in the research site, and all established shrubs of this species, without exception, are intensively colonized by ECM fungi (Nara et al., 2003a). In total, 31 ECM fungal species have been shown to be associated with Salix (Nara et al., 2003a,b).

Because ECM tree species such as Fagus, Quercus, Abies and Tsuga dominate other areas at the same elevation on Mount Fuji (Ohsawa, 1984), it is likely that the research site will eventually become occupied by forests of these species in the distant future. Two timber species, Betula ermanii and Larix kaempferi (hereafter referred to as Betula and Larix, respectively), are the ECM host plants that follow Salix in the successional sequence. As in the case of Salix establishment, these secondary colonizing plants always establish beside a vegetation patch, where the soil is poorly developed compared with inside a vegetation patch.

Although nitrogen-fixing plants play an important role in some areas of primary succession (Walker & del Moral, 2003), this is not the case on Mount Fuji. In this area, there were two N-fixing plant species. A relatively late-colonizing legume, Hedysarum vicioides, was observed as a minor constituent in c. 20% of vegetation patches. An alpine alder, Alnus maximowiczii, is also rare, and only two saplings were found in 21 ha of research area. The total area covered by these two N-fixing plants was <0.001% of the total vegetation. The absence of N-fixing plants is common at the initial stage of primary succession (Sprent & Sprent, 1990).

Spatial patterns of timber establishment

Because naturally established saplings of timber species were rare, I used a larger study area (c. 21 ha; Fig. 1) than the 5.5-ha quadrat used in previous studies (Nara et al., 2003a, 2003b). The position of all vegetation patches >1 m in smallest diameter distributed in the study area was recorded using GPS. The area of each vegetation patch was estimated from the largest and smallest diameters.

Figure 1.

Distribution pattern of vegetation patches and occurrence of timber species (Betula ermanii and Larix kaempferi), with special reference to early-established willow (Salix reinii) shrubs in the volcanic desert on Mount Fuji, Japan.

All vegetation patches were surveyed in October 2004 for established Betula and Larix individuals. Because vegetation patches are physically separated by some distance by scoria desert, one patch is unlikely to affect another. This condition allowed examination of the natural establishment patterns of timber trees, and the conditions for this establishment, by determining the differences between patches that contained timber trees and those that did not. This study focused on the presence and absence of early-established ECM hosts (Salix shrubs) in each vegetation patch.

Sampling and treatment of ECM root tips

In early October 2004, 10 saplings of each of the two timber species were selected. The range of stem diameters of selected Larix saplings was 15–40 mm at ground level. For Betula, in most cases multiple stems had sprouted from each individual, and the largest stem diameters ranged from 14 to 46 mm. The age range of selected saplings was estimated at 7–20 yr for Betula and 7–22 yr for Larix by analysing the growth rings of four individuals of each species (data not shown).

A 20-cm section of main root (c. 3–5 mm diameter) that was traced to the stem was sampled in two different positions for each sapling. Each main root section and all attached root tips were placed in a plastic bag with a small amount of scoria, and kept at 4°C until use. If Salix roots were present in the same sampling positions, they were treated in the same manner. In total, I sampled 19, 20 and 14 root systems of Betula (no traceable root found at one position), Larix and Salix, respectively.

Each root system was washed carefully with tap water. I arbitrarily subsampled c. 100 root tips from a root system. If a root system contained <100 root tips, all root tips were used. The subsamples were observed under a dissecting microscope to determine the percentage of ECM root tips. ECM tips were then classified into morphotypes based on surface colour, texture, emanating hyphae and rhizomorphs, using the method from a previous study (Nara et al., 2003b).

Molecular identification of ECM fungi

Two to five ECM root tips in a morphotype in each root system, depending on the number of root tips in the morphotype, were individually placed into 2.0-ml microtubes as replicates. I tried to use at least duplicate samples for molecular identification, but in some cases only one tip was available. These replicate ECM samples were dried and used for DNA extraction, as described by Nara et al. (2003b). Polymerase chain reaction (PCR) was conducted to amplify the internal transcribed spacer (ITS) region of rDNA of ECM fungi using a KOD plus kit (Toyobo, Osaka, Japan). The PCR primers were ITS1F and ITS4 (Gardes & Bruns, 1993), which were labelled with fluorescent Beckman dyes (Proligo Japan Co., Kyoto, Japan) D4 and D3, respectively. To adjust the fluorescence intensity between D4 and D3, the labelled ITS1F primer was diluted with the same amount of nonlabelled ITS1F primer. The PCR product (2 µl) was digested in 10 µl HinfI solution (1061A, Takara Shuzo, Shiga, Japan) at 37°C for 8 h, and diluted twice with sterilized MQ water (Elix UV10, Millipore, Billerica, MA, USA). As another fragment, I used a PCR product that was amplified using the ITS3 and ITS4 primer set; ITS3 was labelled with D2 fluorescent dye (Proligo). The latter PCR product was diluted five times with sterilized MQ water. Then 1 µl of both diluted solutions were mixed as a sample solution for terminal restriction fragment length polymorphism (T-RFLP) analysis (Dickie et al., 2002b; Zhou & Hogetsu, 2002; Nara et al., 2003b).

DNA fragments in sample solutions were purified and diluted in 25 µl SLS solution (PN608082, Beckman Coulter, Fullerton, CA, USA) containing 0.25% of the 600-bp size standard mix (PN608095, Beckman Coulter). For T-RFLP, capillary gel electrophoresis was conducted in a sequencer (CEQ8800, Beckman Coulter) using the default setting for 600-bp fragment analysis. Three separate peaks, which represented D2-, D3-, and D4-binding fragments, were obtained from each sample in most cases. In some cases, a sample contained multiple peaks for a fluorescent dye because of contaminating soil fungi that appeared irrespective of morphotype. If the contaminating peaks were indistinguishable, all peaks of this dye were excluded from further analysis.

To identify the ECM fungal species, each T-RFLP pattern was compared with the patterns of sporocarps described in a previous study (Nara et al., 2003b). If the T-RFLP patterns did not match any of the sporocarp species, they were grouped into T-RFLP types. T-RFLP patterns with all three fragments showing similar sizes within ±2 bp were regarded as single T-RFLP types. Some samples showed no peaks in T-RFLP analysis because of unsuccessful PCR amplification. These samples were regarded as the same fungal species identified within the same morphotype of each root system or, rarely, within the same morphotype of different root systems.

Two different samples from each T-RFLP type were selected for sequencing. PCR products using nonlabelled ITS1F and ITS4 primers were cleaned and used for sequencing reaction with DTCS Quick Start Master Mix solution (PN608120, Beckman Coulter). Three sequencing primers (ITS1F, ITS3 and ITS4) were used individually for each sample. Sequence data obtained were combined for each sample using atgc ver. 4.2 (Genetyx Co., Tokyo, Japan). The combined sequence data were compared with the sequences of known species in the DDBJ/EMBL/GenBank database using blast.

In total, 7192 root tips were observed microscopically, 587 ECM tips were used for T-RFLP analyses, and 42 were sequenced.

Data analyses

The species richness of ECM fungi on each host species was estimated from the observed species counts using Chao2 and bootstrap estimators, using estimate s ver. 7.5 (Colwell, 2005; Mao & Colwell, 2005). To compare the number of observed species against species known to exist in this area, a species-accumulation curve was created using estimate s, where the mean of the expected number of species in pooled samples was plotted with 95% confidence ranges after 50 randomizations with replacement (Colwell et al., 2004). Significant differences in richness among the three host species were tested using anova. Simpson's and Shannon's indices were used to analyse the diversity of ECM fungi on each host species. Similarities in the ECM fungal communities on Betula and Salix, and on Larix and Salix, were evaluated using the Sørensen similarity index after abundance-based estimation for unobserved species using estimate s (Chao et al., 2005).

Because the three host species examined in this study belong to three different families, ECM fungal species that appeared on two or three host species were regarded as generalists. ECM fungi shown to colonize Salix in previous studies were also included in determining host specificity (Nara et al., 2003b). In contrast to generalists, ECM fungi that appeared on only one host species were separated into two groups, specialists and unknowns. The specialists comprised species the specific associations of which have been phylogenetically demonstrated in other studies (Molina et al., 1992; Kretzer et al., 1996; Bruns et al., 2002; Grubisha et al., 2002; den Bakker et al., 2004). The remaining ECM fungi were included in the unknown category. Because the frequency of each ECM fungus in the unknown category was low, their occurrence on only one host species may be an effect of limitations in sample size (Horton & Bruns, 2001). To test for significant differences in ECM fungal composition (generalists vs specialists) among the three host species, the exact χ2 test was performed using spss 11.5 (using the exact test option).


Spatial pattern of timber species establishment

There were 509 vegetation patches >1 m in smallest diameter. The total area of vegetation patches (Atotal) was 12 743 m2, indicating that 6% of the ground was covered by vegetation. Of the 509 vegetation patches, 161 contained Salix (Fig. 1). The total area (ASalix) covered by vegetation patches containing Salix was 8461 m2.

In total, 39 individuals of Betula were found in 24 vegetation patches (Fig. 1). All these were accompanied by early-established Salix shrubs. To test for a significant effect of Salix presence on Betula establishment, I assumed the null hypothesis that Betula individuals had established evenly in all vegetation patches, but were detected only in Salix-containing patches by chance. The probability of this null hypothesis was as follows: PBetula = (ASalix/Atotal)N, where N represents the number of Betula individuals. Note that ASalix was used instead of the total area of Salix cover because Salix roots usually extend far from the above-ground cover, and sometimes spread through the whole area of a vegetation patch. Because N = 39 and ASalix/Atotal = 0.664, PBetula was 0.00000012. Thus the null hypothesis was rejected (P < 0.001).

Twenty-six individuals of Larix were found in 20 vegetation patches that contained Salix (Fig. 1). The same probability analysis was conducted, and the null hypothesis for Larix was also rejected (PLarix = 0.000024).

Another ECM timber species, Pinus densiflora, also appeared in the research area. Only two individuals were found, and both were accompanied by early-established Salix.

Salix that had recruited secondarily colonizing tree species ranged from 0.03 to 89 m2 in coverage area. These Salix shrubs were substantially larger than the recruited secondary colonizer individuals in all cases, indicating the earlier establishment of Salix. Salix shrubs that were accompanied by secondary colonizers were significantly larger than those lacking secondary colonizers (P = 0.006, exact Mann–Whitney test). Vegetation patches that contained secondary colonizers were significantly larger than those containing no secondary colonizers (92.7 ± 11.5 vs 19.9 ± 1.5 m2, respectively; P < 0.001, exact Mann–Whitney test).

ECM fungal communities

Molecular identification of ECM tips on 54 root systems found 36 fungal species from the three host species (Table 1; Appendix). The total species richness values estimated using Chao2 and bootstrap methods were 37.3 (36.2–44.9, 95% confidence range) and 39.9, respectively. These estimated values of species richness in the study area were close to from 36 (the actual number of species identified). The species-accumulation curve started to level off, indicating that most ECM fungal species were detected (Fig. 2). The species richness per root sample was not significantly different among host species (P = 0.312; Table 1).

Table 1.  Species richness (S) of ectomycorrhizal (ECM) fungi colonizing three host species during early primary succession on Mount Fuji, Japan
HostsECM tips examinedObserved speciesSpecies per sample*Estimated S by Chao2Estimated S by bootstrap
  • *

    Number of ECM species detected in a root system shown as mean ± SE.

  • Estimated values at the actual number of samples after 1000 randomization without replacement using estimate s ver. 7.5.

  • ns, Not significantly different among the three host species (P = 0.312, anova).

Betula ermanii2581263.2 ± 0.3 ns26.028.0
Larix kaempferi2481232.8 ± 0.3 ns23.125.1
Salix reinii1636213.4 ± 0.2 ns21.122.8
Total6698363.1 ± 0.237.339.9
Figure 2.

Species-accumulation curve for ectomycorrhizal (ECM) fungi in the early successional volcanic desert on Mount Fuji, Japan. The mean of the accumulated number of expected species in pooled samples was plotted with 95% confidence ranges after 50 randomizations with replacement using estimate s ver. 7.5. Because most ECM fungi were shared among the three host species, all samples (root systems) of the three host species were combined.

Twenty-one ECM fungal species were found associated with Salix (Table 1). Most of these were also detected in a previous study (Nara et al., 2003b), whereas four additional species were found in the present study (Table 2).

Table 2.  Relative abundance (RA) and frequency of ectomycorrhizal (ECM) fungi colonizing three host plants in the early successional volcanic desert on Mount Fuji, Japan
ECM fungal species*Betula ermaniiLarix kaempferiSalix reinii
RA (%)Frequency (/19)RA (%)Frequency (/20)RA (%)Frequency (/16)
  • The relative abundance of an ECM fungus at a sampling point was pooled and used to calculate RA at all sampling points for each host species to reduce the effect of differences in the number of ECM root tips between sampling points. Frequency, number of sampling points that contained each fungus in relation to the total number of sampling points (in parentheses) for each host species.

  • *

    ECM fungi colonizing the pioneer host plant species Salix reinii show a clear successional pattern with its growth. The sere starts with first-stage fungi, followed by second- and third-stage forms. Thus all fungi in these stages were confirmed to be compatible with S. reinii. See Nara et al. (2003a, 2003b) for details. ECM fungi that did not appear in the above stages were regarded as later-stage forms.

First-stage fungi
Inocybe lacera (Fr.) Kumm.0.003.512.72
Laccaria amethystina Cooke29.4927.2922.89
Laccaria laccata (Scop. Fr.) Berk. & Br.1.642.4312.15
Second-stage fungi
Laccaria murina Imai11.145.039.15
Scleroderma bovista Fr.0.410.000.41
Third-stage fungi
Boletus cf. rubellus Krombh.0.710.000.00
Cenococcum geophilum0.000.620.00
Cortinarius decipiens (Pers. Fr.) Fr.0.720.002.72
Hebeloma leucosarx Orton0.004.531.72
Hebeloma mesophaeum (Pers.) Quél.2.155.441.51
Hebeloma pusillum Lange0.000.002.32
Hebeloma sp. 10.310.000.00
Inocybe calospora Quél.2.521.012.31
Inocybe dulcamara (Pers.) Kumm.1.420.004.11
Inocybe sp.
Inocybe sp. 21.022.442.02
Laccaria sp. 10.710.000.00
Russula pectinatoides Peck0.212.020.00
Russula sororia (Fr.) Romell2.324.230.00
Sebacina sp. 13.310.003.42
Tomentella sp. 120.261.632.43
Tomentella sp. 25.130.811.82
Unidentified D11.921.211.31
Later-stage fungi
Cortinarius sp. 10.520.510.00
Inocybe sp.
Leccinum sp. 13.820.000.00
Sebacina sp. 24.610.000.00
Suillus grevillei (Klotzsch: Fr.) Sing.0.008.720.00
Suillus laricinus (Berk.) O. Kuntze1.1116.150.00
Thelephoraceae sp. 10.003.320.00
Thelephoraceae sp.
Tomentella sp. 30.410.005.72
Tomentella sp. 42.521.810.00
Unidentified D20.000.008.12
Unidentified D30.000.613.31
Unidentified L11.010.210.00

Betula was associated with 26 ECM fungal species (Table 1). Among these, Laccaria amethystina was the most abundant, occupying 29.4% of ECM tips, followed by Tomentella sp. 1 (20.2%) and Laccaria murina (11.1%; Table 2). Laccaria amethystina was also the most frequently observed species, appearing on nine of 19 root systems (g 2). The ECM community of Betula was quite similar to that of early-established Salix (Sørensen similarity index = 0.798).

Twenty-three ECM fungal species were detected in Larix (Table 1). Laccaria amethystina was the most abundant (Table 2). Two Larix-specific fungi, Suillus laricinus and Suillus grevillei, followed L. amethystina in relative abundance (Table 2). Laccaria amethystina was the most frequently observed species, occurring in nine of 20 root samples (Table 2). As in the case of Betula, the ECM community of Larix showed high similarity to that of Salix (Sørensen similarity index = 0.646).

The majority of ECM fungal species on each host were generalists that were detected in least two plant families: 24 of 26, 20 of 23 and 18 of 21 fungal species in Betula, Larix and Salix, respectively (Table 2). The relative abundance of generalists in the total ECM tips reached 92, 70 and 85% for Betula, Larix and Salix, respectively (Fig. 3). In contrast, specialists were limited to one Leccinum and two Suillus species on Betula and Larix, respectively (Table 2). The relative abundance of specialists was significantly greater for Larix than for the other host species (P < 0.001; Fig. 3). Suillus laricinus was unexpectedly detected from a root sample of Betula that was accompanied by a Larix sapling. However, this fungus was regarded as Larix-specific (see Discussion). This is the first example in which a Larix-associated Suillus species was found to colonize a neighboring ‘nonhost’ species.

Figure 3.

The relative abundance of generalist and specialist ectomycorrhizal (ECM) fungi colonizing three host plant species during early primary succession on Mount Fuji, Japan. Generalists were ECM fungi that were compatible with two or more host plant species belonging to different families. ECM fungi observed on only one host species were classified as specialists or unknowns. See text for details.

Shannon's diversity indices for ECM fungi on Betula, Larix and Salix were 2.41, 2.53 and 2.66, respectively. Simpson's diversity indices on Betula, Larix and Salix were 6.65, 8.05 and 10.11, respectively. Although the variation in estimates of diversity was not as great when using Shannon's index, both indices suggest that ECM fungal diversity was highest on early-established Salix (on which species richness was lowest) because of higher evenness in individual abundances of ECM fungi.


ECM fungal community on secondarily colonizing timber species

In the sere of vegetation succession in the volcanic desert on Mount Fuji, the establishment of Betula and Larix is very important as an initial stage of forest formation. In a previous study, ECM fungi detected on transplanted current-year seedlings of both tree species were common to those on early-established Salix at this site (Nara & Hogetsu, 2004). Here, I found that ECM fungi on naturally established saplings of Betula and Larix, mostly >10 yr old, were still dominated by ECM fungi common to Salix. This indicates that other ECM fungal species that are associated preferentially with secondary colonizers are not dominant fungi, even a decade after their establishment.

ECM associations of current-year seedlings in this volcanic desert are mainly accomplished via extramatrical mycelia that radiate from established hosts (Nara & Hogetsu, 2004; Nara, 2006). The ECM fungal communities on naturally established saplings of the timber species were quite similar to that on Salix. This indicates that these timber species may have been connected to Salix shrubs by common mycelia of the same ECM fungus, called common mycorrhizal networks (CMNs: Newman, 1988; Simard & Durall, 2004). Because L. amethystina was the most abundant and frequent species on each plant species, CMNs of this fungus would have the highest possibility of occurrence. In a previous study, the effects of CMNs on experimentally connected Salix seedlings varied significantly among ECM fungal species (Nara, 2006). Interestingly, L. amethystina showed no positive effects on Salix seedlings via the CMN, whereas the other 10 ECM fungi improved the N absorption and growth of connected seedlings to varying extents (Nara, 2006). Whereas nutrients absorbed by ECM fungi are shared among different host species in a CMN, there is preferential transport to highly compatible hosts, rather than to less compatible hosts (Finlay, 1989). If L. amethystina is physiologically more compatible with Betula and Larix than Salix, the CMNs of this fungus would be favorable for Betula and Larix.

Host ranges of ECM fungal species during primary succession

Observations of sporocarp–host associations in the field have been used repeatedly to evaluate the ecological specificity and host ranges of a variety of ECM fungi (Molina et al., 1992 and references therein). These studies have accumulated a large amount of information on sporocarp–host specificity; however, an association at one site does not necessarily indicate the same association at other sites because field conditions may differ (Molina et al., 1992). To consider the practical effects of ECM fungi on ecological processes, ECM specificity and compatibility should be examined at each site (Harley & Smith, 1983; Molina et al., 1992). A sporocarp approach is unsuitable in most sites for various reasons, including limited sporocarp production; co-occurrence of multiple host species in close vicinity; or great disparity between sporocarp and underground ECM communities. Instead of the sporocarp approach, a molecular approach is now available to confirm ECM colonization on individual hosts species under natural settings (Gardes & Bruns, 1993; Horton & Bruns, 2001). This enables studies of the specificity of ECM fungi between co-occurring tree species belonging to the same family (Horton & Bruns, 1998; Cullings et al., 2000), and an evaluation of host ranges by comparing ECM fungi on taxonomically distant hosts (Horton et al., 1999; Kennedy et al., 2003; Richard et al., 2005).

Generalists are usually defined as ECM fungi that are compatible with various plant families. I defined generalists as ECM fungi that were compatible with at least two of the three plant families examined (Betulaceae, Betula; Pinaceae, Larix; Salicaceae, Salix). In contrast to this robust definition of generalists, the definition of specialists varies among studies, and the demonstration of specificity is usually difficult. This is partly because specific ECM associations include various levels of host specificity (species, genus or family; Molina et al., 1992). Moreover, it is difficult to determine whether a specific occurrence is caused by real specificity or insufficient sample sizes, because ECM communities are usually composed of many rare species detected in only one or a few samples (Horton & Bruns, 2001; Richard et al., 2005). Molecular phylogenetic studies of some fungal genera have demonstrated that specific ECM associations are common on related taxa within a genus that have evolved from a common ancestor (Kretzer et al., 1996; Grubisha et al., 2002; den Bakker et al., 2004). These studies would be valuable for the definition of specialists.

Although the host ranges of dominant ECM fungi after severe disturbance have not been well studied, specialists (including Rhizopogon species) often dominate the postdisturbance ECM community (Horton & Bruns, 1998; Baar et al., 1999; Bruns et al., 2002). Rhizopogon and phylogenetically related genera such as Suillus show high specificity to the Pinaceae (Molina & Trappe, 1994; Kretzer et al., 1996; Bruns et al., 2002) and adapt well to disturbances by forming dormant spore banks (Baar et al., 1999). The persistence of dormant spores of other ECM genera has not been demonstrated clearly. Thus the dominance of specialists may be restricted to areas where the Pinaceae can dominate following disturbance. Although the effect of specialist dominance on tree succession remains unknown, it should prevent the invasion of incompatible plant species. Therefore, Pinus–Rhizopogon relationships may allow both partners to continue to prosper in repeatedly disturbed sites.

In the volcanic desert on Mount Fuji, however, generalists clearly dominated the ECM community. This is completely different from specialist-dominated ECM communities after disturbance. Because there is no dormant spore bank in primary successional settings, the specialist–generalist patterns of ECM fungi may differ fundamentally between primary and secondary succession. In addition, the host-range patterns may vary with pioneer host species, where conifers are associated with specialists and broad-leaved species are associated with generalists. Although little is known about the effects of generalist dominance on succession, this work shows that the dominance of generalists contributes to ECM associations in secondary colonizing Betula and Larix, and possibly in late-successional tree species such as Fagus, Quercus, Abies and Tsuga.

The presence of a few specialists was also confirmed: Leccinum sp. 1 was found on Betula, and two Suillus spp. were found on Larix. Suillus laricinus, a notable Larix-specific fungus, was also detected from a root system of Betula that was associated with a Larix tree. Under suitable experimental conditions, especially if exogenous sugars are abundant, ECM fungi can colonize ecologically nonhost plants (Finlay, 1989; Molina et al., 1992), although such ecologically incompatible associations are not fully functional (Finlay, 1989). Thus the mycelia of S. laricinus may infect Betula roots because of carbon support from the neighboring Larix. In accordance with other studies, I defined this fungus as a specialist because its occurrence was significantly biased to Larix (P < 0.001, exact χ2 test where S. laricinus ECM root tips in Betula and Salix were pooled and compared with those in Larix), and phylogenetic studies of related taxa support its specificity (Kretzer et al., 1996; Bruns et al., 2002; Grubisha et al., 2002). Consequently, specialists (two Suillus species) represented 25% of the relative abundance of ECM root tips in Larix. Because these specialists were not shared with early-established Salix, Larix could receive exclusive benefits from these fungi. Moreover, specialists may transfer more N to hosts than generalists (Hobbie et al., 2005). Therefore, the relative contribution of specialists to timber establishment may not be proportional to their relative abundance in ECM communities.

Contribution of ECM fungi to tree succession

Under secondary successional settings, remaining host plants sometimes facilitate subsequent recolonization of ECM trees (Perry et al., 1987). This facilitation can be derived from many biotic and abiotic factors (Callaway & Walker, 1997). ECM fungal symbionts may contribute partly to establishment because ECM fungal communities on seedlings near the remaining hosts are different from those on distant seedlings (Horton et al., 1999; Kranabetter, 1999; Dickie et al., 2002a; Ashkannejhad & Horton, 2006). Many remote seedlings, however, are usually colonized by some ECM fungi, irrespective of the early-colonizing hosts in these studies. Thus the facilitated seedling establishment cannot easily be attributed to the effect of ECM fungi unless different ECM communities are experimentally shown to have different effects at each site.

In contrast to these secondary successional sites, ECM colonization itself is nearly impossible for current-year seedlings of host species in the primary successional volcanic desert on Mount Fuji, unless seedlings are accompanied by early-established Salix (Nara & Hogetsu, 2004). Furthermore, facilitated seedling establishment of Salix was solely attributable to ECM colonization in a field inoculation experiment in which all abiotic and biotic conditions were uniform, except for ECM fungi (Nara, 2006). However, the ECM contribution to vegetation succession was uncertain because established Salix did not improve the performance (aside from ECM colonization) of Betula and Larix seedlings in a transplant experiment (Nara & Hogetsu, 2004). Here, I found only 39 and 26 established individuals of Betula and Larix, respectively, in the 21 ha of the study area approx. 300 yr after the last volcanic eruption. Although the establishment of Betula and Larix appears to have accelerated in recent years, the natural establishment of both timber species appears to be too episodic to be studied using experimental approaches.

The natural establishment of Betula and Larix coincided spatially with Salix, without exception. What is the most likely mechanism that explains the observed establishment pattern? On average, secondary colonizers occurred more often in association with large Salix shrubs and in large vegetation patches. Because soil nutrient availability and organic matter content are correlated with patch development in this volcanic desert (Hirose & Tateno, 1984), soil development may be attributable partly to the observed establishment patterns. However, a very small Salix patch (0.03 m2) within a small patch (8.4 m2) was confirmed to have recruited a Larix seedling (1 yr old) where the soil had developed poorly. Thus soil development alone may not be a requisite for the establishment of timber species. Because both Betula and Larix depend obligately on ECM fungi, these timber species must be colonized by some ectomycorrhizal fungi during early establishment. Therefore, compatible and accessible ECM fungi provided by early-established Salix would be an important mechanism that potentially could explain the observed pattern of timber establishment.


I thank Mohamad Wadud for help with the field surveys, and Takahide Ishida for help with the molecular analyses. This work was supported in part by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan.


Table A1.  Ectomycorrhizal fungi identified using ITS-T-RFLP and ITS sequence analyses
Fungal speciesITS terminal fragment lengthClosest match*
ITS1F–HinfIHinfI–ITS4ITS3–4Accession no.Accession no.Speciese
  • *

    If ITS-T-RFLP patterns did not match any sporocarp sample (Nara et al., 2003a) or any species in my T-RFLP database (Nara et al., 2003b), the sequence data were compared with sequence data of known fungal species in DDBJ/EMBL/GenBank. The accession number, species name and e value of the closest match in blast search are shown for each unknown T-RFLP type. Sequences were also studied to confirm the correctness of T-RFLP analyses.

  • Accession numbers in italics are from our previous studies (Nara et al., 2003b; Nara, 2006).

Boletus cf. rubellus272–276157682–683    
Cenococcum geophilum334–335AB211277   
Cortinarius decipiens324.0307–308376    
Cortinarius sp. 1324–325314–315382–384AB244040AJ534712Cortinarius sp.e-134
Hebeloma leucosarx398336–337408–409AB211268   
Hebeloma mesophaeum401–402334–338404–409AB211272   
Hebeloma pusillum399–400334–335405–406AB211274   
Hebeloma sp. 1397–398329–331401–402    
Inocybe calospora370–371311–313378–381    
Inocybe dulcamara428–429337407–408    
Inocybe lacera387–390331–333400–404AB211269   
Inocybe sp. 1160–161304–305371–372    
Inocybe sp. 2402–403318382–383    
Inocybe sp. 3382175–176376AB244041AY751558Inocybe sp.7e-95
Laccaria amethystina387–388335–339406–412AB211270   
Laccaria laccata391–392329398AB211273   
Laccaria murina388–390328–329399–400AB211271   
Laccaria sp. 1352313385AB244042AF204814Laccaria laccata0.0
Leccinum scabrum160420489–492AB244043AY538849Leccinum scabrum0.0
Russula pectinatoides373233–234413–415AB211276   
Russula sororia374–375347419AB211275   
Scleroderma bovista283–285247–248428–433AB211267   
Sebacina sp. 1323–324321389–391AB244044AY296254Sebacinaceae sp.0.0
Sebacina sp. 2162327396–398AB244045AY112923Sebacinaceae sp.4e-52
Suillus grevillei320–321217–218423AB244046M91616Suillus grevillei0.0
Suillus laricinus225197419–421AB244047L54102Suillus laricinus0.0
Thelephoraceae sp. 1362–363202–203411–414AB089959  e-150
Thelephoraceae sp. 2360–361337402AB244048AF184744Thelephoraceae sp. 
Tomentella sp. 1358–361198–200407–413AB089960U83480Tomentella ramosissimae-162
Tomentella sp. 2358–361200404AB244049AF272913Tomentella ellisie-102
Tomentella sp. 3360335–336403–405AB244050U83482Tomentella sp.e-179
Tomentella sp. 4357–359343–345412–414AB244051U83482Tomentella sp.0.0
UD-1, Sordariaceae?177278–279343AB244052  
UD-2, Sordariaceae?229279342–343AB244053