Fine-scale distribution of pine ectomycorrhizas and their extramatrical mycelium


Author for correspondence: Ian Alexander Tel: +44 1224272697 Fax: +44 1224272703 Email:


  • • In order to clarify the functional role of individual ectomycorrhizal (EcM) fungal species in the field, we need to relate their abundance and distribution as mycorrhizas to their abundance and distribution as extramatrical mycelium (EMM).
  • • We divided each of four 20 cm × 20 cm × 2 cm slices of pine forest soil into 100 cubes of 2 cm × 2 cm. For each cube, ectomycorrhizas were identified and the presence of EMM of the EcM fungi recorded as ectomycorrhizas was determined by terminal restriction fragment length polymorphism (T-RFLP) analysis of ITS rDNA.
  • • Ectomycorrhizas and EMM of seven EcM species were mapped. Spatial segregation of mycorrhizas and EMM was evident and some species produced their EMM in different soil layers from their mycorrhizas.
  • • The spatial relationship between mycorrhizas and their EMM generally conformed to their reported exploration types, but EMM of smooth types (e.g. Lactarius rufus) was more frequent than expected. Different EcM fungi foraged at different spatial scales.


To date, analysis of EcM fungal communities has been largely based on the ectomycorrhizas themselves. However, we know from field observations (Ogawa, 1977) and laboratory experiments with defined host–fungal associations in sheet microcosms (Read & Perez-Moreno, 2003) that, for many EcM fungi, an extensive external mycelium grows out from the mycorrhiza into the surrounding substrate. This extramatrical mycelium (EMM) is an important sink for host carbon, and is the part of the association that is functionally important in water and nutrient capture. This is also where the EcM fungus interacts most intimately with the saprotrophic microflora and the fungivorous microfauna. Several estimates of the extent of EMM have been obtained in microcosms (e.g. Colpaert et al., 1992). Using ingrowth bags in the field, Wallander et al. (2001) have shown that EMM can account for 80% of the total EcM biomass (700 kg ha−1 in a pine stand) and Hogberg & Hogberg (2002) have demonstrated that the EMM is at least 30% of the microbial biomass in boreal forest soils. Taken together, these studies show that the EMM is an extensive system which consumes carbon, and which is active in nutrient and water capture. However, virtually all current information about EMM has been derived either from single species in the artificial conditions of sheet microcosms, or from hyphal assemblages of unknown composition growing into uncolonized sand cores. Information on the species composition and extent of the EMM of mixed communities of EcM fungi developing in natural substrates in the field is essential if we are to move to the next stage of understanding how the EcM symbiosis works.

It has been known for some time that there are marked differences in the amount and nature of EMM produced by different fungi. Agerer's (2001) proposal to recognize ‘exploration types’ of ectomycorrhizas, on the basis of the development and differentiation of their EMM, has provided a tool to analyse EcM fungal communities with respect to functional patterns of exploration and nutrient exploitation below-ground. Agerer's classification ranges from the smooth, or ‘contact’ exploration type characteristic of many members of the Russulaceae, to the extensive rhizomorphs of genera such as Suillus and Paxillus. These rhizomorph-forming fungi have been particularly popular for use in the microcosm studies on which so many of our notions of EcM functioning are based (Read & Perez-Moreno, 2003). They are relatively easy to culture, and amenable to laboratory manipulation. However, they are not the fungi that are found, in most studies, to form the majority of ectomycorrhizas in the field. In the field, members of the Russulaceae, Thelephoraceae or Corticeaceae usually dominate (Horton & Bruns, 2001). These fungi are difficult or impossible to culture and are rarely, if ever, used in laboratory experimentation. Consequently, extrapolation from microcosms to the field situation is bedevilled by untested assumptions. It may be that fungal species that are present in small numbers as mycorrhizas, but that form an extensive EMM, are functionally very important in carbon and nutrient cycles.

The distribution of EMM in relation to ectomycorrhizas has been studied in microcosms (Lindahl et al., 2001; Agerer & Raidl, 2004; Donnelly et al., 2004), and the distributions of EcM fungi as root tips (Tedersoo et al., 2003; Rosling et al., 2003) and EMM (Dickie et al., 2002; Koide et al., 2005a) have been measured separately in natural forest soils. However, until we know how the abundance of specific species as mycorrhizas in the field relates to their abundance as EMM, we cannot begin to address questions of the functional roles of individual species in mixed-species communities of EcM fungi.

The aim of this study was to relate the fine-scale distribution of ectomycorrhizas to that of corresponding EMM in a natural EcM community. In particular, we wanted to determine (1) whether observed spatial relationships between ectomycorrhizas and their EMM conform to their apparent exploration type; (2) whether ectomycorrhizas occur in the same substrates that their EMM exploit; and (3) whether there are differences in the scale of spatial exploration by the EMM of different EcM fungal species. Our approach was to sample forest soil in a spatially explicit manner, by extracting 20 cm × 20 cm × 2 cm slices of soil and dividing them into a matrix of 100 cubes of 2 cm × 2 cm. Within each cube, we identified ectomycorrhizas by morphotyping and molecular typing, and recorded the corresponding distribution of EMM of the same fungal species by terminal restriction fragment length polymorphism (T-RFLP) analysis of total soil DNA extracts.


Site description

Culbin Forest is located on the southern shores of the Moray Firth in the north-east of Scotland (57°38′08″ N, 03°42′07″ W). This forest was chosen as a study site because the homogenous sandy mineral horizon (old dune sand) allowed extraction of the thin (2-cm) slices of soil required for this fine-scale study. Within the forest, a uniform 20 m × 20 m plot of 125-year-old Pinus sylvestris L. was selected. The soil has a weakly developed podzolic profile composed of organic horizons [bryophyte/litter (L; c. 0–2 cm), fermentation (F; c. 2–4 cm) and humic (H; c. 4–12 cm)] over an aeolian sand mineral horizon (Gauld, 1981). There is a heavy bryophyte layer, composed primarily of Rhytidiadelphus triquetrus (Hedw.) Warnst. and Hylocomium splendens (Hedw.) B.S.G. Other than the bryophytes, there was no ground layer of vegetation.

Sampling method

Within the 20 m × 20 m plot, four vertical soil slices (20 cm × 20 cm × 2 cm) were extracted 3 m away from four randomly chosen trees, and > 3 m from any other tree, in late May. The slices included a cross-section of horizons, well into the mineral horizon. The soil was first cut with a sharp knife and guide (Fig. 1a) before the slicer was inserted (Fig. 1b). Once extracted, the front plate was removed and the slice placed in a cutting guide (Fig. 1c), where the slice was first divided into 2-cm layers (A–J) and then 2-cm cubes to yield 100 soil cubes per slice. The cubes were individually sealed in plastic bags and immediately stored in a portable freezer at −20°C.

Figure 1.

Design of sampler used to extract soil slices from the forest plot. The soil was first cut with a sharp knife and guide (a) before insertion of the slicer (b). Once extracted, the front plate was removed and the slice placed in a cutting guide (c), where the slice was divided into 20-mm cubes.

Analysis of ectomycorrhizas

After thawing, all live roots were manually extracted from soil cubes under a stereo-microscope and the mycorrhizas classified into morphotypes according to colour and morphological characteristics of the mantle, external mycelium and rhizomorphs (Agerer, 1991). Roots were determined to be living based on the health of the stele, and almost all live root tips were colonized by EcM fungi. The distinction between living and dead mycorrhizas was most problematic for black mycorrhizas formed by Cadophora finlandia (Wang & Wilcox) Harrington & McNew. Samples of each morphotype from each soil cube were photographed, and stored in 300 µl of 2% CTAB at −20°C before DNA extraction. Hyphae that detached easily from mycorrhizas and remained in the substrate when mycorrhizas were extracted from soil were treated as EMM, but no attempt was made to detach firmly adhering mycelium. The remaining root-free soil was immediately refrozen at −20°C before DNA extraction.

After freezing of the mycorrhizas in liquid nitrogen and homogenization with a micropestle, DNA was extracted from single representative ectomycorrhizas of each morphotype using a DNeasy Plant Mini Kit (Qiagen Ltd, Crawley, UK), following the manufacturer's protocol. DNA was amplified in a 50-µl reaction mix that contained 1 µl of DNA template, 1 × buffer [16 mm (NH4)2SO4, 67 mm Tris-HCl (pH 8.8 at 25°C), 0.01% Tween-20], 2.0 mm MgCl2, 250 µm dNTPs (Bioline Ltd, London, UK), 20 pmol ITS1F (Gardes & Bruns, 1993) and ITS4 (White et al., 1990), and 2.5 U BIOTAQ polymerase (Bioline Ltd, London, UK). The PCR cycle comprised an initial denaturation step at 95°C for 5 min followed by 29 cycles of 95°C for 30 s, 55°C for 30 s and 72°C for 30 s, followed by a final extension step of 72°C for 5 min on a PTC-220 DYAD™ Thermal Cycler (MJ Research Inc., Waltham, MA, USA). PCR products were purified using a magnetic bead ChargeSwitch PCR clean-up kit (Invitrogen, Paisley, UK) before endonuclease restriction digestion or sequencing. Where putative basidiomycete mycorrhizas produced more than one ITS PCR product, the DNA was amplified with the basidiomycete primers ITS1F and ITS4B (Gardes & Bruns, 1993), before a nested re-amplification with ITS1F and ITS4 to allow subsequent comparison of terminal restriction fragment (TRF) profiles.

In order to confirm morphotype categorization, RFLP patterns were produced for a subsample of mycorrhizas from each morphotype. PCR products were digested for 3 h with HinfI (at 37°C) and TaqI (at 65°C) restriction endonucleases (Promega, Southampton, UK). Each reaction contained 1 µl of PCR product (c. 100 ng of DNA), 5 units of restriction endonuclease, 1 µl of the recommended buffer and 0.1 µl of bovine serum albumin (BSA), made up to 10 µl with d.H2O. RFLP images were imported into gelcompar II V4.01 software (Applied Maths, St-Martens-Latem, Belgium) to allow UPGMA Pearson correlation coefficient cluster analysis with the automatic optimization settings of the program. Where more than one RFLP pattern was produced within a morphotype sample, DNA was extracted and restriction digests performed on all the mycorrhizas of that morphotype to allocate them to the appropriate RFLP type.

PCR products from each RFLP type were sequenced with the primers ITS1F and ITS4, using the BigDye Terminator Cycle Sequencing Kit v3.1 on an ABI PRISM™ 3130xl genetic analyser (Applied Biosystems, Warrington, UK). DNA sequences were manually checked and edited where necessary using the sequencher software package (version 3.0; Gene Codes Corporation, Ann Arbor, MI USA) and matched against sporocarp sequences from the site, using a local nucleotide database compiled with the BioEdit sequence alignment editor (Hall, 1999). Where no good match was found, sequences were submitted to the NCBI and UNITE (Kõljalg et al., 2005; online nucleotide databases, using the blastn algorithm.

To determine the TRFs of each RFLP type, DNA from each EcM RFLP type was amplified with ITS1F labelled with Fam and ITS4 with Hex fluorescent dyes. The PCR products were digested with HinfI and TaqI to produce four unique diagnostic TRFs for each EcM RFLP type. To determine the size (bp) of TRFs, between 12 and 20 ng of the digested products was added to 12 µl of Hi-Di formamide and 0.5 µl of GS-500 ROX size-standard (Applied Biosystems), denatured at 95°C for 5 min and then chilled on ice. Fragment separation was performed on an ABI PRISM™ 3130xl genetic analyser (Applied Biosystems) using POP 4 and a 50-cm column with a 15-s injection time at 1.5 kV for 40 min at 60°C. The TRF sizes were determined with genemapper V3.7 software (Applied Biosystems) (Table 1) and imported into the tramp program (Dickie et al., 2002) to provide a reference database for detection of EMM in the soil samples. For two EcM RFLP types, only three TRFs > 50 bp were produced.

Table 1.  ITS terminal restriction fragment lengths (bp) for ectomycorrhizal (EcM) fungi found as mycorrhizas within  4the soil slices
EcM fungusaHinfITaqI
Cortinarius sp.318301267305
Clavulinaceae sp. I185351286287
Lactarius rufus104377318 68
Cenococcum geophilum 61108187150
Suillus variegatus224213 77
Basidiomycete sp. I350 86298
Cadophora finlandia401263141213

Identification of EMM in soil cubes by T-RFLP

Total DNA was extracted from the root-free soil cubes. For cubes from the organic horizons, all the material was first ground with liquid N2 to homogenize the sample. A subsample of 1 g of soil was then incubated at 65°C for 45 min in 600 µl of 5% CTAB buffer. Further DNA extraction followed the method of Griffiths et al. (2000). For cubes from the sand horizon, DNA was extracted from soil following the method of Griffiths et al. (2000) using FastPrep lysing matrix-E (Qbiogene, Cambridge, UK), and macerated for 2 × 30 s at 5.5 ms−1 in a FastPrep bead beating system. The DNA was amplified with fluorescent-labelled ITS1F and ITS4 as described in the previous section, except for the addition of 1 µl of BSA to facilitate amplification, and digested with HinfI and TaqI restriction endonucleases. After fragment analysis as described in the previous section, all fragments > 50 bp, and with a peak height detection limit of > 50 absorption units, were exported to the tramp program for comparison with TRF profiles derived from the EcM RFLP types. Where more than 50 TRFs were identified for a particular sample, the TRFs were ranked by peak height and the 50 highest exported to tramp. Positive identification of EMM from one of the EcM RFLP types was confirmed if all available TRFs (three or four) were detected within an error of 1.5 bp.

Statistical analysis

Throughout the following text, ‘frequency’ describes the number of soil cubes that contained mycorrhizas or EMM.

Two-dimensional analysis of EMM distributions was conducted by calculating conspecific neighbour frequency distributions. Two approaches were taken: either equal weight was given to each cube, by calculating the proportion of neighbour cubes occupied by conspecific EMM (to minimize edge effects), or greater weight was given to cubes in the centre of slices, by using absolute conspecific neighbour frequency.

To test for differences in the vertical distribution of ectomycorrhizas and EMM of a given species, frequency data were analysed by a general linear model with genstat 7 statistical software (VSN International Ltd, Hemel Hempstead, UK). The model used the ‘logit’ link function to test the mean deviance for row × type against mean deviance for row × type × slice, where ‘type’ was either EMM or mycorrhizas of a given species. This analysis was only conducted where both ectomycorrhizas and EMM of the target species occurred in all four slices.

To test for nonrandom distributions of EMM and mycorrhizas between soil layers (A–J), χ2 tests were performed between observed frequencies and those expected assuming a random distribution across all layers.


In total, 1775 EcM tips were extracted from the slices. Total root-tip abundance was greatest in the bryophyte/litter and F horizons and decreased rapidly with depth, with no ectomycorrhizas found in the lowest layers of the sand horizon (Fig. 2). For all RFLP types, both ectomycorrhizas and EMM were distributed nonrandomly in relation to soil layers (χ2 tests; P < 0.001 for all species).

Figure 2.

Total ectomycorrhizal (EcM) fungal tip abundance distribution with depth.

Seven EcM RFLP types were recognized, and named after their closest ITS sequence match (Table 2). Basidiomycete sp. I had a poor match (< 80%) with reference database sequences, but was most closely related to ectomycorrhizal sequences from Clavulinaceae. Cenococcum geophilum Fr. was the most frequent species as mycorrhizas, whilst C. finlandia was the most frequent as EMM (Fig. 3).

Table 2.  Database matches of ITS sequences from ectomycorrhizas
Tip nameSize (bp)Sequence matchDatabaseSimilaritya
  • a

    Per cent similarity over the entire sequence length.

Cortinarius sp. (AM109900)606Cortinarius sp. (AM109898)Local 99
Clavulinaceae sp. I (AM109903)669Clavulinaceae sp. (AJ534710)NCBI 99
Lactarius rufus(AM109901)725Lactarius rufus(AM109897)Local100
Cenococcum geophilum(AM109905)963Cenococcum geophilum(DQ179119)NCBI 98
Suillus variegatus(AM109899)693Suillus variegatus(UDB000170)UNITE 99
Basidiomycete sp. I(AM109902)694No direct match(see text)
Cadophora finlandia(AM109904)775Cadophora finlandia(AJ534704)NCBI 99
Figure 3.

Frequency distribution of ectomycorrhizas (grey bars) and extramatrical mycelium (EMM; black bars) ranked by tip frequency.

EcM tip and EMM distribution patterns

Cortinarius sp.  Together, mycorrhizas and EMM of Cortinarius sp. were found in 13% of the soil cubes sampled. Mycorrhizas were found in three of the slices, but EMM of the fungus was in all four (Fig. 4). The frequency of EMM was much greater than that of mycorrhizas (F1,6 = 10.86; P < 0.005) (Table 3). Mycorrhizas were only found between 2 and 8 cm depth, in F and H horizons (Fig. 5). EMM was found at all depths, but was most frequent in the bryophyte/litter horizon and F horizons between 0 and 4 cm, i.e. above the main occurrence of the mycorrhizas themselves. No EMM was detected in three of the seven cubes that contained EcM tips (Table 3). Each cube that contained Cortinarius sp. EMM had on average 2.4 neighbouring cubes containing the same fungus. Whilst 18% of the cubes that contained EMM of this species had no neighbouring cubes containing the same species, the modal proportion of conspecific neighbours was 0.4 (Fig. 6). The maximum patch size of this species (EMM and/or mycorrhizas) was 13 cubes (104 cm3).

Figure 4.

(a) Photographs of slices with distribution of ectomycorrhizas (circles) and corresponding extramatrical mycelium (filled squares) of (b) Cortinarius sp., (c) Clavulinaceae sp. I, (d) Lactarius rufus, (e) Cenococcum geophilum, (f) Suillus variegatus, (g) Basidiomycete sp. I and (h) Cadophora finlandia. Open squares identify cubes for which extramatrical mycelium (EMM) data were not obtained.

Table 3.  Total number of soil cubes occupied by each ectomycorrhizal (EcM) fungus as extramatrical mycelium (EMM) or mycorrhizas
 EMM and tipsEMM onlyTips onlyTotal
  1. Total number of cubes = 400.

Cortinarius sp. 446 353
Clavulinaceae sp. I11334893
Lactarius rufus 3102639
Cenococcum geophilum 7265790
Suillus variegatus 0 1 910
Basidiomycete sp. I 110 314
Cadophora finlandia 3701992
Figure 5.

Depth distribution (summed across all slices; maximum frequency = 40) of ectomycorrhiza (grey bars) and extramatrical mycelium (black bars) frequency from row A to J (0–200 mm) (Fig. 4) for each species of ectomycorrhizal (EcM) fungus.

Figure 6.

Histogram of proportion of adjacent soil cubes that contain the same species of extramatrical mycelium for selected ectomycorrhizal fungi.

Clavulinaceae sp. I   Together, mycorrhizas and EMM of Clavulinaceae sp. I were found in 23% of the soil cubes sampled, and both were found in all four slices (Fig. 4). The frequency of mycorrhizas was slightly greater than that of EMM (Table 3). Almost all mycorrhizas were found between 2 and 12 cm depth, in F and H horizons, but EMM was found at all depths (Fig. 5). The vertical distribution of EMM differed from that of mycorrhizas (F9,27 = 6.59; P < 0.001) and no EMM was detected in 81% of the cubes that contained mycorrhizas (Table 3). Each cube that contained EMM of Clavulinaceae sp. I had a mean of 1.6 neighbouring cubes containing the same species. However, a third of the cubes that contained EMM of Clavulinaceae sp. I had no conspecific neighbours, and the conspecific neighbour distribution was bimodal, with a second modal proportion of 0.4 (Fig. 6). The maximum patch size of this species (EMM and/or mycorrhizas) was 31 cubes (248 cm3).

Lactarius rufus (Scop.) Fr.   Although L. rufus had the greatest absolute mycorrhiza abundance of any species present in the slices (699 tips), together, mycorrhizas and EMM of L. rufus were found in only 10% of the soil cubes. Mycorrhizas were present in all four slices, but EMM was only found in two slices (Fig. 4). Twice as many cubes contained mycorrhizas as contained EMM (Table 3). Mycorrhizas were found between 0 and 6 cm depth in L, F and upper H horizons, but were most abundant between 0 and 2 cm, in the bryophyte/litter horizon (Fig. 5). EMM was also mainly found between 0 and 6 cm; however, some EMM was detected in the sand horizon. No EMM was detected in 90% of the cubes that contained mycorrhizas (Table 3). Each cube that contained L. rufus EMM had a mean of 2.0 neighbours which also contained the fungus. Seventeen per cent of the cubes that contained EMM had no conspecific neighbours, and the modal proportion of conspecific neighbours was 0.4 (Fig. 6). The maximum patch size of this species (EMM and/or mycorrhizas) was 12 cubes (96 cm3).

Cenococcum geophilum  Together, mycorrhizas and EMM of C. geophilum were found in 23% of the soil cubes sampled. Mycorrhizas and associated EMM were found in all slices (Fig. 4). Mycorrhizas were almost twice as frequent as EMM (Table 3) and had a different depth distribution (F9,27 = 7.58; P < 0.001). Mycorrhizas occurred between 0 and 10 cm depth, but were most frequent in the F horizon (Fig. 5). The EMM was concentrated in the L and F horizons between 0 and 10 cm, and again in the sand horizon below the main occurrence of the mycorrhizas themselves. These deeper EMM occurrences were only detected in two slices. No EMM was detected in 89% of the cubes that contained ectomycorrhizas (Table 3). Each cube that contained C. geophilum EMM had a mean of 0.63 neighbours which also contained the fungus, and 52% of the cubes that contained EMM had no conspecific neighbours (Fig. 6). The maximum patch size of this species (EMM and/or mycorrhizas) was 33 cubes (264 cm3).

Suillus variegatus (Fr.)  Tubercules of S. variegatus were infrequent, and mycorrhizas and EMM of the fungus were found in only 4% of the soil cubes sampled. Tubercules were found in all slices, but EMM of the fungus was only detected in one cube from slice two (Fig. 4). Mycorrhizas were found between 0 and 14 cm depth, in L, F and H horizons (Fig. 5). No EMM was detected in the cubes that contained mycorrhizas (Table 3). The maximum patch size of this species (EMM and/or mycorrhizas) was two cubes (16 cm3).

Basidiomycete sp. I  Together, mycorrhizas and EMM of Basidiomycete sp. I were found in 4% of the soil cubes sampled. All the mycorrhizas and > 80% of the EMM were found in one slice (Fig. 4). The frequency of EMM was almost three times greater than that of mycorrhizas (Table 3). Mycorrhizas were only found between 6 and 12 cm depth, in the H horizon (Fig. 5). EMM was found between 0 and 20 cm, but was less frequent in the L and F horizons, between 0 and 4 cm. No EMM was detected in three of the four cubes that contained mycorrhizas (Table 3). Each cube that contained Basidiomycete sp. I EMM had a mean of 0.7 neighbouring cubes containing the same species; however, most occurrences had no conspecific neighbours (Fig. 6). The maximum patch size of this species (EMM and/or mycorrhizas) was seven cubes (56 cm3).

Cadophora finlandia  Together, mycorrhizas and EMM of C. finlandia were found in 23% of the soil cubes sampled. Analysis of root tips indicated that living mycorrhizas of this fungus were present in all slices, but EMM was only found in two slices (Fig. 4). EMM occurred almost twice as frequently as the ectomycorrhizas (Table 3) but had a similar depth distribution. Ectomycorrhizas and EMM occurred between 0 and 20 cm depth, but both had greatest frequencies in the H horizon (Fig. 5). No EMM was detected in 86% of the cubes that contained ectomycorrhizas (Table 3). Each cube that contained C. finlandia EMM had a mean of 3.61 neighbouring cubes containing the same species, and only two of the cubes that contained EMM had no conspecific neighbours (Fig. 6). The maximum patch size of this species (EMM and/or mycorrhizas) was 38 cubes (312 cm3).



By sampling in slices we have, for the first time, described the fine-scale two-dimensional relationship between mycorrhizas and their EMM. We have successfully used T-RFLP for high-throughput identification of the EMM of individual species from total community DNA extractions. In contrast to recent studies (e.g. Koide et al., 2005a,b), we have used more stringent criteria to confirm the presence of a species. This gives us high confidence in our positive identifications, but may contribute to the failure to detect EMM in all cubes where a particular species occurs as mycorrhizas. Because we required all TRFs to be present for positive identification, the chance of losing one of the TRFs below the detection limit was relatively high. It must also be noted that T-RFLP analysis detects a limited amount of nucleotide variation and may not be able to separate some cryptic species groups, such as some Cortinarius spp. (Kårén et al., 1997; Horton, 2002; Edwards & Turco, 2005).

The vertical sampling scheme adopted is much more likely to pick up extension of EMM from mycorrhizas in the vertical plane. The large sample numbers necessitated restriction of our study to two dimensions. We decided to sample vertical slices because we had no a priori knowledge of the depth at which most EMM would be found. However, the results suggest that the layered nature of the substrate is more likely to encourage horizontal extension of EMM and a similar analysis, using horizontal slices, would be an interesting addition to the data presented here.

Vertical distribution of mycorrhizas

This study confirms previous observations of spatial segregation of EcM fungi, particularly between soil horizons (Heinonsalo et al., 2001; Rosling et al., 2003; Tedersoo et al., 2003). Mycorrhizas of L. rufus occurred predominantly in the L horizon, those of Cortinarius sp. in the upper H and F horizons, those of Basidiomycete sp. I in the lower H horizon and those of S. variegatus and Clavulinaceae sp. I throughout the H horizon. Mycorrhizas of C. geophilum and C. finlandia were found in all the organic horizons. Rosling et al. (2003) found two-thirds of mycorrhizas (including a Suillus sp.) in the mineral horizons of a northern Swedish boreal forest soil and half the EcM species were restricted to these horizons. That was not the case here, possibly because of the short history of soil-profile development in our site with limited mineral weathering and migration of organic matter in the inorganic horizons. The surface organic horizons are almost the only source of nutrient elements for trees in the Culbin ecosystem (Miller et al., 1979). In addition, our samples were relatively shallow, and it is possible that more roots and more EcM fungal species would have been found had we gone deeper, towards the water table.

Vertical distribution of EMM

The frequencies of EMM of the seven EcM fungi were less clearly different at different depths than were the frequencies of the mycorrhizas themselves. EMM of Cortinarius sp. and L. rufus occurred mainly in the upper bryophyte/litter, F and upper H horizons, although both were also detected in the sand. This preferential association of the EMM of a group of EcM with the lower L horizon has been demonstrated in Pinus resinosa stands (Dickie et al., 2002). The EMM of Clavulinaceae sp. I, Basidiomycete sp. I, C. finlandia and C. geophilum was distributed more uniformly across all depths. Although it was not possible to identify Clavulinaceae sp. I to species, it may be significant that Clavulina cristata (Holmsk.) J. Schrot. was also found to have a relatively uniform distribution across litter, F and H horizons (a so-called multilayer species) by Dickie et al. (2002).

Mycorrhizas of the two ascomycete fungi in this study (C. geophilum and C. finlandia) were mainly found in the organic horizons; however, EMM of both these species was found throughout the sand horizon. C. geophilum is known to be drought tolerant (Jany et al., 2003) and the ability of its EMM to forage down into potentially moister soil may be one mechanism that contributes to this tolerance. Tedersoo et al. (2003) found a high frequency of ascomycete mycorrhizas in mineral soil.

While we have demonstrated some of the same vertical stratification patterns of mycorrhizas and EMM found in previous studies, it is likely that particular site conditions are important in determining the distributions of species, and many more studies are required before any general principles emerge.

Spatial relationships between EMM and mycorrhizas

This is the first study to compare the fine-scale spatial distribution of ectomycorrhizas with their associated EMM in the field. Cortinarius sp. produced the majority of its EMM in the L and F horizons, whereas the few ectomycorrhizas found were deeper, in the upper H horizon. Also, the EMM of this Cortinarius sp. were frequently found in soil cubes that did not contain its mycorrhizas. According to Agerer's (2001) classification of EcM explorations types, Cortinarius spp. usually have medium-distance fringe rhizomorphs (type A) and this is consistent with the EMM distribution pattern observed in our study.

For most species, EMM exhibited a greater depth range than their associated mycorrhizas. This pattern might be expected where the fungal partner is able to utilize and compete within a wide range of substrates, but root-tip distribution is limited by the availability of new root tips. New root-tip availability is determined by a combination of distribution of long roots and direct competition with other EcM fungi for new tips. Competition for root tips can be an important determinant of EcM community structure; for example, in a microcosm experiment, Kennedy & Bruns (2005) have demonstrated priority competition for uncolonized root tips by two species of Rhizopogon.

Cortinarius sp. and L. rufus coexisted as EMM but appeared to be more segregated as mycorrhizas. This may demonstrate competition for root tips by the two fungi. In contrast, Cortinarius sp. and Clavulinaceae sp. I tip distributions overlapped but they were segregated as EMM (χ2 = 9.188; P < 0.005). There are relatively few examples of competitive exclusion between EcMs (Wu et al., 1999; Kennedy & Bruns, 2005). Koide et al. (2005b) detected a negative correlation between the occurrence of C. geophilum and Clavulina cinerea, both as mycorrhizas and as EMM in a Pinus resinosa plantation. However, it will always be difficult to prove competition between EcM fungi from field observations because these same negative correlations may also result from different resource acquisition and survival strategies, for example substrate preference by EMM, avoidance of environmental fluctuations or ability of EcM fungi to differentially utilize substrates in soil layers not occupied by pine roots. Perhaps the deeper distribution of Cortinarius sp. mycorrhizas in relation to EMM reflects a strategy to avoid loss of valuable tips, in dry or cold periods, while still allowing exploration of the active F horizon as EMM. This study focused on the EcM component of the soil fungal community. In terms of mycelial interactions, competition between the EcM and saprotrophic fungal hyphae has been demonstrated (Lindahl et al., 2001) but we do not know how direct competition with non-EcM fungi influenced the differences in EMM distribution observed here.

As noted above, EMM was not detected in many of the soil cubes in which the mycorrhizas were found. The Clavulinaceae sp. I, L. rufus and Basidiomycete sp. I mycorrhizas were all relatively smooth morphotypes with little or no obvious EMM. It is possible that any EMM attached to the mycorrhizas was removed with them when they were extracted from the soil. Even if EMM was not removed with mycorrhizas, low densities of EMM proximal to the mycorrhizas may have been below detection limits following PCR amplification in competition with all other soil fungi. EMM of each of these ‘smooth-type’ mycorrhizas was, however, detected in some cubes, irrespective of the distribution of mycorrhizas. It has generally been assumed that ‘smooth’ or ‘contact’ type mycorrhizas do not produce extensive EMM. However, our spatial data, for example for L. rufus, suggest that EMM can be found distant from mycorrhizas. Of course, we cannot be sure that the cubes with EMM but no mycorrhizas did not have mycorrhizas immediately adjacent to them, but on average we might expect that the nearest tip was at least 1 cm away.

EMM of S. variegatus was not detected in any of the cubes that contained tubercules, and indeed was only detected once in the 390 cubes examined. This is surprising given that Suillus tip clusters have, by Agerer's (2001) classification, a long-distance exploration strategy and produce highly visible thick rhizomorphs. However, we observed that only one such rhizomorph was attached to each tubercule. This was firmly attached to the tubercule and was removed along with the mycorrhizas when roots were separated from the soil, and this would explain the lack of EMM in cubes containing tubercules. In microcosms, Suillus rhizomorphs extend away from the mycorrhiza and hyphal fans proliferate when suitable substrates are found (Bending & Read, 1995); however, it seems that none of the field substrate slices studied here included such a proliferation region. Another possibility is that hyphal fans seen in microcosms are ephemeral structures that are only produced at certain times of the year. We may have missed the hyphal fan period in our one-off sample event.

Patch size

Neighbour analysis demonstrated two general patterns of EMM distribution across the slices. Soil cubes in which EMM of Cortinarius sp., L. rufus, Clavulinaceae sp. I and C. finlandia occurred had a mean of two or more conspecific neighbours and there were many examples where almost half the neighbouring cubes were of the same species. Cubes in which EMM of C. geophilum and Basidiomycete sp. I occurred had a mean of less than one conspecific neighbour and most cubes had no conspecific neighbours. Because all cubes that contained EMM were in effect at the slice boundaries, it was not possible to measure the absolute size of individual EMM patches, even in the two-dimensional analysis our sampling scheme permits. However, Cortinarius sp., Clavulinaceae sp. I, L. rufus and C. finlandia occurred in patches of up to 10, 9, 11 and 38 contiguous cubes, respectively, whilst C. geophilum and Basidiomycete sp. I EMM occurred in patch sizes of only up to 5 and 3 contiguous cubes, respectively. Patch size may differ in the horizontal plane; however, this suggests that the EMM of the first group of EcM fungi forage at a larger spatial scale than those of the latter group. These inferred foraging scales were expected in some cases (e.g. large EMM patches for the rhizomorphic Cortinarius sp.); however, the scale of the EMM of other species (e.g. the contact-type L. rufus and Clavulinaceae sp. I) did not appear to be directly related to the apparent foraging strategy inferred by visual observations of the mycorrhizas themselves. Understanding the distribution of EMM in the real world brings us a step closer to understanding the functional importance of individual EcM fungal species within a community.


We thank Pamela Parkin, Leanne Reid and Allan Wilson for their technical assistance and the Forestry Commission for access to Culbin Forest. We thank Liz Holden for her help with sporocarp identification, David Elston (BioSS, UK) for statistical advice and Daniel Lou-Hing for his help with sorting roots. This work was funded by the Natural Environment Research Council (grant: NER/A/S/2002/00861) and the Scottish Executive Environmental and Rural Affairs Department (SEERAD).