ArticlePDF Available

Diversity and fruiting patterns of ectomycorrhizal and saprobic fungi in an old-growth Mediterranean forest dominated by Quercus ilex L

December 2004
Canadian Journal of Botany 82(12):1711-1729

December 2004
82(12):1711-1729

DOI:10.1139/b04-128

Authors:

Franck Richard

Université de Montpellier

Pierre-Arthur Moreau

Université de Lille

Marc Selosse

Muséum National d'Histoire Naturelle

Monique Gardes

Paul Sabatier University - Toulouse III

We collected and mapped epigeous fruitbodies of both ectomycorrhizal (ECM) and saprobic fungi in an old- growth Quercus ilex L. Mediterranean forest within a permanent transect of 6400 m2 over three consecutive fruiting seasons. Out of 5382 fruitbodies, a total of 234 species were found, including 166 and 68 ECM and saprobic taxa, re- spectively. Both communities were mainly composed of rare species. Two genera, Russula and Cortinarius, accounted for 34.4% of ECM fruitbodies and 50% of species diversity. The three most abundant ECM species were Laccaria laccata (Scop.: Fr.) Berk. & Broome, Inocybe tigrina R. Heim, and Lactarius chrysorrheus Fr. The fruiting ECM com- munity encompassed a few Mediterranean species and numerous broad host range temperate species. We also analysed the fruiting patterns in relation to forest structure, host composition, and natural canopy gaps. The results showed (i )a significant correlation of species richness to tree density, (ii) a richness decrease as the number of vegetation layers in- creases, and (iii) a preferential fruiting of some species near Q. ilex or Arbutus unedo L. Another noteworthy feature was that richness and production were greatly enhanced in canopy gaps. Selective fruiting was also observed among species. These results highlight the importance of forest structure and large woody debris for fungal conservation.

Spatial patterns of fruitbodies using the Besag function (Besag and Diggle 1977).

…

Production and species richness at various distances from individuals of Quercus ilex (n = 108) and Arbutus unedo (n = 227).

…

Figures - uploaded by Marc Selosse

Content may be subject to copyright.

Content uploaded by Marc Selosse

Content may be subject to copyright.

Diversity and fruiting patterns of ectomycorrhizal

and saprobic fungi in an old-growth Mediterranean

forest dominated by Quercus ilex L.

F. Richard, P.-A. Moreau, M.-A. Selosse, and M. Gardes

Abstract: We collected and mapped epigeous fruitbodies of both ectomycorrhizal (ECM) and saprobic fungi in an old-

growth Quercus ilex L. Mediterranean forest within a permanent transect of 6400 m2over three consecutive fruiting

seasons. Out of 5382 fruitbodies, a total of 234 species were found, including 166 and 68 ECM and saprobic taxa, re-

spectively. Both communities were mainly composed of rare species. Two genera, Russula and Cortinarius, accounted

for 34.4% of ECM fruitbodies and 50% of species diversity. The three most abundant ECM species were Laccaria

laccata (Scop.: Fr.) Berk. & Broome, Inocybe tigrina R. Heim, and Lactarius chrysorrheus Fr. The fruiting ECM com-

munity encompassed a few Mediterranean species and numerous broad host range temperate species. We also analysed

the fruiting patterns in relation to forest structure, host composition, and natural canopy gaps. The results showed (i)a

significant correlation of species richness to tree density, (ii) a richness decrease as the number of vegetation layers in-

creases, and (iii) a preferential fruiting of some species near Q. ilex or Arbutus unedo L. Another noteworthy feature

was that richness and production were greatly enhanced in canopy gaps. Selective fruiting was also observed among

species. These results highlight the importance of forest structure and large woody debris for fungal conservation.

Key words: ECM community, saprophytic fungi, holm oak, macromycete fruiting patterns, canopy gaps, fungal conser-

vation.

Résumé : Nous avons récolté et cartographié les fructifications épigées de champignons ectomycorhiziens (ECM) et

saprophytiques dans une vieille forêt méditerranéenne de Quercus ilex L. au sein d’un transect permanent de 6400 m2,au

cours de trois saisons consécutives de fructification. À partir de 5382 fructifications, nous avons reconnu 234 espèces in-

cluant 166 espèces ectomycorhiziennes et 68 saprophytiques. Les deux communautés étaient principalement composées

d’espèces rares. Les genres Russula and Cortinarius représentaient 34.4% des fructifications et 50% de la diversité ecto-

mycorrhizienne. Les trois espèces ectomycorhiziennes les plus abondantes étaient Laccaria laccata (Scop.: Fr.) Berk. &

Broome, Inocybe tigrina R. Heim et Lactarius chrysorrheus Fr. La communauté ectomycorhizienne était constituée de

quelques espèces à affinités méditerranéennes et de nombreuses espèces tempérées à large spectre d’hôtes. Nous avons

également examiné les relations entre les patrons de fructification et la structure forestière, la composition en hôtes et la

présence de trouées naturelles. Les résultats ont montré: (i) une corrélation significative entre la richesse et la densité des

arbres, (ii) une diminution de la richesse en fonction du nombre de strates de végétation et (iii) une fructification plus

importante de certaines espèces près de Q. ilex ou d’Arbutus unedo L. Une autre caractéristique remarquable était que la

richesse et la production étaient plus élevées dans les ouvertures de canopée. Une fructification préférentielle de certaines

espèces était aussi observée dans ces ouvertures. Ces résultats mettent en évidence l’importance de la structure forestière

et des bois morts de grande dimension pour la conservation in situ des champignons.

Mots clés : communautés ectomycorhiziennes, champignons saprophytiques, chêne vert, patrons de fructification des

macromycètes, trouées naturelles, conservation in situ des champignons.

Richard et al. 1729

1711

Received 5 January 2004. Published on the NRC Research Press Web site at http://canjbot.nrc.ca on 8 December 2004.

F. Richard1,2 and M. Gardes.3Unité mixte de recherche 5174 Évolution et Diversité Biologique, Université Toulouse III Paul

Sabatier, 118 Route de Narbonne, 31062 Toulouse CEDEX 4, France.

P.-A. Moreau.4Geobotanische Institut ETH, Zollikerstrasse 107, CH 8008 Zürich, Switzerland.

M.-A. Selosse.3,5 Systématique, adaptation et évolution, Centre national de la recherche scientifique, Université Paris VI, Muséum

national d’histoire naturelle, 43, rue Cuvier, 75005 Paris, France.

1Corresponding author (e-mail: richard.fran@wanadoo.fr).

2Present address: Office National des Forêts, 20250 Corte, France.

3These two authors have contributed equally to the supervision of this work.

4Present address: Laboratoire de Botanique, Faculté des Sciences Pharmaceutiques et Biologiques, Lille II, 3, rue du Professeur

Laguesse, B.P. 83, F-59006 Lille CEDEX, France.

5Present address: Centre d’Ecologie Fonctionnelle et Evolutive, Centre National de la Recherche Scientifique, Unité mixte de

recherche 5175, Equipe co-évolution, 1919, route de Mende, 34 293 Montpellier CEDEX, France.

Introduction

Old-growth forests are rare in Europe, particularly in the

Mediterranean basin, because of ancient anthropic pressures

(Quézel and Médail 2003). These forest ecosystems are gen-

erally characterised by (i) a high diversity of ligneous and

herbaceous plants species, (ii) numerous old and large trees,

and (iii) a mosaic pattern of trees at various stages of devel-

opment. In addition, one of their most noteworthy architec-

tural features is the presence of numerous gaps resulting

from tree falls (Oldeman 1990). These small-scale canopy

gaps play a key role in natural regeneration of unmanaged

forests (Connell 1978; Mc Carthy 2001).

Superimposed on this complex structure, old-growth for-

ests also harbour a high diversity of interacting plants, ani-

mals, and fungi (Berglund and Jonsson 2001). Saprobic and

ectomycorrhizal (ECM) fungi generally form species-rich

guilds that interact with trees (Dahlberg et al. 1997; O’Dell

et al. 1999; Smith et al. 2002). Thus, saprobic fungi are pri-

mary decomposers of dead organic plant material and ECM

fungi develop symbiotic structures, ectomycorrhizae, on fine

root tips of numerous tree species (Smith and Read 1997).

Identification of the fungi is based almost exclusively on

their sexual reproductive structures (commonly referred to as

mushrooms, fruitbodies, or sporocarps).

Ectomycorrhizal fungi are taxonomically diverse (>5000

species) and belong to several clades in the Ascomycotina

and Basidiomycotina (Smith and Read 1997). A large major-

ity of species produce fruitbodies that are visible above

ground, with the naked eye. However, there are ECM fungi

(e.g., Cenococcum geophilum) for which the sexual stage

has not been observed, and others that form inconspicuous

fruitbodies (e.g., resupinate Thelephoraceae and

Sebacinaceae, and truffle-like fungi). Most studies of ECM

fungal communities are based on fruitbody surveys and on

molecular investigations of diversity from mycorrhizal

rootlets. Some reports allow a comparison between fruiting

patterns and ectomycorrhizae. These two approaches show

different and sometimes contrasting results (Gardes and

Bruns 1996; Horton and Bruns 2001). For example, species

that are well represented in the fruiting record are rare below

ground and reciprocally. Despite their limitations, fruitbody

surveys are a useful way to assess the presence of species in

a stand because many ECM fungi produce fruitbodies that

can be easily mapped and identified to the species level. In

addition, fruiting patterns still provide valuable tools to for-

est practitioners interested in the management and conserva-

tion of nonwoody resources such as edible mushrooms.

Factors driving fungal diversity remain unclear, but com-

position and demographic features of the host tree popula-

tion (i.e., changes in ecosystem structure) are likely to act,

particularly on symbiotic fungi (Smith and Read 1997).

ECM fungal species differ in their host breadth, ranging

from narrow to broad spectra (Molina et al. 1992). Similarly,

plants vary widely in their ability to associate with fungal

species. As a result, host composition can shape the diversity

of the fungal guild (Villeneuve et al. 1989; Nantel and

Neumann 1992; Vogt et al. 1992). Fungal richness and com-

position also change in relation to forest succession (Keizer

and Arnolds 1994; Smith et al. 2002). Some ECM species

(referred to as early-stage fungi; Last et al. 1987) establish

early and stop fruiting in old stands, although they can

persist on roots (Smith and Read 1997). During forest matu-

ration, progressive recruitment diversifies the community,

introducing species (referred to as late-stage fungi, Last et

al. 1987) that often fruit in old stands. Indeed, early-stage

fungi with ruderal strategies and late-stage fungi with com-

petitive or stress-selected strategies have been suggested as

functional groups (Deacon and Fleming 1992). Further evi-

dence that host demography may be involved in fungal com-

munity dynamics is provided by the descriptions of major

disturbance effects on diversity. Thus, shifts in the ECM

fungal community have been observed following clear-cuts

(Jones et al. 2003), volcanic eruptions (Nara et al. 2003), or

wild fires (Baar et al. 1999; Dahlberg et al. 2001). Canopy

gaps and subsequent regrowth create a diverse age structure

of trees, so it is likely that old-growth forests will contain

early and late succession ECM fungi. However, to our

knowledge, the local effects of natural canopy gaps on fruit-

ing patterns and species richness have not been explored.

We investigated the fruiting patterns of macromycetes at

local scale in an old-growth forest dominated by Quercus

ilex L. The study was initiated in 1999 within a 6400-m2

permanent plot located in the Fango forest (Corsica Island).

This work opens a new window on Mediterranean diversity,

in a climatic region usually described as a diversity hot-spot

for many other organisms (Cowling et al. 1996; Médail and

Verlaque 1997). In addition, our study plot provides a model

in which the role of natural forest dynamics can be tested

because plant composition and vegetation structure have

been described on a fine scale-grid in a previous study

(Panaïotis et al. 1997). Our major objectives were (i)todoc-

ument the ectomycorrhizal diversity in an unmanaged Medi-

terranean forest ecosystem and (ii) to correlate fruiting

patterns with host-tree distribution and other forest structural

features such as tree density and canopy gaps. We also tested

whether the ECM community could be divided in various

components, based on host specificity and life history traits

of species such as fruitbody size, temporal fruiting patterns,

fruitbody abundance, and spatial frequency. In addition to

the ectomycorrhizal fungal community, we examined the di-

versity of saprobic fungi, which are submitted to the same

abiotic conditions, but lack any direct connection to living

plants.

Materials and methods

Location of the study site and habitat characteristics

The study site is located within the Fango forest

(42°20′N; 8°49′E) at the northwestern edge of Corsica Is-

land (Fig. 1a). This French island extends between sea level

and a maximal altitude of 2710 m. The Fango valley has

been a Man and Biosphere (MAB) reserve since 1973, as it

contains one of the rare old-growth Q. ilex (holm oak) for-

ests of the Mediterranean basin (Quézel and Médail 2003).

The Fango forest covers a 4318-ha area on Hercynian gran-

ite with enclaves of volcanic rhyolites. It is dominated by

sclerophyllous evergreen Mediterranean species, among

which Q. ilex occurs in the mesomediterranean belt (i.e., be-

tween 200 and 900 m above sea level). Soils are alocrisols

(AFES 1995) with mull humus overlying (i) a thick organic

layer with a slightly acidic pH ranging from 5.7 to 6.4 and a

1712 Can. J. Bot. Vol. 82, 2004

C to N ratio ranging from 24 to 28 and (ii) an altered granite

horizon. The climate is subhumid with a mean annual rain-

fall of 750 mm and an average annual temperature of

14.6 °C at 192 m above sea level. Temperatures range from

3.5 °C (mean January minima) to 29.9 °C (mean July max-

ima).

Characteristics of the Q. ilex old growth stand

The study stand is located in a 15-ha Q. ilex old-growth

forest that consists of numerous large trees and a 7-m-high

dense chaparral. This dense understory layer consisted prin-

cipally of Phillyrea latifolia L., Erica arborea L., Arbutus

unedo L. (strawberry tree) shrubs and other rare species

scattered throughout such as Cistus salviifolius L. and Cistus

monspeliensis L. (Gamisans 1999; Panaïotis et al. 1997; data

not shown). Of the woody species, only Q. ilex,Arbutus

unedo, and the two species of Cistus have the ability to asso-

ciate with ECM fungi.

The forest shows numerous canopy gaps of about 100 m2

each (Panaïotis et al. 1995), which occur when old Q. ilex

stems (170 ± 46 years) break and fall down (Fig. 1b). In this

paper, natural gap is defined as the surface of the forest floor

directly located under the canopy opening, following the ter-

minology of Mc Carthy (2001).

Collecting macrofungi and plants and processing of

field data

The fruitbody survey was conducted within a 6400-m2

(160m×40m)permanent transect that extended from the

upper Q. ilex forest limit at 390 m down to 330 m (Fig. 1b).

This transect (established in 1994 by Panaïotis et al. (1997))

was divided into 64 contiguous (10m×10m)plots (la-

belled from A1 to D16, Fig. 1b), each divided into 100

(1m×1m)subplots.

Between 1999 and 2002, epigeous fruitbodies of ECM

and saprobic macrofungi fruiting on soil were searched

weekly over the whole fruiting season, i.e., from September

15th to March 15th. The position of each fruitbody was re-

corded and mapped with a 0.1-m resolution using the GIS

(Geographical Information System) software Arcview 3.2a

(ESRI Inc, Paris). Species forming microscopic, hypogeous,

or resupinate fruitbodies were not taken into account.

Fruitbodies of parasitic or saprobic species fruiting directly

on woody debris, fresh dungs, or living plants were also ig-

Richard et al. 1713

Fig. 1. (a) Location of the study site on Corsica Island (France). (b) Spatial distribution of the individuals of Quercus ilex and Arbutus

unedo in the study plot (6400 m2). 䊉,Q. ilex standing trees; 䊊,Arbutus unedo shrubs; lines, Q. ilex fallen trees; hatched zones, rocks

and screes. (c) Concentric zones (Z1–Z5) created using Arcinfo software to analyse the relationship between fruiting patterns and can-

opy gaps. The grid indicates limits of the 100 m2plots.

nored, as their repartition is obviously determined by the oc-

currence of such substrates.

Fungal collections were identified to species level in most

cases and occasionally to subspecies (= variety) level. Taxa

poorly monographed, at least for this Mediterranean region,

were ascribed to broad taxonomic units according to Bon

(1988) or, alternatively, to unidentified species that were de-

lineated for this study (Appendix A, Tables A1 and A2).

Over 3 years, we only removed a minimal number of fruit-

bodies for identification. Representative voucher collections

for all ECM taxa were deposited in the herbarium of the

Évolution et diversité biologique laboratory (Unité mixte de

recherche 5174, Université Toulouse III Paul Sabatier,

Toulouse, France), as were vouchers of 25 saprobic taxa for

which identification was doubtful in the field (noted by as-

terisks in Table A2). In addition, a picture database of

saprobic species was prepared and deposited in the same

place.

Vascular plants were first recorded in 1994 by Panaïotis

and Gamisans (unpublished data). They were surveyed again

in 2002, together with nonvascular plant species. Quercus

ilex and Arbutus unedo individuals were mapped at a 0.1-m

accuracy, but seedlings were not considered. Large rocks and

loose rock debris were also positioned (Fig. 1b).

All field data were stored using Arcview 3.2a software.

Assignment of ECM status and diversity analysis

Fungal taxa belonging to genera for which ectomy-

corrhizae have been undoubtedly reported in the literature

were ascribed to the ECM guild (Table A1). We included all

genera and species of uncertain ECM abilities (e.g.,

Clavulina,Clavulinopsis,Entoloma, and Hygrocybe)inthe

saprobic guild (Table A2). Species abundance was calculated

as the cumulative number of fruitbodies produced by a given

species in the period 1999–2002 (i.e., during three consecu-

tive fruiting seasons; absence or presence data were used

because counting fruitbodies (instead of biomass measure-

ments) does not affect fruiting patterns and diversity). Spe-

cies frequency corresponds to the percentage of spatial units

(plots or subplots, depending on the analysis) in which a

given species fruited at least once during the entire sampling

period (Bills et al. 1986; Villeneuve et al. 1989). Production

(also referred to as fruitbody abundance) was defined as the

total number of fruitbodies encountered over the period Sep-

tember 1999 – March 2002. Species diversity was estimated

using (i) richness, that is, the total number of taxa (S;Smand

Ssfor ECM and saprobic fungi, respectively), (ii) Simpson’s

diversity index (D), (iii) Shannon–Wiener information index

(′H), and (iv) Fisher’s parameter alpha (Fisher et al. 1943).

The distribution of relative species abundance was analysed

using a rank–abundance curve.

Statistical analyses of spatial patterns

To examine whether differences in community composi-

tion arise as a result of geographical distance, we performed

a spatial autocorrelation test at the (100 m2) plot level. We

examined the correlation between geographic distance and

community similarity among plot pairs. Distances among

plots were measured as the distance between plot centers.

The similarity matrix was constructed using either Jaccard

(J) or Sorensen ( ′S) coefficients that were calculated as fol-

lows:

[1] J=c/(a+b–c)

[2] ′S=2c/(a+b)

where ais the total number of species in one plot, bis the

total number of species found in the other plot, and cis the

number of species the two plots have in common. Correla-

tion between geographic distance and community similarity

was assessed using a Mantel test in the R Package for

Multivariate and Spatial Analysis version 4.0 (Casgrain and

Legendre 2001; www.fas.umontreal.ca/BIOL/Casgrain). The

Mantel’s standardized rvalues range from –1 (negative cor-

relation between similarity matrix and distance matrix) to +1

(positive correlation). The significance level was calculated

by randomization of one matrix and calculation of Mantel’s

rvalue (Mantel 1967) for each random pairing.

To determine fruitbody distribution patterns (i.e.,

aggregative, random, or regular), we used the Besag function

L(r) (Besag and Diggle 1977), calculated as follows:

[3] Lr Kr r() ()/=−π

where K(r) is the Ripley function (Ripley 1977) and ris the

distance. The function L(r) was calculated for distances that

range from 0.5 to 20 m (i.e., half of the transect width) and

correspond to aggregate from 1 to 1250 m2on average. Con-

fidence intervals at 99% for the null hypothesis of spatial

randomness were calculated using the Monte Carlo method

using the software ADE-4 (Thioulouse et al. 1997).

The analysis was performed for (i) the ECM community,

(ii) the saprobic community, and (iii) species with sample

sizes greater than 64 fruitbodies, that is, for those species

that could potentially be encountered in all 64 plots of the

transect. Calculations were not done for less abundant spe-

cies because of the inherent lack of statistical power.

Search for correlations between the fungal community

and structural characteristics of the forest

First, we examined at the plot level the relations between

either fungal productivity or species richness and three pa-

rameters of vegetation structure, that is, tree density, cover-

age, and number of vegetation layers. Fruitbody abundance

and species richness were calculated per plot for both ECM

and saprobic fungi and compiled over the entire transect. In-

dividuals of Q. ilex and Arbutus unedo were mapped and

counted in each 100 m2plot. We distinguished two vegeta-

tion entities: chaparral understory and Q. ilex cohort sensu

stricto. Each entity was further divided into vegetation layers

according to plant height. The chaparral was composed of

layer 1 (i.e., 1.5- to 5-m-high shrubs) and layer 2 (i.e., 5- to

10-m-high shrubs). In the Q. ilex cohort, we distinguished

trees ranging from 5 to 10 m in height (layer 1) from trees

higher than 10 m (layer 2). A coverage value was estimated

for each layer in each plot and expressed as the percentage

of the plot covered by this layer. Finally, the relation be-

tween the macrofungal community and forest structure was

quantified and tested statistically using the Pearson’s corre-

lation coefficient.

Second, we examined the relationships between either

fungal productivity or species richness and the presence of

1714 Can. J. Bot. Vol. 82, 2004

Q. ilex and Arbutus unedo (i.e., the two dominant ECM

hosts in our study site). Circular zones (1 or2minradius)

centred on Q. ilex and Arbutus unedo individuals were simu-

lated using Arcview software. Fruitbodies and species of

ECM and saprobic fungi were counted in each zone,

summed at the plot level, and expressed per square meter

per plot. Values were then compiled over the entire study

site. Finally, Mann–Whitney nonparametric tests were per-

formed for three distance classes (i.e., 0–1 m; 1–2 m, and 0–

2 m) to test for differences in macromycete fruiting patterns

between Q. ilex and Arbutus unedo.

Third, we examined qualitatively and quantitatively the ef-

fect of canopy gaps on the fruiting fungal community. Ten

canopy gaps were identified over the entire transect based on

the presence of fallen Q. ilex trunks above the ground. The

decaying trunk in each gap was mapped at a 0.1-m accuracy.

Five concentric zones (Z1–Z5; 1–5 m in radius) centred

on the tree trunk were simulated using Arcview software

(Fig. 1c). Fruitbody abundances and species richness were

calculated at the subplot level in each zone. A Kruskal–

Wallis nonparametric test was performed with five distance

classes, each corresponding to one of the five zones. This

statistical test was done for (i) the whole ECM fungal com-

munity, (ii) the whole saprobic community, (iii) abundant

species (ECM and saprobic) that produced more than 30

fruitbodies during the period 1999–2002, and (iv) saprobic

species known to have some wood-decaying abilities such as

Armillaria mellea (Vahl: Fr.) Kumm..

Classification of ECM taxa based on life history traits

We documented seven life history traits for each of the

128 well-identified ECM species or subspecies (Table A1).

Three traits were caused by temporal variation in fruitbody

production: (i) duration (days) of the fruiting period, (ii)pe

riodicity (number of days between the production peak and

the beginning of the fruiting season), and (iii) fluctuation

from one year to another. The last four traits included

(1) abundance (number of fruitbodies per 100 m2plot),

(2) spatial frequency (i.e., number of 100 m2plots in which

a species is present), (3) mean diameter of the fruitbody cap,

and (4) host range. Traits 3 and 4 were documented using

compiled litterature from Bon (1988) and P.-A. Moreau’s

unpublished data for rare species. A multivariate Hill and

Smith analysis (Hill and Smith 1976) was performed on

qualitative and quantitative traits using the software package

ADE-4 (Thioulouse et al. 1997). The Ward’s linkage method

based on Euclidian distances (Ward 1963) was then used to

define groups of taxa. Their significance was assessed using

a random permutation test with 1000 permutations (Manly

1997). Finally, using Mann–Whitney nonparametric tests we

investigated the relationship among the life history traits that

we identified as being discriminating using results from the

multivariate analysis.

Results

Diversity of the macrofungal communities

As a result of 78 weekly surveys over three consecutive

fruiting seasons (in the period 1999–2002), we recorded

5382 fruitbodies of macromycetes. A total of 234 species

were found, including 166 ECM taxa (Sm) and 68 taxa (Ss)

of saprobic fungi (Table 1). In contrast, only 19 angiosperm

species were collected out of 36 Embryophytes (data not

shown). In addition to this low plant diversity, the vegetation

survey showed that the ECM plants at our site consisted only

of Q. ilex and Arbutus unedo, with the exception of two

small individuals of Cistus.

The ECM fungal community was represented by 25 gen-

era of Basidiomycetes (22) and Ascomycetes (3), out of

3659 collected fruitbodies (Fig. 2a). Members in the family

Russulaceae accounted for 40.5% of the fruitbodies and

31.9% of the taxonomic diversity. At the genus level, three

main patterns were observed: (i) species-rich and highly pro-

ductive genera, e.g., Russula and Amanita,(ii) species-rich

but unproductive genera, e.g., Tricholoma and Cortinarius

(23.5% of Smand 6.7% of the ECM fruitbody production);

and (iii) species-poor but highly productive genera, e.g.,

Lactarius and Hydnum (Fig. 2a). The ECM community was

also strongly dominated by rare species with more than

60.8% of the taxa (101 out of 166) producing less than 1

fruitbody per 1000 m2on average (Fig. 3, Table A1). At the

other extreme of the rank–abundance curve, the three domi-

nant species, Laccaria laccata (Scop.: Fr.) Berk. & Broome,

Lactarius chrysorrheus Fr, and Inocybe tigrina R. Heim,

produced 32.9% of all fruitbodies (Fig. 3, Table A1).

The saprobic taxa were distributed over 31 genera of

Basidiomycetes (28) and Ascomycetes (3), out of 1723 col-

lected fruitbodies (Fig. 2b). The community was mainly

composed of Marasmiaceae (44.7% of the total fruitbody

number). The two most abundant species were Mycena

vitilis (Fr.) Quél. and Lycoperdon perlatum Pers.: Pers.,

which together produced 30% of saprobic species

fruitbodies. As for ECM fungi, rare taxa also dominated this

community since 58.8% of taxa (40 out of 68) produced less

than 1 fruitbody per 1000 m2on average (Fig. 3, Table A2).

The Shannon–Wiener information index (H′), the

Simpson’s diversity index (D), and the Fisher’s alpha value

were high (relative to other diversity studies) using either

Richard et al. 1715

Diversity index ECM community Saprobic community

Shannon–Wiener information index ( ′

Using abundance 5.48 4.33

Using frequency 5.98 4.98

Simpson’s diversity index (D)

Using abundance 0.953 0.918

Using frequency 0.975 0.948

Fisher’s alpha value (α) 36.51 14.31

Richness (S) 166 68

Table 1. Estimators of diversity for the ECM and saprobic fungi.

fruitbody abundance or frequency (Table 1). They were also

greater for ECM than for saprobic fungi.

Spatial patterns of species and fruitbodies

The Mantel’s test showed a low rvalue and a significant

autocorrelation at the plot level for ECM fungi, using

Jaccard’s and Sorensen’s coefficients (r= 0.079, p= 0.030

and r= 0.076, p= 0.047, respectively). These results suggest

that community similarity remains nearly constant as the

interplot distance increases. No spatial autocorrelation was

found for saprobic fungi (r= –0.011, p= 0.380 and r=

−0019. ,p= 0.310 using Jaccard’s and Sorensen’s coeffi-

cients, respectively).

Fruitbodies were unevenly distributed on the study site.

We found a few highly productive plots and numerous un-

productive plots. For example, clustered fruiting was ob-

served in two groups of adjacent plots (i.e., A3 + B2 + B3

and A7 + B7, Fig. 1b) that harboured 18% of the ECM

fruitbody production (i.e., twice the expected number of

fruitbodies in an assumed uniform area of the same size).

These plots were also species-rich since they included 66%

of the taxa (i.e., 2.4 times the expected number). In contrast,

four groups of plots (i.e., C1 + D1 + D2 + D3 + D4, C7 +

D7, A8 + B8 + D8, and D11, Fig. 1b) covering 1100 m2

(17.2% of the transect) exhibited only 5.1% of the total

ECM fruitbody production (three times less than the mean

production). Such unproductive plots were generally

species-poor and typically contained 6–8 ECM taxa (data

not shown); they were mainly observed in zones with screes

at ground level. For saprobic fungi, three plots (i.e., C4, B7,

and A12, Fig. 1b) covering 4.7% of the total area exhibited

28.8% of the fruitbody production while 32 others (50% of

the total area) accounted for only 10.8% of the production.

Rare species were evenly encountered in the 6400 m2

Q. ilex stand. They were scattered in 87.5% (56 out of 64)

and 62.5% (40 out of 64) of the 100 m2plots for ECM and

saprobic species, respectively (data not shown). In contrast,

abundant species (i.e., those that produced more than 64

fruitbodies over 3 years) were not regularly distributed. Of

the 12 abundant ECM species, nine occurred in half of the

100 m2plots or more, but none fruited in all plots (Table A1,

column 4). As revealed by the Besag function, the distribu-

tion of fruitbodies of these abundant species was mainly

aggregative with the exception of two ECM taxa, Russula

delica and Russula olivacea, and two saprobic fungi,

Mycena vitilis and Mycena alcalina (Table 2). The ECM ag-

gregates ranged from 20 m2for Laccaria laccata to 200 m2

for Russula fragilis. At the community level, aggregates

were also found consistently each year, except year 2 for

saprobic fungi (Table 2). ECM aggregates covered a 12- to

50-m2area in average.

Production and species richness in relation to

vegetation coverage, number of layers, and ECM host

density

Negative correlations were found between the fungal com-

munity and the number of vegetation layers (with the excep-

tion of saprobic production) (Table 3). These results suggest

that the fruiting fungal community may respond to canopy

closure. In addition, ECM richness was negatively correlated

to Q. ilex density but positively correlated to Arbutus unedo

density (Table 3), suggesting a potential (but distinct) role of

these two ECM hosts on fungal diversity. A similar pattern

was also observed for the saprobic fungi as richness was

negatively correlated to Q. ilex density and coverage (Ta-

ble 3). A negative correlation was also found between pro-

duction of saprobic fungi and total vegetation coverage.

Production and species richness in relation to the

presence of Q. ilex and Arbutus unedo

No significant difference was observed in ECM species

richness and fruitbody abundance between Arbutus unedo

1716 Can. J. Bot. Vol. 82, 2004

Fig. 2. Relative abundance of fruitbodies (filled bars) and rela-

tive species diversity (open bars) of the most abundant genera

for ECM fungi (a) and saprobic fungi (b).

Fig. 3. Dominance–diversity curves for ECM (䉱) and saprobic

(ⵧ) taxa of macrofungi. Most frequent species to least frequent

ones are ordered from left to right.

and Q. ilex (Table 4). However, some ECM species fruited

preferentially (significant differences by Mann–Whitney

nonparametric tests) either near Q. ilex (e.g., Russula

decipiens,Russula acrifolia, and Sarcodon cyrneus), or near

Arbutus unedo (e.g., Inocybe geophylla var. lilacina,

Leccinum corsicum,Tricholoma ustale, and representatives

of the genus Inocybe) (data not shown). In addition, taxa

such as Leccinum corsicum or Inocybe cervicolor were al-

ways absent from Q. ilex stem vicinity.

In contrast to ECM fungi, a clear distinction was observed

for saprobic fungi between these two woody plants.

Fruitbody abundance was higher near Arbutus unedo shrubs

(within the first 2 m) than near Q. ilex trees (Table 4). Spe-

cies were also more numerous (Table 4). Many of them such

as Calocybe carnea,Clavulina cinerea, and Clitocybe

odorata tended to fruit more frequently close to Arbutus

unedo than in other areas (data not shown).

Production and species diversity in relation to the

presence of canopy gaps

Fruiting patterns were analysed at various distances

(zones Z1–Z5, Fig. 1c) from fallen Q. ilex trunks. In Z1,

ECM fruitbody production and species richness were respec-

tively 27% and 31% higher than outside gaps (Fig. 4a). Spe-

cies richness and production were also significantly

enhanced for saprobic fungi in Z1, i.e., within the first meter

around the dead trunk (Fig. 4b). In the four following meters

(from Z2 to Z5), values decreased drastically for both eco-

logical groups, suggesting a strong and localized effect in

the nearby proximity of the decaying trunk. Thus, the ECM

community was 42% less productive in Z4 and 40% less di-

verse in Z5 than it was outside gap (at >5 m away; Fig. 4a).

At the species level, differences in fruitbody production

were also detected between gap and non-gap areas (Table 5).

Some ECM species were highly productive in gaps, others

were absent or rarely encountered. For example, Russula

vesca,Russula persicina var. rubrata, and Lactarius

rubrocinctus produced two to nine times more fruitbodies in

Z1 than in the non-gap area. In contrast, another set of abun-

dant taxa (e.g., Laccaria laccata,Russula chloroides, and

Inocybe geophylla) and species of the genus Hebeloma were

absent or weakly represented in gaps. A similar pattern was

observed for saprobic species. Production of well-known

wood-decaying species such as Armillaria mellea,Leuco-

paxillus gentianeus, and Leucopaxillus tricolor was particu-

Richard et al. 1717

Aggregative spatial pattern

Random spatial

pattern

Whole community

Ectomycorrhizal Season 1 (12), 2 (27), 3 (50 and 710), and all three seasons (50)

Saprobic Season 1 (50), 3 (50 and 615), and all three seasons (no clear scale) Season 2

Abundant taxa

Ectomycorrhizal Laccaria laccata (20), Lactarius chrysorrheus (50), Inocybe tigrina (20), Hydnum repandum

(150), Russula acrifolia (30), Hygrophorus russula (50), Russula risigallina (30), Russula

fragilis (200), Tricholoma saponaceum (80), Cortinarius elatior (50)

Russula olivacea,

Russula delica

Saprobic Lycoperdon perlatum (250), Entoloma nidorosum (400), Clavulina cinerea (590), Collybia

butyracea (200), Clitocybe gibba (no clear scale), Collybia dryophila (10), Mycena pura

(250)

Mycena vitilis,

Mycena alcalina

Note: The size (in square meters) of aggregates is given in parentheses.

Table 2. Spatial patterns of fruitbodies using the Besag function (Besag and Diggle 1977).

Fig. 4. Fruitbody production (top bars) and species diversity

(bottom bars) of macrofungi in canopy gaps versus closed can-

opy forest (>5 m from fallen trunk) at the subplot level.

(a) ECM species. (b) Saprobic species. Means followed by the

same letter do not differ significantly at the 5% level. Error bars

are standard deviations.

larly important in gaps (Table 5). For example, 25 out of 26

Armillaria mellea fruitbodies were collected in Z1, and 41

out of 42 Leucopaxillus gentianeus fruitbodies occurred in

Z1, Z2, and Z3 (data not shown). Conversely, despite its

abundance (Table A2), the litter-decaying species Collybia

butyracea was rarely found in gaps, and never in the first

three zones (Table 5).

Grouping of ECM fungi based on host specificity and

life-history traits

The ECM community included species reported to associ-

ate with (i) a large range of angiosperms (e.g., Scleroderma

verrucosum and Russula amoenicolor), (ii) trees in the fam-

ily Fagaceae (e.g., Russula grisea), and (iii) various species

of oaks (e.g., Lactarius chrysorrheus and Hygrophorus

russula) (Table A1). It also included broad host range spe-

cies (e.g., Laccaria laccata and Amanita phalloides)re

ported to associate with both Angiosperms and

Gymnosperms. Furthermore, the community contained a

small group of species restricted to Mediterranean Q. ilex

forests (11 taxa, e.g., Russula laricino-affinis and Sarcodon

cyrneus), chaparrals (5 taxa, e.g., Russula cistoadelpha and

Amanita gioisa), or other thermophilic angiosperm forests

(17 taxa, e.g., Boletus aereus and Cortinarius caligatus).

This entire group of habitat-restricted species encompassed

25.9% of the species richness and 9.9% of the fruitbody pro-

duction.

Out of 128 identified taxa, 3 (i.e., Laccaria laccata,

Inocybe tigrina, and Lactarius chrysorrheus) were clearly

separated from the others based on a Hill and Smith

multivariate analysis (data not shown). These three species

(group A, Table A1) were very abundant in the stand. We

excluded them from a second multivariate analysis aiming to

separate the remnant species. In this new analysis, the first

two factorial axes accounted for 22.2% (F1) and 10.8% (F2)

of the total variability, respectively. Two groups of traits ap-

peared to be opposed along the F1 axis: high abundance,

high frequency, regular fruitbody production from one year

to another, and broad host range (positive scores) versus low

abundance, low frequency, high fluctuation in fruitbody pro-

duction from one year to another, and narrow host range

(negative scores) (Figs. 5aand 5b). The F2 axis further sepa-

rated thermophilic and chaparral linked species from species

with other ecological specificity. The other axes did not add

further information (Eigenvalues < 9%, data not shown).

Among the 125 ECM taxa, five new groups (B to F) were

distinguished using the Ward’s method (Fig. 5c, Table A1).

Random permutation tests performed on these groups

showed that within-group inertia was significantly higher

than between-group inertia (p< 0.001). Groups B, C, and D

(92 taxa) were dominated by species that (i) were mainly

broad host range, (ii) fruited regularly (i.e., observed in ei-

ther 2 or 3 years), and (iii) produced a large number of

fruitbodies. In contrast, groups E (mainly chaparral re-

stricted) and F (mainly Q. ilex linked or thermophilic) en-

compassed 33 species mainly linked to Mediterranean hosts

and producing episodically a few fruitbodies (Table A1).

Mann–Whitney nonparametric tests (including all groups,

from A to F) confirmed that thermophilic species as well as

those frequently encountered in Q. ilex forest or in chapar-

rals (Table A1) had distinctive characteristics. They pro-

1718 Can. J. Bot. Vol. 82, 2004

ECM fungi Saprobic fungi All macrofungi

Production Richness Production Richness Production Richness

Q. ilex density* –0.134 (0.290) –0.298‡(0.017) 0.202 (0.109) –0.340‡(0.006) –0.215 (0.088) –0.350‡(0.005)

Arbutus unedo density* 0.139 (0.274) 0.244‡(0.050) –0.122 (0.336) 0.075 (0.554) 0.004 (0.978) 0.213 (0.091)

Q. ilex coverage –0.153 (0.226) –0.234 (0.063) –0.182 (0.151) –0.342‡(0.006) –0.213 (0.091) –0.302‡(0.015)

Chaparral coverage 0.049 (0.698) 0.216 (0.087) –0.078 (0.543) 0.233 (0.064) –0.021 (0.868) 0.249‡(0.047)

(Q. ilex + chaparral) coverage –0.083 (0.515) 0.033 (0.796) –0.252‡(0.044) –0.045 (0.724) –0.217 (0.085) 0.009 (0.946)

Total number of layers†–0.254‡(0.043) –0.298‡(0.017) –0.237 (0.059) –0.453§(0.000) –0.311‡(0.012) –0.298§(0.001)

Note: Quantification of the relationships was performed using Pearson’s correlations coefficients at the plot level. Probability (p) values are given in parentheses.

*Density was defined as the number of Q. ilex or Arbutus unedo individuals per 100 m2plot.

†For each vegetation entity, we only considered layers that represented at least 10% coverage.

‡Significant correlation at p< 0.05.

§Significant correlation at p< 0.001.

Table 3. Relationships between the fruiting fungal community and forest structure including Quercus ilex and Arbutus unedo density and coverage, and number of layers.

Richard et al. 1719

duced significantly fewer fruitbodies (p= 0.011), less

frequently (p= 0.007), and less regularly (p= 0.003) than

species that are not restricted to a Mediterranean habitat (Ta-

ble A1). They also had a shorter fruiting period (p= 0.014).

No additional difference was found between these two

groups of species in fruitbody size or in other phenological

characteristics.

Discussion

ECM richness, fruitbody productivity, and species

composition

Only a few studies have dealt with fungal diversity associ-

ated with Q. ilex (Signorello 1996; Laganà et al. 1999) but

none of them in old-growth forests. Our results show that es-

timates of taxonomic diversity in Q. ilex old-growth forests

are likely to be tremendously high, especially when consid-

ering the large number of rare species found at our site

(61.4% of the ECM community), the limited size of our

sampling site (6400 m2), and the limited duration of our sur-

vey (3 consecutive fruiting seasons). Thus, species richness

was high (S= 166; Table 1) with respect to the low number

of ectomycorrhizal hosts and the low productivity of the

fungal community during the survey period. However, this

value is consistent with others reported from old-growth co-

nifer forests. For example, O’Dell et al. (1999) and Smith et

al. (2002) found 150 and 133 ECM species in Douglas-fir

(Pseudotsuga menziesii Mirb.) and Western hemlock (Tsuga

heterophylla (Raf.) Sarg.) stands of similar sizes, respec-

tively. Furthermore, in a comparison of Norway spruce

(Picea abies L.) forests of different ages, the highest ECM

fungal diversity was observed in the oldest, 200-year-old

stand (Peter et al. 2001).

In our stand, two ECM hosts (i.e., Q. ilex and Arbutus

unedo) shared 166 fungal symbionts, without taking into ac-

count either hypogeous or resupinate taxa, such as

0–1 m from tree 0–2 m from tree 1–2 m from tree

Q. ilex A. unedo Q. ilex A. unedo Q. ilex A. unedo

ECM fungi

Production 0.379±0.064a 0.477±0.071a 0.396±0.049a 0.427±0.039a 0.415±0.060a 0.409±0.037a

Richness 0.242±0.039a 0.248±0.027a 0.218±0.022a 0.210±0.015a 0.236±0.027a 0.224±0.018a

Saprobic fungi

Production 0.117±0.063a 0.187±0.088a 0.069±0.021a 0.202±0.045b 0.045±0.010a 0.210±0.051b

Richness 0.047±0.011a 0.071±0.018a 0.037±0.006a 0.059±0.008b 0.032±0.007a 0.059±0.007b

Note: Numbers are means per square meter at plot level ± SD (n= 64). For each of three distance classes, values followed by different letters indicate

a significant difference based on Mann–Whitney nonparametric tests (p< 0.05).

Table 4. Production and species richness at various distances from individuals of Quercus ilex (n= 108) and Arbutus unedo (n= 227).

Distance from fallen trunk in canopy gaps

Species Z1 (0–1 m) Z2 (1–2 m) Z3 (2–3 m) Z4 (3–4 m) Z5 (4–5 m)

Closed canopy

forest (>5 m)

ECM species

Inocybe tigrina 10.09 12.9 6.97 2.45 7.26 4.84

Russula vesca 2.29* 0.72 0 0 0.24 0.23

Russula persicina var. rubrata 1.38 1.79* 0.61 0 0.97 0.46

Russula globispora 1.83 0.71 0.61 0.54 0.24 0.71

Russula fragilis 2.29 1.43 0.61 0.27* 0.48 1.38

Cortinarius elatior 1.83 0.36 0 1.09 0.97 1.11

Lactarius rubrocinctus 1.83 1.08 0.91 0.27 0 0.61

Laccaria laccata 0.46 1.43 2.12 1.36 0.97 7.58

Russula chloroides 0 0 1.21 0.27 0 1.23

Inocybe geophylla 0 0.36 0.91 0.27 0.48 1.06

Whole genus Hebeloma 0.46 0 0 0 0 0.63

Saprobic species

Armillaria mellea 11.47* 0 0 0 0 0.02

Leucopaxillus gentianeus 1.38 0.36 10.3* 0 0 0.02

Leucopaxillus tricolor 1.83* 0 0 0 0 0.02

Collybia butyracea 0 0 0 0.82 0 1.86

Note: Five distance classes corresponding to the five concentric zones in Fig. 1cwere identified in canopy gaps. They range from 0 to

5 m from the fallen trunk in each gap.

*Significantly different according to Kruskal–Wallis nonparametric tests at p< 0.05.

Table 5. Mean numbers of fruitbodies per 100 m2in gaps (n= 10) and the closed canopy forest for the most abundant

ECM and saprobic species.

Thelephoraceae and Sebacinaceae. The ECM fungal com-

munity was 4.6 times more diverse than the whole plant

community, including nonvascular plants (data not shown).

Most reported ratios from other forests are lower. They gen-

erally range from 0.2 to 1.8 ECM species per vascular plant

species (e.g., in deciduous forests dominated by alder (Brun-

ner et al. 1992), oak (Schmit et al. 1999), or birch

(Villeneuve et al. 1989)). However, Dahlberg et al. (1997)

found a ratio of 3.7 in an old-growth Norway spruce forest.

Various factors could account for such a high ratio in our

stand: (i) historical legacy of ECM fungal diversity and

(ii) extreme abiotic factors or strong environmental fluctua-

tions leading to species-rich assemblages (Whittaker 1972;

Huisman and Weissing 1999). For example, in a study of

1720 Can. J. Bot. Vol. 82, 2004

Fig. 5. Hill and Smith (1976) multivariate analysis. (a) Correlation circle for quantitative variables. (b) Position of modalities for quali-

tative variables. (c) Distribution of species groups.

hypogeous and epigeous Gasteromycetes from different sites

in the Sonora region, a surprisingly high diversity was found

in the tropical thorn forest, an extremely xeric habitat, 88.6%

of which was covered by a single plant species (Esqueda-

Valle et al. 2000). Additional factors could also drive such

a diversity at local scale by limiting competitive exclusion

among fungal species; they include unrecognized limiting fac-

tors or environmental heterogeneities (Tilman 1992) and

antagonistic interactions (Czárán et al. 2002). Later on, we

discuss environmental heterogeneities in fruiting patterns

linked to forest structure and composition.

Another noteworthy feature was the low fruitbody produc-

tivity of the ECM community (a feature also shared by

saprobic species) (Fig. 3). Most ECM species were repre-

sented by less than 1 fruitbody per 1000 m2(Table A1). This

low productivity together with the high species diversity lead

to particularly high values of diversity indexes (Table 1).

Similarly, Bills et al. (1986) and Villeneuve et al. (1989) ob-

tained high Shannon’s entropy values (ranging from 4.1 to

5.0) for the ECM fruiting community from old hardwood

stands. Peter et al. (2001) also reported a high Simpson’s in-

dex value (0.95) in a 200-year-old Norway spruce forest.

Finally, Smith et al. (2002) demonstrated that old-growth

Douglas-fir stands were as diverse as young forest with

closed canopy, but six times less productive.

The ECM community in our Q. ilex stand was very simi-

lar to that described by Skirgiello (1998) in another Euro-

pean old-growth hardwood forest in Bia»owieóa (Poland).

Striking similarities included (i) the high number of species

of Russula and Cortinarius,(ii) the low productivity of most

Cortinarius species, and (iii) the high abundance of

Laccaria laccata. In our stand, species of Russula and

Cortinarius accounted for 50% of the taxonomic diversity

(Fig. 2a). Species of Cortinarius represented only 6.7% of

the total ECM fruitbody production (Fig. 2a) and Laccaria

laccata was one of three most abundant producers (Table

A1). Similar trends in species composition have been ob-

served in conifer forests. For example, the genera Russula,

Cortinarius, and Inocybe accounted for 49% of fruitbody

collections and about 52% of taxonomic diversity in an old-

growth stand of Douglas-fir (Smith et al. 2002). However,

our results differ from other studies undertaken in coniferous

dominated old-growth stands, in which species of

Cortinarius were the most abundant producers, but neither

species of Russula nor other taxa were abundant (Dahlberg

et al. 1997; Peter et al. 2001). In our forest, the genus

Russula alone represented more than 25% of the number of

fruitbodies and species diversity. In summary, the genera

Russula and Cortinarius often dominate most fruiting ECM

communities. This result may be explained because

fruitbodies of these fungi are usually easy to detect in the

field and because these two genera also encompass a very

large array of species with diverse ecological requirements.

The predominance of Russula in a deciduous sclerophyllous

forest is interesting although this result requires additional

investigations to confirm that Russula spp. also dominate the

nonfruiting ECM community in the northern hemisphere.

Several life-history traits allowed us to separate the ECM

community into six distinct components that a posteriori ap-

peared to differ mainly by fruitbody abundance and host

specificity (Table A1). Two striking results emerged from

this classification: (i) in this Mediterranean forest, the ECM

diversity essentially encompassed broad host range species

with temperate affinities and (ii) a weak Mediterranean com-

ponent was composed of species that fruited less regularly

and less abundantly than temperate taxa. Nearly 90% of

fruitbodies were produced by temperate fungal species. A

similar trend was already reported by Norstedt et al. (2001)

in a polypore survey in endemic Corsican pine forests. More

than 56% of species were also encountered in northern Eu-

rope. The abundance of species with large geographical dis-

tribution may be explained by long-distance fungal dispersal

and (or) historical legacies. Alternatively, taxonomic and

methodological limitations may bias our view of the fungal

diversity and thus the ratio between Mediterranean and tem-

perate fungal taxa. Much of the fungal diversity remains to

be described (Hawksworth 2001) and many cosmopolitan

taxa may in fact cover several cryptic species, including

some restricted to the Mediterranean region. Finally, Medi-

terranean species may have an erratic fruiting pattern that

does not reflect their true diversity and abundance in soil

and on tree roots. Thus, as mentioned earlier, below ground

communities (i.e., those observed from mycorrhizae and my-

celia in the soil) would have to be explored to obtain a com-

prehensive overview of the ECM community.

ECM versus saprobic fungi

Overall, the ECM fungi appear 2.4 times more diverse and

2.1 times more productive than saprobes. This lack of diver-

sity and productivity in decomposers species is surprising

since accumulation of favourable substrates is likely to occur

in an old-growth forest (Ohlson et al. 1997). This result may

be explained by (i) our limited sampling of the decomposer

community that excludes wood-decaying fungi, (ii) differen-

tial responses to dry environmental conditions between ECM

fungi and saprobes, and (or) (iii) the relatively low C/N ratio

of the organic layer (ranging from 24 to 28; data not shown)

that would favour ECM fungi but not decomposers.

Spatial distribution of fruitbodies and species

At both the community and species level, we observed

aggregative distributions of ECM fruitbodies on areas of var-

ious sizes (Table 2). This has been described in a few studies

addressing the spatial distribution of epigeous and

hypogeous ECM species (Fogel 1976; Yamada and Katsuya

2001). For example, Yamada and Katsuya (2001) reported

an uneven distribution of fruitbodies for the most abundant

epigeous ECM species in Pinus densifolia reforested stands.

The presence of fruitbody aggregates among ECM fungi

may be indicative of a higher local activity of mycelia and

mycorrhizae in relation to soil heterogeneity and (or) host

root distribution. As a comparison, the saprobic community

did not show scales of fruitbody aggregation as clear as

those observed for ECM fungi (Table 2), suggesting differ-

ences in fruiting determinism between these two ecological

fungal groups and (or) differences in the spatial distribution

of mycelia.

A significant spatial autocorrelation was found among

ECM taxa using Mantel’s test, but it was lower than values

obtained in other studies of fruiting ECM communities

Richard et al. 1721

(Schmit et al. 1999; Peter et al. 2001). Thus, community

composition did not drift away with distance. This can be

explained by the large number of rare species with a high

spatial equitability (Crawley 1997). Overall, our results en-

tail methodological considerations with regard to the best

sampling strategy as follows: (i) new species appear slowly

as the sampled area increases, (ii) spreading out sampling

plots is unlikely to increase the capturing of rare ECM spe-

cies, and (iii) sampling few adjacent and productive plots

may be as effective as sampling discontinuous heteroge-

neous areas to characterize most of the fruiting ECM com-

munity. These recommendations are applicable to a lesser

extent to saprobic fungi as they did not show any significant

spatial autocorrelation and species appeared somewhat less

clustered than ECM fungi on average (Table 2).

Fungal diversity and productivity in relation to forest

structure and composition

Diversity and productivity of the ECM fruiting commu-

nity decreased significantly as the number of forest layers

increased (Table 3). This response is similar to the well-

known canopy closure effect on fruiting (Jansen and de Nie

1988; Vogt et al. 1992). In addition, species diversity was in-

fluenced by tree and shrub density (Table 3). Species diver-

sity decreased as the number of Q. ilex trees increased and

was positively correlated to the density of Arbutus unedo

shrubs. In addition, some species in the ECM community

were preferentially encountered near Q. ilex or near Arbutus

unedo (data not shown). We concluded that ECM diversity

and productivity are pro parte shaped by the ECM host and

forest structure, as already shown by Nantel and Neumann

(1992) and Såstad (1995) using fruitbodies, and by

Kernaghan et al. (2003) on ectomycorrhizae. No further cor-

relation was found between ECM fruiting patterns and other

vegetation descriptors (Tables 3 and 4).

As observed for ECM species, fruiting patterns of

saprobic species were sensitive to changes in vegetation

structure and composition (Tables 3 and 4). They were

highly diverse and abundant under chaparral canopy and de-

creased drastically as number and coverage of Q. ilex in-

creased, suggesting a possible response of these fungi to

litter quality.

Role of natural canopy gaps on fungal diversity and

productivity

Our results show a strong and localized effect of gaps on

fruiting (Fig. 4; Table 5) for both symbiotic and saprobic

guilds. ECM fruiting was higher quantitatively (+27%) and

qualitatively (+31%) near the decaying tree trunks (Fig. 4a).

A similar pattern was detected for saprobic fungi with wood-

decay abilities (Table 5, Fig. 4b). We also found that some

ECM species (e.g., the early-stage Laccaria laccata) were

poor producers in gaps, while six species belonging to late-

stage ECM genera (Russula,Lactarius, and Cortinarius)

were consistently very abundant within the first meter

around the fallen Q. ilex trunk (Table 5). These later species

might have the ability to acquire carbon from large woody

debris in addition to photosynthates from their living hosts.

Interestingly, abundant litter decomposers such as Collybia

butyracea were always found outside the gaps in contrast to

several saprobic species with wood decaying abilities (Ta-

ble 5). Alternative explanations to the observed diversity

patterns in gaps include shifts in abiotic conditions. Thus,

light level on the forest floor is known to increase with can-

opy opening, including in small gaps (Chazdon and Fetcher

1984; Canham et al. 1990). This, in turn, affects soil temper-

ature, soil nutrients, and soil moisture (Mehus 1986; Mc

Carthy 2001). These changes in abiotic conditions have the

potential to drive the physiological and metabolic activities

of the fungal symbionts and, as a result, the diversity and

productivity of the entire fruiting community. Regardless of

the mechanisms involved, our results highlight the role of

canopy gaps as favourable habitats for ECM fungi in an old-

growth sclerophyllous forest. Moreover, these results con-

firm the primary importance of large woody debris for the

conservation and management of the fungal diversity in

Q. ilex forests, as shown in other ecosystems (Harvey et al.

1978; Goodman and Trofymov 1998; Tedersoo et al. 2003).

Conclusions and perspectives

Our sampling protocol was aimed at producing a fair im-

age of a local macromycete fruiting community in terms of

taxonomic composition. Results suggest that Mediterranean

old-growth forests are of great interest for conservation of

the species diversity of ECM fungi, as shown by the unusual

high richness and diversity of the fruiting community in our

6400-m2study site. The paradox of a few tree species asso-

ciated with a large array of ECM fungal species, already de-

scribed in temperate ecosystems (Malloch et al. 1980),

applies to Mediterranean ecosystems, but we also found evi-

dence of a partial determinism of forest structure and host

dynamics in building various niches, mainly through gaps

created by fallen trees. Future work will include a descrip-

tion of below ground communities to obtain a more compre-

hensive view of the diversity. Comparisons with other sites

should also be attempted on the basis of parameters such as

Fisher’s alpha index or the species accumulation curve. Fur-

ther investigations are required to understand mechanisms

affecting the diversity and fruiting of fungal species. An-

other intriguing question is the ecological role of this ECM

diversity and its possible feedback on host tree dynamics.

Acknowledgements

This study is part of F. Richard’s Ph.D thesis dealing with

the diversity and role of ECM fungi in Q. ilex forests in rela-

tion to plant succession. We thank the Société Mycologique

d’Ajaccio and especially A. Tristani for help with the pre-

liminary identifications of fungal collections, the Parc Natu-

rel Régional de Corse (P.N.R.C), the Office National des

Forêts in Corte, and P. Pozzo di Borgo, M. Figarella, J.

Alessandri, and P. Lepaulmier for providing facilities and

encouragement. We are also grateful to C. Panaïotis who

helped us to initiate this Q. ilex thesis project, to G. Eys-

sartier for identification of numerous Cortinarius species,

to L. Riche for data entry and GIS analyses, to J. Chave for

critical advice on statistical analyses of community diversity

and spatial patterns, and to A. Lecerf for many helpful sug-

gestions with the ADE-4 software. A. Ramelot, J. Jacquot,

and H. Gryta helped us to collect fungi. Funding was

provided by the Office National des Forêts (O.N.F.), the

1722 Can. J. Bot. Vol. 82, 2004

Ministère de l’Ecologie (D.I.R.EN. Corse), and the Col-

lectivité Territoriale de Corse (C.T.C.) to F. Richard. We fi-

nally thank two anonymous reviewers and S. Berch for their

constructive comments that greatly improved this manu-

script.

References

AFES. 1995. Référentiel pédologique. Collections Techniques et

pratiques, INRA Editions, Paris.

Baar, J., Horton, T.R., Kretzer, A.M., and Bruns, T.D. 1999.

Mycorrhizal colonization of Pinus muricata from resistant

propagules after a stand-replacing wildfire. New Phytol. 143:

409–418.

Berglund, H., and Jonsson, B.G. 2001. Predictability of plant and

fungal species richness of old-growth boreal forest islands. J.

Veg. Sci. 12: 857–866.

Besag, J.E., and Diggle, P.J. 1977. Simple Monte Carlo tests for

spatial patterns. Appl. Stat. 26: 327–333.

Bills, G.F., Holtzman, G.I., and Miller, O.K., Jr. 1986. Comparison

of ectomycorrhizal-basidiomycete communities in red spruce

versus northern hardwood forests of West Virginia. Can. J. Bot.

64: 760–768.

Bon, M. 1988. Champignons de France et d’Europe occidentale.

Arthaud, Paris.

Brunner, I., Brunner, F., and Laursen, G.A. 1992. Characterization

and comparison of macrofungal communities in an Alnus

tenuifolia and an Alnus crispa forest in Alaska. Can. J. Bot. 70:

1247–1258.

Canham, C.D., Denslow, J.S., Platt, W.J., Runkle, J.R., Spies, T.A.,

and White, P.S. 1990. Light regimes beneath closed forest cano-

pies and tree-fall gaps in temperate and tropical forests. Can. J.

For. Res. 20: 620–631.

Casgrain, P., and Legendre, P. 2001. The R package for

multivariate and spatial analysis. Version 4.0 d5. User’s manual.

Département des sciences biologiques, Université de Montréal,

Montréal.

Chazdon, R.L., and Fetcher, N. 1984. Photosynthetic light environ-

ments in a lowland tropical rainforest in Costa Rica. J. Ecol. 72:

553–564.

Connell, J.H. 1978. Diversity in tropical rain forests and coral

reefs. Science (Washington, DC), 199: 1302–1310.

Cowling, R.M., Rundel, P.W., Lamont, B.B., Arroyo, M.K., and

Arianoutsou, M. 1996. Plant diversity in mediterranean-climate

regions. Tree, 11: 362–366.

Crawley, M.J. 1997. Plant ecology. 2nd ed. Blackwell Science Inc.,

Malden, Massachusetts.

Czárán, T.L., Hoekstra, R.F., and Pagie, L. 2002. Chemical warfare

between microbes promotes biodiversity. Proc. Natl. Acad. Sci.

U.S.A. 99: 786–790.

Dahlberg, A., Jonsson, L., and Nylund, J.E. 1997. Species diversity

and distribution of biomass above and below ground among

ectomycorrhizal fungi in an old-growth Norway spruce forest in

south Sweden. Can. J. Bot. 75: 1323–1335.

Dahlberg, A., Schimmel, J., Taylor, A.F.S., and Johannesson, H.

2001. Post-fire legacy of ectomycorrhizal fungal communities in

the Swedish boreal forest in relation to fire severity and logging

intensity. Biol. Conserv. 100: 151–161.

Deacon, J.W., and Fleming, L.V. 1992. Interactions of

ectomycorrhizal fungi. In Mycorrhizal functioning, an integra-

tive plant-fungal process. Edited by M.F. Allen. Routledge,

Chapman and Hall, New York. pp. 249–300.

Esqueda-Valle, M., Perez-Silva, E., Herrera, T., Coronado-

Andrade, M., and Estrada-Torres, A. 2000. Gasteromycete com-

position in a vegetation gradient in Sonora, Mexico. An. Inst.

Biol. Univ. Nac. Auton. Mex. 71: 39–62.

Fisher, R.A., Corbet, A.S., and Williams, C.B. 1943. The relation

between the number of species and the number of individuals in

a random sample of an animal population. J. Anim. Ecol. 12:

42–58.

Fogel, R. 1976. Ecological studies of hypogeous fungi. II.

Sporocarp phenology in a western Oregon Douglas-fir stand.

Can. J. Bot. 54: 1152–1162.

Gamisans, J. 1999. La végétation de la Corse. 2nd ed. Edisud,

Aixen-Provence, France.

Gardes, M., and Bruns., T.D. 1996. Community structure of

ectomycorrhizal fungi in a Pinus muricata forest: above- and

below-ground views. Can. J. Bot. 74: 1572–1583.

Goodman, D.M., and Trofymov, J.A. 1998. Comparison of com-

munities of ectomycorrhizal fungi in old-growth and mature

stands of Douglas-fir at two sites on southern Vancouver Island.

Can. J. For. Res. 28: 574–581.

Harvey, A.E., Jurgensen, M.F., and Larsen, M.J. 1978. Seasonal

distribution of ectomycorrhizae in a mature Douglas-fir/Larch

forest soil in Western Montana. For. Sci. 24: 203–208.

Hawksworth, D.L. 2001. The magnitude of fungal diversity: the

1.5 million species estimate revisited. Mycol. Res. 105: 1422–

1432.

Hill, M.O., and Smith, A.J.E. 1976. Principal component analysis

of taxonomic data with multi-state discrete characters. Taxon,

25: 249–255.

Horton, T.R., and Bruns, T.D. 2001. The molecular revolution in

ectomycorrhizal ecology: peeking into the black-box. Mol. Ecol.

10: 1855–1871.

Huisman, J., and Weissing, F.J. 1999. Biodiversity of plankton by

species oscillations and chaos. Nature (London), 402: 407–410.

Jansen, A.E., and De Nie, H.W. 1988. Relations between

mycorrhizas and fruitbodies of mycorrhizal fungi in Douglas fir

plantations in the Netherlands. Acta Bot. Neerl. 37: 243–249.

Jones, M.D., Durall, D.M., and Cairney, W.G. 2003.

Ectomycorrhizal fungal communities in young forest stands re-

generating after clearcut logging. New Phytol. 157: 399–422.

Keizer, P.J., and Arnolds, E. 1994. Succession of ectomycorrhizal

fungi in roadside verges planted with common oak (Quercus

robu r L.) in Drenthe, The Nederlands. Mycorrhiza, 4: 147–159.

Kernaghan, G., Widden, P., Bergeron, Y., Légaré, S., and Paré, D.

2003. Biotic and abiotic factors affecting ectomycorrhizal diver-

sity in boreal mixed-woods. Oikos, 102: 497–504.

Laganà, A., Loppi, S., and De Dominicis, V. 1999. Relationship

between environmental factors and the proportions of fungal

trophic groups in forest ecosystems of the central Mediterranean

area. For. Ecol. Manag. 124: 145–151.

Last, F.T., Dighton, J., and Mason, P.A. 1987. Successions of

sheathing mycorrhizal fungi. Trends Ecol. Evol. 2: 157–161.

Malloch, D.W., Pirozynski, K.A., and Raven, P.H. 1980. Ecologi-

cal and evolutionary significance of mycorrhizal symbioses in

vascular plants (a review). Proc. Natl. Acad. Sci. U.S.A. 77:

2113–2118.

Manly, B.F.J. 1997. Randomization, bootstrap and Monte Carlo

methods in biology. 2nd ed. Chapman and Hall, London.

Mantel, N. 1967. The detection of disease clustering and a general-

ized regression approach. Cancer Res. 27: 209–220.

Mc Carthy, J. 2001. Gap dynamics of forest trees: A review with

particular attention to boreal forests. Environ. Rev. 9: 1–59.

Médail, F., and Verlaque, R. 1997. Ecological characteristics and

rarity of endemic plants from southeast France and Corsica: im-

plications for biodiversity conservation. Biol. Conserv. 80: 269–

281.

Richard et al. 1723

1724 Can. J. Bot. Vol. 82, 2004

Mehus, H. 1986. Fruit body production of macrofungi in some

north Norwegian forest types. Nord. J. Bot. 6: 679–702.

Molina, R., Massicotte, H.B., and Trappe, J.M. 1992. Specificity

phenomena in mycorrhizal symbioses: community-ecological

consequences and practical implications. In Mycorrhizal func-

tioning, an integrative plant-fungal process. Edited by M.F. Al-

len. Routledge, Chapman and Hall, New York. pp. 357–423.

Nantel, P., and Neumann, P. 1992. Ecology of ectomycorrhizal-

basidiomycete communities on a local vegetation gradient. Ecol-

ogy, 73: 99–117.

Nara, K., Nakaya, H., Wu, B., Zhou, Z., and Hogetsu, T. 2003. Un-

derground primary succession of ectomycorrhizal fungi in a vol-

canic desert on Mount Fuji. New Phytol. 159: 743–756.

Norstedt, G., Bader, P., and Ericson, L. 2001. Polypores as indica-

tors of conservation value in Corsican pine forests. Biol.

Conserv. 99: 347–354.

O’Dell, T.E., Ammirati, J.F., and Schreiner, E.G. 1999. Species

richness and abundance of ectomycorrhizal basidiomycete

sporocarps on a moisture gradient in the Tsuga heterophylla

zone. Can. J. Bot. 77: 1699–1711.

Ohlson, M., Söderström, L., Hörnberg, G., Zackrisson, O., and

Hermansson, J. 1997. Habitat qualities versus long-term conti-

nuity as determinants of biodiversity in boreal old-growth

swamp forests. Biol. Conserv. 81: 221–231.

Oldeman, R.A.A. 1990. Forests: elements of sylvology. Springer-

Verlag, Berlin.

Panaïotis, C., Loisel, R., and Paradis, G. 1995. Dating natural gaps

in the holm oak forest (Quercus ilex L.) in Fango MAB reserve

(Corsica) by reading rings of maquis components. Ann. Sci. For.

52: 477–487.

Panaïotis, C., Carcaillet, C., and M’hamedi, M. 1997. Determina-

tion of the natural mortality age of an holm oak (Quercus

ilex L.) stand in Corsica (Mediterranean Island). Acta Oecol. 18:

519–530.

Peter, M., Ayer, F., Egli, S., and Honegger, R. 2001. Above- and

below-ground community structure of ectomycorrhizal fungi in

three Norway spruce (Picea abies) stands in Switzerland. Can.

J. Bot. 79: 1134–1151.

Quézel, P., and Médail, F. 2003. Écologie et biogéographie des

forêts du bassin méditerranéen. Elsevier, Paris.

Ripley, B.D. 1977. Modelling spatial patterns. J. Roy. Stat. Soc.

B39: 172–212.

Såstad, S.M. 1995. Fungi-vegetation relationships in a Pinus

sylvestris forest in central Norway. Can. J. Bot. 73: 807–816.

Schmit, J.P., Murphy, J.F., and Mueller, G.M. 1999. Macrofungal

diversity of a temperate oak forest: a test of species richness es-

timators. Can J. Bot. 77: 1014–1027.

Signorello, P. 1996. Indagini micocenologiche sulle cenosi a

Quercus ilex L. dell’Etna. Micol. Ital. 1: 74–80.

Skirgiello, A. 1998. Macromycetes of oak-hornbeam forests in the

Bia»owieóa National Park– monitoring studies. Acta Mycol. 33:

171–189.

Smith, S.E., and Read, D.J. 1997. Mycorrhizal symbiosis. 2nd ed.

Academic Press, London.

Smith, J.E., Molina, R., Huso, M.M.P., Luoma, D.L., Mc Kay, D.,

Castellano, M.A., Lebel, T., and Valachovic, Y. 2002. Species

richness, abundance, and composition of hypogeous and

epigeous ECM fungal sporocarps in young, rotation-age, and

old-growth stands of Douglas-fir (Pseudotsuga menziesii)inthe

Cascade Range of Oregon, U.S.A. Can. J. Bot. 80: 186–204.

Tedersoo, L., Koljalg, U., Hallenberg, N., and Larsson, K.-H. 2003.

Fine scale distribution of ectomycorrhizal fungi and roots across

substrate layers including coarse woody debris in a mixed for-

est. New Phytol. 159: 153–165.

Thioulouse, J., Chessel, D., Dodélec, S., and Olivier, J.M. 1997.

ADE-4: a multivariate analysis and graphical display software.

Stat. Comp. 7: 75–83.

Tilman, D. 1992. Resource competition and the community struc-

ture. Princeton University Press, Princeton, New Jersey.

Villeneuve, N., Grandtner, M.M., and Fortin, J.A. 1989. Frequency

and diversity of ectomycorrhizal and saprophytic macrofungi in

the Laurentide Mountains of Quebec. Can. J. Bot. 67: 2616–

2629.

Vogt, K.A., Bloomfield, J., Ammirati, J.F., and Ammirati, S.R.

1992. Sporocarp production by basidiomycetes, with emphasis

on forest ecosystems. In The fungal community: its organization

and role in the ecosystem. Edited by G.C. Carroll and D.T.

Wicklow. Marcel Dekker, Inc., New York. pp 563–581.

Ward, J.H. 1963. Hierarchical grouping to optimise on objective

function. J. Am. Stat. Assoc. 58: 236–244.

Whittaker, R.H. 1972. Evolution and measurement of species di-

versity. Taxon, 21: 213–251.

Yamada, A., and Katsuya, K. 2001. The disparity between the

number of ectomycorrhizal fungi and those producing fruit bod-

iesinaPinus densifolia stand. Mycol. Res. 105: 957–965.

Appendix A

Appendix appears on the following page.

Richard et al. 1725

Species FR* A†SF‡D§P|| C¶Host range Group#

Amanita caesarea (Scop.: Fr.) Pers. 3 37 28.1 (18) 19 9 100–200 Thermophilic F

Amanita citrina (Schaeff.) Pers. 3 29 28.1 (18) 43 25 60–100 Broad host range C

Amanita franchetii (Boud.) Fayod 1 1R1.6 (1) 1 35 60–100 Broad host range D

Amanita gioiosa Curreli ex Curreli 1 1R1.6 (1) 1 7 60–100 Chaparral restricted E

Amanita junquillea Quél. 2 34 31.2 (20) 7 14 50–100 Thermophilic E

Amanita malleata (Piane ex Bon) Contu 1 5R7.8 (5) 10 14 60–120 Broad host range D

Amanita pantherina (D.C.: Fr.) Krombh. 3 26 29.7 (19) 25 21 60–100 Broad host range C

Amanita phalloides (Fr.: Fr.) Link. 2 10 7.8 (5) 43 14 60–120 Broad host range D

Amanita rubescens (Pers.: Fr.) S. F. Gray 1 1R1.6 (1) 1 21 100–180 Broad host range F

Amanita spissa (Fr.) Kumm. 1 3R3.1 (2) 3 7 100–180 Angiosperms linked D

Amanita vaginata (Bull.: Fr.) Vitt. 2 5R7.8 (5) 41 21 60–120 Broad host range D

Aureoboletus gentilis (Quél.) Pouz. 3 20 23.4 (15) 26 14 30–70 Broad host range C

Boletus aereus Bull.: Fr. 1 1R1.6 (1) 1 21 150–300 Thermophilic F

Boletus edulis Bull.: Fr. 1 1R1.6 (1) 1 7 120–250 Broad host range F

Boletus erythropus Pers.: Fr. 1 1R1.6 (1) 1 14 150–200 Broad host range F

Boletus queletii Schulz. 2 2R3.1 (2) 1 21 150–200 Angiosperms linked F

Boletus rhodoxanthus (Krombh.) Kallenb. 1 1R1.6 (1) 1 42 120–200 Thermophilic F

Boletus satanas Lenz 1 1R1.6 (1) 1 14 200–300 Thermophilic F

Cantharellus cibarius (Fr.: Fr.) Fr. 2 54 35.9 (23) 40 42 50–120 Broad host range C

Cortinarius balaustinus J. E. Lange 1 3R3.1 (2) 1 7 40–80 Angiosperms linked D