全 文 ：Received 28 Apr. 2003 Accepted 28 Jul. 2003
Supported by the State Key Basic Research and Development Plan of China (G2000046802), the National Natural Science Foundation of
China (30070018) and the Knowledge Innovation Program of The Chinese Academy of Sciences (KSCX2-SW-101C).
* Author for correspondence. Tel: +86 (0)10 62591431 ext. 6223; E-mail: .
http://www.chineseplantscience.com
Spatial Pattern of the Most Common Late-stage Ectomycorrhizal
Fungi in a Subtropical Forest in Dujiangyan, Southwest of China
LIANG Yu1, GUO Liang-Dong2, MA Ke-Ping1*
(1. Laboratory of Quantitative Vegetation Ecology, Institute of Botany, The Chinese Academy of Sciences, Beijing 100093, China;
2. Systematic Mycology and Lichenology Laboratory, Institute of Microbiology, The Chinese
Academy of Sciences, Beijing 100080, China)
Abstract: Late-stage or later-successional ectomycorrhizal fungi, dominant ectomycorrhizal species in
mature forest, are generally important symbiotic partners of dominant tree species in many forest
ecosystems. Spatial patterns of fungal sporocarps of three families, i.e. Amanitaceae, Boletaceae and
Russulaceae, in a subtropical forest in Dujiangyan were examined using second-order analysis in the present
paper. The woody plant compositions of the plots associated with ectomycorrhizal fungi of three families
were also compared using binary logistic regression analysis. Results indicated that presences of non-
ectomycorrhizal and some ectomycorrhizal plants might have negative effects on the occurrence of
ectomycorrhizas (ECM) fungal sporocarps and the characteristics in clonal growth of fungal taxa would not
be the only determinant in the spatial pattern of ECM fungi. We suggest that besides host plants, non-
ectomycorrhizal woody plants and interaction of ECM fungi should also be considered in spatial studies of
ECM fungal communities in natural forests.
Key words: late-stage ectomycorrhizal fungi; Amanitaceae; Boletaceae; Russulaceae; plant composition;
second-order analysis; binary logistic regression
Ectomycorrhizas (ECM) are symbiotic structures be-
tween plant roots and soil fungi. ECM fungi are estimated
to be 5 000 to 6 000 species (Molina et al., 1992), and over
3% of plant species are ectomycorrhizal. Since the majority
of host plants in these associations are trees, including
many dominant species in temperate and boreal forests,
ECM fungi play important roles in inter- or intra-specific
interactions and maintenance of biodiversity in the eco-
systems (Read, 1991; Read, 1997; Simard et al., 1997; 2002).
ECM Fungi were considered as indicators of succes-
sion stage of forests (Frankland, 1998). Some ECM Fungi,
e.g. Laccaria species, dominate in the early stage of suc-
cession and have broad host range. More host-selective
ectomycorrhizal fungi, e.g. species of Suillus and Lactarius,
however, were generally considered as late-stage species
(Frankland, 1998). The structure and shifts of ECM fungal
communities could therefore imply the successional status
in the plant community.
Fungi were also found to be associated with certain
parts of the root system, i.e. the younger roots with early-
succession species and the older part of the root system
with late-stage species (Ford et al., 1980; Last et al., 1984;
1992; Bigg, 2000). In a model of ectomycorrhizal infection
along a root extending from the stem into the soil, Bruns
(1995) hypothesized that the older, more proximal parts of
root systems can support later-successional species and
the young, mostly distal ends support only early succes-
sion species. The distribution of late-stage ectomycorrhizal
fungi could therefore be a proper indicator of spatial pat-
terns of their hosts.
The diversity and spatial patterns of ECM fungi depend
also on: (a) suitable environmental parameters under the
forest canopy, e.g. temperature, moisture, and light (Gong
et al., 1997); (b) edaphic factors, e.g. soil moisture, depth of
organic matter, and soil pH (Erland and Taylor, 2002); (c)
biological traits of the fungi, e.g. genet size, growth habit
(single, scattered, gregarious, or caespitose), and explora-
tion type of ecto-mycorrhizas; and (d) interactions (negative
or positive relationships) between fungal species. In a lo-
cal forest with relatively homogeneous meteorologic and
edaphic conditions, the spatial patterns of ECM fungi may
be helpful for better understanding the biological traits of
the fungi, the interactions between trees and fungi, and
interactions between ECM fungal species.
In this paper, the spatial patterns of three late-stage
ECM fungal families, i.e. Amanitaceae, Boletaceae, and
Russulaceae in a subtropical forest were studied using sec-
ond-order analysis. Woody plant composition of plots
Acta Botanica Sinica
植 物 学 报 2004, 46 (1): 29-34
Acta Botanica Sinica 植物学报 Vol.46 No.1 200430
associated with sporocarps of the ECM fungi was also com-
pared using binary logistic regression analysis. The objec-
tives of this work were: (1) to find the relationship between
plant compositions and the occurrences of fungal
sporocarps; and (2) to compare the spatial patterns of ECM
fungal sporocarps of different families.
1 Materials and Methods
1.1 Study site
The study site, about 0.06 km2 subtropical woodlands,
was on a low hill in Dujiangyan, west edge of the Sichuan
Basin, Southwest of China (103°27 E, 30°44 N). It had a
mean annual precipitation of 1 244 mm and a mean annual
temperature of 15.2 ℃ (provided by Southwest Subalpine
Botanical Garden, The Chinese Academy of Sciences). The
altitude of the site was about 780 m above sea level and the
zonal vegetation type was subtropical evergreen broad-
leaved forest.
The study was conducted along the ridgeline and the
east hillside, varying from 0o to 35o in slope (Fig.1). The
forest was dominated by evergreen broad-leaved trees and
subtropical conifers (Pinus massoniana Lamb. and
Cunninghamia lanceolata (Lamb.) Hook.).
1.2 Environmental parameters
Moisture and bulk weight of topsoil (0-10 cm) were
measured on September 10, 2001. Leaf area index (LAI),
mean leaf angle (a) and the transmission coefficient for the
diffuse penetration (td) of the study site were determined
by Digital Plant Canopy Imager CI-110 (CID Inc., Vancouver,
Washington State, USA). Environmental parameters and
common woody species of the study site are listed in Table
1. Among the common tree species in the study site,
Castanopsis fargesii Franch., Quercus glandulifera Bl., Q.
variabilis Bl., and P. massoniana were generally observed
as ectomycorrhizal plants (Molina et al., 1992; Gong et al.,
1997).
1.3 Sampling procedure
From July to September 2001, sporocarps of the
Amanitaceae, Boletaceae and Russulaceae were collected
twice a week. Every sporocarp was mapped and those ag-
gregated within a circle of 1 m in diameter were resembled
by one point in the map (Fig.1). Some trees along the hill
ridge were positioned by GPS receiver (EtrexC, Garmin Ltd.,
USA) and further adjusted by field measurement. The spo-
rocarps were then mapped according the relative position
to the marked trees. Centered by each sporocarp, a 5 m×5
m plot was set temporarily and trees and shrubs in the plot
were recorded. As to plots associated with ECM fungi of
particular family, one plot was selected randomly from each
group of overlapped plots and only those having no over-
lap with other plots were finally used for further data
analysis. The woody plant composition of the study site
was also determined by randomly selected 5 m×5 m plots
in an investigation in the early September of 2001.
1.4 Logistic regression analysis
The four groups of plots, i.e. plots associated with spe-
cies of Amanitaceae, Boletaceae, Russulaceae, and plots
randomly selected in the study site, were named Group Am,
Group Bo, Group Ru and Group Ra, respectively. Species
richness of woody plant in each group was represented by
the mean number of woody species per plot of that group.
In each plot, woody plants were score 1 for presence
and 0 for absence. To a plot associated with particular kinds
of ECM fungi, it was scored “1” for presence of that kind of
fungi. And in Group Ra, a plot was scored “0” when it did
not overlap with any of the plots associated with the par-
ticular fungi and was ignored when the plot overlapped
partially with any plots scored “1”. The occurrence of par-
Fig.1. Study site and the position of common Ectomycorrhizas (ECM)
fungal sporocarps.
ticular kinds of ECM fungi in this paper, i.e.
Amanitaceae, Boletaceae and Russulaceae, was then
expressed by the presence or absence of woody
plants using binary logistic regression analysis (SPSS
V10.0 for windows, SPSS Inc., USA). A forward Wald
method was used and probabilities for stepwise were
0.05 for entry and 0.10 for removal.
1.5 Spatial patterns of ECM fungi
The spatial patterns of sporocarps of three ECM
fungal families were analyzed using second-order
analysis, one of the most commonly used methods
for studying the spatial pattern of mapped points
(Andersen, 1992). Second-order analyses based on
Ripley’s (1976) K-function. In the analyses a circle
of radius t was centered in each sporocarp and the
LIANG Yu et al.: Spatial Pattern of the most Common Late-stage Ectomycorrhizal Fungi in a Subtropical Forest 31
number of neighbors within the circle was counted. K(t)
was a function of t, with the expected value of p t2 in ran-
domly arranged points. Ripley (1976) gave an approximately
unbiased estimator for K(t) as
K(t) = An-2 ΣΣwij
-1 It (uij) (1)
where n is the number of sporocarps in the analyzed
plot, A is the area of the plot in m2, wij is a weighting factor
to correct for edge effects, uij is the distance between spo-
rocarp i and j, It(uij) is a counter variable, which is set to 1
if uij ≤ t and 0 if uij > t. wij were obtained according to edge
correction methods of Getis and Frankin (1987), and Goreaud
and Pélissier (1999). Another parameter, L(t) was also used
to determine the spatial pattern of the population. L(t) was
estimated by (Thioulouse et al., 1997; Dale, 1999):
L(t) = √k(t)/p - t (2)
According to a series of t, from 0 to 100 m in this study,
large negative values of L(t) indicate overdisperse and large
positive values indicate clumping.
2 Results
2.1 Species richness of woody plants
Species richness of woody plant per plot of each group
was determined at species, genus and family levels, respec-
tively (Table 2). The woody plant richness of the five groups,
at species, genus or family level, showed similar trends, i.e.
Group Ra> Group Am, Group Bo and Group Ru (P<0.05).
Because plots associated with these ECM fungi had rela-
tively lower species richness, it seemed that sporocarps of
Amanitaceae, Boletaceae and Russulaceae tended to ap-
pear in plots with relatively fewer woody plant species.
2.2 Woody plant composition and occurrences of sporo-
carps
From the plots associated with ECM fungi, 38, 45, and
76 plots were score “1” in Group Am, Group Bo, and Group
Ru, respectively. Out of 27 plots in Group Ra, 21, 18, 10
plots were scored “0” for ECM fungi of Amanitaceae,
Boletaceae and Russulaceae, respectively.
Results of binary logistic regression are listed in Table
3. All the plants in the regression equations were nega-
tively correlated with the occurrence of the three kinds of
ECM fungi. The presence of Amanitaceae sporocarps was
negatively affected by C. fargesii (P = 0.008), Eurya sp. (P
= 0.063), and Ardisia japonica Bl. (P = 0.024). Similarly,
the presence of Boletaceae sporocarps was also negatively
affected by C. fargesii (P = 0.010), E. sp. (P = 0.041), and
A. japonica (P = 0.004). Vaccinium sprengelii (G. Don)
Sleum. was also found negatively correlated with the oc-
currence of Boletaceae sporocarps (P = 0.051). As to the
occurrence of Russulaceae sporocarps, only P. massoniana
had negative effects (P = 0.004).
2.3 Spatial patterns of sporocarps
Results of second-order analysis are shown in Fig.2.
The spatial pattern of Russulaceae sporocarps is aggre-
gated to all the distances. The spatial patterns of
Amanitaceae and Boletaceae are similar for distance less
than about 70 m, i.e. aggregated for distances less than
Table 1 Some parameters of soil and canopy in the study site
Source Parameters Values
Soil Moisture (%) 26.39±0.77
Bulk weight 0.940±0.020
(103 kg/m3)
Canopy Leaf area index 1.779±0.049
parameters a (radian) 0.786±0.125
td 0.228±0.009
Common Broad leaved trees Ardisia japonica
woody species and shrubs Camellia gaudichaudii
Castanopsis fargesii
Eurya sp.
Ilex szechwanensis
Myrisine africana
Quercus glandulifera
Q. variabilis
Smilax china
Symplocos laurina
S. stellaris
Vaccinium sprengelii
Viburnum dilatatum
Conifers Pinus massoniana
Cunninghamia
lanceolata
Data are presented in the format of mean ± SE (n=20)
Table 2 Species richness of woody plants per plot in each group
Group index Species Genus Family
Ra 6.24±0.35a 5.96±0.33a 5.20±0.26a
Am 3.96±0.18b 3.84±0.18b 3.80±0.17b
Bo 3.88±0.20b 3.80±0.19b 3.68±0.19b
Ru 4.04±0.26b 3.96±0.26b 3.88±0.27b
Data are presented as the species, genus or family number per plot in
the format of mean ± SE; same letter denotes non-significant dif-
ference while different letters denote a significant difference (a=
0.05). Am, plots associated with ECM fungi of Amanitaceae; Bo,
plots associated with ECM fungi of Boletaceae; Ra, randomly se-
lected plots; Ru, plots associated with ECM fungi of Russulaceae.
Table 3 Binary logistic regression equations of ectomycorrhizas
(ECM) fungi
ECM fungi Regression equation
Amanitaceae Y = 2.343-1.884 CF-1.899ES-2.705 AJ
Boletaceae Y = 4.487-2.938 CF-2.262 ES-1.9 VS-3.43 AJ
Russulaceae Y = 3.205-2.646 PM
AJ, Ardisia japonica; CF, Castanopsis fargesii; ES, Eurya sp; PM,
Pinus massoniana; VS, Vaccinium sprengelii.
i¡Ùj
∧
∧
∧
∧
Acta Botanica Sinica 植物学报 Vol.46 No.1 200432
about 40 m and appeared in a Poisson pattern for a larger
distance. An overdispersal pattern of Amanitaceae sporo-
carps is also found for distances between 70 and 100 m.
3 Discussion
ECM fungal species generally have different host range.
They may form ectomycorrhizas with diverse hosts, with
only angiosperms or gymnosperms, or with plants restricted
to a family or a genus (Molina et al., 1992; Dahlberg, 2001).
Although the structure and dynamics of ECM fungal
communities are influenced by many factors, host plant
composition is generally considered as a major determi-
nant (Dahlberg, 2001). Plants in the forest, including host
plants or non-host plants, could act as either symbiotic
partners or competitors for nutrients and/or space to a par-
ticular kind of ECM fungi. The species richness and spatial
distribution of plants, especially woody plants in the forest,
should therefore be fundamental determinants to the com-
position of ECM community. Nantel and Neumann (1992)
calculated similarity matrices of fungal species abundance
and woody species abundance among stations in south-
ern Québec, Canada. They found that similarity among ECM
fungal communities was strongly and significantly corre-
lated with tree community similarity. Såstad (1995) also be-
lieved that ECM fungi were mainly influenced by plant spe-
cies present in the root zone. Although big trees out of the
investigated plots in our study might also have effects on
the formation of ECM fungal sporocarps, differences in
woody plant composition between ECM fungi-associated
and non-ECM fungi-associated micro-sites could still be
useful for understanding the correlation between ECM fungi
and woody plants.
While environmental parameters and plant compositions
in the study site were suitable for ECM fungi of Amanitaceae,
Boletaceae and Russulaceae, they usually tended to be
mosaic on a scale of micro-sites. As to particular kinds of
ECM fungi, some of the micro-sites were preferable and
some were not. Our results that all the ECM fungi tended to
appear in plots with lower woody species richness might
be due to competition between ECM fungi and non-
ectomycorrhizal plants. Since most plants species in the
study site belonged to non-mycorrhizal or endo-mycorrhizal
families and only four species were ectomycorrhizal accord-
ing to Molina et al. (1992), high species richness implied an
increased probability of occurrences of competitors with
similar niche to the ECM fungi in the micro-environment,
which would inhibit colonization and fruiting of ECM fungi.
For similar reasons the non-ectomycorrhizal plants, i.e.
E. sp., A. japonica and V. sprengelii were found to be nega-
tively correlated with ECM fungal species of Amanitaceae
and Boletaceae, respectively. In this study, C. fargesii was
also found negatively correlated with ECM fungal species
of Amanitaceae and Boletaceae and P. massoniana was
negatively correlated with ECM fungal species of
Russulaceae. Since P. massoniana and C. fargesii are
ectomycorrhizal trees, we assume that: (1) the negatively
correlated plants were not hosts to the particular fungi, i.e.
P. massoniana was not host of Russulaceae ECM fungi
and C. fargesii was not host of Amanitaceae and Boletaceae
ECM fungi in the present study; (2) since the binary re-
gression showed only statistics results, a small fraction of
species in a particular fungal family might also have oppo-
site trends to the whole family; and (3) competition for root
tips between ECM fungal family might also be involved,
that is, although a Amanitaceae species could form
ectomycorrhizas with C. fargesii, it might also be absent in
the plots where Russulaceae species were dominant.
The sporocarps of three ECM fungal families appeared
in different spatial patterns to large distances in the study
site, indicating their differences in vegetative growth ability,
or heterogeneity in host distribution or edaphic traits (the
growth habit affected the spatial pattern only in a very
limited region). As to the vegetative growth of ECM fungi,
species of Boletaceae were generally found to have large
Fig.2. Spatial patterns of fungal sporocarps using second-order
analysis. The dash lines indicate boundaries of 99% confidence
interval. L(t): estimator for L(t), a second order parameter in spa-
tial pattern analysis.
∧
LIANG Yu et al.: Spatial Pattern of the most Common Late-stage Ectomycorrhizal Fungi in a Subtropical Forest 33
genets (Dahlberg and Stenlid, 1994; Fiore-Donno and
Martin, 2001) and species of Russulaceae and Amanitaceae
were found to have relatively small genets (Redecker et al.,
2001; Bergemann and Miller, 2002). It was interesting that
although having a similar genet size in the relative homoge-
neous environment of a local forest, the sporocarps of
Amanitaceae aggregated in relatively short distances
whereas the sporocarps of Russulaceae appeared in an
aggregated pattern to quite larger distances. This could be
explained by the interactions between trees and ECM fungi.
Since the occurrences of ECM fungal sporocarps of
Amanitaceae and Boletaceae were negatively correlated
with C. fargesii, a dominant species in the overstory of the
forest (accounting for 45.37% of the canopy coverage in a
fine scale plant investigation, unpublished data), the size
of aggregations in their spatial pattern should therefore be
limited. As the relative abundance of P. massoniana was
relatively lower (6.54% according to a recent investigation,
unpublished data), the spatial pattern of Russulaceae spo-
rocarps was less inhibited.
Subtropical or tropical forests (pure Eucalyptus or
Dipterocarpaceae forests are not included) are always much
complex. The forests generally have diverse plants, includ-
ing many non-ectomycorrhizal trees, and the mycelia net-
works underground could be discontinuous or appear in a
mosaic pattern. Not only the spatial distribution of host
plants, but also the spatial pattern of non-host plants and
other competitive ECM fungi should therefore be consid-
ered while studying the spatial pattern of ECM fungal com-
munity in subtropical or tropical forests.
Because it is time costing to record the number, height
and canopy parameters of all the plants in each plot, only
binary information, i.e. present or absent of a species, were
used in this study. Further studies based on fixed plots in this
forest (unpublished data) will provide more detailed information,
which enables us to evaluate the relative importance of host
and non-host plants, and other ECM fungi in determining the
spatial pattern of a particular kind of ECM fungi.
Acknowledgements: We are grateful to Southwest Subal-
pine Botanical Garden, The Chinese Academy of Sciences
for providing climate data of the study site and to Dr. GAO
Xian-Ming, Dr. DU Xiao-Jun, Dr. ZHUANG Ping and Mr.
LIU Can-Wei for identification of plant species in the study
site. We also thank Dr. GUO Qing-Feng, Dr. YE Wan-Hui,
Dr. YU Ming-Jian, Dr MA Ke-Ming and three anonymous
reviewers for their suggestions and comments on an early
version of this paper.
References:
Bergemann S E, Miller S L. 2002. The size, distribution, and
persistence of genets in local populations of the late-stage
ectomycorrhizal basidiomycete, Russula brevipes. New Phytol,
156:313-320.
Bigg W L. 2000. Fungal succession and diversity in ectomycorrhizal
associations: a case study approach. USDA Forest Service
General Technical Report. Proceedings of the California For-
est Soils Council Conference on Forest Soils Biology and For-
est Management, Sacramento, California. 43-52.
Bruns T D. 1995. Thoughts on the processes that maintain local
species diversity of ectomycorrhizal fungi. Collins H P, Klug
M J. The Significance and Regulation of Soil Biodiversity.
Dordrecht, the Netherlands: Kluwer Academic Publishers. 63-
73.
Dahlberg A. 2001. Community ecology of ectomycorrhizal fungi:
an advancing interdisciplinary field. New Phytol, 150:555-
562.
Erland S, Taylor A E S. 2002. Diversity of ecto-mycorrhizal fun-
gal communities in relation to the abiotic environment. van der
Heijden M G A, Sanders I R. Mycorrhizal Ecology. New
York: Springer- Verlag. 163-200.
Fiore-Donno A M, Mart in F. 2001. Populat ions of
ectomycorrhizal Laccaria amethystine and Xerocomus spp.
show contrasting colonization patterns in a mixed forest. New
Phytol, 152:533-542.
Ford E D, Mason P A, Pelham J. 1980. Spatial patterns of sporo-
phore distribution around a young birch tree in three succes-
sive years. Trans Brit Mycol Soc, 75:287-296.
Frankland J C. 1998. Fungal succession-unraveling the
unpredictable. Mycol Res, 102:1-15.
Gong M-Q, Chen Y-L, Zhong C-L. 1997. Mycorrhizal Research
and Application. Beijing: China Forestry Publishing House.
35-60. (in Chinese)
Last F T, Mason P A, Ingleby K, Fleming L V. 1984. Succession
of fruitbodies of sheathing mycorrhizal fungi associated with
Betula pedula. Forest Ecol Manag, 9:229-234.
Last F T, Natarajan K, Mohan V, Mason P A. 1992. Sequences of
sheathing (ecto-) mycorrhizal fungi associated with man-made
forests, temperate and tropical. Read D J, Lewis D H, Fitter A
H, Alexander I J. Mycorrhizas in Ecosystems. Wallingford
UK: CAB International. 214-219.
Molina R, Massicotte H, Trappe J M. 1992. Specificity phe-
nomena in mycorrhizal symbioses: community-ecological con-
sequences and practical implications. Allen M F. Mycorrhzial
Functioning: an Integrative Plant-Fungal Process. New York:
Chapman and Hall. 357-423.
Acta Botanica Sinica 植物学报 Vol.46 No.1 200434
(Managing editor: HAN Ya-Qin)
Nantel P, Neumann P. 1992. Ecology of ectomycorrhizal-basidi-
omycete communities on a local vegetation gradient. Ecology,
73:99-117.
Read D J. 1991. Mycorrhizas in ecosystems. Experientia, 47:
376-391.
Read D J. 1997. Mycorrhizal fungi ¡ª the ties that bind. Nature,
388:517-518
Redecker D, Szaro T M, Bowman R J, Bruns T D. 2001. Small
genets of Lactarius xanthogalactus, Russula cremoricolor and
Amanita francheti in late-stage ectomycorrhizal successions.
Mol Ecol, 10:1025-1034.
Såstad S M. 1995. Fungi-vegetation relationships in a P. sylvestris
forest in central Norway. Can J Bot, 73:807-816.
Simard S W, Jones M D, Durall D M. 2002. Carbon and nutrient
fluxes within and between mycorrhizal plants. van der Heijden
M G A, Sanders I R. Mycorrhizal Ecology. New York:
Springer-Verlag. 34-74.
Simard S W, Perry D A, Jones M D, Myrold D D, Durall D M,
Molina R. 1997. Net transfer of carbon between
ectomycorrhizal tree species in the field. Nature, 388:579-
582.