Content uploaded by Andy FS Taylor
Author content
All content in this area was uploaded by Andy FS Taylor on Jan 27, 2014
Content may be subject to copyright.
Plant and Soil 244: 19–28, 2002.
© 2002 Kluwer Academic Publishers. Printed in the Netherlands. 19
Fungal diversity in ectomycorrhizal communities: sampling effort and
species detection
Andy F. S. Taylor
Department of Forest Mycology and Pathology, Swedish University of Agricultural Sciences, PO Box 7026, SE-750
07 Uppsala, Sweden∗
Key words: community structure, ectomycorrhizas, fungal distribution, species richness, vertical distribution
Abstract
A number of recent review articles on ectomycorrhizal (ECM) fungal community diversity have highlighted the
unprecedented increase in the numberof publications on this ecologically important but neglected area. The general
features of these species-rich, highly dynamic and complex communities have been comprehensively covered but
one aspect crucial to our assessment of diversity, namely the sampling of ECM communities has received less
attention. This is a complex issue with two principal components, the physical sampling strategy employed and
the life cycle traits of the ECM fungi being examined. Combined, these two components provide the image that
we perceive as ECM diversity. This contribution will focus primarily on the former of these components using
a recent study from a pine forest in central Sweden to highlight some sampling problems and also to discuss
some features common to ECM communities. The two commonly used elements of diversity, species richness
and community evenness, present rather different problems in the assessment of ECM diversity. The applicability
of using current measures of abundance (number or percentage of root tips colonised) to determine community
evenness is discussed in relation to our lack of knowledge on the size of individual genets of ECM fungi. The
inherent structure of most ECM communities, with a few common species and a large number of rare species,
severely limits our ability to accurately assess species richness. A discussion of theoretical detection limits is
included that demonstrates the importance of the sampling effort (no. of samples or tips) involved in assessing
species richness. Species area abundance plots are also discussed in this context. It is suggested that sampling
strategy (bulk samples versus multiple collections of single tips) may have important consequences when sampling
from communities where root tip densities differ. Finally, the need for studies of the spatial distribution of ECM
on roots in relation to small-scale soil heterogeneity and of the temporal aspects of ECM community dynamics is
raised.
Introduction
Recent years have witnessed an unprecedented in-
crease in the number of studies examining the species
composition of ectomycorrhizal (ECM) communities
below ground. While this can be primarily attributed
to rapid progress in the ability to characterise the fungi
colonising individual root tips, there is also an in-
creasing awareness that soil organisms, in particular
mycorrhizal fungi, can have a significant influence
upon aboveground organism dynamics and on ecosys-
tem processes as a whole (Copley, 2000; Read, 1998;
van der Heijden et al., 1998).
∗FAX No: 18309245. E-mail: Andy.Taylor@mykopat.slu.se
In a recent review on diversity in ECM communit-
ies, Horton and Bruns (2001) gave an excellent and
comprehensive account of the general characteristics
of ECM communities. It is, therefore, not my inten-
tion here to give an overview of ECM fungal diversity
but rather to focus on a number of interconnected
issues relating to the sampling of ECM communit-
ies. There has so far been little attention given to the
most suitable sampling methods for assessing ECM
diversity but it is fundamental to how we perceive
community diversity. Given the high species richness
and the apparent non-random distribution of species,
it is critical that we establish sampling protocols that
can accurately determine ECM diversity.
20
One purpose of the present paper is to examine
the relationship between what we know about ECM
community structure and what it is hypothetically pos-
sible to determine about that structure with different
sampling efforts. I will give a brief summary of how
we define diversity and then use data from a recent
study carried out in Sweden to illustrate a number
of points relating to sampling effort and species de-
tection. This data will also be used to discuss some
general features that appear to be common to many
ECM communities. Although highly relevant for com-
paring differences between ECM communities, I have
not included a discussion on diversity indices in re-
lation to ECM communities. This topic is, by itself,
sufficiently large and complex to warrant a separate
investigation. The general use of these indices has
been covered elsewhere (see Krebs, 1989; Magurran,
1988). The primary aim of the following discussion
is to demonstrate that we must exercise caution in
our interpretation of diversity data, particularly when
comparing the species richness of two communities.
Defining community diversity
Most studies that have examined ECM communities
have done so over rather small areas, usually <1–
2 ha (see Bruns, 1995; Horton and Bruns, 2001). It
therefore seems practical and convenient to use this
as the spatial scale at which to discuss sampling of
ECM communities. It should be apparent that changes
in the scale at which we define communities will have
affects on both our ability to accurately determine the
species within that area and upon the potential number
of species that can occur within the defined area.
Currently, sampling of ECM communities below-
ground usually involves taking bulk samples of soil
with subsequent analysis of the mycorrhizal fungi
present on the root tips. Characterisation of the fungi
may be done using morphology or molecular tech-
niques or now more commonly a combination of
both. A joint approach to characterising the ECM
on roots involving a limited amount of morphotyping
followed by molecular characterisation of individuals
from the morphotypes distinguished has been advoc-
ated by Horton and Bruns (2001). Once characterised,
community diversity is usually reported as two com-
ponents, species richness, the number of species in
the community and community evenness, a measure
of the abundance of each species in the community
(Magurran, 1988).
Most studies involve single sampling events, which
effectively give a snapshot in time of the community.
Under these conditions, it may be important to dis-
tinguish two different aspects of species richness; ap-
parent and cryptic species richness. Apparent species
richness is that which can be distinguishedon the roots
at a particular moment in time for a given sampling
effort. Cryptic species are species either below the de-
tection limits of the study at that time or those existing
only as persistent propagules (spores and sclerotia).
These species may only become apparent after tem-
poral changes in species abundance or, in the case of
the propagules, after perturbation events e.g. wildfire
(Baar et al., 1999; Taylor and Bruns, 1999).
Community evenness is usually reported as the
number or percentage of root tips colonised by in-
dividual species. A small number of studies have
estimated the biomass of species based of estimates
of the amount of fungal material associated with the
mycorrhizal tips (e.g. Stendell et al., 1999) but our
knowledge of the interspecific differences in the ex-
traradical phase of the symbiotic association is such
that these estimates may have limited value. However,
biomass provides an effective estimate of the mass of
roots colonised by individual ECM species and, by
derivation, root-associated fungal biomass.
There is a problem inherent within current
sampling techniques to using the concept of even-
ness. For the purpose of defining evenness, we assume
that within bulk samples, root tips represent individu-
als, but there is a high likelihood that mycorrhizal
tips within a sample identified as supporting the same
fungal species are colonised by the mycelium of one
individual. We are therefore effectively sampling parts
of the same individual. Horton and Bruns (2001) have
discussed some of the problems associated with defin-
ing individuals within ECM species and, in addition,
give an interesting discussion on the use of frequency
versus abundance and demonstrate that both can give
biased views of a species importance by emphasising
different aspect of a species autecology.
Thecasestudy
The study (hereafter referred to as the Riddarhyttan
study) was carried out in a 50-year-old Pinus sylvestris
L. stand ca. 100 km SW of Uppsala in central Sweden.
It is a typical dry, lichen-heath boreal pine forest on
well-drained sand. In 2000, the site was visited on
5 occasions, at two-week intervals, during the main
21
period of sporocarp production. At each visit, the
sporocarps from three plots (30 m ×30 m) were col-
lected and identified to species. In addition, 10 soil
cores (2.8 cm d.) were taken from each plot during
the same period. Each core included the whole of
the organic layer. The mycorrhizal tips were extracted
from the cores and characterised using a combination
of morphotyping (Agerer, 1986–98) and restriction
fragment length polymorphism (RFLP) analysis of the
ITS region (Kåren et al., 1997). For the purpose of
the present discussion, data from the three plots are
combined. In total, 1478 sporocarps were collected be-
longing to 56 species and 5371 mycorrhizal tips were
extracted and characterised into 37 morphotypes, of
which only 19 could be identified to genus or species
level.
It is now well established that the production of
sporocarps, the reproductive structures of most ECM
fungi, is not a good indicator of the abundance of ECM
species belowground (see Dahlberg, 2001). The reas-
ons for this discrepancy are numerous and have been
discussed on several occasions (e.g. Erland and Taylor,
2002). At Riddarhyttan, Cortinarius was the most
abundant genus forming 625 (42.3%) sporocarps rep-
resenting 25 spp. However, belowground only 1.6%
of the mycorrhizal tips examined could be attributed
to this genus. Similar patterns have been reported
previously (Gardes and Bruns, 1996). It is also com-
mon to find greater numbers of species occurring as
sporocarps than recorded belowground (e.g. Dahlberg
et al., 1997). Horton and Bruns (2001) pointed out
this may reflect the common practice, as carried out
here, of repeated observations on the sporocarps but
only single sampling of the mycorrhizal tips. They
suggested that an equally intensive study of the my-
corrhizal tips might lead to greater similarity between
above and belowground views. It is, however, diffi-
cult to see how this effort may be achieved. In the
present study, there were almost four times as many
root tips sampled as sporocarps. If it were to be done
on an area basis, the numbers of root tips would be
impossible to handle. There were between 7 and 72 ×
104mycorrhizal tips per square metre of forest floor in
the present study. Other studies have recorded similar
figures (e.g. Dahlberg et al., 1997).
Vertical distribution of species
As mentioned above, there are a number of possible
reasons for the disparity between above and below-
ground views of ECM communities. One factor that
has so far received little attention is the vertical dis-
tribution of species down the soil profile. A number
of studies have demonstrated that ECM species differ
in their preference for the organic and mineral soil
layers (Danielson and Visser, 1989; Stendell et al.,
1999; Taylor and Bruns, 1999). However, the un-
equal distribution of root tips between the organic and
mineral layers in the soil profile must be taken into
consideration when making observations on distribu-
tion. Recently, Fransson et al. (2000) used a binomial
statistical model that incorporated this feature into an
analysis of species distribution between organic and
mineral soil layers. They showed that Cenococcum
geophilum mycorrhizas were preferentially found in
the organic layers while those of Tylospora fibrillosa
were associated with the mineral soil. The useful-
ness of this model was however restricted by the large
variation in the data from a single species.
Most studies of ECM communities, including the
present one, restrict sampling to the organic horizons.
Species that occur further downin the profile are there-
fore effectively excluded from the analysis. In a recent
study carried out in northern Sweden in a mixed con-
iferous forest growing on a well-developed podsolic
soil, 25 ECM taxa were identified on the roots by
sequencing of the ITS region. Fifteen of these were
found exclusively in the mineral soil (Rosling et al.,
unpublished). This study illustrates, not just a poten-
tial reason for above and below discrepancies, but also
highlights a potential problem in determining the spe-
cies richness of a community. The species restricted
to the mineral soil constitute part of the hidden or
cryptic community and sampling strategies need to
incorporate this phenomenon.
Species rank abundance patterns, theoretical
detection limits and sampling effort
In common with most other studies on ECM diversity
(Horton and Bruns, 2001), the sporocarp and the
mycorrhizal data from the Riddarhyttan pine study
comprised a few common species and a long ‘tail’ of
rarer species (Figure 1a, b). How does this distribution
pattern affect our ability to establish the species rich-
ness of a community and to detect differences between
communities as a result of perturbations? If we know
the relative abundance of a species in a community, it
is possible to calculate the probability (p) of finding
22
Figure 1. Rank abundance patterns for ectomycorrhizal (ECM)
sporocarps (a) and ECM morphotypes (b) using log abundance
(log10 no. of sporocarps formed and log10 no. of tips colonised)
plotted against ranked abundance. Data from a 50-yr-old pine stand
in central Sweden.
Figure 2. Theoretical detection limits when sampling ectomycor-
rhizal communties belowground. Relationship between sampling
effort (number of tips sampled), probability of detection and the
relative abundance of species (percentage of root tips colonized).
The vertical lines indicate the relative abundance at each sampling
effort (10 tips - , 20 tips - , 50 tips - , 100 tips - •) that species
must make up in order to stand a 50% and a 95% chance of being
detected.
that species in a sample of a given size:
probability of not finding sp. Aon a tip..........
p=1−x
probability of not finding sp. Aon yroot tips...
p=(1−x)y
probability of finding sp. Aon yroot tips.......
p=1−(1−x)y,(1)
where xis the proportion of species Ain the com-
munity and yis the number of individuals (in this case
root tips) sampled. By changing formula (1) to:
x=1−y1−p(2)
it is possible to calculate the relative abundance at
which species have a given probability of being found
in samples. Figure 2 shows the relationships between
sampling effort (at 4 different levels; 10, 20, 50 and
100 tips sampled), the probability of detection and
the relative abundance of species. The importance
of abundance obviously increases with increasing de-
mands on the probability of detection. If we accept an
even chance of detecting species in a community, then
we have a 50% chance of detecting species that make
up 6.69, 3.41, 1.37 and 0.69% of the individual my-
corrhizal tips when we sample 10, 20, 50 and 100 tips,
respectively. However, with an increase in the level
of certainty to 95%, there is a steep increase in the
required abundance of species to 25.88, 13.91, 5.82
and 2.95% for the four levels of sampling effort.
How do these levels relate to the relative abund-
ance of ECM species recorded on the root tips at
Riddarhyttan? Figure 3 shows a rank abundance plot
of the mycorrhizal data using the relative abundance
(%) of root tips colonised by each morphotype. The
dotted lines represent the lowest percentage of mycor-
rhizas, at each level of sampling effort, that a species
must form in order to have a 95% chance of being
recorded in an analysis of the community. For ex-
ample when 100 tips are sampled, only seven out of
the 37 morphotypes occur on or above the dotted line
at 2.95% i.e. 30 of the morphotypes have a greater
than 5% chance of remaining undetected. The lowest
recorded abundance of a morphotype at Riddarhyttan
was only 3 tips out of the total of 5371 tips or 0.055%.
The calculated number of tips that would need to be
sampled in order to stand a 95% chance of detect-
ing this species is 5500, which is remarkably close
to the actual number of root tips sampled which was
5371. These theoretical probabilities of detection are
based upon the assumption that species are randomly
distributed within the community. We still know very
little about the spatial distribution of species, but what
we do know suggests that species have a clumped
distribution rather than being randomly distributed.
A clumped distribution would have the effect of de-
creasing even further our ability to obtain accurate
assessments of species richness.
Horton and Bruns (2001) suggested that since
most ECM species typically occur in less than 10%
23
Figure 3. Rank abundance plot, using percentage of root tips colonised, of the ectomycorrhizal morphotypes found on roots of 50-yr-old
pine stand in central Sweden. Overlaid on the plot are the 95% theoretical detection probabilities (dotted lines) at four different sample sizes.
Morphotypes with a relative abundance that occurs on or above these lines have a 95% or better chance of being detected in four sample sizes.
of soil samples, the distribution of many ECM spe-
cies is clustered. At Riddarhyttan, almost 84% of the
morphotypes occurred in three or fewer samples out
of a total of 30. However, the occurrence of species
in samples will be greatly influenced by sample size.
In addition, low frequency of occurrence in samples,
which actually represent a very small proportion of
the forest, does not necessarily mean that a species
is not randomly dispersed across a site. There is no
doubt that where some species occur they form the
great majority of the tips in samples, but this is not
concomitant with a non-random distribution of indi-
vidual mycelia of the species across the site. If the
phenomenon of clustering within samples is signific-
ant then the number of species occurring in individual
samples may deviate from a random distribution. This
was tested on the Riddarhyttan morphotype data using
the Chi-squared test of dispersion (Heath, 1995; Sokal
and Rohlf, 1996). This is a simple analysis that exam-
ines whether the frequency distribution of organisms
in samples is random, clumped or regular. If clustering
of tips of a single species influences the numbers of
species found in samples it should do so such that the
frequency of species in samples should be more reg-
ular than expected from a random distribution i.e. the
ratio of variance/mean (I– index of dispersal) should
be significantly less than 1. The value of Ifor the Rid-
darhyttan data was 0.787. This is used to calculate a
value of χ2using I∗(n-1), giving a value of 22.28,
which is well within the expected range of values
(16.04 – 45.72) at α2= 5% for a random distribution.
However, the distribution of the number of species in
samples will be greatly affected by sample size. De-
creasing sample size to the point where it equals that
Figure 4. Species area plots for ectomycorrhizal morphotypes
found in a study of the mycorrhizal community in a 50-yr-old pine
stand in central Sweden. The cumulative number of morphotypes is
plotted against number of samples (a) and the cumulative number of
root tips (b) examined.
of the mycorrhizal clusters should give a more regular
distribution of species within samples. However, this
is a very unlikely scenario because in order to achieve
this we would have to sample a discrete 3-D block
of soil equal in size to the mycorrhizal clusters. Con-
ventional bulk samples invariably constitute a vertical
section through the entire organic material and will
thus potentially contain many clusters.
24
There are increasing numbers of published reports
on the distribution of genets of individual species (see
Horton and Bruns, 2001). Where these include de-
tailed mapping of individuals on a grid system, it is
possible to analyse the distribution of genets using
the Chi-squared test of dispersion mentioned above.
Gherbi et al. (1999) examined the genetic structure
of a population of Laccaria amethystina (Bolt. ex
Hooker) Murr. in a closed 150-year-old beech (Fagus
sylvatica L.) forest in the Vosges mountains in France.
They found a very high density of small genets (>100
per 100 m2). Analysing the distribution of genets
shown in Figure 5 of Gherbi et al. (1999) by assign-
ing each genet to a single 1 m ×1 m square reveals
that the distribution of genets is clumped. Increas-
ing the sampling area to 2 ×2 m also gives a very
clumped distribution pattern. Gherbi et al. (1999) sug-
gested that L. amethystina repeatedly colonises the
root system of beech more or less on an annual basis.
If this is the case and the main source of inoculum is
from spores released from sporocarps, then a clumped
distribution pattern of the resultant genets is hardly
surprising, since most of the spores will fall in the
vicinity of the sporocarps. A similar analysis of the
data on Russula cremoricolor genets reported by Re-
decker et al. (2001, Figure 4a) also reveals a clumped
distribution. The authors, like Gherbi et al. (1999) for
L. amethystina, suggested that genets of R. cremori-
color were short lived and recolonised from spores. It
is, therefore, not surprising that the two species show
similar distribution patterns.
There are several reports of much larger genets (>
15–20 m) than the two examples given above (e.g.
Suillus bovinus, Dahlberg and Stenlid, 1990; Laccaria
bicolor, Baar et al., 1994; Pisolithus tinctorius, Ander-
son et al., 1998; Cortinarius rotundisporus,Sawyer
et al., 1999). Sampling in communities primarily
composed of large genets should result in a more reg-
ular distribution of individual species within samples.
However, this would only be the case where the integ-
rity of the mycelium of the genets was more or less
intact. Until we are able to examine genet size be-
lowground, we lack knowledge on mycelial integrity
and indeed of genet longevity. The apparent exist-
ence of large differences in genet size poses a number
of problems for sampling ECM communities. If our
objective is to assess the abundance of individuals
within a species, then the minimum distance between
samples must be greater than the size of the individuals
being investigated. Many more studies on the genet
size and life strategies of ECM fungi, including the
same species under different ecological conditions, are
required before we can adapt sampling strategies to
match individual species.
Species area abundance relationships
It is clear from the above examination of theoretical
constraints on detection and the data from Riddarhyt-
tan that the skewed abundance distribution patterns of
species within ECM communities must have a consid-
erable influence upon our ability to gain an accurate
assessment of species richness within communities.
A visual assessment of the accuracy of our observed
estimate of species richness may be gained by plot-
ting the cumulative number of species found against
the cumulative number of samples. This type of plot
may be considered as a species area plot where area is
replaced by sample (see Figure 3, Horton and Bruns,
2001). If our estimate of species richness is close to
the actual value then the plot of these two parameters
should level off indicating that we have sampled suffi-
cient samples to detect most of the species present in
the community. There is little tendency for the species
area plots for the mycorrhizal data from Riddarhyttan
to level off (Figure 4a,b). This is the case when the
cumulative number of species is plotted against either
the number of samples or the number of tips sampled.
This close relationship between sampling and ob-
tained species richness has some important implica-
tion when sampling ECM communities. In particular,
when we compare two or more communities with re-
spect to species richness, it is vital that we have some
idea of how accurate the assessment of species rich-
ness is within each community. For example, if there
are two communities, A and B, that are similar in
structure but B actually contains fewer species, then it
would be easy to assume that the species richness was
the same in both communities if we do not consider
how sampling effort and species richness is related in
these two communities. This is illustrated graphically
in Figure 5a. A sampling effort below the line crossing
the two community plots would result in the conclu-
sion that the two communities do not differ. Increasing
the sampling effort beyond this point would increas-
ingly illustrate that the species richness is actually
quite different. Another possible scenario is illustrated
in Figure 5b. Here the species in community B are dis-
tributed in a more regular pattern than in community A
and this results in more species being found in B for a
given sampling effort. In this case if we sample below
25
Figure 5. The potential influence of sampling below the asymptote of species area plots on the number of species recorded from ectomycorrhizal
communities. (a) Sampling in communties with similar species distributions but which differ in species richness and (b) as (a) but where the
species in community B are more regularly distributed than in community A.
the indicated line, then we could reach the conclusion
that there were more species in B.
Given the species-rich nature of ECM communit-
ies and the high degree of spatial heterogeneity it will,
in most cases, be impossible to sample all species
within a community. For example, even with intens-
ive sampling over a two-year period, only 22% of the
species found fruiting on three Norway spruce sites
were found belowground (Peter et al., 2001a). There
are a number of statistical techniques that can provide
an estimate of the expected number of individuals in
communities from which a subsample has been taken
(e.g. Jackknife estimate, bootstrap method). These
techniques are based on using the community struc-
ture (relative abundance of individuals) observed in
the subsample to estimate the potential species rich-
ness within the community as a whole. However, given
the difficulties in defining individuals within ECM
communties, the use of these techniques could be
questioned.
It is also possible to use the method of rarefac-
tion (Krebs, 1989; Magurran, 1988) to compare the
expected number of species in samples of different
sizes. This technique was employed by Taylor et al.
(2000) to compare ECM diversity in samples of dif-
ferent sizes from Picea abies (L.) Karst. and Fagus
sylvatica L. forests along a N-S transect in Europe.
Peter et al. (2001b) also used this method to estim-
ate the number of expected RFLP patterns likely to
be found in a standard sample size from a fertilization
experiment in a Picea abies forest in Switzerland. Al-
though this method is useful, it can only be used to
compare expected species richness up to the level of
the community (or sample) where the lowest number
of individuals were sampled.
There is another aspect to consider in relation to
species area abundance relationships and that is the
difference between sampling bulk samples containing
many mycorrhizal tips and sampling individual my-
corrhizal tips at random from the same community.
This may be important when comparing two com-
munities that are very similar in all aspects but where
the root tip density at one site is significantly lower
than the other. Changes in root tip numbers are com-
monly recorded as a result of perturbations. Fertil-
ization with N often results in a reduction in root
tips (Wallenda and Kottke, 1998), while liming, as a
counter measure against forest soil acidification, often
results in considerable increases in root tip numbers
(Kreutzer, 1995). Figure 6 illustrates the potential
difference between taking bulk samples and multiple
collections of single tips as samples from two com-
munities, A and B. Community B is found on a site
where there is a lower density of root tips. Due to the
strong relationship between root tips sampled and spe-
cies found (see Figure 4), taking the same number of
bulk samples from each community (Figure 6a) may
result in a lower species richness being recorded for
B. However, taking individual samples consisting of
single root tips from both A and B (Figure 6b), should
in theory result in similar species richness being recor-
ded for both communities. This approach was used to
good effect recently by Peter et al. (2001b) to investig-
ate the effects of fertilization upon ECM communities.
Sampling individual root tips at random to assess
ECM diversity may have inherently greater potential
for determining both species richness and community
evenness than taking bulk samples. It would be very
instructive to compare these two sampling strategies.
26
Figure 6. The potential effect of taking bulk samples (a) and individual tips (b) to assess species richness in two similar ectomycorrhizal
communities, A and B, where the root tip density in B is significantly less than in A.
Spatial heterogeneity and sampling
Heterogeneity, both spatial and temporal, within soil is
high and it has been suggested that this is a significant
factor in maintaining the high diversity of ECM com-
munities (Bruns, 1995; Erland and Taylor, 2002) and
of other soil organisms (Giller, 1996). With the excep-
tion of a small number of studies that have examined
ECM distribution at a gross scale in soil (e.g. Harvey
et al., 1976, 1978; Kropp, 1982), there are as yet no
studies that have specifically examined the microspa-
tial distribution of ECM fungal species belowground
in relation to the physicochemical soil environment.
Bebber (1999) recently urged ecologists to get ac-
quainted with geostatistics as a powerful tool in the
analysis of spatial and temporal heterogeneity. There
have been no studies, as far as I am aware, that have
utilised geostatistics to examine soil heterogeneity,
both horizontal and vertical, at fine scales (cm – dm)
in relation to ECM fungi. Most of what is known about
the variance of soil physical and chemical parameters
at scales of less than 1–10 m comes from studies of ag-
ricultural systems (e.g. White et al., 1987), ecosystems
dominated by plants forming arbuscular mycorrhizas
(e.g. Dushyantha et al., 1997) or from studies ex-
amining the spatial distribution of soil contaminants
(e.g. Bringmark and Bringmark, 1998; Olivier and
Badr, 1995). A number of studies (e.g. Grundmann
and Debouzie, 2000) have examined the distribution of
bacteria populations at the millimetre scale and found
spatial dependence ranging from 2 to 4 mm for dif-
ferent bacterial groups. Although the mycelial nature
of ECM fungi means that they can integrate over a
greater volume of soil than bacteria, it is at this level
of spatial heterogeneity that studies need to be carried
out. Jackson and Caldwell (1993) working in a nat-
ive sage brush steppe site in Utah demonstrated that
soil organic matter, pH, phosphate, potassium, am-
monium, and nitrate all showed strong spatial patterns
at scales of less than 1 m. At a slightly larger scale (2
m), similar variation was found by Lechowicz and Bell
(1991) in a Quercus / Acer forest in Quebec. Many of
these soil parameters have been linked to the distri-
bution of ECM fungi, at least when assessed by the
occurrence of sporocarps (Hansen, 1988; Nantel and
Neumann, 1992). It seems likely that the distribution
of mycelia and the mycorrhizal tips of ECM species
will also be influenced by the same factors. There is
a great need for studies that utilise geostatistical tech-
niques for identifying both the structure of variability
within individual edaphic dimensions and for quanti-
fying the degree of autocorrelation between samples.
Application of these techniques to the analysis of the
microdistribution of individual fungal species in rela-
tion to soil heterogeneity and to examine the structure
of communities sampled in areas of increasing sizes
would provide new insights into how individuals are
spatially arranged within highly complex and diverse
communities.
A fundamental concept within plant and animal
ecology is that of the minimum sampling area for re-
cording a certain proportion of individuals within a
community. No estimates of minimum sampling areas
exist for sampling ECM fungal communities on root
tips. Coupled to the concept of minimum sampling
area is how samples should be distributed spatially
in order to obtain a realistic description of the com-
munity occurring on the root tips. Determining the
autocorrelation between samples will indicate the ne-
cessary distance required between samples when the
27
species abundance within samples is no longer cor-
related. To obtain a random sample of a community,
samples would therefore have to be taken at distances
greater than this. Conversely, the same analysis will
provide an indication of the predictive power of a
sample i.e. how similar will the community be in
another sample taken at a set distance from the first
sample. Within communities with highly skewed spe-
cies abundance distributions (i.e. high dominance by
one or a few species) the predictive power of a sample
of a given size (at least for these species) should be
greater than in communties where species are more
evenly distributed.
Conclusions
It is clear from the above discussion that sampling ef-
fort and sampling strategy can have a major influence
upon how we perceive ECM community structure. In
addition, our ability to detect the effects of perturba-
tions upon changes in species richness may be limited
by the inherent structure of ECM communities. Both
these factors, therefore, need to be considered when
designing sampling strategies during investigations of
ECM diversity. It is also apparent that in order to ap-
preciate the complexity of ECM systems fully, there
is a great need for more information on how species
are distributed on roots and on the scale and rapidity
of temporal shifts in community structure.
Acknowledgements
The author would like to thank the organisers of
the 3rd International Conference on Mycorrhizas
(ICOMIII) for the invitation to prepare this contribu-
tion, the Forestry Research Institute of Sweden for the
use of the Riddarhyttan site and for logistical support
and the Swedish Forestry companies that financed the
project. In addition, thanks are also due to the many
people who provided valuable comments during the
preparation of this contribution, in particular Björn
Lindahl, Roger Finlay, Tom Horton and Martina Peter.
Thanks are also due to the anonymous referees for
constructive comments.
References
Agerer R 1986–1998 Colour Atlas of Ectomycorrhizae.
Schwäbisch-Gmünd, Einhorn-Verlag.
Anderson I C, Chambers S M, Cairney J W G 1998 Use of mo-
lecular methods to estimate the size and distribution of mycelial
individuals of the ectomycorrhizal basidiomycete Pisolithus tinc-
torius. Mycol. Res. 102, 295–300.
Baar J, Ozinga W A and Kuyper Th W 1994 Spatial distribution of
Laccaria bicolor genets reflected by sporocarps after removal of
litter and humus layers in a Pinus sylvestris forest. Mycol. Res.
98(7), 726–728.
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.
Bamforth S S 1995 Interpreting soil ciliate biodiversity. Plant Soil
170, 159–164.
Bebber D, 1999 Spatial autocorrelations. TREE 14(5), 196.
Bringmark E and Bringmark L 1998 Improved soil monitoring by
use of spatial patterns. Ambio 27, 45–52.
Bruns T D 1995 Thoughts on the processes that maintain local
species diversity of ectomycorrhizal fungi. Plant Soil 170, 63–73.
Copley J 2000 Ecology goes underground. Nature 406, 452–454.
Dahlberg A 2001 Community ecology of ectomycorrhizal fungi: an
advancing interdisciplinary field. New Phytol. 150, 555–562.
Dahlberg A and Stenlid J 1990 Size, distribution and biomass of
genets in populations of Suillus bovinus (L.:Fr.) Ruossel revealed
by somatic incompatibility. New Phytol. 115, 487–493.
Dahlberg A, Jonsson L and Nylund J-E 1997 Species diversity and
distribution of biomass above- and below-ground among ecto-
mycorrhizal fungi in an old-growth Norway spruce forest in
south Sweden. Can. J. Bot. 75, 1323–1335.
Danielson R M and Visser S 1989 Effects of forest soil acidific-
ation on ectomycorrhizal and vesicular-arbuscular mycorrhizal
development. New Phytol. 112: 41–48
Dushyantha K, Hutchings W and Hutchings M J 1997 The effects
of spatial scale of environmental heterogeneity on the growth of
a clonal plant: an experimental study with Glechoma hederacea.
J. Ecol. 85, 17–28.
Erland S and Taylor A F S 2002 Diversity of ectomycorrhizal com-
munities in relation to the abiotic environment. In Mycorrhizal
Ecology. Eds. Marcel G A van der Heijden & Ian R Sanders. pp
470. Ecological Studies Series. Vol 157. ch 7. Springer-Verlag.
Berlin.
Fransson P M A, Taylor A F S, Finlay R D 2000 Effects of op-
timal fertilization on belowground ectomycorrhizal community
structure in a Norway spruce forest. Tree Phys. 20, 599–606
Gardes M and Bruns T D 1996 Community structure of ectomycor-
rhizal fungi in a Pinus muricata forest: above- and below-ground
views. Can. J. Bot. 74, 1572–1583
Gherbi H, Delaruelle C, Selosse M –A and Martin F 1999 High
genetic diversity in a population of the ectomycorrhizal basi-
diomycete Laccaria amethystina in a 150-year-old beech forest.
Mol. Ecol. 8, 2003–2013.
Giller P S 1996 The diversity of soil communities, the ‘poor man’s
tropical rainforest’. Biol. Con. 5, 135–168.
Grundmann G L and Debouzie D 2000 Geostatistical analysis of
the distribution of NH4+and NO2−-oxidizing bacteria and
serotypes at the millimetre scale along a soil transect. FEMS
Microbiol. Ecol. 34, 57–62
Hansen P A 1988 Prediction of macrofungal occurrence in Swedish
beech forest from soil and litter variable models. Vegetatio 78,
31–44.
Harvey A E, Larsen M J and Jurgensen M F 1976 Distribution of ec-
tomycorrhizae in a mature Douglas fir larch forest soil in Western
Montana. For. Sci. 22(4), 393–398.
28
Harvey A E, Larsen M J and Jurgensen M F 1978 Comparative dis-
tribution of ectomycorrhizae in soils of three Western Montana
forest habitat types. For. Sci. 25(2), 350–358.
Heath D 1995 An Introduction to Experimental Design and Stat-
istics for Biology. UCL Press, University College London, UK.
372.
Horton T R and Bruns T D 2001 The molecular revolution in ecto-
mycorrhizal ecology: peeking into the black box. Mol. Ecol. 10,
1855–1871.
Jackson R B and Caldwell M M 1993 Geostatistical patterns of
soil heterogeneity around individual perennial plants. J. Ecol. 81,
683–692.
Krebs C J 1989 Ecological methodology, 2nd ed. Harper & Row,
New York. Kreutzer K 1995 Effects of forest liming on soil
processes. Plant Soil 168-169, 447–470.
Kropp B R 1982 Formation of mycorrhizae on nonmycorrhizal
Western Hemlock outplanted on rotten wood and mineral soil.
For. Sci. 28(4), 706–710.
Kårén O, Hogberg N, Dahlberg A, Jonsson L and Nylund J -
E 1997 Inter- and intraspecific variation in the ITS region of
rDNA of ectomycorrhizal fungi in Fennoscandia as detected by
endonuclease analysis. New Phytol. 136, 313–325.
Lechowicz M J and Bell G 1991 The ecology and genetics of fitness
in forest plants. II. Microspatial heterogeneity of the edaphic
environment. J. Ecol. 79, 687–696.
Magurran A E 1988 Ecological Diversity and its Measurement.
Croom Helm, London. Nantel P and Neumann P 1992 Eco-
logy of ectomycorrhizal-basidiomycete communities on a local
vegetation gradient. Ecology 73(1), 99–117.
Oliver M A and Badr I 1995 Determining the spatial scale of
variation in soil radon concentration. Math. Geol. 27, 893–922.
Peter M, Ayer F, Egli S and Honegger R 2001a Above – and below-
ground community structure of ectomycorrhizal fungi in three
Norway spruce (Picea abies) stands in Switzerland. Can. J. Bot.
79, 1134–1151.
Peter M, Ayer F and Egli S 2001b Nitrogen addition in Norway
spruce stand altered macromycete sporocarp production and
below-ground ectomycorrhizal species composition measured by
PCR-RFLP analysis of the ribosomal ITS-region. New Phytol.
149, 311–326.
Read D J 1998 Plants on the web. Nature 396, 69–72.
Redecker D, Szaro T M, Bowman R J and Bruns T D 2001 Small
genets of Lactarius xanthogalactus,Russula cremoricolor and
Amanita francheti in late stage ectomycorrhizal successions.
Mol. Ecol.10, 1025–1034.
Sawyer N A, Chambers S M and Cairney J W G 1999 Molecular
investigation of genet distribution and genetic variation of Cor-
tinarius rotundisporus in eastern Australian sclerophyll forests.
New Phytol. 142, 561–568.
Sokal R R and Rohlf F J 1996 Introduction to Biostatistics. W.H.
Freeman and Company, New York. 363 p.
Stendell E R, Horton T R and Bruns T D 1999 Early effects of
prescribed fire on the structure of the ectomycorrhizal fungus
community in a Sierra Nevada ponderosa pine forest. Mycol.
Res. 103, 1353–1359.
Taylor A F S, Martin F and Read D J 2000 Fungal diversity in ec-
tomycorrhizal communities of Norway spruce (Picea abies [L.]
Karst.) and Beech (Fagus sylvatica L.) along north-south tran-
sects in Europe. In Ed. E-D. Schulze. Ecol. Stud. Vol. 142. pp
343–365. Springer-Verlag, Heidelberg.
Taylor D L and Bruns T D 1999 Community structure of ecto-
mycorrhizal fungi in a Pinus muricata forest: minimal overlap
between the mature forest and resistant propagule communities.
Mol. Ecol. 8,1837–1850
van der Heijden M G A, Klironomos J N, Ursic M, Moutoglis P,
Streitwolf-Engel R, Boller T, Wiemken A and Sander I R1998
Mycorrhizal fungal diversity determines plant biodiversity, eco-
system variability and productivity. Nature 396, 69–72.
Wallenda T and Kottke I 1998 Nitrogen deposition and ectomycor-
rhizas. New Phytol. 139, 169–187.
White R E, Haigh R A and MacDuff J H 1987 Frequency distribu-
tions and spatially dependent variability of ammonium and ni-
trate concentrations in soil under grazed and ungrazed grassland.
Fertil. Res. 11, 193–208.
