Type of disturbance and ecological history determine structural stability

Abstract

This study aims to reveal whether complexity, namely, community and trophic structure, of chronically stressed soil systems is at increased risk or remains stable when confronted with a subsequent disturbance. Therefore, we focused on a grassland with a history of four centuries of patchy contamination. Nematodes were used as model organisms because they are an abundant and trophically diverse group and representative of the soil food web and ecosystem complexity. In a field survey, a relationship between contaminants and community structures was established. Following, two groups of soil mesocosms from the field that differed in contamination level were exposed to different disturbance regimes, namely, to the contaminant zinc and a heat shock. The zinc treatment revealed that community structure is stable, irrespective of soil contamination levels. This implies that centuries of exposure to contamination led to adaptation of the soil nematode community irrespective of the patchy distribution of contaminants. In contrast, the heat shock had adverse effects on species richness in the highly contaminated soils only. The total nematode biomass was lower in the highly contaminated field samples; however, the biomass was not affected by zinc and heat treatments of the mesocosms. This means that density compensation occurred rapidly, i.e., tolerant species quickly replaced sensitive species. Our results support the hypothesis that the history of contamination and the type of disturbance determine the response of communities. Despite that ecosystems may be exposed for centuries to contamination and communities show adaptation, biodiversity in highly contaminated sites is at increased risk when exposed to a different disturbance regime. We discuss how the loss of higher trophic levels from the entire system, such as represented by carnivorous nematodes after the heat shock, accompanied by local biodiversity loss at highly contaminated sites, may result in detrimental effects on ecosystem functions.

Full text

Ecological Applications, 17(1), 2007, pp. 190–202
Ó2007 by the Ecological Society of America
TYPE OF DISTURBANCE AND ECOLOGICAL HISTORY
DETERMINE STRUCTURAL STABILITY
A. W. G. VAN DER WURFF,
1,6
S. A. E. KOOLS,
2
M. E. Y. BOIVIN,
2,3
P. J. VAN DEN BRINK,
4,5
H. H. M. VAN MEGEN,
1
J. A. G. RIKSEN,
1
A. DOROSZUK,
1
AND J. E. KAMMENGA
1
1
Wageningen University, Laboratory of Nematology, Wageningen, The Netherlands
2
Vrije Universiteit, Institute of Ecological Science, Department of Animal Ecology, Amsterdam, The Netherlands
3
National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands
4
ALTERRA, Wageningen University and Research Center, Wageningen, The Netherlands
5
Wageningen University, Department of Aquatic Ecology and Water Quality Management, Wageningen, The Netherlands
Abstract. This study aims to reveal whether complexity, namely, community and trophic
structure, of chronically stressed soil systems is at increased risk or remains stable when
confronted with a subsequent disturbance. Therefore, we focused on a grassland with a history
of four centuries of patchy contamination. Nematodes were used as model organisms because
they are an abundant and trophically diverse group and representative of the soil food web
and ecosystem complexity. In a field survey, a relationship between contaminants and
community structures was established. Following, two groups of soil mesocosms from the field
that differed in contamination level were exposed to different disturbance regimes, namely, to
the contaminant zinc and a heat shock. The zinc treatment revealed that community structure
is stable, irrespective of soil contamination levels. This implies that centuries of exposure to
contamination led to adaptation of the soil nematode community irrespective of the patchy
distribution of contaminants. In contrast, the heat shock had adverse effects on species
richness in the highly contaminated soils only. The total nematode biomass was lower in the
highly contaminated field samples; however, the biomass was not affected by zinc and heat
treatments of the mesocosms. This means that density compensation occurred rapidly, i.e.,
tolerant species quickly replaced sensitive species. Our results support the hypothesis that the
history of contamination and the type of disturbance determine the response of communities.
Despite that ecosystems may be exposed for centuries to contamination and communities
show adaptation, biodiversity in highly contaminated sites is at increased risk when exposed to
a different disturbance regime. We discuss how the loss of higher trophic levels from the entire
system, such as represented by carnivorous nematodes after the heat shock, accompanied by
local biodiversity loss at highly contaminated sites, may result in detrimental effects on
ecosystem functions.
Key words: contamination; density compensation; heat shock; mesocosm; nematodes; species richness;
stability.
INTRODUCTION
Humans have extensively altered the global habitat in
many ways (Chapin et al. 2000), including by contam-
ination of the soil environment (Helgen and Moore
1996, White and Claxton 2004). The long-term response
of soil ecosystems to contamination in terms of function
and complexity is uncertain. Some studies suggested that
only when a toxicity threshold is exceeded, soil
ecosystem functions are severely impaired, while below
toxicity thresholds the composition of the soil commu-
nity may change without significant effects on function
owing to functional redundancy (Pennanen et al. 1996,
Giller et al. 1999, Schwartz et al. 2000). In contrast, a
recent study (Ramsey et al. 2005) on a 93-year-old mine
site showed that function was negatively correlated with
an increase in contamination levels while soil microbial
community structure was not significantly affected. This
discrepancy may stem from the fact that previous studies
looked at ecosystem response on the scale of days to few
decades, while ecologically relevant response times may
be longer. In addition, the relation between soil
community structure and ecosystem functioning may
vary among systems depending on, e.g., dispersal
abilities (Hedlund et al. 2004), the relation between
response and functional traits of focal species (Ek-
schmitt et al. 2001), and disturbance regime.
Here, we investigate in addition an ecosystem with a
long pollution history. Our study aims to reveal whether
complexity, namely, community and trophic structure,
of chronically stressed soil systems is at increased risk or
remains stable when confronted with a subsequent
Manuscript received 20 December 2005; revised 17 May
2006; accepted 25 May 2006. Corresponding Editor: J.
Gulledge.
6
Present address: Business Unit Greenhouse Horticulture,
Applied Plant Research (PPO), Wageningen University and
Research Centre, Kruisbroekweg 5, 2671 KT Naaldwijk, The
Netherlands. E-mail: Andre.vanderWurff@wur.nl
190

disturbance. Stability is defined here as the ability to
withstand or to recover from a disturbance (Griffiths et
al. 2000). We address the importance of ecological
history and disturbance regime on biodiversity and
structural stability of the soil community. Biodiversity is
important with regard to ecosystem functioning, and it
is widely accepted that biodiversity is needed to insure a
stable supply of ecosystem goods and processes (Hooper
et al. 2005).
As initially proposed by Sankaran and McNaughton
(1999), there is increasing support that ecological history
and disturbance regime are principal determinants of
structural stability of communities. For instance, Fi-
scher et al. (2001) showed that communities previously
exposed to high pH regimes, showed an increased
tolerance toward acidification. Community tolerance
may result from the elimination of sensitive species and
successive replacement by insensitive species (Odum
1985). In addition, populations may adapt to specific
disturbances (Blanck and Wangberg 1988, Millward and
Grant 2000) or become tolerant to even a wide range of
disturbances (Molander et al. 1992). However, the
response of chronically stressed communities to a
different disturbance regime is less studied. Studies of
the effects of subsequent stress are crucial since large-
scale disturbances are expected with global change, such
as local heat waves and drought.
The focus of our study is a grassland ecosystem with a
history of centuries of contamination with heavy metals,
predominantly zinc, copper, lead, and cadmium (Boivin
2005, Kools et al. 2005). The distribution of the
contaminants shows a large variability on a scale of
meters (Bosveld et al. 2000). In a field survey,
contaminants and abiotic parameters were inventoried
and linked to nematode and bacterial abundance data.
In a successive experiment, intact field mesocosms were
treated with a contaminant or a heat shock to evaluate
tolerance of the soil community to a subsequent stressor.
As model organisms we used the diverse group of the
nematodes (Nematoda) (Trett et al. 2000): they are
abundant, trophically diverse, and good estimators of
ecosystem functions, i.e., nutrient cycling and decom-
position (Beare et al. 1992, Ferris et al. 2001).
MATERIALS AND METHODS
Location
The grassland Demmerikse polder (The Netherlands:
528130N, 48560E; see Plate 1) is a 280-ha heavily
polluted soil ecosystem. Since the 16th century, heavily
polluted city waste, sludge, farmyard manure, and dune
sand were applied to fertilize, equalize, and pave the soil
top layer for agriculture. This resulted in a patchy
distributed mixture of predominantly zinc, copper, lead,
and cadmium (Bosveld et al. 2000). Nowadays, the top
layer of approximately 40 cm contains clay and sand on
top of peat, dominated by rye grass (Lolium perenne)
and ribwort plantain (Plantago lanceolata).
Nematode extraction and identification
In order to extract nematodes from soil, 100 g of the
10 cm of topsoil was used (Oostenbrink 1960).
Suspension was put on double cotton wool filter (Hygia
milac filter; Hartmann BV, Nijmegen, The Netherlands)
and incubated 48 h. Numbers were estimated as an
average over two to four counts in 1:10 volume
suspension. Nematodes were heat-killed and fixed in
4%formalin. Per sample, about 150 species were
identified under a magnification of 400 –10003to genus
following Bongers (1988). Trophic groups were accord-
ing to Yeates et al. (1993), were hyphal feeders (H),
carnivores (C), bacterivores (B), omnivores (O), and
plant parasites (P).
Abiotic analyses and CFU
Water-holding capacity (WHC), field humidity, ex-
changeable and total fractions of heavy metals, total
organic carbon, lutum, and bacterial colony-forming
units (CFU) were analyzed according to Boivin (2005).
Total metal concentrations were obtained by digesting 1
g dried soil with 2 mL demineralized water, 6 mL HCl
(37%pro analysis grade [p.a.]; Baker, Philipsburg, New
Jersey, USA) and 2 mL HNO
3
(65%; p.a., Riedel-
deHae
¨n, Seelze, Germany) in a microwave (model MDS
81 D; CEM Microwave Technology Ltd., Buckingham,
UK) and determined with an ICP-MS (HP-4500,
Hewlett Packard, Palo Alto, California, USA) using
standard procedures (SOP LAC-M149; National Insti-
tute for Public Health and the Environment, Bilthoven,
The Netherlands). Exchangeable metal concentration
was based on WHC50 and 1:10 DW/CaCl
2
.This
exchangeable fraction was interpreted as ‘‘bioavailable,’’
as it mimics the effect of the plant root exudates to
dissolve matter for uptake (Houba et al. 1996). Organic
matter (OM) was calculated as loss of ignition (LOI) at
4508C and inorganic carbon content by the mass
difference of the samples at 9008C minus 4508C (see
Appendix A, Table A1 for an overview of field
parameters). The numbers of CFU were determined,
by using tryptone soya broth (TSB) agar plates in Petri
dishes (0.3 g/L). Four bacterial dilutions were inoculated
on Petri dishes: 1:3 310
12
, 1:3 310
10
, 1:3 310
8
, and
1:3 310
6
. The plates were incubated in the dark at 258C
for eight days.
Nematode biomass and indices
Fresh mass (FM) body mass of nematodes was
estimated with a modified Andrassy’s formula (Freck-
man 1982) as FM ¼(w
2
3L)/(1.6 310
6
), where FM is
fresh weight (lg), wis width (lm), and Lis total length
of a nematode (lm). Tail-corrected length was applied
as the average length of the most likely occurring species
in grassland in the Netherlands, corrected by the tail
shape as the deviation from a regular cigar shape to the
tail end in millimeters. All morphometric parameters
were from Bongers (1988).
January 2007 191DISTURBANCE, HISTORY DETERMINE STABILITY

The general type of response of nematodes to
disturbances, summarized into so-called cp values, is
indicated in addition to trophic level and nomenclature
throughout the text. The cp values range from 1 to 5, and
reflect a position on the rand Kcontinuum, such as short
generation time and high numbers of offspring for cp1
genera and vice versa for cp5 genera (Bongers 1999). The
B1 class represents a group of bacterivorous enrichment
opportunists, i.e., Rhabditidae and Diplogastridae.
These bacterivores typically have a short generation
time, high numbers of offspring, and are considered as
insensitive to disturbance. They are generally used as
indicators of enrichment (Bongers 1999).
Species richness (D0) was calculated according to
Simpson (1949) as D0¼1/(Rp2
i), where p
i
is the
proportion of category iin the community. The relative
abundance of the enrichment opportunists (EI) were
calculated in the field survey according to Ferris et al.
(2001). The EI represents the weighted ratios of
enrichment opportunists B1 and H2 to the enrichment
opportunists (B1) plus the nematodes that are consid-
ered as early succession nematodes. The NCR (nema-
tode channel ratio; Yeates 2003) represents the ratio of
the fractions of mycovorous nematodes (H2) to the
enrichment opportunists (B1) plus mycovorous nema-
todes (H2), and is a good indicator of the decomposition
pathway (Mulder et al. 2005). Index values were ln-
transformed to meet assumptions of normality. One-way
ANOVA (SPSS, Chicago, Illinois, USA) with Levene’s
test for homogeneity of variances across samples was
used to analyze both index values.
Field survey
Sampling.—The field area was approximately 200 3
30 m and surrounded by ditches. In order to get an
overview of contamination levels, the top 10 cm of the
soil was sampled 120 times along a grid with 5 m
between samples. Samples were dried at 808C. The next
day, concentrations of lead, copper, cadmium, and zinc
were screened with a hand-held XRF-apparatus (XTAC
Analytical, Leiden, The Netherlands). Following, the
upper 10 cm of soil from 10 of the least contaminated
sites and eight of the most contaminated sites were
sampled along a rectangular grid in October 2001.
Nematodes were extracted the following day after
sampling field as well as mesocosms.
Data analyses.—In order to get an overview of
dominant ordinal gradients, a matrix of 18 samples
and 64 species with 487 occurrences was evaluated using
ordination methods as incorporated in CANOCO for
Windows 4.5 (ter Braak and Smilauer 2002). Nematode
abundance data were ln(2xþ1)-transformed and metal
concentrations were ln-transformed to down-weight
high values and approximate a normal distribution.
To account for spatial dependency of the different
samples, spatial coordinates of samples were trans-
formed into principal coordinates of neighbor matrices
(PCNM; Borcard and Legendre 2002) with the program
SpaceMaker 2 (Borcard and Legendre 2002), calculated
as un-truncated Euclidean distances among sites, and
used as covariates to ‘‘partial out’’ spatial patterns.
Detrended correspondence analyses (DCA) by seg-
ments revealed a gradient of 1.2 standard deviation units,
indicating a strong linear response of taxa. Therefore,
principal component analyses (PCA) was used. Data
were of a compositional nature; an amount of approx-
imately 150 nematodes were counted in all samples.
Redundancy analysis (RDA) followed by unrestricted
Monte Carlo permutation (MC) tests with 999 permu-
tations was performed to assess the significance of the
relation between each environmental variable and
community composition (Verdonschot and Braak
1994). RDA is the constrained version of PCA, meaning
that it focuses on the part of the variance explained by
the environmental variables only. For environmental
variables see the chapter on abiotic analyses and CFU in
the materials and methods section.
PCA analysis of the environmental variables indicated
that many variables were colinear. To remove this
colinearity, the procedure was repeated using the
forward selection procedure. Within forward selection,
the variable explaining the largest amount of the total
variance is included into the model. Following, the
variable explaining the largest part of the remaining
variance is included and so on. Thus, variables that are
colinear with variables already included become unim-
portant. MC permutation tests determined the signifi-
cance of inclusion of an extra variable in explaining the
differences in relative species composition between
samples (ter Braak and Smilauer 2002). After this a
RDA was performed using only variables that explained
a significant part of the variation in relative abundance
of the species.
Mesocosm study
Sampling.—In August 2003, two adjacent fields were
sampled 180 times to get an overview in terms of
contamination levels by the XRF apparatus. Following,
79 mesocosms were sampled by means of high-density
polyethylene columns (length 40 cm, diameter 17.5 cm;
Koolhaas et al. 2004) from two adjacent fields of
approximately 20 339 m and 35 339 m, each bordered
by ditches approximately 10 m in width. Thirty-nine
mesocosms were sampled from the lowest (L) and 40
from the highest (H) contaminated spots. Mesocosms
were transferred to a climatic chamber (VU, Amsterdam,
The Netherlands) and placed according to a randomized
block design in temperature-controlled carts (see Plate
1). Climate conditions simulated summer conditions: air
temperature (day : night) 208C:168C; light regime 14:10 h,
soil temperature 108C:108C, relative humidity 60–80%.
All mesocosms received 250 mL artificial rainwater twice
a week according to Koolhaas et al. (2004).
After five weeks of incubation, seven mesocosms, i.e.,
three L and four H, were sampled. Following, 12
mesocosms from L contaminated soils and 12 from H
A. W. G. VAN DER WURFF ET AL.192 Ecological Applications
Vol. 17, No. 1

contaminated soils were treated with 4200 mg/L zinc
(pure ZnCl
2
; Merck, Darmstadt, Germany) applied at
two successive days. Heat stress was applied to another
12 mesocosms from L and 12 from H contaminated soils
by incubation in a room at 408C for 24 h (see Plate 1).
High soil water loss was prevented by keeping air
moisture at .98%. Twelve L as well as 12 H mesocosms
were used as a reference. At 1, 3, 8, and 16 weeks after
stress application, mesocosms were randomly sampled.
The time series contained three L and one reference;
three H and one reference; three L and one zinc; three H
and one zinc; three L and one heat; and three H and
heat.
Data analyses.—In order to investigate the change in
genera composition related to zinc or heat treatment,
principal response curve (PRC; Van den Brink and ter
Braak 1999) analysis was performed. In a PRC diagram,
time is shown on the x-axis, while the first principal
component of treatment effects is shown on the y-axis
(see Fig. 2A as an example). The deviations in species
composition of all treatments to a reference (in this case
the low, reference series) are shown for time series at
week 0 and week 1, 3, 8, and 16. A species weight
diagram shows the affinity of all species with the
response indicated in the diagram. Multiplying the
weight (b
k
) of species kby the canonical coefficient C
dt
of a treatment at a specific time point yields the fitted
change on a ln scale of this species compared to the
reference (low, reference series). The exponent of this
quotient yields the relative abundance compared to the
reference. Significance of the PRC was tested by MC, by
permuting time series in partial (p)RDA, using an F-type
test statistic based on the eigenvalue of the component.
Significance of the background contamination (L vs. H)
as well as zinc and heat was tested individually as well as
their interaction by MC as described by Van Wijngaar-
den et al. (2005).
The restricted maximum likelihood (REML) repeated
measurement analyses (Corbeil and Searle 1976; linear
mixed models, SPSS) based on 1 310
6
iterations was
used to analyze a time series of nematode species
richness (D0) and biomass. The factors time with levels
1, 3, 8, and 16 weeks posttreatment; contamination
background with levels H and L; and treatment with
levels reference, zinc, and heat, were used as fixed effects
in a full-factorial design. REML repeated-measurement
analysis was preferred since mesocosms are sampled
with unequal time intervals and the approach allows an
evaluation of time-dependent effects. The fit of the
REML covariance models did not differ significantly for
any of the dependent variables according to the LRT
based on 2 restricted log likelihoods (2RLL
ML
)
values. Therefore the simplest model (scaled identity)
was chosen. This model treats the variances as constant
and covariance’s not correlated among trials (in week 1,
3, 8, and 16).
Mesocosms were sampled only once. Therefore, in
order to use REML, different mesocosms were desig-
nated to one subject by means of pPCA analyses on total
lead content with time and treatment as covariates.
Total lead content was used instead of zinc, since
additional zinc was applied. Furthermore, according to
partition coefficients, the addition of zinc does not affect
lead concentration levels (Jonker et al. 2004).
RESULTS
Field survey
The pPCA on the environmental variables justified a
classification according to ‘‘high-contaminated’’ and
relatively ‘‘low-contaminated’’ samples (Fig. 1A), fur-
ther referred to as H and L, respectively. The H and L
assignment was validated by the first ordination axis,
explaining 83.6%of the total variation in levels of
environmental variables. The pRDA, using the nema-
tode data set and OM and zinc as environmental
variables, confirmed the H and L classification. One
sample, however, was separated from the others, owing
to relatively low levels of exchangeable cadmium and a
high WHC.
The mosaic-like distribution of the contaminants
showed a large variability on a scale of meters.
Nevertheless, two ordinal environmental gradients were
revealed by the forward selection procedure related to
differences in the nematode genera composition, namely
total zinc content and OM (conditional effects, P,
0.05). Both were negatively correlated (Fig. 1B). In total,
13 soil parameters were significant when tested alone
(marginal effects, P,0.05). This discrepancy between
marginal and conditional effects indicated that many
environmental variables were confounding. Partial
correlation analyses with PCNM as covariates con-
firmed that total and exchangeable metal contents were
correlated (partial correlation: 0.51 r0.85, P,
0.05) and inversely related to OM, pH, and humidity
(partial correlation: 0.69 r0.77, P,0.05).
The differences in nematode communities were related
to a higher abundance of genera with a high cp value in
the L samples such as the plant parasitic Trichodorus
(P4) and Tylenchorhynchus (P3), the bacterivorous
Alaimus (B4), the carnivorous Mononchus (C4), and
the omnivorous Oxidirus (O5).
The highly contaminated field samples contained a
higher number of genera belonging to the family of the
Rhabditidae and Diplogastridae (B1) as suggested by
the EI (ANOVA: EI
L
¼56.35, EI
H
¼72.33; F
1,16
¼9.89,
P,0.01). In addition, the NCR index, i.e., the ratio of
the fractions (H2) mycovorous nematodes to the
bacterivorous (B1) plus mycovorous (H2) nematodes,
is higher in the higher polluted sites (ANOVA: NCR
L
¼
0.47, NCR
H
¼0.28, F
1,16
¼4.79, P,0.05).
Mesocosm study
Zinc treatment.—Treatments with zinc resulted in a
zinc load of approximately 750 mg/kg soil dry mass in
the mesocosms to the end of the experiment. Exchange-
able fractions were much lower, specifically the high,
January 2007 193DISTURBANCE, HISTORY DETERMINE STABILITY

zinc series contained 50 mg/kg soil dry mass; while the
low, zinc series contained 15 mg/kg soil dry mass.
PRC analyses (Fig. 2A) showed neither effects of zinc
on the communities nor effects of the zinc treatment
related to the contamination background (H or L; MC: P
0.05). The omnivorous species Aporcelaimellus (O5)
was approximately 1.6 times more abundant at week 16 in
the H samples when compared to the low, no zinc samples.
The plant parasitic Paratylenchus (P2) flourished in the H
mesocosms regardless of zinc treatment as numbers were
10-fold higher than in the low, no zinc series (Fig. 2A).
Although the zinc treatment triggered an inverse
response of trophic groups between H and L meso-
cosms, i.e., predominantly between omnivorous (O4)
and bacterivorous (B1) nematodes, both recovered after
16 weeks (Fig. 3A).
Similar to the PRC results, REML analysis showed
that both L and H samples reacted not significantly to
FIG. 1. Ordination diagrams representing low- and high-contaminated samples (n¼18 samples) of the grassland ecosystem.
Relatively low-contaminated samples are indicated by open squares, high-contaminated by filled squares. Only variables with a
species fit range between 25%and 100%are shown for clarity. Exchangeable and total metal contents of zinc, copper, and lead were
strongly correlated, while OM was negatively correlated with metals and positively correlated with pH and WHC. (A) Partial
(p)PCA biplot with a focus on samples showing the absolute differences in levels of different environmental variables. Of the overall
variation in absolute levels of the different variables, 83.6%is displayed on the first axis and another 15.6%on the second. Metals
are represented as exchangeable and total concentration; WHC is water holding capacity; humidity is expressed as percentage of
soil dry mass; OM is organic matter. (B) Partial (p)RDA (redundancy analysis) triplot of the relative differences in nematode
abundance values explained by the environmental variables that were significant in the forward selection procedure, namely, total
zinc concentration and organic matter. Of the explained variance, 56.9%is displayed on the first axis, and another 43.1%on the
second.
A. W. G. VAN DER WURFF ET AL.194 Ecological Applications
Vol. 17, No. 1

FIG. 2. Principal Response Curves (PRC) diagram showing the response over time of the nematode communities from either
low- or high-contaminated soils to zinc and heat treatment during 16 weeks. The weight (b
k
) indicates the relative contribution of
genera to the community response curve. Taxa with a weight between 0.5 and 0.5 are not shown for clarity. Trophic groups are
indicated as H, hyphal feeders; B, bacterivores; and O, omnivores. Values that accompany the trophic groups represent colonizer–
persister (cp) values. (A) Analyses of the response to the zinc treatment. Of all variance, 10%could be attributed to the sampling
date; this is displayed on the horizontal axis. Twenty-seven percent of all variance could be attributed to treatment level. Of this
variance, 38%is displayed on the vertical axis. (B) Analyses of the response to the heat treatment. Of all variance, 10%could be
attributed to the sampling date; this is displayed on the horizontal axis. Thirty-four percent of all variance could be attributed to
treatment level. Of this variance, 38%is displayed on the vertical axis.
January 2007 195DISTURBANCE, HISTORY DETERMINE STABILITY

the zinc treatment in terms of species richness (D0),
although H samples showed a temporary decrease at
eight weeks posttreatment (Fig. 4A). In addition, no
effect of zinc was observed on total estimated biomass
(REML: F
2,47
¼2.241, P0.05; Fig. 4B), nor on
biomass of carnivorous nematodes (Fig. 5).
PRC confirmed the results of the field survey since it
showed a difference in composition of the nematode
community between the H and L reference samples
(MC: P,0.01). Specifically, the L samples showed a
higher number of plant parasitic genera (P2 and P3) and
mycovorous genera (H2) (Fig. 2A). In addition, species
FIG. 3. Principal Response Curves (PRC) diagram showing the response over time of trophic groups from either low- or high-
contaminated soils to zinc and heat treatment during 16 weeks. The weight (b
k
) indicates the relative contribution of trophic groups
to the community response curve. Taxa with a weight between 0.5 and 0.5 are not shown for clarity. Trophic groups are: H,
hyphal feeders; B, bacterivores; and O, omnivores. Accompanying values represent colonizer–persister (cp) values. (A) Analyses of
the response to the zinc treatment. Of all variance, 17%could be attributed to the sampling date; this is displayed on the horizontal
axis. Thirty-three percent of all variance could be attributed to treatment level. Of this variance, 53%is displayed on the vertical
axis. (B) Analyses of the response to the heat treatment. Of all variance, 9%could be attributed to the sampling date; this is
displayed on the horizontal axis. Thirty-four percent of all variance could be attributed to treatment level. Of this variance, 51%is
displayed on the vertical axis.
A. W. G. VAN DER WURFF ET AL.196 Ecological Applications
Vol. 17, No. 1

richness (REML: F
1,47
¼8.24 , P,0.01; Fig. 4A) and
total estimated biomass (REML: F
1,47
¼41.158, P,
0.001; Fig. 4B) differed between H and L reference
samples. Furthermore, the analysis of bacterial numbers
in H and L, in terms of CFU/g soil dry mass, showed
higher numbers in the L sites (REML: F
1,47
¼40.55, P,
0.001).
Heat treatment.—PRC analysis of response to the
heat shock showed, in contrast to the zinc treatment,
that the heat treatment significantly affected the
community structure (MC: P0.001, Fig. 2B).
However, the difference in background contamination
level did not influence the response of the communities
to the treatment (MC: P.0.05). The principal response
was dominated by Rhabditis sp. (B1) (Rhabditidae). Its
abundance was roughly 20 times lower in the heat
treated samples at week 16 when compared to the
reference series at week 16 (Fig. 2B).
The principal response of trophic groups to the heat
shock (Fig. 3B) showed no effects of the H communities.
FIG. 4. REML (restricted maximum likelihood) analyses on nematode community parameters in response to background
contamination level (low vs. high) and treatments during 16 weeks. Individual time points are represented as means calculated
under the Scaled Identity model of covariance structure (variance ¼constant; covariance ¼unrelated). (A) Change in species
richness (D0¼Simpson’s diversity) in reference and treated mesocosms. (B) Change in estimated total nematode biomass in
reference and treated mesocosms.
January 2007 197DISTURBANCE, HISTORY DETERMINE STABILITY

In contrast, the heat shock resulted within the L
communities in a decrease in omnivorous (O4) nema-
todes, bacterivores (B1), and omnivores (O5).
As with the PRC analyses on trophic groups, the
REML analyses of the heat treatment showed an
interaction with contamination background, i.e., species
richness from H samples decreased significantly after
heat treatment in week 8 and 16 according to estimates
of fixed effects (REML: D0
wk8, heat treatment 3H, t
47
¼
2.543, P,0.05; D0
wk16, heat treatment 3H: t
47
¼3.77,
P,0.001; Fig. 4A). These results were supported by
nematode specific indices, i.e., maturity index (Bongers
and Ferris 1999) and structure index (Ferris et al. 2001;
see Appendix A, Fig. 1, and Fig. 2, respectively). The
apparent increase in species richness in the L reference
samples was not significant (REML: D0
wk8, reference 3L,
t
47
¼1.221, P0.05; D0
wk16, reference 3L, t
47
¼0.323, P
0.05; Fig. 4A).
Similar to the zinc treatment, no significant effect of
the heat shock treatment on total biomass with respect
to contamination background was observed (REML:
F
2,47
¼2.241, P0.05; Fig. 4B).
The highest trophic group in our analyses, exemplified
by the nematode-eating carnivorous nematodes such as
Trypila (C3), and several monochids, i.e., Clarkus (C4),
Mylonchulus (C4), and Mononchus (C4), were eliminated
when exposed to the heat shock regardless of contam-
ination background (Fig. 5). The carnivorous nema-
todes were present in the mesocosms that were sampled
prior to the start of the treatments (L ¼109.80 672.46,
H¼28.27 656.55; indicated as average biomass in lg/
50 g soil dry mass 6SD).
The effect of time on CFU numbers in the REML
analyses was significant (REML: F
3,47
¼13.49, P,
0.001): The initial bacterial biomass based on CFU was
restored with greater speed after the heat shock when
compared to reference and zinc-treated samples.
Evaluation of abiotic parameters.—Since pH and
water content could have influenced nematode commu-
nities directly or indirectly by releasing contaminants
from the soil matrix, additional REML analyses on
these parameters were performed. Water content was
not affected by any treatment (REML: F
2,47
¼2.82 , P
0.05) although it differed between H and L (REML:
F
1,47
¼17.05, P,0.001) and was affected throughout
the mesocosm incubation of 16 weeks (REML: F
3,47
¼
17.05, P,0.001): it slightly increased toward week 16,
namely from 30–40%of WHC100 in week 1 to 40–50%
in week 16 post treatment. The pH was not affected by
any treatment (REML: F
2,47
¼0.07, P0.05) nor time
(REML: F
3,47
¼1.66, P0.05), although it differed
significantly between H and L samples (REML: F
1,47
¼
54.281, P,0.001).
PRC analyses of the metal data showed that zinc and
heat treatments and incubation time of the mesocosms
did not significantly affect metal concentrations (MC: P
0.05).
DISCUSSION
The biodiversity of the nematode communities from
the contaminated soils remained stable when exposed to
elevated levels of a contaminant. In contrast, the
biodiversity of the communities from the highly
contaminated soils did not recover from a heat shock.
Therefore, our results contribute to the hypothesis that
the type of disturbance in combination with the
ecological history determines the structural stability of
the soil community (Sankaran and McNaughton 1999).
The history of contamination must have initiated an
adaptive response of the community members since
communities reacted minimally to a high amount of
4200 mg/L zinc. This is in sharp contrast to various
other studies where field assays with treatments of 400 to
1600 mg/L zinc result in acute effects on nematode
FIG. 5. Total biomass (mean þSD; n¼3 samples) of carnivorous nematodes per 50 g of soil (dry mass) in reference and treated
mesocosms from low- and high-contaminated sites throughout 16 weeks.
A. W. G. VAN DER WURFF ET AL.198 Ecological Applications
Vol. 17, No. 1

communities not previously exposed to pollutants
(Korthals et al. 1996). In addition, even a concentration
as low as 50 mg/L zinc is reported to have a significant
long-term impact (Korthals et al. 2000, Smit et al. 2002).
Although the mesocosms received an acute shock of
4200 mg/L, a lower zinc load of 1600 mg/L and 700
mg/L was detected in the L and H samples, respec-
tively, after three weeks and onwards. The majority of
the applied zinc must have been flushed out of the
mesocosms. The L samples retained a higher zinc load
probably as a result of higher fractions of OM. We
analyzed this by calculating the ratio total measured
zinc concentration to the exchangeable, i.e., ‘‘bioavail-
able,’’ concentration. This is roughly 750:35 mg/kg
(total : exchangeable) and is indeed strongly correlated
to the fraction OM (correlation: r¼0.758, P,
0.001). As a result, the long-term bioavailable fraction
was higher in the H samples while the L samples
received only a short and intense zinc load. Neverthe-
less, nematode communities from both L and H
contaminated samples were able to deal with a high
and acute zinc stress, which is documented to cause
significant effects on community structure. Therefore,
we conclude that both communities from both H and L
contaminated sites exhibit increased tolerance toward
contaminants.
It is shown here that communities of high-contam-
inated soils, when compared to the communities from
low-contaminated soils, are more vulnerable to the heat
shock in terms of biodiversity. The trophic composition
is, however, not affected when H samples were exposed
to the heat shock (Fig. 3B). The decrease in omnivores
of the L samples was expected based on their high cp
classification, i.e., their K-strategist features. However,
numbers of the opportunistic bacterivores (B1), such as
Rhabditis (Rhabditidae), were reduced remarkably.
These nematodes are regarded as stress tolerant and a
reduction in abundance was not expected. They may
have been eliminated indirectly by a decline in specific
bacteria. Although data on CFU numbers suggests that
the bacterial community was restored within eight
weeks, it may not include the type of bacteria that
these bacterivores use as a food source, i.e., no single
method in microbial ecology can characterize a
microbial community completely (e.g., Garbeva et al.
2004).
Another indirect effect is presented by the plant
ectoparasite Paratylenchus (P2). It flourishes when heat
as well as zinc is applied (see also Georgieva et al. 2002).
Increase of species in abundance with increasing
contamination level is a poorly understood phenomenon
and may result from indirect effects such as stimulated
hatching of nematode eggs by chemicals (Clarke and
Shepherd 1966), competition and top-down, bottom-up
effects by less sensitive taxa (Hendrix and Parmelee
1985).
Indirect effects of contamination were in addition
observed in the field survey. The abundance of
Rhabditidae and Diplogastridae (B1) was negatively
correlated with OM content. This is remarkable since
these nematodes are generally expected to flourish at
spots with high energy levels, such as OM and
accompanying bacteria. This phenomenon could be
explained by the lower predation pressure of carnivores
and/or lower number of fungal antagonists in the H
samples relative to the L samples (Georgieva et al. 2002);
i.e., the L soils contained relatively higher numbers of
mycovorous (H2) species and carnivores, i.e., Tripyla
(C3) and Mononchus (C4). The abundance of another
group of bacterivores in the L samples, i.e., the
cephalobids (B2), may be explained by the lower WHC
in these samples (Griffiths et al. 1995).
PLATE 1. (Left) The grassland ecosystem ‘‘De Ronde Venen’’ where samples were taken. (Right) Laboratory incubation of the
mesocosm samples. Photo credits: S. A. E. Kools.
January 2007 199DISTURBANCE, HISTORY DETERMINE STABILITY

The variety of indirect effects as described here
emphasizes the importance of studies that incorporate
various players of the soil food web, such as nematodes,
bacteria, fungi, protozoa, and bacteria to study the
effect of (secondary) disturbances on horizontal, bot-
tom-up and top-down mechanisms.
Although communities from high- and low-contami-
nated soils reacted similarly to a high zinc application,
the community response to the heat shock reenforces
our conclusion that contaminant levels had an impact
on the soil communities. We hypothesize that the
toxicity pressure of zinc, copper and lead conferred
elevated tolerance to increased sensitivity toward an
additional stressor or that contaminants and heat stress
act in an additive way.
Total zinc is identified as an important driver for
nematode communities in the field survey (Fig. 1B),
however, the analysis of marginal and conditional
effects in the MC tests indicated that zinc, copper and
lead contents were confounding. The heavy metals zinc,
copper, and lead are reported to act additively (Sandifer
and Hopkin 1996, Korthals et al. 2000). Therefore, the
effects of the contaminants can be treated as cumulative,
although tolerance mechanisms may be refined to reflect
the difference in specific toxicity between different
metals (Blanck 2002).
We can only speculate on the similarity of mode of
action of elevated temperature and contaminants. Both
act on general stress mechanisms of species, for instance
proteins that repair intracellular damage or remove
toxicants, such as heat shock proteins and metallothio-
neins, while the response is fine tuned by specific actions
of genes (Liao et al. 2002).
Remarkably, the total biomass of nematodes re-
mained stable throughout the mesocosm incubation
time irrespective of treatment (Fig. 4B). The decrease of
the low-abundance and larger-body-sized omnivores
and carnivores was apparently quickly balanced by the
rapid increase in stress-resistant nematodes, which
typically have a short generation time, such as bacter-
ivores and small plant parasites such as Paratylenchus.
This suggests that total biomass is not a sensitive
indicator in risk assessment.
The combined impact of pollutants and another
environmental stress on natural soil communities is in
general understudied. To our knowledge, this study is
among the first to describe effects of sequential
disturbance on long-term chronically stressed natural
soil communities.
Risk of contaminated ecosystems
Ecosystems that are contaminated with heavy metals
are inherently different from naturally disturbed systems
because the effects of heavy metals are highly persistent
in time. Below a toxicity threshold, there may be a
predictable and constant selection pressure on species
for decades (Van Straalen and Timmermans 2002),
whereas effects of natural abiotic disturbances, such as
flooding or fire, tend to be temporally restricted. As a
result of the persistent heavy metal pressure in time,
tolerance mechanisms may develop. These may be
specific or broad scale and depend on the mode of
action of the contaminant (Blanck 2002). Thus, in the
case of contaminants, there may not be a general
stability of the ecosystem in either complexity, i.e.,
species and interaction richness, or functioning (Tobor-
Kaplon et al. 2005). Hence, the stability of communities
is linked to the disturbance regime and ecological
history.
The interdependence of type of disturbance, commu-
nity response, and ecosystem function can be further
investigated within a framework where species are
classified into response and functional classes (Naeem
and Wright 2003). Response classes may be extracted
from the literature while functional classes may be
determined experimentally (Heemsbergen et al. 2004).
The association of response and effect traits of
organisms may have great consequence for ecosystem
function. For instance, if both are interrelated, the
species that contribute most to an ecosystem function
are the first to go extinct (Larsen et al. 2005).
We have shown here that contamination history
resulted in a biased response of soil nematode commu-
nities toward a different disturbance regime. The patchy
distribution of contaminants in this grassland area may
conserve biodiversity when the system is exposed to a
subsequent disturbance. The less polluted spots may
function as a source for colonizers to compensate species
loss in highly polluted sites. This is in accordance with
the spatial insurance hypothesis as proposed by Loreau
et al. (2003). However, the loss of higher trophic group
from the entire system in combination with biodiversity
loss may have detrimental effects on ecosystem func-
tions. Elimination of top-down control may decrease
grazer diversity, which, in its turn, may limit production
(Duffy et al. 2005). In general, the loss of specialists,
keystone species, or entire trophic groups can signifi-
cantly effect ecosystem processes (Elmqvist et al. 2003,
Ebenman and Jonsson 2005). Therefore, the above-
described alarming effects of declined biodiversity and
loss of higher trophic groups urge the need for research
on the effects of sequential disturbances on ecosystem
functions and on restoration of chronically stressed
ecosystems.
ACKNOWLEDGMENTS
We thank the Dutch organization for scientific research
(NWO) and the Stimulation program on System-oriented
Ecotoxicological research (SSEO) for funding the project
0123.060/061 (NWO/SSEO). A. Doroszuk was funded by
TRIpartite Approach to Soil system processes (TRIAS), a
program of Delft Cluster, the Foundation for Development and
Transmission of Soil Knowledge (SKB) and NWO as project
835.80.083. BLGG, Oosterbeek, The Netherlands, and Henri-
ette Hilberink are acknowledged for help in identifying part of
the mesocosm nematode samples. We thank two anonymous
reviewers for their help in improving this manuscript.
A. W. G. VAN DER WURFF ET AL.200 Ecological Applications
Vol. 17, No. 1

LITERATURE CITED
Beare, M. H., R. W. Parmelee, P. F. Hendrix, W. X. Cheng,
D. C. Coleman, and D. A. Crossley. 1992. Microbial and
faunal interactions and effects on litter nitrogen and
decomposition in agroecosystems. Ecological Monographs
62:569–591.
Blanck, H. 2002. A critical review of procedures and
approaches used for assessing pollution-induced community
tolerance (pict) in biotic communities. Human and Ecological
Risk Assessment 8:1003–1034.
Blanck, H., and S. A. Wangberg. 1988. Induced community
tolerance in marine periphyton established under arsenate
stress. Canadian Journal of Fisheries and Aquatic Sciences
45:1816–1819.
Boivin, M. E. Y. 2005. Diversity of microbial communities in
metal-polluted heterogeneous environments. Vrije Universi-
teit, Amsterdam, The Netherlands.
Bongers, T. 1988. De nematoden van Nederland. Pirola,
Schoorl, The Netherlands.
Bongers, T. 1999. The maturity index, the evolution of
nematode life history traits, adaptive radiation and cp-
scaling. Plant and Soil 212:13–22.
Bongers, T., and H. Ferris. 1999. Nematode community
structure as a bioindicator in environmental monitoring.
Trends in Ecology and Evolution 14:224–228.
Borcard, D., and P. Legendre. 2002. All-scale spatial analysis of
ecological data by means of principal coordinates of
neighbour matrices. Ecological Modelling 153:51–68.
Bosveld, A. T. C., T. C. Klok, J. M. Bodt, and M. Rutgers.
2000. Ecologische risico’s van bodemverontreiniging in
toemaakdek in de gemeente de ronde venen. 151, Alterra,
Research Instituut voor de Groene Ruimte, Wageningen,
The Netherlands.
Chapin, F. S., et al. 2000. Consequences of changing
biodiversity. Nature 405:234–242.
Clarke, A. J., and A. M. Shepherd. 1966. Inorganic ions and the
hatching of Heterodera spp. Annals of Applied Biology 58:
497–508.
Corbeil, R. R., and S. R. Searle. 1976. Restricted maximum
likelihood (REML) estimation of variance components in the
mixed model. Technometrics 18:31–38.
Duffy, J. E., J. P. Richardson, and K. E. France. 2005.
Ecosystem consequences of diversity depend on food chain
length in estuarine vegetation. Ecology Letters 8:301–309.
Ebenman, B., and T. Jonsson. 2005. Using community viability
analysis to identify fragile systems and key-stone species.
Trends in Ecology and Evolution 20:568–575.
Ekschmitt, K., A. Klein, B. Pieper, and V. Wolters. 2001.
Biodiversity and functioning of ecological communities—
why is diversity important in some cases and unimportant in
others? Journal of Plant Nutrition and Soil Science 164:239–
246.
Elmqvist, T., C. Folke, M. Nystrom, G. Peterson, J. Bengtsson,
B. Walker, and J. Norberg. 2003. Response diversity,
ecosystem change, and resilience. Frontiers in Ecology and
the Environment 1:488–494.
Ferris, H., T. Bongers, and R. G. M. de Goede. 2001. A
framework for soil food web diagnostics: extension of the
nematode faunal analysis concept. Applied Soil Ecology 18:
13–29.
Fischer, J. M., J. L. Klug, A. R. Ives, and T. M. Frost. 2001.
Ecological history affects zooplankton community responses
to acidification. Ecology 82:2984–3000.
Freckman, D. W. 1982. Nematode contributions to ecosystem.
Pages 81–97 in D. W. Freckman, editor. Nematodes in soil
ecosystems. University of Texas Press, Austin, Texas, USA.
Garbeva, P., J. A. van Veen, and J. D. van Elsas. 2004. Microbial
diversity in soil: selection of microbial populations by plant
and soil type and implications for disease suppressiveness.
Annual Review of Phytopathology 42:243–270.
Georgieva, S. S., S. P. McGrath, D. J. Hooper, and B. S.
Chambers. 2002. Nematode communities under stress: The
long-term effects of heavy metals in soil treated with sewage
sludge. Applied Soil Ecology 20:27–42.
Giller, K. E., E. Witter, and S. P. McGrath. 1999. Assessing risks
of heavy metal toxicity in agricultural soils: Do microbes
matter? Human and Ecological Risk Assessment 5:683–689.
Griffiths, B. S., et al. 2000. Ecosystem response of pasture soil
communities to fumigation-induced microbial diversity re-
ductions: an examination of the biodiversity–ecosystem
function relationship. Oikos 90:279–294.
Griffiths, B. S., I. M. Young, and S. Caul. 1995. Nematode and
protozoan population dynamics on decomposing barley
leaves incubated at different soil matric potentials. Pedobio-
logia 39:454–461.
Hedlund, K., B. Griffiths, S. Christensen, S. Scheu, H. Setala,
T. Tscharntke, and H. Verhoef. 2004. Trophic interactions in
changing landscapes: responses of soil food webs. Basic and
Applied Ecology 5:495–503.
Heemsbergen, D. A., M. P. Berg, M. Loreau, J. R. van Haj,
J. H. Faber, and H. A. Verhoef. 2004. Biodiversity effects on
soil processes explained by interspecific functional dissimi-
larity. Science 306:1019–1020.
Helgen, S. O., and J. N. Moore. 1996. Natural background
determination and impact quantification in trace metal-
contaminated river sediments. Environmental Science and
Technology 30:129–135.
Hendrix, P. F., and R. W. Parmelee. 1985. Decomposition,
nutrient loss, and microarthropod densities in herbicide-
treated grass litter in a Georgia piedmont agroecosystem. Soil
Biology and Biochemistry 17:421–428.
Hooper, D. U., et al. 2005. Effects of biodiversity on ecosystem
functioning: a consensus of current knowledge. Ecological
Monographs 75:3–35.
Houba, V. J. G., T. M. Lexmond, I. Novozamsky, and J. J.
vanderLee. 1996. State of the art and future developments in
soil analysis for bioavailability assessment. Science of the
Total Environment 178:21–28.
Jonker, M. J., R. Sweijen, and J. E. Kammenga. 2004. Toxicity
of simple mixtures to the nematode Caenorhabditis elegans in
relation to soil sorption. Environmental Toxicology and
Chemistry 23:480–488.
Koolhaas, J. E., C. A. M. Van Gestel, J. Rombke, A. Soares,
and S. E. Jones. 2004. Ring-testing and field-validation of a
terrestrial model ecosystem (TME): an instrument for testing
potentially harmful substances. Effects of carbendazim on
soil microarthropod communities. Ecotoxicology 13:75–88.
Kools, S. A. E., M. van Roovert, C. A. M. van Gestel, and
N. M. van Straalen. 2005. Glyphosate degradation as a soil
health indicator for heavy metal polluted soils. Soil Biology
and Biochemistry 37:1303–1307.
Korthals, G. W., A. D. Alexiev, T. M. Lexmond, J. E.
Kammenga, and T. Bongers. 1996. Long-term effects of
copper and pH on the nematode community in an agro-
ecosystem. Environmental Toxicology and Chemistry 15:
979–985.
Korthals, G. W., M. Bongers, A. Fokkema, T. A. Dueck, and
T. M. Lexmond. 2000. Joint toxicity of copper and zinc to a
terrestrial nematode community in an acid sandy soil.
Ecotoxicology 9:219–228.
Larsen, T. H., N. M. Williams, and C. Kremen. 2005.
Extinction order and altered community structure rapidly
disrupt ecosystem functioning. Ecology Letters 8:538–547.
Liao, V. H. C., J. Dong, and J. H. Freedman. 2002. Molecular
characterization of a novel, cadmium-inducible gene from the
nematode Caenorhabditis elegans: a new gene that contrib-
utes to the resistance to cadmium toxicity. Journal of
Biological Chemistry 277:42049–42059.
Loreau, M., N. Mouquet, and A. Gonzalez. 2003. Biodiversity as
spatial insurance in heterogeneous landscapes. Proceedings of
the National Academy of Sciences (USA) 100:12765–12770.
January 2007 201DISTURBANCE, HISTORY DETERMINE STABILITY

Millward, R. N., and A. Grant. 2000. Pollution-induced
tolerance to copper of nematode communities in the severely
contaminated restronguet creek and adjacent estuaries,
Cornwall, United Kingdom. Environmental Toxicology and
Chemistry 19:454–461.
Molander, S., B. Dahl, H. Blanck, J. Jonsson, and M. Sjostrom.
1992. Combined effects of tri–normal-butyl tin (tbt) and
diuron on marine periphyton communities detected as
pollution-induced community tolerance. Archives of Envi-
ronmental Contamination and Toxicology 22:419–427.
Mulder, C., J. B. Dijkstra, and H. Setala. 2005. Nonparasitic
nematoda provide evidence for a linear response of func-
tionally important soil biota to increase livestock density.
Naturwissenschaften 92:314–318.
Naeem, S., and J. P. Wright. 2003. Disentangling biodiversity
effects on ecosystem functioning: deriving solutions to a
seemingly insurmountable problem. Ecology Letters 6:567–
579.
Odum, E. P. 1985. Trends expected in stressed ecosystems.
BioScience 35:419–422.
Oostenbrink, M. 1960. Estimating nematode populations by
some selected methods. Pages 85–102 in J. N. Sasser and
W. R. Jenkins, editors. Nematology. The University of North
Carolina Press, Chapel Hill, North Carolina, USA.
Pennanen, T., A. Frostegard, H. Fritze, and E. Baath. 1996.
Phospholipid fatty acid composition and heavy metal
tolerance of soil microbial communities along two heavy
metal-polluted gradients in coniferous forests. Applied and
Environmental Microbiology 62:420–428.
Ramsey, P. W., M. C. Rillig, K. P. Feris, N. S. Gordon, J. N.
Moore, W. E. Holben, and J. E. Gannon. 2005. Relationship
between communities and processes; new insights from a field
study of a contaminated ecosystem. Ecology Letters 8:1201–
1210.
Sandifer, R. D., and S. P. Hopkin. 1996. Effects of ph on the
toxicity of cadmium, copper, lead and zinc to Folsomia
candida Willem, 1902 (Collembola) in a standard laboratory
test system. Chemosphere 33:2475–2486.
Sankaran, M., and S. J. McNaughton. 1999. Determinants of
biodiversity regulate compositional stability of communities.
Nature 401:691–693.
Schwartz, M. W., C. A. Brigham, J. D. Hoeksema, K. G.
Lyons, M. H. Mills, and P. J. van Mantgem. 2000. Linking
biodiversity to ecosystem function: implications for conser-
vation ecology. Oecologia 122:297–305.
Simpson, E. H. 1949. Measurement of species diversity. Nature
163:688.
Smit, C. E., A. J. Schouten, P. J. van den Brink, M. L. P. van
Esbroek, and L. Posthuma. 2002. Effects of zinc contamina-
tion on a natural nematode community in outdoor soil
mesocosms. Archives of Environmental Contamination and
Toxicology 42:205–216.
ter Braak, C. J. F., and P. Smilauer. 2002. CANOCO reference
manual and CANODRAW for Windows user’s guide:
software for canonical community ordination (version 4.5).
Microcomputer Power, Ithaca, New York, USA.
Tobor-Kaplon, M. A., J. Bloem, P. Romkens, and P. C. de
Ruiter. 2005. Functional stability of microbial communities
in contaminated soils. Oikos 111:119–129.
Trett, M. W., B. C. Urbano, S. J. Forster, J. D. Hutchinson,
R. L. Feil, S. P. Trett, and J. G. Best. 2000. Terrestrial
meiofauna and contaminated land assessment. Environmen-
tal Science and Technology 34:1594–1602.
Van den Brink, P. J., and C. J. F. ter Braak. 1999. Principal
response curves: analysis of time-dependent multivariate
responses of biological community to stress. Environmental
Toxicology and Chemistry 18:138–148.
Van Straalen, N. M., and M. Timmermans. 2002. Genetic
variation in toxicant-stressed populations: an evaluation of
the ‘‘genetic erosion’’ hypothesis. Human and Ecological
Risk Assessment 8:983–1002.
Van Wijngaarden, R. P. A., T. C. M. Brock, and P. J. Van
den Brink. 2005. Threshold levels for effects of insecticides
in freshwater ecosystems: a review. Ecotoxicology 14:355–
380.
Verdonschot, P. F. M., and C. ter Braak. 1994. An
experimental manipulation of oligochaete communities in
mesocosms treated with chlorpyrifos or nutrient additions:
multivariate analyses with Monte Carlo permutation tests.
Hydrobiologia 278:251–266.
White, P. A., and L. D. Claxton. 2004. Mutagens in
contaminated soil: a review. Mutation Research-Reviews in
Mutation Research 567:227–345.
Yeates, G. W. 2003. Nematodes as soil indicators: functional
and biodiversity aspects. Biology and Fertility of Soils 37:
199–210.
Yeates, G. W., T. Bongers, R. G. M. DeGoede, D. W.
Freckman, and S. S. Georgieva. 1993. Feeding habits in soil
nematode families and genera: an outline for soil ecologists.
Journal of Nematology 25:315–331.
APPENDIX
Average contamination levels, soil parameters, and nematode structural responses in soil samples in the grassland ecosystem
(Ecological Archives A017-009-A1).
A. W. G. VAN DER WURFF ET AL.202 Ecological Applications
Vol. 17, No. 1