Environmental Pollution 158 (2010) 245251

Contents lists available at ScienceDirect

Environmental Pollution
journal homepage: www.elsevier.com/locate/envpol

Effects of soil properties on food web accumulation of heavy metals

to the wood mouse (Apodemus sylvaticus)
Nico van den Brink*, Dennis Lammertsma, Wim Dimmers, Marie-Claire Boerwinkel,
Annemariet van der Hout
Alterra, Wageningen UR, Box 47, NL-6700AA, Wageningen, The Netherlands
Soil properties signicantly affect accumulation of heavy metals to wood mice so; risks cannot be based on total concentrations.

a r t i c l e i n f o a b s t r a c t

Article history: Effects of soil properties on the accumulation of metals to wood mice (Apodemus sylvaticus) were
Received 9 March 2009 evaluated at two sites with different pH and organic matter content of the soil. pH and organic matter
Received in revised form content signicantly affected accumulation of Cd, Cu, Pb and Zn in earthworms and vegetation. For Cd, Cu
10 July 2009
and Zn these effects propagated through the food web to the wood mouse. Soil-to-kidney ratios differed
Accepted 12 July 2009
between sites: Cd: 0.15 versus 3.52, Cu: 0.37 versus 1.30 and Zn: 0.330.83. This was conrmed in model
calculations for Cd and Zn. Results indicate that total soil concentrations may be unsuitable indicators for
Keywords:
risks that metals pose to wildlife. Furthermore, environmental managers may, unintentionally, change
Bioavailability
Wildlife soil properties while taking specic environmental measures. In this way they may affect risks of metals
Contamination to wildlife, even without changes in total soil concentrations.
Environmental risk assessment 2009 Elsevier Ltd. All rights reserved.

1. Introduction and dissolved organic content on accumulation patterns to preda-

Wildlife is generally orally exposed to soil pollutants through
food web accumulation. Numerous studies are available that relate
soil concentrations of metals to concentrations in wildlife, in order
to assess the risks that they may pose (Beyer et al., 1985; Hamers
et al., 2002; Hendriks et al., 1995; Pascoe et al., 1996; Rogival et al.,
2007; Van den Brink et al., 2003; Wijnhoven et al., 2007). In
addition to eld based studies, modelling approaches have been
developed to assess food web accumulation of contaminants
(Gorree et al., 1995; Hendriks et al., 1995; Hunter et al., 2003;
Kooistra et al., 2001). Studies are generally based on total concen-
trations of pollutants in the soil, although the bioavailable fraction
has also been used to base risk assessments on (Rogival et al., 2007;
Torres and Johnson, 2001). It has been shown that bioavailability is
affected by soil properties like pH, organic matter content ([OM]) or
clay content ([clay]) (Bradham et al., 2006; Hobbelen et al., 2004;
Ma, 2004; Spurgeon et al., 2006; Tack et al., 1996; van Wezel et al.,
2003). Nevertheless, little is known of such effects on the food web
accumulation of metals to wildlife in terrestrial food webs. This is in
contrast to aquatic food webs, for which effects of for instance pH
* Corresponding author. Tel.: 31 317 485796; fax: 31 317 419000.
E-mail address: nico.vandenbrink@wur.nl (N. van den Brink).
tors are reviewed (Scheuhammer, 1991).
Several studies on terrestrial small mammals report linear or
log-linear relationships between total metal concentrations in soil
and in wildlife (Sharma and Shupe, 1977; Shore, 1995), but, others
report no or only weak relationships (DHave et al., 2007; Torres and
Johnson, 2001), or relationships that are dominated by other factors
such as age or bodyweight (Wijnhoven et al., 2007). Only few papers
report on levels of bioavailable fraction based on specic extraction
approaches (Rogival et al., 2007). When reviewing the literature on
accumulation of metals, it is evident that no clear relationships
between total soil concentrations and animal tissue concentrations
can be derived. Nevertheless, total soil concentrations of contami-
nants are very often used to derive their risks for wildlife. Further-
more, it is unknown whether effects of soil properties on the
availability of metals at the lower trophic levels are propagated to
the higher trophic levels. To address this we studied accumulation
of metals to the wood mouse (Apodemus sylvaticus) and its food
items at two sites. Both sites are historically contaminated with
similar levels of metals, but with different soil properties like pH and
[OM]. In the current paper we discuss effects that soil properties
have on the accumulation of metals in organisms lower in the food
web, and whether these effects propagate to the wood mouse,
higher in the food web. Since the design of the study was limited to
the two sites we used simple food web accumulation models to
interpret the results in a more generic context.

0269-7491/$ see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.envpol.2009.07.013
246 N. van den Brink et al. / Environmental Pollution 158 (2010) 245251
2. Materials and methods
2.1. Site description
2.1.1. Heteren
Heteren is the name of a oodplain area of the Neder-Rijn River near the village
Heteren in The Netherlands (51160 01N, 5 240 07E). This oodplain oods regularly
and sediments are deposited in the area. Historically, these sediments were
contaminated with heavy metals and other contaminants from upstream industrial
areas (Van den Brink et al., 2003). This process resulted in a spatially variable levels
of contaminants (Kooistra et al., 2001). Soils of these oodplains are further char-
acterised by relatively high pH-levels and high clay content. This site is managed
under an agricultural regime, with pastures and cornelds. Fields are surrounded by
hedges and trees.
2.1.2. Plateaux
The site Plateaux is located in the southern part of The Netherlands, south of the
city of Eindhoven (5157049N, 5 440 28E). It is characterised by sandy soils with
relatively low pH and organic matter content. At this site the soil is historically
polluted by atmospheric deposition of heavy metals from a zinc-smelter nearby
(Ma et al., 1983). Plateaux is a nature conservation area with low density grazing,
with solitary trees and shrubs.
2.2. Collection of soil, vegetation and earthworms
Soil was collected from the top 20 cm, after removal of visible organic material
like roots. The soil was stored in a poly-ethylene bag at 20  C prior to metal
analysis. The pH was assessed within 24 h of collection. Samples of vegetation were
clipped with stainless steel scissors and also stored in poly-ethylene bags at 20  C.
Only above ground parts were collected, since wood mice predominantly feed on
vegetative plant parts. Samples from stinging nettle (Urtica dioica), grasses and
herbs were collected at both sites and analysed separately. At both sites stinging
nettle was a dominant species. No other species were as dominant as stinging nettle,
so the remaining species were grouped to either grass or herbs. The group of grasses
included all monocotyledonous species; the group of herbs covered the dicotyle-
donous species, excluding stinging nettle. In Plateaux, additional samples were
collected from grass seeds and berries because these items can be of importance in
the diet of the wood mouse. Adult earthworms (Lumbricus rubellus) were hand
collected in the upper 20 cm of the soil, when possible from the sampled soil. This
was the only species that could be collected in high enough numbers at both sites for
proper chemical and statistical analyses. Worms were starved on a lter paper for
two days to remove their gut content. After this, the earthworms were stored in
a glass jar at 20  C. Before analysis all samples were weight to assess fresh weight.
After this the samples were freeze dried and weight to assess dry weight. Wet-to-
dry weight ratios were determined as the ratio between wet weight and dry weight.
2.3. Collection of wood mice
Wood mice (A. sylvaticus) were collected using Longworth life traps (Van den
Brink and Bosveld, 2005). Traps were baited with peanut butter and cat food, and
contained hay as bedding. Animals are caught alive in these traps, so side-catches
could have been released. Animals were sedated with CO2, and sacriced by cervical
dislocation. All procedures involving the handling of animals were conducted by
certied persons, and reviewed by an Animal Ethics Committee according to Dutch
legislation on the protection and welfare of vertebrate animals used for experimental
and other scientic purposes. In Heteren, wood mice were captured between 20th and
28th of September 2006 and 27 animals in Plateaux between 10th and 25th of August
2006. In the lab, the animals were weighed and dissected under clean conditions.
2.4. Chemical analyses
Soil was dried at 40  C, all other samples were freeze dried prior to chemical
analysis. Samples of soil, earthworms and kidneys were destructed with aqua-regia
in a microwave in teon vessels. Vegetation samples were destructed rstly with
uoric acid at room temperature to remove the silica-skeleton in the plants. After
this, the samples were destructed with aqua regia similarly to the other samples
(Novozamsky et al., 1996). Samples were analysed for cadmium (Cd), copper (Cu),
lead (Pb) and zinc (Zn) using inductively coupled plasma atomic emission spec-
trometer (ICP-AES). When concentrations were below detection limits, analyses
were performed with inductively coupled plasma mass spectrometry (ICP-MS). All
concentrations are reported as dry weight concentrations, unless stated otherwise.
For quality assurance, reference samples from clay and sandy soils and from grass
and strawberry were analysed (respectively ISE 989, ISE 949, IPE100 and IPE 125,
WEPAL, www.wepal.nl).
2.5. Statistical analyses
Differences between sites and vegetation types were analysed with Analysis of
Variance (ANOVA), with Least Signicant Differences (LSD) as post-hoc test, using
Genstat version 11.1 (www.genstat.co.uk). Generally, data on concentrations in
biotic samples are not normally distributed, a prerequisite for the use of parametric
methods, like ANOVA. To obtain normality of the data they were transformed into
their natural logarithm prior to statistical analysis. In those cases the geometric
mean will be reported, combined with the range of observations. The number of
observations is sometimes low, which could demand for non-parametric methods.
In those cases both parametric (ANOVA) and non-parametric (KruskalWallis
ANOVA) tests have been applied. In no case the results differed between the tests,
and for sake of unity we report the parametric ANOVA results.
2.6. Accumulation modelling
2.6.1. Earthworms
Ma (2004) reports a regression model Equation (1) that can be used to calculate
concentrations in L. rubellus from total soil concentrations in combination with soil
properties, using metal specic parameters for Cd, Cu, Pb and Zn (Table 1). The
database on which the parameters are based was very extensive, and covered a wide
range of soil types (Ma, 2004). This allows to apply these parameters at both sites.
logHMworm  a b  logHMsoil  c  pH d  logOM e  epigeic (1)
Equation (1), (Ma, 2004); epigeic worms live more or less in the litter. We consider
L. rubellus to be epigeic (Ernst et al., 2008).; [HMworm]: heavy metal concentration in
worm (mg/kg d.w.); [HMsoil]: heavy metal concentration in soil (mg/kg d.w.); [OM]:
% organic matter in soil; [clay]: % clay content in soil.
The r2s differ between metals, with highest r2 for Cd and Pb, and relatively low r2
for Cu and Zn (Table 1).
2.6.2. Vegetation
For grass, accumulation models have been developed similar to the previous
ones for earthworms (Equation (2), with metal specic parameters (Table 2). The
parameters are for Cd and Zn are derived based on a Dutch national wide database,
but for Cu the dataset was more or less limited to a dataset on oodplain soils (van
Wezel et al., 2003). For Cd and Zn the parameters are applicable at both sites, for Cu
the application of the models for Plateaux should be considered with some caution.
 
log HMvegetation a b  logHMsoil  c  pH d  logOM e  logclay (2)
Equation (2), (van Wezel et al., 2003); [HMvegetation]: heavy metal concentration in
vegetation (mg/kg d.w.); [HMsoil]: heavy metal concentration in soil (mg/kg d.w.);
[OM]: % organic matter in soil; [clay]: % clay in soil.
For Pb no signicant parameters could be established (van Wezel et al., 2003).
This was related to the fact that Pb generally occurs in organic complexes, and is thus
not available for uptake by vegetation. In this parameter set it can be seen that the
effect of pH on the accumulation of Zn is only limited (pH parameter is only 0.09)
while the effect of [OM] appears to be larger ([OM] parameter is 1.09), similar to
(Tack et al., 1996). For Cd and Cu the relative contribution of pH in the accumulation
patterns is larger.
2.6.3. Food web accumulation
In the current study the food web accumulation is modelled for Zn and Cd, an
essential and non-essential metal. The modelling is not performed for Pb because no
parameters are available to include soil properties in modelling the accumulation to
grasses (see above). For essential metals the modelling was restricted to Zn due to
lack of proper parameters to model the uptake of Cu by the wood mouse in the
literature. The uptake of Cd and Zn by wood mice is modelled based on their Total
Daily Intake (TDI, mg/day). TDI is calculated by Equation (3).
      
TDI pseed  HMseed pworm  HMworm pberry  HMberry pgrass
 
 HMgrass  Cons rate (3)
Equation (3), TDI calculation; pseed: fraction of seed in diet; pworm: fraction of
earthworms in diet; pberry: fraction of berries in diet; pgrass: fraction of grass in diet;
[HMseed]: geometric metal concentration in seed (mg/kg w.w.); [HMworm]:
geometric metal concentration in earthworm (mg/kg w.w.); [HMberry]: geometric
metal concentration in berries (mg/kg w.w.); [HMseed]: geometric metal concen-
tration in seed (mg/kg w.w.); Cons_rate: daily consumption rate (g/day).
Table 1
Metal specic parameters to describe accumulation of different metals from soil to
earthworm with Equation (1) (Ma, 2004).
Metal a b c d e r2
Cd 2.920 0.747 0.210 0.534 0.78
Pb 2.850 0.843 0.461 0.347 0.62
Cu 0.936 0.499 0.061 0.311 0.275 0.46
Zn 2.800 0.224 0.064 0.34
 N. van den Brink et al. / Environmental Pollution 158 (2010) 245251 247

Table 2 we applied a fraction of 15% that is deposited in the kidneys, based upon Anderson
Metal specic parameters to describe accumulation of different metals from soil to et al. (1994).
grasses with Equation (2) (van Wezel et al., 2003). For Zn the fraction of [Cacc] that is deposited in the kidneys is 0.017 (based on
control animals in Huber and Gershoff, 1970). We applied 0.28 as the dry-to-weight
Metal a b c d e r2 ratio for kidneys of the wood mouse (data not shown).
Cd 0.17 0.49 0.12 0.28 0.53
Cu 1.4 0.83 0.18 0.65 0.37
Zn 2.06 0.41 0.09 1.09 1.05 0.49
3. Results

3.1. Soil

Concentrations of earthworm and grass were analysed in the two studied areas.
Berries and grass seeds were only analysed in Plateaux. Ratios between concen-
Table 4 lists the soil concentrations and soil properties per site.
trations in berries, grass seed and grass were calculated for Plateaux and used to The pH and OM percentage were signicantly higher in the soil
calculate heavy metal concentrations for berries and seeds in Heteren. It was from the oodplains near Heteren when compared to the Plateaux
assumed that the ratios between berries, grass seeds and grass were similar between (p < 0.001). Levels of Pb and Zn did not differ signicantly between
the sites. Vegetation and earthworm specic dry-to-wet weight ratios were assessed
the sites (p > 0.05), while concentrations of Cd were signicantly
to calculate the concentration on a wet weight base.
Wood mice predominantly feed on seeds and vegetative parts of plants, but may lower near Heteren (p < 0.001). In contrast to this were the levels of
supplement this with other food items, including animal species like earthworms Cu signicantly higher in soils from the oodplains (p < 0.001).
and insects (Canova and Fasola, 1993; Rogers and Gorman, 1995; Rogival et al., 2007;
Tew et al., 2000; Watts, 1968). Watts (1968) reports a seasonal shifting diet, which
3.2. Earthworms
was also observed by Rogers and Gorman (Rogers and Gorman, 1995). The latter
study was performed on set aside land, which may be best comparable to the sites in
our study. Wood mice consumed a high fraction of grasses in spring and early Concentrations of heavy metals in earthworms differed signi-
summer, while later in the summer the amount of seeds increased (Rogers and cantly for all metals between sites (p < 0.001 for Cd, Cu and Pb,
Gorman, 1995). In spring and summer, the diet also contained animal material, both
p 0.003 for Zn). Most concentrations were higher in the area near
insects and other. The choice of diet items in the modelling was limited to the items
for which metal data are available in order to being able to compare the modelling
Plateaux, except for Cu which showed higher concentrations in
results with measured data. The diet composition used in the current study is Heteren (Table 5).
presented In Table 3. Earthworms were included in the same ratio as other animal
in Rogers and Gorman (1995) in order to reect non-vegetarian uptake, apart from
3.3. Vegetation
insects. Insects like beetles have a specic accumulation pattern, because metals
may be excreted, for instance during metamorphosis (Scheier et al., 2002). For
insects no suitable accumulation model could be selected for application, and The vegetation types included were stinging nettle, grasses and
furthermore no site specic data on metal concentrations in insects were available herbs (Table 6). Concentrations of metals in the vegetation differed
for validation. Therefore, insects were not included in the diet used for modelling.
between sites, although this was not always signicantly for all
Total consumption rate was assumed to be 4 g/day (Toal et al., 2001). The modelled
mice are born at April 17th at which a large part of the adult population was
three vegetation types. Cd concentrations were signicantly higher
expected to be have been sexually active for several weeks (Dickman and Doncaster, in all vegetation types from the Plateaux (p < 0.001 grass and herbs,
1989). The modelled animals are caught September 31st (similar to the eld p 0.003 for stinging nettle), comparable to Zn (p 0.012 for
collected animals) and are feeding the expected diet (Table 3) from the beginning of stinging nettle, p 0.027 for grass, p < 0.001 for herbs). For Cu this
the modelling.
was the opposite; concentrations in stinging nettle from Heteren
The amount of metal accumulated in the body of the wood mouse from the food
is modelled as follows: were signicantly higher (p < 0.001). Concentrations of Pb were
higher in vegetation from Plateaux, although this was only signif-
TDI$c up out$tv

icantly for stinging nettle and herbs (p < 0.001, Table 6).
Cacc  1  ec (4)
c out
In Plateaux, additional samples of berries and seeds were ana-
This equation uses metal specic kinetic parameters: c_up (g/g) and c_out (g/g/
lysed for metal content (Table 7). It is shown that concentrations
day). For Cd we applied two scenarios: Cd_1 and Cd_2. In scenario Cd_1, c_up was set
at 0.005 g/g (Lehman and Klaassen, 1986) and c_out at 0.0035 g/g/day (Gorree et al., differ between the vegetation parts. Concentrations of Cd, Pb and
1995). In a second scenario (referred to as Cd_2, based upon Anderson et al. (1994)) Zn are generally lowest in berries and highest in grass, although the
c_up was set at 0.0075 g/g and c_out at 0.0035 g/g/day. For Zn, c_up is depending on differences are not always signicant. Cu concentrations are similar
the concentrations in the food (Weigand and Kirchgessner, 1980). From their data in for all three types (p > 0.05).
Fig. 3 we derived Equation (5).

c up 1:0864  e0:0082Cons
TDI
rate0:15 (5) 3.4. Wood mice
Equation (5), Calculation of c_up for Zn; TDI Total Daily Intake (ug/day w.w.);
Cons_rate is set at 4 g/day; 0.15 is set to be the average dry-to-wet weight ratio of the In the two areas, several species of small mammals were caught,
food (Fig. 3 (Weigand and Kirchgessner, 1980)). but only wood mice were sampled in high enough numbers to
This dependence between concentration in the diet and the uptake efciency is allow for statistical analyses. In Table 8, concentrations of metals in
part of the internal regulation of zinc levels. For Zn, c_out is set at 0.045 (g/g/day)
kidneys of wood mice are listed. Concentrations of Cd were
(Huber and Gershoff, 1970).
In order to calculate kidney concentrations the fraction of [Cacc] that is deposited signicantly higher in wood mice from the Plateaux (p < 0.001).
in the kidney is calculated as follows: in scenario Cd_1 approximately 80% of [Cacc] of Concentrations of Cu, Pb and Zn did not vary signicantly between
Cd is accumulated in the liver and kidneys (Scheuhammer, 1987); the concentrations sites (p > 0.05).
in the kidneys are two times the concentrations in the liver (Scheuhammer, 1988);
the average weight of kidneys and liver for the captured wood mouse was respec-
tively 0.23 and 0.96 g (data not shown). Hence the fraction of Cd that is deposited in 4. Discussion
0:232
the kidneys is as follows: 0:8  0:2320:96 0:259. In the second scenario (Cd_2)
4.1. Soil concentrations
Table 3
Diet composition used in modelling of metals to the wood Mouse (fraction). Soil concentrations of the heavy metals in the oodplain near
Heteren are lower than, or similar to those found in moderately
pdiet Grass Seed Worm Fruit
item
contaminated oodplains of the River Waal, the contributory to the
AprilJuly 0.725 0.175 0.1 0.0 river Neder-Rijn (Hendriks et al., 1995; Van den Brink et al., 2003).
AugSept 0.45 0.45 0.0 0.1
The concentrations of Cd, Zn and Pb found in the soil from the
248 N. van den Brink et al. / Environmental Pollution 158 (2010) 245251

Table 4
Geometric means of concentration of heavy metals, and arithmetic means of soil properties in soils from the studied areas (mg/kg dw.; ranges in brackets) and signicances of
differences between sites (ANOVA, n.s.: not signicant).

pH OM % Cd mg/kg Cu mg/kg Pb mg/kg Zn mg/kg

Heteren (n 4) 7.1 (7.17.2) 8.6 (7.89.5) 1.23 (1.021.66) 32.5 (26.841.0) 57.9 (51.573.4) 207 (176273)
Plateaux (n 29) 4.6 (4.05.7) 4.5 (3.96.1) 2.23 (1.004.58) 13.6 (8.925.7) 61.9 (30.7146.9) 154 (53470)
Signicance <0.001 <0.001 <0.001 <0.001 n.s. n.s.

Table 5
Measured (geometric means) and modelled concentrations of heavy metals in earthworms from the studied areas (mg/kg dry weight; ranges in brackets) and signicances of
differences of the measured means between sites (ANOVA).

Area Cd Cu Pb Zn

Measured Modelled Measured Modelled Measured Modelled Measured Modelled

Heteren (n 4) 14.7 (11.620.4) 9.8 16.6 (12.820.5) 17.4 3.5 (2.37.0) 2.0 523 (432643) 732
Plateaux (n 29) 46.7 (16.5106) 73 9.9 (7.313.9) 19.6 13.5 (4.957.2) 46 922 (5371629) 990
Signicance <0.001 <0.001 <0.001 0.003

Plateaux are in the same range as in an earlier study performed 25
year ago in the same region (Ma et al., 1983). Current soil properties
in Plateaux are also similar to that earlier study. OM and pH in
Heteren are in the same range as in another study concerning river
oodplains from the river Waal (Hendriks et al., 1995). This indi-
cates that Heteren can be considered as a representative for river
oodplains of the rivers Waal and Neder-Rijn in the Netherlands,
while Plateaux may stand for the situation in the Kempen in the
South of The Netherlands, which is an area that is contaminated by
a nearby zinc smelting factory (Ma et al., 1983).
4.2. Accumulation from soil to earthworm
When addressing the accumulation of the different metals from
soil to earthworms, it is evident that only Cd and Zn show higher
concentrations in the worms compared to soil. Accumulation of Cd,
Pb and Zn in earthworms is higher in Plateaux than in Heteren, for
Cu this difference in accumulation is less obvious. Such differences
in accumulation have been reported before, indicating that total
soil concentrations are not the only factor governing the accumu-
lation of metals (Janssen et al., 1997). pH and OM affect the
bioavailability of metals considerably (Bradham et al., 2006; Jans-
sen et al., 1997; Ma et al., 1983; Sample et al., 1999; Spurgeon et al.,
2006) which may account for the elevated accumulation of metals
in the earthworms at Plateaux, the site with the lower pH and [OM]
(Table 4). When applying the accumulation model (Equation (1)
and Table 1), the differences in measured Cd concentrations are also
reected in the modelled concentrations (Table 5), although the
modelled difference is larger. This is similar for Pb. The measured
Cu concentrations are signicantly lower in earthworms from
the Plateaux in comparison to Heteren, but this is not reected
in the modelled concentrations. This may be due to the fact that
the predictive power of the Cu specic parameters is relatively
low (r2 0.46, Table 1). The difference in measured Zn concen-
trations between the sites is also shown in the modelled Zn
concentrations, although the difference between the measured
concentrations appears to be larger.
These modelling results show that accumulation patterns of Cd,
Pb, and Zn from soil to earthworm are affected by soil properties, in
addition to total soil concentrations. For Cu this is less the case,
perhaps this is due to other soil properties that may be more
important in governing the uptake, for instance dissolved organic
carbon (Janssen et al., 1997; Ma, 2005).
4.3. Accumulation from soil to vegetation
For all three vegetation types, metal accumulation appears to be
highest in the Plateaux (Table 9), comparable to earthworms.
Generally, the uptake of metals by vegetation is much lower than by
earthworms except for Cu. Herbs appear most efcient in accu-
mulating Cd, Pb and Zn, while stinging nettle and grass are similarly
effective. For Cu all three types show similar accumulation levels.
Differences in accumulation patterns between sites are greatest in
case of Zn, which may have been caused by differences in soil
properties (Table 9). In an earlier study it was shown that accu-
mulation of Zn in stinging nettle was affected by [OM], but not by
pH (Tack et al., 1996).

Table 6
Measured (geometric means, ranges in brackets) and modelled concentrations of heavy metals in vegetation from the studied areas (mg/kg dry weight; n.a.: not analysed; n.m.:
not modelled) and signicances of differences of the measured concentrations between sites (ANOVA; n.t.: not tested; n.s.: not signicant).

Vegetation type Area Cd Cu Pb Zn

Measured Modelled Measured Modelled Measured Modelled Measured Modelled

Stinging nettle (n 4) Heteren 0.032 (0.0230.067) n.m. 10.3 (9.211.8) n.m. 0.24 (0.200.34) n.m. 41 (3944) n.m.
Grass (n 4) Heteren 0.047 (0.0260.155) 0.124. 6.7 (5.310.3) 5.8 0.83 (0.375.35) n.m. 39 (3159) 45
Herbs (n 4) Heteren 0.096 (0.0550.185) n.m. 7.5 (6.010.8) n.m. 0.85 (0.671.02) n.m. 45 (3553) n.m.

Stinging nettle (n 7) Plateaux 0.247 (0.0910.821) n.m. 5.3 (4.57.9) n.m. 1.3 (0.92.7) n.m. 154 (49458) n.m.
Grass (n 8) Plateaux 0.415 (0.1180.914) 0.412. 6.0 (4.17.6) 12.3 1.4 (0.81.8) n.m. 96 (40225) 105
Herbs (n 7) Plateaux 1.782 (0.7074.471) n.m. 8.6 (6.414.2) n.m. 1.8 (1.12.3) n.m. 231 (106545) n.m.

Signicance
Stinging nettle 0.003 <0.001 <0.001 0.012
Grass <0.001 n.s. n.s. 0.027
Herbs <0.001 n.s. <0.001 <0.001
 N. van den Brink et al. / Environmental Pollution 158 (2010) 245251 249

Table 7
Geometric means of concentrations of metals in berries, grass seed and grass from the Plateaux (mg/kg d.w., range in brackets), and the dry-to-wet weight ratios. Cells with
similar capital in column are not signicantly different (ANOVA, a: 0.05).

Cd Cu Pb Zn Dry-to-wet weight
Berries (n 2) 0.11 (0.060.20) A 4.5 (4.34.7) A 0.4 (0.30.7) A 31.7 (25.240.0) A 0.15
Grass seed (n 5) 0.15 (0.051.84) AB 8.0 (5.317.6)A 0.6 (0.21.6) A 61.2 (3.4810.53) AB 0.45
Grass (n 8) 0.42 (0.120.91) B 6.0 (4.17.6) A 1.4 (0.81.8) B 95.6 (40225) B 0.11

Table 8
Measured (geometric mean, ranges in brackets) and modelled concentrations of heavy metals in kidneys of wood mice in two areas in The Netherlands (mg/kg dry weight) and
signicances of differences of measured concentrations between sites (ANOVA, n.s.: not signicant). Modelled concentrations of Cd according to two scenarios Cd_1 and Cd_2.
In bold the percentile at which the modelled concentration is located in the distribution of the measured concentrations at that site (concentrations log-transformed).

Area Cd Cu Pb Zn

Measured Modelled Measured Measured Measured Modelled

Heteren (n 9) 0.18 (0.061.81) 1.18 (Scenario Cd_1) P94 12.0 (1364) 1.62 (0.0622.0) 68.2 (1625) 77.3 P53
0.9 (Scenario Cd_2) P91
Plateaux (n 27) 9.85 (0.8286.6) 8.48 (Scenario Cd_1) P50 17.6 (12.726.2) 1.87 (0.5115.4) 122 (52.4941) 105.3 P40
6.6 (Scenario Cd_2) P43
Signicance <0.001 n.s. n.s. n.s.

4.4. Food web accumulation from soil to wood mice

Table 9
Ratios between concentrations in vegetation and in soil for different metals (mg/kg
dry weight)/(mg/kg dry weight). In earlier paragraphs it is shown that the accumulation of metals
in earthworms and vegetation is affected by total concentrations in
Vegetation type Site Cd Cu Pb Zn
the soil and soil properties like pH and [OM]. It is however unclear
Stinging nettle Heteren 0.03 0.32 0.00 0.20
whether effects of soil properties are propagated through the food
Plateaux 0.11 0.39 0.02 1.00
web to the wood mouse, especially for essential metals which are
Grass Heteren 0.04 0.21 0.01 0.19 physiological regulated by organisms. Several other factors may
Plateaux 0.19 0.44 0.02 0.62
also affect the food web accumulation, for instance the foraging
Herbs Heteren 0.08 0.23 0.01 0.22 ecology of the receptor (e.g. diet composition, food preferences,
Plateaux 0.80 0.63 0.03 1.50 searching behaviour), the habitat preferences of the species (e.g.
spatial distribution, prey availability), and the spatial/temporal
occurrence of food items (van den Brink, 2004). These ecological
factors, in combination with physiological feed back mechanisms
In Table 6 it is shown that the differences between the sites as like metabolism and excretion of contaminants may obscure direct
found for Cd and Zn are also reected in the modelling based on relationships between soil concentrations and properties on the
Equation (2) and Table 2, similar to the earthworms. For Cu, the one hand and food web accumulation patterns on the other. In
modelling results appear to differ between the sites, but this is not Fig. 1, the concentrations of Cd, Cu, Pb and Zn in soil and kidneys of
the case for the measured concentrations. This lack of agreement wood mice are shown for the two sites. This gure illustrates that
between modelled and measured concentrations in case of Cu was the metal specic accumulation patterns differ considerably
also noticeable in the earthworm modelling. This may point to between sites, except for Pb. It is evident that apart from Pb, the
other soil properties than the ones measured here, that may affect mice/soil ratios for the other metals are higher in Plateaux when
Cu accumulation in vegetation as well as in earthworms (i.e. dis- compared to Heteren, ranging from 1.24 times (Pb) to 24.05 times
solved organic carbon). Nevertheless, it can be concluded that also in case of Cd. These differences indicate different patterns of
for vegetation the differences in accumulation patterns for Cd, Zn accumulation, i.e. the efciency of accumulation is much higher at
and likely also Cu can be described better with inclusion of the the Plateaux.
appropriate soil properties. This however appears to be less Rogival et al. (2007) found linear relationships between total
applicable for Pb. metal concentrations and the bioavailable fraction in the soil and

Fig. 1. Concentrations of Cd(a), Cu(b), Pb(c) and Zn(d) in soil (black bars) and kidneys of wood mice in Heteren and Plateaux (grey bars) (geometric means (mg/kg dry weight), for
ranges and further details see Tables 4 and 8).
250 N. van den Brink et al. / Environmental Pollution 158 (2010) 245251
Table 10
Ratios between geometric mean concentrations in wood mouse, grass and earth-
worms from Heteren and Plateaux (mg/kg d.w.)/(mg/kg d.w.).
Cd Cu Pb
Wood mouse/grass
Heteren 3.8 1.8 2.0
Plateaux 20.3 3.0 1.6
Wood mouse/worm
Heteren 0.01 0.7 0.5
Plateaux 0.2 1.8 0.2
levels in liver and kidneys in wood mice. Total concentrations and
their bioavailable fractions in that study were highly correlated
with each other, so it is likely that the ranges of soil properties were
relatively small and did not play a major role in changing the
bioavailability of the metals between sites. Nevertheless, the soil-
to-kidney ratios of our study are in the same range as Rogival et al.
(2007), although our ratios for Cd are in their lower range.
Modelling the concentrations using Equations (1)(5) results in
Cd concentrations similar to the measured concentrations in case of
the Plateaux. The different modelling scenarios for Cd render
similar trends between the sites, although the concentrations in the
kidney from the wood mice from Plateaux are slightly under-
estimated in case of the second scenario (scenario Cd_2, Table 8).
The modelled Cd concentrations in the kidneys of the wood mice
from Heteren are slightly too high for both scenarios, but due to the
low measured concentrations this difference is relatively large
(modelled concentrations at the 9194th percentile of the distri-
bution of the measured concentrations Table 8). The measured
difference between the locations is reected in the modelled
concentrations, although the measured difference is somewhat
larger. This is also the case for Zn for which the measured difference
is larger than the modelled difference (Table 8). However, the
measured difference was not signicantly due to a high variability
in Zn concentrations (p 0.13). These results show that the
differences in accumulation between the locations are larger for
a non-essential metal like Cd than for Zn, an essential metal which
is regulated. In the models, the uptake efciency of Zn in the gut is
depending on the concentrations of Zn in the food (cf. Weigand and
Kirchgessner, 1980), which is part of the internal regulation of Zn
levels in organisms. Despite this regulation, the effects of different
soil properties are still detectable at the trophic level of wood
mouse, both for Zn and Cu (Table 8).
When modelling the food web accumulation of metals, the
uptake between trophic levels is often described by Bio-Accumu-
lation Factors (BAFs) and Bio-Magnication Factors (BMFs)
(Hendriks et al., 1995; Traas et al., 1996). This is the ratio between
the concentrations in the soil, food items and the predator or
herbivore. In Table 10 the ratios between the concentrations in the
kidneys of the wood mouse and grass or earthworms are listed. Our
results show that accumulation of metals in the lower food web is
directly affected by soil properties that govern the bioavailability of
these metals. This effect is propagated through the food web to
wood mice, higher up in the food web. This is even the case for
essential metals like Cu and Zn that are expected to be physiolog-
ically regulated to maintain homeostasis (Hunter and Johnson,
1982). This indicates that risks of metals or other pollutants for
wildlife are not necessarily related to the total soil concentrations,
which is currently often used in risk assessments. This illustrates
the need to include bioavailability of pollutants while assessing
such risks for wildlife. This study shows that accumulation is not
well described by simple BAFs or BMFs. These are applied to derive
generic environmental quality standards (EQSs) in for instance the
Netherlands (Van Vlaardingen and Verbruggen, 2007). For generic
EQSs a worst case scenario may be used, based on the most efcient
food web route of accumulation, but also assuming that pollutants
are bioavailable. When addressing site specic risks of pollutants,
local concentrations and soil properties, and species specic
Zn information should be used to assess the local risks of the pollut-
ants for the specic species residing in the area of concern.
1.8 Furthermore, knowledge on the effects of soil properties on food
1.3
web accumulation of pollutants is of importance in relation to land
use management. Soil properties often vary with land use and land
0.1 management, and it is possible that land managers affect risks
0.1
associated with soil contaminants not because they change soil
concentrations but because soil properties may change due to their
activities, and thereby the bioavailability of the contaminants.
Acknowledgements
This research was supported by the European Interreg IIIB
program (BERISP project). Funding was also obtained from the
Dutch Ministry of Agriculture, Nature Conservation and Food
Quality (from the Strategic research program Sustainable spatial
development of ecosystems, landscapes, seas and regions and the
BO-cluster (BO-02-011-007)). Michel Hendrix of Natuurmonu-
menten kindly provided access to Plateaux.
References
Anderson, O., Nielsen, J.B., Sorensen, J.A., Scherrebeck, L., 1994. Experimental
localization of intestinal uptake sites for metals (Cd, Hg, Zn, Se) in vivo in mice.
Environmental Health Perspectives 102, 199206.
Beyer, W.N., Pattee, O.H., Sileo, L., Hoffman, D.J., Mulhern, B.M., 1985. Metal
contamination in wildlife living near two zinc smelters. Environmental Pollu-
tion 638, 6386.
Bradham, K.D., Dayton, E.A., Basta, N.T., Schroder, J., Payton, M., Lanno, R.P., 2006.
Effect of soil properties on lead bioavailability and toxicity to earthworms.
Environmental Toxicology and Chemistry 25, 769775.
Canova, L., Fasola, M., 1993. Food habits and trophic relationships of small mammals
in six habitats of the northern Po plain (Italy). Mammalia 57, 189199.
DHave, H., Scheirs, J., Covaci, A., Van den Brink, N.W., Verhagen, R., De Coen, W., 2007.
Nondestructive pollution exposure assessment in the European hedgehog (Eri-
naceus europaeus): IV. Hair versus soil analysis in exposure and risk assessment
of organochlorine compounds. Environmental Pollution 145, 861868.
Dickman, C.R., Doncaster, C.P., 1989. The ecology of small mammals in urban
habitats. II. Demography and dispersal. Journal of Animal Ecology 58, 119127.
Ernst, G., Zimmermann, S., Christie, P., Frey, B., 2008. Mercury, cadmium and lead
concentrations in different ecophysiological groups of earthworms in forest
soils. Environmental Pollution 156, 13041313.
Gorree, M., Tamis, W.L.M., Traas, T.P., Elbers, M.A., 1995. BIOMAG: a model for
biomagnication in terrestrial food chains. The case of cadmium in the Kempen,
The Netherlands. Science of the Total Environment 168, 215223.
Hamers, T., Smit, L.A.M., Bosveld, A.T.C., van den Berg, J.H.J., Koeman, J.H., van
Schooten, F.J., Murk, A.J., 2002. Lack of a distinct gradient in biomarker
responses in small mammals collected at different distances from a highway.
Archives of Environmental Contamination and Toxicology 43, 345355.
Hendriks, A.J., Ma, W.C., Brouns, J.J., De Ruiter-Dijkman, E.M., Gast, R., 1995.
Modelling and monitoring organochlorine and heavy metal accumulation in
soils, earthworms and shrews in Rhine-delta oodplains. Archives of Environ-
mental Contamination and Toxicology 29, 115127.
Hobbelen, P.H.F., Koolhaas, J.E., van Gestel, C.A.M., 2004. Risk assessment of heavy metal
pollution for detritivores in oodplain soils in the Biesbosch, The Netherlands,
taking bioavailability into account. Environmental Pollution 129, 409419.
Huber, A.M., Gershoff, S.N., 1970. Effects of dietary zinc and calcium on the retention
and distribution of zinc in rats fed semipuried diets. Journal of Nutrition 100,
949954.
Hunter, J.G., Burger, J., Cooper, K.R., 2003. Use of an integrated mercury food web
model for ecological risk assessment. Journal of Environmental Science and
Health Part A Toxic Hazardous Substances and Environmental Engineering 38,
12011214.
Hunter, B.A., Johnson, M.S., 1982. Food chain relationships of copper and cadmium
in contaminated grassland ecosystems. Oikos 38, 108117.
Janssen, R.P.T., Posthuma, L., Baerselman, R., Den, H.H.A., Van, V.R.P.M.,
Peijnenburg, W.J.G.M., 1997. Equilibrium partitioning of heavy metals in Dutch
eld soils. II. Prediction of metal accumulation in earthworms. Environmental
Toxicology and Chemistry 16, 24792488.
Kooistra, L., Leuven, R., Nienhuis, P.H., Wehrens, R., Buydens, L.M.C., 2001. A
procedure for incorporating spatial variability in ecological risk assessment of
Dutch River oodplains. Environmental Management 28, 359373.
Lehman, L.D., Klaassen, C.D., 1986. Dosage-dependent disposition of cadmium
administered orally to rats. Toxicology and Applied Pharmacology 84, 159167.
 N. van den Brink et al. / Environmental Pollution 158 (2010) 245251
Ma, W.C., 2004. Estimating heavy metal accumulation in oligochaete earthworms:
a meta-analysis of eld data. Bulletin of Environmental Contamination and
Toxicology 72, 663670.
Ma, W.C., 2005. Critical body residues (CBRs) for ecotoxicological soil quality
assessment: copper in earthworms. Soil Biology and Biochemistry 37, 561568.
Ma, W., Edelman, T., van Beersum, I., Jans, T., 1983. Uptake of cadmium, zinc, lead,
and copper by earthworms near a zinc-smelting complex: inuence of soil pH
and organic matter. Bulletin of Environmental Contamination and Toxicology
30, 424427.
Novozamsky, I., van Eck, R., Houba, V.J.G., van der Lee, J.J., 1996. Solubilization of
plant tissue with nitric acid-hydrouoric acid-hydrogen peroxide in a closed-
system microwave digestor. Communications in Soil Science and Plant Analysis
27, 867875.
Pascoe, G.A., Blanchet, R.J., Linder, G., 1996. Food chain analysis of exposures and
risks to wildlife at a metals-contaminated wetland. Archives of Environmental
Contamination and Toxicology 30, 306318.
Rogers, L.M., Gorman, M.L., 1995. The diet of the wood mouse Apodemus sylvaticus
on set-aside land. Journal of Zoology 235, 7783.
Rogival, D., Scheirs, J., Blust, R., 2007. Transfer and accumulation of metals in a soil-
diet-wood mouse food chain along a metal pollution gradient. Environmental
Pollution 145, 516528.
Sample, B.E., Suter, G.W., Beauchamp, J.J., Efroymson, R.A., 1999. Literature-derived
bioaccumulation models for earthworms: development and validation. Envi-
ronmental Toxicology and Chemistry 18, 21102120.
Scheier, R., Gomot de Vaueury, A., Toussaint, M.L., Badot, P.M., 2002. Transfer and
effects of cadmium in an experimental food chain involving the snail Helix
aspersa and the predatory carabid beetle Chrysocarabus splendens. Chemo-
sphere 48, 571579.
Scheuhammer, A.M., 1987. The chronic toxicity of aluminium, cadmium, mercury
and lead in birds: a review. Environmental Pollution 46, 263295.
Scheuhammer, A.M., 1988. The dose-dependent deposition of cadmium into organs
of Japanese Quail following oral administration. Toxicology and Applied Phar-
macology 95, 153161.
Scheuhammer, A.M., 1991. Effects of acidication on the availability of toxic metals
and calcium to wild birds and mammals. Environmental Pollution 71,
329376.
Sharma, R.P., Shupe, J.L., 1977. Trace metals in ecosystems: relationships of the
residues of copper, molybdenum, selenium, and zinc in animal tissues to those
in vegetation and soil in the surrounding environment. In: Drucker, H.,
Wildung, R.E. (Eds.), Biological Implications of Metals in the Environment.
Technical Information Center Energy and Development Administration.,
Springeld, pp. 595608.
Shore, R.F., 1995. Predicting cadmium, lead and uoride levels in small mammals
from soil residues and by species-species extrapolation. Environmental Pollu-
tion 88, 333340.
Spurgeon, D.J., Lofts, S., Hankard, P.K., Toal, M., McLellan, D., Fishwick, S.,
Svendsen, C., 2006. Effect of pH on metal speciation and resulting metal uptake
251
and toxicity for earthworms. Environmental Toxicology and Chemistry 25,
788796.
Tack, F.M., Verloo, M.G., West, H.K., Davies, M.S., Morgan, A.J., Herbert, R.J., 1996.
Metal contents in stinging nettle (Urtica dioica) as affected by soil character-
istics. Science of the Total Environment 192, 3139.
Tew, T.E., Todd, I.A., MacDonald, D.M., 2000. Arable habitat use by wood mice
(Apodemus sylvaticus) 2. Microhabitat. Journal of Zoology 250, 305311.
Toal, M.E., Copplestone, D., Johnson, M.S., Jackson, D., Jones, S.R., 2001. A dynamic
compartmental food chain model of radiocaesium transfer to Apodemus syl-
vaticus in woodland ecosystems. Science of the Total Environment 267, 5365.
Torres, K.C., Johnson, M.L., 2001. Bioaccumulation of metals in plants, arthropods, and mice
at a seasonal wetland. Environmental Toxicology and Chemistry 20, 26172626.
Traas, T.P., Luttik, R., Jongbloed, R.H., 1996. A probabilistic model for deriving soil
quality criteria based on secondary poisoning of top predators .1. Model
description and uncertainty analysis. Ecotoxicology and Environmental Safety
34, 264278.
van den Brink, N.W., 2004. Soil and higher organisms: from bottom-up relations to
top-down monitoring. In: Doelman, P., Eijsackers, H.J.P. (Eds.), Vital Soil, Func-
tion Value and Properties. Developments in Soil Science. Elsevier, Amsterdam,
pp. 215244.
Van den Brink, N.W., Bosveld, A.T.C., 2005. Alkoxyresorun-O-deethylase activities
and polychlorinated biphenyl patterns in shrews as biomarkers in environ-
mental risk assessments: sensitivity and specicity. Environmental Science and
Technology 39, 73377343.
Van den Brink, N.W., Groen, N.M., De Jonge, J., Bosveld, A.T.C., 2003. Ecotoxicological
suitability of oodplain habitats in The Netherlands for the little owl (Athene
noctua vidalli). Environmental Pollution 122, 127134.
Van Vlaardingen, P.L.A., Verbruggen, E.M.J., 2007. Guidance for the Derivation of
Environmental Risk Limits Within the Framework of International and National
Environmental Quality Standards for Substance in The Netherlands (INS).
Report 601782001/2007. RIVM, Bilthoven.
van Wezel, A.P., de Vries, W., Beek, M., Otte, P.F.M., Lijzen, J.P.A., Mesman, M., van
Vlaardingen, P.L.A., Tuinstra, J., van Elswijk, M., Romkens, P.F.A.M., Bonten, L.T.C.,
2003. Bodemgebruikswaarden voor landbouw, natuur en waterbodem. Tech-
nisch wetenschappelijke aeiding van getalwaarden. Report 711701031/2003.
RIVM, Bilthoven, The Netherlands.
Watts, C.H.S., 1968. The foods eaten by wood mice (Apodemus sylvaticus) and bank
voles (Clethrionomys glareolus) in Wytham Woods, Berkshire. Journal of Animal
Ecology 37, 2541.
Weigand, E., Kirchgessner, M., 1980. Total true efciency of zinc utilization: deter-
mination and homeostatic dependence upon the zinc supply status in young
rats. Journal of Nutrition 110, 469480.
Wijnhoven, S., Leuven, R.S.E.W., van der Velde, G., Jungheim, G., Koelemij, E.I., de
Vries, F.T., Eijsackers, H.J.P., Smits, A.J.M., 2007. Heavy-metal concentrations in
small mammals from a diffusely polluted oodplain: importance of species-
and location-specic characteristics. Archives of Environmental Contamination
and Toxicology 52, 603613.