Download full-text PDF

Phytotoxicity of nickel in a range of European soils: Influence of soil properties, Ni solubility and speciation

Article · February 2007with88 Reads
DOI: 10.1016/j.envpol.2006.04.008 · Source: PubMed
Abstract
We investigated the influence of soil properties on Ni toxicity to barley root and tomato shoot growth, using 16 European soils. The effective concentration of added Ni causing 50% inhibition (EC(50)) ranged from 52 to 1929mgkg(-1) and from 17 to 920mgkg(-1) for the barley and tomato test, respectively, representing 37- and 54-fold variation among soils. Soil cation exchange capacity was the best single predictor for the EC(50). The EC(50) based on either the Ni concentration or free Ni(2+) activity in soil solution varied less among soils (7-14 fold) than that based on the total added Ni, suggesting that solubility of Ni is a key factor influencing its toxicity to plants. The EC(50) for free Ni(2+) activity from the barley test decreased with increasing pH, indicating a protective effect of protons. The results can be used in the risk assessment of Ni in the terrestrial environment.
Phytotoxicity of nickel in a range of European soils: Influence of
soil properties, Ni solubility and speciation
Corinne P. Rooney, Fang-Jie Zhao*, Steve P. McGrath
Agriculture and Environment Division, Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, UK
Received 12 January 2006; received in revised form 3 April 2006; accepted 6 April 2006
Soil cation exchange capacity explains variation among soils in the toxicity threshold values of nickel to plants.
Abstract
We investigated the influence of soil properties on Ni toxicity to barley root and tomato shoot growth, using 16 European soils. The effective
concentration of added Ni causing 50% inhibition (EC
50
) ranged from 52 to 1929 mg kg
1
and from 17 to 920 mg kg
1
for the barley and tomato
test, respectively, representing 37- and 54-fold variation among soils. Soil cation exchange capacity was the best single predictor for the EC
50
.
The EC
50
based on either the Ni concentration or free Ni
2þ
activity in soil solution varied less among soils (7e14 fold) than that based on the
total added Ni, suggesting that solubility of Ni is a key factor influencing its toxicity to plants. The EC
50
for free Ni
2þ
activity from the barley
test decreased with increasing pH, indicating a protective effect of protons. The results can be used in the risk assessment of Ni in the terrestrial
environment.
Ó2006 Elsevier Ltd. All rights reserved.
Keywords: Nickel; Bioavailability; Plant toxicity; Speciation; Soil properties; Risk assessment
1. Introduction
Risk assessments are being carried out in the European
Union for existing chemicals, including metals and their com-
pounds, in relation to their effects on human health and terres-
trial, aquatic and atmospheric ecosystems due to concern
regarding the potential risks of existing chemicals in the envi-
ronment (European Commission, 2003). The risk assessment
process includes assessment of effects, exposure assessment
and risk characterisation, and the procedure is described in de-
tail in the Technical Guidance Document (TGD) (European
Commission, 2003). The results of a risk assessment provide
the principal basis for legislative decisions to reduce the risks
arising from manufacture, transport, storage, formulation into
a preparation or other processing, use and disposal or recovery
of existing substances. Therefore, these risk assessment exer-
cises concern the sustainability of current and future use of
a substance rather than the identification and remediation of
previously contaminated sites. For terrestrial ecosystems,
risk assessments require toxicity evaluations of the effects of
contaminants on plants, soil microbes and invertebrates. The
focus of this study is on the toxic effect of Ni to higher plants
evaluated with a range of soils representing major soil types in
Europe.
Bioavailability of Ni varies with soil properties and this has
to be taken into account in the risk assessment. Soil properties
that influence Ni bioavailability to plants include pH, and the
contents of organic matter, clay and Fe oxides/hydroxides
(Wallace et al., 1977a; Sauerbeck, 1991; Weng et al., 2003,
2004). There have been a number of studies on the toxic effect
of Ni to plants grown on soils (Wallace et al., 1977a; Khalid
and Tinsley, 1980; Dixon, 1988; Sauerbeck and Hein, 1991a;
Weng et al., 2003, 2004). For example, Wallace et al.
(1977b) determined the effects of Ni on seedlings of several
* Corresponding author. Tel.: þ44 1582 763133; fax: þ44 1582 760981.
E-mail address: fangjie.zhao@bbsrc.ac.uk (F.-J. Zhao).
0269-7491/$ - see front matter Ó2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.envpol.2006.04.008
Environmental Pollution 145 (2007) 596e605
www.elsevier.com/locate/envpol
plant species grown in a loam soil adjusted to pH values from
4.2 to 8.2. The greatest reduction in yield of corn (Zea mays),
soybean (Glycine max), barley (Hordeum vulgare) and bush
bean (Phaseolus vulgare) occurred at the lowest pH levels.
In an additional experiment, the reduction in yield of barley
was similar with 25 mg Ni kg
1
added to soil at pH 5.6 and
400 mg Ni kg
1
to soil at pH 7.2. Sauerbeck and Hein
(1991b) showed that an addition of 50 mg Ni kg
1
to an acidic
sandy soil caused damage to several plant species, whereas
100 mg Ni kg
1
added to a neutral Luvisol was not toxic to
plants. Weng et al. (2004) reported a 32-fold variation in the
EC
50
(effective concentration causing a 50% inhibition) of to-
tal added Ni, from 40 to 1275 mg kg
1
soil, in three soils in
which pH was artificially varied by the addition of lime. How-
ever, the published data are insufficient to be used to ade-
quately quantify the effects of soil properties on the
bioavailability of Ni to plants. Achieving this goal requires
a systematic study using standard test methods and the same
plant species, on soils with a wide range of physicochemical
properties.
This study was designed to derive toxicity thresholds for Ni
in test plant species grown on a range of European soils. Our
main aim was to establish empirical relationships between soil
properties and toxicity thresholds for Ni, which can be used in
the risk assessment of Ni in the terrestrial environment. A fur-
ther aim was to investigate whether variation in plant toxicity
response among soils can be better explained by the differ-
ences in the solubility of Ni or by the speciation of Ni in
the soil solution alone, without using soil properties.
2. Materials and methods
2.1. Soil sampling and treatment
Sixteen soils were sampled from locations throughout Europe (Table 1).
These soils are representative of the major soil types in the region and cover
a wide range in soil properties expected to affect the bioavailability of Ni.
Agricultural soils were collected from the plough layer and natural or undis-
turbed soils from the surface horizons after clearance of litter (Oorts et al.,
2006b). After sampling, the soils were air-dried and sieved to <4 mm and
stored in drums at ambient temperature until use. Table 1 shows the locations
of soils and selected soil properties.
Each soil was amended to obtain a range of seven Ni concentrations in
a geometric series (control þsix Ni doses; twofold differences between adja-
cent doses). The maximum target concentration varied from 50 to
3200 mg Ni kg
1
depending on soil properties (soil organic matter and pH).
Nickel was added to soil by spraying appropriate volumes of a NiCl
2
solution
(50 g Ni l
1
) diluted with deionised water in a total volume of 50 ml kg
1
oven
dry soil. This method ensured an even distribution of the Ni solution on the
soil and that the pore structure of the soils remained intact. After amendment,
each sample was thoroughly mixed on a plastic sheet by hand. Control samples
were treated in a similar manner using deionised water only. All treatments
were equilibrated for 7 days before use in plant bioassays.
2.2. Plant assays
Two plant toxicity tests were performed: a root length assay, based on In-
ternational Organization for Standardization (ISO) 11269-1 (International Or-
ganisation for Standardisation (ISO), 1993) and a shoot growth assay, based on
ISO 11269-2 (International Organisation for Standardisation (ISO), 1995).
A preliminary germination test was carried out using six plant species
(H. vulgare,Lycopersicon esculentum,Z. mays,Brassica napus,Triticum
aestivum and Lepidium heterophyllum). This showed that L. esculentum
(tomato) and H. vulgare (barley) had the best growth across all 16 soils, and
these two species were used in the assays. Plant tests were carried out in
a glasshouse with natural light supplemented with 600 W SON-T sodium
lamps (OSRAM, Langley, Berkshire, UK) to maintain a minimum light inten-
sity of 350 mmol m
2
S
1
for a duration of 16 h per day, and temperature of
20 2C (day) and 16 2C (night). Soil moisture content was maintained
at w60% water holding capacity.
2.2.1. Barley root length assay
Soil was packed into plastic cylinders (diameter 45 mm and length
100 mm) to approximate the bulk density of each soil. Five pre-germinated
barley (H. vulgare cv. Regina) seeds with radicle <2 mm in length were
planted in each cylinder. Each Ni treatment of the 16 soils was replicated in
three cylinders. To reduce evaporation of moisture from the soils, the cylinders
were placed inside propagators with transparent plastic covers and slits open to
allow air exchange. After 4 days plants were removed intact and the length of
Table 1
Selected properties of the soils used in the plant toxicity tests
Soil No. Location (country) Soil pH
(in 0.01 M CaCl
2
)
Ambient
soil Ni
Fe oxide Organic C CaCO
3
Soil CEC Exch. Ca Exch. Mg Sand Silt Clay
(mg kg
1
) (%) (cmol kg
1
) (%)
1 Houthalen (Belgium) 3.6 1 902 1.73 0.0 1.84 0.11 0.03 94.9 4.8 0.4
2 Zegveld (The Netherlands) 4.1 26 14299 33.05 0.0 52.75 70.50 5.55 47.8 18.2 34.0
3 Montpellier (France) 4.1 16 610 0.25 0.0 8.39 2.59 2.88 63.3 11.4 25.3
4 Rhydtalog (UK) 4.2 3 2202 12.52 0.0 11.91 8.85 1.32 36.8 50.5 12.7
5 Jyndevad (Denmark) 4.5 1 2188 1.32 0.0 1.84 1.31 0.06 95.0 3.5 1.5
6Ko
¨vlinge II (Sweden) 5.1 2 2679 2.47 0.0 4.31 3.33 0.40 82.6 13.4 3.9
7 Aluminusa (Italy) 5.6 19 1941 0.99 0.0 19.26 10.61 4.69 29.3 23.7 46.9
8 Borris (Denmark) 5.6 3 2758 1.33 0.0 4.91 4.31 0.38 78.6 17.1 4.3
9 Woburn UK 6.1 39 17774 4.30 0.0 28.87 21.36 5.02 40.7 24.0 35.3
10 Ter Munck (Belgium) 6.7 11 2848 1.09 0.0 7.80 6.65 0.62 11.0 79.4 9.6
11 Souli (Greece) 7.0 81 919 0.45 1.5 12.85 9.25 1.16 52.4 14.4 33.2
12 Marknesse (The Netherlands) 7.6 19 4340 1.14 11.5 19.44 19.51 0.57 12.3 67.7 19.9
13 Bre
´cy (France) 7.5 113 3693 1.37 19.1 23.57 21.62 0.48 11.4 39.4 49.2
14 Cordoba 2 (Spain) 7.6 24 660 0.49 34.8 35.26 33.66 0.77 23.0 21.6 55.4
15 Cordoba 1 (Spain) 7.6 18 797 0.53 21.8 13.35 11.52 1.12 46.3 33.9 19.8
16 Guadalajara (Spain) 7.7 11 243 0.31 29.0 13.27 12.40 0.72 55.2 27.6 17.2
597C.P. Rooney et al. / Environmental Pollution 145 (2007) 596e605
the longest root on each plant was recorded. Each replicate value represents
mean root lengths of five plants per pot.
2.2.2. Tomato shoot growth assay
Twenty tomato seeds (L. esculentum, cv. Moneymaker) were sown in three
replicate pots (10 cm diameter) of each Ni treatment of each soil. Again, soil
was packed into the pots to approximate the bulk density of each soil and the
pots were randomised. Following emergence, seedlings were thinned to five
plants per pot, which were grown for 21 days from the time of emergence.
Soil moisture content was maintained at 60e70% of water holding capacity
by additions of deionised water. To prevent possible nutrient deficiencies in
some of the soils, a dilute nutrient solution was added to all pots on three oc-
casions during the growth period. The nutrient solution contained 3 mM
NH
4
NO
3
, 3 mM KNO
3
and 1 mM KH
2
PO
4
and was applied at a rate of
50 ml per pot in each application. At harvest, tomato shoots were cut just
above the soil surface, washed with deionised water, and dried at 70 C for
48 h, and the dry biomass was recorded.
2.3. Chemical analyses
Total soil Ni concentration was analysed by aqua regia digestion (McGrath
and Cunliffe, 1985), followed by determination using inductively coupled
plasma atomic emission spectrometry (ICP-AES; Fisons ARL Accuris, Ecu-
blens, Switzerland). Blanks and reference materials (National Institute of Stan-
dard & Technology standard reference soil No. 2711, Gaithersburg, MD, USA
and an in-house standard soil material) were included for quality assurance.
Actual concentrations of added Ni were calculated for each Ni amendment
by subtracting the total Ni concentration of the unamended control. Soil pH
was measured in 0.01 M CaCl
2
(1:5 soil:solution ratio). Total carbon was mea-
sured by ignition with a Variomax CN analyzer (Pro-Tech, Coolum Beach,
Queensland, Australia). Organic carbon was calculated as the difference be-
tween total and inorganic carbon content. The carbonate-C was determined
from the pressure increase after addition of HCl to the soil in closed con-
tainers, including FeSO
4
as a reducing agent. The concentration of amorphous
iron, aluminium and manganese was determined by an ammonium oxalate ex-
traction in the dark (Schwertman, 1964). Iron, aluminium and manganese in
these extracts were analysed by ICP-AES. The silver thiourea method (Chha-
bra et al., 1975) was used to measure the CEC and exchangeable cations at the
pH of the soil. Silver and the exchangeable cations were determined in the ex-
tract by ICP-AES.
At the beginning of the plant bioassays, remaining soil was equilibrated for
a further 2 days at a moisture content equivalent to 80% of water holding ca-
pacity, and then soil solution was extracted. Solutions were collected by centri-
fugation at 3000 gfor 15 min using a two-chamber method. Each Ni treatment
for each soil was extracted and analysed in duplicate. Immediately after extrac-
tion, soil solution samples were filtered using 0.45 mm cellulose acetate mem-
brane filters and solution pH was measured. The concentrations of Ni and other
major and trace elements were determined by ICP-AES. Dissolved organic
carbon (DOC) was measured using a DOC analyzer (Thermalox, Analytical
Sciences, Cambridge, UK). Major anions were determined using continuous
colorimetric flow analysis (SAN
PLUS
, Skalar, Breda, The Netherlands).
2.4. Metal speciation
Free Ni
2þ
activities in soil solutions were calculated using the WHAM
(Windermere Humic Aqueous Model) model VI (Tipping, 1998). Inputs to the
model included solution pH, and concentrations of fulvic acid, Ni, and other cat-
ions and anions. The concentration of fulvic acid was calculated from DOC by
assuming that the dissolved organic matter (DOM) contained 50% C, and that
65% of the dissolved fulvic acid was active for metal binding (Tipping et al.,
2003). In addition, estimated values of free Fe
3þ
and Al
3þ
activities were in-
cluded in the input data set, calculated from the solubility products of
Fe(OH)
3
(log K
so
¼2.7) and Al(OH)
3
(log K
so
¼8.5) (Tipping et al., 2003).
To investigate the soil properties that control Ni solubility, a pH-dependent
Freundlich equation was used (Tye et al., 2004):
log½NiSolid=Ni2þnF¼k1þk2pH þk3log½Ið1Þ
where [Ni
Solid
] is the concentration of Ni in soil solid phase (after solution
phase Ni was subtracted), (Ni
2þ
) the activity of free Ni
2þ
calculated by the
WHAM model VI, [I] the ionic strength of the soil solution, k
1e3
the regres-
sion coefficients, and n
F
the power term of the Freundlich isotherm. Soil solu-
tion pH was used in Eq. (1). The inclusion of the ionic strength term was
necessary because ionic strength varied substantially as a result of the addi-
tions of different concentrations of NiCl
2
. In addition, [Ni
Solid
] was expressed
on the basis of either the whole soil (mol Ni kg
1
soil), soil organic C
(mol Ni kg
1
organic C), or soil CEC (mol Ni mol
1
CEC). Model optimisa-
tion was undertaken using the data analysis tool Solver in Microsoft Excel. Er-
rors in the prediction of pNi
2þ
(¼log(Ni
2þ
)) were minimised and expressed
as residual standard deviations (RSD) (Tye et al., 2004).
2.5. Statistics
A logelogistic curve (Haanstra et al., 1985) (Eq. (2)) was fitted to the
doseeresponse data for each of the soils tested using Genstat for Windows
version 8 (VSN International, Hemel Hempstead, UK):
Y¼Y0
1þeðbðXMÞÞ ð2Þ
where Y¼barley root length or tomato shoot biomass, X¼log
10
(actual
concentration of added Ni), and Y
0
,Mand bare parameters to be fitted, where
M¼log
10
(EC
50
), EC
50
being the effective concentration of added Ni that re-
duced root length or shoot biomass by 50% (i.e. when Y¼50%Y
0
). The zero
metal dose in the control soil was attributed a very small value
(1 10
3
mg kg
1
) to allow log transformation prior to curve fitting. The ad-
equacy of the model was checked by examining the distribution of the resid-
uals and the relationship between the residuals and the fitted values. The metal
dose causing 10% and 50% inhibition (EC
10
and EC
50
) and their 95% confi-
dence intervals were derived from the fitted curve parameters and standard
errors according to Haanstra et al. (1985). Single and stepwise multiple regres-
sion analysis was used to determine relationships between EC
x
(x¼10 or 50)
data and soil properties. All EC
x
and soil properties except pH were log trans-
formed prior to regression analysis.
3. Results
3.1. Doseeresponse curves and toxicity threshold values
for soil Ni
Plant growth in the unamended control varied among soils
(Figs. 1 and 2), with indications that low soil pH and heavy
soil texture resulted in small barley root length or tomato shoot
biomass. Barley root length and tomato shoot biomass were
significantly inhibited by Ni additions in all soils. Relative to
the unamended control, the inhibition on barley root length or
tomato shoot biomass by the maximum concentration of Ni
added to soil was >70%. Hormesis, i.e. a stimulation of re-
sponse occurs at low doses followed by inhibition at high doses,
was significant ( p<0.05) in two soils in the barley root length
test (Houthalen and Marknesse; Fig. 1) and one soil in the to-
mato shoot growth test (Marknesse; Fig. 2). It was possible to
fit logelogistic curves (Eq. (2)) to the doseeresponse data
from both tests in all soils, with the model explaining 85%
of the variation in the data (Figs. 1 and 2).
The toxicity thresholds based on added Ni varied among the
16 soils, with the EC
50
values ranging from 52 to
1929 mg kg
1
and from 17 to 920 mg kg
1
for the barley
root length and tomato shoot growth tests, respectively, repre-
senting 37- and 54-fold variations among soils (Fig. 3). The
range of EC
10
was from 31 to 1101 mg kg
1
and from 10 to
598 C.P. Rooney et al. / Environmental Pollution 145 (2007) 596e605
599 mg kg
1
for the barley and tomato test, respectively, rep-
resenting 36- and 60-fold variation among soils (Fig. 3). In 15
out of the 16 soils, the EC
50
values were greater for barley root
length than for tomato growth, indicating that tomato shoot
growth was generally more sensitive to Ni toxicity than barley
root growth.
3.2. Relationships between Ni toxicity threshold values
and soil properties
Single regressions were carried out between Ni toxicity
threshold values (EC
x
) and various soil properties (CEC, ex-
changeable Ca or Mg, pH, organic C content, Fe oxide
concentration and clay content; based on log-transformed
data except for pH). Soil CEC and exchangeable Ca concen-
tration were the two best single predictors of the variation of
EC
x
among soils for both plant tests (Table 2). Soil CEC ex-
plained 85% and 60% of the variance for the barley and
tomato EC
x
, respectively (Table 2). Exchangeable Ca concen-
tration explained a similar percentage of variance for the to-
mato EC
x
, but a slightly lower percentage (>82%) of
variance for the barley EC
x
. The lower r2
adj values in the to-
mato data set were due to two apparent outliers (the Montpel-
lier and Aluminusa soils). When these two soils were excluded
from the regression analysis, r2
adj increased to 86e88% for soil
CEC and to 71e76% for exchangeable Ca concentration,
Woburn
Zegveld
Houthalen
Jyndevad
Kovlinge
Mean root length per pot (mm)
0
20
40
60
80
100
0
20
40
60
80
100
Aluminusa
Borris
Brecy
Souli
Ter Munck
Cordoba 1
Cordoba 2
Guadalajara
Added soil Ni (mg kg-1)
Marknesse
Montpellier
Rhydtalog
10-1 10010110210310410-1 10010110210310410-1 100101102103104
Fig. 1. Doseeresponse curves for the barley root length test in 16 European soils. Symbols represent all replicated data points and lines are the fitted logelogistic
curves.
0.0
0.1
0.2
0.3
0.4
0.5
Brecy
Cordoba 1
Kovlinge
Added soil Ni (mg kg-1)
0.00
0.05
0.10
0.15
Jyndevad
Souli
0.00
0.05
0.10 Houthalen
Montpellier
Shoot biomass per pot (g DW)
0.0
0.2
0.4
0.6
0.8
1.0
Borris
Ter Munck
Rhydtalog
0.0
0.5
1.0
1.5
Marknesse
Woburn
Zegveld
0.00
0.05
0.10
0.15
Aluminusa
Cordoba 2
Guadalajara
10-1 10010110210310410-1 10010110210310410-1 100101102103104
Fig. 2. Doseeresponse curves for the tomato shoot growth test in 16 European soils. Symbols represent all replicated data points and lines are the fitted
logelogistic curves.
599C.P. Rooney et al. / Environmental Pollution 145 (2007) 596e605
respectively. The relationships between soil CEC and the EC
x
values for both plant tests for all soils are shown in Fig. 4. The
two outlier soils have relatively low pH and organic C content.
Therefore, the CEC of the soils is contributed mainly from
clay (Helling et al., 1964). Because binding of Ni to clay is
much weaker than to organic matter (Weng et al., 2004), the
adsorption of Ni to these two soils normalised for their CEC
would be lower than the other soils. This would explain the
smaller EC
x
than expected from the regression line shown in
Fig. 4.
Other soil properties that significantly explained the vari-
ance for the barley and tomato EC
x
values include clay content
and exchangeable Mg concentration (for the barley EC
x
only;
Table 2). Soil pH explained only 27% of the variance in the
tomato EC
x
(p<0.05), and was not a significant predictor
for the barley EC
x
. In contrast, neither organic C content nor
(b)
Soil No. (Increasing pH)
EC50 for added Ni (mg kg-1)
10
100
1000
10000
(a)
EC10 for added Ni (mg kg-1)
1
10
100
1000
Barley
Tomato
Barley
Tomato
12345678910111213141516
Fig. 3. Variation among 16 European soils in the EC
10
(a) and EC
50
(b) for the
total concentration of Ni added to the soils. Vertical bars represent 95% con-
fidence interval. See Table 1 for soil number.
Table 2
Single regression equations between the EC
x
of total added Ni and soil properties (n¼16)
Dependent variable Independent variable Regression equation r2
adj (%) Significance level ( p)
Barley root length
EC
50
CEC log EC
50
¼1.60 þ0.89 log (CEC) 90.4 <0.001
EC
50
Exch Ca log EC
50
¼2.03 þ0.57 log (Exch Ca) 88.6 <0.001
EC
50
Clay log EC
50
¼1.92 þ0.54 log (Clay) 68.4 <0.001
EC
50
Exch Mg log EC
50
¼2.58 þ0.47 log (Exch Mg) 52.1 <0.001
EC
10
CEC log EC
10
¼1.35 þ0.87 log (CEC) 84.9 <0.001
EC
10
Exch Ca log EC
10
¼1.78 þ0.56 log (Exch Ca) 81.8 <0.001
EC
10
Clay log EC
10
¼1.68 þ0.52 log (Clay) 60.6 <0.001
EC
10
Exch Mg log EC
10
¼2.31 þ0.46 log (Exch Mg) 47.6 <0.01
Tomato shoot biomass
EC
50
CEC log EC
50
¼1.06 þ1.04 log (CEC) 70.2 <0.001
EC
50
Exch Ca log EC
50
¼1.56 þ0.68 log (Exch Ca) 70.4 <0.001
EC
50
Clay log EC
50
¼1.47 þ0.61 log (Clay) 47.1 <0.01
EC
50
pH log EC
50
¼0.98 þ0.20 pH 28.4 <0.05
EC
10
CEC log EC
10
¼0.93 þ0.99 log (CEC) 55.9 <0.001
EC
10
Exch Ca log EC
10
¼1.41 þ0.64 log (Exch Ca) 60.2 <0.001
EC
10
Clay log EC
10
¼1.35 þ0.56 log (Clay) 36.3 <0.01
EC
10
pH log EC
10
¼0.77 þ0.20 pH 28.7 <0.05
(a)
ECx for added Ni (mg kg-1)ECx for added Ni (mg kg-1)
10
100
1000
EC10
EC50
EC10
EC50
(b)
Soil CEC (mmol kg-1)
1
1
10 100
10
100
1000
Fig. 4. Relationship between soil CEC and the EC
10
and EC
50
for the barley
root length (a) and tomato shoot growth (b) test.
600 C.P. Rooney et al. / Environmental Pollution 145 (2007) 596e605
Fe oxide concentration alone was a significant predictor of the
EC
x
values (data not shown).
Stepwise multiple regression analysis was performed to test
whether inclusion of other soil properties further improved the
prediction of Ni toxicity thresholds. The results showed that
the improvement was relatively small: for the barley root
length test, r2
adj increased by 3e5% units for EC
10
and EC
50
with the addition of Al oxide concentration in addition to
CEC. A second term did not consistently improve the predic-
tions for tomato EC
x
(data not shown).
3.3. Solubility and speciation of Ni
At each Ni dose, the concentration of Ni in soil solution
varied by 2e3 orders of magnitude among the 16 soils tested
(data not shown). Nickel was most soluble in the acidic and
sandy Houthalen and Jyndevad soils; the lowest solubility
was observed in six calcareous soils (Souli, Marknesse, Brecy,
Cordoba 1, Cordoba 2, Guadalajara). Free Ni
2þ
activity varied
by over 6 orders of magnitude, from 1.9 to 8.0 pNi
2þ
(data not
shown). In most soil solution samples, Ni bound to dissolved
organic matter (DOM) generally accounted for <50% (mostly
<10%) of soluble Ni.
Free Ni
2þ
activity (pNi
2þ
) in soil solution calculated by the
extended pH-dependent Freundlich equation (Eq. (1)) agreed
well with that calculated by the WHAM VI model (Table 3
and Fig. 5). Both soil solution pH and ionic strength were
highly significant parameters for the optimisation of the pH-
dependent Freundlich equation. The agreement was further
improved when [Ni
Solid
] was expressed on the basis of soil or-
ganic C or CEC, compared to when [Ni
Solid
] was expressed on
the basis of whole soil. In particular, [Ni
Solid
] based on soil
CEC gave the best prediction of pNi
2þ
(Fig. 5), with the dif-
ference between the pNi
2þ
calculated by the WHAM VI model
and that predicted by the extended pH-dependent Freundlich
equation being smaller than 0.9 of a unit in all Ni treatments
of all 16 soils.
3.4. Ni toxicity threshold values based on soil solution Ni
concentration and free Ni
2þ
activity
The EC
10
and EC
50
values were derived for both soil solu-
tion Ni concentration and free Ni
2þ
activity from the fitted
logelogistic curves (Eq. (2)) relating barley root length or
tomato shoot biomass to these Ni measurements. The fitted
curves described the doseeresponse data satisfactorily, ex-
plaining 90.1e99.5% and 85.6e99.4% of the variance in the
case of soil solution Ni and free Ni
2þ
activity, respectively.
The EC
50
based on soil solution Ni concentration ranged
from 9.5 to 65.3 mg L
1
and from 1.2 to 15.8 mg L
1
for
the barley and tomato tests, respectively, corresponding to
a 7- and 14-fold variation among the 16 soils (Fig. 6). For
the EC
10
, variation among soils was 21- and 41-fold for the
barley root length and tomato growth tests, respectively.
The EC
50
values based on free Ni
2þ
activity ranged from
3.50 10
5
to 3.54 10
4
M (pNi
2þ
3.5e4.5) for barley
root length and from 9.25 10
6
to 7.38 10
5
M (pNi
2þ
4.1e5.0) for tomato shoot biomass (Fig. 7), representing
ranges between soils of 10- and 8-fold, respectively. There-
fore, variability in the EC
50
for free Ni
2þ
activity among soils
was similar in magnitude to that for soil solution Ni, and
smaller than that for total added Ni. The EC
50
values based
on both the concentration of soil solution Ni and free Ni
2þ
ac-
tivity were lower for the tomato test than for the barley root
elongation test (Figs. 6 and 7).
3.5. Relationships between toxicity threshold values for
soil solution Ni or free Ni
2þ
and soil properties
No consistent relationships were found between the tomato
EC
50
or EC
10
based on soil solution Ni concentration or free
Ni
2þ
activity and soil properties. For the barley root length
test, either soil pH or soil solution pH at the EC
x
doses ex-
plained a significant percentage of the variation among soils
(Eqs. (3e6) for soil pH):
EC
x
based on soil solution Ni concentration:
log EC50 ¼2:05 0:11 Soil pH r2
adj ¼0:49;p¼0:001 ð3Þ
log EC10 ¼1:68 0:15 Soil pH r2
adj ¼0:38;p¼0:006 ð4Þ
EC
x
based on soil solution free Ni
2þ
activity:
log EC50 ¼2:85 0:17 Soil pH r2
adj ¼0:74;p<0:001 ð5Þ
log EC10 ¼3:08 0:23 Soil pH r2
adj ¼0:50;p¼0:002 ð6Þ
Note that the coefficient for the soil pH term was negative,
indicating that soluble or free Ni
2þ
in the soil solution became
more toxic with increasing soil (or soil solution) pH. Fig. 8
shows the relationship between barley and tomato EC
50
based
on free Ni
2þ
activity and soil pH.
4. Discussion
Tomato shoot growth was generally more sensitive to Ni
toxicity than barley root elongation. This difference was oppo-
site to that observed for Cu in a study using a similar set of
soils (Rooney et al., 2006), which showed that barley root
elongation was more sensitive to Cu than tomato shoot growth.
The difference between the two metals may be because roots
are particularly sensitive to Cu toxicity (De Vos et al., 1991;
Arduini et al., 1995), and/or Ni is more readily transported
from roots to shoots than Cu (Yang et al., 1996). Despite these
differences, the EC
50
or EC
10
values for total added Ni derived
Table 3
Coefficients and residual standard deviation (RSD) for the Ni solubility model
based on a pH-dependent Freundlich equation: log½NiSolid=Ni2þnF¼
k1þk2pH þk3log½I
[Ni
Solid
] based on n
F
k
1
k
2
k
3
RSD r2
adj
Whole soil 1.36 3.51 0.64 2.27 0.57 0.85
Organic C 1.25 2.72 0.82 1.91 0.48 0.89
CEC 0.77 0.93 0.26 0.84 0.35 0.94
601C.P. Rooney et al. / Environmental Pollution 145 (2007) 596e605
from the two plant tests correlated closely among the 16 soils,
suggesting that, although sensitivity differs between the two
tests, soil factors affected the bioavailability of Ni to the two
plants in a similar way.
The EC
x
values for total added Ni varied widely among the
16 soils tested. Soil CEC was found to be the best single pre-
dictor for toxicity values from both plant tests (Table 2 and
Fig. 4). Similarly, studies on the soil factors affecting Ni tox-
icity to soil microbial processes also reported a strong relation-
ship with soil CEC (Doelman and Haanstra, 1984; Oorts et al.,
2006b). The positive correlation between the EC
x
and CEC
suggests that sorption of Ni by soil solid phase plays a key
role in controlling its bioavailability, and hence toxicity, to
plants. When a pH-dependent Freundlich equation was fitted
to the relationship between free Ni
2þ
activity and the quantity
of sorbed Ni, it was found that a normalisation of sorbed Ni on
the basis of CEC gave the best fit (Fig. 5), again indicating the
role of CEC in controlling Ni solubility. Although CEC is
a measure of the capacity for exchangeable cations, it may
correlate closely with the amount of metal binding sites. The
EC
x
for both plant tests also correlated with clay content,
but surprisingly, not with organic C content. This may be be-
cause that the capacity of metal sorption on soil organic matter
is dependent on pH. In this study, CEC was measured at soil
pH using the unbuffered silver thiourea method (Chhabra
et al., 1975). This method gives lower CEC values for soils
with high organic matter and lower pH than the method using
ammonium acetate buffered at pH 7.0. The relationship be-
tween CEC, measured by the unbuffered method, and organic
C content is highly dependent on soil pH (Chhabra et al.,
1975). CEC measured by unbuffered methods can be consid-
ered to be the effective CEC (as compared to potential CEC
(a)
pNi2+ calculated by WHAM VI
123456789
pNi2+ calculated by WHAM VI
123456789
pNi2+ calculated by WHAM VI
123456789
Predicted pNi2+ using
solubility model
1
2
3
4
5
6
7
8
9(b) (c)
Fig. 5. Prediction of free Ni
2þ
activity from a pH-dependent Freundlich equation assuming that [Ni
Solid
] is based on the whole soil (a), soil organic C (b) or soil
CEC (c). See Table 3 for the coefficients of the equation.
(b)
Soil No. (Increasing pH)
12345678910111213141516
EC50 for solution Ni (mg L-1)EC
10 for solution Ni (mg L-1)
0.1
1
10
100
Barley
Tomato
(a)
0.1
1
10
100
Barley
Tomato
Fig. 6. Variation among 16 European soils in the EC
10
(a) and EC
50
(b) for the
soil solution concentration of Ni. Vertical bars represent 95% confidence
interval.
(b)
EC50 for free Ni2+ activity (M)
Barley
Tomato
(a)
EC10 for free Ni2+ activity (M)
10-3
10-4
10-3
10-4
10-5
10-6
10-5
10-6
10-7
Barley
Tomato
Soil No. (Increasing pH)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Fig. 7. Variation among 16 European soils in the EC
10
(a) and EC
50
(b) for the
soil solution free Ni
2þ
activity. Vertical bars represent 95% confidence
interval.
602 C.P. Rooney et al. / Environmental Pollution 145 (2007) 596e605
measured at a buffered pH). It has been argued that unbuffered
methods are preferable for studies of environmental problems
related to soils where information on the CEC at the pH of the
soil in the field is of prime importance (Hendershot et al.,
1993). Since CEC, exchangeable Ca concentration and clay
content correlated closely with each other (r¼0.90e0.93,
p<0.001), their close correlation with EC
x
probably points
to the same mechanism, i.e. sorption of Ni as a controlling fac-
tor of Ni toxicity to plants. There were only small improve-
ments in EC
x
predictions when multiple soil properties were
considered, supporting the use of soil CEC for normalising
toxicity threshold values based on the total concentration of
added Ni.
The difference among the 16 soils in the EC
x
values of the
total concentration of added Ni may be due to the large differ-
ence in Ni solubility or speciation in the soil solution. Indeed,
the variation among soils was substantially reduced by consid-
ering EC
x
on the basis of soluble Ni or free Ni
2þ
in soil solu-
tion. Variation in the EC
50
was 37- and 54-fold based on total
added Ni (barley root length and tomato shoot biomass tests,
respectively); 7- and 14-fold for soil solution Ni; and 10-
and 8-fold for free Ni
2þ
in soil solution. In this regard, Ni
and Cu behaved differently, as soil solution Cu or free Cu
2þ
did not narrow the soil to soil variation in the EC
x
to plants
(Zhao et al., 2006). The fact that the extent of the EC
x
varia-
tion among soils was similar for soluble Ni and free Ni
2þ
sug-
gests that solubility was more important than solution phase
speciation in controlling Ni toxicity to plants. This is perhaps
not surprising considering that complexation of Ni by dis-
solved organic matter was not important in the majority of
the solution samples in the present study, accounting for
<10% of the total dissolved Ni.
Weng et al. (2003, 2004) showed that soil pH was the most
important factor controlling the bioavailability of Ni to plants.
In their studies, either one or three soils were amended with
various amounts of calcium hydroxide to obtain different pH
values. The EC
50
value for total added Ni, derived from an
oat growth test, increased markedly with increasing pH in
each soil. In our study, there was no significant correlation be-
tween soil pH and the barley EC
x
values based on total added
Ni; a significant but not very strong relationship was found for
the tomato test (Table 2). Our study differed from those of
Weng et al. (2003, 2004) in that the variation in pH was not
created by the addition of calcium hydroxide, and that, in ad-
dition to pH, other soil properties also varied widely among
the soils used. Our results showed that the pH effect on Ni bio-
availability was much less consistent when a wide range of
soils were used, probably because other soil properties (e.g.
CEC) were more influential. Another explanation for a lack
of a strong pH effect is given below.
Soil pH can have two contrasting effects on metal bioavail-
ability to soil organisms. On the one hand, protons compete
with metal ions for soil sorption sites, thus increasing metal
solubility and free ion activity. This was clearly demonstrated
by the effect of pH on Ni solubility as shown in Table 3 and
Fig. 5. On the other hand, protons also compete with metal
ions for the biotic ligands on the cell surfaces of organisms,
thus reducing metal uptake or toxicity (Plette et al., 1999;
Weng et al., 2003). When toxicity threshold values are based
on free metal ion activity, an inverse relation between the
EC
x
and pH is evidence of the protective effect of protons
(Lofts et al., 2004). Such inverse relationships have been
shown for Cu, Pb, Zn, Cd and Ni for different ecotoxicological
endpoints (Weng et al., 2003; Lofts et al., 2004; Zhao et al.,
2006). In the present study, a significant inverse relationship
between the EC
x
for free Ni
2þ
and pH was obtained for the
barley root elongation test (Fig. 8). The slopes of approxi-
mately 0.2 (Eqs. (5) and (6)) were similar to those obtained
for Ni derived from three soil microbial assays (Oorts et al.,
2006b). The slopes for Ni were smaller in the absolute term
than those for other metals (Cu, Pb, Zn and Cd) (Lofts
et al., 2004; Oorts et al., 2006b; Zhao et al., 2006), suggesting
that the protective effect of protons was smaller for Ni than for
other metals. The effect of pH on the EC
x
of free Ni
2þ
activity
appeared to have offset the pH effect on Ni solubility, resulting
in an overall insignificant effect of pH on the EC
x
of total
added Ni for the barley root elongation test. A similar obser-
vation was reported for Ni toxicity to microbial processes in
the same 16 soils (Oorts et al., 2006b). However, in the case
of the tomato growth test, pH did not significantly influence
the EC
x
based on free Ni
2þ
activity (Fig. 8), and the overall
effect of pH on Ni toxicity reflects its effect on Ni solubility.
Our toxicity tests were based on standard methods (Interna-
tional Organisation for Standardisation (ISO), 1993)(Interna-
tional Organisation for Standardisation (ISO), 1995). A
criticism of these methods is that they probably overestimate
metal toxicity because of the fresh additions of soluble metal
salts (Smolders et al., 2003; Stevens et al., 2003; Oorts et al.,
2006a). Additions of soluble metal salts increase ionic strength
in soil solution and can also decrease soil solution pH, result-
ing in a larger metal solubility. This effect leads to a lower EC
x
value based on total added metal, but does not influence EC
x
based on soluble metal concentration or free metal activity
in soil solution (Stevens et al., 2003). Additions of large
amounts of soluble metal salt could also cause a direct toxic
effect of salinity on plant growth (Stevens et al., 2003). This
effect was unlikely to occur in the barley test because barley
Soil pH
345678
EC50 for free Ni2+ activity (M)
10-3
10-4
10-5
10-6
Barley
Tomato
Fig. 8. Relationship between barley and tomato EC
50
based on free Ni
2þ
activity and soil pH.
603C.P. Rooney et al. / Environmental Pollution 145 (2007) 596e605
is relatively tolerant to salinity, and the electrical conductivi-
ties in the soil solutions at the EC
50
dose of Ni addition
were all below the threshold (8 dS m
1
) for this plant species
(Australian and New Zealand Environment and Conservation
Council eAgriculture and Resource Management Council
of Australia and New Zealand, 2000). In the case of the tomato
test, a direct salinity effect could occur at the EC
50
dose of Ni
addition in four soils, in which the electrical conductivity
exceeded the threshold value of 2.3 dS m
1
. Metal toxicity
and the potential salinity effect generally decrease in field
aged soils, due to leaching of excess solutes in the soil solution
and the shift of metals from more soluble to less soluble phases.
This ‘‘aging’’ effect has been assessed by comparing phytotox-
icity in soils receiving freshly added Ni (as in the present study)
and in those left to leach and equilibrate outdoor after Ni addi-
tions. The results will be presented separately.
In conclusion, the present study has shown that bioavail-
ability of Ni to plants varied widely among 16 soils that rep-
resent major soil types in Europe. This variation was
substantially reduced when toxicity threshold values were
based on soil solution Ni or free Ni
2þ
activity. However, using
soil solution or free Ni
2þ
measurements would not be appro-
priate for site-specific risk assessment, as variability among
soils was not adequately explained, and because these mea-
surements are not routinely performed. When a soluble Ni
salt was added to soils and allowed to equilibrate for a short
period, sorption of Ni by the soil solid phase appeared to be
the key factor controlling Ni bioavailability and toxicity to
plants, and the toxicity threshold values correlated closely
with soil CEC. We propose that CEC, as measured by unbuf-
fered methods, can be used to normalise Ni toxicity data ex-
pressed as total added Ni to account for soil to soil
variation, allowing a more realistic and site-specific risk as-
sessment of Ni in the terrestrial environment.
Acknowledgements
Funding for this work was provided by the Nickel Pro-
ducers Environmental Research Association. We thank
E. Smolders and K. Oorts, Laboratory for Soil and Water Man-
agement, K.U. Leuven, Belgium, for soil collection, process-
ing and analysis of soil properties. Rothamsted Research
receives grant-aided support from the Biotechnology and Bio-
logical Sciences Research Council of the United Kingdom.
References
Arduini, I., Godbold, D.L., Onnis, A., 1995. Influence of copper on root
growth and morphology of Pinus pinea L. and Pinus pinaster Ait. seed-
lings. Tree Physiology 15, 411e415.
Australian and New Zealand Environment and Conservation Council e
Agriculture and Resource Management Council of Australia and
New Zealand, 2000. Australian and New Zealand Guidelines for Fresh
and Marine Water Quality. In: National Water Quality Management Strat-
egy 4. <http://www.deh.gov.au/water/quality/nwqms/pubs/volume3-9-2.pdf>.
(Canberra, Australia).
Chhabra, R.J., Pleysier, J., Cremers, A.C., 1975. The measurement of the cat-
ion exchange capacity and exchangeable cations in soils: a new method.
In: Proceedings of International Clay Conference. Applied Publishing,
Wilmette, USA/Mexico City, Mexico, pp. 439e449.
De Vos, C.H.R., Schat, H., De Waal, M.A.M., Vooijs, R., Ernst, W.H.O., 1991.
Increased resistance to copper induced damage of the root cell plasma-
lemma in copper tolerant Silene cucubalus. Physiologia Plantarum 82,
523e528.
Dixon, R.K., 1988. Response of ectomycorrhizal Quercus rubra to soil cad-
mium, nickel and lead. Soil Biology and Biochemistry 20, 555e559.
Doelman, P., Haanstra, L., 1984. Short-term and long-term effects of cad-
mium, chromium, copper, nickel, lead and zinc on soil microbial respira-
tion in relation to abiotic soil factors. Plant and Soil 79, 317e327.
European Commission, 2003. Technical Guidance Document on Risk Assess-
ment. European Commission Joint Research Centre, Ispra, Italy.
Haanstra, L., Doelman, P., Voshaar, J.H.O., 1985. The use of sigmoidal dose
response curves in soil ecotoxicological research. Plant and Soil 84,
293e297.
Helling, C.S., Chesters, H.G., Corey, R.B., 1964. Contribution of organic mat-
ter and clay to soil cation-exchange capacity as affected by the pH of the
saturating solution. Soil Science Society of America Proceedings 28,
517e520.
Hendershot, W.H., Lalande, H., Duquette, M., 1993. Ion exchange and ex-
changeable cations. In: Carter, M.R. (Ed.), Soil Sampling and Methods
of Analysis. Lewis Publishers, Boca Raton, USA, pp. 167e176.
International Organisation for Standardisation (ISO), 1993. Soil Quality e
Determination of the Effects of Pollutants on Soil Flora ePart 1: Method
for the Measurement of Inhibition of Root Growth. No. 11269e1. Interna-
tional Organization for Standardization.
International Organisation for Standardisation (ISO), 1995. Soil Quality e
Determination of the Effects of Pollutants on Soil Flora ePart 2:
Effects of Chemicals on the Emergence and Growth of Higher Plants.
No. 11269e2. International Organization for Standardization.
Khalid, B.Y., Tinsley, J., 1980. Some effects of nickel toxicity on rye grass.
Plant and Soil 55, 139e144.
Lofts, S., Spurgeon, D.J., Svendsen, C., Tipping, E., 2004. Deriving soil crit-
ical limits for Cu, Zn, Cd, and Pb: a method based on free ion concentra-
tions. Environmental Science and Technology 38, 3623e3631.
McGrath, S.P., Cunliffe, C.H., 1985. A simplified method for the extraction of
the metals Fe, Zn, Cu, Ni, Cd, Pb, Cr, Co and Mn from soils and sewage
sludges. Journal of the Science of Food and Agriculture 36, 794e798.
Oorts, K., Bronckaers, H., Smolders, E., 2006a. Discrepancy of the microbial
response to elevated copper between freshly spiked and long-term contam-
inated soils. Environmental Toxicology and Chemistry 25, 845e853.
Oorts, K., Ghesquie
`re, U., Swinnen, K., Smolders, E., 2006b. Soil properties
affecting the toxicity of CuCl
2
and NiCl
2
for soil microbial processes in
freshly spiked soils. Environmental Toxicology and Chemistry 25,
836e844.
Plette, A.C.C., Nederlof, M.M., Temminghoff, E.J.M., van Riemsdijk, W.H.,
1999. Bioavailability of heavy metals in terrestrial and aquatic systems:
a quantitative approach. Environmental Toxicology and Chemistry 18,
1882e1890.
Rooney, C.P., Zhao, F.J., McGrath, S.P., 2006. Soil factors controlling the ex-
pression of copper toxicity to plants in a wide range of European soils. En-
vironmental Toxicology and Chemistry 25, 726e732.
Sauerbeck, D.R., 1991. Plant, element and soil properties governing uptake
and availability of heavy metals derived from sewage sludge. Water Air
and Soil Pollution 57, 227e237.
Sauerbeck, D.R., Hein, A., 1991a. The nickel uptake from different soils and
its prediction by chemical extractions. Water Air and Soil Pollution
57e58, 861e871.
Sauerbeck, D.R., Hein, A., 1991b. Nickel uptake from different soils and its
prediction by chemical extractions. Water, Air, and Soil Pollution
57e58, 861e871.
Schwertman, U., 1964. Differenzierung der eisenoxide des bodens durch ex-
traction mit ammoniumoxalat lo
¨sung. Zeitschrift fu
¨r Pflanzenerna
¨hrung
und Bodenkunde 105, 194e202.
Smolders, E., McGrath, S.P., Lombi, E., Karman, C.C., Bernhard, R.,
Cools, D., Van Den Brande, K., Van Os, B., Walrave, N., 2003.
Comparison of toxicity of zinc for soil microbial processes between
604 C.P. Rooney et al. / Environmental Pollution 145 (2007) 596e605
laboratory-contamined and polluted field soils. Environmental Toxicology
and Chemistry 22, 2592e2598.
Stevens, D.P., McLaughlin, M.J., Heinrich, T., 2003. Determining toxicity of
lead and zinc runoff in soils: salinity effects on metal partitioning and on
phytotoxicity. Environmental Toxicology and Chemistry 22, 3017e3024.
Tipping, E., 1998. Humic ion-binding model VI: An improved description of
the interactions of protons and metal ions with humic substances. Aquatic
Geochemistry 4, 3e48.
Tipping, E., Rieuwerts, J., Pan, G., Ashmore, M.R., Lofts, S., Hill, M.T.R.,
Farago, M.E., Thornton, I., 2003. The solidesolution partitioning of heavy
metals (Cu, Zn, Cd, Pb) in upland soils of England and Wales. Environ-
mental Pollution 125, 213e225.
Tye, A.M., Young, S., Crout, N.M.J., Zhang, H., Preston, S., Zhao, F.J.,
McGrath, S.P., 2004. Speciation and solubility of Cu, Ni and Pb in contam-
inated soils. European Journal of Soil Science 55, 579e590.
Wallace, A., Romney, E.M., Cha, J.W., Soufi, S.M., 1977a. Nickel phytotox-
icity in relationship to soil pH manipulation and chelating agents. Commu-
nications in Soil Science and Plant Analysis 8, 757e764.
Wallace, A., Romney, R.M., Cha, J.W., Soufi, S.M., Chaudhry, F.M., 1977b.
Nickel phytotoxicity in plants. Communications in Soil Science and Plant
Analysis 8, 757e764.
Weng, L.P., Lexmond, T.M., Wolthoorn, A., Temminghoff, E.J.M., Van
Riemsdijk, W.H., 2003. Phytotoxicity and bioavailability of nickel: chem-
ical speciation and bioaccumulation. Environmental Toxicology and
Chemistry 22, 2180e2187.
Weng, L.P., Wolthoorn, A., Lexmond, T.M., Temminghoff, E.J.M., Van
Riemsdijk, W.H., 2004. Understanding the effects of soil characteristics
on phytotoxicity and bioavailability of nickel using speciation models.
Environmental Science and Technology 38, 156e162.
Yang, X., Baligar, V.C., Martens, D.C., Clark, R.B., 1996. Plant tolerance to
nickel toxicity. 2. Nickel effects on influx and transport of mineral nutri-
ents in 4 plant species. Journal of Plant Nutrition 19, 265e279.
Zhao, F.J., Rooney, C.P., Zhang, H., McGratht, S.P., 2006. Comparison of soil
solution speciation and diffusive gradients in thin-films measurement as an
indicator of copper bioavailability to plants. Environmental Toxicology
and Chemistry 25, 733e742.
605C.P. Rooney et al. / Environmental Pollution 145 (2007) 596e605