Soil solution extraction techniques for microbial ecotoxicity testing: a comparative evaluation

Abstract

The suitability of two different techniques (centrifugation and Rhizon sampler) for obtaining the interstitial pore water of soil (soil solution), integral to the ecotoxicity assessment of metal contaminated soil, were investigated by combining chemical analyses and a luminescence-based microbial biosensor. Two different techniques, centrifugation and Rhizon sampler, were used to extract the soil solution from Insch (a loamy sand) and Boyndie (a sandy loam) soils, which had been amended with different concentrations of Zn and Cd. The concentrations of dissolved organic carbon (DOC), major anions (F⁻, Cl⁻, NO₃⁻, SO₄²⁻) and major cations (K⁺, Mg²⁺, Ca²⁺) in the soil solutions varied depending on the extraction technique used. Overall, the concentrations of Zn and Cd were significantly higher in the soil solution extracted using the centrifugation technique compared with that extracted using the Rhizon sampler technique. Furthermore, the differences observed between the two extraction techniques depended on the type of soil from which the solution was being extracted. The luminescence-based biosensor Escherichia coli HB101 pUCD607 was shown to respond to the free metal concentrations in the soil solutions and showed that different toxicities were associated with each soil, depending on the technique used to extract the soil solution. This study highlights the need to characterise the type of extraction technique used to obtain the soil solution for ecotoxicity testing in order that a representative ecotoxicity assessment can be carried out.

---

Introduction

The assessment of soil quality is not possible based solely on chemical criteria. It is now widely accepted that a risk assessment for soils should be based on the total content of pollutants and their mobile and bioavailable fractions.^1,2 It is therefore essential that chemical and ecotoxicological analyses are integrated in order that a more meaningful assessment of soil quality can be obtained.

The extraction of the interstitial pore water of soil (soil solution) is an important step in an ecotoxicity assessment of soil. Soil solution chemistry is important for understanding the behaviour and fate of environmental pollutants, the availability of nutrients to plants and many other fundamental soil chemical processes. The soil solution may be defined as the aqueous liquid phase of the soil and its solutes. Under natural conditions, the soil solution composition varies as a function of time and position in the soil.³ Furthermore, the chemical composition of the soil solution also depends on the soil type, pH, ionic strength, ion complexation by ligands and ion competition.⁴ Several methods have been used to obtain soil solutions either in the field or in the laboratory.^5,6 The methods commonly used to obtain soil solutions are based upon the principles of pressure, vacuum, displacement, and centrifugation, and are summarised in Table 1. However, there is a body of evidence showing that the chemical composition of soil solutions is influenced by the extraction technique used,^20,25,30 although how this may impact on an ecotoxicity assessment of soil has not yet been investigated.

Table 1 Summary of the techniques commonly used for extracting soil solutions

Method	Summary	Reference
Column displacement/miscible displacement	Requires a large sample volume	7, 8
	Slow sample turn-around
	Not suitable for soils with a sandy and loamy sand texture
	Special skill is required in packing the column
Vacuum displacement	Modification of the column displacement method	9, 10
	Involves the use of a vacuum extractor to obtain soil solution
	Lower time requirement and faster sample turn-around
Immiscible displacement	Displacing solution is an organic solvent	11–13
	Alters the chemical characteristics and speciation of compounds in the soil solution
	Displacing solutions are highly toxic to soil flora and fauna
	Unsuitable for ecotoxicity testing
Centrifugation	Requires a sample holding device that enables the solution to be isolated from the soil	5, 14–20
	Reliable and rapid technique
	Low cost
	Easy to use and requires no special skill
	Non-destructive method
	Can be applied in conjuction with the displacement method
Suction technique	Suction cup—	6, 21–25
	Requires a cup made from a porous material (commonly ceramic, aludum or Teflon)
	Non-destructive method
	Preconditioning of the surface of the suction cup is required to avoid strong sorption effects
	High cost of some items (e.g., Teflon cups)
	Porous plastic tube—
	Inert nature and small pore size (<0.2 µm) of plastic tubing ensures no sorption problems or microbial and colloidal contamination	^26–29
	Non-destructive method
	Low cost

---

Assessing the bioavailable fraction of the pollutant in an ecotoxicity test requires the application of rapid, low cost, novel techniques. The bioavailable fraction of metals in soils has been attributed to the free ionic species in the soil solution.^31–34 Luminescence-based microbial biosensors have been shown to respond to the bioavailable fraction of metal pollutants in soil solutions,^31,35,36 and are therefore useful tools in a soil ecotoxicity assessment. In a luminescence-based assay, bacterial luminescence is negatively correlated with an increase in the toxicity of a pollutant since bacterial luminescence is linked to electron transport³⁷ and therefore provides a measure of the metabolic activity of the cell.

The aims of this study were to evaluate two widely used soil solution extraction techniques (centrifugation and Rhizon sampler) on the ecotoxicity assessment of two soils amended with Zn and Cd at different concentrations.

Materials and methods

Soils

Two soils were used for this study, Boyndie, a loamy sand of the Boyndie series (Fragiorthod/Iron Polozol) and Insch, a sandy loam of the Insch series (Dystrochrept/Dystric Cambisol). These soils were sieved moist (<3 mm), homogenised and then stored at 3–5 °C prior to use. The characteristics of these soils are summarised in Table 2. Soil pH was determined using a 1 ∶ 2 soil to soil water ratio.

Table 2 Soil characteristics

	Boyndie	Insch
a WHC, water holding capacity.
Series	Boyndie	Insch
Texture	Loamy sand	Sandy loam
Sand (%)	79.1	57.7
Silt (%)	14.3	30.79
Clay (%)	6.53	11.51
Organic matter (%)	3.57	3.75
pH	5.5	6.1
WHC^a (%)	44	69

---

Metal amendment of soils

Triplicate samples of each soil were amended with Zn(NO₃)₂·6H₂O (Sigma, Dorset, UK) and Cd(NO₃)₂·4H₂O (Aldrich, Dorset, UK) to obtain total metal concentrations of 75, 150 and 300 mg kg⁻¹ dry weight for Zn and 1.5, 3 and 6 mg kg⁻¹ dry weight soil for Cd. These concentrations represented 50%, 100% and 200% of the maximum metal concentrations currently permitted under the UK sewage sludge guidelines.³⁸ Control soils were those which were not amended with metals. The soils were raised to 50% of their water holding capacity (WHC) with distilled water, placed in plastic pots, and left to equilibrate for 30 days.

Soil solution extraction techniques

All glassware and plastic was washed in 1% HNO₃ and rinsed twice in distilled water prior to use. All soils were raised to 80% WHC with distilled water before the soil solution was extracted. The following techniques were used for the extractions.

Centrifugation. A modification of the centrifugation technique of Adams et al.⁷ was used to collect the soil solution. The extraction unit consisted of a sample container cup, a solution collection cup, a filter, and a plastic bag to collect the solution. A plastic bag (12 cm × 8 cm) was put inside the container cup. A soil sub-sample (150 ± 0.1 g) was placed onto Whatman No 42 filter paper in the sample container cup, which was connected to the collection cup. The soil solution was extracted by centrifuging for 1 h at 4 °C and 1818g in an MSE Coolspin 2 centrifuge. The soil solution was collected into a plastic bag, transferred to a clean glass bottle and stored at 4 °C prior to analysis.

Rhizon sampler. Rhizon soil moisture samplers were obtained from Rhizosphere Research Products (Wageningen, The Netherlands). These samplers consisted of a 10 cm length of an inert porous polymer tube, capped at one end with nylon, and attached to a 10 cm length of a poly(vinyl chloride) (PVC) tube at the other end. A 15 cm stainless-steel strengthening wire is present inside the porous polymer tube and the PVC tube is joined to a female Luer lock. The Rhizon samplers were washed by forcing 50 ml of 1% HNO₃ through the probe followed by 50 ml of distilled water, and then drying before use. A soil sub-sample (225 g dry weight) from each pot was placed into a plastic container and a small glass rod used to make a small primary hole in the soil prior to the Rhizon samplers being gently inserted. The soil solution was extracted over a 24 h period by applying a suction pressure using a syringe connected to the Luer lock. The soil solution collected in the syringe was transferred to a glass bottle and stored at 4 °C prior to analysis.

Chemical analysis and metal speciation

The pH values of the soil solutions were measured immediately using a standard pH electrode (Hanna HI 8424, Norlab Instrument Ltd., Aberdeen, UK). The DOC was measured using an automated Labtoc UV digestion analyser (Pollution and Process Monitoring, Kent, UK). The concentrations of F⁻, Cl⁻, NO₃⁻, SO₄²⁻ were determined by ion chromatography (Dionex Series 4500I, Dionex UK Ltd, Surrey, UK). The separation column used was an Ionpac A54 ASC ion exchange column with a pre-guard column and the mobile phase was 1.7 mmol NaCO₃^–1 mmol NaHCO₃. Sub-samples of the soil solution were acidified with 1% HNO₃ and used to determine the total concentrations of various metals. Total concentrations of Mg, Ca and Zn were determined by flame atomic absorption spectrometry (FAAS; Alpha Models 4, Baird Atomic Ltd, UK), K by flame atomic emission spectrometry (FAES; Alpha Models 4) and Cd by graphite furnace atomic absorption spectrometry (GFAAS; Perkin-Elmer 3300, Perkin-Elmer, Buckinghamshire, UK). The last of these involved the addition of 0.2 mg PO₄³⁻ as a matrix modifier.

The concentrations of Zn²⁺ and Cd²⁺ were determined using the method described by Holm et al.³⁹ The method involved measuring the total concentration of metal before and after equilibrium with a calcium-saturated cation exchange resin. Ground and sieved Amberlite IR120 plus (100–200 mesh) cation exchange resin (Sigma) was used and converted from the Na-form to the Ca-form by the sequential addition of 1000 ml of 1 M Ca(Ac)₂, 0.01 M Ca(Ac)₂, and 0.01 M Ca(NO₃)₂. An aliquot (100 mg) of Amberlite IR120 was weighed into a 50 ml polypropylene centrifuge tube and 10 ml of soil solution added. The centrifuge tube was placed on an orbital shaker for 24 h at 25 °C. Sub-samples of the supernatant were then acidified with 0.3% HNO₃ and the concentration of Zn and Cd remaining in the soil solution determined. A reference solution was prepared with the same final concentrations of Ca, Zn and Cd as the soil solutions and treated in the same way as the soil solutions. The proportions of Zn²⁺ and Cd²⁺ were calculated by comparison with the reference solution as described by Holm et al.³⁹

Soil ecotoxicity assessment

The toxicity of the metals was determined using the luminescence-based microbial biosensor Escherichia coli HB101 pUCD607. E. coli HB101 was marked with the lux CDABE genes, isolated from Vibrio fischeri, using the multicopy plasmid pUCD607.⁴⁰ Cultures of E. coli HB101 pUCD607 were freeze-dried using standard protocols⁴¹ and stored at −20 °C. Freeze-dried cultures were resuscitated by resuspending the cells in 10 ml of 0.1 M KCl and incubating at 25 °C for 1 h with shaking. Aliquots (900 µL) of each soil solution were transferred to luminometer cuvettes (Clinicon, Petworth, West Sussex, UK) and a 100 µL aliquot of the resuscitated cells added and mixed into each sample at 15 s intervals. The luminescence of the samples was measured after a 15 min exposure time using a 1 s measurement on a Bio-Orbit 1253 luminometer (Labtech International, Uckfield, UK). All the assays were carried out in triplicate. The luminescence of the samples was expressed as a percentage of the luminescence of the control samples.

Statistical analyses

Two-way analyses of variance (ANOVA) were carried out on the data using the statistical package Minitab for Windows, release 12.1 (State College, PA, USA). Significant differences between treatments were elucidated using least significance difference (LSD) values. The relationship between the free and the total metal concentrations in the soil solutions was determined using linear regression analyses on log-transformed values. The response of the biosensor to the total and free metal concentrations in the soil solutions was determined using non-linear regression analyses. Sigmoid functions were fitted to the biosensor data and EC₂₅ (effective concentration causing a 25% decrease in luminescence) and EC₅₀ (effective concentration causing a 50% decrease in luminescence) values calculated. All regression analyses were performed using SigmaPlot for Windows version 5 (Jandel Corporation, CA, USA).

Results and discussion

Chemical analyses

The soil solution extracted from the Boyndie soil using the centrifugation technique had a higher pH than the soil solution extracted using the Rhizon sampler technique (Table 3). The opposite was true for the Insch soil, the pH of the soil solutions extracted using the centrifugation technique being lower than those extracted using the Rhizon sampler technique (Table 3). The higher the metal addition applied the lower the pH of the soil solution compared with control soils, irrespective of the extraction technique employed (Table 3). The concentration of DOC was significantly (p < 0.05) higher in soil solutions extracted using the centrifugation technique over those extracted using the Rhizon sampler technique, for both Boyndie and Insch soils (Table 3). Overall, there was a significant (p < 0.05) increase in the concentrations of K, Mg and Ca with an increase in the level of metal added (Table 3). This is not surprising as the Zn and Cd ions in solution would have exchanged with ions such as K, Mg and Ca, which were absorbed onto the surfaces of soil particles. The concentrations of K, Mg and Ca were significantly (p < 0.05) higher in soil solutions extracted using the centrifugation technique compared with those extracted using the Rhizon sampler technique.

Table 3 Soil solution pH and concentrations of DOC, K, Mg, and Ca

Extraction technique	% of metal conc. limit	Boyndie soil solution conc./mg L^−1					Insch soil solution conc./mg L^−1
		pH	DOC	K	Mg	Ca	pH	DOC	K	Mg	Ca
a LSD, least significant difference (p = 0.05). b ND, not determined.
Centrifugation	0	6.49	28.60	0.67	5.60	26.18	5.12	18.15	7.04	7.95	78.84
	50	5.81	17.35	16.46	29.69	89.23	4.85	12.85	17.77	16.40	145.99
	100	5.77	15.15	23.86	43.18	186.81	4.58	13.50	28.64	23.26	202.03
	200	5.50	26.65	36.61	42.43	213.14	4.43	17.53	41.39	36.70	314.35
Rhizon sampler	0	6.26	30.10	1.25	7.14	19.90	5.74	9.53	10.96	4.34	38.50
	50	5.30	10.65	21.68	34.95	112.90	5.48	6.75	24.87	17.33	142.61
	100	5.44	10.50	28.20	42.53	175.70	4.80	8.63	32.12	22.92	193.82
	200	4.48	17.60	40.81	43.52	237.29	4.87	9.35	44.43	36.05	295.99
	LSD^a	ND^b	2.38	1.99	6.74	10.38	ND^b	2.50	1.62	0.93	12.24

---

The concentration of F⁻ in the soil solutions was, overall, not significantly different between the two extraction techniques (Table 4). Overall, the concentration of Cl⁻ was significantly higher in soil solutions extracted using the centrifugation technique compared with those extracted using the Rhizon sampler technique, for both Boyndie and Insch soils (Table 4). Overall there was a significant (p < 0.05) increase in the concentration of NO₃⁻ in the solutions extracted from soils which were amended with high levels of metals using the centrifugation technique compared with those extracted using the Rhizon sampler technique (Table 4). However, the concentrations of NO₃⁻ in the solution extracted from soils amended with no metals were significantly (p < 0.05) higher using the Rhizon sampler technique compared with those extracted using the centrifugation technique (Table 4). The concentrations of SO₄²⁻ in the soil solutions extracted using the Rhizon sampler technique were generally significantly (p < 0.05) higher than those extracted using the centrifugation technique (Table 4). The significant increase in the concentration of nitrate in soil solutions with an increase in the level of metal added reflected the fact that the metals were added as nitrate salts (Table 4).

Table 4 Soil solution anion concentration

Extraction technique	% of metal conc. limit	Boyndie soil solution conc./mg L^−1				Insch soil solution conc./mg L^−1
		F^−	Cl^−	NO3^−	SO4^2−	F^−	Cl^−	NO3^−	SO4^2−
a DL, detection limit. b LSD, least significant difference (p = 0.05). c ND, not determined.
Centrifugation	0	0.11	47.46	163.90	5.86	<DL^a	15.41	182.01	6.15
	50	0.29	41.51	648.08	3.59	<DL^a	18.61	690.15	2.01
	100	0.29	40.88	1082.97	0.94	<DL^a	14.61	909.54	1.62
	200	1.15	41.58	1863.32	2.16	0.36	18.83	1319.13	1.64
Rhizon sampler	0	0.17	44.80	636.90	7.66	<DL^a	12.20	737.33	6.53
	50	0.39	35.94	670.04	3.34	<DL^a	16.23	760.25	2.21
	100	0.34	33.78	1011.07	1.33	<DL^a	11.98	903.89	0.71
	200	0.82	39.12	1292.62	5.65	0.37	14.48	1064.50	7.31
	LSD^b	0.17	3.26	25.20	0.65	ND^c	1.36	94.13	0.60

---

Metals and their speciation

The concentrations of Zn and Zn²⁺ in soil solutions extracted from Boyndie soils using the centrifugation technique were significantly (p < 0.05) higher than those extracted using the Rhizon sampler technique (Table 5). The concentrations of Zn and Zn²⁺ in soil solutions extracted from Insch soils using the centrifugation technique, compared with those extracted using the Rhizon sampler technique, were only significantly (p < 0.05) higher in soil amended with 300 mg kg⁻¹ Zn (200% of the metal concentration limit) (Table 5). The concentrations of Cd and Cd²⁺ in Insch and Boyndie soil solutions extracted using the centrifugation technique were only significantly higher (p < 0.05) than those extracted using the Rhizon sampler technique in soils amended with 6 mg kg⁻¹ Cd (200% of the metal concentration limit) (Table 5).

Table 5 Soil solution concentration of total Zn, Cd, Zn²⁺ and Cd²⁺

Extraction technique	% of metal conc. limit	Boyndie soil solution conc./mg L^−1				Insch soil solution conc./mg L^−1
		Zn	Zn^2+	Cd	Cd^2+	Zn	Zn^2+	Cd	Cd^2+
a DL, detection limit. b LSD, least significant difference (p = 0.05).
Centrifugation	0	<DL^a	<DL^a	<DL^a	<DL^a	<DL^a	<DL^a	<DL^a	<DL^a
	50	13.19	8.61	0.0475	0.0291	3.82	0.78	0.0073	0.0062
	100	26.16	21.21	0.1150	0.0682	9.28	4.68	0.0199	0.0163
	200	74.00	66.42	0.6400	0.4333	45.47	36.37	0.2370	0.1493
Rhizon sampler	0	<DL^a	<DL^a	<DL^a	<DL^a	<DL^a	<DL^a	<DL^a	<DL^a
	50	3.04	1.01	0.0182	0.0114	1.82	0.51	0.0084	0.0026
	100	17.98	8.36	0.0805	0.0463	3.14	0.88	0.0177	0.0091
	200	66.38	48.05	0.4477	0.3520	11.22	4.35	0.0379	0.0167
	^bLSD	5.47	7.16	0.0697	0.0651	8.16	9.67	0.0331	0.0179

---

There was a linear increase in the concentration of Zn²⁺ in the soil solution with an increase in total Zn in the soil solution, irrespective of the extraction technique employed (Fig. 1a). A similar relationship between the Zn²⁺ and total Zn in the soil solutions was observed by Chaudri et al.³⁵ The concentration of Cd²⁺ in the soil solution also increased linearly with an increase in the total Cd in the soil solution, irrespective of the extraction technique used (Fig. 1b).

Fig. 1 Relationship between Zn²⁺ and total Zn (a), and Cd²⁺and total Cd (b) in soil solution extracted using different techniques: ●, centrifugation and ○, Rhizon sampler.

Soil ecotoxicity assessment

The toxicity of aqueous solutions of Zn and Cd to the E. coli HB101 pUCD607 biosensor are of the same order of magnitude (data not shown). The concentrations of Cd in the soil solutions were two orders of magnitude lower than the concentrations of Zn in the soil solutions, therefore the response of the biosensor could be attributed solely to the presence of Zn in the soil solutions (Table 5). Since the biosensor was responding to the concentration of Zn in the soil solutions, irrespective of the extraction technique used, it was possible to combine the toxicity data obtained using both extraction techniques for each soil. Single relationships showing the decrease in the luminescence of the biosensor with an increase in the concentrations of total Zn and Zn²⁺ in the soil solutions were therefore obtained for Boyndie and Insch soils (Figs. 2 and 3, respectively). The toxicities of the soil solutions extracted using the centrifugation technique were generally greater due to the higher concentrations of Zn present (Figs. 2 and 3). The latter was particularly apparent for Insch soil, in which the luminescence of the biosensor did not even drop to half its control level (Fig. 3).

Fig. 2 Relationship between the luminescence of E. coli HB101 pUCD607 (expressed as a percentage of the control) and Zn²⁺ (a) and total Zn (b) in soil solution extracted from Boyndie soil using different techniques: ●, centrifugation and ○, Rhizon sampler.

Fig. 3 Relationship between the luminescence of E. coli HB101 pUCD607 (expressed as a percentage of the control) and Zn²⁺ (a) and total Zn (b) in soil solution extracted from Insch soil using different techniques: ●, centrifugation and ○, Rhizon sampler.

The non-linear regression analyses showed that the concentrations of Zn²⁺ and total Zn in Boyndie soil solution accounted for 81 and 85% of the variance in the luminescence response of the biosensor, respectively (Fig. 2a and 2b). The concentrations of Zn²⁺ and total Zn in Insch soil solution accounted for 96 and 97% of the variance in the luminescence response of the biosensor, respectively (Fig. 3a and 3b). The EC₂₅ and EC₅₀ values for Zn and Zn²⁺ were of the same order of magnitude as those quoted by Chaudri et al.³⁵ (Table 6).

Table 6 Critical Zn values resulting in a decrease in luminescence of the E. coli HB101 pUCD607 biosensor

Soil solution	Zn^2+/mg L^−1		Total Zn/mg L^−1)
	EC25	EC50	EC25	EC50
a NA, not available.
Boyndie	NA^a	11.52	NA^a	18.62
Insch	3.87	5.41	6.11	6.62

---

Evaluation of extraction techniques

An ideal soil solution extraction technique would enable the solution phase to be extracted from the solid phase in sufficient quantities for the desired chemical and biological assessment. There are a number of extraction techniques available and their advantages and disadvantages have been summarised (Table 1). In this study, it was shown that the composition of the soil solution extracted using the centrifugation technique was significantly different from that derived by the Rhizon sampler technique. In assessing a suitable extraction technique the following key attributes were considered important: (a) the technique should enable the extraction of a representative soil solution; (b) sample preparation should be rapid, reproducible and simple to administer; (c) the technique should be reliable and inexpensive; and (d) there should be minimal adsorption of material onto the surfaces of the equipment.

The centrifugation technique resulted in a soil solution with a higher concentration of Zn and Cd than the Rhizon sampler technique. These differences are certainly due to the mechanism involved in extracting the soil solution. In the centrifugation technique a variable matric potential, which is a function of the force of centrifugation, is imposed on the sample to extract the solution from a broad range of pore sizes. However, in the Rhizon sampler technique, a constant matric potential lower than that of centrifugation is used to extract the soil solution. Soil water is removed from the pores and the physical structure of the sample remains intact, hence the desorption/sorption characteristics of the matrix are controlled.

One of the advantages of the Rhizon sampler technique, in comparison to centrifugation, is that it can be successfully applied in the field. Furthermore, the Rhizon sampler technique has been described as mimicking the physical aspects of plant uptake; with the porous plastic rod inserted into the soil acting like a root. The Rhizon sampler technique is particularly applicable where a large number of samples need to be extracted simultaneously since the labour and equipment requirements would be less than for the centrifugation technique. On this basis, the Rhizon sampler technique is a more versatile extraction technique and more suitable for determining the bioavailable fraction of metals in soils.

Conclusion

There is an urgent requirement to assess the efficiency of techniques used to extract interstitial pore waters from soils, both for nutrient and toxicity assessment. In this study it was shown that soil solutions extracted using centrifugation had different chemical and ecotoxicological characteristics to soil solutions extracted using a Rhizon sampler. Furthermore, the trends observed between the two extraction methods differed between soils. It is therefore crucial that the method used to extract the soil solution is carefully considered prior to estimating the level of pollution in soil. The application of ecotoxicity tests to soil solutions can only be adequately interpreted after a comprehensive understanding of the benefits and limitations of the extraction techniques commonly used to obtain them has been achieved.

References

1. U. Wahle and W. Kördel, Chemosphere, 1997, 35, 223.
2. R. Debus and K. Hund, Chemosphere, 1997, 35, 239.
3. D. L. Sparks, Soil Sci. Soc. Am. J., 1984, 48, 514.
4. G. Sposito, The chemistry of soils, Open University, New York, 1989..
5. J. C. Vedy and S. Brucket, in Constituents and properties of soils, Academic Press Inc., New York, 1982, pp. 184–213..
6. J. Grossmann and P. Udluft, J. Soil Sci., 1991, 42, 83.
7. F. Adams, C. Burmester, N. V. Hue and F. L. Long, Soil Sci. Soc. Am. J., 1980, 44, 733.
8. J. W. Wendt, Commun. Soil Sci. Plant Anal., 1992, 23, 764.
9. J. Wolt and J. G. Graveel, Soil Sci. Soc. Am. J., 1986, 50, 602.
10. G. G. S. Holmgren, R. L. June and R. C. Geschwender, Soil Sci. Soc. Am. J., 1977, 41, 1207.
11. A. Mubarak and R. A. Olsen, Soil Sci. Soc. Am. J., 1976, 40, 329.
12. B. R. Whelan and N. J. Barrow, J. Environ. Qual., 1980, 9, 315.
13. D. G. Kinniburgh and D. L. Miles, Environ. Sci. Technol., 1983, 17, 362.
14. G. P. Gillman, Div. Soils Rep. No. 16, CSIRO, Australia, 1976..
15. D. H. Thibault and M. I. Sheppard, Commun. Soil Sci. Plant Anal., 1982, 23, 1638.
16. J. A. Kittrick, Soil Sci. Soc. Am. J., 1983, 47, 1045.
17. D. C. Edmeases, D. M. Wheeler and O. E. Clinton, Aust. J. Soil Res., 1985, 23, 151.
18. E. A. Elkhatib, O. L. Ennett, V. C. Baligar and R. J. Wright, Soil Sci. Soc. Am. J., 1986, 50, 297.
19. E. A. Elkhatib, J. L. Hern and T. E. Staley, Soil Sci. Soc. Am. J., 1987, 5, 578.
20. D. L. Ross and R. J. Bartlett, J. Environ. Qual., 1990, 19, 108.
21. G. Wagner, Soil Sci., 1962, 94, 379.
22. E. A. Hansen and A. R. Harris, Soil Sci. Soc. Am. Proc., 1975, 39, 528.
23. C. N. Smith and R. F. Carsel, Soil Sci. Soc. Am. J., 1986, 50, 263.
24. W. H. Hendershot and F. Courchesne, J. Soil Sci., 1991, 42, 577.
25. C. L. Creasey and S. J. Dreiss, Soil Sci., 1988, 145, 93.
26. R. D. Morrison, Soil Sci., 1982, 134, 206.
27. J. Yanai, S. Araki and K. Kyuma, Soil Sci. Plant Nutr., 1993, 39, 737.
28. D. L. Jones and A. C. Edwards, Plant Soil, 1993, 150, 157.
29. B. P. Knight, A. M. Chaudri, S. P. McGrath and K. E. Giller, Environ. Pollut., 1998, 99, 293.
30. R. A. Dahlgren, Soil Sci. Plant Anal., 1993, 24, 1783.
31. S. P. McGrath, B. Knight, K. Killham, S. Preston and G. I. Paton, Environ. Toxicol. Chem., 1999, 18, 659.
32. B. P. Knight and S. P. McGrath, Environ. Toxicol. Chem., 1995, 14, 2033.
33. R. Petersen, Environ. Sci. Technol., 1982, 16, 443.
34. W. G. Sunda, D. W. Engel and R. M. Thuotte, Environ. Sci. Technol., 1978, 12, 409.
35. A. M. Chaudri, B. P. Knight, V.L. Barbosa-Jefferson, S. Preston, G. I. Paton, K. Killham, N. Coad, F. Nicholson, B. J. Chambers and S. P. McGrath, Environ. Sci. Technol., 1999, 22, 1880.
36. G. I. Paton, E. A. S. Rattray, C. D. Campbell, M. S. Cresser, L. A. Glover, J. C. J. Meeussen and K. Killham, in Bioindicators of Soil Health; Use of genetically modified microbial biosensors for soil ecotoxicity testing, ed. C. S. Pankhurst, B. Doube and V. Gupta, CAB Int., Wallingford, Oxfordshire, 1997..
37. J. W. Hastings, C. J. Potrikus, S. C. Gupta, M. Kurfürst and J. C. Makemson, Adv. Microb. Physiol., 1985, 26, 235.
38. MAFF/DoE, Review of the rules for sewage sludge application to agricultural land. Soil fertility aspects of potentially toxic elements, Report of an independent scientific committee, MAFF Publications, London, 1993..
39. P. E. Holme, T. H. Christensen, J. C. Tjell and S. P. McGrath, J. Environ. Qual., 1995, 24, 183.
40. E. A. S. Rattray, J. I. Prosser, K. Killham and L. A. Glover, Appl. Environ. Microbiol., 1990, 56, 3368.
41. S. Sousa, C. Duffy, H. Weitz, L. A. Glover, E. Bär, R. Henkler and K. Killham, Environ. Toxicol. Chem., 1998, 17, 1039.

---

Footnote

† Presented at the Whistler 2000 Speciation Symposium, Whistler Resort, BC, Canada, June 25–July 1, 2000.

---