Phthalate and alkyl phenol concentrations in soil following applications of inorganic fertiliser or sewage sludge to pasture and potential rates of ingestion by grazing ruminants

Abstract

Soil concentrations of dioctyl phthalate (DOP) and the alkyl phenols, octyl phenol (OP) and nonyl phenol (NP), after repeated surface applications of sewage sludge to pastures, were investigated. Liquid sludge was applied at a rate of 2.25 tonnes dry matter (DM) per hectare to each of three treated (T) plots on three occasions during the summer and two occasions in the early spring over a period of 2.5 years. Control (C) plots were treated with inorganic fertiliser containing amounts of nitrogen equivalent to those applied to the treated plots. At between 69 and 81 days after the application of sludge, 15 separate soil samples were collected from one half of each of the plots (Experiment 1). Concentrations (µg g⁻¹) of DOP were higher (P < 0.001) than those of NP, while those of OP were generally below detectable levels. Mean soil concentrations of DOP were not significantly different in T and C plots [0.233 vs. 0.155 µg g⁻¹; standard error of the difference (SED) = 0.046; not significant (NS)], partly because there was already a relatively large amount of DOP present. NP concentrations were, however, significantly higher in T than in C plots (0.021 vs. 0.013 µg g⁻¹; SED = 0.002; P < 0.05). There was no consistent change over time in the mean soil concentrations of these compounds when sampled at intervals of 3–6 months. Concentrations in soil samples collected at monthly intervals following sludge application indicated that the variation in concentrations of these endocrine-disrupting compounds (EDC) was unrelated to time since sludge application. Rates of soil ingestion, expressed as the percentage of DM intake represented by soil, were higher during the winter than the summer (5.40 vs. 1.17; SED = 0.360; P < 0.001) and estimated daily intakes of DOP and NP were up to 150 µg and 8 µg, respectively. It is concluded that the application of sewage sludge to pasture does not increase soil concentrations of phthalate (as DOP) or alkyl phenols. Thus, the risk of increased exposure to these EDC as a result of sludge application is small. However, the small effect of sludge application on soil concentrations may be largely a reflection of the relatively high concentrations of DOP already present in the soil, which may be biologically significant.

---

Introduction

Following the ban in Europe on the disposal of sewage sludge at sea, increasing amounts of sewage sludge are likely to be applied to farmland, including pasture grazed by ruminants.¹ In countries such as Scotland, in which a relatively high proportion of the agricultural land is pasture, recycling of sludge to pasture may be particularly important. In comparison with normal environmental levels, sludge contains relatively large amounts of both inorganic pollutants, such as heavy metals, and organic pollutants. The latter includes endocrine-disrupting compounds (EDC), such as alkyl phenols, phthalates, polychlorinated biphenyls and organochlorine pesticides. These are potentially hazardous to the health of wildlife, farm animal species and humans.^2,3

The formulation of sustainable practices involving the application of sludges to land requires a knowledge of the patterns of degradation/elimination of pollutants from the soil, and of their accumulation following repeated sludge application. It is generally considered that these organic pollutants represent a relatively low risk to soil fertility, as a result of their long-term accumulation, or to humans, as a result of their entry into the human food chain.⁴ However, this conclusion is based on the known chemical and toxicological properties of these compounds, rather than on field observations of soil concentrations and tissue accumulation in animals. There is already evidence of the significant bioaccumulation of phthalate, in the form of dioctyl phthalate (DOP) in the liver of animals, entering the food chain.⁵ Furthermore, assessments of potential hazards posed by EDC are frequently based on their known toxicity and carcinogenicity and not their endocrine-disrupting effects. Such endocrine-disrupting effects may be exerted at much lower concentrations than those resulting in toxic or carcinogenic effects.⁶

While the effects of sludge application on heavy metal accumulation in soil have long been recognised,⁷ relatively little is known about the patterns of accumulation of different types of EDC in soil following repeated applications or of their rates of loss or degradation. Thus, the relative contribution of sludge to the soil EDC burden and its effect on the magnitude of the associated risk of EDC ingestion by grazing ruminants are unknown. In this study, representatives of two groups of compounds, alkyl phenols [octyl phenol (OP) and nonyl phenol (NP)] and phthalates (DOP), with endocrine-disrupting effects ^6,8,9 and known to be present in relatively large amounts in sewage sludge [approximately 100 µg g⁻¹ dry matter (DM)],^10,11 were investigated.

The risk of accumulation of EDC in the topsoil and the risk of ingestion by grazing ruminants are dependent on the procedure used to apply the sludge to land. The protocol used in this study was considered to be that likely to result in the maximum rate of contamination of the topsoil and the maximum likely risk of exposure of grazing animals to EDC through their food.

The objectives of the study were to: (i) determine the range of soil EDC concentrations present within a typical soil under pasture; (ii) test the hypothesis that sludge application would (a) increase soil concentrations of selected EDC and (b) increase ingestion rates of these EDC by grazing ruminants; (iii) determine the variation in soil EDC concentrations associated with season, time from sludge application and soil properties; (iv) assess the potential risk of ingestion of these compounds by grazing ruminants through the ingestion of soil, and therefore the potential risks of endocrine disruption. (Rates of soil intake by grazing sheep were measured twice a year and under contrasting climatic conditions, and the daily rates of ingestion of the selected compounds were then estimated.)

A knowledge of the natural and induced variability in soil concentrations and of the associated risk of ingestion by grazing animals is essential if changes in the pattern of environmental exposure resulting from altered environmental conditions or inputs are to be detected.

Materials and methods

Experimental design and application of sludge

The experiments were conducted at the Macaulay Land Use Research Institute research station at Hartwood, Scotland, approximately 20 miles east of the industrialised city of Glasgow. The soil is predominantly a non-calcareous gley and is typical of large areas of grazed land in the Central Belt of Scotland. The mean annual rainfall is approximately 1100 mm.

Six experimental plots, each comprising an area of approximately 9 ha, were used in the study. Commencing in mid-summer 1997, liquid, digested sewage sludge was applied twice annually (early spring and mid-summer) to three replicate plots (T) until five separate applications had been made (summer 1999). On each occasion (two occasions per year), sludge was applied at a rate of 2.25 tonnes DM per hectare to the whole of each plot using a pivot irrigation system, so that more than 95% of the surface area was estimated to be covered by the liquid sludge in a series of adjacent, sometimes overlapping, circles.

Animals were not allowed to graze the pasture until a minimum of 3 weeks after sludge application. This was achieved by dividing each plot into two half-plots. While one half-plot was treated over a period of 4–6 days, the animals grazed the other half. After the 3 week period had elapsed, the process was repeated with the animals grazing the previously treated half. After a further 3 weeks, the animals were given access to the whole plot.

Samples taken from each delivered load of sludge were pooled to create a representative sample of the total applied to each half-plot (in one case, the half-plot was further subdivided for the purpose of sludge application). The pooled samples were analysed to determine phthalate and alkyl phenol content. The rate of sludge application used resulted in the application of about 225 kg of nitrogen (N) ha⁻¹ year⁻¹, which was consistent with normal management practice. Control (C) plots were treated with an equivalent amount of inorganic N applied in three lots during each application period.

Collection of soil and herbage samples

All samples were collected and stored in glass washed with distilled water or in aluminium foil to ensure that the samples were not contaminated with phthalates from plastic containers or alkyl phenols from detergents.

Soil samples (3 cm deep) were collected using a 5 cm diameter corer. All samples were stored in glass jars at 4 °C until analysis.

Herbage samples were collected by cutting near to ground level. Care was taken to ensure that soil contamination of the herbage was restricted to that which occurred naturally as a result of rain splash or the actions of grazing animals. Samples were air-dried and stored in aluminium foil until analysis for titanium.

pH and organic matter determinations

Soil pH was determined using 15 g of soil in a one-to-three mixture of soil to 0.01 M CaCl solution.

Soil samples were oven-dried at 105 °C. The percentage of organic matter was determined from the weight loss on ignition at 450 °C of oven-dried (105 °C) soil.

Determination of sludge and soil concentrations of DOP, NP and OP

Although several phthalates were likely to be present in the samples, only DOP was measured as a representative of this group of compounds.

Field moist soil (15 g) was sieved to pass through a 5.6 mm mesh, weighed into a glass jar and an internal standard solution [2-ethylanthracene (EA); 75 µg] was added. The contents of the jar were mixed thoroughly and allowed to stand overnight to ensure equilibration of the contents. Dichloromethane (DCM; ‘Baker analysed’, J. T. Baker Ltd., Deventer, The Netherlands; 50 ml) was added and the contents were agitated for 6 h using an automatic roller system. Samples of the sludge (1 g) were dried at 50 °C, ground using a mortar and pestle and refluxed with DCM (100 ml) for 3 h. All samples were filtered through Whatman No. 4 filter paper containing anhydrous sodium sulfate, the solids were washed with DCM and the filtrate was reduced to a small volume using a stream of nitrogen prior to analysis. To correct for background concentrations of each EDC, blanks containing no sample (i.e. solvents and reagents only) were included with each set of samples and taken through the extraction and analysis procedures.

Recoveries of each EDC from the soil were determined by spiking samples with pure EDC at a rate of 5 mg g⁻¹. Mean [±standard error (SE)] rates of recovery were 86% ± 7% for OP, 89% ± 2% for NP and 134% ± 0% for DOP. The internal standard for all analyses was EA.

EDC concentrations were determined using gas chromatography linked to mass spectrometry (GC-MS) operated in the single ion recording (SIR) mode. The instrument used was a TRIO 1 quadrupole mass spectrometer (VG Masslab, Altrincham, Cheshire, UK) linked to a Fisons 8000 gas chromatograph fitted with an AS800 autosampler. Separations were effected on an HP5 fused silica capillary column [30 m × 0.25 mm (id)] coated with 95% dimethyl-/5% phenyl-polysiloxane with a phase thickness of 0.25 µm (Hewlett Packard Ltd., Stockport, UK). The operating temperature of the column was 150–190 °C at 7 °C min⁻¹, to 200 °C at 1 °C min⁻¹, to 305 °C at 10 °C min⁻¹. Helium was used as the carrier gas and samples were injected on to the GC column at a split ratio of 25∶1.

The mass spectrometer was operated in the electron ionisation mode at 70 eV and a source temperature of 300 °C. The ions monitored for each compound were as follows: m/z 206 (OP), m/z 135 (NP), m/z 150 (DOP) and m/z 191 (EA). Response factors relative to the internal standard EA were calculated for each component.

Quality control was achieved by repeated analysis of bulked samples of sewage sludge and soil. Mean values and ranges (±2σ) were calculated for each of the analytes. These quality control samples were then included with each batch of experimental samples analysed. The limit of detection for each of the analytes was 0.01 µg g⁻¹.

Experiment 1. Assessment of variability of EDC

The spatial variation in EDC concentration within plots, owing to uneven sludge application or differential degradation or retention, was assessed using 15 individual soil samples collected at regularly spaced intervals on transects across one half of each experimental plot, so that samples were collected from all parts of the half-plot. Samples were collected on a single occasion, between 69 and 81 days after the application of sludge. The samples were analysed individually, using the method described previously, to determine the concentrations of DOP, NP and OP present.

Experiment 2. Temporal variation in soil concentrations

During the course of the study, approximately 50 soil samples were collected every 3–6 months from each whole plot. The samples were collected at regular intervals on transects across each plot. The first set of samples was collected before the application of the first batch of sewage sludge. All samples collected from each plot on each date were pooled and analysed.

The temporal pattern of change in soil concentrations of the selected EDC following the application of sludge was estimated, to determine whether there was a decline in concentrations over time and, if so, whether the pattern differed with season and associated climatic conditions. During each month, for periods of 3 and 4 months following applications in the autumn and spring/summer, respectively, additional groups of 15 soil samples were collected from one half of each plot and pooled before analysis; the same half-plot was sampled on each occasion.

Concentrations were related to soil temperatures (107 temperature probe; Campbell Scientific Ltd., Shepshed, UK), recorded automatically at 10 s intervals and averaged daily, measured at a depth of 10 cm at a separate site within the same farm.

Experiment 3. Estimation of rates of EDC ingestion by grazing ruminants

The rate of soil ingestion by sheep grazing the plots was measured in order to estimate the potential rate of exposure of grazing ruminants to EDC.

Ten ewes from a single T plot and nine ewes of a similar range of ages from a single C plot were used, at two contrasting times of year (early March and late June), when the respective mean (±SE) pasture sward heights were short (T: 3.70 ± 0.144 cm; C: 3.65 ± 0.127 cm) and moderately long (T: 7.93 ± 0.320 cm; C: 8.48 ± 0.266 cm), and when the ewes were in mid-pregnancy and mid-lactation, respectively. Both of these stages of the reproductive cycle are associated with high rates of feed intake. The animals were dosed with intraruminal controlled-release devices (CRDs) (Captec ALKANE*, Captec NZ Ltd, New Zealand) which were designed to deliver n-dotriacontane (C₃₂ alkane) and n-hexatriacontane (C₃₆ alkane) at constant rates over a period of about 20 days to sheep weighing 25–80 kg. Over a 5 day period, beginning 10 days after CRD insertion, rectal-grab faeces samples were obtained daily from the animals. Samples of herbage, representative of ingested material, were collected by hand-plucking from the experimental plots during each of the faecal sampling periods. At the same time, soil was sampled to a depth of 3 cm. The samples of faeces from each animal were bulked over each collection period and, together with the herbage samples, stored at −20 °C prior to analysis; soil samples were air-dried.

Faeces and herbage samples (freeze-dried and milled) were analysed for n-alkanes using the method of Mayes et al.,¹² with modifications described by Salt et al.¹³

Herbage intake was calculated according to Mayes et al.,¹² using C₃₂ and C₃₃ alkanes as the respective dosed and herbage markers; the daily dose rate of C₃₂ was that quoted by the manufacturer of the CRD (50 mg). The daily output of faeces was estimated from the dilution rate of the C₃₆ alkane which was released from the CRD (50 mg day⁻¹). A correction was made to account for incomplete recovery (0.95) of C₃₆ in faeces.¹⁴

Soil intake was estimated using titanium as an indigestible marker.¹⁵ The titanium contents of soil and faeces were determined by X-ray fluorescence spectroscopy. Titanium is abundant in most inorganic soils, yet its uptake by plants is negligible, and so Ti recovered in faeces can be considered to originate from the ingestion of soil.¹⁶ The intake of soil was calculated from the daily faecal DM output and the concentrations of Ti in the samples of soil and faeces.

Statistical analyses

In Experiment 1, the variance in soil EDC concentrations within and between plots was assessed using restricted maximum likelihood analysis. EDC concentrations were correlated with soil organic matter content and pH.

In Experiment 2, the concentrations of EDC in soils collected sequentially, following sludge application, were analysed by a two-way analysis of variance, blocking for plot, and with main effects of treatment and season. Within each season, samples were analysed using a two-way analysis of variance, blocked for plot, with main effects of treatment and date. Data were transformed using log₁₀(value), where appropriate, to correct for skewed distribution.

In Experiment 3, the effects of treatment and season on herbage intake and digestibility were analysed using a two-way analysis of variance. Soil intake data were log₁₀ transformed before analysis to meet the requirement for constant variance.

Correlations, analysis of variance and restricted maximum likelihood analysis were performed using Genstat 5.3.¹⁷

Results

Concentrations of DOP, NP and OP in the sludge exhibited considerable variation. NP was present at the highest concentrations, followed by DOP, and OP was present at lower concentrations (Table 1).

Table 1 Mean (±SE) concentrations (µg g⁻¹) of dioctyl phthalate (DOP), nonyl phenol (NP) and octyl phenol (OP) in batches of sludge applied to three plots (n = 30)

	Mean	SE	Range
DOP	95.6	6.50	58.3–208.6
NP	145.9	10.6	73.5–283.5
OP	0.277	0.140	<0.001–3.68

---

Experiment 1. Assessment of variability of EDC

Random soil samples within plots exhibited a wide range of soil concentrations of DOP and NP (Fig. 1).

Fig. 1 Distribution of soil concentrations (µg g⁻¹) of (a) dioctyl phthalate (DOP) and (b) nonyl phenol (NP) in soil samples from plots treated with inorganic fertiliser (control; solid bars) or sewage sludge (treated; open bars).

The overall mean soil concentrations of DOP were not significantly higher in T than in C plots [0.233 vs. 0.155 µg g⁻¹; standard error of the difference (SED) = 0.046; not significant (NS)]. Mean soil concentrations of NP exhibited a similar trend (0.021 vs. 0.013; SED = 0.002; P < 0.05), but concentrations were much lower. Soil OP concentrations were generally below detectable levels (<0.01 µg g⁻¹; Fig. 1).

Contributions to the variability in soil EDC concentrations from both within and between replicate plots are shown in Table 2. The ratio (W/B) describes the relative contributions from the two sources, namely within and between plots. In all four cases, the ratios were greater than unity, implying that, for both treated and control areas, contribution to the variability from within plots was greater than that from between plots. The ratio gives an estimate of the number of samples required for the variance of the sample mean to contain equal contributions from within and between plot variance components.

Table 2 Comparison of estimated variance components for ‘within plot’ and ‘between plot’ for DOP and NP concentrations. Standard errors are shown in parentheses. Values shown are 10⁴ × (µg g⁻¹)²

	Within plot	Between plot	Ratio (W/B)
DOP
Treated	169.2 (36.9)	41.4 (52.7)	4.1
Control	97.9 (21.4)	3.8 (10.4)	25.0
NP
Treated	0.642 (0.140)	0.0110 (0.0545)	58.3
Control	0.299 (0.065)	0.0246 (0.0447)	12.2

---

For both DOP and NP, the within plot variance is greater for T plots than for C plots. The variance ratios are 1.73 (P < 0.05) and 2.15 (P < 0.01) for DOP and NP, respectively (42 d.f. for both numerator and denominator degrees of freedom).

Differences between samples in EDC contents were not significantly correlated with the pH or organic matter contents of the samples. The correlations between pH and soil DOP and NP and between soil organic matter and soil DOP and NP were low, irrespective of treatment (Table 3), and none was statistically significant.

Table 3 Correlation coefficients between soil pH and soil DOP and NP concentrations and between soil organic matter contents and soil DOP and NP concentrations in sludge-treated (T) and control (C) plots (n = 45 per treatment)

	Treated		Control
	DOP	NP	DOP	NP
pH	0.226	−0.038	−0.276	−0.034
Organic matter	0.100	−0.207	−0.008	−0.249

---

Experiment 2. Temporal variation in soil concentrations

Long-term changes in soil EDC concentrations. The mean soil concentrations of DOP and NP (values pooled within treatment) over time are presented in Figs. 2(a) and 2(b), in conjunction with the periods of sludge application and mean weekly soil temperature profiles (at 10 cm depth). There was no sustained increase during the study in the mean soil concentrations of either DOP or NP, and similar concentrations of DOP were recorded in C and T plots.

Fig. 2 Mean concentrations (µg g⁻¹) of (a) dioctyl phthalate (DOP) and (b) nonyl phenol (NP) in soils from sludge-treated (◊) and control (✦) plots throughout the study and associated soil temperature profiles. Sludge was applied during the periods marked with arrows.

The overall mean log₁₀ concentrations (µg g⁻¹; back-transformed means in parentheses) of DOP were more than 50% higher in soil from T than C plots (0.157 (0.44) vs. 0.108 (0.28), SED on log scale = 0.020; P < 0.1), but there was a wide range of concentrations in both T and C samples. The overall mean concentrations of NP (back-transformed means in parenthesis) were much lower, but exhibited a similar, but statistically significant, trend towards higher concentrations in T plots (0.011 (0.023) vs. 0.007 (0.016), SED on log scale = 0.001; P < 0.05). As many samples had OP concentrations below detectable levels, mean concentrations were not calculated.

There was no evidence that the mean soil concentrations of DOP and NP were related to patterns of rainfall, which showed marked short-term variation, but no prolonged periods of rainfall or drought (data not shown). However, significant differences in mean concentrations of DOP (P < 0.001) and NP (P < 0.05) were recorded for samples collected at different times. In general, the higher and lower concentrations were associated with relatively low and high soil temperatures, respectively (Fig. 2).

Soil DOP concentrations at the first soil sample collection time after sludge application were generally above basal levels in T plots, but the effect was most marked when the sludge was applied during periods of low or declining soil temperatures.

Short-term attenuation of EDC in soil. There was no effect of either treatment or season (sample date) on the overall mean concentrations of DOP (0.211 µg g⁻¹) or NP (0.025 µg g⁻¹) in soils collected at monthly intervals following sludge application. The concentrations of OP in many samples were undetectable (<0.01 µg g⁻¹) and concentrations were considered to be too low and variable to allow meaningful analysis.

Comparisons of mean values within seasons indicated that there were significant differences (P < 0.001) in mean NP with month of sample collection at both times of the year, although differences in mean DOP concentrations (P < 0.001) were observed only during the spring (April), when the mean soil concentration at the first month after sludge application was higher (Fig. 3); this trend was recorded in both T and C plots. No such trend was apparent following application in the summer (July). By contrast, the highest soil concentrations of NP were present at the third month after sludge application in summer (Fig. 3). There were no significant treatment × season interactions.

Fig. 3 Mean concentrations (µg g⁻¹) of (a) dioctyl phthalate (DOP) and (b) nonyl phenol (NP) in soils from sludge-treated and control plots (3 plots for each treatment) at monthly intervals following sludge application. Mean date of application of sludge (Treated) or inorganic fertiliser (Control) is indicated by arrows.

In the T plots, there was no consistent relationship between the mean soil concentrations of either DOP or NP and the time since sludge application (Fig. 4).

Fig. 4 Mean concentrations (µg g⁻¹) of (a) dioctyl phthalate (DOP) and (b) nonyl phenol (NP) in soils from sludge-treated plots in relation to time from application (3 plots for each season).

Experiment 3. Estimation of rates of ingestion of soil and associated EDC by grazing ruminants

There was no evidence of treatment or season effects on mean herbage digestibility (81.2%; SED = 0.015) or intake (2.66 kg DM day⁻¹; SED = 0.264), but there was a higher mean log(daily intake (g) of soil) in March than in June (2.13 vs. 1.45; SED = 0.061; P < 0.001) and a much higher percentage of DM intake represented by soil (5.40 vs. 1.17; SED = 0.360; P < 0.001) (Table 4).

Table 4 Mean daily intakes of soil (g DM; back-transformed means) and estimated mean daily intakes (µg) of DOP and NP by ewes grazing sludge-treated (T) or control (C) plots in March (late winter; low sward height) and June (summer; moderate sward height). These data are based on the mean concentrations recorded in soil samples collected at 3–6 month intervals during the study

	March		June
Soil intake	135.5		28.5
(Range)	(70.1–314.0)		(15.4–47.8)
% soil in diet	5.40		1.17
(Range)	(3.0–8.8)		(0.60–2.45)

---

	March		June
	Control	Treated	Control	Treated
Estimated intakes
DOP	43.5	66.5	9.1	13.8
(Range)	(21.5–96.4)	(32.9–147.3)	(4.7–14.7)	(7.22–22.4)
NP	2.41	3.68	0.50	0.77
(Range)	(1.19–5.34)	(1.82–8.16)	(0.26–0.81)	(0.40–1.24)

---

Extrapolating from the measured concentrations of EDC in soil, calculated daily rates of intake of DOP and NP are up to approximately 150 and 8 µg, respectively, with large differences in estimated values both between animals and between seasons (Table 4).

Discussion

The concentrations of phthalate (as indicated by the concentrations of DOP) and alkyl phenols in the sludges applied to the pastures in the present study were broadly consistent with those reported previously, both in terms of mean concentration and variation with time or batch.^10,11 The concentrations in the soil of sludge-treated pasture were also similar to those reported previously.¹⁸

There was substantial within plot variation in concentrations in both T and C plots, indicating that multiple samples were necessary for an accurate measure of soil EDC concentrations. However, as the variance amongst the 15 samples within each half-plot was broadly similar to the between plot variance component, it was concluded that the number was sufficient for the assessment of mean EDC concentrations. It was concluded, on the basis of these observations, that the collection and pooling of approximately 50 samples from entire plots, at intervals during the study, was sufficient to provide a meaningful measure of EDC concentrations in whole plots.

The low correlations between EDC content and organic matter content and pH suggest that the variation in soil EDC concentrations could not be attributed to the variation in these factors, despite the fact that these compounds are generally considered to be hydrophobic and tend to associate with organic matter.¹⁹ However, the ranges of organic matter content and pH in the soils studied were relatively small, although probably representative of lowland pasture, and so it is concluded that these variables are unlikely to be important determinants of soil EDC concentrations within fields.

The slightly elevated DOP and NP concentrations in T plots at 75 days after sludge application were recorded at a time which was approximately 50 days after the end of the ‘no-grazing’ period, and so animals grazing these pastures were potentially exposed to increased levels of EDC. While no such treatment difference in overall mean concentrations was recorded in Experiment 2, similar trends were observed. There were significant changes in both NP and DOP with time, with the higher concentrations being recorded during or shortly after periods of low soil temperature. However, there was no evidence of a long-term accumulation of these compounds in the soil as a result of repeated applications of sludge.

The changes in mean concentration with time were presumably a function of reduced microbial degradation, volatilisation, leaching or sequestration, resulting in an extension of the normal half-lives of these compounds of days (NP) or weeks (DOP).²⁰ Laboratory-based studies of microbial degradation of phthalates²¹ support this suggestion, indicating that degradation rates increase as the temperature is increased, even from relatively high initial temperatures (23 °C).

Higher concentrations were recorded in soils of T than C plots on some, but not all, occasions following sludge application. However, there were also occasions when DOP concentrations in T and C plots changed in parallel and not in response to sludge application. This suggests that sources other than sludge application may be important determinants of soil concentration. It is suggested that the transient increases in EDC concentration observed in Experiments 2 and 3, which could not be attributed to sludge application, may reflect variations in aerial deposition, which is known to be a significant source of some EDC.²² The effects of such changes were apparently sufficient to mask any changes in concentration as a result of leaching, volatilisation, degradation or sequestration following sludge application, at least 1 month after the time of application. Furthermore, as the magnitude of these changes was broadly similar to those observed in T plots, in which there was a substantial input of EDC via sludge, it is concluded that the putative aerial inputs may also be substantial and potentially biologically significant.

While the mean concentrations of NP and DOP in sludge were of a similar level, the soil concentrations of NP in both T and C plots were markedly lower. This suggests that the relative rate of loss of NP from the soil, as a result of degradation, leaching, evaporation or other loss, was higher than that of DOP, and so the relative risk of exposure of grazing animals to the former was therefore lower.

In a review of the potential effects of repeated sludge application on soil concentrations of organic pollutants,¹⁹ it was concluded that the effects were minimal and that concentrations were unlikely to be in excess of those considered to be ‘safe’. The results of the present study also indicate that there is no long-term accumulation of these EDC in soil, following repeated sludge application. However, this study also shows that soil DOP concentrations in the area are relatively high, even in C plots, and the potential remains for ingestion by grazing animals to be biologically significant. Furthermore, soil concentrations were generally higher during the winter when the herbage mass was generally reduced and the intake of soil was shown to be highest, and so there may be an enhanced risk of EDC ingestion at this time as a result of two different factors.

Many factors other than soil concentrations of EDC are likely to determine the rate of exposure of grazing animals. Studies of patterns of uptake and accumulation of EDC in plants indicate that significant uptake of these compounds via the root system is unlikely to occur,^23–25 and so it is unlikely that there is significant transfer to grazing animals via this route. However, the possibility remains of transfer of EDC to ruminants through the surface contamination of the pasture,²⁰ either as a result of accumulation on the leaf surface, following the application of sludge, or through the transfer of soil on to plant surfaces as a result of rain splash^26,27 and the actions of grazing animals, particularly in wet soil conditions.²⁷

The results of Experiment 3, which indicated that soil ingestion by sheep differed with season, but could be as high as 300 g day⁻¹, are entirely consistent with those reported previously.²⁰ Based on the mean concentrations of DOP in soil from T plots, this could represent an ingestion rate of about 150 µg day⁻¹ of DOP, but very much smaller amounts of alkyl phenols.

There are few data concerning the ‘safe’ levels of EDC in the diet of farm animals. Based on a No Observed Effect Level (NOEL) of 25 µg kg⁻¹ body weight day⁻¹ derived from human studies,²⁸ assuming a 60 kg live weight and applying a margin of safety factor of 100 times the NOEL, levels of phthalate intake in excess of 15 µg day⁻¹ may be deemed to be worthy of concern. Most of the estimated intakes which exceeded this value were associated with the T plots. It is concluded that the potential rate of ingestion of DOP by ruminants, under certain climatic and soil conditions, may exceed levels that are currently considered to be of concern, and that the application of sludge to pasture may slightly increase levels of exposure. The estimation of the biological significance for farm animals of the estimated levels of exposure will require the determination of rates of accumulation in tissue, patterns of metabolism and excretion and sensitivity of the target organs to each type of EDC.

It should be noted that, as there are large variations in DOP concentrations within plots, together with large variations in soil ingestion rate, individual animals may ingest larger amounts than those estimated, depending on the precise areas grazed. Furthermore, for the same reasons, levels of ingestion in individual control animals may also exceed the levels which are deemed to be safe.

Rat studies have provided a NOEL of 15 mg kg⁻¹ live weight day⁻¹ for alkyl phenols.²⁹ Using this figure and assuming a 60 kg live weight and a 100-fold safety margin, daily doses up to 9 mg day⁻¹ would be deemed to be safe with respect to risk of endocrine disruptive effects. It is therefore concluded that the likely levels of exposure of grazing ruminants to NP are unlikely to be of any biological significance.

Although a possible risk of ingestion of significant amounts of EDC has been identified, as there is little change in soil EDC concentrations from about 1 month after sludge application, it is unlikely that an extension of the period of exclusion of grazing animals would result in a significant change in exposure to EDC through soil ingestion.

Acknowledgements

The provision of sludge by the West of Scotland Water Authority and their co-operation and advice in the planning and conduct of this study are gratefully acknowledged. This work was funded by the Scottish Executive Environment and Rural Affairs Department.

References

1. Commission of the European Communities (1994), Directorate Generale XI, Waste Management – Sewage Sludge. Part 1. Survey of Sludge Production, Treatment, Quality and Disposal in the European Union, EU Commission, Brussels, 1994.
2. J. Toppari, J. C. Larsen, P. Christiansen, A. P. Giwercman, P. Grandjean, L. J. Guillette, B. Jegou, T. K. Jensen, P. Jouannet, N. Keiding, H. Leffers, J. A. McLachlan, O. Meyer, J. Muller, E. R. Meyts, T. Scheike, R. Sharpe, J. Sumpter and N. E. Skakkebaek, Environ. Health Perspect., 1996, 104(Suppl. 4), 741.
3. T. M. Crisp, E. D. Clegg, R. L. Cooper, W. P. Wood, D. G. Anderson, K. P. Baetcke, J. L. Hoffman, M. S. Morrow, D. J. Rodier, J. E. Schaeffer, L. W. Touart, M. G. Zeeman and Y. M. Patel, Environ. Health Perspect., 1998, 106(Suppl. 1), 11.
4. R. B. Dean and M. J. Suess, Waste Manage. Res., 1985, 3, 251.
5. M. L. Boerjan, S. Frejnagel, S. M. Rhind and G. A. L Meijer, Anim. Sci., 2001, 74, in press.
6. C. A. Harris, P. Henttu, M. G. Parker and J. P. Sumpter, Environ. Health Perspect., 1997, 105, 802.
7. J. C. Baxter, M. Aguilar and K. Brown, J. Environ. Qual., 1983, 12, 311.
8. R. White, S. Jobling, S. A. Hoare, J. P. Sumpter and M. G. Parker, Endocrinology, 1994, 135, 175.
9. P. C. Lee, P. Arndt and C. Nickels, Endocrine, 1999, 11, 61.
10. P. H. Brunner, S. Capri, A. Marcomini and W. Giger, Water Res., 1988, 22, 1465.
11. M. D. Webber and S. Lesage, Waste Manage. Res., 1989, 7, 63.
12. R. W. Mayes, C. S. Lamb and P. M. Colgrove, J. Agric. Sci., 1986, 107, 161.
13. C. A. Salt, R. W. Mayes and D. A. Elston, J. Appl. Ecol., 1992, 29, 378.
14. H. Dove and R. W. Mayes, Aust. J. Agric. Res., 1991, 42, 913.
15. H. F. Mayland, A. R. Florence, R. C. Rosenau, V. A. Lazar and H. A. Turner, J. Range Manage., 1975, 28, 448.
16. S. Jagger, J. Wiseman, D. J. A. Cole and J. Craigon, Br. J. Nutr., 1992, 68, 729.
17. Lawes Agricultural Trust, Genstat 5 Committee, Genstat 5 Release 3, Reference Manual, Clarendon Press, Oxford, 1994.
18. G. A. O'Connor, in Sewage Sludge: Quality Aspects and Risk in Connection with Land Application, Swedish Water and Wastewater Association, Stockholm, 1994, pp. 1–12.
19. S. R. Smith, in Agricultural Recycling of Sewage Sludge and the Environment, CAB International, Wallingford, 1995, ch. 10, pp. 207–236.
20. B. A. Stark and J. E. Hall, in Effects of Organic Contaminants in Sewage Sludge on Soil Fertility, Plants and Animals, ed. H. E. Hall, D. R. Sauerbeck and P. L'Hermite, Commission of the European Communities, Luxembourg, 1992, pp. 134–157.
21. J. C. Inman, S. D. Strachan, L. E. Sommers and D. W. Nelson, J. Environ. Sci. Health B, 1984, 19, 245.
22. T. Colborn, F. S. vom Saal and A. M. Soto, Environ. Health Perspect., 1993, 101, 1.
23. G. A. O'Connor, R. L. Chaney and J. A. Ryan, Rev. Environ. Contam. Toxicol., 1991, 121, 129.
24. S. R. Wild and K. C. Jones, Sci. Total Environ., 1992, 119, 85.
25. G. Fries, J. Anim. Sci., 1995, 73, 1639.
26. M. Dreicer, T. E. Hakonson, C. C. White and F. W. Whicker, Health Physics, 1984, 46, 177.
27. T. G. Hinton, J. M. Stoll and L. Tobler, J. Environ. Radioact., 1995, 29, 11.
28. EU Commission, Synoptic Document No. 7. Draft of Provisional List of Monomers and Additives Used in the Manufacture of Plastics and Coatings Intended to Come into Contact with Foodstuffs (Updated May 1994), Document CS/DM/2356, EU Commission, Brussels, 1994..
29. S. Muller and C. Schlatter, Pure Appl. Chem., 1998, 70, 1847.

---