Cadmium Accumulation in Small Mammals: Species Traits, Soil

Jul 19, 2011 - The results of this field study show that species-specific traits of small mammals are important determinants of accumulation of cadmiu...
0 downloads 0 Views 827KB Size
ARTICLE pubs.acs.org/est

Cadmium Accumulation in Small Mammals: Species Traits, Soil Properties, and Spatial Habitat Use Nico W. van den Brink,* Dennis R. Lammertsma, Wim J. Dimmers, and Marie Claire Boerwinkel Alterra, Wageningen UR, PO-Box 47, 6700 AA Wageningen, The Netherlands ABSTRACT: In this study, the impact of species-specific spatial habitat use, diet preferences, and soil concentrations and properties on the accumulation of cadmium in small mammals was investigated. The results show that for the wood mouse (Apodemus sylvaticus), a mobile species with a large range in diet composition, accumulation of cadmium was not related to local soil concentrations or soil properties, but to diet preferences. For the common vole (Microtus arvalis), a nonmobile, specific feeding species, accumulation of cadmium was related to local soil concentrations or properties. For the bank vole (Myodes glareolus), a species with a smaller home range than the wood mouse but a broader diet spectrum than the common vole, both local soil properties and diet appeared to affect the cadmium accumulation in the kidneys. The results of this field study show that species-specific traits of small mammals are important determinants of accumulation of cadmium on a local scale. For site-specific assessment of risks of contaminants, such information is essential in order to understand exposure dynamics.

’ INTRODUCTION To assess risks that environmental contaminants may pose to wildlife it is essential to get insight into the degree to which animals are exposed. In terrestrial ecosystems, wildlife is mainly exposed to contaminants from the soil, and to a lesser extent to atmospheric or aquatic sources. Soil contaminants are accumulated by terrestrial receptors through the diet, or by direct ingestion of soil.13 Several studies have assessed relationships between concentrations in wildlife and concentrations in the soil, with varying success.48 This variable success in relating soil concentrations directly to concentrations in the organisms is due to the fact that several other factors may interfere with the accumulation of environmental contaminants. First, soil properties may affect the bioavailability of, for instance, metals,9,10 which can affect uptake in wildlife.7 Second, age of the animals may also influence internal concentrations.11,12 Furthermore, since environmental chemicals are generally accumulated through the diet, food web dynamics and diet composition may affect accumulation patterns significantly.13,14 Factors affecting accumulation may vary in time and space (seasonality), so traits of species that govern their spatial and temporal (foraging) behavior may affect accumulation patterns.1519 Several modeling studies have focused on the integration of these confounding factors in accumulation processes.16,2022 However, only a few experimental laboratory or field studies are available in which detailed information on soil concentrations and properties, diet composition, and spatial habitat use of the receptors are integrated in an assessment of accumulation of environmental contaminants in terrestrial wildlife.15,23 To overcome this, a study was designed in which accumulation of r 2011 American Chemical Society

cadmium (Cd) from soil to small mammals was related to their species-specific spatial habitat use and diet composition, and to soil concentrations and properties. Cd accumulation was assessed in three different small mammalian species, wood mouse (Apodemus sylvaticus), bank vole (Myodes glareolus) and common vole (Microtus arvalis), known to have different diet preferences and spatial habitat use.17,2427 Spatial habitat use was assessed using injectable transponders and automated receivers, while diet preferences were assessed using stable isotope signals of nitrogen and carbon (δ-15N and δ-13C).28

’ EXPERIMENTAL SECTION Study Area. The study area is situated near the village of Renkum, The Netherlands (51°950 58.9200 N, 5°450 16.6200 E). The landscape consisted of a mosaic of hedgerows, arable land, forest, and pastures on a sandy soil. Crops grown on the arable land were maize and wheat. Forest was coniferous dominated by Scotch-pine (Pinus sylvestris) without a shrub layer, or deciduous forest, dominated by oak (Quercus robur) and birch (Betula pendula). The farmed pastures were abandoned about 10 years ago and were transformed into meadows with vegetation dominated by long grass and herbs (e.g., Holcus lanatus, Taraxacum officinalis, Rumex acetosella, Senecio jacobea, Agrostis spp., and Trifolium repens). Received: March 16, 2011 Accepted: July 19, 2011 Revised: June 21, 2011 Published: July 19, 2011 7497

dx.doi.org/10.1021/es200872p | Environ. Sci. Technol. 2011, 45, 7497–7502

Environmental Science & Technology

ARTICLE

Table 1. Statistical Output for Relationships between Occurrence of Wood Mice and Bank Vole in Relation to Distance to Hedgerow, Season, and Species (Figure 1)a estimate

tprob

constant

2.369

p < 0.001

season spring

1.306

p = 0.005

season summer

2.014

p < 0.001

distance

0.4143

p < 0.001

wood mouse distance*wood mouse

0.524 0.1974

p = 0.040 p = 0.026

parameter

a

GLM-analysis with Poisson as link function;31 significant dependent variable: occurrence (n/receiver); independent variables: season (Spring (AprilJune), Summer (JulySeptember), Fall (OctoberDecember)), distance, species and interaction distance (m)*species (wood mouse and bank vole); reference: season: Fall, species: bank vole).

Figure 1. Number of observations of animals per receiver at different distances of the hedgerow where the animals were caught. (A) wood mouse; (B) bank vole (n/receiver; for 0, 5, and 10 m n = 3 antennas, for 15 m n = 1 antenna). (Spring: AprilJune; Summer: JulySeptember; Fall: October December; Winter JanuaryMarch, too low numbers in winter for detailed analyses). Total number of wood mouse caught: 187; bank voles: 44.

Collection of Small Mammals. Three studies were performed to assess (1) spatial extent of habitat use by the species, (2) habitat preferences of the species, and (3) diet and accumulation assessment. Spatial Extent of Habitat Use. To assess the spatial habitat use by small mammals, 187 wood mice and 44 bank voles were captured in a hedgerow (May 8, 2006July 17 2007) using Longworth live traps.29 Only a single common vole could be captured in this habitat and will not be discussed in this experiment. Traps were baited with peanut butter and cat food, and contained hay as bedding. Animals were caught alive in these traps and marked individually with a transponder (TROVAN, Identify UK Ltd., Yorkshire, UK). Spatial habitat use was determined by placing receivers at 0, 5, 10, or 15 m from the hedgerow (total 10 receivers) from June 16, 2006 until November 29, 2007. Customized single-loop, ring-shaped antennas with a diameter of 50 cm and height of 10 cm were used to automatically detect marked animals within close distance (< 20 cm), and the data were stored in a logger (LID650, EID Aalten bv, Aalten, The Netherlands). Per receiver, the total number of animals per day (24 h) could be calculated. Animals that were recorded more than once per day were regarded as a single reading. For each receiver the total number of animals per receiver per day (no./receiver/day) was calculated per species.

Habitat Preferences. To determine the habitat preferences of the small mammals, 32 additional wood mice and 7 bank voles were captured at different trap lines in different habitats. Animals were also marked with a transponder and tracked with the similar receivers as the former experiment, placed in different habitat types around the trap lines. Animals were detected similarly to the other experiment. Accumulation Assessment. Twenty-seven wood mice, 13 bank voles, and 7 common voles from the same locations, with an additional location on arable land, were sacrificed in order to assess the cadmium concentrations in their kidneys ([Cd]kidney) and stable isotopes of carbon and nitrogen in the muscles of the hind leg. Only adult specimens were used in order to minimize age-related variation. Animals were sedated with CO2, and sacrificed by cervical dislocation. All procedures involving the handling of animals were conducted by certified persons, and reviewed by an Animal Ethics Committee according to Dutch legislation on the protection and welfare of vertebrate animals used for experimental and other scientific purposes. At each sampling location, soil samples were collected to assess the Cd concentrations in the soil ([Cd]soil). Furthermore, diet items (grass (vegetative parts and seeds), acorn, and earthworm) were collected at random locations for stable isotope analyses. Metal Analysis. Metal analyses were performed according to ref 7. In short, soil was dried at 40 °C and the kidneys were freezedried prior to chemical analysis. Samples were digested with aqua-regia in a microwave in Teflon vessels. Samples were analyzed for Cd using inductively coupled plasma atomic emission spectrometer (ICP-AES). When concentrations were below detection limits, analyses were performed with inductively coupled plasma mass spectrometry (ICP-MS). Limits of detection were 0.012 mg/kg dry weight for the soil and 0.075 mg/kg dry weight for kidney samples. Recoveries were in range of 95111%. All concentrations are reported as dry weight concentrations, unless stated otherwise. For quality assurance, reference samples from clay and sandy soils were analyzed according to criteria of WEPAL (ISE 989, ISE 949, WEPAL, www.wepal.nl). Furthermore, reference liver tissue (BCR-185R, Community Bureau of Reference, the former reference materials program of the European Commission) was analyzed as quality assurance for the kidneys. Diet Assessment Using Stable Isotope Analysis. All samples were freeze-dried, ground by a ball mill (Retsch MM 2000, Retsch GmbH & Co., Haan, Germany) and 13C and 15N enrichment was 7498

dx.doi.org/10.1021/es200872p |Environ. Sci. Technol. 2011, 45, 7497–7502

Environmental Science & Technology

ARTICLE

Figure 3. Biplot of δ-13C and δ-15N (%; averages and standard deviations) of the three small mammal species (filled dots) and some major diet items (open dots).

δ-values in per mill (%) deviations from the standards. All analyses were carried out in duplicate. Statistical Analysis. All statistical analyses were performed with GenStat, version 13 (www.genstat.co.uk). Relationships between occurrence of small mammals and distance to the hedgerow were analyzed with General Linear Models (GLMs) with a Poisson link function.31 Differences between habitat types and species in [Cd kidney] δ-13C and δ-15N and ratios between [Cd]kidney and [Cd]soil were analyzed with Analysis of Variance (ANOVA), with Least Significant Differences (LSD) as posthoc test. Homogeneity of variance was checked with Bartlett’s test, which is rather sensitive for departures from normality; we therefore also checked residuals posthoc for deviations. Relationships between factors governing [Cd]kidney and [Cd]soil were analyzed with GLMs. [Cd]kidney and [Cd]soil were log10-transformed prior to statistical analyses, but the signals for δ-13C and δ-15N were not.

’ RESULTS AND DISCUSSION

Figure 2. Number of wood mice (A) and bank voles (B) caught in different habitat types in different seasons in different habitat types (n/ receiver/day; note different scales between A and B).

measured after combustion in an elemental analyzer with a continuous flow isotope ratio mass spectrometer (Delta C, Finnigan MAT, Bremen, Germany). Data were expressed as atom percent (atom %). As standard for C-enrichment, the VPDB standard was used, a redefined standard in relation to the PeeDee Belemnite formation (South Carolina, USA).30 For N-enrichment, standards supplied by the International Atomic Energy Agency (IAEA, Vienna) were used. All reference gases used in the analyses were obtained directly from IAEA. Isotope ratios are expressed as

Spatial Habitat Use. In Figure 1, the relationship between the occurrence of marked animals and the distance to the hedgerow in which they were caught is shown. Too few common voles were captured for detailed assessment, so data on this species are not included in the following analysis. Statistical analysis revealed significant effects of distance, season, and species (Table 1). Bank voles were detected in higher numbers, especially considering the lower number of bank voles captures in this experiment. This is most pronounced within the hedgerow. Numbers of detection are lowest in fall, which is likely related to the fact that in spring and summer the arable habitat provided more cover and possibly food in comparison to fall. In a study on small-scale movements of wood mice, both cover and food availability were important drivers of local spatial habitat use.27 It is evident that numbers of detection of marked animals were lower at greater distance from the hedgerow. The interaction between species and distance is significant (p = 0.026) indicating that the wood mouse and bank vole showed different relationships between occurrence and distance from the hedgerow (Figure 1). Wood mice occurred at relatively higher numbers at greater distance in comparison to the bank vole. Furthermore, marked wood mice were detected at lower frequency than bank voles, which would imply that a higher proportion of the captured wood mice were outside the area 7499

dx.doi.org/10.1021/es200872p |Environ. Sci. Technol. 2011, 45, 7497–7502

Environmental Science & Technology

ARTICLE

Table 2. δ-C and δ-N % in Different Small Mammal Species in Different Habitats (%)a δ-15N (%) wood mouse arable

2.87 ( 3.09

forest

1.00 ( 2.69

meadow a

0.41 ( 1.3

δ-13C (%)

bank vole

common vole

2.40

20.24 ( 0.23

wood mouse

bank vole

25.56 ( 1.76

30.21

1.72 ( 1.64

25.87 ( 1.18

26.72 ( 1.29

3.78 ( 0.77

25.34 ( 0.79

26.78 ( 0.21

common vole 29.75 ( 0.52

Average ( standard deviation, in case of lacking standard deviation just one observation is available.

covered by the receivers. Wood mice were detected in all habitats (Figure 2), while bank voles only showed up in habitats with cover: long grass, shrubs, open forest, and hedgerow (arable sites not included in this experiment). The total number of wood mice marked in this experiment was greater than bank voles, which is reflected in the number of detection of marked animals. Within hedgerows, bank vole numbers were in the same range as those of the wood mouse; for the other habitat types the number of bank voles was much lower in comparison to the wood mouse (note the different scales in the two figures). The results show that wood mice appeared to occupy larger areas and are less strict in their habitat preferences. This is in confirmation with refs 32 and 33 who deduced larger home ranges for wood mouse in comparison to the bank vole in similar habitats. The common vole was only detected in two habitats, short grass and arable land, indicating a restricted distribution, which is in agreement with ref 33. These findings indicate species-specific use of the habitat, which may affect the potential of the different species to accumulate Cd.20 Species-Specific Diet. Accumulation patterns may be related to the diet of the animals in the different habitats, because prey species and vegetation may show habitat-specific availability.34 Because accumulation efficiency can differ among soil invertebrate species in different habitats,35 habitat-specific diets, reflected in the stable isotope signal, may have a big impact on the accumulation patterns. Figure 3 shows the biplot of δ-13C and δ-15N in the muscles of the small mammal species and some diet items. There are significant differences between species and habitats in δ-15N and δ-13C (Table 2; δ-15N: Species Fprob = 0.005, habitat: Fprob < 0.001;δ-13C: Species Fprob < 0.001, Habitat Fprob = 0.003, ANOVA). Species are separated by δ-13C (common vole versus wood mouse and bank vole) and also by δ-15N (bank vole versus common vole and wood mouse, Figure 3). Trophic transfer of diet items results in enrichment of the δ-13C and δ-15N signals,36,37 on average 0.51% for δ-13C and 3.5% for δ-15N. When applying an average increase of 0.75 for δ-13C and 3.5 for δ-15N it becomes evident that the isotope signal in common voles indicates vegetative grass as a major diet item (Figure 3). For the wood mouse, the isotope signal indicates acorn and earthworm as potential diet items, while grass seeds and/or earthworm appear to be important in the diet of the bank vole. For the bank vole and wood mouse the variation in the isotope signal was rather high (Table 2), indicating high variance in diet composition for these species. This is in agreement with Watts,24 who reports for both the wood mouse and bank vole in woodland edge habitat, a wide range of diet items including animals, but a preference of the wood mouse for seeds and the bank vole for vegetative parts. The common vole showed little variation in both δ-13C and δ-15N indicating a narrow diet. Accumulation of Cadmium and Relationships with Soil Concentrations, Soil Properties, and Stable Isotopes. [Cd]soil ranged from 0.06 to 0.44 mg/kg dry weight (mean: 0.26, standard

Figure 4. Relationships between [Cd]kidney and δ-C (A) and δ-N (B) (mg/kg dry weight). All species and habitat types included in regression line.

deviation: 0.14 mg/kg), kidney concentrations varied more than 2 orders of magnitude (range: 0.0814.4 mg/kg dry weight, mean 4.0, standard deviation: 3.1 mg/kg). Cadmium concentrations were significantly lower in kidneys of the common vole in comparison to the other species (ANOVA, p < 0.001). Kidney concentrations are similar to the ones found in small mammals from Dutch river floodplains,7,15 although in the latter study the concentrations in the bank vole were somewhat lower. When combining all species, log([Cd]kidney) was positively related to δ-13C, and negativelyrelated to δ-15N (GLM, p < 0.001, Figure 4). This is most likely related to a relative high proportion of earthworm in the diet, a diet item with relatively high Cd concentrations7, and a relatively high δ-13C but low δ-15N signal (Figure 3). Habitat type, log([Cd]soil), and pH were dropped from the analyses because their effects were not significant. Average [Cd]kidney in the wood mouse was highest at the meadow locations in comparison to forest and arable land, resulting in high ratios between [Cd]kidney and [Cd]soil for that species in the meadow (Table 3). This was similar for the bank vole, which indicates that accumulation patterns differed among 7500

dx.doi.org/10.1021/es200872p |Environ. Sci. Technol. 2011, 45, 7497–7502

Environmental Science & Technology

ARTICLE

Table 3. Ratios between [Cd]kidney and [Cd]soil for Different Species in Different Habitatsa wood mouse

bank vole

common vole

arable

87 (n = 10) a; B

7 (n = 1) ; 

9 (n = 7) ; A

forest meadow

191 (n = 15) a; A 1384 (n = 5) b; A

380 (n = 9) a; A 1086 (n = 3) b; A

 (n = 0) ;   (n = 0) ; 

a

Within habitats (rows), species with different capital characters are significantly different. Within species (columns), habitats with different lower case characters are significantly different. ANOVA, p < 0.05 (: no statistical analysis possible due to low number or lack of observations)

habitats. Within the arable habitat, the accumulation of Cd is significantly higher in the wood mouse in comparison to the common vole (Table 3), indicating additional species specific accumulation patterns, irrespective of [Cd]soil. In meadows with high Cd accumulation, the δ-15N signal in wood mice and bank voles was significantly lower than in the other habitats. Similarly, the δ-15N signal in wood mice from arable was also lower than in the common vole, which was also related to an increased accumulation of cadmium. Apparently, cadmium accumulation was most efficient on a diet with low δ-15N, i.e., with grass seeds or earthworm (Figure 3). Since earthworm generally contain much higher [Cd] in comparison to grass seeds,7 it is likely that effects of the decrease in δ-15N in the meadow on the increase in [Cd]kidney was caused by an increase of earthworms in the diet. However, it should be noted that the current study was not focused on a full analysis of the diet of the different species, but merely uses stable isotopes as a proxy of the diet, with reference to samples of diet items. Hence, the focus was on a few general food items (acorns, vegetation, and seeds24,25) and on an item with relatively high Cd concentrations (earthworms). But, for instance, insects and fruits and berries were not included, although these items may also be important in the diets.25 The fact that δ-15N relates well with accumulation efficiency does imply an effect of diet composition on the accumulation of cadmium. However, the conclusion on which diet item causes this relationship may not be conclusive without further study. Additionally, the pH is low in the soils from the forest locations (3.9 in forest compared to 4.4 in meadow and 5.8 in arable land), which may cause the availability of Cd to increase,10,38 resulting in a relatively high ratio between [Cd]soil and [Cd]kidney in forest habitat. When analyzing data for the individual species, different factors affect [Cd]kidney in a species specific way. [Cd]kidney in common voles was significantly lower than in the other two species, and significantly negatively related to log([Cd]soil) and to pH (ANOVA for both: p = 0.010, 72% explained variance) but not to either δ-13C or δ-15N. This species was caught at only two trap locations, with different [Cd]soil and pH. Because [Cd]soil and pH were confounded, it is impossible to separate these factors statistically. Since a negative relationship between [Cd]kidney and pH can be explained by an increase of the bioavailability of Cd at lower pH,39 it is likely that this was the most important factor regulating [Cd]kidney in common voles. In case of the wood mouse, δ-13C and δ-15N were significantly related to [Cd]kidney (Multiple linear regression: δ-13C: p = 0.002, δ-15N: p < 0.001, percentage variance accounted for 55%), indicating a major influence of the diet on the accumulation patterns in this species. For the bank vole, [Cd]kidney was significantly related to δ-15N and pH (multiple linear regression, pH: = p 0.017, δ-15N: p < 0.001), indicating influence of both diet and soil properties. The results show that [Cd]soil or soil properties are not indicative for

kidney concentrations in mobile species like the wood mouse. This species spatially integrates cadmium over a larger area, and consumes a wider range of diet, which varies among habitats. In case of the bank vole, an intermediate mobile species with a similar range in diet to the wood mouse, both diet and pH are of importance. This indicates that for this less mobile species, local soil properties affect [Cd]kidney, combined with the habitat specific diet. The common vole is least mobile, and shows a very narrow range in diet. For this species, local soil properties (likely pH) were governing [Cd]kidney. However, for this species samples were only analyzed from animals from arable land, which may account for the narrow range in diet. This may possibly lead to an underestimation of the importance of diet in this species. However, if only wood mice from arable locations are used in the statistical analyses, just δ-15N and δ-13C are significantly related to [Cd]kidney (poverall: 0.042, pδ-N: 0.019, pδ-C: 0.038, linear regression, percentage variance accounted for: 48%), similar to the results based on full data set on wood mice. The significance of the relationship is somewhat lower in comparison to the full data set, which is related to the lower degree of freedom, the percentage accounted is similar (55% versus 48%). This illustrates the importance of diet on the accumulation of cadmium to the wood mouse, even in single habitats. The results of this study clearly illustrate that accumulation of Cd is affected by species traits in a logical manner. Local soil properties are most important for predicting Cd accumulation in the nonmobile species, while diet composition predicted Cd accumulation for the species with a variable diet. However, it should be noted that this is scale dependent, and depending on the variation in [Cd]soil and soil properties. On a larger spatial scale for instance, it has been shown that [Cd]soil may be predictive for Cd levels in kidneys and liver of small mammals,8,40 although this can be hampered by differences in availability of cadmium.7 It has also been suggested that differences in feeding behavior of animals among sites may be of importance.41 The current study experimentally validates this finding, and it may be concluded that in addition to soil concentrations and soil properties, species traits affecting the habitat use of small mammals and their diet preferences can be of prime importance when assessing exposure of small mammals to cadmium.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT Field work of this study was funded by Ministry of Economic Affairs, Agriculture and Innovation, coordinated by Bas Volkers (project BO-02-011-007). Further funding was obtained from the INSPECT project funded by the SNOWMAN network (www.snowmannetwork.com) and Ministry of Economic Affairs, Agriculture and Innovation (project KB-01-015-014-ALT). Leon de Jonge (Wageningen University) analyzed the stable isotope. We are grateful to the Dutch State Forestry Service for access to their properties and to Dick Belgers (Alterra, Wageningen UR) for his support. ’ REFERENCES (1) Blankenship, A. L.; Zwiernik, M. J.; Coady, K. K.; Kay, D. P.; Newsted, J. L.; Strause, K. D.; Park, C.; Bradley, P. W.; Neigh, A. M.; Millsap, S. D.; Jones, P. D.; Giesy, J. P. Differential accumulation of polychlorinated biphenyl congeners in the terrestrial food web of the 7501

dx.doi.org/10.1021/es200872p |Environ. Sci. Technol. 2011, 45, 7497–7502

Environmental Science & Technology Kalamazoo River superfund site, Michigan. Environ. Sci. Technol. 2005, 39 (16), 5954–5963. (2) Stansley, W.; Roscoe, D. E. The uptake and effects of lead in small mammals and frogs at a trap and skeet range. Arch. Environ. Contam. Toxicol. 1996, 30 (2), 220–226. (3) van den Brink, N. W. Soil and higher organisms: From bottomup relations to top-down monitoring. In Vital Soil, Function Value and Properties, 29th ed.; Doelman, P., Eijsackers, H. J. P., Eds.; Elsevier: Amsterdam, 2004; pp 215244. (4) Beyer, W. N.; Pattee, O. H.; Sileo, L.; Hoffman, D. J.; Mulhern, B. M. Metal contamination in wildlife living near two zinc smelters. Environ. Pollut. 1985, 638 (1), 63–86. (5) Pascoe, G. A.; Blanchet, R. J.; Linder, G. Food chain analysis of exposures and risks to wildlife at a metals-contaminated wetland. Arch. Environ. Contam. Toxicol. 1996, 30 (3), 306–318. (6) Van den Brink, N. W.; Groen, N. M.; De Jonge, J.; Bosveld, A. T. C. Ecotoxicological suitability of floodplain habitats in The Netherlands for the little owl (Athene noctua vidalli). Environ. Pollut. 2003, 122 (1), 127–134. (7) Van den Brink, N.; Lammertsma, D.; Dimmers, W.; Boerwinkel, M.-C.; van der Hout, A. Effects of soil properties on food web accumulation of heavy metals to the wood mouse (Apodemus sylvaticus). Environ. Pollut. 2010, 158 (1), 245–251. (8) Shore, R. F. Predicting cadmium, lead and fluoride levels in small mammals from soil residues and by species-species extrapolation. Environ. Pollut. 1995, 88 (3), 333–340. (9) Bradham, K. D.; Dayton, E. A.; Basta, N. T.; Schroder, J.; Payton, M.; Lanno, R. P. Effect of soil properties on lead bioavailability and toxicity to earthworms. Environ. Toxicol. Chem. 2006, 25 (3), 769–775. (10) Spurgeon, D. J.; Lofts, S.; Hankard, P. K.; Toal, M.; McLellan, D.; Fishwick, S.; Svendsen, C. Effect of pH on metal speciation and resulting metal uptake and toxicity for earthworms. Environ. Toxicol. Chem. 2006, 25 (3), 788–796. (11) Dip, R.; Stieger, C.; Deplazes, P.; Hegglin, D.; Muller, U.; Dafflon, O.; Koch, H.; Naegeli, H. Comparison of heavy metal concentrations in tissues of red foxes from adjacent urban, suburban, and rural areas. Arch. Environ. Contam. Toxicol. 2001, 40 (4), 551–556. (12) Sanchez-Chardi, A.; Marques, C. C.; Nadal, J.; da Luz Mathias, M. Metal bioaccumulation in the greater white-toothed shrew, Crocidura russula, inhabiting an abandoned pyrite mine site. Chemosphere 2007, 67 (1), 121–130. (13) Ma, W. C. Effect of soil pollution with metallic lead pellets on lead bioaccumulation and organ/body weight alterations in small mammals. Arch. Environ. Contam. Toxicol. 1989, 18 (4), 617–622. (14) Morrissey, C. A.; Elliott, J. E.; Ormerod, S. J. Diet shifts during egg laying: Implications for measuring contaminants in bird eggs. Environ. Pollut. 2010, 158 (2), 447–454. (15) Wijnhoven, S.; Leuven, R. S. E. W.; van der Velde, G.; Jungheim, G.; Koelemij, E. I.; de Vries, F. T.; Eijsackers, H. J. P.; Smits, A. J. M. Heavy-metal concentrations in small mammals from a diffusely polluted floodplain: Importance of species- and location-specific characteristics. Arch. Environ. Contam. Toxicol. 2007, 52 (4), 603–613. (16) Hope, B. K. Performing spatially and temporally explicit ecological exposure assessments involving multiple stressors. Hum. Ecol. Risk Assess. 2005, 11 (3), 539–565. (17) Canova, L.; Fasola, M. Food habits and trophic relationships of small mammals in six habitats of the northern Po plain (Italy). Mammalia 1993, 57 (2), 189–199. (18) Faber, J.; Ma, W. C. Observations on Seasonal Dynamics in Diet Composition of the Field Vole Microtus-Agrestis with Some Methodological Remarks. Acta Theriol. 1986, 31 (2741), 479–490. (19) Plesner Jensen, S. Temporal changes in food preferences of wood mice (Apodemus sylvaticus L.). Oecologia (Heidelberg) 1993, 94 (1), 76–82. (20) Schipper, A. M.; Loos, M.; Ragas, A. M. J.; Lopes, J. P. C.; Nolte, B. T.; Wijnhoven, S.; Leuven, R. S. E. W. Modeling the influence of environmental heterogeneity on heavy metal exposure concentrations for terrestrial vertebrates in river floodplains. Environ. Toxicol. Chem. 2008, 27 (4), 919–932.

ARTICLE

(21) Kooistra, L.; Leuven, R.; Nienhuis, P. H.; Wehrens, R.; Buydens, L. M. C. A procedure for incorporating spatial variability in ecological risk assessment of Dutch River floodplains. Environ. Manage. 2001, 28 (3), 359–373. (22) Loos, M.; Ragas, A. M. J.; Plasmeijer, R.; Schipper, A. M.; Hendriks, A. J. Eco-SpaCE: An object-oriented, spatially explicit model to assess the risk of multiple environmental stressors on terrestrial vertebrate populations. Sci. Total Environ. 2010, 408 (18, Sp. Iss. SI), 3908–3917. (23) Drouillard, K. G.; Norstrom, R. J. The influence of diet properties and feeding rates on PCB toxicokinetics in the ring dove. Arch. Environ. Contam. Toxicol. 2003, 44 (1), 97–106. (24) Watts, C. H. S. The foods eaten by wood mice (Apodemus sylvaticus) and bank voles (Clethrionomys glareolus) in Wytham Woods, Berkshire. J. Anim. Ecol. 1968, 37, 25–41. (25) Rogers, L. M.; Gorman, M. L. The diet of the wood mouse Apodemus sylvaticus on set-aside land. J. Zool. 1995, 235 (1), 77–83. (26) Fitzgibbon, C. D. Small mammals in farm woodlands: The effects of habitat, isolation and surrounding land-use patterns. J. Appl. Ecol. 1997, 34 (2), 530–539. (27) Tew, T. E.; Todd, I. A.; MacDonald, D. M. Arable habitat use by wood mice (Apedomus sylvaticus) 2. Microhabitat. J. Zool. 2000, 250, 305–311. (28) Ben-David, M.; Flynn, R. W.; Schell, D. M. Annual and seasonal changes in diets of martens: Evidence from stable isotope analysis. Oecologia 1997, 111 (2), 280–291. (29) Van den Brink, N. W.; Bosveld, A. T. C. Alkoxyresorufin-Odeethylase activities and polychlorinated biphenyl patterns in shrews as biomarkers in environmental risk assessments: Sensitivity and specificity. Environ. Sci. Technol. 2005, 39 (18), 7737–7343. (30) Werner, R. A.; Brand, W. A. Referencing strategies and techniques in stable istope ratio analyses. Rapid Commun. Mass Spectrom. 2001, 15, 501–519. (31) Lane, P. W. Regression analysis. In Genstat 5 Release 3: Reference Manual; Genstat 5 Committee; Oxford Science Publications: Oxford, UK, 1993; pp 359460. (32) Todd, I. A.; Tew, T. E.; Macdonald, D. W. Arable habitat use by wood mice (Apodemus sylvaticus). 1. Macrohabitat. J. Zool. 2000, 250 (3), 299–303. (33) Wijnhoven, S.; van der Velde, G.; Leuven, R.; Smits, A. J. M. Modelling recolonisation of heterogeneous river floodplains by small mammals. Hydrobiology 2006, 565, 135–152. (34) Churchfield, S.; Sheftel, B. I.; Moraleva, N. V.; Shvarts, E. A. Habitat occurrence and prey distribution of a multi-species community of shrews in the Siberian taiga. J. Zool. 1997, 241 (1), 55–71. (35) Vermeulen, F.; Van den Brink, N. W.; D’Have, H.; Mubiana, V. K.; Blust, R.; Bervoets, L.; De Coen, W. Habitat type-based bioaccumulation and risk assessment of metal and As contamination in earthworms, beetles and woodlice. Environ. Pollut. 2009, 157, 3098–3105. (36) Vander Zanden, M. J.; Rasmussen, J. B. Variation in delta15N and delta13C trophic fractionation: Implications for aquatic food web studies. Limnol. Ocean 2001, 46 (8), 2061–2066. (37) Post, D. M. Using stable istotopes to estimate trophic position: Models, methods, and assumptions. Ecology 2002, 83 (3), 703–718. (38) Ma, W. C. Estimating heavy metal accumulation in oligochaete earthworms: A meta-analysis of field data. Bull. Environ. Contam. Toxicol. 2004, 72 (4), 663–670. (39) Janssen, R. P. T.; Posthuma, L.; Baerselman, R.; Den, H. H. A.; Van, V. R. P. M.; Peijnenburg, W. J. G. M. Equilibrium partitioning of heavy metals in Dutch field soils. II. Prediction of metal accumulation in earthworms. Environ. Toxicol. Chem. 1997, 16 (12), 2479–2488. (40) Talmage, S. S.; Walton, B. T. Small mammals as monitors of environmental contaminants. Rev. Environ. Contam. Toxicol. 1991, 119, 47–145. (41) Walton, K. C. Fluoride in Bones of Small Rodents Living in Areas with Different Pollution Levels. Water Air Soil Pollut. 1987, 32 (12), 113–122.

7502

dx.doi.org/10.1021/es200872p |Environ. Sci. Technol. 2011, 45, 7497–7502