Environ. Sci. Technol. 2004, 38, 2298-2303
Biomagnification of r- and γ-Hexabromocyclododecane Isomers in a Lake Ontario Food Web G R E G G T . T O M Y , * ,†,‡ W E S B U D A K O W S K I , † THOR HALLDORSON,† D. MICHAEL WHITTLE,§ MICAHEL J. KEIR,§ CHRIS MARVIN,+ GORDIA MACINNIS,+ AND MEHRAN ALAEE+ Department of Fisheries and Oceans, Freshwater Institute, Winnipeg, Manitoba R3T 2N6, Canada, Department of Chemistry, University of Manitoba, Winnipeg, Manitoba, R3T 2N2, Canada, Department of Fisheries and Oceans, Great Lakes Laboratory for Fisheries and Aquatic Sciences Burlington, Ontario, L7R 4A6, Canada, and National Water Research Institute, Environment Canada, Burlington, Ontario, L7R 4A6 Canada
The extent of bioaccumulation of hexabromocyclododecane (HBCD) isomers (R, β, and γ) was determined in the Lake Ontario pelagic food web using liquid chromatography tandem mass spectrometry (LC/MS/MS). Concentrations of the R-isomer were consistently higher than that of the γ-isomer. The β-isomer was below method detection limits in all samples. Whole body concentrations (ng/g, wet wt) of R- and γ-HBCD were highest in the top predator lake trout samples ranging from 0.4 to 3.8 ng/g for the R-isomer and 0.1 to 0.8 ng/g for the γ-isomer. For the prey fish species, the trends in R- and γ-HBCD levels were slimy sculpin > smelt > alewife. Mean concentrations of total (Σ) HBCD (sum of R- and γ-isomers) in the macrozooplankter Mysis relicta (0.14 ( 0.02 ng/g wet wt) and in the benthic invertebrate Diporeia hoyi (0.16 ( 0.02 ng/g, wet wt) were similar and approximately twice as high as in plankton (0.06 ( 0.02 ng/g, wet wt). A strong positive linear relationship was found between ΣHBCD concentrations (wet wt) and trophic level based on δ15N suggesting that HBCD biomagnifies in the Lake Ontario food web. The trophic magnification factor (TMF ) 6.3) derived from the slope of the ΣHBCD - trophic level relationship was slightly higher than TMFs for p,p′-DDE (6.1) and ΣPCBs (5.7) found previously. Biomagnification factors (BMF, calculated as the ratio of lipid corrected concentration in predator/lipid corrected concentration in prey) were variable between feeding relationships and ranged from 0.4 to 10.8 for the R-isomer and from 0.2 to 10 for γ-isomers.
* Corresponding author phone: (204)983-5167; fax: (204)984-2403; e-mail:
[email protected]. † Freshwater Institute. ‡ University of Manitoba. § Great Lakes Laboratory for Fisheries and Aquatic Sciences Burlington. + Environment Canada. 2298
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 8, 2004
Introduction Brominated flame retardants (BFRs) are a structurally diverse group of chemicals that are incorporated into a variety of consumer and industrial products to increase their flame resistance. BFRs are either aromatic, aliphatic, or cycloaliphatic compounds containing between 50 and 85% bromine by weight (1). Estimates suggest that global demand for BFRs is on the rise; from 1990 to 2000, usage increased from approximately 145 to 310 kt (2). Hexabromocyclododecane (HBCD, C12H18Br6) is the most widely used of the cycloaliphatic BFRs. In 1999, the global demand for HBCD (∼16 kt) was more than double that of the pentabromodiphenyl ether (∼8.5 kt) mixture (3). HBCD is the principal flame retardant in extruded (XEPS) and expanded (EPS) polystyrene foams used as thermal insulation in the building industry (4). Secondary uses of HBCD are in upholstery textiles including residential and commercial furniture, upholstery seating in transportation, draperies, and wall coverings (3, 4). Like the brominated diphenyl ethers (BDEs), HBCD is an additive type BFR. Thus, HBCD is not covalently bonded to the material leading to the risk of migration out of the product during use or disposal. Technical HBCD is industrially synthesized by the addition of bromine to cis-trans-trans-1,5,9-cyclododecatriene (4, 5). The resulting technical mixture contains three diastereoisomers (R, β, and γ, Figure 1) and tetrabromocyclododecene as an impurity (70%), followed by the R- and then β-isomer. The selection of HBCD grade used depends on the usage of the end product. The physical-chemical properties of HBCD are similar to those of BDEs (7) and other persistent organic pollutants, including polychlorinated biphenyls (PCBs), which are known to be persistent and bioaccumulative (3, 8). In fact, estimates of the log Kow of HBCD (5.6) places it in the optimum range for bioaccumulation (9). Despite these properties and its widespread use, there is little known about the fate and environmental levels of HBCD. Sellstro¨m et al. reported microgram/gram levels of HBCD in sediment and pike downstream of textile manufacturing industries from the River Viskan in Sweden (8). Bioavailability of ΣHBCD, calculated as the ratio of the concentration in the fish (corrected for lipid content) to the concentration in the sediment (corrected for carbon content), was 15, which was similar to BDE-47 (6.6-19) and BDE-99 (17) (8). Sublethal effects of HBCD were examined in juvenile rainbow trout (Oncorhynchus mykiss) by intraperitoneal injection (10). After 5 days, HBCD was found to induce catalase activity, and an increase in liver somatic index was observed after 28 days. These changes are indicative of peroxisome proliferators. HBCD also had an antagonistic effect on CYP1A. In an in vitro study using mammalian cells, HBCD was linked to carcinogenesis by inducing genetic recombination (11). To our knowledge, no attempt has been made to quantify the extent of bioaccumulation of HBCD in an aquatic food web. The objectives of this study were to examine the bioaccumulation and biomagnification of the three isomers of HBCD in a pelagic food web from Lake Ontario by a recently developed isomer specific method based on liquid chromatography tandem mass spectrometry (LC/MS/MS) (12). 10.1021/es034968h CCC: $27.50
2004 American Chemical Society Published on Web 03/04/2004
FIGURE 1. Structures of the r-, RR,SR,RS (left), β-, RR,SR,SR (middle), and γ-, RR,RS,SR (right) HBCD isomers.
Materials and Methods Chemicals. All three native (R, β, and γ) and labeled [deuterated (d18-γ-HBCD) and carbon-13 mass labeled (13C12γ-HBCD)] HBCD isomers (each >98% purity) were provided by Wellington Laboratories, Inc. (Guelph, ON, Canada). The synthesis of the labeled isomers has been described by Arsenault et al. (13). HPLC Optima grade methanol, water, and acetonitrile were obtained from Fisher Scientific (Nepean, ON, Canada). Samples. Lake trout (Salvelinus namaycush, whole fish, n ) 5), a top predator fish species, and alewife (Alosa pseudoharengus, composites of 5 fish, n ) 3), rainbow smelt (Osmerus mordax, composites of 5 or 20 fish, n ) 3), slimy sculpin (Cottus cognatus, composites of 10 or 15 fish, n ) 3), forage fish species, and major diet items of lake trout were collected between June and September 2002 at offshore stations in Lake Ontario. Adult lake trout were collected via overnight sets of nylon gillnet with stretched mesh sizes of 7.5-11.5 cm (equivalent to commercial mesh sizes of 3.04.5 in.) in 30 to 40 m of water. Forage fish species were collected with a modified 9 m Biloxi bottom trawl towed at depths of 80-100 m. All fish samples were stored whole on dry ice immediately after collection. Invertebrate samples including mysids (Mysis relicta, composites of >100 individuals, n ) 2) and amphipods (Diporeia hoyi, composites of >100 individuals, n ) 2) were collected with a modified epibenthic sled towed a depths of 50-80 m at the same general offshore sites where fish were collected. Net plankton grabs (n ) 2), containing mainly zooplankton fauna, copepods, and cladocerans, hereafter referred to as plankton, were collected with 0.5 m diameter Wisconsin type nets constructed of 153 µm Nitex mesh that were towed horizontally at a subsurface depth of 1 m. Invertebrate samples were sorted immediately after collection and stored in solvent rinsed containers on dry ice. The invertebrates (plankton, Mysis, and Diporeia) and forage fish (alewife, smelt, and sculpin) were processed as composites of whole individuals, whereas all lake trout were individual whole fish. Further details on samples are given in the Supporting Information. Analytical Method. Extraction and detection of HBCD in environmental samples has been described in detail elsewhere (12). In brief, samples were extracted by accelerated solvent extraction (ASE) and cleaned up using GPC (14) and Florisil. Prior to GPC cleanup, lipid content was determined gravimetrically from an aliquot of the extract. Sample sizes ranged from 8 to 15 g wet wt for the various fish species and 20 g wet wt for the invertebrate samples. Separations were performed on a C18 analytical column (5.0 cm × 2.1 mm i.d., 4 µm particle size) at a flow-rate of 300 µL/min. LC/MS/MS analyses were performed on a Sciex API 2000 (MDS Sciex, ON, Canada) using electrospray ionization (ES) in the negative ion mode. MS/MS detection used multiple reaction monitoring (MRM) conditions for the m/z 640.6 ([M - H]-) f Brreaction (both isotopes), utilizing unit resolution on the first and third quadrupoles and a 200 ms dwell time. Quantitation was based on the ion signal from the m/z 640.6 f m/z 79 MRM transition. Collision activated dissociation gas pressure was 8 au, and the collision energy was -50 eV. QA/QC. Procedural blanks were analyzed for each batch of samples. Recoveries were determined by addition of d18-
γ-HBCD and 13C12-γ-HBCD at the point of sample extraction. MS/MS detection of d18-γ-HBCD and 13C12-γ-HBCD was based on the analogous ([M - H(D)] -) f Br- reaction monitored for the native HBCD [m/z d18: 657.6 ([M - (D)] 13 - and C: 652.4 ([M - H] )]. All results were both blank and recovery corrected. In addition, duplicate samples of sculpin and alewife were processed to verify repeatability of the analytical method. Method detection limits (MDLs) (15) were determined by spiking known amounts of R- and γ-HBCD into an ASE cell (n ) 4) packed with 10 g of hydromatrix and cleaned up in the same manner as samples. Separate injections of the extracts were then made. The ion signals obtained for both R- and γ-HBCD were then adjusted to estimate concentrations that would give a signal-to-noise ratio of 3:1. Food Web and Biomagnification Factor Calculations. Two types of trophic transfer terms were calculated for the Lake Ontario food web. The first determined trophic magnification factors (TMFs) for the entire food web based on the relationship between δ15N and contaminant concentration:
ln ΣHBCD concentration (wet wt) ) a + (b × trophic level) (1) Converting δ15N to trophic level (TL) was done using the relationship (16-18)
TL ) 1 + (δ15Nconsumer - δ15Nplankton/3.8)
(2)
The slope b of eq 1 was used to calculate TMF using
TMF ) eb
(3)
TMFs around zero imply that the chemical is moving through the food chain without being biomagnified, whereas a TMF of >1 indicates that a chemical is biomagnifying (16, 17). Negative values indicate that a chemical is not taken up by the organism or is metabolized. The second method determined biomagnification factors (BMFs) for individual species using
BMF ) [predator]/[prey]
(4)
where [predator] and [prey] are the lipid corrected concentrations in the predator and prey species, respectively.
Results and Discussion The average recoveries of d18-γ-HBCD and 13C12-γ-HBCD in samples were 64.8 ( 11.6 and 83.5 ( 17.5%, respectively. Concentrations in biota were all blank and recovery corrected based on the mean recovery of d18-γ-HBCD and 13C12-γ-HBCD in each sample. Respective MDLs for R- and γ-HBCD were 17 and 6.7 pg. The MDL for the β-isomer was estimated to be 30 pg based on its response factor. Duplicates of alewife and sculpin, analyzed to check for repeatability of the method, were within >80% of each other for both R- and γ-HBCD isomers. R- and γ-HBCD were detected in samples from all three trophic levels of the Lake Ontario pelagic food web. The β-isomer was below method detection limits for all species. The ion chromatograms of R- and γ-HBCD in selected biota samples from Lake Ontario are shown in Figure 2, and whole body concentrations (ng/g, wet wt) of R- and γ-HBCD in biota are shown in Figure 3. R- and γ-HBCD levels were highest in the lake trout ranging from 0.4 to 3.8 ng/g (wet wt) for the R-isomer and 0.1 to 0.7 ng/g (wet wt) for the γ-isomer. Plankton had the lowest level of the R-isomer [0.03 ( 0.01 ng/g (wet wt): arithmetic mean ( standard deviation], while Mysis had the lowest concentration of the γ-isomer (0.010 VOL. 38, NO. 8, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2299
2300
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 8, 2004
FIGURE 2. Elution profiles of r- and γ-HBCD isomers in lake trout, smelt, alewife, and Mysis.
FIGURE 3. Blank and recovery corrected mean concentrations (ng/ g, wet wt) (( standard error) of r- and γ-HBCD in a Lake Ontario food web. ( 0.01 ng/g, wet wt). For the forage fish, the trends in R- and γ-HBCD levels were slimy sculpin > smelt > alewife. Because of their benthic association and higher lipid content, it is not surprising that the levels of HBCDs are highest in the sculpin [R: 0.1-0.4 and γ: 0.02-0.2 ng/g (wet wt)] as compared to the other two forage fish species. In contrast, alewife, which is a planktivore species, had the lowest concentrations of both R-HBCD [mean: 0.11 ( 0.03 ng/g (wet wt)] and γ-HBCD [mean: 0.015 ( 0.002 ng/g (wet wt)]. Mean concentrations of the R- and γ-HBCD isomers in Mysis and Diporeia were similar. It is clear that there are differences in the relative abundance of the R- and γ-isomers among species. This can be expressed numerically as the fraction of R-isomer in the samples: concentration of R-isomer divided by the sum of the concentrations of R- and γ-isomers. For example, the fraction of R-HBCD in plankton, Mysis, and Diporeia are 0.49, 0.78, and 0.69, respectively. For the forage fish, respective fractional amounts of the R-isomer are 0.88, 0.76, and 0.86 for alewife, sculpin, and smelt. In lake trout, the R-isomer accounts for 82% of the total. Comparing the fractional amounts of the R-isomer for the Lake Ontario samples to that of technical mixtures might be useful in tracking sources of HBCD into the environment.
However, this is complicated by the fact that the isomeric composition of the technical HBCD mixture, although wellcharacterized, changes during application of the flame retardant (5, 6). For example, when the technical mixture is incorporated into XEPS or EPS, temperatures above 160 °C are employed, and this effectively changes the abundances of the R- and γ-isomers. The thermal rearrangement or isomerization of HBCD isomers was examined by Peled et al. (6). Interestingly, they found that between 160 and 200 °C the R-isomer becomes the most predominant isomer (∼78%). This result was also independent of the isomer ratio of the starting material. So whereas the γ-isomer is dominant in the technical mixture at room temperature, during elevated temperatures (such as those present during XEPS and EPS processing), a rearrangement occurs resulting in a predominance of the R-isomer (6). It is also conceivable that small changes in the temperature of the treatment procedure from one manufacturer to the next will also affect the proportions of the isomers present in the end product. In addition to exposure to materials containing different proportions of HBCD isomers, other factors that might contribute to differences in the concentrations of HBCD isomers in biota include ecological differences (i.e., trophic level), differences in the physical chemical properties among isomers (e.g., Kow), bioaccumulation parameters (e.g., uptake rates, half-lives), exposure to an environmentally weathered product, source of exposure, direct exposure to the technical mixture, and selective environmental transformation reactions (e.g., metabolism, biotransformation). Because of the variability of HBCD isomer compositions present in technical mixtures and the myriad of processes that can affect isomer compositions in the environment, fingerprinting sources of HBCD into the aquatic environment is not likely to be trivial. Trophic Magnification Factors (TMFs). The trophic status of the Lake Ontario pelagic food web has been elucidated previously using stable isotopes (19). The relative trophic status, as defined by δ15N is plankton f mysids and amphipods (benthic inverterbrates) f alewife, smelt, slimy sculpin (prey species) f lake trout (top predator). Although the samples analyzed in this study were collected at a different time to that of Kiriluk et al. (19) study, the food web δ15N values still provide a reasonable estimate of the trophic status of the lake. VOL. 38, NO. 8, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2301
FIGURE 4. Mean (( standard error) ΣHBCD (r + γ) concentrations (ng/g, wet wt) - trophic level relationship for the Lake Ontario food web. Regression analysis: ln[ΣHBCD] ) -9.070 + 1.843 (TL) (r2 ) 0.7192, p < 0.0001). The δ15N values were taken from Kiriluk et al. (19). Plankton was not included in the plot because of the large variation in the δ15N values (19).
TABLE 1. Lipid normalized BMFs for r- and γ-HBCD and p,p′-DDE and ΣPCBs in a Lake Ontario Food Weba predator/prey
r-
γ-
p,p-DDEb
ΣPCBsb
trout/alewife trout/smelt trout/sculpin sculpin/Diporeia sculpin/Mysis smelt/Mysis smelt/Diporeia alewife/plankton
4.8 1.0 1.1 3.5 9.7 10.8 4.0 0.4
7.5 1.5 0.8 2.5 9.9 5.5 1.4 0.2
3.3 1.9 2.6 2.1-5.5 2.8-3.2 3.8-4.4 2.9-3.7 8.9-11.2
3.1 2.1 1.6 3.7-5.9 3.6-5.3 2.7-4.0 2.8-4.4 3.7-8.7
a Lipid normalized BMFs for p,p′-DDE and ΣPCBs were calculated from concentrations reported by Kiriluk et al. (19). b Concentrations of p,p′-DDE and ΣPCBs in samples were taken from Kiriluk et al. (19). Mean ΣPCB and p,p′-DDE concentrations were used in the calculation and were lipid normalized using the reported mean lipid % from the Kiriluk et al. study. Lipid normalized BMFs were calculated based on the ratio of the concentration in the predator to concentrations in the prey.
A highly significant relationship (p < 0.0001) was found from the plot of natural log concentration of ΣHBCD (sum of R- and γ-HBCD) versus trophic level values for the Lake Ontario food web (Figure 4). Plankton was purposely omitted from the plot because of the large variation in the δ15N values reported by Kiriluk et al. (19). We also chose to plot total HBCD instead of individual HBCD isomers because it is still unclear whether isomerization or biotransformation takes place in biota. If isomerization or biotransformation does occur, this would skew the relationship between individual isomer concentration and trophic level. From the slope of Figure 4, the TMF for ΣHBCD was calculated to be 6.3. For comparison, mean p,p′-DDE and ΣPCBs concentrations taken from Kiriluk et al. were plotted in the same manner. Calculated TMFs for p,p′-DDE (6.1) and ΣPCBs (5.7) suggests that HBCD biomagnifies to the same extent as these persistent organochlorines (19). Predator/Prey Biomagnification Factors (BMFs). Lipid normalized BMFs for R- and γ-isomers found from this study are shown in Table 1. BMFs of p,p′-DDE and ΣPCBs from Kiriluk et al. (19) are also listed for comparison. For most feeding relationships, BMFs for both isomers were greater than 1, suggesting biomagnification between trophic levels. The highest BMFs for both isomers were found for the sculpin and smelt to Mysis predator/prey relationship, and this might 2302
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 8, 2004
be driven by the low lipid normalized HBCD concentrations in Mysis. High BMFs were also observed between lake trout to alewife for both isomers. Alewife currently make up >90% of the diet of adult lake trout from Lake Ontario. For this predator-prey, the γ-isomer was approximately 1.5 times higher than the R-isomer. It is not clear at this stage if this arises because of the low levels of the γ-isomer in alewife or because of the preferential biotransformation or excretion of the R-isomer in the lake trout. When comparing BMFs for p,p′-DDE and ΣPCBs to Rand γ-HBCD isomers, the largest discrepancy was observed between the alewife to plankton feeding relationship. A combination of the low lipid normalized concentrations of R- and γ-isomers in alewife (as compared to the other forage fish species) and the high lipid normalized concentrations of HBCD in plankton probably contributed to this difference. Future Research. It is clear that there are discrepancies in the isomer burden of HBCD in biota and that these differences can arise in part by the combination of many environmental processes. Identifying pathways that lead to these changes in the field are complicated, and the environmental behavior of isomers may best be gleaned from laboratory studies. Work is underway to examine the bioaccumulation parameters of individual isomers in fish to try and understand the behavior and fate of isomers and to test the possibility of biotransformation and isomerization. Debromination of brominated diphenyl ethers (BDEs) and cytochrome P450 monooxygenase mediated metabolism of BDEs to their hydroxylated analogues has already been reported in the literature (20-22). The R- and γ-isomers have been previously detected in beluga blubber samples from the Canadian Arctic (G. Tomy, unpublished data). This implies that these isomers have physical-chemical properties that make them amenable to long-range transport. It is well-known that the behavior of R- and β-HCH in the environment is directly linked to differences in the physicochemical properties. In light of this, measurements of the physical-chemical properties of HBCD isomers have begun. Work is also warranted on the possibility of microbial mediated transformations in the sediment compartment. Past studies have shown that isomerization of γ-HCH to R-HCH in aquatic sediments can occur (23). Last, transformation of the isomers in the air and water compartments should also be considered.
Acknowledgments We thank Brock Chittim and Gilles Arsenault (Wellington Laboratories Inc., Guelph, Canada) for the native and labeled HBCD congeners. Collin Allchin (Centre for Environment, Fisheries and Aquaculture Science, U.K.), Sheryl Tittlemier (Health Canada, Ottawa, Canada), and Aaron Fisk (University of Georgia, Athens, GA) are thanked for their helpful comments on an earlier version of the manuscript. The manuscript benefited greatly from the comments of two anonymous reviewers.
Supporting Information Available A table with information on the samples and concentrations of R- and γ-HBCD. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Younes, M. Proceedings, 2nd International Workshop on Brominated Flame Retardants; Stockholm, Sweden, May 14-16, 2001, 21-22. (2) Arias, P. A. Proceedings, 2nd International Workshop on Brominated Flame Retardants; Stockholm, Sweden, May 14-16, 2001, 17-19.
(3) de Wit, C. A. Chemosphere 2002, 46, 583. (4) American Chemistry Council. Data Summary and test plan for hexabromocyclododecane (HBCD). AR201-13459A. (5) Reyes, J. D.; Scheinert, J.; Georlette, P. HBCD: Advancing performance through innovation. Recent Adv. Flame Retard. Polymer Mat. 1997, 8, 390. (6) Peled, M.; Scharia, R.; Sondack, D. Thermal rearrangement of hexabromocyclododecane (HBCD). Adv. Organobrom. Chem. II 1995, 92. (7) Tittlemier, S. A.; Halldorson, T. H. J.; Stern, G. A.; Tomy, G. T. Environ. Toxicol. Chem. 2002, 21, 1804. (8) Sellstro¨m, U.; Kierkegaard, A.; de Wit, C.; Jansson, B. Environ. Toxicol. Chem. 1998, 17, 1065. (9) MacGregor, J. A.; Nixon, W. B. Hexabromocyclododecane (HBCD): Determination of n-octanol/water partition coefficient. Brominated Flame Retardant Industry Panel: Arlington, VA, 1997. (10) Ronisz, D.; Farmen Finne, E.; Karlsson, H.; Fo¨rlin, L. Proceedings, 2nd International Workshop on Brominated Flame Retardants; Stockholm, Sweden, May 14-16, 2001, 271. (11) Helleday, T.; Tuominen, K.-L.; Bergman, Å.; Jenssen, D. Mut. Res. 1999, 439, 137. (12) Budakowski, W. R.; Tomy, G. T. Rapid Commun. Mass Spec. 2003, 17, 1399. (13) Arsenault, G.; Chittim, B.; Halldorson, T.; Konstantinov, A.; McAlees, A.; McCrindle, R.; Tomy, G.; Yeo, B. Organohal. Compds. 2003, 61, 267.
(14) Stalling, D. L.; Tindle, R. C.; Johnson, J. L. J.sAssoc. Off. Anal. Chem. 1972, 55, 32. (15) Winefordner, J. D.; Long, G. L. Anal. Chem. 1983, 55, 712A. (16) Broman, D.; Na¨f, C.; Rolff, C.; Zebu ¨ hr, Y.; Fry, B.; Hobbie, J. Environ. Toxicol. Chem. 1992, 11, 331. (17) Fisk, A. T.; Hobson, K. A.; Norstrom, R. J. Environ. Sci. Technol. 2001, 35, 732. (18) Hobson, K. A.; Welch, H. E. Environ. Sci. Technol. 1992, 26, 9-18. (19) Kiriluk, R. M.; Servos, M. R.; Whittle, D. M.; Gilbert, C.; Rasmussen, J. B. Can. J. Fish. Aquat. Sci. 1995, 52, 2660. (20) Tomy, G. T.; Palace, V. P.; Halldorson, T.; Braekevelt, E.; Danell, R.; Wautier, K.; Evans, B.; Brinkworth, L.; Fisk, A. T. Environ. Sci. Technol., in press. (21) Stapleton, H. M.; Alaee, M.; Letcher, R. J.; Baker, J. Environ. Sci. Technol. 2004, 38, 112. (22) Asplund, L. T.; Athanasiadou, M.; Andreas, A.; Bergman, A° .; Bo¨rgeson, H. Ambio 1999, 28, 67. (23) Benezet, H. J.; Matsumuar, F. Nature 1973, 243, 480.
Received for review September 4, 2003. Revised manuscript received January 28, 2004. Accepted February 2, 2004. ES034968H
VOL. 38, NO. 8, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2303
