Environ. Sci. Technol. 2006, 40, 1439-1446
Mercury Bioaccumulation in Forage Fish Communities Invaded by Rainbow Smelt (Osmerus mordax) H E I D I K . S W A N S O N , * ,† T H O M A S A . J O H N S T O N , ‡,| DAVID W. SCHINDLER,† R. ANDREW BODALY,§ AND D. MICHAEL WHITTLE‡ Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6G 2E9, Canada, Department of Fisheries and Oceans, Great Lakes Laboratory for Fisheries and Aquatic Sciences, Burlington, Ontario L7R 4A6, Canada, and Department of Fisheries and Oceans, Freshwater Institute, Winnipeg, Manitoba, R3T 2N6, Canada
We compared total mercury concentrations ([Hg]) among 6 forage fish species in 25 central Canadian lakes and related [Hg] to adjusted-δ15N (an index of trophic position), δ13C, growth rate, and a suite of environmental variables. Growth rates were also compared among species and related to environmental variables. We found that rainbow smelt (Osmerus mordax), a recent invader in many of these lakes, had intermediate [Hg] and growth rates relative to other species. Forage fish growth rates differed significantly among species and were related to latitude (inverse relationship) and lake conductivity (positive relationship). Mercury concentrations also differed significantly among species and the strongest predictors were growth rate and lake conductivity; [Hg] was significantly and negatively related to both. Adjusted-δ15N explained very little variation in [Hg] and was significant only when the analysis was restricted to biotic variables. These results indicate that biomagnification may not be observed at fine scales of trophic differentiation and that rainbow smelt are unlikely to cause post-invasion [Hg] increases in most predatory fish species.
Introduction Rainbow smelt are anadromous forage fish that are native to coastal regions of North America and isolated freshwater lakes of the St. Lawrence drainage system (1). The species’ freshwater range has recently expanded due to a series of introductions through the Laurentian Great Lakes and Hudson Bay drainage systems (2, 3). There are now naturalized populations in lakes throughout Ontario and the Nelson River system in Manitoba. Previous studies have reported that rainbow smelt invasion causes increased contaminant concentrations in predatory fish (e.g., lake trout, walleye, northern pike) by length* Corresponding author phone: 780-436-1996; fax: 506-648-5811; e-mail:
[email protected]. † University of Alberta. ‡ Great Lakes Laboratory for Fisheries and Aquatic Sciences. § Freshwater Institute. | Present address: Ontario Ministry of Natural Resources, Cooperative Freshwater Ecology Unit, Laurentian University, Sudbury, Ontario P3E 2C6, Canada. 10.1021/es0510156 CCC: $33.50 Published on Web 01/21/2006
2006 American Chemical Society
ening the food chain to these species (4, 5). Empirical evidence for rainbow smelt-induced increases in contaminant concentrations is equivocal, however. A study conducted on oligotrophic lakes in Ontario and Que´bec found that smelt had significantly higher trophic positions and mercury concentrations ([Hg]) than other forage fish and that lake trout in smelt-invaded lakes had significantly higher trophic positions and [Hg] than those in reference lakes (4). In contrast, a study conducted in the recently invaded Hudson Bay drainage found that although smelt were trophically elevated relative to other forage fish species, they had relatively low [Hg] (6). As well, although the invasion did have a positive and significant effect on predator trophic position in these lakes, the effect on [Hg] was not statistically significant (7). The fact that [Hg] was not related to trophic position in rainbow smelt-invaded forage fish communities (6) seems to conflict with studies of biomagnification that predict a positive relationship between [Hg] and trophic position (8, 9). Most studies that have documented and quantified biomagnification, however, have examined multiple trophic levels (e.g., from primary consumers to top predators). It is possible that at fine scales of trophic differentiation, such as within the trophic guild of “forage fishes,” factors other than trophic position determine inter-species differences in [Hg]. Previous authors have argued that longevity and growth rate may be as important in explaining fish [Hg] as biomagnification from prey (10, 11). The relationship between [Hg] and growth rate is often inverse because exposure times to contaminants are limited in fast-growing, shorter-lived fish (10, 11). Also, growth dilution may occur; fast-growing fish may dilute their contaminant burdens by producing more flesh per unit energy and contaminant intake than slowgrowing fish (12-15). Anticipating the effects of ecosystem stressors (e.g., contaminant inputs, climate change, species additions/ deletions) on freshwater lakes often involves predictions of contaminant concentrations with food web data. As illustrated by Swanson et al. (6), this may not be appropriate in study systems where overall trophic variation and/or disruption is small. In this study, we attempt to determine the best predictors of forage fish [Hg]. We examine interspecies differences in forage fish [Hg] and growth rate, and relate [Hg] to a suite of abiotic (latitude, lake conductivity, pH, maximum depth, and lake area) and biotic (growth rate, δ13C, and adjusted-δ15N (an index of trophic position)) variables. On the basis of previous results that illustrated decoupling of trophic position and [Hg] in forage fish communities (6), we hypothesized that growth rate and water chemistry would be stronger predictors of forage fish [Hg] than adjusted-δ15N. If true, this would explain how rainbow smelt invasion could lengthen the food chain to predatory fish without affecting parallel increases in predator [Hg].
Materials and Methods Field Sampling. Twenty-five lakes were sampled for forage fishes (defined as fishes 0.05. As well, analyses were performed to assess whether changing the body size of comparison (within the range of captured body sizes common to all species) would affect the results of inter-species comparisons. Among-lake differences in raw (not LSmean) [Hg] were determined for each species with one-way ANOVA. Species-specific residuals were then regressed against loge wet mass and slopes and intercepts of the relationships were compared among species. A different procedure was necessary for growth rate because wet mass is part of the absolute growth rate calculation and within an age class there is a 1:1 relationship between growth rate and wet mass. Amonglake differences in loge wet mass were therefore determined for each species with one-way ANOVA and species-specific residuals were regressed against age. Mass-at-age regressions were then compared among species.
Results [Hg] and Growth Rate Comparisons Among Species. Before standardization for body size or calculations of means, total [Hg] ranged from 0.05 µg/g (rainbow smelt and round goby in the Great Lakes) to 1.81 µg/g (emerald shiner in Lake Minnitaki). Emerald shiner and spottail shiner had the highest mean [Hg], while round goby had the lowest mean [Hg] (Table 2). Least-squares means [Hg] varied significantly among the six species analyzed statistically (ANOVA, F ) 11.84, P < 0.0001, R2 ) 0.75, df ) 5,53; Figure 2). Emerald shiner had significantly higher [Hg] than all other species and spottail shiner had significantly higher [Hg] than cisco, rainbow smelt, trout-perch, and yellow perch (Tukey’s test, P < 0.05; Figure 2). Mean growth rates of slimy sculpin, spottail shiner, and emerald shiner were lowest, while those of cisco and yellow perch were highest (Table 2). Of the six species analyzed statistically, there were significant differences among LSmean growth rates (ANOVA, F ) 25.01, P < 0.0001, R2 ) 0.79, df VOL. 40, NO. 5, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1441
TABLE 2. Mean [Hg] and Growth Rates for All Species Captureda
species
mean total [Hg] (µg/g dry mass)b
standard deviation
Alewife Bluegill Cisco Common shiner Emerald shiner Golden shiner Logperch Round goby Pumpkinseed Rainbow smelt Rock bass Slimy sculpin Spottail shiner Trout-perch Yellow perch
0.20 0.26 0.26 0.37 0.86 0.34 0.32 0.07 0.32 0.30 0.26 0.18 0.44 0.29 0.30
0.27 0.23 0.13 0.29
0.15 0.17 0.03 0.18 0.17 0.12
mean absolute growth rates (g/year) 3.94 5.02 7.29 3.31 2.32 5.93 5.02 3.63 6.17 3.63 5.55 2.10 2.79 4.14 7.31
standard deviation 0.36 0.66 0.31 0.24
0.13 0.25 0.21 0.16 0.30 0.31
N (number of lakes) 2 1 7 5 6 1 1 1 3 24 1 2 12 16 20
a For species captured in more than one lake, means of lake-specific least-squares means were calculated (standardized mass of 8 g). For species captured in only one lake, least-squares means (standardized mass of 8 g) are presented for that lake. b Mean % moisture was 80% for all species.
FIGURE 2. Least-squares mean (standardized mass of 8 g) total [Hg] ( SE for the six species analyzed statistically. Emerald shiner had significantly higher [Hg] than all other species, and spottail shiner had significantly higher [Hg] than cisco, rainbow smelt, trout-perch, and yellow perch (Tukey’s test, P < 0.05). Letters indicate significant pairwise differences.
FIGURE 3. Least-squares mean (standardized mass ) 8 g) loge growth rates ( SE for the six species analyzed statistically. Yellow perch and cisco had significantly higher growth rates than all other species (Tukey’s test, P < 0.05) and emerald shiner had significantly lower growth rates than all other species except spottail shiner (Tukey’s test, P < 0.05). Letters indicate significant pairwise differences.
) 5,53; Figure 3). Yellow perch and cisco had significantly higher growth rates than all other species (Tukey’s test, P < 0.05), and emerald shiner had significantly lower growth rates than all other species (Tukey’s test, P < 0.05) except spottail shiner. Inter-species comparisons of [Hg] and growth rate were generally consistent across the range of captured body sizes common to all species (Figure 4 a and b). Regardless of body size, emerald shiner had higher [Hg] than all other species and yellow perch and cisco had higher mass-at-age than all other species (Figure 4 a and b). The only pairwise comparisons of [Hg] and growth rate that changed with body size were those that involved cisco (Figure 4 a and b). Slopes of [Hg]-body size relationships and wet mass-age relationships differed significantly among species (ANCOVA, F > 3.78, P > 0.0022). Relationships Between Growth Rate and Environmental Variables. Within species, growth rate differed significantly among lakes (ANCOVA, 1.9 < F 3.78, P < 0.0022, df ) 5,741; 5,766).
TABLE 3. Summary of Pearson Product-Moment Correlations between Intraspecific Growth Rates and Lake Characteristics (abiotic variables) in the Six Species Analyzed Statisticallya species
variable
correlation coefficient
P-value
Cisco Emerald shiner Rainbow smelt Spottail shiner Trout-perch Yellow perch
pH maximum depth conductivity latitude conductivity latitude
0.59 0.43 0.42 -0.64 0.69 -0.52
0.410 0.473 0.083 0.062 0.013* 0.025*
a
The variables presented are those that were most closely related to intraspecific loge growth rates. If nothing was significant, the variable with the lowest P-value is presented. All correlations were based on least-squares means calculated at a standardized mass of 8 g. *Relationships were significant at R ) 0.05.
and negatively related to [Hg] (multiple regression, F ) 11.66, P < 0.0001, R2 ) 0.29, df ) 3,77). Analyses of residual variance revealed, however, that of the 29% of variation accounted for by biotic variables, growth rate was responsible for 22% while adjusted-δ15N and δ13C each accounted for ∼3%. The relative strength of these relationships is illustrated in Figure 5 (a-c).
Species-specific correlations between [Hg] and abiotic variables revealed significant relationships between [Hg] and conductivity in emerald shiner, rainbow smelt, trout-perch, and yellow perch (P < 0.027; Table 5). The relationship between spottail shiner [Hg] and conductivity was negative and close to significant (P ) 0.0556; Table 5). There were also negative relationships between trout-perch [Hg] and pH, and emerald shiner [Hg] and lake area (Table 5). When interspecific [Hg] was related to abiotic variables in a stepwise multiple regression, the only significant variable was conductivity (multiple regression, F ) 20.34, P < 0.0001, R2adj ) 0.22, df ) 1,67); this relationship was negative. In a final interspecific model with both biotic and abiotic variables included, growth rate and conductivity were the only two significant variables and both relationships were negative (multiple regression, F ) 23.15, P < 0.0001, R2adj ) 0.40, df ) 2,65). Because growth rate was also significantly related to conductivity, it was important to verify whether growth rate had an effect on [Hg] that was independent of conductivity. To test this, we followed the example of Greenfield et al. (35), and examined the relationship between total [Hg] and growth rate within lakes (where conductivity is constant). Growth rate remained significant (ANCOVA, F ) 8.47, P ) 0.0053, df ) 1,51) and so is interpreted to have a significant and independent effect on total [Hg]. Conductivity, however, was the stronger predictor of [Hg]. When a categorical species variable was included with growth rate and conductivity in a general linear model, 60% of the variation in forage fish [Hg] was explained (GLM, F ) 13.14, P < 0.0001, R2adj ) 0.60, df ) 7,60). Neither of the interaction terms (species-growth rate or species-conductivity) was significant (P > 0.05), meaning that the slopes of the [Hg]-growth rate and [Hg]-conductivity relationships were not significantly different among species.
Discussion Growth Rates. At a standardized mass of 8 g, we found that cisco and yellow perch had the highest growth rates and spottail shiner and emerald shiner had the lowest growth rates. Inter-species differences in growth rate did not appear to be affected by the fish size and, where data were available, corresponded to differences in maximum size (20). Both intraspecific and interspecific analyses showed that forage fish growth rates were negatively related to latitude and positively related to conductivity. These results were expected. Growth may be slower in high-latitude lakes because these lakes are generally colder and have fewer growing degree-days. Optimal temperatures for fish growth vary with consumption rate, but yellow perch and cyprinid species tend to prefer water temperatures >20 °C (21, 22). VOL. 40, NO. 5, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1443
FIGURE 5. Interspecific linear regressions of loge [Hg] vs (a) loge growth rate, (b) adjusted-δ15N, and (c) δ13C. Data for adjusted-δ15N and δ13C were taken from Swanson et al. (18). When the variables were entered in a multiple regression, all three were significant but growth rate explained the majority of the variation (R 2 ) 0.22 (growth rate), 0.025 (adjusted-δ15N), and 0.035 (δ13C)). Values plotted were LSmeans calculated for each of the six species analyzed statistically in each lake at a standardized mass of 8 g.
TABLE 5. Summary of Pearson Product-Moment Correlations between Intraspecific loge [Hg] and Environmental Variables in the Six Species Analyzed Statisticallya species
variable
correlation coefficient
P-value
Cisco Emerald shiner
maximum depth conductivity loge area conductivity conductivity conductivity pH conductivity
0.65 -0.92 -0.96 -0.65 -0.69 -0.67 -0.65 -0.57
0.166 0.027* 0.009* 0.003* 0.056 0.017* 0.031* 0.021*
Rainbow smelt Spottail shiner Trout-perch Yellow perch
a The variables presented are those that were most closely related to intraspecific loge [Hg]. If nothing was significant, the variable with the lowest P-value is presented. All correlations were based on leastsquares means calculated at a standardized mass of 8 g. *Relationships were significant at R ) 0.05.
The positive relationship between forage fish growth rate and conductivity is most likely an indirect indication of productivity effects on growth. In general, lakes in this study that had relatively high conductivity also had relatively high primary productivity (assessed from literature sources and limited chlorophyll-a and total phosphorus data) and previous authors have reported positive relationships between growth rate and conductivity because of greater prey availability (23, 24). Mercury Concentrations. Differences in [Hg] among species were consistent with previous findings. Swanson et al. (6) found that spottail shiner had significantly higher total [Hg] than trout-perch, yellow perch, rainbow smelt, and cisco in 10 northwestern Ontario lakes. We confirmed these results over a much larger geographic range and with a slightly different array of species. At a standardized body mass of 8 g, emerald shiner and spottail shiner had significantly higher [Hg] than all other species and these differences were consistent across the range of captured body sizes. These inter-species differences, along with observed [Hg]-body size relationships, have implications for predicting the effects of rainbow smelt invasion on predator fish [Hg]. Similar to Swanson et al. (6), we suggest that predicting the magnitude and direction of predator [Hg] changes requires knowledge of the species and size composition of pre- and post-smelt invasion diets. In the lakes of this study, predator [Hg] would only have increased post-invasion if the diet switch was from cisco to rainbow smelt and if the predators were feeding on relatively large prey (>12 g). The most likely predator fish to experience this diet shift is lake trout because it is a large, long-lived species that feeds almost exclusively on pelagic prey. 1444
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 5, 2006
When converted to wet weight concentrations, many individual emerald shiners and spottail shiners had total [Hg] that approached the Canadian human consumption guideline of 0.5 µg/g. This is striking given that emerald shiner and spottail shiner are forage rather than predator species. Although most humans do not consume these species, fish [Hg] in this range can pose health risks to fish-eating birds; 0.3-0.4 µg/g (wet) has been shown to impair reproductive function in common loons (Gavia immer) (25). In contrast to emerald shiners and spottail shiners, round goby had extremely low [Hg]. This may be explained by their diet. Round goby has recently invaded the Great Lakes region and is known to feed on zebra mussels (Dreissena polymorpha) (26). As filter-feeding primary consumers, zebra mussels may have relatively low [Hg] compared to other invertebrate prey for forage fish, but further research would be required to confirm this. Overall, forage fish [Hg] was best explained by interspecies differences, lake conductivity, and growth rate. In the intraspecific analyses, adjusted-δ15N was not a significant predictor of [Hg] for any of the species and δ13C was significant for only one species. In the interspecific analyses, adjustedδ15N and δ13C were significant variables in the biotic model only and even then explained very little variation. Previous authors have attributed negative [Hg]-δ13C relationships to benthic invertebrates at the base of the littoral food chain having lower [Hg] than pelagic zooplankton (27, 28). This may also explain why organisms feeding opportunistically in both the benthic and pelagic food chains (such as many forage fish) have weak [Hg]-adjusted-δ15N relationships; they may be feeding on prey that have similar adjusted-δ15N but different [Hg]. We expected the [Hg]-adjusted-δ15N relationship to be weak, but we did not expect it to be negative. This was most likely an artifact of the species assemblage that was analyzed. The relationship was strongly influenced by the contrast of emerald shiners, which had higher [Hg] than all other forage species but a relatively low adjustedδ15N (18), to rainbow smelt, which had high adjusted-δ15N (18) but relatively low [Hg]. If this study had included a more complete forage fish species assemblage we expect that the slope of the [Hg]-adjusted-δ15N relationship would be near zero. Both intra- and interspecific analyses showed that growth rate was the best biotic predictor of [Hg] and, as expected, the relationship was negative. Fish that grew more slowly and were older at the standardized size of 8 g had higher [Hg] than fish that grew faster and were younger at 8 g. This may be because fast-growing, younger fish had shorter exposure histories to Hg and/or because there was a growth dilution effect in fast-growing fish (10-12). Conductivity was the best overall predictor of forage fish [Hg]. The relationship between [Hg] and conductivity was
negative and there are two possible reasons for this. First, the drainage basins of the study lakes were either totally within or partially within the Canadian Shield. Lakes in this region show strong positive relationships between conductivity and concentrations of major cations (either Ca2+ or Mg2+) (29). Higher concentrations of Ca2+ and Mg2+ tend to decrease direct uptake and toxicity of metals such as Hg because they compete for uptake sites on the gills and alter gill permeability and electric charge (30, 31). The negative relationship between [Hg] and conductivity could also result from the positive relationship between conductivity and primary productivity in lakes examined in this study (assessed from literature sources and limited chlorophyll a and total phosphorus data). There are a number of reasons why the relationship between fish [Hg] and primary productivity is negative (14, 15). First, fish growth rates tend to increase with primary productivity. It was demonstrated above, however, that there were independent effects of conductivity and growth rate on fish [Hg]. Another possible reason for a negative [Hg]-productivity relationship is that high primary productivity dilutes Hg at the base of the food chain, in algal cells. In more eutrophic systems, the burden of bioavailable Hg in the water column is distributed among a greater number of algal cells and those cells are growing more quickly (32, 33). This dilution effect has been shown to result in lower contaminant concentrations in zooplankton and fish. As well, higher sedimentation rates in more eutrophic systems may increase the flux of contaminants out of the water column and into the sediment (32). The results of our study indicate that biomagnification processes are not readily discernible within a community of relatively short-lived, trophically similar organisms. The maximum among-species difference in adjusted-δ15N was ∼2 ‰ (18). This is less than the average value of 3.4‰ for one full trophic level (34). At this fine scale of trophic differentiation, forage fish [Hg] was best predicted by growth rate and conductivity, not adjusted-δ15N. Somewhat similar results have been presented for yellow perch (35), but this is the first study that has related [Hg] to adjusted-δ15N, δ13C, growth rate, and lake characteristics in more than one forage fish species. Future research is needed to address how predictors of fish [Hg] change with the scale of trophic differentiation among species. Although adjusted-δ15N or trophic position may be the best predictor of fish [Hg] at the scale of whole food webs, water chemistry and growth rate may be more important variables within trophic guilds or communities. This has implications for anticipating the effects of ecosystem stressors such as species invasions. When applied to the issue of rainbow smelt invasion, the results indicate that despite a lengthened food chain to top predators (4-6), pre-to post-invasion diet switches to rainbow smelt should only result in predator [Hg] increases if predators are feeding on relatively large prey and switch from a ciscodominated diet to a rainbow smelt-dominated diet. The most likely predator to experience this is lake trout and it is thus not surprising that post-invasion [Hg] increases have most often been reported in lake trout (4, 5).
Acknowledgments Funding for this study was provided by Manitoba Hydro, Canadian Circumpolar Institute, Alberta Ingenuity Foundation, and NSERC (Discovery Grant to D.W.S. and PGS-A scholarship to H.K.S.). Simcha Snell provided assistance in the field. Samples from northern Manitoba were provided by North-South Consultants and samples from Lake Champlain were provided by Vermont Fish and Wildlife. Angela Sommerville, Neil Strange, and Anne McGeachy performed laboratory analyses. John Brzustowski, Vince St. Louis, Erik Allen, and three anonymous reviewers provided constructive criticisms on early drafts of the manuscript.
Literature Cited (1) Scott, W. B.; Crossman, E. J. Freshwater Fishes of Canada. Bull. Fish Res. Board Can. 1973, 184, 966. (2) Evans, D. O.; Loftus, D. H. Colonization of inland lakes in the Great Lakes region by rainbow smelt, Osmerus mordax: Their freshwater niche and effects on indigenous fishes. Can. J. Fish. Aquat. Sci. 1987, 44, 249-266. (3) Remnant, R. A.; Graveline, P. G.; Bretecher, R. L. Range extension of the rainbow smelt, Osmerus mordax, in the Hudson Bay drainage of Manitoba. Can. Field-Nat. 1997, 111, 660-662. (4) Vander Zanden, M. J.; Rasmussen, J. B. A trophic position model of pelagic food webs: Impact on contaminant bioaccumulation in lake trout. Ecol. Monogr. 1996, 66, 451-477. (5) Franzin, W. G.; Barton, B. A.; Remnant, R. A.; Wain, D. B.; Pagel, S. J. Range extension, present and potential distribution, and possible effects of rainbow smelt in Hudson Bay drainage waters of northwestern Ontario, Manitoba, and Minnesota. N. Am. J. Fish. Manage. 1994, 14, 65-76. (6) Swanson, H. K.; Johnston, T. A.; Leggett, W. C.; Bodaly, R. A.; Doucett, R. R.; Cunjak, R. A. Trophic positions and mercury bioaccumulation in rainbow smelt (Osmerus mordax) and native forage fishes in northwestern Ontario lakes. Ecosystems 2003, 6, 289-299. (7) Johnston, T. A.; Leggett, W. C.; Bodaly, R. A.; Swanson, H. K. Temporal changes in mercury bioaccumulation by predatory fishes of boreal lakes following the invasion of an exotic forage fish. Environ. Toxicol. Chem. 2003, 22, 2057-2062. (8) Kidd, K. A.; Hesslein, R. H.; Fudge, R. J.; Hallard, K. A. The influence of trophic level as measured by δ15N on mercury concentrations in freshwater organisms. Water, Air, Soil Pollut. 1995, 80, 1011-1015. (9) Cabana, G.; Rasmussen, J. B. Modelling food chain structure and contaminant bioaccumulation using stable nitrogen isotopes. Nature 1994, 372, 255-257. (10) Huckabee, J. W.; Elwood, J. W.; Hildebrans, S. G. In The Biogeochemistry of Mercury in the Environment; Nriagu, J. O., Ed.; Elsevier: New York, 1979; pp 277-302. (11) de Freitas, A. S. W.; Qadri, S. U.; Case, B. E. In Proceedings of the International Conference on Transport of Persistent Chemicals in Aquatic Ecosystems; de Freitas, A. S. W., Kushner, K. J., Qadri, S. U., Eds.; National Research Council of Canada: Ottawa, ON, 1974; pp III.31-III.36. (12) Meinertz, J. R.; Stehly, G. R.; Gingerich, W. H.; Allen, J. L. Residues of malachite green in eggs and fry of rainbow trout Oncorhynchus mykiss, after treatment of eggs. J. Fish Dis. 1995, 18, 239-247. (13) Rask, M.; Jarvinen, M.; Kuoppamaki, K.; Poysa, H. Limnological responses to the collapse of the perch population in a small lake. Ann. Zool. Fenn. 1996, 33, 517-524. (14) Kidd, K. A.; Paterson, M. J.; Hesslein, R. H.; Muir, D. C.; Hecky, R. E. Effects of northern pike (Esox lucius) additions on pollutant accumulation and food web structure, as determined by δ13C and δ15Ν, in a eutrophic and an oligotrophic lake Can. J. Fish. Aquat. Sci. 1999, 56, 2193-2202. (15) Essington, T. E.; Houser, J. N. The effect of whole-lake nutrient enrichment on mercury concentration in age-1 yellow perch. Trans. Am. Fish. Soc. 2003, 132, 57-68. (16) Hendzel, M. R.; Jamieson, D. M. Determination of mercury in fish. Anal. Chem. 1976, 48, 926-928. (17) Nielsen, L. A., Johnson, D. L., Eds. Fisheries Techniques; American Fisheries Society: Blacksburg, VA, 1983. (18) Swanson, H. K.; Johnston, T. A.; Schindler, D. W.; Cunjak, R. A.; Whittle, D. M. Is the trophic ecology of forage fishes related to ecosystem size? Oikos, submitted for publication. (19) Cabana, G.; Rasmussen, J. B. Comparison of aquatic food chains using nitrogen isotopes. Ecology 1996, 93, 10844-10847. (20) Froese, R., Pauly, D., Eds. FishBase. http://www.fishbase.org. Accessed 2004. (21) Hayward, R. S.; Margraf, F. J. Eutrophication effects on prey size and food available to yellow perch in Lake Erie. Trans. Am. Fish. Soc. 1987, 116, 210-223. (22) Jobling, M. Fish Bioenergetics; Chapman and Hall: Bury St. Edmunds, Suffolk, 1994. (23) Donald, D. B.; Anderson, R. S.; Mayhood, D. W. Correlations between brook trout growth and environmental variables for mountain lakes in Alberta. Trans. Am. Fish. Soc. 1980, 109, 603610. (24) Trippel, E. A.; Beamish, F. W. H. Lake trout (Salvelinus namaycush) growth potential predicted from cisco (Coregonus artedii) population structure and growth. Can. J. Fish. Aquat. Sci. 1989, 46, 1531-1538. VOL. 40, NO. 5, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1445
(25) Scheuhammer, A. M.; Atchison, C. M.; Wong, A. H. K.; Evers, D. C. Mercury exposure in breeding common loons (Gavia immer) in central Ontario, Canada. Environ. Toxicol. Chem. 1998, 17, 191-196. (26) Ray, W. J.; Corkum, L. D. Predation of zebra mussels by round gobies, Neogobius melanostomus. J. Great Lakes Res. 1997, 50, 267-273. (27) Tremblay, A. In Mercury in the Biogeochemical Cycle: Natural Environments and Hydroelectric Reservoirs of Northern Quebec (Canada); Lucotte, M., Schetagne, R., Therien, N., Langlois, C., Tremblay, A., Eds.; Springer-Verlag: Berlin, Germany, 1999; pp 89-113. (28) Power, M.; Klein, G. M.; Guiger, K. R. R. A.; Kwan, M. K. H. Mercury accumulation in the fish community of a sub-Arctic lake in relation to trophic position and carbon sources. J. Appl. Ecol. 2002, 39, 819-830. (29) Armstrong, F. A. J.; Schindler, D. W. Preliminary chemical characterization of waters in the Experimental Lakes Area, northwestern Ontario. J. Fish. Res. Board Can. 1971, 28, 171187. (30) Pagenkopf, G. K. Gill surface interaction model for trace-metal toxicity to fishes: role of complexation, pH, and water hardness. Environ. Sci. Technol. 1983, 17, 342-347.
1446
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 5, 2006
(31) Rodgers, D. W.; Beamish, F. W. H. Water quality modifies uptake of waterborne methylmercury by rainbow trout, Salmo gairdneri. Can. J. Fish. Aquat. Sci. 1983, 40, 824-828. (32) Larsson, P.; Andersson, A.; Broman, D.; Nordback, J.; Lundbert, E. Persistent organic polluntants (POPs) in pelagic systems. Ambio 2000, 29, 202-209. (33) Pickhardt, P. C.; Folt, C. L.; Chen, C. Y.; Klaue, B.; Blum, J. D. Algal blooms reduce the uptake of toxic methylmercury in freshwater food webs. Proc. Natl. Acad. Sci., U. S. A. 2002, 99, 4419-4423. (34) Post, D. M. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 2002, 83, 703718. (35) Greenfield, B. K.; Hrabik, T. R.; Harvey, C. J.; Carpenter, S. R. Predicting mercury levels in yellow perch: use of water chemistry, trophic ecology, and spatial traits. Can. J. Fish. Aquat. Sci. 2001, 58, 1419-1429.
Received for review May 30, 2005. Revised manuscript received December 6, 2005. Accepted December 12, 2005. ES0510156
