Selenium and mercury concentrations in some fish species of the Madeira River, Amazon Basin, Brazil

Samples of 7 species of piscivorous, omnivorous, and herbivorous fish caught at 12 different sites on the Madeira River, Amazon Basin, were analyzed for selenium and mercury. Selenium was determined by anodic stripping voltammetry and mercury by cold vapor atomic absorption spectrophotometry. The means for selenium concentrations ranged from 0.49 to 3.11 nmol/g and for mercury from 0.41 to 6.66 nmol/g depending on the fish species. The molar ratios of Hg:Se increased according to the fish trophic level. Piscivorous species had the highest mean ratio (4.0) and herbivorous species the lowest (0.9). There was a positive and statistically significant correlation between selenium and mercury concentrations for the herbivorous species (r = 0.716;p = 0.0088) not seen for omnivororus and piscivorous species (r = -0.2032;p = 0.3407). These findings are significant for the fish-eating population of the Madeira River because the ingestion of mercury would always be in excess of selenium.     

Selenium:mercury molar ratio in eared grebes (Podiceps nigricollis) as a possible biomarker of exposure

Mercury and selenium have adverse effects on health, and in the past their individual levels in tissues have been used as biomarkers of environmental contamination. These selenium:mercury molar ratios has been proposed as an alternative way to anticipate possible health risks to organisms. We examine selenium and mercury levels, their molar ratios, and variability in the ratios in the brain, liver, muscle, and feathers of a common waterbird, the eared grebe (Podiceps nigricollis), at several locations and phases of their annual cycle. We found: (1) Mean total mercury, for any site or tissue, ranged from 0.15 ppm in the brain to 29.2 ppm in breast feathers; (2) In any tissue, mean mercury levels varied by as much as 10-fold while selenium varied by 3-fold; (3) Mercury and selenium levels were correlated only in liver; (4) Selenium:mercury molar ratios varied significantly in regular patterns among tissues (less than 1 in feathers, up to 23 in brain), sites, and stages of annual cycle; (5) Molar ratios were affected by body weight (but not age), and the heaviest birds had the lowest ratios; and (6) Molar ratios varied more for brain than in other tissues. Low molar ratios are generally considered harmful, although no threshold ratio has been identified. Despite wide variation of molar ratios, field studies of eared grebes have not detected overt adverse effects. Before being adopted as a biomarkers, we suggest that selenium:mercury molar ratios be used in conjunction with studies of individual metal levels, and in accordance with detailed studies of selected species, to provide a baseline of variation in different organisms and tissues.     

Selenium and Mercury in Pelagic Fish in the Central North Pacific Near Hawaii

Protective effects of selenium against mercury toxicity have been demonstrated in all animal models evaluated. As interactions between selenium and mercury and their molar ratios in seafood are essential factors in evaluating risks associated with dietary mercury exposure, considering mercury content alone is inadequate. In this study, the absolute and molar concentrations of mercury and selenium were determined in edible portions from 420 individual fish representing 15 species of pelagic fish collected from the central North Pacific Ocean near Hawaii. Selenium was in molar excess of mercury in almost all fish species evaluated. The rank order of mean Se/Hg molar ratios was striped marlin (17.6) > yellowfin tuna (14.1) > mahimahi (13.1) > skipjack tuna (12.8) > spearfish (11.4) > wahoo (10.8) > sickle pomfret (6.7) > albacore tuna (5.3) > bigeye tuna (5.2) > blue marlin (4.1) > escolar (2.4) > opah (2.3) > thresher shark (1.5) > swordfish (1.2) > mako shark (0.5). With a Se/Hg molar ratio of less than 1, mako shark was the only fish containing a net molar excess of mercury. A selenium health benefit value based on the absolute amounts and relative proportions of selenium and mercury in seafood is proposed as a more comprehensive seafood safety criterion.     

Selenium:Mercury Molar Ratios in Bullfrog and Leopard Frog Tadpoles from the Northeastern United States

Vertebrates experience adverse effects from methylmercury, largely obtained through their food. Selenium has the potential to reduce the toxic effects of methylmercury (and vice versa). In this paper, we examine the selenium:mercury molar ratios in tadpoles (Lithobates sphenocephalus, Lithobates catesbeianus (formerly Rana), and a newly documented leopard frog species currently referred to as R. sp. nov.) and fully formed leopard frog metamorphs. There were no significant differences in metal levels between the two leopard frog species, and data were therefore combined. Selenium:mercury molar ratios varied from 19 to 38 for bullfrog tadpoles, from 16 to 330 for leopard frog tadpoles, and from 7 to 17 for leopard frog metamorphs. Leopard frog tadpoles with less than 45 days exposure to field conditions had significantly higher molar ratios than other tadpoles and leopard frog metamorphs. There were significant locational differences for the molar ratios of bullfrogs, and leopard frog tadpoles with more than 45 days of field exposure. At the sites where we were able to sample both leopard frog tadpoles and leopard frog metamorphs, there were significant differences between the two distinct life stages. Most of the variation in the ratio was accounted for by selenium levels, field sites, and exposure period.     

Quantification and speciation of mercury and selenium in fish samples of high consumption in Spain and Portugal

Mercury (Hg) and selenium (Se) determinations were carried out to evaluate human exposure to those elements through fish consumption in Spain and Portugal. Atomic fluorescence spectroscopy (AFS) was applied in a cold vapor mode for total mercury quantification and was also hyphenated to gas chromatography (GC) to achieve the speciation of organomercurial species in fish samples. The results obtained show the highest concentration of Hg in swordfish and tuna (0.47+/-0.02 and 0.31+/-0.01 microg g-1, respectively), and a much lower concentration in sardine, mackerel shad, and octopus (0.048+/-0.002, 0.033+/-0.001, and 0.024+/-0.001 microg g-1, respectively). The determination of alkyl mercury compounds revealed that 93-98% of mercury in the fish samples was in the organic form. Methylmercury (MeHg) was the only species found in the three fish species with higher mercury content.Total selenium concentration was high in sardine, swordfish, and tuna (0.43+/-0.02, 0.47+/-0.02, and 0.92+/-0.01 microg g-1, respectively), but low in mackerel shad and octopus (0.26+/-0.01 and 0.13+/-0.01 microg g-1, respectively). Speciation of selenium compounds was done by high-performance liquid chromatography in conjunction with inductively coupled plasma mass spectrometry (LC-ICP-MS). Selenomethionine (SeMet) was the only selenium compound identified in the fish samples with higher selenium content.Among the fish species studied, sardine had the most favourable Se:Hg and SeMet:MeHg molar ratios; therefore, its consumption seems to be preferable.     

Elemental stoichiometry of freshwater fishes in relation to phylogeny, allometry and ecology

Twenty species of freshwater fishes were collected from Minnesota, Iowa and Michigan and their whole-body carbon, nitrogen and phosphorus contents and the respective C:N:P ratios were determined. Patterns were examined in intra- and interspecific variation, allometry and variation caused by habitat and trophic level in whole fish while controlling for the role of phylogeny. Stoichiometric variation was greater across than within species, C:N:P allometry was species-specific, nutrient content within a species was somewhat habitat-specific and P concentration showed a strong phylogenetic signal. Stoichiometric relationships with allometry and feeding guild were observed but were not significant in an analysis accounting for non-independence of closely related species. Supportive evidence for the hypothesis that the considerable variation in whole fish phosphorus concentrations could be ascribed to differences in bone and scale development, as previously suggested, is shown. Whole fish Ca:P ratios had a nearly constant stoichiometry consistent with the chemical signature of bone. This result combined with a phylogenetic signal for fish P indicated that the great stoichiometric variability among fish taxa in P content was derived almost entirely from skeletal investment.     

OPTIMUM N:P RATIOS AND COEXISTENCE OF PLANKTONIC ALGAE1

The optimum atomic ratio of N to P, the ratio at which one nutrient limitation changes over to the other, was determined in seven species of freshwater planktonic algae. The ratio varied over a wide range among species; the average for these species was 17. If the cellular nutrient ratios in marine species are comparable with those in freshwater organisms, Redfield's ratio of 15 is remarkably close to the average. Cellular N:P ratios varied over a 24-h period under a light:dark cycle. The variation of the optimum ratio between species and diel change in cellular N:P ratios within a species could play an important role in population dynamics by enhancing the probability of coexistence of species.     

Selenium Health Benefit Values as Seafood Safety Criteria

Selenium (Se) is absolutely required for activity of 25–30 genetically unique enzymes (selenoenzymes). All forms of life that have nervous systems possess selenoenzymes to protect their brains from oxidative damage. Homeostatic mechanisms normally maintain optimal selenoenzyme activities in brain tissues, but high methylmercury (MeHg) exposures sequester Se and irreversibly inhibit selenoenzyme activities. However, nutritionally relevant amounts of Se can replace the Se sequestered by MeHg and maintain normal selenoenzyme activities, thus preventing oxidative brain damage and other adverse consequences of MeHg toxicity. Findings of studies that seem contradictory from MeHg exposure perspectives are entirely consistent from MeHg:Se molar ratio perspectives. Studies that have reported dose-dependent consequences of maternal MeHg exposures on child development uniformly involved seafoods that contained much more Hg than Se. Meanwhile more typical varieties of ocean fish contain much more Se than Hg. This may explain why maternal MeHg exposure from eating ocean fish is associated with major IQ benefits in children instead of harm. Therefore, instead of being avoided, ocean fish consumption should be encouraged during pregnancy. However, the safety of freshwater fish consumption is less certain. In freshwater fish, MeHg bioaccumulation and toxicity are both inversely related to Se bioavailability. Their Se can be far lower than their MeHg contents, potentially making them more dangerous than pilot whale meats. Therefore, to provide accurate and appropriate regulatory advice regarding maternal consumption of seafoods and freshwater fish, Hg:Se molar ratios need to be incorporated in food safety criteria.     

Ontogeny, diet shifts, and nutrient stoichiometry in fish

Most stoichiometric models do not consider the importance of ontogenetic changes in body nutrient composition and excretion rates. We quantified ontogenetic variation in stoichiometry and diet in gizzard shad, Dorosoma cepedianum , an omnivorous fish with a pronounced ontogenetic diet shift; and zebrafish, Danio rerio, grown in the lab with a constant diet. In both species, body stoichiometry varied considerably along the life cycle. Larval gizzard shad and zebrafish had higher molar C:P and N:P ratios than larger fish. Variation in body nutrient ratios was driven mainly by body P, which increased with size. Gizzard shad body calcium content was highly correlated with P content, indicating that ontogenetic P variation is associated with bone formation. Similar trends in body stoichiometry of zebrafish, grown under constant diet in the laboratory, suggest that ontogeny (e.g. bone formation) and not diet shift is the main factor affecting fish body stoichiometry in larval and juvenile stages. The N:P ratio of nutrient excretion also varied ontogenetically in gizzard shad, but the decline from larvae to juveniles appears to be largely associated with variation in the N:P of alternative food resources (zooplankton vs detritus) rather than by fish body N:P. Furthermore, the N:P ratio of larval gizzard shad excretion appears to be driven more by the N:P ratio at which individuals allocate nutrients to growth, more so than static body N:P, further illustrating the need to consider ontogenetic variation. Our results thus show that fish exhibit considerable ontogenetic variation in body stoichiometry, driven by an inherent increase in the relative allocation of P to bones, whereas ontogenetic variation in excretion N:P ratio of gizzard shad is driven more by variation in food N:P than by body N:P.     

Body length variation within multi-species fish shoals: the effects of shoal size and number of species

Theory predicts that selection should favour phenotypic homogeneity in fish shoals, and field studies have indeed confirmed that variation in body length within fish shoals is significantly lower than expected from a random distribution of fish among shoals. We investigated the extent to which variation in fish body length within shoals is determined by the shoal mean of body length, the number of species in a shoal, and the overall shoal size. We collected 34 fish shoals, ranging in size from 6 to 776 individuals, from the littoral zone of a Canadian lake. Shoals consisted of up to four different species, with multi-species shoals being larger and more frequent than single-species ones. The strongest determinant of body length variation within shoals was the shoal mean of body length, followed by the number of fish species in a shoal; i.e. multi-species shoals were less size-assorted than single-species ones. A more detailed analysis showed that the higher body length variation observed in multi-species shoals was due to increased body length variation both within and between component species. Shoal size had no significant effect on body length variation within shoals. Potential explanations of the positive relationship between body length variation and the number of species in a shoal are suggested. The implications of the above results for the evolution of multi-species shoals are discussed. Received: 6 May 1997 / Accepted: 14 October 1997     

Geographical variation in male genitalia in Brachyrhaphis episcopi (Poeciliidae): is it sexually or naturally selected?

Male poeciliid fishes inseminate females using an intromittent organ called the gonopodium. Here we report on natural variation in gonopodium size both within and between 12 populations of the freshwater fish Brachyrhaphis episcopi (Poeciliidae) in Panama. We show that males from sites with more predatory fish species have, on average, a relatively longer gonopodium than males inhabiting sites with fewer predatory fish. Gonopodium length was not correlated with the site-specific adult sex ratio and the average sex ratio was more strongly female biased at sites with more predatory fish. The gonopodium exhibited lower phenotypic variance than the average for sexually selected traits and it generally showed negative allometry. Our results are similar to those reported for the guppy Poecilia reticulata . Two alternative hypotheses for these findings are discussed. First, that population differences are sexually selected. Second, that they are an incidental consequence of environmental differences between sites. Specifically, that higher water flow rates select for enlarged fin size and stockier bodies in downstream sites where predatory fish are more common.     

From Ethnobiology to Ecotoxicology: Fishers’ Knowledge on Trophic Levels as Indicator of Bioaccumulation in Tropical Marine and Freshwater Fishes

Information on fish trophic levels is important to assess fishing impacts and to better understand the bioaccumulation of pollutants within aquatic food chains. The local ecological knowledge held by small-scale fishers can fill knowledge gaps in fish trophic ecology. We estimated the trophic levels of 69 tropical and subtropical fish species (33 coastal and 36 freshwater species) using data on fish diets from the literature and obtained from interviews with Brazilian fishers. The fish trophic levels estimated from fishers’ knowledge were positively correlated with the trophic levels estimated using data from biological studies for both coastal and freshwater fish. The fishers’ knowledge also indicated bioaccumulation patterns, as the fish trophic levels estimated from fishers’ knowledge were positively related to the mercury (Hg) content in fish muscle (wet weight, from literature data) in 41 fish species (15 coastal and 26 freshwater). These findings reveal the potential for new applications of fishers’ knowledge to ecotoxicology, which could improve management of aquatic ecosystems and strengthen fishers’ political status.     

Carbon,nitrogen and phosphorus ecological stoichiometry of Lateolabrax macultus and Acanthogobius ommaturus in the Estuary of Yangtze River,China

Most stoichiometric models do not consider the importance of ontogenetic changes in body nutrient composition. Among aquatic animals, body elemental ontogenetic changes may be particularly important in fish due to bone formation at larval–juvenile development stage. We investigated the stoichiometric characteristics and ontogenetic variation of carbon (C), nitrogen (N), phosphorous (P) and calcium (Ca) in the whole fish without gut of Lateolabrax macultus and Acanthogobius ommaturus from the Yangtze Estuary. Some differences were found in the whole-fish nutrient content across species as well as fish length and mass groups. Mean body content of L. macultus was 41.2% for C (34.9–44.9%), 11.8% for N (10.3–12.9%), and 3.3% for P (1.4–5.4%) and the molar ratio of C:N:P was 32:8:1. A. ommaturus had the following mean compositions: carbon, 43.4% (40.0–45.9%); nitrogen, 13.2% (11.7–14.1%); phosphorus, 2.9% (1.6–6.0%) and molar ratio of C:N:P was 38:10:1. The variation of C content in of L. macultus was minimum (0.05), followed by N (0.06), and the variation of P content was the highest (0.22). The variation of C content in A. ommaturus was minimum (0.03), followed by N (0.04), and the variation of P content was also the highest (0.27). Both species varied the least in C, moderate in N and the most in P. Comparing the two species, A. ommaturus was significantly higher for C, N, C:P, N:P and significantly lower for P and C:N than L. macultus. The positive correlation between C and N contents was significant (P < 0.01) for both species. In contrast, C and P concentrations showed little correlation (P > 0.05). The contents of C, N in L. macultus showed significant negative correlation with total length and wet mass, respectively, while C:N ratio showed significant positive correlation with them. As for A. ommaturus, the content of C and the ratio of C:N showed significant negative correlation with total length and wet mass, respectively. The two estuarine fishes in this study mainly consumed fish, shrimps and crabs, which could be the reason that the stoichiometry of L. macultus and A. ommaturus are higher than other freshwater fishes that have been reported previously. In both species, the content of calcium was highly correlated with the content of body P, which indicated that the change of the content of P is associated with bone formation in ontogeny. Fish shape, scale type, the ratio of head length/standard length and the degree of ossification of the internal skeleton and outer integument might have caused the results that L. macultus had higher contents of P and Ca than A. ommaturus.     

Temperature-dependent sex determination in fish revisited: prevalence, a single sex ratio response pattern, and possible effects of climate change

Background

In gonochoristic vertebrates, sex determination mechanisms can be classified as genotypic (GSD) or temperature-dependent (TSD). Some cases of TSD in fish have been questioned, but the prevalent view is that TSD is very common in this group of animals, with three different response patterns to temperature.

Methodology/Principal Findings

We analyzed field and laboratory data for the 59 fish species where TSD has been explicitly or implicitly claimed so far. For each species, we compiled data on the presence or absence of sex chromosomes and determined if the sex ratio response was obtained within temperatures that the species experiences in the wild. If so, we studied whether this response was statistically significant. We found evidence that many cases of observed sex ratio shifts in response to temperature reveal thermal alterations of an otherwise predominately GSD mechanism rather than the presence of TSD. We also show that in those fish species that actually have TSD, sex ratio response to increasing temperatures invariably results in highly male-biased sex ratios, and that even small changes of just 1–2°C can significantly alter the sex ratio from 1∶1 (males∶females) up to 3∶1 in both freshwater and marine species.

Conclusions/Significance

We demonstrate that TSD in fish is far less widespread than currently believed, suggesting that TSD is clearly the exception in fish sex determination. Further, species with TSD exhibit only one general sex ratio response pattern to temperature. However, the viability of some fish populations with TSD can be compromised through alterations in their sex ratios as a response to temperature fluctuations of the magnitude predicted by climate change.     

Biomagnification of Mercury and Selenium in Blue Shark Prionace glauca from the Pacific Ocean off Mexico

The aim of this study was to determine the biomagnification of mercury through the principal prey of the blue shark, Prionace glauca, off the western coast of Baja California Sur, Mexico, as well as the relationship between mercury and selenium in blue sharks. High levels of mercury were found in shark muscle tissues (1.39?±?1.58?μg/g wet weight); these values are above the allowed 1.0?μg/g for human consumption. The mercury to selenium molar ratio was 1:0.2. We found a low correlation between mercury bioaccumulation and shark size. Juveniles have lower concentrations of mercury than adults. Regarding the analyzed prey, the main prey of the blue shark, pelagic red crab, Pleuroncodes planipes, bioaccumulated 0.04?±?0.01?μg/g Hg wet weight, but the prey with higher bioaccumulation was the bullet fish Auxis spp. (0.20?±?0.02?μg/g wet weight). In terms of volume, the red crab P. planipes can be the prey that provides high levels of mercury to the blue shark.     

Fatty acid content of some freshwater fish of commercial importance from tropical lakes in the Ethiopian Rift Valley

Fifty samples of freshwater fish, representing eight tropical species, were collected from Ethiopian Rift Valley lakes in order to study the variation of lipids and fatty acids (FA) both within and between species. Most specimens (36 samples) were low in fat, ≥5% of dry weight (dw). Medium- (nine samples) and high-fat fish (five samples) contained ≤6% and >10% dw, respectively. The extent of variation was more pronounced in the herbivore Oreochromis niloticus than in the omnivorous (e.g. Barbus sp.) or carnivorous–piscivorous ( Clarias gariepinus ) fishes. Twenty-eight FA of various chain lengths and saturation levels were identified. Most FA were unsaturated and long-chained. The major individual FA were palmitic acid (16:0), stearic acid (18 : 0), oleic acid (18 : 1ω9) and docosahexaenoic acid (DHA) (22 : 6ω3). The ω3/ω6 ratio varied considerably (1·1–7·6) and O. niloticus from Lake Haiq was found to be superior in lipid quality to all the tropical fish species considered in this study. The data show that tropical freshwater fish are comparable to temperate freshwater fish as sources of polyunsaturated FA.     

Heavy metal and selenium concentrations in liver tissue from wild American alligator (Alligator mississippiensis) livers near Charleston, South Carolina

Liver samples from 33 wild American alligators (Alligator mississippiensis) livers from the Charleston, South Carolina, area were analyzed for arsenic (As), cadmium (Cd), cobalt (Co), chromium (Cr), mercury (Hg), nickel (Ni), lead (Pb), and selenium (Se) concentrations. Alligators are top predators and are considered a good biomonitoring species for various toxins, including heavy metals. Alligators from other areas in the US have shown high concentrations of mercury and other heavy metals, but the Charleston area, which is highly industrialized, has not been investigated. We found wide variation in hepatic heavy metal and selenium concentrations among alligators. Length and sex did not show a strong relationship with any metal based on statistical analysis. However, cluster analysis revealed three groupings of alligators based on liver metal concentrations. Alligators with low Se:Hg ratios also had high concentrations of Hg. Due to the wide variation in metal concentrations among individual alligators, we postulate that individual diet and microhabitat usage could be the cause for this variation.     

The chemical forms of mercury and selenium in whale skeletal muscle

Human exposure to potentially neurotoxic methylmercury species is a public-health concern for many populations worldwide. Both fish and whale are known to contain varying amounts of methylmercury species. However studies of populations that consume large quantities of fish or whale have provided no clear consensus as to the extent of the risk. The toxicological profile of an element depends strongly on its chemical form. We have used X-ray absorption spectroscopy to investigate the comparative chemical forms of mercury and selenium in fish and whale skeletal muscle. The predominant chemical form of mercury in whale is found to closely resemble that found in fish. In the samples of skeletal muscle studied, no involvement of selenium in coordination of mercury is indicated in either whale or fish, with no significant inorganic HgSe or HgS type phases being detected. The selenium speciation in fish and whale shows that similar chemical types are present in each, but in significantly different proportions. Our results suggest that for equal amounts of Hg in skeletal muscle, the direct detrimental effects arising from the mercury content from consuming skeletal muscle from whale and fish should be similar if the effects of interactions with other components in the meat are not considered.     

Large-scale intraspecific variation in life-history traits of European freshwater fish

Aims To test the magnitude and direction of the effects of large‐scale environmental factors (latitude and habitat type: lotic or lentic) on the intraspecific variations in multiple life‐history traits, across multiple European freshwater fish species. To test the relevance of defining species traits by quantifying the magnitude of interspecific vs. intraspecific variability in traits. Location Europe. Methods We obtained estimates of 11 fish traits from published sources for 1089 populations of 25 European freshwater fish species. Traits were: longevity, maximal length, growth rate, asymptotic length, mortality rate, age and length at maturation, fecundity, egg size, gonadosomatic index, and length of breeding season. We described population habitats by latitude and habitat type (lotic or lentic), when available. For each species we tested the combined effect of latitude and habitat type on the intraspecific variation of each trait using analysis of covariance (ancova ). We compared the intraspecific variation in traits with the variation between species using an analysis of variance (anova ) for each trait, all species pooled. Results We found a consistent effect in direction of latitude on six traits, but we showed that this effect is not always significant across species. Higher‐latitude populations often grew more slowly, matured later, had a longer life span and a higher maximal and asymptotic length, and allocated more energy to reproduction than populations at lower latitudes. By contrast, we noted only a slight effect of habitat type on the intraspecific variation in traits, except for Salmo trutta. All traits varied significantly between species. However, traits such as growth rate, mortality rate and length of breeding season varied more between populations than between species, whereas fecundity and traits associated with body length varied more between species. Main conclusions Latitude, in contrast to habitat type, is an important factor influencing several traits of geographically widely dispersed populations of multiple European freshwater fish species. Species traits that vary more between species than between populations are attractive variables for understanding and predicting the responses of stream fish communities to their environment.     

Mercury and selenium in whole blood and serum in relation to fish consumption and amalgam fillings in adolescents

Mercury and selenium in whole blood and serum of 245 17-year old Swedish adolescents were analysed. The relationships between these elements' concentrations and the consumption of fish as well as the number of dental amalgam fillings were studied. The geometric means (GM) of the mercury concentrations were 1.1 microg/L in blood and 0.43 microg/L in serum. The mean selenium concentration in blood was 110 microg/L and the GM of the serum selenium concentration 110 microg/L. Fish species with dietary restrictions due to elevated mercury Levels (i.e. pike, perch, pikeperch, burbot, eel and halibut) were consumed on average 0.7 times/month and fish species without such restrictions 4.1 times/month. Despite this comparatively low fish consumption, the adolescents' blood mercury concentrations were positively correlated with fish consumption. Of the adolescents, 39% had amalgam fillings (mean 2 +/- 1.5). Serum mercury was influenced by the number of amalgam fillings, by fish consumption, blood and serum levels of selenium and the residential area. Blood and serum selenium concentrations were not influenced by fish consumption, but were positively associated with the serum mercury concentration.