Arch. Environ. Contam. Toxicol. 45, 235246 (2003) DOI: 10.

1007/s00244-003-0207-1

A R C H I V E S O F

Environmental Contamination a n d T oxicology

2003 Springer-Verlag New York Inc.

Mercury Biomagnication in a Tropical Black Water, Rio Negro, Brazil

A. C. Barbosa,1 J. de Souza,2 J. G. Dorea,3 W. F. Jardim,4 P. S. Fadini5
1 2

Instituto Brasileiro do Meio Ambiente (IBAMA), Braslia, Brazil Instituto de Qumica, Universidade de Braslia, Braslia, Brazil 3 Faculdade de Ciencias da Saude, Universidade de Braslia, C. P. 04322, 70919-970 Braslia, DF, Brazil 4 Instituto de Qumica, Universidade de Campinas (UNICAMP), Campinas, Brazil 5 PUC, Campinas, Brazil

Received: 10 September 2002 /Accepted: 8 March 2003

Abstract. The population living along the riverbanks of the Amazon basin depends heavily on sh for nutritional support. Mono-methyl-mercury (MMHg) concentrates in sh, which can contaminate humans, the risk depending not only on sh MMHg concentration but also on the amount of sh consumed. We sampled nine locations of the Rio Negro basin, differing in water pH, Hg concentrations, and dissolved organic carbon (DOC), and determined total Hg from 951 sh samples of species representative of the food web: herbivorous, detritivorous, omnivorous, and piscivorous. Mercury concentrations varied widely in all species but showed a trend that depended on sh feeding strategies. The highest mean concentration was found in the piscivorous species (688.90 ng/g 1), followed by omnivorous (190.30 ng/g 1), detritivorous (136.04 ng/g 1), and herbivorous (70.39 ng/g 1). Fish Hg concentrations exceeding current safe limits (500 ng/g 1) for human consumption were found mainly in the piscivorous species (60%). Signicant positive correlation between sh weight and Hg concentration was seen for the piscivorous Serrasalmus spp. (n 326; r 0.3977; p 0.0001), Cichla spp. (n 125; r 0.4600; p 0.0001), and Pimelodus spp. (n 12; r 0.8299; p 0.0008), known locally as Piranha, Tucunare, and Mandi, respectively. However, a negative correlation was seen for nonpiscivorous Potamorhina latior (n 30; r 0.3763; p 0.0404) and Leporinus spp. (n 44; r 3987; p 0.0073), known as Branquinha (detritivorous) and Aracu (omnivorous). Fish-Hg concentrations in the acidic waters (pH range, 4.09 6.31) of the Rio Negro habitat, with its wide gradient of Hg concentrations (3.4 11.9 g/L 1) and DOC (1.8515.3 mg/L 1) but no history of gold mining activityare comparable to other Amazonian rivers. Opportunity sh catches in the Rio Negro habitat show high muscle-Hg derived from natural sources, but no systematic association with site-dependent geochemistry.

Mercury (Hg) is widespread in the environment and is an important food contaminant that, under special circumstances, can cause

Correspondence to: J. G. Dorea; email: dorea@rudah.com.br

neuromotor disturbances and neuropathies. It occurs naturally in three oxidation states Hg0, Hg 1, and Hg 2. A series of complex chemical transformations allows Hg to cycle in the environment, but methylation is the most important step for Hg entry to the aquatic food chain. Fish is the highest bioaccumulator of monomethyl-mercury (MMHg) and serves as an indicator of Hg contamination of aquatic systems. Mercury concentration in sh depends on species feeding strategies and, within a species, the age and size of the sh as well as water parameters related to acidity and Hg speciation. Therefore, contamination of sh-eating populations will depend not only on the quantity of sh consumed, but also on the species of choice. MMHg bioaccumulation in aquatic systems varies considerably with food-chain structure and length. It can be simplied with plankton at the base (zooplankton as the primary consumer), followed by small forage shes (secondary consumers) and piscivorous shes (tertiary consumers) where higher MMHg concentrations are expected (Nichols et al. 1999). However, even within a food trophic category, sh species present a variety of feeding strategies that inuence MMHg acquisition. In the Amazonian ecosystem, and the Rio Negro in particular, sh feeding strategies may change on a seasonal basis (Goulding et al. 1988). Fish physiology and environmental interactions with regards to Hg bioaccumulation are not well understood, and in the Amazonian ecosystem, little is known regarding Hg trophic transport along the food chain. Artisanal gold extraction from alluvial deposits in rivers of the Amazonia is carried out without regulations and causes great environmental damage, regardless of country. The methods used involve hydraulic dismount of river terraces and dredging of watercourses. In terrace dismount, miners use water jets to remove land strips, while pumps are used in water courses. Mercury is used to improve gold recovery by amalgamation and all discarded material is put back into rivers, causing turbidity and shoaling (Kligerman et al. 2001). In areas of intensive gold-mining activity, a gradient of sh-Hg concentration was suggested (Uryu et al. 2001). Some have done research, attempting to relate the direct impact of gold-mining activities and sh-Hg concentration (Lima et al. 2000; Hylander et al. 1994). Although the environmental damage due to alluvial gold extraction is enormous, with disruption and de-

236

A. C. Barbosa et al.

struction of aquatic ecosystems (Kligerman et al. 2001), the Hg balance showed the metal discharged from gold-mining activities accounted for less than 1% of the total Hg (Fadini and Jardim 2001). Furthermore, extensive deforestation and agricultural developments (that also occurred in the 1970s and 1980s), were environmentally destructive and may have impacted on the release of Hg from soil (Roulet et al. 1999). Recent work is unveiling the dynamics of environmental Hg in the Amazonian ecosystem showing that Hg is released from sediment and soil. Roulet et al. (2001) showed that Hg concentrations of the Amazon basin rivers (Arapiunas, Tapajos, and Amazon) are governed by the concentrations of suspended particles. They also found that the dominant stock of Hg in the aquatic ecosystem is derived from soil erosion. Indeed, Oliveira et al. (2001) showed that the upper horizon of the Amazon soil receives an important contribution from anthropogenic activity95% of the total. This includes re deforestation, agriculture, and gold-mining activities. However, a complex hydrological cycle in which some ooded plains can be covered by 6 15 m of water for six or more months adds confounding variables to the origin of Hg in the aquatic food chain. Riverbank inhabitants of the Amazon are heavily dependent on sh for their daily nutritional sustenance and therefore are exposed to the easily absorbed MMHg consumed in contaminated sh. For these populations, sh is a culturally important food resource occurring naturally and in abundance. A survey of Rio Negro riparians showed that sh is eaten at least once a day (7.1%), but that most of them (78.6%) consumed it at least twice a day (Barbosa et al. 2001). Per capita sh intake has been estimated as 200 g/day 1 (Barbosa et al. 1995). These populations benet greatly from this excellent source of highquality protein, macronutrients, and micronutrients, especially its unique source of highly unsaturated fatty acids. To identify the sources and fate of environmental Hg, and to enrich the Hg ecotoxicological studies in Amazonia, a monitoring program was established in 1995 in the Rio Negro basin, aimed at sampling soils, water, sediment, air, rainwater, sh, and hair from riverine populations. Studies of hair-Hg speciation of the riverbank population and environmental Hg massbalance have already been published (Barbosa et al. 2001; Fadini and Jardim 2001). In order to understand sh-Hg bioaccumulation we studied muscle-Hg concentrations in a large number of sh representative of all stages of the trophic chain.

pling soils, sediments, water, and respective Hg concentrations. Pertinent data regarding water pH, Hg concentrations, and DOC are summarized in Table 1. The sh were caught by local professional sherman, identied (Ferreira et al. 1998), weighed, and measured for length. Approximately 20 g of muscle samples were immediately taken, frozen to be transported, and analyzed in the laboratory at the University of Braslia. The samples were digested with concentrated HNO3. For 0.4 g of sample, 4 mL of acid was added and then digested for 20 min in a microwave system DGT-100 Provecto Sistemas Analticos, with power between 0 and 800 W. After digestion, the sample is cooled until the temperature reached 25C (room temperature), transferred to a 25-mL volumetric ask. Then, 1.5 mL of a 6% KMnO4 solution and 0.8 mL of a 1% hydroxylamine solution were added to the volumetric ask. All measurements were carried out by cold vapor atomic absorption spectroscopy (AAS-CV), using a Hg monitor of LDC Analytical, Model 1255. All glassware used in the analytical protocol was washed clean, rinsed with both KOH and double-distilled water, and left to rest in 50% HNO3 for 24 h. It was then rinsed again in double-distilled water, and dried at 100C for 12 h. Precision and accuracy of Hg determinations were assured by the use of internal standards prepared in our laboratory and used in intercalibration exercises among Brazilian laboratories. Our laboratory is also a participant in the Mercury Quality Assurance Program for Fish (Department of Fisheries and Oceans, Central and Artic Region, Fresh Water Institute, Winnipeg, Canada), that began in 1992. Our performance was considered acceptable with 94% 2 s.d. (n 16) for sh. Summarized data (mean, s.d., and ranges), Pearson correlation between variables, and analysis of variance, were performed using a SAS (SAS Institute, Cary, NC, USA). We also summarized and compared data of total sh-Hg concentration from other water bodies in Amazonia and Central Brazil.

Results
Results of sh-Hg concentrations and respective range of sh weight are organized according to feeding strategies in Table 2. Fish weight and Hg concentrations in this habitat show a wide variation. However, the species at the top of the food chain had the highest concentrations of muscle-Hg. Median Hg concentrations were higher in predatory and lowest in herbivores species. In the piscivorous category, Tucunares (Cichla spp.), Piranhas (Serrasalmus spp.), and Peixe Cachorro (Hydrolycus scomberoids) comprised 70% of the catch and showed the highest mean Hg concentrations. Bioaccumulation of Hg is further illustrated by its distribution according to sh trophic level presented in Figure 2. Most of the sh (75%) in the herbivorous group showed Hg concentrations below 100 ng/g 1, while most of the samples (65%) in the piscivorous group showed Hg concentrations above 400 ng/g 1. A summary of physicochemical characteristics of the sampled aquatic habitat is integrated with sh-Hg concentrations in Figure 3. Abiotic parameters that inuence sh-Hg concentrations showed variability among sites with regards to pH (4.09 6.31), DOC (1.8515.3 mg/L 1), and water-Hg concentration (3.4 11.9 g/L 1). Sorting sh-Hg by feeding strategy did not show a consistent trend as a function of these water parameters, except for the piscivorous category. Collectively, predatory species showed a tendency to increase muscle-Hg with water-Hg concentrations and dissolved organic carbon (DOC). In

Materials and Methods

The Rio Negro basin has distinctive physicochemical characteristics and does not have a recent history of gold-mining activities along its tributaries (Figure 1). Its catchment spreads over an area of 690 103 km2, representing around 14% of the total area of the Brazilian Amazon, and it has well-dened seasons of high and low waters, with an average annual rainfall of 2000 mm. The Rio Negro joins the Rio Solimoes to form the Rio Amazonas (Figure 1), with an average ow of 29,000 m3/s 1 at Manaus. As part of an ongoing monitoring project for Hg contamination of the Amazon basin we surveyed the Rio Negro from March of 1998 to August of 2001. We sampled sh at nine specic locations (see map, Figure 1) and determined total Hg in sh species. The sh sampled were representative of the feeding hierarchy: herbivorous, detritivorous, omnivorous, and piscivorous. A previous publication (Fadini and Jardim 2001) described the protocol for sam-

Mercury Biomagnication in Rio Negro, Brazil

237

Fig. 1. Map of Rio Negro basin indicating sampling sites

Table 1. Location and physicochemical characteristics of the Rio Negro sampling sites Water parameters Local names (1) (2) (3) (4) (5) (6) (7) (8) (9) Rio Marauia Rio Padauar Rio Demini Foz do Rio Demini Barcelos Carvoeiro Vila da Cota Camanau Maependi Latitude S 0.29910 0.13460 0.44400 0.77040 0.96676 1.33350 1.24862 1.95500 2.12905 Longitude W 65.20425 64.10531 62.86000 62.93433 62.90600 62.06350 61.88785 61.20983 61.03116 Hg (ng/kg 4.68 4.25 3.4 3.4 7.6 9.5 4.23 4.8 6.9
1

DOC (mg/L 6.0 9.5 6.0 3.8 15.3 10.7 1.9 13.1 15.4

pH 5.48 4.09 4.93 4.93 4.2 4.8 6.3 5.3 4.8

Site numbers correspond to Figure 1. DOC is dissolved organic carbon.

spite of a three-fold increase in water-Hg concentration and a eight-fold increase in DOC, there was no apparent impact of muscle-Hg concentrations of nonpredatory species. The correlation data between sh weight and Hg concentration by species is presented in Table 3. Signicant correlations were seen for the piscivorous Serrasalmus spp. (n 326; r 0.3977; p 0.0001), Cichla spp. (n 125; r 0.4600; p 0.0001), and Pimelodus spp. (n 12; r 0.8299; p 0.0008). Among other sh groups, signicant correlations were seen positively for Triportheus elongates (n 33; r 0.4588; p 0.0073), and negatively for Potamorhina latior (n 30; r 0.3763; p 0.0404). To consolidate information on sh-Hg concentration in the

Amazon, a summary of reports is organized in Table 4 as a function of water bodies (rivers, lakes, and reservoirs). No salient feature in sh-Hg concentration can be distinguished from rivers with intense gold-mining activities.

Discussion
The sh-Hg concentrations in the nonpolluted freshwater of the Rio Negro habitat are in the upper range of values reported for the Amazon rain forest and central wetlands with recent history of gold mining (Table 4). A wide variation of sh-Hg bioac-

238

A. C. Barbosa et al.

Table 2. Fish weight and total mercury concentration in sh species according to feeding strategies Weight (g) Common name Piscivorous Agulha Barba-chata Bico-de-pato Jacunda Mandi Mandube Peixe-cachorro Pescada Piranha Sorubim Trara Tucunare Omnivorous Acara Acara preto Acaras Aracu Cangat Orana Piaba Sardinha Detritivorous Acari-bodo Arraia Branquinha Jaraqui Herbivorous Cubiu-orana Pacu Pacu branco Scientic name Potamorrhaphis spp. Goslinia platynema Sorubim lima Crenicichla reticulata Pimelodus spp. Ageneiosus brevilis Hydrolycus scomberoides Plagioscion spp. Serrasalmus spp. Pseudoplatystoma fasciatum Hoplias malabaricus Cichla spp. Acarichthys heckellii Heros spp. Satanoperca spp. Leporinus spp. Parauchenipterus galeatus Hemiodus immaculatus Hemigrammus spp. Triportheus elongatus Liposarcus pardalis Potamotrygon scorbina Potamorhina latior Semaprochilodus taeniurus Anodus melanopogon Mylossoma aureum Myleus torquatus n 16 7 4 4 12 28 83 26 327 3 24 26 15 28 20 44 10 18 4 33 3 8 30 18 4 25 34 Min 68 269 61 296 17 52 19 202 18 433 124 102 102 46 65 40 143 22 39 59 54 1,500 41 248 104 94.0 70.0 Max 420 1,235 121 680 745 390 783 790 1,653 830 2,020 5,000 560 525 469 682 217 99 66 442 690 15,000 334 570 149.0 836.0 324.0 Median 1,065.5 881.4 226.7 181.0 125.2 493.2 652.4 401.5 320.0 349.4 301.7 448.5 132.0 179.0 180.0 95.9 93.1 102.0 203.5 32.0 64.0 112.4 85.9 106.8 39.5 18.1 22.6 Hg range (ng/g 1) 254.81,408.0 448.51,572.7 113.9388.6 153.0606.8 60.42,024.0 242.61,321.2 114.55,437.4 160.41,064.3 15.01,713.7 204.40436.0 120.11,592.1 39.22,441.0 69.0778.4 10.0506.6 47.8677.0 10.0516.0 40.4114.7 32.0170.0 124.0368.0 5.3476.5 21.071.0 47.1189.1 23.0450.1 16.2207.3 24.349.7 2.2186.0 10.085.0

cumulation is expected because of the diversity in sh feeding strategies involving mobility, foraging location, migratory traits, as well as differences in Hg metabolism for each species. In spite of a wide range of Hg bioconcentration between and within sh trophic rank, this work clearly shows that the dominant feature in Hg propagation is an increase in muscle-Hg with increasing hierarchy of sh feeding strategy. Most of the Hg found in sh is MMHg and in the Amazonian ecosystem it represents 62100% of total Hg (Lacerda et al. 1994; Akagi et al. 1995; Malm et al. 1995b; Palheta and Taylor 1995; Guimaraes et al. 1999; Brabo et al. 2000). Fish-Hg variation in rivers of the Amazon basin is wide and a systematic comparison is difcult, not only because of habitat diversity, but also because of incomplete information (Table 4). Fish local names may change from region to region, and there is also insufcient knowledge of sh feeding hierarchy, that is, some sh may change food acquisition strategies during its life cycle (Goulding et al. 1988). Coupled with that, most studies do not report sh size (length or weight), nor age, important in controlling the random nature of sh sampling. Such difculties limit comparison and causes confusion when drawing conclusions on the diversity of the Amazonian ecosystem. The complexity of inherent variability of sh-Hg bioaccumulation is best illustrated by reports on Tucunares (Cichla spp.), a sh commonly found and consumed throughout Amazonia (Table

4). With regards to sample size, while we analyzed 126 specimens, others studied modest numbers such as four (Akagi et al. 1995), six (Bidone et al. 1997a; Guimaraes et al. 1999), ten (Santos et al. 2000a), 11 (Hacon et al. 1997), or 17 (Santos et al. 2000b). Such small numbers compromise conclusions that gold-mining activities impact on Hg concentrations for this species (Guimaraes et al. 1999). In tertiary-consumer sh, Hg is frequently reported to increase as a function of age and size (length or weight). This association was found in the freshwater Amazon ecosystems, although in a seemingly inconsistent way. While no signicant correlation was reported for Tucunares in impacted and nonimpacted lakes of north Amazon (Guimaraes et al. 1999) or in the Rio Madeira (Malm et al. 1997), some studies reported a positive correlation for Tucunares (Castilhos et al. 2001), Trara, and Caratinga (Brabo et al. 2000). Contrary to this, signicant negative correlation between sh-Hg concentration and sh weight, were reported by Lima et al. (2000) in piscivorous species (Trara and Tucunare). We also found a positive correlation between weight and Hg concentration for Tucunares and Piranhas (Table 3) with a larger number of samples (126 and 327, respectively), but not for Trara (n 24). We found a negative correlation only for Branquinha (Potamorhina latior), a detritivorous sh. Irrespective of trophic level, sh-Hg concentration varies considerably, even when sh age (size or weight) is considered. In this study, not all predator

Mercury Biomagnication in Rio Negro, Brazil

239

Fig. 2. Percent distribution as a function of Hg concentrations in sh samples organized by feeding strategies

species showed a signicant positive correlation between muscle-Hg and sh weight (Table 3). This may be due to variability in prey-Hg levels, the Hg chemical form in the water column and in

food material, or differences in sh physiology (Hg-depuration rate). Rate constant of Hg loss in studied species is very low and inversely correlated with sh size (Trudel and Rasmussen 1997). Stafford and Haines (2001) discussed circumstances under which sh-Hg bioconcentration and biodilution occur and recommended that sh age should be included in contamination studies. The origin of Hg in the Rio Negro waters does not include industrial activities and there is no history of intensive goldmining in upstream tributaries. The physicochemical characteristics of the Rio Negro basin were discussed in a parent publication indicating that 99.74% of its Hg would be of natural sources (Fadini and Jardim 2001). Therefore, sh-Hg contamination cannot be directly related to Hg released during gold amalgamation. Instead, other anthropogenic activities related to deforestation by re and agricultural practices may play an important role in the release of natural Hg from soils. As a consequence, abiotic parameters (pH, DOC, water-Hg concentrations) are the most important determinants of Hgmethylation and its acquisition and retention by sh. In the Amazonian waters of Surinam, Mol et al. (2001) observed that piscivorous species from freshwater habitats showed higher Hg concentrations than those from estuarine and ocean habitats, while estuarine habitats showed the lowest Hg concentrations for both piscivorous and nonpiscivorous species. Natural Hg release and methylation potential are associated with geochemistry and complex human activities, while Hg acquisition, accumulation, and biomagnication by sh are associated with the Hg chemical form in food material and sh physiology. The Hg released during alluvial goldmining was previously thought to contaminate the river biota, especially sh, and ultimately the riverine population through their sh consumption. Such studies used Hg concentration in sh muscle and human hair as biomarkers of Hg pollution following gold-mining activities in the Amazonian rivers. A gradient of sh-Hg concentrations were proposed for the Rio Tapajos (Uryu et al. 2001)presently showing the most intense gold mining activityto address Hg pollution that was assumed to be caused by gold-mining activity. In the biochemical complexity of a tropical rain forest like the Amazon, environmental predictors of Hg acquisition by primary consumers, or biomagnication in tertiary consumers (predatory sh), are difcult to trace to Hg released from gold-mining activity. Furthermore, the rapid changes in land use coincided with the alluvial gold rush. Table 4 shows that the ranges of mean sh-Hg concentrations in the Tapajos and other impacted rivers are not sys tematically higher. Mercury concentrations in freshwater sh of the Rio Negro basin are comparable to those of other Amazonian rivers (Table 4). Kehrig and Malm (1999) showed that Hg concentrations in piscivorous species of the Balbina reservoir (also not impacted by gold mining, but in the Rio Negro catchment) are among the highest in the Amazonia. The interplay of Hg geochemistry among sampling sites showed that gradients of pH, water-Hg, and DOC were not coincident (Figure 3). Therefore sh-Hg (by food-chain rank) and site-dependent, Hg geochemistry could be singled out. It is apparent that preformed MMHg from animal sources is the primary modulator of sh-Hg bioconcentration. The gradients of either Hg (threefold) or DOC (eightfold) in the water column had no proportional impact on nonpredatory species. Instead a tenfold increase in sh-Hg concentration between extremes of feeding ranks (herbivorous versus piscivorous) seemed to exist along all levels

240

A. C. Barbosa et al.

Fig. 3. Hg concentrations in sh samples organized according to water parameters

Mercury Biomagnication in Rio Negro, Brazil

241

Table 3. Summary of data correlating weight and mercury concentration in sh Common name Agulha Cachorro Mand Mandube Pescada Piranha Trara Tucunare Acara Acaras Acara-preto Aracu Orana Sardinha Braquinha Jaraqui Pacu Pacu branco P piscivorous; O Scientic name Potamorrhaphis spp. Hydrolyeus scomberoides Pimelodus spp. Ageneiosus brevilis Plagioscion spp. Serrasalmus spp. Hoplias malabaricus Cichla spp. Acarichthys heckellii Satanoperca spp. Heros spp. Leporinus spp. Hemiodus immaculatus Triportheus elongatus Potamorhina latior Semaprochilodus taeniurus Mylossoma aureum Myleus torquartus omnivorous; D detritivorous; H Feeding strategy P P P P P P P P O O O O O O D D H H herbivorous. n 16 83 12 28 26 326 24 125 15 20 28 44 18 33 30 18 25 34 r 0.2514 0.064 0.8299 0.3171 0.0129 0.3977 0.3629 0.4600 0.4755 0.2448 0.1644 0.3987 0.1038 0.4588 0.3763 0.0413 0.1318 0.2366 p 0.3477 0.9536 0.0008 0.1002 0.9502 0.0001 0.0813 0.0001 0.0733 0.2983 0.4033 0.0073 0.6820 0.0073 0.0404 0.8709 0.5301 0.1780

of the water physicochemical parameters (pH, water-Hg, and DOC). In this context, the constant challenge of water-Hg and acidity is much less important than the preformed MMHg present in the food material. Indeed, as already observed, Hg uptake for all sh species occurs mainly from food, with biomagnication being dominant in predatory species (Vigh et al. 1996). The retention of Hg in muscle is operative in aquatic and terrestrial species, man included. Reviewing the subject, Sweet and Zelikoff (2002) showed that sh is more tolerant to Hg toxicity than humans and coincidently shows a half-life of Hg in muscle one order of magnitude higher than humans. Fish consumption advisories must take into consideration the fundamental aspects of human ecology of the Amazonias riverine population, which depends on sh for food security. As a constant dietary item, sh is a good source of nutrients, high in protein and available lysine, and with a good biological value (Batterham et al. 1979) that balances the protein-poor starchy food consumed by many (Araujo et al. 1975). With regards to other nutrients, small sh are rich in available iodine (Toure et al. 2001) and other trace elements (copper, zinc, iron, and manganese), as well as calcium (Larsen et al. 2000), which compensates for the lack of dairy products in riverines diet (Giugliano et al. 1984). Fish is also known to enhance zinc (Garcia-Arias et al. 1993) and vegetableiron absorption (Larysse et al. 1968). Amazonian sh of higher trophic level accumulate selenium (Dorea et al. 1998), which counteracts the toxic effect of Hg. According to Inhamuns and Franco (2001), Amazonian sh are also a specic source of omega-3 polyunsaturated fatty acids (PUFA; decosahexanoic [22: 6] acid and eicosapentaenoic [20:5] acid). As well as this, in the human ecology of the Rio Negro ribeirinhos, sh have also medicinal uses (Begossi et al. 2000). As a result of legitimate concern over the extensive environmentally-destructive alluvial gold mining and Hg release into the Amazon rivers, there are sh consumption advisories aimed at diminishing risk of MMHg intake. Work with the sh-eating population of the Amazon has suggested that Hg contamination could be reduced by changing the pattern of sh consumption

(Akagi et al. 1995; Boischio et al. 1995; Lebel et al. 1997). Balancing the risks and benets of sh consumption for indigenous populations was highlighted by Egeland et al. (1997). Clarkson (1995) casts doubt on the application of advised safe upperlimits of MMHg naturally bioaccumulated in sh. Indeed the MMHg concentrations in aquatic mammals, such as the whale (Clarkson 1995), are relatively higher than in most piscivorous species reported in the Amazon basin (Table 4). The threat of destruction of the fragile Amazonian ecosystem by gold-mining activities is serious enough and raised world concern for curbing additional release of Hg (Hylander 2001). Nevertheless, advisories of sh restriction for the ribeirinhos may prove disruptive for their strategies of food security and maintenance of nutritional status with unforeseen consequences. With regard to their current health problems, changes of sh consumption for the Rio Negro ribeirinhos does not guarantee them better health.

Conclusions
The Rio Negro has no history of gold-mining activity. Nevertheless, its sh-Hg concentrations are among the highest in the Amazon rain forest. Fish-Hg concentration is a consequence of preformed MMHg in the natural material of the sh food chain. Although relatively high sh-Hg concentrations above 500 ng/g 1 (wet weight) do occur in older (larger) predatory species, Hg concentrations above 1,000 ng/g 1 (wet weight) are infrequently caught when subsistence shing. Given the health benets of sh consumption to the Rio Negro ribeirinhos, advisories for changes in food habits that curtail sh intake may bring unforeseen consequences.

Acknowledgments. We thank the Rio Negro population for support and participation in the study. We also thank Ms. Fatima Barretto of

242

A. C. Barbosa et al.

Table 4. Summary of Hg concentrations (ng/g tanal) Sampling site Amapa Rivers Guimaraes et al. 1999 Feeding strategy P P P P O

) in muscle of sh caught in the water bodies of the Amazonia and central wetlands (Pan-

Range of mean Hg (342650) (240269) 197549 305742 351,225

Local name Tucunare (range) Tucunare (range) Trara, Piranha, Tucunare Tucunare, Piranha, Pirarucu, Aruana, Itu, Jacunda, Jeju, Uena Aracu, Cachorro de Pedra, Cara, Manduba, Matrinxa, Pacu Branco, Taumata Tucunare, pescada Acara Tucunare, Cachorro, Piranha Acara-Tinga Tucunare, Cachorro, Piranha, Trara Acara Surubim Sardinha Trara, Pescada Sardinha Mandube, Trara, Surubim, Bico de Pato, Mandi, Pescada NK NK NK NK Piranha, Trara, Pescada, Tucunare Tucunare, Traira, Piranha, Mandube Orana preta, Sardinha (See details in Boischio et al. 2000) (See details in Boischio et al. 2000) Cara-Acu, Filhote, Piranha, Cachorro, Pescada, Jacundaa, Apapa, Surubim, Barba Chata, Tucunare, Bico de Pato, Jeju, Piramitaba, Dourada, Jandira, Caparar, Trara, Pirarucu Tamoata, Cascudo, Jaraqu, Curimata, Bodo, Branquinha, Ubarana, Mapara Cubiu, Mandi, Cara, Cuiu-cuiu, Sardinha, Aruana, Pintadinho, Pirarara, Aracu, Matrinxa Pacu, Bacu, Pirapitinga, Jatuarana, Tambaqui Piranha, Barba-Chata, Dourada, Tucunare Aruana, Mandi, Sardinha, Aracu, Cara, Jatuarana, Pacu, Pirapitinga, Tambaqu, Curimata, Jaraqu, Cascudo, Branquinha NG NG Tucunare, Cachorro, Piranha, Mapara, Pacu, Sardinha, Piau Branquinha, Matrinxa Dourado, Filhote, Piranna, Tucunare, Pirarucu Curimata, Jatuarana NG NG NG Cachorro, Bagre, Dourado, Pintado, Surubim Jatara, Tambaqui, Pacu Cachorro, Bagre, Dourado, Pintado, Surubim

Weight (kg) (0.311.05) (0.40.54) NG NG NG

Bidone et al. 1997b

Andira River Eve et al. 1996 Balbina Reservoir Kehrig et al. 1998 Kehrig and Malm 1999 Colombian Rivers Olivero et al. 1997 Olivero and Solano 1998 Olivero et al. 1998 French Guyana Rivers Frery et al. 2001*

P O P O P O P N-P P N-P P P O D H P P N-P

5101,390 100 130700 60 320 60 5161 48150 195322 386 501,130 (1,4258,736) (256556) (116242) (8115) 1801,150 2201,113 99230

NG

NG NG

NG NG NG NG NG NG NG NG NG NG 0.0186 NG

Mol et al. 2001 Richard et al. 2000 Madeira Basin Barbosa et al. 1995 Barbosa et al. 1997 Boischio and Henshel 2000

2301340

0.1826

D O H P O H D P P P O D P NP P O P P H P

110900 1701440 ND360 220740 1201,900 100360 110240 670 846 570 150 180 280420 702,700 80210 700 130 846 491,522 994 7121,819

0.070.53 0.114.9 0.318.3 0.235 0.10.97 1.1408.290 0.070.56 NG NG NG NG 3545 cm 0.4520 0.3430.520 NG NG NG 0.1216.3 0.74.5 NG

Boischio et al. 1995

Gali 1997 Guimaraes et al. 1999 Kehrig and Malm 1999

Lechler et al. 2000 Malm et al. 1990 Malm et al. 1995a Malm et al. 1997 Maurice-Bougoin et al. 2000 Maurice-Bougoin et al. 1999

Mercury Biomagnication in Rio Negro, Brazil

243

Sampling site

Feeding strategy O D P NP P NP P NP

Range of mean Hg 8129 55 702,700 210 702,700 80210 3401,330 120490 391,544 51102 610 90 1811635 18203 803,820 170280 100 100690 52100 3787 174419 95120 42112 60690 16100 36149 1284 NG NG 2802,750 8080 3001000 10170 140810 50 90800

Local name Griso, Panete, Sabalo, Pacu Sabalo Sabalo Dourado, Filhote, Pintado, Pirarucu Curimata, Dourado, Filhote, Pintado, Tucunare, Pirarucu Curimata, Jatuarana Tucunare, Pintado, Piranha, Peixe-cachorro, Traira, Apapa, Dourado, Filhote Acara, Mapara, Mandi, Curimata Surubim, Piranha Mandi, Branquinha Tucunare, Aruana, Cachorro, Piranha Branquinha, Cara, Aracu (See Table 2) (See Table 2) Dourada, Jau, Piraba, Mandube, Cachorro, Trara, Apapa, Pescada, Tucunare, Filhote, Pirarucu Acara, Aruana Pacu Tucunare, Piranha, Cachorro, Surubim, Jacunda, Mandi, Pescada, Trara, Piracmuntaba Acaratinga, Aracu, Matrinxa Pacu, Jaraqui (same as Brabo et al. 2000) Tucunare, Trara, Barbado, Piranha, Surubim, Aruana, Caratinga, Jandia, Aracu, Mandia Jaraqu, Pacu, Apapa, Cachorro, Dourada, Filhote, Jacunda, Mandi, Pescada, Piramutaba, Saranha, Piranha, Surubim, Trara, Tucunare Acara-Acu, Acara-Tinga, Aracu, Curimata Jaraqui, Mapara, Matrinxa Pacu, Tambaqui Tucunare (See Castilhos et al. 2001) Piraiba Jau, Pescada, Pintado, Dourada, Trara, Piranha, Tucunare Pacu, Curimata. Tucunare, Jau, Pintado Pacu, Matrinxa, Curimata Tucunare, Trara Aracu Apapa, Barbado, Filhote, Mandi, Pescada, Piracatinga, Piranha, Sarda, Surubim, Jiju, Tucunare, Jacunda, Dourada, Cachorro, Trara, Pirarucu Aracu, Caratinga, Pacu Caraucu, Mandube, Saranha, Sardinha Dourada, Cachorro, Pescada, Piranha, Sarda, Surubim, Tucunare Pacu, Pirapitinga, Tambaqu, Aracu NG NG Pirarucu NG Tucunare, Trara Acara, Cachorrinho de Padre, Trara Pacu, Piaba NG

Weight (kg)

Pfeifer et al. 1989 Pfeifer et al. 1991 Reuther 1994

NG NG NG NG NG NG NG NG NG

Manu River Gutleb et al. 1997 Negro Basin Kehrig and Malm 1999 This study Tapajos Basin Akagi et al. 1995

P NP P D P NP P O H P O H

0.140 0.160.515 1.43 NG

Bidone et al. 1997a

Brabo et al. 1999 Brabo et al. 2000

Castilhos et al. 1998

P O H P O D H P P P H P H P H P

NG

NG

Castilhos et al. 2001 Castilhos and Bidone 2001 Hacon et al. 1997

NG

Hacon et al. 2000 Kehrig and Malm 1999 Lebel et al. 1997

0.32.8 0.13.5 NG NG NG

Lima et al. 2000

H O P H P P P NG P O H P

40110 130400 125306 3069 690 482 (25776) 92,575 110610 30210 40 190 529

0.311,420 0.2280.71 NG NG 20 NG NG

Malm et al. 1995b Malm et al. 1997 Martinelli and McGrath 1999 Moreton and Delves 1998 Palheta and Taylor 1995

Santos et al. 2000a

NG

244

A. C. Barbosa et al.

Sampling site Santos et al. 2000b

Feeding strategy NP P D H P

Range of mean Hg 106 375839 121 5184 922,348

Local name NG (non-piscivorous) Dourada, Filhote, Surubim, Pescada, Tucunare, Sarda. Curimata Aracu, Pacu, Tambaqui Apapa, Peixe-cachorro, Traira, Pescada, Piranha, Piramutaba, Barbado, Piraiba, Pirarucu, Jau, Tucunare, Dourado, Mandube, Pintado. Caratinga, Cabeca-gorda, Piau, Curvina, Mapara, Aruana. Acari, Jaraqui, Pacu, Tambaqui, Branquinha, Aracu, Curimata. Trara, Cachorro, Jau, Piranha Matrinxa, Pacu Branco Tucunare, Piranha, Pescada Piau Curimata Piaba Trara Mand Piranha, Surubim, Pescada, Trara Mandi Pintado Pintado Surubim Pintado Piranha Piranha, Bagre, Trara Cara, Piranha Surubim non-piscivorous; NG not given; NK

Weight (kg) 0.53.74 0.7 0.3551.702 NG

Uryu et al. 2001

O H Tocantins Basin Lacerda et al., 1994 Tucuru Reservoir Porvari 1995 P H P O D H P O P O P P P P P P P P

110930 8113 1301560 10160 4102600 220 60 390 120 520 1,04012,310 7,780 130540 120300 263429 76403 114813 70120 40160 68189

NG

NG

Venezuela Rivers Nico and Taphorn 1994 Central wetlands (Pantanal) Alho and Vieira 1997 Hylander et al. 1994 Hylander et al. 2000

(68277 mm) 240 mm NG NG NG

Kehrig and Malm 1999 Lacerda et al. 1991 Parana Moraes et al. 1997

NG NG 0.2480.616 not known in

P piscivorous; O omnivorous; D detritivorous; H herbivorous; N-P Portuguese; * dry weight; () range of individual values.

the Department of Chemistry for the technical help with analysis of samples and Dr. Coni Pimentel for redactorial suggestions.

References
Akagi H, Malm O, Kinjo Y, Harada M, Branches FJP, Pfeiffer WC, Kato H (1995) Methylmercury pollution in the Amazon, Brazil. Sci Total Environ 175:8595 Alho CJR, Vieira LM (1997) Fish and wildlife resources in the Pantanal wetlands of Brazil and potential disturbances from the release of environmental contaminants. Environ Toxicol Chem 16:7174 Araujo TM, Lago ES, Bion FM, Nascimento JS, Costa LP, Antunes NL, Chaves N, Mello AV (1975) [Biological value of manihot our enriched with sh our and soybean isolated protein, and casein.] Rev Bras Pesqui Med Biol 8:139 142 (in Portugese) Barbosa AC, Boischio AA, East GA, Ferrari I, et al. (1995) Mercury contamination in the Brazilian Amazon. Environmental and occupational aspects. Water Air Soil Poll 80:109 121 Barbosa AC, Garcia AM, Souza JR (1997) Mercury contamination in hair of riverine population of Apiacas Reserve in the Brazilian Amazon. Water Air Soil Poll 97:1 8 Barbosa AC, Jardim W, Dorea JG, Fosberg B, Souza J (2001) Hair mercury speciation as a function of gender, age, and body mass index in inhabitants of the Negro River Basin, Amazon, Brazil. Arch Environ Cont Toxicol 40:435 444

Batterham ES, Murison RD, Lewis CE (1979) Availability of lysine in protein concentrates as determined by the slope-ratio assay with growing pigs and rats and by chemical techniques. Br J Nutr 41:383391 Begossi A, Hanazaki N, Peroni N (2000) Knowledge and use of biodiversity in Brazilian hot spots. Environ Develop Sustain 2:177193 Bidone ED, Castilhos ZC, Cid de Souza TM, Lacerda LD (1997a) Fish contamination and human exposure to mercury in the Tapajos River Basin, Para State, Amazon, Brazil: A screening approach. Bull Environ Contam Toxicol 59:194 201 Bidone ED, Castilhos ZC, Santos TJS, Souza TMC, Lacerda LD (1997b) Fish contamination and human exposure to mercury in Tartarugalzinho River, Amapa State, Northern Amazon, Brazil. A screening approach. Water Air Soil Pollut 97:9 15 Boischio AA, Henshel D (2000) Fish consumption, sh lore, and mercury pollutionrisk communication for the Madeira River people. Environ Res 84:108 126 Boischio AA, Henshel D, Barbosa AC (1995) Mercury exposure through sh consumption by the upper Madeira River population, Brazil1991. Ecosystem Health 1:177192 Brabo ES, Santos ED, Jesus IM, Mascarenhas AF, Faial KF (1999) [Mercury levels in sh consumed by the Sai Cinza indigenous community, Munduruku Reservation, Jacareacanga County, State of Para, Brazil]. Cad Saude Publica 15:325331 (in Portugese) Brabo ES, Santos ED, de Jesus IM, Mascarenhas AFS, Faial KD

Mercury Biomagnication in Rio Negro, Brazil

245

(2000) Mercury contamination of sh and exposures of an indigenous community in Para State, Brazil. Environ Res 84:197203 Castilhos ZC, Bidone ED, Lacerda LD (1998) Increase of the background human exposure to mercury through sh consumption due to gold mining at the Tapajos River region, Para State, Amazon. Bull Environ Contam Toxicol 61:202209 Castilhos ZC, Bidone ED (2001) Hg biomagnication in the ichthyofauna of the Tapajos River Region, Amazonia, Brazil. Bull Envi ron Contam Toxicol 64:693700 Castilhos ZC, Bidone ED, Hartz SM (2001) Bioaccumulation of mercury by Tucunare (Cichla ocellaris) from Tapajos River region, Brazilian Amazon: A eld dose-response approach. Bull Environ Contam Toxicol 66:631 637 Clarkson TW (1995) Environmental contaminants in the food chain. Am J Clin Nutr 61:682S 686S Dorea JG, Moreira MB, East G, Barbosa AC (1998) Selenium and mercury concentrations in some sh species of the Madeira River, Amazon Basin, Brazil. Biol Trace Elem Res 65:211220 Egeland GM, Middaugh JP (1997) Balancing sh consumption benets with mercury exposure. Science 278:1904 1905 Eve E, Oliveira EF, Eve C (1996) The mercury problem and diets in the Brazilian Amazon: Planning a solution. Environ Conserv 23:133139 Fadini PS, Jardim WF (2001) Is the Negro River Basin (Amazon) impacted by naturally occurring mercury? Sci Total Environ 275: 71 82 Ferreira EJG, Zuanon JAS, Santos GM (1998) Peixes Comerciais do Medio Amazonas: Regiao de Santarem, Para. Edicoes IBAMA, Brasilia (in Portugese) Frery N, Maury-Brachet R, Maillot E, Deheeger M, de Merona B, Boudou A (2001) Gold-mining activities and mercury contamination of native Amerindian communities in French Guiana: Key role of sh in dietary uptake. Environ Health Perspect 109:449 456 Gali PA (1997) Contaminacao mercurial em peixes carnivoros dos Rios Madeira, Jaci Parana e Jamari-Rondonia. B.Sc. Thesis. Fed eral University of Rondonia, Porto Velho, Brazil Garcia-Arias MT, Castrillon AM, Navarro MP (1993) Bioavailability of zinc in rats fed on tuna as a protein source in the diet. J Trace Elem Electrolytes Health Dis 7:29 36 Giugliano R, Shrimpton R, Marinho HA, Giugliano LG (1984) Estudos nutricionais das populacoes rurais da Amazonia. II. Rio Ne gro. Acta Amazonica 14:427 449 Goulding M, Carvalho ML, Ferreira EG (1988) Rio Negro, rich life in poor water. SPB Academic Publishing, The Hague Guimaraes JRD, Fostier AH, Forti MC, Mel JA, Kehrig H, Mauro JBN, Malm O, Krug JF (1999) Mercury in human and environmental samples from two lakes in Amapa, Brazilian Amazon. Ambio 28:296 301 Gutleb AC, Schenck C, Staib E (1997) Giant otter (Pteronura brasiliensis) at risk? Total mercury and methylmercury levels in sh and otter scats, Peru. Ambio 26:511514 Hacon S, Rochedo ER, Campos R, Lacerda LD (1997) Mercury exposure through sh consumption in the urban area of Alta Floresta in the Amazon Basin. J Geochem Explor 58:209 216 Hacon S, Yokoo E, Valente J, Campos RC, da Silva VA, de Menezes ACC (2000) Exposure to mercury in pregnant women from Alta Floresta-Amazon Basin, Brazil. Environ Res 84:204 210 Hylander LD, Silva EC, Oliveira LJ, Silva AS, Kuntze EK, Silva DX (1994) Mercury levels in Alto-PantanalA screening study. Ambio 23:478 484 Hylander LD, Pinto FN, Guimaraes JR, Meili M, Oliveira LJ, Castro e Silva E (2000) Fish mercury concentration in the Alto Pantanal, Brazil: Inuence of season and water parameters. Sci Total Environ 261:9 20 Hylander LD (2001) Global mercury pollution and its expected de-

crease after a mercury trade ban. Water Air Soil Pollut 125:331 344 Inhamuns AJ, Franco MRB (2001) Composition of total, neutral and phospholipids in mapara (Hypophtalmus sp.) from the Brazilian Amazonian area. J Agric Food Chem 49: 4859 4863 Kehrig HA, Malm O, Moreira I (1998) Mercury in a widely consumed sh Micropogonias furnieri (Demarest, 1823) from four main Brazilian estuaries. Sci Total Environ 13:263 271 Kehrig HD, Malm O (1999) Methylmercury in sh as a tool for understanding the Amazon mercury contamination. Appl Organomet Chem 13:689 696 Kehrig HD, Malm O, Akagi H, Guimaraes JRD, Torres JPM (1998) Methylmercury in sh and hair samples from the Balbina reservoir, Brazilian Amazon. Environ Res 77:84 90 Kligerman DC, La Rovere EL, Costa MA (2001) Management challenges on small-scale gold mining activities in Brazil Env Res 87:181198 Lacerda LD, Bidone ED, Guimaraes AF, Pfeiffer WC (1994) Mercury concentrations in sh from the Itacaiunas-Parauapebas River system, Carajas region, Amazon. An Acad Bras Cienc 66:373379 Lacerda LD, Pfeiffer WC, Marins RV, Sousa CMM, Rodrigues S, Bastos WR (1991) Mercury dispersal in water, sediment, sand aquatic biota of a gold mining tailing deposit in Pocone, Brazil. Water Air Soil Pollut 55:238 294 Larsen T, Thilsted SH, Kongsbak K, Hansen M (2000) Whole small sh as a rich calcium source. Br J Nutr 83:191196 Layrisse M, Martinez-Torres C, Roche M (1968) Effect of interaction of various foods on iron absorption. Am J Clin Nutr 21:1175 1183 Lebel J, Roulet M, Mergler D, Lucotte M, Larribe F (1997) Fish diet and mercury exposure in a riparian Amazonian population. Water Air Soil Pollut 97:31 44 Lechler PJ, Miller JR, Lacerda LD, Vinson D, Bonzongo JC, Lyons WB, Warwick JJ (2000) Elevated mercury concentrations in soils, sediments, water, and sh of the Madeira River basin, Brazilian Amazon: A function of natural enrichments? Sci Total Environ 260:8796 Lima APD, Muller RCS, Sarkis JED, Alves CN, Bentes MHD, Brabo E, Santos ED (2000) Mercury contamination in sh from Santarem, Para, Brazil, Environ Res 83:117122 Lodenius M, Malm O (1998) Mercury in the Amazon. Rev Environ Contam Toxicol 157: 2552 Malm O, Castro MB, Bastos WR, Branches FJP, Guimaraes JRD, Zuffo CE, Pfeiffer WC (1995a) An assessment of Hg pollution in different goldmining areas, Amazon Brazil. Sci Total Environ 175:127140 Malm O, Branches FJ, Akagi H, Castro MB, Pfeiffer WC, Harada M, et al. (1995b) Mercury and methylmercury in sh and human hair from the Tapajos river basin, Brazil. Sci Total Environ 175:141 150 Malm O, Guimaraes JRD, Castro MB, Bastos WR, Viana JP, Branches FJP, et al. (1997) Follow-up of mercury levels in sh, human hair and urine in the Madeira and Tapajos basins, Amazon, Brazil. Water Air Soil Pollut 97:4551 Malm O, Pfeiffer WC, Souza CMM, Reuther R (1990) Mercury pollution due to gold mining in the Madeira River, Brazil. Ambio 19:1115 Martinelli MC, McGrath D (1999) Mercury accumulation in the pirarucu Arapaima gigas Cuvier (1918) in the Lower Amazonian varzea. Bol Mus Par Emilio Goeldi-Zool 15:722 Maurice-Bourgoin L, Quiroga I, Guyot JL, Malm O (1999) Mercury pollution in the Upper Beni River, Amazonian Basin: Bolivia. Ambio 28:302306 Maurice-Bourgoin L, Quiroga I, Chincheros J, Courau P (2000) Mercury distribution in waters and shes of the upper Madeira rivers and mercury exposure in riparian Amazonian populations. Sci Total Environ 260:73 86

246

A. C. Barbosa et al.

Mol JH, Ramlal JS, Lietar C, Verloo M (2001) Mercury contamination in freshwater, estuarine, and marine shes in relation to smallscale gold mining in Suriname, South America. Environ Res 86:183197 Moreton JA, Delves HT (1998) Simple direct method for the determination of total mercury levels in blood and urine and nitric acid digests of sh by inductively coupled plasma mass spectrometry. J Anal At Spect 13:659 665 Moraes LAF, Lenzi E, Luchese EB (1997) Mercury in two sh species from the Parana River oodplain. Parana, Brazil. Environ Pollut 98:123127 Nico LG, Taphorn DC (1994) Mercury in sh from gold-mining regions in the upper Cuyuni River system, Venezuela. Fresenius Environ Bull 3:287292 Nichols J, Bradbury S, Swartout J (1999) Derivation of wildlife values for mercury. J Toxicol Environ Health B Crit Rev 2:325355 Oliveira SMB, Mel AJ, Fostier AH, Forti MC, Favaro DIT, Boulet R (2001) Soils as an important sink for mercury in the Amazon. Water Air Soil Pollut 126:321337 Olivero J, Navas V, Perez A, Solano B, Acosta I, Arguello E, Salas R (1997) Mercury levels in muscle of some sh species from the Dique Channel, Colombia. Bull Environ Contam Toxicol 58:865 870 Olivero J, Solano B (1998) Mercury in environmental samples from a waterbody contaminated by gold mining in Colombia, South America. Sci Total Environ 217:83 89 Olivero J, Solano B, Acosta I (1998) Total mercury in muscle of sh from two marshes in goldelds, Colombia. Bull Environ Contam Toxicol 61:182187 Palheta D, Taylor A (1995) Mercury in environmental and biological samples from a gold mining area in the Amazon region of Brazil, Sci Total Environ 168:63 69 Pfeiffer WC, de Lacerda LD, Malm O, Souza CMM, da Silveira EG, Bastos WR (1989) Mercury concentrations in inland waters of gold-mining areas in Rondonia, Brazil. Sci Total Environ 87/88: 233240 Pfeiffer WC, Malm O, Souza CMM, de Lacerda LD, da Silveira EG, Bastos WR (1991) Mercury in the Madeira River ecosystem, Rondonia, Brazil. Forest Ecol Manag 38:239 245 Porvari P (1995) Mercury levels of sh in Tucurui hydroelectric

reservoir and in River Moju in Amazonia, in the state of Para, Brazil. Sci Total Environ 175:109 117 Reuther R (1994) Mercury accumulation in sediments and sh from rivers affected by alluvial gold mining in the Madeira River Basin, Brazil. Environ Monit Assess 32:239 258 Richard S, Arnoux A, Cerdan P, Reynouard C, Horeau V (2000) Mercury levels of soils, sediments and sh in French Guiana, South America. Water Air Soil Pollut 124:221244 Roulet M, Lucotte M, Canuel R, Farella N, Goch YGD, Peleja JRP, et al. (2001) Spatio-temporal geochemistry of mercury in waters of the Tapajos and Amazon rivers, Brazil. Limnol. Oceanogr 46: 11411157 Roulet M, Lucotte M, Farella N, Serique G, Coelho H, Passos CJS, et al. (1999) Effects of recent human colonization on the presence of mercury in Amazonian ecosystems. Water Air Soil Pollut 112: 297313 Santos EO, Jesus IM, Brabo ES, Loureiro EC, Mascarenhas AF, Weirich J, et al. (2000a). Mercury exposures in riverside Amazon communities in Para, Brazil. Environ Res 84:100 107 Santos LS, Muller RC, de Sarkis JE, Alves CN, Brabo ES, Santos EO, Bentes MH (2000b) Evaluation of total mercury concentrations in sh consumed in the municipality of Itaituba, Tapajos River Basin, Para, Brazil. Sci Total Environ 261:1 8 Stafford CP, Haines TA (2001) Mercury contamination and growth rate in two piscivore populations. Environ Toxicol Chem 20: 2099 2101 Sweet LI, Zelikoff JT (2002) Toxicology and immunotoxicology of mercury: A comparative review in sh and humans. J Toxicol Environ Health 4:161205 Trudel M, Rasmussen JB (1997) Modeling the elimination of mercury by sh. Environ Sci Technol 31:1716 1722 Toure F, Stoecker BJ, Lucas E (2001) Fish and shrimp provided bioavailable iodine for rats fed cassava/millet-based diets. FASEB J 15:A635 (Abstract) Uryu Y, Malm O, Thornton I, Payne I, Cleary D (2001) Mercury contamination of sh and its implications for other wildlife of the Tapajos Basin, Brazilian Amazon. Conserv Biol 15:438 446 Vigh P, Mastala Z, Balogh KV (1996) Comparisons of heavy metal concentration of grass carp (Ctenopharyngodon idella Cuv. et Val.) in a shallow eutrophic lake and a sh pond (possible effects of food contamination). Chemosphere 32:691701