Interspecific Variation of Metal Concentrations in Three Bivalve Mollusks fromGalicia

Y. Saavedra,1 A. Gonzalez,2 P. Fernandez,1 J. Blanco3

1 Centro de Control do Medio Marino, Conselleria de Pesca e Asuntos Marıtimos, Vilaxoan, Spain2 Departamento de Quımica Analıtica, Nutricion y Bromatologıa, Facultad de Quımica, Santiago de Compostela, Spain3 Centro de Investigacions Marinas, Consellerıa de Pesca e Asuntos Marıtimos, Vilanova de Arousa, Spain

Received: 20 July 2003 /Accepted: 3 March 2004

Abstract. There has been growing concern about the inflow ofmetals to the coastal areas because they can be toxic to aquatic andhuman life. Some studies have demonstrated the existence ofspecies-specific differences in the metal concentrations of mol-lusks. We compared metal concentrations between Mytilus gallo-provincialis, used as a water quality indicator, and two otherbivalve species collected for human consumption (Venerupis pul-lastra and Cerastoderma edule) in different locations on theGalician coast (northwest Spain). M. galloprovincialis was foundto be the best zinc and lead accumulator, whereas silver andarsenic were preferentially accumulated by V. pullastra and chro-mium and nickel by C. edule. Bivalve concentrations of mercury,cadmium, chromium, arsenic, silver, and zinc appeared to belinearly related to environmental concentrations, but this was notthe case with copper, nickel, and lead in some species, whichindicated that there is a nonlinear accumulation of these metals oran influence of the environmental conditions on species accumu-lation. The relationship between metal concentration in musselsand in the two other species varied with the metal and the species.In some cases the correlation was high, making it possible to usemussels as bioindicators for the other species. In other cases thecorrelation was moderate or low, therefore rendering mussels oflittle or no use in predicting the metal concentrations in the twoother species.

Commercial fisheries of bivalves are an important source ofincome to many coastal populations. Nevertheless, achievingcompatibility between harvesting these resources and industrialdevelopment is a difficult task. Some industrial activities canproduce significant heavy-metal contamination of mollusks as wasevidenced by the mercury accidents in Japan, Sweden, Iraq, etc.(Mance 1987; Fergusson 1990). Consequently, the implementa-tion of heavy-metal monitoring programs for shellfish is requiredto ensure the quality of the edible species. The European Unionhas dealt with this requirement through Council Directive 91/492/

EC. Most national monitoring programs do not analyze the metalconcentration in each harvested species and instead use the one ora few species that are assumed to be bioindicator species. Follow-ing this approach, metal concentrations in a bivalve species (usu-ally mussels) are used to assess water quality, and it is assumedthat all bivalves in good-quality water are safe for human con-sumption. This method does not hold when the metal accumula-tion characteristics of the target species are not similar to those ofthe bioindicator species, making the estimation of the risk incor-rect. In Galicia, and in many other European areas, cockles(Cerastoderma edule) and clams (Venerupis pullastra) are speciescommercially important because of their high production andprice. To prevent heavy-metal intoxication in these and otherspecies, mussels (Mytilus galloprovincialis Lmk.) from the areawere analyzed for heavy-metal content and used as bioindicatorsof metal contamination. However, the usefulness of M. gallopro-vincialis as an indicator of metal contamination in C. edule and V.pullastra has not been checked and—taking into account thatseveral studies have demonstrated different bivalve species todiffer substantially in their capability to accumulate various metals(Reinfelder et al. 1997; Wang and Fisher 1999; Chong and Wang2000; Wang 2001)—it should be checked. Therefore, there is aneed to evaluate both the usefulness of mussels as bioindicatorsand the actual concentrations of metals in these two species toguarantee their safety for consumption. In this study, we attemptedto determine the how the metal concentrations of two poorlyknown species (V. pullastra and C. edule) compare with those ofM. galloprovincialis in an area (the Galician coast) with diversesources of metal pollution. We thus obtained an idea of the riskthat these species represent as well as the usefulness of theconcentrations found in mussels as an indicator for the other twospecies.

Materials and Methods

Sampling and Sample Processing

Samples of three species of bivalves—the mussel M. galloprovincialis,the cockle C. edule, and the clam V. pullastra—were obtained fromnine locations in the Galician Rıas chosen to represent the harvestedCorrespondence to: Y. Saavedra; email: ysaavedra@cccmm.cesga.es

Arch. Environ. Contam. Toxicol. 47, 341–351 (2004)DOI: 10.1007/s00244-004-3021-5

A R C H I V E S O F

EnvironmentalContaminationa n d Toxicology

© 2004 Springer Science�Business Media, Inc.

area of the coast and to allow the collection of the three species fromthe same place (Figure 1). To avoid differences in metal contentbecause of size or reproductive stage, commercial-size individuals ofeach species were collected during a period of the annual cycle outsideof the spawning and reserve-building periods. Samples of 30 individ-uals were collected and kept frozen until processed. Analyses of metalconcentrations of the total edible portion of the bivalves were carriedout on pooled samples consisting of 30 specimens of each species. Thethawed bivalves were opened and the soft tissues were extracted andpooled in batches of 30 individuals. The water content of the bivalveswas calculated by drying aliquots at 105°C for 24 hours after discard-ing the shells. The other aliquots of bivalve tissues were lyophilizedand pulverized in a mixer mill with zirconium oxide balls and jar. A0.5-g subsample of the homogenate was digested with nitric acid byheating it in a microwave (CEM MDS 2000).

Analytical Procedures

The analytical determination was performed using several atomicabsorption spectrometry techniques. Copper and zinc contents wereanalyzed using flame atomic absorption spectrometry with a PerkinElmer 4100 Atomic Absorption Spectrometer. Cadmium, lead, chro-mium, nickel, arsenic, and silver were analyzed by electrothermalatomic absorption spectrometry with Zeeman background correctionand stabilized temperature platform furnace) using a Perkin ElmerAtomic Absorption Spectrometer 4100ZL (Autosampler As70 cou-pled). A Perkin Elmer 4100 Atomic Absorption Spectrometer, with anAs 90 autosampler and a FIAS 400 flow injection system, was used todetermine mercury content by cold-vapor atomic absorption spectrom-etry. Two analytic replicates were done for all samples.

Validation of the mussel samples was performed with intralaboratoryquality control using certified reference material CRM 278: trace metals(except nickel and silver) in mussel tissue from the Community Bureau ofReference. The quality of the method was externally assessed with bian-nual participation in the Quality Assurance of Information for MarineEnvironmental Monitoring in Europe program with a result of Z �2(satisfactory results of intercalibration, Thompson and Wood 1993).

Because no certified reference material was available for metalconcentrations in clams and cockles, we checked the recovery and theexistence of any obvious matrix effect by spiking known referencesolutions to soft tissues of the two species. Recovery percents between90% and 110% were obtained indicating no matrix effect.

Statistical Treatment of Data

The relationships between the concentration of the different metals,their possible grouping, and origin were studied by means of principalcomponent analysis (PCA). The differences of concentration betweenthe different metals, locations, and species—as well as their interac-tions—were checked using three-way and two-way analysis of vari-ance (ANOVA) and the differences of each metal between species bymeans of paired Student t test. All of these statistical procedures werecarried out using the statistical package MINITAB 13.3.

The relationship between the environmental concentrations of met-als and those in bivalves was inferred by the analysis of the residualsof the ANOVA and by the “structural equation with latent variables”approach (Bollen 1989) using the simultaneous nonlinear least-squaresfit of the metal concentration data for the three species to an imaginaryset of environmental concentrations. The method roughly consisted ofassuming that the metal concentrations found in the bivalves werederived from a latent variable, the environmental concentration, whichcould be estimated if the structure (functional form of the relationship)were known.

We assumed that the concentration of any metal in the bivalves wasrelated to the environmental concentrations by one of four equations—(1)MO � a � b � MA (straight line); (2) MO � a � eb � MA (exponential); (3)MO � a � (1 � eb � MA) (asymptotic exponential); or (4) MO � a �b(MA � c)2 (polynomial of degree 2) where MO and MA were metalconcentrations in the organisms and in the environment, respectively, and“a,” “b,” and “c” were parameters that had to be estimated. By allowingvariation of the environmental concentrations, the parameters of the ac-cumulation curve, and the equation, we found a “relative environmentalconcentration” (the latent variable). The equations that produced the bestfit were considered to give an approximate representation of the actualrelationship between the concentrations in the environment and those inorganisms. Minimization was carried out using Matlab and the MatlabOptimization Toolbox.

Results

Sources of Variability in Metal Concentration

Species-specific metal accumulation was the main factor regulat-ing the metal concentration in the bivalves studied. PCA extractedthree components, with the first two explaining more that one halfthe total variance (53.2%). The first one was defined mainly by thecovariation of five of the nine metals studied: chromium, nickel,copper, arsenic, and silver. Even when these five metals covari-ated, chromium and nickel varied inversely to copper, silver, andarsenic as can be observed from the different signs of theirloadings on the first principal component (PC1; Figure 2). Thismeans that in the samples and species, the concentrations ofchromium and nickel were higher when those of arsenic, silver,and copper were lower and vice versa. It can be observed (Figure

Fig. 1. Sampling locations in Galician coast (northwest Spain)

342 Y. Saavedra et al.

2) that chromium and nickel were associated with C. edule and thethree others with M. galloprovincialis and V. pullastra. No asso-ciation of these two groups of metals with the sampling locationcould be found.

The second principal component (PC2) was mostly definedby zinc, cadmium, and lead in one direction and by mercuryand silver in the other, indicating again that the two groupsassociated with the component are inversely related, i.e., whenone group increases in concentration, the other decreases.Again, the basis of this component and therefore of this metalgrouping seemed to be species because each group was asso-ciated with one species: zinc, cadmium, and lead with M.galloprovincialis and mercury and silver with V. pullastra.

A third component was also extracted, obviously with lessimportance that the first two, and was linked mainly to mercury(Figure 3). This was mostly defined by a sampling location thatseemed to have an exceptionally high incidence of mercury,probably because of a nearby paper mill, but also by the specialcharacteristics of another location (Baldaio, a coastal lagoon)that determines different concentrations of cadmium.

Metal Concentrations in Species and Locations

The metal concentrations were different for the different met-als, species, and locations (significant main effects in the three-way ANOVA; Table 1). The way in which the metal, location,and species affect the concentration was, in each case depen-dent on the other two factors (significant interactions in thethree-way ANOVA; Table 1).

For most metals, there were significant differences betweenspecies, but for only some of them were there differencesbetween locations (two-way ANOVA; Table 2, and Figures 4and 5).

The highest concentrations of most metals were recorded inRedondela, in the inner part of the Rıa de Vigo, but some othershad their maximum concentration in different locations: cad-mium in Corcubion, mercury in Campelo, and nickel andchromium in Carino (Figure 4).

Significant differences of metal concentrations between spe-cies (two-way ANOVA; Table 2) were found for cadmium,chromium, arsenic, silver, nickel, and zinc. The highest inter-specific differences were detected for chromium and nickel incockles; zinc in mussels; and arsenic and silver in clams. Incontrast, the lowest values were found for cadmium in cockles(Figure 5). Normally distributed residuals from the two-wayANOVAs were obtained for all but three of the metals studied,thus indicating a linear relationship between ambient metalconcentrations and organism metal concentrations. The rawdata for the other three metals (copper, nickel, and lead) werenot suitable for ANOVA because the statistical residuals ob-tained did not have a normal distribution, thus pointing to theexistence of a nonlinear effect of the pair of factors, location–species. The logarithmically transformed data were used tomake the corresponding comparisons. The best fit of the con-centration in the three species to the “relative” environmentalconcentrations (see Material and Methods section) is shown inFigure 6.

Relationship Between Metal Concentrations in Musselsand Other Species

The magnitude and even the sign of the regression coeffi-cient between metal concentrations in mussels and in theother two species varied substantially with both the metal

Fig. 2. Loading of species, metal, and location on the first twoprincipal components (PC1 and PC2) extracted from the correlationmatrix of the data

Fig. 3. Loading of species, metal, and location on the second and thirdprincipal components (PC1 and PC3) extracted from the correlationmatrix of the data

Metal Concentrations in Three Bivalve Mollusks 343

and the species. In this study there was good correlationbetween the mercury and nickel concentrations in musselsand those in the two other species. Nevertheless, the corre-lation was poorer for chromium concentrations in clams andcockles and for lead and copper in clams, and almost non-existent for the remaining metal–species pairs (Table 3 andFigure 7). In some cases in which there was a bad correla-tion between M. galloprovincialis and the other species,there were no significant differences between metal concen-trations in mussel and the other species, or the metal con-centrations in mussels were significantly higher than thosein cockles and clams (Table 4).

Discussion

Sources of Variability in Metal Concentration

Variability of the metals studied in M. galloprovincialis, C. edule,and Venerupis pullastra was mainly associated with the species-specific accumulation characteristics of each metal. Location wasnotably less important in conditioning this distribution, indicatingthat under moderate pollution, the metal concentrations in thethree species were defined by the capability of each species toaccumulate a particular metal. Mussels appeared to accumulatemore zinc and lead than the other two species, and the same wastrue for arsenic and silver in clams and for chromium and nickelin cockles. In contrast, clams and cockles seemed to accumulateless lead, and cadmium and arsenic, respectively (see InterspecificVariation in Metal Concentration subsection below).

The importance of location in determining the overall dis-tribution of metals in the area’s bivalves was only marginal

compared with that of species. The only relevant locations (inrelation to interspecific variation) were Campelo and, to alesser extent, Baldaio, in part because of anthropogenic actions.

Geographic Distribution of Metal Concentration

As was to be expected from the different sediments and thedifferent urban or industrial activities carried out at the locationssampled, location was found to have an effect on the concentrationof each metal. Some aspects of this influence are particularlyrelevant. The mercury concentrations found were especially highin Campelo (Ria de Pontevedra), probably because of the contam-ination originating from a paper pulp manufacturing plant thatuses mercury cells to obtain chlorine. In other studies (unpub-lished data), we found mercury levels in M. galloprovincialis thatwere 10% higher than at other sites on the Galician coast. Thehighest concentrations of chromium and nickel were found inorganisms from the northern Rıas. Higher values of chromium andnickel have also been found in northern compared with southernsediments, a phenomenon that has been related to substrate lithol-ogy (a predominance of schist and gneiss, rocks that are richer inthese metals, in the north) by Carral et al. (1995). These investi-gators inferred that the body-metal contents of sediment-associ-ated organisms are to some extent influenced by watershed lithol-ogy, although a close relationship has not been established.Moreover, the high occurrence of chromium at the Carino locationcan be associated with the closed mining of dunite rocks, whichcontain chromium as a minor component. Chromium concentra-tions are also important at the Carril location, and a high presencein the sediments has been detected (Belzunce Segarra et al. 1997).In this case, the most likely explanation is anthropogenic activitybecause a few miles upstream, tannery wastes historically have

Table 1. Three-way ANOVA of metal concentrations considering metal, species, and location factors

Factors df Seq SS Adj SS Adj MS F pa

Site 8 0.7188 0.7925 0.0991 6.29 0.000Species 2 0.0982 0.1665 0.0833 5.29 0.006Metal 8 172.2821 172.9761 21.6220 1373.72 0.000Site-metal 64 9.2507 9.3884 0.1467 9.32 0.000Site-species 16 0.8869 1.0133 0.0633 4.02 0.000Species-metal 16 12.4490 12.4490 0.7781 49.43 0.000Error 127 1.9989 1.9989 0.0157Total 241 197.6845

a p indicates significance of differences in metal concentrations with each factor and of interactions between factors. p � 0.05 was consideredsignificant difference or interaction.Adj: Adjusted, df: degrees of freedom, Seq: sequential.

Table 2. Two-way ANOVA of metal concentrations with site and species

p(Hg) p(Cd) p(Pb) p(Cr) p(Ni) p(As) p(Ag) p(Cu) p(Zn)a

Site 0.000 0.006 0.182 0.004 0.115 0.024 0.012 0.197 0.516Species 0.205 0.001 0.076 0.033 0.001 0.000 0.001 0.037 0.000

a p indicates the significance of differences in metal concentrations with site or species. p � 0.05 was considered significant.Ag: silver, As: arsenic, Cd: cadmium, Cr: chromium, Cu: copper, Hg: mercury, Ni: nickel, Pb: lead, Zn: Zinc.

344 Y. Saavedra et al.

been discharged into the water. Nevertheless, the lithology in thiszone has similar characteristics to those found in the northernRıas, and this natural origin cannot be ignored.

Redondela appears to be a major point of contamination forsome metals, most likely because it is situated in the section of theRıa having the highest industrial and urban activity in Galicia. Thelarge copper concentrations observed are probably caused by theproximity to shipyards and harbors where copper-based, antifoul-ing paints are used. According to the lead concentrations in M.galloprovincialis quoted by Besada et al. (1997), an extraordinar-ily high lead concentration in bivalves, the highest on the Galiciancoast, was found in that study. This is also in keeping with the highconcentrations found in sediment by Vilas et al. (1995). Theintense maritime and terrestrial traffic in the zone are considered tobe the main uptake routes of this metal.

Interspecific Variation in Metal Concentration:Differences in Concentration

Interspecific variation in the concentration of each metal ingeneral had more importance than geographic distribution.

Despite being exposed to the same environment, the threespecies studied did not accumulate all metals to the sameextent. Mercury was an exception because differences in theconcentration of this metal among the three species were small(compared with that found between locations).

From the data obtained, mussels M. galloprovincialis werestrong accumulators of lead and zinc and poor accumulators ofarsenic and silver compared with the other bivalves studied.This species and a closely related one, M. edulis, have beenshown to accumulate lead because they were found to ex-crete limited amounts of this metal when placed in a cleanenvironment (Regoli and Orlando 1994) and to accumulatelarge amounts when placed in polluted areas (Riget et al.1997) or when experimentally exposed to high concentra-tions of this metal (Schulz-Baldes 1974). M. galloprovin-cialis was the best zinc accumulator of the three speciesstudied. This capability to accumulate zinc has not beendemonstrated consistently in previous studies. Fowler andOregioni (1976) found high zinc concentrations in M. gal-loprovincialis in polluted areas of the northwestern Medi-terranean Sea, but Regoli and Orlando (1994) reported con-trary results indicating that zinc was not accumulated by M.

Fig. 4. Two-way ANOVA. Main ef-fects plot of location on metal concen-trations

Metal Concentrations in Three Bivalve Mollusks 345

galloprovincialis when this species was transplanted from aclean to a polluted environment. This inconsistency in theaccumulation of this metal was also observed in M. edulis(Philips 1976a; Riget et al. 1997).

The poor accumulation of arsenic was consistent with theresults obtained by Unlu and Fowler (1979) and in M. edulis byLangston (1984), who found that these mussels accumulatearsenic poorly, probably because of the active secretion of thismetal in the byssal threads. The small accumulation of silver inM. galloprovincialis has been documented only by Berthet etal. (1992).

The concentrations of most metals, with the exception ofnickel and chromium in cockles, were low compared withthose in clams and mussels. This finding supported the obser-vations made by Bryan et al. (1985) who reported that cocklesC. edule were not particularly good metal accumulators withthe exception of nickel.

The clam V. pullastra presented the highest silver and ar-senic concentrations in this study. No other studies of thesemetals are known for this species. Two species of deposit-feeding clam (Scrobicularia plana and Macoma balthica) werestudied by Bryan et al. (1980) who found that the two clam

species had higher concentrations of silver than other bivalves,including M. edulis and C. edule.

Inference of the Relationship Between Concentration inthe Bivalves and in the Environment

The comparative study of the metal concentrations in the threespecies allowed the inference of their accumulation characteris-tics. The actual concentrations measured were the result of apply-ing the accumulation properties of each species to the range ofenvironmental concentrations. If there was a linear relationshipbetween the environmental concentration and that in the bivalves,then the residuals of the two-way ANOVA of species-locationswere normally distributed. If the relationship in any of the specieswas not linear, then the distribution of the obtained residuals wasnot normal. In our study, all of the species appeared to accumulatemercury, silver, cadmium, zinc, arsenic, and chromium in thesame linear way because the distribution of the residuals wasnormal. The residuals of the remaining metals were not normallydistributed residuals, but normalization was achieved by logarith-mically transforming the concentration.

Fig. 5. Two-way ANOVA. Maineffects plot of species on metal con-centrations

346 Y. Saavedra et al.

The causes of the nonnormality of these metals very likelyderived from a nonlinear relationship between the accumu-lated metal and the environmental concentration. Even when

this later concentration was not measured, it could have beenconsidered a “latent variable” that is expressed in a differentway by each species, and it could have been estimated from

Fig. 6. Relationships between metal concentrations inorganisms and those estimated in the environment

Metal Concentrations in Three Bivalve Mollusks 347

those responses. Obviously, the “latent variable” was notstrictly the real environmental concentration but rather animaginary variable closely related to it (this constituted thebases of the statistical techniques named “structural equa-tions with latent variables”; see Bollen 1989).

The accumulation of lead in mussel, and to a lesser extentin clam, seemed to be an exponential function of the envi-ronmental concentration, where in cockle it was parabolic.Even being exponential, the responses in mussel and clamwere not very different from a straight line, which was foundin previous experiments in both M. galloprovincialis (Mar-tincic et al. 1992) and in the closely related species M. edulis(Schulz-Baldes 1974). No information exists about the ac-cumulation of lead by V. pullastra, but linear relationshipswith environmental concentrations have been found in clamspecies Circenita callipyga (Zorba et al. 1992) and Scro-bicularia plana (Bryan et al. 1980) species. In contrast,cockle showed a strongly nonlinear response. There was noclear explanation for the parabolic response found in thisspecies because the accumulation levels observed were fartoo low to have an adverse effect on accumulation. Problemswith cockles as lead bioindicators have also been reportedby Bryan et al. (1985); they also reported associated chro-mium problems not noted here.

The bivalve concentration of nickel seemed to be exponentiallyrelated to that in the environment. This response is not supportedby the few available references. Wilson (1983) found that nickelconcentrations in C. edule tissues correlated with water concen-trations at any given time, but the concentration range used by thisinvestigator was much wider than that found in our samples.Similar results have also been found for M. edulis (Friedrich andFilice 1976), but there are no available data about M. gallopro-vincialis, V. pullastra, or other clams.

The estimated response of the three species for copper was

similar to that found for lead, with the difference that in M.gallovincialis and V. pullastra the relationship was linear in-stead of slightly exponential. The response in M. galloprovin-cialis did not agree with other studies that found a poorcorrelation of concentration in mussels with copper environ-ment in this species (Martincic et al. 1992) and also in M.edulis (Phillips 1976a).

There is no information about V. pullastra, but in other clamsa linear relationship was found (Corbicula fluminea by Co-lombo et al. 1997 and Potamocorbula amurensis and Macomabalthica by Brown and Luoma 1995).

The poor correlation of copper concentrations in C. edule withthe ambient concentrations found here has also been cited by otherinvestigators who associated it with high copper regulation in thisspecies (Bryan et al. 1985; Cheggour et al. 2001).

Interaction Between Species and Geographic Location

The significance of first-order interactions in the three-wayANOVA show that both the effect of species and the effect ofeach metal are dependent on location. This is most likely relatedto the influence of environmental factors on the accumulationprocess. Some studies have shown that the relative proportions ofaccumulated loads of metals were dependent on the metal con-centration available in the environment but probably on the localenvironment conditions. O’Leary and Breen (1997) showed thatspecies-specific metal bioaccumulation occurs and is site-specificin some species. In our study, neither the season nor the positionof the animal in the water column varied with the sites, andconsequently other factors such as salinity, temperature, chemicalstate of the metals, as well as the simultaneous occurrence ofseveral of them, would affect the uptake and accumulation of the

Table 3. Equations of the regression lines describing the relationship between metal concentrations in mussels (m) and the other species,clams (cl) and cockles (co), and percentages of explained variability by regressions

Equations Regression R-Sq % R-Sq (adj) % pa

Cd:cl–Cd:m Y � 0.0280361 � 0.555422X 38.30 29.50 0.075Cd:co–Cd:m y � 0.0264460 � 0.282230X 32.80 24.40 0.083Pb:cl–Pb:m Y � 0.0462467 � 0.389224X 63.80 58.60 0.010Pb:co–Pb:m Y � 0.253303 � 0.106963X 2.70 0.00 0.651Hg:co–Hg:m Y � 0.0047402 � 0.664132X 90.60 89.50 0.000Hg:cl–Hg:m Y � 0.0061966 � 0.756410X 95.20 94.50 0.000Cr:cl–Cr:m Y � 0.0739210 � 0.594748X 69.00 64.60 0.006Cr:co–Cr:m Y � 0.122618 � 2.19330X 68.30 64.40 0.003Ni:cl–Ni:m Y � 0.181410 � 1.60886X 95.20 94.50 0.000Ni:co–Ni:m Y � 0.796569 � 9.79922X 84.40 82.50 0.000As:cl–As:m Y � 1.31625 � 0.717395X 17.40 5.50 0.265As:co–As:m Y � 1.16172 � 0.261573X 3.20 0.00 0.623Ag:cl–Ag:m Y � 0.128352 � 0.302964X 3.00 0.00 0.658Ag:co–Ag:m Y � 0.0406257 � 0.461182X 13.30 2.40 0.300Cu:cl–Cu:m Y � 1.28725 � 2.30768X 67.60 62.90 0.007Cu:co–Cu:m Y � 0.923588 � 0.0945145X 0.50 0.00 0.840Zn:cl–Zn:m Y � 12.2648 � 0.0305654X 1.80 0.00 0.734Zn:co–Zn:m Y � 9.28814 � 0.0338983X 2.70 0.00 0.648

a p indicates the significance of the regression.Ag: silver, As: arsenic, Cd: cadmium, Cr: chromium, Cu: copper, Hg: mercury, Ni: nickel, Pb: lead, Zn: zinc.

348 Y. Saavedra et al.

Table 4. Capability of the mussel to serve as indicator of metal concentration in clam (V. pullastra) and cockle (C. edule)

Metal

Clam Cockle

Concentrationa Regressionb Concentration Regression

Hg � Good � GoodCd � Bad Lower BadPb Lower Good � BadNi Higher Good Higher GoodCr � Moderate � ModerateAs Higher Bad � BadAg Higher Bad � BadCu � Moderate Lower BadZn Lower Bad Lower Bad

a Indicates if the species has significant higher or lower concentration than mussel or if there are not significant differences (�).b Indicates if the regression of the concentration in mussels with that in the species is good (R2 � 80%), moderate (R2 between 50% and 80%),or bad (R2 � 50%).Ag: silver, As: arsenic, Cd: cadmium, Cr: chromium, Cu: copper, Hg: mercury, Ni: nickel, Pb: lead, Zn: zinc.

Fig. 7. Relationships describing the prediction of metal concentrations in V. pullastra and C. edule from metal concentrations in M. gallopro-vincialis

Metal Concentrations in Three Bivalve Mollusks 349

different species selectively (Phillips 1976b, 1977, 1978; Martin-cic et al. 1992).

Mussel as an Indicator of Clam and Cockle MetalConcentrations

Mussels are useful indicators of the levels of some metals inclams and cockles but not of others (Table 4). There are twomain ways in which a species can be used as an indicator ofmetal concentrations in others: (1) by a good linear relation-ship or (2) by having systematically lower concentrationthan the other species. Mercury and nickel concentrations inmussels are good predictors of those in the two other speciesbecause the regression between the concentrations in musseland in the other species is good. The same happened withlead in clam and, to a lesser extent because of moderatefitting, with chromium in the two species and copper inclam.

Mussel had significantly higher concentrations of zinc thanthe two other species, of copper and cadmium than cockle, andof lead than clam. It can therefore be used as an indicator of theconcentrations of those metals in the cited species, at least atthe levels found in this study.

Mussel, therefore, cannot be used as an indicator of silver andarsenic in clam and cockle, of cadmium in clam, or of lead incockle. In these cases, each species should be independentlymonitored.

References

Belzunce Segarra MJ, Helios-Rybicka E, Prego R (1997) Distributionof heavy metals in the river Ulla and its estuary (North-West,Spain). Oceanol Stud 2–3:139–152

Berthet B, Amiard JC, Amiard-Triquet C, Martoja M, Jeantet AY(1992) Bioaccumulation, toxicity and physico-chemical speciationof silver in bivalve molluscs: Ecotoxicological and health conse-quences. Sci Total Environ 125:97–122

Besada V, Fumega J, Cambeiro, B (1997) Variacion anual de lasconcentraciones de Hg, Pb, Cd, Cu y Zn en mejillon silvestre dela Ria de Vigo. Procesos Biogeoquımicos en sistemas costerosHispano-Lusos. VII Seminario Iberico de Quımica Marina 95–98

Bollen KA (1989) Structural equations with latent variables. Wiley,New York, NY, p 514

Brown CL, Luoma SN (1995) Use of the euryhaline bivalve Pota-mocorbula amurensis as a biosentinel species to assess trace metalcontamination in San Francisco Bay. Mar Ecol Prog Ser 124:129–142

Bryan GW, Langston WJ, Hummerstone LG (1980) The use of bio-logical indicators of heavy metal contamination in estuaries. MarBiol Assoc UK 1:1–73

Bryan GW, Langston WJ, Hummerstone LG, Burt GR (1985) A guideto the assessment of heavy-metal contamination in estuaries usingbiological indicators. Mar Biol Assoc UK 4:1–92

Carral E, Villares R, Puente X, Carballeira A (1995) Influence ofwatershed lithology on heavy metal levels in estuarine sedimentsand organisms in Galicia (North-West Spain). Mar Pollut Bull30:604–608

Cheggour M, Chafik A, Langston WJ, Burt GR, Benbrahim S, TexierH (2001) Metals in sediments and the edible cockle Cerastoderma

edule from two Moroccan Atlantic lagoons: Moulay Bou Selhamand Sidi Mousa. Environ Pollut 115:149–160

Chong K, Wang W-X (2000) Assimilation of cadmium, chromium,and zinc by the green mussel Perna viridis and the clam Rrudi-tapes philippinarum. Environ Toxicol Chem 19,6:1660–1667

Colombo JC, Brochu C, Bilos C, Landoni P, Moore S (1997) Long-term accumulation of individual PCBs, dioxins, furans, and tracemetals in Asiatic clams from the Rio de la Plata Estuary, Argen-tina. Environ Sci Technol 31:3551–3557

Fergusson JE (1990) The heavy elements: chemistry, environmentalimpact and health effects. Part IV. The toxicity of heavy elementsto human beings. Pergamon, New York, pp 533–567

Fowler SW, Oregioni B (1976) Trace metals in mussels from the N.W.Mediterranean. Mar Pollut Bull 7:26–29

Friedrich AR, Filice FP (1976) Uptake accumulation of the nickel ionby Mytilus edulis. Bull Environ Contam Toxicol 16:750–755

Langston WJ (1984) Availability of arsenic to estuarine and marineorganisms: A field and laboratory evaluation. Mar Biol 80:143–154

Mance G (1987) Pollution threat of heavy metals in aquatic environ-ments. Elsevier, New York, NY, pp 1–8

Martincic D, Kwokal Z, Peharec Z, Margus D, Branica M (1992)Distribution of Zn, Pb, Cd and Cu between seawater and trans-planted mussels (Mytilus galloprovincialis). Sci Total Environ119:211–230

O’Leary C, Breen J (1997) Metal levels in seven species of molluscand in seaweeds from the Shannon Estuary. Biology—an Envi-ronment. Proc R Irish Acad 97:121–132

Phillips DJH (1976a) The common mussels Mytilus edulis as anindicator of pollution by zinc, cadmium, lead and copper. II.Relationship of metals in the mussel to those discharged byindustry. Mar Biol 38:71–80

Phillips DJH (1976b) The common mussel Mytilus edulis as an indi-cator of pollution by zinc, cadmium, lead and copper. I. Effects ofenvironmental variables on uptake of metals. Mar Biol 38:59–69

Phillips DJH (1977) The common mussel Mytilus edulis as an indica-tor of trace metals in Scandinavian waters. I. Zinc and cadmium.Mar Biol 43:283–291

Phillips DJH (1978) The common mussel Mytilus edulis as an indica-tor of trace metals in Scandinavian waters. II. Lead, iron andmanganese. Mar Biol 46:47–156

Regoli F, Orlando E (1994) Accumulation and subcellular distributionof metals (Cu, Fe, Mn, Pb and Zn) in the Mediterrarean musselMytilus galloprovincialis during a field transplant experiment.Mar Pollut Bull 28:592–600

Reinfelder JR, Wang W-X, Luoma SN, Fisher NS (1997) Assimilationefficiencies and turnover rates of trace elements in marine bi-valves: A comparison of oysters, clams and mussels. Mar Biol129:443–452

Riget F, Johansen P, Asmund G (1997) Uptake and release of lead andzinc by blue mussels. Experience from transplantation experi-ments in Greenland. Mar Pollut Bull 34:805–815

Schulz-Baldes M (1974) Lead uptake from sea water and food, andlead loss in the common mussel Mytilus edulis. Mar Biol 25:177–193

Thompson M, Wood R (1993) The international harmonized protocolfor the proficiency testing of (chemical) analytical laboratories.Pure Appl Chem 65:2123–2144

Unlu MY, Fowler SW (1979) Factors affecting the flux of arsenicthrough the mussel Mytilus galloprovincialis. Mar Biol 51:209–219

Vilas F, Nombela MA, Garcıa-Gil E, Garcıa-Gil S, Alejo I, Rubio B,et al. (1995) Contenido y distribucion de metales pesados. Car-tografia de Sedimentos Submarinos, Rıa de Vigo, pp 35–38

350 Y. Saavedra et al.

Wang W-X (2001) Comparison of metal uptake rate and absorptionefficiency in marine bivalves. Environ Toxicol Chem 20:1367–1373

Wang W-X, Fisher NS (1999) Effects of calcium and metabolicinhibitors on trace element uptake in two marine bivalves. J ExpMar Biol Ecol 23:149–164

Wilson JG (1983) The uptake and accumulation of Ni by Cerasto-derma edule and its effects on mortality, body condition andrespiration rate. Mar Environ Res 8:129–148

Zorba MA, Jacob PG, Al-Bloushi A, Al-Nafisi R (1992) Clams aspollution bioindicators in Kuwait’s marine environment: Metalaccumulation and depuration. Sci Total Environ 120:185–204

Metal Concentrations in Three Bivalve Mollusks 351