Otolith chemical composition as a useful tool for sciaenid stock discrimination in the south-western Atlantic

1 Centro de Estudios Transdisciplinarios del Agua, FVET, Universidad de Buenos Aires, Av. Chorroarín 280, Ciudad de Buenos Aires (C1427CWO). 2 Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET). 3 Departamento de Biodiversidad y Biología Experimental, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales, Ciudad Universitaria Pabellón 2, (1428). Buenos Aires Argentina. E-mail: volpedo@bg.fcen.uba.ar

Striped weakfish and whitemouth croaker are two major components of the fisheries of the Argentinean, Brazilian and Uruguayan coastal region (Haimovici et al., 1989;Norbis, 1995;Lasta and Acha, 1996).In Argentina, striped weakfish annual catches are approximately 16000 tons, and whitemouth croaker annual catches are in the order of 18000 to 20000 tons (Suquele and Colautti, 2003;SAGPyA, 2004).These commercial coastal species are the most important ones in Argentine fisheries (24% of the total landed catch of coastal commercial species) and are an important socioeconomic resource for local communities (SAGPyA, 2004).
High concentrations of fish from July to September and the high vulnerability of these species to low net selectivity, pose several difficulties for fisheries management (Ruarte and Aubone, 2004).Furthermore, these species are common fishing resources between Argentina (Río de la Plata estuary) and Uruguay (Lasta et al., 2000) and hence, fish stock species identification is needed for fisheries management.
The numerous definitions of "stock" as applied to fisheries mostly involve defining population(s) of fish that maintain their genetic integrity over time (Booke, 1981).Fish stocks may be identified and discriminated by different methods (e.g.population parameters, capture-mark-recapture studies, physiological and behavioural characters, morphometric and meristic characters, cytogenetic characters and biochemical characters).As all these methods have limitations in their application, alternative approaches are continuously evaluated (Milton and Chenery, 2001).The success of capture-mark-recapture techniques for stock identification purposes is dependent on representative tagging and recapture efforts.These studies are generally expensive and time consuming (Begg and Waldman, 1999).Both physiological and behavioural characteristics (e.g.fecundity at age, spawning time, sexual maturity, growth rate) have frequently been used to discriminate stocks (Begg and Waldman, 1999), but applying them requires a knowledge of the fish life history.Phenotypic characteristics do not provide direct evidence of genetic isolation between stocks; however, these can indicate the prolonged separation of postlarval fish in different environmental regimes (Begg et al., 1999).Genetic variation between stocks can provide a direct basis for stock structure, but can prove inadequate in situations where even low levels of larval or adult mixing occur (Begg and Waldman, 1999), and high levels of gene flow between stocks are found (Levy et al., 1998).
Previous studies (Figueroa and Díaz de Astarloa, 1991;Díaz de Astarloa and Bolasina, 1992;Machi et al., 1992;Díaz de Astarloa and Ricci, 1998) in Argentina suggested morphologic and morphometric differences in striped weakfish and whitemouth croaker from the Buenos Aires coastal zones.Díaz de Astarloa and Bolasina (1992) applied morphometric and meristic characters of striped weakfish, suggesting the existence of another group in the "El Rincón" area.The same result was obtained for whitemouth croaker applying morphometric and meristic characters (Figueroa and Díaz de Astarloa, 1991;Díaz de Astarloa and Ricci, 1998).
In recent years, otoliths have been used as a tool for stock identification, along with variation in morphometric and morphologic characteristics (Campana et al., 1995;Griffiths, 1996;Gillanders, 2001).In Argentina, otolith morphology and morphometry has been used to discriminate between different stocks of sciaenids from the coast of Buenos Aires province (Volpedo, 2001).
Otolith chemical composition represents a natural tag or fingerprint that may provide information on natal origin and geographic association (Edmonds et al., 1989;1991;Kalish, 1990 a;Secor et al., 1991;Thresher et al., 1994, Thresher, 1999;Proctor et al., 1995;Campana et al., 1994;Campana et al., 1995;Campana, 1999).The technique relies on the assumption that certain elements present in the otoliths are related to their physical and chemical environment, and that resorption or alteration of these elements during ontogeny is minimal (Campana, 1999;Campana et al., 2000).This method is increasingly being used to help researchers understand the environmental history of fish movements (Radtke and Shafer, 1992), population structure (Thresher et al., 1994), and identification of spawning estuaries (Milton et al., 1997).It has been used to discriminate between fish stocks of southern hemisphere species (Edmons et al., 1991;Gunn et al., 1992;Gillanders and Kingsford, 2003;Patterson et al., 2004).
In the present study, otolith chemical composition of striped weakfish and whitemouth croaker was analyzed for catches from three coastal sites to test its use as a stock discriminator.It is the first time that this methodology has been applied in South America and may be an important tool for the sustainable exploitation and management of commercially important fisheries.

Study area
The marine coast of Buenos Aires extends from San Antonio Cape (36°21'S), Río de la Plata external limit, to the mouth of the Río Negro (41°12'S).The coastal shelf slopes gently.In summer, coastal waters have a temperature between 15°C and 25°C, while in winter the temperature range is between 10°C and 12°C (Hoffman et al., 1997) and is influenced by three water masses: subantarctic, subtropical and estuarine masses (Piola and Rivas, 1997).The presence of two coastal ocean fronts, the Subtropical/Subantartic front and the "Frente El Rincón" (Acha et al., 2004), results from the confluence of continental and marine waters, and the tidal effect (Guerrero and Piola, 1997).The oceanic fronts have important biological characteristics, where high biodiversity and biomass production are observed (Lasta and Acha, 1996;Olson, 2002;Acha et al., 2004).
The coastal sites within the study area (Partido de La Costa, Mar del Plata and San Blás Bay) have different geomorphological and environmental characteristics (Fig 1).Partido de La Costa (36°12´S to 36°43´S, 57°45´W to 57°40´W) is constituted by a group of 11 small towns, whose main activity is tourism in summer (four months) due to the presence of extensive, gently sloping coasts.Surface area: ca.3000 km 2 , spring-summer temperature: 20°C-22°C and spring-summer salinity: 20-30 (CEADO, 2004).

Sample Collection
Striped weakfish and whitemouth croaker were collected from sampling stations along the coast of Buenos Aires province, during spring-summer 2002-2003.Striped weakfish (n=20) were captured in Partido de La Costa and San Blás Bay, and whitemouth croaker (n= 48) were collected in Partido de La Costa, Mar del Plata and San Blás Bay (Fig. 1).Sample sites were selected taking into account their importance for commercial fisheries.These samples are representative of fisheries, based on minimum total lengths of 320 mm and 340 mm for striped weakfish and whitemouth croaker respectively.Fish total length (TL) was measured (mm).Otoliths were removed, cleaned of adhering tissue, weighed (OW) (in g) on an analytical balance (to 0.1 mg), digested in 10 % nitric acid and rinsed in Milli Q water.
Calcium was determined by titration using EDTA (APHA, 1993).Trace elements were analyzed by inductively coupled plasma-optical emission spectrometry (ICP-OES, Perkin Elmer Optima 2000).Typical operational conditions were as follows: power (1400W), nebulizer Ar flow rate (0.5 mL min -1 ), auxiliary Ar flow rate (0.7 mL min -1 ), plasma Ar flow rate (15 mL min -1 ) and sample nebulizer flow (1.2 mL min -1 ).The method Multicomponent Spectral Fitting (MSF) was applied to improve the sensibility and the detection limit (Nölte, 1999).Internal standards of Ytrio 371.029 (1 mg L -1 ) were used to correct matrix interference in the ICP-OES.The wavelengths used for the determinations were: Cd (226.502 and 228.801),Cu (324.752),Mg (285.210 and 279.075),Mn (257.610),Sr (232.235) and Zn (213.857).The standard reference materials (SRM) MERCK were used for calibration.Verification and analytic methods validation were performed with the SMR Fish Otolith (No. 22) from NIES (National Institute for Environmental Studies).Detection limits, which were calculated from the concentration of analyte yielding a signal equivalent to three times the standard deviation of the blank signal, were 0.3 µg/g (Cd); 0.36 µg/g (Cu); 0.24 µg/g (Mg); 0.02 µg/g (Mn); 0.12 µg/g (Sr) and 0.24 µg/g (Zn) respectively.Blank samples were prepared in the same manner, but no otolith was present.Determinations were performed in triplicate with a relative error of <1.0% for all of them.The Element:Ca ratios (micromoles/mol) were calculated.

Statistical analyses
The mean values of the Element:Ca ratios for each site were compared using analysis of variance (ANOVA) (Sokal and Rholf, 1995) and multiple comparisons test Tuckey's Honestly significant differences (HSD) among sites (Zar, 1999); P< 0.05 was considered statistically significant.Otolith weight (OW) was included as a covariate.Normality and homogeneity of variance assumptions of ANOVA were tested by the Kolmogorov-Smirnov test and the Levene statistic respectively (P< 0.05).The Cd/Ca and Cu/Ca were logarithmic transformed to normalize the variances (Sokal and Rholf, 1995;Zar, 1999).The Element:Ca ratio of each sciaenid species in Partido de La Costa and San Blás Bay were compared using ANOVA.The software used in statistical analysis was Statistica 5.1 (Statsoft®, 1999).
Multi-elemental compositions were analyzed by non-parametric analysis of similarity (ANOSIM) permutations test (Clarke, 1993).Triangular dissimilarity matrices were calculated using Euclidean dissimilarity measures on double square root transformed data.The ANOSIM permutations test was used to assess the significance of differences among estuaries.A global R-value (=test statistic) was obtained based on average dissimilarities within replicate samples and average dissimilarities between different samples.The global R-value was tested by evaluating random rearrangements (permutations) of the data (Clarke, 1993); in this case, the procedure was repeat-ed 5000 times.Multiple comparisons of individual pairs were also carried out in the same way.All analyses were performed using PRIMER programs (Clarke and Warwich, 1994).
Non-metric multidimensional scaling (nMDS) plots were used as the ordination method.The dissimilarity matrix was calculated using Euclidean distances.The goodness of fit of the data points in the nMDS was measured by the stress coefficient, where stress tends to zero when data are perfectly represented (Clarke, 1993).The final nMDS explains the absence of axis scales and labels in the figures (Clarke, 1993).Data was pooled by species for comparing coastal sites.

RESULTS
Fish total length and otolith weight of striped weakfish and whitemouth croaker are shown in Table 1.Otolith chemical analysis showed that four elements (Mn, Mg, Sr and Zn) were detectable in otoliths from striped weakfish and six elements (Cd, Cu, Mn, Mg, Sr and Zn) were detectable in otoliths  from whitemouth croaker.These elements were chosen for the analyses.Element:Ca ratios for striped weakfish from Partido de La Costa and San Blás Bay and Element:Ca ratio for whitemouth croaker from Partido de La Costa, Mar del Plata and San Blás Bay are presented in Figure 2.
The ANOVA results for striped weakfish showed significant differences in Mg/Ca, Mn/Ca and Sr/Ca ratios between coastal sites, while no significant differences were observed for Zn/Ca (Table 2).
Significant differences among the sampling sites were observed for Cd/Ca, Cu/Ca, Mg/Ca, Sr/Ca and Zn/Ca ratios of whitemouth croaker otoliths, while the Mn/Ca ratio showed no significant differences (Table 2).
The multiple comparison test Tuckey´s Honestly significant differences (HSD) applied to ln Cd/Ca, ln Cu/Ca, Mg/Ca, Sr/Ca and Zn/Ca showed significant differences (P<0.05) in the whitemouth croaker otolith chemical composition between San Blás Bay and the other coastal locations.However, no significant differences were observed in whitemouth croaker otolith chemical composition between Partido de La Costa and Mar del Plata (ln Cd/Ca: P= 0.382; ln Cu/Ca: P= 0.495; Mg/Ca: P= 0.45; Sr/Ca: P= 0.867 and Zn/Ca: P= 0.643).
The multivariate composition of trace elements was analyzed.Significant differences were found between all coastal sites for the chemical composition of striped weakfish otoliths (ANOSIM, global-R=0.636,P<0.001).However, pairwise comparisons of sample sites for ANOSIM suggest that Partido de La Costa samples differ from San Blás Bay.The nMDS plot showed that in striped weakfish, the otolith samples could be grouped in Partido de La   Costa and in San Blás Bay, based on chemical composition (Fig. 3A).Significant differences were found between all coastal sites for chemical otolith composition of whitemouth croaker (ANOSIM, global-R=0.301,P<0.001).However, pairwise comparisons of sample sites for ANOSIM suggest that Partido de La Costa did not differ from Mar del Plata, whereas both of them differed from San Blás Bay.The nMDS plot showed that in whitemouth croaker the samples could be grouped into two sets based on otolith chemical composition, one related to Partido de La Costa and Mar del Plata, and the other one to San Blás Bay (Fig. 3B).
These results suggest the existence of two different fish stocks for striped weakfish and whitemouth croaker.
The otolith chemical composition of the sciaenid species was different in each coastal site.In Partido de La Costa and San Blás Bay, the sciaenid species presented significant differences in the ratios of Mg/Ca and Mn/Ca, whereas the Sr/Ca and Zn/Ca ratios showed no differences (Table 3, Figs.2C, F).

DISCUSSION
The otolith chemical composition of striped weakfish and whitemouth croaker showed significant differences between coastal sites.In other studies, differences in otolith composition among different geographic areas were also found (Campana et al., 1995;Gillanders andKingsford, 2000, 2003).These differences are frequently used to discriminate between fish stocks.
Significant differences in Mg/Ca, Mn/Ca and Sr/Ca ratios in the otoliths of striped weakfish in the coastal province of Buenos Aires suggest the existence of two stocks, one originating in the North (Partido de La Costa) and the other in the South (San Blás Bay).
Significant differences in the ratios of Cd/Ca, Cu/Ca, Mg/Ca, Sr/Ca and Zn/Ca in the otoliths from whitemouth croaker from the 3 sampling stations suggest that the fish may have occupied different environments during their lives.Differences in element concentration among fish from different stations are not due to size differences, because in the ANOVA test otolith weight was the covariate.These facts let us to discriminate between different fish stocks: one originating in the north (Partido de La Costa and Mar del Plata area) and another originating in the south (San Blás Bay).The northern stock is probably constituted not only from fish from Partido de La Costa and Mar del Plata, but also those from Samborombón Bay, based on their geographical proximity and previous reports (Figueroa and Díaz de Astarloa, 1991;Díaz de Astarloa and Bolasina, 1992;Díaz de Astarloa and Ricci, 1998;Volpedo, 2001).On the other hand, the southern stock is probably constituted from fish species from San Blás Bay and those from the "El Rincón" area.
The existence of two separate fish stocks of striped weakfish and whitemouth croaker may be explained by the occurrence of an oceanographic event.Along the South American Atlantic coast other studies of the discrimination of sciaenids stocks have been carried out (Paiva Filho and Cergole, 1988;Isaac, 1988;Vazzoler and Phan, 1989;Vazzoler, 1991).Some of these studies suggest that the differentiation of stocks of these species, may be due to upwelling in Cabo Frio, Brazil (23°S), which could function as an oceanographic barrier, facilitating diversification processes (Paiva Filho and Cergole, 1988;Isaac, 1988).It is possible that in Argentina, the two different fish stocks of striped weakfish and whitemouth croaker may be separated by another oceanographic barrier, the "Frente El Rincón" (39°-39°30´ S) (Fig. 1).
The "Frente El Rincón" was characterized by Piola and Rivas (1997), Guerrero and Piola, (1997) and Acha et al. (2004), as an area whose water mass characteristics (temperature, salinity, density) differ to those of the rest of Buenos Aires province coastal waters.The coastal regime in the "El Rincón" front is characterized by vertical homogeneity due to tidal forcing, and this coastal front is separated from the coast by diluted coastal waters from the Negro and Colorado rivers (960 m 3 s -1 total average discharge), and from shelf waters (Acha et al., 2004).The salinity gradient increases due to the presence of high saline waters originating in San Matías Gulf.The front encloses an area of 10000 km 2 , and has weak seasonality (Acha et al., 2004).The chemical composition of the otolith is the result of the accretion of trace elements during the fish's life cycle.Otolith chemical composition variability depends on different factors associated with physicochemical water characteristics and with the fish itself (ecology, physiology, diet, microhabitat uses and others).Thus otolith chemical composition may vary throughout the different stages in the fish life history, or when environmental changes occur (Gillanders and Kingsford, 2000).Changes in water chemistry may be due to continuous oceanographical events, environmental pressure associated with the El Niño Southern Oscillation (ENSO) or freshwater inputs in coastal zones.
In summary, significant differences in Cd, Cu, Mg and Mn otolith concentrations were observed between striped weakfish and whitemouth croaker.These interspecific variations in otolith chemical composition may be explained by trophic level, use of microhabitats, the species' behaviour, or the depth and the residence time in the coastal area of the two sciaenids species (Kalish, 1990b;Sadovy and Severin, 1992;Fowler et al., 1995;Kinsgford and Gillanders, 2000;Sánchez Jerez et al., 2002).
Further studies, including expanded sampling along the Buenos Aires coast in order to determine the variability among sites and sampling over a number of years, to determine temporal variability in otolith composition are in progress.
Otolith chemical composition allowed us to identify two different stocks of striped weakfish and whitemouth croaker in the Buenos Aires coastal zone.Commercial exploitation of striped weakfish and whitemouth croaker, which accounts for up to 24% of the total landed catch of coastal commercial species, should be managed differently in the future, where fishery efforts should be regulated in relation to each fish stock.This technique has not been previously used in Argentina and may constitute a simple, quick and useful tool for studying fish stocks of commercially significant species.

TABLE 1 .
-Sampling data for fish collected during the study.n: sample size, TL: mean total length ± standard deviation, OW: otolith weight.

TABLE 3 .
-Results of ANOVA of Element: Ca ratios for striped weakfish and whitemouth croaker of Partido de La Costa and San Blás Bay.