Color profile: Disabled Composite Default screen

361

Stability of elemental signatures in the scales of spawning weakfish, Cynoscion regalis Brian K. Wells, Simon R. Thorrold, and Cynthia M. Jones

Abstract: We quantified elemental signatures in scales of ages-1 and -2 weakfish, Cynoscion regalis, collected during the spawning season in Pamlico Sound, Chesapeake Bay, and Delaware Bay in 1998. We compared these signatures with elemental signatures from scales of juvenile weakfish collected while still resident in natal estuaries at five locations along the Atlantic coast in 1996 and 1997. Although Mg/Ca and Mn/Ca were lower in the juvenile portion of scales from adults compared with scales from juvenile fish, Sr/Ca and Ba/Ca were similar in the three age groups. We compared scale and otolith chemistries from juveniles and adults to determine if relative concentrations of elements/Ca in scales remained consistent, even if absolute levels were altered. Scale Mn/Ca and Ba/Ca remained correlated with those in otoliths of adult fish. Finally, we examined the ability of elemental signatures in scales to act as natural tags of natal estuaries in spawning weakfish. Allocation of fish to natal estuaries based on geochemical signatures in scales and otoliths from age-1 fish was similar; however, allocation was different for age-2 fish. Elemental signatures in scales degraded after the juvenile period and after maturation were insufficiently stable for use as a natural tag of natal origins in weakfish. Résumé : Nous avons quantifié les signatures d’éléments chimiques dans les écailles d’acoupas royaux, Cynoscion regalis, d’âges 1 et 2, récoltés durant la saison de fraye dans le détroit de Pamlico et dans les baies de Chesapeake et de Delaware en 1998. Nous avons comparé ces signatures à celles d’écailles d’acoupas juvéniles récoltés alors qu’ils résidaient toujours dans les estuaires où ils sont nés, à cinq sites le long de la côte atlantique en 1996 et 1997. Bien que les rapports Mg/Ca et Mn/Ca soient plus bas dans la portion juvénile des écailles des adultes par comparaison aux rapports mesurés chez les jeunes poissons, les rapports Sr/Ca et Ba/Ca sont semblables dans les trois classes d’âge. Nous avons comparé les données chimiques des écailles et des otolithes des jeunes et des adultes pour voir si les concentrations relatives des éléments demeurent stables par rapport au Ca, même si les concentrations absolues ont changé. Les rapports Mn/Ca et Ba/Ca des écailles restent reliés à ceux des otolithes chez les adultes. Enfin, nous avons évalué le potentiel des signatures d’éléments chimiques dans les écailles pour servir d’étiquette naturelle pour identifier l’estuaire d’origine des acoupas sur les frayères. Les identifications des estuaires d’origine d’après les signatures géochimiques des écailles sont similaires à celles des otolithes chez les poissons d’âge 1, mais elles sont différentes à l’âge 2. Les signatures d’éléments chimiques dans les écailles se dégradent après la période juvénile et, après la maturation, elles n’ont plus la stabilité nécessaire pour servir d’étiquette naturelle des sites d’origine des acoupas. [Traduit par la Rédaction]

Wells et al.

369

Introduction Geochemical tracers in the calcified structures of fishes and marine invertebrates are becoming more commonly used as natural tags of population affinities and natal origins (e.g., Pender and Griffin 1996; DiBacco and Levin 2000; Thorrold et al. 2001). The use of isotopic and elemental markers as natural tags is based on the assumption that variations in the physical and chemical environment are recorded in skeletal tissue of individuals that are geographically isolated for at least some part of their life histories. Otoliths are generally considered the best option for use of geochemical signatures

as natural tags in fish populations because concentrations of at least some elements in otoliths are highly correlated with environmental levels (Bath et al. 2000; Wells et al. 2003) and because of the unique patterns of daily and annual increments visible in sectioned otoliths (Campana and Thorrold 2001). However, otolith methods also have significant disadvantages because the fish must be sacrificed and mutilated to recover the otoliths. Scales may represent a nonlethal, nondestructive alternative to otoliths for use as a natural tag of natal origins. Concentrations of at least some elements in fish scales are highly correlated with environmental chemistry (Wells et al. 2000a,

Received 9 August 2002. Accepted 8 March 2003. Published on the NRC Research Press Web site at http://cjfas.nrc.ca on 13 May 2003. J17038 B.K. Wells,1,2 S.R. Thorrold,3 and C.M. Jones. Center for Quantitative Fisheries Ecology, Department of Biological Sciences, Old Dominion University, Norfolk, VA 23529, U.S.A. 1

Corresponding author (e-mail: [email protected]). Present address: NOAA/NMFS, 110 Shaffer Rd., Santa Cruz, CA 95060, U.S.A. 3 Present address: Biology Department, Woods Hole Oceanographic Institute, Woods Hole, MA 02543, U.S.A. 2

Can. J. Fish. Aquat. Sci. 60: 361–369 (2003)

J:\cjfas\cjfas60\cjfas6004\F03-028.vp Tuesday, May 13, 2003 10:21:32 AM

doi: 10.1139/F03-028

© 2003 NRC Canada

Color profile: Disabled Composite Default screen

362

2003). For example, Wells et al. (2000b) demonstrated that elemental signatures in scales of juvenile weakfish (Cynoscion regalis) were specific to natal estuaries along the east coast of the United States, which was similar to results based on the otoliths from the same fish (Thorrold et al. 1998). However, the usefulness of scale geochemistry as a natural marker is dependent on the stability of the elemental signatures to the age when determination of natal location is desired. Although otoliths are generally considered to be metabolically inert (Campana 1999), the stability of scale chemistry has not been properly tested. It is known, for instance, that scales are vulnerable to resorption during times of increased calcium demand (Bilton and Robins 1971; Bilton 1975). Scales may also continue to crystallize after circuli are initially formed (Fouda 1979). The aim of our research was to determine if scale chemistry was sufficiently stable to be used as a natural tag of natal origins of adult weakfish. Our analysis of scale chemistry stability was in two parts. First, we examined correlations between the core chemistry of otoliths and scales from adult (1- and 2-year-old) weakfish. We then estimated the natal origins of adult weakfish based on elemental signatures in scales. The accuracy of these assignments was tested by comparing the results with estimates obtained from geochemical signatures in otoliths from the same fish.

Methods Study species Weakfish support recreational and commercial fisheries along the Atlantic coast of the United States. Adults follow a seasonal migration, moving south and offshore during the fall and winter and north and inshore during the spring and summer spawning season (Nesbit 1954; Wilk 1979). Spawning occurs throughout the species range in estuarine and nearshore waters. Larvae remain in their respective nursery areas through spring and summer by using selective tidal transport (Rowe and Epifanio 1994). In the fall, juveniles migrate from estuaries to coastal marine waters south of Cape Hatteras, North Carolina, U.S.A., before reinvading estuaries the following spring. Most weakfish reach sexual maturity at age 1, with remaining individuals becoming sexually mature at age 2 (Lowerre-Barbieri et al. 1996). Importantly, in the context of the present study, each fish experiences at least two distinctly different environments as juveniles in the estuaries and adults in the coastal marine zone. Therefore, the stability of the natural elemental signatures in the scales of juvenile weakfish can be inferred from a lack of significant alteration of the estuarine elemental signature while resident in the marine environment. Quantification of the natal-estuary elemental signatures In a previous study (Wells et al. 2000b), we quantified elemental signatures (Mg/Ca, Mn/Ca, Sr/Ca, and Ba/Ca (Mg, magnesium; Ca, calcium; Mn, manganese; Sr, strontium; Ba, barium)) in scales of juvenile (age-0+) weakfish from the 1996 and 1997 cohorts using laser ablation inductively coupled plasma mass spectrometry (ICP-MS). We found significant differences in elemental signatures among five estuaries along the Atlantic coast of the United States (Doboy Sound, Pamlico Sound, Chesapeake Bay, Delaware Bay, and Peconic

Can. J. Fish. Aquat. Sci. Vol. 60, 2003

Bay). However, before these signatures could be used to determine natal origins of adult fish, we first needed to assess the stability of these signatures in adult weakfish. Adult data collection Age-1 and age-2 weakfish were collected between June and September 1998 from Pamlico Sound, Chesapeake Bay, and Delaware Bay using a long-haul seine, a pound net, and a 10m otter trawl, respectively (Fig. 1). Ages were determined from validated annuli in thin-sectioned otoliths (LowerreBarbieri et al. 1994). Scales were removed from the fish and decontaminated for elemental analyses. The cleaning process included sonication for 5 min in 3% ultrapure hydrogen peroxide to loosen organic debris followed by two rounds of washing with an acid-washed electric rotary toothbrush and Milli-Q water (Millipore, Billerica, Mass.). This methodology was identical with the preparation of juvenile scales for elemental analysis in our previous study. Scales were then secured onto petrographic slides with mounting tape for elemental analysis by laser ablation ICP-MS. Otoliths from the same fish were analyzed as part of a larger study examining natal homing of adult weakfish. Details on the otolith geochemical analyses can be found in Thorrold et al. (1998) and Thorrold et al. (2001). Chemical analysis of adult scales The elemental composition of scales from adult weakfish was analyzed using a Thermo Finnigan Element2 sector field ICP-MS (San Jose, Calif.) and a New Wave LUV266 laser ablation system (Fremont, Calif.) operating at 266 nm (Thorrold and Shuttleworth 2000). A square 0.3-mm2 raster was ablated through the osseous layer from the core toward the anterior edge representing the same period of growth assayed in the juvenile scales so that direct comparisons could be made with the juvenile scales assayed in our earlier study (Wells et al. 2000b). The ablated material was swept by a carrier gas (He) into a dual-inlet quartz spray chamber. The He stream was then mixed with a wet aerosol (1% HNO3) from a 20 µL·min–1 self-aspirating nebulizer. A total of five isotopes were quantified (25Mg, 48Ca, 55Mn, 86Sr, and 138Ba), all in analog mode and in low resolution (r = 300). A blank sample consisting of the 1% HNO3 wet aerosol and a matrix-matched laboratory standard were introduced every 10 samples through the nebulizer to account for deposition on cones and changes in elemental mass bias through time. Measurement precision (relative standard deviation) of our lab standard, uncorrected for variations in elemental mass bias, was as follows (N = 33): Mg/Ca, 4.6%; Mn/Ca, 1.1%; Sr/Ca, 1.7%; Ba/Ca, 2.8%. These estimates are likely to be conservative because scale samples were corrected for mass bias using the standard. We calculated detection limits as 3σ values of 1% HNO3 sample blanks (N = 33) that were run throughout the analyses. These limits were <1% of the overall mean blank-corrected intensities obtained from the scale analyses for all five isotopes. Quantification of the elemental ratios followed the procedure outlined by Rosenthal et al. (1999). Isotopic counts were blank-corrected and converted to elemental intensities by multiplying percent natural occurrence of the isotopes. All data were standardized to Ca to account for variability in laser energy and the amount of ablated material. Finally, © 2003 NRC Canada

J:\cjfas\cjfas60\cjfas6004\F03-028.vp Tuesday, May 13, 2003 10:21:33 AM

Color profile: Disabled Composite Default screen

Wells et al.

363

Fig. 1. Five estuaries on the Atlantic Coast were sampled for juvenile weakfish, Cynoscion regalis, in 1996 and 1997. These locations were Doboy Sound, Pamlico Sound, Chesapeake Bay, Delaware Bay, and Peconic Bay. Age-1 and age-2 weakfish were collected during the 1998 spawning season from Pamlico Sound, Chesapeake Bay, and Delaware Bay.

data were converted to molar ratios by comparison of the intensities from samples with intensities from standards. Samples were analyzed in blocks of six, with one randomly chosen fish from each estuary and each age. Statistical analysis Stability of elemental signatures at the scale core through age 2 was evaluated by correlating elemental concentrations in scales with levels in the core of otoliths from the same fish. Values of Mn/Ca, Sr/Ca, and Ba/Ca in the scales of juvenile weakfish were significantly correlated with elemental concentrations in the otoliths (r = 0.55, 0.59, and 0.72, respectively, N = 253; Wells et al. 2000b), but Mg/Ca values between the two structures were uncorrelated. If scale chemistry remained stable over time, we would expect that elemental concentrations at the core of scales and otoliths of individual adult weakfish should remain similarly correlated.

We also examined the accuracy of elemental signatures in scales as natural tags of natal origin. The proportion of adult fish in the sample from each natal estuary was determined with maximum likelihood estimation (MLE; Millar 1990a) parameterized with elemental signatures from the baseline known-origin juvenile data sets (Mg/Ca, Mn/Ca, Sr/Ca, and Ba/Ca; Wells et al. 2000b). We used the maximum likelihood estimator (Millar 1990a, 1990b) and estimated errors with 1000 bootstrap-sampled baseline and adult sets (Fournier et al. 1984). Although MLE is the preferred approach for estimating overall proportional stock composition (Millar 1990b), the technique is not appropriate to assign individual fish to their natal source. Therefore, we used linear discriminant function analysis (LDFA) with elemental signatures from the juveniles as a baseline to determine the natal origins of individual adults. Before determining the contribution of each natal estuary © 2003 NRC Canada

J:\cjfas\cjfas60\cjfas6004\F03-028.vp Tuesday, May 13, 2003 10:21:33 AM

Color profile: Disabled Composite Default screen

364

to the adult sample either with MLE or LDFA, the elemental data were transformed and standardized. Residual analysis was used to test assumptions of homogeneity of variances and normality for the elemental data from juveniles (Winer 1971). To meet these assumptions, it was necessary to loge transform Ba/Ca and Mn/Ca ratios for the juvenile scale data (Wells et al. 2000b), and therefore, concentrations from the adults were also transformed before the mixed-origin compositions and associated errors were estimated. The element/ Ca concentrations in scales from juveniles and adults were standardized to a mean of zero and standard deviation (SD) of one (standard value = (observed – group mean)/SD). Each cohort and age group was standardized independently. This practice accounted for absolute differences in element/Ca concentrations between the age groups. Finally, the results from the LDFA and MLE analyses were compared with our best estimates of true natal origins. Allocations of adults determined from elemental signatures in scales were compared with estimates of natal origin from geochemical signatures in otoliths from the same adult fish, classified using baseline otolith signatures in juvenile fish of known origin as was done with the scales (Thorrold et al. 1998, 2001). Otolith signatures were based on the same elemental ratios as were analyzed in the scales, but included stable carbon (δ13C) and oxygen (δ18O) isotopes. Cross-validated accuracy of LDFA analysis of the 1996 juvenile otolith reference set using both elemental and stable isotope ratios ranged from 86 to 93% (Thorrold et al. 1998). Based on these data and on the metabolically inert nature of otoliths, we assumed that the otolith results represented the true natal origin of the adult weakfish for comparison with natal origin allocations based on scale chemistry from the same fish.

Results Age-1 (N = 46) and age-2 (N = 43) weakfish were collected for both scale and otolith chemical analyses (Table 1). Based on histological analysis of ovaries, all females used for analyses either showed evidence of spawning or were preparing to spawn, and macroscopic examination of testes confirmed that males were mature. The most obvious pattern in the elemental data was the sharp decline in Mg/Ca and Mn/Ca concentrations between scales from juvenile and adult weakfish. Indeed, Mg/Ca and Mn/Ca values in scales from adults were one-third of the values observed for juveniles (Fig. 2). The large declines in Mg/Ca and Mn/Ca in scales clearly suggested that there had been postdepositional changes in scale chemistry between the juveniles leaving their natal estuary and returning to spawn later. However, not all elements in the scales showed obvious evidence of metabolic reworking. Levels of both Sr/Ca and Ba/Ca were similar in magnitude between juvenile and adult scales, suggesting that these elements were relatively stable over time (Fig. 2). Correlations between scale and otolith chemistry A further test of temporal stability of elemental signatures in scales was provided by correlating elemental ratios in scales and otoliths from the same adult fish (Fig. 2). Because Mg/Ca levels were not significantly correlated in scales and otoliths of juvenile weakfish, we had no reason to anticipate

Can. J. Fish. Aquat. Sci. Vol. 60, 2003

any correlation in adult fish. Indeed, Mg/Ca ratios were not significantly correlated in adults. Despite the observation that Mn/Ca levels were significantly lower in the scales of adult weakfish compared with those of juveniles, Mn/Ca ratios in scales and otoliths remained strongly correlated for age-1 (r = 0.89, N = 46, P = 0.0001) and age-2 (r = 0.61, N = 43, P = 0.0001) fish. Ba/Ca levels in the scales of adult and juvenile weakfish were of similar magnitude, and Ba/Ca ratios in adult scales and otoliths were significantly correlated for age-1 (r = 0.76, N = 46, P = 0.0001) and age-2 (r = 0.44, N = 43, P = 0.0028) fish. Finally, Sr/Ca concentrations in the scales of juvenile and adult weakfish were of similar magnitude, but Sr/Ca levels in otoliths and scales of adult weakfish were not significantly correlated for age-1 fish (r = 0.18, N = 46, P = 0.2875) and were only marginally correlated for age-2 fish (r = 0.34, N = 43, P = 0.0261). Determination of adult natal origins Scale chemistry was clearly altered some time between juvenile residency in estuaries and the subsequent return of spawning adults. Nonetheless, the observation that some scale and otolith element/Ca levels remained significantly correlated, even if absolute levels in scales were lower, suggested that it might still be possible to recover information on natal origins from chemical signatures in scales. To test this possibility, we used MLE and LDFA to classify natal origins of the adult weakfish based on scale chemistry. We initially determined the natal origins of the adult sample using geochemical signatures in the otoliths of the adult fish and baseline signatures from juvenile weakfish (N = 253 from 1996 cohort and N = 367 from the 1997 cohort) from each of the five estuaries. The proportion of fish in the sample from each of the five natal estuaries was determined using MLE, and natal origins of individuals was determined using LDFA. Otolith-based MLE allocations of natal origins found estuarine contributions to the age-1 sample ranged from 9 (Doboy Sound) to 43% (Delaware Bay) and contributions to age-2 fish ranged from 3 (Doboy Sound) to 32% (Pamlico Sound; Table 2). We then tested the accuracy of allocations from scale chemistry, assuming that the otolith results were accurate. Maximum likelihood estimation, parameterized with the baseline scale data from the juvenile collections (N = 146 from 1996 cohort and N = 125 from the 1997 cohort; Wells et al. 2000b), was used to determine the proportion of adults in the sample from each of the five estuaries represented in the baseline juvenile data set. Scale allocation of age-1 fish was not significantly different from otolith allocation (χ 2 = 4.83, df = 4, P = 0.3049, 1 – β = 0.19; Table 2) and neither scale nor otolith allocations was random (df = 4, P < 0.01). However, scale allocation of age 2 overestimated Doboy Sound and Peconic Bay and underestimated Chesapeake Bay (χ 2 = 41.93, df = 4, P < 0.0001; Table 2). We also estimated individual allocations of age-1 and age2 fish from scale and otolith chemistry. Almost one-half (41%) of the 46 age-1 fish was classified to the same natal location by geochemical signatures in otoliths and scales (Table 3). However, only 7% of the 43 age-2 fish was allocated similarly by the two structures (Table 3). The probability of an individual fish being randomly allocated to the same natal location by both scale and otolith signatures is © 2003 NRC Canada

J:\cjfas\cjfas60\cjfas6004\F03-028.vp Tuesday, May 13, 2003 10:21:33 AM

Color profile: Disabled Composite Default screen

Wells et al.

365 Table 1. Adult weakfish, Cynoscion regalis, were collected from Pamlico Sound (PS), Chesapeake Bay (CB), and Delaware Bay (DB) during June to September 1998. Age 1

Age 2

Location

Date

N

TL

SE

N

TL

SE

PS CB DB

1 July – 1 Sept. 16 June – 16 Aug. 25 June – 20 July

17 15 14

20.7 21.2 21.4

0.37 0.19 0.23

14 12 17

23.5 25.2 26.0

0.29 0.39 0.36

Note: Collection date, sample size (N), mean total lengths (TL, cm), and standard errors (SE) are shown for each group of data.

Fig. 2. Plots of (a) Mg/Ca, (b) Mn/Ca, (c) Sr/Ca, and (d) Ba/Ca ratios (mmol·mol–1) in core regions of otoliths and scales from individual age-1 (open circle) and age-2 (solid circle) weakfish, Cynoscion regalis, collected from June to September 1998 from Pamlico Sound, Chesapeake Bay, and Delaware Bay. Significant correlations between adult otolith and scale data were detected for Mn/Ca and Ba/Ca ratios. Ellipses cover 95% of data points from otoliths and scales of juvenile weakfish collected in all five estuaries in 1997 (open ellipse) and 1996 (shaded ellipse).

approximately 4% (0.202) if there was an equal distribution of fish across all natal locations. To examine directly alterations in the scale signatures, we estimated mean element/Ca values in scales for age-1 and age-2 fish allocated to each natal estuary by geochemical signatures in otoliths (Fig. 3). Generally, the natal signatures in scales from age-1 fish were similar to signatures in juveniles, even if absolute concentrations were altered. For instance, Mn/Ca and Ba/Ca ratios in juvenile scales were highest in Peconic Bay and Chesapeake Bay, respectively, and we found the same pattern for both elements in scales from age1 fish. In contrast, we found remarkably little variation among

natal locations for any of the elements in age-2 fish, suggesting that original differences in juvenile scale signatures had been almost entirely overwhelmed by the time the fish had reached this age.

Discussion Analysis of trace elements in scales has been used to identify poached fish (Belanger et al. 1987), understand stock structure (Pender and Griffin 1996), determine river of origin (van Coillie and Rousseau 1974; Wells et al. 2003), determine marine residency (Bagenal et al. 1973), and identify © 2003 NRC Canada

J:\cjfas\cjfas60\cjfas6004\F03-028.vp Tuesday, May 13, 2003 10:21:33 AM

Color profile: Disabled Composite Default screen

366

Can. J. Fish. Aquat. Sci. Vol. 60, 2003 Table 2. Estimated contribution of each natal estuary (±1 standard deviation) to a sample of age-1 (N = 46) and age-2 (N = 43) weakfish, Cynoscion regalis, collected from Pamlico Sound, Chesapeake Bay, and Delaware Bay in 1998. Hard part

DS

Age-1 allocations (%) Scale 7 (±1) Otolith 9 (±6) Age-2 allocations (%) Scale 27 (±13) Otolith 3 (±7)

χ2

PS

CB

DB

PB

P

13 (±10) 23 (±7)

17 (±12) 16 (±7)

48 (±15) 43 (±9)

15 (±7) 9 (±6)

4.83

0.3049

19 (±10) 32 (±8)

2 (±5) 21 (±8)

13 (±14) 19 (±8)

39 (±10) 26 (±7)

41.93

0.0001

Note: A maximum likelihood procedure, parameterized with baseline juvenile collections from each of the five estuaries, was used to determine natal origin based on the elemental composition of scales (Mg/Ca, Mn/Ca, Sr/Ca, and Ba/Ca), and the stable isotope and trace element composition of otoliths (δ13C, δ18O, Mg/Ca, Mn/Ca, Sr/Ca, and Ba/Ca). We also calculated χ2 values for comparisons between allocations from scale and otolith chemistry (df = 4). Natal origins are labeled as Doboy Sound (DS), Pamlico Sound (PS), Chesapeake Bay (CB), Delaware Bay (DB), and Peconic Bay (PB).

Table 3. Summary of individual age-1 (N = 46) and age-2 (N = 43) weakfish, Cynoscion regalis, allocations to natal location based on geochemical signatures in otoliths (δ13C, δ18O, Mg/Ca, Mn/Ca, Sr/Ca, and Ba/Ca) and trace element signatures in scales (Mg/Ca, Mn/Ca, Sr/Ca, and Ba/Ca). Otolith allocation Scale allocation Age-1 allocations DS PS CB DB PB Age-2 allocations DS PS CB DB PB

DS

PS

CB

DB

PB

0 1 1 2 0

0 5 2 2 1

1 0 4 2 2

5 1 1 8 1

0 3 1 1 2

0 0 1 0 1

3 1 1 1 7

0 2 0 2 5

2 4 1 1 1

5 0 1 3 1

Note: Natal origins were determined using linear discriminant function analyses (LDFA) parameterized with baseline data from juvenile weakfish collected in each of the natal estuaries in 1996 and 1997. Estuarine origins shown are Doboy Sound (DS), Pamlico Sound (PS), Chesapeake Bay (CB), Delaware Bay (DB), and Peconic Bay (PB). Bold values on the main diagonal indicate individuals that were allocated by LDFA to the same natal estuary by both otolith and scale approaches.

nursery areas (Coutant and Chen 1993). Little work has been done comparing scale and otolith chemistries (Wells et al. 2000b; Gillanders 2001; Wells et al. 2003). Until the present study, no work has addressed the stability of scale chemistry by examining the same region of the scale across years. Several lines of evidence suggested that the minor and trace element chemistry of weakfish scales deposited while juveniles were resident in natal estuaries was significantly altered sometime before the fish returned to spawn in estuaries 1 and 2 years later. First, Mg/Ca and Mn/Ca values at the center of adult scales were considerably lower than in the same region of scales from juvenile weakfish. Second, correlations between otolith and scale chemistries in adult fish were lower than those found between the same two structures in juveniles. Finally, although scale-based allocation on age-1 weakfish to natal estuaries was similar to otolith-based estimates defined using the same elements and δ13C and δ18O

values, the statistical power of the test was low. Age-2 scale allocations were, however, significantly different from otolith allocation. Direct examination of mean element/Ca levels in scales of adults classified to each natal estuary by otolith chemistry confirmed that the signatures degraded over time. Taken together and assuming that otolith chemistry was stable over the same time period, we concluded that elemental signatures in scales were altered after the juvenile period and that relative elemental concentrations between natal signatures were altered after maturation. There are at least two possible causes for the alteration of the elemental signatures in scales over time. The first relates to the potential for postdepositional crystallization of hydroxyapatite. Scales consist of a distal layer composed of an organic framework impregnated with hydroxyapatite (van Oosten 1957; Hamada and Mikuni 1990) and a proximal layer that is an uncalcified fibrillary plate (van Oosten 1957; Fouda 1979). The osseous layer forms by accretion at the margin and does not thicken as the fish ages (van Oosten 1957; Fouda 1979). However, even minimal continued crystalization of central regions of the scale may have substantial effects on estuary-specific signatures. Fouda (1979) found that P/Ca decreased from the scale anterior edge to the core of scales. Several other studies also suggested that the ratio of Mg to Ca in hydroxyapatite crystals decreased as the hydroxyapatite formed (e.g., Gedalia et al. 1982; Bigi et al. 1992; Aoba et al. 1992). Further, additional calcification in the marine environment after initial deposition would lead to a decrease in Ba/Ca and Mn/Ca ratios if fewer Ba and Mn ions relative to Ca were present during this process. Similarly, there would be an increase of Sr/Ca if more Sr ions relative to Ca were present. We found that Mg/Ca and Mn/Ca in the scales from adults were significantly lower than in those from juveniles. Reductions in Ba/Ca in the scales were subtle after the first year but became more apparent after 2 years. Finally, Sr/Ca values were generally higher in adult scales. Given these observations, we suggest that scale crystallization in the region representing juvenile growth was the primary mechanism for alteration of elemental concentrations in the scales during the first year following the juvenile period. The second possible mechanism for the lack of stability of chemical signatures in scales is founded on a change in Ca demand during the reproductive season. Ca demand is relatively high during gonadal development compared with dur© 2003 NRC Canada

J:\cjfas\cjfas60\cjfas6004\F03-028.vp Tuesday, May 13, 2003 10:21:34 AM

Color profile: Disabled Composite Default screen

Wells et al.

367

Fig. 3. Mean element/Ca ratios (mmol·mol–1, ±1 standard error) in scales of juvenile (shaded bar) and adult (solid bar) weakfish, Cynoscion regalis, with natal origins of Doboy Sound (DS), Pamlico Sound (PS), Chesapeake Bay (CB), Delaware Bay (DB), and Peconic Bay (PB). Adult origins were determined with linear discriminant function analysis results from geochemical signatures in otoliths (δ13C, δ18O, Mg/Ca, Mn/Ca, Sr/Ca, and Ba/Ca) parameterized with baseline data from otoliths of juvenile weakfish collected in 1997 and 1996. Shown are (a, e) Mg/Ca, (b, f) Mn/Ca, (c, g) Sr/Ca, and (d, h) Ba/Ca in scales from juveniles from 1997 (a–d) and 1996 (e–h) and the age-1 (a–d) and age-2 (e–h) fish collected in 1998 from Pamlico Sound, Chesapeake Bay, and Delaware Bay.

ing nonreproductive periods (Mugiya and Watabe 1977; Kalish 1991). Kalish (1991) also showed an increase of Ca in the blood plasma during the spawning season. Importantly, experimentation has shown that injections of estradiol lead to mobilization of Ca from the scale to the blood plasma preferentially over other hard parts (Mugiya and Watabe 1977). In the present study, selective mobilization of Ca from scales would have led to increases in element/Ca ratios, whereas Mg/Ca and Mn/Ca ratios decreased dramatically in adult scales. It was possible, therefore, that we observed selective mobilization of Mg and Mn, at least compared with Ca. Interestingly, Sr/Ca and Ba/Ca ratios were similar in both juvenile and adult scales, suggesting that mobilization rates of both elements were similar to that of Ca. Although there is some evidence that Sr may, indeed, be at least as stable in the scale as Ca (Norris et al. 1963), very

little is known about the residence times of other elements in fish scales. Further studies will therefore be needed to test the veracity of this hypothesis. However, our data were consistent with the hypothesis that alteration of scale elemental signatures between ages 1 and 2 was due to the reworking of scale material during the first spawning season and subsequent remineralization in coastal waters during the resting period. Based on the data that we have presented, there appears little doubt that scale chemistry is significantly altered after initial deposition. Nonetheless, both Mn/Ca and Ba/Ca in otoliths and scales remained significantly correlated in adult weakfish. The Mn data were particularly interesting, as adult Mn/Ca values were some 70% less than levels in juvenile scales. This suggests that Mn mobilization in scales is relatively constant among individuals and (or) that little Mn is © 2003 NRC Canada

J:\cjfas\cjfas60\cjfas6004\F03-028.vp Tuesday, May 13, 2003 10:21:34 AM

Color profile: Disabled Composite Default screen

368

added to scales during additional calcification. Some elements (e.g., Mn and Ba) may be more abundant in estuarine waters, where juvenile weakfish initially reside, than in coastal waters, where further calcification is presumably occurring. We might predict, therefore, that these elements would be found in lower concentrations in adult scales. The observation that Mn/Ca values were lower is consistent with such an explanation, and yet Ba/Ca ratios showed only moderate change between juvenile and adult scales; this underscores that we still know very little about processes determining minor and trace element composition of marine fish scales. However, Ca was shown to be selectively mobilized during metabolic reworking of biogenic apatite in other vertebrates, leading Ba/Ca ratios in the apatite to increase with age (Manea-Krichten et al. 1991). A similar process may be occurring in fish scales. We used a correlative approach to investigate the stability of elemental signatures in weakfish scales. Several lines of evidence argued that the elemental signatures were not stable over time. However, a more definite study could make use of the observation that scales could be removed from juvenile weakfish without the need to sacrifice the individual. The fish could then be recaptured at some later stage, and scales collected and analyzed to determine stability directly. Elemental signatures in scales may still contain useful information even if there is significant alteration after initial deposition, providing relative abundances of the element of interest are retained in the scale. Such a situation appears to be the case for Mn/Ca in weakfish scales. Finally, it is worth noting that elemental signatures in weakfish otoliths were not much better predictors of natal origin than were elemental signatures in scales. Thorrold et al. (1998) found an increase of -25% in accurate classification when carbon and oxygen stable isotopes (δ13C and δ18O) were analyzed in addition to minor and trace elements in otoliths of juvenile weakfish. We expect that the additional analyses of δ13C and δ18O, and perhaps stable isotopes of N, S, and Sr, will provide similar improvements to the determination of natal origins from scale chemistry (Kennedy et al. 1997). However, the stability of stable isotope signatures in scales remains unknown. Significant mobilization of scale material should not affect the isotopic composition of those ions remaining in the scale, but secondary calcification could be problematic if elements such as Sr are also deposited after initial scale formation. Nonetheless, the ability to collect samples from fish in a nonlethal manner remains a powerful motivation for further work on elemental and isotopic signatures in scales.

Acknowledgements This work is a result of research sponsored in part by National Oceanic and Atmospheric Administration (NOAA) Office of Sea Grant, U.S. Department of Commerce, under grant NA56RG0489 to the Virginia Graduate Marine Science Consortium and Virginia Sea Grant College Program and also National Science Foundation grants OCE-9876565 and OCE-9871047 (C.M.J., S.R.T.). The U.S. Government is authorized to produce and distribute reprints for government purposes notwithstanding any copyright notation that may appear hereon. Laser ablation ICP-MS analyses of juvenile

Can. J. Fish. Aquat. Sci. Vol. 60, 2003

scales were conducted by S. Wang (Elemental Research Inc.) and adult scales were analyzed by C. Latkoczy (LITERODU). We thank L. Barbieri, L. Daniel, J. Fortuna, P. Geer, C. Grahn, M. Greene, S. Lowerre-Barbieri, and T. Targett for collection of weakfish. We also thank the anonymous reviewers for helpful comments on the manuscript.

References Aoba, T.E., Moreno, C., and Shimoda, S. 1992. Competitive adsorption of magnesium and calcium ions onto synthetic and biological apatites. Calcif. Tissue Int. 51: 143–150. Bagenal, T.B., Mackereth, F.J.H., and Heron, J. 1973. The distinction between brown and sea trout by the strontium content of their scales. J. Fish Biol. 5: 555–557. Bath, G.E., Thorrold, S.R., Jones, C.M., Campana, S.E., McLaren, J.W., and Lam, J.W.H. 2000. Sr and Ba uptake in otoliths of marine fish. Geochim. Cosmochim. Acta, 64: 1705–1714. Belanger, S.E., Cherry, D.S., Ney, J.J., and Whitehurst, D.K. 1987. Differentiation of freshwater versus saltwater striped bass by elemental scale analysis. Trans. Am. Fish. Soc. 116: 594–600. Bigi, A., Foresti, E., Gregorini, R., Ripamonti, A., Roveri, N., and Shah, J.S. 1992. The role of magnesium on the structure of biological apatites. Calcif. Tissue Int. 50: 439–444. Bilton, H.T. 1975. Factors influencing the formation of scale characters. Int. North Pac. Fish. Comm. Bull. 32: 102–108. Bilton, H.T., and Robins, G.L. 1971. Effects of starvation, feeding, and light period on circulus formation on scales of young sockeye salmon (Oncorhynchus nerka). J. Fish. Res. Board Can. 28: 1749–1755. Campana, S.E. 1999. Chemistry and composition of fish otoliths: pathways, mechanisms and applications. Mar. Ecol. Prog. Ser. 188: 263–297. Campana, S.E., and Thorrold, S.R. 2001. Otoliths, increments, and elements: keys to a comprehensive understanding of fish populations? Can. J. Fish. Aquat. Sci. 58: 30–38. Coutant, C.C., and Chen, C.H. 1993. Strontium microstructure in scales of freshwater and estuarine striped bass (Morone saxatilis) detected by laser ablation mass spectrometry. Can. J. Fish. Aquat. Sci. 50: 1318–1323. DiBacco, C., and Levin, L.A. 2000. Development and application of elemental fingerprinting to track the dispersal of marine invertebrate larvae. Limnol. Oceanogr. 45: 871–880. Fouda, M.M. 1979. Studies on scale structure in the common goby Pomatoschistus microps Kroyer. J. Fish Biol. 15: 173–183. Fournier, D.A., Beacham, T.D., Riddell, B.E., and Busack, C.A. 1984. Estimating stock composition in mixed fisheries using morphometric, meristic, and electrophoretic characteristics. Can. J. Fish. Aquat. Sci. 41: 400–408. Gedalia, I., Mayer, I., Giron, J., and Deutsch, D. 1982. Fluoride deposition in the bones of rats determined by fluoride and X-ray diffraction analysis. Arch. Oral Biol. 27: 823–825. Gillanders, B.M. 2001. Trace metals in four structures of fish and their use for estimates of stock structure. Fish. Bull. 99: 410–419. Hamada, M., and Mikuni, A. 1990. X-ray diffraction analysis of sardine scale ash. Nippon Suisan Gakaishi, 56: 947–951. Kalish, J.M. 1991. Determinants of otolith chemistry: seasonal variation in the composition of blood plasma, endolymph and otoliths of bearded rock cod Pseudophycis barbatus. Mar. Ecol. Prog. Ser. 74: 137–159. Kennedy, B.P., Folt, C.L., Blum, J.D., and Chamberlain, C.P. 1997. Natural isotope markers in salmon. Nature (Lond.), 387: 766–767. © 2003 NRC Canada

J:\cjfas\cjfas60\cjfas6004\F03-028.vp Tuesday, May 13, 2003 10:21:34 AM

Color profile: Disabled Composite Default screen

Wells et al. Lowerre-Barbieri, S.K., Chittenden, M.E., Jr., and Jones, C.M. 1994. A comparison of a validated otolith method to age weakfish, Cynoscion regalis, with the traditional scale method. Fish. Bull. 92: 555–568. Lowerre-Barbieri, S.K., Chittenden, M.E., Jr., and Barbieri, L.R. 1996. The multiple spawning pattern of weakfish (Cynoscion regalis) in the Chesapeake Bay and Middle Atlantic Bight. J. Fish Biol. 48: 1139–1163. Manea-Krichten, M.C., Patterson, C.E., Miller, G.E., Settle, D., and Erel, Y. 1991. Comparative increases of lead and barium with age in human tooth enamel, rib, and ulna. Sci. Total Environ. 107: 179–203. Millar, R.B. 1990a. A versatile computer program for mixed stock fishery composition estimation. Can. Tech. Rep. Fish. Aquat. Sci. No. 1753. Millar, R.B. 1990b. Comparison of methods for estimating mixed stock fishery composition. Can. J. Fish. Aquat. Sci. 47: 2235–2241. Mugiya, Y., and Watabe, N. 1977. Studies on fish scale formation and resorption. II. Effect of estradiol on calcium homeostasis and skeletal tissue resorption in the goldfish, Carassius auratus, and the killifish, Fundulus heteroclitus. Comp. Biochem. Physiol. A, 57: 197–202. Nesbit, R.A. 1954. Weakfish migration in relation to conservation. U.S. Fish Wild. Serv. Spec. Sci. Rep. Fish. No. 115. Norris, W.P., Chavin, W., and Lombard, L.S. 1963. Studies of calcification in a marine teleost. Ann. N.Y. Acad. Sci. 109: 312–336. Pender, P.J., and Griffin, R.K. 1996. Habitat history of barramundi Lates calcarifer in a North Australian river system based on barium and strontium levels in scales. Trans. Am. Fish. Soc. 125: 679–689. Rosenthal, Y., Field, M.P., and Sherrell, R.M. 1999. Precise determination of element/calcium ratios in calcareous samples using sector field inductively coupled plasma mass spectrometry. Anal. Chem. 71: 3248–3253. Rowe, P.M., and Epifanio, C.E. 1994. Tidal stream transport of

369 weakfish larvae in Delaware Bay, U.S.A. Mar. Ecol. Prog. Ser. 110: 101–114. Thorrold, S.R., and Shuttleworth, S. 2000. In situ analysis of trace elements and isotope ratios in fish otoliths using laser ablation sector field ICP-MS. Can. J. Fish. Aquat. Sci. 57: 1232–1242. Thorrold, S.R., Jones, C.M., Swart, P.K., and Targett, T.E. 1998. Accurate classification of nursery areas of juvenile weakfish (Cynoscion regalis) based on chemical signatures in otoliths. Mar. Ecol. Prog. Ser. 173: 253–265. Thorrold, S.R., Latkoczy, C., Swart, P.K., and Jones, C.M. 2001. Natal homing in a marine fish metapopulation. Science (Wash., D.C.), 291: 297–299. van Coillie, R., and Rousseau, A. 1974. Composition minérale des éscailles du Catostomus commersoni issu de deux milieux différents: étude par microsopie électronique analytique. J. Fish. Res. Board Can. 31: 63–66. van Oosten, J. 1957. The skin and scales. In The physiology of fishes. Edited by M.E. Brown. Academic Press, New York. pp. 207–244. Wells, B.K., Bath, G.E., Thorrold, S.R., and Jones, C.M. 2000a. Incorporation of strontium, cadmium, and barium in juvenile spot (Leiostomus xanthurus) scales reflects water chemistry. Can. J. Fish. Aquat. Sci. 57: 2122–2129. Wells, B.K., Thorrold, S.R., and Jones, C.M. 2000b. Geographic variation in trace element composition of juvenile weakfish scales. Trans. Am. Fish. Soc. 129: 889–900. Wells, B.K., Rieman, B.E., Clayton, J.L., Horan, D.L., and Jones, C.M. 2003. Relationships between water, otolith, and scale chemistries of westslope cutthroat trout from the Coeur d’Alene River, Idaho: the potential application of hard-part chemistry to describe movements in fresh water. Trans. Am. Fish. Soc. 132: 409–424. Wilk, S.J. 1979. Biological and fisheries data of weakfish, Cynoscion regalis (Block and Schneider). NMFS Sandy Hook Laboratory, Highlands, New Jersey. NOAA Tech. Ser. Rep. No. 21. Winer, B. 1971. Statistical principles in experimental design. 2nd ed. McGraw-Hill, New York.

© 2003 NRC Canada

J:\cjfas\cjfas60\cjfas6004\F03-028.vp Tuesday, May 13, 2003 10:21:34 AM