The Giant Pacific Oyster (Crassostrea gigas) as a modern analog for fossil ostreoids: Isotopic (Ca, O, C) and elemental (Mg/Ca, Sr/Ca, Mn/Ca) proxies

Authors


Abstract

[1] Modern analogs are an essential part of palaeoclimate studies, because they provide the basis for the understanding of geochemical signatures of fossils. Ostreoids are common in many sedimentary sequences and because of their fast growth, high temporal resolution sampling of past seasonal variability is possible. Here, two shell structures of modern Giant Pacific Oysters (Crassostrea gigas), the chalky substance and foliate layers, have been sampled for trace element distributions (Mg, Sr, Mn) and stable isotope variability (C, O, Ca). Oxygen isotopes exhibit a clear seasonal signature. Mean carbon isotope values of different oysters agree within 0.1‰, but ontogenic variability is complicated by shell growth patterns and potential small vital effects. The calcium isotope ratios are found to be constant throughout ontogeny within analytical precision at a value of δ44/40Ca = 0.68 ± 0.16‰ (2 sd) SRM–915a which is consistent with other bivalve species. Calcium isotope ratios in oyster shell material might thus be a possible proxy for palaeo seawater calcium isotope ratios. Element/Ca ratios are significantly higher in the chalky substance than in the foliate layers and especially high Sr/Ca and Mn/Ca ratios are observed for the first growth season of the oysters. Mg/Ca ratios in the chalky substance show a negative correlation with δ18O values, compatible with a temperature dependence, whereas this correlation is absent in the foliate layers. Seasonal changes of Sr/Ca are controlled by metabolic processes, whereas for Mn/Ca an additional environmental control is evident.

1. Introduction

[2] Stable isotope (Ca, O, C) and elemental (Mg/Ca, Sr/Ca, Mn/Ca) distribution patterns in shell materials of calcifying organisms are controlled by seawater chemistry and environmental conditions. This fact enables the reconstruction of past seawater conditions and palaeoenvironments from carbonate fossils, such as brachiopods, bivalve mollusks, belemnites, and foraminifers [e.g., Buchardt, 1978; Veizer et al., 1999; Zachos et al., 2001; Steuber and Veizer, 2002; Farkaš et al., 2007; Prokoph et al., 2008; Blättler et al., 2012]. Some of these organisms, however, show species dependent fractionations of isotope and element ratios (vital effects) during the formation of their shells. The extent of vital effects in ancient specimens can be addressed by investigations of phylogenetically related modern species, in culture or from the field. Such calibrations are essential, because element concentrations and isotopic ratios may be proxies for different environmental parameters, depending on the studied species [e.g., Klein et al., 1996; Vander Putten et al., 2000; Lorrain et al., 2005; Freitas et al., 2006; Ford et al., 2010; Heinemann et al., 2011; Schöne et al., 2011; Freitas et al., 2012].

[3] The biomineralization of oysters is of particular interest because of their large economic value [e.g., Pauley et al., 1988] but also because of their common occurrence as fossils in the sedimentary record [e.g., Korte and Hesselbo, 2011]. Numerous studies investigated aspects of biomineralization of oysters, focusing on the role of organic molecules [e.g., Lee and Choi, 2007; Lee et al., 2011; Marie et al., 2011], bulk shell and amino acid composition [Almeida et al., 1998] and element concentrations in foliate and prismatic layers [Dauphin et al., 2013]. Studies reporting both, isotopic and elemental compositions of oyster shell material, especially of the chalky substance, however, are rare (but see Goodwin et al. [2013]).

[4] Here we focus on oyster shells from a well monitored area, where environmental factors have been studied continuously for more than 20 years, so that external influences on the organism can be approximated. We present high resolution data sets for Mg/Ca, Sr/Ca, and Mn/Ca ratios as well as isotope ratios of C, O, and Ca in two different shell structures, the chalky substance and foliate layers. We show the primary variability of the chemical and isotopic composition and investigate, which information about past environments can potentially be derived from the analysis of the respective tracers.

2. Materials and Methods

[5] Oysters of the species C. gigas were sampled by Karsten Reise (Alfred Wegener Institute, Sylt) in the fall of 2007 from the List Tidal Basin, Germany (Figure 1a). This semienclosed basin, covering an area of about 400 km2, is located in the Northern Wadden Sea and connected to the North Sea by a tidal inlet between the islands of Sylt, Germany and Rømø, Denmark (Figures 1a and 1b) [van Beusekom et al., 2008]. The specimens were sampled from a shallow subtidal oyster bank close to the Island of Sylt. Monthly measurements of water temperatures and salinities in the relevant time interval from 2003 to 2007 showed variations between 0.2 to 20.3°C and 27.2 to 31.4 psu [van Beusekom et al., 2008] (Figure 1c). The salinity changes are mainly controlled by discharge variations of the rivers Elbe and Weser [van Beusekom et al., 2008].

Figure 1.

(a and b) Sample location in the List Basin. (c) Salinities and temperatures at the sample location with monthly resolution.

2.1. Shell Material

[6] Oysters are macrobenthic, filtering, and opportunistic suspension feeders [Leal et al., 2008; Dauphin et al., 2013]. They grow preferentially on hard substrates in estuaries in the intertidal and shallow subtidal zones [Quayle, 1969; Pauley et al., 1988] and prefer water temperatures of 4–24°C and salinities >20 psu [Pauley et al., 1988]. For a more comprehensive picture of oyster ecology, see Quayle [1969] and Pauley et al. [1988].

[7] Shells of juvenile and adult oysters consist mainly of calcite [Lee et al., 2006]. The calcitic part of the shell of Crassostrea gigas comprises the outermost prismatic layer, foliate layers, and chalky substance (Figure 2) [Miyawaki, 1954; Carriker and Palmer, 1979; Carriker et al., 1980; MacDonald et al., 2009; Dauphin et al., 2013]. The chalky substance (Figure 2a) is specific to oysters [Dauphin et al., 2013] and apparently an early feature of oyster evolution. Its presence has been inferred for Cretaceous [Chinzei, 1986; Ayoub-Hannaa and Fürsich, 2011] and even Jurassic oysters [Komatsu et al., 2002]. Contrasting to the network of microcrystalline calcite laths of the chalky substance, the foliate layers are composed of interfingering bladelike calcite crystals (Figures 2a and 2b; see also Ullmann et al. [2010]).

Figure 2.

SEM images of shell structures of C. gigas: (a) chalky substance, (b) foliate layers, and (c) millimeter-sized sheet of organic material in contact with foliate layers. A thin package of foliate layers attached to the organic material is visible in the central part. Scale bars are 50 µm.

2.2. Sample Preparation

[8] The left valve of three shells was cut through the umbo and a ∼2 mm thick section of the umbo was mounted on a glass slide with epoxy resin for further sampling (Figure 3). One of the investigated profiles stems from the same oyster that was analyzed by Ullmann et al. [2010] (Figures 3a and 3b). Foliate layers and chalky substance [Carriker and Palmer, 1979; Carriker et al., 1980] were sampled to assess the potential effects of the differing ultrastructure on the geochemical tracers. A total of 91, 110, and 70 samples, respectively, were taken along traces through the sections of specimens vl, 8l and 1l, with an average resolution of ∼0.2 mm per sample. Aliquots were prepared by scratching shell calcite parallel to the growth bands using a stainless steel scalpel. Each sample was homogenized with a spatula and subdivided into subsamples for carbon and oxygen isotopic analyses and element ratio determinations. From 14 of the homogenized aliquots of specimen vl, calcium isotope ratios were analyzed.

Figure 3.

Sections through the umbo of the investigated specimens of C. gigas: (a) section of specimen vl used for this study, (b) parallel section of specimen vl studied by Ullmann et al. [2010], (c) section of specimen 8l, and (d) section of specimen 1l. Age estimates for the shell material are based on oxygen isotope results discussed in Ullmann et al. [2010] and in sections 3 and 4. Sample positions of the two profiles through specimen vl were correlated using growth bands. Dotted lines indicate approximate boundaries between growth bands attributed to different years.

2.3. Carbon and Oxygen Isotope Preparation

[9] The carbon and oxygen isotope measurements on specimen vl were carried out on ∼100–500 µg sample powder using the MAT-253 THERMOTM Finnigan Isotope Ratio Mass Spectrometer at the Freie Universität Berlin using the routines described in Ullmann et al. [2010]. The external reproducibility of the measurements was 0.07‰ (2 sd) for δ13C and 0.12‰ (2 sd) for δ18O as defined from the results for a laboratory reference material measured together with the samples (n = 20). The results were correlated with the higher resolution data set from Ullmann et al. [2010] by matching the growth bands of the investigated sections and assigning the new samples to the closest spaced sample from the higher resolution profile. No further adjustments in the spacing were applied. Carbon and oxygen isotope measurements of specimens 8l and 1l were conducted at the University of Copenhagen with a Micromass IsoPrime Isotope Ratio Mass Spectrometer, using the routines described in Ullmann et al. [2013]. Sample amounts were ∼400–800 µg. The external reproducibility of the analyses is 0.08‰ (2 sd) for δ13C and 0.16‰ (2 sd) for δ18O as defined by multiple analyses of a laboratory reference material (n = 56).

2.4. Trace Element Preparation

[10] Splits for element ratio determinations of specimen vl were processed in the clean air laboratory of the Department of Earth Sciences, University of Oxford, UK. The carbonate powders were transferred to acid cleaned tubes and weighed. The material was reacted with 500 µL of 0.3 M HNO3 and diluted with 0.3 M HNO3 to a calculated concentration of 100 µg/g calcium. Aliquots were measured using a PESCIEX ELAN 6100 DRC Quadrupole ICP-MS at the University of Oxford analogous to the method described in Ferguson et al. [2008] using a standard sample bracketing protocol with a synthetic multielement solution as a reference material. The element concentrations of the calibration solution were measured against the Merck ICP multielement solution VI for accuracy control. Repeatability for the calibration solution is better than 4% for Mg/Ca and Sr/Ca ratios and better than 5% for Mn/Ca ratios (2 sd, n = 113).

[11] Element ratios of samples from specimens 8l and 1l were measured at the University of Copenhagen using an Optima 7000 DV ICP-OES. Sample powders were reacted with 0.2 M HNO3 and Ca concentrations adjusted to a nominal concentration of 25 µg/g by dilution with 0.2 M HNO3. Element concentrations were determined using a three-point calibration with two matrix matched synthetic calibration solutions and a blank solution. Spectral intensities of Ca (317.933 nm), Mg (280.271 nm), and Sr (407.771 nm) were analyzed in radial mode, while Mn (257.610 nm) was measured in axial mode for better sensitivity. Precision and accuracy of the analyses were determined by multiple measurements of JLs-1 (limestone). Repeatability was better than 2% for Mg/Ca and Sr/Ca ratios and better than 5% for Mn/Ca ratios (2 sd, n = 24). Average results of the element ratios of JLs-1 were 13.9 mmol/mol for Mg/Ca, 0.343 mmol/mol for Sr/Ca, and 29.8 µmol/mol for Mn/Ca, which compares to published values of 15.3 mmol/mol, 0.343 mmol/mol, and 30.0 µmol/mol, respectively [Imai et al., 1996].

2.5. Calcium Isotope Measurements

[12] Calcium isotope ratios were measured with a Finnigan Triton TI TIMS (Thermal Ionization Mass Spectrometer) at the Geomar Helmholtz-Zentrum für Ozeanforschung Kiel, following the method described in Heuser et al. [2002] and Tang et al. [2008]. Aliquots of 0.1 mg were dissolved in 2.2 N ultrapure HCl, dried down and redissolved in 2.2 N HCl with a Ca concentration of 160 ng/µL. The sample solutions were mixed with a 43Ca/48Ca double spike (with the spike contributing about 90 % of 48Ca) and about 300 ng of each sample-spike mixture was loaded with TaCl5 activator on an outgassed zone-refined Re filament. Measurements were made on single filaments at a temperature of about 1500°C and a typical 40Ca signal intensity of 50 pA (5 V). Data acquisition was performed in dynamic mode. The acquired data were fractionation corrected online to the spike 43Ca/48Ca ratio of 0.748429 [Gussone, 2003] using the exponential fractionation law. The double spike correction was carried out with the iterative algorithm described by Heuser et al. [2002].

[13] The isotope values of Ca are reported as δ44/40Ca values relative to the NIST standard SRM915a, where δ44/40Ca = [(44Ca/40Ca)sample/(44Ca/40Ca)SRM915a − 1]×1000. The samples are normalized to the mean 44Ca/40Ca of four SRM915a analyses, run on the same turret. The sample precision is given as two times the standard error of the mean determined (2 err) by repeated measurements of the sample stock solutions. All samples were measured at least twice. The average 2 err is 0.12‰. The spike corrected mean 44Ca/40Ca of NIST SRM915a was 0.021185(3) (standard deviation, n = 17). Total Ca blanks for the isotope analyses were less than 1%.

3. Results

[14] The morphologies and succession of foliate layers and chalky substance in the umbo of the three specimens vary considerably (Figure 3), but isotopic and element profiles show strong similarities in trends and absolute values (Figure 4). The carbon and oxygen isotope values of the oysters (including the data from Ullmann et al. [2010]) vary between −1.9 and +1.0‰ and between −3.0 and +1.3‰, respectively (see supporting information). The data sets for specimen vl of the present study and of Ullmann et al. [2010] are in a good agreement (Figure 4). All four annual cycles of the years 2004–2007 reported in the previous study are represented here at about half of the former spatial resolution. The reduced resolution is due to the larger sample sizes necessary, and because of a less favorable section through the umbo of the shell (Figure 2; see sections 2.1 and 2.2 for details). Oxygen isotope profiles through specimens 8l and 1l indicate 5 growth seasons (2003–2007) (Figure 4).

Figure 4.

Geochemical results for shell carbonate plotted against distance from the earliest growth bands. Sample positions of specimen vl from this study are correlated with the previous study of Ullmann et al. [2010] by matching the growth bands. (a–e) Specimen vl; (f–j) specimen 8l; and (k–o) specimen 1l. White diamonds: chalky substance; gray diamonds: mixed samples containing material from foliate layers and chalky substance; black samples: foliate layers. Gray bands in Figures 4a and 4b represent isotope data presented in Ullmann et al. [2010].

[15] Where sufficient material was available for submonthly sample resolution, minimum δ18O values between the analyzed specimens agree within 0.4‰ for all growth seasons (Figure 4). Larger disagreements for annual maxima and average δ18O values are due to shell morphology and differential shell growth (Figure 4).

[16] Carbon isotope values show systematic variations in all three specimens and average δ13C values of the specimens of −1.27 to −1.17‰ agree within 0.1‰. An increase of the δ13C values at the beginning of each growth season is indicated, but annual trends differ in details.

[17] Trace element data are given in Figure 4 and the online supporting information. All the samples are composed of low-Mg-calcite with Mg/Ca ratios varying between 2.5 and 14.6 mmol/mol (Figures 4c, 4h, and 4m). The median Mg/Ca ratios for the foliate layers—excluding the first growth season—range from 3.7 to 4.1 mmol/mol (Figure 5). Mg/Ca ratios of the foliate layers are not correlated with δ18O (Figure 6a). The Mg/Ca ratios in the chalky substance, compared to the foliate layers, are enriched in Mg more than twofold with median Mg/Ca ratios of 9.4–12.3 mmol/mol (Figure 5a). In addition, Mg/Ca ratios in the chalky substance show a seasonal pattern with the highest ratios in the warm season (Figures 4c, 4h, and 4m). This seasonal pattern is expressed in a strong, negative, correlation of Mg/Ca ratios of all specimens with δ18O values of the form Mg/Cachalky = −2.85 ± 0.52 × δ18O + 6.19 ± 0.83 (r2 = 0.60, Figure 6a).

Figure 5.

Box plots of element ratios in foliate layers and chalky substance of the three specimens. The boxes represent the second and third quartile with the median value as a black line and the whiskers show the full range of observed ratios.

Figure 6.

Element ratios plotted against oxygen isotope values. Trend lines of correlations with r2 > 0.2 are plotted with 95% confidence intervals as gray bands.

[18] The Sr/Ca ratios of the specimens range from 0.60 to 1.14 mmol/mol (Figures 4d, 4i, and 4n). For all specimens, the highest Sr/Ca ratios are measured in the samples of the first growth season. A seasonal pattern of Sr/Ca ratios can be recognized but is not very pronounced (Figures 4d, 4i, and 4n) and weak, negative, correlations of Sr/Ca ratios with δ18O values exist for the samples from the respective first growth seasons (r2 = 0.30), foliate layers (r2 = 0.28), and chalky substance (r2 = 0.27) (Figure 6b). The equations with 95% confidence intervals for these correlations are Sr/Ca1st season = −0.0069 ± 0.0035 × δ18O + 0.858 ± 0.068; Sr/Cafoliate = −0.0048 ± 0.0015 × δ18O + 0.708 ± 0.021; and Sr/Cachalky = −0.0046 ± 0.0015 × δ18O + 0.771 ± 0.026. The median Sr/Ca ratios of the chalky substance of 0.82–0.87 mmol/mol are similar for the three specimens (Figure 5b). Also the median Sr/Ca ratios of the foliate layers of 0.71–0.80 mmol/mol (excluding data from the first growth season) are similar but 9–13% lower than the respective values of the chalky substance (Figure 5b).

[19] The Mn/Ca ratios exhibit the highest relative variability of the three element ratios and vary from 6 to 190 µmol/mol (Figures 4e, 4j, and 4o). Analogous to the Sr/Ca ratios, the Mn/Ca ratios decrease strongly from the highest ratios measured for the shell in the first growth season. A seasonal pattern with higher Mn/Ca ratios in the warm season is observed. Like Sr/Ca ratios, Mn/Ca ratios of foliate layers and chalky substance correlate negatively with δ18O values (r2 = 0.30 and 0.40, respectively; Figure 6c). As for Sr/Ca and Mg/Ca, median Mn/Ca ratios of the respective shell materials of the three specimens are similar, but median Mn/Ca ratios of the chalky substance are ∼20–50% higher than median Mn/Ca ratios of the foliate layers (Figure 5c).

[20] The calcium isotopic composition of specimen vl is constant within analytical uncertainty (Figure 7), precluding a visible ontogenic trend. No correlation with the oxygen isotope ratios is found and the chalky substance (δ44/40Ca = 0.66 ± 0.04‰) and foliate layers (δ44/40Ca = 0.70 ± 0.07‰) have indistinguishable compositions.

Figure 7.

Ca isotope values and respective 44Ca/40Ca fractionation factor (αcc-water) for shell calcite of C. gigas. White triangles: chalky substance and gray triangles: foliate layers. Black line: mean value with 2 err uncertainty.

4. Discussion

4.1. Carbon and Oxygen Isotopes

[21] The assumption of Ullmann et al. [2010] that C. gigas records a high fidelity δ18O signature is supported by the reproducibility of the signature within the same specimen (Figure 4a) and in other specimens (Figures 4f and 4k). The excellent agreement between the isotope profiles through two sections of the umbo of specimen vl (Figures 4a and 4b) indicates that the carbon and oxygen isotope ratios of coeval growth bands are uniform. The slightly smaller amplitude for the reproduced profile can be explained by time averaging [Goodwin et al., 2003], generated by the larger samples compared to the sample amounts that were taken for carbon and oxygen isotope analyses in Ullmann et al. [2010].

[22] The generally low variability of the δ13C values within the specimens of ∼0.7‰ (2 sd) and the very good agreement between the average values of different specimens attest that vital effects in C. gigas are not expressed as large isotope fluctuations and biases differing between individual specimens. Moreover, a correlation between oxygen and carbon isotope ratios, suggesting kinetic isotope fractionation during incorporation of the carbonate ion into the crystal lattice [e.g., Parkinson et al., 2005, for brachiopods], is not found here. These findings suggest that C. gigas records the δ13C signature of the water DIC without major bias. In the absence of direct measurements of δ13CDIC in the List Basin, this hypothesis cannot be further strengthened. It remains unclear, if increasing δ13C values in our shell material, coinciding with the timing of phytoplankton blooms in the late spring [van Beusekom et al., 2009], are related to changes in the ambient DIC [Wang et al., 1995; Goodwin et al., 2013] or are controlled by metabolic processes.

4.2. Ca Isotopes

[23] The lack of a correlation with temperature or seasonal effects in C. gigas, expressed by the uniformity of the calcium isotope ratios (Figure 7), is in agreement with an apparently small temperature dependence of Ca isotope fractionation in other biogenic carbonates [Steuber and Buhl, 2006; Farkaš et al., 2007; but see Immenhauser et al., 2005]. The δ44/40Ca of open ocean seawater was found to be globally homogeneous at a value of 1.88 ± 0.04‰ versus SRM915a [Hippler et al., 2003]. This is also true for most marginal seas, including the North Sea near Sylt and the Baltic Sea [Heinemann et al., 2008], the Mediterranean Sea [Böhm et al., 2006], and the Red Sea [Kisakürek et al., 2011]. The mean of the measured values of δ44/40Ca = 0.68 ± 0.04‰ versus SRM915a can therefore be used to define a fractionation factor αcc-water between the shell calcite and sea water (Figure 7). The fractionation factor determined for C. gigas from the δ44/40Ca value of seawater and the average δ44/40Ca value of the shell calcite is αcc-water = 0.99880 ± 0.00006. This result is in good agreement with the fractionation factor calculated for a calcite shell layer of Mytilus eduliscc-water = 0.9989) collected from the same region [Heinemann et al., 2008]. It further agrees with the average fractionation factor for marine biogenic calcite at ∼15°C from Gussone et al. [2005], αcc-water = 0.9989. The fractionation factor for C. gigas for the List Basin is significantly higher than the value of 0.9985 found by Steuber and Buhl [2006] for modern oysters from the North Sea. The differing fractionation factor found by Steuber and Buhl [2006] might be due to more pronounced freshwater influx changing the δ44/40Ca ratio of the ambient water [Heinemann et al., 2008; Holmden et al., 2012]. Also comparatively high Mn/Ca ratios of the specimens of Steuber and Buhl [2006] point to a freshwater influenced setting [Almeida et al., 1998].

4.3. Element Ratios

[24] Contrasting to the isotopic signatures of the oyster specimens, element ratios in the shell materials differ significantly between chalky substance and foliate layers (Figures 4-6). For Sr/Ca and Mn/Ca, the highest ratios are observed in the first growth season. The high Sr/Ca and Mn/Ca ratios recorded by specimen vl in 2004 are not observed in specimens 8l and 1l, for which 2004 represents the second growth season. This difference evidences that the enrichments of the respective first growth seasons are related to metabolic processes and not to an environmental control. The differing incorporation behavior of Mg into foliate layers and chalky substance strengthens this hypothesis. For the interpretation of element ratios, the samples from the first growth season are therefore treated separately.

[25] Mg/Ca ratios for oysters published so far range from 5.5 to 27 mmol/mol [Lerman, 1965; Ohde and Kitano, 1984; Almeida et al., 1998]. The average values of our specimens ranging from 5.8 to 6.8 mmol/mol fall into this range, but most of the samples from foliate layers show lower ratios (Figures 4c, 4h, 4m, 5a, and 6a). Salinities are too high and stable (Figure 1c) [van Beusekom et al., 2008; Ullmann et al., 2010] to result in a potential major effect on Mg/Ca ratios, either by changes in water composition [Dodd and Crisp, 1982; Carré et al., 2006], or salinity-related changes in element fractionation [Eisma et al., 1976]. The oxygen isotope values of the investigated oyster vary strongly throughout the year, following primarily the temperature of the ambient water (see Ullmann et al. [2010] for detailed discussion). The strong, negative, correlation of Mg/Ca ratios with δ18O values of the chalky substance (r2 = 0.60) indicates that Mn/Ca ratios is controlled largely by temperature, pointing to Mg/Ca ratios in the chalky substance of C. gigas as a promising proxy for temperatures. Mg/Ca temperature proxies using bivalve calcite [e.g., Vander Putten et al., 2000; Ford et al., 2010; Heinemann et al., 2011; Schöne et al, 2011; Freitas et al., 2012] can be interesting for environmental reconstructions, if the covariations of Mg/Ca with temperature are very strong. Further studies, including δ18O measurements of ambient water and using environmental data with better time resolution might show, if the chalky substance of C. gigas is suitable for such reconstructions.

[26] The Mg/Ca ratios of the foliate layers of C. gigas do not show a strong covariance with δ18O values (r2 = 0.05) and their averages of 4.1–4.3 mmol/mol (excluding the first growth season) agree within 5% for the three investigated specimens. Mg/Ca ratios of foliate layers might thus be used to estimate seawater Mg/Ca ratios. Together with a seawater Mg/Ca ratio of 5.2 mol/mol [de Villiers and Nelson, 1999], our observed average Mg/Ca ratios for the foliate layers of the three specimens translate into a Mg distribution coefficient of 0.00080 ± 0.00003 (2 err, n = 3) for C. gigas in the List Basin. The internal variability of each specimen, however, is on the order of 50% (2 rsd), necessitating the measurement of numerous or large samples to derive a reliable mean value.

[27] The average Sr/Ca ratio of our oyster material of 0.82 mmol/mol (n = 271) compares well with data published for oyster shell material of Malleus albus (0.81 mmol/mol, n = 1), C. gigas (0.83 mmol/mol, n = 268), Crassostrea virginica (0.86 mmol/mol, n = 10), and Saxostrea gigas (1.06 mmol/mol, n = 1) [Lerman, 1965; Ohde and Kitano, 1984; Almeida et al., 1998]. Weak negative correlations of Sr/Ca ratios with oxygen isotope values (Figure 6b) indicate that seasonal effects may play a role in the incorporation of Sr into the shell of C. gigas. Together with a seawater Sr/Ca ratio of 8.54 mmol/mol [de Villiers, 1999], and assuming an average δ18O value of the ambient water in the List Basin of −1.4‰ SMOW [Ullmann et al., 2010], calibrations for foliate layers and chalky substance can be derived from these empirical covariations. The resulting equations are DSr foliate = 0.0907 ± 0.0027 − 0.0056 ± 0.0017 × (δ18Ocalcite − δ18Owater) for foliate layers and DSr chalky = 0.0978 ± 0.0033 − 0.0053 ± 0.0020 × (δ18Ocalcite − δ18Owater) for the chalky substance of C. gigas. The above distribution coefficients are consistently lower than for Pinna nobilis [Richardson et al., 2004], Pecten maximus [Lorrain et al., 2005; Freitas et al., 2006], and brachiopods living in water depths of less than 100 m [Brand et al., 2003], a relationship that is also observed for Early Jurassic specimens [e.g., Korte and Hesselbo, 2011]. The Sr distribution coefficients for C. gigas in the List Basin might therefore serve as approximations for the derivation of past seawater Sr/Ca ratios using ostreoid materials.

[28] The Mn/Ca ratios in our oyster specimens, ranging from 6 to 190 µmol/mol (n = 271), are in very good agreement with the values reported for modern Portuguese oysters of 12–93 µmol/mol [Almeida et al., 1998], considering that only 6 values > 100 µmol/mol were measured in our samples. Among other features, peak Mn/Ca ratios in the second half of 2006 and comparatively low Mn/Ca ratios in 2007 are common to all three specimens (Figures 4e, 4j, and 4o), indicating that these systematic trends may be partially controlled by changing Mn/Ca ratios of ambient water. Such an environmental control on Mn/Ca ratios has also been observed for Pecten maximus [Freitas et al., 2006]. The high Mn/Ca ratios of the oyster's first growth season (Figures 4e, 4j, 4o, and 5) and the differing Mn incorporation trends of foliate layers and chalky substance demonstrate that an additional biological control on Mn uptake, as also invoked by Freitas et al. [2006], is an important factor.

[29] Mn concentrations of biogenic carbonates are often taken as quality markers for fossil preservation [e.g., Prokoph et al., 2008; Korte et al., 2009]. Mn/Ca ratios of up to ∼100 µmol/mol [Almeida et al., 1998; Steuber and Buhl 2006], as well as Mn/Ca ratios of up to 450 µmol/mol in Mytilus edulis [Vander Putten et al., 2000], are evidence for potentially elevated Mn/Ca ratios in calcite shells that are not necessarily generated by post-depositional alteration (see also Korte and Hesselbo [2011]). Also, significant negative correlations of Mn concentrations with δ18O values are taken as markers for diagenetic trends [e.g., Brand and Veizer, 1981]. The negative correlations of Mn/Ca ratios with δ18O values found in modern C. gigas shells underline that the interpretation of such empirical covariations in fossil materials may not always be straight forward. In studies on fossil materials, additional proxies for alteration that are independent of Mn concentrations should be employed to address such ambiguities.

5. Conclusions

[30] C and O isotope trends and average values in all three specimens of modern C. gigas are similar, indicating that disequilibrium isotope fractionation plays a minor role during shell formation.

[31] Ca isotope ratios are constant within analytical precision throughout the lifespan of a single oyster and a fraction factor for 44Ca/40Ca of αcc-water = 0.99880 ± 0.00006 is derived. The isotopic uniformity of Ca suggests that palaeo-seawater calcium isotopic compositions can be deduced from shells of C. gigas (and other oysters).

[32] Element/Ca ratios show metabolically controlled enrichments in the first growth season and are lower in the foliate layers than in the chalky substance. Significant correlations of element/Ca ratios with oxygen isotope ratios and r2 > 0.20 are observed for all element ratios apart from Mg/Ca ratios in the foliate layers. An Mg distribution coefficient of DMg = 0.00080 ± 0.00003 is found for the foliate layers of C. gigas. The covariation of Mg/Ca ratios in the chalky substance of C. gigas with Mg/Cachalky = −2.85 ± 0.52 × δ18Ocalcite + 6.19 ± 0.83 can mainly be explained by temperature dependent element incorporation. Empirical Sr distribution coefficients are DSr foliate = 0.0907 ± 0.0027 − 0.0056 ± 0.0017 × (δ18Ocalcite − δ18Owater) for foliate layers and DSr chalky = 0.0978 ± 0.0033 − 0.0053 ± 0.0020 × (δ18Ocalcite − δ18Owater) for the chalky substance. Trends in Mn/Ca ratios are likely controlled by a combination of metabolic factors and the changing Mn/Ca ratio of the ambient water.

Acknowledgments

[33] The authors acknowledge the editor Louis Derry, an anonymous associate editor, and two anonymous reviewers for comments and suggestions that helped to greatly improve the manuscript. The authors acknowledge Karsten Reise for taking the samples in the field, Cees-Jan DeHoog for ICP-MS measurements at the University of Oxford, Bo Petersen for carbon and oxygen isotope analyses at the University of Copenhagen, and Stephen Hesselbo for providing accommodation and assistance during the work stay in Oxford. The project was funded by the Center for International Cooperation, Freie Universität Berlin (travel grant for CVU) and by the DFG (project Ei 272/13). We thank Anton Eisenhauer and Folkmar Hauff (GEOMAR) for help and support for the TIMS analyses at Geomar.

Ancillary