Interpreting the role of pH on stable isotopes in large benthic foraminifera: Towards a Broader Perspective on Ocean Acidification Research Part 2 A special issue of the ICES Journal of Marine Science

&NA; Large benthic foraminifera (LBF) are prolific producers of calcium carbonate sediments in shallow, tropical environments that are being influenced by ocean acidification (OA). Two LBF species, Amphistegina gibbosa (Order Rotaliida) with low‐Mg calcite tests and Archaias angulatus (Order Miliolida) with high‐Mg calcite tests, were studied to assess the effects of pH 7.6 on oxygen and carbon isotopic fractionation between test calcite and ambient seawater. The &dgr;18O and &dgr;13C values of terminal chambers and of whole adult tests of both species after 6 weeks were not significantly different between pH treatments of 8.0 and 7.6. However, tests of juveniles produced during the 6‐week treatments showed significant differences between &dgr;18O and &dgr;13C values from control (pH 8.0) when compared with the treatment (pH 7.6) for both species. Although each individual's growth was photographed and measured, difficulty in distinguishing and manually extracting newly precipitated calcite from adult specimens likely confounded any differences in isotopic signals. However, juvenile specimens that resulted from asexual reproduction that occurred during the experiments did not contain old carbonate that could confound the new isotopic signals. These data reveal a potential bias in the design of OA experiments if only adults are used to investigate changes in test chemistries. Furthermore, the results reaffirm that different calcification mechanisms in these two foraminiferal orders control the fractionation of stable isotopes in the tests and will reflect decreasing pH in seawater somewhat differently.


Introduction
Carbonate chemistry of surface seawater is changing as it maintains equilibrium with increasing atmospheric carbon dioxide concentration [CO 2atm ] (Doney et al., 2009).Models suggest that when [CO 2atm ] reaches 1300 ppm, average pH of global seawater could be as low as 7.6, with a calcite saturation state (V c ) of 2.0 (Po ¨rtner et al., 2014).These projected changes in both pH and V c are collectively referred to as ocean acidification (OA; Caldeira and Wickett, 2003), and laboratory experiments suggest that OA will significantly influence calcifying marine organisms with potential impacts soon (Doney et al., 2009).Meta-analysis of these biological responses, however, has shown great complexity with high variability between individual species (Kroeker et al., 2010).
Large benthic foraminifera (LBF) are symbiont-bearing protists that are prolific calcium carbonate (CaCO 3 ) producers in both modern ( 130 Mt year 21 ; Langer, 2008) and past (BouDagher-Fadel, 2008) oceans.Recent studies have demonstrated a variety of significant biochemical and morphological responses to OA (e.g.Uthicke and Fabricius, 2012;Keul et al., 2013;McIntyre-Wressnig et al., 2013;Doo et al., 2014 and references therein;Knorr et al., 2015;Prazeres et al., 2015).These variations, in part, are rooted in processes involved in test calcification, which varies between orders (Erez, 2003;Knorr, 2014).Less clear are the effects of the chemical changes associated with OA on the stable isotopic composition of LBF.This is problematic because stable isotopic composition of foraminiferal tests (expressed as d 18 O and d 13 C values) are used in many applications to constrain a broad spectrum of paleoenvironmental elements, which include characterizing paleotemperature (McCrea, 1950;Emiliani, 1955), calculating global ice-volume fluctuations (Shackleton and Opdyke, 1973), stratigraphic correlation (Imbrie et al., 1984), interpreting organic C reservoirs (Shackleton, 1977), and reconstructing life processes (Grossman, 1987).In this study, we report on analyses of the tests of Amphistegina gibbosa d'Orbigny, 1839 (Order Rotaliida), subsequently abbreviated as A. gibbosa, and Archaias angulatus Fichtel and Moll, 1798 (Order Miliolida), subsequently abbreviated as A. angulatus, to assess the effects of low pH and higher CO 2 on stable isotopes as proxies of OA and the potential implications in paleoenvironmental reconstruction using these species.We chose culture experiments at pH 7.6 as a target value when [CO 2atm ] reaches 1300 ppm (Stocker et al., 2013;Po ¨rtner et al., 2014).

Background on calcification mechanisms in Amphistegina and Archaias
Representatives of Amphistegina and Archaias were investigated to determine the effects of higher CO 2 and lower pH on light stable isotopes in their calcareous tests in part because members of their respective families are among the most prolific sediment producers (Hallock, 1981;Hallock et al., 1986a;Langer et al., 1997).In addition, these taxa have quite contrasting methods of calcification.Amphistegina gibbosa produces hyaline low-Mg calcite tests and calcifies through biological mediation (de Nooijer et al., 2014), whereas A. angulatus produces porcelaneous, high-Mg calcite tests and calcifies using a less mediated, vacuolization process (de Nooijer et al., 2009).Erez (2003) reported that calcification rates in Amphistegina lobifera increased essentially linearly between pH 7.0 and 8.3, levelling off then declining at pH .9.In contrast, the larger miliolid Amphisorus hemprichii exhibited an exponential increase in calcification rates at pH between 7.0 and 9.5 (Erez, 2003).
Archaias angulatus (Figure 1) is a shallow-water (,30 m) species common throughout the western Atlantic and Caribbean.The test is planispiral and involute in early whorls, and is constructed of imperforate, porcelaneous calcite.Members of this species host the chlorophyte algal endosymbiont Chlamydomonas hedleyi (Lee et al., 1974).The surface of an A. angulatus test is covered with pseudopores that enhance light transmission into the network of chamberlets.The test wall consists of two layers of high-Mg calcites; the outermost test layer consists of a continuous sheet of rhombohedral calcite tiles; the more voluminous, inner layer consists of randomly oriented calcite needles (Lynts and Pfister, 1967;Reid and Macintyre, 1998).Wetmore (1999) documented that chamber formation in A. angulatus occurs through the accretion of new chambers at the end of the test; overgrowth of older chambers may occur, but is an incidental consequence of growth geometry.When building a new chamber, the foraminifer first forms an organic membrane cyst with its reticulopodia, encapsulating and isolating space for new chamber construction.Meanwhile, a calcitic supersaturated solution is concentrated in seawater-filled vesicles within the cytoplasm, primarily the result of photosynthetic pH increases.This solution ultimately precipitates high-Mg calcite rods in the vesicles (Figure 2).The cytoplasm then transports the calcite rods to the cyst, where they are deposited.First, a single layer of calcite rhombohedra is placed along the extent of the new chamber; then, the remainder of the chamber wall is constructed, consisting of an unordered mass of calcite rods.Consistent with Erez's (2003) observation of pH influence of calcification rates in Amphisorus, Knorr et al. (2015) found a 50% reduction in growth rate of A. angulatus at pH 7.6.
Amphistegina gibbosa (Figure 3) is also a common western Atlantic and Caribbean shelf (,100 m) species.Its test is perforate, involute-trochospiral with a hyaline calcite composition.The test is divided into a series of interconnected chambers in which pennate diatom endosymbionts can be housed (Talge and Hallock, 2003).On the basis of the study of A. lobifera, Erez (2003) proposed that vacuoles of seawater, which originate by endocytosis, provide internal pools into which Ca 2+ and CO 2− 3 are concentrated by active transport that exchange Ca 2+ ions for H + ions, thereby raising the pH within the vacuoles (Figure 4).
Amphistegina gibbosa grows using a biologically mediated process of the addition of overlapping new chambers at the end of the test (Hemleben et al., 1986;Lowenstam and Weiner, 1989;Erez, 2003).The foraminifer extends its reticulopods around the test, forming a cytoplasmic substrate for calcification (Lowenstam and Weiner, 1989).The internal high pH pool, derived from both inorganic and metabolic carbon (ter Kuile et al., 1989), supplies Ca   24.82688N, 80.81448W) and from nearby Tennessee Reef (24.7758N, 80.7398W), respectively.Both species were maintained in culture in St. Petersburg, FL, as described in Knorr et al. (2015).Briefly, groups of approximately 20 individuals from both species were photographed and placed into six plankton mesh bags, randomly consigned to six 38-L tanks maintained at target maintenance levels of 248C, salinity of 34.9, and alkalinity of 2325 mmol kg 21 .The three control tanks were maintained at a headspace [CO 2(air) ] average of 412 + 3 ppm and pH T of 8.0, whereas the three experimental tanks were maintained at a headspace [CO 2(air) ] average of 700 + 87 ppm before CO 2 injection and 1200 + 265 ppm after injections with an average pH T of 7.6.
On the basis of typical growth rates determined for A. angulatus (Hallock et al., 1986a) and for A. gibbosa (Hallock et al., 1986b;Talge and Hallock, 1995;Dettmering et al., 1998), samples were grown in control and experimental tanks for 6 weeks, which allowed sufficient time to precipitate new CaCO 3 chambers as well as to asexually produce juvenile specimens.Mesh bags containing cohorts of A. angulatus and A. gibbosa were examined for adults from the original collection and for juvenile specimens that were produced during the experiment.These specimens were removed separately and prepared for isotope ratio mass spectroscopy.To compare the efficacy of selectively removing only the terminal chambers, which had formed under the control or experimental conditions, they were extracted from the tests of the adults (Knorr et al., 2015) and combined to reach the 200 mg target mass needed for analysis.This 200 mg of adult specimens' terminal chambers was designated as the "last chamber" or "terminal chamber" sample.Whole tests from the juvenile specimens  were analysed since they had only experienced conditions within the tank in which they were produced and grew.

Measurement of carbonate chemistry
Target pH in the control (pH 8.0; n ¼ 3) and experimental (pH 7.6; n ¼ 3) tanks were achieved using a continuously monitored, automated pH-controlled CO 2 injection apparatus, which periodically injected CO 2(g) into each tank headspace (Knorr, 2014).CO 2 was mixed with the tank water by a submersible aerator.The injection apparatus pH probes were calibrated 1 -2 times weekly relative to concurrent total-scale pH spectrophotometric measurements.Spectrophotometric analyses of each tank pH T and total alkalinity (A T ) were performed once or twice weekly and water chemistry was adjusted as necessary.The pH T was measured with an Ocean Optics USB4000-UV-VIS fiber optic spectrophotometer and a thymol blue indicator following established methods (Clayton and Byrne, 1993;Dickson et al., 2007).Total alkalinity (A T ) was measured using an open-cell titration with an Ocean Optics USB4000-UV-VIS fiber optic spectrophotometer and a bromocresol purple indicator (Yao and Byrne, 1998).The calibration of A T was performed using Dickson certified reference materials (Dickson et al., 2003).When necessary, reagent-grade NaHCO 3(s) was dissolved in 50 ml of tank water and added to increase A T values back to the target value.A second measurement of A T was made 10 min after the end of the CO 2 injection cycle following the addition of NaHCO 3 , at which time additional dilutions or additions were made as necessary.Previous trials determined that the addition of 100 mg of NaHCO 3 increased A T by 80 mmol kg −1 sw .Kruskal -Wallis testing revealed no significant difference (H 23 ¼ 7.31, p ¼ 0.2) between the A T of the six tanks before and after NaHCO 3 addition.

Stable isotopic analyses
All stable isotopic analyses were completed at the University of South Florida's School of Geosciences Stable Isotope Laboratory on a Thermo Delta V isotope ratio mass spectrometer with the Gasbench II preparation system.Isotopic values for oxygen and carbon (d 18 O and d 13 C values) are reported on the delta scale (d¼ R sample /R standard 2 1) in per mil notation (‰, where R is the ratio of heavy to light isotope for sample and standard).equilibrium conditions, d 18 O values must be converted between VPDB and VSMOW scales.This conversion was done using the following equation: (1) Water samples from Dolphin Cove and tank experiments were collected for stable oxygen isotopic analysis of water and carbon isotopic analysis of dissolved inorganic carbon (DIC).Serum bottles (125 ml) were completely filled, fixed with 50 ml of saturated mercuric chloride, and sealed with greased Teflon crimp caps.Analyses of d 18 O of H 2 O were completed by equilibrating 200 ml of sample with 12 ml headspace of 0.3% CO 2 -He mixture in Labco Exetainers and allowing it to react for 24 h.Analyses of d 13 C of DIC were completed by injecting 1 ml aliquots of sample water into 12 ml vials that were pre-flushed with He and pre-filled with 1 ml of 85% H 3 PO 4 followed by equilibration for 24 h.Juvenile and adult foraminiferal specimens from each cohort were sonicated in 25 ml of distilled, deionized water (DI) for 2 min, rinsed with 500 ml of DI water, dried at 608C, and consolidated until the required analysis mass ( 200 mg) was reached.Analyses of d 18 O and d 13 C of test calcite (n ¼ 30) were completed by flushing 6 ml Labco Exetainers, containing 100 mg of sample, with He gas, adding 2.5 ml of 103% phosphoric acid (H 3 PO 4 ), and equilibrating for 24 h.In each analysis, the relevant isotopic composition (d 18 O or d 13 C values) of the headspace CO 2 was measured after equilibration was attained under controlled temperature conditions, and analyses normalized to internal reference standards.
Analytical precision (2s) was ,0.1‰ for d 18 O and d 13 C values of all water, DIC, and carbonate reference materials used as internal standards.Internal standards were anchored to the VPDB and VSMOW scales using international reference materials (NBS-19 and NBS-18 calcite, and VSMOW2, and SLAP2 water).Results were considered significant at p , 0.05; values are shown as (mean + standard error).

Data analyses
Analysis of variance (ANOVA) with Tukey's pairwise comparisons was used to evaluate significant differences in the mean temperature, salinity, and alkalinity between experimental and control tanks.The general linear model (GLM) was used to evaluate the effect of pH on the oxygen and carbon isotopic ratios in the adult terminal chambers, adult whole tests, and juvenile whole tests of A. angulatus and A. gibbosa, in which the null hypothesis of primary interest was that there was no significant effect of pH on the isotopic values of tests.Using this approach, the factors (in nested order) were species, pH treatment, sample type (i.e.adult terminal chamber, adult whole test, and juvenile whole test), and replicate tanks, all of which were fixed.Data were analysed using the Minitab 17 statistical software and an a level of 0.05 was used in all tests.

Calculations
Unless influenced by vital effects, the d 18 O value of precipitated LBF calcite is highly dependent on that of local [HCO − 3 ] and [CO 2− 3 ], precipitating as the weighted average of the two ions, which are in equilibrium with seawater (Zeebe, 1999).For precipitation of calcite from seawater, the d 18 O value of calcite in oxygen isotopic equilibrium with seawater can be predicted by mass balance: This equation assumes enrichment factors between carbonate species (x) and seawater (1 x−H2O ) of 34.3‰ for HCO − 3 and 18.4‰ for CO 2− 3 [values calculated at a seawater temperature of 198C, as in Zeebe (1999)].HCO − 3 and CO 2− 3 distribution at each pH are calculated using CO2calc (Robbins et al., 2010), with inputs consisting of the mean temperature, salinity, and pH of the culture tanks, and carbonic acid dissociation constants (pK 1 and pK 2 ) from Lueker et al. (2000) and HSO − 4 dissociation constants from Dickson (1990) (Table 1).
Thus, at a seawater d 18 O value of 0, and pH of 7.6: and at pH 8.0: These calculations show that if the pH is reduced from 8.0 to 7.6 (at constant temperature), the d 18 O value will be enriched by 0.8‰.d 13 C values do not follow a similar theoretical relationship, but can be predicted if the isotopic composition of C sources (atmospheric CO 2 , experimental CO 2 , and seawater DIC) are known, using equilibrium carbon isotopic fractionation factors between these sources, assuming carbonate system equilibrium conditions.

Results
Seawater and inorganic carbon parameters for the tanks are summarized in Table 1.As reported in Knorr et al. (2015), there were no statistically significant differences in the measured temperatures, at the 98% confidence level.The main effects of pH treatment, type of sample, and tank replicates on isotopic composition are presented in Tables 3 and 4, where the pH treatment (by sample type and tank) and sample type (by tank) are significantly related to oxygen and carbon isotopic composition for A. gibbosa and oxygen isotopic composition for A. angulatus.As previously indicated by other tests, the replicates are not significantly different.

Experimental vs. control
Adult whole-test analysis of A. angulatus showed little difference between the experimental and control pH on d 18 O values (VPDB; p , 0.62) with a lack of significance for d 13 C values (p .0.9).The d 18 O and d 13 C values for adult terminal chambers of A. angulatus were insignificant between experimental pH and controls (VPDB; Figures 5a and 6a; n ¼ 41, p .0.3 and 1.0, respectively).No significant differences (p .0.99) were observed between experimental (pH 7.6) means and control mean d 18 O (Figure 5b) or d 13 C values (Figure 6b) for either whole tests or adult terminal chamber of A. gibbosa (Figure 6b).Juveniles of both species exhibited significant differences in mean d 18 O values between pH treatments (p , 0.005 for both; Figure 5a and b).In contrast to juvenile A. angulatus, which did not show significant differences for d 13 C values between experimental and control conditions (p ¼ 0.25, Table 4), juvenile    A. gibbosa displayed highly significant difference (p , 0.001) for mean d 13 C values between experimental and control tanks (Figure 6b).

Discussion
Both equilibrium and disequilibrium kinetic effects on isotopes of inorganic carbon species have been shown to contribute to the total observed fractionation of carbon and oxygen isotopes between CO 2(aq) and the anions involved in carbonate precipitation, et al., 1997;Zeebe, 1999).At equilibrium, [CO 2− 3 ] is 18 O-depleted by as much as 16‰ relative to [HCO − 3 ], and [HCO − 3 ] is 18 O-depleted by 24‰ relative to [CO 2(aq) ].Thus, as pH decreases the [CO 2− 3 ] : [HCO − 3 ] ratio, the carbon speciation effect increases both d 18 O and d 13 C values.Carbon isotopic composition is subject to metabolic and abiotic kinetic fractionation as well as relatively small temperaturedependent equilibrium fractionation factors, whereas oxygen isotopic composition primarily reflects large equilibrium fractionation factors and variation in the total fractionation with inorganic carbon speciation (Spero et al., 1997).Rotalids, such as A. gibbosa, store ions from multiple sources in an internal carbon pool for use in calcification (Erez, 2003), while miliolids, such as A. angulatus, precipitate calcite within vesicles containing relatively unaltered seawater (ter Kuile and Erez, 1988).Therefore, distinct speciation effects on stable isotopic composition are predicted to follow these different metabolic processes; the results of this experiment reflect these divergent calcification mechanisms.
Isotopic disequilibrium that arises during precipitation of LBF carbonate tests can be caused by respiration, ontogenetic effects, symbiont photosynthesis, and other vital effects (Wefer and Berger, 1991;Bemis et al., 1998;Rohling and Cooke, 2003).Previously reported stable isotopic values for modern A. angulatus tests from natural environments range between approximately 21.0 and +1.0‰ for d 18 O, and between +1.0 and +4.5‰ for d 13 C (VPDB scale); specimens from the same location can exhibit variations of 1.0‰ (Gross, 1964;Wefer, 1985;Brasier and Green, 1993;Knorr et al., 2015;VPDB scale).Moreover, stable isotopic composition for Amphistegina spp.can vary between approximately 23.5 and 21.0‰ for d 18 O and approximately 22.0 and +2.0‰ for d 13 C (Vinot- Bertouille and Duplessy, 1973;Saraswati et al., 2004;Rollion-Bard et al., 2008;Knorr, 2014;VPDB scale).Intraspecies variations of .2‰are common (Vinot-Bertouille and Duplessy, 1973), and little or no correlation between d 18 O and d 13 C values has been observed (Saraswati et al., 2004).The analytical results on the juvenile specimens presented here agree with previous work on the relationship between carbonate speciation and oxygen isotopic fractionation (Spero et al., 1997;Zeebe, 1999;McConnaughey, 2003).Specifically, analyses of juvenile tests demonstrate the relative enrichment of d 18 O values in the lower pH treatment (7.6) relative to the control (8.0) of both rotalids ( 0.7‰) and miliolids ( 0.9‰), consistent with the 0.8‰ 18 O-enrichment predicted by the mass balance calculation of equilibrium speciation effects (Zeebe, 1999).The agreement between observed and predicted d 18 O values of the juvenile tests indicates that equilibrium fractionation plays the dominant role in determining the d 18 O values, with little or no variation in the biological kinetic 18 O-discrimination involved in the selection of ions for use in calcification.In contrast, neither whole adult tests nor terminal chambers show the expected oxygen isotopic enrichment (Figure 5a and b) after 6 weeks at low pH treatment.

Stable isotopic responses of LBF
Miliolids, including those with algal symbionts such as A. angulatus, precipitate carbon directly from seawater (ter Kuile and Erez, 1988).In contrast, rotalids produce internal carbon pools containing carbon derived from seawater, algal endosymbiont photosynthesis, and metabolic processes (de Nooijer et al., 2014).The invariance of miliolid d 13 C values in different pH treatments can be attributed to the lack of a relationship between pH and the d 13 C of dissolved inorganic carbon species of seawater, and the subsequent direct precipitation of carbonate, with no significant biological isotopic effects.
The significant difference found in the d 13 C of juvenile A. gibbosa from different pH treatments may be the result of a vital effect.Growth in A. gibbosa is stimulated at modestly elevated levels of [CO 2atm ] ( Knorr et al., 2015).Their diatom endosymbionts can benefit from increased concentrations of CO 2 and HCO − 3 , a consequence of the lower seawater pH associated with OA (Engel et al., 2015).The experimental results of the carbon isotopic composition (Figure 6b) indicate that A. gibbosa's internal carbon pool is relatively 13 C-enriched under low pH conditions, a likely consequence of preferential uptake of 12 C for photosynthesis by the diatom symbionts.However, the adult A. gibbosa specimens (both whole test and terminal chamber values) do not show the expected carbon isotopic enrichment (Figure 6a   speciation effects in the surrounding seawater; this speciation is primarily dependent on pH (Pilson, 1998).However, the relationship of seawater pH and test d 13 C values is more nuanced; miliolid d 13 C values are independent of pH, whereas rotalid d 13 C values are influenced by pH.Comparing the results from representatives of two very different foraminiferal lineages, in which biomineralization appears to have independently evolved (Pawlowski et al., 2013;Mikhalivich, 2014), provides insights into how different taxa may respond to changes in the marine carbonate system and confirm using stable oxygen isotopes use in paleoenvironmental reconstruction of paleo pH.Furthermore, the importance of appropriate selection of laboratory specimens becomes apparent when stable isotopic values of adult whole tests, adult terminal chamber, and juveniles are compared.

Adults vs. juvenile specimens
Observed differences in oxygen isotopic speciation between adult (both whole test and terminal chamber) and juvenile tests likely reflect the fact that the adult specimens originated from field samples, so early chambers precipitated in equilibrium with isotopic composition of their harvest localities in the Florida Keys, while juveniles calcified only under experimental conditions (Table 1).At the conclusion of the experiment, preparation of the adult foraminifera for analysis involved dissection of terminal chambers, which were presumed to have been precipitated during the experiment.Since labelling the original chambers was not performed, we cannot be completely confident that the sampling of the adults only included those chambers precipitated during the experiment, despite the use of photographs to document each individual and their growth periodically over 6 weeks, which was then used as a guide for terminal chamber dissection.More likely, however, bias may have been introduced as a function of dissection of the terminal chambers, since precise extraction of individual chambers from the others was impossible.This problem is particularly apparent for A. gibbosa, in which a new layer of calcite is deposited over previously formed ones with each chamber addition.Not only might dissected chambers have pieces of chambers produced earlier in ontogeny still attached, which could bias the isotopic results, but also the deposition of new calcite would alter the signal from the older part of the test.Potential contamination may be more prevalent depending on the original size of the test with more original chambers increasing that potential, subsequent growth rate of chambers onto previously existing chambers during the treatment period, and duration of the experiment.More precise analysis of individual chambers may be feasible with smaller analytical sample sizes (tens of mg), and in particular using new micro-analytical approaches such as secondary ion mass spectrometry (Vetter et al., 2013(Vetter et al., , 2014)).
In our experiments, the difference in d 18 O values of the native seawater and tank experiments was 4.6‰ and the averaging of that was seen when comparing the adults with the juveniles.This indicates that if whole, mature specimens are selected for calcification experiments, the growth of only one or two chambers during the pH treatment is insufficient to dilute the original stable isotopic signals.As a result, our data indicate that isotopic analyses of juvenile specimens produced during an OA experiment are more likely to reveal differences than analyses of adult LBF that originated from field samples or stock cultures.

Conclusions
Can stable isotopes in LBF be used as a proxy for OA?The effects of lowered seawater pH on the stable oxygen and carbon isotopes of two prolific sediment-producing species of foraminifera demonstrated distinct enrichments in stable oxygen isotopes of juveniles produced under low pH conditions.Moreover, isotopic results from the culture experiments using adult specimens, or even dissected "last chambers/terminal chambers" of adult specimens, are likely to be of limited use because of overlapping test construction.
However, stable isotopic analysis of juvenile foraminifera grown under low pH conditions responded as expected, showing an enrichment in both d 18 O and d 13 C values when compared with control values.As CO 2atm approaches 1300, our data suggest that the lower pH associated with OA will more significantly impact the miliolids, such as A. angulatus.Further knowledge of the calcification mechanisms of different species, especially how carbon and oxygen are incorporated during the biomineralization of a foraminiferal test, is critical in interpreting results.
primary source for this internal carbon pool (ter Kuile and Erez, 1988) although the details of the ion acquisition mechanism of the carbon are ambiguous (de Nooijer et al., 2014).The vacuoles containing concentrated Ca 2+ and CO 2−

Figure 5 .
Figure 5. (a and b) Box-and-whisker plots of d 18 O values (VPDB scale) from adult and juvenile specimens of A. angulatus and A. gibbosa cultured at pH 8.0 and 7.6.Boxes show 1st-3rd quartiles with central line at median values; whiskers show the range of values; circles show mean value.
Studies of benthic foraminifera have documented varied isotopic responses to changes in pH and pCO 2 .Amphistegina spp.follow the predicted speciation effect of decreased d 18 O values with increased pH and [CO 2− 3 ] (Rollion-Bard et al., 2008).Similarly, the culture study of the symbiotic planktonic foraminifer Orbulina universa d'Orbigny by Spero et al. (1997) was the first to demonstrate the speciation effect, finding that at that lower pH, the decrease in 18 O-depleted [CO 2− 3 ] and increase in 18 O-enriched [HCO − 3 ], resulted in lower d 18 O (and d 13 C) values in cultured foraminifera.
and b); this result is discussed below.The results of this experiment indicate that the d 18 O values of both rotalid and miliolid LBF are dependent on the carbonate

Figure 6 .
Figure 6.(a and b) Box-and-whisker plots of d 13 C values (VPDB scale) from adult and juvenile specimens of A. angulatus and A. gibbosa cultured at different pH 8.0 and 7.6.Boxes show 1st-3rd quartiles with central line at median values; whiskers show the range of values; circles show mean value.
The standard reference scale used is Vienna Standard Mean Ocean Water (VSMOW) for d 18 O values for water and Vienna Pee Dee Belemnite (VPDB) for d 13 C and d 18 O values for calcite.For comparison of d 18 O of the calcite (on VPDB scale) and water (on VSMOW scale) under

Table 1 .
Carbonate system parameters that were analysed during experiment.Mean isotopic values for adult whole tests, adult terminal chambers, and juvenile whole tests for A. gibbosa and A. angulatus in the control and experimental tanks after the 6-week experiment are displayed in Table 2. Tukey's pairwise comparisons revealed that d 18 O values of adult whole tests were significantly different from adult terminal chambers for both A. gibbosa and A. angulatus (VPDB scale; Uppstrom (1974)2000)0)arameters in treatments during the 6-week experiment.Values for DIC, pCO 2 , and V calcite were calculated using CO2calc (version 1.3)(Robbins et al., 2010)using pH-alkalinity data pair, carbonic acid dissociation constants fromLueker et al. (2000), HSO 4 2 constants fromDickson (1990), and borate dissociation constant ofUppstrom (1974).adjusted p , 0.001), but not for d 13 C values (VPDB scale, p .0.5)

Table 2 .
Oxygen and carbon stable isotopic values for adult (terminal chamber analysed and whole test) and juvenile (whole test) LBF at pH 8.0 and 7.6 (average + s.d.).

Table 3 .
Results of ANOVA GLM of A. gibbosa oxygen and carbon isotopes.
a Sample type refers to juvenile, terminal chamber adult, and whole chamber samples.

Table 4 .
Results of ANOVA GLM of A. angulatus oxygen and carbon isotopes.Source D.F. F-value p-value a Sample type refers to juvenile, terminal chamber adult, and whole chamber samples.