Experimental Design

King scallops Pecten maximus with a shell height of 84.1–103.0 mm (mean = 94.4 mm) were collected by SCUBA from Sea Box SM16, Sound of Mull (GPS 56.6109 N, 6.3905 W), Scotland, in March 2016. Scallops were then transported on ice to the Leibniz Center for Tropical Marine Research (ZMT, Bremen, Germany), where they were acclimated to laboratory conditions (temperature ± SE = 9.0 ± 0.0°C, salinity ± SE = 34.6 ± 0.0 ppt, pH ± SE = 8.1 ± 0.0) for 14 days and subsequently reared for 74 days (April–June 2016) under two temperatures crossed with three pCO₂ regimes, corresponding to present day conditions (temperature ± SE/pCO₂ ± SE = 9.0 ± 0.0°C/362 ± 6 ppm; 12.3 ± 0.0°C/454 ± 12 ppm), the IPCC RCP 6.0 year 2100 scenario (IPCC 2014, temperature ± SE/pCO₂ ± SE = 9.0 ± 0.0°C/860 ± 31 ppm; 12.3 ± 0.0°C/946 ± 42 ppm), and a year 2500 scenario (T ± SE/pCO₂ ± SE = 8.9 ± 0.0°C/2,639 ± 78 ppm; 12.1 ± 0.1°C/2,750 ± 110 ppm). Temperature treatments represent average spring (9°C) and summer (12°C) seawater temperatures at Oban, Scotland (Curry 1982), which is near the scallop collection site. Each scallop was fed 2.5 mL of a microalgae-based commercial shellfish food (Shellfish Diet 1800) diluted in 7.5 mL seawater. Scallops were fed once per day, at which time, the seawater filtration systems were bypassed for 30 min to allow scallops to filter the suspended algae. Five replicate tanks were used per treatment, with each replicate tank housing one scallop. Calcification rates were estimated from the percent change in buoyant weight of the specimens between the beginning and end of the experiment. Microelectrode measurements of scallop EPF pH were obtained over a 7-day period at the end of the experiment.

Measurement and Control of Seawater Carbonate Chemistry

Compressed CO₂-free air and compressed CO₂ were blended using solenoid-valve mass flow controllers to produce gas mixtures formulated at the target pCO₂ conditions. These gas mixtures were then bubbled into six 284-L flow-through experimental seawater systems, each consisting of two seawater reservoirs and five 2-L replicate aquaria. Temperature was controlled with 1650-W aquarium chillers (Aqua Medic Titan 2000). Experimental treatment water was refreshed with natural seawater (obtained from Spitzbergen, Norway, and stored at ZMT) at a rate of 0.6 L/h. This rate was sufficiently slow to allow near-equilibrium to be reached between the mixed gases and the experimental seawaters, yet fast enough to prevent material alteration of the dissolved constituents of the experimental seawaters through calcification or waste excretion. Temperature, pH, and salinity of all replicate tanks (n = 5) were measured three times per week using a combination probe kit (WTW Multi 3430 Set K). Before each use, the pH electrode was calibrated with pH 4.00 and 7.00 NBS buffers traceable to NIST standard reference material at both treatment temperatures, and the conductivity (salinity) probe was calibrated with certified seawater reference material of 33.347 salinity (Dickson CRM batch #154) at both treatment temperatures. Seawater samples for the analysis of DIC and total alkalinity (TA) were collected weekly from each replicate tank, with TA samples stored in 50-mL polypropylene centrifuge tubes and DIC samples stored in 25-mL borosilicate glass vials sealed with rubber septae. After collection, DIC and TA samples were poisoned with 10 and 20 μm, respectively, of saturated mercuric chloride solution and refrigerated until analysis. Total alkalinity was measured by open-cell potentiometric Gran titration (accuracy = ±20 μmol/kg). Dissolved inorganic carbon was measured with a Shimadzu DIC analyzer (accuracy = ±10 μmol/kg). Seawater pCO₂, pH (seawater scale), carbonate ion concentration ([CO₃^2–]), bicarbonate ion concentration ([HCO₃^–]), aqueous CO₂, and calcite saturation state (Ω_c) were calculated from measured TA and DIC with the program CO₂SYS (Lewis & Wallace 1998), using Roy et al. (1993) values for the K₁ and K₂ carbonic acid constants, the Mucci (1983) value for the stoichiometric calcite solubility product, and an atmospheric pressure of 1.015 atm (Table 1).

Measurement of Calcification Rates

Calcification rates were calculated from the change in buoyant weight of the scallop specimens (n = 30) between the beginning and end of the experiment. Buoyant weights were obtained by placing scallops on a platform hung from a bottom loading scale (Mettler Toledo, precision = 0.01 g), with the platform suspended at a fixed depth in seawater of constant temperature and salinity. Scallops were weighed three times to obtain a mean weight. Scallops were reweighed if repeat measurements varied by more than 0.05 g. A standard of known weight was measured every five samples to ensure that the scale was functioning properly (i.e., no instrumental drift). The relationship between the dry and buoyant weights of the scallop shells was empirically defined by plotting dry weight of the shells (determined at the completion of the experiment) against their corresponding buoyant weights at that time (Fig. 1; method previously established for bivalves by Ries et al. 2009). This relationship can be defined for the king scallops from this experiment with the following algorithm:

Figure 1.

Linear relationship between measured buoyant weights and dry weights of the king scallop shells at completion of experiment (y $ 1.537x + 2.500, P < 0.001, R² $ 0.986, SE $ 0.034).

TABLE 1.

Average measured seawater parameters: salinity, temperature, pH on NBS scale (pH_NBS), total alkalinity (TA), and DIC.

continued

where the precision of the scallop buoyant weight as an estimator of their dry weight is equivalent to the SE of the regression (0.034 g). This relationship was then used to convert measured buoyant weights to dry weights, with net calcification rates expressed as the %-change in dry weight of the shell to control for the allometric relationship between scallop shell size (mass) and calcification rate.

Microelectrode Measurements of Scallop EPF pH

Scallop EPF pH was measured during the final week of the experiment, shortly after obtaining the final buoyant weight of each specimen. A 3-mm hole was drilled into the central rib of the right valve of each scallop at a distance from the umbo of approximately 1/3 the shell height. The drilling site was flushed with seawater during drilling to reduce friction and heating. Immediately after drilling, a 1-mm-diameter plastic pipette tip was cut to a length of 5 mm and epoxied into the drilled hole to serve as a port for inserting the microelectrode into the EPF. The open end of the port was sealed with paraffin film to prevent seawater from entering the EPF. Scallops were then returned to their tanks and reacclimated to their treatment conditions for 2 days before EPF pH measurement.

The pH microelectrodes were produced at the Max Planck Institute of Marine Microbiology (MPIMM) using the technique described by de Beer et al. (1997). Briefly, green soda lime glass tubes (Schott model 8516) were exposed to a heated coil and pulled into microcapillary tubes with a target tip diameter of ca. 10 μm. The microcapillary tubes were then silanized to obtain a hydrophobic surface to which the LIX membrane would adhere. Microelectrodes were filled with ca. 300 μm of degassed, filtered (0.2-μm Millipore membrane) electrolyte (300 mM KCl and 50 mM sodium phosphate adjusted to pH 7.0) using a plastic syringe with a 0.1-mm tip. The capillaries were then backfilled with LIX containing a polyvinyl acetate epoxy to prevent leakage of electrolyte. This was performed by dipping the capillary tips in the LIX and applying suction until 100–200 μm of the polyvinyl acetate–containing LIX was drawn into the capillary tip. Micro-capillary tubes were encased in a Pasteur pipette for shielding, with the pulled tip of the microcapillary tube protruding ca. 2 cm beyond this casing. The casing was filled with a 0.3 M KCl solution and connected to the reference electrode with an Ag/ AgCl wire to minimize electrical noise in the system. Microelectrodes with a tip diameter of between 8 and 20 μm were filled with LIX-positive and LIX-negative pH membranes. Microelectrodes were left for 24 h after construction to allow stabilization of the LIX membranes.

All microelectrode equipment (millivolt meter, National Instruments DAQ Pad 6020E, laptop, cables, micromanipulator, VT80 Micos motor arm, laboratory stands, Zeiss Stemi SV6 binocular microscope) was set up adjacent to the experimental tanks to minimize the stress of transporting the scallops. Two reservoirs of seawater, sourced from the corresponding experimental treatment tanks, were established next to the microelectrode system. These reservoirs were bubbled with the corresponding treatment gases and maintained at the treatment temperature using aquarium chillers. The seawater was circulated between the two reservoirs through two 5.4-L flow-through chambers (30 × 12 × 15 cm). All pH microelectrode measurements were performed within these smaller flow-through chambers.

The pH microelectrodes were calibrated before and after EPF measurements using pH 7.00 and 9.00 NBS buffers (to obtain the slope of the pH-mV calibration) traceable to NIST standard reference material, and a Dickson seawater CRM (pH_SW = 7.817; to obtain the y-intercept of the pH-mV calibration). All buffers and standards were maintained at the two treatment temperatures. The microsensor was vertically positioned with a computer-controlled motorized micromanipulator with a precision of 1 μm.

Extrapallial fluid pH was measured in three scallops per experimental treatment. Scallop shells were maintained in a slightly opened position (ca. 5 mm) during measurement of EPF pH by inserting a 5-mm-thick strip of plastic between their shells. This ensured that EPF pH was measured while the scallops were in typical resting position (i.e., feeding with shells open and with scallop tissue in direct contact with circulating seawater), rather than the closed position that they maintain for only short intervals when stressed or threatened—which could cause EPF chemistry to deviate from typical conditions because of build-up of respired CO₂ and other waste products. The pH microelectrode was lowered through the paraffin film, through the microsensor port, and ultimately into the EPF of the scallop. Once the voltage signal stabilized (ca. 5 min), a pH profile of the EPF, port, and seawater adjacent to the scallop was generated by moving the microelectrode out of the port in 50-μm steps (Fig. 2).

Measurement of Scallop Condition Factor (Fulton's K)

All tissue was extracted from the scallop shells at the end of the experiment, blotted dry, and weighed. Shell height was measured to the nearest millimeter with calipers. Scallop condition factor (Fulton's K) was then calculated as follows:

where “TW_WET” is the wet weight of the whole tissue and “SH” is the shell height (distance from umbo to ventral margin). Shell height was cubed according to standard allometric scaling practice (Lucas & Beninger 1985) so that measurements of both tissue and shell varied on a cubic scale.

Figure 2.

Diagram illustrating insertion of the proton-sensitive LIX microelectrode through the shell port into the EPF of a king scallop (right) and the corresponding pH_NBS profile (left) obtained by moving the microelectrode upward from the mantle, through the EPF, shell port, and ultimately into the surrounding seawater.

Data Processing and Statistical Analysis

The voltage output of the microelectrode measurements was converted to pH using the calibration regressions described previously. Fluid pH profiles were then generated by plotting measured pH against distance of the microelectrode from its starting position. The pH profiles, annotated with relative location data recorded during the microelectrode measurements, were then divided into three spatial zones: EPF, shell port, and external seawater (Fig. 1). Mean pH was calculated for each zone.

All statistical analyses were performed using the program R Studio (version 0.99.903). General linear models (GLM) were used to test how calcification rate, EPF pH, and condition factor respond to pCO_2, temperature, and the interaction between these predictors. Various models were run in the order of decreasing complexity, and Akaike Information Criterion (AIC) was used to identify the optimal model (Tables A1–A4). Models were run with and without EPF pH as a predictor of calcification rate because this variable was measured for only a subset of individuals. Owing to the low sample size within treatments (n = 5 for models excluding EPF pH, n = 3 for models including EPF pH), an alpha of 0.1 was used to reduce the potential for type 2 errors (Gotelli & Ellison 2013). When predictor variables exhibited an obvious difference in trend under the two temperature regimes, or when the interaction term P value was less than 0.2, the models were run for each temperature treatment separately to permit interpretation of the main effects at each level of the interaction (Carey 2013). To investigate the seasonal impacts of ocean acidification, a paired t-test was used to test whether the slopes of each response to ocean acidification, and their direction, were different under the two seasonal temperature scenarios.

The assumptions of each GLM were tested by generating diagnostic plots to examine homoscedasticity and normality of residuals. A plot of residuals versus fitted values and a scale location plot were used to examine homoscedasticity, and a quantile–quantile plot was used to examine normality. Normality of residuals was further tested for each model with a Shapiro–Wilk test, and all models met this assumption.

Calcification Rate

Calcification rate was best predicted (pursuant to AIC) by a model that included the effects of pCO₂, temperature, and their interaction (Table A1). Because the response of calcification rate to pCO₂ was markedly different under the two temperature regimes (Fig. 3A), consistent with the low P value of the temperature–pCO₂ interaction term (P < 0.2, see Methods), the effects of pCO₂ on calcification rate were tested separately at 9°C and 12°C (Table 2). At 9°C, calcification rate (Fig. 3A) increased significantly with increasing pCO₂ (Table 2, GLM, P = 0.097), whereas calcification rate did not significantly vary with pCO₂ at 12°C (Table 2, GLM, P = 0.377). The slope of the scallop calcification response to increasing pCO₂ was significantly more positive at 9°C than at 12°C (Table 3, paired t-test, P = 0.051, t = 1.7, df = 26).

Figure 3.

Treatment pCO₂ versus scallop calcification rate (A), EPF pH_NBS (B), and condition factor (C) of the king scallop Pecten maximus at 98C (circle) and 128C (triangle). Vertical error bars represent 1 SE of each response variable and horizontal error bars represent 1 SE of the calculated pCO₂ across replicate tanks. Trend lines in panel B represent measured seawater pH at 98C (solid black line) and 128C (dashed gray line). (A) Increasing pCO₂ significantly increased scallop calcification rate at 98C (P $ 0.097), but did not impact calcification rate at 128C (P $ 0.377). (B) Increasing pCO₂ significantly decreased EPF pH at 98C (P$ 0.001) but did not impact EPF pH at 128C (P $ 0.577). (C) Increasing pCO₂ significantly increased condition factor at 128C (P$ 0.089) but did not impact condition factor at 98C (P$ 0.210).

EPF pH

Extrapallial fluid pH (Fig. 3B) was best predicted (pursuant to AIC) by pCO₂ alone (Table A2). This analysis showed that EPF pH and pCO₂ were negatively correlated (Table 2, GLM, P = 0.048). Because the response of EPF pH to pCO₂ was markedly different under the two temperature regimes (Fig. 3B), the effect of pCO₂ on EPF pH was also tested separately at each temperature. EPF pH was negatively correlated with pCO₂ at 9°C (Table 2, GLM, P = 0.001) but showed no correlation with pCO₂ at 12°C (GLM, P = 0.577). EPF pH variance was also substantially higher at 12°C than at 9°C (9°C, SE = 0.00005; 12°C, SE = 0.00016). The slope of the EPF pH response to increasing pCO₂ was significantly more positive at 12°C than at 9°C (Table 3, paired t-test, P = 0.092, t = –1.37, df = 26)—opposite of the calcification responses of scallops to pCO₂ at the two temperatures.

The correlation between EPF pH and calcification rate was also explored (Figure 4, Table A3). In this model, increasing EPF pH and decreasing pCO₂ were significantly correlated with decreasing calcification rate at 9°C (Table 2, GLM, P = 0.033). At 12°C, calcification rate was best predicted by a model that included the interaction between EPF pH and pCO₂; however, neither term, nor their interaction, was significantly correlated with calcification rate (GLM, pCO₂: P = 0.187; EPF pH: P = 0.397; interaction: P = 0.190).

Scallop Condition Factor

A GLM that included the effects of pCO₂, temperature, and their interactive effects best predicted condition factor (Fig. 3C, Table A4). The effect of pCO₂ on EPF pH was also tested separately at each temperature because the response of condition factor to pCO₂ was markedly different under the two temperature regimes (Fig. 3C). At 9°C, condition factor was not significantly impacted by increasing pCO₂ (GLM; P = 0.210), whereas at 12°C, condition factor significantly increased with increasing pCO₂ (GLM, P = 0.089). The slope of the condition factor response to increasing pCO₂ was significantly more positive at 12°C than at 9°C (Table 3, paired t-test, P = 0.016, t = –2.28, df = 26)—similar to the EPF pH responses to pCO₂ at the two temperatures. Notably, condition factor decreased significantly with increasing temperature (GLM, P = 0.036).

Impact of Ocean Acidification on King Scallop Shell Production

King scallop calcification exhibited relative resilience to the prescribed seawater acidification, with calcification rates remaining constant with increasing pCO₂ under the summer temperature condition and increasing with pCO₂ under the spring temperature condition (Fig. 3). These findings contrast most prior studies that report strong declines in bivalve calcification under future ocean acidification scenarios (e.g., Michaelidis et al. 2005, Melzner et al. 2011, Amaral et al. 2012, Gazeau et al. 2013, Waldbusser et al. 2014), although those investigating the combined stressors of CO₂ and temperature found a weaker response to increasing pCO₂ under elevated temperatures (Waldbusser et al. 2011, Kroeker et al. 2013). Whereas many bivalve species exhibit linear declines in calcification rate in response to declining saturation state at values greater than 1 (Ries et al. 2009, Gazeau et al. 2010, Barton et al. 2012), others only exhibit a decline in calcification rate when the calcium carbonate saturation state of seawater falls below 1 (i.e., undersaturated conditions favoring shell dissolution; Ries et al. 2009, Waldbusser et al. 2014, Waldbusser et al. 2015). It is, therefore, possible that the lack of negative calcification response to ocean acidification in the present study may arise from the calcite saturation states of all experimental treatments remaining above 1 (i.e., supersaturated).

TABLE 2.

Summary of AIC-selected models for subsequent analysis, along with their corresponding statistical parameters.

Few studies have addressed the impacts of ocean acidification on pectinids (Ries et al. 2009, Talmage & Gobler 2009, Andersen et al. 2013, Sanders et al. 2013, Schalkhausser et al. 2013, White et al. 2013). Two prior studies found that the growth of scallop larvae decreased under acidified conditions (Talmage & Gobler 2009, Andersen et al. 2013), contrasting the results of the present study. Differences between larval and adult scallop responses to ocean acidification should not be surprising given their differences in physiology, energy reserves, ability to isolate their EPF, shell polymorph mineralogy, shell function, and calcification mechanisms, as well as inherent allometric differences, such as shell surface area-to-volume ratio (Pörtner & Farrell 2008). A prior study on the adult bay scallop Argopecten irradians revealed that their calcification rates declined linearly with CO₂-induced ocean acidification (Ries et al. 2009). These different outcomes are further evidence of the wide range of responses to ocean acidification exhibited by different taxa, even within a single family (Pectinidae), and highlight the need for further comparative research into the causes of the resilience or vulnerability of marine calcifiers to this critical aspect of global oceanic change.

TABLE 3.

Summary of t-tests investigating differences in the response of king scallop Pecten maximus calcification rate, EPF pH, and condition factor to ocean acidification under spring (9°C) vs. summer (12°C) temperature conditions.

Impact of Ocean Acidification and Temperature on King Scallop Condition Factor

Condition factor is commonly used to assess the quality and profitability of commercial fish and shellfish and also to assess the general physiological performance and stress level of an organism (Lucas & Beninger 1985). In the present study, condition factor of the king scallops varied significantly with pCO₂, temperature, and the interaction between seawater pCO₂ and temperature. Specifically, condition factor was not significantly correlated with pCO₂ at 9°C, was significantly positively correlated with pCO₂ at 12°C, and was significantly negatively correlated with increasing temperature—collectively suggesting that king scallop condition factor is relatively resilient to ocean acidification, but vulnerable to warming.

When organisms are stressed, more of their energy budget is diverted toward basal physiological processes, such as maintenance and repair of tissue (Pörtner 2002), which diverts energy from reproduction, production of new tissue, and build-up of lipid reserves and polysaccharides—thereby decreasing their condition factor. Ocean acidification and warming appear to have complex effects on the condition factor of marine bivalves. Thermal stress has been shown to negatively impact condition factor and energetic reserves in a number of bivalve species (Hiebenthal et al. 2013, Ivanina et al. 2013), whereas the impact of ocean acidification on bivalve condition factor appears more nuanced. Although increased pCO₂ can negatively impact the condition factor of some species (e.g., the eastern oyster Crassostrea virginica, Beniash et al. 2010, Dickinson et al. 2012), other species have exhibited no change in condition factor in response to elevated pCO₂ (e.g., the blue mussel Mytilus edulis, Thomsen & Melzner 2010). Other studies, including the present one on king scallops, have observed an improvement in condition factor in response to increasing pCO₂ (Beniash et al. 2010, Fernández-Reiriz et al. 2011) and mitigation of the impacts of thermal stress under elevated pCO₂, and vice versa (Hiebenthal et al. 2013). Given the ecological and economic importance of bivalve species, further investigation of how temperature and ocean acidification interact to influence their condition factor is needed to predict how bivalves will respond to future global oceanic change.

King Scallop Response to Ocean Acidification Varies with Seasonal Temperature

The combined effects of ocean acidification and temperature (across the seasonal range of a species) on bivalves have received little previous study. Notably, the present study showed that all measured performance parameters (EPF pH, condition factor, and calcification rate) exhibited significantly different responses to ocean acidification under the two temperature regimes (Fig. 3, Table 3). Furthermore, calcification rate and condition factor showed opposite responses to ocean acidification under each temperature regime, with calcification rate exhibiting a more positive response to pCO₂ than condition factor at 9°C, and condition factor exhibiting a more positive response to pCO₂ than calcification rate at 12°C (Fig. 5). This opposition of the calcification rate and condition factor responses to ocean acidification at each of the two temperatures (in favor of calcification at 9°C and in favor of condition factor at 12°C) suggests that king scallops exposed to ocean acidification divert excess energy toward tissue production under summer temperatures (12°C) and toward shell production under spring temperatures (9°C).

Although the cause of this apparent diversion of energy from shell to tissue production (with increasing pCO₂) between spring and summer temperatures cannot be unequivocally deduced from the data at hand, it is possible that it reflects seasonal differences in the physiological demands of king scallops. The diversion of energy toward shell production with increasing pCO₂ under spring temperatures may arise from their need to produce new shell material to replace shell lost to dissolution during the cooler months when external seawater approaches undersaturation with respect to the calcite mineral of scallop shells. Likewise, the apparent diversion of energy toward tissue production with increasing pCO₂ under their summer temperature may reflect their need to accumulate lipid reserves, polysaccharides, and other tissue in support of reproduction in the warmer months—which is also when bivalve shells are less vulnerable to dissolution because of the higher calcite saturation state of their surrounding seawater.

Role of EPF Carbonate Chemistry in King Scallop Calcification Response to Ocean Acidification

Most prior studies exploring links between calcification site pH and calcification response to ocean acidification have investigated scleractinian corals (e.g., Ries 2011a, McCulloch et al. 2012a, Holcomb et al. 2014), with few studies to date focusing on bivalves (Ramesh et al. 2017). In the present study, king scallop EPF pH declined significantly with increasing pCO₂ at 9°C but was uncorrelated with pCO₂ at 12°C (Fig. 3B). Notably, king scallop calcification rate increased with decreasing EPF pH at 9°C and exhibited no change with decreasing EPF pH at 12°C. Ries (2011a) previously published a generalized model of calcifying fluid chemistry showing that a CO₂-induced decline in calcification site pH and concomitant increase in calcification site DIC should reduce calcium carbonate saturation state at the site of calcification (and thus calcification rate) for calcifiers that remove a small number of protons from their calcification site (i.e., “weak proton pumpers”), but should increase saturation state for calcifiers that remove a large number of protons from their calcification site (i.e., “strong proton pumpers”) (see also McCulloch et al. 2012a, Holcomb et al. 2014, Sutton et al. 2018). Although the observed decline in king scallop EPF pH with decreasing seawater pH is consistent with this model, the observed resilience of king scallops to ocean acidification (increasing calcification rate with increasing pCO₂ at 9°C; no change at 12°C) is not consistent with the predictions of the model—assuming that the low EPF pH observed in king scallops (i.e., below seawater pH; Fig. 3B) indicates they are “weak proton pumpers.” Crenshaw (1972) showed that DIC of the EPF within three species of bivalves was approximately double that of seawater, which raises the possibility that scallops may actually be strong proton pumpers to the extent that they substantially elevate EPF pH relative to EPF pH that is in equilibrium with such highly elevated DIC. If king scallops are indeed strong proton pumpers, then their positive correlation between pCO₂ and calcification rate at 9°C, and lack of correlation at 12°C—both indicating relative resilience to ocean acidification—would be consistent with the predictions of the proton pumping model of ocean acidification response (Ries 2011a). Future studies should aim to measure both DIC and pH of the scallop EPF under a range of pCO₂ and temperature scenarios to further explore this hypothesis.

Figure 4.

Relationship between king scallop EPF pH_NBS and net calcification rate for 128C (triangles) and 98C (circles) treatments, across all pCO₂ conditions (grayscale intensity proportional to pCO₂). Calcification rate was inversely correlated with EPF pH at 98C (LM, P$ 0.033) when controlling for the effects of pCO₂ on calcification rate (LM, P $ 0.052) and not significantly correlated with EPF pH at 128C (LM, P $ 0.397) when controlling for the effects of pCO₂ on calcification rate (LM, P$ 0.187).

It should be noted that the positive correlation between pCO₂ and calcification rate, and the inverse correlation between pCO₂ and EPF pH at 9°C, may drive much of the apparent inverse correlation between EPF pH and calcification rate (Fig. 4). Yet, even after controlling for the relationship between pCO₂ and calcification rate at 9°C via the GLM (Table 2), there still exists a significant inverse correlation between EPF pH and calcification rate. Thus, even among replicate specimens within pCO₂ treatments at 9°C, calcification rate was inversely correlated with EPF pH. This observation provides further support for the assertion that some other component(s) of the EPF carbonate system beyond pH—such as DIC (Crenshaw 1972)—must be modified relative to seawater conditions to elevate CaCO₃ saturation state of the EPF in support of shell mineralization. Interspecimen differences in EPF DIC, e.g., could arise from differential rates of respiration, which translate into differential rates of calcification as the scallop elevates EPF pH relative to equilibrium pH conditions associated with increased EPF DIC—thereby driving the inverse correlation between EPF pH and calcification rate within pCO₂ treatments in a manner similar to that discussed previously for CO₂-induced increases in EPF DIC across pCO₂ treatments.

Extrapallial fluid pH of adult bivalves is lower than seawater pH under present-day pCO₂ conditions for all bivalve species whose EPF pH has been measured directly (Crenshaw 1972) or inferred from the boron isotopic compositions of the shells of bivalves (a proxy of EPF pH; Sutton et al. 2018). By contrast, Ramesh et al. (2017) found that the pH of the calcification site larvae of the blue mussel Mytilus edulis was higher than that of seawater. Shell formed during the larval stage serves a protective function and also acts as ballast and attachment cement during larval settlement, rendering it particularly critical for larval survival and recruitment. Furthermore, larval bivalves are known to produce their shells exclusively from the more soluble aragonite polymorph of CaCO₃, whereas adult bivalves produce them from the less soluble calcite polymorph (as is the case with scallops), from the aragonite polymorph, or from a complex multilayered structure of both the calcite and aragonite polymorphs (Lowenstam & Weiner 1989, Marin et al. 2012, Gazeau et al. 2013). It, therefore, makes sense that bivalve larvae, which produce a more soluble polymorph of CaCO₃ and use it not only for protection from predators but also for ballast and attachment during settlement, would direct a greater proportion of their energy toward elevating pH at their site of calcification than their adult counterparts that often produce a less soluble polymorph of calcium carbonate and use their shell primarily for protection from predators (although adult oyster shell is also used for substrate attachment). Ramesh et al. (2017) also found that ocean acidification reduced calcification site pH of larval blue mussels, which is consistent with the results of the present study on adult king scallops.

Figure 5.

Scatterplot showing reversal of calcification rate and condition factor responses to increasing pCO₂ between spring (98C) and summer (128C) temperatures. Data markers corresponding to equivalent pCO₂ treatments have been slightly offset to improve visibility. Vertical bars denote SE.

Although no visible signs of stress (e.g., shell closure, mortality, and reduced feeding) were observed in scallops after implantation of the ports or insertion of the microelectrodes, these are invasive procedures and may have elicited undetected stress responses in the scallops that impacted their ability to regulate EPF pH. The impact of these procedures on the scallops should have been equivalent under all treatment conditions, so should not be responsible for the different trends observed between EPF pH and pCO₂ at the two temperatures. Nevertheless, future studies aimed at quantifying EPF chemistry of bivalves should assess the impact of these procedures on the physiological status of the study organisms by obtaining physiological measurements, such as respiration rate, before and after port implantation and microelectrode insertion.