The influence of food supply on the response of Olympia oyster larvae to ocean acidification

Increases in atmospheric carbon dioxide drive accompanying changes in the marine carbonate system as carbon dioxide (CO2) enters seawater and alters ocean pH (termed “ocean acidification”). However, such changes do not occur in isolation, and other environmental factors have the potential to modulate the consequences of altered ocean chemistry. Given that physiological mechanisms used by organisms to confront acidification can be energetically costly, we explored the potential for food supply to influence the response of Olympia oyster ( Ostrea lurida) larvae to ocean acidification. In laboratory experiments, we reared oyster larvae under a factorial combination of pCO2 and food level. ElevatedpCO2 had negative effects on larval growth, total dry weight, and metamorphic success, but high food availability partially offset these influences. The combination of elevatedpCO2 and low food availability led to the greatest reduction in larval performance. However, the effects of food andpCO2 interacted additively rather than synergistically, indicating that they operated independently. Despite the potential for abundant resources to counteract the consequences of ocean acidification, impacts were never completely negated, suggesting that even under conditions of enhanced primary production and elevated food availability, impacts of ocean acidification may still accrue in some consumers.


Introduction
Global mean atmospheric carbon dioxide (CO 2 ) concentrations have increased by nearly 40 % from 280 parts per million (ppm) during preindustrial times to approximately 390 ppm by the year 2010 (Doney, 2010).The equivalent of nearly 30 % of total human-derived CO 2 emissions have entered the world's oceans, altering ocean chemistry and leading to more acidic seawater with lower carbonate ion [CO 2− 3 ] concentrations (Sabine et al., 2004).Studies have demonstrated that this "ocean acidification" affects a spectrum of organismal traits, including but not limited to growth, reproduction, development, shell strength, skeletal composition, and survival (e.g., Kurihara et al., 2007;Gaylord et al., 2011;LaVigne et al., 2013), across a diversity of taxa (Doney et al., 2009;Kroeker et al., 2010Kroeker et al., , 2013;;Gazeau et al., 2013).There have been fewer studies examining how impacts might be influenced by accompanying shifts in other environmental factors; most have focused on combined effects of altered pH/pCO 2 and temperature (reviewed by Byrne, 2011).However, other environmental factors, such as salinity, dissolved oxygen, and patterns of primary production are also predicted to diverge from present values (Behrenfeld et al., 2006;Harley et al., 2006;Riebesell et al., 2007;Diaz and Rosenberg, 2008;Doney, 2010).
The consequences of such multifaceted changes may be particularly dramatic in nearshore environments.This could be especially true in areas where the upwelling process, which brings waters that are naturally pCO 2 -rich to the surface, already exposes coastal benthic marine invertebrate Published by Copernicus Publications on behalf of the European Geosciences Union.
species to waters low in pH and carbonate concentrations (Feely et al., 2008(Feely et al., , 2010;;Barton et al., 2012;Gruber et al., 2011Gruber et al., , 2012)).Moreover, some models and time series data project an enhancement in the intensity of upwelling, as may be the case for the California Current System of the eastern Pacific (Bakun, 1990;Snyder et al., 2003;García-Reyes and Largier, 2010).Altered water temperature and nutrients, coupled with ocean acidification, may influence phytoplankton productivity (Rose et al., 2009), a food source for pelagic planktotrophic larval stages (Phillips and Gaines, 2002).
In marine systems, the planktonic food environment influences larval condition, and is one of a suite of interacting variables that affects larval and juvenile quality (Phillips, 2002;Boidron-Métairon, 1995;Pace and Manahan, 2007;Giménez, 2010;Vargas et al., 2006Vargas et al., , 2013)).Furthermore, the ability of planktotrophic larvae to ingest a range of food particle sizes and types (e.g., Baldwin and Newell, 1995) could influence larval responses to ocean acidificationinduced shifts in the biochemical and taxonomic composition of phytoplankton.There are mechanisms by which ocean productivity could moderate ocean acidification impacts on planktotrophic larvae.First, ocean acidification may induce changes in phytoplankton growth, inorganic carbon uptake, nitrogen fixation, and elemental ratios (Berge et al., 2010;Reinfelder, 2012), each of which could reduce the nutritional food quality for planktotrophic larvae and other consumers (Riebesell et al., 2007;Rossoll et al., 2012;Reinfelder, 2012).Secondly, ocean acidification could also alter the taxonomic composition of phytoplankton assemblages through a combination of direct and indirect effects (Rose et al., 2009;Caron and Hutchins, 2013).Such shifts may be important because larval performance in oysters and other invertebrates is often sensitive to phytoplankton composition (Thompson et al., 1993;Talmage and Gobler, 2012;Vargas et al., 2013).These shifts would operate in addition to upwelling-induced phytoplankton blooms directly altering seawater chemistry by drawing down total CO 2 through enhanced rates of photosynthesis, shifting carbonate and pH levels back towards conditions favorable for marine invertebrate shell and skeleton-building (Hauri et al., 2009).
Surmounting low pH conditions can incur increased energetic costs, and thus changes in food supply might modulate the effects of ocean acidification on marine taxa across life history stages.Nevertheless, few experiments have been performed to determine whether improved nutrition might ameliorate the negative effects of ocean acidification or whether impaired nutrition might exacerbate these effects.Among the studies that have been conducted on bivalves, Melzner et al. (2011) found greater internal shell dissolution of adult mussels when individuals were exposed to high pCO 2 and low food density.Examining a different life stage of the same mussel species, Thomsen et al. (2013) found that mussels were able to overcome exposure to high pCO 2 during the juvenile phase if there was sufficient food available in the environment.
A remaining unknown is how larval responses to elevated pCO 2 might change as a function of food availability.Ocean acidification has been shown to affect larvae and the larvalto-juvenile transition in bivalves (e.g., Talmage and Gobler, 2009;Parker et al., 2010;Barton et al., 2012), and research indicates that energetic stresses imposed during the larval phase (including those from elevated pCO 2 ) can influence latter life stages (e.g., Rodriguez et al., 1990;Videla et al., 1998;Pechenik, 2006;Hettinger et al., 2012Hettinger et al., , 2013)).Less clear is the degree to which plentiful food might counteract any or all of these impacts.To address this, we exposed larval Olympia oysters (O.lurida) to a factorial combination of food and pCO 2 regimes, tracked growth over the full pelagic duration, and determined percent metamorphosis.

Study system
Olympia oysters (Ostrea lurida) are native to bays and estuaries along the northeast Pacific coast from Alaska to Baja California Sur, Mexico (Baker, 1995).Olympia oysters are a brooding species, and females collect sperm from the water column to fertilize their eggs.Developing trochophore larvae are brooded for approximately 10 days in the mantle cavity before being released by the female as veliger larvae.Veliger larvae develop for 2-3 weeks before they settle and metamorphose into juveniles cemented to hard substrate with one valve of the shell (Baker, 1995).

Larval culture maintenance
Adult Olympia oysters (n = 140) were collected from Tomales Bay, California (38 • 06 58 N, 122 • 51 16 W) in June 2011, transported to Bodega Marine Laboratory (BML), Bodega Bay, California, cleaned of all epiphytes, and distributed evenly among four 100 L culturing cylinders.A similar average size of adult oysters was maintained among cylinders.Every day, adults in each cylinder were fed microalgae (Isochrysis galbana) at a concentration of approximately 200 000 cells mL −1 ; such high food levels encourage larval release.Seawater, filtered to 0.45 µm and held at 18-22 • C, was changed every other day.Seawater in the cylinders was not altered with respect to pCO 2 concentrations, and was bubbled with air to maintain oxygen levels and induce water motion.Adults released larvae in two of the four culturing cylinders after 72 h.Within 12 h of their release, larvae were distributed by pipette to 4.5 L glass jars on day 1 of the experiment (n = 1000 larvae per jar), for subsequent rearing through the full pelagic larval period.Each jar held 2 L of filtered seawater at the appropriate pCO 2 level, resulting in initial rearing concentrations of 1 larva per 2 mL of seawater.
In subsequent experimental trials, we employed a target elevated seawater pCO 2 concentration of 1000 µatm (∼ 7.76 pH T ).This level of pCO 2 is found currently in upwelled waters off the northern California coast (Feely et al., 2008), and these upwelled waters can enter Tomales Bay during the time when oyster larvae are in the water column, bringing low pH waters as far as 8-10 km into the bay (A.D. Russell, unpublished data).The target control seawater pCO 2 concentration was 500 µatm (∼ 8.05 pH T ), approximating an average seawater pCO 2 level that occurs presently during the summer months in Tomales Bay (A.D. Russell, unpublished data).All seawater used during larval rearing was pre-adjusted, prior to addition of larvae, to the appropriate pCO 2 concentrations in 20 L carboys by bubbling filtered seawater for 2-3 days with NIST-traceable CO 2 air mixtures (carboy water).To minimize net CO 2 exchange across the seawater surfaces in the jars and help maintain seawater pCO 2 concentrations at target levels within the jars, the same CO 2 gas mixtures used to maintain carboy water were pumped continuously into sealed air spaces above the free surfaces of the seawater in the culture jars (hereafter "headspaces").Five replicate jars per pCO 2 concentration shared a given headspace.Carboys and jars were held in seawater tables maintained at 20 • C (±0.2 • C), within the natural temperature range in our study region during summer months when Olympia oyster larvae are in the water column (Smith and Hollibaugh, 1997).Additional details of the larval culturing apparatus can be found in Hettinger et al. (2012).
Every other day, 90 % of the seawater in each jar (jar water) was removed using reverse-filtration through 125 µm mesh and replaced with carboy water at the target pCO 2 concentration.Immediately following each water change, microalgae (I.galbana) were added to each jar to generate final densities in each jar of: 100 000, 50 000, or 10 000 cells mL −1 (n = 2 pCO 2 levels × 3 food levels × 5 replicate jars = 30 jars).The three food levels were the equivalent of initial daily algal cell to larva ratios of 50, 25, and 5 cells per larva per day.The highest level is known to encourage rapid growth and high survival of larval and juvenile oysters in culture (Strathmann, 1987).In our study region during the time when Olympia oyster larvae are in the water column, field measurements suggest chlorophyll a concentrations of 10-15 mg m −3 arise commonly, the lower limit of which approximates our high food level (100 000 cells mL −1 ; ∼ 10 mg m −3 ) (Kimbro et al., 2009;B. S. Cheng, unpublished data).
To take into account that elevated pCO 2 concentrations and the algal growth phase can induce shifts in phytoplankton biochemical composition, with possible consequences for consumer nutrition (e.g., Engel et al., 2008;Bellerby et al., 2008;Reinfelder, 2012), we cultured I. galbana under control pCO 2 conditions and added algae to our larval cultures when algal cultures were at peak densities.Examination of a subsample of algal cells and cell counts were performed every other day under the microscope prior to preparing algae to feed larvae.Algal cultures were not used if they ap-peared to have many dead cells or any contamination by other microorganisms, an indication that cultures were in decline.Presumably, the majority of I. galbana cells were consumed before any potential effect of elevated pCO 2 significantly influenced their quality.I. galbana remaining in larval cultures were removed during water changes every other day.
A set of modified glass jars (substrate jars) was constructed for ease of quantifying larval settlement and metamorphosis by removing the bases from a new set of jars, replacing the bases with grey PVC plates (5 mm thick) attached with aquarium-safe silicone, and conditioning the new jar bases in filtered seawater for 2 days.Each PVC plate was roughened with sandpaper to encourage settlement, and a 1 cm 2 grid was drawn onto each base to assist with quantifying settlement and metamorphosis.On day 9 of the experiment, prior to the commencement of settlement and metamorphosis, larvae were transferred into substrate jars.Following metamorphosis, the PVC plates were removed from the jars and percent metamorphosis was quantified (see Sect. 2.6 for details).

Seawater chemistry during larval culturing
Jar water (previously containing larvae and microalgae) and carboy water (no larvae or microalgae) were sampled for total alkalinity (TA) and dissolved inorganic carbon (DIC) every other day when a water change was performed.Seawater pH and temperature were measured using a potentiometric pH/temperature meter (Accumet Excel XL60).Raw pH readings (mV) were calibrated using two seawater buffers (2amino-2-hydroxymethyl-1,3-propanediol ("TRIS") and 2aminopyridine/HCl ("AMP") in synthetic seawater).These buffers were made in-house following Dickson et al. (2007) and checked against a certified TRIS buffer (A.Dickson, Scripps Institute of Oceanography, La Jolla, California).pH was monitored primarily as a real-time indicator of changes in the carbonate system.Salinity was determined using a YSI Professional Plus multiparameter instrument with a conductivity probe calibrated in a NIST-traceable conductivity calibration solution (YSI, Yellow Springs, OH).TA was measured using automated Gran titration with duplicates (Metrohm 809), and standardized using certified reference material (CRM) from A. Dickson.The offset between measured and certified TA values for Dickson CRM (Batches 104, 107, 111) was −0.01 (±2.81) meq kg −1 (n = 91).A subset of samples was analyzed for DIC at the University of Georgia's infrared CO 2 analysis facility (Cai and Wang, 1998).Other carbonate system parameters (calcite and aragonite saturation states: calcite , aragonite and seawater pCO 2 ) were calculated using the carbonate system analysis software CO2SYS (Lewis and Wallace, 1998) employing DIC and TA as the primary input variables, with equilibrium constants K1 and K2 taken from Mehrbach et al. (1973) refit by Dickson and Millero (1987), and KSO 4 from Dickson (1990).

Sampling of larval growth
Larval oysters in the culture jars were sampled at key time points to quantify growth through the larval duration.On day 1, prior to their placement in the culture jars, larvae were collected haphazardly by pipette (n = 20), placed on a 125 µm Nitex plankton filter, rinsed twice with distilled water, with all excess distilled water wicked immediately through the back of the filter, and left to dry at room temperature for 24 h.Larvae were photographed individually under a dissecting microscope (Leica M125 with DC290 camera) for analysis using ImageJ software (ver.1.37, National Institutes of Health) to determine the initial projected area of the shell.Larvae were sampled and their projected areas were determined similarly on day 5, 9, and 11 post-larval release (n = 10 larvae per jar at each time point).Larval growth was estimated as the change in projected shell area since larval release (day 1).

Sampling of larval total dry weight
Total dry weight (body plus shell, µg) of the oysters was also determined on days 5, 9, and 11 post larval release.Larvae were collected haphazardly by pipette from each jar (n = 25), placed on a 125 µm Nitex plankton filter, and were rinsed twice and dried as described above.Total dry weights of larvae sampled were determined by transferring larvae individually to aluminum vessels pre-ashed at 500 • C for 3 h, drying the larval sample at 50 • C for > 24 h, and weighing on a microbalance (Sartorius Ultramicro, Goettingen, Germany).Larvae were dried in their vessels a second time, reweighed to verify their weights, and combusted at 460 • C for 4 h in a muffle furnace (Thermo Scientific FB1415M) to remove all organic matter (see also Gaylord et al., 2011).Ash-free dry tissue weights were determined from the weight difference before and after combusting.

Sampling of percent metamorphosis
Settlement of larvae and metamorphosis into benthic juveniles was assessed daily starting on day 11, two days after larval transfers were made into the substrate jars.When < 5 % of larvae remained swimming (used as our assay point for settlement) in each pCO 2 treatment, the bases of each substrate jar were removed, and the proportion of metamorphosed individuals was determined by first subtracting dead and non-metamorphosed larvae (e.g., pediveligers) from the total initial number of larvae, then dividing by the total initial number of larvae (1000).Any metamorphosed individuals on the walls of the glass jars were also included in the proportion of metamorphosed individuals.

Statistical analyses
Larval growth, total dry weight, and percent metamorphosis were analyzed using separate, partly nested ANOVAs.
The measurements we took on individual larva were subsamples, thus we considered the jars to be our sampling unit, and analyzed mean responses from each jar using ANOVA to avoid pseudoreplication.We employed a split-plot design, where pCO 2 level was the whole-plot factor, food level was the sub-plot factor, and headspaces were nested within pCO 2 level (Underwood, 1997).Any biological processes occurring within each jar and any resultant effect on seawater chemistry within each jar operated independent of the headspace.However, because the jars were nested within the headspaces, we included the term "headspace[pCO 2 ]" in our design as a test for any potential differences.The water chemistry data were analyzed using this same ANOVA structure, and separate ANOVAs were conducted for pH and TA.For each jar, all measurements of jar water and carboy water were averaged across the experiment, and an ANOVA was conducted on these estimates of the average conditions in each jar.In all analyses, data fulfilled assumptions of normality and homogeneity of variance, tested using Shapiro-Wilks' and Bartlett's test, respectively, and thus untransformed data were used.The statistical software JMP (ver.8.0.1, Statistical Analysis Software) was used for all analyses.

Seawater chemistry during larval culturing
pH values in the pCO 2 treatments differed from one another (ANOVA, pCO 2 , F 1,12 = 22 850.5, p < 0.0001; Table 1).pH of jars within headspaces assigned to the same treatment also varied slightly (ANOVA, headspace [pCO 2 ], F 4,12 = 6.66, p < 0.005).Specifically, mean pH in the jars from headspace 1 in the 500 µatm pCO 2 treatment differed from headspaces 2 and 3 by 0.009 pH unit.pH levels across the three headspaces in the 1000 µatm pCO 2 level did not differ.pH values also differed as a function of food level (ANOVA, food, F 2,12 = 33.82,p < 0.0001; Table 1).pH units decreased approximately 0.01 unit between each of the subsequent food levels in both pCO 2 levels (i.e., the high food level had a pH unit ∼ 0.01 lower than the medium food level and the medium food level had a pH unit ∼ 0.01 lower than the low food level).This pattern of decreasing pH with increasing food levels (i.e., algal concentrations) was likely from increased respiration in treatments with higher concentrations of algae, given the low-light conditions that characterized our cultures.Total alkalinity did not differ between the pCO 2 treatments (ANOVA, pCO 2 , F 1,12 = 0.38, p = 0.547), among food levels (ANOVA, food, F 2,12 = 0.41, p = 0.675) (Table 1), or among the headspaces assigned to the same treatments (ANOVA, headspace[pCO 2 ], F 4,12 = 0.30, p = 0.871).DIC samples were used simply to overconstrain the carbonate system and were not analyzed statistically.

Larval growth
Larval growth on day 5 post-larval release was reduced in low compared to medium food levels (ANOVA, p = 0.039), but did not differ between the two pCO 2 concentrations (ANOVA, p = 0.092; Table S1, Fig. 1a).Similar patterns were exhibited on day 9; however, at this time point, growth was reduced in low compared to medium and high food levels (ANOVA, food, p = 0.002; pCO 2 , p = 0.175; Table S2, Fig. 1b).At day 11, growth was reduced in low food compared to medium and high food levels (ANOVA, p = 0.048) and also in high compared to control pCO 2 conditions (ANOVA, p = 0.047; Table S3, Fig. 1c).Indeed, larval growth was 9 %, 5%, and 14% higher in control compared to high pCO 2 conditions, for the high, medium, and low food treatments, respectively.The effects of food on larval growth were consistent regardless of pCO 2 concentration (ANOVA, pCO 2 × food, Day 5, p = 0.778; Day 9, p = 0.962; Day 11, p = 0.720; Table S1-S3).Larval growth did not vary between headspaces assigned to the same treatment (Table S1-S3).

Larval total dry weight
On day 5 post-larval release, total dry weight was reduced in low compared to medium food levels (ANOVA, p = 0.001) and was also reduced in elevated compared to control pCO 2 conditions (ANOVA, p = 0.002; Table S4, Fig. 2a).On day 9, low food decreased total dry weight compared to medium and high food levels (ANOVA, p = 0.003), but pCO 2 concentration did not have an effect (ANOVA, p = 0.206; Table S5, Fig. 2b).The same patterns were exhibited on day 11 for low food (ANOVA, p < 0.0001) and pCO 2 concentration (ANOVA, p = 0.089; Table S6, Fig. 2c).The effect of pCO 2 on total dry weight was dependent on food level on day 5 (ANOVA, pCO 2 × food, p = 0.001), but not day 9 (ANOVA, p = 0.448) or day 11 (ANOVA, p = 0.782; Table S4-S6).
The significant interaction on day 5 is likely from the particularly high weights in the ambient pCO 2 /medium food level treatment.

Percent metamorphosis
Settlement occurred on day 16 in the high and medium food treatments in both pCO 2 concentrations, on day 19 in the low food/control pCO 2 treatment, and on day 22 in the low food/elevated pCO 2 treatment.On average, across all food treatments, percent metamorphosis was 70 % higher in control compared to high pCO 2 concentrations (ANOVA, p < 0.0001; Table S7, Fig. 3).There was also a significant effect of food level on percent metamorphosis, which was reduced in the low compared to the medium and high food levels (ANOVA, p = 0.01; Table S7, Fig. 3).Percent metamorphosis was 40 % higher in the control pCO 2 /high food compared to the control pCO 2 /low food treatments, and 54 % higher for the high pCO 2 /high food compared to the high pCO 2 /low food treatments.The effect of pCO 2 level on settlement was consistent across the three food levels (ANOVA, pCO 2 × food, p = 0.226; Table S7).

Discussion
Ocean acidification-induced reductions in growth and calcification have been reported across all life stages of bivalve species (e.g., Ries et al., 2009;Miller et al., 2009;Waldbusser et al., 2010;Gaylord et al., 2011;Hettinger et al., 2012;Gazeau et al., 2013).Several mechanisms are thought to drive such responses including altered metabolic rates and modified energy budget allocation (Pörtner et al., 2004;Michaelidis et al., 2005;Wood et al., 2008;Melzner et al., 2011;Waldbusser et al., 2013).However, studies explicitly testing whether increased food supply might counteract the negative effects of elevated pCO 2 focus primarily on adult or www.biogeosciences.net/10/6629/2013/Biogeosciences, 10, 6629-6638, 2013 juvenile life stages (e.g., Holcomb et al., 2010 (adult   are among the most productive regions in the world's oceans, and their response to ocean acidification is therefore of considerable interest.As discussed above, strong upwelling brings nutrient and pCO 2 -rich water to shallow nearshore environments (Feely et al., 2008;Kudela et al., 2008).Plentiful nutrients fuel rapid population growth of phytoplankton, which operate as a food source for larvae while simultaneously drawing down pCO 2 concentrations (Kudela et al., 2008).The timescale of the draw down can also be important: in the initial stages of upwelling, phytoplankton communities may still be in the early stages of population growth, and may not have yet reached high abundance.Under such conditions,  low pH may accompany low food supply for larvae, whereas in days hence, the opposite pattern may arise.Thus, during the strong upwelling season along the northern California coast, which typically occurs from May-August, limited food might often accompany high pCO 2 /low pH conditions, and vice versa.Moreover, upwelled waters have been documented to enter the mouth of bays and estuaries, including those in Oregon and California (e.g., Smith and Hollibaugh, 1997;Barton et al., 2012;A. D. Russell, unpublished data).In these instances, even species such as oysters, whose larvae spend much of their planktonic existence within restricted bodies of water, can experience variable combinations of food and pH.
Whether due to temporally varying upwelling phenomena or other processes, many organisms experience temporary periods of stress and/or reduced nutrition during early development with consequent effects on succeeding performance and life history trajectories (Phillips and Gaines, 2002).Our study revealed that low food supply and elevated pCO 2 can each lead to decreased oyster larval growth, weight, and percent metamorphosis.However, the effect of elevated pCO 2 concentrations on each response did not depend upon the food supply.Conversely, we found that a benign food environment (i.e., one in which food is not limiting) can partially ameliorate the negative effects of larval exposure to high pCO 2 conditions.Melzner et al. (2011) found similar results of high pCO 2 and reduced food levels on adult mussels.In this case, shell growth was lower in the high pCO 2 and low food treatments, but the effect of pCO 2 on shell growth did not depend upon the food level.Thomsen et al. (2013) inves-tigated the effects of food supply and ocean acidification on an earlier life stage of the same mussel species, and juvenile mussels were able to overcome exposure to high pCO 2 in high resource environments.The response of juvenile mussels to these interacting stressors could conceivably change if individuals were exposed to varying resources and high pCO 2 during the larval stage.
Previous work demonstrated that larval carry-over effects (also referred to as latent effects; Pechenik, 2006) resulted in a magnified, negative response in juvenile Olympia oysters, and effects persisted well into juvenile life (Hettinger et al., 2012(Hettinger et al., , 2013)).A similar trend was exhibited in this study with high pCO 2 having a much larger effect on percent metamorphosis (70 %) than larval growth (5-14 %).The detected reduction in percent metamorphosis might have resulted from larvae starting the energetically expensive process of metamorphosis with depleted energy (i.e., protein and lipid) reserves.Carry-over effects often arise through energetic deficits accrued in prior life history phases.By the completion of the larval phase, oysters exposed to acidified conditions may reach metamorphosis in an energy-depleted state.Many life processes incur metabolic costs during the larval life phase, and because metamorphosis is also an energetically demanding process (e.g., Videla et al., 1998), any energetic deficits accumulated during the larval phase could be magnified in subsequent life phases.The magnitude of carry-over effects might depend in part on the mode of larval development, especially dependence on exogenous versus endogenous food supply.Olympia oysters are a brooding species that rely on external resources during larval development.Planktotrophic species such as this might be more strongly impacted by changes in phytoplankton abundance and quality compared to lecithotrophic species that are provided with a source of nutrition upon larval release.However, since the maternal environment has been shown to influence offspring performance (Marshall, 2008), any stress from high pCO 2 conditions or altered food supply might equally impact planktotrophic and lecithotrophic species directly or indirectly through maternal effects.Early life stages often act as population bottlenecks (Gosselin and Qian, 1997); thus, adult demographics can be influenced by larval carry-over effects that persist into juvenile and adult life.
Our results suggest that Olympia oyster larvae do not demonstrate the ability to fully counteract exposure to elevated pCO 2 conditions in high food environments.Furthermore, low food environments and elevated pCO 2 conditions both lead to reduced larval growth and total weights, but these stressors operate additively rather than synergistically.This study shows that even when food supply is abundant, larvae exposed to elevated pCO 2 are smaller and weigh less than larvae in ambient conditions.This difference in larval size and weight will likely influence juvenile growth through larval carry-over effects, and could have significant implications for adult populations.

Fig. 1 .
Fig. 1.Effect of pCO 2 and food supply on larval growth in Olympia oysters (Ostrea lurida).Data are shell growth on (A) day 5, (B) day 9, and (C) day 11 post-larval release in control (black bars)and elevated (grey bars) pCO 2 levels.Food levels in each treatment relate to the following concentrations: 100 000 cells mL −1 (High), 50 000 cells mL −1 (Med), 10 000 cells mL −1 (Low), and refer to the density of Isochrysis galbana in each jar per treatment.Values are means across all jar replicates of a given treatment + 1 SE.Shared letters above bars indicate food levels that did not differ significantly (Tukey HSD, p > 0.05).

Fig. 2 .
Fig. 2. Effect of pCO 2 and food supply on total dry weight in Olympia oyster larvae (Ostrea lurida).Data are on (A) day 5, (B) day 9, and (C) day 11 post-larval release in control (black bars) and elevated (grey bars) pCO 2 levels.Food levels in each treatment relate to the following concentrations: 100 000 cells mL −1 (High), 50 000 cells mL −1 (Med), 10 000 cells mL −1 (Low), and refer to the density of Isochrysis galbana in each jar per treatment.Values are means across all jar replicates of a given treatment + 1 SE.Shared letters above bars indicate food levels that did not differ significantly (Tukey HSD, p > 0.05).

g. 3 .
Effect of pCO 2 and food supply on percent metamorphosis in Olympia oysters (Ostrea ida).Data are the percent of larvae that metamorphosed and settled in control (black bars) and vated (grey bars) pCO 2 levels.Food levels in each treatment relate to the following ncentrations: 100,000 cells/mL (High), 50,000 cells/mL (Med), 10,000 cells/mL (Low), and er to the density of Isochrysis galbana in each jar per treatment.Values are averaged across jar replicates of each larval pCO 2 treatment + 1 SE.Shared letters above bars indicate food els that did not differ significantly (Tukey HSD, p > 0.05).

Fig. 3 .
Fig.3.Effect of pCO 2 and food supply on percent metamorphosis in Olympia oysters (Ostrea lurida).Data are the percent of larvae that metamorphosed and settled in control (black bars) and elevated (grey bars) pCO 2 levels.Food levels in each treatment relate to the following concentrations: 100 000 cells mL −1 (High), 50 000 cells mL −1 (Med), 10 000 cells mL −1 (Low), and refer to the density of Isochrysis galbana in each jar per treatment.Values are averaged across all jar replicates of each larval pCO 2 treatment + 1 SE.Shared letters above bars indicate food levels that did not differ significantly (Tukey HSD, p > 0.05).

Table 1 .
A table of seawater properties of the two seawater pCO 2 regimes during the culturing experiment (±s.d.computed from averages of the jar replicates).The number of samples for TA and DIC are indicated in parentheses.Temperature and salinity (mean ± s.d.) were 19.4 ± 0.05 • C and 33.4 ± 0.46, respectively.TA, total alkalinity; DIC, dissolved inorganic carbon.