Effect of increased p CO 2 level on early shell development in great scallop ( Pecten maximus Lamarck ) larvae

As a result of high anthropogenic CO 2 emissions, the concentration of CO 2 in the oceans has increased, causing a decrease in pH, known as ocean acidification (OA). Numerous studies have shown negative effects on marine invertebrates, and also that the early life stages are the most sensitive to OA. We studied the effects of OA on embryos and unfed larvae of the great scallop ( Pecten maximus Lamarck), at pCO2 levels of 469 (ambient), 807, 1164, and 1599 μatm until seven days after fertilization. To our knowledge, this is the first study on OA effects on larvae of this species. A drop in pCO2 level the first 12 h was observed in the elevatedpCO2 groups due to a discontinuation in water flow to avoid escape of embryos. When the flow was restarted, pCO2 level stabilized and was significantly different between all groups. OA affected both survival and shell growth negatively after seven days. Survival was reduced from 45 % in the ambient group to 12 % in the highest pCO2 group. Shell length and height were reduced by 8 and 15 %, respectively, whenpCO2 increased from ambient to 1599 μatm. Development of normal hinges was negatively affected by elevated pCO2 levels in both trochophore larvae after two days and veliger larvae after seven days. After seven days, deformities in the shell hinge were more connected to elevated pCO2 levels than deformities in the shell edge. Embryos stained with calcein showed fluorescence in the newly formed shell area, indicating calcification of the shell at the early trochophore stage between one and two days after fertilization. Our results show that P. maximusembryos and early larvae may be negatively affected by elevated pCO2 levels within the range of what is projected towards year 2250, although the initial drop inpCO2 level may have overestimated the effect of the highestpCO2 levels. Future work should focus on long-term effects on this species from hatching, throughout the larval stages, and further into the juvenile and adult stages.


Introduction
The increase of anthropogenic CO 2 emissions since the industrial revolution has lead to an increase of carbon dioxide (CO 2 ) concentration and a decrease in pH in the oceans termed as ocean acidification (OA).Over the next 50-100 yr the CO 2 values are projected to reach 700 ppm (Zondervan et al., 2001;IPCC, 2001).Caldeira and Wickett (2003) showed through their geochemical model that with no reduction in fossil fuel emission of CO 2 into the atmosphere, there may be a rapid decline in pH in surface ocean waters over the next 100 yr of as much as 0.4 units (a doubling of H + -ion concentration).And within less than 250 yr the ocean surface pH may drop as much as 0.7 units, which may be the lowest pH value experienced during the last 300 million years (Caldeira and Wickett, 2003).
Depending on their distribution and habitats, marine organisms are exposed to various levels of pH fluctuations.In general, deep-sea species live in a more stable environment compared to shallow-living species that experience both seasonal and daily fluctuations in physio-chemical water parameters.In areas with natural high CO 2 supply (upwelling of deep-sea CO 2 -rich water and volcanic CO 2 vents) marine organisms are exposed to seawater CO 2 values as high as 2000 ppm, corresponding to a pH of 7.4-7.5 (Hall-Spencer et al., 2008;Thomsen et al., 2010).Some organisms can experience even lower pH values in nature, as shown for the oyster Ostrea chilensis (Chaparro et al., 2009).During valve closure Published by Copernicus Publications on behalf of the European Geosciences Union.and isolation of the brood chamber, the brooded veligers were exposed to pH values as low as 7.0.
The numerous reports on effects of OA are slowly building up a picture of how marine coastal ecosystems may be affected (biodiversity, functioning, and service) in the future (Raven et al., 2005).It is crucial to gain more knowledge about the effects on different marine organisms in order to get realistic projections of future changes of the marine food web.A negative effect on early life stages may not only be detrimental to recruitment and endanger the species survival but also result in economic loss due to a collapse in global shellfish aquaculture production (Gazeau et al., 2010).Together with mussels and oysters, several scallop species are economically important in global aquaculture production (Bourne, 2000;FAO, 2010) in addition to being important calcifiers in the marine ecosystem.
The great scallop Pecten maximus Lamarck is distributed from the Iberian Peninsula in the south up to Lofoten in Norway in the north, and is found in highest abundances at 5-50 m depth in shell sand areas (Strand and Parsons, 2006).The larvae are pelagic for about 15-32 days at temperatures of 15-18 • C (Comely, 1972;Le Pennec et al., 2003).The great scallop is a highly valued sea food product, and has been cultured for more than 25 yr in Norway.The extended research efforts during the same period make it an excellent candidate for OA studies (Andersen et al., 2011).To our knowledge, only the study of Schalkhausser et al. (2012) has shown the effect of OA on P. maximus.They found a negative effect of OA on adult P. maximus shell-clapping performance and aerobic scope.However, no studies have focused on the effect of OA on P. maximus larvae.Studies on other scallop species as Argopecten irradians, Placopecten magellanicus and Chlamys nobilis show a negative effect of increased pH in most life stages, including fertilization, embryos, larvae, and juveniles (Desrosiers et al., 1996;Talmage and Gobler, 2009, 2010, 2011;Liu and He, 2012).
Most OA studies on mollusc larvae focus on effects on shell formation (Kurihara et al., 2007(Kurihara et al., , 2008;;Kurihara, 2008;Ericson et al., 2010;Waldbusser et al., 2010;Byrne, 2012;Gaylord et al., 2011;Gazeau et al., 2011).The shell that starts to evolve at the trochophore stage consists of proteins (Bellioli et al., 1993;Casse et al., 1998).Onset of calcification in the initial organic larval shell was detected during the trochophore larval stage in Mercenaria mercenaria and Crassostrea gigas larvae (Weiss et al., 2002).Casse et al. (1998) did not detect calcification in P. maximus trochophores, and according to Cragg (2006) it is likely that there is initially little or no calcification in pectinid larvae, but that the calcification starts within less than a day after the initial shell formation.If calcification is the actual process being affected, effects on the shell formation should be detected only after the stage when the calcification starts.The onset of calcification can be determined by the use of an epifluorescent dye, calcein, which binds to calcium in metabolically active individuals (Körbl and Vydra, 1958;Day et al., 1995;Lucas et al., 2008;Chaparro et al., 2009).
The aim of the present study was to report the effect of elevated CO 2 levels on scallop P. maximus Lamarck larvae survival, shell size, and shell development from fertilized eggs to 7-day-old veligers.The exact initiation of calcification in P. maximus larvae is still unknown, and thus a pilot study using calcein staining as a method to elucidate this question was also conducted.

Material and methods
Local broodstock were collected in January 2012 from the outer Hardangerfjorden, located on the south-western coast of Norway.Ten scallops were conditioned for eight weeks at 12.0-12.5• C and fed 13-15 cells µL −1 of a standard diet containing Isochrysis galbana (Tahitian), Pavlova lutheri and Chaetoceros mulleri (Andersen et al., 2011).Spawning was induced on 21 March by increasing the temperature to 16-17 • C. Pecten maximus is a simultaneous hermaphrodite, and some sperm will remain in the genital tract, resulting in self-fertilization (Mackie, 1984).Self-fertilization rate was checked before cross-fertilization, and egg batches with less than 10 % self-fertilization were used in the experiment.Eggs from one individual were cross-fertilized with sperm from three other individuals.Egg : sperm ratio was Table 1.Measured and calculated water parameters given as mean ± sd for four different pCO 2 groups (µatm).Carbon chemistry values were computed based on daily measurements (0-6 days) of pH NBS in all replicates (n = 4, * n = 3), means of hourly temperature measurements in three tanks (n = 468), salinity (daily), and total alkalinity (11 analyses from 3 dates) in seawater running into the lab.Coefficient of variation (CoV) range (sd as % of mean) is shown for variation in CO 2 (ppm) between days (d) for the replicates, and between replicates (r) for all days.1).The pH levels were used based on the predicted drop of 0.5 units from present day to year 2250 (IPCC, 2001).Mean temperature based on hourly measurements in three replicates (± sd) was 15.6 ± 0.3 • C (n = 495).Seawater with a salinity of 35 ppm was pumped from 160 m depth and filtered through a sand filter before temperature was adjusted in a heat pump.The water was aerated and finally filtered through a 50 µm filter.The experiment was conducted in a flow-to-waste system consisting of circular fiberglass tanks with slightly conical bottom, resulting in a depth difference of 22 mm between tank centre and edge.The inner tank diameter was 500 mm, and water depth was 226-257 mm at the tank edge.Four replicates were used per pH level, and the flow rate was set to 10 L h −1 per exposure tank.One replicate was lost after three days in the group exposed to a mean pH NBS of 7.64 (1184 ppm) due to overflow.To prevent escape of embryos, water flow in the experimental tanks was turned off overnight at the incubation of fertilized eggs.A slow flow (< 10 L h −1 ) was re-established after 10.5 h, and then increased to 15 L h −1 (flow meter readings) after another 3.3 h.Slow bubbling of pressurized air was introduced at the tank bottom through a perforated plastic tube.Larvae were not fed during the experimental period.
Seawater at different pH levels was produced by mixing seawater with an acid stock solution of pH NBS 5.80, made from mixing CO 2 gas and seawater with an ambient pH NBS of 7.95.The pH NBS in each mixing tank was continuously adjusted to preset levels by addition of stock solution with dosage pumps (IWAKI) controlled by feedback from pH electrodes to pH transmitters (Endress & Hauser).The different water qualities were distributed by gravity to the exposure tanks via four high-mounted header tanks.The water level in header and mixing tanks was kept constant by flotation valves, and water was pumped in a loop between the mixing tanks and header tanks.Water flow in the loop was much higher than the flow from the header tanks to the exposure tanks in order to keep the pH level constant in the header tanks.
Exposure tanks were illuminated continuously by fluorescent tubes (Philips TL-D 18w/830) 110 cm above the tanks.Light was dispersed through an opaque sheet of polycarbonate 70 cm above the exposure tanks.
The pH level in each exposure tank was measured daily in a 100 mL sample using a Mettler Toledo equipped with a Mettler Toledo InLab ® ExpertPro pH-probe, calibrated with 4.00 and 7.00 buffers (Certipur ® buffer solutions, Merck KGaA, 64271 Darmstadt, Germany) traceable to standard reference material from NIST (NBS).Total alkalinity (A T ) was analysed in four replicate samples on 9, 15, and 26 March 2012 by a TitraLab radiometer.Salinity was checked daily using a WTW LF330 conductivity meter, and temperature recorded every hour using a four-detector (one in air and three in exposure tanks) EBI -1 Ebro 4 temperature S. Andersen et al.: Effect of increased pCO 2 level on early shell development in great scallop logger.The pCO 2 values (µatm) corresponding to the pH NBS values were calculated from the daily average temperature ( • C) based on hourly recordings, salinity, A T and pH using the macros in Ernie Lewis' "CO2SYS.BAS" basic program (Pierrot et al., 2006).
Larvae samples were taken after carefully stirring from bottom to top using a plunger.A 50 mm diameter pipe was lowered to the bottom, and both ends were plugged before the pipe was pulled out of the water, collecting a volume of 400 mL.On day 2 a volume of 800 mL from each replicate was withdrawn to measure shell size and classify larval deformities.When terminating the experiment after seven days, all tanks were drained and a total sample from each replicate was collected and preserved to determine final shell size, survival, and classify deformities.To determine timing of shell calcification, larvae from ambient pH in an additional tank were stained for 18-20 h from day 1 to day 2.
When collected for preservation, embryos and larvae were retained on a 35 µm mesh screen and preserved in 4 % formalin (diluted from 36.5 % formaldehyde solution, Sigma-Aldrich, cat.no.33220) in PSB for two hours, rinsed twice in 70 % ethanol, and then kept in 70 % ethanol until the ratio of developmental stages was investigated.The formalin was made from mixing 8.14 g of disodium hydrogen phosphate dihydrate (Na 2 HPO 4 • 2H 2 O, Merck 1.06580.1000)and 4.00 g of sodium dihydrogen phosphate monohydrate (NaH 2 PO 4 • H 2 O, Merck 1.06346.1000)with 100 mL of 37 % formalin in 900 mL of distilled water.
A calcein stock solution was made from mixing 1 g of calcein (Sigma C0875-5G) in 1 L of distilled water adjusted to pH NBS 7.0 using a sodium bicarbonate (NaHCO 3 , Merck 1.06329.1000)solution of 2.4 g in 400 mL of distilled water (Moran, 2000;Moran and Marko, 2005;Fitzpatrick et al., 2010).For staining of calcium, 20 mL of calcein stock solution was added to 400 mL of seawater with larvae, giving a final calcein concentration of 22 mg L −1 .Only larvae at ambient pH NBS were stained, as the aim of this pilot study was to elucidate when the calcification of the larvae shell is normally initiated.We assumed that the pH was not affected during staining since the calcein stock solution was less than 5 % of the total water volume; however the pH was not measured.Calcein has little or no effect on mollusc larvae survival or growth (Moran, 2000;Moran and Marko, 2005;Chick, 2010;Fitzpatrick et al., 2010).
Live (calcein stained) and preserved larvae were photographed using a Canon EOS 5D Mark II camera attached to an Olympus BX60 microscope.Final shell length (parallel to the hinge) and height (perpendicular to the hinge) were measured in 21-60 individuals to the nearest 4.8 µm from printed photographs.Survival was based on the initial number of fertilized eggs, and was estimated by concentrating all day-7 larvae in 25-110 mL and counting the number in 10 subsamples of 50 µL.Preserved larvae photographed at 100 magnifications using bright field were classified according to shell shape and counted.The number of trochophores classi-fied on day 2 was 76-150, and 46-178 veligers were examined on day 7. Since we were unable to find any published study on deformities in scallop larvae, deformities were categorized according to deformities as described in larvae of mussel (Mytilus sp) and the Pacific oyster (C.gigas) (His et al., 1997).We observed the same "convex hinge" and "indented shell margins" in our larvae, but called the categories "hinge deformity" and "edge deformity", respectively.We classified preserved larvae into four categories according to shell shape, and counted (1) normal (Fig. 1a), (2) hinge deformity (Fig. 1b-c), (3) edge deformity (Fig. 1d-e), and (4) both edge and hinge deformity (Fig. 1f).Trochophore larvae on day 2 were only classified into category 1 and 2, since shell edge was not visible until the shell valves covered the whole larva.In veligers on day 7, all four categories were used.
Statistica version 11 (Statsoft Inc.) was used to run statistical tests.Arcsine-transformed survival data and shell dimension data were not confirmed to normality (Shapiro-Wilk's W test), and significant effects were tested using Kruskal-Wallis ANOVA by ranks.Differences between groups were tested using multiple comparison p values (two-tailed).The rest of the data were tested using one-way ANOVA, followed by unequal N HSD post hoc test to find differences between groups.The significance level used in all tests was set to 0.05.

Initial variation in pCO 2 (µatm)
Measured and calculated water parameters for the four pCO 2 groups are given in Table 1.Salinity and total alkalinity (A T ) showed little or no variation, including in periods before and after the experiment (February-May).The pCO 2 values (µatm) for the elevated groups showed a sharp decrease during the first night after incubation due to a discontinuation in seawater flow in the exposure tanks (Fig. 2).This drop in pCO 2 (delta, pCO 2 ) increased with increasing pCO 2 levels, and fell from a mean of 2489 to 1203 µatm ( pCO 2 of 1286 µatm) in the highest pCO 2 group.The ambient group showed only an insignificant pCO 2 of 21 µatm.Naturally, the variation inpCO 2 values given as the coefficient of variation (CoV), increased with increasingpCO 2 levels, and was higher between days for replicates than between replicates on any day (Table 1).

Survival
The median survival of day-7 veligers based on the initial number of incubated eggs reared at ambient pCO 2 was 45 % (range 25-46) and decreased with elevated pCO 2 levels (Fig. 3).At the highest pCO 2 level the median survival was 12 % (range 6-13 %).There was a significant effect of pCO 2 on survival (p = 0.013), but only the group at the highest pCO 2 level was significantly different from the control. 3285

Shell size
Shell length and height on day 7 were both negatively affected by an increase in pCO 2 level (p < 0.001 for both) showing a decrease in median values of 8 and 15 %, respectively, when pCO 2 increased from 469 µatm (ambient) to the highest pCO 2 of 1599 µatm (Fig. 4).The median shell length and height were 115 and 91 µm, respectively, in the ambient group.All the pCO 2 groups were significantly different for both parameters, except for the two pCO 2 groups 807 and 1164 µatm.

Shell deformities
On day 2 an average of 85 % (82-88 %) of examined individuals had developed a shell, and only 1 % had not developed further from egg stage.Around 28-68 % of all individuals in the different pCO 2 groups on day 2 had developed a normallooking straight hinge (Fig. 5).On day 7 the percentage of normal hinges was higher: 69-97 % for all pCO 2 groups.The percentage of larvae with a normal hinge was negatively affected by pCO 2 on both days, (p = 0.004 and p = 0.021 for d2 and d7, respectively), and the lowest percentages were found at the highest pCO 2 level (Fig. 5).The percentages of all four shell shape categories on day 7 were affected by increasing pCO 2 level (p < 0.000-0.005)(Fig. 6).The percentage of normally shaped larval shells decreased from 64 to 28 % with an increase in pCO 2 level from ambient to 1599 µatm.The effect of elevated pCO 2 level seemed to follow no clear pattern as the ambient group was significantly different only from the least (807 µatm) and the most elevated pCO 2 level (1599 µatm).Shell edge deformities accounted for 30-57 % in the different pCO 2 groups (Fig. 6), and the control was significantly different only from the least elevated group.The percentage of larvae with hinge deformities, and both hinge and edge deformities, increased when the pCO 2 level increased.Only the most elevated pCO 2 group was significantly different from the ambient group for these two shell deformity categories.Variation within replicates was high, and the coefficient of variation (CoV) ranged between 5 and 70 % for all pCO 2 groups and categories.

Calcification
Trochophore larvae in ambient seawater (Fig. 7a, b(i)) stained with calcein overnight (day 1-2) showed green fluorescence when exposed to UV light (Fig. 7a, b(ii), (iii)).The fluorescence was highly concentrated in a small area on each valve, closer to the hinge than the shell margin centre.Also, there was a weak fluorescence from the total shell area, indicating that calcification of the organic shell started between day 1 and day 2.

Non-feeding regime
In the present study, scallop larvae were not fed during the 7-day experimental period.Lack of food is known to affect the energy status of scallop larvae after only three days (Andersen et al., 2012); however, the survival and growth in the control group indicate that the unfed larvae were healthy and viable.Previous OA experiments on bivalve larvae support that the decreased survival, smaller size, and increased percentage of deformed P. maximus larvae found in the highest pCO 2 group was caused by the change in water chemistry rather than the feeding regime (feeding or non-feeding) (Table 2).Thomsen et al. (2010) showed that calcifying keystone species may be able to handle the change in seawater chemistry projected for the end of the century as long as food supply is sufficient.The lack of food can be regarded as a stress factor that may push the larval energy status towards a point where the effects will be visible as slower growth, increased mortality, and increased shell deformity.We believe that in nature, bivalve larvae may experience shorter or longer periods of food scarcity; thus unfed larvae may not be uncommon, and most species may be quite well adapted to such conditions.More studies are needed to elucidate how well P. maximus larvae are able to handle the future pCO 2 levels at different feeding regimes.

Initial drop in pCO 2 level
The initial drop in pCO 2 level was most likely due to both the discontinuation in seawater flow and gas exchange at the water surface due to slow aeration during the first 12 h.This drop may have added additional physiological stress to the scallop larvae, especially in the higher pCO 2 groups.Both the delta pCO 2 and the initial maximum values may have contributed to an overestimation of the effect on all parameters (survival, growth, and deformities) that were given for the average pCO 2 levels, especially for the two groups exposed to the highest levels.Surface waters (upper 100 m) of the open ocean vary between pH 7.9 and 8.3 depending on season and geographic region, and coastal areas may even show higher fluctuations and relatively rapid changes in pCO 2 levels may be expected (Raven et al., 2005;Blackford and Gilbert, 2007;Kerrison et al., 2011, and references therein).P. maximus is most abundant at 5-50 m depth in Norwegian coastal waters, and may be exposed to such variable levels in pH in nature.However, short-term fluctuations within hours have not yet been quantified to our knowledge.Comparing our results with OA studies using more or less the same pCO 2 or pH levels indicates that the effects we measured were most likely due to elevated pCO 2 over time rather than a result of the sudden drop in the first hours of the experiment (Talmage andGobler, 2010, 2009;Watson et al., 2009).Experiments with rapid changes in pH can elucidate how variation experienced under natural conditions may affect marine organisms.However, such experiments should be conducted under controlled conditions, not as a result of methodological constraints.
Table 2. Overview of ocean acidification (OA) effect studies.The focus is on bivalves, but some studies on other calcifying invertebrates are included."Volume" is the experimental unit volume, "n" is the number of replicates, "water exchange" is the flow rate in the flow-through system, and if the system is stagnant "water exchange" gives the exchange rate of the total volume.NGinformation not given in the original publication.

Inter-and intra-species response and synergistic factors
In the present study scallop larvae showed a reduction in shell size, survival, and percentage of normally developed larvae from the ambient group to the highest pCO 2 group (discussed in further detail in the sections below).These responses to OA seem to vary little between bivalve species (Table 2).However, the magnitude of the response vary, although the pH window, temperature, and rearing volume used during experiments vary, and one should be careful in drawing conclusions about species-specific effects between studies.Comparative studies of two (Sunday et al., 2011;Parker et al., 2010;Talmage and Gobler, 2010) or several species (Liu and He, 2012;Ries et al., 2009;Talmage andGobler, 2009, 2011) mostly conclude that species have different sensitivity to OA. Within-species and betweenpopulation variation was reported by Parker et al. (2011) when they investigated differences between a wild population and selectively bred populations in sensitivity to OA exposure.Their results indicated that sensitivity to rearing conditions may have been different between the populations, as growth rate in the ambient group was higher for the selected group.This may have overestimated the differences in sensitivity to OA exposure alone.Many factors may contribute to a synergistic response of a species to OA exposure (Table 2).Some have been investigated for early life stages of bivalves: food availability (Thomsen et al., 2010(Thomsen et al., , 2013;;Melzner et al., 2011), temperature (Brennand et al., 2010;Courtney et al., 2013;Parker et al., 2010;Talmage and Gobler, 2011), salinity (Dickinson et al., 2012), parental exposure to OA (Parker et al., 2012), fertilization under OA conditions (Parker et al., 2010), and genetic selection (populations or families, Parker et al., 2011Parker et al., , 2012)).Additionally, there are other external factors that have not yet been investigated that may impact OA effects through reducing or improving the energy level: rearing volume, diet, day length, UV radiation, etc.
A large part of the work undertaken to investigate effects of OA on marine mollusc larvae has been carried out in volumes of 50 mL-200 L (Table 2).Wernberg et al. (2012) points out that it may be problematic that so many experimental studies of ocean acidification effects are undertaken in relatively small volumes.Although the larvae are small in size, small rearing volumes may not give a sufficiently stable rearing environment, and result in negative environmental stress.The volume of exposure tanks in the present study was 38 L, and we cannot rule out that these tanks are too small to keep variations in the rearing environment at a low level.However, the low variation in survival in the control group indicates that the experimental conditions were within acceptable limits.
None of the studies conducted on bivalve larvae are directly comparable with the present study (Table 2).When pCO 2 was relatively similar, the experiments were con-ducted in stagnant systems with different rearing volumes and densities, or the larvae were fed.As discussed above, all these factors may be synergistic.Larvae in our experiment were reared from eggs of one individual and sperm from three different individuals (i.e. three half-sibling groups) after being well mixed and distributed in equal densities in the experimental tanks.We assume that the half-sibling groups were distributed more or less equally in the tanks, and the variation we found between tanks is most likely not a result of family differences (intra-species differences).

Larvae survival
The survival of day-7 veligers for the ambient group replicates in this study was 25-46 %.In comparison, the average survival of day-3 veligers (referred to as hatching) in a Norwegian hatchery was 36.7 % of 12 broodstock groups (Magnesen et al., 2006).This indicates that the larval group used in the present study was viable and healthy.In the treatment groups, median survival was reduced by 41-74 %, which was much higher compared to larvae of hard clam (M.mercenaria), bay scallop (A.irradians), and eastern oyster (C.virginica) (Talmage and Gobler, 2009).The highest reduction in survival on day 6 for these larvae was 20 % for the oyster larvae at 64 and 150 Pa CO 2 (pH 7.83-7.85 and pH 7.48-7.50, respectively) when ambient conditions were 36 Pa (pH 8.02-8.08).A study on abalone larvae (Haliotis kamtschatkana) showed that both survival and growth were reduced by 40 % on day 8, when CO 2 was elevated from an ambient 400 ppm up to 800 ppm (Crim et al., 2011).These variations in survival may indicate inter-species differences, but we cannot rule out that the variation at least partly may be a result of the differences in experimental design (culture volume, water regime, feeding regime).
In an OA study on the Pacific oyster, C. gigas, Gazeau et al. (2011) manipulated the A T levels in one of their treatment groups and showed that the saturation level of CO 2− 3 had a more severe effect on development and shell growth in oyster embryos than did the pH or pCO 2 level.However, none of the other OA experiments on bivalve larvae have manipulated the A T level, and thus elevated pCO 2 levels seem to be the most plausible cause of reduced survival in these studies.Although survival of bivalve larvae in OA studies shows a significant variation, a clear trend of a negative effect due to elevated pCO 2 levels or lower pH is found (Table 2).Of six studies calculating survival on larvae, all reported a negative effect on survival (Table 2).

Larvae shell size
On day 7 larvae size (shell length and shell height) was smaller in the treatment groups compared to the group held in ambient water, but no difference was found between the two mid-treatment groups (807 and 1164 µatm The reduction in shell growth found in the present study was at the lower end of the scale compared to other studies.This may reflect an inter-species difference, and indicates that P. maximus shell growth is not the most sensitive to OA compared to other bivalves.However, the high reduction in survival may indicate that other physiological processes are highly affected by OA.The duration of the study was only seven days, and keeping the scallop larvae in the same conditions throughout the larval cycle will allow for elucidation of how OA affects growth, survival, and the critical metamorphosis over a longer term.

Shell deformities
In the present study the percentage of normally developed larvae on day 7 was 64 % in the control group.The common ratio of normal shell development in P. maximus larvae has been poorly described and is not known.In a control group of oyster (C.gigas) larvae, Kurihara et al. (2007) described that 68 % developed into normal D-shaped veligers, and 72 % of these larvae had fully mineralized shells after 48 h.This was explained as rather low values for the control group, and was supposed to be a result of using gonadal stripping to produce the embryos.In a study on M. galloprovincialis, Kurihara et al. (2008) reported abnormal morphology in < 1 % of control larvae, showing that the variation in percentage of normally developed larvae may vary between species, although other conditions may play a significant role.
The percentage of deformed larvae varied highly between the treatments in our study, and the results showed a tendency of more shell deformities with increasing pCO 2 levels.The negative effects observed have been reported also in other OA studies on bivalves (Table 2).Watson et al. (2009) found that in 8-day-old S. glomerata larvae, both shell morphology and shell surface characteristics were negatively affected by reduced pH (7.8 and 7.6).Talmage and Gobler (2010) found that the level of CO 2 strongly affected the development of the hinge structure in larvae of M. mercenaria and A. irradians using CO 2 levels of 250 (pre-industrial), 390 (ambient), 750, and 1500 ppm.They found a decline in size, integrity, and connectedness of the hinge with increasing CO 2 levels.As the authors pointed out, the hinge facilitates opening and closing of the shell, which again is crucial for both feeding and excretion.Losing the ability to perform these basic functions most certainly has a negative effect on larval survival.In the present study, the higher percentage of normally developed hinges on day 7 compared to day 2 in all larvae groups was most likely a result of increased mortality of abnormal larvae from day 2 to 7. The final sample at the end of the experiment did not show a high amount of deformed empty shells to confirm this assumption.However, empty larvae shells are extremely fragile and dissolve fast (and even faster in low-pH water), which may explain why we did not find high numbers of them in the final samples.In future experiments, more frequent sampling may reveal whether the deformed larvae die off throughout the experimental period.
It is known that other factors like self-fertilization, high egg : sperm ratio, temperature, and pollutants may cause an increase in larval deformities (Conroy et al., 1996;Spangenberg and Cherr, 1996;His et al., 1997;Krassoi et al., 1997, Concha et al., 2011;Kobayashi and Kijima, 2010).It has also been reported that factors in the rearing environment such as small rearing volumes, stagnant systems, addition of antibiotics, or pulse feeding may cause larvae deformities in control groups (Kurihara et al., 2007(Kurihara et al., , 2008;;Talmage and Gobler, 2009, 2010, 2011;Waldbusser et al., 2010;Gaylord et al., 2011;Gazeau et al., 2011).Although we cannot rule out the effect of small rearing volumes, other factors should play a minor role in the present study as we cross-fertilized the eggs to minimize the amount of self-fertilization, egg : sperm ratio was within the recommended numbers (Martinez et al., 2007;S. Andersen, personal communication, 2012), temperature was within the optimum range (Davenport et al., 1975), we used a flow-to-waste system, and no antibiotics or feed were added.
The ambient water used for the control group in the present study had a pH of 7.98, which was 0.19 units lower than the pH values Talmage and Gobler (2010) used when mimicking the pre-industrial level to investigate past, present and future ocean CO 2 concentrations.They found improved results at pre-industrial pH levels compared to ambient pH levels of 8.05 and 8.04.The shell deformities we found in our control group may therefore at least partly be a result of an already elevated CO 2 level; thus future studies on P. maximus should include experimental groups kept in pre-industrial CO 2 conditions to elucidate whether the ongoing OA already may have an impact on scallop larvae.
For the seven OA studies looking at the normal development of bivalve embryos and larvae (a total of seven different species), all reported a negative effect of elevated CO 2 levels (Table 2).Although the results in our study showed a high variation in shell edge deformities between treatment groups, the percentage of hinge deformities increased with increasing pCO 2 levels.As the determination of shell edge deformities can be a bit challenging due to damage caused by handling the samples, calculating the amount of hinge deformity is probably a better estimate of abnormalities in our study.

Calcification
In the present study we confirmed that calcein bonded to the calcium in 1-2-day-old trochophore larvae, confirming similar timing for calcification in P. maximus larvae as in larvae of M. mercenaria and C. gigas (Weiss et al., 2002), but at an earlier stage than was suggested by Cass et al. (1998) and Cragg (2006).We were unable to determine the exact time of initiation of calcification in the larvae since the staining period was 18-20 h.Staining embryos and larvae frequently (e.g.every 5 h) for a shorter period (e.g 3 h) from day 0 to day 2 may reveal more details on the exact onset of the calcification process.Comeau et al. (2009) used calcein staining of the pteropod Limacina helicina to compare shell extension in animals held at pCO 2 levels of 350 and 760 µatm over a period of five days.Although too few organisms were used to get statistically robust quantitative estimations, the results showed that shell extension in pteropods held at the highest pCO 2 levels were lower compared to those held at 350 µatm.Future OA experiments on P. maximus larvae should elucidate whether calcein staining can be used to calculate calcification either as shell extension or size of shell area in larvae held at different pCO 2 levels.

Concluding remarks and future work
The rather severe effects of elevated future pCO 2 levels on a number of bivalve larvae seem undisputable (Table 2).Talmage and Gobler (2010) showed that the CO 2 levels of today may already affect bivalve larvae negatively compared to pre-industrial levels, and hence, negative effects of OA are happening today rather than in the future.Also, OA should be seen in combination with the increase in global ocean temperature (Blunden and Arndt, 2013).Effects of a temperature increase may amplify the OA effects or reduce them.Talmage and Gobler (2011) showed that the negative effects of elevated CO 2 and temperature on larvae were additive.
Long-term studies on larvae of the great scallop P. maximus are needed to elucidate the effect of OA throughout the larvae stages, metamorphosis, and settlement, and also the potential of an evolutionary response in larval development rate.Such studies should involve assessment of the energy budget, take place in facilities that allow a normal development, and include studies on temperature effects.Effects on important processes for survival, such as immune defence functions, feeding rate, and shell mass and strength should also be investigated.More studies using calcein at the earliest larvae stages (day 0-2) are needed to increase knowledge about the onset of the larval shell calcification process, which is crucial to the existence of calcifying organisms.

Fig. 1 .Fig. 2 .
Fig. 1.Shell shape categories observed in Pecten maximus larvae.(a) Normal shell shape in formalin-preserved day-7 veliger larva; (b) hinge deformity in formalin-preserved day-2 trochophore larva and (c) in day-3 veliger larva; (d) edge deformity in shell edge centre of a live day-7 veliger larva and (e) in shell edge side of a live day-6 veliger larva; (f) both hinge and edge deformity in a live day-7 veliger larva.The scale bar is 100 µm.

Fig. 3 .
Fig. 3. Survival of day-7 scallop larvae given as median ± quartile in the four pCO 2 groups as percentage of eggs incubated.* * Significantly different from the other pCO 2 groups.

www.biogeosciences.net/10/6161/2013/ Biogeosciences, 10, 6161-6184, 2013 ies
Parker et al. (2010)8)h is in accordance with previous stud-(Table2).Gaylord et al. (2011)found that when CO 2 concentration increased from ambient 380 to 970 ppm, the shell area of mussel larvae (Mytilus californianus) on day 5 and day 7 was reduced by 7 and 5 %, respectively.Kurihara et al. (2007)showed the shell length of oyster (C.gigas) larvae after 48 h to be ca.20 % smaller at a pCO 2 of 2268 µatm (pH 7.42) than at the ambient 348 µatm (pH 8.21).Kurihara et al. (2008)also found that the shell length of Mytilus galloprovincialis larvae was reduced by about 20 % after six days when pH was reduced from the ambient 8.05 (14.3 µmol CO 2 kg −1 ) to 7.43 (83.3 µmol CO 2 kg −1 ).Parker et al. (2010)found a synergistic effect of elevated pCO 2 and suboptimal temperatures on larvae size of the two oyster species Saccostrea glomerata and C. gigas, the first being the most sensitive to the exposure.