Response of benthic foraminifera to ocean acidification in their natural sediment environment : a long-term culturing experiment

Calcifying foraminifera are expected to be endangered by ocean acidification; however, the response of a complete community kept in natural sediment and over multiple generations under controlled laboratory conditions has not been constrained to date. During 6 months of incubation, foraminiferal assemblages were kept and treated in natural sediment withpCO2-enriched seawater of 430, 907, 1865 and 3247 μatmpCO2. The fauna was dominated by Ammonia aomoriensisand Elphidium species, whereas agglutinated species were rare. After 6 months of incubation, pore water alkalinity was much higher in comparison to the overlying seawater. Consequently, the saturation state of calc was much higher in the sediment than in the water column in nearly allpCO2 treatments and remained close to saturation. As a result, the life cycle (population density, growth and reproduction) of living assemblages varied markedly during the experimental period, but was largely unaffected by thepCO2 treatments applied. According to the size–frequency distribution, we conclude that foraminifera start reproduction at a diameter of 250 μm. Mortality of living Ammonia aomoriensis was unaffected, whereas size of large and dead tests decreased with elevated pCO2 from 285 μm (pCO2 from 430 to 1865 μatm) to 258 μm ( pCO2 3247 μatm). The total organic content of living Ammonia aomoriensishas been determined to be 4.3 % of CaCO 3 weight. Living individuals had a calcium carbonate production rate of 0.47 g m−2 a−1, whereas dead empty tests accumulated a rate of 0.27 g m−2 a−1. Although calc was close to 1, approximately 30 % of the empty tests of Ammonia aomoriensis showed dissolution features at high pCO2 of 3247 μatm during the last 2 months of incubation. In contrast, tests of the subdominant species, Elphidium incertum, stayed intact. Our results emphasize that the sensitivity to ocean acidification of the endobenthic foraminifera Ammonia aomoriensis in their natural sediment habitat is much lower compared to the experimental response of specimens isolated from the sediment.

As a result, the life cycle of living assemblages was largely unaffected by the tested pCO 2 treatments.Growth rates, reproduction and mortality, and therefore population densities and size-frequency distribution of Ammonia aomoriensis varied markedly during the experimental period.Growth rates varied between 25 and 50 µm per month, which corresponds to an addition of 1 or 2 new chambers per month.According to the size-frequency distribution, foraminifera start reproduction at a diameter of 250 µm.Mortality of large foraminifera was recognized, commencing at a test size of 285 µm at a pCO 2 ranging from 430 to 1865 µatm, and of 258 µm at 3247 µatm.The total organic content of living Ammonia aomoriensis has been determined to be 4.3 % of dry weight.Living individuals had a calcium carbonate production rate of 0.47 g m −2 yr −1 , whereas dead empty tests accumulated at a rate of 0.27 g m −2 a −1 .Although Ω calc was close to 1, some empty tests of Ammonia aomoriensis showed dissolution features at the end of incubation.In contrast, tests of the subdominant species, Elphidium incertum, stayed intact.This species specific response could be explained by differences in the elemental test composition, in particular the higher Mg-concentrations in Ammonia aomoriensis tests.Our results emphasize that the sensitivity to ocean acidification of endobenthic foraminifera in their natural sediment habitat is much lower compared to the experimental response of specimens isolated from the sediment.

Introduction
Benthic foraminifera are the most diverse group of hard-shelled protists.They live at the sediment-water interface, especially in the uppermost 0-1 cm, or within oxygenated sediments down to > 12 cm depth (Corliss, 1985).It is expected that calcifying foraminifera will be adversely affected by the ongoing acidification of the oceans.
Surprisingly, previous experimental studies did not report a consistent and uniform response when living specimens were subjected to simulated ocean acidification.Most benthic foraminifera were negatively affected by high pCO 2 (Le Cadre et al., 2003;Kuroyanagi et al., 2009;Allison et al., 2010;Dissard et al., 2010;Fujita et al., 2011;Haynert et al., 2011;Uthicke et al., 2013).In contrast, some species showed an indistinct sensitivity under elevated pCO 2 (Vogel and Uthicke, 2012;McIntyre-Wressnig et al., 2013).Keul et al. (2013) revealed that not pCO 2 , but rather CO 2− 3 is the parameter which affects the test size and dry weights of Ammonia species.All these studies cultured living benthic foraminifera as isolated specimens without any natural sediment.
To date, no ocean acidification studies were reported to pursue culturing experiments with benthic foraminifera in their natural sedimentary environment.The field study of Haynert et al. (2012) in Flensburg Fjord exhibited that the carbonate chemistry of sediment pore water differed strongly from the conditions in the overlying near-bottom water.Sediment pore water pCO 2 was constantly high, ranging from 1244 to 3324 µatm during the entire year.Nevertheless, as a consequence of higher alkalinity, Ω calc was slightly supersaturated.Under these conditions the benthic foraminiferal community was not affected by seasonally elevated bottom water pCO 2 .
Benthic foraminifera are common in Kiel Fjord, western Baltic Sea, although seawater carbonate concentrations are permanently low and seasonally undersaturated with respect to Ω calc (Thomsen et al., 2013).High biological activity and nutrient inputs characterize the area, and organic-rich mud prevails in Kiel Fjord (Nikulina et al., 2008).Degradation of organic matter between the sediment-water interface influences the underlying sediment chemistry (Graf et al., 1984), and therefore the habitat of benthic Figures with Parafilm to avoid contamination by dust or evaporation.Afterwards, the samples were immediately brought to the laboratory at GEOMAR, where they were acclimated to culturing conditions at 17 • C room temperature for four weeks.During that acclimation period, stock cultures were aerated with compressed, humidified air and fed with 200 µL DT's Premium Blend containing living algae of Nannochloropsis oculata, Phaeodactylum tricornutum and Chlorella once a week.
After four weeks of acclimation, the seawater of the stock cultures was sucked off except for a few millimeters above the sediment surface.The sediment was homogenized with a spoon for a few minutes.This technique was a simple and gentle way to achieve an even distribution pattern of benthic foraminifera inside the sediment.After homogenization, each culture vessel was filled with 0.4 cm of sediment by using a spoonshaped plastic spatula.At the end of the experiment, the measured sediment height was 0.36 cm on average, which results at a given culture vessel area of 42.25 cm 2 in a mean sample volume of 15.2 cm 3 .We prepared 84 culture vessels, three replicates for four pCO 2 lines (380, 1120, 2400 and 4000 µatm), combined seven replicate sets per pCO 2 treatment were terminated.After spreading of the sediment, the culture vessels were transferred into the experimental set up and filled up with seawater from Kiel Fjord.For acclimatization and in order to restore the natural pore water chemistry in the sediments, the culture vessels were kept for 20 days at temperature of 17 • C and a salinity of 16 until the start of the experiment.This time was deemed necessary to achieve a full recovery of foraminiferal microhabiat pattern in a sub-cm scale (Ernst et al., 2002).During that time, 400 µL DT's Premium Blend were added weekly into each culture vessel.

Experimental design
The culturing of benthic foraminiferal assemblages from Kiel Fjord was performed in a closed flow-through system modified after Hintz et al. (2004) and Haynert et al. (2011).The culture vessels (6.5 × 6.5 × 4.5 cm) with an area of 42.25 cm 2 were filled with a 0.4 ± 0.1 cm sediment layer and an overlying seawater column of 4 cm.Introduction

Conclusions References
Tables Figures

Back Close
Full A nepheloid detritus layer of 0.1 cm separated the seawater from the underlying sediment pore water.The resulting pore water stagnation equates the conditions in their natural habitat, where foraminifera concentrate in the upper oxic surface layer of the sediment (Haynert et al., 2012).During field sampling we observed that the oxic sediment layer varied between 0.5 and 0.8 cm.
Each culture vessel was flushed with 25 µm cartridge-filtered and UV-sterilized seawater from Kiel Fjord.Four pCO 2 levels were adjusted by aeration of 5 L compact jerrycans with compressed and CO 2 enriched air.The target pCO 2 levels were 380, 1120, 2400 and 4000 µatm.Treatment levels were chosen in order to simulate present day and future pCO 2 peaks, which are transiently occurring today and are going to prevail for longer periods in future in Kiel Fjord (Thomsen et al., 2010(Thomsen et al., , 2013;;Melzner et al., 2012).The preconditioned seawater from each tank flowed through the culture vessels, which were three times replicated for each pCO 2 level.The overflow seeped through the fissure between lid and vessel, and draining off to a sink.The flow rate was adjusted to 0.16 mL s −1 , which is sufficient to replace the water volume of the vessels 1.4 times per hour.The overflow drained off to a 60 L catchment tank.In the catchment tank, the water was sparged with compressed air at a high rate, in order to remove the excess CO 2 before the water was pumped back to the compact jerrycans.In order to avoid evaporation by aeration of the seawater with compressed and thus dry air, the gas was humidified in gas-washing bottles which were inserted in each compressed air connection.Evaporation of one liter per week was compensated by refilling the system with deionised water.Food (400 µL DT's Premium Blend) was added to each culture vessel every week.In order to monitor changes of the foraminiferal community during the incubation time, the sediment of three culture vessel replicates was analyzed monthly for foraminiferal assemblage composition.The experiment ended after six months.

Determination of water chemistry parameters
Seawater pH NBS , temperature and salinity were determined with a WTW 340i pH analyzer and a WTW Cond 315i conductivity meter.The pH NBS -electrode was calibrated with standard buffer solutions of pH 4.01, 7.00 and 10.00 (WTW standard, DIN/NIST buffers L7A).Precision was ±0.01 for pH NBS and ±0.1 • C for temperature.The precision of the conductivity meter was ±0.1 salinity units.Nutrient concentrations were analyzed in sterile-filtered (0.2 µm pore size) water samples.Phosphate (PO 3− 4 ) and silicate (Si) were analyzed colorimetrically in a spectrophotometer (U 2000; Hitachi-Europe) at a wavelength of 882 nm and 810 nm according to Koroleff and Grasshof (1983).The measurement precision was ±0.2 µmol L −1 for phosphate and, in dependence of the concentrations, 2.5 to 6 % for silicate.Total dissolved inorganic carbon (C T ) was analyzed in sterile-filtered (0.2 µm pore size) water samples taken from the culture vessels during the incubation time using an AIRICA autoanalyzer (Maranda GmbH, Kiel, Germany) with a precision of 2-4 µmol kg −1 .The accuracy of the C T measurements was ensured by using certified reference material provided by Andrew Dickson of the Scripps Institution of Oceanography.Seawater carbonate system parameters pCO 2 , total alkalinity (A T ) and omega for calcite (Ω calc ) were calculated from pH NBS , C T , temperature, salinity, PO 3− 4 and Si values using CO2SYS software by Lewis and Wallace (1998).Dissociation constants K 1 and K 2 were chosen according to Mehrbach et al. (1973) as refitted by Dickson and Millero (1987) and KHSO 4 dissociation constant after Dickson (1990).
Water chemistry parameters of pH NBS , A T , temperature and salinity were analyzed Introduction

Conclusions References
Tables Figures

Back Close
Full sediment samples were transferred into 50 mL centrifuge tubes and centrifuged at 3000 rpm for 15 min in order to separate the sediment pore water from the sediment.The extracted pore water was transferred through 0.2 µm sterile filters into 20 mL PVC bottles.pH NBS and temperature were measured using the WTW 340i pH analyzer, salinity was determined with WTW Cond 315i conductivity meter.Total alkalinity (A T ) was determined with a Metrohm titration instrument according to Ivanenkov and Lyakhin (1978).A greenish-brown Methyl-Red and Methylene-Blue indicator was added, and titration was performed with 0.02M HCl and finished until a stable light pink colour occurred.During titration, the sample was degassed by continuously bubbling argon through the solution in order to remove the generated CO 2 or H 2 S. The measured values were standardized using an IAPSO seawater solution.The precision of the alkalinity measurements was 0.37 %.Carbonate parameters pCO 2 , C T and Ω calc of seawater and sediment pore water were calculated from measured pH NBS , A T , temperature, salinity, PO 3− 4 and Si values according to dissociation constants as specified above.All chemical parameters are reported as mean values of replicate measurements (Table 1).

Foraminiferal processing
Benthic foraminiferal sediment samples were transferred in 100 mL Kautex TM wideneck containers, preserved and stained with Rose Bengal ethanol (94 %) solution of 2 g L −1 for three weeks following Lutze and Altenbach (1991).This period is sufficient to stain the protoplasm completely with Rose Bengal in all tests of foraminifera (Schönfeld et al., 2012).Samples were first passed through a 2000 µm screen in order to remove mollusks shells and pebbles.Subsequently the samples were gently washed with tap water through a 63 µm sieve.The 63-2000 µm and > 2000 µm fractions were dried at 60 • C for at least 24 h.The size fraction 63-2000 µm was picked completely for living and dead foraminifera.Samples were not split.All Rose Bengal stained foraminifera were considered as living at the time of sampling, whereas unstained tests were consid-Introduction

Conclusions References
Tables Figures

Back Close
Full ered as dead.Living and dead specimens were sorted by species.All species, except Ammonia aomoriensis, were mounted in Plummer cell slides with glue, counted and measured using an eyepiece reticle on the Wild M3C dissecting microscope.Ammonia aomoriensis was kept in single cell slides for further analyses.Test morphometry of Ammonia aomoriensis was analyzed with an automated image analysis system using a Leica Z16 APO microscope, and analySIS software (version 5.0) at the University of Angers, France (cf.Bollmann et al., 2004;Clayton et al., 2009).Resolution of images is 2.69 µm −2 .Due to the oval shape of Ammonia aomoriensis tests, the size of tests is given as mean diameter, calculated from the minimum and maximum diameter.
The census data were standardized to population density given as individuals/tests per 10 cm −3 sediment.Histograms were created depicting the proportion of living A. aomoriensis in 11 size classes from < 50 to > 500 µm in 50 µm intervals and presented as the average of three replicates taken each month during the six months incubation time.
The dry weight of dead and living A. aomoriensis was measured using a microbalance with an accuracy of 1 µg (Sartorius, M3P-000V001).Dry weight per individual/test was calculated from the total dry weight divided by the total number of individuals/tests.
Finally, tests of Armorella sphaerica and corroded specimens of A. aomoriensis were photographed with a MiniPixie (MPX2051UC) digital microscope.

Electron microprobe analysis
The two ranked species, A. aomoriensis and E. incertum, were prepared for Electron MicroProbe (EMP) X-ray microanalysis (JEOL JXA 8200).We focused on the visual elemental distribution of Ca and Mg on cross-sections of the tests, which were cultured at pCO 2 of 3247 µatm and retrieved after six months incubation.The advantage of this technique is high spatial resolution (±1 µm), so that single foraminiferal tests can be analyzed (Kellner et al., 1998;Glock et al., 2012).Introduction

Conclusions References
Tables Figures

Back Close
Full Tests of both species were embedded under vacuum into Araldite TM epoxy resin by using a CitoVac TM vacuum embedding system (Struehrs TM ).Small drops of resin were used to fill the inner part of the chambers.The tests were set under pressure to collapse air inclusions inside the resin and hardened in a drying cabinet at 60 • C.
Using the Tegra-Pol-21 system (Struehrs TM ), the surface of the resin was ground down by hand with alumu-silica grinding paper until the proloculus of the specimen became visible.The surface was polished with different grades of alumo-silica and diamond paste (until 1 µm grain size) on a self rotating polishing plate.
Prior to the measurement, each cross-section was carbon coated.The microprobe was operated in a wavelength dispersive mode by using different Ka X-ray lines for each element.Two different detector crystals were used for the elements Ca (crystal: PETJ) and Mg (crystal: TAPH).The acceleration voltage was 15 kV and the beam current was 20 nA.The tests of A. aomoriensis and E. incertum were mapped by 0.5 µm step size and a dwell time of 500 ms.The results are illustrated as maps of relative measured intensities for both elements.

Calculations
Carbonate production and accumulation of the major calcifying species A. aomoriensis have been calculated from the difference of dry carbonate weight of individuals or tests produced or accumulated during six months incubation time.The production or accumulation rate in grams refers to an area of 1 m 2 per year.
For 100 living A. aomoriensis from size fraction 200 to 300 µm, the total organic content, including cytoplasm and organic linings of the test, was determined by combustion off the organic material from the carbonate tests in a muffle furnace at 500 • C for 6 h.
The amount of total organic content is calculated from the weight loss and given as a percentage of the total dry weight of the individuals (Table S6 in the Supplement).
This value was subtracted from total weight to obtain the dry weight of shell carbonate (Table S5 in the Supplement).Introduction

Conclusions References
Tables Figures

Back Close
Full

Statistics
Two-factorial ANOVA was performed with the two parameters pCO 2 and incubation time with STATISTICA 8 (Table 4).All measured raw data of each treatment were used for statistical analyses.The data are given as mean ± standard deviation.

Carbonate chemistry
Seawater temperature and salinity were stable during the experimental period.Average temperatures ranged from 16.4 to 17.2 • C and salinities from 15.1 to 15.9 (Table 1).
Phosphate and silicate concentrations changed markedly over the course of the experiment.Phosphate concentration increased 31-fold during the first three months from 0.19 to 5.86 µmol L −1 (Table 1).Mean silicate concentrations were in general high with 191.98 µmol L −1 and exhibited a threefold increase from 138.94 to 367.19 µmol L −1 after two months (Table 1).By the end of the experiment, concentrations of phosphate and silicate decreased again and achieved the initial values that were observed at the beginning of the experiment.In dependency to seawater pCO 2 , mean pH values ranged between 8.21 and 7.17 (Table 1) and were stable throughout the incubation period.
Total inorganic carbon increased with increasing pCO 2 from 2210.5 to 2464.2 µmol L −1 (Table 1).The calculated mean pCO 2 levels of the treatments were thereby 430, 907, 1865 and 3247 µatm (Table 1).It has to be noted that CO 2 concentrations in the unchanged air effected a higher pCO 2 , while the partial pressures at higher concentrations led to markedly lower pCO 2 levels than the target values.In the following, we refer our results to the pCO 2 levels that were prevailed in the culture vessels (Haynert et al., 2011).Mean seawater alkalinity was 2379.3 µmol kg −1 and did not differ significantly between the treatments (Table 1).At a pCO 2 of 1865 µatm, seawater Introduction

Conclusions References
Tables Figures

Back Close
Full Ω calc values decreased partially below 1. High pCO 2 of 3247 µatm caused permanent undersaturation of Ω calc with 0.57 on average (Table 1).The carbonate chemistry measurements revealed strong differences of pCO 2 and Ω calc between seawater and sediment pore water in the culture vessels (Fig. 1).The sediment pore water was characterized by higher pCO 2 and lower pH than the supernatant seawater in the culturing vessel (Fig. 1a and b).At the same time, pore water alkalinity (3127.8µmol kg −1 ) was much higher than the bulk seawater alkalinity (0-2 cm water depths) of 2113.1 µmol kg −1 and the water overlaying the sediment (2-4 cm water depths) with 2442.3 µmol kg −1 .The accumulation of A T in the sediment caused a higher saturation of Ω calc and even in the highest pCO 2 treatment a slight supersaturation, Ω calc > 1 (Fig. 1c and d, Table S1 in the Supplement).
The fauna was dominated by A. aomoriensis with 99 % of the living fauna and 85 % in the dead fauna, whereas E. incertum was rare with 1 % of living individuals and 7 % of dead tests (Tables 2 and 3).All other species were very rare or were only occasionally found in the foraminiferal assemblages.Accordingly, we will focus on the dominant calcareous species A. aomoriensis in the following.

Ammonia aomoriensis
The population density and size-frequency distribution exhibited strong variations due to growth, reproduction and mortality of foraminiferal faunas during the six months incubation.Incubation time had a significant effect on test diameter of living A. ao-Introduction

Conclusions References
Tables Figures

Back Close
Full moriensis (F(5,45) = 18.14, p < 0.01) whereas pCO 2 had no effect as a single factor (F(3,45) = 0.61, p > 0,05,Table 4).However, the interaction of both, pCO 2 and incubation time was significant (F(15,45) = 2.14, p < 0.03).The size distribution revealed growth cohorts which were characterized by a fixed, simultaneously reproduction period.A further character of a cohort was the similar increase of test diameter and the loss of individuals by death during the same period.Initial population density of living A. aomoriensis fauna varied from 342 to 583 ind. 10 cm −3 (Table S2 in the Supplement) and the mean test diameter per treatment ranged from 169 to 182 µm (Table S2 and 8 in the Supplement).Lowest mean population densities were observed in the 430 µatm treatment with 295 ind. 10 cm −3 , intermediate with 367 and 407 ind. 10 cm −3 in 907 and 3247 µatm pCO 2 -treatments, and highest mean densities of 524 ind. 10 cm −3 at a pCO 2 of 1865 µatm (Table S2 in the Supplement).Whereas the population density at 430 µatm was relatively stable, the density fluctuated considerably in the 907, 1865 and 3247 µatm pCO 2 -treatments (Fig. 3).Ammonia aomoriensis density declined strongly during the first month at pCO 2 of 1865 and 3247 µatm.An explosive increase of living Ammonia aomoriensis up to 610, 1378 and 518 ind. 10 cm −3 was observed at pCO 2 of 907, 1865 and 3247 µatm after three months.A further increase up to 602 ind. 10 cm −3 followed at 3247 µatm after five months (Fig. 3a, Table S2 in the Supplement).The strong increase of population densities correlated with dominant reproduction events at 1865 µatm after three months, and two sequent reproduction events at pCO 2 of 3247 µatm after three and four months as indicated by the occurrence of high numbers of small individuals with a test diameter < 100 µm (Fig. 4, Table S3 in the Supplement).A less pronounced reproduction event in the background was observed at a pCO 2 of 907 µatm after four and five months, though population densities were not affected (Fig. 4, Table S3 in the Supplement).
Ammonia aomoriensis was able to reproduce from size class of 250 µm up to 350 µm, at 350 µm 40 % of the cohorts had reproduced.After reproduction, the growth of individuals at 1865 µatm led to an increase of the mean test diameter from 185 to 287 µm until the end of the experiment (Table S4 in the Supplement).The mean growth of A. Introduction

Conclusions References
Tables Figures

Back Close
Full aomoriensis varied between 25 and 38 µm per month during incubation, at the beginning specimens grew by up to 50 µm within one month.

Abundance and test diameter of dead Ammonia aomoriensis
The abundance of the dead fauna ranged from 24 to 61 tests 10 cm −3 .At pCO 2 -levels of 430 and 907 µatm, their abundance increased steadily until the 4th and 5th month, and subsequently decreased until the end of experiment (Fig. 3b, Table S2 in the Supplement).In contrast, at pCO 2 of 1865 and 3247 µatm, the number of A. aomoriensis tests did not significantly change until the 4th month (Fig. 3b).Afterwards, the number of empty tests increased by 80 % at 1865 µatm until the end of the experiment (Table S2 in the Supplement).These results correlated with a strong decline of population densities from 1378 to 338 ind. 10 cm −3 from the third to the fifth month (Fig. 3a).At 3247 µatm, test abundance decreased to 32 tests 10 cm −3 at the end of the experiment (Table S2 in the Supplement).
The mean test diameter ranged between 290 and 283 µm at pCO 2 from 430 to 1865 µatm.At higher pCO 2 of 3247 µatm, test diameter was clearly lower with 258 µm (Table S4 in the Supplement) which indicated a frequent mortality of large and adult foraminifera up to 1865 µatm, whereas at higher pCO 2 a frequent mortality of smaller and younger individuals was recognized.Test diameter of dead A. aomoriensis was significantly effected by incubation time (F(5,46)= 16.65, p < 0.01) but not by pCO 2 (F(3,46, p > 0, 05) (Table 4).

Dry test weight
The initial dry weight per individual of living A. aomoriensis was 1.56 µg on average (Fig. 6, Table S5 in the Supplement).With increasing test diameter, the dry weight per living individual increased constantly, but no significant differences of dry weightdiameter ratio were observed between the different pCO 2 treatments (Fig. 5a).Dry and 3247 µatm respectively until the first month (Table S5 in the Supplement).At 430 and 3247 µatm, dry weight per individual did not change significantly any more until the end of the experiment (Table S5 in the Supplement).In contrast, the dry weight of A. aomoriensis increased up to 7.30 µg at a pCO 2 of 907 µatm after two months (Table S5 in the Supplement), which correlated with an increase in test diameter by 116 µm on average (Fig. 4).At a pCO 2 of 1865 µatm, dry weight per individual decreased significantly to 1.30 µg after three months (Table S5 in the Supplement).These results correlated with the reproduction events and resulting dominance of small foraminifera of size classes < 50 up to 200 µm (Fig. 4).
In comparison to living A. aomoriensis, mean dry weight per empty test was higher with 6.18 µg on average (Fig. 6, Table S5 in the Supplement).This probably results from the more frequent reproduction or mortality of large specimens.Similar to the dry weight of specimens from the living fauna, no effect of any pCO 2 was observed with empty tests (Fig. 5b).In accordance to changes in test diameter, the average dry weight per test changed during the experiment.At 430 and 907 µatm, dry weight per test increased by 60 and 50 % until the second month and remained stable with mean values of 5.63 and 5.84 µg until the end of the experiment (Table S5 in the Supplement).
In pCO 2 -treatments 1865 and 3247 µatm, dry test weight increased by 41 and 71 % during the first month (Fig. 6b), followed by a further slight increase up to 7.73 µg at 1865 µatm after five months (Table S5 in the Supplement).These observations were in agreement to the decline of population densities of A. aomoriensis after the first and fifth month at 1865 µatm (Fig. 3a).

Carbonate production and accumulation
The total organic content of living individuals was found to be 4.3 %, which was subtracted from total dry weight in order to determine the absolute amount of CaCO 3 (cf.Movellan et al., 2012) (Table S5 in the Supplement).Production is considered as weight increase measured at subsequent sampling.Average production during the entire incubation period is given as mean value of monthly weight increases.Living 9537 Introduction

Conclusions References
Tables Figures

Back Close
Full A. aomoriensis exhibited the lowest mean CaCO 3 production of 595 µg at the lowest pCO 2 of 430 µatm, following by 886 µg at the highest pCO 2 of 3247 µatm after six months incubation.The three replicate treatments of 907 and 1865 µatm showed similarly higher mean CaCO 3 production of 1269 and 1256 µg with a total culture vessel area of 42.25 cm 2 and an incubation time of six month.The calculated total carbonate production rate for living individuals was 0.47 g m −2 yr −1 (Table S6 in the Supplement).
The CaCO 3 accumulation of empty tests varied between the pCO 2 treatments (Table S6 in the Supplement).The highest rate of 908 µg CaCO 3 was observed at 907 µatm, followed by 670 µg at 430 µatm.With increasing pCO 2 of 1865 and 3247 µatm, CaCO 3 rates decline to 435 and 239 µg after six months incubation with the above specified settings.The calculated total carbonate accumulation of empty tests was 0.27 g m −2 yr −1 (Table S6 in the Supplement).

Observations of test structures and elemental distribution
At high pCO 2 of 3247 µatm, light micrograph images of dead A. aomoriensis showed that most of the tests were completely destroyed during the last two months of the experiment.The strongest effect of test degradation was the charistic star-like appearance, which left only the umbilicus area and the suture of the chambers intact.
Cross-sections of test walls of A. aomoriensis (Fig. 7a) and E. incertum (Fig. 7b) revealed that the tests of both species were composed of a single calcium carbonate layer, which contained high Ca and low Mg calcite.In comparison to E. incertum, Introduction

Conclusions References
Tables Figures

Back Close
Full

Discussion
This study describes the impact of elevated pCO 2 on a natural foraminiferal community from the southwestern Baltic Sea which was exposed to 4 different pCO 2 levels for six months.It is the first study to investigate the multi-generational response using natural sediment with living foraminiferal fauna of natural composition.The foraminiferal community developed and reproduced normally in all treatments, as the specific sediment chemistry caused, compared to the bulk seawater, a relative high saturation state with respect to omega calcite, and thereby prevented the tests of living individuals from dissolution.

Sediment chemistry
The sediment used in this experiment was sampled from Kiel Fjord, western Baltic Sea, which is rich in organic matter (Nikulina et al., 2008).During the experiment, a layer of organic detritus formed at the sediment-water interface.Organic substances influence the microbial activity, the rates and pathways of organic matter remineralization and nutrient recycling (Jonsson et al., 1990;Conley and Johnstone, 1995), and thereby influenced the underlying sediment chemistry (Graf et al., 1984).The remineralization of organic matter is mainly attributed to a wide array of hydrolytic and fermentative bacteria (Turley et al., 2000), which break complex multi-carbon compounds down to smaller, more soluble and digestible substances.Bacterial activity is often attributed as a limiting step in degradation of organic matter and extent of degradation (Tyson, 1995;Arnosti, 2004).
It is conceivable that silicate and phosphate concentrations were highly variable in the stagnated water body of the culture vessels throughout the experiment due to the high microbiological activity.Concentrations of both nutrients increased during the first two months and afterwards decreased until the end of the experiment.
One aspect could be the disturbance of the sediment after homogenization at the beginning of the experiment.This process could cause a remobilization of the bounded Introduction

Conclusions References
Tables Figures

Back Close
Full nutrients in the pore water, which diffused gradually through the filter of the nepheloid layer.This in turn would increase nutrient concentration increased in the overlying seawater.However, silicate concentrations remain relatively high.Microscopic analysis showed that the detritus-layer was composed of centric diatoms, which also dominated in the natural environment.The algal material of Nannochloropsis oculata, Phaeodactylum tricornutum and Chlorella added as food did not accumulate on the sediment surface.This in turn could imply that the added algae material was directly dissolved, which could explain the high silicate concentrations.Otherwise, increasing silicate concentrations may result from remineralization of diatoms (Conley and Johnstone, 1995).
Sediment samples were taken in April, when high nutrient and productivity is usually encountered (Wulff et al., 1986(Wulff et al., , 1990;;Elmgren, 1989;Wollast, 1998;Thomas et al., 2003;Pätsch and Kühn, 2008).The spring bloom, ranging from the end of February to early April in the southwestern Baltic Sea, is dominated by diatoms, therefore maximal diatom biomass and biogenic silica flux to the sea floor (Wasmund et al., 2005(Wasmund et al., , 2006)).After deposition at the sediment surface, remineralization of the dead diatoms may have caused a substantial increase of silicate in the culture vessels during the experiment (Conley and Johnstone, 1995).
The high degradation of organic matter such as diatoms may have also caused an increase of phosphate concentrations (Balzer, 1986).In oxic sediments dissolved inorganic phosphate (DIP), is usually bound to calcium, chemisorbed by ironoxyhydroxides in distinct iron compounds and causes high phosphate accumulations in the sediment (Nissenbaum, 1979;Filipek and Owen, 1980;Krom and Berner, 1981).However, high O 2 consumption by degradation of organic matter may have reduced the oxygen concentrations in the sediment, resulting in mobilization of DIP from sediments (Krom and Berner, 1981).The decrease of both nutrients until the end of the experiment may have arisen from bacterial microbiological activity of bacteria, although this aspect has not yet been sufficiently studied.Introduction

Conclusions References
Tables Figures

Back Close
Full Degradation processes of organic matter at the sediment-water interface consumed O 2 and produced CO 2 , and thus resulted in higher sediment pore water pCO 2 in comparison to the seawater.Similarly, the sediment was characterized by much higher alkalinity values than the seawater (3127 vs. 2379 µmol kg −1 ).Both aerobic or anaerobic degradation of organic matter release dissolved inorganic carbon (DIC).Whereas aerobic degradation has no significant effect, anaerobic degradation increases alkalinity (A T ) by denitrification and sulfate reduction (Yao and Millero, 1995;Thomas et al., 2009).Therefore, we assume that O 2 consumption at the sediment-water interface exceeded the delivery of O 2 via diffusion, which could induce anaerobic conditions in the sediment pore water.A dark grey sediment layer on the bottom of the culture vessels supported our assumption.This in turn enhances the CO 2 buffer capacity and consequently causes a relatively high Ω calc (Thomas et al., 2009).These differences of the carbonate chemistry between water column and sediment was in agreement with observations from field studies (Thomas et al., 2009;Haynert et al., 2012).

Foraminiferal communities
The species composition of living and dead assemblages which were used in the experiment were similar to the assemblages prevailing in Kiel and Flensburg Fjords (Nikulina et al., 2008;Polovodova et al., 2008Polovodova et al., , 2009;;Haynert et al., 2012).At the beginning of the experiment, the population density of the living fauna (433 ind. 10 cm −3 ) was very similar to the mean population density of 448 ind. 10 cm −3 , found in the nearby sampling station PF15-13 in December 2005 (Nikulina et al., 2008;Polovodova et al., 2008).The living assemblages at both stations were dominated by A. aomoriensis (A. beccarii of Nikulina et al., 2008).In December 2005, E. excavatum clavatum was common and E. excavatum excavatum and E. incertum were rare, whereas in April 2011, the living fauna consisted almost exclusively of A. aomoriensis.

Conclusions References
Tables Figures

Back Close
Full The dominance of living A. aomoriensis is probably explained by the diatom dominated spring bloom, which enables high food supply and thereby facilitates reproduction.A similar relationship between diatom spring blooms and reproduction events was reported for E. excavatum clavatum at deeper waters off Boknis Eck, southwestern Baltic Sea (Schönfeld and Numberger, 2007).
In addition, the arenaceous species Armorella sphaerica occurred by incident in the set up.The species was recorded in Kiel Bay by Rhumbler (1935) and later by Brodniewicz (1965) in the southern Baltic Sea.Its sporadic occurrence and sudden appearance in our experiment could be explained by the activation of dormant propagules (Alve and Goldstein, 2002).Alternatively, the species may be widespread but rare in Kiel Fjord, and so have been overlooked in previous studies.In our experiment, however, suitable conditions, such as food supply and oxygenation could have induced the activation of resting stages.

Response of Ammonia aomoriensis life cycle
In the first month of the experiment, population densities of living A. aomoriensis declined strongly at a pCO 2 of 1865 and 3247 µatm.This decline might indicate conditions of Ω calc < 1 in the sediment.The sediment was put in the vessels 20 days prior to CO 2 manipulation.Nevertheless, the remineralization processes in the sediment were still ongoing as indicated by the strong increases of seawater phosphate and silicate concentration during the first two months.Therefore, high pore water alkalinity and thereby supersaturation of calcium carbonate might have not been established in the early phase of the experiment.Undersaturation has been already shown to decrease growth of A. aomoriensis and thereby increase mortality and test dissolution (Haynert et al., 2011).Similarly, a decrease of Ammonia species growth was observed with decreasing [CO 2− 3 ], respectively Ω calc < 1 (Keul et al., 2013).
During the course of the experiment, however, the progressing remineralization processes in the sediment may have increased saturation state above 1.Accordingly, A. Introduction

Conclusions References
Tables Figures

Back Close
Full aomoriensis was relatively unaffected by high pCO 2 later on and continued to grow and reproduce, even under highly elevated pCO 2 conditions (Haynert et al., 2012).
The current study underscores, that endobenthic foraminifera from marginal low saline habitats responded completely different under elevated pCO 2 in culturing experiments, depending whether they were kept with or without their natural sediment.Furthermore, high or low organic matter content of the substrate, grain size composition (Conley and Schelske, 1989), bacterial remineralization (Turley et al., 2000), and oxic or anoxic conditions (Jonsson et al., 1990) play a certain role in sediment and pore water chemistry, and thereby may influence the living conditions of benthic communities.In fine-grained sediments with a high content of organic substances these processes play an important role for the foraminiferal microhabitat.It is reasonable to assume that the effects on carbonate chemistry are less pronounced in coarse sediments.These results emphasize the importance to understand the sediment carbonate chemistry in the context of their natural settings.
During the six months incubation, the populations of all treatments revealed strong variations of their density and size-frequency distribution due to growth, mortality and reproduction events.The shift of mean size distribution between the monthly sampling intervals enabled the calculation of growth rates which varied between 25 and 38 µm per month and agreed with earlier growth rates estimates reporting 13 to 39 µm (Haynert et al., 2011).This growth corresponds to the addition of one or two chambers per month, as one added chamber equates 27 ± 16 µm in test diameter (Haynert et al., 2011).An earlier study described, growth rate as being uniform in all size classes.Introduction

Conclusions References
Tables Figures

Back Close
Full Small individuals grow more rapidly, whereas adult individuals slow down the growth before reproduction (Bradshaw, 1957).Furthermore, changing habitat conditions, in particular food supply, play an important role at chamber formation (Bradshaw, 1961).It appears plausible that larger and a greater number of chambers will be formed under favorable conditions.
Size distributions of tests revealed that most of the individuals reproduced or died at a mean size of 285 µm at a pCO 2 ranging from 430 to 1865 µatm.At the highest pCO 2 of 3247 µatm, mean test diameter was slightly lower with 258 µm.Apart from growth, reproduction events had a pronounced effect on the mean size of the population.Cohorts of small, juvenile individuals (< 100 µm) could be identified.Whereas reproduction events could not be detected at 430 µatm, strong cohorts with high numbers of individuals were observed at 1865 and 3247 µatm, with a strongest variability at 1865 µatm.
Ammonia aomoriensis reproduced at sizes between 250 µm and 350 µm.At 350 µm, 40 % of the cohorts either have reproduced or died.These results agreed with a previous study of Bradshaw (1957) where the mean test diameter of Streblus beccarii var.tepida ranged between 266 and 357 µm at reproduction.In dependency of environmental conditions, reproduction follows approximately 28-day intervals with about 28 young juveniles per parent under optimal conditions (Bradshaw, 1957).In the current study we observed reproduction events, but the different cohorts made it impossible to determine exactly either the intervals of reproduction or the number of juveniles per parent.
In dependency of asexual and sexual reproduction, the generations consisted of different forms as influenced by test dimorphism (Murray, 2012).The megalospheric form is the result of asexual reproduction, whereas the microspheric form is the product of sexual reproduction (Lister, 1903).However, the reproduction of A. aomoriensis is insufficiently studied to date and further field and laboratory experiments would be necessary to better understand the life cycle of A. aomoriensis, and the species specific variability of Ammonia in general.However, since reproduction was observed in all pCO 2 treatments, the conditions prevailing in the sediment were sufficient to allow Introduction

Conclusions References
Tables Figures

Back Close
Full successful reproduction even at 3247 µatm pCO 2 .This supports findings that natural foraminiferal communities in the southwestern Baltic Sea can withstand permanent high pCO 2 levels of above 2000 µatm when they are fully sheltered in fine-grained sediment (Haynert et al., 2012).

Carbonate production of Ammonia aomoriensis
The carbonate production of benthic foraminifera was described in previous studies (e.g.Phleger and Soutar, 1973;Muller, 1974;Wefer and Lutze, 1978;Hallock, 1981;Bosence, 1989;Langer et al., 1997) which determined the foraminiferal carbonate content of sediment samples from different habitats.In contrast, in our study the main focus was on living benthic foraminifera containing cytoplasm.Little is known about the fractions of cytoplasm and tests weight in benthic foraminifera.The estimated total organic content in living A. aomoriensis is 4.3 % of total dry weight which is in agreement with the estimated protein content of 20 % and 5 % organic dry content in Ammonia tepida (Movellan et al., 2012).The study of Wefer and Lutze (1976) observed a ratio of 1 : 1 of protoplasm and test weight in benthic foraminifera which referred to wet weight (see Table 1 in Wefer and Lutze, 1976) The cytoplasm water content plays an important role in this determination as TEM micrograph images documented a high amount of seawater vacuoles in the foraminifera chambers (see Fig. A3.3.2 in Glock, 2011).Therefore, and in agreement with A. Movellan (personal communication, 2013), we assume a water content of approximately 76 % in foraminiferal cytoplasm.However, the percentage of water and cytoplasm content varies in relation to test size and weight, and due to the natural variability of the microenvironment and the species-specific morphometry (Movellan et al., 2012).The carbonate production of living A. aomoriensis amounts 0.47 g m −2 yr −1 , whereas the carbonate accumulation of dead tests equates 0.27 g m −2 yr −1 during the incubation time.This estimate of CaCO 3 production is much higher than those reported for natural foraminiferal assemblages in the western Baltic Sea.Production rates of calcareous benthic foraminifera rates ranged from 0.01 to 0.03 g m −2 yr −1 in the shallow Introduction

Conclusions References
Tables Figures

Back Close
Full water habitat from 5-15 m depth of Kiel Bay (Wefer and Lutze, 1978).However, this low carbonate production was probably caused by a strong seasonality of food supply and mechanical stress on shoals with lag sediments, where these samples were taken.Muddy sediments such as in Kiel Fjord, with high organic matter content as potential source of food, showed a markedly higher foraminiferal carbonate production of up to 3.12 g m −2 yr −1 (Wefer and Lutze, 1978;Nikulina et al., 2008).Therefore, both favorable food supply and sustaining high population densities during the incubation facilitated the observed high CaCO 3 productivity of A. aomoriensis.Not less than 36 % of the produced tests are accumulated in the sediment, which is obviously higher as reported by Wefer and Lutze (1978), the accumulation rate ranged between 0 and 4 % in the natural habitat.In the transitional environments from fine sand to mud, the accumulation rate varied from 1 to 1.2 % (see Table 2 in Wefer and Lutze, 1978).Under natural conditions, mechanical forces play an important role and affected the tests, as well as dissolution and ingestions by macro-and meiofauna (Wefer and Lutze, 1978).In the present study, A. aomoriensis was cultivated in a protective habitat without any mechanical impacts of the empty tests, therefore a higher accumulation rate was recorded.

Impact on foraminifera tests
Living A. aomoriensis and E. incertum exhibited no dissolution features during the whole incubation time.The conditions in the natural sediment create a protective microhabitat for benthic foraminifera with a relatively high calcium carbonate saturation state.These observations are in contrast to previous laboratory studies, where A. aomoriensis and A. beccarii exhibited a relationship between pCO 2 , respectively pH, and test degradation (Le Cadre, 2003;Haynert et al., 2011).In comparison to the present study, living foraminifera were isolated from their natural sediment, therefore the tests of living specimens were directly affected by the carbonate chemistry conditions of the seawater.Introduction

Conclusions References
Tables Figures

Back Close
Full However, some empty tests of A. aomoriensis were destroyed at a pCO 2 of 3247 µatm during the last two months of incubation.This may be because high pCO 2 and unfavorable Ω calc < 1 reduced the average diameter.In contrast, empty tests of E. incertum showed no signs of dissolution.This species specific response agrees well with our field observation from Flensburg Fjord (Haynert et al., 2012).In order to explain these differences, cross sections of both species were analyzed using EMP.The element maps suggested that tests of both foraminiferal species are composed of a single layer of secondary calcite, which is characterized by a relative low Mg-content (Erez, 2003).In comparison to E. incertum, the test of A. aomoriensis displayed a higher Mgcontent and was highly porous.The crystals, rich in Mg are oriented in a structure of small radial needles (Hansen and Reiss, 1972;Bellemo, 1974), which are sensitive and dissolved first at undersaturated conditions.In combination with a greater test surface, A. aomoriensis was therefore less resistant to dissolution than E. incertum.EMP observations are thus considered to provide valuable constraints for the sensitivity of a species towards dissolution stress.

Conclusions
In the present ocean acidification study, benthic foraminifera were cultured in their natural sediment over a period of several generations.Under those laboratory conditions, the alkalinity (A T ) and therefore the CO 2 buffer capacity in the sediment differed strongly from the conditions in the water column.Thereby the sediment chemistry created a microhabitat that supported the growth and development of a benthic foraminiferal community even at highly elevated pCO 2 .Fine-grained sediments with high organic matter content facilitated a high carbonate production of A. aomoriensis.Growth, reproduction and mortality of A. aomoriensis were unaffected by elevated pCO 2 .Consequently, under the current microhabitat conditions, the dominant Ammonia aomoriensis could maintain an important role in benthic carbonate production and accumulation in the southwestern Baltic Sea.However, at high pCO 2 and slight undersaturation of Ω calc ,

Conclusions References
Tables Figures

Conclusions References
Tables Figures

Back Close
Full   Full  Full  Full  Full  Full  Full  Full  Full  Full  Full Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Seawater temperature, salinity and pH NBS (National Bureau of Standards pH-scale) were measured weekly in each culture vessel.Total dissolved inorganic carbon (C T ), Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | weight per individual increased by 68, 57, 70 and 62 % at a pCO 2 of 430, 907, 1865 Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Wollast, R.: Evaluation and comparison of the global carbon cycle in the coastal zone and in the open ocean, in: The Global Coastal Ocean, edited by: Brink, K. H. and Robinson, A. R., John Wiley & Sons, Havard University, 213-252, 1998.Wulff, F., Ertebjerg, G., Nicolaus, G., Niemi, A., Ciszewski, P., Schulz, S., and Kaiser, W.: The changing pelagic ecosystem of the Baltic Sea, Ophelia Supplement, 4, 299-319, 1986Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Fig. 1 .Fig. 2 .
Fig. 1.Gradient of carbonate chemistry parameters of (A) partial pressure of CO 2 (pCO 2 ), (B) pH NBS , (C) total alkalinity (A T ) and (D) saturation state of calcite (Ω calc ) in the seawater and sediment pore water for 4 pCO 2 -treatments after six months incubation time.

Fig. 3 .
Fig. 3. (A) Population density and (B) abundance of living and dead A. aomoriensis at 4 pCO2-levels during six months incubation time.The symbols represent the mean in three replicates (mean ± standard deviation).Reproduction events are indicated by asterisk and bold R.

Fig. 3 .Fig. 4 .
Fig. 3. (A) Population density and (B) abundance of living and dead A. aomoriensis at 4 pCO 2levels during six months incubation time.The symbols represent the mean in three replicates (mean ± standard deviation).Reproduction events are indicated by asterisk and bold R.

Fig. 4 .Fig. 5 .
Fig. 4. Histogram of size class proportions of living A. aomoriensis during six months incubation.Light grey bars represent relative proportion of 11 size classes ranging from < 50 until > 500 µm in 50 µm intervals.Mean test diameter is displayed by the black diamonds and dashed lines.The number in each figure represents the number of individuals (mean ± standard deviation of three replicates).Reproduction events are indicated by asterisk and bold R.

Fig. 5 .
Fig. 5.Total dry weight of (A) living specimensincluding cytoplasm and (B) empty test of A. aomoriensis in relation to mean test diameter for the 4 tested pCO 2 -treatments.

Fig. 6 .
Fig. 6.Mean dry weight per individual including cytoplasm/empty test and standard deviation of three replicates of (A) living and (B) dead A. aomoriensis of the 4 tested pCO2-levels during six months incubation.

Fig. 6 .
Fig. 6.Mean dry weight per individual including cytoplasm/empty test and standard deviation of three replicates of (A) living and (B) dead A. aomoriensis of the 4 tested pCO 2 -levels during six months incubation.

Table 1 .
Carbonate chemistry parameters of each pCO 2 -treatment in the culture vessels.Total alkalinity (A T ), partial pressure of CO 2 (pCO 2 ) and saturation state of calcite (Ω calc ) were calculated from measured temperature, salinity, pH NBS , total carbon (C T ), phosphate (PO

Table 2 .
List of living foraminifera collected from each culture vessel during six month incubation time, size fraction 63-2000 µ m.

Table 3 .
List of dead foraminifera collected from each culture vessel during six month incubation time, size fraction 63-2000 µm.

Table 4 .
Two-way analysis of variance (ANOVA) revealed the effect of pCO 2 and incubation time on test diameter of A. aomoriensis.Significant results are represented in bold.SS = Sum of Squares, d.f.= degrees of freedom, MS = Mean Squares.