Interactive effects of oxygen, carbon dioxide and flow on photosynthesis and respiration in the scleractinian coral Galaxea fascicularis

The symbiosis between scleractinian corals and phototrophic dinoflagellates (zooxanthellae) has been studied intensively (reviewed in Furla et al., 2005; Weis, 2008; Osinga et al., 2012). Factors reported to influence the efficiency of photosynthetic processes inside corals include light (reviewed in Osinga et al., 2012), temperature (reviewed in Smith et al., 2005), nutrients (reviewed in Osinga et al., 2011) and gas exchange (e.g. Dennison and Barnes, 1988; Finelli et al., 2006). Gas exchange involves both diffusion of the primary substrate for photosynthesis, carbon dioxide, into the coral and diffusion of photosynthetically produced oxygen out of the coral tissue.

Several authors have reported positive effects of water flow on photosynthesis in corals (Dennison and Barnes, 1988; Lesser et al., 1994; Nakamura et al., 2005; Finelli et al., 2006; Mass et al., 2010; Schutter et al., 2011). In a flume experiment with Galaxea fascicularis, Schutter et al. (2011) found an instantaneous increase in photosynthesis upon an increase in flow velocity from 5 to 20 cm s⁻¹, which supports the idea that the effects of flow on photosynthesis are related to an improved exchange of dissolved gases. However, it remains unclear whether this flow effect relates to a higher influx of carbon dioxide into the coral tissue (Dennison and Barnes, 1988; Lesser et al., 1994) or to a higher efflux of oxygen out of the coral tissue (Nakamura et al., 2005; Finelli et al., 2006; Mass et al., 2010; Schutter et al., 2011). Mass et al. (2010) demonstrated that the onset of flow led to a decrease of oxygen supersaturation from 500% to 200% inside illuminated tissue of the scleractinian coral Favia veroni. This decrease in internal oxygen was accompanied by an increase in the rate of net photosynthesis. Mass et al. (2010) concluded that a flow-induced, enhanced efflux of oxygen out of the coral tissue increases the efficiency of photosynthetic carbon fixation by reducing the rate of photorespiration. Photorespiration is a reaction in which oxygen instead of carbon dioxide binds to RuBisCO, the initial carbon dioxide-fixing enzyme in the Calvin–Benson–Bassham (CBB) cycle (Peterhansel et al., 2010). RuBisCO then converts oxygen and ribulose bisphosphate into 2-phosphoglycerate (2PG), instead of converting carbon dioxide and ribulose bisphosphate into 3-phosphoglycerate (3PG). Although 3PG can be regenerated from 2PG, this process requires several reaction steps, which consume more metabolic energy than regenerated 3PG can return after its uptake in the CBB cycle (Peterhansel et al., 2010). Hence, photorespiration decreases the efficiency of the photosynthesis process.

So far, only RuBisCO type II, which is the type of RuBisCO that is most sensitive to photorespiration, has been found in zooxanthellae (Rowan et al., 1996), which supports the view that flow modulates photosynthesis mainly through a reduction in photorespiration. Notwithstanding this, Smith et al. (2005) stated that photorespiration in corals is not likely to play a major role because of the presence of an efficient carbon-concentrating mechanism (CCM) in Symbiodinium (Leggat et al., 1999).

The contradicting views on the effects of dissolved oxygen on photosynthesis and photorespiration in corals and, consequently, on the mechanism that underlies flow-enhanced photosynthesis prompted us to address the following research questions. (1) How does the ambient oxygen concentration in seawater affect photosynthesis in corals? (2) Does water flow stimulate photosynthesis in corals through stimulating the uptake of carbon dioxide or through reducing the negative effects of oxygen? To answer these questions, we studied the effect of flow on net photosynthesis and dark respiration in the stony coral Galaxea fascicularis (Linnaeus 1767) under three different levels of dissolved oxygen (hypoxia, 20% saturation; normoxia, 100% saturation; and hyperoxia, 150% saturation) and two different values for dissolved carbon dioxide (9.5 and 19.1 μmol l⁻¹), as reflected by two corresponding values for pH (8.1 and 7.84). The applied values for pH and oxygen saturation are within the range experienced by corals in nature: natural diel fluctuations in pH (8.7–7.8) and oxygen saturation (27–247%) have been reported for reef flats and tropical lagoons (Ohde and Van Woesik, 1999). Our results suggest that in G. fascicularis, flow stimulates photosynthesis through a more efficient uptake of carbon dioxide rather than through a decrease in photorespiration, whereas respiration is stimulated by flow through an increased uptake of oxygen.

Fragments of a captive-bred specimen of the encrusting coral G. fascicularis (originally obtained under CITES no. 52139) were provided by Burgers' Zoo BV (Arnhem, The Netherlands). All experiments were conducted at Wageningen University (Wageningen, The Netherlands). No approval from an ethics committee was required as scleractinian corals are exempted from legislation concerning the use of animals for scientific purposes in the European Union (Directive 2010/63/EU). Six genetically identical fragments of G. fascicularis were obtained from one parent colony. The fragments were glued to PVC plates of 25 cm², and were allowed to recover from fragmentation for a period of 3 weeks before the start of the experiments. The coral fragments were held in an aquarium of 356 l, wherein flow velocity, temperature and salinity were maintained at constant values (flow velocity: 10–20 cm s⁻¹, temperature 25°C, salinity 35.0 g l⁻¹). Corals were subjected to a 12 h:12 h light:dark cycle at a quantum irradiance (E) of 280 µmol quanta m⁻² s⁻¹. Alkalinity and calcium concentrations were measured 5 times a week and were adjusted when necessary by adding NaHCO₃ and CaCl₂·H₂O, respectively. Total alkalinity was maintained at the current natural oceanic value of 2.4 mequiv. l⁻¹ and calcium at 400 mg l⁻¹. Corals were fed 30 ml of a concentrated suspension of freshly hatched Artemia nauplii (∼3000 nauplii ml⁻¹) on a daily basis. Artificial seawater was made with enriched-blend, phosphate- and nitrate-free Reef Crystals (Aquarium Systems, Sarrebourg, France) and was aerated a few days before use.

Prior to the incubations, the projected surface area of the coral fragments was determined according to Wijgerde et al. (2012a) from top-view pictures of the fragments using the software Image Tool for Windows.

Interactive effects of flow, CO₂ and O₂ on net photosynthesis (P_n) and dark respiration (R_d) were studied by subjecting corals to different combinations of these parameters in a 2×2×3 factorial design for repeated measures. Rates of P_n and R_d were determined in 1500 cm³ Perspex cylindrical incubation chambers using an oxygen evolution technique as described in Schutter et al. (2008). Temperature in the chambers was maintained at 26°C by pumping freshwater of 26.5–27°C (maintained with a TC20 Aquarium Cooler, Teco, Ravenna, Italy) through a water jacket surrounding the cylindrical chamber.

Flow was created in the incubation chambers by a magnetic stirrer, using two different-sized stirring bars to obtain two distinct flow regimes: high flow velocity and low flow velocity. The actual range of velocities within each flow regime was determined by video recordings of neutrally buoyant particles that were moving over the experimental coral fragments perpendicular to the camera. The flow velocity of at least five particles per flow regime was determined by dividing the travelled distance of the particle by the time needed to travel that distance. For the low flow regime, flow velocity within a chamber varied between 1 and 1.6 cm s⁻¹, whereas for the high flow regime, flow velocity within a chamber varied between 4 and 13 cm s⁻¹.

Two values for dissolved CO₂ concentration were applied: 9.5 and 19.1 μmol l⁻¹. The former is representative of current atmospheric and oceanic conditions; the latter represents a doubling of the current atmospheric concentration. Three oxygen saturation levels were applied, 20%, 100% and 150%, which were obtained by flushing seawater with either N₂ gas or pure O₂ gas.

Following Riebesell et al. (2010), it was calculated that at prefixed, oceanic values for total alkalinity (2.4 mequiv. l⁻¹), salinity (35‰) and temperature (26°C; representative for coral reefs), the desired levels of dissolved CO₂ of 9.5 and 19.1 μmol l⁻¹ correspond to pH values of 8.10 and 7.84, respectively. We applied these pH values to experimentally control the concentration of dissolved CO₂ by measurement of pH. Hence, the two values for pH (8.10 and 7.84) will be used in the subsequent text to indicate the two experimental CO₂ treatments.

Before the start of each incubation, water from the aquarium in which the corals were maintained was specifically prepared for that particular treatment in a 10 l⁻¹ bucket. First, the seawater was titrated with a 1 mol l⁻¹ solution of HCl to acquire a pH of 8.10. Subsequently, total alkalinity was measured and NaHCO₃ was added to restore the alkalinity to the current oceanic value of 2.4 mequiv. l⁻¹. The water was then further prepared by addition of CO₂, O₂ and/or N₂ to achieve the particular combination of dissolved gases needed for each treatment, and afterwards the prepared water was carefully poured into the incubation cells.

During preparation of the water, the O₂ saturation was monitored by measuring the dissolved O₂ concentration with an optical oxygen probe (LDO101, HQ 40d multi, Hach-Lange GmbH, Düsseldorf, Germany; calibrated prior to use in water-saturated air) and the CO₂ concentration was measured indirectly by measuring pH in the water sample. All measurements of seawater pH were done with a pH probe (Hach) that was calibrated using two NBS pH calibration fluids (WTW, Weilheim, Germany) representing a pH of 7.0 and a pH of 10.0, which were supplemented with 25.4 g l⁻¹ NaCl (or 10 g l⁻¹ Na⁺), to approximate the ionic strength of seawater at a salinity of 35 g l⁻¹ (corresponding to 10 g l⁻¹ Na⁺; Wijgerde et al., 2014).

Corals were placed in the incubation chambers immediately after filling the chambers with the customised seawater, after which the chambers were sealed airtight by tightening the lid. Two corals were measured at the same time, in parallel with one control (incubation cell only containing seawater) to correct for background photosynthesis/respiration or possible chemical water composition changes caused by the addition of the gases. Control values were subtracted from incubations with corals.

To avoid light limitation effects, P_n was measured at an E of 500 µmol quanta m⁻² s⁻¹, which is considered to be saturating for this clone of G. fascicularis (Schutter et al., 2008). Aquablue Spezial T5 fluorescent bulbs (ATI, Hamm, Germany) were used to ensure a good light spectrum. E was measured with a Li-1400 light meter (Li-Cor, Lincoln, NE, USA). P_n and R_d were assessed by measuring the change in dissolved O₂ concentration in the incubation cells by optical oxygen probes (Hach) that had been calibrated prior to use in water-saturated air (see above). O₂ concentration was measured every minute for a period of 80 min. After an acclimation period of 10 min, a subsequent period of 30 min was applied to determine P_n. Then, the lights were turned off and the incubation cells were darkened to measure R_d. R_d was assessed by measuring the dissolved oxygen concentration for 30 min after a second acclimation period of 10 min. Throughout the incubations, oxygen saturation levels never deviated more than 5% from the initial values. These small deviations did not notably affect P_n or R_d, as regression coefficients for the observed increases or decreases in oxygen over time were always above 0.99.

In total, 72 measurements were performed using six replicate corals, which were subjected to each of the 12 treatments. The sequence of the experiments was randomised using Microsoft Excel. By using such a repeated measures design, variability due to differences in size and shape of the experimental colonies could be minimised.

An additional experiment was performed in which a more extreme O₂ saturation was chosen to further elucidate the effects of O₂ on P_n. The same six experimental colonies of G. fascicularis were subjected to both 100% O₂ saturation and 280% O₂ supersaturation, which resembles the daily maximum of 247% reported by Ohde and Van Woesik (1999) for Indo-Pacific reef flats. Other experimental conditions were identical to those of the previously described experiment.

Data were analysed in SPSS Statistics version 19 (IBM, Armonk, NY, USA). Data were checked for normality (Shapiro–Wilk's test) and sphericity (Mauchly's test; only applied to data regarding different oxygen saturation levels) to test whether the assumptions for ANOVA were met. All data for R_d and P_n were normally distributed based on the skewness and kurtosis of residual histograms. The assumption of sphericity was met for oxygen saturation (P>0.050). A three-way factorial ANOVA for repeated measures was applied to analyse both data series. Significant interactive effects were followed up by simple contrasts, to determine the effect of a given factor at various level combinations of the remaining two factors. The supersaturation control experiment was analysed by a paired t-test. P<0.05 was considered statistically significant.

Average rates of R_d and P_n under the 12 experimental treatments are presented in Fig. 1. Table 1 shows the corresponding results of the three-way ANOVA. Data for R_d are graphically presented as negative oxygen production (i.e. oxygen consumption). Throughout the text, absolute values will be used to express rates of R_d.

Values for R_d ranged between 0.050 and 0.087 mg O₂ cm⁻² h⁻¹). There was a significant main effect of flow velocity on dark respiration (Table 1). No significant main effect of pH on dark respiration was observed. Oxygen saturation, in contrast, did have a significant main effect on R_d rates (Table 1). A significant interaction was found between the effects of oxygen and flow velocity on R_d in G. fascicularis (Table 1). Under hypoxic (20% O₂) conditions, flow velocity had a strong and significant effect on R_d (42% increase, P=0.008 at low pH and 43% increase, P=0.014 at high pH). Under normoxic conditions (100% O₂), flow also significantly enhanced R_d (16% increase, P=0.004 at low pH and 30% increase, P=0.035 at high pH). The flow effect was not significant under hyperoxic (150% O₂, P=0.388 and P=0.874, respectively) conditions. No three-way interaction was found (Table 1).

Values for P_n in G. fascicularis (Fig. 1) ranged between 0.178 and 0.256 mg O₂ cm⁻² h⁻¹, the rates of P_n being 2–5 times higher than the rates of R_d. Flow and pH exerted a significant main and interactive effect on P_n (Table 1). At high flow, there was a significant effect of pH on P_n (14% increase at pH 7.84; P=0.026), whereas the pH effect was not significant at the low flow velocity (P=0.113). Conversely, the effect of flow was significant (P=0.001) at pH 7.84, but not at pH 8.10 (P=0.066). The interaction is graphically depicted in Fig. 2. In contrast, no main effect of oxygen saturation on P_n was found and no interactive effect of oxygen with the other factors (flow, pH) could be observed (Table 1). Furthermore, in the supersaturation control experiment, the additional oxygen (280% supersaturation) did not significantly affect the rate of P_n (paired t-test, P>0.05), the rates being 0.17±0.0191 and 0.174±0.026 mg O₂ cm⁻² h⁻¹ (means±s.d.) for 100% and 280% saturation, respectively.

Our data show that photosynthesis in this particular genotype of G. fascicularis is not affected by ambient oxygen levels. Changes in oxygen saturation (both hypoxia and hyperoxia, up to 280% supersaturation) did not influence the P_n output of G. fascicularis under saturating light. As we will outline below, this result implies that photorespiration is not of quantitative importance in this genotype of G. fascicularis. A conceptual picture (Fig. 3) illustrates how we deduced this conclusion from our results.

Oxygen produced inside the coral tissue through photosynthesis (gross oxygen production) has three potential sinks: (1) release into the external environment via diffusion (oxygen evolution, also termed P_n); (2) respiration, a process referred to as light respiration under illuminated conditions and dark respiration in darkness; (3) photorespiration. Our results show that P_n in G. fascicularis is not affected by ambient oxygen. Hence, an increase in light respiration and photorespiration under high ambient oxygen, as found in Favia veroni (Mass et al., 2010), would only be possible if gross photosynthesis increased as well (Fig. 3B, scenario 1). An increase in gross photosynthesis under high oxygen is not likely, as photosynthetic oxygen production is not limited by oxygen. In contrast, very high oxygen levels (>800% supersaturation) may even decrease the photochemical yield of the zooxanthellae (Finelli et al., 2006), thus leading to a lower gross productivity. Therefore, it seems reasonable to assume that gross photosynthesis was not affected by ambient oxygen, which excludes scenario 1 in Fig. 3. With both gross and net photosynthesis remaining unchanged, it is not very likely that light respiration and photorespiration changed under the influence of ambient oxygen. Both processes would in that case have to show an opposite response (scenarios 2 and 3, Fig. 3C,D). We conclude from this analysis that photorespiration is not quantitatively important in G. fascicularis (scenario 4, Fig. 3E), as otherwise we should have observed a decrease in P_n at high ambient oxygen.

This conclusion is in agreement with work by Muscatine (1980), who found no effect of oxygen on photosynthesis by zooxanthellae in hospite. Zooxanthellae in hospite may not be very susceptible to photorespiration because of efficient CCMs that operate within the coral–zooxanthellae symbiosis (Smith et al., 2005). Notwithstanding this, ex hospite photosynthesis of zooxanthellae isolated from Pocillopora damicornis was very sensitive to oxygen (Downton et al., 1976), which suggests that the CCM may be largely host related. Indeed, recent work by Tansik et al. (2015) showed that in several coral species, the dissolved inorganic carbon (DIC) influx into the coral tissue is host controlled. These corals exhibit external carbonic anhydrase activity (i.e. bound to the outer cell membrane of the coral ectoderm cells) that mediates the influx of carbon dioxide into the coral tissue. The G. fascicularis–Symbiodinium symbiosis has both a carbonic anhydrase-driven CCM and transport pathways for HCO₃⁻ (Al-Moghrabi et al., 1996; Goiran et al., 1996), thus enabling an efficient supply of DIC to the site of photosynthesis.

The contrast between our results and the study by Mass et al. (2010) shows that effects of oxygen on coral photosynthesis are species specific. Sensitivity to high oxygen levels may depend on the effectiveness of the CCM in the coral. In contrast to Favia veroni (Mass et al., 2010), G. fascicularis apparently has a highly efficient CCM, making photosynthesis in this species rather insensitive to changes in ambient oxygen.

Interestingly, oxygen supersaturation does affect calcification in this genotype of G. fascicularis. A 20–33% decline in light calcification rates was found when colonies of this genotype were exposed to oxygen saturation levels of 150% and higher (Wijgerde et al., 2012a). The authors attributed this decline to the formation of reactive oxygen species (superoxides) under high oxygen levels. The differential response of photosynthesis (this study) and calcification (Wijgerde et al., 2012a) to high oxygen may relate to the spatial separation of the two processes. Calcification occurs in the ectodermal layer, whereas photosynthesis takes place within zooxanthellae located in the endodermal layers of the coral (Jokiel, 2011). Endodermal layers have been shown to exhibit higher superoxide dismutase (SOD) activity (Richier et al., 2003). SOD can counteract the toxic effects of superoxides. The difference in sensitivity to oxygen between calcification and photosynthesis might thus be explained by a difference in defence possibilities against superoxide toxicity between ectodermal and endodermal cells.

In contrast to dissolved oxygen, the supply of additional CO₂ into the seawater did stimulate P_n in G. fascicularis, an effect that was significant only under the high flow regime of 4–13 cm s⁻¹. This finding suggests that the often-reported positive effect of flow on coral photosynthesis is, in the case of G. fascicularis, due to a flow-induced increase in the uptake of CO₂ rather than an increased efflux of oxygen out of the coral tissue.

Increased influx of DIC under high flow has previously been suggested to stimulate photosynthesis in the symbiotic coral species Acropora formosa (Dennison and Barnes, 1988). These authors based their suggestion upon an observed increase in photosynthesis under stirred conditions versus non-stirred conditions. As no effect of stirring was observed at the compensation point (i.e. no net gas exchange), the observed effect was ascribed to diffusion limitation. Dennison and Barnes (1988) concluded that diffusion limitation of CO₂ rather than inhibition by accumulation of oxygen was responsible for the flow effect on photosynthesis. Their conclusion was based on a preceding study (Crossland and Barnes, 1977) on a related coral species, in which no photosynthetic response to varying oxygen levels had been found. Our results provide direct evidence for this limiting role of CO₂ diffusion, as we studied the effects of CO₂ and oxygen simultaneously on the same species.

Nevertheless, our results contradict with other literature. According to Tansik et al. (2015), coral photosynthesis is not likely to be DIC limited because of the highly efficient external carbonic anhydrase system. This assumption is in line with several other studies (e.g. Reynaud et al., 2003; Schneider and Erez, 2006; Marubini et al., 2008) in which no beneficial effects of P_CO2 on coral photosynthesis were found.

The apparent contrast with our observations probably relates to the specific combination of abiotic parameters that was applied. In our study, light availability was more than twofold higher than in the study by Tansik et al. (2015), at a level that was saturating for photosynthesis (see Schutter et al., 2008), which may thus have invoked DIC limitation. The other studies (Reynaud et al., 2003; Schneider and Erez, 2006; Marubini et al., 2008) used higher irradiance levels than Tansik et al. (2015), but those studies do not provide details on the flow conditions that were applied in the experiments. As flow can augment the P_CO2 effect as observed in this study, incubations performed under low flow conditions may not show detectable effects of P_CO2 on photosynthesis. Additionally, the studies by Marubini et al. (2008) and Reynaud et al. (2003) lasted for periods of 1 and 5 weeks, respectively, whereas the current experiment investigated the acute effects (30 min exposure) of CO₂. Possibly, there is an initial effect of CO₂ that is later countered by adaptation of the coral. For example, coral may adapt to a higher P_CO2 by lowering carbonic anhydrase activity (Lesser et al., 1994), leading to a lower efficiency of the CCM.

There is one earlier study that describes effects of dissolved inorganic carbon on P_n in G. fascicularis (Goiran et al., 1996), in which an optimal pH of 8.8 was reported for P_n. Our study disagrees with this result for reasons that remain to be investigated.

We conclude that our study represented a situation in which the influx of DIC limited coral photosynthesis. DIC limitation of photosynthesis may occur in shallow water where light is saturating (Osinga et al., 2011). Our results demonstrate that under such conditions, high flow and increased levels of carbon dioxide can alleviate DIC limitation and stimulate coral photosynthesis.

Changes in P_CO2 and corresponding changes in pH did not change rates of R_d in G. fascicularis. Apparently, a decrease in ocean pH to 7.84 does not quantitatively affect catabolism in this coral.

Flow and ambient oxygen saturation level both stimulated R_d of G. fascicularis, in an interactive way. The positive effect of flow on R_d found here is in line with earlier research (Patterson et al., 1991; Sebens et al., 2003), which showed that flow enhances R_d in corals within the range of flow velocities applied in our study (1–13 cm s⁻¹). The significant positive main effect of oxygen saturation on R_d supports the often-proposed theory (e.g. Chalker and Taylor, 1975; Rinkevich and Loya, 1984; Ip et al., 1991; Colombo-Pallotta et al., 2010; Wijgerde et al., 2012a) that dark calcification in corals is limited by a lack of metabolic energy due to R_d being oxygen limited. Both flow and ambient oxygen saturation influence the availability of oxygen inside the tissue by affecting diffusion rates. The effect of flow was significant under hypoxic and normoxic conditions, but not under hyperoxic conditions. This indicates that R_d is no longer diffusion limited under such a high availability of dissolved oxygen. An earlier study on the effect of flow on R_d in G. fascicularis under normoxic conditions showed that flow rates above 10 cm s⁻¹ did not further promote R_d (Schutter et al., 2010), indicating a similar boundary to the relief of diffusion limitation by flow.

The rates of P_n measured in this study are among the highest ever reported in scleractinian corals (Table 2). In addition, light utilisation was very efficient. At the highest rate of P_n measured in this study, 23 quanta were used per molecule of eluted oxygen, an efficiency only exceeded by shade-adapted cave corals (Anthony and Hoegh-Guldberg, 2003). It should be noted here that our corals were grown at an E of 200 µmol quanta m⁻² s⁻¹ and incubated at an E of 500 µmol quanta m⁻² s⁻¹. Hence, one might argue that the observed rates of P_n reflect a stress response. However, the values for P_n measured in this study correspond well to values within photosynthesis:irradiance (P:I) curves determined earlier for the same clone of G. fascicularis (Schutter et al., 2008). The shape and values of those earlier P:I curves are in good agreement with P:I curves measured in the field for several coral species, thus indicating that the values reported here reflect normal photosynthetic responses to the light levels that were applied.

Under all treatments in this study, photosynthetic oxygen production by G. fascicularis was more than 2 times higher than its consumption of oxygen in darkness. This indicates that there was a surplus of photosynthetically derived organic carbon available for G. fascicularis and its zooxanthellae under all experimental circumstances. Such a positive carbon balance implies that an organism has scope for growth (SfG). SfG is defined as the amount of organic carbon (acquired autotrophically and/or heterotrophically) that is available to an organism after subtracting the daily respiratory demand and excretory losses (Anthony and Fabricius, 2000). Subtracting daily R_d from daily P_n gives SfG′, the amount of photosynthetically derived organic carbon that is available to the coral holobiont for growth and excretion. As an example, we calculated SfG′ for the experimental treatment with the highest P_n (0.256 mg O₂ cm⁻² h⁻¹). This value was measured at 150% O₂ saturation, 4–13 cm s⁻¹ and pH 7.84. A corresponding value for R_d of 0.084 mg O₂ cm⁻² h⁻¹ was measured under these conditions. Taking into account the 12 h:12 h light:dark cycle that was applied in our study, and assuming a molar ratio of 1:1 for the number of carbon molecules fixed per molecule of oxygen eluted, daily P_n could be estimated at 1.15 mg C cm⁻² and daily R_d at 0.38 mg C cm⁻². These values yield a SfG′ of 0.77 mg C cm⁻². For comparison, a SfG′ of 0.13 mg C cm⁻² day⁻¹ was calculated for Porites cylindrica and 0.06 mg C cm⁻² day⁻¹ for Goniastrea retiformis (Anthony and Fabricius, 2000). Those values are an order of magnitude lower than the values found for G. fascicularis in the current study. Hence, we conclude that under the experimental circumstances applied, G. fascicularis and its zooxanthellae had a high surplus of carbon available for processes such as somatic growth, reproduction and repair of tissue damage. Indeed, growth rates reported for G. fascicularis are very high, specific growth rates sometimes exceeding 0.020 day⁻¹ for young colonies (Wijgerde et al., 2012b). These high initial growth rates were obtained at low levels of E of 40–60 μmol quanta m⁻² s⁻¹, which confirms that G. fascicularis is very efficient in collecting light. Similarly, Schutter et al. (2008) reported an initial specific growth rate for G. fascicularis of 0.025 day⁻¹ at an E value of 90 μmol quanta m⁻² s⁻¹. In addition, they found that G. fascicularis was still capable of growing with a specific growth rate of 0.006 day⁻¹ at a level of E as low as 39 μmol quanta m⁻² s⁻¹.

The results obtained in this study show that G. fascicularis has a high net productivity that is insensitive to the ambient level of dissolved oxygen. Photorespiration is not likely to play a quantitatively important role in this coral. The high productivity of G. fascicularis leads to a saturated light respiration and a positive SfG. Under saturating light, photosynthesis in G. fascicularis may become DIC limited, as shown by the positive, interactive effects of flow and pH on net photosynthesis.

We thank the personnel of De Haar Vissen for logistic support and Marleen ten Napel for executing preliminary studies preceding this work.