The world oceans are vast reservoirs of carbon dioxide (CO₂) (Feely et al., 2004; Sabine et al., 2004). The atmospheric CO₂ concentration has increased rapidly over past decades due to the burning of fossil fuels and other anthropogenic activities (Intergovernmental Panel on Climate Change, 2007). Dissolution of CO₂ in seawater shifts the carbonate equilibrium, increasing the H⁺ ion concentration (i.e., pH) and decreasing the CO₃ ^2- concentration. Increased pCO₂ has led to a 0.1 unit decrease in surface ocean water pH over the past 200 years (Caldeira and Wickett, 2003) and it is projected to decrease by 0.3 to 0.4 units by 2100 (Caldeira and Wickett, 2003; Raven et al., 2005). Reductions in seawater pH have detrimental effects on the development and reproductive processes of many marine organisms (Portner et al., 2004, 2005; Raven et al., 2005). Also, increases in CO₂ have been suggested to not only affect individuals but also the entire living ecosystem (Widdicombe and Spicer, 2008).

Mimicking ocean acidification (OA) has been difficult to study its effects on ecosystems and animals, although several experimental systems have been designed to estimate the effects of OA. Some groups have designed microcosm systems in which concentrated CO₂ was manipulated by air-CO₂ gas mixing systems (Findlay et al., 2008; Arnold et al., 2009; Egilsdottir et al., 2009). Other groups have used different molecular acids (H₂SO₄) to control pH manually, with pH being monitored once a day (Zalizniak et al., 2009). The designs of the pH-controlling systems noted above either require large budgets and space or well-trained workers to obtain reliable data. In the present study, we constructed a pH controller that was a simple, cost-effective, and automatically operated system for maintaining water pH accurately.

Until now, studies on the effects of OA have mainly focused on the growth and development of calcareous marine organisms, including corals (Reynaud et al., 2003; Langdon and Atkinson, 2005), mollusks (Michaelidis et al., 2005; Ellis et al., 2009), echinoderms (Dupont et al., 2008; Havenhand et al., 2008), copepods (Kurihara et al., 2004), and amphipods (Egilsdottir et al., 2009). Since crustaceans are considered one of the most vulnerable groups because of their dependence on the availability of calcium and bicarbonate ions for the mineralization of their exoskeleton. The effects of lower pH (either from direct exposure to HCl or CO₂) have been noted on the survival and growth of shrimps Penaeus monodon (Allan and Maguire, 1992), Penaeus borealis (Bechmann et al., 2011), Palaemon pacificus (Kurihara et al., 2008), and Homarus americanus (Arnold et al., 2009). Little attention has been paid to the synergistic effects of salinity and CO₂-mediated pH on hatching rates for elucidating the effects of CO₂ on coastal and estuarine ecosystems.

In the present study, we developed a cost-effective pH-control system that can maintain preset seawater pH by automatically regulating CO₂ injections. This system was reliable for various animal experiments related to OA. Once the reliability of the system had been confirmed, four different pH treatments (7.0, 7.3, 7.6, and control) combined with three salinity levels (15, 25, and 35 ppt) were maintained and the hatching rate of Artemia was measured.

Commercially available Artemia franciscana cysts (Red Fish brand; Golden Sea Aquatic Products Co., Ltd ., China) were purchased and stored at 4℃ until use. A. franciscana is considered the ideal crustacean model due to the ease of culture and maintenance and its short life cycle of 3-4 weeks.

A CO₂ control system was constructed by assembling a CO₂ gas tank, pH sensors, a pH monitor, and an automatic pH controller. Fig. 1 depicts how different pH levels were maintained over the experimental period by the automatic pH controller. The pH level was set at the onset of an experiment and a pH sensor was placed in each hatching cylinder (1 L). Another sensor from the pH controller was linked to the CO₂ container and regulated the amount of CO₂ injected into the water. A solenoid valve connected to the pH controller allowed the CO₂ input to be switched on and off automatically to achieve a constant pH, which could be checked with the pH monitor.

An experiment was conducted to investigate the effects of pH changes that were induced by CO₂ and salinity on Artemia hatchings. For this purpose, three pH treatments were chosen:

7.0, 7.3, and 7.6. Untreated seawater was used as a control (pH 8.0 ± 0.1). Three replicates were conducted for each experimental treatment. One gram of Artemia cyst was used for each pH treatment. Cysts were incubated in 500 mL of seawater in a 1,000 mL glass cylinder with continuous vigorous aeration. Salinity was measured by a temperature corrected refractometer (YSI Inc., Swedesboro, New Jersey, USA ) and salinity was adjusted by adding deionized water to autoclaved stock seawater solution. Temperature was maintained at 27 ± 0.5℃ by a thermostat and a light intensity of 2000 Lux was supplied. The hatching rate was assessed after 24 h using the counting method of the Laboratory of Aquaculture & Artemia Reference Center, Ghent, Belgium. With a Pasteur’s pipette, 50-mL aliquots were taken from each cylinder. Hatching rate was determined as the number of hatched cysts versus unhatched cysts or partially hatched cysts × 100 (Clegg and Conte, 1980). Hatching rates were evaluated statistically by t-tests using the Sigma plot program and values of P < 0.05 were considered statistically significant.

To confirm the reliability of the pH-control system as described above, blanks without Artemia were used. After 2 days of experimental operation, salinity and pH were still being adjusted properly. pH and temperature were maintained at 8.00 ± 0.1 and 27 ± 0.5℃, respectively, indicating that the system could be used to perform animal tests of OA (Fig. 1).

After 24 h of cyst incubation, CO₂-mediated pH changes had significant effects on the hatching rate of Artemia. The

maximum hatching rate was observed with the control pH (~8.0) and decreased significantly to the lower pH (7.0), irrespective of salinity (Fig. 2A-2C). At a salinity of 15 ppt, the mean hatching rate decreased to 17% with pH 7.0, whereas it was 46% with the control pH (~8.0). Both the 25 and 35 ppt salinity treatments exhibited similar patterns of decreased hatching rates, dropping to 22% and 18%, respectively, in the lowest pH (7.00) treatment. Hatching rate decreased by 2.7, 2.5, and 2.9 times compared to the control pH with salinity levels of 15, 25, and 35 ppt, respectively (Fig. 2D). A study on the effects of H₂SO₄ acidified seawater on the hatching and survival of A. franciscana at 10 different pH levels (4 to 8.5) showed that the hatching rate was highest at pH 7.3 and decreased significantly at either lower or higher pH levels (Doyle and McMahon, 1995). In contrast, we found the highest hatching rate with the control pH (~8.0) and the rate decreased significantly as pH was reduced to 7.0. The difference between the two results may have been due to the different methods used to lower pH. Previous experiments used mineral acid (HCl), but in this study, we used molecular CO₂ to alter the pH. The effects of CO₂/HCl acidified seawater were observed on the fertilization rate of two sea urchins, Echinometra mathaei and Hemicentrotus pulcherrimus (Kurihara and Shirayama, 2004); the fertilization rate of eggs decreased linearly in high pCO₂ seawater, but it only decreased at pH < 7.0 in HCl acidified seawater. The reason for this difference was likely the diffusion capacity of CO₂ and protons. Molecular CO₂ diffuses directly through the biological cell membrane faster than protons (Gutknecht et al., 1977); hence it causes a faster decrease in the intracellular pH of eggs compared to HCl/H₂SO₄. The low intracellular pH may prevent fertilization and subsequent embryonic development (Kurihara, 2008). Both the hatching rate and nauplius survival in two marine copepods (Acartia steueri and Acartia erythraea) were significantly reduced at pH 6.8, which was the lowest level observed. Similarly, early development in the oyster Crassostrea gigas can be delayed because of exposure to acidified seawater with pH 7.4 (Kurihara et al., 2007). However, the survival, growth, morphology, and development of Acartia tsuensis eggs were unaffected at all stages when reared under 2000 uatm pCO₂, which is equivalent to pH 7.3 (Kurihara and Ishimatsu, 2008).

In this study, the hatching rate of A. franciscana was affectphotoed by reduced pH regardless of salinity and no notable synergistic effects were observed. Embryonic development in the amphipod Echinogammarus marinus was investigated using CO₂ acidified seawater with different salinity levels; low pH (7.5) resulted in prolonged embryonic development regardless of salinity, but reduced salinity, not lower pH, had a significant effect on the number and calcium content of hatchlings (Egilsdottir et al., 2009). The highest hatching percentage in horseshoe crab, Tachypleus gigas, was observed with salinity levels from 25 to 35 ppt, but no hatched nauplii were found at salinity levels of 15 and 20 ppt (Zaleha et al., 2011). Similar findings regarding hatching rate were observed in A. urmiana and parthenogenetic Artemia (Asem and Rastegar-Pouyani, 2010). These results support our findings that 25 ppt resulted in the highest hatching rate.

In conclusion, we constructed a cost-effective pH-control system involving automatic CO₂ injection that could be used to study the effects of CO₂-induced reductions in pH on marine animal physiology. We also demonstrated that pH and salinity changes had independent effects on the hatching rate of A. franciscana. This suggests that organisms spawned in coastal regions and eggs drifted from deep seas to coastal areas, where salinity and pH changes occur frequently, may be challenged by the adverse effects of OA. However, synergistic effects of salinity and pH may not occur and would be influenced more by species-specific traits of physiology.