The stability of pH values and similarity for pH observed among the treatments could be due to the fact that seawater is strongly buffered. The sand-substrate acts as an internal biofilter and the biofilter started functioning sooner in the R system; therefore, alkalinity in the R system started to decline soon after the first month of the experiment (April) and rapidly declined in May and June; as a result, the alkalinity was significantly lower in the R system as compared to the F system in April. In contrast, alkalinity in the F system kept almost stable during the first two months of the experiment and just started to go down during the last two months of the experiment because the high flow-through rate of new sea-water in the F system slowed down the sand substrate to function as biofilter during the first two months. Since the biofilter in the F system only started to function in May, no significant differences in alkalinity were observed between treatment FL and both dietary treatments in system R during May and June, while there were significant differences in alkalinity only between the treatment FC and treatment RC during May and then between the treatment FC and both dietary treatments in system R during June. This may be due to higher nitrification process in the treatment diet L compared to diet C. Ammonia levels were lower and more stable in the recirculation system with low-water exchange (system R) than in the flow-through water system (system F). This demonstrates that the biofilter function was much better in the system R than in the system F.