The percentage dietary protein had a significant effect on the concentration of FAN at each tank position. Thus, within the range of protein levels tested an increase in dietary protein level of 21% resulted in 50% higher FAN concentrations. In H. midae, growth has been shown to be independent of dietary protein, provided that dietary energy is maintained above 13.5 MJ kg−1 and that it is provided mostly in the form of carbohydrates (Green et al. 2011). The diets used here conformed with these requirements, in that energy levels were 15.6 MJ kg−1 and lipid content was low (3%). There was also no difference in growth between the two diets for position 1 where the flow indices were above those shown to cause reductions in growth as a function of water quality (Naylor et al. 2011). Thus, it is hypothesized that differences in growth were the result of water quality or environmental conditions.
Despite being exposed to lower FAN concentrations, abalone in the P26 treatment did not gain weight or length faster at any of the positions within the serial-use raceways. However, the reduction in growth as a result of increasing water use for both dietary treatments was more pronounced in abalone fed the high-protein diet from the second to the third tank in the series. It is hypothesized that a low-protein diet may become more beneficial as the intensity of water use increases. Abalone are sensitive to FAN, and have demonstrated a large reduction in growth rate (Harris et al. 1998; Basuyaux & Mathieu 1999; Huchette et al. 2003; Reddy-Lopata et al. 2006) well below the 50–200 μg L−1 safe limit described for many marine finfish species (Russo & Thurston 1991; Person-Le Ruyet, Galland, Le Roux & Chartois 1997; Lemarie, Dosdat, Coves, Dutto, Gasset & Person-Le Ruyet 2004). Although no information is available regarding the EC5 for FAN in H. midae, Reddy-Lopata et al. (2006) determined that the EC50 for FAN in juvenile H. midae was <7 μg L−1 and speculated that the EC5 would be lower than the 4 μg L−1 as was reported by Huchette et al. (2003) for H. rubra. As there was no difference in growth between the P26 and P33 treatments at position 3, where FAN concentrations were 1.0 ± 0.9 and 2.4 ± 1.3 μg L−1, respectively, suggests that FAN concentration in the P33 treatment was below the level that can significantly reduce growth. This may be due to the larger size class used (69.5 ± 2.5 mm versus 10–25 mm used by Reddy-Lopata et al. 2006), as larger abalone are less sensitive to high concentrations of FAN (Fallu 1991; Reddy-Lopata et al. 2006). Other environmental conditions, especially pH (Thurston et al. 1981), may also influence the estimation of EC5. As the average FAN concentration could not explain differences in growth between tank positions, other water quality variables may have influenced growth.
Using predictions by Britz, Hecht and Mangold (1997), the decrease in temperature between position 1 and 3 accounts for only 11–12% of the difference in weight gain observed and therefore may have influenced growth, but this was not likely the most limiting variable. Similarly, mean dissolved oxygen concentration (7.3–7.5 mg L−1 O2) and percentage oxygen saturation (90.5–92.8%) were similar between positions and treatments and are unlikely to have caused the differences in abalone growth. This was supported by the low correlation between oxygen concentration and growth, and as oxygen concentration, within the range tested here, did not contribute to a predictive model. The EC50 for dissolved oxygen in H. laevigata has been estimated at 5.91 mg L−1 and 77% saturation (Harris, Maguire, Edwards & Johns 1999). Although nitrite–N concentrations increased with increasing serial-use, reaching a maximum of 34.1 μg L−1, these values were below those shown to affect the growth in other abalone species. Harris, Maguire, Edwards and Hindrum (1997) recorded growth reductions in H. laevigata at nitrite–N concentrations above 540 μg L−1, whereas Basuyaux and Mathieu (1999) estimated safe concentrations at <5000 μg L−1 in H. tuberculata L. There was, however, a significant correlation between the reduction in growth and the decrease in water pH. In addition, water pH contributed the most to a multiple regression model to predict growth as a function of water quality. The hypothesis that low pH caused a reduction in growth is supported by the negative effects of chronic hypercapnia on growth rate and shell dissolution shown in echinoderms, gastropods and bivalves (Shirayama 2002; Michaelidis et al. 2005). In this study, the decrease in pH was most likely caused by the respiration of the abalone and subsequent release of carbon dioxide. As carbon dioxide hydrates in water, it dissociates into aqueous carbon dioxide (CO2 aq), carbonic acid (H2CO3), bicarbonate (HCO3 −), carbonate (CO3 −2) and hydrogen ions (H+), in a pH-equilibrated reaction (Smith 1988). The lower the pH, the greater the proportion of CO2 aq, which could lead to reduced CO2 diffusion at the gills and subsequent blood acidosis and altered oxygen-haemocyanin affinity (Sanni & Forsberg 1996; Harris, Maguire, Edwards & Hindrum 1999). In H. corrugata (Wood), Burnett, Scholnick and Mangum (1988) showed a decrease in haemocyanin oxygen affinity with a decrease in pH. In the marine mussel, Mytilus galloprovincialis, exposure to pH 7.3 caused a significant reduction in haemolymph pH and metabolic rate and increased haemolymph CO2-concentrations, when compared to controls held at pH 8.05 (Michaelidis et al. 2005). A decrease in pH also reduces the availability of essential CO3 −2 ions (Kleypas et al. 2006; Lopez et al. 2011) and it lowers the saturation state of calcite and aragonite, which are calcium carbonate minerals essential for the formation of skeletal structures and shell in many marine organisms (Feely, Sabine, Lee, Berelson, Kleypas, Fabry & Millero 2004). In the greenlip abalone, Haliotis laevigata, the EC5 and EC50 values for pH have been estimated as 7.78 and 7.39, whereas in the blacklip abalone, Haliotis rubra, EC5 and EC50 values were 7.93 and 7.37 (Harris, Maguire, Edwards & Hindrum 1999). The pH values recorded in this study were therefore within the range known to affect other abalone species. They were also well below the mean pH recorded by Yearsley (2008) for morning (pH 8.06) and afternoon (pH 8.13) samples of influent water quality for this specific farm during 2007, representing the pH levels of this species' natural environment.