SHORT COMMUNICATION Leaching of nutrients from Australian sardines used as feed for southern bluefin tuna farming Milena Fernandes Aquatic Sciences Centre, South Australian Research & Development Institute and Aqua¢n CRC, Henley Beach, Australia Correspondence: M Fernandes, Aquatic Sciences Centre, South Australian Research & Development Institute and Aqua¢n CRC,PO Box 120, Henley Beach, SA 5022, Australia. E-mail: fernandes.milena@saugov.sa.gov.au Small or low-price ¢sh (‘bait¢sh’) are still widely used as aquaculture feed in Asia, and for the farming of large carnivorous species such as tuna (Wu 1995; Leung, Chu & Wu 1999; Aguado-Gime¤ nez & Garc|¤a-Garc|¤a 2005; Fernandes, Lauer, Cheshire & Angove 2007). In comparison with pelletized feeds, bait¢sh present the disadvantage of having a lower nutritional value and a high proportion of soluble nutrients (Qian, Wu & Ni 2001; Nakada 2002). In a previous study, we investigated the release of ammo- nium, nitrate/nitrite, total nitrogen and phosphate from Australian sardines ( Sardinops neopilchardus ) into arti¢cial seawater (Fernandes, Angove, Sedawie & Cheshire 2007). This species is used as feed in the farming of southern blue¢n tuna ( Thunnus maccoyii ), the second most valuable aquaculture industry in Australia (ABARE 2007). The use of10^20 mm pieces of Australian sardines for assessing nutrient leach- ing rates represented the main shortfall of this earlier work as a substantial fraction of uneaten bait- ¢sh might sink whole through the water column, in this way minimizing nutrient release. Here, I address this gap by investigating the leaching of nutrients from whole Australian sardines into natural sea- water, and extend the range of analysis to include total phosphorus. Although the results were in- tended for the tuna industry, a similar situation is likely to occur in other aquaculture industries using bait¢sh, and therefore the data are applicable beyond tuna farming. In this work, I used frozen Australian sardines ob- tained from a commercial tuna operator.To assess the initial water and nutrient contents, 12 ¢sh were freeze-dried and homogenized. The nitrogen content was determined by continuous £ow stable isotope ratio mass spectrometry using either a ANCA-SL ele- mental analyser coupled to a Geo 20-20 mass spec- trometer (Europa Scienti¢c, Crewe, UK), or a EA-CN preparation system coupled to a Hydra 20-20 mass spectrometer (SerCon, Crewe, UK). Phosphorus con- tent was determined in aVarianVista Axial ICP-AES (US Environmental Protection Agency1991). To assess nutrient leaching rates, experiments were conducted in a controlled environment room where temperature was maintained at 18.6 Æ 0.3 1C. In South Australia, tuna farming occurs in water temperatures from12 to 22 1C, with a seasonal aver- age of 18 1C (Fernandes, Angove et al . 2007). Seawater used in the experiments was ¢ltered through an iso- pore polycarbonate ¢lter (Millipore,0.4 mm) and kept in the controlled environment room overnight. Four bait¢sh were thawed overnight in a fridge. These ¢sh weighed 26 Æ 5g and had a fork length of 129 Æ 7 mm (mean Æ SD). Four beakers were ¢lled with 500 mL of the ¢ltered seawater (salinity of 36.2), and an initial 20 mL water sample was taken to deter- mine the background nutrient concentrations. One bait¢sh was placed in each beaker and the water was stirred once with a glass rod. After 2min, a 20 mL water sample was taken from approximately 2^3 cm o¡ the surface for consistency throughout the experiment. After the sample was taken, the water was again stirred once. This procedure was re- peated at 5, 30, 60, 240 and 1440 min. Experiments were conducted over 24 h to minimize microbiologi- cal growth. Fish were still intact at the end of this per- iod, with only minor signs of decomposition of the skin. Each water sample was ¢ltered (0.45 mm) and Aquaculture Research, 2009, 40, 968^972 doi: 10.1111/j.1365-2109.2008.02158.x r 2009 The Authors 968 Journal Compilation r 2009 Blackwell Publishing Ltd