Pond study: The study was conducted in ten earthen mini ponds, with a water depth of 1 m and 50 m2 surface area. Excavation, predator elimination and liming were done to maintain the water level, water quality and to protect the extruder in to the ponds. Juvenile M. rosenbergii were stocked at 3 m-2 at the end of July 2003. The experiment was carried out from August to November, 2003. Before stocking all the prestocking measurements like pond repairing, liming and removal of predator were done.
Selection and analyses of feed ingredients: Locally available feed ingredients (Table 1), considering both protein supplement and price were selected. The ingredients were analyzed at FMRT lab to determine the level of crude protein, crude lipid and ash according to AOAC (Anon, 1984).
Feed formulation and feeding: In the experiment, five different diets (the Control and the 4 other diets) were formulated by using 'Pearson square' method (De Silva and Anderson, 1995) using a locally made pelleter powered by electricity (Table 2). Four experimental feeds were prepared by replacing fish meal protein by soybean meal as follows - D1: 30% replacement, D2: 40% replacement, D3: 50% replacement; D4: 80% replacement. Feed was given 2 times a day, early in the morning (before sunrise) and just after sunset. Initially, feed was given 5 to 7 % of the total body weight by using feeding tray placed 18-20 cm above the pond bottom.
Variable measured: Growth, survival, SGR (Specific Growth Rate), and FCR (Feed Conversion Ratio) were evaluated. Proximate analyses of the experimental feed were done following standard procedures (Anon, 1992). Water quality parameters like temperature, pH, dissolved oxygen, alkalinity, soil pH and total nitrogen were measured. DO (Winkler method, determined with Azide modification of iodometric method (Anon, 1992) and temperature were measured two times a day at 5:00 and 17:00 h for the duration of culture period. Water pH (pH meter Woon Socket, HANNA) and alkalinity (measured with titrimetric method) were measured once a week at 9:00 h and soil pH and total nitrogen were measured fortnightly.
Statistical analysis: The growth responses of the prawn were calculated for each pond (mean body weight, biomass, growth rate, survival, food intake, food conversion, protein intake and protein efficiency ratio (PER) and were subjected to analysis of variance (ANOVA) followed by Duncan’s multiple range test (DMRT).