Experimental fish and procedures: Pabda were used as a study fish because of its high value and farming in Bangladesh in recirculating aquaculture system (RAS) for the last 5 years. Fry of the fish was collected from Sharnalata Matshya Hatchery, Trishal, Mymensingh, Bangladesh. The experiment was undertaken from July to October 2017 in RAS and closed aquaculture system (CAS) for four months.
Culture systems: Three tanks of RAS (10,000 L capacity) were used by filling with 10,000 liters tap water, while three concrete tanks were used for CAS as the experimental systems. In CAS, each small tank was 10 × 6 × 21/4 ft, filled with 650 liters of tap water.
Culture station: The fish was reared in RAS in the Agro 3 Fish Hatchery and Culture Farm, Mymensingh. However, the CAS experiment was undertaken in the wet laboratory of Zoology Section, Biological Research Division, Bangladesh Council of Scientific and Industrial Research (BCSIR), Dhaka.
Experimental design: This was a 2 × 4 factorial study in triplicate. The fixed factors were RAS and CAS and four culture months. Growth performances were measured by ADG (average daily weight gain), SGR (specific growth rate), FCR (feed conversion ratio) and well-being or condition factor (K).
Stocking density: Each RAS tank and CAS tank was stocked with a fish fry at the rate of 1 fish/L, i.e., 10,000 and 650 fries, respectively. The initial average weight and length of a fish were 0.35 g and 1.25 cm, respectively. Before stocking, the fry was acclimatized for 6 hours in a flow-through tank system in clean water. After that, they were slowly released into the tanks. No feed was given on the first day of stocking.
Feeding of fish: In both systems, fish were fed the same commercial diet (Tongwei, China). Fish were fed with powdered feed at 10% of body weight (BW) in the 1st month and pellet feed at 4% BW in the 2 nd & 3rd months and 2% BW in the 4 th month.
Sampling design: Fish samples (10 fish/tank) were drawn every month with a fine meshed net. Individual length (cm) was measured by using a measuring scale and weight (g) by an electronic balance. At the end of the experiment, the final length (cm) and weight (g) of the individual fish were also recorded.
Water quality variables: Water quality variables of CAS were monitored at every 7 days interval. The dissolved oxygen (DO) concentration and water temperature (°C) of each tank were measured by using a portable DO meter. The level of pH was measured by a digital pH meter. Light intensity was measured by using a light intensity meter.
Fish growth performances: Final body weight and length of 10 fish were recorded from each tank after harvest. The observed body weight and food intake data were used to calculate the following growth indices:
1. Average Daily Weight Gain (ADG) = W2 − W1/ T2 – T1
2. Specific Growth Rate (SGR) = (lnW2 − lnW1/ T2 – T1) × 100%
3. Feed Conversion Ratio (FCR) = Feed (g) consumed by fish/ Weight (g) gain of fish
4. Condition factor (K) = (W/ L 3 ) × 100
Where, W2 = Final body weight (g) of fish at time T2 (day), W1 = Initial body weight (g) of fish at time T1 (day), W= Body weight in gram (g), L= Body length in centimeter (cm).
Biochemical composition of the fish muscle: The studies fish (Pabda) of our experiment was a small indigenous species (SIS), so we collected the whole muscle to make a pull sample for analysis. Moisture and ash were measured following the methods described by AOAC (2000). The crude protein content was assayed by measuring nitrogen (N, 6.25) using the Kjeldahl method (Kjeldahl 1883), and the crude lipid content was measured by the Floch method using chloroform: methanol (2:1). The amino acid profiles were determined using an automatic amino acid analyzer following the protocol established by SYKAM S4300, The Czech Republic. Fat was extracted into the ether and then methylated to fatty acid methyl esters (FAMEs). Gas chromatography (GC) was used to measure FAMEs quantitatively (Dodds 2005).
Statistical analysis: Percent data were transformed into the square root before statistical analysis. Means between systems were compared by t-test, while among months by ANOVA followed by Tukey’s HSD post hoc for multiple comparisons. The Statistical Package for the Social Sciences (SPSS) v. 20.0 software package (SPSS, SAS Institute Inc. Gary, USA) and Microsoft office excel 2007 were used for statistical analysis. The level of significance was at 5%.