Sample collection: Depending on the availability of the adult fish in the local markets, four Mystus species namely M. gulio, M. bleekeri, M. cavasius and M. vittatus were selected for this study. M. gulio was collected from local market of Khulna City Corporation and the rest were obtained from a local market (Lakhmipur bazar) of Rajshahi City Corporation during the month of June, 2005 and were identified following the works of Talwar and Jhingram (1991). Keeping into ice, the fish were brought to the laboratory for analysis within 24 hours of collection. After gutted and dressed the body flesh was used for experiments.
Sample preparation: The fish-fillet samples of four mystus species were homogenized separately in a micro cutter and then mixed uniformly with pre-cooled petroleum ether in a homogenizer and filtered through a clean muslin cloth. The process was repeated at least twice in order to obtain lipid-free homogenate. Then each filtrate was clarified further by centrifugation at 8x103 r.p.m for 10 min. Finally, these precipitated mass were air-dried at room temperature (Yeasmin et al., 2001). For iron estimation fish fillet is not needed to be defatted. For estimation of Na, K, Ca, Mg, Cu and Zn, sample solution was prepared by following the standard dry ash method (Anon, 1992). In the case of iron (Fe) estimation few ml. of Bromine water was added to the sample solutions in order to convert ferrous (Fe2+) ion to ferric (Fe3+) ion. Wet oxidation was performed in each Mystus species for P estimation (Didar-ul- Alam et al., 1991).
Estimation of macro (Ca, P, Na, K, and Mg) and micro (Fe, Cu, and Zn) minerals: The concentration of Na, K, Mg, Cu and Zn were measured by atomic absorption spectrometry (Perkin-Elmer, model-3110, England) method (Anon, 1990 and Anon, 1992). Iron content of Mystus species were estimated Spectrophotometrically (Erma, AE-300) by thiocynate colourimetric method (Vogel’s, 1989). Colourimetric method Arsenazo III (2,2-[1,8-Dihydroxy-3,6-disulpho-2,7-napthalalene-bis-(azo)] dibenzenearsonic acid) was used to determine the amount of calcium at pH 8.5 (Bauer, 1981). Phosphorus content in each species was determined Spectro-photometrically by the Phosphovanadomolybdate method (Vogel’s, 1989).
Proximate composition: Defatted homogenate samples (1g) were digested at 390 ºC in a digestion unit (2040 digester) for two hours. The digested mass of each species were then distilled in a KJELTEC system 1002 distilling unit and titrated with 0.05 M NaOH. Crude protein was obtained by multiplying the total nitrogen content by a conversion factor of 6.25 (Peterson, 2002). Lipid from each sample was extracted with petroleum ether (40-60 ºC) in a soxhlet apparatus (Southcombe, 1971). A weighed quantity of samples was ignited in a muffle furnace (Perkin-Elmer, model-3110) at 600 ºC for about 6 hours. Moisture content was determined by drying each sample in a thermostat oven at 105 ºC until a constant weight was obtained (Khunna, 1978).
Data analysis: Proximate and mineral composition of these four species were compared by one-way analysis of variance (ANOVA) followed by Duncan’sMultiple Range Test (DMRT, p<0.05) by using Statistical Packages for Social Science (SPSS) software. Each experiment was replicated three times.