The present experiment was conducted for a period of four months (120 days) from March to June 2012 in Kaptai Lake near to Riverine Sub-Station (Lake Fisheries) of Bangladesh Fisheries Research Institute (BFRI) in Rangamati Hill District of Bangladesh. In this experiment, 12 floating net cage each having an area of 3 m × 3 m × 2 m (L × W × H, 18 m3) made of synthetic nylon net (mesh size 1.1 cm) has been installed in the lake. Each net cage was tied and hanged with bamboo pole frame and covered at the top with another piece of plastic net (mesh size 2.5 cm) to prevent escape of fish by jumping and bird predation. The bamboo poles were covered with long pieces of wooden raft for easy movement, feeding and sampling of the experimental fish on the cage structure. Empty vacuum plastic drums of 250 liters size were used as cage float for buoyancy of cage structure. The outer side of each experimental net was also covered with a fine meshed net (20 cm depth) according to surface of the lake water level to inhibit floating feeds to go out from the net and to reduce the entrance of non-caged fishes from wild sources. The whole cage structure was tied with anchors at both shore sides by nylon rope to facilitate easy floating and moving of whole cage structure with 12 individual cages depending on water level. In the present study, we tested four different stocking densities of monosex tilapia like 50 fish/m3, 75 fish/m3, 100 fish/m3 and 125 fish/ m3 designated as T 50 , T75, T100 and T125, respectively, in triplicates for each treatment group. In brief, hatchery produced and hormonally sex-reversed juvenile monosex male tilapia, Oreochromis niloticus (GIFT strain) averaging 15.20 ± 0.15 g (mean ± SD) were transported to the experimental site by oxygenated polyethylene bags and they were kept in three net hapas for 24 hours for acclimation with environment and then initial length and weight of fish were recorded individually in ‘cm’ and ‘g’ with the help of a measuring scale and a digital electronic balance (OHAUS, Model CT 1200-S, USA), respectively. Finally, the cages were randomly stocked with monosex tilapia and the number of fish stocked in each cages (18 m3) were recorded simultaneously. A commercial floating pellet feed was used in the experimental period. The diet’s pellet size was 1.8-2.2 mm, and the feed contained 28.76% crude protein, 5.95% crude lipid, 14.23% crude ash and 9.43% moisture. Feeding was done at 3-5% of the initial body weight. The total amount of feed was divided into two equal rations for feeding at 8 a.m. and 4 p.m.. Feeding was done by hand very slowly and carefully to ensure ingestion of feed completely. Around 20% of fish in each treatment were sampled fortnightly in order to determine fish weight and the feed amounts were adjusted as per changes of body weight in each trial. A record of supplied feed was maintained to determine the food conversion ratio (FCR). During the study period, dead fish were recorded and removed quickly. The cages were lifted from water at every 15 days interval to check the net and for cleaning purposes. Cages were cleaned with soft brush to remove algae, sponges and other organisms. Loose twine, net mesh torn by predators, anchors and sinkers were checked routinely and immediately mended or replaced as needed. Physico- chemical parameters of water such as temperature (°C), transparency (cm), pH, dissolved oxygen, DO (mg/L), hardness (mg/L) and total alkalinity (mg/L) were monitored weekly in the morning between 7 and 8 a.m. during the whole culture period. Water temperature was recorded with a glass Celsius thermometer, water transparency was recorded with a Secchi disc, pH and DO were measured using a digital pocket pH meter and DO meter, respectively. Other chemical parameters were measured using a HACH kit box. After 120 days of trial, the whole cage structure moored to shore of the lake and all fish were harvested by repeated scoop netting and then fish were counted, measured and weighed for each cage. Proximate composition analyzes of experimental diet and whole fish body were performed by the standard methods of Association of Official Analytical Chemists. After the completion of experiment, three fishes from each treatment were collected for further analyzes of body carcass composition. For determining moisture content, samples of diets and wet fishes were dried at 135°C for 2 h. Ash content was determined by using a muffle furnace (600°C for 4 h). Crude lipid content was determined by the soxhlet apparatus using Soxtec system 1046 (Foss, Hoganas, Sweden) and crude protein content by Kjeldahl method (N × 6.25) after acid digestion, distillation and titration of samples. The samples were analyzed in the Fish Nutrition Lab, Department of Aquaculture, Faculty of Fisheries, Bangladesh Agricultural University, Mymensingh, Bangladesh. The mean values for water quality parameters, growth, survival and production of different treatments were subjected to one-way ANOVA followed by Duncan’s New Multiple Range test. All statistical analyzes were performed using SAS software version 9.1. Standard deviation of each parameter and treatment was determined and expressed as mean ± SD. Treatment effects were considered with the significant level at P<0.05.After the termination of experiment, an economic analysis was performed to estimate the net return and benefit–cost ratio on the basis of different stocking densities of monosex tilapia. At the end of the experiment, all fish were sold and the prices of fish were attributed to the Rangamati local fish market price in July 2012 and expressed in Bangladeshi taka (1US$=80 BDT). The local price per kg of monosex tilapia was 250 BDT for near or more than 250 g fish (T 50 ), 200 BDT for 200-250 g fish (T75 and T100 ) and 180 BDT for less than 200 g fish (T125 ). In the study, fish selling price was higher than normal price, because every year from April/May to June/July, Kaptai Lake is closed by the government authority for capture fishing.