Since 1976, a series of net-cage--culture experiments of tilapia and grass carpwere carried out in Dianshan Hu Lake, Qingpu County, Shanghai, in order to verifythe appropriate technology.In 1976, a high yield of 306 kg/36M²(8. 5kg/M²) in 135 days was obtained, butthe conversion rate of feed was very low. Therefore, another experiment in culturingtilapa was carried out in 1977. The aim of this experiment was to find out the appro-priate technique of cage culture suitable to our country. Nine 2×2×2.5 M cages wereinstalled in the Dianshan Hu Lake for the experiment. After 80 days rearing, theyields obtained ranging from 10.6 kg/M² to 35.15 kg/M². The rate of conversionrange from 2.7 to 5.2.The size of the net cage was 2×2×2.5M, and the stocking density was 4,200individuals/M², the average weight of the fingerling was 1.68g. Feeding with drypellet composed of 10% fish meal, 40% fine rice bran and 50% soybean cake, at afeeding rate of 12.2% to 2.4% fish biomasa per day. After 120 days 388 kg was harvested. The rate of conversion was 2.16, In 1980, the same procedure applied torear grass carp in cage, and an average yield of 25.09 kg/M² (net yield of 14.5 kg/M²) obtained. The Conversion rate was 2.49. The economic value of grass carp ismuch higher than tilapia.The key factors to obtain high yield can be concluded as follows: (1) The cageshould be of medium--sized (20 to 60 M²). A feeding stage is necessary to set inside thecage in order to save the waste of the feed. (2) The optimum stocking density rangesfrom 5 to 10kg/M². Mono-culture of a high valued fish is preferable. (3) The diss-olved oxygen should not be less than 3 mg/L. (4) Nutritive feed with a total protiencontent over 35% should be available. (5) According to the difference of sizes, thefeeding rate ranges from 10% to 2% of fish biomass per day. (6) Effective controlof pest and fish diseases.