This trial was carried out with community oartjcipation in 2 perennial reservoirs i.e. Chandrikawewa and Kiri-Ibbanwewa in Sri Lanka In this study, an attempt was made to find out suitable, low cost, fanner made aqua-feed that has cetter growth and higher % survival L. rohita. Six types of fanner made aqua-feed (Fl=4fm:3cf:3rb; F2=4fm:4cf:lrb; F3=2fm:5cf:lrb; F4= 4fm:3sbm:3rb; F5=4fm:4sbm:lrb and F6 = 2f n:5sbm:lrb) were tested in triplicate in a Completely Randomized Design with L. rohita in floating cages (6m ) in these 2 reservoirs. These feed types were prepared in situ using rice bran (Rb), oil extracted soyabean meal (Sbm), locally prepared fish-meal (Fin) and cassa\a flour (Cf) in different ratios and according to the 10% of body weight of the Fish in respective cages. The amount of feed was adjusted bi-weekly according to the bod) weight of Fish, feed were provided twice per day, in the morning (0830 hrs) and in the evening (1700 hrs). Two sets of 28 days - old L. rohita fry with different mean lengths, i.e. 2.08+0.1403, 2.9+0.2797 cm and mean weight;., i.e. 0.102+0.03.1, 0'3543+(jxf78bg that were stocked in Kiri-Ibbanwewa * and two sets of L rohita with mean length of 4.9:^0.6148, 1.98+0.4782 cm and mean weight of 2.6055±0.3351, 0.1143+0.0745g that were stocked in Chandrika wewa were used for the four trials. The rearing period for each trial, varied fromj '80 - 90 days. One-way ANOVA revealed that the Specific Groy/th Rate in weight (SGR-W), Average Daily Growth (ADG) and % Survival of tfje fish fed on 6 feed (:ypes were signiFicantly different (p>0.05) in cages in both reservoirs. The pair-wise comparison by the Fisher’s test showed that the ADG, SGR-W aid % Survival with 6 feed types are signiFicantly different in the two reservoirs, respectively. According to the normal distribution of ADG, SGR-W and % Survival of the Fish with respect to these 6 feed types; FI, F3 and F6 could be considered as suitable feed for Lrohita seed production through cage culture.