The rates of growth and nitrogen efflux (total nitrogen and ammonia) of individual C. gariepinus ($\bar{x}~=~32.2$ g; S.D. = 4.8 g) kept under 4 feeding regimens, following a 48 h imposed fast (phase l), were measured periodically. In phase 2 (35 d), groups A, B and C were fed a 49.75%, 45.55% or 41.10% protein diet respectively at a ration of 0.5% body weight (1 d−1) Group D were not fed. In phase 3 (25 d) all groups were fed the 41.10% diet. In phase 1, the ammonia efflux rates were lower than any of the values found in either phase 2 and 3. In phase 2, group A (49.75%) had a higher mean ammonia efflux rates than the other groups and ammonia comprised 60-100% of the total nitrogen efflux in all groups. Group D showed a direct relationship between ammonia efflux rate and length of fast. In phase 3, refeeding with the 41.1% protein diet caused the ammonia efflux rates of groups B and D to converge within 7 days to values no different from those of group C, but group A maintained a significantly higher mean value until day 25. During phase 2, the growth rates in B (45.55%) were greatest, but, none of the among groups differences were significant. Group D fish (unfed) lost approximately 30% of their initial weight during phase 2. Nitrogen efflux rates, notably ammonia, showed a pattern of excretion that was directly related to the protein content of the diet, but that the source of dietary protein, dietary energy and the total available energy also influenced nitrogen metabolism. The small differences in growth found were related to between diet differences in composition. The weight loss in the unfed group of fish was probably attributable to the utilisation of lipid and/or protein reserves as metabolic fuels.