We tested the mode of water exchange in tanks containing Clarias gariepinus to determine if different methods yielded differing results for feed consumption or feed conversion. The purpose of the research was to better design future trials comparing feeds and different rates of water exchange. Twelve plastic tanks with capacity 1m 3 were each stocked with 25 subadult catfish, Clarias gariepinus , averaging 151±2g. Water exchange was designed to use the same total amount of water (500 litres per day), but on three different schedules: treatment 1 was a one-time 500L water exchange per day, treatment 2 was a one-time 1000L single exchange every two days and treatment 3 was a continuous (flow through) slow rate of exchange at a rate of 500L of water per day. The water source was a borehole located on the farm. All tanks drained their water from the bottom. Fish were fed every afternoon to satiation for 87 days on feed of 42% crude protein, followed by 33% crude protein feed for a total of 115 days. Treatments 1 and 3 resulted in significantly higher fish growth and better feed conversion than treatment 2. Even though the same amount of water exchange was applied in all treatments, a steady rate or smaller, more frequent exchang es worked better for intensive Clarias production.