Kibos Fry Production Centre was constructed in 1984, 10 years before the trails began. The farm is located in the sugar belt area of the Kano Plains, 10 km inland from the eastern shore of Lake Victoria. The farm is at an altitude of 1 220 m, and located 6 km south of the equator. Air temperatures vary between 18 and 28 C. Day length is an almost constant 13 hours. Soils at the farm are black cotton clay.
Kibos was designed with the primary objective of supplying tilapia fingerlings to small scale rural farmers in Western Kenya. In 1994, the centre began rearing fingerlings of Clarias gariepinus for distribution to the farmers. At the same time, integrated farming with chickens (layers) was begun as demonstration of the technique for private fish farmers.
Given the high demand for the catfish, 6 ponds with integrated farming were used for the production of clarias fingerlings. The trial ponds were identical in size measuring 23 × 58 metres or 1 334 m2 with a water depth of 60 cm at the inlet gradually increasing to 85 cm at the outlet. The ponds had previously been used in clarias fingerling production with rice bran as a feed, and it was assumed that any residual fertilization would be common to all ponds, and in any event, was not of significant amounts.
Pond water was supplied from 2 bore hole pumps with a combined yield of 60 m3/hour. Ponds were filled over 2 to 3 days. Water temperature at the source was a constant 24 C, pond temperature varied between 22 C and 28 C. Total volume of pond water was maintained at 940 m3.
Laying hens were housed in 6 separate wooden structures, each constructed inside separate ponds and located at the water inlet. The houses measured 14.3 m in length, 1.26 m width for a floor surface area of 18 m2. The floors were made of pine wood slates, 1" × 2" leaving a gap of 2 cm between each slat. This allowed the fresh chicken manure and any waste feed to fall directly into the pond water. Drinkers and feeders were of standard design. The chicken houses were built to be 90 cm above the water surface. The density of birds in the houses, including drinkers, laying boxes, and feeders, was 5.5/m2. Initially, each house contained 105 birds for 1300 m2, or at a ratio of 770 per ha.
The birds were transferred to kibos on March 24, 1994 at the age of 4 months and fed a locally available commercial layers mash (Unga Feeds Ltd). Feeding began immediately and continued through out the experimental period of 8 months. Feeding rates were calculated on a basis of 150 g/bird/day, or 16 kg per chicken house per day. The birds began laying eggs about 1 month after introduction. Feeding, watering, and egg collection were done twice daily. Recommended vaccines and vitamin supplements were given to the chickens when needed. Records were kept of chicken mortality/theft, the numbers of eggs collected, broken, and sold, and the total amounts of feed given.
Stocking material was obtained from a hatchery located at the farm. Sexually mature females of Clarias gariepinus weighing a minimum of 200 g and a maximum of 400 g were selected for induced reproduction on the basis of the ability to extract mature sized, translucent eggs from the abdominal vent when applying moderate pressure (Viveen et al, 1985, Verheust, pers.comm.).
Standard methods for artificial reproduction for Clarias gariepinus were used (Viveen et al, 1985). For each pond, 2 to 4 females of 200 to 400 g were used. The fish were injected with a pituitary gland from a male C. gariepinus, the injection made between 16 and 17 h to allow ovulation at the ambient temperature (24 C) by the next morning.
At 8:00 to 9:00 h the next day, eggs were stripped from the females and fertilized. Eggs were then incubated in concrete troughs with bore hole water at a constant 24 C. After 4 days, the yolk sac was largely absorbed and the larvae were swimming freely. At that point, the larvae were transferred to the trial ponds.
The number of larvae was estimated by weighing the eggs, multiplying by 700 per g, and using a hatching rate of 50 %. With this rough method, at least 20 000 and no more than 50 000 larvae were stocked in each pond.
For comparative purposes, one pond without chickens measuring 15 × 46m (700 m2) was used. Stocking rates were between 15 000 and 30 000 larvae. For each trial, the pond was initially manured at a rate of 15 kg/are of composted deep litter (rice chaff) from a chicken house. One week after stocking, the fish were fed with between 1 and 2 kg per day with a processed (hammer milled) feed composed of 10 % fish meal, 20 % cotton seed cake, and 70 % rice bran.
In both the experimental and control ponds, fish were harvested 18 to 45 days after stocking. The ponds were emptied into a 4 mm mesh net placed at the outlet of the pond, the fingerlings were then transferred into circular concrete tanks with a continuous water flow. After about 1 hour, the mud and pond debris were eliminated from the tank. An estimation of the number of fingerlings was made by using a plastic strainer of approximately 500 ml capacity. Three volumetric samples were taken and the fish counted. The mean value of the three samples was then used to determine the mean number of fingerlings per container. The entire harvest was then measured and the total number of fingerlings harvested was estimated by extrapolation. Usually, within a few days, the fingerlings were physically recounted one by one as they were sold to farmers allowing verification of the numbers. This sampling method was found to be accurate with less than 0.5 % difference between the sampled and physically counted fish.
