The above demonstrates that it is possible to spawn Clarias gariepinus on a small scale using semi-artificial techniques and very minimal equipment and facilities. The expected yield is between 3 000 to 5 000 larvae. All indications are that 10 m2 ponds completely protected with a fine mesh net are very suitable for rearing these relatively small numbers of larvae. With both aspects, the technology is simple, inexpensive, and easy to master, ft remains to be tested in the field with small scale rural farmers. This will be a slow process requiring close support Initial results may be discouraging as the farmers learn the techniques. We recommend that only 2 or 3 farmers be initially selected and a series of 3 to 5 cycles made at each farmers' location. Refinements can then be made before the technique is spread further.
Secondly, from the point of view of applied research, the use of very small ponds should prove a very efficient way to conduct further trials on larval rearing. The number of small ponds can easily be expanded, allowing researchers to quickly conduct the numerous trial repetitions necessary for comparative treatments. With a relatively small surface area of about 400 rn2, a series of 20 small ponds of 1 × 10 m can be quickly built. By using artificial reproduction, these 20 ponds could all be stocked the same day with the spawn from 1 or 2 fish which would give more control over the experiments. All the ponds can be harvested on the same day. We recommend that a series of 20 ponds measuring 1 × 10 m be constructed at Kibos for future experiments. For these trials, the following areas should take priority.
To determine the optimum number of days to leave the Clarias larvae in the ponds. At what point is growth limited? At what point do differential growth rates and cannibalism come intoplay? What is the optimum size to harvest the fish?
In western Kenya, a variety of feeds stuffs suitable for larval fish are available, including blood meal, fish meal, brewers yeast, vitamin supplements, rice and wheat bran, etc, The quantities needed for larval rearing are quite small as is the corresponding cost; an entire cycle for a small scale farmer would only require 4 or 5 kg of such a feed, Various feed combinations should be tested.
Further work is needed on stocking densities in relation to survival and growth, and at what stage does larval mortality occur and why.
The introduction of a simple method of spawning Clarias to rural farmers may prove to be a catalyst for further development of aquaculture in the region. Several hundred fish farmers already raise this fish in polyculture with Tilapia, the demand is increasing, and shortages of Clarias fingerlings already occur.
In addition, and perhaps particular to the area immediately around Lake Victoria, there is a substantial market for small, 5 to 10 g Clarias for use as bait in the commercial nile perch fishery. Already, hatchery produced Clarias fingerlings are sold for this purpose; some fishermen buy the very small clarias and then grow them for 1 or 2 months in small ponds before using them as bait. If the above system can be successfully modified to produce fingerlings of at (east 5 g, a potentially significant, small scale industry could develop around the lake.
