3. RESULTS AND DISCUSSION

The primary objective of Kibos fry production centre is the production in mass of fingerlings and the supply of these fish to the local farmers. Consequently, practical considerations of selling fingerlings, logistical bottlenecks, etc. influenced the timing of pond harvests and stocking, and to some degree the results. It was not possible to maintain all the fish for all trials the same number of days in the ponds. Some ponds remained empty for 1 or 2 weeks, others were refilled within a few days.

In the 8 months of the study, 186 000 were produced in the integrated ponds and 25 900 in the control pond. Table I and II present the data from ponds with integrated farming and those trials with fish feed.

Table I

Results of raising larval Clarias gariepinus
In ponds with integrated farming at 744 chickens per ha.
Pond surface area 1340 m².

* Those marked “counted” were physically counted, those marked “sample” are those extrapolated from volumetric measurement.

** Counted and weighed, mean weight determined by total weight/total number

Table II

Ponds fed daily with 1 to 2 kg fish feed
Pond surface area 700 m²

3.1 Stocking Density and Survival Rates

Within limits, the stocking rate of the larvae in rearing ponds influences the survival as well as the growth rate. De Graff et al (in press) found that at lower stocking densities, 8 to 40 larvae/m², the individual growth rate was faster but yielded less individual fish than at medium stocking rates of 40 to 100 /m². Higher stocking rates (200/m²) gave much the same results in terms of numbers and weight as the medium rate, with an expected increase in the mortality rate. Janssen (pers.comm.) has found little correlation between stocking rates and harvest results: using a stocking rates of between 20 to 100 larvae /m², the total numbers and weights of fish harvested are much the same, with an average harvested number of 10/m².

With the current series of trials, the initial stocking rates ranged from about 15 to 38 larvae /m². These values are somewhat lower than those quoted in the literature (Viveen et al, 1985, De Graaf et al, in press). However, most of the ponds used by these authors are much smaller; 100 or 200 m². Direct extrapolation from stocking densities of small ponds to those several times larger would not necessarily yield corresponding results. There is very likely an influence of the ratio of water surface area to the pond dike area. There is an unknown factor involved here; the larvae may need to seek occasional shelter, or find the environment near the bank area more conducive to survival and growth (De Graaf, pers.comm.). Trials where larvae were protected from predation by being kept in very large hapas and away from the banks yielded lower growth (Nugent, pers.comm).

As stocking densities were only roughly calculated, it is only possible to give a rough estimate of survival rates. With stocking densities of 20 000 to 50 000 (mean value 35 000) and harvested results varying between 700 and 29 000 (mean value 7 700), an approximate, mean survival rate of 22 % was achieved. However, the extreme variation in the results would tend to indicate that there is little if any improvement in the survival rate of the larvae when comparing the results of smaller and larger ponds. In fact, the overall mean result of from the smaller, 700 m² ponds (7.4/m²) would appear to be better than those of the larger 1 330 m² ponds (5.8/m²), although this is not significant due again to the large variation in results from both series of trials (p> 0.05).

3.2 Duration of the Rearing Time in the Ponds

It is not clear at what time or stage of development that larval or fry mortality occurs in the ponds. Theoretically, a significant amount of mortality should occur fairly soon after stocking the larvae. Other experiments (Obuya, not published) have shown a 50 % mortality occurring in the first 14 days in a pond protected from predation by completely covering the pond with a fine mesh, suspended net. In the first few days some of the fish may not find adequate food, environmental extremes of low oxygen at night might cause problems before the branchial organs are functional, and there may be some individuals where physiological or developmental defects would cause death. With the relatively high fecundity of the species, the survival rate of the larvae in nature is obviously very poor.

Many authors, however, attribute most of the problems of low and variable survival in the ponds to predation and cannibalism (Micha, 1975; Hogendoorn, 1979; Janssen, 1985; De Graaf et al, in press; De Graaf and Nugent, pers.comm). When cannibalism begins, the difference in individual growth rates leads to lower numbers at the end of the harvest, with two distinct size classes, one considerably larger than the other (De Graaf et at, in press). Predation by frogs or birds should, however, be somewhat more of a constant factor with a number of fish removed on a daily basis. In any event, with both cannibalism and predation, the longer the fish remain in the pond, the fewer the number that should be found at the harvest.

This was not the case in the current series of trials. The number of fingerlings harvested correlated with the number of days spent in the ponds show a random distribution (r2 = 0.023), See figure 1.

FIGURE 1
Number of Clarias Fingerings Harvested Versus Number of Days in the Pond

3.3 Growth Rates

A value for the mean weight of the harvested fingerlings was not systematically determined; at the time of the trials, the farm was not equipped with a scale of the required accuracy. At a later date, when a more accurate scale was available, a weight/length determination was made and it is safe to estimate that the mean size of the fingerlings from the ponds was generally between 4 and 15 g. The smaller fish occurred when the harvested number was higher and the larger value at lower densities. The results of trial 17 are of some interest. Forty-five days from stocking the larvae, 952 fingerlings weighing a total of 40 kg (balance accurate to 50 g) were harvested. The resulting mean weight of 42 g gives an indication of the growth capacity for this species at theoretical optimum conditions of very low density (<1 m²) and abundant food.

3.4 Chickens

The results obtained from the chicken operation were not encouraging (see Table III). The laying percentage of the birds was only above 60 % in 4 of the 8 months of the trial. Table III does not include the cost of the day-old chicks, the early medicines and vaccines needed, and the feed, housing, and labour costs necessary to produce a 4 month old bird before transferring them to the chicken houses over the ponds. The commercial price for 4 month old layers is currently about Ksh 250 a piece (1US$=Ksh50), and the residual value from the sale of the 18 month old birds is approximately Ksh 150. If these factors are entered in the financial equation, as well as the depreciation costs of the chicken houses, it is clearly uneconomical. The major cause of this poor performance is the quality of the birds used. The highest monthly laying percentage recorded was 64 %, the break-even point is above 70%.

3.5 Pond Production

Integrated fish farming, or the rearing animals and fish at the same time, is a technique that is common throughout South East Asia, China, and other countries where fish farming has a long history. It is one of the most economical means of producing fish. The feed used by the animals is “used again” to feed the fish in the form of the animal manure. The increased nutrient load in the pond encourages the development of natural fish feeds, which further benefit the fish. Almost all integrated farming is practiced, however, to produce larger, table sized fish where they are kept in the ponds for several months. Yields from integrated farming using laying chickens and Tilapia species vary from 4 to 12 t/ha/yr or 40 to 120 kg/are/yr (Burns and Stickney, 1980, Delmondo, 1980, Cowx, 1992). Results elsewhere in Western Kenya (Ongadi, unpublished) with Clarias gariepinus show very good results with chicken/catfish culture, 15 g fingerlings grew to 200 g in 4 months with a survival rate of 90 % at a stocking density of 3/m², a production rate of 90 kg/are/yr.

Table III

Kibos Fish Farm
Integrated farming : Chickens and Clarias fingerlings
Production figures for chickens, 8 months

The extrapolation to annual production rates is only of limited usefulness, and comparing production rates of fingerling systems with grow out systems is misleading. The fingerlings are in the ponds for only 3 to 6 weeks and they are at the fastest stages of growth doubling their size every few days. In production ponds, growth rates are relatively slower and fish remain undisturbed in the ponds for 4 to 8 months.

This comparison does have some value when comparing the use of on-farm resources: is it more economical for the farm to use the pond facilities for the production of table size fish, or fingerlings for sale. This comparison must include the times the ponds are empty. In actual practice, the maximum possible pond-days in the trial period would have been 1 464, in practice, there were only 764 days when the ponds were full of water and fish, or roughly half of the time. The ponds require 1 day to drain and harvest, 2 or 3 days to dry and insure that there are no remaining clarias, and 2 or 3 days to fill. This cycle is repeated every 30 days. There are also logistical problems with timing the arrival of the larvae with empty ponds, weekends, etc. and some ponds stay idle for several more days than necessary. In the integrated ponds, if one combines the actual dates of the trial period with the actual number of ponds used and the actual number of fingerlings produced (at a mean weight of 7g), the production was only 2.4 t/ha/yr, or 24 kg/are/year. This is low.

The sale value of the clarias fingerlings is Ksh 1.00 a piece at the farm gate. During the 8 month trial period, Ksh 186 000 were realized from the sale of fingerlings which were produced in the integrated ponds. The sale price of adult Clarias in the local markets is about Ksh 70/kg. The same monitory value could have been received from the sale of 2.65 t of adult Clarias, which is equivalent to 49 kg/are/year (4.9 t/ha/yr). For integrated farming systems, this is a relatively low value and could easily be improved. It may then prove to be more profitable to use the larger ponds at Kibos for production of table sized fish.