3. OBSERVATIONS, CONCLUSIONS AND RECOMMENDATIONS

3.1 Kajansi Experimental Station

During the period of the expert's assignment the following facilities were added at the Kajansi Fish Farm:

1. A series of twenty ponds of 1/8 acre each were constructed and brought into operation.

2. Two permanent concrete troughs for temporary storage of fingerlings were constructed.

3. Two 1/50-acre ponds were fenced with asbestos sheets for frog control trials.

4. Based on the results obtained from the ponds mentioned in (c) above, one ¼-acre pond was fenced with sheets, and operated.

5. A small aquarium of 24 tanks, each with a capacity of 250 litres, was set up in one of the rooms at the farm laboratory, with piped water, controlled drainage and illumination facilities, for fish behaviour studies.

During the course of the project, the Uganda Fisheries Department changed the name of Kajansi Fish Farm to Kajansi Experimental Station, to specify more clearly the functions of the farm and attached laboratory. It is expected that the Station will be devoted entirely to experimental work on breeding and selection of suitable brood stock. The species and hybrids found suitable for fish farming should be made available to the regional fry production centres, and these centres should take over as far as possible the distribution of fry to farmers.

With changes in the functions of the Station, a re-organization of its staff may be required. The main object of re-organization should be to separate research and extension work from routine administration of the Station. An administrative officer may be appointed to look after the routine administrative work under the general supervision of the Fisheries Officer.

The Fisheries Officer should be responsible for extension and experimental work, but he should be assisted by two Senior Fisheries Assistants, one for extension work and the other for experimental studies. It has to be ensured that the results of successful experiments are implemented in the field, through the fry centres, as soon as possible. A trained technician should be appointed to take charge of the laboratory and aquarium, and to be responsible for routine chemical and biological analyses.

3.2 Carp Culture Experiments

3.21 Control of Predation

It has already been reported (FAO/UN, 1965) that predation by insects (especially Notonectids) and frogs (Xenopus sp.) were observed to be responsible for the low survival rate of carp (Cyprinus carpio) fry. In order to determine the relative importance of these predators, a series of experiments was carried out in which the spawning ponds were treated as follows:

1. 1 ppm of 3 percent B.H.C. (to destroy the insects)

2. 1 ppm of 3 percent B.H.C. plus fencing of the pond (the latter to prevent the entry of frogs into the ponds). (See Fig. 1).

Two ponds of 60 m² each were used for each treatment. The experiment was repeated twice. A control pond was maintained without any treatment.

The results of the experiment showed that the frogs were responsible for 86 percent of the mortality. In the second experiment (b) above, carp eggs spawned by females (3–4 kg in weight) were introduced into the fenced ponds treated with B.H.C. A sample of eggs from each spawning was used and the total survival per spawning was estimated by extrapolation. The results showed that there was a survival of 40,000 to 50,000 fry per spawning. The experiment was repeated on a larger scale, in which the survival was determined from the spawn of one female in a fenced ¼-acre pond, and was found to be 40,000 fry. This survival of 40,000 fry per female may be used as a standard for assessing the effect of predation on spawn under uncontrolled conditions. It may be mentioned here that the above survival compares very favourably with the survival obtained in carp-breeding countries, such as Israel and Yugoslavia. Allowing adequate time for drying, disinfecting, etc., after each spawning, a ¼-acre pond can be used for three spawnings per year and thus produce a total of 120,000 fry annually.

Fig. 1-Carp Spawning Pond Fenced against Frogs

Following these successful results in predator control, fencing of ponds was done at Kisizi Centre in Kigezi District and is now being done in other centres also. A site has been chosen adjacent to Lake Mulehe for the construction of a fenced pond for production of carp fry for stocking the lake.

3.22 Multiple Stocking

Replication of experiments conducted in 1962–64 (FAO/UN, 1965) has reconfirmed that the most appropriate rate of stocking carp in unfed ponds would be 250–400 fish per acre.

Further experiments on multiple stocking, with a combination of carp and Tilapia hybrids in the same ponds, confirmed the results of previous work, viz., that better growth rates of carp were obtained in polyculture.

These results support the conclusions that neither the Tilapia hybrids nor the carp has a significant advantage over the other. The carp is important as a supplementary fish in Tilapia hybrid ponds. The results are typical of most areas of Uganda, except for the Kigezi district where the low water temperature (15–20°C) is a limiting factor for the growth of Tilapia and the carp has a significant advantage in this respect.

3.23 Observations on Carp in Natural Bodies of Water

For many years there has been great opposition to the introduction of carp in East African waters, on the grounds that the species would compete with, and destroy the breeding sites of the indigenous species. Despite these objections, carp were eventually stocked at Kajansi, 57 specimens being introduced in 1957. Earlier experiments showed that although the carp spawned successfully, fry mortality was considerable. Later work (see 3.21) showed that the mortalities were due to intensive predation by frogs and insects.

Stocking of natural bodies of water was done two years after the introduction of the species at Kajansi, and several lakes in Kigezi District were stocked. Contrary to expectations, but in keeping with the results of the work at Kajansi, although the growth of the stocked fish was rapid, the species failed to establish itself in these lakes as self-sustaining populations. It was observed to spawn in the lakes, but commercial fishing of the species declined after an initial high level. This was probably due to low survival rate at the fry stage. To maintain the fishery, restocking was necessary.

Observations in the stocked lakes showed an abundance of the predators responsible for high fry mortality in ponds. It seems very probable that recruitment is very adversely affected by intense predation by the abundant frogs and insects in the lakes.

The control of fry predators may present practical difficulties in large lakes, but in the small lakes of Kigezi it may be advantageous to initiate a stocking program to establish a carp fishery. Suitable sites for the construction of carp breeding ponds are available adjacent to the main lakes of Kigezi, especially Lakes Bunyoni, Mutanda and Mulehe. Fenced breeding ponds like those in Kajansi may be constructed and carp fry raised to a sufficiently large size for release into the lakes.

3.3 Tilapia Hybridization

3.31 Cross Breeding Experiments

In attempts to produce monosex or sterile offspring of Tilapia, many crosses were made in the Kajansi Experimental Station in the years of 1962–66. The results of this work are summarized below. Some of them have already been described in the report of the expert's earlier assignment (FAO/UN, 1965), but these descriptions have beem amplified and amended in the light of further observations and nomenclatural changes. The species so far used for hybridization are:

1. Tilapia leucosticta (Trewavas, 1933)

2. T. nilotica (Linnaeus), from Lake George

3. T. aurea (Steindachner), from Israel

4. T. hornorum Trewavas, from Zanzibar

5. T. nigra (Günther), from Kenya

6. Tilapia sp. aff. nilotica, from Lake Rudolf

7. T. variabilis Boulenger, from Lake Victoria

In the first experiment, one adult male of one species and three females of another species were placed in an aquarium tank. When breeding occurred, the offspring were reared under controlled conditions, either in another aquarium or in asbestos cement tanks, to sexual maturity. The sex of each individual of the entire batch was determined from external characters (genital papillae) and in doubtful cases, by dissection. Morphological characteristics of all the individuals were recorded. The second phase was to use groups (instead of individuals) of two species by releasing several males of one species with several females of the second species in one pond. The results of the group crossings are given below.

Breeding occurred in the hybridization pond within 15–20 days after releasing the brood stock. At Kajansi there seems to be no definite breeding season for Tilapia; the fish appear to spawn all the year round.

A fertility test was made on the hybrids obtained from cross No.2 (see page 5) by back crossing with females of T. nilotica. The hybrids were found to be fertile.

Male hybrids from all crosses, when sexually mature, display the breeding colours and dig nests in the ponds even in the absence of females.

Crosses made

The hybridization of T. aurea and T. nilotica was repeated on a larger scale than previously. Four hundred hybrids of that cross were sexed and found to be 100 percent males. The results of other hybridizations carried out are as follows:

Male T. hornorum was crossed with female T. mossambica and T. nilotica. As a large percentage of the hybrids of these crosses were males, it may be concluded that high yields of male hybrids is a specific character of crosses between distinct species.

3.32 Growth Potential of Hybrids

In order to study the growth potential of hybrids in relation to that of parent species, comparison was made of the growth rate of the hybrid ♂T. hornorum × ♀T. nilotica, T. nilotica, T. hornorum, and T. nigra. The results are summarized below which clearly show the higher growth potential of the hybrid.

A comparative study was carried out of the 100 percent male hybrids obtained by crossing male T. hornorum with female T. nilotica and male T. aurea with female T. nilotica. Experiments did not reveal any significant differences between the growth rates of the two hybrids. However, it would appear that only the ♂ T. hornorum × ♀ T. nilotica cross should be recommended for large-scale use in fry centres outside the experimental station, because of the easily determined distinction between the ♂hybrids and ♂T. hornorum. It was difficult to distinguish the hybrids from ♂T. aurea.

Unmistakable differences in external appearance of the hybrids and the males of the parent species is essential if back crossing is to be prevented. In back crossing the hybrid with the ♀ parent, the resulting offspring in the F₂ generation are a mixture of males and females of both parent species and of the hybrids.

It was observed that the growth rate of Tilapia hybrids in ponds in Kigezi was greater than that of other Tilapia spp. in the same ponds.

Efforts to obtain herbivorous hybrids of T. zillii × T. melanopleura were continued during the period under report, but without any success.

3.33 Feeding Experiments

In order to determine the response of ♂T. hornorum × ♀ T. nilotica cross to supplemental feeding, an experiment was carried out in 500 m² ponds stocked at rates varying from 1,000 to 4,000 fish per hectare. Feeding was done with maize meal at the rate of 3 percent of body weight, 6 days a week. To study the influence of stock density on response to feeding, four ponds were used (two for feeding and two controls) and the experiment lasted 130 days. The results are summarized below.

In view of the high price of maize meal in Kampala, such supplemental feeding cannot be considered economical. Further experiments using less expensive feeds, different rates of feeding and stock densities would be desirable.

3.34 Sex Ratio

The sex ratio of the following species of Tilapia used for hybridization was studied for comparison with that of the hybrids.

The sex ratio was found to be 1:1 with a variation not exceeding 1 percent in the populations of the above species. It has already been observed that hybridization between distinct, identifiable species, results in changes in this ratio. If sex ratio in Tilapia species is as stable as observed by the expert and disturbed sex ratio the characteristic of crosses between distinctly distinguishable species, this character may be valuable in judging the validity of species in the genus Tilapia, particularly in cases where meristic features are not of sufficient diagnostic importance. For example, in a cross between specimens suspected to belong to two different species, a disturbed sex ratio of progeny may serve to confirm that the parents actually belonged to two distinct species. Similarly, a 1:1 sex ratio of the offspring may indicate that the parents belonged to the same species.

A test of this hypothesis was made with Tilapia sp. from Lake Rudolf, which has the same meristic characters (number of spines, rays, gill rakers, etc.) as T. nilotica, but differs from it in general appearance and in colouration. The F₁ generation of a cross between these two species consisted of 95 percent males, which appears to indicate - on the above evidence - that the parents really belong to different species.

3.35 Improvement of the Techniques of Hybridization

To avoid back crossing between male hybrids and parent female T. nilotica in hybridization ponds, the ponds have to be drained completely every five months and the hybrids transferred to nursery ponds. This is to be done before the hybrid fry attain sexual maturity, and likely to cross with the female parent. The need for regular draining of spawning ponds appears to be a handicap in hybrid production in the up-country fry centres, where close technical supervision of operations is difficult.

To avoid the periodical draining of ponds, the breeding of Tilapia was tried in pens made of welded steel mesh of 1–1.5 inch mesh size, in the hybridization ponds. In a half-acre pond, a breeding pen measuring 29.5 × 19.7 ft was constructed, and stocked with six male T. hornorum and eight female T. nilotica. Male Tilapia are known to chase away small fish from the breeding ground. As the wire mesh was about 1.5 inch in size, the small hybrid fish could be chased out of the pens. These same fish could not enter the pen to breed with the female parents when they became sexually mature after 4 to 5 months, as they are then too large to pass through the mesh.

Further trials are required to determine the most suitable size of pens for the spawners, and the optimum number of spawners to be introduced in the pens. However, it is clear that this method permits the hybrids to be nursed in the same pond as the parents by preventing them from back crossing with the female parent. This cuts out the necessity of transferring the hybrids to nursery ponds before they attain sexual maturity, and therefore leads to better utilization of space. The nursery ponds may be used as additional spawning-cum-nursery ponds with enclosed pens, or for other purposes.

With the old method, fry were transferred to the nursery ponds when they were 5 to 10 g in weight. At this size a mortality of 25 to 30 percent was found to occur in the nursery ponds. Extended transportation at this size from nursery ponds to peasant fish ponds, taking up to 3 or 4 hours, caused a mortality greater than 75 percent. The new method permits the fry to be kept in the spawning ponds at densities of up to 10,000 fish per acre until they attain a weight of 30 g when the growth is arrested. When the fry are allowed to remain in the spawning ponds up to this size, the mortality is reduced to 5 to 10 percent. When transferred from the spawning ponds to peasant fish ponds at 30 g weight, mortality is reduced considerably. The daily gain in weight is higher when fish of 30 g weight are stocked in ponds than when 5 g fish are stocked, and therefore the time taken to grow to marketable size is shorter. This in turn leads to the more efficient utilization of available space in fish ponds and permits a greater weight of fish to be produced in any one year.

For the new method to be used, the small fish farmer has to incur additional expenditure for the purchase of frames and wire mesh for making the pens. So, it may be easier for the new method to be adopted in the fry centres. Large numbers of fry of about 30 g weight can be produced at these centres and transported to rural fish ponds with less mortality.

3.36 Change of Water in Ponds as an Aid to Growth

It was observed that in ponds stocked at the rate of 800 fish/acre, Tilapia hybrids ceased growing in 4 to 5 months when they had attained a weight of about 150 g. This represents a total production of only 250 to 300 lbs/acre, with the fish being below marketable size. Because of this, it was recommended that the stocking rate for Tilapia hybrids be reduced to 400 fish/acre. It was, however, observed later that if the water in ponds stocked at 800 fish/acre was changed when growth ceased at 150 g, growth would re-start, probably due to removal of accumulated growth inhibiting substances, and continue till the individual fish attained a weight of 200 to 250 g, which is a marketable size. The total weight of fish produced in these ponds was recorded to be approximately 500 lbs/acre. Changing the water in ponds when growth ceases would, therefore, permit stocking rates to be increased from the recommended rate of 400 fish/acre to 800 fish/acre.

3.4 Nile Perch (Lates niloticus) Investigations

3.41 Collection of Fingerlings

A search was made for suitable areas for the collection of Nile perch fingerlings in Lakes Albert, Kyoga and Rudolf (Kenya). The most suitable size range of perch fingerlings for transportation has been found to be between 8 and 14 cm (FAO/UN, 1965). Fingerlings of this length were found in the lagoons of Lake Albert, but never in large numbers. A fishing unit of three men collected 250–300 per day in the lagoon at Butiaba. Lake Kyoga, because of its unseineable shores, was not suitable for collecting perch fingerlings. Trapping was tried on the lake, but without success.

Large concentrations of fry, 10 to 30 mm in length, were found near Ferguson's Gulf on Lake Rudolf. Further search failed to show any concentrations of bigger fry. The small fry have been found to withstand transportation at densities of 150–250 in bottles containing 6–7 litres of filtered water and 12 litres of oxygen, for not longer than 8 hours. Reduction of the density of fish/bottle did not increase the survival time proportionately. Owing to the remoteness of Lake Rudolf, the collected fry have to be flown out by air from Ferguson's Gulf if this source is to be exploited. The cheapest way of doing this is to charter a small aeroplane from Entebbe at a cost of Sh. ¹ 1,500/per flight.

As already reported (FAO/UN, 1965), attempts to breed the Nile perch in ponds were not successful. It will therefore be desirable to study the possibility of inducing them to spawn in confined waters by hormonal injections or other techniques.

¹ 1 East African Shilling (Sh) = ca. $0.14

3.42 Storage of Fingerlings

Storage of Nile perch fingerlings was tried at densities of 500–1,000 fish/acre. The experiment was carried out in 60 m² ponds, without artificial feeding. It was found that the fingerlings could only be kept for periods not exceeding 4–5 weeks at those high densities. Survival declined rapidly after this time, due to food shortage in the ponds.

Nile perch fingerlings were added in small numbers to all Tilapia hybrid ponds for the control of wild fish. They were found to be very effective for this purpose, but could not be used on a large scale due to non-availability of fingerlings in adequate numbers.

3.5 New Species for Culture

3.51 Bagrus docmac

Specimens of this species obtained from Lake Victoria were successfully bred in ponds at Kajansi. This is the first record of the species spawning in captivity. Fingerlings produced at Kajansi were stocked with Tilapia sp. in an attempt to establish balanced predator/forage populations. These experiments are now in progress.

Bagrus docmac was chosen instead of other species of similar behaviour and ecological requirements such as Clarias sp., because of consumer preference.

3.52 Grass Carp (Ctenopharyngodon idella)

As experiments in the biological control of weeds in dams with the indigenous herbivorous fish Tilapia zillii were not successful, the Fisheries Department decided to try the grass carp (Ctenopharyngodon idella). About 400 fingerlings from Hong Kong were introduced in the Kajansi Farm. They are expected to attain maturity by the end of 1967, when experiments can be initiated to induce spawning. If the spawning experiments are successful, further experiments will have to be carried out on the weed control ability of the species in dams and ponds. If found suitable, the species should be widely used for weed control in weed-choked ponds and dams throughout the country.

3.53 Fresh-water Prawn (Macrobrachium lepidactylum)

The Athi River prawn (Macrobrachium lepidactylum) has a high market value, as a full grown specimen has as much as 20 g of tail meat, which is almost 50 percent of the total body weight. The ratio of 1:1 for edible tail meat to waste is superior to that of the Louisiana crayfish, Procambarus clarkii, (previously introduced by the Fisheries Department to Kajansi), the edible part of which is only about 20 percent of the body weight. In view of this, it was decided to conduct experiments in the culture of Macrobrachium.

Experiments were carried out in the collection and transport of fry in February, 1965. Further studies were required to ascertain the possibilities of culturing them.

3.6 Fish Farming Policy

3.61 Areas Suited for Fish Farm Development

As a follow-up of the recommendations made on the future fish farming policy in Uganda (FAO/UN, 1965) a survey was made to determine the profitability of, and priority to be given to, fish farming in various parts of the country. The survey showed that the country could be divided into three categories of areas from the point of view of fish farming as follows:

1. Areas where there is no economic justification for official encouragement of fish farming, due to the ready availability of fish at low prices from nearby lake sources. The price of lake fish in such areas is lower than that at which pond fish can be produced (e.g. Masaka district of Uganda).

2. Areas where fish farming currently produces fish at prices competitive with those of lake fish, but where unsuitable topography limits the expansion of fish farming (e.g., Bugisu, Eastern Uganda).

3. Areas where fish farming is not carried out on a large scale at the present time, but where it can be developed as an important source of fish in the future. A typical example of this area is Kigezi.

Because of the very great potential for fish farming in the Kigezi District, the requirements and implementation of fish farming policy in this area may be discussed in greater detail.

In his study of fish marketing in Uganda, Crutchfield (FAO/UN, 1959) considered that it was impossible to bring fish into Kigezi at economic prices, because of the distance from Lakes Edward and Nakivali. He was of the opinion that the only real hope for large increases in fish supplies lay in a determined program of pond construction and research. The conditions observed by Crutch-field were found to prevail also during the period of the present survey.

Kigezi has a population of 500,000 and has the lowest per capita income in Uganda (1/10 of that in Buganda in 1961). Because of this, and the prevailing high prices of fish from the main lakes, the regular purchase of fish is beyond the means of the majority of people and their consumption of fish is negligible. To increase the consumption up to the country average of 20 lbs per capita, a production of 5,000 tons per year will be required.

As suggested by Crutchfield, fish farming on a large scale could be developed as a major source of cheap fish for the population. Observations in the existing ponds show that the yield may be as high as 600–700 lbs/acre/annum without feeding. Large tracts of swamps suitable for the construction of fish ponds, occur in the district. In view of these, it is clear that this is the most important district in the country for large-scale fish farming.

At present there is a steady increase in the number of ponds in the district, but these are constructed manually and are of small size, between 1/20 and 1/10 acre. So, the production of pond fish contributes very little to the marketable supplies in the district, and the cash return to the individual farmer from his ponds is very small.

The development of fish farming on a large scale in the past has been hindered by the lack of machinery for pond construction. Under the existing conditions, it is very unlikely that a program of fish farming expansion can be implemented in the area without government assistance. The Government should set up a pond construction unit consisting at a minimum of a fish farm engineer, surveyor and mason, with essential construction equipment.

Stocking and management of ponds should be brought under the general supervision of the officer in charge of the regional fry centre. This will enable the acceleration of pond construction in the area and facilitate the construction of ponds of standard designs, so as to enable proper management and consequent increased production. It must be emphasized here that fish farming can form the major source of fish in the district, with the fisheries of Lake Bun Bunyoni, Lake Mulehe and Lake Mutanda as supplemental sources.

3.62 Economic Feasibility of Fish Farming

The cost of pond construction with mechanical equipment in Uganda has been calculated to vary from Sh. 2,000/- to Sh. 2,200/- per acre, depending on soil conditions, on the basis of the cost of pond construction carried out by the Uganda Company in its estate in Toro, under the direct supervision of the expert, using mechanical equipment in typical low-lying swampy areas in Kigezi district. The highest figure may be taken into account for costing purposes. This, as well as the cost of fry and pond maintenance, are more or less the same in most parts of Uganda. The variable factors in different areas appear to be the expected yield of fish in potential markets. These factors are of significant importance in the development of planned fish farming.

The economics of fish culture under local conditions may be summarized as follows:

As the maintenance labour for the ponds is contributed by the owner and his family members, total cash expenditure per annum works out to Sh. 405/-, leaving Sh. 195/- cash income per annum per acre.

It may be mentioned here that in some parts of Kigezi where topographical conditions are not very favourable and transportation is difficult, the cost of pond construction may be as much as Sh. 3,000/- per acre. The ponds can even then be operated on a profitable basis, but other terms of financing, e.g. longer periods of repayment, may be arranged in such areas to maintain an attractive level of annual income.

Production from fish ponds can be expected to be raised above the 600/lb/acre/annum given above, but supplemental feeding with brewery wastes and low grade grains unsuitable for human consumption will have to be adopted for this purpose.

The area available for fish pond construction in Kigezi is considerable. In a report by Sir. A. Gibbs (1955) entitled “Water resources survey of Uganda”, the total area of swamps in Kigezi has been given as 27,920 acres, and that of the area reclaimable for agriculture as 10,820 acres. It appears that a larger area than reported is available for reclamation into fish ponds, especially in regions where draining could reduce the level of the water table. A new survey, however, would be required to determine this.

Swamp reclamation for agriculture has been undertaken by the Water Development Department in Kigezi at a cost of Sh. 2,000/- per acre, which is very similar to the cost of fish pond construction. However, when agricultural production from reclaimed land is compared with fish production from ponds constructed in the area, it is seen that reclamation of swamps into fish ponds is more economic.

The main agricultural crops raised on reclaimed land in Uganda are: sweet potatoes, maize, surghum and beans. The cash income from these crops is low, ranging from Sh. 136/- per acre for maize to Sh. 340/- per acre for sweet potatoes, and averaging only Sh. 289/- per acre as compared with the Sh. 600/- per acre obtained by fish farming, as mentioned earlier. Furthermore, the estimated income from fish farming takes into account the repayment of construction cost and interest on it over a period of 10 years, whereas the income from agriculture has been calculated without providing for these. It is important to note that though sweet potatoes yield the highest cash income amongst the crops, this income is attained only in the third year of cultivation, the income being only half of this in the first two years. The fact that the agricultural produce consists of low-protein high-carbohydrate food is of special significance in view of the need for improving the nutrition of the local population. It is also to be noted that swamp lands when reclaimed for agriculture are under production only for six months in a year, whereas when utilized for fish culture, the land is utilized for food production more or less continuously.

3.7 Training Program

The success of a large-scale fish farming program in Uganda will depend to a large extent on the availability of trained personnel. The Fisheries Training Institute, now being established at Entebbe, is expected to start functioning in 1967. It is recommended that a course in practical fish farming be included in the curriculum of this Institute. The course should be mainly for the junior staff employed at Kajansi Station and the regional fry centres.