Productive inland fisheries based on small fish are found in various parts of the world. A well-known example is the fishery for the small cyprinid Mirogrex terrasanctae on Lake Kinneret which dates from biblical times (Petr and Kapetsky, 1983). Another interesting example is from the Philippines where highly productive fisheries were based on the goby Mistichthys luzonensis which, at 12.5 mm mean length, is the smallest commercially exploited fish (Gindelberger, 1981).
Small fish have not been intensively exploited in African waters although there are some traditional fisheries which utilize them. The long-established one based on the Lake Tanganyika clupeids is well-known and Poll's (1952, p. 154) description of it as “de quelque sorte le sport national des indigènes du Tanganyika” suggests that is played an important social as well as economic role. Clupeids were also taken in the West African rivers, as in the “Atalla” fishery of the lower Niger (Awachie and Walson, 1978). In Lake Chilwa, the small cyprinid Barbus paudinosus was a major component of the fishery (Eccles, 1970) and Barbus spp. are regularly exploited in African rivers.
There has been a recent upsurge of interest in the small pelagic fish of African lakes and their commercial value is becoming more apparent. There appears to be two main reasons for this. The first is that the rapid growth of Africa's human population has led to increased exploitation of many inshore fish stocks. In Lake Victoria, for example, the tilapia stocks were thought to have been over-exploited as early as 1960 (Fryer and Iles, 1972) and increased fishing pressure since then has left the pelagic stocks, mainly the cyprinid Rastrineobola (Engraulicypris) argentea, as the only possible avenue for fisheries expansion apart from Haplochromis.
The second factor is the considerable increase in surface water resources brought about by the construction of several large reservoirs during the last three decades. In some of these, productivity was depressed because no indigenous fish could colonize their pelagic waters. Pelagic fish of the great lakes seemed to be good candidates for introduction into these lakes and the success of the Tanganyika sardine Limnothrissa miodon in Lake Kariba (Marshall, 1984) showed the possibilities of translocating and commercially fishing such species. The establishment of such fisheries might compensate local populations, to some extent, for the loss of traditional homes and other social problems caused by large reservoirs (Balon, 1978).
Pelagic communities in African inland waters consist, as a rule, of small zooplanktivorous species and a group of predatory fish which feed upon them. This paper is concerned mainly with the small fish which are the basis of the system. In addition to being small these fish usually have a short life-cycle and high productive potential. There are four major families which include species with actual or likely value as pelagic fish, and these are discussed in more detail below.
Small characids, mostly Alestes spp., are widely distributed in African waters but have generally not become pelagic. In Lakes Malawi and Tanganyika, for example, they tend to be restricted to inshore areas or affluent rivers (Jackson et al., 1963; Poll, 1953). Nevertheless they might have been expected to become pelagic in reservoirs, and Balon (1974) suggested that Alestes lateralis would have done so in Lake Kariba had it not been for the introduction of Limnothrissa. This view was disputed on the grounds that the original samples were taken in shallow water, that there were no data to show that A. lateralis was planktivorous and that ample time had elapsed for it to have become pelagic in the 10 years between Kariba's creation and the sardine's introduction (Marshall, 1984).
Diet is probably the main reason why Alestes have not become pelagic. Most species examined ate a wide variety of food but were mainly insectivorous, with flying terrestrial insects being a significant component (Poll, 1953; Reynolds, 1973; Marshall and vand der Heiden, 1977; Paugy, 1978, 1980/1980a). Although plankton was occasionally taken, only A. baremoze was able to utilize it extensively (Paugy, 1978). A consequence of this insectivory is to confine these fishes to inshore areas where insects are likely to be most abundant.
The breeding habits of Alestes are poorly known but may be another reason why they have not become pelagic. Some, like A. imberi, seem to be potamodromous and require running water in which to breed (Bowmaker, 1973). Others, like A. lateralis, remain in weed beds throughout the year and may require vegetation in which to spawn (Balon, 1971; Bowmaker, 1973). Either of these reproductive strategies would inhibit the development of pelagic forms.
The cyprinids are widely distributed in Africa and occur in almost all water-bodies. Several pelagic species of Engraulicypris and Rastrineobola have developed in Lakes Malawi, Tanganyika, Victoria, Mobutu Sese Setu and Turkana (Turner, 1982a). The only two to have been commercially exploited are R. sardella of Lake Malawi and R. argenteus of Lake Victoria (Petr and Kapetsky, 1983).
The numerous small cyprinids of African rivers generally have not become pelagic in reservoirs but one species which has flourished in small dams in Zimbabwe is Neobola (formerly Engraulicypris) brevianalis. It has been translocated to provide food for black bass and trout in some reservoirs but has not been commercially exploited (unpublished data).
The numerous small Barbus spp. have not produced pelagic forms in any large African reservoirs, although Reynolds (1973a) studied them in Lake Volta in the hope that they might have some fisheries potential. In most lakes they are confined to inshore areas or affluent rivers, possibly because most seem to be strongly rheophilic and require riverine conditions. An exception is Lake Le Roux in South Africa where B. anoplus occurs in all areas (Cambray, 1983). This reservoir is narrow and fjord-like with a mean breadth of only 1.4 km (Allanson and Hahndiek, 1979); it is also very turbid and has extreme and irregular water-level changes. It thus resembles a very large river and so is suitable for Barbus.
Most cyprinids seem to be strongly rheophilic which affects their distribution in large lakes and reservoirs (Poll, 1953; Jackson et al., 1963; Begg, 1974). Even those which can penetrate deep water, such as the mpasa Opsaridion microlepis of Lake Malawi still have to move up rivers to breed (Tweddle, 1983). The small Barbus spp. have catholic feeding habits with zooplankton as a relatively minor component of their diet (Cambray, 1983). These facts suggest that the evolution of pelagic cyprinids might be a relatively recent phenomenon and could explain why some cyprinids seem to be less efficient pelagic species than the clupeids (Turner, 1982a).
The cichlids are widespread and very successful in African waters. The adaptive radiation of Haplochromis in the great lakes is a distinctive feature of the African freshwater fish fauna (Fryer and Iles, 1972) and they have been able to exploit almost every food source, including zooplankton. Pelagic forms occur in several lakes and the best-known are the “Utaka” of Lake Malawi, primarily Haplochromis quadrimaculatus but with other Haplochromis spp. as well (Jackson et al., 1963). The “Utaka” are abundant in inshore areas but this decline rapidly with distance from shore and these fish do not fully occupy pelagic waters (Fig. 1).
The reason why cichlids have not become truly pelagic is probably their specialized breeding habits. Haplochromis males, like most other cichlids, establish territories marked by a “nest” to which females are attracted. After a brief courtship the females take the fertilized eggs into their mouths and move away (Fryer and Iles, 1972). This behaviour means that cichlids will be restricted to inshore areas, especially as their breeding seasons may extend over several months. Cichlids can, of course, establish themselves in the pelagic waters of shallow lakes, as in Lake George (Gwahaba, 1975). The recent discovery of mid-water spawning by Haplochromis chrysonotus in Lake Malawi (Eccles and Lewis, 1981) suggests that true pelagic behaviour may be developing in some cichlids.
This is by far the most significant pelagic family and the one which has been studied in most detail. It is primarily a planktivorous marine family which has colonized the rivers of West Africa and the Congo system (Beadle, 1981) right up to Lakes Tanganyika and Mweru where specialized pelagic forms have evolved. The most important group are the 22 species of the sub-family Pellonulinae occuring throughout West and Central Africa (Table 1).
Two specialized lacustrine species, Limnothrissa miodon and Stolothrissa tanganicae, have developed in Lake Tanganyika and form the basis of the pelagic fishery there (Petr and Kapetsky, 1983). Three other lacustrine forms occur in Lake Mweru but virtually nothing is known about them although Jackson (1961) suggested they were neither large enough nor abundant enough to be of commercial value. Further investigation of these species is clearly required.
The remaining pellonulines are riverine but appear to have the potential for becoming pelagic. This is well illustrated in West Africa where the riverine clupeids developed pelagic stocks in Lakes Kossou (Pellonula afzeliusi), Volta (P. afzeliusi, Sierathrissa leonensis) and Lake Kainji (P. afzeliusi, S. leonensis and Cynothrissa mento) (Vanderpuye, 1973; Otobo, 1979). This suggests that some of the other species listed in Table 1 might do the same and this should be borne in mind if reservoirs are proposed in rivers where they occur. Some might also be suitable for translocation and further research on these species would be valuable.
The success of the clupeids as pelagic fishes is probably due to several factors. They seem to be primarily planktivorous, but able to utilize other food if neccessary (Matthes, 1968; Otobo, 1979; Begg, 1974a; Cochrane, 1978; Gliwicz, 1984; Reynolds, 1970). It is interesting to note the Stolothrissa, which Matthes (1968) considered to be the most specialized planktivore, did not establish itself in either Lakes Kivu or Kariba whilst the more generalized Limnothrissa did (both species must have been present when the fry were introduced).
Another interesting aspect of clupeid biology is that they seem to be able to adjust their growth rates to the food supply. This is well illustrated by Limnothrissa which in Kariba only grow to half the size that they do in Kivu or Tanganyika (Fig. 2). Marshall (1984) postulated that food supplies in Kariba were poorer because the reservoir was less fertile than the other two (Kariba conductivity = 80 μS cm-1; Tanganyika = 500; Kivu = 1 200). Linked to this is a reduction in breeding size; in Tanganyika Limnothrissa mature at 75 mm (Ellis, 1971) but in Kariba they rarely reach this size and probably mature at 40 mm (Begg, 1974a).
It is not known if other clupeids show similar plasticity but it seems likely as the morphology of the South African estuarine species Gilchristella aestuarius varies according to its food supply (Blaber, Cyrus and Whitfield, 1981). The ability to change growth rates and body form is probably a mechanism to enable these fish to maintain a high biomass in the face of limited food resources. This would enable them to adapt to a wide variety of water-bodies.
The most important characteristic of small pelagic species is their very high potential productivity. This is the factor that makes their economic exploitation possible. Their high productivities is a result of their short life cycles and consequently very high P/B ratios which are inversely related to longevity (Leveque, Durand and Ecoutin, 1977). The P/B ratio for Stolothrissa was 3.9 over a 12-month period but mortality rates are so high that only 2.5% of the original number were survivors at the end of this period (Coulter, 1981). The implication is that production is about 4 times greater than the biomass and that very high yields are possible. Coulter (1970) suggested that commercial yield could well be equal to production for these species.
This concept can be quantified by using Pauly's (1982) equation, in which
MSY = 2.3 w-0.26 Bv | ||
| Where MSY | = | maximum sustained yield (kg ha-1yr-1) |
| w | = | mean weight (g) |
| Bv | = | unexploited biomass (kg ha-1) |
It can be seen that yield increases exponentially as individual weight decreases, assuming Bv to be equal at all size ranges (Fig. 3).
This can be seen in most pelagic communities where unfished stocks support very high predator populations. In Lake Tanganyika predators made up about 60% of the catch to begin with (Coulter, 1970) whilst in Lake Malawi predatory Rhamphochromis spp. were about as abundant as their prey, Engraulicypris (Walczak, 1982). The predators in Lake Tanganyika decreased under fishing pressure although the total catch increased. These were presumably the fish that had previously been consumed by the predators (Coulter, 1970). This interpretation is, however, somewhat open to question because the period of observation was not sufficently long to include cyclic fluctuations in abundance of predators. Roest (pers. comm.) notes that over the period 1955–80 the abundance of Lucio lates, actually increased while Lates decreased in the exploited areas.
Small pelagic species may also have an impact on nutrient dynamics, especially in reservoirs. Some evidence suggests that Limnothrissa plays a major role in Lake Kariba as a nutrient store and so reducing losses throught the outflow (Marshall, 1970; 1984). It has also been postulated that this fish contributed to the decline of the floating fern Salvinia molesta, which had been so troublesome in the early days of Kariba, by retaining nutrients which would otherwise have been available to the plant (Marshall and Junor, 1981). Much more research needs to be done on this aspect but it is interesting to speculate on the contribution small fish might be able to make to the control of troublesome plant growths.
A number of lakes and reservoirs have pelagic stocks, but their fisheries are unevenly developed (Table 2). Various factors have promoted or inhibited this process and each can be considered in more detail.
The oldest and best-developed African pelagic fishery is on this lake and currently produces about 73 000 t or 22.5 kg ha-1 annually (Petr and Kapetsky, 1983). The establishment and expansion of this fishery was undoubtedly facilitated by the fact that a traditional fishery already existed (Poll, 1952) and that very high yields were possible.
Stolothrissa tanganicae is the most important species in the fishery and has consequently been studied in greatest detail (Matthes, 1968; Ellis, 1971; Chapman and Van Well, 1978; Roest, 1978). The biology of Limnothrissa is less well-known except for some details on growth and breeding (Matthes, 1968; Ellis, 1971). Some workers have investigated the biology of the four predatory Lates spp. (Chapman and Van Well, 1978a; Coulter, 1977; Ellis 1978) whilst relationships within the pelagic community as a whole were described by Coulter (1970). Productivity and yield estimation have also been investigated (Coulter, 1981) as have the economic aspects and future prospects of the fishery (Herman, 1978).
The most outstanding feature of the Tanganyika fishery is its very high potential productivity. Coulter (1977; 1981) attempted to estimate production by using Gulland's (1971) equation in which
| Y = M · X · B | |
| where | Y = yield (kg ha-1) |
| M = Natural mortality | |
| X = proportion of the stock available to the fishery, usually X = 0.5 | |
| B = unexploited biomass (kg ha-1) |
Sardine biomass is estimated to be about 160 kg ha-1 and M = 5.2 (Coulter, 1977) which would suggest that the yield might be about 400 kg ha-1. The P/B ratio is about 4 and production has elsewhere been estimated at 350 kg ha-1 (Coulter, 1981). The yield might then amount to 1.1 million tons for the whole lake, which is far in excess of the present yield. However, much depends on how representative early estimates of biomass really are. Given long-term cyclic fluctuations in biomass and the possibility that the estimates were made during the cycle when biomass was at a peak, then the 1.1 million ton estimate for yield could be too high. Coulter (1977) suggested that the stock might be sufficiently resilient for yield to be determined empirically by fishing without causing long-term transformations. It may not, of course, be economically feasible to establish and operate such an intensive fishery again, because long-term fluctuations in stock availability would create economic difficulties for fishing intensively when on the low side of the abundance cycle.
The fishery on this lake is poorly-developed except in the south, but studies suggest that an annual yield of 30–40 kg ha-1 may be possible (FAO, 1982). In inshore waters most of this would be “Utaka” (Haplochromis spp.) but Engraulicypris sardella would be most abundant in open waters.
The potential of Lake Malawi seems very low in comparison to Lake Tanganyika and Turner (1982a) postulated that this was because Engraulicypris was not an efficient planktivore. He concluded that most pelagic production was being channelled into large zooplanktonic animals or the lake fly Chaoborus, neither group being preyed upon by Engraulicypris. By contrast, these organisms were absent from Lake Tnaganyika having probably been eliminated by the pelagic clupeids.
Turner (1982a) felt that the introduction of the Tanganyika clupeids into Lake Malawi might greatly improve its productivity. He acknowledged that this might have unforeseen effects on the lake's fauna and was not a step to be lightly contemplated.
The fishery for Rastrineobola (Engraulicypris) argenteus on this lake would appear to be potentially very valuable but is poorly-developed. Provisional figures show that about 9 500 t were taken from the Kenyan sector in 1982 but it was not apparently exploited in Uganda (Ssentongo, 1983).
A major drawback is that there are presently very little data available on E. argenteus or any pelagic Haplochromis spp. This is urgently required before any further developments can take place. Productivity is difficult to estimate but a conservative estimate of 30 kg ha-1 is probably realistic; the total yield would then reach 200 000 t annually. Turner's (1982a) comments about introducing clupeids into Lake Malawi may also apply to Lake Victoria and this, again, is an area that would repay further study.
This lake is comparable to a reservoir in many respects as it is geologically very young, with few fish species of which none were pelagic. The first artificial introduction of Tanganyika clupeids was made into Lake Kivu, between 1958 and 1960, when large numbers of fry were stocked (Collart, 1960). It was hoped that Stolothrissa would become established but it was, in fact, the less specialized Limnothrissa which did so (Frank, 1977). Recent studies have shown that it is well-established and the artisanal fishery that has been established has attained yields of 42 kg ha-1 which would amount to 13 500 t if the whole lake was fished (Spliethoff, de Iongh and Frank, 1983).
The population around the lake has no tradition of fishing, fish are not popular as food and fishing material is scarce and expensive (Petr and Kapetsky, 1983). These factors have inhibited rapid development of this fishery, but once they have been overcome it might be possible to attain the yield of 30 000 t (or 110 kg ha-1) estimated by Welcomme (1972) since Kivu would appear to be at least as rich as Lake Tanganyika.
This was the first reservoir to be stocked with clupeids and the successful fishery that developed is now well-known (Marshall, 1984; Marshall, Junor and Langerman 1982). Limnothrissa was the only species introduced; it colonized the lake within 2 years and commercial fishing began after 6 years.
Sardine production reached 12 000 t in 1981 but declined in 1982 and 1983 (Fig. 4). This suggests that the population was fully-exploited but it may also be a result of the hydrological regime during this period. Sardine abundance in the lake seems to be closely related to river inflow and nutrient supplies (Marshall, 1982). The 1981–82 and 1982–83 rainy seasons were very poor and the river flows the lowest on record. Because sardine abundance decreases along with river flows the reduced catches were expected (Marshall, 1981).
Studies to estimate Kariba's potential are in progress and some data suggest that it is very high (Marshall, 1984). These data should be treated with caution, however, as they are still to be verified. In the last good year, 1982, the eastern Sanyati basin produced 7 400 t or about 74 kg ha-1 (Marshall and Shambare, 1983). If this was extrapolated over the whole lake then the catch would reach 40 000 t or about 4 times the present yield.
A secondary effect of the sardine in Kariba was that the Tigerfish Hydrocynus vittatus became partially pelagic, preying on the small fish in open water. Although it declined under intensive fishing pressure, it increased in the gillnet catches (Junor and Marshall, 1979). Other species, such as Synodontis and Eutropius, also now feed on sardines and the survival of these fish has probably been enhanced which may contribute to an improvement of inshore productivity. This is an important factor in Kariba because production from this zone is low (Marshall, Junor and Langerman, 1982).
The colonization of this reservoir by Limnothrissa from Kariba is an interesting example of its hardiness and adaptability. Soon after their establishment live sardines were found in the Zambezi river below the Kariba dam, having survived passage through the hydroelectric turbines, and it was suggested that they might move down the river to Cahora Bassa (Junor and Begg, 1971; Kenmuir, 1975). This in fact happened and there is now a substantial stock in the lake (Bernacsek and Lopes, 1984) but for various socio-economic reasons a fishery has not yet been established.
The potential yield has been estimated to be about 8 000 t or 30 kg ha-1 (Bernacsek and Lopes, 1984). A recent acoustic survey produced estimates of a 2 800–17 600 fish per hectare (Lindem, 1983); since these fish weigh about 1.0 g this amounts to 2.8–17.6 kg ha-1. The mean was about 10 kg ha-1 which was very low and suggests that production might not exceed 30 kg ha-1 (assuming their P/B ratio to be about 3.0).
Two factors may have influenced this estimate, however. The first is that Limnothrissa may be as highly seasonal in Cahora Bassa as it is in Kariba (Marshall, Junor and Langerman, 1982). February is a period of low abundance in Kariba and this may also be the case in Cahora Bassa. On the other hand productivity may in fact be this low because of the heavy clay load in the water which reduces light penetration and may inhibit primary production (Gliwicz, 1984; Bernacsek and Lopes, 1984).
Riverine clupeid populations existed in the Niger and Volta rivers before these reservoirs were constructed. Once the lakes formed, these fish populations grew and occupied the pelagic waters (Vanderpuye, 1973; Otobo, 1979). These fish are taken to a small extent in the littoral artisanal fisheries but large-scale industrial fisheries have not yet been established. It will probably be very difficult to establish such a fishery on Volta because the lake bed was never cleared of trees and these would interfere with large nets.
Clupeid biomass on Volta is unknown (C.J. Vanderpuye, personal communication) but an estimate of 3 140 t mean biomass was obtained from Kainji (Otobo, 1979). Assuming that the P/B ratio of these fish is similar to those in Tanganyika then the yield might be as high as 10 000 t (or 79 kg ha-1). This is close to the yield from the Sanyati Basin of Lake Kariba and so is probably realistic.
Production in Volta might be similar or possibly even higher since it has a higher morpho-edaphic index than Kainji (Henderson and Welcomme, 1974). Thus a yield of 60 000 t might be possible (at least when the lake is at its full surface area of 8 482 km2).
From the foregoing it is clear that pelagic populations in most African waters are still greatly under-exploited. There are also other pelagic populations, such as the Rastrineobola (Engraulicypris) spp. in Lakes Turkana, Edward and Mobutu Sese Seku or the clupeids in Lake Mweru, but very little is known about their biology, abundance or commercial potential.
The possibility of introducing clupeids into some of the Great Lakes in order to improve their pelagic production has already been mentioned. The dangers of interfering with the biota of these lakes has been discussed (Fryer, 1972) and, although the argument for introducing clupeids into them is sound, most fisheries biologists still have reservations about doing so (Turner, 1982a).
These inhibitions are less strong where reservoirs are concerned as reservoirs have extensive environmental impacts in any event (Petr, 1978). The success of Limnothrissa in Kariba has shown the value of such introductions and Eccles (1975) has suggested that a number of other fish from the Great Lakes could also be introduced. There seems to be no reason why these fish should not be stocked into those large reservoirs which do not have them, like Nasser-Nubia. Medium-sized reservoirs, such as Itezhi-Tezhi and Kafue Gorge in Zambia, Nyumba ya Mungu in Tanzania or Kamburu in Kenya might also be suitable.
Indeed, in view of the adaptability of the clupeids it might be possible to introduce them into any reservoir regardless of size. There is an urgent need for further studies on this family especially to assess their commercial potential and suitability for stocking. Most clupeids occur in relatively warm waters and their temperature tolerances are not known. This needs to be investigated and some may be suitable for high altitude lakes, such as Lake Tana or the reservoirs in Madagascar, which presently have a very limited fish fauna.
This brief review of small pelagic fish has shown that they have an important role in increasing supplies of fish in Africa. The clupeids are the family with the most potential, and further studies on this group are urgently required. The clupeids are also more easily harvested because they are readily attracted to lights, which is the commonest method of catching these fish.
The need to use light attraction is perhaps one of the major drawbacks in exploiting pelagic fish. Large, relatively expensive boats are needed to increase catches significantly and to be able to venture into the open waters of large lakes. These capital requirements are often beyond the means of local inhabitants and in many countries the prohibitive cost or non-availability of gear such as lighting plants or netting materials constrains development. These factors have probably retarded industrial-scale fisheries in many areas. Artisanal nearshore fisheries can be practised with relatively simple gears, such as kerosine lights in place of electric plants for fish attraction.
There are many advantages in utilizing these fish. Their high productivity means that catches will be good enough to repay the initial capital investment in a short time. A very palatable and easily marketed product can be prepared at very low cost by sundrying and costly processing facilities are not required. Finally, the bycatch of predators, such as Lates in Tanganyika or Hydrocynus in Kariba, is often more valuable than the catch of the small species as they are sought-after and highly palatable fish.
