As gillnet yields of Nile perch stagnated in the middle of the 1990s, the hook and line fishery intensified in Lake Victoria. Frame surveys show that the number of hooks used in this fishery grew tenfold in the Tanzanian waters between 1992 and 2002. The increased use of endemic fish species as bait has gone largely unnoticed, but is thought to impede the recovery of these species in the lake. A survey was conducted in March 2003, covering 51 sites, to assess the impact of this fishery targeting Nile perch on the endemic species. The species used as bait were juvenile Clarias (41%), haplocromines (37%), dagaa (Rastrineobola argentea) (10%), Synodontis (3%), Mormyrus (2%), and Labeo (1%).
The total annual weight of the mostly undersized and illegally caught (beach seining, gillnets below 3″ mesh) baitfish was estimated at 7,465 t. The size of the Nile perch caught is bait-dependant, with catches using Clarias bait having a modal length at 61–62 cm TL of which 20% is above the legal upper size limit. Catches with haplochromines and dagaa bait have a mode at 49–50 cm TL, with 62% below the legal lower size limit. Seasonal variations in the fishery and management implications of its legalization are discussed, while the threats to biodiversity conservation and the stability of the ecosystem are addressed.
The original fish fauna of Lake Victoria comprised over 28 genera with more than 350 species. Cichlids formed the majority of these, with about 300 haplochromine and the two tilapiine species Oreochromis esculentus and O. variabilis (Graham, 1929; Greenwood, 1974, 1981; Witte and Van Oijen, 1990). In the 1950s, the fishery targeted the tilapiines and Labeo victorianus (Worthington, 1933; Cadwalladr, 1965; Fryer, 1973). Other local fisheries existed for Protopterus aethiopicus, Bagrus docmak, Clarias gariepinus and species like Barbus altianalis and Mormyrus kannume (Garrod, 1961). In 1969/1971, haplochromines constituted 80% of the estimated 750,000 t fish biomass of Lake Victoria (Kudhongania and Cordone, 1974).
In order to compensate for the dwindling abundance of the endemic tilapia species, Tilapia melanopleura, Tilapia zilii, Oreochromis leucostictus and Oreochromis niloticus were introduced into the lake (Welcomme, 1968). Nile perch was also introduced in the late 1950s and the early 1960s (Hamblyn, 1961; Arunga, 1981; Welcomme, 1984). Catch rates and total yields continued to decrease for the next twenty years. Haplochromines were the only under-fished stocks, due to their small size and bony texture (Scully, 1975). The decrease in the catch rates in the late 1960s led to the introduction of a number of additional gears such as small-meshed gillnets (1.5–2 inches) which caught smaller fish such as Synodontis spp., Schilbe intermedius, Barbus profundus, Brycinus spp. and the haplochromines (Marten, 1979; Scully, 1975). Beach seines were increasingly used in the early 1970s to catch the haplochromines, but also caught large numbers of spawning tilapiines and their juveniles (Marten, 1979), escalating the effects of overfishing. Simultaneously, light fishing for dagaa (Rastrineobola argentea) developed in Tanzania and Kenya (Marten, 1979; Okedi, 1981). This fishery uses very small mesh sizes (8–13 mm), detrimental to juveniles of both haplochromines and tilapiines. The use of baited hooks also increased, but received little attention. By 1987, Nile perch contributed 60% to the total yield in Tanzanian waters and haplochromines less than 10% (Ligtvoet et al., 1988; Bwathondi, 1990; Witte et al., 1995). Although Nile perch preyed predominantly on haplochromines, all other species were also recorded in its diet (Gee, 1964, 1969; Hamblyn, 1966; Okedi, 1970; Ogari, 1988; Hughes, 1986; Ogari and Dadzie, 1988; Ogutu-Ohwayo, 1990; Mkumbo and Ligtvoet, 1992).
From 1995, heavy fishing for Nile perch was beginning to take its toll as water quality parameters showed some improvement (Lipiatou et al., 1996; Mkumbo, 2002). As a result, a number of endemic species were recovering (Bayona et al., 1998; Mkumbo and Ezekiel, 1999; Katunzi et al., 2003).
Currently, these endemic species are in high demand as baitfish for the long-line fisheries, and it is feared that their recovery may be short-lived. This practice may disrupt the trophic interactions (Hart, 1995) and efficiency of the ecosystem as they are feeding on the lower levels of the food web (Wanink, 1998). The mixed trophic impact routine (ECOPATH II, Moreau, 1995) showed that a decrease of the biomass of these recovering species may result in an increase in zooplankton and phytoplankton and worsen the eutrophication problem in the lake. At the present time, there is a considerable commercial fishery for haplochromines during the full moon using dagaa seines (Mkumbo, LVFO, Jinja, Uganda, personal observation in Speke Gulf). This study attempts to assess the impact of the bait fishery on the native stocks and recommends management strategies and/or alternative sources of bait (Mkumbo and Mlaponi, 2003).
Study area and data collected
This study was conducted at 48 sites, 45 of which are landing beaches on Lake Victoria; three additional sites were located along river tributaries, where bait was collected (Figure 1).
To assess the extent of the hook fishery, the total number of fishing canoes and the number of canoes targeting the resurging species, the type and source of bait and the means used to catch the bait were recorded as well as the hook size. Beach Management Unit (BMU) leaders and long-line fishermen were interviewed. Their general perception of the sustainability of the bait and long-line fishery was also assessed.
To determine the population structure of the catch and to assess the performance of the fishery, the length of individual fish was measured in cm TL and total weight was recorded where possible. A weighing scale and a measuring board were used. Fishermen were interviewed in order to determine the types and quantities of bait used for each canoe landing, as well as the frequency of fishing, using the specified baits. This information was later used to work out the average number of baitfish used per canoe and per week.
Estimation of the importance of the bait fishery in the Tanzanian waters of Lake Victoria
The fisheries statistics of Lake Victoria, Tanzania are partly based on the results of Frame Surveys. Reports of these surveys show a marked and continual increase in the numbers of fish hooks used between 1992 and 2002 (Table 1). The first enormous increase was observed in the middle of the 1990s as gillnet yields decreased and fishermen adopted additional long-line fishing.
A total number of 4,684 fishing canoes were recorded on the 45 landing beaches surveyed. These represent about 30% of all fishing canoes (15,489 canoes) recorded during the Frame Survey of 2000. From the sampled canoes, 54% (2,552 canoes) took part in the baited hook fishery. In prominent landing sites for long-line fishing, 80–90% of all the canoes use baited hooks.
The 83 long-line fishermen interviewed reported to use 150–2000 hooks per canoe. In total 48,022 hooks were recorded targeting Nile perch and using endemic species as bait, with approximately 600 hooks per canoe. The hook size used depends on the size and type of bait and ranges from 5 to 13. The most commonly used were sizes 9 to 11; sizes 5 – 7 are used with Mormyrus and Labeo baits from paddled canoes. Sometimes, as little as size 1–3 hooks are used.
Types and sources of bait
Juveniles of Clarias species rank first as bait for the Nile perch hook fishery (used by 41% of the canoes). These included juvenile Clarias gariepinus, Clarias alluaudi and Clarias wermeri. They are used mostly during the dry season (May to October). At some sites and in most of the islands they are used throughout the year.
Haplochromines from the rocky areas, mostly Haplochromis nyererei, Neochromis nigricans, and Paralabidochromis ‘yellow rockpicker’ and those from the inshore areas, which include H. nigricans, H. laparogramma and H. microdon, contributed about 37% of the bait reported in the hook fishery for Nile perch. These are used whenever Clarias is not available and in the minor landing sites.
Dagaa (Rastrineobola argentea) was also reported to be used as bait for the Nile perch fishery with 10% frequency, but only when the other two bait species are not available.
Other endemic species are also used but only depending on their availability. Synodontis species were recorded (3%). Mormyrus kannume (2%) and Labeo victorianus (1%) were recorded at beaches with adjacent rocky areas or rocky islands. At only one landing site, chopped pieces of Oreochromis niloticus and chicken were reported to be used.
Source of baitfish and collection techniques
Juvenile Clarias are collected in almost all water bodies where they occur naturally, while some reportedly originate from aquaculture abroad. Beaches bordering Kenya received additional bait from across the border. On the eastern and southern sides of the lake, Clarias baitfish are collected in main rivers, their tributaries and swamps as well as in seasonal rivers. Sometimes, passenger seats are removed from minibuses which transport 70–80 30-litre buckets, holding about 1000 baitfish each.
At some beaches juvenile Clarias are reported to be transported from hatcheries in Rwanda and Uganda. Trucks from Rwanda reportedly carry up to 200 containers of 30 litres (referred to as diabas), each with up to 1000 baitfish. These were to be distributed by canoe to different sites. Six to eighteen diabas were reported to be transported daily across the Nyakaliro ferry, each with 1000 baitfish, except during heavy rains and very dry spells. By 1998, 70 to 80 20 litre buckets, each holding about 500 fish, were received on a daily basis. Traps, cloth like bed sheets, mosquito seine nets or handpicking are used to catch/collect baitfish. These are kept in 20 or 30 litre buckets of which the water is exchanged once a day. They are fed biscuits and can stay alive while waiting to be used.
Haplochromines are collected from the main lake either by using beach seines or gillnets of 3/3 -1″ mesh sizes or by handlining in the rocky areas, using earthworms and Caridina nilotica.
Synodontis species are collected from bays on the Tanzanian side and from rivers and swamps in Kenya. Beach seines and small meshed gillnets are used in the rainy season along almost all river mouths.
Mormyrus and Labeo are collected by gillnets of 1½ – 3″ mesh size from the main lake in areas with submerged rocky substratum or near rocky islands, mainly at the height of the rainy season.
Size range, numbers and estimated weight of the endemic baitfish species used
Sizes range from about 11 – 20 cm TL, with weights of 8 – 34 g and a mean weight of 17.5 g. Fishermen report that only half of the hooks in a canoe would be baited with Clarias, due to scarcity of the baitfish and its cost (they are sold at TShs. 70/ = to 120/ =).
With an approximate number of hooks of 4,000,000, estimating that 41% of all baitfish used is Clarias and only half of the hooks in a canoe being baited, the total number used per day would be: 4,000,000 × 0,41 × 0,5 = 820,000 Clarias individuals. Full-time fishing takes place in the period from May to October each year. The total estimated number of Clarias used in a year would then be: 6 months × 30 days × 820,000 Clarias = 147,600,000 individuals, and their total weight 2,583 tons.
Sizes used range from 7.8 cm to 11.8 cm TL, with weights of 7.0 g to 26.0 g and a mean weight of 15.85 g.
When using haplochromines, all the hooks are baited, but fishermen alternate, collecting bait one day and fishing the next day. At an estimated frequency of use of 37%, the total number of haplochromines baitfish used in a day is: 4,000,000 hooks × 0,37 = 1,480,000. Assuming that haplochromines are used in the six months when Clarias is not used, the total numbers of haplochromine baitfish per year would be: 1,480,000 × 6 months × 30 days = 266,400,000; this represents a total weight of 15.85 g × 266,400,000 = 4,222.44 tons.
Little direct information is available on the use of dagaa as baitfish. Their lengths are known to range from 21 mm to 76 mm TL, which, according to the relationship W = 0.00007L2.4375corresponds to weights of 0.072 g to 4.56 g, with a mean weight of 1.54 g. The frequency of use of dagaa as baitfish is about 10% and its duration 4 months, but only during dark moon days, as they are fished with light. The estimated annual number used would therefore be: 4,000,000 hooks × 0.10 × 4 months × 17 days = 27,200,000 fish, weighing 27,200,000 × 1.54 g = 41.89 tons.
Synodontis are reportedly used at 3% and only once a week during the rainy season, so the estimated number of fish used is 0.03 × 4,000,000 hooks × 5 months × 4 days = 2,400,000 individuals per year. At an estimated mean weight of 200 g, the biomass estimate is = 2,400,000 × 200 = 480 tons yr− 1.
The size range of Mormyrus used was 27 to 34 cm TL, weighing 520 to 1000 g, averaging 737 g. Only 2% of the canoes use Mormyrus as bait, and only 1 hook out of 60 is baited with this species, as canoes in this fishery do not employ more than 10 hooks. Assuming 4 days of lightning and thunder per (rainy) month, the number of baitfish and their weight can be estimated as: 0.02 × 4,000,000 × 1/60 hooks × 5 months × 4 days = 26,667 individuals, weighing 26,667 × 737 g = 19.64 tons yr− 1.
The frequency of utilisation of Labeo was only 10% and like the latter species only during the stormy/rainy season. Estimates, therefore, are 0.10 × 4,000,000 × 1/60 hooks × 5 months × 4 days = 13,333 individuals. At a mean weight of 500 g, they weigh 13,333 × 500 = 6.67 tons yr− 1.
Total estimated biomass of the endemic species used as bait
The combined estimates calculated for the individual species groups (Clarias 2,583 t yr− 1, haplochromines 4,222.44 t yr− 1, Dagaa 41.89 t yr− 1, Synodontis 480 t yr− 1, Mormyrus 19.64 t yr− 1 and Labeo 6.67 t yr− 1) show that about 7,464.86 t of the endemic species are used annually as bait in the Nile perch hook fishery. Excluding dagaa, about 7,422.97 t yr− 1 of the threatened endemic species are, therefore, used in the bait fishery.
Population structure of Nile perch caught in the bait fishery
Length frequencies of Nile perch in the catch was only recorded for three bait species, Clarias spp., haplocromines and dagaa. Their combined frequency distribution shows a mode at 55–56 cm TL, while 49% of the catch is below 50 cm TL and 18% is above 85 cm TL (Fig. 2a). The slot-size under regulation is 50–85 cm TL.
Separated by baitfish type, Nile perch caught using Clarias had a mode at 60–61 cm TL (Fig. 2b), with 41% of the catch below 50 cm TL and 20% above 85 cm TL. Haplochromine and dagaa baitfish caught relatively smaller Nile perch with a mode of 49–50 cm TL and 61% of the catch below 50 cm TL, and only 7% (dagaa) – 12% (haplochromines) was above the 85 cm TL (Figure 2c and Figure 2d). It was not possible to get enough length distribution samples for the other baitfish species Synodontis, Mormyrus and Labeo, but fishermen report catching bigger Nile perch with these than when using Clarias bait.
Other fisheries impacting the endemic species
Other fisheries exist which target the endemic species for domestic consumption and not for use as bait. A hook and line fishery using earthworms, haplocromines and pieces of washing soap targeting Clarias gariepinus was also recorded. The size range of the catch is shown in Figure 3a, with a mode below 39 cm TL.
Gillnets with mesh sizes below 5 inches (with a majority of 21/2”) catch a variety of the endemic species in sheltered bays or river mouths. These include Clarias gariepinus with a mode at 38 cm TL (Figure 3b), Oreochromis esculentus with a modal length below 18 cm TL (Figure 3c) and lungfish (Protopterus aethiopicus) with mode below 62 cm TL (Figure 3d). Other species encountered in the gillnet catches were Schilbe intermedius, Brycinus sadleri and B. jacksonii, Mormyrus kannume, Synodontis victoriae, Synodontis afrofischeri and Labeo victorianus.
Status of the baited hook fishery in the Tanzanian part of Lake Victoria
The decline of species diversity in Lake Victoria is usually associated with the increased predation of Nile perch, overfishing and environmental change (Goldschmidt and Wanink, 1993; Acere, 1988), but the impact of the baited hook fishery had so far been completely overlooked. This hook fishery was of little commercial importance before 1994 and targeted only the tilapiine species. The increased demand for Nile perch led fishermen to diversify their fishing techniques and since then there has been a continuous increase in the number of hooks in the Tanzanian part of Lake Victoria (Ligtvoet et al., 1995). In this period, the number of fish processing factories increased to five (Gibbon, 1997; SEDAWOG, 1999; Medard, 2002). With the increase in gear theft, especially gillnets which are more expensive, some fishermen have permanently engaged themselves in hook fishing, despite the seasonal differences in the availability of fish to different gears and in the survival time of baitfish. Some prominent fishermen own both gillnets and hooks, which are used interchangeably during the periods of high catches of the respective gear. Currently, 54% of all the canoes on Nile perch fishery are using baited hooks.
The number of hooks actually used depends on the availability of baitfish and the experience of the baiter, as the time needed to set the hooks depends on experience. However, due to scarcity of baitfish, fishermen are now forced to use even dead specimens. Following the imposition of a lower (and upper) size limit in the Nile perch catch, fishermen currently use hooks of numbers 7–11, as the smaller hooks of numbers 12 and above catch small-sized Nile perch. The small sized hooks, numbers 13 and above, are only used for Oreochromis niloticus, sometimes using haplochromines as bait.
The seasonality observed in the bait fishery is explained by the seasonality of the rainfall and of the environmental parameters. Generally the dry season extends from June to September and the rainy season from October to May, with a dry spell in January/February (Mkumbo, 2002; Akiyama et al., 1977). The rainy period before the dry spell is characterized by light rains (October–December) while the heavy rains are in March–May. Heavy rains are associated with strong currents and abrupt mixing. This may result in the sudden upwelling of deep anaerobic waters, and lead to massive fish kills (Ochumba, 1990; Kitaka, 1972). This phenomenon is commonly reported in Lake Victoria when Nile perch are found dead on the surface after heavy, windy rains (Ochumba, 1987, 1990; Mkumbo, 2002). Heavy rains can also cause high turbidity, which is known to limit the shoreward distribution of some fish, avoiding the erosion of gill filaments by accumulated sediments (Laevastu and Hayes, 1981). This is also the period when bait fishing is less practised, restricted to limited areas as the bait dies once hooked, before a predator is attracted. However this is the period when gillnets have high catches as the fish are restless and migrate, thus increasing their chance of encountering a net. The bait fishery represents a latent threat to the stocks as the estimated weight of the endemic species used represents 25% of that of the total haplochromis catch of 2000 (17,161 t), 235% of that of Clarias (1,101.5 t), 593% of Synodontis (81.4 t), 44% of Labeo (15.2 t) and 20% of Mormyrus (99.6 t) catches of the previous year (Mkumbo 2002). These very high proportions of Clarias and Synodontis in relation to the recorded catch illustrate the potential negative impact of this fishery on the recovery of these native stocks and the high demand for these species in the bait fishery. The threat is severe for Clarias gariepinus (Lm50 = 40 cm TL), as the majority of the juveniles observed belong to this species. C. alluaudi and C. wermeri originating from swamps, dams or rivers outside the lake basin and even outside the country, mature at much smaller sizes (Lm50 = 14 cm TL and 21 cm TL respectively). The other bait species Synodontis, Mormyrus and Labeo mature first at respectively 9–10 cm FL, 34 cm TL and 15–19 cm SL, and the majority of the fish observed are within these ranges. Nevertheless, the techniques of catching them by blocking river mouths with traps or small meshed nets and in rocky areas during the rainy season interferes with spawning activities as they are caught before releasing the eggs. This has major implications for the replenishment and sustainability of the stocks.
The import of species from outside the lake basin poses the additional threat of introducing exotics into Lake Victoria. Although this is indeed against government regulations, traders either transport them at night or bribe local authorities to get through. With the increasing importance of the baited hook fishery it is almost impossible to stop the supply of baitfish. Therefore, management measures must be taken ensuring both the supply of the baitfish and the sustainability of the stocks, taking into consideration the risks of exotics finding their way into the main lake.
Sizes of fish caught in the bait fishery in relation to existing regulations
There are marked differences in the sizes of fish caught with Clarias bait and with the other species. The mode of the distribution (60–61 cm TL) of the Nile perch catches from Clarias bait is within the slot size range (50–85 cm TL), but 20% of the catch is above the upper size limit. The latter can affect the sustainability of the Nile perch fishery. As little is known of its stock—recruitment relationship, this needs to be studied before a bait fishery using Clarias is encouraged.
When haplochromines and dagaa are used as baitfish, almost 61% of the Nile perch catches are below the lower limit. Baitfish themselves are caught in beach seines and gillnets of less than 3 inches, which are all illegal gears. The Fisheries Department could perhaps demarcate areas or bays to allow these gears to operate specifically for bait. This would require considerable additional monitoring and surveillance, as the catching and trading of juvenile fish are prohibited.
Recommendations for conservation and management
The most used bait is Clarias spp., collected from almost all rivers, tributaries, swamps and ponds in the Lake Victoria watershed. They, together with the other bait species: haplochromines, Synodontis, Mormyrus and Labeo are either harvested immature or their harvesting technique is illegal. There is a serious threat of introducing exotic species into Lake Victoria from outside the lake basin.
Any viable management measure to regulate the hook and line fishery should take into account its socio-economic implications for the 25,092 fishermen fully engaged in the bait fishery. An immediate complete ban of hook fishing, aiming at the conservation of the threatened species will be a catastrophe. It is therefore recommended:
– to start an immediate detailed investigation of the selectivity of hooks, bait type and size so that clear advice can be formulated to manage this fishery;
– to conduct a study of the stock—recruitment relationship to monitor the impact of the significant percentage of Nile perch above the slot size, so as to enable early warnings of any threat to recruitment overfishing;
– to ensure a functioning system on Monitoring, Control and Surveillance (MCS) in Lake Victoria so that illegal fishing of juveniles, including baitfish are controlled;
– to ensure a sustainable supply of Clarias baitfish, of which 62% of the Nile perch catch falls within the slot size, by disseminating the technology of culture/farming and artificial propagation of the Clarias species endemic to Lake Victoria.
The authors wish to express their gratitude to the Lake Victoria Environment Management Project (LVEMP) for financial support. The Director General of Tanzania Fisheries Research Institute is acknowledged for permission granted to the scientists to conduct the study. The efforts undertaken by the Project Coordinator, Fisheries Research Component to solicit funds are much accredited. The cooperation extended by all stakeholders especially long-line fishermen, BMUs and leaders at the beaches visited is much appreciated.