This study describes the inshore fishery on the littoral cichlids in Lake Tanganyika, to evaluate the potential impact of fishing gear on the cichlid fishes and their diversity. The fishing activity at the northwestern end of Lake Tanganyika in Uvira was examined from March to July 2010. In total, the catch examined was composed of 3,555 cichlid fish: 47 species and 30 genera identified representing 10 of the 16 tribes of Lake Tanganyika cichlids. Fishermen in Uvira, exploiting littoral cichlids, used three main fishing gears including beach seine, frame gillnet and mosquito net, varying according to the substrates of the littoral habitat at different study sites. They generally captured immature fishes (63%) and small-sized fish due to their small mesh size and unsuitable fishing techniques. The beach seine caught more species on the sandy substrate while the frame gillnets caught more individuals. The cichlid species Aulonocranus dewindti was the most abundant in the catches.
Lake Tanganyika, one of the greatest lakes of the Albertine Rift and the oldest of the African Great Lakes, is well known for its exceptional biodiversity (Poll, 1956; Brichard, 1989; Coulter, 1991; Patterson and Makin, 1998; Fermon, 2007). This lake has fish fauna especially rich and diversified in benthic and littoral areas (Nakai et al., 1994) with around 325 (Steenberge et al., 2011) species of fish including a very high endemism, particularly in the Cichlidae family (98%), but also in non-Cichlidae (46%) (Brichard, 1978; Kawabata and Mihigo, 1982; Hori et al., 1983; Poll, 1986; Kawanabe et al., 1992). Important taxonomic studies on Lake Tanganyika cichlids were carried out by Boulenger (1915) and Poll (1956, 1986); later the discovery and description of new varieties and species were mostly executed by ornamental fish hobbyists (Snoeks, 2000). Many authors emphasize that the littoral environment is very complex: rocky, sandy and mixed rocky/sandy shores alternate, offering a variety of habitats (Patterson and Makin, 1998). This lake contains one of the largest inland water fisheries in Africa, especially in its pelagic zone (Moreau et al., 1991; Breuil, 1995; Mölsä et al., 1999, 2002). Although many studies have focused on pelagic fisheries, fishing activity near the coast was given little scientific attention. Over the years, the number of fishermen has increased and fishermen have increased the frequency of daily trips in order to compensate reduced unit catch rates (Chokola and Cikwanine, 2004). However, studies on Lake Tanganyika cichlids, especially along Congolese coasts, remain fragmented and few of them deal with cichlid fishing activities despite the current increase of riparian human population in the region. To ameliorate filling this gap of knowledge, the present study examines the status of cichlid fish diversity and fishing gears used for commercial capture in the northwestern part of Lake Tanganyika.
Study area and methodology
The northwestern part of Lake Tanganyika off Uvira is located in the Bujumbura sub-basin (between 03° 28′ 00 S and 29° 17′ 00 E). This region has a tropical humid climate characterized by an alternation of wet and dry seasons. Four beach fishing sites were selected near Uvira for sampling due to their accessibility and intensive fisheries activities, namely Kilomoni, Kasenga, Mulongwe and Kalundu beaches (Figure 1). Sampling of commercial cichlid fish catches was conducted during two visits per month for each sampling site from March to July 2010. Fish samples with a selection of species were usually bought from fishermen, and preliminary sorting was carried out at sampling sites because fishermen could not deliver the large fish of great commercial value of their catches. Lengths and widths of fishing nets and their mesh sizes were measured using a measuring tape. The type of fishery and net, the numbers of fishermen involved per trip and per fishing unit were recorded and are discussed in relation to the type of gear used in the littoral habitat.
Regarding littoral habitat, three major substrates are recognized in Lake Tanganyika: rocky substrate, sandy substrate and rocky/sandy (mixture or mixed) substrate (Axelrod et al., 1977; Coulter, 1991; Patterson and Makin, 1998) and fishermen are aware of these different habitats in selecting their target species. Each sample was sorted into species following books and/or available keys (Fryer and Iles, 1972; Axelrod et al., 1977; Brichard, 1978, 1989; Eccles, 1992) in the Laboratory of Biology at the Centre for Hydrobiological Research/CRH-Uvira as final sorting. For each species caught, length frequencies were recorded depending on coastline habitat and the net type used. Measurements and measuring techniques (e.g. SL and TL) of fish were executed according to Barel et al. (1977), Eccles (1992) and Kaningini (1995) for maturity stages. Data analysis (descriptive statistics and ANOVA) was conducted using STATVIEW and SPSS.
Total length (F = 66.7, p < 0.0001) and standard length (F = 66.2, p < 0.0001) of specimens of species captured varied according to the sampling sites. The largest fish sizes of cichlid fishes were observed in Mulongwe site followed by Kilomoni site (Table 1), due to the rocky habitat at Mulongwe.
Beach seines, which were longer and wider (Figures 2a and b) and thus swept a larger surface area than gillnets and mosquito nets, caught significantly greater quantities of fish. However, the number of fishermen per fishing unit was also highest for beach seines varying between four and five fishermen (Figure 2d) although at certain sites these numbers may be higher. For the four sampling stations, mesh sizes (Figure 2c) in gillnets were larger than in the other two net types used for cichlid fishing.
The characteristics of fishing gear used varied with different habitats in which littoral cichlid fish live in northwestern Lake Tanganyika (Figure 3). Fish caught in the rocky habitats were the largest in size (Figures 3a and b), apparently because the usual fishing gear used in this habitat was gillnet which had the largest mesh size (Figure 3c) and was more selective. So, many rocky shore fishes were therefore caught as adults with mature gonads (Figure 3c) of present species in that habitat. In addition, it was observed that preferred fishing net in sandy habitats was the beach seine which also was the longest net used (Figures 3d and e). The beach seine required a high number of fishermen per unit and per trip, resulting in higher labour investment on the sandy habitats (Figure 3f). Number of hauls per daily trip and per fishing unit was highest in predominantly rocky habitat where gillnets were the main fishing gear (Figure 3g).
In all, 47 species of Cichlidae were caught on the rocky, sandy or mixed littoral habitats. However, fish species variety was rich on sandy habitat (44 species) compared to that from other habitats. The cichlids Lepidiolamprologus profundicola, Petrochromis polyodon and Lobochilotes labiatus were absent in sandy areas. Forty species were caught with gillnets (Table 2). The most common species found in all substrates were Aulonocranus dewindti, Trematocara kufferathi, and Xenotilapia nigrolabiata.
Aulonocranus dewindti was most abundant in sandy and mixed areas. The low abundance of this fish species was remarkable in rocky substrate, while Limnotilapia dardennei was more common on rocky substrates and was mostly captured by gillnet. Around 50% of fish species were completely absent from rocky areas in our total catch. This can likely be attributed to a combined effect of differences in habitat preferences of the fish species and gear performance which contributes to species-level composition in our catches. The low proportion of capture on the rocky substrate can also be explained by the fact that fish find shelter among the rocks and avoid being caught in the fishing nets. A net can also less completely scrape a rocky area than it would do on sandy or mixed substrates.
Females were slightly dominant over males. The average sex ratio (males: females) in our sample was 0.9:1. Over half of the cichlids were caught at an immature stage (Figure 4). Percentage of immature fish in the catch was even higher (63.10%). Also, the most abundant species as in the entire catch presented the smallest average lengths compared to all other species caught.
The size at first maturity, varied according to the maximum size of each species (Table 3). The size at first maturity was estimated in only twenty-one species that had mature gonads in our littoral catch and during the study period. Considering the maximum standard length of adult individuals examined, gonads matured at an early stage. The size at first maturity varied from Neolamprologus pleuromaculatus (29 mm) with the smallest size, to Oreochromis niloticus (135 mm) showing the largest maturity size for specimens caught of each species. However, among the fish species examined, Bathybates minor was the largest fish (300 mm) presenting precocious size at first maturity (70.7 mm) with n = 51.
The smallest average lengths were recorded for Aulonocranus dewindti (36 ± 19 mm) and Trematocara kufferathi (37 ± 5 mm), which were captured with average mesh sizes of 4.5 ± 4 mm and 6.9 ± 4.7 mm, respectively (Table 4). The greatest lengths were shown by Petrochromis polyodon and Lepidiolamprologus profundicola, but they were represented by a few individuals only. The average capture mesh size ranged from 3.8 ± 1.0 mm (Xenotilapia flavipinus) up to 20.0 mm (Hemibates stenosoma).
Lake Tanganyika littoral cichlid fishery is considerable, especially along the Congolese northwestern coastline where fishing methods and gears identified are not sustainable. Inshore fisheries are complex in this part of the lake. They are multi-species, and multi-gear, and involve artisanal fishermen as well as those who practice subsistence or traditional fishing. Most fishing areas near the coastline (depth 0–20 m) and adjacent areas of high human influence are already under pressure from a range of gears (Fermon, 2007). However, this pressure would be much better illustrated if we knew the reproduction mode of each species fished in the area; this is not yet the case given the few studies available on reproduction of Lake Tanganyika cichlids. As far as is known in Lake Victoria, fecundity, spawning areas, spawning periods and nursery grounds can be strikingly different among species (Wootton and Potts, 1984; Kaufman and Cohen, 1993; Eccles, 1986) as it is was observed in Lake Tanganyika where biodiversity is higher (Brichard, 1989; Nakai et al., 1994). It is known that many cichlid species are mouth-brooders (Brichard, 1978, 1979, 1989). Each species, such as sand- or rock-dwellers, is therefore well adapted to a particular habitat and can use spawning sites like sand bottoms and sheltered nests (Lowe-McConnell, 1956; Wootton and Potts, 1984; Coulter, 1991; Nakai et al., 1994). As for the harvested cichlids, up-to-date, no reports on the behavior of the fish in the lake have been published, although spawning of several cichlids in the aquarium has been reported by aquarium hobbyists (Brichard, 1979).
The present study inventoried three main types of Lake Tanganyika Cichlidae fishing gear on a total of four coastline sites in the Uvira territory where inshore fisheries represent an important economic activity. This is a traditional fishery with a tendency towards artisanal techniques. There are beach seines, gillnets and variants of mosquito nets. These gears represented quite different fishing methods and, beach seine requires more numerous persons to pull it from shallow water up to the land. It was observed that the beach seine was used throughout the year both in the daytime and at night, but usually during the day especially in littoral and sub-littoral sandy shores. However, its catch was composed of littoral fish (mainly cichlids). Each beach seine was used not only at one beach but this fishery was very dynamic; canoes and fishermen moved from one beach to another along the coastline fishing wherever was possible. Catch varied considerably between sampling sites depending on the number of fishing trips and dragnets by trip, trip duration (often about 5 h), and the net size. Beach seining was most common on the predominantly sandy Kasenga and Kalundu sites.
The frame gillnet can also be flexibly used at different locations along the coastline in Lake Tanganyika. This net was a sheet mounted on two boltropes: an upper equipped with floats and a lower with leads or stones. Net length was sometimes longer than 100 m and depth from 10 m to several meters and the mesh sizes ranged from 4 to 30 mm. Fishing was practiced during sunrise and consisted of a non-motorized canoe, sometimes with two strings of net attached, 1–3 paddles and bamboo sticks for movement along the coastline. The gillnet fishery caught especially various sizes of cichlid fishes, crabs, and fishes belonging to the families Mochocidae and Bagridae and, it was practiced on all substrates.
For the mosquito net (seines), material used is a very dense mosquito tablecloth (mesh size less than 2 mm) of various lengths and a depth of about 1.5 m. Mosquito nets were used as different varieties ranging from a simple net to beach seine. They caught any fish encountered in their path.
The gears were described and analyzed in relation to several factors (e.g. length, width and mesh size of nets) with which every fisherman sought to increase the catching efficiency. Fish caught in a rocky habitat were relatively large; usually because large-meshed gillnets were used in those sites. However, the type of substrate also played a great role in determining the fish communities in littoral areas, and rocky habitat contained more varieties of fish (Matthes, 1960; Fryer and Iles, 1972; Lowe-McConnell, 1975; Brichard, 1978; Hori et al., 1983). This is also valid to rocky communities in Lake Malawi where they are well known (Lévêque et al., 1988, 2008; Eccles, 1986). Also, in Lake Tanganyika, as is known in Lake Victoria, each species has a specific habitat (depth range, wind exposure etc.) and dietary ecology (Witte and van Oijen, 1990; Fermon, 1996).
Cichlidae species composition was broadly consistent with that found by Kawabata and Mihigo (1982) near Uvira and by Ntakimazi (2007) in the Ruzizi Delta. The 47 Cichlidae species in our catch from Uvira shores were distributed in 30 genera and 10 tribes out of 12 tribes listed by Poll (1956, 1986) and revised to 16 by Takahashi (2003; in Steenberge et al., 2011) from this lake. The Ectodini, with 16 species and 8 genera, were abundantly represented, followed by Lamprologini and Limnochromini. These observations agree with those previously reported by Amundala et al. (2002) and Nanzige (UOB, Bukavu - DR Congo, pers. comm.). However, we should note that the abundance of a community in the catch depended on the substrate of littoral habitat, method and fishing technique and the range of food available and also, each community is characteristic of certain habitats (Matthes, 1960; Fryer and Iles, 1972; Lowe-McConnell, 1975; Hori et al., 1983; Poll, 1956, 1986).
Nets inventoried had mesh sizes ranging from 2 (mosquito nets) to 20 mm (gillnets). The latter caught large fish on rocky substrate, comparing favorably with the beach seine and the mosquito net, also because of its high mobility at all sampling sites. On the other hand, it is assumed that beach seine and mosquito net catch anything of any size and destroy nests and habitats of Cichlid species on the sandy substrate along the Uvira coastline and are illegal by law. Such high percentage of immature fish as we found, was also reported by Muterezi (C.R.H.-Uvira, DR Congo, pers. comm.) for a beach seine variant in Uvira. He explained this by the fact that beach seine was actively used near shore zone and it is not very selective – that was also observed in our results. For fishery management, the Lake Tanganyika cichlids constitute a problem because this group comprises over 184 species (De Vos et al., 1994) with a wide ecological diversity and speciation. Accordingly, in our study different species were caught according to area and depth. The longevity of at least some of the cichlid species (Egger et al., 2004) makes them particularly vulnerable to intense fishing. Our cichlid fishing is an outstanding example of a multispecies fishery in which each species has a specific ecology in its natural habitat in the lake. Considering these species as a single unit for fishery research and management is not warranted (Axelrod et al., 1977). However, management of such a fishery at the species level would be unfeasible as noted by Pauly (1980) in Lake Victoria for the haplochromine species.
This study focused on the importance of species diversity and characteristics in Lake Tanganyika cichlid fisheries in the northwestern sector Uvira. In fact, it identified the different types of gear and fishing techniques used, as well as the species richness of littoral stocks targeted based on their natural habitat found in the inshore areas of the lake.
Littoral fishery of cichlids is large-scale and important with a high value which supports a significant number of fishermen. Hence, there is a need for a biodiversity programme for monitoring and management. In the absence of an ecosystem approach involving all stakeholders fishing activities will cause cichlid fishing catches per unit and per effort to decrease drastically, especially in areas heavily exploited such as the littoral zone of Uvira in the northwestern part of Lake Tanganyika.
Our thanks go to the scientific staff of the Faculty of Sciences at the State University of Bukavu (UOB), particularly Professors Muhigwa and Kahindo for their advice and data analysis. We also thank all researchers from the Department of Biology of CRH-Uvira for their assistance in the laboratory.