The major commercial fish species of Lake Victoria at the present time are Lates niloticus, Oreochromis niloticus and Rastrineobola argentea. The contribution of Caridina nilotica in their diet was studied in the Tanzanian waters of Lake Victoria between March, 1999 and January, 2002. Stomach samples were collected during routine bottom trawl surveys in Tanzania. The results show that haplochromines dominate the diet of Nile perch, followed by C. nilotica, R. argentea, juvenile Nile perch, fish remains and other prey. Caridina nilotica predominance in diets decreased as size of the Nile perch increased, as this species switched to haplochromines. Larger perch also feed on their own juveniles. The importance of C. nilotica in the diet of L. niloticus was relatively greater in deeper water than that in shallower stations.
The diet of O. niloticus was predominantly algae followed by C. nilotica, dagaa, Chaoborus, Odonata and others. O. niloticus has shifted its diet from strictly herbivory to a more omnivorous diet, feeding opportunistically on the most available food material.
The overall diet of R. argentea was predominantly copepods, followed by Chaoborus, Cladocera, C. nilotica, Chironomids and insects. In the present food web, C. nilotica is an important food source for the fish stocks of Lake Victoria. The sustainability of the fisheries of Lake Victoria depends among other things on the abundance and availability of C. nilotica.
Currently, the main commercial fish species of Lake Victoria are Lates niloticus, Oreochromis niloticus, Rastrineobola argentea and the haplochromine cichlids (Budeba, 2003). The diet of L. niloticus has been the subject of many detailed studies (Ligtvoet and Mkumbo, 1990; Mkumbo and Ligtvoet, 1992; Ogutu Ohwayo, 1985, 1990a, 1990b; Hughes, 1986; Ogari and Dadzie, 1988; Owili, 1999). Their results indicate that its food preference has changed from haplochromine cichlids to C. nilotica, R. argentea, juvenile L. niloticus and insects. In the research undertaken so far, the importance of C. nilotica in the diet of the main fish species has received little attention (Budeba, 2003).
The objective of this paper is to assess the spatial and temporal importance of C. nilotica in the diet of various size classes of the major commercial fish species in the Tanzanian waters of Lake Victoria.
Materials and methods
The Tanzanian part of Lake Victoria is usually divided into three sampling zones (A, B, C) (Figure 1). Zone A stretches from Kome and Buhiru islands in the southwest, and northeastwards to the south west of Ukerewe island, including the Speke and Mwanza Gulfs. Zone B starts from the Tanzania-Kenya border southwards to the north west of Ukerewe Island. Zone C covers the Kagera waters from Rubafu Bay on the Tanzania-Uganda border southwards to Kome Island, including the Emin Pasha Gulf. These three zones were sampled over 14-day periods, on a quarterly basis.
The diet study was conducted between March, 1999 and January, 2002. Fish stomachs were collected during routine quarterly bottom trawl surveys in each of the three sampling zones. Trawling was done during the daytime for a period of fourteen days per area per month, using RV Lake Victoria Explorer, (length 17 m; 250 H.P.). The catch was sorted into commercial categories. Depending on the size of the catch, from one or two hauls, specimens of Nile perch, Nile tilapia, dagaa and haplochromines were dissected on a daily basis for sex/maturity and dietary analysis. Entire guts were removed and individual stomachs examined separately. For the stomach analysis, the points method (Hynes, 1950; Hyslop, 1980) was used to determine the contribution of each prey item to the diet (Figure 2). The food items were identified to the generic level and awarded points proportional to the total number of points awarded to the stomach. Stomach fullness was noted before opening the gut. Each stomach was awarded an index of fullness ranging from 0 (empty) to 1 (full). The contribution of each prey item was estimated as a fraction of the fullness in decimal points. To determine changes in diet with size, the points awarded to each prey type were summed within each 10 cm length category. Monthly geographical variations in the diet were similarly assessed.
R. argentea has no distinct stomach. Its Z-shaped intestine forms three equal loops in the body cavity (Wanink, 1998). Of these, only the anterior loop was removed and its content flushed into a Petri dish which was then examined under a low-power stereo microscope.
In order to assess the temporal and spatial variations in the diets of all species, data from each zone were grouped by month. Within each zone, the data were also grouped into depth ranges. For practical purposes, L. niloticus were grouped into 10 cm size classes.
A total of 4263 stomachs of L. niloticus were examined. This species feeds on a wide variety of organisms, 18 of which were identified. They can be divided into haplochromines (47.8%), C. nilotica (38.6%), dagaa (7.5%), juvenile Nile perch (2.3%), fish remains (2.0%) and other prey items (2.5%). C. nilotica showed consistent seasonal variation in all three zones (Figure 2).
In zone A, the overall diet of Nile perch was predominantly haplochromines (60.2%), followed by C. nilotica (29.2%), dagaa (4.0%), juvenile Nile perch (1.6%), fish remains (1.8%) and unidentified other items (3.4%). In zone B, C. nilotica (53.6%) dominated, followed by haplochromines (35.6%), dagaa (8.5%), juvenile Nile perch (3.7%), fish remains (0.8%) and other items (1.8%). In zone C, C. nilotica dominated as well (42.6%), followed by haplochromines (27.5%), dagaa (13.8%), juvenile perch (2.5%), fish remains (3.7%), and other items (1.7%). C. nilotica was more important during the dry season (June – September); when they were absent, haplochromines formed the alternative food for the Nile perch.
Changes in diet with size of Nile perch differed between zones (Figure 3). Diets of fish below 20 cm TL were always dominated by C. nilotica. Juveniles of its own species contributed only 14% in stomachs above 100 cm TL in zone A, against 61.5% in zone B and 80% in zone C.
There was marked variation in the contribution of C. nilotica to the diet of Lates niloticus with depth (Figure 4). At depths below 10 m in zone A, haplochromines predominated, followed respectively by C. nilotica, dagaa and juvenile Nile perch.
A total of 616 stomachs of O. niloticus were examined. Its diet was predominantly algae (65.0%), C. nilotica (17.0%), dagaa (8.5%), Chaoborus (5.3%), Odonata (1.5%) and others (2.8%). The latter component was made up of bivalves, plant material, shells, oligochaeta, haplochromines, Barbus, and zooplankton. The diet showed temporal and spatial variations over the study period (Figure 5). There was a slight increase in the contribution of C. nilotica with increased predator length (Figure 6).
In the 1–10 m depth range, the diet of O. niloticus consisted of algae (65.3%), C. nilotica (15.7%), dagaa (9.1%) and Chaoborus (5.6%). At 11–20 m depths, the diet was predominantly algae (71.9%) and C. nilotica (20.8%). Three items dominated the food at 21-30 m depth: algae (38.9%), C. nilotica (34.9%) and dagaa (16.7%). The contribution of C. nilotica to the diet increased from shallow to deep-water stations, as algae decreased (Figure 7).
A total of 638 stomachs of R. argentea were examined. The overall diet was composed of copepods (36.1%), Chaoborus (30.8%), Cladocera (17.0%), C. nilotica (11.4%), Chironomid larvae (4.6%) and insects (2.6%). The contribution of C. nilotica fluctuated per month, but generally in low proportions (Figure 8). C. nilotica occurrence was relatively higher in stomachs of larger fishes in each of the three sampling areas (Figure 9). Diet composition varied little with water depth (Figure 10). Copepoda, Chaoborus, Cladocera and C. nilotica were the main items caught at all depths. The contribution of C. nilotica was relatively higher in the deep and offshore water stations than in the shallow and inshore water stations (Figure 10).
Diet and feeding pattern of Nile perch
Nile perch was introduced in Lake Victoria in 1954 (Fryer, 1960). At first, it was predominantly feeding on haplochromines, which were the most abundant species (Gee, 1964, Hamblyn, 1966; Okedi, 1971). Kudhongania and Cordone (1974) reported 83% of the ichthyomass to be of haplochromine cichlids. The abundance of Caridina nilotica stocks was low at the time, and it was consumed mainly by Schilbe, Mormyrus, Barbus and Brycinus (Corbet, 1961). After the L. niloticus boom in the Mwanza Gulf, which started in 1983, the haplochromine cichlids virtually disappeared (Witte et al., 1992) and by 1986 large quantities of C. nilotica were observed.
C. nilotica was reported to be an important prey for L. niloticus in Lakes Albert and Turkana where L. niloticus is endemic (Hamblyn, 1966; Gee, 1969), and contributes significantly to the diet of L. niloticus in Lake Chad (Hopson, 1972).
Lates niloticus, being an opportunistic predator, quickly adapted its feeding habits, starting with zooplankton and then switching to insects, C. nilotica and finally to fish (Ogari and Dadzie, 1988; Ligtvoet and Mkumbo, 1990; Ogutu-Ohwayo, 1990a, 1990b; Mkumbo and Ligtvoet, 1992; Hughes, 1992; Mkumbo et al. 1996; Lowe-McConnell, 1997; Agullo, 1999; Owili, 1999). Owili (1999) reported that zooplankton is important in the diet of L. niloticus of 1–3 cm TL and that C. nilotica becomes dominant from 4 cm TL onwards. Ogari and Dadzie (1988) reported that zooplankton, and especially Cladocerans, Chironomids, Copepods and Povilla adusta comprised the food of L. niloticus below 5 cm TL. As very small fish were not present in the trawl samples, no zooplankton was found in L. niloticus stomachs. The dominance of juvenile Nile perch in the catches (Mkumbo and Ezekiel, 1999; Mkumbo et al., 2005; Mkumbo, 2002; Mkumbo et al., 2002) reflects their availability as prey for the larger Nile perch.
Over the past three decades the abundance of C. nilotica in Lake Victoria has increased tremendously (Budeba, 2003). The present study shows that C. nilotica has become the dominant prey for L. niloticus especially for fish up to 60 cm TL (Figure 2).
The present dominance of haplochromine cichlids in the stomachs of L. niloticus above 60 cm TL reflects the recovery of the haplochromine cichlids and the return to their former role when they were the most abundant species in the lake. They formed an important by-catch in the dagaa catches at Igombe beach (Budeba, 2003). From 1991 on, Seehausen et al. (1997) reported a recovery of the zooplanktivores Yssichromis pyrrhocephalus and Y. laparogramma in the Mwanza Gulf. In Speke Gulf, Witte et al. (2000) reported that Y. laparogramma comprised 64% of the weight of the catch of the lift net fishery targeting dagaa, while 33% consisted of R. argentea and 3% of juvenile Nile perch.
Fish may switch to other prey in response to changes in their relative abundance (Murdoch and Oaten, 1975). In Zone A, Nile perch feed mostly on haplochromines, which appear to be more abundant (Witte et al., 2000; Budeba, 2003), while in zones B and C where haplochromine abundance appears to be low, they feed predominantly on C. nilotica. The size range of L. niloticus feeding mostly on C. nilotica appears to have increased from 1–50 cm TL in the past to 60 cm TL at present. This change in food selection with increased predator size is in agreement with the findings of Ogari and Dadzie (1988) in the Kenyan waters of Lake Victoria and Hamblyn (1966) in Lake Albert, who found that invertebrates dominated in the stomachs of L. niloticus below 60 cm TL, and that fish above 60 cm TL were piscivorous. The majority of L. niloticus caught in the Tanzanian waters of Lake Victoria were 5–45 cm TL in size (Mkumbo and Ezekiel, 1999; Mkumbo et al., 2005; Mkumbo et al., 2002). The whole Nile perch stock therefore predominantly feeds on C. nilotica.
Diet and feeding pattern of Nile tilapia
Oreochromis niloticus, which used to be a strictly herbivorous fish feeding on a variety of algae and higher plant materials, has diversified its diet to C. nilotica, zooplankton, R. argentea and Chaoborus larvae. This confirms the trophic plasticity of the species (Lowe-McConnell, 1958). The switch from a strictly herbivorous to a more generalist diet reflects the ecological changes that have occurred in the lake. The anaerobic environment in Lake Victoria favoured the flourishing of C. nilotica and other micro-invertebrates (Hecky, 1993; Mwebaza-Ndawula, 1998; Mwebaza-Ndawula et al., 1999; Budeba, 2003). In the present study, blue-green algae, green algae and diatoms formed the dominant food of O. niloticus. The relative contribution of C. nilotica was, however, also important. O. niloticus is a shallow water fish, which prefers to live in aquatic macrophytes like papyrus (Cyperus papyrus), lake bottom grass-beds and even underneath the water hyacinth Eichhornia crassipes mat. The same areas are used by C. nilotica as feeding grounds as well as nursery habitats. It is probable that this habitat overlap is responsible for the feeding relationship between these organisms. In Ugandan waters, Balirwa (1998) reported detritus and insects, especially chironomid larvae to be the most important food items of O. niloticus. Njiru (1999) found insects as the dominant food of O. niloticus followed by algae, fish, higher plant material, zooplankton, C. nilotica, bivalves and oligochaeta in the Kenyan part of the lake. On the other hand, Owili (1999) reported phytoplankton and zooplankton to be the most important food items in the Kenyan waters but this was for fish below 10 cm in length.
The present study indicated seasonal variations in the contribution of C. nilotica to the diet of O. niloticus (Figure 5). These may reflect variations in the abundance of both C. nilotica (Budeba, 2003) and its zooplankton food in the lake (Wanink, 1989, 1998; Mwebaza-Ndawula, 1998). In the dry season, Caridina nilotica tends to migrate away from the shallow inshore waters to the deep offshore waters to breed and comes back during rainy season (Budeba, 2003). Sample sizes in some months are rather small.
The introduction of O. niloticus in 1961 (then restricted to very shallow littoral waters) caused a strong competition for food with the phytoplanktivorous haplochromines (Moreau et al., 1993). A wider food spectrum is one of the factors which enabled O. niloticus to be successful in Lake Victoria (Fryer and Iles, 1972).
Diet of Rastrineobola argentea
Although Rastrineobola argentea is one of the most important prey items of L. niloticus, it manages to co-exist with it in Lake Victoria, even in the absence of haplochromines cichlids. R. argentea has replaced the zooplanktivorous haplochromines cichlids in Lake Victoria (Goldschmidt et al., 1993).
In the daytime, R. argentea feed on zooplankton, which concentrates near the bottom of the lake. They migrate to the surface at night. The zooplankton community of Lake Victoria is dominated by cyclopoid copepods (Branstrator et al., 1996; Mwebaze-Ndawula, 1998; Wanink, 1998; Omondi, 1999; Owili, 1999; Budeba, TAFIRI, Dar-es-Salaam personal observations). During the nights, with many emerging Chaoborus and chironomids at the surface, R. argentea catches these insects (Wanink et al., 1989). In the dry season, zooplankton and C. nilotica tend to migrate for breeding from the shallow inshore waters to deep offshore waters. In case of oxygen depletion in the lower water column, adult R. argentea concentrate just above the deoxygenated layer during the daytime (Wanink, 1998). The same area is occupied by C. nilotica during thermal stratification (Budeba, 2003).
The present study identified six food items of R. argentea, i.e. C. nilotica, Cladocera, Copepoda, Chaoborus, insects and chironomids. These results are in agreement with Wanink (1988, 1989, 1998) and Hoogenboezem (1985) in the Mwanza Gulf of Lake Victoria, but differ somewhat from Kenyan and Ugandan parts of Lake Victoria. Owili (1999) reported the dominance of Cladocera and absence of C. nilotica in the diet of R. argentea in Kenyan waters of Lake Victoria. Mwebaza-Ndawula (1998) reported the absence of C. nilotica in the stomachs of R. argentea in the northern part of Lake Victoria. The contribution of C. nilotica to the diet of R. argentea shows seasonal variation without defined trend. This can be attributed both to the seasonal fluctuations in R. argentea abundance as reflected in the commercial catches (Budeba, 2003), and to seasonal variations in the abundance and availability of its key food C. nilotica and other macro-invertebrates.
The contribution of C. nilotica to the diet of R. argentea was found to increase with fish size. Juvenile R. argentea hide in shallow waters, especially in aquatic weeds such as water hyacinth (E. crassipes), papyrus (C. papyrus), and Ceratophyllum; feeding on insects, which are more abundant there. As the fish increase in size, C. nilotica and Chaoborus become important as food. Later, R. argentea becomes truly pelagic and moves offshore. Even here, it tends to hide near the bottom in the daytime and comes to the surface at night. C. nilotica shows exactly these same diel migrations.
In order to assess the possible impact of prey switching on dagaa, Wanink (1998) established the energy content of different prey types in function of their size. He observed that stomachs filled with prawns contain 15 times more energy than those with zooplankton. The energy content of C. nilotica is 20.34 J mg− 1 for unberried and 28.01 J mg− 1 for berried ones (Hart, 1981; Wanink, 1998; Budeba, 2003).
In the present food web, C. nilotica is an important food source for the fish stocks of Lake Victoria. Therefore the fisheries of Lake Victoria, and their sustainability, depends among other things, on the abundance and availability of C. nilotica in the lake.
This paper formed part of Yohana Budeba's PhD research on the role of Caridina nilotica in the Lake Victoria fisheries, Tanzanian waters. We express our sincere thanks to the European Union for funding this study through the Lake Victoria Fisheries Research Project II. We thank the Director General of TAFIRI for institutional support. Thanks also go to the crew members of RV Victoria Explorer for data collection, handling and analysis and to my collegues for moral support during this study.