In their adaptation to the rapidly shifting environment of the African Great Lakes, small pelagic fish appear to be the most successful fish species. This paper examines the fecundity, size at maturation and the growth and population parameters of Rastrineobola argentea in Lake Victoria.
R. argentea may have a low absolute fecundity of a few thousand eggs, but its relative fecundity by weight is enormous: 70 times higher than for L. niloticus and almost 4,000 times that of the tilapia species O. niloticus and O. esculentus. Its food consists mainly of zooplankton, which is superabundant in the environment and for which there is not much competition.
In response to increased predation and fishing, the species has reduced its size at maturity as well as its maximum size. Population parameters show a typical r-selected life strategy.
The reproductive potential and diet alone do not seem to have enabled R. argentea to thrive as larval densities of R. argentea can be significantly related to a combination of abiotic factors like water conductivity, temperature, Secchi depth and dissolved oxygen concentration.
The decline in endemic fish species of Lake Victoria has among other things been attributed to papyrus encroachment and habitat degradation (Balirwa and Bugenyi, 1980; Ochumba, 1984), as well as pollution (Ochumba, 1984; Ochumba and Kibaara, 1989). After the disappearance of the haplochromine cichlids from the Winam Gulf of Lake Victoria, Nile perch (Lates niloticus Linn.) shifted its diet to freshwater shrimp (Caridina nilotica Linn.), L. niloticus juveniles and the small pelagic cyprinid (Rastrineobola argentea Pellegrin), the latter being the third most important prey contributing some 10–20% to the diet (Ogari, 1984, 1985; Ogari and Dadzie, 1988). A number of studies are available on the impact of the introduced L. niloticus and Oreochromis niloticus (Linn.) in Lakes Victoria and Kyoga (Ogari, 1984, 1985; Ogari and Dadzie, 1988; Asila and Ogari, 1988; Ogutu-Ohwayo, 1990a, 1990b; Ligtvoet and Witte, 1991; Wanink et al., 1998).
It is now clear that small pelagic fish species are more successful in the fast changing environment of the African Great Lakes region than other groups of fish. These include Limnothrissa miodon and Stolothrissa tanganicae in Lake Tanganyika, of which the former was successfully introduced into Lakes Kivu (Roest, 1999), Kariba and Cabora Bassa (Gliwicz, 1984). Previous studies include the spawning season and migration activity of larvae of Limnothrissa miodon (Mwenyimali, 1993) and its exploitation (Mughanda and Mutamba, 1993) in Lake Kivu. Studies on the same species in Lake Kariba include fitting the Von Bertalanffy growth model to length-at-age data of larval fish (Mtsambiwa, 1993), life history (Chifamba, 1993), factors explaining the small body size (Marshall, 1993) and catch trends (Lupikisha, 1993). Limnothrissa miodon and Stolothrissa tanganicae have been well studied in Lake Tanganyika (Roest, 1977, 1993; Bayona, 1993; Katonda, 1993; Bazolana, 1993; Phiri, 1993). In Zambia there is a rich pelagic fishery in Lakes Mweru-Luapula, Mweru-wa-Ntipa and the Bangweulu swamps targeting Poecilothrissa, Microthrissa, Barbus and Engraulicypris, but their biology is still very poorly documented. Kapasa and van Zwieten (1993) provided a preliminary report on the biology of Poecilothrissa moeruensis while Thompson et al. (1993) described the growth of usipa (Engraulicypris sardella) in Lake Malawi.
The small pelagic fish species of the African Great Lakes have also increasingly become commercially and ecologically important (Reynolds, 1993). In Lake Victoria, R. argentea was the second most important commercial species of Lake Victoria fishery up to 1992, which was then dominated by high catches of the introduced Nile perch (L. niloticus), while the Nile tilapia (O. niloticus) is the third most important commercial fish species. These three species currently form the basis of the commercial artisanal fishery. According to Mannini (1993) and the detailed studies of Wanink et al. (1998), R. argentea plays a crucial role within the ecosystem, serving as a trophic intermediary between zooplankton and the apex predator(s).
Ample scientific information exists in the literature on the biology and ecology of R. argentea. Preliminary observations by Okedi (1973) on its breeding ecology and fecundity in Lake Victoria are considered pioneering work on the biology of the species. Recent studies on R. argentea include those of Wanink (1989), Wanink et al. (1998), Wandera (1992, 1993a, b), Manyala (1991), Chitamwebwa (1992; 1995a), Katunzi (1992) and Manyala et al. (1992; 1995a, b) on the general biology and ecology of the species in Lake Victoria. Mannini (1992; IFIP, 1992) attempted a comparative review of R. argentea in Lake Victoria and other small pelagic fish species from the African Great Lakes.
In consideration of the fast development of this promising fishery, the sixth session of FAO/CIFA Sub-committee for the Development and Management of the Fisheries of Lake Victoria (CIFA, 1992) assigned the highest priority to research work on population dynamics of R. argentea. Gaps in knowledge were identified by Manyala (1994) and Wandera (1993b) on the the biology and ecology of R. argentea in Lake Victoria. Wanink (1998) summarizes and discusses all relevant information on the central role of this species in the disrupted ecosystem of Lake Victoria.
Materials and methods
The main sampling stations for the Kenyan side of Lake Victoria are shown in Figure 1. These are the permanent sampling stations established during the Kenya–Belgium Project on Freshwater Ecology in Lake Victoria (1993–1997). Here, biological samples were taken from Rastrineobola catches in order to estimate growth parameters, the size at maturation as well as the fecundity of the species. The fecundity data were compared to published information on the fecundity of several other fish species in Lake Victoria. For each species, the number of eggs per unit weight was determined and plotted on a log scale.
The growth performance of Oreochromis niloticus, Lates niloticus and Rastrineobola argentea was compared using a L∞ –K plan (Mannini, 1992). The data was plotted to determine their grouping according to the hypothesized life history strategy.
Size at massive maturity was also examined and the data superimposed on the size frequency distribution of commercial and experimental catches. This data was used to compare the 50% selection length with the length at massive maturity (Lc50%).
Results and discussion
Fecundity estimates (Figure 2) show that the absolute fecundity per individual fish is higher in Lates niloticus, Clarias gariepinus, Protopterus aethiopicus and Labeo victorianus than in R. argentea. However, when fecundity was compared on the basis of weight, then R. argentea had the highest fecundity per unit weight, followed by Lates niloticus, Clarias gariepinus and some Haplochromis spp. (Figure 3). From these results we can address the following issues:
i) Can the survival of R. argentea in Lake Victoria be explained through its reproductive and ecological strategies?
ii)How does the relative reproductive potential of R. argentea (based on weight) compare with that of Nile perch?
Graham (1929) reported on the planktonic nature of the eggs of R. argentea in Lake Victoria. Okedi (1973) was the first to determine its fecundity (see Table 1). Wanink (1989), Manyala et al. (1992) and Wanink (1998) report on the fecundity. The latter author noted that the absolute fecundity has halved since Okedi's study and ascribes this to the dwarfing of this species in Lake Victoria.
The biomass of R. argentea in Lake Victoria was estimated at about 5,276.5 tonnes, or about 18.19% of the total fish biomass in the lake. About 30% of the biomass of R. argentea (1,462 t) should be sexually mature. At a sex ratio of 1:1, egg production can be estimated at about 1,260 tonnes. At 12.5 million eggs kg−1, we expect an annual production of 6.6 × 1013 eggs from R. argentea. This estimate could be used, together with mortality/survival rates, to estimate total annual recruitment. Unfortunately, an attempted egg and larval survey on the Kenyan side of Lake Victoria yielded very few larvae and not a single egg (Manyala, 1995b). Knowledge of the distribution of eggs in the water column then and now is still too limited to allow a meaningful design of such a survey.
Size at massive maturation
Okedi (1973) reported a size at maturity of R. argentea at 6.3 cm TL (total length) for males and 5.4 cm TL for females. For this study, he analyzed 604 specimens from Winam Gulf, Mwanza, Bukoba and Musoma and found a sex ratio of 53:34 (females/males). The size at maturity of R. argentea in Pilkington Bay has been determined to be 40–41 mm SL (standard length) for males and 43–44 mm SL for females (Wandera, 1993). For the Kenya sector of Lake Victoria, the size at massive maturity has been estimated at 34 mm SL for males and 36 mm SL for females (Manyala, Moi University, Kenya, personal observation). Wanink (1998) found values of 33 mm for females and 46 mm for males in 1988 in the Mwanza Gulf. These lengths correspond approximately to an age of one year (Manyala, 1995b) (Table 2). Wanink (1998) remarks that the increased predation pressure on R. argentea has not led to an increased fecundity. Instead, dagaa has tended to mature at a much smaller body size. Figure 4 shows the current size at maturity compared to previous values against the background of population structure.
Due to predation and fishing, the increased mortality of R. argentea has indeed led to a reduction in its size at massive maturity in Lake Victoria.
The best life strategy for this species would seem to be the r–selected behaviour. It is necessary to look at its growth characteristics in order to understand whether this strategy has indeed been adopted.
Population parameters of R. argentea for 1988 were found to be L∞ = 64.5 mm SL, K = 0.92 yr−1, M = 2.37 yr−1, F = 1.22 yr−1 and Z = 3.59 yr−1 (Wandera, 1992). Manyala (1991) found the following values: L∞ = 67.8 mm TL, K = 0.58 yr−1, M = 0.88 yr−1, F = 1.98 yr−1 and Z = 2.86 yr−1 for 1989/1990 and L∞ = 63.4 mm SL, K = 0.94 yr−1 and Z = 3.23 yr−1 (Manyala, 1993) (Table 3).
Wanink examined 1030 specimens collected in 1983, and after correction for gear selectivity (Wandera and Wanink, 1995), used the ELEFAN routine to determine the growth parameters K = 1.42 yr−1and L∞ = 61 mm SL. Similarly, for 1988 (n = 4522) he obtained the values K = 3.00 yr−1and L∞ = 53 mm SL.
L∞ is the asymptotic length, SL is standard length, K is the Von Bertalanffy growth coefficient, M is the natural mortality, F the fishing mortality and Z is the total mortality coefficient.
The characteristics of r–selection include inhabiting upwelling or pelagic zones (which is observed in many clupeids), a high Von Bertalanffy growth coefficient (K), being short lived (2–3 years), having a high natural mortality rate independent of population size (rarely reaching the maximum carrying capacity), high rates of egg production at low trophic levels and high P/B ratio with small body size and rapid turnover rate.
Both intraspecific and interspecific interactions are generally lax in such species. Figure 5 shows the growth performance of R. argentea compared to that of O. niloticus and L. niloticus.
Abundance and environmental conditions
The chemical characteristics of Lake Victoria have changed considerably over the last fifty years or so. In 1950, the conductivity varied between 95–98 μ S cm−1, while the pH ranged from 8.2 to 9.0. The conductivity in the lake subsequently increased to its present level of 132–140 μ S cm−1 and the pH now ranges from 5 to 8, indicating an increased ionic concentration and acidification.
There have been shifts in diatom abundance, linked closely with changes in pH but the phytoplankton of Lake Victoria is presently dominated by the blue–green algae which form very widespread blooms in the lake during the rainy season. These changes result from increased eutrophication of the lake, as well as from pollution from agricultural, municipal and industrial effluents.
In the absence of quantitative data on stock–recruitment relationships, the data of Manyala (1993, 1995b) allow carrying out a multiple regression between some of these environmental parameters and the measured larval densities of R. argentea. Larval density (D) was found to be significantly related to electrical conductivity, ambient water temperature, Secchi depth, and dissolved oxygen (but not to water depth and pH), in the following relationship (p < 0.05, r2 = 0.746):
D = −42.0 – 0.117 x Conductivity + 2.0 x Temperature – 7.82 × Secchi + 2.65 × Dissolved oxygen
The absence of water depth as an influential abiotic factor is expected since R. argentea larvae are caught only at the surface. Yet most of the larvae were found in the sheltered bays and near river mouths, all of which are shallow areas. Figure 6 shows the distribution of the fish larvae at various sampling stations in the Kenyan part of Lake Victoria.
There is evidence that predation by the Nile perch alone is not the only cause of loss in fish diversity in Lake Victoria. Inappropriate fishing gears and methods and environmental degradation also contributed to the loss.