Two mesh sizes (5 and 10 mm) usually used to harvest Rastrineobola argentea in Lake Victoria were assessed from April to September 2001 to determine their suitability. It was observed that 13.2% of fish caught in 5 mm and 11.2% caught in 10 mm for April, May, June and July were of a size below maturity (44 mm standard length, [SL]). Mean size of fish caught with the 5 mm mesh was 46.8 ± 1.2 mm and that of fish caught with the 10 mm mesh was 46.8 ± 0.5 mm (SL). Catches per unit effort (CPUE) from the two meshes were not significantly different, with Haplochromis laparogramma (50.1%) dominating the catch followed by R. argentea (48.9%). Juveniles of Nile Perch, Brycinus jacksonii, Brycinus sadleri, and Barbus spp. (0.9%) were caught in small numbers. About half of the females matured at 43.5 mm and the males at 46.6 mm SL. Fish caught with the 5 mm mesh had on average 1050 ± 400 eggs and those caught with the 10 mm mesh had 918 ± 281 eggs, but the difference was not significant. However, a significant difference was noted between non-infected fish (1050 ± 400 eggs) and fish infected by the cestode Ligula intestinalis (663 ± 316 eggs). Given the absence of differences in key parameters under investigation between the fish caught with the 5 and 10 mm mesh sizes, the two meshes can be used interchangeably.
Rastrineobola argentea (Pellegrin, 1904) is an indigenous fish species of Lake Victoria, East Africa. It is a resilient species that withstood radical changes in the ecosystem and the modified lake fishery from a multi-species fishery to one comprised only of three commercially important fish species namely, Lates niloticus, R. argentea and Oreochromis niloticus. Rastrineobola argentea (“Dagaa” in Tanzania) is reported to spawn both inshore and offshore and to have a life span of 1.5 to 2 years (Wanink, 1992; Wandera, 2000). In terms of catch tonnage it is the most important species, and in terms of value of the catch it is second after the Nile Perch (Lates niloticus) (Lake Victoria Fisheries Organization [LVFO], 2008a).
This fishery started in the 1960s using beach seines and scoop nets with mesh sizes of 10 mm. It was intensified in the late 1980s after the introduction of new fishing gear; i.e. small seines (harry-up) brought in from Kenya and lift nets using catamaran boats adopted from Lake Tanganyika where the nets were used to fish Limnothrissa miodon and Stolothrissa tanganicae. Fishing with small seine nets can occasionally be done during the day when shoals of R. argentea are spotted near the shores. The importance of this fishery has continually increased after the decline of other fish species and the expansion of the local market and the export markets to neighbouring countries.
By 1995, the main gears used for fishing R. argentea in the Tanzanian waters of Lake Victoria were the beach seine, lift net, small seine net and scoop net (Witte and Van Densen,1995). Beach seine (kokoro), which was initially used to fish haplochromine cichlids and R. argentea in the 1980s, later became very important during the boom of the Nile Perch fishery. Most beach seines, particularly those targeting
R. argentea, were exchanged for either lift nets or small seine nets in the 1990s. They were further reduced after declining catches of R. argentea in the inshore waters and even further reduced during their ban in 1995. The introduction of new fishing gear was necessitated by demand of fish and scarcity of R. argentea in the inshore waters. These new fishing techniques enabled fishers to venture into more off-shore waters which had relatively abundant fish. Associated with these developments were changes in mesh sizes from the initial 10 mm to 5 mm.
This led to a reduction in the size of R. argentea as reported by Manyala (1993), Wanink and Witte (2000) and Wandera et al. (2006) and to a decrease in the size at maturity elsewhere in the lake, tempting the Government to prohibit mesh sizes of less than 10 mm. However, a majority of the fishers continued to fish illegally with the 5 mm mesh. In this article we will evaluate the differences between the 5 mm and the legal 10 mm mesh in exploitation patterns, catch per unit effort and fecundity of R. argentea.
Materials and Methods
Study area and methods
Experimental fishing with 5 mm and 10 mm mesh sizes for R. argentea was carried out from April to September, 2001 at Igombe in Lake Victoria. In order to check the landings of the artisanal fishers, visits were made to other important R. argentea beaches at Kibuyi in the Mara region and Nyamkazi in the Kagera region (Figure 1).
Experimental fishing was done using two small seine nets with mesh sizes of 5 mm and 10 mm and 19.5 meters long. Each of the two boats used four pressure lamps, and an outboard engine fixed on one of the boats was used to tow the second boat without an engine to and from the fishing grounds where local fishers were fishing. Fishing was conducted approximately 5 km away from the shore for ten days in each of the consecutive six months. On arrival at the beach, catches were sorted into species and weighed. A 0.5 kg sample was scooped randomly, packed and preserved in 4% formaldehyde. In the laboratory, samples were measured individually to the nearest mm below (standard length, SL) and weighed to the nearest gram below. The same was done for samples collected from artisanal fishers. Sampled fishes were dissected to examine sex, fecundity and gonadal maturity stages (coded according to Mannini ). Fecundity was defined at maturity stage IV as recommended in Bagenal (1978).
Due to seasonality effects, comparison of mean fish size between mesh sizes by month was made using a paired t-test. The same test was used to compare percentages of immature fish but was performed on ArcSine transformed percentages. Analysis of variance (ANOVA) was used to compare catch rates (CPUE) and numbers of Dagaa eggs caught in 5 mm and 10 mm meshes.
Size structure and abundance of R. argentea
Monthly length frequency distribution patterns for R. argentea caught with the 5 mm and 10 mm meshes were plotted in Figure 2a and 2b. The percentage of fish below size at maturity in the months of April–July (15.6% for 5 mm mesh; 12.8% for 10 mm mesh) was higher than in August–September (10.6% for 5 mm mesh; 9.8% for 10 mm mesh). Catches by the two mesh sizes did not differ significantly in terms of monthly changes in fish size distribution. The monthly determined mean lengths had significant difference in variance (F test: F = 6.06, P(F< = f) one-tail = 0.035, F Critical one-tail = 5.05). Therefore, t-test assuming unequal variance revealed that there was no statistically significant difference between the mean lengths of fish caught by the 5 mm and 10 mm mesh sizes (t-test (one tail): t = −0.0597, df = 7, p = 0.406). Likewise, percentages of fish below size at maturity (44 mm) were not significantly different between the two mesh sizes (t-test (one tail): t = 0.3534, df = 5, p = 0.7382).
Catch per unit effort (CPUE) as an index of fish abundance showed an average of 3.4 kg lamp−1 hr−1 and 3.1 kg lamp−1 hr−1 landed in nets of 5- and 10 mm mesh, respectively. A seasonal decreasing trend in CPUE was observed from 8.0 kg lamp−1 hr−1 in April to 0.1 kg lamp−1 hr−1 in August in both nets, which went back up to 7.6 kg lamp−1 hr−1 in September for the catch with the 10 mm mesh and 2.6 kg lamp−1 hr−1 for the catch with the 5 mm mesh (Figure 3).
Haplochromine cichlids are now re-emerging as an important by-catch in the R. argentea fishing. They suffered a sharp decline from over 80%, prior to the Nile Perch boom to about only 1% of fish biomass in Lake Victoria in the late 1990s (Ogutu-Ohwayo et. al., 1998). Haplochromis laparogramma (mostly caught in offshore waters) and H. pyrrocefalus (fished in the littoral areas) are the two species normally caught in the R. argentea fishery (Witte, et al.,2000). In this study H. laparogramma was abundantly caught. Its catches increased with increasing moonlight when compared to R. argentea. Over the study period, haplochromines contributed 50.1% while R. argentea made up 48.9% of the total catch. The contribution of other fishes like Brycinus jacksonii, Brycinus sadleri, Barbus spp. and juveniles of Nile Perch was very minimal (0.9%).
Size at maturity, fecundity and sex
As expected, sizes at maturity were similar regardless of mesh size. Fifty percent of R. argentea caught with the 5 mm mesh were found to mature at 43.5 (females) or 46.6 mm SL (males). In the 10 mm mesh catches, 50% maturity in females was attained at 44 mm and in males at 46.5 mm. These values were statistically not different at p < 0.05.
Mean numbers of eggs per individual female varied little between fish caught with the 5 mm and 10 mm meshes, but individual fecundity varied more dramatically with infection by the parasite Ligula intestinalis. Non-infected R. argentea caught with the 5 mm net had a mean number of 1050 ± 400 eggs and those caught with the 10 mm net had 918 ± 28 eggs, a difference lacking statistical significance (p > 0.05). In contrast, the difference between the non-infected (1050 ± 400 eggs) and the infected females (663 ± 316 eggs) was statistically significant (p < 0.05). In the combined data for both mesh sizes, the fecundity of non-infected fish increased strongly (R2 = 0.515; N = 568; p < 0.01) with increasing fish length unlike the infected R. argentea in which the correlation of fecundity with fish size was weaker (R2 = 0.368; N = 171; p < 0.01). Sex ratio of R. argentea over the study period was 1.5:1 in favour of females.
Visual examination of size structure of the catches in the two mesh sizes suggests that the selectivity of the two nets is very similar. This is understandable given the nature of light attraction used in R. argentea fishery. Retrieving all size ranges of fish gathered in a particular area of water is likely because the bigger fishes in the fishing grounds tend to fill up and clog the net. Most of the small sized fish become trapped deep inside the net and are unable to escape, as might have happened in April-July. Hamley (1975) tested efficiency in gillnets and observed that regulating the mesh size of the net, may only be partially effective in minimizing the catch of undersized fish. A similar concept can also apply for R. argentea nets, where their efficiency increases through net clogging.
The average size of fishes caught with both mesh sizes during this study was above the size at first maturity (44 mm), in contrast to what Wandera (1992) reported from Ugandan waters. His observations were based on catch samples collected from beach seines which are normally operated in the inshore waters where the juveniles stay. Small-sized fish observed in catches from both meshes during April, May and June represented recruits to the fishery. In these months, particularly in March and April, most areas on the Tanzanian side of the lake experience the peak rainy season, and after May rain is either scarce or sometimes absent. The appearance of these small-sized fish in the fishery during this particular time of the year reflects a strategic timing in the life history of R. argentea, in which reproduction is associated with the booming of zooplankton organisms which follows the phytoplankton bloom (Mwebaza-Ndawula, 1998).
The catch per unit effort (CPUE), which is considered as an index of fish abundance was another parameter for comparing the 5 mm and 10 mm mesh sizes. CPUE was almost the same for both mesh sizes, indicating that any observable differences in catches could have been attributed to other factors like, personal effort and skill of the fishers rather than the mesh size. Nevertheless, slightly higher CPUE was observed in the 5 mm mesh as expected, presumably because it is easier to operate than the 10 mm mesh net which allows most fish to be gilled. There is possibility, however, that before the fish clog the net bag, a relatively small proportion of small sized fish is lost through the meshes accounting for the slightly lower CPUE in the 10 mm mesh net.
On the other hand, the CPUE of R. argentea in both net meshes showed a decline while that of haplochromines was increasing during April to August in both net meshes. Most probably haplochromines were abundant on the fishing grounds at those nights feeding particularly on Chaoborus larvae that are attracted by light. Chaoborus larvae are also an important food for R. argentea. In this scenario where two different species compete for the same type of food, the more powerful, i.e. the haplochromines, remained in the lighted area to become available to fishers. The larvae usually tend to move towards lighted areas for search of zooplankton as their food (Witte, 1984; Witte et al., 2000).
This temporary decline in R. argentea, however, should not be mistakenly concluded to be a result emanating from over-fishing because the CPUE went up again towards the end of this study in September. Although haplochromines were caught as a by-catch of R. argentea fishery, they formed about 50.1% of the total catch which may be in response to the decline of the Nile Perch stock reported recently (LVFO, 2008b) as a result of the fishing pressure and effects of eutrophication of the lake ecosystem (Kolding et al., 2008). During the period of booming of Nile Perch, haplochromines formed the main food item for Nile Perch. Haplochromines were also abundantly caught in Speke Gulf during the day by artisanal fishers using small seines in July–August (personal observations).
Fecundity does not always relate to reproductive success (fertility), but it can provide an objective measure of reproductive effort (Moyle and Cech, 2000). At times, fecundity may vary annually basing on food supply (Treasurer, 1981), variation in environmental conditions including, temperature and stress, and fish size. It is also widely accepted that larger fish with a greater visceral space for egg development have larger ovaries, and thus more eggs than smaller fish (Thomas et al., 2006). In the same vein, we originally expected in this study that the fecundity of fish caught with the 10 mm mesh would be higher than in the 5 mm mesh, but we could not find a significant difference in size of the fish and consequently in their fecundity. The mean number of eggs per female we found is similar (1050 ± 400 eggs) than the 1073 ± 313 eggs reported by Wanink (1991), and much lower than that (2292 eggs) reported by Okedi (1974). The sex ratio in the R. argentea population was 1.5:1 in favour of females, as reported earlier by Okedi (1974), Wanink (1988), and Wandera (1992). According to Nikolskii (1969), a dominance of female fish would occur where low fecundity prevails in the fishery.
The size structures of R. argentea in the catches of the 5 mm and 10 mm meshed nets proved to be similar indicating a likeness in retention capability of the two nets. During April, May and June all nets landed small-sized fish with some bigger fish, which mostly had attained maturity. However, the small portion which had not reached maturity should not be ignored because it is this part that will form the future fish stocks.
For the purpose of protecting the juveniles, the joint Council of Ministers of the Lake Victoria Fisheries Organization (LVFO) held a meeting in Dar es Salaam on 27th February 2009 and harmonized regulations for R. argentea fishing in Lake Victoria. The resolution of the meeting allowed fishing for Dagaa at a distance of at least 2 km from the littoral to offshore using the 10 mm and above mesh sizes, while lower mesh nets were restricted as a precaution. However, according to the results of the present study, the 10 mm and 5 mm mesh nets did not differ in terms of fish sizes being caught and could therefore be used interchangeably.
The authors would like to thank Dr. Martin Van der Knaap, the Coordinator for Lake Victoria Fisheries Research Project (LVFRP) which provided the funds for this research. Thanks are also extended to Professor P. O. J. Bwathondi of Tanzania Fisheries Research Institute (TAFIRI) and to Mr. E. F. B. Katunzi the TAFIRI Mwanza Centre Director, for permitting the use of the necessary laboratory equipment at the Mwanza centre and for their continued support as immediate supervisors of this work.