The objective of the study was to investigate some aspects of the breeding patterns of the cyprinid Barbus amphigramma in Lake Naivasha. The study was carried out from February to October, 2003. Fishing was done with the use of gillnets of mesh size 0.5 inch to 3 inch and a beach seine of mesh size 1 inch. Six gonad maturity stages have been described visually, based on morphological features, and validated by histological features of the ovary, as well as examination of the oocyte diameter. Females dominate the population at all sizes classes. Barbus amphigramma spawns all year round, but with discernible peaks in March, July and October. The peaks in March and October correspond with the beginning of the long and short rains, respectively, in Kenya.
Barbus amphigramma appeared in the lake in the early 1980's (Harper, 1984). It is the only fish species in Lake Naivasha which has been described as a natural invader, being a small riverine fish which migrated down to the lake from the River Malewa. Although B. amphigramma does not frequently feature in the fish markets around the lake, its exploitation forms a minor subsistence activity of the local fishing community.
Ecologically, B. amphigramma has been reported to be one of the food items for the macro-predator species Micropterus salmoides in Lake Naivasha (Aloo, University of Nairobi, Kenya, unpublished). It was under exploitation and of high commercial value in the mid 1980's with records showing that in 1985, 60.0 tonnes of B. amphigramma were landed, followed by 62.9 tonnes in 1986 (Anonymous, 1987). However, the catch decreased significantly in 1987 to 26.1 tonnes. Only a small number of fish was caught in 1989 (Anonymous, 1989). There have been no records on its exploitation thereafter, or any detailed ecological study describing its biology and interaction with the abiotic and biotic factors in Lake Naivasha (Figure 1). This gap in the lack of knowledge and data, for nearly twenty years, stimulated the interest of the present study in Lake Naivasha, situated in the Great Rift Valley, Kenya. The study was carried out from February to October, 2003.
Materials and Methods
Fleets of gillnets, each with a depth of 1.5 m, a length of 45 m and mesh sizes ranging from 0.5 to 3 inches, were used for fishing. The nets were set overnight parallel to the shore at each of the sampling stations (Figure 1) for 12 hours and examined the following morning. The catch was sorted and identified. B. amphigramma specimens (1247 caught) were weighed to the nearest gram on a top loading electrical balance and their total length measured on a fish measuring board to the nearest mm. They were then dissected, sex determined and the ovary was assigned a maturity stage following a gonad maturation staging scheme described by Ntiba (University of Nairobi, Nairobi, Kenya, unpublished) and Ntiba and Jaccarini (1990). Each gonad was weighed to the nearest 0.001 g on an Ohaus Adventurer analytical balance. Gonadosomatic index (GSI) was calculated from the expression: GSI = WO/(WF-WO) × 100, as described by Hickling (1970) and Ntiba and Jaccarini (1990), and where WO is weight of ovary and WF is the weight of fish. From one of the lobes of the ovary, a portion from the anterior, mid and posterior was cut, rinsed in physiological solution and preserved in Bouin's fixative for a minimum period of 24 hours, then rinsed in 50% alcohol and preserved in 70 % alcohol. This material was then dehydrated in a series of graded alcohols, cleared in xylene and impregnated in molten wax at 58–60°C (Steedman, 1960). The impregnated material was embedded in a block of wax, mounted on a microtome and cut into sections of 5–10 μ m. The sections were mounted on slides and stained using haematoxylin and eosin (H and E) staining technique. The slides were then examined under a standard microscope, and fitted with a micrometer in the eye piece to study the internal appearance of the ovaries at different stages of maturity. Standard statistical tests (Snedecor and Cochran, 1989) were carried out.
The monthly variations in sex ratio of the B. amphigramma are shown in Table 1. The χ2 value was calculated to be 20.342; df. = 8, p < 0.5 shows that in this population females formed a bigger portion of the fish sampled meaning that the sex ratio is significantly different from the expected 1:1 ratio (p < 0.05).
Maturity stages and the breeding cycle
The scheme of Gonad maturation is given in Table 2.
Figure 2 shows the gonad weight relative to body weight at each maturity stage for males and females. These plots more or less confirm the gross morphological criteria used in determining the maturity stages in B. amphigramma by visual inspection.
The weights of the testes and ovaries as a percentage of body weight exclusive of gonads, calculated on a monthly basis, are plotted in Figure 3.
These results show that while there is no congruency in changes of the gonad relative to body weight between the males and females the GSI values remain relatively high and mostly above 2. This alludes to the fact that this fish spawns all year round with probable peaks in March, July and October.
Seasonal occurrence of maturity stages
The percentage monthly occurrence of maturity stages in females alone is shown in Figure 4. All stages (stage I and VI) were present albeit in varying proportions throughout the study which is characteristic of a fish that spawns round the year. The peak spawning periods probably are preceded by a peak in the numbers of stage V ovaries in March, May, August and October. Spent fish are proportionally very low in catches probably because individual fish are multiple spawners carrying batches of oocytes at different stages of development.
The most important conclusion of this study is the clear demonstration that the population of B. amphigramma of Lake Naivasha, Kenya spawns throughout the year. The evidence for this is that morphologically ripe gonads occurred throughout the year. Additionally, gonadosomatic indices remained clearly high, mostly above 2, throughout the year, but also showed peaks during the wet and in the dry season. This coincided with high levels of dissolved oxygen and low temperatures, which made the conditions suitable for spawning. Seemingly young juveniles were present throughout the sampling period, an indication of all year recruitment (Mutia, University of Nairobi, Kenya, unpublished). According to Manyala (1992) the cyprinid, Rastrineobola argentea, from Lake Victoria also breeds throughout the year with peaks in August and December-January. In Lake Kariba, Begg (1974) reported that breeding in Limnothrissa miodon takes place all year round, a fact supported by the presence of larvae, juveniles and ripe adults. Furthermore, Begg identified two breeding peaks that coincided with periods of increased nutrient availability; the first during the annual turnover (June–August) and then again when the tributary rivers are fast-flowing in January–March. Kenya experiences two rainy seasons during March–June and October-December, respectively. This study has showed that the breeding pattern of B. amphigramma is characterized by high (GSI value > 2) gonad activity throughout the year. However, even if breeding takes place throughout the year, peak spawning activities occur in March, July and October. The two peaks in March and October coincide with the beginning of the short and the long rains in Kenya. It was of interest to note while fish with ripe ovaries (stage IV) occurred in the samples almost throughout the year fish with running ovaries (stage V) were found in all the stations sampled during the peak spawning time.
Rainy season spawning is well documented among other Barbus species in Africa. Gaigher (1976) found that spawning in Barbus kimberleyensis, in the Hardap Dam of South West Africa, coincided with peak seasonal rains. Several seasonally spawning Barbus have well-defined seasonal run up the rivers and streams to spawn at the start of or early into the rainy period. Payne (1975) found that Barbus liberiensis showed movement upstream and a single, discrete, breeding season coinciding with the early part of the rainy seasons.
In a small tributary stream of Lake Victoria near Jinja, Uganda, Welcomme (1979) reported well-defined seasonal migrations in Barbus kersteinii, Barbus apleurogramma and Barbus paludinosus. He classified these three species as potamodromus in that they depend on the river for breeding, early development and on the lake for survival of the adult fish. In Rivers Nzoia and Sondu, both tributaries of Lake Victoria in kenya, Whitehead (1959) reported that Barbus altianalis, Barbus nummifer and Barbus doggetti were anadromous, where their breeding coincided with seasonal flooding. However, he classified B. apleurogramma as non-anadromous since its adults and juveniles occurred in rivers, streams, swamps, as well as in the floodwater pools of Lake Victoria during the dry season. Furthermore, he also found they spawned amongst the grasses in one of the lakeside streams during the dry season. According to Sempeski and Gaudin, (1995) adults and juveniles of B. apleurogramma occur in the heavily vegetated swamps surrounding Lake Nabugabo in Uganda during both the dry and wet season. This probably suggests that the reproductive seasonality and migratory tendencies of the population(s) of this Barbus species could change depending on their geographical location and environment.
The significant difference in sex ratio could be attributed to several factors. First, there is a possibility of sex reversal where the males reverse their sex in response to changes in environmental parameters. Secondly, there is a possibility that there is higher male mortality probably attributed to greater reproductive investment. Thirdly, food supply could be a determining factor, where females predominate when food is abundant, while males predominate in oligotrophic environments (Nikolsky, 1963). A more detailed study could unravel the bias of the population towards females.
It is concluded that, although gonad recrudescence showed spawning peaks in March, July and October, several lines of evidence have indicated that B. amphigramma of Lake Naivasha, Kenya, spawns throughout the year. The spawning peaks in March and October are clearly associated with the onset of short and long rains, respectively, in Kenya. Finally, a more detailed study is certainly required to unravel the bias of the population in terms of sex ratio.
This study was supported through the Kenya-Belgium Inter-University Cooperation Programme (VLIR-IUC-UON) and University of Nairobi to whom we are extremely grateful. We also wish to thank the Fisheries Department, Kenya and Lake Naivasha Riparian Association for being very supportive during the study. John Mutembei assisted dedicatedly in finalizing the draft paper.