a b s t r a c tDue to the production of toxins, cyanobacteria may adversely affect economically important fish such as Nile tilapia Oreochromis niloticus in tropical lakes. We studied the diet composition and factors affecting the diet of Nile tilapia in two tropical lakes where cyanotoxins were present. Particle-bound microcystins were present in all analysed water samples, ranging in concentration from 0.00012 to 1.11 and from 0.006 to 0.254 mg L À 1 in Murchison Bay in Lake Victoria and Lake Mburo, respectively. Detritus and phytoplankton were the main dietary components of the Nile tilapia, with phytoplankton contributing to over 30% by volume of stomach contents. The cyanobacteria Microcystis spp., which are also the most likely source of microcystins in the lakes, accounted for more than 80% of ingested phytoplankton. Microcystis spp. were also the most abundant cyanobacteria in both lakes (460%). We found no significant relationship between the contribution of phytoplankton in Nile tilapia diet and the concentration of microcystins in the water but we found a close association between water transparency and the contribution of insects to Nile tilapia diets in Murchison Bay. Our results further show that none of the other measured environmental variables was a good predictor of diet items in Nile tilapia. Adult Nile tilapia in our study lakes, rely heavily on filter feeding, particularly under conditions of low water transparency, trapping detritus and phytoplankton cells especially colonies. They can ingest more mobile prey like insects and insect larvae when the water transparency and visibility increases.

A study aimed at investigating the temporal variation of phytoplankton assemblages in Lake Nyamusingiri was carried out during the period of December 1997–May 1998. Uganda’s freshwaters are ecologically diverse but a few are intensively studied. Research on phytoplankton has been restricted to large water bodies. There is little information on phytoplankton of the western Uganda crater lakes, which are important water and biodiversity resources. This study provided baseline data on phytoplankton, which will serve as a basis for monitoring the effects of human activities on the lake that might result in ecological transformations like loss of biodiversity because of overexploitation. A laboratory thermometer and Winker’s method were used to determine temperature and dissolved oxygen concentration, respectively. Lake transparency was measured by using the Secchi disc. A Van Dorn sampler was used to collect water samples. Nutrient and chlorophyll a concentrations were determined by using facilities at the Fisheries Resources Research Institute (FIRRI), Jinja. The Sedgwick‐Rafter counting chamber was used to analyse phytoplankton. Variation in temperature was small (25.4–26.2°C). Stable thermal stratification was not evident. The Secchi disc transparency was less than unity. The chlorophyll a value was high. Biomass was found to be light‐limited by nonalgal materials. Dissolved oxygen concentration was more than 100% in the surface waters but declined to <20% at the bottom, which reflected the eutrophic nature of the lake. Diversity indices were low. Eighteen species and five classes of phytoplankton were revealed by this study. The phytoplankton flora was dominated by chlorococcal green algae characteristic of the large eutrophic East African lakes.

Temperature, oxygen and nutrients were determined in the permanently stratified Lake Bukoni from July 2004 to December 2005. Meromixis has resulted from the nature of the surrounding hilly topography and the almost vertical crater walls that exclude the effects of light winds. These circumstances and the tropical temperature regime ensure permanent stratification in the offshore part of the lake. From about 20 m to the bottom, the water was out of circulation and devoid of oxygen. Water samples from such depths had a strong smell of hydrogen sulphide gas. The inshore site, with greater turbulence and littoral area, had a uniformly oxygenated water column compared with the offshore site. According to information from members of the community around the lake, ‘fish kills’ have taken place but on irregular occasions. The nutrients (N, P and Si) were well distributed in the inshore site. In the offshore site, the concentrations were generally lower in the uppermost 20 m but more than double near the mud surface. Soluble reactive phosphorous was the least abundant nutrient while soluble reactive silicon was the most abundant nutrient. The high Si : P ratios have led to the dominance of species like Synedra ulna Ehr.

Lake Mburo and Lake Kachera are shallow, eutrophic lakes in mid‐western Uganda. Lake Mburo recorded higher values of Secchi and eutrophic depths and lower extinction coefficient (k) values. The lakes showed a ‘red shift’ phenomenon in maximum light transmission. The average values of electrical conductivity in Lake Mburo and Lake Kachera were 136 and 244 μS cm−1, respectively. The pH values indicated high photosynthetic activity. Dissolved oxygen concentration averaged 6.9 and 7.8 mg l−1 in Lake Mburo and Lake Kachera, respectively. The lakes had high total nitrogen (TN) : total phosphorus (TP) ratios averaging 200 and 280 in Lake Mburo and Lake Kachera, respectively. The lakes are dominated by cyanobacterial blooms that reduce light penetration to less than 1 m. Lake Mburo had a lower algal biomass than Lake Kachera. Chlorophyll a concentrations correlated positively (r = 0.73, P < 0.05) with the extinction coefficient in Lake Mburo but not in Lake Kachera. The correlations between chlorophyll a and TN and TP were also high. Both lakes recorded high primary productivity, Lake Mburo showing higher values. The study highlighted the need to investigate the organism–community interrelationships in the two water bodies.