Sri Lanka has more than 12.000 artificial reservoirs and no natural lakes as is true for many other areas in SE Asia. These reservoirs cover a total surface area of around 175.000 ha. They are shallow with maximum depths often less than 5 m and exhibit considerable fluctuations in water level. In general the fish production, estimated from commercial catches, is high in these reservoirs, but lower than would be expected on the basis of the high primary production. This situation changed drastically when around 1952 the exotic tilapias O.mossambicus and O.niloticus were introduced. From that moment onward the fish yield increased and the proportion of inland fisheries to total commercial fisheries rose from almost 0% up to the current 20%. This freshwater fish yield is dominated by the introduced tilapias. The cause of this successful introduction of the tilapias has long been debated. Two different hypotheses exist. Firstly, it has been suggested that the relatively low fish production (based on the yield) in the reservoirs before the introduction of the tilapias was probably due to the origin of the fish, mainly indigenous riverine carps, which may not be optimally adapted to their new conditions. According to this hypothesis the introduced tilapias were competitively dominant and were able to capture niches formerly occupied by the indigenous fish species. Secondly, the alternative hypothesis assumes that the tilapias are able to occupy an empty niche when introduced in the reservoirs and consequently did not compete with the indigenous fish species. Determination of the correct theory is important, both from a theoretical and applied (fisheries management) point of view. This is attempted by studying, for the first time, the effects of the exotic lacustrine tilapias on the indigenous riverine fish species, within the framework of a tropical fish community.
The present study was conducted at Tissawewa reservoir. Tissawewa is a typical shallow lowland irrigation reservoir (average depth = 1.2 m) with a surface area of about 200 ha in the dry SE corner of Sri Lanka. In this study the total biological fish production was estimated at 2430 kg/ha/yr of which about 20% consisted of the exotic tilapias. In contrast, about 80% of the commercial catches of 242 kg/ha/yr consisted of these species. This strongly suggests that the importance of tilapias in terms of biological fish production was markedly overestimated not only in Sri Lanka but most probably in all of SE Asia because all estimates of fish production were based on commercial catches. The gross primary production was estimated at about 13.000 kg C/ha/yr of which 1.9% is transferred into fish. This efficiency is high compared to efficiencies between 0.2% and 1.6% reported for other lakes and reservoirs. This high transfer efficiency of fish production can be explained by the short food chain. Most of the fish biomass (ca 45%) consists of a small indigenous cyprinid, A.melettinus, which feeds predominantly on phytoplankton and detritus. The tilapias (which make up ca. 9% of the fish biomass) are also herbivorous. These primary consumers represent about 64% of the biological fish production in the reservoir. The secondary consumers represent 36% of the biologic production while only 0.3% of the fish production is made up by the third trophic level, that of the tertiary consumers.
The structure and fuinctioning of the fish community can be regulated according to several mechanisms such as a deterministic versus a stochastic regulation or, when deterministically regulated, by bottom-up (regulated by primary producers) versus top-down (regulated by predators) control. A deterministically regulated community is generally at an equilibrium, with population levels at carrying capacity determined by resource limitations and coexisting species avoiding competitive exclusion through biotic interactions such as resource partitioning. In contrast, a community is stochastically regulated when the environment is not stable enough to allow an equilibrium to persist. The abundances of species in such a community are determined largely through differential responses to unpredictable environmental changes, rather than through biotic interactions. Based on the disturbances, caused by water level fluctuations, acting on the fish community and community characteristics such as resistance and resilience, the Tissawewa fish community is considered to be deterministically regulated. However, extreme low water levels or the drying up of the reservoir which occur occasionally can cause a relatively brief time of stochastic regulation. Both these phenomena could be studied because during the sampling period, due to a long period of extreme drought, the water level decreased until the reservoir was completely dry. After filling up a different ecosystem had evolved. Before the drought (Period 1) Tissawewa had vegetation only in the shallow, littoral zone and a high turbidity due to resuspension of the thick layer of detritus on the bottom. After the drought (Period 2) vegetation was found all over the reservoir, covering the entire water column, and the water was significantly clearer due to a lower concentration of suspended detritus and phytoplankton. During Period 1 the average fish biomass was about 1700 kg/ha, while during Period 2 this was only 770 kg/ha. The community was at equilibrium at the beginning of the sampling period but as the water level decreased below 1.5 in maximum depth, this equilibrium was disturbed. About one year after the reservoir refilled, a new equilibrium different from the former equilibrium was established. The species mainly affected by these environmental changes was A.melettinus. Its share in the total fish biomass decreased from on average 45% during Period I to 9% during Period 2. This was mainly caused by a decrease of its main food source, that of suspended detritus, from respectively 7.3 g C/m 3to 2.1 g C/m 3. This also illustrates the importance of bottom-up regulation for the structure of the Tissawewa fish community.
The fact that the structure of the Tissawewa fish community is deterministically regulated together with the importance of bottom-up regulation implies that competition and resource partitioning are important mechanisms determining the abundance of the different populations. Therefore the niche occupation of the ten most important fish species was determined along three dimensions: the trophic, the spatial and the temporal dimension. The position along the trophic dimension depends on the diet of the species, the position along the spatial dimension on the distribution over the various habitats in the reservoir and the position along the temporal dimension depends on the time of day the species were actively foraging. Because the body size of fish can increase several orders of magnitude during ontogeny which, in turn, can have considerable consequences for the niche occupation of an individual, each species was subdivided into size-classes for which niche occupation was determined separately. Not only the size-specific niche occupation along each dimension was determined in this study, but also the interaction between the different dimensions. In order to comprehend the impact of these interactions two types of competition should be distinguished: exploitative competition and interference competition. The former describes how species affect each other through the exploitation of the same food sources, the latter deals with a more direct interference through the capturing of space.
Exploitative competition mainly depends on the interaction between the trophic and spatial dimension. In other words: what food is taken from where while the time at which this happens is unimportant. Interference competition depends on the interaction between the spatial and temporal dimension: where and when is a certain species active. With regard to the time of day a species is actively foraging, seven diurnal species were observed, two nocturnal piscivores while one species was mainly active during dusk and dawn. When actively foraging the distribution of a fish is determined by a trade-off between foraging profitability and risk of predation. For those species or size-classes subject to predation, avoidance of predation is the main factor determining their distribution even if this results in suboptimal foraging. The juveniles of all species avoid predation by using the cover provided by the vegetation. The main adult prey, A.melettinus, H. gaimardi and R.daniconius avoid predation by evading the bottom layer of the reservoir where the piscivores reside. For those individuals which are not foraging their distribution is aimed at avoiding interference competition and predation. This is realised by all fish through a migration upward along the vertical plane. Environmental changes affecting the risk of predation and foraging rate are reflected in the distribution patterns of the fish species involved, confirming the ability of species to facultatively respond to these factors. Because fish density and consequently interference competition was lower after the drought, the upward migration also decreased.
For the Tissawewa fish community a correlation was observed between the morphology of a fish and its diet. This diet was apparently determined by only a few morphological characters. The gape width constrained the maximum size of the food particles, the gut length was related to the potential of utilising vegetable matter and the orientation of the mouth and the presence of barbels were correlated with the position of the food along the vertical gradient. Assuming functional relationships, the potential niches of the species were established, and within each species the main ontogenetic differences could be explained. The realised trophic niches of each population both before and after the drought coincided with what could be expected from the relative positions of the potential niches of these populations along the trophic resource dimension assuming niche segregation.
It was observed that size-specific measures of calculating niche breadths and niche overlaps, in which ontogenetic changes can be incorporated, present considerable advantages to the conventional measures when studying intra- and interspecific competition. For example, niche breadth which can be considered an indicator of the flexibility of a species to adjust to changes in food availability is markedly affected by ontogenetic changes; species with large ontogenetic changes would be considered more flexible then they actually are. Species of which the juvenile stages appeared to compete for resources would not be considered potential competitors based on the conventional measures. Also, possible competition for food was observed between the zooplanktivorous juvenile stages of several species with non- zooplanktivorous large adults, and the adult zooplanktivores: the juvenile competitive bottleneck. The abundance of a species is determined by the combination of intra- and interspecific competition. The calculated niche overlap measures were compared: 1) within one species between ontogenetic stages; 2) between species; 3) for the entire fish community between the periods before and after the drought. This strongly suggested that: 1) intra-specific competition is lower for species with large ontogenetic changes; 2) inter-specific competition is lower for specialists; and 3) competition for resources exists. The latter was concluded because during Period 1 when fish biomass was highest, niche segregation was highest and consequently average niche overlap was lowest. For the Tissawewa fish community segregation along the trophic dimension was more important than along the spatial or temporal dimension.
Based on niche overlap along three resource dimensions, the main potential indigenous competitor of the exotic introduced tilapias is the small pelagic cyprinid, A . melettinus. However, a more detailed analysis of the foraging behaviour of these potential competitors shows that A.melettinus mainly feeds on the fine particulate detritus and phytoplankton, suspended in the water column, while the tilapias graze the deferral aggregate from the bottom. This implies that A . melettinus is not a true competitor of the tilapias and might even benefit from the tilapias breaking down the flocculant material from the bottom into smaller, better suspendable, particles. Therefore, the successful introduction of the lacustrine tilapias can be explained from the ability of these species to occupy a previously unfilled niche, and as a consequence the indigenous fish community was most probably not harmed by this introduction.
The acquired knowledge of the Tissawewa ecosystem pertaining to the effects of, among others, competition, predation and water level fluctuations on the fish community is used to discuss the consequences of various management measures on the fish production and fish yield. It was concluded that an additional exploitation of van the small pelagic fish species, either directly, or through introduction and exploitation of a large (controllable) piscivore, will most probably enhance the existing fishery on tilapias. Based on the Tissawewa case study it is concluded that at least for Tissawewa, but most probably for all similar tropical waterbodies a markedly higher utilisation efficiency (yield/ production) can be achieved than the 10% that is currently assumed.