In Sri Lanka, there is a great potential for the development of culture-based fisheries because of the availability of around 12 000 non-perennial reservoirs in the dry zone (<187 cm annual rainfall) of the island. These reservoirs fill during the north-east monsoonal period in October to December and almost completely dry up during August to October. As these non-perennial reservoirs are highly productive, hatchery-reared fish fingerlings can be stocked to develop culture-based fisheries during the water retention period of 7-9 months. The present study was conducted in 32 non-perennial reservoirs in five administrative districts in Sri Lanka. These reservoirs were stocked with fingerlings of Indian (catla Catla catla Hamilton and rohu Labeo rohita Hamilton) and Chinese (bighead carp Aristichthys nobilis Richardson) major carps, common carp Cyprinus carpio L., genetically improved farmed tilapia (GIFT) strain of Nile tilapia, Oreochromis niloticus (L.) and post-larvae of giant freshwater prawn, Macrobrachium rosenbergii De Man, at three different species combinations and overall stocking densities (SD) ranging from 218 to 3902 fingerlings ha )1 , during the 2002-2003 culture cycle. Of the 32 reservoirs stocked, reliable data on harvest were obtained from 25 reservoirs. Fish yield ranged from 53 to 1801 kg ha )1 and the yields of nonperennial reservoirs in southern region were significantly (P < 0.05) higher than those in the northern region. Naturally-recruited snakehead species contributed the catches in northern reservoirs. Fish yield was curvilinearly related to reservoir area (P < 0.05), and a negative second order relationship was evident between SD and yield (P < 0.05). Chlorophyll-a and fish yield exhibited a positive second order relationship (P < 0.01). Bighead carp yield impacted positively on the total yield (P < 0.05), whereas snakehead yield impact was negative. Bighead carp, common carp and rohu appear suitable for poly-culture in non-perennial reservoirs. GIFT strain O. niloticus had the lowest specific growth rate among stocked species and freshwater prawn had a low return. K E Y W O R D S : culture-based fisheries, fish yield, genetically improved farmed tilapia, non-perennial reservoirs, species combinations, stocking density.

Sri Lanka is blessed with a large number of irrigation reservoirs. Depending on their hydrological regimes they are broadly categorized into perennial and seasonal reservoirs, and are secondarily used for inland fisheries. The Ministry of Fisheries and Aquatic Resources in Sri Lanka initiated a programme to introduce culture-based fisheries in minor perennial reservoirs (< 250 ha) whereby 15 minor perennial reservoirs situated in eight administrative districts were selected for introduction of culture-based fisheries in its initial year of implementation. These reservoirs were stocked with fish fingerlings of catla Catla catla, rohu Labeo rohita, common carp Cyprinus carpio, Nile tilapia Oreochromis niloticus and post-larvae of freshwater prawn Macrobrachium rosenbergii. Stocking commenced in 2004; stocking data from 2004, 2005 and 2006 are considered in this study. Stocking densities varied from 217 to 870 ha )1 year )1 , with an average density of 425 ha )1 year )1 . Fish catch statistics for 2005, 2006 and 2007 were considered; high variability in fish yield among reservoirs as well as in some reservoirs between the years were observed. Mean annual fish yield from all 15 reservoirs increased from 57.3 to 208.1 kg ha )1 year )1 after the introduction of culture-based fisheries. Overall fish production in all 15 reservoirs increased by 263%, ranging from 42.8 to 1344%. Overall contributions of Nile tilapia, catla, rohu, common carp and freshwater prawn to fish harvest were 47.4, 27.2, 16.9, 4.3 and 0.7%, respectively. Biomass gain per fingerling stocked with respect to catla, rohu, common carp and freshwater prawn are 0.51, 0.42, 0.26 and 0.04 kg, respectively, indicating attractive monetary gains through stocking of these species. Results indicate the viability of culture-based fisheries in minor perennial reservoirs. Possible reasons for high variability in fish yield and areas for followup studies are discussed.

No abstract

Sri Lanka is endowed with more than 12,000 reservoirs. Depending on their hydrological regimes, they are broadly categorized into perennial and seasonal reservoirs and are secondarily used for the development of inland fisheries. This paper discusses the importance and status of culture-based fisheries (CBF) in reservoirs of Sri Lanka and assess the impacts of introduction of CBF. The CBF in seasonal reservoirs was initiated in the 1980s and it is well documented. There are around 200,000 ha of perennial reservoirs in Sri Lanka and CBF in perennial reservoirs is a recent development. Introduction of CBF into perennial reservoirs has resulted in significant increases in fish production, increased availability of fresh fish to rural communities, enhanced livelihood opportunities and income for fishers and strengthening the rural economy. Progress achieved so far in respect of seasonal reservoirs has not reached the envisaged levels. Nile tilapia, catla, rohu and mrigal are the main contributory species for fish production through CBF in perennial reservoirs. Although the contribution to fish catches is low, stocking of freshwater prawn (Macrobrachium rosenbergii) is financially feasible. Adequate stocking with suitable species, the existence of active Community Based Organizations (CBOs) and their involvement in planning and implementation of CBF, and the existence of a legal framework which facilitate implementation of activities pertaining to CBF are the key factors for successful implementation and sustainability of CBF in perennial reservoirs.

Gillnet selectivity of two indigenous fish species, Ompok bimaculatus and Puntius dorsalis was determined in a small-scale fishery a t Uruwal Oya, in the Gampaha district. The. optimal lengths (Lopt) of the two species for different mesh sizes were determined by the Baranov-Holt method. The relationships between Lnptand mesh size (M) of gillnet for the two species are described by the following equations: The mean sizes of maturity of 0. bimaculatus (21.6 cm for males and 26.1 cm for females) and P. dorsalis (13.4 cm for males and 14.8 cm for females) approximately correspond to Lo,, in stretched gillnet mesh sizes of 5.8 cm and 3.8 cm respectively according to the above relationships. Therefore there seems to be a need for regulating mesh size of gillnets in this fishery in order to prevent immature 0, bimaculatus and P. dorsalis being caught.