REEF FISH ASSEMBLAGES IN NORTH-WESTERN SRI-LANKA - DISTRIBUTION PATTERNS AND INFLUENCES OF FISHING PRACTICES

The distribution and abundance of reef fishes in relation to habitat s tructure were studied within Bar Reef Marine Sanctuary (BRMS) and on a n adjacent reef, disturbed by destructive fishing techniques, in north western Sri Lanka, by visually censusing 135 species groups using fift y metre belt-transects. Two types of continental shelf patch-reefs are found in the study area: coral reefs and sandstone reefs, which are d ivided into distinct habitats, four for the coral reef (shallow reef f lat, shallow patch reef, deep reef flat and Porites domes) and two for the sandstone reef (structured sandstone-reef and flat sandstone-reef ). Fish assemblages varied in structure between reef types and among h abitats within reef types. Functional aspects of habitat structure and composition, such as available food and shelter, seemed to be importa nt factors influencing distribution patterns. The strongest separation in the organisation of fish assemblages in BRMS was between reef type s: 19% of all species were confined to the coral-reef patches while 22 % were restricted to the sandstone reef patches and 59% were represent ed on both reef types. In terms of distribution among habitats, 21% of all species were restricted to one habitat while only 1.5% were prese nt in all. The highest density of fish was in the coral reef habitats while highest species diversity was found in the most structurally com plex habitat: the structured sandstone-reef. This habitat also had the highest proportion of species with restricted distribution. Planktivo res were the most abundant trophic group in BRMS, and the species comp osition of the group varied among habitats. The comparison of the dist urbed reef with BRMS suggested that habitat alteration caused by destr uctive fishing methods has strongly influenced the fish community. Wit hin the fished area the structure of the fish assemblages was more het erogeneous, fish abundance was lower by an order of magnitude and spec ies numbers were lower than in BRMS.