The influence of several substratum variables, particularly structural complexity and Live coral cover, on coral reef fish species richness and abundance was investigated. By surveying fish species with varying strengths of association with the substratum and at different sampling scales, this study aims to assess the scale-dependence of correlations between substratum characteristics and fish comn~unity parameters. Fish from 5 families (Acanthuridae, Chaetodonhdae, Labridae, Pomacentridae and Scaridae) were counted along transects by visual census. Substratum charactenstics were measured from stereophotographic transects. Three measures of structural complexity were used: (1) an index of surface area, (2) the vanety of different substrata (biological diversity), and (3) numbers of holes (of hfferent sizes). Although the depth of peak specles richness varied between families, overall there was a trend of increasing species richness with increasing depth. Surface index was a poor predictor of overall species richness but was significantly correlated with pomacentrid species richness and abundance on 200 m transects. In contrast, biological diversity of the substratum (H') was highly correlated with overall fish species nchness. Numbers of holes of 3 sizes taken together accounted for 77 O/ O of the variance in overall fish abundance on the 200 m transects. None of the measures of structural complexity were significantly correlated with species nchness or abundance of small, strongly siteattached fish. Live coral cover was not significantly correlated w t h species richness or abundance of fish from any family (except for a weak correlation w~t h abundance of obligate corallivorous chaetodontids).

We studied the coral-reef fish communities of Saba Manne Park (Netherlands Antilles) and Hol Chan Marine Reserve (Ambergns Caye, Belize) in the Canbbean to assess differences between them and adjacent ecologically sirmlar sites after 4 yr of protection from fishing Forty-five percent of target species commonly recorded in visual censuses in Belize (23 % of all recorded target species), and 59% at Saba (22 %) showed greater abundance slze or biomass in shallow protected sltes These differences are considered prlmanly to reflect increased survivorship with the cessation of f~shing mortahty The greatest estimated biomasses were observed in locally protected snapper (Lutlanidae) in Belize and Saba and grunt (Haemuhdae) at Saba In both protected areas the local stock of vls~ble demersal target f~shes was 1 9 to 2 0 tlmes greater in biomass and 2 2 to 3 5 times greater in commercial value than in fished sites Larger local stock of many target species is likely to support higher egg output from the protected areas while larger predator biomass will mean more intense predation at the protected sites

No abstract

No abstract