Traditional Antillean arrowhead fish traps were deployed around and on a shallow, 8-ha coral reef of south-western Puerto Rico in two 1987 studies: Study I, 2–18 March, a comparison of 30 trap catches and 30 visual transect counts; Study II, 8 April–18 June, a tag and recapture study of 95 trap hauls. Study I considered model taxa and conditions for calculating effective fishing area (EFA) of the traps on the basis of catch per haul and fish density from transects. Of the 185 fish of 24 species caught, EFA could only reasonably be calculated for three taxa: redband parrotfish, Sparisoma aurofrenatum, 25 m 2 and 90 m 2; stoplight parrotfish, Sparisoma viride, 24 m 2; ocean surgeon, Acanthurus bahianus, 93 m 2. Other taxa were considered unsuitable because of near-zero density or catche, e.g. white grunt, Haemulon plumieri, was often caught but seldom seen. Ability to assess EFA depended on trap location; future surveys of reef fishes will necessarily have to utilize stratified sampling based on habitat characteristics. Study II considered fish movements within a five-trap, cross-shaped array deployed on a back reef area. Of 702 captures from 28 species, redband parrotfish, white grunt and ocean surgeon dominated. Of 272 fish tagged, 65 were recaptured once and 24 were recaptured twice or more. In the short term, traps seemed to sample a discrete population. Except for one individual, all tagged fish were recaptured at the same trap or at a nearby trap of the array, i.e. no more than 25 or 35 m away; ocean surgeon were always caught at the trap where tagged. Using the mark-recapture data, population estimates were calculated for dominant species. These were not used to estimate EFA because the area sampled by the traps could not be reliably estimated. However, properly designed marking studies could be used to obtain this information. It was concluded that using EFA in trap-based assessments is a promising technique for residents like redband parrotfish or ocean surgeon. Mark and recapture techniques may also return useful density estimates for cryptic or diurnally migrating residents such as white grunt.