Photoquadrats were used to determine coral abundance and condition and to estimate coral cover. In each location, 30 m transects are extended parallel to depth contours (Figure 1) . A total of 10 quadrats are placed along each transect at 3 m intervals, with each quadrat (70 x 100 cm; 0.7 m² area) positioned such that the narrow (70 cm length) side is parallel to the line. Numbered tags are placed next to each coral within the quadrat to assist in identification of the corals (Figure 2) . For each quadrat, one diver records the coral species associated with each number, and notes the condition of the coral. A second diver photographs the quadrat. All quadrats are photographed from a planar perspective at a vertical distance of 1.2 m above the quadrat. This distance is maintained by using a 1.2 m stick. Although variation from true vertical to the surface may introduce cover errors, these are slight. Cover is underestimated because the photo is a planar projection of three-dimensional coral.

Coral cover is determined for each quadrat using Deneba Canvas that has an area measuring function. Initially, individual corals are identified on the photograph and the perimeter of the coral is marked with a pen. The photograph is then scanned and individual corals are traced on the photograph and measured to determine the area in cm² relative to the area of the print . All data are entered into a spreadsheet in order to calculate percent cover by species and total cover for each quadrat and for the total transect, average colony size, and number of colonies.

In this study, "coral cover" is defined as the amount of the total bottom area occupied by living coral colonies within an entire area that includes sand channels and bare areas.

Whenever possible, corals identified within quadrats were recorded to species level. Although the Montastraea annularis group (species complex) have been subdivided into three species as proposed by Knowleton, ¹⁹⁹² many of the surveys used in this paper were made before the introduction of new species. Therefore, the cover is designated the Montastraea annularis group and the individual species are not distinguished. Forms or morphotypes of Agaricia agaricites, Porites porites and Meandrina meandrites were not differentiated. In addition, encrusting colonies of Madracis were recorded as Madracis decactis. In this study, we only recognized one species of Stephanocoenia (Stephanocoenia intersepta) and one species of Dichocoenia (Dichocoenia stokesii). The percent cover by all Mycetophyllia species and Diploria species was so low that both are recorded only to genera.