Three adjacent pairs of reefs, located in the central region of the Great Barrier Reef were chosen for this study. One reef of each pair had been impacted by a crown-of-thorns starfish (COTS) outbreak in the previous 5 to 7 years. The reefs chosen were: the mid-shelf reefs, Grub Reef (impacted) and Centipede Reef (non-impacted) and 2 pairs of outer-shelf reefs, Yankee Reef (impacted) and Bowl Reef (non-impacted), and Dip Reef (impacted) and Coil Reef (non-impacted). Surveys were carried out on the windward reef slope of each reef at depths between 3 m and 7 m.Substratum cover and the density and size of herbivorous fish were surveyed on three occasions (June 1991, January 1992 and March 1993). On each occasion 6 sites were selected at each reef and 3 replicate, 45 x 6 m transects surveyed at each site. Seven species of roving herbivorous fish were selected for analysis of density and size: Acanthurus nigrofuscus and Zebrasoma scopas (Acanthuridae), Scarus frenatus, Chlorurus sordidus, and Scarus niger (Scaridae), and Siganus corallinus and Siganus vulpinus (Siganidae). An observer swam slowly along each transect (10 m/min), estimating size classes and counting all fish occurring 3 m on either side of the transect as it was being laid out.On completion of the fish surveys, the observer swam back along the transect identifying the substratum cover at a point at each 1 m interval. This resulted in 46 point intercepts (0 to 45 m). To distinguish between intensively grazed surfaces and damselfish territories, turf algal cover was split into 2 arbitrary categories: =1 mm.Feeding rates of two species of herbivorous fish (Acanthurus nigrofuscus and Scarus frenatus) were determined by counting the number of bites taken during a 5 min observation period for each individual. Seven replicate observations were made at each of 8 sites within each reef between 1000 hrs and 1600 hrs. Fork length of Acanthurus nigrofuscus was estimated in 1 cm size classes. Observations of Scarus frenatus were made on initial phase females only.The diet of Acanthurus nigrofuscus was analysed for 53 specimens collected from impacted reefs and 41 from non-impacted reefs. Fish were collected with spear guns and monofilament nets (15 mm mesh size) at all reefs on 2 separate occasions, April and October 1992. The stomach was dissected out and preserved in 70% seawater formalin. In the laboratory, the gut was split open and the contents washed into a petri dish, mixed with tap water and broken up with a dissecting needle. A 'clump' of this material was placed on a microscope slide, and further teased apart under low power. At 100x magnification, 4 areas of each slide were examined using an ocular micrometer with 25 intersection points. Any alga falling across one or more of these points of intersection was identified and counted. This procedure was repeated 3 times for each stomach sample, resulting in a total of 300 points of intersection.Acanthurus nigrofuscus were also collected with spear guns and monofilament nets (15 mm mesh size) at all reefs on 3 separate occasions, April and October 1992 and April 1993 for age and growth studies. Fish were measured (fork length (cm)) and weighed (g). The sex of individuals was determined according to appearance of gonads, and selected gonad samples preserved for histological examination. The head of each specimen was tagged and frozen for later removal of otoliths.The length and width of each otolith were measured using a dissecting microscope and an ocular micrometer. All age determinations were based on transverse sections of the sagittae. All sections were examined under a dissecting microscope and counts were made in the same region, which showed identifiable and readable growth rings. Transverse sections were photographed under a compound microscope (40x magnification) using colour 200 ASA film.In May 1992, 300 individuals were tagged using tetracycline on an isolated coral bommie at Bowl Reef to validate age estimates of Acanthurus nigrofuscus. Fish were collected and initially placed in a submerged holding tank, after which they were transferred a few at a time to the boat, measured (FL), injected with the required dosage of tetracycline, and released. All specimens were tagged externally with fine T-bar anchor tags inserted in the dorsal musculature between the 2nd and 3rd dorsal spines. In April 1992, 1 specimen was recaptured and a further 4 specimens were recaptured in November 1993. Sagittae were examined under a compound microscope with an ultraviolet light source to identify the fluorescent band resulting from the treatment with tetracycline.
