A total of 18 BRUVS sets were made at Mandu, Point Cloates and Mauds Landing in depths ranging from 66-96 metres. Most of the sets were made where the habitats in the immediate field of view were flat sediments, with or without visible bioturbation for the activities of small burrowing animals. Three of the six sets made off Point Cloates had significant cover of benthic sponges, fan and whip gorgonians, and other biotic or abiotic structures. Only two of the nine sets made offshore from Mauds Landing had such cover, and it was much sparser in the field of view of the cameras. All sets at Mandu were in flat, sandy habitats. All the BRUVS sets deeper than 80m provided very gloomy, grainy footage that could be interrogated only by adjusting the brightness and colour balance of the screen on the tape player. At this time of year, and in these coastal waters, these depths are at the lower limit of penetration of daylight. Beyond them artificial lighting must be employed to obtain useful footage. There were 409 individuals from 52 taxa in 25 fish and elasmobranch families recorded on the 18 BRUVS sets (Table 2). Most taxa were identified to species level, with the notable exception of the very small juveniles of threadfin breams (Nemipterus sp), and the small (<60cm total length) sharpnose or sliteye sharks (R. taylori or L. macrorhinos). Published diagnostic features for these species could not be resolved on the footage. The diversity and abundance of fish differed greatly amongst the coarse habitat types recognised on the BRUVS footage (Table 3). The sets in megabenthos had more than twice as many species and individuals as the sandy habitats, even though fewer sets occurred in such habitat types. There appeared to be a trend in increasing diversity and abundance in habitats with increasing grain size for the sets off Point Cloates, but not for the sets off Maud?s Landing. The increase in species diversity and fish abundance apparent in Table 3 may reflect the larger number of sites visited at Mauds Landing. The measures of the various depths, the habitat classes, and the fish species abundance were analysed from multiple BRUVS sites, or sets. Multivariate analyses are required to examine the multidimensional relationships among the sites, the environmental variables and fish species. These analyses allowed us to describe the major groupings in the BRUVS data, measure the strength of fish-habitat associations, and identify indicator species defining these patterns.
