The major groups of motile cryptofauna inhabiting dead coral substrata across a transect at Davies Reef were quantified in terms of abundance, biomass and density.Samples of dead corals of similar size were collected by enclosing a piece of protruding dead coral substrate with a plastic bag and detaching a piece with a masonry hammer. Collections were made at 7 locations in November 1985: on the front slope at 10 m depth; on the front crest; on the main flat; where the main flat borders the lagoon at 10 m depth; on the outer back reef on the flat; and on the outer back reef at 10 m depth. Five replicate samples were collected at each of 6 sites (3 inside, 3 outside damselfish territories) where possible, at each location. Stegastes apicalis territories were present in the reef flat and reef crest locations. The front slope location contained Plectroglyphidodon lacrymatus territories. Two locations did not have damselfish territories. Sampling was repeated in August 1986 at 2 locations (reef flat and reef crest), to provide an estimate of seasonal variation in cryptofaunal abundance. Ten replicate samples were collected both within and outside fish territories at each location.Cryptofauna were extracted in the laboratory and any animals too large to be washed through a sieve of 1 cm mesh were excluded from the analysis. Individuals within each of the 5 most common taxonomic groups (amphipods, copepods, other crustaceans, polychaetes and molluscs) were counted. Separate counts of crustaceans were also made to a finer taxonomic level (gammarideans, caprellids, isopods, cumaceans, tanaids, ostracods, decapods). Biomass was measured (ash-free dry weight) for the 5 common taxonomic groups for the November 1985 samples only.Chlorophyll a was used as an index of algal biomass present on the coral rock. Following pigment extraction, the surface areas (SA) of coral blocks were estimated using a latex-coating method. Blocks were photocopied, and their projected surface area (PSA) measured by digitising those images. The ratio of PSA to SA was taken as a measure of surface complexity.The expected food consumption rate of cryptofauna for each sample was calculated using the published allometric relationships between measured ingestion rate, individual body mass and the organic fraction of the food source (epilithic algal community), for a range of small macroinvertebrates representative of the taxa collected in the study. Ingestion rates were converted to cryptofaunal group rates based on the average density of that group for all locations.