Gravel habitats on continental shelves around the world support productive fisheries but are also vulnerable to disturbance from bottom fishing. We conducted a 2-year in situ experiment to measure the rate of colonization of a gravel habitat on northern Georges Bank in an area closed to fishing (Closed Area II) since December 1994. Three large (0.25 m2) sediment trays containing defaunated pebble gravel were deployed at a study site (47 m water depth) in July 1997 and recovered in June 1999. The undersides of the tray lids positioned 56 cm above the trays served as settlement panels over the same time period. We observed rapid colonization of the gravel substrate (56 species) and the settlement panels (35 species), indicating that colonization of gravel in this region is not limited by the supply of colonists. The species composition of the taxa found in the trays was broadly similar to that we collected over a 10-year period (1994-2004) in dredge samples from gravel sediments at the same site. The increase in abundance of animals in the gravel colonization trays was rapid and reached a level in 2 years that took 4.5 years to achieve in the surrounding gravel sediments once fishing had stopped, based on data from dredge sampling at this site. The increase in biomass of animals found in the sediment trays paralleled the trend of biomass increase observed in dredge samples over the same period (1997-1999) but was lower in value. These data suggest that after rapid initial increase in abundance of organisms, succession proceeded by increasing individual body size. A comparison of settlement panel and tray faunas revealed that the mean biomass of structure-forming epifauna (sponges, bryozoans, anemones, hydroids, colonial tube worms) on the panels was 8 times that found on the trays. Structure-forming taxa constituted 29% of the mean biomass of the panel fauna but only 5.5% of the tray fauna. By contrast, the mean biomass of scavengers (crabs, echinoderms, nudibranchs, gastropods) in the trays was 32 times that on the panels. Colonization of the tray gravel was more rapid for free-living species (many of which are prey for fish) than for structure-forming epifauna, though colonists of the latter species were present. The reduced success of structure-forming species in colonizing the tray gravel possibly is related to factors such as intermittent burial of the gravel by migrating sand and low survival of new recruits due to the presence of high numbers of scavengers on the gravel. These two factors might explain, to varying degree, the slow recolonization of gravel habitats by structure-forming species in Closed Area II of the northern part of Georges Bank. ?? 2009 Elsevier Ltd.