The survey was carried out to describe the types of sublittoral habitats and communities adjacent to the coast in the region of the eastern boundaries of the Skomer Marine Nature Reserve. Surveys were carried out on two days. On October 10th 1982, 11 sites were surveyed in the region of Gateholm. On October 24th 1982, 3 sites were surveyed to the east of Martins Haven. Survey staff included biologists and divers recruited from the Pembrokeshire BS-AC Branch 417. Broad descriptions of habitats and communities were made based on the recording card developed for the South-West Britain Sublittoral Survey. Areas to the west and south of Gateholm were found to be predominantly rocky with some areas of sand. Rich plant and animal communities were observed, but it was not considered that any especially unusual communities or species were present. The sea bed to the north of Gateholm included extensive areas of sand but where rock was present, species characteristic of open-coast conditions were more common than to the south. Areas to the east of Martins Haven were remarkably similar in the habitats and communities present. Here, the bedrock and boulder slope with fringing areas of small boulders, stones and shells, extended only about 30 m from the shore and to depths of about 14 m. Offshore, there was plain of muddy sand charaterised by a fairly epifaunal community. Some alterations to the habitats mapped by P. Hunnam in 1976 were made and it was also found difficult to match descriptions from the present work to the small number of categories included on the 1976 map. It was concluded that the absence of any sublittoral habitats or communities of outstanding interest south of Gateholm would allow the termination of the reserve boundary at the western point of Gateholm and that, to include the areas of offshore rocky sea bed, there could usefully be a straight-line boundary from Gateholm Point to 500 m south of the Mew Stone. On the northern coast, it was suggested that, providing the boundary extended at least as far east as High Point, there were no biological grounds for precisely determining its eastern extent beyond that. The presence of very extensive areas of muddy sand with similar communities from about 40 m offshore provoked questioning of the offshore extent of the reserve boundary which could be as little as 200 m to include the main communities of conservation interest. However, practical points of navigation and enforcement would be important considerations and no recommendations have been made. It was also pointed out that a straight-line boundary from 500 m north of the Garland Stone towards the east would be practical. Records currently considered sensitive have been removed from this dataset.