Originally conceived as a modified Schnabel (1938) design mark-recapture study, the unique random sampling regime of this long line tagging study has allowed us to describe habitat correlates of Pacific halibut (Hippoglossus stenolepis) and other demersal fishes. Pacific halibut and other fish were captured by longline sets of constant length and hook number distributed in a random stratified fashion. General Position System (GPS) location and depth were recorded at one-second time intervals during setting to obtain a depth profile and track of the set. Hooks were counted during pulling and the hook number was recorded for each fish and invertebrate to allow accurate determination of its capture location and depth. All fish were measured. Sediment type was determined from a grab sample taken at the mid-point of each set. Kruskall-Wallis ANOVAs were conducted to relate abundance of nine common groundfish species to substrate and to depth. Six species had significant relationships with substrate while all nine had significant relationships with depth. There were no significant relationships between fish size and substrate, and only Pacific halibut exhibited a significant relationship between size and depth (Kruskall-Wallis H-22.8, n=974, p<0.001).