North American Pacific salmon (Oncorhynchus spp.) are heavily exploited in coastal fisheries of the gantlet class (Paulik and Greenough, 1966). The Canadian fishery of Juan de Fuca Strait, British Columbia, is a particularly complex example involving four gear types: gillnet, seine, troll and sport which harvest, at various times, large numbers of all salmon species. Because salmon are highly available to fishing gears, exploitation must be carefully regulated. This study, based on various field data and catch statistics, documents factors affecting exploitation: seasonal timing of exploitable salmon, distribution and amount of fishing gear, relative gear efficiency, accessibility of salmon to the gear, vulnerability of salmon to the gear. All species and gears are covered to varying degrees. Each species has a characteristic seasonal timing, but species vary in run duration and timing consistency between years. There are considerable overlaps in species timing which complicate intraseasonal management. In general sockeye (O. nerka) enter in July and August followed by pink (O. gorbuscha) from mid-August to early September, coho (O. kisutch) in September, and chum (O. keta) in October. Chinook (O. tshawytscha) migrations intermingle with all species. Additionally, chinook and coho are exploited on oceanic migrations. Fishing gears are distributed over ninety linear miles from the Bonilla-Tatoosh net line to Victoria. During the August-October net fishery seines fish within five to ten miles of the net line; gillnets fish offshore, from the net line to Sheringham Point, the eastern commercial boundary. Sports fishermen are clumped near shore, east of Sheringham Point, in close proximity to launching or marina facilities. Gear types showed obvious overall differences in relative gear efficiency, based on catch and effort statistics from two or more gear types operating at the same time in a particular area. For example on coho, one seine equals 265 sport units; one gillnet, 63 sport units; and one troller, 8 sport units. Migrating salmon of all species favoured offshore Canadian waters except near Sooke; all species avoided waters east of Race Rocks where 30 per cent of the sport fleet fishes, the discrepancy was least pronounced for chinook. Based on troll catches using standardized gear, coho favoured surface waters above 27 meters; chinook were most abundant below 36 meters. During periods of spawning migration activity, all species favoured the 18-36 meter depth stratum. Gillnets were directionally size selective for all species, but direction and intensity of selection varied between species and between months within species. Because fleet mesh distribution remains relatively constant each year, changes in fish size will have a pronounced effect on gillnet exploitation. Troll gear was species and size selective; however of importance, subtle fishing techniques have a significant effect on selectivity of lures and may be a serious source of bias in empirical lure studies. Coho decreased in susceptibility to hook and line gear between mid-August and mid-September, apparently due to decreased feeding intensity; this has the effect of lowering hook and line catch success for constant abundance. Future studies on lure selection should stress selective mechanisms rather than empirical description. The complexity of inter seasonal and intraseasonal management strategies applicable to the Juan de Fuca fishery undoubtedly are best studied using techniques of systems analysis. However, present gantlet fishery simulation models (Royce et al., 1963; Paulik and Greenough, 1966, detailed in Greenough, MS 1967), although highly sophisticated, lack sufficient generality for direct application to the Juan de Fuca situation. / Science, Faculty of / Zoology, Department of / Graduate