A persistent period of low abundance in what was once the second largest fishery for sockeye salmon Oncorhynchus nerka in British Columbia has kept the Rivers Inlet fishery closed since 1996. Initial speculation about the cause of the decline focused on factors such as reduced egg-to-fry survival, declining quantity and quality of spawning habitat, and reduced fry-to-smolt survival in Owikeno Lake (the only nursery lake in Rivers Inlet). We developed an index of juvenile sockeye salmon abundance by combining direct estimates of abundance from trawl surveys with indirect estimates of abundance inferred from density-dependent growth of juvenile sockeye salmon. Juvenile growth data were available as either direct samples of presmolt weight or as measurements of freshwater growth from the scales of returning adults. Collectively, these data do not indicate a long-term decline in juvenile sockeye salmon abundance since the 1950s. Throughout the 1970s and 1980s and even more recently (1991 and 1994 brood years), the juvenile abundance index exceeded the long-term mean. If freshwater abundance was either untrended or increasing, the most likely cause of the population decline would have been lower survival after the fry stage, which would have been noticeable in the 1970s and especially from 1992 to 1998. Poor marine survival is the most parsimonious explanation for the declining fry-to-adult survival in Owikeno Lake, particularly in light of coincident declines in sockeye salmon returns per spawner at Long Lake (a nearby pristine watershed) and declines in adult sockeye salmon abundance in other populations to the north of Rivers Inlet.