Four brood years of juvenile spring chinook salmon Oncorhynchus tshawytscha were reared in conventional and baffled raceways at various rearing densities and loads at Willamette Hatchery, Oregon. A period of rapid linear growth occurred from August to November, but there was little or no growth from November to March when the fish were released. Both fall and winter growth rates were inversely related to rearing density. Final weight and length were also inversely related to rearing density. No significant relationship between load and any growth variable was observed. Fish reared at lower densities in conventional raceways tended to develop bimodal length distributions in winter and early spring. Fish reared in conventional raceways showed significantly larger growth rates and final lengths and weights than those reared in baffled raceways. Food conversions and average delivery times for feed were significantly greater in baffled than in conventional raceways. No significant relationships were observed between either rearing density or load and condition factor, food conversion, or mortality. Mortality was not significantly different between the two raceway types. When fish were transported to seawater for further rearing, there were no significant relationships between mortality in seawater and rearing density or load, but fish reared in baffled raceways had significantly higher mortality than those reared in conventional raceways.