We describe the migratory behavior and physiological development of subyearling fall chinook salmon Oncorhynchus tshawytscha migrating through John Day Reservoir on the Columbia River, Washington and Oregon. Fish were freeze-branded and coded-wire-tagged at McNary Dam, Oregon, from 1991 to 1994, to determine travel time to John Day Dam and subsequent adult contribution. Stepwise multiple regression showed that 47% of the variation in subyearling fall chinook salmon travel time was explained by the reciprocal of minimum flow and fish size. Smoltification, as measured by gill Na+-K+ adenosine triphosphatase (ATPase) activity, was not important in explaining variability in travel time of subyearling chinook salmon. Fish marked early in the out-migration generally traveled faster than middle and late migrants. Seawater challenges were used to describe physiological development and showed that osmoregulatory competence of premigrants in the Hanford Reach of the Columbia River increased with fish size and gill ATPase activity. Once active migrants began passing McNary Dam, fish generally had survival exceeding 90% and were able to regulate their blood plasma Na+ in seawater. Gill ATPase activity increased as premigrants, reared in nearshore areas of the Hanford Reach, reached a peak among active migrants in late June and early July then decreased through the remainder of the out-migration. Salinity preference also peaked in subyearling fall chinook salmon during late June to mid July in 1995. Return of adults from marked groups showed no consistent patterns that would suggest a survival advantage for any portion of the juvenile out-migration. Presumed wild migrants from the middle and late portions of the out-migration were primary contributors to all fisheries, except the Priest Rapids Hatchery. As such, fishery managers should take action to ensure the survival of these fish, especially because they migrate under more unfavorable environmental conditions than early migrants.