The abundance of the 1982 brood of juvenile coho salmon (Oncorhynchus kisutch) was determined in August 1983, and January and April 1984 at 20 study sites spread throughout Knowles Creek, an Oregon coastal watershed. The timing of emigration of juvenile coho from the watershed was monitored from October 1983 through June 1984. Condition factor, fork length, and gill (Na+K)-ATPase activity were measured in migrants, a captive group of Knowles Creek juvenile coho held in the laboratory, and nonmigrant fish periodically sampled from the stream. Skin guanine levels were also measured in migrant and nonmigrant groups. Juvenile coho abundance in January was significantly correlated with abundance in August. Wood volume and amount of undercut streambank were the pair of physical variables that best explained variation in the number of fish per square meter or per cubic meter in January. Two debris torrent ponds in the middle of the watershed contained large amounts of woody debris and were the most heavily used overwintering habitats for juvenile coho in the Knowles Creek. Few juvenile coho overwintered in the lower half of watershed, an area lacking woody debris. Peaks in outmigration occurred in November and May. Approximately 24% of the total number of migrants emigrated in November. Fish that reared in two of three third-order areas in summer, together with fish from the lower (fifth-order) half of the mainstem, were the first to leave the watershed. While lack of winter habitat may have been the cause of migration from the lower mainstem, low summer streamflows may have caused early migration from the low order sites. Gill (Na+K)-ATPase activity of migrants rose gradually from a low in January to a peak at the end of the study in June. Mean gill (Na+K)-ATPase activity of nonmigrants was only significantly lower than that of migrant fish during April. Gill (Na+K)-ATPase of captives was similar to that of nonmigrants until it peaked during the last two weeks in April, after which the activity fell below that of migrants or nonmigrants. Condition factor of nonmigrant fish was higher than either migrants or captives throughout the study. Migrant skin guanine levels rose sharply during the first two weeks in April and continued to rise until the end of the study in June. Approximately 8,300 juvenile coho, 44% of the estimated number of juvenile coho present in Knowles Creek in August, migrated from the watershed by the following June. An estimated 9% of the August population migrated as smolts after April 1. / Graduation date: 1986

from the summer dry season to the winter wet season. Such movement that connects summer and winter habitats may be particularly important for coho salmon, O. kisutch, because availability of overwintering habitat can limit freshwater survival for this species. Here, I describe basin-scale variability in the spatial pattern of fall movement for juvenile coho salmon between mainstem and tributary streams during the fall of 2002, 2003, 2004, and 2005. Juvenile coho salmon were tagged with a passive integrated transponder (PIT) and could be detected at five stationary detection sites, two located in perennial tributaries, two in intermittent tributaries, and one in the upper mainstem of the West Fork Smith River, Oregon. For each detection site, I compare the likelihood of detection during the fall by juvenile coho salmon from tagging locations over a multi-kilometer range of distances in each direction away from the tributary confluence. I developed logistic regression models with data from each detection site to estimate: 1) the relative likelihood of immigration into a tributary as compared to emigration out of the tributary, and 2) the relative likelihood of immigration into a tributary from the mainstem downstream of the tributary confluence as compared to immigration from the mainstem upstream of the confluence. For each pair of directions at each detection site, I also compare the change in the likelihood of detection with increasing distance for each direction. Overall, at the two upper-river detection sites, juvenile coho salmon were more likely to emigrate than to immigrate. At the remaining detection sites, juvenile coho salmon were no more likely to emigrate than immigrate. Of these detection sites, fish that immigrated into the mid-river perennial stream were more likely to come from the mainstem downstream of the confluence, whereas fish that immigrated into the two lower-river intermittent tributaries were more likely to come from the mainstem upstream of the confluence. Fall movement of juvenile coho salmon between tributary and mainstem habitat can occur over relatively long distances. This case study demonstrates variation among tributaries in the overall likelihood of emigration and immigration and in the source of immigrants from the mainstem, which may be related to spatial context that combines the physical characteristics and network position of tributary streams. The demonstrated variation in fall movement that connects summer and winter habitat within a stream network is a first step in exploring how complexity in movement interacts with the spatial arrangement and quality of seasonal habitats. More research on the causes of variation in the expression of fall movement will improve our understanding of how the spatial arrangement of habitat within a stream network influences the survival of juvenile coho salmon over the whole freshwater life cycle. / Graduation date: 2013