A study of the population ecology of Columbia River fall chinook salmon, Oncorhynchus tshawytscha (Walbaum), was made in an attempt to determine the cause of a serious decline in this run which occurred in the early 1950's. Fluctuations in abundance of major salmon runs the North Pacific were examined to detect any coastwide pattern. Only chinook salmon in Cook Inlet, Alaska, and chum salmon from Oregon to southwestern Alaska showed a similar trend. The following life history stages broken down into pre- and post-decline years were examined: (1) marine life including distribution and migration, growth and maturity, survival rate, oceanography, and commercial and sport fisheries; (2) upstream migration including river fisheries, gear selectivity, size and age composition of the run, escapement, and influence of dams, diseases, and water quality; (3) reproduction and incubation including spawning areas and spawning and incubation conditions; and (4) downstream migration which included predation, dams and reservoirs, diseases, flow, turbidity and temperature, and estuary life. Salient points of the analysis were: (1) a change in the maturity and survival pattern based on tagged and fin-clipped fish recovered before and after 1950; (2) a significant negative correlation between sea-water temperature during a year class' first year at sea and subsequent survival; (3) a large increase in the ocean fisheries coincident with the decline in the run; (4) catch-effort statistics of the ocean fishery show a near classic example of the effect of overexploitation; (5) estimates of the contribution of Columbia River chinook to the ocean fisheries based on tag recoveries could be underestimates rather than overestimates; (6) a significant inverse correlation between estimated ocean catch of Columbia River fall chinook and numbers entering the river; (7) size and age composition of the ocean and river catches decreased coincident with the decline in the run; (8) the gill-net fishery shows little size selectivity by age, size, or sex in the dominant group; (9) fluctuations in abundance of hatchery stocks are related to differences in survival between fingerling and adult; (10) hatchery, lower river, and upriver populations fluctuate in abundance in much the same pattern; (11) optimum escapement is between 90,000 and 100,000 adults, a value that was exceeded during most years; (12) a highly significant negative correlation between numbers of spawners and return per spawner; (13) most of the early dams had no direct effect on fall chinook and the decline in productivity occurred when river conditions were relatively stable; (14) temperatures at time of migration and spawning for fall chinook have not increased enough to be a serious mortality factor; (15) little relationship between flow, turbidity, and temperature at time of downstream migration and subsequent return was evident except that high temperatures and high flows (and turbidities) tended to produce poorer runs during certain time periods; and (16) predation and delay of smolts in reservoirs are largely unknown factors, but circumstantial evidence suggests that they were not important in regulating fall chinook numbers during the period of the study. Finally, variables that appeared to bear some relationship to fluctuations in abundance of fall chinook were submitted to multiple regression analysis. For the predecline period (1938-46 brood years), sea-water temperature and ocean troll fishing effort were significant variables (Rp2s = 0.74). For post decline years (1947-59 broods), troll had the most influence on total return with ocean temperature and escapement having lesser effects. For the combined years, troll intensity and ocean temperature were the significant variables (Rp2s = 0.572). Entering interaction of river flow at downstream migration with the other variables brought Rp2s to 0.754 which means that 75% of the variability in the returning run could be accounted for by these three factors. Return per spawner was so heavily influenced by numbers of spawners that the other factors assumed negligible importance. Equations were derived that predicted the returning run in close agreement with the actual run size. Substituting a low and constant troll fishing effort in the equation resulted in the predicted run maintaining the average predecline level. The increase in ocean fishing was the main contributor to the decline of the Columbia River fall chinook run as shown by correlation, by analogy, and by the process of elimination. To demonstrate why other chinook runs have not shown similar declines, it was shown that due to several unique features in Columbia River fall chinook life history they are exposed to much more ocean fishing than other populations. It was emphasized that these conclusions should not be extrapolated to the future or to other species or runs of salmon.