Concentrations of three trace gases-H2, CO, and CH4-were measured in the upper waters and in the near-surface atmosphere at the Controlled Ecosystem Population Experiment (CEPEX) site in Saanich Inlet, B. C., Canada. The upper 20 m of water contained these reduced gases at concentrations significantly higher than atmospheric equilibrium values in the presence of high levels of dissolved oxygen. Because vertical mixing was quite weak within the CEPEX container, distributions of CO, CH4, and H2 were most strongly influenced by in situ production and comsumption. Concentrations of hydrogen and carbon monoxide were generally highest at or near the surface in all containers. Daily cycles of CO concentrations in the CEPEX enclosures and in the water outside the enclosures were observed. A similar, but less pronounced, H2 cycle was seen only inside the enclosure. In the CEPEX containers that received no nutrient additions, a methane maximum was commonly found at depths between 8 and 12 m. Methane concentrations and distributions in this container showed little change over a period of several weeks. Methane distribution did undergo significant change in the nutrient-enriched containers. In Saanich Inlet, the CH4 maxima occurred deeper in the water column. No excess of H2 or CO was found in this methane-rich deeper water. Within the containers, changing hydrogen concentrations seem related to phytoplankton abundance, but the hydrogen distribution suggests zooplankton play a role. Carbon monoxide abundance varies with the amount of illumination received by the waters. Probably a variety of factors influence the distribution of each of the gases. Seawater samples collected at the site and incubated in dark containers, at in situ temperature, showed rapid decrease of CO and H2 levels to below air equilibrium. Carbon monoxide in these incubated samples reached subequilibrium levels within 12 hours. Shallow and deep water samples that had been briefly exposed to sunlight during sample handling subsequently showed higher levels of dissolved CO than did samples that were carefully shielded from outside light. Care is needed to avoid such artifacts when studying diurnal variations in dissolved CO.