Sixteen years of annual censuses (1990-2005) of an assemblage of tide pool sculpins at Tatoosh Island, Washington, USA, revealed relative constancy in the numbers of recruits and adults. However, the dominant species as a recruit (Clinocottus globiceps) was always replaced numerically as an adult by another species (Oligocottus maculosus). When mechanisms of coexistence were evaluated, little support was found for either a lottery model, because competitive interactions are hierarchical, or a storage-effect model of coexistence, because the relative ranking of recruitment varied little among years and did not covary among species with environmental variables such as ocean temperature, upwelling indices, or an estimator of ocean productivity (oyster condition index). There was also little evidence of niche partitioning based on habitat affinities. Additionally, predation-mediated coexistence had little support, given that the competitive dominant did not have the greatest rates of mortality. Instead, a competition-colonization trade-off may contribute to the coexistence of these species, where C. globiceps always recruits in the greatest numbers, while O. maculosus dominates the adult assemblage via competitive processes. The importance of post-recruitment processes in this assemblage is further suggested by some negative associations among species, the presence of density dependence, some habitat affinities, and previously published experimental work that demonstrated the competitive dominance of O. maculosus over C globiceps.