ABSTRACT: Although euphausiids of the genus Thysanoessa (‘krill’) are a key zooplankton taxon in the Bering Sea ecosystem, the processes controlling variation in the standing stock of these animals are not well understood. Both forcing by temperature (‘bottom-up’) and predation (‘top-down’) have been proposed. If strong top-down forcing were present, a negative relationship would be expected between the standing stock of euphausiids and that of walleye pollock Gadus chalcogrammus, their single most important predator in the Bering Sea. We developed multiple regression models using survey data collected over several years to test the hypothesis of a negative relationship between local (mesoscale) densities of these animals. We used the models to evaluate the relative importance of pollock biomass and water temperature in predicting euphausiid biomass, and found that temperature was a far better predictor, a result that is not consistent with dominant top-down control of euphausiids by pollock predation.