Abstract Two hypotheses related to effects of juvenile pink salmon (Oncorhynchus gorbuscha) foraging behaviour and size on their predation risk were evaluated using field data collected in Prince William Sound, Alaska 1995–97. My results supported the hypothesis that low macrozooplankton density leads to dispersion of juvenile salmon from shallow nearshore habitats and greater predation risk, but zooplankton type was an important factor. When the biomass of large copepods (primarily Neocalanus spp.) declined, salmon dispersed from shallow nearshore habitats, and mean daily individual predator consumption of salmon increased by a factor of 5. A concomitant five-fold increase in the probability of occurrence of salmon in predator stomachs supported the notion that increased predation on salmon was caused by a greater overlap between predator and prey when salmon dispersed offshore, not an increase in the number of salmon consumed per feeding bout. The results also generally supported the hypothesis that the timing of predation events modifies the nature of size-dependent predation losses of salmon to different predator groups (small and large planktivores and piscivores). Size-dependent vulnerabilities of salmon to predators were a function of both predator and prey sizes. When simulated predation was shifted from May to June, the vulnerability of salmon became more dependent on their growth than initial size. But, the size- and growth-dependent vulnerabilities of salmon differed more among predator groups than between May and June, suggesting that changes in the composition of predator fields could more strongly affect the nature of size-dependent predation than changes in the timing of predation losses.