Predation is a major source of mortality in the early life stages of fishes and a driving force in shaping fish populations. Theoretical, modeling, and laboratory studies have generated hypotheses that larval fish size, age, growth rate, and development rate affect their susceptibility to predation. Empirical data on predator selection in the wild are challenging to obtain, and most selective mortality studies must repeatedly sample populations of survivors to indirectly examine survivorship. While valuable on a population scale, these approaches can obscure selection by particular predators. In May 2018, along the coast of Washington, USA, we simultaneously collected juvenile quillback rockfish Sebastes maliger from both the environment and the stomachs of juvenile coho salmon Oncorhynchus kisutch. We used otolith microstructure analysis to examine whether juvenile coho salmon were age-, size-, and/or growth-selective predators of juvenile quillback rockfish. Our results indicate that juvenile rockfish consumed by salmon were significantly smaller, slower growing at capture, and younger than surviving (unconsumed) juvenile rockfish, providing direct evidence that juvenile coho salmon are selective predators on juvenile quillback rockfish. These differences in early life history traits between consumed and surviving rockfish are related to timing of parturition and the environmental conditions larval rockfish experienced, suggesting that maternal effects may substantially influence survival at this stage. Our results demonstrate that variability in timing of parturition and sea surface temperature leads to tradeoffs in early life history traits between growth in the larval stage and survival when encountering predators in the pelagic juvenile stage.