Downstream migrant traps are a widely applied fishery management tool for sampling anadromous Pacific salmon Oncorhynchus spp. and steelhead O. mykiss smolts along theWest Coast of North America and elsewhere, yet predation on juvenile salmonids in traps has not been studied quantitatively.We assessed the frequency of occurrence and abundance of juvenile salmonids in the stomachs of coastal cutthroat trout O. clarkii clarkii, coho salmon O. kisutch, steelhead, and prickly sculpin Cottus asper (>70 mm fork length) captured in traps and in nearby stream habitats. All four predator species took juvenile salmonids with much greater frequency in traps than in stream habitats. Among free-swimming predators, only coastal cutthroat trout were observed with salmonid fry in their stomachs, but they took fewer salmonid prey and appeared to rely more heavily on insect prey than did coastal cutthroat trout captured in traps. Predators consumed up to 25% of the available prey over a broad range of prey abundances. Over the course of the study, predators consumed 2.5% of all salmonid fry captured in traps, but this fraction ranged from less than 1% to more than 10% in any given year. The number of prey taken in traps increased with predator length and with prey abundance in traps, and predation in traps peaked during the period of most intense downstream migration by salmon fry. In contrast, live-box design and trap location had little or no effect on the total number of prey taken by individual predators.We estimated that the predation mortality of juvenile salmon increased by 0.5-1.0% due to in-trap predation (i.e., a 9-10% relative increase over natural predation rates). We found no evidence that predators selected for prey on the basis of species. These results should motivate additional research on methods that reduce or eliminate predation in trap live-boxes and protocols for efficiently measuring predation associated with the trapping of downstream migrants. ?? American Fisheries Society 2011.