We studied the persistence of fishing impacts on coral reef fish assemblages by sam- pling 2 protected and 2 fished locations, using random replicate stereo baited remote underwater video. At each location we sampled a variety of coral reef habitats on 4 separate occasions be - tween 2006 and 2007. We tested for consistency in differences in the biomass of target and non- target species, trophic groups and overall assemblages. Generally, target species were more abundant and/or larger at protected locations. Many non- target species were either more abun- dant or depleted at protected sites, and some of these species were significantly larger or smaller. Trophic groups such as piscivores, piscivore invertivores, invertivores and planktivores were con- sistently more abundant within protected locations. Generally, greater numbers of species and individuals were found on protected reefs, though this was not consistent in all cases across all 4 sampling periods. These findings are consistent with the theory that protected areas can increase the abundance of not just target species but overall fish assemblages. These findings provide evi- dence for ecological mechanisms such as predator-prey interactions, competitive release, and benefits to invertivores, omnivores and other non-target groups from the presence of more abun- dant and larger target species within marine protected areas.