In diverse tropical webs, trophic cascades are presumed to be rare, as species interactions may dampen top-down control and reduce their prevalence. To test this hypothesis, we used an open experimental design in the Galápagos rocky subtidal that enabled a diverse guild of fish species, in the presence of each other and top predators (sea lions and sharks), to attack two species of sea urchins grazing on benthic algae. Time-lapse photography of experiments on natural and experimental substrates revealed strong species identity effects: only two predator species–blunthead triggerfish (Pseudobalistes naufragium) and finescale triggerfish (Balistes polylepis)–drove a diurnal trophic cascade extending to algae, and they preferred large pencil urchins (Eucidaris galapagensis) over green urchins (Lytechinus semituberculatus). Triggerfish predation effects were strong, causing a 24-fold reduction of pencil urchin densities during the initial 21 hours of a trophic cascade experiment. A trophic cascade was demonstrated for pencil urchins, but not for green urchins, by significantly higher percent cover of urchin-grazed algae in cages that excluded predatory fish than in predator access (fence) treatments. Pencil urchins were more abundant at night when triggerfish were absent, suggesting that this species persists by exploiting a nocturnal predation refuge. Time-series of pencil urchin survivorship further demonstrated per capita interference effects of hogfish and top predators. These interference effects respectively weakened and extended the trophic cascade to a fourth trophic level through behavioral modifications of the triggerfish-urchin interaction. We conclude that interference behaviors capable of modifying interaction strength warrant greater attention as mechanisms for altering top-down control, particularly in speciose food webs.