Release from herbivory is a factor underpinning coral replacement by macroalgae. Once macroalgae have achieved dominance, shifts back to the coral-dominated state can be hindered by stabilizing feedbacks. Thus, restoring herbivore assemblages alone can be insufficient to trigger coral recovery. However, herbivores could control macroalgal recovery in the aftermath of physical disturbances removing macroalgae. Diadematid urchins at Moorea (French Polynesia) have collapsed in the last decade. By means of a manipulative field experiment, we tested the interactive effects of physical disturbance and increased diadematid densities on macroalgae inside the lagoon. Massive Porites colonies, referred to as ‘bommies’, were assigned to 3 different macroalgal removal treatments (removal of stipes and fronds of the canopy-forming macroalgae Turbinaria ornata and Sargassum pacificum , total removal of erect macroalgae or untouched) and exposed to 3 different urchin densities (absent, low [~0.5 ind. m -2 ], and intermediate [~1 ind. m -2 ]). After 1 yr, sea urchins had no effect on the covers of S. pacificum and T. ornata when macroalgal canopies were left untouched. Urchins could control the recovery of S. pacificum on total macroalgal removal bommies, but not that of T. ornata . However, urchins, even when at intermediate densities, did not generate major changes in the structure of benthic assemblages on experimental bommies. Our study indicates that a moderate increase in diadematid densities is unlikely to reduce the extent of macroalgal stands in Moorea back reefs unless associated with the recovery of other herbivore guilds able to remove adult macroalgae (i.e. browsers).