Ecologists modeling predator reintroduction have typically assumed that restoring an apex predator to a landscape will, through trophic cascade, suppress herbivore populations enough to allow suppressed vegetation to recover, as occurred following wolf reintroduction in one widely cited river valley. Long-term monitoring of several other reintroduction sites complicates this template: herbivore populations in these sites declined only modestly, yet vegetation recovered nearly as strongly as at the celebrated site. The explanation appears to lie not in reduced herbivore numbers but in altered herbivore behavior—grazing animals, sensing predation risk, avoided browsing in exposed river corridors and forest edges regardless of how many of them the landscape still held, concentrating their feeding instead in cover where they were harder to ambush. Vegetation recovery, in other words, tracked a landscape of fear reshaping where herbivores chose to feed, not a body count reshaping how many herbivores there were to feed at all. Where ecologists had folded both mechanisms into a single population-suppression story, the newer sites reveal that behavioral avoidance can drive the same outcome largely on its own.