In the last decades, marine reserves have dramatically increased in number worldwide. Here I examined the potential of no-take marine reserves to reestablish lost predatory interactions and, in turn, cause community-wide changes in Mediterranean rocky reefs. Protected locations supported higher density and size of the most effective fish preying on sea urchins (the sea breams Diplodus sargus and D. vulgaris) than unprotected locations. Density of sea urchins (Paracentrotus lividus and Arbacia lixula) was lower at protected than at unprotected locations. Size structure of P. lividus was bimodal (a symptom of predation on medium-sized urchins) only at the protected locations. Coralline barrens were less extended at protected than at unprotected locations, whereas turf-forming and erect-branched algae showed an opposite pattern. Erect-unbranched and erect-calcified algae and conspicuous zoobenthic organisms did not show any pattern related to protection. Tethering experiments showed that predation impact on urchins was (1) higher at protected than at unprotected locations, (2) higher on P. lividus than on A. lixula, and (3) higher on medium-sized (2-3.5 cm test diameter) than large-sized (.3.5 cm) urchins. Sea urchins preyed on by fish in natural conditions were smaller at unprotected than at protected locations. The analysis of sea urchin remains found in Diplodus fish stomachs revealed that medium-sized P. lividus were the most frequently preyed upon urchins and that size range of consumed sea urchins expanded with increasing size of Diplodus fish. These results suggest that (1) depletion and size reduction of predatory fish caused by fishing alter patterns of predation on sea urchins, and that (2) fishing bans (e.g., within no-take marine reserves) may reestablish lost interactions among strongly interactive species in temperate rocky reefs with potential community-wide effects.