Tropical reefs are in global decline with seaweeds commonly replacing corals. Negative associations between macroalgae and corals are well documented, but the mechanisms involved, the dynamics of the interactions, and variance in effects of different macroalgal-coral pairings are poorly investigated. We assessed the frequency, magnitude, and dynamics of macroalgal-coral competition involving allelopathic and non-allelopathic macroalgae on three, spatially grouped pairs of no-take Marine Protected Areas (MPAs) and non-MPAs in Fiji. In non-MPAs, biomass of herbivorous fishes was 70-80% lower, macroalgal cover 4-9 fold higher, macroalgal-coral contacts 5-15 fold more frequent and 23-67 fold more extensive (measured as % of colony margin contacted by macroalgae), and coral cover 51-68% lower than in MPAs. Coral contacts with allelopathic macroalgae occurred less frequently than expected by chance across all sites, while contact with non-allelopathic macroalgae tended to occur more frequently than expected. Transplants of allelopathic macroalgae (Chlorodesmis fastigiata and Galaxaura filamentosa) against coral edges inflicted damage to Acropora aspera and Pocillopora damicornis more rapidly and extensively than to Porites cylindrica and Porites lobata, which appeared more resistant to these macroalgae. Montipora digitata experienced intermediate damage. Extent of damage from macroalgal contact was independent of coral colony size for each of the 10 macroalgal-coral pairings we established. When natural contacts with Galaxaura filamentosa were removed in the field, recovery was rapid for Porites lobata, but Pocillopora damicornis did not recover and damage continued to expand. As macroalgae increase on overfished tropical reefs, allelopathy could produce feedbacks that suppress coral resilience, prevent coral recovery, and promote the stability of algal beds in habitats previously available to corals.