There has been much discussion and some controversy regarding the role of nutrient enrichment versus other factors, such as altered rates of herbivory, in the degradation of coral reef ecosystems. The resolution of this controversy has been hampered by the lack of manipulative field studies testing the effects of ecologically relevant levels of nutrient enrichment on coral reef communities. We present a new method for adding ecologically relevant levels of nutrients to experimental substrates on coral reefs. The method elevates nutrients for sustained periods of time (>41 d without replenishment of nutrients) and allows for testing interactions of nutrients and altered levels of herbivory. Results from an offshore reef in Key Largo, Florida, show strong effects of excluding large herbivorous fishes, negligible effects of nutrient enrichment–;or effects that are opposite of predictions, and no interaction between nutrient levels and herbivory in affecting algal abundance. Patterns observed for this reef did not confirm predictions of previously proposed models that frondose macroalgal or crustose algal abundance would be enhanced with nutrient enrichment or that dominance of filamentous turfs would be greater in unenriched conditions. In contrast to previous predictions, the abundance of larger macroalgae at this site was not increased by elevating nutrients above predicted threshold response levels of 1.0 µM for total inorganic nitrogen or 0.10 µM for soluble reactive phosphate. Also conflicting with some models, filamentous, nitrogen-fixing cyanobacteria were enhanced, rather than suppressed, by nutrient enrichment.