Professor Mark Hay and co-author Deron Burkepile of the Georgia Institute of Technology constructed 32 cages on a coral reef at Key Largo, Fla. Then into each 140-cubic-foot cage, Hay and Burkepile placed different numbers of two types of herbivorous fish.

During the next 10 months they measured changes in coral cover and seaweed growth. The scientists explained seaweeds compete with coral reefs, and are eaten by a variety of fish. In the absence of fish, reefs decline as corals are replaced by seaweeds.

"For the cages in which we mixed the two species of herbivores, the fish were able to remove much more of the upright seaweeds, and the corals in those areas increased in cover by more than 20 percent during the 10 months," said Hay. "That is a dramatic rate of increase for a Caribbean reef."

Corals caged with just one type of fish or no fish lost up to 30 percent of their cover during the same period.

The study appears in the Proceedings of the National Academy of Sciences.