THE DECLINE OF TREE DIVERSITY ON NEWLY ISOLATED TROPICAL ISLANDS - A TEST OF A NULL HYPOTHESIS AND SOME IMPLICATIONS

Six islands, each less than a hectare in area, were isolated in about 1913 from the mainland of central Panama by the rising waters of Gatun Lake. By 1980, the diversity of trees on all but one of these islands was far lower than on mainland plots of comparable size. A restricted subset of tree species has spread on these islands, notably Protium panamense, Scheelea zonensis, Oenocarpus panamanus and Swartzia simplex. We constructed a null model to predict how chance would change tree diversity and the similarity of tree species compositions of different islands, assuming that each mature tree has equal chances of dying and/or reproducing, regardless of its species. This model cannot account for the diminished diversity of the changes in vegetation on these islands: some factors must be favoring a particular set of tree species. Two factors, exposure to wind and absence of mammals, seem needed to bring about the vegetation changes observed on these small islands. Their vegetation shows many signs of wind damage and of adaptation to resist wind, reflecting its exposure to dry season winds and storm winds sweeping across the lake from the west. Their most common tree species appear to have spread because mammals rarely visit these small and isolated islands. Seed of these common species are normally much eaten by mammals and do not need burial by mammals to escape insect attack. A thorough grasp of plant-animal interactions is needed to understand the events that have taken place on these islands. Identifying those 'keystone animals' essential for maintaining plant diversity is a necessary element of reserve design and forest management in the tropics.