PATTERNS AND PREDICTION OF ALPHA-DIVERSITY AND BETA-DIVERSITY OF AQUATIC PLANTS IN ADIRONDACK (NEW-YORK) LAKES

Patterns in alpha diversity (species richness) of submersed and floati ng-leafed aquatic plants were investigated for 45 lakes in the Adirond ack region of northern New York State. Multiple linear regressions wer e used to build predictive models of species richness; the independent variables included lake surface area, approximate littoral zone area, pH, and measures of isolation and connectedness. The regression model s that explained the most variance in species richness incorporated pH , a connectedness predictor, and either lake surface area or littoral zone area (r(2) = 0.74). The two area measures accounted for a majorit y of the variance in species richness (about 57%), but neither was a s uperior predictor to the other. When acidic lakes were systematically removed from the multiple regression model, r(2) tended to increase. T he coefficient of determination was maximized (at r(2) = 0.822) when w e limited the data set to lakes with pH > 5.6. As there was no signifi cant correlation between pH and any of the isolation or connectedness measures we used (maximum r = 0.18), acidic lakes are not more isolate d than circumneutral lakes. Point beta diversity (the heterogeneity of species composition among sites) was measured for each pair of lakes. It was not strongly correlated to either the difference in surface ar ea or the difference in pH between pairs of lakes. Standardized nested ness was calculated following Wright and Reeves (1992, Oecologia 92: 4 16-428). Adirondack lakes are moderately nested. An implication of thi s is that species that are found in less diverse lakes have a 44% prob ability of being found in more diverse lakes.