Computing β-diversity from species-area curves

A measure of beta-diversity is proposed for equal-sized presence-absence vegetation sample plots. The method begins by computing the minimum variance unbiased estimator of the expected species richness E[S-n] as a function of sample size n for a set of N plots. Using a semilogarithmic relation, beta-diversity is then determined from the slope of the best fitting line of the E[S-n] vs. log n plot. beta-diversity is thus interpretable as the linear relation between expected species richness and log sample size. The method is illustrated using 15 square plots of 100 m(2) each from a Matorral community in the Arid Chaco of central Argentina. Further, some analogies between the proposed method and Whittaker's multiplicative model of beta-diversity are discussed.