THE INTERMEDIATE DISTURBANCE HYPOTHESIS DOES NOT EXPLAIN FIRE AND DIVERSITY PATTERN IN FYNBOS

The intermediate disturbance hypothesis is a widely accepted generaliz ation regarding patterns of species diversity, but may not hold true w here fire is the disturbance. In the Mediterranean-climate shrublands of South Africa, called fynbos, fire is the most importance disturbanc e and a controlling factor in community dynamics. The intermediate dis turbance hypothesis states that diversity will be highest at sites tha t have had an intermediate frequency of disturbance and will be lower at sites that have experienced very high or very low disturbance frequ encies. Measures of diversity are sensitive to scale; therefore, we co mpared species richness for three fire regimes in South African mounta in fynbos to test the intermediate disturbance hypothesis over differe nt spatial scales from 1 m(2) to 0.1 hectares. Species diversity respo nse to fire frequency was highly scale-dependent, but the relationship between species diversity and disturbance frequency was opposite that predicted by the intermediate disturbance hypothesis. At the largest spatial scales, species diversity was highest at the least frequently burned sites (40 years between fires) and lowest at the sites of moder ate (15 to 26 years between fires) and high fire frequency(alternating four and six year fire cycle). Community heterogeneity, measured both as the slope of the species-area curve for a site and as the mean dis similarity in species composition among subplots within a site, correl ated with species diversity at the largest spatial scales. Community h eterogeneity was highest at the least frequently burned sites and lowe st at the sites that experienced an intermediate fire frequency.