The relationship between local and regional species richness in birds of the Caribbean Basin

Abstract

Srivastava (1999) recently evaluated the relationship between local and regional species richness as a tool for determining whether local assemblages of organisms are ecologically saturated or open to invasion. Srivastava also pointed out several pitfalls of sampling and interpretation in such studies. One problem arises when regional species pools are estimated by summing species richness over habitats (Method 1). In this case, the regional pool includes species that may not be capable of living in a particular local community and the species pool is thus overestimated. Terborgh & Faaborg (1980), Ricklefs (1987), and Wiens (1989) used this approach to analyse local and regional richness in the avifauna of the West Indies, where each island was considered a separate region. The local-regional tests in those studies correspond to Srivastava's Method 3, which relates diversity in 'identical habitats in geographically different regions' to the regional species pools. One may nonetheless object to these analyses because regional species richness was estimated by summing over habitats within islands and because individual islands within an archipelago are not statistically independent, i.e. 'regional', samples. The problem is largely one of scale. Loreau (2000) has pointed out that the relationship between local and regional species richness depends on the scale at which each is measured. Thus, when the concepts of 'local' and 'regional' are similar, for example in terms of habitat breadth, a large proportion of the regional heterogeneity is sampled within each locality and local diversity tends to vary in proportion to regional diversity. This gives the appearance of non-saturation (Cornell & Lawton 1992; Caley & Schluter 1997). When a narrower concept of 'local' is applied, local diversity is relatively independent of regional diversity and the appearance of saturation is a more likely outcome. In Pearson's (1977) study of birds in tropical forests in six regions, concepts of 'local' and 'regional' did not differ greatly with respect to habitat breadth. Consequently, local communities contained