The ant faunas of three remote Polynesian islands were censused using hand collecting techniques. Known ant species richnesses were increased by factors of 2.3, 3.7, and 4.3, and total species richnesses were estimated with a first-order jackknife estimator. The large increase in species numbers is apparently due to inadequate earlier censuses (which trussed localized and cryptic species) rather than recent immigrations. Tests of species associations revealed more positive than negative interactions among species, on both a pairwise and community-wide basis. There is no evidence that ant species on these islands exclude each other from islands or from communities within islands, with the exception of three very aggressive species. A multiple regression analysis of known ant species richness against sampling effort and area for Polynesian islands which have been differentially censused for ants by various collectors revealed sampling effort was highly significant, while area was not significant, in explaining variation in known ant species numbers. On Pacific islands that have been surveyed relatively thoroughly for ants, multiple regression analyses of known ant species richness on area and distance showed that area was always highly significant, but distance was only marginally significant (depending on the regression model used). Thus remote Polynesian islands appear neither to be as depauperate as previously thought in numbers of ant species present, nor possess an unusual potential for evolutionary increase in species numbers.

Document Type


Stable URL


Rights Information

© 1996 Ecography

Publication Date


Journal Title