Island area and 2 parameters of environmental diversity could explain c90% of the variation in the number of ant species. Effects of isolation were statistically significant, but added only 3% to the variance already explained by area. Both the equilibrium model, involving parameters on area and distance, and the habitat diversity approach were equally suitable to describe the distribution of ants in this archipelago, which runs parallel to the coast. An alternative sigmoid model, based on the logistic growth equation, fits the data as well as the double log-transformed equilibrium model. The sigmoid model also provided an estimation of the species source pool, which agreed with the size of complete regional ant faunas on the mainland of NW Europe. The presence of most ant species was related to available area of suitable habitat, but some rare species were distributed at random over the islands.-from Authors