Larger refuges or islands generally will preserve more species than a series of small refuges of equivalent total area. The claim of Simberloff and Abele (1976) that several small refuges may contain more species than a single large refuge is valid only for islands which contain a very small fragment of the total available species pool. Such areas are inappropriate as permanent wildlife refuges. A system of small tracts of land, such as a series of small urban parks, may be designed specifically to preserve only a small fraction of the total fauna in each park but a large diversity among parks. Two small islands may contain more species than a single island of equivalent area as the result of one or both of 2 effects: a sampling phenomenon or a faunal exchange with a mainland source pool. Both effects are reduced with increasing island size and as a result of realistic assumptions about the relative colonizing abilities of species. Possibly, the results obtained by Simberloff and Abele are not the result of taxonomic idiosyncracy of arthropods; their findings reflect that the archipelago they studied contained only a small fraction of the total mangrove arthropod fauna.