Image: Courtesy of JONATHAN B. LOSOS

Naturalists have long recognized that large islands typically house more species than do comparable, smaller islands. More than three decades ago, E.O. Wilson of Harvard University and the late Robert MacArthur of Princeton University proposed that this relationship between an island's size and the number of species it contains reflects the balance between its immigration and extinction rates, the so-called equilibrium theory of island biogeography. What they didn't know, owing to a lack of data, was how importantly evolution can figure in this equation if the island is large. New research, published today in the journal Nature, suggests that "in-house" speciation contributes more species to the sum total on the island than does immigration--but only if the island is big enough.

To test this hypothesis, biologists Jonathan B. Losos of Washington University and Dolph Schluter of the University of British Columbia studied 143 species of Anolis lizards (right) on 147 islands in the Caribbean, concentrating on the four largest--Cuba, Hispaniola, Jamaica and Puerto Rico. Using a DNA-based family tree of the lizards, they estimated the numbers of immigration and speciation events that had taken place on the islands. Comparing these figures with island area, the researchers confirmed that, as predicted, larger islands witnessed more speciation events. "At some level this is intuitive," Losos remarks, "but it has never been demonstrated before that differences in the rate of speciation, of evolution, can produce the species-area relationship."

Intriguingly, though, Losos and Schluter only found within-island speciation on islands larger than 3,000 square kilometers. "We don't know why the threshold is there," Losos admits. "The islands of Guadeloupe and Martinique are quite large and vegetationally diverse; there are plenty of habitats for lizards to exploit, and yet speciation has not occurred there." And even among the four largest islands, they note, Puerto Rico has experienced surprisingly few speciation events, considering its history and topographic diversity. Perhaps similar studies of other species in other regions will someday explain the pattern.