Ecology. Crafting the pieces of the diversity jigsaw puzzle.

A fter his encounter with the Atlantic rainforest of Brazil, Charles Darwin observed that “the land is one great, wild, untidy, luxuriant hothouse, made by Nature for herself ” (1). Increasing knowledge of the forest fauna over the past 30 years has only underscored Darwin’s point: 978 species of beetles from six trees in Venezuela, 160 species of grasshoppers in 60 trees in Amazonia, 218 species of ants from 19 trees in Sabah—the list goes on (2). Dobzhansky, with characteristic prescience, asked “why?” as early as 1950 (3). He suggested that the animals of the Tropics were fundamentally different from those of temperate regions: In general, tropical species were more specialized, with niche limits defined by interactions among species rather than by physical or chemical factors. This has proved very difficult to test; indeed, it moved surreptitiously from hypothesis to dogma without any very critical evaluation. This difficulty is nowhere more acute than in the case of tropical herbivores. Two reports in this issue, one by Novotny et al. (4) on page 1115 and the other by Wilf et al. (5) on page 1112 shed light on both old and new aspects of these issues. The relationships between the diversity of tropical trees and that of the associated arthropods have been central in the exciting if somewhat circular debates on global species richness. It is relatively easy to count the number of tree species in a forest and to extrapolate this count to a region, continent, or even planet. If each tree species supports a certain number of herbivorous insects, then we can extrapolate to a global estimate of herbivore richness. With a few further assumptions, we can obtain an overall number of all arthropod species. That the estimates range from 5 to 100 million (6) suggests that this is not an exact science. The Novotny/Basset group based in Papua New Guinea has been one of only two or three groups that have measured tropical host specificity empirically by field collecting, larval rearing, and choice experiments. Based on Herbivorous insect species are more numerous in tropical than in temperate forests. Studies of present-day forests and fossil leaf patterns show how plant diversity controls insect diversity. Crafting the Pieces of the Diversity Jigsaw Puzzle