Do the geographical range sizes of species decline from the poles to the tropics, as Rapoport's rule suggests? And if so, is the smaller size of species' ranges in the tropics responsible for the greater species richness observed there? With biodiver-sity on the decline, these questions have much urgency. To minimize the loss of species, there is a pressing need to understand what's behind observed patterns of diversity.
When George Stevens formulated the rule he named after Rapoport in 1989 (American Naturalist, vol. 133), he hypothesized that the latitudinal gradient in species richness might be explained by a latitudinal gradient in range size. He noted the species that were exceptions to the rule, which has been controversial from the outset, and argued that because the same species are also exceptions to the latitudinal gradient in species richness, they helped demonstrate the link between range size and diversity. Until now, though, few datasets have been complete enough to test these ideas.
A new study, published in the July issue of PLoS Biology, maps the global distribution of an entire taxonomic group—birds—and shows that the spatial variation in geographic ranges concerns more than a simple latitudinal gradient. And the pattern that emerges does not coincide with that of species richness.
The authors of the study, led by David Orme of the Imperial College London, found that most of the 9505 bird species studied have smaller than expected range areas, with the smallest occurring in three regions: islands, tropical and subtropical mountains, and the Southern Hemisphere. Stevens argued that it is the latitudinal extent of a species' geographic range that is the key component of range size, but the PLos Biology study did not find a decline in latitudinal extent of ranges toward the tropics, either.
The more general trend is for range areas to decline gradually from north to south, with a dip at the equator, indicating that Rapoport's rule applies to species in the Northern Hemisphere but not to those in the Southern Hemisphere. Species richness, on the other hand, does follow a latitudinal pattern, with the highest diversity found in the tropics, as expected, as well as at higher altitudes, such as peaks in the Andes, Himalayas, and African Rift Valley. Compared with the Northern Hemisphere, species richness in the Southern Hemisphere drops more steeply as latitude increases. These results confirm that Rapoport's rule does not apply generally, the authors conclude, and that patterns observed in the better-studied northern temperate zones don't necessarily apply elsewhere.
Stevens did not expect birds to support Rapoport's rule, because many species are migratory and capable of escaping the selective climatic pressures he thought were responsible for the latitudinal gradient in ranges. It will be interesting to look at similar data for other taxa as they become available to see if there is a latitudinal pattern in nonmigratory species.
Another pattern that can now be investigated with new analytic tools is phylogenetic tree imbalance, that is, when the tree for a major group comprises lineages (clades) with different numbers of species. Identifying the processes underlying clade diversification is also an important step toward understanding and protecting diversity.
Geographical range size is just one of the characters that has been hypothesized to explain the wide variation in clade richness within the class Aves. Others include body size, sexual selection, life history, ecological specialization, and ecological generalization. Although some of the associations have been shown to be significant, the explanatory power of these tests has been low, leading some scientists to argue that diversification is driven by random processes.
Albert Phillimore, David Orme, and Ian Owens, of Imperial College London, and Robert Freckleton, of Oxford University, have just published an analysis of bird lineages that identifies two ecological variables—annual dispersal and feeding generalization—as significant determinants of avian diversification rates (American Naturalist, vol. 167).
The scientists tested a number of variables for their ability to explain the variation among clade diversification rates: female weight, clutch size (as an index of life history), sexual dichromatism (as an index of sexual selection), number of breeding habitats and food types (as indexes of habitat and feeding generalization, respectively), adult dispersal, geographic range size, and island dwelling. As a whole, the multivariate model explained 53 percent of the variation in diversification rate, as compared with the 10 to 25 percent explained by earlier models.
The most significant predictors of diversification rate were annual dispersal, accounting for 24 percent of the variation, and feeding generalization, accounting for 17 percent. The mechanism linking cladogenesis with high levels of dispersal is not clear, nor is the influence of feeding generalization understood, but the importance of ecological factors in explaining evolutionary patterns is clear.