Reality botany: data ease doubts about plant species.

Despite the doubts of some botanists, plant species aren't just some arbitrary human classification scheme, says a team of evolutionary biologists. What's more, plants don't deserve their reputation of being outrageously promiscuous, breeding across species boundaries, because some animals can do so even more freely.

Animal species can cross boundaries to mate more readily than plants do, says Loren H. Rieseberg of Indiana University in Bloomington. Among records of plant-and animal-hybridization tests, Rieseberg's team found that 31 percent of attempted plant combinations readily yielded fertile offspring, whereas 61 percent of animal-species crosses did so. "That was a surprise to me," says Rieseberg.

His team has been investigating the concept of plant species. Citing, for example, the alleged tendency of plants to hybridize, many botanists have concluded that plant species are arbitrary groupings, says Rieseberg. He and his colleagues compiled data from earlier statistical analyses of traits for species in nearly 700 plant species. In the March 23 Nature, the researchers argue that the data divide plants into distinct groups.

The idea of a species has changed over the years, says herbarium director Brent Mishler of the University of California, Berkeley. Before Charles Darwin, people considered each species a basic unit of creation. Darwin's emphasis on how populations gradually change gave the notion of species a more arbitrary quality: Species had whatever boundaries taxonomists chose. The idea of a species as a population of individuals that breed mostly with each other comes from 20th-century, theorists.

Rieseberg and his colleagues reviewed 218 studies that have classified organisms by quantifying traits such as stem height or leaf shape and then calculating how similar various populations are.

In more than 80 percent of plant and animal genera with at least five species studied this way, the species made tidy clusters instead of forming a continuum, Rieseberg and his colleagues report. The clusters argue for species as distinct entities.

To see whether members of the numerically derived clusters breed mostly within a cluster, Rieseberg and his colleagues searched for studies of hybridization. They examined results of more than 1,000 experimental crosses of plants and more than 600 of animals. Ferns had the least successful mating across species boundaries. Birds had the most.

John Kress, curator of botany at the Smithsonian Institution in Washington, D.C., says that Rieseberg and his colleagues' study makes progress, but he doubts that it "will put an end to the controversy over species as arbitrary constructs versus objective entities."

Mishler calls the group's approach "clever" but argues that plant classification has already moved beyond such issues. The most modern approach to classifying plants, he says, depends on sorting out the family histories of lineages before defining species.