Botanists uproot their old tree of life.

A shrub so obscure that most botanists have never seen one appears to represent the long-sought first branch at the base of the family tree of flowering plants.

The lineage of the seemingly ho-hum Amborella, often forced to nestle with the avocado and sassafras in the family tree of living relatives, deserves to be moved way down the trunk, according to several lines of genetic evidence.

Some 200 botanists from 12 countries have been collaborating for the past 5 years to refine various sections of the tree of life. Members of this Green Plant Phylogeny Research Coordination Group unveiled their conclusions this week at the XVI International Botanical Congress in St. Louis.

"One of the most practical things to know about a plant is what it's related to," explains one of the group's co-leaders, Brent Mishler of the University of California, Berkeley. Evolutionary relationships offer clues to a plant's properties. Mishler points out that when cancer-drug prospectors found promising qualities in the bark of a yew tree, they turned to related species to find a more readily available compound.

Some of the phylogeny group focused on the origin of flowering plants, which Charles Darwin called an "abominable mystery." In the past 10 years alone, botanists have published at least 15 arrangements of early flowering-plant lineages.

Now, a collaboration including Pamela Soltis and Douglas Soltis of Washington State University in Pullman and Yin-Long Qiu of the University of Zurich reports analyses of the similarities across more than a hundred species for at least six genes in chloroplasts, mitochondria, and cell nuclei. Their work pushes Amborella to the bottom of the flowering-plant tree. Water lilies branch out next, and then comes a group with star anise.

Amborella, essentially a living fossil, grows wild only on the island of New Caledonia in the South Pacific. The Pullman botanists know of just one plant in the United States; it's in Santa Cruz, Calif.

The surge of genetic data thrills Michael J. Donoghue of Harvard University. "I think we've finally got a resolution on a problem that looked as if it was not going to be resolved ever," he says.

He's found some other surprises in the tree that's emerging from the 200-scientist collaborative effort. The lotus, Nelumbo, appears not to be related to water lilies as previously thought. Instead, several genetic analyses place it with sycamores. "It's bizarre," shrugs Donoghue.

Looking with an outsider's perspective, mycologist John W. Taylor of the University of California, Berkeley cautions, "It's a huge advance, but it's not over." The first comparisons of a fungal gene yielded a tidy family tree, he recalls, but as more people analyzed more genes, conflicts developed.

The boom in phylogeny may swamp the traditional Linnean system of plant names, Mishler warns. The intricate relationships scientists are discovering are overwhelming that hierarchical system, with its families, orders, phyla, and so on. Mishler favors dropping these ranks entirely.

Despite the challenges of modern, fast-moving phylogeny, Donoghue says, "this is truly the greatest time to be alive with respect to these problems."