Lowering lilies on the tree of life.

New genetic analyses are shaking the lowest living branch on the family tree of flowering plants. The question is: Should we move water lilies down to that ancient lineage?

Last summer, presentations at the International Botanical Congress in St. Louis redrew the notoriously puzzling base of the tree. On the lowest branch that still has a living sprout, they placed a living fossil that few botanists have ever met. Amborella, a white-flowering shrub, grows wild only on the island of New Caledonia in the South Pacific (SN: 8/07/99, p. 85).

Three papers now reconsider that bottom branch. Amborella is not alone, argues Todd J. Barkman of Pennsylvania State University in State College. He, Claude W. dePamphilis, and their colleagues collected a large set of data and compared three methods of tree drawing.

The results differed, and one placed water lilies, or Nymphaeales, beside Amborella on that bottom branch. The water lilies were on a slightly higher branch in the other drawings and previous analyses.

To explore this discrepancy, the researchers reexamined the data, first applying a technique called RASA, which pinpoints and eliminates highly inconsistent, or noisy, data. From this point of view, water lilies belong on the very lowest branch, Barkman and his colleagues say in a paper scheduled for the PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES.

This new detail of the tree's early branches could change ideas about the history of flowering plants, Barkman and his team argue. For example, if Amborella alone represents the first branch, they'd bet that the ancestral plants segregated sex organs onto different flowers just as Amborella species do today. However, add water lilies' two-sex blooms and that supposition doesn't look as strong.

Not so fast there, protests another group that has also looked at the lowest-branch question with a large set of data. It's true that some ways of doing the analysis yield a basal lineage of Amborella plus Nymphaeales, Pamela Soltis of Washington State University in Pullman and her colleagues report in the November supplement to INTERNATIONAL JOURNAL OF PLANT SCIENCES (IJPS).

However, they also still find "strong support" for Amborella alone, says coauthor Douglas Soltis, also of Washington, and he says he's leaning toward that configuration.

A paper in the same issue of IJPS draws trees using different data. Michael Donoghue of Yale University (SN: 10/23/99, p. 268) and Sarah Matthews of the University of Missouri in Columbia narrowed down the choices for lowest branch of the tree to either Amborella alone or Amborella with water lilies. Donoghue says he doesn't see either position as a clear winner yet.