The mustard war wasn't so racy after all.

The battle's been raging for almost 9 million years, but contrary to a beloved evolutionary theory, there was no arms race between the plant and its pest.

Molecular biology's much-studied mustard, Arabidopsis thaliana, and the plant's old enemy, a Pseudomonas bacterial rot, have not been trumping each other's defenses with escalating innovations, report Eli A. Stahl of the University of Chicago and his colleagues.

Instead, analysis of DNA regions around the plant's rot-resistance gene suggests that pretty much the same old weapons have gone through cycles of use and disuse in what the researchers describe as "trench warfare." For the mustard-rot war, "we reject the arms race hypothesis," they state flatly in the Aug. 12 NATURE.

Their manifesto challenges what evolutionary ecologist Peter M. Kareiva from the National Oceanic and Atmospheric Administration in Seattle calls "one of the more compelling metaphors of biology." The idea resounds through tales of interplay between diseases and hosts and between herbivores and food plants.

"I don't think there has ever been that much empirical evidence for any sort of escalating arms race," Kareiva says. "The power of the idea was mostly the power of the metaphor."

Much of the support for the notion came from observations of intricate ways that one organism overcomes another's defenses, Kareiva points out. For example, he describes a plant covered with tiny spikes that stab invading bugs. One successful attacker, however, spins a web to traverse the spikes and reach a naked leaf edge, where it can nibble unscathed.

For a rigorous look at the history of such conflicts, Stahl and his colleagues sequenced DNA from mustard plants collected in 26 locales ranging from Indiana to Kazakhstan. Twelve populations had a gene, rpm1, that allows the plant to recognize the rot and mount a physiological defense. The others had no resistance gene and developed soft, mushy lesions when exposed to the pest.

Had there been an arms race, the currently effective resistance would have evolved recently. The researchers instead found that rpm1 has been around for about as long as the species itself. The diversity in the inactive DNA flanking the resistance gene indicates its ancient origin, argues coauthor Joy Bergelson, also of Chicago.

"It was very surprising," Bergelson says. This evidence suggests to her that raging epidemics favor the spread of the resistance genes until so many plants are protected that the rot runs out of victims and recedes. Then rpm1 itself, which might drain some of the plant's resources and thus exact some maintenance cost, wanes in the population until the next epidemic.

"It's very clever, and it's probably right," comments Kareiva. However, he muses that there may be too little evidence yet to accept a new model. For example, "it's often hard to identify a cost of resistance," he cautions.

Regardless, the mustard's history is hardly an arms race, says Barbara A. Schaal of Washington University in St. Louis. In the classic scenario, any variation is transient. Yet in the mustard's, there's "good evidence that natural selection is operating to maintain diversity," she says.

Without more molecular tests, Bergelson doesn't claim to know whether other conflicts follow the same pattern as the mustard-rot scenario. As Kareiva puts it, however, now "there will be a lot of labs looking at sequencing data."