Dinosaurs' swan song: out with a bang.

Paleontologists have grown frustrated during the last dozen years as physicists, chemists and other researchers outside their specialty have speculated that a huge meteorite may have wiped dinosaurs from the face of the planet 65 million years ago. Many of those who actually study dinosaurs believe other scientists have oversimplified the issue without consulting the fossil record to see whether the great beasts died out gradually or abruptly.

A group of paleontologists has now completed an intensive study indicating that dinosaurs did indeed die off rapidly at about the same time as the proposed impact. "We found no evidence of a gradual decline," says study leader Peter M. Sheehan of the Milwaukee Public Museum.

He and his co-workers addressed the extinction question by conducting a dinosaur census in the Hell Creek formation, a group of rocks from the last 2.5 million years of the Cretaceous period, which ended with the dinosaurs' extinction. Located in North America's upper Great Plains, this formation has for decades yielded an exceptional record of the last known dinosaurs.

Over three summers, paleontologists and volunteers surveyed undisturbed parts of Hell Creek in western North Dakota and eastern Montana. Walking 10 feet apart in search-party style, they scoured the area for dinosaur bones protruding from rock. Paleontologists then identified the bones and surrounding sediments, making sure the fossils had not been washed away from their original locations. The crew found some 4,100 bones representing almost 1,000 individual dinosaurs from eight taxonomic families.

Analysis of the bones showed the presence of all eight families in the lower, middle and upper portions of the formation, suggesting no gradual decline, the team reports in the Nov. 8 SCIENCE. Moreover, the relative strengths of these families remained constant from the earliest portion to the latest, indicating that ecological diversity was not waning, Sheehan says.

The new findings contradict previous studies, conducted at different sections of Hell Creek, that suggested the dinosaurs died off gradually over millions of years. Sheehan says his results fit with an abrupt-extinction scenario, but he adds that the data cannot resolve whether the extinctions unfolded over weeks or 100,000 years.

Past studies of Hell Creek pollen fossils, however, have revealed that many plant species in this region died out extremely rapidly at the end of the Cretaceous, suggesting that the purported impact caused the die-offs, says Douglas J. Nichols of the U.S. Geological Survey in Denver.

David Weishampel, a paleontologists at the Johns Hopkins University in Baltimore, says Sheehan's study represents a significant improvement over earlier work because the research team devised better methods for addressing problems inherent in fossil collecting. Nonetheless, he says, "I don't think their data are sufficiently strong to make the case they're trying to make."

Weishampel says the researchers need to collect more data on dinosaur bones in floodplain sediments, a type of deposit that yielded equivocal results in the new study.