Ian Buckle said that "for bridge engineers and owners in the eastern and central U.S., the Kobe earthquake has perhaps even greater significance than recent earthquakes in California, such as Loma Prieta and Northridge."

A professor of civil engineering at UB, Buckle made reconnaissance visits to the Kobe region following the earthquake.

While bridges in California are generally monolithic continuous structures, he noted, bridges in both the Kobe region and the eastern and central U.S. tend to be constructed of several discontinuous parts. These include a concrete deck slab and steel girders supported by steel bearings that sit on concrete columns and foundations. The purpose of these bearings is to allow the bridge superstructure to expand and contract in response to seasonal temperature changes.

In Kobe, many of these bearings failed, said Buckle, causing some bridge decks to fall from their supports.

"We've been suspicious of these devices for some time," he said. "We've seen them fail before, but assumed that it was simply a matter of improving the design."

In Kobe, he added, reconnaissance teams saw literally hundreds of failures of steel bearings.

In past earthquakes, engineers have attributed failures of bridge bearings to inadequate design details and poor workmanship.

"But Japan's standard of engineering is exceptionally high and yet they still failed," Buckle said, "even in new bridges."

Determining the cause of these failures presents a difficult problem for engineers, Buckle said.

UB researchers and others have tested individual bearings in customized test machines in which hydraulic actuators apply loads that mimic those expected to occur in an earthquake. However, in these laboratory tests, the bearings did not fail.

"Based on our test results, you wouldn't have expected these bearings to fail," Buckle said. "They appear to have adequate capacity for resisting seismic loads."

He and John Mander, UB associate professor of civil engineering, are now proposing to shift the emphasis of their research to focus on the demand that earthquakes place on bridge bearings.

"We don't understand well enough the forces that an earthquake delivers to a bridge bearing," Buckle said. "They may be much greater than we have calculated."

According to Buckle, bearing failures may happen because of an uneven distribution of loads among the many bearings that support a bridge deck.

"Some bearings may be resisting far greater loads than we expect while others resist far less load than we expect," he said.

Once the bearing with the greatest load fails, the load distribution changes, he explained. Since fewer bearings are now carrying the same total load, another bearing is likely to fail, causing another redistribution of load and so on.

"Bearings could be failing progressively one after the other like falling dominoes," said Buckle.