PORTLAND -- It would take a massive spill inside a Umatilla Chemical Depot bunker to give construction workers the symptoms they complained of in September 1999, an Army meteorologist testified Tuesday in U.S. District Court.

"It would take a spontaneous spill of every munition in that bunker," said Mike Myirski, an expert in projecting how chemical agents travel through the air.

That would mean 2,000 weapons filled with sarin had to spill at the same time to sicken workers in a building 80 yards away. Myirski's analysis assumed the sarin bunker's doors were closed and the air inside nearly stagnant with the chemical agent leaking through the ventilation system.

If the doors had been open, however, just a liter of sarin spilled on the ground would have caused the same illnesses.

Myirski was called by the Army to the witness stand on the sixth day of the federal trial to help refute the workers' claims they were exposed to sarin while building the incinerator plant that will be used to destroy the Army's 3,717 tons of sarin and mustard gas.

Several workers became simultaneously ill on Sept. 15, 1999, complaining of eye, throat and chest pains. Many were coughing, vomiting and having difficulty breathing.

Myirski said he was asked to estimate how much agent would have to have escaped K Block to explain the injuries.

He said his analysis was based on "conservative presumptions."

"This is a rough guess at best?" questioned Keith Dozier, a Portland-based attorney representing the plaintiffs.

"An educated guess," answered Myirski.

"An educated guess without knowing the configuration of the bunker or the outside conditions?" Dozier asked.

"Correct," Myirski replied.

The Army contends the workers were sickened by some type of industrial accident unrelated to the chemical weapons.

District Court Judge Dennis Hubel also heard testimony Tuesday that an Army news release distributed just a few hours after dozens of workers became ill claimed chemical agents were not involved in the incident. However, the interior of the incinerator building wasn't tested for chemical agents until more than a week later, said Darrel Johnston, a laboratory manager for Southwest Research Institute, who worked at the depot at the time of the incident.

Johnston said he was told to obtain air samples from within the incinerator building, but he didn't receive the stainless steel, vacuum-sealed canisters needed to collect the air samples for several days.

Those canisters were placed in the building nearly a week after the incident, he said.

Johnston said he and co-workers went into the incinerator building after the accident to look for possible causes, and none of his team experienced the symptoms the construction workers were experiencing.

The Army also used its portable monitoring equipment to test along the perimeter of the building and inside one room, but the hoses used to collect air samples weren't long enough to reach deep inside the building, where most of the injured workers were at the time of the accident.

But that testing also was done after the Army's first news release notifying the public of the accident was sent out.

Another Army news release sent out two days after the accident, said chemical agent had been ruled out as a cause. But Johnston said he couldn't rule out chemical agent based on the air monitor's results.

The trial is expected to continue through next Tuesday.