The mystery of why backup cooling pumps were shut and locked out the morning of the accident at Three Mile Island deepened yesterday when four technicians testified they left the pump valves open after testing them two days before the accident.

"I reopened those valves and those valves were reopened," There Mile Island shift foreman Carl L. Gutherie told the President's Commission on the Accident at Three Mile Island. "I opened the valves myself and three of us went into the control room and checked the lights that said they were open. I then verified it and signed off on a written checklist that the valves were opened."

Gutherie and three technicians who closed, tested and said they reopened the valves leading to the backup pumps were asked repeatedly if they had any idea why the valves were closed so that the pumps failed to supply cooling water to the reactor for the first eight minutes of the March 28 accident.

"No, sir, I have no idea," Gutherie replied. "There was no reason other than the surveillance checks we gave them 42 hours before the accident to close them."

Martin V. Cooper, one of the three technicians who assisted Gutherie in the test and who testified the valves were left open, said the backup pumps and valves were tested often enough for operators to be familiar with the procedure of reopening them. Said Cooper: "It's critical that those valves be left open. They must be left open."

The Nuclear Regulatory Commission has said that if the backup pumps had come on when the main pumps failed, there might have been no accident at Three Mile Island. Locking the valves and pumps out for longer than three hours is a violation of NRC rules, punishable by fines, suspension or loss of license.

Following Gutherie and his three man crew to the witness stand yesterday were the four operators on duty in the control room the morning of the accident. They testified they did not noticed that the pump valves were closed until eight minutes into the accident, in part because 200 alarms went off on the control panel in the first two minutes.

"I'd like to have thrown away that alarm panel," said Edward Frederick, one of the control room operators. "It wasn't giving us any useful information."

Chief operator Craig Faust said he finally noticed lights on the control panel indicating that the backup pumps weren't working.

"One of my immediate actions was to take control of those valves and open them up," Faust said. "Things had happened so fast up to then that I thought 60 seconds must have passed before I noticed those valves were closed. It turned out to be eight minutes."

Once they heard the technicians testify about the locked-out backup pumps, the members of the president's commission bore down on the four control room operators, wanting to know why it took them so long to realize that a relief valve was stuck open, allowing cooling water to escape from the reactor and uncover the atomic core inside.

The operators said the relief valve leaked on so many occasions that they paid no attention to gauges telling them the valve was overheating. These should have told them that steam and hot water were pouring out of the reactor and through the open valve.

"That valve had always run hot and we'd had prior leaks that ran up temperatures in the neighborhood of 200 degrees," shift supervisor William Zewe testified, "so that when we saw temperatures in that valve up to 230 degrees we thought they were the normal leaks. We had no idea the valve was stuck open."

Zewe said a computer-operated alarm system that could have warned the operators that the valve was so hot it was stuck open failed in the first two hours of the accident. A computer printout of the alarms was so overburdened that it was running an hour and a half behind.

"There was another indicator of the valve's condition, a drain tank indicator where the dumped water was going," said Frederick. "But it's not easily visible. You have to walk behind the panel facing the control room to even look at it."

The operators were asked what their reactions would have been had they known that temperatures in the relief valve were really as high as 280 degrees. Said Frederick: "It would have concerned me to a larger degree."

The four operators testified for almost four hours, giving the commission the clear indication they had no idea at the time of how serious the accident was.

"We don't train for accidents like this and we were really getting a lot of contradictory signals," Frederick said. "We need something in that control room that gets us as close to the nuclear core as we can get."

Faust, Frederick and Zewe testified they were unaware that the core had been badly damaged until three or four days after the accident.

"None of our instruments showed the core temperatures [2,000 degrees] we'd need to have to indicate core damage," Zewe said. "Even when we got the first radiation alarms, we though it might be crud burst due to core shock."

There is always some uranium dust stuck to fuel rods inside the core, experts have said, that can come loose and escape with the cooling water if the coolant undergoes frequent temperature changes. This is what Zewe meant by "crud burst" and "core shock."

Halfway through the hearing, commission Chairman John Kemeny interrupted the operators and said: "I have to ask this question." He pointed out that March 28 was the first anniversary of the day Three Mile Island produced its first chain reaction. "Were you having a party?" he asked. "No, sir," Zewe replied."We didn't even know it was the first anniversary."