The entry stops at the control room. By Wednesday evening the reactor had been covered again and the damage was done; nothing that happened inside the plant over the next five days made it worse. What happened outside the plant is why the accident is remembered as it is, and the President’s Commission, which took most of its testimony on these days, told it in detail. This piece follows the Commission’s account from Thursday 29 March to Monday 2 April, with the words of the people involved as the Commission’s report gives them.

Thursday

Thursday, the Commission wrote, “seemed a day of calm.” The log of the Dauphin County emergency office reads like a crisis passing: at 7:55 in the morning the state emergency agency reported “no danger to public”; at 3:30 in the afternoon, “situation is improving”; at 9:55 that night, “no health risk off-site, no emergency, bringing reactor to cold shut down.” Readings across the river were one to three millirems an hour.

Late in the afternoon the governor, Richard Thornburgh, held a press conference at which the NRC’s inspector on the island, Charles Gallina, told reporters the danger was over for people off the island. Thornburgh distrusted it. At 6:30 that evening Gallina and a reactor inspector, James Higgins, got the result of an analysis of the reactor’s coolant, which showed that the core damage was far greater than either had supposed; at ten o’clock Higgins telephoned the governor’s office with the news and said the chance of radiation releases was greater than had been thought. “Nothing had changed inside the plant,” the Commission wrote, “only NRC’s awareness of the seriousness of the damage.”

Friday morning

The reactor’s let-down and make-up systems ran through the auxiliary building, and NRC officials suspected leaks in them for the sporadic releases of the previous two days. Water from the let-down line went to a make-up tank, where the gases dissolved in it at high pressure came out of solution, “much as the gas bubbles in a pressurized carbonated beverage appear when the bottle is opened,” and were normally compressed into two waste gas decay tanks. The NRC’s worry on Thursday night was that if those tanks filled, their relief valves would open and the release would be continuous. That worry, and what Kemeny afterwards called a “horrible coincidence”, made Friday morning.

About halfway through his shift James Floyd, Unit 2’s supervisor of operations, decided to transfer gas from the make-up tank to a decay tank, because the pressure in the make-up tank had risen so high that water could no longer flow into it. He knew the transfer would release radiation through the system’s leaks. He did not check with anyone, and ordered it begun at 7:10. Thirty-four minutes later he asked for a helicopter to take readings. It reported 1,000 millirems an hour at 7:56 and 1,200 at 8:01, 130 feet above the Unit 2 vent stack.

At NRC headquarters in Bethesda, Lake Barrett, a section leader in the environmental evaluation branch, had helped the previous evening to calculate “a hypothetical release rate” for the decay tanks’ relief valves opening. Shortly before nine on Friday he was told of a report from the island that the tanks had filled, and was asked to brief the senior staff: Lee Gossick, the executive director for operations; Harold Denton, the director of reactor regulation; Victor Stello, the director of operating reactors; John Davis, the acting director of inspection and enforcement; and Harold Collins of the office of state programmes. Asked what the release would mean as a dose off site, he did a quick calculation and gave the figure of 1,200 millirems an hour at ground level. “Almost at that moment,” the Commission wrote, “someone in the room reported a reading of 1,200 millirems per hour had been detected at Three Mile Island.” “It was the exact same number,” Barrett testified, “and it was within maybe 10 or 15 seconds from my first 1,200 millirems per hour prediction.” The report that the tanks had filled was wrong, and the helicopter’s reading had nothing to do with them. Nobody in the room knew whether the 1,200 had been measured on or off the site, at ground level or in the air, or from what.

Collins described “an atmosphere of significant apprehension,” and “the question was being raised in the minds of many as to whether or not those people up there would do the right thing at the right time, if it had to be done.” Denton directed Collins to tell Pennsylvania that senior NRC officials recommended the governor order an evacuation. Collins telephoned Oran Henderson, the director of the Pennsylvania Emergency Management Agency, and, choosing the distance himself, recommended evacuating everyone as far as ten miles downwind. Henderson called Scranton. An aide told Thomas Gerusky, the director of the state’s Bureau of Radiation Protection, who knew of the 1,200 reading, telephoned an NRC man at the plant, concluded no evacuation was needed, could not get through to the governor’s office by telephone, and went there in person to argue against it.

Kevin Molloy, the Dauphin County emergency director, had been told of the release by Floyd at 8:34 and by the state at 8:54, with the word that no evacuation was needed. At 9:25 Henderson called to say an official evacuation order would follow in five minutes. Molloy alerted every fire department within ten miles and broadcast on WHP that an evacuation might be called. On the island, shortly afterwards, Gallina was confronted by a Metropolitan Edison employee: “What the hell are you fellows doing? My wife just heard the NRC recommended evacuation.” Gallina checked the readings on and off the site, was told by a reactor inspector that “things were getting better”, and telephoned Region I and Bethesda “to call back that evacuation notice.”

Shortly after ten Thornburgh spoke to Hendrie, who assured him no evacuation was needed and suggested that everyone within five miles downwind stay indoors for half an hour; the governor later that morning advised everyone within ten miles to stay inside. He asked Hendrie for a single expert on the island on whom he could rely. About an hour later President Carter telephoned the governor, having talked to Hendrie, and promised him the expert, who would be Denton, and a dedicated communications link between the plant, the governor, the White House and the NRC.

At about 11:40, as the governor met his aides, Hendrie called again, and, by Gerusky’s recollection of the conversation over the speaker phone, apologised for the staff’s error in recommending evacuation. Just before, an aide of the state’s health secretary, Gordon MacLeod, had renewed MacLeod’s recommendation that pregnant women and children under two be moved. Thornburgh told Hendrie. “If my wife were pregnant and I had small children in the area,” Hendrie said, as Gerusky recalled it, “I would get them out because we don’t know what is going to happen.” Shortly after 12:30 Thornburgh advised pregnant women and preschool children within five miles of the plant to leave, and closed the schools in that area.

“There are known risks, I was told, in an evacuation,” he told the Commission afterwards. “The movement of elderly persons, people in intensive care units, babies in incubators, the simple traffic on the highways that results from even the best of an orderly evacuation, are going to exert a toll in lives and injuries. Moreover, this type of evacuation had never been carried out before on the face of this earth. . . . When you talk about evacuating people within a 5-mile radius of the site of a nuclear reactor, you must recognize that that will have 10-mile consequences, 20-mile consequences, 100-mile consequences, as we heard during the course of this event. This is to say, it is an event that people are not able to see, to hear, to taste, to smell.”

Many people had already gone on their own. “On March 29 of this year, my wife and I joyously brought home our second daughter from the hospital; she was just 6 days old,” one man told the Commission.

Friday afternoon

The company’s standing with the press ended that morning. John Herbein, Metropolitan Edison’s vice president for generation, arrived at eleven to brief reporters in Middletown and opened by saying the release had been measured at about 300 to 350 millirems an hour. The reporters knew the figure of 1,200; he did not. “I don’t know why we need to . . . tell you each and every thing that we do specifically,” he said at one point, and the Commission judged that it was that remark that “essentially eliminated any credibility Herbein and Met Ed had left with the press.” The next day the White House proposed that Denton alone brief reporters on the technical side, and the company agreed.

Denton reached the island that afternoon. He had been given estimates before leaving Bethesda that the bubble of hydrogen in the top of the reactor vessel, which had been found on Thursday, could not ignite of itself for five to eight days, and he set to work on ways of removing it. At about 8:30 that evening he briefed the governor in person for the first time: the fuel damage was extensive; the bubble was a problem for cooling the core; no immediate evacuation was necessary. The two men then held the first of their joint press conferences. The governor lifted his advice that people within ten miles stay indoors, and kept the advisory for pregnant women and preschool children.

Saturday

“The great concern about a potential hydrogen explosion inside the TMI-2 reactor came with the weekend,” the Commission wrote. “That it was a groundless fear, an unfortunate error, never penetrated the public consciousness afterward, partly because the NRC made no effort to inform the public it had erred.”

The hydrogen had been made on Wednesday morning, when the steam met the overheated zirconium of the fuel cladding. For it to burn or explode inside the vessel there would have to be oxygen, and the fear was that radiolysis, the breaking of water molecules by radiation, was supplying it. At about 9:30 on Friday night Hendrie asked Roger Mattson, the director of the NRC’s division of systems safety, to find out how fast. “He said he had done calculations,” Mattson said in his deposition. “He was concerned with the answers.” Two teams worked through the weekend, with laboratories and consultants outside the agency, one on the oxygen, the other on whether the mixture could burn.

At noon on Saturday Hendrie told Denton by telephone that he was concerned oxygen was building up, and that the governor should be told. Shortly after one Mattson had his first answers, and more an hour later. “I had an estimate there was oxygen being generated, from four independent sources, all with known credentials in this field,” he said. “The estimate of how much oxygen varied, but all estimates said there was considerable time, a matter of several days, before there was a potential combustible mixture in the reactor coolant system.” At a Commission hearing later, under questioning from Commissioner Thomas Pigford, he conceded that from the information available at the time the agency could have found that no excess oxygen was being made and there was no real danger. By Saturday night, though, his consultants were telling him the oxygen was at the threshold of flammability. Late that evening James Taylor of Babcock & Wilcox repeated to the NRC what another of the company’s engineers had said on Thursday night, that no excess oxygen was being generated. “That information,” the Commission wrote, “never reached him.”

At 2:45 on Saturday afternoon Hendrie met reporters in Bethesda and said that a precautionary evacuation out to ten or twenty miles might be necessary if engineers tried to force the bubble out of the reactor. An Associated Press reporter, Stan Benjamin, followed up with two NRC officials, Edson Case and Frank Ingram, and an unnamed source, and learnt of the concern inside the response centre that the bubble could become explosive within days, perhaps two. He read much of his story to Case and Ingram, who agreed it was accurate. It moved at 8:23 that night, the first notice to the public that NRC officials feared the bubble might explode on its own.

Denton heard of the story shortly before he joined the governor and the lieutenant governor for a late press conference in Harrisburg. “There is no imminent catastrophic event foreseeable at the Three Mile Island facility,” Thornburgh said. “There is not a combustible mixture in the containment or in the reactor vessel,” Denton said. “And there is no near-term danger at all.” Asked about the contradiction with Bethesda, he said, “No, there is no disagreement. I guess it is the way things get presented.” The Commission’s next sentence is: “But there was disagreement, and Denton wanted it resolved.” The President had announced that evening that he would visit the plant the next day. Denton told Stello to take the oxygen question to outside experts. Stello had already had a call from a presidential aide about the AP story, and had told the White House he did not share the concern in Washington. In the counties around the plant the emergency offices were swamped with calls, and the flow of information from the state to the local level had largely stopped.

Sunday

On Sunday morning Mattson and other staff met three of the NRC’s commissioners and agreed, by Mattson’s deposition, that five percent oxygen was a realistic flammability limit and eleven percent a detonation limit, that there could be no spontaneous combustion below 900 degrees, that oxygen was being produced at about one percent a day, and that the bubble now held five percent. Hendrie and Mattson then drove to the plant.

Stello had spent the morning on the other side. A pressurised water reactor runs with free hydrogen in its coolant, and that hydrogen takes up the oxygen that radiolysis frees; the process was the same now, he argued, and there was no danger of an explosion. Minutes before the President’s one o’clock arrival at Harrisburg airport, Hendrie and Mattson met Denton and Stello in a hangar. Mattson and Stello had not spoken since Friday morning. Mattson laid out the headquarters conclusions. “And Stello tells me I am crazy, that he doesn’t believe it, that he thinks we’ve made an error in the rate of calculation,” Mattson told the Commission’s staff. “Stello says we’re nuts and poor Harold is there, he’s got to meet with the President in 5 minutes and tell it like it is. And here he is. His two experts are not together. One comes armed to the teeth with all these national laboratories and Navy reactor people and high faluting PhDs around the country, saying this is what it is and this is his best summary. And his other [the operating reactors division] director, saying, ‘I don’t believe it. I can’t prove it yet, but I don’t believe it. I think it’s wrong.’”

Denton briefed the President on the state of the plant and the uncertainty about the bubble. Carter was driven to the island, put on yellow plastic shoe covers, and toured the plant and the Unit 2 control room with Mrs Carter, the governor and Denton. Hendrie, Stello and Mattson went to the NRC’s temporary offices and spent the afternoon on the telephone to experts at Westinghouse, General Electric and elsewhere. “By three o’clock, we’re convinced we’ve got it,” Mattson said. “It’s not going to go boom.” In Bethesda the staff reached the same conclusion later in the day.

In mid-afternoon new measurements showed the bubble diminishing: the gas was still there, but spread through the system in smaller pockets that would be easier to remove. “Why this occurred, no one knows,” the Commission wrote. “But it was not because of any intentional manipulation by Met Ed or NRC engineers.” By late Sunday the NRC knew there was no danger of an explosion and that the bubble was going. “It was good news, but good news unshared with the public.” No announcement was made that day of the error or of the bubble’s shrinking, partly because the agency’s experts were not yet certain, and the governor was not told of the miscalculation.

Monday

On Monday morning a company official had already told a reporter the bubble was essentially gone when Denton and Mattson met the press. Denton acknowledged a “dramatic decrease in bubble size” and said more analysis was needed “to be sure that the equations that are used to calculate bubble size properly include all effects.” Of the explosion, he said that “the oxygen generation rate that I was assuming yesterday when I was reporting on the potential detonation inside the vessel is, it now appears, to have been too conservative.” He used the word “conservative” throughout, and never said the agency had been wrong. The tone, Mattson said, had been settled at a meeting that morning: “We wanted to go slow on saying it was good news. We wanted to say it is good news, do not panic, we think we have got it under control, things look better, but we did not want to firmly and finally conclude that there was no problem. We had to save some wiggle room in order to preserve credibility. That was our judgement.”

Schools outside the five-mile area reopened on Wednesday 4 April. The governor withdrew the advisory for pregnant women and preschool children on 9 April. The Commission’s finding on the week was that the amount of radiation received by anyone outside the plant had been very low, and that its most serious health effect had been “severe mental stress, which was short-lived,” highest among those within five miles and in families with preschool children.

Sources

• President’s Commission on the Accident at Three Mile Island (John G. Kemeny, chairman), The Need for Change: The Legacy of TMI, Washington, October 1979, “Account of the Accident”, the sections for Thursday 29 March to Monday 2 April and the epilogue (pages 132 to 158), and the overview’s findings on the health effects. Every quotation here (Barrett, Collins, Gallina, Hendrie as Gerusky recalled him, Thornburgh, Herbein, Mattson, Denton) is as the Commission’s report gives it, from testimony, deposition or interview. A United States government work. archive.org