“Bert? You want to comment?”

Seger, his eyes brilliant, shrugged his narrow shoulders. “We made the best call we could; we did all the tests. We got it wrong. Next time we fly a NERVA, we’ll fix it.”

These things happen. Not an answer to satisfy the Presidential Commission, Michaels thought sourly.

“Go on, Hans.”

“By now,” Udet said, “the crew was aware that the thrust had died, after that first shove. We were only a few seconds after the first glitch in the flow. Now the hydrogen flow increased, markedly,” Udet continued. “It was like a spurt, from the faulty piping. The hydrogen passed its nominal flow rate and effectively flooded the core. We withdrew the moderator further—”

“And this is another point at which the standard procedure said shutdown,” Muldoon said harshly. “The moderators’ control margin was too low now; we didn’t have full control of the core. But again, we overrode the automatics.”

“We tried to save the mission,” Udet said.

“All right. Let’s stick to the facts for now; we can justify ourselves later. What next?”

“Now the flow of coolant into the core stopped altogether,” Udet said. “Perhaps at this point the piping failed completely.”

“This is the key moment, Fred,” Muldoon said. “You have a reactor that’s already unstable. The hydrogen flood has made the core isothermal — that is, at the same temperature throughout — so any changes happen all over the core, simultaneously. And the coolant flow has stopped; the core’s main heat sink, the flow of hydrogen through the jackets, has gone.”

“So it starts to heat up.”

“So it starts to heat up. Uniformly. And a lot faster than before.”

Udet said, “We tried to shut down. But the moderator was too far out of the core to have any immediate effect. The hydrogen in the core and the jacket boiled quickly and started to expand…”

“And now you’ve got a runaway,” Muldoon continued. “Because the reactor was designed with a positive temperature coefficient.”

Michaels sighed and locked his hands behind his head. “Just pretend I don’t know what you’re talking about.”

Muldoon grinned tightly. “I know. It took me a while to figure this stuff out. Look: suppose the temperature of your core rises. And suppose that the core is designed so that when it heats up, the reactivity drops — that is, the reaction rate automatically falls. That’s what’s meant by a ‘negative temperature coefficient.’ In that case you have a negative feedback loop, and your reaction falls off, and the temperature is damped down.”

“Okay. It’s kind of self-correcting.”

“That’s right; the whole thing is stable. That’s how they design civilian reactors. But in the case of NERVA, that coefficient was positive, at least for some of the temperature range. So when the temperature went up, the reactivity went up, too—”

“And the rate of fission increased, leading to a further temperature rise.”

“And so on. Yes.”

Michaels glared at Udet. “I can see the fucking headlines now, Hans. Why the hell did we fly an unstable reactor?”

Udet sat forward, his face pale, a muscle in his neck rope-taut with anger. “You must understand that we are not building a reactor to supply domestic electricity, here. We are not heating coffeepots. NERVA 2 is a high-performance booster, a semiexperimental flight model. Stability is not always the condition we require.”

Michaels frowned. And you just hate having to answer these asshole questions, don’t you, Hans? “Why do we need instability? What do you mean?”

Seger put in, “It’s like a high-performance aircraft, Fred. A ship that’s too stable will wallow like a sow. So you might design for instability. If a bird’s unstable, it can flip quickly from one mode to another; if you can control that, you’ve gained a lot of maneuverability”

“But that’s a big if, Bert. And evidently, when it got to the wire, we couldn’t control it. Hans, why didn’t you beef up the control system to cover for this?”

Udet punctuated his words by thumping the edge of his hand on Michaels’s desk. “Because — of — unacceptable — weight — penalties.”

Michaels dreaded having to put this man in front of the commission. “Let’s move on. What next?”

Udet said, “Events unfolded rapidly. The power output began to rise exponentially, doubling in a fraction of a second. The fuel pellets — which are uranium carbide coated with pyrolytic carbon — shattered from the thermal shock of the sudden power rise. The flow passages within the core melted. The moderator systems became inoperative. There was a hydrogen explosion, which ruptured the pressure shell and the biological shield—”

“All right.” Michaels found himself shuddering. “We know the rest.” Jesus. What a mess. “So the whole damn thing was caused by faulty hydrogen pipelines.”

Bert Seger nodded, and then he startled Michaels by saying: “It’s actually not as bad a scenario as you might have feared.”

“Not as bad? What the hell are you talking about, Bert?”

“The glitches in the hydrogen flow came from a simple component failure. What you had was ruptures in a six-foot length of stainless-steel fuel line, five-eighths of an inch in diameter, carrying liquid hydrogen from the tank into the nuclear engine. That’s all. So it’s easy to fix.”

“Why did the damn pipe rupture?”

“Well, we were flying with a new innovation,” Seger said, “that was supposed to guard against the effects of vibration. Each length of the pipeline had two vibration-absorbing ‘bellows’ sections in it, with wire-braid shielding on the outside. When the new line was put through vibration tests on the ground, it worked perfectly.”

“So how come—”

Udet said, “It turned out that in the atmosphere, the liquid hydrogen running through the pipe caused ice to gather on the braided shield. And that altered the characteristics of the pipe, enough to enable it to dampen out the most severe vibrations in the bellows during our testing.”

“Oh,” Michaels said. “But in the vacuum, no ice could form.”

“And those little bellows sang like a rattlesnake,” Joe Muldoon said. “When the Saturn first stage started its pogoing, the bellows couldn’t handle it. They just fell apart.”

Michaels asked Udet, “But how come you didn’t pick up the ice thing when you ran vacuum ground tests on the bellows?”

Udet faced Michaels squarely; he looked calm, somehow confident. “We did not run vacuum tests on this component. We did not anticipate the necessity.”

Michaels held his gaze for long seconds, but nothing more was forthcoming: no more data, no justification, no apology. “Well, I will be dipped in shit. Joe?”

Muldoon leaned over the desk and tapped the report. “This is where we show ourselves as culpable, Fred. Those goddamn bellows were Criticality One components: that is, their failure was liable to cause the loss of the spacecraft. But we didn’t test them out under true flight conditions. And, what’s worse, we’ve now dug out evidence of bellows problems on the S-NB’s previous unmanned test flight, although in that case we didn’t lose the mission.”

I’m dead meat, Michaels thought.

They could have anticipated the fault, and that was always deadly. And, it was always the way, some obscure little asshole technician somewhere at Marshall or the Cape would have written a report predicting precisely the failure they’d suffered, a report which no doubt had been laughed off and suppressed by NASA senior management, a report which was no doubt falling into the hands of some congressman even then…

“Culpable. Jesus. How I hate that word.”

Michaels got to his feet. He crossed to his window and folded his hands behind his back as he stared out over Washington. The light was fading from the sky, softened and stained by smog.

“I don’t want to minimize the impact of this, gentlemen. Quite apart from losing the crew, this is a genuine catastrophe. I have the ecology lobby around the world jumping up and down on my back. We’ve even been criticized for bringing a radioactive Command Module back into the atmosphere. There was strong opposition to flying nuclear materials into space even before the flight. And now the Russians have a fucking Soyuz up there taking pictures of the out-of-control glowing radioactive core we’ve abandoned in orbit.