James Hallock discovered just how little it takes to bring down a space shuttle.
He did it by playing with pencils.
As a member of the Columbia Accident Investigation Board, the pear-shaped, bewhiskered expert on flight safety had a New Englander's flinty skepticism and a physicist's distaste for untested accident theories.
On this day, Hallock, 62, scowled at specifications for the reinforced carbon panels that shielded the leading edge of Columbia's wings from the heat of reentry.
If one of the $800,000 panels had cracked, it might have been the flaw that on Feb. 1 caused the $1.8-billion spacecraft and its crew of seven astronauts to plummet in a shower of molten debris across six states.
Hallock brooded over a simple question: What would it take to break one?
Engineers gave him the original 25-year-old NASA specifications, which said the panels must withstand an impact equal to the kinetic energy of 0.006 foot-pounds.
What did that mean? Hallock twiddled a yellow No. 2 pencil between his fingers. How far, for instance, would he have to drop the pencil to generate that kind of impact?
In the mailroom at the board's makeshift headquarters in Shirlington, Va., he found a box of pencils and a postal scale.
He weighed the pencils, then calculated the mass of the average pencil. With that number, he worked out how much punch each pencil would pack.
"The answer," he would say later, "is that a No. 2 pencil dropped from about 6 inches equals the kinetic energy number they had."
The panels, in actual manufacture, were much stronger. But that remarkably low standard, Hallock believed, was a tangible measure of how confident NASA engineers were that nothing would hit the leading edge of the wing.
"The number didn't matter to them," Hallock said, "because they assumed nothing would ever hit the shuttle."
Call it forensic engineering or, more plainly, detective work.
The Columbia accident investigation was the most exhaustive scientific inquest ever undertaken.
Suspicion led down a hundred blind alleys. Investigators quarreled. Mission insiders tried to control the probe. Outsiders railed about secrecy.
The investigators mustered the most sophisticated techniques that science could devise -- X-ray scanners, neutron beam machines, hypersonic wind tunnels.
They also used red food coloring, a bicycle pump, a hobby-shop hacksaw and a steam iron.
They conducted some tests in classified military laboratories, others in the nearest kitchen sink.
In arcane debates about trapped-gas analysis, radar cross sections and spatter metallurgy, they stalked answers to wrenching questions of guilt, shame and responsibility.
Every imperfection they found revealed a human face.
All that they discovered reinforced what Hallock learned by dropping his No. 2 pencil.
The space shuttles are by design unsafe.
As the most complex flying machines ever assembled, each shuttle contained more than 2.5 million parts, 230 miles of wiring, 1,060 valves and 1,440 circuit breakers. All of it had to function properly at extremes of speed, heat, cold, gravity and vacuum -- the interaction of its parts just at the edge of human understanding and control.
From liftoff to landing, the shuttles flew in peril.
In orbit, they maneuvered through a hailstorm of 10,000 man-made objects larger than a softball and millions of smaller pieces of debris. At orbital velocities, an object no larger than a pea carried the force of a falling 400-pound safe.
During NASA's first 75 shuttle flights, technicians had to replace 60 cockpit windshields -- at $40,000 each -- because of pitting from debris.
But launch was even more dangerous. Orbiting junk at least could be tracked by radar and avoided.
For all their efforts, shuttle engineers could not stop ice and chunks of insulating foam from falling off the shuttle's 15-story external fuel tank and striking the spacecraft during its eight-minute ascent into orbit.
Agency engineers could not fix this fundamental flaw, nor could they craft a safer vehicle. They dared not abandon the only vehicle the country had to carry people into space.
So NASA continued to launch the shuttles, gaining confidence each time the crew returned safely.
"The program had been put in this box they could not get out of," said Scott Hubbard, director of NASA's Ames Research Center in Mountain View, Calif., and a member of the investigating board.
Blind to the consequences, they had constructed a trap and baited it with ambition.
In the winter darkness, Caltech astronomer Tony Beasley, 38, shivered in his front yard.
According to the announced flight path, Columbia would soon swoop over moonlit mountain slopes draped with snow and above the white, upturned antenna dishes of the university's Owens Valley Radio Observatory.
His wife and mother-in-law, wrapped in a rug to ward off the February chill, stood beside him. Stifling yawns, cupping hot mugs of tea, they scanned the star fields over Bishop for the returning shuttle.