Among the icy crevasses of Antarctica, Mark Devlin lost a little white box that might have contained the best work of his career.
The box, no bigger than a PC, had been recording images from a special telescope aboard a helium-filled balloon that Devlin and colleagues had built in West Philadelphia and that had flown high above the icy continent.
Devlin thought he might have captured evidence for a sort of hidden universe - a vast collection of stars and galaxies that no one had seen before. Such a finding could challenge astronomers' view of the universe and how it came to be.
But something went terribly wrong. After the landing, an errant parachute dragged the telescope more than 100 miles, scattering its invaluable pieces over some of the planet's most treacherous terrain. Any potential discoveries were probably lost.
Devlin and his colleagues at the University of Pennsylvania felt their careers blowing away with the Antarctic wind.
Finding it seemed unlikely, but Devlin had no intention of giving up.
It all started with a mysterious glow that several NASA satellites detected in the 1990s.
The glow was not visible to the eye, but came in a hard-to-observe wavelength known as sub-millimeter, somewhere between radio and infrared.
Human eyes are tuned to a narrow band of light wavelengths - the visible range - but the universe shines across a much wider spectrum. Radio waves, microwaves, infrared, ultraviolet, and gamma rays each reveal a different set of phenomena - a different face of the cosmos.
Sub-millimeter waves could potentially reveal objects otherwise obscured by dust, which absorbs visible light. Devlin and other scientists wondered if that mystery glow was coming from galaxies buried in cosmic dust.
To find out, they'd need a telescope designed to take images in this sub-millimeter wavelength.
Since the earth's atmosphere blocks sub-millimeter waves, the only way to detect them would be to get above it. NASA had plans to do this with a satellite, but Devlin had a different idea: He thought he'd do it faster and cheaper with a helium balloon.
Scientific balloons can be as big as football fields and can fly up to 125,000 feet high. That would be good enough to loft a telescope above 99.7 percent of the atmosphere.
The white box was lost from Devlin's second attempt to capture the sub-millimeter universe. In 2000, he began building an earlier version of the project - dubbed BLAST, for Balloon-borne Large-Aperture Sub-millimeter Telescope.
Much of it was assembled in the Left Bank building in West Philadelphia. It had a 6-foot-diameter carbon fiber mirror - almost as wide as that flown on the Hubble Space Telescope. The apparatus stood 25 feet high and weighed two tons.
It was launched from a balloon facility in Sweden. But the mirror absorbed water, which froze, distorting the view. Then it landed hard in a remote part of Canada. "It was sad," Devlin said. "The whole thing was in tiny pieces on the ground."
As team leader, Devlin said, he couldn't show signs of despair. He had too many colleagues and students who had put too many hours into this.
"If you're a principal investigator and something gets screwed up, you have to forge a path," he said. So they recovered whatever pieces they could and started rebuilding.
When it was finished in 2006, they launched it from a balloon facility in Antarctica.
The frozen continent was perfect, he said, because in the southern summer, there's 24-hour daylight and a reasonably constant temperature to keep the balloon at a steady altitude. And the winds go predictably around the pole. Devlin's brother Paul came with him to make a documentary.
The launch was just before Christmas. The team happily watched the balloon disappear over the horizon. Everything seemed to be working. It stayed aloft for 12 days, making a loop around the South Pole.
Devlin flew home. Then, around New Year's Day, he got a call.
It was the head of the balloon facility, saying he was sorry but, on landing, the parachute had failed to drop off and was dragging the telescope along the ground.
Later, Devlin learned that the wind had dragged it 120 miles, leaving pieces of the telescope over the whole treacherous path.
The empty gondola finally came to rest, wedged in a crevasse.
Somewhere in that path was the data they had collected. The data were stored in something like a black box, he said, except that it was painted white - a mistake in retrospect.
NASA and the National Science Foundation pitched in with a search plane. After what Devlin recalls as three or maybe four tense days, the pilot finally spotted something near the end of the trail, not far from the crevasse that finally caught the frame.
Devlin calls it pure luck. A more religious colleague saw it as divine intervention.
The hardware was demolished but the hard drives looked OK. He sent them to a recovery company. And back came a new perspective on the universe.
Revealed for the first time was a previously dust-shrouded population of exotic galaxies radiating from the early universe.
Called starburst galaxies, they crackle with the production of hundreds of new stars a year - prolific compared to the three stars our galaxy makes in the same time.
If you lived in a starburst galaxy, said Devlin, you'd see many more of the bright stars that make up Orion and other prominent constellations.
These galaxies really shouldn't be there according to the favored theory of cosmology, said Ian Smail, an astronomer from Durham University in the north of England.
Because they are looking so far out in space, they are also seeing back in time, more than halfway back to the big bang. And at that stage, there should be galaxy fragments, not big, bright galaxies.
The findings were officially released today in the journal Nature. "This is one of those big-impact things," said Smail. "They're my competition, and this field is very cutthroat, but at some point you have to say they've done something pretty unique."
"For the theorists, it's a big headache," Smail added. "But that's good. It's one of those creative conflicts that really drives the field."
Devlin said that if the pilot had failed to find the box, he would have asked the National Science Foundation to send out snowmobiles. He would have gone out himself but said the authorities wouldn't allow it since they don't like people to die driving into crevasses.
The next step will be to verify his data.
That could happen soon, as NASA expects to launch its long-planned sub-millimeter satellite, called Hershel, in May. Devlin said his smaller project had to come first to have an impact.
If they hadn't turned up that white box, there would have been no third chance.
by Paul Devlin.
When: Wednesday from 3:30 to 5:30 p.m.
and from 6 to 8 p.m.
Where: Cohen Auditorium, Claudia Cohen Hall, 249 S. 36th St., on the University of Pennsylvania campus.
Cost: Admission is free.
Web: View the trailer at www.blastthemovie.com EndText