Imagine a particle of light, zipping along at 186,000 miles a second on its trip from Earth past the constellation Virgo into the great infinity of the universe. It has been on this trip for a few billion years when wham! All of a sudden it’s headed back towards Earth without pulling a U-turn.

A footnote in a paper written by Penn astrophysicists Max Tegmark and Angélica de Oliveira-Costa raises this possibility.

The footnote in question speculates that the universe is not infinite and isotropic—the same in all directions. It may well be finite and narrower in one direction than in others—a giant cosmic doughnut, if you will.

They doubt it, but that hasn’t stopped scientists and media from The New York Times to the BBC from picking it up and running with it.

The possibility arose when Tegmark, assistant professor of physics and astronomy, and his wife Oliveira-Costa, research assistant professor of physics and astronomy, analyzed data from the Wilkinson Microwave Anisotropy Probe (WMAP), a satellite that measures microwave background radiation released not long after the Big Bang, some 13 billion years ago.

Tegmark and Oliveira-Costa set out to scrub the static from the data the WMAP satellite produced. “You have lots of emissions from our galaxy, and to make an accurate measurement, you have to remove [those] emissions,” Oliveira-Costa said. “When we processed those, we got an interesting result.”

What they found when they mapped the data was that the distribution of hot and cold spots in the universe was asymmetric. “It was definitely less clumpy in one direction, towards the constellation Virgo,” Tegmark said.

They didn’t think much of this finding. “We wrote a paper saying, Here’s the best map of the microwave background. We then put a footnote at the end saying, Here’s an interesting puzzle that people might want to solve.”

Neither Tegmark nor Oliveira-Costa believe the doughnut theory will pan out. “My Ph.D. thesis was trying to prove that we don’t live in a doughnut,” she said. “We were using data from COBE”—the Cosmic Background Explorer, the satellite that produced the first usable microwave background data in 1992—”to rule out the model, but people have this intrinsic fascination with the model.”

Tegmark said, “The map will be useful no matter what, because we have a really clean map of what’s out there, and other people will use it for things we haven’t really thought of.

“As for the puzzle, we’re going to have to ask if there really is a puzzle. Another possibility is, maybe it’s just a fluke. [And] if there is something funny going on in space on a larger scale, it doesn’t have to be the doughnut theory. There are more complicated theories, such as pretzels, and a lot of people are going to be keeping themselves busy over the next year analyzing the data. But the smart money’s on there not being a doughnut.”