It was last year's "breakthrough of the year," according to the journal Science. Two teams of astronomers reported that the universe was not only expanding, but that the expansion was accelerating. The discovery could profoundly shape astronomers' notions about the nature of the universe. And it has tantalizing historical interest: It means that a mathematical error that Einstein called his biggest blunder might be correct after all.

New observations reported at a scientific workshop in Aspen, Colorado, however, raise some questions about the evidence that led to the discovery.

The discovery was made by observing similar supernovae, or exploding stars, discovered at different distances from Earth. The conclusion that the universe is accelerating rests on the premise that these supernovae are all approximately the same brightness. Astronomers call these Type Ia supernovae "standard candles." That means that changes in the brightness observed from Earth can be used to determine their distance--the dimmer they are, the farther away they are.

But Adam Riess and Alex Filippenko of the High-z Supernova Search Team at the University of California at Berkeley now report that nearby Ia supernovae are increasing in brightness more slowly than distant Ia supernovae. That could mean that the supernovae are not of equal brightness. That, in turn, would raise questions about the calculations that showed the universe is accelerating.

The notion of the expanding universe is related to a mathematical quantity Einstein devised called the "cosmological constant." He eventually discarded it, deciding it was a mistake. If the universe is accelerating, however, it means that Einstein's cosmological constant was correct.

Filippenko says he thinks the accelerating-universe idea will hold up, and here's why: First, the new report could be the result of some measurement discrepancy.

Secondly, it's possible that the supernovae all reach the same brightness eventually. "On the other hand, if the evolution means that the distant explosions are more powerful, then the case for acceleration becomes even stronger," Filippenko says.

Saul Perlmutter, head of the Supernova Cosmology Project at Lawrence Berkeley National Laboratory, says this is only one of a number of hints that have suggested differences among Type Ia supernovae. "So far, every time we've followed up one of these things, it's turned out we're not seeing anything different," he says. He is confident the accelerating-universe idea will hold up. "Scientists get excited when we see this," he says. "This is the kind of thing we all live for."