A massive plume of oil trapped deep underwater in the Gulf of Mexico may not have broken down as rapidly as some researchers predicted, according to new research published today.

Appearing in the journal Science, the survey -- the first published, peer-reviewed study of the subsurface plume -- traced an invisible, finely diffused swath of oil nearly the size of Manhattan. Researchers also found there was not an "appreciable" surge in bacteria growth and oil consumption in the plume.

Woods Hole Oceanographic Institution scientists traced under-surface oil that stretched a mile wide and snaked at least 20 miles southwest of the crippled BP PLC well. The plume may have stretched even farther, but the scientists had to cut short their late-June research journey in the face of hurricane threats before they found the end of the plume.

To track the 3,000-foot-deep plume, the researchers used a robot submarine that recorded information about the water's biological and chemical composition. They also gathered water samples and made deepwater measurements using instruments cabled to a ship.

The study's release comes in the midst of battle over just how much oil from the massive spill is left in the Gulf, what the long-term effects of the spill might be on the ecosystem and the accuracy of recent federal estimates that the oil may be quickly dissipating (see related story).

A team of government and independent scientists released a report earlier this month that estimated nearly three-fourths of the oil had been captured, skimmed, burned, dispersed or broken down by natural processes. But other scientists have questioned those numbers. University of Georgia scientists said this week that 79 percent of the oil has not yet been recovered and remains a threat to the ecosystem.

Today's report found the spill did not fuel an "appreciable" surge in bacteria growth, a finding that could indicate there is more oil below the surface. But the Woods Hole researchers sought during a Washington news briefing to deflect questions about the overall oil budget, noting their paper only shows the existence of a plume but makes no conclusions about the amount of oil in the Gulf or in the plume itself.

"What we have done is an estimate -- science is incremental," said Richard Camilli, a lead scientist on the Woods Hole team. "We don't know how much oil was in it or the plume length. ... Later, we may see if this was a penny in a very big checking account."

He added, "We have to balance patience with urgency."

At best, the study is a snapshot that reveals the migrations of oil that had escaped from BP's Macondo wellhead and diluted into deep water over five days. The study does not speak to current conditions in the Gulf, and researchers are uncertain how the environment has shifted since BP plugged the leak a month ago.

But the study does gather the first hard data about the plume.

During a House hearing today, Ian MacDonald, an oceanography professor at Florida State University, called the report "some of the best science I've seen yet out of this process."

Reporters Paul Voosen and Katie Howell contributed.