Documents released under a court order show that a government investigator studying the stranding of 37 whales on the North Carolina coast last year changed her draft report to eliminate all references to the possibility that naval sonar may have played a role in driving the whales ashore.

The issue of sonar's effects on whales is a sensitive topic for the U.S. Navy. It has clashed with environmentalists in several court suits seeking to limit use of the technology because of its possible effects on marine mammals and other sea creatures.

The January 2005 stranding occurred shortly after naval maneuvers in the area -- which is off North Carolina and in the region where the Pentagon wants to build a controversial underwater sonar training range.

In her initial April 2005 preliminary report on the deaths, Teri Rowles, coordinator of the National Marine Fisheries Service's stranding response program, described injuries to seven of the whales that "may be indicative" of damage related to the loud blasts of sound from active sonar.

She also noted that one of the injuries -- air bubbles in the liver of a pilot whale -- had been reported in mass strandings in the Bahamas and Canary Islands associated with sonar activity.

That report was made public this week after a federal judge in New York ordered its release to the Natural Resources Defense Council (NRDC), an environmental group, which had sued the agency over its refusal to release information on the whales' stranding on North Carolina's Outer Banks.

But before it was released by NRDC, the National Oceanic and Atmospheric Administration released an updated report -- by Rowles and others -- that did not mention sonar. In a cover letter to that report, NOAA officials said the initial draft that mentioned sonar "contains early information that was later found to be inaccurate."

NRDC attorney Andrew Wetzler said that the second report "seems a lot more like spin than science." He said the absence of any reference to sonar was surprising because the evidence suggesting that sonar might have played a role hardly changed between the first and second drafts. What changed, he said, was some limited analysis by Rowles.

In an interview yesterday, Rowles said the references to sonar were removed because it was just one of several possible causes of the strandings. "Sonar has not been implicated or eliminated -- it remains one of many possible causes," she said. "We wanted to put out a report that included our most up-to-date information."

Most important, she said, was the conclusion after further analysis that the presence of air bubbles in one animal's liver had not been conclusively confirmed. Air bubbles were found in the organs of several whales that stranded in the Canary Islands after a sonar exercise, leading some researchers to conclude that the animals swam to the surface too rapidly and suffered a version of the bends. If air bubbles were present in the whales that beached in North Carolina, it could suggest that sonar caused their stranding, as well.

The federal court order to release the report came at an awkward time for NOAA and the Navy, which has been holding public hearings on its controversial plan to build an underwater sonar training range.

The public record on that issue will close at the end of the month, and some activists have complained that officials are trying to withhold information about the stranding until after that time. In its court filings, NRDC argued that it was unfair to complete the hearings before information about the strandings was released.

Navy officials say that the sonar training range is essential, and that active sonar is increasingly important because of a growing threat from diesel submarines that cannot be detected using traditional passive sonar.

The Navy has also acknowledged that sonar can harm whales. A Navy-NOAA investigation found that sonar from Navy ships was the most plausible explanation for the stranding of 17 whales in the Bahamas in 2000. The report found that sonar-induced damage to the ears of some animals may have disoriented them and caused them to swim onto the shore.

Researchers are also studying the ears of some animals that stranded in North Carolina, but Rowles said those results will not be known for some time. The final report, she said, is scheduled to be released in March.

© 2006 The Washington Post Company