nly a year ago, Michael A. Bellesiles was well on his way to becoming an academic superstar. He had just published a book with a startling thesis: very few people owned working guns in colonial America. Stepping into the ferocious national debate over guns and the meaning of the Second Amendment, Mr. Bellesiles, a history professor at Emory University in Atlanta, caused a sensation. Legal scholars said his prize-winning book could influence federal court cases challenging gun laws; gun-control advocates championed the research as proof that America's gun culture is, as Mr. Bellesiles put it, "an invented tradition"; angry gun owners saw it as an insidious attack, a calculated effort to prove that the Constitution's framers could not have intended the "right to bear arms" to apply to individuals if so few people owned them.

Now many of Mr. Bellesiles's defenders have gone silent. Over the past year a number of scholars who have examined his sources say he has seriously misused historical records and possibly fabricated them. They say the outcome, when all the evidence is in, could be one of the worst academic scandals in years.

Mr. Bellesiles (pronounced buh-LEEL) has denied that the errors in "Arming America: The Origins of a National Gun Culture" are more serious than the ones found in any lengthy and serious work of scholarship, and he has repeatedly said the attacks against him are politically motivated. Mr. Bellesiles, who owns five guns and likes to shoot skeet and target-shoot in his spare time, said he never intended his book to become a cause célèbre for gun control advocates. "When I saw that the flap copy said, 'This is the N.R.A.'s worst nightmare,' I was horrified," he said. "I feel like I'm a historian who accidentally stepped into a minefield."

Indeed, after the National Rifle Association alerted its members about the book, Mr. Bellesiles said, he began receiving hate mail and threats by phone, e-mail, fax and letter. He was forced to get an unlisted number and to change his e-mail address, he said. Earlier this year, two American historical societies passed special resolutions condemning the harassment.

Without doubt, Mr. Bellesiles's research would not have received such careful scrutiny if he had not stepped into the politically and ideologically charged struggle over guns. Yet the scholars who have documented serious errors in Mr. Bellesiles's book — many of them gun-control advocates — do not appear to have any sort of political agenda.

They were struck by his claim to have studied more than 11,000 probate records in 40 counties around the country. He found that between 1765 and 1790, only 14 percent of estate inventories listed guns, and "over half (53 percent) of these guns were listed as broken or otherwise defective." Those claims are featured prominently in the book and were cited in many positive reviews as the core of its argument.

But those who tried to examine the research soon found that they could not, because most of Mr. Bellesiles's records, he said, had been destroyed in a flood. The records they could check showed an astonishing number of serious errors, almost all of them seemingly intended to support his thesis. In some cases his numbers were off by a factor of two, three or more, said Randolph Roth, a history professor at Ohio State University.

To use one example: in his book, Mr. Bellesiles writes that of 186 probate inventories from Providence, R.I., recorded between 1680 and 1730, "all for property-owning adult males," only 90 mention some form of gun, and more than half the guns were "evaluated as old and of poor quality."

At least three scholars have independently examined the same archive and found that 17 of the estates in question were owned by women; that some estates lacked inventories, and that of those that had them, a much higher percentage than Mr. Bellesiles reported contained guns; and that only 9 percent of the guns were evaluated as old and of poor quality.

"The number and scope of the errors in Bellesiles's work are extraordinary," Mr. Roth said. They go well beyond the probate record data, he added, affecting Mr. Bellesiles's interpretation of militia returns, literary documents and many other sources.

Confronted with serious errors in his research, Mr. Bellesiles has acknowledged that there are problems with the way he used probate record data, and he even made some changes in the paperback edition that came out earlier this year. But he said that the data were only a small part of the book. "I wish I had taken them out entirely," he said.

Jack Rakove, a Stanford University historian who has been supportive of "Arming America," agreed: "The book raises a host of interesting questions about the role firearms have played in American life and culture, and it goes well beyond the probate data."

But Mr. Rakove conceded that he had not looked at the research that has been questioned, and he said it was important that Mr. Bellesiles respond to his critics more fully than he has so far.

Mr. Bellesiles's failure to explain himself has led to the most serious accusations against him, which were outlined in The Boston Globe this fall. Earlier this year, when the criticism of his book became more intense, he asked Mr. Roth to help him defend himself. Mr. Roth wrote back, saying that if Mr. Bellesiles would tell him what records he looked at in Vermont, he would go to the archive on his own time, and that if the records matched, he would defend him. Mr. Bellesiles never responded to that offer, Mr. Roth said.

Those who have pressed him hardest for details say they have been led on a bizarre scholarly car chase, with Mr. Bellesiles offering new memories about where he got his records as soon as the old ones were discredited.

He has said from the start that he took notes on the thousands of colonial-era probate records with tick marks in pencil on yellow legal pads. That fact alone was surprising to many of his fellow historians, who tend to use a database when working with such large amounts of information.

Almost all of those notebooks were destroyed when his office at Emory was flooded in May 2000, Mr. Bellesiles said.

James Lindgren, a professor at Northwestern University Law School and by far the most thorough of Mr. Bellesiles's critics, asked him last year where he had done his research on probate records. Mr. Bellesiles responded with a number of locations, including the San Francisco Superior Court, where he said he had found probate records from the 1850's.

Mr. Lindgren, who has done extensive work in probate data, called the courthouse and was told that all the records for that decade were destroyed in the 1906 earthquake and fire. They were not available in two other Bay Area libraries, either. Mr. Bellesiles now says he must have done the research somewhere else and cannot remember where.

But Kathy Beals, former director of the California Genealogical Society, who has worked extensively with probate records from that era, said: "Nobody knows of those records being in existence, and if they are, there are hundreds of people who would like to look at them."

In September, Mr. Bellesiles offered a new defense. Mr. Lindgren and a reporter from The Globe, David Mehegan, found additional serious errors on Mr. Bellesiles's Web site, where he had been posting probate records in an attempt to replace what he said had been lost in the flood. He conceded the errors and responded to The Globe, and later said someone had altered his Web site, presumably a computer hacker.

But several scholars, including one of Mr. Bellesiles's colleagues at Emory, said they doubted that story. Robert A. Paul, the interim dean at Emory College, said, "I can neither independently confirm nor deny that Professor Bellesiles's Web site was hacked."

In September, James Melton, the chairman of the Emory history department, asked Mr. Bellesiles to write a "reasoned, measured, detailed, point by point response to your critics" in an appropriate professional forum. Mr. Bellesiles did publish a response in the November issue of the Organization of American Historians newsletter, but it focused on harassment rather than charges of serious misconduct.

Mr. Bellesiles's supporters have said they expect a fuller response to emerge in a special issue of the William and Mary Quarterly to be published next month.. A draft of the lengthy response Mr. Bellesiles wrote for that issue, supplied by the journal's editor, concedes some mistakes and challenges others, but leaves many serious errors unaddressed.

It is not clear what will happen to Mr. Bellesiles or his book if the scholarly community reaches a consensus that "Arming America" is a seriously flawed or even fraudulent book. The Emory College dean, Mr. Paul, said, "If there were scholarly fraud, we would take that very seriously." Alan Brinkley, the chairman of the history department at Columbia University, said similar questions had never been raised about a book that had won the prestigious Bancroft Prize in American History and Diplomacy. Although there has been no discussion of disciplining Mr. Belles iles or revoking the prize, a spokesman for Jonathan R. Cole, the provost and dean of faculties at Columbia University, said he had distributed copies of the documents detailing Mr. Bellesiles's mistakes to this year's three Bancroft jurors and asked them to examine it.