When a Senate committee asked Interior Secretary Gale A. Norton questions about caribou in the Arctic National Wildlife Refuge, she sought answers from the agency in her department that runs the refuge.

But when Norton formally replied to the committee, she left out the agency's scientific data that suggested caribou could be affected by oil drilling, while including data that supported her case for exploration in the refuge, documents show. Norton also added data that were erroneous, stating that caribou calving has been concentrated outside the proposed drilling area in 11 of the last 18 years, when the opposite is true.

The Arctic refuge was already the first issue to pierce the bipartisan consensus that prevailed on Capitol Hill since last month's terrorist attacks, with drilling proponents pushing for a vote on national security grounds and opponents arguing for delay.

Norton's behind-the-scenes rebuff of the Fish and Wildlife Service -- which is the Interior Department's front-line environmental agency but is not yet staffed by Bush administration political appointees -- is likely to heighten tensions over the nation's most disputed patch of tundra.

Norton spokesman Mark Pfeifle said her error was simply that -- an inadvertent substitution of "outside" for "inside" -- and noted that she has preached peer-reviewed science ever since she got to Washington.

As for other disparities between her response to Congress and Fish and Wildlife's proposed response, Pfeifle said, she relied on external input as well, especially a peer-reviewed caribou study from the Wildlife Society Bulletin concluding that oil development has little impact on caribou. That study's acknowledgments thank the oil company BP Exploration for funding, but Pfeifle said that makes no difference.

"One of the cornerstones of this department is to reach out and listen to a wide variety of people and sources to determine the best information and the best policy," Pfeifle said. "Sometimes we look for guidance inside the department; sometimes we look outside as well."

A Fish and Wildlife spokeswoman referred all questions to Pfeifle.

President Bush's nominee to head the agency, Steve Williams, had his confirmation hearing on Wednesday, so the agency should get political direction soon. For now, Norton and her aides oversee the bureaucracy, and the two sides clearly approach the Arctic refuge from different perspectives.

Fish and Wildlife is a purely environmental agency that does not concern itself with economic development or America's dependence on imported oil; it's no secret that most of its employees oppose industrial development in the refuge. Norton is a conservative Cabinet member who has led the administration's pro-drilling campaign.

Still, some Fish and Wildlife officials said that they were careful to send Norton a complete and balanced portrait of the science, regardless of personal biases, and that she cherry-picked the data that suited her.

"If Congress is going to have a serious discussion on the future of the Arctic refuge, it ought to have the whole story, not a slanted story," said one agency employee, who requested anonymity. "We tried to present all the facts, but she only passed along the ones she liked. And to pass along facts that are false, well, that's obviously inappropriate."

The disagreement has its roots in a series of questions posed to Norton on May 15 by Sen. Frank H. Murkowski (R-Alaska), chairman of the Energy and Natural Resources Committee and a fierce advocate of drilling in what is called the 1002 area of the refuge's coastal plain.

The four questions related to caribou -- two about calving habits, two about the potential impact of oil development -- were referred to Fish and Wildlife. On May 24, biologists at the refuge provided their recommended answers, which were approved by five offices up the agency's chain of command without substantive changes, agency sources said.

When the answers arrived at Norton's office, they were rewritten with input from her congressional affairs director, David Bernhardt, and her senior counselor, Ann Klee. The before-and-after documents were provided to The Washington Post by Public Employees for Environmental Responsibility (PEER), which has battled Norton since her nomination.

The scientific data the agency gave Norton offered fuel for both sides of the debate. The bottom line was that caribou gravitate toward the 1002 area -- and that they tend to avoid oil fields when calving -- but that they have mostly prospered over the last three decades despite the development of sprawling oil infrastructure.

On the calving questions, Fish and Wildlife did suggest that the 130,000-strong Porcupine Caribou Herd prefers to calve in the 1002 area, which is about as far as it can get from predators and which is rich in nutritious food. The agency noted that the herd has calved there in 27 of the last 30 years, a statistic Norton did not use.

But Norton did use the agency's statement that the migratory Porcupine do not concentrate most of their calving in the 1002 area in years of late snowmelt, because they can't reach it in time. Then she added that "concentrated calving occurred primarily outside the 1002 area in 11 of the last 18 years," when she should have said "inside."

Norton also neglected to mention Fish and Wildlife's data illustrating that "calf production and early survival of calves are lower in such years" when the caribou do not calve in the 1002 area.

"The long-term evidence shows that's where the caribou want to be. But I guess if you're going to be an advocate, you're not going to mention that if it doesn't help your case," said another Fish and Wildlife official, also on the condition of anonymity. "They're not denying facts. Just ignoring them."

On the oil questions, Fish and Wildlife reported that the Central Arctic Herd has grown dramatically since the onset of oil field development in its historical calving grounds, from an estimated 5,000 in 1975 to about 27,000 today. It also included two sentences noting that since the development of the Kuparuk field in the early 1980s, the oil industry has been more sensitive to caribou, sentences that Norton copied into her response.

However, the agency again included several caveats, warning that the consequences of further development would be "very difficult to predict." It pointed out that in the harsh-weather years from 1988 to 1994, the overall caribou numbers declined, and that the herd's birth rate was 64 percent in developed areas and 83 percent elsewhere.

Norton included no such caveats, stating that the "data do not support the hypothesis that oil fields adversely affect caribou productivity."

Pfeifle said the Interior officials believed that Fish and Wildlife had relied too heavily on one source. He said the officials also objected to statements in the agency's proposed answers that included the words "suggesting" and "apparently," because they indicated too much doubt.

The Wildlife Society Bulletin article also included such phrases as "as suggested previously," "we believe" and "seems most likely," but Pfeifle emphasized that it was peer-reviewed first.

Environmentalists have never been fond of Norton. The former Colorado attorney general endured countless comparisons to former interior secretary James Watt during her confirmation hearings, and yesterday PEER's national field director, Eric Wingerter, called on her to resign. He said the inside-outside error sounded too conveniently pro-drilling to be an accident.

"This went way beyond spin," Wingerter said. "They manipulated the data in an attempt to manipulate Congress. Norton's big mistake here was getting caught."

But there are often tensions between political appointees and the bureaucracies beneath them, and there is always a fine line between strengthening an argument and twisting the facts.

"I'm not sure it's a perversion of the facts," a third Fish and Wildlife official said. "It's just selective of the facts. I guess it all depends what the meaning of 'is' is."

Interior Secretary Gale A. Norton erroneously testified that caribou calving has been concentrated outside the proposed drilling area in 11 of the last 18 years.The Fish and Wildlife Service reported to Norton that its data suggested caribou in the ANWR could be affected by oil drilling. Norton did not use that information but did include data supporting the case for exploration in the refuge.