The Environmental Protection Agency is preparing to take "formal action against DuPont soon" over allegations that the giant chemical company failed to report possible health problems connected with a key ingredient used in making Teflon, EPA spokeswoman Cynthia Bergman said yesterday.

For more than a year, EPA has been investigating whether the company violated the Toxic Substances Control Act, which requires companies that produce chemicals to report to EPA any information indicating that a substance it uses "presents a substantial risk of injury to health or the environment."

Also yesterday, an advocacy group released internal DuPont e-mails obtained by lawyers suing the company in West Virginia over claims that the chemical, perfluorooctanoic acid (also called PFOA or C-8), had contaminated drinking water in two communities near a DuPont plant. In the e-mails released by the Environmental Working Group, a DuPont lawyer expressed concern that the company had underestimated the health risks associated with C-8 and then failed to fully report them to local water officials.

Perfluorooctanoic acid is a soaplike chemical that serves as an essential processing agent in making stain- and stick-resistant surfaces and materials.

DuPont officials deny any health problems linked to C-8 and said they are vigorously contesting the class action lawsuit.

But company documents indicate that DuPont began exploring the possible hazards as far back as 1981, when animal studies indicated C-8 could cause cancer and birth defects in rats. DuPont began conducting tests on female workers handling C-8 and discovered that two female employees who worked with C-8 had given birth to children with defects similar to those reported in the animal tests.

That year, DuPont told the West Virginia Division of Water Resources that it had reassigned "female personnel of childbearing capability to areas outside those in which" C-8 was being made and handled, but it did not mention that the chemical had been found in the area's groundwater. They did not inform EPA about the reassigned workers.

Company officials said that EPA requires companies to report suspected health risks associated with a chemical only if they can ". . . reliably ascribe the effect to the chemical," and that there was no such link between the two children's birth defects and C-8.

Also in 1981, according to documents provided by DuPont, the company told a regional EPA office that low levels of C-8 had been detected in a local aquifer.

Three years later, the company detected the chemical in tap water taken from a store in Little Hocking, Ohio, which is across the Ohio River from the Washington Works plant, outside Parkersburg, W.Va., but it did not inform local or federal officials.

DuPont spokesman Clif Webb said yesterday that follow-up tests failed to detect any C-8 in the water and that the company did not test again until 2002, after it entered into a consent agreement with the West Virginia Department of Environmental Protection.

In the agreement, it promised to provide drinking water if levels of C-8 in the water exceeded specified thresholds.

Environmental Working Group Senior Vice President Richard Wiles, whose group pressed EPA to begin investigating DuPont in 2003, said the company should be penalized for not fully disclosing possible dangers associated with C-8.

"They're willing to lie to the community," Wiles said. "The administration needs to enforce the law to the maximum extent possible, because the company is clearly a repeat violator."

In December 2001, DuPont lawyer Bernard J. Reilly wrote in an e-mail to his son that the company had learned that its method of detecting C-8 in water "has very poor recovery, often 25 %, so any results we get should be multiplied by a factor of 4 or even 5. However that has not been the practice, so we have been telling the agencies results that are surely low. Not a pretty situation, especially since we have been telling the drinking water folks not to worry . . . Ugh."

DuPont officials said yesterday they were confident they had complied with federal law and had taken appropriate precautions in disposing of C-8, cutting releases by 98 percent since 1999.

"We're not aware of any evidence of any health effects associated with PFOA," said Robert Rickard, DuPont's chief toxicologist. He said that while DuPont took women who could become pregnant off the C-8 production line in 1981, they were allowed to return after the company determined that the animal studies indicating birth defects had been flawed.

But that has failed to convince residents near the Washington Works plant, many of whom have become embroiled in litigation with DuPont.

Della Tennant, who lives in Parkersburg and settled a lawsuit with the company in 2001, said: "My whole family's been affected by it. I believe the C-8 was one of the factors . . . How are we going to know how much that contaminated water has affected our family?"

She said she has lost several head of cattle that grazed near the company's Dry Run Landfill and had suffered from respiratory problems she blamed on contaminated water and eating beef from her cattle. Her high-school-age daughter's gums decayed and required skin grafts, she said.

Bob Griffin, general manager of the Little Hocking Water Association, whose facility serves 12,000 area residents, said, "We're just trying to find a solution."

Griffin told EPA officials at a meeting last year: "Twenty years ago, DuPont knew C-8 contaminated our water, but we didn't find out about it until about a year-and-a-half ago. Since then, we have learned that C-8 not only contaminates the aquifer from which we pump our drinking water, but it also contaminates our soil and the air we breathe."