When the Senate killed the proposed Clear Skies bill in 2005, Bush's EPA turned to rule-making.

"We moved to Plan B," said Connaughton, Bush's senior aide.

That spring, Johnson signed two of EPA's most significant administrative rules. Both closely tracked elements of the failed Clear Skies agenda.

The first was called the Clean Air Interstate Rule (CAIR). It became a cornerstone of the Bush administration's environmental policies, and was based in part on EPA's successful Acid Rain program of the 1990s.

EPA said CAIR would have reduced total sulfur dioxide and nitrogen oxide emissions from power plants in 28 states by 70 percent. CAIR involved a market-based system called cap-and-trade that set a limit for the total amount of pollution from all plants.

Under cap-and-trade, factories whose pollution fell below EPA limits would earn so-called allowances. Plants whose pollution exceeded EPA limits could then purchase these allowances. In other words, utilities faced a choice: invest in expensive pollution control "scrubbers," or make arrangements to buy pollution allowances from a company that had.

The Bush administration said CAIR would trigger almost immediate pollution reductions, but environmental groups said that in reality the program would delay serious pollution controls for 10 to 20 years.

EPA projected that CAIR would prevent 17,000 premature deaths annually by 2015, and trigger related health benefits of $70 billion.

Power companies and others in industry did not relish spending billions to erect scrubbers, but they favored the Clear Skies element of CAIR that created the cap-and-trade program because it offered more flexibility and time to meet the coming regulations.

Environmental advocates thought CAIR set weak long-term goals but nonetheless were eager to see the scrubbers installed.

Shortly after Johnson signed CAIR in March 2005, he issued the second key EPA regulation, the Clean Air Mercury Rule.

That rule, which targeted the nation's coal-fired power plants, was far more controversial. Like the CAIR provision, it created a cap-and-trade market for mercury pollution. This alarmed environmentalists because it meant some communities would be more exposed to mercury, a dangerous neurotoxin, than others.

The EPA defended its mercury calculations and projections. They said the numbers were based on sound economic principles.

But an EPA inspector general's report said the EPA's process was flawed.

Investigators found that the EPA's political staff instructed career scientists to flatly set mercury levels at 34 tons nationally per year, instead of using existing best-technology data to determine an appropriate level.

Thirty-four tons, it turned out, was the amount of mercury power plants expected to emit anyway under the related CAIR rule. The EPA, the inspector general said, should try again, and make an "unbiased determination."

"It was so clearly an ass-backward way of doing it, so transparently political," said Temple University environmental law professor Amy Sinden. "That's why EPA has had the kind of record it's had in the courts."

Burnett, the former associate deputy administrator who worked on the mercury rule, conceded that the EPA's own lawyers had warned the agency that it was likely to lose in court.

"On the mercury rule, we were told in no uncertain terms about the legal vulnerability," Burnett said. "But the decision was to risk this, given the strongly desired policy outcome."

Queen of Hearts