The Alberta government asserts that the river is not being contaminated—that anything found in the river or in its delta, at Lake Athabasca, comes from natural bitumen seeps. The river cuts right through the oil sands downstream of the mines, and as our chopper zoomed along a few feet above it, McEachern pointed out several places where the riverbank was black and the water oily. "There is an increase in a lot of metals as you move downstream," he said. "That's natural—it's weathering of the geology. There's mercury in the fish up at Lake Athabasca—we've had an advisory there since the 1990s. There are PAHs in the sediments in the delta. They're there because the river has eroded through the oil sands."

Independent scientists, to say nothing of people who live downstream of the mines in the First Nations' community of Fort Chipewyan, on Lake Athabasca, are skeptical. "It's inconceivable that you could move that much tar and have no effect," says Peter Hodson, a fish toxicologist at Queen's University in Ontario. An Environment Canada study did in fact show an effect on fish in the Steepbank River, which flows past a Suncor mine into the Athabasca. Fish near the mine, Gerald Tetreault and his colleagues found when they caught some in 1999 and 2000, showed five times more activity of a liver enzyme that breaks down toxins—a widely used measure of exposure to pollutants—as did fish near a natural bitumen seep on the Steepbank.

"The thing that angers me," says David Schindler, "is that there's been no concerted effort to find out where the truth lies."

Schindler, an ecologist at the University of Alberta in Edmonton, was talking about whether people in Fort Chipewyan have already been killed by pollution from the oil sands. In 2006 John O'Connor, a family physician who flew in weekly to treat patients at the health clinic in Fort Chip, told a radio interviewer that he had in recent years seen five cases of cholangiocarcinoma—a cancer of the bile duct that normally strikes one in 100,000 people. Fort Chip has a population of around 1,000; statistically it was unlikely to have even one case. O'Connor hadn't managed to interest health authorities in the cancer cluster, but the radio interview drew wide attention to the story. "Suddenly it was everywhere," he says. "It just exploded."