In 1956, an oil-company geologist named M. King Hubbert made a bold prediction. U.S. oil production outside of Alaska would peak in the early 1970s and decline forever after. In 2001, Princeton University geophysicist Kenneth S. Deffeyes, a former colleague of Hubbert's, took the analysis a step further, predicting that world production would do the same sometime between 2004 and 2009. Three years later, in a lecture at a meeting of the American Geophysical Union, he asserted that the peak would come sooner rather than later. In fact, we may already be past the peak by the time you read this.

Then, in his 2006 State of the Union Address, President George W. Bush drew headlines by announcing that America is "addicted to oil" and that the time had come to do something about it. He pledged to invest in alternative energies (including ethanol and hydrogen fuels) and to reduce Middle Eastern oil imports by 75 percent by 2125. He also promised increased funding for alternative energy, including hydrogen-powered cars.

In combination, these announcements raise two alarming questions: (1) Are we about to run out of oil? (2) What are we going to do about it if we are?

Let's start with the "are we?"

The short answer is "no." We may run low, but not "out." That's because as oil gets scarcer, price rises will force us to cut back. At least a modicum will remain in the ground for a long time.

A better question is when we will be forced to cut back. Conventional wisdom says that we can figure out how much oil we have left by dividing the amount that remains by the annual rate of consumption, with fudge factors for anticipated growth, both in demand and supply. Right now the world has about a trillion barrels of known reserves and is using somewhere around 30 billion per year. That sounds like a thirty-plus year supply, but to Hubbertians it's not. There comes a point, they say, when the oil gets harder and harder to pump out of the ground. Even if there's several more years' supply there, we simply can't get it out fast enough.

To many people, this was Hubbert's primary contribution: the realization that the rate of extraction increases, peaks, and then declines, long before the oil is totally gone. In any given oil field, there's initially a lot of oil available, but the field is underdeveloped. Then come more wells and increased production--until yields begin to decline no matter how many technological improvements you make.

All of that makes intuitive sense: There's "easy oil" and "hard oil," and by definition, the easy oil comes out of the ground faster. But Hubbertians take it a step further and apply the same analysis to the entire world. Initially, there's a lot of oil available, but extraction techniques are crude. There's not much demand yet, either. Then as technologies adjust, demand soars, mass-production becomes feasible, and the rate of extraction skyrockets. Until, that is, you can no longer keep up.

Hubbert asserted that this process more or less follows a bell curve. His reasoning was complex (one 1982 journal article covered 125 pages), but once you've said "bell curve," you know that peak consumption comes at the middle, when you've used up half of the world's total endowment.

Remember those trillion barrels of reserves? Well, we've already used 900 billion (a bit more by the time you read this), so if that's all there is, we're getting close to the midpoint. Once we get there, Hubbertian theory says, our oil-consuming lifestyle will change, like it or not. Worse, it may change pretty quickly. David Goodstein, vice provost at California Institute of Technology, estimates that if Hubbert's peak hits today, we'll slide down the other side of it so quickly that world oil production will be halved within a decade. Deffeyes projects a 90 percent decline by 2019. That doesn't give us a lot of time to adjust.

"We have a civilization firmly rooted in a mathematically impossible promise of an endless supply of oil," Goodstein said in late 2004. "The oil will run out. The only question is when."