If you read High Country News, you probably know the statistics: Global greenhouse gas emissions have increased 70 percent since 1970, and our energy-squandering ways are to blame. Coal-fired power accounts for nearly a third of all energy-related greenhouse gas emissions in the U.S.; natural gas power another 10 percent. If our emissions continue to climb, the oceans will rise and submerge island nations, species will go extinct, and millions of acres of fertile soil will turn to punishing dust.

We may be able to avert the worst of these consequences, say a team of scientists in the journal Nature, if we cut our carbon emissions to half of what they were in 1990. In the U.S. alone, that means reducing carbon by 3.5 billion metric tons, the equivalent of retiring all of our coal-fired and natural gas power plants.

In many environmental circles, these facts have been used to support a simple argument: Anything done to displace coal-fired power is good. We need big solutions and massive investments in energy projects, centralized and large like the old ones, only cleaner. Small solutions will take too long.

And maybe that's true. A 2007 Department of Energy report estimates that 12 million independently owned small onsite generators, or "DG units," currently operate in the U.S., with a combined capacity of only 2,000 megawatts. Roughly half of those megawatts are from solar photovoltaic cells, which produce electricity by absorbing the sun's photons and liberating electrons. In a country with a little more than 1 million megawatts of generating capacity -- slightly more than 90 percent of it fossil-fueled or nuclear -- those numbers seem hardly worth discussing.

But even that report, which Congress ordered the Energy Department to complete in accordance with the 2005 energy bill, finds few reasons not to scale down more of our power and move it closer to home. The authors argue that the days of "economies of scale," as they apply to electrical power generation, are over: The improved technology and efficiency of small-scale systems means that it no longer costs less per kilowatt to build giant plants.

That's not to say that we don't need large-scale renewable energy to help replace fossil fuels. Electricity generated at big wind farms costs 5 cents per kilowatt-hour, well below the national retail electricity average of 10 cents; with production tax credits, the price drops to 3 cents per kilowatt-hour, which is hard to beat. Photovoltaic solar still costs 30 cents per kilowatt-hour compared to 15 cents per kilowatt-hour for large-scale concentrating solar power (CSP), which uses sun-tracking mirrors to focus the sun's energy, and can be scaled up to hundreds of megawatts.

Nor does it mean that distributed energy alone can solve all, or even most, of our energy problems. The precise benefits of millions of small generators infiltrating the U.S. power supply are difficult to predict, even for the Energy Department. (Doing so would require "a complete dataset of the operational characteristics for a specific site," and success would be "highly improbable.") But to trivialize distributed generation's potential is to assume that our problems stem only from the fuel we use to make electricity, not from the model we use for generating it.

There are many indications that large solutions will take longer than small ones. Even leaving aside the land-use issues and lengthy permitting processes that may bog down large-scale renewable projects for years, the Midwestern Independent System Operator estimates that just to bring wind energy from the Great Plains to cities will require building $80 billion worth of transmission lines, amounting to more than half the circumference of the earth. We're also in a deep recession.

"When the capital market returns, the small, fast, modular projects will recover much faster than the big, slow, lumpy ones," says Rocky Mountain Institute founder Amory Lovins, who has been advocating for distributed generation for 30 years. "Central power plants with capital costs in the billions can't get financing at all." Collectively, the benefits of distributed energy -- Lovins counts 207 of them -- "increase the value of smaller systems by a factor of 10. That's enough to flip any investment decision."

As with Dan Fink's experiment with Skippy and the wheel, the transition to small and local after a century of large and distant may take some imagination. "What you're really talking about," Lovins says, "is the flowering of a million miniature solutions."