Our dependence on foreign oil has researchers and policymakers taking another hard look at weeds and corn stalks as sources of home-grown fuel.

President Bush's Advanced Energy Initiative, announced last month, calls for research into biofuels from "cellulosic" plant waste, "to displace up to 30 percent of the nation's current fuel use." Indeed, in his State of the Union address, the president suggested that one solution to the nation's "addiction" to oil could be fuel derived from switchgrass, a tall plant native to U.S. prairies. Reinforcing that vision, more than one recent study has suggested that biomass could eventually play a significant role in U.S. transportation energy needs, and do so without adding to the carbon in the atmosphere.

But are these realistic scenarios -- or just wishful thinking? The idea of using biomass for energy isn't new of course. Already, about four billion gallons of ethanol are produced yearly in the United States by fermenting corn and distilling out its energy-rich alcohol. But the amount that can be produced is limited by the land required to grow the corn. What's more, the process for producing ethanol is inefficient, requiring nearly as much energy to make as is available in the final product.

Advances in genetic engineering, however, now have many experts feeling optimistic about dramatically increasing the amount of biomass that can be harvested from an acre of land, by using microbes to convert leaves and stalks, not just corn, into liquid fuels. They believe this can be done efficiently, too, without exhausting available land and water, and also predict that production costs could be competitive with gasoline.

A recent report by the National Resources Defense Council and researchers at Dartmouth and Princeton projects that by 2050, in part through harvesting both protein and cellulose from corn and switchgrass, existing agricultural land could both supply our food needs and replace gasoline with ethanol.

Unless it's done carefully, however, deriving fuels from biomass could destroy crop lands through erosion, increase air pollution -- and even increase our dependence on fossil fuels. For example, one of the steps in processing biomass, distillation, requires heat. In the short term, inexpensive coal may appear to be a good energy source for this, says John Reilly, associate director for research at MIT's Joint Program on the Science and Policy of Global Change. But this would cancel out one of the primary benefits of biomass: carbon released by burning biofuels is offset by the carbon captured by growing crops, leading to near-zero total carbon emissions. Using coal for distillation would destroy this balance. In one scenario, Reilly says that "68 percent of the carbon you think you're saving is actually being emitted through other processes." Likewise, using gasoline or diesel to transport biomass from widespread farms and other agricultural facilities to processing centers would change the overall carbon equation.