Biofuels from crops are emerging as a Jekyll & Hyde – promoted by some as a means to offset fossil fuel emissions, denigrated by others as lacking sustainability and taking land from food crops. It is frequently asserted that plants convert only 0.1% of solar energy into biomass, therefore requiring unacceptable amounts of land for production of fuel feedstocks. The C4 perennial grass Miscanthus×giganteus has proved a promising biomass crop in Europe, while switchgrass (Panicum virgatum) has been tested at several locations in N. America. Here, replicated side-by-side trials of these two crops were established for the first time along a latitudinal gradient in Illinois. Over 3 years of trials, Miscanthus×giganteus achieved average annual conversion efficiencies into harvestable biomass of 1.0% (30 t ha−1) and a maximum of 2.0% (61 t ha−1), with minimal agricultural inputs. The regionally adapted switchgrass variety Cave-in-Rock achieved somewhat lower yields, averaging 10 t ha−1. Given that there has been little attempt to improve the agronomy and genetics of these grasses compared with the major grain crops, these efficiencies are the minimum of what may be achieved. At this 1.0% efficiency, 12 million hectares, or 9.3% of current US cropland, would be sufficient to provide 133 × 109 L of ethanol, enough to offset one-fifth of the current US gasoline use. In contrast, maize grain from the same area of land would only provide 49 × 109 L, while requiring much higher nitrogen and fossil energy inputs in its cultivation.