Breeding perennial grain crops

One-third of the planet's arable land has been lost to soil erosion in recent decades, and the pace of this degradation will increase as the limits of our food production capacity are stretched. The persistent problem of worldwide soil erosion has rekindled interest in perennial grain crops. All of our current grain crops are annuals; therefore, developing an array of new perennial grains - grasses, legumes, and others - will require a long-term commitment. Fortunately, many perennial species can be hybridized with related annual crops, allowing us to incorporate genes of domestication much more quickly than did our ancestors who first selected the genes. Some grain crops - including rye, rice, and sorghum - can be hybridized with close perennial relatives to establish new gene pools. Others, such as wheat, oat, maize, soybean, and sunflower, must be hybridized with more distantly related perennial species and genera. Finally, some perennial species with relatively high grain yields - intermediate wheatgrass, wildrye, lymegrass, eastern gamagrass, Indian ricegrass, Illinois bundleflower, Maximilian sunflower, and probably others - are candidates for direct domestication without interspecific hybridization. To ensure diversity in the field and foster further genetic improvement, breeders will need to develop deep gene pools for each crop. Discussions of breeding strategies for perennial grains have concentrated on allocation of photosynthetic resources between seeds and vegetative structures. However, perennials will likely be grown in more diverse agro-ecosystems and require arrays of traits very different from those usually addressed by breeders of annuals. The only way to address concerns about the feasibility of perennial grains is to carry out breeding programs with adequate resources on a sufficient time scale. A massive program for breeding perennial grains could be funded by diversion of a relatively small fraction of the world's agricultural research budget.