Abstract
During the origin of agriculture in the Fertile Crescent, the broad spectrum of wild plant species exploited by hunter-gatherers narrowed dramatically. The mechanisms responsible for this specialization and the associated domestication of plants are intensely debated. We investigated why some species were domesticated rather than others, and which traits they shared. We tested whether the progenitors of cereal and pulse crops, grown individually, produced a higher yield and less chaff than other wild grasses and legumes, thereby maximizing the return per seed planted and minimizing processing time. We compared harvest traits of species originating from the Fertile Crescent, including those for which there is archaeological evidence of deliberate collection. Unexpectedly, wild crop progenitors in both families had neither higher grain yield nor, in grasses, less chaff, although they did have larger seeds. Moreover, small-seeded grasses actually returned a higher yield relative to the mass of seeds sown. However, cereal progenitors had threefold fewer seeds per plant, representing a major difference in how seeds are packaged on plants. These data suggest that there was no intrinsic yield advantage to adopting large-seeded progenitor species as crops. Explaining why Neolithic agriculture was founded on these species, therefore, remains an important unresolved challenge.