For thousands of years mankind has altered the genomes of crop plants to improve their yields and qualities, first by domestication of suitable plants and selection of landraces, followed by classical plant breeding strategies during the last 150 years, later complemented by cell-and tissue-culture techniques, and most recently by molecular breeding. For many of today’s crop plants domestication started 9,000 to 13,000 years ago by selection of wild plants with certain traits meeting the needs of humans.1 The repeated cultivation and maintenance of these selected plants under site-specific conditions led to landraces highly adapted to particular growing conditions and production methods. In some areas with extremely harsh conditions these landraces are still of certain relevance. Starting in Europe in the first half of the 19th century, landraces were in many parts of the world gradually replaced by higher yielding cultivars generated by classic breeding strategies.2 These modern plant breeding strategies have led to a multiplication of yield and considerable quality improvements in many important crop plants. At the same time the continuous selection for specific traits and higher yields has caused a loss in diversity in modern breeding material compared to landraces or the wild ancestors of cultivated plants. The domestication and breeding process usually represents a kind of genetic bottleneck. On the other hand, diversity is a precondition for breeding. Without a broad base of heterogeneous plant material it is impossible for plant breeders to produce cultivars that meet the changing needs regarding adaptation to growing conditions, resistance to biotic and abiotic stresses, or quality requirements. Therefore, the most efficient way to further improve the performance of crop varieties is still to have access to a large and diverse pool of genetic variation.