In maize (Zea mays L.), as in other cross-pollinated crops, a high mutational load of deleterious recessive alleles hampers the use of landraces in hybrid breeding. The in vivo doubled haploid (DH) technology meanwhile offers great promise in purging such alleles. This approach was employed to three European flint type landraces. Eighty DH lines were evaluated for testcross performance under three farming conditions. Mean testcross performance of DH lines was similar to that of their parental landraces but 22-26% lower than that of present elite flint lines. Highly significant genetic variance existed in each of the three DH-line groups. Coefficients of genetic variation ranged from 6.7 to 7.4. DH-lines derived from ÔLucq de Be´arnÕ displayed specific adaptation to N deficiency. The 10% (out of 80) best DH lines surpassed the first-cycle lines by 11% in testcross grain yield and remained only 13% below the elite lines. Altogether, the DH approach proved to be a highly effective tool for gaining access to the hitherto hardly tapped broad genetic diversity of landraces in research and breeding.

Leaf rust (Puccinia recondita Rob. ex. Des.) is the most frequently occurring leaf disease in German winter rye (Secale cereale L.). To test the usefulness of race-specific resistance genes, the effects of increased host diversity and complexity by producing two-and fourline synthetics from inbred lines carrying different resistance genes were analysed. Thirty-three synthetics along with two full-sib families and one hybrid variety were tested in 17 environments in Germany under natural infections. For comparison, the parent lines of the synthetics were evaluated in 11 environments. Only two synthetics and the full-sib families were resistant across all environments. Observed resistance levels of the synthetics were highly correlated (r ¼ 0.83, P ¼ 0.01) with those predicted from the parental values. Host complexity had a minor effect in two-line synthetics only. In conclusion, the effectiveness of race-specific leafrust resistances among environments, and increasing the host complexity and diversity does not lead to a higher resistance level than that expected from the resistances of the parents.