Three winter wheat (Triticum aestivum L.) composite cross populations (CCPs) that had been maintained in repeated parallel populations under organic and conventional conditions from the F 5 to the F 10 were compared in a two-year replicated field trial under organic conditions. The populations were compared to each other, to a mixture of the parental varieties used to establish the CCPs, and to three winter wheat varieties currently popular in organic farming. Foot and foliar diseases, straw length, ear length, yield parameters, and baking quality parameters were assessed. The overall performance of the CCPs differed clearly from each other due to differences in their parental genetics and not because of their conventional or organic history. The CCPs with high yielding background (YCCPs) also yielded higher than the CCPs with a high baking quality background (QCCPs; in the absence of extreme winter stress). The QCCPs performed equally well in comparison to the reference varieties, which were also of high baking quality. Compared to the parental mixture the CCPs proved to be highly resilient, recovering much better from winter kill in winter 2011/12. Nevertheless, they were out yielded by the references in that year. No such differences were seen in 2013, indicating that the CCPs are comparable with modern cultivars in yielding ability under organic conditions. We conclude that-especially when focusing on traits that are not directly influenced by natural selection (e.g. quality traits)-the choice of parents to establish a CCP is crucial. In the case of the QCCPs the establishment of a reliable high-quality population worked very well and quality traits were successfully maintained over time. However, in the YCCPs lack of winter hardiness in the YCCP parents also became clearly visible under relevant winter conditions. Keywords: baking quality; climate change; dynamic management; evolutionary breeding; heterogeneous populations; resilience; sustainable agriculture; Triticum aestivum;