The need for increased sustainability of performance in cereal varieties, particularly in organic agriculture (OA), is limited by the lack of varieties adapted to organic conditions. Here, the needs for breeding are reviewed in the context of three major marketing types, global, regional, local, in European OA. Currently, the effort is determined, partly, by the outcomes from trials that compare varieties under OA and CA (conventional agriculture) conditions. The differences are sufficiently large and important to warrant an increase in appropriate breeding. The wide range of environments within OA and between years, underlines the need to try to select for specific adaptation in target environments. The difficulty of doing so can be helped by decentralised breeding with farmer participation and the use of crops buffered by variety mixtures or populations. Varieties for OA need efficient nutrient uptake and use and weed competition. These and other characters need to be considered in relation to the OA cropping system over the whole rotation. Positive interactions are needed, such as early crop vigour for nutrient uptake, weed competition and disease resistance. Incorporation of all characteristics into the crop can be helped by diversification within the crop, allowing complementation and compensation among plants. Although the problems of breeding cereals for organic farming systems are large, there is encouraging progress. This lies in applications of ecology to organic crop production, innovations in plant sciences, and the realisation that such progress is central to both OA and CA, because of climate change and the increasing costs of fossil fuels.

Marker-assisted selection (MAS) provides opportunities for enhancing the response from selection because molecular markers can be applied at the seedling stage, with high precision and reductions in cost. About each of 50 genes conferring monogenic resistances and hundreds of quantitative trait loci (QTL) for quantitative disease resistances have been reported in wheat and barley. For detecting single-major gene resistance, MAS could be easily applied, but is often not necessary because the resistances are selected phenotypically. In quantitative disease resistances, MAS would be very useful, but the individual QTL often have small effects. Additionally, only a few monogenic resistances are durable and only a few QTL with high effects have been successfully transferred into elite breeding material. Further economic and biological constraints, e.g., a low return of investment in small-grain cereal breeding, lack of diagnostic markers, and the prevalence of QTL–background effects, hinder the broad implementation of MAS. Examples in which MAS has been successfully applied to practical breeding are the wheat rust resistance genes Lr34 and Yr36, the eyespot resistance gene Pch1, the recessive resistance genes rym4/rym5 to barley yellow mosaic viruses, mlo to barley powdery mildew, and two QTL for resistance to Fusarium head blight in wheat (Fhb1 and Qfhs.ifa-5A). Newly identified broad-spectrum resistance genes/QTL conferring resistance to multiple taxa of pathogens offer additional perspectives for MAS. In the future, chip-based, high-throughput genotyping platforms and the introduction of genomic selection will reduce the current problems of integrating MAS in practical breeding programs and open new avenues for a molecular-based resistance breeding.

Fusarium culmorum and F. graminearum Groups 1 and 2 cause seedling blight, crown rot, foot rot and head blight in wheat and rye that may affect grain yield and quality for baking and feeding. This review starts with an analysis of Fusarium populations with regard to their genetic variation for aggressiveness, mycotoxin production, and isolate‐by‐host genotype interaction. To assess resistance in the different host growth stages, quantitative inoculation and disease assessment techniques are necessary. Based on estimated population parameters, breeding strategies are reviewed to improve Fusarium resistance in wheat and rye. Epidemiological and toxicological aspects of Fusarium resistance that are important for resistance breeding are discussed. F. culmorum and F. graminearum display large genetic variation for aggressiveness in isolate collections and in naturally occurring populations. The production of mycotoxins, especially deoxynivalenol and its derivatives, is a common trait in these populations. Significant isolate‐by‐host genotype interactions were not found across environments in wheat and rye.Artificial infections in the field are indispensable for improving Fusarium crown rot, foot rot and head blight resistance in wheat and rye. For a reliable disease assessment of large populations, disease severity ratings were found to be the most convenient. The differentiation of host resistance is greatly influenced by an array of nongenetic factors (macro‐environment, microclimate, host growth stage, host organ) that show significant interactions with host genotype. Selection for environmentally stable resistance has to be performed in several environments under a maximum array of different infection levels. Selection in early growth stages or on one plant organ does not in most cases allow prediction of resistance in adult‐plant stages or another plant organ.Significant genetic variation for resistance exists for all Fusarium‐incited diseases in breeding populations of wheat and rye. The patho‐systems studied displayed a prevalence of additive gene action with no consistent specific combining ability effects and thus rapid progress can be expected from recurrent selection. In wheat, intensive testing of parental genotypes allows good prediction of the mean head blight resistance after crossing. Subsequent selection during selfing generations enables the use of transgression towards resistance. In hybrid breeding of winter rye, the close correlation between foot rot resistance of inbred lines and their GCA effects implies that selection based on the lines per se should be highly effective. This is not valid for F. culmorum head blight of winter rye caused by a greater susceptibility of the inbred lines compared to their crosses.For both foot rot and head blight resistance, a high correlation between the resistance to F. graminearum and F. culmorum was found in wheat and rye. Mycotoxin accumulation occurs to a great extent in naturally and artificially infected plant stands. The correlation between resistance traits and mycotoxin contents are medium and highly dependent on the environment. Further experiments are needed to clarify whether greater resistance will lead to a correlated reduction of the mycotoxin content of the grains under natural infection.