Use of multivariate statistics to separate allelopathic and competitive factors influencing weed suppression ability in winter wheat

P>There is an increasing interest in an integrated, less herbicide-dependent approach to weed management, based on combinations of cultivation practices and competitive crops. More competitive cultivars are essential to this approach and breeders need to know how to breed such. It is known that early vigour and extensive early season vegetative growth are important traits. The questions are, which is most efficient to use in breeding and if allelopathy should be bred for? Weed suppressive ability (WSA) of 12 winter wheat, two rye and two triticale cultivars was studied in a 2-year field trial with natural weeds and under-sown Apera spica-venti and rapeseed (Brassica napus). Crop biomass, shoot length, straw length and plantlet root length and root weight were analysed. Biomass growth in autumn and spring was studied in weed-free plots using multispectral analysis. Grain yield, winter hardiness and maturity data were collected from official variety tests in Sweden. Allelopathic activity was assessed with use of a bioassay. Factor analysis showed that early season crop biomass and allelopathy were two traits explaining most of the variance in weed biomass. It was predicted by partial least square regression models that weed biomass could be reduced by 60%, if both crop biomass and allelopathy are improved to the levels of triticale. An improvement of allelopathy alone should reduce weed biomass by 18-28%. The factor analysis and multivariate models showed that WSA could be improved without a negative trade-off in grain yield or winter hardiness.