AbstractWe report preliminary results of a series of experiments designed to explore the importance of interspecific competition within arable weed communities at different scales. Competition hierarchies were apparent from a pot experiment with different levels of nutrients and water. Two field experiments looked atBromus sterilis, Galium aparineandPapaver rhoeasin winter wheat in the field, in a range of combinations and management treatments, and a fourth field experiment included a wider variety of species. There was little effect of fertilizer on population behaviour in the the field.Bromusincreased around ten fold per year on minimum‐tilled plots, regardless of other treatments.Galiumincreased on organically‐fertilized and minimum‐tilled plots, but only in the absence ofBromus. Papaverdensities remained low, but again were depressed in the presence of high densities ofBromus. Taken together, the experiments demonstrate the existence of competition between weed species. However, as the design of the experiment increased to include greater levels of environmental variation, so competition became more difficult to detect, and less useful for interpreting the results than knowledge of the biology of the individual species. At the scale of interest to the farmer, the level of competition is not a good predictor for weed population dyn