Yields of arable crops are commonly lower on the crop margins or headlands, but the nature of the relationship between yield and distance from the crop edge has not been clearly defined, nor have the reasons for lower marginal yields. Surveys of 40 winter wheat headlands were carried out in 2 years to determine how yield changed with distance, and what factors might influence this relationship. Two field experiments were also conducted over 3 years in winter cereal headlands, in which the effect of distance was measured under conservation headland and conventional (fully sprayed) management.

Yields in the headland surveys varied from 0·8 to 10·2 t/ha. An inverse polynomial regression model was fitted to yield and weed data. Best fits were obtained by using separate parameters for each site. Adjusting yields to take account of weed dry matter improved the non-linear fit between yield and distance from crop edge. Field experiments provided similar results but the non-linear relationship was not as apparent.

There was a negative relationship between soil compaction, as measured by a cone penetrometer, and yield in one field experiment, where soil density values were relatively constant. No relationship was found between pattern of nitrogen fertilizer application and yield. Conservation headland management resulted in lower yield at one experimental site, especially in the third year, but not at the other site. Where yields were affected, weed dry matter was higher in conservation headland plots than in fully sprayed plots.

Although greater weed competition appears to account for at least part of the observed yield reductions on headlands, the role of other factors, particularly soil compaction, needs further study. Increased weed infestation may be an indirect result of reduced crop competition caused by other adverse conditions.