In field trials in 1987/88 near Pangbourne, England, wheat (Triticum aestivum) and field beans (Vicia faba) were grown in an organic farming system as sole crops and additive intercrops. The sole crops were grown at 25, 50, 75, 100 and 150% of the recommended density (RD) for conventionally grown crops. The intercrops consisted of all density combinations of wheat and beans from 25 to 100% RD in a factorial experiment. The grain yield of sole cropped wheat and beans increased significantly as their density was increased. The highest yield of both was achieved at 100% RD, indicating that the conventional recommendation was the optimum when applied to organically grown crops. Land equivalent ratio (LER) values for the intercrops were significantly greater than 1·0 when the wheat was sown at > 5% RD and beans at > 50% RD. The highest LER of 1·29 was achieved when wheat and beans were both sown at 75% RD. There was resource complementarity, expressed as relative yield total (RYT) > 1·0, in all of the density combinations. There was a significant decrease in resource complementarity with increasing wheat and bean density. The nitrogen content of the wheat grain and whole plant biomass was significantly increased when the density of beans in the intercrops was increased; this was reflected in a significant increase in grain protein at harvest. The total amount of N accumulated by the wheat, however, decreased with increasing bean density due to a reduction in the biomass of wheat. Beans also showed a significant increase in %N as the density of the other component increased and a decrease in total N accumulation due to reduced biomass. All of the intercrops accumulated more N than the sole cropped wheat, but did not exceed that accumulated by sole-cropped beans. The biomass of weeds was greater under beans than under wheat. Weed biomass in intercrops was significantly reduced when the density of wheat and beans was increased, resulting in a lower weed biomass in the intercrops than was achieved in either the sole cropped wheat or beans. The N content of weeds was significantly reduced with increasing wheat density but was significantly increased with increasing bean density. The total amount of N accumulated by weeds per unit area was reduced significantly by increasing the density of both components. The levels of disease on the wheat were low, but mildew (Erysiphe graminis) increased significantly as bean density increased. The incidence of chocolate spot (Botrytis fabae) increased significantly with increased bean density. The experiment demonstrated that it was possible to harvest the crop with a combine harvester and the wheat and beans can be planted separately mechanically, therefore this system is suited to mechanized agricultural systems.