Crop Cropping system Average for 2003–2006, kg/ha % increase ANOVA, variable F P Grain yield     Maize M(C) 12,813c 0 Cropping 54.1 <0.0001 M&F(C) 18,905a 49 Year 25.2 <0.0001 M&F(R) 17,363b 37 Cropping x year 1.5 0.2133     Faba bean F(C) 4,291b 0 Cropping 9.6 0.0010 M&F(C) 5,242a 22 Year 48.2 <0.0001 M&F(R) 5,716a 29 Cropping x year 2.9 0.0308 Above-ground biomass     Maize M(C) 26,918c 0 Cropping 53.9 <0.0001 M&F(C) 39,993a 49 Year 31.2 <0.0001 M&F(R) 36,989b 38 Cropping x year 3.0 0.0254     Faba bean F(C) 10,379b 0 Cropping 9.0 0.0014 M&F(C) 12,659a 22 Year 64.0 <0.0001 M&F(R) 13,003a 21 Cropping x year 4.3 0.0053

M(C), continuously cropped maize; F(C), continuously cropped faba bean; F&M(C), faba bean and maize continuously intercropped with the same crop on the same strips of land; F&M(R), faba bean and maize rotationally intercropped with each crop rotated to the other crop's strips each year (see Methods). Each treatment had three replicates, and all plots received 225 kg/ha N fertilizer and 40 kg/ha P₂O₅ fertilizer. Values in the same column within the same crop followed by different superscript letters are significantly different (P < 0.05) based on least-squared difference multiple comparison. The percent increase values are (M&F-M)/M x 100 or (M&F-F)/F x 100. The ANOVA was a two-way analysis with cropping systems and year as variables, where the df is 2 for the cropping system, 3 for the year, and 6 for the interactions between cropping and year.