Interference between plots

Previous chapters have dealt with methods of statistical design and analysis for improving the precision of genotype comparison by controlling sources of environmental variation, particularly those due to soil heterogeneity. Plot interference is another potential source of experimental error which occurs when plot yield is affected not only by the genotype grown in the plot, but also by the particular genotypes in neighbouring plots. Interference is most likely to occur when the plots are small and unbordered, as in the early stages of a plant selection programme. In contrast to plot errors arising from underlying field variation, interference error can lead to a systematic bias in genotype effects that persists across trials and is not reduced by randomization and replication.