A been variety trial was used by the author to investigate the problem as to whether or not it is necessary to use guard rows for this type of experiments. The inner rows and the guard rows of each plot of the experiment, were harvested separately and statistical analysis was carried out on their separate yields and also on that of the plot as a whole. A satisfactory agreement was found between the F values and the coefficients of variation for the inner rows, for the guard rows, and for the whole plot. An F-test on the yield differences between central rows and guard rows of each plot was made and found not to be significant. Also, the correlation between the yields of the inner rows and the guard rows was very high. Based on the results of this investigation, the author suggests that guard rows in bean variety trials should be used only when the possibility of strong competition betweetn adjacent rows of different varieties is suspected.