The results show that, in 1921, there was increased yield in the outside and middle border rows for the winter wheat varieties where the alleys were not cropped and for the spring grains separated by alleys cropped to winter wheat. The effect on the inside border rows (the third from the outside of the plots) was nil for the winter wheat and slight for the spring wheat. Cropping the alleys between the plots of spring grains to winter wheat reduced border effect so that it was not plainly evident at any time before grain was harvested. This reduction is reflected in the relatively lower yields of the outside border rows of spring wheat, expressed in percent based on the yields from the central rows, as compared with the results secured in 1918 and 1917. For the oat and barley varieties, this reduction in border effect is not apparent from the yields of the outside rows expressed in percentages of the central rows because the fact that there was unavoidable loss of grain in harvesting the lodged central rows with the binder. This resulted in yields lower than they should be from the central rows and consequently higher percentages for the outside rows based on the yields from the central rows. The three-year average yields of the outside rows of oats, spring wheat and barley expressed in percent based on the yields of the central rows is 199.8 and that of the middle border rows (inside rows in 1917) is 138.0. Results secured in 1917 with Kubanka wheat, under dryer conditions, were 182.4 percent for the outside and 127.7 percent for the second rows based on the yields of the six central rows. All portions of the plots were harvested by hand. The effect did not extend to the third rows in the plots. The uncropped alleys varied from 16 inches to 38 inches and one was a cultivated roadway several feet in width. There appears to be some correlation between width of alley and amount of border effect but the number of determinations was to small to give conclusive results. The amount of increase in yield in the border rows in this trial is very similar to the three-year average under Minnesota conditions. The yields of the outside and inside border rows in the variety plots separated by alleys were in almost every instance considerably higher than the yields from the central rows. Where variety or rate of seeding plots have been planted without the intervention of alleys, border rows have given yields both considerably above and below the yields for the central rows. This point is brought out in work reported from Nebraska. Although no statement is made regarding the plan of the plots in the variety and rate of seeding trials, it is assumed that there were no alleys between these plots. Therefore, the plants in contiguous plots came into direct competition and both higher and lower yields resulted in the border rows as compared with the yields from the central rows. Available data emphasize the necessity of considering border effect seriously in variety and rate of seeding trials, both when no alleys intervene and when the plots are separated by cropped or uncropped alleys. The results for the varieties separated by alleys in plots without and with border rows removed for a three-year period, show that the removal of the outside border row from either side of each plot reduced yields approximately 10 percent and when two border rows were removed from either side of each plot the yields were reduced approximately 17 percent. When the border effect extended to the third row (inside border row) it was relatively unimportant. That border effect in plots separated by uncropped alleys does make a different interpretation of results necessary in some instances has been shown. In 1921, the third year that work of this kind was carried on, one variety was moved from below to above the discard point for the year by the removal of border rows. The fact that on each of the three years, one or more varieties or rates of seeding was moved from o