256 Journal of Agricultural Research voi. xv. no. 4 



Plots -4^ and y^-^j acre in size and each having a width to length ratios 

 of I to II have in ends and sides 68.69 ^^'^ 35-35 P^r cent, respectively, 

 and in the sides only 64.65 and 32.83 per cent, respectively, of their total 

 area in a 12-inch strip within their margins. Likewise, plots -^ acre in 

 size ^^'ith a ratio of i to 4 for width to length have only approximately 

 half as great areas in 12-inch strips within their margins as those -^^ acre 

 in size and of the same shape. 



Increased size then in plots of the same shape reduces the area exposed 

 to alley effect. Increase in size of plots where shapes are widely different 

 may not bring about a reduction of the area in the marginal strip exposed 

 to alley effect. As an example of this, plots -^ acre in size and 24 by 

 33 feet compared with those -^ acre in size and 17 by 66 feet have, re- 

 spectively, 13.89 and 14.44 per cent of their total area in a 12-inch mar- 

 ginal strip. 



As shown in Table II, plots ^^ acre in size but having width to length 

 ratios of i to 22 and i to 5.5 have 67.68 and 37.37 per cent, respectively, 

 of their total area in 12-inch marginal strips. Plots ^ acre in size with 

 the width to length ratios of i to 16 and i to i (approximate) have 24.69 

 and 11.59 per cent, respectively, of their areas in 12-inch marginal strips. 

 The nearer plots approach the form of a square, the lower the percentage 

 of their total area is exposed to border effect (z). 



A consideration of the percentages given in Table II brings out the 

 fact that yields from the same varieties or treatments obtained under 

 like conditions, but from different-sized plots, surrounded by alleys are 

 not comparable. 



If end and side border effect are similar, it should also be noted in 

 this connection that the removal of end borders only from plots of 

 several different sizes surrounded by alleys may render yields from them 

 still less comparable. A single example using the data included in 

 Table II for plots ^ acre in size, but the width to length approxi- 

 mately I to 16 and I to I, respectively, will suffice to make this 

 point clear. With no borders removed the percentage of a 12-inch 

 strip within the margins of the plots to the total area of each are 24.69 

 and 11.59, and with end borders removed, 18.89 and 5.70, respectively. 

 With no borders removed the percentage of the plots 8.5 by 132 feet 

 exposed to alley effect is approximately twice as great as for those 34 

 by 32 feet. When the end bo,rders are removed, these percentages are 

 18.89 and 5.70, respectively, the one being approximately three times 

 as great as the other. 



In Table III are given the yields of oats, wheat, and barley grown in 

 plots of the same length, but varying in width. The yields for each 

 crop from the y^-acre plots are considerably higher than those from 

 the ^V-acre plots and the yields from the ^-acre plots are somewhat 

 higher than those from the -jijj-acre plots. 



