484 Journal of Agricultural Research voi. xxi. n«. t 
number and location of controls, alleys between plots, etc., (3) the technic 
of harvesting and thrashing, and (4) the interpretation of the results. 
Carleton (4), Piper and Stevenson (17), Thorne ( 20 ), and others have 
considered practically the entire subject in a general way and have 
offered suggestions looking toward the improvement of technic in plot 
tests. Lyon ( 12 ) used original data and that of others to emphasize 
some of the essential considerations in plot tests. 
Harris ( 8 ) has used the original data of Montgomery (15) and that of 
Mercer and Hall (rj) to show the need of considering variations in the 
soils of fields used for experiment when interpreting the results from 
them. In a later article ( 9 ) the same writer used the original data of 
Kiesselbach (jj) ;and that of others to further demonstrate heterogeneity 
of soil in fields selected for their uniformity. Emphasis is laid on the 
necessity for greater care in the technic of plot tests and 011 the use of 
statistical methods in the analysis of the data. 
Surface and Pearl ( 19 ) propose a method for use in correcting for soil 
heterogeneity in plot tests. This method is useful only when the plots 
are located in blocks several in extent each way. 
Smith ( 18 ) emphasizes the value of replication and of carrying the tests 
over a period of years in securing dependable results. He also mentions 
that tiie plants in one plot may have an influence on those growing in 
adjacent plots. Montgomery ( 14 ) has shown that the plants of one 
variety may have a considerable effect on those of another variety grow¬ 
ing near. Hayes and Arny (10) report 011 the effect that plants in 
rod rows spaced a foot apart may have on each other. 
That the yields of plots flanked by cultivated alleys are higher because 
of the effect of the additional space on the outside rows has been shown 
by Amy and Steinmetz ( 2 ). Tjhe higher yields of border rows as com¬ 
pared with central rows in plots flanked by alleys has been shown by 
Arny and Hayes (r), and in addition, the effect of border rows 011 tlie 
rank of particular varieties in tests is emphasized. 
Since the publication of the earlier article ( 2 ) several papers which 
relate to this subject have come to the attention of the writer. Fletcher 
(7) reports that crops growing on the border of a fallow yielded at a rate 
as much as 10 times as great as the rate in the center of the plot. This 
is accounted for largely by the absence of toxic substances on the fallow 
side of the outside rows. The width of the fallow area is not mentioned. 
Chittenden (5) gives results for turnips planted in plots 33 feet long 
and 7 feet 6 inches wide. The rows of turnips were 18 inches apart, and the 
outer rows of any two plots were 4 feet 6 inches apart. This gives 36 
inches of additional space to the two outside rows Hanking each alley. The 
rows extended east and west. For the crop planted May 25 and weighed 
July 29, the weights of the tops of the outer rows averaged 94 as compared 
with the middle row’s weight of 78, both on the basis of 100 for the 
heaviest row. On the same basis, the roots from the outside rows aver- 
