METHOD OF CONDUCTING THE EXPERIMENTS. 9 



The size of the plats was left to the cooperator. In the work from 

 1906 to 1910 the plats consisted for the most part of five rows each, 

 making with the one intermediate row which was discarded at har- 

 vest, 11 rows in the experiment. In the work of 1911 most of the 

 experiments were conducted on four plats of five rows each, with 

 three intermediate rows, one between every two plats, which were 

 discarded at harvest. The length of the plats varied with the con- 

 venience of each man, the average length being about 250 feet. 



After planting, one plat or series of plats was to receive absolutely no 

 cultivation, the weeds and grass being removed at frequent intervals 



Fig. 1.— Typical appearance of the soil on a cultivated plat of corn. 



by means of a sharp hoe. This hoe was to be used with a horizontal 

 stroke to cut off the vegetation at the soil surface, and particular 

 care was to be paid not to stir the soil any more than was absolutely 

 necessary. The other plat or series of plats was to receive the ordi- 

 nary cultivation and in addition to have the weeds and grass removed 

 by chopping so as to eliminate the weed factor. Figure 1 shows the 

 typical appearance of the soil on a cultivated plat and figure 2 on an 

 uncultivated plat when the plants had reached the tasseling stage. 

 These illustrations show the tests at the Arlington Experimental 

 Farm on a stiff clay soil where the surface of the uncultivated plats 

 was nearly as dry and hard as a floor. 

 51092°— Bui. 257—12 2 



