52 Journal of Agricultural Research voi. xviii. no. 2 



The crop was the selected strain of Beldi barley used in much of the 

 previous work of this laboratory. The plants were grown on a silty clay 

 loam soil (ic) in 191 6 and on a fine sandy loam soil (i^) in 191 7. 



In the 1 91 7 experiment the soil, made as homogeneous as possible by 

 sifting and mixing, was installed in 8 tight galvanized iron boxes or con- 

 tainers with a surface area of 30 by 60 inches each and a depth of 1 8 inches. 

 Two of the containers were equally divided by a lateral partition, giving 

 a total of 10 independent compartments. The soil in one of the large con- 

 tainers was used as a control plot, and the other plots were used for growing 

 the crop. The plants were placed 6 inches apart in the row with 6 ifiches 

 between rows. Thus there were 50 plants to each of the large containers 

 and 25 plants to each of the small containers, with equal volumes, weights, 

 and superficial areas of soil per plant. The plants were cut at intervals of 

 two weeks throughout the season, beginning with the third week from 

 planting. At the same time samples were drawn and water extractions 

 made from soil in the control plot container and from the containers being 

 cropped. No further soil samples were taken from any container after 

 the plants were removed therefrom. The containers were compactly 

 arranged, insulated from heat fluctuations except at the surface, and 

 protected from birds and rodents. All plots were watered with distilled 

 water only and protected during the infrequent rains by rubber covers 

 placed on wire frames. The mechanical arrangement and methods of 

 sampling soils were in all respects the same as those described in greater 

 detail in connection with other experiments from this laboratory (8). In 

 order to obtain sufficient material for the chemical studies, v/e used the 

 contents of two of the large containers for the first cutting, one large con- 

 tainer for each of the three succeeding cuttings, and one small container 

 each for the last four cuttings. The minimum number of plants embraced 

 by any cutting was 23 at the fifteenth week from planting, in which 

 instance 2 plants failed to develop after thinning. 



The procedure in the 191 6 experiment was similar to that of 1 917. A 

 minor difference which should be noted was that the containers for the 

 cropped soil in 191 6 were of wood and varied several inches in each 

 dimension from the iron containers used in the work of the following year. 

 A more important difference lies in the fact that in attempting to econo- 

 mize equipment and soil, which was brought from a distance at consider- 

 able expense, we used two containers of soil, one cropped and one 

 uncropped, which had been in place and were grown to a crop of barley 

 in the year 191 5. One of these was used as a control plot and one for the 

 determination of total yield at the end of the season. Subsequent 

 determinations of the water-extractable matters from these soils showed 

 somewhat higher concentrations than those from the soils in the newer 

 containers from which the successive cuttings were made. This could 

 have had no effect on the data obtained for the plant, inasmuch as the 

 periodic cuttings were all made on soil assembled, mixed, and installed 



