342 



Journal of Agricultural Research 



Vol. XII, No. 6 



In figures 8 to 20 it is seen that the graphs given on the duplicate con- 

 tainers agree excellently. In all the soils except No. 9, 12, and 7 there 

 was an appreciable increase of potassium, which occurred from four to 

 six weeks after the time of planting. In all except the same soils, and 

 in addition No. 14, there was a liberation of nitrates which, like the 

 potassium, began to fall soon after the crop was rapidly developing. 



The method employed for calcium had proved so inaccurate that the 

 results were excluded from the final charts. It was known that the 

 potassium method had a high factor of error in the lower concentrations. 

 A definite increase of nitrates had been shown in the earlier periods of 

 growth. It was very clear that before any interpretation of these figures 

 could be made it was necessary to have not only more accurate methods 



100 8A 



ot,h'jKjl^ liiio 



5jElft;^o;i 1916 



Fig. 7. — Graphs of the yield of grain in 1915 and 19:6, expressed as a percentage of the maximum yield. 

 The heavily shaded portions in 1915 represent difference between duplicates. 



of determination but also a comparison between the conditions which 

 were found in the planted soil and those which would have occurred in 

 the same soil had the crop not been present. 



The first season's work therefore established the following facts: 

 (i) That a satisfactory duplicate crop could be grown under the con- 

 ditions employed; (2) that the extraction technic followed yielded prac- 

 tically duplicate graphs for the elements extracted. The period of a 

 year's treatment under uniform conditions had also tended to bring the 

 soils into normal relation to each other. It had given an opportunity 

 for the soils to recover from the sifting and handling which they had 

 received and be more comparable to a varied group of field soils exposed 

 to the same climatic conditions. 



