12 



as selective absorption. Most salts belong to the first or second 

 class and it is of the greatest importance that we know whether 

 arsenic remains fixed in the soil or is washed out by the rains. If 

 it is inactive then it may be considered harmless but if active 

 toward substances already present then a judicious use is 

 imperative. 



The first series of experiments to determine the action of 

 arsenic upon the soil from a chemical view point, were made with 

 the use of large glass percolators, 800 grams of soil being placed in 

 each of six percolators. The soils used were the same three as 

 used in the pot experiments. In this case 500 cc of a solution of 

 sodium arsenite containing 184 parts per million arsenic was 

 poured into the percolator every day for 3 days and the analysis 

 of the filtrate made to determine how much arsenic had been 

 taken out by the soil. These simple tests indicated a strong affinity 

 of soils for arsenic, the first portions of the filtrate being absolutely 

 free of this chemical. Later portions passing the soil contained 

 slight traces. 



In view of the above results a more thorough test was made 

 upon these same soils with lysimeters, using about 25 pounds of 

 soil. A solution of sodium arsenite of the same strength as used 

 in spraying was prepared and about 3 litres (approximately 3 

 quarts) poured on the soil in each lysimeter. Two lysimeters were 

 filled with each soil, to one of which the 3 litres of arsenic spray 

 was applied and the lysimeter allowed to stand about 2 months. 

 Samples of the soil were then taken at different depths to deter- 

 mine how deep the arsenic had penetrated and to what extent it 

 had distributed itself at various depths. To the second lysimeter 

 one quart of water was added every other day for several weeks, 

 then allowed to remain until dry enough to sample. Samples were 

 then taken at different depths to determine to what extent the 

 arsenic is carried down by the movement of the water. 



