2 9 



( i ) . Ferric hydroxide and a solution containing phosphoric 

 acid in large excess of the amount necessary to form the normal 

 salt, were allowed to react for weeks. The amount of phosphoric 

 acid absorbed was then determined. 



Fe(OH) 3 used . 0.2500 grs. 



P 2 O 5 absorbed 0.0275 " 



P 2 O 5 required to form FePO 4 0.1664 " 



Even in the presence of a large excess of phosphoric acid, 

 the ferric hydroxide absorbed little more than one tenth of the 

 amount necessary to form the normal salt. 



(2). The compound formed was treated with the 5 per 

 cent solution of potassium oxalate already described. After 

 filtering the filtrate was tested for iron, the absence of which 

 showed the absence of normal ferric orthophosphate. 



The compound formed by fixation is evidently a highly basic 

 phosphate of iron. The compound appears to be formed even 

 when only a limited amount of the base, is present. In the soil, 

 the bases would usually be present in large excess with respect 

 to the phosphoric acid. 



In the experiments just described, the lime was brought into 

 intimate contact with the phosphate of iron. Its action there- 

 fore was probably at a maximum. It was thought advisable to 

 supplement this with artificial soil experiments under conditions 

 which usually obtain in soils, conditions presumably less favora- 

 ble to the action of lime. 



For this purpose 200 gram portions of clean sand were ex- 

 tracted with hydrochloric acid and washed free from acid with 

 cold water. A blank experiment showed that the product pos- 

 sessed no absorptive properties for phosphoric acid. Each por- 

 tion was mixed with 4 grams of ferric hydroxide and placed in a 

 six inch cylinder. Superphosphate solution was then added and 

 absorption allowed to continue for ten days. During the mean- 

 time the mixture was kept moderately moist. After ten days 

 the phosphoric acid unabsorbed was washed out with cold water 

 and determined, the amount absorbed being calculated from the 

 data thus obtained. The mixture then represented a soil con- 

 taining 1.5 per cent Fe 2 O 3 , and 0.12 per cent of P 2 O 5 . 



