928 EXPERIMENT STATION RECORD. 



burner to drive off organic matter. Get the soil as much as possible on one side of 

 the dish and put in 1 or 2 cc. of hydrofluoric acid. Allow the soil to come in contact 

 with the acid very slowly to avoid loss by sputtering, using a platinum stirring rod. 

 After the violent action has ceased, place on a steam bath and evaporate to dryness. 

 Repeat this operation one or two times and then take up with a little hydrochloric 

 acid and water. Filter and wash into a 100 cc. flask, place the filter and contents 

 into the platimuu dish and, after drying over the flame, ignite the paper. There 

 will be a small amount of the coarse mineral, which is transferred to an agate mortar 

 to be ground, after which it is again digested in hydrofluoric acid until there is no 

 insoluble residue left. Take up in hydrochloi-ic acid and water, and add to the 

 original washings." 



The method of mechanical analysis used "was to some extent improvised, it being 

 in the main the beaker decantation, or what is more generally known as the Osborne 

 method. The radical change was a method devised for the disintegration of the 

 sample. Instead of pestling with a rubber-tipped pestle, the sample was agitated in 

 water by means of a shaking machine. Twenty grams of soil were put into a cylin- 

 drical bottle, the ordinary 8-oz. sterilizing bottle being used, with about 150 cc. of 

 ■water, and the bottle shaken about one hour at the rate of 150 strokes per minute. . . 

 Samples shaken one hour and one week gave constant results, disproving any grind- 

 ing of particles. The siftings were made through sieves and bolting-cloth, the decan- 

 tations controlled by the microscope. Below 0.006 mm., the decantation could be 

 made with no accuracy, and resort was had to an arbitrary floating method. The 

 period of 18 hours was selected." 



In the studies of availability it was found that digestion for 5 houi's in irJo normal 

 hydrochloric acid gave results more closely agreeing with the amounts of fertilizing 

 constituents taken uj) by the crop in the pot tests than any other method tested. 

 "In this method the ratio of substance to solvent was 200 gm. per liter. However, 

 a liter flask was used in which 186 gm. of soil were placed, and the solvent added 

 up to the mark. This varied only 1 or 2 cc. from 930, which Avas the ratio desired. 

 After the digestion the whole Avas shaken and emptied on to a fluted filter sufficiently 

 large. After draining, the volume of the filtrate did not vary 10 cc. from 800. This 

 was shown in so many cases that the filtrate was no longer measured, but taken as 

 800 cc, corresponding to 160 gm. of the soil. This expedited the work greatly, and 

 avoided the recording of figures and the chance for mistake. To this filtrate was 

 always added 1 or 2 cc. of nitric acid for the double purpose of decomposing any 

 ammonium chlorid which may have been formed in the digestion of organic matter, 

 and also to oxidize any organic matter in solution. After evaporating to dryness, 

 hydrochloric acid was added repeatedly and evaporated until there was no further 

 evidence of the presence of nitrates. The residue was now transferred to a smaller 

 porcelain dish, and diluted to about 50 cc. with water. To this was added 2 cc. of 

 platinum chlorid, according to the method of direct estimation of potash (E. S. R., 

 10, p. 408). The solution was slowly concentrated until the potassium platinichlorid 

 could be crystallized on the sides of the dish, after which it was set off to cool and 

 solidify. It was then treated with acidified alcohol as described in the method 

 referred to, washed onto a paper filter, and washed with plain alcohol and then 

 with the half-saturated solution of ammonium chlorid, according to the usual 

 method. After drying the salt, it was dissolved and washed through with hot water 

 into small platinum dishes, evaporated to dryness, dried at 100°, and weighed. The 

 filtrates were set aside after the washing with ammonium chlorid, stirred up and 

 the ammonium platinum chlorid allowed to settle over night. Most of the liquid 

 can then be decanted into porcelain evaporators. The residue is washed onto a filter 

 with alcohol 3 or 4 times, the washings being added t« the original solution for 



