V. THE ANALYSIS AND COMPOSITION OF SOILS 89 



the organic matter charred or burnt, cooled, moistened with strong 

 hydrochloric acid, taken up with water, and filtered ; the residue is 

 ignited and weighed, its weight, "soluble silica," being added to the 

 other " silica," already separated. 



The filtrate and washings are made up to 100 cc. (or 250 if more 

 convenient). 25 cc. (or 50 if from 250 cc.) are then taken, boiled 

 with a few -drops of nitric acid, in order to oxidise the iron, and mixed 

 with just sufficient ammonia to neutralise the free acids present. A 

 slight excess of ammonia is then added and the whole boiled until the 

 free ammonia is nearly all expelled. The precipitate is filtered off, 

 washed, thoroughly dried, ignited and weighed. The weight is equal 

 to the ferric oxide, alumina and phosphoric acid present. The aggre- 

 gate weight of ferric oxide and alumina is then obtained by deducting 

 the weight of phosphoric acid, calculated from the results of its direct 

 determination (vide p. 90). 



Another portion of the filtrate from the silica is taken for the 

 determination of the total potash. This can be done conveniently 

 and accurately by Tatlock's method. 



25 cc. (or 50 cc.) of the hydrochloric acid solution are placed in 

 a porcelain dish, and platinum tetrachloride in sufficient quantity to 

 convert all the potash, soda and magnesia into double chlorides is 

 added (in general, about 0*3 gramme will be sufficient), and the liquid 

 slowly evaporated on the water bath. The residue is washed, first 

 with a little platinum chloride solution, which dissolves and removes 

 sulphates, phosphates, etc., also the double chlorides of platinum with 

 calcium, sodium, magnesium, etc., and then by decantation, in the 

 dish, with alcohol (96 per cent, not methylated), the washings being 

 passed through a small filter. When the washings are colourless, the 

 precipitate, which should consist of bright orange crystals, is washed 

 with alcohol into a weighed porcelain crucible, and the alcohol poured 

 off as completely as possible through the filter. The crucible is then 

 placed in a warm place for a short time, heated in a steam bath for 

 two hours, and then weighed. The small filter, which should only 

 contain traces of precipitate, is then burnt in platinum wire, its ash 

 added to the crucible, and the whole again weighed. The last increase 

 is taken as being due to 2KC1 4- Pt. It is calculated to K 3 PtCl 6 and 

 added to the weight of the main quantity of the precipitate. From 

 the sum of these weights the amount of potash in the soil can be cal- 

 culated, knowing that 94 of potash (K 2 O) correspond to every 485 of 

 the double chloride. 



This method works very well and is a great saving in time and 

 labour over the old method requiring the preliminary removal of the 

 iron, aluminium, lime and magnesium. Unfortunately it prevents the 

 simultaneous determination of the sodium. 



The so-called " total potash " obtained in this way is not necessarily 

 the whole quantity of potash which the soil contains, for hydrochloric 

 acid does not remove all the potash from silicates. 



If the real total potash be required, it is necessary to analyse the 

 residue insoluble in hydrochloric acid, exactly as is done with a 

 mineral silicate. However, the potash in a soil, insoluble in hydro- 



