POTASH, NITRITE OP 597 



The Austrian method has also been used by some, but it is quite inadmissible as a 

 general working process. It consists in ascertaining the temperature at which the 

 (solution crystallises. 



Gossart's method consists in determining the value of the nitre oy measuring its 

 power of oxidation. The latter is accomplished by finding the quantity of protoxide 

 of iron which it can convert into peroxide. If to an acid solution of protosulphate 

 of iron nitric acid or a nitrate be added, the proto is converted into a persalt at the 

 expense of a portion of the oxygen of the nitric acid. 



M. Pelouze endeavoured to improve the above process by using such an excess of 

 the protosalt of iron that the nitre added should be able to convert only a portion of 

 it into a persalt. The remaining protoxide was then converted into persalt by 

 means of a -solution of permanganate of potash of known strength. The data so 

 obtained enabled the value of the nitre to be estimated. But even this process is liable 

 to variations, so much so, indeed, that Messrs. Abel and Bloxam obtained in eleven 

 experiments made with pure nitre as many different results. 



The next process which we shall notice is that which the chemists generally have 

 settled upon as yielding the best results. It is that of M. Gay-Lussac. It depends 

 on the fact that if nitrate of potash be heated with charcoal, or, in fact, any carbona- 

 ceous matters in excess, the nitrate is converted into carbonate of potash, the amount 

 of which may be accurately estimated by means of a standard solution of sulphuric 

 acid. The chlorides which may be present are unacted upon by the charcoal, and do 

 not, therefore, influence the result ; but if sulphates be present, they are reduced by 

 the carbon to sulphides, which, in consequence of being decomposed by the sulphuric 

 acid, may cause serious errors. Fortunately the amount of sulphuric acid present 

 in nitre is seldom sufficient to cause any great error. Any nitrate of soda present 

 would come out in the final result as nitrate of potash, and thus become another 

 source of error ; in practice this is seldom likely to occur. The original process con- 

 sists in weighing out 20 grammes (308'69 grains) of crude saltpetre, and mixing it 

 with 5 grammes (77'17 grains) of charcoal, and 80 grammes (1234'7 grains) of 

 chloride of sodium. The mixture is thrown little by little into a red-hot crucible, 

 and, when the decomposition is over, allowed to cool. The residual mass is dissolved 

 in water, filtered, and water passed through the filter until it amounts to 200 cubic cen- 

 timeters (12'2 cubic inches). The amount of alkali is then ascertained with a burette 

 and standard sulphuric acid. (See ALKALIMETER.) Messrs. Abel and Bloxham have 

 minutely and laboriously studied this operation, and detected its sources of difficulty 

 and error. Their researches have led them to employ the following modification : 



Twenty grains of the sample are to be well mixed in a platinum crucible with 30 

 grains of finely-powdered resin and 80 grains of pure dry common salt. The heat of 

 a wire-gauze flame is then applied, until no more vapour is given off. The crucible is 

 then allowed to cool down a little, and 25 grains of chlorate of potash are added. A 

 gentle heat is then applied until most of the chlorate is decomposed ; the heat is then 

 raised to bright redness for two or three minutes. The mass should be fluid, and free 

 from floating charcoal. The mass, when cool, is removed to a funnel, and the 

 crucible, &c., washed with boiling water. The mass is then dissolved in hot water, and 

 the entire solution, coloured by litmus, is neutralised with the standard acid. In 

 the annexed table 20 grains of pure nitre were taken for each experiment : 



Exp. Nitre found - Nitre per cent. 



1 20-00 100-00 



20-00 

 19-97 

 19-97 

 20-08 

 20-08 

 20-08 



100-00 



99-85 



99-85 



100-40 



100-40 



100-40 



These chemists, not yet satisfied, made 53 more experiments by this method. The 

 mean result with pure nitre was 997 per cent. The mean of 25 of the above experi- 

 ments was 98'7 per cent. The mean of the remainder was 100-7. 



Subsequent experiments showed that greater accuracy might be obtained by substi- 

 tuting, for the resin, pure ignited finely -divided graphite, prepared by Professor Brodie's 

 patented process. To perform the process, 20 grains of the nitre are to be mixed 

 with 5 grains of ignited graphite and 80 grains of salt. The general process is con- 

 ducted in the manner described in the operation with resin. The results are very 

 exact, and apparently quite sufficient for all practical purposes. 



POTASH, NITRITE or. KO.NO 3 (KJTO 2 ). When ordinary saltpetre, or nitrate 

 of potash, is heated with sulphuric acid, in the cold, no special reaction becomes evi 



