CHEMISTRY. 457 



(.'{) The oxidation method of Macli and PassonJ 



It appears that as a rule the first gives the Inghest results, the other 

 two following in order. 



An important source of error is claimed to be the i)re(-ipitation of 

 silicic acid if the temperature is allowed to rise above 80 to 85° 0. 



On the quantitative determination of potash, J. H. Vogel and 

 H, Haefcke ( Landw. Vers. iStat., 47 {1S96), No. 2-S,])p. 97-113, figs. 6). — 

 The various methods which have been j)roposed for the determination 

 of potash are reviewed and their faults pointed out, especial attention 

 being given to (1) the Freseuius short method^ and (2) the Lindo-Glad- 

 ding method as adopted by the Association of Ofiflcial Agricultural 

 Chemists. In the first the suli^huric acid is precipitated in acid solu- 

 tion, carefully avoiding an excess of barium; and the double salts of 

 sodium, calcium, and magnesium are washed out of tlie final precepi- 

 tate by means of alcohol. The author's experiments indicate that this 

 can be effectually done, but that if more than 75 cc. of alcohol is used a 

 certain amount of the potassium-platinum salt is also dissolved. The 

 averages of 4 tests of the solubility of this salt are: 1 : 35,190 for abso- 

 lute alcohol, 1:31,523 for 95 per cent alcohol, and 1: 20,760 for 80 per 

 cent alcohol. It is further urged against this method that the weigh- 

 ing on the filter or after dissolving on the filter, evaporating, and dry- 

 ing in a weighed dish is inaccurate. In the latter case the water is not 

 completely expelled from the large crystals of the double salt, and an 

 error is thus introduced which varies with the size of the crystals. 

 There is also danger of error from an excess of barium. The double 

 salt of barium is said to be decomposed by alcohol, and barium chlorid 

 is formed which is insoluble in that medium. 



By extraction of 5 gm. of the barium sulphate precipitate with hydro- 

 chloric acid there was found from 0.09 to 0.15 per cent of potassium 

 chlorid which had been occhuled. 



The Lindo-Gladding method is condemned on the following grounds : 

 (1) The addition of sodium chlorid increases the amount of washing 

 necessary and the chances of error from solution of the potassium- 

 platinum chlorid in the ammonium chlorid solution. (2) The precii)itate 

 not being ground, the thoroughness of washing varies with the size of 

 the crystals. Such being the case, it follows that (3) it is inaccurate to 

 weigh the double salt directly, but it should be reduced and the washed 

 and dried platinum weighed.^ 



The method proposed by the authors is as follows : In case of potash 

 salts dissolve 10 gm. in 300 cc. of hot water, make up to 500 cc, 

 evaporate 50 cc. of the solution to dryness, add 20 cc. of neutral 



» Ztscbr. angcw. Cliein., 1896, No. 5, p. 129 (E. S. R., 8, p. 23). 



^Quaiititiitive Analyses, 6tli ed., vol. 2, p. 292. 



'It will be seen that these objections arc based principally on a feature of the 

 original Lindo-Gladding method, viz, the addition of NaCl, which has long since 

 been abandoned by the Association of Official Agricultural Chemists. The method 

 described ia that adopted by the Asaociatiou iu 1887. 



