ON THE ESTIMATION OF POTASH AND PHOSPHORIC ACID. 31 



chloride added to cause the precipitate to appear distinctly reddish. The 

 liquid is boiled, well filtered, and the precipitate washed slightly. It is 

 redissolvcd on the filter in hydrochloric acid, tartaric or citi'ic acid added 

 judiciously to the solution, and then a tolerable excess of ammonia. The 

 alkaline sohftion should be greenish, not reddish. " Magnesia mixture " is 

 then added in moderate excess, the liquid stirred and left over night. The 

 precipitated double phosphate is then filtered off and treated in the ordinary 

 manner. The method gives results agreeing well with the average of 

 chemists of repute. On repetition, the results of the two estimations agree 

 ■within 4 to -2 per cent. 



Phosjpiates of Iron and Aluminium. 



The precipitate of iron and aluminium phosphates, produced by treating 

 the cold solution of a sample containing the above metals with an alkaline 

 acetate (or with ammonia and excess of acetic acid), can be very con- 

 veniently analyzed by the following method, contributed by Mr. H. Waring- 

 ton : — " The precipitated phosphates of iron and aluminium are washed, 

 ignited, and weighed, redissolved in strong hydrochloric acid, and the iron 

 determined volumetrically with stannous chloride and iodine (see Fresenius). 

 Erom the iron the quantity of ferric phosphate in the precipitate is cal- 

 culated, the phosxihate of aluminium found by difi'erence, and thus the iron, 

 aluminium, and phosphoric acid in the precipitate are obtained. A little 

 phosphoric acid is liable to be removed from the precipitate during washing, 

 and basic salts are thus reckoned in the calculation as of normal composi- 

 tion." 



Estimation hy Uranium. 



The removal of iron and aluminium by addition of an alkaline acetate in 

 the cold, with determination of the phosphoric acid in the filtrate by means 

 of uranium *, is a method which appears to deserve more extended employ- 

 ment. The use of an acetate in a slightly acid solution brings the liquid 

 into just the condition required for the use of the uranium process. The 

 volumetric estimation by uranium is very highly spoken of by some chemists 

 as convenient and fairly accurate, while others consider it very unsatisfactory. 

 The conflict of opinion is very great, and special experiments on this process 

 appear desirable ; but the following seem to be the precautions necessary for 

 successful working. 



The proportions of acetic acid and alkaline acetate employed and the 

 volume of the solution should be approximately constant. The uraniimi 

 nitrate should be standardized with an acetic-acid solution of pm-e precipi- 

 tated ammonio-magnesium phosphate or tricalcic phosphate, instead of with 

 phosj)hate of sodium as is commonly done. 



The titration should be converse, the solution of the phosphate being added 

 to that of the uranium. The latter should be mixed with a constant 

 proportion of acetic acid and heated on a bath of boiling water. The 

 indicator should be powdered potassium ferrocyanide on a white plate. 

 Owing to the reversal of the usual process, the brown colour of the ferro- 

 cyanide of uranium becomes gradually fainter till the end of the titration. 



This method, which is recommended and employed by some authorities of 

 great experience, is said to be capable of giving results of every desirable 

 accuracy. 



* It is universally admitted that the estimation by uranium is untrustworthy unless 

 any iron or aluminium present in the original solution is first removed. 



