416 EXPERIMENT STATION RECORD. 



precipitation should be allowed to stand over night. This plus error, 

 however, is so small that it may be neglected in rapid methods designed 

 for industrial work. The transformation of the tri magnesium phos- 

 phate into ammonium-magnesium phosphate is very slow in the pres- 

 ence of ammonium chlorid alone. It is advisable, therefore, in every 

 case to add the required amount of citrate. The precipitation of mag- 

 nesium in the presence of an excess of ammoniacal phosphate gives a 

 minus error as great as the plus error observed when the phosphoric 

 acid is precipitated in the same solution. It thus appears that this 

 classic method for the determination of magnesia is always inaccurate. 



Volumetric estimation of phosphoric acid, B. W. Kilgore (North 

 Carolina Sta. Bui. 110, pp. 123-128). — This is an account of comparative 

 tests of the Pemberton method as modified by Kilgore and others 1 on 

 North Carolina, South Carolina, Florida, Tennessee, and Pennsylvania 

 phosphates; cotton-seed meal, tankage, sodium phosphate, aluminum 

 phosphate, Thomas slag, bone meal, acid phosphate, and mixed fer- 

 tilizers, to determine the accuracy of the method as applied to sub- 

 stances of widely varying composition, and to ascertain the amount of 

 washing necessary to free the yellow precipitate from acid. The method 

 gave as good results as the gravimetric method on all the samples 

 analyzed. In fact, on materials containing large amounts of iron it 

 apparently gave better results than the gravimetric method. 



The influence of the amount of wash water was tested by using 200 

 and 500 cc. of water and of wash solutions. The results showed that 

 200 cc. of water was sufficient in all cases and that the use of 500 cc. 

 of water did not lower the results at all or only very slightly. 



"When 200 cc. of 3 per cent ammonium and potassium nitrate solutions were used 

 for washing the results were practically the same as those obtained when the two 

 quantities of water were used; but when they were washed with 500 cc. the results, 

 greatly to my surprise, were much lower. The results were so surprising that quite 

 a number of them were repeated, and while they were not uniform in all cases they 

 were always low. We had hoped to be able to wash with a very large volume of 

 ammonium nitrate without appreciably dissolving the precipitate or causing it to 

 run through the filter. The filtrates from these 500 cc. ammonium and potassium 

 nitrate washes were perfectly clear, but on evaporation the ammonium phospho- 

 molybdate was found to be in solution."' 



The determination of potash by reduction of potassium-plati- 

 num chlorid with sodium formate, B. S.tollema (Chem. Ztg., 21 

 (1897), No. 71, pp. 73'.), 710). — The author briefly discusses the various 

 sources of error in the determination of potash by the methods gener- 

 ally employed, and recommends a modification of the method of Coren- 

 winder and Contamine,- 1 which is as follows: A portion of the potash 

 solution, corresponding to 0.5 gram of substance, is slightly acidified 



'U. S. Dept. Agr., Division of Chemistry Buls. 43, p. 68; 47. p. til', ami 49, p. 75; 

 North Carolina Sta, Bui. If9; Jour. Amer. Chem. Soc., 16 (1894), p. 765; 17 (1895), 

 p. 941 (E. S. R., 6, pp. 180, 376. 502; 1, pp. 264, 741;. 



a Bui. ttoc. Ind. du Nord, 1879. 



