240 Gooch and Austin — Determination of 



proportioned to the ammonium chloride present, and the 

 mean error of the results for the phosphate when the ammonium 

 chloride reached 20 grains — 0-0007 grra. — is considerably less 

 than the mean error —0*0018 grm. — when the amount of the 

 ammonium salt was less than 20 grim 



Experiments were also made according to the modifications 

 suggested by Munroe,* viz., the boiling of the manganous salt 

 with an excess of microcosinic salt until the precipitate be- 

 comes crystalline and just neutralizing with dilute ammonia ; 

 but we have been unable to find the conditions of this treat- 

 ment by which uniform results may be obtained in even mode- 

 rate agreement with the theory. 



We have tried also the effect of substituting ammonium 

 nitrate for ammonium chloride in the conversion process; but, 

 so far as our experience goes, the nitrate is not so effective 

 weight for weight in producing the change of the tri manga- 

 nous phosphate to the ammonium manganese phosphates, while 

 the solubility of the product in the solution of the ammonium 

 nitrate becomes appreciable more rapidly with the increase of 

 the amount present than is the case when the ammonium salt 

 is the chloride. 



In the light of the experiments described it would seem to 

 be reasonable to expect the best results from the phosphate 

 method for determining manganese when the conditions are 

 so arranged that precipitation may take place in the cold solu- 

 tion in the presence of but little free ammonia, and of enough 

 ammonium chloride to bring about the rapid conversion of the 

 precipitate to the crystalline condition. Under such circum- 

 stances it should be possible to secure the conversion of the 

 phosphate to the ideal constitution as completely as possible 

 without danger of subsequent decomposition by the prolonged 

 action of the hot free ammonia. In carrying out this idea, 

 the solution of manganese chloride was treated as before with 

 microcosmic salt and a large amount of ammonium chloride, 

 the precipitate first formed was redissolved in hydrochloric 

 acid and precipitation again brought about by the very careful 

 addition of dilute ammonia in slight but distinct excess. The 

 mixture was heated only until the precipitate became silky 

 and crystalline, when it was allowed to stand and cool for a 

 half hour. The precipitate was filtered off upon asbestos in 

 a perforated platinum crucible under pressure, ignited and 

 w T eighed. Table V comprises the results of experiments made 

 in this manner. In those of section A the precipitation was 

 made in platinum vessels: in those of section B the treatment 

 was in glass. 



*Loc. cit. 



