238 G-'ooch and Austin — Determination of 



pitate first thrown down was rediseolved, reprecipitated and 

 filtered after cooling; in those of section B, the precipitate was 

 filtered after cooling without resolution and without reprecipi- 

 tation . and in those of section C, the first precipitate was 

 filtered at once while the solution was still hot. The length 

 of digestion hefore filtering and the indications of manganese 

 in the filtrate are recorded in the table. 



It was observed in these experiments that when the amount 

 of ammonium chloride is present in considerable quantity a 

 fine crystalline condition is got much more readily than when 

 the amount of that salt is small: with maximum amounts of 

 ammonium chloride the change from the flocky to the crystal- 

 line condition is almost immediate; even in the cold the 

 change takes place to a marked extent in a few seconds. No 

 manganese was found in the filtrate by boiling with bromine 

 and ammonia — a test which is capable of indicating 0*0001 

 grams of manganous sulphate in 500 ciu3 of water containing 60 

 grams of ammonium chloride — until the ammonium chloride 

 amounted to 20 per cent of the mass, or to 40 grams in 200 cmS 

 of the liquid, and even then but once in three trials : even 

 when the proportion was 30 per cent — 60 grams in 200 cma — 

 the solvent action of the ammonium chloride upon the man- 

 ganese salt was trifling. The pyrophosphate residues obtained 

 in these experiments, as well as in all those recorded in this 

 paper, were dissolved in nitric acid and tested for contamina- 

 tion by a chloride ; in no single case did silver nitrate produce 

 more than an inappreciable opalesence in the solution. It is 

 plain, therefore, that the variations of the results from theory 

 are occasioned by variation in the degree of conversion of the 

 trimanganese phosphate to the ammonium manganese phos- 

 phate, and that, while the ammonium chloride shows no appre- 

 ciable solvent action on the precipitate in the presence of the 

 precipitant, its effect in the process of conversion is plainly 

 evident. For the smaller amounts of the manganese salts 

 (equivalent to 0*15-12 grams of the pyrophosphate) the effect 

 of the ammonium chloride reaches a maximum when that salt 

 amounts to 10 per cent of the solution ; for twice that amount 

 of manganese salt, the best results were obtained by doubling 

 the amounts of ammonium chloride. Either line of treatment 

 yields under the most favorable conditions, results which are 

 passably good, but the advantage inclines slightly to the first 

 method in which the first precipitate was dissolved and repreci- 

 pitated while the liquid was cooled before filtering. 



In Table III are recorded results obtained by precipitating 

 the cold acid solution of the manganese salt 'and the microcos- 

 mic salt with a strong excess of ammonia. The mixture was 

 heated to boiling for from five to ten minutes and filtered hot. 



