190 Gooch and Austin — Constitution of the 



the conditions. In nearly all the experiments to be detailed 

 use was made, therefore, of faintly ammoniacal solutions and 

 wash -water. 



In Table III are given the results obtained in a study of the 

 effects of varying proportions of ammonium chloride and the 

 soluble phosphate upon the constitution of the precipitate. 

 All precipitates were gathered upon asbestos in the filtering 

 crucible, washed in faintly ammoniacal water, and ignited as 

 usual. In every case the precipitation was practically com- 

 plete ; for, upon allowing the filtrates with the wash-water to 

 stand for several days after further addition of microcosmic 

 salt, nothing but insignificant traces of a precipitate — not ex- 

 ceeding: 0*0001 grm. — ever appeared. In the experiments of 

 section A precipitations were made in the cold by the action 

 of microcosmic salt in considerable excess upon the solutions 

 of magnesium nitrate containing varying amounts of ammo- 

 nium chloride. In experiments (1) to (5) the liquid was made 

 faintly ammoniacal after the addition of the precipitant and 

 the precipitate was filtered off immediately after complete 

 subsidence ; in experiments (6) to (10) the precipitate first 

 thrown down was redissolved in a very little hydrochloric acid 

 and reprecipitated by dilute ammonia (the operation being 

 repeated several times) with a view to improving the crystal- 

 line condition of the precipitate, and this treatment introduced, 

 of course, a small amount of ammonium chloride, probably less 

 than a gram. It will be observed that errors of excess appear 

 in all of these determinations, those being the greatest in the 

 experiments in which the largest amounts of the ammonium 

 salt were present. 



In the experiments of section B the manipulation was so 

 changed that the supernatant liquid was poured off (through 

 the filtering crucible which was to be used subsequently to col- 

 lect the phosphate) after the precipitate had subsided and the 

 insoluble phosphate was dissolved in hydrochloric acid and 

 brought down again, after dilution, by the addition of a faint 

 excess of dilute ammonia. By thus removing the supernatant 

 liquid after the first precipitation, the excess of the precipitant 

 and the amounts of ammonium chloride originally present were 

 reduced to relatively low limits, so that their effects in the 

 reprecipitation were at a minimum ; and by adding varying 

 amounts of ammonium chloride, or none at all, before the 

 reprecipitation, it became possible to demonstrate the individual 

 effect of that reagent apart from that of an excess of the 

 microcosmic salt. It will be noted that in experiments (11) 

 and (12), in which no ammonium salt was added after the 

 decantation from the first precipitate, the results are ideal, and 

 that the errors of excess advance as the amounts of ammonium 



