OF ARTS AND SCIENCES. 59 



few drops of a solution of ammonic nitrate in ammonia, dried on the 

 vacuum pump, ignited on a platinum crucible-cover, at first gently 

 until fumes of amnionic nitrate no longer appeared, then at a full 

 red heat until the spreading of a glow over the whole residue indi- 

 cated the formation of the magnesic pyrophosphate. 



The results of the experiments are tabulated below. The figures 

 of the first column show the weights of magnesic pyrophosphate 

 found, those of the second the corresponding weights of phosphoric 

 oxide, those of the third the weight of phosphoric oxide required by 

 the determinations of the standard of the solution used, those of the 

 fourth the absolute error in the determination of phosphoric oxide, 

 and those of the fifth the percentage error referred to the actual 

 amount of phosphoric oxide present. The Roman numeral standing 

 against the record of each experiment shows which phosphate solution 

 was used in that experiment. 



Alkaline Phosphates. 



Experiments (1) to (10) inclusive were made with 40 cm. 8 each 

 of the phosphate solution. Portions (1) and (2) were precipitated 

 with 20 cm. 3 of magnesic sulphate mixture. Portions (3) to (6) were 

 treated with 20 cm. 3 of a magnesic sulphate mixture containing only 

 a small amount of free ammonia, but the same weights of magnesium 

 and ammonium salts as the regular mixture. The first precipitates of 

 (3) and (4) were dissolved in hydrochloric acid and again thrown 

 down by ammonia in presence of the excess of the precipitant, and 

 those of (5) and (6) were dissolved and reprecipitated in this manner 

 twice. Portions (7) to (10) were precipitated with 20 cm. 3 of 

 magnesic sulphate mixture after being acidulated — (7) and (8) with 

 hydrochloric acid, (9) and (10) with 0.5 gr. of citric acid — and made 

 alkaline with ammonia. 



Per cent 

 error. 



+2.54 

 -j-4.16 

 +3.70 

 +4.62 

 +3.23 

 +3.39 

 +1.51. 

 +3.02 

 +7.17 

 +7.92 



