OF AETS AND SCIENCES. 169 



it necessary to insure the acidity of the residue of evaporation by add- 

 ing a drop or two of acetic acid before repeating the treatment with 

 methyl alcohol. 



On the whole, methyl alcohol shows itself to be an excellent agent 

 by which to secure the volatilization of boric acid, 



To retain free boric acid, magnesium oxide naturally suggests itself. 

 According to Mai'ignac* it is effective, and, if in the course of analysis 

 it may liave been partly converted to the chloride, it is easily regener- 

 ated by the action of heat and moisture. Marignac, it will be remem- 

 bered, makes use of magnesia mixture — the chlorides of ammonium 

 and magnesium with free ammonia — to fix the boric acid, evaporating 

 the solution to dryness, igniting, extracting with boiling water, filter- 

 ing, and weighing the residue, while the filtrate is again treated as 

 before to recover traces of the borate which have yielded to the solvent 

 action of the water. During the drying and ignition the magnesium 

 chloride yields hydrochloric acid, and it would seem scarcely possible 

 that the magnesium borate should fail to show some loss of boric acid 

 when both hydrochloric acid and moisture exert their action. Further, 

 the presence of ammonia during evaporation does not prevent the 

 volatilization of boric acid,t and Marignac regards the addition of it 

 from time to time as of doubtful use. So it appears natural to look 

 for some loss under such conditions, and Marignac fully recognizes 

 the fact that the apparent accuracy of his method is due to the bal- 

 ancing of errors, the inclusion of foreign matter by the magnesium 

 borate and the deficiency of the magnesia when precipitated as am- 

 monio-magnesium phosphate together compensating for the loss of 

 boric acid by volatilization. To bring the matter to the test, the fol- 

 lowing experiments were made. In them and in all succeeding exper- 

 iments the boric acid was weighed in solution, the standard of this 

 having been fixed by dissolving in a known weight of water a known 

 weight of fused boric anhydride prepared in a state of purity by 

 frequent recrystallization. The magnesium oxide employed was made 

 from the pure chloride by precipitating by ammonium carbonate and 

 igniting, and was free from lime and alkalies, and as far as could be 

 determined otherwise pure. The whole operation of each experiment 

 was conducted in one vessel, so as to avoid transfers. In all cases a 

 weighed platinum crucible of 100 cm.^ capacity received a weighed 

 portion of magnesia, and after ignition and subsequent weighing the 

 weighed solution of boric acid was introduced. In experiments (1) to 



* Zeit. fiir Anal. Chem., i. 40G. t Rose, Fogg. Ann., Ixxx. 262. 



