﻿108 Pen field and Harper — Chemical 



cipitate contained the bulk of the beryllium) were ignited in 

 the same crucible and fused with sodium carbonate. The 

 fused material was soaked out in water, phosphoric acid 

 obtained in the solution by the usual method, while the insol- 

 uble beryllium oxide was dissolved in hydrochloric acid, pre- 

 cipitated with ammonia, weighed, and the trace of P 2 5 con- 

 tained in it separated with ammonium molybdate. The P 2 & 

 and bases in analysis III, on which we place the greatest con- 

 fidence, were separated and determined in the above manner. 



Other determinations are given to show the accuracy of the 

 method. The CaO in II was all obtained in the filtrate from 

 the basic acetate precipitation and was not partially precipitated 

 as gypsum. The BeO in the same analysis was tested for 

 A1 2 3 , but not more than a questionable trace could be de- 

 tected. The P 2 5 in I was from a direct determination with 

 ammonium molybdate, but a very slight mechanical loss was 

 incurred. The fluorine was determined in III by driving off 

 the silicon fluoride and titrating the hydrofluosilicic acid by 

 means of a standard alkali.* After making the determinations, 

 fresh U-tubes were connected with the apparatus and the 

 aspiration carried on for several hours, but no more silicon 

 fluoride was driven off. As our alkali had just been standard- 

 ized by means of sodium carbonate and test experiments on 

 fluor spar, we feel very confident that the fluorine determina- 

 tion is correct. The water determinations are somewhat sur- 

 prising- Mackintoshf made no tests for water, and considered 

 the mineral to be an anhydrous phosphate with fluorine. Wink- 

 ler;); made no determination of fluorine and obtained no satis- 

 factory tests for that element by etching glass. He obtained 

 a loss by ignition, however, of 6*59 per cent in the Stoneham 

 mineral by strong ignition, which he regards as water. Grenth§ 

 obtained 6*04 per cent of fluorine by direct determination, and 

 0*61 per cent of water by heating the mineral with lead oxide 

 to bright redness. Mackintosh] obtained good tests for fluorine 

 by etching and found a loss of 6*03 per cent by strong ignition. 

 He proved conclusively that fluorine was given off by strong 

 ignition, and concludes that it is replaced in part at least by 

 oxygen. We are scarcely willing to believe that such a reac- 

 tion would take place on heating an anhydrous phosphate con- 

 taining fluorine. Our first water determination was made by 

 weighing out the mineral in a boat, placing it in a combustion 

 tube containing a layer of dry sodium carbonate, igniting the 

 tube to full redness and collecting the water in a weighed 

 chloride of calcium tube. We obtained 0*69 percent; after- 

 ward the mineral was dissolved, and analysis II completed 



* Am. Chem. Journal, i, 27. f This Journal, III, xxvii, 135. 



\ Neues Jahrbuch fur Mineralogie, 1884, ii, 134. 



§ Proo. Amer. Phil. Society, xxi, 1884, 694. || This Journal, III, xxviii, 401. 



