Related Physical Properties of Topaz. 389 



sium carbonates. The fusion was soaked out, filtered and 

 washed with hot water. To the hot filtrate five to ten grams 

 of ammonium carbonate were added and after cooling, still 

 another addition of the same reagent. After standing in the 

 cold for twelve hours the precipitate was filtered off, the excess 

 of ammonium carbonate expelled from the filtrate by heating 

 in a platinum dish on the water bath, and an ammoniacal solu- 

 tion of zinc oxide added. After evaporating until the odor of 

 ammonia had disappeared the zinc oxide precipitate was 

 removed by filtration, the filtrate heated and dilute nitric acid 

 added until the excess of alkali carbonate was nearly decom- 

 posed. To the slightly alkaline, boiling solution an excess of 

 calcium chloride was added and from this point the precipitate 

 was treated as previously described by one of us.* That deter- 

 minations made by this method are satisfactory was proved by 

 the following : In an experiment with topaz the residue result- 

 ing from soaking out the alkali carbonate fusion, and the pre- 

 cipitates formed by the ammonium carbonate and zinc oxide 

 were united, mixed with a fresh portion of alkali carbonates, 

 fused and treated as before. The amount of fluorine that was 

 obtained by this second treatment was only 0*07 per cent, 

 showing that practically all can be extracted by one fusion. 

 Moreover, fluorine determinations were made by the above 

 method in an artificial mixture of cryolite, cyanite and quartz, 

 taken in proportions to correspond with the composition of 

 topaz with the following results. 



Cryolite taken. Fluorine calculated. Fluorine found. Loss. 



•3325 '1805 -1788 -0017 



This result indicates only a slight deficiency, and it is probable 

 that the determinations in the regular topaz analyses are not 

 over 0*20 per cent low, as usually a gram and sometimes a 

 gram and a half of the mineral are taken for a determination. 

 In the above no allowance has been made for what might be 

 recovered by a second fusion of the residues, and probably some 

 of the lo6S is occasioned by volatilization during the alkali car- 

 bonate fusion, but since the crucible was kept covered this must 

 have been very slight. Since water is present, it is evident 

 that attempts that have "been made to determine fluorine by loss 

 on ignition, assuming that silicon fluoride is given off, can not 

 have given reliable results. For the determination of water, 

 the mineral has been fused with dry sodium carbonate and the 

 water absorbed in a weighed sulphuric acid tube. The method, 

 which has been carefully tested, gives accurate results. The 

 mineral is completely decomposed and it is impossible for acid 



* This Journal, III, xlvii, p. 190, 1894. 



