190 Penfield and Hoioe — Chemical Composition 



After having definitely determined the crystallographic 

 character of the minerals they were pulverized and sifted to a 

 uniform grain and separated from the gangue and other 

 impurities by means of the barium-mercuric-iodide solution. 

 Thanks to this accurate method of separation we have had an 

 advantage over all previous investigators in being able to ob- 

 tain an abundance of material for the chemical analyses. Each 

 product that was obtained was nearly uniform in specific 

 gravity and almost absolutely pure, as shown by examination 

 with the polarizing microscope. 



Method of analysis. — Silica, fluorine and bases were deter- 

 mined, in one portion, usually of from one to two grams, which 

 was fused with a mixture of sodium and potassium carbonates 

 and treated according to the Berzelius method of analyzing- 

 silicates containing fluorine. On soaking out the fusion with 

 water in a platinum dish it was found that, owing to the basic 

 character of the mineral, only a small quantity of silica had 

 gone into solution, and this was separated directly by means of 

 zinc oxide dissolved in ammonia, thus avoiding the customary 

 precipitation with ammonium carbonate. The solution filtered 

 from the zinc oxide and silica, and containing the alkali carbo- 

 nate and fluoride, was nearly neutralized with hydrochloric 

 acid, heated to boiling, and calcium fluoride and carbonate pre- 

 cipitated by addition of calcium chloride. After collecting 

 the precipitate it was ignited to low redness in a weighed 

 platinum crucible and the calcium carbonate or oxide removed 

 by means of acetic acid. At this point we have usually pro- 

 ceeded as follows : To the ignited precipitate in the crucible 

 water and 1 or 2 CC of acetic acid were added, the whole was 

 then digested for some time on the water bath with the cruci- 

 ble covered, finally the cover was removed and the excess of 

 acid evaporated. The dry salts were then extracted with hot 

 water and the solution filtered through a small filter. After 

 washing, the paper was burned in the same crucible, theTesi- 

 due treated again with water and a small quantity of acetic 

 acid, the operation being repeated until all of the calcium car- 

 bonate and oxide had been extracted. Our experience has 

 shown that a large excess of acetic acid used at once gives too 

 low results. This method is not very tedious and is satisfac- 

 tory, as shown by the following test experiments where deter- 

 minations were made with known weights of pure fluorite 

 mixed, with chrysolite in the proper proportion to give amounts 

 of fluorine and magnesium silicate about equal to those in 

 chondrodite. 



Equivalent of 

 Chrysolite. CaFe taken. fluorine. Fluorine found. Loss. 



1-3782 -2942 '1438 -1416 "0022 

 1-3794 -3012 -1473 -1456 -0017 



