S. L. Penfield — Methods for Determination of Water. 37 



carbonate in the form of fine dust, but its passage into the 

 weighed absorption apparatus was prevented by a loose plug 

 of fine asbestus, slipped into the tube after the boat was in 

 place. 



Three determinations of water, made on the same sample of 

 talc that was used for testing the previous method, gave 446, 

 4*37 and 4*44 per cent, a mean of 442. At the conclusion of 

 an experiment the platinum foil can readily be removed from 

 the glass by heating and plunging into cold water, and may be 

 flattened by rolling with a glass tube. This method is one 

 that is applicable for the determination of water in all cases 

 of mineral analysis, and if carefully executed the results are 

 very accurate. In its essential details, decomposition by fus- 

 ing with an alkali carbonate and collecting the water in a 

 weighed absorption apparatus, it is similar to the methods of 

 Lndwig* and Sipocz.f The former suggested the use of a 

 specially constructed platinum tube in which the mineral was 

 fused with a mixture of dry sodium and potassium carbonates, 

 the latter made use of a porcelain tube and the fusion was 

 made with the mixed carbonates in a platinum boat. The use 

 of a glass tube has a decided advantage over platinum and 

 porcelain, since the operation may be watched and the intense 

 heat that can be obtained with the charcoal furnace renders it 

 possible to use sodium carbonate for making the fusion, thus 

 avoiding the use of the very hygroscopic potassium carbonate. 

 The furnace may also be found convenient for other experi- 

 ments where an intense heat is required. 



This investigation was undertaken for the purpose of finding- 

 some simple means for the direct determination of water in 

 refractory minerals, and the results that have been given indi- 

 cate that the methods are accurate. The closed tube method 

 is almost as simple as the determination by loss on ignition, 

 and a direct weighing of the water is always a satisfaction, 

 since otherwise there is a possibility that some volatile con- 

 stituent may go off and render the result inaccurate. For 

 example the sample of talc that was used in the previous ex- 

 periments, and found to contain 4*42 per cent of water, lost 

 4*87 per cent by direct ignition the high result being due to 

 the presence of a little fluorine and the liberation of acid 

 water probably carrying some silica. 



Laboratory of Mineralogy and Petrography, 

 Sheffield Scientific School, New Haven, Conn., April, 1894. 



*Tschermak's Min. Mitth., 1875, p. 214. 



f Sitzb. der K. Akad. der Wiss., \Vien, lxxvi, p. 51, 1877. 



