236 PROFESSOR W. A. TILDEN ON THE SPECIFIC HEATS OF METALS, 



These preliminary experiments showed that with a little practice it is possible to 

 get results which are uniform to the third place, and could therefore be trusted as a 

 basis for calculating the atomic weight if the pure metal were used. 



Experiments on Pure Cobalt. 



Preparation. Commercially pure cobalt nitrate was dissolved in cold water and 

 fractionally precipitated by the addition of weak solution of bleaching powder, 

 leaving at least a quarter of the cobalt in solution. The oxide was collected, washed 

 with hot water, dissolved in moderately strong hydrochloric acid, and the solution 

 boiled to expel a part of the chlorine. Solution of ammonia was then added in excess, 

 and the solution filtered through paper to remove a small brown precipitate containing 

 iron and probably a little alumina. 



The solution on being evaporated gave a deep red crystalline precipitate of 

 purpureo-cobaltic chloride, leaving a pale mother liquor. The precipitate, drained, 

 well washed with hydrochloric acid, and dried, was heated strongly in a platinum 

 dish till fuming ceased. The dark blue crystalline mass of cobalt chloride was then 

 dissolved in water and precipitated hot by excess of pure sodium carbonate. The 

 precipitate was filtered off, washed, and heated to redness in a platinum dish. The 

 resulting oxide was washed with hot water till free from alkali and dried. The oxide 

 was then packed in a glass combustion tube and reduced at a red heat in pure 

 hydrogen. The spongy metal was allowed to cool in hydrogen, and was then 

 compressed into cylindrical blocks by a Spring compressor under a pressure of 64 tons 

 to the square inch. 



The rods thus formed were heated in a vacuum, when they gave off a little water 

 and about their own volume of gas, consisting almost wholly of carbon dioxide, 

 apparently absorbed from the air. 



On exposing this metal to contact with steam and subsequently drying, it was 

 found to increase appreciably in weight. Moreover, on attempting a calorimetric 

 estimation with the metal, the first considerable increase of weight due to condensa- 

 tion of steam was followed by a small but continuous loss of weight, probably due to 

 the slow expulsion of occluded gas. No accurate determination of specific heat was 

 therefore possible. The metal was therefore supported upon a block of pure lime and 

 melted in an oxyhydrogen flame. On the first attempt it was found that the melted 

 mass on cooling ejected a considerable quantity of gas, giving rise to hollow 

 excrescences upon the surface, the phenomenon closely resembling the expulsion of 

 oxygen from melted silver in the process of solidification. After several experiments, 

 it was found best to employ a considerable excess of oxygen at the end of the fusion, 

 and though some loss of metal was incurred through oxidation of the fused button, 

 much less escape of gas occurred on cooling. The buttons obtained were bright, but 

 often hollow, the sides of the cavity being white and silvery. The metal breaks with 

 a brilliant crystalline fracture. 



