NOTES ON SILVER SMELTING. 



123 



In furnacing the material, matte, averaging 200 ozs. silver 

 per ton and 30 per cent, of copper, together with lead carrying 

 over 600 ozs. of silver per ton, was produced. The proportion of 

 lead to matte was as 1 is to 2. The furnaces worked well ; an 

 excellent slag was made (of type closely approaching type D, of 

 lies), and tlie daily assay of samples taken in the usual manner 

 (by drawing samples from the molten slag on a iron rod, and 

 plunging into cold water) showed a fairly low loss of silver, 

 averaging from 2 to 3 ozs. per ton. A calculation of results 

 however, revealed a great loss of silver, and in searching for the 

 cause, the slag dump was sampled and assayed, with a result 

 shewing a much larger silver content than the daily assays had 

 shewn. At first it was thought that the slag-breakers had been 

 careless in throwing away matte with the rejected slags, but 

 examination proved this not to be the case, and the clean crystal- 

 lized slag assayed as high as tlie bulk. At the same time a similar 

 material was being smelted in reverberatory fui'naces, producing 

 matte only, but of an equally high grade, and the slags from them 

 were I'emarkably clean, in fact much cleaner than those from 

 furnaces producing a lower grade matte. In blast furnaces, as is 

 well known, the matte often runs out with the slag, and is found 

 at the bottom of the pot forming a cake separated when cold by 

 a marked division from the slag above it. In reverberatory 

 furnaces the slag is skimmed from the molten matte in a liquid 

 state into beds to cool, while the matte is tapped into other and 

 separate beds. The writer was led to think that possibly the 

 blast furnace slag assays were correct, and that although when 

 hot and liquid the slag might contain little silver, yet by cooling 

 in contact with the rich matte, silver might pass into it fi^om the 

 matte. 



For some days assays were made of the slags both hot and cold 

 with the view of testing this surmise, and the following tigui-es 

 were obtained : — ■ 



