86 TRANSACTIONS OF THE CANADIAN INSTITUTE. [VOL. II. 



shews the scales again. This process is, in fact, Plattner's assay for 

 Cobalt and Nickel carried out on a large scale. 



I now come to the treatment of the Sulphides. This is a more tedious 

 operation, as the ores are generally not so rich as the Arsenides. The same 

 principle is pursued, namely, to collect all the Nickel in a matte and then 

 treat the matte further. The ore is first roasted to remove part of the 

 Sulphur, and is converted partially into oxides. This product is then 

 smelted, and the Nickel combines with the remainder of the Sulphur, 

 while the Iron in conjunction with the Silicates which are already present 

 in the ore, or are added purposely, forms a slag. It is impossible to 

 effect this separation of Iron and Nickel in one operation, because it 

 would produce such an enormous amount of slag at once ; the operation 

 must therefore be repeated several times. The resulting matte is then 

 smelted under a blast in a reverberatory furnace or in a Bessemer con- 

 verter, and then further refined with Baryta and Quartz or with Saltpetre 

 and Soda in order to be finally treated in the wet or dry way. 



The reaction of the Baryta and Quartz on Iron Sulphide is interesting. 



3 Ba SO, +4 Si O. + Fe S + Ni S. 



=3 Ba Si 0.,+3 S O^ + Si 0, + Fe S + Ni S. 



3 S Og in the high temperature breaks up into 4 S O. +0. 



=Ba S I 0,+4 S O..+0+Si 0..+Fe S+Ni S. 



A ' 



=-3 Ba S I 0,+4 S 0, + Fe Si Og+Ni S. 



The method of treating the Sulphides can be divided into the following 

 parts : — 



1. Roasting the ores. 



2. Smelting the roasted ore to matte. 



3. Roasting the matte. 



4. Smelting the roasted matte. 



5. Blowing the concentrated matte. 



6. Refining, smelting of matte. 



The ore is generally roasted in heaps, with wood as fuel. Analyses of 

 different parts of the roast heaps shew that a concentration towards the 

 middle and bottom of the heap takes place, which can probably be ex- 

 plained by assuming that the easily fusible Sulphides of Copper and 

 Nickel sicker out through the mass, grown porous by the roasting. 



We have perhaps here an explanation of the fact frequently observed 

 in the deposits of Nickel-bearing pyrites, namely, that the deposit 

 becomes richer in Nickel as one goes down. Assuming that the molten 

 mass of Iron and Nickel Sulphide was originally homogeneous, as it 



