April, ’89.] history society of Wisconsin. 
199 
of the metal with different fluxes and different forms and 
materials for crucibles and electrodes. Some of the many 
points to consider in attempting to obtain A1 by the 
direct electrolysis of its salts are the following. 
1. Cheapness of the salt or material used. 
2. A salt whose melting point is above the melting 
point of the A1 and boiling point below that of Al. 
3. That will to the least extent attack the material of 
the crucible and electrodes. 
4. Which offers the least transfer resistance. 
5. Which does not dissolve the metal when once liber¬ 
ated at the cathode. 
6. That does not facilitate diffusion of the ions and 
thus permit their reuniting. 
The salts that have been found in practice necessary to 
use, are the anhydrous chlorides and fluorides of Al, more 
commonly the double salts with sodium. The dou e 
chlorides, as given in Hoffmann’s reports for 18/-, cos 
at the factory in Salindres, France, 22 cents per poun , 
which would make its cost for a pound of Al $2.20. e 
Webster Castner works at Oldbury, near Birmingham, 
England, now have facilities for manufacturing the c ou o e 
chloride in much larger quantities than heretofore, an 
have undoubtedly made improvements in minor etai s , 
but the general process remains the same as it 1 
1872, and no great reduction in cost can be expecte . 
double chloride can be made quite fiee from Si icon 
but always contains some iron (Fc). Cryolite is 
sarilv used as a flux when the double chloiide is re u 
bv Na. The action of the cryolite has been explained as 
dissolving the thin coating of oxide that forms arou 
the reduced particles of Al. . 
It may be, however, that the action of t ie ciyo i * 
not that of a solvent of the oxide of the metal simply, bu 
that it assists in decreasing the surface tension c 
ute globules of metal, or otherwise mechanica } ^ 
