On some of the relations of salts of zinc and alumina to soda and potassa