For this estimation, solutions of the chloride as well as those of the sulphate (ammonium, iron, alum) may be used in the manner previously described. The electrolysis is best effected in the presence of a sufficient quantity of ammonium oxalate; no separation of any iron compound takes place. The iron is deposited in the form of a bright, steel gray, firmly-adhering mass on the platinum dish. The iron may be exposed to the air for several days without any noticeable oxidation taking place. DETERMINATION OF ZINC. Zinc may be separated from a solution of the double salt fully as easily and rapidly as the previously mentioned metals were. The reduced zinc has a dark gray color, and adheres very firmly to the electrode. The separated metal is dissolved by using dilute acids and heating. It is only removed with difficulty, and generally leaves a dark coating on the dish, which is separated by repeated ignitions and treatment with acid. DETERMINATION OF MANGANESE. It is already known that manganese may be separated as the peroxide from its nitric acid solution. We find, however, that the precipitation is only completely effected when the quantity present is small; the amount of nitric acid must also be slight, and it is necessary to wash the dish without interrupting the current. If the manganese is converted into the soluble double salt, prepared by adding an excess of potassium, and submitted to the electric current, the whole of the manganese will be deposited at the positive electrode. When ammonium oxalate is used, the complete precipitation does not take place. As the separated peroxide does not adhere firmly to the electrode, it is necessary to filter it and convert it, by ignition, into the trimangano-tetroxide (Mn_3O_4). DETERMINATION OF BISMUTH. This separation presents considerable difficulty, because the metal is not precipitated as a compact mass on the platinum. The bismuth is always obtained in the same form, no matter whether it is precipitated from an acid solution, or from the double ammonium oxalate, or, finally, from a solution to which potassium tartrate has been added. As large a surface as possible must be used, and the dish piled to the rim; then, if the quantity of bismuth is small, the washing with water, alcohol, and ether may be effected without any loss of the element. If small quantities of the metal separate from the dish, they must be collected on a tared filter, and determined separately. In our experiments, an excess of ammonium oxalate was added to a nitric acid solution of bismuth. During the electrolytic decomposition, a separation of the peroxide was observed at the positive electrode, which, however, slowly disappeared. In order to prevent the reduced metal from oxidation, the last traces of water are completely removed by repeated washings with alcohol and anhydrous ether. DETERMINATION OF LEAD.