422 
PROFESSOR J. W. MALLET ON THE ATOMIC WEIGHT OF GOLD. 
the barely necessary amount of the solvent liquid, but afterwards adding a further 
quantity, so as to have potassium cj^anide in excess. Potassium argento-cyanide was 
prepared by precipitation of a solution of pure silver in nitric acid with the purified 
potassium cyanide, washing the precipitate, and re-solution with the aid of the 
necessary quantity of potassium cyanide, of which finally a moderate excess was 
added. The solutions of the gold and silver salts were made of equivalent strength, 
for the most part at the rate of 7 grm. of metallic gold for each 100 c.cm. 
of solution, and an approximately corresponding amount of silver, taken atom 
for atom. Both solutions received the same excess of potassium cyanide, generally 
equal to one-half of that already present in the double salt, but in some of the 
experiments it was found necessary to add yet more during the electrolysis in order to 
preserve the purely metallic character of the surface of the plates. As an additional 
security against admixture of auri-cyanide with the auro-cyanide of the gold solution, 
it was subjected for some time to electrolysis with unweighed gold plates immersed, 
these being reversed two or three times in position, just before the introduction of 
the weighed plates for a quantitative experiment. A number of attempts were made 
to substitute for the solution of potassium auro-cyanide one of sodium auro-thio- 
sulphate, of potassium or sodium auri-chloride, and of simple auric chloride, in the 
last two cases employing at the same time a solution of silver nitrate, but these efforts 
led to no success. 
In many of the experiments made with the double cyanide solutions the cathode 
plates, both of gold and silver, after removal from the electrolysis cells and thorough 
washing, were found to curl up on being heated, the deposit, which in these cases was 
rather hard and brittle, swelling up in a remarkable w'ay, with formation and^bursting 
of little blebs or minute bubbles of the metallic surface, and parting off to some extent 
of the deposit from the original plate underneath. When the heating was carried out 
in the Sprengel vacuum small but Cjuite appreciable amounts of hydrogen were found 
to be given off, having been occluded in the metal deposited. It seemed necessary to 
throw aside the results in all cases in which this condition of the deposit was well 
marked. Other experiments were vitiated by the gold deposit not being thoroughly 
compact, and still others by the surface not being clearly metallic, aiu’ous cyanide 
making its appearance from the solution. It was hoped that in the experiments, 
free from apparent defect, any irregular behaviour of the gold solution at first might 
be got rid of by continued electrolysis, with reversal of the anode and cathode plates 
when necessary, until the ratio of gold to silver deposited should become constant; 
but confidence in this was greatly shaken when an instance occurred, followed after¬ 
wards by others, of sudden change in this ratio, attended with much less loss from the 
anode gold plate than the gain of the opposed cathode plate, pointing to deposition of 
gold from the auro-cyanide with simultaneous formation of auri-cyanide in the 
solution.'" 
* Hitiorf (‘ POGOEXDOKFF, Aiiiialen,’ [4], vol. 16, p. 523), in tbe simnltaneons electrolysis of gold and 
