PROFESSOR J. W. MALLET ON THE ATOMIC WEIGHT OF GOLD. 
421 
argento-cyanide, and. the substitute for the free sulphuric acid of the copper experi¬ 
ments was an excess of potassium cyanide. The results obtained were essentially 
similar to those of the copper experiments, the effect of difference in temperature 
between the two solutions being, however, less decided, and the slight effect of 
difference in the size of the plates (“ density of current”) less constant and distinct. 
In all the preceding experiments it was found that the most constant results 
under otherwise similar conditions were obtained by using feeble currents rather than 
those of greater strength, especially in the case of the silver solutions, 'there seemed, 
however, to be a limit to this. On the whole, the most satisfactory results were 
obtained (both in these preliminary experiments, and in those aiming at the atomic 
weight determination) with currents not exceeding y^th of an ampere per square 
centimetre of surface of (one side of) the opposed plates, and in some cases a current 
but one-fifth of this maximum was used. 
Having in view the indications afforded by the preliminary experiments, it was 
detei’inined to use tolerably strong solutions of the metals to be deposited, with 
not moi’e than a moderate excess of free acid, or, in the case of the double cyanide 
solutions, excess of potassium cyanide, to maintain the same temperature in both the 
electrolysis cells and to have this temperature as low as possible (about 2° C.), and to 
have the plates of the two metals to be compared equal in size, and at equal distances 
apart, using a weak electric current, and keeping watch over its strength by means of 
an ordinary hydrogen voltameter in the circuit. 
In the actual experiments on the deposition of gold as compared with silver it 
was originally proposed to use a solution of potassium auri-cyanide against one of 
potassium argento-cyanide, with the expectation that 3 atoms of silver would be 
thrown down for 1 atom of gold. But the first attempts made showed clearly that 
this reaction could not be obtained. The comparison as to gain in weight of the gold 
and silver cathode plates gave results leading to an atomic weight for gold impossibly 
high if the silver deposited were taken to represent 3 atoms, and much too low if it 
were taken to represent but 1 atom. Hence it appeared that the potassium auri- 
cyanide had been partially, but not completely, reduced to auro-cyanide by the action 
of the current, and an intermediate result obtained as to the equivalent quantity of 
silver between that due to the one or the other gold salt if exclusively present. 
A change was therefore made to the auro-cyanide in the preparation of the solution 
to be electrolysed. A pure form of potassium cyanide was prepared with the aid of 
alcohol, and carefully tested as to the absence of any metal capable of deposition 
from the watery solution on electrolysis. Auric chloride was precipitated by ammonia, 
the fulminating gold, after washing, dissolved in a strong solution of this potassium 
cyanide with the aid of heat, and the auro-cyanide crystallized out by cooling. The 
crystals were washed, redissolved in water, aurous cyanide separated from the 
solution by evaporation with hydrochloric acid, and the crystalline powder after 
cautious washing again dissolved in potassium cyanide solution, using for the purpose 
