424 
PROFESSOR J. W. MALLET OX THE ATOMIC WBTG-HT OF GOLD. 
quite the same, so that the weights of the two metals thrown down may not be 
strictly equivalent."^ It was, therefore, deemed important to work with feeble 
currents, and, while heating all the plates in a Sprengel vacuum before weighing, to 
reject the results of all th(jse experiments in which the quantity of gas thus discharged 
amounted to more than the merest trace. But, if the source of error in question still 
exist at all, it might affect the atomic weight of gold in comparison with that of 
silver, either by making the former appear higher or lower than the truth. 
The source of error most to be feared, however, in connection with the application 
of this electrolytic method to the determination of the atomic weight of gold, is the 
uncertainty of having all the gold throughout the process in the form of potassium 
auro-cyanide in the solution, in view of the transition observed to auri-cyanide during 
electrolysis, although change in the opposite direction occurs with even greater ease. 
Each of the two salts appears to admit of electrolytic decomposition, and the presence 
of any traces of the auri-cyanide, in which the gold has triad character, while the 
calculation is based on the supposed presence of monad gold only would, of course, tend 
to make the atomic weight of the metal appear lower than the truth. 
Sixth Series of Exi^erimenis. 
These experiments consisted merely in the further application of electrolysis to the 
deposition of metallic gold from a solution of potassium auro-cyanide, comparing the 
weight of the metal thrown down, however, not with the weight of silver, but with the 
volume of hydrogen gas liberated by the action of the same current, the object being 
to thus secure, with an assumed knowledge of the density of hydrogen,^ a direct 
comparison of the atomic weight of gold with that of the element most generally 
taken as the basis of the numerical constants in question. 
A cell containing the same solution of potassium auro-cyanide as was used in the 
* As beai-ing on the question of tlie simultaneons decomposition of two electrolytes in the same 
solution, the follovying results may be recorded of an experiment made with a solution of mixed zinc and 
copper sulphates, with excess of potassium cyanide, the anode plate being of brass and the cathode plate 
of platinum, and an analysis made of the proportions of the two metals in the anode plate, in the solution 
as first taken, and in the alloy deposited on the cathode plate and subsequently dissolved off from it by 
means of nitric acid. 
Proportion of copper 
to zinc. 
In the brass anode 
plate. 
In the solution elec¬ 
trolysed. 
In the alloy deposited 
on the cathode plate. 
Copper .... 
68-74 
13-81 
71-34 
Zinc .... 
:H-26 
80-19 
26-60 
lUU-00 
lOU-00 
100-00 
Different results would undoubtedly have been oblained by substituting some other metal for one of 
those taken. 
