PROFESSOR J. W. MALLET ON THE ATOMIC WEIGHT OP GOLD. 
397 
known with greater certainty than at present. But until this condition is fulfilled 
the result for gold cannot be depended upon as of the first rank in exactness. In any 
renewed attempt to apply this method several questions would have to be examined 
as to the precise nature of the solution used, and of the reaction itself. 
[Berzelius* * * § also precipitated gold by means of a known quantity of phosphorus 
from a solution of the chloride used in excess, and his results, as calculated by 
F. W. Clarke,! lead to the atomic weight 195'303 for gold ; but this process appears 
ill adapted to give very exact results, even in such hands_ as those of the great 
Swedish chemist, and the value obtained is certainly too low in the light of more 
modern researches.] 
B. Experiments of Berzelius, 1844.|—In these experiments potassium auri- 
chloride, which, it was found, could not be completely dried without loss of chlorine, 
was ignited in hydrogen, and the residue was treated with water to dissolve potassium 
chloride, the quantity of which was determined, as well as that of the metallic gold 
left undissolved. Five experiments were made, and the aggregate amounts obtained 
of potassium chloride and gold were 3*7800 grm. and 9*9685 grm. respectively. 
These figures, if we assume K = 39*03 and Cl = 35*37, give for the atomic weight of 
gold 196*20, the lowest result from one of the individual experiments being 196*11, 
and the highest 196*27. 
Among possible sources of error in this process we may note as deserving considera¬ 
tion the conceivable retention by the potassium auri-chloride of hydrogen auri chloride, 
and the difficulty of directly determining with accuracy the potassium chloride 
extracted by water. The former would lead to a higher result for gold than should 
be obtained; the latter might either give too low a result in consequence of imperfect 
drying, or too high if there were partial loss by volatilization, either during the 
ignition in hydrogen or in subsequently recovering the potassium chloride from 
solution. The quantities of material used were smaller than is probably desirable. 
C. Experiments of Levol, 1850.§—A weighed quantity of gold was dissolved as 
auric chloride, the metal reduced from the solution by means of sulphur dioxide, and 
the sulphuric acid formed was determined as barium sulphate. In two expei’iments, 
reported as giving exactly the same result, 1 grm. of gold gave 1*782 grm. of barium 
sulphate. Hence, if Ba be taken = 136*86, S = 31*98, and O = 15*96, we have for 
the atomic weight of gold the number 195*86. 
Of the sources of error to which this method is liable probably the most important 
are atmospheric oxidation of sulphurous to sulphuric acid and imperfect washing out 
of soluble compounds of barium from the barium sulphate. Both would tend to give 
too low a result for gold. 
* Berzelius, ‘ LeErbuch,’ 5. Aufl., vol. 3, p. 1188. 
t “ Tbe Constants of Nature,” ‘ Smitlisoniau Miscellaneous Collections,’ Washington, D.C. 
4 Berzelius, ‘ Lebrbucli,’ 5. Aufl., vol. .3, p. 1212. 
§ ‘ Annales de Cbimie,’ [3], vol. 30, p. 355. 
