PROFESSOR J. W. MALLET ON THE ATOMIC WEIGHT OF GOLD. 
899 
being observable in the results obtained, no further record was made in the remaining 
series of the history of the gold used in these. 
After correction of the weighings for atmospheric buoyancy in such cases as seemed 
to the author to involve a correction worth noticing, the following results were 
calculated from the figures obtained, these figures agreeing in general closely with 
each other in each series : — 
Series a. Mean of 8 experiments. Atomic weight of gold = 196’022 
„ h. „ 4 „ „ „ lPG-143 
„ c. „ 9 ., „ „ = 196741 
„ d. „ 5 „ „ „ = 196743 
,, e. „ 4 ,, ,, ,, = 196-619 
„ / „ 4 „ „ „ = 196-620 
Leaving out the results of series h on the ground of the very small quantity of 
sublimed auric chloride available, and the considerable discrepancy of one of the 
results (that in which most material was used) from the rest, the author calculates 
from the remaining 30 experiments the general mean 196-669 ; but, taking into 
account the greater or less closeness of agreement of the figures obtained by the 
several methods, he comes to the conclusion that 196-64 may better be assumed as 
the true atomic weight of gold. In these calculations Ag was assumed = 107-660, 
Cl = 35-368, Br = 79-750, and K = 39-040. 
As regards possible sources of constant error in Kruss’s experiments, it may be 
observed that—• 
1. In series h very small quantities of sublimed auric chloride were used—the 
whole amount available for all four experiments being only about seven-tenths of a 
gramme—and it is probable that a little free chlorine may have been physically 
retained by the chloride in spite of the long-continued passage over it of dry air. 
The experiment in which the largest quantity of material was used gave the atomic- 
weight = but 194 - 79 . On these grounds the author himself excludes the series from 
consideration in calculating his general mean. 
2. In series c, d, and e the evidence is pretty strong, but perhaps not conclu¬ 
sive, to show that potassium auri-bromide can be rendered absolutely dry by exposure 
to air in a vessel containing phosphorus pentoxide, either at ordinary or higher 
temperatures, without, at the same time, undergoing any loss of bromine. The attain¬ 
ment of constant weight by the salt does not positively prove the entire removal of 
water. If moisture were retained the atomic weight of gold found would be brought 
out lower than it should be. 
3. Krdss himself observed that in all cases in which he dissolved potassium auri- 
bromide in water a small residue of metallic gold was left, and, determining in a single 
experiment the amount of this (about -05 per cent.), he used it as a corection for all 
