414 
PROFESSOR J. W. MALLET ON THE ATOMIC WEIGHT OF GOLD. 
Experiment. 
Character of 
gold used. 
Fraction of crystahized 
auri-bromide used. 
Gold. 
Silver required to 
precipitate Br. 
I. 
A, b 
First 
grm. 
5*7048 
arm. 
12A851 
11. 
A, b 
Second 
7'9612 
17-4193 
TTI. 
B 
First 
2‘4455 
5-3513 
IV. 
B 
Second 
4T632 
9-1153 
Of the tendencies to constant error which may be imagined in connection with the 
experiments of the first two series, and which have been noticed above, the first may 
probably be considered as not applying to the method 2 :)ursued in this third series, 
while the second and third might still be applicable. But the superior stability of 
the double salt constitutes an advantage in its favour, and, as it formed the chief 
material for the experiments of Kruss and of Thorpe and Laurie, a comparison 
with their results is desirable, the mode of treatment pursued by me in ascertaining 
the composition of the salt not having been quite the same as that used by these 
chemists. 
Fom'th Series of Experiments. 
A weighed quantity of trimethybammonium auri-chloride [N(CH 3 ) 3 HAuCl 4 ,] 
was decomposed by heating in the air, and the weight of the residual metallic gold 
determined. This trimethylamine salt was selected because the base is of simple and 
well established constitution, and may with reasonable probability be counted upon 
as obtainable in a state closely approaching purity, and because the gold salt is easily 
crystallized, possesses a considerable degree of stability, and contains approximately 
lialf its weio'ht of e’old, so as to offer the most favourable chance of detenninino' with 
accuracy the ratio between the metal left behind and the sum of the remaining 
constituents driven off on ifjnition. Althouo’h its use in fixino- the atomic weio-ht of 
O O O 
gold involves the atomic weights of three other elements—carbon, nitrogen, and 
chlorine—all three of these constants deserve to be ranked amongst those already 
known with the nearest approach to precision at present attainable. 
In order to obtain pure trimethyl-ammonium chloride, the impure commercial salt, 
derived from the vinasse of beet-root sugar making, was used, first setting free and 
diTilling off a considerable cjuantity of trimethylamine and condensing at about the 
right temperature, and subsequently purifying the product by Hofmann’s metbod of 
treatment with etliyl oxalate and renewed distillation. The purified trimethylamine 
was several times fractionally distilled, and the portion of correct and most constant 
boiling-point finally neutralised with pure hydrochloric acid. The concentrated solu¬ 
tion of trimethyl-ammonium chloride was now precipitated by a strong solution of 
auric chloride, the mother liquor decanted off, and the gold salt redissolved in hot 
water, and recrystallized several times. The bright yellow crystalline powder was 
