402 
PROFESSOR J. W. MALLET OS THE ATOMIC WEIGHT OF GOLD. 
General Results of Former Determinations most deserving Confidence. 
TJiese recent researches, unquestionably by far the most valuable up to the present 
time, give us, when taken separately and together, the followung values for the atomic 
weight of gold :— 
1 . General mean of 5 series by Keuss, as calculated by himself . 196‘640 
2. ,, ,, 3 series by Thorpe and Laurie, as calculated 
by themselves.. . . 196‘852 
3. ,. ,, 1 and 2, giving equal value to each .... 196’746 
Difficulties to he overcome in Determining the Atomic Weight of Gold. 
Besides the special difficulties connected with each method which may be adopted, 
the determination of any high atomic weight ^vith a degree of accuracy wdiich enables 
the result to be accepted to a given decimal place is cleaiiy a much less easy matter 
than would be the attainment of an apparently equal degree of precision for an atomic 
weight represented by a small number. In obtaining the atomic weight of lithium, 
the first with which, many years ago, I had any personal experience, a difference of 
unity in the first decimal place corresponded to about 7 -, 7 th of the whole value con¬ 
sidered to be correct. In getting the atomic weight of aluminum, worked on later, a 
like difference represented approximately -a^tli of the whole value. But, in the case 
now considered, of the atomic weight of gold, unity in the first decimal place moans 
but about TWi ffih of tlie whole value. So that, looking at the matter in this light, it 
may be said that a degree of precision is demanded more than seven times as great as 
in the case of aluminum, and twenty-eight times as great as in the case of lithium. 
There is also to be noticed, as the most obvious general difficulty to which all 
methods for determining the atomic weight of gold are more or less exposed, the 
instability of compounds of this metal ; not merely the ease with which complete 
decomposition occurs, with separation of free gold, but the much more insidious and 
less easily detected troiible arising fi'om the comparative ease with which aurous pass 
into auric compounds, and the reverse. 
New Experiments hy the Author. 
The general difficulties just alluded to, and the special points to be investigated in 
regard to each method of determination tried, have demanded much time and work, 
and I cannot feel even now that all has been done that is desirable and possible; but 
the experiments projected have been so far completed as to seem to justify publication, 
and I am not likely soon to be able materially to extend them. 
