430 
PROFESSOR J. W. MALLET ON THE ATOMIC WEIGHT OF GOLD. 
volume of hydrogen ; but this, on account of the time required, would have made an 
experiment exceedingly troublesome and difficult. 
In the work of this series the same unsatisfactory need for selecting only such 
results as came fairly close to the figures expected, and rejecting several others on the 
ground of very considerable departure therefrom, and the same sources of possible 
constant error in regard to the gold deposit present themselves which have already 
been noticed under the head of the fifth series. As regards the hydrogen, one is led 
to consider j^ossible diffusion of hydrogen and oxygen between the two little volta¬ 
meter tubes, and slight imperfection in the drying of the hydrogen obtained. The 
former would, on the whole, probably tend to diminish the volume of gas collected, 
and lienee to raise the apparent value of the atomic weight of gold. The latter would 
have the opposite tendency. That neither can have had more than an extremely 
minute influence was fairly proved by testing a part of the hydrogen obtained, on the 
one hand by passing it through a red-hot glass tube, and on the other by submitting 
it to more extended drying by contact with phosphorus pentoxide both before and 
after such heating; in neither case was there appreciable change of volume. 
Notwithstanding the desirability of comparing the atomic weight of any other 
element dhectly with that of hydrogen, the difficulty is not to be overlooked of doing 
this for an element having so high an atomic weight as that of gold. There is a 
manifest objection to the necessity of dealing with such minute quantities of hydrogen 
as those concerned in these experiments. A very small error in the determination of 
the hydrogen greatly affects the value found for an atomic weight nearly two hundred 
times as large. It is true that the measurement of the volume of the hydrogen admits 
of being made with such precision as to leave room for but an extremely minute error 
in the corresponding weight, yet this measurement is not one of limitless delicacy, 
particularly if the difficulty be properly appreciated of ascertaining with certainty the 
precise temperature of the gas at the time its volume is read. Moreover, in measuring 
the volume of the gas, and thence deducing its weight, there is need not merely for 
a knowledge of changes of temperature and pressure, but for absolutely correct 
readings of the barometer and thermometer, so that there must usually be a degree of 
hesitation in accepting the readings of even fairly standard instruments, when tem¬ 
perature and pressure come to be placed in comparison with these conditions as 
afiecting the results of Regnault for gaseous density. Nor can the results of that 
great physicist be assumed as themselves free from all possible need of further 
correction. 
The error of direct comparison with so small an atomic weight as that of hydrogen 
is, however, after all only masJeed by substituting an indirect comparison through 
some larger atomic weight, since the assumed value of the latter is uncertain within 
limits which depend iqDon its conq^arison with the atomic weight of hydrogen. 
