PROFESSOR J. W. MALLET ON THE ATOMIC WEIGHT OF GOLD. 
427 
tapping of the apparatus to get rid of the effect of irregular adhesion of the liquid to 
the walls of the voltameter tubes. 
Fig. 6. 
It will be seen that, with the arrangement described, the electrolysis could he 
effected of acidulated water, thoroughly deprived in advance of dissolved air, and in 
quantity so small as to be capable of i-etaining in solution but infinitesimal quantities 
of the hydrogen and oxygen electrolytically sej^arated. As the decomposition 
proceeded, the quantity of liquid in the voltameter tubes could be maintained constant 
by opening the stopcock c, with a supply of mercury in the funnel a, and then 
cautiously opening f, so as to feed forward a little of the air-free water from the 
cylinder h, thus leaving the proportion of acid unaltered. The surface presented by 
the platinum wire electrodes was so small as to allow of occlusion of the gases to 
only an extremely minute extent, and both the hydrogen and oxygen were allowed to 
escape for some time before any was collected for measurement. 
The hydrogen only was collected and measured. I had hoped to apply this form of 
voltameter to a more exact determination of the relative volumes of hydrogen and 
oxygen derived from water by electrolysis than is possible with the voltameters of 
more common construction. But I have not yet seen my way to getting over the 
difficulties connected with the presence of ozone, hydrogen dioxide, Berthelot’s per- 
sulphuric acid, or other by-products in the oxygen gas evolved at the positive pole. 
If this could be accomplished, a useful contribution might possibly be made to the 
question, revived and worked upon of late by several chemists, of the exact atomic 
weight of oxygen. The vessel for collecting and measuring the hydrogen, shown in 
fig. 7, consisted of a spherical globe of tolerably stout glass, with a capacity of about 
250 c.cm., having a neck of about 1 cm. internal diameter, and 22 cm. long. This 
neck had etched upon it a simple linear scale of millimetres. At the mouth it was 
fitted witfi a well ground perforated glass stopper, forming part of a glass stopcock 
with an outer orifice of about 1 mm. bore. The exact capacity of the whole globe 
and neck was ascertained by heating it in an air-bath to remove air and moisture 
condensed on the interior surface, drawing dry air through with an aspirator, then 
filling the globe vith heated mercury, allowing it to cool to an accurately noted 
3 I 2 
