PHYSICS: ROSA AND VINAL 
61 
have been settled so that at present, we believe no important questions 
are still outstanding. A few of the most important developments dur- 
ing the recent voltameter work may be mentioned. 
It was well known before the Bureau of Standards began its work 
that the deposits in the Rayleigh or filter-paper form of voltameter ex- 
ceeded those in the porous cup voltameters which were used in series 
with them. No satisfactory explanation of this difference had been 
given, and its existence had indeed been questioned. The most re- 
cent investigation prior to the London Conference appeared to show 
that under certain conditions the two forms of voltameter agreed. It 
was discovered at the Bureau of Standards that the difference could 
be doubled if the amount of filter paper was doubled and that when 
the filter paper was wrapped around the porous cup the results were 
the same as if the porous cup were not present. In short, it was defi- 
nitely proved that the filter paper itself was a source of serious error in 
the voltameter. A chemical study showed that the cellulose fibers 
of the filter paper are slightly oxidized in the air and the minute amounts 
of oxycellulose are soluble in water. These give rise to strong reducing 
agents in silver nitrate solutions and colloidal silver is formed. The 
colloidal particles carry electrical charges much smaller proportionally 
to their mass than do the univalent silver ions and therefore give rise 
to deposits in excess of the true deposit according to Faraday's law. 
They also alter the form and appearance of the deposit. 
Methods have been developed at the Bureau for the preparation 
and testing of silver nitrate of the exceptional purity which is necessary 
for voltameter work, because very minute impurities produce exag- 
gerated effects in the voltameter deposits. All forms of voltameter 
(except the filter paper form and forms using linen, cotton, silk, etc., 
as a septum) are in substantial agreement when used with solutions of 
the highest purity and with proper precautions against anode slime. 
The temperature coefficient with pure solutions was shown to be zero 
as it ought to be. The precaution taken by some observers to soak 
the silver deposits over night in distilled water to remove the last traces 
of electrolyte was shown to be harmful because it was discovered that 
silver in contact with platinum is appreciably soluble in distilled water. 
This was shown by repeated tests. The silver and the platinum differ 
slightly in potential so that a current passes from the silver to the 
platinum through the water. Although this current is very small, it 
was demonstrated by a sensitive galvanometer and the loss of silver 
from the deposit was measured. 
