266 
PROFESSOR H. A. WILSON ON THE EFFECT OF HYDROGEN ON THE 
7. A Comparison of the Negative Leak in Hydrogen with the Positive Leak 
in Oxygen. 
The positive leak in oxygen has been studied very fully by Professor 0. W. Richardson 
(‘Phil. Trans.,’ A, 413, 1906), who attributes it to the presence of oxygen in the 
surface layer of the wire, and this view is undoubtedly in accordance with the facts. 
I attribute the negative leak in hydrogen in the same way to the presence of hydrogen 
in the surface layer of the platinum. If this view is correct, the behaviour of the 
positive leak in oxygen might be expected to be closely analogous to the behaviour of 
the negative leak in hydrogen. As a matter of fact, there is an extremely close 
analogy between the two, as will be shown presently. 
In my previous paper I showed that treating a wire with nitric acid and heating it 
in air diminished the negative leak to a value very small compared with that usually 
obtained on heating a wire in a vacuum. This diminution was explained as being 
due to the removal of traces of hydrogen or other substances from the wire, and I 
suggested that possibly a complete removal of such traces would entirely destroy the 
negative leak. Since then Richardson ( loc. cit.) has described an experiment in 
which the leak from a platinum tube in air at atmospheric pressure was measured 
while hydrogen was allowed to diffuse through the platinum from the inside of the 
tube out into the air. He found that the small negative leak was unchanged by the 
hydrogen. I have repeated this experiment and confirmed his result. As Richardson 
points out, this experiment shows clearly that the very small negative leak given by 
platinum in air is not due to traces of hydrogen in the platinum. In addition, 
measurements of the negative leak from clean platinum wires in air and other gases, 
except hydrogen, give fairly concordant results, so that I think there is now no doubt 
that there is a small negative leak due to the platinum itself. In fact, since traces of 
hydrogen in the surface of the platinum must be burnt up at once when it is heated 
in air, it is almost impossible that such traces could be the cause of the small leak 
given in air. This conclusion, of course, does not ajiply to the much larger leaks 
which are obtained when an ordinary platinum wire is heated in a vacuum without 
any special precautions. I think there is no doubt that they are due, as I suggested, 
to traces of hydrogen. In the present paper, when the negative leak in hydrogen is 
referred to, it is not intended to imply that there is not a small negative leak due to 
the platinum itself. 
To return to the comparison of the positive and negative leaks, the variation with 
the pressure at constant temperature will be taken first. Richardson shows that the 
positive leak increases with the pressure of the oxygen at low pressures rapidly, but 
at high pressures more slowly. The relation between the leak and the pressure is 
x — ap n /(/3+p n ), where a, j3, and n are constants. The constant n varies with the 
temperature, being jr at 820° C. and 1 at 1170° C. Assuming that the leak depends 
on the amount of oxygen in the surface layer of the platinum, this shows that the 
