m THE PRESENCE OP DUST-FREE AIR AND OTHER GASES. 
2fi9 
surface of an indiarubber cork K. This closes the opening of the tube in a perfectly 
air-tight manner, when the air above it is at atmospheric pressure, and the pressure 
in F is a small fraction of this, as it always was maintained during a series of 
experiments. 
K can be pulled up by means of the vertical wire shown in the diagram, thus 
rapidly making free communication between D and G and causing the expansion. 
The tap T must of course previously be closed. In order that one may be able to 
work the arrangement from the outside without admitting air, the wire passes up 
through a thin-walled indiarubber tube, M, closed at its upper end by a cork in 
which the end of the wire is fixed. A cord attached to this cork and passing over 
a smooth peg fixed vertically above it, enables the observer to make the expansion, 
while watching the behaviour of the air in the cloud chamber A. 
The tubes D and G had an internal diameter of about 8 millims., so that the fall of 
pressure in the air-space in the upper part of B must be very rapid. The valve E, 
which prevented the water in the bell-jar B from following the air into the tube I), 
was made by cutting a thin slice from the end of an indiarubber cork and supporting 
it with the smooth surface uppermost on a piece of cork just thick enough to float it. 
It was fixed by means of a wire hinge to the lower surface of the indiarubber cork 
of the bell-jar, in such a position that when raised by the water reaching it, it 
covered the hole bored for the tube D. The latter did not quite reach the lower 
surface of the indiarubber cork, so that when the valve closed the contact was 
between two indiarubber surfaces. 
This valve was found to work j^erfectly when the excess of pressure below was 
sufficient, and smooth indiarubber surfaces were used. 
The bell-jar had a diameter of about 14 centims. and was about 30 centims. high. 
The inner vessel had a diameter of about 9 centims. A vertical glass scale, divided 
into millimetres, fixed by means of sealing-wax to the outside of the inner vessel, 
enables the observer to note with the aid of a telescope the level of the water before 
and after expansion. From a subsequent calibration the initial and final volumes are 
obtained. 
To make visible any condensation in the form of fog or rain, the light from a 
luminous gas-flame is brought by means of a convex lens to a focus at the centre of 
the cloud chamber A. Any condensation which may result is most distinctly seen 
when the eye is placed just out of reach of the directly transmitted light. This 
method of illumination was used both by Aitken and R. v. Helmholtz. The 
experiments were performed in a dark room. When an expansion was made the 
only source of light was that mentioned above. After the result had been noted, 
light was admitted by raising a shutter in order that the readings of volume might 
be made. 
The inside of the inner vessel, A, was cleaned before use, first with caustic potash 
and then with nitric acid, and well washed with distilled water. After this treatment 
