10 
MR. J. W. CAPSTICK ON THE RATIO OF THE SPECIFIC HEATS 
If the dust is uniformly spread, this want of symmetry has no influence on the 
measurements. 
The method I employed for putting in the dust was to draw through the tube a 
dry cloth which cleaned out what was left from the previous experiment and slightly 
electrified the glass. The tube was then placed in a sloping position, and the powder 
poured in at the end through a small funnel, and allowed to run through gently in a 
narrow stream. It was then turned end for end, and the dust poured through in the 
opposite dmection, the result of which was that the small electrification caused as 
much to adhere to the glass in a narrow uniform strip as served for the purpose of 
the experiment. 
When it was placed in position and the figures were to be made the tube was 
turned round, so that the dust did not lie along the bottom but was a little way up 
the side; then, on rubbing the vibrator with a piece of wet flannel, the powder ran 
down to the lowest point everywhere but at the nodes, which were left as clear 
spaces, narrow and sharply bounded, separating rectangular patches of dust of great 
regularity. 
Fig. 2 shows part of a set of figures obtained with isopropyl bromide. 
Fig. 2. 
For the measurement of the figures two parallel platinum wires were carried 
on a framework sliding along a steel scale divided to millimetres (fig. 2). These 
wires were placed so that the tube lay between them, and their plane passed through 
the centre of the clear space at the node, and the position of the framework was read 
on the scale, tenths of a millimetre being estimated with the help of a lens. When 
the figures Avere of average quality the setting could be repeated so that the positions 
did not vary by more than two or three tenths of a millimetre. 
Table I. gives a typical set of measurements. They were made in one of the 
methyl bromide experiments, and are chosen as being neither the best nor the worst 
of the sets, but a fair average. 
The first column gives the scale-readings, and the second column the half-wave- 
lengths got by subtracting the consecutive readings from each other. The first half- 
