336 
MR. T. ROYDS ON THE CONSTITUTION OF THE ELECTRIC SPARK. 
are then rotated until they pass through the ends of a fine spectral line. A rotating 
photograph being substituted for the stationary one and the lines of the ruling again 
set parallel to the direction of travel of the microscope stage, readings can now be 
made. Adjusting by means of the stationary photograph also possesses the advantage 
that it is not necessary to assume that the air lines appear simultaneously at the top 
and bottom of the spectrum. 
Latterly the velocities have been obtained by taking enlargements (about 10 times) 
on bromide paper of the negatives and measuring these with the millimeter scale. 
The accuracy with which this can be done is, except for short lines indicating a large 
velocity, not so great as the error of experiment. Moreover, more detail can be seen 
in the enlargements than in the microscope even when using small powers, and the 
measurements can be made more rapidly. 
Accurate measurements of wave-length are not required, the finer air lines and a 
few well-known metallic lines were used to draw a dispersion curve from which any 
line could be identified. Measurements of the distances of the metallic lines from the 
air lines have, however, been made on the stationary photographs for comparison with 
the moving. 
5. The Photographs.—In most cases the photographs have been taken in the 
following manner. A spark was made with the film stationary. The motor having 
been started, the handle of the Wimshurst was turned by an assistant until about 
six sparks had passed. The interval between successive sparks would be about 
a quarter of a minute. Occasionally the stationary photograph was taken last. 
In all the photographs reproduced in the plates, the spark length was 8 millims., 
the upper electrode being the initially positive one. In each case the capacity of the 
circuit was 00306 mfd., and the self-induction that of the leads, the period of the 
oscillations being 3'8xl0 -6 second. The speeds of the film at the centre of the 
spectrum were about 90 metres per second. 
It was soon found that the most regular structure in the moving photograph was 
given by the lead lines XX 4387, 4245, and the bismuth line X 4260. The lead 
spectrum photographed on the moving film is shown enlarged about six times in 
Plate 28, fig. 2, and the bismuth spectrum in Plate 28, fig. 4, together with the 
photographs on the stationary film. 
There is a slight loss in the definition of the air lines in the moving photographs 
due to a broadening caused by the air particles remaining luminous for a short time. 
The air lines shoot straight across the spark gap, whilst the metallic lines are obliquely 
inclined at each electrode. The inclination of the metallic lines indicates the velocity 
with which the particles of metallic vapour travel away from the electrodes. The 
discharge is oscillating, the period of the oscillations being large enough for them to 
be separated in the photograph. 
Let us examine in detail the photograph of the two lead lines, XX 4387, 4245, shown 
enlarged about ten times in Plate 29, fig. 5. It is noticed that the streamers start 
