MR. T. ROYDS ON THE CONSTITUTION OF THE ELECTRIC SPARK. 
342 
expected that the velocity of the vapour in the spark should be anything but 
approximately inversely proportional to the square root of the vapour density, for the 
disturbing influences may be considerable. Eddy currents of considerable magnitude 
apparently take place, as seen from the irregularities not only of the structure of the 
spectral lines but also of the “aureole” in the photograph of the spark itself. Since 
the vapour does not flow into vacuous space, # the velocity of the vapour will be 
retarded, though it is difficult to estimate the magnitude of this effect. The 
experiments on effusion show that the velocity does not fall appreciably until the 
back pressure approaches one-half that causing the effusion. However, if we increase 
this forward pressure by increasing the amount of matter vaporised, we shall reduce 
the effect of back pressure, and the velocity will probably, therefore, be dependent on 
the boiling-point of the metal. Attempts were made to find whether the amount of 
matter present had any effect on the velocity by introducing a small globule of 
mercury on one of the electrodes. The globule was small and probably disturbed the 
electric field, but it was seen that the velocity was 125 met./sec., compared with 
936 met./sec. when a cup of mercury formed the electrode. The amount of matter 
vaporised is probably least for metals of high boiling-points, and the order of boiling- 
points is found to be approximately the reverse of that of velocities. 
A second hypothesis is that the particles of vapour are charged. It is well known! 
that the resistance of the spark varies with the material of the electrodes between 
which the spark passes. We must, therefore, suppose that the metallic particles are 
participating in the carrying of the current, i.e., that the vapour is ionised, or that 
the quantity of air which is ionised varies with the material of the electrodes. The 
period of the oscillations depends on the resistance of the circuit, and a change in the 
resistance should produce a corresponding change in the period. The periods for 
different electrodes are given in the following table; all lines in the same spectrum 
show the same period :—- 
Magnesium . . . 3 • 70 x 10~ 6 sec. 
Aluminium „ . . 3‘89 
Cadmium . . . . 3 • 7 4 
Tin. 3-73 
Antimony .... 3 - 76 x 10 _G sec. 
Mercury. 3'88 
Lead.3 • 84 
Bismuth.3 ■ 88 
It cannot, therefore, be said with certainty that the period varies with the material 
of the electrode. 
Now, if the metallic vapour is ionised, then, as the direction of the electric field 
between the electrodes alternates, the vapour should be alternately impelled in one 
direction during one half of the period and in the opposite direction during the second 
half. There might be some lag of the reversal of the velocity behind the reversal of 
the field, for the kinetic energy already generated by the previous oscillation would 
* There will be a partial vacuum due to the sound-wave, 
t Fleming, “ Principles of Electric-wave Telegraphy.” 
