MR. T. ROYDS ON THE CONSTITUTION OF THE ELECTRIC SPARK. 
341 
The effect of self-induction has been found by lengthening and shortening the 
leads from the condensers to the spark electrodes. The results are given in Table III. 
In the first case the leads were as short as possible. 
Table III.—Effect of Self-Induction on the Velocity. 
Bismuth. Capacity = 0'03 mfds. 
Length of gap. 
Period = 2 • 82 x 10 6 sec. 
= 3 - 88 x 10 6 sec. 
= 4"98 x 10 6 sec. 
centim. 
0-8 
V = 242 
= 283 
= 288 
7. Constitution of the Spark. 
(a) Origin of the Velocity. —We now proceed to discuss the problem of the 
constitution of the electric spark, and deal first with the propagation of the vapour 
into the spark gap. We have the following possibilities :— 
(1) That the vapour diffuses towards the centre of the gap by the pressure 
suddenly developed at the electrodes. 
(2) That the particles of vapour are charged and move under the action of the 
electrical field. 
(3) That the particles are projected towards the centre. 
The velocity of diffusion (or effusion) of the metallic vapour will be proportional to 
the molecular speed and independent of the amount of matter vaporised ; indeed, the 
velocity will, if there is no loss of energy in eddy currents, by cooling or by internal 
friction, and if the pressure developed is sufficiently great, be identical with the 
velocity of sound in the vapour, as has been verified in effusion experiments. This is 
given in metres per second by the relation 
V = 64 (T fpff 
where T is the absolute temperature of the vapour and p its density compared with 
that of hydrogen, if we neglect the ratio of the specific heats. Assuming that the 
temperature of the vapour is not very different in different sparks, we should expect 
that the velocity of the vapour would be inversely proportional to the square root of 
the density. In the last column of Table I. are given the values of T calculated from 
this equation on the assumption that the metallic vapours are monatomic. The high 
values for lines of type II. show that the velocity of the particles emitting these lines 
is due to some other cause than diffusion. 
That the velocity is due to diffusion is certainly the simplest explanation of the 
phenomenon, and diffusion must, in any case, play a great part. It was not to be 
