ON ELECTRICAL DISCHARGES THROUGH RAREFIED GASES. 
213 
yoking a response from the other terminal. And this method shows us that in the 
ordinary case of the air-spark discharge it passes right through the tube before this 
response comes, if in fact it can be said to come at all. One of the most important 
consequences that necessarily follows from this is that the discharges at the two 
terminals of the tube are entirely independent, and are primarily determined each by 
the conditions at its own terminal and only in a secondary degree, if at all, by the 
conditions that exist at the opposite terminal. We shall hereafter recur to this point, 
and shall show still more conclusively the complete independence of the discharge at 
either terminal of a tube, but it will be well to call attention here to the very 
important consequences that flow from this principle. Since the discharges are 
not identically the same at both terminals, the tube must contain different free charges 
at different times.* It is therefore in no respect like a conductor, but is an indepen¬ 
dent electrical system, holding much the same position as the air-vessel in a forcing- 
pump. All the electricity that comes into it goes out again, but the truth of this can 
only be asserted when we consider the whole discharge from the beginning to the end, 
and it may not be even approximately true during a small finite time. 
This independence of the discharges from the two terminals is a most important fact 
in the analysis of the mode of passage of electricity through rarefied gases. It 
dissipates the error of seeking analogies in metallic conduction ; and shows that any 
appearance of obedience to regular laws as to change of potential as we proceed along 
the tube, resistance, &c., must arise from the fact that the effects measured are really 
average effects over a space of time enormously long compared with the duration of 
* In order to show this conclusively, a vacuum tube was enclosed in a metal canister (the wires passing 
to its terminals through tubes of insulating material inserted in small holes in the canister) and a telephone 
was placed in circuit between the canister and the earth. When a discharge with an air-spark in the 
external circuit was sent through the tube a sound was heard in the telephone similar to that made by the 
air-spark. By a fundamental proposition in electricity the free electricity on the surface of the canister 
(and which escaped through the telephone to the earth) was at any instant equal to the excess of one 
kind of electricity over the other in the space within the canister. Had the discharge been in the nature 
of conduction, as in a galvanic current, there would at no instant have been an excess of either kind of 
electricity, and therefore there would not have been any sound in the telephone. The existence of a sound 
testified to variations in the algebraical sum of the free electricity in the tube. To show that this was not 
due to anything depending on the wires leading to the tube, the same experiment was repeated with a tube 
in which the middle portion was connected with the two end-portions by very narrow passages. The 
middle portion was placed in the canister and the narrow parts passed through small holes made in its 
side, so that only a portion of the complete tube was within the canister. The same results were obtained 
with this arrangement. 
By these experiments, together with those previously described, in which a telephone is placed in circuit 
between the earth and a piece of tinfoil upon the tube, we are able to obtain from the telephone direct 
evidence of the intermittence of the discharge at the time of its actual passage through the tube. This 
evidence fully confirms that previously obtained by the use of the revolving mirror. It is, however, of 
inferior value, for though it shows conclusively that there are rapid periodic alterations of intensity or 
character in the induction round the tube, it does not enable us to decide whether these are due to inter■= 
missions or only to fluctuations in the discharge. 
