220 
MESSRS. W. SPOTTISWOODE AiSTD J. FLETCHER MOULT OX 
This experiment, then, proves that the instantaneous increase of tension arrives at 
the tinfoil nearer the air-spark end, which we will call A, sooner than at the other, 
which we will call B. Now what is the state of the tube at B at the moment when 
the rise of tension occurs at A ? We have shown that it is not affected by the rise of 
tension at A, but we have not otherwise ascertained its electrical state. This is ascer¬ 
tained in the following way. Fasten the piece of tinfoil B to a glass rod and let the 
wore that connects it be a fine wire running near to the tube during the whole of its 
course, so that the capacity of the system relieving A may depend almost exclusively 
on the capacity of B (Plate 20, fig. 26). Now if we move B as far as possible from 
the tube, keeping the wire straight, it will be found that the capacity of B to relieve 
A is diminished. In other words, by placing B on the tube we place it in a neigh¬ 
bourhood where there is a rising demand for the very electricity which is driven 
off from A. Suppose, for example, that the air-spark is in the positive. The elec¬ 
tricity that leaves A is positive ; and hence it follows that at the part of the tube 
upon which B rests there is a demand for positive, which amounts to saying that 
there is a rising negative potential there ; or in other words the quantity of negative 
electricity within the tube at B must be increasing. 
There would seem to be but one interpretation of this phenomenon. The negative 
terminal (in this case) is connected with its source of electricity during the whole of the 
interval between the pulses, and it must be pouring its electricity into the tube during 
all that time, thus raising gradually the quantity of negative electricity in the tube. 
When the pulse of positive comes ft sweeps up all the negative that it encounters and 
rushes out at the opposite terminal probably leaving the tube in a positive state. 
Indeed, this is so marked that we often see indications of a reverse current so soon as 
we introduce an air-spark, showing that the positive pulse is so violent that it causes 
over-relief. When this pulse has passed, the non-air-spark terminal, or (as we may 
better term it) the connected terminal, commences to pour out its electricity, undoing 
the effect of the pulse on the tube and gradually filling it with negative electricity till 
the next pulse comes. 
So much then for the condition of the tube in front of the electric pulse. We must 
next inquire what is the condition of the tube behind the pulse during the discharge. 
This is much more difficult, because we can no longer use the test above described, as 
we cannot tell whether the reaction which undoubtedly follows the relief-effect occurs 
during or after the passage of the pulse. 
If, however, we put several pieces of tinfoil on the tube and connect the one nearest 
to the air-spark terminal to earth,.placing a telephone in circuit, we shall find that the 
loudness of the tone is increased by connecting the other pieces of tinfoil with the first 
piece. This shows that their relief demands are additive, and as we know that the 
sign of the electricity in the relief-pulse is the same for all, and consequently that the 
pulse which constitutes the re-action must also necessarily be of the same sign for all, 
we see that the last piece must be in a condition demanding relief before the first has 
