G32 
MESSRS. W. SPOTTISWOODE AND J. FLETCHER MOULTON 
and the relief-phosphorescence is found to be swept downward along the tube, while 
the interference between the two streams is greatly increased for the reason that the 
main parts of both come into collision with each other. 
Two other experiments in connexion with interference should also be considered in 
dealing with this matter. If a shadow of a small piece of conducting material, such as 
a piece of wire enclosed within the tube, be cast on the side of the tube by relief- 
phosphorescence and a conductor be brought in contact with the tube near to one end 
of the wire, the shadow of the whole wire, including the other end of it, will bulge 
considerably. This bulging is exceedingly black, showing that the whole of the 
streams that passed close to the wire on their way to produce the relief-phos¬ 
phorescence have been diverted from their course. Thus the discharge from the wire 
must have continued during practically the whole of the time that the relief molecular 
streams were passing it. Now the discharge from the wire must have commenced at 
the same time as that from the side of the tube, for they were both in response to the 
advancing positive electricity, and they will presumably last for an equal time ; hence 
the time required for the molecular streams to cross the tube and arrive at the wire 
is not an important part of the period during which the discharge lasts, for otherwise 
the relief molecular streams that passed after the discharge from the wire had ceased 
would probably have shown themselves in the form of phosphorescence on the bulged 
part of the shadow. 
The indications derivable from the second experiment to which we are about to 
refer are yet more distinct. In the experiment described in Section XXII., in which 
a phosphorescent image was formed of a small hole in an intermediate terminal (the 
air-spark being in the positive), it was found that this image was splayed oat by the 
finger being placed on the tube. Now a magnet displaced it as a whole without any 
splaying out (Plate 28, fig. 23). This, then, pointed to a variation in the relative strength 
of the interfering stream and the stream interfered with, and such variation must have 
occurred during the period that they were encountering one another, and were moving 
in the ordinary way of such streams, for it showed itself in a variation in the extent to 
which the streams from the negative terminal were diverted. We may hence conclude 
that the time requisite for the molecules to move the length of the tube was decidedly 
less than that occupied by the discharge, but was sufficiently comparable with it to allow 
the diminution of intensity of the streams from the side of the tube to make itself 
visible before the streams from the negative terminal experienced a similar diminution. 
Thus far we have only been dealing with positive relief. But the phenomena of 
negative-special are equally important in the demonstration of the truth of this theory. 
In that case we know that the impulses that cause discharge arrive at the negative 
terminal of the tube and at the tinfoil synchronously, for the difference (if any) in the 
time required to pass from the machine along the wires to the two places is incom¬ 
parably smaller than any of the quantities with which we have to deal.* And yet we 
* This can be seen by the fact that no difference is produced by making the path to the tinfoil longer 
