196 
MESSRS. W. SPOTTISWOODE AND J. FLETCHER MOULTON 
sary before we identify it with the familiar phenomenon of repulsion which lias been 
described as so characteristic of positive relief-effects. 
To establish the identity of this effect with the relief-effect when the air-spark is in 
the positive, the connexion between the negative terminal and the tinfoil was removed, 
and the tinfoil was connected to earth and the air-spark placed in the positive external 
circuit. On varying the size of the air-spark as in the case just described, precisely 
the same series of phenomena were presented except for very minor peculiarities 
of definition. 
Inasmuch as in neither of these effects had there been any appearance of the blue 
discharge which, as we have seen, is so characteristic of the relief-effect when the air- 
spark is in the positive circuit, we examined the tube and found that it did not give 
this effect, or, at all events, it did not give it with the kind of current that was then 
being used. We therefore took a nitrogen tube in which we had often observed it, 
and repeated the experiment. No sooner was the air-spark interval of sufficient length 
than the blue discharge appeared, whether the one or the other of the two effects was 
being observed. In both cases the effects were at first repulsion, but when the action 
became more violent it passed into the blue discharge effect. The identity of the 
effects in the two cases was thus placed beyond all doubt. Indeed, the only difference 
that can be detected is that the definition is not equally sharp in the two cases. 
On testing these tubes by moving the ring of tinfoil up and down the tube it 
was found that the blue discharge had a greater tendency to appear when the ring was 
near the positive than when it was near the negative terminal of the tube, thus 
confirming the law arrived at in the former case : that relief-effects are strongest near 
to the air-spark terminal, and non-relief effects are strongest at a distance from that 
terminal. 
On looking closely at the blue discharge under the tinfoil, it is seen that it does not 
lie close to the glass, but is separated from it by a little blank space, precisely as the 
luminosity on the negative terminal of a tube is separated from the terminal itself by 
the small non-luminous space known as Crookes’ space. Thus the inside of the glass 
beneath the tinfoil acts precisely as a negative terminal. In contrast to this we find 
that in the cases in which the effects are produced by positive pulses, the luminosity 
starts sharply from the surface of the glass itself, just as the glow on the positive 
terminal appears to start from the very surface of the terminal itself. Nor is this the 
only point of resemblance between the effects produced by external negative pulses 
and the appearances of negative terminals. Any one familiar with the appearances 
presented by negative terminals shaped like a ring will know that they form round 
them a blank space, driving off the glow till it at length appears at the centre of the 
ring only, where it forms a bright line. The resemblance between these appearances 
and the effects above spoken of, produced by negative pulses, is so close that, so far as 
the authors of this paper are able to judge, these last-named effects are substantially 
identical with what would be produced were there a terminal in the tube consisting of 
