ON THE SENSITIVE STATE OF VACUUM DISCHARGES. 
603 
the importance of these considerations is greatly enhanced by the fact that this relief- 
phosphorescence occurs in tubes of every degree of high exhaust, showing conclusively 
that when the air-spark is in the positive the discharge is carried in these tubes by 
bursts of positive electricity, which pass throughout the whole length of the tube just 
as in tubes of lower exhaust. No stronger confirmation of the results of Section XIX. 
could be desired so far as regards positive intermittence. 
Nor are these the only important conclusions that can be drawn from the appear¬ 
ance of this relief-phosphorescence. It shows that it is not necessary that there 
should be a discharge actually passing from out of a solid body to cause these streams 
of molecules. The discharge in question comes from the inner surface of the glass, 
not from the interior of its mass. This would go to show that the action takes place 
at the bounding surface of the terminal, or perhaps in the layer of gas that lies 
immediately upon it, and forms a kind of border-land between the solid and the 
gaseous, and that it is really an action between the gas and the solid terminal. It 
would seem clear that it does not lie in the free gas itself, or consist of an action 
between the particles of the gas merely, as we should then expect to find that there 
was phosphorescence on the surface of the glass from which the discharge proceeds, 
caused by the backward recoil of the particles of gas in that layer derived from their 
violent disruption from those of their fellows that go to form the molecular streams. 
And there is certainly no trace of this, for as we shall presently see, the spot from 
which the discharge comes is denuded even of the phosphorescence that it would 
receive from other sources. 
Before we pass on to examine this relief phosphorescence in other particulars we 
may remark that it is well shown, even in tubes of comparatively low vacua, by the 
use of the positive unipolar or double unipolar arrangement. By this method we are 
able to obtain intermittence of a much sharper and more violent type than with an 
effective current, so that the violence of the relief-effects is proportionally increased. 
A finger placed on such a tube will give well-marked phosphorescence. This serves, 
as in the former case, to show the correctness of the conclusions as to the cause of the 
unipolar phenomena, since it demonstrates that there are sharp periodical bursts of 
positive electricity into the tube. The relief-phosphorescence with these unipolar 
discharges is extremely bright, and the whole of the other relief-effects seem greatly 
intensified. By the use of this double unipolar arrangement we are able to produce 
phosphorescence in tubes of much lower exhaust than with an effective current, but 
both of these methods are inferior in this respect to the method described in Section 
XVII., which is a combination of a unipolar with an inductive discharge. 
Another very convenient arrangement for showing the phenomenon of relief-phos- 
phoresence consists in the use of a tube in which the terminals are near together at 
one end. The introduction of an air-spark will fill the remaining portion of the tube 
with a luminous discharge or haze (according to the degree of exhaustion), which is in 
fact a unipolar discharge (Plate 27, fig. 12). If the air-spark be in the positive circuit, 
