596 
MESSRS. W. SPOTTISWOODE AND J. FLETCHER MOULT OH 
obtain direct evidence of the truth of the proposition in question.* If we place a 
somewhat broad piece of tinfoil round the tube, and connect two wires from it to the 
ends of a suitable tube of moderate exhaust, we shall get clear signs of the appropriate 
double unipolar discharge (Plate 27, fig. 11). The sole drawback of this test is that 
it is only applicable in cases where the action is of considerable violence, so that it 
gives no result where the air-spark is very small, or where the tinfoil is very near to 
the terminal remote from the air-spark. This proof can, however, be extended to 
cases to which it is not directly applicable, by passing to them in a continuous maimer 
from cases in which it can be used. Thus, where the action is sufficiently violent near 
the air-spark terminal to enable us to use the unipolar test, but not sufficiently strong 
at points further removed from that terminal, we can show that the nature of the 
electrical disturbances at the latter is the same as at the former, by passing the finger 
along the tube and observing that the appearances at the different points are sub¬ 
stantially identical. 
There is another method by which we can raise a strong presumption as to the 
applicability to high vacua of the principles we established in our previous paper. 
This is by observing continuously the phenomena presented by the discharge during 
the process of passing from a low vacuum to a high vacuum while the tube is being 
exhausted. If an intermittent current of either type be allowed to pass through the 
tube during the whole of the operation, the phenomena observable in low exhausts 
will pass in such a gradual and continuous way into those which we are accustomed to 
meet with in high exhausts, that it becomes well nigh impossible to doubt that the 
modus operandi of the discharge is the same throughout. And in the same way we 
can extend the test from the case of a large air-spark to that of a small one. And if 
it were not that the direct methods of which we are about to speak render these less 
direct evidences unnecessary for the establishment of the truth of the proposition in 
question, such considerations as these would be of the highest value as raising a strong 
presumption in favour of the radical identity of the modes of discharge in the two 
cases. As it is, however, we need not dwell on them further, and we have only 
referred to them in order to show that the methods of our former paper would have 
enabled us to solve the difficulties of the new subject-matter with which we are dealing 
had it been necessary that we should have recourse to them. 
All the foregoing evidence, though valuable as confirmation of the theory, and 
interesting in connexion with our previous results, is insignificant in importance com¬ 
pared with the direct evidence afforded by the standard-tube method described in 
Section XY. A tube of high exhaust is taken, and its terminals are connected with 
those of a Holtz machine. A patch or ring of tinfoil is placed anywhere upon the tube, 
except in immediate contact with either of the terminals, and a wire is taken from it 
to a ring of tinfoil upon the standard-tube. No effect will be produced on the 
standard-tube unless the high vacuum tube is of a nature to cause by its own action 
* See Phil. Trans., 1879, p. 216. 
