ON ELECTRICAL DISCHARGES THROUGH RAREFIED GASES. 
195 
the latter case are directly due ro the inductive action of the actual electric variations 
within the tube near the. tinfoil, and will thus be proportional to the extent of these 
variations. In the former case they are directly due to the elevations of potential at 
the positive terminal which cause the discharge. In this case, therefore, they will be 
of equal strength throughout the tube ; in the other case they will be so only if the 
extent of the electrical variations is the same throughout the tube, which we shall 
hereafter show is in all probability not the case. And hence we find, on comparing 
the observations given above, that while in the special effect the intensity increases 
cceteris 'paribus with the distance from the air-spark terminal, in the relief-effect it is 
stronger the nearer it is to that terminal. We shall show hereafter that the instan- 
taneous variations of potential produced by the discharges are greater nearer the air- 
spark terminal than at a greater distance from it, so that it is natural that the relief- 
effects should be most marked in the neighbourhood of that terminal, while it is not 
strange that the non-relief-effects, in which the forces at work outside the glass are 
the same whatever part of the tube be selected, should be most marked at the parts 
of the tube where the electrical actions in opposition to which they act are the 
weakest. 
Having thus demonstrated the identity of these two classes of effects there remain 
the other two, viz.: the relief-effect when th e positive is the air-spark terminal, and the 
special or non-relief-effect, when the negative is the air-spark terminal. Both these 
should be produced, according to the theory, by pulses of negative electricity rushing 
to the tinfoil, caused in the one case by the inductive action of the positive discharge 
within the tube, and in the other case coming directly from the terminal of the tube 
which is in metallic connexion with the air-spark interval. 
To examine these effects a tube was taken, and a narrow ring of tinfoil placed round 
it which was connected with the negative terminal, and the air-spark interval was 
placed in the part of the external circuit leading from the machine to the negative 
terminal of the tube. The effect was to constrict the luminous column by the forma¬ 
tion of a dark space between the part of the interior of the tube directly under the 
tinfoil and the luminous column (Plate 17, fig. 13). The section of this dark space 
through the axis of the tube was in the form of a segment of a circle of which the 
centre would be in or near the tinfoil or within the glass immediately beneath it, so 
that the whole figure was that of a section of an anchor-ring by a co-axial cylinder. 
As the air-spark was increased the dimensions of this space also increased, till nothing 
was left of the luminous column within the part of the tube surrounded by the tin- 
foil but a very bright central line (Plate 17, fig. 14). 
It will be observed that this phenomenon is precisely what would be produced by a 
repulsion of the luminous column, the peculiar form being occasioned by the fact that 
the tinfoil was used in the form of a ring, so that the repulsion must take place 
symmetrically from all sides, and hence must produce constriction of the column 
instead of displacing it entirely. We shall, however, see that some caution is neces- 
2 c 2 
