NoOvEMBER 12, 1896] 
found with the Wimshurst used in this way are, first, that after | 
a rest or complete discharge the polarity of the machine is apt 
to reverse, the cure being either (2) to completely discharge 
again and trust that the desired reversal will take place on re- 
starting, a very uncertain plan; (4) to interchange the wires, 
probably best by disconnecting them at the tube and turning 
the tube suitably ; or (c) by discharging completely and starting 
the electrical action of the machine by holding in some position, 
found easily by experiment, a body electrified and whose elec- 
tification is of known sign—e.g. a glass rod rubbed quickly 
and lightly on a piece of silk or flannel. The second fault 
is that in this way of using the Wimshurst, whether the 
condensers are in circuit or not, after a time the character | 
of the discharge is apt to alter. Beautiful patches and streaks of 
very bright grass-green make their appearance on the inner sur- 
face of the glass ; and when one of these has formed, the \-ray 
discharge becomes weaker, and two or three are altogether fatal 
to it, even if the machine be in excellent working order. When 
once a patch is formed, too, it seems to cling obstinately to 
the place where it develops. It was suggested to me that these 
patches were invariably associated with bubbles in the glass of 
‘the tube, but on touching the tube lightly with a camel’s-hair 
brush dipped in rather dry Indian ink, and thus painting over 
these green spots whilst they were visible, and afterwards re- 
moving the ink spot by spot and examining for bubbles, I found 
in many cases they were not present, and also that there were 
many bubbles where no fluorescent spot had been painted. These 
spots lie in the shortest lines which can be drawn on the glass 
from the kathode to the anode. They are far more numerous 
mear the kathode, forming near it, and often gradually elongating 
towards the anode. Blowing gently on the middle part of the 
tube causes their retreat towards their birthplace, and often ex- 
tinguishes them altogether. Their forms seem to be those which 
slightly divergent jets of matter would take if they were thrown 
out from different points on the sharp edges of the kathode and 
impinged on the inner surface of the bulb very near the kathode, 
exciting fluorescence so long as the matter maintained a velocity 
above a certain limit. If this is so, and the matter is negatively 
charged, and if the glass surface acts to it as a rough surface does 
to a stone, perhaps on account of its high electrical resistance, it 
is easy to see why the patches are small and slowly spread when 
a higher E.M.F. is used, and how it is that the dissipation of 
the positive charge which lies on the outside of the bulb leaves 
these jets of negatively electrified particles more free to take a 
shorter path tothe anode. Thus, if they hit the glass at all, 
they will do so further and further from the kathode ; and this 
explains why the fluorescent patches travel towards the anode 
region of the bulb, when the neighbourhood of the kathode is 
deprived of its positive charge by gently breathing on the glass 
in this part of the bulb. Similarly it also explains why the | 
patches travel kathode-wards when the middle and anodal region 
is drained in any way of its positive charge, for then the at- 
traction in the kathodal region is practically strengthened by the 
destruction of the attraction to other glass parts of the tube. 
Not only do these jets impinge on the inner side of the glass just 
by the kathode under ordinary circumstances, when no pains are 
taken to drain the external positive charge of the tube, but they 
are partly reflected, and, owing to the attraction of the positive 
charge, the path of the reflected particles is bent towards the 
glass surface, and in many cases a fainter patch of longer 
shape is formed further from the kathode, and I believe a 
third may be, so that the inner surface of the tube tends to 
become filled with patches smallest, brightest, nearest together 
and least elliptical, near the kathode; and becoming pro- 
gressively larger, fainter, further apart, and longer in proportion 
the nearer they approach the anode. This has a very important 
application, for seeing that the formation of the patches is, by 
experiment, inimical to the creation of the X-rays, anything 
which tends to their formation must be avoided; and if, as lam 
convinced, they are due to the sharp edges and small irregulari- 
ties of the circular kathode, no pains should be spared to give 
that electrode a perfect polish on the side facing the anode, and to 
make its edge circular, and not square, as the circular kathode’s 
edges are in my tubes. Another way of partly overcoming this 
difficulty is to make the tube very wide round the kathode, or, 
which amounts to the same thing, push the kathode well forward 
into the globular part of the tube. This last is done in some 
tubes, I believe, with excellent results. A striking experiment 
_is to arrange a tube and Wimshurst to give the green patches, 
and then to breathe rather strongly on the kathode half of the 
NO. I41I, VOL. 55] 
WA TRORE ai 
re) 
tube two or three times if necessary, and watch how the phos- 
phorescent patches fly towards the anode as the external positive 
charge is dissipated and, at last, just meet behind the anode (which 
region becomes brightly fluorescent); finally, if the positive 
charge is sufficiently dissipated, part creeping and part bounding 
along the glass tube which, in my tube, encloses the anode’s wire 
support, into which support, or the actual back of the anode, 
they finally yield their negative charge. I think that any one, 
who will take the trouble to make the very easy experiments 
here recorded for himself, will feel convinced that this is the true 
explanation of many of the—so far as I know—widely observed, 
but unexplained variations in the behaviour of the X-ray tubes, 
and the last experiment suggests that it would be a good thing 
to (i.) completely surround the anode stem and the back of the 
anode itself with some insulator—glass would probably prove best 
—and if the anode be platinum, to make the anode’s edges very 
blunt and smooth ; and it may be well, but of this I cannot feel 
certain without experiment, to make a small area, large enough to 
include the point from which the X-rays seem to emanate, rough 
with, say, platinum black, in order that over this area, to put it 
in old-fashioned language, the density of the positive electrical 
charge may be as great as possible. 
In these ways, perhaps, the stray negative ‘‘ jets ~ may be pre- 
vented, and a tube made to emit the X-rays more easily and 
steadily than any at present used. 
It seems natural, from these conclusions, to suppose that 
pushing both anode and kathode far into the tube so that they 
are fairly close together, is a very good plan. for it would cer- 
tainly tend to prevent leakage to the internal surface of the 
glass. I have heard that very excellent results have been 
obtained with tubes in which the electrodes are but a very few 
millimetres apart. The same reasoning would indicate that it 
would be well to make the kathode convex towards the anode, 
and fairly small, but not very small ; for each tube there will be 
aspecial size which will be best. 
Whatever the nature of the ‘‘ jets” may be —and I suppose 
most will be inclined to believe (with Crookes) that they are par- 
ticles of the residual air, for there is very little evidence of any 
scattering of an aluminium kathode, save just round its edge—it 
is clear that the effect of the external positive charge is to create 
a higher vacuum in the central portion by drawing the particles 
to the sides of the tube; and this accounts for the action of a 
flame on the glass bulb, which is two-fold—for it not only drives 
off from the sides, and possibly from the electrodes, particles of 
moisture or occluded gases held by some force other than 
electrical, but also speedily dissipates the external charge, and 
thus frees the ‘‘ electrically bound” molecules, distributing the 
matter more evenly over the internal space, and thus making the 
passage of electricity between the electrodes easier. At the 
same time the phosphorescence, which is produced by their 
impact on the glass, becomes evenly distributed over the tube, 
and itis whilst hot that the X,-rays are emitted (to which wood is 
transparent, but flesh opaque); and it should be particularly 
noted that these X,-rays are in this case emitted when the tube 
offers less resistance to the electrical discharge than when the 
X,-rays (which penetrate flesh, but to which bone is fairly opaque) 
are being emitted. 
Several observers—in particular, Mr. Swinton—have men- 
tioned that under some conditions a tube is capable of emitting 
rays to which bone is almost as transparent as flesh. These I 
shall call throughout the rest of this paper X3, and in the experi- 
ment Iam about to describe it will be found that it is possible— 
and, indeed, very easy—to cause a tube to emit either X,, Nj, or 
X,-rays ; and not only so, but inasmuch as the change from X, 
through X, to X, is perfectly continuous, it is simplest to believe 
they differ, not in kind, but only in some one inherent quality, 
such as frequency, which varies continuously. I may say, before- 
hand, that the following experiment seems to me a very im- 
portant one, and the study of it likely to lead to some definite 
conclusions as to the undulatory nature, and even evaluation of 
the wave-lengths of the X-rays. ; 
Two small Leyden jars then are attached to the Wimshurst in 
the ordinary way; two well-coiled insulated wires, which are 
supported on insulating posts, terminate in brass knobs at the 
ends next the prime conductors of the Wimshurst, and in care- 
fully-made small smooth loops of bare wire at the four ends. 
There are thus four spark gaps (and it is impossible to over- 
tate the importance of the adjustment of these spark gaps in 
using the Wimshurst for this, or any other experiments where 
fluorescent screen or photographic effects are require’/—why, 
