c7O IMAGE RL 
i int entered still further into the convergent kathode beam, 
yne-half of the ring disappeared, as Fig. 9, while, when the 
»bstacle—which, it should be remembered, blocked only one- 
quarter of the circular area of the tube—was brought close up 
to the kathode, only about one-quarter of the ring remained, as 
in Fig.,10. 
Further ex 
both in the « 
exhausted to 
yeriments were tried with the aluminium obstacle 
livergent and convergent but with the tube 
different degrees of vacuum, when it was observed 
cones, 
that when the obstacle was in the divergent cone, a portion of the 
ring was cut off exactly proportional to the angle subtended by 
the sides of the obstacle ; while when the obstacle was placed 
in the convergent cone, a much larger proportion of the ring was 
cut off in each case, this being much more eds with a high 
vacuum, when the diameter of the ring was small, than with a 
low vacuum, when the diameter of the ring was large. 
The C Higher 
The carbon anti-kathode screen was found useless for investi- 
gating the conyersen cone of kathode rays at anything but a 
very low vacuum, by the reason of the well-known difficulty in 
mivergent Cone at Vacua. 
getting any discharge to pass when the distance between the 
electrodes is less than the thickness of the dark space; and for 
the further reason that if the anti-kathode screen was not con- 
nected to the anode, it became itself negatively charged, and 
acted as an additional kathode when brought into the space 
between the kathode and the focus. ; 
Under these circumstances, it was thought that possibly some 
additional information might be obtained with regard to the 
form of the convergent cone at high vacua, by making the con- 
cave kathode itself of carbon. A tube was therefore constructed 
having a concave carbon kathode, the diameter of which was 
I inch, and the radius of curvatureo-75 inch. The appearance 
of the kathode with this tube is shown for a fairly high vacuum 
in Fig. 11, in which the kathode itself is shown in section, so as 
to let the form of the discharge better seen. As will be 
»bserved under this condition of vacuum, which was too high to 
be 
show any divergent cone, the cone of convergent rays appears 
to be contracted in diameter at its base, and to come off from 
the central portion of the kathode only, the remaining surface 
f the kathode being apparently inactive. This was found to be 
ill more the case at higher vacua, as will be seen from Fig. 12, 
which shows in a similar manner the form of the kathode dis- 
NO. 1433, VOL. 55] 
| 
| 
| 
| 
[APRIL 15, 1897 
charge in a tube exhausted to a very high vacuum. In this 
case, as will be observed, the whole of the kathode rays appear 
to come off from a very small spot in the centre of the kathode. 
Further, that this small spot is, at any rate, the source of most, 
if not all, activity, was evident from the fact that it became 
luminescent exactly in the same manner, but in a less degree, 
arbon surface upon 
Whether this 
than had previously been observed with a c 
uthode 
which kz rayS were concentrated. sur- 
face luminescence of the kathode carbon, at the point where 
the kathode rays leave it, is due to the violent tearing away of 
particles of carbon, or to some other cause, it is difficult to say ; 
but the fact that at high vacua the kathode rays come of entirely 
Fics. 11, 12 
—or, at any rate, almost entirely—from only a very small por- 
tion of the centre of the kathode, explains the observed fact that, 
within limits, large kathodes have no advantage over small 
kathodes in X-ray tubes. 
During the carrying out of the above experiments with a 
carbon kathode, very bright sparks were occasionally seen 
coming off the kathode and passing through the focus, and it 
was consequently thought that possibly by placing two concave 
carbon kathodes facing one another, such particles, by being 
caused to rebound backwards and forwards continuously between 
the two, might render the form of kathode stream visible at very 
high vacua when the stream itself becomes otherwise invisible. 
Fics. 
13, I4. 
With this view, a tube was made with two concave carbon 
kathodes, similar to those employed in the last experiment, were 
placed exactly opposite one another. The anode was placed in 
an annex, and the two kathodes were connected together by 
means of a wire outside the tube. At a very high exhaustion, 
this tube gave very beautiful effects, and showed clez arly the form 
of the kathode discharge at a degree of exhaustion when it is 
usually in itself quite invisible. Immediately on the current 
being turned on and the discharge passing, a straight and thin 
streain of bright golden coloured particles of apparently imcan- 
descent carbon passed between small luminescent spots at the 
