JUNE 3, 1915] 
NATURE 
3/1 

entirely obviates the use of a ‘‘flux”’ glass; and in 
view of the present needs, it seems advisable to bring 
this method again before the notice of those who 
may find it to their advantage to use it. A general 
description of the process appeared in the Electrician 
of July 4, 1913; but the method does not seem to have 
become so popular as might have been expected from 
the simplicity and ease of its application. As to the 
results obtained, the following facts from my own 
experience may be mentioned as affording proof of its 
absolute perfection. 
In the course of some research worl: which is being 
carried on at present in Prof. Gray’s laboratory, 
involving the examination of certain resolved spectral 
lines, a Wehnelt electrolytic interrupter was used in 
connection with an 18-in. induction coil. This inter- 
rupter was employed on account of the fact that, with 
it, the discharge can be allowed to run for hours 
without attention. The positive electrode was simply 
a small length of platinum wire sealed through the 
end of a glass tube. The wire conveying the current 
dipped into mercury surrounding the upper end of 
the sealed wire. At first the sealing of the platinum 
wire into the glass tube was done by means of a 
“flux ’’ glass in the usual way. It was noticed, how- 
ever, that after a few hours’ working, as a result of 
the disruptive action of the interruptions, this glass 
round the anode became pitted out, a crater being 
formed with the wire projecting from its centre. 
This action may go on until a leak occurs. Again, 
for the large currents used—more than 10 amperes— 
a fairly heavy wire is necessary, and the heating 
effect frequently causes fracture of the glass, by small 
radial cracks starting from the wire, or even from 
the junction of the two glasses. Both these effects 
are, of course, undesirable. The pitting out of the 
glass exposes a greater amount of the conducting 
material, and consequently alters the frequency; while 
leakage due to fracture considerably increases the 
amount of ‘steady current,’ that is, current spent 
in mere ordinary electrolytic effects. 
To overcome both these defects, I used a piece of 
Jena glass tubing, with a small length of heavy 
platinum wire sealed directly into it by the Burnside 
method. This anode has been in use now for several 
hours daily for the past five weeks, and there is 
neither any fracture nor any sign of the pitting out 
of the glass as formerly. This hard glass is less 
liable than the soft sealing-in glass to be damaged 
by the disruptive action of the current, and the chilling 
process has the effect of toughening the skin of the 
glass as well as of perfecting the seal between the 
glass and the wire. ; 
In connection with the same research, a large 
number of small vacuum discharge tubes were re- 
guired. In most of these tubes, the electrical 
conducting wire has been sealed in by the Burnside 
process, and in every case the result has been entirely 
satisfactory. The spectral lines under observation are 
of rather low intensity, and, in consequence, exposures 
of four or five hours’ duration are required to give 
a good photographic presentment. Generally, each 
tube has been used for two photographs. Thus the 
discharge has been passed in one direction for four 
hours or more, continuously, and then in the opposite 
direction, for a similar continuous period of four or 
five hours. In many cases, during the process of 
exhaustion, the tubes have been strongly heated by 
means of a bunsen burner; during some of the expo- 
sures, when the vacuum has been high, the tube has 
become very hot; in some cases the discharge has 
fractured the glass near the discharge end of the 
kkathode; but in no single case has cracking occurred 
at the seal, nor has there been any sign of leakage. 
NO. 2379, VOL. 95] 

Considering the large number of tubes used, I think 
these facts afford striking evideace of the trustworthi- 
ness of the seal. 
The essential part of Mr. Burnside’s process of 
sealing consists of the repeated immersion of the 
leading-in wire and the glass surrounding it, in a 
bath of oil, fat, or wax. If a wire is to be led into 
a tube, the tube should be drawn down until the 
bore is just large enough to admit the wire. The 
conductor having been inserted, the glass around it is 
strongly heated in the blow-pipe flame until perfect 
cohesion has taken place between the metal and the 
glass. The wire should then be drawn out slightly, 
carrying the adhering glass with it, and this glass 
heated again. This may be repeated until a “neck” 
of glass about two or three millimetres long, and 
having the end well rounded, is formed around the 
wire. The seal is then withdrawn from the flame, 
and, when the red glow has entirely disappeared, the 
end of the tube carrying the wire is cooled by several 
immersions in the oil. This may best be done by 
bringing the bath containing the oil—a short, wide 
test-tube, say—up round the seal. Each immersion 
should last about two or three seconds. The depth 
of immersion is increased with each successive. dip, 
until the seal is completely cooled out. In the case 
of vacuum tubes where the electrode requires a small 
inner support tube, the conducting wire can first be 
fused into the support tube, leaving a fairly long 
thick neck of glass in contact with the wire. This 
support is then well fused into the bulb of the vacuum 
tube in the usual way, and, finally, the seal is cooled 
out as described above. Not only does this cooling 
process render perfect the seal between the metal and 
the glass, but it also improves the junction of the 
glass of the inner tube with that of the outer. The 
leading-in wire of one of the vacuum tubes, imme- 
diately after being cooled out in oil, has been suddenly 
heated in the blow-pipe flame until it was at a bright 
red heat up to within about 1 mm. from the glass, 
without the seal being affected. 
Compared with previous methods of sealing in 
electrodes, where a special ‘“‘flux’’ glass had to be 
used, where care had to be taken to prevent the 
ordinary glass of the tubes from coming directly into 
contact with the metal, and where, unless special 
precautions as to annealing were observed, fractures 
frequently ensued, the present method is much simpler 
and easier, and more effective. It also enables us to 
make discharge tubes of Jena or other hard combus- 
tion glass. This was formerly quite impossible on 
account of the fact that no ‘‘ flux’’ was known which 
was suitable for sealing wires through glass of this 
nature. 
Mr. Burnside recommends that sperm oil be used, 
and that it be heated slightly before the glass is 
immersed in it. For the above work, I used the 
ordinary machine oil from the workshop—simply 
because it was most readily obtainable—and used it 
at room temperature. In cases where the position of 
the seal prevents its being conveniently immersed in 
a bath of oil, the process may be effected by bringing 
vertically upwards a piece of fat or wax cut to a 
suitable size and shape, and pressing it against the 
hot seal. 
It may be mentioned also, that not only can 
platinum be sealed directly through Jena or other 
hard resistance glasses, and through fused quartz, 
but that perfect seals may be made between the more 
easily oxidisable metals, such as copper or iron, as 
well as platinum, and glasses having comparatively 
low fusion points, e.g. lead glass and German glass, 
or any glass of this nature. For large currents the 
conductor should be tubular, because it can be» more 

