of the other part.

These considerations lead us to a closer examination of the joins where
the dissimilar glasses are introduced, and we find that in no case is
the internal seal made between dissimilar glasses, but that a soda-glass
extension is joined on to both the uranium-glass tube and the lead-glass
tubes at a point about half an inch before the internal seal commences.
Careful examination of these joins shows that the change from one glass
to another is not abrupt but gradual. Such a transitional joint may be
made by taking a length of soda-glass tubing, sealing the end and fusing
a minute bead of the other glass on to the sealed end, the end is then
expanded and another bead of the other glass added, this bead is
expanded and the operation is repeated, thus building up a tube, and,
finally, the tube of the other glass is joined on to the end of this.

We are now concerned with the question of the insertion of the
uranium-glass spiral into the bulb (see p. 38). Obviously the spiral is
too large to pass through the necks of the bulb, and it is difficult to
imagine that the spiral was obtained by the insertion of a length of
straight tubing which was bent after entering the bulb; therefore, the
only remaining method is that the spiral was made first and the
soda-glass extensions fastened on, and that the bulb was blown, cut in
halves and the spiral inserted, and the two halves were then rejoined.
That this was actually the case is confirmed by traces of a join which
are just visible round the middle of the bulb. The insertion of the
spiral and the making of the first internal seal are shown by _b_, and
_c_.

There is one detail in making the second join of the spiral to the bulb
which calls for attention, and the small branch, similar to an
exhaustion branch, at the side of the bulb provides a clue to this. If
an attempt were made to complete the second internal seal through a
closed bulb it would be impossible to obtain a good result, as the
air-pressure in the bulb would not be under control when once union was
effected, and further heating of the air in the bulb would cause
expansion and perforate the wall near the second internal seal; we
therefore make a small branch which can be left open and through which
such air-pressure as may be found necessary can be maintained.

The third join, by which the lead-glass tube is joined to the soda-glass
is made in stages similar to those in which the soda-glass and
u