om the surface of the main tube
(Fig. 28).

Fig. 28.

Fig. 29.

When this stage is reached, again heat the tube all round till it
nearly softens, and by means of the other hand heat the end of the
other tube which it is proposed to weld. Just before the main tube
actually softens, turn it so as to heat the edges of the aperture, and
at the same time get the end of the side tube very hot. Take both out
of the flame for an instant, and press the parts together, instantly
slightly withdrawing the side tube. If the operation is well
performed, it will be found that the point of maximum thickness of
glass is now clear of the main tube. The joint is then to be heated
all round and blown out--a rather awkward operation, and one
requiring some practice, but it can be done.

Fig. 30.

If great strength is wanted, heat the main tube all round the joint
bit by bit, and blow each section slightly outwards. If the operator
is confident in his skill, he should then heat the whole joint to the
softening point, blow it out slightly, and then adjust by pulling and
pushing. Cool first in the gas flame, and then plunge the joint into
the asbestos and cover it up--or if too large, throw the asbestos
cloth round it.

In the case of soda glass this final "general heat" is almost
essential, but it is not so with flint glass, and as the general heat
is the most difficult part of the job, it will be found easier to use
lead glass and omit the general heating. With soda glass a very small
irregularity will cause the joint to break when cold, but flint glass
is much more long-suffering. It is easy to perform the above
operation on small tubes. For large ones it will be found best to
employ flint glass and use the clip stands as in the case of direct
welds, treated above, but, of course, with suitable modifications.
Never let the main tube cool after the hole is made until the work is
done.

Sec. 34. Inserted Joints.

In many instances the performance of apparatus is much improved by
joints of this kind, even when their use is not absolutely essential.

There are two ways in which inserted joints may be made. The first
method is the easier, and works well with flint glass; but when one
comes to apply it to soda glass there is a danger of the glass
becoming too thick near the joint, and this often leads to a cracking
of the joint as the glass cools.

Fig. 31.

Suppose it is desired to insert the tube B into the tube