wo seconds,
and the tube then drawn out as described, so as to change the bulb to a
tube. The drawing must proceed with care: portions nearest the unchanged
tubes are the first to reach the proper diameter, and must be given time
to just set at that point before the center of the bulb is finally drawn
into shape. The drawing is perhaps best done intermittently in a series
of quick pulls, each drawing the tube perhaps 1/16 inch, and each taking
place as the thumbs and first fingers grasp the tube for a new turn in
the rotation. If the tube is not rotated during the blowing, the bulbs
will be lop-sided and it will be impossible to get a joint of uniform
wall-thickness; if rotation is omitted during the drawing, the tube
will almost invariably be quite crooked.

If the lump still shows distinctly after the operations described, the
cross-section of the tube will be as in _h_, and the tube will be likely
to break if ever reheated at this point after it becomes cold. The
operations _d_, _e_, _f_, and _g_ may be repeated upon it, and it may be
possible to get it to come out all right.

Care must be taken not to blow the bulbs _d_ and _f_ too thin as they
then become very difficult to handle, and the joint is usually spoiled.
The wall-thickness of these bulbs must never be much less than that of
the original tube. If the joint as completed has thinner walls than the
rest of the tube, it will be more easily broken. It should be remembered
that the length of the finished tube must be exactly the same as that of
the original piece, if the walls of the joint are to be of their
original thickness. Therefore the pushing together during the two
operations _c_ and _d_ must shorten the tube just as much as the final
drawing (_f_ to _g_) lengthens it.

The interval between the removal of the work from the flame and the
beginning of the blowing must be made as short as possible, or else the
portions next the main parts of the tube will set before they can be
blown out, and cause irregular shrunken areas.

EXERCISE NO. 2

JOINING TWO TUBES END TO END--SECOND METHOD

The method described in Exercise No. 1 is very satisfactory for joining
short lengths of straight tubing, but becomes inconvenient or impossible
when the pieces are long or bent, on account of the difficulty in
uniformly rotating such work. In such cases, this second method is
used. It does not usually give as smooth and pretty a joint as the first
method, and takes a