sfactory, it may be
shrunk again and reblown. Such shrinking should begin in a large yellow
flame, with just enough air to give it direction. The amount of air may
be gradually increased as the bulb shrinks and the walls become thick
enough to bear it without collapsing. If the bulb starts to collapse at
any time, it must be immediately blown enough to regain its convex
surface, before the shrinking proceeds further.

=Discussion.=--In collecting the glass for the bulb, enough must be
gathered to give the walls the desired strength. Since the area of a
sphere is proportional to the cube of its diameter, it is evident that
doubling the size of a bulb diminishes the thickness of its walls to a
very large extent. The limit of diameter for a strong bulb on ordinary
1/4-inch tubing, collecting the glass as above, is about 1-1/2 inches,
and the beginner will do well not to blow his bulbs more than an inch in
diameter.

The collection of the glass is one of the most important parts of the
process. If the mass of glass be twisted, furrowed or ridged, or
lop-sided, it is very difficult to get a good, even, spherical bulb, no
matter how many times it is shrunk and blown. The greatest care should
therefore be taken to get a uniform cylinder, on the same axis as the
main tube; and to this end the rotation of the tube must be carried on
very evenly. For method of holding the tube, see page 14.

If a very large bulb is required, it will often be economical to seal on
the end of the tube a short piece of a large tube, provided with the
proper tail, and use the glass in the large tube for the bulb instead of
attempting to collect it from the small tube. In this case part of the
small tube will usually be included in the bulb, so that the joint comes
in the latter, and not where it joins the tube. As the amount of glass
carried on the end of the tube increases in weight and size the
difficulties of heating it uniformly, keeping it in the proper position
and handling it increase rapidly.

In collecting glass, it is usually best not to leave the part of the
cylinder next the tube with too thick walls. This is always the coolest
part during the preparation for blowing the bulb, consequently it does
not get blown out, and causes an ugly thickened appearance on that end
of the bulb.

If the bulb grows too long or pear-shaped, it may be easily shortened by
heating to the blowing temperature, and then blowing gently with the
main tube in