t then be heated with constant rotation in the flame. If this final
heating of the end _e_ be done skilfully, the glass will probably
collapse and flatten, as at _F_. The end must then be gently blown into
the form shown at _G_.

If a flat end to the tube be desired, the tube may be left in the
condition shown by _F_, or a thin rounded end may be flattened by
pressure on a plate of iron.

If a concave end be wished for, it is only necessary to gently suck air
from the tube before the flattened end has become solid.

In each case, _immediately_ after the tube is completed, it must be
closely wrapped in cotton wool and left to cool. With good lead glass
this last process, though advantageous, is not absolutely necessary; and
as glass cools slowly when enveloped in cotton wool, this precaution may
frequently be neglected in the case of apparatus made from lead glass.

[Illustration: FIG. 9.]

In order to draw out tubes for sealing, close to one end, and thus to
avoid waste of material, it is a good plan to heat simultaneously the
end of the glass tube _A_ which is to be sealed, and one end of a piece
of waste tube _E_ of about the same diameter, and when they are fused to
bring them together as at _DD_ (Fig. 9). _E_ will then serve as a handle
in the subsequent operations on _A_. Such a rough joint as that at _D_
must not be allowed to cool too much during the work in hand, or _E_ and
_A_ may separate at an inconvenient moment. Or the glass at the end of
the tube may be pressed together to close the tube, and the mass of
glass may be seized with a pair of tongs and drawn away.

=Choking, or Contracting the Bore of a Glass Tube.=--If it be not
desired to maintain the uniformity of external dimensions of the tube
whilst decreasing the diameter of the bore, the tube may be heated and
drawn out as described in the description of sealing tubes on pp. 32-35.
This may be done as shown at _A_ or _B_ in Fig. 8, according to the use
to which the contracted tube is to be put.

[Illustration: FIG. 10.]

Greater strength and elegance will be secured by preserving the external
diameter of the tube unchanged throughout, as shown in Fig. 10. For this
purpose heat the tube with the pointed flame, if it be small, or in the
brush flame if it be of large size, constantly rotating it till the
glass softens and the sides show an inclination to fall together, when
this occurs, push the two ends gently towards _A_. If the tube should
b