ould always be rotated and kept in a straight line
until the glass has set, so that the capillary may have the same axis as
the main tube. This capillary or "tail" is often a very necessary handle
in glass-blowing, and if it is not straight and true, will continually
make trouble.

In drawing out very large tubing, say from one to two inches in
diameter, it is often necessary to draw the tube _in the flame_,
proceeding very slowly and at a lower temperature than would be used
with small tubing. This is partly on account of the difficulty of
heating large tubing uniformly to a high temperature, and partly in
order to prevent making the conical part of the tube too thin for
subsequent operations.

=Constricting a Tube.=--Where a constriction is to be made in a tube,
the above method must be modified, as the strength of the tube must be
maintained, and the constricted portion is usually short. Small tubes
are often constricted without materially changing their outside
diameter, by a process of thickening the walls. The tube is heated
before the blast lamp, rotating it about its axis as later described,
and as it softens is gradually pushed together so as to thicken the
walls at the heated point, as in _a_, Fig. 1. When this operation has
proceeded far enough, the tube is removed from the flame, and the ends
cautiously and gently drawn apart, continuing the rotation of the tube
about its axis and taking care not to draw too rapidly at first. The
resulting tube should have a uniform exterior diameter, as shown in _b_,
Fig. 1.

[Illustration: FIG. 1.--Constricting a tube.]

This method of constriction is not suited to tubes much over 1/4 inch in
diameter, since the mass of glass in the constricted part becomes so
thick as to be difficult to handle when hot, and likely to crack on
cooling. Larger tubes are therefore constricted by heating in a narrow
flame, with constant rotation, and when soft, alternately gently pulling
the ends apart and pushing them together, each motion being so regulated
that the diameter of a short section of the tube is gradually reduced,
while the thickness of the wall of the reduced portion remains the same
as that of the rest of the tube, or increases only slightly. This
pulling and pushing of the glass takes place _in the flame_, while the
rotation is being continued regularly. The result may appear as
indicated in _c_, Fig. 1. The strength of the work depends upon the
thickness of the walls o