of sudden cold to any
point on the zone of hot glass will usually start a crack, which, if
necessary, may be continued in the usual manner. The last three methods
are chiefly useful in dealing with the largest and thickest tubes, and
with bottles.

A fairly stout copper wire, bent into the form of a bow so that it can
be applied when hot to a considerable surface of a glass tube, will be
found superior to the point of hot glass or metal usually employed, for
leading cracks in glass tubes. With such a wire a tube can be cut so
that the cross section of the end is at any desired angle to the axis of
the tube, with considerable precision. I am indebted for this suggestion
to Mr. Vernon Boys and Dr. Ebert.

=Bending Glass Tubes.=--The blow-pipe flame is not a suitable source of
heat for bending tubes, except in certain cases which will be mentioned
in a subsequent paragraph. For small tubes, and those of moderate size,
a fish-tail burner, such as is used for purposes of illumination, will
answer best. Use a flame from one to two inches in breadth--from _A_ to
_A_ (Fig. 6), according to the size of the tube which is to be bent. If
the length of tube that is heated be less than this, the bend will
probably buckle on its concave side.

[Illustration: FIG. 6.]

The tube to be heated should be held in the position shown in Fig. 6,
supported by the hands on each side. It should be constantly rotated in
the flame, that it may be equally heated on all sides. In the figure the
hands are represented above the tube, with their backs upwards. A tube
can be held equally well from below, the backs of the hands being then
directed downwards, and this, I think, is the more frequent habit. It is
difficult to say which position of the hands is to be preferred. I
lately observed how a tube was held by three skilful amateurs and by a
professional glass-blower. All the former held the tube with the hands
below it. The latter, however, held it from above, as in Fig. 6. He,
however, was working with a rather heavy piece of tube, and I am
inclined myself to recommend that position in such cases. During a long
spell of work, the wrist may be rested from time to time by changing the
position of the hands.

When the tube has softened, remove it from the flame, and gently bend
it to the desired angle. The side of the tube last exposed to the flame
will be slightly hotter, and therefore softer, than that which is
opposite to it. This hotter sid