und the tube, and not trust to its running straight when the tube is
pulled apart.

On the whole, the best method in this case is to employ a flame
pencil, which should be kept ready for use. This merely consists of a
bit of glass tube of about the same dimensions as an ordinary lead
pencil, drawn down to a very fine jet at one end. The jet must not be
very long or thin, or the glass will soon fuse up. A few trials will
enable the operator to get the proper proportions, which are such that
the tube has the general appearance of a pencil normally sharpened
(say with a cone of 60'). This tube is best made of hard glass.
Connect it to a gas supply by light flexible tubing, and turn down the
gas till the flame from the end of the jet is not more than one-tenth
of an inch long. Then apply the jet, beginning from the end of the
crack, and gradually draw it (the crack) round the tube. The
operation will be assisted if a rubber ring is slipped on the tube to
begin with, so that the eye has some guide as to whether the flame is
being drawn round properly or not. The ring must, of course, be far
enough away to escape the effect of the flame. The crack will be
found to follow the flame in the most docile manner, unless the tube
is thick or badly annealed. Some operators recommend a pencil of
glowing charcoal, but the flame is undoubtedly better.

Sec. 30. To cut very thick Tubes.

A large number of methods have been proposed, and nearly everybody has
his favourite. The following has always succeeded with me. First
mark on the tube, by means of a little dead black spirit paint,
exactly where the cut is to be. Then sharpen the glass knife and
scratch a quite deep cut all round: there is no difficulty in making
the cut one-twentieth of an inch deep. It will be proper to lubricate
the knife with kerosene after the first mark is made. [Footnote: The
edge of the knife may be advantageously saved by using an old file
moistened with kerosene for the purpose. I find kerosene is not
worse, but, if anything, better than the solution of camphor in
turpentine recommended by Mr. Shenstone.]

If the glass is about one-eighth of an inch thick, the scratch maybe
conveniently about one-twentieth of an inch deep, but if the glass is
anything like one-quarter of an inch thick, the scratch must be much
deeper, in fact, the glass may be half cut through. To make a very
deep scratch, a wheel armed with diamond dust, which will be d