eparating them until a rod of the desired
size has been obtained.

If, on the other hand, it is desired to produce a tube from the mass of
heated glass, the mass should be blown hollow before the pipe-stems
supporting it are separated.

_Methods of Manufacture._--When the student has familiarised himself
with the more common operations and processes used in glass-blowing, he
will be in a position to increase his skill and knowledge of special
methods by a critical examination of various examples of commercial
work. There are few exercises more valuable than such an examination,
combined with an attempt to reconstruct the stages and the methods by
which the article chosen for examination was made.

Obviously, it is impossible to give full details of all constructions in
a small text-book; but it is easy to give an example of the
constructional methods employed in the making of almost any piece of
light blown-glass apparatus, and these methods should prove of special
value when apparatus of a new pattern has to be evolved for the purposes
of research. That is to say, one designs the apparatus required, applies
known methods of construction as far as possible, and, by the
examination of commercial apparatus having similar features, evolves the
new methods required. For an exercise in such a process of
reconstruction we may well take an ordinary commercial vacuum tube, such
as that shown by _a_, Fig. 17.

[Illustration: Fig. 17]

In the tube from which this drawing was made, it was found that the
spiral in the middle bulb was of a slightly yellowish colour and gave a
green fluorescence when the electric discharge was passed through the
tube; that is to say, the spiral is made of uranium-glass, which is
usually a soda-glass containing trace of uranium, and hence differing
slightly in composition from the ordinary glasses. The two enclosed
tubes which are bent into a series of S bends gave a pink fluorescence,
which indicates lead-glass; and the remainder of the tube fluoresced
with an apple-green colour; this suggests ordinary soda-glass. We have,
therefore, a piece of apparatus in which three dissimilar glasses are
joined, while, at the same time, that apparatus contains a number of
internal seals, and it is not probable that the dissimilar glasses will
have their coefficients of expansion so nearly alike as to permit of a
stable internal seal being made if one part of the seal consists of a
glass differing from that