d be from 3/32 to
1/8 inch in diameter, and the tail drawn on the tube should have a
slightly less diameter at the point (about 1/8 inch or less from the
tube) where it is cut off. There are now two ways of sealing in the
wire. (1) The wire with the globule of glass is placed inside the tube
and the latter revolved until the end of the wire sticks out of the cut
tail (_a_, Fig. 18). The latter is now gently heated, and the two glass
surfaces fused together, taking care to use only the end of the hissing
flame, if the special glass contains lead. (See Chapter I, page 1.) The
whole circumference of the tube is then heated and annealed carefully.
(2) The end of the wire which is to be outside the tube is attached to
the end of a thin scrap of glass, by heating the glass and thrusting
the wire into it a very little way. Using this piece of glass as a
handle, the wire is inserted in the cut tail (_b_, Fig. 18) and the
globule brought near to the end of the tail. (If the main tube is cold,
it must of course first be warmed.) With the end of the hissing flame,
as in the first method, the globule of glass is melted and the end of
the tail softened. The wire is now pushed into place, the handle removed
by heating the end and withdrawing it, and the tail reheated a little if
necessary to make it shrink back into line with the walls of the tube.
The whole circumference of the tube is heated at that point and annealed
as usual.

[Illustration: FIG. 18.]

The use of this special glass is not absolutely necessary if the
platinum wire is small (1/4 millimeter or less in diameter), and in fact
it is often better in such cases not to use it, unless the apparatus is
to be subjected to a very high vacuum. On small tubes, especially, it is
undesirable to use the special glass, as a lump of it will usually cause
the tube to crack on cooling. When such glass is not at hand or is not
to be used, the procedure is altered somewhat. The tail which is drawn
out is very fine, having only a sufficient diameter so that when it is
cut off the wire can be inserted in it. Such a fine tail is readily made
by heating a small spot on the tube, touching it with a warm platinum
wire, removing from the flame and drawing out the tail with the wire.
After cutting off the tail the wire is inserted in it, being held on a
scrap of glass as in the previous case, and the wire and tail heated
until the latter shrinks back into line with the walls of the tube. If
t