se
proceeding as follows:--

Join a piece of tube somewhat larger than _M_ to its end _A_, draw out
the other end of the larger tube, and blow a bulb _L_ as directed on p.
47. Then seal the electrode _R_ into the bulb _L_ (p. 55).

Blow a similar but larger bulb _N_ from a large piece of tube sealed
between two tubes of similar size to _M_, as described at p. 50. Cut off
one of the tubes at _B_, and join the bulb _N_ to _M_ at _B_. Form the
bulb _Q_ in the same manner as in the case of _L_, seal into it the
electrode _R_, and add the tube marked by the dotted lines at _F_.

Seal a narrow tube _P_ to the end of a larger tube, and blow out the
tube at the joint till the glass is thin and regular. Take a tube _O_,
of similar size to _M_, slightly longer than _P_, contract its mouth
slightly to meet the wide end of _P_ at _D_, and after loosely
supporting _P_ inside _O_ with a cork, or otherwise, close the end _N_
of _O_ by sealing or corking it, and join _P_ to _O_ at _D_. Cut off _O_
just above _D_ at _E_, and join it to the bulb _Q_, closing either _O_
or _F_ for the purpose. Cut off the end of _O_ at _C_ parallel to the
end of _P_, and connect _O_ to _N_, using _F_ for blowing the joint at
_C_. _F_ may be used subsequently for introducing any gas into the tube,
and, when a vacuum has been established, may be sealed before the
blow-pipe.

[Illustration: FIG. 29.]

=Modes of combining the Parts of Heavy Apparatus.=--It is often
necessary to connect pieces of apparatus which are too heavy to be
freely handled before the blow-pipe, and which, therefore, cannot be
welded together as described on p. 39, by some more effective method
than the ordinary one of connecting by india-rubber tubing. For example,
apparatus which is to be exhausted by a Sprengel air-pump must be
attached to the pump by a joint as perfectly air-tight as can be
obtained. This, indeed, often may be done by welding the apparatus to be
exhausted to the air-pump before the blow-pipe. But such a method is
open to the obvious objection that it is very troublesome to connect and
disconnect the parts as often as may be necessary, and that there is
some risk of accidental breakages. Nevertheless it may be done on
occasion, especially if there be no objection to the use of the
flexible spiral tubes already alluded to. When the use of a spiral
connecting-tube is not admissible the difficulty is considerably
increased. For example, the author has lately requ