ius of curvature is anything more than one inch and
the disc not more than one inch in diameter the cathode can be struck
at once from the flat as described. For very deep curves no doubt it
will be better to make an intermediate pair of dies and to re-anneal
the aluminium after the first striking.

When the tube is successfully prepared so far as the glassblowing goes
it may be rinsed with strong pure alcohol both inside and out, and
dried. The straight part of the side tube is then constricted ready
for fusing off and the whole affair is placed on the vacuum pump.

In spite of the great improvements made during recent years in the
construction of so-called Geissler vacuum pumps--i.e. pumps in which
a Torricellian vacuum is continually reproduced--I am of opinion that
Sprengel pumps are, on the whole, more convenient for exhausting
Crooke's tubes. A full discussion of the subject of vacuum pumps will
be found in a work by Mr. G. S. Ram (The Incandescent Lamp and its
Manufacture), published by the Electrician Publishing Company, and it
is not my intention to deal with the matter here; the simplest kind of
Sprengel pump will be found quite adequate for our purpose, provided
that it is well made.

Fig. 42 is intended to represent a modification of a pump based on
the model manufactured by Hicks of Hatton Garden, and arranged to suit
the amateur glass-blower. The only point of importance is the
construction of the head of the fall tube, of which a separate and
enlarged diagram is given. The fall tubes may have an internal
diameter up to 2 mm. (two millimetres) and an effective length of 120
cm.

Free use is made of rubber tube connections in the part of the pump
exposed to the passage of mercury. The rubber employed should be
black and of the highest quality, having the walls strengthened by a
layer of canvas. If such tube cannot be easily obtained, a very good
substitute may be made by placing a bit of ordinary black tube inside
another and rather larger bit and binding the outer tube with tape or
ribbon. In any case the tubing which comes in contact with the
mercury should be boiled in strong caustic potash or soda solution for
at least ten minutes to get rid of free sulphur, which fouls the
mercury directly it comes in contact with it. The tubing is well
washed, rinsed with alcohol, and carefully dried.

Fig. 42.

The diagram represents what is practically a system of three Sprengel
pumps, though t