DISCHABGE FROM METALLIC SURFACES IN DIFFERENT GASES. 
443 
the light only traversed a small thickness of the gas’ in the vessel before reaching 
the electrode. 
The second apparatus, used principally for testing the ultra-violet light, is shown 
diagrammatically in fig. 3. It was of much smaller volume than the other one, and in 
this apparatus the electrode, Z, was always of zinc. The a|)paratus v/as oidy used 
at low pressures and it was seldom necessary to 
repolish the zinc, it being found that at pressures 
below, say, a tenth of an atmosphere, zinc showed 
little or no signs of fatigue when illuminated by 
ultra-violet light. Further description of this 
apparatus is unnecessary, the lettering of fig. 2 
applying also to this diagram. 
Each leak apparatus (as they may be termed for 
brevity) was connected to a Topler pump and to a 
barometer column. The pressures were read by 
noting the difiei'ences in height between these and 
true barometer columns placed beside them. The 
main apparatus was also connected to a McLeod 
gauge wdiich enabled pressures down to O’OOl milliiii. 
to be accurately read. Suitable drying bulbs contain¬ 
ing phosphorous pentoxide were included between 
the pump, McLeod gauge, and leak apparatus. A 
side tube closed by a good stop-cock was attached to the tu1)e connecting the pump 
and leak apparatus, hy means of which any desired gas could be admitted into the 
apparatus or tlie wliole exhausted down to any pressure above about 20 millims. by 
connecting to a water jiump. 
The apparatus was so arranged that the spark gap was in the focus of a quartz 
convex lens of about 30 centims. focal length, and by turning the lens slightly a 
parallel beam of rays could be thrown on to the sensitive electrode of either leak 
apparatus. Each of these electrodes was connected to the negative terminal of a 
battery of small secondary cells, the number of which could be varied up to 180. 
The other electrodes, the gauzes, were connected through a three-way key to one 
pair of quadrants of an electrometer and to earth, except when a key, which we 
will term the electrometer key, was raised. The other pair of quadrants of the 
electrometer, the positive terminal of the battery, and the metal case of each leak 
apparatus, were all permanently earthed. All wires leading to the electrometer 
were run through earthed brass tubes, and when additional capacity was introduced 
into the electrometer circuit, a parallel plate condenser enclosed in a metal-lined box 
was used, the casing of the box being also connected to earth. 
The electrometer was of the ordinary type, and was used with electric lamp 
filament and scale. The needle was usually charged to a jjotential of 400 volts, 
3 L 2 
