224 Rey. P. J. Kirkby on the Passage of Electricaty 
restored by a wire network which fitted tightly into the 
circular hole. Upon the plate ZN rested a glass vessel G 
containing pentoxide of phosphorus, the purpose of which 
was to absorb any water-vapour which might be produced. 
The glass vessel G was kept in its place by a rectangular 
strip of brass which was screwed to the zinc plate at ZN; a 
brass rod CD was in metallic connexion with the strip of 
brass, and passing through the ebonite plug P in the neck of 
the bell-jar could connect the plate ZN to the electrometer. 
Ebonite pillars coated with sulphur supported ZN and in- 
sulated it from AB. Ordinary precautions were taken to 
avoid the influence upon the electrometer of charges which 
might creep up the glass from the plate AB. The brass 
tube T, through which the gas was passed into the bell-jar, 
was fastened into the ebonite plug P. The cylindrical part 
of the bell-jar was about 12 cm. in height and 6°5 in dia- 
meter. The diameter of the plate ZN was about 5 cm. 
All the joints were sealed by means of elastic glue, which 
proves remarkably air-tight. 
The plate AB was supported on ebonite pillars which kept 
it insulated. Ultra-violet light was produced between alu- 
minium points immediately below the quartz plate Q by 
means of the discharge of a leyden-jar whose electrodes were 
connected to the terminals of the secondary circuit of a 
Ruhmkorff coil. 
Hydrogen and oxygen were prepared by the electrolysis 
of a solution of caustic potash in the equivalent proportions 
of two volumes to one, and after passing through drying- 
vessels were introduced into the bell-j -jar by means of the 
tube T. The pressure of the mixed gases was sufliciently 
reduced to bring it within the range of the McLeod gauge, 
the limit being about 7 mm, of mercury. 
In order to render small falls of pressure easily measurable, 
a stopcock 8 had been introduced into the apparatus so as to 
cut off the bell-jar, the McLeod gauge, and the ordinary 
manometer from the rest of the apparatus. By this means 
the volume within which a fall of pressure took place was 
greatly diminished, and the accuracy with which such a fall 
could be measured was equally increased. 
Fig. 2 is a diagram of this arrangement. The volume of 
the apparatus to the right of 8, and cut off by the stopcock 8, 
within which the falls of pressure took place, was eventually 
measured in the following way. The apparatus was first ex- 
hausted to a low pressure, and, the stopcock 8 having been 
closed, the glass tube R was cut and a small glass bulb fused 
on to it.. This bulb could be filled with air at atmospheric 
