DISCHARGE FROM METALLIC SURFACES IN DIFFERENT GASES. 441 
owing to the rapid volatilization of the alnmininm and consequent lengthening of the 
arc, thus causing an alteration in the intensity of the light emitted. 
The airangeinent described above, although proving’ satisfactory as a rule, required 
much attention at times, and a modified form has recently been adopted which, 
though not so simple in construction, affords a miicli more relialile source of ultra¬ 
violet light. The improvement consists in having tlie iron terminals in mi atmosphere 
of pure hydrogen, the spark apparatus being 
shown diagrammatically in fig. I. A brass tube 
B, 12 centims. in length and 3 centims. in 
diameter, was closed at one end by a quartz 
window Q. The iron terminals TT were 
admitted through the side tubes SS, and were 
surrounded by glass tubes GG, as shown in 
the figure. This was necessary in order to 
jirevent short-circuiting across the brass vessel, 
and long side tubes were necessary, as it was 
found that a metallic film was deposited on 
surfaces near the spark, which was also liable to 
cause short circuits unless some such precaution 
was taken. All joints were made airtight by 
sealing wax. A continuous slow stream of 
hydrogen was passed through the apparatus, 
entering at A and leaving at E. The hydrogen 
was prepared in a Kipp’s apparatus from pure 
zinc and hydrochloric acid, and was dried liy hubhling slowly througli concentrated 
sulphuric acid and passing tiirough long tubes packed with calcium cliloride. It was 
necessary to ensure that the hydrogen was jjerfectly free from oxygen, or water 
vapour was formed wiien tlie discharge passed, and this being opaque to ultra-violet 
rays would quickly rob tlie light passing out through the quartz window of these 
rays. The length of spark used was about 6 millims. 
The ultra-violet light obtained liy this method remained practically unaltered in 
intensity for hours or even days, the sjiark being usually run for 10-second periods 
every few minutes. 
The later series of experiments were made using this source of light, tliough the 
simpler arc in air is quite satisfactory unless very long series of ohsei'vations are 
needed or great accuracy is required. 
Apparatus and Method Employed. 
In order to be able to test the constancy of the ultra-violet light employed, two 
separate pieces of apparatus were used, in one of which the conditions under which 
VOL. CCII.—A. 3 L 
