502 
Mr Horton , The spectrum of the discharge 
luminous discharge would appear at a pressure of a few milli- 
metres. The spectrum showed the nitrogen bands brilliantly and 
also a faint mercury spectrum. On still further reducing the 
pressure the mercury lines increased in brightness and the nitrogen 
bands gradually faded away. Three new lines were then observed 
to show up in the orange part of the spectrum. These increased in 
brightness as the pressure was lowered and finally stood out sharply 
on a dark background. By raising the temperature of the cathode 
the current through the tube was increased, the orange lines 
became more brilliant and two new lines appeared in the red. 
The wave-lengths of these lines were measured by means of a 
direct wave-length measuring spectroscope and found to be : 
The current through the tube was now 30 milli-amperes and 
the mercury spectrum was brilliant*. No other bright lines were 
observed, but the hydrogen red line showed up faintly at times. 
The three orange lines were of about equal intensity. The red 
line 6715 was not quite so bright and 6912 was still fainter and was 
always the last to appear as the current was increased. These lines 
could not be found recorded in any table of wave-lengths of mercury 
lines, nor, so far as I could find, of lines in the spectrum of any 
element. At first it was thought that they might be due to 
calcium or barium from the cathode. There is a strong calcium 
line at 6122 in the solar spectrum but it is accompanied by 6103 
and 6162 which Kayser and Bunge regard as making up one of the 
calcium series of triplets. By throwing the spectrum of lithium 
into the field of view of the spectroscope it was shown that the red 
line measured as 6715 was not the lithium line 6708. 
It seems therefore that the lines are due to mercury, and are 
produced either by the special method of causing the luminosity 
used in these experiments, or they are due to some new substance 
formed from mercury during the passage of the discharge. Experi- 
ments were therefore made to see if these lines could be obtained 
in the absence of mercury. For this purpose a long vertical spiral 
of glass tubing was inserted between the mercury pump and the 
discharge tube. This spiral was immersed in liquid air and the 
apparatus pumped down. This would effectively prevent mercury 
vapour coming over from the pump to the spectrum tube. At a 
low gas pressure it was now very difficult to get the discharge to 
6121V orange 
* The mercury spectrum referred to consisted of the lines 5790, 5770, 5461, 
4916, and 4359. The orange line 6152 was only seen in the presence of helium gas. 
