386 
DR. T. R. MERTON AND MR, S. BARRATT 
the ends. Since the lines of the Balmer series are strongly enhanced relatively to the 
secondary spectrum by very small traces of impurities, it is evident that the phenomena 
can be explained by an accumulation of the oxygen or other impurities in the central 
portions of the discharge tube. When however the total amount of impurity is exces¬ 
sively small, the accumulation of the greater part of it in the centre of the capillary 
may not be sufficient to weaken the secondary spectrum appreciably, and the capillary 
thus appears of a uniform white colour throughout its entire length. There are a number 
of other observations which appear to be related to these effects, but in the absence of 
any theory we do not venture to discuss them in the present communication. 
(11) On Impurities in Vacuum Tubes. 
It may perhaps be considered remarkable that any profound influence on the spectrum 
of a gas can be exerted by impurities which are present in such small quantities that 
then- presence cannot be detected by any characteristic lines or bands in the spectrum. 
The difficulty of eliminating such impurities as those which give rise to the Angstrom 
carbon bands is of course familiar to all who have worked with vacuum tubes, but with 
the aid of charcoal cooled with liquid air there should be no difficulty in preparing 
tubes containing Hydrogen or Helium which would show no lines or bands other 
than those peculiar to these gases. This is indeed the case under the conditions usually 
obtaining, when the gases are contained in the tubes at pressures of a few millimetres ; 
but it has been found that the difficulties are very much greater when it is desired to 
obtain vacuum tubes containing Helium at higher pressures up to 60 mm., which 
show no trace of impurities. The relative intensities of lines and bands due to 
impurities are enormously enhanced as the pressure increases, and the form in which 
the impurities appear is also often unusual. The influence of Helium on the secondary 
spectrum of Hydrogen is by no means unique, and a remarkable instance has been 
observed in the case of tubes containing Helium at high pressures and a very small trace 
of some carbon compound. If any considerable quantity of carbon is present the 
Angstrom and Swan bands can be seen, but with a very small trace of carbon the 
“ Cometbands first observed by Fowler (‘ Monthly Notices R.A.S.,’ vol. 70, p. 484, 
1910) appear quite brightly in the bulbs of the tubes just outside the capillary. This 
effect was first observed by one of us in an investigation in collaboration with Dr. T. 
Tak amine, to whom we are indebted for a photograph which shows the “ Comet ” 
bands almost free from other bands associated with carbon compounds. Fowler 
(loc. cit.) has found that these bands appear with the greatest relative intensity in tubes 
in which the pressure is so low (0-01 to 0*005 mm.) that the luminosity of the discharge 
is very small, and yet we find them here in tubes containing Helium at pressures from 
15 to 50 mm. This is only one example of the changes which may occur, and a number 
of other lines and bands have been observed in the case of other impurities. Further 
investigation of the phenomena is required, and we do not venture to discuss them in 
the present communication. 
