370 
DU. T. E. MERTON AND ME, S. BARRATT 
agreed that both the Balmer series and the secondary spectrum are to he attributed to 
Hydrogen. The relative intensities of the two spectra vary in a surprising manner. 
Such traces of Hydrogen as are necessary to exhibit the earlier members of the Balmer 
series are indeed difficult to eliminate from luminous sources, but the secondary spectrum 
only appears in sources in which considerable quantities of Hydrogen are present, and 
its intensity relative to that of the Balmer series is greatly enhanced by the complete 
removal of all impurities. 
It has long been known that in the spectrum of water vapour the lines of the secondary 
spectrum are very weak, and many investigations have shown that the intensity of the 
secondary spectrum is greatly reduced by the presence of small traces of mercury 
vapour and. other impurities. The relative intensities of the two spectra are affected by 
variations in the electrical excitation, but for a high relative intensity of the secondary 
spectrum purity of the gas is essential. The appearance of the discharge in vacuum 
tubes containing Hydrogen of the highest degree of purity is indeed strikingly different 
from that observed in hydrogen tubes prepared without special precautions, the former 
being of an almost white colour whilst the latter show the familiar red glow, which is 
due to the predominance of the red line, Ila, of the Balmer series.* In a recent 
investigation, Wood (‘ Roy. Soc. Proc.,’ A, 97, p. 455, 1920 ; £ Phil. Mag.,’ 6, 251, 
p. 729, 1921) has described a number of interesting phenomena which he has observed 
in long vacuum tubes containing Hydrogen which were excited by a high potential 
transformer. Many of these observations cannot at present be explained fully, but 
the influence of traces of impurities is shown to be an important factor. 
It is well known that the secondary spectrum appears under less energetic conditions 
of excitation than the Balmer series, the latter alone being found in vacuum tubes 
excited by powerful condensed discharges, and important investigations by Fulcher 
( £ Astrophys. Journ.,’ 34, p. 388, 1911 ; 37, p. GO, 1913) have shown that when Hydrogen 
is excited by the impact of cathode rays the relative intensity of the secondary spectrum 
increases as the velocity of the cathode rays is reduced. Fulcher also found similar 
variations in intensity amongst the lines of the secondary spectrum itself, and identified 
a number of lines as characteristic of low potential discharges. It was found that these 
low potential lines exhibited regularities somewhat resembling those associated with band 
spectra. These regularities and their relation to other phenomena will be discussed in 
a later section. Any method by which the lines of so complex a spectrum can be 
separated into different physically related groups, cannot fail to yield results which 
will prove of assistance in theoretical investigations. 
It has been pointed out in a previous communication (Merton, £ Roy. Soc. Proc.,’ A, 
96, p. 382, 1920) that the relative intensities of the secondary lines are affected by the 
pressure in the discharge tube, the Fulcher bands being enhanced at low pressures, but 
* These remarks do not apply without amplification to the case of tubes excited by discharges of exceed¬ 
ingly low current density. In the presence of water vapour the intensity of the Balmer lines, relative to 
the secondary spectrum, increases very rapidly with the current density. 
