Vol. 7, 1921 
PHYSICS: R. A. MILLIKAN 
291 
of having identical spectra in all particulars between 1000 A and 235 A, 
while showing altogether distinct characteristic spectra between 1700 A 
and 6000 A, a behavior which could only mean that they themselves had 
no lines in the region, between 235 A and 1000 A, all of the observed lines 
being those of the common impurity oxygen — an interpretation supported 
by the fact that the well known oxygen lines of the visible region were 
also found in both spectra. It was with the aid of this discovery and the 
careful comparison of the spectra of otherwise pure metals, or of metals 
containing known impurities that the origins of practically all of the hun- 
dreds, even thousands, of new lines have been established with but very 
little uncertainty. 
b. The spectrum due to the 4 L-ring electrons of the carbon atom be- 
gins upon our plates at 360.5 A and extends with much complexity and 
strength up to 1335.0 A where its strongest line, which, in harmony with 
the foregoing convention will be called its LJ line is found. Above 1335.0 
A, the carbon lines are widely scattered and relatively weak. Since the 
K line of carbon is accurately computed to be at 44.4 A, the ratio of the 
K to the L frequency in carbon is about 30. 
It is interesting that practically the whole group of lines which Lyman 
obtained from his condensed discharge in helium 7 are due to carbon, oxygen, 
nitrogen and hydrogen. Thus those whose wave-lengths he gives as 
599.0, 643.7, 702.9, 718.2, 796.8 and 834.1 are all due to oxygen; those 
whose wave-lengths he gives as 904.6, 977.2, 1010.6, 1037.0, 1175.9 and 
1247.9 are due to carbon ; those whose wave-lengths he gives as 916.7, 991.1 
and 1085.5 are due to nitrogen ; while those whose wave-lengths he gives as 
972.7, 1026.0, 1216.0, are due to hydrogen. 
c. The spectrum due to the 5 L-ring electrons of the nitrogen atom 
(atomic number 7) was obtained by using Al electrodes with ammonium 
nitrate in their cores and observing the new lines which were not due 
to aluminium, oxygen or hydrogen. Before this experiment was tried, 
it was predicted that the L a line of nitrogen would have to lie between 
the L a line of oxygen and that of carbon. It was found that the spectrum 
due to the nitrogen atom was very simple, beginning on our plates on 
the short wave-length side at 685.6 A and reaching a maximum in the 
line of wave-length io8j.j A, precisely as predicted. This 1085.3 line is 
taken, in accordance with the foregoing convention, as the L a line of the 
nitrogen atom. The only other strong lines which we have obtained 
which are due to this atom have the following wave-lengths: 685.6, 916.2, 
991.1. The pair of nitrogen lines found by Lyman at 1492.8 and 1494.8 
appear upon our plates, but they are very faint in comparison with the 
foregoing lines, the nitrogen spectrum thus showing a behavior quite like 
that due to the atoms of oxygen and carbon. Since the K a line of nitro- 
gen is at 31.2 A, the ration of the K a to the L a frequency is 34.8. 
