1916] on The Spectra of Hydrogen and Helium 651 



This old observation was kept in remembrance during occasional 

 experiments on helium extending over several years, until at last 

 the line 4686 was recovered very faintly in a tube which I had 

 prepared for a British Association lecture in the summer of 1912. 

 With this beginning, a systematic investigation resulted in due 

 course in showing how the line might be produced at will, and with 

 almost any desired intensity. Generally speaking, the whole secret 

 was to apply condensed discharges of sufficient strength. The denser 

 the gas, the greater the intensity of the discharge required to pro- 

 duce the line, and the brighter and broader the line obtained. If 

 we make the experiment, you will see that the treatment is rather 

 severe ; indeed, the tubes very frequently break down under the 

 strain. In Fig. 5 the spectrum given by a heavy condensed dis- 

 charge in a tube at about 5 mm. pressure is compared with that 

 given by an ordinary uncondensed discharge in the same tube, and it 

 will be seen that the line 4686 appears with great intensity. At 

 lower pressures the line is less bright, but more sharply defined and 

 better suited for accurate measurement ; the resulting wave-length 

 was 468.5 "98, in close agreement with the corresponding line in the 

 chromosphere, stars, and nebula. It was also believed to be in 

 sufficient accordance with the wave-length 4687*88, calculated by 

 Rydberg for the supposed first Principal line of hydrogen. 



Identifications resting on a single line are justly looked upon 

 with some suspicion, but in this case confirmation was apparently 

 given by the presence of other lines calculated by Rydberg at 

 2733, 2385, and 2253. These lines were not in absolute agreement 

 with Rydberg's calculations, but very nearly so, and within the 

 possible range of error of the formulae. 



Having now obtained the long-sought lines, considerable em- 

 barrassment was caused by the fact that more lines were obtained than 

 were expected. Besides the Rydberg lines, there was an inter- 

 mediate set of lines at 3203, 25H, and 2306. However, as these 

 lines could be put in a series having the same limit as the 4686 series, 

 and as hydrogen might be supposed to have still another peculiarity, 

 they were provisionally regarded as forming a second Principal series 

 of hydrogen. 



The experiments were less successful as regards the reproduction 

 of the Pickering lines, but three of them were obtained with suf- 

 ficient brightness for approximate measurement ; their wave-lengths 

 were 5411, 4541, and 4200. The difficulty with regard to these lines 

 is, I believe, that the conditions which would be expected to brighten 

 them cause them to become so diffuse as to be unrecognizable, for this is 

 certainly what happens in the case of the ultra-violet members of the 

 4686 series. The stars manage these experiments rather better than 

 we do. 



Since hydrogen was always present as an impurity in the helium 

 tubes employed, the new lines were at first attributed to hydrogen, in 



