PHENOMENA RELATING TO THE SPECTRA OF HYDROGEN AND HELIUM. 261 
Hydrogen and Helium under similar conditions is possible. We are thus led directly 
to a problem which has many other important implications, some of which are stated 
below. 
In the earlier paper* we suggested that the neutral-tinted wedge, as an accessory 
to the spectroscope, virtually increased the resolving power, and that in combination 
with the interferometer it should be able to solve problems which are beyond the 
power of pure spectroscopic analysis in its ordinary forms. Some experiments have 
been performed in connection with what is perhaps the most urgent of these problems 
at the present time—the elucidation of the nature of the Balmer series of Hydrogen. 
It has been generally accepted for some time that the series is one of doublets—we 
are not, at the present juncture, considering some very recent theoretical work by 
Sommerfeld— but the exact .nature of the series has always been an open question. 
For if it corresponds to a subordinate series (Diffuse or Sharp) in other elements, 
the separations in wave-number of the two components should be identical along the 
series. If, on the other hand, it corresponds to a Principal series, the separations 
should decrease rapidly towards the violet, after a manner which can be calculated 
from the separation in PI a , the first member of the series. There is at present no precise 
knowledge of the nature of these separations. MichelsonI found H a to be a doublet 
of separation O' 14, the intensities of the components being in the ratio 10:7, and the 
stronger one being on the red side. On the other hand, in H^ he found a separation 
only half as great, and the lines were so broad that no great degree of accuracy 
could be attached to this value. Others who have investigated the separation of 
H a have obtained very discordant values. Thus Houstoun^ gives 0'065, Fabry and 
Buisson§ 0'132, and Paschen and Back|| 0'20, so that there is no certain knowledge 
even as regards the separation of H a . Fabry and Buisson have also stated that 
H^j is double, with a separation in accordance with the theoretical value. This 
statement is somewhat vague, especially as no numerical estimate was given, and the 
theoretical separation depends very much on the view we take as to the character of 
the series. For example, with Michelson’s value for H a , §X = 0'14, the separation in 
H s as a Diffuse or Sharp member would be 
SX' = 0'14 (486l) 2 /(6563) 2 = 0'075 A.U., 
but as a member of a Principal series would only be 
d\" = 3W/4 3 = 0'031 A.U. 
An experimental arrangement, therefore, of more sensitiveness than any used 
* Loc. tit. 
t ‘Phil. Mag.,’ (5), vol. 34, p. 282, 1892. 
I ‘ Phil. Mag.,’ (6), vol. 7, p. 456, 1904. 
§ ‘ Journ. cle Phys.,’ June, 1912. 
|| ‘ Ann. der Phys.,’ 39, p. 897, 1912. 
2 o 
VOL. COXVII.—A. 
