PHENOMENA RELATING TO THE SPECTRA OF HYDROGEN AND HELIUM. 
265 
and the corresponding wave length differences are 
<S\i = 0-013 A.U., SX 2 = 0'039 A.U. 
The diffuse character of the components of H„ would render it practically impossible 
to detect so small a quantity as by any method, and since SX 2 gives a weak 
component falling inside the main components, which themselves nearly overlap, the 
satellite corresponding to SX 2 could not be expected either. 
The structure for 11^ is shown in fig. 3, and is more complex. The values of the 
separations indicated in the figure are :— 
Ai = 4V A h , A 2 = Af A h , A 3 = ^ A h , 
and all the satellites are even more'difficult to detect than those of H a . 
The most fundamental test of the theory, however, in a preliminary form, is to 
decide whether the main separation is again <L H . If it is smaller and of the magnitude 
required in a Principal series, the theory would apparently not hold. We confine 
ourselves for the present to a presentation of the results, and further discussion 
follows after the description and measurements of the plates. 
(XIV.) The Lummer Gehrcke Plate. 
The main theoretical features of this somewhat complex problem are now defined, 
and we may proceed to the experimental method. This consists merely in the use of 
a Lummer Gehrcke plate in addition to the former apparatus, the interference fringes 
being brought to a focus on the slit through the neutral wedge. The final pattern 
which is photographed—again with the use of the process screen—for any line such 
as Ha consists of a set of fringes of the various orders. These consist of similar 
curves, with a general parabolic appearance, and approximately equidistant. From the 
shape of the contour of any one of these curves, we can deduce mathematically its 
analysis into components, from which by knowledge also of the density of the wedge 
the separations and relative intensities on an absolute scale of the components in the 
original light can be found, 
The ordinary theory of the action of a Lummer Gehrcke plate on monochromatic 
light is, of course, familiar, but in its extension to a problem of this nature it becomes 
somewhat cumbrous. We have found it possible, however, to devise a more simple 
treatment of the theory of the plate which can be applied not only to the present case 
but apparently to all practical cases in which the use of the plate is necessary. 
We append, therefore, an account of this method, which is to some extent empirical, 
although it is not difficult to show that it is equally accurate. It is rendered the 
more necessary in that even the strict theory has apparently not been worked out 
completely for lines which are not monochromatic, but have a definite breadth as in 
the present experiments, where in fact the breadth is itself a, matter of investigation. 
