486 Mr. J. Barnes on the Analysis of — 
circumstances of temperature, pressure, strength of current, 
size and shape of electrodes, diameter of vacuum-tube, &e.’’ 
After experimenting a few months with both the Michelson 
and the Fabry and Perot interferometer the author was fully 
convinced that the Fabry and Perot method possessed the 
advantage for the problems in view, since it shows directly 
the structure of a given radiation by the simple inspection of 
the system of fringes. Hach fringe is in fact a true spectrum 
of the source and the conditions are the same as those existing 
in the spectra obtained by the use of a grating having a small 
number of lines but where the spectra employed are of a 
very high order. During the progress of the experiments 
the method proposed by Lummer appeared. While I have 
not been able to use this method exactly, I used, before 1 
read his paper, one which is very similar to it. This method 
and results obtained will be described below. 
Method. 
The method involved in this production of interference- 
fringes will be first briefly considered as it will assist 
towards a clear conception of the results. 
Consider a ray of monochromatic light incident at an 
angle 0 upon two glass plates whose inside surfaces A and B 
(fig. 1) are slightly silvered and separated from one another 
Ries i 

a distance D. If the silvered surfaces are parallel we have 
on account of the multiple reflexions a number of transmitted 
rays coming from the same source, whose differences of path 
increase in arithmetical progression. The differences of path 
with respect to the first are 2D cos 6, 4D cos @..'. 2nD cos @. 
