1911-12.] Absorption of Light by Inorganic Salts. No. V. 43 
(1) Rigidity : the only moving part is the mirror. 
(2) Automatic focussing : we have only to move the plate carrier 
towards N until the sodium lines are in focus at a particular point on the 
ground glass screen, P say. Then, if N be rotated, whatever part of the 
spectrum falls on P is in focus there, and owing to the plate holder being 
cut at the proper angle the whole spectrum is in focus all over. 
(3) The quartz prism can be replaced by a glass prism or rock-salt 
prism without any structural alterations. In that case the mirrors must 
of course be set at a different angle, and the angle of the plate carrier 
altered somewhat if the whole spectrum is to be in focus at once. 
(4) The instrument can be used as monochromator, and can be em- 
ployed for spectral work with a radio-micrometer or similar stationary 
instrument. 
The apparatus was first used with the thermopile in the infra-red, a 
flint prism and silver mirrors being employed for this part of the work. 
The micrometer screw was calibrated, as described in the first article of the 
series, with emission lines in the visible spectrum and water bands in the 
infra-red. Stray heat was eliminated by means of screens before the 
thermopile. This method is not quite so thorough as the use of an 
auxiliary prism, but saves a little more energy and is above all faster in 
working. The other details were as formerly. 
After finishing the infra-red I attempted to use the thermopile in the 
ultra-violet with a spark between metal electrodes as source. But with the 
induction coils at my disposal the deflections were not large enough to 
make the method practicable. I therefore fell back on photography. 
The photometer already described by John S. Anderson * and myself 
suffers from the disadvantage that its range is limited. The illumination of 
each of the quartz plates cannot be altered in a greater ratio than that of 
one to four. In my efforts to remedy this defect I built an arrangement, the 
plan of which is shown in fig. 2. 
Let us suppose the rays to issue from the slit of the spectrograph S. 
They first fall upon a quartz rhomb R, which divides them into two pencils 
as shown by the elevation, fig. 3. The lower pencil is reflected by the 
totally reflecting quartz prism P, and meets the diffusely reflecting surface 
B. The upper pencil keeps straight on and meets the similar diffusely 
reflecting surface A. These surfaces are large enough to stop all the rays 
issuing from the instrument. An iron arc L moves along a scale between 
A and B, and the illumination of these surfaces varies inversely as the square 
of their distances from the arc. The light from A and B forms two spectra 
* Proc. Roy. Soc. Edin., xxxi. p. 550 (1911). 
