548 CAPTAIN ABNEY AND MAJOR-GENERAL FEST1NU 



colour whose luminosity was to be measured, illuminating the shadow of the rod 

 thrown on the coloured paper by the spectrum colour, and that thrown on the white 

 card by the white light reflected from the surface of the first prism ( XXVI). This 

 did away with any reduction or calculation ; but still an objection remained, as, for 

 definite comparison, it was almost necessary that the same observer should always 

 make the measurement. 



XXVIII. Revised Method. 



One of us having to measure the colour of various water-colour pigments for a 

 Government enquiry on their fading, it became important to introduce some other plan 

 by which the same end could be attained. Various artifices were tried, but finally we 

 came to the conclusion that a spectrum photometer was necessary, and on these lines 

 the following various modifications of our original apparatus were devised by one of us 

 (Captain ABNEY) : The collimator, prisms, and camera were at first kept as in the 

 colour photometer ; but for the camera lens was substituted a lens divided into equal 

 segments, which could be centrally separated, as in a heliometer. The light coming 

 through the last prism fell as a square patch on this divided lens, and the two segments 

 were separated so that two spectra fell on the focussing screen, one above the other. 

 A slit in a card was then passed across this double spectrum, and any required ray 



Fig. 13. 



i 



-^sf 



was isolated. P is a right-angled prism attached by a rod to the top half of the slit 

 so as to reflect the ray from the top spectrum to one side, whilst the ray of the same 

 colour from the bottom spectrum traversed the slit unimpeded and fell on the lens L m , 

 forming a patch of monochromatic light on the screen. The ray which was reflected 

 by P was again reflected by a mirror M 2 , and fell on another lens L IV , by which a similar 

 patch of monochromatic light could be made to fall over the patch formed by L m . Each 

 of these monochromatic rays cast a shadow of a rod, placed in front of the receiving 

 screen, and the shadow cast by each spectrum was illuminated by light of the same 

 colour coming from the other. To measure the value of a coloured paper, the screen 

 was made half with a white card and half with the coloured paper, as in the figure. 

 The shadows were made to touch at the intersection of the card and coloured 

 paper. In front of the light which illuminated the shadow cast on the white card was 

 placed a motor rotating movable sectors, as described in our paper recently read before 



