u.\ tiM.'M |{ I'llMl'o.MKl |;y 541) 



the Royal Society.* Two methods presented themselves of equalising the illumhmt ! >n 

 of the shadows first, by moving the slit across t IK.- spectrum whilst the sectors rotated 

 \\ ith a fixed aperture ; or, secondly, by placing the slit at known places in the spectrum 

 and equalising the illumination of the shadows by altering the aperture of the sectors. 

 In cases where the absorption of the spectrum by the colour increased rapidly, the 

 first method was most convenient, but where the absorption was very gradual, the 

 latter method was found to be most accurate, and was most usually adopted. 



* 



Fig. 14. 



This plan of producing the two spectra, at first sight, seemed everything that could 

 be wished, but a difficulty occurred which rendered a further modification advisable, 

 for it was found that the two spectra were not of proportionate intensity throughout. 

 This was discovered in following out the necessary order of experiment ; which was, 

 first, to compare the luminosity of the spectrum on the coloured paper with that on 

 the white card, and then to compare the values of the two spectra to one another 

 by throwing both shadows on white card. If the spectra were of proportionate 

 intensity throughout, it should only have been necessary to measure the relative 

 values of any one ray, and the same ratio ought to have been obtained for any other. 

 When trying this, however, it was found that if the two shadows were cast by rays 

 in the red end of the spectrum, there was decreasing value in one of the spectra 

 towards the violet. In fact, in the extreme violet the spectrum of one was only 

 about three-fourths as bright as that of the other. The cause of this difference 

 became apparent when examining the matter. The half lens which focussed one 

 spectrum received the rays which had passed through the thinnest part of the 

 prisms, whilst that focussing the other had passed through the thickest parts. The 

 difference in the ratio of the brightness at different parts of the spectrum was 

 traced to the different amounts of light absorbed by the different thickness of 

 glass traversed. Any slight shift in the position of the line of separation of the lens 

 altered the nitio of absorption, and, as in some cases such a shift could not well be 

 avoided, the method, though practicable, was scarcely practical. It became evident, 

 then, that some means must be adopted of forming each spectrum with the light 

 which had traversed the same thickness of glass. 



" Photometry of the Glow Lamp," ' Roy. Soc. Proc.,' vol. 43 



