510 Capt. \V. de W. Abney. On the Limit of [May 14 



brightnesses could be equalised, and the intensities of the light 

 transmitted passing through the paper estimated. 



Now there is a ray very near D in the spectrum, whose colour i 

 very closely, if not quite, identical with the colour of the light emitted 

 by the burning amyl acetate, and for making the measures this ray waa 

 used. When the measure had been made, the screen, with the square 

 aperture, was placed in the position of the ground glass, and the- 

 amyl acetate lamp placed on the side of the screen, away from the- 

 colour patch, and the rod placed in position to cast the shadow 

 necessary. The rotating sectors were then placed between t' 

 spectrum and the screen, and the light reduced so that the illnmi 

 tion of the translucent and opaque white square, viewed from the side 

 of the lamp, was equalised. Knowing the distance of the lamp 

 the two cases, and the aperture of the sectors, the relative illumina- 

 tion of the two surfaces was asceiiained. For convenience, the- 

 aperture of the ground glass was limited by means of a diaphragm,, 

 or by placing a diaphragm in front of the first prism. 



Two sets of measures showed that if the illumination of the ground 

 glass be represented by 1, the illumination of the card at the end of 

 the tube was -^ ; that is, any light falling on the ground glass wa* 

 diminished to that extent. 



The actual measures were -^fa and yj-y , but we may take 7-^ as- 

 sufficiently close to the truth. 



The colour-patch apparatus to which reference is made is described 

 in the Bakerian Lecture, 1886 (Abney and Testing, " Colour Photo- 

 metry"). The only addition to it that was made was to use an 

 adjustable slit to move through the spectrum. There was thus a 

 treble means of altering the intensity of the light, viz., by altering- 

 the aperture of the slit of the collimator, by altering that of the slit of 

 the slide, which was shifted at will into different parts of the spectrum,, 

 and by the rotating sectors placed in front of the spectrum. The- 

 mode of proceeding to measure the luminosity at which light 

 disappeared was as follows : The dullest part of that portion 

 of the spectrum which it was desired to extinguish was allowed 

 to pass through the slit in the spectrum, and a patch was form 

 on the ground glass, which, it may be remarked, had a tn 

 fitted over it, to prevent any chance of extraneous light reaching it. 

 The card at the end of the square box was viewed, and the slits closed 

 till all trace of light disappeared. (It may be as well to call to mind 

 what is well known, that faint light of all colours appears as white.) 

 In some sets of experiments the sectors were set at fixed angles, and 

 rotated in front of the patch, and the slit in the spectrum moved from 

 a position in which faint light appeared to one in which it just dis- 

 appeared, the position in the spectrum being noted by the scale at the 

 back of the moving slide carrying the slit.* In other cases the s 



