562 CAPTAIN ABNEY AND MAJOR-GENERAL TESTING 



and of the coloured sector 275. In the third observation they are 235 and 247 

 respectively. (The dotted lines in fig. 18 show the luminosity curves of the different 

 colours for R. on the normal scale of wave-lengths.) 



XXXIX. Comparisons by Gas Light. 



Our observations were extended to gas light comparisons. Our first endeavour was 

 to obtain a comparison between the intensity of the crater of the positive pole of the 

 electric light and gas. In order to do this we used a collimator, as described in 

 No. 232 of the ' Proceedings of the Royal Society,' 1884. On one slit the light of 

 the positive pole was focussed by means of a lens, whose aperture was reduced to 

 about 1 mm. in diameter ; on the other was focussed the brightest part of the flame 

 of gas in an ARGAND burner. The spectra from the two sources appeared on the 

 focussing screen of a camera, one above the other, and just touching. A card with a 

 slit was passed through the spectra to isolate any part required. The two spectra 

 were viewed by a RAMSDEN eye-piece, and the intensity of the electric light reduced 

 by means of the movable rotating sectors we have already described till equality was 

 established. In the brightest parts of the spectrum the light was too intense to be 

 readily compared ; so, in order to diminish the brightness, a photographic plate, on 

 which an even grey tint had been produced by development, was interposed between 

 the eye and the eye-piece. The light would then be sufficiently reduced in 

 intensity to allow fairly accurate and concordant measures to be made. It may be 

 here remarked that a series of three tints was prepared, from a very light grey, which 

 cut off about l/4th of the light, to one which cut off 19/20ths, and according to the 

 brightness of the part of the spectra under measurement, so was the darkness of the 

 interposing glass increased. There is an intensity in each case which gives the greatest 

 facility of accurate measurement, and this we endeavoured to obtain. (For the sake 

 of convenience, in the following table the ordinates of the luminosity curve of the 

 light reflected from white card have been increased in such a proportion that the total 

 luminosity from the gas light is equal to the total luminosity found for the electric 

 light. The curves of the colours have been calculated on the assumption that, as 

 before, the white sector had an angular value of 133'7.) 



