April 6, 1893] 



NA TURE 



539 



ciples. A box, similar to Fig. 2, was prepared, but having 

 two apertures, one at each side. Through one the 

 coloured ray was reflected, and through the other a white 



out for the ordinatcs'; each curve is therefore made 

 on a scale ten times Aat of its neighbour, counting from 

 the centre. 



In th« diagram the sodium light of 

 the spectrum before extinction was 

 made of the luminosity of the amyl- 

 acetate lamp (hereafter called al), which 

 is about -8 of a standard candle, at i 

 foot distance from the source. Before 

 it ceased to cause an impression on the 

 eye, the illumination had to be reduced 



10,000,000 

 E light 10 - 

 F light 

 G light 

 C light 

 B light 



65 



10,000,000 



ISO 



10,000,000 



3000 



15 



,000,000 



or 3 



10,000,000 10,000 



' ^-!^££_ or _LL^ 

 10,000,000 10,000 



70,000 7 



' ' or ' 



10,000,000 looo 



Fig. I.— Extinction of Srectrum Colours. 



beam of light to a white screen. Both beams were 



diminished, and when the white and coloured patches 



appeared the same hue, the amount of illumination was 



calculated. Fig. i shows graphically the reduction of 



illumination, when the D light of the spectrum is the 



same intensity as one amyl-acetate lamp at one foot 



from the screen. To measure the extinction of light, a 



box was made as in the diagram, 



closed at each end, but having 



two apertures as shown, Fig. 2 ; 



— E is a tube through which the 



eye looks at S, which is a black 



screen with a white spot upon it, 



and which can be illuminated by 



light comingthrough the diaphragm 



D first falling on a ground glass 



which closes the aperture, and 



reflected on to it by M a mirror. 

 The patch of light of any colour 



being thrown on D,rotating sectors, 



the apertures of which could be 



opened and closed at pleasure, 

 were placed in the path of the 

 beam, thus enabling the intensity 

 of the patch to be diminished. 

 I) could be made of any desired 

 aperture, and thus the illumination 

 of the ground glass would be 

 diminished at pleasure. After 

 keeping the eye in darkness lor 

 some time, the eye was placed 

 at E, when the white spot illumi- 

 nated by the colour thrown on D 

 was visible, and the sectors closed 

 till the last scintilla of light was 

 extinguished. This was repeated 

 for rays at different parts of the 



spectrum, and the results are shown in Fig, 3 by the con- 

 tinuous curved lines. The diagram would have been 

 too large had the same scale been adopted through- 



NO. 1223, VOL. 47] 



'^ There was one objection which 

 might have been offered to this method, 

 and that was to the use of the rotating 

 sectors, and perhaps to the ground 

 glass. This objection was met by 

 first of all reducing the light by 

 means of a double reflection of the beam foiming the 

 patch from one or two plain glass mirrors, and also by 

 using a plain glass mirror in the box instead of a silvered 

 glass. By this i)lan the light falling on the first plain 

 glass mirror was reduced, before it reached the end of 

 the box, 1000 times ; and again, by narrowing the slit of 

 the collimator, and'also the slit placed in the spectrum, 



I another similar reduction would be effected. .All rays 

 thus enfeebled were within the range of extinction. It 

 was found that neither ground glass nor rotating sectors 



