﻿Phenomena in Photometry . 129 



Nevertheless, in order to give an idea of the order of 

 illumination used, the red lamp was compared against a 2 C.P. 

 Methven gas standard, usingtheLummer-Brodhun photometer, 

 and its candle-power worked out to something of the order of 

 0'25 candle-power. 



The relative illuminations used in the curve are expressed 

 in terms of this figure. 



It will be seen it is only when the illumination has been 

 reduced to about 0*2 candle-metres, that the ratio of the red 

 to the green light begins to decrease. 



The same experiment was tried on several occasions with 

 lights of different candle-power, and it was always found that 

 fairly consistent results were obtained until the illumination 

 fell to something of the order specified above, when the 

 accentuation of the green light began to appear. 



At such illuminations it is very difficult to obtain readings 

 at all, and the mean of a great many should be taken. The 

 eye seems to be in a state of uncertainty, and one receives 

 the impression that the lights are continually varying. In 

 fact the eye seems to be in a state analogous with that of 

 unsaturated iron on the steep part of the magnetization curve. 



It may safely be assumed, therefore, that in all ordinary 

 cases, where the illumination would be at least ten or twelve 

 times as great as that employed above, where the field of view 

 would subtend a comparatively small angle at the eye, and 

 where we should never meet with such an extreme colour 

 contrast as in this case, the Purkinje phenomena will not 

 materially influence the results. 



(4) In order to obtain the curve of vertical distribution of 

 light from arc lamps, &c, the beam of light is often reflected 

 in the desired direction by means of a 45 degrees mirror*. 

 When this is done, the question arises, whether the coefficient 

 of reflexion of the mirror will be the same for lights of 

 different colour. 



The following experiment was carried out with a silvered 

 glass mirror of this type. Two 200-volt 32 C.P. lamps were 

 run in parallel off a constant P.D. of 200 volts. The two 

 lamps were compared against each other direct. They were 

 then compared when the light from one of them was reflected 

 along the bench from the mirror. In this way, the coefficient 

 of reflexion of the mirror is easily obtained. 



This was repeated when the lamps were screened with red 

 and green glass, successively. 



* A mirror, that is, which rotates about an axis making an augle of 

 45° with its plane. 



Phil. Mag. S. 6. Vol. 12. No. 68. Aug. 1906. Iv 



