446 Intelligence and Miscellaneous Articles. 



of their tints. The employment of a spectrophotometer permits 

 this difficulty to be surmounted. 



Let us suppose the simple radiations which constitute the light 

 emanating from an electric regulator and that from a standard 

 Carcel lamp spread out in two contiguous spectra. If the distances 

 of the two sources from the photometer be such that their mean 

 illumination is the same, the two spectra will be far from present- 

 ing the same aspect : that of the carbon points is more intense 

 towards the violet extremity, and less luminous towards the red, 

 than that of the lamp. The temperature of emission of the electric 

 light being much higher than that of the lamp, the ratio of the 

 intensities of the simple radiations of the electric light to the in- 

 tensities of the corresponding lights of the spectrum of the flame 

 of the lamp, will with equality of mean illumination, be represented 

 by a fraction greater than unity towards the violet ; but if the red 

 be approached, the ratio will gradually diminish, following the law 

 of continuity, and will be less than unity at the red end. 



There exists, then, a determinate simple radiation (whose wave- 

 length depends on the nature of the lights compared) for which that 

 ratio is exactly equal to unity. If this radiation be precisely known, 

 the measurement of the ratio of its intensities in the two spectra 

 will give exactly the ratio of the total intensities of the two sources. 

 I have constructed a very simple apparatus which permits the 

 practical realization of the theoretic conception of which I have 

 just spoken. Let us illuminate the half-disks of the screen of a 

 Eoueault's photometer with an electric light and that of a standard 

 Carcel lamp : the two illuminated regions being placed tangentially, 

 it is very difficult to judge of the equality of the illumination, one 

 of the moieties having a bluish tint in comparison with the other, 

 which appears to be of an orange-yellow ; and the standard lamp can 

 be displaced within pretty wide limits without the eye being able 

 to judge if equality of illumination is really obtained. 



Let us look at the screen, placing in front of the eye two Xicol 

 prisms the cross sections of which are rectangular, and between 

 which is placed, perpendicular to the axis, a quartz plate of 9 mil- 

 lim. thickness. The eye then sees the two fields coloured of a 

 green tint washed with white ; and if we vary the distance' of the 

 standard lamp, a moment arrives when, equality of intensity being 

 obtained, the line of demarcation of the two fields disappears. A 

 very slight variation of the distance of the lamp then suffices to 

 give rise to a very sharp contrast between the two regions. 



The thickness of the quartz plate has been calculated so that its 

 interposition between the two rectangular nicols give rise, in the 

 spectra of the two sources, to two large interference-bands situated 

 in the two extremities of the spectrum. On passing from these 

 two bands towards the middle of the spectrum, the intensity of 

 the different preserved radiations varies as the square of the cosine 

 of the angle made by the cross section of the second nicol with the 

 planes of polarization of the different radiations, which have under- 

 gone rotatory dispersion in the quartz plate. There is therefore 

 one of them for which the square of the cosine is equal to unity, 

 and which suffers no weakening. It is conceived that it is pos- 



