When equalized the sector is read off, and tells us how much brighter one spectrum is than the other. Thus suppose in the direct beam the sectors had to be closed to an angle of 80°, to effect this, the bottom spectrum would be 180/80, or 2·25 times brighter than the bottom spectrum. It should be noted that as the two spectra are formed by the identical quality of light, this same ratio will hold good throughout their length. If it does not, it shows that the double-image prism is not in adjustment, and that the same rays are not coming through the slit in the slide, and it must be rotated till the readings throughout are the same. One of the most sensitive tests for adjustment is to form a patch with orange light, when the slightest deviation from adjustment will be seen by the two patches differing in hue. We can now place the coloured transparent object in the path of the beam which is most convenient, and by again equalizing the shadows, measure the amount it cuts off; this we can do for any ray we choose. As both right-angled prisms can be attached to the card or slide which moves across the spectrum, nothing besides the card need be moved. In the following diagram we have the proportion of rays transmitted by the three different glasses, red, green, and blue, in terms of the unabsorbed spectrum. Take for instance on the scale of the spectrum the number 11. The curve shows that at that particular part of the spectrum which lies in the blue, the blue glass only allowed 4/100 or 1/25 of the ray to pass, whilst the green glass allowed 10/100 or 1/10 to pass. So at scale No. 4 in the orange, through the blue only 2% was transmitted, through the green glass 4%, and through the red 20%. Fig. 17.--Absorption by Red, Blue, and Green Glasses. Fig. 18.--Light reflected from Metallic Surfaces. Fig. 19.--1. Vermilion 2. Carmine. 3. Mercuric Iodide. 4. Indian Red. From such curves as these we can readily derive the luminosity curves of the spectrum, after the white light has passed through the coloured object. All we have to do is to alter the ordinates of the luminosity curve of white light in the proportion to the intensities of the rays before and after passing through the object. It will be seen that when the luminosity curve of the spectrum of _any_ source is known, this method holds good. Fig. 20.--1. Gamboge. 2. Indian Yellow. 3. Cadmium Yellow. 4. Yellow Ochre.