The curves for the annexed figure (Fig. 28) were derived from measures taken in this manner. If the rays of a May-day sun are taken at 100, it will be seen what a rapid diminution there is in the green and the blue rays in gaslight. Gaslight only possesses about 20% of the green rays, whilst of the violet hardly 5%. On the other hand the light which comes to us from the sky shows a very marked falling off in the yellow and red rays. A very easy experiment will convince us of the difference in colour between skylight and gaslight. If we let a beam of daylight fall on a sheet of paper at the end of a blackened box, and cast a shadow with a rod by such a beam, and then bring a lighted candle or gas-flame so that it casts another shadow of the rod alongside, one shadow will be illuminated by the artificial light, and the other by the daylight. The difference in colour will be most marked: the blue of the latter light and the yellow of the former being intensified by the contrast (see page 198). Fig. 29.--Comparison of Sun and Sky Lights. By a little trouble the blue light from the sky may be compared with sunlight. A beam of light B (Fig. 29) is reflected by a silvered glass mirror from the blue sky into the box HH, at the end of which is a screen E. Another mirror A, which is preferably of plain glass, reflects light from the sun on to a second unsilvered mirror G (shown in the figure as a prism), which again reflects it on to the screen, and each of these lights casts a shadow from the rod D; K are rotating sectors to diminish the sunlight, and we can make two equally bright shadows alongside one another. The bluer colour of the sky will be very evident. CHAPTER IX. Colour Mixtures--Yellow Spot in the Eye--Comparison of Different Lights--Simple Colours by mixing Simple Colours--Yellow and Blue form White. The colour of an object in nature, without exception we might almost say, is due, not to one simple spectrum colour, or even to a mixture of two or three of them, but to the whole of white light, from which bands of colour are more or less abstracted, the absorption taking place over a considerable portion or portions of the spectrum. Notwithstanding this we shall now experimentally show that every colour can be formed by the simple admixture of not more than three simple colours, if they be rightly chosen, and from this we shall make a deduction regarding vision itself. We are in a position to obtain three simple colours by means of a slide containing three slits. Now for our purpose we require that the three slits can be placed in any part of the spectrum, and that they can be narrowed or widened at pleasure. Instead of a card the writer uses a metal slide, as shown in Fig. 30. Fig. 30.--Slide with slits to be used in the Spectrum.