164 LORD RAYLEIGH ON THE COLOURS OF THIN PLATES. 



thus the total of this component reaches only 178, while the two other com- 

 ponents are present in fair quantity. The resulting colour is a good orange. 



As V increases, the dark band moves down the spectrum. When V = 1951, 

 the centre of the band is at (44) ; thus nearly all the green is eliminated, 

 and the colour is a rich purple. Again, when V = 2328, the centre of the band 

 is at (24), the resulting colour is a rich blue. This band then moves out of the 

 visible spectrum ; but a new one presently makes its appearance, and begins 

 to invade the spectrum from the violet end. When V = 2 x 1688 or 3376, the 

 ray (68) is again extinguished, and the colour is the yellow of the second order. 

 For higher values of V, there may be two or more dark bands simultaneously, 

 as appears in Table II., when V = 6800. 



§ 5. Any sequence of colours may conveniently be represented on Newton's 

 diagram, in the manner adopted by Maxwell for the particular sequence found 

 in the spectrum. Such a curve would represent, for example, the colours of an 

 absorbing medium, as the thickness traversed varies from nothing to infinity. 

 In all suck cases the cui've starts from the point white, and ends at the point 

 representative of that ray of the spectrum to which the medium is most trans- 

 parent. For many coloured media the curve would not depart widely from a 

 straight line ruled ont wards from white to a point on one of the sides of the 

 triangle. But when the medium is dichromatic, as for example a solution of 

 chloride of chromium, the curve might start in one direction and ultimately 

 come round to another. Thus in the case referred to the course of the curve 

 from white would be towards the middle of the blue side of the triangle, then 

 after a good progress in that direction it would bend round through yellow, and 

 ultimately strike the triangle at a point near the red corner representative of 

 the extreme visible rays at the lower end of the spectrum. The principal object 

 of the present investigation was to exhibit in a similar manner upon Newton's 

 diagram the curve of the colours of thin plates. To find the point correspond- 

 ing to the retardation 1688, we imagine weights proportional to the numbers 

 2 - 04, 1*49, "18 to be situated at the three angular points of the triangle, and 

 construct the centre of gravity of such weights. This point represents the 

 colour due to retardation 1688. 



§ 6. The diagram (Plate X.) embodies the results of Table III., so far as the 

 quality of the effects is concerned. When the thickness, or retardation (V), is 

 infinitely small, the amount of light reflected of course vanishes, but the colour 

 approaches a limit, found by combining the constituents in quantities propor- 

 tional to \~' 2 , the limit of sin 2 (7rV/\). This limiting blue of the first order 

 would be the blue of the sky, according to the theory which attributes the light 

 to reflection from thin plates of water in the form of bubbles. The blue of the 

 sky is, however, really a much richer colour than this, and corresponds more 

 nearly to that calculated on the supposition that the disturbance is due 



