Mr. J. Barton on the Inflexion of Light. 



an equation from which it is evident that the values of A or of 

 a may be determined, the values of the other quantities being 

 known. 



Thirdly. The theory of Fresnel is quite irreconcileable with 

 another observation of Newton's, which I have frequently 

 verified. When the interval between the two knife-blades is 

 very much diminished, the spectrum thrown on the paper ex- 

 hibits a dark space in its centre. " When the distance of the 

 edges was about the four hundredth part of an inch," says 

 Newton, " the stream of light parted in the middle, and left 

 a shadow between the two parts. This shadow was so black 

 and dark that all the light which passed between the knives 

 seemed to be bent, and turned aside to the one hand or to 

 the other. And as the knives still approached one another, 

 the shadow grew broader, and the streams shorter at their in- 

 ward ends which were next the shadow, until upon the con- 

 tact of the knives the whole light vanished, leaving its place 

 to the shadow *". 



The most satisfactory way of performing this experiment 

 is to employ two razor-blades, whose edges are slightly con- 

 vex ; for then the coloured bands will be seen 

 running round the dark space on either side, 

 as in the annexed figure. The position of the 

 dark space, and of the greatest divergence of 

 the coloured bands, answers of course to the 

 point where the razor-blades approach one an- 

 other most nearly. 



Now by the theory of Fresnel, so far from 

 the centre of the spectrum being occupied by 

 a dark space, it should be the brightest and 

 most luminous part of the whole; as will appear 

 on calculating the intensity of the light at dif- 

 ferent distances by the rule which he has laid down. Suppose, 

 for instance, the distance of the knife-blades from the hole in 

 the window-shutter to be 30 inches ; the distance from the 

 paper 21 inches; the interval between the two blades one four 

 hundredth part of an inch : then we have 



/ T&+b) = J_ / 2 (30 + 21) = ' 



C sJ abX 400 V 30x21 x -00002512 



And if x represent the distance of any point from the centre 

 of the spectrum, the intensity of the light at that point should 

 be as (y d v cos q v 2 ) 2 + (fdv sin q v*)* 



• Optics, Book III. Obs. vi. 



