VOL. XXIX.J PHILOSOPHICAL TRANSACTIONS. 23f 



pendicularly through ab ; and the rest will emerge through bc, and by refrac- 

 tion make the imperfect spectrum de ; as in fig. 10. 



If the sun-beam enter ac perpendicularly, and in the middle of it, the light 

 will be all reflected, as in fig. 13; some of it by the plane bc to r, and the rest 

 by the plane ab to ^. But if the beam fall nearer to a (still perpendicularly) it 

 will all be reflected by the plane ab ; if nearer to b, it will be all reflected by 

 the plane bc 



In order therefore to have the coloured spectrum as it ought to be, care must 

 be taken that the emerging coloured light may make the same angle with the 

 plane bc, as the emerging light does with the plane ac; that is, the angle aeh 

 must be equal to bdg, as was said before, fig. Q ; which may also be seen on 

 the enlightened dust in the air. But the best way is to turn the prism on its 

 axis, and at the same time look at the coloured spectrum, which will rise and 

 fall, and become longer or shorter, as you turn the prism ; and between the 

 ascent and descent of the image, it will appear stationary : there stop the prism, 

 and the reflection will be such as is required for all the experiments hereafter 

 mentioned. 



In order to have the prism move freely on its axis, and stop any where, I fixed 

 each end of it into a triangular collar of tin, from the end of which came a 

 wire, which was the axis of the prism produced ; and thus I laid it on two 

 wooden pillars, with a notch on the top to receive the wires, and fixed it to a 

 small board, just broad enough to stand fast ; as in fig. 14. 



Exper. V. — I took the prism cd, and through it looked at the coloured 

 spectrum rp, which appeared then round and white as at s, just as if it had been 

 the sun's light received on a paper from the hole h, and seen with the naked 

 eye. In this case the prism cd must be held in a direct line with ab, and the 

 refracting angles in the two prisms must be equal. This spectrum appearing 

 v/hite but just in one point, is not so readily found ; but the best way is to look 

 through the same prism ab which makes the spectrum, which may easily be 

 done if it be pretty long, and then rp will be seen white and round, and as at 

 s, as if coming directly from h, in fig. 15. 



Exper. VI. — I held a broad lens l1, ground to a radius of 24^ feet, in such a 

 manner, that the whole coloured spectrum fell upon it ; and after refraction all 

 the colours appeared to converge, when received on a paper at pp ; but when 

 the paper was held in the focus at f, in the position ttPtt, the spectrum was 

 round and perfectly white, by the union of all the coloured rays. When the 

 paper was held at nn, the colours appeared to diverge from each other, but 

 then the red was uppermost, which before used to be the lowest, and so on in 

 an inverted order. 



HH 2 



