300 I'HILOSOPHICAL TRANSACTIONS. [aNNO 1728. 



Exper. 7. — The Doctor took the cube of fig. 8, and looking obliquely 

 through it at the hole of the window of his dark chamber (the sun shining or 

 not shining) the hole appeared entirely colourless, as did also a candle; both 

 appearing fringed with colours when seen through the prism. Then holding 

 two prisms together, as in fig. 10, if the liole of the dark chamber be at k, it 

 appears white to the eye at e ; but if the angles of the prisms at bp be a little 

 separated, while the points ad touch, the hole will appear coloured : when the 

 surfaces are separated at ad, and touch at bf, the colours appear in an inverted 

 order ; but if the surfaces ad and bf are parallel, whether they touch or not, 

 the hole will appear white. 



N. B. In this case the prisms must be similar, that the surface fg may be 

 parallel to ac ; otherwise ab and dp must be so inclined to each other, as to 

 render ac and fg parallel. Indeed if one of the prisms be very far removed 

 from the other, the heterogeneous light which entered in at fg, may be so far 

 spread by the separation of the differently refrangible rays, that the prism abc 

 will not take it all in ; then the eye behind the second prism may see colours, 

 as it seems Kizzetti did. See page 79 of his book. 



If the ray of light Rabcdd, fig. 15, changing its direction in the manner 

 abovementioned, makes an angle of about 45 degrees with the perpendicular pa; 

 on the removal of the lower prism, the ray will be turned up again, as in fig. 14. 

 But if the angle pan be greater, the ray will still be turned up again in a curve, 

 as abcdef, fig. l6, notwithstanding the lower prism is at dfg; but if that prism 

 be brought up closer to the surface ab, the curves will be destroyed where the 

 prisms touch, and all the rays in the place of contact brought down through 

 the lower prism. 



The most refrangible rays consist of smaller particles than the least refrangible 

 rays, and therefore must have least momentum, the velocity of all the rays be- 

 ing the same ; and consequently are more easily turned out of the way by at- 

 traction or repulsion, which makes the curves made by the purple and violet 

 rays, under the surface ab, to be less and nearer the said surface, than the 

 curves made by red and orange rays. 



Suppose a violet ray moving in the direction Rr, fig. 1 6, to be so bent under 

 the surface ab, that at the vertex of the curve, or where its tangent cc is 

 parallel to ab, there still remains a small space between the curve and the line 

 nn, where the limits of attraction, contracted by the proximity of the under- 

 most prism DPG, end, that ray will be turned up again in the curve def, and so 

 reflected in the line ii, the directions having been successively changed, as in 

 fig. 14. But a red ray, with the same inclination, would pass on into the lower 

 prism, as was explained in fig. 15. Because the momentum of the red ray be- 



