458 ON THE ABSORPTION OF LIGHT BY COLOURED MEDIA, 



one extremely faint bluish-green, and one almost evanescent 

 violet J its type being in consequence as in Fig. 18. 



It is needless to insist on the advantage that may be taken 

 of these and similar properties of coloured flames and media, 

 in optical researches. The power they afford of insulating 

 rays of several species, of a refrangibility perfectly definite, 

 and capable of being identified at all times, offers every facili- 

 ty for a more exact examination than has hitherto been under- 

 taken, of the dispersive action of media on the intermediate 

 rays, as well as for the direct determination of the dispersions 

 of the extreme ones. A method which I have sometimes 

 practised with success for the latter purpose, is as follows : Let 

 two parallel slits A A, BB, the one half the length of the other, 

 and each about two-tenths of an inch in breadth, be made in 

 a screen, and being placed horizontally in a window, examin- 

 ed with a prism of known refracting angle, having its edge ho- 

 rizontal. The eye being defended by a thickness of blue glass 

 (such as is described in Articles 3. and 5.), sufficient to stop the 

 green, yellow, and most refrangible red ; two images a a, a! a\ 

 and b b, b' b', of each slit are seen a red and a violet of each ; 

 the red being perfectly sharp and well defined^ the violet also 

 pretty well defined at its most refracted edge. The eye and 

 prism being now gradually withdrawn, the violet image a' of 

 the longer slit will approach the red b of the shorter one, and 

 at length be seen projected upon it, (the apparent angle be- 

 tween the slits diminishing, while the angle of dispersion re- 

 mains invariable). When arrived at such a point that (the 

 prism being placed in a position of minimum deviation) the 

 corresponding edges n' a! and bb form one straight line, which 

 may be hit with very considerable exactness ; it must be stop- 

 ped, and its distance from the slit measured. In this situa- 

 tion, the difference of deviation of the extreme rays, is equal 



to 



