AND ON THE COLOURS EXHIBITED BY CERTAIN FLAMES. 459 



to the angle subtended by the interval between the corres- 

 ponding edges AB of tlie slits ; and this being known, the dis- 

 persive power is easily calculated. In fact, if i be the interval 

 between the coiTesponding edges of the slits observed, and d 

 their distance from the prism, A the angle of the prism, D the 

 deviation of the extreme red rays, ^o, the refractive index for 

 those rays, and p the dispersive power, we have 



p = ~~ where S^o- ~ 



cos 



(^) 





For example, to determine the dispersions of 8 specimens of 

 glass, I found by this method as follows : 



These results will doubtless appear extraordinary, after all 

 that has been written on the subject of dispersive powers. 

 The highest value of p, for flint-glass, given in Dr Brewster's 

 Table, is only 0.052, which is f th lower than any of these re- 

 sults ; yet I have no doubt of the correctness of my observa- 

 tions. The agreement of the dispersions of the three first spe- 

 cimens of flint, within a 260th part of their whole quantity, 

 while the refractive indices differ so considerably, is not a little 

 remarkable. The prisms used were made at once, by the same 



artist, 



