for the Dispersion of Light. 119 



log k = 3-87943 



log / = 3-65263 



log m = 3-94185. 

 In any particular medium then, taking the logarithms of the 

 differences of the three indices, we easily obtain the fourth by 

 the above formula. 



For the corresponding relation of the other rays we must 

 take a formula analogous to (21.), which will be as follows: 



r (/^C-'^b) (■^g'-TS") (^E^—^i") (^G'—^i") 



= <| -(f^E -l-u) (^G^-^i^) (^^'-^i') ijo'-^Z') 



L+(^g-/*b) C^i'-^B^) {-^c'-^b') {r^'-^c"-) 

 which for brevity may be written as before, 



= Qj.^ -[^^) k'- (f*j,-/^B) V + {fx.o -l^s) m'. 



The coefficients k' I' m' may be found exactly as before 

 from Fraunliofer's values of t^ t^, t^ t^. which give 



Tc^ = 17-045, -ri^ = 26-437, t^^ = 39-705. 



By means of these we obtain 



log A' = 3-54623 

 log /' = 2-93137 

 log ?«'= 2-28794. 



1 will only add at present that I am now engaged in deter- 

 mining by observation the indices for various media, especially 

 those of a highly dispersive nature : and in the few attempts 

 as yet made to verify the theory in these cases, (in which it is 

 manifestly put to a more severe test than in any of the cases 

 hitherto given,) I have found the method just explained by 

 far the most preferable. The approximate method in any 

 form appears to me at once more troublesome and less satis- 

 factoiy. The values of the constants applicable to nil media 

 above given, may be useful to those who may engage in such 

 calculations, or in verifying those already pertbrmed by other 

 methods. 



Oxford, June lytli, 1836. 



