29 



Combining this with (4) Art. 1.. p. 130, of the paper just quoted 

 we obtain 



dtV^J a'eos^n{t—a)' ^^ 



From this we infer that, to an observer receiving the light, the 

 revolution in the ellipse is 



(4) 



right-handed I ,. ( «i > «2 I 



[ according as | . 



left-handed j [ a^ < a, j 



Again, from (1) in Art. 1 above it follows 



we niay therefore conclude that the revolution is 



right-handed I ,. { x^ — x, < I ,„. 



^ according as ^ ^ [. (6) 



left-handed J [ Xi — Xg > ^ j 



§ 4. Whether the revolution in the ellipse is right-handed or 

 left-handed will ultimately depend, as we have seen, upon the sign 

 of the difference Xj — x» of the two coefficients of extinction of 

 the medium. On the other hand, from equation (2) in Art. 1 above 

 it appears that the sense of the rotation of the major axis of the 

 ellipse [i. e. the sign of our angle xp) is determined by the sign of 

 the difference of the velocities with which the two circular com- 

 ponents of the wave are propagated. The question now arises, Can 

 any connexion be traced between the sign of the rotation ip and 

 the direction of the revolution in the ellipse? In the case of cryst- 

 alline media this question was long ago answered by Babinet 

 who formulated the well-known rule: «le rayon lemoins ab- 

 sorbs est celui qui se propage le plus vite». This rule 

 however does not bear the test of examination. In the case with 

 which we are here concerned, that is in the case of circular di- 

 chroism in naturally-active media, there can be no doubt, as we 

 shall presently see, that Babi net's Rule as a general law i& 

 inaccurate. 



To a courteous private communication from Mr A. Cotton we 

 owe the important remark that one of the results we have arrived 



