ON LIGHT. 369 



of them be slowly turned round in its own plane till its 

 axis comes to be situate at right angles to that of the 

 other, they will gradually decrease in intensity and at 

 length disappear altogether when this rectangularity is 

 precisely attained. In the first case, then, the rays have 

 interfered in the last, not : while in the intermediate 

 states a partial interference takes place, the more com- 

 plete the nearer the axes are to parallelism. How tJiis 

 is operated we shall now proceed to explain. 



(148.) Cii'cidar and elliptic polarization. If we regard 

 the vibratory movement of any single particle of an elas- 

 tic medium in its most general mode of conception, we 

 shall find that it may always be considered as capable of 

 resolution into three rectilmear vibrations in three planes 

 at right angles to each other, each going on as if the 

 others had no existence : and its place in space at any 

 instant will be had by estimating its distance on one side 

 or the other of its neutral or central position (those of 

 perfect equilibrium and rest), reckoned along each of the 

 three lines in which these planes intersect (which, after 

 the manner of geometers, may be considered as three 

 rectangular axes, or co-ordinate lines), which it would 

 have attained at that instant in virtue of each separately, 

 and independent of the others. This is nothing more 

 than the enunciation of one of the simplest of mechanical 

 laws, that of the composition and resolution of motions. 

 But the theory of movements propagated through elastic 

 media (a theory far too elevated and intricate to admit 

 of any explanation in these pages, and whose results the 

 reader must take for granted) further teaches us that a 



2 A 



