374 ON LIGHT. 



only difference : viz., that the resultant rectilinear vibra- 

 tion will be performed along the other diagonal, H f, of 

 the same square. 



(151.) It appears, then, that a change of phase in the 

 vibrations of one of the component rays, of half an un- 

 dulation, exactly reverses the polarization of the com- 

 pound ray, and causes its vibration to be performed 

 along the diagonal H f instead of G e. Let us now 

 examine by what sort of grciditions the one of these 

 movements passes into the other, when the phase of one 

 of the vibrations C ^ is changed gradually. Suppose, 

 for instance, the vibration a b (so, for brevity, we will 

 designate it) to be in advance of the vibration a b by 

 one-twentieth part of a complete undulation, so that at 

 the moment w^hen c starts from c in the direction c a, 

 c shall have already got to i in the direction c a. Then 

 at that moment our molecule O will not be at O but at y. 

 After the lapse of one-twentieth more of a period, C will 

 have got to i in the direction C A, and c to 2 in the 

 direction c a, and O, actuated by both movements, will 

 have arrived at z, having of course described in the in- 

 terval a line y z, connecting these two extremities of tlie 

 diagonal of the rectangle xyz. And exactly in the same 

 way, at the expiration of the next twentieth of a period, 

 it will be found in z/, the extremity of the diagonal of 

 the next rectangle and thus tracing its course step by 

 step through the whole twenty, which constitute a 

 period, we shall see that it will have described a narrow 

 ellipse, having in n for its shorter axis, and e g for the 

 direction of\\.'s> longer, and touching the four sides of the 



