CIRCULAR POLARIZATION. ]f;f) 



to tlicir original coincidence, or by increasing llin discrepancy between tliem by 

 repeated rcllcctioas until it amounts to halt' an undulation. In the first case, if 

 the reflection were total, as at the second surface of glass, the plane of polari- 

 zation, after restoration, would be unaltered. In the second case, on similar sup- 

 positions, it would be changed 90°. When the reflection is not total, the, 

 resultant plane, after the reunion of the displaced nodes, will difler from the 

 origind plane in consequence of the unecjual losses experieuced by the two com- 

 ponents in reflection. 



Both those methods have been employed in experimentally determining the 

 dislocatiiig effects of different media, at different incidences, upon the rectangular 

 componeats of plane polarized light. Many of Mr. Jamin's more recent and 

 elaborate researchers have been made by the meLhod first mentioned. The essentia 1 

 part of his apparatus consisted in a double prism, formed of two equal acute 

 wedges of rock crystal, cut parallel to the axis, and combined as shown in the 

 ,^ figure annexed. The wedge ABC has the edge AB y,"//-- 



atlcl to the axis, and the wedge ADO has the edge DC 

 jjcrpen'dicular to the axis. The surfaces AB and DC arc 

 both ixirallel to the axis. If a plane polarized ray, PQ, 

 pass at right angles to AB through the middle of this 

 system, wliere the wedges are equally thick, it will remain 

 plane polarized, and the position of its plane Avill remain 

 unaltered v/hatever be the azimuth of incidence ; the dis- 

 locating effect of the double refraction of one of the wedges being compensated 

 by an equal and opposite effect of the other. But if the .system bf moved to 

 the right or to the left, the two opposite effects will no longer bo equal. This 

 being a positive crystal, the component of molecular movement parallel to AB 

 will be in retardation of that perpendicular to AB, for the position of the ray, 

 P'Q'. and will be in advance for the position P' Q". But if the ray has been 

 already dislocated by reflection, some part of the prism may be f )und Avhicli 

 will produce au equal and contrary effect, so as to restore the plane polarization 

 All that is necessary, then, to make this instrument a measure of the dislocation 

 is to connect it Avith a scale and a screw movement, and to determine the value of 

 the scale divisions. This last determination is easy, since a run which converts 

 a plane polarized ray into a ray plane polarized with a dift'erence of azimuth of 

 90*^, is equivalent to half an undulation. Mr. Jamin's apparatus accomplishi'd 

 this change in twelve complete turns of the screw. Smaller divisions were 

 measured by the graduation. of the screw-head, of which there were two hun- 

 dred divisions to the revolution. The delicacy of the contrivance may be appre- 

 ciated from this statement. It is known as Mr. Jamin's "compensator." 



It would occupy too much space to attempt here to give a full account of i\[r. 

 Jamin's interesting researches, o*r of the methods employed by him auxiliary to 

 that just described. The most important results of his investigations are the 

 following : 



Nearly all transparent bodies produc.-, by reflection from their surface?, a 

 difference of phase between the component waves polarized in the two princip,d 

 planes. All whose indexes of refraction exceed 1.414, advance the phase of the 

 component polarized in the plane of incidence. All those whose indexes are 

 below that value retard the phase of the same component. 



The difference of phase augments with the incidence from ^l at 0' to x at 

 90°, and becomes fA at the polarizing angle. The variations are slow and 

 almost insensible for some distance from either 0^ or 90^^. They usually become 

 sensible near the polarizing angle. The limits are nearer together as th(i polari- 

 zation under that angle is greater. 



Beyond these limits the polarization is plane, but imperfect. Within them it 

 is elliptic. 



The two limits are nearer to each other as the ind<^x of refraction approaches 



