Astronomical and Nautical Collections. 397 



on one side of their intersection, and to become continua- 

 tions of each other beyond that intersection . 



Thus, let PP' represent the 

 primitive plane of polarisa- 

 tion of the incident rays, OO' 

 the principal section of the 

 crystallized plate, and SS' 

 that of the rhomboid through 

 which it is viewed. Now, 

 the incident light, at first po- 

 larised in CP, is divided^ as 

 it passes through the plate, 

 into two parts, one which 

 undergoes the ordinary refraction, and receives a new po- 

 larisation in CO, the other the extraordinary, being polarised 

 in CE, perpendicular to CO. Let us represent the former 

 by F„ and the latter by F,. The passage through the 

 rhomboid divides F„, which is polarised in CO, into two 

 other systems of undulations, one polarised in the direction 

 of CS, the principal section, which may be called F„^, the 

 other in the perpendicular direction CT, and this may be 

 called F,^^. In the same manner F,, polarised in CE, is 

 divided by the rhomboid into two systems, the first F,^^, 

 polarised in CS, the other F,^^ polarised in CT. If we 

 examine the motions of the planes of polarisation of the two 

 pencils F„^ and F,^^, which contribute to the formation of- 

 the ordinary image, considering them on one side of C, we 

 see that, beginning from CP, they separate, and take the 

 directions CO and CE', and then meet in CS. Now, in this 

 case, the ordinary image will answer precisely to the differ- 

 ence of the paths described at the same moment by the 

 ordinary and extraordinary rays transmitted by the plate. 

 And following, in the same manner, the progress of the 

 planes of polarisation of the two pencils constituting the ex- 

 traordinary image, F,^^ and F,^,, we see that they both set 

 out from CP, and having assumed, in the plate, the direc- 

 tions CO and CE', instead cf approaching each other, they 

 continue to separate, until they form continuations of each 

 other in the directions CT and CT'; so that, according to 

 the rule given, we must add half an undulation to the dif- 

 OCT.~DEC. 1828, 2 E 



