210 Sir D. Brewster on the Connexion between 



the light reflected and transmitted by a single film by means 

 of a fine prism placed in front of a telescope, or examine the 

 prismatic spectrum produced by such an apparatus when it is 

 reflected or transmitted by the film in question. When we thus 

 examine the reflected tints of the three first orders of colours, 

 we find them to consist of that part of the spectrum which 

 gives the predominating colour of the tint mixed with the rays 

 on each side of it. The reflected green of the third order, for 

 example, consists of the green part of the spectrum, bounded 

 on one side with some blue, and on the other side with some 

 yellow rays, all the rest of the spectrum being wanting, having 

 passed, as it were, into the transmitted beam. In analysing, 

 therefore, the transmitted beam, its spectrum is found to con- 

 sist only of the violet and blue, and the orange and red spaces, 

 a dark band corresponding to the reflected spectrum separa- 

 ting it into two parts. In the higher orders of colours the 

 reflected spectrum consists of two or more portions separated 

 by perfectly dark bands, while the transmitted light exhibits 

 analogous bands, which are much less dark in consequence 

 of the tint being diluted with a portion of white light. The 

 coloured bands of the reflected spectrum occupy the same 

 place among the fixed lines of the spectrum as the dark bands 

 of the transmitted one ; and if the two spectra were superposed 

 they would form a perfect spectrum, whose rays when united 

 would form white light. Hence the reflected and the trans- 

 mitted tints are complementary to each other. 



When this analysis is made with a highly magnified spec- 

 trum, the numerous lines of which are distinctly seen, it 

 forms one of the most splendid experiments in optics. The 

 spectrum is crossed throughout its whole extent with alternate 

 dark and coloured bands, increasing in number and diminish- 

 ing in magnitude with the thickness of the plate by which the 

 tint is produced. 



If we use a thin film of mica, of such a thickness as polar- 

 izes the isohite of the first order, the transmitted spectrum will 

 be crossed by upwards of three hundred dark and three hun- 

 dred luminous bands, thirty-four of each being included be- 

 tween the lines C and D of Fraunhofer, a space less than one 

 tenth of the whole spectrum. 



W 7 hen we use polarized light, and interpose a doubly re- 

 fracting plate, and subsequently analyse the transmitted beam, 

 the spectrum is crossed with an analogous series of bands, 

 which are still more splendid and more perfect than those 

 given by a singly refracting film. The bands in the comple- 

 mentary spectra are equally and perfectly dark; and when 

 the tints are pure as in calcareous spar, the colours are nearly 



