226 Dr Brewster on the reflection and decomposition of' 



itself highly probable, and which has been already adopted to 

 explain the loss of part of an undulation in several of the phe- 

 nomena of interference. In such a case the reflection of the 

 light will commence at a line where the density or elasticity of 

 the ether in the first medium begins to change, and will continue 

 till the ray has penetrated to that part of the second medium 

 where the density or elasticity of the ether is uniform. In this 

 theory, therefore, the preceding facts may be regarded as prov- 

 ing the variable condition of the ether near the surfaces of 

 bodies, and of establishing the beautiful and sagacious deduc- 

 tion of Dr Young, that the part of an undulation lost is a 

 variable fraction depending on the nature of the contiguous 

 media. 



II. We come now to consider the second class of pheno- 

 mena, or the existence of periodical colours at the confines of 

 certain media of the same and of dififerent refractive powers. 



That the periods of colour arise, as in all similar pheno- 

 mena, from the interference of two portions of light cannot be 

 questioned ; though it does not appear how these interfering 

 pencils are generated. If we adopt the hypothesis of the re- 

 iflecting forces shown in Fig. 4, we may conceive the light re- 

 flected about C D to be interfered with by the light reflected 

 about C D^5 so that the same effect nearly might be produced 

 as if C D, C D' were the limits of a thin plate. If this sup- 

 position is not admissible, we may hazard the conjecture, coun- 

 tenanced by some facts which will presently be stated, that an 

 invisible film, differing in refractive power from the plate glass, 

 has been formed upon its surface. 



There is one phenomenon which has been more than once 

 mentioned, and which requires some farther notice ; namely, 

 the decrease in the intensity of the pencil as the incidence be- 

 comes more oblique. In re-examining this very perplexing fact, 

 which takes place in the prism B, though it does not produce 

 periodical colours, I have observed at a great incidence a dis- 

 tinct change of colour, from a bluish gray to a blue ; so that 

 I have no doubt that in this case the tints are those of a long 

 period approaching slowly to its minimum. This considera- 

 tion led me to suppose that in the case of balsam of capivi and 



