formed in the Prismatic Spectrum. 83 



(8.) When the plate is very thin, another set of appear- 

 ances presents itself. 



On splitting a piece of mica to such a tenuity that only a 

 few indistinct broad bands were barely visible with a flint- 

 glass prism, I observed at the same time another set of very 

 Jlne but extremely faint bands, evidently independent of the 

 former. 



(9.) When the film of mica was still thinner, the broad 

 bands ceased to appear altogether, and only \\izjine set were 

 visible. To show these bands the films must be so thin as 

 to be nearly iridescent : it is difficult to succeed in tearing them 

 off sufficiently fine. 1 have sometimes used drops of water 

 between glass plates pressed hard together. These bands are 

 always very faint; but they are somewhat more conspicuous 

 with prisms of the more dispersive oils, and always require a 

 strong light to be seen. 



(10.) There is, however, a more remarkable circumstance 

 connected with this set of bands ; they continue to be formed 

 when the edge of the thinjilm is towards the thicker side of the 

 prism. 



(11.) In pursuing the theoretical explanation, we have to 

 consider the conditions which may affect the rays situated 

 towards the opposite sides of those primary homogeneous 

 pencils, into which the incident beam of light is separated, 

 and which converge in the eye to form the several points in 

 the spectrum, both in the case of the prism, and of interfe- 

 rence from grooves or gratings. 



Now in either case a distinction of this kind is deducible 

 from the wave-theory, on comparing the length of undulatory 

 route of the two extreme rays of any primary pencil ; from 

 which it appears that one of these rays is always more retarded 

 than the other, as well in the prismatic as in the interference 

 spectrum : that side of the pencil which is previously the least 

 retarded, being that to which the plate is applied \u the original 

 form of the experiment. This distinction, combined with the 

 general principles of explanation at first referred to, appear 

 to me not only sufficiently to account for the ordinary phae- 

 nomena, but in my modification of the experiment to assign a 

 reason why a similar effect should be produced on the oppo- 

 site side. ,_ 



(12.) With regard to the mathematical investigation, in the 

 case of the prism, without going into a formal discussion, it is 

 sufficient to observe, that on the principles of mathematical 

 optics, when a diverging pencil of homogeneous light is re- 

 fracted through a prism in the position of minimum deviation, 



G 2 



