3o6 ON LIGHT. 



medium corresponds to the series of larger balls in this 

 illustration. This ought to be so, for the velocity is less 

 in the denser medium as it is in the larger of two balls 

 after their collision ; and because, as already remarked, 

 the ether in such media must be either denser in propor- 

 tion to its elastic force, or somehow encumbered by 

 their material atoms. And hence w^e finally conclude 

 that in the act of the reflexion of light on the surface of 

 a rarer medium, the phase of the undulation clmnges, 

 and a scini-iuididation is lost (or gained it matters not 

 which) : but not so when the light is reflected from a 

 denser medium. 



(89.) To return now to the case of a thin pellucid 

 film. If its thickness, i.e.^ the interval separating its 

 two surfaces, be any number of j"(?;;^/-undulations; double 

 that number, i.e., an exact number of entire waves, will 

 have been lost by the wave reflected from the second 

 surface at its re-emergence from the first, by reason of 

 its greater length of path ; and thus were no part of an 

 undulation lost or gained in the act of rcflcxitm., it would 

 start thence in exact harmony with the first reflected ray. 

 r)Ut the second reflexion being made at the surfiice of a 

 rarer medium, an additional scmi-widulation will have 

 been lost, so that the two reflected rays will really start 

 from the first surface in complete disco7'dance, and destroy 

 each other. The same is the case if the thickness be 

 A'/7, or so excessively minute as to be much less than the 

 length of a wave, as at the vertex of a soap-bubble wlien 

 just about to burst. Here also will the same mutual 

 dcstruciiou of the reflected waves take place. And thus 



