330 OPTICAL PROJECTION 



on the nozzle of the lantern and turning it into the proper 

 position (see Chapter XXII.) and the disappearance of the 

 rings, though there is light upon the screen, will demonstrate 

 that the reflection from both surfaces is necessary to their 

 production. 



Finally, the light destroyed by interference can be shown 

 by spectrum analysis of a slit crossing the rings. In blue 

 or red light such a slit would give an appearance like B or n 

 in fig. 181. It has been shown already that in blue light the 

 bands B are closer than in red light, hi which they appear 

 comparatively as at E. If, then, the slit across the rings is 

 dispersed into a spectrum, the largest and smallest rings 



must be connected 

 by curved lines 

 somewhat as in 

 fig. 181. The ar- 

 rangements for 

 the experiment 

 are shown in fig. 

 FIG. isi 182. The slit 



may be cut out 



of black paper and fixed to the front of the glasses, as at L, 

 being illuminated by the parallel beam from the flange nozzle 

 of the lantern N, the slit being focussed by the lens P and 

 dispersed by the prism P as usual. In this case the lantern 

 must be turned quite away from the screen, and the light 

 fall on the lenses very nearly normal, as shown in the figure ; 

 else the incident light will not be able to get down to the film 

 and out again through the slit, owing to the thick glass. 

 An easier way is to place the slit on the nozzle of the lan- 

 tern, with the parallel beam sent through it, and, bringing 

 the glasses as close to it as possible, to focus the reflection 

 of the slit in the glasses. A spectrum will in either case 

 appear, crossed by beautiful parabolic dark bands. 



A soap film is easily analysed in the same way, using the 



