366 Mr. G. Dodd on 



3. Impinge nearer to base than to apex hy ^th. 



Inclined 5** from base. 



"Jinclin, 



direct ray 6 > =50 /toward 

 1">1'5* Jbase. 



Sinclin. 

 3,13, 19=5°Sfrom 

 3 base. 



Inclined 5° towards base. 



"Jinclin. 

 3,9, 13, 19=5° [toward 

 3 base. 

 ■Jinclin. 

 direct ray, 6, 10=5° Sfrom 

 3 base. 



Thus, it appears, that if a homogeneous ray, red for 

 instance, were traced through a prism, the positions of the 

 different emersions with respect to each other would be 

 most materially altered by any variation in the point of its 

 entry, or the direction of its passage. 



Having treated of a simple homogeneous ray, I will now 

 proceed to a ray of white light, under circumstances more 

 consonant with general experiment. 



Let us assume, as before, that there are but three colours, 

 and three refractive indices : — then let the prism be equi- 

 lateral and of flint glass, with the refractive indices of the 

 mean, and the two extreme rays, respectively 1*620, 1*600, 

 1*580 (which is about a mean between Fraunhofer's table 

 and those of other observers) : then, let a ray of white light 

 impinge midway between the apex and the base, and at such 

 an angle, that the central ray shall pass parallel to the 

 base : with these data I have traced the progress of all the 

 three rays through twenty reflexions, and fig. l.(a) repre- 

 sents the points at 'which all the emersions would take 

 place at the posterior face ; in fact, it would be the form of 

 the spectrum immediately on leaving the prism. 



Let the white ray now impinge at a point nearer to the 

 apex than to the base by J^th, but let the central ray still 

 pass parallel to the base : it will now be found that the 

 spectrum immediately on leaving the prism would be shorter 

 than before, and not only so, but the relative disposition of 

 the colours very different, as in fig. 1 {h). 



Let the case be now reversed, and the ray impinge 

 nearer to the base m the same proportion, the parallelism 

 being still preserved : the same process of tracing the ray 

 through twenty reflexions would give us a spectrum as in 

 fig. 1 (c). 



