BY A. L. McAfLAY. n.Sc. I!. A.. IMi.D. 



101 



fore, necessary to use the echelon in conjunction wi .'i an 

 ordinary spectrometer or other device for selecting a narrow 

 range of wave length. 



S 



e 



— ■m" 



— r 



L 

 K 



rMt«i--»e inn. 



Dispersion. — Let HK (Figure VIII.) be a wave front of 

 wave length \ + <i \ of the m th order. From general con- 

 siderations, it is obvious that the dispersion will be large, for 

 if the refractive index of glass were a constant, LK would 

 equal m ci ,\ [the ray BL would have an optical path m X 

 and the ray BK a path m i\ • <1 \) longer than the ray AH], 

 and thus with m lii;;cil^^ must be relatively large. Actually 

 the conditions are complicated by the fact that m is a func- 

 tion of \, fj- being the refractive index of glass. It can very 

 easily be shown, however (see references above), that the di.s- 



persion .1 o\\ \ '" ' '' >^''^ ^ 



Resolving Power. — Evidently to find th*^ angle, <1 '>. 

 between the maximum of a spectral line and its first mini- 

 mum, the discussion of section 8 holds without alteration, but 

 with the restriction that is always very small, and there- 

 fore that c(p« tf ', very nearly. Therefore, .1 (' ^ 



