100 rHOI'KRTIKS OK OlTKAl- OIlATINflS. 



Why cannot a two line grating resolve two specti-al lines 



for which is less than 6, although the expression ob- 



tained above would give G as ths resolving power in its third 

 order spectrum? 



SECTION 11. APPLICATION TO THE ECHELON. 



In this instrument some twenty plates of optically plane 

 parallel slabs of glass, all accurately of the same thickness 

 (about 1 cm.), ai'e piled above one another, each one over- 

 lapping the one beneath it by about a millimetre. Light is 

 passed normally through the pile, emerging from the over- 

 lapping ledges, which behave as the clear spaces of a grating, 

 the beam coming from one ledge being retarded many 

 thousands cf wave lengths behind that from the ledge next 

 it, owing to its passage through a greater thickness of glass. 

 The spectra observed are thus of about the tsn thousandth 

 order. 



The following brief discussion should be supplemented by 

 reading. Treatments will be found in Schuster, Theory of 

 Optics, p. IIG; Baly, Spectroscopy, p. 190; Wood, Physical 

 Optics, p. 274; and in other text books. 



Figure VIII. shews two plates of an echelon. HL is a 

 wave front for wave length ^ in the m th order. 



The Visibility Curve. — The width of the slil is about 1 

 mm. Thus (sections 3 and 4) the breadth of the central 

 maximum of the visibility curve is given by - wln-ic «iii ft 



X/e <V'<JO~' Therefore -iff- about 4 minutes. Out- 

 ' 1 .; 1".-' 



side this narrow range the spectra have not v.ufficicnt inten- 

 sity to be observed, consequently the echelon is only useful for 

 examining the fine structure of a spectral line or determining 

 the separation of two lines very close together. 



Separation of the Orders. — Let HK (Figure VIII.) be a 

 wave front for wave length \ m 'In- m + I th order, e is 

 about 1 mm. Therefore, •\ (> l.K/IIL \/e the angle 



between the two orders, is about !_? ^-L. i.e., is half the 



•1 



breadth of the central maximum of the visibility curve. The 



result is that there are in general two orders of every wave 



length in view in the field, and that the different lines are 



piled on one another in an inextricable jumble. It is, '.here- 



