August 19, 1898.] 



SCIEIWE. 



209 



point of observation through one opening 

 in the grating has travelled exactly one 

 wave more than it has in passing from 

 source to point of observation through the 

 nest adjacent opening. The spectrum of 

 the second order is that spectrum which is 

 formed when this difference of path amounts 

 to exactly two waves, etc. Thus when one 

 is observing the spectrum of the second or- 

 der, say, he has matters so arranged that 

 the light from the source is divided into 

 numerous thin beans, each of these beams 

 having to travel two waves farther to reach 



polishing a plate of optical glass until its 

 opposite surfaces are plane and parallel to 

 each other, and then sawing this plate into 

 pieces of the requisite size. 



Suppose now a beam of parallel light to 

 fall on this pile of plates perpendicular to 

 the face at A. In passing through the first 

 block from a to b the beam will be retarded 

 bj', say, 20,000 waves, the number depend- 

 ing on the thickness and the index of re- 

 fraction of the glass. Part of the light 

 then comes out into air through the narrow 

 opening he and the rest goes on through the 



Fig. I 



its goal than its adjacent beam on one 

 side of it, and two waves less than its ad- 

 jacent beam on the other side of it. If, 

 then, one wishes to obtain spectra of the 

 order 20,000, say, he must arrange the con- 

 ditions of the experiment so that this dif- 

 ference in the optical path for various sec- 

 tions of the light shall be 20,000 waves in- 

 stead of two. This result is accomplished 

 in the Echelon spectroscope by building up 

 a flight of steps of optical glass of a given 

 thickness. The diagram (Fig. 1) shows a 

 plan of this arrangement. 



block cd. The part of the incident beam 

 which comes out through the opening de 

 has also been retarded 20,000 waves over 

 the part that came through be and so on. 

 It is thus easily seen that this instrument, 

 when looked through along the axis from B 

 to ^4, gives the required conditions for the 

 production of a diffraction spectrum of high 

 order ; for we have the light between the 

 source and the point of observation divided 

 into a number of beams, each of these beams 

 having to travel 20,000 waves farther to 

 reach its goal than its adjacent beam on 



Fig. 2. 

 The blocks of glass composing this ' stair- one side of it, and 20,000 waves less than its 

 case ' must be of exactly the same thick- adjacent beam on the other side of it. 

 ness, a result which is accomplished by The cut (Fig. 2) represents a spectro- 



