92 



parallel and of the same thickness — each one a little wider than 

 the preceding as in Fig. 6. 



Fig. 6. 



Thus, if the pile has forty plates, each one centimeter thick, 

 the retardation will be about ten thousand light-waves; and the 

 revolving power would be forty times this, or four hundred 

 thousand — which is about four times as great as that of a six- 

 inch diffraction grating of the usual form. The number of ele- 

 ments might be increased till the absorption of the glass brought 

 a limit. A diflficulty, which appears long before this limit is 

 reached, is due to the loss of light by repeated reflections be- 

 tween the many surfaces. This has been very ingeniously over- 

 come by Mr. Twyman, of the firm of Hilger & Company, by 

 pressing the plates tigether to actual contact, when the re- 

 flection vanishes. It is likely that the echelon under these con- 

 ditions may be used by reflection instead of transmission (the 

 plates being silvered for the purpose) with the advantage of 

 quadrupling the resolving power for the same number of plates 

 and eliminating the absorption. 



An illustration of the efficiency of the echelon spectroscope 

 is furnished by the following photographs of the spectrum of 

 green radiations from mercury vapor. The first of the figures 

 shows the spectrum of the second order of a diffraction grat- 

 ing whose ruled surface is nine inches by four and a half — the 

 largest in existence. The second is by an echelon of thirty 

 plates, each an inch and a fourth think (30 mm). The cor- 

 responding lines are similarly lettered in the three figures. The 



