﻿316 Diffraction Gratings with Controlled Groove Form. 

 the nicol through a right angle the entire red end of the 



© o o 



spectrum above the dark band disappears, the portion o£ the 

 spectrum vanishing in this case being at least five or six 

 times the width of the dark band. 



A study was also made of the polarization of all of the 

 spectra and the central image. Though I doubt if these 

 data will prove of material assistance in the development of 

 a theory accounting for the bright and dark bands, I will 

 include them for the sake of making the study as complete 

 as possible. 



The grating was set at normal incidence which brought 

 the dark band to wave-length 65 b\ The incident light was 

 unpolarized, and the spectra examined through a nicol. In 

 the first order spectrum all of the light on the red side of the 

 dark band was polarized J,, the rest of the spectrum showed 

 an excess o£ = polarization. In the second and third orders 

 on the same side, the red end was nearly, completely polarized 

 = , the remainder showed an excess of X polarization. The 

 fourth order was completely polarized =. 



On the other side of the central image the very bright 

 first order spectrum showed little or no trace of polarization. 

 In the second order the red end was completely polarized X. 

 Third order no polarization, and fourth order all X. The 

 central image, examined with a spectroscope, showed a dark 

 band at 650, and all longer waves were found to be polarized 

 X as in the faint first order spectrum. The X and — signs 

 referred to above denote the direction of the electric vector 

 with respect to the direction of the groove. 



Jt is a matter of some importance to ascertain what 

 becomes of the light which is absent in the spectrum. If 

 one spectrum shows a dark band we should expect some 

 other spectrum to show a bright band at the same point. 

 This, however, is not the case, and we may infer from this 

 that the absent energy is distributed amongst the other 

 spectra, the amount received by each being too small to be 

 noticeable. 



It appears to me that the present paper contains sufficient 

 experimental data to thoroughly test any theory which may 

 be developed to account for these remarkable anomalies. It- 

 seems evident, that they are to be referred to the action cf 

 the sharp ridges and not to the bottoms of the grooves, as 

 they are profoundly modified by the lightest possible rubbing 

 of the grating with soft chamois skin. 



The extraordinary narrowness of the bright line (3 A.E.) 

 undoubtedly requires the co-operation of a large number of 

 lines, and it is to be noticed that its wave-length agrees 



