﻿Gratings with Controlled Groove Form. 313 



with celluloid deeply stained with an aniline dye, which was 

 sectioned with a microtome and examined under the micro- 

 scope. The grating was mounted on the table of a large 

 Fuess spectrometer so arranged that the table turned with 

 the graduated circle. At normal^ incidence it showed a 

 black band in the red about 30 Angstrom units in width, 

 the centre of which was at 6560, and a bright line 3 

 Angstrom units wide at 6400. 



The telescope and collimator were not used, the spectrum 

 being viewed by the eye directly. The source of light was 

 a Nernst filament situated at a distance of about 3 metres 

 from the grating, with a very brilliant neon tube imme- 

 diately behind it. A Nicol prism was mounted in front of 

 the lamp, for the bands appeared at their best with polarized 

 light (electric vector perpendicular to the grooves) » Behind 

 the neon tube an arc lamp was mounted for producing the 

 very intense illumination necessary for the experimental 

 determination of the wave-length of the light passing off 

 from the grating at grazing emergence. On viewing the 

 spectrum in the grating, the dark band or the bright line 

 could be brought into coincidence with some one of the 

 numerous bright neon lines, by changing the angle of 

 incidence. 



As a check on this method of determining the position 

 of the bands the arc was lighted, and a Schmidt and 

 Haensch pocket-spectroscope, provided with an illuminated 

 scale of wave-lengths, was directed at the spectrum seen in 

 the grating. The dark band then appeared in the spectrum 

 seen in the small spectroscope, and its wave-length could be 

 read off at once. The width of the bright band was deter- 

 mined with the large spectrometer with fixed telescope and 

 collimator of 2*5 metres focus, by comparison with the 

 D lines. 



Considering the grating mounted as in fig. la, the faint 

 first order spectrum in which the abnormalities are seen is 

 viewed in the direction A. The first order spectrum on the 

 other side (direction B) is extremely bright, being favoured 

 by the broad, gently sloping sides ot' the groove. The 

 spectra passing off at grazing emergence are very bright 

 in the direction C and very faint in the direction D, as is to 

 be expected, for the steep, narrow sides of the groove favour 

 the direction C. The circle of the spectrometer was sot at 

 zero, and the grating turned on the table until the incidence 

 was normal, as determined by reflecting the imago o\' tho 

 ruled spot back to the arc. 



Phil Man. S. 6. Vol. 23. No. 134. Feb. 1912. V 



