SUPERPOSITION OF PARAGENIC SPECTRA, ETC. 229 



and indistinct. The light being incident perpendicularly, and the gratings turned 

 round AM on a plane perpendicular to AM, the fringes do not increase in number 

 or greatly change, if the motion is accurately in a plane perpendicular to AM. 



When the gratings are turned in the plane of the horizon passing through 

 AM, the side NM approaching the eye, the fringes on the left-hand spectra de- 

 scend, increasing rapidly in number, and when the side MN recedes from the eye, 

 the fringes ascend, increasing in magnitude and diminishing in number, and are 

 highly coloured. At a certain angle, they become parallel to the grooves, when 

 by continuing the rotation they move downwards increasing in number and be- 

 coming parallel to the grooves. 



In the preceding experiments, the bands are seen on the surface of the gratings, 

 but when the grooved surfaces are in contact, and the grooves parallel, bands of 

 an entirely different kind are seen, not on the surface of the gratings, but by rays 

 diverging from the luminous disc. If we use a long and narrow bar of light, such 

 as the opening between the window-shutters, then, when the grooves are parallel 

 to the bar, and the grooved surfaces perpendicular to the plane of incidence, the 

 bands are parallel to the bar and its spectra. By inclining the grooves to the 

 luminous bar, the bands are inclined to the spectra, dividing each of them into a 

 great number of spectra, and at an azimuth of 45° the bands become perpen- 

 dicular to the spectra. At all these inclinations the bands on the second spec- 

 trum are double those on the first, the number increasing in arithmetical progres- 

 sion on succeeding spectra. 



When the angle of incidence is increased, the bands increase in number, but 

 very slightly with gratings of 1250 divisions in an inch. 



By increasing the distance between the gratings, the bands also increase in 

 number. 



Bands similar to those now described are produced with interesting pheno- 

 mena by a single grating placed as in fig. 8, so that the image of the grooved 

 surface AB, reflected from MN, the lower surface of the glass is superposed as it 

 were upon the grooved surface itself. 



1. When the plane of reflexion is perpendicular to the grooved surface, and 

 the grooves in the same plane, the bands on the spectra are parallel to the bar 

 of light AB, those on the second spectrum being double those on the first. They 

 are seen at all angles of incidence, and are larger and more distinct at small angles. 



When the grating is turned round in its own plane, at any angle of incidence, 

 so that the grooves form different angles with the bar of light, the bands cross 

 the spectra and become perpendicular to them in the azimuth of 45°. The 

 paragenic spectra are thus divided into a great number of spectra, the number in- 

 creasing as formerly on each succeeding spectrum. 



2. When the grooves are parallel to the bar of light, and the plane of reflexion 

 perpendicular to the grooves, the bands are apparently segments of concentric 



