488 SCIENTIFIC RECORD FOR 1882. 



torj^ accordance of experiment with theory. {Am. J. ScL, May, III, 

 XXIII, p. 395.) 



Rowland has pirtjlished a preliminary notice of the results accom- 

 plished in the manufacture and theory of gratings for optical purposes. 

 Having devised a plan for making a practically perfect screw, which 

 proved successful, the ruling machine was at once constructed. No 

 error as great as the one-hundred thousandth part of an inch has been 

 detected in any part of this screw, and it has no appreciable periodic 

 error. Gratings have been ruled on the machine having 43,000 lines 

 to the inch, and a ruled surface has been made with it haA'ing 160,000 

 lines, 20,000 to the inch, in size 6^ x 4:^ inches. Eowland then made an 

 investigation of the theory of concave gratings and found their laws to 

 be very simple. Draw the radius of curvature of the mirror to the cen- 

 ter of the mirror, and from its central point, with a radius equal to half 

 the radius of curvature, draw a circle. This circle thus passes through 

 the center of curvature of the mirror and touches the mirror at its cen- 

 ter. If the source of light is anywhere in this circle the image of this 

 source and the difierent orders of the spectra are all brought to focus 

 on this circle. This leads to a mechanical contrivance by which we can 

 move from one spectrum to the next and yet have the api)aratus always 

 in focus. It consists simj^ly in attaching the slit, the eye-piece, and 

 the grating to three equal arms pivoted together at their other ends. 

 The most interesting case is when the bars carrying the eye-piece and 

 grating are attached end to end, forming a diameter of the circle, the 

 eye-piece being at the center of curvature of the mirror, and the rod 

 carrying the slit alone movable. In this case the spectrum viewed by 

 the eye-piece is normal ; and when a micrometer is used, the value of a 

 division on its head in wave-lengths does not depend on the position of 

 the slit, but is proportional simply to the order of the spectrum. More- 

 over, all the superimposed spectra are in exactly the same focus, and it 

 is a beautiful sight to see the lines appear colored on a nearly white 

 ground. A list of some of the flat and the concave gratings is given, 

 together with the results obtained with them. {Mature, June, xxvi, p. 

 211; November, xxvii, p. 95; Phil. Mag., July, Suppl., V, xiii, p. 4G9.) 



Frolich has studied the light which is reflected from very fine metal- 

 lic gratings, comparing it with that reflected from the non-striated sur- 

 face. When the light was polarized, the plane of polarization making 

 an angle of 45° with the plane of incidence, he observed that the two 

 portions of the black fringe of the compensator did not coincide, and 

 that the plane of polarization, being restored, is not the same for the 

 two halves of the field, the diflerence being 10° for an incidence of 70°, 

 the lines vertical. If the angle of polarization of the incident light q) 

 be varied, for the light reflected by the ruling the angle 4' of the re- 

 stored plane of polarization does not satisfy the equation tan q) cot ^' = 

 const. If cp be taken as 0, i/' = 5o 40' when i = GOo, the lines vertical. 

 If a lens of 10 meters radius be superposed on the grating, the center 



