38 THE INTERFEROMETRY OF 



Two micrometers, one at M and the other at N, were installed, and moved 

 forward in alternate steps, within a range of over 2 cm., naturally without 

 modifying the fringes. These are now observed on both sides (A/" and M) , each 

 with the micrometer which is manipulated. One may note in passing that the 

 two screws are being incidentally compared. To set the 30 prism properly it 

 would have to be provided with a fine fore-and-aft, right-and-left slide adjust- 

 ment, in order that its edge may be set sharply in the line where the two 

 component rays intersect. An attempt was made to increase the dispersion 

 by allowing a spectrum to fall on the first prism (20), but without success. 



It is noteworthy that the 30 prism at P r is no marked improvement as 

 to range of displacement over the 90 prism at P', previously used. In other 

 words, the effect of decreasing the angle of reflection 6 at M is, unexpectedly, 

 of small importance, in relation to the range of displacement atM. This result 

 already treated in 16 will be accentuated in other ways below (Chapter II). 



19. Non=reversed spectra. Restricted coincidence. In figure 25, the white 

 ray L from the collimator is diffracted by the grating G and the two spectra 

 a and a', thereafter reflected by the parallel opaque mirrors M and N, to 

 be again diffracted by the grating G' '. The rays are observed by a telescope 

 at T. If the gratings G, G' have the same constant, it is obvious that the field 

 of the telescope will show a sharp white image of the slit, for each mirror. 

 If M N G G' are adjusted for symmetry by aid of the adjustment screws on 

 each and the rulings are parallel, the two white slit images will coincide 

 horizontally and vertically. If now a direct-vision spectroscopic prism or 

 a direct-vision prism-grating G" is placed in front of the telescope, the super- 

 posed white slit images will be drawn out into overlapping non-reversed 

 spectra, which will usually show a broad strip of interference fringes. 



In my first experiments, G and G' were film gratings of high dispersion 

 (D= I67XIQ- 6 ). The field was therefore too dark and the fringes were 

 obtained with difficulty because of the cumulative distortion of images from 

 the gratings. When found, the fringes were very fine parallel lines, filling 

 an irregular strip in the orange-yellow region, and it was already obvious 

 that an enormous play of the micrometer-screw at M would be permissible. 



A number of film gratings were now tested and the best samples selected 

 (D = lysXio- 6 ), although the dispersion was still too large and the D lines 

 not clear. To secure more light, a beam of sunlight 15 cm. in diameter was 

 condensed to a focus on the slit by a large lens of about 2 meters in focal 

 distance. The illumination was now adequate and the fringes were found 

 at once, as they should be; for they are always present, though not in the 

 same colored region. These fringes, found with more accurate adjustment, 

 were also larger than before. 



Figure 25 shows, if ab, a'b' and cd, c'd' are pairs of corresponding rays of 

 the same order but different wave-lengths, X and X' respectively; that for 

 the given position of G and G' only the rays a a' issue coincidently at T. 

 The rays cd, c'd' issue at e\, e'i, and, though brought to the identical focus 



