34 THE INTERFEROMETRY OF 



is not as bright as desirable, owing to the strong dispersion. When the grating 

 G' was replaced by a ruled-glass grating (Z) = 352Xicr 6 cm.), the dispersion 

 was not much reduced, but the light was better. The fringes were now found 

 after some searching and seemed to be of D\Dz breadth, a strip of oblique 

 lines of the usual character. But they were not brilliant and were hard to 

 recover when lost. The Fraunhofer lines were still disagreeably blurred. 



On exchanging the gratings (ruled-glass grating at G and film at '), though 

 the dispersion was smaller, the brilliancy of spectra was greatly improved. 

 The fringes came out fairly sharp. However, on cutting down the incident 

 beam at the collimator and near G to a breadth of not more than 0.5 cm., the 

 fringes were acceptable and capable of high magnification. They remained 

 visible for a displacement of 5 mm. at the micrometer at M. With fore-and- 

 aft motion of G', the fringes rotated as usual from fine vertical hair-lines, 

 through the horizontal (probably arrow-shaped forms of maximum size), 

 back again to hair-lines. Here the excursion of G' was about 1.5 cm. On 

 tilting the grating G' in its own plane and readjusting M, the rotation is 

 through the vertical maximum (the linear phenomenon). With a slotted 

 screen (0.5 cm.) at the collimator, the slit may be widened until the Fraun- 

 hofer lines just vanish. If the slit is but 0.2 cm., the fringes become bulky 

 and the play at M is but 2 mm. 



The film grating may be used by reflection, on adapting the apparatus in 

 figure 1 2 for this purpose, by supplying a ruled grating or prism at P and the 

 film grating (with its ruled side toward P) at G. If a ruled grating is put at P, 

 the spectra and fringes are good; but naturally there is deficient illumination. 

 Nevertheless a strong telescope may be used and a range of displacement of 

 4 mm. at M is available. This may be increased indefinitely by using a 

 micrometer at M and N alternately. The chief difficulty was the (incidentally) 

 unequal brightness of spectra. 



Again, the method of figure 14, apart from the drawbacks to which that 

 method is incident, succeeds almost perfectly, both in the first- and second- 

 order spectra. The fringes are strong and clear. An Ives grating of high 

 dispersion (D = 167 X icr* cm.) was tested. 



The method of figure 20, with auxiliary mirrors m and n to accommodate 

 the dispersion of G, was also successfully tried. Here G was originally a 

 concave reflecting grating. It was replaced by a film grating used as a reflect- 

 ing grating, with entire success. The ruled side of the film should be free 

 (without cover-glass), but the reversed side cemented on plate-glass as usual 

 and the latter placed towards the telescope at T. The prism P, in other 

 words, admits an abundance of light, so that even the loss in reflection from 

 the film is not serious. Sunlight should be used without a condensing lens ; 

 or, if the latter is added, the light leaving the telescope is to be narrowed 

 laterally. 



18. Non=re versed spectra. The prismatic method of cleaving the incident 

 beam of white light is available for the superposition of non-reversed spectra, 



