REVERSED AND NON-REVERSED SPECTRA. 47 



The beam of light, L, either white or homogeneous, as the experiment may 

 require, is furnished by a collimator (not shown), which, with the telescope at 

 T (placed in plan, in figure 31, at T or D}, are the usual parts of a spectro- 

 scope. The collimator with slit is always necessary for adjustment. It may 

 then be removed if the phenomenon is to be studied in the absence of the 

 slit. The telescope is frequently replaced to advantage by a lens. White 

 light is to be furnished by the arc lamp (without a condenser), by sunlight, 

 or by an ordinary Welsbach burner. Both spectra are naturally very intense. 

 A sodium flame suffices for the work with homogeneous rays. 



The adjustments in case of white light are simple and the interferences 

 usually very pronounced, large, and striking. Brilliant spectra, channeled 

 with vertical narrow black lines, are easily obtained when the longitudinal 

 axes are placed accurately in coincidence by rotating the plate h carrying 

 the grating H, on the plate /, around the axis g. If the gratings are quite 

 identical the sodium lines will also be in coincidence. Otherwise the two 

 doublets, DiD 2 and D'lD'?, of the two spectra (nearly identical in all their 

 parts and in the same direction) are placed in coincidence by rotating either 

 grating around a vertical axis. Thereupon the strong fringes will usually ap- 

 pear for all distances, e, less than 2 cm. These fringes are nearly equidistant 

 and vertical and intersect the whole spectrum transversely. They are not 

 complicated with other fringes, as in the experiments of the next section. 

 They increase in size till a single shadow fills the field of view, in proportion 

 as the distance e is made smaller and smaller to the limit of complete contact. 

 With the two adjustments carefully made, finally, by aid of the fringes 

 themselves, further trials for parallelism are not necessary. Two film grat- 

 ings, or even films, give very good fringes. During manipulations great care 

 must be taken to keep the angle of incidence, i, rigorously constant; i.e., 

 to avoid rotating both gratings together or the apparatus as a whole, as this 

 displaces the sodium doublets relative to each other and seriously modifies 

 the equations. 



20. White light. Colored fringes. The two sodium doublets seen in the 

 arc spectrum are usually equally brilliant, and but one set of strong fringes 

 is present in the field of the telescope. Relatively faint fringes may some- 

 times occur, due, no doubt, to reflection, as investigated in the next section. 



If both gratings are rotated, changing the angle of incidence from o to i, 

 the fringes disappear from the principal focal plane, but reappear strongly 

 in another focal plane (ocular forward or rearward). In such a case the D 

 lines are no longer superposed. To be specific, let i and i', 6 and 6', be the 

 angles of incidence and diffraction at the two gratings in question, the angle 

 between their ruled faces being i-i'. Let D and D' be the two grating con- 

 stants, and nearly equal. Then for a given color, X, in relation to the individual 

 normals of the two gratings, 



sin 6 sin i = \/D sin 6' sin i' = \/D' 



