Febeuaey 25, 1916] 



SCIENCE 



283 



distinct from the first. In the case of crossed 

 rays mN and nM, however, a uniformly 

 striated field is only incidental. There is al- 

 ways a second periodicity present, distinct 

 from the first, even if concealed. The stria- 

 tions are grouped in parallel strands. It is 

 now quite possible to obtain the linear phe- 

 nomenon with a wide slit and the occurrences, 

 when homogeneous light and the wide slit are 

 used, are merely a rhythmic reproduction of 

 the linear phenomenon, parallel to the slit; 

 i. e., transverse to the spectrum. 



To make this clearer, suppose the original 

 or regular striations are vertical and that 

 sodium light is used. Then the typical pattern 

 is of the kind shown in Fig. 2o. It looks like a 

 parallel set of thick twisted cords, hung side 

 by side and equidistant. It is often much 

 more complicated, though adhering to this de- 



positions of the grating for which this occurs), 

 on slightly passing one slit image of the other, 

 to different degrees, micrometricaUy, the ap- 





sign. The evolution of this pattern is obtain- 

 able on moving one of the slit images in the 

 field in different amounts micrometricaUy over 

 the other, keeping the longitudinal axes of the 

 spectra in coincidence. Then if the fringes 

 are originally nearly vertical apparently uni- 

 form striations, figure 25, they change by rota- 

 tion into the form figure 2c, and at the same 

 time enlarge. In other words, the lines h con- 

 sist of individual parts, behaving similarly but 

 independently, as if they were a set of mag- 

 netic needles. The same results may be ob- 

 tained with the single strand of the linear phe- 

 nomenon and white light and here the rotation 

 may be carried through nearly 180°, between 

 infinitely small sizes. It is fairly tumbling 

 in its mobility when of maximum size and 

 horizontal. Again, suppose the original reg- 

 ular fringes to have been horizontal and appar- 

 ently uniform. Then if the phenomenon is 

 made of maximum coarseness (there are two 



pearance presented is given in Fig. 3. The 

 fringes have become nodules, half black and 

 half brilliant, strung transversely like bean- 

 shaped beads on parallel strings and hung 

 vertically against a neutral (non-interfering) 

 yellow background of sodiiun light. In inci- 

 dental cases the black shadows may be a line, 

 and the field is then an apparently uniform 

 coarse grid; but generally they are separated 

 as in Fig. 3. Sometimes the central strand is 

 strongest and the intensity diminishes on the 

 right and the left. More frequently the two 

 central strands are equally strong. Five or six 

 strands may be present. On moving the 

 mirror M parallel to itself, these strands move 

 to right or to left as a whole, in accordance 

 with the equations of displacement inter- 

 ferometry. In fact, in view of the individual- 

 ity of the strands, the apparatus is a useful 

 displacement interferometer. 



The occiu-rence of these parallel strands for 

 crossed rays and homogeneous light is diffi- 

 cult to explain. I have tried a great variety 

 of things (slightly wedge-shaped compensators 

 and other methods of superposing special in- 

 terferences, etc.) to produce them with paral- 

 lel rays mM and nN, or to break them with 

 crossed rays mN and nM, without avail. There 

 is no focal plane effect, nor any polarization 

 effect. It is therefore necessary to confront 

 the case, at its face value, as in Fig. 4. Here 

 8 and 8' are the traces of two vertical, longi- 

 tudinally coincident, reversed spectra, drawn 

 apart for distinction, the region of the D 

 lines only being used. The light is homo- 

 geneous to this extent and the slit wide, so 

 that there is oblique incidence. Then every 

 point of 8 should (on adjustment) interfere 

 with every point of 8', the result showing as a 



