434 C. Barus — Interference of Reversed Spectra. 



geneous diffracted rays condensed at b, that about '5 cm of spec- 

 trum, d'a' and ad on either side of a, is chiefly active in 

 modifying the resulting diffraction pattern. "Within this, the 

 homogeneous rays, cc' and dd\ are capable of interference. 

 Although the wave fronts entering b are slightly spherical, 

 their radius is about r = 1 meter and they may therefore be 

 regarded plane. In such a case the angular width dx of the 

 illuminated strip at b, for a width of screen dd' — l cm , between 

 two extinctions, may be written 



dx X 60X10" 6 



dO = — = -5-=- = — ^— - = 6X10- 6 

 r dd 1 



whereas the angular breadth of the D X D^ doublets is about 

 37xl0 -5 ; i.e., the rays from d and d\ if in phase, should 

 cease to illuminate b at a breadth of about 1/6 the distance 

 between the sodium lines. The rays within dd' would corre- 

 spond to greater widths ; those from cc ', for instance, '5 milli- 

 meters apart, would illuminate twice the estimated width, so 

 that as trip at b, with a breadth of one-third the interval D^D„, 

 is a reasonable average. All rays, however, would produce 

 illumination at b. As the screens are narrower, not only would 

 the fringe be broader, but more lines would appear because 

 there is less overlapping. All this is in accord with observation. 



With regard to waves of slightly different lengths, focussed 

 at b\ each is there superposed on a wave of different length 

 from its own and appreciable interference ceases for this 

 reason. If the slit is widened the phenomenon (with white 

 light) also vanishes by overlapping. The case of the screen 

 with two spaces has already been treated in relation to fig. 12 

 and corresponds to the diffraction of a rod. 



This seems to be an approach to the real nature of the phe- 

 nomenon, though it does not give an adequate theory for it. 

 The actual difference of wave length within which interference 

 can take place, is naturally within the spectrum breadth of the 

 diffraction pattern, h\ = 2XlO" 8 cm. observed, or within about 

 5XlO _9 cm. 



To return finally to the original inference, it appears that 

 beating wave trains have not been observed, but that the strik- 

 ing scintillations are due to an exceptional susceptibility of the 

 apparatus to laboratory tremors, when exhibiting the phenome- 

 non in question. What has certainly been observed is the 

 interference of a D 1 or D 2 line with a reversed D' or Z>/ line, 

 both having the same source and longitudinal axis. One can 

 only assert, therefore, that light of the wave length interval of 

 the breadth of these lines is capable of interference, when the 

 line is reversed. The phenomena, as a whole, are diffractions 

 of symmetrical half wave fronts, each of which may be sepa- 

 rately controlled by the corresponding micrometers. 



Brown University, Providence, R. I. 



