89 



the purpose of a standard, it became evident that the interfer- 

 ence method might l)e made to yield infomiation concerning the 

 distribution of hght in an approximately homogeneous source 

 when such observations would be entirely beyond the power of 

 the best spectroscopes. To illustrate, suppose this source to 

 be again the double radiation from sodium vapor. As the wave 

 lengths of these two radiations differ by about one part in a 

 thousand, then at a difference of path of five hundred waves 

 (about 0.36 mm.) the bright fringes of one v/ave-train would 

 cover the dark fringes of die other, so that if the two radia- 

 tions were of equal intensity all traces of interference would 

 vanish. At twice this distance they would reappear, and so on 

 indefinitely, if the separate radiations were absolutely homo- 

 geneous. As this is not the case, however, there would be a 

 gradual falling off in the clearness or visibility of the bands. 

 Inversely, if such changes are observed in actual experiment, 

 we infer that we are dealing with a double source. Further, 

 from the distance between the maxima of distinctions, we may 

 determine (and with extraordinary accuracy) the ratio of wave 

 lengths of the components ; from the ratio of maxima to minima 

 we may infer the ratio of their intensities ; and, finally, the grad- 

 ual falling off' when the distance becomes large gives accurate 

 information of the "width" of the corresponding spectral lines. 

 In this way it was found that the red line of hydrogen is a 

 double with components about one-fortieth of the distance apart 

 of the sodium lines. Thallium has a brilliant green radiation 

 which is also double, the distance being one-sixtieth that of the 

 sodium lines. Mercury shows a brilliant green line, which is 

 highly complex, but whose chief component is a doublet, whose 

 separation is only one seven-hundredth of that of sodium. The 

 interference fringes are still visible when the difference of path 

 is of the order of five hundred millimetrs, corrsponding to 



