WOOD. — ACHROMATIZATION OF INTERFERENCE FRINGES. 599 



wide ranges of wave-lengths ou opposite sides of the absorption-band. 

 The fringe sliifts of the two spectral regions will be in opposite direction.-^, 

 while the diifts of the jioints of maximuni visibility will be in the same 

 direction. It appeared as if this might increase the width of the region 

 over which fringes could be observed^ for the red-orange fringes are 

 shifted in one direction and the yellow-green in the o{)posite. Each set 

 would be more or less perfectly achromatized, and in the region in which 

 they overlapped we should expect a periodic visibility, owing to the dif- 

 ference in the widths of the fringes of the two systems. 



To test the point it seemed best to work with a narrow range of the 

 spectrum symmetrical about the D lines. This was obtained by opening 

 the slit of the monochromatic illuminator, bisecting it with a wire, and ad- 



FlGURE 3. 



justing the prisms so that the region of the D lines was screened off by the 

 wire. By means of a small screen either of the two narrow portions of 

 the spectrum bordering the D lines could be screened off. 



The effect of the sodium vapor on the fringes formed when the inter- 

 ferometer was illumined by either one or both of the two portions of the 

 spectrum could then be studied at leisure. 



It was found that when a considerable amount of the vapor was 

 present the apparent centre of the greenish-yellow fringe system was 

 widely separated from the centre of the orange-yellow system. 



When both sorts of light were used at once there was a periodic visi- 

 bility in the region in which the two systems overlapped, the. appearance 

 in the three cases bein^ shown in Figure 3. 



The case is a little more complicated when white light, or the entire 

 spectrum, is used, but it does not differ materially from the special case 

 just considered. 



