87 



and dark, according to the relative retardation of the two re- 

 flected light -beams. 



If this retardation change (by slowly increasing the distance 

 between the surfaces (, the center of the ring system goes through 

 alternations of light and darkness, the number of these alterna- 

 tions corresponding exactly to the number of light-waves in 

 twice the increase in distance. Hence the measurement of the 

 length of the waves of any monochromatic light may be obtained 

 by counting the number of such alternations in a given distance. 

 Such measurements of wave lengths constitutes one of the most 

 important objects of spectroscopic research. 



Another object accomplished by such measurement is the es- 

 tablishment of a natural standard of length in place of the arbi- 

 trary standard at present in use — the meter. Originally it was 

 intended this should be the ten-millionth part of an earth-quad- 

 rant, but it was found that the results of measurements differed 

 so much that this definition was abandoned. The proposition 

 to make the ultimate standard the length of a pendulum which 

 vibrates seconds at Paris met with a similar fate. 



Shortly after the excellent gratings made by Rutherford ap- 

 peared, it was proposed (by Dr. B. A. Gould) to make the length 

 of a wave of sodium light the ultimate standard; but this idea 

 was never carried out. It can be shown also that it is not sus- 

 ceptible of the requisite degree of accuracy, and in fact a num- 

 ber of measurements made with a Rowland grating have been 

 shown to be in error by about one part in thirty thousand. But 

 modern conditions require a much higher degree of accuracy. 

 In fact, it is doubtful if any natural standard could replace the 

 arbitrary standard meter, unless it can be shown that it admits 

 of realization in the shape of a material standard which can not 

 be distinguished from the original. 



One of the most serious difficulties encountered in the at- 



