758 BELL SYSTEM TECHNICAL JOURNAL 



into motion, the fringes march off sidewise, growing more curved as 

 they go; and the number passing any fixed marker set up in the field 

 of vision is the double of the number of wave-lengths comprised in the 

 distance through which the mirror moves. 



(^onsider now for a moment what must be observed, if wave-trains 

 of many wave-lengths fall upon the mirrors, instead of pure mono- 

 chromatic light. Each kind of light produces its own pattern of 

 fringes; but since the breadth of a fringe depends upon the wave- 

 length, those of one kind cannot fall perfectly — light upon light and 

 shade upon shade — upon those of another; and in most parts of the 

 field of view, if not in all, the various patterns must blot one another 

 out. Yet there is one exception; returning to equation (18) one sees 

 that for any value of r the phase-difiference <p must vary with X, 

 unless t = — in which exceptional case ip = <po = 180° whatever the 

 wave-length.^^ \\'hen the real mirror and the virtual one coincide 

 perfectly, the field of view is black, however many wave-lengths are 

 contained in the incident light; and when the real mirror and the 

 virtual one intersect, the line of intersection is marked with a black 

 fringe. Moreover, in the neighborhood of the central fringe there is 

 a brilliant display of colors. Words are too feeble to describe them, 

 but there would be no great scientific advantage to be gained from a 

 description; for the tint observed in any particular direction is not a 

 pure prismatic hue, but results from mixture of the wave-lengths not 

 completely extinguished by interference in that direction, and depends 

 therefore upon the physiology of the eye. What interests us as 

 physicists is the service of the central black fringe and the com- 

 panioning glory of colors in marking the point and moment when the 

 real and the virtual mirrors intersect or coincide. This service was 

 essential in the measuring of the standard metre, a process which we 

 will now examine. 



The interferometer used n the process is sketched as seen from 

 above in Fig. 5. The mirrors M and M' are at the two extremities of 

 the intermediate standard; they are made parallel with the greatest 

 possible exactness, and the further is elevated above the level of the 

 nearer. As for the other mirrors, D is the movable "reference" 

 reflector; the purpose of N will appear directly; N' may be ignored. 



The instrument is now so adjusted that I) is strictly parallel to the 

 virtual image of N, and intersects that of M at a small angle. Red 

 cadmium light being used, a part of the field of view is occupied by the 

 circular fringes due to the cooperation of D and N and part l)y the 



'' Also <p is indciKMidcnt of X if cos r = 0, a case which may occur if we have a 

 stratum of air between glass plates, and choose an angle of incidence near the angle 

 of total reflection. 



