a Light-wave as Ultimate Standard of Length. 1 83 



corresponding to !Na 2 a correction of one-third the phase-dif- 

 ference between the two was added. The sodium flame gave 

 good interference rings with E the longest standard used ; but 

 when longer standards are used it will be better to use the 

 vacuum tube. 



The lithium flame gives very indistinct interference circles 

 but is valuable as a check on account of its great wave-length. 

 . The yellow mercury light has the inconvenience of being 

 double ; but the two radiations are of nearly equal intensities. 

 "When, however, the phase-difference amounts to nearly half a 

 wave it may be necessary to add or subtract half a wave from 

 the observed result. 



The brilliant green line gives beautifully clear circles even 

 with a difference of path of half a million waves, so that in all 

 probability this will be the wave to be tised as the ultimate 

 standard of length. So far as mere errors of setting are con- 

 cerned, it is easy to obtain results consistent to a fiftieth of a 

 wave, and with care to a hundredth. 



The violet is also perfectly clear but the light much fainter. 

 Its short wave-length makes it very valuable as a check. 



As has been elsewhere shown the refractometer gives precisely 

 the same interference phenomena as two surfaces enclosing a 

 plate of air. One of these surfaces corresponds to one of the 

 mirrors on the standard ; the other to the image of another 

 mirror. This image will be called the "reference plane." 



The standard, A, let us say, which has already been carefully 

 adjusted for length and parallelism is placed in the refractom- 

 eter and the distance of the reference plane adjusted till the 

 interference rings are equally distinct on the upper (rear) and 

 the lower (front) surfaces of A. Then the reference plane is 

 made to bisect the (extremely small) angle between the upper 

 and lower surfaces. This is accomplished by adjusting till there 

 is no alteration in the diameter of the rings on either surface 

 when the eye is moved vertically or laterally (if the angle is 

 zero) ; or till the alteration is equal and opposite on the two 

 surfaces. 



This adjustment is effected by means of the sodium flame, 

 without the observing telescope. Temperature and barometric 

 pressure are then taken. The reference plane is then moved 

 by a fine adjustment till the central spot is black (phase =zero) 

 on the upper (mirror) and the diameter of the first ring- 

 measured by the Troughton micrometer ; and the telescope is 

 then lowered and the width of the first ring on the lower sur- 

 face measured. Then the phase difference at the lower surface 

 is adjusted to zero, and diameter of first ring measured ; and 

 then the telescope is raised and width of first ring on upper 

 mirror is measured. 



