ee 
280 W. LeConte Stevens—frecent Progress in Optics. 
has for convenience been called a single component. Each of 
the two sodium lines is itself a double whose components are 
separated by an interval about one-hundredth of that between 
the long known main components; and an interval yet less 
than one-fifth of this has been detected between some of the 
components of the green line of mereury. Indeed Michelson 
deems it quite possible to detect a variation of wave length 
corresponding to as little as one ten-thousandth of the interval 
between the two main sodium lines.* 
This new found complexity of radiation, previously thought 
to be approximately if not quite simple, proved to be a tempo- 
rary barrier to the accomplishment of the plan of using a light 
wave as a standard of length. It necessitated careful study of 
all those chemical elements which give bright lines that had 
been supposed to be simple. The red line of cadmium has © 
been found the simplest of all those yet examined. ‘The vapor 
in a rarefied state is held in a vacuum tube through which the 
electric spark is passed, and under this condition the difference 
of path for the interfering beams in the refractometer may be 
a number of centimeters. A short intermediate standard, 
furnished with a mirror at each end, is now introduced into 
the comparer and moved by means of the micrometer screw. 
Its length is thus measured in terms of the cadmium wave 
length. A series of intermediate standards of which the 
second is double the first, the third double the second, ete., are 
thus compared, and finally in this way the value of the meter 
is reached. 
The feasibility of this ingenious method having been made 
apparent, Michelson was honored with an invitation from the 
International Bureau of Weights and Measures to carry out 
the measurement at the observatory near Paris with the col- 
laboration of the director, M. Benoit. After many months of 
labor, results of extraordinary accuracy were attained. For 
the red line of cadmium at an air temperature of 15° C. and 
pressure of 760™", two wholly independent determinations 
were made. From the first a meter was found equal to 
1553162°7 wave lengths; from the. second, 15531643 wave 
lengths,—giving a mean of 1553163°5, the deviation of each 
result from the mean being very nearly one part in two 
millions. A determination by Benoit’ from the first series 
gave 1553163°6, which differs but one-tenth of a wave length 
from the mean of Michelson’s measurements. 
The direct comparison of the lengths of two meter bars, 
though not easy, is a simple operation in comparison with the 
* Astronomy and Astrophysics, February, 1894, p. 100. 
+ Travaux et Mémoires du Bureau internationale des Poids et Mésures, vol. xi, 
p. 84, 1894, 
