476 
PHYSICS: A. A. MICHELSON 
Proc. N. a. S. 
A MODIFICATION OF THE REVOLVING MIRROR METHOD 
FOR MEASURING THE VELOCITY OF LIGHT 
By a. a. Michelson 
Ryerson Physical Laboratory, University of Chicago 
Read before the Academy, April 26, 1920 
The accuracy of any measurement of the velocity of Hght by the re- 
volving mirror depends on that of three factors : 
(1) The distance between stations. 
(2) The speed of rotation. 
(3) The angle described by the revolving mirror while light goes to the 
distant station and returns. 
It is estimated that (1) and (2) may be determined to within one part 
in 200,000 and possibly one in a million. 
The third measurement may be eliminated by constructing a revolving 
mirror in the form of a prism with a square or octagonal base, the angles 
of which may be made equal to 90° (or 135°) to this same order of 
accuracy. 
For this a test angle is made by cementing a prism of approximately 
90° (or of 45°) to a true plane. The square (or octagon) is applied to the 
two surfaces, producing interference bands when illuminated by the light 
from a mercury lamp. 
If these are straight and parallel to the axis of rotation, the angles are 
•equal; if not the difference, a, is measured by e/26, if e is the fraction of 
the fringe width and h the width of the mirror face. Thus if 6 = X/20 
and 6 = 15 mm., a = 0.000001, which will, therefore, be the order of 
accuracy of the measurement under these conditions. For a wider face 
the accuracy would be still greater. If the angle of the test prism is 
and the measured differences are ao, ai . . . then completing 
the 360° gives Bo = (360-2Q;)/n, whence d, = do + au ^2 = + ^2, 
€tc. 
The faces are corrected until the angles are all equal within the limit 
of error assigned. This method of constructing reflecting surfaces in- 
clined at equal angles, which may readily be extended to any number of 
surfaces, may prove useful in the construction of standard test angles, 
which may thus be made accurate to a tenth of a second or less; and also 
in the subdivision of meridian circles, etc., in which the limit of error is of 
the order of one or two-tenths of a second. 
Such a revolving mirror has been constructed in form of an octagon with 
faces of 15 by 20 mm. and with air under pressure of one atmosphere, 
