Improvements in the apparatus formerly used include the use of an eleven-stage electron multiplier tube, the control of the transmitter frequency by the laboratory standard, the substitution of a 1000-watt water-cooled mercury arc for the light source, a new type of Kerr cell, changes in the mirror supports and base line measurements, and the use of an automatic recorder. The apparatus has been completely rebuilt and converted to a.c. operation. The optical system has been changed to permit simpler measurements of the path difference, and electrical circuits have been devised to smooth out the fluctuations due to voltage variations and other causes.Group velocity is discussed as a correction factor in this and previous measurements. The correction is shown to amount to as much as 7 km/sec. in some cases. Electron transit time is shown to be a limiting factor for this method of measuring the velocity of light. The final result of 2895 observations is given as 299,776±14 km/sec. This includes a group velocity correction and should not be compared with previous results without taking this into consideration. The conclusion is reached that the velocity of light is a constant as nearly as we can measure it at present.
WILMER C. ANDERSON, "Final Measurements of the Velocity of Light," J. Opt. Soc. Am. 31, 187-197 (1941)