The great Galileo Galilei (1564-1642) tried to measure the velocity of light by means of lantern signals between mountain tops. Naturally he failed. Light travels about 186,270 miles (more than seven times the circumference of Earth) in one second. In modern physics, light is regarded as the fastest thing in the universe, and its velocity in empty space as a fundamental constant of nature.
First good figure for light's speed was obtained in 1676 by a Danish astronomer, Ole Roemer, who measured the variations in eclipse times of Jupiter's satellites according to Earth's distance from that planet. His calculation was only about 3% too high. First terrestrial measurement was made in 1849 by Armand Hippolyte Louis Fizeau of France, who passed a beam of light through the teeth of a spinning cogwheel. The light struck a mirror, bounced back to the wheel. The wheel had been timed to move just enough in the brief interim for the teeth of the wheel to intercept the light as it was reflected. By timing the revolutions per second of the cogwheel and measuring the distance to the mirror, it was easy to calculate the speed of the light.
The late great Albert Abraham Michelson, in his final experiments, reflected light back & forth ten times in a mile-long vacuum tube from the faces of a rapidly spinning, 3 2-sided mirror. Velocity measurements completed by his successors after Michelson's death yielded an average figure of 186,270.75 miles per second. But in individual runs there were unexplained, periodic variations up to twelve miles a second. At first this caused excitement over possibility that the speed of light might not be constant (TIME, Dec. 25, 1933). The clamor was quieted by attributing the variations to "experimental error." So the velocity of light was re-established as a constant in good standing.
Last week Harvard University announced development of a new apparatus for refining measurements of light's speed still further. It is compact enough to be housed in a small laboratory room and hallway, it eliminates friction as a source of error, and the measurement is automatic—that is, the human eye is not a factor (the Michelson crew aimed their beams by eye) and the clocking is done, in effect, by a photoelectric cell.
The principle is that of cutting a light beam up into a certain number of sections per second, then measuring the length of one section. This is like clocking a freight train when you know the length of the cars. If the cars are 30 feet long and you see that two of them pass a given point every second, you know the speed is 60 feet per second.
The train of light emitted from a 1,000-watt lamp is "sectioned" by a formidable-looking device called a standard frequency generator (see cut), also developed at Harvard, which alternately brightens and dims the beam 19,200,000 times a second. This is like nicking at regular but very close intervals a cable which is rapidly being paid off a drum. The light beam is split. One part is conducted over a long course (185 yd.), the other over a short course (about 2 yd.). Both are reflected back to a photoelectric cell. On the beam which has been over the long course the brightness peaks (nicks) lag somewhat behind those on the other. From the amount of this lag the length of one section is calculated.