# It's official: light travels exactly 299,762,458 meters per second

It's official now. The speed of light is exactly 299,762,458 meters per second. There's no ambiguity about it, none of those ''error bars'' with which scientists indicate the uncertainties inherent in every measurement. The value of this fundamental natural constant has been arbitrarily set by international fiat.

Yet in adopting a fixed value for the speed of light, the General Conference on Weights and Measures is not trying to preempt nature. Instead, on Oct. 20, it gave the world a more accurate standard of length, in which the meter now is derived from the color of a laser beam with the help of the agreed-upon light speed.

In doing this, the conference has taken a giant step toward simplifying our system of weights and measures so that at least some of the basic units are based directly on time.

The meter now is officially defined as the distance light travels, in vacuum, in the incredibly short time span of one second divided by 299,762,458. Thus the units of length used to describe the height of a mountain or the reach of your forearm are tied directly to the second. Someday the kilogram, for example, may also be based on time.

Astronomers have long been accustomed to using time to define a unit of length with the help of the speed of light. The light-year, by which they specify the distance to a star, is the distance light travels in a year. But that has not hitherto been the case for the standard basic units (such as meter, kilogram, second, or ampere) which, by international agreement, underlie all other systems of measurement. Even the light-year is referenced ultimately to the standard meter.

Up to now, that standard has basically been a length. It once was literally the distance between two marks on a platinum-iridium bar kept by the International Bureau of Weights and Measures in Paris. By 1960, however, this had become much too crude for the precision measurements of physicists and astronomers. So the meter was redefined to be 1,650,763.73 wavelengths of the orange-red light emitted by a krypton-86 lamp (krypton-86 being one of several forms of that element). Now this standard has also become too imprecise for the needs of science.

This search for ever-greater precision is what has driven metro-logists to base the unit of length on the second rather than on an actual physical distance. ''The main reason for doing this is that the second is the most accurate of all the base units,'' explains Kenneth W. Evenson of the US National Bureau of Standards (NBS). It can be measured to better than one part in 10,000 billion. The krypton meter was accurate to about 4 parts in a billion.

The key tool in tying the meter to the second is the laser. The wavelength of light is related mathematically to its frequency - wavelength is just the speed of light divided by frequency. And frequency, Evenson says, can be measured 1, 000 to 10,000 times as accurately as can wavelength.

A laser provides the kind of pure, stable light source needed for this precision work. Then, with the use of a mirror, this light can be made to interact with itself to produce a characteristic pattern of bright and dark lines called fringes. The spacing of these fringes is directly related to the wavelength of the light. And since wavelength is accurately calculated from the light's frequency, a metrologist has only to count the appropriate number of fringes - that is, the appropriate number of wavelengths - to lay out a standard meter, Evenson explains.

Since the measurement of frequency is directly linked to the measurement of time, the accuracy of frequency measurements, and hence the precision of the standard meter, now is tied directly to the precision of the atomic clock, the most precise of all present standards for weights and measures.

The work of Evenson and his colleagues at the NBS laboratory in Boulder, Colo., helped encourage the General Conference on Weights and Measures to adopt the new definition of the meter. Already, he says, the standard meter is 10 times as accurate as when it was based on the wavelength of krypton light. He adds that there should be little difficulty in gaining another tenfold improvement in accuracy.

But what of other units? Is it realistic to try to base the kilogram on time measurements, for example, now that this has been done for the meter? Evenson says he can't imagine how to do this for degrees of temperature or amperes of electric current. But he says it may be possible to do it for the kilogram by measuring the distance between atoms in crystals of silicon. If the number of atoms in a standard volume of such a crystal could be accurately measured, this could be a way of defining mass (that is the kilogram) in terms of length. Thus, Evenson says, with the meter and kilogram tied to the second, ''you could do a fairly good job of reducing the number of base units in terms of the second.''