THE SPEED OF LIGHT IS INDEPENDENT OF FREQUENCY within a factor of 6x10^-21. Bradley Schaefer of Yale (203-432-3806, schaefer@grb2.physics.yale.edu) bases this estimate on the observed arrival of gamma rays from distant explosive events in the cosmos, such as gamma-ray bursters. If the speed of light (c) were slightly different for the different frequency ranges, then some light waves would show up before the others, but this is not the case. The best previous effort to locate a frequency dependency for c, deduced from light coming from the Crab pulsar, was at the 5x10^-17 level. Why would c vary with frequency? Einstein's theory of relativity, and its insistence on a universal light speed, might be at fault. Or photons might have mass. Schaefer*s analysis addresses this issue, and puts an upper limit of 10^-44 g on any putative photon mass, not quite as sharp a limit as those based on the observed strength of the galactic magnetic field (a nonzero photon mass would allow the fields to decay away). The new sharper limits on any possible frequency-dependency for c is a vindication of relativity. By the way, the prefix for anything as small as 10^-21 is "zepto" (Shaefer, Physical Review Letters, 21 June; journalists can obtain the article from AIP.)