The dreams of warp drive and faster-than-light communications will have to remain just that. According to the scientists who have demonstrated superluminal group velocity, the reports of the death of special relativity have been greatly exaggerated. The group has found that the superluminal speeds it demonstrated in the lab can be explained with theories in accordance with relativity and are unlikely to apply to information carried by light. In its report in the July 20 issue of Nature, the team at NEC Research Institute Inc. describes how it used laser pumping to excite cesium gas in a 6-cm-long, paraffin-coated glass cell to one of its 16 possible quantum mechanical states. A near-Gaussian, 3.7-µs probe pulse from a diode laser was injected into the cell and detected at the far side with a high-sensitivity avalanche photodiode. The researchers found that the peak of the light pulse emerged from the cell 62 ns earlier than it would have had it traveled the same distance through a vacuum. Effectively, it took negative time to make the journey, showing a pulse advance 310 times the vacuum transit time. They attribute the effect to interference of the frequency components of the pulse as it passes through the gas. The excited cesium gas is a so-called anomalous dispersion medium that modifies light waves in such a way that, as shorter wavelengths enter the cell, they are elongated and longer wavelengths are shortened. When a wave exits the medium, its original wavelength is restored. Considering a pulse as a number of waves in phase, the effect of the gas is to bring the light's components back into phase ahead of where they would have been had they passed through a normal dispersion region or a vacuum. Lijun Wang, a research scientist at the institute and lead author of the report, said his team's rough calculations show that the phenomenon seen in the smooth pulses does not apply to light conveying information. Citing Leon Brillouin's assertion that the speed of information should be understood as the frontal velocity of a step function on a pulse, the researchers believe that it may accelerate information transfer speeds up to c, but not beyond it. Wang said that his group, like that of Raymond Y. Chiao of the University of California in Berkeley in 1998, has demonstrated the effect in an electronic system, where typical speeds are far less than c. Because the superluminal effect can be understood in terms of the wave nature of light, Wang stressed that it is not at odds with special relativity. "It is not our intention to claim that we have defeated Einstein's theory," he said. "Some members of the popular press have reported that we are rebels tearing down relativity, but this is not the case. The results are consistent with relativity." The team plans to perform a similar experiment, using a small number of photons to see what effects anomalous dispersion will have. "We suspect that there is going to be additional noise," Wang said. "This will most likely reduce the signal-to-noise ratio, preventing a real signal being sent faster than c.