Ball lightning - a slow-moving ball of light about 30 centimetres across that is occasionally seen at ground level during thunderstorms - has puzzled scientists for centuries. Benjamin Franklin, for instance, made several attempts to observe this mysterious phenomenon over 200 years ago. In one attempt he tried to make ball lightning from ordinary lightning - inspired by reports of a Russian contemporary who had been killed by such a ball after directing a lightning strike into his laboratory.
So what progress have we made in understanding ball lightning? It has been especially difficult to explain the longevity of the ball (which ranges from several seconds to several minutes), its stability and the way it releases energy.
Two years ago James Dinniss and I proposed a model of ball lightning based on the oxidation of silicon nanoparticles in the atmosphere following a lightning strike (Nature 2000 403 519). We proposed that the silicon nanoparticles were formed as a result of the reaction of silicon oxides and carbon in the soil: at the high temperatures created by the lightning strike, the carbon in the soil chemically reduces the silicon oxides to the metallic form of silicon. Laboratory experiments confirmed that nanoparticles containing silicon could be produced in this way.
In the April issue of Physics World, John Abrahamson of the University of Canterbury, New Zealand, surveys the current theories of this weird phenomenon.