Lightning, according to the Greek philosopher Empedocles in the 5th century BC, consists of rays of the sun that are intercepted by clouds. Anaxagoras, at about the same time, said it was the gleam of the fiery upper ether which had descended from above, and thunder was the hissing noise of its being extinguished by the cloud. Aristotle, a hundred years later, produced a remarkably perceptive objection to this theory which, he said, "involves the view that lightning is actually prior to thunder and does not merely appear to be so".
Aristotle had realised that light travels faster than sound - though the way he put it was to say that sight is quicker than hearing. "The rowing of triremes illustrates this," he wrote. "The oars are going back again before the sound of their striking the water reaches us."
He believed that thunder and lightning were part of a "dry exhalation" from clouds. "If any of the dry exhalation is caught in the process as the air cools, it is squeezed out as the clouds contract, and collides in its rapid course with the neighbouring clouds, and the sound of this collision is what we call thunder." Lightning is the burning of this same exhalation.
The discovery of electric charges more than 2,000 years later rather spoilt this elegant theory, leaving in its place a highly complex, and still incomplete picture of what really happens within a thundercloud.
Between about 70km and 3,000km above the earth's surface is the ionosphere, which is positively charged because of the action of solar ultraviolet radiation. The earth's surface, however, is negatively charged during fine weather, so cloud droplets acquire an induced positive charge on their lower sides and a negative charge on their upper sides. This undoubtedly plays a part in the electrification of a cloud, but a far greater contribution is probably made by collisions between ice crystals and warmer pellets of hail.
Generally, the higher parts of a cloud will hold positively charged ice crystals, the middle parts will hold a mixture of negatively charged hail and positively charged raindrops caught in an updrift, and the lowest part will have falling hail or raindrops, both negative. But an area of positive charge observed at the bottom of clouds is still a mystery.
Lightning should be thought of as a spark between not-quite-connected parts of an electrical circuit. It usually starts just before the storm hits the ground, as the negatively charged melting hail comes close enough to the positively charged ground for the circuit to be bridged with a spark. The first stroke comes from cloud to ground, followed almost instantaneously by a return stroke from the ground. A simple path of least resistance between cloud and ground results in sheet lightning, but when the flash takes several paths from a main channel, we see the more attractive forked lightning. The extremely rare ball lightning, which appears as a bright sphere less than a foot across, and bead lightning, which appears as a succession of small balls, have so far evaded explanation.
The difference in electrical potential bridged by a lightning flash is usually hundeds of millions of volts - and that explains why lightning flashes generate such heat. The heat causes a rapid expansion of the air around it, and the resulting compression wave produces thunder as it breaks the sound barrier. Since sound waves travel at about one- third of a kilometre per second, count the seconds between lightning (which is seen almost instantaneously) and the thunder that follows, then divide by three to determine how far away the lightning flash occurred in kilometres.