Studiot said:
Surely it's pretty obvious that in intercloud lightning one cloud must be positive, the other negative?
Positive charges usually collect in the upper regions of a thundercloud, and negative in the lower. There may be a small amount of positive charge at the very bottom.
Cloud lightning is any lightning that does not connect to earth, and constitutes the majority of all lightning. Cloud discharges can be subdivided into intracloud, intercloud, and cloud-to-air flashes. These are not usually studied because they are not of great practical interest - do not cause forest fires, power outages or ill effects of cloud-to-ground lightning - and are not easily studied. They are of interest to aircraft designers. But investigators sometimes differ as there may be a number of phenomena at work here.
Intracloud lightning discharges typically occur between positive and negative cloud charges, and takes place over about the same duration as cloud-to-ground discharges, about half a second. A typical cloud discharge transfers tens of coulombs of charge over a spatial extent of 5-10km. It's thought to consist of a continuously propagating leader that generates weak return strokes called recoil streamers, with which are associated electric fields termed K-changes, when the leader contacts pockets of charge of space charge opposite its own. The cloud K-changes are similar, but usually of opposite polarity, to the K-changes associated with K-processes that occur in the interval between return strokes in ground discharges.
In some cases, flashes that are produced primarily within the cloud, and are best characterized as cloud flashes, produce, seemingly as an unimportant by-produce, a channel to ground.
All the above information comes from Martin Uman's book, "The Lightning Discharge". I apologize if all this is more than you really wanted to know about lightning. Me, I'm buggy over it.
Respectfully submitted,
Steve