Originally posted by jj
I think another explanation that would seem to have some basis in fossil records is that when an organism evolves something new (say for instance primitive vascularization in plants), the new species have so much more viability that the rapid spread looks instantaneous.
Every organism has a range in which conditions favor its survival, and populations tend to become stable within that range. The type of 'sweep' of a gene pool you suggest depends not only on a mutation producing a significant advantage, but also on it producing a relatively drastic change -- this is highly disfavored probabilistically.
Niles Eldrige explained this very well in
Reinventing Darwin: The Great Debate at the High Table of Evolutionary Theory.
As you approach the outer edges of an organism's range, conditions generally become less favorable, until you reach a point at which any individuals able to survive at all are unlikely to reproduce successfully. Even if some mutation occured which would better equip offspring for survival at the edge of the range (or even beyond), that mutation would (by definition) make the offspring
less suited to life nearer the middle of the range, so it can't take hold unless a sub-population becomes reproductively isolated from the main population. Evolving independently of the main population, such a daughter population might then chance on something that would be a significant advantage back in the home range.
Another way of looking at it is that the main population had reached an adaptive peak, but it wasn't the highest one around. Reaching the higher peak meant backing up so as to permit passage through a fitness valley, but backing up is an evolutionary no-no. The sub-population, though,
started in a fitness valley (being under-adapted to their environment, which was
outside the ancestoral range) and climbed a different peak (which turned out to be higher).
A lot of the debate around PE seems to have centered on whether it is anything all that exciting a new; such allopatric speciation is easily handled by mainstream Darwinism.