Hope I'm not jumping in here but I have read this thread with interest, some good points raised/questions asked.
It does strike me though that there is a fine line between adaption and evolution. As a hugly expensive/unethical/never-going-to-happen experiment a way to look at this may be to take a sample of non-runners who have no history of running in family. Train them up as hard as possilble over a numbr of years, recording all their progress - This would be the adaption data. Then, and this is the really unethical bit, couple them off and train up their kids following exactly the same regimes.... (!) If you repeated this process a few times you would be able to look at genetic ability/adaption too.
Nuts I know.
But seriously, there has to be a limit on the ability to reach a certain performace level by a certain individual, otherwise "Olympic Athlete" would be a viable career choice for everybody - imagine the careers advisor at school saying, "Well, as long as you put the hours in to train then you can be one."
But defining what these limits/reasons/factors are is an almost imposible task IMO. Hence the silly "experiment" example above.
I don't agree however, that the african nations are genetically more adapted to distance running, I think that it is the case as has already been raised that they are conditioned this way from an earlier age, and in a lot of ways have more drive to achieve the desired results (at the risk of stereotyping) escape from povery and lack of oportunitys. And the role-model issue as raised above. Yes they are good, no argument there but there are enough non-africans at the top to prove that it isn't all about genetics, (Radcdcliff, Motram, Jelena Prokopcuka etc al).
FIN
Talking of unethical experiments - did anyone see the documentary on that poor Indian kid who was being made to run crazy mileages each week by his "coach"? Apparantly he has "a natural ability" according to his "coach".