I wonder if diet plays a role. The only way you're getting 20 miles' worth of glycogen is from eating grains- which were domesticated just recently in human history. I'm not sure if that type of glycogen density was available before then. Maybe now we have the luxury of burning up 5kcals for a very long jog, but 100,000 years ago?
Diet almost certainly plays a role. The diet of the Tarahumara Indians (the subject of the 'Born to Run' book mentioned in the article) is practically meatless and consists of about 75% corn, with the rest made up mostly of beans and various types of squash. As a result their diet is about 80% complex carbohydrates, which is what allows them to run for those insane distances.
This doesn't seem to jive with the persistence hunting theory, since meat is one of the worst fuels for distance running (or any endurance sport).
