According to my field guide, Osage-orange has a limited natural range in the Red River region of east-central Texas, southeastern Oklahoma, and adjacent Arkansas. Indians used to travel hundreds of miles for the wood, prized as the finest for making bows. Then European settlers planted it widely as living fences, taking advantage of the tree’s ability to spread via shoots from lateral roots. But Osage-orange persisted, and became widely naturalized long after the invention of barbed wire rendered them useless to farmers. The tree can now be found in 39 states and Ontario. If Osage-orange does so well elsewhere, why was it restricted to such a small area?

The answer likely lies in the disappearance of its primary disperser. Without mammoths, groundsloths, and other megafauna to transport its seeds uphill, the range of the species gradually shrank to the Red River region. In fact, fossils tell us that Osage-orange was much more widespread and diverse before the megafaunal extinctions. Back then, Osage-oranges could be found north up to Ontario, and there were seven, not just one, species in the Osage-orange genus, Maclura.

Another anachronistic tree is the Kentucky coffeetree, so named because early Kentucky settlers used its beans as a coffee substitute. Coffeetrees have tough, leathery pods with large, toxic seeds surrounded by a sweet pulp. Water cannot penetrate the thick seed coat to begin germination unless it is abraded or cut. Sounds like mammoth food to me. The natural range of coffeetrees is concentrated in the Midwest, but without its megafauna disperser, it is generally rare and mostly limited to floodplains.

The increasing disappearance of young calvaria trees, is suggested due to Cylindraspis tortoises, fruit bats or the Broad-billed Parrot could have been dispersing the seeds.

http://en.wikipedia.org/wiki/Mascarene_Islands