According to the dates you gave earlier, the downsizing of the Channel Island pygmy mammoths took place over 100,000 years. I saw the figure of 2,000,000 years given for the Mediterranean dwarf elephants and wondered if there were intermediate sized fossils there to show the progression.
If modern humans came out of Africa 100,000 or 200,000 years ago, there would appear to be ample time for people to have developed various genetic differences (e.g., skin color) that gave some small advantage to their possessors. It is too bad that humans didn't invent refined sugar 100,000 years ago, so that we could now eat as much as we want without suffering adverse effects.
I wonder about the origin of human pygmies. Apparently the negroid pygmy found in Africa is also found on islands in the Far East, but I have no idea whether their small size comes from their ancestors having been located on limited-resource islands in the past. The mainland pygmies haven't had time to revert to normal size or else there is insufficient driving force for it to occur quickly.
A few years ago I was one of the organizers of Pacific Section Joint Convention of the American Association of Petroleum Geologists, Society of Economic Paleontologists and Mineralogists, and the Society of Petroleum Engineers. (What a mouthful - that's why they use acronyms!) The late Professor Pete Weigand at Cal State Northridge edited one of the two volumes from the convention which included a good short article by Dr. Larry Agenbroad, entitled "New Pygmy Mammoth (Mammuthus exilis) Localities and Radiocarbon Dates from San Miguel, Santa Rosa, and Santa Cruz Islands, California." The radiocarbon dates for Agenbroad's spectacular M. exilis discovery were reported at 12,840 +/- 410 B.P. The importance of this 1994 discovery was that the skelton war nearly complete and that it was clearly in a "primary depositional context." These were also the last of the mammoths on the islands. Scattered remains are probably re-worked, so accurate dates for the fossil can not be obtained from the surrounding strata, but only from the fossil itself.
One recently discovered site yielded an "intermediate form" (somewhat larger than the other M. exilis specimens) dated at 47,110 BP. Uranium-thorium dates have been acquired on mammoth teeth from Santa Rosa Island report 68,000 +/- 4,000 BP, 64,500 +/- 7,400 BP, and 29,000 +/- 2,000 BP. However, these dates are suspect. I once suggested using the amino acid racimization dating technique that had been used on marine mollusks, but who listens to me?