Among the startling discoveries about Clovis people is the speed of their spread. All Clovis sites in the US dated by the most advanced radiocarbon techniques were occupied for only a few centuries, in the periodjust before 11,000 years ago. A human site even at the southern tip of Patagonia is dated at about 10,500 years ago. Thus, within about a millenium of emerging from the ice-free corridor at Edmonton, humans had spread from coast to coast and over the entire length of the New World.

Equally startling is the rapid transformation of Clovis culture. Around 11,000 years ago Clovis points are abruptly replaced by a smaller, more finely made model now known as Folsom points (after a site near Folsom, New Mexico, where they were first identified). The Folsom points are often found associated with bones of an extinct wide-horned bison, never with the mammoths preferred by Clovis hunters.

There may be a simple reason why Folsom hunters switched from mammoths to bison: there were no more mammoths left. There also were no more mastodonts, camels, horses, giant ground sloths, nor several dozen other types of big mammals. In all, North America lost an astonishing seventy-three per cent, and South America eighty per cent, of their genera of big mammals around this time. Many paleontologists do not blame this American extinction spasm on Clovis hunters, since there is no surviving evidence of mass slaughter—only the fossilized bones of a few butchered carcasses here and there. Instead, those paleontologists attribute the extinctions to changes of climate and habitats at the end of the ice ages, just around the time that Clovis hunters arrived. That reasoning puzzles me for several reasons. Ice-free habitats for mammals expanded rather than contracted as glaciers yielded to grass and forest; big American mammals had already survived the ends of at least twenty-two previous ice ages without such an extinction spasm; and there were far fewer extinctions in Europe and Asia when the glaciers of those continents melted around the same time.

If changing climate had been the cause, one might have expected opposite effects on species preferring hot and cold climates. Instead, radiocarbon-dated fossils from the Grand Canyon show that the Shasta ground sloth and Harrington's mountain goat, derived from areas of hot and cold climates respectively, both died out within a century or two of each other, around 11,100 years ago. The sloths were common until just before their sudden extinction. In their softball-sized dung balls, still well-preserved in some southwestern US caves, botanists have identified remains of plants on which the last sloths chomped: the Mormon tea and globe mallow, which still occur around those caves today. It is highly suspicious that both those well-fed sloths and the goats of the Grand Canyon disappeared just after Clovis hunters reached Arizona. Juries have convicted murderers on the grounds of less compelling circumstantial evidence. If climate really was what did in the sloths, we would have to credit those supposedly stupid beasts with unsuspected intelligence, since they all chose to drop dead simultaneously at just the right moment to deceive some twentieth-century scientists into blaming Clovis hunters.

A more plausible explanation of this 'coincidence' is that it really was a case of cause and effect. Paul Martin, a geoscientist at the University of Arizona, describes the dramatic outcome of hunter-meets-elephant as a blitzkrieg'. According to his view, the first hunters to emerge from the ice-free corridor at Edmonton thrived and multiplied, because they found an abundance of tame, casy-to-hunt big mammals. As the mammals were killed off in one area, the hunters and their offspring kept tanning out into new areas that still had abundant mammals, and kept exterminating the mammal populations at the front of their advance. By the time that the hunters' front finally reached the south tip of South America, most of the big mammal species of the New World had been exterminated.

Martin's theory has attracted lots of vigorous criticism, most of it centring on four doubts. Could a band of 100 hunters arriving at Edmonton breed fast enough to populate a hemisphere in a thousand years? Could they spread fast enough to cover the nearly 8,000 miles from Edmonton to Patagonia in that time? Were Clovis hunters really the first people in the New World? And could stone-age hunters really pursue hundreds of millions of big mammals so efficiently that not a single individual survived, while nevertheless leaving little fossil evidence of their hunts? Take first the question of breeding rates. Populations of modern hunter-gatherers on even their best hunting grounds number only about one per square mile. Hence, once the whole western hemisphere had been settled, its population of hunter-gatherers would have been at most ten million, since the New World's area outside of Canada and other areas covered by glaciers in Clovis times is about ten million square miles. In modern instances where colonists have arrived at an uninhabited land (for instance, when the H.M.S-. Bounty mutineers reached Pitcairn Island), their population growth has been as rapid as 3.4 % per year. That growth rate, which corresponds to each couple having four surviving children and a mean generation time of twenty years, would multiply 100 hunters into ten million in only 340 years. Thus, Clovis hunters should easily have been able to multiply to ten million within a millenium.

Could the descendants of the Edmonton pioneers have reached the south tip of South America in a thousand years? The overland straight-line distanced slightly under 8,000 miles, so that they would have to average eight miles a year. That is a trivial task—any fit hunter or huntress could have fulfilled the year's quota in a day and not moved for another 364 days. The quarry from which a Clovis tool was made can often be identified by its local type of stone, and we know in that way that individual tools travelled up to 200 miles. Some of the nineteenth-century Zulu migrations in southern Africa are known to have covered nearly 3,000 miles in a mere fifty years. Were Clovis hunters the first humans to spread south of the Canadian ice sheet? That is a harder question, and it is extremely controversial among archaeologists. Primacy claims for Clovis are inevitably based on negative evidence: there are no unequivocal human remains or artifacts with universally accepted pre-Clovis dates anywhere in the New World south of the former Canadian ice sheet. Mind you, there are dozens of claims of sites with pre-Clovis human evidence, but all or almost all of them are marred by serious questions about whether the material used for radiocarbon dating was contaminated by older carbon, or whether the dated material was really associated with the human remains, or whether the tools supposedly made by humans were just naturally shaped rocks. The two most nearly convincing of those claimed pre-Clovis sites are Meadowcroft Rock Shelter in Pennsylvania, dated to about 16,000 years ago, and the Monte Verde site in Chile, dated to at least 13,000 years ago. Monte Verde is described as having amazingly good preservation of many types of human artifacts, but those results have not yet been published in detail, so they cannot yet be properly evaluated. At Meadowcroft there has been an unresolved debate about whether the radiocarbon dates are in error, especially because the plant and animal species from the site are ones expected to have been living there only much more recently than 16,000 years ago. In contrast, the evidence for Clovis people is undeniable, is to be found in all forty-eight contiguous states, and is accepted by all archaeologists. Evidence for the still earlier settlement of the other habitable continents by more primitive humans is also unequivocal and universally accepted. At one Clovis site after another, you can see a level with Clovis artifacts and bones of numerous large extinct mammal species; immediately above (that is, younger than) the Clovis level, a level with Folsom artifacts but with the bones of not a single large extinct mammal except for bison; and immediately below the Clovis level, levels spanning thousands of years before Clovis times, reflecting benign environmental conditions, and full of the bones of large extinct mammals, but with not a single human artifact. How could people possibly have settled the New World in pre-Clovis times and not left behind the usual trail of abundant evidence that convinces archaeologists, like stone tools, hearths, occupied caves, and occasionally skeletons, with unequivocal radiocarbon dates? How could there have been pre-Clovis people who left no trace of their presence at Clovis sites, despite such favourable living conditions? How could people have arrived at Pennsylvania or Chile from Alaska, as if by helicopter, without leaving good evidence of their presence in all the intervening territory? For these reasons, I find it more plausible that the dates given for Meadowcroft and Monte Verde are somehow wrong, than that they are correct. The Clovis-first interpretation makes good sense, but the pre-Clovis interpretation just does not make sense to me.