COMMENTARY Holocene underkill Donald K. Grayson* Department of Anthropology, University of Washington, Seattle, WA 98195 nthisissueofPNAS,Joneset al. However, not all were large: the short- vival well into the Holocene became (1) address an issue that has been faced skunk (Brachyprotoma), the Aztlan clear in 1976, when G. V. Morejohn with us, in one way or another, for hare (Aztlanolagus), and the diminutive (11) reported C. lawi bones in an ar- some 200 years. Around the year pronghorn (Capromeryx)weresmall,the chaeological site north of Santa Cruz, 1800,I French paleontologist Georges latter weighing no more than 15 kg. Al- California dated to between 5,400 and Cuvier established the reality of verte- though the timing of the losses is un- 3,800 14Cyearsago.Heestimatedthat brate extinction by using animals so clear, there is no reason to think that the extinction of this bird had occurred large that their future discovery on the any lasted significantly after Ϸ10,500 between 2,500 and 3,000 years ago. hoof was highly unlikely (2). Included 14Cyearsago,whereas16ofthemam- Jones et al. (1) now confirm that esti- on the list were some now known to mals are known to have existed beyond mate, presenting new radiocarbon dates have been late Pleistocene in age, in- 12,000 14Cyearsago.Some,however, showing the survival of Chendytes to cluding mammoth and mastodon. As cannot be shown to have survived the 3,000 14C, or Ϸ2,500 calibrated, years time went on, the list of extinct Ameri- last glacial maximum, some 22,000 to ago. Younger sites within what was once can and Eurasian mammals of this age 14 18,000 Cyearsago(7). Chendytes territory do not contain grew hand in hand with our understand- From the beginning, the fate of spe- Chendytes remains, suggesting that this ing of the geological deposits in which cies on islands has been critical to the date is likely to be very close to the the fossils were embedded. As this hap- overkill argument. Martin observed that pened, it became harder and harder to time of extinction in this area. understand why all of these extinctions The early literature on C. lawi under- standably focused on its morphology had occurred. Island extinctions in the Things became even more complex in and skeletal similarities and differences 1859, not because of the publication of Pacific were associated with other waterfowl and other flightless Darwin’s On the Origin of Species,but birds. Morejohn (11) speculated that the because that year also saw the recogni- with massive landscape bird likely bred on offshore islands tion that the animals had actually coex- where it would be relatively immune to isted with humans (3). Such coexistence modification. predation. This speculation was con- raised the possibility that the extinctions firmed with Daniel Guthrie’s report (12) had happened at least in part because of of both immature individuals and egg- shells in late Pleistocene sites on San human hunting. Soon after 1860 many in prehistoric island settings extinctions scientists agreed that people must have seemed to follow human colonization Miguel Island, one of California’s played some role in causing the losses. with great speed (moas in New Zealand, Channel Islands. So we now know that the bird sur- Pleistocene Overkill sloths in the Caribbean, flightless rails in Oceania); all seemed to have been vived until 2,500 years ago and that it During the ensuing century, it was gen- quickly doomed by the arrival of hu- nested on islands. We also know that erally thought that the extinctions were mans. Human hunters and naı¨ve prey coastal California, including its islands, most likely caused by a combination of could not, he argued, coexist in such has an archaeological record that ex- human predation and climate change. tends back into the latest Pleistocene All of that changed in 1967, with the settings: people arrived and extinction quickly followed. (13). The simple presence of a publication of a powerful paper by pa- Chendytes bone in the lowest level of leoecologist Paul S. Martin (4). Martin Chendytes lawi Daisy Cave, on San Miguel Island, dated observed that the North American ex- to 10,000 14Cyearsago,doesnotmean tinctions seemed to coincide with the The extinct flightless sea duck Chendytes lawi,discussedbyJoneset al. (1), pro- that it represents an animal hunted by first arrival of people here, a phenenomen people or even introduced into the site known as Clovis and dated to 11,000 14C vides an important counter-example. by them. However, it does show that (Ϸ13,000 calibrated) years ago, plus or Apparently most closely related to the people and the bird were at the same minus a few centuries (5). It was the eiders (Somateria)(8),C. lawi was de- place at the same time. In addition, the sudden arrival of human predators in fined by Loye Miller (9) in 1925 based Chendytes the Americas, he argued, that put an on two specimens from late Pleistocene fact that remains have been end to a diverse array of herbivores and deposits near Santa Monica, California. found at so many island and mainland the carnivores dependent on them, a He recognized it as a large species of archaeological sites, sometimes in great process that became known as ‘‘Pleisto- extinct waterfowl but because only leg number, strongly suggests that people cene overkill.’’ He soon argued that all elements were known, he could not tell were preying on them, although tapho- of this happened within a few hundred that it was flightless. That understanding nomic analysis of all of this material years, a ‘‘prehistoric blitzkrieg’’ (6). came in 1947 when Hildegarde Howard would be helpful. As Jones et al. (1) The loss of some 35 genera of North (10) announced the discovery of wing note, even the specimens from mainland American mammals toward the end of bones that showed the bird could the Pleistocene continues to be hotly not fly. debated. Most of the mammals involved For many years, it was widely as- Author contributions: D.K.G. wrote the paper. were huge: a ground sloth with the sumed that Chendytes had been lost to- The author declares no conflict of interest. height of a giraffe and the bulk of an ward the end of the Pleistocene, even See companion article on page 4105. elephant (Eremotherium)andabeaver though there were early reports of mate- *E-mail: [email protected]. the size of a black bear (Castoroides). rial from archaeological contexts. Sur- © 2008 by The National Academy of Sciences of the USA www.pnas.org͞cgi͞doi͞10.1073͞pnas.0801272105 PNAS ͉ March 18, 2008 ͉ vol. 105 ͉ no. 11 ͉ 4077–4078 sites likely reflect hunting activities on individualism’’ is the general rule, not turalists, but this is not quite true. Moas nearby islands. the exception. If that is the case, the were lost on both South and North Is- knotty problem of understanding the land, New Zealand, but most of South Holocene Underkill North American extinctions is not likely Island was too cold to support the tropi- This work establishes that people coex- to be solved until each species involved cal food plants so important elsewhere isted with, and fairly obviously preyed has been provided with its own history. in Oceania. However, island extinctions on, a flightless, ground-nesting bird for This process is well under way in Eur- in the Pacific were associated with mas- some 8,000 years. In the Greater Anti- asia, but has barely begun in North sive landscape modification, including lles, there is strong evidence that people America (19). burning and the introduction of exotic coexisted with ground sloths for 1,000 The intellectual history of Chendytes animals, most notably the Pacific rat years (14). We are thus now learning shows the pattern well. Once thought to (Rattus exulans)anddog.Thehuman that rapid extinction is not the only pos- have been part of a wave of Pleistocene arrival on islands in other parts of the sible outcome for such vulnerable spe- extinctions, it is now known to have world was also routinely associated with cies as flightless ducks and huge sloths, lasted well into the Holocene. The same significant habitat alteration. All agree even if it is a common one. We are also can be said for many other species, giant that it was not just human hunting that learning that late Pleistocene species deer and muskox in Eurasia, for in- contributed to species loss in island set- that became stranded on islands did not stance, and the North American vam- tings, but the many and varied impacts always require a human presence to end pire bat Desmodus stocki,allnowknown that people had on the landscape. For breeding populations of Chendytes their existence. Mammoths became ex- to have lasted into, and sometimes deep on the Channel Islands, one such addi- tinct on St. Paul Island in the Bering into, the Holocene (17, 20, 21). Al- tional impact is provided by the fox, Sea after 5,700 14Cyearsagoeven though it is unlikely that any of the Urocyon littoralis,seeminglyintroduced though people did not arrive there until North American genera involved in the by people by at least the middle Holo- historic times (15, 16). A similar event debate survived this long, it remains cene, before the bird’s extinction (ref. 23 occurred in Ireland and on the Isle of fully possible that substantial losses and T. Rick, personal communication). Man, where the giant deer Megaloceros occurred long before the 11,000 year 14 As Jones et al. (1) observe, dogs may was lost shortly after 11,000 Cyears date that has been the focus of the have played a role as well, as might the ago, again before people arrived (17).