In regard to the wildness of birds towards man, there is no way of accounting for it except as an inherited habit . . . both at the Galapagos and at the Falklands, [many individuals] have been pursued and injured by man, but yet have not learned a salutary dread of him. We may infer from these facts what havoc the introduction of any new beast of prey must cause in a country before the instincts of the indigenous inhabitants have become adapted to the stranger's craft or power.
Charles Darwin, Journal of Researches . . . During the Voyage of the Beagle
As we have seen, the basis for the overkill model is what Ross MacPhee calls the "deadly syncopation" of human arrivals and megafaunal extinctions in new lands. Before we explore some of the arguments raised against this model, it will be useful to review that syncopation in more detail.
Although geochemical dates are not available for the times of extinction of all the target species (those whose disappearance is of interest relative to human arrival), the trend is quite clear. The fossil record shows no concentrated extinction of large mammals until we reach near time (Alroy 1999). Then, within the last 50,000 years—where we have the advantage of a much more refined time scale, thanks in large part to radiocarbon dating—megafaunal extinctions pop up independently in different parts of the world (Martin and Steadman 1999). From a modest start in Africa and Eurasia one to two million years ago (reaching Flores by one million years ago; Morwood and others 1998), they erupt in near time in the following sequence: Australasia, the Solomon Islands, continental America, the West Indies, and Pacific islands from New Caledonia east to Hawaii and Rapanui (Easter Island). They end (apart from historic losses) in New Zealand and Madagascar, with moa extinction in the former approximately 500 years ago (Bunce and others 2003) and hippopotamus extinction in eastern Madagascar 200 years ago.
Large-animal extinctions on the continents of human origin, Africa and Asia, were relatively few and episodic, not only in near time (see totals at bottom of table 3) but over the last several million years (Klein 1999). In Africa, extinctions of large mammals were last apparent around 1.5 to 2.5 million years ago, during the early evolution of the genus Homo. A moderate extinction pulse involving a small number of species blighted Europe and Asia over the last 70,000 years. Some of these were simply extirpations, or local extinctions; close relatives survive elsewhere (Stuart 1999). Mammoths disappeared gradually in Eurasia, over thousands of years (Stuart and others 2002). In less than a millennium they disappeared from North America, leaving a small group of survivors on the Pribilof Islands in the Bering Sea. Even when summed, the Afro-Asian losses extending over more than a million years do not match the number of near-time extinctions in either Australia, the Americas, Madagascar, or New Zealand.
The large number of large mammals in Africa, many more than in America, has long been accepted as a basic fact of zoogeography, a natural condition and a baseline. Historically, no genus of large mammal in the New World features more than a few species, and no part of the Americas could begin to match the game plains of Africa for numbers of large mammals, especially artiodactyls. Above all, the New World lacks anything to match Africa's elephants, black and white rhinoceroses, hippopotamuses, and giraffes, five familiar species of living megaherbivores.
Rarely considered, at least until recently, is the fact that in the late Quaternary North America north of Mexico had at least 10 species of megaherbivores (five proboscideans, two ground sloths, two glyptodonts, and a camelid). This is twice Africa's quota. Beyond that, South America was more than twice as rich as North America, with an extraordinary assemblage of 25 species of megaherbivores, those exceeding 1,000 kilograms (2,200 pounds). These include four species of proboscideans, nine of ground sloths, five glyptodonts, five notoungulates (all but one in the genus Toxodon), and two camelids (Lyons, Smith, and Brown 2004). In addition, South America was the evolutionary center for pho-
rusrhacoid birds, highly specialized predators with a beak like that of a hawk or an eagle but much larger (Murray and Vickers-Rich 2004).
For its part the Afro-Asian fauna is much richer than that of the Americas, either in the present or in near time, in medium-sized herbivore species, those ranging from about 0.5 to 45 kilograms (1 to 100 pounds) in body mass (see figure 1 in Lyons, Smith, and Brown 2004). For example, Africa has 18 species of duiker (genus Cephalophus), small antelopes ranging from the mass of a hare to that of a small female elk. Similarly, 16 species of gazelle (species Gazella) weighing between 12 and 85 kilograms (25 to 185 pounds) are found in Africa and parts of Asia. Nothing like them exists among the larger (over 45 kilograms) or small mammals in the Americas. In comparison, the most species-rich genus of small artiodactyls in the New World is the brocket deer (Mazama), with four (possibly six) species. Though the current distribution of mammalian fauna has long been accepted as simply the way things are, it has been greatly influenced by different extinction rates on different continents over the last two million years and especially in near time.
Under the overkill model, two arguments may explain the lower extinction rate on the continents of human origin. In the Old World ho-minids evolved with other mammals, which accordingly developed both fear of human predators and other defenses against them. As our hominid ancestors moved into drier or colder and often less hospitable areas, such as the high plateaus of central Asia and the coast of the Arctic Ocean, contact between large mammals and Paleolithic people was limited.* Taxa of Rangifer (caribou or reindeer), Ovis (mountain sheep), Bison (bison), Cervus (red deer), and Alces (moose, elk) had a lengthy exposure to Paleolithic hunters and may have evolved more resistance to them than any of the genera of large mammals suddenly encountering Stone Age humans entering the Americas, especially South America, where genetic influence from Old World fauna was negligible. It is no surprise that South American large mammals were obliterated.
Data from Australia, the first continent to be invaded by humans, are crucial to the overkill model. In the 1960s the first radiocarbon dates appeared showing the potential time of extinction of Australia's megafauna. The youngest dates on diprotodonts were 7,000 and 13,000 years ago.
*In Africa and Asia, human diseases such as sleeping sickness and tropical malarias may also have favored the survival of large mammals. Elephants, rhinoceroses, and hippos survived in more tropical areas but not in temperate, boreal, and/or subarctic parts of Eurasia, where human pathogens are less numerous or less virulent.
The dates were overshots and not replicated by later geochronology. It was the same sort of problem I encountered in trying to assemble a chronology of megafaunal extinction in North America over 40 years ago. The first dates released were seriously inaccurate, a common failing of bone samples not properly treated.
Recently Australian geochemists and paleontologists released a series of thermoluminescence dates on their extinct fauna (Roberts and others 2001). The method lacks the precision of radiocarbon. Nevertheless, none of the dates were younger than 46,000 years. While generally accepted, an age of 46,000 for contact between humans and extinct fauna is challenged in one stratified site at Cuddie Springs in New South Wales (Wroe and others 2004). But the vast majority of Australian evidence indicates that large mammals, large birds, and large reptiles became extinct around 42,000 to 48,000 years ago (Gillespie 2002; Roberts and others 2001). And Homo sapiens reached Australia roughly coincident with these extinctions, as we will see in chapter 7 (Mulvaney and Kamminga 1999). This discovery contradicted those who once assumed, decades ago, that not until the early or mid-Holocene, just a few thousand years ago, would ancestral Aborigines have been clever enough to make the boats needed to cross the ocean from Southeast Asia to Australia.
In the Americas, the time of extinction of about half the lost genera remains to be determined critically. However, there is no chronological indication that any of the extinct American genera listed in tables 2 and 3 endured after 13,000 calendar years ago. These genera are absent from numerous fossil deposits (including bone beds) of the last 10,000 radiocarbon years excavated in North, Central, and South America. In addition, the youngest radiocarbon dates on these genera from many localities terminate at around 11,000 radiocarbon years ago. The boundary is a sharp one. Pending further geochronological testing, the extinction of ground sloths and Harrington's mountain goats appears to be particularly well dated.
From other extinct animals in caves in the Southwest, Ken Cole, Donna Howell, Jim Mead, Geof Spaulding, Bob Thompson, Tom Van Deven-der, and other Desert Lab researchers have also recovered dates not appreciably younger than 11,000 radiocarbon years ago. As we have seen, the youngest reliable dates from both Argentina and Chile are quite similar (see table 5).
Back in the 1950s and 1960s I flirted with much younger (and much older) radiocarbon dates on extinct North American megafauna. Many cultural associations linked artifacts with bones of extinct mammals. Ar chaeologists reported human artifacts or at least cultural charcoal mixed with the bones of extinct North American animals radiocarbon dating from 2,000 to 30,000 years ago (Martin 1958b), a result that proved erroneous at both ends. There were bugs to work out, especially with charcoal, which was not always in a true association with the animal bone investigators dearly hoped to date. The problem lingers. In youthful abandon I did what armchair consumers of published radiocarbon date lists often did in those early days. We had not visited, much less excavated, the fossil sites or dealt with any of the fossil and stratigraphic evidence. We had no basis on which to presume to pick and choose. If professional archaeologists or geologists found associations of artifacts with remains of extinct species, I accepted their dates.
Taken literally, those dates indicated that people and extinct animals had coexisted for thousands of years. From Science, for example, I extracted a Florida date on charcoal with the lab catalogue number Lam-ont 211 that was said to be associated with extinct mammals. If accurate, this would be extraordinary: extinct large animals still alive in Florida only 2,000 years ago (Martin 1958b)!* As far as I know, L-211 still lies unmourned in the sizable geochronological graveyard of anomalous, undefended, or unreplicated radiocarbon dates. In the 1950s, more dates suggested that mammoths and other extinct megafauna lived as late as 8,000 radiocarbon years ago. Libby's C-222 date of 8,500 years on Shasta ground sloths at Gypsum Cave was an example. But this date could not be replicated (Poinar and others 1998). And with new suites of radiocarbon dates, such as the series on ground sloths, extinct mountain goats, and condors from the Grand Canyon and elsewhere in the Southwest, and the steady input of new dates around 11,000 radiocarbon years on extinct horses, camels, saber-toothed cats, and mastodons, the chronology supporting extinctions 8,000 or fewer years ago crumbled (Martin 1990; Stuart 1991).
Much additional chronological work is needed in the West Indies,
*An outraged Florida archaeologist of impeccable credentials, R. P. Bullen, concluded, based on his experience excavating many archaeological sites, that extinct mammals could not possibly have survived in Florida until only 2,000 years ago. No good associations of cultural remains with extinct animals that late, or even thousands of years older, were known to exist in Florida. The problem with L-211 might well have been intrusion of charcoal from a younger hearth into a much older deposit of extinct animal bones. Though in most cases charcoal is a reliable source of carbon for radiocarbon dating, an association problem of this type could lead to misinterpretations. Whatever the explanation, no defenders of L-211 emerged, and ten years later, without comment beyond citing Bullen's article, I washed my hands of it (Martin and Wright 1967).
which, like mainland North and South America, seem to lack robust associations between cultural material and bones of extinct species. At least some of the dwarf ground sloths there lasted until roughly 5,000 years ago, and archaeologists place the first human arrival in the Caribbean Islands at about 6,000 years ago (Wilson in Fagan 1996). The fossilif-erous asphalt seep in Matanzas Province in Cuba includes four genera of dwarf ground sloth (Iturralde-Vinent and others 2000). It also includes Ornimegalonyx, a nearly flightless giant raptor or "walking owl" up to 18 kilograms (40 pounds) in weight—possibly one of Cuba's top terrestrial carnivores, second in size only to the terrestrial crocodiles. It includes, as well, an extinct condor, Gymnogyps varonai—perhaps Cuba's largest extinct scavenger—and an extinct crane, Grus cubensis. Ross MacPhee supplied samples of a dwarf ground sloth, Parocnus brownii, from the seep for radiocarbon dating. The results ranged from 4,960 ± 280 years at the youngest to 11,880 ± 420 years at the oldest. At 4,960 to 6,330, MacPhee's youngest dates are "apparently in good agreement with the trend of the earliest archaeological dates for this island" (MacPhee, Flemming, and Lunde 1999).
After the West Indies, extinctions next erupted in Madagascar. Recent chronological refinement suggests that Madagascar's giant birds, hippos, giant tortoises, and giant extinct lemurs vanished around the time of human arrival from Borneo over 2,000 years ago. Recent work by David Burney (Burney and others 2004) indicates extinctions within the last 2,400 years in the megaherbivores on Madagascar's west coast. Presumably extinction of hippo and the elephant bird, Aepyornis, accounts for a reduction in the dung fungus Sporormiella and a rise in charcoal. The discovery of the sensitivity of this spore type to the presence of dung and dung production by large animals may prove to be the most valuable tool in the methodology of those seeking to detect and date both extinctions and change in biomass.
Around the same time, as we have seen, extinctions struck remote islands in the Pacific, followed by New Zealand less than 1,000 years ago. For the relatively few islands undiscovered by prehistoric voyagers, such as the Galapagos in the southeastern Pacific, the Commander Islands in the north Pacific, the Mascarenes in the Indian Ocean, and the Azores in the Atlantic, evidence of prehistoric extinctions is minimal or unknown. In some cases there is evidence that the same species went extinct on different landmasses at different times, coincident with human arrival. For instance, in New Caledonia, the last records of the horned turtle, Meiola-nia, are associated with cultural remains about 1,500 years old; this genus had gone extinct tens of thousands of years earlier in Australia (Martin and Steadman 1999, 27-28).
Recent discoveries of island extinctions are particularly interesting. Almost without notice, beginning in the mid-1970s, avian paleontologists of the Smithsonian Museum reported bones of new taxa of an extinct goose-sized flightless duck and a flightless ibis from Hawaii (Olson and Wetmore 1976). Since then, thanks especially to paleo-ornithologist David Steadman and his archaeological collaborators, a completely unexpected flood of fossil finds, especially the bones of extinct birds, has turned up on islands and archipelagoes in the remote Pacific. Investigators from Darwin onward (summarized in Quammen 1996) had simply missed finding the fossils. Therefore, only recently has it been possible to evaluate these rich faunas within an archaeological framework (Kirch and Hunt 1997). As it turns out, humans played an unexpectedly traumatic role in a severe prehistoric reduction of the fauna of Pacific islands (Burney and others 2001; Kirch and Hunt 1997; Steadman 1995, n.d.).
Endemic Pacific island birds included taxa of parrots, pigeons, doves, megapodes or bush turkeys, and especially flightless rails. Thousands of these taxa, as well as many seabird colonies, are no more. Over 2,000 taxa of flightless rails (Rallidae) alone have been lost (Steadman 1995, 1997, n.d.). The 800 Micronesian, Melanesian, and Polynesian islands over a square kilometer in area have together lost roughly 8,000 species, taxa, or indigenous populations of land birds (Martin and Steadman 1999, 29). Other terrestrial vertebrates and numerous taxa of endemic land snails have also disappeared. Islands in many other places as well have lost small endemic mammals, including some that had evolved from larger continental mammals. Many of these endemics disappeared in the Holocene, typically when brought into contact with continental species of mice or rats (Rattus).
Pacific island fossils occur in cave deposits or open sites on all islands inhabited prehistorically. When they have been radiocarbon dated, most of the extinctions fall in the last few thousand years, and all of these correlate with the arrival either of humans or of the Pacific rats that accompanied the voyagers (MacPhee and Marx 1997), often within a few hundred years (Martin and Steadman 1999, 29). From west to east across the Pacific, beginning 3,000 years ago in New Caledonia and Tonga, passing through Hawaii and Rapanui (Easter Island) 1,500 years ago, and ending only 700 years ago in New Zealand, extinction marked the spread of our species.
In fact, directly or indirectly, human colonization resulted in more near-
time vertebrate extinctions on Pacific islands than on the continents (Steadman 1995, n.d.).* According to Dave Steadman, "On any remote Pacific island with a respectable fossil record, one can expect to find at least two to three times the number of land bird taxa before contact than after" (Martin and Steadman 1999, 29). The evidence of syncopation on particular islands is quite clear. For instance, there were far fewer if any near-time extinctions on the Hawaiian islands before colonization than after (Martin 1990). In New Zealand, radiocarbon dating suggests that "human hunting and habitat destruction drove the 11 species of moa to extinction in less than 100 years after Polynesian settlement" (Holdaway and Jacomb 2000). (The moa fauna has recently been reduced from 11 to 10 species by recovery of nuclear DNA sequences, and further reduction is expected [Huynen and others 2003].) All 14 New Zealand birds weighing over 9 kilograms (20 pounds) disappeared, including moas up to 180 kilograms (400 pounds), along with about 28 of the 140 species of birds under 9 kilograms (A. Anderson 1997; Worthy and Holdaway 2002).
Exceptions to the general pattern, not only in the Pacific but also in the other oceans of the world, are instructive. Islands lacking artifacts or other evidence of prehistoric inhabitants (including severe prehistoric extinctions) include the Azores and Bermuda in the Atlantic, the Mas-carenes in the Indian Ocean, Lord Howe Island east of Australia, the Commander Islands in the Bering Strait, and the Galapagos west of South America. Such islands serve as controls for an anthropogenic extinction model.
In addition there are "islands of doom," occupied and abandoned pre-historically. These islands, including Norfolk, Henderson, and Pitcairn in the South Pacific and Nihue in the Hawaiian chain, had been colonized and then, apparently after exhaustion of their resources, abandoned. The most dramatic example of resource depletion was on Rapanui (Easter Island), where the wood for canoes was depleted and surviving settlers engaged in internal warfare. This suggests prehistoric humans' capacity to exhaust resources in a touch-and-go.
Worldwide, the body size of animals scales to the size of the landmass under consideration: the largest animals on smaller landmasses are smaller than those on large landmasses. This is also true of the body size
*MacArthur-Wilson island biogeography (developed by eminent ecologists Robert MacArthur and E. O. Wilson) employs mathematical models to explain why certain islands supported richer faunas than others. I suggest that we recognize "Olson-Wetmore island biogeography" after Storrs Olson and Alexander Wetmore, the first to show massive extinctions on oceanic islands accompanying human arrival.
of extinct animals. For example, in North and South America the preponderance of extinct mammals exceeded 45 kilograms (100 pounds), while Madagascar, the size of Texas, lost mammals down to 9 kilograms (20 pounds). The West Indies lost many more medium to small mammals than either continental North America or Madagascar, and oceanic islands, especially in the remote Pacific, lost animals down to the size of land snails.
This scaling is consistent with the overkill model. If we make the reasonable assumption that human hunters took the path of least resistance, they would first have gone after those species that were easiest to track, find, and kill and that provided the most food (or prestige) for the least effort. In general, these would have been the largest animals in any given region. Human foragers would have worked their way down the size scale until they reached species that were difficult to kill. In addition, it would have been easier to find smaller animals in smaller areas. Finally, any other ecological disruption caused by humans (e.g., an increase in fires) would have had greater impact in smaller areas. For all these reasons, the size scaling of the extinctions supports the idea of overkill.
In short, the global pattern of extinctions in near time appears to be just what one might expect if people played the major role in triggering them. If this "deadly syncopation" was a coincidence, if the extinctions had nothing to do with our species and its global hegira, their "true" cause will be the greatest geological discovery of the new millennium.
The syncopation we have seen, however, would be consistent not only with overkill but with "overill": the idea that the extinctions were caused not only by human hunting but also by other debilitating changes introduced by humans, such as new commensals (Steadman 1995) or pathogens (MacPhee and Marx 1997). In some cases it seems clear that introduced rats caused extinctions.
A dramatic case has emerged from New Zealand. Some 55 species of flightless or ground-nesting birds vanished after Pacific rats (Rattus exulans) appeared 2,000 years ago (Worthy and Holdaway 2002). Very likely as stowaways in double-hulled sailing canoes crewed by Polynesian explorers, the rats crossed the vast Pacific, jumping ship on islands free of terrestrial mammals. While not endorsed in all quarters, recently obtained radiocarbon dates on Pacific rat bones indicated their presence in New Zealand before human settlement (Holdaway and others 2002). These rats apparently eliminated the smaller members of New Zealand's extinct fauna, such as flightless wrens and "giant" insects. In addition, presumably because of rat predation, 57 species of ground-nesting pe trels, Storm Petrels, and other Procellariiformes (an order of predominantly pelagic birds) abandoned breeding colonies in New Zealand. Most managed to survive by virtue of nesting colonies on rat-free islands elsewhere in the Southern Hemisphere (Steadman 1995, n.d.).
For whatever reason, the Polynesians themselves seem not to have settled New Zealand until at least a millennium after their initial landfall. Then they rapidly eliminated 10 species of moa. Human hunting forced moa extinctions; the loss of smaller species may involve the Polynesian rat, introduced dogs, and other side effects of human colonization, such as increase in wildfire.
Of the various factors that could have contributed to overill, rats probably receive the most scholarly attention. Microorganisms are more controversial, partly because they are harder to detect in the fossil record. Corbett (1973) proposed that prehistoric humans had unwittingly spread viruses akin to Ebola, which can burn explosively through populations of large animals. Similarly, Ross MacPhee and Preston Marx (1997) have proposed hyperdisease, the inadvertent introduction by the first human invaders of highly lethal pathogens able to jump the species barrier. They note that disease appears to be a factor in at least some historic extinctions, as illustrated by the loss of two endemic species of Rattus on Christmas Island south of Java. But most of the "extinction-ists" I have consulted consider hyperdisease at best an unlikely cause of the worldwide extinctions of large mammals. Its role appears to be testable by mitochondrial DNA analysis, however, and MacPhee has assembled a paleovirology team and is searching for lethal diseases in possible victims, such as the youngest known mammoths, those found on Wrangel Island.
Other indirect anthropogenic causes of extinction have also been proposed. These include Dan Janzen's theory that "the Pleistocene hunters had help," which Charles Kay has refined using Canadian data on wolfmoose predation (Kay 2002). Under this model, as the proboscideans declined, predators such as the giant short-faced bears, scimitar cats, and American lions would have turned to smaller prey (Janzen 1983; Kay 2002). Bob Dewar (1997) has proposed that wild cattle helped force extinctions in Madagascar. And in some cases, the precise reason that a species died off after human colonization is simply not known. For example, the giant rats of the Galapagos went extinct shortly after the Spanish discovered the islands. There are no rat kill sites, but some sort of anthropogenic linkage, rat-borne diseases included, is strongly suspected.
Overkill does not, of course, purport to explain every extinction since the evolution of Homo sapiens. Although firmly established examples are few, some late-Quaternary extinctions apparently occurred before human arrival on the landmasses in question and thus must be ascribed to other causes. To be certain of this, robust chronologies and, ideally, DNA evidence are needed.
In my initial brief treatment of the West Indies, for example, I assumed that all known extinctions corresponded with the arrival of prehistoric people. More critical appraisal reduced the number of such cases. For instance, according to MacPhee and others (1989), extinction of the giant rodents Amblyrhiza in Anguilla and Clidomys in Jamaica, which I assumed reflected overkill (Martin 1984), predated human arrival. If Manuel Iturralde-Vinent and Ross MacPhee (1999) are correct that most land mammal lineages entered the Greater Antilles around the time of the Eocene-Oligocene transition, or at any time in the Tertiary, a large number of Tertiary turnovers (extinctions) predating near time, and hence human arrival, can be expected. In the majority of cases, however, the extinctions of endemic West Indian mammals are not chronologically separated from possible human presence and from prehistoric human activities. Undoubtedly, significant extinctions accompanying evolution occurred on all the large and persistent island platforms rising out of deep water. Such change is not incorporated in the overkill model and may be difficult to detect if the fossil record is poor, as is often the case for insular faunas that predate near time. Overall, insular faunas of extinct vertebrates provide a valuable opportunity to test anthropogenic and other extinction models.
Was this article helpful?