What Happened to the Megafauna?

The Overkill Hypothesis

The overkill hypothesis states: ‘‘the majority of animal extinctions that have taken place over the Quaternary have resulted directly or indirectly from human hunting.’’ Proponents of overkill do not claim that humans caused all animal extinctions of the Quaternary; nor do they claim that all human-caused extinctions resulted from over-hunting of extinct prey. For example, extinctions of large carnivores might have resulted from competition with humans rather than direct human predation of carnivores. The argument is simply that prehistoric human hunting dramatically modified animal communities, and had humans not colonized the planet, most of the species that suffered extinction during the Pleistocene and Holocene would still exist today. Although the origins of overkill can be traced to the nineteenth century, overkill was championed in the latter half of the twentieth century by Paul Martin, a palaeoecolo-gist at the University of Arizona.

Figure 1 Line drawing reproductions of Upper Palaeolithic cave paintings of extinct Pleistocene fauna. Irish elk, Lascaux cave a, woolly mammoth, Rouffignac cave b, cave bear c, and two-horned rhinoceros d, Chauvet cave.

The keystone to Martin’s argument was the apparent widespread occurrence of ‘first contact’ extinctions across the globe. With the important exceptions of the continents of Africa, Europe, and Asia, waves of animal extinctions immediately folLowed initial human colonization of many regions of the world (Figure 2). In the 1970s, when considering extinctions in the Americas, Martin developed his most explicit formulation of the overkill hypothesis, a model he called ‘blitzkrieg’. With mathematician James Mosimann, Martin constructed a computer simulation which showed how humans could have colonized the entirety of the unglaciated New World from southern Canada to Patagonia within 1000years (Figure 3). Rapid population growth and migration would have been fueled by hunting of now-extinct Pleistocene mammals. These animals which had never experienced human hunters would have been naive and easily killed.

Although the overkill hypothesis (or variations thereof) applied to oceanic islands extinctions is generally accepted today, whether overkill explains extinctions on continents remains highly controversial. The most serious obstacle to overkill is that in most regions archaeological evidence for human exploitation of extinct taxa is scarce. In North America, for

Figure 2 The relative timing of human colonization and the major wave of animal extinctions for various continents (left) and islands (right). Solid colors and lines indicate certainty about age. Dashed lines and colors indicate uncertainty about age. Age axis is log-scaled.

Figure 3 The Martin/Mosimann ‘blitzkrieg’ model of Pleistocene extinctions. Humans enter North America via the ‘Ice-free corridor’ at approximately 13 400 BP, and rapidly colonize the New World within 1000 years. Human range expansion takes the form of a traveling wave. Along the wave front, naive Pleistocene fauna are hunted to extinction. Redrafted from Figure 2 from Martin PS (1973) The discovery of America. Science 179, 969-974.

Example, although 33 genera of large-bodied mammals suffered extinction during the Late Pleistocene, fewer than five can be shown to have been utilized by humans. If humans caused the extinction of North American species by over-hunting, then they must have killed thousands if not millions of animals, which begs the question, ‘Where is the archaeological evidence?’ Martin has argued that if blitzkrieg-type overkill happened very quickly, little archaeological evidence would be expected. For others, however, this lack of direct evidence has meant that perhaps we should be seeking explanations for extinctions elsewhere, such as in the dramatic swings in global climate that have occurred during the Quaternary.

Climate Change

Variations in the Earth’s orbit over the last two million years have caused climate to oscillate between Glacial and interglacial conditions. There were more than 20 instances when continental glaciers expanded and contracted over the Northern Hemisphere (Figure 4). As continental glaciers waxed and waned, oceanic and atmospheric temperatures cooled and warmed, sea levels fell and rose, precipitation regimes were greatly modified, and plant and animal species migrated and reorganized ecological communities. Because swings in climate had major impacts on biological organisms, there is no doubt that animal extinctions could have resulted, but there is considerable argument about how climate change might cause mass extinctions, and whether climate change explains the observed extinctions of the Quaternary.

Figure 4 The last five glacial cycles recorded in stable oxygen isotopes from benthic foraminifera recovered from a core of the Pacific Ocean floor. Oxygen isotopes of marine foraminifera record the isotopic composition of ocean water which is a direct reflection of the volume of glacial ice in terrestrial settings. High values of S18O indicate glacial periods. Low values of S18O, like those observed in modern times, indicate interglacials. Data from Mix AC et al. (1995) Benthic foraminferal stable isotope stratigraphy of site 846: 0-1.8 ma’. In: Pisias NG et al. (eds.) Proceedings of the Ocean Drilling Program. Scientific Results 138, College Station, TX, USA. 839-854.

A handful of models have been proposed which attempt to link Quaternary climate change to extinctions. Unlike overkill models, climate change hypotheses are typically constructed with reference to specific geographic regions because extinctions occurred at different times in different parts of the world. The most explicit climatic/ecological extinction models have considered North America. For example, Dale Guthrie, a vertebrate palaeontologist at the University of Alaska, has argued that the transition from glacial to interglacial climate in North America (20 00010 000 years ago), resulted in vegetational communities that would have been detrimental to many species of Pleistocene megafauna. Lower levels of seasonality and longer growing seasons in Pleistocene ecosystems produced high-diversity mosaics of vegetation that would have permitted large herbivores to easily meet their nutritional requirements. With the transition to the Holocene, North American biomes were transformed into large-scale low-diversity patches wherein generalist feeders with simple digestive tracts, such as mammoths, mastodons, sloths, and horses would essentially starve, or due to anti-herbivory toxins, be poisoned to death by being forced to eat large quantities of toxic foods, to which these animals had little physiological resistance. According to Guthrie, only large mammals with specialized diets or physiologies, such as bison, pronghorn, and moose, thrived in Holocene ecosystems.

Climate-based hypotheses, like overkill, suffer from many problems. Most commonly, it is questioned why it was a particular climate change that caused the extinctions. For example, in the Americas, extinction correlates with the most recent transition from glacial to interglacial climate, but this was just one of many such transitions that occurred during the Pleistocene. Proponents of overkill would argue that this episode of climate change just happens to correlate with the first arrival of human migrants; proponents of climate hypotheses would argue that this period of climate change was uniquely severe. Also, proponents of climate hypotheses must explain why extinctions seemingly always correlate with human colonization. If humans had no causal role in Pleistocene extinctions whatsoever, then the timing of extinction should be independent of human global colonization, and yet human migration and animal extinction seem to be highly correlated (Figure 2).

Hyperdisease

Due in part to problems with overkill and ecological-climatic hypotheses, a third major contending explanation for Pleistocene extinctions has come to the fore - hyperdisease. This idea, proposed by palaeontologist Ross MacPhee and virologist Preston Marx, can explain two aspects of Quaternary extinctions with which overkill and climate change struggle. Climatic models fail to explain the global pattern of ‘first contact’ extinctions, and overkill struggles to explain the paucity of archaeological evidence demonstrating human hunting of extinct fauna. MacPhee and Marx propose that extinctions were caused by the introduction of ‘hypervirulent, hyperlethal diseases’, which they call ‘hyperdiseases’, by humans or domestic dogs to naive populations of animals, naive in the sense of having no immunity to such pathogens because they had no prior experience with them. The hyperdisease hypothesis predicts that extinction should follow a first contact pattern and should produce little archaeological evidence of human exploitation of extinct animals.

However, the disease model has yet to successfully overcome a range of potential theoretical setbacks: (1) Foremost, it is unclear whether a disease is capable of driving any species to extinction under natural conditions. (2) It is also unknown if a single disease could infect such a broad range of species as those lost during the Quaternary. (3) Finally, this hypothesis is very difficult to test. Efforts are in progress to attempt to recover pathogens directly from remains of extinct animals or from their fecal matter, but even if pathogens are discovered, it will be difficult to determine if in fact the identified pathogen is the sought-after hyperdisease.

The Keystone Herbivore Hypothesis

South African ecologist Norman Owen-Smith, an expert in ‘mega-herbivores’ (>1000kg), recognized that the extinction of the largest of the Pleistocene fauna must have had dramatic effects on past vegetation. Extant terrestrial mega-herbivores, the African and Asian elephants and rhinoceroses, through foraging and trampling can dramatically impact ecosystems, turning forests and woodlands into scrubland, savannah, or grassland. The removal of mega-herbivores by hunting or climate change, therefore, could have dramatic effects on the structure of vegetational communities. If mega-herbivores served as keystone species in Pleistocene environments, their removal could have cascading effects causing the extinction of numerous other animal taxa. Owen-Smith calls this idea the ‘keystone herbivore hypothesis’. The Achilles’ heel of the keystone herbivore hypothesis is the prediction that mega-herbivore extinction should have preceded the extinction of other taxa, and to date there is very little evidence to indicate that this actually happened.

Sitzkrieg

Jared Diamond argues that most Pleistocene extinctions can be attributed to humans, but diverges from traditional overkill models with respect to cause. Using observations of historically recorded extinctions, Diamond notes that many human-induced animal extinctions are not caused by hunting, but have instead resulted from secondary effects. Diamond contrasts the traditional ‘blitzkrieg’ model to what he calls ‘sitzkrieg’. Diamond suggests that extinctions may not have been a lightning-quick predatory assault on Pleistocene fauna by humans, but instead a slow ‘war of attrition’ where extinctions were by-products of secondary impacts, such as felling of forests, anthropogenic burning, and the introduction of non-native competitors. Like the disease model, the ‘sitzkrieg’ model predicts that extinctions should correlate with or postdate human colonization, and that there should be relativelY little archaeological evidence for human hunting of extinct fauna. As such, it remains a very difficult model to distinguish from hyperdisease, but some researchers find this idea very compelling for explaining extinctions on oceanic islands.