The Case of the Arboricidal Megaherbivores

Elin Whitney-Smith Ph.D.

Holmes and Watson Investigate Pleistocene Extinctions

I have chosen to present this model of the extinctions at the end of the Pleistocene in the form of a detective story. In the traditional detective story the reader is asked to consider a wide variety of apparently unconnected clews which the detective then pulls together in a convincing story of how and why the crime was committed. This paper also asks the reader to consider a widely divergent set of facts, some of which have been used to support extinction theories, which I want to use in a new way.

Scientists reading this paper will undoubtedly hold views on which theory he or she supports. This makes it difficult for the informed reader to look at the evidence separately from the known and accepted interpretation. Because the detective story asks for a suspension of disbelief it is an effective mode for the presentation of complex models. A precedent for this use of the genre was established by Albert Einstein and Leopold Infeld in their work, The Evolution of Physics (1938).


Our scientific investigators, Skylark Holmes (great great grand niece of the famous detective) and her friend Dr. Janet Watson have taken on the case of the New World Megaherbivore Extinctions of the Pleistocene.

"Well, Holmes, How do we proceed."

"First we should survey the existing evidence, make a note of everything we think is important, and organize it into something easier to understand."

"How do we know what is important?"

"Well, it seems that we need to know how today's environments, the environments of the Holocene, differ from the environments when the animals existed, those of the Pleistocene."

"Ah, yes. Good idea."

"To make it easier to understand the solutions which have already been proposed we will name them and enter brief summaries into our data base." Holmes types and Watson follows along on her screen.

Temperature Change

The first proposed solution - temperature change - is based on the observation that the extinctions coincided with the wasting of the last glaciers. It posits that global warming was the cause of the large mammal extinctions. However, since it was proposed, it has become evident that today's mean annual temperature is no warmer than that of previous interglacials (Davis 1976, Andersen 1973, Ashworth 1980, Birks & Birks 1980, Birks 1973, Bradely 1985). Thus warmer temperature alone is not a sufficient explanation for the extinctions since the large mammals survived similar temperatures in previous interglacials.

"OK, Holmes. What now?"

"The dismissal of the temperature change prompted the question: What makes this interglacial different than previous interglacials? One answer is the entrance of Homo sapiens to the New World. This led to the overkill hypothesis which pins it on those found at the scene of the crime."


The overkill hypothesis suggests that as humans crossed into the New World they hunted the herbivores to extinction. The carnivores which depended upon those herbivores went extinct because their prey was gone (Martin, 1963, 1967, 1984, 1986, 1988, Reeves 1978, Scott 1984). This is based on the observation that the pattern of extinctions follows the march of humankind: 1- animals go extinct soon after the introduction of humans and 2- extinctions are greatest where humans have arrived last - the New World - and are least where humans were first - Africa. This is based on the reasoning that in Africa animals learned and evolved as H. sapiens developed new hunting techniques. The animals of new territories were naive about hunting and were more vulnerable - there was a greater gap between the ability of the hunter to kill and the ability of the animal to evade being killed. This predicts that the later H. sapiens entered a territory the more animals will be killed. (Foley 1984, Dawkins & Krebs 1979).

The overkill hypothesis has been supported by the evidence of projectile points found imbedded in the bones of extinct animals, by bones of extinct animals found in association with archaeological sites (Brumley 1978, Frisson 1974, 1978, Hester 1967) and by a computer simulation by Whittington & Dyke (1989) based on the relative recruitment rates of H. sapiens and the various animals.

Watson is dissatisfied with this solution. She says, "I say, Holmes, Martin (1979, Mossiman &Martin 1989) is suggesting that humans found naive fauna which were easy to hunt and that they swept across the two continents of the New World in 1000 years, killing as they went, but there's something about it that just doesn't seem plausible."

Holmes answers, "That's because it runs counter to our sense of how the predator/prey relationship works. In formal models it is generally assumed that predators and prey are linked in a mutually causal loop (May 1968, Smuts 1979, Roff 1975, Hilborn & Sinclair 1979, Schaffer 1988, Gluckenheimer et al 1976, Caugley 1970). Thus, as prey population increase predators increase and as prey populations decrease so do predator populations. This makes it difficult if not impossible for predators to kill off all their prey. (Figure 1). In simple ecosystems scientists have observed the kind of oscillations shown in the formal model. In more complex ecosystems as a given prey species become scarce predators start hunting other prey which are more plentiful. This smooths out the oscillations and makes it even less likely that the prey will go extinct."

Watson studies the graph and nods.

Holmes continues, "If our notion of predator/prey relationships is correct, then as herbivores were over- hunted, human populations would have contracted. This would allow prey populations to increase. The overkill hypothesis suggests that despite contracting herbivore populations people continued to over-hunt. They continued to herd animals off cliffs and into arroyos, they didn't change their prey or their life style. This makes us feel dissatisfied because it suggests that humans don't behave like other predators, that they are somehow different. That goes against the principle of uniformitarianism - that the dynamics of today are applicable to past times."

Watson asks, "Does this allow us to dismiss the argument?"

"Well, it certainly indicates that there is more work to be done. The correlation of the introduction of humans and the extinctions is hard to dismiss. There must be more evidence which will allow us to verify, dismiss, or, perhaps, alter the model." Holmes replies.

Our detectives search some more. They enter what they found in their data base:

In the last 15 years it has become more and more evident that we cannot use current environments as a simple analogy for the environments of the Pleistocene (before 10,000 BP) because the pattern of where animals and plants lived is so different (Guthrie & Matthews 1971, Guthrie 1968, 1980, 1990, Ashworth 1977, 1980, Hoffman & Jones 1970). The literature from many different fields is full of references to "defunct species associations" (Slaughter 1967), "communities without modern or extant counterparts" (Matthews 1979), and "disharmonious species associations" (Graham 1976, Graham & Lundelius 1989). These all indicate that not only were there were more species but species which today have disjunct ranges (are allopatric) had overlapping ranges (were sympatric) during the Pleistocene.

"Holmes?" Watson asks, "I don't understand."

"Ah, Let me explain graphically."

She brings up a drawing program and says, "Suppose this is our continent. There are two different species who live on the continent. One is a warm loving species, signified by these beach balls and one is a cold loving species signified by these snow flakes. The range of the cold loving species is limited in the south by the heat of the summer. The range of the warm loving species is limited in the north by the cold of the winter. During the Pleistocene the species ranges looked like this (Figure 2 - Pleistocene Range) Now, if all that had happened was that mean annual temperature had changed you would expect that the ranges would continue to overlap but would move north and perhaps the change sizes (Figure - 2 - Expected Holocene Range). However what happened is that the ranges stopped overlapping which suggests that winters are colder, which limits the warm loving species and summers are hotter which limits the cold loving species (Figure - 2 - Actual Holocene Range). This is referred to as an increase in continentality."

range illustration


"Now this kind of thing also happened in other dimensions. During the Pleistocene there were species which today are confined to woodland who lived with species who today live on prairies. This suggests that the vegetative environment was less sharply partitioned between grassland and woodland than it is today. This is referred to as a decrease in patchiness or a simplification." Holmes continues entering data into the data base.

Continentality and Environmental Simplification have both been used to support hypothesis of extinction.


An elaboration of the climate change theory, the Continentality Hypothesis, suggests that the herbivores went extinct because they were unable to tolerate the extremes in temperature (King & Saunders 1989, Grahm & Lundelius 1989, Bryson et al 1970).

The problem with this hypothesis is that, although the New World climates are more continental today than they were in the Pleistocene they are not more continental than the Pleistocene of Siberia where megaherbivores prospered (Frenzel 1968, Flerow 1967, McDonald, 1989).

"Well, Holmes," Watson says, "That seems to dismiss that."

"Yes, It does. And besides, Watson, it would seem that large mammals would be able to migrate seasonally if they found the temperature too extreme (Pennycuick 1979). But, Watson, we must be careful not to dismiss too much. We have to hold on to our clews. We have uncovered another difference between today's environments (the environments of the Holocene) and the previous interglacials."

"Umm... Ahaa.. What, Holmes?"

"That in the New World today's environments are more continental."

"Ah, yes let me start a list file." As Watson works at the computer she says, "So now we know that there was a temperature change, an increase in continentality, and the introduction of H. sapiens. What next?"

"We have to examine the proposed mechanisms of extinctions related to these observed differences in the environment between the Pleistocene and today."

Reproductive Consequences of Continentality

Slaughter (1967) and Axelrod (1967) argue that, along with colder winters and hotter summers, rainfall became less predictable. This tended to restrict the amount of time which was favorable for reproduction. They suggest that large animals with longer, "inflexible" mating periods produced young at an unfavorable season.

Kiltie (1989) observes that the timing of gestation and the periods of available resources is more important than the actual length of gestation per se. He notes that an increase in continentality is due to a decrease in rainfall, which, in turn, reduces the reliability of rainfall. The main way for large placental mammals to adapt to this unpredictability would be to breed conservatively, so that young are produced at the most favorable period of time. This can occur when the animal relies on indirect cues to the probable quality of the environment tied to breeding, or if they have a sufficiently short gestation period so that adjustments to the seasonality can be made. He observes that extinct species have gestation periods which exceed one year, suggesting that they may have been unable to adjust their breeding period to take advantage of the most favorable timing.

The evidence of disjunct species associations is also used to support another view of Pleistocene environments - that they were more patchy than today's environments.


The shift from Pleistocene to Holocene environments has been described by Guthrie (1988) as a shift from plaid to striped environments (Figure 3).

Environments or biomes of the Pleistocene were plaid, i.e. more patchy less homogeneous then today's environments or biomes, which are by contrast striped with big homogeneous stripes. (Bernabo & Webb 1977, Coope 1967, 1977, Davis 1976, Delcourt & Delcourt 1987, Hoffmann & Jones 1970, Guthrie 1968, Hibbard 1970, Hopkins 1967 Lowe & Walker 1984, Matthews 1968, 1975, Morgan & Morgan 1979, 1981, Peck 1973, for a discussion of patchiness see Picket & White 1985).

Today's tundra is a product of permafrost. It is typified by cold, poorly drained soils. It could not support the complement of animals supported in the Pleistocene. The arctic of the Pleistocene, the Mammoth Steppe Biome (Guthrie, 1968), seems to have been a parkland with well drained soils supporting a large and diverse complement of animals dominated by mammoth, bison, horse, and their predators, (Guthrie, 1989, Batzli et al 1980, Calef 1984, Felerow 1967, Frenzel 1968, Fuller & Bayrock 1965, Hoffmann & Taber 1967, Hopkins et al 1969, Kurten 1972, 1988, Matthews 1975, Petrov 1967, Pewe & Hopkins 1967, Pruitt 1984, Repenning 1967, Ritchie & Hare 1971, Sainsbury 1967, Sorenson 1977, Yurtsev 1972).

During the Pleistocene the geographic region which is today associated with prairie or grassland was more of a mixed woodland/parkland. (Anderson 1974, Birks 1973,Graham & Lundelius 1989, Bryson et al 1970, Brumley 1978, Hibbard 1970, Hopkins et al 1969, Morgan & Morgan 1979, 1981, Wendorf 1970)

During the Pleistocene the geographic regions today associated with closed canopy forests were more patchy - a mixed woodland/parkland environment (Guilday 1989, Bernabo & Webb 1977, Davis 1976, Dillon 1956, Whitehead 1973, Wright 1976, 1984).

"OK, Holmes, Let me catch up with our list of what we have so far. We know that the tundra had better drained soils and was more productive and that the distinction we currently see between grassland and woodland was much less pronounced. Might that indicate that those biomes were more productive as well?"

"Very good, Watson. You anticipate me."

Nutritional Consequences of Simplification

Graham & Lundelius (1989), Guthrie (1989), and Guilday (1989) attribute the extinctions to simplification of the environment. Graham & Lundelius argue that because of the change in continentality and the decrease in the patchiness of the floral environment, herbivores were unable to find the plants with which they had co-evolved and thus fell prey to the anti-herbivory toxins of the plants they were able to find. Guthrie (1980, 1989) argues that extinction of large herbivores and the dwarfing of many others was due to changes in the length of the growing season. He notes that large ruminants (bison) fared better than monogastrics (horses and elephants) because they were able to extract more nutrition from limited quantities of high fiber food, and were better able to deal with antiherbivory toxins (Hoppe, 1978). In sum, monogastric ungulates were unable to find a sufficient mix of high quality food in the simplified environments of the Holocene and went extinct.

After considering for a while, Holmes says, "I like the observation that ruminants extract more energy from their forage. Certainly when I have seen cow pies and they have far less grass than horse or elephant scats, but somehow, Watson, I still see some problems with these arguments:

First, today there are feral horses successfully dealing with the same Holocene environments, finding sufficient mix of food to deal with toxins, extracting sufficient nutrition to live and reproduce effectively even though they have gestation periods in excess of one year. Second, season length is not simply a condition of temperature and humidity; it is also a function of latitude. Thus, by moving south, holarctic herbivores could have found areas with longer, more favorable growing seasons for food and breeding and survived. This is especially true of large mammals."

"Are you suggesting that there wasn't a change in continentality or the patchiness of the environment, Holmes?"

"No, that seems to be well documented. It's just that it doesn't seem to account for the extinctions. Also, no one seems to be thinking about the relationship between continentality and patchiness or why this interglacial is associated with a change in continentality and patchiness and other interglacials have not. What is different about this interglacial? What causes a simplification of the environment? Why did patchiness decrease?"

"Well, it..... Ummm, I suppose .... what do you think, Holmes?"

"I'm not sure. It seems that even after the change in temperature, patchiness can be reestablished. For example, Guilday (1989) observed that the red fox (Vulpes vulpes) initially retreated to the north as the mid-latitude Appalachians became a closed canopy forest but re-entered the area as colonial land clearing recreated patchiness. In the plains, now that great herds of bison no longer graze at will, trees are re-invading the long grass prairies indicating that a woodland could be re-established."

"Hummm.... bison, yes...well but there are bison."

"Yes, there are... Watson, why are there bison? I mean why are the bison as they are?"

"Why are bison....Holmes, do you feel all right?"

"Yes, fine. What I mean is why are bison grazers? What is it that makes them grazers? During the Pleistocene bison were different. They were mixed feeders. They were larger. They lived in smaller groups. Today they are grazers, living in large herds. In fact, although it is often overlooked, there are two kinds of bison and they fit the new simplified environments. There are woodland bison (B. bison athabascae) and plains bison (B. bison bison). But, leaving the woodland bison aside for a moment, why are plains bison grazers?"

"Well, plains bison are grazers because they live on the prairie and eat grass. What are you talking about Holmes?"

"Yes, but why is there prairie?"

"Well... well because... I don't know Holmes. What are you driving at?"

"There is prairie because of the bison. Without the bison, trees re-invade the prairie. Yes?"


"This is an example of a co-evolutionary equilibrium between the grass and the bison, like the maintenance of grazing lawns by herbivores first observed in Africa (Vesey-Fitzgerald 1961, McNaughton 1976, Leuthold 1978, Arnold 1964, Bell 1984, Jarman & Sinclair 1979, Macfadyn 1964, Maddock 1979)."

"Yes. This is not new, Holmes."

"But we're not finished yet. This doesn't tell us anything about how the relationship got started in the first place - what caused the simplification of the environment."

"No... Holmes, but I don't see why it bothers you. Everyone else seems to accept it."

"Yes, Watson, that's the problem, everyone accepts it."

She wanders away and picking up and putting down various books and papers which are spread all over the study. She picks up a book and then an article and finally starts entering data into the data base.

When Holmes is finished, Watson asks, "Well, Holmes, have you figured out what caused the simplification of the environment?"

"No, but I've found out more about how the current relationship between plains bison and the prairie is maintained." Watson looks at what has been typed and reads:

McDonald (1981, 1989) suggests that the animals that became extinct should actually have prospered during the shift from a mixed woodland/ parkland to prairie since their primary food source, grass, was increasing rather than decreasing (McDonald 1981, 1989, Birks & West 1973, Birks 1973).

He accepts the overkill hypothesis and suggests that bison managed to survive the overkill in widely scattered small groups. It was uneconomic for humans to search for these few bison. As populations re-established themselves, there was selective pressure on the survivors to adopt a strategy which protected them from overkill. The small size of the breeding population would allow them to benefit from the founder effect (any beneficial trait is more quickly incorporated into a small population than a large population). The prairie did not have enough resources to tempt humans to live on it full time. Humans would take a certain number of animals on the fringes of the prairie but the combination of the size of the herds and the vastness of the plains protected the species as a whole from extinction. He sees the morphological changes in the species as indicators of a shift to this new life style (McDonald & Lammers 1992).

"So, Watson, we can see that once the current ecosystem was established it would remain stable, or did remain stable until Europeans wiped out the bison herds. But, Pleistocene bison were mixed feeders, eating browse and grass, McDonald's stability theory doesn't tell us how the current relationship got started. His focus is on how bison have managed to survive overkill through a change in life style."

"What do we do now, Holmes?"

"I'm not sure. We have established certain things and dismissed others. We know that this interglacial - the Holocene is associated with extinctions, increased continentality, decreased patchiness, the introduction of humans, and a shift in bison morphology and life style. But we're not satisfied with any of the proposed hypotheses of extinction."

"Now, what happens?"

"Well, Watson, For the detective the job is done when the identity and the motivation of the culprit is known. For us, the job sill be done when we have a full-scale dynamic hypothesis. This should take into account all the evidence and paint a picture of a logically coherent world. According to Einstein and Infeld," she reads,

"We seem to be at that point. Unfortunately, Watson, I neither smoke nor play the violin." She sits back in her chair and thinks.

Watson fidgets some and then leafs through some of the scholarly papers on the floor. She finds one and reads, then she looks up at Holmes who is still in deep thought.

"I say, Holmes. There was a suggestion that humans were able to continue hunting because they could subsist on gathered food. They didn't depend on hunted food to keep them from starvation, so the ordinary predator /prey relationships don't apply."

"Yes, yes. I know." Holmes says abstractedly, "It doesn't really help. As prey became scarce and harder to find they still would have had to invest more and more energy to find the animals. If they had enough food they would eat the food they had instead of spending energy hunting. If they didn't have enough food then we're back with the original difficulty. Either way it wouldn't be energetically economic."

"No, it makes it sound like humans wanted to kill off all the herbivores - that they hunted herbivores as if they had a vendetta against them. Silly, really."

"What!" Holmes says suddenly, "What did you say?"

"Silly, really?"

"No, before that!"

"They had a vendetta?"

"Yes! A policy!

"A policy, Holmes? Why would they?"

"No, no, of course not. Never mind, I'm going out." she says, grabbing for her muffler.

Over the next few days Watson finds Holmes reading books and papers from many different fields. She looks at the new entries Holmes has added:

Climate - Continentality

Continentality is largely a factor of the relative aridity of the environment. Where clear cutting has occurred in existing rain forest of South America and the Northwestern United States there has been a considerable reduction in humidity. (It is estimated that 50% of the water in a rain forest is recycled (Deshmukh, 1986, Jacobs 1988, Longman & Jelmnick 1987)

Concluding note: Vegetation influences the degree of moisture and hence the relative continentality.


Where there is sufficient land, elephants create and maintain patchiness of the environment. Wing and Buss (1970) studied the forest elephant and found that elephants contribute to landscape building in three ways:
  1. They knock down older trees which allow smaller animals to reach the leaves
  2. They create gaps which allows grasses and seedlings to establish themselves. This floral diversity contributes to animal diversity. (Multi-storied forests support more animals than single- storied forests.)
  3. Elephant feces, urine, and the action of their feet contribute to soil formation and fertility. Although they eat a great deal of vegetation much of it is returned to the environment.
Where there is not sufficient land, elephant populations become dense and turn forest into grassland. Wing and Buss (1970) observe that the combination of fire and elephants turned forest into a grassland in only 50 years (Anderson & Walker 1984, Barnes 1985, Laws 1970, Leuthold 1977, Yoaciel 1981, vanWyk & Fairfall 1969, Short 1981, Walker et al 1987)


Current day tundra is a product of permafrost. Permafrost produces boggy and poorly drained soils. Tundra vegetation must not only be able to withstand extreme cold, it must be able to get enough nutrition from shallow cold soils where only a small top layer has either sufficient usable oxygen or sufficient nutrients (Brown et al 1980, Chapin et al 1984, Colinvaux 1967, 1978, Gersper et al 1980, Giterman & Golubeva 1967, Hoffman & Taber 1967). Therefore there is not sufficient vegetation to support the wide variety of herbivores.

In the arctic there are areas where trucks and bulldozers have broken up the permafrost. This has resulted in better drainage and increased productivity and diversity of the plant and animal community (Kershaw 1984, Webber et al 1980). These new, truck-produced communities are made up of flora and fauna which are normally disjunct. They become stable communities at different times depending on the severity and timing of the impact and the surrounding environment (Webber et al, 1980).

An increase in diversity and productivity also occurs where entire herds of caribou trample vegetation on narrow trails which they use to get to new ranges or to escape mosquitoes. (Caribou migrating across a wide landscape do not disturb permafrost because their foot loading is light compared to similar sized animals in other environments.)

Concluding note: The African and Arctic evidence taken together suggest that big animals cause patchiness.


At the end of the last century the introduction of rinderpest severely reduced herbivore populations. There was a rash of man-eating lions. Populations stabilized with both herbivores and lions at a lower level. With the eradication of rinderpest in the 1960s the herbivore populations recovered and the number of lion prides in the Serengetti increased from 18 to 24 and the number of lions per pride increased from 15 to 19 (Schaller, 1968, 1972, Sinnclair, 1979, Scinclair & Norton Griffeths, 1979, Hanby & Bygott, 1979, also generally Bertram 1979, Bourliere & Verschuren 1960).

Hypothesis: prey are removed and predators die - Overkill (Figure 4)

Hypothesis: predators are removed and prey increase then die - Second Order Overkill (Figure 5)

Concluding note: This suggests that there are two ways in which overkill could have caused megaherbivore extinctions. Either through direct means - killing herbivores or indirect means - killing predators. We must examine the consequences of each to determine which might be true!

A few days later Watson finds the litter of books and papers is gone and Holmes is writing in the study. Watson is about to leave her alone but Holmes motions her to come in and sit down. "I've just about finished. Fix us a drink so we can celebrate."

Watson fixes the drinks and pulls up their easy chairs by the fire. Holmes prints out her file and joins her. "Well, here's to the Pleistocene."

"So, you've solved it."

"Yes, I don't know that I can prove it. But I know what happened."

"Can you tell me?"

"Yes, I realized early on that the crux of the matter was the patchiness. Everyone was trying to explain the extinctions. No one questioned how the change in patchiness occurred.

We know the change in patchiness isn't accounted for by temperature change because patchiness didn't change in previous interglacials. We also know that continentality is partially a function of relative humidity and that relative humidity is partially a function of the number of trees.

The African evidence suggests that patchiness would be re-established in Holocene forests if mammoth and mastodons were re-introduced. Introducing mammoths to the arctic tundra would break up the permafrost, producing better drained warmer soils and greater floral and faunal diversity. In short, mammoths would change the nature of today's tundra and forests."

"But wait, Holmes, I know you are basing the tundra part of this on the evidence of trucks driving on tundra and creating well drained soils. But can you compare trucks to mammoths and mastodons? Trucks aren't biological."

"Ah, but the analogy we are drawing is between trucks, as heavy bodies on the tundra, and mammoths, as heavy bodies on the tundra. The principle of uniformitarianism says that the forces or mechanisms have operated uniformly. Thus, a heavy body will impact the land in the same way as another heavy body. If we now went on to say something about the digestive system of the mammoth based on trucks we would off base, but as heavy bodies they are similar."

"I see."

"So, where was I?" she consults her notes, "Oh yes, ....simple to create patchiness in the forest and in the tundra but it is a different matter all together to create prairies so we have to proceed carefully.

It seems clear that the prairies of today are maintained by today's bison. As bison and megaherbivores are removed trees reinvade. Thus, logically, if there were no bison and no megaherbivores there would be uniform tree cover.

"Y.. y.. yes? But why go about it in such a round about way? I don't understand."

"Let me try another tack. Let us consider the overkill hypothesis in relation to prairies. Suppose, Watson, you have a mixed woodland/ parkland Pleistocene environment and you remove all the mammoths in a short period of time as suggested by the overkill hypothesis. What happens to the trees?"

"The trees ... the trees multiply and fill in the gaps as they did in the Appalachians?" (Figure 6)

"Exactly, Watson. And the bison?" Holmes asks.

"The bison continue to be browsers and grazers? Or, perhaps the bison become more adapted to browse because there is a net increase in trees."

"Right, good."

"But that's not what happened, Holmes."

"No, but we have managed to advance our inquiry. And, if you buy the notion that mammoths create and maintain patchiness you have thrown serious doubt on the overkill hypothesis because there is prairie and bison are not mixed feeders or browsers."

"Oh." Watson says tentatively.

"Current day bison are a confusing factor if we are mentally trying to recreate the environments of the Pleistocene because they are a product of the shift in environments. Later they will be a clarifying factor again as they indicate how the shift must have occurred."

"I suppose so."

"Yes. Look, there was a stable patchy environment in the Pleistocene, and there is a stable simplified environment in the Holocene - Guthrie's shift from plaid to striped environments. Right?"


"Both Pleistocene and Holocene environments are able to support bison and horses. Bison made it through the shift, horses didn't. This suggests that there was something during the shift, but not before or after, which was stressful for horses. there was some kind of a bottleneck."

"Because there are horses in both environments."

"Exactly. Now, bison made it through the shift by changing their life strategy. They went from being mixed browsers and grazers to being grazers. This suggests that the shift exerted pressure away from browsing toward grazing as a bison life style."

"Yes obviously. But, what happened, Holmes?"

"Well, this is where I was stuck when you made your statement and it all became clear, Watson."

"Which?..The statement about a vendetta?"

"Yes, or a policy. If we look back at our elephant data from Africa we see that elephants change environments in two ways. When their population is not too large they knock over trees and maintain patchiness, but when the population of elephants is too large for the geographic area they experience food stress and turn forest into prairie in an amazingly short period of time. Now, given that, if we think about predator/ prey models, what is one way in which prey populations could become too large?"

"If there were no predators?"

"Yes, and this is where the notion of policy comes in. Only human beings can have a policy of killing. Imagine this scenario.

Homo sapiens enters the New World

Holmes looks up from her notes at Watson with a grin, "Well what do you think?"

"Wonderful, Holmes, It accounts for why there can be horses today and there could be horses during the Pleistocene but that horses went extinct during the shift. They couldn't compete with the bison who could utilize the earliest grasses."

"Right, Watson. Horses require more breeding time and a higher quality of forage. If some horses did survive and reproduced, their offspring would be born into an environment which was already dominated by bison. Once the bison populations stabilized, there would have been room for the horses again but they didn't make it that far."



"How did the humans kill off the predators?"

"Well, if they had a policy they might trap, poison, search out dens and kill off the young."

"Yes, but the carnivores who are extinct are those that were more specialized not those who would be likely to eat poisoned food. The generalists like wolves are still around.

Also, I wouldn't have thought that mammoths would have recruited that fast, (Blueweiss et al.1978) or that predators would be what kept mammoth populations in check. Doesn't this throw doubt on the model?"

"Ah, very good. It is true that today elephant populations are controlled more by environmental quality and the amount of range available more than by predation. There is evidence that Dinobastis, the scimatar cat hunted young mammoths (Agenbroad 1989, Evans 1961) so mammoth populations, unlike elephant populations, may have been affected by predation. But, it doesn't matter, the model works even if predation was not a major factor in controlling mammoth populations, and even if the only predators hunted were generalists.

Let's do a thought experiment. If the wolves are the major control of bison and if wolves are hunted down to small relict populations then what happers?"

"Well, then the bison populations would explode."

Exactly, Watson. The mammoths need not have caused the food crisis, even though they knocked down the trees. In a stable environment the animals partition the environment into a grazing succession. Some animals prefer to eat the dry overburden, some the grass sheath, some the new second growth shoots and so on (Vesey-Fitzgerald 1960, McNaughton 1979, 1984, McNaughton & Georgiadis 1986, Bell 1971, Hirst 1975, Norton-Griffiths 1979). But if there were an increase in the populations of animals in competition with mammoth, like bison, the neat partitioning of the grazing succession would break down. All the animals would eat what ever they could find. There would have been increased use of all resources. Mammoths would have eaten more and more browse as they were excluded from grazing resources by animals with faster recruitment rates, such as bison and horses. As the mammoths are more browze they put pressure on mastodons.

This may also account for some of the extinctions which don't seem to be related to either climate or to over-hunting. Hansen's (1978) study of the diet of the Shasta ground sloth shows that over time the animal ate less and less of various dietary supplement plants and less of its major staple, globemallow (Sphaeralcea ambigua) while it ate more mormon tea (Ephedra nevadensis). Mormon tea is not eaten by most herbivores possibly because of more effective antiherbivory defense compounds. This suggests that the animal was unable to find the food it needed to maintain a balanced diet and instead ate to fill its stomach and so died." (Phillips 1988)

The bison made it through the bottle neck because it was able to survive on less. The wolf population could re-establish itself whereas the predators who were specialists had lost their prey. The extinction is a result of the environmental exhaustion not directly due to what had been hunted."

"Yes, I see, and I like the model. I think that's how it must have happened."

"Yes? but?"

"But aren't you doing the same thing Martin is doing? Suggesting that humans do not behave in accordance with the usual predator/ prey relationship?"

"I suppose it could be seen that way. But I don't think I am.

Martin's model suggests that humans were killing off their own food supply, but that this didn't affect the size of human populations. In this way he ignored the dynamics of predator prey models. I'm not suggesting that they killed off their food supply. I am suggesting that humans eliminated animals which were seen as competitors and a threat. This is consistent with what we observe of those same animals today. Let one lamb or cow be killed and ranchers are out trying to put a bounty on mountain lion.

We know that humans in Europe killed carnivores from the evidence of cave paintings. Now, why did they kill them? It wasn't for food - in animals that high up on the food chain the build up of vitamin A is toxic to humans so for the most part it isn't economic to kill predators. They take as much or more energy to hunt and are less useful.

Martin doesn't mention carnivores and we are left with the impression that with their preferred food supply dwindling predators just slipped off to die. When rinderpest reduced the herbivore populations in Africa lions, who were sophisticated in the ways of humans, turned to eating humans. This would be much more likely for animals which had no inbred fear of humans."

"Yes that's true." She thinks a while then says, "Holmes?"


"What about the human populations? You said that they would have crashed too. But they didn't become extinct."

"No, but their culture did which is what you would expect to find if their populations were sufficiently reduced. The paleoindians - the people of the Pleistocene - made beautiful fluted points. Similar points have been found throughout the New World suggesting that there was a uniform or connected culture. However, after the close of the Pleistocene the points disappear as does the uniformity of style. The points of the Archaic, period which follows the Pleistocene, are not as well made, and styles are different in different geographic areas. This, together with our theory, suggests that the uniform strong culture was destroyed and disorganized by the shift. The remnant populations had to learn their skills all over again."

"A paleo dark ages?"

"Yes, perhaps. If this was a general pattern at the end of the Pleistocene it may account for the buffalo jumps we see in later times. The people may have realized that bison herds could become unstable if they were allowed to breed withoug any kind of control. It may also be what set the stage for those who turned to horticulture and pastoralism - because they suddenly couldn't trust the animals to be there when needed. But, this is speculation and not part of our problem."

"I like the model, Holmes, but can you prove it?"

"No, but we can present it. And we can suggest questions, experiments, and observations for scientists in various fields, which will tend to support or falsify the hypothesis that the extinctions caused a simplification of the environment through a boom/ bust mechanism. Pursuing these will take us further in our understanding."

First, it is possible to do computer simulations which will test the assumptions which have been made.

Next, there are experiments which test the environmental relationships.

  1. I propose an experiment based on some of the relationships specified in the model. Current day tundra is a product of permafrost. We know trucks break up permafrost. If we designated a given area and had trucks drive over the tundra, simulating the action of mammoths, and we reintroduced bison and horses, could we reproduce the Mammoth steppe?
  2. A mathematical experiment which would shed some light on the relationship between trees and climate. Contentality is a function of the relative aridity of the environment. Where clear cutting has occurred in the rain forests of South America and the Northwest there has been a considerable reduction in humidity. (It is estimated that 50% of the water in a rain forest is recycled (Deshmukh, 1986, Jacobs 1988, Longman & Jelmnick 1987).

    If we measure the amount of humidity a tree gives off into the atmosphere and multiply that by the estimated number of trees per square area, we can estimate how many trees needed to raise the humidity of the Great Plains sufficiently to reduce contentality. Does the hypothesized woodland/ parkland of the Pleistocene fit this model? Could species which were sympatric during the Pleistocene and allopatric during the Holocene live together in this environment?

  3. Next, the model presented here proposes a definite sequence of events relative to the evolution of modern plains, B. bison bison and wood bison B. bison athabascae. Therefore observations which are out of this sequence in time or in phylogeny will falsify this model.

    1. Time If extinct predators are found in association with modern bison or if they date to the same time period, the model would be falsified.
    2. Phylogeny There are two major hypotheses of bison phylogeny:

      The Beringian hypothesis suggests that modern wood bison B. bison athabascae originated in Eastern Siberia or Beringia, dispersed southward, and represents either an end point in evolution, or gave rise to modern plains bison B. bison bison (Skinner & Kaisen 1947, Guthrie 1970, Flerow 1971, Harrington 1984, van Zyll de Jong 1986, Fuller & Bayrock, 1965.).

      The New World hypothesis suggests that both B. bison bison and B. bison athabascae have a common New World origin from Bison antiguus (McDonald 1989)

      These hypotheses can be tested using cladistic methods using Bos as an out group to polarize traits of extinct New and Old World bison and extant New World bison.

      If modern bison are found to be more closely related to Asian bison than extant or extinct New World forms - the Beringian hypothesis - they would have evolved prior to the closing of the Bering land bridge. This would suggest that the selective pressure was not a consequence of the boom/ bust scenario proposed here and would falsify the model. (For discussion of issues related to bison phylogeny McDonald 1981, Hughs 1978, Guthrie 1990, Kowalski 1967, Krasinski 1967, Kurten & Anderson 1980, Reher 1978, Borowski et al 1967, Banfield 1960)



"None of these constitute proof."

"Well, what do you mean by proof?"

"Well.. you know the truth. That it's accepted."

"Ah, but science is not about truth and it's not about how many accept it. It's a way to approach truth - it's a process."

"Like a conversation?"

"Yes. One person offers an idea or explanation and others try to falsify it. Science doesn't move by support it moves by falsifications. We have suggested some new ideas.

We have suggested some tests and, over time, others will suggest more tests. The theory will be changed as others make it their own, or, of course, it may be falsified. Either way the conversation of science will move on."

"Holmes, the implications could be interesting."

"Yes, among other things it implies that there is no primeval pure environment. Humans have altered the global environment before and come close to extinction, and we could do so again. If we do, it will be the same kind of indirect means."

"Yes. Well, Watson, it looks like it's time for bed."

"Yes, good night, Holmes."

"Good night."


Return to the Extinctions Menu Page preloaded