Real Audio Recording   

David Hurst Thomas:

. . . would someone want to just launch into this?

Russell Graham:

I guess I have a question for Paul Martin. I'm still bothered by the idea that we don't find Clovis sites with other extinct taxa. The early arguments in the 1960s is that this was site visibility -- Gary Haynes has alluded to this -- but I don't think that argument really holds any water anymore. Because all of the Clovis sites that we do find are associated with the proboscidians -- and as we're suggesting in our paper, the proboscidians were the last to go extinct, and perhaps they were the only things here for the Clovis people to hunt in any numbers. Any response?

Paul Martin:

Yes, I have to live with that. Incidentally, in back of what was already said by Dave Thomas, I want to take this opportunity to mention one more aspect of this 30-year traveling road show debate about the cause of extinction at the end of the Pleistocene. And that's the collegial aspect of it, which I am very impressed with. The people that I've run into -- and some of them are bitter adversaries at the academic level -- have proved good beer-drinking friends at the bar level. That is a precious feature. But Russ's question has to be dealt with now. And he's right on. There are no good associations with ground sloths. I'd certainly make a lot of arm wave about it if there were. However, I'm also impressed, as I indicated, with the fact that this is a unique time in the fossil record. And John Alroy has amplified that argument in a magnificent way.

So that drives the argument against the need to find. . . . The reason this idea occurred to ecologists first, and not to archaeologists, was because ecologists weren't smart enough to know that you had to have a kill site to make an argument for an extinction. But maybe the archaeologists and the paleontologists overlooked the fact that human impacts could have amplification at some level -- and this plays straight into Ross MacPhee's search for what seems so improbable, that disease factor introduced into new lands. Or maybe its amplification by a predator pit, as was suggested by Norm Owen-Smith, in addition to his earlier suggestion that he's published about amplification of habitat change by the extinction of mammoths. I'm arm-waving again -- that's where you end up with on this business. The question that was asked is a good one.

David Hurst Thomas:

I'd like to just make a comment about a point that Dr. Graham just raised. One thing that concerns me about the discussion over terminal dates is that the terminal dates by themselves don't mean anything, unless you have what you might call "background dates," or sort of positive absences -- dates for localities where you're sure that stuff isn't around. And that's been a big problem in other parts of the paleontological record, where people have been talking about extinction events, and haven't really dealt with the question of sampling before and after. So with respect to this issue of what was the exact order of extinctions, I would like to see some discussion of the quality of the records for the other megafauna -- the horses and camels, and so on -- and some discussion of whether you would expect to have found them if they had been around in sites that could date their apparent extinction events. That's not a criticism -- I'm just asking for more data.

Russell Graham:

Actually, we have that data, and we went through it pretty fast. But in the horse data, in particular, it's very impressive. I mean, the whole purpose of this project -- it was an NSF project to actually try and date terminal Pleistocene extinctions. So that's why we looked for horses that were right at the boundary, or tried to find their highest stratigraphic occurrence. There are lots of other background dates that you could add to that, which would show that they were around well before the terminal Pleistocene extinction. But in that graph, we did date a bunch of horses that presumably were Holocene survivors. And this has been a large argument -- in that we get people calling us finding these in Indian mounds, in a variety of things, and that they survived into the Holocene. Every one of those horses that we dated was modern -- and we've got that data. And, in fact, one of the Pleistocene horses -- a type specimen of Equus laurentius -- we dated is modern. And the same thing is true for the yak in Alaska. Dale Guthrie has dated all of the yak specimens; they are all modern cows. So I think there's a lot of data. And I think one of the strongest arguments, actually, for the really tight chronology is the archeological record, is that if you look in the Holocene sites, there's tons of dates on these, and very few of them, if any of them, have extinct fauna in them. So I think it is a very tightly constrained, very rapid event.

David Hurst Thomas:

Well, what I'm specifically after is the density of the record of these particular species before the event. If you have a species that's very rare before the event, and its terminal date is a little bit old, you simply would not -- that's what you would expect, based on sampling. It might have nothing to do with the actual date of termination. So, to answer that question, you need to understand: What is the probability of sampling that particular species? And, to do that, you need to have the background dates. Not just the terminal, the highest stratigraphic level dates.

Russell Graham:

I'd agree to that with a certain extent. I mean, part of this also depends upon the fossils themselves. Because there are a particular number of taxa -- a thing like Homotherium, the scimitar cat -- which is a very rare fossil here in North America. And we could date all of those specimens, and we may not get a so-called terminal date. So, yes, it is a problem.

David Hurst Thomas:

These seem the kinds of issues that are dealt with by people studying other extinction events, like the K-T event, and they turn out to be very important for understanding whether events are truly catastrophic or staggered. Anybody else want to jump in?

Q: This is a question to Russell Graham, to ask: What is it about climate that could potentially produce patchy or heterogeneous habitats?

Russell Graham:

Well, I think there are two potential explanations for that. One of them are the arguments that we've generated before -- what's called the "equitability model" -- and that we've tried to explain these nonanalog faunas, that the seasonal extremes were reduced during most of the Pleistocene. The basis of that argument is that the ranges of species are controlled by seasonal extremes, not mean annual temperatures nor mean annual precipitation, but it's seasonal. And so if you relax the seasonal extremes -- particularly if you relax the summer extremes -- and I think that's the important point, is maybe the winters were the same. But if you relax those summer extremes, you'll allow those northern taxa to disperse further south, and then you get an integration of those.

Now, how would an equitable climate produce a more patchy environment? I think the explanation of that is that you've relaxed the climatic effects of it, and you've allowed edaphic factors, and competition, and other biological and physiographic aspects to be more of a limiting factor on the distribution of taxa, which would allow for sort of a more fine-grained patchy type of environment.

I think there's another possible explanation for that, which relates to that Greenland ice core, in that you see in the Pleistocene, throughout the Pleistocene, this repeated and very rapid fluctuations in climate that you don't see in the Holocene. And I think, by having those repeated fluctuations, that you're continually stirring the environment, and so that you're again creating much more patchiness, much more heterogeneity by the stirring factor.

Q: I don't see any connection between temporal and spatial patchiness. That's the connection I can't make. I can see the edaphic variability -- something about geological processes could make soils more variable, it could come in. But, more generally, from a botanical point of view, disturbance is fire, herbivore impact. Those are the kinds of things that produce the small-scale patchy heterogeneity within landscapes, within some kind of a broad pattern by climate, including its variability, which is regional.

Russell Graham:

Well, I don't think the climate is totally regional. I think we're looking at a global scale here when we're looking at those ice cores. I mean, you see regional variation, and we actually see that when you look at the distribution of taxa. And I think that's one of the factors that limits these distribution, is that we see gradience in moisture, and we see gradience in temperature -- that would, again, separate these faunas. And we can do a cluster analysis, where you get these provinces, but the provinces show a much greater variety of composition. So I think you do see the greater patchiness, due to this continual stirring and edaphic characteristics, and biological interactions.

I would agree with you that the large animals were a major factor. And one of the things we do see in looking at the Pleistocene is that the forests of eastern United States were much more open than we see today. And if you look at the data that were published by Eric Grimm -- again, George Jacobson, Eric Grimm -- what they show is the greatest rate of change in vegetation for the last 18,000 years ago started at 11,000 years. And I've asked them, one -- I think that that's initially driven by climate change, but I think the extinction also is a major factor, and that you have disrupted the communities. And so the evolution of these Holocene communities is both a factor of climate and a factor of the extinction of the megafauna. Because it's not only the disruption of those environments, but the ability to spread nutrients, seed-dispersal agents, a wide variety of things that could have accompanied the extinction, based upon the animals.

Ross MacPhee:

Gentlemen, I wonder if I can intervene for a moment. In order to fulfill the expectation that we're also going to talk about more recent extinctions, I'd like to pose the following point to Norman Owen-Smith and to Russ Graham, in particular. Clare Flemming and I recently finished a paper on recent extinctions of mammals -- mammals being lost in the last 500 years, a précis of which appears in last month's Natural History magazine. Of the 90 or so species that disappeared in that five-century period, 8 or 9% would qualify as megafaunal under Paul Martin's definition, somewhere around 100 pounds or better. I don't see a single case in which range reduction -- in terms of continental-style extinctions, as opposed to insular -- I don't see a single case in which range reduction seems to have made any impact whatsoever.

I'd like to also ask for your observations -- and, perhaps, particularly from Norman, since he deals with these animals -- why he thinks it is that, in the last 500 years, there have been no extinctions among the groups that were the heaviest losers -- by proportion or by number of species -- at the end of the Pleistocene. Thinking here particularly of Perissodactyls -- not one dead; not one has died out in the last 500 years. I'm thinking of Artiodactyls -- three or four species having disappeared. Overwhelmingly, what we have seen disappearing around us are the smaller mammals -- bats, insectivores and rodents. How could it be, accordingly -- if it is correct that range reduction, and that sort of impact, was chiefly responsible for these Pleistocene reductions -- that we don't see an analog of that today, given the amount of change and modification of habitats worldwide that's been committed by our species?

John Alroy:

Could I make a very quick comment on that question? I'd like us all to remember that captive breeding and creation of reserves has focused, for the last hundred years, on protecting the very large mammals that are the focus of the question. There are many large-mammal species -- such as the mountain zebra, Arabian oryx, etc., etc., etc. -- that would certainly be extinct, and the American bison is in that category, if it weren't for heroic conservation efforts over the last century.

Russell Graham:

Well, I guess I would argue just the opposite of you, and that your data have shown that range and habitat are very important, because what you're looking at is extinctions on these island faunas, where you have very constricted geographic regions. Now --

Ross MacPhee:

Yes, but we're not talking about island extinctions this morning -- we're talking about continental-scale extinctions. I don't see that you can have it both ways. I can see how you might want to extend the arguments developed earlier by Norman, that you have all of these virtual islands of suitable habitat, in which the megaherbivores could survive. That gets impacted, in some degree, in some fashion, by climate change -- and then they're gone.

What I would like you to point out to me -- what I would like you guys to point to are cases in the record of the last few hundred years in which things have been brought to grief as a result of range reduction exclusively. Now, that doesn't mean that there couldn't have been overexploitation as a component of that -- but the notion that, by driving larger mammals into smaller areas, that you tip them down the slippery slope.

There are countervailing cases -- and here I'm going to respond to John Alroy's points. Of course, you can look at bison as an example of recovery on a scale that is nearly unimaginable. The figures that are quoted in the literature is that in the 1820s there were something like 60 million bison running across the Great Plains of Canada and the U.S. By the late 1880s there were 500 or 600. That is an incredible scale of loss. It is true that, thanks to conservation efforts, they were brought back. My point being, however: Can you frequently, in the natural scheme of things, reduce populations by tens of thousands of percent and have the species rebound -- again, under natural conditions? Are there examples of that? Well, cheetahs have been pointed to as one possible example of having gone through an extremely narrow bottleneck and having come back under natural conditions. Northern elephant seal -- another good example, its population having been brought down to a few hundred.

My point in saying all of this is that it actually seems like some of these large mammals that we know about today can withstand the most incredible scales of impact and still bounce back, whether or not there's a great deal of human mediation. So what I want you to think about, as you build your patterns for the Pleistocene: How it is that these Pleistocene mammals seem to have been so much less successful in rebounding. That's the part I don't get.

Russell Graham:

Ross, may I respond to that? I'll give you three examples, all from Eurasia: Bison bonasus, Bos taurus, and Equus caballus. All three became, after a protracted period of range contraction, became extinct in the wild, the most recent being Equus caballus. It's possible that caballus bactrianus is about to that point.

Now, before they became extinct in the wild, they were taken into captivity. They survived as wards of humans at this point -- but it seems to me those three or four examples are the kind of thing you were asking about.

Norman Owen Smith:

My perspective is going to be African, and, obviously, the kinds of extinctions we can look at mostly are going to be local extinctions, because there haven't been many rangewide extinctions. But I want to mention one of those first, and that is, of course, the blue antelope, which had a very restricted distribution around the southwestern Cape -- which is one of the first large-mammal species to go extinct in Africa following colonization by Western humans.

A very good recent example is one that I mentioned briefly in my presentation -- the roan antelope, which is a small marginal population near the edge of its range in Kruger Park, that has been there for a long time. And through causes not directly related to humans is now on the verge of extinction -- a combination of climatic stress, local pressure by predators -- and is now down to 50 -- hasn't gone extinct sitting at that level. . . . I mentioned some species had gone extinct locally within the area. There were small populations -- with the habitat change the [inaudible], one of the species has gone extinct locally. But they occur elsewhere where habitat conditions are suitable. So I think one could find a number of examples of small populations popping up. Hunter's antelope in east Africa is another one that they've had to translocate to avoid extinctions.

Ross MacPhee:

I really don't want to bring this conversation only to the question of modern analogs, but please -- let's use a terminology that we can all agree on. A "local extinction" is not an extinction -- that's an oxymoron. An extinction is when everyone within a species is gone; there aren't any more. If we want to talk about local diminutions, then I think we should use the word "extirpation." Of course, you can list off hundreds of examples of populations that have been brought down to very low numbers, have disappeared locally and so on -- but that doesn't necessarily signify what I'm talking about, or doesn't lead you to a necessary conclusion that they must go extinct. That is, of course, party line.

My point once again -- this is the last time that I want to beat you over the head with it. What I would like you to do is to go back now to the Pleistocene and tell me why it is that elephants all over the world suffered extinction, to the extent that you have two surviving species. There's something about that that doesn't make intuitive sense. If they are the surviving megaherbivores that can survive under almost any conditions in Africa -- they have no predators except us -- how is it that you have a number of species disappearing, apparently rapidly, in the time periods that you've been talking about today, all across the world? In tropical regions -- which, in fact, we didn't talk about today, but they certainly disappeared from tropical parts of the Americas. They also disappeared from high latitudes. They disappeared in places where they were integrated with one set of faunas and with other sets of faunas; in patchy areas, nonpatchy areas. You get my point.

Norman Owen Smith:

Just to clarify the mechanisms . . . what you have to realize is that the process of extinction involves a series of local extirpations, eventually ending up with extirpating the last population of that species -- not a fundamentally different process. Extinction of a species occurs through a whole set of small populations popping off one by one. Normally that does not lead to extinction, or recolonization can occur -- population models we develop in conservation biology.

So to understand the process of species rangewide extinctions, we have to understand what causes the sequence of local extirpations, one by one, until there's nowhere, within the whole range of -- former range of that species -- where any population can survive. Now, we have very few examples of species extinctions in recent times -- none of large mammals, excepting, you know, in Africa the blue antelope, which probably wasn't a species.

Ross MacPhee:

But we have many examples of range reduction -- that's my point. Range reduction frequently leads to extinction. Why has it not happened already? Why is it always something for the next [inaudible]?

Russell Graham:

I think that relates to the magnitude of the range reduction, and it's a series of maps that I didn't show. But if you'd like to look at this in really deep time, there are species that, you know, basically went up into the Arctic regions, and by the end of the Pleistocene they're restricted to basically the Gulf coastal regions. So I consider that a significant range reduction. Also, the climate changes of the Late Holocene -- the last 500 years -- are nothing compared to what you saw in the Pleistocene . . . because they're very significant. The frequency of those changes, the magnitude of those changes, you know, have been very different from the Pleistocene to the Holocene. And so I think that what we're seeing is a very long-term trend. I think the extinction probably was well underway during a good portion of what we call the Rancho La Brea in the last 300,000 years, and it's just culminated at the end of the Pleistocene because the ranges had reduced so much, and the heterogeneity of those areas had changed so much. So what we need to do -- and if NSF is listening -- we need to do a FAUNMAP that captures the rest of the Pleistocene, so we can look at the changes that occur at that time.

John Alroy:

I've done that already. And my analog on the FAUNMAP database shows. . . . One thing is that the dating is extremely difficult when you get past the point where thermoluminescence and carbon dating doesn't work anymore. There are very few decent dates for the North American Pleistocene before the last 100,000 years. That makes it very hard to talk about sequential range changes.

Another point is that I showed a map with my localities for a 5-million-year interval, and then Dr. Graham showed a map with his localities for the very last bit, where you do have the carbon dates, and his map was a lot better than mine. Okay. The point there is that although my sampling is okay, it's not even vaguely comparable to the FAUNMAP sampling intensity. It would be much harder to establish, with any type of certainty, what the geographic distributions of any species were at any one particular time prior to the last glacial.

Russell Graham:

Well, I wouldn't agree with that, necessarily. You can see -- and I think it was obvious in your data -- that with increasing time our sampling goes up. Because I think we've got a thousand sites that date to the Late Holocene, which are basically archeological sites, so we do have a better sample of that. Our Early Holocene -- to a certain extent, our Mid-Holocene and the Late Holocene -- we have very limited samples, and then the late glacial, at 10 to 15,000, we have very good.

And I would agree with you, and I think it's one of the problems with your analysis, is that the dating, actually beyond 40,000, is very difficult -- and you're lumping things that go from, say, 1.6 million up to maybe 200,000. And I don't think you would see any change in looking at it in that big a time slice for the Pleistocene, because you're incorporating a good portion, then, of the previous glacial-interglacial cycles.

And the other point to be made is, it's been argued that the Pleistocene is the same throughout, and that we see all the glacials and interglacials the same. That's not true. In the earlier parts of the Pleistocene, the fluctuations were on a 40,000-year cycle, and it's only for the last five or six glacials that we see a 100,000-year cycle. So I think there are differences.

The age dating is a real problem, and there are other ways to maybe get around that -- they become somewhat circular -- is using biochronology, like microtine sequences, things like that, where you can divide them up into relatively thin time slices.

I would agree that we're not going to be able to get the temporal resolution, and that's why we restricted the FAUNMAP database to 40,000 -- because even the 3 or 4,000-year slices we have are very gross compared to what we showed today, where the extinction probably occurred in 400 years radiocarbon, which may be 250 years real time.

John Alroy:

I agree that it was a wise move to restrict your database to that interval because of the dating issue, but for what I wanted to argue -- which has directly to do with this issue of whether there were range contractions -- to the extent that I have lousy dates and mixed-up sets of faunas, that's going to create associations. Or it's going to create a relatively not larger number of disharmonious associations, if there were any to be found. And that's because, in the category I called "early and middle Pleistocene," I'm including interglacials and glacials; I'm including the older Pleistocene interglacials and glacials that were less severe and were on a different periodicity; and the later ones together. And if Dr. Graham is right, I should be including intervals when geographic distributions were at their greatest extent -- thereby creating a lot of disharmonious associations. But that's not what you saw. Relatively speaking, there are less disharmonious associations in all of the Early and Middle Pleistocene thrown together than in the Late Wisconsinan.

David Hurst Thomas:

I'd like to intervene at this point and ask you some of the questions from the audience in our last five minutes. I'd like you to answer quickly, if that's possible, so that we could get in as many of these as we can. Dr. Smith -- Norman Owen-Smith -- is there similar evidence of keystone herbivores in North and South America. By comparison to the African data that you talked about, does the keystone-herbivore concept apply to the North and South American context?

Norman Owen Smith:

The best way of characterizing the modern megaherbivores in Africa, elephants and rhinos, is with ecosystem engineers -- that they don't simply respond to habitats as they find them, they actively transform those habitats. And with the demise of species in their body size range in the Americas in recent times, you don't have any counterparts in terms of that kind of keystone effect.

David Hurst Thomas:

But, presumably, they would have existed at one time in the Pleistocene, if the idea is general.

Norman Owen Smith:

Yes, the species were there formerly.

David Hurst Thomas:

Dr. Graham -- is there any fossil record that shows the extinction of plants due to the extinction of megaherbivores?

Russell Graham:

Not in North America. There is one interesting record that's been recently found -- some spruce cones from Louisiana have been found that are extremely large, and they don't fit any living species. And this is being worked on by Steve Jackson at the University of Wyoming, and he may be suggesting that this may be an extinct species of spruce. But, in large, during the Late Pleistocene anyway, you don't see any real plant extinctions. The other interesting one is that there are no real insect extinctions, either.

David Hurst Thomas:

Yes, good point. Or nonmammalian vertebrates, other than some scavenging birds.

Russell Graham:

And some large reptiles.

David Hurst Thomas:

And completely unknown is the rate of loss in freshwater fish at this period, one might add. For Paul Martin: Is there any research that indicates a change in extinction rates or types as humans shifted from hunter-gathering to agriculture?

Paul Martin:

A change in extinction rates from hunting-gather to agriculture in North America -- I don't think of any right off the bat. Ross, do you?

Ross MacPhee:

No.

David Hurst Thomas:

That's what I call an answer. John Alroy, can your analyses discern whether trophic patterns of extinction occurred that may relate to changes in habitat structure due to megaherbivore extinction? That is, do megaherbivore extinctions immediately precede extinction of large numbers of other grazing, open-habitat herbivores?

John Alroy:

I presume that the question refers to the Cenozoid database, where the temporal resolution is on the order of a million years. The problem there is that you would expect any such effect to operate on a scale of a few thousand years. So these would be extinctions of megaherbivores and other things happening in the same bin -- in the same million-year bin. So all we're going to be able to show is a correlation of those rates of extinction. And that's discouraging, because then you can't show a sequence.

The bigger problem, however, is that megaherbivores have never been very diverse on any continent, including North America. Although they were present in North America from the Middle Eocene for about the last 40 or 45 million years, there's never been a time when there were more than maybe a half a dozen species of really big megaherbivores -- like proboscidian-sized. So computing a rate of extinction for megaherbivores is really dicey. You're talking about one over six or two over three -- that type of thing. It's hard to say what that type of an extinction rate really means.

David Hurst Thomas:

Last question, and this one I'll direct to Tony Stuart: What is the composition of the deposits that preserve the remains of extinct species, and how did this material survive through time?

A.J. Stuart:

Well, the fossils occur in a whole range of different environments. You've got river deposits, you have lake deposits, you have cave deposits. And I suppose what they mainly have in common is that they should be rich in lime, calcareous sediments, that preserve [inaudible]. Is that the point wanted, or am I missing something?

David Hurst Thomas:

No, I think we'd all agree what get preserved, gets preserved, and that's the only rule.