Prehistoric Extinctions in Hawaii: The Search for Causes
Helen F. James, Smithsonian Institution
Douglas Siegel-Causey, National Science Foundation
Helen James presents first, and then Douglas
One of the rationales for bringing this group together is to ask
how the extinctions that we study in the fossil record inform
us about our current societal concerns about the biodiversity
crisis. And I want to address myself rather specifically to that
in my talk, because our main concern is that on continents we
are transforming natural habitats into habitat archipelagos.
And, at the same time, there are emerging signs of global climate
change. And will these two factors interact in some catastrophic
way to increase the already rising rate of extinctions and loss
Now, islands are places where these fears have already been confirmed,
because in the human era there have been catastrophic losses
of diversity. And I'm going to use the Hawaiian Islands as an
example of this, and hope, basically, to show that -- or to convince
you that -- indeed, there are things that we can learn about the
island extinctions that may inform us about biodiversity crises,
both with respect to managing habitat islands and with respect
to how serious this crisis is. Because the island extinctions
are seamless with the current rise in extinction rates -- it's
really one event. It has all taken place in the human era. I
think you've seen quite a bit of radiocarbon data that point
to that fact already. So we'll begin with the Hawaiian archipelago.
These islands -- the major diversity in vertebrates in these islands
is birds -- birds are a vagile taxon, and they are mainly the
species that have been able to colonize and diversify there.
Stuart Pimm -- not meaning to steal his thunder, but with his
permission -- I can quote some figures that he has published.
He's attempted to estimate the number of species of birds that
were in the Hawaiian Islands before humans arrived, perhaps 2,000
years ago, and the number that have survived the extinction event
today. He estimates that between 125 and 145 species of birds
were in the islands, and that perhaps 10% have essentially survived
the extinction event. They are in decent condition -- they're
not imminently threatened with extinction or, indeed, extinct.
So this is a catastrophe. And I'm going to try to address possible
causes of the catastrophe.
It turns out that the species that were able to colonize Hawaii
2,000 miles from any continental land mass were all either water
birds, raptorial species, or passerines. And this histogram gives
you an idea of the prehistoric extinctions which we've found in
the fossil record of the decline in diversity.
Now, before we launch into causes, just a quick bestiary. The
islands did not have megaherbivores, but they had a mesoherbivore
guild of waterfowl that had become flightless and large in body
size. There were six species of these that became extinct. The
paintings are by artist Julian Hume. The plumages are fanciful,
but the ecological setting and the morphology -- the body form
of the birds -- we've tried to reflect faithfully.
Ibises. Ibises had colonized and become woodland species, flightless
woodland species. And this painting depicts one of the major ways
that bones had been preserved in the islands. Flightless birds
fall into lava tube caves, into vertical openings, and then are
preserved inside the caves. And we later go and find their bones.
There were, as Stuart Pimm will also mention, in the Pacific there
were a tremendous number of flightless rails. Every island had
one to three species of flightless rails, and these have largely
Although, looking at the recent Hawaiian, the surviving
Hawaiian avifauna, you would think that predation was not a factor
ecologically, an important factor ecologically, in the fossil
record there is a group of extinct predators, including a genus
of owl, which was specialized to eat Hawaiian birds, and a harrier
that had specialized on eating birds.
There is emerging evidence of a mini-radiation of crows. There
were crows, undescribed crows, with hammering adaptations in the
beak and with high-arched bills. They are all large, raven-sized crows,
and I know of at least five species of these.
Now, the extinctions in Hawaii -- obviously, if we have only perhaps
10% of diversity remaining, these extinctions have reached much
deeper into the faunal list than the extinctions we've spoken
about in North America. Many small species have disappeared.
And there was a great diversity of songbirds. This is a reconstruction
of a finch-like Hawaiian honeycreeper. Bill shapes were extremely
variable in this radiation -- this is one with perhaps a fly-catching
adaptation in the beak. And, in addition to the species that
have been described, we continue to find more odd, extinct animals.
In this slide you see a reconstruction, at the bottom, of an undescribed
species which is similar to what we call the "heterobills,"
with this very odd bill form -- with the long, overhanging maxilla
and the short mandible. And this is an undescribed one with the
largest honeycreeper, with a very long upper mandible and short
lower mandible. So diversity is continuing to emerge.
The extinctions, just to characterize them briefly, did not simply
randomly remove species from Hawaiian ecosystems, but they had
differential effects. I have very roughly grouped the prehistorically
extinct animals into feeding guilds -- and, as you can see, predators
disappeared practically completely. Terrestrial herbivores largely
disappeared. There was a guild of ibises and rails that fed on
the forest floor, in detritus and on mollusks, and on everything
that was available that have almost all disappeared. Granivores
took a very hard hit. Nectarivores did pretty well.
So ecosystems in Hawaii, as we see them now, have a different
trophic structure. They're not just missing some members, but
they're radically different. And this carries that point a little
farther. This is a summary of an owl pellet fauna from the island
of Molokai, and we have the modern birds that still occur on Molokai
-- of Hawaiian honeycreepers -- representing their bill forms
here -- nectarivorous species, insectivorous species that also
take nectar; one frugifore here. The seed-eating species disappeared
entirely. And if you look at the abundance of those species
in the owl pellet fauna, in terms of minimum number of individuals,
notice that the granivores had been extremely abundant -- at least
in the selection made by this owl -- and that this guild had been
fairly rare. And the birds that are common now in Hawaii are
absent. This is very typical of the Hawaiian fossil record.
Often species that were widespread and abundant have disappeared -- or, at least, that we infer were widespread and abundant.
Now, to explain the extinctions. Back in 1981, when Storrs Olson
and I first realized that there was this tremendous diversity of
birds that had disappeared, and we began to wonder why -- and
we saw that the radiocarbon dates were pointing to human-era disappearance
-- we proposed, basically, a laundry list of possible causes.
And these are things -- this is basically the arsenal of possible
causes that came in with Polynesians: hunting and overharvesting;
fire and other types of landscape change; the Pacific rat, which
Richard Holdaway talked about, was brought in; domesticated dogs;
the Polynesian pig; jungle fowl, which could have been an important
introduction. And then this is something that we have always
given lip service to, but never really given any serious consideration
to -- and I think it's about to be raised more seriously by Ross
MacPhee -- but it ought to be considered, and that's introduced
diseases. Now, I've tried to show that the difficulty is in actually
connecting any of these possible causes with the actual extinctions
So, to review the progress that has been made towards trying to
make some of those connections, I want to organize the rest of
the talk by addressing how the various hypotheses, or possible
explanations of extinction -- which the other presenters have
very nicely educated us about already today -- may apply in the Hawaiian
situation. So if we begin with climatic forcing -- we're going
to look at whether some climatic change has affected the extinctions;
we're going to look at Paul Martin's blitzkrieg hypothesis; the idea
of gradual overharvesting; and the effects of biological invasions;
introduced predators, competitors and parasites; the ideas that
Norman Owen-Smith very beautifully explained about the ecological
web and possible interactions between the extinction of important
herbivores causing habitat changes -- and, of course, this
extends to predator-prey relationships, and a lot of ecological
interactions that may have exacerbated the situation. And then, finally, what has been the most prominently promulgated
idea is that habitat destruction, habitat loss, reduced the ranges
of these species, and that was the major force behind the extinctions.
To address the climate change idea, what we want to do
is find out when the extinctions occurred so that we can relate
that to possible periods of climatic change. Now, this site represents
an excavation in a cave, where there was an alluvial fill, with
very fine stratigraphic layers. And in about five months' time,
I and my helpers were able to peel back these layers and make
a very extensive collection of bird bones from one site. Now,
I worked with Tom Stafford on this, and so I have some sanctified
Tom Stafford radiocarbon chronology here, showing that, at the
bottom of this sediment -- we begin at about 9,000 years ago,
and there's a remarkably linear progression of the dates, up to
modern. Looking at the 16 most abundant species in that stratigraphy,
I have recorded here the first and last occurrence. And what
I want you to notice -- this red bar indicates human presence,
and the dashed red bar indicates possible human presence. What
I want you to notice is that, with one exception, there appears
to be a decapitation of this fauna in the human era. And that
dates to about 800 to 2,000 years ago; we're in that ballpark.
Now, if we really want to know whether climate change is playing
a role in the extinctions, we don't just want to know when the
birds went extinct. But we want to know, during periods of
climate change, did we also record extinction events, and how
severe were they? These are the data that I just showed you from Puu Naio Cave, representing
species richness over the 9,000 years of the Holocene, and showing
that sharp decline after human arrival. Now, I'm borrowing data
here from David Steadman, to look back tens of thousands of years.
This is a fossil site that he excavated on 'Eua, in Tonga. And
he was able to date this fauna. We're looking at just a point
here, a point here and a point here. We don't have a full diversity
curve for these data. But 60,000 to 80,000 years he recorded a fauna.
Then he had an archeological fauna which was a small sample,
and possibly had missed some species. And then he had the modern
fauna. And, as you can see, on that time scale you also have
this sudden decapitation in the human era.
This is perhaps the most interesting site. This is a site on
Oahu that dates -- in the Hawaiian Islands it dates to . . . I'm
going to have to hurry. This site dates to over 120,000 years old.
So the fauna here has gone through a complete cycle of glaciation
and deglaciation. It covers the Late Pleistocene period of maximum
climate change. And, as you can see, the fauna comes straight
through, with no dramatic decline, until the Late Holocene and
human arrival, and then you lose it again.
These are the radiocarbon dates for all extinct species and populations
that Tom has provided us with -- and, as you can see, they have
a Holocene distribution. They come down to within the human era,
and many of them go well beyond the period of 2,000 to 1,000 years
ago. In fact, the most recent dates we have on extinct populations
are 500 years ago. Now, this suggests that there may have been
a period of coexistence of the Polynesian populations with the
now-extinct animals, for a period of perhaps 1,000 years -- and,
if that is true, then it doesn't look good for the blitzkrieg
hypothesis, because we would have had to see the extinctions occurring
I'll hurry through some of that. . . . Gradual overharvesting would
look like a more reasonable hypothesis, in light of the radiocarbon
evidence. So we have sort of discarded climatic forcing and blitzkrieg,
and gradual overharvesting looks plausible. There's a problem
with the archeological record. It is certainly not full of remains
of these birds. The birds that are extinct are generally rare
in archeological collections so far. The Nene, which happens to not have gone extinct, is fairly
abundant, and so is the dark-brown petrel, which is severely endangered.
So some populations may have been restricted, but there's not
overwhelming evidence for a focus of the human culture on exploiting
Now, moving on to biological invasions. I said a few things about that. A couple of months ago, I was sitting in my office, looking at bird bones, when Douglas Siegel-Causey came in and explained that he had been doing research for two years in avian epidemiology, and we began a discussion of the possible role of disease in the prehistoric extinctions. And I think that I have covered these ideas well enough that I would like to turn the podium over to Doug Siegel-Causey, and he will give us an idea of something that may not have been important, but is another idea that has not been seriously investigated, and should be.
Disease is the manifestation of an infection by a pathogen, and
the types of pathogens that are likely to play a role in extinction
are going to be viruses and fungi. And I'm going to focus, in my very
brief comments, just on viruses. When you have an infection,
there is an initial response, where you have a manifestation of
the disease -- and then, soon after, you have an immune response.
Based on how pathogenic that organism is, you can have the full
manifestation of the disease occur before an immune response --
that is, you die really quickly.
And so there is almost a counterintuitive relationship. Right here on the
graph, this is illustrated on the horizontal axis here, is how dangerous
a certain disease organism may be -- from hardly having any effect on the organism,
to being very dangerous, right here. And the line we're looking at is
this very dark line. If the disease pathogen does not cause much effect
on the host, then many of the individuals in the population -- in fact,
all of them -- can harbor the disease pathogen and not show any ill effects.
If it is very dangerous, then everyone dies, and then you also have the
same effect -- that is, you have hardly any of the population harboring
the pathogen still alive, because they're all dead. So, if you'll look
at this curved line right here, it's counterintuitive, but that the disease
pathogen -- which is likely to have the greatest effect -- is something
which is not the most dangerous. The most dangerous pathogen will kill
everything off. And so, as horrible as a disease like that caused by the Ebola
virus may seem to us all, if all of us in this hall came down with Ebola,
we would all die, and we probably would all die in this room. But New
York City probably would be okay, because none of us would leave this
room. That's a comforting thought, I know. . . .
All right. And so it turns out, through our studies of co-evolution,
we can show that it's unlikely that the type of diseases which
are going to cause these sorts of epidemic effects are going to be
those that are evolutionarily associated with the host. Instead, we
expect these to occur through transmission of an existing disease
in some other organism to another species. That is, you will
be experiencing a novel or an emergent disease.
And we don't have enough time to go into this, but the next question
is: Well, if disease may have played a role in extinctions in
Hawaii, where did these diseases come from? And there were basically
two sources. You could have them occurring naturally, by naturally
occurring immigrations or invasions of other species, or they
could be introduced -- by humans, or by the animals that they
brought with them.
And one of the first things that Helen and I thought about were
chickens, and this slide here was my third attempt to show the
diseases that are vectored by chickens. I ran out of space on
the first two times. You could almost say most of the diseases
we understand to affect Hawaiian birds, or birds in general, are
vectored by chickens.
And we can summarize the sources of the diseases that may have
been introduced in the Polynesian Islands -- chickens, ducks, and
geese are probably the most likely candidates; pigs -- in my next
and last slide, Ross -- I'll show you play a very important role;
insects -- hardly likely. The ecological theory will show us
that, when you introduce something like an insect that will transmit
a virus or a disease, you get very chaotic patterns that probably
will not produce epidemics of the sort that cause extinctions.
In humans, we know of no diseases that we can give to other animals,
to cause any substantial disease.
So which of all these possible diseases might play a role? Well,
we do have historical accounts in Hawaii of diseases like fowl
pox, an avian malaria, which did cause major reductions in populations.
But there is another candidate which is much more interesting,
and has, I think, a much more important role in what Ross is going to
talk about, which is influenza. And I now will give you a one-minute
description of how influenza works.
Influenza is a virus. It has many components, but we can classify
all the various strains of influenza by virtue of looking at the
type of neuramidase gene it has and the type of hemogluten gene,
and we have codified these in terms of H types and N types.
There are 14 H types and there are 9 N types. The combination of
these two produce a virus strain. And so, for example, here --
the strain H1N1 -- is an influenza strain that caused the 1918
Spanish flu epidemic, pandemic, which killed over 18 million people.
The H5N2 strain is an avian influenza. It's responsible for
the largest and most widespread epidemic that we currently know
about -- and it's occurring as we speak. Over 20 million birds
are dying throughout Mesoamerica by this flu strain. H7N7 is
a seal plague.
How do these occur? How do these different strain combinations
occur? I'm showing you, in this diagram, that there is a mixing
-- a genetic recombination -- that occurs, and it occurs in mammals.
Now, for this to occur we must have high densities of the original
vectors of the influenza types, and we must have high densities
of the mammalian vessel, if you will, for recombination.
First of all, the place where all of these subtypes occur are in birds
-- and, in fact, some species, through our own work on migratory birds
show that some species have all types. For example, mallards generally
carry all types. And, so, if you have high densities of waterfowl and
high densities of mammals -- like you would with pigs -- and high densities
of humans -- which is what you would have on a farm or in domestication --
you have the setup for not only human influenza epidemics, but wildlife
epidemics. A very famous virologist was speaking on Ebola, but, in general,
about something like the Andromeda Strain. And he says: One wonders what
might happen if and when the day comes that an agent of this type and
degree of lethality -- like Ebola or the Andromeda Strain, which was featured
in a novel -- obtains the right kind of receptors, or whatever is needed,
to replicate in the respiratory tract and be transmitted. This would be,
I think, the most lethal of influenzas. . . .