Revisiting Pimm et al. 1988

In a paper published in The American Naturalist in 1988, Stuart Pimm, Lee Jones and Jared Diamond explored existing mathematical models and developed theoretical predictions about how risk of extinction would vary between species. Pimm and colleagues then put these predictions to test using a database of annual breeding censuses of birds carried out by volunteer birdwatchers on islands off the coast of Britain. Twenty-eight years after the paper was published, I spoke to Stuart Pimm about the match between theory and data, how he got interested in the topic of extinction, and his shift from being a theoretical ecologist to a full-time conservation biologist. 

Citation: Pimm, S. L., Jones, H. L., & Diamond, J. (1988). On the risk of extinction. The American Naturalist, 132(6), 757-785.

Date of interview: 17th August 2016 (on Skype)


Hari Sridhar: I would like to start by asking you about your motivation to write this paper. From looking at your publication profile, it seems like this was your first article on extinction. What got you thinking about this topic at that time?

Stuart Pimm: This was in the very early days of conservation biology. For obvious reasons, a lot of people, were interested in how long small populations would persist. It is a very straight forward management problem. If you have a tiger reserve that has only 4-5 tigers in it, how long will that population be expected to live? It obviously has something to do with the generation time of the animal — individual tigers live a long time. But even if you are measuring generation times, how many generations does a small population persist? I knew – because I had spent a lot of time when I was a teenager on some of these islands –  that, off the coast of Britain and Ireland, were a large number of islands where birdwatchers visited very routinely in the spring. These birdwatchers not only kept lists of how many species they saw and how many species were breeding there, but they had done so for 30-40, in one case 50, continuous years. So, one could directly answer the question of how long these small populations lived. And moreover, you could begin to test some of the mathematical models that suggest what the relationship of the population size would be to how long the populations lived. A fair bit of the paper is trying to fit those models. So, the motivation was a combination of understanding that it was an important problem and having been to the places where I knew that the data existed to solve it.


HS: You mention that this dataset had been analysed earlier by Jared Diamond in 1984. Was this planned to be a follow-up on that analysis? Could you tell us how this group of authors – yourself, Lee Jones and Jared Diamond – came together for this piece of work?

SP: I knew that Jared had pulled together some of the data. He was more interested in the island biogeographical aspects of this. These islands had bird observatories on them. They were usually a small field station – maybe a small house – where you could stay. These bird observatories would publish annual reports. It wasn’t as if all the data were stored in one place. It was a whole bunch of individual reports. For example, Bardsey, which was one of the islands that had quite a number of years [of data] would have a report for 1950, another for 1951. You had to go to the individual reports and extract the individual numbers. Jared and his student had done some of that, but I knew there was a lot more data sitting around at the British Trust for Ornithology (BTO) in Britain. What we were looking for were not average values; we wanted to have long-term data on how long a particular species persisted. I knew Jared reasonably well and he was obviously a person to work with as a collaborator.


HS: Did you develop most of the theory that is in this paper? Were you mathematically inclined right from early in your career?

SP: In my early career I was very much an empirical person – going out into the field to watch and ring birds. Then I started getting into theory because I thought the existing theory wasn’t very good; sort of a funny way of doing it. But I was sufficiently mathematically inclined even at that point, that the theory of how long these populations would last was not difficult. There were different ideas about what the shape of the relationship between size and longevity was.

I knew that Jared had analysed this data, but I had also done my BSc at Oxford, where I spent a lot of time in the Edward Grey institute. I knew about David Lack’s work, about Mark Williamson’s work – I had spent some time at York. So, the people who had looked at these data before were people I knew.


HS: Was Lee Jones Jared Diamond’s student at that time?

SP: I think so. I’ve never met Lee, but I believe he was one of Jared’s students.


HS: Did you and Jared Diamond ever meet during this study and the writing of this paper?

SP: Yes. In fact, there is a wonderful story there. I met up with him and we finished this paper just as his wife went into labour with twins. She wasn’t expecting to have the babies quite as early as she did. I visited Jared because I knew that, at some stage, we needed to talk face to face to finish off the paper. I arrived to stay with Jared and Marie. Marie was in hospital. Whenever the phone rang, Jared would jump two metres up in the air thinking that the babies were due. Jared and I both remember the writing of this paper very clearly. I talked to Jared just the other day – we both remember it very well because we finished it on the day his twins were born.


HS: Where was Jared Diamond based at that time?

SP: He was where he has always been, at UCLA [University of California, Los Angeles].


HS: Do you remember how long the project took – both doing the analysis and writing it up?

SP: I don’t really, I’m afraid. My guess is a few months.


HS: Apart from this one time when you met up with Jared Diamond, were all your other discussions over the phone?

SP: By letters! Remember, this is 1986-87, so a lot of stuff was done the old-fashioned way by exchanging letters.


HS: You would share drafts over snail mail?

SP: Yes, I think we did share drafts over snail mail.


HS: Do you remember how you made the figures for this paper?

SP: I know exactly how I made the figures. I had a light table and I had very good black pens. That figure was drawn by hand. In fact, looking at it now, enlarging it, I can make out that it is drawn by hand; in one place it doesn’t connect up very well. But this was just before the ability to do figures on computers became available.


HS: Were all the figures drawn by hand?

SP: Yes.

Stuart Pimm in his office at the University of Tennessee, Knoxville (photo taken in 1988; © Stuart Pimm)

HS: You mention a person by the name of Timothy Reed who collected all the data. Could you tell us who Timothy Reed was?

SP: Tim did his PhD, at Oxford, I think, I’m not quite sure. I didn’t know him at Oxford but we met up soon after. In all the follow-up papers Tim was a co-author. Not on this one, but on the subsequent papers that we wrote on these data.


HS: This paper doesn’t have an Acknowledgements section. I’m guessing that was the format in The American Naturalist at that time?

SP: I’m not sure.  We probably should have acknowledged Tim. We say somewhere in here that Tim was helpful. However, we didn’t exactly get a lot of help from anyone else. We certainly didn’t get any grant money for this. Most of this was just sort of going to these reports and pulling out the data. It was a very tedious business. I would go to a report and use a highlighter pen to indicate, for example, how many pairs of buzzards nested on Lundy in a given year, put those data probably onto punch cards and go on from there.


HS: Did you access all this data from hard copies of the reports in the BTO library?

SP: Funny thing is, just yesterday afternoon when I was in my office, I wondered if I still have all the data in there. I have a big filing cabinet, in which I have a draw of these particular files. The data were from a mixture of places but I think that the BTO library was the main source. I seem to recall that at some stage I went to the BTO headquarters and just photocopied a lot of documents.

There is an interesting point in all of this. Several I think. One of them is that Britain has an extraordinarily rich tradition of natural history. People have been collecting observations on things in Britain since Victorian times. So, when it comes to studies of how global warming is moving species about, the British have, by far, the best data, because they can map out where butterflies and crickets and birds and all sorts of things are now, and where they were 50 or 100 years ago.

Of course, that tradition spilled over with the British Raj in India, and India also has an amazing tradition. Krithi Karanth did her PhD with me, looking at how big mammals have disappeared from India. And there is this incredible dataset of mammals – most of them were shot across all of India for 150 years. These kinds of data are very idiosyncratic, they are very peculiar, but they are incredibly valuable for understanding what happened to populations, what’s happening to our environment. These individual bird observatories were mostly or entirely populated by amateur bird watchers who would go out to these islands and spend a week or two bird watching, and collect all these data that are so valuable now.


HS: Do these observatories still exist?

SP: Yes and no. I think some of them are still there. I went looking to see if these observatory reports were being produced online. I don’t think they are in the way they used to be, say 30 years ago, when people were simply printing out the hard copies. That’s a real shame, because long-term data on small islands is a very valuable guide to how small populations are surviving in other areas.


HS: When was the last time you visited any of these islands?

SP: Good question. Not for a long time. I may not have visited any in 40 years. These were the sort of places I visited when I was in my teens. That seems an awfully long time not to have visited some of them. Some of them are very hard to get to, I might add. Bardsey is hard to get to. Hilbre, you can walk over to at low tide. I probably visited that more recently. Skokholm and Skomer take a bit of getting to. Fair island, I’ve never been to; it’s very remote. But these were certainly places I would go to and stay, and camp when I was in my teens and as an undergraduate.


HS: Over what period was this data collected? I don’t think that is mentioned in the paper.

SP: That’s unfortunate. We did write up some follow up papers which may have been more informative. On some of these islands, people were counting things before the second world war. Bardsey I think has a record going back into the 1930s. There was a break during the war years and they started again thereafter. But some of these islands had 30-40 years’ worth of data.

SP: I can hear a Koel calling in your garden.

HS: Yes, these days they are calling all the time, and they start very early in the morning, sometimes as early as 3 am.

SP: That’s wonderful. Back to the subject of British birds! So, we did some other papers that were follow-ups of this one. I don’t know if you have seen those but I will dig them out and send them to you when I get off the Skype call. Those might have more detail about the years that we used.


HS: Would you know if these islands have changed a lot since the time these data were collected? Are surveys still done on the islands, and if yes, would you know how bird numbers have changed on these islands?

SP: We did a paper on long-term turnover. The lead author was Gareth Russell who was my graduate student. There was the idea that if you measure turnover over a short period of time it would look to be high and then it would gradually sort of asymptote. But Gareth and I found that the turnover sort of kept on increasing. That’s because there were long-term changes on the islands – grazing by sheep and rabbits were the main drivers. A lot seemed to depend on whether there were rabbits on the island and if so how many, and whether they were controlled by Myxomatosis. It also seemed to depend on sheep; if people put sheep on the island it made a difference. That paper’s title is “A Century of Turnover: Community Dynamics at Three Timescales” and it is published in Journal of Animal Ecology in 1995. That will tell you a lot about the longer-term datasets. In that paper, we only used the seven islands in which the maximum interval was greater than 80 years. So, it was an extraordinarily long period over which these islands were studied.


HS: Did this paper have a relatively smooth ride through peer-review? Was The American Naturalist the first place you submitted it to?

SP: I think The American Naturalist was the first place we submitted it to. Regarding peer-review – we can get an idea from the dates. We submitted it in June 1987, we got comments back quite quickly obviously because we revised it in October and then it was accepted in February and published in December. Back then that was pretty fast. In this day and age, with online publishing, we would have got it out much faster.


HS: So, I’m guessing the revisions were minor and the final published version was not very different from the original submission?

SP: Yes, I doubt if it changed much from the first submitted draft.


HS: How was this paper received in academia when it was published?

SP: You know, I’ve learnt so much in my career and I have changed a lot of the way that I do things. If I had published this paper today I would have talked to the media, discussed a lot about why the paper was important, what its significance was and all the rest of it. Back then, we didn’t. We just published the paper and let it go. But in fact, you will see that it is a highly significant publication in terms of the number of people who have looked at it. We did do a follow-up paper – “Times to extinction for small populations of large birds”, which we published in PNAS a few years later. And that paper, very specifically, addressed management issues of the Hawaiian crow, northern spotted owl etc. We drew much tighter connections of the results to practical conservation problems in that.

We wrote this paper because we wanted it to be an important paper on conservation. This was at a turning point in my career, where I was going from being somebody who was a theoretician more than anything else, to becoming a full-time conservation biologist. So, while the paper is very strong on theory at the time, it was very weak in the connections to conservation, connections that I would make now. The second sentence of the paper is very interesting. I start by saying: “Understanding this risk is of theoretical interest to population biologists,” and then add “but it is a central practical issue for conservation biologists and wildlife managers concerned with saving populations.” If I were writing that now, I wouldn’t care about the theoreticians; it would all be about the practical side! When I wrote the paper in PNAS a few years later, I started by saying that a major practical problem in conservation is to predict the survival times of small populations. So, within five years of this first paper, I’m publishing a paper where I’m emphasising the practical conservation side of things that matters. These papers were written at a time when I was undergoing a transition from being a theoretician and doing a little bit of experimental work, to becoming a full-time conservation biologist.


HS: What made you shift to becoming a full-time conservation biologist?

SP: I’m just trying to look up when the Society for Conservation Biology was formed. Yes, here it is – The Society was founded in Michigan on 8th May, 1985. I was there when that happened. It was a Thursday afternoon, very nice and conveniently organised just after teatime. That’s when a group of us decided to get together and create the Society for Conservation Biology. While that decision had been made around the time I was writing the American Naturalist paper, it wasn’t until really a few years later that it became clear that conservation was going to be a big thing. I was very fortunate to get a Pew scholarship in conservation and the environment in 1993. That told me that I could do conservation full-time, that conservation can be the most important thing in my life, that I don’t have to justify my existence by doing ecological theory. I had never looked at this before, but the beginning sentences of those two papers capture that transition nicely. One says theory and conservation, the second one, five years later, says conservation. Getting the Pew fellowship was fantastic because it meant I had four years of funding to develop a conservation programme, and that’s exactly what I did.


HS: This, in a way, leads into the next question I want to ask you – if you were to do this study and write this paper today would you do anything differently? From what you have just said, I’m guessing that the connection to conservation in the paper would be much more explicit.

SP: Yes, it would be. Back then there was a lot of discussion about conservation, whether it was a separate science and whether it was good science. There were people who were saying – look, if you can do ecology, you do ecology; if you can’t do ecology then you do conservation. They were basically saying conservation is less rigorous. It’s softer, squishier, it’s not serious ecology.

The international prizes I have won — the Heineken Prize and Tyler prize — are for showing how science is essential for conservation. They are for making conservation safe for scientists, as it were, by looking into very important problems that need a good understanding of ecological ideas and theory. I think this paper illustrates that very well, i.e. that there is a strong theoretical basis for how the relationship between population size and its time to extinction might unfold, and understanding that relationship was clearly going to be important for practical problems. The questions that conservation pose are difficult ones – they are theoretically challenging ones.

At the same time, the other challenge to conservation was the following. People would say – I’m an ecologist and therefore everything that I do is important for conservation. Clearly, that’s not true. Ecology and conservation do not entirely overlap. Conservation isn’t a subset of ecology. It involves a lot more. It involves different kinds of techniques, different kinds of challenges. Equally, a lot of what ecologists do isn’t particularly relevant to conservation problems. So, this paper came at a time when we were fighting to develop the identity of conservation as a field. And I think I have been given these international awards for showing the unique scientific content of conservation. This is a good paper to illustrate that.


HS: Today, 28 years after the paper was published, would you say that the main conclusions from the paper – both in terms of the theory and the empirical findings – still hold true, more-or-less?

SP: Yes, I think they do. The theory is pretty simple stuff. There is a huge amount of theory that has been done since, but this paper does capture the idea that there are problems that small populations face –demographic accidents that you cannot get around. These are the unavoidable vagaries of sex and death. If you have a small population you are going to have these demographic problems. But even if you have a large population, you are going to have stochastic effects, environmental disturbances, catastrophes etc. and you have to recognise that all these processes are going on. Now, I didn’t invent these concepts; those ideas were around. But I do think that, with this paper, Jared and I were able to begin to look at how real populations behave. And we had a unique dataset and it remains a unique dataset. There are few places that have been studied quite so consistently for such a long period of time as these islands off the coast of Britain.


HS: One interesting prediction from your theory is regarding the differences between the small-bodied and large-bodied species, and how their likelihoods of extinction interact with population size. Has there been more work on that subsequently?

SP: I think people realise that these extinction problems do relate to body size, and they are pretty clear why. Small birds are going to breed only one or two years, and a year is approximately the generation time of the species, whereas the bigger birds can last for several years and they have more than one chance of breeding. There are differences in the dynamics that follow from that. I think people understand that.


HS: You say you left out five species from the analysis because there were no extinctions recorded for them. These include red grouse, chough, oystercatcher, rock pipit and twite. Do you know if since your study any of these species have gone extinct on these islands?

SP: I’m trying to look at the 1995 Journal of Animal Ecology paper. No, I can’t find it there. But, in general, as we produced more papers, we acquired more data. The most recent one we did with Christine Stracey. Christine looked not just at the breeding records but also at the birds that turn up on the islands as migrants. That’s published in Journal of Biogeography in 2009. I sent Christine to England and she acquired more data to add to our compilation of data.


HS: This paper was your first with Jared Diamond. Did you continue to work with him after this piece of work?

SP: Yes, we work on a number of things. We haven’t worked on the island stuff for a while. Well, we are working on other aspects of islands, looking at checkerboard distributions and things like that. In fact, I have just published a book on that, not with Jared, but with Jim Sanderson about the debate between Jared and Dan Simberloff. But Jared and I have written other papers together, as recently as a year or two ago.


HS: Have you ever read this paper after it was published?

SP: Have I read it from start to finish? Probably no. But in the sense of revisiting it and extending it, yes. The work that I did with Christine Stracey is relatively recent. We published that in 2009. So, I have revisited the data and thought about what they mean. This has been a paper that generated a Research Program, in the sense that, once we did this, we realised that there are a lot of other papers that we could write that would follow-on from this. The paper with Christine marks 21 years of further papers as a consequence of this one. I’m very encouraged by that. I think that tells me that this was an important paper in terms of shaping a research agenda.


HS: If you compare this paper to the ones you write now, do you notice any striking differences, maybe in style?

SP: Oh yes, I’m sure. Developing a writing style is hard. It is hard to write simply and accurately, but I’ve had almost 30 years more practice now! Writing a popular book – “The World according to Pimm”. was great training to try and explain complicated ideas in simple language. I would like to think my writing has improved in the last 30 years.


HS: Do you have a writing routine? Do you write at a particular place and time of day?

SP: Yes, I always like to write first thing in the morning, and then I go into the office at lunch time, and either have lunch with my students or other people, usually some sort of a working lunch, and then I meet with my students in the afternoon. I do a lot of teaching in the fall and the autumn. In the spring, I have much lighter teaching and that is when I tend to do my serious writing.


HS: Would you count this paper as one of your favourites among all the papers you have written?

SP: I definitely would. Not just because it is one of my most highly cited papers, but for the reasons that we have explored. It was a paper that marked my transition from being an ecologist to being a conservation biologist. I might have called myself a conservation biologist when I wrote this paper, but only just. Soon thereafter, I would say – I’m a conservation biologist. For example, I go to conservation meetings, I don’t go to the Ecological Society of America meetings. That’s not out of spite or anything. I can only go to a limited number of meetings because they are so expensive. It’s the conservation meetings that are important to me. Its conservation and action that matter to me now and this paper marked that transition. But I also like this paper because it very clearly illustrates what I have done that I’m most proud of, which is bringing good science into conservation.


HS: You mentioned that it is one of your most cited papers. Do you know what it mostly gets cited for?

SP: No, but I guess I can take a peek to find out. It’s been cited over a thousand times. The papers and books that cite it are looking at the fate of small populations, why they go extinct, and what one can do about that.


HS: What would you say to a student who is about to read this paper today? What should he or she take away from this paper published 28 years ago?

SP: I hope they will take away the connection between ecological theory and observation and their relevance to practical management. The results in this paper were vitally important in managing at least two threatened species of bird – the alala or the Hawaiian crow and the spotted owl. That is much clearer in the PNAS paper. But the ideas that these small populations are only going to last a few years and that they are probably going to last a little bit longer if the species are larger-bodied are actionable results. This is science that can make a difference. So, I think it is the connection between theory, an extraordinary and exceptional dataset and the application to real-world problem, is what students should really be taking away from reading this paper.






Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s