Interview with EICES Director Shahid Naeem

On a recent visit to Bangalore, India, EICES Director Dr. Shahid Naeem was interviewed by the National Center for Biological Sciences. The interview discusses Dr. Naeem’s research on the role of biodiversity in ecosystem functioning over 20 years ago, and the importance of his publication in Nature in 1994. This article, “Declining Biodiversity Can Alter the Performance of Ecosystems”, went on to become a “citation classic” among ecologists and biologists, and is recognized as one of the earliest papers to explore biodiversity as a parameter rather than an outcome.

His interview was published on the Ecology Students’ Society webpage at the Centre for Ecological Sciences, Indian Institute of Science, and on the blog for the International Network of Next-Generation Ecologists.

Read his interview with Hari Sridhar below:

Hari Sridhar: I’d like to start by talking about the history of this project and your motivation for doing the work presented in this paper. What took you to Silwood Park?

Shahid Naeem: Silwood Park – in Imperial College, London – was extremely well-known, what with people like Charles Godfray,John Lawton and Mick Crawley.  As an American, I was quite influenced by the research that they were doing. And then they advertised for three post-doctoral positions, with a possibility for working with others on the Ecotron, even though it was still being built. I was a postdoc. at the University of Michigan at that time, and when I saw this opportunity, I thought I would love to go to England and meet European scientists; till then I had met mostly American scientists. It was a great opportunity, and I applied and got it. But they didn’t know what they wanted to do with the Ecotron, and so they asked if I would be willing to take the lead on it, and come up with a project.

The Ecotron was a series of growth chambers in which you could control all the parameters – light, temperature, rain – everything could be controlled. We naturally thought we will do a global warming experiment – setup some chambers at ambient temperature, some chambers a little bit warmer, some even warmer and measure the response of artificially created biological communities.

But you know, the Earth summit in Rio happened in 1992 – around the same time we were planning these experiments – and one of the things that had come out of that summit was that biological diversity was considered, sort of like – what did they call it – like the underlying foundation of sustainable development. Why that was I’m not sure, but they definitely felt that, if you didn’t have biological diversity you wouldn’t be able to get environmental sustainability, and without environmental sustainability you could not achieve sustainable development. There were a number of people who believed that, but there was no actual proof. If you think about ecology, most of what was done in ecology worked the other way around – why is it that if you change latitude biodiversity changes? Why is it that if precipitation changes biodiversity changes? Why is it that if temperature changes biodiversity changes? So biodiversity was always a response variable – a victim of whatever the physical environment was. But the public was focussing on the opposite – that without biodiversity we couldn’t achieve these environmental goals.

Around this time, John Lawton – my post-doc. supervisor at Silwood Park – asked me to read his contribution to a symposium volume on the topic of whether biodiversity is important. And John was actually arguing that most species were redundant, and that if you lost them it probably doesn’t matter that much. And I read this and thought – I can’t believe people ask this question. My own feeling was – how could it not matter? I mean, if you lost biodiversity, surely an ecological system would just not work as well, right? And so there I was, with the Ecotron, and I suddenly thought – we could answer this question with the Ecotron. At our next weekly meeting, I got the people together, including John Lawton, and I said – instead of manipulating temperature, why don’t we manipulate biodiversity and see how that affects ecosystem function? We will set the temperature and other parameters identical for all the chambers, and then we will have one third of the chambers with high biodiversity, one third with medium biodiversity, and one third with low biodiversity. Then we will measure ecosystem function in these different levels of biodiversity. All of us post-docs got very excited about this, but John Lawton was not convinced. He said that what I was proposing was not something anybody does or thinks about, and that global warming was a big deal and very topical. He also said that, as postdocs, we needed to think about our careers and getting professorial positions. He just wanted to make sure that we had thought through this completely. And it was Sharon Lawler – who is now a professor at UC Davis – who said she liked “shooting from the hip”, and that my idea sounded more exciting. Lindsey Thompson – the other post-doc. – also agreed, after which John said we would have his full support. It was a very expensive project – it had three engineers, two technicians and the three of us as postdocs, who all had be paid. I forget how much it cost, but it was a vast amount of money to pay for running the project – getting all that soil, sterilizing it, buying a lot of equipment for the new idea – but John fully supported it.

It is interesting the way this idea came up. It didn’t really come from the scientific community initially. It was because of the Rio summit sort of getting everybody to believe that conservation is an important part of sustainable development. Before that, if you looked at sustainable development, people were worried about things like housing, food, health, water etc. Biodiversity did figure, but it was seen as something to protect and save. The links between biodiversity and food, health, water etc. were not talked about at all. Also, the view from conservation biology was not very utilitarian. It was mostly about the moral and aesthetic and cultural reasons to conserve biodiversity. The typical ecologist’s way of thinking was that we have to change things to save biodiversity.  In contrast, we were thinking that we have to change biodiversity to save things that are important to us. That was the motivation for this project. It’s a long story, but you can see it was somewhat serendipitous.

HS: You said there were three postdocs, including you, on this project, and John Lawton. Who is the fifth author on the paper?

SN: Richard Woodfin was a research scientist. He had a master’s in ecology but he was an engineer by training. And we really needed an engineer, because this was mostly “mechanical ecology”. Working in growth chambers is much more difficult than working in the field. Three in the morning, something would breakdown and you have to run in there. The rain would go off at the wrong time and we had to figure out why. Or there would be a power surge and the lights would blow. I slept at the station many nights just nursing the thing through. We had set “the sun to go down” in the middle of the day, so if I wanted to get there at “sunrise” that would be two a.m. So, I just slept in the office most of the time.  Richard Woodfin, our go-to person for troubleshooting the system, was unique because he was both a biologist and an engineer. He was considered as much an author of the project as others.

HS: You touched upon something I wanted to ask you next- what was a typical day like for you during this project?

SN: It was a long experiment and it was intense every day. We met weekly, and we would write out all the duties we had to do in a very regular fashion on the wall. We were collecting soil, arthropod and vegetation data, and monitoring all those environmental parameters. We really couldn’t go away anytime, because the machine was running all the time. Back then, I had never done growth chamber research. I was a field ecologist, so, to me, the idea that we would have this experiment upstairs, right there, running all the time, was strange. There was lot of maintenance too – the earthworms would all crawl out, so we had one person whose job, every morning, was to put all the earthworms back in. Or, the snails would leave the chamber, or get under the railings, and you had to find them all and put them back. Things like that. It was a lot of maintenance.

But this was England, where people don’t work on the weekends and have a tea break in the morning and in the afternoon, no matter what. When I complained to John Lawton that the libraries weren’t open during the weekend – we didn’t have the internet back then – he said nobody stays here late at night, or comes in on weekends, except the crazy Americans! I was one of those crazy Americans.

HS: So you were in the lab most of the time, during this period?

SN: Most of the time, yes. We did take time off. The engineers had those devices, what do you call them…

HS: Pagers?

SN: Pagers, they had pagers. And the Ecotron would actually automatically buzz them if there was a problem. The three engineers would be on call, in rotation. When they were on call, they were not allowed to drink. And to tell a British person that he can’t go to the pub is pretty hard on him. If the pager went off in the middle of the night, because the machine was overheating or something, whoever was on call would have to come in to trouble shoot. Some of it they could try to control remotely, or sometimes I would get a call – we have an alarm going off in chamber 6, could you just go and take a look to see if it’s a false alarm? If I was there I would run up and check, but if nobody was there they would have to come all the way in, sometimes only to find out it was a false alarm. But it could be quite serious – if the machine started to overheat, all of a sudden, 100 days into a 200-day experiment, you could lose the whole thing, because heat could over-stress the plants, insects, worms, snails, and all the other organisms in the chambers.

HS: How long did the experiments take?

SN: I think it was about 210 days. There was no season – we basically had the same conditions every single day for 210 days; that’s two-thirds of a year. So, in a way, it was like a very long summer.

HS: How did you decide on what species to use in the communities you constructed?

SN: That’s a really difficult question. Anybody who sets up a microcosm or mesocosm experiment struggles with two things: one is this desire to mimic nature – in your chamber or in your bottle. To capture nature in miniature. But at the other end of that spectrum, you just want a biological analog to test an idea you have. If you think that biodiversity matters, for primary productivity or carbon sequestration, then you simply want to have biodiversity that varies from low to high, and not worry about species identities. Also, we needed to pick species that would survive in the Ecotron. In preparation for the experiment, for a couple of years, we tried out lots of candidate plants in a mock chamber, and picked the ones that survived best.  But then you will always get the people who say – why did you put that species with that species when they don’t co-occur in nature? So we did the best we could, given these constraints. We chose what John Lawton used to call, the “weedy meadow”, as the model for our communities. Apart from plants, we also wanted other trophic levels, to simulate a real biological community – decomposers, herbivores, predators of the herbivores, and a below-ground community including Collembola and mites and earthworms. We also wanted soil bacteria, for which we took soil from the meadow, shook it up in water and then filtered the water, so that the bacteria could go through, but all the little insects and other invertebrates get removed.  What we finally had was very simple – a patchwork community with all the ingredients. But it wasn’t nature in a bottle. Rather, it was all the processes of nature in a bottle.

We had people coming to see the Ecotron all the time, and most of them were disappointed. Even Mahesh [Sankaran] said that he was really shocked when he saw the Ecotron; he expected something amazing and the Ecotron looked like a bunch of meat lockers. I think people expected to see some sort of miniature rainforest inside the Ecotron, with maybe parrots flying out! Once we had the local gardening club visit us, because John Lawton thought we should have good relations with our neighbours. I remember there was this group of gardening enthusiasts standing around, all eager to see what’s inside, and when I opened the door they were all shocked.  One elderly gentleman looked at it and said – “That’s a gardener’s nightmare!” I will never forget that. What he saw were weedy plants and slugs and aphids and white flies; all these pests. To us, thinking abstractly, it had all the essential ingredients to of an ecological community, and was beautiful. But to the visitors it was a nightmare.

We also had to vary diversity within each of the trophic levels in different treatments. That presented a special challenge, because we had to be really careful about always going from low to medium to high diversity when working in the chambers. Because if you went from high to low, there was the risk of accidentally introducing a new species – some tiny aphid in your hair or on your clothes – into the low diversity treatments. So we wore these Hazmat suits and slippers whenever we were in the chambers.

The other thing was we couldn’t open a chamber after three o clock in the afternoon because the sun had set in our artificial communities. And it was a real sunset – we actually shifted the red to far-red, and dimmed the lights to get the right balance to mimic a sunset. So, if we opened a chamber door after the sunset, all this light would come in, and there were a lot of plants that would respond to that, and it could alter their flowering and growth patterns. So, at three pm every day, the buzzers would go off, warning us that the sun was setting, and we would always be behind, so we would have to rush to complete whatever we were doing and get out. The other thing was the rain – again we had buzzers warning us that the rain was about to start, and we had to get ourselves and all our equipment out quickly before they got drenched!

HS: Sounds straight out of a sci-fi movie! Did the results of your experiments surprise you?

SN: There was a British Ecological Society (BES) meeting happening around that time, for which abstracts were due eight months in advance. We were only maybe halfway through the experiment, and John Lawton was already convinced that biodiversity wouldn’t matter that much. He said, as long as you had plants and herbivores, why would it really matter if you had two species or 16. But I was convinced of the opposite. I had no particularly empirical or theoretical basis for this – it was more a gut sense that diversity must matter. Anyway, by the time the abstract was due, nothing that had been measured was showing differences across treatments – nutrient loss, soil chemistry, growth rate, the total amount of standing biomass. So we sent in an abstract to BES which actually said that biodiversity didn’t matter! But at the meeting, which was 8 months later, when we had completed the experiments, I delivered a talk which said biodiversity did matter. Nobody seemed to notice that the title of the abstract was the other way around! The room was packed and it was the biggest lecture hall in the conference. I had never given a talk with so many people. Once, I gave a talk on the reproductive biology of a histophagus protozoan, and I think there were six people in the room. And they were all my friends! But at this talk I had to actually step over people to get to the podium. Everybody wanted to find out what we had found out in the Ecotron. Because it was controversial. It was a very expensive experiment, at a time when funding for research in England was pretty tight -right after Margaret Thatcher’s reign. So, I gave the talk, and the controversy started right then. People said terrible things – about the quality of the research, that we were prejudiced by our fondness for biodiversity, that we were seeing results that weren’t really as clear as we felt they were, or that these experiments were too artificial and did not reflect what’s going on in nature

Regarding your question whether the results were a surprise –I was surprised by how difficult it was to show. I obviously thought biodiversity mattered. And 22 years later, I am still surprised by how difficult it is to show. There are 8.7 million species on this planet, and they all play different kinds of roles – making nutrients cycle, energy flow, conditioning the atmosphere, conditioning the soil, conditioning the water – and they are all linked to one another. There have been numerous food web studies that have shown that biological communities are much more tightly linked than random. Yet, in spite of nature consisting of a tightly connected web of life, it was surprising how hard it was to demonstrate that if you dismantle it nature will behave differently.

To illustrate the issue, I have actually done this in class – only a few times because it is difficult to arrange – I bring a computer into class, take the back off, take a pair of pliers, and say I am going to pull a part out. And the computer is running. I wear rubber gloves, so I don’t get a shock. I randomly pull a part out, and ask the students how much would you pay for this computer now. The computer is obviously still working fine, but because I ripped a part out, the students feel that it has dropped down in value, usually by about a quarter. Around the sixth or seventh part, they say they are not going to buy it, no matter what. Yet, when we do this to nature (allow species or the parts of ecological systems to be lost), people don’t seem so concerned. Now, of course, the argument is that the computer is designed by an engineer and nature is not. But my argument is that nature does have a design to it.  For example, you don’t have to believe in a creator or deity, but you could say that, after 3.5 billion years of evolution, a bird has a design. It is a remarkable flying machine, and it could do this long before we figured out how to fly. I feel that ecosystems, after 100s of millions of years of biological interactions, have a design. I feel that our knowledge of how ecosystems work is incredibly primitive. A computer engineer can look at a computer and tell you this is a bad design, or that part needs to be replaced, or I know why this is not working. We know next to nothing about how our trillion tonne biomass machine, made of 8.7 million different kinds of parts, and actually trillions and trillions of interconnected parts, functions. And yet we are allowing it to become dismantled. I feel that the public can kind of grasp the idea, but it’s been very difficult to show why biodiversity matters. So, that’s my surprise. I am never really surprised when I hear a more biodiverse system is more stable, or more efficient, or more productive, or more resistant to invasive species. Those things don’t surprise me. What surprises me is how difficult it is to prove that. No one has actually shown that more biodiversity is bad. That would be problematic for me. But it has been very hard to show that biodiversity is good. And that remains the case.

HS: At the time when you did the experiments, did you have any inkling of how important the work and the paper might be?

SN: Not at all. Like most post-docs., we thought if we could just get a Nature paper it would really help our careers. So, we were very excited about that. In hindsight, we can also say that we all did really well by this paper.

But at that time, we did not expect that it would become so big, because the field was new, and nobody had thought about this. We didn’t think it would attract as much attention, say as a Nature paper on a well-studied subject, like climate change. Biodiversity-ecosystem functioning – even the term is so long and ugly that we call it BEF; we still don’t have a short term for it. But in a few years – I think two years later – it became the fifth most cited paper in global change research. Sharon Lawler and myself and Lindsey, we were really surprised, but even more so by the reasons for which it was getting cited. It was being cited so much because it was very controversial. There were many people who were pointing to it as misleading – overselling biodiversity. Or that it was badly designed. There were many critics of it. And in fact, a lot of the literature that then followed was addressing whether the results were being interpreted properly. Although, we ourselves had recognised, before the experiment was even done, that there were two ways in which biodiversity could improve ecosystem functioning; many assumed we were not aware of the two different ways to explain the results. One was that the high diversity plots simply had the key species in them, the ones which had the big impacts. Of course, it could work the other way too, i.e., high biodiversity plots could hold the species that had big negative impacts. The other possibility is that a more diverse community has more kinds of functions co-occurring. So you get more efficiency. In the chamber, if you have a tall plant, a short plant, a crawling plant, a plant that sort of grows on the side walls, and grasses that grow more centrally, it is going to be very hard for a photon to get to the ground, without being captured by a leaf. Instead, if you have a bunch of identically-structured plants, even if they have a decent amount of biomass, the light will go through and hit the ground. So, high biodiversity could have a positive impact on ecosystem function, either because it has the big impact species, or because it has a greater diversity of functional types. It did not occur to us, that people would feel, that if it is the former – the likelihood that you simply had big impact species in your high biodiversity – that that’s not a biodiversity effect; that it is simply a statistical effect. In fact, people who didn’t like our experiment called this the “hidden treatment”, implying that we had foolishly designed an experiment with a treatment we didn’t know existed and which sort of snuck up on us. It is no longer called that – it is now called the sampling effect or the selection effect. The functional diversity effect is now called complementarity. Like in the rhyme “The butcher, the baker the candlestick maker..”, a neighbourhood that has butchers, bakers and candlestick makers is better than a neighbourhood that has only butchers. A neighbourhood of people who complement one another, a diverse neighbourhood, is better than a homogeneous one. We recognised both these possibilities, but didn’t think it would become such a big deal. That’s why we didn’t even talk about it in our paper. I was actually surprised how more and more literature started to snowball on this.

HS: Was the paper’s journey through review and publication smooth?

SN: Yes. I think we were very moderate in the way we worded it. In fact, I think, in the end, I said something to the effect “to the extent that our results reflect what’s really going on in nature, they suggest that”. I think our closing paragraph was cautious. We recognised that it was a growth chamber experiment, and, therefore, inferences for the real-world need to be drawn cautiously. And Nature demands that.  When writing a Nature paper, you need to get as close to the edge of saying that you found something dramatic, but you can’t cross the line. It went through review quite smoothly. We did not have much resistance. Later on, partly because of the controversy created by our paper, and also because more people started working in this area, it became much more difficult to publish on biodiversity and ecosystem functioning. And I think that’s pretty standard in science. Like, in molecular systematics, earlier you could use a single mitochondrial gene to do phylogenetic analysis, but these days that just won’t do. The demands for phylogenetic research have just gone up and up and up. Gone are the days of Linnaeus, when you could just simply write your dichotomous trees. A friend of mine – he was a fellow graduate student – who has been very successful doing molecular phylogenies throughout his career, freely admits that he would never accept a paper, today, which does a phylogenetic reconstruction on a single mitochondrial gene. And yet, that’s how his career started.

Students don’t like to hear this, but I feel that we were very fortunate that when we started it was a brand new discipline. And we could get away with a lot. If we tried to submit this paper today I am sure it will be rejected, because it was not aware of the things that followed. So, I think it’s a universal thing about science that it progressively gets more and more complicated. So, if you get in in the early days, life is a little simpler.

HS: It’s been 22 years since the paper was published. At any time during this period have you gone back to the paper – read it again – for any reason?

SN: I use it in teaching. I am pleased that, now when I go to an ecology meeting, there is always a session or two or three on biodiversity and ecosystem functioning. It’s become sort of a standard part of ecology, and there are a couple of textbooks as well, that mention our work. I never imagined that my work would appear in a textbook! So, it’s become kind of a standard way of thinking. If I were to ask a student why is biodiversity important, most of them today will talk about complementarity. So, when I teach, I very often say let’s take a look at the first experiment on this topic, because it was very simple  – we just had high, medium and low biodiversity, and we measured carbon dioxide sequestration, which back then was very difficult to measure, but now it’s much easier to measure because there is better technology. Our study is also one of the few to have multiple trophic levels. Most of the biodiversity experiments since then manipulate only one trophic level, usually plant diversity and nothing else.

Thomas Kuhn says that a lot of the early experiments in science try to articulate an idea. I think ours is one such experiment – some people said biodiversity is related to the properties of ecosystems, so we said – okay, we will articulate that in the form of an experiment – manipulate biodiversity and measure the way the ecosystem functions. Another paper of mine – Bob Costanzawas the lead author – said that nature provides so much in the way of services that its monetary value would be twice the global GDP. That paper was also incredibly controversial. There are as many people who hated it as people who liked it. Bob Costanza said that the real value of that paper was not necessarily its content – there are much better data now, and we understand some of the mistakes we might have made – but that it provoked discussion. I often think of him saying that, and I feel that the great thing about the Ecotron paper also was that it provoked discussion. It’s funny – I am an ISI highly influential scientist, and I think it is primarily because of these two papers and a few others. But these are papers that don’t really need to be cited, except to say that these questions have been addressed for many years. I don’t think that many of the people who cite these papers today even read them – they cite them to put a date on the beginnings of the field.

HS: Is that what this paper mostly gets cited for these days – as the first one to test the idea?

SN: Yes, it mostly gets cited for being one of the first papers. At this point, there are so many more recent papers on the very same issue. For example, we now have papers that can actually statistically separate sampling effect from complementarity. Like I said earlier, while we were aware of these two effects, we didn’t discuss it at all in our paper, so there is no need to cite our paper when discussing sampling effect and complementarity. As an aside, we actually did a second experiment in the greenhouse, which was published later, to tease apart sampling or selection and complementarity, although we didn’t call it that.  It was done with the same species in the Ecotron but with only the plants. I don’t think it has been cited much. So, apart from saying that this question has been of interest for 22 years, there is not much reason to cite our 1994 paper. If you wanted to say that biodiversity effects consist of selection and complementarity you would probably go to Hector & Loreau, or something more recent. Because they are better examples. If you wanted to say biodiversity influences nitrogen cycling you would probably go to Tilman or Peter Reich, because they had much bigger experiments – hundreds of plots – and they were outdoors. So, our paper is mostly of historical interest.  If you look at the track record of most papers – I think about 99% of them are cited a very small number of times over the space of a few years – 3-4 years – and then they are hardly ever cited again. And then, there are a small number of papers which continue to be cited long after they were published. I was telling students about an old paper –Hairston, Slobodkin and Smith (1960) –  HSS, as it is known – it is still cited today, but mostly to say that this idea, about why the world is green, is still around, and it’s been there for a long time. Papers like that are very few. I don’t think our paper had that kind of impact, because the idea was a very broad one, and there are better studies, much bigger studies now.

HS: What would you say to a student reading your 1994 paper today? What should he or she take-away from it?

SN: I would say, to notice that, compared to a lot of the literature that you read in ecology today, what we did was fairly simple. Also, that most of what you do in science consists of looking at stuff on bivariate plots. We tend to think X versus Y, and usually biodiversity is on the Y axis, and X is something like habitat fragmentation, ocean acidification, elevated CO₂ etc. What was really bizarre about us was we had biodiversity on the x axis, and CO₂ on the Y axis, and I think that is what really got people upset. To say that atmospheric CO₂ was being driven by biodiversity seemed almost stupid – like we had got the R code backwards or something (you didn’t mean to plot it that way, right?). But, no, we did mean to plot it that way. Nowadays, it’s not uncommon at all to have number of species on the x axis. Nobody will get surprised by that at all. But then, it was a big deal.

What I really liked about what we did was, not so much the actual ecological or environmental issue but, the fact that we thought differently, and that we were all excited about thinking differently.  Earlier, I told you something that nobody knows – that the experiment originally was, in fact, to have biodiversity on the y axis and temperature on the x axis. But then we had this idea, all by accident – John Lawton said take a look at this paper, proofread it, and that got me thinking – it came out of nowhere and we decided to pursue it.  It was risky, and John Lawton recognised that right away – he was going to support it, but he did recognise it was risky for our careers, and made sure we realised that.

These PhD interviews are going on right now in NCBS, all these professors are questioning them, and they know they have to master the state of current knowledge, but what will really advance their career is if they can actually breakaway from the current state and try to do something that will change that. There are three different qualities we value in research: one is to just do excellent work – this is the definitive study, it was beautifully done. The second is to do synthesis, where you say – well, somebody was thinking about this chemical phenomenon, and somebody was thinking about this biological phenomenon, and we put the two together. And the third is discovery. But it’s very hard to write a grant proposal and say – give me money for 5 years and, by accident, I will come up with (discover) something new. And yet, there are all these incredibly important discoveries – the idea of natural selection came to Wallace in a malarial fever dream, the fellow who discovered the structure of the benzene ring got it from a dream about monkeys, many discoveries in chemistry came from accidents in the lab – somebody tipped something and the colour changed and the whole colour industry in Germany was born. But we also know that “Chance favours the prepared mind”, so all these students need to prepare their minds, so that if something really interesting comes along, by accident, they seize the opportunity. It is very hard to design a research programme that optimises this. All you can do is be prepared to make use of unforeseen and unplanned opportunities.

HS: Recently, I read somewhere about the importance of “limited sloppiness” in science. I have forgotten who coined the phrase – it was a famous chemist or biologist. 

SN: I haven’t heard that one, that’s a great term.  Paul Feyerabend’s philosophical approach was completely different from Karl Popper’s. The Popperian method says – construct your null hypothesis, have an alternative, if you reject the null, favour the alternative. You never really wander very far, you just keep going through this chain of bifurcating hypotheses. Feyerabend, on the other hand, said – he didn’t use the term ‘limited sloppiness’ – it was important to be able to, first, recognise what constitutes a surprising event, and then capitalize on it. Don’t be constrained by orthodox methodology. I will give you an example. Brad Cardinale and Martin Solan had a bunch of data in which they were looking at marine species – marine biodiversity in the sediment. They had dropped cameras into the sediment and photographed invertebrates.  Martin said he had all this data and was interested in doing something on biodiversity-ecosystem functioning – sediment turnover was the function in this case.  So we got together – I was not on the paper, but I had a grant to bring people together to do this sort of stuff. So, what happened was Brad, who took the lead in analysing the data, got an asymptotic curve on the biodiversity ecosystem function graph. But he got two curves instead of one. He first thought that there was something wrong with the simulations, so he kept trying to rewrite the code, but he couldn’t get rid of the second curve. Finally he went to Martin, who is the marine biologist, who took a look at the data, and you know what they discovered? The difference between the two curves was a keystone species.  If the keystone species was present the whole system moves to another level, when it was absent it dropped! This turned out to be a really interesting finding, the paper was published in Science, and attracted quite a bit of attention. Just imagine if Brad had managed to “clean up the data”! The first reaction to something anomalous is that there’s a problem with the study. This might be true, but be open to the possibility that something really novel is actually what is going on.

HS: One final question – maybe a difficult one- is this your favourite paper?

SN: No, not really. It is probably the most important paper in my career, but it is not my favourite. You know, my most favourite paper was rejected so many times I gave up on it. It hasn’t been published. I had used a method called angular statistics, on predator-prey cycles, and was able to show that they really weren’t cycles. I thought it was neat, and nobody had ever done that before. But it’s been rejected and rejected and rejected. So my favourite paper is only mine!

I think, what I didn’t like about the Ecotron experiment, was that it was indoors, in artificial conditions. I like being in the field. I have done a lot of laboratory research in my career, but I would say that my dissertation work, where I was in the rainforest of Costa Rica, working on 30-40 species of insects inside Heliconia flowers, that was when I really felt I was doing what motivates me – understanding how nature works.  And to me, to have worked in the middle of the jungle, collected all that data, and discovered a pattern that helped me understand this remarkable thing called  a tropical rainforest, was so much more rewarding than working with the “gardener’s nightmare”!