In a paper published in Ecology in 1997, Oswald Schmitz, Andrew Beckerman and Kathleen O’Brien experimentally tested the relative importance of direct and indirect effects of predation in causing tropic cascades, Using an old-field system that included herbaceous plants, a herbivorous grasshopper and a spider predator, Schmitz and colleagues showed that indirect effects of predation, expressed in the form of antipredator behaviour, can result in trophic cascades that are similar in form and strength to those resulting from lethal effects of predators. Nineteen years after the paper was published, I spoke to Oswald Schmitz about his motivation to do this study, his memories of field work and what we have learnt since about the role of direct and indirect predation effects on trophic cascades.
Citation: Schmitz, O. J., Beckerman, A. P., & O’Brien, K. M. (1997). Behaviorally mediated trophic cascades: effects of predation risk on food web interactions. Ecology, 78(5), 1388-1399.
Date of interview: 30 November 2016 (via Skype)
Hari Sridhar: I wanted to start by asking you a little bit about your motivation to do this work. A lot of your initial work was on mammals and you did a PhD on whitetail deer. I think it was the early 90s when you published your first paper on spiders. What was your motivation to do this particular piece of work in 97?
Oswald Schmitz: It’s an interesting story. I did my PhD on what was then called risk-sensitive foraging. That was basically a body of theory that predicted how animals should balance starvation risk against surviving in inclement conditions. It was really thinking about how birds managed to survive overnight because of fat levels in their body and whatever. I extended the theory to study mammals in winter, like whitetail deer that, both, are reproductive but also have to survive cold and not very good climate. And so, you know, I was developing this theory and testing it, and all the while people were starting to think about predation risk as an alternative way of thinking about risks and behavioral ecology. By the time I finished my PhD, the whole idea of risk-sensitive foraging became passé – and it was rather quick – in maybe five or six years it became passé. And so, you know, I finished a PhD dissertation that really wasn’t going to have an impact, and, so, I was a little frustrated by it. I’ve always been interested in predator-prey interactions – I did Master’s work on that – but I wanted to do things more experimentally. At the time, there was a paper written by a person named Tom Schoener, who had published a synthesis of food webs and was really trying to understand food webs. And I said, this is where I want to go. So that’s why I started pursuing this. The reason why I work with spiders and grasshoppers is because my PhD advisor had worked on optimal foraging and community structure with grasshoppers, and I’d helped him in the field a couple of times, and I figured this would be a really nice model system to work with, and it has paid off.
HS: What was your PhD supervisor’s name?
OS: His name was Gary Belovsky. He was in fact, a PhD student of Tom Schoener. What’s interesting, though, to get into the risk effects paper, at the time, people in aquatic ecology were testing ideas of trophic cascades, and people were saying that in terrestrial environments, trophic cascades were not going to operate. And, you know, I said to myself, there weren’t really a lot of tests in terrestrial systems. And so, I decided to test trophic cascades theory in terrestrial systems with these experimental communities that you could, actually, see fast responses with. And what’s really interesting is I had vowed to just give up on behavioral ecology and get into population and community ecology, and when we had done some of the early experiments, like in this paper, we had done the manipulations – you know, how you would test for a trophic cascade – and we found that the predators did not have a significant effect on the grasshopper densities, but there were significant changes in the vegetation. This, at first didn’t make sense, in light of the classic theory that if predators killed their prey then there are fewer prey to eat the vegetation. And then it dawned on me that, you know what, there’s a predation risk effect going on. And if you look at the time budgets and habitat selection, the story all came together. And what was serendipitous there is, the fact that, in my PhD, I’d done the behavioral ecology, I knew that body of theory and I was able to draw on it. Had I not done the behavioral ecology, I would have probably written it off as a failed experiment! And so, these are some of these serendipitous things that happen in your career – you vow to give up on it and later on you end up discovering that what you’ve learned is really, really helpful to interpret other kinds of things.
HS: Stepping back a bit, could you tell us how you got interested in behavioral ecology?
OS: My PhD advisor had developed some mathematical models in optimal foraging theory and applied it to large mammal foraging. He did a classic study on moose. And I had learned about his work when I was taking a wildlife nutrition course as an undergraduate and came upon this work and I was just fascinated, captivated by the fact that you can use economic theory to predict nutritional intake of animals. I wanted to go and work with him and do optimal foraging and really try to understand animal foraging behavior and habitat use using that body of theory. I was very interested in mathematical modeling and being able to test predictions from models using empirical analysis. So that’s why I got into looking at the risk-sensitive foraging in the deer.
HS: Was this paper the first time you did this kind of experiment, i.e. gluing the spider’s chelicerae to prevent feeding?
OS: Yes. That was the origin. And again, the reason we did that was to really validate the fact that it is the predation risk that’s driving most of the response.
HS: Could you tell us a little bit about how you came up with the idea?
OS: I’d seen a poster at a conference by somebody who was interested in just the behavioral ecology of fear. I forget who it was, but it was a poster at a conference and they had, actually, used paraffin wax and done short term behavioral trials and showed how prey responded just to the mere threat of predation. They were testing the predation risk ideas that were out there in behavioral ecology. And so, I said, we can adopt this method, but we need something else, because the wax isn’t as long-term. We needed something that would last an entire field season and that’s why we used glue, it amounted to Krazy Glue, but it was a surgical cement at that time because it’s non toxic. Andrew Beckerman, who was my graduate student and an author on this paper, had a friend in the medical school. I guess, when they do measures on people or animals they glue things to them, but they want to avoid toxicity, so they use this fast-drying durable cement. And I don’t know what the product name was, but it turns out Krazy Glue has the same properties and Krazy Glue is not very toxic either, so we were able to use that eventually.
HS: Did you try other stuff before finally deciding on Krazy glue?
OS: No. We wanted something that we knew would be durable and that the spiders couldn’t just rip off, you know; the paraffin wax they can. What we needed was to be sure that they were glued for the entire duration of the long-term experiment over the season.
HS: Was this poster you saw at the conference also on spiders?
OS: Yeah, it was on spiders. I think the spiders were hunting crickets, or something like that, and it was done in a lab setting. So again, you know, it’s the value of going to conferences and seeing what other people are doing.
HS: Do you remember which conference this was and in which year?
OS: It would have been an Ecological Society of America conference, probably, around 1995.
HS: Could you tell me how the three authors on this paper came together and what each person brought to the study?
OS: Yeah. Andrew Beckerman was a doctoral student of mine and Kathleen O’Brien was a Master’s student. Kate had done the bench-top behavioral ecology and Andrew had helped me with the field experiment, keeping it running and managing the mesocosm part of it.
HS: Tell us a little bit about the area where this study was done, the Yale-Myers Research Forest. At this point, had you already been working in this area for a while and how did you choose to work here?
OS: It’s owned by our School of Forestry and Environmental Studies. It’s a nice place to go where you have access to old fields, where we could do this research without having to, you know, get landowner permission. It was also a safe place to work because it’s restricted access – it’s owned by Yale – so people couldn’t go in and vandalize the experiments. It’s often hard to do it in farmer fields because farmers, typically, don’t want all their land used up for an experiment or some people might view it as unsightly. So, it was really key to be able to work in this forest.
HS: Had you already been working earlier on the spider and grasshopper species used in this study?
OS: Now that was serendipitous, too, you know. Well, certainly, Melanoplus femurrubrum, the grasshopper, yes, because I’d worked on that with my advisor in Montana. That grasshopper species is ubiquitous across North America. The choice of spider, again was really, really lucky. There’s serendipity there, again, because we had just gone in the field and found the most dominant or most abundant species, which happened to be a sit-and-wait predator, and decided, okay, we’ll just work with those as a starting point. And that was the most abundant predator in the field, you know. Had it been one of the other spider species that we’ve used, we would have never discovered the behavioral shift, you know, because other spiders don’t cause that. And so again, you know, there’s a lot of luck involved in your research. And this was lucky again. It basically opened up a whole research career.
HS: When you went there looking for the spider, did you already have this experiment in mind?
OS: Yes, that we did.
HS: Could you give us a sense of what this work involved? What was your daily routine when this work was being carried out?
OS: Yeah.It was sporadic work. Basically, you have to build the cages, which can take a long time if they’re made from aluminum screening, and then you have to sink them into the ground to seal them so that animals can’t escape out the bottom. So you have to dig small trenches and then sink them in and then close off the trench against the screening to hold them tight into the ground. And then you have to go and use sweep nets to catch the spiders and the grasshoppers, and you have to do that all fairly quickly, in the early part of the year, so that you can catch them in time that they will actually have an effect over the growing season. You have to get it done early on so that they all develop together in the field and in the cages. So it’s two to three weeks of frenetic work at the beginning of the summer, usually, starting at the end of May till the middle of June, and then, every week or so, we would go in and just do a count, where you actually get on the ground, you would blow into the cages to get the spiders and grasshoppers to rise, and then you’d count their abundances in the cage and monitor their survivorship over time. That’s what generated those survivorship curves and allows you to understand the trajectory of predation and predation risk effects. And then at the end of the field season you, basically, tear down the cages, clip all the vegetation, sort it to species and then dry it and weigh it. So, pretty much, you know, it’s just monitoring from the middle of July till September, when we took the experiment down. And then the behavioral trials happen in cages, but we put them on wooden bases, and then we cut out sod from the field and put the sod into the cages and then put the animals in. The cages were gridded so we know where they were in space, and we could look at how they were interacting with each other.
HS: Was everything done in the field site?
OS: In the field site. Yeah. The behavioral trials were done outdoors on a bench-top. We wanted ambient conditions, not fully enclosed laboratory conditions.
HS: Could you give us a timeline for this study?
OS: Well, you know, we did some pilot work, just to see whether things would work out. And then, you know, we got a sense of how the system was working. And then that begs the questions about the risk and all that. So then we designed a bigger factorial experiment where we had plants by themselves, plants with the grasshoppers, and then plants with predation spiders, and then plants and grasshoppers with risk spiders. And that was replicated. Also, we wanted to make sure that the result was consistent from one year to the next, so we replicated across years – patiently did it over a couple of years.
It was controversial that trophic cascades occurred in terrestrial systems to begin with, and then to actually assert that it’s behavior driving it, is even more controversial, because, at the time, people thought that behavior is one of these stimulus response things and it would attenuate very quickly. And so it was reasoned that behavior shouldn’t matter over the course of a whole growing season. But the paradigm was wrong, because it’s really chronic behavioral response to predation that is driving it. So, you know, the study revealed that current thinking in behavioral ecology as well as in population community ecology wasn’t sort of appreciating the importance of the role of chronic risk behavior in these systems.
HS: Were all the authors involved in doing the fieldwork and the lab work?
OS: Yeah, we helped each other out. But, Kate, for example, did most of the work on the behavioral trials on the bench-tops. Andrew and I were involved in stocking the cages and monitoring them. Andrew spent a lot of time doing the monitoring on a week by week basis.
HS: Could we go over the Acknowledgements to get a sense of how you knew these people and how they helped?
OS: Yeah.
HS: A. Joern.
OS: Anthony Joern. Yeah. He is an eminent grasshopper ecologist and he provided helpful discussion and I was bouncing ideas off him.He was an inspiration for the work and also another valuable individual to discuss grasshopper ecology with.
HS: K. Johnson.
OS: Kris Johnson was another doctoral student. He helped, you know, a little bit with fieldwork once in a while, especially sinking cages, but he also, I think he read a draft of the paper and gave comments on it.
HS: G. Mittelbach.
OS: Mittelbach is somebody who has done a lot of work on predation risk and population ecology in fishes. He was also somebody that we bounced ideas off of. He might have actually been a reviewer on the paper.
HS: K. Rothley.
OS: Kristina Rothley was a doctoral student in the lab. And again, you know, she probably reviewed the paper and gave us comments on it before we submit
HS: L. Rowe.
OS: Locke Rowe. Again, somebody who read the paper and gave us critical feedback. He’s a professor at University of Toronto.
HS: D. Skelly.
OS: David Skelly is a colleague here at Yale. He also read the paper and gave comments.
HS: M. Uriarte.
OS: She was a master’s student. She would have helped, maybe, sink the cages a little bit but also, you know, provided discussion and read the manuscript and gave us comments on it.
HS: In the next sentence you name some of these people again for “help with the ‘crunch periods’ during the field season. I’m guessing that’s to do with setting up the cages and all that.
OS: That’s right. Exactly.
HS: In one place in the paper you cite unpublished data from ‘Schmitz and Beckerman’ to say you that spiders are least active at a particular time in the day. Was this published later?
OS: Andrew did that as part of his PhD dissertation, so the data would be in his dissertation. But I don’t think they ever got published in a journal article.
HS: In another place you say, “A detailed, complementary foraging study on this system showed that grasshoppers also exhibited a significant diet shift when they were exposed to the same levels of predation risk as in this study.” For this, you cite unpublished data by Rothley and Schmitz. Again, was this published later?
OS: Yeah, that eventually did get published in a journal called Evolutionary Ecology.
HS: Was Figure 1 hand-drawn?
OS: Yes. I draw all my figures. All the animals. If you look at all my publications, those are all hand drawn by me. Yeah, that’s sort of my art-natural history bent.
HS: Do you remember how long it took you to write up the paper and when and where you did most of the writing?
OS: It came together very quickly. I think we wrote it in the course of a couple of months in the fall of the year before we submitted it. It was done on university campus, going back and forth with my co-authors. I’d do a draft and then they would do editorial comments and we’d go back and forth.
HS: Did this paper have a relatively smooth ride through peer review and was Ecology the first place you submitted it to?
OS: Yes, that’s the first place we submitted it. The journal rankings have changed but, at that time, Ecology was the very best journal you could publish something in. The culture in ecology has changed. The more specialized journals now have much higher impact.
HS: Did it have a smooth ride through peer review?
OS: I think it was normal. It was fairly smooth. The people understood and appreciated the work that was done, they gave some comments to improve the quality of it. It wasn’t really an onerous thing. It was just sort of modest revision and then it was accepted.
HS: At the time where it was published, do you remember if the paper attracted a lot of attention, both within academia and in the popular press?
OS: It attracted attention in academia. Because, again, it was one of the early demonstrations of trophic cascade in a terrestrial system;and also because of the behavioral ecology driving it. But there was no public attention, really. Some of the later papers, where we showed more widespread effects of the behavior, did get some press interest. But, you know, in those days, we didn’t really care about letting the press know. It’s a very different ballgame nowadays. The media wasn’t really as heavily involved in communicating science.It was more about publishing in excellent journals.
HS: You mentioned that the idea of a behavior-mediated cascade was new. Was it considered controversial then and it did it get any push back?
OS: Yeah. There were doubters. There was a theory paper written by a guy named Peter Abrams, who sort of laid out the theory, why it should work. So, there was mathematical theory to argue for it. This paper presented a test of that. And I think people sort of bought the theoretical arguments, but they didn’t necessarily believe that it was universal. A few of the detractors thought the result was rather idiosyncratic; that it wasn’t going to be a general thing.
HS: Did you anticipate at all that that this piece of work would have the impact it has had?
OS: No. I’ve taken the attitude that you can’t really bet on which paper is going to really have the impact and which isn’t. I published another paper, actually, that same year in Ecology, which I thought would have the bigger impact, simply because of the way people were thinking about predicting food web interactions. Totally opposite outcome that this paper has had a much bigger impact than the other one. So it’s really hard in this game to judge which paper is going to catch on. If you really look at the trajectory of it, it had a slow citation amount, because, first of all, there weren’t many people interested in ecology doing trophic cascades work, and second, thinking about the overall ecology and population community ecology synergistically wasn’t part of mainstream thinking. So there weren’t very many people really working in that area. I don’t want to say I was ahead of my time, but it was pushing the envelope a little bit. Nowadays, obviously, we have these ideas of landscape of fear and everything, where this kind of thinking is just common practice. It was really not out there back then.
HS: Do you have a sense of what it gets cited for?
OS: Early on, I think it was cited for both the fact that trophic cascades work in terrestrial systems and that predation risk can drive that. It doesn’t get cited much anymore. But in the heyday of predation risk and food chain interaction, when these ideas started really coming to forth, which I would argue is sort of late 1990s to the mid 2000s, it was cited a lot as an early example of predation risk mattering to food chain interactions.
HS: Did this paper have any kind of direct impact on your career? And was it important paper in influencing the trajectory of your research?
OS: Oh, most definitely. It inspired me to, really, think, philosophically, about how does reductionism explain the whole and what sort of mechanisms should we focus on to really try to understand whole ecosystem functioning. Plus, it opened up a door to connect evolutionary ecological thinking with ecosystem function, which is what I’m doing now, really thinking about traits. And the work I’m doing now on predation stress and physiological changes and how that drives nutritional demand, it has really opened up this stuff for me, to use the reductionism to explain and predict the context-dependency I see among different study systems.
HS: Today, would you say the main conclusions of this paper still hold true more or less?
OS: Yes, they would, most definitely. The main conclusion is that a sit-and-wait predator induces risk, that chronic risk changes the behavior and physiology of the prey, and has cascading effects on plant biomass. And it is a generalizable principle.
HS: If you were to redo these experiments today, would you would you change anything about them, given the advances we’ve seen since then in technology, the development of theory, etc.
OS: I still use the same methodology. I’m not a big technocrat when it comes to trying to understand nature; the experimental approach is the same. What I would love to try and do is to not use mesososms as much anymore, and be able to track individuals just unencumbered by mesocosms in the field, and look at their movements a lot more and their spatial ecology. And that’s where, if we could get some sort of way to remotely sense these individuals and monitor their movements, that would be be incredible. We have tried it in the past, you know, you dot a color onto their carapace or whatever and watch them move. For me, being able to do movement ecology and really follow them over large distances would be really really fun to do. And if we had the technology to do that, that would be cool.
HS: Do you continue to work in this field site today?
OS: Yes, although, it’s slowly filling in with trees. Old fields undergo succession and eventually fill in. But what’s been amazing is, you know, you look at these little creatures, they seem benign, like, when you walk in a field, you don’t see them right away. They’re so small relative to the vegetation. But they hold back succession.The trophic interactions really arrest succession, and that’s why I’ve been able to work in those fields still today, because they haven’t completely filled in. But if those insects were gone, the field would fill in way faster. We know that from excluding herbivores and spiders and subsequent plot experiments where it just quickly filled in and succession really went gangbusters.
HS: Do you still work on the grasshopper and the spider species you used in this study?
OS: Yeah. They’ve been the workhorses for my main research program.
HS: What about experimental setup you used – the cages and the terrarium and all of that?
OS: We’ve modified the designs to make the building more efficient and getting them into the ground more efficiently, but, essentially, we use the same apparatus.
HS: You say “We do not yet have adequate replication to present the results of experiments involving adult grasshoppers.” Was this something that you replicated more and wrote up later on?
OS: Yes, there’s a companion paper in the Proceedings of the National Academy of Sciences on which Andrew Beckman is the lead author. That one talks about what the adults do. What happens at the adult stage is they’ve escaped that potential window where the predators could be effective in killing them.They’ve become so big that the predators can’t kill them. Yet, the grasshoppers still do respond. So,the predators were a threat. I don’t know why they do that, but it’s just really interesting that they still do.
We’ve repeated the experiment with varieties of predators and shown that different hunting modes of predators do different things, and what top predator is present can determine what the plant community structure looks like and has effects on ecosystem functioning, like nutrient cycling. We’ve been, really, able to build mechanistically from that.
HS: You point out that some of the results are only marginally significant and more replication is required. Did you later add more data?
OS: Yeah.We have done that. And the other interesting thing is what we’ve repeated these experiment over years – you know, the core experiment, we might have pursued other aspects in a factorial design – but the core experiment was repeated probably seven to 10 times over the years. And it turns out that as the environment warms you get stronger effects. So the years that we did those experiments were on the cooler side. When it becomes warmer, those marginal effects go away, with the same level of replication. But where the power comes in our study also is we were able to replicate the same qualitative outcome time and again.
HS: Why do you think there is a link with temperature?
OS: The grasshoppers can forage more, right; they are getting closer to their thermal optimum. As it warms up, it can be much more responsive to predators and forage quality.
HS: Has this phenomenon been documented in other taxa?
OS: Oh, yeah. It’s part of mainstream food web ecology now, and trophic cascades literature. People have shown it in agricultural systems, where you have pests and predators. They’re starting to show up with large mammals now, too. And certainly the aquatic ecology literature was already abundantly aware that these risk effects matter. I can’t think of a taxon nowadays that, probably, doesn’t demonstrate some sort of risk response in a chronic way.
HS: In the paper you say that these two independent lines of research, the direct and indirect effects, must be integrated. At that time, you say that conventional models of community dynamics do not account for the behavioral indirect effects. Has that happened since?
OS: Yeah. People include behavior in the food chain models nowadays. And we’ve even developed ecosystem models to look at carbon and nitrogen cycling that includes stress and behavioral ecology in them. So, yeah, the theory has also kept pace. There are a whole variety of papers looking at those kinds of things. I have papers on individual-based modeling where it shows that behavior causes emergent effects. In terms of the food chain, like the ecosystem-level affect, that’s theory developed by a person named Shawn Leroux and myself. People like Andy Sih and Bob Holt and Peter Abrams have spent a whole career developing mathematical models looking at how predation risk drives community dynamics.
HS: Have you ever read the paper after it was published?
OS: Once in a while. I haven’t read it in a long, long time.
HS: In what context do you go back to it?
OS: Just to see how my writing has improved over time.
HS: When you compare this paper to papers you write today, do you notice any striking differences in the way you write?
OS: Yeah, I mean, first of all, the norm in the field is very different. Back then you’d developed a very scholarly introduction to demonstrate your command of the literature and the papers were a lot more long-winded than they are nowadays. There was room for a little more speculation. But, nowadays, you have to cut right to the chase, mainly because journals want shorter articles, but, also, the readership isn’t as patient. And that’s a big frustration of mine, nowadays, that this younger generation is not as scholarly as they should be. And that’s a sad reflection driven by the way we publish articles. But then, the other challenge also is to do a better job of storytelling. And that’s something I’ve learned, that when you’re writing a scientific paper it’s no different than if you’re writing fiction, in the sense that you want to have a nice compelling narrative line. The difference between fiction and the science paper is that you build in evidence, you know, you actually support your story with data and information. That’s, sort of how I’ve evolved my writing, to recognize I’m telling people a story and trying to embed the evidence in a storyline.
HS: Would you count this as one of your favorite pieces of work?
OS: Yeah. I’m especially proud of that paper.
HS: Why?
OS: It was one of my first really truly integrative papers, where I was able to look at behavior, population biology, trophic interactions, and really put a coherent story together. And, I was able to write it in a short format and get to the point, and, actually, report on a really exciting discovery.
HS: What would you say to a student who’s about to read this paper today? Would you add any caveats?
OS: Um, no. Maybe just look at it in a historical context. Try to understand the context and that some of the caveats that we have in the paper are really in there because of the more skeptical readership at that time than there would be now for that topic. And really try and understand the history of that paper in relation to the way the field was thinking about things.
HS: Would you also point them to more recent papers to read along with this. If they had to read a follow-up paper, which one would it be?
OS: Um, probably the 2009 Ecology paper. Also, Schmitz and Suttle in 2001 where we looked at multiple predators. You know, and that’s where we came to the realization that different hunting modes of predators cause different effects. Then there’s a Schmitz 2009 that really of starts looking at how these things go from the community to the ecosystem level.
0 Comments