In a conceptual paper in the American Zoologist in 2001, Judith Bronstein highlighted the need to understand the costs of mutualisms along with its benefits, and laid out a research agenda for future empirical work in this area. Sixteen years after the paper was published I asked Judith Bronstein about the origins of her interest in this topic and how research on costs of mutualism has progressed since the writing of this paper.
Citation: Bronstein, J. L. (2001). The costs of mutualism. American Zoologist, 41(4), 825-839.
Date of interview: Questions emailed on 14th December 2017; responses received by email on 16th December 2017
Hari Sridhar: I would like to start by asking you about the origin of this paper. In a footnote you say this paper was from a symposium on plant-animal interactions at the Annual Meeting of the Society for Comparative and Integrative Biology in 2000 in Atlanta. Why did you decide to speak on this topic at the symposium? I also notice that this was around the time when you wrote a chapter on Mutualisms in the edited volume “Evolutionary Ecology: Concepts and Case Studies” and a couple of other conceptual papers on mutualism. What triggered your interest in topic at this time?
Judith Bronstein: These were the questions I had long been interested in – really, since I was a grad student. But I was only just gaining in confidence to write big conceptual papers. I had received tenure – which depended on showing my university I could write original data papers, not reviews and conceptual pieces. I was on sabbatical at Duke University, talking to very conceptually oriented colleagues like Bill Morris and Will Wilson. I was starting to feel as if I had some things I wanted to say of a general nature. Things like this aren’t too easy to publish in regular journals, so most of my conceptual work (in that era) was published as invited papers linked to symposium talks, or as book chapters. At the same time, I was publishing data and modeling papers the usual way, i.e., submitting them to the big journals.
HS: Stepping back a bit, could you tell us how you got interested in the fig-fig wasp system, which you studied for your PhD?
JB: I remember this very specifically. I knew the general questions I was interested in studying in graduate school, but I was at a total loss as to what system to work on. I kept trying to find a system to fit my exact questions, and nature doesn’t tend to comply with that approach! I actually knew a fair amount about figs and fig wasps because W.D. Hamilton was a professor at University of Michigan, where I was in grad school, and I had heard him lecture on them. But it never really caught my attention – the aspects that fascinated him, like fig wasp sex ratios and behavior, were pretty far from my own interests, which were centered in plant/pollinator community ecology. In my second year in graduate school, I took the OTS tropical ecology course in Costa Rica. I was terribly worried that I didn’t know what I was doing in grad school. And I was very reluctant to share my concerns with anyone. But I really got along well with one particular professor who came to work with our course: Steve Hubbell (subsequently the founder of the neutral theory of ecology). I sat with him one evening and poured out my concerns, and tried to explain my research interests. He said to me – in exactly these words – “have you ever thought about figs?” And everything Hamilton taught me about figs, and that I had learned more casually on my OTS course, came together magically. I knew that THIS was my system.
Steve then invited me to come work with him for the summer in Panama (mapping his famous tree-diversity plot), and suggested that I could work with the fig researchers there as a way to get to know the system. That’s what I did. For many reasons I ended up going back to Costa Rica, in fact back to the site where I had that first conversation with Steve Hubbell, to do my dissertation research on Ficus pertusa. I owe him a huge debt!
My long-term conceptual interests in mutualism as a whole, and in the costs of mutualism in particular, had their origin in a single question I was asked after a talk I gave during my final month of graduate fieldwork in Costa Rica in 1983. The person who asked the question was renowned evolutionary biologist Douglas Futuyma. In my dissertation research, I had been looking at fig reproduction (the number of seeds they made and the number of pollen-carrying wasps that develop at the costs of some seeds) strictly from the plant perspective. During the question-and-answer session, Doug asked, “What about the evolutionary interests of the fig wasps? Shouldn’t they be in conflict with the interests of the fig trees?” I had no answer. At the time there was simply no one thinking about whether or not the fitness interests of mutualists aligned or didn’t align. That really became the centerpiece of all my subsequent research on all the systems I’ve ever studied. You can see this theme clearly in this 2001 paper.
HS: A part of this paper is based on field data you collected on Ficus aurea from February to December 2001. Could you tell us a little more about this fieldwork – where was it conducted, who all were involved, where was the lab work done, what was your daily routine during this period etc.?
JB: I think you mean 1991. The work was conducted between 1991 and 1993, in collaboration with Martine Hossaert, who had an appointment as research scientist at University of Miami, Florida. The work was conducted on a number of very large Ficus aurea trees that grow right on the beautiful, landscaped campus. These are convenient because they have very low branches, allowing us to bag figs, expose figs to pollinators, and so on without needing to work on ladders. It was very low-tech work! We would daily record the developmental stage of figs on tagged branches of 10-20 trees. We could only do experimental pollinations of those figs when individual figs (syconia) were in two specific stages: when some were releasing pollen-carrying wasps, and others were receptive to pollination. This was rare and you had to be ready for it. When we sensed that this was coming, we would take figs about to release wasps into the lab (Martine’s lab in a nearby biology building), and leave them in covered petri dishes. The wasps leave the figs, and mill around in the dish. The dish would then be taken into the field, and individual wasps picked up with a tiny paintbrush and placed on figs that looked receptive. If we were right (the wasps knew; we couldn’t tell), then the wasps would burrow into the fig. We would allow 1, 2 or 5 wasps into the fig, then label and cover the fig. Weeks later, when the fig was ready to release wasps, we’d collect it, bring it to the lab, and allow the wasps to depart. We would count them, as well as the number of seeds they left behind. So, a lot of conditions had to be exactly in place to get even one usable data point (i.e., one fig with a known number of foundresses introduced and an accurate count of the number of wasps and seeds that matured within it). It was low-tech work, as I say, but tedious and very time-consuming. The graduate student on the project, Charlotte Anstett (see below), was single-handedly responsible for getting this experiment to work.
HS: Do you continue to work in this field site? When was the last time you visited it? In what ways has it changed since the time you worked there for this study? What has happened to the lab you worked in?
JB: I haven’t worked on figs at all since the mid-1990s. It’s not a system that is local for me – the nearest figs are in some rather dangerous areas of Mexico, and Miami is a few thousand miles away. I could only do this work with a local collaborator, since the trees flower year-round and can’t be studied (at least for the things I want to know) via short visits. Martine Hossaert, who I worked with (see below) and whose lab I relied upon, no longer is there. I visited Miami and my study site periodically until about the time this paper was published; I had a courtesy professorship from University of Miami, and served on several students’ dissertation committees. I understand that my fig trees are still flourishing.
HS: You acknowledge a number of people at the end of your paper. Could you tell us a little more about how you knew them and they helped in this study:
Martine Hossaert-McKey: Martine was my collaborator on the Florida fig research (Ficus aurea), and co-principal investigator on the grant that funding the empirical work reported in this paper. She is a plant reproductive biologist. It was a great collaboration that produced many papers, of which this one was about the last. Our collaboration likely would have continued, but towards the end of our study, Martine and her husband Doyle McKey left Florida for permanent positions in France. This is one major reason why my fig research started wrapping up in the late 1990’s, well before this paper was published.
Marie-Charlotte Anstett: Charlotte was a French graduate student, advised by Martine, who did that actual hard work to introduce fig wasps one by one into figs to get the data shown in this paper. I never believed that this experiment was even possible! Charlotte is fully responsible for its success.
Allen Herre; Finn Kjellberg; Olle Pellmyr: All of these guys gave me information, or allowed me to double-check information, displayed in the long table quantifying costs of different mutualisms. Allen was, and remains, the “god of fig biology”. He has been working on figs in Panama since before I started working with the system. Finn Kjellberg works on figs in France and paleotropics. He was my postdoctoral advisor during 1987. It was during my stay in France that I first got interested in working with theoreticians on mutualism, something that turned into a long-term, important element of my career. And Olle was the “god of yucca biology”. He knew everything about yuccas, and trained every single one of the current generation of researchers who study them. Olle died long before his time, on 5 December 2017. His legacy as a scientist, mentor and friend will not be forgotten.
Bill Morris: Bill was my host for my sabbatical at Duke University, which is where I was when I wrote the paper. I spent that sabbatical working on ecological models of population dynamics of mutualisms in the presence of cheaters (including three papers published in 2003 (1, 2, 3). The functions we use in our model of how the benefits of mutualism change with partner abundance are based in part on the relationships I discuss in this paper.
Nat Holland & Ted Fleming: Nat was my postdoctoral fellow during the period in which I was writing this paper. He was specifically studying how costs and benefits of mutualism change as population size changes. A related paper to this one is Holland, DeAngelis and Bronstein 2002, in American Naturalist. Nat and Ted Fleming (who was his doctoral advisor) worked extensively on another pollinating seed parasite system, the senita-senita moth mutualism, and their work provided the data I used in the table. Nat and Ted (a professor at the time at University of Miami, now retired in Tucson) provided comments on the manuscript.
Greg Dimijian: Interesting that I thanked Greg! He must have given me comments. He was a retired physician and fantastic photographer who took photos that I’ve regularly used in my talks. He took the wonderful photo that is on the cover of the mutualism book I published in 2015.
Goggy Davidowitz: Goggy’s my husband. He’s a professor of entomology at University of Arizona. He and I have done some work together on plant-insect interactions, though not on mutualisms. He always gives me great feedback on my talks and manuscripts.
HS: How long did it take you to write this paper and where and when did you write it? Was this part of a planned special issue on the symposium?
JB: I don’t remember much about this. I think it was pretty easy, because it was based on a symposium talk (where I could talk about whatever I wanted), it wasn’t peer reviewed, it wasn’t going to appear in a high-power place, and no one cared if I repurposed data collected for other reasons. I wrote it while I was on sabbatical at Duke University.
HS: What kind of attention did this paper receive when it was published?
JB: None whatsoever, as far as I remember! I don’t know, in fact, that it gets any attention now. I’ve always assumed that the other conceptual papers I published in 2001 (one on cheating in Ecology Letters, and one general overview of the evolutionary ecology of mutualism as a book chapter) were the ones that got the attention.
HS: What kind of impact did this paper have on your career and the future course of your research?
JB: Writing this paper offered me the opportunity to systematize my thinking about the costs of mutualism in a way that led me to pick the questions and empirical systems that I focused on subsequently. This paper brought home to me that most of our empirical knowledge of cost-benefit balances emerged from pollinating seed parasite (“nursery pollination”) mutualisms alone. This drew me to a different system, in which the pollinator (Manduca sexta) is a very large moth that lays eggs on its hosts’ leaves; the larvae consume the plant (Datura wrightii). It’s specialized, but not nearly as specialized as the pollinating seed parasite systems, so I think it has considerably more generality. On the theoretical side, the relationships between costs, benefits, and net effects that I formalized in this paper form the basis for most of the ecological and evolutionary models I’ve subsequently collaborated on.
More generally, I published three conceptually oriented papers in 2001: this one on costs, one on cheating, and one on the evolutionary ecology of mutualism. All of them, I think, put me on the radar in ecology as a big thinker about mutualism. I didn’t supplant anyone – this was a very new field – so maybe it’s more accurate to say that I was the first person recognized for arguing that big thinking about mutualism could be interesting and fun. Many of the most exciting people working on mutualism today are a younger generation who wanted to work with me or discuss their insights with me. I think their interest emerged from their exposure to these papers (as well as to three modeling papers I published in 2003) as undergrads and grad students. Working with them has been among the most fulfilling elements of my career.
HS: In many places in the paper you describe the state of mutualism research at that time and point to fruitful future avenues for research. Today, 16 years after the paper was published, could you reflect on the progress we have made in these aspects:
a. “My aim here is to encourage a rigorous comparative approach to studying the costs of mutualism…Equally importantly, however, I argue for the value of conducting considerably more detailed studies of the phenomena generating costs within individual pairwise interactions” – To what extent has this happened?
JB: Interestingly, I just gave a symposium talk in Germany that dealt with this exact issue. I did a literature search that shows we now have measurements of costs and benefits in a far wider range of mutualisms than we had 16 years ago. More generally, I went through my own reference database (I obsessively track advancements in this field) and found the following articles on different aspects of costs of mutualism:
Identifies a cost (54, 27%)
Measures a cost (89, 45%)
Ecological model of costs (11, 6%)
Empirically weighs costs and benefits (28, 14%)
Evolutionary model of costs (8, 4%)
Conceptual contribution (6, 3%)
What you can see here is that there are a lot of empirical studies (relatively speaking) that identify and/or measure one cost or another, and a fair number that weigh costs against benefits. But there is still very little incorporation of costs into ecological or evolutionary models of mutualism dynamics, and almost no conceptual work since this 2001 paper.
b. “The data [post-dispersal seed predation, dispersal to unsuitable microsites] simply do not exist yet to attempt such a demographic analysis [i.e. comparing different sources of seed mortality] for figs”. Do we have such data now and has such a demographic analysis been attempted?
JB: Not as a far as I know, but I should say that I don’t follow the fig literature to the extent that I could answer this. The center of fig research has shifted to India and China, and great work is coming out of there now; I just don’t know the people doing it the way I once did.
c. “A more detailed comparison should take this approach to examine whether (for instance) the costs of fig seed consumption are more similar among more closely related figs, due to shared ancestry rather than shared ecological pressures”. Has such a phylogenetic analysis been done?
JB: Again, not that I know of – but wouldn’t it be an interesting thing to do?! I doubt that fig researchers even know of this particular paper, so I doubt that they’ve come across this suggestion.
d. You highlight four methodological issues with interpreting any single measure of a cost of mutualism – getting an accurate measure, the issue with using an average measure, the problem of interpreting the significance of a cost, and the problem of using a unilateral measure. How much progress have we made in dealing with these issues?
JB: I forgot I said that stuff. I’ll have to think more about that way of framing the question – I like it! No, I don’t think any progress has been made on this at all. As the above data suggest, there is still very little conceptual work on mutualism costs, even while a great deal more data quantifying those costs has been conducted and published.Let me add that the newer work is technologically impressive in many cases, with a far more rigorous physiological basis than the work I summarized in 2001. As just one example, readers may want to look at Beth Pringle’s article in New Phytologist entitled “Integrating Plant Carbon Dynamics into Mutualism Ecology” (2016, 201:71-75).
I’ve been thinking of writing a new paper on this topic, and answering your questions is getting me more and more excited about doing so.
HS: In Table 1 of your paper, you summarise empirical estimates of the “cost of mutualism” in pollinator seed parasite mutualisms. If you were to make such a table today, what would be the significant changes?
JB: As I say, there’s a much greater diversity of systems now for which we have measures of costs. There’s been quite a lot of work on ant-plant defensive mutualisms in particular (I would direct interested readers especially to research coming out of Megan Frederickson’s lab), and a growing number of studies look at symbiotic mutualisms. When I wrote this paper, fig and yucca mutualisms were the standard model systems for studying mutualism. These days, the model system of choice is often a rhizobial or mycorrhizal mutualism. There are some very interesting studies of costs in those interactions.
HS: In concluding your paper you highlight three themes that will be useful to explore “via studies within individual pairwise interactions”:
- Magnitude of costs experienced by an individual mutualist in relation to its traits, partner’s traits and partner’s abundance.
- Whether costs of mutualism experienced by an organism are functionally linked to benefits received both by it and its partner
- Conflicts of interest arising from links between costs experienced by one partner and benefits gained by other partner
To what extent have these aspects been explored within individual pairwise interactions?
As a thought experiment, if you were to rewrite a paper titled “Costs of mutualism” today, what will its main takeaway be?
JB: Well, I have a new list of questions that I think are the unanswered ones about costs. I presented these at a symposium talk in November 2017, without having reread this 2001 paper, I confess! Some of what I offered is more or less the same questions, and some are new. The new ones are more evolutionary in focus, and reflect the amount of time I’ve spent working with evolutionary biologists in the past decade. All of them, I think, are very exciting for someone to pick up:
- Why are the costs of mutualism the magnitude they are? (This is basically the same question as “a” above.)
- When are the costs to one partner linked to the benefits received by the other, and what are the consequences (this is “b” and “c” above, and this question has driven some of my empirical work on the Datura-Manduca interaction mentioned previously.)
- Which are more variable in time and space: the costs or the benefits of mutualisms, and why? Should it matter, for mutualism’s evolutionary and ecological dynamics?
- Do the costs of mutualism evolve, and if so, how?
- When the net effects of mutualism increase over time, is it because the benefits increase, the costs decrease, or both?
- Are the costs of mutualism tolerated, or are they resisted?
- Does cost control arise during mutualism evolution, or is it a precondition for mutualism evolution?
The take-home message of that talk was as follows: “Research on why costs of mutualism are the magnitudes they are, as well as how selection has shaped them, is wide open.” I think that was the message of the 2001 paper too! The audience (all of them researchers on mutualism) seemed to agree that this was interesting, and no one challenged me that it wasn’t in fact a gap in the literature.
As another thought, there’s been a huge amount of work over the past sixteen years on ‘cheating’ in mutualism. Papers on cheating make up about one-third of all publication on mutualism evolution right now. At the time I wrote this paper, the concept of cheating and the concept of the cost of mutualism were pretty separate; I wrote papers on both these things in 2001, and they don’t make a lot of reference to each other. Now, I think, these topics are merging in many ways. Cheating from within the mutualism is one of mutualism’s major costs, and investments necessary to keep “outside” cheaters out often requires significant investments. If I were to rewrite this paper today (and, as I say, I might), I would stress this point.
HS: Have you ever read this paper after it was published? If yes, in what context?
JB: I haven’t read it word for word in quite a while (except in the context of this interview!), but I have used the figures from it very regularly over the years to illustrate that one can quantify the costs of mutualism and weigh them against the benefits. I also regularly use them to talk about context-dependency of the cost: benefit ratio.
HS: Would you count this paper as a favourite, among all the papers you have written?
JB: That’s an interesting question. There are some aspects of this paper, particularly the way I displayed the data in Figures 2 and 3, which I think I intentionally oversimplified nature in a way that has always left me uneasy. (Note that there are no error bars …) But the thinking I did about costs in the context of putting this paper together was really important in shaping my current ideas about mutualism dynamics. I’m glad I had the opportunity to write this paper.
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 written 16 years ago? Would you add any caveats?
JB: I’d say – we have many more good empirical measures of costs now, but not one of the big conceptual questions I posed in this paper has been answered yet! Go for it!! Mutualism is a very young field, and there’s a lot of low-hanging fruit for an enterprising biologist. Then I’d refer him or her to our 2015 Mutualism book (Oxford University Press) for even more research ideas.