In 1990, Michael Ryan, James Fox, Walter Wilczynski and Stanley Rand published a paper in Nature showing that, during courtship, male Tungara frogs exploit a pre-existing neural bias in females towards lower-pitched calls. This experiment provided evidence in support of a third hypothesis for the evolution of female choice (in addition to runaway sexual selection and natural selection): ‘sensory exploitation’. Twenty-six years after the paper was published, I interviewed Michael Ryan about the making of this paper and what we have learnt since then about the sensory exploitation hypothesis.
Hari Sridhar: Before this study, you had already done a lot of work on Tungara frogs and sexual selection. You had even written a book about it. Can you tell us what the specific motivation was to do the work presented in this paper, on the neural basis for female preference?
Michael Ryan: Well, when I started to work with frogs for a Master’s thesis, before I went to Cornell to do my PhD, the dogma was – most of which was true – that frog calls evolved to indicate the species – the species identity of the caller. And that females were under strong selection to mate with males of the same species. That’s certainly true. And then when I went to Cornell and started working with bullfrogs – I was not working with calls then; I was studying territoriality – I was struck by how variable their calls were. I mean a bull frog call sounds like a bull frog call, but I could clearly tell males apart. I would hear a male and know that he was on a neighbouring territory the night before. That he’d moved over. That’s how I became interested with the frog call, as it had to do with mate selection within a species. The thought then would have been that the variation is just noise; that it’s not very meaningful. Then I came across Trivers’s papers on sexual selection, read Darwin and then got very excited about studying the calls as a sexually-selected character. But when I went to Cornell, I was down the hall from this faculty member, Robert Capranica. Capranica had done ground-breaking work on how frogs extract biologically meaningful information from the call. He was a very serious neurobiologist who had received his PhD at MIT, then was at Bell labs for a number of years, before he went to Cornell. Capranica and a whole series of students had shown how the frog’s brain was organised to cause the female to find calls of its own species the most attractive. I thought that if this brain is wired to detect calls, and if there is a lot of variation in calls, then some calls must sound more attractive than other calls. Then the same year I was hired in the zoology department in Texas, just by chance, they hired a post-doc from Capranica’s lab – Walt Wilczynski – in the psychology department. We knew each other from before, though we weren’t close friends. But we both felt that looking at this question – variation in the neural basis of sexually-selected calls – would be very much worth doing. But before that, in my Tungara frog book, I actually sketched out a little model for how neural biases in the Tungara frog might cause females to prefer lower frequency calls. What I suggested then was wrong, but that’s when I started thinking that the variation in the auditory system could explain variation in female preferences for male calls.
HS: Did you test your model in this paper?
MR: We didn’t test that model specifically. I tested that model behaviourally and showed that it was wrong. I had worked with Walt Wilczynski on another frog – a cricket frog – looking at variation among populations. That’s where Walt and I started this neuro-ethological approach. I am just going to quickly glance at my publications to see if we started publishing on cricket frogs before Tungara frogs. I am sorry, this will just take a few seconds. We were doing this research at the same time. Ah, here it is. In 1988 we had a paper in Science: “Coevolution of sender and receiver: Effect on local mate preference in cricket frogs”. That study did not demonstrate that the brain was generating sexual selection within populations, but instead it was influencing how calls and preferences evolve amongst different populations. Then with the Tungara frog we asked that question at a different level. Instead of variation among populations we were asking about variation within a population. But the neural work on the Tungara frogs came after Walt and I started working on the brain and behaviour of the cricket frog. We had already been interested in the Tungara frog, so that was a very natural progression.
HS: Going back one step, can you tell us how you first got interested in the Tungara frog?
MR: I searched the literature for some species to work on for my PhD thesis. I wanted to look at how variation in calls influenced mate preference within the species, and I decided to work on red-eyed tree frogs in Panama. But they were very high up in the canopy and I was having a difficult time recording the males and a difficult time watching the matings. And when I was trying to record these males, all these Tungara frogs would be calling at my feet. They were very common and I would always be kicking the frogs to shut them out so I could hear the red eyed tree frogs calling. That’s when I thought that since these Tungara frogs are always calling, maybe I should study them instead. That’s how I started to study them.
HS: You spoke about Walt Wilczynski’s role in this paper. Can you tell me about the roles of the other two authors – James Fox and Stanley Rand?
MR: Stanley Rand is the most important person in this story. Stan had studied the Tungara frog in the 1960s, although he had not published anything on them. He was very well-known for his work on reptiles in Panama, but he had also done a lot of work on frog calls in Brazil. So when I went to Panama, to the Smithsonian institution, he was my mentor. However, during my PhD, I worked fairly independently of Stan. Because I was on Barro Colorado island, I would only see him for a short time, maybe once a week, when I went into the city. But after I received my PhD, he and I started this very serious and very long-term collaboration, working not only in Panama but all over south America. Stan had done the initial work in the Tungara frogs, describing the call variation. But like I said earlier, he had not worked on these frogs for 10 or 15 years before I started to. Stan then became a very very close collaborator and a very dear friend for many years. So the importance of Stan in all of this cannot be overemphasized. And Walt, who I also wrote a small Animal Behaviour textbook with a few years ago, he and I are also very close friends. As were Walt and Stan. So the three of us collaborated a lot and were also personally close, besides being professionally close.Jim Fox was a graduate student of Walt’s lab. He was doing auditory neurophysiology for his PhD thesis and was also doing some computer modelling. Walt had recorded the tuning curves for the auditory system of the Tungara frog. So when I went to Walt and said that we needed to do the simulations that we did in this paper, Walt said Jim would be a very good person to do that part of the work.
HS: Did you bring the frogs back to Texas for the lab work?
MY: Yes we did. We brought them back from Panama. In that paper we have just the Tungara frog, but I think we published another paper in which we had nine different species of Physalaemus. Maybe there were eight, I forget, but it included all of the closest relatives of the Tungara frog and some outlier species. For those, Stan and I went all over South America – Ecuador, Peru, Venezuela and maybe Colombia – and brought those frogs back to the lab where Walt did the neurophysiology studies.
HS: In the Acknowledgements, you thank the governments of Panama and Ecuador. Did you samples the other species you use in this study – coloradorum – in Ecuador?
MR: Yes. I spent, probably, 4-5 months in Ecuador, working with Physalaemus coloradorum, as well as two other species of Physalaemus. There’s one in the Amazon – Physalaemus petersi – which we have done a lot of work with, and then another Physalaemus species from the Pacific coast – Physalaemus pustulatus. Then we added to that Physalaemus enesefae from Venezuela and then Physalaemus – I have forgotten what we called it, kaikai I think – it’s still an undescribed species. Then we had another species – Physalaemus roraima – that was from Brazil. We did many long and arduous field trips to record those animals in the field and bring them back to Texas in good shape and, of course, alive.
HS: Why did you pick P. coloradorum for the comparison in this paper? Is it the closest relative?
MR: No, it’s not the closest relative. The closest relative is something called Physalaemus petersi, which we did study, but I didn’t want to compare with petersi because there had been reports that some populations had complex calls and other populations did not have complex calls. So we didn’t know then if the complex call evolved independently in Tungara and petersi or if it was the common ancestor of those two. Those two are sister species. So then we decided to go kind of over to the next little group. And that’s where we started the work with coloradorum.
HS: You say you recorded calls from 54 males. Did you make all the recordings yourself?
MR: Yes, all of the acoustic work I did myself. Included in that is a transect that Stan and I did one time. We drove from Texas, through all of Central America, to Panama, and recorded Tungara calls from maybe 25 to 30 populations. And I did all the recording and all the acoustic analysis.
HS: Was the equipment you used for recording then very different from what you use now?
MR: Oh yes. It has changed a lot. In the very beginning I was using a Nagra tape recorder – big heavy reel-to-reel tape recorder – and occasionally another make called Uher. From that we moved to high quality cassette recorders and then onto digital recorders.
HS: Did you come up with the terms “whine” and chuck” to describe the Tungara frog’s calls?
MR: No, ‘whine’ and ‘chuck’ are terms that Stan had already come up with. Again, the work wasn’t published. What’s interesting is that when I went to Panama these were not called Tungara frogs. They were called mud-puddle frogs. And then I asked a Panamanian if there is a Panamanian name for the frogs. He said ‘Tungara’. The word is onomatopoetic. Tungara sounds like a whine and two chucks – Tooon ga ra. Later, after I started to use that name, I found a field guide to the frogs of Nicaragua where the author independently was also calling this frog Tungara.
HS: Did you do most of the writing for this paper?
MR: Yes. In all of the frog work I did with Stan, in all of those papers – I don’t know if there are 50 or 80 or 100 – I did all the writing. I would say up until maybe 4 years ago, I did all the writing of papers myself. In the last 10 years, we have had a lot of grad. students working on the Tungara frogs and now they do a lot of the writing.
HS: In the abstract, you say your results allow you “to reject two popular hypothesis for the evolution of this female preference (runaway sexual selection and natural selection) in favour of a third: sexual selection for sensory exploitation.” Were your results one of the first evidences in support of this mechanism?
MR: Well yes, because the term sensory exploitation had never been used before. There are actually two papers – this one and a larger, more conceptual, paper in Oxford Surveys of Biology in 1990 – where I laid out this idea of sensory exploitation. Now, like most ideas, it wasn’t like brand new. It was very much related to Tinbergen’s idea of supernormal stimulus, to some of Mary Jane West-Eberhard’s ideas about – I am blanking on the term she used – sensory traps. Then, when I sent some of this work to John Endler, he wrote back and said this was exactly their idea of sensory biases. So Endler came up with sensory biases about the same time I came up with this idea of sensory exploitation.
HS: You say “In all studies to date in which female mate preferences based on call frequency are expressed, tuning of the peripheral auditory system predicts the preference by indicating frequencies that would most strongly stimulate the auditory system. There are no reported examples of any mechanisms overriding this peripheral bias”. Since then, have there been such examples?
MR: I don’t think so. There was a study by Gerhardt and Schwartz where they reviewed a lot of the data and showed that the dominant frequency of the male calls and the tuning of the female’s auditory system were fairly well correlated for a number of species. I think the R2 might have been 0.8 or 0.85. Now, I probably wouldn’t state that exactly the same way now, because there are frogs in which the females are much more attentive to temporal variation, for example, the pulse rate. I would have to check; I don’t know if people have played pulse rate against preferred dominant frequency and seen what happens.
HS: You say that “As the chuck falls below optimal frequency, neural stimulation decreases only slightly, but there is a greater decrement in stimulation when the frequency of the chuck is increased above the optimal chuck”. Do you get the same asymmetry in preference when you play a range of frequencies around the optimum to the female?
MR: Do we get the same asymmetry..hmmm… The short answer is I don’t know. We have done the experiment, but I am not sure we have published the data in that context. We have a big paper in Animal Behaviour – Wilczynski, Rand and Ryan – on preferences of spectral features, and I don’t know if the asymmetry pops up in those preferences or not.
HS: Can you give us a sense of how long the whole project took. Of course, the fieldwork on the frogs had been going on for a long time, but can you tell us when all the lab work and writing was done for this specific paper? Would you remember how long it took from idea to execution and publication?
MR: Yes, I have a pretty good idea, because I had stopped working with Tungara frogs for a couple of years and then I came to Texas for my faculty position. So I think it was about 1986 that Stan and I started to work together, and it was for this idea. We weren’t using the word sensory exploitation at the start. The term only came in 1990 when we used it in the titles of three papers – this one, the Oxford series and another paper in Evolution. So all that work started in 1986, and it probably took Walt a good year to do the recordings. I would say about four years from the beginning, with the idea already taking form in 1986.
HS: How long did just the writing take?
MR: I write very quickly. I’m guessing it took a couple of months. Just using this paper as an example – Stan and Walt and I would constantly talk about what was going to be in the paper. So by the time I started writing – and this is how I write always – the outline was already there. It was merely a matter of putting it on paper. I know other people write very differently. They think as they write, but I usually do most of the thinking before.
But just to put one thing in context. I am not sure exactly when we started to use the internet, but in ‘86 when Stan and I started this work, we didn’t have any internet. Stan was in Panama, so we would have to mail manuscripts back and forth. That really slowed things down.
HS: Would Stan, Walt and you meet often?
MR: Walt is not here now, but he was in the psychology department here in Austin then. He was also my closest friend here, so we would see each other a lot. And Stan would come to Austin twice a year for visits. And, of course, I was in Panama every summer, and Walt would come to Panama occasionally. So we would meet a lot and constantly bounce ideas off one another.
HS: Did all four authors ever meet as a group?
MR: Oh yes. I forget exactly the time of the year, but I remember Stan was up for a visit and he met Jim Fox and Jim showed him the results of the simulation. Once Fox was done analysing those data, then the paper was fairly well outlined, and Stan met up with all three of us.
HS: Do you remember how the figures in the paper were drawn?
MR: Those were drawn with a ruler, a Rapidograph and Indian ink. And they might have been labelled with press type. It’s possible that an artist might have helped with some of this, but if so, she must have done it the same way, you know, by hand.
HS: In the Acknowledgements you thank a few people for comments on the manuscript. Can you tell us how you knew these people?
MR: James Bull is an evolutionary biologist who works on virus evolution. He’s down the hall from me. Mark Kirkpatrick is pretty famous for his theoretical models of sexual selection. Kirkpatrick and I were post-docs. together at Berkeley. There, we started talking about sexual selection, and he taught me a lot about evolutionary genetics and runaway sexual selection. He was also in my department. Susan Weller was a good friend and an entomologist, who was very interested in phylogenetics and the comparative method. So she was a lot of help reading this paper. And Cocroft was a graduate student of mine at the time.
HS: Did this paper have an easy ride through peer-review?
MR: It did. I was in England after submitting it to Nature and somebody – I’d rather not say who – came up to me and told me he or she had reviewed it and really liked it. So yes, we didn’t have to argue or anything about it.
HS: It was submitted in the end of May and accepted by the end of October. So it took just five months.
MR: I guess I was impatient in those days, but it did seem long then! But a lot of people disliked that paper when it came out.
HS: Why was that?
MR: Well, it was different and it was new. I first talked about this paper at a behavioural ecology congress in Uppsala in 1990. I was asked to give a plenary talk – this is before I had published anything on this. So I gave the talk and Maynard Smith gave a talk later. That talk of Maynard Smith became famous in the field because in it he acknowledged that he had made a mistake criticizing Zahavi’s Handicap principle. Zahavi and I are close friends, so I was sitting next to Zahavi when Maynard Smith was giving that talk. But then, at the end of his talk, after acknowledging his mistake, he said – ‘And then Mike Ryan gave this talk and..’ and he pauses and I’m getting very nervous and he says – ‘And.. ah, I’m just not sure what to think about that’ Now I knew Maynard Smith, because he had come to Panama for a long time and we had become close friends. So I talked to him later on and he said he was very intrigued by my idea. He told me about work by Magnus Enquist on neural networks and said that I should contact him because he had the same ideas that I did. Then there was this group at Cornell, which is where I got my PhD thesis, who were very upset by this paper. These included Paul Sherman, Kern Reeve, and I know Steve Emlen told me that no one at Cornell saw any validity in the idea. But Sherman especially was very very critical in print, and we went back and forth for a while. I think the problem was that people like Sherman – who I like a lot and have immense respect for – came out of those sociobiology wars. He was a graduate student of Richard Alexander and I think he was kind of an extreme adaptationist who viewed this sensory exploitation idea as anti-adaptation. But it’s not really anti-adaptation; it only has a very modular view of adaptation. I’m just finishing to write a book now called “The Nature and Neuroscience of Sexual Beauty”, in which the basic idea is this: the brain determines what is sexually appealing to it. These sexual preferences in the brain, e.g. the way we perceive beauty, some of that certainly evolves for adaptive mate choice, but the brain does many different things and sex is only one of them. I think Paul didn’t like that idea at all. He also didn’t think that historical contingency played much of a role and that the strength of natural selection would overpower any phylogenetic inertia. So I think Paul saw people like myself who were pushing this idea as falling more into the camp of, let’s say, Stephen Jay Gould than of E.O. Wilson.
HS: Did the paper attract attention in the popular press when it was published?
MR: It did, but not all that much. You know papers in Nature and Science always get some attention, but I don’t think it got attention in the press for the right reasons. The New York Times did a profile of me, when I was 50 – which would have been in 2003 – and that’s when, for the first time, I got the feeling that at least some people in the media were understanding what people like myself and John Endler and Mary Jane West-Eberhard were talking about. For most people in the media, I think our message was too subtle.
HS: Its now 27 years since this paper was published. Given all the work you have done since then, do you think the main findings of this paper still hold true?
MR: Yes, it does. One of my colleagues here – Molly Cummings – and I wrote a review for Annual Review of Ecology, Evolution and systematics, maybe in 2013. We reviewed over a couple of hundred studies that were, you know, supportive of these basic ideas of sensory bias and sensory exploitation and sensory traps – they all kind of blend into one another. But there’s a couple of 100 studies that we saw as support for these ideas, I should say.
HS: If you were to repeat this experiment today would you change anything, given the advances in technology, theory and analytical techniques?
MR: Well, probably. I mean, now, most of the brain work is done using Immediate Early Genes (IEG). There is no doubt these genes have told us a lot about organisation in the central Nervous System, but these gene expression studies would not give us the information we got from putting an electrode in a cell and testing its characteristics. We are now interested in multimodal communication – how visual signals combine with acoustics. So, as we speak, we are doing IEG studies, but I also have a grad. student working in a medical school with one of my old postdocs because we feel that for some of these questions we are asking we still need to use electrodes and characterise cells. The gene expression studies can’t do that. So I think if we redid that study today, we would add gene expression studies to the neurobiology, but I don’t think we would have replaced it.
And in terms of the phonotaxis experiments – we do them in big fancy walk-in chambers, and videotape all the responses, so they can be double-checked. For the female to go from the middle of the arena up to the speaker – we scored it as a choice when she is actually 10 cm in front of the speaker – if we were to scale that response to human body size, it would be the equivalent of travelling 80 metres. So this was a very robust response even then.
We have adapted to new technologies to do the analysis of the calls, the recording of the calls and the way we do the experiments, but the basic experiment is the same – we ask the females to go to the calls and tell us which one she likes better.
Can I just tell you another small incident about not everyone liking this paper?
HS: Sure, go ahead.
MR: I went to a workshop at Davis once, which had a lot of people working on communication. Peter Marler, who wrote a forward for my Tungara frog book and was a great hero of mine, had organised this. They had discussion groups you could join, and I was surprised to find one discussion group on sensory exploitation. I went there, and in that group of people was W.D. Hamilton. I had met him a few times before, said hello, but we weren’t friends. In that discussion group, Hamilton said – I don’t really see anything – I forget his word; enlightening or important or worthwhile, something like that – in this idea. Then a couple of other people in the group said: ‘Ya, I agree with Bill’. ‘I agree with Bill’. Well, in those days, who would not agree with Bill Hamilton! I mean he was a genius.
HS: Which year was this?
MR: It was early 90s. But I have to make one thing clear: he wasn’t saying it at all in an aggressive way or a mean way. He was just stating what he thought. I met Hamilton another time and we talked about other things and he gave me some very good suggestions about some fish studies I was doing. But for someone like me who, you know, had recently published this idea, to have one of the greatest minds in evolutionary biology tell you that he didn’t think there was anything there – that was a bit deflating.
HS: How much of an influence did this particular piece of work have on the future course of your research? Did it open up new lines of research?
MR: Oh yes. I had a paper in Science in 1980, and I’m pretty sure that that was the first experimental demonstration of sexual selection by female mate choice. It was a year before those beautiful studies by Malte Andersson on the widowbirds. But what this idea said to us is, you know, if sensory exploitation is right, we should be able to synthesise any kinds of calls that we want. This is one of the beauties of working with acoustic displays. So we decided we’ll just see how we can exploit the female’s sensory system. Can we add things to a whine that are not chucks but are as attractive as chucks? After Stan passed away a few years ago, my student and I, along with Stan as posthumous author, brought together a whole lot of this research and showed all these different kinds of sounds that would have made males much more attractive if they possessed them. Our paper came out a little bit after a paper by James Gould at Princeton, where he did pretty much the same thing with fish. He presented females with models of male fish with all these constructed variation in visual display. So yes, for us it opened up a very different approach, through which we have been exploring the female’s – now, I wouldn’t say auditory system – whole perceptual system. Because we are including the brain now. The work we have been doing since 2010 has been asking very similar questions, but at the level of cognition. So we had a paper in Science in 2010 on Weber’s law and how that influences preference. I think since 2010 – if a paper we have submitted now gets accepted and it is very likely it will – we would have had five papers in Science, all of them based on exploring these biases, not just in the sensory system, but in the whole perceptual and cognitive systems of the females.
HS: I know that Stan Rand passed away a few years ago. Do you continue to work with the other two authors of this paper?
MR: Not with Jim Fox. I am not quite sure what became of Jim. But I do continue to work with Walt. Like I said, we even published a small Animal Behaviour text book together a few years ago. He left University of Texas maybe 8 years ago but he still has his house here. Just last week we had dinner together and we were talking about new things we might start to work on. He still serves on th committees of my students who have neurobiology components in their work. I am not a real neurobiologist. Walt is now at Georgia State University. He went there to become the director of a brain and behaviour research institute.
HS: Have you gone back to these field sites after this study? Do you continue to work there?
MR: Yes, I continue to work in Panama, although I don’t work on the island anymore. We work on the mainland right across from the island. So yes, we’ve been working there for many years. Just to give you an example – the most basic female preference we have studied is this preference for “whine-chuck over “whine” I think we have about 20 years of data on the strength of the preference for “whine-chuck” over just “whine”. We almost always have sample sizes over 100 so we have very good measures of the strength of that preference. Because of my accident, this is the first summer in many years that I have not been in Panama. Since we have stopped for a year, maybe I will start putting all that work together. We have lots of long-term data of this kind.
HS: If you don’t mind my asking, why have you stopped working on the island?
MR: Well for few reasons. We work in this small town called Gamboa. When we started working there, there weren’t many scientists, but since then the Smithsonian Tropical Research Institute has built a big lab there. The very first lab building they built had our frog lab in there. And I still have it – the Smithsonian lets me keep it after Stan passed away. And there are many more Tungara frogs there than on the island. Also, as the phonotaxis experiments started to become more sophisticated, we needed to be closer to electronic shops and computer shops. From Gamboa it is easier to get to Panama City.
HS: Have you ever read the paper after it was published?
MR: I generally don’t go back and reread my papers. I look at them, of course, when I reference them, to make sure I’m remembering correctly what is in the paper. But I did read that paper and the one in Oxford Surveys a few years ago and I was pleasantly surprised that I liked them so much.
HS: Do you think the way you write has changed since this paper?
MR: Not too much. The last two years I have been working on a popular book that I already mentioned, and that’s having a little bit of an influence on my writing now – I try to make things a little bit broader maybe. But, in general, I don’t think it has changed too much.
HS: Do you know what this paper mostly gets cited for?
MR: I certainly did, when we wrote the review I told you about, a couple of years ago. It’s usually cited by studies that are also finding evidence for sensory exploitation, but it’s also been cited by studies that are more broadly looking at the neurobiological basis of behaviour.
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?
MR: What they should take away from the paper is this: in sexual selection by mate choice, if you want to understand what the chooser – of course, in this case its females choosing males, but it can be the opposite – is doing, it is, of course, important to understand the fitness consequences. But to understand the true details of it, we really need to understand how the brain is processing these signals. And every stimulus processed by the brain is processed in a biased way, because all sensory filters, whether they are ears or noses or eyes or touch sensors, have tuning curves that are biased. They don’t perceive all stimuli equally. For example, we don’t see ultra-violet or infra-red. And within the range of colours that we do see, we are more biased to the ones to which our cones are tuned. So if we want to understand biases in behaviour, we just have to look at biases in the stimuli that trigger those behaviours. This is similar to something my friend Mark Kirkpatrick once told me, when people started to study ‘good genes’ in sexual selection. He said: ‘If people want to understand ‘good genes’, at some point they are actually going to have to study genes’. And I feel the same way. If we want to understand biases in behaviour, we need to understand the neural biases that lead to the behavioural biases.
HS: Is this one of your favourite papers, among those you have written?
MR: It is certainly among them. You know, I very much like the first paper I published in Science on the Tungara frog. Even though, in that paper, I do suggest there could be a ‘good genes’ explanation to those preferences, but now, it’s not that I don’t believe that’s true, but we certainly don’t have any evidence that its true. But this paper is also certainly among my favourites.