In a paper published in Evolution in 1982, Mark Kirkpatrick showed, theoretically, that strong female mating preference for a male trait that reduces viability is neither selected for or against, but the mating advantage it provides to the males it prefers can lead to maintenance of that male trait in the population. Thirty-four years after the paper was published, I spoke to Mark Kirkpatrick about how he got interested in this topic, his memories of carrying out this work, and what we have learnt since about this topic.
Citation: Kirkpatrick, M. (1982). Sexual selection and the evolution of female choice. Evolution, 1-12.
Date of interview: 30th November 2016 (via Skype)
Hari Sridhar: What was the motivation to do this work for your PhD?
Mark Kirkpatrick: A fellow graduate student and friend, Greg Butcher, came walking into my office during my first year of graduate school. He said, I’ve been reading these pages about sexual selection in Fisher’s famous book, The Genetical Theory of Natural Selection, and I don’t understand what he’s talking about; can you help me figure this out? And he said, wouldn’t it be useful to try to really work out a proper model. I had no previous background or interest in sexual selection, but it was a very good question, and I went off and did some reading. One person had done modeling, Peter O’ Donald in Cambridge. I started messing around with his equations, and discovered some things that he hadn’t noticed. I then thought, well, maybe there’s more to do here.
HS: When you joined graduate school did you already have an idea of what you wanted to do?
MK: No. As a matter of fact, I wasn’t certain whether I was going to go into evolution or into community ecology. After I started working on this project, I was still trying to make up my mind. I had an opportunity to do research on fishes in the Amazon. I was in Brazil for my second field season and it was a total catastrophe because the government prevented our research team from going out into the field. So I spent about three or four weeks in Manaus writing out the equations that finally became that paper.
HS: Who was your PhD supervisor?
HS: This was one paper in your PhD. What else went into it?
MK: Ah, well, that’s a bit of a joke. This was the first chapter in my PhD. And chapters two and three had nothing to do with chapter one or with each other. My committee had the enlightened idea that if you had three good publications, it didn’t matter that they didn’t go together, so they approved a thesis that had just a single chapter. The Graduate School didn’t like this at all, and we got into a bit of a fight about that. But my committee went to bat for me and I was able to submit the other two publications as appendices.
HS: What were chapters two and three about?
MK: One was on punctuated equilibrium – how you can get a very rapid change in a trait. There was a big controversy at the time as to whether some sort of “genetic revolution” was necessary. I did a model that showed that slight shifts in ecological conditions can result in a rapid evolutionary change. The third chapter was a model of a population’s demography when birth and death rates are based on body size rather than on age.
HS: Stepping back a bit, were you always mathematically inclined?
MK: As an undergraduate I was trained as a biologist. I became interested in population genetics in my last year of undergraduate work. I was hanging around the lab of Dick Lewontin, and there were a lot of very interesting, bright people doing experimental work and theoretical work. That was a wonderful place to see you could actually do both and that got me interested in the possibility of doing modeling. I didn’t have any experience with it. In graduate school, I tried a few other things that didn’t work out so well. And so I did what, by default, out turned out okay.
HS: If you don’t mind my asking, how come your PhD supervisor(s) weren’t authors on this paper? Was that common then?
MK: It was more common then than it is now, but I think it also varies between disciplines and between advisors, and also between countries. I do remember asking Monty Slatkin if he would be coauthor. He gave me a tremendous amount of help, obviously, on my first theoretical paper. But he said, no, this was your idea, and you did most of the work, and you should take the credit. He has had a very enlightened, very generous attitude for his entire career with all of his students. I know that in some fields, and in some countries, it’s actually a sign of disrespect of the advisor if he or she does not want to be associated with the paper. It’s considered an insult to the student – the advisor is effectively saying that I’m washing my hands of this work. That was certainly not my case. Monty’s attitude was, if you come up with the idea and you do most of the work, then it’s yours.
HS: Who are the other people you were talking to about these ideas at the time? Were there papers you read that were critical in shaping your thinking about this topic?
MK: This is the strangest thing. I’d like to give you a deep answer that sounds like I was reflecting on bigger things. But, in fact, I just sort of followed my nose on this one question and this one set of models. The field of sexual selection has sort of fallen into disrepute, or at least it had become a backwater. My paper just happened, by good luck, to appear at a time when a lot of other people, for independent reasons, were becoming very interested in the field. Sexual selection as a topic in the mid-80s really took off, and a whole bunch of people for different reasons became interested in it. The timing just happened to work out for me and for the paper.
HS: Before Greg Butcher were you interested or even aware of this literature at all?
MK: I hadn’t thought about sexual selection. I had not thought about behavior. No, this was a whole new field for me.
HS: So, you looked up Lande’s model and Peter O’ Donald’s work only subsequently?
MK: Well, I started working on this quite before Russell’s models were published. As a matter of fact, I think I might have submitted my paper before he did. But my paper had a bit of a rough time getting published.
HS: Lande’s paper is 80-81.
MK: Right. I submitted my paper in 1980. And well, this is sort of a boring story…
HS: We’ll come to the story of its publication. So, you are saying those papers weren’t even published at the time.
MK: Yeah, that’s right. As a matter of fact, I knew Russ from when he was a graduate student. I was an undergraduate hanging out in the same lab as his PhD advisor. And so, even though I knew him, I didn’t realize that he was working on sexual selection until we’d both, basically, developed our models. We saw each other at a scientific meeting and discovered that we were both working on the same thing.
HS: Apart from your supervisor and Joe Felsenstein, were there other people you were discussing these ideas with? Can we go over names in the acknowledgments, to get a sense of how you know these people and how they helped?
MK: Well, I don’t think I’ve looked at those names since 1981 or so, but sure.
HS: First you thank M. Slatkin and J. Felsenstein. And then, S. Arnold.
MK: Steve Arnold.
HS: L. Heisler
MK: Lorraine Heisler.
HS: And G. Butcher
MK: Yes. Greg butcher. He’s the guy who should get a huge amount of credit for suggesting the whole project in the first place.
HS: What about the others?
MK: Steve Arnold was with Russ Lande at the University of Chicago at that time. Steve was actually the person who got Russ interested in sexual selection. I visited Chicago at the time when I was finishing up the analysis and starting to write my paper, when I discovered that those guys were working on sexual selection. I had a lot of discussions with them that were really, really useful and valuable.
I’ve known Lorraine Heisler for a long time and I can’t quite reconstruct the social and intellectual interaction at that time. But she was also doing work on sexual selection, as a behaviorist on the empirical side. So, she was one of the people I was talking to about what does sexual selection mean. Just to emphasize there was a very small community of people who were interested in sexual selection then. And I was a second year graduate student, so it wasn’t like I had access to a greater community of people.
HS: The empirical work on sexual selection came a little after this, e.g. Malte Andersson’s widowbird study…
MK: That’s a perfect example. Malte Andersson and several other people were starting work on sexual selection for independent reasons — I don’t know what was in the air or in the water. It all just sort of came together in the early 80s. Jack Bradbury would be another person who was in that mix. He was working on sage grouse and getting interested in all these questions. You know, reading the few pages in Ernst Mayr and the few pages in Fisher and trying to figure out what Wallace and Darwin had in mind. It was really like trying …
HS: Do you have a sense of why there was this sudden interest in the topic in the early 80s?
MK: I don’t. I am not a historian of science. I’m sure there are people who’ve thought more about that than I have. But, for me, it just strikes me as a remarkably wonderful coincidence that a lot of this stuff came together at the same time.
HS: How difficult was it to do this work? Can you give us a sense of how long you took and where and when you did most of this work?
MK: Sitting in a house in Manaus, Brazil. I don’t remember how long it took. Basically, 98% of the math in that paper is high school algebra; just a lot of it. And with Mathematica or other computer software, you could do it all now in about five minutes. I’m quite sure, in my office, I’ve got the binder with all of my original calculations, and it’s probably about three inches thick, because I was really not very good at it at the time. There was a lot of trial and error. So, I would guess probably six months was when most of it was done, but I’m not sure.
HS: So, part of it in Brazil and part of it in Seattle?
HS: Was all the work done in 1980?
MK: I started graduate school in fall of 1978. I believe I had that conversation with Greg sometime that winter or following spring. And then, I was in Brazil the following summer. So I think that I was doing the math in ‘79, and then wrote up a very short paper, which I submitted to Science and was rejected. I think that was early 1980.
HS: Now you can tell us the ‘boring story’ about its publication.
MK: The paper submitted to Science was probably inscrutable because it had to be extremely short. I can’t remember what it said, but it could not quite have explained anything clearly. I then submitted a much revised version to Evolution, which was at the time the premier journal in the field. It went out for review, which took an excruciatingly long period of time. I think it was in review for more than nine months. Finally, Monty said it would be appropriate to ask what’s going on. Doug Futuyma was the editor at that time, and he said, you’re absolutely right, this is ridiculous. And he took control of the paper and shepherded it through. I really have a huge debt to Doug for getting closure on the whole project, ultimately. I think it was well over a year from submission to publication.
HS: Did the paper change a lot from the first submitted version?
MK: I don’t remember, I can’t tell you that.
HS: Do you remember if it attracted a lot of attention when it was published?
MK: It drew a lot of attention as the momentum of the topic developed. It wasn’t that it initially generated a lot of controversy – maybe it did later – but when it first came out there was so little of a worldview or entrenched view about the subject. I think that there was a lot of interest in sexual selection as a developing field, but it wasn’t considered controversial at the time.
HS: And later on?
MK: Yeah. I guess it was controversial to the extent that behavioral ecologists couldn’t quite wrap their head around the idea that a female mating preference might not be adaptive. And that is still the case. There’s, sort of a religious worldview among a certain slice of behavioral ecologists that everything is adaptive, including female mating preferences. One of the messages of this paper is that female mating preferences can cause a male trait to evolve to ridiculous extremes without the preference itself being an adaptation. A lot of people find that a very difficult concept. That remains a big controversy in the field.
HS: Did you anticipate at all the kind of impact the paper has had?
MK: No! When you’re beginning graduate student, you have no idea. It’s not like you have five choices of research projects and you’re trying to make some strategic decision. You have, what you think, is a really fun idea and you go after it, for its own sake. So, clearly no anticipation of what would happen.
HS: So, are you surprised by its impact?
MK: Yes. I would say surprise is an understatement.
HS: Do you have a sense of what it gets cited for?
MK: That’s an interesting question. No, I don’t. I think that probably the majority of citations are not because of anything specific it says, but because if you’re writing a paper about sexual selection you should probably have a couple of theoretical references in there. And this is one of the older papers that’s in that field. So, it’s sort of like, you know, a historical touchstone. And then, there are the people who would like to really try to figure out whether good genes or runaway process or a neutral process is responsible for female mating preferences, and this is maybe the simplest way to understand how and why that plays out as it does.
HS: Would you say that this had a huge influence on your career? I mean, did it figure in, for example, postdoc interviews?
MK: Sure, definitely. I think what we all work on is a combination of what most interests us and what we can do, but then also what we do that gets positive feedback. And this did attract a lot of attention. And so, for the first third of my career, probably half of my publications were on sexual selection. And if this had gone off into the ether and never been cited, I’m sure I never would have thought twice about the subject again.
HS: You said in the first third. Do you work less on the topic now?
MK: Yeah, well, I guess it depends on what you consider to be sexual selection. The biggest project that we have in my lab right now is on the evolution of sex chromosomes. And one of the things that we think is very important there is what’s called sexually antagonistic selection. Sexual selection is one of the things that can cause sexually antagonistic selection which could cause sex chromosomes to evolve. So that’s several steps away, but, well, the word sex is still in there. I don’t think there’s a linear path from the 1982 paper to what we’re doing now. But my attention has always been attracted to that set of themes. So, it surely has had an influence up until now.
HS: Did you continue working on the other topics that were in your PhD – punctuated equilibrium and the demographic model?
MK: The punctuated equilibrium paper was using a type of model that Russell Lande popularized – his quantitative genetic model – that had a huge impact on population genetics in the 1980s. And so, that was the first paper that I did using that modeling technology, and I worked on other topics using that approach. So, that was also very important in giving me research trajectories in other areas as well.
HS: What about the demographic model?
MK: That was a one-off. I mean, it was fun and interesting at the time but didn’t quite have the same impact.
HS: Was the work you did in your first postdoc a continuation, of sorts, of the work in this paper?
MK: I had a Miller postdoc at Berkeley, and I was sponsored by Dave Wake, an incredibly broad-minded and generous scientist who was the director of the Museum of Vertebrate Zoology at Berkeley. I am the only theoretical biologist to have a postdoc at that museum. So, I interacted with all those people and had a really interesting and very fun time. The postdoc was not dedicated to a particular research topic or theme. I did do a follow up project leading to paper that in retrospect is not a very good, about the so-called “sexy son hypothesis”. I think, probably, the biggest single thing that I worked on during my postdoc is a project that I still am very interested in: how do traits that are measured by a function, like a growth trajectory or reaction norm, how do those evolve? That work was an extension of the Lande quantitative genetic framework, and so, that theme that I started as part of my thesis was fundamental to that work.
HS: It’s now 34 years since that paper.
MK: Ah, don’t tell me that!
HS: I’ll read out what you say is the primary conclusion of the paper.
MK: Good. Remind me, because I have no idea.
HS: You say, “The primary conclusion of the present paper is that the initial selective advantage for the female preference assumed by Fisher, O’ Donald, and many later authors are not necessary for either the origin or subsequent elaboration of mating preferences for traits associated with reduced survivorship.” Today, has your thinking about this model changed in any way or is it, more or less, the same?
MK: Wow. Thank you for reading that. Well, I guess, maybe, it’s a testimony to my own lack of mental flexibility, but I guess I’d have to say that that’s pretty much the way I still think.
HS: That’s great!
MK: Well, it’s either great or horrible! My views have shifted a bit. I think what that paper did get right is mating preferences can evolve for non-adaptive reasons. It is mentioned in that paper that almost certainly any gene that affects a mating preference is going to have pleiotropic effects, perhaps on things not having to do with mating at all. Every mutation, every allele that has a phenotypic effect has pleiotropic effects. It’s almost inevitable that those pleiotropic effects that can have some influence on how the mating preference evolves. So I do think that there are a lot of “mating preferences”- female biases – that are responsible for the evolution of extremely elaborate male traits, that are there for reasons completely disconnected with mating. That much I think is correct. In this paper, and in some later papers, I was probably a little bit too militant about arguing against the so-called ‘good genes’ or indirect benefits. Much after this paper, I worked on the reinforcement of mating preferences, when you have two species that are hybridizing and producing low-fitness offspring. This is an idea that goes back to Dobzhansky. When you have that situation, you can have reinforcement of mating preferences that strengthen barriers and decrease the amount of hybridization. And I believe all, or virtually all, evolutionary biologists agree. And if you believe that reinforcement works, then you believe in indirect benefits, because reinforcement of mating preferences during hybridization is basically an extreme version of ‘good genes’. If you think about mating with a heterospecific as being a really bad genetic choice, that’s just an exaggerated version of picking a really bad male of your own species. So, it’s really a matter of degree. And I’m more open, I think, now to the possibility of genetic benefits, not just with reinforcement of hybridization between species, but with conspecifics as well. So, yeah, maybe I’m getting old and soft, but I’m more open to that being a viable hypothesis than I was.
HS: When you start describing the model you make assumptions about the natural history of the organism. First, polygynous males are contributing only gametes to the future generation. And two, there is no direct relationship between a female’s mating preference and her survivorship or fecundity. Today, would you still think those are reasonable assumptions, given what we have learnt empirically in this context?
MK: Those assumptions are what you might think of as the Hardy-Weinberg equilibrium assumptions of sexual selection. They’re not meant to be literally true in any actual system, but selectively neutral mating preferences do provide a good null hypothesis. It’s an important empirical question, as to the ways are that these assumptions are violated, and what generalizations we can make about the natural world in terms of how they are violated.
HS: You talk about another implication, about the finding that there is no stabilizing force intrinsic to sexual selection. You say, “It may be very easy for allopatric populations to diverge in their secondary sexual characters, and hence in their isolating mechanisms, leading to accelerated speciation.” Have we seen evidence of this?
MK: Oh, that has been a topic of controversy for a long time. I have the same intuition that almost all biologists have, which is that sexually selected clades seem to have a lot of species in them. I have not kept up with research in that area, but I know that there’s ongoing work by a number of groups on the connection between speciation and sexual selection. I don’t have any strong empirical reason to think there is, but intuition does suggest (and most people feel) like there should be. It’s sort of surprising that it’s been so difficult to nail down. That comment strikes me as something that I probably might have added after reading Russ’s paper. His 1981 paper is not really about speciation, but it has speciation in the title, so I’m wondering …
HS: You actually say it in the paper. You say, “This possibility first suggested by Fisher has been demonstrated by Lande”.
MK: Oh, good, I did the right thing.
HS: In this paper, you use a haploid model, but you mention a diploid model and mention you found some support for the diploid model. Did you work further on the diploid model subsequently?
MK: Well, after Greg came in and asked me, what’s with this passage in Fisher, and what do these papers by Peter O’ Donald mean, I wrote some simulations of O’ Donald’s diploid recursion equations, and I discovered that if you started them with different initial conditions, you came to different equilibria, which O’ Donald had not noticed. And that was the first time I thought, Oh, my goodness, there’s something more to this than people have figured out. I don’t remember for a fact but I would bet that Monty was the person who said, you really should work it out in a haploid first, because one of the things you do as a theoretician is to try to understand the simplest case before you make it more complicated. And turns out the diploid model in this setting is actually pretty tricky. Richard Gomulkiewicz, a few years later, showed that dominance can qualitatively change the outcome and destabilize the neutral curve. Some funny things happen, but the equations get really complicated also. So, I think, that from probably some good advice and some good luck, the haploid model turned out to be the way to go. You know, the algebra is already bad enough for the haploid model and it gets really bad with the diploid.
HS: In these 34 years, have you ever read the paper again?
MK: Well, I certainly haven’t within memory. I’m not someone who actually likes to go back and read what I’ve written. It’s a little bit painful. Even if you don’t remember the prose, you remember the conclusions. And really, the excitement is in what you’re now thinking about, now working with on. I guess, historians and sociologists are interested in back stories. Different people have different things that excite them.
HS: Would you count this as one of your favorite pieces of work?
MK: I guess I would. I do feel it has had a positive impact. And I do think it has some interesting conclusions. Yeah, it’s very satisfying that way. Nobody wants to do work and not have it have an impact. There’s no use writing something and burying it under a rock. And even if you do not get public attention or promotion or accolades for it, you would like what you do in this world to have some impact out there. And I think that paper has, probably, mostly, for the right reasons.
HS: You said you like it for the impact it had. Do you also enjoy doing this kind of theoretical work?
MK: Yes, I really love having a new abstract idea and working it out. I mean, that’s like having a new toy. So yeah, that was my first new toy of that mode. And that’s still what gets me really, really excited about science. Doesn’t have to be, necessarily, a bunch of algebra. If there is a really interesting concept for me that I work on using the tools that are in that paper, I can be happy for days. That’s a kind of research that I really love doing. Still do.
HS: What would you say to a student who’s about to read this paper today? Would you guide their reading in any way or add any caveats?
MK: I’m sure they’re going to pick and choose what they want to believe anyway!
HS: Would you point them to other papers they should read along with this
MK: Maybe a way to ask that question is, what would be a better paper to recommend than this one, if you had to read one paper in this field? And I’m not sure what that would be. And this is embarrassing because I do teach undergraduates and so I should be able to say, here’s what you should read if you want to get introduced to this field.
This is a pretty technical paper, so I don’t think that – other than people really interested in sexual selection, or people are interested in modeling, or people who have a historical interest, such as yourself – I’m not sure that you actually have to read this paper. If you want to get interested in sexual selection, there are going to be more general, and certainly more up to date, bigger overviews. So, I honestly don’t know if people are actually reading this paper anymore. There’s a certain kind of paper that gets cited by people who have never looked at it. There’s a hilarious anecdote about a typo in a citation in E.O. Wilson’s book, Sociobiology.The incorrect page numbers in this one citation have been repeated thousands of times. So we now know how many people have not actually read the original paper! I have no idea if anyone’s actually reading this paper. Maybe it doesn’t matter.