In a paper published in Nature in 2002, Alan Bond and Alan Kamil showed, with experiments using real jays and digital moths, that frequency-dependent predation led to the evolution of greater crypticity and phenotypic variation. Fourteen years after the paper was published I spoke to Alan Bond about how he got interested in studying effects of predation on prey coloration, his collaboration with Alan Kamil, memories of lab work and what we have learnt since about this topic.
Citation: Bond, A. B., & Kamil, A. C. (2002). Visual predators select for crypticity and polymorphism in virtual prey. Nature, 415(6872), 609-613.
Date of interview: 25 August 2016 (via Skype)
Hari Sridhar: I’d like to start by talking a little bit about your motivation for doing this work. This is not the first piece of work you did on this particular study system, or even your first in virtual ecology. So, I’d like to step back a little bit and ask you how you got interested in the topic of search images.
Alan Bond: Sure. I first got interested in searching image when I was in graduate school. I took a seminar course – searching image was one of the topics – and I got hooked and started reading the literature. And I realized that there was a deep flaw in the explanations that people were giving. Apostatic selection, where a predator switches back and forth between several different prey targets, developing a searching image for one and then switching to another when the current one becomes less common,simply couldn’t work the way people were explaining it. They were describing it as a balance between learning the appearance of the prey and then gradually forgetting it after a time and having to relearn it. That didn’t make any sense at all. Animals don’t forget the appearance of their food and certainly not in a short enough period of time to cause this kind of dynamic to occur, in terms of switching from one prey to another. So, I got interested in it and I applied for a postdoctoral fellowship from NIH to work in a psychology department, because I thought that psychologists might have a clearer idea about how this mechanism might work. I went to Donald Riley’s lab at UC Berkeley, and I did my postdoc on pigeons, where I set them up to look for cryptically colored grains (beans vs wheat) on a background of gravel. It’s not a very difficult task for pigeons; all you have to do is get them to eat — you don’t have to train them to do that. And they show very nice switching effects, concentrating first on the most common grain type and then switching to the rarer one. But pigeons only do this when the grains are on a cryptic background. You put the grains on a conspicuous background, and the pigeons take them in random sequences, in proportion to their relative abundance. So what that told me was that there really was a searching image, and that it probably would, under the correct circumstances, produce apostatic selection. But because I couldn’t actually see the sequence in which the pigeons were eating the grains (the birds were too fast) and couldn’t determine the time intervals between successive discoveries, I was only able to show mathematically that their behavior was consistent with the use of searching images. So I put it on the back burner and I went off to other things. My wife and I came out here to Nebraska, initially, for her job. But a friend of mine, Al Kamil, had come out about the same time to a faculty position at the School of Biological Sciences. This was in about 1990-1991. And we got to talking and he speculated that it should now be possible to do searching image research with graphical computer images. Essentially, the hardware and software capabilities had matured enough to make it feasible to ask the searching image question with great precision. So we decided on a collaboration. He’d done all of this work for years involving blue jays looking for moths in slide images, so I thought, well, all I have to do is become creative and do some graphics programming, and we’ll see what we come up with. So we did that original experiment, which was published in, I think, 1998. And what that demonstrated to me was that if you set up a system with a set of fixed morphs – I think were four in that first study – and set it up so that the morphs didn’t interbreed but just reproduced asexually in accordance with their population numbers – that you got perfectly good apostatic selection with a nice oscillatory pattern. It was clear that the birds were, in fact, changing their attentional focus depending on how frequently they were seeing these different morphs. So we sat around and talked about it and put out a paper in Nature, and then we said,you know, we really have to ask the question: As a consequence of this effect, would birds that were searching an initially uniform distribution of interbreeding prey,would they, in fact, select for morph differences? Would the prey population become more diverse under predation? So that’s how we got into the issue. At that point, I had to learn genetic programming and a bunch of other fun things, but it was it was actually very cool. And it really got to the base of the question: given that this underlying mechanism is operating, is that mechanism capable of generating polymorphism? So, that was where the idea came from.
HS: Was the phrase “virtual ecology” used for the first time in this paper? Was this one of the first experiments that used this approach?
AB: Yes to both questions. I mean, it is possible someone else used the term; I didn’t look for it, I just used it. As far as I know, it’s the first and was the only experiment that combined live animals with a genetic algorithm, let them, actually, cause changes in the artificial ecology by making choices.Sort of this hybrid of robotic and living components that made it so fascinating to people, just the fact that, you know, it had that edge, which made it seem really exciting. So, yeah, ‘virtual ecology’ was our creation.
HS: It’s surprising that it hasn’t been used more, because it seems like such a powerful approach.
AB: Well, that’s a good question. Part of it is that the laboratory procedures involved in training animals to do this, and even the business of getting the artificial genetics right, is very laborious.It was very hard work. And I think that that combination of animal behavior plus computer programming and simulation just discourages a lot of people from wanting to get into it. There are a number of studies where people have used a human interface to look at visual selection. Humans are easy because you don’t have to train them to do anything. They just tap the screen. The interface was originally designed for human access. So then it comes down to whether your genetics are realistic or not. And that’s certainly a problem. We designed the virtual genome for this experiment with a great deal of care. It’s actually based substantially on what’s known about the genetics of wing coloration in butterflies. I even checked it out with a couple of people who are authorities in the field, and they said, well, it’s sort of like a cartoon, but, yes, that is more or less the way it works. So, we felt comfortable that whatever we concluded from it might have at least some applicability to real animals. As it happens, of course, the result was not what we expected, but, you know, that often happens.
HS: Tell us more about that. Why wasn’t it what you expected?
AB: Let me take a couple steps back. Apostatic selection was well known as an explanation for stabilizing classic polymorphisms, where you have a limited set of discrete morphs that are very distinctive from one another and that do not change over time. Each animal’s morphology is fixed in the genome.When you actually look at the genetics of morph coloration, however, it is far from a single locus switch. Most of themse polymorphic species have very peculiar complicated mechanisms that essentially prevent the morphology, the coloration, from getting modified by recombination. For example, there’s a grasshopper population here in the midwestern US with something like 30 different morphs, and they’re all very distinctive, and they behave as if they were based on a couple of simple Mendelian loci. In fact, they aren’t; it’s very complicated, but the DNA structure suppresses recombination in such a way that the color morphs act as a set of single competing genes. So what happens when you do what we did,which is just sort of turn the birds loose on searching for prey with a simple genome, something like 120 genes, that did not force the prey into a limited number of distinctive morphs – what happened was that the birds, by using searching image and switching from looking for one kind of morph to another, what you got was a great increase in variance. They prey were more cryptic, but they were also more variable. The result did not boil down to just a limited number of morphs. And that’s in spite of the fact that I included in the virtual genome mechanisms that should have enabled selection for the evolution of recombination inhibitors. We always got at least some recombination, so there were always distinctive, conspicuous forms that would suddenly emerge. The final population was not stable, at least not in the same way that real, classic polymorphisms are stable. So, what that says to me, looking back on it now all these years later, what it says is that predators can stabilize these polymorphisms and can maintain a set of differently appearing prey items, but you have to have some preexisting genetic mechanism in order to force the variation into a set of distinct morphs. I’m guessing that it’s actually relatively difficult to evolve, because if you look across taxonomic groups, you see that discrete polymorphisms tend to occur in particular taxonomic families,and there’s often evidence suggesting that it’s the same basic genetic mechanism that is responsible across related species,that having this mechanism allows all of these different species to become polymorphic. But if they don’t have it, what you get is continuous variation. You get what’s been called massive polymorphism, where the number of different forms proliferates, but they’re not discrete. Otherwise,the experiment was very successful. I mean, it had all of the effects that we had hoped it would. And people were certainly happy with it. And it’s, you know, been included in textbooks and websites all over at this point. But it’s not the final answer. Eventually, someone will figure out an even better way of doing it and I’ll be very impressed.
HS: Was your training and most of your research until this point mostly in psychology?Did you learn all the genetics and programming specifically to do this work?
AB: Oh, no. I’d had a lot of prior experience with research on predatory behavior. My PhD was in zoology and animal behavior, and I’d always been interested in behavioral mechanisms. I did a lot of simulation for my doctoral work, which was on insect predators. Then, after I finished my postdoc, I spent the best part of a decade working in the real world, doing contract biomedical programming, analysis, and simulation. So, I’d had quite a bit of computer experience before getting involved with jays and virtual ecology.
HS: Can you give us a sense of the programming you did for this study – what computer platform did you use, how long did running these programs take at that time, etc?
AB: We were a bit constrained by the apparatus. The birds are in a box with a touch screen on the front of it. The touch screen and the reward mechanism that provided the birds with food pellets,those were run by beater PCs, old IBM machines, because it’s a dirty environment. I mean, you’ve got birds and bird poop and bird feathers and dust all over. I wanted equipment that was going to be tough and that, if it went soft on me, I could replace it pretty easily. So we used cheap machines. The whole thing was run with C++ programs on DOS. Very robust, very solid. Initial configurations were accessed from a Linux box in a clean location elsewhere, and the results were written back over the same network links. I could access the Linux box remotely to keep track of how experiments were going, but the individual DOS machines were totally secure – the code itself was not visible from the internet. Moth evolution occurred in stages, one generation of 100 moths per day. A new generation was produced by random breeding within the old one and downloaded each morning when the machines were started up. So, the system had to be capable of doing the evolution of the new morphs each morning when you turned it on. It didn’t take very long. I mean, it was a pretty efficient bit of code. I don’t know, it took maybe about 10 minutes or so, when you first started it. It worked okay. I showed the code to people and they said, oh my god, you did this in DOS! I had all these old DOS tricks to wedge all these great big matrices into the 512 kb memory limits. But it all worked.
HS: What about the touchscreen? Was that something you built in-house or was it commercially available?
AB: It’s still commercially available, I believe.What was most important was that it had to be an infrared touch screen, because bird beaks are unlike human fingers –they won’t make a capacitive contact. So, we had to a system where the peck breaks a beam in order to register a touch. And these people made touch screens as interfaces for, like, industrial environments with lots of grime and dust. The advantage of infrared is that it’s pretty sturdy, and you don’t have to worry about damage to the screen. So, I think that it just worked out really well. It was almost like they were designed for working with birds. Decent resolution. And they had a software package that that allowed me to interact pretty cleanly with it. It was straightforward, really.
HS: Do you know what’s happened to this setup?
AB: Oh, you mean all the equipment or just the specs?
HS: The equipment?
AB: Now that all the old folks have retired, the lab has passed to Jeff Stevens, who is faculty in the Psychology department. He’s continuing to work with jays (though not doing virtual ecology), so he has at least four or five of those boxes in operation at the moment. I don’t know how long he’s going to keep it up. We’re going through a difficult place with respect to funding in the US. It’s basically gotten very hard to get governmental support for doing lab work on animal behavior, particularly any animals that aren’t rats or pigeons. I think that, ultimately, the lab is going to shut down. What will happen then, I’m guessing, is that they’ll just break up the units and turn them over to other people to do other things with.
HS: Maybe all of this should maybe go into a museum somewhere.
AB: That’s a clever thought! We could make a virtual ecology museum. And yeah, that’s very cool. I’ll have to mention that to Jeff to make sure that he hangs on to at least one of the units so that we can make an exhibit of it.
HS: All these pieces of equipment that have been designed for experiments are part of the history of our discipline.
AB: I hadn’t even thought about it. You’ve got to recognize that scientific equipment doesn’t look all that great — it’s all put together with bits of wire and duct tape; not exactly impressive! But it’s an interesting notion. When I go back to campus next week, I will drop by and see what remains of the lab and whether it can be preserved. The system doesn’t look like something that should be artistically interesting to anyone, but I think you’re correct. I’m not sure how much more of this kind of work is going to be done in the future.The funding is so poor, and the work involved is very demanding. I’ve got some plans to do some experiments with moths and backgrounds, but I would be doing them on iPads with human subjects. I think I’m out of the laboratory bird business. My wife and I have done a lot of fieldwork on birds. We’re have just published a book on the behavior of wild parrots. But we’re not going to bring parrots into the lab if we can avoid it.
HS: Which species was your model for the digital moth, and how did you pick it?
AB: The basis for the original digital moths was Catocala relicta, an owlet moth from Northeastern US. They have a five morph system, ranging from almost pure white animals to ones that look very dark and similar to tree bark. And the reason we chose them is that they are, in fact, actually eaten by blue jays in the wild. Al Kamil had done a bunch of his original experiments on searching image back in the 1970s. He used specimens of real moths of this species pinned to logs in the forest, taking color photographs of logs with and without moths. Then he showed the slide images to blue jays in the lab. So it seemed logical for us to use the same species. In the virtual ecology set-up, it was necessary to make the moths quite small to make the task difficult enough, because blue jays are really good at this. We only had 640*480 screens, only so many pixels to work with. So we ended up with 16*16 pixel targets, which on those computer screens are about the size of a housefly. It’s often a challenge for the birds to find them. But you need to make the task hard enough that the birds will make mistakes – they will overlook at least some prey items that are actually present. From a human point of view, the targets do look like little insects. If you enlarge them so that you see the 16 by 16 pixels they don’t look much like anything, but when they’re shrunk down to natural size it’s believable. The resemblance to moths, of course, isn’t necessary. People always ask whether the birds think of the targets as insects, and, I have to tell them that that is very unlikely. Birds have four different visual pigments rather than just three the way we do.They see into the ultraviolet, they see into the infrared,and what they’re actually seeing when they look at a computer screen, god only knows. I’m guessing that what they’re looking at is textures in shades of purple. But in any event, they don’t think that they are moths.These birds have been hand raised in the laboratory. They’ve never even seen a real moth before. So what they’re learning is just that these visual patterns, if they are pecked, will cause food to drop into their cup. And that’s all they need to know.
HS: Where did you capture the blue jays you used in this study?
AB: We captured them locally here in Lincoln, Nebraska. Blue jays are difficult to work with in the laboratory because they’re very wired. They’re very intense. If you catch them as adults, they never tame, they never relax in captivity. They will beat themselves up on the wires. So what we have done– this is the way Al Kamil developed it years ago — was that you find a blue jay nest and you go and watch it for the point at which the young hatch, and you look for the moment at which the nestling’s eyes are just opening. It’s about 10 days after they hatch. At that point, you take the babies out of the nest, bring them to the lab and then you hand-feed them. Essentially, they associate people with food for the rest of their lives. But they never get pleasant about it. They don’t like you and they will bite you on a regular basis, but they do behave properly in the apparatus.
So, they were locally caught here in Lincoln. Some of the best places are nests in people’s backyards, or there are several parks and graveyards.Cemeteries are very good for blue jays and we’ve picked a number of them from there. Many of the last set of birds we acquired are still at work. They will live up to 20 years in the laboratory. You can’t release them, obviously; if you raised them by hand, you can’t release them. So even though we’re no longer running blue jay experiments, the birds have been adopted by researchers in Canada and in Minnesota and Iowa. Some experiments with them are still continuing.
HS: Do you mean the four individuals used in this study are probably still being used an experiment somewhere?
AB: Oh my. This study was 2002. I’m pretty sure that all of those individuals must have died by now, or at least fully retired from experiments.
HS: Did the birds have names?
AB: Well, I tried to discourage the lab techs from giving them names. They all had numbers on their leg bands. And that was the whole point, that all the people running the experiment needed to do was to start the computer, enter in the bird band number, check to make sure that was in fact the correct bird in the box and press the start button. Everything else the computer handled. But I know the people who fed and cared for the birds gave them names. People always name lab animals.
HS: Were you and Al Kamil present, along with the lab technicians, when the experiments were happening?
AB: I was, certainly at the beginning. And I ran test sessions on myself to make sure everything was working; I don’t know how many sessions I went through having to search for hundreds of obscure models on backgrounds. It’s not fun for human beings, I’ll tell you that. People don’t have the patience that the birds do. I set the program up running within batch files — complicated arrays of batch files — so that, in fact, people who were comfortable handling birds but knew nothing of programming or of the of the issues being tested could come in, put the bird in the box, push the button, and make the session work. Usually, we had about four or five technicians and one lab manager who would report back to me when things went wrong. Of course, they would go wrong periodically at the beginning. Software is never bug free. But the system was rock solid after about a week’s worth of practice runs. Sometimes using ancient software has advantages.
HS: Did you do most of the writing? Do you remember how long it took to write the paper?
AB: Actually, Al and I collaborated very extensively on the writing. I did a lot of the analysis, and then I had to explain it to him and justify it. And then we went back and forth, and he came up with other ideas that I hadn’t thought of. So, it was pretty collaborative. I think that, if you combine the analysis and the writing, it took the best part of a year.
HS: Could you give us a sense of how you and Al Kamil shared and discussed drafts? Did you sit together to work on the paper and the analysis, or was it done over email?
AB:Oh, we were right down the hall from one another. We didn’t have to do it remotely. We’d sit and look at screens and look at printout and make suggestions back and forth. It was a hands-on collaboration. It worked very well at that level. I realize that most of the underlying programming and experimental structure was mine, but the basic issues were a real collaboration. And, of course, it incorporated all of his methods of dealing with the birds and the touchscreens and raising them and training them and so on. It was a joint project.
HS: Were you a postdoc at that point?
AB: Oh, lord, no. I was 40 years old. I was a research professor in the School of Biological Sciences, which meant that, the grant provided my income, but otherwise I was faculty.
HS: Did this paper have a smooth ride through peer review? Was Nature the first place you submitted this to?
AB: Yeah, Nature was our first choice. I think we had to do some revision.Very few papers are accepted right off the bat anymore. But we only really had to do about one revision in defense. And then it went straight out. It was very popular and there was a lot of media attention given to it. For at least a while. We even got a full page display of digital moths in Wired magazine. That suggested we had caught on with at least some of the software people in the media.
HS: It has attracted quite a bit of attention in academia as well; it is used as a textbook example. At the time when you’re doing the work did you anticipate that it would have such an impact?
AB: Oh, did we guess that we were going to become famous, or at least famous for 15 minutes? No, I’m not sure we were at all aware of making history, in any sense. All we were trying to do was to ask the basic question, which is, whether the predators could cause moth populations to become polymorphic purely as a result of selecting similar forms. So, really we were just totally fascinated with the issue. Although, at the National Science Foundation, which covered most of the costs of this project, the people there were really interested in it and they thought we were going to become famous. Maybe they knew something we didn’t.
HS: Do you have a sense of what this paper mostly gets cited for?
AB: It’s definitely cited for its indication of the effect of predation on variance in the prey population. And the evidence is fairly strong that predation can, in fact, select for crypticity as well as for high variability. We did a subsequent study looking at this process on heterogeneous backgrounds –mixed dark and light backgrounds — that work came out in PNAS several years later. It has also been cited fairly heavily because of its indication of the role of background textures in maintaining polymorphic prey coloration.
HS: Could we go over the list of people you acknowledge to find out more about how you knew them and how they helped?
AB: Sure. I don’t remember who’s in there, but okay. Who’s in the Acknowledgments?
HS: J. Allen
AB: Yes. John Allen was the editor of the Journal of the Linnean Society at the time. I’d never actually met him in person, but I’d known of him and we’d corresponded about this and he actually looked at early versions of the paper and gave some suggestions for other ways of describing the analysis.
HS: R. Balda
AB: Russ Balda was, at the time, the third faculty member for the Center for Avian Cognition, which is what we called the big lab where we did this work here at UNL. Russ was much more a field ornithologist, so he had less to do with the details, but he certainly gave us lots of suggestions and was very encouraging. Right now, he’s retired too. Of course. Everybody’s retired.
HS: J. Endler
AB: John‘s a good friend. I had known him since – oh, I don’t remember when – I think we first met at a Society for the Study of Evolution conference years ago. But he’s what I consider to be a one of the true geniuses in talking about animal coloration and crypticity, and he’s given me a lot of feedback, both on this work and on several other things I’ve done. My wife and I wrote a book called Concealing Coloration in Animals — came out in 2013 — and he was a major reviewer of that publication for Harvard University Press.
HS: T. Getty.
AB: Tom Getty is another old friend. He was at Michigan State and was basically a mathematical ecologist.But he did a lot of experimental work on how birds go about finding cryptically-colored food items. And we had several long discussions about this work. I gave a presentation at his lab at the Kellogg Biological Station, as well as interacting a lot by email and on the phone.
HS: L. Harshman
AB: Larry Harshman was in the faculty here at UNL. He’s a fruit fly geneticist. And he was one of the people that I talked to extensively about the validity of the artificial genome.
HS: S. Louda
AB: Svata Louda was also faculty here in Lincoln. She’s retired too. She worked in plant-animal interactions, but she had a number of suggestions about looking at the way we were talking about birds foraging for prey items. I remember, particularly, a long conversation with her on the topic.
HS: D. Pilson
AB: Diana Pilson is also a botanist.She was running the ecology and evolution seminar series at the time we did this work. And she had some very clear and to-the-point criticisms of the way I was talking about the relationship between what we were seeing here and what is ordinarily thought of as apostatic selection. So it made us much more conscious about how we discussed it in the text.
HS: S. Shettleworth
AB: Sara Shettleworth was a psychologist and old colleague of ours. She’s from Canada. And she’s done a lot of work with birds in apparatus like this. I talked to her at one set of meetings very extensively about how we were running these touch screens, and she had a number of suggestions about little tweaks or changes in the way the food was being delivered and so on that, I think, improved the apparatus considerably.
HS: What kind of an impact did this paper have on the future course of your research? In the last part of the paper, you say, “In the next stage of our explorations of virtual ecology, we will introduce a range of heterogeneous backgrounds to examine the evolution of more classical, discontinuous polymorphisms. This will allow a direct test of the hypothesis that high phenotypic diversity is, as has been suggested, a step along the evolutionary path toward more discrete and genetically integrated polymorphisms.” Is this something that happened and formed a major part of your research subsequently?
AB: At the time we published that paper, we were already beginning work on background heterogeneity. That’s the PNAS paper that came out subsequently. Again, it was a good story. The results were quite clear.If you have dark backgrounds and light ones mixed together, you do, in fact, get greater heterogeneity and it tends to compartmentalize the population into light and dark morphs to some degree, but the moths still do not become simply dimorphic. What the selection did was it operated differently on dark and light backgrounds, so that you got populations that were a mixture of dark and light moths, but you did not, in fact, get morphs that were uniformly dark or uniformly light. But it was a very good story, and this one has also been cited quite a bit. It’s more technical and harder for people to grasp than the original 2002 paper. So, I don’t think it’s been put in textbooks. But it’s been very popular.
HS: Would you say that the main conclusions of this paper still hold, true more or less, today?
AB: Yes. In terms of the main conclusions, yes, I think so. The question is, what else is involved, because clearly something else is involved. We ran several other experiments looking at alternative factors that might affect the evolution of discrete morphs. So we gave our virtual moths genetically encoded behavior, so that you could select for moths that that were more or less likely to land on backgrounds that they tended to match. And you could select for moths that tended to mate with other moths that looked like them. So, we were looking for any indications of other external selective factors that could encourage the development of discrete polymorphism. And the answer was that behavioral factors did not work either. There’s clearly something that has to happen at the genetic level that predisposes moths to form discrete morphs. We were not able to access it. So, we did not have the final answer. And that’s not exactly unusual in science. I don’t feel disappointed with the work at all. What ultimately happened is that after that set of experiments, I got more and more interested in just the question of perception in birds and what determined what they attended to, and how long that attentional state persisted,and what sorts of consequences the features of the background, like the degree of complexity, or the number of distracting stimuli on the screen and so on, how those affected the efficacy of searching image. I guess I’ve got, like, three or four more experiments where the analysis has been done but the paper hasn’t been written yet. So, I’ve still got quite a bit of work to do, even though I’m retired. But we didn’t do anything more on polymorphism.
HS You have partly answered my next question, which is, if you were to redo this experiment again today, would you do anything differently, given the advances in technology, theory, statistical approaches etc?
AB: Well, the first thing I would do, now that I know what I know about how the dynamics of searching image works, is to simulate the whole system. I wouldn’t just simulate the moth population. I would simulate the birds, too. I could make what I call a robojay, which you could put up against the initial stimulus configuration, look at its evolutionary effects, and then compare the simulation results to the effects of living birds. It would be a much stronger study, statistically, as well as producing a broader range of hypotheses to investigate. If I ultimately get time, I’m going to go into doing some more extensive simulations and see if I can tease out a better way to ask the question about what determines the selection for discreteness. I think that some more extensive simulations and more careful design of the genome might prove to be valuable. I don’t know that I would run another captive bird study. It’s just too much work and I can no longer get the funds for it. But I can certainly do simulations. And I can certainly do similar kinds of experiments using human subjects that may ultimately get around to testing the same hypotheses. I still have a few more years.
HS: In the paper you say, “To our knowledge, these results provide the first evidence of searching image effects where prey appearance was continuously variable.”Subsequently,have there been other such examples?
AB: You know, I can’t think of any. There’s a lot of new literature that is consistent with predatory effects on discrete polymorphism. There is not a lot in which people have looked at the actual nitty-gritty of how predators interact with continuously variable prey populations.
HS: And in another place you say, “Instances of phenotypically diverse or massively polymorphic cryptic species are by no means uncommon in the literature”, and you cite examples from brittle stars, bivalves, grasshoppers, fish, and noctuid moths. Have any more such examples been documented, subsequently?
AB: We mention several of them in the coloration book. I wouldn’t guess how many. It’s very interesting. There’s a lot of this work that’s been done on desert animals that become selected for matching particular differently colored backgrounds. It’s not a discrete polymorphism; it’s a consequence of high levels of local selection to particular background colors. There’s a lot of literature on this now.
HS: Have you ever read the paper after it was published?
AB: I’m sure I’ve read it at least a couple of times. It’s funny how you get attached to particular papers and you will suddenly, for whatever reasons, come across them again, and read them again. The familiarity is very attractive. It’s always enjoyable to read a paper that really worked well; when it was your paper, obviously! I have even re-read books I have written and enjoyed them. You look back on these things and think about how much work they were (our parrot book was five solid, grinding years in the making). But if you are lucky, you can still think, gee, yeah, this was good. This was this was fun to do. It was all worth it.
HS: If you compare this paper to the papers you write today, do you notice any striking differences in the style of writing?
AB: As a consequence of writing our books, which are specifically designed to be accessible to the general public as well as to other scientists, my writing style has become much simpler and clearer. I try to be much more aware of easier ways of talking about phenomena instead of just coming up with the first pompous piece of jargon that occurs to me. So, I do think that with experience I’m getting better at writing clear prose. That doesn’t necessarily mean that everything’s going to be accessible to everyone, but I think that I’m doing better.
HS: Would you count this as a favorite among all the papers you published?
AB: Oh, that’s like asking which of your children you prefer. It’s certainly one of my favorites, mostly because there was so much of it that we were creating absolutely from scratch. We were really the first people to think about it this way and to do it this way. And you don’t get many opportunities to be that innovative. Most science is actually derivative, to some degree. It’s derived from other people’s work; it’s derived from your earlier work or whatever. There aren’t that many occasions where you could say, gee, you know, we created this from the very beginning. That always makes you feel good.
HS: What would you say to a student who’s about to read this paper today? What should he or she take away from this paper that was written 14 years ago?
AB: There are two things that occur to me. One of them is the degree to which you can use computer interfaces as a replacement for an ecological interaction, and thereby ask some very sophisticated questions. I think that the methodology itself is open to some really potentially very exciting things, if people were willing to put in the effort.You can make whole worlds that animals and people can interact with. And you can control a lot of things that are completely inaccessible to you in the real world. I mean, when you’re talking about ecological phenomena, what you can do in terms of field experiments is really limited, in comparison to what you can do in interaction with a computer model, provided the computer model has the requisite complexity and at least similar dynamics. You can ask really cool questions. And I would hope that students who read it think about what could they do that could make use of this kind of system and ask this kind of question. And, and in some ways, that aspect of it, the virtual ecology aspect, is, I think, of longer-term interest than the specific question we went after. Really, the issue is, you know, what can you do in this kind of a hybrid system where you’re combining real animals and simulated environments that you can’t manage looking at real animals in real environments because you’ve got much better control. In some ways, I think that that’s really the most lasting story that I would like people to be able to carry away from it.
The second issue is the fact that natural selection is not just the interaction between predators and prey as if they were molecules or balls colliding. There’s the underlying psychology of the predators and there’s defensive responses in the prey. And the result is a very complicated interaction that, in fact, has major influence on how species evolve. The idea that there’s a psychological dimension to behavioral ecology has grown very slowly in the literature, but it is now, really, very broadly accepted. It’s clear that you have to take into consideration how the animals think, and what they do with the information that they take in, in order to understand how they’re going to interact and how they’re going to affect evolution in one another.
0 Comments