In the summer of 1976, having just completed his PhD from Oxford and with some time to spare before his next job, Nick Davies decided to do a short study of speckled wood butterflies in Wytham Woods, where he was living at the time. Davies observed that male speckled woods seemed to defend sunspots in the woods and chase away any intruding males from them. He decided to investigate why they do so, using a mix of natural history observations and simple removal experiments. The results of this study were published in a paper in Animal Behaviour in 1978. Thirty-eight years after the paper was published, I spoke to Nick Davies about his motivation to do this study, what he found, memories of field work, and what we have learnt since about territoriality in butterflies.
Citation: Davies, N. B. (1978). Territorial defence in the speckled wood butterfly (Pararge aegeria): the resident always wins. Animal Behaviour, 26, 138-147.
Date of interview: 3rd August 2016 (on Skype)
Hari Sridhar: What was your motivation to do this study? Was it to test the theoretical ideas of Maynard-Smith and Parker (1976)?
Nick Davies: No, absolutely not. This came about completely by chance. I had just finished my PhD, a study of the behaviour of a little bird, the pied wagtail. My pied wagtails defended territories along a river, and I noticed that every so often, a territory owner would leave his territory to feed elsewhere, particularly on days when there wasn’t much food on the territory. But periodically throughout the day, he would keep coming back to check if there was anyone intruding on his territory, and if there was, he would chase them off. I was very interested to know why he was doing this. I wondered if he was trying to prevent any newcomer spending sufficient time on the territory that it would get to learn its characteristics and, in effect, think this is a jolly nice place to live. If the owner came back periodically and chased off a newcomer before it had time to learn about the territory, it would be easier to get rid of him. To test this idea, I wanted to catch the owner and keep him away for a sufficient length of time for a newcomer to learn the characteristics of the territory, then put the original owner back and ask if he would then find it harder to win back the territory. I tried for a winter and failed completely simply because the birds were too hard to catch.
I was then living in a little chalet in Wytham Woods, on the edge of Oxford, where David Lack started his famous studies of the great tit. I had a spare summer because I had just finished my thesis and my next job didn’t start till the autumn. This was 1976, so 40 years ago, and it was romantic living up in the woods. Every day was sunny which was highly unusual for England. People still remember that glorious summer. I think we had a three month spell with cloudless skies and the wood was full of butterflies. And I just noticed some butterflies doing little spiral flights in sunny patches, and I thought, well, that looks like territory defence. Maybe I can do the experiments on them instead of the wagtails because butterflies would be much easier to catch. At that time, I did not know the Maynard Smith and Parker paper. The whole thing came through natural history curiosity and serendipity, I guess.
HS: Were you already in the Edward Grey institute by this time?
ND: I did my PhD at the Edward Grey institute, which is for studies of birds. So, I was playing truant a little bit by studying butterflies. But the same theory of territory defence can be applied to birds and butterflies, so I thought it would be fun to play around with butterflies. I’m a naturalist and I like watching animals and wondering why. At first, I had no idea what species of butterfly it was and had to look in a book to discover they were speckled woods. I had never watched butterflies before. I was a complete newcomer. And in many ways, I think that was the advantage – to come completely fresh at a system, just ask simple questions. I reread the paper this morning, for first time in 40 years, and I was pleasantly surprised. Often with past papers you are a little bit embarrassed about what you have written but I thought this still reads quite nicely. I don’t think I could a better job now, despite all this extra experience.
HS: Yes, it is a lovely read! It is very different from papers of today. There is a lot of really nice natural history and it is written in a very engaging style.
ND: Thank you. That’s very kind.
HS: You spent a little over a month – 25 July-30 August 1976 – doing this study. Can you give us a sense of what your daily routine was like during this time?
ND: I spent the whole time – every day, all day – watching butterflies. I almost never went into town, except to get shopping at the end of the day. I would just wander around with a net and a notebook and a stopwatch. That’s all I used. The first thing I did was to mark a lot of individuals so they could be recognised. I remember being delighted how easy this was compared to trying to catch pied wagtails. I would see a butterfly in a sunny spot, capture it in a net and mark it with felt tip pen through the net. I didn’t actually have to handle the butterfly at all. This way I marked several hundred during the study.
On my first day I decided to watch one individual for the whole day, just to try and get a feel for what it was like to be a speckled wood butterfly. I watched one male for six hours or so, in one little sunny patch. As the sun moved across the sky, the sunny patch moved over the woodland floor, and the male followed it faithfully, travelling something like 20 or 30 metres during the day, always staying in his little sunny spot. I noticed that he didn’t spend any time feeding there. He just perched and whenever anything came close by he would sally out and inspect it. Often it was another insect, perhaps a ladybird or another species of butterfly, and he’d then very quickly ignore it and return to his perch. But if it was another speckled wood butterfly he was very interested. If it was a male, they would have a brief spiral flight, and then one of them would retreat. It was always the owner who would come back to the territory. If it was a female he would follow and court her, but almost always she said no. She would fold her wings and depart, and he would then leave her alone. Subsequently, I discovered that these females which said ‘No’, were already mated and laying eggs. In butterflies, females can say ‘No’ and that is the end of it. There is nothing the male can do. But occasionally, a female would say ‘Yes’ and the male would then leave the territory and fly up and mate with her up in the canopy. These simple observations suggested to me that what the male was doing in these sunny spots was looking out for females.
The next question I asked was whether these little chase, these little spiral flights, were territory defence. You might think, well it is obvious they must be! But there was some controversy back in the 60s and 70s about whether butterflies really did defend territories or whether these little interactions were simply a case of butterflies spacing out amongst the available habitat. So I thought it was very important to test whether these other intruding butterflies really wanted to settle in the territories. So I removed the original owners, kept them in a net, and when one of these butterflies from the canopy came down I saw what they did. I discovered that in all cases where I had removed an owner, the new comer settled in the territory and began to defend it. And that suggested to me that these spiral flights had previously kept them out, so it really was territory defence. I was very pleased with that very simple removal experiment; these other butterflies were clearly keen to get the territories when the vacancy arose.
HS: Do you remember the colour coding scheme you used to mark the butterflies?
ND: I always put two colour spots on one wing, and used several colours. Speckled woods are a dark brown butterfly with pale yellow spots which are perfect for putting coloured spots on. I didn’t wonder whether particular colours might affect the behaviour of the butterflies. In retrospect, that would have been a good thing to check, to see whether the ones that I had given red spots, or whatever, did better. I never looked for that and should have.
HS: In one place in the paper you talk about sitting in a tree for seven and a half hours! Can you tell us a little more about that? Was that seven and half hours at one stretch?
ND: Yes, I still remember that day in an ash tree! Butterflies defending sunspot territories sat on low perches, often on brambles or other leaves in the ground layer. Butterflies that were intruding came down from the canopy. I noticed that the canopy males searched for females in a different way. Whereas the sun spot males perched in the sunshine and waited to inspect things that passed by, the ones up in the canopy were patrolling. The canopy was more or less completely in the sun, so perhaps the best thing to do was to just patrol around and try and find females by active search. On days that were cloudy, but still warm enough for the butterflies to be active, I noticed that all the butterflies were in the canopy patrolling.
These different tactics for searching for females were fascinating. What I wanted to test was why patrolling males were so keen to get sun spot vacancies that arose. Was the woodland floor the best place to look for females? This is why I sat up in a tree all day! I simply scored the number of females that visited sun spot territories below, and the number of females that I saw being harassed by patrolling males above in the canopy where I was sitting. I found that the sun spot males encountered many more females than the canopy ones. Though statistically the evidence wasn’t very strong, together with the behaviour of the males, it convinced me that the males were competing for best places to look for females.
Another very important finding was that because territory vacancies arose frequently, most of my canopy males got a sunspot territory in the end. This suggested to me that a sun spot wasn’t an incredibly valuable resource, unlike my pied wagtails where only a small proportion of the population held territories. Here, it was almost like seats on a bus – everybody had a turn, if only they waited patiently for a little bit.
HS: You are the sole author on this paper. Is that because this project was done at a time when you were working more-or-less independently, when you were in between a PhD and a post-doc?
ND: Yes. I was also sole author on my PhD papers, except for some I did together with a fellow student who was a theoretician. In those days supervisors didn’t put their names on student papers, and throughout my career I have also never put my name on students’ papers. I know the tradition has now changed.
HS: But at the time of this study, did you have a peer-group of some sort with whom you were discussing this idea?
HS: In the Acknowledgements, you thank a person called Michael Brooke.
ND: Yes, he was a friend and fellow student who read a draft of the paper for me.
HS: Staying with the acknowledgements a little bit, I would also like to know how the other people you name contributed. Was Christopher Perrins the director of the Edward Grey institute at that time?
ND: Yes, he was the director of the Edward Grey institute at that time, so it was a courtesy to thank him for giving me the space to work there.
HS: Lawrence Packer?
ND: He was an undergraduate at Oxford doing a butterfly project as part of his degree – butterfly diversity in the wood. He encountered some of my marked butterflies and told me where they were, so I thanked him for that.
HS: Maggie Norris?
ND: This will seem really old fashioned! Maggie Norris was the secretary at the Edward Grey Institute. In those days there were professional typists, who would type up your manuscripts on paper, with a Carbon Copy underneath. The typed manuscript would get sent off to a journal by post and then you would wait for two or three months to hear whether it has been accepted or not. And all the figures were hand drawn, with Letraset spots for the data points on the graphs. Computers and the digital world had still not yet appeared.
HS: You say “This work was financed by a Natural Environment Research Council Studentship”. Was this a grant you obtained specifically for this piece of work?
ND: No, this was a grant for my PhD, and I snuck this extra project into it, in the few months I had left after I had finished my PhD.
HS: There is one figure in the paper that was drawn by Tim Halliday.
ND: Tim is a friend of mine. I did a study of toads with him a year after, and he was a good artist. I had sketched out this figure and he drew it much more beautifully then I could.
HS: Was he also in the Edward Grey Institute?
ND: He was actually in the Animal Behaviour group with Richard Dawkins, but it was all part of the same zoology building. Just separate research groups. We all had coffee together every day.
HS: Did this paper have a smooth ride through peer-review? Was Animal Behaviour the first journal you submitted this to?
ND: Yes. After I gave the talk I mentioned earlier, Richard Dawkins, who was the editor of Animal Behaviour at the time came rushing up and told me that I must submit the paper to Animal Behaviour. And so I did. I had two referee reports which were both encouraging. I don’t know who the referees were. I revised the paper and that was it.
HS: It didn’t change much from the first submitted draft?
ND: No, not at all. I might have made some minor changes.
HS: Do you remember how the paper was received when it was published?
ND: The theoretically minded people jumped on it as an example of a dispute settled by an uncorrelated asymmetry. I need to go back a little bit here. The key of the Maynard Smith and Parker paper was showing that in theory, at least, ownership could solve disputes even when it was not correlated with fighting ability. Even if it was a totally arbitrary asymmetry. In retrospect, I think I, too, jumped on this interpretation too quickly. Subsequent work has shown that an arbitrary asymmetry is unlikely to apply in most cases in nature. The problem is that in theory, whereas ‘owner wins’ can be an arbitrary asymmetry, so can ‘intruder wins’. This then produces the crazy situation where you get a cycle of nobody ending up with the territory because intruders always displace owners. And that should immediately set alarm bells ringing and make you think that there must be something very simplistic about this model.
In fact in most cases, ownership is going to be correlated with some real ability, even something as simple as an ability to find a territory quicker than others, or some real difference in fighting ability. Subsequent studies have shown that where territories are scarcer and not everybody gets a turn at ownership, then the very brief spiral flights which I observed, can escalate into longer contests in which acrobatic ability might be being demonstrated (Wickman and Wicklund 1983, Animal Behaviour, 31, 1206-16). Maybe acrobatic ability is a good way of settling contests, because when a female comes along, the most acrobatic male might be the first to intercept her. So if a male can detect that another male is better in flight than he is, the best thing to do might be to retreat. And yet, in my study, there was no doubt that these contests were settled very quickly within a few seconds. These spiral flights were literally brief contact, where the intruder almost seemed apologetic and would retreat almost immediately. I still think the reason the intruder retreated so quickly in my study was that basically it wasn’t worth having a big fight for a resource which was going to become vacant soon anyway.
Actually, in my Discussion, there was another idea which got lost, which I still like very much. In my pied wagtails, the spatial knowledge of a territory was likely to be very important. Any bird that could get to know this little place well would be more likely to value it and fight hard in a contest. These butterfly sun spots were different. As the sun moved across the sky, the territory moved across the woodland floor. Spatial knowledge could never really build up because your territory was always on the move. In the last paragraph of the Discussion I suggested that this may be why spatial cues or getting to know a territory’s characteristics was not a viable option in these butterflies, and they instead used a temporal cue for settling contests.
HS: So one hypothesis that emerges from this is that intruders that spend a longer time on the territory will be less likely to be kicked out by the resident who has left?
ND: Yes, they should be harder to get rid of. If we go back to my initial experiment, what I did was to remove an owner, wait for an intruder to settle on the territory, and then release the original owner back to ask will he regain his territory? To my amazement – this completely surprised me – in every case the newcomer won the contest, even though the original owner may have been in residence for several hours and the newcomer had only been there for a few minutes. That was a shock to me. I still had pied wagtails in my mind, and therefore I expected that if the intruder had been there for a short time, the owner would quickly chase him off. And the longer the intruder had been allowed on the territory, the more difficult it would be to get rid of him. But I found even a minute or two of occupancy was enough to cause this role reversal. That was a complete surprise to me, and that’s when I went down this theoretical “Resident wins” explanation for settling the contest. This is an example of where you do an experiment and the result completely shocks you. You think: my goodness, what on earth is going on?
But of course, the explanation for why the original owner lost, could be that this is just an experimental artefact. The period when I kept him in the net might have reduced his ability to defend his territory. I tried my best to control for that by re-releasing owners back onto an empty territory, and I discovered they all quickly regained their territory. Some subsequent studies suggested that the reason you got these role reversals is because the hotter male always wins. Spending time in a territory enabled you to warm up. When I removed a male, even a brief period in the net might have caused him to cool down and this temperature asymmetry might have caused rapid role reversal. That’s a good idea and the reason I hadn’t considered it was because the males in the canopy were in warm sunshine too. So it seemed unlikely to me that there would be a temperature asymmetry. Subsequent studies have cast doubt on whether a temperature asymmetry could really be the reason and I suspect that ownership is usually linked to some real difference in fighting ability.
ND: Yes, after my butterfly study, John Krebs did a beautiful test of this with great tits ( 1982, Behavioral Ecology and Sociobiology, 11, 185-94). He removed an owner and gave intruders various lengths of time to stay on the territory before the original owner was released back. He discovered that the longer intruders had been in residence, the harder they were to get rid of. That’s the result I initially expected with my butterflies.
HS: You say that it is possible that there are other traits that are different between territorial and non-territorial males. Do you have an idea about what those traits might be?
ND: Subsequent studies have shown that, on average, patrolling and perching males differ in darkness and wing spots (Shreeve, T.G. 1987, Animal Behaviour, 35, 682-90; Van Dyck, H. et al 1997, Animal Behaviour, 53, 39-51). So phenotypic variation might reflect adaptation to different mate-searching behaviour. Nevertheless, my removal experiments did show that at least some of those patrolling males were keen to get a territory because when you created a vacancy they would take over the opportunity. That’s why that initial experiment, though very simple, was so crucial.
HS: I wanted to ask you about the impact of this paper on your subsequent research. From your publication profile it looks like you published only one other paper on speckled wood butterflies. Is that correct?
ND: Yes, my studies throughout my career have mainly been on bird behaviour – on territorial systems in pied wagtails, on mating systems in dunnocks and on cuckoo-host interactions. So this was very much a side project.
HS: Do you think doing this study had any impact on your career and the course that your subsequent research took?
ND: Well, that’s hard to tell. Every publication you do when you are young adds to your short publication list, so is likely to count when you are applying for jobs. But I think, looking back at my career, the papers I’m most proud of are the work on mating systems in dunnocks, using DNA fingerprinting to unravel sexual conflicts, and those on cuckoo-host interactions, which I have been studying for the last 30 years. But my style of research is still evident in this early butterfly paper: watching, wondering why and then trying to do experiments to tease apart what is going on. That experimental approach is something I have specially enjoyed.
HS: This was one of the first tests of Maynard Smith & Parker’s idea. Following your study, did this idea get more empirical attention?
ND: Around this time, people were becoming very interested in game theory in general and how that influences animal behaviour. I don’t think anybody thinks that arbitrary, uncorrelated asymmetries are important now. I think asymmetries for settling contests are likely to be tied to real differences in fighting ability. And as I have admitted, I think I was too quick to jump on arbitrary asymmetry as an explanation for what’s going on in my butterflies. I think what was going on here is just: territories were abundant, contests were short, and when territories become scarcer then contests escalate.
I think the main lesson for naturalists from game theory has been to show how we should expect variability in the natural world. When I was a student, it was often assumed that each species would have one best way to behave. Game theory showed us that we should expect variability. So when in nature you see different animals within a population doing different things, that’s not just noise; that’s exactly what you should expect from competition. This expectation for variability was a whole new world, which I think we hadn’t really suspected. And this is exactly what I found in the dunnocks, of course, tremendous variability in mating systems as a result of sexual conflict. So for naturalists, I think, game theory opened up this whole new way of looking at variation in nature, and that was the most exciting thing.
HS: Hypothetically, if you were to do this experiment today, would you do it differently?
ND: Yes, I would film the spiral flights, to see exactly what was going on. One of the experiments I was most proud of was trying to introduce a second male onto a territory which was already occupied. This was incredibly difficult to do because, usually, as I was releasing the second butterfly, the owner would quickly spot it and chase it off. If I remember right, in only five cases was I successful in quietly introducing an intruder onto the same territory. Then, when the two males eventually met, there was an extraordinarily escalated fight. It was really exciting to see that! If I was doing this study again, I would film that escalated fight to see what was going on. And I would pay much more attention to who came back and occupied the territory. I didn’t record that, to my embarrassment now. I was so thrilled by this escalated fight that I just let them waltz off into the skies out of view and forgot them! It was rather like watching two people going simultaneously for the same parking space and having an argument. The crucial asymmetry of ‘owner-intruder’ was no longer available, so you got this escalated contest. Theoreticians got very excited about that.
HS: This paper has been cited over 500 times. Do you have a sense of what it mostly gets cited for?
ND: I’m pleased other people have cited it, but no, I don’t know.
HS: Did you go back to Wytham Woods after this study? When was the last time you visited this place?
ND: Well, I got married in this wood three years later. We had our party among the speckled wood butterflies! After settling in a new territory in Cambridge since 1979, I have been back to Wytham Woods just for visits, but have not done any more studies there. As far as I know, people haven’t studied speckled woods there. They should! It would be very interesting to see if, in summers where territories are in shorter supply, the contest behaviour is different I hope someone will do that and film the butterflies to see what is going on.
HS: But there has been other work on speckled wood butterflies in other places right?
ND: Yes, particularly in Sweden by Per Wickman and Christer Wiklund, absolutely beautiful work, and way more sophisticated than mine. But I hope that my simple study encouraged experimental studies of butterfly territorial behaviour. If my study is credited just for that, I will be very pleased.
HS: You said you were reading this paper for the first time in 40 years. Do you think your style of writing has changed since that time?
ND: I think it’s the same, but editors these days might not let me get away with such a chatty style. But I enjoy reading papers where you feel you are out there with the researcher, watching the animal and eavesdropping on the study. These days, of course, statistical analysis and quantification and use of computers has changed the way we do research for the better. There is absolutely no question about that. But I do think that people quite often get beguiled by numbers and statistics and lose track of the natural history, which I think should always be the starting point. I think observation from natural history is the most interesting thing and if you then can combine that with statistics and modelling then that’s great. But complicated statistics and numbers for the sake of it just makes the whole study dull.
HS: I really like the style of the paper. A couple of lines that really stayed with me. You say “Rather they are a short conventional display where, to put it anthropomorphically, the owner says `I was first here’ and the intruder says `Sorry, I didn’t know there was anyone occupying the sunspot, I’ll retreat back to the canopy’”. Just these lines, for me, summarise the main finding of this paper so well.
ND: Well, I’m all for anthropomorphism. When I was a student, we were told that anthropomorphism was a sin. And that’s quite right, in the sense you shouldn’t put human emotions onto animals. There’s the famous example of monkeys which were sent up in space capsules and when they came back grinning people said look how happy these monkeys are to have been in space. But these were fear grins! This is a wonderful example where transposing human emotions, or reading animals as having human emotions, can take you completely down the wrong avenue.
But imagining what you might do if you are a butterfly or a dung fly or even a bacterium, in the sense of weighing up costs and benefits and choosing the best option, that’s a very profitable way of thinking, because all you are doing is simulating in your brain exactly what natural selection is doing, namely choosing the best genetic option across the generations. You are simulating that in behavioural time in your brain. And I think that can be a very useful guide to thinking about the choices natural selection faces or thinking about what sorts of experiments you might do to tease out what is the best option. So I’m all for anthropomorphism, provided it’s done knowing its pitfalls.
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?
ND: I’d say: Come with me into the woods, watch these butterflies displaying, think of the sorts of questions you would ask, think critically about whether the experiments I did provided convincing evidence for the “Resident wins” idea and then think critically about whether the interpretation I gave from the current theory at the time would be a useful interpretation today.