In 1992, Dolph Schluter and Don McPhail published a paper in The American Naturalist in which they provide evidence for ecological character displacement among species of stickleback fish that live in the lakes of coastal British Columbia. In the paper, Schluter and McPhail also provide a conceptual framework in the form of six criteria that need to be satisfied to demonstrate the occurrence of character displacement. Twenty-four years after the paper was published, I spoke to Dolph Schluter about his motivation to do this study, the collaboration with Don McPhail, and how his stickleback research programme has developed over the years since the publication of this initial paper.
Citation: Schluter, D., & McPhail, J. D. (1992). Ecological character displacement and speciation in sticklebacks. The American Naturalist, 140(1), 85-108.
Date of interview: 2nd August 2016 (on Skype)
Hari Sridhar: The first thing I want to talk about is your motivation for doing this particular study. From looking at your publication profile, I learnt that you did your PhD on the Galapagos finches, and this study seems have come immediately after that. Was this work that you did in your first post-doc?
Dolph Schluter: It began during my second post-doc. After defending my thesis, I did some work on the evolution of continental finch assemblages, both to allow comparison with the kinds of patterns that had evolved on the Galapagos, but also the Hawaiian finches, and also to compare the continents with each other. And I was hoping to be able to say something about differences that might be involved in assembling finch communities on mainland compared with Archipelagos. That’s what I started with, and I did some field work in Kenya and also in South Western United States. A former student of mine, Dick Repasky, did his PhD on finches and sparrows in communities in California. And I also explored some sites in South America. But as I was doing this work, I was also starting to think about how one would do experiments in a system like that of the Galapagos Islands. In that system, we have patterns that become evident because we have replication. Different islands have acquired, by luck or by resources, different assemblages of bird species and those species have evolved, adapting to the island characteristics, but also to each other. Much of my thesis work focused on trying to determine the extent to which interactions between species, especially competition between finch species for food, had contributed to the diversity of forms that are seen today on the Galapagos. But it was just not possible to do experiments on the Galapagos Islands. First, no one really wants to tamper with the Galapagos Islands, but also finches aren’t necessarily ideal. I had contemplated establishing an archipelago from old Aircraft carriers, dumping dirt on them and seeding them with -literally – seeds.Of course, that was just a crazy dream, but in the meantime I was looking for a system like that – where again one had an archipelago of isolated islands, where it might be possible to do experiments – really simple-minded experiments. One of the topics I had become interested in in the Galapagos was the phenomenon of character displacement. This was something Peter [Grant] had worked on earlier in his career, and we had written some papers on this topic in the Galapagos finches, finding comparative evidence that whether species beak sizes had become adjusted on islands according to which other species were also present. David Lack, for example, pointed out that the two smaller ground finches – the small-beaked ground finch and the medium-beaked ground finch – were different on islands where they occurred together, but where they occurred alone, their beak sizes were intermediate. I discovered something similar in these three-spined sticklebacks when I came to the University of British Columbia. They seemed to possess many of the features of the system that I was looking for. In this case the islands are actually small lakes surrounded by land and they were isolated from each other, but all had been colonised from the sea by three-spined sticklebacks. The phenotype that was the phenotype of the original colonist still exists today in the marine anadromous three-spined stickleback. So that first paper was sort of describing the patterns that were evident in these species pairs. Don McPhail was my co-author on that paper because the work really began when I had conversations with him and his students about these species pairs of three-spined stickleback that he’d discovered through his career. That is what I began to work on, and that was that first paper that described the patterns and present the evidence for character displacement, and also to tell something about the natural history of these forms, and the curious pattern where some lakes had two species and most lakes just one.
HS: When did you join the University of British Columbia?
DS: I came here originally in 1983 on a post-doc to work with Jamie Smith. He had done some work on the Galapagos and I did some work with him on song sparrows. He had been influenced by the Galapagos finch work and was initiating a long-term study of song sparrows on a small island in BC [British Columbia]. And I also did work in the American southwest and spend time in Tom Schoener‘s lab at the University of California at Davis. Then in 1985, I got through a five-year super post-doc here. That became a real job in 1989.
HS: Was Don McPhail a post-doc supervisor or more like a colleague?
DS: Yeah, colleague. He was a professor in my department, who had been working on stickleback for many years. And he had told me about this pattern as well, that suggested character displacement.
HS: What were your respective roles in this study? Did you do most of the work?
DS: I did. Don McPhail had done a series of papers, each of which was basically describing one of the species pairs, describing what he had learnt about them, their natural history, some of their genetic differences and so on.He had also told me about this pattern that suggested character displacement, but had not thought of putting it all together in a comparative study of species pairs compared with single species. With his blessing, I went out into the field and made my own collections of these populations, sampled them in the breeding and non-breeding seasons, measured them, looked at their diet etc. So all of the results that are in that paper are based on data collected by myself, but based on a pattern that McPhail recognised but hadn’t yet himself quantified. So it seemed logical that the two of us should get together to write that paper.
HS: In the paper you say that most lakes contain only one species but coexisting pairs had recently been found. Were these pairs found by McPhail?
DS: They were. They were brought to his attention by various people. For example, in the government, they might have been doing fish surveys of lakes and they might have pulled some fish out of a lake, and gone: Woah, this is curious! These sticklebacks look unusually variable. And so they send them to McPhail, who was already then known as the stickleback guy or the fishes of British Columbia guy. Then he went back to those lakes and collected them himself.
HS: How long did the study take? Somewhere, you mention that some of the sampling was done in 1988. Was all the sampling done that year?
DS: Yeah, the data are primarily or entirely from collections made in a single year, earlier in the year in the breeding season and later in the year in the non-breeding season.
HS: You say that external traits were measured on the left side of the fish. Can you tell us why on the left side?
DS: That’s a good question. I don’t remember my precise reasoning for that.. it may have been simply because that’s how McPhail did it.
HS: How long did the writing of this paper take? Did you do most of the writing?
DS: Yes I did. That is not something I remember very well, how long it actually took me. I don’t believe it took me a long time to write, maybe a couple of months, something like that. I can’t remember the details.
HS: Did you and McPhail discuss drafts in person or over the phone?
DS: That part I do remember, and what I remember was that I never actually got my comments back from McPhail on the paper. He was very inspirational in the development of the work, in the early stages, but we didn’t work very closely on that paper.
HS: Did you have a writing routine – a particular place and time when you wrote?
DS: No, I don’t think I had a particular time. I probably wrote most of it sitting at my dining table at home. These days, I’m a little older and I find that the morning is the best time for writing. But in those days I don’t think it mattered very much. I just wrote any time.
HS: How were the figures in this paper made?
DS: I think I used the software Systat in those days, and that’s because it was one of the first packages that were available that would produce graphs.
HS: I want to go over the list of people you acknowledge, to get a sense of how you knew them and how they helped. Could we do that?
HS: T. Gullison?
DS: Ted Gullison was his name, and he was an undergraduate here at UBC, and became my first field assistant. He had previously worked with me in Africa, when I was working on birds in Kenya. And he stuck with me and we went out to the field together, and I made a lot of the collections with Ted.
HS: Do you know if he continued in research after this?
DS: He ended up doing a PhD with Steve Hubbell on forest dynamics in Bolivia. And then he had a job for a while at a university in the UK. But now he is not an academic; he runs his own business, consulting.
HS: G. Haas and G. Ross?
DS: Gordon Haas was also an undergraduate.. or he might have been a master’s student at the University of British Columbia. His role was not large, and I can’t remember exactly what it was.
HS: You say help with collection of data..
DS: Maybe he came out with me in the field or maybe…no wait I think his role was… I collected the species pairs, but I used some collections that McPhail had also made.. it’s coming back to me. And Gordon Haas had helped him make some of those collections. I was wrong when I said that all of the specimens came from me. Some of those single species populations were collected by McPhail.
HS: G. Ross?
DS: I don’t know who G. Ross is. He might have been another individual who helped McPhail with some of the collections.
HS: You thank H. Diggon and N. Scott for “permitting access to lakes”.
DS: Yes, Harold Diggon worked with Texada mines. I’m not sure exactly what the name of the company was then, but he was our contact person at the mine. Texada Island is basically all limestone, and it’s mined there. So Harold Diggon was the contact. The Paxton lake itself is on private land, and so to get access to the lake we needed their help. Also, Enos lake is on private land, and Mr. Scott facilitated access to that lake.
HS: And then you thank a number of people for their comments on the article. First, M. Bell?
DS: Michael Bell
HS: And you thank him again later on for careful readings of the manuscript.
DS: He was a reviewer of the paper at the American Naturalist. He wrote a long and helpful review.
HS: C. Benkman?
DS: Craig Benkman is a former post-doc of mine. He works on finches; on crossbills.
HS: J. Coyne?
DS: Jerry Coyne was the editor on the paper.
HS: S. Foster?
DS: Susan Foster is another stickleback researcher.
HS: P. Grant was, of course, your former supervisor.
DS: I sent him the paper for his comments.
HS: T. Hatfield?
DS: Todd Hatfield was my first graduate student working on stickleback.
DS: Tara Law was an undergraduate working in the field
HS: J. Losos
DS: Jonathan Losos. He works on anole lizards in the Caribbean, and I sent the paper to him for comments.
HS: A. McCune
DS: Amy McCune was another reviewer of the paper.
HS: You say your “work was supported by the Natural Sciences and Engineering Research Council (Canada).” Was this specifically for this study?
DS: I had an NSERC research grant. The way NSERC research grants work is they are sort of given for all of the work carried out by a supervisor of a lab. In the early days of the project, I believe that I started work on an NSERC grant that was written mostly about finches – mainland and continental finch assemblages. There is no requirement at NSERC that you spend research funds on precisely what you said you would in your grant application. That was useful because I was able to obtain some preliminary results on sticklebacks, before I wrote my next NSERC grant, which was then to do work on stickleback.
HS: Did this paper have a smooth ride through peer-review?
DS: It did, as far as I can recall, yeah.
HS: Was American Naturalist the first place you submitted it to?
DS: It was.
HS: Do you remember if it changed much from the first submission to the final published version?
DS: I don’t remember major changes taking place. I think Mike Bell suggested some additional analyses that influenced how we analysed the data, especially the body size correction.
HS: Do you remember how this paper was received, in academia and the media, when it was published?
DS: I don’t remember any news in the popular press. But I think it received a reasonable amount of interest in academia, people started citing the paper and so on. The thing that I remember most about its reception is that I received a hand-written letter from Ernst Mayr about the paper, which I still have and treasure. He was curious, in particular, because I had talked about our thinking on the origins of these species pairs and whether they had resulted from a double invasion process, or whether they had evolved in sympatry. Sympatric speciation was, of course, a topic of great interest to Ernst Mayr, and it was not something that he believed ever happened. So I think he was pleased by our double invasion interpretation.
HS: Did this paper have a big impact on your career?
DS: It did. It sort of launched my research of the three-spined stickleback system and that work is still ongoing today. Also, in those days, the American Naturalist was considered the ultimate place to publish a paper. I can remember, when I was a graduate student, it was the journal that all of us students talked about. That’s because it was a very influential journal.It was often where a lot of the very fundamental concepts in our field had first been published. So I was pleased to have a paper in the American Naturalist, and to have it appear there meant that many people would see it and read it. And also that it would be a good thing when applying for grants and so on, to have had a paper in the American Naturalist.
HS: This paper has been cited over 700 times. Do you know what it mostly gets cited for?
DS: I think it’s mostly cited for the character displacement part. It doesn’t make very much progress on the speciation question- how these forms actually originated. It introduced the system and described what we knew about the geology, and their possible originsHowever, it did propose that the process of competition and character displacement might be one factor contributing to, and driving the evolution of, reproductive isolation between the sympatric stickleback. There is still a great deal of interest in the connection between biotic interaction and speciation, and this was an example pointing to the possible role of resource competition.
But the most important part of the paper, the part of the paper it’s remembered for is, maybe, two-fold: One is the example of character displacement in the stickleback, because it’s a reasonably careful description. It tries to cover a lot of the evidence that is needed really to make a case, to at least make it a strong candidate. And the other reason that the paper is cited is that we also wrote down, based on what previous people had written, six criteria that ought to be investigated before one can claim that this is a convincing case of character displacement. In the time that this paper appeared, community ecology was in a great ferment and there was considerable discussion and criticism in the literature about the role of competitive interactions between species in the evolution of their differences. I first became aware of, or became embroiled in, this controversy because some of it focussed on the Galapagos finches. The most forceful arguments were made by people like Dan Simberloff and Donald Strong. They had analysed data, especially on body sizes, of various kinds of organisms, in an effort to determine whether the magnitudes of differences between species we see when we survey ecological communities, whether those differences are greater or less than you would expect simply by chance. So most of their argument concerned Criterion No. 1., which was that if you want to present evidence that competitive interactions were important, you ought to be able to at least, at the very least, rule out a null model in which you are able to say what the differences would have looked like if competition had not been involved, but simply the differences had been produced “by chance”, i.e. everything else that might influence beak size except competition. So, that’s the sort of environment in which this paper was published. I was very aware of the literature and how weak many of the examples up to that point really were. I had worked hard on our previous example on Galapagos finches, and then the importance of this example was that we had amassed a reasonable amount of evidence that made it at least a candidate for ecological character displacement. We also sort of spelt out a list of criteria that others have used subsequently. And then we also scored the stickleback on how well those criteria had been met to date, and said something about what else needed to be done.
HS: I wanted to talk a little more about these six criteria you list in the paper. Even today, are these the criteria used to investigate character displacement?
DS: Pretty much, yes. Particularly for studies of character displacement that are based, not on experiments but, on observational evidence. Studies of patterns in nature.
HS: In the paper you say that you find evidence for five out of the six criteria. Subsequent to this paper, did you also find evidence for the sixth, i.e. independent evidence that similar phenotypes actually compete?
DS: Right. Upto that point we had no direct evidence of competition between stickleback species. One of the reasons why this paper was important for my research program was that it presented the patterns,but then also motivated the experimental studies that followed. Studies that would test whether or not stickleback species competed more strongly when their morphological differences were less. The subsequent experimental work tried not only to test the idea that similar phenotypes actually compete for food, but also then to go beyond these criteria and test whether competition for food actually generated natural selection favouring divergence.
HS: Could you summarise what you have found in this regard?
DS: Yes.Two kinds of experiments took us beyond this paper. One of them was evidence that during the process of character displacement not only did competition occur, but it was weaker after divergence than before. That is an expectation from mathematical models of ecological character displacement, and it’s also fairly intuitive. As these forms become more different, competitive interactions would become reduced. The reason we were able to test that idea is because the observational data suggested how these pairs of species actually originated, or at least which phenotypes came before and which phenotypes came after. There are few lakes that have pairs of species; most lakes have just one. We focussed mostly on lakes that are similar in species composition to the lakes that contain pairs of stickleback species, except for differences in the numbers of stickleback species. Lakes that contain pairs of stickleback are all relatively similar in elevation – relatively low elevation – they are all relatively similar in size – all fairly small lakes – and they contain stickleback, cutthroat trout and no other fish species. Most other lakes in BC also have other species such as Prickly Sculpin, and they may have salmon in them, and various other fishes. We tried to keep things simple by focusing only on those lakes that are otherwise very similar to the lakes that contain species pairs. We think that lakes today that contain a single species of stickleback that is morphologically intermediate between the sympatric species, the limnetic and the benthic that we see in two-species lakes –and the marine form represent the first stage of the process that produced two species from one. That’s our working model, and it continues to be our working model. We imagine that stickleback species pairs formed when the marine colonised freshwater a second time, and it would have then been in sympatry in a small lake with another form that was at that time morphologically intermediate. Under the hypothesis of character displacement, competitive interactions caused a displacement of the intermediate form towards, what is today, the benthic species. And the second invader – the marine form of the second invasion – would not have evolved toward the centre, but instead would have evolved into what is today the limnetic species. When we set up our first competition experiment, we had two treatments.One of them was the marine form with an intermediate population as the target. The other treatment was the marine form against the modern benthic species from one of the lakes. In the first case the two forms were only about half as different as the present day limnetic and benthic species.In the second case, by using the benthic species as the competitor, we were able to test whether competitive interactions between them had diminished.The focus of that experiment was actually the marine form- we measured the extent to which competition experienced by the marine form had changed in the transition from an intermediate form to a benthic form. That tested CriterionNo. 6, but then also tested another feature of ecological character displacement, which is that competition actually goes down with time as the forms evolve greater differences.
The next series of experiments (1, 2) focussed on the question I was really interested in, and the one that really got me searching for a system in which one could do experiments, and that was to ask whether competition causes natural selection, and whether divergence is favoured. To do that we used an intermediate form once again. In the first version of the experiment, we used an intermediate form in the control and in the experimental treatment we used the intermediate form plus an added competitor, in this case the marine form. Once again, one of the treatments in this experiment attempted to recreate, as best one can with modern day materials, to recreate the forms that were believed to have been present when the marine form colonised the freshwater a second time. Then we asked whether the individuals within the intermediate form were differentially affected by the addition of the marine competitor. In other words, did the average growth decrease and were those phenotypes most similar to the added competitor differentially impacted, with the result that those individuals with the highest fitness were those that were farthest away from the added competitor, phenotypically. So we’ve done a series of experiments like that and each of them has demonstrated that competition generates selection, and specifically it generates divergent selection in this system. The phenotypes that do best are always the ones most different from the added competitor.
HS: Do you have a good phylogeny of these species now?
DS: No, we don’t. Why not you might ask? We should have one fairly soon, because now we have complete genome sequences of all these populations.But the picture emerging from stickleback is one that other researchers have seen in other systems, like Heliconius butterflies for example,which is that different genes have different phylogenetic histories. We think that a lot of the genes that are different, that distinguish limnetic and benthic sticklebacks in sympatry, are the same genes in different species pairs. Not only that, but that many of these genes are present in other freshwater stickleback populations that are out there. So, a lot of the genetic differences, we think – this is still being analysed – that differentiate limnetics and benthics, are older than the lakes themselves. And that many of these genes are geographically widespread in the Pacific and occur in other freshwater populations. Not all of them, because there are no true bona fide benthics in these other populations, that the benthic species in these species pair lakes just contained more of those benthic-like alleles, which make benthic forms so different from the limnetic and also from the marine species. So there’s a lot of genetic variation, and that genetic variation came to the lake when the lakes were originally colonised.What natural selection has done is really just to sort the variation that was available at the time. And that’s how things could happen so fast. That’s our growing understanding.
You asked about phylogeny. The mitochondrial DNA, for example, suggests that the stickleback in each of these lakes colonised independently from the sea, and that after double colonisation, mitochondria moved between the limnetics and the benthics in sympatry. That’s one history. The other history is, genes that got into the lake are actually the same genes as occur in other lakes, and natural selection resulting from adaptation in a similar environment, and to similar competitive interactions, caused the fixation of alleles over and over again in multiple lakes. A phylogeny based on those genes alone would be completely different from one suggested by the mitochondrial DNA. And also suggested by neutral regions of the genome. So what does phylogeny mean in that case?
HS: You mention this other species – the cutthroat trout – and say it is a predator of the stickleback. Is it an important player in stickleback evolution? Have you studied its role?
DS: We don’t have a full understanding of the role of cutthroat trout. One of the curious patterns that we see in the species pair lakes is that, whenever a pair of stickleback species is present, one of them has more bony defensive armor than the other. That includes bony lateral plates and spines – dorsal spines and pelvic spines. That’s evolved over and over again. Not only that – if you go to the single species population – lakes that are otherwise similar that contain only a single stickleback form, but also contain cutthroat trout – we see that the total amount of armor possessed is about the same as the amount of armor that limnetics possess in the species pairs. So the sympatric forms are divergent in armor. Not only that, they are more divergent in armor than randomly paired allopatric populations. And based on what we see in single species populations it is the benthic that has lost armor, compared to the limnetic form. We think that results because, when pairs of species are present, their vertebrate predators – vertebrate predation is thought to be the major selective force favouring more armor, more plates and spines – focus their predation on the limnetic. We think that vertebrate predation would be caused by cutthroat trout which is native but also by diving birds. Loons breed on these lakes every year and they eat many stickleback. We think that the limnetic form experiences higher predation and that is the explanation for why they have more armor than the benthic form. There is also another hypothesis.The benthic form might actually be under countervailing selection favouring reduced armor because of a greater exposure to insect predators.
HS: Insect predators?
DS: There are insects in the littoral zone of these lakes where benthics spend most of their time, dragonfly larvae and backswimmers that prey on sticklebacks, but we are not sure how important a selective force they are. They tend to prey only on the youngest stickleback, so it’s not clear how important that is. Whereas the birds prey on the old stickleback, the larger stickleback. We are not exactly sure why insects would select for reduced armor. One possibility is that it gives them something to hold onto, but there is very little evidence for that. The other possibility is that it’s simply costly to produce all that armor, and that as a result it slows the growth of stickleback and keeps them vulnerable to insect predators for longer. Some sort of an indirect cost. So we are not exactly sure why benthics have reduced armor, but it is repeatable, it happens in every species pair. And in two of the species pairs- the pair in Paxton lake and one other – the reduction is almost complete. There is very little bony armor, at all, in the benthic species. They’ve lost the first dorsal spine, and they’ve lost the entire pelvic girdle. It is still polymorphic, i.e. those traits are still present at low frequency in the benthic forms. We think that they are sustained there by hybridisation between the two species. The benthics continue to acquire armor genes from the limnetic species, but they continue to be selected against in the benthic. So, what’s the explanation for that? Well, we think its greater vertebrate predation on the limnetics, and so we have been testing this. A student of mine, Diana Rennison, has just completed an experiment that shows that the presence of cutthroat trout actually does select for greater armor, compared with control treatments where cutthroat trout are absent. The reason we think trout play a role in the divergence- there is a very simple reason – is that it is just that it is a downstream consequence of the competitive interactions that cause ecological character displacement. Ecological character displacement is thought to be the process that caused these sympatric forms – limnetic and benthic – to become so different in their diet and habitat use, but the secondary consequence of that divergence is that one of the species is more subject to vertebrate predation than the other, which might be more subject to insect predation. As a result, a second interaction now – predation – is coming into play that has exaggerated the divergence between the limnetics and benthics still further.
There is another possible explanation or contributing factor that we haven’t been able to test yet. The presence of a limnetic form out in the open waters, again, thought to be in part the result of ecological character displacement. The presence of that form actually helps benthics, reducing the amount of vertebrate predation and causing a reduction in body armour – costly body armour. So the first possibility I gave you is that predation differences are just an indirect consequence of habitat and diet divergence that competition has brought about. But the second one is a little more interesting because it suggests that there might actually be another kind of interaction between limnetics and benthics, besides competition. It might be kind of a commensalism, where benthics benefit from the presence of limnetics, in terms of the selection caused by predators.
HS: In the Discussion you say “In all, these results satisfy four of the six commonly listed criteria for ecological character displacement. They therefore support the hypothesis over several alter- native explanations (i.e., chance, nongenetic morphological shifts, biased extinctions, and absence of a link to resource use). What other mechanisms might account for the pattern?” You talk about differences in resource availability between the one and two species lakes.
DS: Oh yeah, that’s Criterion No. 5
HS: And you say one way to control for this might be to include a comparison of adaptive surfaces, or to do multiple regressions, or to do experimental manipulations. Subsequent to this study, did you do any or all of these?
DS: No. We haven’t, I think, got enough information really to estimate adaptive surfaces. That’s something I did as part of my PhD work on the Galapagos finches based on seed distributions and the relationships between traits and foraging success,especially beak size and foraging success. We were able to, sort of, estimate the relationship between fitness and beak size on Galapagos Islands. I haven’t really convinced myself that I have the ability to do that yet with the stickleback. That’s partly because of the spatial separation of these food resources, which makes it a little bit difficult to calculate sort of the consequences of alternative phenotypes and foraging success in the alternatives. So, we haven’t succeeded in doing that. There has been some work done. There was a paper a few years ago where a student in one of my colleagues’ labs went to a number of these lakes and measured a whole pile of differences like pH, calcium, dissolved oxygen and dissolved organic carbon, a variety of features to ask whether there is a systematic difference between lakes containing one species and lakes that contain two. And on that basis, we can say that there is no obvious difference between one and two species lakes, in their turf growth characteristics, characteristics of the chemistry of the water, the nutrients and so on. Nobody has identified really anything that makes the two species lakes different from one species lakes. The geology of this region is very complicated, and there are strips of different kinds of bedrock underlying lakes in this part of the world. Two extremes might be lakes that are sitting on limestone, such as Paxton lake on Tuxeda Island, and those lakes tend to have high pH and high calcium and they are very productive. The other extreme might be lakes that are sitting on granite, which causes a lower pH, closer to neutral pH, and lower nutrients and reduced productivity. We have stickleback species pairs in both those kinds of lakes. So it is very difficult to identify any features of the lakes that separate one and two species ones. But of course ignorance is no defence! There is always the possibility that there is some crucial feature that we are not understanding. So we cannot rule that out. So, instead of trying to simply prove that lakes with one and two species are the same, our focus has been on carrying out experiments to test mechanisms that underlie ecological character displacement. We do those experiments in smaller bodies of water, in ponds, where we can ask is it true that competition generates selection, for example.
HS: You say “We note that within-population variation in morphology also has a strong heritable component (Hagen 1973; Bell 1984; Baumgartner 1986; Lavin and McPhail 1987; D. Schluter, unpublished data). For example, comparison of wild-caught limnetic parents from Enos Lake with their offspring raised in the lab yielded an estimate of h2 = 0.45 in size-adjusted gill raker length (D. Schluter, unpublished data). Traits within populations are therefore expected to respond quickly to changing selection pressures.” Was this data published in a separate paper later?
DS: It was published in 1996. The title of that paper was “Adaptive radiation along genetic lines of least resistance”. That study had somewhat different goals but it includes an estimate of genetic variation in one stickleback population.
HS: In another place in the discussion you say “In a lake largely devoid of other fish species, an increment of divergence along the benthos-water column habitat gradient may yield the greatest decline in interspecific competition together with the least compromise in resource availability. The question merits more careful documentation and study because it may stand as one of the few instances in which early patterns of adaptive radiation are predictable.” Did you research this further subsequently?
DS: I think what I was thinking about, at that time, was just that this gradient, from more zooplanktivorous at one extreme to more benthic at the other extreme, is one that happens over and over again. That those two environments really do represent two main resources available to fish in lakes, in general. We see limnetic – benthic divergence in many other fishes as well. McPhail and I have another paper in which we talk about this.That paper is mostly just a documentation of how frequent we see divergence along a benthic -limnetic axis in post-glacial fishes. The idea is that stickleback is actually just one example of a very general phenomenon happening in low diversity lakes, where a similar split is often seen. That paper was just a short one in 1993 called “Character displacement and replicate adaptive radiation”.
Through discussions with McPhail, I realised that he had this immense knowledge of all these other cases of divergence, and the origin of closely-related species pairs of fishes in lakes of previously glaciated regions. That sort of made me think that this is a very fundamental resource gradient. And it has a large influence on the evolution of fishes, and that’s particularly conspicuous in these lakes that contain very few fishes to begin with.
HS: You say “Reproductive isolation is incomplete even today: current estimates are that roughly 1% of adult fish are hybrids (McPhail 1992a), although we lack an estimate of gene flow.” What’s the status today?
DS: We haven’t published anything on this yet, but we have undertaken a genetic survey of two of the species pairs – Priest lake and Paxton lake – using modern sequencing methods, and I can now tell you that about 2% of individuals in lakes are F1 hybrid individuals.
HS: So it has increased?
DS: No, it was badly estimated the first time. Our initial estimate was really just based on morphological measurements. It’s very difficult to unambiguously or convincingly determine hybrid status. Now we have been able to identify, from a large sample from these two lakes, individuals that are F1 hybrids, individuals that are backcrosses, and even one, possibly two, F2 hybrids. So we now know how often, or how frequently, hybrids are produced. We are currently using this to try to estimate how strong selection is against hybrids, in particular which hybrid classes suffer the brunt of selection.
HS: Given the problem with the earlier estimate, does it mean that you don’t know how much this percentage has changed over the years?
DS: That’s correct. We don’t know how it’s changed over time. I can only tell you what it is currently. We do have evidence of gene flow in the past, just based on an observation I mentioned earlier, which is that there is evidence that limnetics and benthics share mitochondrial DNA haplotypes.We now know also, from the complete genomes, that putatively neutral regions of the genome are also shared in sympatry. We interpret that also as a legacy of past gene flow. I think there are methods based on genome scale data that would allow you to estimate how much gene flow, not just how much currently, but how much gene flow has occurred over time. We haven’t yet applied these methods, but we would like to try to. These methods of course make assumptions that might be difficult to verify, but it might be insightful.
HS: In the last section of your paper you say “The two components of adaptive radiation are speciation and divergence. Results presented here suggest that competition for food played a hand in one of these processes (divergence). The sticklebacks present an opportunity to examine its potential role in speciation also.” Is the latter a line of enquiry you pursued in this paper?
DS: It is. The original reason I started working on the stickleback system – the species pairs and the single species populations – was because I saw it as a way to carry out experiments to test, more fully, theoretical predictions of ecological character displacement. To sort of go beyond the six criteria. But as I worked on these species more and more, I became very much interested in the speciation question. How could these new species form? Was natural selection also involved? How could it have occurred so rapidly? And so on. Some of that work has been done already, and we are continuing to pursue some of those questions using the genome sequences. One of the things that my group worked on, after the character displacement work, was an attempt to answer the question of whether natural selection had been involved in the origin of these species (1, 2, 3). I had already become interested in that question in the Galapagos finches.That was because it kind of blew my mind that it was possible to imagine how fluctuating environments, say on those Galapagos islands, may cause hybridisation between species to be low at some points in time and high at other points in time. I was impressed by the idea that environments could thereby determine the fates of species, whether they persisted indefinitely or whether they collapsed. The idea is that certain kinds of speciation can be reversible and that the persistence of species in sympatry may require the presence of ecological selection. So I started to investigate the extent of our understanding of how ecological environments and adaptation to those environments by populations was linked to the origin of species. This was a very Darwinian view of speciation, but it was not, at that time, the most popular view of speciation. Back in the day when I was a graduate student, the consensus seemed to be, although it certainly hadn’t convinced everyone, there were certainly critics of this – but there seemed to be widespread acceptance of the idea that random genetic drift played a large role, or some combination of drift and selection played a large role in the origin of species. The more I looked into it, the more I read, the more I realised that the role of natural selection, in particular, the role of adaptation to ecological environment, was almost completely unknown. These ideas were out there, there were experiments, for example in Drosophila, which suggested it was theoretically plausible. But just the very basic question that species might have originated by this process was very poorly understood. I might even say that despite the simplicity of the idea, an idea that’s been kicking around for a long time, we were then unable really to point to even two species in all of nature and say that they had formed in this way. I became interested in coming up with ways of testing whether ecological selection had been involved in the origins of stickleback species. There were a whole slew of tests that my students and I carried out to investigate that question (1, 2, 3, 4, 5, 6). The overall conclusion so far seems to be that adaptation to ecological environment is one of the most important aspects underlying the origins of these species.
HS: It is now 24 years since this paper was published. Would you say that the main conclusions are still more-or-less true?
DS: Yup, I think so. We’ve certainly filled in a lot more details and we’ve carried out experiments to test these aspects, and so far I’d say that the results still stand. There’s been no reason yet to revise any of the main conclusions of the paper. But at the same time, I still feel that the evidence for ecological character displacement in general is incomplete. Even the selection experiments that I carried out in the past, when we first began to work on the stickleback, the selection experiment where we asked does competition lead to divergent natural selection, have used surrogates for fitness. We have never actually truly measured fitness differences among phenotypes or genotypes in a target experimental population. That’s something we still aim to attempt. What’s also come up in this conversation is just the extent to which the role of predation from cutthroat trout and diving birds and insects played in the divergence of these forms. In the latest experiment we did (carried out by my students Diana Rennison and Seth Rudman and which is still unpublished) to address the role of predation, specifically cutthroat predation, we were able to look not only at whether the presence of cutthroat trout generates selection in a target population, but also show that this effect is transmitted between generations. We’ve been able to show that the presence of cutthroat trout affects allele frequencies of known genes that underlie armor differences between species. So, our abilities to look at evolutionary consequences of competitive interactions, or any interactions between species, is now much facilitated by all of the genetic resources available for stickleback. Our aim is to continue to pursue this question in this system – What are the roles of interactions between species? We’ve been looking at predation, but I think it is time to look at competition again as well, look more finely, at the regions of the genome that are affected by competition, and look at the fitness effects of competitive interactions between species – How strong are they? What genes are affected? Is this transmitted between generations? and so on. This is work that’s been impossible up to now.
HS: I’m guessing you continue to work in the lakes you sampled for this study. Have they changed in any way from the time you worked there for this study?
DS: They have. Certainly, this part of the world is warming, and the lakes have also warmed. We have not investigated the consequences of that. So we don’t actually know what effects of warming there have been up to now. But it’s an interesting question for the future. The more conspicuous changes that have taken place is that two of the lakes have suffered invasions from non-native species.In both cases, the result was the extinction of the species pair. When this study was conducted – Schluter and McPhail (1992) – we knew only four pairs of species, or at least pairs of species in only four drainages. One of the drainages has, sort of, a chain of lakes, but we treat them as the same pair. Since then a fifth has been discovered. Sorry, let me backtrack a bit here. Not sure if there were four or just three at that time. Did we have Hadley lake in this paper? I think so, I want to confirm. Right, Hadley lake was there. Enos, Hadley, Paxton and Priest/ Emily are the four drainages we knew about then. There’s now a fifth. Little Quarry lake on Nelson Island has been discovered. In the meantime, Enos lake and Hadley lake species pairs are gone. Enos lake was invaded by a crayfish that is not native. Enos lake is on Vancouver Island, and that crayfish is actually a native crayfish to this part of the world but did not occur on the islands. It was introduced into Vancouver Island and eventually made its way into Enos lake. What happened was that the two species began hybridising more and one of the species during this process – the limnetic species – became rarer. The two forms eventually collapsed into a single hybrid swarm that’s more benthic-like than intermediate. It’s a hybrid swarm, but it’s not a reversion back to the intermediate state. Because of some idiosyncrasies of the collapse the result is actually a higher proportion of benthic genes than limnetic genes, but it’s nevertheless a true hybrid swarm. Then on Hadley lake on Lasqueti Island, somebody introduced cat fish, as a result of which there are no stickleback in that lake at all anymore. As a result of those catastrophes, the species pairs are now federally listed endangered species.
HS: In all the lakes?
DS: Yes. All the lakes are now federally protected.
HS: What impact have advances in technology had on your work?
DS: Advances in technology have changed what we do. A number of years ago I started a collaboration with David Kingsley’s group at Stanford university, and he has been the driver behind the development of many stickleback genetic resources, including a really high quality reference stickleback genome and many genome re-sequences from various populations . I’m currently working with Felicity Jones, a former post-doc of his, on the complete genome sequences for limnetics and benthics. The development of RADseq methods have been really useful as well. This new work has provided us a number of things. First, the ability to identify who is a hybrid and who is not. The ability to identify some of the genes that distinguish species has enabled us to say that, for example, the variation has risen by new mutation or instead whether the differences that have evolved between limnetics and benthics were the result of genetic variation that pre-dates the origins of limnetics and benthics, pre-dates the lakes they occur in. The genetic data will give us a clearer picture of the phylogenies of different regions of the genome.The development of these genetic resources has also given us the ability to measure evolutionary changes, even across single generations in our experimental studies. One of the reasons motivating my collaboration, my desire to help David Kingsley and his group do what they do so well, was that I felt that the genetic resources that would result from that effort would allow us to address some of these old questions much better.
HS: What about other aspects – sampling diet, habitat use, and trapping of fish – do you use more-or-less the same methods today as you did then?
DS: Yes, that’s correct. The field methods have not changed at all.
HS: Have you ever read the paper after it was published?
DS: I’ve never read it start to finish again. Should I? But I have flipped back through it just to remind myself of things people ask me, e.g. well, what do you know about the diets of these things. And I’d go – Well, I think I did both breeding and non-breeding diets in this paper and the diet differences are much greater in the non-breeding season.Then I’d have to go back and see – Did I really show that in that paper or not? Or was it someplace else. So I have often gone back to refresh my memory about certain facts.
HS: Would you say your style of writing has changed from the time of this paper?
DS: That’s a good question. I don’t think so. When I wrote this paper, what was really important was getting the Introduction right, where all the ideas are laid down. And most of the work that we have done on stickleback – the initial work in this paper and the work that’s come subsequently – is sort of the realisation that this provides an amazing system to address really fundamental questions, and often it’s easy to state those fundamental questions in the Introduction. The reason for doing that is that it makes the paper more broad. It makes the ideas motivating the work really clear and spells them out in such a way that people working in other systems, besides stickleback, might find this work interesting, might find that they are actually interested by similar problems or indeed are working on similar problems. So that is the style that I used when I wrote this paper, and it’s the one that I continue to write with today and impress upon my own students.
Stickleback are really not just about stickleback. The things that are going on in these small lakes not very far from where I live are helpful in understanding really fundamental questions. And character displacement at the time that paper was written was controversial and was considered a really fundamental question. I still think that it is a fundamental question, particularly its role in the evolution of reproductive isolation, which we remain interested in and which I don’t think I have tested definitively (but plans are being made). But yes, the writing style I haven’t really changed over the years.
HS: Would you count this as one of your favourite papers?
DS: Oh definitely. It really launched the research program that I’m still working on today.
HS: What would you do say to a student who is about to read this paper today? What should he or she takeaway from this paper written 24 years ago?
DS: Well.. looking back.. there are some very specific things about this paper – like how does one go about testing, particularly when experiments aren’t possible yet, a general idea like character displacement. How does one find a system that allows one to address some of these questions?S o yes, one of the messages would be: find the right system for the right question. I started working on stickleback not because I found them compelling organisms in their own right – which of course they are – but because I was interested in these general ideas, and I needed to find a system where it would be possible to address them. A system that’s got very few species, that has replication of instances, of one species versus two species, (like Galapagos finches, which are slightly more than one versus two species, but still very low diversity communities) where it’s possible to discover important patterns. Then to use the patterns as a launching board for more detailed studies that attempt and test the questions more fully. It’s kind of different, in that sense maybe, from the way many people approach their work, or at least the first systems that they begin work upon, which is to actually hunt around for a while to find one that has the properties that one desires to make progress on a question. And, although it doesn’t come out in this paper, as I said earlier, I was looking for a system in which it would be possible to do experiments. We could take these stickleback into the lab, we could breed them in large numbers, we could hybridise them so we could generate experimental populations as mixtures of other populations to help us test some of these things. That might be the major takeaway from this paper, I guess.
Also what was really important for me in thinking about this system – because many people would ask me why I don’t work on Drosophila. Wouldn’t it be much easier to do the kinds of experiments that I do in Drosophila? My answer has always been: well, Drosophila experiments in a laboratory tells us what’s plausible. If you want to know what has actually happened in nature you need to work on a natural system. And that was really important for me at this stage in my career, that I find a natural system with interesting patterns where I could design comparative studies and eventually experiments to answer questions, so that we would know what actually had happened, not just what might happen.