Peter R. and B. Rosemary Grant lecture Darwin s Finches

Transcription

1 Darwin s Finches October 28, 2009 Peter R. and B. Rosemary Grant lecture presented as part of the Darwin Distinguished Lecture Series. These events are sponsored by Arizona State University, Office of the President, College of Liberal Arts and Sciences, School of Life Sciences, and the Center for Biology and Society. Transcript Quentin Wheeler: [0:00] Good evening, welcome. My name is Quentin Wheeler. I'm university vice president and Dean of the College of Liberal Arts and Sciences, and I'm delighted to welcome you here this evening. [0:10] This is the Darwin Distinguished Lecture Series, and it's part of a continuing celebration of the 150th anniversary of the publication of "The Origin of Species" that we've been having for, well, a little bit over the last year. And we've just had a fantastic series of events and tonight will be no exception. [0:29] Before I begin, I want to remind you that there will be a book signing afterwards at the back of the room. And if I could ask you to muzzle your cellphones for the benefit of everyone, I'd appreciate that. [0:42] Perhaps no group of species on earth is more immediately and universally associated with the theory of evolution than Darwin's finches. And certainly no scientists are more immediately or universally associated with Darwin's finches than the Grants. So this is a superb opportunity to have insight into the on-the-ground kind of studies that they have made a career of. [1:06] Peter and Rosemary Grant are professors emeriti at Princeton University, fellows of the National Academy of Sciences, the American Academy of Arts and Sciences and the Association for the Advancement of Science. [1:20] I had the pleasure of leading a tour in May of the Galapagos Islands, my first visit there, and it was a joint fund raising venture of an institute that I run called the International Institute for Species Exploration, and our Institute for Human Origins. And experiencing the Galapagos is just a life altering event, and we'll have the pleasure of seeing some images of that tonight. But I certainly recommend it to you. [1:44] And our vessel did do one lap around Daphne Major. And I was thinking of the fortitude of field biologists working for more than three decades on that island. This requires considerable love and devotion, I think, to one subject matter. [1:59] Beginning some 35 years ago, the Grants went to the Galapagos to study the birds that had so intrigued Darwin. They lived there part of each year, they banded, tracked and observed literally thousands of finches, and raised two daughters in the process. [2:14] Darwin, like most of us today, did not fully recognize the variability inherent in the finches of the Galapagos. These birds vary from vegetarians to vampires, and reveal adaptations of a range rarely matched in other groups that have been studied. Darwin Distinguished Lecture Series page 1 - ASU School of Life Sciences Grass Roots Studio

2 [2:31] During Darwin's brief visit to the Galapagos, he didn't appreciate what these birds might tell us about evolution, and regrettably his notes and his labeling of his specimens was not quite what it might have been. And upon returning to London, it became obvious as he worked with ornithologists just the full implication of what these birds might mean in terms of revealing something to us of their evolutionary history. [2:54] But it was the Grants who would pick up where Darwin left off and reveal the full potential of the Galapagos finches to tell us about evolution. Their powers of observation are legendary. The staff at the Darwin Research Institute in the Galapagos reported to have said that only God and Peter Grant can recognize Darwin's finches. [3:14] It was the Grants' careful observations that elevated the finches to the paradigmatic case study for natural selection that they are today. Meticulously measuring beak sizes and shapes, and tracing such traits through genetic lineages for literally generations, the Grants have elevated natural selection in the field laboratory from theory to incontrovertible fact. [3:37] The Grants and their finch studies formed the basis of a non-fiction book, "The Beak of the Finch: A Story of Evolution in Our Time" by Jonathan Weiner, which won the Pulitzer Prize in In his book, Weiner writes that the Grants are the first scientists equipped with enough patience, stubbornness, ground and sea support, computer power, airplane power and staying power to watch the process actually happen. [4:03] For their scientific contributions, Peter and Rosemary Grant have been recognized this year with a Kyoto Prize for lifetime achievement. They have received the Darwin-Wallace Prize, given only once every 50 years by the Linnean Society of London. They were awarded the American Institute of Biological Sciences Outstanding Scientist Award, and Balzan Prize for population biology in The Balzan Prize is reputed to be equivalent to the Nobel Prize, created to honor scientific fields not recognized by the Nobel committee. And both received the Senior Scientist Award of the Humboldt Foundation, the highest given to foreign scientists by the German government. [4:42] So please join me in extending a warm Darwinian welcome to Peter and Rosemary Grant as they share with us what they've learned about evolution through the study of Darwin's finches. [4:51] [applause] Peter Grant: [5:07] Well thank you very much for that over-the-top introduction. [laughs] It's been a great pleasure to come back to Phoenix after an absence of no less than 55 years. That time, we experienced 115 degree temperatures. I could not believe it at 3:00 one afternoon. And it's much more comfortable to be here at whatever it is, 62 degrees, today. [5:27] Anyway, it's a great pleasure to be here, and to tell you about what we know about evolution as a result of the studies that we and our associates have done on Darwin's finches. [5:37] Now the title of this book is simply a restatement of "On the Origin of Species." The fact that 150 years ago, there was a title to a book written by this man tells us that we are still confronted by the same series of questions that Darwin was 150 years ago, namely, how as modern evolutionary biologists do we explain the enormous biological diversity in the world? Darwin Distinguished Lecture Series page 2 - ASU School of Life Sciences Grass Roots Studio

3 [6:05] And I'm really referring to two aspects of diversity as illustrated by these Hawaiian honey creeper finches. One is the variety of them, they look very different. And the other one is the numbers. More than 50 species evolved in a relatively short space of time. [6:22] And evolutionary biologists like Rosemary and myself would like to be able to understand this diversity. How is it that so many species have evolved from a common ancestor in a relatively short space of time? [6:34] Well the core process of that generation of biological diversity is speciation, and I'm going to illustrate that with an oversimplified diagram of how speciation occurs. It is the process of splitting of a single population or lineage into two, which diverge, and after a period of time they are no longer capable of interbreeding, at which point they are unequivocally two species. [7:03] Darwin was familiar, or rather he understood the early stages of this process much better than the later ones. And I'm going to give a quote to show his understanding. Those cases he wrote in which a speciation splits into two or three or more new species, I should think near perfect separation would greatly aid in their specification, to coin a new word. [7:29] Well fortunately the word "specification" did not catch on. We now say "speciation" but the fundamental importance of geographical separation of derived populations has persisted to the modern day in the process of speciation. [7:45] How then does it end? To obtain a succinct statement of a general understanding by biologists of how that process ends, one has to move fast forward into the modern era of Mendelian genetics. [8:00] And here I'm going to show a quotation from Hermann Muller, a population geneticist who put it like this: "Thus, a long period of non-mixing of two groups is inevitably attended by the origination of actual immiscibility, that is to say genetic isolation." And he was referring specifically to random processes alone. Random genetic drift would do it, aided and abetted perhaps and probably by natural selection. [8:29] Now the theme of our talk is that there are many more interesting things going on evolutionarily between the starting point and the end point than is reflected in these two statements. And I make that assertion on the basis of our long-term study of Darwin's finches. On the Galapagos Islands there are 14 species in total, one on Cocos Island off the coast of Costa Rica derived from a communal ancestor in, again, a short space of time, two to three million years only. [9:00] None of the species have become extinct as far as we know as a result of human activity, and we can't say that about many groups of birds elsewhere in the world. And not only that, but there are some islands that are in as close to a natural or pristine state as it is possible to find. And therefore, whatever we can find out through out studies about the relationship between organisms, the finches, and the environment in contemporary time can be extrapolated -- without any qualification about human influence -- can be extrapolated backwards in time to the times when the evolutionary diversification took place. [9:40] So, here then, is an outline of our talk. I will first introduce the diversity of the species with an illustration of four only. Talk about speciation and the adaptive radiation. Give you evidence for evolution when the environment changes. And then Rosemary will address the Darwin Distinguished Lecture Series page 3 - ASU School of Life Sciences Grass Roots Studio

4 highly important critical question of what exactly constitutes the barrier to interbreeding between two species. Well, that barrier occasionally leaks, species hybridize, and she will tell you what we know about the causes of hybridization as well as the consequences. [10:16] So, here then, is a brief overview of diversity, starting with perhaps the most famous of all of the species, the woodpecker finch. The woodpecker finch uses a tool, a stick it fashions itself in order to obtain a cryptic larvae from the environment. But the beak itself can be considered a tool. And the beaks amongst the species vary enormously, and I'd like you to concentrate on the beaks, their sizes, and their shapes, less on the plumage. There's a little bit of variation in plumage color, hardly any unpatterned, but the beaks undoubtedly are the centerpiece of the radiation because they vary so much. Here's a water finch, little eight gram bird. Another bird bird, a large tree finch, has a completely different beak, and it functions in a very, very different way. And then finally, here's the beak of a cactus finch with a tool appropriate for probing cactus flowers. [11:11] Now, it's called an adaptive radiation because one can easily see a correspondence, or a match, between the nature of the beaks, the tools for dealing with foods, and the foods themselves. A few years ago, Dosch Luter and I tried to go a little bit further than just making that kind of matching correspondence. We asked, "Is it possible to be able to take information from what is available in the matter of food supply on each of the islands and so understand the way in which birds use the foods that we could predict, island by island, how many species there would be on that island and which particular beaks they would have?" [11:53] So it seems like a very difficult task. To be able to do that, you take the birds out of the system and use the system itself minus the birds to predict what the birds would be doing. We, never the less, managed to do this by developing an index of ecological opportunity on each of the islands, and it is an index composed of two relationships. The first one is the relationship between the size of seeds eaten by the graniferous species and their beak sizes. And unsurprisingly big beak birds can crack open bigger seeds than small beak birds. And then the second relationship is between the amount of food supply in the environment and the number, or biomass, of finches supported by that food. And here, the data are summarized by islands and you can see the unsurprising positive correlation between the two. [12:46] So putting these two together, we came up with an expected population density of finches of the graniferous sort in relation to beak sizes along that horizontal axis. Looking at what would be the expected density a small beak bird, a medium beak bird, and a large beak bird on these three islands. And the first thing that strikes you straight away is that the profiles of ecological opportunity, sometimes we refer to those by the jargon term of adaptive landscapes, the first thing that strikes you is that they're peaked. It's not a flat distribution, but there are one, two, even three peaks or more on an island. [13:31] So now the question is, where do these finches fit into these islands given the expected peak population densities, and the answer is very close alignment, as indicated by the symbols put underneath these peaks. There are three other features I would like to point out about this figure. First of all, there is only one species aligned with each peak. You don't see two species for any one peak. Secondly, there's no major peak that lacks a species. The ecological opportunity has been responded to in evolutionary time by the diversification of the finches. And the third thing is that the same peak on different islands may be occupied by different species. Darwin Distinguished Lecture Series page 4 - ASU School of Life Sciences Grass Roots Studio

5 And that you can see by the different symbols under the small peak, the peak associated with small beak size, on the middle and lower profiles. [14:25] So I think there's a very strong, controversial argument that in the course of the radiation multiple speciations have resulted in several species exploiting different aspects of the food environment. It's much more difficult though to answer the question, "How exactly did speciation take place?" [14:46] And so, now I'm going to move to that topic and give you a simple Darwinian cartoon-type diagram of the major views, not the sole views, but the major view about how speciation takes place. It involves, in Darwin's language, three steps: colonization of a new area through dispersal, adaptation to the new environment, and then finally, the coming together of two derived products of this process separated by a barrier to interbreeding. Now what is that barrier? How is it constituted? [15:22] Well in this little cartoon, I have shown this with an arbitrarily chosen example on the right-hand island of San Cristobal with a small figure of a finch and a large figure. So these are symbols for two populations that have come together. And now the barrier to breeding may have arisen solely as a result of adaptive change in allopatry, that is geographical isolation. And so, when they came together they were fully formed different species. [15:51] But more likely when they came together, there would be some adjustment. They wouldn't be so completely different that they did not recognize each other as potential mates. And then natural selection might be expected to work to the disadvantage of the intermediate birds because they might interbreed with each other with low fitness and then might compete with their parental species and suffer as a disadvantage because they are in the minority. That second process, that divergence as a result of competitive interactions, is known as character displacement, and I'm going to give you an example of that at the end of what I'm going to say. [16:30] Now Rosemary's going to talk about the reproductive interactions at this crucial stage of secondary contact. I'm going to start by giving you an example of the allopatric phase with a species that is on the northern islands, not on the southern islands, as we see them here. So, bear in mind on different islands, within the same species, populations are likely to differ according to the nature of the food supply. That's the theme that I'm now going to develop. [16:58] We start with this example of a sharp beaked ground finch -- that's the name of the species -- high up on the island of Pinta occupying a forest where there's a predominance of hard seeds, hard, cuticled insects, and mollusks - little snails - that this bird eats with a robust... it's sharp, pointed, but it is, nevertheless, a robust beak. [17:22] So, that's on that island, but on the low northeast island of Genovesa, there is no such food. Another population of the same species feeds in a different way. Here the seeds are much smaller, and they're softer, and the birds probe cactus flowers for nectar and pollen. Correspondingly, and here's the association between morphology and diet, correspondingly, these birds are small and with small beaks. [17:48] Now, on the island of Wolf in the northwest of the archipelago, another population of the same species has adapted to a different food supply. Here they have a rather robust beak, and they use that as a tool for breaking open the eggs of sea birds that are called boobies. Darwin Distinguished Lecture Series page 5 - ASU School of Life Sciences Grass Roots Studio

6 [18:07] If that doesn't work, if hammering away at this egg doesn't work, they use behavioral stratagems to achieve the same end. Can you imagine that egg rolling along, and then falling off a cliff, smashing, and thereafter, there is a free-for-all for the food. [18:24] Well, if that's not bizarre enough, on this same island, the same species does the following. It drinks blood, which it draws by pecking at the base of developing wing feathers of this hapless bird, called a booby, sprouting forth blood as a result, and the bird will then drink the blood. A habit we believe, and I think it's a reasonable supposition, that this probably derived from feeding on hippoboscid flies. [18:53] Now, hippoboscid flies are like mosquitoes. They fly up inside the feathers of sitting birds, boobies in particular, and suck the blood. So, they're blood-filled insects that finches probably fed on initially, and then since sometime later, and who knows how, they switched to feeding on the boobies themselves, and in so doing, they shortened the food chain. [19:18] So, you get the point. Different islands have different food supplies, different opportunities for exploiting those different foods, and the species have changed in beaks as a result of evolutionary processes in an adaptive way with a corresponding match between the nature of the tool, and the nature of the task to be completed by it. [19:39] That's the kind of argument that many of us evolutionary biologists make because we can't go any further. But, it would be much better if we could show that natural selection as a process producing the adaptive change that we only infer as a result of our studies to have occurred in history with this comparison that I've just made. [19:59] And as a process, natural selection has been demonstrated on this little island that Quentin Wheeler circled once and wondered how anybody could want to go there. This is the island of Daphne Major in the center of the archipelago, and people ask us, "What is it like to work on this island?" So, the first two slides before I get back to science, I'm going to show you brief little views of what it's like to be working on this island. [20:25] The first one shows the bedroom -- I was going to nearly say the bathroom [laughs] -- the bedroom, the dining room, and the laboratory where we do the work. And the second slide shows how we do it. We put up nets. We capture birds. We take them out. We measure them in various ways. Beak dimensions, we measure. Leg, wing, and we weigh them. [20:48] And then having banded them with a unique combination of color bands and a numbered metal band, we release them. We started in In 1988, we developed a technique for taking blood without the need of liquid nitrogen, and since that year, from 1988 onward, we've been bringing blood specimens, a single drop of blood, a minute amount, back to the United States for DNA analysis. But when we started in 1973, that technique hadn't been invented, and we measured and watched the birds to see what they did. [21:21] Now, one of the questions right at the beginning was this. The finches differ markedly in morphology. They've evolved divergently from common ancestors, so there must have been some genetic variation for that process to take place because the species must be genetically different. So, that implies any one population had genetic variation and could be added on by the forces of selection shifting the population one way or the other. Darwin Distinguished Lecture Series page 6 - ASU School of Life Sciences Grass Roots Studio

7 [21:51] How would we know about genetic variation governing traits like beak size, for example? Well, the obvious way is to look for an association between parents and their offspring because if there are traits that are inherited from parents by offspring, that should show up as a positive relationship between, in this case, beak size of the offspring and the parents. [22:15] So, we did a lot of work to establish the following relationship -- the average of the mother's and the father's in beak depth, in this case. The average beak depth of the mother's and the father's is a very good predictor of the average beak depth of the sons and the daughters. [22:34] That relationship has a slope of.74, which is a measure of the inheritability of the trait, is very high, and comparable to the inheritability height in humans. We humans have the same kind of propensity to transmit genetic factors to our offspring to that degree, approximately three quarters of the total maximum. [22:59] So, there's a high degree of inheritance, therefore, in beak depth and other traits as well, and then one can then take the question a little bit further and ask, "Well, what are the genetic factors involved?" I'd like a pointer here, but I don't have it. Anyway, I don't need pointer to ask you to look at the vertical scale and the high end of the vertical scale. I thank you very much. So, well, I can't...i need another one. [laughs] [23:26] [laughter] Peter: [23:26] I'm sorry, that's greedy, isn't it? OK, notice the low end of the spectrum there again. [23:33] [laughter] Peter: [23:34] So, what is it that is genetically different about these individuals here from those individuals at the top end of the spectrum here? We don't know. We don't know, but we're heading in the right direction with a study of contrasting species to find out the genetic differences between small-beaked and large-beaked species, and that is a stepping stone towards being able to find out what makes the genetic difference within species, the genetic propensity, therefore to evolve. [24:05] So, I'll tell you a little bit about what we know about genetic differences between species. And here I'm referring to collaborative work that we have done with two geneticists, Cliff Tabin and Arhat Abzhanov at the University of Harvard Medical School of Genetics. They are the professionals. We've collaborated with them on everything except doing the molecular work. [24:27] And the question outstanding right at the beginning was, "What are the genes that make a difference between a big-beaked bird like the large ground finch on the top, and the cactus finch down at the bottom. So, they started by taking the known list of genetic factors responsible for cranio-facial feature development in vertebrates, mammals in particular, and us humans even in particular within that particular category. [24:56] And they found one, a factor called a bone morphogenetic protein number four -- Bmp4, for short -- that is expressed in our early embryonic development and mice and chickens and a lot of other things as well. And they found that the expression at Day five in embryological development is graded in a fashion exactly parallel with the variation in adult beaks between the large beaks on the top and the most shallow beak of the cactus finch on the bottom, suggesting there's more than just association. Darwin Distinguished Lecture Series page 7 - ASU School of Life Sciences Grass Roots Studio

8 [25:34] I'll show you the pattern that shows that. Here's a diagram, and I'd like you to ignore the right-hand panel, concentrate on the left one, and notice that Bmp4 expression at day five in the upper mandible only -- we're looking at a plain, vertical slice through the upper mandible -- is very intense at day five in that one, you hardly see any expression in that one, and you have to imagine a graded variation in both of those. Cause and effect, or just simply correlation? A perennial question for biologists. [26:10] Well, to get at the possibility of a causal relationship, our collaborators did a nifty thing of injecting, with a retroviral vector, Bmp4 from magnirostris, into a chicken at the same stage of development to see if the Bmp4 from magnirostris, would convert a normal chicken into a kind of magnirostris chicken. It did, as you can see from the middle one of these three panels. [26:38] The top one is the control. The red double-headed arrow shows the normal beak, there is the experimental treatment, and this bizarre horrible thing at the bottom, unfortunate thing has been treated with another gene called noggin that antagonizes Bmp4. So the evidence is very strong of a causal relationship, therefore, between Bmp4 and beak expression in this dimension as well as, not shown in this picture, in the width. [27:04] That leaves the question of what governs legs. Here they use the micro-ray technique to go on a fishing expedition for genes that might be involved, and they found one, a very strongly operating one. It's called calmodulin. It is expressed very intensely in these two species with very long beaks; The Cactus and the Large Cactus Finch as you can see from there. It is graded, you can hardly see that in this picture, but it is graded all the way from the long beak ones to the deep beak ones. [27:37] So two genes make a difference in beak shape, and it turns out that this one is operating exactly the same time as Bmp4, but completely independently. So, any beak shaped different can be achieved by a change of just one genetic factor. I'm not suggesting that there are only two genes involved in the development of beaks there are lots, and lots, and lots. So this is really only just the beginning of an investigation into genetic factors controlling beak size and beak shape variations. [28:07] This will eventually allow us to understand why there is this genetic variation within a single population. Our guess is that the significant genetic variation doesn't lie in Bmp4 exactly; we think that Bmp4 is probably very structurally similar amongst the species, but it's in the whole complex regulatory apparatus governing the expression of this gene; that core ingredient resulting in different adult sizes, lies in the regulation rather than the gene itself. [28:42] Anyway, a very exciting beginning. That then tells you that the traits are heritable. They genetically, therefore, potentially, malleable or labile. Now, what about environmental factors that might be operating on this genetically variable trait? We're right at the beginning of our study. As I said earlier, we banded off the birds, we watched them without having to capture them again, we saw what they did; here's one cracking his seed, we new by his bands what his beak dimensions were. We found the unsurprising thing that big beak birds could crack open seeds that the small beak birds could not. [29:19] Then in 1977, four years after we began, there was a severe drought in the archipelago. Eighty-five percent of the medium ground finch population on the Island of Daphne died, and the 15 percent that survived were a non-random sample of the total. Large beak birds survived much better than small beak birds, as shown in this diagram. Owing to the fact that they were able to Darwin Distinguished Lecture Series page 8 - ASU School of Life Sciences Grass Roots Studio

9 crack open the large and hard seeds of Tribulus, this thing here measuring about a centimeter long, a woody tissue in which there are hidden some seeds. [29:58] Darwin would not have been at all surprised. He was challenged by one of his correspondents to explain exactly what he meant by natural selection. Darwin chose an example of birds to give an answer, and this was his answer: "I would now say," he said, "that of all birds annually born, some will have a beak shape longer and some shaped shorter. And that under conditions or habitats favoring longer beaks, all the individuals with beaks a little longer, will be more apt to survive than those with beaks shorter than average. [30:38] Well substitute beak depth for length, and basically that's what we found. One can ask rhetorically, was it worth our while doing all that work if he knew it back in 1867? He also knew that natural selection, which, by the way, is measured as that difference there between the start and the finish of an episode like that; he knew that natural selection would have no evolutionary consequences if the trait was not heritable. And I said earlier on the trait is highly heritable; there's a high degree of genetic characterization underpinning the variation in beak. And so, with that information it is possible to predict what the evolutionary change would be from one generation to the next. And here is the so-called breeders equation that gives away a predicting. [31:28] R, the evolutionary response to S selection is a function of selection strength and the heritability or degree of inheritance. What you see above the dotted line, are the surviving adults. They reproduced and produced a set of offspring, which, when fully grown, we measured in order to compare the actual offspring average beak depth whereas the predicted one, which is the broken line here, and there is the answer. Extraordinarily close, the prediction of an evolutionary change was very closely matched by observations. Evolution therefore, had occurred by natural selection on a genetically varying trait; functionally significant, functionally interpretable, when the environment changed. But we learned a lot more about natural selection by following the fate of this population through many environmental changes. [32:27] And the one I am now going to talk about is the change that took place as a result of the extraordinary amount of rain that fell in 1983, associated with an El Nio event, unprecedented in the last 400 years, according the coral core data. But notice that two years later there's virtually no rain; and then things changed. [32:49] So first of all, rather than show you figures, or tables of numbers, here's some pictures to show you what happened. Here's the normal dry season view of Daphne. The rain comes about January, February, March, or April. The island becomes green, and in the foreground, one can see Tribulus. This is the plant whose fruits, woody things like this, made the difference between survival and non-survival in Only the big beak birds can crack open and get out the seeds. [33:21] Well that plant was smothered in El Nio by the growth of grasses, first of all, but then vines, and vines grew upon vines; vines grew over the bushes, they grew over the trees. They grew over the bushes so much that the bushes disappeared. And even in the year after El Nio, one could still see the aftermath of this extraordinary production of plants; principally small-seed producing plants. [33:49] So the island environment have been converted from one dominated by large and hard seeds, to one dominated by small, soft seeds. And now the selective pendulum swung the other way. Now the birds that have the selective advantage were the small beak birds. Small body Darwin Distinguished Lecture Series page 9 - ASU School of Life Sciences Grass Roots Studio

10 sizes, birds with big tools for picking up and processing little seeds, small, soft seeds, at a rapid rate to meet their daily energy requirements. And that happened repeatedly. [34:17] And I'm now going to finish, before handing over to Rosemary, with a summary of the long-term view of natural selection in this environment. The orange arrows indicate droughts. They occur about as often as El Nios, and at each drought, there are high levels of mortality, and survival depends upon the nature of the food supply produced in the wet years immediately before the drought years. [34:43] So, that pattern of recurring droughts implies oscillating directional selection, and that is actually observed. The figure here is a little complicated, so I ask you to follow me through. Fortis at the top and scandens are the medium ground finch and the cactus finch. And the panels indicate selection coefficients -- the strength of selection on body size and beak size and beak shape. [35:11] If there were no selection, you would see a line from zero running all the way across each panel, but you don't. You see histogram bars above and below the zero line indicating fluctuating selection, and that in turn implies that morphologically these birds are not static. They're not unchanging. Rather, they do change as a result of this oscillating selection, and once again, that is matched by observations. [35:40] The observations here laid out in the same panel, only instead of selection coefficients, what you're looking at here are the mean values for body size, beak size, and so on. They would stay within these parallel lines if there were no change across the time, but, in fact, they don't. They move out of the initial confidence intervals, and so the remarkable thing is the birds now on the island are not the same that they were when we started our study. Morphologically, they've underdone change with this exception here, which I will put in a different frame here. [36:17] Here is the one and only exception. This is a trait that increased the beak size, increased markedly in the drought of '77, and then relaxed back again later. But, actually, this exception is more apparent than real, because we had a drought on the island in , 2004, into early and by coincidence, the amount of rain that fell in 2004 was exactly the same as in So, all other things being equal, we would expect here another rise in the average beak size, but instead, what we saw was an unprecedented drop in average beak size of the finches. This is the medium ground finch population. And why? [37:02] Well, all other things were not equal. There had been an invasion of another species onto the island. Starting in 1983, the large ground finch established a breeding population built up in numbers, I'll put it here, very gradually, and then getting faster and faster towards the year 2007, so that by the time the drought struck, there were more than a couple of hundred large ground finches on the island. [37:28] So what happened was this. As the food crunch hit those populations, magnirostris, a super competitor at crunching those Tribulus fruits, out-competed the larger members of the medium ground finch population, so that the survivors, the ones who survived best...all the individuals were at risk of dying, but the small beaked individuals in the medium ground finch population had the survival advantage. [37:55] So, here's the summary statement. I'll do it in two parts, one on this side. It goes like this. Under stressful conditions where feeding on Tribulus makes a difference between survival and Darwin Distinguished Lecture Series page 10 - ASU School of Life Sciences Grass Roots Studio

11 non-survival, in the absence of magnirostris, the large beak members of the medium ground finch population have the selective advantage over the small beak members, and the average population size of beak depth goes up. [38:23] Now in the presence of magnirostris, on the other hand, under the same conditions, now these ones are at a selective disadvantage. They're out-competed by these ones, and although the small ones are dying, they're dying at a slower rate, so that the average beak size goes down. [38:37] And the difference between medium ground finches and the large ground finches has now expanded. It's increased. That is character displacement as envisaged in that little cartoon I showed you of the speciation cycle. So, at this point, I'm going to hand over to Rosemary. This is the point on the island which is the easiest one to climb on. Rosemary Grant: [39:02] So, Peter has left us with two questions. How is the genetic variability maintained, because he's shown that a selection oscillates backwards and forwards, and under those conditions you would expect genetic variability to be eroded, but it's not. [39:18] So, the next question is, "What is the barrier to interbreeding between these two species, fortis on one hand and scandens on the other?" And the answer to both these questions is completely related, as I will be about to show you. [39:33] So, first, what is the barrier to interbreeding between species? Well, a very quick answer is song and morphology, because the two species differ in song, and as Peter has shown you, in morphology. They're very similar in all other respects. They build similar nests. They have similar courtship display. And they're similar in plumage - males are black, and females are brown. [39:56] So, we first asked, "Can individuals discriminate on the basis of beak shape and body size in the absence of any vocal cues?" And to do this, what we did is, we took two dead, stuffed museum specimens -- a female fortis, and a female scandens -- and put them on either side of a pole, took that pole into a territory, and asked if the territory owner could discriminate between his own species female and the other species female. [40:28] And the answer was a resounding, "Yes." They courted vigorously their own species, and completely ignored the other one. So, we then asked, "Can individuals discriminate on the basis of song in the absence of any morphological cues?" [40:43] And we took recorded song, and we played back first of all fortis song. Fortis came into the loudspeaker; scandens completely ignored the loud speaker. Played back scandens song, scandens came in, and fortis ignored the song. So, clearly they can discriminate both on morphology and song. [41:04] And it turns out that between these two species, fortis and scandens, that song is very important. So, what do we know about song? Well, song differs between a species. Scandens on the left-hand side sings, "Cha cha cha cha cha," and fortis we call the "Swiss bird" because he sings "mosly, musly, mosly, musly, mosly, musly." [41:27] [laughter] Rosemary: [41:28] [laughs] These birds sing just a short, simple song. There is individual difference, so in the "musly, musly" one you get variations of "muesli, muesli" -- some are faster, Darwin Distinguished Lecture Series page 11 - ASU School of Life Sciences Grass Roots Studio

12 some are slower. And in scandens you get variations in this "cha cha cha cha cha," and some are faster and some are lower. But each individual bird only sings one song. [41:52] We also know more, and this is work done by Bob Bowman many years ago, when you could take finches off the Galapagos, and he took some to California and had a breeding population going. And he found that song is learnt in association with morphology during a very short sensitive period of time early in life between Day 10 and Day 40. [42:15] And this period, Day 10 to Day 40, corresponds with the last few days that the birds are in the nest, and when they are outside their nest being fed by their parents and mainly their father. [42:32] Now, once this song is learnt, it is absolutely retained for life. We have many hundreds of birds recorded every single year, and we have never known one change its song, and lifetime can last for as long as 16 years. So this is a typical Laurentian imprinting process whereby species specific song is learned in association with parental morphology early in life and it influences behavior later in life because a bird's mate according to their species specific song. [43:14] So the consequence of this is a pre-mating barrier between the species. But we can ask how robust is this barrier, does it ever leak? Because after all song is learnt and this imprinting process would be very vulnerable to disruption if a young bird hears and learns another species' song during its short, sensitive period. And this does happen, it happens very occasionally, it happens typically if, scandens being a bit larger than fortis, sometimes takes over the fortis' nest, usually it turfs out all the eggs, but occasionally an egg is left and that egg grows up as a fortis in a scandens' nest and learns its foster father's song. [44:01] It also happens with death of the father. Females do not sing, so if the father dies, and the rest of the plutch is reared by the mother then what the birds will hear and learn is their neighbor's song, their natal neighbor's song, and if that natal neighbor is a bird of another species, it will learn that species' song. So we get these occasional learning of the wrong species' song. And this rare learning of another species' song has led to hybridization. And I want to stress at this point, hybridization has occurred throughout our whole study at this rare low rate. And it is low, it is usually about one percent or perhaps even less. And these individuals mate according to their song type so we get hybridization. And this hybridization gives us the opportunity to test the degree of genetic incompatibility between the species which is post-mating isolation. [45:06] So we asked how fit are the hybrids? Are they viable? Are they fertile? And for the first ten years of this study we had these hybrids -- one percent of each breeding time we got these hybrids -- but none of the hybrids survived to breed. So we thought oh maybe this is because they are genetically incompatible. Scandens and fortis are not sister species so they're not very closely related. So we thought maybe this was genetic incompatibility but we also thought there could be another reason. Hybrids are intermediate in beak size between the parental species and during this time, as Peter showed you, the majority of seeds in the seed bank during the drought were these large, hard Tribula seeds. And we watched hybrids trying to crack these large, hard Tribula seeds, but none of them were successful. They could crack opuntia seeds, which is what scandens will eat in the dry season, but they took two to three times as long as scandens did to do this. [46:12] And it seems that this was the correct explanation because after the El Nio of 1983 when the island was converted from a large hard seed producer to the small soft seed and they became Darwin Distinguished Lecture Series page 12 - ASU School of Life Sciences Grass Roots Studio

13 very abundant, hybrids survived. And they survived extremely well. But we can still ask, how well did the hybrids survive relative to their parental species? So what we did is we took all the birds that were born in 1983, 1987 and 1991 and we chose these years because there was a large amount of rainfall, large numbers of birds were produced and we had a decent number of hybrids. [46:57] And this is a survival curve for these three cohorts of the parental species and fortis and scandens over their lifetime and the cohort 1983, 1987, and Now if the hybrids survived less well than the parental species we would expect the survival curve of the hybrids to lie below that of the parental species, and instead, this is what we found. Hybrids survived very well, they survived as well as the parental species if not better, and I say that because there was no statistical significance between the survival, they also produced as many eggs, as many offspring and as many recruits. So hybrids were doing extremely well and clearly there was no evidence of incompatibility. [47:51] So how did these hybrids mate? Let me just take you through this. We take this one fortis-scandens here, and this is the result of a fortis father that had learnt a scandens song and so his son also sang the scandens song and mated with a scandens female to produce the first back cross generation, also singing a scandens' song mated with a scandens female to produce the second back cross generation and the same on the fortis side. Now if this fortis-scandens was a daughter and not a son, even though females do not sing, they do actually learn their father's song and they still mated with the scandens. So the pattern is exactly the same whether it's for males or for females. [48:43] So the message is hybrids back cross according to the song pipe of their father, genes will trickle from one species to another but the two species are kept apart by song. So this implies that fortis and scandens are converging genetically and to test this what we did was to standardized the difference, we took out 16 micro satellite loci, we standardized the difference between the species in 1982 before there was any intragression, hybridization and intragression to one. And then we followed the population through. And we found that the scandens' population became increasingly similar to fortis in genetic composition and we did exactly the same with our morphological measurements and we found exactly the same that scandens population became increasingly similar in morphology. And the two lines almost exactly parallel each other. [49:41] So does this mean that the process of speciation is collapsing? Well, yes it does, but it doesn't mean that it will go to complete collapse because the ecological food conditions, allowing the hybrids to survive only occur episodically. For the first ten years of our study, the food was not suitable for birds of intermediate condition, and they died. So what we see is a fission in the first ten years of our study and a fusion in the last 20 years of our study. [50:13] And so we have this fission/fusion dynamics. And during times of fusion, a fusion we feel is genetic variation and this top slide shows an increase in heterozygocity and it also enhances the morphological variation and you can see this through the beak shape variation increase and so today scandens, and actually it's the same with fortis but I'm using scandens as an example, it's really as though you were looking at a three foot or seven foot human, that is the amount of variation by which it has increased. And this increase in genetic and morphological variation on which selection can act, explains the answer to my first question, so it's episodic intragression will explain why the genetic variation is not eroded. Darwin Distinguished Lecture Series page 13 - ASU School of Life Sciences Grass Roots Studio

14 [51:08] So we can then ask, when does the pre-mating barrier never leak? When does hybridization never occur? And the answer is after the species have become substantially different in morphology and I can give you an example of this from the colonization of the big beak bird, Geospiza magnirostris, that arrived on the island, as Peter said, in 1983, and the population built up so that now there are over 200 individuals on the islands. [51:38] And during this time, eight fortis imprinted or learnt a magnirostris song, but none actually bred with a magnirostris and we were at first a little surprised about this, because fortis and magnirostris are sister species. Genetically they're very closely related, much closer than scandens and fortis. [51:58] So we thought that it could be recognition. When the species are very different in size, and magnirostris is almost twice the size of fortis, that they do not interbreed. [52:10] But it could also be harassment, because when these poor little eight fortis opened their mouths and sang magnirostris's song instead of singing "mosly, musly, mosly, musly," they went "chee-a, chee-a" and every time they open their mouth to sing "chee-a" a magnirostris would whip in from nowhere and just beat the living daylights out of them. [52:33] [laughing] [52:35] Now there was one poor little fortis that sang a magnirostris song and he sang and got beaten up, he sang and he got beaten up, he sang a little quieter, he still got beaten up. He went behind a bush and he sang a little quieter and he did this for years. And after about, I think it was after about four or five years, he stopped singing altogether and then he got a mate. He got a fortis mate but not a magnirostris. So the pre-mating barrier is robust when the size difference between species is large, even learning the heterotypic song didn't lead to hybridization. [53:16] But let's go back to this case where the barrier leaks when the size difference between the species is small. But this only happens when the ecological conditions are suitable for survival of birds with intermediate bill size. [53:31] So how general is this process of episodic hybridization and introgression. We asked, "Does it only happen on Daphne?" We saw that it also happened on Genovesa, another island where we had done a long-term study. But we asked, "Could we see the footprints of hybridization and introgression in the recent past on other islands in the Galapagos Archipelago?" [53:53] And we asked the question in the following way, or tried to answer it in the following way - we said, "If the transfer of genes occurs episodically, species A should be genetically more similar at neutral loci, like microsatellite loci, to closely related species B on the same island, than to species B on a different island." [54:14] That was a mouthful, but I can explain it a little bit better. I can say that if fortis and scandens, who know have been introgressing genes on Daphne, so would fortis and scandens on Daphne be more similar genetically at these microsatellite loci than fortis is to scandens on any other island? [54:36] And we did this for all species pairs of ground finches and we found most species of ground finches are genetically more similar to a closely related sympatric species than they are to Darwin Distinguished Lecture Series page 14 - ASU School of Life Sciences Grass Roots Studio