thank you so much lows for the very kind introduction and thanks to all of you for coming today I’m really excited to tell you a little bit about the work that I’ve been doing to untangle the complicated evolutionary history of convergent color pattern and new world snakes and how their example may serve as a window into understanding both pattern and process and evolutionary convergence more generally now to set the stage I want to explain that this is more than just a pretty title slide so the two clades of snakes that are represented here this is a venomous elapid coral snake on your right and I’ll harmless colubrid snake on your left these two clades of snakes diverged from each other just after the end of the Cretaceous about 65 million years ago right about the same time as the major crown diversification of most orders of modern mammals so these are two very very different snakes the individuals that are depicted here are illustrated from two real museum specimens these two individuals were collected together at the same field site on a ummz expedition to Guatemala in 1940 and yet despite the great length of time since their most recent common ancestor and the fact that the vast majority of snakes in the world are some variation of brown these two species now coexist together in space and time in almost perfect replicas of the same very conspicuous color pattern so today I’m going to tell you a little bit about how and why we think that came to be which does come with some surprises for perhaps more excitingly how likely such a thing is to be true millions of years from today and so to help you it to have some context for understanding why I’ve pursued research on mimetic convergence I want to say a few brief words about the major themes of my overall research program so my research interests ban a broad range of topics across both ecology and evolution as well as some applied conservation biology I’ve worked on a number of different systems in a number of different questions over the course of my research career so far however one common characteristic that unites most of the research that I do is asking big picture questions about trait evolution especially about the ecological factors that drive the origin and stability of novel traits over time now I study all kinds of traits which often co-vary so I tend to rely pretty heavily on interdisciplinary he approaches to the questions that I ask now I spent most of my early career investigating the origin of cooperative nuclear family social groups in lizards in which I described an independent orange origin of complex sociality and vertebrates now here you see one of the species that I studied this is the desert night lizard that inhabits the joshua tree woodlands of california and i combined experimental experimental manipulations in both the lab and the field with molecular analyses of relatedness and to look at not only the mechanisms but also the fitness consequences of the transition between solitary and social behavior y becomes social i also spent about four years investigating variation in mating systems and the role that alternative mating strategies can play in promoting reproductive isolation so here you can see six polymorphic throat color phenotypes found in a single population of lizards and these these phenotypes correspond to three different mating strategies that cycle over time through rock paper scissors dynamics in which each strategy can beat and is beaten by one of the other strategies in the system now I investigated the reproductive consequences of shifts between a polymorphic system and a monomorphic system in two independent origins of rock-paper-scissors dynamics in the side watch lizard in the US and Mexico and in the common lizard in Europe now all of these projects whether they’re on sociality or mating systems or mimicry they’re all fundamentally driven by my desire to understand how and why we see shifts to new and sometimes very distinct states and trait space and especially how reversible these transitions are over both ecological and evolutionary timescales and this focus has led to my appreciation for the power of evolutionary convergence multiple independent origins of the same trait for not just characterizing species specific patterns of change but for actually having statistical power replicates in order to test hypotheses about trade evolution in nature which is what I’m going to talk about today so my talk will be divided into two main parts first I’m going to get some background about one of the ultimate examples of convergence and that’s mimicry a little bit about what we know and what we predict about how mimicry should operate and how we can use snake mimicry to test hypotheses about evolutionary trajectories of traits across all of New World snakes then we’re going to look at a single played in depth which allows us to follow the mimetic phenotype over the landscape of both time and space and test among mechanisms producing the patterns that will see in part one and then briefly at the end I’ll talk about some of my ongoing and future research goals and especially how these projects can interface with both my teaching and my outreach as an assistant professor here at the University of Michigan this

is a photo of our of the University of Michigan and Peruvian research team that was on our recent expedition to the Amazon and I’ll talk a little bit about that today as well so first I want to make sure that everybody is on the same page when I use the term mimicry so there are two main types of mimicry each name for the man that proposed it fritz Muller and Henry Bates which perhaps not coincidentally both based their work on observations they made in the Amazon during the 1800s which in this artist’s representation involves getting attacked by ferocious flocks of two cans which I can assure you doesn’t actually happen so in the first kind of mimicry mullerian mimicry multiple toxic or noxious species converge on the same warning phenotype for signaling their toxicity to their predators now a great example of this is bees and vested wasps they’re not closely related but yet they’ve converged upon a similar yellow and black banded coloration which is an honest signal of their toxicity to other organisms so in this kind of system it’s sort of like every species is simultaneously a model and a mimic and the other kind of mimicry PhD in mimicry there is a toxic species that has an honest signal but there are also harmless species that are called mimics which are giving a dishonest signal because although they have the exact same color pattern as the toxic models they’re completely palatable now in both cases this phenomenon results in multiple independent origins of the same trait which is the key its key to the test that we want to conduct now it’s this latter type of mimicry that I’ll be focusing on today and specifically the basting mimicry of New World coral snakes most of which are in the genus cur’us this is one species indicated here now coral snakes are highly venomous and they’re complete not only with bright warning coloration but also with a striking behavioral warning display involving a lot of tail waving and body flattening and thrashing while the snake hides its head there are at least 80 species of coral snakes distributed across almost every biome in the Western Hemisphere their main predators are tetra chromatic birds with excellent color vision in contrast to the coral snakes themselves which have much more limited capacity for color discrimination now although there is variation across coral snakes in warning coloration all species have a color pattern that some permutation of red and black banding most excitingly there have been more than 150 species identified with the same warning coloration but that appear to be quite harmless to predators these are mostly in the family colubrid a I mean they represent an unknown number of origins so now because i’m going to show a lot of photos in this talk that looked like this i think it’s important for me to emphasize here why this color pattern is something special so for those of you who don’t spend a lot of time trying to find snakes in nature it’s easy to get the impression that this is just another typical snake color pattern but herpetologist literally spend thousands of hours staring at leaf litter like this trying to hone the perfect search image in order to visually resolve the snake that’s sitting right here this is a fair to Lance which is a kind of Viper coiled up against the buttress of a tree and so that’s why when early natural is for tromping through the Amazon finding a bright red and black snake was something of note because they’re so easy to detect compared to something like this and this observation was even more remarkable when it was realized that actually many of those brightly colored snakes were pretty harmless to humans and which was in start contracting stark contrast the coral snakes that could easily kill you so why has coral snake mimicry been controversial well that has to do with the theoretical expectations of how mimicry is supposed to operate so Alfred Russel Wallace was actually the first to write a formalization of men Chris theory in the 1860s and he settled upon three logical requirements that should be true for mimicry to persist over time and these laws despite their age have actually held up quite well for insects models and mimics are generally found in the exact same geographic area especially in the tropics where Klein’s of phenotypic variation tend to be quite sharp the mimetic phenotype is almost always more rare than a cryptic phenotype and the mimics are almost always less abundant than the models now it’s important here to note that nearly all of what we know about batesy and mimicry the theory the observations everything comes from butterflies but there are a wide variety of systems that have batesy and mimicry frogs hoverflies even fish but my favorite examples among invertebrates are organisms that mimic ants where we see examples of convergence that are so extreme that they are very difficult to explain through any other process besides batesy and mimicry here you can even see this is a tree hopper with part of its thorax in a surprising surprisingly close rendition of one of the sim Patrick ant

species and these kind of examples have led to a very inclusive perspective perspective to the study of mimicry and insects we’re pretty much every example of striking pattern matching to something that’s toxic pretty much considered mimicry at least until proven otherwise but this has not been the case for snakes it’s been a very exclusive perspective and lots of controversy that’s led to a debate that’s been raging for about a hundred years so many researchers actually advocate the perspective that coral snake mimicry doesn’t exist at all or if it does it’s an extremely limited phenomenon restricted to just a handful of species in just a few places now things came to a head in 1981 when Harry green and Roy MacDermid wrote a paper in science that they actually had to title coral snake mimicry does it occur in which they are you among other things that for at least some places these three rules are indeed satisfied by coral snakes in their mimics and that this coloration is indeed conspicuous which had been a point of major debate they have a gorgeous figure in their paper showing variation in in color pattern matching across Central America in each of these species pairs here the lethal coral snake is here on the left with a rounder stretch snout and the totally harmless mimic is here on the right with a more pointy snout now although this paper convincingly showed that there are circumstances under which these these rules are met they did not fully quiet the critics who were noting legitimate problems and the problems are mainly with rule 1 and 3 so lots of researchers have pointed out that there appear to be mimics outside the geographic range of the models both north of the distributional limit of coral snakes in the United States but also outside of the Western Hemisphere and in fact snakes have the only known example of an undefended mimic completely outside the range of any model species this is the California mountain kingsnake lampropeltis inada this geographic range is shown here in black which never gets closer than about 200 kilometers to the nearest coral snake shown here right and in terms of individual abundance these colubrid mimics can sometimes outnumber venomous coral snakes by quite a lot now these observations create a legitimate problem for explaining the persistence of such a conspicuous phenotype on a completely palatable animal especially over evolutionary time but yet we can’t ignore the striking pattern of phenotypic pattern matching that we see in snakes like in the picture that Harry green and Roy McDermott had in their paper and so elegantly illustrated so the legacy of this controversy has been a whole lot of lost opportunity snakes have simply been left out of nearly all major reviews of mimicry over the last 20 years there are no comparisons really being made to insect systems and this is a shame because snakes actually represent the best and most well-studied comparison to butterfly mimicry and there’s a lot to be gained to understanding both now this tension between theory and observation has led to two possible alternative explanations other than coral snake mimicry doesn’t exist at all and you can categorize these as either better tests or different rules so maybe the mimicry theory that we’ve developed for insects especially our ways of interpreting costs to predators in making an identification mistakes are simply inadequate for predicting the dynamics across the full spectrum of mimicry systems you have big lethal models and innate avoidance of those models by birds which has definitely been found those just might need modified theory secondly if you’re looking for correlations between models and men and mimics in terms of phenotypes we know that species do not represent statistically independent data points so first species are not randomly distributed across space there are continental patterns of diversity like the latitudinal gradient and strong spatial not independence relative to other species especially within the same clay secondly we know the distributions of phenotypic traits often have phylogenetic signal so you have to account for that kind of non independence as well this is a figure from a classic paper by Ted garland showing the importance of phylogenetically informed tests of trade correlations and reminding us that traditional statistics assume a star phylogeny which is rarely a realistic assumption so what we had was just a handful of point observations this model mimic species pair over here seems to conform to expectations this other one over here seems not to but we didn’t have with a systematic unification of all the data but knots what I decided to do the first comprehensive statistical test of mimicry across the entire western hemisphere and the 80 million years that snakes have been in the New World I used two complementary approaches to test for a correlation between models and mimics over space and over time and I’m going to describe both approaches here and then present all of the results so first we phenotype to all of the New World snake species for coral snake coloration which is best thought of is simply red and black banding that we had a few reviews that had helped us

out with this but we also updated and standardized all of the taxonomy we added new species that hadn’t been available to the original authors and added more information about variation and coloration next we constructed geographic range polygons for as many species as we could now for the coral snakes we hand constructed these range ranges from published sources mostly books but for most of the snakes we made these ranges based on museum data now I want to highlight here that this was a major undertaking so there’s no database that you can go to and just download the geographic ranges of the snakes of the world even for most countries in the Americas there’s not even a book that’s the reptiles of that country and in fact for the majority of snakes the only information about their distribution is what’s present in the collections of natural history museums around the world and it’s thanks to the huge digitization their efforts of these museums so you get their specimens publicly their information for their specimens publicly available that we were able to find usable data for almost three hundred thousand specimens of snakes across the New World then we want to do something very simple just simply compare the patterns of species richness we’re looking for positive correlations between models and mimics while also accounting for patterns of overall species richness been for the phenotype or a phylogenetic approach we augmented our phenotypic data set across all old world species specifically looking for this red and black banding then simply we reconstructed the origins of the mimetic phenotype or at least the red and black banded phenotype and look at changes in rates of phenotypic evolution but specifically we wanted to compare the timing when does this red and black coloration arise and colubrid relative to coral snakes simply asking do coral snakes predict mimicry and if so how well so what are the data look like what you see here is an overlay of every species range polygon that then have raster counts for every latitude and longitude grid so these counts are then colored is a heat map where places with red have more species and places with what should be kind of tannish have fewer and what do we see well we see a classic pattern of a latitudinal diversity gradient where the center of snake diversity is here in the tropics what about for coral snakes we see a somewhat similar pattern except it seems to be more restricted here to the Western Amazon basin so in Ecuador and Peru and not spread across the entire the entire Amazon so what about mimics so at first glance it seems that mimics generally look pretty similar to all snakes you find lots of mimics where you get lots of snakes and that’s pretty much it when you run a spatial autoregressive model on this this is just a regression that accounts for spatial autocorrelation we find that mimetic richness is positively correlated with both total species richness and coral snake species richness but we weren’t a hundred percent confidence in these results this is a parametric tests and we weren’t sure how how well it was doing at handling non normality and the distributions of our species richness counts because they were decidedly non-normal we also weren’t quite sure how they were handling the complex shapes shapes of continents because the weighting matrices for distance sort of assume a square and that’s not the shape that we had at all so we used a permutation test what we did is we randomized the species ID tags that are associated with all geographic ranges that were not coral snakes doing this we could simulate a null expect a shin of how many mimics you should expect in the locality based only on species richness then we want to visualize the residual variation left over after you remove this effect so we’re asking where are there too many mimics given you know that this effect occurs and so what you’re seeing here is instead of species counts you’re looking at standardized residuals we’re and we’re specifically looking for places where the observed value of mimics is much higher than the expected value and that’s what you’ve seen that the threshold for significance is two standard deviations above the mean although we did see extreme values that were almost 25 standard deviations above the mean and moreover this these extreme values had strong strong Geographic signal that matched coral snake richness patterns very well so if you remember no other slide from this talk remember this one despite all the potential sources of error we had in our geographic range reconstructions of which there were new numerous sources coral snakes pretty much explain all of the residual variation in the distribution of mimetic colubrid once you account for overall species richness and I think that’s pretty amazing so to look at these results in another way here in grey are the simulated expected mean values of mimics for all grid cells with the corresponding number of coral snakes what you see here is

just the correlation between coral snakes and mimetic snakes that you would expect based solely on species richness alone but the black dots are the observed data and what we see is a strikingly increased correlation such that there are two to three times more mimetic species than expected and actually the relationship is surprisingly linear with a slope of about two suggesting that for every additional coral snake species the system can support two more mimetic species which is way too many mimics but what about the abundance of individuals which is perhaps the better test just because there are multiple mimetic species if they’re super rare and coral snake species are super common the abundance situation could still be that coral snakes are more common than their mimics so remember we have 300,000 museum records and so we can actually test this we can plot the ratio of mimics two models my grey line is a little light there sorry but that’s a reference line at 0 representing locations where models and mimics are equally common so if you have a situation where there where mimics are more abundant you should see that locality showing up on this side of the reference line is if coral snakes are more common will be on that side so what do the data actually look like and we plot all of the grid cells we find this although there are definitely places where coral snakes specimens are actually more common than their mimics and sometimes three to four times more common if you note the log to scale the overarching pattern is that mimics are more abundant in collections than coral snakes from the same area sometimes two great excess and as an independent estimate I’ve also overlaid here the ratio of mimics to models that I observed on our recent trip to Peru and you see that it’s about average but definitely mimics were more common so I think that there’s a discussion to be had about whether collections are a good proxy for real ecological abundance but these ratios are so skewed that I think it’s hard to explain these results by collection bias alone and i think that the general pattern is a real one such that the overabundance of mimics is actually the common state rather than the exception so now that we’ve looked at spatial correlations what about temporal ones so what we have here super great either is a time calibrated phylogeny of all snakes the major clades here are marked so remember that coral snakes are lapid so they’re going to be here kind of at the base of this plate and what I’m referring to is colubrid are basically these subclades up here the dip students to nature scenes in the kalu brains so we’re looking for a correlation between models and mimics and time we reconstructed colorevolution across all snakes using three different methods these methods were generally concordance so I’m only going to show the results from the maximum likelihood and stochastic character mapping so now this tree not throwing up well sorry guys it’s got a number of different things marked so first you can see these yellow circles so snakes are primarily an old world played that has made multiple independent arrivals in the new world and those are each marked by these yellow yellow circles then every node that we reconstructed as having the red and black banded phenotypes the descending branch is then colored by either red if it’s a tax on that’s in the New World or this brown color if the tax on is in the old world and what do we see we see that there is indeed a background transition rate not surprisingly just like any other trait this red and black bandits phenotype was present across the entire tree to low frequency however those transition rates skyrocket in the New World colubrid now a lot of the transitions in color phenotype were recovered at the tips of the tree overall be recovered about 60 origins and about 30 losses but those could be influenced pretty heavily by taxonomic uncertainty if they’re at the tips of the tree so the numbers that i give here for origins and losses that we found are all deeply nested transitions that are clearly robust and independent i think that there are two important takeaways from this analysis first the only clade in which red and black banded coloration spans the transition between the old world and the new world is in the models the coral snakes in every case of an origin within new world colubrid the coloration evolved after they were sympatric with coral snakes in the new world which is the key prediction of BT and mimicry but secondly we see that it’s not just once a mimic always a mimic there’s a lot of phenotypic change and that changes in both directions and this is a particular note because despite a lot of treat evolution studies in insect mimicry systems no losses have ever been recovered at least not irrefutably and the strength of this empirical evidence has led a number of researchers to hypothesize that mimicry is a one-way street or where in which know mimics ever revert back to the cryptic

phenotype now there have been two reasons that have been hypothesized to explain this so first it’s possible that the geographic ranges of insects are oddly stable but i’m not sure that there’s a whole lot of data supporting that and perhaps what’s more likely is that it’s due to the genetic architecture of coloration components so in all the baton mimicry systems for which we understand the genetics of the mimetic coloration the multiple pigmentation genes that are involved in making the complete signal are all physically linked in small chromosomal regions called super genes and given this architecture it may be very difficult for butterflies too quickly deconstruct the mimetic signal when an ala patrie to their model and in that case such allopatric populations would quickly go extinct and thus be invisible when you’re making a phylogeny of extant taxa now I think that this is a an interesting idea and I’m not sure we have a whole lot of evidence for this either but what we can say for sure is that this one-way street is not what we see in snakes so we find lots of losses and I don’t have time for all the analyses that I did but I can say that they aren’t clearly related to a la patrie with coral snakes which is a bit surprising but bringing this back to our set of alternative explanations for why sneak mimicry systems seem so different is it better tests or different rules that are accounting for this conflict between predicted and observe distributions of the mimetic phenotype and I would argue that it’s actually both some of the conflict is explained through better tests and some of the conflict is even worse than historically suggested so coral snakes explain way too much of the variation in the distribution of the red and black banded phenotype to argue that coral snake mimicry doesn’t exist or even as a very restricted phenomenon and most importantly if you’re looking for correlations between models and mimics you can’t treat each species each species as an independent data point although i would say at least superficially this non-independent seems to be more important in snakes than it isn’t insects but despite the overwhelming support that we found for batesy and mimicry we also found or confirmed patterns that you just don’t see an insect mimicry systems suggesting that the rules may be different too and specifically these requirements for an increased abundance of models or for sympathy between models and emic seemed to just not be applicable to snakes and that the evolutionary lability of the red and black bandage phenotype seems to be pervasive and extreme and includes numerous losses so I would argue that if you really want to understand mimicry snakes provide a very important comparison to insects and especially to butterfly mimicry and that those comparisons are probably of the greatest importance going forward to figure out why these discrepancies exist but I’ve oversimplified everything that I’ve shown you so far and there’s one key factor in particular that I left out and that’s that snake mimicry systems just like insect mimicry systems have spectacular color polymorphism in which multiple very different color types coexists and patrick lee within single populations of single species of both models and in mimics so the species shown here is one of my main study species this is the western ground snakes Sonora semi-annually de and under a single rock you can find four different color phenotypes a mimetic form of red and black bands ones with just the black bands ones with just the red and one that seemed to have once seemed to have neither color component so in conducting our global phenotyping across all snakes we absolutely paid attention to this color variation in addition to simply noting whether one of them works had this red and black behind it phenotype and in doing so we determined that at least a third of all mimics and a fifth of all coral snakes have a major color polymorphism for either the red or the black banding and that’s almost assuredly an under estimate given how this kind of polymorphism is reported in the literature so that brings us to our second major theme in which we use this color variation is an opportunity that allows us to follow the mimetic phenotype over time and space in a focused way looking at a single claimant depth and specifically there are two goals here first focusing on the ecological interactions between the models the mimics and the predators what is the relationship between predator and prey do predators avoid the red and black phenotype as predicted by mimicry theory and secondly focusing on a single clade allows us to investigate the genetics of pigmentation many cities have suggested that biases in the direction and frequency of phenotypic change may be due to the genetic mechanisms that are underlying trait development but how to sneak mimicry compared to butterfly mimicry do they have super genes too and as an overarching idea over all of this what is the role of the color polymorphism so the clade that I looked at in depth is this tribe so noreen e there are a clade of small secretive snakes in North and Central America there are about 25 species they have various levels of color polymorphism and of mimicry although the extent of both is pretty

unknown before I started working on them but the reason I chose that chose them is because their northern distribution and the fact that they specialize on eating insects means that they reach abundances that are more typical of lizards than of snakes that means they’re relatively easy to catch in very large numbers which is not true for most of the tropical species so the original heat map that I showed you a coral snake speeches species richness made it seem that all the excitement was in South America and while that’s true coral snakes in North America are still quite interesting with lots of color variation among species and complicated patterns of geographic range overlap both with other coral snake species and with their mimics so this is that all of the species of Sona rini that are also mapped I’ve simply randomly color ranges by species here with transparency so you can visualize some of the complexity now the sooner a species do have a similar distribution to coral snakes although there are more of them but I mostly focused on this top species here in yellow which has areas of both a la patrie and simp at Rita coral snakes and they overlap it the part of the ranges at least three different species of coral snakes so in order to test hypotheses about selection and the genetic control of coloration I compiled two different data sets a phenotypic data set and a genetic data set now soon or are mainly captured through rock turn surveys as I’m demonstrating here where you flip over a lot of rocks and take snakes out from underneath from our work we estimate you have to flip about a thousand rocks for every Sonora that you catch so we augmented our field data collection with some museum specimens as well we ended up with a total data set of almost 3,000 individuals we also collected three different genetic data sets today I’m only going to be showing the data from the aflp analyses but you should know that the other data sets to exist as well so first looking at the distribution of morphs across space here you’re going to see I’m data for a hundred and eighty counties in the United States and the height of each of the bars is proportional to the number of individuals in that county with that phenotype so first we see that the mimetic phenotype is distributed all across the continental United States within the range of the species even up in Idaho we’re actually the original specimen that was found there in 1923 was actually misidentified as a coral snake however by far the most common worse seems to be this uniform brown more and the one with only read more importantly coloration seems to be completely decoupled from any environmental variable that we can find so it doesn’t matter it matter whether you’re looking at the presence or absence of morphs the proportion of morphs the proportions of ones that just have black once it just have read and it doesn’t matter whether you’re looking at latitude longitude sympatric models anything we can think of we don’t see any correlation so this immediately begs the question is this phenotypic variation neutral or is it driven by selection so the way that we tested that was just simply compare genetic and phenotypic variation for all the populations for which we had at least 20 contemporary samples we need tissue samples so for example if you’re looking for a signature of either genetic bottleneck sword rift you can imagine a situation where if a population has gone through a bottleneck you get reduced genetic variation you should have lost phenotypic variation as well it’s you should get a positive correlation between these two variables what do we see nope no correlation they are either so what about genetic distance versus phenotypic distance you can imagine if the populations are sharing a lot of migrants they’re going to be pretty similar genetically and they should also be pretty similar phenotypically do we see a positive relationship here now we don’t in fact it’s only when you compare FST s well we found very little genetic variation among populations the average is about point O six but phenotypically populations were quite different and this kind of pattern is the classic signature of phenotypic traits that are under selection but what kind of selection so the way that we tested that was in two different approaches first we compared to age classes within a single generation do you see different frequencies of the color morphs and juveniles versus adults so here what you can see is for every gray bar that’s a set of juveniles and for every black bar that’s your adults and we see that there are quite different phenotypic frequencies across the different color classes which is another classic signature of selection but what about it changes in more frequency over many generations anytime you have a color polymorphism that seems to be maintained by selection you probably should be thinking about negative frequency dependent selection basically an increase in selective advantage when the morph is rare this is a classic mechanism for maintaining polymorphism within populations so the way to look for that is to simply for every population where you have a long capture history from your museum specimens use multinomial logistic regressions to

basically look at the change in more frequencies over time and what do we find so yes we found very convincing evidence of negative frequency dependent selection so each of these lines is different one of them works and that the probability is simply a proxy for the frequency the real frequency of them worked within the populations and you see very clear cycling of more frequencies over time with the morph that was the most rare at the beginning of the time window becoming one of the more common ones by the end so we didn’t just find this in a single population but we found it in every population for which we had a long enough time sampling interval to actually perform this test so the reason why this is important is because it makes morphs hard to lose so even in a la patrie to the models if a morph is under strong selection positive selection when it’s rare can be very hard to eliminate from a population entirely now this is an exciting study because it’s the first direct assessment of selection of color pattern and a snake mimicry system and I want to be clear that what I mean is that it’s not using clay replicas so rather than focusing on how predators interpret visual stimuli looking at their attack rate on a clay replica you put out on the ground what you’re actually looking at is the real response of prey populations over time to predation which is pretty exciting so the one other thing that you can do with this data set is to examine the genetic architecture of the mimetic signal so although breeding stat studies in the laboratory are always the gold standard because of the kind of control you have within a laboratory setting not all tax are amenable to large-scale breeding studies to address things like genetic control over coloration and snore is definitely one of those pieces that’s not super amenable to that but with a whole lot of population of data population level data on the observed more frequencies and some basic population genetics you can make some inferences about the likelihood of some very simple models of genetic and epigenetic inheritance and these are the simplest models that can conceivably produce the four phenotypes that you see here so one locus model where you have three alleles one for red one for black one for null there codominance so if you are given are and a B then your mimetic if you have two nulls your uniform and so on and so forth it could be that red and black are actually are actually two different loci in which case you have four alleles are a red a null for the red a black and a null for the black or it could be the seas tu lo sai are quite linked you can imagine a situation under total leakage disequilibrium you would find populations that have all mimetic and all uniform but not any of the other two types so you can simply calculate these expected phenotypic frequencies under higher hardy-weinberg for these different models and do a quick comparison of the fit of the models to the observed data and generate ase scores so which model is more likely so what I’m going to show you here the Delta AIC scores for each of the 40 populations that we had with enough data to actually perform this test so here’s the reference line here at zero any population that falls along the reference line would be pretty uninformative it doesn’t support one model or the other it’s above the line in this plot it’s going to support a to locust model over a one locus and if it’s above the line and this one is going to support the unlinked model versus the tightly linked model or supergene model so what are the data look like well there are lots of populations that are actually completely uninformative mostly these are populations along the Great Plains if you think back to the 3d maps that are highly that have a high frequency of the uniform work and that really doesn’t tell you a whole lot about the the control of the red and the black but for the populations that did have high information content we see very strong very strong support for red and black being controlled by separate low side and support that those loose I are not linked and in fact the model with the very stronger supergene linkage was actually a worse fit than the ma than the one locus model and this is interesting because finding color components maintained on separate chromosomes has only ever been described for mullerian mimics and we’ve in all the baiting mimics that have been looked at all we ever find our super genes so it’s possible that senora are an anomaly right they represent an atypical breakdown of genetic linkage but if that’s not true if these are an accurate representation of the coloration genetics across snakes more widely then there are two important ramifications of this finding first because the mimetic signal can be broken down through simple recombination you should expect the trait to evolve very quickly and you should be able to predict what kinds of works that you see that are in components of the of the total signal and both of these may work together to be part of the explanation for the extreme lability that we see and mimetic phenotype across an entire

phylogeny so bringing it back to some of the spatial analyses that we did instead of looking at just the residual variation of mimics this is now looking at the mimics that are actually polymorphic where do you see too much color polymorphism given total species richness and we actually find that polymorphism seems to be too high both in places with high coral snake species richness and in places with the lowest coral snake species richness and to me that suggests that there’s a mixture of processes and well it’s not a perfect correlation it seems that the cryptic mark seem to be much more common in a la patrie and so this ends up being a key trade it’s not just does polymorphism exist what is the actual nature of the polymorphism which ends up being critical for understanding how phenotypes are interacting especially from the perspective of a predator but I think the overall message here is that this color polymorphism appears to be a key step in both the origin loss and maybe even maintenance of mimicry so my hope is that everyone takes away from this talk a somewhat new view of mimicry moving away from the oversimplified easy to explain version of mimicry as illustrated here by two species of frogs in which there’s a single mimetic species that’s perfectly imitates the color pattern of a single model species and that we instead embrace the dynamic and complex reality that we’re just beginning to fully appreciate where single species can actually have exceptional diversity of quickly evolving phenotypes and this represents an exciting opportunity for understanding how evolutionary convergence operates over time in space so where do we go from here so i have several projects for which i’ve been collecting preliminary data and that i currently envision is both feasible and fundable moving forward they’re a good mix of field lab museum work that are all geared towards the goal of better understanding the mechanisms and dine amex behind novel phenotypes now only think I have time to briefly outline one of these projects is sort of a representative of how my work promotes educational collaborative opportunities for students at Michigan and beyond so bringing it back for a moment to Mueller in base so some of you in the audience may already have recognized that all of the coral snake mimicry problems whether it’s the abundance problems or the distributional mismatch between models and mimics or the curious lack of genetic linkage all of these concerns suddenly disappear if the system isn’t actually batesy and it’s mullerian if all of the species are toxic then it no longer matters which species are allopatric to others or what the abundances of each species are reinforcement of predator avoidance would then be everywhere that the snakes fur so this is not a new idea but it’s certainly not been well tested I want to make it clear that no new world colubrid snakes have the same kind of advanced venom delivery system that you see in coral snakes they don’t have front things they don’t have muscular control over over their venom glands and all of those kinds of adaptations that can clearly kill predators but a surprisingly large proportion of mimics do you have a system for incapacitating their prey so they have enlarged teeth they’re just in the rear of their mouth and those teeth are connected to specialized glands that sit in the same place as real venom glands on elapids and on vipers and they’re actually hooked up to the teeth now the components of these glandular secretions have never been really well examined systematically I mean certainly not across hypothesized transitions to a venomous state but it the question begs to be asked how toxic actually are the minute mimics so before we actually formal and preserve specimens for collection we’ve been removing these glands that you weren’t always glance this is a picture of johanna larsson doing one of these gland removals on a coral snake mimic you simply go in through the mouth in between the teeth and the cheek and pull out the gland and actually the specimens look great afterwards and you can’t even tell what we’ve done so now we’re ready to analyze the RNA in these glands to examine transcript differences across species and especially across evolutionary transitions and this will be the first such broad scale a say in colubrid so linking venom components to tooth evolution so even without getting venom glands from every single species of new world snake we can take advantage of the huge collection of preserved specimens at the ummz from the perspective of skeletal changes so an upcoming senior thesis students Aaron West Indies currently a junior is using a combination of micro CT scanning and geometric geometric morphometrics to phylogenetically model tooth evolution across new world colubrid and here you can see one of her scans this is the front of the snake here this is the hinge of the jaw here and I have false colored the enlarged maxillary teeth here in red so you can see you can see how big they are relative to the rest of the tooth in the mouth and so you can ask a bunch of questions here how many of the teeth in there in the rear of the mouth are enlarged how enlarged are they

relative to the others is there grooving on the teeth which is the conduit for the venom I’m coming down out of the gland what’s the curvature like and most excitingly are these transitions correlated with changes and to the art red black and banded phenotype nobody’s ever approached this question quantitatively and I think that if you want to study something like cranial morphology across snakes University of Michigan is an ideal place to be because it’s the largest university affiliated collections of herps in the world so although university michigan is clearly great home for me i think it’s also important to think about giving back what opportunities do i provide so you’ve seen some some demonstration of some of the student opportunities that are available through my research and also because this work is across Latin America there’s a lot of opportunity for international collaboration of outreach but one of the most exciting opportunities for uniting research and teaching is my vision for adding more field based courses to our catalog so in graduate school as a teaching assistant for 13 courses many of them with a strong field component and then as a postdoc I was a primary instructor for two field courses one of them the flagship course on vertebrate Natural History at UC Berkeley and another was a destination field course to Kenya and I think that my experience teaching field courses hit has led me to really appreciate the incredible city for field courses to inspire students both about doing their own science and also for drawing in new majors to Evie and I think that there’s a great opportunity for more of that here that I would be really excited to leave so with a few minutes that I have left I just want to make some important acknowledgments to people who helped make this work a reality so first to Christian so he was a first year PhD student when I was finishing my PhD and we found out about each other when we started requesting the same tissues from the same museums because we both realized the exciting capacity thats Nora had for answering questions about mimicry and polymorphism and so instead of getting territorial we combined forces and work together and it’s ended up being a fantastic collaboration with each of us bringing very different strengths to the table my postdoc advisor Jimmy McGuire I’m hosted me at Berkeley and helped add a really important Southeast Asian Cappetta fauna view to some of our trade evolution studies and also when you’re struggling through the difficulties of ancestral state reconstruction it’s really helpful to have a macro evolutionary model model are sitting at your kitchen table but I also want to say highlight some work by two of our eb graduate students pascal title who was instrumental for the construction of the of the range maps 4,000 new world snake species and also iris holmes who contributed to some of the genetic analyses that went into my work and especially thanks to the EV department as a whole you guys have all been incredibly supportive and i thank you so much for coming today i think i have a couple minutes for questions no question yeah I but the question was about if the color polymorphism might be for evading predators so I would say probably the prediction in that situation is that you shouldn’t have a conspicuous polymorphism so you could have all different kinds of cryptic phenotypes but what if problem is once you get one that kind of looks like it like a stop sign but is perfectly palatable to predators and the idea is that once they eat one they learn that that’s a really great thing to eat and then they can very easily easily specialized their search image and take them all out of the population so I would say probably the best predictors is how conspicuous time yep so I would probably I mean for some butterfly groups I think we do know geographic range is pretty well but I think if you wanted to know like well the ranges of all the butterflies that are in the Amazon I you know if the data don’t exist for snakes they don’t exist for butterflies and so I think it’s a really exciting test and I think with more investments in digitizing our invertebrate collections as a natural history community yeah I think that’s a

really important test to conduct and I’m really excited to see what that looks like compared to the results that I found its name so within oh definitely they’re totally they even get to all four works within a single clutch yeah absolutely no no reproductive isolation that we can find buy more so we tested that and I and I haven’t analyzed the data yet so some of you might remember Kinsey Brock she was a master student SN re and so she had a summer here before she went under her ph.d program and so we had a giant red tailed hawk that was mounted that we swooped in on the snakes that we had in the lab to look at escape behavior and she videotaped it all and no one’s gone through all the video yes but yeah we were interested in that because I’m they’re definitely some classic papers on correlational selection between color pattern and behavior and so I think that’s a really great question so that’s a great question too and I would say we don’t have a whole lot of data on exactly which birds are eating snakes so predation events are pretty difficult to see in the wild and so most of the data we have at least for Raptors comes from a place up on the Snake River where they have a long-term Raptor study and they go and check all the nests and there are little bits of things that get flung out of the net nests and they pick it all up and try to identify it and so we know we have some information about some of the snake or some of the birds that eat snakes but mostly we kind of wave our hands so even Robins for example are known to eat snakes so sometimes it’s not always the birds that we’re expecting and to try to get a really good handle on which are providing them the majority of the selection pressure yeah we don’t know but that’s a good question so it depends how strong you think the selection should be I mean if you’re a population geneticist and you think something that’s that conspicuous and that edible should be taken out pretty quickly by selection within a few generations use the mismatch between between the the coral snakes and the and the mimics even if it’s been a retraction for an expansion was several thousand years ago and so you would expect it to be gone but I would argue that yet it’s hanging on yeah and and that needs to be explained an evolutionary lag sure so at least the Pacific route is a known migratory route for birds I think the fact that you get them going up pretty high into the Great Plains I’d love an ornithologist perspective on migration routes here but I’ve heard that that’s a little controversial about whether that’s actually a migration route or whether Birds mostly follow water so you could absolutely test if you knew bird migration routes you could actually test very very clearly whether or not those overlap one test that was conducted by one of my labmates it at Santa Cruz was to actually compare the response of predators to clay models within the Sierra Nevada mountains and within the Coast Range and those have different proportions of resident versus migratory birds and he found no difference so an attack rates and so it’s it’s hard to say whether the residents are actually behaving in a different way than the migrants but yes the working hypothesis would be that migration could explain at least some of the distributional mismatch but it strikes me that cool think at least a really consistent and like poured all the way down to the Amazon with red and black so I believe that could it just be that the birds that are ranging over large areas learned that red and black think is bad kind of no matter where they are so I didn’t show pictures of the coral snake color polymorphism but it’s rampant so we get because though actually the species that are in the US are pretty conserved we get this idea that that yeah it’s pretty much always banding and we have little rhymes about the order of the bands telling you whether or not it’s a safe snake to touch but no there are coral snakes that are completely red there are coral snakes that are completely black there are coral snakes that are black and white they’re usually some permutation but they can they might not

have bands they could be all black they could be all red and so but but yes I mean a detailed analysis of exactly how many are tricolor verses by color versus unicolor might be able to tease some of that apart but you start to lose power when you start dividing your ad coral snakes up by color pattern so we didn’t do it here but yeah I think that’s an exciting future target thanks guys

You Want To Have Your Favorite Car?

We have a big list of modern & classic cars in both used and new categories.