Community ecology: Size, space, distributions

♪ Music ♪ In this second of three lectures on community ecology, Dr. Dan Simberloff overviews theories of community composition and changes in community composition over space and time He draws on historical and contemporary empirical examples of studies that have contributed to ideas about how closely related species develop within a community over time including theories of minimum and equal size ratios and character displacement and release He also discusses displacement in morphospace rather than time and presents extensive discussion and examples of community assembly rules from checkerboard to allopatry patterns He concludes the discussion about the need to distinguish between the presence of species, the number of each species present, and the distribution of species within a community Another version of too similar to coexist has to do with the sizes of coexisting species and the idea was for, most prominently the early expression by Hutchinson who’s often viewed as – he was MacArthur’s advisor actually and he’s viewed as a towering figure in the history of modern ecology, there’s a biography Dr. Peter Grant applied this to islands arguing that, well I’ll show what this was about So here’s that famous paper we discussed yesterday that one of the great titles of scientific papers, “Homage to Santa Rosalia, or Why Are There So Many Kinds of Animals?” So it was all about this underlying question that community ecologists kept asking, why is it community have x species instead of 4x species? And why do many communities have so many species? And what Hutchinson did here was he looked at a series of species that are congeneric, that is pairs or in one case trios of species in the same genus, all of which have ranges such that their ranges largely do not overlap, we call that allopatric, but they overlap in one part of the range of each, so one might be in the east and one might be in the west, and there might be a zone of overlap, which we call sympatry and I’ll probably lapse into the jargon of my jargon, and say allopatry and sympatry He looked at the ratio of measurements of the food gathering organ of these species where they were sympatric, where they were together and where they were allopatric as where the same species, the parts of their range where they’re not coexisting with the same species, so these are weasels, this is the least weasel and the short tailed weasels, these are field mice, Apodemus, these are two birds rock nut hatches, and I’ll refer to them again They turned out to be important in another development and these are Galapagos finches And he didn’t stress so much the fact that they were more different where they were sympatric than allopatric, although he recognized it, but the reason that – one of the main reasons that he became famous is that he noticed these values for the size ratios of the food gathering apparatus when they were sympatric and he said well look, the mean is 1.28 and so he annunciated something which came to be called Hutchinson’s Rule or Hutchinson’s Law, sometimes, that where they co-occur, a mean ratio is roughly 1.3, so this is one of these sort of magic numbers in ecology If they’re 1.3 or more similar in size, the ratio of one over the other, they can’t make it and if they’re greater they can So this is the idea of minimum size ratios and other people have looked at other morphological traits, so here’s a very nice paper by Mike Bower and Jim Brown on rodents in the southwestern deserts and said well if you look at local communities and which species, there’s a lot of rodents, but local communities always have some subset

of them and they said well if you look closely at the sizes and here are the sizes with nice pictures of dead rodents next to them of four of these local communities, they’re not similar in size ever, they said But for them size was body mass and the ratio in their particular case was 1.5, so it’s sort of a version of Hutchinson’s rule, but with a different trait and so a different number So these are all about like the minimum size ratios that can be, that two species can have and still coexist in the same community and not be too similar There’s a somewhat different idea involving size ratios and it’s captured nicely in this paper by C. Levitt Smith on fishes in a coral reef in the Caribbean You know, here he lines them all up and he shows from smallest to largest, and his argument, he wasn’t the first to point to a case like this, but this was a nice one, was that at least for most of the range of these sizes there’s a very regular progression and at the size ratios are, tend to be equal, that is the ratio of the smallest to the next smallest is the same as size ratio of the next smallest to the next smallest etc., etc., etc It breaks down at the very end and he explains why that might be, but the idea of equal size ratios and the underlying ideas, well however many species are present, competition between them will cause them to be different They will either evolve to be different, if they weren’t already different or they won’t make it or if you view it as sort of the local community having constant comings and goings of individual species and the regional species pool, the ones that make it will be fortuitously the ones that are sort of equidistant between two existing species and so the size, of course there’ll be less competition, so it’s a competition driven idea Levitt Smith as I said was not the first one to point this out, neither was Jared Diamond, but his is the most famous case, so it’s a very important book in community ecology, Ecology and Evolution of Communities published in 1975 edited by Martin Cody and Jared Diamond, and this, one of the papers was 130 pages long and that was Diamond’s paper, and it was all about how birds coexist or do not coexist on islands in Melanesia and the Bismarck Archipelago and in the Solomon Islands especially And one of his two main observations or contentions is that where species in the same guild or the same genus, by guild he meant species that are very similar in how they get their resources, that’s a concept that’s, you know, widely used in ecology; mostly he applied it to species in the same genus assuming they would be in the same guild, a couple cases he defined larger guilds including more than one genus But here he didn’t just have one branch, but these, so he repeated that different birds, and here he had all these birds that are pigeons eating fruit and here he has the weights of each one And his observation was the size ratio of this one to this one is the same as the size ratio of this one to this one, pretty much, which is again almost the same as the size ratio of that one to that one etc., etc., so equal size ratios His, he related this to the idea that they’re using different foods and that’s how they coexist and they’re eating fruit, they all eat fruit of different sizes and he, this is a picture he drew; there’re no points and he never published any data to support this idea, but this is what he said they eat and so that competition for food forces them to

diverge enough so that the size ratios are constant And here with one of my students Bill Boecklen, who’s now a professor at New Mexico State, we looked at both of these ideas – minimum size ratios and equal size ratios and the odd thing was until we did this, no one did any statistical test to ask whether these ratios are any different than you would’ve expected if their points were independently distributed on the line So like if there were five species and you random, well each one individually randomly uniform in, we may distribute on the line, you know, would the size ratios be more or less equal than they are and would you have minimum size ratios any larger or smaller than you are? And we looked at all the papers that did this, and almost none of them except for this rodents in the southwestern desert were any different than you would expect for if the species were completely independent of one another The size ratios were exactly, in most cases, almost exactly what you’d expect if they were independent, so it’s not really a rule of nature at all This is not to say that some cases aren’t very specially arranged or, you know, that not everything can be explained, not every case can be explained by just independence among the species, so with a former post-doc of mine Tamar Dayan, who’s now a professor at Tel Aviv University, we looked at those Heteromyidae rodents that Bowers and Brown had weighed and we looked at them in two different deserts and the Heteromyides are usually put in one guild; their special thing is they have these external cheek pouches and they’re all granivores, they ate seeds, and what they do is they pick up the seed and then they very rapidly remove the seed coat and they put it in their external cheek pouches and carry it to some place and store it And they’re mostly kangaroo rats and pocket mice, there’s also a kangaroo mouse, and they all in any one place in southwestern deserts you can find a number of species of these, and we got some films of them eating, of four species of these, both pocket mice and kangaroo rats, and we slowed the film down to see exactly how they did this I mean if you watch them do it, it’s very fast, pick up a seed, that’s it, you know, you can’t see anything, but you slow it down, what you see is that they use their incisors, the two incisors as a chisel really and they use that to, they chisel a groove around the seed coat and throw it out, and then they take the seed, but they do it really, you can only see it if you slow the film down a lot, but all four of them were doing exactly the same thing So here’s two of the kangaroo rats with their chisel-like incisors, so our thought was well if they’re really competing and they’re competing for the size of the seed, which is what Brown and Bowers thought they were probably doing, you should see something about the size ratios of the incisors, and so we measured a whole bunch of things that people have measured in these size ratio studies, so we measured body weight as BW, greatest total length of the skull, condyle-premaxillae length is another skull measurement, these are tooth row areas then tooth row length, mandible length; yeah these are the areas and these are the lengths of the tooth row, which they used in the end they use that to grind up some of the seeds, anyway, and this is a log scale line, so it’s a logarithmic line, so because it’s a logarithmic line equal spacing would be the same as equal ratios, right Because the log of A over B is the log of A minus the log of B, and the capital letters above the line are the kangaroo rats – this is just for the Chihuahuan Desert; we also did the Great Basin desert and found the same thing And the small letters below the line are the pocket mice

And, you know, if you look through this even without doing and we did statistical and there are, I won’t describe it, but part of the, a large part of the toolkit that community ecologists came to use involves statistical tests of patterns, and so we either, well we found in the literature tests for things like equality of ratios and even without doing the test, you can see that for incisor width there that really you’d never independently throw these points out on a line and get anything like this I mean surely you can see that and for all the other traits, there’s nothing like that There’re pairs of species, sometimes several pairs of species that are too similar; it’s only that one trait that, and I said, you know, Brown and Bowers and others have suggested they’re probably competing for the size of seeds, you know, so some cases, you know, if applied judiciously you find something striking that fits with the idea of limiting similarity and in this case equal size ratios manifesting competition The island’s case was developed by Peter Grant following Hutchinson’s idea of minimum size ratios and here Peter talks about the percent difference between the two species, but in other cases he talks about the ratio of one to the other, it’s the same thing, the difference in the real size, and his point here is well for islands the percent difference or the size ratio is greater than on the mainland in most cases, you know, twice as many cases as vice versa You look at lots of pairs of species that are found or sometimes sister species where one is on the island, the other’s on the mainland and they both coexist with the same species, so his idea was that islands because they’re resource poor the competition must be more intense and so they have to be more different from one another, but, you know, Bill and I showed, but you, if you think about it, so here’s the island and here’s this, this is the size say of Bill’s, which is what Peter usually used, but, you know, islands always have fewer species than the mainland, that’s one of the famous patterns that you can actually see in nature Islands have fewer species than mainland, and it should be obvious that if you randomly draw some subset of points from these existing points, of course the average is going to be greater, right, if you have just two or three of them, of course the size ratio or percent there is going to be greater So it turns out when you actually look at those same data, they’re pretty much what you would expect You don’t find a percent difference or the size ratios to be on average greater on islands In some cases you even find them to be less than, but the whole distribution essentially, you don’t really think there’s a pattern Anyway this idea of minimum size ratios is related to another idea that actually preceded it and that Peter cited in his size ratio paper, and this is the idea of character displacement and release, and this was published first in, I think it was 1954 by Bill Brown and Ed Wilson and it’s a very prominent paper in the evolutionary literature as well as in the ecology literature, and this returns to those rock nut hatches that were also among the cases looked at by Hutchinson in Homage to Santa Rosalia So here they are and here’s the range of one and here’s the range of the other and so there’s a narrow zone of overlap and this is a geographic, line through, these are geographic patterns from west to east, and

here they’re looking at bill length and their observation was that if you look at allopatric population of species, populations from parts of the, where they’re not coexisting with the other species, they’re very similar, and where they’re together is that they’re displaced, they diverged, and in their paper they gave many, many examples of this kind of thing with a lot of different taxa, birds were not the only one And so they had examples from amphibians, from mammals, from birds, from insects, etc And they gave two explana – and they called the phenomenon character displacement and it’s, in the opposite, you know, what happens, that’s what happens here and when they’re separate it’s character release as it’s released from whatever’s happening here, and they gave two explanations for it – one is competition that if they’re too similar, they can’t coexist, so they evolve to be different in the different, where they’re together sympatrically and that’s the only way they can coexist and that’s the only way they can actually have overlapping ranges and their other explanation for some cases was a reproductive character displacement cause if they’re similar they might hybridize and yet they’re different species, so the offspring they would either, there would be no offspring or the offspring themselves would be sterile and so it’d be a waste of reproductive effort and so they’d be selected to be different in order not to hybridize And most of the literature on character displacement since then has been on competitive character displacement, which is the idea for these two rock nut hatches There’s lots and lots of papers on character displacement Peter Grant pointed out a really important fact and it’s surprising it took, you know, about 20 years before someone pointed it out, but he did He said well, you know, just the fact that they’re different where, more different where they’re together than where they’re apart wouldn’t by itself mean they’re influencing one another and he gave an example like this where for each species, this is what we call clinal variation, and as morphology varies in some way, in this case it would be some measure of size with geography for whatever reason and all ecologists and evolutionists know many cases of clinal variations So well suppose Peter’s idea, well suppose you had these two species and both of them had parallel clinal variation and they overlapped in this area, well what would you see? You’d see that in the area of overlap they’re more different than allopatric populations are, and that wouldn’t in his view indicate character displacement unless you thought the cline itself was caused by competition say, and he said well, you know, if enough of the same cline is seen outside the range of coexistence, you wouldn’t think that the cline itself is caused by the interaction between them, so I said you really have seen more than that, and his, this is from one of his two papers and he said you’d have to see probably something like that or it could be tilted, but you’d have to see some change, right, in the pattern for each of the species exactly where they become sympatric Yeah, well character displacement, so there’re lots and lots of papers showing or purporting to show, and many of them actually do show real character displacement that sometimes can be fairly cogently attributed, I think, too Okay, so that’s really a primitive version, these sort of size ratio characters, but at the very least ideas, of an idea that’s often viewed as sort of an advance on that, a separate idea, but it’s really only an extension of it and that’s the idea of constructing what’s called a morphospace We look at lots of features of the morphology of a species

or of a population of a species, and you construct an N dimensional space for which, and N is the number of features you’ve measured and then you use some sort of factor analysis like principal components is often used to reduce the dimensionality of the space and then you locate each species or population in that morphospace, and this is sort of a tool, you know, Margaret asked us about tools, what tools are used? This is a tool that’s used in a lot of community analyses now, not only for this idea, but for some others, so somehow you construct N dimensional space that captures the probably functional or morphological essence of a species; you reduce it by factor analysis and then you look at the relationship of the locations of the species and the idea is they can’t be too close, so it’s really still a version of Darwin’s naturalization hypothesis The underlying hypothesis if they’re too close in that world of space, they’re not going to make it The first major paper of this sort, if you don’t count the size ratios, which would be like a one-dimensional morphospace was by Rob Ricklefs and Joe Travis, and here you can see the, they measured eight things on a bunch of birds and they reduced the space by a factor analysis, they found three factors explain much of the variation, then they located each one of the species in a local community in this N dimensional space, and since in those days they didn’t have great readily accessible software to show N dimensional spaces, they did it comparing each factor by each other one and they separated them by the different species, by different families of birds, so here’s six families of birds And they look to see where these species were in the space and their general conclusion was that they’re pretty far apart; that really captures I guess the conclusion of that paper, and then they extended it to look at islands versus mainland and they looked at sets of island species by the same method, and the mainland from which they were thought to have come, and you can see their results So they viewed it as species packing again on small islands greater morphological distance between nearest neighbors, so the statistic they used in that case was nearest neighbor in the morphospace, greater on the islands than on the mainland, and they thought it was probably because of the, here’s the species interaction, but they think that it’s competition More important on an island, cause islands are resource poor, but if you think again about what I said with respect to the size ratios, there’re more species on the mainland than islands, so if you take any one of these families like these flycatchers here and if this were a mainland and you took only two of them to represent an island, of course the average nearest neighbor distance, if you did that randomly, oh it would be greater So it doesn’t, the pattern is there, but the interpretation doesn’t automatically make sense Moulton and Julie Lockwood did something a little different that does make more sense although there were some problems with it They just looked at introduced species in Hawaii and Hawaii is a spectacular uncontrolled experiment People have introduced about 85 species of birds to Hawaii No one’s ever happy with species, you know, so they – The Asians in Hawaii introduced Asian birds, Europeans introduced European birds, North Americans introduced North American birds, even birds from Australia, and of these 85 or 86 birds, around 45 or 50 have survived and they just looked at finches of these four families and they looked on Oahu; I don’t remember why they looked at 25 and 23 species, but anyway what they looked at

is which ones survived and which ones didn’t And so here, the open circle, open squares represent in morphospace and here they did a principal component analysis to reduce the dimensionality of space The ones that survived are the open squares and the ones that didn’t are the closed black circles, and their argument is well look, if you randomly drew 15 survivors from all these points and you took some statistic that captures how far apart they are, they used a statistic that was then available called minimum spanning tree, which is the single line that’s the shortest that connects all the points in a set, called minimum spanning tree in this software So that you would, these are over-dispersed compared to random sets of the same number of points, the open squares and they were right The difficulty here is there’s some native birds that compete with them on Oahu and they didn’t include those, but, you know, the underlying hypothesis was in a test that made sense, and so they took that as a competitive interaction, they had to be more different from one another in order to all to coexist There’s also a priority effects issues as these birds were, all use different dates and we know the dates they were introduced, and there is some literature on the priority effects on dozens of birds in Hawaii, but I won’t get into it Anyway, so there’re been a number of papers doing versions of this, looking at introduced species and which ones survive and which ones don’t; this is a paper I published with a bunch of colleagues recently where we looked at the fishes that came from the Red Sea into Mediterranean when the Suez Canal was opened; they’re called Lessepsian Migrants after the Lesseps who built the canal And it’s just a staggering change in the fish fauna of the Mediterranean, there’s now almost a hundred Red Sea fishes in the Mediterranean and there’re also a few from the Atlantic who’ve come in, but very few, like eight or so, and I don’t know if you, you know, within the last two months either celebrated the widening of the Suez Canal and deepening it in area called the Bitter Lake, so that more ships can go through, and it’s almost certain we weren’t going to bring more of those fishes, but anyway what we did was we looked at the, looked at about 90 of the fish for which you get really good data We know the dates they first appeared in the Mediterranean, a lot is known for fishes about, about, well like four of these people about exactly what they do, so it was divided into guilds that are widely recognized by ichthyologists, and we asked well the ones that survived, did they tend to be within guilds? On the outside of a morphospace and the way we constructed the morphospace is, it’s not principal components now, so it’s a rather similar technique and we reduce the space and what we then did was for all the species in a group, we constructed what’s called a, well it’s a polygon, it’s called a convex hole, and it’s an N dimensional polygon that has all the points in it and the convex hole is the minimum N dimensional polygon that has all of the points And the next thing we did was we divided that polygon into parts that are also polygons, they’re called Voronoi polygons, each one of which has one species, and it consists, has one species and each species has within its Voronoi polygon all the points in that space that are closest to that point rather than to any other point And so here’s just, you know, one example and we’re expressing it into space, you know, of a convex hole divided into polygons and the point is the actual location in N space of that species and we asked well if a species would fit here, is it less likely to survive

or become common compared to a species that’s on the outside and, you know, so it’s not similar, and, you know, it’s pretty interesting We, here’s one example This species’ become extremely common and a major ecological problem in the Mediterranean, but anyway it’s part of a guild and the red ones are the ones that have become common and the yellow ones are usually not common, they’re never abundant and the green ones didn’t make it at all Here there’s no green ones In this guild, and we know the years they came in, this one there’re only two species, one that survived, but it’s not a major component of any community and this one has just become a horror in the Eastern Mediterranean, now the central Mediterranean Here’s a larger guild and you can see through time what happened, you know, at the beginning it looked like we were right on The red ones, you know, they were all on the outside of the, they were different from all the others and they became really common and then even through 75 we’re in, some others were introduced that were not, that were very similar to existing ones in morphospace and they didn’t become common Here’s one that broke the rule, it’s quite different from any others, it does sort of the same thing, but it didn’t survive it, so you don’t see it in the next, but in a fairly general way the pattern held You had a question? What’s a guild? A guild is a bunch of species that get the same kind of resource in the same way, and in this particular guild, you know, all of them forage in the same way and they’re eating about the same kind of food There is one other idea, I think that sort of springs from Darwin’s naturalization hypothesis and that’s the idea of species combinations that exist or don’t exist and this also comes from Diamond’s long paper in this book, and what he, in addition to the size ratio the other big idea in this paper is what he called community assembly rules and that is if you have a lot of, in this case islands, each one has a group of species; there’s some combinations that you don’t see and that’s because they’re species that are too similar to one another, they compete with one another too much And the main one of these rules, I mean these are sort of all versions of the same rule is this one, and you see pairs of species never coexist on any size island, those you call checkerboard It’s another one of the great ecological metaphors, you know, there’s like Liebig’s barrel and there’s too many, I once gave a whole lecture on metaphors in ecology Anyway, Diamond used extensive bird data he had gathered from the Solomon’s and the Bismarck Archipelago and from that he derived these rules But and here’s one example that he published in his book These are two species of Macropygia, which are pigeons, they’re both the same genus, and you said well look, they’re found in the Bismarck’s, and some islands have one, some have none, and others have the other species, but none of them have both And, well the question I asked pretty quickly was well suppose they were independently strewn about islands, you have hundreds of species of birds there, what would be the likelihood you’d see a certain number of checkerboard distributions? And we had a problem, was Diamond never would give us the data and never published the data, which species were where, he just gave the examples that he felt made his case, but he told us that we could find published data for what was Vanuatu, which is now, which was then New Hebrides, and we did We found someone else who published a list of which species on which islands, so my then student, Ed Connor, who’s now a faculty member at San Francisco State, and I used that measure, those data, and we constructed a matrix of binary metrics, species and sites, so a one means the species is present on that site, zero means it’s not present, and a checkerboard distribution would be two rows, two species where there’s never,

well this wouldn’t be one cause that site has both, but where they’re complimentary, in other words And we randomized the matrix keeping the row and columns on the same, and we found that in New Hebrides, you know, the number of missing species combinations was exactly what, number of checkerboards was exactly what you would’ve expected, and Diamond said well you shouldn’t just look at all of them, you should look within genera, so we look within genera later and found the same And we could never, ever get the data, but then in 1991, in 2001 this book was published and it didn’t actually have a matrix, but it had the data, if you could find the islands, and with Michael Cones, who’s in my grad student, we spent a long time finding islands on maps It was an incredibly labor intensive job, but we found them and we constructed a matrix and did a randomization, we found a very interesting thing This is also in that same book by Mayr and Diamond; 10,000 years ago when sea levels were lower both of these archipelagos and many fewer islands and lots of islands were connected by land and basically these five big groups here in the Solomons and they were basically four We actually got a hold of the data for these places and looked We found a remarkable thing that the few cases that we found that where they had mutually exclusive distributions, so checkerboards, they weren’t really, the metaphor of checkerboard didn’t fit; they were regionally allopatric in addition to not being found in the same island, they weren’t found in exactly the same region with respect to these groups of islands So here’s just one example this accipiter hawks; five species of accipiters there and there’re five checkerboards, well four of those checkerboards are formed by this one species of accipiter with only found in the rennell group and none of the others have found in the rennell group, and that happens to be the one species that actually came not from New Guinea, but there’s the other four species did or one came across the Pacific, but from Australia So that’s four of the five checkerboards; they’re regionally allopatric, one is here and the others are here It’s not like they’re scattered throughout and this is the fifth checkerboard, one of them is sort of their version of our goshawk, it’s the same, it’s like a subspecies, the same goshawk we have, and this is a Melanesian one, well, you know, this one is found on these islands and this one is found on these islands plus a couple small islands around there and they’re not regionally allopatric because this island was probably connected to those, although some people say it wasn’t; there’s a trench there that’s sort of marginal, but that’s the degree of non-regional allopatry We look at all of the congeneric checkerboards there weren’t many of them, and every case there was regional allopatry And I guess the last thing that I’ll say before we take a break, if you want to take a break, is that throughout this I haven’t had anything about numbers of individuals, it’s just a presence and absence matrix, and there’s not a huge amount of real theory on the numbers of individuals and different species within communities There were three major sort of ideas that that distribution fits one or another kind of distribution especially log series and log normal distribution, but there’s no like real mechanism for it and it was sort of, it wasn’t associated with these other questions in community ecology There’ve been a few attempts to make this not just a binary matrix, but to put numbers in here for relative abundances and I haven’t led too far, you know, the advent of supercomputing allows the manipulation of matrices in ways that weren’t available for, lots of people in other fields have been very interested in binary matrices, so you can figure out, you know, things like how many

checkerboards you would expect the, you know, from other literature, mathematicians have been very interested in that because it relates interesting graphs theoretic questions, but, you know, once you put numbers and there’s not as much theory, it involves, you know, lots of simulations and so I don’t think that’s gone too far yet, but people are looking at it ♪ Music ♪