The Monkey's Voyage

by Alan de Queiroz

  When it comes to explaining distributions broken up by oceans, biogeography has been as fickle as fashion. Back and forth the pendulum has swung, from the likes of Darwin to Schuchert to Simpson to Nelson to the present, from dispersal to vicariance to dispersal to vicariance and, finally, today again toward dispersal (but now including a healthy dose of vicariance). Given this history, why should we think that this most recent development has any more validity than the others? Why should we believe that we’re finally getting it right?

  The answer goes back to the Hawaiian fruit flies and bristletails, the many other examples I’ve described, and, most generally, to all the studies that have piled up over the past twenty years or so: it’s all about following the evidence.

  This may sound presumptuous. After all, weren’t the scientists of previous generations following the evidence too? To an extent, of course, they were, but as the biogeographer John Briggs has said, too often the field has been “beset by attempts to make the facts fit the theory instead of consulting the evidence in the first place.” People adopted biogeographic theories and then saw what they wanted to see, dismissing observations that were inconvenient for their pet hypotheses.

  For example, although the land-bridge advocates were rightfully swayed by Paleozoic and Mesozoic fossils that indicated former land connections (which we now explain by continental drift), they tended to conjure up bridges on the flimsiest of geologic evidence, turning faint hints into the pillars of an argument. Charles Schuchert really did use the existence of granite on Tristan da Cunha, St. Helena, and Ascension to argue for an Atlantic land bridge, although there was in fact no reason to believe that granite could only form within continental crust. In the extreme, land-bridge enthusiasts used similarly weak reasoning to concoct vast land connections to the unlikeliest of places, including Hawaii, Samoa, and other remote volcanic islands. In those cases, facts no longer seemed to matter; everything became subservient to the theory.

  On the surface, it might seem that the vicariance movement of the 1970s was, in contrast, deeply grounded in evidence, but that wasn’t really the case. It is true that vicariance biogeography was founded in part on the conclusive evidence supporting the theory of plate tectonics. However, although plate tectonics provided a compelling mechanism for the fracturing of landmasses, other evidence was required to show that this fragmentation actually explains the piecemeal distributions of living things. To say that vicariance biogeography was based on evidence because it was tied to the validation of plate tectonics is like saying that oceanic dispersal is based on evidence because it’s connected to the undeniable observation that rafts of vegetation can float on water. Identifying a way that something could have happened, while significant, is not the same as showing that is how it did happen; “how possibly” should not be equated with “how actually.”

  As described in Chapter Two, evidence that vicariance is the best explanation for disjunct distributions, including those broken up by oceans, was actually scant. There was Brundin’s wonderful early example of the chironomid midges, and then . . . not much. The American Museum ornithologist Joel Cracraft used Gondwanan breakup to explain disjunctions in several groups, but none of these cases was as convincing as the midges (and some, such as the southern beeches and ratite birds, clearly haven’t held up as pure examples of vicariance). It probably isn’t coincidental that Donn Rosen’s findings on Middle American swordtail fishes and their relatives were analyzed over and over again and became almost legendary among the vicariance crowd; Rosen’s study was an excellent piece of work, to be sure, but one wonders if its importance got blown out of proportion because so few other good examples of vicariance materialized in the 1970s and 1980s.

  If evidence for the ubiquity of vicariance was lacking, then what drove this scientific movement? As noted in Chapter Two, there seem to have been two influences at work: forceful personalities and the inherently attractive nature of the theory itself. The forceful personalities included Gary Nelson, the “hub” at the American Museum; Colin Patterson, the “voice of God” at the British Museum; and Léon Croizat, who was viewed by some as a sort of biogeographic messiah. Take away those three and a few others, erasing both their publications and personal influence, and perhaps there would have been no extreme vicariance movement.

  Beyond this social aspect, however, vicariance may have been for many people an inherently seductive explanation, regardless of the evidence. In particular, it combined two traits that many scientists strive for in their theories: simplicity and generality. The basic idea of areas and their biotas fragmenting in concert was comprehensible to the average eight-year-old and, when coupled with plate tectonics, could be applied to the entire world.57 Tectonic vicariance also had a feeling of almost Newtonian inevitability about it: start with lineages on a landmass, let rifting and seafloor spreading take their course, and the predictable end-­product is vicariance. Long-distance dispersal, in comparison, was a messy proposition. By definition it involved random, unpredictable events—its other names were “chance” and “occasional” dispersal—not to mention a reliance on unverifiable assumptions about things such as the velocity of ocean currents in the Oligocene, or a beetle’s ability to survive a tumultuous, freezing trip in the jet stream.

  Finally, if vicariance was accepted as the dominant process, it led to the notion that the whole geography of life could be distilled into a series of fragmentation events, the sequence of Gondwanan breakup being the archetype. This was perhaps the grandest, cleanest idea of all, because it meant that one could reduce biogeographic history to cladograms, unambiguous branching diagrams that represented these area-splitting events. Hell, it was so clean, it was barely biological. The final product, the general area cladogram, didn’t even have to mention living things, just the names of regions—southern South America, Australia, New Zealand, or whatever. This was the logical conclusion of the dictum that “Earth and life evolve together”; if that was indeed true in the way Croizat had envisioned, then, in a cladogram, areas of the Earth could stand in for the living groups within them. It was so simple, and so general, and so orderly.

  If this interpretation is correct, the rise of vicariance biogeography had more to do with the beauty or elegance of the theory than with any flood of supporting observations. A pretty theory can, at least for a time and for some people, trump the evidence.58 In great contrast, just the opposite was true of the next and latest swing of the pendulum, back toward dispersal. Evolutionary biologists who started their careers in the 1970s, 1980s, and 1990s, especially in English-speaking countries, were often trained to be suspicious of long-distance dispersal as an explanation and to view it as anything but elegant. Even if they weren’t directly connected to the hard-core vicariance school, they had heard the rhetoric about dispersal hypotheses being the product of a muddy, storytelling mindset, that these hypotheses were unfalsifiable and therefore unscientific.

  Then, something happened, namely, the polymerase chain reaction (PCR) created a giant pile of DNA sequence data, which in turn created a flood of molecular dating studies, putting ages on evolutionary branches. Initially, few scientists were thinking that this was going to shift biogeography away from vicariance and toward dispersal. In fact, some of the people doing these timetree studies were coming from the vicariance side and were anticipating that the molecular evidence would just clarify or confirm exactly which fragmentation events had been important for their chosen study organisms. For instance, Miguel Vences had aimed in his PhD project to clarify the vicariant origins of Malagasy frogs, and Matt Lavin had expected to confirm that an Early Tertiary land bridge and a subsequent cooling climate had left related woody legumes on both sides of the Atlantic. As we have seen, though, the clarification and confirmation of vicariance did not materialize. The ages estimated from the DNA sequences were almost always too young, turning both Vences and Lavin away from vicariant explanations and toward long-distance, overwater dispersal.
>
  For other scientists, a loss of confidence in vicariance came before the molecular dating evidence arrived, but ultimately it was this evidence that shifted their thinking, not some prior faith that dispersal had to be the answer. Steve Trewick, who has done some of the key studies on the biogeography of New Zealand and, especially, the Chatham Islands, had what was probably a fairly common experience. “Early on I took the vicariance model as the default,” he said, “and yes, I have to say when I first arrived [in New Zealand for graduate school] I probably did believe it. Why? Because that is what everyone talked about and I assumed that it was based on evidence. Quite quickly it became apparent that there is almost no direct evidence.” Molecular dating studies—his own and others—and geological support for the partial or total drowning of landmasses eventually convinced him that most of the plants and animals of New Zealand and the entire biota of the Chathams has been derived from overseas colonists.

  In short, the shift toward long-distance dispersal hypotheses was driven by evidence, not by any preconceived notions that dispersal was a particularly attractive, elegant theory. If anything, the unpredictable nature of long-distance dispersal and its disconnection from the grand events of Earth history made it an ugly duckling in comparison to vicariance. It was evidence, not theorizing, that turned it into a swan. And it is evidence—a mountain of evidence—that distinguishes this latest swing of the pendulum from the ones that have gone before.

  FINALLY, A PARADIGM?

  In his seminal book, The Structure of Scientific Revolutions, the philosopher Thomas Kuhn describes a typical revolution beginning with some kind of anomaly (or anomalies), that is, an observation or experiment that isn’t well explained by the current paradigm. For instance, under the view that the Sun, the other planets, and the stars revolve around the Earth, it was hard to explain the fact that some planets, when observed over a period of several days, sometimes appear to reverse their direction of movement (so called “retrograde motion”). Such anomalies, if they persist, can lead to the formulation of a new theory that better accounts for them. In this case, the theory that followed was Copernicus’s heliocentrism, in which retrograde motion was explained by the idea that the Earth orbits the Sun faster than do the planets farther out, so that these outer planets seem to move backward as the Earth catches up with and then passes them.59 Eventually, corroboration of the new theory piles up and a “paradigm shift” takes place, with the result that everyone who is considered a serious investigator in the field believes the new theory rather than the old one. The shift occurs partly because individuals are converted, but also because those who cling to the older view eventually die and are replaced by younger investigators who hold the new view.

  To a degree, the shifting views about distributions broken up by oceans follow this model of anomaly leading to revolution. For instance, the frequent occurrence of very closely related fossil groups in widely separated regions—like the Triassic vertebrates of Africa and South America, or the Glossopteris flora on various Southern Hemisphere landmasses—was anomalous under the Darwinian view that distributions had come into being against a background of fixed continents and ocean basins. However, the various swings of the pendulum—from Darwinian dispersalism to land bridges to dispersalism again (the New York School version) to vicariance biogeography—clearly do not fit Kuhn’s view of typical revolutions in at least one way: none of them resulted in a complete or even nearly complete shift of the discipline from the old view to the new. In particular, “dispersalism” persisted through both the land bridge and more recent vicariance movements.

  Kuhn’s views, of course, should not be taken as gospel. It is possible, for instance, that his model of scientific revolutions applies much better to the physical sciences that were his focus than to evolutionary biology. However, what he had to say about situations in which multiple views have been tossed about, without any of them completely taking over the field, seems to describe extremely well the history of explanations of distributions broken up by oceans. If his interpretation of such situations is right, it makes sense of more than 150 years of confusion and disagreement among scientists interested in the geography of life.

  The word to remember here is “pre-paradigm.” Specifically, what Kuhn said is that the period before a discipline adopts its first paradigm is “regularly marked by frequent and deep debates over legitimate methods, problems, and standards of solution, though these [debates] serve rather to define schools than to produce agreement.” This seems a perfect description of the history of biogeography, especially when it comes to explaining piecemeal distributions.60 For instance, dispersalists criticized land-bridge advocates for erecting vast land connections on illusory geological evidence; in other words, the former questioned the latter’s “standards of solution.” Similarly, vicariance biogeographers thought that much of the evidence used by dispersalists, such as the fossil record and Darwin’s seed-survival experiments, was basically worthless. Scientists in every discipline have disagreements, of course, but what is striking about biogeography is how deep the divisions ran, to the point that members of different schools of thought seemed, as Kuhn might have put it, to be living in different worlds. Thus, vicariance scientists thought that the approach of dispersalism was unscientific, because of faulty evidence and because dispersal hypotheses supposedly could not be falsified, and, further, that it was based on the erroneous notion that species and other taxa necessarily spread out from centers of origin. These were not the kinds of minor disagreements that could be resolved with a bit of new data; practitioners of vicariance biogeography were basically saying that dispersalists weren’t even performing legitimate scientific research.

  Kuhn also wrote that “Throughout the pre-paradigm period when there is a multiplicity of competing schools, evidence of progress, except within schools, is very hard to find. This is the period . . . during which individuals practice science, but in which the results of their enterprise do not add up to science as we know it.” It doesn’t add up to science because there is little that becomes an accepted foundation upon which the whole discipline can build. Instead, the different, conflicting schools develop independently, as if they aren’t even dealing with the same subject.

  With this in mind, recall the disdain that Croizat and Nelson had for Darwinian biogeography, describing it as “a world of make-believe and pretense”; clearly, they did not see Darwin’s ideas as any sort of foundation for their own work. Similarly, in the dispersalist William Diller Matthew’s criticism of land-bridge advocates, there was little sense of shared progress between the two schools, apart from the gathering of new information on distributions.

  The disconnect between different biogeographic schools jumps out very baldly in communications I had with the New Zealand botanist Michael Heads and with John Briggs, an emeritus professor at the University of South Florida. Heads, as described earlier, is a staunch proponent of panbiogeography, and Briggs, although he has written books about the importance of plate tectonics in biogeography, has argued strongly for the significance of long-distance dispersal. When I asked Heads, “Who do you consider the most important contributors to the development of biogeography in the last fifty years?” he wrote, in an email: “Apart from Croizat I don’t think there has been much development over the last 50 years. I think biogeography has gone downhill. In many ways the 19th century writers were better informed than the moderns and I enjoy reading them much more. I study lots of molecular phylogenies but don’t usually read the text of the papers as they’re so predictable.” In other words, from Heads’s perspective, the biogeography of the past half-century (except for the work of Croizat and his followers) hasn’t gone anywhere. There hasn’t been any progress to speak of.

  In great contrast, when I asked Briggs about the significance of Croizat’s work, he replied, in no uncertain terms, “I don’t think he made any lasting contribution.” Similarly, Briggs had this to say about cladistic vicarian
ce biogeography: “Cladistics by itself has been a useful systematic procedure but the vicariance inclusion was a distraction that led to a lot of argument and wasted time. Except for stimulating the interest of some who were not acquainted with biogeography, I don’t think the vicariance movement had any lasting benefit.” Thus, again, an entire segment of the discipline—a rather large segment, at that—had been banished to irrelevance.

  In short, Heads thinks that Croizat made the only significant contributions to biogeography in the past fifty years, Briggs thinks that Croizat made no contribution at all, and neither of them seem to have much use for the cladistic vicariance biogeography advocated by Nelson, Rosen, and others. Heads, with his panbiogeography, and Briggs, with his “balanced” view (more or less the view taken in this book, although I don’t agree with his statement about the irrelevance of cladistic vicariance biogeography), are technically researchers in the same field, but they seem to share little common ground from which to move forward.

  According to Kuhn, such fundamental disagreements occur not only in the period before a discipline’s first paradigm has been established, but also when an accepted paradigm falls apart and different views are jostling to take its place, that is, during the first stages of revolution. However, the history of biogeography points to the pre-paradigm explanation: the disagreements have been deep and persistent, basically going back in one form or another to Darwin’s criticism of land bridges, and that kind of lasting discord signifies not a scientific revolution in the making but a field searching for its initial consensus. If this view is correct, historical biogeography has been spinning its wheels for the past 150 years in a pre-­paradigm state, an immature condition in which the most consistent agreement has been that the “other guys”—dispersalists, land-bridgers, cladistic vicariance scientists—were fundamentally wrong in their approach to the science and, therefore, also in their conclusions about the world.