The Monkey's Voyage

by Alan de Queiroz

  Even so, the conversion was not far off. Within a few years, extensive new data from the ocean floor and from lava flows on land validated the idea of reversals in the Earth’s magnetic field, one of the underpinnings of Vine and Matthews’s work. Far more complete magnetic recordings across several mid-oceanic ridges were published, with stunning “zebra-­stripe” diagrams showing many bands of rock alternating between the current field orientation and its reverse, the patterns on the two sides of the ridge looking like mirror images. Where the theory predicted that tectonic plates should be sliding against each other (as the Pacific and North American Plates are doing at the San Andreas Fault), seismic studies showed exactly that kind of motion. The formation of chains of oceanic volcanoes, such as the Hawaiian Islands and the seamounts extending northwest from them, was elegantly explained by the movement of a plate over a hotspot, a region where a stationary plume of magma rises up from the mantle.

  2.4 Wegener vindicated: reconstructions of the globe emphasizing the breakup of Gondwana. AF = Africa, AN = Antarctica, AU = Australia, IN = India, SA = South America. Redrawn and modified from Scotese (2004).

  Wegener, had he survived (he would have been in his mid-eighties at the time), might have been especially pleased to see two new developments. One was a computer-generated fit of the continents bordering the Atlantic Ocean, based on a mathematical theorem for the movement of a rigid plate on the surface of a sphere. The fit was remarkably good and remarkably similar to the reconstructions of Pangea that Wegener had done by eye. The other was a study looking at the continuity of rock formations in North America and Europe, based on similarity in the age of the rocks as measured by radiometric dating, and using the new computer-­generated map as a foundation. The result was a striking alignment of the formations on opposite sides of the Atlantic. It was Wegener’s matching “lines of print” argument all over again.

  In November 1966, a symposium on continental drift was held in New York City, with many of the most influential geologists in North America present. By the end of that two-day meeting, it was clear that the new theory had won over the doubters. When the scientist who was supposed to summarize arguments against seafloor spreading and continental movement was called away on other business, no one bothered to stand up in his place. The weight of evidence—more than fifty years after publication of Wegener’s book, almost forty since Holmes’s paper on convection in the mantle, and four since Hess’s “History of Ocean Basins”—had finally brought down the building and raised a new one.

  I have, admittedly, spent a long time telling the story of the validation of plate tectonics and continental drift. Partly that’s because I like the story and find Harry Hess, in particular, a sympathetic figure. He seems to have been that rare thing, a brilliant but modest scientist. (Unfortunately, he was also that not-so-rare thing, a heavy drinker and smoker, and died of a heart attack in 1969, when he was sixty-three. Unlike Wegener, however, he at least lived long enough to see his theory widely accepted.) Obviously, plate tectonics is also a vital part of the vicariance revolution, providing a mechanism for fragmentation on a global scale. Beyond both of those points, though, there is a message in the prolonged resistance to the notion of continental drift that will be echoed in other parts of this book. The message is this: Acceptance of theories in science, even theories that we think do a great job of explaining the evidence, is not a straightforward process; the clean, textbook version of how science proceeds is almost invariably misleading. Often, things are said that are ignored or outright rejected, even ridiculed, yet in hindsight turn out to be exactly right (at least to our current understanding). And we’re not just talking here about Galileo or Copernicus, but about science in the twentieth and twenty-first centuries. In the case of continental drift, Wegener, Holmes, Hess, Vine, and Matthews all presented theories and/or evidence that we now see as substantially correct and crucial, yet their work was either ignored or, as Vine said about his paper with Matthews, went over like a lead balloon. As we shall see, notions about oceanic dispersal have had a similar history of rejection and ridicule, especially in the past few decades. In fact, if anything, “ridicule” is too soft a word to describe some of the things that have been said (and continue to be said) about long-distance dispersal.

  THE NEW NEW YORK SCHOOL

  When Lars Brundin was writing his midge monograph, plate tectonics had not yet been widely accepted. Holmes and Hess had published their theories, but the definitive zebra-stripe magnetic orientation diagrams and other evidence were just coming out. Brundin, though, was not a man afraid of new ideas. He saw that continental drift could explain the phylogenetic patterns he was seeing in the midges, saw, in fact, that the explanation of those patterns almost had to be drift. It probably also helped that he was from Europe, where the new geological theory was gaining traction faster than in the United States.

  In plate tectonic reconstructions of the past locations of landmasses, both New Zealand and Australia were connected in the Mesozoic to southern South America via parts of Antarctica. That was the key to explaining why Brundin’s midges showed transantarctic relationships. Antarctica had been at the center of the southern part of Gondwana and thus shared with those other areas a good part of its biota, apparently including the midges. When the land connections were eliminated by continental drift, and Antarctica was transformed into an isolated, largely midge-free block of ice, the tiny flies were still left in New Zealand, Australia, and South America. More than this, the connection between New Zealand and South America, through West Antarctica, had been severed well before the connection between Australia and South America, which had been linked through East Antarctica. The expectation, then, was that in a cladogram that included midges from New Zealand, South America, and Australia, those from New Zealand would branch off first, reflecting the sequence of geological breakup (see Figure 2.5). To put it in cladistic terms, New Zealand midges would be sister to a group that included both South American and Australian midges. This is exactly what Brundin found, and not just once, but over and over, in different subgroups within the Chironomidae. The repetition was important because a pattern seen just once, even if it matched the history of Gondwanan fragmentation, could have been a coincidence, and might still have been explained by long-distance dispersal; the chironomid midges showed the pattern far too frequently and consistently for that explanation to hold water.

  What Brundin had done was the first really impressive study in modern vicariance biogeography. The ancient land connections were there to be read, and Brundin, having had the patience to look at the pupal characteristics of hundreds upon hundreds of midges and build cladograms from them, had found the patterns that revealed those connections. He had shown that the transantarctic relationships of midges were explained by the fragmentation of southern Gondwana. And, moreover, he had found “not the slightest evidence of chance dispersal over wide stretches of ocean.” Specifically, because New Zealand and Australian taxa were never sister groups, there was no indication that midges had ever colonized one of these areas from the other by crossing the Tasman Sea that lay between them.

  2.5 Vicariance meets cladistics: Lars Brundin’s scenario for the geographic history of chironomid midges. Upper: distribution of a group of midges through time in the context of the breakup of New Zealand (NZ), Australia (AU), West Antarctica (WA), East Antarctica (EA), and South America (SA). Initially, the midges occurred widely in New Zealand, West Antarctica, and South America. Subsequently, New Zealand broke away from the other landmasses, and, later still, midges colonized Australia from South America by normal, “garden-variety” dispersal (arrow). In the present, midges are absent from Antarctica because of the severe climate, but occur in the now widely separated landmasses of New Zealand, South America, and Australia. Lower: the evolutionary trees at the same three stages in the group’s history. Note that Brundin’s scenario does not require any long-distance overwater dispersal, only normal dispers
al coupled with the tectonic separation of landmasses.

  Unlike some of the biogeographers he inspired, Brundin did not see long-distance dispersal as something that either never happens or is impossible to study.14 For instance, he wrote, “No one denies . . . that the transporting capacity of gales, hurricanes, and tornados is impressive and that small insects can cross wide expanses of sea with the aid of air-borne dispersal.” Nonetheless, he had sharp words for what he saw as the fuzzy and misguided approach of the likes of George Gaylord Simpson, Ernst Mayr, and P. J. Darlington—the dispersalists of the “New York School.” He chastised them for their ignorance about phylogenetic relationships, that is, for not understanding how to build cladograms and not realizing that this was the key to understanding biogeographic history. He scolded them for clinging to the assumption that the positions of continents and ocean basins are fixed, or that, if continental drift had occurred, it had happened too long ago to be relevant for modern groups. And he was condescending about their facile reliance on chance dispersal. “Several troubled biogeographers,” he wrote dismissively, “have found consolation and relief in the thought of a raft with a sufficient store of food and water put at the disposition of a pregnant female at the right moment.” These improbable dispersal stories smacked of deus ex machina and seemed to Brundin “comparable with a confession of failure.” Brundin’s venom, his disdain, poured off the page: “There is something negative, sterile, and superficial involved in the above [dispersalist] approach which offends a critical mind,” he wrote. His own emphasis was instead on the general pattern revealed by phylogeny, typically a pattern produced by the fragmentation of areas, that is, by vicariance.

  I am trying to imagine what it must have been like for the young Gary Nelson, reading Brundin. The strengths of Brundin’s work were many to someone open to new ideas. By the standards of the time, Brundin’s Hennigian approach to inferring phylogenetic relationships was a great step forward, and he was clearly right in saying that phylogeny was the key to understanding biogeographic history. The case of the midges, with their repeated phylogenetic pattern matching Gondwanan breakup, also showed the power of the resuscitated theory of continental drift for biogeography. All of that must have impressed Nelson. On top of that, the whole thing was radical, a slap in the face of the grand old men of the New York School, who seemed to be doing everything wrong where Brundin was doing it right. Brundin even made that affront an explicit point, writing, “Biogeography has a great future. But the rate of progress is strongly dependent on due distrustfulness of authorities.” I imagine that this attitude struck a chord with the rebellious young Nelson. Take up the cause of the fresh and clear-minded, or stick with the muddled establishment. Seems like an easy choice. Hell, I’m neither young nor particularly rebellious, but reading Brundin now I almost wish I had been there too, ready to join the revolution.

  As it turns out, Brundin was working in the Swedish Museum of Natural History, where Nelson read the midge paper, so Nelson had a chance to meet him soon afterward. Brundin was then in his late fifties, a thin, somewhat severe-looking man. Nelson found him to be polite, maybe more so than one might have expected from his sharply worded monograph. Oddly enough, according to Nelson, the two of them didn’t really talk about science, either then or on the few other occasions when their paths crossed. Nonetheless, Nelson had discovered a great intellectual influence, one that he has carried with him ever since. Not long ago I emailed him and, in a follow-up, he wrote, without any provocation from me, “If you have not read the first 50 pages of Brundin [i.e., the midge monograph], I would suggest you do. He is much clearer than Hennig; and unlike Hennig, he has potent critiques of the dominant paradigms of systematics and biogeography.” (There might have been a subtext to that message too, namely, that I was misguided and needed to read Brundin to straighten myself out—Nelson knew that I thought chance dispersal was important.)

  Nelson was on a path now, some would say a warpath. As a maven, he had acquired the special information in Brundin’s monograph; now he would switch to his roles of connector and salesman. His first stop, while still on his research fellowship, was the British Museum of Natural History in London. During a weeklong visit there, he discussed Brundin’s approach with three other ichthyologists: Humphry Greenwood and Colin Patterson, who were the resident curators, and Donn Rosen, who was visiting from the American Museum of Natural History. All of them were smokers, so, to avoid the possibility of setting the alcohol-preserved fish collections on fire, they would meet outside of the building under the colonnade to smoke and chat. At the time, Nelson was the only one who had read Brundin; thus, despite the fact that he was the youngest and only unestablished member of the group, he was in the role of teaching the others about cladistics and its implications for studying biogeography. How far he convinced them during that visit is a little unclear, but even if he only made them look into Brundin’s work for themselves, he had planted a seed. Not long afterward, both Patterson and Greenwood became prominent converts, despite the fact that accepting cladistics invalidated much of the research they had already done. Patterson, in particular, who was known for his dramatic talks delivered in an impressive baritone—the “voice of God,” as it has been described—would become a fervent advocate of cladistics and vicariance biogeography.

  Returning to the United States, Nelson began a job as a curator at the American Museum of Natural History in New York, ironically the very place from which William Diller Matthew and George Gaylord Simpson had spread the dispersalist view against which Brundin was now railing. Nelson wasted no time in continuing his evangelical work. He started by continuing his discussions with Donn Rosen, haranguing his new colleague as they walked the halls of the museum’s research area. Other staff members closed their doors to shut out the loud, heated arguments reverberating in the corridors. Like Patterson and Greenwood, Rosen had already published research papers that would be rendered obsolete by cladistics. But no matter. He eventually saw the light and became one of Nelson’s closest intellectual allies and a valued confidant. Within the next few years, many others at the museum followed, all of them becoming cladists, and most also adopting the related approach of vicariance biogeography as defined by Hennig and Brundin and in later papers by Nelson, Rosen, and others.

  They were definitely onto something, as Brundin’s midge work had indicated. The importance of using phylogenetic relationships in an exact way quickly became obvious to most scientists studying biogeography, even to old-school types like Ernst Mayr. (However, Mayr ceded this point only grudgingly, writing, “The Hennigian methodology may be more rigorous than the frequently rather superficial analyses of earlier authors, but the basic approach is the same.” In this he exemplified a common reaction to a new theory or approach: claim that your side had it more or less right all along.) Also, by the late 1960s, plate tectonics had become the accepted paradigm in geology, and no one could doubt that continental drift had influenced the distributions of living things, although people continued to argue about just how pervasive that influence had been. Already the dispersalist P. J. Darlington’s assessment of continental drift, from his 1965 book Biogeography of the Southern End of the World, sounded quaint and ridiculous: “I have therefore become a Wegenerian, but not an extreme one. I doubt the former existence of a Pangaea or Gondwanaland, and I think that the movements of continents have been simpler and shorter than most Wegenerians suppose.” Darlington seemed to imply that the continents had always been apart, as they are today, just not quite so far apart. He was, of course, completely wrong.

  One thing vicariance biogeography did not quickly produce was a wealth of clear examples to show that the fragmentation of some area, by plate movements or other processes, explained disjunct distributions. For instance, thumbing through Systematics and Biogeography, the thick 1981 book written by Nelson and an American Museum arachnologist, Norm Platnick, one is struck by how few real-world cases of vicariance are included. Through t
ree diagram after tree diagram, the authors show how to construct cladograms and use them in biogeographic studies, but the book gives little reason to believe that fragmentation actually explains many piecemeal distributions.

  There were a few compelling examples, however, to add to Brundin’s work on the midges. One came from Nelson’s ichthyological buddy Donn Rosen, who had been studying swordtails and other guppy relatives in Middle America since before his conversion to cladism. Looking at two groups of these fishes, Rosen found that the geographic histories implied by the cladograms of the two groups were strikingly similar. Specifically, when one replaced the fish taxon names on the cladograms with the areas in which each taxon was found—producing what are known as area ­cladograms—the patterns were largely the same for the two groups. Rosen then took his analysis a step further by calculating the probability of obtaining the observed level of agreement by chance and found that the probability was very low. The exact probability he came up with, 1 in 105, was later shown to be inaccurate, but the general point he made seemed reasonable to most biogeographers: the similar area cladograms of the two groups suggested a common history of range fragmentation rather than chance dispersal, which would have produced a more haphazard pattern. A hole in Rosen’s example was that it wasn’t clear what geological or other events had fragmented the fishes’ distributions and given rise to the matching histories, but he could at least point to some possible causes, such as the rise, in the Pliocene, of an east-west belt of volcanic mountains in Mexico (the range that includes Mexico’s highest peaks, such as the Pico de Orizaba and Popocatépetl). His case was strengthened by the fact that at least parts of his fish area cladograms matched those of other, completely unrelated groups, such as box turtles and red-bellied snakes. The common cause of vicariance should affect many organisms in the same way—the uplift of a new mountain range, for instance, would isolate populations of many species on the two sides—and that was just what Rosen seemed to be seeing.