The Tangled Tree

by David Quammen

  Why had he changed his mind? “Recent work has made more evident the profound differences of organization between bacterial cells and those of other organisms.” Whose recent work? He cited the 1962 paper by Stanier and van Niel. They had added those two words to his scientific vocabulary and those two categories to his view of life: prokaryote and eukaryote. It seemed undeniable now, to Whittaker, among others, that bacteria and single-celled protists are distinct from each other, belonging to utterly different kingdoms, because the cells of protists contain nuclei and the cells of bacteria do not. There were other big differences too, such as the mitochondria and chloroplasts within eukaryotes, the capacity for mitosis, the flagella. Those contrasts, he now conceded, echoing Stanier and his two textbook coauthors, defined “the clearest, most effectively discontinuous separation of levels of organization in the living world.” Whittaker accepted that discontinuity as convenient for defining his first kingdom, Monera, and then essentially ignored it for everything else. The prokaryote condition served a purpose within his “broad classification.” The eukaryote condition didn’t.

  Whittaker’s prickly pear of life, 1969.

  Still, he had implicitly recognized those two major divisions of life, even while explicitly defining five kingdoms. How had such a huge evolutionary leap, from prokaryote to eukaryote, occurred? One good hypothesis, he added, lay in the idea of “ancient cellular symbioses.” At that point, in 1969, Whittaker’s thinking converged with the work of Lynn Margulis.

  44

  Convergence is a theme that will be salient throughout the rest of this book. Convergence is something the limbs on a tree never do—not in the natural world, not in the normal course of growth among oaks, elms, maples, hickories, pines, larches, sycamores, beeches, banyans, baobabs, or other actual trees living in the wild. Limbs diverge, they don’t converge. Branches diverge. Flannery O’Connor, the extraordinary Southern short-story writer and novelist famed for her dark humor and her pet peacocks, once published a story titled “Everything That Rises Must Converge.” It was a grim tale of racism and rage, labeled ironically with those five words, which she had drawn from a hopeful quote in the writings of Pierre Teilhard de Chardin, a French Jesuit and paleontologist. Chardin’s woozy philosophical writings were popular among liberal Catholics (but not at the Vatican) in the late 1950s and the 1960s. Flannery O’Connor was a liberal Catholic. The story title became the book title for a collection released after her early death. It’s an interesting axiom, her everything that rises must converge—not always applicable to biology, but sometimes. The path of Robert Whittaker’s efforts to classify life converged with Lynn Margulis’s arguments on endosymbioses, for instance, and in 1978 the two published a paper together.

  At the heart of their collaboration was a friendly disagreement about how to define Protista, the everything-else kingdom. They proposed a compromise. Their discussion of the particulars was long and technical, reflecting the fact that this paper was first presented (by Whittaker, speaking for both authors) as a keynote address to the 1977 meeting of the Society for Evolutionary Protistology, a group that Margulis had helped found. Margulis knew much more about protists than Whittaker did, having spent much of her career to that point gazing at them through microscopes. But she had liked Whittaker’s 1969 paper on the five kingdoms, shared his frustrations at trying to teach new biological discoveries within old-fangled categories, and chose to throw in with him for this joint effort, which emphasized their agreements—not just about Protista but also about all five kingdoms and the entire enterprise of classification.

  One point they agreed on was that kingdoms of life are hard to define. The lines dividing one kingdom from another are inescapably blurry. This had always been a problem, with some creatures left in the borderlands no matter how those borders were drawn, and the problem hadn’t gone away. Fungi: Are they plants or something else? Blue-green algae: Are they, in fact, algae, or are they bacteria? Sponges: Can they be animals, despite the fact that they don’t move around and have no nervous, digestive, or circulatory systems? Protozoans: They’re not really “zoans,” proto or otherwise. Placozoans: What the hell are those flat little things? Whittaker and Margulis noted that classification is a human endeavor, not an inherent reality of the natural world—it’s a matter of discovery plus decision—and that simplicity is important. You can’t have sixty-one or ninety-three major kingdoms of life, it just isn’t practical, if you want to organize biological knowledge and teach it.

  The second point on which Whittaker and Margulis agreed was that classifications of life should, as far as possible, reflect evolutionary events and relationships. They should be phylogenetic. The third point of agreement was that phylogenetic classification isn’t always possible.

  So they compromised, resigning themselves to the embarrassment of polyphyletic taxa. A polyphyletic taxon is any group, such as a kingdom, that encompasses creatures from more than one evolutionary lineage. The phrase itself seems a bit paradoxical, because taxon means a unit of classification—but it might be classification by convenience (as in the old days before Darwin) rather than classification by evolutionary ancestry. If you chose to call “marine vertebrates” a taxon, for instance, it would be polyphyletic because whales belong to one distinct lineage of descent, sharks belong to another, and saltwater crocodiles to still another. Each of those three, though indisputably a marine vertebrate, shares closer ancestry with some terrestrial animals than the three share with one another. Therefore a polyphyletic taxon can’t be drawn as a tree limb, with all its branches diverging. It’s something else: a human construct, an organizing principle, that allows the convergence of lineages. In this picture, branches come together.

  Some lineages do converge, even in the real world—by endosymbiosis, for instance, Margulis’s favorite process. At the time Whittaker and Margulis were coauthoring their paper, though, that sort of convergence was taken as a rare event. Most biologists considered it anomalous. In later years, and thanks in great part to Carl Woese as well as Lynn Margulis, convergence of genetic lineages would be seen differently. Classification itself would be seen differently. In 1978 the main issue was how to draw lines, between one kingdom and another, one class of creatures and another, that were “natural” insofar as possible yet still orderly and convenient. To the evolutionary classifier, Whittaker and Margulis admitted, polyphyletic taxa are “unwelcome”—but not quite so unwelcome as the prospect of having just too many kingdoms, too many classes, too many small but precisely defined groups.

  This dilemma explains why Whittaker drew his systems—the four kingdoms of 1959, the five kingdoms of 1969—as prickly pear cactuses, not as trees. He had recognized all along, and represented candidly, the complication of polyphyletic groups. You can see this if you look closely at his five-kingdom prickly pear. There are multiple lines, like veins, running from one pad to another. Three lines cross from the Protista pad into the Plantae pad, implying three separate origins of the plant kingdom. Five lines cross from Protista into Fungi: five separate origins of the fungi. Two lines from Protista into Animalia, one of which leads only to the sponge group: indicating that sponges evolved from nonanimals along a lineage separate from that of all other animals. Animalness was invented twice. The two distinct lineages converged upon animal identity, at least as Whittaker (and Margulis with him in 1978) chose to define that identity.

  Contrary to Flannery O’Connor’s title, contrary to Tielhard de Chardin, not everything that rises must converge. But some limbs of the tree of life do. You’ll see much more of this in what follows.

  45

  The years 1977 and 1978 marked a notable stage in the tree-building saga because of Whittaker and Margulis’s paper and one other: the announcement, by Carl Woese and George Fox, in November 1977, of a third kingdom of life. Both these offerings were highly influential, and both imprinted themselves on biology textbooks. The striking thing about them in retrospect is that they have so little in common. They don’t
agree, and they don’t disagree. Five kingdoms, according to Whittaker and Margulis; three kingdoms, according to Woese and Fox—and not just the numbers of kingdoms but also the kingdoms themselves were discrepant. The two pairs of authors were talking past each other. They were worlds apart.

  Fox, the young postdoc who had served as Woese’s main collaborator on the discovery of the archaea and the 1977 paper, had by then left the Woese lab in Urbana and accepted a job, his first independent academic position, at the University of Houston. He was an assistant professor, thirty-two years old, without tenure, and besides teaching and establishing a lab of his own, he needed to publish more work. The sooner he did that, the better his prospects of a career. Fortunately, he was still much engaged with Woese, by telephone and mail, on the effort to classify life-forms, especially bacteria and archaea, and deduce their deep relationships, using the evidence of ribosomal RNA.

  Up in Urbana, in Woese’s lab, another young researcher was filling the role Fox had played, and the data continued to pile up. Woese got more catalogs of 16S rRNA, that very special molecule he had hit upon as his Rosetta stone for the early history of evolution, and those catalogs revealed things that microscopy and biochemistry couldn’t. New patterns emerged. There was more to be said than simply that the archaea are a separate form of life. Woese felt it was urgent that he and his team publish an overview. His informal name for that project, in correspondence with Fox and others, was “big tree.”

  Late in November 1977, just weeks after his Warhol moment on the front page of the New York Times, Woese wrote to Fox in Houston, enclosing the catalog of still another organism, voicing concern about certain aspects of their data-analysis method, and then raising a larger subject: “Please give big tree the top priority. Unless we get that out soon, you will find our credibility will be eroding.” They had based their proclamation of the kingdom Archaea on just four organisms, a shockingly small sample to represent such a large category of life, and he was eager to present more of their data. They needed the overview paper. “Just rough out the text and generate the tree,” Woese told Fox. “That will get the ball rolling. The situation I think is more critical from your point of view than you may realize. If you become controversial to your peers, you’ll find that the scientific friends you’ve made at Houston will tend to become enemies.”

  Soon afterward, they began drafting a paper.

  It would be a highly collaborative publication, culminating a decade of work, coauthored by Fox and Woese and a list of their partners. What “coauthorship” means on a scientific publication is that these people have contributed in their various ways. Some might have done painstaking, dangerous laboratory chores. Some might have donated microbes, from cultures in their own lab, or offered fruitful discussion and ideas. A senior researcher, the group’s leader, most likely established the context, mentored at least some of the team, and provided the funding. Another scientist—call this the primary active researcher, who was possibly a grad student or a postdoc—might have chosen the topic, conferred with the senior researcher on conceptual details and experimental design, and done a good share of the hands-on work. Then somebody, usually just one person or a few, almost certainly including that primary active researcher, would do most of the writing. In the purely biological sciences, the primary active researcher would likely appear as first author; the senior researcher, the mentor and sponsor, would appear as final author, a position known to suggest overarching responsibility and ultimate credit. For the field of biophysics, though, in which Woese had been trained, the opposite was true—and that may have contributed to a small conflict between Woese and Fox.

  In this case, the two men worked closely together on the text. Fox had functioned as primary active researcher in Urbana, handling some of the more difficult tasks, with support from others of Woese’s lab team. Fox had labeled microbes with radioactive phosphorus and extracted their ribosomal RNA. Linda Magrum had cut those molecules into fragments and run the fragments through electrophoresis. Magrum had printed films of the stampeding fragments, from which Woese could deduce their sequences, creating catalogs. It was Fox again who had devised the similarity coefficient by which those catalogs could be analyzed, and he had written the computer code to do it. (That skill derived from his training as a chemical engineer: unlike any other biologist in Woese’s lab, and almost any in America at the time, he knew the computer language Fortran.) He had entered the code and the data onto hundreds of IBM punch cards, state of the art in those days, and run them through a mainframe machine. He had sketched the resulting trees. Now he and Woese produced a series of drafts, which were converted to clean typescript by a departmental typist in Urbana. Some of those early typescripts, which have survived, show many scribbled revisions and handwritten additions by Fox.

  The first page of the first version began with a sentence that was accurate but not catchy: “For at least a century microbiologists have attempted to determine the relationships among the myriad microorganisms that inhabit almost every conceivable niche on the planet.” They could do better. Woese wanted more sense of drama from the get-go. A second version began: “Biology has reached an important turning point with regard to the study of evolution.” Better, but with a pencil, he changed it to: “An important turning point has been reached in the study of evolution.” That put some oomph into his first four words: an important turning point. At the top of the page, too, now appeared Woese’s informal title: “Big Tree.”

  Fox’s fourth version, as the effort progressed, contained twenty-five handwritten pages of new and reworked material. It went to the departmental secretary for typing, with a cover sheet that said: “Fox. Draft. Need ASAP. Put ahead of all my other work. Priority 1.” The tinkering continued until, in a seventh version, the opening sentence was replaced by a new one, its initial four words setting the tone better still: “A revolution is occurring . . .”

  There would be more versions, more back-and-forth of typescripts between Houston and Urbana during 1978 and well into 1979, more changes and tweaks and at least one caustic dispute. But from that point, version seven, the opening paragraph of the Big Tree paper was cast in a form that Carl Woese saw no need to improve:

  A revolution is occurring in bacterial taxonomy. What had been a dry, esoteric, and uncertain discipline—where the accepted relationships were no more than officially sanctioned speculation—is becoming a field fresh with the excitement of the experimental harvest. For the most part the transition reflects the realization that molecular sequencing techniques permit a direct measurement of genealogical relationships.

  This was a sonorous understatement. Measuring genealogical relationships with molecular techniques was what Francis Crick had proposed back in 1958. It opened a new perspective on the evolutionary past, equivalent in revelatory scope to all the fossils in all the museums of the world. It was a revolution not just in bacterial taxonomy but also in something broader: the way scientists understand the shape of the history of life.

  46

  The typescript of version seven also carried a new title: “The Phylogeny of Procaryotes.” It showed a roster of authors, including Ralph Wolfe, Linda Bonen, Linda Magrum, and fourteen others. George Fox’s name led the list, and Carl Woese’s came last, signaling his role as the senior researcher. That order of names would become an issue as vexed as any disagreement over content. Somewhere along the way, Woese had proposed that he should be first author, and that Fox should be satisfied with the honor of last.

  George Fox talked candidly about this disagreement as we sat in the pizza parlor in Urbana. “I told him I wanted to be a first author on that paper. Because, I mean, I typed all those goddamn IBM cards. I made all those tree diagrams, right? Participated in all the discussion.” Then he had gone to Houston and, as new junior faculty, faced some exigent expectations. Thirty-six years after the fact, he seemed to remember those pressures as though he were still in the moment. “Look, I’m in a different university now, right? I�
�m trying . . .” He paused. Trying to establish myself, were the words that went unspoken. “I’m an assistant professor. I need to get tenure, and I need to ultimately become an associate professor and full professor. You know,” he said, “it’s not doing me any good to collaborate with him if I don’t get any serious credit for it.”

  On August 27, 1979, Woese wrote to Fox about several “potential points of conflict.” The first involved Fox’s forceful request for first authorship, an important distinction for a young scientist. “I agreed to that,” Woese wrote, “and still do. However, you should know how I feel, as the matter is touchy.” Then, mindful of all those long hours and days spent blurring his eyes before the light board, gazing at films of galloping fragments, taxing his brain to infer their sequences and recognize their patterns, he unloaded: