Wonderful Life: The Burgess Shale and the Nature of History

by Stephen Jay Gould

  3.16. Camera lucida drawing of the key specimen of Marrella that settled the major problem in reconstructing the head anatomy. Only this specimen shows the two pairs of appendages (labeled a1 and a2) attached separately to the head shield.

  Consider now the dilemma that Whittington faced as he began to compose his monograph on Marrella. He took for granted the old view that fossils fall within major groups and that life’s history moves toward increasing complexity and differentiation. Yet Marrella seemed to belong nowhere. Whittington had found that the legs of the body segments are not sufficiently trilobite-like to warrant classification in this group. He had established a sequence of head appendages—two pre-oral and none post-oral—not only unlike the one pre-oral and three post-oral of trilobites but also completely unknown among arthropods. What was he going to do with Marrella?

  Today, this situation would cause no problem. Harry would simply smile and say to himself—ah, another arthropod beyond the range of modern groups, another sign that disparity reached its peak at the outset and that life’s subsequent history has been a tale of decimation, not increasing variety in design. But this interpretation was not available in 1971. The conceptual cart could not push this lead horse; in fact, the cart hadn’t even been constructed yet.

  In 1971, Harry was still trapped in the concept that Burgess fossils, as old, must be primitive—either generalized members of large groups that later developed more specialized forms, or even more distant precursors that combined features of several groups and could be interpreted as ancestors to all. He therefore toyed with the idea that Marrella might be a kind of precursor for both trilobites and crustaceans—trilobites for the vague similarity in leg structure, crustaceans for the characteristic two pairs of pre-oral appendages on the head shield. (A weak argument even in its own terms, for Whittington had shown important differences in detail between the legs of Marrella and those of trilobites, while crustaceans also have three post-oral appendages on the head shield, and Marrella has none.) Still, Whittington was stuck with a conventional notion of primitivity, and he could offer no more to Marrella. He wrote: “Marrella is one of the fossils indicating the existence of an early arthropod fauna, characterized by serially uniform, generally trilobite-like limbs … and by a lack of jaws, features associated with particle and detritus feeding” (1971, p. 21).

  But Whittington still had to classify Marrella. Again, a quandary, for Marrella possesses unique features that violate the key characters of every group of arthropods. Harry, on the brink of a transforming insight, chose caution and tradition this one time—and placed Marrella in Størmer’s Trilobitoidea, as the title of his monograph proclaims. Yet, as he did so, he felt the pain of betraying his own better judgment. “I had to put something at the top,” he told me, “so I put ‘Trilobitoidea.’” Yet, in the interval between submitting his manuscript and receiving printed copies, Whittington realized that he would have to abandon Trilobitoidea as an artificial group, a “wastebasket” hiding the most interesting story of arthropod evolution. He said to me: “When I saw Marrella printed with ‘Trilobitoidea’on top, I knew it was a bust.” But Marrella, in fact, had been the beginning of a boom—and the documentation of this anatomical explosion would soon transform our view of life.

  Yohoia: A SUSPICION GROWS

  On his cautious journey through the Burgess arthropods, Whittington meant to proceed in order of abundance. Canadaspis stood next in line, but Harry wanted a research student to handle the entire group of arthropods with bivalved carapaces (Derek Briggs would do this work with brilliant results, as Act 3 will show). Burgessia and Waptia, the two genera that Størmer had united as his subclass Pseudonotostraca, followed in terms of abundance. But Whittington had allocated these genera to his colleague Chris Hughes (who published a study of Burgessia in 1975, but has yet to finish his work on Waptia). Hence, Whittington tackled the next most abundant arthropod (with some four hundred specimens)—the interesting genus Yohoia, namesake of the national park that houses the Burgess Shale.

  Whittington’s second monograph, his 1974 study of Yohoia, marks a subtle but interesting transition in his thinking, a necessary step toward the major transformation to come. Whittington had struggled with Marrella, and had come to the correct empirical conclusion—that this most common Burgess genus fits into no known group of arthropods. But he lacked the conceptual framework for thinking of Burgess organisms as anything other than primitive or ancestral—and he certainly had no inclination to construct a new guidepost for only one example that might not be typical. But one is an oddity, and two a potential generality. With Yohoia, Whittington made his first explicit move toward a new view of life.

  Yohoia is a very peculiar animal. It looks “primitive” and uncomplicated at first glance (figure 3.17)—an elongate body with a simple head shield, and no funny spines or excrescences. Walcott had placed Yohoia among the branchiopods, Størmer as an uncertain genus tacked to the end of Trilobitoidea. Yet, as Whittington proceeded, he became more and more puzzled. Nothing about Yohoia fitted with any known group.

  The preservation of Yohoia left much to be desired by Burgess standards, and Whittington had difficulty resolving the order and arrangement of appendages—a crucial factor in arthropod taxonomy. He finally decided that the head probably bears three pairs of uniramous walking legs—nothing unconventional here, since this is the standard trilobite pattern and consistent with Størmer’s placement in Trilobitoidea. But the most curious anomaly of all stands just in front—the large pair of grasping appendages, composed of two stout segments at the base and four spines at the tip. This design is unique among arthropods, and Whittington found no name in the panoply of available jargon. With elegant simplicity, he opted for the vernacular and called this structure the “great appendage.”*

  3.17. Reconstruction of Yohoia by Whittington (1974). Note the unique great appendage (labeled rga and lga) attached to the head.

  Yohoia bears no other appendages on its head shield—no antennae, † no feeding structures (the so-called jaws and mouth parts of insects and other arthropods are modified legs—the main source for our feelings of bizarreness or discomfort when we view films of enlarged insects eating). The first ten body segments behind the head bear lobate appendages fringed with setae, or hairlike extensions (figure 3.18; see also figure 3.17). The appendage on the first segment may have been biramous, including a walking leg as well—but Whittington was not able to resolve the appendages satisfactorily due to poor preservation. Segments 11–13 are cylindrical and carry no appendages, while the last, or fourteenth, segment forms a flattened telson, or tail. Again, this arrangement of segments and appendages departs strongly from the standard trilobite pattern of biramous limbs on each body segment. Yohoia, with its great appendage in front, and curious arrangement of limbs behind, was an orphan among arthropods.

  Whittington (interview of April 8,1988) remembers his study of Yohoia as a turning point in his thinking. He had assimilated Marrella, despite its uniquenesses, under the two reigning p’ s—” primitive” and “precursor.” But Yohoia forced a different insight. This basically simple, elongate animal with many segments did have a primitive look in some respects. “This animal,” he wrote, “resembles Snodgrass’hypothetical primitive arthropod in that the alimentary canal extended the length of the body” (1974, p. 1). But Whittington did not shunt the uniquenesses aside, particularly the form of the great appendage. He had attempted a reconstruction of Yohoia as a working animal—showing how the lobate body appendages with their setal fringes might have been used for swimming, for breathing (as gills), and for transporting food particles; and how the great appendage might have captured prey with its spiny tips and then folded back to bring food right to the mouth.

  All these features were unique anatomical specializations that probably helped Yohoia to work in its own well-adapted way. This animal was not a precursor with a few oddities, but an entity unto itself with a mixture of primitive and derived characters. “In t
he exoskeleton and appendages,” Whittington wrote, “Yohoia tenuis is clearly specialized” (1974, p. 1).

  Thus, as the crucial year 1975 dawned, Whittington had completed monographs on two Burgess arthropods with the same curious result. Marrella and Yohoia didn’t fit anywhere—and they were specialized animals apparently living well with their unique features, not simple and generalized creatures from the dawn of time, ripe for replacement by more complex and competent descendants.

  3.18. Yohoia. Drawn by Marianne Collins.

  3.19. The fateful first expression of doubt. Whittington (1974, p. 4) still placed Yohoia in the Trilobitoidea, but expressed his doubt about the status of Størmer’s group.

  Whittington remained too cautious to translate these suspicions into hard taxonomy. He still, and for one last time, placed Yohoia in Trilobitoidea, but with two crucial differences. He did not use Størmer’s category in the title of his monograph, and he inserted a fateful question mark after the designation in his formal taxonomic chart (1974, p. 4)—the first overt sign of challenge to the old order (figure 3.19). Whittington wrote: “I am doubtful whether Yohoia should be placed in Trilobitoidea” (1974, p. 2). Never doubt the conceptual power of a question mark.

  ACT 2. A New View Takes Hold: Homage to Opabinia, 1975

  Harry Whittington began his 1975 monograph on Opabinia with a statement that should go down as one of the most remarkable in the history of science: “When an earlier version of figure 82 [reproduced here as figure 3.20] was shown at a meeting of the Palaeontological Association in Oxford, it was greeted with loud laughter, presumably a tribute to the strangeness of this animal” (1975a, p. 1). Are you baffled by my claim? What is so unusual about this inoffensive sentence that doesn’t even abandon the traditional passive voice of scientific prose? Well, you have to know Harry Whittington, and you have to be steeped in the traditions of style for technical monographs. Harry, as I have stated many times, is a conservative man.* I doubt that he had, in all the several thousand pages of his output, ever written a personal statement, much less an anecdote about a transient event. (Even here, he could bring himself to do so only in the passive voice.) What, then, could possibly have persuaded Harry Whittington to begin a technical monograph in the Philosophical Transactions of the Royal Society, London with a personal yarn that seems about as fitting in this format as Kareem Abdul-Jabbar in Lilliput? Something really unusual was about to happen.

  3.20. Reconstruction of Opabinia by Whittington (1975). (A) Top view, showing the five eyes on the dorsal surface of the head. (B) Side view: note the orientation of the tail fins relative to the body; the dorsal surface is at the right.

  In 1912, Walcott had described Opabinia as yet another branchiopod crustacean. Its curious design, particularly the bizarre frontal nozzle (figure 3.21), had made Opabinia a center of Burgess attention. Many different reconstructions had been attempted, but all authors had found a place for Opabinia within a major group of arthropods. Opabinia, as the most puzzling of all Burgess arthropods, stood as a challenge and a logical next step for Harry Whittington after two monographs on common genera (Marrella and Yohoia), and one on the structure of trilobite limbs (1975b).

  3.21. Opabinia, showing the frontal nozzle with terminal claw, five eyes on the head, body sections with gills on top, and the tail piece in three segments. Drawn by Marianne Collins.

  Whittington began his study of Opabinia without any doubt about its status as an arthropod. He soon received the surprise of his life, though the lesser oddities of Marrella and Yohoia had prepared him for astonishment from the Burgess. Whittington presented his first reconstruction of Opabinia to the annual meeting of the Palaeontological Association* in Oxford in 1972.

  Laughter is the most ambiguous of human expressions, for it can embody two contradictory meanings. Harry recognized the laughter of his colleagues at Oxford as the sound of puzzlement, not derision—but it really shook him up nonetheless. Both Simon Conway Morris and Derek Briggs, his two superb students, agree that this Oxford reaction marked a turning point in Harry’s work on the Burgess Shale. He simply had to resolve and diffuse that unanticipated and incongruous laughter. He had to overwhelm his colleagues with a reconstruction of Opabinia so incontrovertible that all its peculiarities could pass into the realm of simple fact, and never again disturb the courts of science with the spirit of Milton’s L’Allegro:

  Haste thee nymph, and bring with thee

  Jest and youthful Jollity, …

  Sport that wrinkled Care derides,

  And Laughter, holding both his sides.

  Although Opabinia is a rare animal with only ten good specimens (Walcott found nine, and the Geological Survey of Canada added another in the 1960s), Walcott established its importance as a centerpiece in interpreting the Burgess fauna. He awarded Opabinia pride of place, describing this genus first among the Burgess arthropods (see table 3.1). Walcott put Opabinia at the head of his classification because he regarded the elongate body, composed of many segments without prominent and complex appendages, as “very suggestive of an annelidan ancestor” (1912, p. 163). Since the Annelida, or segmented worms (including terrestrial earthworms and marine polychaetes), are the presumed sister-group of the Arthropoda, an animal that combined characters of the two phyla might stand close to the ancestry of both and act as a link between these great invertebrate groups. To Walcott, Opabinia was the most primitive Burgess arthropod, the closest model for a true ancestor of all later groups.

  But what arthropod features did Walcott discern in Opabinia? He had little to offer for the head, since he could find no appendages. The frontal “nozzle” might be interpreted as a pair of fused antennae, and the eyes were consistent with arthropod design (Walcott noted only two eyes, but Whittington found five—two paired and one central). Walcott admitted that “none of the heads … show traces of antennules, antennae, mandibles or maxillae. If these appendages were large they have been broken off; if small they may be concealed beneath the crushed and flattened large posterior section of the head” (1912, p. 168). I regard this statement as a lovely example of apparently unconscious bias in science. Walcott “knew” that Opabinia was an arthropod, so the animal had to have appendages on its head. Since he didn’t find any, he provided explanations for their absence—either they were so large that they always broke off, or they were so small that they became hidden beneath the head. He never even mentioned the obvious third alternative—that you don’t see them because they didn’t exist.

  (Walcott, by the way, also made another error—see the next paragraph—that may seem merely amusing or tangential but underscores the serious point that we observe according to preset categories, and often cannot “see” what stares us in the face. A set of empirical anomalies may have instigated the Burgess revision by Whittington and colleagues, but as we shall see, the conceptual framework of the new view, coalescing between 1975 and 1978, established a novel context that allowed further observations to be made. I preach no relativism; the Burgess animals are what they are. But conceptual blinders can preclude observation, while more accurate generalities guarantee no proper resolution of specific anatomies, but can certainly guide perceptions along fruitful paths.)

  Walcott, following our primal biases of gender, found two specimens that appeared to lack the frontal nozzle. (Walcott thought that the nozzles were truly absent on these specimens, but Whittington later proved, by dissecting one of the specimens and finding the jagged edge of the break point, that the nozzles had been broken off.) On one specimen, Walcott found a slender, two-pronged structure in the same location as the nozzle. (This turned out to be a fragment from an unrelated worm, but Walcott interpreted it as a genuine part of Opabinia, in the same position as the nozzle of other specimens.) Walcott therefore concluded that he had discovered sexual dimorphism in Opabinia: the strong and stout nozzle belonging to the male (naturally), and the slender structure to the more delicate female. He wrote that these supposed females “differ from the male … in having
a slender, bifid frontal appendage instead of the strong appendage of the male.” He even foisted the stereotypes of active and passive upon his fictitious distinctions, arguing that the nozzle “was probably used by the male to seize the female” (1912, p. 169).

  Walcott’s main justification for regarding Opabinia as an arthropod lay in his interpretation of the paired body segments. He read these flaps as the gill branches of ancestrally biramous appendages. He thought that he had observed two or three “rather strong, short joints” (1912, p. 168) at the base of each flap, followed by the broad lobe bearing the gills. He hoped to find the inner leg branches as well, but he could never fully persuade himself, and eventually concluded that the walking legs probably existed in an “insignificant or rudimentary” form (1912, p. 163).

  Walcott was clearly troubled by the failure of Opabinia to preserve any smoking gun of arthropod affinity. He even took some modern anostracans and crushed them between plates of glass, trying to simulate the conditions of Burgess fossilization. This mayhem provided some solace, because such treatment often destroyed all evidence of the delicate appendages. He wrote: “After flattening specimens of Brachinecta and Branchipus between plates of glass and studying them, I am greatly surprised that any distinct characters of the appendages are preserved in the fossils in a recognizable condition” (1912, p. 169). Walcott had shown the cardinal skill of his adopted profession—administration. He had put the best face upon adversity. Opabinia would remain an arthropod.