Wonderful Life: The Burgess Shale and the Nature of History

by Stephen Jay Gould

  3.28. The enigmatic Nectocaris, looking mostly like an arthropod in front and like a chordate with a tail fin behind. Drawn by Marianne Collins.

  What can be done with such a chimaera—a creature that looks mostly like an arthropod up front (with possibly unjointed appendages casting some doubt), and mostly like a chordate (or a creature of unknown design) behind? Not much more, when you have but one specimen. So Conway Morris wrote a short, provocative paper and dropped Nectocaris into the great holding bin of taxonomy—phylum Uncertain. The title of a taxonomic paper traditionally lists the broad affiliation of the animal being described, but Conway Morris chose a conspicuously noncommittal approach: “Nectocaris pteryx, a new organism from the Middle Cambrian Burgess Shale of British Columbia.” His final words express no surprise at such a peculiar beast, but hint instead at an emerging generality: “The failure to resolve definitely the affinities of this creature need not be a source of surprise. Current research is showing that a number of species from the Burgess Shale cannot reasonably be accommodated in any extant phylum” (1976a, p. 712).

  2. Odontogriphus. Conway Morris mounted one rung higher on the ladder of evidence with his second treasure of 1976. He still had only a single specimen, but this time he found both part and counterpart. Walcott had at least set Nectocaris aside and supplied a photograph to signal its importance. But Odontogriphus—appropriately endowed by Conway Morris with a name meaning “toothed riddle”—was a true discovery, an entirely unnoted specimen, with part and counterpart in separate sections of Walcott’s collection. Conway Morris began his paper in the conventional passive voice, but his personal pride and passion come through beneath the stylistic cover-up:

  During a search … through the very extensive collection of Burgess Shale fossils … a sawn slab bearing the specimen described here was noticed and set aside for further study. Shortly afterwards the counterpart was found elsewhere in the collections. The specimen had evidently never been noted by any other worker. No other specimens have been found (1976b, p. 199).

  The fossil of Odontogriphus is not well preserved and few structures can be distinguished, but these few are strange indeed. This highly flattened, elongated, oval animal is about two and a half inches long, and marked behind its frontal region with a series of fine, transverse parallel lines, spaced about a millimeter apart. Conway Morris regards these marks as annulations, not separations between true segments. He found no appendages or indications of hardened areas, and assumes that Odontogriphus was gelatinous.

  The body includes only two resolvable structures, both on the ventral surface at the head end (figure 3.29). A pair of “palps” (probably sensory organs) occupies the corners of the animal’s front end. These are shallow rounded depressions formed by up to six platelike layers of tissue parallel to the body surface. The more interesting feature, presumably a mouth surrounded by a feeding apparatus of some kind, lies just forward of the palps, but right in the midline. The structure has the form of a shallow, squashed U, opening toward the front. Along the trackway of this U, Conway Morris found some twenty-five “teeth”—tiny pointed, conical structures less than half a millimeter in length. Since these teeth were far too small and fragile to rasp or bite, Conway Morris made the reasonable conjecture that they acted as supports for the bases of tentacles, and that the tentacles, serving as food-gathering devices, surrounded the mouth in a ring.

  Such a ring of tentacles would strongly resemble a lophophore—the feeding structure of several modern phyla, notably the bryozoans and brachiopods. Hence, Conway Morris tentatively placed Odontogriphus among the so-called lophophorate phyla. But no modern lophophores grow internal teeth to support their tentacles, and nothing else about Odontogriphus recalls the form or structure of any other lophophorate animal. “Toothed riddle” remains a fine designation.

  Those who follow high-risk strategies must accept the embarrassment of error with the joys of chancy victory. Simon’s decision to publish on the rarest and oddest specimens, and to range widely in his interpretations, almost guaranteed some significant mistakes. These come with the territory, and are not badges of dishonor. Simon “made a beauty,” as we Yanks used to say, in trying to judge the wider implications of Odontogriphus. He couldn’t help noticing that its “teeth” bore a vague resemblance to conodonts, then the most enigmatic objects of the fossil record. Conodonts are toothlike structures, often quite complex, that occur abundantly in rocks spanning the great geological range from Cambrian to Triassic (see figure 2.1). They are among the most important of all fossils for geological correlation, but their zoological affinities had long remained mysterious, thus fueling the most famous and long-standing of all paleontological puzzles. Obviously, conodonts are the only hard parts of a soft-bodied animal. But the creature itself had never been found—and what can you tell from some disarticulated teeth?

  3.29. The flattened swimming animal Odontogriphus. The mouth surrounded by tentacles and the pair of palps are shown on the underside of the head. Drawn by Marianne Collins.

  Conway Morris thought that the “teeth” of Odontogriphus might be conodonts, and that, perhaps, he had discovered the elusive conodont animal. He even took a chance and placed his toothed riddle in the class Conodontophorida. What a potential coup for a beginner—to discover the secret of secrets, and resolve a century of debate! But Simon was wrong. The soft-bodied conodont animal has since been found—with undeniable conodonts lying just in the right place at the forward end of the gut. This creature was also discovered in a museum drawer—in a collection made during the 1920s from a Carboniferous Lagerstätte in Scotland known as the Granton Sandstone. The conodont animal, now ranking as one of the few post-Burgess oddballs, looks nothing at all like Odontogriphus. Derek Briggs participated in the original description and thinks (though I am not convinced) that the conodont animal may be a chordate, or member of our own phylum (Briggs, Clarkson, and Aldridge, 1983).

  3. Dinomischus. Simon’s third mystery animal carried him another rung up the ladder of evidence. Again, Walcott had set aside and photographed a specimen, but published nothing and left no notes. But this time Conway Morris found himself wallowing in a virtual sea of evidence, for he had three specimens—Walcott’s in Washington, another in our collection at Harvard, and a third discovered on Walcott’s talus slope by the Royal Ontario Museum in 1975.

  All animals discussed so far have been mobile and bilaterally symmetrical. Dinomischus represents another major functional design: it is a sessile (fixed and immobile) creature with radial symmetry, suited to receiving food from all directions, like many sponges, corals, and stalked crinoids today. Dinomischus looks much like a goblet attached to a long thin stem, with a bulbous holdfast at the bottom to anchor the animal to the substrate (figure 3.30). The entire creature scarcely exceeds an inch in length.

  The goblet, called a calyx, bears on its outer rim a series of about twenty elongate, parallel-sided blades, called bracts. The upper surface of the calyx contains both a central and a marginal opening, presumably mouth and anus by analogy with modern creatures of similar habits (figure 3.31). A U-shaped gut, with an expanded stomach at the base, runs between the two openings through the interior of the calyx. Strands radiating from the stomach to the inner surface of the calyx may have been suspensory fibers (for the gut) or muscle bands.

  A number of superficial similarities may be noted with bits and pieces of various modern animals, but these are probably broad analogies of similar functional design (like the wings of birds and insects), not detailed homologies of genealogical connection. Conway Morris found closest parallels with a small phylum called the Entoprocta (included with bryozoans in older classifications), but Dinomischus is basically a bizarre thing unto itself. Conway Morris showed some hesitation in his original paper (1977a, p. 843), but his latest opinion is unequivocal: “Dinomischus has no obvious affinity with other metazoans and presumably belongs to an extinct phylum” (Briggs and Conway Morris, 1986, p. 172).

  4. Amisk
wia. With Amiskwia, Simon finally tackled a mainstream Burgess organism, though one of the rarest. Five specimens had been discovered, and Walcott had formally described the genus—as a chaetognath, or arrow worm—in 1911. Amiskwia had also been a source of some published debate, though none outside the accepted framework of homes within modern phyla. Two articles in the 1960s had suggested a transfer from the chaetognaths to the nemerteans. These phyla are not household names, but both are staples of modern taxonomy.

  3.30. Original reconstruction of Dinomischus by Conway Morris (1977a). Part of the calyx is broken away to show the interior anatomy of the organism. Note the U-shaped gut going from the mouth (labeled M.) to the anus (An.), and the muscle bands (Sus. Fb., for “suspensory fibers”) anchoring the gut to the wall of the calyx.

  3.31. Three specimens of the stalked animal Dinomischus. One bends toward us, showing the openings of the mouth and anus on the top of the calyx. Drawn by Marianne Collins.

  Amiskwia, as a compressed, probably gelatinous animal with no outer carapace, did squash flat on the Burgess rock surfaces. Hence, these fossils are truly preserved in the mode that Walcott incorrectly viewed as normal for all Burgess organisms—as a flat sheet. Without the three-dimensional structure that Whittington found for arthropods, and that Simon confirmed for several other oddballs, little of Amiskwia’s anatomy can be well resolved—though enough has been preserved to preclude a place in any modern phylum.

  The head region bears a pair of tentacles, inserted on the front ventral surface (figure 3.32). The trunk sports two fins, unsupported by rays or any other stiffening device, in the plane of body flattening—lateral (at the sides) and caudal (forming a tail). (The chaetognaths often have fins in roughly similar positions, hence Walcott’s designation. But a true chaetognath also has a head with teeth, hooks, and a prominent hood—and no tentacles. Nothing else about Amiskwia even vaguely suggests chaetognath affinities, and the rough similarity of fins represents separate evolution for similar function in swimming.) Amiskwia is probably one of the few Burgess animals that did not live in the bottom community engulfed by the mudslide. It was presumably a pelagic (or swimming) organism, living in open waters above the stagnant basin that received the Burgess mudslide. This different mode of life would explain the great rarity of Amiskwia, Odontogriphus, and a few other creatures that may have lived in open waters above the grave, but away from the original home, of the main Burgess community. Only a few animals of the water column above would have died and settled into the sediments below during the short time when the mudslide was coalescing into a layer of sediment in the stagnant basin.

  3.32. The flattened swimming animal Amiskwia, with a pair of tentacles on the head, and side and tail fins behind. Drawn by Marianne Collins.

  Within the head, a bilobed organ may represent cerebral ganglia, while the gut can be traced as a straight tube from an enlarged region at the head to an anus at the other end of the body, just in front of the caudal fin (figure 3.33). The head, lacking the characteristic proboscis with a prominent fluid-filled cavity and muscular walls, looks nothing like that of a nemertean—the other candidate for a conventional taxonomic home; while the caudal fin exhibits only superficial similarity (in nemerteans, the fin is bilobed, and the anus opens at the very tip of the body). Conway Morris, now becoming quite comfortable with the idea of taxonomic uniqueness at high anatomical levels, concluded:

  While Amiskwia sagittiformis is certainly not a chaetognath,.… the worm cannot be placed within the nemerteans either. The relative similarity … [to nemerteans] is regarded as superficial and merely a product of parallel evolution. Amiskwia sagittiformis does not appear to be more closely related to any other known phylum (1977b, p. 281).

  5. Hallucigenia. We need symbols to represent a diversity that we cannot fully carry in our heads. If one creature must be selected to bear the message of the Burgess Shale—the stunning disparity and uniqueness of anatomy generated so early and so quickly in the history of modern multicellular life—the overwhelming choice among aficionados would surely be Hallucigenia (though I might hold out for Opabinia or Anomalocaris). This genus would win the vote for two reasons. First, to borrow today’s vernacular, it is really weird. Second, since names matter so much when we are talking about symbols, Simon chose a most unusual and truly lovely designation for his strangest discovery. He called this creature Hallucigenia to honor “the bizarre and dream-like appearance of the animal” (1977c, p. 624), and also, perhaps, as a memorial to an unlamented age of social experiment.

  3.33. Reconstruction of Amiskwia by Conway Morris (1977b). (A) Bottom view: note the insertion of the tentacles (labeled Tt.), the position of the mouth (Mo.), the path of the gut (Int.) to the anus (An.), and the structure interpreted as possible cerebral ganglia (Ce. Ga.). (B) Side view.

  3.34. Hallucigenia, supported by its seven pairs of struts, stands on the sea floor. Drawn by Marianne Collins.

  Walcott had assigned seven Burgess species to Canadia, his principal genus of polychaetes. (Polychaetes, members of the phylum Annelida, the segmented worms, are the marine equivalent of terrestrial earthworms, and are among the most varied and successful of all animal groups.) Conway Morris later showed (1979) that Walcott’s single genus was hiding remarkable disparity under one vastly overextended umbrella—for he eventually recognized, among Walcott’s seven “species,” three separate genera of true polychaetes, a worm of an entirely different phylum (a priapulid that he renamed Lecythioscopa), and Hallucigenia. Walcott, mistaking the strangest of all Burgess creatures for an ordinary worm, referred to this oddball as Canadia sparsa.

  How can you describe an animal when you don’t even know which side is up, which end front and which back? Hallucigenia is bilaterally symmetrical, like most mobile animals, and carries sets of repeated structures in common with the standard design of many phyla. The largest specimens are about an inch long. Beyond these vaguest of familiar signposts, we are forced to enter a truly lost world (figure 3.34). In broad outline, Hallucigenia has a bulbous “head” on one end, poorly preserved in all available specimens (about thirty), and therefore not well resolved. We cannot even be certain that this structure represents the front of the animal; it is a “head” by convention only. This “head” (figure 3.35) attaches to a long, narrow, basically cylindrical trunk.

  Seven pairs of sharply pointed spines—not jointed, arthropod-like appendages, but single discrete structures—connect to the sides of the trunk, near the bottom surface, and extend downward to form a series of struts. These spines do not articulate to the body, but seem to be embedded within the body wall, which extends as a sheath for a short distance along the top of each spine. Along the dorsal midline of the body, directly opposite the spines, seven tentacles with two-pronged tips extend upward. The seven tentacles seem to be coordinated with the seven pairs of spines in an oddly displaced but consistent way: the first tentacle (nearest the “head”) corresponds to no spine below. Each of the next six tentacles lies directly above a pair of spines. The last pair of spines has no corresponding tentacle above. A cluster of six much shorter dorsal tentacles (perhaps arranged as three pairs) lies just behind the main row of seven. The posterior end of the trunk then narrows into a tube and bends upward and forward.

  3.35. Original reconstruction of Hallucigenia by Conway Morris (1977c).

  How can a taxonomist proceed in interpreting such a design? Simon decided that he must first try to figure out how such an animal could operate; then he might gain some further clues to its anatomy. Searching for analogies, Simon noted that some modern animals rest upon, and even move with, spines attached to their bottom sides. “Tripod” fish support themselves upon two long pectoral spines and one tail spine. The elasipods, a curious group of deep-sea holothurians (sea cucumbers of the echinoderm phylum), move in groups along the bottom, supported by elongate, spiny tube feet (Briggs and Conway Morris, 1986, p. 173). In Hallucigenia, the two spines of each pair meet at an angle of some seventy degrees, an excell
ent arrangement for a series of struts supporting the body in fair stability. Conway Morris therefore began by supposing that the seven pairs of spines permitted Hallucigenia to rest on a muddy substrate. This assumption defines both a mode of life and an orientation: “Dorsal and ventral surfaces are identified on the assumption that the spines were embedded in the bottom sediments” (Conway Morris, 1977c, p. 625).

  So far, so good; Hallucigenia could rest on the bottom in fair stability. But the animal couldn’t stand there in perpetuity like a statue; bilaterally symmetrical creatures with heads and tails are almost always mobile. They concentrate sensory organs up front, and put their anuses behind, because they need to know where they are going and to move away from what they leave behind. How in heaven’s name could Hallucigenia move on a set of spikes fixed firmly into the body wall? Conway Morris did manage to suggest a plausible model, in which strips and bands of muscle anchor the proximal end of the spine to the inner surface of the body wall. Differential expansion and contraction of these bands could move the spines forward and back. A coordinated wave of such motion along the seven pairs might propel the animal, if a bit clumsily. He was not thrilled with the prospects for such a mode of locomotion, and suggested that “Hallucigenia sparsa probably did not progress rapidly over rocks or mud, and much of its time may have been spent stationary” (1977c, p. 634).