THE ECOLOGY OF THE BURGESS FAUNA

  In 1986, a year after his monograph on Wiwaxia, Simon Conway Morris published a “blockbuster” of another type—a comprehensive ecological analysis of the entire Burgess community. He began with some interesting facts and figures. About 73,300 specimens on 33,520 slabs have been collected from the Burgess Shale. Ninety percent of this material resides in Washington, in Walcott’s collection; 87.9 percent of these specimens are animals, and nearly all the rest are algae. Fourteen percent of the animals have shelly skeletons; the remainder are soft-bodied.

  The fauna contains 119 genera in 140 species; 37 percent of these genera are arthropods. Conway Morris identified two main elements in the fauna: (1) An overwhelmingly predominant assemblage of benthic and near-bottom species that were transported into a stagnant basin by the mudslide. Conway Morris inferred, from abundant algae needing light for photosynthesis, that this assemblage originally lived in shallow water, probably less than three hundred feet in depth. He called this element the Marrella-Ottoia assemblage, to honor both the most common substrate walker (the arthropod Marrella) and the most common burrower (the priapulid worm Ottoia). (2) A much rarer group of permanently swimming creatures that lived in the water column above the stagnant basin, and settled amidst the animals transported by the mudslide. Conway Morris called this element the Amiskwia-Odontogriphus assemblage, to honor two of his pelagic weird wonders.

  He found that the Burgess genera, despite their odd and disparate anatomies, fall into conventional categories when classified by feeding style and habitat. He recognized four major groups: (1) Deposit-feeding collectors (mostly arthropods)—60 percent of the total number of individuals; 25–30 percent of the genera. (This category includes Marrella and Canadaspis, the two most common Burgess animals, hence the high representation for individuals). (2) Deposit-feeding swallowers (mostly ordinary mollusks with hard parts)—1 percent of individuals; 5 percent of genera. (3) Suspension feeders (mostly sponges, taking food directly from the water column)—30 percent of individuals; 45 percent of genera. (4) Carnivores and scavengers (mostly arthropods)—10 percent of individuals; 20 percent of genera.

  Traditional wisdom, with its progressionist bias and its iconography of the cone of increasing diversity, has viewed Cambrian communities as more generalized and less complex than their successors. Cambrian faunas have been characterized as ecologically unspecialized, with species occupying broad niches. Trophic structure has been judged as simple, with detritus and suspension feeders dominating, and predators either rare or entirely absent. Communities have been reconstructed with broad environmental tolerances, large geographic distributions, and diffuse boundaries.

  Conway Morris did not entirely overturn these received ideas of a relatively simple world. He did, for example, find comparatively little complexity in the attacking and maneuvering capacities of Burgess predators: “It seems plausible that the degree of sophistication in styles of predation (search and attack) and deterrence in comparison with younger Paleozoic faunas was substantially less” (1986, p. 455).

  Still, his primary message made the ecology of the Burgess Shale more conventional, and more like the worlds of later geological periods. Over and over again, when the full range of this community could be judged by its soft-bodied elements, Conway Morris found more richness and more complexity than earlier views had allowed. Detritus and suspension feeders did dominate, but their niches did not overlap broadly, with all species simply sopping up everything edible in sight. Rather, most organisms were specialized for feeding on particular types and sizes of food in a definitely limited environment. Suspension feeders did not absorb all particles at all levels in the water column; the various species were, as in later faunas, “tiered” in assemblages of complex interaction. (In tiering, various forms specialize, confining themselves to low, medium, or high level of the water column, as communities diversify.) Most surprising of all, predators played a major role in the Burgess community. This top level of the ecological pyramid was fully occupied and functioning. No longer could the disparity of early form be attributed to reduced pressures of an easy world, devoid of Darwinian competition in the struggle for existence, and therefore open to any contraption or jury-rigged experiment. The Burgess fauna, Conway Morris argued, “shows unequivocally that the fundamental trophic structure of marine metazoan life was established early in its evolution” (1986, p. 458).

  Conway Morris had reached the same conclusion for the entire Burgess ecology that Briggs and Whittington had established for arthropod life styles. The “ecological theater” of the Burgess Shale had been rather ordinary: “It may transpire,” Conway Morris wrote, “that the community structure of the Phyllopod Bed was not fundamentally different from that of many younger Paleozoic soft-bodied faunas” (1986, p. 451). Why then was the “evolutionary play” of these early times so different?

  THE BURGESS AS AN EARLY WORLD-WIDE FAUNA

  Nothing breeds scientific activity quite so effectively as success. The excitement generated by recent work on the Burgess Shale has inspired an outburst of interest in soft-bodied faunas and the history of early multicellular life. The Burgess Shale is a small quarry in British Columbia, deposited in Middle Cambrian times, after the celebrated explosion of the Lower Cambrian. As long as its fauna remained geographically confined, and temporally limited to a mere moment after the main event, the Burgess Shale could not tell a story for all of life. The most exciting development of the past decade, continuing and accelerating as I write this book, lies in the discovery of Burgess genera all over the world, and in earlier rocks.

  The first and most obvious extension occurred close to home. If a mudslide down an unstable slope formed the Burgess, many other slides must have occurred in adjacent regions at about the same time; some must have been preserved. As previously discussed, Des Collins of the Royal Ontario Museum has pioneered the effort to find these Burgess equivalents, and he has been brilliantly successful; during the 1981 and 1982 field seasons, Collins found more than a dozen Burgess equivalents in areas within twenty miles or less of the original site. Briggs and Conway Morris joined the field party in 1981, and Briggs returned in 1982. (See Collins, 1985; Collins, Briggs, and Conway Morris, 1983; and Briggs and Collins, 1988.)

  These additional localities are not mere carbon copies of the Burgess. They contain the same basic organisms, but often in very different proportions. One new site, for example, entirely lacks Marrella—the most common species by far in Walcott’s original quarry. The champion here is Alalcomenaeus, one of the rarest creatures, with only two known examples, in the phyllopod bed. Collins also found a few new species. Sanctacaris, as already noted, is especially important as the world’s first known chelicerate arthropod. Another specimen, a weird wonder, has yet to be described; it is “a spiny animal with hairy legs, of unknown affinities” (Collins, 1985).

  Above all, Collins has supplied the most precious themes of diversity and comparison to supplement Walcott’s canonical find. His additional localities include five assemblages sufficiently distinct in mix and numbers of species to be called different assemblages. Significantly, these additional sites include four new stratigraphic levels—all close in time to the phyllopod bed, to be sure, but still teaching the crucial lesson that the Burgess fauna represents a stable entity, not an unrepeatable moment during an early evolutionary riot of change.

  A few basically soft-bodied Burgess species have lightly skeletonized body parts that can fossilize in ordinary circumstances—notably the sclerites of Wiwaxia and the feeding appendages of Anomalocaris. These have long been known from distant localities of other times. But a few bits do not make an assemblage. The Burgess fauna, as a more coherent entity, has now been recognized away from British Columbia, in soft-bodied assemblages in Idaho and Utah (Conway Morris and Robison, 1982, on Peytoia; Briggs and Robison, 1984, on Anomalocaris; and Conway Morris and Robison, 1986). These contain some forty genera of arthropods, sponges, priapulids, annelid
s, medusoids, algae, and unknowns. Most have not yet been formally described, but about 75 percent of the genera also occur in the Burgess Shale. Many species once known only for a moment in time, at a dot in space, now have a broad geographic range and an appreciable, stable duration. Writing about the most common Burgess priapulid, Conway Morris and Robison mark the “notable geographic and stratigraphic extensions of a previously unique occurrence.… Ottoia prolifica has a range through much of the middle Cambrian (?15 million years) during which time it shows minimal morphological changes” (1986, p. 1).

  More exciting still has been the recognition of many Burgess elements in older sediments. The Burgess Shale is Middle Cambrian; the famous explosion that originated modern life occurred just before, during the Lower Cambrian. We would dearly like to know whether Burgess disparity was achieved right away, in the heart of the explosion itself.

  Even before the most recent discoveries, a few positive hints were already in hand, notably some Burgess-like elements in the Lower Cambrian soft-bodied Kinzers fauna of Pennsylvania, and a suspected weird wonder from Australia, described as an annelid worm in 1979. Then, in 1987, Conway Morris, Peel, Higgins, Soper, and Davis published a preliminary description of an entire Burgess-like fauna from the mid-to-late Lower Cambrian of north Greenland. The fauna, like the Burgess itself, is dominated by nontrilobite arthropods. The most abundant creature, about a half inch in length, has a semicircular bivalved carapace; the largest, at about six inches, resembles the Burgess soft-bodied trilobite Tegopelte. Existing collections are poor, and the area is, as we say in the trade, “difficult of access.” But Simon will be visiting next year, and we can expect some new intellectual adventures. In the meantime, he and his colleagues have made the crucial observation, confirming that the Burgess phenomenon occurred during the Cambrian explosion itself: “The extension of stratigraphic ranges of at least some Burgess Shale–like taxa back into the early Cambrian also suggests that they were an integral part of the initial diversification of metazoans” (1987, p. 182).

  Last year, my colleague Phil Signor, knowing of my Burgess interests, sent me a spare reprint from a colleague in China (Zhang and Hou, 1985). I could not read the title, but the Latin name of the subject stood out—Naraoia. Chinese publications are notorious for poor photography, but the accompanying plate shows an unmistakable two-valved, soft-bodied trilobite. A key Burgess element had been found half a world away. Far more important, Zhang and Hou date this fossil to the early part of the Lower Cambrian.

  One creature is tantalizing; but we need whole faunas for sound conclusions. I am delighted to report—for it promises to be the most exciting find since Walcott’s original discovery itself—that Hou and colleagues have since published six more papers on their new fauna. If the djinn of my previous fable (see page 62) had returned five years ago and offered me a Burgess-style fauna at any other place and time, I could not have made a better choice. The Chinese fauna is half a world away from British Columbia—thus establishing the global nature of the Burgess phenomenon. Even more crucially, the new finds seem well dated to a time deep in the Lower Cambrian. Recall the general anatomy of the Cambrian explosion: an initial period, called Tommotian, of skeletonized bits and pieces without trilobites—the “small shelly fauna”; then the main phase of the Cambrian explosion, called Atdabanian, marked by the first appearance of trilobites and other conventional Cambrian creatures. The Chinese fauna comes from the second trilobite zone of the Atdabanian—right in the heart, and near the very beginning, of the main burst of the Cambrian explosion!

  Hou and colleagues describe a rich and well-preserved assemblage, including priapulid and annelid worms, several bivalved arthropods, and three new genera with “merostomoid” body form (Hou, 1987a, 1987b, and 1987c; Sun and Hou, 1987a and 1987b; Hou and Sun, 1988).

  The Burgess phenomenon, then, goes right back to the beginning of the Cambrian explosion. In a preliminary report, based on admittedly uncertain dating, Dzik and Lendzion (1988) describe a creature like Anomalocaris and a soft-bodied trilobite from Eastern European strata below the first appearance of ordinary trilobites. We can no longer doubt that Walcott found products of the Cambrian explosion itself in his slightly later strata of British Columbia. Burgess disparity is astounding enough for a time just 30 to 40 million years after the beginning of the Cambrian. But we cannot even view the Burgess range as accumulating steadily during this relatively short period. The main burst occurred well down in the Lower Cambrian—and probably produced the full Burgess range, if the Chinese fauna proves to be as rich as preliminary accounts suggest. The Burgess Shale represents the slightly later period of stabilization for the products of the Cambrian explosion. But what caused the subsequent decimation, and the consequent pattern of modern life, marked by great gaps between islands of extensive diversity within restricted anatomical designs?

  THE TWO GREAT PROBLEMS OF THE BURGESS SHALE

  The Burgess revision poses two great problems about the history of life. These are symmetrically disposed about the Burgess fauna itself, one before and one after: First, how, especially in the light of our usual views about evolution as a stately phenomenon, could such disparity arise so quickly? And second, if modern life is a product of Burgess decimation, what aspects of anatomy, what attributes of function, what environmental changes, set the pattern of who would win and who would lose? In short, first the origin, second the differential survival and propagation.

  In many ways, the first is a juicier problem for evolutionary theory. How in heaven’s name could such disparity have arisen in the first place, whatever the later fortunes of its exemplars? But the second problem is the subject of this book, for the decimation of the Burgess fauna raises the fundamental question that I wish to address about the nature of history. My key experiment in replaying the tape of life begins with the Burgess fauna intact and asks whether an independent act of decimation from the same starting point would yield anything like the same groups and the same history that our planet has witnessed since the Burgess maximum in organic disparity. Hence, I shall shamelessly bypass the first problem—but not without presenting a brief summary of possible explanations, if only because one aspect of the potential solution does bear crucially on the second problem of differential fate.

  THE ORIGIN OF THE BURGESS FAUNA

  Three major kinds of evolutionary explanation are available for the explosion that led to Burgess disparity. The first is conventional, and has been assumed—largely faute de mieux—in almost all published discussions. The last two have points in common and represent recent trends in evolutionary thinking. I have little doubt that a full explanation would involve aspects of all three attitudes.

  1. The first filling of the ecological barrel. In conventional Darwinian theory, the organism proposes and the environment disposes. Organisms provide raw material in the form of genetic variation expressed in morphological differences. Within a population at any one time, these differences are small and—more important for the basic theory—undirected.* Evolutionary change (as opposed to mere variation) is produced by forces of natural selection arising from the external environment (both physical conditions and interactions with other organisms). Since organisms supply only raw material, and since this raw material has been judged as nearly always sufficient for all changes occurring at characteristically stately Darwinian rates, environment becomes the motor for regulating the speed and extent of evolutionary alteration. Therefore, according to conventional theory, the maximal rates of the Cambrian explosion must indicate something odd about environments at that time.

  When we then inquire about the environmental oddity that could have engendered the Cambrian explosion, an obvious answer leaps at us. The Cambrian explosion was the first filling of the ecological barrel for multicellular life. This was a time of unparalleled opportunity. Nearly anything could find a place. Life was radiating into empty space and could proliferate at logarithmic rates, like a bacterial cell alone on an agar plate. In the bu
stle and ferment of this unique period, experimentation reigned in a world virtually free of competition for the one and only time.

  In Darwinian theory, competition is the great regulator. Darwin conceived the world in metaphor as a log with ten thousand wedges, representing species, tightly hammered in along its length. A new species can enter this crowded world only by insinuating itself into a crack and popping another wedge out. Thus, diversity is self-regulating. As the Cambrian explosion proceeded, it drove itself to completion by filling the log with wedges. All later change would occur by a slower process of competition and displacement.

  This Darwinian perspective also addresses the obvious objection to the model of the empty barrel as the cause of the Cambrian explosion: Life has suffered some astounding mass extinctions since the Cambrian—the Permian debacle may have wiped out 95 percent or more of all marine species—yet the Burgess phenomenon of explosive disparity never occurred again. Life did rediversify quickly after the Permian extinction, but no new phyla arose; the recolonizers of a depleted earth all remained within the strictures of previous anatomical designs. Yet the early Cambrian and post-Permian worlds were crucially different. Five percent may not be a high rate of survivorship, but no mode of life, no basic ecology, was entirely wiped out by the Permian debacle. The log remained populated, even if the wedges had become broader or more widely spaced. To shift metaphors, all the big spheres remained in the barrel, and only the pebbles in the interstices needed a complete recharging. The Cambrian barrel, on the other hand, was flat empty; the log was unscathed, with nary a woodsman’s blow nor a lover’s knife scratch (see Erwin, Valentine, and Sepkoski, 1987, for an interesting, quantitative development of this general argument).