Wonderful Life: The Burgess Shale and the Nature of History

by Stephen Jay Gould

  The canonical attitude of scientists then and now—and the argument that finally secured our legal victory before the Supreme Court in 1987—holds that science and religion operate in equally legitimate but separate areas. This “separationist” claim allots the mechanisms and phenomena of nature to scientists and the basis for ethical decisions to theologians and humanists in general—the age of rocks versus the rock of ages, or “how heaven goes” versus “how to go to heaven” in the old one-liners. In exchange for freedom to follow nature down all her pathways, scientists relinquish the temptation to base moral inferences and pronouncements upon the physical state of the world—an excellent and proper arrangement, since the facts of nature embody no moral claims in any case.

  To Walcott, this separationist view was anathema. He longed to find moral answers directly in nature—his kind of answers, to support his conservative view of life and society. He wished to bring science and religion together, not carve out separate domains in mutual respect. In fact, he charged that the separationist argument had fanned Bryan’s anti-intellectual flame by driving people to the suspicion that scientists really wanted to dispense with religion entirely (but settled, as a temporary and practical matter, for the banning of religion from the affairs of nature). Walcott therefore decided to combat Bryan and his ilk by publishing a statement, signed by a group of respected traditionalists like himself, on the connections between science and religion—particularly on the manifestation of God’s handiwork in the pathways of evolutionary change. Canvassing for signatures, he circulated a letter among his friends:

  Unfortunately through the action of radicals in science and in religion, men of the type of mind of William Jennings Bryan have seen a great danger coming to religion through the teaching of the facts of evolution.

  A number of conservative scientific men and clergymen have been asked to sign a statement to be given much publicity, on the relations of science and religion.

  The statement, published in 1923, two years before the Scopes trial, bore Walcott’s name as first signer, and included Herbert Hoover and such scientific leaders as Henry Fairfield Osborn, Edwin Grant Conklin, R. A. Millikan, and Michael Pupin. “In recent controversies,” the statement held, “there has been a tendency to present science and religion as irreconcilable and antagonistic domains of thought.… They supplement rather than displace or oppose each other.”

  Walcott’s statement went on to argue that the fundamentalist assault could only be quelled by showing the unity of science with religious truths that most Americans viewed as basic to their personal equanimity and social fabric. The primary evidence for this unity lay in the ordered, predictable, and progressive character of life’s history—for the pathways of evolution displayed God’s continuous benevolence and care for his creation. Evolution, with its principle of natural selection leading to progress, represented God’s way of showing himself through nature:

  It is a sublime conception of God which is furnished by science, and one wholly consonant with the highest ideals of religion, when it represents Him as revealing Himself through countless ages in the development of the earth as an abode for man and in the age-long inbreathing of life into its constituent matter, culminating in man with his spiritual nature and all his God-like power.

  In this key passage, the shoehorn becomes an instrument of God. If the history of life shows God’s direct benevolence in its ordered march to human consciousness, then decimation by lottery, with a hundred thousand possible outcomes (and so very few leading to any species with self-conscious intelligence), cannot be an option for the fossil record. The creatures of the Burgess Shale must be primitive ancestors to an improved set of descendants. The Burgess shoehorn was more than a buttress to a comfortable and convenient view of life; it was also a moral weapon, and virtually a decree of God.

  The Burgess shoehorn and Walcott’s struggle with the Cambrian explosion

  If Walcott had never encountered a Cambrian rock before discovering the Burgess Shale, his persona and general attitude toward evolution would by themselves have generated the shoehorn. But Walcott also had highly specific reasons for his view, based upon his lifelong commitment to Cambrian studies, particularly his obsession with the problem of the Cambrian explosion.

  I devoted the first chapter of this book to documenting the influence of iconography upon concepts. I showed how two basic pictures—the ladder of progress and the cone of increasing diversity—buttressed a general view of life based on human hopes, and forced a specific interpretation of Burgess animals as primitive precursors. In the present chapter, my two previous sections, on Walcott’s persona and attitude toward evolution, invoke the ladder; his more specific argument about the Cambrian rests upon the cone.

  Evolutionary trees as the standard iconography for phylogeny had been introduced in the 1860s by the German morphologist Ernst Haeckel. (Others, including Darwin in his single drawing for the Origin of Species, had used botanical metaphors and drawn abstract, branching diagrams as general guides to relationships among organisms. But Haeckel developed this iconography as the preferred representation of evolution. He drew numerous trees with real bark and gnarled branches. And he placed an actual organism on each twig of his copious arborescences.) To native speakers of English, Haeckel’s name may not be so well known as Thomas Henry Huxley’s, but he was surely the most dogged and influential publicist that ever spoke for evolution. Those trees, the mainstay of instruction when Walcott studied and taught paleontology, embody the themes of ladder and cone in both flamboyantly overt and deceptively subtle ways.

  To begin, all of Haeckel’s trees branch continually upward and outward, forming a cone (Haeckel sometimes allowed the two peripheral branches in each subcone to grow inward at the top, in order to provide enough room on the page for all groups—but note how he carefully preserved the general impression of up and out whenever he used this device). Haeckel’s placement of groups reinforces the great conflation of low with primitive, thus uniting the central themes of cone and ladder.

  Consider, for example, Haeckel’s treatment of vertebrate phylogeny (figure 4.4; all figures from Haeckel appear in his Generelle Morphologie of 1866). The entire tree branches upward and outward, forming two levels, with greater diversity at the top. The lower tier, for fishes and amphibia, clearly denotes limited spread and primitivity; the upper, for reptiles, birds, and mammals, implies both more and better. Yet fishes and amphibians live still, whatever their time of origin—and fishes are by far the most diverse of vertebrates both in range of morphology and number of species. Haeckel’s tree of mammals (figure 4.5) dramatically illustrates the conflation of high with advanced, and the misrepresentation of relative diversity that may arise when a small twig is equated with an entire upper level of progress. On this tree, the highly diverse and morphologically specialized artiodactyls (cattle, sheep, deer, giraffes, and their relatives) are squeezed together in the lower tier. By contrast, the primates, forming a comparatively small group, occupy nearly half the upper level on the culturally favored right-hand side. The most diverse of all mammals, the rodents, must squash into a little bubble of space, caught in limbo between the two main layers—for there is no room for them to spread out at the top, where Haeckel’s two favored groups—carnivores (for general valor) and primates (for smarts)—hog all the space.

  4.4. Haeckel’s evolutionary tree of the vertebrates (1866). Fishes (Pisces) actually encompass more disparity than all the rest of the vertebrates combined, but this false iconography, based on the cone of increasing diversity, confines them to a lower branch that gains in breadth as it expands upward.

  4.5. The evolutionary tree of mammals according to Haeckel (1866).

  Echinoderms provide the test case for the iconography of the tree, for in well-preserved hard parts already well-documented in Haeckel’s time, they tell the same tale as the Burgess Shale—maximal early disparity followed by decimation. Note how Haeckel acknowledges this maximal early disparity with a fo
rest of primary stems at the geological beginning (figure 4.6). But the cone decrees that trees must spread outward as they grow, so all these early groups are shrunk into the insignificant space available at the outset. The radically decimated modern tree concentrates nearly all its diversity in two groups of strictly limited range in design—the starfish (Haeckel’s “Asterida”) and the sea urchins (his “Echinida”). Yet Haeckel’s iconography conveys the impression of a continuous increase in range.

  Finally, consider Haeckel’s tree of annelids and arthropods (figure 4.7), the framework upon which Walcott would hang all the Burgess organisms that have fueled our new interpretation. Upon this ultimate expression of up and out, Walcott put all the Burgess arthropods on two adjacent branches of the lower tier—Sidneyia and its relatives in Haeckel’s “Poecilopoda” with horseshoe crabs and eurypterids, and nearly all other forms on the branchiopod-trilobite branch.

  4.6. The evolutionary tree of echinoderms as depicted by Haeckel (1866), in accordance with the cone of increasing diversity. This group actually displays the Burgess pattern of maximal early disparity followed by decimation, but Haeckel’s iconography conveys the impression of continuously increasing diversity and range.

  4.7. The evolutionary tree of arthropods and their relatives as depicted by Haeckel (1866), once again in accordance with the cone of increasing diversity.

  Walcott followed all these iconographical conventions in the three sketchy trees that represent his only published attempts to draw a phylogeny for Burgess organisms. All appear in his major paper on Burgess arthropods (Walcott, 1912). Considered in their original order, they beautifully illustrate the restriction of ideology by iconography. His first chart (figure 4.8) claims to be a simple description of “stratigraphic distribution” in a phylogenetic context. Yet even here, both conventions of cone and ladder conspire to confine Burgess disparity within the limits of a few recognized major groups. The ladder acts to compress one group of five “merostomoid” genera into a single line: by treating Habelia–Molaria–Emeraldella–Amiella–Sidneyia as a structural sequence of ancestors for eurypterids and horseshoe crabs, Walcott conveyed an impression of temporal succession for these contemporaneous (and, we now know, quite unrelated) genera.

  4.8. Walcott’s first chart showing the phylogeny of Burgess arthropods (1912). Walcott forcibly shaped his data in accordance with the cone and ladder by drawing speculative lines of convergence toward a common ancestry in his hypothetical Lipalian interval. He also minimized the explosion of disparity in the Burgess itself by lining up, in an apparently temporal sequence, five forms that were actually contemporaneous (right) and by drawing a hypothetical line at the left boundary to suggest continuing diversity after the Burgess where no evidence exists.

  The cone then forces all other genera into two major groups—the branchiopod and the trilobite-to-merostome lineages. All these genera were contemporaneous, but Walcott framed the entire picture with two vertical lines, implying that later ranges continued to match recorded Burgess disparity—although no direct evidence supports this assumption. Note, especially, that the left-hand boundary line corresponds to no organism at all—the line is an iconographical device added to guide the eye into seeing a cone. Without this line, disparity would be maximal in the Burgess, and markedly decreased thereafter. Never doubt the power of such tiny and apparently insignificant moves. In a way, everything that I am trying to say in this book achieves an elegant epitome in this one vertical stroke—added to represent a philosophy of life, not the empirical record of organisms.

  As a second device, buttressed by no data and added to support a traditional interpretation, Walcott drew the origin of Burgess genera at different levels within a Precambrian interval that he called Lipalian. He connected these levels with two slanted lines that point downward toward a distant Precambrian ancestor for the entire tree. This device provides the tree with a root, in an early period of restricted disparity. But Walcott had no evidence at all—and we have none today—for such evolutionary order among the Burgess arthropods.

  Walcott’s second chart (figure 4.9) illustrates the tyranny of the cone in an even more striking manner. Walcott claimed that five distinct lineages could be recognized among Burgess arthropods—the extinct trilobites, and four prominent groups of organisms inhabiting modern waters. Again, he used two devices to compress Burgess disparity into the narrow end of a cone. First, he showed all five lineages as converging toward the bottom (subtly for four, perhaps because he felt sheepish about making such an assertion with no supporting data at all; more boldly, with a distinct angular bend, for the merostome lineage, where he adduced some evidence—see below). Second, he placed all these contemporaneous fossils at different positions on his vertical branches, implying that they represented evolutionary diversification through time. On the merostome branch, he lined up eight genera (five of which are known only as contemporaries in the Burgess Shale) to forge a hypothetical link between merostomes and crustaceans: “Such forms as Habelia, Molaria and Emeraldella serve to fill in the gap between the Branchiopoda and the Merostomata as represented by Sidneyia and later the eurypterids” (1912, p. 163). Finally, figure 4.10 shows Walcott’s last and most abstract phylogeny for the Burgess arthropods. Even larger groups are lined up on vertical branches, and the entire tree converges to a branchiopod root.

  These phylogenies embody the crucial link between Walcott’s interpretation of Burgess arthropods and the previous focus of a career that had spanned more than thirty intense years—the study of Cambrian rocks and the problem of the Cambrian explosion. The linkage between the Burgess and Walcott’s view of the Cambrian explosion provides a final, and more specific, explanation for his inevitable embrace of the shoehorn as an interpretation for Burgess fossils.

  4.9. Walcott’s second chart showing the phylogeny of Burgess arthropods (1912). Again, the lineages converge toward a hypothetical common ancestry, and several contemporaneous forms are placed in ladder-like order, on the left-hand and middle lines.

  In short, Walcott viewed the Burgess arthropods as members of five major lineages, already stable and well established at this early Cambrian date. But if life had already become so well differentiated along essentially modern lines, the five lineages must have existed at the inception of the Cambrian explosion as recorded by fossil evidence—for evolution is stately and gradual, not a domain of sudden jumps and mad eruptions of diversity. And if the five lineages existed as well-differentiated groups right at the beginning of the Cambrian, then their common ancestor must be sought far back in the Precambrian. The Cambrian explosion must therefore be an artifact of an imperfect fossil record; the late Precambrian seas, in Darwin’s words, must have “swarmed with living creatures” (1859, p. 307).

  Walcott thought that he had discovered why we have no evidence for this necessary Precambrian richness. In other words, he thought that he had solved the riddle of the Cambrian explosion in orthodox Darwinian terms. The ordering of Burgess arthropods into five well-known and stable groups cemented his solution:

  4.10. Walcott’s third and last attempt at depicting arthropod evolution (1912). The lineages now converge to a common point, and major groups are lined up, one above the other, on one of the three diverging branches.

  The Cambrian crustacean fauna suggests that five main lines or stems … were in existence at the beginning of Cambrian time and that all of them had already had their inception in Lipalian time or the period of the Precambrian marine sedimentation of which no known part is present in on the existing continents (1912, pp. 160–61).

  We must remember that the Cambrian explosion was no ordinary riddle, and its potential solution therefore no minor plum, but something more akin to the Holy Grail. Darwin, as already noted, had publicly fretted that “the case at present must remain inexplicable; and may be truly urged as a valid argument against the views here entertained” (1859, p. 308).

  Two different kinds of explanations for the absence of Precambrian ances
tors have been debated for more than a century: the artifact theory (they did exist, but the fossil record hasn’t preserved them), and the fast-transition theory (they really didn’t exist, at least as complex invertebrates easily linked to their descendants, and the evolution of modern anatomical plans occurred with a rapidity that threatens our usual ideas about the stately pace of evolutionary change).

  Darwin, making his characteristic (and invalid) conflation of leisurely, gradual evolution and change by natural selection, rejected the fast-transition theory out of hand. He insisted that any complex Cambrian creature must have arisen from a lengthy series of Precambrian ancestors with the same basic anatomy: “I cannot doubt that all the Silurian [Cambrian, in modern terminology] trilobites have descended from some one crustacean, which must have lived long before the Silurian [Cambrian] age” (1859, p. 306).

  Accordingly, Darwin searched for a believable version of the artifact theory, finally proposing that, in Precambrian times, “clear and open oceans may have existed where our continents now stand.” Such tracts of uninterrupted water would have received little or no sediment. Hence our current continents, containing all rocks available to our view, rose from an area that accumulated no strata during the crucial span of late Precambrian faunas, while regions of shallow water that did receive Precambrian sediments now lie in inaccessible oceanic depths.