Ever Since Darwin: Reflections in Natural History

by Stephen Jay Gould

  The last century of argument has produced only two basic strategies for a scientific explanation of the Cambrian explosion.

  First, we may argue that it is a false appearance. Evolution was really slow and gradual, as Western biases dictate. The so-called explosion only marks the first appearance in the fossil record of creatures that had been living and developing for a long part of the Precambrian. But what prevented the fossilization of such rich faunas? Here we have a variety of proposals ranging from the absurdly ad hoc to the eminently plausible. To cite just a few:

  (1)The Cambrian represents the first preservation of unaltered rocks; Precambrian sediments have been subjected to such heat and pressure that their fossil remains have been obliterated. This is empirically false, beyond any doubt.

  (2)Life evolved in terrestrial lakes. The Cambrian represents the migration of this fauna to the sea.

  (3)All early metazoans were soft-bodied. The Cambrian represents the evolution of fossilizable hard parts.

  The popularity of this first strategy has plummeted with the discovery of abundant Precambrian fossil deposits devoid of anything more complex than algae. Nonetheless, the argument based on hard parts probably contains an element of truth, though it cannot provide the entire answer. A clam without a shell is not a viable animal; you cannot clothe any simple soft-bodied organism to make one. The delicate gills and the complex musculature clearly evolved in association with a hard outer covering. Hard parts often require a simultaneous and complex modification of any conceivable soft-bodied ancestor; their sudden appearance in the Cambrian, therefore, implies a truly rapid evolution of the animal they cover.

  As a second strategy, we may claim that the Cambrian explosion is a real event representing the extremely rapid evolution of complexity. Something must have happened to the environment of simple, soft-bodied precursors of Cambrian metazoans in order to engender such a rapid burst of evolution. We have only two overlapping possibilities: changes in the physical or in the biological environment.

  In 1965, Lloyd V. Berkner and Lauriston C. Marshall, two physical scientists from Dallas, published a famous article proposing that levels of oxygen in the earth’s atmosphere exerted a direct physical control on the Cambrian explosion of life. Geologists agree that the earth’s original atmosphere contained little or no free oxygen. Oxygen built up gradually as a result of organic activity—the photosynthesis of Precambrian algae. Metazoans require high levels of free oxygen for two reasons: directly, for respiration; indirectly because oxygen, in the form of ozone, absorbs harmful ultraviolet radiation in the upper atmosphere before it reaches life on the earth’s surface. Berkner and Marshall simply proposed that the base of the Cambrian marks the first time that atmospheric oxygen reached a level sufficient for respiration and the shielding of harmful radiation.

  But this attractive notion has foundered on the geologic evidence. Photosynthesizing organisms were probably abundant more than two and a half billion years ago. Is it reasonable to suppose that some two billion years were required for the buildup of sufficient oxygen for respiration? Moreover, many extensive deposits between one and two billion years old contain large volumes of strongly oxidized rocks.

  Berkner and Marshall’s hypothesis embodies an attitude all too common among nonbiologists who lack sufficient appreciation for the complexity that makes a machine a poor model for a living organism. Physical models often employ simple, inert objects like billiard balls that respond automatically to the impress of physical forces. But an organism cannot be pushed around so easily; it certainly does not evolve automatically. Berkner and Marshall’s hypothesis relies upon the billiard-ball thinking that I term “physicalism”—metazoans arise immediately and automatically when a physical barrier to their existence is removed. The presence of sufficient oxygen, however, does not guarantee the immediate evolution of everything that could breathe it. Oxygen is a necessary but woefully insufficient requirement for the evolution of metazoans. In fact, enough oxygen probably existed for a billion years before the Cambrian explosion. Perhaps we should look to biological controls.

  Steven M. Stanley of Johns Hopkins University has recently argued that a popular ecological theory—the “cropping principle”—may provide such a biological control (Proceedings of the National Academy of Sciences, 1973). The great geologist Charles Lyell argued that a scientific hypothesis is elegant and exciting insofar as it contradicts common sense. The cropping principle is just such a counterintuitive notion. In considering the causes of organic diversity, we might expect that the introduction of a “cropper” (either a herbivore or a carnivore) would reduce the number of species present in a given area: after all, if an animal is cropping food from a previously virgin area, it ought to reduce diversity and remove completely some of the rarer species.

  In fact, a study of how organisms are distributed yields the opposite expectation. In communities of primary producers (organisms that manufacture their own nutrients by photosynthesis and do not feed upon other creatures), one or a very few species will be superior in competition and will monopolize space. Such communities may have an enormous biomass, but they are usually impoverished in numbers of species. Now, a cropper in such a system tends to prey on the abundant species, thus limiting their ability to dominate and freeing space for other species. A well-evolved cropper decimates—but does not destroy—its favorite prey species (lest it eat itself to eventual starvation). A well-cropped ecosystem is maximally diverse, with many species and few individuals of any single species. Stated another way, the introduction of a new level in the ecological pyramid tends to broaden the level below it.

  The cropping principle is supported by many field studies: predatory fish introduced in an artificial pond cause an increase in the diversity of zooplankton; removal of grazing sea urchins from a diverse algal community leads to the domination of that community by a single species.

  Consider the Precambrian algal community that persisted for two and a half billion years. It consisted exclusively of simple, primary producers. It was uncropped and, for that reason, biologically monotonous. It evolved with exceeding slowness and never attained great diversity because its physical space was so strongly monopolized by a few abundant forms. The key to the Cambrian explosion, Stanley argues, is the evolution of cropping herbivores—single-celled protists that ate other cells. Croppers made space for a greater diversity of producers, and this increased diversity permitted the evolution of more specialized croppers. The ecological pyramid burst out in both directions, adding many species at lower levels of production and adding new levels of carnivory at the top.

  How can one prove such a notion? The original cropping protist, perhaps the unsung hero of the history of life, probably was not fossilized. There is, however, some suggestive indirect evidence. The most abundant producer communities of the Precambrian are preserved as stromatolites (blue-green algal mats that trap and bind sediment). Today, stromatolites thrive only in hostile environments largely devoid of metazoan croppers (hypersaline lagoons, for example). Peter Garrett found that these mats persist in more normal marine environments only when croppers are artificially removed. Their Precambrian abundance probably reflects the absence of croppers.

  Stanley did not develop his theory from empirical studies of Precambrian communities. It is a deductive argument based on an established principle of ecology that does not contradict any fact of the Precambrian world and seems particulary consistent with a few observations. In a frank concluding paragraph, Stanley presents four reasons for accepting his theory: (1) “It seems to account for what facts we have about Precambrian life”; (2) “It is simple, rather than complex or contrived”; (3) “It is purely biological, avoiding ad hoc invocation of external controls”; and (4) “It is largely the product of direct deduction from an established ecological principle.”

  Such justifications do not correspond to the simplistic notions about scientific progress that are taught in most high schools and advanced by most media. St
anley does not invoke proof by new information obtained from rigorous experiment. His second criterion is a methodological presumption, the third a philosophical preference, the fourth an application of prior theory. Only Stanley’s first reason makes any reference to Precambrian facts, and it merely makes the weak point that his theory “accounts” for what is known (many other theories do the same).

  But creative thought in science is exactly this—not a mechanical collection of facts and induction of theories, but a complex process involving intuition, bias, and insight from other fields. Science, at its best, interposes human judgment and ingenuity upon all its proceedings. It is, after all (although we sometimes forget it), practiced by human beings.

  15 | Is the Cambrian Explosion a Sigmoid Fraud?

  RODERICK MURCHISON, urged on by his wife, gave up the joys of fox hunting for the more sublime pleasures of scientific research. This aristocratic geologist devoted much of his second career to documenting the early history of life. He discovered that the first stocking of the oceans did not occur gradually with the successive addition of ever more complex forms of life. Instead, most major groups seemed to arise simultaneously at what geologists now call the base of the Cambrian period some 600 million years ago. To Murchison, a devout creationist writing in the 1830s, this episode could only represent God’s initial decision to populate the earth.

  Charles Darwin viewed this observation with trepidation. He assumed, as evolution demanded, that the seas had “swarmed with living creatures” before the Cambrian period. To explain the absence of fossils in the earlier geologic record, he apologetically speculated that our modern continents accumulated no sediments during Precambrian times because they were covered by clear seas.

  Our modern view synthesizes these two opinions. Darwin, of course, has been vindicated in his cardinal contention: Cambrian life did arise from organic antecedents, not from the hand of God. But Murchison’s basic observation reflects a biological reality, not the imperfections of geologic evidence: the Precambrian fossil record is little more (save at its very end) than 2.5 billion years of bacteria and blue-green algae. Complex life did arise with startling speed near the base of the Cambrian. (Readers must remember that geologists have a peculiar view of rapidity. By vernacular standards, it is a slow fuse indeed that burns for 10 million years. Still, 10 million years is but 1/450 of the earth’s history, a mere instant to a geologist.)

  Paleontologists have spent a largely fruitless century trying to explain this Cambrian “explosion”—the steep rise in diversity during the first 10 to 20 million years of the Cambrian period. (see essay 14). They have assumed, universally, that the puzzling event is the explosion itself. Any adequate theory, therefore, would have to explain why the early Cambrian was such an unusual time: perhaps it represents the first accumulation of sufficient atmospheric oxygen for respiration, or the cooling down of an earth previously too hot to support complex life (simple algae survive at much higher temperatures than complex animals), or a change in oceanic chemistry permitting the deposition of calcium carbonate to clothe previously soft-bodied animals with preservable skeletons.

  I now sense that a fundamental change in attitude is about to take hold within my profession. Perhaps we have been looking at this important problem the wrong way round. Perhaps the explosion itself was merely the predictable outcome of a process inexorably set in motion by an earlier Precambrian event. In such a case, we would not have to believe that early Cambrian times were “special” in any way; the cause of the explosion would be sought in an earlier event that initiated the evolution of complex life. I have recently been persuaded that this new perspective is probably correct. The pattern of the Cambrian explosion seems to follow a general law of growth. This law predicts a phase of steep acceleration; the explosion is no more fundamental (or in need of special explanation) than its antecedent period of slower growth or its subsequent leveling off. Whatever initiated the antecedent period virtually guaranteed the later explosion as well. In support of this new perspective, I offer two arguments based on a quantification of the fossil record. I hope not only to make my particular case but also to illustrate the role that quantification can play in testing hypotheses within professions that once eschewed such rigor.

  The day-to-day work of field geology is a painstaking exercise in apparent minutiae of detail: the mapping of strata; their temporal correlation by fossils and by physical “super-position” (younger above older); the recording of rock types, grain sizes, and environments of deposition. This activity is often pooh-poohed by hotshot young theorists who regard it as the dog work of unimaginative drones. Yet we would have no science without the foundation that these data provide. In this case, our revised perspective on the Cambrian explosion rests upon a refinement of early Cambrian stratigraphy established primarily by Soviet geologists in recent years. The long Lower Cambrian has been subdivided into four stages and the first appearances of Cambrian fossils have been recorded with much greater accuracy. We can now tabulate a finely divided sequence of first appearances where previous stratigraphers could only record “Lower Cambrian” for all groups (thus accentuating the apparent explosion).

  J.J. Sepkoski, a paleontologist at the University of Rochester, has recently found that a plot of increasing organic diversity versus time from the late Precambrian to the end of the “explosion” conforms to our most general model of growth—the so-called sigmoidal (S-shaped) curve. Consider the growth of a typical bacterial colony on a previously uninhabited medium: each cell divides every twenty minutes to form two daughters. Increase in population size is slow at first. (Rates of cell division are as fast as they will ever be, but founding cells are few in number and the population builds slowly toward its explosive period.) This “lag” phase forms the initial, slowly rising segment of the sigmoidal curve. The explosive, or “log,” phase follows as each cell of a substantial population produces two fecund daughters every twenty minutes. Clearly this process cannot continue indefinitely: a not-too-distant extrapolation would fill the entire universe with bacteria. Eventually, the colony guarantees its own stability (or demise) by filling its space, exhausting its nutrients, fouling its nest with waste products, and so on. This leveling puts a ceiling on the log phase and completes the S of the sigmoidal distribution.

  It is a long step from bacteria to the evolution of life, but sigmoidal growth is a general property of certain systems, and the analogy seems to hold in this case. For cell division, read speciation; for the agar substrate of a laboratory dish, read the oceans. The lag phase of life is the slow, initial rise of latest Precambrian times. (We now have a modest fauna of latest Precambrian age—mainly coelenterates [soft corals and jellyfish] and worms.) The famous Cambrian explosion is nothing more than the log phase of this continuous process, while post-Cambrian leveling represents the initial filling of ecological roles in the world’s oceans (terrestrial life evolved later).

  A typical sigmoidal (S-shaped) curve. Note slow beginning (lag phase), middle period of rapid increase (log phase) and final tapering off.

  If the laws of sigmoidal growth regulated the early diversification of life, then there is nothing special about the Cambrian explosion. It is merely the log phase of a process determined by two factors: (1) the event that initiated the lag phase well within Precambrian times and (2) the properties of an environment that permitted sigmoidal growth.

  As Johns Hopkins paleontologist S. M. Stanley wrote in a recent review (American Journal of Science, 1976): “We can abandon the traditional view that the origins of major fossil taxa near the start of the Cambrian … represent a major enigma. What remains as the ‘Cambrian Problem’ is the delay of the origin of multicellularity until the Earth was nearly 4 billion years old.” We may deny the Cambrian problem by casting it back upon an earlier event, but the nature and cause of this earlier episode remains as the enigma of paleontological enigmas. The late Precambrian origin of the eukaryotic cell must provide an important determinant. (I argue in es
say 13 that efficient sexual reproduction required a eukaryotic cell with discrete chromosomes, and that complex organisms could not evolve without the genetic variability that sexual reproduction supplies.) But we haven’t the slightest idea why the eukaryotic cell arose when it did more than 2 billion years after the evolution of prokaryotic ancestors. In essay 14, I advocated Stanley’s “cropping” theory for the initiation of sigmoidal increase following the evolution of eukaryotic cells. Stanley argues that Precambrian prokaryotic algae had usurped all available space in their potential habitat, thus precluding the evolution of anything more complex by denying a foothold to any competitor. The first eukaryotic herbivore, in the course of its copious, if unvaried, worldwide feast, freed enough space for the evolution of competitors.

  Speculation may be intriguing, but we have little concrete to say about my first factor—the cause that initiated sigmoidal increase. We can, however, do better for the second—the nature of an environment that permitted it. Sigmoidal growth is not a universal property of natural systems; it occurs only in one kind of environment. Our laboratory bacteria would not have increased in an S-shaped curve if their medium had already been densely populated or devoid of nutrients. Sigmoidal patterns occur only in open, unconstrained systems, where food and space are so abundant that organisms grow until their own numbers limit further increase. The Precambrian oceans clearly formed such an “empty” ecosystem—plenty of space, abundant food, no competition. (The early eukaryotes could thank their prokaryotic ancestors not only for an immediate supply of food but also for their prior service in supplying oxygen to the atmosphere through photosynthesis.) The sigmoidal curve—with the Cambrian explosion as its log phase—represents the first stocking of the world’s oceans, a predictable pattern of evolution in open ecosystems.