This View of Life

Play It Again, Life: Natural History, February 1986

If the great experiment were rerun, the ending might be a world of wiwaxiids or a sea filled with little penis worms

By Stephen Jay Gould

Prediction must be the most difficult of human arts, a source of constant embarrassment for the best-laid guesses of mice and men. My special badge of merit goes to Abraham Lincoln for his unintended taunt to a hundred million schoolchildren forced to memorize the Gettysburg Address: "The world will little note nor long remember what we say here…."

Understanding after the fact confers no special perspicacity. The real test for any diviner can only lie in grasping the outcome at the outset. Correct predictions, in themselves, offer no proof of true wisdom, for how can we distinguish dumb luck from horse sense? The only good experiment is, alas, the most undoable of all intriguing thoughts in a world of irrevocable history-to run back the tape and play it again, Sam.

This famous old thought experiment applies with special fascination to life's history-and for a particular, parochial reason. Humans arrived just yesterday on the stage not noticeably prepared for our advent. Would we appear at all if we could rewind the tape to an appropriate beginning (say, the origin of modern phyla in the Cambrian explosion more than 500 million years ago) and let it run again? If anything like human intelligence graced the replay, we might feel more secure in a certain sense of inexorability. If the rerun began with trilobites and ended with trilobites, we might treasure our fragility and thank our lucky stars.

Our usual suspicion that we rule by right (and necessity) arises from the common conviction that, however noble the sentiment, the Preacher was no prophet when he proclaimed: "the race is not to the swift, nor the battle to the strong" (Eccles. 9:11). Does not evolution (at least for vertebrates) pass through a sequence of advance from fish to amphibian to reptile to mammal to human? If evolution has moved along such a linear path of progress, how could the story have unfolded in any basically different way? What ever is, is right. We bolster this dubious interpretation with the cardinal misperception of retrospective history. From a vantage point at the top of the pile, the slopes seem to obviously inclined toward the summit. But again I ask: Can anyone read the final form of a drip castle in the first handful of wet sand?

I believe that the greatest impediment to understanding this thought experiment lies in our lamentable tendency to interpret complex issues as dichotomies with only two possible and contradictory outcomes. In thinking about the rerun of life's tape, we usually entertain two alternatives-either the replay will generate something close to the original story (leading us to argue that life's history makes sense and had to unfold as it did), or the second showing will differ markedly from the first enactment (leading us to conclude that life's history doesn't make much sense and could have ventured down a million other pathways). The history of life, in this dichotomy, is either sensible or random, orderly or chaotic.

I wish to suggest a third, or middle, position. I even believe that I can present some good, if inferential, evidence for it. Suppose that the rerun varies greatly from the original, but that the new version makes perfectly good sense too. Perhaps our world is not random. The race does go to the swift, but each time you play the tape, different lineages win the relay for wondrously diverse reasons. We cannot know who will be swift and why, but winners prevail for good cause. Actual events are accidents in the important sense that any outcome records but one possible result among millions, but each pathway yields to reasonable interpretation. This third view represents historical contingency in its deep meaning-the enormous complexity and multifarious opportunities that make history quite unpredictable, but sensible after the fact. If we understood contingency in this light, we might steer a proper course between the reefs of strict repetition (to preserve sense) and randomness (to acknowledge the likelihood of different outcomes). Unpredictable does not mean chaotic.

Since our fossil record is largely the story of marine invertebrates (dinosaurs are famous, but rare by contrast), let us consider our current faunas and ask whether a Cambrian observer could have predicted the outcome. Marine animals are amazingly diverse, but clumped at several nodes of overwhelming success. You will find a few aplacophorans, a kinorhynch here and there, and even, if you are lucky (or happen to be hovering over a deep-sea "vent" fauna), a pogonophoran-but the main inhabitants of the ocean floor reside in great abundance within a few groups: clams, snails, echinoids, crabs, and polychaete worms. All these creatures do very well, and biologists have demonstrated the functional basis of their success, but could we always have senses their victory?

As we press the rewind button and watch the tape go back, our obvious stopping point must be the middle Cambrian Burgess Shale, our unique "window" on soft-bodied marine life just after the Cambrian explosion (and subject of my last two columns as well). As we saw last month, C.D. Walcott, discoverer of this fauna early in our century , shoehorned all its members into modern groups. Since he also tended to force Burgess organisms into well-known (read abundant) classes and phyla, the Burgess bestiary favored groups that still dominate today-and we might grant a kind of stately inevitability to life's tape. But both of my earlier columns also emphasized, as their common theme, the exciting reinterpretation that has recognized a fauna of surpassing strangeness in the Burgess, an assemblage containing a dozen creatures so different from anything living today that we must recognize them as new phyla in our classifications (the highest level of taxonomic distinction below the kingdom Animalia itself).

How does this new reinterpretation speak to our thought experiment in rewinding life's tape? We might ask two related questions: First, were the unique Burgess phyla doomed by inadequate design to their brief existence as failed experiments in the first flowering of animal life? Second, for living groups with Burgess representatives, would we have known, at this outset, which were destined for domination and which for peripheral status in the nooks and crannies of an unforgiving world? Consider an example of each question-as an illustration for the "might-have-beens" of evolution.

Case 1: The fate of Burgess phyla. The Burgess fauna will reorient the thinking of anyone who suspected that modern animals must have exploited all possible designs because creatures can eat and move in just so many ways. Consider Opabinia, with five eyes at the base of a "head" stalk armed with teeth at its front edge; Anomalocaris (the "odd shrimp") with one pair of appendages in front, a mouth surrounded by a circle of plates below and an expanded wedge-shaped shield, housing the gills, above and behind; or Hallucigenia (the name speaks for itself), with seven pairs of fixed spines supporting a trunk with a globular "head" at one end and a spiked "tail" at the other .

[NOTE: after this paper was published, it was realized that Hallucigenia had been reconstructed incorrectly: upside down. It has the spines on its back, sticking up, and has stubby legs with claws. It has now been recognized as a spined relative of the velvet worm, Peripatus, Phylum Onychophora]

Wiwaxia corrugata, the subject of my first Burgess column, lies among the oddballs. Its simple, worm-like body sported a nearly complete cover of elliptical plates, called sclerites, and two rows of spines along the back. Simon Conway Morris, author of a new monograph on Wiwaxia (the initial inspiration for this trilogy of columns), places this animal among the Burgess enigmas but proposes a potential distant relationship with mollusks. For Conway Morris discovered two bars of teeth, inserted at the front end of the gut and similar in form and position to the distinctive radula of mollusks.

Clams and snails represent one of life's greatest success stories, yet only a few specialists have ever heard of Wiwaxia. But who can say that a replay of life's tape would not see the mollusks dive to an early death and wiwaxiids flourish. We know nothing about design or abundance of mollusks and wiwaxiids that would lead our hypothetical Burgess zoologist to specify outcomes without the benefit of hindsight. Wiwaxia was not a fragile rarity in Burgess times. Its sclerites fossilized, and we can therefore trace its abundance and distribution beyond the Burgess. Conway Morris cites six genera of described Cambrian sclerites that might be wiwaxiids. Mollusks, moreover, had not yet proved their success. We know only two genera of tiny clams from the Cambrian, while the larger array of small, twisted, or cap-shaped shells may be snails or monoplacophorans. But we have no definite proof that all are even mollusks.

Conway Morris, after pondering the operation and distribution of wiwaxiids for many years, could not convince himself that any predictable inadequacy had sealed their fate. He wrote:

In conclusion, if the clock was turned back so metazoan diversity was allowed to rerun across the Precambrian-Cambrian boundary, it seems possible that the successful body plans emerging from this initial burst of evolution may have included wiwaxiids rather than mollusks.

Moreover, the current domination of clams and snails within the mollusks seems equally uncertain from a Cambrian perspective. Clams presumable evolved from rostroconchs, an extinct group with an outer covering that seems to include two symmetrical valves (as in clams) but actually grow from a single starting point and then folds over along a pseudohinge. My colleague John Pojeta has just told me that rostroconchs (see picture below) outnumbered clams for the first 100 million years of their coexistence. Our Burgess zoologist might well have placed his money on the ultimate losers.

Case 2: Relative abundances then and now. Consider the current distribution of two phyla sharing the most common of invertebrate body plans-the flexible, elongate, bilateral symmetry of "worms". Polychaetes, the major marine component of the Phylum Annelida (including earthworms on land), represent one of life's great success stories. The best modern epitome, Sybil P. Parker's McGraw-Hill Synopsis and Classification of Living Organisms, devotes forty pages to a breathless summary of their 87 families, 1,000 genera, and some 8,000 species. Polychaetes range in size from less than one millimeter to more than three meters; they live nearly everywhere, mostly on the sea floor, but some in brackish or fresh water, and a few in moist earth. Their life styles also span the range of the thinkable: most are free living and carnivorous or scavengers, bit others dwell commensally with sponges, mollusks or echinoderms, and some are parasites.

By contrast, consider the priapulids, literally, little penis worms. Priapulids are burrowing worms with bodies divided roughly into three parts; a rear end with one or two appendages; a middle trunk; and a retractable front end or proboscis. Both the form of the proboscis and its power of erection from the trunk, inevitably reminded early male zoologists of something else to which they were, no doubt, firmly and fondly attached-hence the burden of nomenclature for these creatures of Priapulus, or the "little penis".

The armature of the priapulid proboscis might give some cause for alarm in unwarranted analogy. The lower portion sports twenty-five rows of little teeth, or scalids (in most species). An unarmed collar, or buccal ring, surmounts the scalids, while the upper end contains several inscribed pentagons of teeth surrounding the central mouth. Most priapulids are active carnivores, capturing and swallowing their prey whole, although one species may feed on detritus.

But when we turn to Parker's compendium, we find but three pages devoted to priapulids, with a leisurely description of each family. Priapulids just don't contribute much to an account of organic diversity. Zoologists have found only nine species, allocated to six genera and three families. For some reason, priapulids do not rank among the success stories of modern biology.

An examination of priapulid distribution provides a clue to their relative failure. All priapulids live in unusual, harsh or marginal environments-as if they fail to compete in the shallow, open marine environments frequented by most "standard" organisms of the sea and can hang on only where ordinary creatures don't bother. Two priapulid families (with one species each) include miniaturized worms grown so small that they live among sand grains in the rich and fascinating (but decidedly "unstandard") world of the so-called interstitial fauna. Most priapulids (seven species in four genera) belong to the family Priapulidae, larger worms (up to twenty centimeters) of the macroscopic sea bottom. But these priapulids do not inhabit the richest environment of the shallow water tropics. They live in the coldest realms, at great depths in tropical regions, but rising to shallow waters in the cold climates of high latitudes. They can also tolerate a variety of unusual conditions-low oxygen, hydrogen sulfide poisoning, low salinity, sharp fluctuations in saltiness, and unproductive regions that impose long periods of starvation. It does not strain the boundaries of reasonable inference to argue that priapulids have managed to keep a toehold in a tough world by opting for difficult places devoid of sharp competition.

We might assume that such a striking difference in today's seas recorded something so intrinsic about the relative mettle of these two groups that their geological history should be an uninterrupted tale of polychaete prosperity and priapulid struggle. If so, we are in for yet another surprise from the redoubtable Burgess fauna. This first recorded beginning of modern soft-bodied life contains six genera of polychaetes and six genera of priapulids. (See Simon Conway Morris's monographs on both groups-Special papers of the Palaeontological Association of the United Kingdom, no. 20, 1977, for the priapulids; Philosophical Transactions of the Royal Society of London, vol. 285,1979, for the polychaetes).

This difference between past and present extends well beyond this bald numerical statement, impressive enough in itself. The Burgess priapulids are, numerically, a major component of the fauna and, perhaps, the earth's first important soft-bodied carnivores. Ottoia prolifica, most common of the Burgess priapulids, swallowed its prey whole. Hyolithids (conical shelled creatures of uncertain affiliation) were favored as food. Thirty-one specimens have been found in the guts of Ottoia, most swallowed in the same orientation (and, therefore, almost certainly hunted and consumed in a definite style). One Ottoia had six hyolithids in its gut. Another specimen had eaten one of its own-the first recorded example of cannibalism in our fossil record.

Polychaetes, by contrast, are much rare numerically, if equal to priapulids in taxonomic diversity (and range of exploited environments, with active swimmers burrowers, and sediment eaters among the polychaetes). Simon Conway Morris remarks: "In comparison with the situation in many modern marine environments, the Burgess shale polychaetes had a relatively minor role."

Obviously, something dramatic (and disastrous) has happened to priapulids since the Burgess. Once they had no rivals for abundance of soft-bodied forms, exceeding even the proud polychaetes of current majesty. Now they are few and forgotten, denizens of the ocean's spatial and environmental peripheries. The entire modern world contains no more genera of priapulids than the single Burgess fauna from one quarry in British Columbia-while Burgess priapulids occupied center stage, not the tawdry provinces. What happened?

We do not know. It is tempting to argue that polychaetes had some obvious biological leverage from the start, and were destined for domination, however modest their beginning. But we have no idea what such an advantage might be. Conway Morris makes the intriguing observation that Burgess polychaetes had no jaws and that these organs of successful polychaete predators did not evolve until the subsequent Ordovician period. Perhaps the origin of jaws was what gave polychaetes their edge over the previously more abundant priapulids.

This tale is plausible and may be true, but we do not know, and a correlation (jaws with the beginning of dominance) need not imply a cause. In any case, our hypothetical Burgess geologist would not have known that the modest polychaetes would evolve jaws fifty million years hence.

The distribution of modern priapulids does indicate a basic failure relative to Burgess abundances, but who knows the whys or wherefores? And who can say that a rerun of life's tape would not yield a modern world dominated by priapulids, with a few struggling jawless polychaetes at a tenuous periphery? What did happen makes sense: our world is not chaotic. But many other plausible scenarios would have satisfied any modern votary of progress and good sense, and priapulid dominance lies firmly among the might-have-beens.

Are these Burgess fancies a common property of life's history throughout or an oddity of uncertain beginnings superseded by later inexorability? Consider one more might-have-been. When dinosaurs perished in the Cretaceous debacle, they left a vacuum in the world of large-bodied carnivores. Did the current reign of cats and dogs emerge by predictable necessity or contingent fortune? Would an Eocene paleontologist, surveying the vertebrate world fifty million years ago, have singled out for success the ancestors of Leo, king of beasts?

I doubt it. The Eocene world sported many lineages of mammalian carnivores, only one, ancestral to modern forms and not especially distinguished at the time. But the Eocene featured a special moment in the history of carnivores, a pivot between two possibilities: one realized, the other forgotten. Mammals did not hold all the chips. In 1917, the American paleontologists W. D. Matthew and W. Granger describe in the Bulletin of the American Museum of Natural History the skeleton of a giant predacious bird from the Eocene of Wyoming, Diatryma gigantea. They wrote of "this magnificent and quite unexpected bird skeleton":

Diatryma was a gigantic bird, ground living and with vestigial wings. In bulk of body and limbs it equaled all but the largest of the moas and surpassed any living bird….The height of the reconstructed skeleton is nearly 7 feet. The neck and head were totally unlike any living bird, the neck short and very massive, the head of enormous size with a huge compressed beak.

The gigantic head and short, powerful neck identified Diatryma as a fierce carnivore, in sharp contrast with the small heads and long slender necks of more peaceful ratites (ostriches, rheas, and their relatives). Like Tyrannosaurus with its diminutive forelimbs, Diatryma must have kicked, clawed and bitten its prey into submission.

Diatrymids, distant relatives perhaps of cranes but no kin to ostriches and their ilk, ranged over Europe and North America for several million years. The plum of dominant carnivory could have fallen to the birds, but mammals finally prevailed, and we do not know why. We can invent stories about two legs, bird brains, and no teeth as necessarily inferior to all fours and sharp canines, but we know in our heart of hearts that if birds had won, we could tell just as good a tale about their inevitable success. A. S. Romer, leading vertebrate paleontologist of a generation just past, wrote in his textbook, the bible of the profession:

The presence of this great bird at a time when mammals were, for the most part, of very small size (the contemporary horse was the size of a fox terrier) suggests some interesting possibilities-which never materialized. The great reptiles had died off, and the surface of the earth was open for conquest. As possible successors there were the mammals and the birds. The former succeeded in the conquest, but the appearance of such a form as Diatryma shows that the birds were, at the beginning, rivals of the mammals.

In all these speculations about rerunning life's tape, we lament our lack of any controlled experiment. We cannot instigate the actual replay, and our planet provided only one run-through. Wiwaxiids are gone forever, and priapulids fell on a global scale. But the crucial Eocene pivot between birds and mammals provides more and different evidence. For once, our recalcitrant and complex planet actually performed a proper experiment for us. This particular tape did have a rerun in South America-and this time the birds won or at least held mammals to a respectable draw.

South America was an island continent, a kind of super-Australia, until the Isthmus of Panama rose just a few million years ago. Most animals usually considered as distinctively South American-jaguars, llamas and tapirs, for example-are North American migrants of post-isthmian arrival. The great native fauna of South America is largely gone (or surviving as a poor, if fascinating remnant of armadillos, sloths and the Virginia opossum, among others). No placental carnivores inhabited this giant ark. Most popular books tell us that the native South American carnivores were all marsupials, the so-called borhyaenids. They neglect to say that another prominent group-the phororhacids, or giant ground birds-fared just as well if not better. Phororhacids (see also: Carl Zimmer, 1997: Terror, take two. Discover, June 1997); online without figures click here ) also sported large heads and short, stout necks, but hey were not closely related to Diatryma. Birds had a second and separate try as dominant carnivores in South America, and this time they won.

From the smug outlook of our placental-centered parochialism, we may say that birds could triumph in South America only because marsupials are inferior to placentals and did not offer the kind of challenge that conquered predacious ground birds in Europe and North America. But can we be so sure? Borhyaenids could also be large and fierce, ranging to bear size and including such formidable creatures as Thylacosmilus, the marsupial sabertooth. We might also sneer and point out that, in any case, phororhacids quickly snuffed it (along with borhyaenids) as soon as superior placentals flooded over the rising isthmus. But this common saga of progress will not wash either. As G. G. Simpson, our greatest expert on the evolution of South American mammals, wrote in one of his last books (Splendid Isolation, Yale University Press, 1980):

It has sometimes been said that these and other flightless South American birds…survived because there were long no placental carnivores on that continent. That speculation is far from convincing….Most of the phororhacids became extinct before, only a straggler or two after, placental carnivores reached South America. Many of the borhyaenids that lived among those birds for many millions of years were highly predacious….The phororhacids …were more likely to kill than to be killed by mammals.

We must conclude, I think, that South America does represent a legitimate replay, run two for the birds.

One may properly ask, in conclusion, if all these reveries really make any difference. Who cares, in the old spirit of America at its pragmatic best? It is fun to imagine oneself as a sort of divine disk jockey, sitting before the tape machine of time with a library of cassettes labeled priapulids, pogonophorans, polychaetes and primates. But would it really matter if all the reruns discussed in this essay produced their unrealized opposite-and we inhabited a world of wiwaxiids, a sea littered with little penis worms, and forests full of phororhacids. We might be shucking sclerites instead of opening shells in clambakes. Our trophy rooms might vie for the longest Diatryma beak, not the richest lion mane. But what would be fundamentally different?

Everything, I suggest. Our literary metaphors have correctly captured the theme of contingency-from the old saw about kingdoms lost for want of a horseshoe nail, to Marty McFly struggling to unite his prospective parents in Back to the Future. The divine tape player holds a million scenarios, each perfectly sensible. Little quirks at the outset, chosen for no particular reason, unleash cascades of consequences that make a particular future seem inevitable in retrospect. But the slightest early nudge contacts a different groove, and history veers into another plausible channel, diverging continually from its original pathway. The end results are so different, the initial perturbation so apparently trivial. If little penis worms ruled the sea, I have no confidence that Australopithecus would ever have walked erect on the savannas of Africa. For ourselves, we can only exclaim: O brave- and improbable-new world, that has such people in it.

 Stephen Jay Gould teaches biology, geology, and the history of science at Harvard University.