92 S C I E N T I F I C A M E R I C A N R e p r i n t e d f r o m t h e O c t o b e r 1 9 9 4 i s s u e
ome creators announce their inventions with grand éclat. God proclaimed, “Fiat lux,” and then flooded his new universe with brightness. Others bring forth great discoveries in a modest guise, as did Charles Darwin in defining his new mechanism of evolu-
tionary causality in 1859: “I have called this principle, by which each slight variation, if useful, is preserved, by the term Natur- al Selection.”
Natural selection is an immensely powerful yet beautifully simple theory that has held up remarkably well, under intense and unrelenting scrutiny and testing, for 135 years. In essence, natural selection locates the mechanism of evolutionary change in a “struggle” among organisms for reproductive success, lead- ing to improved fit of populations to changing environments. (Struggle is often a metaphorical description and need not be viewed as overt combat, guns blazing. Tactics for reproductive success include a variety of nonmartial activities such as earlier and more frequent mating or better cooperation with partners in raising offspring.) Natural selection is therefore a principle of local adaptation, not of general advance or progress.
Yet powerful though the principle may be, natural selection is not the only cause of evolutionary change (and may, in many cases, be overshadowed by other forces). This point needs em- phasis because the standard misapplication of evolutionary the- ory assumes that biological explanation may be equated with devising accounts, often speculative and conjectural in practice, about the adaptive value of any given feature in its original en- vironment (human aggression as good for hunting, music and religion as good for tribal cohesion, for example). Darwin him- self strongly emphasized the multifactorial nature of evolu- tionary change and warned against too exclusive a reliance on natural selection, by placing the following statement in a max-
imally conspicuous place at the very end of his introduction: “I am convinced that Natural Selection has been the most impor- tant, but not the exclusive, means of modification.”
Reality versus Conceit NATURAL SELECTION is not fully sufficient to explain evo- lutionary change for two major reasons. First, many other caus- es are powerful, particularly at levels of biological organization both above and below the traditional Darwinian focus on or- ganisms and their struggles for reproductive success. At the low- est level of substitution in individual base pairs of DNA, change is often effectively neutral and therefore random. At higher lev- els, involving entire species or faunas, punctuated equilibrium can produce evolutionary trends by selection of species based on their rates of origin and extirpation, whereas mass extinc- tions wipe out substantial parts of biotas for reasons unrelat- ed to adaptive struggles of constituent species in “normal” times between such events.
Second, and the focus of this article, no matter how ade- quate our general theory of evolutionary change, we also yearn to document and understand the actual pathway of life’s his- tory. Theory, of course, is relevant to explaining the pathway (nothing about the pathway can be inconsistent with good the- ory, and theory can predict certain general aspects of life’s geo- logic pattern). But the actual pathway is strongly underdeter- mined by our general theory of life’s evolution. This point needs some belaboring as a central yet widely misunderstood aspect of the world’s complexity. Webs and chains of historical events are so intricate, so imbued with random and chaotic elements, so unrepeatable in encompassing such a multitude of unique (and uniquely interacting) objects, that standard models of sim- ple prediction and replication do not apply.
The history of life is not necessarily progressive; it is certainly
not predictable. The earth’s creatures have evolved through
a series of contingent and fortuitous events
evolution of life on earth
By Stephen Jay Gould
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 93
History can be explained, with satisfying rigor if evidence be adequate, after a sequence of events unfolds, but it cannot be pre- dicted with any precision beforehand. Pierre-Simon Laplace, echoing the growing and confident determinism of the late 18th century, once said that he could specify all future states if he could know the position and motion of all particles in the cosmos at any moment, but the nature of universal complexity shatters this chimerical dream. History includes too much chaos, or extremely sensitive dependence on minute and unmeasurable differences in initial conditions, leading to massively divergent outcomes based on tiny and unknowable disparities in starting points. And his- tory includes too much contingency, or shaping of present results by long chains of unpredictable antecedent states, rather than im- mediate determination by timeless laws of nature.
Homo sapiens did not appear on the earth, just a geologic second ago, because evolutionary theory predicts such an out- come based on themes of progress and increasing neural com- plexity. Humans arose, rather, as a fortuitous and contingent outcome of thousands of linked events, any one of which could have occurred differently and sent history on an alternative pathway that would not have led to consciousness. To cite just four among a multitude: (1) If our inconspicuous and fragile lin- eage had not been among the few survivors of the initial radia- tion of multicellular animal life in the Cambrian explosion 530 million years ago, then no vertebrates would have inhabited the earth at all. (Only one member of our chordate phylum, the genus Pikaia, has been found among these earliest fossils. This small and simple swimming creature, showing its allegiance to us by possessing a notochord, or dorsal stiffening rod, is among
the rarest fossils of the Burgess Shale, our best preserved Cam- brian fauna.) (2) If a small and unpromising group of lobe- finned fishes had not evolved fin bones with a strong central axis capable of bearing weight on land, then vertebrates might nev- er have become terrestrial. (3) If a large extraterrestrial body had not struck the earth 65 million years ago, then dinosaurs would still be dominant and mammals insignificant (the situa- tion that had prevailed for 100 million years previously). (4) If a small lineage of primates had not evolved upright posture on the drying African savannas just two to four million years ago, then our ancestry might have ended in a line of apes that, like the chimpanzee and gorilla today, would have become ecolog- ically marginal and probably doomed to extinction despite their remarkable behavioral complexity.
Therefore, to understand the events and generalities of life’s pathway, we must go beyond principles of evolutionary theory to a paleontological examination of the contingent pattern of life’s history on our planet—the single actualized version among millions of plausible alternatives that happened not to occur. Such a view of life’s history is highly contrary both to conven- tional deterministic models of Western science and to the deep- est social traditions and psychological hopes of Western culture for a history culminating in humans as life’s highest expression and intended planetary steward.
Science can, and does, strive to grasp nature’s factuality, but all science is socially embedded, and all scientists record pre- vailing “certainties,” however hard they may be aiming for pure objectivity. Darwin himself, in the closing lines of On the Ori- gin of Species, expressed Victorian social preference more than
SLAB CONTAINING SPECIMENS of Pteridinium from Namibia shows a prominent organism from the earth’s first multicellular fauna, called Ediacaran, which appeared some 600 million years ago. The Ediacaran animals died out before the Cambrian explosion of modern life. These thin,
quilted, sheetlike organisms may be ancestral to some modern forms but may also represent a separate and ultimately failed experiment in multicellular life. The history of life tends to move in quick and quirky episodes, rather than by gradual improvement.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
nature’s record in writing: “As natural selection works solely by and for the good of each being, all corporeal and mental endowments will tend to progress towards perfection.”
Life’s pathway certainly includes many features predictable from laws of nature, but these aspects are too broad and general to provide the “rightness” that we seek for validating evolution’s particular results—roses, mushrooms, people and so forth. Organisms adapt to, and are constrained by, physical princi- ples. It is, for example, scarcely surpris- ing, given laws of gravity, that the largest vertebrates in the sea (whales) exceed the heaviest animals on land (elephants to- day, dinosaurs in the past), which, in turn, are far bulkier than the largest ver- tebrate that ever flew (extinct pterosaurs of the Mesozoic era).
Predictable ecological rules govern the structuring of communities by principles of energy flow and thermodynamics (more biomass in prey than in predators, for example). Evolutionary trends, once started, may have local predictability (“arms races,” in which both predators and prey hone their defenses and weap- ons, for example—a pattern that Geerat J. Vermeij of the University of California at Davis has called “escalation” and doc- umented in increasing strength of both crab claws and shells of their gastropod
prey through time). But laws of nature do not tell us why we have crabs and snails at all, why insects rule the multicellular world and why vertebrates rather than persistent algal mats exist as the most complex forms of life on the earth.
Relative to the conventional view of life’s history as an at least broadly pre- dictable process of gradually advancing complexity through time, three features of the paleontological record stand out in opposition and shall therefore serve as organizing themes for the rest of this ar- ticle: the constancy of modal complexity throughout life’s history; the concentra- tion of major events in short bursts inter- spersed with long periods of relative sta- bility; and the role of external impositions, primarily mass extinctions, in disrupting patterns of “normal” times. These three features, combined with more general themes of chaos and contingency, require a new framework for conceptualizing and drawing life’s history, and this article therefore closes with suggestions for a dif- ferent iconography of evolution.
The Lie of “Progress” THE PRIMARY paleontological fact about life’s beginnings points to pre- dictability for the onset and very little for the particular pathways thereafter. The earth is 4.6 billion years old, but the old- est rocks date to about 3.9 billion years
because the earth’s surface became molten early in its history, a result of bombard- ment by large amounts of cosmic debris during the solar system’s coalescence and of heat generated by radioactive decay of short-lived isotopes. These oldest rocks are too metamorphosed by subsequent heat and pressure to preserve fossils (al- though some scientists interpret the pro- portions of carbon isotopes in these rocks as signs of organic production). The oldest rocks sufficiently unaltered to retain cellular fossils—African and Aus- tralian sediments dated to 3.5 billion years old—do preserve prokaryotic cells (bacteria and cyanophytes) and stroma- tolites (mats of sediment trapped and bound by these cells in shallow marine waters). Thus, life on the earth evolved quickly and is as old as it could be. This fact alone seems to indicate an inevit- ability, or at least a predictability, for life’s origin from the original chemical constituents of atmosphere and ocean.
No one can doubt that more com- plex creatures arose sequentially after this prokaryotic beginning—first eu- karyotic cells, perhaps about two billion years ago, then multicellular animals about 600 million years ago, with a relay of highest complexity among animals passing from invertebrates, to marine vertebrates and, finally (if we wish, albeit parochially, to honor neural architecture as a primary criterion), to reptiles, mam- mals and humans. This is the conven- tional sequence represented in the old charts and texts as an “age of inverte- brates,” followed by an “age of fishes,” “age of reptiles,” “age of mammals,” and “age of man” (to add the old gender bias to all the other prejudices implied by this sequence).
I do not deny the facts of the preced- ing paragraph but wish to argue that our conventional desire to view history as progressive, and to see humans as pre- dictably dominant, has grossly distorted our interpretation of life’s pathway by falsely placing in the center of things a relatively minor phenomenon that arises only as a side consequence of a physical- ly constrained starting point. The most salient feature of life has been the stabil- ity of its bacterial mode from the begin-
94 S C I E N T I F I C A M E R I C A N D I N O S A U R S A N D O T H E R M O N S T E R S
PROGRESS DOES NOT RULE (and is not even a primary thrust of) the evolutionary process. For reasons of chemistry and physics, life arises next to the “left wall” of its simplest conceivable and preservable complexity. This style of life (bacterial) has remained most common and most successful. A few creatures occasionally move to the right, thus extending the right tail in the distribution of complexity. Many always move to the left, but they are absorbed within space already occupied. Note that the bacterial mode has never changed in position, but just grown higher.
Left wall of minimal complexity
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
ning of the fossil record until today and, with little doubt, into all future time so long as the earth endures. This is truly the “age of bacteria”—as it was in the be- ginning, is now and ever shall be.
For reasons related to the chemistry of life’s origin and the physics of self- organization, the first living things arose at the lower limit of life’s conceivable, preservable complexity. Call this lower limit the “left wall” for an architecture of complexity. Because so little space exists between the left wall and life’s initial bac- terial mode in the fossil record, only one direction for future increment exists—to- ward greater complexity at the right. Thus, every once in a while, a more com- plex creature evolves and extends the range of life’s diversity in the only avail- able direction. In technical terms, the dis- tribution of complexity becomes more strongly right skewed through these oc- casional additions.
But the additions are rare and epi- sodic. They do not even constitute an evo- lutionary series but form a motley se- quence of distantly related taxa, usually depicted as eukaryotic cell, jellyfish, trilo- bite, nautiloid, eurypterid (a large relative of horseshoe crabs), fish, an amphibian such as Eryops, a dinosaur, a mammal and a human being. This sequence can- not be construed as the major thrust or trend of life’s history. Think rather of an occasional creature tumbling into the empty right region of complexity’s space. Throughout this entire time, the bacteri- al mode has grown in height and re- mained constant in position. Bacteria rep- resent the great success story of life’s path- way. They occupy a wider domain of environments and span a broader range of biochemistries than any other group. They are adaptable, indestructible and astoundingly diverse. We cannot even imagine how anthropogenic intervention might threaten their extinction, although we worry about our impact on nearly every other form of life. The number of Escherichia coli cells in the gut of each hu- man being exceeds the number of hu- mans that has ever lived on this planet.
One might grant that complexifica- tion for life as a whole represents a pseudotrend based on constraint at the
left wall but still hold that evolution with- in particular groups differentially favors complexity when the founding lineage begins far enough from the left wall to permit movement in both directions. Em- pirical tests of this interesting hypothesis are just beginning (as concern for the sub- ject mounts among paleontologists), and we do not yet have enough cases to ad- vance a generality. But the first two stud- ies—by Daniel W. McShea of the Uni- versity of Michigan on mammalian ver- tebrae and by George F. Boyajian of the University of Pennsylvania on ammonite suture lines—show no evolutionary ten- dencies to favor increased complexity.
Moreover, when we consider that for each mode of life involving greater com- plexity, there probably exists an equally advantageous style based on greater sim- plicity of form (as often found in para- sites, for example), then preferential evo- lution toward complexity seems unlikely a priori. Our impression that life evolves toward greater complexity is probably only a bias inspired by parochial focus on ourselves, and consequent overattention to complexifying creatures, while we ig-
nore just as many lineages adapting equally well by becoming simpler in form. The morphologically degenerate parasite, safe within its host, has just as much prospect for evolutionary success as its gorgeously elaborate relative cop- ing with the slings and arrows of outra- geous fortune in a tough external world.
Steps, Not Inclines EVEN IF COMPLEXITY is only a drift away from a constraining left wall, we might view trends in this direction as more predictable and characteristic of life’s pathway as a whole if increments of complexity accrued in a persistent and gradually accumulating manner through time. But nothing about life’s history is more peculiar with respect to this com- mon (and false) expectation than the ac- tual pattern of extended stability and rapid episodic movement, as revealed by the fossil record.
Life remained almost exclusively uni- cellular for the first five sixths of its his- tory—from the first recorded fossils at 3.5 billion years to the first well-doc- umented multicellular animals less than
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 95
NEW ICONOGRAPHY OF LIFE’S TREE shows that maximal diversity in anatomical forms (not in number of species) is reached very early in life’s multicellular history. Later times feature extinction of most of these initial experiments and enormous success within surviving lines. This success is measured in the proliferation of species but not in the development of new anatomies. Today we have more species than ever before, although they are restricted to fewer basic anatomies.
Anatomical Diversity Ti
STEPHEN JAY GOULD taught biology, geology and the history of science at Harvard Uni- versity from 1967 until his death in 2002 at age 60. The influential and provocative evo- lutionary biologist had a Ph.D. in paleontology from Columbia University. Well known for his popular writings, in particular a monthly column in Natural History magazine, he was the author of more than a dozen books, including Full House: The Spread of Excellence from Plato to Darwin and The Mismeasure of Man. T
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
600 million years ago. (Some simple multicellular algae evolved more than a billion years ago, but these organisms be- long to the plant kingdom and have no genealogical connection with animals.) This long period of unicellular life does include, to be sure, the vitally important transition from simple prokaryotic cells without organelles to eukaryotic cells with nuclei, mitochondria and other com- plexities of intracellular architecture— but no recorded attainment of multicel- lular animal organization for a full three billion years. If complexity is such a good thing, and multicellularity represents its initial phase in our usual view, then life certainly took its time in making this cru- cial step. Such delays speak strongly against general progress as the major theme of life’s history, even if they can be plausibly explained by lack of sufficient atmospheric oxygen for most of Precam-
brian time or by failure of unicellular life to achieve some structural threshold act- ing as a prerequisite to multicellularity.
More curiously, all major stages in or- ganizing animal life’s multicellular archi- tecture then occurred in a short period be- ginning less than 600 million years ago and ending by about 530 million years ago—and the steps within this sequence are also discontinuous and episodic, not gradually accumulative. The first fauna, called Ediacaran to honor the Australian locality of its initial discovery but now known from rocks on all continents, con- sists of highly flattened fronds, sheets and circlets composed of numerous slender segments quilted together. The nature of the Ediacaran fauna is now a subject of intense discussion. These creatures do not seem to be simple precursors of later forms. They may constitute a separate and failed experiment in animal life, or
34. Sidneyia 35. Odaraia 36. Eiffelia 37. Mackenzia 38. Odontogriphus 39. Hallucigenia 40. Elrathia 41. Anomalocaris 42. Lingulella 43. Scenella 44. Canadaspis 45. Marrella 46. Olenoides
22. Emeraldella 23. Burgessia 24. Leanchoilia 25. Sanctacaris 26. Ottoia 27. Louisella 28. Actaeus 29. Yohoia 30. Peronochaeta 31. Selkirkia 32. Ancalagon 33. Burgessochaeta
11. Micromitra 12. Echmatocrinus 13. Chancelloria 14. Pirania 15. Choia 16. Leptomitus 17. Dinomischus 18. Wiwaxia 19. Naraoia 20. Hyolithes 21. Habelia
1. Vauxia (gracile) 2. Branchiocaris 3. Opabinia 4. Amiskwia 5. Vauxia (robust) 6. Molaria 7. Aysheaia 8. Sarotrocercus 9. Nectocaris