The Structure of Evolutionary Theory

by Stephen Jay Gould

  In a general argument strikingly similar to Blackburn's for the evolution of squamate viviparity, Smith (1994) holds that gradualistic assumptions have stymied our understanding of evolutionary processes at the small scale of ecological immediacy in deep-sea faunas. No other environment has been so conventionally associated with plodding, incremental change through substantial periods of time. Smith begins his article by noting the “the deep-sea floor is traditionally perceived as a habitat where low food flux and sluggish bottom currents force life to proceed at slow, steady rates. In this view, benthic com­munity structure is controlled by equilibrium processes, such as extreme lev­els of habitat partitioning, made possible by remarkable ecosystem stability” (Smith, 1994, p. 3).

  As indicated by the title of his article — “Tempo and mode in deep-sea ben­thic ecology: punctuated equilibrium revisited” — Smith holds that we must revise this traditional view, and reconceive the deep-sea as a punctuational domain where “endogenous disturbances may be relatively frequent, and where pulses of food reach the seafloor from the upper ocean” (p. 3). In what he labels as a “parallel argument” to our punctuated equilibrium from a much lower scale of analysis — in other words, as a claim for conceptual homology of constraining structural principles (in the language of this sec­tion) — Smith discusses three examples of “pulsed events that 'punctuate' the apparent 'equilibrium' of the deep-sea floor” (p. 3), and that “may sub­stantially influence processes of modern and past ecological significance in­cluding (1) maintenance of macrofaunal diversity and population structure, (2) deposit-feeder-microbe interactions and associated trace production, and (3) dispersal and biography of chemosynthetic communities at the deep-sea floor” (p. 3).

  First, Smith documents the importance of “pulsed physical disturbance” in benthic faunas of the Nova Scotian Rise (4750-4950 m) — particularly of erosional “storms” that scour and redeposit sediments “to depths of millimeters-centimeters over areas encompassing at least tens of square kilometers” (p. 7), and that strongly influence both the composition and successional stage of local faunas.

  In a second microbiotal example, Smith documents the importance of “phytodetrital pulses” in nutrition for the deep-sea macrofauna. The “slow and steady 'drizzle'” usually regarded as the gradual (and meager) planktonic [Page 948] contribution to sustaining deep-sea life “can be punctuated by downpours of 'phylodetritus' (i.e., detrital material composed primarily of relatively fresh phytoplanktonic remains), during which the flux of labile particulate organic carbon to the seafloor temporarily exceeds biological demand, yielding a car­pet of’ food'” (p. 7).

  Finally, and to add a third punctuational source of maximally different character from the physical and microfloral cases discussed above, Smith argues (p. 10) that “whale falls” produce occasional and (obviously) “huge local pulses” of organic matter that may decay to produce distinctive “chemosynthetic habitats” supporting faunal associations much like those documented at deep-sea vents. For example, in 1987, his team discovered a 21-meter whale skeleton at a depth of 1240 m: “The bones were covered with mats of sulfur bacteria and clusters of small mussels and limpets; nearby sedi­ments harbored large vesicomyid clams” (p. 10) — for a total of 42 macro-faunal species, only nine of which also inhabited surrounding sediments. Smith concludes that “sunken whales may provide dispersal stepping stones for at least some of the species dependent on sulfide-based chemosynthesis.”

  Strong circumstantial evidence indicates considerable temporal and spatial influence for this source that most of us would surely have regarded as dubi­ous, if not risible, at apparent face value of relative importance. A fossilized chemosynthetic community has been reported from a 35 million year old whale fall on the Northeast Pacific ocean floor (p. 10). “Whale skeletons,” Smith concludes (p. 10), “may be the dominant source of chemosynthetic habitats over the vast sediment plains constituting most of the ocean floor.”

  At the opposite end of a hierarchy in spatial and temporal scales, punctuational models continue to gain in strength and acceptability for events that impact entire biotas at regional or even planetary scales — with catastrophic mass extinction as a “flagship” notion, spurred by nearly conclusive evidence for bolide impact as the trigger of the K-T global dying 65 million years ago (see Chapter 12 for full treatment). An expansion of research away from the extinctions themselves, and towards the subsequent recovery phases as well, has strongly accentuated the episodic and punctuational character of this most comprehensive signal in the history of life.

  Even after the Alvarez's impact hypothesis forced paleontologists to acknowledge the potentially catastrophic nature of at least some mass extinc­tions, students of fossils usually assumed that subsequent recoveries of global faunas must have been tolerably gradual. This expectation has not been fulfilled, and episodes of recovery from maximum decimation at the extinc­tion to full reestablishment of previous levels of diversity occur more quickly, and in a much shorter percentage of the “normal” time (until the next mass extinction), than previously suspected. (Of course, no one expects that recov­eries which require successive events of branching can be nearly as rapid as truly catastrophic extinctions, which can feature truly simultaneous kill­ings — so the complete record of an extinction-recovery cycle will surely re­main asymmetric. But the recoveries now seem to occur rapidly enough, in [Page 949] most cases, to invoke the central concept of punctuational change: origin in a tiny fraction of later existence in stasis.)

  For example, Kerr (1994) begins his report on Peter Sheehan's work (in a commentary entitled “Between extinctions, evolutionary stasis”) by writing (Kerr, 1994, p. 29): “More and more, paleontologists are learning that the full measure of a mass extinction can't be found in its immediate toll. Just as important is the wholesale reorganization of living communities that takes place afterward. And those brief recovery periods, lasting just a few million years, are all the more important because during the tens of or hundreds of millions of years that follow, until the next mass extinction, not much may happen.”

  Sheehan divides the last 640 million years into six major faunal packages that he calls EEU's, or Ecologic Evolutionary Units. Each lasts for 35 to 147 million years, and each ends at a mass extinction. The subsequent recovery periods for the new units occupy only 3 to 8 million years.

  This recent affirmation of a strongly punctuational character for change (primarily extinction) at the highest level has led to a tendency, probably overextended — and I blame myself, in part, for propagating the theme, see Gould, 1985a — for ascribing a dualistic character to the pulse of evolution, with punctuations of mass extinction alternating with a more stately flow in “normal” times between these macropulses. But this view may prove to be overly simplistic, although not wrong. When we assess each level of change by its own appropriate measuring rod (scaled to emphasize the relevant unit or units), all may be punctuational. We must dismiss as irrelevant and mis­leading the fact that punctuations at a small scale may “smooth out” to more gradual and continuous trends when inappropriately measured at too large a scale to reveal the causal mechanics, or even to identify the relevant unit, of change — a theme that I have emphasized throughout this chapter, in such ex­amples as punctuated bacterial anagenesis, viewed as gradual when sampled too infrequently to note the steps of mutational sweeps; and cladal trends, viewed as anagenetic when sampled too broadly to discern the speciational jumps of punctuated equilibrium.

  In a provocative work, Raup (1992) played devil's advocate by asking if all extinctions at all levels, from single local populations to global faunas, might be catastrophic — for he could not reject the “null hypothesis” of his “field of bullets” model (random and catastrophic removal, triggered by “bolides” of various sizes randomly shot towards the earth at frequencies inversely pro­portional to their size and effect) in favor of the traditional Darwinian model of gradual declines mediated by competitive inferiority in biotic in
teractions. I do not believe that such extreme punctuationalism could rule so completely (see full discussion of this argument in Chapter 12, pp. 1323–1326). But finer analysis of the most famous cases of supposedly gradual, and biotically con­trolled, events may well require such a punctuational reinterpretation. Most outstandingly, perhaps the two most widely discussed and most generally ac­cepted examples of geologically slow global diversification — the Ordovician [Page 950] spread of the great Paleozoic marine invertebrate fauna, and the Mesozoic “modernization” of invertebrate predators and prey (the classic example of a supposed biotic and gradualistic “arms race”) — now seem to occur far more abruptly in each separate geographic region, with the previous impression of gradual construction based on a blurring of the different times of transition in each region (Jablonski, 1999, p. 2114).

  In an important paper, Miller (1998) has generalized this claim by summarizing the increasing evidence for punctuational tempos in faunal change (both locally and regionally, and for both extinctions and the necessarily slower rediversifications) — with our conventional notions of gradual flux, particularly for build-ups, arising as an artifact of summation over displaced timings for rapid pulses in several regions. Miller first states the general ob­servation and emerging principles (1998, pp. 1158-1159): “In recent years, local and regional studies of marine faunal patterns have converged on a sim­ilar theme — that biotic turnover occurred episodically through investigated stratigraphic intervals. There were comparatively broad intervals with little net turnover, punctuated by narrower intervals in which many taxa either emigrated or became extinct and were replaced by a roster of taxa that either originated in the area or immigrated into it. ... Episodicity appears to be a general feature of regional stratigraphic packages.” He then uses this find­ing to correct what may be a substantial error in traditional views (1998, p. 1159): “Thus, major faunal transitions in global-scale compilations, which seem to have transpired over protracted intervals of geological time, took place far more rapidly and episodically when evaluated regionally or locally. The transitions only appear gradual on a global scale because of variations in their timing from venue to venue.”

  Finally, Miller asserts a general “fractal” conclusion about punctuational change (ibid., p. 1159): “The processes that produced major mass extinctions simply represented the largest and most globally extensive of a spectrum of perturbations that produced episodic biotic transitions.”

  As a closing note in this context, Miller also offers a similar punctuational reinterpretation for the putatively best documented and most widely accepted case of global, geologically gradual, and broadly progressive change in life's history — the pattern that Vermeij (1987) has called “escalation” (largely, and with good reason, to avoid false implications and arguments in the tradi­tional notion of “progress”), based on relayed “arms race” between preda­tors and their prey, and on other kinds of similarly reciprocal biotic inter­action through extensive time. This entirely sensible concept of escalation seemed to provide the best available argument for two deeply rooted and strongly held themes of traditional Darwinian extrapolationism: the predom­inant power of biotic interactions to shape patterns in the history of life, and trends towards the slow accumulation of biomechanically improved designs in major lineages.

  The general argument sounds so reasonable, but when we rethink macroevolution as a process based upon geologically rapid production of higher-level individuals by punctuational speciation as the primary units of change, [Page 951] then the mechanics of this usual interpretation of escalation become elusive. The pattern certainly exists — especially for Vermeij's (1977) classic case of in­creasing strength and efficiency in crab claws matched by growing intricacy and sophistication of adaptive defenses in molluscan shells. But how can such an arms race operate if the full trend proceeds by stepping stones of punctuational speciation for any increment, and not in the style of tit-for-tat anagenetic escalation, based on immediate organismal competition and more familiar to us through human models of “anything you can do, I can do better” — a point that Vermeij himself recognizes and finds puzzling (1987)?

  Miller, on the other hand, affirms the gradual trend to escalation in biomechanical improvement — and I don't think that any party to this debate de­nies the reality of the pattern (for we have been arguing about mechanisms) — but finds the same unconventional (and punctuational), finer-scale theme upon “dissecting” the full result into component causal units. Again, each step in escalation seems episodic in each region, with the full trend thus rendered as a summation of punctuational events. Miller writes (p. 1159): “Although the case for these kinds of transitions over the sweep of the Phanerozoic is difficult to deny, the manner in which they transpired over shorter intervals is less certain. There is little evidence of gradual escalation through stratigraphic intervals at local or regional levels. The introduction of escalated forms appears to have occurred episodically, in concert with the broader class of changes in taxonomic composition discussed earlier, which suggests a role for physical mediation.”

  Two general points provide a fitting close for this section:

  1. The probable generality of punctuation and stasis as a powerful — if not predominant — style of change across all scales must lead us to reassess our previous convictions about “important” and “interesting” phenomena in evolutionary theory and the history of life. Kerr admits the potential general­ity, in reporting punctuation at lower levels to complement Sheehan's similar claim for the broadest scale. But he closes his report by writing (1994, p. 29): “Sheehan sees these intervals as analogous to his longer, global EEU's reaf­firming that stability — as boring as it may be — is the evolutionary norm.” But, to restate my mantra, and to emphasize its implications for understand­ing the history of earthly life and the psychology of human discovery, stasis is data — and data of such high generality, such unanticipated occurrence, and such theoretical interest simply cannot be boring.

  2. The ubiquity — and the possibly canonical character — of punctuational change at all scales, from the shortest trends of bacterial anagenesis in single clonal lineages over weeks to months, to the broadest patterns of global wax­ing and waning of biotas through the history of life in deep time, can only re­call the familiar tale, by now a cliche, of the Eastern sage who revealed the nature of the cosmos to his disciple: the globe of the earth rests on the back of an elephant who stands, in turn, on the back of a turtle. When asked by the disciple what one might find under the turtle, from its feet to the ultimate source of being, the sage simply replies: “it's turtles all the way down.” I suspect that it is also punctuational change all the way down, from Permian [Page 952] extinctions to mutational sweeps through little laboratory populations of E. coli.

  Punctuational models in other disciplines: towards a general

  theory of change

  PRINCIPLES FOR A CHOICE OF EXAMPLES. In their symposium for the American Association for the Advancement of Science, and in their subse­quent book, Somit and Peterson (1992) explored the wider role of punctu­ated equilibrium in suggesting similar modes of change in other disciplines. (Their edited book bears the title: The Dynamics of Evolution: The Punctu­ated Equilibrium Debate in the Natural and Social Sciences.) In discussing the “manner in which punctuated equilibrium theory renders its greatest con­tribution to the behavioral sciences” (1992, p. 12), they suggested (loc. cit): “By providing a different metaphor for explaining social phenomena, the the­ory may assist us in better understanding human behavior in all of its mani­festations.”

  I don't question either the widespread invocation or the extensive utility of the metaphorical linkage, and I list elsewhere (pp. 976–979) a range of such invocations across disciplines from economics to cartooning to guidelines for the self-help movement. But in discussing the application of punctuated equi­librium to other disciplines, I am more interested in exploring ways in which the theory might supply truly causa
l insights about other scales and styles of change, based on conceptual and structural “homologies” (as defined and discussed on pp. 928–931), rather than broader metaphors that can surely nudge the mind into productive channels, but that make no explicit claim for causal continuity or unification. Thus, in discussing the influence of punctu­ated equilibrium upon other disciplines, I will focus upon two kinds of poten­tially homological proposals.

  First, where authors proceed beyond simple claims for broad similarities in jerky tempos of change to identify additional and explicit overlaps in the set of collateral principles that I called “conjoints” (see p. 930) in defining con­ceptual homology vs. analogy — including, for example, (1) claims that link punctuations to the origin of discretely individuated units arising by branch­ing (a conceptual homolog of speciation), (2) discussions of the difference be­tween punctuational and saltational modes, and (3) proposals about active causes for the maintenance of stasis. And second, where authors use the simi­larities between punctuated equilibrium and punctuational tempos in their own discipline to advance more than vaguely metaphorical suggestions for general theories about the nature of change in systems that may be said to “evolve,” and to display historical continuity.