The Structure of Evolutionary Theory

by Stephen Jay Gould

  Such broad arguments about environmental correlations have been notoriously difficult to document because, even when the effect can be validated as both real and pervasive, so many other factors will be operating in any partic­ular case (including immediate and local influences able to overwhelm the smaller impact of the generality under test) that the signal may be lost in surrounding noise. But I am strongly attracted to Sheldon's plus ca change hypothesis for two primary reasons. First, the concept makes good sense of patterns that have often been noted empirically, but regarded as con­fusing in interpretation — particularly the common finding of pronounced sta­sis through major climatic fluctuations, including Pleistocene ice age cycles (Cronin, 1985, on ostracodes; Coope, 1980,1994 on beetles), and the largest climatic crash in Tertiary North America (Prothero and Heaton, 1996). The presentation of a hypothesis like Sheldon's prompts researchers to focus stud­ies on interesting issues, and to seek wider implications. For example, Wei (1994) used Sheldon's hypothesis to explain the link of stasis to intensifica­tion of ice-age climatic fluctuations in the planktonic foram Globoconella inflata. Wei (1994, p. 81) suggests that stasis may represent “a compromise for the species as an attempt to meet with both glacial and interglacial ex­tremes.”

  Second, Sheldon's hypothesis predicts a large suite of definite correlations subject to empirical test. Plus ca change predicts linkages of different relative frequencies for punctuated equilibrium and gradualism to geographic gradi­ents (with more punctuated equilibrium expected in temperate areas, and more gradualism in the topics), environmental distinctions (with more punc­tuated equilibrium in near shore shallow-water strata and more gradualism in offshore regions, as Johnson had earlier predicted), and evolved responses of organisms and populations (with, ceteris paribus, more punctuated equilib­rium in eurytopes and r-strategists, and more gradualism for stenotopes and K-strategists). Needless to say, ceteris paribus does not always hold — but with so many expected consequences, the probability of finding patterns (if they exist) does rise substantially.

  Finally, as for any good hypothesis in science, Sheldon's plus ca change suggests several interesting extensions. For example, Sheldon raises an intriguing argument for linking these putative correlations with patterns of genuine se­lection at the species level or above:

  Perhaps the most important (and perhaps the most controversial) mechanism I am suggesting here is a type of lineage selection with two stages: [Page 874] (1) if an established or an incipient species experiences a widely fluctu­ating environment on geological timescales, the evolutionary response (morphological change) tends to become damped with time, and (2) those species that are least sensitive to environmental change (the most “generalized” in a long-term sense) are the ones that tend to persist, remaining in morphological stasis until a threshold is reached (Sheldon, 1996, p. 218).

  In a second extension, Sheldon makes an almost quizzical, but oddly compelling, argument based on another important source of potential correla­tions, previously unaddressed here but perhaps quite important: time itself, expressed either as the absolute time of particular intervals in the earth's his­tory, or as the relative time of distinctive segments in the general “ontogeny” of a species's duration. Many biologists have noted the apparent paradox that so little sustained and directional evolution (as opposed to abundant evidence for rapid and adaptive fluctuations in such characters as bill form in Darwin's finches or wing colors in peppered moths) has been noted for species in his­toric, and recent prehistoric, times during the tenure of modern humans (who have also remained in stasis) on earth. I would, of course, attribute this phe­nomenon mostly to a general prediction for stasis in the vast majority of lin­eages at any time (while charging our puzzlement only to the false equation of evolution with gradual change). But Sheldon raises the interesting ancillary argument that this general expectation may now be enhanced by special ad­vantages for stasis in the regimes of strong and rapid worldwide climatic fluc­tuation that our earth has been experiencing in these geologically unusual times: “Given the Quaternary climate upheavals, relatively little evolution may be occurring worldwide at present (except for evolution induced by hu­mans)” (Sheldon, 1996, p. 209). I can only hope that the more punctuated equilibrium induces change in our evolutionary views, the more things will not be the same in our interpretations of the history of life.

  The Broader Implications of Punctuated Equilibrium for

  Evolutionary Theory and General Notions of Change

  WHAT CHANGES MAY PUNCTUATED EQUILIBRIUM INSTIGATE

  IN OUR VIEWS ABOUT EVOLUTIONARY MECHANISMS AND

  THE HISTORY OF LIFE?

  The explanation and broader meaning of stasis

  As emphasized throughout this chapter, the stress placed by punctuated equilibrium upon the phenomenon of stasis may emerge as the theory's most important contribution to evolutionary science. The material world does not impact our senses as naturally and objectively parsed categories. We can make accurate observations and measures of particular “things,” but the or­dering of “things” into categories must be construed largely as a mental oper­ation based on our theories and attitudes towards “reality.” Moreover, we [Page 875] must also apply mental screening to select “things” meriting our attention within nature's potential infinity, and even to recognize a configuration of matter as a “thing” in the first place. Therefore, phenomena without names, and without theories marking them as worthy of notice, will probably not be recognized at all.

  The phenomenon always existed “out there” in nature, of course, but punctuated equilibrium largely “created” the category of stasis as an impor­tant item in evolutionary theory through a four-step process of (1) defining stasis as a positive “thing” with properties and boundaries, a phenomenon rather than an unnamed and unrecorded absence of evolution; (2) bringing stasis to visibility as the expectation of a particular theory of evolutionary modalities; (3) suggesting methods for the active and rigorous study of stasis, so that the concept could be operationalized as a subject for empirical re­search; and (4) granting interest and importance to stasis as a controversial topic with broad implications for revising traditional modes of thought in evolutionary biology.

  Before Eldredge and I published our first paper in 1972, most paleontolo­gists treated stasis as an embarrassment, imposed by the poverty of the fossil record upon hopes for recording evolution (defined as gradualistic anagene­sis), and therefore as not meriting active study, or even explicit recognition as a discrete phenomenon. Just a decade later, the situation had changed so dra­matically that Wake, Roth and Wake (1983) could write, “perhaps no phe­nomenon is as challenging to evolutionary biologists as what has been termed 'stasis'” (p. 212), defined by them as “the maintenance of a standard mor­phology over vast periods of time during which much environmental change has taken place” (p. 211). Illustrating my claim that a phenomenon becomes interesting only in the light of defining theories, Wake et al. (1983, p. 212) then stated: “With natural selection operating in a changing environment as an agent of adaptation, we expect to see changes at the organismal, ulti­mately physiological and morphological, level. How, though, can we explain the paradoxical situation in which environments change, even dramatically, but organisms do not?”

  As I now survey the subject, a quarter century after our initial presentation and definition, stasis has become an even more general and important issue in evolutionary theory for three principal reasons:

  FREQUENCY. Once the phenomenon had been named, and criteria estab­lished for recognition and study, researchers documented stasis at far too high a relative frequency to represent anything other than an evolutionary norm and expectation. Such predominance also implicates stasis as a property ac­tively maintained by species — thus leading to a substantial literature (dis­cussed at the end of this subsection) on the causes of non-change. Several au­thors, notably Paul (1985) and Jackson and Cheetham (1994, also Cheetham and Jackson, 1995), develo
ped models and data sets to prove that stasis oc­curs too frequently for explanation under random models (including pure neutralism with no natural selection), and therefore must be caused by active [Page 876] forces promoting such a result, either directly, or as a consequence of some important linked property of organisms or populations.

  This growth in emphasis has been so vigorous since 1972 that geological gradualism, once the unquestioned expectation of evolution itself, is now generally regarded as an infrequent, if not anomalous, phenomenon requiring a special explanation in the light of anticipated stasis. Geary (1990, p. 507) after documenting a case of gradualism within a clade showing a much higher frequency of stasis, wryly noted: “Given that past studies were assumed com­plete only if gradual change was apparent, it seems somewhat ironic that un­seen mechanisms or events, however realistic, must now be invoked in order to explain an instance of gradual change!” Gradualism, in short, has become both a rarity and a puzzle. (Much as I take a rather wickedly and secret per­sonal pleasure in this sea change, I'm not sure that I can, in good scientific conscience, regard such a priori mental downgrading of gradualism as a “good thing.” A posteriori downgrading based on documented rarity repre­sents nature's chief signal in my view, but I do think that any study should be­gin with equal potential welcome for either result!)

  GENERALITY. The interest in stasis, originally generated by punctuated equilibrium for inquiries at the appropriate level of species durations through time, has since expanded to other domains of size and time, and to more com­prehensive questions about the nature of change itself. Causes operating at punctuated equilibrium's proper scale will not explain other forms of stasis, but the generalized definition and inquiry did arise by expansion from our theory (at least as a sociological phenomenon), while we may also anticipate the identification of some common causes or constraints (see further dis­cussion on conceptual “homology,” pp. 928–931) — that is, in evolutionary parlance, causal parallelisms, based on structural homologies, rather than convergences or mere analogies of appearance — behind the deeper generality (with different immediate forces producing similar and partly homologous results at various levels). I shall discuss some of these other scales in Part B of this subsection. These extensions include: punctuational anagenesis for di­rectional changes in lineages of asexual organisms by clonal sorting (in a do­main below punctuated equilibrium, which, sensu stricto, only operates at the level of speciation to explain trends in multicellular sexual lineages by species sorting); longterm morphological stability for basic anatomical fea­tures of larger clades (at a level above punctuated equilibrium and within monophyletic lineages — see Chapter 10); putative “lock-step” stasis for the great majority of defining species within larger faunas through significant geological intervals (at a still higher level above punctuated equilibrium and across genealogical lineages to a consideration of faunal dynamics — see pp. 916–922). Interest has also extended beyond evolutionary systems to the meaning and causes of stasis in stairstep patterns of ontogenetic growth, stubbornly persistent plateaus followed by thresholds of rapid change in re­sponse to continuous input in human learning, and active stasis followed by [Page 877] punctuational breakdown in the history of human ideas and social organiza­tion (see pp. 952–967).

  CAUSALITY. Fruitful debate about the causes of stasis must first specify the level manifesting the common phenomenology. (Obviously, causes of learning plateaus in piano playing cannot be strict homologs of ecological reasons for joint stability of species in coordinated stasis, even though the graphed pat­tern of change may manifest the same geometrical form.) In this section, I confine my discussion to punctuated equilibrium sensu stricto, and not to the general pattern of punctuational change at any level — that is, to proposed reasons for the observed high relative frequency of stasis during the full geo­logical range of metazoan species as preserved in the fossil record.

  But first, and as an example of how discussion can proceed at cross pur­poses when proper scales have not been specified, I must note that most of the literature proclaiming punctuated equilibrium as “old hat” (Lewin, 1986) or something long known and merely hyped by ill-informed paleontologists, has only analyzed ecologically rapid anagenesis in populations rather than the relevant phenomenon of cladogenesis by speciation scaled against subsequent geological duration in the stasis of species so generated. Most notably (in terms of subsequent commentary), two papers of the mid-80's (Newman, Co­hen and Kipnis, 1985; and Lande, 1986) developed mathematical models to show that single populations could move rapidly (in the “ecological time” of a human career) from one adaptive peak to another in the absence of environ­mental change. (The major previous stumbling block had been set by prob­lems in envisaging how a population could move down an adaptive peak, against any force of selection, to inhabit a valley, and therefore become sub­ject to selection up an adjacent peak. The basic solution — that the descent must be rapid — allows sufficient impetus against selective forces, and also links the models to themes of speedy anagenesis.)

  Lewin (1986) used these studies to write a news and views feature for Sci­ence entitled “Punctuated equilibrium is now old hat,” while also recognizing that ecologically rapid anagenesis does not address the scale or level (not to mention the reality of changing environments in our actual world) of punctu­ated equilibrium's central concern. We welcome such plausible models of eco­logically punctuated anagenesis as a contribution to understanding the pano­ply of causes that yield punctuational change at other levels. But this smaller-scale phenomenon, however fascinating and important, bears little relevance to the causes of stasis within species during geological time (or to the cladogenetic sources of geological punctuation as a slow branching event in eco­logical time).

  We may order the major propositions for explaining stasis at the scale of punctuated equilibrium as an array running from conventional resolutions based on Darwinian organismic selection to more iconoclastic proposals in­voking either higher levels of causation or less control by selection and adap­tation. (Much of the genuine interest in the otherwise tedious and tendentious [Page 878] debate on the theoretical novelty of punctuated equilibrium lies in the legiti­mate weights that will eventually be assigned to the various proposals of this array.)

  Stabilizing selection. For most evolutionists who chose to see nothing new in punctuated equilibrium, the previously unacknowledged fre­quency of stasis (admitted, albeit sometimes begrudgingly, as an unexpected finding) could only indicate a stronger role than previously envisaged for the conventional mechanism of stabilizing selection. Although this putative explanation of stasis within paleospecies achieved an almost canonical sta­tus among evolutionists who tried to forge complete compatibility between punctuated equilibrium and the Modern Synthesis, and although we all ac­knowledge stabilizing selection as too important and pervasive a phenome­non to hold no relevance for this issue, a complete explanation of stasis in these conventional terms seems implausible both on empirical grounds, and also by the basic logic of proper scaling.

  As often emphasized in this chapter, if stasis merely reflects excellent adaptation to environment, then why do we frequently observe such profound stasis during major climatic shifts like ice-age cycles (Cronin, 1985), or through the largest environmental change in a major interval of time (Prothero and Heaton, 1996)? More importantly, conventional arguments about stabilizing selection have been framed for discrete populations on adaptive peaks, not for the totality of a species — the proper scale of punctuated equilibrium — so often composed of numerous, and at least semi-independent, subpopulations. A form of stabilizing selection acting among rather than within subpopu­lations may offer more promise — as Williams (1992) has proposed (see dis­cussion under point 6) — but such forms of supraorganismal selection fall into a domain of heterodoxies, not into this category of conventional explana­tions that would leave the Modern Synthesis entirely unaffected by the recog­nition of stasis as a paleontolo
gical norm.

  Developmental and ecological plasticity. If stabilizing selec­tion holds that species don't change because they have achieved such excel­lence in current adaptation, this second proposal (of Wake, Roth and Wake, 1983) proposes that species don't change (in an evolutionary and genetic sense) because they can usually accommodate to environmental alteration by exploiting the plasticity (behavioral and developmental) permitted within their existing genetic and ontogenetic system — thus calling upon the physiol­ogist's entirely different meaning of the term “adaptation” (improvement in functionality by exploiting possibilities within a norm of reaction, as in the enlarged lungs of people who inhabit the high Andes), rather than the usual evolutionary meaning in our profession.

  (Although I have roughly ordered this list of proposed explanations for sta­sis from Darwinian conventionality towards more challenging proposals, I don't regard any item as excluding any other — indeed, I would be surprised if all cannot claim at least some measure of validity, for once again we deal with an issue of relative frequencies and differential circumstances — and I don't re­gard any pair as establishing a contradiction. In particular, these first two [Page 879] proposals, although different in implications about styles and reasons for lim­ited change in species, remain primarily complementary in their common at­tribution of stasis to reasons based on satisfactory current status — the first on immediate optimality of overt features, the second on inherent plasticity within a current, and presumably adaptive, norm of reaction.)