A phenomenon marking the disruption of normality holds a very different philosophical status than a phenomenon representing the ordinary architecture of biological space and time. Evolutionary change, regarded as an occasional disrupter of stasis, requires a different set of explanatory concepts and mechanisms — a different view of life, really — from evolutionary change, defined as an anagenetic expectation intrinsically operating in most populations most of the time. Punctuated equilibrium proposes that the macroevolutionary key to this new formulation lies in speciation, or the birth of new higher-level individuals at discrete geological moments (corresponding to long intervals at the scale of a human lifetime). Macroevolution, in this view, becomes an inquiry into modes and mechanisms for breaking the stasis of existing species, and generating new species, conceived and defined as discrete higher-level Darwinian individuals — and not a question about how species-individuals gradually change their parts and constitutions through time (as in conventional Darwinism). But even if this particular formulation at geological scales eventually yields more limited impact or utility than proponents of punctuated equilibrium suspect, the more general redefinition of evolution as a set of rare incidents in the breaking of stasis, rather than the pervasive movement of an expected and canonical flow, still poses an interesting challenge for rethinking a fundamental proposition about the nature and history of life.
Punctuation, the origin of new macroevolutionary individuals,
and resulting implications for evolutionary theory
I have argued throughout this chapter that sets of related implications for expanding and reformulating the structure of Darwinian theory, particularly in applications to macroevolution, flow from each of the two major components of punctuated equilibrium — stasis as a norm for the duration of paleospecies, and punctuation (on geological scales) for their cladogenetic origin. The punctuational origin of species by cladogenesis provides our strongest rationale for regarding species as true evolutionary individuals in Darwin's causal world — rather than as arbitrarily delineated segments of transforming continua, and therefore not as genuine entities at all (a position maintained by both Darwin and Lamarck in some of their most forceful passages). If, following what I called the “grand analogy,” species represent “items” or “atoms” of macroevolution in the same sense that organisms operate as fundamental interactors for natural selection in microevolution (see pp. 714–744), then many features of the mechanics and patterning of macroevolution must be reformulated. For macroevolution then becomes a process irreducibly fueled by the differential birth and death of species (just as microevolution, [Page 886] under natural selection, is powered by the differential reproductive success of organisms) — and not, as Darwin and his successors have long held, a phenomenology ultimately built by, and extending causally from, the accumulating consequences of continuous organismic adaptation in transforming populations.
In this sense, punctuated equilibrium — by crowning the case for stable species as atoms of macroevolution — challenges all three legs of the essential Darwinian tripod: the first leg of organismal focus most directly, by establishing the higher-level species-individual as a potent causal agent of evolution as well; the second leg of functionalism more indirectly by affirming, as generators of macroevolutionary patterns, several modes of explanation that do not flow from organismal adaptation, or even rest upon an adaptational base at all; and, most comprehensively, the third leg of extrapolationism by validating a hierarchical view of pattern and causality, and by denying that the mechanisms of macroevolutionary change all flow from our uniformitarian understanding of how natural selection, working in the organismal mode, can alter populations on the scale of human observation in historical time.*
To illustrate the expansive and reformative potential of the species-organism as a causal agent in macroevolution, I will discuss the three major topics that punctuated equilibrium has helped to redefine during the past two decades:
TRENDS. In Chapter 8, I proposed that trends among species in clades may differ substantially from trends among organisms in populations as an “allometric” result of varying weights attached to the three major causal processes at disparate scales of organism-individuals and species-individuals — drives, and the two sources of sorting, drift, and selection (see pp. 714–744 for full development of an argument only summarized here). At the conventional organismic level, drives from below assume little importance because the [Page 887] organism-individual so effectively suppresses the selective proliferation of lower-level individuals within its own body. In most circumstances, the sorting process of drift also contributes little to sustained trends because population sizes (of organisms in demes) usually exceed the small numbers required for maximal efficacy of such a stochastic force. Thus, of the three potential mechanisms, trends at the organismal level usually arise by selection.
But this understandable, and theoretically defendable, domination of selection at the focal level favored by traditional Darwinism does not extrapolate well to the higher causal level of species (as Darwinian individuals) in clades (as populations). When we shift our focus to this upper level, all three processes can claim significant potential weight in theory. (We cannot yet estimate the actual empirical weights due to paucity of research on a topic so recently defined — but see Wagner, 1996, for a breakthrough study based on quantitative and statistical discrimination of all three modes for various trends in the evolution of Paleozoic gastropods — see pp. 733–735 for a summary of his particular conclusions.) Since the species-individual does not preferentially suppress its own transformation by directional alteration of subparts (organisms), macroevolutionary trends may often be propelled by drives from below. Such drives may arise either by the orthodox route of anagenetic transformation in populations via organismic natural selection (“ontogenetic drive” in my terminology of Table 8-1), or by the unorthodox process of directional speciation (“reproductive drive” in my terminology).
When Wright's Rule holds (see pp. 731–735), and species arise at random with respect to the direction of a sustained trend in a clade, then we must invoke sorting processes among species. Sorting by drift can be highly effective at the species level because N tends to be small in relevant populations (species within clades), in contrast with the traditional Darwinian level (organisms within demes), where the magnitude of N usually precludes effective drift for major traits of organismal phenotypes.
A traditional Darwinian perspective might therefore lead us to denigrate the efficacy of the species level as a locus of causation for trends. If species do not marshall sufficient “strength” to stifle their own transformation by drives from below, or sufficient numbers to “prevent” the propagation of a cladal trend by random sorting, then species must pale as evolutionary agents before the strength of organisms (which manifest enough functional integrity to resist any differential proliferation of subparts, and also maintain sufficient population size to forestall random, and potentially nonadaptive, transformation of their collectivities).
I would, however, suggest that such an attitude stymies evolutionary theory as a restricting bias in the category that Francis Bacon called idola theatri, or idols of the theater, in his brilliant early 17th century analysis of mental impediments to understanding the empirical world. Bacon defined idols of the theater as constraining mental habits bred by allegiance to conventionalized systems of thought. In the present case, we fall into the bad habit of reading susceptibility to drive and drift as signs of weakness in an evolutionary individual [Page 888] because the Darwinian agent that we understand best, and that we have previously deemed exclusive — the organism — happens to resist these modes of change as an active consequence of its inherent structure. But nature builds her scales with strong allometry, and not in a fractal manner with every higher level formed as an isometrically enlarged version of each lower level enfolded within (Gould and Lloyd, 19
99).
I suggest that we challenge this idol of our traditional Darwinian conceptualizations, and at least open ourselves to the opposite view that the species-individual's capacity for change by drive and drift, as well as by selection, defines a potential source of strength for this hierarchical level as an exploiter of the full panoply of available causes for trends. Perhaps we should pity the poor organism for its self-imposed restrictions. Or perhaps, rather, we should praise the organism for managing to achieve so much with such a limited range of mechanisms! (Pardon my metaphorical lapses. I am, of course, suggesting that we view the different interplay of potential forces at various hierarchical levels as sources of distinctiveness and strength for each. We will gain a better understanding of evolutionary mechanics when we try to identify the particular capabilities of each level rather than attempting to establish a single “gatekeeper's” criterion for ranking levels in linear order by their quantity of a single enabling power analogous to such fictions as IQ.)
In any case, this allometric expansion of potential mechanisms for trending at the species level offers significant promise for fracturing by redefinition (rather than solving in conventional terms) one of the great conundrums of paleontology — an issue much fretted over, and bruited about, but usually (and finally) cast aside with vague statements of hopeful confidence that traditional explanations will suffice once we finally record enough details in any given case. At least in terms of dedicated pages in our professional literature,* trends represent the cardinal subject of macroevolution (with differential waxing and waning of diversity within and among clades, especially as influenced by episodes of mass extinction, as the second great theme of evolutionary discussion in paleontology).
Paleontology has long been trapped in the dilemma of recognizing only one conventional model for the explanation of trends, and then finding little credible [Page 889] evidence for the model's adequacy. By the expectations of all three central precepts in Darwinian logic, and by our habits of restricting explanations of sustained organismic trends to selectionist causes (given valid arguments for rejecting the alternatives of drift and drive at the organismic level, as discussed above), increasing adaptation of organisms must also propel macro-evolutionary trends under extrapolationist premises.
(All Darwinians understand, of course, that natural selection only yields adaptation to immediate environments — a notion not conducive to sustained directional trends through geological time, given the effectively random fluctuation of most environmental configurations through substantial geological intervals. Consequently, most sustained trends have been interpreted as generalized biomechanical improvements conferring advantages across most or all experienced environments, and arising from Darwin's own preference for domination of biotic over nonbiotic competition in the history of life. See the discussion of Darwin's rationale for this defense of “progress” in evolution, Chapter 6, pp. 467–479.)
Discourse about trends dominates the traditional literature of evolutionary paleontology, both at the most general level of universal phenomenology (Cope's law of increase of size, Dollo's law of irreversibility, Williston's law of reduction and specialization of modular segments, etc.), and as a dominating theme for the history of almost any individual clade. We all know the particular tales for textbook groups — increasing brain size in hominids; larger body size, fewer toes, and higher crowned cheek teeth in horses; increasing symmetry of the cup in Paleozoic crinoids, with eventual expulsion of the anal ray to the top of the calyx; complexification in ammonite suture lines; reduction in number of stipes in graptolite colonies. These summary themes for clades, all based on the concept of general biomechanical improvement through time, distill the essence of traditional paleontology.
I do not deny that generalized organismic advantage may explain some of these classic trends. I do not, for example, challenge the traditional notion that increasing perfection of radial symmetry may confer adaptive benefits in feeding upon sessile organisms like stalked crinoids (see Moore and Landon, 1943, for the classic statement). But I also note that other classic trends, apparently ripe for explanation in biomechanical terms, have stubbornly resisted any reasonable hypothesis ever proposed — most notably the complixifying ammonite suture, which does not clearly confer greater resistance to shell crushing, and does not evidently aid the growing animal by increasing surface area of tissues covering the septa.
But other trends, despite their prominence, have never generated even a plausible hypothesis of biomechanical advantage. Why should fewer-stiped graptolite colonies “do better” in any usual sense of organismic (or, in this case, astogenetic) advantage? In such circumstances, we need to expand our explanatory net by considering alternative causal resolutions based on differential success of species as Darwinian individuals engaged in processes of sorting. Instead of focussing upon the putative biomechanical virtues of fewer stipes, we should be asking how and why such a character might correlate [Page 890] with the propensities of species for branching or for resistance to extinction — the “birth” and “death” processes that regulate sorting at this higher level.
I am not even confident that we should preferentially attribute traits with more plausible organismic advantages — including the enlarging brain of hominids as an obvious example — to conventional microevolutionary explanations, without seriously considering unorthodox possibilities based on causal correlations of such traits with propensities for speciation, or on the sheer good fortune of nonadaptive hitchhiking due to fortuitous presence in the subclade growing to domination for other reasons. We should be paying more attention to interesting and plausible proposals like Sacher's (1966) on lifespan and developmental timing as the primary target of selection in hominid evolution, with large brains on a facilitating causal pathway to advantageous retardations of development (Gould, 1977b).
In any case, and most generally, the need to describe trends — when punctuated equilibrium dominates the geometry of evolutionary change within a clade — as differential success of stable species, rather than as extrapolated anagenesis of populations, requires, in itself and as a “one liner” of extensive reformatory power, a radical reformulation (in the literal sense of reconstruction from the very radices, or roots, of the subject on up) of the primary topic in our macroevolutionary literature. Jackson and Cheetham (1999, p. 76) conclude: “Granted the prevalence of punctuated equilibria, macroevolutionary trends must arise through differential rates of origination and extinction, and not by adaptive evolution within single species. All of this is compatible with traditional neodarwinian evolutionary biology, but was unexpected before the theory of punctuated equilibria.”
In summary, the efficacy of drifts and drives, in addition to selection, for generating trends at the hierarchical level of species as Darwinian individuals, suggests a rich, and virtually unexploited, domain of alternative explanations that might break through the disabling paradox of our current inability to resolve such a salient phenomenon in our preferred mode of adaptive advantages to organisms. Species-level explanations of trends in organismal phenotypes add at least two categories of potential resolution to our usual search for organismic benefit.
First, the trending character may be causally significant not for its phenotypic consequences to the organism, but for its role in influencing rates and directions of speciation in populations of organisms bearing the trait. If fossorial features of burrowing rodents (Gilinsky, 1986), or nonplanktonic lifestyles of marine molluscan larvae (Hansen, 1978, 1980), help to generate populational traits that enhance speciation rates, then a trend spreading such organismic features through a clade may arise by positive sorting of species rather than by general adaptive advantage of the phenotype itself. The organismal phenotypes may enjoy no general advantage at all, and may only produce adaptation to relatively ephemeral habitats within the clade's potential range. In fact, both fossorial rodent species (for reasons of small population size) a
nd nonplanktonic molluscan species (for limited geographic ranges) [Page 891] may experience a reduced average geological longevity relative to surface dwelling (for rodents) and planktonic (for mollusks) clade members. But these impediments may be overbalanced by enhancement of speciation rates, thus driving the trend.
Second, both the low N for species in clades (relative to organisms in most populations), and the remarkably (and, for most people, counterintuitively) high frequency of fortuitous but significant correlations between pairs of traits in systems built by genealogical branching (Raup and Gould, 1974) virtually guarantee that trending by drift will be much more common in sorting of species-individuals than in conventional sorting of organism-individuals. After all, branching evolution imparts a set of autapomorphic traits (through a unique common ancestor) to any subclade of species — and we can scarcely believe that each of these traits establishes the basis of selective existence and success for each species in the entire monophyletic group. Therefore, any process that favors the relative proliferation of any subclade for any reason will automatically engender a positive trend for any included autapomorphic character, whatever the causal basis of the general trend.
When we combine this spur to drift by hitchhiking with the observation that many successful clades go through severe bottlenecks (often as single surviving species) during their geological existence, we obtain even more compelling reasons for considering drift as a major source of macroevolutionary trends, however much we may reject analogous processes as substantial generators of trends in phenotypic characters controlled by organismal selection. For example, ammonites endured two severe bottlenecks at two major mass extinction events — suffering reduction, perhaps to two surviving lineages in the Permo-Triassic debacle, and to a single lineage in the closing Triassic event. In this light, why should we regard explanations based on general biomechanical advantages for organisms as preferable to the obvious blessing of good fortune upon any trait belonging to the phenotype of single lineages that manage to squeak through such profound bottlenecks? Few other evolutionary processes can promote traits from partial representation to exclusivity within a population (of species-individuals within a clade in this case) so quickly and so decisively.
The Structure of Evolutionary Theory Page 141