Wake et al. (1983), for example, document how salamanders, artificially raised to encounter only fixed potential prey, will learn to eat immobile objects, thus contradicting “the widespread assumption that amphibians feed only on moving prey” (p. 216), and also permitting substantial “adaptation” (physiologist's sense again) to feeding regimes without disturbing the stasis of evolved form. In a thought-provoking conclusion, Wake et al. (1983, p. 219) site (and cite) stasis as one component of a more general attitude towards stability of systems and preference for non-change, with evolution conceptualized as a “default option” in the history of life — in contrast with the usual view of active and normative change embodied in the first explanation of stabilizing selection:
Stasis is but the most rigid form of the stability that pervades living systems. Thus organisms have evolved as systems resistant to change, even genetic change. While changing environmental conditions may ultimately necessitate change in the system, until some critical point the system remains stable and compensating. The living system is sometimes envisioned metaphorically as a kind of puppet, with enormous numbers of strings, each controlled genetically, or as a blob of putty that can flow in any direction given sufficient force (selection). Our metaphor is the living system as a balloon, with the environment impinging as countless blunt probes. The system compensates environmental and genetic changes, and persists by evolving minimally.
Developmental constraint. This proposal veers more towards heterodoxy in ascribing stasis to an internally specified inability to change (thereby implying frequent suboptimality of adaptation), rather than to lack of adaptive impetus for change due to current optimality (explanation one) or flexibility within a current constitution (explanation two). (This notion of inability stands forth most clearly in the strict definition — too strict in my view (see Chapter 10) — of constraint as absence of genetic variation for a particular and potentially useful alteration, as in the consensus concept of Maynard Smith et al., 1985.)
In our original paper on punctuated equilibrium, Eldredge and I (1972), basing our arguments partly on Mayr's (1954, 1963) concept of genetic revolutions in speciation of peripherally isolated populations, but more on Lerner's notions (1954) of ontogenetic or developmental, but especially of genetic, “homeostasis,” proposed such constraint as the primary reason for stasis. We wrote (1972, pp. 114-115):
If we view a species as a set of subpopulations, all ready and able to differentiate but held in check only by the rein of gene flow, then the stability [Page 880] of species is a tenuous thing indeed. But if that stability is an inherent property both of individual development and the genetic structure of populations, then its power is immeasurably enhanced, for the basic property of homeostatic systems, or steady states, is that they resist change by self-regulation. That local populations do not differentiate into species, even though no external bar prevents it, stands as strong testimony to the inherent stability of species in time.
This proposal became one of the most widely controverted aspects of punctuated equilibrium, especially in linkage with other, largely independent concepts like the prevalence of neutral change (Kimura, 1968), and the exaptation of originally nonadaptive spandrels (Gould and Lewontin, 1979), also viewed as challenges to the more strictly adaptationist concept of Darwinian evolution then prevalent. I now believe that these criticisms, with respect to the issue of stasis in paleospecies through geological time, were largely justified — and that the theme of constraint, while not irrelevant to the causes of stasis in punctuated equilibrium, does not play the strong role that I initially advocated. (However — and perverse as this may seem to some detractors — my conviction about the general importance of constraint vs. adaptationism at other more appropriate scales has only intensified, particularly in the context of revolutionary findings in developmental genetics — see Chapter 10.)
I have changed my initial view for two primary reasons. First, the arguments of Mayr and Lerner, the intellectual underpinnings of our initial proposals about constraint, have not held up well under further scrutiny, particularly in the privileging of small populations as especially, if not uniquely, endowed with properties that permit the breaking of stasis. Further modelling has led most evolutionists to deny that any major impediment for such change can be ascribed to large populations. Second, I now realize that my arguments for the channeling of potential direction and limitation of change apply primarily to levels above species — to aspects of the developmental Bauplan of anatomical designs that usually transcend species boundaries, rather than to resistance of populations against incorporating enough genetic change to yield reproductive isolation from sister populations.
The ecology of habitat tracking. This explanation for stasis, long favored by my colleague Niles Eldredge (1995, 1999), offers a first alternative (in this list) based on the structuring of species-individuals as ecological entities, rather than on adaptations or capacities of component organisms — thus taking explanation to a higher descriptive level of the evolutionary hierarchy. Otherwise, however, habitat tracking ranks as a conventional Darwinian explanation in calling upon stabilizing selection to confer stasis upon populations that react to environmental change in their geographic locale not by evolutionary alteration to new conditions, but rather by moving with their favored habitat to remain in an unchanged relationship with their environment of adaptation. Eldredge writes (1999, p. 142): “Paradoxically (and contrary to at least superficial Darwinian expectations) . . . stabilizing natural selection will be the norm even as environmental conditions change — [Page 881] so long, that is, as species are free to relocate and 'track' the familiar habitats to which they are already adapted. Rather than remaining in a single place and adapting to changing conditions, species move. And so they tend to remain more or less the same even if the environment keeps on changing.”
I place this otherwise conventional explanation towards the heterodoxical end of my list because habitat tracking embodies the remarkably simple and obvious (in one sense), yet profound and unconventional view (in another sense) that evolutionary change represents a last resort, and not a norm for most times, in the response of populations to their environments. (The second explanation of plasticity also invokes this theme, but from the organism's, rather than the population's, perspective.) Habitat tracking also emphasizes the cohesion, and evolutionary reality, of supraorganismic individuals — an essential theme in the hierarchical reconstruction of Darwinian theory (see Chapter 8). This subtly unconventional notion of change as a last resort or default option puts one's mind in a much more receptive state towards the reality of stasis as a genuine and fundamental phenomenon in evolutionary theory.
The nature of subdivided populations. With this fifth category, we finally enter the realm of truly — that is, causally — macroevolutionary explanations based on the reality of supraorganismal individuals as Darwinian agents in processes of selection. In a brilliant paper that may well become a breakthrough document on this perplexing subject, Lieberman and Dudgeon (1996) have explained stasis as an expected response to the action of natural selection upon species subdivided (as most probably are) into at least transiently semi-autonomous populations, each adapted (or randomly drifted) to a particular relationship with a habitat in a subsection of the entire species's geographic range.
Lieberman and Dudgeon derived their ideas (see also McKinney and Allmon, 1995, for interesting support) in the context of Lieberman's extensive multivariate morphometric analysis of two brachiopod species from the famous Devonian Hamilton fauna of New York State (see pp. 916–922). Lieberman noted profound stasis (with much morphological “jiggling” to and fro but no net change) over 6 million years (Lieberman, Brett, and Eldredge, 1994, 1995); but he also studied samples of each species from each of several paleoenvironments through time. Paradoxically (at least at first glance), Lieberman documented several cases of measurable change in single discrete and continuous pal
eoenvironments through the section — but not for the entire species integrated over all paleoenvironments (an argument against habitat tracking, explanation 4 above, as a primary explanation for stasis). “It was found,” Lieberman and Dudgeon write (1996, p. 231), “that more change occurred through time within a single paleoenvironment than across all paleoenvironments.”
Interestingly, such a conclusion also builds a strong argument against the standard explanation of stabilizing selection (number one of this list) for stasis in paleospecies — because demes tracking single and stable environments through time should show no, or at least less, change than the species as a [Page 882] whole, not more. Lieberman and Dudgeon write (p. 231): “If stabilizing selection played a prominent role in maintaining stasis one would expect to find relatively little morphological change through time within a single environment.” Williams (1992) has made a similar argument, at a lower scale, against stabilizing selection by emphasizing that the copiously, and lovingly, documented efficacy of natural selection in short-term situations of human observation — from beaks of Darwin's finches to industrial melanism in Bistort betularia — makes stabilizing selection doubtful as a general explanation for such a pervasive phenomenon as stasis within paleospecies.
But when we consider this finding in supraorganismal terms, with demes as Darwinian individuals, an evident and sensible interpretation immediately emerges. A temporally coherent population may adapt gradually and continually while tracking one of several paleoenvironments inhabited by a species. But how can these anagenetic changes spread adaptively through an entire species composed of several other subpopulations, each adapted to (and tracking) its own paleoenvironment through time? No single morphology can represent a functional optimum for all habitats. In this common, and probably canonical, situation for species in nature, stability emerges as a form of “compromise” in most circumstances, a norm among “competing” minor changes that are, themselves, probably distributed more or less at random around a standard configuration, with each particular solution generally incapable, in any case, of spreading through all other demes of the species in the face of better locally adaptive configurations in most of these demes.
Of course, one can think of several obvious alternative structures where gradual change might be noted — lack of metapopulational division, with the entire species acting as a single deme, or some accessible and general biomechanical advantage that might be adaptive in all demes. But such circumstances may be uncommon — however important by cumulation in the overall history of life — in any general sample of species within a clade at any given time, thus accounting for the predominant relative frequency of stasis among all species, and for the relative rarity of anagenetic change within species as well.
Lieberman and Dudgeon summarize their proposed explanation by writing (1996, p. 231)
Stasis may emerge from the way in which species are organized into reproductive groups occurring in separate environments.... The morphology of organisms within each of these demes may change through time due to local adaptation or drift, but the net sum of these independent changes will often cancel out, leading to overall net stasis . . . Only if all morphological changes across all environments were in the same direction in morphospace, or if morphological changes in a few environments were very dramatic and in the same direction, would there be significant net change in species morphology over time . . . Thus, as long as a species occurs in several different environments one would predict on average it should be resistant to change. [Page 883]
The theoretical modelling of Allen, Schaffer and Rosko (1993) offers intriguing support in an implication not discussed by the authors. Allen et al. argue that the demic structure required for Lieberman's explanation of stasis strongly buffers species against extinction in chaotic ecological regimes. As an evident corollary, species selection must favor this architecture for species if such chaotic circumstances often prevail (or even just occur sporadically enough to impact a species' fate) over the geographic and temporal ranges of most species in nature. Thus, stasis would attain a predominant relative frequency among paleospecies because higher-level selection so strongly favors the persistence of species composed of multiple, semi-independent demes — the architecture that, as a consequence, engenders stasis by Lieberman's argument. Allen et al. (1993, p. 229) write:
Even when chaos is associated with frequent rarity, its consequences to survival are necessarily deleterious only in the case of species composed of a single population. Of course, the majority of real world species . . . consist of multiple populations weakly coupled by migration, and in this circumstance chaos can actually reduce the probability of extinction . . . Although low densities lead to more frequent extinction at the local level, the decorrelating effect of chaotic oscillations reduces the degree of synchrony among populations and thus the likelihood that all are simultaneously extinguished.
Normalizing clade selection. I cite Williams's (1992, p. 132) term for what most evolutionists would identify as a form of interdemic selection within species. (Williams uses “clade selection” as a general descriptor for all forms of selection among gene pools rather than among genes or gene combinations in organisms.) Williams also notes, as did Lieberman in a different context, the paradox of such strong empirical evidence for predominant stasis in the light of abundant data on substantial change within populations during the geological eyeblink of human careers in observation and experiment.
Williams therefore proposes, using Bell's work on stickleback fishes as a paradigm, that the environments of many demes within most species tend to be highly transient in geological terms, whereas one primary environment (often the original context of adaptation for the species) often tends to be highly persistent. (This phenomenon, however well recorded in sticklebacks, need not extend to a generality for species in nature, as Williams would readily admit in citing sticklebacks as a paradigm, not a claim for nature's normality. Sticklebacks exhibit this pattern because they generate successful, but also transient, freshwater demes from a persisting saltwater stock of lower population density.) Williams (1992, p. 134) therefore argues: “Clade selection acts against freshwater populations either because they cannot compete in mature freshwater faunas or because their habitats and ecological niches are ephemeral. The freshwater forms come and go in rapid succession, but the species complex endures in much the same form for long periods of time . . . [based on] the implied rapid extinction and intense clade selection [Page 884] against all but the conservative marine form . . . The appearance of stasis in the fossil record would result from an enormous variability in the persistence of ecological niches.”
I am more attracted to Lieberman's suggestions, based on averaging among demes with no net change among persistent demes adapted to differing habitats, than to Williams's hypothesis, based on differential survival of one stable deme in a persistent habitat — if only because Lieberman has generated empirical evidence for longterm survival in several habitats within his two brachiopod species, whereas Williams's stickleback example may represent an unusual situation in the drastically different habitats (fresh vs. marine waters) of his transient vs. persistent demes. Still, I applaud these two suggestions for stasis based on the structuring of species-individuals as collections of deme-individuals, with differential selection acting upon demes in an irreducibly macroevolutionary mode. These proposals therefore occupy the heterodox end in a spectrum of proposed explanations for stasis — for they challenge the Darwinian orthodoxy of primacy or exclusivity for organismal selection. I especially appreciate Williams's openness towards explanations in this form, given his previous and highly influential preferences for formulating all evolutionary explanation, except when absolutely unavoidable, at the level of genic selection (in his famous book, Williams, 1966, as discussed on pp. 550–554).
In summary, then, the assertion of predominant stasis in the geological history of most paleospecies — one of the
two primary claims of punctuated equilibrium — has provoked an interesting debate in evolutionary theory, with implications for some of the most basic concepts and perspectives in our science. First, and if only as a comment about the contemporary sociology of science, the recognition of stasis as a norm of controlling relative frequency at the level of punctuated equilibrium (at least for conventional sexual species of Metazoa), has spurred general interest in phenomena of stability and non-change throughout other levels of evolutionary inquiry (see, for example, Maynard Smith, 1983). We do not yet know (see fuller discussion on pp. 928–931) whether or rather how much, stasis across all scales might be attributed to structural similarity in nature's materials and processes — thus rendering this common pattern as an interesting parallelism (to use our evolutionary jargon) with genuinely homologous causal elements across scales, rather than a fortuitous convergence of similar overt patterns for disparate and merely analogous reasons. But at least we stand at the threshold of such an inquiry.
Second, and even more generally, the validation of predominant stasis as a norm would impel us to recast the basic problematic of evolution itself. If, following our conventional assumptions from Darwin to now, change represents the norm for a population through time, then our task, as evolutionary biologists, lies in specifying how this expected and universal phenomenon operates. But if, as punctuated equilibrium suggests, stasis represents the norm for most populations at most times; and if, moreover, stasis emerges as an active norm, not merely a passive consequence (as the modelling of Jackson and [Page 885] Cheetham, 1995, strongly suggests in documenting stasis at too high a relative frequency for models based on neutralism, directional selection, or any set of assumptions that do not include some active force promoting stasis directly) — then evolutionary change itself must be reconceptualized as the infrequent breaking of a conventional and expected state, rather than as an inherent and continually operating property of biological materials, ecologies and populations.
The Structure of Evolutionary Theory Page 140