Punctuated Equilibrium and Phyletic Gradualism Peter R Sheldon, The Open University, Milton Keynes, UK

Secondary article Article Contents . Punctuated Equilibrium and Phyletic Gradualism: Competing Models . The Perception of Microevolutionary Patterns

Punctuated equilibrium and phyletic gradualism are contrasting patterns of evolution among a spectrum of patterns found in the fossil record. In punctuated equilibrium, species tend to show morphological stasis between abrupt speciation events, whereas in phyletic gradualism species undergo more continuous change.

Punctuated Equilibrium and Phyletic Gradualism: Competing Models In 1972, Niles Eldredge and Stephen Jay Gould stimulated controversy by publishing their model of punctuated equilibrium, according to which most evolutionary change in the history of life is concentrated in brief speciation events. The belief that fossil species remained in morphological stasis throughout their existence had prompted them to propose this pattern as an alternative to what they saw as the pervasive paradigm of gradual evolution within lineages, which they called phyletic gradualism. In phyletic gradualism, evolution occurs slowly and steadily, without the particular association of morphological change with speciation; as much or more evolutionary change takes place between times of branching as at the branching points themselves (Figure 1a). By contrast, in punctuated equilibrium a new species originates rapidly (a punctuation) and then remains virtually unchanged (in equilibrium or stasis) for the rest of its existence (Figure 1b). The new species might itself bud off another one, but the parent species remains in stasis until its extinction. Thus, in punctuated equilibrium, significant evolutionary change occurs at events of branching speciation (cladogenesis) and not during the in toto transformation of lineages (anagenesis). Eldredge and Gould maintained that the pattern of punctuated equilibrium was more or less universal, whereas phyletic gradualism was rare or absent. Eldredge and Gould based their model on observations of fossils, especially Devonian trilobites and Pleistocene land snails, and on modern speciation theory. They invoked Mayr’s concept of peripatric speciation, a form of allopatric speciation involving the geographic isolation of small, localized populations peripheral to the main species range. Such founder populations, often with a somewhat random or unusual sampling of the parental gene pool, would undergo rapid selection to local conditions, and soon become reproductively isolated from the parent population. Eldredge and Gould argued that various ‘constraints’, especially those on genetic variability and developmental pathways, maintained stasis in the parent population. These constraints, however, became

. Examples from the Fossil Record . Recent Developments . Conclusion

broken down in small, isolated populations, allowing rapid change by natural selection and genetic drift, possibly facilitiated by macromutations. The possible genetic mechanisms involved were then, and remain, controversial. Eldredge and Gould argued that, if peripatric speciation is the norm, the rise of new species is expected to be episodic, local and rapid – as opposed to continuous, widespread and slow. The chances of finding in the fossil record intermediate forms recording a speciation event is therefore bound to be low. Previously, such a lack of intermediates had largely been accounted for by incompleteness of the record, or by grossly inadequate sampling. But if peripatric speciation prevails, the fossil record can be taken more at face value; it would explain as real phenomena both the sudden appearance of new fossil species and the rarity of intermediate forms ‘caught in the act’ of speciation. Eldredge and Gould stated that the species-forming event would normally take less than 1% of that species’ later existence in stasis. This means, for example, that for a new species subsequently lasting a million years, reproductive isolation and associated morphological evolution might take the order of 5000–10 000 years – ample time for these two processes to occur from the perspective of a neontologist. This is an awkward timespan: too long for direct observation, yet normally too short for details to be resolved in the fossil record. Net, long-term stasis, by contrast, can readily be verified simply by finding very similar morphologies separated by extensive time intervals; its demonstration does not require a high-resolution record. In punctuated equilibrium, fossiliferous strata should typically show a species in stasis, with sharp morphological change recording the migration of the descendant species from the peripheral, isolated area in which it developed. Cases of introduced species in modern times (e.g. the starling in North America) demonstrate the rapidity with which a species may spread across large areas. In the strict version of punctuated equilibrium, the presumed ancestor is then expected to persist for a while alongside its descendant species. In the toned-down version – not

ENCYCLOPEDIA OF LIFE SCIENCES © 2001, John Wiley & Sons, Ltd. www.els.net

1

Time

Punctuated Equilibrium and Phyletic Gradualism

Morphology (a)

Morphology (b)

Figure 1 The contrasting idealized evolutionary patterns of phyletic gradualism (a) and punctuated equilibrium (b). The top of each evolutionary tree represents the most recent time depicted, such as the present day. The horizontal distance between lineages (lines of descent) indicates in a general way the degree of morphological difference. In punctuated equilibrium (b), all morphological change is concentrated in branching events, with stasis thereafter. Note that, for comparative purposes, the two patterns have been drawn with a similar number of branching points. In addition, if the patterns are superimposed on each other, corresponding lineages have the same morphology at the latest occurrence (i.e. at their extinction or at the top of the diagram), but the morphological routes by which lineages reach those points are very different in each pattern. Fine-scale zig-zags in morphology will occur in both phyletic gradualism and punctuated equilibrium, but are not shown in these simplified depictions.

strictly punctuated equilibrium – punctuated change may occur within a lineage, but without lineage branching (speciation); this is sometimes referred to as punctuated anagenesis or punctuated gradualism (see below). Eldredge, Gould, Stanley and others claimed that stasis was a major, hitherto largely unrecognized or ignored, feature of the fossil record, and should be studied in its own right. ‘Stasis is data’, became the punctuationists’ motto. As Gould and Eldredge (1993) wrote, ‘palaeontologists never wrote papers on the absence of change in lineages before punctuated equilibrium granted the subject some theoretical space’. Punctuated equilibrium differs radically from traditional neo-Darwinism, not in claiming that stasis is common, but mainly by proposing that (1) species generally evolve new morphologies only when a small population becomes a new, reproductively isolated species, and (2) long-term trends in morphology among taxa above the species level are often the consequence not of natural selection favouring genotypes within species, but of a higher-level selection process among species (see below). The concept of punctuated equilibrium has certainly exerted a strong influence on the investigation and interpretation of evolutionary patterns (see review by Gould and Eldredge, 1993). Many supposed textbook cases of gradualism dissolved in the wake of punctuated equilibrium and in the climate of more rigorous analysis that it produced. Some have argued, however, that Eldredge and Gould’s version of phyletic gradualism was to some extent a straw man, because they identified one of its tenets as being that ‘transformation is even and slow’. Punctuationists have also been criticized as portraying the modern synthesis as more reductionist, i.e. denying 2

hierarchical processes, than it in fact was. Although it is true that the prevalence of morphological stasis in fossil species was not widely predicted by the modern synthesis, many earlier authors, such as Simpson, believed that evolution embraced a wide spectrum of rates. Darwin’s position in The Origin of Species was essentially gradualistic, especially in explaining the evolution of complex adaptations by the step-by-step accumulation of relatively small changes. Interestingly, however, in response to criticism by Hugh Falconer, a palaeontologist, Darwin inserted a passage concerning the stability of species in later editions, and in chapter 15 there is this statement: ‘Many species when once formed never undergo any further change but become extinct without leaving modified descendants; and the periods during which species have undergone modification, though long as measured by years, have probably been short in comparison with the periods during which they retain the same form.’ Darwin was thus not a caricature phyletic gradualist in that he did not insist there was constant, steady evolution within lineages between speciation events.

The Perception of Microevolutionary Patterns The ideal requirements for establishing microevolutionary patterns include large numbers of complete fossil specimens from successive, small time intervals; a precise and accurate framework of absolute ages; samples that span the entire geographical and temporal range of all closelyrelated species; a correct understanding of phylogenetic

Punctuated Equilibrium and Phyletic Gradualism

relationships; and statistical data on many characters from different ontogenetic stages of each species. Some of these requirements are rarely if ever met, and many relevant hypotheses, such as ancestor–descendant relationships, are never truly verifiable; as usual, we must assess the relative probabilities of competing hypotheses. A knowledge of geographical variation is highly desirable because spurious patterns of change may arise in local areas owing, for example, to the immigration or emigration of intraspecific variants tracking their favoured environment. The critical test of punctuated equilibrium (sensu stricto) is whether morphological change is usually accompanied by lineage splitting, i.e. cladogenesis or true speciation – as opposed to chronospeciation, in which sufficient anagenetic evolution occurs within a single lineage for forms to be considered different species and given a different name. One obvious problem is that speciation cannot be recognized in the fossil record by the aquisition of reproductive isolation, and is instead necessarily equated with the initiation of considerable morphological difference in preserved features. The most that palaeobiologists can do is try to make species they describe live up to a definition such as ‘Species are morphologically distinct groups within which variation is of the magnitude expected in interbreeding populations, and between which the differences are of the kind and degree expected to result from reproductive isolation in natural populations.’ There are actually very few indisputable cases of lineage splitting observed in the fossil record, and sister species have to be inferred but cannot be proven. Demonstrating stasis in a species is not the same as demonstrating punctuated speciation. In theory, for example, a peripheral isolate could evolve gradually (and undetected) to a new species which, on migration, appeared abruptly alongside its parent species, thereafter remaining in stasis. Darwin recognized that gaps in the geological record, both vertically (in time) and laterally (in space), can generate an impression of abrupt evolutionary change (though not stasis). In addition, various biases can arise during the collection and description of fossils that generate an impression of punctuation and stasis, irrespective of the real pattern (Sheldon, 1996). These biases include the requirement to apply binominal taxonomy (i.e. discrete Linnean names) to fossils as well as living organisms; the amalgamation of fossils collected from strata of different ages, thereby submerging any evidence of change within one large sample; a natural reluctance to use terms implying uncertainty when identifying species; an absence of formal nomenclature to signify small morphological differences between samples of different age; the plotting of species time ranges in charts as vertical lines; and cladistic methods of phylogenetic analysis. The almost universal textbook depiction of gradualistic evolution as straight-line change (as in Figure 1a) reflects another bias that has hindered understanding of microevolutionary patterns. This is the predisposition not to

expect evolutionary reversals or to dismiss them as ecophenotypic effects, the migration of ecotypes, genetic drift or sampling problems – all of which are possible influences on individual patterns. The simplest, most tempting way of connecting successive data points – a straight line – is, in fact, highly unlikely in nature. Straightline evolution, that is prolonged unidirectional evolution at a constant rate, is, in fact, a very special case, like drawing one random walk and expecting evolution to follow that particular path (Sheldon, 1996). Reversals can be expected to occur in any long record of a single character, as long as that character maintains some variation. Cases of gradualism will indeed seem rare if the theoretical pattern is set up as unidirectional change. High-resolution fossil records show that patterns generally become increasingly dynamic (fluctuating, with reversals) as finer timespans are resolved; how patterns are perceived can thus depend on sampling strategy and the extent to which samples from different timespans have been amalgamated. The pervasiveness of reversals also explains the general observation of an inverse relationship between measured rates of evolution and the timespan under consideration.

Examples from the Fossil Record In a survey of 58 detailed studies made since 1972, covering a wide range of taxa and different geological periods, Erwin and Anstey (1995) concluded there was no overwhelming signal, but a wide variety of patterns. These certainly include punctuated equilibrium and phyletic gradualism, but also many cases of punctuated anagenesis (stasis and abrupt change in lineages without branching), with some cases of gradual multiple branching, punctuated multiple branching, and so on. Patterns can be complex within individual lineages, let alone groups. For example, Geary (in Erwin and Anstey, 1995) describes a Cenozoic gastropod lineage that underwent stasis for 7 million years, followed by prolonged directional change to a new species over about 2 million years. One of the most convincing examples of punctuated equilibrium, with stasis and with abrupt speciation, is that of the bryozoan genus Metrarabdotos described by Cheetham in 1986 and subsequent papers (see Cheetham and Jackson in Erwin and Anstey, 1995). Cheetham measured up to 46 characters in colonies of this sessile marine organism from Cenozoic strata in the Dominican Republic, and applied a method of discriminant analysis to determine patterns of morphological change in the numerous samples. Most of the 19 species concerned appeared suddenly, changed very little (if at all) through some millions of years, and tended to persist alongside their inferred parent species. This case was lent extra weight by studies of Cheetham and Jackson (see their paper in Erwin and Anstey, 1995) on living species of a closely related 3

Punctuated Equilibrium and Phyletic Gradualism

genus, Stylopoma. They investigated the extent to which species of Stylopoma that had been differentiated according to the morphology of their hard parts alone corresponded with genetic differences determined by protein electrophoresis. There was an excellent correspondence between morphology and genetics, so that in Stylopoma – and by inference Metrarabdotos species (which are either extinct, or rare and hard to study) – equating morphospecies with true species seems justified. In a remarkable demonstration of stasis, several species of living bivalves studied in great morphological detail by Stanley and Yang (1987) show no more evolutionary change over the past 17 million years than there is geographic variation within each species at the present day. Brett and Baird (in Erwin and Anstey, 1995), studying shallow marine Devonian faunas, found a pattern of species occurring in discrete packages that appeared and disappeared more or less in synchrony, a phenomenon they coined ‘coordinated stasis’. Once originated, new biotas persisted more or less intact in both morphologies of constituent species and ecological associations for several million years, despite environmental fluctuations. Only the occasional, more extreme perturbations produced change when thresholds were exceeded. Other clear examples of at least approximate stasis have been documented in invertebrate groups such as ostracods, brachiopods, gastropods, echinoids and corals (see references in Erwin and Anstey, 1995). The average duration of marine invertebrate species is about 5 million years. In general, notwithstanding the biases mentioned above, the fact that Linnean taxonomy can be applied more or less successfully to such fossils, as well as to living forms, is itself good evidence for the prevalence of approximate stasis. A study by Sheldon of about 15 000 Ordovician trilobites from central Wales revealed a pattern of broadly parallel, gradualistic evolution in eight lineages from a moderately deep, relatively stable marine setting (see references in Sheldon, 1996). Over a period of about 2 million years, a variety of changes took place at different times in different lineages, with patterns far closer to phyletic gradualism than to punctuated equilibrium. There was no evidence of lineage branching. The ‘missing links’ between previously described, successive trilobite species were no longer missing: intermediate horizons yielded specimens of intermediate morphology. The end members of most lineages had been assigned to different species and, in one case, to different genera. The apparent success of earlier Linnean nomenclature (with its implications of discrete species) could easily have been misinterpreted as evidence of punctuation and stasis. Many of the other biases mentioned above could be seen in retrospect to have obscured evidence of gradual evolution, and temporary reversals made it impractical to subdivide each lineage into discrete successive chronospecies. A wide spectrum of patterns has been observed in vertebrates, as documented in a comprehensive survey by 4

Carroll (1997). The average duration of mammal species in the fossil record is about 0.5–1 million years, i.e. shorter than for marine invertebrates. There seems to be a tendency for mammals to show more continuous evolution than many invertebrate groups, but there are numerous exceptions. Prothero and Heaton (1996) document the prevalence of stasis in some North American mammals during a period of intense climatic change in the mid Cenozoic. Unlike many Quaternary organisms, the mammals apparently did not even migrate in response to climatic changes. Badgley and Behrensmeyer (1995, for reference see Sheldon, 1996) found that, at another time in the Cenozoic, the mammals of Pakistan underwent their greatest faunal turnovers, with considerable gradual turnover too, during a long interval in which no climatic change is discernible. They argue that these results, and others showing persistent evolution of early Cenozoic mammals during stable or gradually changing climates, support the Red Queen model of evolution driven by continuous biotic interaction, but they emphasize that major climatic change also initiated substantial faunal change. There are few high-resolution data on land plants, but stasis appears to be common in the deltaic floras of the Carboniferous tropics. Quaternary beetles are well known for showing remarkable stasis throughout glacial–interglacial cycles. Marine microorganisms such as planktonic foraminiferans and radiolarians show a tendency for slow, gradualistic evolution, but again a wide spectrum of patterns has been discovered. Although some correlations are emerging between individual groups of organisms and particular evolutionary patterns, most patterns will have to remain suggestive at best until many more high-resolution studies are undertaken.

Recent Developments A major criticism of punctuated equilibrium is ample evidence that populations can evolve in response to selection without speciating. Many fossil sequences reveal evidence of stasis within species, but not direct evidence for the association between morphological change and speciation. Gould and Eldredge (1993) agree that speciation does not cause morphological change, but they maintain that the two are strongly associated with each other in the fossil record. To explain this, they accept the suggestion of Futuyma (see also Futuyma, 1998) that speciation facilitates long-term morphological divergence not by accelerating morphological change but by preventing such changes being broken down by recombination when divergent populations eventually come into contact with ancestral phenotypes. Reproductive isolation allows changes to be ‘locked up’; speciation thus provides

Punctuated Equilibrium and Phyletic Gradualism

morphological changes with enough permanence to be registered in the fossil record. There are several possible mechanisms by which stasis could be maintained, including stabilizing selection, the close tracking of favoured habitats by migration, and genetic constraints, but their relative contributions are not yet understood. According to Eldredge (1995), stasis is an outcome of both habitat tracking and, more importantly, the organization of species in the wild as envisaged by Sewall Wright in his ‘shifting balance theory’. Where species are split up into so many local populations in slightly different habitats, evolution cannot proceed in a lock-step, gradualistic transformation of the entire species because local populations are evolving in different directions. Eldredge places great importance on factors that affect the probability of survival of fledgling species, and argues that the rate of successful speciation is higher in the tropics because, being easier to specialize, it takes less adaptive change at speciation for a newly evolved species to establish ecological distance from its ancestor. Given the variety of patterns reported above, the challenge is to find a better synthesis incorporating elements of both punctuated equilibrium and phyletic gradualism, and encompassing evidence from the spectrum of patterns obtained so far. There is now a wealth of evidence that contradicts the widely held, intuitive expectation that evolution is the normal response of species to environmental change, and that stable environments lead to stable morphology. According to the Plus c¸a change model (Sheldon, 1996), the more that the physical environment changes, the more that species stay the same. Morphological stasis over geological time scales thus tends to arise not from the stability of physical environments, but from their instability. In this model (Figure 2), stasis is the usual response to widely fluctuating physical environments over geological time scales, until thresholds are reached (when punctuational change or extinction occurs). The lineages that survive in an environment of widely fluctuating sea levels, climate, substrate, salinity, etc., become relatively inert to each environmental twist and turn, in contrast to more evolutionarily sensitive lineages in less changing environments. Generalists in this long-term sense – i.e. most species found in the fossil record – are species with properties that enable them to survive throughout wide environmental fluctuations for millions of years; they are distinct from ecological generalists (eurytopes). The relationship between evolution and environmental change is thus far from simple; the impact of an event is contingent on the long-term history of the system. If a species has been subjected many times to large physical disturbances, yet another such disturbance might yield no response at all (or might send it over a threshold). Conversely, smaller, less easily detectable, events following a long stable interval may often promote evolution. So, in this perspective, small

Figure 2 The Plus c¸a change model, which combines aspects of punctuated equilibrium and phyletic gradualism. It proposes that, over geological time scales (e.g. a million years), gradualism (right) is characteristic of narrowly fluctuating, relatively stable environments such as a tropical rainforest and the deep sea. By contrast, punctuated equilibrium (left) is expected to prevail in the physically more unstable environments that dominate the fossil record, especially shallow seas. In a widely fluctuating environment (left), a threshold might occur not only when some environmental variable exceeds wide reflecting boundaries, but also when it contracts to become narrowly fluctuating (dotted line). ‘Environment’ indicates some long-term physical aspect of environment, e.g. mean temperature, sea level or type of sediment forming the sea floor. ‘Morphology’ indicates some aspect, or aspects, of morphology. Highfrequency environmental oscillations such as annual cycles are not shown. From Sheldon (1996), modified after Sheldon (1990).

abiotic events and biotic interactions tend to drive evolution, but so also do some major perturbations. The Plus c¸a change model includes the following predictions: (i) continuous, gradualistic evolution tends to occur on land in the tropics (e.g. tropical rainforest), in the deeper sea, in settings with many biotic interactions, and at times of global climatic stability; (ii) stasis, and occasional punctuation, tends to occur in shallow marine settings, in temperate latitudes, in settings subject to wide abiotic fluctuation, and at times of global climatic instability. Any such model will have many exceptions (among which appears to be the evolution of Homo sapiens). If true as a tendency, the model has some important implications (Sheldon, 1996). Punctuated equilibrium may be being mistakenly perceived as the overwhelming pattern in the history of life because the environments in which gradualism is expected to prevail (e.g. a tropical rainforest and the deeper sea) are rarely represented in the fossil record. The vast majority of fossils come from dynamic shallow marine environments, so it is not surprising that many fossil lineages show approximate stasis and occasional punctuations. Another implication is that many of today’s species, especially in temperate latitudes, may be relatively inert to environmental twists and turns, having lived through 2 million years of Quaternary climate 5

Punctuated Equilibrium and Phyletic Gradualism

upheavals, and that, except for human-influenced environments, relatively little microevolution may be occurring globally compared with more stable times. The last 10 000 years of relative climatic stability in the Recent may be an order of magnitude too short an interval to relax the influences promoting net stasis. The claim of punctuated equilibrium that there are higher-level selection processes among species has proved highly controversial. Is microevolution really decoupled from macroevolution? Do species behave as units, having properties that influence large-scale patterns of evolution independently of natural selection acting on individuals in populations? Can a group of species show variation that affects the probability with which they speciate, and that can be inherited by daughter species? The answer to the latter is yes, in principle. Despite much early confusion in the literature on hierarchical thinking, it seems increasingly likely that species-wide properties such as population size, geographical range and dispersal mode can and do influence the probabilities of speciation and extinction. Stanley (1998) gives some cases in which species have apparently been ‘selected’, or at least sorted over geological time intervals. In general, however, it seems that trends may have multiple causes. For a useful discussion of these issues and the causes of evolutionary trends, see Futuyma (1998).

Conclusion Punctuated equilibrium and phyletic gradualism have both been documented in the fossil record, as have a variety of other patterns spanning many different groups of organisms. Morphological changes, whether abrupt or gradual, need not be associated with branching speciation, but reproductive isolation may confer permanence on such changes. Approximate morphological stasis is commonly seen in many fossil species, especially the shallow marine invertebrates that dominate the fossil record, and this may be related to the instability of shallow marine environ-

6

ments over geological time scales. Conversely, gradualism may prevail in more stable, though rarely (if ever) preserved, environments such as a tropical rainforest. The claim of punctuationists that higher-level selection processes can act among species remains controversial, but species-wide properties can probably influence the probabilities of speciation and extinction.

References Carroll RL (1997) Patterns and Processes of Vertebrate Evolution. Cambridge, UK: Cambridge University Press. Eldredge N (1995) Reinventing Darwin. New York: Wiley. Eldredge N and Gould SJ (1972) Punctuated equilibria: an alternative to phyletic gradualism. In: Schopf TJM (ed.) Models in Paleobiology, pp. 82–115. San Francisco: Freeman and Cooper. Erwin DH and Anstey RL (eds) (1995) New Approaches to Speciation in the Fossil Record. New York: Columbia University Press. Futuyma DJ (1998) Evolutionary Biology, 3rd edn. Sunderland, MA: Sinauer. Gould SJ and Eldredge N (1993) Punctuated equilibrium comes of age. Nature 366: 223–227. Prothero DR and Heaton TH (1996) Faunal stability during the Early Oligocene climatic crash. Palaeogeography, Palaeoclimatology, Palaeoecology 127: 257–283. Sheldon PR (1990) Shaking up evolutionary patterns. Nature 345: 772. Sheldon PR (1996) Plus c¸a change – a model for stasis and evolution in different environments. Palaeogeography, Palaeoclimatology, Palaeoecology 127: 209–227. Stanley SM (1998; originally published 1979; with new introduction) Macroevolution – Pattern and Process. Baltimore, MA: Johns Hopkins University Press. Stanley SM and Yang X (1987) Approximate evolutionary stasis for bivalve morphology over millions of years: a multivariate, multilineage study. Paleobiology 13: 113–139.

Further Reading Clarkson ENK (1998) Invertebrate Palaeontology and Evolution, 4th edn. Oxford: Blackwell Science. Kemp TS (1999) Fossils and Evolution. Oxford: Oxford University Press. Ridley M (1996) Evolution, 2nd edn. Oxford: Blackwell Science. Skelton PW (ed.) (1993) Evolution – A Biological and Palaeontological Approach. Wokingham: Addison Wesley.