0.9, with no evidence of morphologically intermediate forms, and all ancestral species but one survived unchanged along with their descendants” (Jackson and Cheetham, 1994, p. 420).
By dense sampling in both vertical sequence and geographic spread, Nehm and Geary (1994) demonstrated the punctuational origin of the gastropod Prunum christineladdae from its ancestor P. coniforme in a small part of its [Page 844] Caribbean range, during a short interval (0.6 to 2.5 percent) of ancestral persistence in stasis. Following the descendant's origin, ancestors continued to thrive in central areas of the range.
In a common pattern found in many taxa, punctuated equilibrium can be confirmed, even in local sections, and even when ancestors do not occur in the same strata as their descendants. Frequently, a population from an ancestral species of known and widespread geographic range branches punctuationally to a descendant that maintains exclusive occupancy of the range for a time, but then becomes extinct. The ancestor subsequently reinvades the range, thus establishing earlier coexistence during the descendant's geological tenure. For example, Bergstrom and Levi-Setti (1978) documented the threefold reappearance of the Middle Cambrian trilobite Paradoxides davidis
9-18. The best-documented, indeed already canonical, example of punctuated equilibrium as an invariant pattern for an entire clade across its full geographic range — the research of Cheetham on Tertiary and Quaternary Caribbean species of Metrarabdotos. Each point depicts a multivariate centroid based on all characters, not just a single feature. All species express stasis, several for extended periods and a large number of samples. Ancestors persist after the origin of descendants in 7 of 9 cases where Cheetham felt confident enough to assert a claim for direct filiation.
[Page 845]
davidis following local and allopatric origins of derived taxa that then become extinct at diastems, with the ancestor reappearing in strata just above.
Similarly, Ager (1983) traced the allopatric origin of late Pliensbachian brachiopod species from the central stock of Homoeorhynchia acuta, and the later Toarcian migration of the descendant H. meridionalis into the ancestral region. Williamson's (1981) celebrated and controverted study (see pp. 769–771) of punctuational origin for several pulmonate snail species in African Pleistocene lakes invokes the same kind of evidence — as the ancestral species migrates back (in several separate episodes, moreover) after a coalescence of lakes and the extinction of descendant species that had originated in previous times of isolation.
When all evidence derives from a restricted region, the separation of punctuation in situ from migrational incursion (with origin elsewhere at an unspecified tempo) becomes more difficult, but some criteria of admittedly uncertain inference may still be useful. For example, Smith and Paul (1985) argue that the sudden appearance of the descendant echinoid Discoides favrina in strata still holding ancestral D. subucula may represent an event of punctuational speciation on morphological grounds — for the descendant species, though visually distinct in many features, can be easily derived, given allometric patterns shared by both forms, through a simple heterochronic process of hypermorphosis.
In graptolites, the pattern of ancestral survival after cladogenetic origin of a descendant taxon has been noted frequently enough to inspire its own terminology as the concept of “dithyrial populations” (Finney, 1986), or samples from the same stratum containing two directly filiated and noninter-grading species.
The widespread geographic distribution of many late Tertiary and Quaternary mammalian lineages provides several examples of geographically resolvable allopatric origin followed by later survival with the ancestral species. For example, Mammuthus trogontherii, the presumed ancestor of the woolly mammoth M. primigenius, first appears in northeastern Siberia while its presumed ancestor, M. meridionalis, continued to survive in Europe (Lister, 1993a, p. 209).
Other forms of evidence can lead to strong inferences from data of ancestral survival to origin of descendants by punctuated equilibrium, even in the absence of such firm geographic data. I previously mentioned the growing evidence for rapid cladogenesis as the primary pattern in hominid evolution (see p. 833), based on several criteria, including the high relative frequency of observed overlap, the limited time available for cladogenetic origin (even when place and geological moment have not been clearly specified), and our confidence that all events (at least preceding the origin of Homo erectus) occurred in Africa. In his review, McHenry (1994, p. 6785) stated “ancestral species overlap in time with descendants in most cases in hominid evolution, which is not what would be expected from gradual transformations by anagenesis.” McHenry's summary diagram (reproduced here as Fig. 9-19) shows a clear pattern of dominant relative frequency for rapid cladogenesis — [Page 846] a weight that has only increased in the light of discoveries since then (see Leakey et al., 2001), particularly for a vigorous phase of cladogenesis 2-3 million years ago, leading to at least half a dozen hominid species (see Johan-son and Edgar, 1996).
On the same subject of punctuational and cladogenetic reformulations for classic evolutionary trends previously framed (and widely celebrated in both textbook and story) as exemplars of anagenetic gradualism, the phylogeny of horses has been rewritten as a copious cladogenetic bush replete with ancestral survival in the very parts of the sequence once most firmly read as a tale of linear progress. For example, Prothero and Shubin (1989) have shown that the Oligocene transition from Mesohippus to Miohippus conforms to punctuated equilibrium, with stasis in all species of both lines, transition by rapid branching rather than phyletic transformation, and stratigraphic overlap of both genera (one set of beds in Wyoming has yielded three species of Mesohippus
9-19. From McHenry (1994). The hominid record is spotty, but the basic pattern of substantial stasis within several species — particularly A. afarensis — and numerous branching points with persistence of putative ancestors lends support to the model of punctuated equilibrium.
[Page 847]
and two of Miohippus, all contemporaries). Prothero and Shubin conclude: “This is contrary to the widely-held myth about horse species as gradualistically-varying parts of a continuum, with no real distinctions between species. Throughout the history of horses, the species are well-marked and static over millions of years. At high resolution, the gradualistic picture of horse evolution becomes a complex bush of overlapping, closely related species.”
To end this section with a particularly instructive example, punctuated equilibrium has frequently been saddled with the charge that inherent limitations of paleontological data yield biased results, artificially and superficially favorable to the theory — with Darwin's classic argument against a literal reading of punctuations as the conventional antidote. However, an opposite bias may also be significant, and may lead to serious underestimation of punctuated equilibrium in a circumstance likely to be quite common: when a descendant, fully distinct at its origin but initially rare, enters the ancestral area, and then increases steadily in relative abundance as the ancestor declines to extinction. The true evolutionary pattern will be fully punctuational, with stasis in both ancestor and descendant throughout, and with abrupt geological origin of the descendant. But if we misread the event as a tale of anagenetic transformation, and if the two species overlap extensively in ranges of variation, then we will misinterpret the full pattern as transformation by anagenesis, rather than replacement with steadily increasing relative abundance of the descendant species.
The important distinction between these interpretations can be made with appropriate statistical tools applied to samples of sufficient size — but the punctuational alternative must be conceptually available to suggest such a test. In this subtle sense, among so many other more overt reasons explored in this book, expectations of gradualism seriously restrict our range of potential explanations for evolutionary modes and tempos — and punctuated equilibrium therefore becomes both suggestive and expansive, whether or not the hypothesis holds in
any particular case.
In an elegant demonstration of this principle, Heaton (1993 and 1996 for data in extenso) showed that a classic case of supposedly gradualistic anagenesis in Oligocene rodents from the western United States really represents a case of replacement. Heaton writes (1993, p. 297): “Statistical investigation of large samples suggests instead that two closely related species coexisted, and the shift in mean size that was thought to represent anagenesis actually represents replacement.”
Heaton demonstrated the distinct character of the two taxa both by bimodality in their joint occurrences (Fig. 9-20), and by showing that the two species maintained distinctly different geographic ranges (with overlap in Nebraska and eastern Wyoming, but only the descendant taxon living at the same time in the Dakotas — see Figs. 9-21 and 9-22). The small species, Ischyromys parvidens, predominates in the early Orellan, although the larger I. typus already occurs low frequency in the same strata. I. typus then increases,
[Page 848]
9-20. From Heaton (1993). Data that had, in the past, been interpreted as a gradualistic evolution of increasing size within a single species actually represent a change in relative abundance of two species, each stable throughout its interval — with the species of larger body size gradually becoming more common in the local section.
as ancestral I. parvidens declines, throughout the remainder of Orellan times.* Interestingly, I. typus does undergo a small anagenetic increase following the extinction of I. parvidens, “but this change is minor and not deserving of chronospecies recognition” (Heaton, 1993, p. 297), and the species, in any case, becomes extinct soon thereafter — a common pattern,
[Page 849]
9-21. Note that both species of Ischyromys live sympatrically and remain in stasis in some parts of their range, particularly in Nebraska and Wyoming.
[Page 850]
also strongly implicating punctuated equilibrium as the major generator of larger trends, if only because the fruits of anagenesis get plucked so quickly unless they can be “locked up” in cladogenetic iterations.
The “dissection” of punctuations to infer both existence
and modality
Once a literal punctuation has been noted, and a cladogenetic origin inferred by such criteria as the documentation of ancestral survival, further testing of punctuated equilibrium as the mode of origin for the new species may be achieved by several standards that might be characterized (somewhat metaphorically)
9-22. In other areas of the range, as here in South Dakota, only the descendant species lives during the entire interval.
[Page 851]
as devices for “dissecting” the punctuation by revealing an internal “fine structure” with probative value for inferences about evolutionary causes. Three major modes of dissection have been featured in the existing literature (although the theme has not been organized in this manner before), each explicitly invoked as a tool for the potential validation or refutation of punctuated equilibrium.
TIME. I discussed the operational definition of punctuation as scaled to periods of stasis (see pp. 765–774). The obvious barrier to testing this primary requirement of the theory lies in our inability to specify requisite information about time in “standard” paleontological situations, where the duration of speciation lies beneath the resolving power of our basic operational “moment” — the bedding plane. Therefore, to achieve a proper dissection, we must search for unusual situations that permit an adequate resolution of time in one of two manners dictated by the logic of the problem: either by finding a way to date individual specimens compressed on a single bedding plane, or by locating situations of unusually rapid sedimentation, where a sequence of events usually collapsed onto a single bedding plane can be expressed in true temporal order through a vertical sequence.
The first tactic can be applied only in highly unusual circumstances effectively limited to nearly modern bedding planes with specimens that can be dated individually, for the error bars associated with most radiometric techniques exceed the entire duration of most bedding planes (except for isotopes with very short half lives, which can then only be applied to Pleistocene or Holocene specimens). However, in a recent example (discussed more fully on p. 771 and Fig. 9-4), Goodfriend and Gould (1996) traced a species transition by hybridization in the land snail Cerion on the Bahamian island of Great Inagua. We found all specimens jumbled together on a modern mudflat (a bedding-plane-to-be, if you will), and we then used a combination of radiocarbon dating and amino acid racemization to determine that the smooth and complete species transition occupied 15,000 to 20,000 years — a reasonable figure for a punctuational event (here compressed, as usual, into a single geological “moment,” which we, thanks to the rare combination of recent occurrence and availability of dating techniques, were able to disaggregate and resolve).
In the far more common situation of sedimentation rates high enough to spread the usual compressions of single bedding planes into resolvable vertical sequences, assessments have been made in both relative and absolute terms. I previously cited Fortey's (1985) conclusion in the relative mode (see p. 769), based on calibrating punctuational origins against gradual transitions observed for other taxa in the same strata (thus obviating the usual claim that literal punctuations probably represent a geologically slow gradualism that extremely spotty sedimentation cannot record). With this technique, Fortey's found about a 10:1 ratio for punctuated vs. gradual origins of species in Ordovician trilobites from Spitzbergen.
The best examples in the more satisfactory absolute mode do not arise [Page 852] from direct paleontological records, but as firm inferences based on species flocks in lakes or on islands of known and recent origin (with African cichlid fishes as the classic case of modern evolutionary biology). These evolutionary “explosions” often produce several hundred species in just a few thousand years, and must be ranked as punctuational with a luxurious vengeance! But such circumstances do not represent a norm for most speciation in most clades, and such an unusual phenomenon, however stunning and however well documented, cannot suffice to validate a proposed generality.
The punctuational origin of many species can be accurately timed with direct paleontological data. Lister (1996) calculated a maximum of only 5000 years for the Quaternary evolution of dwarfed woolly mammoths on Wrangel Island, and 6000 years for the dwarfed red deer of Jersey. The punctuational origin of the marginellid gastropod Prunum christineladdae, based on the study of Nehm and Geary (1994), took 73,000 to 275,000 years, and spanned 0.6 to 2.5 percent of the full duration of the ancestral species. Reyment (1982) calculated outside limits of 100,000 to 200,000 years (perhaps a good deal less) for origin of the Cretaceous ostracode Oertiella chouberti from its ancestor, O. tarfayaensis.
GEOGRAPHY. I have already discussed this important tool for validating punctuated equilibrium by gathering data at a more inclusive and finer scale than the local documentation of a literal punctuation. On pp. 840–845, I described cases where geographic data affirmed an allopatric and punctuational event of cladogenesis, thus demonstrating that the abrupt appearance of a descendant species truly represents punctuated equilibrium, and not just a migrational incursion of a species that originated by an uncertain mode in an unknown place (Wei and Kennett, 1988, for protistans; Williamson, 1981, for mollusks; Lister, 1996, and Heaton, 1993, for mammals, among many others. Albanesi and Barnes (2000) present a particularly well documented case both for allopatric and punctuational origin of new taxa and survival of ancestors in their original regions for a lineage of Ordovician conodonts).
MORPHOMETRIC MODE. By quantitative study of patterns in morphological transition between ancestral and descendant species, several criteria of inference can increase our confidence in the identification of punctuated equilibrium, both by establishing a case for direct filiation rather than simple replacement by a taxon evolved elsewhere, and by indicating a punctuational mode for the cladogenetic event. As i
llustrations of this approach, consider three effective morphometric arguments:
1. Visually extensive change (supposedly requiring many independent inputs expressed over substantial time) can arise as coordinated consequences of one, or few, generating factors, and can therefore readily be accomplished at a punctuational tempo. This “standard” argument has a long pedigree, and serves many purposes, in evolutionary theory (see Chapter 10, for example, on “positive” constraints). In the context of punctuated equilibrium, [Page 853] this theme establishes plausibility for temporal compression of visually substantial change into a single cladogenetic event at a punctuational tempo. The argument also proceeds in several modes, including inductive approaches based on “covariance sets” (Gould, 1984b) of correlated characters transformed together along a single multivariate axis. For example, the transition documented by Goodfriend and Gould (1996) involved several measures coordinated by a single change in direction and rate of growth along the axis of coiling — a coherently correlated pattern running orthogonal to, and therefore independent of, the standard covariance set representing shell size alone. Since the cladogenetic event altered shape (as expressed along the axis of coiling), but not size, this multivariate separation established the source of morphometric change and also revealed its unitary nature in modified growth.
Among examples of the opposite deductive approach, based on fitting an apparent complexity of observed changes to a simple model of underlying generation, Smith and Paul (1985) recognized a suite of alterations through a punctuational event in Cretaceous echinoids as coordinated consequences of a single heterochronic change; and Benson (1983) explained a “punctuational event” (1983, p. 398) in the ostracode Poseidonamicus as a set of secondary, and mechanically automatic, accommodations of the carapace to a primary change in shape.
The Structure of Evolutionary Theory Page 135