9-23. Multivariate changes based on discriminant analysis of 10 characters throughout 14 to 20 temporal units in the evolution of seven molluscan genera in Miocene strata of Maryland. From Kelly, 1984. Stasis prevails within a large majority of species. For most lineages where a descendant replaces an ancestor, stair-step punctuation characterizes the transition. In a particularly interesting case of three successional species, ancestral Anadara does seem to move anagenetically towards the morphology of its descendant, which then remains quite stable. But the third and uppermost species, arising punctuationally, returns virtually to the morphology of the initial form.
[Page 860]
Two additional features enhance the methodological value of this study: first, these species belong to the best known and most intensely studied of all molluscan faunas; secondly, all species either still exist (12 of 19 cases) or can be compared with a close living relative (almost surely the immediate descendant in 4 cases, and perhaps directly filiated in the other 3). Thus, in the most important innovation of this study, temporal variation can be directly scaled against current geographic variation of the same species, or of a close relative. In testing whether temporal fluctuations exceed the limits of stasis, comparison with the range of geographic variation among current populations of the same species should serve as our best anchor and standard.
Using eigenshape analysis for multivariate representation of shell form, Stanley and Yang first compared variation among modern populations for each species with differences between these modern populations and early Pliocene (circa 4 million years old) samples of the same species. In a convincing demonstration of stasis properly scaled to realized intraspecific variation, Figure 9-24 (from Stanley and Yang, 1987, p. 124) shows histograms for overlap of eigenshape areas in comparing modern geographic variation with 4 million year distances between early Pliocene and modern samples. The temporal mode slightly exceeds the geographic value, but the ranges overlap completely, and the difference in central tendency is very small. The authors conclude (p. 113) that “with minor exceptions, the distribution of morphologic distances between 4 million year old and Recent populations resembled the distribution of distances between conspecific Recent populations.” “Approximate morphological stasis has been the rule for the taxa considered” (p. 124).
Stanley and Yang then extended their study (for species with available data) back to Miocene samples up to 17 million years old. Even for this extended duration, they found the same pattern of mild fluctuation, rarely extending outside the range of modern geographic variation, and with no
9-24. Non-overlap percentages based on comparisons of eigenshapes — from Stanley and Yang, 1987. The top histogram shows all conspecific pairs for the geographic variation of Recent populations. The bottom histogram shows differences between Recent populations and their presumed early Pleistocene ancestors. Note that the spread for the full geological range barely exceeds the spread for differences among geographic variants of living populations.
[Page 861]
accumulative directional effect. For example, Figure 9-25 (from Stanley and Yang, p. 132) shows the temporal distribution of mean values for each of the 24 characters over 17 million years in the venerid bivalve Macrocallista maculata. For most characters, the full temporal range lies within the variational scope of living populations (noted by the “forks” for separate geographic samples at the top of the trajectory). They conclude (p. 113): “We calculated net rates of evolution separating pairs of populations that belong to single lineages. For all intervals of time, the distribution of differences between population means for individual variables is remarkably similar to a comparable distribution representing the comparison of pairs of conspecific Recent populations from separate geographic regions . . . Evolution has followed a weak zigzag course, yielding only trivial net trends.”
A particularly impressive study by Prothero and Heaton (1996) documents the overwhelming dominance of punctuated equilibrium in a full tabulation of one of the most prominent fossil faunas — a study that also gives us good insight into how biased reporting in general, and Cordelia's dilemma in particular (p. 763), can so strongly skew tabulated results to appearances of equal frequency or only mild domination by punctuated equilibrium. These authors studied one of the world's richest and best known mammalian sequences — the upper Eocene and Oligocene White River Group of the American High Plains, particularly as exposed in the Big Badlands of South Dakota — “one of the densest and most complete records of mammalian evolution anywhere in the world . . . The spectacularly stark and beautiful outcrops ... have been a Mecca for fossil collectors ever since the first fossils were described in 1846 . . . Enormous collections have accumulated, and White River fossils are found in nearly every rock shop and mineral show across the country” (p. 259). This large mammalian assemblage seems to possess sufficient long-term coherence (from Duchesnean strata of the late middle Eocene into Arikareean strata of late Oligocene times) for designation as the White River Chronofauna (Emry, 1981).
The authors spent more than a decade conducting “an unbiased survey of
9-25. From Stanley and Yang, 1987. A history of change during 17 million years for each of 24 measured characters in the bivalve Macrocallista maculata. For the great majority of characters, the entire temporal spread lies within the scope of variation in the geographic range of living populations (represented by the “forks” for separate samples at the top of the trajectory). This form of comparison provides an excellent documentation of stasis by the criterion of scaling to the full range of geographic variation at a single time within the same taxon.
[Page 862]
all fossil mammal lineages . . . which have large enough sample sizes and recent systematic revision” (p. 258) for the seven million year period (37-30 million years ago) across the Eocene-Oligocene transition (Chadronian to Whitneyan North American land mammal “ages”). The protocol also includes the two other factors — good geographic spread and temporal resolution — most essential for proper studies of punctuated equilibrium, but all too rarely realized: “This study considers geographic variation over a wide area (from western Montana and North Dakota in the north and west, to Colorado in the south), with very fine-scale chronostratigraphic control from magnetic stratigraphy and 40Ar/ 39Ar dating” (p. 258). Finally, and fortunately, this interval includes a major global climatic change (with no disruption of continuity in sedimentation), thus permitting researchers to study how such an external input influences rates of speciation and styles of phyletic change.
Prothero and Heaton found near exclusivity for punctuated equilibrium in the 177 well-documented mammalian species of this fauna. “Most species are static for 2-4 million years on average, and some persist much longer” (p. 257). “Only three examples of gradualism can be documented in the entire fauna, and these are mostly size changes” (p. 257). The details of these three cases also illustrate the exceptional status of gradualism, even at the smaller scale of their own taxonomic context:
1. The lagomorph Falaeolagus undergoes reduction in size of upper molars, accompanied by loss of their roots, during the early Orellan, but maintains stasis for much longer invervals both above and below: “Chadronian falaeolagus shows about 2 m.y. of stasis, followed by gradual reduction in size and development of rootless upper molars during the early Orellan. From Orella B onward, several species of Falaeolagus are present, and except for slight changes, they are static for several million years” (p. 273).
2. The artiodactyl Leptomeryx experiences “subtle, gradual change in a number of characters” (p. 263) in the transition from L. speciosus to L. evansi, but both species show stasis throughout most of their substantial history — so we do not here witness the “classic” continuous anagenesis that supposedly makes the definition of species so arbitrary in temporal sequences: “While the transition from L. speciosus to L. evansi is not stratigraphically instantaneous, it occurs in a r
elatively short time compared to the long durations of both species.”
3. The oreodont Merycoidodon does seem to undergo extensive and gradual dwarfing (30 percent size reduction) over a one million year interval in the early Orellan. I accept this case as a good example of extended gradualism (see Fig. 9-26 taken from Prothero and Heaton, 1996, p. 262), but also note that the trend occurs within a common genus, including several species otherwise showing predominant stasis — and that the dwarfing trend only involved the labile character of size, without concomitant changes of shape, a common finding among exceptional cases of gradualism in faunas dominated by punctuated equilibrium (see previous discussion of Jurassic bivalves on page 855).
Such exhaustive and unprejudiced tabulations can give us insight into the [Page 863] limited value — that is, for establishing proper relative frequencies, not for resolutions of particular cases — of trying to infer the quantitative distribution of rates and modes for all taxa from previously published research carried out within the “best case” tradition, usually with strong (and unacknowledged) preference for defining evolution only as geologically gradual change. Before
9-26. A rare case of gradualism amongst the overwhelming domination in relative frequency for stasis in 177 well-documented mammalian species of the Big Badlands fauna of South Dakota. This species of Miniochoerus (previously known as Merycoidodon) does undergo gradual dwarfing to 30% size reduction over a one million year interval. From Prothero and Heaton, 1996.
[Page 864]
the punctuated equilibrium debate began, how would an evolutionary paleontologist have treated the White River Fauna? Almost surely, any expert on these strata would have selected the cases of apparent gradualism for study and publication, while ignoring the others as negative instances of no evolution, worth only a side comment at best, if noted at all, and suited for explicit mention (but without any quantitative analysis) only in formal taxonomic treatises. Thus, readers with no personal knowledge of the entire fauna — especially non-paleontological readers unaware of strong signals for stasis and punctuation in virtually all faunas — would almost surely assume that the three reported studies characterized the usual situation for the history of fossil species, rather than representing the only examples of a rare phenomenon.
Prothero and Heaton (1996, p. 258) raise the important point that the examples of gradualism most widely featured, and most frequently cited to urge the general case against punctuated equilibrium, derive from such faunas — where they stand as unusual examples against an unstudied (or simply non-discussed) but overwhelming prevalence for stasis and punctuation among all species. They remind us, for example, that Gingerich's most famous half dozen or so cases from lower Eocene beds of the northern Bighorn Basin “are just part of a fauna of over a hundred genera. Detailed monographs by Bown (1979), Schankler (1980), and Gingerich (1989) [in his very own taxonomic work] have shown that stasis is prevalent among most of the taxa not featured by Gingerich (1976, 1980, 1987)” (p. 258). Of another famous claim for gradualism (one that I do not challenge as a single case, while asking that relative frequencies also be acknowledged), Prothero and Heaton (1996, p. 258) write: “Krishtalka and Stucky (1985, 1986) reported a gradualistic transformation in the early Eocene artiodactyl Diacodexis. However, this is a single lineage from the same faunas described by Schankler, Gingerich, and Bown, so these studies do not address the overall prevalence of gradualism vs. stasis.”
Finally, although this issue belongs more to the forthcoming discussion of faunal stasis as an extension of punctuated equilibrium (see pp. 916–922), Prothero and Heaton's (natural) experimental design in choosing the White River Chronofauna for such intensive study included the existence, in the midst of the fauna's duration, of one of the most profound and rapid climatic changes in Tertiary North America — “the earliest Oligocene climatic crash” (p. 257) at 33.2 million years ago, where “vegetation changed from dense forests to open forested grassland, mean annual temperatures dropped 13°C, and conditions got much drier and more seasonal” (p. 257). The nondisturbance of stasis — indeed, the virtual “ignorance” of this event by most species, at least by observable changes in diversity or skeletal anatomy — also illustrates the strength of stasis and the apparently active (rather than merely passive) sources of its maintenance. Prothero and Heaton write (1996, p. 257): “Only a few mammalian lineages speciated, a few more went extinct, and the vast majority (62 out of 70) persisted through this climatic event with no observable response whatsoever.” The authors then end their paper by throwing down the gauntlet to supporters of traditional evolutionary views [Page 865] about response to environmental change: “Evolutionary theory has come a long way since Eldredge and Gould (1972) first pointed out that stasis is the norm in the fossil record, and the data cannot be simply dismissed or explained away ... In fact, stasis and resistance to change is so ingrained that species can actually pass through the most significant climatic change of the last 65 million years as if nothing happened.”
In recent years, studies of stasis and punctuation in entire faunas have blossomed, especially with the introduction and testing of two partly complementary, but partly dissonant, explanations for apparently concerted stability of entire faunas over substantial intervals — the turnover-pulse hypothesis of Vrba (1985), and the theory of coordinated stasis, developed by Brett and Baird (1995 and several other works), and extensively (often contentiously) treated in the symposium of Ivany and Schopf, 1996. I shall treat the theory itself in the last section of this chapter (pages 916–922), but will record here the convincing documentation of extensive faunal stasis that established the evidentiary base for these ideas.
The famous Middle Devonian (Givetian) Hamilton fauna of the Appalachian Basin has provided a “type” case for coordinated stasis. The Hamilton fauna includes more than 330 species of mostly typical Paleozoic invertebrates, ranging through about 9 million years of strata in a series of about twenty identifiable “communities” or biofacies. About 80 percent of these species persist throughout the entire Hamilton, while fewer than 20 percent carry over from the fauna just below. Ever since Cleland's original and wistful comment in 1903 (quoted on p. 750), students of the Hamilton fauna have recognized the overwhelming signal of stasis presented by almost every species of the assemblage — although this original wistfulness has now ceded to considerable positive interest! Brett and Baird write (1995, p. 301): “Individual lineages within particular biofacies of the Hamilton biotas appear to display very little morphological change, and that which is observed is neither progressive nor directional.”
Several taxa of this fauna have now been analyzed in great morphometric detail, beginning with Eldredge's classic study of the trilobite Phacops rana, one of the “founding” examples of punctuated equilibrium (see Eldredge, 1971; Eldredge and Gould, 1972). Eldredge found stasis in more than 50 characters, and directional, but punctuational, change only in one — reduction in rows of eye facets in two punctuational steps during a 5-6 million year period otherwise marked by stasis for this feature as well. Other quantitative studies of stasis in Hamilton species include Pandolfi and Burke (1989) for tabulate corals, Lieberman (1995) on trilobites, and Lieberman, Brett, and Eldredge (1994) on brachiopods. The last study considered 8 characters in two species using principal components and canonical discriminant analysis. The authors found fluctuating variation, correlated neither with age nor facies, throughout the interval. However, and ironically given past expectations of gradualism, the uppermost samples plotted closer to the lowermost than to any intervening population. For the entire fauna, Brett and Baird conclude (1995, p. 303): “Taken together, these studies indicate that a [Page 866] majority of Hamilton lineages display virtual stasis from oldest to youngest samples. Slight nondirectional change is observed in some cases. Such variation seemingly records very minor evolutionary fluctuation . . . However, it clearly does not lead to development of major new grades of morphological
development . . . Several of the species appear abruptly in the Appalachian Basin near the beginning of the Hamilton fauna or become locally extinct at its end.”
Relative frequencies for entire clades
We add another component to studies of relative frequencies when, in addition to the thoroughness provided by assessing all lineages within a given time or region, we add the phylogenetic component of complete coverage for clades (preferably monophyletic of course, but sometimes paraphyletic in the existing literature). Obviously, we feel most secure about such phylogenetic assertions when truly cladistic, or at least stratophenetic, criteria have been used for definitions, but many studies in this mode employ a standard that, albeit and admittedly less preferable, probably provides as much confidence in practical utility: investigations of distinctive taxa known on reliable bio-geographic grounds to be restricted to a region exhaustively studied. Clades confined to isolated islands, lakes, or other such distinct and coherent places and environments constitute our best cases under this criterion.
I have discussed nearly all the best examples in this mode under other headings of this chapter, and will only make brief reference here. Several “classic” mammalian lineages fall into this category of excellent cladistic definition and overwhelming domination by the punctuational pattern of stasis within species and geologically abrupt transitions between — all despite (or rather, in a punctuational reformulation, because of!) such celebrated evidence for sustained and important trends. I include here the excellent evidence for horses (see p. 905) and the spottier but still persuasive data on hominid evolution (see pp. 908–916) — in each case, for clades well delimited both by morphology and geography.
The Structure of Evolutionary Theory Page 137