Naturwissenschaften (2005) 92:1–19 DOI 10.1007/s00114-004-0586-9

REVIEW

L. D. Martin · T. J. Meehan Extinction may not be forever

Published online: 16 November 2004 Springer-Verlag 2004

Abstract Here we review the phenomenon of ecomorph Introduction evolution and the hypothesis of iterative climatic cycles. Although a widely known phenomenon, convergent evo- Here we review the hypothesis of iterative climatic cy- lution has been underappreciated in both its scope and cles, which states that the evolution of faunas and floras commonality. The power of natural selection to override and their extinction has a predictable pattern. South genealogy to create similar morphologies (even among American pollen and North American mammalian as- distantly related organisms) supports classical Darwinian semblages exhibit convergent evolution in repeating A-B- evolution. That this occurs repeatedly in stratigraphically C cycles (van der Hammen 1957, 1965; Martin 1985; closely spaced intervals is one of the most striking features Meehan and Martin 2003), and these consecutive A-B-C of Earth history. Periodic extinctions followed by re- communities form a chronofauna. The stability and then evolution of adaptive types (ecomorphs) are not isolated extinction of these communities have been correlated to occurrences but are embedded within complex ecological cycling sedimentary and temperature profiles. This pat- systems that evolve, become extinct, and repeat them- tern is reflected in simultaneous radiations of convergent selves in temporal synchrony. These complexes of radia- adaptive types (ecomorphs) on separate continents, indi- tion and extinction bundle the biostratigraphic record and cating that it results from natural selection caused by provide the basis for a global stratigraphy. At this scale, global climatic change, as opposed to genetic or com- climatic change is the only mechanism adequate to explain munity biotic factors. An iterative pattern of hierarchical the observed record of repeating faunas and floras. Un- climatic cycles may form the underlying basis for bios- derstanding of the underlying causes may lead to predic- tratigraphy and explain most evolutionary trends and tive theories of global biostratigraphy, evolutionary pro- extinctions. The cycles recognized so far appear to re- cesses, and climatic change. present equal units of time—2.4 Ma for each A, B, and C cycle, and 7.2 Ma for this chronofauna triplet. Whether the cycles represent equal units of time is integral to understanding the cause of these cycles, but is not integral to the main thrust of the argument; repeating ecomorph evolution and extinction among different lineages occurs synchronously, and only climatic change has broad en- ough effects to produce this pattern. L. D. Martin ()) It would be very important to establish that climatic Natural History Museum and Biodiversity Research Center, change and evolutionary processes, including extinction, Department of Ecology and Evolutionary Biology, University of Kansas, are due to random historical accidents, but it would also 1345 Jayhawk Blvd, Lawrence, KS 66045–7561, USA be a scientific dead end. A predictive model of climatic e-mail: [email protected] change and correlated evolutionary processes is obviously Fax: +1-785-8645335 much more desirable, but can such a model be con- structed? While the basis of stratigraphy and almost all T. J. Meehan geology is the Law of Superposition, we must use fossils Division of Science, Chatham College, Buhl Hall, Woodland Rd, Pittsburgh, PA 15232–9987, USA to construct a regional or global stratigraphy. The units of biostratigraphy are unique combinations of last and first T. J. Meehan appearances of organisms. If such events were randomly Research Associate, distributed over the rock column, boundaries would result Carnegie Museum of Natural History, solely from historical accidents, and these boundaries Pittsburgh, Pennsylvania, USA 2 might change with the whims of prevailing academic these changes will have a global signature. Some version politics. It seems obvious that it would be better if a of this scenario must lie at the root of biostratigraphy. boundary corresponded to a recognizable global event. A climatic-evolutionary connection seems fundamental to biostratigraphy, and Krasilov (1974) advocated a climat- Ecomorphs ic-based system in his causal biostratigraphy. He impli- cated overall ecological change as the key to under- The similarities among organisms are a greater theoretical standing biostratigraphy and stated that the succession of problem than differences. Differences can and should ecosystems is controlled by climatic cycles. This implies result from random events over time. Historical accidents that there are bundles of time that can be recognized on come into full play when we examine how organisms the basis of unique biological compositions and the im- differ, but how are we to explain characters that are pact of climate on the sedimentary record. There is a conserved over vast intervals of geologic time? We see school of thought that claims that there is no visible ev- immense amounts of conserved similarity in genetic idence for climatic impact on evolution (e.g., Prothero composition and cellular processes over the hundreds of 1999; Alroy 2000), but most workers see a clear con- millions of years that organisms have inhabited Earth. nection, and Darwinian evolution seems to demand such a Without the action of natural selection, such similarities result. There are numerous studies of modern organisms would soon have fallen prey to random processes. There correlating climate and adaptation, and natural selection are also not an infinite number of solutions to biological resulting from climatic change has been observed in a problems. In fact the number of solutions seems to be three-decade study of Darwin’s finches (Grant and Grant quite small judging from the number of times that the 2002). From the fossil record, we have empirical evidence same solution is evolved independently. A computer that climate changes over time and that biota are directly simulation of “organisms” shows that with strong selec- affected. For instance, Europe was fully tropical in the tion, convergent evolution of even a complex trait is Middle Eocene, as evidenced by the famous deposits at common (Lenski et al. 2003). Natural selection produces Messel where we find rainforest trees and animals whose suites of coordinated similarity resulting from shared closest analogues are in Africa today (Schaal and Ziegler activities, rather than shared phylogeny. If the shared 1988), while fossils of 18,000 years ago indicate a tundra activity is an integral part of an “ecological occupation,” flora and fauna (von Koenigswald and Hahn 1981). The it may predict other similarities related to that occupation. overall global trend during the past 55 Ma has been to- Cuvier’s great contribution to comparative anatomy was wards cooling, but we are presently experiencing a re- the principle of correlation—the idea that changes in one versal of that trend. There is relatively little evidence of anatomical suite required concordant changes in others long periods of climatic stasis, and all long-term patterns and that a lifestyle could be predicted from a subset of are interrupted by periodic fluctuations where the general correlated structures. It is this ability of ecological posi- trend is reversed. tion to predict anatomy independent of phylogeny that is The usual measures for climatic regime are tempera- the basis of the ecomorph concept (Martin and Naples ture and its effects as expressed in terms of water. Hot 2002). worlds are wet worlds because of increased energy and Convergent organisms have been called ecomorphs water surface area for evaporation, while cold worlds are (ecological morphotypes; Williams 1972), which can be dry. Wet worlds favor canopy strategists (trees) and cold correlated to Van Valen’s (1971) adaptive zone—an or- worlds pioneering plants and open vegetational structures ganism’s resource space together with relevant predation (Martin 1994). Wet worlds favor sedimentary deposition and parasitism. An important property of adaptive zones with increased plant cover and a higher base level, is that they exist as opportunities within the ecological whereas dry worlds are characterized by increased ero- framework and are defined by specific resources. The sion. Climate is mediated locally, and it is possible for a existence of an adaptive zone creates an opportunity for local region to be dry when global averages are wetter. the development of a specialized organism to occupy it, Ultimately weather patterns are interconnected, and cli- but does not require or imply that such an organism exists mate cannot change greatly over any large region without (Martin and Naples 2002). Simpson (1953:161) antici- affecting all regions. Usually environmental change is bad pated this view: “Possible ways of life are always re- for established organisms. Changing rainforest to grass- stricted in two ways: the environment must offer the land may be bad for monkeys, while changing grassland opportunity and a group of organisms must have the to forest may be bad for wildebeest. Global change is possibility of seizing this opportunity.” Ecologies contain likely to have a negative impact on organisms best opportunities that define circumscribed morphologies, adapted to the status quo. On the other hand, organisms including physiology and behavior. Convergent forms that occupy marginal habitats may find their habitat evolve independently in separated but similar ecologies, greatly expanded as climate changes (e.g., Martin 1994). demanding only that adequate predecessors for the mor- We may predict that rapid environmental change will phological type be present. Similar ecomorphs generally result in nearly simultaneous extinction of many taxa and do not evolve in the same region at the same time, but dramatic biogeographic reorganization of others and that require either geographic or temporal separation (Martin and Naples 2002). 3

Fig. 1A, B Correlated climatic cycles of North and South America. America as compared to Stout’s (1978) sedimentary/climatic cycles A Top graph: iteration of A, B, and C community types of pollen of Nebraska correlate to the triplet climatic cycles of van der ecomorphs and the cyclic nature of abundances of the palm paly- Hammen (1961). Martin’s interpretation (1985) of Stout’s cycles is nomorph group, Monocolpites medius, in South America (modified shown, as well as the independent evolution of dirktooths within from van der Hammen 1961: Fig. 1). Increasing abundances of M. these cycles. The Barstovian and Hemphillian are poorly repre- medius (shaded) indicate decreasing temperatures (towards bottom sented in Nebraska. The base of each sequence is characterized by of graph). Note that more severe cooling characterizes the end of heavy fluvial incision and deposition, and the top, by eolian and/or subcycle C. Bottom graph: correlated North American subcycles caliche horizons. The paucity of cat ecomorphs in the Hemingford based on mammalian assemblages (Martin and Meehan 2002). Cycle is referred to as the “cat gap.” Note that in North America Each A-B-C community represents a vadh climatic cycle, and this formational boundaries are not necessarily the exact correlates of triplet forms a stout climatic cycle. B Climatic cycles and dirktooth biostratigraphic boundaries, so that these climatic cycles correlate iterative evolution. Stratigraphic distribution of latest Eocene– with faunal assemblages (NALMAs or subages), and not with de- Pleistocene (Chadronian-Irvingtonian) mammalian faunas of North fined rock boundaries. Modified from Martin 1985: Fig. 5 4 Convergent adaptive types plagued taxonomic work in base of each sequence is dominated by fluvial incision Cuvier’s time and continue to do so. Ecomorph analysis is and deposition, the middle cycle by mixed fluvial and “free of taxonomy” (Damuth et al. 1992) and yields a eolian infilling, and the terminal cycle by eolian deposi- perspective on evolutionary processes not attainable from tion and an increase in caliche paleosols. This triplet a phylogenetic approach. Organisms live and evolve pattern reflects a wet climate at the base, a moderate within communities, so perhaps taxonomic units are the climate in the middle, and a much drier climate at the top. wrong units for understanding causal mechanisms (Mee- The White River Cycle as reflected in deposits of Ne- han and Martin 2003). Large-scale ecological analyses in braska (latest Eocene to early Oligocene) is an excellent deep time are rare even though the benefit of such work is example of this pattern (Fig. 1B). The Chadron Formation widely advocated (e.g., van der Hammen 1965; Krasilov is predominantly channel and floodplain deposits, with 1974; Fischer 1981; Damuth et al. 1992). Large-scale occasional lake and pond deposits. The Orella Member of paleontological analyses of South American pollen (van the Brule Formation consists primarily of channel and der Hammen 1957; Leidelmeyer 1966) and North floodplain deposits, with rare pond deposits and some American mammals (Martin 1985; Meehan and Martin eolian influence. The Whitney Member of the Brule 2003) reveal repeated evolution and extinction of eco- Formation in Nebraska is a loess deposit. Schultz and morph types, suggesting the existence of global iterative Stout (1980) stated that major mammalian extinctions climatic cycles. occur at the unconformable boundaries in these sequences and suggested that these unconformities indicate dry, cool periods with relatively little plant cover and more erosion. Hypothesis of iterative climatic cycles Soil changes, as well as faunal and floral changes, support this climatic interpretation (Clark et al. 1967; Retallack Van der Hammen (1957, 1961) discovered an iterative 1983). This triplet pattern matches van der Hammen’s A- pattern in a succession of Late Cretaceous to late Ceno- B-C cycles, including the terminal cooling and drying, zoic pollen assemblages from oil well core samples in which is most severe at the end of subcycle C. Colombia, South America. Pollen shapes can be highly In relating Stout’s sedimentary cycles to mammalian convergent and are not assignable to species level, but are faunas, Martin (1985) recognized an iterative A-B-C classified as form genera (palynomorphs). Pollen classi- pattern in mammalian communities. Through dispersal fication is basically a functional one of adaptive types, and adaptive radiation, a community type would develop similar to leaf shape ecomorphs (e.g., Wolfe 1985). Van and become extinct—only to redevelop in the same place der Hammen (1957) described pollen abundances of during the next sedimentary cycle. Comparing faunas palms, other angiosperms, and ferns that had synchronous from the same position within different sedimentary cy- minima and maxima at apparently regular intervals, rep- cles, Martin showed that they resembled each other in resenting three assemblage types. Pollen community A diversity and ecomorph composition, being more similar was succeeded by a B community and the B succeeded by to one another than to the subcycles immediately below a C before the pattern repeated with the re-evolution of a and above in the stratigraphic sequence. Figure 1B com- convergent A community. This A-B-C triplet pattern was pares Martin’s interpretation of North American land best expressed by palm pollen of the Monocolpites medius mammal “ages” (NALMAs) and their subdivisions with group (Fig. 1A). Stout’s climatic/sedimentary cycles. The extinction of a Abundance of M. medius pollen reflected temperature dirktooth ecomorph at the end of one cycle and its re- trends during the Cenozoic, with a relatively cool period evolution in the next cycle exemplify repeating ecomorph in the early Paleocene, extensive warming into the early replacement. Martin (1985) proposed that there was an Eocene, and then overall cooling up to the Recent (van iterative evolution of mammalian communities with a der Hammen 1961). On a smaller scale, minima and maxima of pollen abundances within each community type represented cyclic warming/cooling periods on the Fig. 2A–D The A-B-C pattern in dominance turnover of mam- order of 2 Ma (Fig. 1B). Each subcycle begins relatively malian ecomorphs in the North American Cenozoic. A Dominant cool, warms considerably, is somewhat stable, and then herbivore ecomorphs of the order Artiodactyla: s Oreodontidae I radiation; D= Oreodontidae II radiation; Moschidae and Dro- cools sharply at the end. Extinction at the community momerycidae; o Antilocapridae; l Bovinae and Cervinae; B level occurs near this temperature minimum. At the end of Fossorial ecomorphs of the order Rodentia (extensive burrowers, subcycles A and B, the temperature drop is similar in such as pocket gophers, mountain beavers, and prairie dogs): n magnitude, and at the end of subcycle C, the cooling is Cylindrodontidae; o Aplodontidae and Allomyidae; ' Castoridae/ o greater. Subcycle A is warmest, B almost as warm, and C Palaeocastorinae and Geomyidae/Entoptychinae; Mylagaulidae; s Sciuridae/Sciurinae and Geomyidae/Geomyinae; C All cat much cooler. ecomorphs: s Oxyaenidae; early Nimravidae; n early Felidae Stout (1978) and Schultz and Stout (1980) recognized and late Nimravidae; D= late Felidae; D Shrew ecomorphs (cryptic, sedimentary cycles in the central Great Plains of North leaf-litter insectivores): D Cimolesta; l Nyctitheriidae; o Sorici- America that reflected regular changes in climate since dae/Heterosoricinae; Soricidae/Soricinae; minor shrew eco- morph radiations not represented for graph clarity. The 27 Ceno- the mid-Cenozoic. Studies of fluvial terraces, soil hori- zoic units are North American land mammal “ages” or subages that zons, and loess deposits show a repeating pattern of valley reflect vadh climatic cycles (Martin 1985; Martin and Meehan cut-and-fill sequences related to fluctuating aridity. The 2002); see Table 2 for abbreviations 5 6 period of about 7 Ma. Martin also showed that the dom- cycle (= stout), lake sedimentation thickness for each inance turnover in selenodont artiodactyls reflects this subcycle (= vadh) was approximately equal; and (2) from iteration (Fig. 2A). He further argued that this pattern was a theoretical consideration, a given pollen community climatically controlled, showing a correlation with the type would likely take a similar amount of time to de- cyclic temperature curve of Wolfe and Hopkins (1967). velop. Although van der Hammen’s criteria are not Meehan and Martin (2003) reported other mammalian compelling, radiometric dates from associated mam- iteration examples in hippo, dog, and bone-crushing dog malian faunas in North America do suggest equal units of ecomorphs. This pattern is also reflected in the dominance time (Martin 1985). Martin also found comparable turnover of fossorial rodent, all catlike, and shrew eco- changes in diversity of late Cenozoic mammalian com- morphs (Fig. 2B–D). munities that suggested regular cycling (Fig. 3). The Martin and Meehan (2002) preliminarily extended this terminal extinction events, which appear to be synchro- pattern of climatic cycles based on North American nous and rapid (Woodburne 1987; Webb 1989), were of mammalian faunas to the early Cenozoic and calculated a similar magnitude, and generic diversity returned to a more refined interval. Recent radiometric dates associated similar level (Fig. 3A). Although adaptive radiation with mammalian faunas suggest a subcycle duration of within North America generates much of the diversity, 2.41 Ma and 7.23 Ma for the triplet. A subcycle thus dispersal is an important source of radiations (Martin approximates the duration of a marine zone, and a cycle a 1985). For example, the earliest Dromomerycidae and marine stage. Benton (1995) reported an average Cervinae/Bovinae artiodactyls (refer to Fig. 2A) dispersed Phanerozoic stage duration of 7.4 Ma. The similarity of from Asia. The timing of these dispersals appears to be biostratigraphic systems (e.g., shells, mammals, and pol- related to Stout’s cycles, suggesting that climatic change len) and, in particular, their correlation on a regional, and is a controlling factor, with dispersal into middle and low sometimes global scale, reflects an underlying basis. latitudes representing time-transgressive range extension In studying Pleistocene terrace sequences and recog- from higher latitudes (Martin 1985, 1994). With global nizing correlations of cut-and-fill sequences across central temperature generally decreasing since the Eocene, cool- North America, as well as worldwide correlated strati- adapted immigrants from higher latitudes and their de- graphic boundaries and climatic events, Schultz and Stout scendants were at an advantage, and the percentage of (1980) concluded that these sedimentary cycles represent Holarctic immigrants and their descendants in North global climatic change. Van der Hammen (1961) came to America continually increased through the late Cenozoic the same conclusion concerning the cause of pollen (Martin 1994). community patterns, as did Wolfe (1978) concerning Based on the K/T boundary, van der Hammen (1965) fluctuations in leaf ecomorph assemblages. Martin and estimated that a cycle lasted 7 Ma, making a subcycle Meehan (2002) concurred that any force able to simul- 2.33 Ma. In 1985, Martin had access to more radiometric taneously affect communities in North and South America dates to test whether these subcycles were of equal du- would have global ramifications. Iterative evolution of ration. He plotted 72 dates correlated to mammalian ecomorphs and community structures requires repetition faunas (latest Eocene through Pleistocene) against the of environmental parameters, which suggests that climatic biostratigraphic divisions. The data formed a highly linear cycles form the underlying basis for a global stratigraphy relationship, suggesting equal units of time. Martin de- (van der Hammen 1957, 1961, 1965; Martin 1985; Martin termined that the radiometric age midrange points for and Meehan 2002; Meehan and Martin 2003). Because each unit were not statistically different from an ideal stratigraphic systems tend to be more localized and carry duration calculated from a radiometric date near the base with them significant historical baggage, Martin and of the Geringian (late Oligocene) divided by the number Meehan (2002) proposed a new nomenclature for these of subsequent vadhs (28 Ma divided by 12 vadhs). This cycles. Each subcycle of about 2.4 Ma is termed a “vadh” gave an interval of 2.33 Ma. for T. van der Hammen, who first characterized this Another approach is to regress radiometric dates pattern and recognized its importance. The complete A-B- against vadh units so that the slope of the regression line C cycle is termed a “stout,” after T.M. Stout, who dis- represents the likely vadh duration. Over the past several cerned this pattern in sedimentary cycles of the Great decades, radiometric dating analyses have gone from Plains and recognized its global significance in relation to whole-rock to single-crystal, which has improved preci- marine stages. Vadhs A, B, and C form a stout in this sion. Although the radiometric dates of Evernden et al. terminology for climatic cycles and may be measures of (1964) were obtained by whole-rock analysis, their data time analogous to days, years, and Milankovitch astro- set has the advantage of being processed in one lab and is nomical parameters. still the most comprehensive, single report for dated NALMAs. Updating these radiometric age estimates (n=55) with the new potassium–argon half-life standard Do these cycles represent equal units of time? (Dalrymple 1979), and then regressing these ages against the 27 Cenozoic vadhs, yields a coefficient of 0.991 and Van der Hammen (1961) considered these repetitions in vadh duration estimate of 2.44 Ma (Martin and Meehan community types to be of equal duration. He assumed 2002). Repeating the same exercise using recent (post- equal units of time based on two criteria: (1) within each 1985), higher precision radiometric dates yields a coef- 7

Fig. 3A, B Generic (ecomorph) counts of terrestrial mammals in value for terrestrial genera is estimated to be approximately 165 the North American Cenozoic. A Generic abundance per vadh. (per 2.4 Ma). The North American mammalian fossil record is Each genus is assigned an ecomorph type, so generic abundance is principally from middle latitudes of the West, so preservational/ equivalent to ecomorph abundance. Equilibrium in mammalian sampling biases may preclude an accurate estimate. Number of communities appears to have been reached 10 Ma after the K/T recognized Cenozoic genera is currently 1,301 (excluding bats and extinction, and the average number of known genera from the marine mammals), which represents 3,375 distributional points Eocene–Pleistocene (Gray-Irvi) is 131. Sample quality generally across the 27 vadhs. B Percentages of surviving/extinct genera. n increases the younger the rock unit, and the average from the mid- percentage of genera that survived into the next vadh; o per- Miocene to Pleistocene (Bars-Irvi) is 151. Modern generic diversity centage of extinct genera. Some units are not well represented in correlates with land mass area (Flessa 1975), which has been fairly the geologic record (e.g., Monroecreekian and Lapointian). The stable for North America since the Paleocene. The equilibrium average extinction rate since the early Eocene is 32% ficient of 0.999 and vadh of 2.41 Ma (Martin and Meehan Not even a regression coefficient of 99.9% statistically 2002; Fig. 4). It is important to note that vadh divisions demonstrates that the 27 vadhs are of equal duration and inferred from mammalian faunas are biostratigraphic represent the best linear fit for dividing Cenozoic bios- boundaries defined by other workers who did not suspect tratigraphy—there are too few radiometric dates (n=99). that these divisions might represent equal units of time. Also, many of these dates are clumped so that regressions 8 Fig. 4 Recent (post-1985) ra- diometric dates regressed against Cenozoic vadhs. Re- gression of dates associated with North American mam- malian faunas against 27 units yields an estimated vadh dura- tion of 2.41 Ma. Updated from Martin and Meehan 2002: Fig. 2

Table 1 Extrapolated Cenozoic epoch boundary dates versus predicted vadh values Epoch Harland 1989 Berggren et al. 1995 Gradstein Ogg 2004 Janis et al. 1998 Predicted Marine dates Marine dates Marine dates Terrestrial dates vadh ages Pleistocene a 2 1.3 1.81 1.8 2.41 Ma Pliocene 5 5.3 5.33 4.5 4.82 Ma Miocene 24 23.8 23.03 23.0 24.10 Ma Oligocene 36 33.7 33.9 33.4 33.74 Ma Eocene 57 55.5 55.8 55.5 55.43 Ma Paleocene 65 65.0 65.5 65.1 65.07 Ma Vadh estimate 2.44 Ma 2.41 Ma 2.44 Ma – – R-squared 0.991 1.000 0.999 – – 95% CI 2.35–2.53 2.36–2.45 2.36–2.51 – – a There is a climatic cooling event recorded worldwide at 2.4 Ma, and some workers have advocated that the base of the Pleistocene be moved to this position, which agrees with the terminal cooling and boundary location predicted by vadh climatic cycles. The ideal vadh ages are estimated from radiometric dates associated with mammal deposits in North America (Martin and Meehan 2002; Fig. 4). Vadh duration estimates from the marine record are equal within a 95% confidence interval (CI). are heavily influenced by a few horizons, such as the K/T The Plio-Pleistocene boundary is the most divergent boundary. Although radiometric date regressions do not point (Table 1), but unlike the Pleistocene–Holocene prove the existence of 27 units of equal duration, the high boundary, it is not defined by a significant global climatic coefficients imply equal units, and radiometric ages event. It is a political boundary where a committee drove cannot be used to argue against this portion of the hy- a “golden spike” at a local foram appearance/extinction pothesis. There is, however, an independent, partial test of (Harland 1989:68). This golden spike philosophy was linearity of these data via epoch boundary ages from a applied due to great controversy among workers, yet no different source: marine sediments (Martin and Meehan dissension arises among marine and terrestrial workers 2002). Assuming epoch boundaries from the marine time about a sharp, global cooling event 2.4 Ma ago, incurring scale are correlated correctly with NALMAs, one would great biological consequences. This climatic change is predict the same linearity when marine boundary ages are evident in such events as (1) a drop in ocean tempera- plotted against the vadhs. Using extrapolated epoch tures; (2) 65% extinction of the tropical western Atlantic boundary ages from marine radiometric dates (Harland et mollusc species; (3) abrupt change in carbonate produc- al. 1989) yields a regression coefficient of 0.99 and vadh tivity and preservation; (4) weather pattern changes in the duration of 2.44 Ma (Martin and Meehan 2002). Running Middle East as indicated by dust deposition; (5) a great the same analysis from two other stratigraphic charts increase in tundra habitat and loess deposition; and (6) (Berggren et al. 1995; Gradstein and Ogg 2004) yields 5C cooling in Colombia as indicated by pollen assem- coefficients of 1.00 and 0.99 with estimated vadh dura- blages (Clark et al. 1980; Liu et al. 1985; Wolfe 1985; tions of 2.41 and 2.44 Ma, respectively. These three es- Stanley 1986; Jansen et al. 1988; Curry et al. 1990; timates are equivalent within a 95% confidence interval Kennett and Barker 1990; Crowley and North 1991; De- (Table 1). Menocal et al. 1991; Hooghiemstra and Ran 1994). Var- 9 ious workers have advocated that the Plio-Pleistocene concluded that Phanerozoic extinction patterns were most boundary be defined by this cooling event (e.g., Liu et al. likely caused by widespread environmental upheaval. 1985), and in Europe, some workers have been using this In contrast, Stucky (1995) suggested that community event for decades as the Quaternary/Tertiary boundary interaction was a more important factor than the physical because it is the start of the first pronounced glacial pulse environment in determining survivorship. Stucky further (e.g., van der Hammen et al. 1971). noted a contemporaneous global trend of increased hyp- sodonty among mammals. Convergent trends of geo- graphically isolated faunas indicate a cause that cannot be Patterns consistent with iterative climatic cycles from competition or some other community interaction. Congruence of faunas across continents by simultaneous The fundamental pattern of evolution is one of relative appearance of the same taxa, similar grade of evolution, stasis bounded by abrupt evolutionary change due to ex- or convergent community structure indicates that climatic trinsic factors. Communities rapidly evolve, remain stable change is a dominant force in large-scale evolution. for long periods, rapidly go extinct, and then a new Where Stucky’s data reflect stasis in community struc- community replaces the previous one. This pattern is so ture, we would argue that it is due to relative stasis in commonly recognized that it has been given many de- climate. scriptors: typostatic and typogenetic/typolytic phases of evolutionary cycles (Schindewolf 1950), chronofaunas (Olson 1952), iterative climatic cycles (van der Hammen Sedimentary cycles 1957; Martin 1985), biomere boundaries (Palmer 1965), punctuated equilibrium (Eldredge and Gould 1972), eco- Sedimentary cycles have been recognized throughout the logical-evolutionary units (Boucot 1975), turnover pulses geologic record. In some cases, as in the Milankovitch (Vrba 1985a), ecosystem model (Krasilov 1987), faunistic cycle, their duration and mechanisms seem well under- cycles (Pascual and Jaureguizar 1990), coordinated stasis stood. In other cases, their duration and periodicity is (Brett and Baird 1995), and stepwise climatic change/ questionable. Ross and Ross (1985) described over 50 extinction (e.g., Benson et al. 1984; Prothero 1989). global sedimentary cycles from the Paleozoic that re- Sometimes no specific term was used to describe this present sea level rising and falling (transgression–re- pattern, except for periodicity (e.g., Newell 1952), or the gression) with an average duration of 2 Ma. The variation pattern was implied in recognition of climatic or sedi- is estimated to range from 1.2 to 4 Ma; however, these mentary cycles (e.g., mesothems or megacycles; Busch durations are highly extrapolated, being based on few and West 1987; Kemper 1987). Brett and Baird (1995) radiometric dates. As in other trans/regressive sequences, noted that in the late 1840s, d’Orbigny, the “father of there is a slow rise in sea level followed by a fast drop biostratigraphy,” recognized that genera and species (Ross and Ross 1985). If this pattern is true, then it is changed little within his defined packages of strata. Ob- consistent with the asymmetric temperature profile of servation of evolutionary stasis bounded by abrupt change vadhs, as well as temperature profiles from oxygen iso- due to environmental perturbation was perhaps first rec- tope data (e.g., Stott and Kennett 1990). ognized in a comprehensive manner by Olson’s (1952) In marine stratigraphy, sedimentary cycles of various idea of chronofaunas. A chronofauna was initially defined scales (e.g., synthems, mesothems, cyclothems, and PAC as a “geographically restricted, natural assemblage of sequences) have been recognized across the globe, and as interacting animal populations that has maintained its early as 1888, Suess suggested that a global stratigraphy basic structure over a geologically significant period of could be based on trans/regressive units (Busch and West time” (Olson 1952:181). 1987). Recent workers are increasingly using temporal/ Missing time due to erosion or nondeposition is always climatic models in stratigraphy, and perhaps the most present at various scales, leading some workers to hy- encompassing model is the hierarchal genetic stratigraphy pothesize erroneously that evolution makes large of Busch and West (1987). They defined a hierarchy of “jumps.” Whenever paleontologists study fossiliferous trans/regressive units bounded by transgressive and cli- strata in finer detail, they discover that change may have matic surfaces that can be correlated using lithological occurred more rapidly than usual, but intermediate mor- and ecological data on a regional, and possibly global, phological steps are represented (e.g., Martin 1984). The scale. These trans/regressive units have periodicities on fundamental pattern of stasis bounded by rapid change the order of 225–300 Ma (first order) down to 50–130 caused by extrinsic factors is consistent with hypotheses thousand years (sixth order). Periodicities of third and of climatic cycles. The importance of climatic cycles has fourth order are comparable to vadhs and stouts (Martin been recognized by numerous workers, and many have and Meehan 2002), and these units have synchronous suggested that stratigraphy be based on climatic change unconformities, indicating global factors linked to cli- because it affects biota and sedimentary processes si- matic change (Busch and West 1987). multaneously. On a large scale, disruption of entire community patterns by extrinsic factors implies that bi- otic factors, such as competition are minor, as concluded by Benton (1983). Raup and Boyajian (1988:109) also 10 Stepwise extinction and iterative evolution The two major Cenozoic radiations of forams have a pattern of extinction at a cooling interval, re-evolution of Stepwise extinction reflects stepwise climatic change, and similar ecomorphs, and stepwise change over the long this pattern is widely recognized in the fossil record (e.g., term (Cifelli 1969). This iterative pattern in forams is seen Kohler et al. 1988; McGhee 1988; Holland 1989; Janis et throughout the Paleozoic and Mesozoic (Stanley 1987). al. 1998). Keller (1986) noted five pulses of extinction Cifelli (1969) noted that foram faunas of widely different over approximately 15 Ma throughout the late Eocene– ages were sometimes more alike than those from adjacent early Oligocene, which resulted in two-thirds replacement strata. He also stated that ammonoid evolution was highly of foram species. Keller (1986:274) stated the following: iterative, paralleling foram evolutionary patterns and that convergence of ecomorphs is so high that ammonoids are Paleontological research has made it increasingly clear sometimes classified on stratigraphic grounds because the that both faunal and climatic changes are characterized same ecomorph type cannot be easily distinguished tax- by long periods of stability separated by brief episodes onomically. This synchronous ecomorph iteration in ma- of rapid faunal turnover and climatic fluctuations. rine faunas resembles the terrestrial record. During middle Eocene to early Oligocene each faunal turnover is characterized by replacement of tropical marine faunas and floras by cooler subtropical and Chronofaunas and faunistic cycles temperate elements as observed by [many authors]. Recently, Berger et al. (1981) discussed major faunal Olson (1975) concluded that succeeding communities did turnovers at the Cretaceous–Tertiary and Eocene– not evolve gradually from the previous one, but were Oligocene boundaries and the late Miocene in terms of replaced in part or whole and that most major evolu- major steps in Cenozoic evolution. Keller (1983a, tionary change took place as new communities formed 1983b) studied one of these “steps” at the Eocene– under rapid environmental change. Some described Oligocene boundary and observed that faunal changes mammalian chronofaunas closely correspond to a stout occurred in a series of yet smaller steps related to climatic cycle. For example, Webb (1969) recognized a successively cooler climatic conditions. Such stepwise White River Chronofauna composed of the Chadronian, faunal changes were also observed by Kauffman Orellan, and Whitneyan faunas and a Clarendonian (1984a, 1984b) in late Cretaceous invertebrate faunas Chronofauna as composed of the Valentinian, Clarendo- and he redefined the late Cretaceous mass extinctions nian, Kimballian, and Hemphillian faunas (refer to as “stepwise mass extinctions” occurring over a period Fig. 1). Chronofaunal or stout characters are also present of 1–3 Ma. in Pennsylvanian/Permian coal swamp communities. Di- Michele and Phillips (1995) recognized stasis in eco- Using the interpolated time-scale boundaries of morph structure for millions of years, rapid turnover due Berggren et al. (1995), these Eo/Oligocene turnovers to severe climatic change, and a new community forming occurred at 40.1, 38.3, 35.3, 33.5, and 29.4 Ma. The hy- by ecomorph replacement. pothesis of iterative climatic cycles predicts seven rela- Pascual (1992) stated that South American land tively significant extinction events during this interval, mammal “ages” represent relatively balanced communi- occurring at 41.0, 38.6, 36.1, 33.7, 31.3 and 28.9 Ma. ties during times of stasis and that these communities Predicted values are close to extrapolated ages of these were disrupted during periods of severe climatic change. foram extinctions, except that there was no notable ex- A pattern of sedimentary bundles correlating with epi- tinction observed at about 31 Ma. sodes in evolution has been recognized since the begin- The stepwise biotic and climatic nature across the Eo/ ning of South American mammalian biostratigraphy, and Oligocene is seen in other marine faunas, paleosols, leaf assemblages of major bundles have the ecological struc- assemblages, mammalian faunas, oxygen isotope data, ture of chronofaunas (Pascual 1992). These sedimentary and ice sheet formations (Berggren et al. 1985; Wolfe bundles and extinction horizons correlate with the marine 1985; Ehrmann and Mackensen 1992; Miller et al. 1991; record. In turn, changes in sedimentation and evolution in Zachos et al. 1993; Diester-Haas et al. 1996; Bestland et marine and terrestrial provinces correlate with global al. 1997). Interpolated ages of late Eo/Oligocene mam- climatic changes and regional effects. Pascual described malian faunal turnovers in North America (Prothero extinction periods within these sedimentary bundles, but 1989) are also in close agreement with predicted vadh termed these “internal episodes” because these disconti- boundaries. A paleosol series from deposits of South nuities did not break up the basic continuity of the Dakota is inferred to show climatic steps at 37, 34, 32, chronofauna, which is consistent with vadh extinctions 29.5 Ma (Retallack 1983), which are close to predicted within a stout. Pascual and Jaureguizar (1990) recognized vadh estimates of 36.1, 33.7, 31.3, and 28.9. In addition, four hierarchical faunistic cycles with durations of 2.5– each step is terminated by highly dry and seasonal cli- 25 Ma. mates, in which reduced plant cover presumably led to increased erosion, as described in Stout’s (1978) climatic/ sedimentary cycles. 11 Trilobite biomeres and mammalian faunas, as argued here. Williamson (1981) documented nearly synchronous morphological Biomeres were originally defined as regional biostrati- change followed by relative stasis in several mollusc graphic units bounded by extinctions in the dominant lineages in the Turkana basin. Vrba (1985a, 1985b) elements of one phylum (Palmer 1965). Rapid extinctions demonstrated comparable pulses of change and long-term bound stable, ecological units of benthic trilobite faunas stability in African bovids and hominids. Vrba (1985a) of North America, and replacement is iterative, with each coined the term “turnover pulse,” and her hypothesis new trilobite community radiating from the same immi- agrees with vadh characters. Besides the mammalian ex- grant ecomorph from open water (Palmer 1965; Stitt tinction in Africa occurring at 2.4 Ma, community re- 1975). These extinction boundaries are not diachronous, covery was rapid; antelope niches were filled by radiation nor restricted to benthic trilobites as first suggested. Other and dispersal within 300,000 years (Vrba 1988). Vrba extinctions at biomere boundaries include inarticulate argued that abrupt climatic change breaks down stable brachiopods, conodonts, agnostoid trilobites, and in the plant and mammal communities causing rapid evolu- case of conodonts, the extinction has been discerned to be tionary change and stated the following (1985a:232): global (Hood 1989). Most workers have proposed that these extinctions were the result of an abrupt temperature Speciation does not occur unless forced (initiated) by drop. From one biomere boundary that retained original changes in the physical environment. Similarly, forc- isotopic signatures, the extinction was associated with a ing by the physical environment is required to produce rapid 4–5C cooling (Hood 1989). After analyzing sug- extinctions and most migration events. Thus, most gested causes of these turnovers, Hood (1989) concluded lineage turnover in the history of life has occurred in that abrupt cooling caused extinction of these tropical pulses, nearly (geologically) synchronous across di- communities (and that an anoxic water event associated verse phylogenies, and in synchrony with changes in with this cooling may have been a factor). the physical environment. Biomeres encompass multiple trilobite zones, and the Upper Cambrian represents three biomeres and about This pattern exactly describes the terminal Pleistocene 7 Ma (Sundberg 1996; Harland et al. 1989). Palmer et al. extinction of North America, which occurred over several (1995) reported the zones averaging 2.7 Ma and biomeres thousand years (e.g., Guthrie 1984; FAUNMAP group averaging 7 Ma, but Bowring and Erwin (1998) reported 1996). Cambrian zones averaging 1.5 Ma and biomeres 4 Ma and that the zones are not of equal duration. It is difficult to ascertain which of these absolute ages is more accurate, Suggestions for further work but the pattern of zones and biomeres is consistent with vadhs and stouts, and what is needed is to test for iterative A perennial problem in studying the geologic past is re- evolution of trilobite ecomorphs and to create detailed liability of correlations, particularly on a global scale. temperature curves spanning at least 15 Ma. An abrupt Standard correlation charts have changed significantly on 5C cooling causing trilobite community extinction a decade basis, and workers still disagree on many as- (Hood 1989) is consistent with the end of a vadh cycle, pects. The frequency and duration of missing time in and cooling of this magnitude has been reported in ter- marine strata may be unappreciated or not even tested restrial and marine records at some extinction boundaries (Aubrey 1995). Terrestrial sequences have far less con- (e.g., Vella 1968; Kennett and Shackleton 1976; tinuity than marine sequences, with much less than 50% Hooghiemstra and Ran 1994). of time preserved in most deposits (Clark et al. 1967; Retallack 1984). In addition, many biostratigraphic boundaries correspond to times of erosion. Coordinated stasis and turnover pulse hypothesis Most of these problems could be mitigated by better absolute age control. Unfortunately radiometric dates may Defining subunits of Boucot’s (1975) ecological-evolu- not be directly associated with biostratigraphically inter- tionary units, Brett and Baird (1995) argued that stasis of esting assemblages and are not systematically distributed Appalachian Basin benthic communities in the Middle over the geologic record (Harland 1989; Berggren et al. Paleozoic occurred on the order of 3–7 Ma. These com- 1995). In North American mammalian biostratigraphy, munities went extinct due to major, rapid environmental Evernden and others (1964) provide the only compre- change, and most boundaries were associated with global hensive dating sequence from a single study. This is an climatic events and extinctions. Replacement of com- early study using techniques with comparatively low munities occurred within 100,000–500,000 years and was precision, but even modern dating may show significant followed by long intervals of stasis. This pattern agrees variation among labs. A concerted effort needs to be made with the one predicted by iterative climatic cycles, and its to control for intralab error and extend the scope and time scale is on the order of vadhs and stouts. density of absolute dates, especially for the Cenozoic Brett and Baird (1995) renamed this pattern “coordi- where such a framework could definitively decide if nated stasis” and argued that it is seen in such varied vadhs and stouts are chronostratigraphic. communities as freshwater molluscs, trilobite biomeres, 12 Another correlation difficulty results from time-trans- resolution. For example, the climatic history of the past gressive climatic effects. This is particularly noticeable in 500,000 years as determined by pollen assemblages high latitude marine faunas, which are first and most around a lake in Japan remarkably correlates with tem- severely affected by climatic change (Stanley and Rud- perature records of the Caribbean and Pacific Oceans, diman 1995). Some marine biostratigraphic definitions sedimentary cycles of the Mediterranean, climatic trends have been changed as more has been learned of the time- of Central Europe, and sea-level changes of Japan and transgressive nature of extinctions and first appearances. New Guinea (Fuji 1988). The Eocene warming event as Innovations seem to occur at high latitudes first and expressed by carbon isotopes from various sources ap- progress to lower latitudes as climate cools. Mathematical pears synchronous, as well as the rapid marine and ter- modeling of foram species origination/extinction events restrial biotic turnovers at this time (Koch et al. 1992). indicates that extinctions are more deterministic than first If we assume a vadh duration of 2.4 Ma, the general appearances (Patterson and Fowler 1996), suggesting that temperature trend as described by van der Hammen extinction events define sharper biostratigraphic bound- (1961) shows overall warming to about 1.7 Ma into the aries, as advocated by Martin (1985). vadh, followed by 700,000 years of rapid cooling, with Because the terrestrial mammalian record is mainly greater cooling at the end of a C vadh. Ice volume one of middle latitudes, the time-transgressive nature of changes in a C vadh, the Pleistocene, reflect this pattern. first appearances is not easily observed. Hickey et al. During the last 700,000 years, ice volume was two times (1983) concluded that many vertebrate taxa inhabited greater than the previous 2 Ma (Barendregt and Irving northern Canada 2–4 Ma earlier than they occurred in 1998). Temperature profiles at the Paleo/Eocene and Eo/ middle latitudes and that floral displacement was much Oligocene boundaries show similar magnitude, direction, greater, although this was disputed (Flynn et al. 1984). and duration (cooling over 500,000 years; Stott and Not only are high latitudes affected first, but low latitudes Kennett 1990), and a number of oxygen isotope data sets act as refugia for warm-adapted taxa, sometimes with indicate a 5C drop at extinction boundaries (e.g., Vella lineages persisting much longer after most relatives be- 1968; Kennett and Shackleton 1976; Guilderson et al. come extinct at higher latitudes (e.g., Webb 1989). The 1994). These global temperature changes should result in description of global ungulate distributions (Janis 1989) environmental modification and concordant evolutionary provides an example of latitudinal climatic offlap. Low- change. On a small scale, Chiba (1998) described a pat- seasonally-adapted hindgut fermenters (such as horses tern of rapid, synchronous morphologic changes with and rhinos) accounted for over 50% of the Eocene un- longer periods of stasis in five different snail lineages on gulate fauna and occurred at high latitudes. During the different islands over the past 40,000 years, concluding Oligocene, their high latitude abundance dropped to about that synchrony in convergence and lineage extinction was 25% as climate became cooler. By the mid-Miocene due to climatic change. lower latitude abundances also dropped to 25% as global We need to better demonstrate which events co-occur. temperature decreased (Janis 1989). Martin (1984) demonstrated that species lineages of a A strong argument for high latitudes being a center of sabertooth felid (Megantereon-Smilodon) and muskrat evolution in the Northern Hemisphere is seen in the (Pliopotamys-Ondatra) showed a similar pattern of body sudden, simultaneous appearance of “immigrant” taxa at size change, with an exponential increase during the last middle latitudes of Asia, North America, and/or Europe 700,000 years of the Pleistocene vadh C. He further with no immediate ancestors known (see Woodburne demonstrated the same pattern in a sabertooth nimravid 1987). Recognition of the Holarctic as a center for evo- lineage (Sansanosmilus-Barbourofelis) in the previous lution has been noted since early studies. As stated by vadh C (Kimballian). These three lineages slowly evolved Matthew (1939:7), “...the present distribution of mam- overall larger body sizes in vadhs A and B, and then mals is due chiefly to migration from the great northern towards the end of a vadh C, their body sizes increased land mass, and the connection of this southward march exponentially. This evolutionary change is concordant with progressive refrigeration in the polar regions was with global cooling, including the glacial pulse starting made more than a century ago (1778) by Buffon.” This 700,000 years ago. Mammal body size can be highly climatic offlap sequence of taxa is particularly variable correlated with climate (e.g., Davis 1981; Zeveloff and with respect to immigrants. Climatic offlap has been Boyce 1988; Smith et al. 1995), and the body size in- better recognized in the marine record where high latitude crease in the armadillo-like Holmesina (Hulbert and faunas are better known (e.g., Jenkins 1974; Stanley Morgan 1993) exhibits a similar exponential rate at the 1987). The timing of mammalian dispersals into North end of the Pleistocene, exemplifying the breadth of this America appears to be related to climatic events, with phenomenon. dispersal into lower latitudes representing time-trans- The power of climatic change to influence evolution gressive range extension (Martin 1985, 1994). High lati- may be further shown in humans. The evolution of the tude data need to be extended and refined. genus Homo about 2.4 Ma ago has been attributed to the Also, marine data need to be more closely related with appearance of more open habitat (e.g., Stanley 1995). the terrestrial record. One promising note in correlating This is a period of cooling and increased aridity based on events is that climatic effects as reflected in the marine a wide variety of evidence (and the timing of this climatic and terrestrial records appear synchronous, even at a fine shift is predicted by vadh cycles). Ruff et al. (1997) de- 13 scribed the pattern of brain size evolution in Homo as a was first proposed by van der Hammen (1957) and has the rapid increase from 600,000 to 150,000 years ago, which advantage of generating many predictions. was preceded by stasis on the order of 1.8 Ma. This Van der Hammen proposed that global climatic cycles pattern is concordant with the exponential growth seen in exist at many scales and can be tested with numerous data muskrat and sabertooth body sizes. A remarkable coin- sets. In his 1965 paper, he reported that Jurassic am- cidence if there is no underlying factor. Naples and monoid stratigraphy reflected cycles of 7 Ma composed of Martin (1998) hypothesized that the evolutionary trend three 2.33 Ma cycles, and these in turn could be divided towards larger brain sizes among ungulates, carnivores, into cycles on the order of 0.8 Ma. He also inferred a and primates throughout the Cenozoic is due to increased 70 Ma cycle. In this last paper on his climatic cycle hy- seasonality and habitat openness that resulted from global pothesis, van der Hammen reiterated that climatic cycles cooling. As in the other major mammalian evolutionary define stratigraphy, providing ideal chronostratigraphic trends of the Cenozoic, such as increased body size and units. degree of hypsodonty, brain size increase is predicted to The recognition of climatic cycles in the rock record is be highly correlated with mean global temperature trends. a common theme. For instance, many workers (Clark et Geist (1983) noted that many Ice Age mammalian lin- al. 1967; Wolfe 1978; Collinson et al. 1981; Kemper eages (e.g., moose, mammoth, and Homo) in more sea- 1987; Frakes et al. 1992) have proposed the existence of sonal environments (e.g., alpine and arctic) evolved or- temperature cycles on the order of 10 Ma as reflected in nate, giant members with larger brains and more gener- marine and terrestrial data. Zubakov and Borzenkova alized niches than close relatives of more southern, (1990) reported a climatic cooling rhythm at 3.7 and equable climates. One would predict that these lineages 11 Ma. Kemper (1987) reported climatic cycles of 2.2 and underwent the most rapid evolutionary change during the 9 Ma bounded by abrupt temperature drops throughout vadh cooling intervals from approximately 3.1 to 2.4 Ma most of the Cretaceous marine deposits of the Sverdrup ago and from 700,000 to 11,000 years ago. Analyzing Basin, in northern Canada. These hierarchal cycles may evolutionary rates within a species lineage may provide a be equivalent to vadhs and stouts. simple and powerful tool for testing the existence of these Kemper (1987) suggested that the cycles may be due to climatic cycles. Besides A-B-C dominance turnover pat- orbital parameters, but noted that solar variation cannot be terns (Fig. 2), any robust, long-term data set with a cli- ruled out. Very small-scale solar changes, such as sunspot matic signal, either more direct, such as oxygen isotope cycles, are reflected in the sedimentary/climatic record data, or less direct, such as the extinction pattern of highly (e.g., Wymstra et al. 1984; Kerr 1996). There is a growing specialized ecomorphs, provides a test for these cycles. body of research showing solar forcing as a viable Although vadh C has a more severe terminal cooling, mechanism for Quaternary/Holocene cycles (e.g., Crow- there is not more generic extinction at its end than in ley and Kim 1996; Bond et al. 2001; van Geel et al. vadhs A or B (Fig. 3B). What we do see is that a few 2003), but do large-scale solar mechanics create vadhs, highly specialized ecomorphs characteristically become stouts, and other hierarchal cycles? extinct in a vadh C (e.g., 12 of 15 sabertooths, 2 of 2 cheetah ecomorphs, and 4 of 4 aye-aye ecomorphs in the North American Cenozoic). Extinction may not be forever

There are many uses of the term extinction. In its strictest Iterative climatic cycles—a useful hypothesis? sense, it refers to the termination of a lineage, but it is also used more generally for the end of an adaptive type such As listed above, many hypotheses recognize the funda- as sabertooth “cats.” In this usage it may not be final, and mental pattern of relative stasis bounded by rapid change in the absence of human intervention, we might expect within the fossil/sedimentary record. Though some have sabertooths to re-evolve, as they have done for the past different theoretical bases, they share many features. Is 50 Ma years. When we compare mammal extinction be- there a significant difference among the concepts of tween the two C vadhs of the Pleistocene and latest biomere, faunistic cycle, ecological-evolutionary unit, Miocene (Fig. 5), we see not only ecomorphs re-evolving, and the older idea, chronofauna? The term punctuated- but also the same ecomorphs going extinct, such as in equilibrium (Eldredge and Gould 1972) became widely proboscideans (Amebelodon/Mammut), giraffe–camels used, but Schindewolf (1950), Simpson (1953), and many (Aepycamleus/Titanotylopus), and scimitartooths (Nim- before have stated that evolution occurs at different rates. ravides/Homotherium). Extinction is not only similar in As mentioned previously, the “father of stratigraphy” terms of adaptive types, but also in scope—about one- (d’Orbigny 1849) recognized the “punctuated” evolution third of ecomorph genera become extinct in each vadh and extinction bounding packages of strata. Given the (Fig. 3B). The community structure between these two commonality of these ideas, it is worthy to consider that vadh C assemblages is comparable, except that the they are all related and may be partial recognition of a Pleistocene organisms are adapted to colder, more open single underlying mechanism. Meehan and Martin (2003) environments, reflecting the general Cenozoic climatic suggested that these hypotheses might be encompassed by trend. a hypothesis of iterative climatic cycles. This hypothesis 14 Fig. 5 Convergent assem- blages. Representative mam- mals that became extinct at the end of two vadh C climatic cy- cles (Kimballian and Irvingto- nian) in North America. These intervals are 7 Ma apart. Modi- fied from Martin 1985: Fig. 4

Evolutionary patterns are shared in many cases by dents, and with them, form one of the first rodent bur- ecomorphs, which start from ecologically similar ances- rowing communities. At the end of the earliest Miocene, tors and progress through similar adaptive stages, usually these dry land beavers become extinct along with gopher- showing similar change rates. These changes are not like geomyoids. A new fossorial rodent community constant, being slower in the A and B vadh cycles and dominated by aplodontoids and another group of gopher- then rapidly increasing in rate in the last third of the C like geomyoids replaced this first burrowing community. cycle (see Martin 1984; Meehan and Martin 2003), cor- Some version of this community persisted until a similar, relating to the temperature cycles of van der Hammen modern community of burrowing squirrels and pocket (1961). gophers replaced it. The diversity of burrowing rodents The community extinction and replacement pattern for each community was similar, but their origins were also implies that the fossil record contains many examples different. We might have expected rodents that first oc- of rare taxa that through subsequent diversification filled cupied burrowing niches would have continued in them to the empty adaptive zones of their ecological predecessors. the present day. This was not the case. The modern North In North America at middle latitudes, open habitat first American fossorial rodent community is composed became widespread in the late Oligocene, and rodents largely of rodents that came into this niche in the last took advantage of this new adaptive zone. Beavers 7 Ma. (Castoridae) are rare animals with little diversity in the Convergent/parallel characters are so prevalent among late Eocene–early Oligocene, but in the late Oligocene– ecomorphs in these iterative communities that an unini- early Miocene, evolve many fossorial forms (Fig. 2B). tiated observer would have difficulty in telling them apart, They are accompanied by a radiation of geomyoid ro- despite that they are not phylogenetically close. Many Table 2 Cenozoic North American land mammal “ages” correlated to vadh climatic cycles Epoch NALMAsa NALMA subdivisionsa Vadh climatic cyclesb Vadh type Basal datesa Basal datesc Predicted vadh ages Pleistocene Rancholabrean Irvingtonian Rancholabrean (Rlb) and Irvingtonian (Irvi1–3) = Irvingtonian (Irvi) C 1.8 1.8 2.41 Pliocene Blancan Blancan (Bl) = Blancan (Blan) B 4.5 4.5 4.82 Miocene Hemphillian L Hemphillian (Hh3) = Hemphillian (Hemp) A 6.0 6.0 7.23 E Hemphillian (Hh1–2) = Kimballian (Kimb) C 8.9e 8.8 9.64 Clarendonian E and L Clarendonian (Cl1–2) = Clarendonian (Clar) B 12.5e 11.0 12.05 Barstovian L Barstovian (Ba2–3) = Valentinian (Vale) A 14.5 12.5 14.46 E Barstovian (Ba1) = Barstovian (Bars) C 16.8f 15.8 16.87 Hemingfordian L Hemingfordian (He2) = Sheepcreekian (Shcr) B 17.9 17.5 19.28 –g E Hemingfordian and L L Arikareean (Ar4-He1) = Marslandian (Mars) A 21.0 19.2 21.69 E L Arikareean (Ar3) = Harrisonian (Harr) C 23.0 23.0 24.10 Oligocene Arikareean L E Arikareean (Ar2) = Monroecreekian (Mocr) B 25.0 27.7 26.51 E E Arikareean (Ar1) = Geringian (Geri) A 29 29.4 28.92 Whitneyan Whitneyan (Wh) = Whitneyan (Whit) C 30.7 31.9 31.33 Orellan Orellan (Or) = Orellan (Orel) B 32.4 33.4 33.74 Chadronian Chadronian (Ch1–3) = Chadronian (Chad) A 38.0 37.1 36.15 Eocene Duchesnean (Late) Duchesneand (Du) = Lapointian (Lapt) C – – 38.56 (Early) Duchesneand (Du) = Pearsonian (Pear) B 42.0 39.5 40.97 Uintan L Uintan (Ui2) = Mytonian (Myto) A 44.6 41.3 43.38 E Uintan (Ui1) = Wagonhoundian (Wghd) C 48.0 45.9 45.79 Bridgerian Gardnerbuttean, Blacksforkian, and Twinbuttean (Br1–3) = Bridgerian (Brid) B 51.0 50.4 48.20 Lostcabinian (L Wasatchian; Wa4) = Lostcabinian (Lcab) A 52.0 53.5 50.61 Wasatchian Lysitean (M Wasat.; Wa3) = Lysitean (Lysi) C 54.5 54.2 53.02 Sandcouleean and Graybullian (E Wasatchian; Wa1-Wa2) = Graybullian (Gray) B 57.3 55.5 55.43 Paleocene Clarkforkian Clarkforkian (Cf1–3) = Clarkforkian (Clfk) A 58.8 56.0 57.84 Tiffanian Tiffanian (Ti1–6) = Tiffanian (Tiff) C 62.7 60.9 60.25 Torrejonian Torrejonian (To1–3) = Torrejonian (Torr) B 65.0 63.8 62.66 Puercan Puercan (Pu1–3) = Puercan (Puer) A 66.4 65.1 65.07 a NALMAs are from Woodburne (1987: Fig. 10.1). b Correlated vadhs from Martin (1985) and Martin and Meehan (2002); only minor changes in the importance of previously recognized biostratigraphic boundaries were made to define these climatic cycles inferred from mammalian faunas. c Data from Janis et al. (1998). d Late and early Duchesnean units were not recognized in Woodburne (1987), but see Lucas (1992). e Amended extrapolated boundary ages from Whistler and Burbank (1992). f Amended extrapolated age from Lindsay (1995). g Basal Marslandian = uppermost Arikareean (Martin 1985). E Early; M Middle; L Late. The base of the Puercan correlates to the base of Paleocene and the top of the Irvingtonian to the top of the Pleistocene, but recent correlations (Prothero 1995; Woodburne and Swisher 1995) place remaining epoch boundaries slightly offset to indicated mammalian ages by several hundred thousand years or less. Predicted ages are based on a vadh duration of 2.41 Ma (Fig. 4). 15 16 such examples can be demonstrated. The most striking sequence can only be taken as evidence that a controlling example is the repetitive evolution of dirktooths (Martin environmental factor (climate) is fluctuating from one set 1985; Fig. 1B), but the whole catlike community is in- of parameters to another and then back again in a geo- volved, with scimitartooth and leopard ecomorphs be- logically short period. It is less easy to demonstrate that coming extinct and re-evolving in the same dominance this pattern is truly periodic, as has been implied for many turnover pattern (Fig. 2C). This pattern can also be seen in of the intervals where it is recognized. Unfortunately, the dog and hippo ecomorphs (Meehan and Martin 2003), as paucity of radiometric dates bearing on this question well as in dominant artiodactyl herbivore ecomorphs presently prevents a resolution. It should be remembered (Martin 1985) and shrew ecomorphs (Fig. 2A, D). The that the various dating methods often vary amongst starting point for these radiations is usually a combination themselves when dating the same sample and that samples of survivors (of the previous extinction event) and im- usually date volcanic events rather than biostratigraphic migrants, as Olson (1975) described in chronofaunal boundaries so that radiometric ages are younger or older succession. The resemblance between ecomorphs is than the event they are supposed to date. Even with these greatest among taxa at the end of evolutionary sequences caveats, the subcycle/cycle system of triplets first pro- rather than at their beginnings. For instance, Barbouro- posed by van der Hammen (1957) and extended by Martin felis fricki, a lion-sized carnivore, is very similar to its and Meehan (2002) into vadhs and stouts, predicts lion-sized ecomorph Smilodon fatalis, but less so to the boundaries that agree with the various time scales for the Smilodon ancestor Megantereon that is closest to Bar- North American terrestrial Cenozoic as well as these time bourofelis in time. We would infer from this pattern of scales agree with each other (Table 2). Their agreement succession that competition was an unlikely contributor to with the marine epochal boundaries is even more striking these extinction events; instead we find that ecological (Table 1). In general vadhs and stouts are comparable to replacement is a fundamental pattern of the fossil record. the durations of marine zones and stages. If in fact this Numerous lines of evidence indicate that environ- periodicity is genuine, it severely limits the possible mental parameters fluctuate globally throughout Earth causal mechanisms. A truly regular mechanism is likely history. In many cases these fluctuations reflect changes to be astronomical, and this may include varying solar in global heat and water balance. Extinctions have often mechanics and heat output. If the time intervals are more been associated with declining temperatures and in- variable, they may reflect a cycling feedback system in- creasing aridity. Increased temperatures and moisture are volving some factor such as atmospheric carbon dioxide. reflected in increased species diversity and community In any case, climatic cycles appear to form the ultimate complexity. Although this pattern has been observed basis of biostratigraphy and have a controlling effect on globally through most of the geologic record, its details the overall pattern that evolution has taken. may be best observed in the well-preserved and well- studied North American terrestrial Cenozoic. The most Acknowledgements We wish to thank A. Seilacher, M. Dawson, characteristic part of this pattern is the concordant ex- and two anonymous reviewers for editorial comments. T.J.M. wishes to thank his PhD committee at Kansas University: L.D. tinction of a variety of organisms followed by the reap- Martin, D. Miao, R.W. Wilson, R.M. Timm, P. Wells, and D.W. pearance of the same ecomorphs in subsequent intervals. Frayer. This phenomenon is not ambiguous, extends across community structure, and is largely independent of phy- logenetic relationships. Only climate can simultaneously References control evolutionary patterns in so many diverse and often phylogenetically distant organisms. This interpretation is Alroy J (2000) New methods for quantifying macroevolutionary further supported by the consistency of evolutionary re- patterns and processes. Paleobiology 26:707–733 Aubrey M-P (1995) From chronology to stratigraphy: interpreting sponses across continents to what must be a global trend the lower and middle Eocene stratigraphic record in the At- in environmental change. If ungulates show adaptations lantic Ocean. In: Berggren WA, Kent DV, Aubry M-P, Hard- consistent with cooler, more open habitat, so will rodents. enbol J (eds) Geochronology, time scales and global strati- It is the common climatic signature that makes recog- graphic correlation. SEPM Spec Publ 54:213–274 nizing global epoch and era boundaries among such dis- Barendregt RW, Irving E (1998) Changes in the extent of North American ice sheets during the late Cenozoic. Can J Earth Sci parate groups as molluscs, pollen, and mammals possible. 35:504–509 In most cases, extinction is not random, but tends to be Benson RH, Chapman RE, Deck LT (1984) Paleoceanographic of similar scope and affects similar ecomorph types in events and deep-sea ostracodes. Science 224:1334–1336 each episode. For this to be true, these ecomorphs would Benton MJ (1983) Large-scale replacements in the history of life. Nature 302:16–17 have to re-evolve time after time. That this is true is one Benton MJ (1995) Diversity and extinction in the history of life. of the most remarkable facts that we can ascertain from Science 268:52–58 the fossil record. It may be easy to think of some reason, Berggren WA, Kent DV, van Couvering JA (1985) The Neogene; real or imagined, that could result in the extinction of a part 2, Neogene geochronology and chronostratigraphy. Mem Geol Soc Lond 10:221–260 particular lineage, but it is more difficult to understand Berggren WA, Kent DV, Aubry M-P, Hardenbol J, eds (1995) how following an extinction, the same ecomorph could Geochronology, time scales and global stratigraphic correla- re-evolve almost immediately—often from an unrelated tion. SEPM Spec Publ 54:1–364 lineage. That this happens many times in the geological 17 Bestland EA, Retallack GJ, Swisher CC III (1997) Stepwise climate Evernden JF, Savage DE, Curtis GH, James GT (1964) K-Ar dates change recorded in Eocene–Oligocene paleosol sequences from and Cenozoic mammalian chronology of North America. Am J Central Oregon. J Geol 105:153–172 Sci 262:145–198 Bond G, Kromer B, Beer J, Muscheler R, Evans MN, Showers W, FAUNMAP group (1996) Spatial response of mammals to late Hoffmann S, Lotti-Bond R, Hajdas I, Bonani G (2001) Per- Quaternary environmental fluctuations. Science 272:1601– sistent solar influence on North Atlantic climate during the 1606 Holocene. Science 294:2130–2136 Fischer AG (1981) Climatic oscillations in the biosphere. In: Boucot AJ (1975) Evolution and extinction rate controls. Elsevier, Nitecki MH (ed) Biotic crises in ecological and evolutionary Amsterdam time. Academic, New York, pp 103–131 Bowring SA, Erwin DH (1998) A new look at evolutionary rates in Flessa KW (1975) Area, continental drift, and mammalian diver- deep time: uniting paleontology and high-precision sity. Paleobiology 1:189–194 geochronology. GSA Today 8:1–8 Flynn LJ, MacFadden BM, McKenna MC (1984) Land-Mammal Brett CE, Baird GC (1995) Coordinated stasis and evolutionary Ages, faunal heterochrony, and temporal resolution in Cenozoic ecology of Silurian to Middle Devonian faunas in the Ap- terrestrial sequences. J Geol 92:687–705 palachian Basin. In: Erwin DH, Anstey RL (eds) New ap- Frakes LA, Francis JE, Syktus JI (1992) Climate modes of the proaches to speciation in the fossil record. Columbia University Phanerozoic. Cambridge University Press, Cambridge Press, New York, pp 285–315 Fuji N (1988) Palaeovegetation and palaeoclimate changes around Busch RM, West RR (1987) Hierarchal genetic stratigraphy: a Lake Biwa, Japan during the last ca. 3 million years. Quat Sci framework for paleoceanography. Paleoceanography 2:141– Rev 7:21–28 164 Geel B van, van der Plicht J, Renssen H (2003) Major delta-14C Chiba S (1998) Synchronized evolution in lineages of land snails in excursions during the Late Glacial and early Holocene: changes oceanic islands. Paleobiology 24:99–108 in ocean ventilation or solar forcing of climate change? Quat Int Cifelli R (1969) Radiation of the Cenozoic planktonic For- 105:71–76 aminifera. Syst Zool 18:154–168 Geist V (1983) On the evolution of Ice Age mammals and its Clark DL, Whitman RR, Morgan KA, Mackey SD (1980) Stratig- significance to an understanding of speciations. Walton Dis- raphy and glacial-marine sediments of the Amerasian Basin, tinguished Lecture Series, Mountain Lake Biology Station, central Arctic Ocean. Geol Soc Am Spec Pap 181:1–57 30:109–33 Clark J, Beerbower JR, Keitzke KK (1967) Oligocene sedimenta- Gradstein FM, Ogg JG (2004) Geologic time scale—why, how, and tion, stratigraphy, paleoecology, and paleoclimatology in the where next! http://www.stratigraphy.org/scale04.pdf Big Badlands of South Dakota. Fieldiana Geol Mem 5:1–158 Grant PR, Grant BR (2002) Unpredictable evolution in a 30-year Collinson ME, Fowler K, Boulter MC (1981) Floristic changes study of Darwin’s finches. Science 296:707–711 indicate a cooling climate in the Eocene of southern England. Guilderson TP, Fairbanks RG, Rubenstone JL (1994) Tropical Nature 291:315–317 temperature variations since 20,000 years ago: modulating in- Crowley TJ, Kim KY (1996) Comparison of proxy records of cli- terhemispheric climate change. Science 263:663–665 mate change and solar forcing. Geophys Res Lett 23:359–362 Guthrie RD (1984) Mosaics, allelochemics, and nutrients: an eco- Crowley TJ, North GR (1991) Paleoclimatology. Oxford Monogr logical theory of late Pleistocene megafaunal extinctions. In: Geol Geophys 16:1–339 Martin PS, Klein RG (eds) Quaternary extinctions. University Curry WB, Cullen JL, Backman J (1990) Carbonate accumulation of Arizona Press, Tucson, pp 259–298 in the Indian Ocean during the Pliocene: evidence for a change Hammen T van der (1957) Climatic periodicity and evolution of in productivity and preservation at about 2.4 Ma. Proc Ocean South American Maastrichtian and Tertiary floras. Bogota Bol Drilling Prog Sci Results 115:509–515 Geol 5:49–91 Dalrymple GB (1979) Critical tables for conversion of K-Ar ages Hammen T van der (1961) Upper Cretaceous and Tertiary climatic from old to new constants. Geology 7:558–560 periodicities and their causes. NY Acad Sci Annu 95:440–448 Damuth JD, Jablonski D, Harris JA, Potts R, Stucky RK, Sues H-D, Hammen T van der (1965) Paloklima, Stratigraphie und Evolu- Weishampel DB (1992) Taxon-free characterization of animal tion. Geol Rundsch 54:428–441 communities. In: Behrensmeyer AK, Damuth JD, DiMichele Hammen T van der, Wijmstra TA, Zagwijn W (1971) The floral WA, Potts R, Sues H-D, Wing SL (eds) Terrestrial ecosystems record of the late Cenozoic of Europe. In: Turekian KK (ed) through time. University of Chicago Press, Chicago, pp 183– The late Cenozoic glacial ages. Yale University Press, New 203 Haven, Conn., pp 391–424 Davis SJM (1981) The effects of temperature change and domes- Harland W, Armstrong BR, Cox A, Craig L, Smith A, Smith D tication on the body size of Late Pleistocene to Holocene (1989) A geologic time scale. Cambridge University Press, mammals of Israel. Paleobiology 7:101–114 Cambridge DeMenocal P, Bloemendal J, King J (1991) A rock-magnetic re- Hickey LJ, West RM, Dawson MR, Choi DK (1983) Arctic ter- cord of monsoonal dust deposition to the Arabian Sea: evidence restrial biota: paleomagnetic evidence of age disparity with for a shift in the mode of deposition at 2.4 Ma. Proc Ocean mid-northern latitudes during the late Cretaceous and early Drilling Prog Sci Results 117:389–401 Tertiary. Science 221:1153–1156 Diester-Haas L, Robert C, Chamley H (1996) The Eocene–Oligo- Holland CH (1989) Synchronology, taxonomy, and reality. Philos cene preglacial transition in the Atlantic sector of the Southern Trans R Soc Lond B 325:263–277 Ocean (ODP Site 690). Mar Geol 131:123–149 Hood KC (1989) Geological significance of biomere boundaries. DiMichele WA, Phillips TL (1995) The response of hierarchically Dissertation,Kansas University, Lawrence structured ecosystems to long-term climatic change: a case Hooghiemstra H, Ran ETH (1994) Late Pliocene–Pleistocene high study using tropical peat swamps of Pennsylvanian age. Studies resolution pollen sequence of Colombia: an overview of cli- in heophysics: effects of past global change on life. National matic change. Quat Int 21:63–80 Academy Press, Washington, pp 134–155 Hulbert CH Jr, Morgan GS (1993) Quantitative and qualitative d’Orbigny AD (1849) Cours elamentaire de paleontologie et de evolution in the giant armadillo Holmesina (Edentata: Pam- geologie stratigraphiques, vol I. Masson, Paris patheriidae) in Florida. In: Martin RA, Barnosky AD (eds) Ehrmann WU, Mackensen A (1992) Sedimentological evidence for Morphological change in Quaternary mammals of North the formation of an east Antarctic ice sheet in Eocene/Oligo- America. Cambridge University Press, Cambridge, pp 134–204 cene time. Palaeogeogr Palaeoclimatol Palaeoecol 93:85–112 Janis CM (1989) A climatic explanation for patterns of evolution- Eldredge N, Gould SJ (1972) Punctuated-equilibria: an alternative ary diversity in ungulate mammals. Palaeontology 32:463–481 to phyletic gradualism. In: Schopf TJM (ed) Models in paleo- Janis CM, Scott KM, Jacobs LL (eds) (1998) Evolution of Tertiary biology. Freeman and Cooper, San Francisco, pp 82–115 mammals of North America, vol I. Terrestrial carnivores, un- 18 gulates, and ungulatelike mammals. Cambridge University Newell ND (1952) Periodicity in invertebrate evolution. J Paleontol Press, Cambridge 26:371–385 Jansen E, Bleil U, Henrich R, Kringstad L, Slettemark B (1988) Olson EC (1952) The evolution of a Permian vertebrate chrono- Paleoenvironmental changes in the Norwegian Sea and the fauna. Evolution 6:181–196 northeast Atlantic during the last 2.8 my: deep-sea drilling Olson EC (1975) Permo-Carboniferous paleoecology and mor- project/ocean drilling program sites 610, 642, 643, and 644. photypic series. Am Zool 15:371–389 Paleoceanography 3:563–581 Palmer AR (1965) Biomere—a new kind of biostratigraphic unit. Jenkins DG (1974) Paleogene planktonic foraminifera of New J Paleontol 39:149–153 Zealand and the Austral region. J Foraminiferal Res 4:155–170 Palmer AR, Stitt JH, Thomas RC (1995) Biomere boundaries. Keller G (1986) Stepwise mass extinctions and impact events: late Northwest Geol 24:67–75 Eocene to early Oligocene. Marine Micropaleont 10:267–293 Pascual R (1992) Episodes in South American land mammal evo- Kemper E (1987) Das Klima der Kreide-Zeit. Geol Jahrb Reihe A, lution and sedimentation: discrete responses to the same causes. Hannover, Germany IGCP Second Meeting Programs and Abstracts, Paleoweath- Kennett JP, Barker PF (1990) Latest Cretaceous to Cenozoic cli- ering Records and Paleosurfaces, October 1992, p 21 mate and oceanographic developments in the Weddell Sea, Pascual R, Jaureguizar EO (1990) Evolving climates and mammal Antarctica: an ocean drilling perspective. Proc Ocean Drilling faunas in Cenozoic South America. J Hum Evol 19:23–60 Prog Sci Results 113:937–960 Patterson RT, Fowler AD (1996) Evidence of self-organization in Kennett JP, Shackleton NJ (1976) Oxygen isotope evidence for the planktic foraminiferal evolution: implications for intercon- development of the psychrosphere 38 Myr ago. Nature nectedness of paleoecosystems. Geology 24:215–218 260:513–515 Prothero DR (1989) Stepwise extinctions and climatic decline Kerr RA (1996) A new dawn for sun-climate links? Science during the later Eocene and Oligocene. In: Donovan SK (ed) 271:1360–1361 Mass extinctions, processes, and evidence. Columbia Univer- Koch PL, Zachos JC, Gingerich PD (1992) Correlation between sity Press, New York, pp 217–234 isotope records in marine and continental carbon reservoirs near Prothero DR (1995) Geochronology and magnetostratigraphy of the Paleocene/Eocene boundary. Nature 358:319–322 Paleogene North American land mammal “ages:” an update. In: Koenigswald W von, Hahn J (1981) Jagdtiere und Jger der Eiszeit. Berggren WA, Kent DV, Aubry M-P, Hardenbol J (eds) Theiss, Stuttgart Geochronology, time scales and global stratigraphic correla- Kohler M, Moya-Sola S, Agusti J (1988) Miocene/Pliocene shift: tion. SEPM Spec Publ 54:305–315 one step or several? Nature 393:126 Prothero DR (1999) Does climatic change drive mammalian evo- Krasilov V (1974) Causal biostratigraphy. Lethaia 7:173–179 lution? GSA Today 9:1–5 Krasilov V (1987) Periodicity in the development of life. Paleontol Raup DM, Boyajian GE (1988) Patterns of generic extinction in the J 3:9–15 fossil record. Paleobiology 14:109–125 Leidelmeyer P (1966) The Paleocene and lower Eocene pollen flora Retallack GJ (1983) A paleopedological approach to the interpre- of Guyana. Leidse Geol Meded 38:49–70 tation of terrestrial sedimentary rocks: the mid-Tertiary fossil Lenski RE, Ofria C, Pennock RT, Adami C (2003) The evolu- soils of Badlands, National Park, South Dakota. Geol Soc Am tionary origin of complex features. Nature 423:139–144 Bull 94:823–840 Lindsay EH (1995) Copemys and the Barstovian/Hemingfordian Retallack GJ (1984) Completeness of the fossil record: some esti- boundary. J Vert Paleontol 15:357–365 mates using fossil soils. Paleobiology 10:59–78 Liu TS, An Z, Yuan B, Han J (1985) The loess–paleosol sequence Ross CA, Ross JRP (1985) Late Paleozoic depositional sequences in China and climatic history. Episodes 8:21–28 are synchronous and worldwide. Geology 13:194–197 Lucas SG (1992) Redefinition of the Duchesnean land mammal Ruff CB, Trinkaus E, Holliday TW (1997) Body mass and en- “age,” late Eocene of western North America. In: Prothero DR, cephalization in Pleistocene Homo. Nature 387:173–176 Berggren WA (eds) Eocene–Oligocene climatic and biotic Schaal S, Ziegler W (1988) Messel–Ein Schaufenster in die Ge- evolution. Princeton University Press, Princeton, N.J., pp 88– schichte der Erde und des Lebens. Verlag Waldemar Kramer, 105 Frankfurt am Main Martin LD (1984) Phyletic trends and evolutionary rates. Spec Publ Schindewolf OH (1950) Grundfragen der Palontologie. E Carnegie Mus Nat Hist 8:526–538 Schweizerbart’sche Verlagsbuchhandlung, Erwin Ngele, Martin LD (1985) Tertiary extinction cycles and the Pliocene– Stuttgart Pleistocene boundary. Inst Tertiary–Quaternary Stud Symp Ser Schultz CB, Stout TM (1980) Ancient soils and climatic changes in 1:33–40 the Central Great Plains. Trans Nebr Acad Sci 8:187–205 Martin LD (1994) Cenozoic climatic history from a biological Shackleton NJ, Blackman J, Zimmerman H, Kent DV, Hall MA, perspective. Inst Tertiary–Quaternary Stud Symp Ser 2:39–56 Roberts DG, Schnitker D, et al (1984) Oxygen isotope cali- Martin LD, Meehan TJ (2002) Recognizing a global stratigraphy. bration of the onset of ice-rafting and history of glaciation in Inst Tertiary–Quaternary Stud Symp Ser 3:175–185 the North Atlantic region. Nature 307:620–623 Martin LD, Naples VL (2002) Environmental controls for multiple Simpson GG (1953) The meaning of evolution. Yale University evolutionary events. Inst Tertiary–Quaternary Stud Symp Ser Press, New Haven, Conn. 3:37–52 Smith FA, Betancourt JL, Brown JH (1995) Evolution of body size Matthew WD (1939) Climate and evolution. Spec Publ NY Acad in the woodrat over the past 25,000 years of climate change. Sci 1:1–223 Science 270:2012–2014 McGhee GR Jr (1988) The Late Devonian extinction event: evi- Stanley SM (1986) Anatomy of a regional mass extinction: Plio– dence for abrupt ecosystem collapse. Paleobiology 14:250–257 Pleistocene decimation of the western Atlantic bivalve fauna. Meehan TJ, Martin LD (2003) Extinction and re-evolution of Palaios 1:17–36 similar adaptive types (ecomorphs) in Cenozoic North Ameri- Stanley SM (1987) Extinction. Freeman, New York can ungulates and carnivores reflect van der Hammen’s cycles. Stanley SM (1995) Climatic forcing and the origin of the human Naturwissenschaften 90:131–135 genus. In: Kennett J, Stanley S (eds) Studies in geophysics: Miller KG, Wright JD, Fairbanks RG (1991) Unlocking the ice effects of past global change on life. National Academic Press, house: Oligocene–Miocene oxygen isotopes, eustasy, and Washington, D.C., pp 233–243 margin erosion. J Geophys Res 96:6829–6848 Stanley SM, Ruddiman WF (1995) Neogene ice age in the North Naples VL, Martin LD (1998) Cenozoic brain evolution as a result Atlantic region: climate changes, biotic effects, and forcing of global cooling. Institute of Tertiary–Quaternary Studies factors. In: DiMichele WA, Phillips TL, Stanley SM, Knoll AH, Symposium Abstracts, Annual Meeting, Lawrence, Kansas, p 4 Kennett JP (eds) Effects of past global change on life. National Academic Press, Washington, D.C., pp 118–133 19 Stitt JH (1975) Adaptive radiation, trilobite paleoecology, and ex- age (late Miocene) in California. Bull Geol Soc Am 104:644– tinction, Ptychaspid Biomere, Late Cambrian of Oklahoma. 658 Fossils Strata 4:381–390 Williams EE (1972) The origin of faunas. Evolution of lizard Stott LD, Kennett JP (1990) The paleoceanographic and paleocli- congeners in a complex island fauna: a trial analysis. Evol Biol matic signature of the Cretaceous/Paleogene boundary in the 6:47–89 Antarctic: stable isotopic results from ODP Leg 113. Proc Williamson P (1981) Paleontological documentation of speciation Ocean Drilling Prog Sci Results 113:829–846 in Cenozoic molluscs from the Turkana Basin. Nature 293:437– Stout TM (1978) The comparative method in stratigraphy: the be- 443 ginning and end of an ice age. Trans Nebr Acad Sci 6:1–18 Wolfe JA (1978) A paleobotanical interpretation of Tertiary cli- Stucky RK (1995) Problems and speculation in mammalian pale- mates in the Northern Hemisphere. Am Sci 66:694–703 oecology and evolution: Western North America during the Wolfe JA (1985) Distribution of major vegetational types during Paleogene. Fossils of Arizona, vol III, Proc 1995 SW Paleontol the Tertiary. Geophys Monogr 32:357–375 Soc Mesa SW Museum, Arizona, pp 97–103 Wolfe JA, Hopkins DM (1967) Climatic changes recorded by Sundberg FA (1996) Morphological diversification of Ptychopari- Tertiary land floras in northwestern North America. In: Hatai K ida (Trilobita) from the Marjumiid biomere (Middle and Upper (ed) Tertiary correlation and climatic changes in the Pacific. Cambrian). Paleobiology 22:49–65 11th Pacific Sci Congr Symp, Tokyo 25:67–76 Van Valen LM (1971) Adaptive zones and the orders of mammals. Woodburne MO (ed) (1987) Cenozoic mammals of North America: Evolution 25:420–428 geochronology and biostratigraphy. University of California Vella P (1968) Inferred temperature fluctuations at the beginning of Press, Berkeley the Taranaki Epoch (Upper Miocene). Tuatara 16:45–49 Woodburne MO, Swisher CC III (1995) Land mammal high-reso- Vrba ES (1985a) Environment and evolution: alternative causes of lution geochronology, intercontinental overland dispersals, sea- the temporal distribution of evolutionary events. S Afr J Sci level, climate and vicariance. In: Berggren WA, Kent DV, 81:229–236 Aubry M-P, Hardenbol J (eds) Geochronology, time scales and Vrba ES (1985b) African Bovidae: evolutionary events since the global stratigraphic correlation. SEPM Spec Publ 54:335–364 Miocene. S Afr J Sci 81:263–266 Wymstra TA, Hoekstra S, Vries BJ de, van der Hammen T (1984) Vrba ES (1988) Late Pliocene climatic events and hominid evo- A preliminary study of periodicity in percentage curves dated lution. In: Grine FE (ed) Evolutionary history of the “robust” by pollen density. Acta Bot Neerl 33:547–557 australopithecines. Aldine de Gruyter, New York, pp 405–426 Zachos JC, Lohmann KC, Walker JCG, Wise SW (1993) Abrupt Webb SD (1969) Extinction–origination equilibria in late Cenozoic climate change and transient climates during the Paleogene: a land mammals of North America. Evolution 23:688–702 marine perspective. J Geol 101:191–213 Webb SD (1989) The fourth dimension in North American terres- Zeveloff SI, Boyce MS (1988) Body size patterns in North trial mammalian communities. In: Morris DW, Abramsky Z, American mammal faunas. In: Boyce MS (ed) Evolution of life Fox BJ, Willig MR (eds) Patterns in the structure of mam- histories of mammals. Yale University Press, New Haven, pp malian communities. Texas Tech University, Lubbock, Spec 123–146 Publ Mus 28:181–203 Zubakov VA, Borzenkova II (1990) Global palaeoclimate of the Whistler DP, Burbank DW (1992) Miocene biostratigraphy and late Cenozoic. Dev Palaeontol Strat 12:1–456 biochronology of the Dove Spring Formation, Mojave Desert, California and characterization of the Clarendonian mammal