the climate changes were large, and simulations of past climates by climate models have reproduced many of the patterns of change in the data (COHMAP, 1988; Wright et al., 1993).
Mapping studies of the fossil pollen data at regional and local scales also show variations in composition, location, and extent of vegetation. At these scales, variations in elevation and soil type become important along with climate in shaping the development of the vegetation (Davis et al., 1980; Webb et al., 1983; Ritchie, 1987; Gaudreau et al., 1989; Woods and Davis, 1989; Jackson and Whitehead, 1991). Studies at each spatial scale show the independent behavior of individual taxa and the resultant changes in vegetational composition. Many of the past pollen assemblages have no analogues among modern pollen samples (Baker et al., 1989; Overpeck et al., 1992). Each of these types of vegetational change has occurred during the switch from full glacial to interglacial climates, and is likely to have occurred each time that such shifts in climate occurred in Earth history (see Figure 7.2 in Stanley and Ruddiman, Chapter 7, this volume). During the past 700,000 yr, the major changes have occurred seven times, and the estimated total change in the global mean temperature is 5° ± 1°C (Webb, 1991). During times with lesser degrees of global climate change, the changes in vegetation were also less dramatic but probably still involved significant changes in community composition that produced assemblages without modern analogues. As Figure 7.2 in Stanley and Ruddiman (Chapter 7, this volume) shows, the climate has been changing continuously for millions of years; therefore, the vegetation is a continuously changing set of variables chasing a continuously changing set of other variables, namely, climate (Webb and Bartlein, 1992).
Webb (1986) and Prentice et al. (1991) have interpreted these compositional changes and consequent no-analogue assemblages as resulting primarily from the different climatic response of each taxon to the changing mixture of climatic variables as climates have varied temporally. They argue for the taxa being in dynamic equilibrium with climate (Prentice, 1986; Webb, 1986). Studies matching observed pollen maps with those simulated from climate model output provide support for this interpretation (Webb et al., 1987; COHMAP, 1988). Other researchers argue for disequilibrium conditions between plant taxa and climate. They have given major emphasis to the role of biotic factors, such as differing dispersal rates and time lags for populations growth, when interpreting the development of no-analogue assemblages and patterns of species migration (Bennett, 1985; Birks, 1986). Recognition is now developing that a hierarchy of factors is operating, and that the importance of different factors varies with time and space scale (Davis, 1991). Biotic factors are most evident over short time and small spatial scales, and climatic impact is most evident over long time and large spatial scales.
No matter which interpretation is favored (equilibrium or disequilibrium), the pollen record shows major changes in plant assemblages at all spatial scales with major plant assemblages (i.e., formations) having an average life time of ca. 10,000 yr in response to orbitally driven climate change. Consideration of the record of climate forcing for the past 2.8 million years (m.y.) (Figure 7.2 in Stanley and Ruddiman, Chapter 7, this volume) reveals that this forcing has been long-term, large, and continuous (Webb and Bartlein, 1992). The net result has been a continuously changing ecological theater for the evolutionary play (Hutchinson, 1965), and individualistic behavior has produced a continuously changing role, setting, and cast of associated characters for each taxon. Despite all this environmental and ecological change, most species have survived. Evidence from the fossil record suggests that the average longevity of species is 1 to 10 m.y. (Stanley, 1985). One reason for the longevity may be the relatively high frequency of mixing (induced by changes in species abundance, distribution, and association) that prevents long-term isolation of genetically distinct populations (Coope, 1978; Webb, 1987; Bartlein and Prentice, 1989).
Gould (1985) and Bennett (1990) discuss how "progress in life's history" may be thwarted, and selection over 10,000 years or less ("ecological time") is erased or lost by longer-term processes. In a well-argued paper, Bennett (1990) identifies orbitally forced climate change, which occurs at time scales of 20,000 to 100,000 yr, as the key longer-term process. As stated by Bartlein and Prentice (1989), "The paleoecological record of the past 20,000 years demonstrates that orbitally induced climatic changes produce changes in the distribution of organisms, leading to the quasi-cyclical alternation between allopatry and sympatry, commonness and rarity, continuous distribution and fragmentation." Furthermore, if recognition is given to the orbital control of long-term variations in monsoonal climates that depends on land-sea contrast (Kutzbach, 1981; COHMAP, 1988; Kutzbach and Webb, 1991), then the orbital pulse to climate can be seen to be possible in the absence of large ice sheets and to have a long history throughout the geological record. Crowley et al. (1986) and Ruddiman and Kutzbach (1989) have explored the potential implications of known tectonic changes on this mechanism for climate change, and Barnosky (1984), Olsen (1986), and others (see Berger et al., 1984) have documented orbitally driven climate changes during several intervals of the Phanerozoic. The long-term occurrence of