Paleosols as trace fossils of ecosystems

Australopithecus afarensis is known from body fossils, such as the partial skeleton "Lucy" (Johanson et al. 1982), as well as from trace fossils, such as the footprints of Laetoli (Leakey and Harris 1987). The soils of A. afarensis also are known, especially at the "first family'' site near Hadar, Ethiopia (Radosevich et al. 1992). Here a troop of at least 13 individuals, young and old, died, rotted, and were partially disarticulated, before being interred in flood deposits on a crumb-structured soil of grassy streamside woodland (O Figure 13.7). The paleosol is not only a matrix to the bones but a trace fossil of their ecosystem. Furthermore, paleosols by definition are in the very place they formed, not redeposited. Unlike the skeleton of "Lucy'' found in the sandstone of a former river channel (Johanson et al. 1982), and thus transported some distance from its natural habitat, the first family was found where it died and had lived (Radosevich et al. 1992). Thus, paleosols give a finer resolution of primate paleoenvironments in time (O Figure 13.6) as well as space (O Figure 13.7).

O Figure 13.7

Reconstruction of paleosols at the "First Family site'' for Australopithecus afarensis at Hadar [data from Radosevich et al. (1992)]

O Figure 13.7

Reconstruction of paleosols at the "First Family site'' for Australopithecus afarensis at Hadar [data from Radosevich et al. (1992)]

Trace Hunter Ward

The various paleosols containing Miocene ape fossils in southwest Kenya can also be used to constrain their habitats (O Figure 13.8). The fragmentary and weathered nature of most of these fossils is evidence that they accumulated through natural processes of death and decay on the paleosols in which they are found (Pickford 1986a). The great diversity of fossil apes in this region (Gommery et al. 2002; Harrison 2002; Ward and Duren 2002) is in contrast to the low diversity of great apes today (Fleagle 1998), leading to the idea that Miocene apes, defined from apelike dentition, were ecologically more like both

O Figure 13.8

Paleosols of Miocene apes from southwestern Kenya [data from Retallack (1991a) with taxonomy after Harrison (2002), Retallack et al. (2002), Ward and Duren (2002)]

Nyanzapithecus

apes and monkeys today (Andrews 1996). Analysis of their occurrence in paleosols shows that there was some ecological separation of different species to different soil types, but still high diversity within a soil type (O Figure 13.8). In the 20 Ma sites of Koru and Songhor, for example, one taxon (Ugandapithecus) shows little habitat specificity through a variety of tropical dry forest habitats, but small taxa (Kalepithecus, Mabokopithecus) are in upland soils and larger taxa (Proconsul, Rangwapithecus, Dendropithecus, Nyanzapithecus) remained in lowland forests closer to water. One paleosol type (Kiewo pedotype) has as many as six taxa: three likely suspensory feeders (Limnopithecus, Dendropithecus, Ugandapithecus from small to large) and three likely overbranch feeders (Nyanzapithecus, Proconsul, Rangwapithecus, from small to large). The contrasting sizes and other differences between these taxa suggest niche partitioning of forest canopy tiers.

Diverse sympatric catarrhine communities persisted into the dry woodland landscapes of Rusinga Island at 17.8 Ma, when paleosols with the crumb peds and iron-manganese nodules of dambo grasslands (Yom pedotype) appear, but are rare and barren of primate fossils (Retallack et al. 1995). Other evidence for grasslands of about the same age are abundant bunch grasses at the Ugandan fossil site of Bukwa (Pickford 2002b). Yom paleosols of dambo grassland are much more common by 14.7 Ma on Maboko Island (Retallack et al. 2002), where they contain abundant vervet-like monkeys (Victoriapithecus: note change of scale for this exceptional collection in O Figure 13.8). These seasonally inundated grasslands of dry climates were not encouraging to fossil apes, which were more common in riparian woodlands (Nyanzapithecus, Limnopithecus, Mabokopithecus and Simiolus of Dhero paleosols). More wide ranging was Kenyapithecus, found in both riparian woodland (Dhero) and nyika shrubland (Ratong), which it exploited more effectively than other apes because of its thick enameled, large molars useful for tough foods (Martin 1985) and its macaque-like limbs and feet (McCrossin et al. 1998). A similar pattern of wide ranging Kenyapithecus and forest-dependent other apes (Oreopithecus, Simiolus, Proconsul) persisted in grassland mosaics of Fort Ternan and Kapsibor at 14.4 Ma, when well-drained short-grass, wooded grassland was widespread. The appearance of grasslands so encouraging for victoriapithecine ancestors of vervets and colobines, was not so encouraging to apes, which remained rare components of the fossil fauna.

Reconstruction of rainfall from paleosols implies also vegetation belts (O Figure 13.6a), by comparison with Holocene climatic ranges of plant formations (Anhuf et al. 1999). There was rainforest in central Africa during the past 20 million years as indicated by rare finds of fossil plants (Bancroft 1932,1933), but evidence of rain forest has not yet been found in the East African areas of hominoid fossils (Retallack 1991a, Jacobs 2002). These interpretations (O Figure 13.6a) are well in accord with indications of vegetation from paleosol classification, profile form and root traces (Retallack 1991a), as evidence that the climatic range of most vegetation types did not change over the past 20 million years.

An exception is the evolution of grasslands, which expanded their climatic range to displace extinct kinds of woodlands (O Figures 13.6 and O 13.9). There is not yet any East African evidence of grasslands before 17.8 Ma, when crumb-textured, brown, simple (A-Bk) profiles of dambo were rare at Rusinga Island (Retallack et al. 1995) and bunchgrasses grew luxuriantly at Bukwa (Pickford 2002b). Well-drained, short-grass, sod-grasslands were widespread by 14.4 Ma (Retallack 1991a, Retallack et al. 2002) and well-drained, tall-grass, sod-grasslands expanded their climatic range considerably by 7 Ma (Wynn 2004a, b). Grasslands were a newly coevolved ecosystem of the Cenozoic, with grasses uniquely suited to grazing by virtue of their intercalary meristems, modular growth, basal tillering, and sod formation, and grazers uniquely suited to coarse grassy fodder by virtue of their wide muzzles, hypsodont teeth, and hard hooves (Retallack 2001a). A world without grasslands was transformed over some 20 million years to a Plio-Pleistocene world with grassland covering at least a quarter of the land surface. Holocene humans spread grassy agroecosystems to almost all parts of the world (Retallack 2001a). Neogene expansion of grasslands within the paleoclimatic belt roughly defined by the 300-750 mm per annum isohyet enabled grasslands to capture the planetary modal rainfall belt and most fertile soils, with consequences for global change including a significant contribution to global cooling (Retallack 2001b).

Before the expansion of the grasslands, an extinct woody vegetation occupied their climatic range (O Figure 13.6). I call these extinct dry woodlands pori (O Table 13.2), from a Hadza word for bush (Woodburn 1968). A good example of a pori ecosystem is the Tek paleosol of Rusinga Island (O Figures 13.3 and O 13.4), which has yielded fossil primates and other mammals, snails, and plants (Pickford 1995; Retallack et al. 1995). Other examples of pori ecosystems include Tut, Choka, and Kwar pedotypes of Songhor and Koru dated at 20 Ma (Retallack 1991a). From the soil perspective, these paleosols have no clear modern analog because they are red and clayey, with large root traces and blocky structure like woodland soils, yet have shallow calcareous horizons like those found in modern African semiarid to subhumid grassland soils. Modern African soils with such shallow carbonate have very different crumb structure, fine root traces, and dark brown organic-rich surface horizons from abundant grasses.

From the paleoanthropological perspective, these ancient communities have no modern analogs because they have so many fossil hominoids, as many as six species in the Kwar pedotype (O Figure 13.8). No community has so many species of hominoids today. Nor do modern hominoids live in such dry climates. Mt. Assirik in Senegal with 956 (854-1224) mm mean annual precipitation is the driest climate with chimpanzees (Kappelman 1993), although Kingdon (2003) gives anecdotes of chimpanzees in wooded grassland. It is now clear that Miocene apes filled a variety of niches like those today filled by vervets, baboons, and colobines as well as apes (Retallack et al. 2002). Pori ecosystems such as the Kwar paleosol at Koru (20 Ma) also show peculiar associations of other mammals, including a mix of dry climate taxa, such as mole rats (Bathyergoides), with wet climate taxa such as flying squirrels (Paranomalurus), giant elephant shrews (Miorynchocyon clarki), tenrecs (Protenrec tricuspis), golden moles (Pro-chrysochloris miocaenicus), and chevrotains (Dorcatherium songhorense. Retallack 1991a). Similarly the Tut and Choka paleosols at Songhor (20 Ma) and Tek paleosols on Rusinga Island (17.8 Ma) have wet climate flying squirrels and tenrecs as well as dry climate mole rats and spring hares (Retallack 1991a; Retallack et al. 1995). These nonanalog combinations of fossil mammals can be explained by a theory of evolutionary replacement of pori with grassland within semiarid to subhumid regions. Before the advent of grasslands, woody vegetation became smaller in stature and biomass from wet to dry regions. This continuum was disrupted as grasslands evolved to usurp the climatic range of pori woodland. Grasslands expanded their range to create a biogeographic divide between Nyika shrubland and miombo woodland (O Figure 13.9).

Fossil primates of East Africa not only coped with changing mixes of animals but with changing climate and vegetation (O Figure 13.6). Wynn (2004b) has introduced the concept of evolutionary entropy to explain effects of climate and vegetation change on hominoid diversity. Climatically dry episodes encouraged grassland mosaic environments with a more varied landscape of open grassland and local woodland, and thus greater landscape disorder or negentropy. Wet episodes of forest vegetation presented more uniform landscapes of higher entropy. My compilation of hominoid diversity, originations, and extinctions (O Table 13.1, O Figure 13.10) supports the view that dry episodes correspond with diverse primates, whereas wet episodes lead to extinctions, particularly of specialized arid-adapted taxa. The concept of ecosystem entropy in hominoid evolution is similar in some respects to Vrba's (1999) "turnover pulse hypothesis,'' but ecosystem entropy presents diffuse and long-term selection pressures, rather than episodic crises or "turnover pulses.'' Recent compilations of mammalian

O Figure 13.9

A scenario for stepwise evolution of East African grasslands with modern precipitation tolerances of African mammals related to those found fossilized in paleosols (Tut, Choka, Kwar) of pori woodlands that preceded the expansion of grasslands. The advent of grasslands disrupted formerly overlapping ranges of apes, bush babies, flying squirrels, mole rats and spring hares. Climatic ranges of modern mammals are from Kingdon (1971, 1974a, b, 1979) and of paleosols from Retallack (1991a); Retallack et al. (1995)

Fruits Miombo Ecosystems

O Table 13.2

Comparison of extinct pori woodland with extant East African vegetation

O Table 13.2

Comparison of extinct pori woodland with extant East African vegetation

Feature

Pori

Miombo

Nyika

Savanna

Vegetation

Dry woodland

Dry woodland

Dry bushland

Wooded grassland

Key genera

Celtis

Brachystegia

Acacia

Combretum

Floral origins

Zambezian

Zambezian

Zambezian

Eurasian

Spinosity

Unarmed

Unarmed

Spinose

Spinose

Leaf set

Semideciduous

Deciduous

Deciduous

Deciduous

Fruit size

Large

Large

Small

Small

Snails

Cerastua

Limicolaria

Achatina

Pupoides

Snail origins

Somalian

Somalian

Somalian

Somalian

Mammals

Apes, rodents

Antelope

Antelope

Antelope

Ungulates

Walangania

Aepyceros

Tragelaphus

Connochaetes

Primates

Proconsul

Cercopithecus

Papio

Papio

Mammal origin

Zambezian

Zambezian

Zambezian

Eurasian

Fire frequency

Low

High

High

High

Soil organics

Low

Low

Low

High

Soil fertility

High

Low

Low

High

Soil type

Alfisol

Oxisol, Vertisol

Aridisol

Mollisol, Vertisol

Parent material

Volcanic

Granitic

Granitic

Volcanic

data from the Turkana region do not show such crises (Bobe et al. 2002), revealing instead an oscillating diversity compatible with less synchronized selection by ecosystem entropy.

A major caveat for such theories is the generally inferior fossil record of climatic wet phases because their soils and sediments are noncalcareous and so not favorable to the preservation of bone (Retallack 1998). We still have no primate fossil record from paleoclimatic wet phases of the early Miocene, but there are discoveries of wet climate human ancestors from 13 Ma (Hill et al. 2002), 6 Ma (Brunet et al. 2002; Galik et al. 2004) and 4-3 Ma (Carroll 2003). The soil-taphonomic bias against wet climate fossils makes the search difficult, not impossible (Peterhans 1993).

Each fluctuation in climate and vegetation presented new crises and opportunities to primates. My own correlation of climatic events with critical adaptations (O Table 13.1) is only an outline of a new research agenda, to be fleshed out with further studies of the critical intervals. The late Miocene paleosols and primate fossils of the Tugen Hills, for example, remain very poorly known compared with those of the Lake Victoria and Turkana basins. Nevertheless, there are general themes apparent from this compilation. We did not evolve from apes in one seminal event, but by a protracted process of growth and pruning of our evolutionary tree. Some specialized features such as procumbent incisors at 18 Ma evolved in dry grassy woodlands, but did not survive succeeding forest expansions (McCrossin and Benefit 1997). Some specialized features such as long arms

O Figure 13.10

Mean annual precipitation and hominoid diversity, extinctions and originations in East Africa over the past 20 million years. The paleoprecipitation curve is from O Figure 13.6. Hominoid data is from Pickford (1986b, 1987), Harrison (2002), Ward and Duren (2002), and Carroll (2003)

O Figure 13.10

Mean annual precipitation and hominoid diversity, extinctions and originations in East Africa over the past 20 million years. The paleoprecipitation curve is from O Figure 13.6. Hominoid data is from Pickford (1986b, 1987), Harrison (2002), Ward and Duren (2002), and Carroll (2003)

by 20 Ma for suspensory locomotion in forests did not persist through succeeding grassland expansions (Harrison 2002). Other forest adaptations such as a short-stiff back by 16 Ma (Pickford et al. 1999), erect stance by 6 Ma (Senut et al. 2001; Galik et al 2004), and flat face by 3.5 Ma (Leakey et al. 2001) proved advantageous in the long term, just as did grassland adaptations, such as thick enamel by 18 Ma (Martin 1985; McCrossin and Benefit 1997), adducted hallux by 14.7 Ma (McCrossin et al. 1998), and long legs for endurance running by 1.8 Ma (Bramble and Lieberman 2004). Although each of these ideas could be debated individually, the general concept of human evolution as a generalist path through a gauntlet of environmental challenges (Potts 1996) is increasingly supported by a burgeoning fossil record (Carroll 2003). There will always be a need for dating and finding more human ancestor fossils, but paleosols now provide new evidence of evolutionary selection pressures with high temporal and spatial resolution.

Past hypotheses of a Miocene pluvial, lake and rain forest (Kent 1944; Leakey 1952; Andrews and Van Couvering 1976; Andrews 1996) and late Miocene grassland (Cerling 1992; Cerling et al. 1997a, b) find, as already reviewed, a counterpart in long-standing theories linking late Miocene evolution of human upright stance or large brains with hunting prowess (Darwin 1872), vigilance against predators (Dart 1926), manipulation of small seeds (Jolly 1970), minimization of sun exposure (Wheeler 1984), long-distance walking (Rodman and

McHenry 1980) or running (Bramble and Lieberman 2004), squat feeding on the ground (Kingdon 2003), or moving between scattered fruiting bushes (Sanford 2003). Forest explanations of upright stance allowing erect-back climbing (Tuttle 1974), hands free to care for premature infants (Lovejoy 1981), phallic display to females (Tanner 1981), or intimidation displays to rivals (Jablonski and Chaplin 1993) move the event back into the "Miocene rain forest'' (of Andrews and Van Couvering 1975; Andrews 1996), for which there is little evidence at hominoid sites in East Africa (O Figure 13.6a). All these views can be reassessed in light of the improved record of East African paleosols, which suggests that there were many alternating habitats in East Africa, not just one seminal environmental shift. Darwin's (1872) idea that erect stance was linked to tool use and brain expansion has been out of favor since the discovery of "Lucy," when it became clear that erect stance preceded tool use and brain expansion by millions of years (Johanson et al. 1982). Erect stance now appears to have occurred in wooded habitats by 6 Ma (Pickford and Senut 2001; Vignaud et al. 2002), perhaps selected by the use of hands in nest provisioning (Lovejoy 1981). We are a mosaic of a complex evolutionary history and no longer need settle for simple or single allegories of human evolution.

Was this article helpful?

0 0

Post a comment