Frozen Fauna of the Mammoth Steppe: The Story of Blue Babe
Page 26
The last wild horses in Europe, the tarpan (Equus ferus), lived on these same eastern European steppes, away from intense agricultural development of central Europe, until they too were eventually displaced by human use of the land within historic times. The wild horse in Mongolia (Equus przewalskii) is closely related to the domestic horse and interbreeds freely with it, despite a difference in one chromosome. The habitat of the Mongolian horse was high arid steppes, where it has not been seen for over a decade. Horses closely related to these ranged into the northernmost parts of Siberia and eastern Beringia in the late Pleistocene. Likewise, hemionids (Equus hemionus), adapted to the high dry steppes, from Iran to Mongolia, were also present in northernmost areas of Siberia and eastern Beringia. The bones of Equus, both horses and hemionids, are frequent fossils throughout all Beringian habitats where bones are preserved, both uplands and lowlands. This seems to be the case for all times during the radiocarbon range within the Pleistocene. All living members of the genus Equus are specialized grazers, all occur in plains or open environments, and their Pleistocene bones are found in these same environments throughout the Holarctic.
Unlike bison and horse, mammoths have no close living analogue, at least not within the same genus. We do know that the main food of the two living species of proboscidians is grass (Laws, Parker, and Johnstone 1975; Olivier 1982). Judging from mammoth morphology, the woolly mammoth seems to have been a more grass-dependent proboscidian than either species of living elephant (Maglio 1973), but we need not rely solely on analogy and morphological features for a dietary reconstruction of these three species. Mummies of the Beringian bison, horse, woolly rhino, and woolly mammoth do exist, and we can see what these individuals ate before they died.
Several caveats must be mentioned before attempting to reconstruct past diets from gut contents. These are cautionary notes that come from having performed and seen performed over two thousand mammalian necropsies, including gut analysis. First, when animals take large mouthfuls of small-sized items they invariably incorporate secondary material along with their target food. Whatever is adjacent is likely to be ingested. A grizzly bear inadvertently eats many leaves and twigs along with berries in the fall, and grazers ingest a variety of plants in one bite. We cannot assume that the presence of a plant species as a minority dietary item in stomach contents means it was actively selected.
Second, animals choose from a variety of items ranging up and down a scale of preference, depending on what is available. Under starvation conditions they resort to the most unpalatable, uncharacteristic part of their diet. At the bottom of the scale are foods so toxic, fiberous, or low in nutrients that they cannot be digested. Ironically, most large herbivores that die of starvation are found with bulging-full rumens. This puzzled biologists until they realized the animals did not die from simple lack of food, but were reduced to eating undigestible food which clogged their digestive tract.
Third, there are different rates of decomposition in the gut; woody plants with an undigestible lignatious skeleton remain in the gut longer than leaves and can thus be overrepresented in a sample of gut contents. Also, animals often shift from their staple, representative diet for a short season to take advantage of transient resources. Most northern large ungulates are adapted to very specific winter diets, but their summer resources are more variable, with a higher interspecific overlap. Season of death thus becomes important in appraising gut contents. These warnings do not mean we cannot study gut contents to understand diet, but only that we must approach with caution.
As discussed in the first chapter of this book, large mammal mummies from the Soviet Union have been found with identifiable gut contents. At least two frozen mammoth mummies have voluminous quantities of identifiable plants in their stomachs. Both deaths occurred during autumn, in an interstadial period (Boutellier Interval, the same as Blue Babe). There is also a horse mummy, the Selerikan horse from Siberia, dating in the same time range. Vereshchagin and Baryshnikov (1982) and Hopkins et al. (1982) in their syntheses have explained the concentration of mummies in this interstade by the greater quantities of silt moving downslope. We have to assume that the Boutellier Interval interstade was wetter in summer (most but not all of the carcasses are autumn deaths) than the Duvanny Yar (glacial), but not as wet as today, presumably with a more incomplete ground cover and exposed silt. Greater quantities of silt are needed to cover a large carcass than to cover individual bones. Of course this is only a statistical bias, as some frozen mummies and bones with soft tissue attached come from other times, even from peak glacials.
The Beresovka mammoth stomach contained mostly grass by volume (Ukraintseva 1981; Tikhomirov 1958) (additionally, samples taken from the stomach contents contained mostly grass pollen). There were, in addition, small percentages of a wide variety of other plants, including a variety of forb macrofossils. The other almost complete frozen carcass, the Shandrin mammoth, from the U.S.S.R. (Ukraintseva 1981), had gut contents that consisted of 90% grasses and sedges. In addition, there were twig tips of some woody plants—willow (Salix), larch (Larix), birch (Betula), and alder (Alnus). With the above evidence and the distinguishing morphology of mammoths, Olivier (1982) followed every other zoologist who has studied mammoths and concluded that the animals were grazers. Megaherbivore species such as mammoths, using the coarsest end of the grazing spectrum, make use of a poor-quality resource, one relatively undefended by antiherbivory compounds but deficient in nutrients and minerals (Guthrie 1982, 1984a; Olivier 1982). Agenbroad (1984), studying the plant remains in dung of late Pleistocene mammoths from the Great Plains, found grass and a supplement of minor herb and arboreal components. Elephants, like horses and bison, use poorer quality fibrous resources as a dietary staple and add supplemental dicots to complete their diets. Woody species in the frozen mammoth stomachs are twig tips, the very growth stage according to Bryant and Kuropat (1980) highest in nutrients. These poorly defended twig tips are high in the nutrients grasses lack (Olivier 1982). However, to avoid poisoning, the animal must limit consumption of any one species of these dicots so as not to exceed its ability to detoxify the secondary plant compounds (Bryant et al. 1985). That is especially true for cecal digesters (like horses and mammoths) which do not have rumen to assist in detoxification (Guthrie 1984b).
We can see this same pattern, using grasses as a dietary staple with supplemental dicots, in the stomachs of frozen horse and bison mummies. A Pleistocene horse, found at Selerikan in the Indigirka River basin and dating about 37,000 years before present, had a stomach and intestines containing 90% herbaceous material, of which Festuca grasses predominated, along with the sedge, Kobresia. The latter was identifiable to species, Kobresia capilliformis, by seeds (Ukraintseva 1981). This xeric sedge species, Ukraintseva remarks, is not present in the Indigirka area today, but it is a typical plant in the high dry mountains of central Asia, the Middle East, and Mongolia. She concluded that it had been abundant in the area where the horse died. Macrofossils in the gastrointestinal tract included small quantities of willow (Salix), dwarf birch (Betula nana) twigs, and very small quantities of moss (Ukraintseva 1981). This diet is reasonable for a northern grazer, especially given the interstadial vegetation available.
Stomach contents of a woolly rhino, Coelodonta, mummy from Yukutia (Churapachi) also contained mostly grass (Vereshchagin and Baryshnikov 1984). Woolly rhinos have the most complex teeth of any rhino, even more so than the African white rhino, Ceratotherium, which is a strict grazer; the Siberian and Polish woolly rhino mummies have a wide front lip for grazing, like those of African white rhinos.
Only one Pleistocene bison, located on the Krestovka River in the Kolyma Basin, has been found in Siberia with stomach contents. No percentages have been described in detail regarding volume of different plant groups, other than to say that graminoids predominated. Among vascular plants, Kobresia macrofossils, Poaceae, Cyperaceae, and Ericaceae were identified from the stomach (Ukraintseva 1981). For this same bison Korobkov and F
ilin (1982) looked at the pollen and spore content of the gut contents. In their largest sample the Graminae predominated with 77.7%, Artemisia with 18.3%, and Cruciferae with 0.9%; in another sample Graminae was 84% and Artemisia 10.9%. In another sample Artemisia predominated, and in another Chenopodiaceae was the most abundant. This information strongly resembles a diet typical of extant bison on a mixed grassland habitat (Peden et al. 1974).
Colinvaux and West (1984) proposed that some animals on the Mammoth Steppe were “itinerant” and ventured above the tree line of the boreal forest only for brief episodes. But this model of a sparse Beringian tundra, lacking enough grass to support a grazing proboscidian, led Colinvaux and West to transform the woolly mammoth into a “browser,” despite the fact that mammoth gut contents clearly show that woolly mammoths were grazing specialists.
Dentition of the bison, horse, and mammoth also provides clear evidence that these species, particularly the Beringian representatives, were specialized grazers. In special situations, when these species are not eating siliceous grasses, it is quite apparent from their teeth. Paleontologists (e.g., Simpson 1951) have recognized there is a high correlation among ungulates between the development of hypsodont, high-crowned teeth with elaborate crown patterns and a grassy diet. Grasses, especially when used for winter range, are very low in energy and nutrients, high in fiber, and quite siliceous because opaline phytoliths are embedded in the leafy tissue. These characteristics mean that grass must be thoroughly chewed to reduce particle size and facilitate digestion of fibers. A grazer actually chews a greater volume of this low-quality food than would a comparable browser. Abrasion from this longer mastication time and larger volume of food causes the teeth of grazers to wear at a faster rate. Mammals invading the grazing niche have responded to increased tooth wear by an evolutionary increase in the height of the tooth and by elaborating the crown pattern in tooth enamel. These changes increase the chewing surface, making chewing more efficient and allowing the tooth to last longer (McNaughton et al., 1985.)
Woolly mammoths in Beringia had the most complex teeth of any proboscidian past or present (fig. 9.13) and have been judged to be the most extreme form of grazing specialist (Maglio 1973; Olivier 1982). In another study based on cementum root annuli (Guthrie, unpub.), I showed that mammoths lived about as long as extant elephants and that they wore out teeth at a similar succession rate, despite much more complex tooth crowns and more hypsodont molars. Likewise, among bovids, bison teeth are some of the most complex; among rhinos, woolly rhinos had by far the most complex dental patterns. Horses and hemionids have the most complex teeth of any fossil equid; this is especially true in Beringia where they have an unusually elongated protoloph.
Tooth wear changes when grazers use a less abrasive dietary staple. An example discussed earlier showed slower tooth wear among bison that eat more sedges than grasses, and the crown surface assumes a vaulted steeple character (Haynes 1984). As we have already seen, teeth of the Beringian bison wore at a rapid rate, indicative of an abrasive grass diet. Again, this is in contradiction to the bison diet portrayed by Ritchie (1984).
The evolution of complex-crowned hypsodont teeth among grazers is associated with the evolution of silica defenses in grasses (McNaughton et al., 1985). There are two divergent pathways in the evolution of plant defenses. Dicots have depended on elaborate secondary chemical defenses, and almost all pharmacologically active plants are part of that radiation. These dicotyledonary plants have their region of active growth on the terminal meristems, unlike graminoids which grow from the base and thus keep active growth tissue near the surface of the ground, protected from large herbivores. Grasses have a depauperate secondary chemistry: instead of producing poisons, they defend leafy photosynthetic tissue with a relatively cheap soil constituent—silicon—in the form of angular opaline crystals. This material is harder than tooth enamel and is the main determinant of tooth wear (Baker, Jones, and Wardrop 1959). Fiber content of different plants seems unrelated to tooth wear rates (Barnicot 1957). A woody-stem dicot eater such as a moose has very polished tooth crowns, but a rather slow rate of wear in its low-crowned teeth.
Fig. 9.13. Molar complexity. Large grazers of the Mammoth Steppe had very complex molars, adapted for chewing sparse winter forage into smaller, digestible fractions. Overall surface area of the crowns are similar; increased trituration is achieved by a more complex enamel pattern, similar to switching the grating surface of a food processor. (Left) Coelodonta, the woolly rhino, has the most complex enamel pattern among rhinos. The second most complex is the African white rhino, Ceratotherium, also a grazer. And on the bottom is the African black rhino, Diceros, a browser. (Right) the top molar is from a woolly mammoth (Mammuthus primigenius) of the Mammoth Steppe. The middle molar is from an Asiatic elephant, Elaphas, which is similar to the southern mammoth, M. columbianus. On the bottom is a molar from the African elephant, Loxodonta, a species that consumes much browse during the dry season.
McNaughton et al. (1985) have stated, with regard to grass phytoliths, that no other plant trait putatively ascribed a protective role against herbivory due to coevolution has simultaneously such well-documented detrimental effects on herbivores, strong support from the fossil record, and evidence of natural selection in contemporary communities. Teeth of mammoths, horses, and bison in the Beringian Pleistocene allow us to state with as much confidence as one can ever obtain from the fossil record that a large component of the diet of these three species was graminoids, mainly grasses.
Small Feet, Firm Substrates, and Steppe Mammals
Another mammalian character available to help us evaluate paleoenvironmental conditions is the form and shape of the hoof. Animals respond evolutionarily to different conditions by developing larger or smaller feet, adjusting foot-loading characteristics to the softness of the substrate. Foot characteristics are one geographic difference between subspecies of extant ungulates, and there are chronological variations in foot shape among related fossils (Eisenmann 1984). Today the boggy lowland tundras of Beringia are not easily negotiable in the summer by small hoofed ungulates. Horses in a packtrain frequently get stuck (Guthrie, pers. obs.).
Plains mammals such as saigas, horses, and bison use speed to outdistance predators. But running over moist muskeg or tundra is simply not possible unless an ungulate has special foot and leg morphology similar to the caribou, musk-oxen, and moose, which is well adapted to negotiating wet landscapes. Neither horses, bison, saigas, nor mammoths have such morphology. Although it is difficult to compare leg morphology (such as the ability of a moose foot to be withdrawn vertically and the ability of the phalanges to fold along a narrow central axis), we can easily compare the loading on feet for a rough estimate of the requirements of summer substrate. Fortunately an easy comparison is possible because most of these foot-loading differences have been reviewed and published (Kuz’mina 1977; Telfer and Kelsall 1984).
Caribou are extremely well adapted to boggy substrate, with loadings of 140–80 (g/cm2), compared to the heavy loadings of horses (625–830), saigas (600–800), elephants (510–660), and, very worst of all, bison (1,000–1,300). Medium-sized mammals found on lowland tundras today include wolves (89–114), wolverines (20–35), and arctic foxes (40–60). For an intuitive sense of what this means, I have a foot loading of about 200 g/cm2 when barefoot. These are figures for walking animals; when these animals run, foot loading is increased considerably.
These figures do not tell the whole story, however, because other adaptations can affect ease of foot withdrawal and walking on boggy ground, especially for caribou and moose. For example, moose (420–560 g/cm2) and musk-oxen (325–400) both have large dew-claw hooves and hooves with the ability to spread apart and form a broad surface. This is, of course, unavailable to horses and mammoths, which have rather fixed foot surfaces.
Pleistocene horses and bison in Beringia did not have larger hooves than their counterparts now living on firm substrates farther south. We know this from well-pres
erved mummies, as well as from comparisons of distal foot bones. Woolly mammoth feet, for example, are not larger than the feet of living elephants.
Ritchie (1984) and Colinvaux and West (1984) have argued that instead of being a firm-substrate grassland, Beringia was simply an expansion of dry tundra, more like a polar desert. Low Arctic tundra beyond the latitudinal margin of the trees, and even most upland tundra, is quite mesic, composed of either wet sedge meadows, tussock-producing sedges, mosses, or cushion plants. Farther north, in the high Arctic, bogs become fewer mainly because of a rocky substrate, a condition not characteristic of the loess-covered Beringia. Colinvaux (1984) argues that much of Beringia was unsuitable for bison, horse, and mammoth and that it was only along the rivers that one could find these grazing mammals. But in interior Alaska, more large mammal fossils are found in the uplands. A map of fossil localities of large mammals in Alaska or Siberia shows that bones are found everywhere conditions are suitable for fossil preservation, from valley bottoms to well above the present tree-line.
The Mammoth Steppe: A Rangeland That Produced Giants
The Beringian tundra as portrayed by Colinvaux and West (1984) and Ritchie (1984) is marginal habitat with almost no plant biomass to support a community of large mammals; further, they suggest that the few large mammals present were beyond their range of optimum adaptation. Animals on marginal summer range (growing season) can be quickly identified by their morphology; these animals are small bodied, and parts of their body not directly related to survival, such as horns and antlers, are poorly developed (Geist 1971b). Do the fossil remains of Beringian large mammals exhibit such signs of marginal summer habitat? The answer is mixed—no and yes. Ruminant and caecalid fossils show opposite trends.