The most common opportunity for preservation is a basic or calcarious environment. For this reason, limestone caves are our most important source for fossil bones. Without limestone caves the Pleistocene record of mammals would be meager. Caves buffer seasonal fluctuations in temperature and moisture, and their floor sediments are seldom acidic. Thus the caves of South Africa, caves and rock shelters of southern France, and limestone sinkhole caves of Florida are important localities for finding Pleistocene mammals. Many caves actively carry water underground. During dry episodes when water pressure decreases, the roofs of these caves sometimes collapse, forming a sinkhole, a natural trap for anything that might step in the hole.
To create most fossils, a bone must be covered so that aerobic decomposers cannot feed on the organic parts. Alkaline ponds and stream oxbows provide such conditions. Bones submerged in the mud of a pond or lake edge are likely to be preserved. Seasonal rains of the Rift Valley have given us a wealth of information about African Plio-Pleistocene mammals in this way.
Fig. 2.21. Few animals become fossils. This elephant died a natural death away from depositional environments out in the open African savanna. Most easily edible parts have been removed by scavengers. Only the thick, dried hide and bones remain. The ears, feet, and tip of trunk are thin-skinned and were the first to be eaten. Eventually, what remains will be scattered by hyenas, but it is unlikely that any portion will become fossilized. This is true for most large mammals.
Other situations, such as oil seeps, also preserve fossils. The loss of volatile short-chain molecules from oil seeps concentrates sticky, long-molecule tars and creates shallow entrapment pools. The La Brea tar pits in California are a famous example. And, as mentioned in chapter 1, a peculiar combination of tar seeps associated with salt domes occurs in Starunia where woolly rhinos were trapped and preserved.
Most conditions that produce fossils preserve, at best, only bones. In very arid regions soft body parts occasionally become so dehydrated that decomposers cannot live and multiply; shriveled mummies are the result. Preservation of Pleistocene fossils in parts of the unglaciated far north is exquisite, but we are only beginning to understand why so many fossils were preserved in Beringia. The cartoons showing woolly mammoths frozen in clear glacier ice misrepresent where mummies are found—they occur in silt deposits across unglaciated portions of the north. Mummification is not just a matter of freezing; carcasses must be buried by some rapid process. The geology of Beringian deposits and the fossils themselves can tell us how this usually happened. The character of vegetative cover is also an important requisite to understanding surficial geology. During former warm and wet episodes (interstadials and interglacials), trees moved north and grew in most of the interior as they do today, and unforested areas were covered by wet tundra. But during drier glacial intervals the Alaskan interior was an arid grassland (fig. 2.22).
The geology of Pleistocene deposits in the Fairbanks area has been thoroughly examined by Péwé (1975a). This information was obtained mostly from placer mine exposures, but also from natural river cut-banks (fig. 2.23). Péwé’s interpretation of the reworking of primary loesses into valley bottom deposits of dark, putrid-smelling organic silt, rich in ice and bones, was derived from analogies with what we see today. At present silt moves downslope only above the treeline, where solifluction—a slight annual slippage of thawed soil downslope over the yet frozen portion beneath—is taking place. Péwé (1975a) implied that solifluction created the fossil-rich silt in some areas of the Fairbanks lowlands. Solifluction movements form a characteristic lobate pattern on steep hillsides and can be easily identified by shrubs that grow along the lower edge of the lobe. On the other hand, as Péwé saw, silt from full glacial episodes is evenly and thinly bedded. I argue that these fans of silt, washed from an incompletely vegetated grassy landscape, were the most important depositional agent in the deep deposits of the valley bottom.
Like all ground cover in interior Alaska, except for high-alpine talus and river bars, solifluction lobes are thoroughly vegetated. Bones lying in such a vegetation mat are rapidly incorporated by plant overgrowth (usually by moss), then leached and destroyed by root acids. The rate of movement of the solifluction lobes is not fast enough to incorporate many bones. Additionally, only a small margin of the solifluction lobe would work to cover bones in any one year. It is difficult to imagine a large mammoth bone or skull being incorporated intact. Solifluction movement would take several decades to cross the specimen, leaving one end decomposed, which is not the case with most Beringian fossils. One has to invoke a different, rapid depositional environment for large mammal mummies, and yet there seems to be a continuum between mummies and the smallest bones.
Fig. 2.22. The two faces of Alaska. Alaska has been a critical biotic link between Eurasia and North America. Glacial Alaska was virtually Eurasian; interstadial and interglacial Alaska was the most northern part of North America. During warmer, wetter interglacials and interstadials, arboreal vegetation and fauna (shown here as spruce distribution) recolonized Alaska from the south; at the same time, high sea levels separated Alaska from Siberia. During these wetter times, Alaska was separated by the Laurentide and Cordillerian ice sheets. Lowered sea levels provided a broad connection between Alaska and Siberia. The cool, arid climate pushed tree distribution far to the south.
Fig. 2.23. Natural exposure created by Birch Creek. In the dark part of this cliff, large ice wedges, the size of houses, are melting out. The stream removes the thawed slump and keeps the exposure fresh all summer. Many Pleistocene mammal bones are found at sites similar to this one.
It seems more accurate to say that there are few modern analogues to the downslope silt movement that occurred in the Pleistocene. We should imagine a different soil surface from the present thick layer of moss, lichen, or partially decomposed plants. Today, any silt that begins to move downslope is immediately filtered out and the water runs clean.
Evidence from an array of studies (Guthrie 1982; Hopkins et al. 1982) suggests an incomplete surface cover for Pleistocene soils. There may have been a closed root system underground, but the rootless cryptogamus mosses and lichens, with their above-ground biomass, were much less abundant during cold phases of the Pleistocene. Areas of bare soil seem to have been exposed between plants, like grasses, which retain most of their biomass below ground. A rough analogue is found today on the American High Plains, where soil surface is exposed amid growth of short grasses. Erosion would have been most pronounced during spring or autumn, when naturally poor soil cover was further depleted by grazing or burning.
This exposed-soil grassland is quite unlike most tundras, which have a complete vegetation mat. But thick or continuous ground cover is not a measure of food quality. In fact, the present vegetation in interior Alaska is not subject to heavy grazing by large mammals because it is largely composed of very conservative and highly defended species rich in antiherbivory compounds (Guthrie 1982, 1984b). These same toxins slow decomposition, and old plant tissue accumulates to form an insulating layer that prevents deep thaw and hence restricts drainage, which in turn creates muskeg lowlands. These muskegs seldom bum, but when they do it is a “top bum” that does not greatly affect humus-forming plants. Even after severe wildfires there is little sheet erosion of silt.
During the Pleistocene, however, there seems to have been frequent transport of exposed silt downslope (Wu 1984). We have several measures of the most extreme degrees of this transport but do not know the usual amounts. Most silt in valley bottoms is finely bedded, indicating that the normal amount transported may have been rather modest. But during heavy rain or rapid snowmelt, enormous quantities could be moved over a short time. As previously mentioned, silt has the ability to go into suspension easily and to remain suspended as long as there is water movement. Once movement slows, silt particles come out of suspension and are deposited as fans or bars (fig. 2.24). I can imagine silt washing downslope during the Pleistocene, varying from broad sheets to
channeled streams flushing over their banks. Despite the ease with which silt can be eroded, there are few artifacts of steep-walled stream channels in Pleistocene deposits. To me this suggests that most Pleistocene precipitation occurred as infrequent, large-drop downpours—short bursts of rain that exceeded the rate of soil absorption, as it does today in the shortgrass plains. This situation contrasts with the frequent, small-drop rain showers we now have in the interior.
The majority of Alaska Pleistocene fossils were buried by the kind of deposition just described. Nevertheless, judging from the volume of silt required and the season of burial (evidence I discuss later), Blue Babe was probably buried in early spring in a different way. It seems most likely that drifted snow melted rapidly from the uplands, carrying enough silt to cover the bison. Blue Babe was found in the throat of a small basin 6–7 km2 in size, so the stream draining the basin could not have been very large. And if we look immediately above the bison, at the area from which silt could flow, there is only 0.3 km2 (fig. 2.25).
Fig. 2.24. Silt flow.
At Fox, a few kilometers north of Fairbanks, a long tunnel has been cut into the permafrost. Originally made for engineering research, the tunnel provides a unique underground look at the permafrost phenomenon (Hamilton, Craig, and Sellmann 1988). One feature exposed in the tunnel directly relates to how frozen mummies are preserved. The tunnel bisects inactive fossil ice wedges, some of which are capped by ice from a small pond. This pond formed when the top of the ice wedges melted during a warmer climatic episode. The interesting point about this pond ice is that it is so clear; this clear ice indicates that enough silt was deposited to keep the pond below freezing. Any thawing of the pond would have incorporated some overlying silt and resulted in dirty ice. The pond has been preserved as it existed in its winter state. For upslope sediments to thaw sufficiently that they wash downslope and cover a pond, without the pond itself melting, the pond has to be buried with a thick cover of silt insulating it from summer’s warmth.
There is a brief time in the spring when all this could happen. I propose that snowmelt from the first warm days of spring, say April, washed down exposed soils on the slopes above, further thawing the silt and carrying it in suspension toward the valley floor. This deposition must be surface wash, because there is a very narrow window of time in which the pond is still frozen yet erosion from melting can occur.
If such a small frozen pond can be buried by an insulative cover so that it would not thaw in subsequent summers, we can begin to imagine how a frozen mummy might be preserved. In Blue Babe’s case, however, slipped hair and partial decomposition shows that, unlike the pond, the bison carcass thawed somewhat during the first summer or two.
Snowmelt can be abrupt in the Yukon-Tanana uplands—a result of two factors. The spring sun gains height rapidly at this latitude. Day length increases at a rate of seven minutes a day, yet below-freezing temperatures persist well past the equinox. The product is bright spring days in a frozen world. The inertia of winter frost and springtime insolation finally reaches a tilting point, and the snow melts in a burst. River ice lifted from its anchor by the flush disintegrates as it begins to move downstream, creating a special season known as breakup. Of course, abruptness of breakup varies from year to year; some years it occurs smoothly and gently, while other years it is explosive and violent. Frozen mummies probably originated in the latter kind of breakup, as seven month’s precipitation flushed across a landscape of exposed soils in a few intense days.
Fig. 2.25. Pearl Creek drainage. This map of the Blue Babe locality shows the drainages that could have carried silt to cover the mummy. The upstream drainage is of moderate size and could have furnished considerable silt if the bison were buried by over-bank alluvium, but this does not seem to have been the case. Apparently, the silt that buried Blue Babe came from a much more limited area, washing down the slope at a right angle to the creek bed. Gold mining activity follows the old Pleistocene streambed; note that the trail of tailings is different from the existing stream, especially up valley.
In the drier Pleistocene landscape, a heavy rain and spring snowmelt would leave contrasting runoff patterns. Rain would produce a more even sheet of water flowing down broad flanks of the hills. Modest Pleistocene snows, however, probably accumulated in drifts blown from slopes into swales (Guthrie 1982), as indeed snow does today at slightly higher altitudes in the Tanana Hills. Snowmelt thus flows in concentrated rivulets, or channels, downslope from such drifts. However, Pleistocene snowdrifts would contain dust from the exposed soil surface, and this dusty component would have greatly changed albedo and increased the rapidity of spring thaw.
Hamilton, Craig, and Sellmann (1988) presented another interpretation of interior Alaskan silts, based on their work in the Fox permafrost tunnel. They think silt was deposited by two processes—slow accretion of wind-bom dust and massive redeposition of this reworked loess—and propose that eolian accretion occurred during cold, drier, glacial episodes, while redeposition processes were active only during wetter interstades and interglacials. These latter redeposited silts dominate the sedimentary record, in their estimation. I think this interpretation overlooks silt reworking during the glacials and overemphasizes interstade—interglacial contribution to the sedimentary record. There are simply too many well-preserved bones and large skulls that radiocarbon date to full glacial (Duvanny Yar). These could not be preserved by a few millimeters of annual eolian loess-fall; their preservation required large quantities of reworked silt. The frozen silt we find enclosing these glacial aged bones is water saturated, not dry, drifted eolian material. For example, the entire skull with tusks of the Colorado Creek mammoth was buried some 15,000 years ago (Thorson and Guthrie, in prep.), prior to Hamilton, Craig, and Sellmann’s postglacial episode of sediment reworking. They do portray an interval of mass silt wastage, between 36,000 and 30,000 years ago, that occurred due to increased moisture during the last interstade, which corresponds to my interpretation of the processes that buried Blue Babe during that time period.
I suggest that we are seeing two quite different Pleistocene patterns of silt redeposition in interior Alaska. In one, summer rains produced a broad sheetwash of silt in the millimeter and centimeter scales (on rare occasions much deeper), often over a shorter distance, with perhaps most not reaching the valley floor. In another, rapid spring snowmelt moved in more confined channels, probably taking large amounts of silt all the way to the valley floor. It is near these valley bottoms that one finds Pleistocene mammals buried.
Vereshchagin and Baryshnikov (1982) have proposed that late Pleistocene fossils, particularly the frozen mummies, are most likely to come from two time periods—the first around 11,000 years ago and the second 35,000–40,000 years ago. These episodes may have been unique times in Beringia, when two taphonomic factors coincided: sufficient moisture for silt transport and exposed soil surfaces. During full glacial conditions, there was probably not enough moisture, especially deep snowdrifts, for rapid silt transport. Silt moved downslope, and animal parts, but not many large carcasses, were entombed. During the Boutellier Interval (the last interstadial—isotope stage 3) and during the early Holocene there was an increase in moisture, and as moisture increased, so did erosion and redeposition—as long as bare soil existed. On the other hand, additional moisture favored mesophytic plant species, which quickly created a complete ground cover and prevented silt movement. There must have been a narrow window after moisture increased but before vegetational change slowed erosion when conditions were ideal for preserving large fossils. Erosion may also have increased at the end of the mesophytic episodes. Bison mummies from Siberia and Alaska appear most commonly within these two periods. It is the older, interstadial, window which concerns us in our reconstruction of Blue Babe’s death and burial more than 35,000 years ago.
Blue Bones
Although Pleistocene bones from silt deposits in Beringia are themselves stained dark from organic chemicals and minerals (they range in
shades from ivory to jet black), their surface is often covered with a dusting of brilliant blue. As our bison mummy dried, its gray surface also turned blue. This blue was the same color as the dust used to chalk billiard cues or to powder a mason’s line—an incongruous hue on the remains of so old an animal.
The blue that forms on Beringian fossils is actually a mineral known as vivianite, an iron phosphate; whitish gray in its unoxidized condition, vivianite soon turns blue when exposed to air. This mineral occurs where organic remains of animals, low in iron but high in phosphates, are buried in damp silt that is relatively rich with iron but phosphate poor. Palynologists see vivianite layers in pond sediments (H. E. Wright, pers. comm.). Presumably, this situation arises from a similar combination of iron oxides (which precipitate out in standing water) and phosphate-rich organic chemicals from the remains of pond organisms. Such bands of vivianite are white when the pond is first cored, but like Pleistocene bones, they turn blue when exposed to air. Chinese potters used iron phosphates in their glazes to obtain the blue-green celadon which imitated jade.
Vivianite appears as an irregular dusting on many fossil bones I have collected. Sometimes this coating is so thin as to be barely visible; on other specimens it is a thick crust that totally masks the underlying bone. In the latter case all one sees is a blue form in the shape of a bone. Blue Babe was not only covered with a dusting of vivianite, but additional wartlike growths (fig. 2.26) of vivianite crystals occurred in clusters on the skin (the size of 0.5–1.0 cm). These blue warts were especially apparent on the head. When they were removed in the preservation process, the underlying skin showed pock-shaped erosion craters, undoubtedly the result of chemical breakdown of phosphorus-containing skin proteins, particularly collagen, in the underlying dermis.
Frozen Fauna of the Mammoth Steppe: The Story of Blue Babe Page 8
