Pleistocene-early Holocene diet

I've spent a fair amount of time during my morning walks thinking about what the world of grizzly bears might have been like during the late Pleistocene and early Holocene in North America. Much of my curiosity has focused on their likely well as their likely predators and competitors. The foundation for all of these imaginings is necessarily a reconstruction of the environment and the foods, whether animal or vegetal. Here I focus on environments and diet between roughly 20,000 and 9,000 years before present, from the peak of the last glaciations (the Last Glacial Maximum or LGM) and the onset of serious heat and drought during the Altithermal. I start below with a synoptic overview of broad-scale environments, then briefly consider vegetal foods, and conclude with a consideration of animal foods, especially meat from large herbivores. For more detail on climates during this period see Early prehistory. For more detail of probable diets during the centuries immediately preceding the arrival of Europeans, see Pre-European diets.

The World of Then

The map above summarizes an amalgam of information from a number of sources, all pertaining to the climate and environment of North America roughly 20,000 (or 20k) to 18k years ago, during the LGM. The continental icesheet margins are from Arthur Dyek; the mountain icecap delineations from Ken Pierce; the isopleths of summer temperature differences from San-Ik Shin and his colleagues; forest cover from models by John Williams; and approximate vegetation boundaries and types synthesized from a host of sources. I've surrounded the map with photos of vegetation that I speculate was typical of much of what is now the western US during the Pleistocene, arrayed so as to depict environments progressively from the north to the south: from dry tundra, to a tundra-parkland mix, to grassland and shrubland with enclaves of protected woodland and forest, to conifer woodland.


A couple of features are noteworthy. First, it was a lot colder. Along the margin of the continental icesheets it was roughly 12 degrees C (or 22 degrees F) colder during the growing season--cold enough to have likely maintained polar desert. The degree of cooling progressively lessened the farther south you around 4 degrees C (7 degrees F) colder near the Mexican border. Second, vegetation in the western half of the ice-free continent was quite open, probably even more so than today. Most models show that much of the West received about the same amounts of precipitation as it does today. However, a number of researchers have shown that much lower levels of atmospheric carbon dioxide during the Pleistocene (roughly 190 ppm versus 290 ppm during 1700s) shifted vegetation, all else equal, from woody forms to herbaceous forms. The result would have been more grass and forbs than one might have expected with a given climate. Less atmospheric CO2 also probably resulted in greater day-night temperature swings, much of it driven by colder nights.


Putting this all together, the world of Pleistocene grizzlies south of the continental icesheets and west of the eastern forests was colder and probably typified by lots of grass and forbs and fewer shrubs and trees--which would be consistent with the number of large herbivores roaming the western part of the continent during the Pleistocene (see below). The mid-continental grasslands warrant a comment. In past decades it was common for paleoecologists to claim that the Great Plains were covered by woodlands, parklands, and even forest during the Pleistocene. The current weight of evidence suggests that this region was largely open then much as it is today. Even so, there were probably stringers of white spruce, birch, cottonwood, and aspen in draws or on protected slopes in the northern part of the plains, along with limber pine on rocky ridges and outcrops. 


Focusing for a moment on the Yellowstone region, you will see the prominent ice cap that covered all of the Yellowstone highlands (see Yellowstone paleontology for more details). All of the evidence I've been able to find suggests that north of the ice cap you entered into an especially cold world of tundra and polar desert. South of the ice sheet was more clement, and typified by woodlands, parklands, grasslands, and sagebrush steppe. One can imagine extensive grassy areas with varying amounts of sagebrush, protected or moister microsites supporting whitebark pine, limber pine, aspen, spruce, and lodgepole pine, and somewhat drier warmer areas with greater amounts of limber pine, Douglas-fir, and common juniper.  

The figure to the left is the result of an analysis I undertook to better understand the relationship between site moisture, ambient annual temperature, and vegetation productivity. I compiled a number of productivity estimates that had been made for sites in the northern US Rocky Mountains and for which I could roughly determine annual temperature and site moisture. Because moisture is hard to precisely measure, I used an index that ranged from dry (xeric, or a value of '1') to wet (hydric, or a values of '6'). The sites spanned lower-elevation grasslands and marshes to alpine tundra.


The lines in the top graph depict the relationship between productivity and site moisture at three different average annual temperatures (the coldest [20 degrees F or -7 degrees C] corresponds with tundra; the warmest [40 degrees F or 4 degrees C] with cold temperate grasslands). The blue bars below show the decline in productivity at each level of site moisture for a 10 (light blue) and 20 (dark) degree F decline in temperature (i.e., declines of 6 and 11 degrees Celsius).


The key results are, first, that productivity increases exponentially with site moisture; second, that productivity increases with annual temperature (at least in this region); and, third, that reductions in productivity with cooling of ambient temperature are most dramatic on dry sites--to virtually nil on the driest coldest sites.


What does this all mean? I suspect that vegetation productivity was disproportionately concentrated in wetter lower-lying sites during the Pleistocene. These areas would have included swales, draws, and river bottoms. If so, then the large Pleistocene herbivores might have been likewise highly concentrated--in ways that would have made them comparatively easy to find by predators and scavengers such as grizzly bears. All of this would have heightened competition among both herbivores and carnivores. 

Vegetal bear foods?

I was interested in coming up with an evidence-based guesstimate of the vegetal foods that Pleistocene grizzlies might have eaten. Given that no one was around taking notes during the last ice age, I was left to rely upon contemporary data, more specifically, data from over 3000 sites where feeding activity by Yellowstone grizzly bears had been investigated between 1977 and 1996. Following the logic outlined above, I assumed that food species occurring at higher elevations and on wetter sites would be candidate Pleistocene grizzly bear foods, at least in the region surrounding the Yellowstone ice cap.


The graphs at right show the results of my analysis of the Yellowstone bear data, specifically, the elevational and site moisture distributions of all the major foods (barring fish and larger herbivores; also, as you might suspect, I used elevation as a surrogate for ambient temperature). The dots vertically aligned with each food are medians and the gray vertical lines bounding each dot the interquartile range.


My conclusions? Grizzlies in what is now the Yellowstone region were probably grazing horsetail (Equisetum sp.), cow-parsnip (Heracleum sp.), native clovers (Trifolium sp.), fireweed (Chamerion angustifolium), and grasses and sedges (Gramineae and Cyperaceae). Alongside, they were also probably pursuing and even excavating voles (Microtus sp.), but probably not earthworms, given that most contemporary earthworms are non-natives introduced from Eurasia. In nearby uplands they were very likely digging biscuitroots (Lomatium sp.) on exposed drier sites and eating whortleberries (Vaccinium scoparium) and seeds of whitebark pine (Pinus albicaulis) in stringers and patches of forest. Maybe even seeds of limber pine as well.


Other than that, who knows? least when it comes to the vegetal and smaller animal foods. A bit farther south grizzlies could have been excavating pocket gophers (Thomomys talpoides) and pocket gopher food caches, digging yampa (Perideridia gairdneri) and sweet-cicely (Osmorhiza sp.) roots, and, perhaps, if they had lived far enough south, seeds from white pines and acorns from oaks such as Quercus gambelii (but see my speculations under Early prehistory).

Meat from large herbivores

I suspect that the amount of meat in local diets of Pleistocene grizzlies varied widely, primarily as a function of the competition they faced from other carnivores. However, I also suspect that the meat in grizzly bear diets increased substantially during the late Pleistocene-Holocene transitions as most of their competitors (and predators) went extinct, and despite the demise of most species of large herbivores as well.


Why do I think this? As I describe in the page devoted to Early Prehistory, there were a lot of large carnivores around during the Pleistocene, including lions, short-faced bears, dire wolves, and saber-tooth and scimitar-tooth cats. Grizzlies would not have fared well trying to either protect a kill or contest a found carcass when confronting such competitors. In fact, they probably sometimes ended up as prey of the largest of these predators, as seems to have been the case for cave bears in Pleistocene Europe and as continues to be the case for brown bears in areas occupied by Siberian tigers. On the other hand, in the absence of such competitors brown and grizzly bears can be quite carnivorous, as is the case for contemporary populations in meat-rich environments (see Pre-European diets).


But getting back to the Pleistocene. I suspect that competition with and predation by other large carnivores could have been intense enough to limit most grizzlies to marginal areas such as the swath of tundra along the continental icesheet margins (see above). But given such limits, grizzlies were probably locally adaptive in response to variations in competition and the meat resource. Herve Bocherens describes precisely such a situation in Pleistocene Eurasia, where brown bears coexisting with the largely herbivorous cave bears in Europe seem to have been quite carnivorous, while grizzlies hanging around meat-eating giant short-faced bears in Beringia were much more herbivorous. 

With that in mind, it is interesting to look at the late Pleistocene distributions of large herbivores. We don't know exactly what those distributions were, but the distributions of fossilized remains probably provide a decent approximation.


Running with that assumption, the maps at left show the locations of Pleistocene fossils compiled in the FAUNMAP database, which is probably the most comprehensive resource of its kind. Dots of different shapes and colors denote different species, genera, or even families of megaherbivores. Looking first at the top map, the stars denote camelids, the diamonds, horses (Equus sp.), the inverted diamonds, stag-moose (Cervalces), and the circles, ground sloths (Megatherium and Nothrotheriops). Blue dots in the bottom map show the locations of elephant remains, both mastodons (Mammut) and mammoths (Mammuthus sp.). For additional reference I've shown ice sheet margins during the LGM and approaching the terminus of most megaherbivore extinctions (21ka and 13.5 ka, respectively), forest cover during the LGM (in shades of green), and the approximate limit of tundra (the dashed line from the map above). I show tundra because of my speculations about the extent to which grizzlies might have been associated with this environment during the Pleistocene.


The main noteworthy patterns are, first, the relative dearth of remains in what was probably the tundra and polar desert north and east of the Yellowstone ice cap; second, the abundance of remains from herbivores other than elephants in more open environments of the West (this holds especially for the remains of camels, but with the opposite holding true for forest-dwelling stag-moose and mastodons); and, third, the abundance of elephant remains in the zone that would have been ice-free only between 21k and 13k years ago--presumably one of the last hold-outs of especially mastodons.


What might this mean for Pleistocene grizzlies? First, there was probably a lot of meat around in the form of carrion in most areas of the West. Even though the competition for this resource was probably intense, a fast-moving lucky grizzly might have still been able to scavenge of lot. Second, smaller-bodied camels and horses might have been an especially important resource given that competing for meat from a fallen giant such as an elephant was probably particularly hazardous for a grizzly. And, third, the concentration of elephant remains along recently-melted margins of the ice sheet would have likely increasingly contributed to the sustenance of grizzlies in these dynamic rapidly revegetating environments of the East. This would fit my speculations about the distribution of early Holocene grizzly remains in this region, the extent to which grizzlies seemed to track recently melted environments, and the apparent absence of eastern grizzlies after about 9k years ago (see Early prehistory).

Big changes at the Holocene boundary

So let's look a little more explicitly at changes that occurred over time. The figures above show changes in relative abundance of large herbivores during the course of the Late Pleistocene and early Holocene using data obtained (again) from the FAUNMAP database. The width of the various bands in the figure top left (Panel A) is proportional to the number of remains attributable to the corresponding taxon. I used these proportions and species-specific estimates of body size to, in turn, estimate the average size of carcasses that might have resulted from one of these herbivores dying, with obvious relevance to scavengers such as grizzly bears. The gray dots and line chart these changes in body size over time. The figure populated by a lot of dots to the right (compiled by R. Dale Guthrie) shows representative changes in the abundance of five large herbivores in more detail. Each dot represents one set of remains found for the taxon identified below it, plotted against time. The coverage is specific to Beringia for the period 18k-9k before present, using uncorrected radiocarbon dates. Finally, the figure in Panel B, to the bottom left, shows extinction dates for a number of mega-herbivores and carnivores, denoted by the terminal dots on the horizontal bars. I obtained these dates from a more-or-less definitive compilation by Stuart Fiedel.


The main patterns in the figures above pertain to the massive changes that occurred during the late Pleistocene and early Holocene transition, when virtually all of the largest-bodied herbivores and carnivores went extinct. Perhaps needless to say, the species that went extinct were the big losers, whereas bison and deer were the big winners. In Beringia, there was much the same pattern, except that elk (Cervus canadensis) and moose (Alces alces) were additional big winners, at least for a few millennia. It is interesting to note that mammoths and mastodons were apparently in decline well before the end of the Pleistocene, coincident with a proportional uptick in the numbers of horses, deer, and camels. Without taking on the onerous task of trying to explain these pre-extinction patterns, the main point is that average carcass sizes began to decline well before terminal extinctions, whatever the causes. I have little personal doubt that extinctions of the large carnivores were driven largely by the disappearance of large-bodied prey or carrion. There is ample evidence from contemporary studies that a close correlation exists between predator and prey sizes, plus lots of compelling theory suggesting that big-bodied predators are at a major disadvantage when competing with predictably more numerous smaller-bodied predators (or scavengers) for smaller-bodied prey (or carrion).


Turning more explicitly to grizzly bears: As I note in the introduction to this page, I suspect they were also big winners as a result of the changes that occurred at the end of the Pleistocene. Most of their big-bodied competitors--and predators--went away. So, barring competition with proliferating wolves and coyotes, they had more carrion to themselves. Grizzlies probably didn't benefit a whole lot from carrion on mammoths or mastodons during the Pleistocene because of the predictable ease with which the other large(r) carnivores could have dominated a large-bodied carcass to the exclusion of a grizzly. This may have even held to some extent for the carcasses of horses and camels. However, this dynamic probably reversed after the massive extinctions roughly 11k years ago because carrion from the remaining large-bodied herbivores (i.e., bison) would then have been readily available for grizzlies to dominate and consume. I suspect that carrion from bison became--and remained--a key part of grizzly bear habitat early on in the Holocene (see The bison factor).