The figure above illustrates what taxonomists call a "consensus tree" describing relationships and timing of divergence among species within the family Ursidae, as well as (in the inset) relationships and divergence times for various subgroups of the species Ursus arctos (the brown and grizzly bears). The polar bear Ursus maritimus is included here for reasons that will be addressed shortly. This so-called consensus has been reached only within the last few years, and may be revised in light of new research. But the main part of it will likely hold over time. An important feature to be noted, in addition to the main branching, is the green arrows. These arrows indicate on-going gene flow between "species," which has led to some of the confusion and disagreements over relations and dates of divergence among different species. Note the gene flow from polar bears to grizzly bears, and from grizzly bears to black bears, which suggests inter-fertile breeding, albeit under presumably rare circumstances.
Surviving bear species of the family Ursidae are, for the most part, relatively recent in origin. The most ancient surviving derivative is the giant panda (Ailuropoda melanoleuca) which, although most closely related to bears, is so different as to have led taxonomists to put into a separate family, or even cluster it with raccoons. Another main split is between species of the subfamilies Tremarctinae and Ursinae. This split happened about 6-13 million years ago (mya), giving rise to bear species that evolved and survived mainly in North America (the Tremarctine bears), and those that evolved primarily in Eurasia (the Ursine bears). The only surviving bear of the Tremarctine lineage is the Spectacled bear (Tremarctos ornatus) of South America, although prior to the end of the last Ice Age a number of Tremartine bears had existed and flourished. Perhaps the most spectacular of these bears was the Giant short-faced bear (Arctodus simus), which was a giant as much as 6 feet tall at the shoulder that survived until roughly 13,000 plus or minus 300 years ago (this in years corrected for bias in carbon dating).
But the main focus here is the species Ursus arctos, which encompasses the grizzly bears living in Yellowstone. Grizzlies and their kin the Polar bear and Eurasian cave bear (U. spelaeus) split from the lineage giving rise to the American black bear (U. americanus) roughly 3 mya. These species all shared a common ancestor called U. minimus, which gave rise as well to an intermediary form called U. etruscus. Barring the American black bear, these species are all best represented by fossils that have been found in Europe and western Asia.
More recently, the branch giving rise to the cave bears split from the brown bear lineage in Eurasia roughly 1-3 mya and, more recent yet, the polar bear lineage split from brown bears about 200-500 thousand years ago (kyBP). There is persisting disagreement among those who study phylogeny regarding when "the" polar bear split happened--some arguing as long as a million years ago, others as recent as 300 thousand years ago. The waters remain muddy because this so-called split has continued to be blurred by interbreeding among brown and polar bears. Whatever the conclusion, polar bears are closely related to brown bears.
But even before the divergence of polar bears, brown bears began to diversify into lineages, or clades, most of which have survived to the present. The major split between brown bears that live in Europe (Clade 1, along with an anomalous bunch on the ABC Islands of Alaska--Clade 2a) and brown bears that live elsewhere (Clades 3-6) occurred around 300-900 kyBP. The reasons for this complex of Clades has a lot to do with how various lineages of brown bears were split and more-or-less isolated during the course of the Ice Ages, which is the story covered under Evolutionary biogeography. Of particular relevance to Yellowstone's grizzlies: all of them belong to Clade 4, which appears to have been isolated in central North America when, according to Chris Stokes and collaborators, the Pleistocene ice sheets of North America coalesced across the northern span of the continent for the last time roughly 70 kyBP, albeit with what seems to have been a temporary opening around 55 kyBP. Clade 4 split from the various lineages of Clade 3 (currently concentrated in Asia and Alaska) around 200-350 kyBP. Clade 4 currently survives nowhere other than the in the center of North America--and on the island of Hokkaido, which testifies to the Eurasian ancestry of all brown bears, including bears of this clade.
The graph to the left provides another view of the timing of divergence among different brown bear clades--along with an estimate of long-term temperature anomalies (from gases trapped in Greenland ice cores) and estimates of the total brown bear effective population size (the red line). This latter estimate was made by geneticists based on analysis of brown bear genomes. I personally give it little credence, but it is an interesting scientific result nonetheless.
There is no consistent pattern in divergence times relative to global temperature anomalies. Clade 1 diverged from Clade 2 in western Eurasia at a time of comparative warmth. Clade 4 (Yellowstone's bears) diverged from Clade 3 in Asia at a time of comparative cold...as was the case for diversification within Clades 2 and 3 roughly 120-220 kyBP. Likewise, Clade 2a became isolated on the ABC islands during near peak coldness.
In some ways the diversifications that happened during cold periods make sense. It is at these times that different subpopulations are thought to have become isolated in refugia, within which genetic differentiation would then plausibly proceed (see Biogeography).
The shrinkage of brown bears
One interesting parenthetical observation about brown bears is that they seem to have diminished in size over the course of the last ice age (the Pleistocene). This was first observed by Bjorn Kurten, a world-famous Finnish paleontologist, based on measurements of fossil brown bear skulls collected in Europe.
More recently, Adrian Marciszak and associates published results, also based on measures of bear skulls found in Europe, that seem to have pretty much nailed this trend. These results are summarized at left, arrayed according to time, from left to right, differentiating males from females and measures of skull length from skull width. (The acronym MP refers to "middle Pleistocene," progressing in time to the recent, "R"). One feature of these results is the suggestion that males dwarfed more substantially than did females.
This begs the question: Why? Hemispheric patterns in the size of contemporary brown bears (see Morphology) suggest that bears are larger where they eat more protein, typically in the form of meat. There was certainly a lot of meat around during the Pleistocene, perhaps in the form of carrion able to be obtained from the carcasses of the numerous large herbivores dying from any number of causes (see History). According to Herve Bocherens, an expert on reconstructing diets from the analysis of ancient tissues, brown bears were especially carnivorous during the Ice Age, especially in the steppe tundra of Eurasia. This is all consistent with the larger size of Ice Age brown bears being a consequence of ingesting more meat. Even the differences between trends in male and female sizes is consistent with a meat effect given that males tend to eat comparatively more meat whenever it is available, which means they would have been comparatively more affected by the rapid decline of the meat resource caused by widespread extinctions of the large herbivores towards the end of the last Ice Age.