The genetics of refugia

4 minute read

The NY Times gave a short writeup earlier this week to a paper about ancient DNA from arctic foxes:

"We wanted to know what happened with the Arctic foxes over the transition from the ice age to the current warm period," Dr. Dalen said. "When the tundra shifted up to Scandinavia and Siberia, did they move too?"
The researchers analyzed DNA from fossilized fox bones found at European ice age sites, and compared it with DNA from the current Scandinavian and Siberian populations. They found that there was no connection between the ancient and modern populations.
"They didn't move," Dr. Dalen said of the European animals. "That whole population is extinct."

The paper itself is a simple, 3-page read. The newsworthy element of the paper is its relationship to climate change -- with the implication that the current genetic diversity of many species will be lost if climate change restricts them to a limited part of their ranges. In the Pleistocene, habitat changes happened without humans getting in the way. So the observation that foxes didn't "track" the movement of their habitat as the glaciers receded means that today's species are very unlikely to do so, if climate zones move in the near future.

But I found a different implication to be more interesting. Arctic foxes that live in northern Scandinavia today are essentially occupying a refugium -- a shrunken fragment of their original habitat. The ancient DNA shows that foxes across northern France, Germany, and Russia were not mtDNA ancestors of today's Scandinavian foxes. While foxes occupy an interglacial refugium, we can look at their mirror image in the glacial refugia of European species; notably Neandertals.

During the height of the last glaciation, Neandertals appear to have been constrained to the southern tier of Europe -- possibly limited at some times to Iberia, Italy, Croatia and points further south. Usually, when people talk about these refugia, they mention northern European populations moving south. But Southern Europe was already full of Neandertals, and there probably was no moving south for the increasingly marginal populations of France, Germany, and other parts of northwest Europe when the climate deteriorated. If they followed the fox model, then northern Neandertal populations may have simply became extinct during each glacial maximum. Southern populations may have undergone substantial demographic turnover also, since glacial and interglacial conditions would have selected for different phenotypes for some characters.

There are many differences between arctic foxes and Neandertals in geographic range and life history, so it is certainly possible that Neandertals moving south maintained greater population continuity than the foxes moving north. Small mammals may grow their population faster than they can disperse over long distances. For Neandertals, that may be less true -- long-distance dispersal would certainly have been possible; the question is whether the population density of the southern refugia would have allowed it.

In general, I think the fox analogy is probably a good one, since carnivores have large home ranges and dispersal distances for their size. The home range size of arctic foxes today averages between 20 and 50 km2 (Eberhardt et al. 1982; Landa et al. 1998), depending on the local habitat. Pups disperse outside their parents' home range for the most part, with a dispersal distance between 20 and 40 km (Strand 2000). The home ranges of Neandertals were larger,

The periodic reduction of Neandertals to glacial refugia in southern Europe would have set up a pattern of extinction and recolonization across most of Europe. This must have been a very important demographic force in Neandertal evolution -- since the continent underwent repeated episodes of climate change, possibly on a submillenial basis during the Late Pleistocene. During each range restriction, a nonrandom sample of southern European Neandertals survived and increased their relative gene frequencies compared to other Neandertals. Whenever conditions were suitable, this southern European population was the "first on the scene" to expand into the empty habitat of central and northwestern Europe. We can imagine some strong gene flow from outside Europe also, but the demographic growth of southern Europeans made their relative allele frequencies increase with a pulse every time the population expanded.

Hence, most of Europe was a pulsed population sink. Significant source populations in southern Europe may have maintained substantially distinct allele frequencies than contemporary populations outside Europe, both as a result of restricted gene flow during glacials and as a result of strong selection for dispersal and colonizing ability. These conditions would explain a relatively strong mtDNA distinction between Neandertals and some contemporaries, in comparison to relatively slight autosomal and X chromosomal differences. The mtDNA is much more strongly affected by restricted and temporally intermittent gene flow because of its smaller effective size. In a growing and shrinking population, the effective size of mtDNA would have made it more strongly affected by drift than other loci. In effect, it may be the strongest signature of this evolutionary pattern.

That's what seems to be the story with the foxes. The ancient mtDNA shows a lack of close relationship between today's arctic foxes in Scandanavia and Pleistocene populations further south. The range contraction had a strong effect on mtDNA. It will be of interest to see if autosomal genes show a similar effect, or whether instead they share the Neandertal-modern pattern of slight differences.

So far, so good. But there are a couple of kinks. More later.


Dalén L and 8 others. 2007. Ancient DNA reveals lack of postglacial habitat tracking in the arctic fox. Proc Nat Acad Sci USA 104:6726-6729. doi:10.1073/pnas.0701341104

Eberhardt LE, Hanson WC, Bengtson JL, Garrott RA, Hanson EE. 1982. Arctic fox home range characteristics in an oil-development area. J Wildlife Management 46:183-190.

Landa A, Strand O, Linnell JDC, Skogland T. 1998. Home-range size and altitude selection for arctic foxes and wolverines in an alpine environment. Can J Zool 76:448-457.

Strand O, Landa A, Linnell JDC, Zimmermann B, Skogland T. 2000. Social organization and parental behavior in the arctic fox. J Mammal 81:223-233. doi:10.1644/1545-1542(2000)081<0223:SOAPBI>2.0.CO;2