Arctic biota underwent dramatic changes during Quaternary glacial-interglacial periods, and these historical processes are reflected in the genes and gene pools of Arctic species.

Geographic isolation in different glacial refugia generated genetic divergence between separated gene pools. Species responded to the glacial-interglacial periods by retreat and colonization, often associated with demographic changes (Hewitt 1996). The processes of distribution changes affected the geographic distribution and amount of genetic diversity. Phylogeography, the analysis of geographic distributions of alleles whose phylogenetic relationships are deduced (Avise 1998), provides insight into historical factors generating extant patterns of genetic variation. Comparisons of genetic patterns across several codistributed species can separate general and species-specific effects and reveal alterations at the level of community and landscape associations.

Isolation in glacial refugia was an important historical factor for the diversification of Arctic species in extensively glaciated North America. The traditional view is that species survived the Pleistocene glaciations in two refugial areas: the Eastern Beringia, large unglaciated land west of the Mackenzie River, and the periglacial tundra south of the main ice sheet (Pielou 1991). Postglacial colonization from these refugia was suggested for the brown lemming Lemmus trimucronatus (Macpherson 1965). Thus, two genetically distinctive groups can be expected within the present distribution range of the brown lemming. On the other hand, paleogeographical data provide indications for the existence of other glacial refugia in ice-free areas north of the main ice sheet in the High Arctic Archipelago and the coastal part of northern Greenland.

Accordingly, the glacial survival in the Eastern Beringia and unglaciated areas of the High Arctic was proposed for the Arctic specialist, the collared lemming Dicrostonyx groenlandicus (Macpherson 1965). There is little fossil and paleoecological evidence yet available from the High Arctic. Nevertheless, phylogeography studies revealed several genetic groups in the collared lemming (Fedorov and Goropashnaya 1999) and in the Arctic plant Dryas integrifolia (Tremblay and Schoen 1999). Geographic distributions of these groups furnish some support for the glacial survival in the High Arctic. Phylogeography study using fast evolving genetic markers and adequate sampling design can locate refugial areas in the High Arctic.

The aim of the project

This project is a continuation of genetic studies in lemmings from the Eurasian (Fredga et al. 1995; Fedorov 1999) and North American (Fedorov and Goropashnaya 1999; Ehrich et al. in press) parts of the Arctic. The aim of the project is to evaluate the effects of Quaternary environmental fluctuations on the present patterns of genetic variation in the two genera of lemmings, Lemmus  and Dicrostonyx. Molecular genetic data in combination with paleoecological information are used to reveal biotic responses and, thus, the history of these small rodents. Comparison of these results with findings from the Eurasian Arctic will shed light on general trends in the biotic responses of lemmings to Quaternary environmental fluctuations on a circumpolar scale.

The field-work

In the field the aim was to collect collared lemmings for genetic sampling from all expedition sites and brown lemmings from sites situated south of the Parry Channel, the northern limit of species distribution. Due to low lemming density (Angerbjörn et al.) and limited research time in most localities, this objective was only partly achieved, despite massive trapping efforts. To increase probability of capture, snap traps were set selectively at spots showing signs of lemming activity. This approach was more effective for trapping collared lemmings in High Arctic localities where active burrows were easily detected. Selective trapping was a supplement to a large standardized trapping programme Conducted by the team of ecologists to obtain estimates of lemming density and demography (Angerbjörn et al.). Systematic trapping by transects with numerous snap traps was particularly productive in rocky and tussock habitats on the Canadian mainland where signs of recent lemming activity could not be detected. Systematic trapping as well as live trapping for feeding experiments (Agrell and Berteaux)  and chromosome research (Fredga and Fedorov) provided a number of lemmings for the phylogeography project. In total, tissue samples for DNA analysis were taken from 87 collared lemmings from all localities, except for sites 7 and 15. Sampling size per locality varies from 1 to 19 collared lemmings with a mean of 5.8. Brown lemmings were sampled only from sites 9, 10 and 11, with sampling size from 1 to 14 and a mean of 6.3 lemmings per locality. In addition, a number of dead lemming remains were collected in the field and these specimens are reasonable for DNA analysis (see below).

Preliminary results

Selection and development of appropriate genetic markers are essential steps for phylogeography and population genetic studies. Previously we used nucleotide variation in the mititochondrial DNA cytochrome b gene (about 900 nucleotide base pairs) for phylogenetic reconstructions in collared lemmings Dicrostonyx (Fedorov and Goropashnaya 1999) and true lemmings Lemmus (cf. Fedorov 1999). To increase phylogenetic resolution in the North American collared lemming D. groenlandicus I recently developed specific primers for polymerase chain reaction and the sequencing of highly variable mtDNA in the control region (about 1000 nucleotide base pairs). The use of these primers makes it possible to amplify and sequence relatively long (up to 400 nucleotide base pairs) segments of the control region with minor amounts of DNA extracted from the dry remains of lemmings found dead in the field.

More laboratory work is needed before new results can be reported. However, it is possible to mention some findings from the Tundra Northwest 1999. D. Ehrich sequenced part of the control region in ten collared lemmings collected from Liverpool Bay (site 11) and one collared lemming from Amundsen Gulf (site 8). The same mtDNA haplotype was found in all individuals and only this haplotype was previously reported in collared lemmings from two other localities in the same geographic area (Ehrich et al. in press). The Holocene forest advances to the Beaufort Sea coast and retreats are well documented in this area. Thus, the lack of mtDNA variation in the tundra specialist, the collared lemming, is probably a result of regional bottleneck events due to range contractions during the Holocene warming events. This finding is consistent with results reported in lemmings from the Eurasian Arctic (Fedorov 1999).