Before Taltheilei people entered the Beverly range, it had been recently vacated by an Inuitlike people who had moved south from the Arctic Coast under extreme cold which destroyed coastal sea-mammal hunting but who returned when conditions ameliorated. They were the Pre-Dorset and had tools very unlike Taltheilei or the even earlier Shield Archaic and Plano Indian tools in manufacture, material and specialization. Their tiny, finely retouched fine-grained chert tools include end-and sideblades used in the tips and sides of harpoons, arrows or spears to facilitate wound penetration, microcores and blades, scrapers, quartzite knives (few more than 10cm), many slotting and planing burins and their spalls, chithos, wedges, adzelike tools, hammerstones and gravers. Many tools are depicted and analyzed in Gordon (1975).
Pre-Dorset is the Canadian equivalent of the Arctic Small Tool tradition which originated in the Siberian Neolithic (Irving 1970:341). It is widespread in the High Arctic, where it began over 4000 years ago. Originally, Pre-Dorsets lived on the coast, moving inland 500 years later when maritime hunting was adversely affected by rising ice thickness from the prolonged cold. In their quest for meat, these arctic-adapted people were forced further south, ultimately linking with migrating caribou herds. As only a tenth of caribou winter on the tundra, hunters adapted to spending their winters in the forests like later Taltheilei Indians. With treeline retreat in this cold period, they even reached Alberta, Saskatchewan and Manitoba, but not as far south as the Churchill River due to the presence of Middle period Plains Indians wintering in the forest while hunting bison and caribou.
As the Pre-Dorset adapted to living hundreds of km south of treeline I doubt if a Inuitlike claustrophobic fear of the forest or their unique tundra adaptation limited southern expansion. Tundra site dispersion is wide and dense like that of Taltheilei, with clusters at Warden's Grove, the middle Thelon and Dubawnt and north Artillery, Whitefish-Lynx-Howard, Sid-Mary-Mosquito, Damant-Jarvis, east Firedrake and Grant Lakes. Forest sites are limited to Pike's Portage-south Artillery, Rennie-west Firedrake, Athabasca and Black Lakes. Many hunters stayed in the range until 2650 years ago when climate warmed. Many returned to the coast to merge with Dorset. Others vanished.
The 3450-2650 year-old Pre-Dorset sites are mostly small surface tool scatters, but a dozen buried levels, two very prolific, allow dating and inclusion of many similar surface artifacts. Far northwest are Ekalluk River tools at NiNg-1, 7 and 10 on Victoria Island, their Late and Early phases dating 2600-2910 and 2980-3180 years ago (Taylor 1967,1972). Near the Mackenzie River, Pilon's Hyndman Lake Early phase site extends the Early period to 3390 and 3470 years ago, but he (pers. comm., 1992) thinks his site has western rather than eastern affiliation based on bipointed points. East of the Beverly range, Nash's Late phase Seahorse Gully site at Churchill, Manitoba in the Kaminuriak range, dates 2900 years ago.
In the Beverly range, Late and Early Pre-Dorset phases are dated using 2-3 sublevels each in KjNb-6 and 7, but dates are absent in forest sites. The Late phase dates 2785-2890 years ago in tundra sites and 2745-2800 years ago in areas forested before and after Pre-Dorset. The Early phase dates 3085-3550 years ago, the earliest on KdLw-1 bottom level Pre-Dorset chert flakes. The northern Late phase sites of KkLn-4 at Grant Lake and KjNb-6 at Warden's Grove complement southern dates at Mary and Mosquito Lakes. Combining and rounding all Barrenland estimates, the Late and Early phases date 2650-2950 and 2950-3450 years ago.
An understanding of treeline shift under plunging temperatures is needed to appreciate Pre-Dorset human adaptation as represented in tools. With treeline retreat came an accumulation of dead trees and buried stumps which could be used as fuel, but not for making handles, sledges or tent poles. Campfire remains occur but any wooden tools originating in the shrunken northern forest were not preserved in tundra or forest sites. Despite this, wood-working tools are in the south range in permanent forest, i.e., forest that existed from Pre-Dorset to the present. Stone tools are modified to forest living like other Barrenland traditions. Thus, the question of where the treeline existed is not only of botanical interest, but it also demarcates past human and caribou seasonal adaptations.
East of the study area in the similarly low and rolling Kaminuriak range, treeline retreated 50-75 km (Sorenson et al. 1971:471). Extrapolating Pre-Dorset treeline retreat in the eastern Beverly range, the Firedrake Lake cluster had forest and tundra portions. Northwest and southwest arms were in forest while the middle narrows and east shore were on tundra. At Pike's Portage and Artillery Lake in the western Beverly range, treeline retreat was altitudinal rather than southerly (Larsen (1971), with only valley trees remaining. Using nearby palynology and macrofossil data, Nichols (1976:66) suggests a 28 km retreat at Thompson Landing west of Pike's Portage. As elevation at Thompson Landing is half that at Pike's Portage (Timoney et al. 1992:Fig.2), and slightly more than half that at Artillery Lake, retreat at the latter of only 14 km but many meters vertically, is expected and places all Artillery Lake Pre-Dorset sites on tundra and all Pike's Portage and Fort Reliance sites barely in the forest.
As the treeline retreated, the southern border of forest advanced 100 km onto the prairie, giving winter haven for bison. Middle Plains Indian hunters, therefore, came close to Pre-Dorset peoples. Some of the Middle Plains projectile points were even carried north into the Beverly caribou range.
from Gordon 1996. People of Sunlight; People of Starlight: Barrenland Archaeology in the Northwest Territories of Canada. Archaeological Survey of Canada, Mercury Series Paper 154. Canadian Museum of Civilization, pp.149-152.