The Thule Migration: A Culture in a Hurry?

3.2.1 The Arctic Environment

The region into which the Thule expanded has high mountain ranges in north Alaska and along the eastern margin of the Canadian Arctic and in many parts of Greenland, separated by relatively low and flat regions in between. The sea forms a vital part of the Arctic environment and along the Alaskan coast they would have been familiar with the wide strip of “landfast” ice that forms during the winter and meets the open ocean at the “floe edge.” The sea ice there is an extremely dynamic environment, frequently broken by leads, linear stretches of open water that are created when cracks in the ice are forced open by wind or currents. In these kinds of winter ice conditions, many kinds of sea mammals can be found in leads or at the floe edge. The Thule would find similar conditions along the eastern edge of the Canadian Arctic and in Greenland, but within the relatively narrow channels separating the islands of the Canadian Arctic Archipelago, the sea freezes in almost unbroken expanses and therefore only a limited range of sea mammals remain available during the winter and different techniques are needed to hunt them. On the land, caribou could be found in most of this vast area, but their numbers and migration patterns, and thus their seasonal availability, differed regionally and between the Barren-ground subspecies in the south and the Peary subspecies in the north.

In parts of their Alaskan homeland and along the Arctic coastline at least as far as Coronation Gulf, the Thule had access to driftwood logs washed down major rivers from below the treeline. This abundant wood provided construction material for their winter houses and for other aspects of their material culture, as well as providing fuel for cooking and heating. But throughout the rest of the Eastern Arctic, driftwood was scarce, so as the migrants moved eastwards and northwards, they will have needed to meet their fuel and manufacturing needs to a far greater extent by animal products. A final difference as they expanded northwards was daylight. At the latitude of Utqiagvik and along the Arctic coast, they would have been familiar with getting through 9 weeks in midwinter when the sun did not arise above the horizon. But as they moved northwards within the Arctic Archipelago to Ellesmere Island and the north end of Greenland, the length of the midwinter dark period, with all its constraints on activities, would have extended to 15 weeks. Any model of the Thule expansion from Alaska throughout these vast and varied new regions must take their technologies into account, consistent with migrating groups reliably meeting their food, fuel, clothing, and shelter needs, while also needing to repeatedly transport themselves and all their belongings significant distances into unfamiliar territory.

3.2.2 The Logistics of Living and Travelling in this Environment

We have a considerable amount of detailed ethnographic information concerning the late nineteenth and early twentieth century Inuit inhabitants of Alaska, Arctic Canada, and Greenland, collected by anthropologists before Inuit ways of life were too drastically altered by interactions with outside societies. That information allows us to make relatively robust inferences concerning the logistics of Thule adaptations. Starting with transportation, for the brief open-water season the Inuit and their Thule ancestors had two kinds of skin boat which were both propelled by paddling: kayaks, which normally transported just one person, and umiaks, which could carry a dozen or more. For the rest of the year, during most of which the ground was covered with snow, they had dogsleds. But the number of dogs a family could keep was constrained, since in some environments, Inuit families sometimes spent as much time working to feed their dogs as they did to feed themselves (Burch, Jr, 1988, p. 86; Saladin d’Anglure, 1984, p. 491). Therefore, people often limited themselves to small dog teams and pulled alongside the dogs when the sled was especially heavily laden. Almost all of what is known about the traditional use of kayaks, umiaks, and sleds derives from their use for hunting or in annual round movements, rather than one-way migrations. For example, in Greenland at the time of contact, groups made annual seasonal moves between the mouths and heads of fjords, during which women would paddle the umiaks while men would travel in their kayaks (Kleivan, 1984). And each winter in the central part of the Canadian Arctic Inuit would sled far out onto the sea ice, moving periodically to new campsites before returning to the coast in the spring. In both cases, the people were able to cache items not needed during the upcoming months, recovering them when they returned at the end of the season and thus limiting the load that would need to be transported by sled or by boat. In the context of the Thule migration from Alaska, this would not have been an option – they would have had to transport absolutely everything that they could not easily manufacture at each stop along the way. The Thule had a very extensive and complex material culture (e.g. Oswalt, 1976, 1987) that Maxwell (1985, p. 249) characterized as “extreme functional specificity and concern with gadgetry,” so for a foraging society they would have had an unusually large number of unique implements to transport if they did not want to have to re-create many items after each leg of the migration.

Turning to shelter, in the warmer months, they used skin tents, but for the long winters, the Thule built semisubterranean houses, with the roof framework and internal features constructed of wood beams and planks in areas where driftwood was available, or elsewhere with bowhead whale bone rafters, and walls, floor, and sleeping platform constructed from large rocks. Driftwood would have been abundant in the first areas encountered by the Thule as they expanded eastward into the Canadian Arctic, but scarce beyond that. The construction of such houses required large investments of labour and of materials, and several houses would have been required to shelter a migrating group (McCartney, 1979; Park, 1988, 2001). To build a house from scratch, the below-ground portion would need to be dug out, often down into the permafrost, and then all the materials would need to be assembled. The amount of effort they put into moving large stones to build winter houses in the Eastern Arctic, sometimes more than a thousand per house, famously led Fitzhugh (1976, p. 141) to describe the effect of their building activities on the Arctic landscape as “practically a geological event, surpassed only by the Wisconsin glaciation.” But with minimal annual refurbishment, such a house could be reoccupied repeatedly, and McGhee (2009a, pp. 85–86) assumes that Thule winter houses in non-migratory contexts were occupied on average between five and ten winters, and indeed we have evidence of houses being re-used far longer than that (e.g. Houmard & Grønnow, 2017, p. 453). If a migrating Thule group chose not to create semisubterranean houses, they had the ability to overwinter in snow houses, but doing so would have been a much less comfortable way to spend the winter and would leave very few archaeological traces (Savelle, 1984, 1987).

For heat, the Thule could burn driftwood if it was available, or the small branches of the bush-like trees that grow throughout much of the region into which they expanded. But the main way that dwellings and food were heated was by burning sea mammal blubber in lamps, over which stone cooking pots were suspended. Blubber could be obtained from a wide variety of sea mammals. One other vital commodity came from a different prey species: skins from caribou. The very best warm and light winter clothing was manufactured from the skins of caribou hunted in the autumn – the skins of other species, or of caribou hunted at other times of the year, were nowhere near as effective (e.g. Stenton, 1991). Thus, both fuel and clothing could be obtained simultaneously with food, but sea mammal and caribou hunting were tightly constrained seasonally and so could in turn have constrained migration. However, the migrants may have experienced unusually easy hunting in the first few years they spent in any given region, before the unhunted or not recently hunted populations of land and sea mammals learned to be wary of humans (Degroot, 2022, pp. 73–74; Maxwell, 1985, pp. 49–50).

In Alaska, the hunting of enormous bowhead whales was one of the most distinctive activities of the Thule and contributed to shelter, food, and fuel. The northwestern Alaskan Thule hunted primarily during the whales’ spring migration around Alaska from the Bering Strait to the Beaufort Sea, when the whales were confined to leads paralleling the coast at the edge of the landfast ice. Open-water hunting of the whales on their autumn return migration was far less productive because the whales’ movements were not constrained by the ice. Ethnographic descriptions from the large north Alaskan communities in the nineteenth century describe a complex social and logistical organization for whaling that required many participants, with the residents of already large villages supplemented each whaling season by individuals from surrounding regions (Spencer, 1959, pp. 177–181). Each whaling umiak had a crew of seven to ten (Grier, 1999, p. 12; Murdoch, 1892, p. 273; Spencer, 1984, p. 330). Normally the crew were men, but women sometimes participated (Murdoch, 1892, p. 273).

The Thule migrants who entered the Eastern Arctic would continue hunting bowheads, but almost certainly on a vastly smaller scale. A study of the demography of the Arviligjuarmiut in 1938 (van de Velde OMI, Constandse-Westermann, Remie, & Newell, 1993, p. 27) found that males aged between 11 and 50 formed almost exactly one-third of the population, so extrapolating from that very limited example we can suppose that a migrating Thule group of 50 (see below) might have been able to form two umiak crews, far fewer than the number mounted at typical Alaskan whaling villages (Grier, 1999, p. 12; Murdoch, 1892, p. 273; Spencer, 1984, p. 330). The number of bowheads the migrants could catch in a given season would therefore be extremely small, but even one would likely have met their winter food and fuel needs (Whitridge, 1999, p. 108). However, in the Eastern Arctic regions lacking driftwood, bowheads were also the source of the large bones needed for house rafters and, as noted by McCartney (1979, p. 305), the bones from multiple whales would have been required to roof even a single house, whereas a migrating group of 50 people would have needed multiple houses. Therefore, small migrating groups either resided in a location for many years in ad hoc housing while acquiring enough bowhead bones to roof their entire village, or they laboriously transported rafters (logs or bowhead bones) with them.

3.2.3 An Example of an Inuit Migration

Just before exploring how archaeologists have attempted to model the Thule migration consistent with all these logistical considerations, it is useful to conclude by describing a famous nineteenth century migration that is often cited as demonstrating Inuit capacity to make rapid long-distance migrations into unknown territory (e.g. Fitzhugh, 2019, p. 52; Friesen, 2016, p. 682; McGhee, 2009b, p. 161). As summarized in most detail in the study by Rasmussen (1908), who interviewed some of the survivors, and Mary-Rousselière (1980), who assembled all of the disparate sources of information, in the spring of 1859, a group of Inuit under the leadership of a charismatic shaman named Qitdlarssuaq set out from eastern Devon Island to join the Inughuit in northwestern Greenland (in the vicinity of present-day Qaanaaq), of whose existence they had learned from British sailors. As noted by McGhee (2009b, p. 161), the distance across unfamiliar territory that they travelled in just a few years was a significant fraction of the distance that Thule from Alaska would have had to travel to reach the same region. However, the Qitdlarssuaq example illustrates the likely complexity of migration from both a logistical and a social perspective. Using Mary-Rousselière’s (1980) reconstruction of the numbers involved and the chronology, it was found that at least 42 Inuit set out from Devon Island in 1859. They did their main traveling in the spring and early summer, utilizing very large dog teams and sleds that were unusually long in order to be able to carry their kayaks. They advanced in short stages and frequently had to double back for a second load to move their kayaks and the rest of their belongings. When the sledding season came to an end sometime in July, they would locate a place to camp, where later in the summer, they could use their kayaks to hunt sea mammals. Over the summer and autumn, they accumulated provisions, and then constructed houses near their food caches, in which to pass the winter.

After 2 years of travel, disagreements among the migrants caused at least 24 of them to break off and head back. The remaining 18 proceeded on and joined the Inughuit in 1862. Several years later, most of those 18 decided to return to Baffin Island, but during the second winter of their return migration, they chose a bad location to overwinter, and several died due to starvation and interpersonal conflict; the survivors retreated to the Inughuit region. Many of the 24 who in 1861 had headed back to Devon Island appear to have suffered the same fate several years afterwards – only two are known to have survived a terrible winter of starvation. In both return migrations, the first necessary response to prolonged hunting failure was to kill and eat their dogs, which subsequently diminished their ability to hunt and travel. Thus, of the 42 who set out in 1859, more than half appear to have died within a decade due to causes that can be linked directly to the difficulties or risks of migration. If there are lessons to be learned from the Qitdlarssuaq migration, they must include that Inuit groups could indeed migrate rapidly but also that such rapid movements into unfamiliar regions were risky and could have catastrophic results if a single unlucky choice of wintering location were made.

3.3.1 Mode of Travel

With the context provided by all the logistical options available to the Thule, and by the Qitdlarssuaq example, we can now turn to how archaeologists have actually envisaged the Thule migration, although surprisingly few have attempted to describe it in any detail. In just a few sentences, Morrison (1999, p. 150) provides the most comprehensive model, which essentially matches what Qitdlarssuaq did:

it is likely that our Thule migrants did their main travelling during the long days of spring, by dog-sled, with their precious boats lashed on top. Autumn would be spent in intensive hunting, perhaps sometimes at inland locations where muskoxen and caribou may have been more reliable than coastal sea mammals. In winter they hunkered down in newly-built houses, and the next spring were off again.

Thus, Morrison sees the sled as the most important mode of travel for the migration, and other researchers have also emphasized the importance of dog traction for the migration (Hill, 2018, p. 92; Lupo, 2021). However, the boats of Qitdlarssuaq’s group only included kayaks, whereas the Thule migrants also needed to transport their much larger umiaks. A Thule umiak found in northern Greenland (Knuth, 1952) was more than 10 m long, which would have made it an extremely awkward sled load, especially when combined with all their other possessions. Alternatively, Friesen (2013, p. 350) proposes that “movement likely occurred during the summer, by umiak,” as does Mason (2020, p. 315) who notes that “a migration by watercraft squares well with the assumed rapidity of the Thule migration.” Migrating during the summer open-water months would have involved its own logistical complications, because it is unclear if an umiak fully laden with as many people and possessions as possible for the purpose of migration could also have been used for whale hunting, although we should undoubtedly imagine each migrating group as one or more umiaks accompanied by several kayaks. Travel by umiak would see the migrants arriving at their new home in the late summer or early autumn, leaving little time to accumulate supplies there before the winter. Items that in the summer would need to be transported by umiak included their sleds and dogs so, as with Qitdlarssuaq’s spring migration, it is probable that migration via umiak would have required backtracking repeatedly in order to transport everything.

Other researchers do not attempt to give priority to one mode of transport over the other. McGhee (2009b, p. 161) envisions “Thule explorers, traveling by dogsled as well as by umiak during the summer months,” and Pauketat and Sassaman (2020, pp. 298–299) note that “Umiaks enabled rapid migration with the transport of lots of people and gear” and that “Dogsleds appeared at the time of Thule expansion, so travel over land with gear and resources was vastly improved.” These models would seem implicitly to envision travel by sled in the spring, followed by umiak travel in the summer. The distance that could be migrated in a single year by utilizing both modes of transport sequentially would be impressive but doing so would impose serious time constraints on hunting or other activities.

3.3.2 Duration of Occupations

Whatever mode of transportation they employed in their moves, another important consideration is how many winters a migrating group would remain at each over-wintering location, which of course has implications for how rapidly the migration could have proceeded. Morrison’s (1999, p. 150) model quoted earlier would seem to suggest just a single winter, but there are two reasons to question whether such brief stays at overwintering sites would have been the norm. First, the implicit assumption is that they overwintered in the same kind of semisubterranean houses that they used in both Alaska and in the Eastern Arctic, which required considerable labour and material to construct, and in non-migratory contexts, such houses are assumed to have had average use-lives of perhaps five to ten winters (McGhee, 2009a, pp. 85–86). Performing the additional labour of constructing new semisubterranean houses every autumn year after year, on top of the extreme exertion that would have been involved in migrating all their possessions by sled or umiak from their previous winter site, including rafters for the houses, would have been extremely onerous.

3.3.3 Gaining Knowledge of Unfamiliar Regions

Second and more important, such brief stays at each overwintering location would have meant advancing far into new regions potentially possessing differing characteristics without having thoroughly explored them beforehand to learn about their resources, which would have been risky. With respect to sea mammals, the differing coastline configurations and bathymetries of the Canadian Arctic Archipelago mean that the abundance of sea mammals can vary dramatically from locality to locality, as does the timing of freeze-up and break-up of the sea ice. The local availability and abundance of land mammals, especially caribou, varies enormously as well. The Qitdlarssuaq example shows that Inuit sometimes did migrate in this rapid fashion, but the catastrophic outcome for many of those migrants from choosing poor overwintering locations indicates how risky doing so could be. Therefore, it seems likely that some years may have devoted not to migrating but to “frontier scanning” (Hickey, 1984, p. 18), by remaining additional winters at one location and making extended hunting trips from there at different times of the year in the direction in which they eventually intended to migrate, to learn about the availability and distribution of resources before actually moving there. Frontier scanning would have been even more important if they wanted to implement the strategy Friesen (2016, p. 684) proposed:

In particular, migrants would have initially sought locations that allowed accumulation and storage of large amounts of food to be consumed in sedentary winter villages

Given that both the people and their dogs needed to be reliably fed throughout the winter, this strategy indeed seems likely, but identifying such locations in the unfamiliar lands ahead of them would have been difficult without first having made exploratory expeditions at different times of year. Thus, there clearly would have been benefits for a migrating group to reside in a single location for several years between major moves to allow enough time for thorough frontier scanning.

3.3.4 Number and Duration of Migration Events

However, we should not build our model based on the logistics of just a single migrating group. It seems generally accepted that the migration was not a single discrete event, but rather a series of such events that may have taken place over a period of years, decades or even centuries. Friesen (2016, p. 681) observes that the process may have ranged from “a tightly constrained series of a few population movements within a short period” to “continued periodic immigration from Alaska,” while Mason (2020, p. 325) likens the movement to a “continuing trickle of people eastward.” In the absence of direct evidence from single-component sites created by migrating groups, we must rely on inferences derived from ethnography to get a sense of how the groups of migrants would have been organized. On that basis, Friesen (2016, pp. 680–681) infers that “most migrating groups were probably relatively small, ranging in size from perhaps 15 to 50 people.” From ethnographic research conducted from the nineteenth century through the middle of the twentieth century, we know that some Inuit societies temporarily broke up into groups as small as 15–20 during the summer (e.g. Balikci, 1984, p. 422) but, because they normally aggregated into larger groups at other times of the year, it seems likely that the maximal band size for each migrating group would necessarily have been larger, closer to the 50 number.

One relevant implication of this model of several sequential groups of Thule migrating eastward is that later migrants may have been able to take advantage of the winter house sites created by earlier groups, making the amount of labour necessary for each major move less. But this model also raises the possibility that earlier groups may have ceased migrating and settled down as soon as they came to promising uninhabited locations in the western Canadian Arctic. Subsequent migrating groups following in their footsteps, especially if they were migrating precisely to avoid conflict or to escape being an underclass, may therefore have been motivated to pass quickly through these newly settled regions to reach new territories they could call their own.

3.4.1 The Initial Migrant Population

A final indirect but vital source of information concerning the timing and pace of the Thule migration is demography – more specifically, the number of Thule individuals who commenced migrating in Alaska, and the eventual size of their descendants’ populations in Arctic Canada and Greenland. The number of Thule people who undertook the eastward migration is of obvious interest, and there is evidence that the number was quite small, based on what turns out to be a very limited amount of genetic diversity among their Inuit descendants in the Eastern Arctic. Inuit genetic diversity can be used to make direct inferences about the size of the Thule migrant population because the Inuit are entirely descendants of those Thule migrants. The Eastern Arctic had previously been occupied by the Paleo-Inuit, and by the eleventh century, Norse Vikings had colonized part of Greenland, but genetic studies supported by circumstantial archaeological evidence demonstrate that there was no intermarriage, and thus, there are no descendants of the Paleo-Inuit or Norse to be found amongst the Inuit (Lopopolo, Børsting, Pereira, & Morling, 2016, p. 701; Park, 2016; Raghavan et al., 2014; Tackney et al., 2019; Unkel et al., 2022). On that basis, Marchani et al. (2007, p. 283) conducted a pilot study using mitochondrial haplogroup data to conclude that the likely “Eastern Inuit female effective population size” of the Thule migrants was between 25 and 50 women. Mason (2020, p. 344) reports that “The demography of the Western Arctic remains too poorly documented to infer the role of population in the Thule migration,” but the geographic region most often suggested as the source of the Thule population migrating into Canada encompasses Northwest Alaska and, before nineteenth century reductions due to introduced diseases (Milan, 1980, p. 20; Spencer, 1984, p. 326), the Inuit population of that region according to Burch (2005, p. 7) was approximately 4,700, so even a millennium earlier, it plausibly could have been the source of up to a few hundred Thule migrants, but likely not more than that.

3.4.2 Constraints on Population Growth

Models of the Thule migration have all implicitly or explicitly postulated “rapid population increase” (e.g. Morrison, 2009, p. 76) among those initial migrants and their descendants, but the rate of population growth would have been constrained by several well-documented factors. The earliest data collected from Inuit societies in the Eastern Arctic noted long lactation periods and long intervals between births, anywhere from 3–5 years (Scheffel, 1988, pp. 40, 42; van de Velde OMI et al., 1993, pp. 31, 33). Another factor that may have contributed to long intervals between births was the suppression of fertility by high-energy activities such as travel, and on an annual basis, the timing of births indicated that conceptions were fewer in seasons when groups had to be highly mobile (Burch, Jr, 1988, p. 19; Lupo, 2021; van de Velde OMI et al., 1993, p. 33). It would therefore seem probable that the additional mobility requirements of migration, over and above those already entailed in their annual pattern of mobility, would have exacerbated the suppression of fertility. A final constraint on population growth was the fact that hunting accidents produced a high mortality rate among young and middle-aged men (Damas, 1975, p. 412; Scheffel, 1988, p. 42; van de Velde OMI et al., 1993, p. 31). This was documented ethnographically among groups who lived in areas with which they were very familiar. Such losses might have been an even more significant factor for hunters hunting in unfamiliar regions and might have affected their ability to migrate rapidly while simultaneously supporting a growing population.

3.4.3 Inuit Populations at the Time of Contact

With the understanding that these ethnographically documented factors must also have influenced Thule Inuit demography, over the next few centuries, those Thule migrants’ Eastern Arctic descendants nonetheless flourished and populated a vast area, although our knowledge of the ultimate sizes of the Inuit populations prior to the effects of introduced diseases is not precise. Our earliest information comes from Greenland, where in the late eighteenth century, the total Inuit population is estimated to have been around 8,000 (Hamilton & Rasmussen, 2010, p. 45). The earliest estimates of the Inuit population in Arctic Canada come from somewhat later but indicate a population totalling more than 9,000 in the nineteenth or very early twentieth century (Arima, 1984, p. 460; Balikci, 1984, p. 429; Boas, 1888, p. 425; Damas, 1984, p. 400; Mary-Rousselière, 1984, p. 445; Mathiassen, 1927, p. 285; Robinson, 1944; Saladin d’Anglure, 1984, p. 480; Smith, 1984, p. 349; Taylor, 1984, p. 508). Thus, it is possible that by AD 1800, the descendants of those few hundred Alaskan Thule Inuit who had entered the Eastern Arctic numbered up to 17,000.

3.4.4 Modelling Thule Inuit Population Growth

We can combine these population estimates – the initial number of Thule migrants, and the number of their descendants by around AD 1800 – to model the growth of the population over time on the assumption that no dramatic bottlenecks intervened. The rate of growth is largely dependent on when the migration commenced, and Figure 4 models it based on four entirely arbitrary entry dates: AD 1250, AD 1200, AD 1100, and AD 1000. The calculated average annual rates fall between 0.737 and 0.506%, which are plausible given what we know about Inuit fertility and the exigencies of migration. These are averages across the entire Eastern Arctic – presumably some regions would have supported higher rates, while others would have been constrained to lower rates. The useful chronological insight provided by this exercise comes from what it reveals about how many individuals were available to settle in the Eastern Arctic soon after the migrants arrived because that population had to be large enough and in the right locations long enough to produce the archaeological record known from each region. On stylistic grounds, we infer that the initial Thule movement beyond the westernmost portion of the Canadian Arctic included groups who migrated towards the northeast. Beyond Victoria Island, examples of artefact styles that would not be out of place in Alaskan sites, and which are therefore inferred to be “early” in the Eastern Arctic, are found primarily in sites along Lancaster Sound and on the islands north of it (Friesen, 2016, p. 684; Friesen & Arnold, 2008, p. 535; Morrison, 1999, p. 142). Researchers have inferred that Labrador and the coasts of Hudson Bay were entered by the Thule somewhat later (Friesen, 2017, p. Loc 3122; Friesen et al., 2020, p. 145; Maxwell, 1985, p. 297; McGhee, 2009a, p. 84; Whitridge, 1999, pp. 133–134; 2016, p. 831), which has implications for demography, although how much later is disputed. Movement into these regions may have commenced by AD 1350 (Fitzhugh, 2020, p. 95) or AD 1400 (Fitzhugh, 2019, p. 54; Kaplan & Woollett, 2016, p. 857), but others put it even later, at AD 1500 (McGhee, 2009a, p. 84; Ramsden & Rankin, 2013, p. 304).

Figure 4

Hypothetical Thule Inuit population curves predicated on an initial migrating population of 300 that grew to 17,000 by AD 1800, with start dates of AD 1000, AD 1100, AD 1200, and AD 1250.

The one reliably dated event early in Thule history is the encounter no later than AD 1364 between Inuit and Norse somewhere in West Greenland. For that encounter to have taken place there and then, a viably large Thule population had to have already migrated all the way from Alaska to North Greenland, and then continued south past the difficult barrier of Melville Bay, where the inland ice cap reaches the sea. As shown by the population curves in Figure 4, if a migration of around 300 Thule Inuit from Alaska commenced around AD 1250 (e.g. Friesen, 2020, p. 226; McGhee, 2009a, p. 81), then the people who met the Norse in the middle of the fourteenth century were part of an entire Eastern Arctic Thule Inuit population that would have grown to just 700 individuals; if they had started migrating 150 years earlier, in AD 1100, the AD 1364 population would have been twice as large, and an earlier migration would have produced an even more substantial fourteenth-century population.

Is the smallest AD 1364 total Eastern Arctic population based on these models – 700 people – enough to account for the evidence? As a thought experiment, we might imagine that of the original 300 or so migrants starting out from Alaska in AD 1250, perhaps half would have settled in the western portion of the Canadian Arctic, around the Mackenzie Delta, Coronation Gulf, and Victoria and Banks Islands on the rationale that some Thule Inuit would have chosen to settle in those very productive regions, supported by archaeological evidence suggesting that Thule occupations were continuous following initial occupation (Ross & Friesen, 2022). The rest migrated further to where many settled along Lancaster Sound and the islands to its north. The ones who continued beyond there entered Greenland and moved south to where they encountered the Norse. The population growth that would have taken place over the 16 or 17 decades since they left Alaska would have resulted in approximately 350 Thule in the west, perhaps 250 in the eastern part of the Canadian Arctic, and 100 in Greenland. The 250 in the eastern part of the Canadian Arctic would have then needed to provide the groups that would head south to Labrador and Hudson Bay. It is impossible to reject this highly speculative scenario with certainty, but these numbers seem improbably small to have colonized all those regions and created the substantial archaeological record attributed to the early Thule. So, from a demographic perspective, a scenario involving a larger fourteenth-century Thule Inuit population in the Eastern Arctic is more consistent with the evidence.