Palaeo-Eskimo Migrations into Greenland: The Canadian Connection

Full text

INTRODUCTION
The present paper is dedicated to Jørgen
Meldgaard – our ‘grand old man’ in Arctic Archae-
ology.
It focuses on the Palaeo-Eskimo migrations into
Greenland – the easternmost part of the Inuit world
– and the palaeo-demography of this island. Since
the beginnings of Arctic anthropological and
archaeological research, scholars have been fasci-
nated by the initial human colonisation of the
Eastern Arctic (Meldgaard 1996), and hypotheses
have changed like the weather conditions in these
rough environments. However, new archaeological
data from Greenland have been presented during
the last decade and it is time to address the ques-
tions associated with the dynamics of migrations
and demography again. We begin our venture with
a presentation of the main conclusion of this paper:
The North Water Polynia in Nares Strait between
Ellesmere Island and the Thule area attracted
groups of Independence I, Saqqaq, Pre-Dorset,
Early Dorset, and Late Dorset, but only two Palaeo-
Eskimo groups migrated beyond the Polynia and
became ’circum-Greenland populations’: Indepen-
dence I/Saqqaq (2500 BC cal.) and the Early Dorset
group: Greenlandic Dorset (800 BC cal.).
NEW SOURCES, NEW APPROACHES
Archaeological fieldwork during the last ten to fif-
teen years has added substantial new information
about the prehistory of the High Arctic gateway to
Greenland (e.g. Schledermann 1990, Sutherland
1996, Appelt & Gulløv (eds.) 1999). Similarly, discov-
eries in old collections – for example in the Eigil
Knuth Archive – provide us with new data
(Grønnow & Jensen 2003). It is rewarding to com-
bine results of research from both the Canadian
and Greenlandic part of the High Arctic: the models
of cultural chronology of Ellesmere Island as out-
lined by Schledermann for the central part (1990)
and by Sutherland for the northern part (1996) and
Appelt’s research on Late Dorset in Inglefield
Land/Thule Area (Appelt 2003). These studies sug-
gest that the region around the North Water Polynia
and Nares Strait formed a coherent settlement area
and that we here find a key to Greenland’s cultural
history.
In our analyses we have screened the published
radiocarbon dates. Thus the absolute dates
referred to in this paper are primarily on bones of
terrestrial mammals or birds and local plant mate-
rial. The dates are calibrated (OxCal 3,8 2002), and
in this connection we draw attention to the prob-
lems related to a certain plateau on the calibration
curve: radiocarbon dates of the interval 2350 BP –
2580 BP cannot be calibrated with precision, and
consequently the cultural sequence of the time
span from 800 BC to 400 BC cal. must be described
without the basis of absolute dates.
Finally, results of ongoing chaîne opératoire
analyses using methods such as refitting of lithic
inventories and dynamic technological classifi-
cations (Sørensen, this volume), produce new
insight into the technological traditions and social
aspects of the Palaeo-Eskimo cultures and their
relations.
Through detailed analysis and documentation of
each step in the sequence of lithic production from
raw material acquisition to production, use and dis-
card of artefacts, this approach intends to explain
the behavioural and mental schemes, which struc-
ture a material culture. One of the present authors
(Sørensen) has demonstrated that the chaîne
59
Palaeo-Eskimo Migrations into Greenland:
The Canadian Connection
Bjarne Grønnow and Mikkel Sørensen
1
1. SILA – The Greenland Research Centre at the National Museum of Denmark.

60
Core
Axe preform
Axe
Axe
Blank Broad thin flake* Curved flake* Curved flake*
Axe production flakes
Bifacial flakes
Burin (BLT) Side blade Scraper Side scraper
Burin preform 1
Burin preform 2
Base of burin
(BLT) Scraper
Side scraper
Blank Blank Blank
Bifacial flakes
Side blade
* The artefact is reconstructed
Flakes
Blanks
Preform 1
Preform 2
Tool
Rejuvenated
tool
Discarded
tool
Blank
Bifacial knife Harpun point
Bifacial preform
Knife
Base of knife
Tabular core
Thin biface
Scraper preform
Bifacial flakes
Harpun
point
End
Scraper
Side
scarper
Side
blade
Burin preform ?
Bifacial
flakes
Blank
Burin (BLT)
Axe production flakes
Blank
Axe preform
Blank Blank
Bifacial preform Bifacial preform
Curved flake Thin flake
Axe
Bevelled points
(Lanse point
and Knife)
Bifacial
knife
Channel flakes
(From fluting)
Expedient
flake knife
Blank/
nodule
Preform 1
Preform 2
Tool
Rejuvenated
tool
Discarded
tool
Bifacial flakes
Axe Burin (BLT)
Harpun
point
Side
scarper
End
Scraper
Burin (BLT)
Bifacial
knife
Bifacial
knife
Independence II
Dorset I

opératoire of the Saqqaq lithic production is funda-
mentally different from that of the Early Dorset.
There is no direct technological continuity between
these cultures in Greenland (Fig. 6).
On the other hand, a comparison of the lithic
technologies of Independence II and Dorset I,
including key elements such as burin technology,
blade production and morphology, modification of
blades and choice of raw material (morphology and
quality), provides substantial evidence to the con-
clusion that these two cultures cannot be distin-
guished from one another (Fig. 1). Consequently,
they can be united under the term ‘Greenlandic
Dorset’ allowing alternative interpretations of the
second Palaeo-Eskimo migration into Greenland.
THE NORTH WATER CULTURAL SEQUENCE
Today, the High Arctic area on the present national
border between Canada and Greenland is per-
ceived as remote and marginal. However, archae-
ology suggests that the region around Smith Sound
and Kane Basin formed a coherent settlement area
(McCullough 1989; Schledermann 1990). The abun-
dant resources of the North Water Polynia and
adjacent smaller polynias periodically attracted
human settlement from the earliest Palaeo-Eskimos
to historic Inuit. On a large scale, the area formed
the ‘Gateway to Greenland’ (Arneborg & Gulløv
(eds.) 1998), and it holds a key to the cultural history
of the island (Fig. 2).
Schledermann (1990), McCullough (1989) and
Sutherland (1996) have presented outlines of the
cultural sequences of the Canadian side of the
area. However, on the Greenland side of the Nares
Strait evidence is rather limited. In particular this is
true for the Palaeo-Eskimo cultures due to the size
and complexity of the habitable areas, and due to
the rather sporadic research effort during the past
50 years. Large areas along the Nares Strait remain
unsurveyed. Except for the excavations by Holtved
(1944, 1954), relatively few archaeological investi-
gations can be listed, viz. Knuth’s three excavation
seasons on Old Nuulliit in the last three decades of
the 20th Century (Knuth 1978), his investigations in
Hall Land in the 1960s (Grønnow & Jensen 2003:34-
55), Diklev and Madsen’s surveys from Qanaaq to
Inglefield Land (1992), Appelt’s, Gulløv’s, Kapel’s and
Grønnow’s excavations of Late Dorset in Hatherton
Bay (Appelt & Gulløv (eds.) 1999), Appelt’s and
Kapel’s archaeological survey from Dundas to Cape
York (Appelt et al. 2001), and Andreasen’s survey in
Washington Land (2000).
The point of departure of the present approach is
the assumption that the cultural sequence from
Ellesmere Island is mirrored on the Greenland side
of the Nares Strait.
61
Fig. 1a. Concepts of the lithic production in Independence II and Dorset I.
When the two concepts are compared in terms of their different stages, from the choice of blank morphology
and raw material to the rejuvenated and discarded tool, the following should be noticed:
1) Both groups produce the same functional tool types excluding arrowheads and drills. 2) One difference to
be noted is that bevelled points of killiaq (metamorphosed slate) are produced in central West Greenland by
Dorset I. 3) In both concepts, axes, burins, and bifacial tool types such as knives and harpoon points are made
from small natural blanks and not from large flakes or serial-produced blanks. 4) Axes in Independence II are
made from dolorite, while axes, burins and bevelled points in Dorset I are made from killiaq. 5) All other tools in
the two groups are made from local microcrystalline quartzes (agate, crystal and chalcedony in West Green-
land and chert in High Arctic). 6) Burins in both groups, are produced in a similar manner and have similar
identical morphology (long distal ends and precisely asymmetrical shaped bases). The burins are finished and
rejuvenated by grinding, and the cutting edges are ground in a variety of fashions. Retouch and spalling of the
burin edges are only seen very rarely during burin rejuvenation. 7) Bifacial knives and harpoon points have a
similar production process and identical morphologies in both groups. Diagnostic attributes in this production
are large notches on the tool bases, and the production of fluting flakes detached from the right-angled base
edge. 8) End scrapers and side scrapers are unifacial, and made from flakes. The characteristic flaring mor-
phology of the end scrapers is also typical to both groups.
To conclude: the range of functional tool types, the overall strategy of the two concepts, the technical produc-
tion, as well as detailed morphological attributes of all tool types – are similar or identical for the two groups
(see also Figure 1b). The conceptual and technological similarities, in Dorset I and Independence II, leads to
the conclusion that Dorset I and Independence II are one and the same culture. The authors therefore bring
these the two groups together under the term “Greenlandic Dorset”.
▲

In the following, we discuss the cultural
chronology around the North Water Polynia based
on the cultural chronology of Ellesmere Island (Fig.
3).
2.600/2.500-900/800 BC
The radiocarbon dates of four Saqqaq sites on
Ellesmere Island are situated within a wide time
frame from around 2500 to 1500/1400 BC cal.,
whereas the Independence I dates are concen-
trated within a few centuries from 2600/2500 to 2200
BC cal. (Table 1) (Schledermann 1990:343; Schleder-
mann & McCullough 1992:13; Sutherland 1989,
1996). We interpret this as evidence of (sporadic)
visits by Saqqaq groups from West Greenland
during almost a millennium, whereas the Indepen-
dence I mainly settled the area during a pioneer
phase, which was probably followed by sporadic
visits until 1600/1400 BC as documented in Eureka
Upland by Sutherland (1996).
Only a few screened dates exist for the Pre-
Dorset in the area. Dates from Ridge Site suggest a
sporadic presence from 1800-900/800 BC cal.
(Schledermann & McCullough 1992: 13). Based on
Sutherland’s preliminary publication of the Riven-
dell Site, Eureka Upland (1996:281-284), we suggest
that this site represents a Pre-Dorset settlement,
thus adding three screened dates between
1400/1200 and 1000/900 BC cal. to the record.
Accordingly, we agree with Schledermann that the
Ellesmere side of the Strait was sporadically settled
62
Blade core Blades Knives (retouched blades)
Blade production Tools
Blade core Blades Knives (retouched blades)
Blade production Tools
Independence II
Dorset I
Fig. 1b. The concept of blade production in Independence II and Dorset I.
When the two blade concepts are compared great similarities are seen. The proportions and especially the
morphology of the blade core during the step of blade production (the facetted preparation of the platform and
the small angle between platform and front) display evidence of an identical technology. Furthermore, blades
have generally a similar morphology and dimensions within the two groups. In both groups, blades are trans-
formed into knives by retouching the proximal ends of the blades and thereby creating tanged blades.

by three early Palaeo-Eskimo cultures during the
time span 2600/2500-900/800 BC cal. Independence
I and Saqqaq overlap, and during a later stage,
Saqqaq and Pre-Dorset overlap, but this does not
necessarily mean that they occupied the area at
the same time.
Is there evidence for a similar cultural sequence
in the archaeological record of Greenland? Inde-
pendence I is represented in Hall Land and
Warming Land (Solbakken and Memnon Site) and
probably on Old Nuulliit (Knuth 1978; Grønnow &
Jensen 2003:36-45, 51-54). Sites of the Saqqaq Cul-
ture have been located and preliminarily recorded
around Qaanaaq, in Hatherton Bay and in Inglefield
Land (Diklev & Madsen 1992). Information from pre-
liminary surveys indicates that Saqqaq sites are
quite large and complex.
The Pre-Dorset, which is relatively well repre-
sented on the Ellesmere Island side (Schledermann
1990: 91-126), is harder to detect and identify in the
archaeological material from Northwest Greenland.
Moreover, it must be emphasised that the lithic
chaînes opératoires of Independence I and Pre-
Dorset show many similarities. However, based on
a technological evaluation of the lithic material, in
particular the burin re-sharpening technique, raw
material preferences, the end blades, and micro
blades, we have identified a single Pre-Dorset
occupation from Greenland viz. Knuth’s material
from Feature no. 13, a midpassage-dwelling from
Solbakken, Hall Land (Grønnow & Jensen 2003:42-
43). The feature is situated on the 19 m beach ridge,
some 40 m from the linear arrangement of the Inde-
pendence I dwellings at the 21 m ridge, and this
supports the assumption that the feature is younger
than the Independence I settlement at the site. So
far, this is the only evidence of Pre-Dorset on the
Greenland side of the Strait, but probably more
sites will be identified through future surveys and
re-analyses of the archaeological assemblages.
63
Site Lab.nr. Material Cultural
affiliation
C-13 C-14, bp
(uncal)
C-14, BC
cal (one
sigma)*
Reference
Bight site,F.1 TO1555 Muskox
bone
Saqqaq - 3610+/-60 2120-
1880
(Schledermann
1990)
Bight site,F.2 TO1556 Muskox
bone
Saqqaq - 3840+/-70 2460-
2200
(Schledermann
1990)
Topo site TO992 Willow Saqqaq - 3420+/-60 1870-
1630
(Schledermann
1990)
Beacon Site ? Willow Saqqaq - 3160+/-50 1495-
1400
(Schledermann/
McCullough
1992)
Lakeview,F.30 TO993 Willow Independence I - 3940+/-70 2570-
2300
(Schledermann
1990)
Campview, F2 TO994 Willow Independence I - 3990+/-70 2620-
2340
(Schledermann
1990)
Daylight site I-12339 Willow Independence I - 25 3790+/-95 2440-
2040
(Sutherland 1989)
Ridge site,F.2 TO1638 Willow Predorset - 3440+/-50 1880-
1680
(Schledermann
1990)
Rivendell Beta35676 Terrestrial
Bone ?
Predorset - 20 2800+/-70 1050-
840
(Sutherland 1996)
Rivendell Beta35677 Terrestrial
Bone ?
Predorset - 20 2980+/-70 1370-
1100
(Sutherland 1996)
Rivendell Beta35678 Terrestrial
Bone ?
Predorset - 20 3070+/-70 1430-
1250
(Sutherland 1996)
Phalarope site KIA-18940 Caribou
bone
Early Dorset -19.20
+/-
0,10‰
2610+/-25 810-
785
(Damkjar 2003)
Baculum site GEO6072 Plant
material
Early Dorset - 2780+/-
140
1130-
800
(Schledermann
1990)
* Dates are calibrated b
y
Oxcal 3,8
(
2002
)
Table 1. Published radiocarbon dates on terrestrial materials from Palaeo-Eskimo sites on Ellesmere Island.

800-400 BC cal: The ‘ plateau’ on the calibration
curve
The Early Dorset related radiocarbon dates from
sites in the High Arctic emerge within the plateau
on the calibration curve (2580 BP-2350 BP = 800 BC-
400 BC cal.). As mentioned above, this has the con-
sequence that we cannot date Early Dorset pre-
cisely. However, a screened date from Feature 15 at
Phalarope Site, Somerset Island, far west of the
Nares Strait is situated very precisely at 810-785 BC
cal. (KIA18940, one standard deviation). Thus far, it
is the only date of Greenlandic Dorset (or the ‘Tran-
sitional Period’ as described by Damkjar (2003:231))
before the plateau. A single date derived from local
plant material comes from the Early Dorset Bacu-
lum Site on Ellesmere Island: 1130-800 BC cal.
(Schledermann 1990:175-183, 343; Schledermann &
McCullough 1992:13). If we accept this broad date,
the first Dorset groups in the area did not overlap
Saqqaq or Independence I, but could have over-
lapped the Late Pre-Dorset.
We must turn to lithic studies to enlighten the
Dorset problem in the Nares Strait area. We sug-
gest that this ‘transitional period’ should be decon-
structed based on a re-evaluation of the lithic
inventories from the eleven sites included in Schle-
dermann’s Transitional Period (1990:127-173).
According to our interpretation, some of the assem-
blages fit into an Early Dorset inventory (e.g. Long
House Site (Dorset Ridge Component) and Grave
Rib Site (Schledermann 1990:139-142, 159-165)),
some fit into Pre-Dorset (e.g. Cape Hunter Site
(ibid:150-154)) and some seem to consist of chrono-
logically mixed assemblages.
Consequently, we do not find substantial evi-
dence for a cultural transition between Early and
Late Palaeo-Eskimo cultures in the North Water
Polynia Area. Rather, Early Dorset groups origi-
64
Greenland
Ellesmere
Inglefield Ld.
Washington Ld.
Hall Ld.
Peary Ld.
Nares Strait
North
Water
Fig. 2. Satellite image of High Arctic Greenland and Ellesmere Island showing the extent of the North Water
during summer (July 5
th
, 2002). Image: ev25621_Greenland. A2003186.2235, Canadian Meteorological Centre.

nating in other parts of Arctic Canada expand into
the polynia area around 800 BC cal. and the subse-
quent centuries. Analyses of the technology and
morphology of burins and burin-like tools seem to
be a fruitful approach to separate these different
Early Dorset groups (Fig. 4, see also Schledermann
1990:344-356; 1996: 55). From this point of view,
Greenlandic Dorset assemblages characterised by
elongated burin like tools with completely ground
working edges are found within the Nares Strait
region, including Eureka Upland (Sutherland 1989,
1996, 2003), Washington Land (Winstedt Site, Level
2 (Andreasen 2000: 86)) and the Thule area (Diklev
& Madsen1992:24-26). Assemblages from other
Early Dorset Groups characterised by short burins
with retouched and slightly ground edges and
ground distal ends are found on eastern Ellesmere
Island (e.g. Skraelling ASTt 5 and Baculum Site
(Schledermann 1990:154-159, 175-183)) and in
Washington Land (Winstedt Site, Level 1
(Andreasen 2000:85); Tent Bay (Peel & Frykman
1975)). It must be emphasised that the distinction
between ‘Greenlandic Dorset’ and ‘Other Early
Dorset Groups’ in the Nares Strait area is to some
extent based on published data, and detailed tech-
nological analyses of the assemblages are neces-
sary to elaborate upon this division. Only one of the
Early Dorset groups, the Greenlandic Dorset,
expanded beyond the North Water Polynia, i.e. one
migration route went around Peary Land, another
went south across the Melville Bay to West Green-
land.
400 BC-700 AD cal.
The entire High Arctic was completely depopulated
for a remarkably long period – 1100 years – while
Palaeo-Eskimo populations thrived, interacted and
developed in other regions of the Arctic.
700 AD-1.300 AD cal.
The Palaeo-Eskimo cultural sequence of Ellesmere
Island is completed by a remarkable colonisation
phase, the Late Dorset. Groups ascribed to this
65
Independence I
Saqqaq
Pre-Dorset
2500 - 900/800 cal. BC
A
Greenlandic Dorset
Other Early Dorset
800 - 400 cal. BC (plateau)
B
Late Dorset
700 - 1300 cal. AD
C
Fig. 3. Palaeo-Eskimo sites at ‘The Gateway to Green-
land’ (the Nares Strait area). Compiled from pub-
lished literature and reports on file (until 1998).

66
A.
B. C.
D. E.
Fig. 4. Burins from Washington Land and Hall Land, in Northwest Greenland.
A. Burin from Feature 13 on Solbakken, Hall Land. The burin was made on a flake (bulb of percussion is on the
base of the burin). The edge is made by a burin spall and the rejuvenation was done by way of thinning the dis-
tal end by a serial detachment of small flakes from the burin edge. The morphology and the rejuvenation of this
burin are typical for the Early Pre-Dorset Culture.
B. & C. Burins from Winstedt site. Burins are made on small blanks. The distal ends are ground on both sides,
while the working edges are produced by a lateral unifacial retouch followed by a minimal grinding of the thin
edges. The distal ends have very characteristic 40-degree angles between the top and the lateral side (oppo-
site to the working edge). The bases are worked precisely into an asymmetrical morphology with large lateral
notches using pressure retouch.
D. & E. Burins are produced similar to B. & C. but have longer distal ends. E. has a broken distal end. D. has a
working edge shaped by a short burin spall and a unifacial lateral retouch. This edge has no grinding. It is
therefore probable that D. was discarded during rejuvenation.
The technology and morphology of B., C., D., and E. points towards the existence of Early Dorset groups in
Washington Land that are distinct from Greenlandic Dorset.

culture migrated to the North Water Polynia
around 700 AD (Schledermann 1990:201-281;
Sutherland 2003:200-206), and according to Ap-
pelt’s interpretation, established a continuous set-
tlement here for the following six centuries (Appelt
2003). Archaeological remains such as dwelling
structures, ‘longhouses’, hearth rows, as well as
artefacts from this period are rather impressive.
The settlement in the area forms the easternmost
and northernmost node in a complex network of
Late Dorset groups linking together societies
distributed over the entire Eastern Canadian Arc-
tic.
After 1200 AD the Late Dorset is confronted with
two other agents on the High Arctic scene, the Inuit
of the Thule Culture, which continued to expand
beyond the Polynia area, and Norse hunting and
trading parties from the south (Schledermann 1993;
Arneborg & Gulløv 1998; Sutherland 2000; Appelt
2003).
67
Fig. 5. Areas of distribution of the Eastern Arctic archaeological cultures
during seven phases. Note that the time spans are different on the seven
maps and that the outlines of the areas of distribution are rough guide-
lines and not exact or static borderlines between the cultures.

PALAEO-ESKIMO MIGRATIONS INTO GREENLAND
Our interpretation of the cultural sequence of the
Gateway to Greenland has implications for the way
we interpret the history of migrations into Green-
land and the subsequent expansions and contrac-
tions of the Palaeo-Eskimo population areas. In the
following, we divide the early cultural history of
Greenland into seven phases (Fig. 5).
2.500-1.900 BC cal.: the initial migrations
The earliest screened radiocarbon dates of human
presence in Greenland are overlapping each other
within one standard deviation. Thus, the initial
migration into Greenland resulting in a circum-
island population took place within a remarkably
short time span around 2.500 BC cal.
The Independence I migration into High Arctic
Canada and Greenland probably represents a
western migration of Denbigh Culture groups, but
the relations to Saqqaq is not as clear. The lithic
chaîne opératoire of Saqqaq is to a large extent
determined by intensive exploitation of the killiaq
sources on the Nuussuaq Peninsula, West Green-
land (see below). The question of the origin of the
Saqqaq culture is still open to debate, but the early
date of this culture suggests that it arrived as a part
of the initial colonisation of the Eastern Arctic and
developed in central West Greenland.
Turning to the High Arctic, Knuth’s surveys and
finds from Hall Land to Dove Bugt provide us with in-
formation on the character of the Independence I
settlement pattern. During six decades of intensive
archaeological investigations, Knuth recorded 51 In-
dependence I sites containing 244 features in this
vast area (Grønnow & Jensen 2003:328). Only a few
dwellings leave the impression of having been used
over a longer period of time rather than just one or a
few seasons, and only three sites, Pearylandville,
Adam C. Knuth Site, and Deltaterasserne (Grønnow
& Jensen 2003:333-338), can be classified as aggre-
gation sites. Furthermore, we must take the time
perspective into consideration. According to the
screened dates, the features on these 51 sites were
constructed and used within a time span of a mini-
mum of 170 years and a maximum of 630 years
(ibid.:328-333). Thus, we get an even stronger indi-
cation of the extremely low settlement density in the
area of the Independence I culture. We interpret the
evidence as follows. The Canadian High Arctic is-
lands together with northernmost Greenland should
be seen as one vast settlement area (cf. Sutherland
1996), occupied by a single or just a few highly mo-
bile groups of Independence I families. Periodically,
they established a seasonal movement cycle within
a specific area including an aggregation site. After a
few years, they abandoned the area and moved on
to new hunting grounds, which could provide them
with prey, in particular musk oxen, arctic char and
other resources. In this new area, a similar com-
plete settlement system was established. A vulnera-
ble settlement pattern like this could be maintained
for only a few centuries.
In contrast, the archaeological remains from
the Saqqaq Culture reflect a development towards
complex and more stable settlement and re-
source utilisation patterns. This included a high
population density and a formation of territories
linked via exchange and trade networks, probably
68
Fig. 6. The concept of lithic production in Saqqaq and Dorset I.
The Saqqaq lithic technology is primarily based on import of killiaq from the peninsula Nuusuaq, Uummannaq
District, West Greenland. Flakes are serial produced from large bifacial cores serving as pre-forms for the
small tool types (arrowheads, awls, burins, scrapers, harpoon points, saws, and flake knives). Large bifacial
tools (axes, knives, lance points) are made from pre-formed blanks.
When the Saqqaq concept is compared to the Dorset I concept (described in fig. 1a), differences can be ob-
served in all stages of production including raw material preferences and morphology of the functional tool
types. The only similarity is the production of side scrapers and the rare bevelled points from Late Saqqaq in
the Sisimiut area (not shown on the Saqqaq figure).
If a transition from Saqqaq to Dorset I took place, there should be a technological development within the
Saqqaq towards the Dorset I concept which among others should yield: 1) The use of single blanks for the pro-
duction of the small tool types, instead of serial production from bifacial cores. 2) A decline in arrow point pro-
duction. 3) A change from spalled burins to burins made from blanks with ground edges. Since none of these
developments are seen in Greenland, the many differences between the Saqqaq and the Dorset I concept
must be seen as evidence for discontinuity between these two cultures.
▲

69
Flake knife
?
Blank Bifacial core
Bifacial preforms
Flake blank
Flake blank
Arrrow point
preform Awl
preform
Burin
preform 1
Burin preform 2
Harpun point
preform 1
Harpun point
preform 2
Flakes from
bifacial core
Bifacial flakes
Burin spall
Bifacial flakes
Bifacial flakes
Burin spalls
Axe Knife Lanse
point
Arrow
point
Awl
Burin
End
scraper
Side
scraper
Harpun
point Saw
Axe
Knife
Knife
Lanse
point
Arrow
point
Awl
Burin
End
scraper Side
scraper
Burin
Saw
Harpun
point
Burin
Tabular core
Thin biface
Scraper preform
Bifacial flakes
Bifacial flakes
Harpun
point
End
Scraper
Side
scarper
Side
blade
Burin preform ?
Bifacial
flakes
Blank
Burin (BLT)
Axe production flakes
Blank
Axe preform
Blank Blank
Bifacial preform Bifacial preform
Curved flake Thin flake
Axe
Bevelved points
(Lanse point
and Knife)
Bifacial
knife
Channel flakes
(From fluting)
Expedient
flake knife
Blank/
nodule
Preform 1
Preform 2
Tool
Rejuvenated
tool
Discarded
tool
Axe Burin (BLT)
Harpun
point
Side
scarper
End
Scraper
Burin (BLT)
Bifacial
knife
Bifacial
knife
Dorset I
Saqqaq
Preform 1
Preform 2
Tool
Rejuvenated
tool
Discarded
tool
Blank/
nodule/
core

matching the historic Thule Culture (Gulløv 1997).
At least in central West Greenland, this complex
network of the Saqqaq Culture existed for a re-
markably long time span of about 1500 years (e.g.
Møbjerg 1986; Grønnow 1996; Jensen 2000; Appelt
2003).
1900-800 BC cal.: Pre-Dorset visitors in the
North
The Greenland side of the Nares Strait was occa-
sionally visited by Pre-Dorset groups during this
period as indicated by finds at Solbakken (see
above). Thus, a potential for contact between this
culture and Saqqaq existed around the North Water
Polynia. The organic components of the two cul-
tures are almost similar, but their lithic technologies
are different in certain respects, for example in the
choice of raw material, micro blade and burin pro-
duction and burin re-sharpening techniques. In
spite of possible sporadic external contacts,
Saqqaq maintained a complex and remarkably
standardised lithic industry without traces of Pre-
Dorset influence. Saqqaq may in fact be considered
as an isolated group with a remarkably conserva-
tive lithic industry characterised by a distinct lithic
chaîne opératoire developed from the exploitation
of one specific lithic source: the killiaq (metamor-
phosed slate) and the grinding stones (sandstone)
from ‘Slibestensfjeldet’ on Nuussuaq in northern
West Greenland (Jensen & Petersen 1998;
Sørensen & Pedersen 2005).
At the end of this period – about 1000/800 BC cal.
– we observe a contraction of the settlement area
of the Saqqaq Culture, which ultimately is limited to
central Western Greenland (Gotfredsen & Møbjerg,
2005). The complex system of territories and
exchange networks broke down.
800-400 BC cal.: Greenlandic Dorset
All early dates of Dorset in Greenland are within
the ‘plateau’ on the calibration curve. However, as
mentioned above, a single screened date from the
Canadian Archipelago is earlier and before the
plateau supporting our assumption that Green-
landic Dorset is a result of an Early Dorset expan-
sion into High Arctic Canada around 900/800 BC.
Similar to the Independence I and Saqqaq, one
and a half millennia earlier, the initial Dorset en-
compasses the entire island, but unlike the pio-
neers, the migrants enter into a cultural landscape
with signs of earlier human presence, tent rings,
mid-passage dwellings, caches, middens and arte-
facts.
Due to the ‘plateau’ on the calibration curve, it
cannot be determined if there was a chronological
overlap between the newcomers of Greenlandic
Dorset and the last group of Saqqaq. The chaîne
opératoire analyses of the lithic production cer-
tainly do not lend support to the idea of continuity
between the two cultures (Fig. 6). However, one
could argue for a minor Dorset ‘fingerprint’ on the
material culture of Saqqaq, viz. the bevelled
ground end blades from the latest Saqqaq phase
at Nipisat in Sisimiut Municipality (Møbjerg
1998).
From a broader cultural historical perspective,
Greenlandic Dorset in the extreme High Arctic re-
presents a brief moment in time, probably a few
episodes within the time span of the ‘plateau’, 800-
400 BC cal. In total, Knuth registered only 14 sites
including 73 features in Peary Land and adjacent
areas. These archaeological traces probably re-
flect one or a few migrations across northernmost
Greenland and brief visits by some groups of fami-
lies, who utilised the resources for a season or
two (Grønnow & Jensen 2003:331-332, 338-339).
Greenlandic Dorset sites become more frequent as
one moves further south into North East Greenland
(Andreasen 1997, 1998), culminating in a strange
phenomenon which is difficult to explain based on
our present knowledge, the huge aggregation site
found and documented by Knuth at Kap St.
Jacques, Île de France, with more than 400
dwelling features (Grønnow & Jensen 2003:278-
296).
400 BC – 1 AD cal.: contraction of the Green-
landic Dorset
As part of the general process of abandonment of
the High Arctic, the Greenlandic Dorset disappears
from the northernmost areas of the island. Ab-
solute dates are rare, but there is clear evidence
of Dorset presence in other parts of Greenland
during the centuries before the Birth of Christ
(Jensen 2004:144). Similar to Saqqaq, the lithic
technology of this culture seems to be quite con-
servative.
70

1 AD – 700 AD: Total abandonment
We loose all traces of human presence in High
Arctic Canada and on the entire island of Greenland
during these six centuries, and we must conclude
that this reflects a total abandonment of these vast
areas while other Palaeo-Eskimo societies – the
Middle Dorset Culture – thrived in other parts of
Arctic Canada (e.g. LeBlanc 2000; Desroisiers &
Rahmani 2003).
700-1200 AD: The Canadian connection: Late
Dorset
Similar to the previous ones, the Late Dorset expan-
sion into the uninhabited High Arctic around
600/700 AD was very fast. As mentioned above the
migration resulted in the foundation of several ‘core
areas’ forming an exchange network covering the
entire Eastern Arctic. However, the north-eastern-
most node of the network around the North Water
Polynia – dated from ca. 700 to 1300 AD – did not
expand further Northeast or South. About ninety
percent of the island remained uninhabited for
another three or four centuries until the Norse set-
tlement in Southwest Greenland was established
around 1000 AD. Norsemen from the settlements
were in sporadic contact with Late Dorset groups in
the High Arctic at a quite early stage (Sutherland
2000; Appelt 2003). Nevertheless, only a fraction of
the island was populated during the following five
centuries.
1200/1400 AD to historic times: The Thule migra-
tion
The easternmost branch of the Thule expansion
from Alaska reached the North Water Polynia
around 1200 AD (e.g. McCullough 1989; McGhee
2000; Appelt 2003). According to the radiocarbon
dates, the initial migration went southwards
crossing the Melville Bay into West Greenland and,
ultimately, around Cape Farewell (Gulløv 1997). A
few centuries later, groups of Thule families
migrated from the Nares Strait area via Peary Land
into north easternmost Greenland (Knuth 1951;
Grønnow & Jensen 2003:210-217). During the 15th
Century, Peary Land was occasionally utilised as
musk oxen hunting grounds by Thule groups, who
had their winter dwellings further south beyond
Nordostrundingen (Grønnow & Jensen 2003:339-
340). Following this brief episode, the extreme High
Arctic of Greenland was abandoned, and ultimately,
we observe a process of population contraction
similar to the demographic developments during
the Palaeo-Eskimo époque. However, one funda-
mental difference must be borne in mind, the Thule
demography was influenced by contacts to Late
Dorset in the North Water Polynia and the Norse in
Southwest Greenland.
CONCLUSIONS AND PERSPECTIVES
We summarise our interpretation of Greenland’s
cultural history as follows: In line with High Arctic
Canada it is characterised by discontinuity, regional
abandonment and repeated re-settlement of key
regions. The entire island even remained com-
pletely abandoned for centuries.
Greenland’s cultural history and palaeo-demog-
raphy is linked with Canada via the North Water
Polynia, which at times attracted different groups of
the Independence I, Pre-Dorset, Early Dorset and
Late Dorset from different regions in Arctic Canada.
Only two migrations, the initial migration of the
Independence I/Saqqaq about 2,500 BC and the
expansion about 900-800 BC by one of the Early
Dorset groups, which we term ‘the Greenlandic
Dorset’, resulted in full ‘circum-Greenland’ human
colonisation during limited time spans.
The transition from the Early Paleo-Eskimo Cul-
tures to the Dorset Culture did not take place at the
North Water Polynia or in other parts of Greenland
for that matter. Rather, as indicated by archaeolo-
gical evidence at the Gateway to Greenland, spo-
radic and extreme northward migrations of different
Pre-Dorset and Early Dorset groups from Canada,
and Saqqaq groups from West Greenland resulted
in discontinuous settlement of both sides of the
Nares Strait. The Late Dorset migration, however,
resulted in a more continuous settlement system
around the North Water Polynia lasting for six cen-
turies. The Late Dorset groups never settled beyond
this area in Greenland.
An interesting overall palaeo-demographic pat-
tern is observed, viz. the initial migration phases of
the Independence I/Saqqaq, the Greenlandic
Dorset, and the Thule, which consist of remarkably
fast and comprehensive demographic processes
including the entire island. Within a quite limited
71

period of time, a few generations, contraction
processes are apparent in all three cases. The High
Arctic is abandoned or only sporadically visited
after a few generations, and only in Low Arctic
Greenland – in particular central West Greenland –
more continuous or long-term settlement is seen in
the form of two millennia of Saqqaq and probably
eight centuries of Greenlandic Dorset settlement.
The overall demographic patterns are charac-
terised by these radical shifts between expansions,
formations of linked local and regional groups,
demographic concentrations in a few regions and
eventually the abandonment of the entire island.
Naturally, further factual evidence is needed to
support and elaborate our model of the palaeo-de-
mography of Greenland. Series of radiocarbon dates
on terrestrial material are desirable, but for some pe-
riods they are not very helpful. In particular, the de-
bated ‘transitional phase’ and the Early Dorset needs
further analyses using archaeological methods. As
we have seen, archaeological approaches, such as
chaîne opératoire studies applied to lithic invento-
ries, have led to the emergence of new points of
view. We hope that this paper will contribute to a
fruitful academic discussion about the prehistoric
connections between Greenland and Arctic Canada.
72
REFERENCES
Andreasen, Claus
1997 Independence II in Northeast Greenland: Some new
aspects. – In: Gilberg, R. & Gulløv, H. C. (eds.) Fifty Years
of Arctic Research. Anthropological Studies From Green-
land to Siberia. Publications of the National Museum,
Ethnographical Series 18: 23-32.
1998 NEWland: settlement patterns, social organisation and
economic strategies at a High Arctic polynya. – In:
Arneborg, J. & Gulløv, H. C. (eds.): Man, Culture and Envi-
ronment in Ancient Greenland. Danish Polar Center Publi-
cation 4: 198-212.
2000 Paleo-Eskimos in Northwest and Northeast Greenland. –
In: Appelt, M., Berglund, J. & Gulløv, H. C. (eds.), Identi-
ties and Cultural Contacts in the Arctic. Danish Polar
Center Publication 8: 82-92.
Appelt, Martin
2003 De sidste palæoeskimoer. Nordvest Grønland i perioden
800-1300 e.v.t. – Unpublished Ph.D.-thesis. Institut for
Forhistorisk Arkæologi, Aarhus Universitet: 165 pp.
Appelt, Martin & Gulløv, H. C. (eds.)
1999 Late Dorset in High Arctic Greenland. Final report on the
Gateway to Greenland Project. Danish Polar Center Pub-
lication 7: 83 pp.
Appelt, Martin, Kapel, Hans & Zimmermann, Marit
2001 Arkæologisk rekognoscering i Thule Kommune, som-
meren 1998. – Unpublished report on file at the National
Museum of Denmark: 62 pp.
Arneborg, Jette & Gulløv, H. C. (eds.)
1998 Man, Culture and Environment in Ancient Greenland.
Danish Polar Center Publication 4: 212 pp.
Damkjar, Eric
2003 Transitional Occupations of southeastern Somerset
Island, Nunavut. Études/Inuit Studies, 2003, 27(1-2): 213-
238.
Desrosiers, Pierre & Rahmani, Noura
2003 Analyse spatiale et architecturedu site dorsétien GhGk-
63 (Kuujjuarrapik, Nunavik). – Études/Inuit/Studies, Vol.
27(1-2), 2003: 131-154.
Diklev, Torben & Madsen, Bo
1992 Arkæologisk berejsning i Thule, 1991. Unpublished report
on file at Avernersuup Katersugaasivia/Thule Museum:
42 pp.
Gotfredsen, Anne Birgitte & Møbjerg, Tinna
2005 Nipisat – a Saqqaq Site in Sisimiut, Central West Green-
land. – Meddelelser om Grønland, Man & Society
31.
Grønnow, Bjarne
1996 The Saqqaq Tool Kit – Technological and Chronological
Evidence from Qeqertasussuk, Disko Bugt. – In: Grøn-
now, B. & Pind, J. (eds.): The Paleo-Eskimo Cultures
of Greenland. New Perspectives in Greenlandic
Archaeology. Danish Polar Center Publication 1: 17-
37.
Grønnow, Bjarne & Jensen, Jens Fog
2003 The Northernmost Ruins of the Globe. Eigil Knuth’s
Archaeological Investigations in Peary Land and Adja-
cent Areas in High Arctic Greenland. Meddelelser om
Grønland, Man & Society 29: 403 pp.

Gulløv, Hans Christian
1997 From Middle Ages to Colonial Times. Archaeological and
ethnohistorial studies of the Thule Culture in South West
Greenland 1300-1800 AD. Meddelelser om Grønland, Man
& Society 23: 501 pp.
Holtved, Erik
1944 Archaeological Investigations in the Thule District, I-II.
Meddelelser om Grønland 141(1-2): 184 pp.
1954 Archaeological Investigations in the Thule District, III:
Nûgdlît and Comer’s Midden. Meddelelser om Grønland
146(3): 135 pp.
Jensen, Jens Fog
2000 West-Greenland Palaeo-Eskimo Raw Material procure-
ment Elucidated by Geological Analysis. – In Appelt, M.,
Berglund, J. & Gulløv, H. C. (eds.): Identities and Cultural
Contacts in the Arctic. Danish Polar Center Publication 8:
71-76.
2004 Dorset-kulturen.- In: Gulløv, H.C. (ed.) Grønlands Forhis-
torie. Gyldendal: 142-172.
Jensen, Jens Fog & Petersen, Erik Brinch
1998 Raw Material Distribution – Social Space, Social Interac-
tion. – Acta Borealia 15(2): 139-152.
Knuth, Eigil
1951 Et umiak-fund i Pearyland. Nationalmuseets Arbejdsmark
1951: 77-87.
1978 The “Old Nugdlit Culture” Site at Nugdlit Peninsula, Thule
District, and the “Mesoeskimo” Site below it. Folk 19-20:
15-47.
LeBlanc, Sylvie
2000 Middle Dorset (1900-1100 BP). Regional Variability on the
Island of Newfoundland and in Saint-Pierre et Miquelon.
– In: Appelt, M., Berglund, J. & Gulløv, H. C. (eds.), Identi-
ties and Cultural Contacts in the Arctic. Danish Polar
Center Publication 8: 97-105.
McCullough, Karen M.
1989 The Ruin Islanders. Early Thule Culture Pioneers in the
Eastern High Arctic. Canadian Museum of Civilization,
Archaeological Survey of Canada, Mercury Series, Paper
141: 347 pp.
McGhee, Robert
2000 Radiocarbon date and the timing of the Thule Migration.
In Appelt, M. Berglund, J. and Gulløv, H. C. (eds.): Identi-
ties and Cultural Contacts in the Arctic. Danish Polar
Center Publication 8: 181-191.
Meldgaard, Jørgen
1996 The Pioneers: The Beginnings of Paleo-Eskimo Research
in West Greenland. In: Grønnow, B. & Pind, J. (eds.), The
Paleo-Eskimo Cultures of Greenland. New Perspectives
in Greenlandic Archaeology. Danish Polar Center Publi-
cation 1: 9-16.
Møbjerg, Tinna
1986 A Contribution to Palaeo-Eskimo Archaeology in Green-
land. Arctic Anthropology 23(1-2): 19-56.
1998 The Saqqaq culture in the Sisimiut Municipality Eluci-
dated by two new Sites Nipisat and Asummiut. – In:
Arneborg, J. & Gulløv, H. C. (eds.) Man, Culture and Envi-
ronment in Ancient Greenland. Danish Polar Center Publi-
cation 4: 98-118.
OxCal
2002 OxCal, Version 3.8 from the Internet. Copyright by C.
Bronk Ramsey.
Peel, John S. & Frykman, Peter
1975 Archaeological observations in southern Washington
Land, July-August 1975. – Report and accompagnying
chert flakes and bone fragments presented to National
Museum, October 1975. Unpublished report on file at the
National Museum of Denmark: 26 pp.
Schledermann, Peter
1990 Crossroads to Greenland: 3000 Years of Prehistory in the
Eastern High Arctic. Komatik Series No. 2, The Arctic
Institute of North America of The University of Calgary:
364 pp.
1993 Norsemen in the High Arctic? – In: Clausen, B.L. (ed.):
Viking Voyages to North America. The Viking Ship
Museum, Roskilde, Denmark: 54-66.
1996 Voices in Stone. A Personal Journey into the Arctic Past.
– Komatik Series No. 5. The Arctic Institute of North
America of the University of Calgary: 221 pp.
Schledermann, Peter & McCullough, Karen
1992 Buchanan Bay Archaeological Project. The 1991 Field
Season. Interim Report. Beacon Site. Report on file at
The Arctic Institute of North America, Calgary, Alberta: 13
pp.
Sutherland, Patricia D.
1989 An Inventory and Assessment of the Prehistoric Archae-
ological Resources of Ellesmere Island National Park
Reserve. Winnipeg, Environment Canada, Parks Service,
Microfiche Report Series 43.
1996 Continuity and Change in the Paleo-Eskimo Prehistory of
73

Northern Ellesmere Island. I: Grønnow, B. & Pind, J.
(eds.) The Paleo-Eskimo Cultures of Greenland. New Per-
spectives in Greenlandic Archaeology. Danish Polar
Center Publication 1, s. 271-294.
2000 The Norse and Native Norse Americans. – In: Fitzhugh,
W. W. & Ward, E. I. (eds.) Vikings. The North Atlantic
Saga. Smithsonian Institutions Press, Washington &
London: 238-247.
2003 Variability and Change in Palaeo-Eskimo Architecture: A
View from the Canadian High Arctic. Études/Inuit Studies,
2003, 27 (1-2): 191-212.
Sørensen, Mikkel & Pedersen, Buck Kristoffer
2005 Killiaq kilden på Nuussuaq: på rekognoscering efter
Saqqaqkulturens redskabssten. Tidsskriftet Grønland, no.
2-3.
74