A cross-check of radiocarbon dates from Stone Age sites i Northern 
Sweden
Ekholm, Therese
http://kulturarvsdata.se/raa/fornvannen/html/2015_227 
Fornvännen 2015(110):4 s. 227-240
Ingår i samla.raa.se
A cross-check of radiocarbon dates from
Stone Age sites in Northern Sweden
By Therese Ekholm
Ekholm, T., 2015. A cross-check of radiocarbon dates from Stone Age sites in Nor -
thern Sweden. Fornvännen 110. Stockholm.
This study compares the results from radiocarbon measurements performed on
charcoal and burnt bone from the same contexts in order to assess the reliability of
the two materials for dating. The study deals with seven Mesolithic sites in Norr-
land and Dalecarlia, an area where datable organic material is difficult to find. The
bone samples have been chosen for the study and the charcoal samples are mainly
from previous work. The study shows that both materials are suitable for dating as
long as they are sampled correctly and with knowledge of the errors that can occur.
Therese Ekholm, Department of Archaeology and Ancient History, Uppsala University.
Box 626, SE–751 26 Uppsala
therese.ekholm@arkeologi.uu.se 
In the late 1990s a new method made it possible
to carbon date burnt bones (Lanting & Brindley
1998). This has changed our chronology and under-
standing of the arrival of both humans and ani-
mals to Norrland, the northern three quarters of
Sweden. It has been shown that bone is more reli-
able for dating archaeological sites than wood
because of its negligible intrinsic age (Lanting et
al. 2001, p. 249; Hedman 2009, p. 5), which is the
age of the sampled organism. Much radiocarbon-
dated wood from long-lived tree species suffers
from the Old Wood Effect. A similar problem is
curation or storage age, when objects have been
kept in use or circulation for long periods (Brock
& Dee 2013, p. 41). In addition burnt bones found
in a sunken feature are more likely to represent
human action and connect people’s activities to a
particular site, while charcoal can result natural-
ly from, for example, a forest fire at any time dur-
ing a site’s formation process.
One of the problems of dating prehistoric sites
in Norrland is how to argue for an association
between radiocarbon dates and artefacts. Due to
thin vegetation, prehistoric sites are often found
right beneath the top soil, which complicates the
interpretation of the stratigraphy. The assemblages
often represent palimpsests of material from seve-
ral periods making the relation of dates, features
and artefacts difficult. Furthermore, due to the
decomposition of organic materials in the acid
soil, artefacts made of wood and unburnt bones
are rarely found in sunken features (Borg et al.
1994, p. 97 f; Björdal 1999, p. 120 f; Christensson
1999, p. 172 f) where burnt bones do survive
(Lanting & Brindley 1998, p. 7). For decades the
only method available for dating these sites was
radiocarbon dating of charcoal, especially the
sites excavated during the expansion of hydro-
electric power in the 1940s through 70s (Iregren
& Ekman 1984). Most sites are chronologically
mixed, and so even at sites with stratigraphy such
as Rappasundet, Döudden/Varghalsen, Garase-
let (Knutsson 1993, p. 25; Bergman 1995a, p. 143
ff; 111 ff) it is often difficult to distinguish which
feature belongs to which period. At all of these
inland sites it is therefore important to do a care-
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 227
ful contextual analysis of the provenance of arte-
facts and samples for radiocarbon analyses.
This study deals with two problems of the
chronology which can be elucidated by systematic
radiocarbon dating of burnt bones.
1. Due to the difficulty of identifying the
sources of charcoal found at a prehistoric
site, it is uncertain what event is dated.
This study investigates if there is any 
difference in the dates of burnt bone and
charcoal taken from the same feature,
and if the charcoal therefore is connected
to the activity in which the feature was
involved.
2. The intrinsic age of the charcoal sample
is often unknown, which may be one rea-
son why dates obtained from charcoal
often seem misleading. This study inves-
tigates if there is a systematic error here
that can be controlled for by using bone
dates as a baseline.
Bones are however not unproblematic. Bones ly-
ing in the soil can become contaminated by ab-
sorbing bicarbonate from the surrounding sedi-
ments and the results may thus show a later date.
This does not affect bones burnt at more than
600 °C, provided that the bones are burnt before
they end up in the soil (Lanting & Brindley 1998,
pp. 2, 6; Olsen et al. 2008, p. 792). Bones from
animals that live in the sea or feed on animals from
marine habitats may yield dates exceeding their
actual age, due to the reservoir effect. It is therefore
important to choose terrestrial over marine animals
for analysis, as well as animals with a terrestrial diet
over those feeding on marine sources (Lanting &
van der Plicht 1998, p. 151; Lanting et al. 2001, p.
249 f; Hedman 2009, p. 5). This does not, however,
affect burnt bones as much as it does unburnt bo-
nes. Because structural carbonate, which is used
when dating burnt bones, originates from the whole
diet (Lanting & Brindley 1998, pp. 6).
Previous studies
Studies comparing radiocarbon dates from burnt
bones and other burnt materials have been under -
taken several times and with different results, such
as in Lanting et al. 2001, Olsen et al. 2008 and Ol -
sen et al. 2013. In 2001 radiocarbon dates obtai-
ned from burnt bones were compared to measure-
ments on charcoal from the same context. In some
cases the charcoal looked older, which was ex-
plained by reference to a higher intrinsic age of
the sample (Lanting et al. 2001). 
The famous Early Bronze Age oak-coffin grave
from Egtved in Denmark contained both an in-
humation and a cremation in a bark bucket. Ol -
sen et al. (2013) have compared the radiocarbon
date of the jawbone from the cremated individ-
ual, a five-year-old child, with the dendro-chrono-
logical date (1370 BC) of the log coffin. Rather
unusually, they converted the dendro date back-
wards into an uncalibrated BP date by means of
the IntCal09 radiocarbon curve (Reimer et al.
2009) and got 3054±16 BP. Two combined sam-
ples from the child’s jawbone yielded a date in
3127±20 BP. The difference between the coffin
and the bone was 73±26 radiocarbon years, the
bone being older.
Olsen et al. suggest that the child was cremat-
ed in connection with the funeral of the Egtved
woman, because remains of the funeral pyre were
found in the bark bucket. No such remains, they
argue, would have been there if the deposition of
the child’s remains represented reburial of a long-
dead individual (Olsen et al. 2013, p. 31 f). The
charcoal from the cremation found with the burnt
bones were not radiocarbon dated, as these re-
mains were assumed to produce identical re sults.
This was because of the intrinsic age of the wood
used as fuel which, according to the authors, was
transferred to the bones by means of CO2. To
prove this, the radiocarbon date of the burnt bone
was instead compared to the dendrochronology
on the oak coffin. The results showed that the burnt
bone was older than the coffin after the reservoir
effect was excluded. This outcome was interpreted
to mean that the burnt bone had been contami-
nated by the intrinsic age of the wood used as fuel
for the cremation (Olsen et al. 2013, p. 33).
This comparison of radiocarbon dating and
dendrochronology is not unproblematic as the
authors assumed that the two deaths were con-
temporaneous. The bucket with the cremated
material could have been unearthed intact from
its first burial place and placed in the oak coffin
as a secondary burial event. The cremated bone
228 Therese Ekholm
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 228
is, furthermore, an unreliable sample because it
derives from the spongy part of the jaw bone. It
is not advisable to use spongy bone tissue for
radiocarbon dating as its cavities can trap other
materials such as radiocarbon date charcoal and
roots (Borg et al. 1994, p. 97). Solid bones are
more suitable for dating. Since all the bones in
the bucket derive from a child and children’s
bones are overall very spongy, the samples were
not suitable to use as a foundation to build argu-
ments concerning the reliability of radiocarbon
dating.
In a 2008 study, Olsen et al. compare radio-
carbon dates on burnt bones with radiocarbon
dates on other materials from the same contexts,
such as charcoal and pitch, and also with dendro-
chronology. The dendrochronological case is near -
ly the same as in the 2013 study of the Egtved
grave, but with another bone sample from the
same jaw. The pitch was used to seal funerary
urns and these dates were compared with the
dates of the burnt bones inside the containers. In
total six urns were dated and in all cases, except
one where the pitch was older, the results were
consistent. The compared bone and charcoal
samples showed close correspondence since one
comparison differed only by one hundred years
and two by two hundred years, the charcoal being
the oldest in all cases. The interpretation of the
results is that the wood was affected by the Old
Wood Effect, an older intrinsic age. In these cases
no transferred intrinsic age by CO2 obviously
affected the bones as it may have done in the stu dy
of the Egtved grave. A test with several samples
taken from different parts of a skeleton with dif-
ferent degrees of combustion, from black, charred
bones without burn cracks to white bones with
distinct burn cracks, was done and resulted in
various dating results. The samples exposed to
higher temperatures showed earlier dates, from
3576±29 BP on charred bones to 3756±28 BP on
cremated bones (Olsen et al. 2008, p. 795 f). Since
there are two different ways of dating bones,
whether they are burnt or unburnt, a suggestion
is to never use charred bones for the reason that
they are neither quite burnt or unburnt. None of
the studies recounted above strengthen the case
for the idea that the Old Wood Effect might be
able to transfer from wood to burnt bones.
Contract excavations for the E4 highway north
of Uppsala in 2002–2004 occasioned a compara-
tive study (Guinard 2006) of radiocarbon dates
for a number of Stone Age contexts including the
sites Stormossen, Bålmyren and Postboda 2, and
hazelnut shells, seal bones as well as charcoal. The
aim was to study whether any systematic pattern
in the reservoir effect could be discerned. It was
assumed that the seals would have acquired a
reservoir. The dates of the bones from four con-
texts at Stormossen were, as expected, consis-
tently too early. The radiocarbon measurements
of the seal bones from the Mesolithic sites Bål -
myren and Postboda 2, however, gave later dates
than the hazelnut shells from the same contexts,
which Michel Guinard does not discuss further.
In the contexts with the overall earliest dates the
reservoir effect was the greatest, but the climate,
the salinity of the water and local deviations all
contribute to the size of the reservoir effect. Also,
a larger study would be needed before any major
conclusions about the reservoir effect could be
reached.To sum up the basis for a study of the
credibility of radiocarbon dates on charcoal and
burnt bones, in this paper I will consider the fol-
lowing: 1) Bones burnt at more than 600 ˚C, 2)
From terrestrial animals, 3) Avoiding spongy parts
of bones, 4) Controlling for variations in the
burning temperature.
Aims
The aims of this study are to investigate whether
there is a systematic difference between radiocar-
bon dates performed on a) charcoal, and b) burnt
bone from the same context on Stone Age Sites in
northern Sweden. The contexts from which the
samples are taken have to be sealed, which means
that they have been covered right after they were
used. At least there has to be no sign of later
activity disturbing the feature. As mentioned
earlier, there is some difficulty in finding datable
material at these sites. Therefore, it is important
to have several different dating methods avail-
able. If it can be shown that the charcoal and the
bone from secure contexts date the same event,
then charcoal dates from such contexts, from
analyses performed during the decades before
burnt bone could be dated reliably, can be used
with some confidence. This is important for the
229A cross-check of radiocarbon dates from Stone Age sites
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 229
great body of data that we have from inland sites
in northern Sweden (Biörnstad 2006).
By dating selected samples of burnt bone from
contexts which have already been dated with
charcoal, we can begin an interesting discussion
of the usefulness of the early analyses from nor -
thern sites. Besides enhancing the number of dat ed
contexts, the dates can one day be used to inves-
tigate if construction traditions visible in the fea-
tures at a site can be linked to any systematic dif-
ferences in space and time. If so, then undated fea -
tures could be assigned a preliminary date using
a typological strategy. It presumes that the bone
samples, as discussed, have a high quality, are solid,
hard-burnt, and are highly crystallised with no
organic components retained (Ol - sen et al. 2008,
p. 792). The animal species must also be known to
be one without a reservoir effect (Persson 1999, p.
28; Hedman 2009, p. 5). Dating of burnt bone
will, in such a case, be able to provide valuable in-
f ormation about sites that might otherwise not
be reliably dated.
This study is a pilot study for a larger PhD
pro ject addressing the establishment of animal
spe cies in Norrland as the Weichselian ice cap
withdrew. With radiocarbon dates of burnt ani-
mal bone from separate archaeological contexts,
not only the date of habitation, but also the tim-
ing of the species occupation of the landscape
will be identified. Only those species that are
linked to certain cultural traditions and therefore
were deposited on Mesolithic sites in Norrland
are discussed, as they are the only animals visible
in the archaeological record. By studying the fau-
nal bone assemblages from the sites the trajectory
of human adaptation to new environments and
variations in people’s cultural practices of animal
resource exploitation can be studied. By dating
bones from these sites, a higher chronological
resolution for cultural practices, illustrated by
the stone technology, will also be established.
The data provides further new information that
will be merged into recent studies of when peo-
ple came and adjusted to the new landscape, and
where they came from (Sørensen et al. 2013). 
Based on radiocarbon dates the next step is to
explore the possibility of building a typology of
features identified on Mesolithic sites in the north,
where some animal species are associated with
certain characteristics of construction, and ulti-
mately connect these two with the lithic ma te-
rial. The latter would, in a later phase, allow an
analysis of cultural ecology during the Mesolithic
in Norrland. 
Selection of sites 
The analysed sites, in Norrland, Sweden, dis-
cussed in the article, were selected primarily for
their chronology and sealed contexts, where
charcoal and burnt bones were found together.
The three exceptions are Kangos, and feature 8
(A8) at Garaselet, where bones were found in asso-
ciation with the feature, and Lappviken, where
no charcoal has been dated. At Lappviken the
lithics serve to indicate the time of occupation.
All sites are Mesolithic, 8000–4000 cal BC, and
have been selected for geographical coverage. 
The sites
The objects under study were initially bone assem-
blages from nine dated Mesolithic sites (fig. 1).
Seven sites are in inland Norrland: Aareavaara,
Kangos, Dumpokjauratj, Garaselet, Lappviken,
Rastklippan and Högland. Two are in Dalecarlia:
Orsand and Tjärna. Upon further study I excluded
Aareavaara (Raä Pajala 1276–1277) and Rastklip-
pan (Raä Sorsele 1000) due to the high fragmen-
tation rate of the bones, with fragments too small
for radiocarbon dating. The bone samples dated
for the study weighed a minimum of 1 gram.
Kangos 
The site at Kangos (Raä Junosuando 22) was
located at River Lainioälven in Norrbotten. The
artefact finds are characterised by quartz worked
both with platform and bipolar technology
(Östlund 2004, p. 9.). On the site only one fea-
ture (A1) was found, a hearth pit with finds in the
pit and in association with it. Bones were found
in the hearth and down-slope from it, in square
x354/y195 and in the southwest part of square
x354/y196. Most of the lithics outside the hearth
occurred in squares x353/y195 and x354/y196.
Rain water may have moved bones and lithics
down-slope from the hearth (Östlund 2004, p. 6
f). The bones were analysed by Leif Jonsson in
2004. Of reindeer (Rangifer tarandus) there was a
mandible, a talus and the middle toe bone (pha-
230 Therese Ekholm
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 230
Fig. 1. The studied Mesolithic sites. By Michel Guinard.
lanx II) from a dewclaw. These are the small toes
on the front hoof that are attached to the smaller
metacarpals (telemetacarpalia) behind the main
metacarpal. Of pike (Esox lucius) there was the
right-hand part of the cranium (cleithrum) and
left mandible (dentale). Olof Östlund’s interpre-
tation of the site (2004, p. 9) is that it was a tem-
porary hunting/fishing site where tools of quartz,
greenstone and slate were made and at least one
reindeer and one pike were cooked.
Dumpokjauratj 
The site at Dumpokjauratj in Lapland (Raä Arje-
plog 1568) had a cooking pit (A1), two hearths (A2–
3; Bergman 1995b, p. 3), two hearth pits (A5–6),
a waste pit (A7; Bergman 2003a, p. 3) and an
accumulation of charcoal (A4; Bergman, 2001, p.
5). It is located by Lake Dumpokjauratj which is
part of the same water system as Lakes Gublijau-
re, Lullebådne, Kakel and Hornavan. The lithics
are, as at Kangos, both platform and bipolar
(Bergman 2003a, p. 6; 2003b, p. 5 f). The two
hearth pits had been cleaned out several times, and
so Ingela Bergman’s interpretation (2003a, p. 3)
is that the site was a temporary dwelling site used
on multiple occasions. The features of interest
here are A4 and A5. Only charcoal was found in
what was left of A4, but in its reddish top layer –
disrupted by uprooted trees, erosion and site
preparation – a bone scatter was found and inter-
preted as belonging to A4. In A5 burnt bones,
flakes, fire-cracked stones and charcoal were
found (Bergman 2003a, p. 14; 2003b, p. 14). 
The bones were analysed by Mats Eriksson
231A cross-check of radiocarbon dates from Stone Age sites
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 231
and Maria Vretemark (2003). They are mainly
elk (Alces alces; Eriksson ms.) and reindeer, but
also a few fragments of fox (Vulpes vulpes), beaver
(Castor fiber), indeterminate bird and pike (Vrete -
mark 2003, p. 1). 
Garaselet
The site at Garaselet in Västerbotten (Raä Jörn
79) has yielded Mesolithic to Middle Neolithic
dates, with a variety of features in the form of
cooking pits, hearths, pits and more recent struc-
tures such as a forge, a fire pit and features from
Sami activities (Sundqvist ms.). The site is on the
south bank of the Garaselet inlet in River Byske -
älven (Knutsson 1993, p. 17). Features A6 and A8
were of special interest. A6 was a cooking pit with
plenty of bones, fire-cracked stones, charcoal and
flakes mainly of porphyry. A8 was a hearth with
charcoal and bones (Sundqvist ms.). The bones
were analysed by Margareta Backe in 1990 and
consisted of elk, beaver, reindeer, bear (Ursus arc-
tos), otter (Lutra lutra), black-throated diver (Ga-
via arctica, Sw. storlom), goose (Anserinae sp.), div-
ing ducks (Aythyinae sp.), pike, carp (Cypr inidae
sp.), bream (Abramis brama, Sw. braxen), ide (Leu-
ciscus idus, Sw. id), perch (Perca fluviatilis, Sw. ab-
borre) and grayling (Thymallus thymallus, Sw. harr).
Lappviken
The site at Lappviken in Västerbotten (Raä Jörn
66) is near Garaselet on the northern shore of the
Lappviken inlet in the same river. No published
excavation report is available, but Kjel Knutsson
(1993) has discussed the site in a paper.The bones
from Lappviken were analysed by Margareta
Backe and consist of elk, beaver, indeterminate
bird and pike. Only bone lists are available and no
detailed analysis has been done of the material.
Högland 
The site at Högland in Västerbotten county,
southern province of Lapland (Raä Dorotea 181)
was divided into three areas, with two features,
A1–2. A1 was a pit in area 1, with artefacts of
quartz and quartzite and burnt bones. The latter
are too small for radiocarbon dating. A2 was a pit
in area 3 with burnt bones and fire-cracked stones.
Jan Melander and Jan-Erik Wallin (1981, p. 5 ff)
interpreted the site as having been used several
times, where the two features in areas 1 and 3 re-
presented temporary sites where humans hunted
their prey, cooked and made tools. Area 2 was
also a temporary site, but much more ephe meral
than areas 1 and 3.The bones from Högland had
not been analysed by the time of the publication
of this paper.
Orsand 
The site at Orsand in Dalecarlia (Raä Leksand
2001) had a cooking pit A255 and two bone fea-
tures A272 and A268. It was on the eastern shore
of Lake Siljan and was interpreted as a temporary
dwelling site (Guinard 2013). I analysed the bo-
nes from Orsand in 2013, but the results have not
yet been published. The represented species are
indeterminate mammal, indeterminate deer (Cer -
vidae sp.), carp (Cyprinidae sp.) and perch.
Tjärna
The site at Tjärna in Dalecarlia (Raä Borlänge
40) had a Mesolithic feature (A140) which was
damaged by a trench on the western side. In the
destroyed part, fire-cracked stones and slag were
found, probably from an iron manufacturing site
nearby. In the intact eastern part of the feature
were burnt bones and fire-cracked stones (Sand-
berg 2001). The bones were analysed by Barbro
Hårding in 2001 and consisted of cattle (Bos tau-
rus), pig (Sus domesticus), sheep (Ovis aries) and
elk.
Selection of bone samples
I chose seven already dated Mesolithic sites for
this investigation. The burnt bones selected for
dating should, as far as possible, be of the same
species. The different contexts in which the bones
were found should also be similar to one another.
Before the bones were radiocarbon dated the pre-
vious osteological analyses had to be verified to
ensure that the determination in the list corre-
sponded with the chosen bone. When there was
no osteological analysis I did one my self, but only
of the bones to be dated. All radiocarbon meas-
urements were performed with AMS at the Ång -
ström Laboratory in Uppsala and calibrated with
OxCal v4.2.4 (Bronk Ramsey 2013), using the Int-
Cal13 dataset (Reimer et al. 2013). All dates in tab.
1 and 2 are uncalibrated. From Kangos a middle
232 Therese Ekholm
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 232
Tab. 1. Radiocarbon meas-
urements done for this
study. 
       
   
 
       

       
 
       

       
 
       
 
       
  
       
  
       
  
       
  
       
 
       

       

       

       

       
 
 14C analyses of charcoal done for this study, where no charcoal had previously been dated 
toe bone from a dewclaw of reindeer was se lected.
No samples of charcoal had previously been dated,
and so charcoal from A1 was also selected, al-
though the collected charcoal is somewhat unre-
liable according to the report (Östlund 2004, p. 8).
The charcoal was sent to wood identification and
identified as willow (Salix sp., e-mail Ulf Strucke
2013). The chosen bone sample, which was the
only bone fragment determined to species that
weighed enough, was found in square x354/ y195/
z249, 42 masl (F18), in association with A1. Two
bone samples from A1 (»F20» and »anl. 1») had
previously been dated, one of which, F20, was
from the same square as F18 (Östlund 2004, app.
3). The previously analysed samples date to 8720
±60 and 8555±65 BP (Östl und 2004, app. 6:1; cf.
tab. 2). From Dumpokjauratj, two bone samples
from feature A4 were selected. One of them was
a distal part of the upper toe bone (phalanx I) of
elk. The other sample was a distal part of the meta-
carpus of elk. Previously, two charcoal samples of
pine (Pinus sect. Sylvestris) from the same context
had been dated to 8000±80 and 8010±75 BP.
From A5 a burnt reindeer bone had previously
been dated to 8450 ±55 BP (Silvermuseet dnr
1998/015). Three charcoal samples – one from
poplar (Populus sp.) and two from willow or
poplar – from A5 had yielded dates in 8630±85,
8445±90 and 8250±85 BP (Berg man et al. 2004, p.
165). From Garaselet, two bone samples were
selected. The first one was from the distal part of
the middle toe bone of elk, found in the cooking
pit A6. The other was a sesamoid bone from the
hoof of elk found in a bone feature associated
with A8. A8 and the bone feature have previously
been interpreted as coeval. A6 had previously
yielded a date in 5920±80 BP, and A8 one in
6190±90 BP (Knutsson 1993, p. 30).
From Lappviken, two bone samples from the
same bone feature were selected, being the upper
and middle toe bones of elk. No charcoal samples
from Lappviken had previously been dated. Nor
was any charcoal dated in this study, because
none of the charcoal samples were collected from
a sealed context. From a 1993 understanding of
the lithics’ typological development, the site would
date to around 6000 BP (Knutsson 1993, p. 21).
From Högland, two samples from elk bones were
chosen for dating. One was from the distal part of
the upper toe bone (phalanx I) and was collected
233A cross-check of radiocarbon dates from Stone Age sites
Fornvännen 110 (2015)
 
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 233
Tab. 2. All radiocarbon
measurements done on
the features under study
here.
234 Therese Ekholm
Fornvännen 110 (2015)
 
 
 C analyses of charcoal done for this study where no charcoal had previously been dated.
1(Östlund 2004, Appendix 6:1) 2(Bergman et al. 2004, p. 165) 3Silvermuseet dnr 1998/015 4(Knutsson 1993, p. 
30) 5(Melander & Wallin 1981, Appendix) 6(Lindberg & Sandberg 2010, p. 14) 7(Sandberg ms. 2001) 
       
   

       

       

       
 
       
 
       

       

        

        

        

        

        
 
        
 
       

        

       
 
        

       
 
       
 
       
 
       

       

       

       

       

       

        

       

       

14   
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 234
from a bone feature in area 2. The other sample
was from the distal part of a metapodium, collect-
ed from A2. This feature had previously been dat-
ed with charcoal to 7715±115 BP (Melander &
Wallin 1981, p. 7). 
From Orsand, one bone sample from each
bone pit was selected: an undetermined element
of an undetermined mammal from A272 and six
vertebrae from carp fish from A268. From the
latter bone pit a charcoal sample was also select-
ed, and furthermore a charcoal sample from the
cooking pit A255. Both of the charcoal samples
were of pine. In a previous study several indeter-
minate burnt bones had been put together and
dated as one sample without any particular con-
text. This analysis yielded 9089±207 BP. Two
charcoal samples had also been dated previously:
one of pine at 1018±30 BP and one of birch at
1202±30 BP (Lindberg & Sandberg 2010, p. 14).
From Tjärna, a bone was selected from the seem-
ingly undisturbed part of the only feature (A140):
an outer malleolus. The previous species determi-
nation was elk, but after reexamination I chang-
ed this to undetermined deer (Cervidae sp.) be-
cause it did not conform to the reference materi-
al. The feature had previously been carbon dated
with charcoal to 7960±35 BP (Sandberg 2001). 
Results 
All results from the radiocarbon measurements
of burnt bones performed for this study are in
tab. 1. In additions to the dates the δ13C values are
given. Note that burning lowers the δ13C values
(O lsen et al. 2008, p. 796; Snoeck et al. 2014, p. 595).
The dates for this study are presented togeth-
er with previously acquired dates from each con-
text in tab. 2 and fig. 2. They essentially agree with
each other, with the following qualified excep-
tions.
The two radiocarbon dates on burnt bones
from A4 at Dumpokjauratj are consistent with
each other, but compared with the charcoal dates
for the same contexts the bone samples are
300–400 radiocarbon years younger.
Previously, A5 at Dumpokjauratj has been
radiocarbon dated with one burnt bone sample
and three charcoal samples. The bone sample and
the middle charcoal sample in fig. 2c gave fairly
consistent dates, the first charcoal sample being
slightly earlier and the last one slightly later. The
235A cross-check of radiocarbon dates from Stone Age sites
Fornvännen 110 (2015)
Fig. 2. All the radiocarbon dates used
for the study.
Fig. 2 a–c.
Art. Ekholm 227-240_Layout 1  2015-12-03  16:07  Sida 235
Fig. 2 d–f.
difference between charcoal and bones in A5 is
not as obvious as in A4.
A cooking pit, A6, at Garaselet has previously
been radiocarbon dated with charcoal and for
this study also with burnt elk bone. The bone
gave a date that is nearly 600 radiocarbon years
earlier than the charcoal (fig. 2d). A hearth, A8,
at Garaselet has also previously been dated with
charcoal, and for this study also with burnt elk
bone from a bone feature associated with the
hearth. The bone from the bone feature is about
500 radiocarbon years earlier than the charcoal
from the hearth (fig. 2d).
No charcoal samples from Lappviken have
been dated. As noted, however, the site has been
placed typologically around 6000 BP. In this study
two bones from the same context were dated, dif-
fering by only 100–250 radiocarbon years (fig. 2e).
The dates match Knutsson's typological dating.
There was already a carbon date on charcoal
from A2 at Högland. For this study one bone
sample from A2 and one from a bone feature
were selected from different areas on the site. The
dates differ by 100 radiocarbon years between the
two samples from A2, but the sigma is ±105 on
the charcoal sample and ±50 on the bone sample.
Both samples are likely to represent the same
event. The bone sample from the bone feature is
1600 radiocarbon years later than the one from
A2, and is thus from a different use phase (fig. 2f).
At Orsand, two bone samples from two dif-
ferent bone pits and two charcoal samples, one
from the bone pit with carp bones and one from
the hearth were dated for this study. The results
show that the pit with mammal bones (A272) is
2000 radiocarbon years older than the pit with
carp bones (A268). The charcoal sample from
A268 was dated to the same period as the carp
bones, and the charcoal sample from the hearth is
almost 400 radiocarbon years later than the sam-
ples from A268. A radiocarbon date on burnt
bones from a previous study is 800 radiocarbon
years earlier than the earliest date achieved in the
present study (fig. 2g). However, this analysis
was done on an unknown number of combined
samples and is therefore somewhat unreliable.
236 Therese Ekholm
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 236
Fig. 2 g–h.
Discussion 
Where people burned old wood, dates on char-
coal can be expected to be earlier than dates on
bone. This may be the case for A4 at Dumpokjau-
ratj. The two bone samples from A4 at Dumpok-
jauratj gave consistent dates, but compared with
the charcoal dates, they are later. Another expla-
nation may be that the charcoal itself is not the
result of human action but accidentally ended up
in the feature when it was used. If so, the charcoal
must have been burnt before the human activity
that created the bone assemblage. According to
Bergman et al. (2004, p. 166) the stratigraphy in
the features showed that each was used several
times, which may explain the varied dates. The
bones were higher up in the stratigraphy in A4
than the charcoal, and so they may derive from
later use of the feature.
The stratigraphy in Dumpokjauratj A5 de-
mon strates re-use of the feature even more clear-
ly, where the three charcoal samples are earlier,
coeval and later than the bone sample. However,
the earlier charcoal may have a high intrinsic age
and thus be coeval with the burnt bone. The ear-
liest charcoal might also, if the feature has been
re-used, belong to an earlier occasion. As the many
radiocarbon dates from the site shows, Dumpok-
jauratj was used several times (Bergman et al.
2004, p. 165). The results from Dumpokjauratj
may thus help explain the results from the other
sites in the study.
At Garaselet, for example, two or more use
phases have previously been identified: one at
8000 BP and the other at 6000 BP (Knutsson
1993, p. 29 f). Two features from the later phase
were analysed in this study. A6 had previously
been dated with charcoal, and this study adds
burnt bone. It turns out that the burnt bone is
nearly 600 radiocarbon years earlier than the
charcoal. According to Olsen et al. (2008, p. 792),
bones can absorb bicarbonate from the ground,
causing an earlier radiocarbon date. This does
not however affect burnt bones. Old bones that
had been lying on the ground for a long time
were probably not used as fuel, and so the inter-
pretation should be that A6 had been re-used.
A8 at Garaselet, a hearth, had also previously
been dated with charcoal. No bones from feature
8 could be dated because of the fragmentation
level, and so a bone from a bone feature associat-
ed with the hearth was dated. The bone was 450
radiocarbon years earlier than the charcoal and
the hearth and the bone feature seem to have no
relation. Since the bone dates from A6 and the
bone feature match each other and the charcoal
dates from A6 and A8 match each other, the con-
nections in this case are among the charcoal dates
and among the bone dates, respectively. This
237A cross-check of radiocarbon dates from Stone Age sites
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 237
suggests two different phases in the later use peri-
od at Garaselet: a »bone phase» and a »charcoal
phase», where A6 was re-used.
This study suggests that when people needed
a cooking pit or a hearth and there was one already
on the camp site, they would use what was already
there. As long as a place lives in the memory of a
group of people who visit it, it will probably be
used in a similar manner every time. What we see
in the archaeological record may be one feature,
but what actually happened there may have gone
on, on several occasions, for a long time. Mattias
Ahlbeck (1995, p. 42) calls it »contextual con-
temporaneity» (sw. kontextuell samtidighet). The
contexts are not always as simple and sealed as we
would like to think, and the re-use of a feature is
not always visible, especially with early features
which have undergone geological change for thou -
sands of years.
Another reused site is Högland. The dates
show that the charcoal and the bone in A2 derive
from the same event, while the bone feature on
the same site represents an event 1600 years later.
The samples from Kangos and Tjärna show the
opposite. Here the bones and the charcoal repre-
sent the same events.
Conclusion
There is no systematic difference between radio-
carbon dates on charcoal and radiocarbon dates
on burnt bones from the same contexts. But they
do not differ very much in relation to the statisti-
cal uncertainty. This indicates that as long as the
samples are taken with accuracy and with know-
ledge of the error sources, and as long as the samp-
le represents the actual object, it does not matter
if the sample is charcoal or burnt bone. Both ma-
terials date the use of the feature.
Where charcoal and bones did not yield con-
sistent dates, it has been possible to offer two
interpretations:
1. Intrinsic age: old wood used as fuel.
2. Re-used sites and features.
The intrinsic age can be a problem when later
contexts are dated and give precise dates with
short error bars. When earlier sites are dated, the
intrinsic age often fits inside the error bar. The
main problem with charcoal samples from earli-
er sites is that we cannot be sure if they derive
from human action. Only people however burn
bones at high temperatures, so by dating burnt
bones human action is also dated. If the bones are
not found in situ, in association with a sunken
feature or other context, contextualisation of the
act in situ is impossible. The burnt bone sample
itself does derive from human action, but it is
impossible to infer contextual information and
relate the data to artefacts, for example. As the
study shows, a feature may have been re-used, even
if this re-use is not apparent during excavation.
Also one must make sure that what gets dat-
ed is what is intended. Only certain artefacts or
contexts of interest get dated, and if their dates are
to be transferred to anything else, one must be
convinced that they are securely associated. The
most reliable sample is the one that derives from
the object itself or is linked to it (Brock & Dee
2013, p. 41 f). Thus we can lower the risk of unex-
pected results, but such cannot be totally avoid-
ed. The unexpected results need not be seen as
failures. It is important to think about, and be
aware of, the sources of error but to accept and to
not gloss over the results because of errors. Other-
wise the study will be useless.
If unexpected dates come out, then we must
believe in the contexts, but discuss them critical-
ly. This study shows that as long as the sample is
taken from a sealed context it is reliable. Both
charcoal and bone samples are useful for radio-
carbon dating, and when the bones are too frag-
mented, it is possible to date charcoal from the
same context instead. As fig. 3 shows, the samples
dated for this study connect to the previously dat-
ed samples from the same contexts, with within
the margins of error suggested by intrinsic age or
the periods during which each site was visited.
Fig. 3 also shows the contextual contemporaneity
of the features and the sites, where one date rep-
resents one event and several dates represent sev-
eral events during a period of use.
Thanks
Thanks to the Helge Ax:son Johnson Foundation, to
the Gunvor and Josef Anér Foundation, to Åsa Lund-
berg, Susanne Sundström and Anders Olofsson of
Västerbotten Museum, to Lage Johansson and Annika 
238 Therese Ekholm
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 238
Fig. 3. Contextual contemporaneity on the sites. The
diagram shows that sites and features have been used
several times. The line represents 2σ and the box rep-
resents 1σ. (C=charcoal, B=bone).
Sander of Skellefteå Museum, to Ingela Bergman and
Lars Liedgren of the Silver Museum, to Olof Östlund
of Norrbotten Museum, to Leena Drenzel of the Swe -
dish History Museum and to Rita Peyroteo Stjerna of
the Department of Archaeology and Ancient History
Uppsala University.
References
Ahlbeck, M., 1995. Kvartsiell afasi. Stenmaterialet från
den gropkeramiska fyndplatsen Bollbacken utifrån ett
processuellt kunskapsintresse. C-uppsats i arkeologi
Institutionen för arkeologi och antik historia. Upp-
sala universitet.
Bergman, I., 1995a. Från Döudden till Varghalsen. En
studie av kontinuitet och förändring inom ett fångst-
samhälle i övre Norrlands inland, 5200 f.Kr. – 400 e.Kr
Studia Archaeologica Universitatis Umensis 7
Arkeologiska institutionen, Umeå universitet.
– 1995b. Arkeologiska utgrävningar inom Arjeplog sn
Lappland 1994. Silvermuseet, rapport 8. Arjeplog.
– 2001. Människan, elden och landskapet. Arkeologiska
undersökningar av mesolitiska boplatser inom Arjeplogs
kommun, Norrbottens län, 2000. Silvermuseet, rap-
port 28. Arjeplog.
– 2003a. Arkeologiska undersökningar vid Dumpokjau-
ratj och Gublijaure, Arjeplogs socken, Lappland, 2001.
Människan, elden och landskapet. Silvermuseet, rap-
port 34. Arjeplog.
– 2003b. Arkeologiska undersökningar vid Dumpokjau-
ratj, Raä 1568, Arjeplogs socken, Lappland, 2002. Män-
niskan, elden och landskapet. Silvermuseet, rapport
38. Arjeplog.
Bergman, I., Olofsson, A., Hörnberg, G., Zackrisson,
O. & Hellberg, E., 2004. Deglaciation and Coloni-
zation: Pioneer Settlements in Northern Fenno-
scandia. Journal of World Prehistory 18:2. New York.
Biörnstad, M., 2006. Kulturminnesvård och vattenkraft
1942–1980. En studie med utgångspunkt från Riksantik-
varieämbetets sjöregleringsundersökningar. Stockholm.
Björdal, C., 1999. Trämaterial – historiskt och arkeo-
logiskt. Fjæstad, M. (ed.). Tidens tand. Förebyggande
konservering. Riksantikvarieämbetet. Stockholm.
Borg, G.C., Jonsson, L., Lagerlöf, A., Mattsson, E., Ullén,
I. & Werner, G., 1994. Konserveringstekniska studier:
nedbrytning av arkeologiskt material i jord. Målsätt-
ning och bakgrund. Rapport RIK 9. Institutionen för
konservering, Riksantikvarieämbetet och Statens
historiska museer. Stockholm.
Brock, F. & Dee, M.W., 2013. Sample Selection for Radio-
carbon Dating. Shortland, A.J. & Bronk Ramsey,
C. (eds). Radiocarbon and the Chronologies of Ancient
Egypt. Oxford.
Bronk Ramsey, C. & Lee, S., 2013. Recent and Planned
Developments of the Program OxCal. Radiocarbon
55:2. Tucson.
Christensson, E. 1999. Ben, horn och likartade material.
Fjæstad, M. (ed.). Tidens tand. Förebyggande kon-
servering. Riksantikvarieämbetet. Stockholm.
Guinard, M., 2006. Dateringsproblematiken. Guinard
M. & Vogel P. (eds). Stormossen. Ett senmesolitiskt bo -
, platskomplex i den yttre uppländska skärgården. SAU
Skrifter 20. Uppsala.
Hedman, S-D., 2009. Gammal-äldre-äldst. Något om
Norrbottens mesolitikum. Arkeologi i Norr 11. Umeå.
Iregren, E. & Ekman, J., 1984. Archaeo-zoological inves-
tigations in northern Sweden. Early Norrland 8. KV-
HAA. Stockholm.
. Jonsson, L., 2004. Brända ben från Kangos 8:4 (Raä 98)
och Särkeniemi 1:1 (Raä 22) i Junosuando socken,
Lappland. Bilaga 6:2, osteologisk analys. Östlund,
O., 2004. Rapport. Arkeologisk förundersökning. Sten- 
åldersboplats samt skärvstensförekomst Raä 22 samt Raä
. 98, Junosuando socken, Norrbottens län, Västerbotten.
. Norrbottens museum. Luleå.
Knutsson, K., 1993. Garaselet – Lappviken – Rastklip-
, pan. Introduktion till en diskussion om Norrlands
äldsta bebyggelse. Tor 25. Uppsala.
Lanting, J.N., Aerts-Bijma, A.T. & van der Plicht, J.,
2001. Dating of cremated bones. Radiocarbon 43:2.
Tucson.
Lanting, J.N. & Brindley, A.L., 1998. Dating cremated
239A cross-check of radiocarbon dates from Stone Age Sites
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 239
bone: the dawn of a new era. Journal of Irish Archaeo-
logy IX. Galway.
Lanting, J.N. & van der Plicht, J., 1998. Reservoir effect
and apparent 14C-ages. Journal of Irish Archaeology
Vol. IX. Galway.
Lindberg, K-F. & Sandberg, F., 2010. Arkeologisk utred-
ning. Orsandbaden. Bebyggelseförtätning på camping-
området, stenåldersboplats raä 31 m.fl., Leksand socken
och kommun, Dalarna. Arkeologisk rapport 2010:1.
Dalarnas museum. Falun.
Melander, J. & Wallin, J-E., 1981. Rapport över arkeolo-
gisk undersökning av fornlämning Raä 181, Långsele
1:15, Högland, Dorotea sn, Västerbottens län. Väster-
bottens museum. Umeå.
Olsen, J., Heinemeier, J., Bennike, P., Krause, C., Horn-
strup, K.M. & Thrane, H., 2008. Characterisation
and blind testing of radiocarbon dating of cremated
bone. Journal of Archaeological Science 35. London.
Olsen, J., Heinemeier, J., Bennike, P., Krause, C.,
Hornstrup, K.M. & Thrane, H., 2013. »Old Wood»
effect in radiocarbon dating of prehistoric cremat-
ed bones? Journal of Archaeological Science 40. Lon-
don.
Persson, P., 1999. Neolitikums början.Kust till kustböcker
1. Uppsala & Göteborg.
Reimer, P.J. et al., 2013. IntCal13 and Marine13 Radio-
carbon Age Calibration Curves, 0–50 000 Years cal
BP. Radiocarbon 55:4. Tucson.
Snoeck, C., Brock, F. & Schulting, J.R., 2014. Carbon
Exchanges Between Bone Apatite and Fuels Dur-
ing Cremation: Impact on Radiocarbon Dates. Ra -
diocarbon 56:2. Tucson.
Sørensen, M., Rankama, T., Kankaanpää, J., Knutsson,
K., Knutsson, H., Melvold, S., Valentin Eriksen, B.
& Glørstad, H., 2013. The First Eastern Migra-
tions of People and Knowledge into Scandinavia:
Evidence from Studies of Mesolithic Technology,
9th–8th Millennium BC. Norwegian Archaeological
Review 46:1. Oslo.
Vretemark, M., 2003. Osteologisk analys av brända ben
från Dumpokjauratj och Adamvalta, Arjeplog. Rapport
2003:38. Arkeologiskt Naturvetenskapligt Labo-
ratorium. Göteborgs universitet.
Östlund, O., 2004. Rapport. Arkeologisk förundersökning.
Stenåldersboplats samt skärvstensförekomst. Raä 22 samt
Raä 98, Junosuando socken, Norrbottens län, Västerbot-
ten. Norrbottens museum. Luleå.
Unpublished sources
Backe, M. Unpublished bone lists for Lappviken.
Backe, M. Ms. 1990. Osteologisk analys av material
från Garaselet, Jörn sn, Västerbotten.
Ekholm, T. Ms. 2013. Det osteologiska materialet från
Orsandbadet, Leksand sn, Dalarna. Uppsala Uni-
versity.
Eriksson, M. Ms. Osteologisk analys av benmaterialet
från lokalerna: Gublijaure 1999:1, utgrävning år
2000 samt Dumpokjauratj 1986:1 utgrävning år
2000. Lund.
Guinard, M. Ms. 2013. Forskningsgrävning i Orsand.
Uppsala University.
Hårding, B. Ms. 2001. Osteologisk analys av ben från
Tjärna, Borlänge kommun i Dalarna. In Sandberg,
F. Ms. 2001. Tjärna område C. Dalarnas museum.
Falun.
Silvermuseet Dnr 1998/015. Arjeplog.
Sandberg, F. Ms. 2001. Tjärna område C. Dalarnas
museum. Falun.
Sundqvist, L. Ms. Anläggningar Garaselet. Skellefteå
museum.
240 Therese Ekholm
Fornvännen 110 (2015)
Art. Ekholm 227-240_Layout 1  2015-11-26  14:42  Sida 240