Fornvännen;2012(107):4, s. 225-240
Ingår i: samla.raa.se
South Scandinavian flint as a raw material for prehistoric tool manufacture has been investigat- ed in several studies. The focus has been on flint types and their quality, appearance, provenance and availability. Studies based on present-day geo- logical conditions and availability under prehis- toric conditions have been carried out to investi- gate where flint may have been found and how it was transported both within and between regions, such as those with abundant flint in the ground as well as others lacking a natural supply (Knuts- son 1988; Eriksen 2000; Apel 2001; Knarrström 2001; Mjærum 2004; Högberg and Olausson 2007;
Högberg 2009). Some of the better-known stu-
dies of the provenance of flint and interpreta- tions of prehistoric transport concern the exten- sive finds of flint objects, chiefly flint axes, from Bjurselet in northern Sweden (Becker 1952; Knuts- son 1988).
Classification of south Scandinavian flint into different types has chiefly been based on visual inspection (e.g. Becker 1988; 1993). In a study pub- lished in 2007, two of the present authors defined seventeen characteristic types of flint from south- ern Sweden, Denmark and northern Germany on the basis of variables such as colour, appear- ance, properties and availability (Högberg & Olaus- son 2007). Several studies, however, have demons-
Many Different Types of Scandinavian Flint
– Visual Classification and Energy Dispersive X-ray Fluorescence
By Anders Högberg, Deborah Olausson and Richard Hughes
Högberg, A.; Olausson, D. & Hughes, R., 2012. Many Different Types of Scandi- navian Flint – Visual Classification and Energy Dispersive X-ray Fluorescence.
Fornvännen107. Stockholm.
Proceeding from previously published studies of the provenance of southern Scan- dinavian flint based on visual classification and chemical sourcing, this paper pres- ents a comprehensive study of flint provenancing. Existing knowledge of the appear- ance and geological origin of flint types is discussed and reappraised, and new chemical analyses of flint from 25 localities are presented. The results show that although there are certain problems in identifying the provenance of south Scandi- navian flint using geochemical and visual criteria, in most cases these problems can be overcome. The study ends with a discussion of how the results of the study can be applied more broadly in future archaeological research.
Anders Högberg,
Archaeology, School of Cultural Studies, Linnaeus University, SE–391 82 Kalmar anders.hogberg@lnu.se
Deborah Olausson
Dept of Archaeology and Ancient History, Lund University deborah.olausson@ark.lu.se
Richard E. Hughes
Geochemical Research Laboratory, Portola Valley, California rehughes@silcon.com
trated the limitations involved in visual classifi- cation (discussion in Schindler 2010), and in an earlier pilot study we have tested the possibility of supplementing visual diagnosis with geoche- mical sourcing of flint through the use of non- destructive energy dispersive X-ray fluorescence spectrometry (EDXRF; Hughes et al. 2010). We were able to recognise distinctions among flint from different geographical areas. However, the results of the pilot study were preliminary, and we emphasised that further samples must be ana- lysed from these localities to ensure that the dis- tinctions identified remained discrete. Also, we needed to test additional samples from localities not included in the pilot study, to determine the degree to which the chemical signatures are use- ful for identifying flints in this part of the world.
Here we build on these previously published results and synthesise new results from an exten- sive chemical analysis of a selection of Scandina- vian flints (Hughes et al. 2012) together with an updated review of the visual classification of some flint types from southern Scandinavia. It is the most comprehensive study to date of chemi- cal characterisation of flint from south Scandi- navia.
Availability and Visual Classification of Flint Högberg & Olausson’s 2007 classification has been commented on (Bjarke Ballin 2008; Eriksen 2008;
Andrefsky 2009; Schindler 2010) and we have done further research on the availability of various flint types in southern Scandinavia. This applies chiefly to varieties of Falster Flint and Common Kristianstad Flint, and a new flint source at Båstad in northwestern Scania and southern Halland (fig. 1).
Falster Flint is a characteristic type found at Hasselø on the island of Falster, Denmark. Its blue striped appearance in a blue, black or grey matrix makes it is easy to recognise and the flint type is well known among modern knappers for its supreme quality. In 2007 (p. 96 ff) Högberg &
Olausson stated that Falster Flint could be found mainly on the island of Falster, giving it a clear geo- graphical provenance in contrast to many other flint types. This has, however, proven not to be the case. On the north coast of Jutland, Denmark, flint nodules can be found which are impossible
to distinguish by eye from Falster Flint (Eriksen 2008). The same problem applies to certain flint nodules from the beach at Hökholz near Schles- wig in northern Germany. When we visited Hök- holz in 2010 we found several flint nodules which we could classify as Falster Flint, together with an abundance of several other ice- or moraine- transported flint types. This means our previous descriptions of the provenance of Falster Flint (Högberg & Olausson 2007, p. 96 ff) are in need of revision.
The geographical distribution of the visually characteristic Common Kristianstad Flint, with its numerous lighter-coloured spots of various sizes in a black or grey matrix, has not previously been investigated in detail. This type of flint occurs in geological formations and in the till in north-eastern Scania, Sweden, but exactly where it occurs and in what quantities has, until recent- ly, been unclear. Old geological literature that has come to our attention after 2007 makes it clear that Common Kristianstad flint occurs in
situin geological chalk formations in the area around Kristianstad and that it can also be found in ice-transported limestone nappes and in gla- cial till as far south as central Scania, as far west in northern Scania as Vittsjö, and as far east as western Blekinge. Farther east in Blekinge or far- ther to the north in Småland there are no report- ed occurrences of Common Kristianstad Flint (Moberg 1880; 1884; Blomberg 1900; Lunde- gren 1934). This ought to have an impact on in- terpretations of trade in raw material and arte- facts made of Common Kristianstad Flint, par- ticularly in Blekinge and north along the east coast of Sweden up to the Kalmar region and Öland, where many finds of production waste and artefacts made from Common Kristianstad Flint have been made (Lundegren 1934; Magnus- son & Selling 2001).
Another less often discussed occurrence of flint is worth mentioning. Palaeontologist J.C.
Moberg (1886) noted the occurrence of flint and
chalk in north-western Scania and southern Hal-
land. Large-scale surveys revealed a greyish-yellow
homogeneous flint and a dark-grey flint together
with till or limestone. Moberg (1886, p. 365) con-
cluded that a solid chalk system stretching from
the north side of the Hallandsåsen ridge at Bås-
tad and extending to the Halland plain had been destroyed or covered by till during the glacial pe- riod. The occurrence of flint in this area should be investigated more carefully in the future.
Chemical Sourcing
For more than 40 years instrumental techniques have been applied to determine the chemical com- position, or “fingerprint”, of geological flint sour- ces with the goal of using chemical correspon- dences between geological and archaeological ma-
terials to establish the probable geological source of origin of prehistoric artefacts (e.g. Aspinall &
Feather 1972; de Bruin et al. 1972; Bush 1976;
Cowell 1981; Stockmans et al. 1981; Craddock et al. 1983; Kinnunen et al. 1985; Bush & Sieveking 1986; Sieveking & Hart 1986; Sieveking & New- comer 1987; Matiskainen et al. 1989; Luedtke 1992; McDonnell et al. 1997; Schild & Sulgos- towska 1997; Costopoulos 2003; Baltr ūnas et al.
2006; Högberg & Olausson 2007; Allard et al.
2008; Hughes et al. 2010; 2012; Olofsson &
Fig. 1. Map of south Scandinavia with sites mentioned in the text.
Rodushkin 2011). Much of this instrumentally based provenance research has been devoted to studies of flint from England and continental Europe. Despite the long history of interest in flint mines in Denmark and Sweden (e.g., Althin 1951; Becker 1959; Rudebeck 1998) and works focusing on lithic technology in the region (e.g., Olausson 1983; Knutsson 1988; Eriksen 2000;
Apel 2001; Knarrström 2001; Högberg 2009), only a few studies focusing on the chemical charac- terisation of flint from these countries have been published (Micheelsen 1966; Hughes et al. 2010;
2012; Olofsson & Rodushkin 2011).1
Methods
The laboratory analysis conditions and instru- mentation used in our study has been described elsewhere and those interested in additional ana- lytical detail may consult Hughes et al. 2010, p.
21; 2012. The instrument used is a QuanX-EC™
(Thermo Electron Corporation) EDXRF spec- trometer equipped with a silver (Ag) x-ray tube, a 50 kV x-ray generator, digital pulse processor with automated energy calibration, and a Peltier cooled solid state detector with 145 eV resolution (FWHM) at 5.9 keV. The x-ray tube was operated at differing voltage and current settings to opti- mise excitation of the elements selected for analy- sis. Samples were cleaned with distilled water be- fore analysis to remove any noticeable surface contaminants. Special care was taken to avoid di- recting the X-ray beam on to obvious patinated surfaces or calcareous or fossil inclusions. The only other requirement was that each sample should be relatively flat, with a minimum surface size for analysis of >2–3 mm thick and >15–20 mm in diameter.
Previous research showed that X-ray intensi- ties above background for certain trace elements were too low to yield reliable composition esti- mates, so analyses presented in Hughes et al. 2012 focused on major and selected minor elements.
Of the nine elements analysed (Al, Si, S, Cl, K, Ca, Ti, Mn and Fe), using the K
αemission line for each, Si, Ca, and Fe typically generated the high- est count rates (counts/second over background) and total counts, so these elements were employ- ed to characterise differences among flint sour- ces.2 Because of the extremely high SiO2 content
of flint, the Geological Survey of Japan’s JCh-1 chert standard (Imai et al. 1996) was used for pre- cision and accuracy comparison.
Source Criticism
First we must briefly recognise four factors affect- ing the precision and accuracy of non-destructive EDXRF analysis.
1. The homogeneity or heterogeneity of the ma- terial being measured. Volcanic rocks (such as obsidian) are typically quite homogeneous in trace, minor and major element composition and they usually exhibit a comparatively nar- row range of elemental variability. By con- trast sedimentary rocks (like flint and chert) are quite heterogeneous in composition be- cause formation processes over very long pe- riods of geological time compress and com- bine assorted organic and inorganic material.
This diverse parent material expresses itself in sedimentary rocks as chemical heterogene- ity (Luedtke 1992).
2. The physical surface to be analysed. The tar- get surface should be flat and free from inclu- sions. Highly irregular surfaces alter the criti- cal geometrical path between the X-ray tube excitation source and the detector. Highly variable and usually inaccurate results can result. Obsidian artefacts that are relatively flat or lenticular in cross-section typically yield precise data via EDXRF analysis (Hughes 1986, p. 31 ff), and we have had the same result with flint (Hughes et al. 2012).
3. EDXRF is a bulk-area, near-surface technique,
so X-rays excited and emitted from the entire
target (sample) surface are detected and com-
bined to derive a composition estimate for
each element. In the present case, the X-ray
beam was focused on each sample via a 6.8
mm collimator (aperture) within a 30 mm2
excitation area. However, Scandinavian flint
often contains inclusions, and if the X-ray
beam excites an area of such inclusions, the
spectrum shows a noticeable enhancement
(spike) in Ca and Fe composition atypical of
the inclusion-free matrix of the specimen
(Hughes et al. 2012). To avoid this problem
and its impact on correct chemical classifica-
tion of flint, every effort was as mentioned
!"#$%& (%)*+ !,*) !*+ -% .)* -"* /, 01 2&
34#5&6 789:';< 789:';< 789:;< 7$$#< 789:';< 789:';< 7$$#< 7$$#< 7$$#<
!"#$%&"$%'()*+,-$.'/ 0112 33044 20211 145 20216 20787 32 "( 118
9'#'%&%&:");)1 <02= <042 <0221 <12 <021 <021 <44 <5
!"#$%&"$%'()*+,-$.'/ 01=7 33027 20211 116 20272 20753 63 "( 144
9'#'%&%&:");)4 <02= <042 <0221 <12 <021 <021 <11 <5
!"#$%&"$%'()*+,-$.'/ 01>= 33036 20211 141 20241 2075= 86 7 115
9'#'%&%&:");)7 <02= <042 <0221 <12 <021 <021 <44 <4 <5
!"#$%&"$%'()*+,-$.'/ 01>6 3301= 20211 115 20213 20782 32 > 116
9'#'%&%&:");)= <02= <042 <0221 <12 <021 <021 <44 <4 <5
!"#$%&"$%'()*+,-$.'/ 011= 33016 20212 116 20217 20754 88 4 11>
9'#'%&%&:");)> <02= <042 <0221 <12 <021 <021 <44 <4 <6
?'$" 0175 33071 20211 11306 20242 20756 6=06 ". 116
@%$"($,()A'B&$%&:" 0244 20761 C 0221 7032 20225 2022> 5053 ". 4082
D:'--&.&'"%):-)
E$,&$%&:")FGH 150> 706 =01 707 7206 104 803 ". 407
I$%&"$%'()*+,-$.'/ 1084 3508= 2027= 417 20157 20468 178 48 78>2
9'#'%&%&:");)1 <02= <042 <0221 <14 <021 <021 <76 <> <71
I$%&"$%'()*+,-$.'/ 1053 3506= 20277 423 20158 20437 1== 46 7882
9'#'%&%&:");)4 <02= <042 <0221 <14 <021 <021 <46 <> <71
I$%&"$%'()*+,-$.'/ 1084 3508= 20277 42> 20184 2043> 1=6 4> 7851
9'#'%&%&:");)7 <02= <042 <0221 <14 <021 <021 <46 <> <71
I$%&"$%'()*+,-$.'/ 1087 35031 2027= 421 201>3 20463 125 44 785>
9'#'%&%&:");)= <02= <042 <0221 <14 <021 <021 <74 <> <71
I$%&"$%'()*+,-$.'/ 1081 3506> 20277 133 20165 20483 1=> 48 7858
9'#'%&%&:");)> <02= <042 <0221 <14 <021 <021 <46 <> <71
?'$" 1084 35064 20277 42>0= 20153 20463 17502 4>06 7857
@%$"($,()A'B&$%&:" 0244 208= 0221 >087 0211 20225 1804> 4073 8088
D:'--&.&'"%):-)
E$,&$%&:")FGH 107 C 1 105 406 504 404 1408 307 C 1
JDKL1 F>)9'#'%&%&:"*H
?'$" 20621 36071 022> "( 2041> 202=2 1870= 175064 45>8
@%$"($,()A'B&$%&:" 20248 20124 0221 20224 20224 13047 7032 =07>
D:'--&.&'"%):-)
E$,&$%&:")FGH 70= C 1 C 1 C 1 =02 1101 403 C 1
JDKL1 2087= 38061 ", 1= 20441 202=> 163 17= 4=32
F,'.:MM'"('(H
Table 1. Results of the replicated EDXRF analy- sis of the patinated and unpatinated surface of a flint flake. Nd = no data, nc = not computed, nr = not reported. Recom- mended values for JCh-1 from Imai et al. (1996). ±
= 2 sigma estimate (in ppm and weight percent) of X-ray counting uncer- tainty and regression fit- ting error (from Hughes et al. 2012).
Fig. 2. Visualisation of the results in tab. 1. Note that the figure is a graphic representa- tion and only shows the rela- tionship within each analysed element with and without pati- nation. It does not show the relationships between ele- ments. For each element the mean values from the unpati- nated surface and the patinated surface respectively have been divided by the sum of the means from the unpatinated surface and patinated surface, and the quota represented by each bar in the chart. Mn was not computed for the unpati- nated surface, and so this ele- ment is not included.
made to direct the X-ray beam to portions of a sample without any obvious inclusions.
4. Patination and other surface adherents can have serious effects on derived composition estimates. Although we have not yet carried out exhaustive experiments on the effects of patination of flint via EDXRF we have de- monstrated that, in relation to an unpatinated surface, a patinated surface shows SiO2 de- pletion and elevated concentrations of Al2O3, K2O and Fe – particularly Fe. Cl and Ti values are also higher in the patinated samples, al- though CaO values appear to be only slightly affected (tab. 1; fig. 2; Hughes et al. 2012).
These results indicate that post-fracture weathering and deposition environment do introduce chemical changes in the affected surfaces (e.g., Shepherd 1972, p. 114 ff; Luedt- ke 1992; Högberg & Olausson 2007, p. 67 ff).
Silica is leached from the flint and replaced by iron and other substances. With respect to archaeological artefacts, polishing and grind- ing done to artefacts by prehistoric artisans or
“painting” of artefacts with ochre (Van Gijn 2010, p. 177 ff) may have left near-surface residue introduced by polishing media, which could affect measurements. As future EDXRF provenance studies are applied to archaeolo-
gical artefacts, the effects of patination are especially important to recognise.
Results of the Chemical Analyses
The 2010 pilot study involved seven locations restricted to eastern Denmark and southern Swe- den. One question that arose from the results of that study was whether or not the chemical types we labelled Hanaskog, Östra Torp and Stevns Klint would remain discrete if more samples were ana- lysed. To investigate this, we reanalysed the origi- nal specimens we had used to discern the chemi- cal distinctions among Scandinavian flint varie- ties and compared them to new samples from the same source sites.
The results from the additional sample analy- sis are illustrated in fig. 3. The graph shows that the same general CaO and Fe relationship identi- fied among the flint types in the pilot study is apparent in these data. This shows that the analy- ses presented in the pilot study can be repeated, using quantitative composition estimates, with the same results.
With this as a starting point, we deemed it worthwhile to proceed with chemical analyses of south Scandinavian flint from other localities.
Therefore, we widened our geographical focus and
collected flint from an additional 18 localities in
Fig. 3. The pilot study (Hughes et al. 2010) analysed 50 sam- ples from a few localities in eastern Denmark and Scania, demonstrating that it is possi- ble to distinguish different chemical types of Scandinavian flint. As a first step in the extended study presented here, 45 samples from the same local- ities were re-analysed to see whether the pilot results could be repeated. The figure shows that this was accomplished; the three chemically distinct groups / types defined in the pilot study – Hanaskog, Östra Torp, and Stevns Klint – are replicated in this bivariate plot (from Hughes et al. 2012).Denmark, Sweden and northern Germany for analysis (tab. 2). Figs 4–5 show the CaO/Fe rela- tionship among these samples.
Several important new findings emerged.
Composition estimates for these Scandinavian flints refine and quantify the chemical distinc- tions previously identified. Samples from a total of 25 localities (18 new ones and 7 samples from previously analysed localities) add to the number of archaeologically significant chemical signa- tures – and thus varieties – in the region’s flint.
On the basis of CaO and Fe composition, flint from Bjerge-by, Hasselø, Odby, Fakse, Voks- lev, Rørdal and Ellidshøj can be subsumed with- in the Stevns Klint chemical type (fig. 5). Flint
from Öland, Bornholm, Jasmund, Helligkilde, Hising, Klintholm, Mønsted, Sevel, Kinnekulle, Kølbygård and Hanaskog (the high Fe variant, i.e. Black Kristianstad Flint) each have CaO and Fe values distinct from each other and from the Stevns Klint, Östra Torp and Hanaskog types, i.e.
the low Fe variant (Common Kristianstad Flint) previously defined in the pilot study (fig. 4, 5).
Consistent with the variants defined visually by Högberg & Olausson (2007), there appear to be at least two chemical varieties of Hanaskog flint; Common Kristianstad Flint falls within the chemical variety termed Hanaskog in tab. 2 and fig. 3, while Black Kristianstad Flint from the same general location in Scania contains a significant-
Table 2. Quantitative composition estimates for flint samples from south Scandinavia. Recommended values for JCh-1 from Imai et al. (1996). nd = no data. The “Flint Type” column lists visual categories according to Högberg & Olausson 2007.!"#$%&'()* +,-."#%( /0+&+.12&3451*6 7"*5*$%8
"#$%& '()$ ')& *# +$) *,) -( ./ 01
2345!67 2345!67 2345!67 28897 2345!67 2345!67 28897 28897 28897
0,#:41; <,::1#= >%, %5%?@ @@5?? %5%$> % %5%$> %5>%A $% % ?%
0,#:41; <,::1#= >% %5%BB @@5?% %5%$@ &? %5%>A %5>A$ A% % ??
'C,/D(/,E(,/ F=;D,# $&!2&5A!997 %5%GG @@5?$ %5%$? G> %5%>@ %5>H$ &B % ?H
'1/I/(,/ J##(D:K=L $@ %5%?B @@5$% %5>@$ ?$ %5%$@ %5>B& HH % AB
'1/I/(,/ J##(D:K=L $@, %5>B& @@5H> %5%$% /D %5%H$ %5>?$ A> /D GG
'1/I/(,/ J##(D:K=L $@M %5>H& @@5AB %5%$G &% %5%$B %5>$H HB /D >%%
'1/I/(,/ J##(D:K=L &%, %5$%? @@5>B %5%>% >%$ %5%&H %5$%B &G /D H?
'1/I/(,/ J##(D:K=L &% 2&5A!997 %5$>& @@5H$ %5%>% >%@ %5%H? %5$%? H> % AB
FNO1/ P,:9Q/D $A!2&5A!997 %5%AH @G5B? %5%$B ?? % >5%B? HA % ?$
FNO1/ P,:9Q/D $AC!2&5A!997 %5>%B @G5&? %5%$% G> % >5&>% $> & HB
R;SITI,/ +#(/4KI#9 >?M!2U$7 %5&H% @@5%% %5%&% >&@ %5%H& %5$?B ?& ? >&A
R;SITI,/ +#(/4KI#9 >?M5& %5H%? @G5@> %5%&& >>A %5%HG %5$@A ?@ A >H&
R;SITI,/ +#(/4KI#9 >?,!2&5A!997 %5$&G @G5G& %5%$@ >$& %5%&> %5&@& H% A >>&
R;SITI,/ +#(/4KI#9 >?C %5&>A @@5%% %5%$% >$@ %5%$H %5&AB >%& & >$G
R;SITI,/ +#(/4KI#9 >?,, %5>%@ @@5HA %5$H& >%G %5%$$ %5$>> A% /D >&&
V '31D(:K!R1,CK <(:(/O1/ >$ %5%@> @G5A% %5%>? >%@ % >5>@% $H & >$B
V '31D(:K!R1,CK <(:(/O1/ >$, % @G5?? %5%$& G$ % >5>A% $H & >A&
W;,S!R,/D!X,/(,/ +=#MSOY;D H> %5$$> @@5%@ %5%&% >$& %5%&& %5A%$ B@ & >>?
R#,CZ!+;(:4(,/:4,D <,/,:ZIO >,!2&5A!997 %5BBG @B5GG %5%A& >%A %5>&& %5HA? >%G B >B&B R#,CZ!+;(:4(,/:4,D <,/,:ZIO >M!2&5A!997 %5H>A @G5?> %5%?H @> %5%B? %5&H% >&@ ? >??>
R#,CZ!+;(:4(,/:4,D <,/,:ZIO >C!2&5A!997 %5$&> @G5&H %5%&? B% %5%?@ >5%>% ?@ >% >%@B
R;I3/!R;SITI,/ 0,Z:1 H 2&5A!997 %5>&% @@5H> %5%>$ @& %5%$B %5&%& ?B % A?
R;I3/!R;SITI,/ [IZ:#1E $?,!2&5A!997 % @@5A@ %5%$@ ?> %5%$A %5$AH $? $ ?$
R,## RI;/KI#9 A % @@5%$ %5%$> G? % %5B?B AG % HB
R,## RI;/KI#9 A,5& %5%?% @G5@% %5%$% @& %5%$% %5BH? $B & A$
+(//1ZQ##1 +(//1ZQ##1 >>5$!2&5A!997 %5H@A @G5&& %5%B% $?? %5%?A %5H%@ >H> % &?BG
W;,S!R,/D!.,441 '1E1# H@ %5>&A @H5>G %5%$$ >%? % A5&A% HH >& H??
W;,S!R,/D!.,441 '1E1# H@,!2&5A!997 %5>AH @&5G$ %5%>? @& % A5AB% G? >H HH>
W;,S!R,/D!.,441 .=/:41D $> %5>$? @B5?@ %5%$? >BG % >5?H% B> & >BG
W;,S!R,/D!.,441 .=/:41D $>, %5$>% @B5%? %5%$B &%A % >5@A% GH H >B>
F1D\R;I3/!R;SITI,/ )DMS >G % @@5AA %5%$G >$& %5%$> %5$?> % % ?%
F1D\R;I3/!R;SITI,/ )DMS >G, %5>>@ @@5&$ %5%%G ?% %5%>@ %5$G$ &B % ?$
F1D\R;I3/!R;SITI,/ RL1;O1\MS H& %5$?H @G5@B %5%AG @H %5%HA %5&%& BB % @B
F1D\R;I3/!R;SITI,/ <1##(OZ(#D1 A% %5>$% @G5B& %5%>@ $& %5%>A >5%?& HH & >%A
);DIE(C(,/ ]#,/D >H >5>B% G>5H% %5%$A >H%G % >A5@% H$G >AA B>>G
ly greater amount of Fe, plotted as “Hanaskog”
in the legend to fig. 4.
A dramatic difference in CaO/Fe composi- tion sets Öland and Kinnekulle flint apart from all others in this Scandinavian sample (tab. 2; fig.
4). These flint types come from other geological formations and periods, and they are visually dis- tinct from the rest of the analysed flint. Indeed they also come from localities that are geograph- ically distant from other sites (Högberg & Olaus- son 2007, p. 132 ff).
CaO/Fe contrasts provide evidence for two chemical types, Mønsted and Sevel, within the visual type we labelled Grey Band Matte Danian Flint (Högberg & Olausson 2007, p. 108 ff). This
distinguishes them from other flints from north- ern Jutland. Although the two localities produce the same visual variety of flint and are geographi- cally close to each other, they have different che- mical compositions.
Flint from Danian and Senonian levels at Stevns Klint (fig. 6) also contains more CaO than other members of the chemical type of that name defined previously (fig. 3), which may help to dif- ferentiate it in archaeological contexts.
The samples from Bornholm (the so-called Ball Flint), and from Hisingen (labelled West Swedish Beach Flint) do not represent homoge- neous flint from two primary sources but rather stray nodules available on beaches along the coast
Fig. 4. A plot of samples from Kinnekulle, Hanaskog, Sevel and Öland. These four types have distinct chemical compositions.The limits for the chemical types from figure 3 are shown. Note that the Hanaskog chemical type defined in the pilot study consists of Common Kris- tianstad Flint and that the samples labelled Hanaskog in this figure and plot- ted to the right of the Hanaskog chemical type are of the Black Kristianstad Flint variety. Both flint types are found at the Hanaskog site. They have different chemi- cal compositions, in addition to looking different and having different knapping properties. The lower left-hand corner is shown enlarged in fig. 5.
Fig. 5. Flint with a chemical composition within or close to the Stevns Klint type. It is evident that flint from several different localities is subsumed within the same chemical type. Further, that it is possible to distinguish different flints on the basis of their chemical composition. The data points form several clusters, indicating that the Stevns Klint chemical type as defined in the 2010 pilot study needs to be revised. Note that the symbols for Vokslev and Rørdal are hidden behind the symbols for Odby and Hasselø.
of Bornholm and on the west coast of Sweden.
The geological origin of these flint types is not known. Relative CaO enrichment allows the Ball Flint from Bornholm to be separated from the Stevns Klint group (fig. 5). This graph also shows that CaO/Fe values distinguish between Born- holm and the Rügen Flint samples from Jasmund in northern Germany. Chemical analysis of the flint from Hisingen distinguishes it from other types. However, Ball Flint and West Swedish Beach Flint come from beach ridges where they have been affected by water and wind for long periods of time and so may be thoroughly pati- nated, that is, affected all the way through the nodule. Even though we analysed surfaces on freshly knapped flakes whose surfaces did not show any visible patination, we do not know for certain whether the distinct pattern we obtained reflected the chemical type of the flint or the im- pact of weathering.
Fig. 5 also shows that the Reddish-brown Bryo- zoan Flint sample from Helligkilde (Högberg &
Olausson 2007, p. 128 ff) has a distinct chemical composition. However, we have only been able to analyse one sample. A single specimen from Køl- bygård of the Grey Band Danian type also has a distinct CaO/Fe composition. To draw any but the most tentative conclusion, we need to analyse more samples of these flint types.
Other samples analysed have CaO and Fe compositions within the range of the Stevns Klint chemical type. The sample from Rørdal, the specimens from Ellidshøj, from Odby and from Bjerge-by conform to the Stevns Klint chemical type, as corroborated by visual classification as Scandinavian Senonian Flint (Högberg & Olaus- son 2007, p. 89, 105, 129). The characteristics of Falster Flint from Hasselø also correspond to those of the Stevns Klint chemical type (fig.5).
Another visually distinct flint type is the Brown Bryozoan Flint which can be found at Fakse on Zealand and Vokslev in Jutland. The honey-brown or grey matrix with numerous bryo- zoans visible to the naked eye makes it easy to recognise (Högberg & Olausson 2007, p. 120 ff).
However, the chemical analysis places it together with the Scandinavian Senonian Flint (fig. 5), demonstrating that flint with different visual cha- racteristics can have a very similar CaO/Fe geo- chemistry.
Flint from Klintholm, the so-called Bryozoan Flint, Funen Variety, is partly chemically distinct from the Stevns Klint group, but intergrades at lower CaO and Fe compositions in specimens where inclusions are less prevalent (fig. 5). The Bryozoan Flint, Funen Variety from Klintholm is distinct in appearance (Högberg & Olausson 2007, p. 116 ff) and has previously been identi-
Fig. 6. Here three of the localitieswith samples within the Stevns Klint type are plotted at high reso- lution. It is possible to distinguish different localities. Flint from Møns Klint and Södra Sallerup cluster together, whereas flint from Stevns Klint stands out. Note that the flints from Stevns Klint in the plot are of both Danian and Senonian ages. Samples analysed from Danian levels at Stevns Klint contain significantly higher con- centrations of Cl than do those from Senonian contexts, making it possible to distinguish between them.
fied as one of the few south Scandinavian flints which are distinctive as to visual appearance and geological source (Thomsen 2000, p. 35). Our study supports Thomsen’s results; flint from Klintholm is chemically distinct, even though it falls close to the Stevns Klint type in the plots.
Note that some of our data come from a very small number of geological samples, just one per locality in some cases. This means that even though the results of our study show that it is possible to identify flint chemically from some localities in south Scandinavia, it is not yet possible to specify the extent of intra-source chemical variability or potential chemical overlap. Doing so will require additional analysis of more source rocks, and the addition of other well-measured chemical ele- ments for corroboration or modification. Future studies might also benefit from multivariate ana- lyses including several or all analysed elements.
But at present our sample size per group (chemi- cal type) is much too small and uneven to satisfy the statistical preconditions for such analyses.
Note also that chemical groupings of flint do not necessarily imply geographical proximity. The terms “source” and “chemical type” are used here to identify groups of specimens that share a dis- tinct chemical signature independently of geo- graphic location (cf. Hughes 1998, p. 104). In fact, the results of our study provide several examples of the non-concordance between che- mistry and geography, e.g. the Brown Bryozoan Flint from Fakse and Vokslev mentioned above.
Although many in situ flint exposures are known in southern Scandinavia, glacial scouring and ice transport have moved flint from its point of ori- gin to widely dispersed and mixed contexts where it was exploited by those areas' inhabitants (Niel- sen 1997; Knarrström 2001; Högberg & Olaus- son 2007; Eriksen 2010, p. 87 ff). In light of this it is important to determine, as precisely as pos- sible, the geographical dispersal of chemical flint varieties so that informed inferences can be made about the role(s) and significance of human acqui- sition and conveyance of these materials at dif- ferent times in the past.
Archaeological Implications
Our 2010 pilot study showed that flint which looks the same, for example Scandinavian Senonian Flint,
ended up in the same chemical type group. In our more recent expanded study we investigated the extent to which the earlier results applied to flint from a wider area, i.e. flint from all over southern Scandinavia (Högberg & Olausson 2007; Hughes et al. 2012). The ideal result we had hoped for was that visually similar flint from geographically di- stant areas would possess different chemical sig- natures. If so, this could lay a clear foundation for interpretations of how raw material or artefacts were moved during prehistory. Regrettably, that was not the result we obtained. One of the clear- est findings of the expanded study is that classifi- cation based on appearance, geography, geology and geochemistry sometimes groups the samples within the same type, but sometimes not. The con- clusion to be drawn from the studies presented here is that there is no exact correlation between the flint’s visual appearance and properties, the place it comes from, the geological formations it comes from, or its chemical composition. This is an important result, as it shows the difficulty of sourcing flint from southern Scandinavia. It also demonstrates that the grouping of flint into three chemical types that we suggested in our 2010 pilot study is not sufficient for future studies. The che- mical types need revision and expansion and, if chemical types are to be discussed, new ones will need to be delineated.
It is relevant at this juncture to ask how our new results regarding the geochemistry of south Scandinavian flint relate to previous provenance analyses of flint, and what their significance will be for future archaeological studies of the trans- port of flint and flint artefacts.
Based on typological and technological stud- ies, Jan Apel (2001) has suggested that during the Late Neolithic, Scandinavian bifacial knapped flint daggers made of Scandinavian Senonian Flint – the so-called Danish Daggers (Callahan 2006; Nunn 2006) – were produced mainly with- in two areas. One was in northern Jutland. From here he suggested that the daggers were trans- ported to northern Germany and Norway and then from Norway to eastern central Sweden.
The other area where daggers were produced was
in south-eastern Denmark and Scania. Here dag-
gers were made of similar Scandinavian Senon-
ian Flint and transported to northern Germany,
southern Sweden and south-eastern Norway (figs 7–8; Apel 2001, p. 304 ff). In Sweden many such daggers have been found in areas where no flint can be found in the ground (fig. 9). Flints which look alike but come from geological sources located far from each other may differ in their chemical composition. If this proves to be the case, the results could be used to further investi- gate Apel’s suggestions about the trade or distri- bution routes for daggers. In fact, this is an im- portant problem to address with future EDXRF analysis.
Overall, our results are both encouraging and sobering. Samples from different sites where Scan-
dinavian Senonian Flint is available display the same chemical characteristics. Fig. 5, for example, shows that samples which are visually similar, have the same properties when knapped, and have the same chemical composition, can come from dif- ferent geographical areas. But as fig. 6 shows, some samples of Scandinavian Senonian Flint do not plot in exactly the same composition range. For example, flint from Stevns Klint is slightly differ- ent in composition from the flint from Södra Sallerup and Møn. It should also be added to the discussion that Danish daggers were sometimes made out of Danian flint, making future geoche- mical studies on this artefact type even more pro-
Fig. 7. The relative frequency of LN I dagger types indifferent regions in Scandinavia. Type I daggers from the production area in Jutland, Denmark, are marked in white and type II and III daggers from the produc- tion area in eastern Denmark and Scania appear in black (from Apel 2001, p. 306, fig. 9:16).
Fig. 8. Proposed exchange routes from the two main production areas to different parts of Scandinavia during LN I (from Apel 2001, p. 307, fig. 9:17).
mising. These examples show that, while there is no simple result that clearly separates flints from different localities from each other, there is tre- mendous potential to inform future investigations of how raw material and artefacts were trans- ported during prehistory.
There are other important results which may be of value to future research on flint provenance.
Rügen Flint from Jasmund is very similar in visual appearance to Scandinavian Senonian Flint. What sets them apart is that the former has a so-called
“diving cortex”, i.e. the chalk cortex intrudes into the flint from the surface, creating cavities and tunnels in the flint. Larger nodules of Rügen Flint also contain irregular lighter spots of different sizes which are often surrounded by a thin band of even lighter-coloured flint (Högberg & Olaus- son 2007, p. 100 ff). But smaller black nodules of Rügen Flint which lack the diving cortex can easi- ly be mistaken for Scandinavian Senonian Flint.
The coasts of the south-west Baltic Sea with Sca- nia, eastern Denmark and Germany shared basi- cally the same material culture over long periods of prehistory (Apel 2001; Högberg 2009). Thus it would be important to examine any evidence for transports of raw material. Sourcing of Rügen Flint has hitherto rested solely on visual assess- ment, which can be problematic when artefacts lack cortex, so chemical sourcing may be useful in future analyses of finds from this area.
Black Kristianstad Flint has a distinct chemi-
cal composition but it is seldom identified as the
raw material for prehistoric artefacts in Swedish
museum collections (Högberg & Olausson 2007,
p. 78 ff), despite the fact that the flint was avail-
able in a densely populated landscape (Edring
2005). A flake or a nodule of Black Kristianstad
Flint is easy to recognise if it is not patinated and
retains its cortex. Its deep black colour and homo-
geneity, together with the rather thick yellowish
Fig. 9. Daggers and spearheads found in a gallery grave at Skogsbo in Västergötland (Forssander 1936). No flint occurs naturally in the area. The dagger at the far left is c. 28 cm long. Photo by AH.grey cortex, are unique. But if Black Kristianstad Flint has been knapped into an artefact, with no cortex left on the surface, and this artefact has become patinated after thousands of years buried in an archaeological context, then the flint type is much more difficult to recognise. It takes on an appearance that resembles patinated Scandina- vian Senonian Flint. Since the latter type is well known to archaeologists, whereas few Swedish archaeologists are familiar with Black Kristian- stad Flint, a patinated artefact with no cortex made out of this flint type can easily be mistaken for Scandinavian Senonian Flint. Our chemical analyses show clear differences between Black Kristianstad Flint and all the other flint types.
This result could be used in the future to investi- gate whether or not Stone Age tools found in the area around Kristianstad and classified as Scandi- navian Senonian Flint may actually have been made from Black Kristianstad Flint. For example, finds from the many excavated megalithic tombs around Kristianstad could be interesting to con- sider from this point of view (Bagge & Kaelas 1950; 1952). This, in combination with the new knowledge presented above about the distribu- tion and occurrence of Common Kristianstad Flint in northern Scania and parts of Blekinge, may lead to new results regarding how flint in this area was used as a raw material in prehistoric times.
Falster Flint was used a great deal during pre- history. In our analyses this flint type proved to have the same chemical composition as Scandi- navian Senonian Flint, for example from Møn and Ellidshøj. As pointed out above, Falster Flint occurs at more places than Hasselø on Falster;
the type has been reported from northern Jut- land (Eriksen 2008), and our renewed studies show that the flint type is also found at Hökholz in Schleswig. Falster Flint is visually distinctive, with its blue stripes in a blue, black, or grey ma- trix. The stripes are visible even if the flint is pati- nated. Therefore artefacts manufactured from Falster Flint are easy to recognise. However, not all areas of the matrix of this type of flint are striped, and when stripes are absent it is difficult to tell dark varieties of Falster Flint apart from Scandinavian Senonian Flint. This applies to pati- nated flint as well as fresh surfaces. And the re-
sults of the chemical analyses do not help us in making this judgement.
Ball Flint from Bornholm falls within a sepa- rate chemical group (fig. 5). This is an interesting result for future studies of contacts between Bornholm and the east coast of southern Sweden and the south Baltic coast. There is an ongoing archaeological discussion (Skak-Nielsen 2004) as to whether or not Bornholm, located only c. 40 km southeast of the Swedish coast, played an im- portant role in contacts across the Baltic during the Late Mesolithic and Early Neolithic. Future provenance studies of flint could clarify this.
In the introduction to this paper we men- tioned the finds from Bjurselet in northern Swe- den as being among the better-known Scandina- vian examples of transport of flint artefacts in prehistoric times (Becker 1952; Knutsson 1988).
In his raw material studies of the finds, based on visual characterisation, Becker (1952) argued that the flint from which most of the axes in the finds are made came from eastern Denmark.
Using our results, we hope to be able to pinpoint exactly where the raw material for the Bjurselet axes came from (Olausson et al. in press).
Thanks to Jan Apel for kindly letting us use figures from his dissertation. Funding for this research was provid- ed by The Birgit and Gad Rausing Foundation for Re- search in the Humanities and by The Erik Philip-Sören- sen Foundation for Research in Genetics and the Hu- manities, Lund University. English translation by Alan Crozier.
End notes
1. Olofsson & Rodushkin (2011) have published a pi- lot study involving Scandinavian flint. The authors emphasise that their results are preliminary, but we wish to make a few comments. The study is bas- ed on six archaeological artifacts and nineteen re- ference samples. Seven of the reference specimens are geological samples from Denmark. The rest are patinated archaeological samples from sites in Rus- sia (p. 1149, tab. 2). Our work shows the difficulties inherent in using patinated flint as a reference samp- le. This, together with the likelihood that archaeo- logical artifacts may have been transported across vast distances during prehistory (as e.g. the flint artefacts from Vuollerim exemplify), renders such
artifacts highly problematic as reference samples.
Furthermore, the geochemical variation between flint types that we have identified from Denmark makes it impossible to treat this large geographical area as one. In table 2 (p. 1149) a geological refer- ence sample is listed as Danian flint from Møns Klint. But there are no Danian flint layers in the chalk cliffs of Møn, only Senonian flint of different ages (Högberg & Olausson 2007, p. 36 ff). It is also stated that flint mines are present on Møns Klint (p. 1164). As far as we know there are no prehisto- ric flint mines from this area (Högberg & Olausson 2007, p. 51 ff).
2. The analyses for SiO2, CaO and Fe (and other reported elements) were conducted in vacuum un- der three different sets of operating conditions to optimise excitation of major and minor elements of interest. Background subtracted integrated net count rate (counts/second) data were converted to composition estimates (ppm and weight percent composition) using a fundamental parameters algo- rithm incorporating international rock standards after overlapping Kαand Kβline contribution from adjacent elements was stripped and matrix correc- tion algorithms applied. X-ray tube current was scaled automatically to the physical size of each specimen.
References
Allard, P.; Bostyn, F.; Gilignay, F. & Lech, J. (eds). 2008.
Flint mining in prehistoric Europe: Interpreting the archaeological records.BAR Intl Series 1891. Oxford.
Althin, C.A., 1951. The Scanian flint mines. Meddelan- den från Lunds Universitets Historiska Museum1950–
51, IV. Lund.
Andrefsky, W. Jr., 2009. Book Review: Scandinavian Flint. An archaeological perspective. Journal of An- thropological Research65:3. Albuquerque.
Apel, J., 2001. Daggers, knowledge and power: the social aspects of flint-dagger technology in Scandinavia 2350–
1500 cal BC. Kust-till-kust böcker 3. Uppsala.
Aspinall, A. & Feather, S.W., 1972, Neutron activation analyses of prehistoric flint mine products. Archaeo- metry14. Oxford.
Bagge, A. & Kaelas, L., 1950–52. Die Funde aus Dolmen und Ganggräbern in Schonen, Schweden. I–II.KVHAA.
Stockholm.
Baltrūnas, V.; Karmaza, B.; Kulbickas, T. & Ostrauskas, T., 2006. Siliceous rocks as a raw material of pre- historic artefacts in Lithuania. Geologija 56. Vil- nius.
Becker, C.J., 1952. Die nordschwedischen Flintdepots.
Acta ArchaeologicaXXIII. Copenhagen.
– 1959. Flint-mining in Neolithic Denmark. Anti- quity33. Cambridge.
– 1988. Skandinavisk flint i nordsvenske fund fra mellem- og sen-neolitisk tid. XVII Nordiska Arkeo- logmötet i Åbo 1985. ISKOS 7. Helsinki.
– 1993. Flintminer og flintdistribution ved Limfjor- den: Kort- og råstofstudier omkring Limfjorden.
Lund, J. & Ringtved, J. (eds). Limfjordsprojektet vol.
6. Århus.
Bjarke Ballin, T., 2008. Book Review: Scandinavian Flint. An archaeological perspective. Antiquity 2008.
York.
Blomberg, A., 1900. Geologisk beskrifning öfver Ble- kinge län. Jemte redogörelse för stenindustrien inom Blekinge län. Sveriges Geologiska Undersökning. Ser.
C.a. Afhandlingar och uppsatser i 4:0. N:o 1.Stock- holm.
de Bruin, M.; Korthoven, P.J.M.; Bakels, C.C. & Groen, F.C., 1972. The use of non-destructive activation analysis and pattern recognition in the study of flint artefacts. Archaeometry 14. Oxford.
Bush, P.R., 1976. The use of trace elements in the archaeological classification of cherts. Staringia 3.
Pinneberg.
Bush, P.R., & Sieveking, G. de G., 1986. Geochemistry and the provenance of flint axes. Sieveking, G. de G.& Hart, M.B. (eds). The Scientific Study of Flint and Chert.Cambridge.
Callahan, E., 2006. Neolithic Danish Daggers: An ex- perimental peek. Apel, J. & Knutsson, K. (eds).
Skilled Production and Social Reproduction. Aspects on Traditional Stone-Tool Technologies. SAU Studies 2.
Uppsala.
Costopoulos, A., 2003. Prehistoric flint provenance in Finland: Reanalysis of southern data and initial results from the north. Fennoscandia Archaeologica XX. Helsinki.
Cowell, M.R., 1981. The archaeological and geochem- ical implications of trace element distributions in some English, Dutch, and Belgian flints. Staringia 6. Pinneberg.
Craddock, P.T.; Cowell, M.R.; Leese, M.N. & Hughes, M.J., 1983. The trace element composition of pol- ished flint axes as an indicator of source. Archaeo- metry25. Oxford.
Edring, A., 2005. Berget i backens skugga. Det sociala land- skapet på Kristianstadsslätten under tidig- och mellan- neolitikum.Report Series 94. Dept of Archaeology, University of Lund.
Eriksen, B.V. (ed.), 2000. Flintstudier: en håndbog i sys- tematiske analyser af flintinventarer.Aarhus Univer- sity Press.
– 2008. Review of Anders Högberg & Deborah Olausson: Scandinavian Flint – an Archaeological Perspective. KUML 2008. Højbjerg.
– 2010. Flint working in the Danish Bronze Age:
The decline and fall of a master craft. Eriksen, B.V.
(ed.). Lithic Technology in Metal Using Societies. Jut- land Archaeological SocietyPublications 67. Højberg.
Forssander, J.E., 1936. Der ostskandinavische Norden wäh- rend der ältesten Metallzeit Europas. Lund.
Högberg, A., 2009. Lithics in the Scandinavian Late Bronze Age: Sociotechnical Change and Persistence. BAR Intl Series 1932. Oxford.
Högberg, A. & Olausson, D., 2007. Scandinavian Flint – an Archaeological Perspective.Aarhus University Press.
Hughes, R.E., 1986. Diachronic Variability in Obsidian Procurement Patterns in North-eastern California and South-central Oregon.University of California Pub- lications in Anthropology 17. Berkeley–Los Ange- les.
– 1998. On reliability, validity, and scale in obsidian sourcing research. Ramenofsky, A.F. & Steffen, A.
(eds). Unit Issues in Archaeology: Measuring Time, Space and Material.University of Utah Press. Salt Lake City.
Hughes, R.E.; Högberg, A. & Olausson, D., 2010.
Sourcing flint from Sweden and Denmark: a pilot study employing non-destructive energy disper- sive X-ray fluorescence spectrometry. Journal of Nordic Archaeological Science17. Stockholm.
– 2012. The Chemical Composition of Some Archaeo- logically Significant Flint from Denmark and Swe- den. Archaeometry 54. Oxford.
Imai, N.; Terashima, S.; Itoh, S. & Ando, A., 1996.
1996 compilation of analytical data on nine GSJ geochemical reference samples, “sedimentary rock series”. Geostandards Newsletter 20. Paris.
Kinnunen, K.; Tynni, R.; Hokkanen K. & Taavitsai- nen, J-P., 1985. Flint raw materials of prehistoric Finland: rock types, surface textures and microfos- sils. Geological Survey of Finland 334. Espoo.
Knarrström, B., 2001. Flint – a Scanian Hardware. Skåns- ka spår – arkeologi längs Västkustbanan. Riksantik- varieämbetet. Stockholm.
Knutsson, K., 1988. Making and using stone tools: The analysis of the lithic assemblages from Middle Neolithic sites with flint in Västerbotten, northern Sweden. Aun 11. Uppsala.
Luedtke, B.E., 1992. An archaeologist’s guide to chert and flint. Archaeological Research Tools 7. Institute of Archaeology, University of California, Los Ange- les.
Lundegren, A., 1934. Kristianstadsområdets kritbild- ningar. Geologiska Föreningens i Stockholm Förhan- dlingar. Band 56, häfte 2, No 397. Stockholm.
Magnusson, G. & Selling, S. (eds). 2001. Möre – histo- rien om ett småland. E22-projektet. Kalmar.
Matiskainen, H.; Vuorinen, A. & Burman, O., 1989.
The provenance of prehistoric flint in Finland. Ma-
niatis, Y. (ed.). Archaeometry: Proceedings of the 25th International Symposium. Amsterdam.
McDonnell, R.D.; Kars, H. & Jansen, B.H., 1997. Pet- rography and geochemistry of flint from six Neo- lithic sources in southern Limburg (The Nether- lands) and northern Belgium. Ramos-Millan, A. &
Bastillo, M.A. (eds). Siliceous Rocks and Culture.
Granada.
Micheelsen H., 1966. The Structure of Dark Flint from Stevns, Denmark. Meddelelser Fra Dansk Geologisk Forening16. Copenhagen.
Mjærum, A., 2004. Å gi øksene liv. Et biografisk perspektiv på slipte flintøkser fra sørøstnorsk tidlig- og mellomneoli- tikum. Hovedfagsoppgave i nordisk arkeologi. Uni- versity of Oslo.
Moberg, J.C., 1880. Berättelse afgiven till Kongl. Veten- skaps-Akademien om en med understöd af allmän- na medel företagen resa till en del svenska kritloka- ler. Öfversigt af Kongl. Vetenskaps-Akademiens Förhand- lingar 1880, N:o 10.Stockholm.
– 1884. Cephalopoderna i Sveriges kritsystem. I. Sve- riges kritsystem systematiskt framstäldt. Med två kartor. Sveriges Geologiska Undersökning. Ser C. Af- handlingar och uppsatser63. Stockholm.
– 1886. Kritsystem i fast klyft i Halland. G.F.F. Bd. 8.
Stockholm.
Nielsen, L.E., 1997. Raw material provenance in the Early Neolithic: A comparative study of thinbutt- ed flint axes from two regions in Jutland, Den- mark. Schild, R. & Sulgostowska, Z. (eds). Man and Flint.Institute of Archaeology and Ethnology, Polish Academy of Sciences, Warsaw.
Nunn, G., 2006. Using the Jutland Type IC Neolithic Danish Dagger as a model to replicate parallel, edge- to-edge pressure flaking. Apel, J. & Knutsson, K.
(eds). Skilled Production and Social Reproduction:
Aspects on Traditional Stone-Tool Technologies.SAU Studies 2. Uppsala.
Olausson, D.S., 1983. Flint and ground-stone axes in the Scanian Neolithic: An evaluation of raw materials bas- ed on experiment. Scripta Minora Regiae Societatis Humaniorum Litterarum Lundensis 1982–1983 no. 2. Lund.
Olausson, D.; Hughes, R.E. & Högberg, A., in press. A New Look at Bjurselet and the Neolithic Flint Axe Caches from Västerbotten, Sweden, Using Non- Destructive Energy Dispersive X-ray Fluorescence Analysis for Provenance Determination. Acta Ar- chaeologica. Copenhagen.
Olofsson, A. & Rodushkin, I., 2011. Provenancing Flint Artefacts with ICP-MS Using REE Signa- tures and Pb Isotopes as Discriminants: Prelimi- nary Results of a Case Study from Northern Swe- den. Archaeometry 53. Oxford.
Rudebeck, E., 1998. Flint Extraction, Axe Offering, and the Value of Cortex. Edmonds, M. & Richards,
C. (eds). Understanding the Neolithic of North-western Europe. Glasgow.
Schild, R. & Sulgostowska, Z. (eds), 1997. Man and Flint.Institute of Archaeology and Ethnology, Pol- ish Academy of Sciences. Warsaw.
Schindler, B., 2010. Review of Anders Högberg and Deborah Olausson, Scandinavian Flint: An Archaeological Perspective. Lithic Technology 35:1.
San Antonio, Texas.
Shepherd, W., 1972. Flint: Its Origin, Properties and Uses.
London.
Sieveking, G. de G., & Hart, M.B. (eds), 1986. The Scien- tific Study of Flint and Chert.Cambridge.
Sieveking, G. de G., & Newcomer, M.H. (eds), 1987.
The Human Uses of Flint and Chert.Cambridge.
Skak-Nielsen, N.V., 2004. The neolithisation of Scan- dinavia: How did it happen? Adoranten 2004. Ta- numshede.
Stockmans, C.; Bosmans, H. & Vermeersch, P., 1981.
Trace element analysis of Belgium flint mine pro- ducts. Staringia 6. Pinneberg.
Thomsen, E., 2000. Flintens geologi og mineralogi.
Eriksen, B.V. (ed.). Flintstudier: En håndbog i system- atiske analyser af flintinventarer. Aarhus.
Van Gijn, A., 2010. Flint in Focus. Lithic Biographies in the Neolithic and Bronze Age. Leiden.
