New Data On The Exploitation Of Obsidian In The Southern Caucasus (Armenia, Georgia) And Eastern Turkey, Part 2: Obsidian Procurement From The Upper Palaeolithic To The Late Bronze Age, Archaeometry, (http://dx.doi.org/10.1111/arcm.12007).

Abstract

Within the framework of the French archaeological mission ‘Caucasus’, in a previous paper we have presented new geochemical analyses on geological obsidians from the southern Caucasus (Armenia, Georgia) and eastern Turkey. We present here the second part of this research, which deals with provenance studies of archaeological obsidians from Armenia. These new data enhance our knowledge of obsidian exploitation over a period of more than 14 000 years, from the Upper Palaeolithic to the Late Bronze Age. The proposed methodology shows that source attribution can be easily made by plotting element contents and element ratios on three simple binary diagrams. The same diagrams were used for source discrimination. As the southern Caucasus is a mountainous region for which the factor of distance as the crow flies cannot be applied, we have explored the capacity of the Geographic Information System to evaluate the nature and patterns of travel costs between the sources of obsidian and the archaeological sites. The role of the secondary obsidian deposits, which enabled the populations to acquire raw material at a considerable distance from the outcrops, is also considered.

Full text

NEW DATA ON THE EXPLOITATION OF OBSIDIAN IN THE
SOUTHERN CAUCASUS (ARMENIA, GEORGIA) AND EASTERN
TURKEY, PART 2: OBSIDIAN PROCUREMENT FROM THE
UPPER PALAEOLITHIC TO THE LATE BRONZE AGE*
C. CHATAIGNER
Archéorient,UMR 5133,CNRS/Université Lyon 2,7 rue Raulin, 69007 Lyon, France
and B. GRATUZE†
IRAMAT CEB,UMR 5060,CNRS/Université d’Orléans,3 D rue de la Férollerie, 45071 Orléans Cedex 2, France
Within the framework of the French archaeological mission ‘Caucasus’, in a previous paper
we have presented new geochemical analyses on geological obsidians from the southern
Caucasus (Armenia, Georgia) and eastern Turkey. We present here the second part of this
research, which deals with provenance studies of archaeological obsidians from Armenia.
These new data enhance our knowledge of obsidian exploitation over a period of more than
14 000 years, from the Upper Palaeolithic to the Late Bronze Age. The proposed methodology
shows that source attribution can be easily made by plotting element contents and element
ratios on three simple binary diagrams. The same diagrams were used for source discrimi-
nation. As the southern Caucasus is a mountainous region for which the factor of distance as
the crow flies cannot be applied, we have explored the capacity of the Geographic Information
System to evaluate the nature and patterns of travel costs between the sources of obsidian
and the archaeological sites. The role of the secondary obsidian deposits, which enabled the
populations to acquire raw material at a considerable distance from the outcrops, is also
considered.
KEYWORDS: OBSIDIAN PROCUREMENT, LESSER CAUCASUS, ARMENIA, GEORGIA,
EASTERN TURKEY, UPPER PALAEOLITHIC, MESOLITHIC, NEOLITHIC, BRONZE AGE,
LA–ICP–MS ANALYSES, GEOGRAPHIC INFORMATION SYSTEM (GIS)
INTRODUCTION
In Armenia, obsidian represents more than 90% of the material used by prehistoric populations
for their tools and weapons. Indeed, obsidian deposits are plentiful in Armenia as well as beyond
the periphery of its territory, in neighbouring regions such as southern Georgia, western Azer-
baijan and eastern Turkey (Fig. 1).
Analysis of the chemical composition of these sources (Keller et al. 1996; Blackman et al.
1998; Poidevin 1998) and of artefacts coming from approximately 70 Transcaucasian archaeo-
logical sites dating from the sixth to the first millennia bc (Badalyan et al. 2004) have enabled the
establishment of an initial cartography of the movements of obsidian between the Neolithic and
the Iron Age, and confirmation of the great variability in their distribution in the region. The
*Received 28 May 2012; accepted 5 October 2012
†Corresponding author: email gratuze@cnrs-orleans.fr
bs_bs_banner
Archaeometry 56, 1 (2014) 48–69 doi: 10.1111/arcm.12007
© 2013 University of Oxford

villagers obtained their supplies either from a single source or from several sources, and the
nearest deposits were not necessarily the most favoured.
In order to explain these phenomena, complementary studies have been carried out. A series
of chemical characterizations has enriched the database and provided new information on the
exploitation of the material. Amodel of the supply routes between the archaeological sites and the
sources of obsidian has been made using a Geographic Information System (GIS) in order to
assess the real ‘cost’ that the direct acquisition of the material represented for the prehistoric
populations and to better understand how it circulated.
DATA
Archaeological sites
The archaeological samples studied come from sites situated in different regions of Armenia and
relate to periods extending from the final Upper Palaeolithic to the Late Bronze Age/Early Iron
Figure 1 The distribution of the obsidian sources and archaeological sites studied in the southern Caucasus and eastern
Turkey.
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 49
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

Age, from 15 000 to 1 000 cal bc (Table 1). This study enables presentation of the diversity of the
sources of supply and the methods of acquisition through time and space.
The samples come from sites (Kalavan-1, Kmlo, Aratashen, Aknashen and Godedzor) exca-
vated conjointly by the Institute of Archaeology and Ethnography of Yerevan and the French
‘Caucasus’ mission, as well as from earlier excavations (samples provided by R. Badalyan,
Institute of Archaeology and Ethnography of Yerevan).
These different sites will be described in the presentation of the results.
METHODS
On the archaeological sites of Transcaucasia, there are thousands of artefacts in obsidian (more
than 20 000 at Aratashen; Badalyan et al. 2007); but for each site only a small number could
be submitted for chemical analyses, for administrative restrictions (export authorization). The
sample is therefore not representative: the analytical results give only a very partial picture of the
diversity of the supply, as only the major sources are identified.
Visual discrimination
Analysis based on visual examination is commonly considered a low-cost, non-destructive
technique to ‘source’ large numbers of obsidian artefacts on site. However, the major question is
how efficient this method of sourcing is forAnatolian and Transcaucasian obsidian (Frahm 2010).
This method can provide interesting results when the number of obsidian sources exploited is
small, and when their physical and chemical characteristics are clearly differentiated. Unfortu-
nately, this is not the case for Anatolia and Transcaucasia, where the sources of obsidian are
numerous and the same varieties (texture and colour) can be found in different deposits. Indeed,
texture and colour are related to the process of dehydration of the material during the eruption
(Moriizumi et al. 2009; Seaman et al. 2009) and not to its geochemical signature.
Therefore, a single flow can produce obsidian of different macroscopic types (the complex at
Gutansar produces black, grey, transparent, brown, red and black obsidian, etc.) and the same
varieties can exist on different volcanoes (the red mottled with black variety can be found at
Geghasar near Lake Sevan, at Kamakar in the Tsaghkunyats range, and also at Chikiani, in
southern Georgia).
Table 1 The archaeological sites
Site Region Altitude (m) Period Date cal BC
Kalavan-1 Mountains to the north of Lake Sevan 1 630 Upper Palaeolithic 15 300–14 000
Kmlo-2 Eastern piedmont of the Aragats 1 760 Mesolithic 9 000–7 500
Aratashen Ararat plain 870 Late Neolithic 6 000–5 400
Godedzor Syunik mountains (south-east Armenia) 1 800 Late Chalcolithic 3 600–3 300
Gegharot North-eastern edge of the Tsaghkahovit plain,
to the north of the Aragats
2 120 Early Bronze
Late Bronze
2 600–2 400
1 500–1 200
Karmrakar Upper valley of the Pambak (north-west Armenia) 1 790 Early Bronze 2 600–2 400
Lusaghbyur Upper valley of the Pambak (north-west Armenia) 1 780 Early Bronze 2 600–2 400
Hnaberd Northern piedmont of the Aragats massif 2 340 Late Bronze 1 500–1 200
Getashen Southwestern bank of Lake Sevan 1 950 Late Bronze 1 500–1 200
Keti Northern edge of the Shirak plain (north-west
Armenia)
1 900 Late Bronze 1 500–1 200
50 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

A visual discrimination test, enabling the suggestion of provenance, has recently been carried
out on an obsidian collection from the Aegean and Turkey (Carter and Kilikoglou 2007). This
collection was then subjected to geochemical analysis; the result of the experiment is undeniable:
‘The trace-elemental analyses have shown our visually discriminated source assignations to be
deeply flawed . . .’In particular, among the 42 samples attributed visually to a non-Aegean origin,
only five (i.e., only 12%!) were consistent with this criterion, the remaining 37 clearly coming
from the Aegean (Melos).
For this reason, analytical methods remain necessary for certain identification of the origin of
the obsidian—and it is the perfecting of these analytical methods, to provide the possibility of
dealing rapidly with a large number of samples, that will enable a more objective picture of how
and where the obsidian was obtained. Recent studies (Forster and Grave 2012; Williams et al.
2012) have shown that fairly reliable results can be obtained by using portable XRF instrumen-
tation. As pointed out by these authors, there must be a dual approach, combining both field
measurements with portable instrument carried out on a large population of artefacts and labo-
ratory analyses of a selected number of samples. Field measurements done on a large number of
objects will allow a first grouping and a first attempt at source attribution, while more complete
laboratory analysis will enable the separation of overlapping groups and the proposal of more
secure source attribution.
LA–ICP–MS analysis
Analyses of obsidian objects conducted at the Centre Ernest-Babelon of the IRAMAT (Orléans)
were carried out according to the analytical protocol described in Part 1 of this paper (Chataigner
and Gratuze 2013).
GIS modelling
The southern Caucasus is a mountainous region and the factor of distance as the crow flies cannot
be applied. We have thus explored the capacity of the Geographic Information System (GIS) to
integrate spatial data (relief, hydrography etc.) to analyse more realistically the movement of
persons and materials across this territory (Chataigner and Barge 2008).
The first stage of our analysis was to evaluate the nature and patterns of travel costs between
the sources of obsidian and the archaeological sites, in order to understand what was actually
involved for the prehistoric peoples who sought this material (climatic factors, vegetation, rivers
to cross, distance covered, slopes climbed, weight of material transported, ‘political’ boundaries
etc.). The topographical element, with elevations often higher than 3000 m and deep valleys,
appeared to be the main constraint on travelling.
Transportation in the prehistoric periods would mainly have been either on human backs or,
from the Late Neolithic onwards, on the backs of oxen; equids were domesticated locally only in
the Early Bronze Age. From different experiments (Scott and Christie 2004), we inferred that a
reasonably trained walker could accomplish an average speed on the flat of 5 km h–1, for a load
between 25 and 30 kg and a walking time of 7–8 hours per day. According to ethnographic
studies, the pack oxen can carry loads from 50 to 90 kg, with an average speed on the flat of
4kmh
–1; the walking time is reduced to about 5 hours per day for draft animals, as they need to
stop to graze.
The second stage was to calculate, using the ‘Spatial Analysis’ functions of ESRI’s ArcGIS®,
the time needed to journey between the sources of obsidian and the different archaeological sites
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 51
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

and to recreate the most efficient travel pathways through the region. The main travel cost being
the topography, the ‘cost surface’ of the GIS has been defined by the time needed for a walker,
according to the slope and the distance (Eastman 1999). Then the ‘cost-weighted’ and ‘least-cost
path’ analysis functions enabled calculation of the time needed to go from one point to another
on the map (equal to the time between sources and villages), as well as to memorize the pathways
requiring the least effort and the least time in relation to the distance to be travelled and the slope
(that is, the best route to take).
RESULTS AND DISCUSSION
Chemical results from archaeological obsidian
The 136 artefacts studied were analysed using the old (Gratuze 1999) or the recent (as described
in Part 1 of that paper) analytical protocols. The obtained data were plotted on the same diagrams
as used to discriminate the geological sources (Figs 2 and 3). Among these artefacts, 130 could
be related to the various chemical groups defined for the geological samples, while six of them
could not be related to any of the geological groups defined above (Table 2). Twenty-eight
artefacts come from the Arteni complex [Arteni 1 (one), Arteni 2 (five) and Arteni 3 (22)], 23
from Gutansar, 22 from Syunik [Syunik 2 (seven) and Syunik 3 (15)], 19 from Tsgahkunyats
[Tsgahkunyats 1 (17) and Tsgahkunyats 2 (two)], 18 from the surroundings of Sarikamis
0
100
200
300
400
500
600
700
800
900
1000
0 50 100 150 200 250
Zr ppm
Ba ppm
Arteni
Hat i s
Gegham
Gutansar
Akhurian/Sarikamis North
Araxes/Sarikamis South
Syunik
Tsag hkunyats
Arteni 1 artefacts
Arteni 2 artefacts
Arteni 3 artefacts
Gegham artefacts
Gutansar artefacts
Hatis artefacts
Sarikamis North artefacts
Sarikamis South artefacts
Syunik artefacts
Tsaghkunyats artefacts
Unattributed artefacts
Figure 2 The binary diagram for the Zr–Ba contents of the studied artefacts and related outcrops.
52 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

[Sarikamis North 11 (Hamamli six/Handere five); Sarikamis South seven (Sarikamis South 1
one/Sarikamis South 2 six)], 12 from the Hatis Mountain [Hatis 1 (one) and Hatis 2 (11)] and
eight from the Gegham Mountains. In order to simplify the graphical representation, only the
seven volcanic complexes involved in obsidian distribution will be represented on the diagrams
that follow.
The six unattributed artefacts form two groups containing two and four artefacts, respectively.
They are both characterized by high concentrations of barium (407 and 506 ppm, respectively)
and low zirconium contents (57 and 86 ppm). The composition of these artefacts has been
compared with the composition of the obsidian from central and eastern Turkey, which contains
similar barium, zirconium and strontium concentrations (Figs 4 and 5): Erzincan, west Erzurum
1, Digor/Yaglica (Chataigner et al. 2013), Nenezi Dag, Bingöl B, Acigöl and Göllü Dag 1 and 2.
Two of the unattributed artefacts, however, show several similarities with the obsidian from the
Arteni 3 group (Figs 4 and 5). The main difference lies in their barium content, which is higher
in the artefacts while the other element contents are similar (Table 3). We could thus consider that
these two artefacts may either come from unsampled obsidian outcrops of the Arteni mountains
or belong to the obsidian with the variability of Arteni 3, which perhaps shows a continuous
variation in barium concentration, as observed at Chikiani. This group will be referred to as
Arteni 3b. As mentioned above, it demonstrates that a new systematic sampling and a new set of
analyses of the Arteni outcrops is necessary.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 0.5 1.0 1.5 2.0
Nb/ Z r
Y/Zr
Arteni
Hat i s
Gegham
Gutansar
Akhurian/Sarikamis North
Araxes/Sarikamis South
Syunik
Tsag hkunyats
Arteni 1 artefacts
Arteni 2 artefacts
Arteni 3 artefacts
Gegham artefacts
Gutansar artefacts
Hatis artefacts
Sarikamis North artefacts
Sarikamis South artefacts
Syunik artefacts
Tsag hkunyats a rtef acts
Unattributed artefacts
Figure 3 The binary diagram of the Nb/Zr–Y/Zr ratios for the studied artefacts and related outcrops.
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 53
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

Table 2 Repartition of the studied artefacts among the different obsidian sources
Sources
(nb)*
Outcrops
(nb)*
Aratashen
(30)
Gegharot
EBA (13)
Gegharot
LBA (e)
Getashen
(2)
Godedzor
(21)
Hnaberd
(6)
Kalavan
(18)
Karmrakar
(10)
Keti
(7)
Kmlo
(20)
Lusaghbyur
(1)
Arteni (30) Arteni 1 (1) 1
Arteni 2 (5) 2 1 11
Arteni 3 (22) 12 1 2 1 3 2 1
Arteni 3b (2) 1 1
Hatis (12) Hatis 1 (1) 1
Hatis 2 (11) 1 10
Gegham (8) Gegham (8) 1 1 1 4 1
Gutansar (23) Gutansar (23) 5 1 1 3 3 10
Sarikamis
North (11)
Akhurian 1/Handere (5) 1 31
Akhurian 2/Hamamli (6) 2 31
Sarikamis
South (7)
Sarikamis South 1
Sarikamis South 1 (1)
1
Sarikamis South 2
Mescitli/Sehitemin (6)
5 1
Syunik (22) Satanakar or Syunik 2 (7) 6 1
Sevkar or Syunik 3 (15) 15
Tsaghkunyats
(19)
Damlik-Ttvakar or
Tsaghkunyats 1 (17)
86 1 2
Kamakar-Aykasar or
Tsaghkunyats 2 (2)
2
Digor/Yaglica
(4)
Yaglica South (4) 31
*nb, Number of artefacts related to volcanic complexes and obsidian sources.
54 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

The four other unattributed artefacts form a homogeneous group characterized by higher
zirconium and lower Nb/Zr and Y/Zr ratios. They have compositions close to those of several
sources (Acigöl, Nenezi Dag, Göllü Dag 2 and Yaglica South) with regard to their barium,
zirconium, strontium, yttrium and niobium contents (Poidevin 1998). However, these sources can
be distinguished by using other elements such as caesium and tantalum (Fig. 5). It then appears
that the only source that matches the composition of the four artefacts very closely is one of the
outcrops of the Digor area, referred to as Yaglica South (Chataigner et al. 2013). We shall,
however, note that for Yaglica, on the six different obsidian blocks analysed, the Yaglica South
subgroup was defined by only two samples (Table 3). It is therefore difficult to circumscribe the
whole variability of that chemical subgroup. At that time, the relationships established between
these four artefacts and Yaglica South have to be considered as the most probable hypothesis, but
this has to be confirmed by new systematic sampling and a new set of analyses of these outcrops.
Archaeological sites
The archaeological samples studied come from sites in different regions of Armenia. concerning
periods extending from the final Upper Palaeolithic to the Late Bronze Age/Early Iron Age; that
is, from 15 000 to 1 000 cal bc. This study provides evidence of the diversity of the sources of
supply and the methods of acquisition over time in different environmental and socio-economic
contexts.
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.0 0.1 0.2 0.3 0.4 0.5
Nb/Z r
Y/Zr
Arteni 3, Pokr Arteni and Aragats flow
Yaglica
Erzincan
West Erzurum 1
Bingöl B
Sakaeli Orta
Acigöl
Nenezi Dag
Göllü Dag 1
Göllü Dag 2
Arteni 3b artefacts
Unattributed artefacts
Yaglica South
Yaglica Summit
Figure 4 The binary diagram of the Nb/Zr–Y/Zr ratios for the unattributed artefacts and obsidian outcrops with similar
Ba and Zr contents.
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 55
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

Final Upper Palaeolithic: Kalavan-1 At the last glacial maximum (20 000–18 000 bp), most of
the Lesser Caucasus range was covered by glaciers and therefore deserted. During the warming
at the end of the Pleistocene, human populations came to progressively reoccupy the Lesser
Caucasus, as indicated by the site of Kalavan-1, recently discovered in the mountains overlooking
the northern edge of Lake Sevan and dated by 14C to the 15th millennium cal bc (Liagre et al.
2009; Montoya et al. 2013). The lithic industry of Kalavan-1 is linked to the Epigravettian
tradition and has parallels in western Georgia (Ortvale Klde, Dzudzuana and Sabelisopeli), in
Table 3 Average compositions and standard deviations for the different artefacts compositional groups. The number
of samples attributed for each groups is given in brackets. Empty cells in the table are due to the fact that some
elements were not determined in early analyses
Source Li B Na2O MgO Al2O3SiO2K2O CaO Ti Mn Fe Zn Rb Sr Y Zr Nb
Arteni 1
artefacts
Average (1) 58.9 44.4 4.11 0.05 14.5 75.5 4.02 0.58 341 701 3 377 41.7 145 8.5 24.2 45.0 39.7
SD
Arteni 2
artefacts
Average (5) 55.5 42.6 4.14 0.06 13.3 76.1 4.40 0.61 449 568 3 837 44.5 129 15.0 17.3 49.9 29.7
SD 4.5 1.7 0.04 0.006 0.3 1.0 0.07 0.08 49 7 244 1.2 6 3.6 0.7 4.5 1.4
Arteni 3
artefacts
Average (22) 49.3 39.0 4.15 0.07 13.2 75.9 4.48 0.63 557 509 4 540 37.7 117 25.0 15.6 57.9 26.2
SD 3.2 3.9 0.10 0.02 0.6 0.9 0.13 0.05 67 36 581 5.8 7 5.2 1.5 7.1 2.0
Arteni 3b
artefacts
Average (2) 48.0 40.3 4.04 0.08 13.2 75.5 4.29 0.59 592 466 4 567 33.4 110 30.9 13.9 57.4 22.7
SD 1.8 0.7 0.08 0.02 0.1 0.03 0.02 30 25 359 8.5 3 0.6 0.0 3.7 0.6
Gegham
artefacts
Average (8) 73.0 45.3 4.41 0.05 14.0 75.4 4.16 0.61 381 603 3 457 30.3 187 7.5 16.8 46 45.7
SD 11.3 4.6 0.10 0.007 0.01 0.6 0.04 0.07 31 70 367 4.4 17 1.2 2.1 7 3.7
Gutansar
artefacts
Average (23) 62.0 31.1 4.21 0.21 75.0 3.74 1.04 1 014 592 7 928 47.6 132 90.6 16.4 124.3 34.5
SD 5.2 2.6 0.06 0.006 0.4 0.09 0.13 73 41 688 10.7 8 9.9 1.3 11.1 1.9
Hatis 1
artefacts
Average (1) 51.0 30.6 75.0 1.28 656 493 6 548 33.5 110 92.8 12.1 70.5 22.6
SD
Hatis 2
artefacts
Average (11) 42.9 28.4 75.0 1.48 1 213 496 11 558 46.7 93 167.3 10.4 96.9 19.5
SD 2.2 1.6 0.03 56 36 1 165 4.9 4 20.7 0.7 7.3 0.9
Akhurian 1
artefacts
Average (5) 55.7 31.1 4.68 0.06 12.7 75.5 4.16 0.48 572 622 6 557 64.2 136 6.3 33.4 138.0 27.7
Handere SD 8.1 2.9 0.12 0.008 0.7 0.10 0.06 53 919 12 0.5 2.0 25.1 1.9
Akhurian 2
artefacts
Average (6) 54.2 30.2 4.50 0.04 75.6 4.12 0.47 492 7 361 134 1.8 35.5 161 26.7
Hamamlı SD 8.4 1.5 0.22 0.02 0.7 0.11 0.06 23 543 7 0.3 2.1 18 1.1
Sarikamis
South 1
artefacts
Average (1) 65.7 26.2 4.77 0.07 14.0 74.6 4.19 0.74 551 597 9 522 60.2 134 20.6 30.3 130.9 19.4
SD 0.00
Sarikamis
South 2
artefacts
Average (6) 37.7 26.5 3.91 0.07 13.3 75.6 4.16 0.54 532 299 5 398 30.8 120 16.3 15.2 70.1 11.7
Mescitli/
Sehitemin
SD 6.0 1.4 1.1 0.02 11 22 111 1.2 5 1.2 0.7 5.8 0.5
Syunik 2
artefacts
Average (7) 60.5 22.0 75.0 0.48 481 428 3 963 37.2 167 7.4 6.9 62.6 32.4
SD 7.9 6.3 0.1 0.04 27 23 408 2.1 20 0.8 0.4 3.6 3.2
Syunik 3
artefacts
Average (15) 56.2 25.0 75.0 0.53 562 399 4 225 36.2 156 13.7 7.5 72.4 31.6
SD 2.9 3.3 0.1 0.02 31 17 470 3.7 9 1.2 0.5 4.5 1.6
Tsaghkunyats
1 artefacts
Average (17) 42.6 24.6 4.16 0.11 75.7 4.19 0.87 608 409 5 800 33.5 107 112.3 7.8 67.0 20.2
SD 9.4 1.9 0.09 0.03 0.6 0.18 0.05 20 2 671 5.3 7 10.4 0.9 5.3 1.2
Tsaghkunyats
2 artefacts
Average (2) 35.5 22.0 75.0 1.05 776 387 7 947 38.5 78 174.3 5.7 102 17.0
SD 3.6 1.3 0.08 10 0 86 1.4 6 3.4 0.0 1 0.3
Yaglica South
artefacts
Average (4) 33.7 36.5 3.85 0.16 13.9 75.4 4.33 0.74 829 328 7 177 30.3 114 37.4 10.7 86 14.7
SD 9.0 6.5 0.42 0.02 0.6 0.4 0.45 0.03 47 7 206 1.6 3 2.6 0.9 5 0.6
Yaglica South Average (2) 36.5 35.3 4.13 0.13 14.0 75.4 4.14 0.81 861 424 6 819 37.3 121 54.2 12.5 97 18.2
SD 2.5 2.4 0.16 0.01 1,2 1,1 0.13 0.08 35 15 431 2.8 12 3.8 1.4 10 1.2
56 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

contexts of the 17th to the 12th millennia bc (Nioradze and Otte 2000; Nioradze 2001; Mesh-
veliani et al. 2007).
At this site, it is striking that most of the tools are in obsidian (about 66%), although it is an
exogenous rock, while the local siliceous rocks transported abundantly by the Barepat River
played a secondary role. The quantity and the variety of the obsidian found at Kalavan-1
(translucent, smoky grey, sparkling black, very dull black, red, mottled brown etc.) suggest a
diversity of sources. The cultural links (Epigravettian community) between Kalavan-1 and the
Georgian sites of Ortvale Klde or Dzudzuana could also be reflected in the obsidian procurement,
Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Th U Ba/Zr Ba/Sr Nb/Zr Y/Zr
4.1 31 9.3 23.6 2.2 8.1 2.8 0.25 3.2 0.63 4.5 0.91 2.8 0.42 3.2 0.42 2.5 2.5 14.0 7.8 0.68 3.6 0.88 0.54
3.5 125 12.4 29.2 2.7 9.3 2.4 0.25 2.5 0.47 3.2 0.65 2.0 0.29 2.3 0.31 2.8 2.0 12.0 7.3 2.47 8.2 0.60 0.35
0.2 46 1.6 3.0 0.2 0.7 0.1 0.01 0.1 0.04 0.2 0.04 0.2 0.02 0.1 0.02 0.0 0.1 0.3 0.7 0.69 1.3 0.06 0.03
3.1 285.0 17.0 38.1 3.6 12.3 2.5 0.34 2.4 0.44 2.8 0.67 1.8 0.27 2.1 0.32 2.9 1.7 12.5 6.9 4.92 12.2 0.45 0.28
0.4 23.5 3.9 7.2 0.9 2.7 0.1 0.06 0.2 0.03 0.3 0.08 0.2 0.02 0.2 0.01 0.4 0.2 1.5 0.5 0.61 0.9 0.02 0.02
2.8 407.1 19.5 41.9 3.2 10.4 2.5 0.38 2.2 0.40 2.6 0.57 1.7 0.27 2.0 0.29 2.3 1.4 12.0 6.0 7.10 13.2 0.40 0.24
0.2 19.6 1.1 2.6 0.2 1.2 0.2 0.0 0.3 0.1 0.11 0.4 0.015 0.0155
6.8 10 12.6 29.4 2.7 9.5 2.6 4.2 25.1 16.5 0.22 1.4 1.01 0.37
0.8 1 1.8 3.1 0.5 1.5 0.4 0.5 6.4 4.5 0.03 0.1 0.14 0.03
4.7 381 26.3 52.0 4.3 14.8 4.4 2.6 15.9 10.1 3.07 4.2 0.28 0.13
0.4 29 2.5 3.1 0.3 1.1 0.5 0.2 1.6 1.3 0.14 0.2 0.02 0.00
4.3 498.0 23.5 45.8 3.7 12.8 2.8 1.9 15.7 10.1 7.06 5.36 0.32 0.17
3.5 536 27.4 53.0 4.4 14.7 3.4 1.5 13.1 7.9 5.55 3.2 0.20 0.11
0.4 30 1.3 2.2 0.3 0.5 0.5 0.2 0.7 0.8 0.25 0.3 0.02 0.01
4.2 100.3 29.4 65.5 4.1 15.1 4.1 0.26 4.0 0.77 5.2 1.12 3.3 0.50 3.9 0.54 3.9 1.7 15.9 7.1 0.75 16.1 0.21 0.25
0.5 5.1 5.0 8.4 1.0 0.4 0.18 1.1 0.06 0.05
4.1 30 30.6 68.6 16.8 7.1 0.19 16.9 0.17 0.22
0.2 1 2.9 4.7 1.2 0.5 0.03 2.8 0.02 0.02
4.4 373.8 30.4 67.8 6.4 25.3 6.3 0.76 5.7 0.96 6.1 1.26 3.6 0.57 4.0 0.56 4.1 1.2 13.8 6.8 2.86 18.2 0.15 0.23
3.9 437 19.9 42.2 3.9 14.2 2.9 0.40 2.6 0.45 2.9 0.66 1.9 0.31 2.3 0.35 3.1 1.0 16.7 7.4 6.25 27.0 0.17 0.22
0.4 26 1.0 1.9 0.3 1.1 0.3 0.1 3.3 1.1 0.30 2.6 0.01 0.02
4.3 15.9 22.4 40.9 3.0 8.4 2.8 2.0 28.9 10.6 0.26 2.1 0.52 0.11
0.5 3.4 1.5 2.5 0.2 1.1 1.1 0.3 5.4 2.0 0.06 0.3 0.04 0.01
3.8 37.0 27.9 49.9 4.0 10.4 3.1 2.1 29.2 10.2 0.51 2.7 0.44 0.10
0.5 4.8 1.5 1.6 0.3 2.0 0.7 0.5 3.1 0.9 0.08 0.2 0.02 0.01
3.6 597 29.6 54.2 4.2 13.9 2.4 1.5 20.9 9.6 8.95 5.3 0.30 0.12
0.4 28 2.0 2.3 0.1 0.4 0.0 0.1 3.0 0.8 0.59 0.4 0.02 0.02
2.6 916 42.4 73.0 5.3 16.1 3.1 1.1 25.9 9.7 8.97 5.3 0.17 0.06
0.0 26 0.2 2.4 0.1 0.0 0.1 0.0 0.1 0.7 0.34 0.2 0.005 0.000
3.7 506.6 22.7 43.8 3.3 11.1 1.9 0.33 1.5 0.28 1.8 0.39 1.2 0.19 1.5 0.23 2.6 1.1 16.2 7.9 5.93 13.6 0.17 0.12
0.1 12.1 1.2 1.5 0.2 0.7 0.4 0.1 0.2 0.3 0.21 0.6 0.01 0.003
3.9 530 28.8 48.6 4.0 12.7 2.2 0.43 2.5 0.35 2.0 0.43 1.3 0.21 1.6 0.25 2.8 1.2 17.7 7.5 5.46 9.8 0.19 0.13
0.5 33 3.8 5.3 0.4 1.4 0.2 0.11 0.9 0.07 0.2 0.05 0.2 0.03 0.2 0.03 0.3 0.1 1.8 0.6 0.22 0.2 0.01 0.002
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 57
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

the populations of Georgia mainly exploiting the deposits of Chikiani, in the south of the country
(Le Bourdonnec et al. 2012). But the analyses of the origin of the artefacts of Kalavan-1 have
revealed a completely different system of supply.
The 18 artefacts that were analysed by LA–ICP–MS are flakes and not retouched pieces.
The results of the analyses (Table 2) show that most of the supply came from deposits situated
to the west of Lake Sevan (Hatis, Gutansar and Geghasar), while only one sample comes from
south-east of Lake Sevan (Sevkar) (Fig. 6). Modelling of the time taken to cover the distance
between the Kalavan site and the obsidian sources, depending on the relief, shows that between
20 and 30 hours would have been required to reach the three main deposits (Gutansar,
Geghasar and Hatis), or 3–4 days of walking. It thus appears probable that the hunters of
Kalavan first sought a supply of obsidian as they moved through the mountains around Lake
Sevan, in order to prepare adequate weaponry for hunting caprins (mouflons), the bones of
which were found in abundance on the site. At the Upper Palaeolithic site of Ortvale Klde in
Georgia, hunting activities were structured according to the migratory behaviour of the Cau-
casian tur (Capra caucasica), which made them locally abundant on a seasonal basis (Adler
et al. 2006). Likewise, Kalavan-1 could have been a key site for the ambush of mouflon herds
during their seasonal movements between the mountains overlooking Lake Sevan and the low-
lands of the Aghstev Valley.
Mesolithic/Early Neolithic: Kmlo-2 The Mesolithic and Early Neolithic are very poorly known
phases in Armenia. New data have been provided by the Franco-Armenian excavations at
Kmlo-2, a small cave situated in the canyon of the Kasakh River, on the eastern flank of the
0.9
1.6
2.3
3.0
25811
Cs ppm
Ta ppm
Unattributed artefacts
Arteni 3b ar tefac ts
Yaglica South
Yaglica Summit
Arteni 1
Arteni 2
Arteni 3
Erzincan
West Erzurum 1
Bingöl B
Sakaeli Orta
Göllü Dag 1
Göllü Dag 2
Acigöl
Nenezi Dag
Figure 5 The binary diagram for the Cs–Ta contents of the unattributed artefacts and obsidian outcrops with similar Ba
and Zr contents.
58 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

Aragats massif (Arimura et al. 2010). The 14C dates suggest a succession of several occupations,
which form levels IV (end of the 10th and first half of the ninth millennia bc) and III (second half
of the ninth and beginning of the eighth millennia bc).
The lithic industry of Kmlo-2 is almost exclusively in obsidian. Made on the site, as shown by
the numerous debitage products, this material includes a high proportion of microliths (~30%),
as well as artefacts characterized by abrupt, parallel and regular retouch, clearly carried out by
pressure. These ‘Kmlo tools’, which appear at the transition between levels IV and III, are
reminiscent, from a techno-typological point of view, of other artefacts in obsidian, present in the
cultures of neighbouring regions:
• the ‘Çayönü tools’, spread over the northern Near East, in the eighth and seventh millennia bc
(Pre-Pottery Neolithic B and Early Pottery Neolithic); and
• the ‘hook-shaped tools’ of the Pre-Pottery Neolithic culture of Paluri-Nagutny, which devel-
oped on the south-western slopes of the Greater Caucasus, and then in southern Georgia—the
only 14C date known for this culture belongs to the mid-eighth millennium cal bc (Kotias Klde,
Neolithic layer; Z. Matskievich, pers. comm.).
The obsidian found at Kmlo belongs to different varieties: opaque black, opaque grey, red,
red–mottled black, marbled red and black, transparent, translucent. The analysis of 20 ‘Kmlo
tools’ showed that these artefacts were knapped in various types of obsidian, all local (Armenia
and north-eastern Turkey) (Fig. 7).
The source used the most was Gutansar (50% of the samples). In second place was the
Tsaghkunyats range (20% of the samples), the two subgroups Damlik-Ttvakar and Kamakar-
Aïkasar being represented. The Kasakh River flows alongside the chain of Tsaghkunyats in its
Figure 6 Isochrones of 7 hours between Kalavan-1 and the obsidian sources.
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 59
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

upper course and transports numerous blocks of obsidian. Several artefacts from Kmlo have
retained the cortex of the pebbles rolled down by the river; these were then collected in the
Kasakh, which flows at the foot of the Kmlo cave.
The complex of Arteni represents 15% of the samples. The three other sources each repre-
sented by a sample, are the Hatis volcano, near the complex of Gutansar, the Geghasar deposit to
the south-west of Lake Sevan, and the deposits of Sarikamis South (Mescitli, Sehitemin).
Modelling of the routes between Kmlo and these different sources of obsidian shows that they
were divided into three main directions: (a) towards the north and the Tsaghkunyats range, 1
day’s walk (about 7 h) following the Kasakh Valley, but the river itself clearly played the role of
secondary source; (b) towards the east, with the Gutansar complex situated also at a day’s walk
from Kmlo, then the Hatis volcano (close to Gutansar) and, further to the south-east, the Geghasar
highlands at 3 days’ walk; and (c) towards the west with the deposits of Arteni at 2 days’ walk
(~15 h), and then, by crossing the Kars plateau, the obsidian deposits of Sarikamis South, at
5 days’ walk.
The deposits of Sarikamis South are far from Kmlo and the shortest route crosses the obsidian
outcrops of theYaglica Dag volcano, which was not identified among the sources exploited by the
Kmlo human group. Astudy on obsidian procurement in California (Eerkens et al. 2008) showed
that hunter–gatherers had a high degree of mobility and often ignored smaller intermediary
sources where the glass was of poorer quality. In the upper part of theYaglica Dag, the obsidian
is full of inclusions and not very suitable for knapping; but on the southern flank of this volcano,
Figure 7 The ‘least cost paths’ between Kmlo and the sources of obsidian.
60 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

the obsidian is homogeneous and of a high quality (Chataigner et al. 2013). Nevertheless, besides
direct procurement, other hypotheses can be considered: the existence of secondary deposits in
the Araxes Valley (a deposit should exist near the village of Gaziler, according to Poidevin 1998)
or an exchange with populations of the Kars region.
Mobility, exchange and social interaction were integral components of Mesolithic hunter–
gatherer lifeways (Lovis et al. 2006). The use of lithics at 100 km or more from their sources is
commonly attributed to long-distance logistical movements that required either negotiation with
groups local to the source areas or transactions during seasonal aggregation ceremonies (Sulgos-
towska 2006). It must be added that near the confluence of the Akhurian and the Araxes, at
Tuzluca (Gokhp or Koulpi, for the Armenians) is a mountain of gem salt, which played a very
important part in the Middle Ages and in recent centuries in supplying Armenia, Georgia and
eastern Turkey (Ouoskherdjan 1828; Karajian 1920). Beds of tertiary (Miocene) rock salt are
widspread in eastern Anatolia, in particular in theAraxes Valley, between Kagizman and Tuzluca
(Yilmaz 2007). Such salt deposits may have been known in the early Holocene and served as
places of exchange, in this way enabling the redistribution of obsidians from neighbouring
outcrops.
The study of the supply of obsidian by the human group living at Kmlo therefore suggests a
fairly vast territory of routes, but with no link either with the region of Lake Van, where the
northern Mesopotamian cultures possessing Çayönü tools obtained their supply, or with the great
deposit of Chikiani, in southern Georgia, near to which are located the Paluri-Nagutny sites with
‘hook-shaped tools’ similar to those of Kmlo.
Late Neolithic: Aratashen At the very beginning of the sixth millennium bc, populations that
already possessed an advanced mastery of the domestication of plants and animals appeared
in the Kura basin (in Georgia and Azerbaijan) and the Araxes basin (in Armenia). These
were the cultures of Shulaveri-Shomutepe in the Kura basin (Kiguradze and Menadbe 2004)
and of Aratashen in the Araxes basin (Badalyan et al. 2007). Agriculture (wheat, barley, lentils)
and herding (sheep especially, and goats and cattle) were from then on the bases of the
economy.
The two levels of the tell of Aratashen (Badalyan et al. 2004, 2007) belong to the Neolithic
period (sixth millennium bc) and have produced an abundance of lithic tools in obsidian
(more than 20 000 artefacts), flint being extremely rare (fewer than 10 artefacts). The techno-
typological analysis having shown the existence of several methods of debitage (indirect percus-
sion, light pressure and pressure with levering), it was interesting to test whether a correlation
existed at Aratashen between the debitage techniques and the sources of obsidian. Thirty tools
from levels I and II were therefore analysed. Debitage by indirect percussion shows a predomi-
nance of Arteni (60%), then of Sarikamis South (27%) and finally of Gutansar (13%). However,
debitage using light pressure (with a crutch) shows a great variety of sources: 28% from Arteni,
21% from Gutansar, 21% from the sources of Sarikamis South and North, 15% from Hatis and
15% from Geghasar. Thus there would seem to be no obvious link between the origin of the
material and the debitage technique used.
An interesting element is the fact that the most exploited sources (Arteni, Sarikamis South and
Sarikamis North, which represent in all 77% of the material analysed) are situated to the west of
the Ararat plain (Fig. 8). Arteni (50%), which is visible from the site ofAratashen itself, lies about
11 hours’ walk (1.5 days) away. In the region of Kars, the deposits of Sarikamis South (17%) and
Sarikamis North (10%) are at a great distance from Aratashen (about 5 days’ walk). However,
these villagers were able to obtain their supply of Sarikamis obsidian in two other ways:
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 61
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

• Blocks from the sources of Sarikamis North are transported in large quantities by the Kars
River up to its confluence with the Akhurian River and further on. Deposits of obsidian pebbles,
forming a layer of more than 1 m thick, are still visible today in the cliffs of the Akhurian canyon,
especially in the region of Haykadzor, south of the dam lake.
• The important role played by husbandry at Aratashen perhaps obliged the inhabitants
of this village to gather salt often enough at places such as Tuzluca, which is about 1.5 days’
walk away, across the Ararat plain, and where they could meet inhabitants of the Sarikamis
region.
Clearly, the territory of circulation of the inhabitants of Aratashen was oriented towards the
west. And this tendency, which existed for 66% of the material in level IIb, increases to 90% in
level IIa. The only artefact from level I that has been analysed also comes from Sarikamis South.
The presence of the salt mountain of Tuzluca, which may have served as a place for meeting and
trade, can be an element of explanation.
Chalcolithic: Godedzor The village of Godedzor, established at an altitude of about 1800 m in
the mountains of the south-eastern Lesser Caucasus (Fig. 9), was probably the summer encamp-
ment of transhumant populations. Indeed, the region of Godedzor is covered by a thick layer of
snow from October to March, which makes the survival of animal herds very difficult during the
winter, as the ethnographic data show (Mkrtumyan 1974). Light architecture (numerous post
holes) confirms temporary occupation in this place. Moreover, the presence of painted pottery,
Figure 8 The ‘least-cost paths’ between Aratashen and the sources of obsidian exploited.
62 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

characteristic of the basin of Lake Urmiah in the Late Chalcolithic, is evidence of relations with
this region (Chataigner et al. 2010).
On the high plateaus that dominate Godedzor are several large deposits of obsidian (Satanakar,
Sevkar and Bazenk). The torrents flowing across these deposits carry many blocks towards the
Vorotan River, which passes through a canyon below the village of Godedzor. The 21 artefacts
analysed come all from the deposits of these high plateaus, predominantly from the sources of
Sevkar (71%), and the rest from the deposit of Mets Satanakar.
The fact that a certain number of artefacts bear strips of neo-cortex (surface ground down by
movement in the river), and the reduced size of most of them, shows that the pebbles found in the
Vorotan were actually exploited by the inhabitants of Godedzor, who also went on to the high
Figure 9 The main routes of communication between the region of Godedzor and the northern Near East.
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 63
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

plateaus or traded with the local populations, since large cores were found together in a cache on
the site.
The inhabitants of Godedzor, who used their cattle for transporting heavy loads, as is shown
by the pathological deformations observed on the vertebrae and phalanxes of several animals
(Chataigner et al. 2010), probably brought blocks and obsidian artefacts down to their winter
encampment in the basin of Lake Urmia. Indeed, the analyses carried out on the Chalcolithic sites
established around Lake Urmia (Pisdeli Tepe, Yanik Tepe etc.) show a predominance of ‘group
3C’ obsidian, coming from Syunik (Sevkar, Satanakar and Bazenk) (Renfrew and Dixon 1976;
Voigt 1983; Keller et al. 1996; Niknami et al. 2010).
Early Bronze: Gegharot, Karmrakar and Lusaghbyur From 3500 cal bc, the culture of the
Kuro-Araxes (Early Bronze Age) developed in Armenia and experienced great expansion during
the third millennium. The bicoloured pottery (black outside, red or light brown inside) that is
characteristic of this culture spread westwards as far as the Mediterranean Sea and eastwards to
the Caspian Sea. The sites of Gegharot, Karmrakar and Lusaghbyur, which belong to the late
phase of this culture (2600–2400 cal bc), are situated to the north of the Aragats massif.
Gegharot is located on a hill, on the southern flank of the Pambak range, on the north-
eastern edge of the Tsaghkahovit plain. The Early Bronze Age settlement occupies the summit
and the western flank of the hill: a tomb discovered at the western foot of the hill contained
a rough obsidian block weighing 8.860 kg (Badalyan and Avetisyan 2007). The analysis of this
block shows that it probably came from Arteni, nearly 14 hours’ or 2 days’ walk away by
skirting the Aragats massif to the west (Fig. 10). A source so far away for a block as heavy as
this suggests the use of pack-animals (cattle), but such a load could also have been transported
by a walker, for a special purpose. Moreover, most of the supplies at Gegharot come from the
deposits of Damlik-Ttvakar (Tsaghkunyats 1), situated less than 6 hours’ walk from the village.
One sample comes from the Gutansar complex, to the east of the massif, at about 12 hours’
walk.
Karmrakar and Lusaghbyur are located on the southern flank of the Shirak range, on the
northern edge of the Pambak Valley. The Karmrakar settlement sits on a triangular spur and the
Lusaghbyur settlement occupies a hill bordered by ravines. At Lusaghbyur, the only sample
analysed is from Arteni (about 2 days’ walk). As for the inhabitants of Karmrakar, most of their
supply (60%) came from the deposits of Sarikamis North, of which the blocks, transported by the
Kars River, had accumulated in the secondary deposits at the confluence of the Kars and Akhurian
Rivers, 1 day’s walk away (Fig. 10). From there, the inhabitants could cross theAkhurian and go
towards Yaglica Dag (30% of their supply) at 1 day’s walk, and then the deposits of Sarikamis
South (10%) at 2 days’ walk further on. The three sources identified at Karmrakar are therefore
situated in the province of Kars, in north-eastern Turkey.
The settlements of Gegharot and Karmrakar, both situated in north-western Armenia and both
belonging to the late phase of the Kuro-Araxes culture, therefore clearly exploited different
sources. The inhabitants of Gegharot exploited the deposits from around the Aragats massif,
while those of Karmrakar were oriented towards the province of Kars. This division reflects
different territories of routes and trade networks. It reveals profound local differences, masked by
the apparent uniformity of the Kuro-Araxes culture.
Late Bronze Age: Gegharot, Hnaberd, Keti and Getashen After the Early Bronze Age, in
the first half of the second millennium, a period followed (the Middle Bronze Age) that is
characterized by the development of transhumant pastoralism (the general abandonment of
64 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

agro-pastoral settlements) and a clear hierarchization of society (the inhumation of a fraction of
the population in kurgans with rich funerary furniture). Then, towards 1600 bc, a rapid transition
occurs towards the Late Bronze Age, seen in the reappearance of numerous permanent settle-
ments in the form of stone-built fortresses of differing size built atop hills (Smith 2005). The
social inequalities visible in the kurgans of the Middle Bronze Age appear to have been formal-
ized into a tightly integrated socio-political apparatus where critical controls over resources—
economic, social and sacred—were concentrated within the Cyclopean stone masonry walls of
powerful new centres; in addition, vast cemeteries appear coincident with the emergence of Late
Bronze Age polities (Smith 2005). Fortresses and cemeteries are present at the sites of Gegharot,
Hnaberd and Keti, in north-western Armenia, and date to between the 15th and the 13th/12th
centuries bc (Badalyan and Avetisyan 2007).
At Gegharot, the analysis of eight obsidian samples from the fortress has determined that six
samples come from the Damlik-Ttvakar deposits in the Tsaghkunyats range (Tsaghkunyats 1),
and the other two from the Aragats flow of the Arteni complex (Arteni 1) (Fig. 11). The territory
of provisionment for this site had thus hardly changed since the Early Bronze Age. The neigh-
bouring sources of Tsaghkunyats are logically in the majority and the existence of a route west
of the Aragats towards Arteni is confirmed.
The fortress of Hnaberd, which is situated on the southern slope of the Tsaghkahovit plain less
than 10 km from Gegharot, shows evidence of a different provisionment. The deposits of the
Figure 10 The ‘least-cost paths’ between the Early Bronze Age sites (Gegharot, Karmrakar and Lusaghbyur) and the
obsidian sources.
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 65
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

Gutansar complex are clearly more predominant (50%) than those of Tsaghkunyats, Arteni and
Geghasar (Fig. 11).
At Keti, further north, Arteni is the main source of obsidian, and the deposits of Sarikamis
North, via the secondary deposits carried by the Kars River as far as Akhurian, as well as those
of neighbouring Yaglica Dag, played a secondary role (Fig. 11).
Thus the territories of supply for the populations of the Late Bronze Age established to the
north of the Aragats massif were appreciably the same as those of their predecessors of the Early
Bronze Age. They provide evidence of the same diversity, which can be linked in this period to
the well-attested breaking-up of the territory into small political entities (Smith 2005).
The tombs of Nor Getashen, on the south-western bank of Lake Sevan, have produced two
obsidian artefacts, one of which comes from the Gegham mountains, very close by, and the other
from Gutansar, situated on the other side of these mountains.
CONCLUSION
Our study confirms that, whatever the period and the geographical location of the settlements, the
prehistoric populations of Armenia supplied themselves with several sources of obsidian. Clearly,
the distance to the source was not the essential parameter in the choice of the deposit. Other
factors were involved in the choice of sources: the role of the secondary deposits of obsidian,
the importance of contacts and exchange between the prehistoric groups, and the existence of
territories of circulation.
Figure 11 The ‘least-cost paths’ between the Late Bronze Age sites (Gegharot, Hnaberd, Keti and Nor Getashen) and
the obsidian sources.
66 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

The role of the secondary deposits in the river valleys has often been underrated, and yet they
enabled the populations to acquire raw material at a considerable distance from the outcrops. This
is the case, in particular, for the sources of Sarikamis North, whose material is available in
abundance in the valley of the Kars River, as well as in that of the Akhurian. This is also the case
for the Tsaghkunyats obsidian blocks, brought down by the Kasakh River to Kmlo, and for the
Sevkar obsidian pebbles carried by the Vorotan River as far south as Godedzor.
In the Upper Palaeolithic and the Mesolithic, the hunter–gatherers had a high degree of
mobility, but exchange and social interaction were also integral components of their lifeways. At
an inter-regional level, exchanges occurred between adjacent regions that were geographically
distinct and therefore capable of producing specific items (including obsidians), which became
objects of exchange for practical reasons or for their social value (Kind 2006). From the Late
Neolithic onwards, domestication is in evidence in Armenia and another form of mobility is
attested as early as the sixth millennium bc: the practice of semi-transhumance (Badalyan et al.
2010). This custom, which reached its peak in the Bronze Age, is an absolute necessity, because
of the high temperatures and drought conditions that descend upon the lowlands in summer.
Consequently, the obsidian from the deposits at high altitude (Syunik and Gegham) was largely
diffused as a result of the transhumant movements (Chataigner and Barge 2008).
In different periods, the primacy given to remote sources of obsidian in a specific direction
suggests the existence of territories of circulation. In the Late Neolithic, the human group living
at Aratashen obtained its obsidian supply mainly in deposits located in the western part of the
Ararat plain and in the Kars region. The Tuzluca salt mountain, located near the confluence of the
Akhurian and the Araxes, could be an element in the explanation of this movement towards
the west. In the Late Bronze Age, striking differences, observed in the provisionment of obsidian
for sites located near each other, such as Gegharot and Hnaberd, enable the perception of
territorial division into small entities and into distinct trade networks.
ACKNOWLEDGEMENTS
The authors express their gratitude to the French Ministry of Foreign and European Affairs and
the Academy of Science ofArmenia, which provided financial backing for their work in Armenia.
They are sincerely grateful to Ruben Badalyan, Pavel Avetisyan and Boris Gasparyan (Institute
of Archaeology, Yerevan) for providing archaeological samples of obsidian.
REFERENCES
Adler, D., Bar-Oz, G., Belfer-Cohan, A., and Bar-Yosef, O., 2006, Ahead of the game: Middle and Upper Palaeolithic
hunting behaviors in the southern Caucasus, Current Anthropology,47(1), 89–118.
Arimura, M., Chataigner, C., and Gasparyan, B., 2010, Kmlo-2, An early Holocene site in Armenia, Neo-Lithics,2(09),
17–19.
Badalyan, R., and Avetisyan, P., 2007, Bronze Age and Early Iron Age archaeological sites inArmenia. I. Mt Aragats and
its surrounding area, BAR International Series 1697,Archaeopress, Oxford/Maison de l’Orient Méditerranéen, Lyon.
Badalyan, R., Chataigner, C., and Kohl, Ph., 2004, Trans-Caucasian obsidian: the exploitation of the sources and their
distribution, in A view from the Highlands: archaeological studies in honour of Charles Burney (ed. A. Sagona),
437–65, Ancient Near Eastern Studies Supplement 12, Peeters, Leuven.
Badalyan, R., Harutyunyan, A., Chataigner, C., Le Mort, F., Brochier, J.-E., Balasescu, A., Radu, V., and Hovsepyan, R.,
2010, The settlement of Aknashen-Khatunarkh, a Neolithic site in the Ararat Plain (Armenia): excavation results
2004–2009, TÜBA-AR, 185–218.
Badalyan, R., Lombard, P., Avetisyan, P., Chataigner, C., Chabot, J., Vila, E., Hovsepyan, R., Willcox, G., and Pessin, H.,
2007, New data on the late prehistory of the southern Caucasus: the excavations atAratashen (Armenia): preliminary
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 67
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

report, in Les cultures du Caucase (VIe–IIIemillénaires avant notre ère)—leurs relations avec le Proche-Orient
(ed. B. Lyonnet), 37–62, Editions Recherche sur les Civilisations, CNRS Editions, Paris.
Blackman, J., Badalian, R., Kikodze, Z., and Kohl, P., 1998, Chemical characterization of Caucasian obsidian geological
sources, in L’obsidienne au Proche et Moyen Orient: du volcan à l’outil (eds. M.-C. Cauvin, A. Gourgaud,
B. Gratuze, N. Arnaud, G. Poupeau, J.-L. Poidevin and C. Chataigner), 205–31, BAR International Series 738,
Archaeopress, Oxford.
Carter, T., and Kilikoglou, V., 2007, From reactor to royalty?Aegean and Anatolian obsidians from Quartier Mu, Malia
(Crete), Journal of Mediterranean Archaeology,20, 115–45.
Chataigner, C., and Barge, O., 2008, Quantitative approach to the diffusion of obsidian in the ancient northern Near
East, in Layers of perception: proceedings of the 35th International Conference on Computer Applications
and Quantitative Methods in Archaeology (CAA), Berlin, Germany, 2–6 April, 2007 (eds. A. Posluschny, K.
Lambers and I. Herzog), 375, Kolloquien zur Vor- und Frühgeschichte, vol. 10, Bonn (11-11_chataigner_et_al-
obsidian.pdf).
Chataigner, C., and Gratuze, B., 2013, New data on the exploitation of obsidian in the southern Caucasus (Armenia,
Georgia) and eastern Turkey, part 1: source characterization, Archaeometry, in press.
Chataigner, C., Avetisyan, P., Palumbi, G., and Uerpmann, H.-P., 2010, Godedzor, a Late Ubaid-related site in the
southern Caucasus, in Beyond the Ubaid—transformation and integration in the late prehistoric societies of the
Middle East (eds. R. Carter and G. Philip), 381–98, SAOC 63, The Oriental Institute of the University of Chicago.
Chataigner, C., Isikli, M., Gratuze, B., and Çil, V., 2013, Obsidian sources in the regions of Erzurum and Kars (north-east
Turkey): new data, Archaeometry, in press.
Eastman, J., 1999, Guide to GIS and image processing, Clark Laboratories, Clark University, Worcester, MA.
Eerkens, J. W., Spurling, A. M., and Gras, M. A., 2008, Measuring prehistoric mobility strategies based on obsidian
geochemical and technological signatures in the Owens Valley, California, Journal of Archaeological Science,35,
668–80.
Forster, N., and Grave, P., 2012, Non-destructive PXRF analysis of museum-curated obsidian from the Near East, Journal
of Archaeological Science,39(3), 728–36.
Frahm, E., 2010, The Bronze-Age obsidian industry at Tell Mozan (ancient Urkesh), Syria: redeveloping electron
microprobe analysis for 21st-century sourcing research and the implications for obsidian use and exchange in
northern Mesopotamia after the Neolithic, Ph.D. dissertation, University of Minnesota.
Gratuze, B., 1999, Obsidian characterisation by laser ablation ICP–MS and its application to prehistoric trade in the
Mediterranean and the Near East: sources and distribution of obsidian within the Aegean and Anatolia, Journal of
Archaeological Science,26, 869–81.
Karajian, H., 1920, Mineral resources of Armenia and Anatolia, Armen Technical Book Co., New York.
Keller, J., Djerbashian, R., Karapetian, S., Pernicka, E., and Nasedkin, V., 1996, Armenian and Caucasian obsidian
occurrences as sources for the Neolithic trade: volcanological setting and chemical characteristics, in Archaeometry
94: the proceedings of the 29th Symposium on Archaeometry, Ankara 9–14 May 1994 (eds. S. Demirci, A. M. Özer
and G. D. Summers), 69–86, Tübitak, Ankara.
Kiguradze, T., and Menadbe, M., 2004, The Neolithic of Georgia, in A view from the highlands: archaeological studies
in honour of Charles Burney (ed. A. Sagona), 345–98, Ancient Near Eastern Studies Supplement 12, Peeters, Leuven.
Kind, C.-J., 2006, Transport of lithic raw material in the Mesolithic of southwest Germany, Journal of Anthropological
Archaeology,25, 213–25.
Le Bourdonnec, F.-X., Nomade, S., Poupeau, G., Guillou, H., Tushabramishvili, N., Moncel, M.-H., Pleurdeau, D.,
Agapishvili, T., Voinchet, P., Mgeladze, A., and Lordkipanidze, D., 2012, Multiple origins of Bondi Cave and Ortvale
Klde (NW Georgia) obsidians and human mobility in Transcaucasia during the Middle and Upper Palaeolithic,
Journal of Archaeological Science,39, 1317–30.
Liagre, J., Arakelyan, D., Gasparyan, B., Nahapetyan, S., and Chataigner, C., 2009, Mobilité des groupes préhistoriques
et approvisionnement en matières premières à la fin du Paléolithique supérieur dans le Petit Caucase: données récentes
sur le site de plain air de Kalavan 1 (nord du lac Sevan, Arménie), in Le concept de territoires dans le Paléolithique
supérieur européen, UISPP, proceedings of the XV World Congress (Lisbon, 4–9 september 2006) (ed. F. Djindjian,
J. Kozlowski and N. Bicho), 75–84, BAR International Series 1938, Archaeopress, Oxford.
Lovis, W., Whallon, R., and Donahue, R., 2006, Social and spatial dimensions of Mesolithic mobility, Journal of
Anthropological Archaeology,25, 271–4.
Meshveliani, T., Bar-Oz, G., Bar-Yosef, O., Belfer-Cohen, A., Boaretto, E., Jakeli, N., Koridze, I., and Matskevich, Z.,
2007, Mesolithic hunters at Kotias Klde, western Georgia: preliminary results, Paléorient,33(2), 47–58.
Mkrtumyan, Yu. I., 1974, Forms of farming in eastern Armenia, Armyanskaya Etnografiya i Fol’klor,6, 7–90 (in
Russian).
68 C. Chataigner and B. Gratuze
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69

Montoya, C., Gasparyan, B., Balasescu, A., Joannin, S., Liagre, J., Nahapetyan, S., Ollivier, V., Colonge, D., and
Chataigner, C., 2013, The Upper Palaeolithic site of Kalavan-1 (Armenia): an Epigravettian settlement in the Lesser
Caucasus. Journal of Human Evolution, in press.
Moriizumi, M., Nakashima, S., Okumura, S., and Yamanoi, Y., 2009, Color-change processes of a Plinian pumice and
experimental constraints of color-change kinetics in air of an obsidian, Bulletin of Volcanology,71, 1–13.
Niknami, K. A., Amirkhiz, A. C., and Glascock, M. D., 2010, Provenance studies of Chalcolithic obsidian artefacts from
near Lake Urmia, northwestern Iran using WDXRF analysis, Archaeometry,52, 19–30.
Nioradze, M., 2001, Le Paléolithique Supérieur de la Géorgie, in Le Paléolithique Supérieur européen: bilan quinquennal
1996–2001, Commission VIII, XIVeCongrès UISPP (Liège, 2–8 septembre 2001) (ed. P. Noiret), 27–33, Liège,
ERAUL 97.
Nioradze, M., and Otte, M., 2000, Paléolithique Supérieur de Géorgie, L’Anthropologie,104, 265–300.
Ouoskherdjan, J., 1828, Mémoire de Jean Ouoskherdjan, prêtre arménien de Wagarchabad, in Mémoires relatifs à l’Asie
contenant des recherches historiques, géographiques et philologiques sur les peuples de l’Orient (ed. H. J. Klaproth),
228–309, Librairie Orientale, Paris.
Poidevin, J.-L., 1998, Les gisements d’obsidienne de Turquie et de Transcaucasie: géologie, géochimie et chronométrie,
in L’obsidienne au Proche et Moyen Orient: du volcan à l’outil (eds. M.-C. Cauvin, A. Gourgaud, B. Gratuze,
N. Arnaud, G. Poupeau, J.-L. Poidevin and C. Chataigner), 105–23, BAR International Series 738, Archaeopress,
Oxford.
Renfrew, C., and Dixon, J. E., 1976, Obsidian in western Asia: a review, in Problems in economic and social archaeology
(eds. G. d. G. Sieveking, I. H. Longworth and K. E. Wilson), 137–50, London, Duckworth.
Scott, P., and Christie, C., 2004, ‘Optimal’ speed-load combinations for military manoeuvres, International Journal of
Industrial Ergonomics,33, 63–8.
Seaman, S., Dyar, M. D., and Marinkovich, N., 2009, The effects of heterogeneity in magma water concentration on the
development of flow banding and spherulites in rhyolitic lava, Journal of Volcanology and Geothermal Research,
183(3–4), 157–69.
Smith, A., 2005, Prometheus unbound: southern Caucasia in prehistory, Journal of World Prehistory,19, 229–79.
Sulgostowska, Z., 2006, Mesolithic mobility and contacts on areas of the Baltic Sea watershed, the Sudety, and
Carpathian Mountains, Journal of Anthropological Archaeology,25, 193–203.
Voigt, M. M., 1983, Hajji Firuz Tepe, Iran: the Neolithic settlement, Monograph 50,The University Museum, University
of Pennsylvania, Philadelphia, PA.
Williams, P. R., Dussubieux, L., and Nash, D., 2012, Provenance of Peruvian Wari obsidian: comparing INAA,
LA–ICP–MS, and portable XRF, in Obsidian and ancient manufactured glasses (eds. I. Liritzis and C. Stevenson),
75–85, University of New Mexico Press, Albuquerque, NM.
Yilmaz, O., 2007, Geology, mineralogy and petrography of the Kagizman (Kars) and Tuzluca (Igdir) salt beds, Subat,
Izmir.
Obsidian in the Caucasus (Armenia, Georgia) and eastern Turkey, part 2 69
© 2013 University of Oxford, Archaeometry 56, 1 (2014) 48–69