Ancient Kura 2010 2011 The First Two Sea

Deutsches Archäologisches Institut !
Eurasien-Abteilung
Außenstelle Teheran
Archäologische Mitteilungen
aus Iran und Turan
Band 44 2012
*
DIETRICH REIMER VERLAG GmbH BERLIN

*
I–IV, 1–352 Seiten mit 342 Abbildungen, 59 Tabellen
Herausgeber: Svend Hansen und Mayke Wagner

Redaktion: Barbara Helwing und Nikolaus Boroffka
Wissenschaftlicher Beirat: Abbas Alizadeh (Chicago)
David Braund (Exeter)
Henri-Paul Francfort (Nanterre)
Ernie Haerinck (Ghent)
Stefan R. Hauser (Halle/Saale)
Lorenz Korn (Bamberg)
Stephan Kroll (München)
Michael Pfrommer (Trier)
Susan Pollock (Berlin)
Robert Rollinger (Innsbruck)
Miroslav Salvini (Roma)
Mitglieder des Deutschen Archäologischen Instituts und Studenten der Altertumswissenschaften können die
Archäologischen Mitteilungen aus Iran und Turan zum Vorzugspreis von 53,20 a zuzüglich Versandkosten
abonnieren. Bestellungen sind an die Schriftleitung zu richten. Studenten werden um Vorlage einer
Studienbescheinigung gebeten. Die Beendigung des Studiums ist unverzüglich mitzuteilen.
Redaktionsschluss ist der 31. März für den im folgenden Jahr erscheinenden Band. Bei der Abfassung der
Manuskripte sind die ,,Richtlinien für Veröffentlichungen der Außenstelle Teheran der Eurasien-Abteilung
des Deutschen Archäologischen Instituts‘‘ zu beachten, die unter http://www.dainst.org/index.php?id¼7490
in ihrer jeweils aktuell gültigen Form aufgerufen werden können.
Die Redaktion behält sich vor, Manuskripte zu kürzen.
! 2012 by Deutsches Archäologisches Institut, Eurasien-Abteilung

ISSN 1434-2758
Redaktion: Deutsches Archäologisches Institut, Eurasien-Abteilung, Im Dol 2–6, D-14195 Berlin
Satz, Druck und Bindung: Beltz Bad Langensalza GmbH, Neustädter Straße 1–4, D-99947 Bad Langensalza
Kommissionsvertrieb: Dietrich Reimer Verlag GmbH, Berliner Straße 53, D-10713 Berlin
Inhalt III
Inhalt
Aufsätze
L y o n n e t, B., G u l i y e v, F., H e l w i n g, B., A l i y e v, T., H a n s e n, S. und M i r t s k h u l a v a, G.,

Ancient Kura 2010–2011: the first two seasons of joint field work in the southern Caucasus.
Mit Beiträgen von L. Astruc, K. Bastert-Lamprichs, W. Bebermeier, F. Becker, N. Benecke, L. Bouquet,
G. Bruley-Chabot, A. Courcier, M. B. D’Anna, A. Decaix, J. Fassbinder, M. Fontugne, F. Geitel, A. Goren,
C. Hamon, J. Koch, G. Le Dosseur, A. Lincot, R. Link, R. Neef, D. Neumann, V. Ollivier, P. Raymond,
A. Ricci, A. Samzun, S. Schorr, F. Schlütz, L. Shillito, M. Ullrich und J. Wahl . . . . . . . . . . . . . 1
R o u s t a i e, K., Archaeological survey of the Shahroud area, northeast Iran: a landscape approach . 191
L a u, D., Schmuckperlen aus Sohr Damb/Nal, Pakistan. Mit einer Einführung von U. Franke . . . . . 221
T e u f e r, M., Der Streitwagen: Eine ,,indo-iranische‘‘ Erfindung? Zum Problem der Verbindung von
Sprachwissenschaft und Archäologie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
X i n, W. und L e c o m t e, O., Clay sealings from the Iron Age citadel at Ulug Depe . . . . . . . . . . 313
Ç a v u ş o ǧ l u, R. und B i b e r, H., A mysterious archaeological site in the vicinity of the Urartian
capital Tuşpa: Van/Kalecik . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Buchbesprechungen
S. Pollock, R. Bernbeck und K. Abdi (Hrsg.), The 2003 Excavations at Tol-e Bas̆i, Iran: Social Life in
a Neolithic Village. Archäologie in Iran und Turan 10 (Verlag Philipp von Zabern, Mainz 2010)
(L l . W e e k s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Bruno Jacobs/Robert Rollinger (Hrsg.), Der Achämenidenhof/The Achaemenid Court. Akten des 2. In-
ternationalen Kolloquiums zum Thema ,,Vorderasien im Spannungsfeld klassischer und altorienta-
lischer Überlieferung‘‘. Classica et Orientalia 2 (Harrassowitz Verlag, Wiesbaden 2010) (W. Messer-
schmidt). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
IV Inhalt
Table of content
Articles
L y o n n e t, B., G u l i y e v, F., H e l w i n g, B., A l i y e v, T., H a n s e n, S. and M i r t s k h u l a v a, G.,

Ancient Kura 2010–2011: the first two seasons of joint field work in the southern Caucasus.
Mit Beiträgen von L. Astruc, K. Bastert-Lamprichs, W. Bebermeier, F. Becker, N. Benecke, L. Bouquet,
C. Hamon, J. Koch, G. Le Dosseur, A. Lincot, R. Link, R. Neef, D. Neumann, V. Ollivier, P. Raymond,
A. Ricci, A. Samzun, S. Schorr, F. Schlütz, L. Shillito, M. Ullrich und J. Wahl . . . . . . . . . . . . . 1
R o u s t a i e, K., Archaeological survey of the Shahroud area, northeast Iran: a landscape approach . 191
L a u, D., Schmuckperlen aus Sohr Damb/Nal, Pakistan. Mit einer Einführung von U. Franke . . . . . 221
T e u f e r, M., Der Streitwagen: Eine ,,indo-iranische‘‘ Erfindung? Zum Problem der Verbindung von
Sprachwissenschaft und Archäologie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
X i n, W. and L e c o m t e, O., Clay sealings from the Iron Age citadel at Ulug Depe . . . . . . . . . . 313
Ç a v u ş o ǧ l u, R. und B i b e r, H., A mysterious archaeological site in the vicinity of the Urartian
capital Tuşpa: Van/Kalecik . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Book reviews
S. Pollock, R. Bernbeck and K. Abdi (Hrsg.), The 2003 Excavations at Tol-e Bas̆i, Iran: Social Life in
a Neolithic Village. Archäologie in Iran und Turan 10 (Verlag Philipp von Zabern, Mainz 2010)
(L l . W e e k s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Bruno Jacobs/Robert Rollinger (Hrsg.), Der Achämenidenhof/The Achaemenid Court. Akten des 2. In-
ternationalen Kolloquiums zum Thema ,,Vorderasien im Spannungsfeld klassischer und altorienta-
lischer Überlieferung‘‘. Classica et Orientalia 2 (Harrassowitz Verlag, Wiesbaden 2010) (W. Messer-
schmidt) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
Deutsches Archäologisches Institut
EURASIEN-ABTEILUNG
AUSSENSTELLE TEHERAN
A Ancient Kura 2010-2011:

M The first two seasons
I
T
ARCHÄOLOGISCHE MITTEILUNGEN AUS IRAN UND TURAN

Sonderdruck aus Band 44, 2012
Ancient Kura 2010–2011: The first two seasons
of joint field work in the Southern Caucasus
Bertille Lyonnet and Farhad Guliyev, Barbara Helwing and Tevekkül Aliyev, Svend Hansen and
Guram Mirtskhulava
with contributions by L. Astruc, K. Bastert-Lamprichs, W. Bebermeier, F. Becker, N. Benecke, L. Bouquet,

C. Hamon, J. Koch, G. Le Dosseur, A. Lincot, R. Link, R. Neef, D. Neumann, V. Ollivier, P. Raymond, A. Ricci,
A. Samzun, S. Schorr, F. Schlütz, L. Shillito, M. Ullrich and J. Wahl
Schlagwörter: Südkaukasus, Neolithisierung, Landschaftsarchäologie, Chalkolithikum, Bioarchäologie

Keywords: South Caucasus, Neolithization, landscape archaeology, Chalcolithic, bioarchaeology
˚º7\*-ß* æº@-/: 2IBßØ ˚/-Œ/H, B*@ºŁ%ŁH/!Ł', º/B+ł/#%B/' /(ı*@º@ªŁ', 9B*@ºŁ%, ÆŁ@/(ı*@º@ªŁ'
Fig. 1
South Caucasia. Map
showing the working
area of Project ‘‘Ancient
Kura’’ and indicating
the three major
research foci at Aruchlo,
Mentesh Tepe and
Kamiltepe (Map on the
basis of satellite image
CaspianSea.A2004274.
0955.500m.jpg,
PD NASA)
Introduction The special geographical situation of the South-

ern Caucasus lands as both a barrier and a land
Bertille Lyonnet and Barbara Helwing bridge exposed the region to innumerable influences
from north and south, from the Mesopotamian civi-
South Caucasia (Fig. 1), bordered by the Black Sea lizations as from the Eurasian steppe populations.
to the west, the Caspian Sea to the east, and the This consideration of South Caucasia only as
southern slopes of the Greater Caucasus Mountains a land bridge caused researchers to focus their at-
to the north, marks the geographical transition from tention on the assumed centres of cultural devel-
Europe to Asia. This area, however, is at the same opment, Greater Mesopotamia, on the one side,
time also separated from eastern Anatolia by the and the northern Caucasus, with the famous Majkop
Lesser Caucasus Mountains. Two major rivers, the culture, on the other. This perspective eventually
Kura and the Araxes, flow through it, forming broad concealed the originality of cultural developments
flood plains from west to east before they join and in South Caucasia proper. Such a phenomenon of
flow into the Caspian Sea. biased academic perception has equally been ob-
2 Bertille Lyonnet et al.
served in discussions of the early cultures of the There is no doubt that the southern Caucasus
Anatolian highlands.1 lands were home to a number of prehistoric cultures
After becoming included in the former Soviet throughout the millennia. A few sites like Chokh in
Union, the region acquired an academic sideline mountainous Daghestan12 or Kmlo 2 in Armenia13
position in the perception of western scholars that have been discovered, but have not yet enabled
is by no means justified.2 Although Soviet scholars any detection of local developments towards the
greatly contributed to research and even made pio- domestication of plants and animals at the begin-
neering developments in scientific methods, such as ning of the Holocene. However, discoveries of Halaf
use wear analysis,3 palaeobotany4 and palaeometal- sherds in the Araxes Valley and the close similari-
lurgy,5 the integration of their results into an overall ties in the material culture of the Leilatepe/Boyuk
culture-historical framework remains difficult or even Kesik sites with that of northern Mesopotamia dur-
impossible. ing the first half of the 4th mill. BCE14 suggested
Since the collapse of the Soviet Union, the that the Fertile Crescent and Northern Mesopotamia
continuity of local administrative structures was played a major role as a trigger contact in cultural
partly interrupted and has only recently been reor- development. Substantial research and excavations
ganized. During this latent period, archaeological by colleagues from Azerbaijan, Georgia, Armenia
methods underwent considerable developments in and Russia in the 1960–1980s have provided start-
other regions. Reliable calibrated radiocarbon dat- ling evidence for the existence of a number of ori-
ing and an array of computer applications in ar- ginal prehistoric cultures over the millennia. Thus,
chaeological research have become standard proce- different groups have been detected, among which
dures. New fields, like landscape archaeology and are the so-called S̆omu-S̆ulaveri,15 Kültepe16 and
micromorphology, emerged as further specializa- Alikemek,17 Sioni and Tsopi,18 and Leilatepe19 cul-
tions that allow more insight into the archaeological tures, and a material culture comparable to that of
record and the human history behind its formation. the Kura-Araxes groups has been reported at a large
Also, new interpretative frameworks, beyond culture number of newly discovered sites. Nevertheless, the
history and cultural essentialism, oriented towards scarcity of calibrated 14C dates, the lack of precision
microhistories and practice approaches have been in the recording of data, and the small number of
formulated. Rarely have these been applied in South publications sometimes written in the Georgian or
Caucasian research. They have only partly been used Azerbaijan language, or out of print, make it difficult
during the recent salvage excavations along the to utilize all of this evidence. Furthermore, these
Baku-Tbilisi-Ceyhan pipeline,6 which led to the dis-
covery of interesting new sites contemporary to the
Leilatepe Chalcolithic occupation.7 A few western
12
archaeologists have started new excavations em- Amirkhanov 1987.
13
ploying these new methods on sites of the Neolithic Arimura et al. 2012.
14
On the relations between northern Mesopotamia and the Cau-
and Chalcolithic periods in Armenia,8 in Georgia,9 in casus from the 6th to 3rd mill. BCE, see Lyonnet 2007 a.
Azerbaijan10 and in Nakhichevan.11 These new inves- 15
S̆omu-Tepe was the first site to be excavated, and many other
tigations all emphasised the real need for further related sites were discovered between the Georgian border
research on the remarkable and original role played and the Zeyemchaj by I. Narimanov (1987). Located on the
other side of the border are S̆ulaveris-Gora and other sites
by the southern Caucasus lands in the development close by, among which Aruchlo was excavated not long after-
of early cultures. wards, especially by T. Chubinishvili. A chronological frame of
the different cultures mentioned here as well as an internal
chronology of the S̆omu-S̆ulaveri culture had been attempted
by T. Kiguradze (1976) mainly on the basis of the pottery finds.
For a synthesis about these previous excavations on the S̆omu-
1 S̆ulaveri culture see Munchaev 1982; Narimanov 1987; Kush-
On Anatolia, in general, cp. Greaves 2007; for a specific focus on nareva 1993, 29–51; Chataigner 1995; Lyonnet/Guliyev 2010.
the Neolithic period, cp. Özdoǧan 2007. Sites discovered during surveys in the Mil Steppe farther East
2 Lyonnet/Guliyev 2010, 220–221. (¨*ææ*B 1965) were at that time considered to be culturally
3
6*C‚B@-/˚@(@ÆŒ@-/ 1983. related to the S̆omu-S̆ulaveri culture.
4 16
Vavilov/Dorofeev 1992. Abibullaev 1982.
5 17
Chernykh 1992; Selimkhanov 1978. The site, located in the Muğan Steppe east of the Mil Steppe,
6
Maynard 2011. was excavated by F. Makhmudov but has not been published
7 except for short articles. It is usually considered as related to
Akhundov 2007; Makharadze 2007; Museibli 2007; Lyonnet et al. the Kültepe culture in Nakhichevan.
2008. 18
8 The Sioni culture, named after the eponymous site (Menabde/
Badalyan et al. 2007; Badalyan et al. 2010; Arimura et al. 2010; Kiguradze 1981), is mainly known in Georgia (Varazashvili 1992;
Chataigner et al. 2010. Chikovani 1998; Chikovani 1999; Kiguradze 2000; Kiguradze/
9
Hansen et al. 2006; Hansen et al. 2007. Sagona 2003; Chelidze/Chikovani 2008; Chikovani et al. 2010;
10
Lyonnet 2009 b; Lyonnet et al. 2008; Lyonnet et al. 2010; Lyon- Nebieridze 2010), although sites with similar material culture
net et al. 2012; Guliyev/Nishiaki 2012. have also been discovered in Azerbaijan.
11 19
Marro et al. 2009; Marro et al. 2011. Narimanov 1987; Aliev/Narimanov 2001.
Ancient Kura 2010–2011: The first two seasons of joint field work in the Southern Caucasus 3
data do not suffice for the kind of questions now leagues from Azerbaijanian and Georgian research
raised in new archaeological research. institutions, with the intent to investigate some of
The Southern Caucasus, with its rich and ver- the questions formulated above. It is funded by ANR
satile fauna and flora and an abundance of mineral (Agence Nationale de la Recherche) in France and
sources, including metal ores, provided a favorable DFG (Deutsche Forschungsgemeinschaft) in Germany.
basis for the establishment of early human cultures. The major aim was an attempt to investigate the
It was home to wandering foragers since the Palae- interaction and interdependency between humans
olithic period, as shown by the discoveries of the and the landscape from early sedentism to the be-
Azikh Cave or attested by the famous Gobustan ginning of the Bronze Age, that is, from the 6th to
rock art dated to the beginning of the Holocene. the 3rd mill. BCE. The rationale behind the group’s
This highly attractive region was, from this point on, approach was to understand how some of the great
occupied by communities who took up farming and transformations visible in the settlement history of
herding towards the end of the 7th mill. BCE at the the Southern Caucasus took place – Neolithisation,
latest. While we now know that this development the development of short-term settlements from the
did indeed occur, the dynamics behind it still re- 5th mill. BCE onward, and the appearance of osten-
main poorly understood. Is the Southern Caucasus tatious burial sites in the 4th mill. BCE – as well as
part of the primary Neolithization zone of Western the role of environmental factors in these develop-
Asia, or was this process a secondary development ments. It was generally assumed that the favorable
there? Was it related to the movement of humans, conditions in the early Holocene invited human com-
or rather the transmission of a new Neolithic life- munities to settle and develop a subsistence econ-
style to this region? While intensive research in omy in which they produced their own food in the
other areas of Western Asia has led to a re-thinking 6th mill. BCE. The abandonment of Neolithic villages
of old paradigms, the Southern Caucasus still re- and the shift to archaeologically less visible lifeways
mains beyond the realms of this discussion. Subse- afterwards were equally attributed to environmental
quent to this first period of settlement during the constrains. But scientific evidence for these hypoth-
6th mill. BCE, several transformations followed that eses was missing; tentative interpretations that as-
affected both the lifestyle and settlement patterns. sumed a turn to nomadism or an overall population
Sites of the 5th and 4th mill. BCE are barely visible withdrawal from the region were arguments based
in the archaeological record, and the reasons for merely on a lack of evidence.
this remain unknown. Does this absence of settle- The research strategy adopted is a combina-
ment sites reflect ancient reality? Did ancient com- tion of intensive research into selected locales with
munities decide to move on, settle in different loca- large-scale comparative reconstructions of the an-
tions, or did they become more mobile altogether, cient environment and the way it was used by the
and whatever was the case, why? At the beginning human population. The localities where the project
of the 4th mill. BCE new types of monuments such has focused its attention provide a representative
as visible burial mounds or kurgans appear. Some cross-section of both archaeological periods and
of the burials contain metal weapons and symbolic landscapes along the middle course of the River Kura
objects of leadership together with exotic items,20 (‘Kür’ in the Azerbaijan language, and ‘Mstkhari’ in
the circulation of which is noticeable in a wider area Georgian) (Fig. 1): Aruchlo in eastern Georgia, and
beginning at the end of the 5th mill. BCE. This indi- Mentesh Tepe in the Tovuz region of western Azer-
cates that important social transformations took baijan, and the Mil Plain in southern Azerbaijan.
place, with new elite groups emerging. The exis- The climate today ranges from temperate in Aruchlo
tence and the abundance of metal ores, including to semi-arid in the Mil Plain, and comparable dif-
gold, in the Southern Caucasus predestines the references must have existed in the past, as both
gion to play a special role in this crucial transforma- the faunal and the botanical records suggest. The
tion, whose dynamics remain to be understood. archaeological foci in the three regions also com-
plement each other: the oldest phases are now re-
presented in all three regions: Aruchlo is a Neo-
The Ancient Kura Project lithic site that was inhabited from about 5800 to
5400 BCE; Mentesh Tepe has also Neolithic layers
Bertille Lyonnet and Barbara Helwing and provides important insight into sedentary life
in the 5th mill. BCE and more mobile ways of life in
The project ‘‘Ancient Kura’’ was founded in 2010 as the Early Bronze Age; the Mil Plain was occupied
a joint research group, in collaboration with col- by Neolithic communities from ca. 6000–5400 BCE,
and also has evidence for occupation in the 4th mill.
BCE, although settlement behaviour there differs ra-
20
Lyonnet et al. 2008. dically from the other two regions.
The project proceeded by a dual strategy of vey and an extensive sampling program aimed at
intensive archaeological and interdisciplinary inves- gaining data on ancient climate and landscape has
tigations in each of the three focal regions on the been undertaken.22 As a result of the intensive sur-
one hand and comparative studies with regard to vey, a total of 46 sites could be identified during
environmental reconstruction and use of natural re- the two working seasons, out of which 18 yielded
sources on the other. Each individual study consists Neolithic material.23 More detailed work on some of
of both in-depth local investigations in combination the sites discovered anew during the intensive sur-
with work in a comparative perspective. With the vey involved magnetometry survey on sites MPS24
following report, we intend to provide a first and 2, 3, 4, 5, 16, 18, 19.25 Test excavations were carried
naturally preliminary overview on the work currently out on sites MPS 4 and 5 in 2010 and extended in
undertaken during the first two seasons and to pre- 2011, and on site MPS 16 in 2011. Furthermore, sys-
sent the first results of this on-going work. We shall tematic cleaning of pre-existing artificial openings
proceed here by first introducing the research that on sites MPS 18 and 36 was used to collect data
has been carried out within the individual research for period identification and dating.
sites – the Mil Plain and Aruchlo providing data As a first and preliminary result, the Mil Plain
concerning the Neolithic, and Mentesh Tepe which can now be described as an exceptionally well pre-
was occupied much longer providing a sequence served Neolithic micro-landscape extending along
until the Early Bronze Age. Following this, the re- the Kura tributaries of the Lower Qarabagh region.
sults obtained from comparative analyses of envi- Sedentary occupation is attested as a fully devel-
ronmental and climatological research, both from oped Neolithic complex around 6000 BCE, with sub-
bioarchaeological remains and material culture, will sistence based on agriculture and herding. Affilia-
be presented. tions with the larger world of the Southern Caspian
and with northern Iran can be stated on the basis
of material culture comparisons, while material cul-
Part I: ture relations with contemporary sites a mere 200 km
northwest along the Middle Kura seem to have
Field work in the Mil Plain: the 2010–2011 been surprisingly weak. Later periods are much less
Expedition well represented, but possible reasons for this pat-
tern are not yet identified.
Barbara Helwing and Tevekkül Aliyev
Excavations at Kamiltepe
The Mil Plain is located in a triangle circumscribed
by the rivers Kura in the northeast and Araxes in The 2009 excavations in Kamiltepe yielded the re-
the south, and by the slowly ascending mountains mains of a massive mudbrick platform dating to the
of the High Qarabagh to the west. Several small 6th mill. BCE.26 The objectives of the 2010 and 2011
streams descend from the mountains and cross the excavation seasons were, therefore, to clarify the
Mil Plain in west-east direction before they join the nature and extent of this structure and of architec-
Kura River. One of these Kura tributaries flows ture possibly surrounding it. Furthermore, it remaind
through the Qaraçay Valley, where the Mil Steppe to be investigated whether or not a later building
research project is centered. phase had existed there.
As part of the Ancient Kura Research Project, Work in Kamiltepe began in 2010 with magne-
the 2010 to 2011 field work seasons of the Azerbai- tometry investigation27 to test whether the mudbrick
janian-German Expedition to the Mil Plain focussed construction was a massive structure or whether it
upon the Neolithic site of Kamiltepe in the Qarasu contained rooms, compartments, or other hollows.
Valley. The site has been under excavation since Unfortunately, the heavy disturbance of the surface
2009 and has yielded rich data on its mid 6th mill. of the mound prohibited conclusive analysis, and
BCE occupation.21 The central part of the badly da- thus we decided to pursue this question by fully
maged site consists of a large massive mudbrick uncovering the surface of the mound (Fig. 2). Ide-
platform surrounded by domestic dwellings, a fea-
ture so far unknown in the Late Neolithic of West- 22
Preliminary results of these investigations are being presented
ern Asia. In addition to more detailed excavation below by the individual research groups.
and magnetometry survey on the Kamiltepe mound, 23 Ricci, in this article.
an integrated program of intensive geomorphologi- 24 MPS stands for ‘‘Mil Plain Survey’’ and pertains to the indivi-
cal, geological, geophysical and archaeological sur- dual identified sites discovered during the Mil Plain Survey. In
this system, MPS 1 stands for the site of Kamiltepe.
25
Fassbinder et al., in this article.
26
Aliyev/Helwing 2009, 32 Fig. 11; Helwing et al. 2012.
21 27
Aliyev/Helwing 2009. Fassbinder et al., in this article.
Fig. 2
Kamiltepe. Topographic
map (data mapping
M. Ullrich, U. Boettcher,
A. Kwast; photo
M. Tonch; map and
montage A. Kwast)
Fig. 3
Kamiltepe. View from
above, photograph
taken from a crane at a
height of ca. 15 m
(photo H. Taylor)
Fig. 4
Kamiltepe op. 2. The
outer facade of the
mudbrick platform
(photo B. Helwing)
ally, by scraping the surface on a large scale,28 an

outline of the monument would become visible.
The southern half of square P12 and the northern
half of square P11 were, therefore, scraped deeply
as one area (continuing as op. 0129 from 2009)
without intermittent baulks. Unfortunately, the sur-
face proved to be more disturbed by both modern
activities and by Iron Age graves (op. 01, SU 125.3;
125.4; 126.2) than was originally estimated, so the
surface had to be exposed by means of excavation.
It was possible to trace the uppermost preserved
layer of mudbricks in most of the area and, thus, to
reach to the respective ends of the structure but at
unequal elevations and with a very uneven surface.
This operation confirmed the massive construction
of the building with no rooms or compartments
(Figs. 3; 4). Its total size can now be calculated at a
diameter of 24 m. One major Iron Age pit was dug
into the centre of the structure (SU 123; 124). This
pit appears to be bell-shaped in the bulldozer sec-
tion’s eastern profile, and reaches deep into natural
soil below the lowermost layer of mudbricks. The
pit fill included limestone fragments, Iron Age pot- tached to this in the northwest at right angles, but Fig. 5
Kamiltepe op. 13.
tery and some fragmentary bones. We, therefore, do not interlock with the platform. Unfortunately, Upper occupation
assume that this was originally a pit related to a the southwest trench was so heavily disturbed by phase, small room
large burial; several smaller Iron Age graves are Iron Age and Islamic period graves that no architec- with storage jar in situ
preserved on top of the mudbrick structure, close ture related to the platform survived.
to the big pit. Possibly, at the time when the Iron In O12 southwest and northwest (op. 13),
Age funerals were held, the high elevation of Kamil- only the uppermost layer of mudbrick architecture
tepe was attractive to people, who often buried was uncovered immediately under the modern sur-
their dead under visible kurgans and also erected face. Although located outside of the estimated ideal
their own kurgans in the immediate vicinity of the outline of the platform facade (based on the as-
site.30 In this context, the Kamiltepe mound may sumption that it originally was round), this architec-
have been perceived as a ‘‘ready-made’’ kurgan.31 ture seems to represent the uppermost preserved,
Two more 4 " 4 m trenches were opened platform-related architecture: it consists of a series
along the assumed northern outer edge of the plat- of cubicle-like rooms, in which cooking pots and
form in order to confirm its outline; one was the large storage jars were standing in situ (Fig. 5).
southwest quarter of square O11 (included in op. 02); Excavations in O12 northeast (op. 06) began
the second was the northeast quarter of square in 2010 and yielded an upper building phase, with
O12 (op. 06), where we expected to find residues of walls constructed from single rows of mudbrick set
an upper building phase. Two more 4 " 4 trenches perpendicular to the course of the wall (Fig. 6);
were then added in 2011 in O12 northwest and these walls define four rooms that are partly ex-
O12 southwest (together op. 13) to trace the contour posed in the excavated area. The southeastern
of the mudbrick structure where it is preserved. room SU 609 yielded a rich burnt deposit, while the
Furthermore, P12 was extended towards the east other rooms towards the north were almost empty.
and towards the north into P13 (op. 10), until the After removing the walls of the upper building
slope was completely included in the excavated area. phase, a wall that was constructed from two rows
Op. 02 in O11 southwest and northwest un- of mudbricks (SU 625) became visible. A clay floor
covered the edge of the platform with a fully pre- connected to this wall, and a thick layer of burnt
served and undisturbed facade; two walls are at- debris with rich finds lay on that floor. West of wall
SU 625 is another curved mudbrick wall. This wall
is largely within in the baulk, as is the join between
28 A technique applied with good results e.g. at Çatalhöyük, cp. the two walls, so we are not yet able to define their
29
R. Matthews 1996; Tell Brak, cp. R. Matthews 2003. relation. Excavation did not proceed beyond the
Here the work in an excavated area is designated ‘operation’ level of the floor.
or ‘op’.
30
¨*ææ*B 1965. The extension of the excavation area towards
31
Helwing et al. 2012. the east and northeast (op. 10) yielded remains of
In summary, Kamiltepe appears as a more

complex site than previously thought. Most impor-
tantly, its occupation now seems to extend over
two phases. The large massive mudbrick construc-
tion formed its centre during the older phase. The
second occupation phase is so far documented only
to a limited extend in ops. 6 and 13 and consists
of domestic architecture with evidence for storage
and food preparation.
The function of the mudbrick construction of
the earlier phase still remains enigmatic. Light-
weight construction debris like daub and branches
found along its northern edge may point to the ex-
istence of standing architecture atop the platform,
an assumption corroborated by residues of architec-
tural materials observed in the micro-morphological
samples cut from related ash layers in op. 01.32 A
large number of cooking vessel fragments con-
tained within the same levels points to a function
as a place of food preparation, as had been sug-
gested before.33 These two observations are not
Fig. 6 burnt platform debris over a larger area (Fig. 7). Un- mutually exclusive, and the working hypothesis is
Kamiltepe op. 06.
Architecture of
derneath the light gray ashes of the uppermost pre- that there was a massive substruction underneath
upper phase (photo served layer was a layer of debris that contained huts or houses made from largely organic material
B. Helwing) residues of burnt clay (daub). Imprints of organic or wattle-and-daub.34
matter like straw and branches indicate that this in-
cluded vegetal building material and may have Excavation at MPS 4
been the residues of a roof. Under this level is a
layer of burnt mudbrick debris, associated with MPS 4 is a flat and almost invisible site located
abundant ceramic and animal bone fragments. It is 700 metres to the southwest of Kamiltepe on the
Fig. 7 this layer that overlies the uppermost mudbricks northern side of the Qarasu Valley where the terrain
Kamiltepe op. 10. toppled down from the platform. Towards the north, gently slopes towards the river course. The site is
Architectural debris a line of mudbricks was observed. Whether or not distinguished by several small elevations that are
within the ash layers
descending from the these are the remains of a wall of an exterior build- barely visible on the ground, and was recognized
platform ing must still be investigated. through surface finds of Neolithic sherds collected
from fox holes before the beginning of the syste-
matic survey. The ground has not been touched by
mechanized agriculture over the last decades, and
at the time of discovery in 2010, the surface of
the site was fully intact and covered by artemisia
steppe vegetation.
Magnetometry survey35 on seven squares of
40 " 40 metres showed several strong linear ano-
malies, some of which were crosscutting each other.
These were investigated by augering and, as a re-
sult, were identified as gravelly sediments gener-
ated by slowly flowing water. These remains of pa-
laeochannels are probably of anthropogenic origin.
32
Shillito, in this article.
33
Aliyev/Helwing 2009, 32–33.
34
Evidence for standing architecture of the 6th mill. BCE is scanty
throughout Western Asia. House images are documented on
painted pottery of the Halaf period in southeastern Turkey and
may serve as inspiration, cp. Campbell et al. 2003; Tekin 2011.
35
historical age, but it covered two structures of Neo-

lithic date, which were subsequently investigated on
a larger scale: one semi-subterranean round build-
ing that contained workshop residues from shell
bead manufacture, and a system of four concentric
ditches.
Semi-subterranean round building

with residues of a shell bead workshop
in squares D–E 29
A round structure (SU 519; 520) was uncovered

during the two seasons in 2010 (op. 05) and 2011
(op. 08; Figs. 8; 9). This structure measures 2.8 m
in diameter and is dug into natural ground, to a
depth of 2.5 m below the modern surface. The walls
of the pit widen slightly towards the bottom; in the
eastern part they are reinforced by a wall constructed
from a single row of hand-formed cushion-shaped
mudbricks. This wall was preserved to a height of
five rows of mudbrick (Fig. 10). No entrance to this
construction was observed, but since an animal
hole disturbed the northeastern side of the struc-
ture, it cannot be fully excluded that an entry once
existed. The floor consists of beaten earth and was
Fig. 8 renewed four times. Underneath the lowermost floor
Site MPS 4. Semi-subterranean round building
were two small oval pits where shell chipping de-
bris was found. Spots of ochre color were observed
In 2010, a 3 " 3 m sounding was dug in the in parts of the floor.
northwest corner of square E29 (op. 05) to obtain The fill of the round building, as well as the
dating evidence from one of the only faintly visible, soil layers in between the floor layers, contained
linear anomalies. This channel turned out to be of large amounts of debris from shell bead manufac-
Fig. 9
Site MPS 4. Plan of the
semi-subterranean
round building
edres and cones, occasionally also as spheres. One

singular piece is a small cone with incised ‘‘feet’’.
The clay objects are made of very fine clay and are
lightly fired, so that they are dark brown to blackish
in color on the surface. They resemble artifacts
know from other Neolithic sites of Western Asia
that are often referred to as ‘‘tokens’’. Such clay to-
kens are attested, for example, at Neolithic Hajji Fi-
ruz in northwestern Iran, in the shape of tetraeders,
spheres and cones, but also other forms.37
C o n c e n t r i c d i t c h s y s t e m – ‘‘h e n g e ’’
One meter east of the round building, a rectangular
structure extending north-south was observed
throughout the full length of the trench; only when
the lowermost fill of this structure was reached did
it become visible in the section that these were the
remains of a V-shaped ditch that reached 2.7 m be-
low the modern surface. This ditch could then be
matched with an anomaly on the magnetometry im-
age that was part of a multiple-line quarter-circular
Fig. 10 ture. The process involved transforming shells pre- structure in the southeastern part of the site in
Site MPS 4.
Semi-subterranean
liminarily identified as sea shells36 into round disc- squares D/E/F/G 28–29 (ops. 09; 15; 19; Fig. 14). In
round building, the wall shaped beads. The assemblage (Fig. 11) includes 2011, trenches were dug to test the existence of
south of the building more than 3700 fragments of shells and shell further ditches. The location of these trenches was
beads in all stages of work, numerous small chips partly determined through the existence of extensive
of quartzite and flint, and one round flint module recent anthropogenic disturbances (op. 09). Subse-
that probably served as a hammer stone. Traces of quently, two larger areas were opened in D28–29
ochre were observed on the shells. and E/F28–29 (ops. 15; 19).
Besides shell bead residues, a bone pendant These excavations confirmed the existence of
was found in the upper fill of the round building a system of four concentric ditches. On the basis of
(Fig. 12). A series of small clay objects of roughly the magnetometry map, one quarter of this ditch
geometric shape was collected from the building’s system seems to be fully preserved. If the ditches
fill (Fig. 13). These appear in the shape of tetra- were originally circular in layout, a diameter of ca.
Fig. 11
Site MPS 4. Remains of
the bead workshop
Fig. 12 37 As tetraeders, spheres and cones but also other shapes, cp.
Site MPS 4. 36
Bone pendant Identification by N. Benecke. Voigt 1983, 185–186 with lists of comparisons.
50 m can be calculated for the outer ditch. The

magnetometry map shows partly covered anomalies
in squares H29 and E30–31 that could also relate
to the ditch system, if it was oval. If this proves to
be the case, then the oval of the other ditch may
have extended more than 70 m along the long axis.
Description of the ditches – from the

outside to the interior of the circle Fig. 13
Site MPS 4. Clay ‘‘token’’
Ditch 1 – outer ditch: Ditch 1 (SU 2002) is the out-
ermost of the ditches and has been documented
over a larger area in square D28 in outline extend-
ing here in a north-south-direction. No sections have
yet been excavated. The ditch seems to end shortly
before the northern end of the excavated area. It is
in its uppermost part ca. 1.6 m wide. Its fill consists
of dark brown soil, so far without mudbrick frag-
ments.
Ditch 2: Ditch 2 is the V-shaped ditch docu-
mented first in the 2010 excavations (SU 516; 903/
905; Fig. 15). It was now accessible in the anthro-
pogenic trench south of the excavated area, where
a second profile was cut. It has two clearly distin-
guishable contours: in its lower part, it is clearly
V-shaped and has steep sides; farther up, the ditch
widens suddenly towards the east and forms a
rather wide structure. In its upper part, it is 2.2 m
wide and narrows towards the base to 40 cm width.
The base of the ditch was reached 2.7 m below the
modern surface. A single course of mudbricks run-
ning in a west-east direction was set into the ditch
in its northern excavated part.
Fig. 14
When a full section through the ditch was cut Site MPS 4. Magnetometry map with indication of excavated areas and position of ditches
in op. 09, mudbricks could be observed in several (magnetometry by J. Fassbinder et al.)
locations. In the section, it is visible on excavated
steps at the sides of the ditch that these mudbricks
had been set in single rows along the inner and
outer side of the ditch. The function of these mud-
brick rows remains unknown; they might have served
to re-enforce the edges of the ditch, or they might
have been part of a larger construction inside the
ditch.
The fill of ditch 2 allows four separate layers
to be distinguished: below a level of topsoil follows
one layer of very loose and crumbly soil that was
almost undistinguishable from the topsoil and that
may be a post-ditch-phase sediment; the third layer
contains loose fine soil and lime aggregations, and
also had some pottery and mudbrick fragments.
This layer may represent a room fill, since the mud-
brick walls set into and across the ditch indicate
the possibility that the ditch compartments were at
some point used as rooms. However, no walk hori-
zon or even floor was observed. At the base the fill
Fig. 15
consists of mudbrick debris. Site MPS 4. Profile of the second ditch
side, and slopes more gently towards the inside of

the ditch system. It is less deep than ditch 2 and
has a rounded, U-shaped base at about 1.8 m
depth. The fill as observed in the section consists
of a dark brown, clayey soil mixed with mudbrick
debris. Two perpendicular mudbrick walls intersect
the ditch. They were constructed from hand-shaped
mudbricks of varying sizes, about 20 cm in section,
and stand inside the ditch. The setting of the bricks
follows the base of the ditch. The facades were
fairly irregular, but the walls must have stood to a
considerable height, since the better preserved wall
in the south (SU 909), which is constructed of two
parallel mudbrick courses, is preserved to the top
of the ditch. Taken together, the walls divide the
ditch into smaller compartments and form a radial
pattern.
Ditch 4: Ditch 4 (SU 1503/1903) is the inner-
most of the four ditches. It is documented in a
small sounding of 4 " 1.2 m along the northwestern
Fig. 16 Ditch 3: The third ditch (SU 903/905/917/1904; edge of square F29, and in the northern section.
Site MPS 4. Profile of
the third ditch
Figs. 16; 17) was uncovered and excavated in the The southeastern end of the sounding is defined by
southeastern corner of square E29, and was traced a perpendicular mudbrick wall. The further exten-
but not further excavated in F29. In E29, it is docu- sion of the ditch was traced in square F29, but was
mented in one section; the remaining part of the not further excavated. In some spots, an ashy fill
ditch was completely excavated over a length of with mudbrick collapse was observed. Ditch 4 is
about 5 m. This ditch is steep on the western (outer) ca. 1.5 m wide in its upper part, has a V-shape with
rather steep sides, and is filled with mudbrick col-
lapse and, in the lower part, with pottery. Within the
section, a flat lying mudbrick forms the lowermost
part of the fill.
Between the ditches, the only anthropogenic
structure observed was the round building. Since
the two structures do not disturb each other, it re-
mains so far impossible to determine their chron-
ological relationship. Radiocarbon dating is under-
way and points to a date in the early 6th mill. BCE
for both features. The function of the ditch system
so far remains unknown. No traces of postholes or
earthen embankments, as one might expect, for
example, to find in a henge monument, were ob-
served between the ditches. The radial mudbrick
walls may indicate that the ditches were at some
point connected and possibly also covered.
The MPS 4 concentric ditch system is a type
of monument so far not known anywhere in the
Southern Caucasus, and its function remains en-
tirely enigmatic at the moment. We hope that the
on-going botanical and micromorphological analysis
will provide some hints regarding its function, and
most importantly, that large-scale excavation will al-
low an investigation of its layout and of the pur-
pose of the perpendicular walls. But at this moment,
the monument appears unique in the Southern Cau-
casus, an conclusion that most certainly reflects the
Fig. 17
Site MPS 4. Detail of
status of research rather than ancient reality. The
mudbricks in the ditch application of magnetometry was crucial to the de-
tection of the structure, and this technique has

rarely been used elsewhere yet. Individual ditches
have been found occasionally during horizontal
excavation at prehistoric sites, for example, at Kvi-
riatskhali and Damtsvari Gora,38 Aruchlo,39 Imiris
Gora40 and also at Alikemektepesi.41 Ditches were
hence a widespread phenomenon during the Neo-
lithic period and are attested, for instance, also in
Turkish Thrace at Aşaǧı Pınar and at Aktopraklık in
the Marmara Region.42 Various functions have been
proposed for these constructions, including earth
quarrying for construction and irrigation canals,43
but nowhere do we yet possess any secure docu-
mentation from excavated ditches that allow analy-
sis of their function. Complex arrangements of sev-
eral concentrically nested ditches rather known from
Central and southeastern Europe. There, they are
usually surrounded by palisades and/or earthworks
that accommodated the excavated soil. These struc-
tures seem to have had varied functions, both de-
fensive and ritual. Regardless of their function, they
seem to be the outcome of collective work efforts
and are thus considered monumental in nature.44 original sounding to a 4 " 4 m square and by open- Fig. 18
Site MPS 5. View from
Work at MPS 4 has only just begun, and it is ing the northwestern quarter of the square as a south
still too early to provide conclusive interpretations 4 " 4 m trench as well (included in op. 7). Magneto-
of the ditch structure. It definitely represents a ma- metry mapping clearly showed squarish anomalies
jor collective work effort. As the oldest site so far in the central part of the mound, and a curved line-
investigated in the Mil Plain, this large-scale layout ar structure encircled the mound on its eastern
and almost monumental construction appears con- side. Following complete magnetometry mapping, a
ceptually as a rightful predecessor of the later mo- trench measuring 15 " 2 m was added in squares
numental construction at the centre of Kamiltepe. RS54 (op. 14) to test the nature of the anomalies
and of the curved linear structure.
Excavation at site MPS 5 All trenches yielded remains of densely ar-
ranged domestic architecture, with some pottery in-
MPS 5 is a small round mound of about 60 metres ventory in situ conserved. Preservation was excel-
in diameter (Fig. 18), located about 700 m north- lent, since the mound has never been touched by
west of Kamiltepe. It was detected during systema- mechanical agriculture since its abandonment in
tic field walking survey in 2010, when it yielded a the Neolithic period. The structures must have been
very rich surface collection of Neolithic pottery. left to slowly decay. At the moment, it is not known
Subsequent augering revealed ashy soil layers to whether this building level is the only occupational
a depth of 2.5 m below the highest point of the layer in MPS 5, or if older phases exist under the
mound (Fig. 19). A small 3 " 3 m sounding was uppermost building level. Preliminary radiocarbon
therefore placed there in square Q55 southeast dating evidence makes MPS 5 the latest of the ex-
(op. 7). Work proceeded in 2011 by enlarging the cavated Neolithic sites found thus far.45
38 ´/(/H/ł-ŁºŁ ´. 1992, 34–35 pl. II.; a shallow ditch was also Domestic houses in square Q55
noted surrounding the site Şaxtepe, about 9 km north of Kamil-
tepe, and at a second site without name, at the time of the Remains of two separate houses were documented
Iessen expedition, but this observation was not confirmed by (Fig. 20), but none of these were contained fully in
excavations, cp. ¨*ææ*B 1965, 13–14. It seems as if these the excavated area; therefore, the overall layout of
ditches encircling mound sites could be related to the extrac-
tion of earth for building materials. such a house is still not known. Between the two
39
Chataigner 1995, 67 with further references; Hansen et al., in houses is an open space (SU. 763/775), about 1 m
this article, for a new documentation.
40
˚$łB/(*-//¸ŁæŁ!ßB/ 1979, 10; cp. Chataigner 1995, 67.
41
Chataigner 1995, 69. 45
Preliminary dating was carried out on one sample at the Kiel
42
Pers. comm. from M. Özdoǧan and Neçmi Karul, 2011. University radiocarbon laboratory and yielded a date around
43
Chataigner 1995, 69. 5400 BCE. However, results from a larger sample series sub-
44
Furholt/Müller 2011. mitted to the French laboratory in Gif-sur-Yvette are awaited.
Fig. 19
Site MPS 5. Topogra-
phic plan
wide, which is filled with ash, pottery fragments and

bones, thus probably a garbage deposit. This ash
deposit seems to run uninterrupted through both
trenches; whether or not the house remains connect
directly remains to be investigated after removal of
the intervening baulk.
The building material is mudbrick, usually in
an elongated, rectangular shape with rounded edges.
The size of the bricks is highly irregular, about 30–
40 cm in length, and a quadrangular section of 16–
20 cm. These dimensions are very close to those of
the Kamiltepe bricks.
The western house appears in Q55 southeast
only as a row of bricks (SU. 720) that belong to a
wall probably running parallel to the outer wall of
the eastern house. In Q55NE, this line of bricks is
interrupted for about 60 cm, but then continues
farther, and the built area successively widens. The
built area encloses a compartment filled with ash
and debris (SU. 762), and extends towards the
north as a flat area completely laid out with mud-
brick, like a pavement. It cannot be decided cur-
rently whether this represents two walls standing
next to each other, or is indeed a small platform or
pavement.
The eastern built zone is less clear in the south-
ern part. Two rooms were partiall exposed there,
with walls of aligned mudbricks that form cubicles
with one straight and one curved wall. Room SU. 704
contained ceramic vessels in situ on the floor. The
outer facade of the house(s) towards the open
space remains partly unclear: fire installations made
of hand-shaped and fired mudbricks are attached
to the facade and enclose an area with a lot of ash
and debris. In the northeastern square, four aligned
rows of mudbrick again enclose an ash-filled com-
partment; attached to this is a row of perpendicular
mudbricks, which is interspaced with small quad-
rangular spacings filled with stones, ceramic frag-
ments and ash.
In summary, the architecture uncovered in the
Q55 area is built from elongated hand-shaped mud-
bricks aligned in parallel or perpendicular rows.
Ash-filled compartments are a regular feature within
the mudbrick construction, and the walls do not al-
ways extend in a straight line, but sometimes also
curve. It remains to be clarified whether the mud-
brick structures are all walls, or are instead one layer
pavements, fillings or re-enforcements.
In the gray ash-filled space between the houses
the fragment of a figurine (Fig. 21) was found. The
head and one arm are broken; a second arm is
formed as a plump protuberance. The figurine is
made of a very clean, lightly fired, brown clay. Fig. 20
Another singular object is a small, massive, vase- Site MPS 5. Plan of architecture in square Q55
shaped object with incised decoration (KAM11-
804.2; Fig. 22), probably also a schematic figurine.
Fig. 21
Site MPS 5. Figurine
Fig. 22
Site MPS 5. Figurine
Fig. 23
Site MPS 16. Topogra-
phic plan with location
of excavated squares
Excavations at MPS 16 mill. BCE sites in Azerbaijan like Menteshtepe. In ad-

dition, there were some fine, well-fired, almost clin-
MPS 16 is a shallow mounded site of about 2.5 m ky sherds comparable to the Alkhantepe material
height, located 8 km upstream from Kamiltepe on and chaff-tempered sherds of the Leylatepe type.
the right bank of the Qaraçay. The site extends over Since sites from these periods are not yet known in
approximately 230 " 130 m. It was detected in the southern Azerbaijan Mil Plain, the site was se-
2010 during the intensive walking survey and lected for further investigation through magnetome-
yielded rich sherd collections of various periods, try survey.
from Neolithic to Late Antiquity. At that time, the The magnetometry map46 revealed magnetic
whole site was under cultivation, which partly ex- anomalies over the complete extension of the site,
plains the abundance of the surface finds. Several
sherds were collected that showed the characteris-
tic traces of surface combing known from late 5th 46
which however were deeply disturbed by plough-

ing. Augering in two areas of strong magnetic
anomalies, one on the highest part of the site and
one on the southwestern slope, where the concen-
tration of Chalcolithic surface finds had been high-
est, yielded disturbed soils to a depth of 70 cm. It
was thus decided to investigate the extent of the
disturbance through two soundings of 4 " 4 m each
(both op.16), one located in square G57 and the
second in O63 (Fig. 23).
Square G 57
A 4 " 4 m sounding was dug on the southwestern
flank of the site. It yielded a layer of disturbed
(ploughed) topsoil about 50 cm thick. Underneath
was a layer of ancient topsoil with pits and a lot of lected during excavation of SU 1603, outside of the Fig. 24
Site MPS 16. Square
bioperturbation. During excavation, the final layer large pit and close to the western profile. O63, burial 1605
reached was still within the bioturbated zone. Three After reaching the large ash pit, the excava-
limestone slabs aligned along the eastern side of tion area was reduced to the southern half of the
the trench were left in place, possibly covering a trench. Underneath the layer of yellow clay were
grave. more layers of ash. These covered a first preserved
layer of architecture or mudbrick remains (Fig. 25).
Square O63 These mudbricks are 30 cm long and about 10 cm
in section. They lay in parallel rows next to each
A 4 " 4 m sounding was excavated at the highest other, extending in west-east direction. No clear
point of the site, in a location where magnetometry wall edges were visible, so the mudbricks may re-
survey results showed a large anomaly that extended present the remains of a wall collapsed sideways.
farther to the northeast as well. During the survey, Furthermore, tree roots had done much damage to
material from historical periods had also been col- the central part of this structure.
lected on top of the site. Excavation proceeded Excavation was stopped at this layer.
through about 50 cm of plough-zone material. De-
spite the historical material collected from the sur-
face, the sherd assemblage from the plough-zone
soil was mostly Chalcolithic material, indicating that
the ploughing had already damaged part of the
Chalcolithic layers.
Underneath the ploughed soil were layers of
ash that belong to a large pit/or several pits cut
into each other. The outlines of the pits became
clearer farther down; in the largest pit (SU 1602)
layers of yellowish, light clay were interlaced with
the ash layers, and when cut flat formed halfcircle-
like contours. Round holes in the layer of harder
yellow clay seem to derive from extensive bioturba-
tion (tree roots and animal holes). The yellow clay
layer seems to remain from the original pit fill that
had extended deeper in an irregular way towards
the northern edge of the pit. A later pit was defined
only after digging down in the southwestern corner
of the trench. At the bottom of this pit was a child’s
grave (SU 1605; Fig. 24), within a large cooking pot
that lay on its side with the opening facing west.
The vessel is typical of the late Chalcolithic wares
of the region. During profile cleaning above the bur- Fig. 25
ial pit, a painted fragment of Sialk III ceramic was Site MPS 16. Square
O63, collapsed mud-
found, and a second Sialk III fragment was col- brick wall
Magnetometry of neolithic sites in the Mil Plain Magnetometer prospection

of Azerbaijan
Magnetometry, among other geophysical methods,
Jörg Faßbinder, Julia Koch, Roland Linck and is a successful and cost-effective tool for the de-
Florian Becker tailed mapping of large areas within a reasonable
amount of time. For our purposes in the Azerbaijan
Mil Steppe, where it was necessary to reach the
Introduction highest possible sensitivity combined with a maxi-
mum speed of prospection, the so-called ‘‘duo-sen-
Within the framework of the cooperation between sor’’ configuration was chosen.53 In this magnet-
the Eurasian Department of the German Archaeolo- ometer configuration the probes are mounted on a
gical Institute Berlin (Germany), the Department of wooden frame and carried in a zigzag-mode 30 cm
Earth and Environmental Sciences of the Ludwig- above the ground. The profiles are oriented ap-
Maximilians-University in Munich (Germany) and the proximately east-west in order to minimize technical
National Academy of Sciences in Baku (Azerbaijan) disturbance of the magnetometer probes. During
we accomplished a geophysical prospection of ar- the period 2010–2011 solar activity and the diurnal
chaeological sites in the Mil Plain. variation, induced by solar wind, were very low.54
Magnetic prospection – for the first time ap- This situation allowed us to reduce the diurnal var-
plied in 195647 – has become one of the most im- iations to the mean value of all data of each 40 "
portant archaeological methods for the detection 40 m grid.55
and mapping of large archaeological sites.48 The The sampling frequency of the magnetometer
magnetic methods are extremely sensitive with re- (10 readings per second) provided the measurement
spect to the characterization and detection of iron of a 40-m profile of the grid in less than 30 sec-
oxides and much more sensitive than any other onds, maintaining the spatial resolution of approxi-
chemical analysis.49 Therefore, it should be empha- mately 10–15 cm at normal to fast walking speed.
sized here that sometimes many details of the soil Every 5 m, a manual switch sets a marker addition-
layers and archaeological structures in soils can be ally to the magnetic data, which is required for the
discovered and visualized only by the ‘‘magnetic correct interpolation of data during the subsequent
eye’’ and by the full understanding of their mag- laboratory processing work.
netic properties.50 However, we also have case his- The linear changes in the daily variation of
tories, in which the magnetic properties of archaeo- the geomagnetic field can be reduced to the mean
logical features resemble very much those of the value of the 40 m sampling profile or alternatively
adjacent soils and sediments, and, thus, it is im- to the mean value of all data of a 40 m grid. Here
possible to visualize these structures by magneto- it is assumed that the variation of the Earth’s mag-
meter prospecting and the resulting magnetograms.51 netic field during one profile length of 40 m follows
It is self-evident that the entire archaeological inter- a linear increase or a linear decrease in intensity. If
pretation also needs every available archaeological so, it is possible to eliminate this variation for each
background information as well as surface findings. traverse line by a reduction to the mean line value.
Many further crucial details can be derived from a This filters apparent linear structures parallel to the
well elaborated soil magnetic analysis of the data. profile. Alternatively in magnetically quiet areas it
As a result, many new archaeological questions is also useful to calculate the mean value of the
arise with the geophysical prospecting findings. whole 40 " 40 m square and use this value as it is
While for a long time it was a firm conviction described above. To create discrete field values a
of archaeologists that geophysical prospecting re- re-sampling program setting the data to 25 " 25 cm
sults on their own would be only of limited use to was used.
resolve archaeological problems,52 today it has be- In addition, by using this procedure, the dif-
come common sense that the start and the initiation ference between the measurement of both magnet-
of a modern archaeological excavation without pre- ometer probes and the theoretically calculated
vious geophysical prospecting is utterly impossible. mean value of the Earth‘s magnetic field was ob-
tained. This intensity difference gave the apparent
47 Aitken 1958; Belshe 1957. magnetic anomaly, which was caused by the mag-
48 Aitken 1974; Benech 2005; Clark 1996; David et al. 2008; Gaff- netic properties of the archaeological structure, the
ney et al. 2000; Neubauer et al. 1999; Scollar et al. 1990.
49
Dunlop/Özdemir 1997; Faßbinder et al. 1990; Faßbinder 1994.
50 53
Faßbinder/Stanjek 1993; Fröhlich et al. 2003; Schleifer et al. Becker 1999.
2003; Schleifer 2004. 54
http://www.geophysik.uni-muenchen.de/observatory/
51
Faßbinder 2009. geomagnetism.
52 55
Aspinall et al. 2008; Schmidt 2001. Faßbinder/Gorka 2009.
Fig. 26
Mil Plain, Lower Qara-
bağ region (sketch by
A. Ricci). Topographical
map of the survey area.
The locations with
settlements and tells
are marked and num-
bered by black dots;
the eight test sites for
magnetic prospection
are marked in red
soil magnetism and the geology. To cancel the nat- susceptibility -, which is the proportion of the in-
ural micro-pulsations of the Earth’s magnetic field, duced magnetization versus the intensity of the
a band-pass filter in the hardware of the magnet- magnetizing field. If we consider it simplified, this
ometer processor was used. Usually more than 90% value compares to the concentration of ferrimag-
of the magnetometer data in a 40 m grid on archae- netic minerals in the sample and gives a measure
ological sites varies in the range of )10 Nanotesla of the enhancement of magnetic minerals by the
(nT) from the corrected mean value of the geomag- settlement activity and the use of fire.56
netic field. The stronger anomalies can be ascribed
to burned structures or pieces of iron containing
slag or iron rubbish. In situ burning, pieces of iron Results of the prospections
and the traces of fire are easily distinguishable by
their different direction of magnetic dipole anoma- In general, all sites are situated on loamy and
lies, but also by their high intensities (> )50 nT). clayey soils and sediments. The enrichment of mag-
netic minerals is a widespread and typical property
Magnetic volume susceptibility of almost all soils worldwide and a crucial attribute
measurements for the successful magnetometer prospection in ar-
chaeology. However, such a type of magnetic en-
The measurement of the magnetic susceptibility hancement was rarely observed in the soils of the
was done using a commercial magnetic susceptibil- Mil Plain (see contribution and results of the anger-
ity meter or kappa-meter (SM 30 ZH-Instruments, ing-hole magnetic susceptibility measurements by
Czech Republic). The exploring coil has a diameter Ainhoa Lincot or else by our measurements on ex-
of 50 mm, the measuring frequency is 8 kHz, the cavation profiles).
sensitivity þ/––1 " 10 ––7 SI Units, and the measuring
time including a drift correction requires 8–10 sec-
onds. The metered value is the alternating field 56
Thompson/Oldfield 1986.
Fig. 27
Kamiltepe. Magneto-
gram of the site and
the vicinity of the exca-
vation trench in the
centre. Smartmag
SM4G special in duo-
sensor configuration,
total Earth magnetic
field ca. 49 230 Nano-
tesla (7/2010), dyna-
mics þ/%25 Nanotesla
in 256 gray values from
black to white, grid size
40 " 40 meter, sam-
pling density interpo-
lated to 25 " 25 cm,
reduction to the mean
value of the square
of the grid
Fig. 28
Sites MPS 2-3. Mag-
netogram of two tells
and their environment.
One is centred on the
upper right, the other
lower left (north on
top). Smartmag SM4G
special in duo-sensor
configuration, total
Earth magnetic field
ca. 49 250 Nanotesla
(7/2010), dynamics
þ/%40 Nanotesla in
256 gray values from
40 " 40 meter, sam-
value of the square
of the grid
Here we present our processed high resolu-

1 1,5 2 2,5 3 3,5
0
tion magnetograms of eight test sites (red dots,
Fig. 26) with a preliminary interpretation. The final
archaeological interpretation maps, which will in-
-1 clude all other archaeological knowledge, will be
shown in a final publication.
-2
Kamiltepe (Site MPS 1)
Kamiltepe was the aim of our first test site for mag-
depth [m]
-3
netometer prospection in Azerbaijan. It was direc-
ted towards the environment of the already partly
excavated and partly destroyed tell site of Kamil-
-4
tepe (Fig. 27). At the first sight, the resulting mag-
netogram is dominated by topographic irregularities
Fig. 29 -5
of the surface, such as the compaction of the sedi-
Sites MPS 2–3. Mag- ments by the modern road, the traces of modern
netic susceptibility
(" 10%3 SI units) versus pit digging as well as from modern irrigation sys-
profile depth drilled -6 tems and field boundaries. These structures show
into an anomaly of an Susceptibility [10E-3 SI] up by their clear and sharp negative (white) anoma-
ancient river bed
beneath the northern lies. Many ‘‘spike-like’’ anomalies (sharp black and
magnetic susceptibility drilling 2 magnetic susceptibility drilling 3
tell site white peaks) can also easily be identified by the
erratic orientation of their dipoles in the relation to
the Earth’s magnetic field. The excavation trench in
Nevertheless, our kappa measurements show the centre (white area) was excluded from our sur-
comparatively quite high values of ca. 2–3 " 10 ––3 vey. The southern and western part (left and upper
SI units. Such values compare to a relatively high part in our magnetogram) is dominated by the an-
concentration of 0.01–0.02 Vol% of ferrimagnetic cient meandering riverbeds and canals in the deep
minerals in these soils.57 sediment. All of the archaeological features adja-
A noteworthy soil formation combined with cent to the excavation trench show up only as very
the enrichment of magnetic minerals in the topsoil faint and diffuse structures and do not correspond
does not occur. This may explain the difficulty to to the excavated findings. Neither traces of the
trace small and slender archaeological structures adobe architecture, which were already found by
such as mudbrick walls or traces of single posts the excavation, nor other clearly interpretable fea-
and palisades. Even big ditches such like the ring tures are visible at the first view of the magneto-
ditch of MPS 4 (Fig. 30) with a depth of ca. 2 m gram. Only the concentration of vague and shape-
and a width of 4–5 m became only slightly visible less anomalies indicate intensive settlement activity
in the magnetogram. combined with the use of fire in the near environ-
The magnetograms of the Mil Plain are gener- ment of this tell.
ally dominated by the ancient and fossil traces of
meandering rivers as well as by big pits and ditches Sites MPS 2/3
that are filled with midden deposits and burnt ma-
terial. The ancient riverbeds show up by the en- This test site covers an area of ca. 3 ha and reveals
riched and concentrated deposits of heavy minerals a settlement including two tells, elevated about 1–
in the fluvial sediments, which produce magnetic 2 m above the plain. The magnetogram is domi-
anomalies of up to þ/––50 Nanotesla. nated by the strong and broad magnetic anomalies
The landscape survey and surface findings of of the meandering old riverbeds, which show up by
the geo-archaeological team of Andrea Ricci58 dis- the enrichment of magnetic minerals (Fig. 28). Meas-
covered a range of potential archaeological sites in urement of the magnetic susceptibility on augening-
the Mil Plain area (Fig. 26). The focus of our geo- hole samples revealed values of ca. 2.5–3.5 "
physical prospection team was, therefore, at first to 10 ––3 SI units on these river sediments. Compared
verify and clarify a range of these settlements. to the average kappa values of the archaeological
sediments and mudbricks (ca. 1.5–2.2 " 10 ––3 SI),
this value stands for the enrichment of magnetic
minerals by the factor two (Fig. 29). Interestingly,
57
Thompson/Oldfield 1986. these riverbeds do not follow today’s topography,
58
Ricci, in this article. but they seem to enclose the old settlement. It be-
Fig. 30
Site MPS 4. Magneto-
gram of a settlement
site with the traces of a
ring ditch in the central
south and other highly
magnetic pit alignments
(north on top). Smart-
mag SM4G special in
duo-sensor configura-
tion, total Earth mag-
netic field ca. 49 300
Nanotesla (7/2010), dy-
namics þ/%40 Nanotes-
la in 256 gray values
from black to white,
grid size 40 " 40 meter,
sampling density inter-
polated to 25 " 25 cm,
value of the square of
the grid
Fig. 31
gram of the tell site
and its environment.
Smartmag SM4G
(8/2011), dynamics
þ/%40 Nanotesla in
40 " 40 meter, sam-
the grid
came obvious that the first settlement was situated All these small-sized features were not detect-
on a peninsula in the shore of and/or between the able by our magnetometer measurements. This can
meandering riverbeds. Both tells show up as a ro- be explained by the measurements of the magnetic
tunda as well as areas of human activity inherently susceptibility on the profile of the excavation. On
linked with the use of fire. While the northern one adobe bricks as well as on the undisturbed adja-
shows only some diffuse and small features, the cent sediments, the kappa data revealed no discri-
southern one reveals some ditches and linear struc- minating differences in the values (ca. 1.92–2.05 "
tures as well as some large rectangular pits, indicat- 10 ––3 SI units). Only the magnetic filling of the ditch
ing heavily burned objects that are arranged around has kappa values of ca. 2.40 " 10 ––3 SI units and,
the top of the tell. hence, show some slight enhancement of magnetic
Modern and recent disturbances are due to the minerals. These magnetic data indicate that the
traces of irrigation canals, a huge modern waste pit ditch and the whole archaeological feature was
and other dugouts (sharp rectangular features). In backfilled by one quick event, rather than by a slow
this example the modern farm track is only slightly sedimentation.
visible and has no influence on our measurement.
Site MPS 5
Site MPS 4
The area of MPS 5 showed up in the topography as
The area of potential concern, MPS 4, was discov- a tell site of ca. 80 m diameter with a slight increase
ered only by the occurrence and finding of pottery and elevation of ca. 2 meters above the Mil Plain
on the surface. In the topography there is no ele- (Fig. 31). To obtain a better idea of the environment,
vation or clearly visible indication of a settlement. we enlarged our magnetometer survey to the north
A first test measurement was already performed in and to the west of the tell site. Two traces of cart
the year 2010; in 2011 we enlarged the survey area. tracks affect the site in the southwest, while in the
The data revealed a large alignment of oval pits north we see again traces of older riverbeds or ca-
with extremely high magnetic intensity of the mag- nals. The white spot in the centre marks the exca-
netic anomaly of ca. þ/––80 Nanotesla. Only one of vation trench, the deposits of the excavated soil and
these pits shows (by his adjacent negative, white another cart track are visible in the top right square
shadow) a thermo-remanent magnetization, which of the magnetogram. Furthermore, some black and
indicates fire damage. All of the other long and white spots and spikes disturb our magnetogram;
oval-shaped pits are filled and enriched by highly they are generated by pieces of iron on the top of
magnetic, but erratically oriented midden depo- the surface.
sits.59 The pits are aligned from north to south, and The tell itself becomes quite clearly visible,
some others perpendicular to it point from east to and the magnetogram reveals further details and
west and dominate the magnetogram by their high the archaeological structures of a rotunda.
magnetic anomalies. Beneath these features, how- To understand the influence of the induced
ever, we discovered further ditches and earthworks. magnetization on the magnetometer data, we per-
The magnetic intensity of these structures is very formed a magnetic susceptibility survey on a surface
weak (þ/––3 Nanotesla, see Fig. 30 bottom). A more layer of the excavation trench (Fig. 32). Herewith the
detailed analysis of the data revealed a ring-shaped resulting kappagram revealed and traced only the
ditch feature of up to 50 m in diameter, as it is also already excavated archaeological features; the re-
known from Neolithic sites in Europe (see bottom sults, however, confirmed that the archaeological
of the magnetogram Fig. 30). Another ring ditch is material has comparatively low magnetic contrast,
visible in the western part of the magnetogram, and, therefore, only little potential to trace also the
and a further semi-circular ditch encloses the large small-sized and tiny archaeological structures.
and highly magnetic pits.
The archaeological excavations of 2011 veri- Site MPS 16
fied not only the finding of the ditch, but, moreover,
revealed the occurrence of adobe wall structures Further magnetometer prospection on Neolithic
inside the ditch. They were orientated perpendicu- places was undertaken in the southwest and the
lar to the ditch and subdivide the ditch into seg- upper section of the Kura River. A natural elevation
ments. in the valley was used as shelter and heavily af-
fected by a dugout. Moreover and very recently the
area was ploughed very deeply by farmers. The re-
sulting magnetogram is, hence, extremely affected
by this activity (Fig. 33). Stripes of the ploughing
59
Le Borgne 1955; Le Borgne 1960. and farm tracks dominate the magnetogram. Never-
Site MPS 19
Site MPS 19 is a hill site on the slope of the hilly

land on the southern border of the Qaraçay valley
(Fig. 26). The topography of the site compares some-
what to a settlement mound; our magnetometer re-
sults, however, revealed a single phase settlement,
situated on an exposed topography rather than a
multi-period tell site (Fig. 36). The northern part of
the magnetogram is affected by the traces of the
modern car tracks; the other slight lineaments are
traces of the plough. A pit alignment is arranged in
a semicircle around the top of the hill; the pits fol-
low the contour line and seem to enclose some
further structures. (Note: The three black and white
spots in the centre are the anomalies of modern
iron pieces). In the southern part of the hill we dis-
covered a huge semi-circular pit, ca. 35 " 12 m in
size, indicating a highly burnt archaeological struc-
ture. The application of the high-pass filter to the
magnetic data enables us to trace further structures
inside this pit in more detail (Fig. 37). Here these
archaeological structures resemble very much a huge
Fig. 32 theless, some huge pits ca. 8–15 m in diameter be- Neolithic house foundation, which we discovered at
Site MPS 5. Kappagram
of the surface of an
come visible and detectable beneath the ground Asaǧı Pınar, Turkey.60
excavation layer (ca. and indicate the occurrence of a settlement and the
70 cm depth). Kappa intensive use of fire.
meter (SM 30 ZH-Instru- Conclusions
ments, Czech Republic).
The exploring coil has a Site MPS 18
diameter of 50 mm; the In the Mil Plain a total of 8 sites and areas ranging
measuring frequency is The MPS 18 site is a natural elevation and forms from 1–3 hectares in size, were measured by a cae-
8 kHz; the sensitivity
þ/%1 " 10%7 SI; dyna- (similar to MPS 16) a landmark in the valley. Like sium magnetometer in the total field variometer
mics 1–3 " 10%3 SI site MPS 16, it is heavily affected by a series of pits duo-sensor-configuration. This sensor configuration
units in 256 gray val- and dugouts. Meanwhile, the site is used as a farm- enabled us to trace and to detect large archaeologi-
ues from black to
white; grid size land; however, it was not as deeply ploughed as cal features to a depth of up to three meters as
4 " 4 meter; sampling the MPS 16 site. Due to the topographical situation well as near-surface archaeological structures by a
density 20 " 20 cm we tilted our survey grid from the optimal and best high spatial resolution of 25 " 25 cm, combined
orientation (east-west) in order to cover the area in with a high instrumental sensitivity of þ/––10 Pico-
a more appropriate manner (Fig. 34). tesla. For comparison only: The intensity of the to-
The site is covered by many surface finds like tal Earth’s magnetic field in the Mil Plain of Azerbai-
obsidian and pottery; moreover, in the profiles of jan in the years 2010 and 2011 was in the range of
the excavated pits archaeological layers seemed 49200 þ/––200 Nanotesla.
visible. The magnetometer results, however, show However, all of the magnetically extremely
mainly only very vanishing and fading traces of the weak anomalies of the small-sized archaeological
archaeological structures. Only on the very north- structures can be traced only if the area is almost
western edge of the magnetogram and on the topo- flat and undisturbed by modern distortion, such as
graphically most exposed top of the site, some deep ploughing, car tracks or pieces of iron on the
massive rectangular features, ca. 50 " 20 m in size surface. That was the reason why at the first sight,
are visible. In the total field magnetometer meas- the most prominent features that we detected were
urements, these features appear as a single anom- mainly the ancient and hidden meandering river-
aly with intensities of more than þ/––70 Nanotesla, beds. Borehole probing revealed that these river-
indicating an ideal thermo-remanent magnetization beds occur in a depth of up to 2–3 m. Namely, the
(TRM) anomaly. This finding can be ascribed most Kamiltepe site was already extremely destroyed by
probably to extensive fire damage of the archaeo- car tracks, by a bulldozer cut, and by the ongoing
logical feature. The application of a high-pass filter
to these data reveals the archaeological structures
in more detail (Fig. 35,1–2). 60
Faßbinder/Becker 2003.
Fig. 33
Site MPS 16. The mag-
netogram of the set-
tlement site with traces
of pits is strongly af-
fected and destroyed
by deep ploughing and
car tracks. Smartmag
SM4G special in duo-
sensor configuration,
total Earth magnetic
field ca. 49 080 Nano-
tesla (8/2011), dyna-
mics þ/%40 Nanotesla
in 256 gray values from
40 " 40 meter, sam-
the grid
Fig. 34
gram of the settlement
site and its environ-
ment. Smartmag SM4G
(9/2011), dynamics
þ/%40 Nanotesla in
40 " 40 meter, sam-
lated to 25 " 5 cm,
the grid
Fig. 35
Site MPS 18 south.
Magnetogram (1) left
side, in the total field
mode. Smartmag SM4G
(9/2011), dynamics
þ/%40 Nanotesla in
black to white; (2) the
same data but treated
with a high-pass filter,
dynamics þ/%40 Nano-
tesla in 256 gray values
from black to white,
grid size 40 " 40 meter,
sampling density inter-
polated to 25 " 25 cm,
the grid
archaeological excavation. This may explain why it huge and strong magnetic pit anomalies, which by
was utterly impossible to detect the weak magnetic further processing turned out to be large objects
structures of the adobe bricks inside the tell. with some more detailed structures inside. All in all,
At the other archaeological test sites MPS 2– the results of the magnetometer prospection not
MPS 5, the conditions for the magnetometer pro- only trace archaeological structures, but help decisi-
spection were much better (Fig. 36). The tell sites vely to detect, to trace and to discriminate different
appear with some very typical features of rotundas, types of settlements. Moreover, the geophysical re-
but the adobe features remain almost unsearch- sults originate new insights and aspects that can
able. The test sites in the southwest of the survey be essential for the success of any modern scienti-
area, MPS 16–MPS 19, were quite different from fic archaeological excavation.
the former ones. They mainly showed up by their
Fig. 36
gram of the hill, reve-
aling pits and traces of
a settlement site.
Smartmag SM4G spe-
cial in duo-sensor con-
figuration, total Earth
magnetic field ca.
49 370 Nanotesla
(8/2011), dynamics
þ/%40 Nanotesla in
40 " 40 meter, sam-
the grid
formation processes of complex deposits, or those

which are too fine to observe by eye in the field. In
addition, it enables the observation of different
components of deposits simultaneously, which can
aid significantly in the interpretation of deposits.62
Microstratigraphic analysis combines thin section
micromorphology with other high resolution analyti-
cal techniques, and has become well established as
a method for investigating activities within archaeo-
logical sites and settlements at a high spatial and
temporal resolution.63 The term ‘microarchaeology’
has been used to describe this combination of high
resolution techniques. This can be thought of in the
same way as excavation, but under the microscope,
with successive layers being observed and analysed
to understand how the deposits formed and which
were the activities represented.64
For sites that lack structural remains, this ap-
proach has allowed, for instance, the identification
of livestock enclosures65 or, on the basis of ash
residues, hearth areas.66 In early urban sites this
approach has enabled the understanding of se-
quences of activities within buildings, for example,
at Çatalhöyük, Turkey,67 and of formation processes
and, thus, activities within complex midden depos-
its68 and, in S̆eih-e Ābād in the Central Zagros re-
ˇ
gion, the investigation of evidence for early animal
Fig. 37
69
Site MPS 19. Magnetogram of the hill, processed with a high-pass management.
filter revealing the structures inside the huge pit in the south in The 2009 excavations at Kamiltepe revealed
more detail. Smartmag SM4G special in duo-sensor configuration,
total Earth magnetic field ca. 49 370 Nanotesla (8/2011), high-pass
complex deposits in a range of contexts, particu-
filter 10 " 10 (Gaussian), dynamics þ/%15 Nanotesla in 256 gray larly several large ‘ashy’ deposits along the edge
values from black to white, grid size 40 " 40 meter, sampling den- of a platform structure, and multiple floor layers
sity interpolated to 25 " 25 cm
within buildings,70 hypothesised during excavation
as being fill/feasting debris. The microstratigraphic
approach was selected to investigate the formation
processes of these deposits, to better understand
Preliminary microstratigraphic observations of ash activities occurring within buildings and open areas.
deposits and architectural materials at Kamiltepe, Analysis of the large ‘ashy’ deposits in open areas
Azerbaijan will test the hypothesis that these are the remains
of ‘feasting’. Floor layers in the round building
Lisa-Marie Shillito (MPS 4 square E29 op. 5) were selected to investi-
gate the nature of activities occurring in this area
Introduction and to assess the contribution of microstratigraphic
analysis to understanding the use of space and ac-
Our understanding of prehistoric urban societies in tivities within the settlement.
the Near East has traditionally been based on the This report presents the preliminary results
analysis of architecture and artifacts. More recently, from the analysis of seven micromorphology sam-
it has been recognised that the study of sediments
and micro-residues at such sites can be highly infor- 62
W. Matthews et al. 1997.
mative about the lives of the inhabitants, and the 63
W. Matthews et al. 1996; W. Matthews 2005; Shillito et al. 2011 a;
activities that were occurring on site.61 Shillito et al. 2011 b.
Micromorphology is the study of archaeologi- 64 Weiner 2010.
65
cal deposits in situ, in their precise depositional Shahack-Gross et al. 2003.
66
context. It is highly valuable in understanding the Weiner et al. 2002.
67
W. Matthews et al. 1996; W. Matthews 2005.
68
W. Matthews 2005; Shillito et al. 2011 a.
69
W. Matthews 2010; W. Matthews et al. 2010.
61 70
W. Matthews 2005; W. Matthews 2010; Shillito et al. 2011 a. Aliyev/Helwing 2009.
Sample no. Area and context Context type and Reason for analysis
field description
1 Kamiltepe, P12, op. 01, Northern section. Open area Formation processes and
Upper unit, yellowish activities of large ‘ashy’ layer
Middle, grey/white activities of large ‘ashy’ layer
Base, yellowish activities of large ‘ashy’ layer
TSM 5 Kamiltepe, O11, op. 02, South section Hard clay/dark ash Formation processes and
boundary activities represented by ash layer
TSM 6 Kamiltepe, O11, op. 02, South section Base of sequence, Formation processes and
midden like activities represented by ash layer
H3 MPS 4, E29, op. 05, western Sequence of floors Activities in building, nature of
profile SU 519, 520 in Neolithic round building. floor deposits
Tab. 1
Kamiltepe and site MPS H4 MPS 4, E29, op. 05, eastern Ashy layer, fill of a v-shaped Activities in ditch, nature of depo-
4. Samples selected for profile SU 516, 521, 524 ash ditch. sit formation and overlying ashy
micromorphological layer, 526 deposits
analysis
ples.71 More analyses will be undertaken to char- Results

acterize these deposits further, including additional
phytolith and geochemical analysis of the deposits. Thin section analysis has revealed a wide range of
deposit types at Kamiltepe. These can be divided
Methods broadly into architectural materials, bioarchaeologi-
cal residues and microartifacts.
Blocks were collected in the field by cutting them
directly from the section and then wrapping them Architectural materials
in tissue and tape. Samples analyzed for this initial
report are summarized in Table 1. Eight large for- A distinction is made firstly between calcitic and
mat (139 " 67 " 3 mm) thin sections were prepared gypsum based materials, and secondly basing on
by impregnation with crystic resin and polishing to the preparation of the materials and the addition of
a standard thickness of 30 microns. Before impreg- temper. The calcitic materials range from a pale
nation with resin, each block was sub-sampled for gray to pale brown (Fig. 38,1), and are mixed with
further phytolith and geochemical analysis, target- varying percentages of plant temper, identifiable
ing individual layers, so that these results could be both on the basis of pseudomorphic voids and the
directly related with the stratigraphy in thin section. presence of desiccated plant remains. The gypsum
It has been demonstrated that this microsampling based material initially thought to be a plaster ap-
is necessary to distinguish between phytolith and pears with further microscopic observation to be
geochemical signals from different activities, which secondary gypsum that has crystallized around the
are difficult to distinguish otherwise.72 Slides were organic material as a post-depositional process.
observed using a Leica DMLP microscope at magni-
fications of "10 to "400 under plane and cross Plant remains
polarized light. Measurements and images were re-
corded using IM v.3 software. Descriptions were Micromorphology enables observation of a wide
carried out according to Courty et al.73 and Bullock range of plant remains that are missed during rou-
et al.74 tine excavation and flotation.75 Indeed, material re-
covered during flotation represents only a fraction
of the plant material that was originally present on
71
Thin section slides were produced by Julie Boreham at earthsli- a site.76 Plant remains in thin section occur as
des.com. Many thanks to Helen Taylor, Ilia Helt, Georg Neumann pseudomorphic voids within the architectural mate-
and the other team members for their help during excavation
and collecting samples. Many thanks to Wendy Matthews for
rials, as microscopic charcoal, calcitic ashes, desic-
providing incredibly helpful discussions on architectural materi- cated plants and silica phytoliths. Phytoliths are mi-
als and for access to the University of Reading micromorphol- croscopic silica particles that form within and
ogy reference collection.
72
Shillito 2011.
73 75
Courty et al. 1989. W. Matthews 2010.
74 76
Bullock et al. 1985. van der Veen 2007.
Fig. 38
Kamiltepe micromor-
phology. Examples of
deposits observed in
thin section: 1. TSM 6:
gypsum material asso-
ciated with possible
omnivore faeces; 2. H4:
dark brown calcitic ag-
gregate with microchar-
coal and large plant
voids; 3. H3: fine
‘dusty’ lamination on
floor deposit; 4. H3:
thick floor with over-
lying burnt debris;
5. H3: microdebitage
embedded in floor;
6. H4: phytolith layer
between the cells of plants, and their form can be Microcharcoal is a feature in many of the deposits.
distinctive to genus, especially in monocotelyde- The presence of microcharcoal indicates the fre-
nous plants.77 They have been studied in detail at quent use of fire, with possible activities including
Çatalhöyük for example, where they have shown fuel use for heating and cooking. Conversely, pale
the use of grasses and sedges for basketry.78 calcitic ashy deposits with no microcharcoal indi-
Pseudomorphic voids can be seen embedded cate burning at a temperature greater than 500 # C,
in architectural material in Fig. 38,2, where they as charcoal is lost above this temperature.79
indicate the deliberate addition of plant temper.
77
Piperno 2006.
78 79
Ryan 2011. Boardman/Jones 1990.
The orientation of the phytoliths in different Post-depositional alterations

layers indicates there were different pathways that and taphonomy
led to their deposition on site. The partially lami-
nated structure of deposits in H4, from the fill of The deposits at Kamiltepe are subject to different
the V-shaped ditch at MPS 4, for example, suggests degrees of post-depositional disturbance, with the
these plants may have decayed in situ, and could most common being gypsum crystallization, which
represent matting, whilst the random orientation and has caused disturbance to a large proportion of the
embedded related distribution of those in TSM 5 deposits. Within buildings this appears to be more
suggest these are mixed and redeposited, for exam- localised, whereas the open area deposits are much
ple, hearth rake-out. more disturbed. Physical disturbance of some de-
Phytolith types identifiable in thin section in- posits has also occurred from bioturbation.
clude reeds, indicated by keystone bulliforms, and Water-sorted material within voids indicates
leafs/stems of monocotyledonous plants (reeds, fluctuating water levels and the washing of fine se-
sedges, grasses). Further analysis of extracted phy- diments through the deposits. The ‘ashy’ material
tolith samples will be undertaken to accurately of the large open area deposits have been subject
identify these remains and to enable quantitative to organic/iron staining that appears to be asso-
comparisons between different contexts. ciated with the abundant desiccated plant temper
in this layer.
Animal dung and human coprolites
Although there are examples of deposits with mor- Discussion
phology similar to that of animal dung,80 such as
the ashy deposits in H4, the lack of calcareous This study has indicated the importance of micro-
spherulites means that this identification cannot be stratigraphic analysis. In addition to testing hypoth-
unequivocally confirmed. Calcaerous spherulites are eses proposed during excavation, this analysis has
microscopic particles that form in the guts of certain contributed to the understanding of architectural
animals, and are a direct indicator of the presence materials, the use of plant resources, and the types
of animal dung.81 Recent studies have noted that of activities that were occurring in different parts of
these particles are not always present, and the rea- the site at Kamiltepe.
son for their presence or absence remains un-
clear.82 The same is concluded for omnivore copro- K a m i l t e p e o p. 1 , s q u a r e P 1 2 , S U 1 1 1 ,
lites – although there are abundant amorphous External Area northern section
yellow deposits with a sub-rounded morphology si-
milar to coprolites (Fig. 38,1), not all of these con- The deposits at Kamiltepe contain a wide range of
tain diagnostic inclusions, such as plant remains or architectural materials, and suggest that people had
bone fragments that could distinguish them from an advanced technological knowledge, so that they
other decayed organic remains.83 Further analysis deliberately selected specific materials for different
of faecal sterols and bile acids will be carried out purposes.
to identify these conclusively.84 The deposits in the external area (Samples 1,
2 and 3) that were hypothesised to be ash deposits
Microartifacts related to feasting, in fact, appear to be a massive
layer of constructional material, consisting of fine
Microdebitage fragments were observed embedded pale calcitic material with deliberately added large
in one of the floor layers of sample H3 (MPS 4, round volumes of plant temper, which is preserved as par-
building in square D29 op. 5). These are identified tially desiccated plants and phytoliths. There are
as particles of very angular microcrystalline silica, also 5% poorly preserved embedded bone and shell
and some are less than 1 mm thick (Fig. 38,5). fragments, which may relate to the construction of
this material, for example, as inclusions from lime
firing. More analyses combining geochemistry of the
different materials will be carried out to characterise
these further and to investigate the method of pro-
duction. Phytolith analysis will also be carried out to
assess the types of plants that were added to this
80
W. Matthews 2010. calcitic material. Preliminary observations in thin
81
Canti 1998.
82 section suggest monocotyledonous plants including
Lancelotti 2010.
83
W. Matthews 2005; Shillito et al. 2011 b. reeds and grasses. The upper part of the section is
84
Bull et al. 1999; Bull et al. 2005; Shillito et al. 2011 b. very disaggregated from gypsum recrystallization.
Activities within buildings – H3 Following these dusty layers and organic deb-
(site MPS 4, semi-subterranean round ris is a further thick floor layer, consisting of the
building) same fine grained calcitic material as the lower
floor, again containing no voids and very few micro-
The organisation of activities has been a central charcoal inclusions. This thicker layer does contain
question in understanding the use of space and a band of voids at the boundary with the over-
function of buildings. At Neolithic Çatalhöyük in lying debris, which is not continuous, indicating
Turkey, for example, there are clear distinctions be- decayed plant material. It is possible that this is a
tween ‘clean’ and ‘dirty’ floors within buildings fragment of the matting. The debris accumulating
(Matthews et al. 1996, Matthews 2005).85 on this floor is much finer than the previous floor,
The analysis of floor sequences of H3 at Ka- and is replastered with a slightly thicker layer
miltepe shows frequent replastering and a variation 826 (m.
in the types of activities represented over time in Following this replastering is an accumulation
one area of the building investigated. The presence of fine charcoal, and the distinctive layer of micro-
of ‘dusty’ layers (Fig. 38,3) and a sharp, undulating debitage fragments discussed earlier, along with
boundary between the floor surface and overlying carbonate rock pebbles, embedded within the floor.
residues indicates the possible presence of matting, These are concentrated in one layer rather than
with the fine dusty layer a result of very fine materi- being present throughout the sequence, suggesting
al becoming trapped beneath matting. The repeti- that this activity occurred in this area of the build-
tion of these layers suggests the matting was fre- ing during one phase of occupation only, before
quently shaken out, as has been observed for either stopping or moving to a different part of the
example in buildings at Tell Brak, Syria.86 The mi- building. This provides a useful context for the
cro-lamination and in situ fragmentation of grains large volume of debitage recovered from this area
and phosphatic fragments and bones is likely to be during excavation, and suggests that it relates per-
a result of trampling, as observed for deposits at haps to just a few episodes of activity rather than
Tel Dor.87 repeated ones occurring over time. Future analysis
The lowest layer of H3 is a thick prepared of samples from other areas of the building could
floor consisting of a pale brown calcitic fabric with determine whether the location of this moved dur-
no voids and few microcharcoal inclusions. This first ing the history of the building, or if this was a one-
floor layer has an undulating surface and is overlain time occurrence.
directly by a thin layer of organic debris, largely
consisting of grass-derived microcharcoal with an
orientation parallel to the floor. This suggests mate-
rial that is in situ, rather than redeposited, as this
would have disturbed the orientation. There are
also some small shell fragments here. Overlying this
floor and debris are two thinner prepared floor sur-
faces, each with another accumulation of organic
debris. The floor surfaces are ca. 2 mm thick, whilst
the accumulated debris is 1.8 mm. The second of
these two layers is overlain by a sequence of thin
dark dusty layers around 71 (m thick, which sug-
gests the area was covered with matting.88
The accumulated organic debris is a mix of
microcharcoal and a heavy density of tiny crushed
bone fragments and phosphatic material. The pre-
sence of possible calcareous spherulites89 and
bone fragments suggests that this phosphatic mate-
rial derives from omnivore coprolites.
Fig. 39
Site MPS 4. Thin sec-
85 tion H3 showing
W. Matthews et al. 1996; W. Matthews 2005.
86
W. Matthews 2010. (1) floor layers and
87 (2) the connection with
Shahack-Gross et al. 2005.
88 the wall plaster (photo
W. Matthews/Postgate 1994. by Julie Boreham at
89
Canti 1998. earthslides.com)
This is followed by a further lens of microchar- TSM 5 – Kamiltepe, operation 2,

coal, another replastering (this time the plaster con- Square O11, South section
tains a higher percentage of microcharcoal inclu-
sions) and a thicker replastering with the final The lower unit consists of a ‘hard’ material that is
accumulated debris. The final floor layer merges with fine grained with a lot of single crystals of lenticular
the wall plaster – it appears that these were con- gypsum, but few plant voids. The overlying dark
structed as one layer – and consists of a pale gray ash layer is very mixed with random orientation,
calcitic material with plant voids and microcharcoal with lots of microcharcoal giving the gray appear-
(Fig. 39). ance in the field.
Activities within buildings – MPS 4, TSM 6 – Kamiltepe, operation 2,

V-shaped ditch H4 Square O11, South section
The lowest unit consists of a calcitic aggregate, but This sample suffered significant post-depositional
differs from the sample H3 (from the round build- disturbance, and much of the micromorphology has
ing) by containing plant voids and some phytoliths. disaggregated. However, non-disturbed areas can
There is a weak orientation to the inclusions at be seen locally, and these show fine grained calcitic
45# . Overlying the aggregate is a distinctive layer of aggregate with embedded microcharcoal inclusions,
partially articulated phytoliths with a parallel orien- shell fragments and plant voids. Again this appears
tation. This may relate to decayed matting. There somewhat similar to the constructional material in
are some areas of disturbance and occasional H3, but with more temper. The middle unit contains
reddish soil particles. Overlying the phytolith layer a semi-discontinuous layer of gypsum associated
is a mixed deposit with melted silica ‘bubbles’. The with phosphate material and large conjoined husk
boundary between the aggregate and phytolith phytoliths (Fig. 38,1). The unit contains 10–20%
layer is gently undulating. Overlying the bubble charcoal, with a weak orientation, suggesting that it
layer is a large aggregate fragment, heated under was stratified before disturbance. Overall this ap-
reducing conditions, with large plant voids. The la- pears to be dumped ashy bioturbated material sit-
minated phytoliths are associated with phosphatic ting on top of a constructed surface with lots of
material, and the morphology of the layer as a plant voids.
whole appears like dung, but there are no spheru-
lites. Embedded in the ashy debris are burnt, small, Conclusions
rounded to sub-rounded, aggregate fragments and
a burnt seed, perhaps domestic cooking waste? Microstratigraphic analysis of the external ‘ashy’ de-
There are some areas where this ‘massive’ deposit posits has tested the hypothesis suggested during
has banding, which indicates that some of the mix- excavation that these layers are a result of feasting.
ing is post-depositional disturbance. Or it could be This has been refuted, showing instead that the de-
that some layers are large dumping events, with posits relate to architectural material, although the
smaller episodes of sporadic disposal. formation and function of this remains unclear and
There is a build-up of midden like material in will require further investigation. This demonstrates
the pit – the depth of this ‘midden like’ material is the importance of using microstratigraphic methods
quite shallow compared to other Near Eastern sites to correctly understand building construction and
such as Çatalhöyük90 and Sheik-e Abad,91 with a use of space. Analysis of deposits within buildings
lack of large in situ ash deposits as seen in these has revealed cycles of in situ activities, such as
earlier Neolithic sites. This lack of ash in the later lithic working and fuel burning, and has demon-
levels of Çatalhöyük is suggested to be a result of strated the importance of this approach to under-
calcite dissolving and recrystalising due to water standing the use of buildings. Observations of plant
flow at the top of the mound. At Kamiltepe, how- remains have revealed a wide range of types that
ever, it appears that ash was either not being gen- would normally have been lost through flotation;
erated in as large quantities as these other sites and they have shown their use for fuel and matting,
and/or was being reused as mortar, by mixing with in addition to uses already known for pottery tem-
fired lime paste. These hypotheses will be investi- per for example.
gated further. Further analysis of a wider range of rooms
within buildings at the site will enable more com-
parisons of the difference in activities between
buildings and the variation in materials and tech-
90
Shillito et al. 2011 a. nology for floors and surfaces. Geochemical and
91
R. Matthews et al. 2010. phytolith analysis of architectural material and ash
deposits will help to further characterise these and

the technology. In turn, these analyses make an im-
portant contribution to understanding the nature of
settlement in a region that is not well understood
in terms of the broader Near Eastern picture, and
continued analysis in future seasons will enable
comparison with sites in the wider region.
Fig. 40
Site MPS 4. Example
Ceramics of slab-technology on
ceramic fragment
The pottery of the site MPS 4 KAM10 516/8
(photo F. Geitel)
Felix Geitel
The following report provides a brief and prelimin-

ary overview in terms of wares, technology and
shapes and sizes of approximately one-third of the
ceramic material excavated so far at MPS 04 in op-
erations 5, 8, 9, 15 and 19.
Wares
Three main ware groups could be distinguished dur-
ing the analysis of the material.
A fine ‘‘untempered’’ ware with fine mineral
inclusions, presumably naturally present in the clay,
and very small organic inclusions.
A fine chaff-tempered ware with fine organic Fig. 41
Schematic sketch of the
and mineral inclusions (mineral inclusions presum- different slab-joins
ably naturally in the clay). observed at site MPS 4
A coarse chaff-tempered ware with very few (drawing F. Geitel)
mineral inclusions, but a lot of large organic inclu-
sions, mainly chaff-faced.
The surface of most of the MPS 4 pottery of all Shapes
ware categories, even the coarse ones, has a very
fine feel. They are finely smoothed and very often The range of shapes at MPS 04 is very limited. Due
burnished and/or slipped. Many sherds have inclu- to the irregularity of the rim and base sherds and
sions of mica, which are also noticeable on the sur- the degree of sherd fragmentation, the vessel forms
faces. Categorization into sub-variants and variants are difficult to reconstruct (Fig. 42). Most of the as-
as basis for a finer analysis in the future is underway. semblage seems to consist of open bowls with only
small volume capacities. Larger vessels and differ-
Technology ent shapes only appear close to the floor of the
round building (SU. 519, 520) excavated in op. 05.
All pottery from MPS 04 is handmade and appears The presence of greater variety in vessel form may
to be made from materials that were available lo- be related to the lesser degree of fragmentation in
cally. Petrographic analysis in the future will contri- these levels.
bute to testing this assumption. Apparently, almost
all of the pottery was made using slab techniques
(Figs. 40; 41).92 The pottery was fired under oxidiz-
ing conditions, but was not completely oxidized; the
cores remain light to dark gray. On the whole the
pottery appears to be very well fired under quite high
temperatures.
92
Rye 1981; Vandiver 1987.
Fig. 42
Site MPS 4. Pottery (scale 1:3; drawings and photographs by F. Geitel): 1. KAM10 511/1.3, RDm 28 cm, fine chaff-tempered ware, smoothed surface, buff; 2. KAM10
517/7.8, RDm 16 cm, fine chaff-tempered ware, smoothed, partly chaff-faced surface, reddish brown; 3. KAM10 519/4.1, RDm 8,6 cm, fine ‘‘untempered’’ ware, bur-
nished surface with a reddish slip; 4. KAM10 502/2.2, BDm 21 cm, coarse chaff-tempered ware, smoothed surface, buff; 5. KAM10 503/2.1, BDm 10 cm, coarse chaff-
tempered ware, very eroded, partly chaff-faced surface, light brown (drawing and photography F. Geitel)
The pottery of Kamiltepe (MPS 1) to the platform façade. Ceramics are highly fragmen-
ted, and complete vessels are rarely reconstruct-
Maria Bianca D’Anna able. The conclusions presented in the first preli-
minary report on Kamiltepe pottery94 are largely
During the 2011 campaign,93 the most complete confirmed by the present study.
shapes restored in the previous year were exam-
ined along with all material found in some particu- General characteristics of the assemblage
larly rich stratigraphic units of ops. 1, 2 and 6. The
ceramics of Kamiltepe itself (or MPS 1) mainly ori- The Kamiltepe assemblages include both painted
ginate from the layers of ash and debris that slope and unpainted vessels. Although final statistics are
down from the platform’s edges, but materials also not yet available and some undecorated sherds
come from fill layers related to small walls adjoined could originate from painted vessels, undecorated
pots outnumber painted ones. The majority of cera-
mics is made with medium to coarse chaff-tempered
93
During the season 2011, I dedicated two weeks and a half to fabrics. Very fine mica, fine and medium white inclu-
the study of some pottery materials from the sites of Kamiltepe sions (possibly either quartz or limestone) as well as
and MPS 5. The main aim was to gain a first insight in the
main technological and morphological characteristics of the
fine and very fine red or gray mineral inclusions are
assemblages and to plan a future systematic study of all the
ceramic materials. The pottery is described here in terms of
general characteristic of fabrics, manufacturing techniques, sur-
face treatments and firing. A non-analytical description of the
94
shape repertoire and decorations is also presented. Aliyev/Helwing 2009, 33–36.
less common, and they might be natural compo- Single or pairs of knobs are sometimes pre-
nents of the clays chosen to make the pottery.95 sent on the external rims of open and small closed
The vessels are all handmade, usually with containers.98 Knobs can be rounded, elongated,
short coils96 and slabs, which overlap only slightly squeezed or U-shaped (in one case). Horizontal or
(Fig. 43). The points of junction between two coils rounded lugs attached to the lower part of the body
are very weak and thus breakages often occur at occur on closed containers,99 and often have black-
these points. On the external side of some vessels, ish soot on the lower side; thus, they were mainly
fingerprints or small round impressions are some- used on cooking pots.
times visible: in these cases, paddling, possibly with In at least three sherds, holes had been drilled Fig. 43
small pebbles, was part of the modeling process. after firing. In one case, on a painted bowl, the hole Kamiltepe. KAM10
Cores are often not oxidized, and well-fired seems to have been drilled only from one side. In 287/3. Evidence of
coiling technique
vessels are rare. The black part of the core can be another case, two joining sherds with holes were (photo M. B. D’Anna)
either sharply distinguished by the oxidized sides found, which supports the idea that these were re-
or can gradually turn into pink, buff or pink-reddish pair holes.
colors. In the first case, fractures are rather straight
and not flaky. The non-oxidized part of the core Overview of shapes
may either cover the whole cross-section or may be
found just in its central part. Hole-mouthed jars are neckless containers with a
Surfaces are often pink, beige or light red. more or less in-turned wall and often simple or
Frequently vessels appear coated with a thin wash slightly flattened rim (Figs. 44–46). Rims tend to be
of different clay or silt and clay. In these cases they more flattened in larger containers, which present
have whitish surfaces associated with pinkish or thicker walls. The clay almost always lacks mineral
reddish matrices.97 Red slip is also attested, though inclusions, visible with the naked eye or with a
not frequently. 10"-magnifying lens. Mica is seldom present and
In general, surfaces were finished by wet occurs in very small pieces. The fracture is rather
smoothing and/or rough burnishing. However, they flaky and irregular. Clay is tempered with coarsely-
display many cracks, possibly due to a quick drying cut chaff, apparently not with straw, in very large
process. The high presence of vegetal temper often quantities. Chaff pieces generally range from 0.2 to
gives vessels a chaff-faced appearance. 0.5 mm. Pots are manufactured with coils, and the
Some general characteristics about shapes
should be briefly listed. The majority of rims is sim-
ply rounded and seems to originate from more or
less deep containers with straight or slightly in-
turned walls. Bases are always flat. In close con-
tainers, breaks often occur in the lower part of the
body wall, a few cm above the base. This could in-
dicate the use of molds to shape the lower part of
the vessels, but on the external side the clay is ty-
pically pulled from the base upwards to the vessels’
walls. Thus, it is likely that bases were shaped as a
disk and then the vessel was further modeled by
adding coils.
Vessel openings are often irregularly shaped.
Thus, when working mainly with sherds, it is not
easy to determine the inclinations or the diameters
of the openings. Moreover, hole-mouthed jars and Fig. 44
cooking pots may have had oval openings too. This Kamiltepe. KAM09
115/1.2. Hole-mouthed
must be kept in mind when reconstructing shapes jar, coarse chaff-tem-
and dimensions from rim sherds. pered ware, wet
smoothed and slightly
burnished (photo
M. B. D’Anna)
95 Schorr, in this article.
96 ‘‘Tonwülste’’, cp. Aliyev/Helwing 2009, 33.
97
This contradicts the description in the first report, where only
red slip was considered a true slip, cp. Aliyev/Helwing 2009,
33. Future studies will hopefully assess the occurence of slips,
98
and thin sections may be used to detect the presence and Aliyev/Helwing 2009, 34 Fig. 12,6; 35 Fig. 13,3.6.
99
composition of clay wash or slips, cp. Cotkin et al. 1999. Aliyev/Helwing 2009, 35 Fig. 14.
Fig. 45
Kamiltepe. KAM10 320.
Hole-mouthed jar
(drawing M. B. D’Anna)
Fig. 46
Kamiltepe. KAM10
224/146. Hole-mouthed
jar (drawing
M. B. D’Anna)
junction between two coils is often clearly detectable. gerprints are visible in the internal side of these
In larger vessels, a coil junction is regularly detect- jars. What is common in these containers is the fin-
able ca. 3 cm below the rim. At this height the pro- ishing of the surfaces, which were often carefully
file is thickened by a bump, and in some cases fin- wet-smoothed or washed with a finer and lighter
layer of clay or silt. Then, probably after drying, they
were smoothed with a hard tool. This treatment did
not always result in a uniform burnishing (but see
Fig. 47); instead, vessels have a slightly shining,
smooth and uniform appearance, even though there
are often chaff voids or cracks on the surface. The
pressure with which this light ‘burnishing’ was cre-
ated was never forceful, and it is often horizontal
or oblique, especially in the upper part of the ves-
sels, and only rarely vertical.
Fig. 47
Kamiltepe. KAM09 Sherds from pithoi are rare. One complete
111/34.1. White slipped pithos has been excavated in op. 13 (Fig. 5), but
and burnished has not yet been restored. The presence of some
hole-mouthed jar
(drawing M. B. D’Anna) diagnostic pieces (rims) allows some general remarks
on this class of containers. They display triangular,

flattened, or thickened and slightly out-turned rims.
The fabric is coarse but dense, with prevailing chaff
as non-plastic inclusion. The way pithoi are shaped
is quite consistent. Sherds are usually broken, with
a straight fracture parallel to the rim, some 7 to
10 cm from the rim. The second last coil is turned
toward the external side attaching the last coil,
which forms the external part of the rim. Thus, the
upper part of pithoi shows two overlapping layers
of clay. The rim seems shaped and smoothed turn-
ing around the thumb and the first finger. The ves-
sels’ walls are rather thick, but firing is more often
good and cores are frequently fully oxidized. Frac-
tures are pink to reddish brown, while surfaces are
whitish in color, probably because a wash was ap-
plied. Pithoi seem to have always been evenly burn-
ished.
A few rims of deep closed containers present a
depression toward the interior of the pot’s mouth.100
A complete jar with this peculiar opening was found
on an exposed profile of site MPS 18 (Fig. 178), from
where also a painted sherd101 comparable to the
Kamiltepe painted productions comes (Figs. 175;
177).102
Closed vessels with a distinguished neck or
rim are rare. Among these are a few collared jars
with large and more or less globular body;103 they
may have a painted band that decorates the
shoulders.104 Very finely finished are a few bottles,
one almost complete (Fig. 48), with a round body
and a long neck: the exterior surface is burnished
(and possibly also smoothed), white and then de-
corated with dark paint. The internal surface was
finished with a hard tool, and it is dark in color,
ranging from gray to purple, possibly because not
enough oxygen reached this part of the pot during
firing. It is interesting that at least one beaker is
very similar to these bottles in its surface treatment
and style of decoration (Fig. 49).
A significant amount of bases displays use
wear from the vessel having been utilised on a fire.
Complete cooking pots are deep, flat based con-
tainers with slightly closed walls (Fig. 50). The shape
does not differ from the common hole-mouthed
jars, and the fabric is also similar. Cooking pots are
slightly burnished, as the great bulk of Kamiltepe pottery. The bases form a sharp angle with the Fig. 48
Kamiltepe. Bottle
pots’ walls, as in all other close-shaped vessels. KAM 09 206.22 from
100
As Felix Geitel suggested, it may be a facility to grasp the ves- This characteristic is usually considered as not suit- op. 02; fine chaff-tem-
sel for transport. ed to prevent thermal shock. The bases of cooking pered ware, burnished
101 and painted
102
Ricci, in this article. pots are often oxidized, while the lower part of the
Different Neolithic and Chalcolithic occupations form the area walls are covered by a blackish or dark brown soot-
that we now recognize as site MPS 18 (Ricci, in this article). It
may be possible that the area was settled contemporary to ing, which may cover also the lower side of han-
Kamiltepe but also before; in this case, at site MPS 18 an dles, when present. Internal carbonized deposits of
occupational phase bridging the gap between of MPS 4 and charred food are rare, and they cover only the
Kamiltepe would be also documented.
103
Aliyev/Helwing 2009, 35 Fig. 13,3. base, not the pots’ walls. This might be due also to
104
Aliyev/Helwing 2009, 35 Fig. 13,4–5. the sharp angle between base and walls: the base
Fig. 49
Kamiltepe. Beaker
KAM 09 224.1.3 from
op. 02; burnished
and painted
Fig. 51
Kamiltepe. KAM11 626/1. Cooking bowl, chaff-tempered ware smoot-
hed. 1. External sooting on the lower part of the body; 2. use wears
on the internal side (photo M. B. D’Anna)
Fig. 50
Kamiltepe. Cooking pot the base, could allow the pots to be pushed or
with lug and external pulled, but not lifted, especially when they were full
sooting, coarse and hot.
chaff-tempered ware,
burnished A few truncated-conical bowls bear fire cloud-
ing on the exterior, so they might have been used
for cooking. In the best-preserved case examined
so far, internal carbon deposit is present on the
was the hottest spot, thus heating was not easily base, while the lower part of the internal walls
transmitted to all of the contents, which tended to shows traces of abrasions, which may be due to a
burn only on the base. The presence of both oxi- repeated stirring (Fig. 51,1–2). It should be men-
dized patches and chippings on the exterior side of tioned that at Hajji Firuz, large open bowls with
bases suggests that cooking pots were placed di- straight sides105 were used for cooking. Their ca-
rectly in the source of heat, which does not seem pacity ranges between 5 and 27.5 litres. Voigt sug-
to have been a fire of high flames, since dark soot- gests that their ‘‘burnished interiors would have fa-
ing coverage is limited to the lower part of the pots. cilitated the cooking of soups or stews’’.106 In
Cooking pots have rather large, non-constricted Kamiltepe, the interior is not burnished, and the
openings; hence, the food was easily accessible general characteristics of this shape suggest that
during cooking, but evaporation must have been
quite high. In some cases two opposite lugs placed 105
Voigt 1983, 107–110; 119 Fig. 78 (Form 10).
on the lower part of the vessel’s wall, not far from 106
Voigt 1983, 159.
they could have been used as cooking trays.107 In-

deed, the evaporation rate is extremely high in
Fig. 52
open-shaped pots, so they could have been used Kamiltepe. KAM10
either to bake or to quickly cook solid contents or 609/7.2. Lid, chaff-tem-
less moist ingredients than those cooked in the pered ware smoothed
with traces of sooting
deeper pots. (photo M. B. D’Anna)
A large fragment of a lid was found (Fig. 52).
It bears some fire clouding on the lower side that
may be due to its use in cooking activities.
The assemblage of open vessels includes dif-
ferent kind of open bowls108 as well as deep bowls
and beakers, differentiated on the basis of the ratio
between height and opening diameter.109 Frag-
ments of beakers are not easily distinguishable from
those of hole-mouthed jars. As bowls, they may
have one or two knobs on the rim, and painted dec-
oration occurs only on the external side. Bowls
show a marked variety of fabrics and surface finish-
ing, and they vary significantly also in shapes and
dimensions.
Some small- and medium-sized bowls with
short straight or slightly convex walls are roughly
made and finished. They were apparently made by
modeling a single lump of clay. Finer bowls of dif-
ferent sizes are well finished and painted on the in- Fig. 53,1–2
Kamiltepe. KAM09
terior (Fig. 53,1), while larger containers might be 111/10.12. Chaff-tem-
coarser in fabric.110 pered ware painted
Fragments originate from several different (photo M. B. D’Anna)
bowls with open, slightly rounded walls, whose in-
terior surface is painted with large triangles formed
by two large bands of parallel lines with different alternation in the number of lines that vary by one
orientations; in some cases a grid motif alternates element (Fig. 55). Some pieces bear a less rigid
with the parallel lines (Fig. 54). Usually the number decorative motif, and others seem to have been
of lines differs by one line. At the point at which painted by less skilled potters or in any case more
the two groups of lines encounter, by the rim, is a hastily.
fully colored triangle. In one case the triangle is iso- Generally, it can be noted that painted motifs
lated, and some lines are painted parallel to its two occur on the interior and rim of open bowls,111 on
sides. The exterior of beakers can be decorated the exterior of beakers (Fig. 49) or deep containers
with diagonal bands of parallel lines, either straight and closed shapes. Painting and applied knobs can
or forming a zigzag motif. In one case, a larger por- occur on the same vessel (Fig. 53,1–2); beakers,
tion of the vessel is preserved and shows the same deep bowls and jars present knobs and painted
Fig. 54
Kamiltepe. KAM10
206/22. Coarse chaff-
tempered ware painted
(photo and drawing
M. B. D’Anna)
107
Henrickson/McDonald 1983, 631–632.
108
Aliyev/Helwing 2009, 34 Fig. 12,1–2.6.8.
109
Aliyev/Helwing 2009, 34 Fig. 12,3–5.9–11.
110 111
Aliyev/Helwing 2009, 34 Fig. 12,8. Aliyev/Helwing 2009, 34 Fig. 12,8.
to distinguishing between different contexts, people

or foods, and the greater occurrence of decoration
on beakers and bowls rather than on jars supports
this idea. Indication of food preparation and con-
sumption at the site is also supported by the large
quantity of animal bones,116 the botanical re-
mains117 and the analysis of use-wear on stone
Fig. 55 tools.118
Kamiltepe. KAM09
206/19.12 and 22;
210/2.4; 220/7.9; The pottery of the site MPS 5
224/8.4. Chaff-tem-
pered beaker smooth- Maria Bianca D’Anna
ed, burnished and
painted (photo
M. B. D’Anna) Pottery from MPS 5 was collected from exceptionally
well preserved room contents that included vessels
in situ. These materials are mainly large jars, char-
decoration on the external surface.112 On the long-
necked bottle the decoration covers the neck and
the upper part of the body, above the point of max-
imum expansion (Fig. 48), while on jars it seems
limited to a band on the shoulder. The most com-
mon color is dark brown or black, while red paint is
more unusual but nevertheless used, especially on
coarser vessel. In some cases, the paint has very
tiny bubbles and it may be bitumen,113 while bowls
are painted on the internal surface and, possibly,
on the rim.114
To summarize, the Kamiltepe pottery produc-
tion is fairly varied in terms of functional classes,
but less in terms of shape variety, manufacturing
technique and fabrics. The control over firing was
not so steady. The abundant quantity of vegetal in-
clusions should have allowed the pots to quickly
reach a high temperature during firing. Pinkish, red-
dish or darker patches sometimes present on ves-
sels indicate that pottery was not uniformly fired.
Kamiltepe pottery seems to have been fired in open
fires, in an oxidizing atmosphere, for a short period
of time, and it is possible that more vessels were
placed one next to the other and fired together. On
the base of shape analysis, it is likely that pottery
at the 6th mill. site of Kamiltepe was used for stor-
ing diverse quantity of goods as well as for differ-
ent cooking techniques on heat sources.115 The
variety of vessels for eating and drinking appears
to be the greatest. It may be that in practices of
food consumption and sharing, attention was paid
112
Aliyev/Helwing 2009, 34 Fig. 12,6.11; 35 Fig. 13,3.6.
113
This hypothesis needs to be confirmed with scientific analysis
though. Fig. 56
114 Site MPS 5. Fragment of large hole-mouthed jar with impressed de-
Aliyev/Helwing 2009, 34 Fig. 12,6.
115 coration (photo M. B. D’Anna)
The idea that early pottery was mainly related to food cook-
ing and storage has been confronted in data sets that do not
support this hypothesis, cp. Vitelli 1989; for northern Meso- 116
potamia, see Nieuwenhuyse et al. 2010; and Nieuwenhuyse Benecke, in this article.
117
2008 for a discussion on defining Late Neolithic cultures on Decaix, in this article.
118
the base of pottery technologies and styles. Hamon, in this article.
acterized by thick walls and coarse fabrics. One of

these jars presents an impressed decoration: bands
of nail impressions or fingerprints are arranged in
garlands that hang from the rim (Fig. 56). In the
course of the 2011 season, the excavated area was
enlarged, but the materials coming from the new
trench have not yet been processed.
No painted pottery has been found at site 5
so far. Fabrics are vegetal-tempered, but, in com-
parison to Kamiltepe, the chaff is finer in dimensions,
although abundant in quantity. The clay looks in
general more compact than that of Kamiltepe cera-
mics. Mica is present, but even when abundant is
hardly visible. Mineral inclusions are occasional. In
the large containers paste is coarser and mineral in-
clusions are visible. The firing quality is irregular,
but completely oxidized sherds are often recorded.
Surfaces are either wet smoothed or covered with a
thin light-colored wash that may be a self-slip as well.
Red slip occurs too, especially on finer fabrics.
One case of a possible repair hole has been
found at this site as well.
In some cases, traces of modelling a vessel
are visible on the surface: fingerprints or impression
of what seems to be small pebbles were not erased
by further treatments and possibly testify a phase
in which the clay was regularized by paddling. Coil-
ing seems to have been employed, especially in One interesting similarity between MPS 5 and Fig. 57,1–2
Site MPS 5. KAM11
shaping the upper part of containers. The vessels Kamiltepe pottery is testified by a large fragment of 751/4.4. Pot with lug,
are rarely burnished. a container base found at Kamiltepe during the RD 35 cm
It seems that both lugs and knobs are rarer 2011 excavation season: the outer surface of this
than in the Kamiltepe assemblage, and they are lo- vessel bears the imprints of the fingers that shaped
cated a few centimetres below the rim (Fig. 57,1–2). it (Fig. 58). They are set very close to one another.
The pottery of the site MPS 16
Barbara Helwing
Pottery collected from MPS 16 mostly derives from

mixed or disturbed contexts, but is typologically lar-
gely Chalcolithic (Fig. 60; 61). Three major fabric
groups can be distinguished: (1) Fine ware, made
from fine sandy clay with no visible temper and oc-
casionally with fine organic inclusions. It seems as
if the clay had not been well prepared, since the
matrix is gagged and fissured. This fabric is used
for thin-walled, small-size vessels that can be burn-
ished on the outside; colors range from light cream
to greenish or grayish color. Fine ware was used for
the production of small bowls and carinated bowls,
as well as for closed vessels, which are only docu-
mented through rim fragments so far. (2) Medium
sandy ware with fine to medium chaff-temper, hard
Fig. 58 The upper break of the sherds shows that a coil or fired and of light brown to reddish color. Vessels of
Kamiltepe. KAM11
140/27. Chaff-tempered
slab was attached there to model the upper part. this fabric can be wet-smoothed and then display a
ware (photo Inside the vessel the clay seems to have been light-colored slip (‘‘self-slip’’). Typical shapes are
M. B. D’Anna) spread with circular movements to shape the base. jars with out-turning rim and an accentuated inner
carination, as well as bowls with an inner ledge. (3)
Medium mineral-tempered ware with inclusions of
minerals and ground schamotte. Despite their brit-
tleness, these vessels tend to be thin-walled; char-
acteristic is a surface finish by scratching with a
Fig. 59 comb-like tool that appears both on the outside
Site MPS 16. Painted and on the inside of vessels. The scratched-surface
pottery of Sialk III tradi- ware tends to be of dark color, brown or purple.
tion (photo F. Geitel)
No rim fragments or larger diagnostic sherds of this
ware have been documented yet, but it seems that
the scratching was only applied to parts of the
body. A variety of medium mineral-tempered ware
was used for the production of S-profiled jars.
These have thicker walls and a surface finish by
wet-smoothing and slight burnishing. The burial jar
of SU 1605 (Fig. 60) is made from such a mineral-
tempered ware.
Two body fragments of fine light-colored,
mineral-tempered ware with brown paint were ob-
served (Fig. 59). One (1603/5.1) shows three hori-
zontal bands and five vertical bands, of which two
are slightly curved; the pattern may be recon-
structed as a set of circles or volutes. The second
fragment (1600/1.1) has a metope-like field above
a field filled with wavy lines; inside the metope is
an animal depicted in profile. The way the animal is
shown with two legs extended to the front is typical
of Central Iranian pottery of the late Sialk III period
(Sialk III4-5). The dotted body most probably de-
notes a leopard, although the head is missing.
Fig. 60
Site MPS 16. Jar
of burial 1605
(photo F. Geitel)
Fig. 61
Site MPS 16. Chalco-
lithic ceramics (drawing
B. Helwing)
The bead workshop from site MPS 4

Type of fragment No. of fragments %
Ilia Heit
Finished beads 19 0.51
Excavations in the floor levels of the eastern part of
Unretouched blanks 79 2.10
the round building in square E 29 in 2010 yielded
numerous fragments of molluscs together with a re- Rounded retouched blanks 29 0.77
markable amount of miniature flint artifacts of less
than 2 cm in length. In the following year, systema- Broken while retouching 139 3.69
tic fine sieving of the fill layers above the floors in
the western part of the round building resulted in Broken while drilling 562 14.95
an almost tenfold number of findings, and to the Broken while grinding 34 0.90
documentation of shell- and flint finds also from
the room fill. The highest find concentration relates Waste 2904 77.11
to the floors in the centre, while their number de-
creases when approaching the edges of the round
building.
Altogether, 3766 shell fragments were col-
lected from the round building; another 11 frag-
ments were found in the adjacent ditch (SU. 516).
All mollusc artifacts were made from the shell of
the sea species Didacna that occurs in the Caspian
Sea. The shell fragments represent waste from the
production of shell beads: besides finished beads
that amount to only 0.51% of the material, the ma-
jority consists of blanks and deficient fragments in
various stages of production as well as residues de-
rived from the shell bead production (Tab. 2). The
lithic inventory is directly related to the production
of the beads and consists of an abundance of
borers and borer-like points (Fig. 62). The numer-
ous flakes and fragments are probably related to Tab. 2
the manufacture of these tools. Site MPS 4 bead workshop. Frequencies of finished beads, unfinish-
ed blanks, manufacturing failures and waste
Traces of use and working on unfinished beads
allowed the reconstruction of four subsequent work
steps in the manufacture of shell beads, which ulti-
mately were shaped into round discs with a central The third step was the most risky of the bead
hole (Fig. 63,1–2). manufacture: the drilling of the central hole. Numer-
In the first step, the shell must be broken. ous broken fragments attest to these problems. As
Whether this was done through direct percussion a rule, the drilling was located in the centre of the
with a hammer stone or through indirect percussion blank. Apparently, the surface had not been grained
using a punch is not yet known and could possibly or roughened beforehand to provide a better grip
be studied through experiments. Among the larger for the borer. This is evident on some shell frag-
fragments of sea shells are both irregular and regu- ments with punctures on their interior from missed
lar oblong pieces that display a specific break pat- attempts to start the drilling.
tern along the transverse ribs of the outer surface. In 97.11% of the cases, the drilling began on
The second work step aimed at shaping the the smooth interior surface of the sea shell. In
blank for the future bead. Almost the complete those cases where the drilling was run from the ex-
shell could be used for this task, except for the terior surface, these were mainly edge fragments
umbo. Apparently, the flat middle part of the shell with very short ribs on the interior. Apparently, these
was better suited for the production of blanks than ribs were the reason for relocating the drilling to the
the brittle and buckled edge parts. The shell frag- more brittle and rippled exterior.
ments were prepared to a roundish shape by press- The drilling was carried out with a very short
ing with a pointed tool on the projecting points borer or drill. This is attested by the profiles of the
along the edge of the blank, thus removing them. holes that appear in the shape of blunt cones on
In some cases, the round shape was set aside in those blanks left cracked during the process. In two
favour of square or irregular blanks. cases, exceptionally thick shell fragments necessi-
tated a drilling from two sides, since the length of either fine-grained light brownish or from coarse- Fig. 62
the borer may not have been enough to accomplish grained jagged gray-purple flint. Obsidian is not Site MPS4. Borers from
the bead workshop
the drilling in one uninterrupted operation. After attested at all. The most frequent type is repre- (photo I. Heit)
the drilling from one side was complete, the blank sented by short squat points with a thick middle
was turned and the second boring began from the part and a slightly tapering base with irregular re-
other side, serving to equalize the conical opening touche (Fig. 64,3–4). The terminal end is in most
to a cylinder. cases rounded or broken from wear. In addition,
175 borers and borer-like points made from there are two examples of flat long borers with reg-
flint seem to be the special tools used for the pre- ular retouche on both sides (Fig. 64,1.5). A double
paration of shells. They were produced most from borer with two pointed ends is a singular piece
Fig. 63,1–2
Site MPS 4 bead work-
shop. Reconstructing
bead manufacturing
steps, from left to right:
rounded retouched
bead blank; blank with
drilling begun from the
shell interior; per-
forated bead blank
redrilled from the shell
exterior; finished bead
with ground edges
(photo I. Heit)
Fig. 64
Site MPS 4. Drilling
tools from the bead
workshop: 1, 5 elongat-
ed slim drills; 2 double
ended drill; 3–4 short
drills (drawing I. Heit)
(Fig. 64,2). The points of several tools contained with traces of grinding. The final appearance of the
white dust in the fissures, probably residues of shell finished beads was determined through the shape
powder formed by abrasion during the drilling. and size of the blanks. Grinding served only to
In the fourth step, the edges of the bead round up the blank, apparently without an attempt
blanks were ground to a fully round shape. The risk to obtain a standardized size or a perfect circle.
of breaking seems to have been less severe than
during the previous working step, and the number
of finished shell fragments is much less than those
Kamiltepe: mineralogical studies It appears that there is no continuous profile

model because of the nature of soil deposits in the
Susan Schorr and Ainhoa Lincot valley (Fig. 67). The soil profile is constituted by
aeolian glacial deposits layers, loess layers interca-
Aspects of the geological profile lated with flood plain deposits, and clay layers.
The investigation of the soil profiles was motivated Due to a high erosion rate and a poor vegeta-
mainly by the question of the occurrence of clay de- tion cover, the upper soil layer close to the surface
posits nearby the sites, possibly within the riverbeds is in general no thicker than 20 to 30 cm. What
that are dry today. 53 drills in up to 6 m deepness should be noted is that the sedimentary sequence
were performed in order to gain insight into the is highly dependent on the localization of the drill-
geological formations accessible in the Kamiltepe ing hole. Because of the nature of the deposits, it
region (Figs. 65–66). is actually understandable that the topography
Fig. 65
Drillings 8–11 are in the vicinity of site MPS 3, would play an important role in the deposition and Drillings performed in
drillings 13–18 in MPS 4, and drillings 44–46 are erosion equilibrium, even if the variation in altitude the Kamiltepe region
close to MPS 5. Drillings 2–7 as well as 19–28 are is low (variation ')10 m, slight depressions and mapped on a Google
Earth image. The red
near Kamiltepe. The remaining drillings were com- bumps). A positive relief will enhance the erosion point marks Kamiltepe
pleted in neighboring areas. processes, while depressions will allow aeolian de- (MPS 1)
Fig. 66
Overview of the dril-
lings in the Kamiltepe
region. The red dot
marks Kamiltepe. The
green arrows indicate
drillings, in which pot-
tery and/or charcoal
was attested in the soil
samples. Red arrows
indicate the occurrence
of clays with good plas-
ticity, while orange
arrows indicate coarse
clay with poor plasti-
city. Blue arrows mark
the old river bed,
noticed in channel rem-
nants in the drillings or
visible in the landscape
posits to accumulate. Therefore, one and the same the depressions shows that the flow was probably
layer can have great lateral and upright variations a braided system with very low energy,119 which is
in thickness, can form lenses, or can even totally in concordance with the observation of a fine grain
disappear in specific locations. size of the deposits. In some locations we found
Nevertheless, reality is not that simple. deeper channels by drilling. Those sedimentary for-
Throughout the valley we found evidence for the mations showed coarser grains, which is proof of a
existence of multiple riverbeds. Some more recent higher energy hydrographical system in the past.
channels, which were still filled with water during
the last century, but appear to have dried by now,
are still noticeable in the topography. The shape of 119
Bebermeier et al., in this article.
We have found indications of anthropogenic

Fig. 67
activity (pottery, sherds and charcoal) in drillings 9 Kamiltepe region. Exam-
and 10 (close to MPS 2), 12 (close to MPS 3) and ples of lateral and up-
16 (close to MPS 4), but more impressively (also right variations in the
soil profiles (left: drill
bones) in 44, 45, and 46 (close to MPS 5) and no. 3, in the old river
additionally in 43. Charcoal was also obtained in bed; right: drill no. 5 as
drilling 54 in a depth of 320 cm. The profiles of dril- an example for clay oc-
currence). In this repre-
lings 44–46 from the vicinity of MPS 5 are de- sentation, the lighter the
scribed as exemples in the following (cp. Fig. 68): color is, the smaller the
proportion of clay. On
the contrary, when the
Drill 44 color appears to be dark
– 50–170 cm: the color of the material changes brown, the proportion of
from light brown to dark brown, the deeper – loess in the cutting is re-
duced and more clay is
the coarser, and the clay fraction increases. present. It should be no-
– a few carbonate precipitates at about 70 cm, more ted that this is a simpli-
at about 100 cm, then a lot, starting at 150 cm; fied representation ob-
tained from the drilling
numerous charcoal markings between 100 and probes soil (cylinders of
150 cm. 8 cm in diameter and
– 150–200 cm: evolution to brown-ochre color, car- 20 cm high) and visual
observations
bonate precipitates; charcoal, ceramics and red
sherds around 200 cm, still clayey-siltish.
– 200–400 cm: color changes from brown-ochre to
light brown-beige, increasing silt fraction, a few
carbonate precipitations at 350 cm. landscape in drillings 3 and 40. In the latter, the
channel formations may be present, but they are
Drill 45 deeper than or aside of the drilling performed (be-
– 20–80 cm: gray (aerosol, dusty upper layer, deep- low 4 m). This allows a reconstruction of the course
er than usual). of an old riverbed (Fig. 66). It seems that there was
– 80–130 cm: abrupt change to light brown-gray a wide, but flat river course to the south and east
color, then gradually to dark brown. of Kamiltepe, on which the large settlement Kamil- Fig. 68
Kamiltepe region. Sche-
– 130–200 cm: dark brown, clayish, pottery, char- tepe and the smaller sites MPS 2, 3 and 4 were matical profile repre-
coal and carbonate precipitates. situated. sentation of the drills
– 200–250 cm: abrupt change to light brown, The soil material obtained from the drillings 44–46 (for the legend,
see appendix). The
more siltish, still a lot of charcoal and a few car- were mainly sediments, like silt (grain size <4 (m), number on the left side
bonate precipitates, slight darkening at the base. sandy or clayish silt, sand, clay (grain size >4 (m), indicates the drilling
– 250–300 cm: black color, ash, charcoal and cera- sandy or siltish clay. These identifications were depth (in m), the num-
bers on the bottom on
mics. made by visual inspection of the drilling core mate- each profile, the coordi-
– 300–400 cm: dark brown to light brown color, rial. Coarse-grained clay was noticed in drillings 2, nates
increasing silt fraction.
Drill 46
– 40–100 cm: dark brown color, slightly clayish; pot-
tery and obsidians, traces of charcoal between
70 and 90 cm.
– 100–150 cm: ochre color, red ceramic sherds,
some traces of charcoal, a few carbonate precipi-
tations.
– 150–230 cm: abrupt change to dark gray color;
ash, pottery, animal bones and charcoal.
– 230–250 cm: gray-brown color, charcoal and car-
bonate precipitates.
– 250–400 cm: gradual change of color from light
brown to ochre, increasing silt fraction, few car-
bonate precipitates at the top of the layer.
We found indications of the old riverbed from chan-
nel layers (layers with accumulations of heavy min-
erals) in drillings 8 as well as 13 and 14, or of the
, quartz decreases the shrinkage rate of the sherds

during the firing process.
, carbonate minerals (for instance, calcite) act as
flux melting agent and support form stability in
desiccation.
Feldspars belong to the class of silicates. They exist
with a certain chemical variability within the solu-
tion series of orthoclase and albite (alkali feld-
spars) as well as of albite and anorthite (plagio-
clase; Fig. 70). Due to this variability they are
Fig. 69 suitable marker minerals for a certain clay deposit.
Kamiltepe clay samples. Thus, in a first attempt we compared the feldspar
Clay ball formed from mineral content in the pottery sherds and the clay
the clay sample
of drilling 42 samples.
EMPA analysis
5, 6 and 7. These locations are very close to Kamil- The mineralogical composition, especially the dif-
tepe, but the clay is of too poor quality (very ferent feldspar minerals, was determined qualita-
coarse grains, poor plasticity) to be suitable for pot- tively using an electron microprobe analysis system
tery. Much more fine-grained clay with good plasti- (EMPA) equipped with an EDX (energy dispersive
city (Fig. 69) was found in drillings 34, 35 and 41 X-ray spectroscopy) unit. For this investigation pol-
as well as 42. Interestingly, the latter is in the close ished sections of the pottery sherds as well as of
vicinity of drilling 43, where we found ceramic sherds the clay samples were prepared.
and traces of charcoal at a depth of about 180 cm. Fig. 71 shows the feldspar triangle for the clay
samples; Fig. 72 represents the EMPA results ob-
Mineralogical analysis of naturally occurring clay tained on the pottery samples.
and of pottery The main component of the feldspar minerals
in all the investigated clay samples was identified
Mineralogical analyses have been performed on as albite, followed by the plagioclases bytownite and
clay samples from drillings 35 and 42 as well as labradorite. Sometimes andesine occurs. On the al-
on pottery fragments (three fragments originating kali feldspar side, sanidines and anorthoclase were
from Kamiltepe 5, KAM09-209.2.2; KAM09-221-4.4; determined, but in less amount.
KAM09-221-4.5 and KAM09-223.1.1). These investi- The main components of the feldspar miner-
gations were aimed at a first estimate of their min- als in the studied pottery sherds are very similar to
eral composition. those obtained in the clay samples. The main com-
Besides clay minerals like kaolinite, illite and ponents are albite and plagioclases, here mainly
montmorillonite, the following minerals are of im- labradorite, bytownite and andesine. It should be
portance for the production of pottery: noted that there is a slight difference between the
, feldspars act as flux melting agent and lead to an pottery samples KAM09-209.2.2, KAM09-221.4.4,
increased compaction as temperatures increase. KAM09-221.4.5, KAM09-223.1.1, on the one hand,
and the three sherds from MPS 5, on the other.
The latter contain plagioclases mainly in the form
of labradorite and bywtonite with less occurrence of
andesine. For the KAM09 sherds the plagioclases
are mainly labradorite and andesine with less oc-
currence of bywtonite.
Besides, alkali feldspars (sanidines and some-
times anorthoclase) occur, but these are less pro-
nounced.
Neither the clay nor the pottery show any in-
dication for the occurrence of anorthite.
Fig. 70
Thus, it can be concluded that the mineral
Schematic representa- content concerning the feldspar minerals of the clay
tion of the chemical samples from drilling 35 and 42 is fairly close to
variability of feldspar
minerals (feldspar
that of the pottery studied. It could be an indica-
triangle) tion for their common origin.
Fig. 71
Kamiltepe clay samples.
Results of the EMPA
studies of the clay sam-
ples from drilling 35
(from upper left to se-
cond row right. SU1,
SU3, SU4 and clay ball
from 250 cm depth) as
well as drilling 42 (third
row, SU5) represented
in the feldspar triangle
Fig. 72
Kamiltepe ceramics.
Results of the EMPA
studies of the pottery
represented in the feld-
spar triangle. From up-
per left to lower right:
1 KAM09-209.2.2;
2 KAM09-221.4.4;
3 KAM09-221.4.5;
4 KAM09-223.1.1;
5–7 samples from
MPS 5
Fig. 72
cont.
Besides feldspar minerals, other silicates like

quartz, pyroxenes (augite, diopside), amphiboles Fig. 73
Kamiltepe clay samples. XRD pattern of the samples taken from drilling 35 in a depth of
(ferro-actinolithe), mica (biotite and muscovite) and 100 cm, 150 cm, 200 cm and 280 cm (from top to bottom). ! Andesine; * Quartz; ^ Calcite;
garnets were identified; apatite, calcite, dolomite " Illite; 3 Kaolinite
and baryte (the last two less pronounced) occurred

in both the clay and the pottery samples.
XRD analysis
The mineralogical composition was investigated

through X-ray diffraction using a Phillips diffracto-
meter working with CuKÆ radiation. The samples ana-
lysed stem from drilling 35 (depth: 100, 150, 200
and 280 cm) as well as from pottery sherds.
The XRD pattern of the raw material from dril-
ling 35 (Fig. 73) confirmed quartz and calcite to be
the most important non-clay mineral associated
with the feldspar (andesine). Concerning the clay
mineral content, the occurrence of illite and kaolinite
could be verified. The fraction of the clay minerals in-
creases with increasing depth, especially the amount
of illite. Illite is a clay-mica mineral. Its X-ray diffrac-
tion peaks are somehow broadened due to a small
grain size. This matches with the observation that
the clay material of this drilling taken from a depth
of 280 cm shows very good plasticity.
The XRD pattern of the pottery samples
(KAM09-223.1.1, KAM09-221.4.4, KAM09-SU312/1 Fig. 74
and KAM09-SU313/1, see Fig. 74) confirmed quartz Kamiltepe clay samples. XRD pattern of pottery samples (KAM09-223.1.1,KAM09-221.4.4,
in all samples studied. But concerning calcite and KAM09-SU312/1 and KAM09-SU313/1, from top to bottom). ! Andesine; * Quartz; ^ Calcite;
" Illite; 3 Kaolinite; * Wollastonite
clay minerals, the pottery sherds show remark-
able differences. The samples KAM09-221.4.4 and
KAM09-SU313/1 show the diffraction peaks of cal- amount of calcite. But in these two pottery sherds
cite and the clay minerals, whereas the XRD pattern the diffraction peaks of a calcium-alumosilicate oc-
of the samples KAM09-SU312/1 and KAM09-223.1.1 curs (wollastonite). This indicates the reaction of cal-
indicates the absence of calcite or only a small cite and the clay minerals during the firing process.
shifted over time. The Chrami likely flowed closer

to the east of the mound in the Neolithic period, as
indicated by the light-colored terrace edge between
the tell and the river meadow.
Located along the Kura and Chrami rivers is a
group of settlement mounds that represent the ear-
liest agricultural way of life and economy present
in southern Georgia and western Azerbaijan in the
6th mill. BCE. Settlements in this region such as S̆u-
laveris-Gora, Imiris-Gora, Chramis Didi-Gora and Aru-
chlo in Georgia as well as S̆omutepe and Toiretepe
in Azerbaijan are the earliest Neolithic known thus
far. Some of these mounds were investigated in the
1960s and 1970s, but they no longer exist today –
or if so, only in a state of ruins. Aruchlo is the north-
ernmost mound within the sign system defined until
now.
Since 2005 a German-Georgian excavation
team has carried out archaeological investigations
in Aruchlo.120 One central point of interest was di-
rected at the reconstruction of house architecture
and the collection of data related to the economy
and environment in prehistoric times. Several reports
Fig. 75 In order to reconstruct the firing process, small were already published in this series in 2005–
Kamiltepe clay samples.
XRD pattern of fired
cubes (1 cm " 1 cm " 1 cm) were formed using the 2007.121 Field work was interrupted in 2008 due to
clay cubes using a clay clay sample from drilling 35 and treated at different the Georgian-Russian war and resumed in 2009.
sample from drilling 35 temperatures in a furnace (100–800 # C in steps of Since 2010 the excavation has been part of the re-
(150 cm depth). The
firing temperature was
100 K, one sample was taken out the furnace after search program ‘‘Ancient Kura’’ on the development
100 # C, 200 # C, 300 # C, a one-hour heat treatment at the elevated tempera- of settlement and environment during the Holocene
400 # C, 500 # C, 600 # C, ture). The fired samples were investigated by X-ray in the middle Kura River valley.
700 # C and 800 # C
(from top to bottom).
diffraction (Fig. 75). The XRD pattern confirmed that From 2009 to the end of the excavation sea-
The diamonds mark the calcite is present until a firing temperature of 700 # C son in 2011 a total of 18 trenches (each a square
Bragg peaks of calcite is reached. This indicates that the reaction of calcite of 5 " 5 m) were investigated. Four more trenches
with the clay minerals to form alumosilicates takes (H, E, F, A) are located in the area of the former
place between 700 # C and 800 # C. excavations undertaken in the 1980s.122 At the in-
ternational conference ‘‘New Research on the Neo-
lithic in the Circumcaspian Regions’’, organised by
Part II: the Aruchlo team together with the Georgian Na-
tional Museum and held September 27–30, 2011,
The neolithic settlement of Aruchlo. in Tiflis, the first results of our research work could
Report on the excavations in 2009–2011 be discussed with scientists and scholars from ten
countries (Fig. 78). Parallel to the conference was
Svend Hansen and Guram Mirtskhulava an exhibition of finds from Aruchlo in the National
Museum of Georgia, organised by Irina Gamba-
Introduction schidze and Mindia Dschalabadze (Fig. 79, with
Prof. Dschaparidse).
Svend Hansen and Guram Mirtskhulava
The settlement Aruchlo I (Fig. 76) is located a few

hundred metres north on the edge of a terrace 120
Here we would like to express our gratitude to Prof. Dr. David
along a broad riverine plain. There the River Ma- Lordkipanize, General Director of the National Museum of
s̆avera exiting from the mountains joins the Chra- Georgia, for the fruitful collaborative work. We also received
constant support from Prof. Dr. B. Maisuradse (y), former di-
mi River, to then flow into the Kura River. The Cor- rector of the Archaeological Institute of the National Museum
ona satellite image (Fig. 77) shows the settlement of Georgia, who passed away far too prematurely. His succes-
mound’s location in a triangle between the hill coun- sor, Prof. Surab Macharadze was endeavoured to provide us
with support, for which we are very grateful.
try in the northwest and the confluence of both riv- 121
Hansen et al. 2006; Hansen et al. 2007; R/Bæ*B et al. 2009.
ers with their wide meadows, whose courses have 122
Chelidze/Gogelia 2004.
Fig. 76
Aruchlo. Aerial photo-
graph (photo courtesy
of I. Gambaschidze)
Fig. 77
Corona satellite
photograph,
DS1115-2154DF083
(courtesy of the U.S.
Geological Survey)
possible to recognize coeval circular walls when the

bricks of two rings were interjoined with one an-
other, whereas walls added later in time could be
easily recognized through their distinct joins. The
necessary removal of the tall profiles, of course,
aided further in clarifying the association of the cir-
cular walls.
Circular wall (L029) in Building Complex III in
trench L was already uncovered in 2007 (Fig. 81).123
This structure was built of brown mudbricks and –
differing from the assumption in 2007 – had an in-
terior diameter of 3.90–4.00 m.124 The outer diam-
eter was ca. 4.30–4.40 m; the structure had been
repaired several times. There was even an additio-
nal support added to the wall in one place (Fig. 82).
To the north another wall (AR06C031) joins the
large circular wall (Fig. 83), which is intersected by
Ring V (dark blue) and superimposed by Ring I, es-
tablishing the subsequence of Ring V after Build-
ing III. The size of the building represented by the
circular wall is uncertain. It could have been a small
Fig. 78 ring (like building complexes I and II), but an enclo-
Participants in the conference in Aruchlo (photo M. Toderaş) sure of larger size cannot be excluded.
Building Complex X in the southwest (red) has
an interior diameter of 3.54 m and an outer diam-
eter of 4.00 m (Fig. 84). It is still constituted by sev-
en layers of brown bricks that reach a total height
of 0.56 m.
Contemporary with Building Complex II (green)
is Building Complex IV, which comprises a small cir-
cular structure and a large circular wall, most of
which is demolished.125 Further remains of walls
were discovered underneath Building IV, which in
turn relate to an area distinguished by signs of se-
vere burning in the eastern part of the excavated
surfaces.
Building Complex IV overlies wall IX, which has
been revealed almost entirely in trench B (Fig. 85).
The wall measures 3.36 m in diameter and consists
of large yellow or ‘checkered’ mudbricks, which
were joined by a dark binding material (Fig. 86).
The adjacent building adjoining to the north
(AR09B049) could be determined only partially in
trench B (Fig. 81). This structure stood on the na-
tive soil. One sample taken from the floor area was
measured by Dr. Bernd Kromer in the Mannheim
laboratory (Hd-12879: 6919 ) 30BP), in the 2-sig-
ma area, and resulted in a 14C-dating between 5877
Fig. 79 Excavation contexts and 5731 cal. BCE. A certain amount of time must
Opening of the exhibi- have elapsed before Building IX was erected be-
tion in the National
Museum in Tbilisi with Svend Hansen, Guram Mirtskhulava and tween this and the building in the south.
Prof. Otar Dschaparidze Michael Ullrich
(2nd from left;
photo B. Helwing)
The advancing excavation succeeded in sorting out 123
the seeming multiplicity of building remains and in Hansen et al. 2007, 5 Fig. 6.
124
A larger diameter was assumed at that time, which is herewith
recognising interrelated structural complexes, in re- corrected, cp. Hansen et al. 2007, 5.
gard to their building history (Figs. 80–81). It was 125
Hansen et al. 2007.
Fig. 80
Aruchlo. Plan of the younger Neolithic settlement remains (plan M. Ullrich)
Fig. 81
Aruchlo. Plan of the older Neolithic settlement remains (plan M. Ullrich)
The small circular wall/ring wall (AR05A108a),

whose lower edge lies 52 cm above the aforemen-
tioned building IX, also intersects this structure.
One 14C dating (6850 BP) stems from this wall and
falls within the time span of 5770–5660 cal. BCE.
The wall belongs to the younger settlement phase
(Fig. 80), represented by the remains of buildings
XI, IV, II and I. This becomes evident when viewing
the lower edges of the buildings as well as by
stratigraphic observations. Thus, Building Complex
II (green), composed of a smaller circular structure
and a larger adjoining ring wall, overlies the older
Building Complex III (yellow). It existed at the same
time as Building Complex I (red), a small circular
structure with the two wings of an outer ring, which
adjoins the small circular structure and the outer
wall ring of Building Complex II (green). Therefore,
probably it was erected after Building II. One 14C-
dating for this small circular structure lies within the
time span of 5472 and 5229 cal. BCE in the 2-sig-
ma area (Bln. 5950: 6369 ) 46 BP).
The eastern part of circular building XI with
six preserved layers of bricks was revealed in
trench N (Fig. 87). The bricks, set longitudinally, example, now very precise data can be gathered Fig. 82
Aruchlo. Repairs on
were interrupted by two rows of laterally placed about the form of mudbricks (Figs. 92–93), and pro- circular wall III
bricks. Presumably this manner of bricklaying was cesses in the production of bricks can be recon- (AR07L029071;
meant to stabilise the building. Indeed, the wall in structed. Furthermore, the countless impressions of photo S. Hansen)
the northern part of this circular structure had been wood in the burnt clay can be carefully conserved
renewed once (Fig. 88). Inside the structure was a and will provide a much better image of the upper
large container made of unfired clay, and the en- parts of building architecture of the 6th mill. BCE in
trance to the structure, formed with a threshold of the Southern Caucasus.
clay, was found to the west.
Fig. 83
Aruchlo. Intersected
Burnt layers circular wall
(AR06C031063;
Work in 2010 included documenting a collapsed photo S. Hansen)
wall consisting of at least six layers of mudbricks
(Fig. 89). The wall belonged to a circular wall built
of yellow bricks, only part of which could be uncov-
ered, while part continued under the south profile
of the trench (Fig. 90). It still must be clarified, why
at least ten layers of bricks were set at a slant to-
wards the west, where they join to a small circular
structure of ca. one metre in diameter. In the north
the circular wall was damaged by fire (Fig. 91).
There it joins a further circular wall, only a short
stretch of which could be traced until now. Thus
far, numerous burnt architectural elements have
come to light, in particular west of the circular wall.
Moreover, traces of fire are evidenced in several
excavation trenches (especially in trenches M, R, T
and U).
Thus, this is the first time in the history of re-
search on settlements of the S̆ulaveris-S̆omutepe
group that a surface area has been revealed, in
which a larger conflagration took place. This will en-
able a fully new insight into the architecture. For
Fig. 84
Aruchlo. Circular wall X,
with building complex
III (AR07L029293) in
the background (photo
S. Hansen)
Fig. 85
Aruchlo. Circular wall IX
(AR10B053024;
photo S. Hansen)
Fig. 86 Fig. 87
Aruchlo. Section through circular wall IX (right) with ‘‘checkered’’ Aruchlo. View of circular wall XI from the east (photo S. Hansen)
bricks (AR10B053041; photo S. Hansen)
Fig. 88 Fig. 89
Aruchlo. Circular wall XI with repairs and a large container (photo S. Hansen) Aruchlo. Collapsed mudbrick wall (photo S. Hansen)
Fig. 90
Aruchlo. Circular wall
XIII (AR10U013174;
photo S. Hansen)
Fig. 92
Aruchlo. Mudbrick burnt in the conflagration (photo S. Hansen)
Fig. 91 Fig. 93
Aruchlo. Small circular wall, with traces of fire in its northern part (AR10U013047; photo S. Hansen) Aruchlo. Mudbrick burnt in the conflagration (photo S. Hansen)
Fig. 94
Aruchlo. View of the
older excavation
towards the new
excavation of 2005
(photo S. Hansen)
Two ditches
Work of the former excavations was concentrated to

the north of the area, in which our excavations have
been conducted since 2005 (Fig. 94). In order to ra-
pidly reach the oldest layers, the former excavations
installed a long, deep trench, which corresponds with
our system with squares H, E, F and A. In trench A we
encountered two ditches, which had already been
touched in the old excavation. Ditches could be
documented in the northern profile in 2011, on the
periphery of the old excavation (Fig. 95). Likewise, in
the older excavation two ditches were identified at
various places around the settlement mound. At
that time they were assumed to be an inner fortifi-
cation ditch and an outer ditch for storing rainwater
in the spring and irrigating the fields. On the other
hand, it was assumed that the outer ditch was
younger in age, that is, dated to the Bronze Age.126
These results are difficult to comprehend and to con-
firm today, as the original documentation has been
lost. Furthermore, the description of those ditches
does not comply with the ditches uncovered in 2011. The ditches found in 2011 display a V-shaped Fig. 95
Aruchlo. View of the
profile (Figs. 95–97). The ditch in the west is ca. profile of both ditches
7 m wide and 4 m deep; the ditch in the east is al- (photo S. Hansen)
126
Kiguradze 1986, 63–64. most 2 m wide and 2.25 m deep. Whereas both en-
Fig. 96
Aruchlo. Exposing
the smaller ditch
(photo S. Hansen)
Fig. 97
Aruchlo. Profile drawing
of both ditches
(plan M. Ullrich)
trenchments were superimposed by settlement

layers, they themselves cut through already existent
settlement layers. Thus, the ditches were installed
at a certain time during the already existing settle-
ments and then filled up again. This process of fill-
ing them up probably took place quickly, for there
were no signs in them of the typical sedimentation
layers that occur naturally, for instance, through
rainfall. On the other hand, a great amount of char-
coal, animal bones and numerous antler tools127
were found in both, while conversely very little pot-
tery was present.
At this point it is still too early to discuss the
function of these two ditches in Aruchlo. Today,
however, the obviously regular appearance of
ditches in Neolithic settlements is no longer exclu-
sively interpreted as defence trenches. It seems, far
more, that the extraction of earth to make ditches
and then filling them up again played a role in the
ritual sphere.128
The cremation grave in Building Complex III Fig 98

Aruchlo. Macehead
Joachim Wahl and Svend Hansen in its find context
(AR07L029015;
photo S. Hansen)
A ‘walking horizon’ could be identified at a height
of 408.10 m in trench L in Building Complex III.
Further, a macehead was found in the northern part Kurdu in the Amuq plain.129 Ten more cremations Fig. 99
of the building, directly at the wall (Fig. 98). Below are known from Yarim Tepe II, Chagar Bazar and Aruchlo. View of
the ‘walking horizon’ came a circular, light-gray ash Tell el-Kerkh of the 6th mill. BCE. The practice of Building Complex III
with the light-colored
pit (Fig. 99), from which stemmed a body sherd cremating the body was designated as a late Neo- ash feature
and burnt bones that were identified as cremated lithic ‘innovation’.130 Cremation graves are also evi- (photo S. Hansen)
human remains. The cremation was that of a 30–
40-year-old late adult, whose gender, however,
could not be determined due to the lack of compar-
able material.
Unfortunately it can no longer be determined
whether the cremation already existed when the
building was erected, or whether it was buried dur-
ing the use-phase of the building. In both cases the
association of the burial and the building is by no
means coincidental. Moreover, the unusual ring-
shaped macehead, for which no comparable exam-
ple has been found until now, was quite likely as-
sociated with the grave. Burials were also discov-
ered in earlier work in Aruchlo, but they were –
without exception – inhumations in a flexed po-
sition, some with grave goods such as carnelian
beads, etc.
The cremation grave of an adult person was
discovered in the main room of a building in Tell
127 The bone industry is currently being studied by D. Zhvania as

part of his PhD thesis at the University in Tbilisi. The com-
parative study of Aruchlo, Mentesh Tepe and Kamiltepe will
be the responsibility of G. Le Dosseur, cp. Le Dosseur, in this
129
article. Yener et al. 2000, 209.
128 130
On Neolithic ditches, cp. Jeunesse 2011. Akkermans/Schwartz 2003, 148.
Fig. 100 denced during the Neolithic period in Southeastern mound, already ended in the 6th mill. BCE. Appar-
Aruchlo. Plan of
post-Neolithic building
and Central Europe in the 6th mill. BCE.131 ently, the site was not settled afterwards for a long-
remains and pits The cremation grave in Aruchlo is the hitherto er time. Possibly there were limited activities there
(plan M. Ullrich) single example of this type of burial during the in the Bronze Age. However, more comprehensive
Neolithic in the Southern Caucasus. The small num- activities can be attested at this site again only in
ber of comparable contexts quite likely relates to the 1st mill. BCE. These include various installations
the problem of identifying the cremated remains and numerous storage pits, some of which cut deep
during excavation. into the Neolithic settlement layers (Fig. 100). Evi-
dently the brick circular enclosures of Neolithic
The post-neolithic settlement houses were used once again – intentionally – as
storage pits (Fig. 101). In places the loose fill of
Svend Hansen these pits contains ashes and a few pottery sherds,
stones and animal bones. The presumption ex-
According to the data at hand, the Neolithic settle- pressed in the last preliminary report, that the pot-
ment, which had grown to a considerably extensive tery found in these pits would possibly belong to
the Iron Age, could be substantiated by a 14C dat-
ing (Hd-28503). This dating was gained from a
131 For example, Platia Magula Zarkou, Thessaly, cp. Gallis 1996, piece of charcoal (willow or poplar) and falls in the
521 ff. one-sigma area between 756–746 cal. BCE.
The few ceramic finds (Fig. 102) from these

contexts have not been examined closely yet, but
they apparently date to the Achaemenid times.132
One noteworthy piece is the rim sherd of a vessel
bearing an inscription (Fig. 103).
The pottery
Katrin Bastert-Lamprichs
During the three excavation seasons conducted in

Aruchlo in the years 2009 to 2011, a total of
16,702 pottery sherds were quantitatively recorded
(Tab. 3) and distinguished according to ware and
type of form. This material has a total weight of
380.43 kg. Thereby, 2295 pieces were classified as
diagnostic sherds: rims, bases, handles, lids and
specifically decorated body sherds. The remaining
sherds were body pieces without any decoration.
The ratio of diagnostic pieces to body sherds is
1 : 6.3.
The majority of the recorded sherds can be
assigned to the Neolithic period; it dominates with Fig. 101
the percentage amount of 71.94% and derives from Aruchlo. Iron Age pit
in a Neolithic
stratified contexts (Tab. 4). The pottery was strongly circular structure
fragmented and very small in size. Measurements (photo S. Hansen)
of the fragments revealed that, except for an extre-
mely low proportion, all of the sherds were smaller
than 10 cm,133 that is, 55.53% of the sherds be-
longed to the category defined as Size 2, 22.71%
to Size 1, and 71.86% to Size 3.134 Despite this
strong fragmentation, three larger, almost complete
vessels could nevertheless be restored. They are all
slightly bulbous, 20–40 cm high, and show mottling
or traces of fire. For example, the slightly red clay
vessel found in 2011 is 35 cm high (Fig. 104).
Aruchlo Body sherds Diagnostic Total no.

sherds of sherds
2009 5,667 854 6,521
2010 4,484 761 5,245

Fig. 102
2011 4,256 680 4,936 Aruchlo. Large Iron
Age vessel (photo
Total 14,407 2,295 16,702 S. Hansen)
Tab. 3
Aruchlo. Ratio of body sherds to diagnostic sherds
132
Our thanks to Dr. Narimanisvili for this information.
133 In addition to weight and ware, the sherds were classified ac-
cording to size. The following four categories were establish-
ed: Size 1: sherds smaller than 2.5 " 2.5 cm; Size 2: sherds
smaller than 6.0 " 6.0 cm, but larger than 2.5 " 2.5 cm; Size 3:
sherds smaller than 10 " 10 cm, but larger than 6.0 " 6.0 cm; Fig. 103
Size 4: sherds larger than 10 " 10 cm. Aruchlo. Sherd with
134 Almost all of the few sherds of Size 4 derive from later pe-
Aramaic (?) inscription
riods in time. (photo S. Hansen)
Body Body Body Diagnostic Diagnostic Diagnostic Total no. %

sherds sherds sherds sherds sherds sherds of sherds
2009 2010 2011 2009 2010 2011
Neolithic 3,729 2,964 3,483 652 606 582 12,016 71.94

sherds
Tab. 4
Aruchlo. Ratio of body Non-Neolithic 1,938 1,520 773 202 155 98 4,686 28.06
sherds and diagnostic sherds
sherds from Neolithic
and non-Neolithic pot- Total no. 5,667 4,484 4,256 854 761 680 16,702 100
tery found during the of sherds
seasons 2009–2011
pigment could be recognized on the surface of six

sherds,136 which were possibly caused by uneven
firing.
The wares
The wares among Neolithic pottery137 that had been

analyzed during the excavations in 2005–2007 could
be confirmed in the seasons 2009–2011; in 2010
two additional wares (wares 4–5) were distinguished
as well, which are described in the following (Tab. 5).
Ware 4 (Fig. 105) is a thicker, coarsely hand-
made ware, whose outer surface usually displays a
smooth, polished and streaked slip (so-called light-
dark effect) that has flaked off in places. The slip
appears on the outer surface of all sherds recorded
until now; however, in only a few cases it was also
Fig. 104 applied to the inner surface or just a bit over the
Aruchlo. Complete rim towards the inside. This ware is poorly fired
clay vessel
(photo S. Hansen)
and very crumbly. The coarse-grained temper con-
sists of small stones, basalt and straw. The color of
the core is dark brown (7.5YR 5/6þ4/6, while the
A change in color caused by the effects of outer and inner surfaces are dark gray-brown (10YR
heat was observed on more than 30% of the sherds 4/1––4/2). The body sherds are 1 cm thick, whereas
found during all of the excavation seasons. Of the bases are somewhat thicker. Relief decoration
these, ca. 80% had considerable traces of fire and has not been observed on this ware thus far.
17% displayed a mottled surface.135 Traces of red
Ware Body Body Body Diagnostic Diagnostic Diagnostic Total no. %

sherds sherds sherds sherds sherds sherds of sherds
2009 2010 2011 2009 2010 2011
1 1,036 659 842 256 119 181 3,093 25.74
2 2,366 1,882 2,469 366 394 347 7,824 65.12
3 327 354 144 30 62 48 965 8.03

Tab. 5
Aruchlo. Quantitative 4 % 54 26 % 7 6 93 0.77
distribution of Neolithic
wares according to bo- 5 % 15 2 % 24 % 41 0.34
dy sherds and diagnos-
tic sherds, from the Total 3,729 2,964 3,483 652 606 582 12,016 100
2009–2011 seasons
136
Hansen et al. 2007, Figs. 21; 25.
135 Traces of fire and mottling were noted particularly on Neoli- 137 Wares 1–3 were described in detail in the first preliminary re-
thic pottery. port on Aruchlo, cp. Hansen et al. 2006, 15.
Ware 5 (Fig. 106) is likewise an irregular hand-

made and very light-weight ware, whose surface
displays distinct traces of longish, organic temper.
The outer surface is very well smoothed and some-
times even polished. The inner surface, oppositely,
is seldom treated. Inorganic particles were rarely
used as temper; instead, straw and chaff were em-
ployed almost exclusively. The color of the core as
well as the outer and inner surface is light brown–
yellow (7.5YR 7/4,6/4,5/4,5/6,6/6). A damaged or
very indistinct knob decoration was observed on
two sherds. The thickness of the walls varies be-
tween 0.7–0.8 cm.
Quantitative evaluation of the wares
When viewing the distribution of the five different

Neolithic wares according to body sherds and diag-
nostic sherds (Tab. 5), it is noticeable that ware 2
with 65.12% predominates over all of the pottery.
It is followed by ware 1, amounting to 25.74%.
Ware 3 amounts to far less, with 8.03%, and wares
4–5, determined later in 2010, are the least re-
presented with 0.77% and 0.34%, respectively.
Compared to the other wares, ware 2 contains the wares 1, 2 and 3 in the various trenches shows that Fig. 105
Aruchlo. Ware 4
largest number of diagnostic sherds. Sherds of examples of ware 2 found in these excavations pre- (photo V. Iosiliani)
wares 1–3 were found in almost all excavation dominate over the ceramic assemblage of Aruchlo,
areas (Tabs. 6–8). Also, the distribution of sherds of not only as a whole, but also in quantity in all of
the excavation areas.
Fig. 106
A notable quantitative concentration of wares Aruchlo. Ware 5
Area/ 1 2 3 4 5 Total % 1–3 in individual trenches has not been observed (photo V. Iosiliani)
Ware
A % 1 % % % 1 0.02
B 1 11 4 % % 16 0.37
C 55 132 13 % % 200 4.56
D 55 176 26 % % 257 5.87
F 4 7 % % % 11 0.25
G 100 161 % % % 261 5.96
I 336 659 107 % % 1,102 25.15
K 70 349 42 % % 461 10.52
L 41 236 80 % % 357 8.15
M 29 128 38 % % 195 4.45
N 303 434 19 % % 756 17.26
O 144 249 3 % % 396 9.04
P 81 82 6 % % 169 3.86
R 73 107 19 % % 199 4.54
Total 1,292 2,732 357 % % 4,381 100
Tab. 6
Aruchlo. Quantitative distribution of wares in the excavation areas
of season 2009
yet. Conversely, ware 4, which was first detected in

Area/ 1 2 3 4 5 Total %
Ware
2010, derives almost entirely from area B (Tab. 7).
It was found there (AR10B069) under the floor
A % 59 % 2 % 61 1.71 (AR10B070/B071) of one of the earliest circular
B 1 58 % 59 % 118 3.31 houses in Aruchlo.138 Circular structure no. IX
(Fig. 81) has a diameter of 3.36 m and is preserved
C 2 24 2 % 16 44 1.23
to a maximal height of 1 m. It opens to the south-
D 44 186 97 % 14 341 9.55 west, where possibly the entrance was located. It is
E 88 224 47 % 1 360 10.08 constructed of large yellow and ‘checkered’ plano-
convex mudbricks of uniform size (40 " 22 "
I 153 235 32 % % 420 11.77
0.8 cm), which were set with brown binding materi-
K 3 7 1 % % 11 0.31 al. This structure is not founded directly upon the
M % 7 2 % % 9 0.25 native soil, but was built upon the remains of a
preceding building that had been removed and of
N 117 342 58 % % 517 14.48
which only small parts of the wall in the east are
O 109 416 64 % 1 590 16.53 attested. Between the floor of the circular structure
P 64 210 54 % 4 332 9.30
and the native soil is a 10–15-cm thick cultural
layer (AR10B070) composed of dark brown to yel-
R 41 195 38 % 3 277 7.76 low clay. Found in this layer, that is, directly upon
T 79 182 14 % % 275 7.70 the native soil (AR10B071) were a number of stone
Tab. 7
Aruchlo. Quantitative
tools and two bone fragments, each next to a body
U 78 136 1 % % 215 6.02
distribution of wares in sherd of ware 4.139 The floor of the circular struc-
the excavation areas of Total 779 2,281 410 61 39 3,570 100 ture was made of dark gray clay, with some ash
season 2010
and charcoal. Many stone tools, a decorated bone
tool140 and pottery sherds – all of ware 4 – were
found on the floor. The fill of the structure likewise
contained many stone tools and one bone tool,141
but only severely burnt pottery of ware 2 and none
of ware 4.
Area/ 1 2 3 4 5 Total % The diagnostic sherds of ware 4 comprise a
Ware
burnt, thick, coarse flat base of the type 4D
A % 5 % % 1 6 0.15 (Fig. 107, AR10B069.1–4) with a diameter of 5 cm.
AA 2 35 2 % % 39 0.96 It was found directly upon the floor of the circular
structure together with a straight rim sherd with a
B 1 % % % 1 2 0.05
C 18 121 27 % % 166 4.08
D 7 49 8 % % 64 1.57
138
46 sherds of ware 4 were found upon the floor (AR10B069);
F 1 9 % % % 10 0.25 one sherd each of this ware was found under the floor of cir-
cular building IX (AR10B070) and one directly upon the native
K 11 30 19 % % 60 1.48 soil (AR10B071). Seven sherds were recovered from the north-
west corner of area B (ARB058), four sherds during the removal
M 6 36 3 % % 45 1.11
of the northern baulk to area A (AR10B075), and two sherds
N 15 115 20 % % 150 3.69 from the clay strip directly adjoining area A (AR10A143). In the
2011 season sherds of this ware were found almost entirely in
NO % 24 % 31 % 55 1.35 the debris from area NO (cp. Tab. 8).
139 These comprise the following small finds: from AR10B070 –
O 30 142 33 % % 205 5.04 find numbers 1451 and 1461 (stone artifacts); 1472, 1473
and 1527 (fragments of cooking stones); 1594 (fragment of a
P 63 225 50 % % 338 8.32 stone pounder); 1166 (piece of antler with incision); 1168
(fragment of an awl), and from AR10B071 – find numbers
R 5 26 4 % % 35 0.86
1512 and 1513 (fragments of cooking stones) and 1497 (sto-
S 10 96 % % % 106 2.61 ne artifact with a cutting edge?).
140
Finds located upon the floor (AR10B069) include nine stone ob-
T 5 38 % 1 % 44 1.08 jects and one bone tool, designated with the numbers 1409,
1410, 1507, 1508, 1509 and 1533 (fragments of cooking sto-
U 16 27 1 % % 44 1.08 nes), 1408 (fragment of a rubber), 1411 (pounder), 1532 (stone
artifact) and 1154 (decorated bone tool).
V 181 496 21 % % 698 17.17 141
Tab. 8 The fill of the room (AR10B055) contained the following seven
Aruchlo. Quantitative W 652 1,342 4 % % 1,998 49.15 stone objects and one bone tool: numbers 1302 (pounder),
distribution of wares 1308 (fragment of a grinding stone), 1309 (fragment of a sto-
in the excavation areas Total 1,023 2,816 192 32 2 4,065 100 ne), 1362 and 1387 (fragments of a grinding stone), 1399 and
of season 2011 1460 (fragments of cooking stones) and 1148 (bone tool).
mouth diameter of 5 cm, and with a closed rim

sherd of 4 cm diameter. A further diagnostic sherd
was discovered when the north baulk was removed.
This sherd was from the rim of a bowl with a radius
of 6 cm and severely burnt.
Ware 5 was primarily documented in the adja-
cent areas C and D (Tab. 7). Diagnostic sherds in-
clude a flat base of the type 4A (AR10C077.2) with
mat impressions and a diameter of 4 cm. Six further
sherds (AR10C078.1–6) were recovered, which be-
longed to one vessel with a closed rim measuring
5 cm in diameter (Fig. 108, upper left and lower
centre). Nine diagnostic sherds are from area D
(AR10D088.1–4,6; AR10D090.1–3), all of which also
stem from the rim of a vessel with a closed mouth of
ca. 5 cm in diameter. One single sherd (D100-7) be-
longs to a vessel with a straight wall and a mouth-
diameter of 7 cm. The diagnostic sherds from areas Fig. 107
O, P and R are very poorly preserved, covered with Aruchlo. Sherds of a
traces of burning and strongly eroded, so that here base, ware 4
(AR10B069.1–4; draw-
no details will be presented (cp. Tab. 7). ing I. Berdzenishvili)
Quantitative evaluation of types of forms is succeeded by form type 2 (vessels with a straight
wall) with 3.99% and 6.76%, form type 7A (body
The typological spectrum of ceramic forms that was sherds with decoration) with 5.22%, 5.28% and
documented during the excavation seasons of 2005 4.81%, and form type 4B (flat bases without any
to 2007 in Aruchlo has already been presented.142 recognizable mat impressions) with 2.45%, 1.98%
No further types were determined during the sea- and 1.03%. All of the other form types are repre-
sons in 2009 and 2011 (Tabs. 10; 12). In 2010 form sented only by a few individual pieces, for example,
type 4D was added as the sole example (Fig. 107; form type 3C (bowl with an out-curving rim) from
Tabs. 9; 11). This form type 4D is represented by 2009, or form type 3D (miniature vessel) from 2009 Fig. 108
an almost complete flat base, restored from four and 2011; and from 2010, form type 4D (coarse, Aruchlo. Diagnostic
fragments, without mat impressions. It strongly re- thick-walled flat base without recognizable mat im- sherds, ware 5
(C078.4–5/D088.1–3/
sembles form type 4A, but differs through its espe- pressions); form type 5A (horizontal grip), form type C078.1–3; photo
cially coarse production and thick walls (cp. the de- 6B (lid without a recognizable knob grip), and form S. Hansen)
scription in the section here on wares).
As can be seen in the quantitative distribution
of diagnostic sherds according to types of forms de-
noted in the year 2009 to 2011 (Tab. 9), in 2009–
2010 sherds of type 1 (closed vessel form) repre-
sented the greatest proportion of the entire assem-
blage with 37.12% or 38.12%. They were followed
by form type 4A (flat base with mat impressions)
with 25.61% and 25.41%, and by form type 3B
(bowls) with 25% and 20.3%. In the season 2011
this sequence shifted somewhat, with bowls (form
type 3B) becoming more numerous, followed by
closed vessels (type 1) and flat bases with mat im-
pressions (type 4A). These three form types domi-
nate the spectrum of forms with ca. 85% of the en-
tire sherd assemblage. This situation could already
be observed in previous excavation seasons.143 It
142
Hansen et al. 2006; Hansen et al. 2007.
143
Hansen et al. 2006; Hansen et al. 2007.
Type 1 2 3B 3C 3D 4A 4B 4D 5A 6B 7A 7D Total n
2009 242 26 163 3 1 167 16 % % % 34 % 652
% 37.12 3.99 25 0.46 0.15 25.61 2.45 % % % 5.22 % 100
2010 231 41 123 % % 154 12 4 4 2 32 3 606
Tab. 9 % 38.12 6.76 20.3 % 25.41 1.98 0.66 0.66 0.33 5.28 0.5 100
distribution of diagnos- 2011 172 29 176 % 2 165 6 % 2 % 28 2 582
tic sherds according to % 29.57 4.98 30.24 % 0.34 28.35 1.03 % 0.34 % 4.81 0.34 100
form types 1–7
Type 1 2 3B 3C 3D 4A 4B 7A Total n
Area
B % % % % % % 1 % 1
C 16 % 7 % % 7 1 2 33
D 15 1 10 % % 5 1 6 38
F % % 3 % % 3 % % 6
G 22 3 12 % % 11 % 3 51
I 53 5 33 % % 42 1 7 141
K 30 1 16 1 % 16 1 5 70
L 14 1 21 1 % 8 3 % 48
M 4 2 9 1 % 4 6 % 26
N 52 9 23 % % 31 1 3 119
O 20 2 21 % % 22 % 5 70
Tab. 10 P 7 2 3 % % 15 % 2 29
distribution of form R 9 % 5 % 1 3 1 1 20
types in the excavated Total 242 26 163 3 1 167 16 34 652
areas of season 2009
Type 1 2 3B 4A 4B 4D 5A 6B 7A 7D Total n
Area
A 2 % 3 1 % % % % % % 6
B 7 2 1 2 % 4 1 % % % 17
C 7 % 3 2 % % % % % % 12
D 28 4 7 10 % % 1 % % % 50
E 21 5 14 24 % % 1 % 3 % 68
I 20 8 14 17 % % 1 2 5 % 67
K 1 % 2 1 % % % % % % 4
M 1 % 3 % % % % % % % 4
N 32 2 22 21 2 % % % 4 % 83
O 36 9 21 25 % % % % 6 % 97
P 23 5 6 12 3 % % % 4 3 56
R 28 2 14 15 7 % % % 4 % 70
Tab. 11 T 17 3 10 11 % % % % 4 % 45
distribution of form U 8 1 3 13 % % % % 2 % 27
types in the excavated Total 231 41 123 154 12 4 4 2 32 3 606
type 7D (body sherd with a hole). During these three diagnostic sherds (Tabs. 10–12). Form type 2 is
seasons there was no enlargement of the otherwise also present in many areas. The quantitative distri-
very limited spectrum of forms. bution of diagnostic sherds according to different
In almost all of the excavated areas of the form types, as described above (Tab. 9), is also lar-
2009–2011 seasons form types 1, 3B and 4A were gely reflected in the distribution of form types in in-
found, accounting for a proportion of ca. 85% of dividual areas. Thus, the afore-named types form
Type 1 2 3B 3D 4A 4B 5A 7A 7D Total
Area
A % % 2 % % % % % % 2
AA % % 1 % 2 % % % % 3
B % % % % % % % % % %
C 8 1 3 % 6 % % % % 18
D 1 % 6 % 5 % % % % 12
F % % % % % % % 1 % 1
K 4 1 5 % 3 % % 1 % %
M 3 % 1 % 1 1 % % % 6
N 9 % 4 % 6 % % % % 19
NO 2 2 2 % % % % 1 % 7
O 22 1 6 % 12 2 % 4 % 47
P 13 1 17 % 17 % % 1 % 49
R 1 1 2 % % % % % % 4
S 1 % 4 % 2 1 2 2 % 12
T % % 1 % 2 1 % 1 % 5
U % % 2 % 1 % % 1 % 4
V 35 6 33 % 38 % % 7 2 121 Tab. 12
W 73 16 87 2 70 1 % 9 % 258 distribution of form
Total 172 29 176 2 165 6 2 28 2 582 types in the excavated
the largest groups in all of the excavation areas. 15). Flat bases of form type 4A and 4B diverge
Further, during these seasons no notable quantita- from this observation, in that they are attested
tive concentrations were observed. most among coarse ware 2.
The distribution of form types according to wares Decoration
When viewing the distribution of wares according Decoration was observed on almost one-third of
to the most frequently represented form types 1, the entire sherd assemblage. Thereby, 29.14% of
4A, 3B and 2, 4B, and 7A found during these three all documented diagnostic sherds from the exca-
seasons, it becomes apparent that all forms appear vation season in 2009, 24.26% from season 2010
relatively equally among wares 1 and 2 (Tabs. 13– and 19.76% from season 2011 display decoration
Type 1 2 3B 3C 3D 4A 4B 7A Total %
Ware
1 108 13 89 % % 45 1 11 267 40.95
2 122 13 64 2 % 118 12 23 354 54.30 Tab. 13
3 12 % 10 1 1 4 3 % 31 4.75 distribution of form
Total 242 26 163 3 1 167 16 34 652 100 types among wares
1–3 of season 2009
Type 1 2 3B 4A 4B 4D 5A 6B 7A 7D Total %
Ware
1 43 8 28 33 1 % 1 % 5 % 119 19.64
2 148 25 79 109 1 % 3 2 27 % 394 65.02
3 21 7 13 11 7 % % % % 3 62 10.23 Tab. 14
4 1 1 1 % % 4 % % % % 7 1.15 Aruchlo. Quantitative
distribution of form
5 18 % 2 1 3 % % % % % 24 3.96 types among the
Total 231 41 123 154 12 4 4 2 32 3 606 100 defined wares 1–5
of season 2010
Type 1 2 3B 4A 4B 4D 5A 6B 7A 7D Total %
Ware
1 62 7 77 1 35 % % 10 1 193 33.16
2 101 18 83 1 119 4 2 16 1 345 59.28
Tab. 15 3 12 2 12 % 11 1 % % % 38 6.53
Aruchlo. Quantitative 4 % 2 2 % % 1 % 1 % 6 1.03
distribution of form
types among the 5 % % % % % % % % % % %
defined wares 1–5 Total 175 29 174 2 165 6 2 27 2 582 100
of season 2011
al.145 The number of knobs present varies between

Season Knob decoration Relief decoration Total n
one (Fig. 109) and seven or eight (Fig. 110), while
2009 172 18 190 one example even exhibits 23 knobs (Fig. 111). The
2010 131 16 147 arrangement of knobs is usually horizontal along or
directly below the vessel’s rim. Thereby, knobs are
Tab. 16 2011 105 10 115
not necessarily always in close, uniform rows, but
distribution of Total 408 44 452 can also appear in irregular intervals (Fig. 112) in
decorated sherds connection with other applications, such as zigzag
lines (Fig. 113).
(Tab. 16). This corresponds approximately with the Pottery sherds with relief decoration present
proportionate percentages from past excavation somewhat more multifaceted motifs. Designs ob-
campaigns.144 served most frequent are semi-circles (Fig. 114) and
Of all of the kinds of decoration, knob decora- circles (Fig. 115). In addition there are wavy- and
tion is most prevalent, while raised/relief décor ap- zigzag lines (Fig. 116) and geometric patterns that
pears less often (Tab. 17). cannot be defined precisely and are mostly very
No new variations of knob-decorated pottery fragmented. All of these motifs appear in combina-
were found during these excavation seasons; the tion with knobs as well.
finds comply with the previously presented materi- Moreover, during the seasons of 2009 and
2010 a total of three sherds of ware 2 with anthro-
pomorphic depictions were documented. Concerned
Decoration 2009 2010 2011 Total n
are a vessel lid (Fig. 117, AR10I037.5–7) and two
1 Knobs 122 88 61 271 rim sherds of a bowl and cooking pot (Fig. 118,
2 Knobs 36 26 23 85
AR10P036.1; Fig. 119, AR09G012.7), each of which
displays a figural depiction.
3 Knobs 10 13 7 30
4 Knobs 1 1 7 9
5 Knobs 1 1 2 4
6 Knobs % 1 1 2
7 Knobs 1 1 2 4
8 Knobs 1 % % 1
9 Knobs % % 1 1
23 Knobs % % 1 1
Semi-circular relief 6 5 6 17
Circular relief 5 3 1 9
Wavy lines/zigzag 3 3 2 8
lines in relief
Geometric relief (?) 3 3 1 7

Tab. 17
Aruchlo. Quantitative Anthropomorphic relief 1 2 % 3
distribution of decora-
Total 190 147 115 452 Fig. 109
ted sherds, from the
Aruchlo. Knob decoration (photo V. Iosiliani)
seasons 2009–2011
144 145
Hansen et al. 2007, 13 Tab. 8. Hansen et al. 2006; Hansen et al. 2007.
Fig. 110 Fig. 111

Aruchlo. Knob decoration (drawing I. Berdzenishvili) Aruchlo. Twenty-three knobs (AR11W006-1; photo V. Iosiliani)
Fig. 112
Aruchlo. Knob decoration (AR10O061.1–2; drawing I. Berdzenish-
vili)
Fig. 113
Aruchlo. Knob decoration together with wavy line (AR09K061.1–2; K062.1–2; drawing Fig. 114
I. Berdzenishvili) Aruchlo. Semicircular relief (AR10T001.10; photo V. Iosiliani)
The lid sherds AR10I037.5–7 stem from a

large concentration of animal bones and other
Neolithic pottery sherds, which was in area I and is
interpreted as a so-called layer of debris. The cook-
ing-pot sherds AR10P036.1 are from area P, found
upon a surface of yellowish brown clay, mixed
charcoal and some collapsed mudbricks. Sherd
AR09G012.7 was discovered in a small pit together
with many other Neolithic sherds and bones. Part
of the pottery was severely burnt and covered with
sinter.
In all three cases the depiction is the frontal
Fig. 115 view of a person standing upright, whose gender
Aruchlo. Knob decoration together with circle relief (AR09I049.1; drawing I. Berdzenishvili) cannot be identified. The representation on the ves-
sel lid (Fig. 117, AR10I037.5–7) shows the left arm
of a figure, bent to the side and slightly damaged.
The arm is very long and ends in a large hand with
distinctly spread fingers. The preserved trunk and
head are relatively thin and not modelled. The legs
as well as the right arm are no longer preserved.
The second depiction (Fig. 118, AR10P036.1)
is a completely preserved, standing figure with
slightly opened legs and with arms stretched at a
slant from the body, as in the position of prayer or
dance. The trunk is relatively short, and the whole
figure is strongly stylized, but details are not shaped.
By the third figural depiction (Fig. 119,
AR09G012.7) the legs and the right lower arm are
not preserved. The trunk and head are relatively
small. The left arm is bent and the hand lengthened
considerably, as if the figure was holding some-
thing. Perhaps the small size of the figures and their
application directly on the vessel rim or shoulder
are an indication that a row of similar figures was
applied to the entire vessel.
Comparisons can be drawn with three anthro-
pomorphic depictions on pottery sherds that were
found in earlier excavations at Aruchlo. The position
of the arms of a figure depicted on the sherd of a
bowl (Fig. 120, left),146 which was found in the low-
er horizon on the floor of a circular building, resem-
bles the second figure described here (Fig. 118,
AR10P036.1). However there, the trunk is far more
elongated and a phallus is recognisable between
the spread legs. The second relatively well pre-
served comparison is from the upper horizons
(Fig. 120, right)147 and shows a bent arm like that
of the first and third figure described here (Fig. 117,
AR10I037.5–7; and Fig. 119, AR09G012.7).
The third comparison from older excavations
in Aruchlo comes from ditch II (Fig. 121, unpub-
lished. No. N158,1998). The connection of its head
146
Fig. 116 Cp. Kiguradze 1986, 12 Fig. 53,28.
147
Aruchlo. Wavy line and knob decoration (AR10O045.2; photo S. Hansen; drawing I. Berdzenishvili) Cp. Kiguradze 1986, 12 Fig. 52,29.
Fig. 117
Aruchlo. Anthropomor-
phic depiction
(AR10I037.5–7; photo
S. Hansen; drawing
I. Berdzenishvili)
Fig. 118
phic depiction
(AR10P036.1; photo
S. Hansen; drawing
I. Berdzenishvili)
Fig. 119
phic depiction
(AR09G012.7; photo
S. Hansen; drawing
I. Berdzenishvili)
Fig. 120
phic depiction, from
the older excavations,
No. 28/29; from
Kiguradze 1986)
Fig. 121
phic depiction, older
excavation in Aruchlo,
N158 (photo
S. Hansen; drawing
I. Berdzenishvili)
and trunk is not particularly modelled; only the

upper arms are preserved, and they were possibly
bent originally.148
Jewelry and ornaments

Fig. 122 Daniel Neumann
Imiris-Gora, anthro-
pomorphic depiction
(from Kiguradze 1986) During the 2009 and 2010 excavation seasons,
about one dozen small finds were recovered, which
can be assigned to the category of jewelry or orna-
ments (Fig. 124). Half of these objects were found
in contexts that can be dated to the neolithic pe-
riod without any doubt. As in previous seasons,
the drop-shaped pendants made of chalcedony or
rather carnelian are of special significance, although
148
Two further comparable anthropomorphic depictions on pot-
tery sherds are from the coeval S̆ulaveri-S̆omutepe group.
One example stems from building horizons I to I in Imiris-Gora
(Fig. 122; Kiguradze 1986, Fig. 48,5). Displayed but not well
modeled on a red-clay rim sherd of a vessel is an upright stand-
ing figure, without legs, raised bent arms and a wide belt
Fig. 123 around the waist. The other comparison is from Chramis Didi-
Chramis Didi Gora, Gora (Fig. 123; Kiguradze 1986, Fig. 69,1) and was found in
anthropomorphic one of the lower horizons. The head and left arm are missing
depiction (from in this depiction. The right arm hangs downwards; the legs
Kiguradze 1986) and the phallus between are rounded.
the number of pendants and beads made of bone

and other material has increased since then.149
A drop-shaped pendant of carnelian
(AR09K054-771) was found in 2009 while removing
the baulk between trenches K and I. This area is
located south of wall K024 from building complex II
(green), below the ‘walking horizon’ (AR09K050), to
which a small fireplace belongs (AR09K051) and a
likely coeval, small pit (AR09I030) that was filled
with ash and stones (Fig. 80). The ‘walking horizon’
does not show any later disturbances. Although the
sherd assemblage from this context is neolithic,
without exception, basing on the stratigraphic asso-
ciation the context itself is probably not Neolithic.
Likewise found in this area was a barrel-shaped bead
of organic material (AR09K054-916).
During the 2010 season another drop-shaped
pendant of carnelian was retrieved, likewise while
removing a baulk (AR10P024-1129). It derives with Fig. 124
Aruchlo. Carnelian
certainty from a group of layers of younger age, but beads and clay
its assignment to a larger find context has not been statuette (photo
possible yet. This applies to the context of a roughly S. Hansen)
chipped piece of carnelian (AR10P023-1127) as
well. The extent to which these objects represent
displaced finds from the neolithic period must re-
main unresolved at present, due to the contextual
situation. In view of the size and form, the rough Fig. 125
Aruchlo. Unworked
semi-product would have been suited for making piece of carnelian
one, or rather several, drop-shaped pendants and, and pendants
thus, can be documented as possibly the first step (photo V. Iosiliani)
in their production process (Fig. 125).
Ornaments are foremost rooted in their use in
an aesthetic as well as symbolic context, yet they tion below a ‘walking horizon’ present an initial ter-
do allow statements about the course of activities minus ante quem; however, progress in excavation
in their production. Examination of neolithic jewelry until now does not condone any closer chronologi-
and its production as compared with ethnographic cal delimitation. Although the pendants found in
observations suggests four steps: previous years strongly implied placement in the
– preparing the raw material neolithic, this assignment was made basing primar-
– shaping roughly cut pieces ily on comparison with graves found in Aruchlo 1.151
– perforating, and A supra-regional and necessarily diachronic overview
– polishing the pieces. of the manufacture of jewelry shows that forms
The numerous grinding stones found in Aruchlo, made of precious stones from the Caucasus were
used together with various fine sands, can most quite similar throughout various epochs.152 There
likely be connected primarily with shaping and pol- are, of course, differences in details (pertaining to
ishing these objects.150 form and especially the manner of production), but
Thus far, recent excavations in Aruchlo have they are difficult to evaluate on hand of only hither-
not produced clear evidence that the drop-shaped to published pieces. Until now chronological argu-
pendants can be definitely assigned to the neolithic ments, based on ornaments produced from semi-
period. Indeed, the context AR09K054 and its loca- precious stones, can be conducted only on a re-
gionally and temporally limited level.153
149 Drop-shaped pendants of carnelian prevail among the jewelry
151 Kiguradze 1986, 66–68.
from Aruchlo – more so in their formal uniformity rather than
number, see also Hansen et al. 2007, 18–19. They are similar 152 For jewelry forms of the 3rd to 1st mill. BCE, see, for instance,
in form to those pendants found in Kamiltepe, Azerbaijan: Apakidze 1999; Apakidze 2009, 94–100.
Aliyev/Helwing 2009, 36. 153
In his analysis of the Late Bronze and early Iron Age in west-
150
Cp. the results of examinations in Sitagroi/Greece with related ern Georgia, and specifically of Late Bronze Age settlement
literature, Miller 2003. See also Hamon 2008 a, and the contri- contexts from Colchis, J. Apakidze studied jewelry production
bution by Hamon, in this volume. and the development of ornament forms. He regards carnelian
The repertory of ornaments found in Aruchlo

was enhanced in the season of 2009/2010 by a lar-
ger variability on materials. For example, a perfo-
rated pendant made of stone was discovered in the
fill consisting of collapsed bricks that was clearly
from the neolithic circular building D011 (Fig. 126,
AR09D060-0920). Furthermore, several objects
made of bone, which can presumably be addressed
as jewelry, likewise stem from neolithic contexts (for
example, AR10C077-1718). Finally, mention should
be made of a bead made of glass paste, found in
the excavation in 2010, which was discovered close
to the surface and which underlines the use of the
settlement mound in antiquity.
The copper bead

Svend Hansen
Fig. 126
Aruchlo. Agate bead
(photo V. Iosiliani) In 2011 a small ring bead was discovered in Aru-
chlo (Fig. 127). It lay outside of the circular wall in
trench U. The associated context did not contain
post-Neolithic pottery or any other indications of a
later disturbance.
According to the find context, the bead is
among the oldest evidence of copper working in the
Southern Caucasus, together with such finds as the
beads from Aratashen.154 Due to its strongly cor-
roded condition, it will not be possible to clearly de-
termine whether the bead was made by hammering
or casting. An initial examination by means of X-ray
fluorescence (Tab. 18), kindly carried out by I. Gam-
baschidze, revealed components of tin, iron and ar-
Fig. 127
Aruchlo. Copper bead senic in the copper as well. At this point, however,
(photo S. Hansen) comments on this result would be premature. The
amount of tin in the bead can be interpreted var-
iously. It cannot be excluded that the bead was
shifted in location through bioturbation. It is, how-
ever, imaginable that involved here is a natural
polymetallic ore deposit. With regard to the recent
Copper beads emergence of the discussion about the oldest metal-
(larger fragments) lurgy in Southeast Europe, the copper bead found
At. numb. Element Series Intensity Concen- in Aruchlo represents an important piece of advice:
tration not to overlook the importance of Anatolia and the
47 Ag K 2,881 6.050%
Caucasus in the development of metallurgy.155
29 Cu K 94,643 93.950%
as one of the most important materials in jewelry production
Copper beads during the Bronze Age and the early Iron Age and considers
(smaller fragments) carnelian jewelry as a good indicator for dating, since prior to
the 11th –7th cent. BCE no ‘‘high quality’’ objects are attested
At. numb. Element Series Intensity Concen- whereas later-on, carnelian lost its importance in jewelry pro-
tration duction. Apakidze (2009, 96) presents drop- and pear-shaped
pendants from Tli as examples, since such artifacts are not
33 As K 2,142 3.016% attested, according to the excavators B. V. Techov and Koß,
prior to burial phase III (presented after Reinhold 2007). On
50 Sn K 5,838 8.350% the chronology of the Late Bronze and early Iron Age, see
most recently Reinhold 2007, 136–149, with particular atten-
Tab. 18 26 Fe K 7,224 3.643% tion to the Tli cemetery.
Aruchlo. Analysis of the 154
copper bead (National 29 Cu K 208,813 84.991% Badalyan et al. 2007.
155
Museum Tbilisi) Courcier 2007; Pernicka/Anthony 2010; Hansen 2011.
Several pieces of azurite (Fig. 128) show that

presumably raw materials were sought for in the
mountains along the Mas̆avera River. The copper
and gold district of Sakdrisi, where in recent years
a German-Georgian team of archaeologists could re-
search the oldest known gold mine, possibly aroused
people’s interest as early as the Neolithic period.156
Three miniature vessels found in Aruchlo were
possibly used as crucibles: a recent scientific analy-
sis undertaken in the Georgian National Museum in
Tbilisi revealed copper residues in them. Further in-
vestigations on ceramic vessels with the aid of X-
ray fluorescence, neutron activation and thin-sec-
tion examination are necessary in order to sub-
stantiate these findings.
The datings
The available 14C-datings that were carried out in

the Berlin laboratory under J. Görsdorf and in the
Mannheim laboratory under the direction of B. Kro-
mer have set the limits of the duration of settle-
ment to the time span between 5800 and 5300 BCE.
Further 14C-datings are still in process in the Co-
logne laboratory and will be presented by B. We-
ninger. In view of these datings it can be excluded often been presumed on the basis of uncalibrated Fig. 128
Aruchlo. Pieces of azu-
that Aruchlo, and likely the entire S̆ulaveri-S̆omu- dates. The results of the datings are provided here rite (photo S. Hansen)
tepe group, continued into the 5th mill. BCE, as has as a list (Tab. 19; Fig. 129).
Lab-no. Uncalib. ! Calib. BC Calib. BC Material Context no.

Age BP (1! prob.) (2! prob.)
Hd-12879 6919 30 5836-5748 5877-5731 Charcoal AR10B066-298
Bln-5854 6850 35 5759-5674 5835-5661 Bone AR05A108a
MAMS-14734 6844 26 5744-5676 5775-5665 Emmer AR11U045-359/360
MAMS-14736 6814 27 5723-5673 5736-5645 Lentil + Barley AR11AA005-369
MAMS-14738 6800 26 5716-5669 5728-5644 Barley + Emmer AR11AA009-386
MAMS-14737 6788 27 5713-5663 5724-5639 Barley AR11AA008-370
MAMS-14735 6784 26 5712-5661 5722-5638 Emmer + Barley AR11AA007-364
Hd-28506 6650 28 5621-5559 5629-5530 Barley seed AR07M013-181
Hd-28505 6591 22 5557-5491 5611-5485 Barley seed AR07K044-191
Bln-5949 6451 40 5474-5378 5482-5341 Charcoal AR06C021-151
Bln-5950 6369 46 5464-5306 5472-5229 Charcoal AR06D013-146
Hd-28504 4870 25 3692-3640 3699-3637 Barley seed AR07K006-182
Hd-28503 2489 22 756-546 767-522 Charcoal AR07L019-178

(meadow, poplar) Tab. 19
14
Aruchlo. C datings
156
Stöllner et al. 2010.
be clarified without any doubt. At present the con-

texts can be interpreted as indicating that a larger
circular wall of ca. 4 m diameter belonged to a smal-
ler circular building. With that the prerequisites for
a temporal division and a functional analysis in a
settlement of the S̆ulaveri-S̆omutepe group are ful-
filled.
The first-time ever uncovering of a burnt set-
tlement area enables fully new insight into the ar-
chitecture, which began with the process of forming
mudbricks and continuing ultimately to the wooden
construction of the buildings. The discovery of
pieces of azurite and a copper bead in the settle-
ment is evidence that the inhabitants were already
interested in metal, as has often been suggested.157
The extent to which this interest was present must
still be clarified. Evidence of two ditches within the
settlement is a new finding with regard to the S̆u-
laveri-S̆omutepe group. In the light of results from
excavations in other, more or less contemporary
settlements, ranging from the Rhine River area to
Anatolia and no less to Kamiltepe in Azerbaijan, the
presence of these ditches needs not be surprising.
However, in view of the manifold variations of Neo-
lithic ditch enclosures, one should guard against
hasty interpretations.
Fig. 129 Conclusions

Aruchlo. Graphic over- Part III:
view of the available
Neolithic 14C datings
(K. Bastert-Lamprichs) Mentesh Tepe
The excavations conducted in the years 2009–2011
brought forth an abundance of new results that sur- Bertille Lyonnet and Farhad Guliyev, in collaboration
passed those known until now. The settlement with Laurence Bouquet, Gaëlle Bruley-Chabot,
mound achieved its present-day form only in the Michel Fontugne, Pascal Raymond and
Iron Age. A ca. two-meter thick package of layers Anaı̈ck Samzun
with few buildings, various installations and large
storage pits superimposed the Neolithic structures. The site is located in the Tovuz district of Western
Until now the native soil was reached in two areas Azerbaijan (Fig. 1), along an old stream of the
(trenches A and G). According to this, the Neolithic Zeyem Çay, which flows into the Kura River on its
contexts are limited to a ca. 2.5 m thick succession right bank. The site was surveyed by I. Narimanov
of layers with numerous over-lapping circular walls. in the 1960s and described as a small mound that
The new excavations clearly revealed that not was already greatly damaged (ca. 45 m in diameter
only yellow, but also dark brown mudbricks were and 1 m high).158 Narimanov divided the pottery
utilized to a great extent for building measures. Yet, surface finds into three groups, two allegedly be-
they are scarcely distinguishable in the surrounding longing to the S̆omu-S̆ulaveri culture with either
soil of the trenches. Constant monitoring of the pro- mineral or vegetal temper, while the third group
files, which had to be sprayed with water, was in- was not attributed to a culture, but only described
dispensable for recognizing the mudbrick structures. as comb-decorated sometimes with applied pellets.
Building construction proved to be quite dense, Later on, during the 1970s, the site was totally lev-
whereby new circular structures were erected directly eled to give place to two dirt roads and to a vine-
atop the ruins of older buildings. Together with the yard equipped with long concrete poles embedded
many repairs carried out on the circular walls, the deep in the ground; they were finally removed in
impression arises of active and lasting construction the 1980s. Later, in the 1990s, the land was di-
activities.
For the first time now, the architectural asso- 157
Hansen et al. in press.
ciation of the various circular building rings could 158
Narimanov 1987, 32.
vided between two different private owners. We re- Four main periods have been identified and
discovered its approximate position during a survey are confirmed by many 14C dates (Figs. 130; 136):
conducted in 2006–2007, when we were exploring , Period I, Neolithic, 1st half of the 6th mill. BCE
for Chalcolithic sites for excavations.159 A small , Period II, Early Chalcolithic, 1st half of the 5th mill.
sounding made in 2007 showed that architecture BCE
was still visible rather deep under the surface, while , Period III, end of the Middle Chalcolithic and Tran-
the pottery that we found in the gardens was mainly sitional to Late Chalcolithic, 2nd half of the 5th mill.
of the combed-type together with a few Kura-Araxes BCE
sherds. Although at first the site seemed earlier , Period IV, Early Bronze Age, 2nd half of the 4th to
than the period we were interested in, we decided middle of the 3rd mill. BCE
to open excavations there in order to get a better Periods I and II were unearthed only at the end of
understanding of the Sioni culture, to which this last season, and the finds in situ within the archi-
combed material could be related and which was tectural remains are still few in number. Another
very poorly known until now. season of excavations this summer should bring
Together with the excavations and the studies more data, particularly, hopefully, concerning the
made on the finds (palaeobotany,160 palaeozool- apparent gap between period I and II, for which
ogy,161 chipped-stone industry,162 bone industry,163 several hints lead us to suppose that it will at least
macrolithic tools,164 ceramics,165 metallurgy,166 ra- partly be filled in (see below, ceramics). For this
diocarbon dating167), a research study dedicated to reason, our presentation here of these two early
the understanding of the environment was con- periods should be considered as provisional.
ducted in a large area around the site.168 The pre- Period I, discovered at ca. 2 m below the
liminary results of all these studies are presented in actual surface and sealed by a level of yellow clay-
this report. ish sediment is characterized by round architecture
of mudbricks (50 " 25 " 10 cm) (Fig. 131). Already
Excavations and periodisation three levels of construction are visible, each of
them only preserved to a slight height. Two 14C
Despite the unfortunate fate of the mound, four sea- dates place it within the 1st half of the 6th mill. BCE.
sons of excavations at Mentesh Tepe have given ex- The diameter of the round loci varies from 3 to 6 m
ceptional results, showing an almost uninterrupted in diameter and walls forming arcs are connected
occupation from the Neolithic to the Early Bronze to them. The architectural remains seem to be orga-
Age.169 Until now, Mentesh Tepe is the only site nized around an area of stagnant water, on the
known with such a long chronological sequence in edge of which part of the skeleton of an animal to-
the whole area. For the first time in the Southern gether with obsidian debris and tools was discov-
Caucasus, this will allow a comprehensive view both ered. Bone tools seem to increase in number.171
of the evolution from the local S̆omu-S̆ulaveri culture The ceramics are tempered with either grit or vege-
and on the important breaks caused by new rela- tal matter.172 The burial of a child was discovered
tions with other areas. The only major interruption along the internal side of the wall of a house in the
concerns the first half of the 4th mill. BCE (period of intermediary level, or under the floor of the upper
the Leilatepe culture).170 level, which was rebuilt slightly over it. Three to
seven years in age, the interred child had been cov-
ered with branches of the hackberry tree, the fruit-
159
stones of which were identified over his skeleton.173
Lyonnet 2009 a.
160
Made by A. Decaix (Ph.D. Student, MNHN and University of
A jaw of sus and a bone awl were all the goods with
Paris I, together with R. Neef (DAI). which the child was buried.
161 Period II is again a very recent discovery from
Made by R. Benecke (DAI).
162
Made by L. Astruc (CNRS) and A. Samzun (Inrap) together with last year and, therefore, also lacks documentation.
B. Gratuze (CNRS).
163 As for period I, the architectural remains are circular
Made by G. Le Dosseur (post doc.).
164 Made by C. Hamon (CNRS). or slightly oval (Fig. 131). The walls seem narrower
165
Made by B. Lyonnet (CNRS). and could be made of beaten earth. The buildings
166
Made by A. Courcier (post doc.). rest upon a pebble foundation, and their diameter
167
168
Made by M. Fontugne (CEA, CNRS). varies from 2.6 m to over 5 m. Circular hearths are
Conducted by V. Ollivier (post doc.) together with M. Fontugne. associated with them. Radiocarbon dates place this
This should be completed by landscape research carried out
by A. Ricci (DAI) this summer.
169
A few sherds of the Late Bronze Age/Early Iron Age have also
171
been discovered on the site, but could not be related to any See G. Le Dosseur, this volume.
172
structure yet. But see also below, period IV, second kurgan. See B. Lyonnet, this volume.
170 Nonetheless, sites of this period, albeit very small, are known 173 Identification by A. Decaix. The tomb was excavated by our
close to Mentesh Tepe, see Maynard 2011. anthropologists, L. Pecqueur and M. Poulmarc’h.
Fig. 130
Mentesh Tepe. 14C
dates of the Neolithic
and Chalcolithic periods
(2 sigma, calibrated)
Fig. 131
Mentesh Tepe. Neoli-
thic (Period I) and Early
Chalcolithic (Period II)
architecture
(plan P. Raymond)
Fig. 132
Mentesh Tepe. End of
the Middle Chalcolithic
and Transition to
the Late Chalcolithic
(Period III) architecture
(plan P. Raymond)
period within the first half of the 5th mill. BCE. A

short gap seems to have followed it, which led to
the disappearance of its architecture through disin-
tegration and, thus, explains its poor state of pre-
servation.
Period III corresponds to the major occupation
at Mentesh Tepe. Three main architectural phases
have been identified with a total height of ca. 1.60 m
(Fig. 132). The material culture does not show ma-
jor differences. This period was severely damaged
both by the intrusions (pits and burials) of period
IV and by the modern vineyard poles and pits.
Several 14C dates place it within the 2nd half of the
5th mill. BCE.
The earliest phase is characterized by a light
architecture with post-holes and by round hearths
similar to the previous ones. No special plan is visi-
ble. This phase did not last very long.
The middle phase is by far the most impor-
tant and certainly lasted a length of time, as shown
by the fill of doors, the construction of secondary
parallel walls and buildings of slightly differing or-
ientation, and by the successive floors that have sometimes also with pebbles or fragments of querns, Fig. 133
Mentesh Tepe. Silos
been identified. All of the 14C dates place this and it is encircled by a coil of clay (Fig. 134). The dug into hearths of an
phase between 4350 and 4200 BCE. The architec- number of red-burnt layers visible in the sections earlier phase (Period III)
ture is characterized by rectangular buildings made shows that the structures were used repeatedly. (photo Mentesh
Tepe Excavations)
of one row of bricks (27–35 " 45–58 " 7–9 cm), Most of them seem to have been used for domestic
mostly yellow in color with a gray mortar (although activities, as indicated by the presence of perforated
sometimes vice-versa) coated with pisé. Some of pyramidal clay firedogs found near them. This, how- Fig. 134
the walls are well preserved (90 cm high) and stand ever, does not exclude a specialized craft, like me- Mentesh Tepe.
upon a clay foundation of ca. 10 cm in height. The talworking, as pointed out by frequent finds near Typical domestic hearth
of Period III (photo
rooms of the buildings had wooden doors as at- them.175 Besides these hearths, other combustion Mentesh Tepe
tested by the numerous limestone door-sockets of- structures of either rectangular or circular shape Excavations)
ten found in situ. Unfortunately, because of the
constraints required by the landowners, we have
been unable as of yet to gain a clear plan of the
architecture. Radar investigations are planned in
the future to solve this problem.
A great number of large jars have been found
in situ, partly buried, usually close to the walls. The
upper part of almost all of the jars was cut off dur-
ing the last stages of this period. One of the jars was
full of ochre, while a few vessels had been re-used,
that is, turned upside down and used as ovens.
A large number of clay silos were also discov-
ered in situ. They are almost cylindrical in shape
(diameter 40–60 cm) and half buried like the jars
(Fig. 133). One broken vessel was found at the bot-
tom of one silo. The analysis of the content of
these storage structures is in process.174
A large number of combustion structures have
also been unearthed. They represent several circu-
lar open hearths (diameter –60 to 100 cm), often at
a slant. Their base is usually covered with sherds, but
175
Courcier, in this article; a recent analysis of these perforated
clay objects made by M. Tonussi actually does relate them
mostly to metallurgy (communication given at the 8th ICAANE,
174
Decaix, in this article, for further discussion. Warsaw, April 2012).
their small sides. We have no clues yet about their

specific function.176
Among the activities that went on in these
buildings we can mention the presence of a large
stain of red ochre on the floor of a courtyard, and
close to it, the presence of small objects of copper
near a small heap of crushed azurite and another of
obsidian waste. Discovered on the floor of another
room was a caprid horn that had been sectioned in
half lengthwise with preliminary transversal cuts.
The faunal remains have not been studied in
detail yet, but an overview shows that the majority
is ovi-caprids.177
The latest phase was apparently short and
ended before 4000 BCE. It corresponds to a destruc-
tion phase (fallen mudbrick walls) and to an occu-
pation with light structures, as evidenced by post-
holes and the alignment of broken pots along what
could have been a wall made of perishable materi-
al. Circular open hearths similar to those of the pre-
vious phase are also present. No evidence of major
change in the material culture is visible yet.
Period IV covers a long span of time, about
Fig. 135 and with probably vaulted roofs have been discov- one mill. (Fig. 136). Through the material finds it
Mentesh Tepe.
Three pottery kilns of
ered. Three such round ovens (diameter ca. 100 cm) correlates with the Kura-Araxes culture. As this phase
Period III (photo Men- placed one against the other in sort of a circle were is at the level of the actual surface of the plain, all
tesh Tepe Excavations) found in a courtyard; they may have been pottery of the upper structures – if there were any – have
kilns, judging from the great amount of overly burnt been totally levelled for the cultivation of the vine-
sherds found in the area (Fig. 135). Three rectangu- yard; hence, what is left corresponds only to struc-
lar ovens (ca. 128–140 " 80 cm) were found against tures buried into the previous levels. This, of
the exterior of the walls of the buildings. Their base course, greatly affects our comprehension of the
Fig. 136 was covered with small pebbles coated with a thin period. Nevertheless, two main phases can be dis-
Mentesh Tepe. layer of clay; their walls are rather thin (1–3 cm) and tinguished.
14
C dates of the Early
Bronze Age period high (one was preserved up to 70 cm in height). The earliest phase corresponds to a large col-
(2 sigma, calibrated) These ovens open onto a pit full of ash on one of lective burial under a kurgan (Fig. 137). Several 14C
dates place it within the second half of the 4th and
very beginning of the 3rd mill. BCE.178 The Neolithic
and Chalcolithic mound, already unoccupied for the
last 500 years, was apparently taken as a ready-
made kurgan, within which the new population ex-
cavated a large and deep funerary chamber (5.1 "
4.3 " 2 m) with a dromos (1.2 m wide) in the mid-
dle of its eastern side; the chamber was used for
several centuries, until it was set on fire, probably
deliberately. The chamber was delimited by wooden
posts that maintained a rather thin wall (5–10 cm)
made of perishable material, of which only traces
were left (black charcoal and red clay). The sides
were precisely oriented towards the cardinal direc-
176
Analysis of the wall of one of them was made by M. Daszkie-
wicz, but our sample was not sufficient to give a clue for the
temperature reached in the oven. Further analysis is planned.
177
Benecke, in this article.
178
The earlier dates are those of the charcoal from the poles of
the walls and from a possible wooden floor. The other dates
come from charcoal found within the pottery. No date could
be obtain from the bones, which are very poorly preserved.
Fig. 137
Mentesh Tepe. Plan of
the Early Bronze Age
collective burial under
kurgan (plan
P. Raymond)
tions. No remains of the roof have been found, but face during the first year of our excavations. The in-
it must have been sufficiently strong to support a itial shape of this cap is unknown, both because of
cap of large river pebbles, which were found scat- the destruction by fire which made the roof col-
tered over a large area immediately under the sur- lapse, and because of the modern destruction of
damages to the upper layers inside the chamber.

The bones of the interred were transformed into
ashes or were very fragile, although they were pre-
served apparently in a gangue of burnt clay, from
which it was extremely difficult to extract them. A
study of the remains made by two anthropologists
is now almost finished and will be published sepa-
rately.182 Altogether, ca. 40 persons had been bur-
ied in the chamber. Some of the skeletons, or at
least parts of them, were still in anatomical order,
while clusters of long bones and skulls were discov-
ered along the walls. We can, thus, presume that
the chamber was reorganized from time to time to
give room for the next burials. Very little funerary
material was discovered within the chamber. Aside
from three animal skulls (one small bovid and two
ovi-caprids), traces of two baskets, ten perforated
bone items,183 and over 300 tiny black or white
stone beads (probably steatite), the grave goods
consist of in 21 ceramic vessels, none of which is
black burnished (see below). If – as mentioned
above – the final fire were intentional, we cannot
exclude that metal ornaments or objects that may
have accompanied the dead may have been col-
lected for re-use prior to the burning. A few meters
south of the chamber, a small symbolic tomb con-
taining the same kind of odd, naturally faceted,
stone ‘‘buried’’ in a stone cist and covered up by a
pile of small stones like a kurgan was also discov-
ered (Fig. 138,1–2).184
Fig. 138 Until now, no other contemporary structure has
Mentesh Tepe. Early been found that provides evidence that the popula-
Bronze Age small sym- tion who buried their dead at Mentesh Tepe also
bolic burial. 1 bottom
part during excavation; lived nearby.
2 the different stones The later phase is evidenced either by indivi-
found during the exca- dual burials or by pits and hearths (Fig. 139). Nu-
vations (in the back:
small pebbles from the merous 14C dates situate them between ca. 2800 and
cap, in the middle: the 2400 BCE (see Fig. 7). They are scattered through-
buried stone and out all of the excavated area, and no association
around it and in the
front the cist stones) can be made between them stratigraphically except
(photo Mentesh Tepe through their ceramic content. The pits are extremely
Excavations) numerous, often very large (over 2 " 2 m), pear-
shaped, and some of them even reaching the Neo-
the mound. I. Narimanov does not mention a possi- lithic levels more than two meters farther down.
ble kurgan or pebbles or even Kura-Araxes sherds, Their function is not yet clear, although most may
so probably no remains were visible at the time of have been storage pits. Analysis of their botanical
his visit. Two special long stones,179 probably of or faunal contents is in progress. They often con-
symbolical meaning,180 were found, one set at the tain almost complete pots broken into large pieces;
entrance of the dromos into the chamber, the other in one pit a set of large querns as well as a small
on the opposite side.181 The fire caused very severe animal clay figurine were discovered. Besides these
179 The stones are naturally facetted (volcanic stone) and can be
182
found at ca. 10–15 km from the site along the Esrik Çayi. L. Pecqueur (Inrap) and M. Poul’march (Ph.D. student, Univer-
180 sity of Lyon II).
A similar case of ‘‘special’’ stones set at the entrance of the
183
chamber (reportedly semi-anthropomorphic) is mentioned at Usually considered as spindle whorls, but they could also be
Horom, Armenia, also in a (smaller) collective burial (Badaljan buttons.
et al. 1993). 184 Besides these stones, the tomb was totally devoid of other
181
This one was no longer in situ. finds.
Fig. 139
Mentesh Tepe. Early
Bronze Age (Period IV)
pits and burials (plan
P. Raymond)
domestic pits, two individual pit burials were found: areas like Kültepe in Naxiçevan, or Chalcolithic sites
one contained a complete skeleton, while the other like Sioni itself for instance, also possess such
is a secondary burial with only parts of the skele- round structures. Although we have no architectural
ton. The openings of both burials were covered by remains for the Kura-Araxes period on our site, dif-
an oval heap of large river pebbles. ferent excavations have unearthed circular structures
We should also mention here a second kur- for this period as well, but they are usually found
gan, located about 30 m away from the first one together with rectangular ones.186 Thus, circular ar-
and covering an area of about 7 " 7 m. This second chitecture seems to have been a typical trend of
kurgan does not seem as damaged as the other, the local culture in most of this Southern Caucasus
and no traces of fire are visible. At present, its cap area. It is, therefore, particularly interesting to sud-
of river pebbles has only been cleaned from the denly find a rectangular architecture during the sec-
earth covering it, and will be opened during next ond half of the 5th mill. No other site of the so-
excavation season; therefore, we do not know its called Sioni culture187 ever produced such plans,
date yet. Nevertheless, it clearly cuts into the Chal- but a similar and probably contemporary change has
colithic levels, and we can suppose that it is later been observed at Alikemek, Çalağan and Ilanlı Tepe,
than the first kurgan. It may be contemporary with all situated close to the Mil Steppe area.188 Rectan-
the Kura-Araxes pits and individual burials, many of gular architecture is also typical of the later settle-
which are situated not far from it. Conversely, it ment of Leilatepe,189 while it is found associated
could also date to a much later period, like the end with round or oval structures at Boyuk Kesik.190 As
of the Bronze Age/beginning of the Iron Age, as in- the study of the ceramic material of Mentesh Tepe
dicated by a few sherds found nearby. shows,191 possible relations with the eastern part
In addition to these pit structures, several of the Kura River Valley may have been established
small hearths belong to the Kura-Araxes period, and at the end of the Neolithic period, from where this
one of them shows the negative imprints of a three new trend in architecture could have come (see
partite, possibly portable, andiron, which curiously Helwing, this volume). However, the period during
are otherwise not attested at the site. which this rectangular architecture appears at Men-
Besides these structures, no architectural re- tesh Tepe, dates several centuries later and corre-
mains can be related with certainty to this period, sponds to the transitional phase from the Middle
either because of the modern destruction to the Chalcolithic to the very beginning of the Late Chal-
site, or because the ancient populations only lived colithic period, while at Leilatepe it dates to the
there seasonally in light or semi-subterranean con- LC2-3 periods in the chronology of northern Meso-
structions. Further studies on the botanical and fau- potamia. Indeed, these are two phases, during which
nal remains should help solve this question. many influences from this last area have been no-
ticed. We are, thus, inclined to think that this sud-
Conclusions den change should rather be connected with this
phenomenon. Our excavations could not extend as
It is exceptional, in the Southern Caucasus, to find far as expected, but we hope to make geophysical
a site like Mentesh Tepe, where successive occupa- investigations in order to determine the plan of our
tions developing along several millennia can be building. We would not be surprised if it turned out
observed and which, therefore, allows a better un- to be a tripartite-like building.
derstanding of the local evolution and of the inno- Besides rectangular architecture, proof of light
vations. constructions based on post-holes have also been
Our excavations clearly display elements of a noticed prior to and after this Chalcolithic III period
great continuity. This is the case for circular hearths and may be a testimony of some kind of mobility.
paved with stones or sherds and surrounded by a Other and different features pertaining to the Early
coil of clay, as well as earthen circular silos so fre- Bronze Age also seem to reveal a similar way of
quent in our Period III, all of which are already at- life. The collective burial in the kurgan is not asso-
tested in the S̆omu-S̆ulaveri culture.185 This is also ciated with any domestic structure, and settlements
clearly visible in the circular architecture, which con- with similar ceramic material were neither found
tinues for more than one mill., from the Neolithic
(period I, S̆omu-S̆ulaveri culture in the first half of 186 See Kushnareva 1993, 51–71.
the 6th mill.) to the Chalcolithic (period II which 187 For references, see footnote 18.
ends up a little before the middle of the 5th mill.). 188 Chataigner 1995, 61; 63. The absence of circular buildings in
We know that contemporary Neolithic sites in other the actual state of research at the Kamiltepe sites may thus only
be provisional.
189
See Akhundov 2007.
190
Museibli 2007.
185 191
Guliyev/Nishiaki 2012. Lyonnet, in this article.
during our survey nor during the salvage excava-

tions along the BTC pipeline. Mentesh Tepe was
then possibly the burial place of a transhumant
group that returned seasonally to the same place.
Due to the destruction of the upper part of the
mound, it cannot be determined whether or not
there were architectural features that are associated
with the numerous pits dating to the later part of
the Early Bronze Age that have been found. Never-
theless, the large size of the sherds found in the
pits and the frequent reconstruction of almost com-
plete vessels also points in the direction of a sea- Fig. 140
sonal occupation with repetitive discard of the pot- Mentesh Tepe. Neo-
lithic pottery of the first
tery vessels on each departure. The hypothesis of group (S̆omu-S̆ulaveri;
mobility advanced here on the basis of the architec- photo Mentesh Tepe
tural and ceramic remains will be investigated Excavations)
further both through isotopic analysis on animal

and human bones in order to determine their diet pers can be correlated with the S̆omu-S̆ulaveri cul-
and possible transhumances, and through detailed ture of the first half of the 6th mill. BCE.
analysis of the environmental conditions in the val- The second group is also handmade and po-
ley during these periods. lished, but it is mainly characterized by its temper,
which left tiny oval cavities visible on the surface of
the sherds (remains of burnt seeds)193 (Fig. 141).
Mentesh Tepe Pottery The shapes seem more varied (including a low
walled, open shape) and complex (partitioned ves-
Bertille Lyonnet sel), and a few sherds are painted black.194 This
group, for the moment, has no parallels known else-
Pottery is extremely abundant at the site. In four sea- where in South Caucasia. We are inclined to think
sons of excavations, ca. 38,000 sherds were found, that it should date to a later phase of the Neolithic,
mostly dating to the Chalcolithic period. We will after the end of the S̆omu-S̆ulaveri culture. At pre-
briefly describe here the different assemblages from sent, we have no structure associated with it and
periods I to IV. The conclusion will underline the therefore no precise 14C date.
features of continuity and the major innovations. Period II pottery is also sparsely represented,
Period I pottery is still scanty (a little over 50 since few structures of this period, dated to ca.
sherds), as this period only appeared at the end of 4800–4600 BCE, were detected only last year. The
last year and was sealed by a level of compact yel- pottery is handmade with vegetal temper, some-
low clay, which prevented sherds from appearing in
the upper levels. Two main groups can be distin-
guished:
The first group is handmade, apparently using
the technique of large coils (ca. 8 cm wide) or slabs
(Fig. 140). At present, the shapes determined are
very limited and consist mainly of hole-mouth ves-
sels. All of the bases are flat, but slightly raised.192
One sherd presents a round pellet applied under
the rim. Most sherds are carefully polished outside
and often also inside. Their color is irregular, but
usually in light tones (from pink to gray or yellow)
and their section is often light gray. A few sherds Fig. 141
present a red slip all over. Two major kinds of tem- Mentesh Tepe. Neo-
per are attested within this group. A minority is lithic pottery of the
second group with its
grit-tempered, but the majority is plant-tempered, typical vegetal temper
displaying long and thin cavities (remains of burnt (photo Mentesh Tepe
chopped plants). This pottery group with both tem- Excavations)
192 None of those discovered up until now shows basket impres- 193
Seeds of poaceae according to A. Decaix.
194
sions, frequent in the S̆omu-S̆ulaveri culture. The sherds are too small to determine any motif.
Three major groups have been recognized:

common ware, tableware and kitchen ware. The
pottery in each group is handmade, using large
bands about 8 cm wide, the joins of which are ea-
sily visible at the fracture. Almost all of the vessels
were smoothed out with a comb on the outer sur-
face, and sometimes inside the neck.198 The temper
is mostly vegetal, and its size varies according to
the function and dimensions of the vessels.199 Only
cooking ware is mainly tempered with obsidian
and/or minerals, although vegetal is not excluded
either. The firing is irregular, so that the colors of
the pots vary from yellow to red and brown, and
their section is always black. Many of the small
vases, though, are often completely gray or even
black,200 and kitchen ware is mostly dark brown
and sooted. No specific slip has been identified,
Fig. 142 but burnishing is extremely frequent; it usually con-
Mentesh Tepe. Early sists of more or less horizontal lines close to one
Chalcolithic mangal. – another. The shapes are simple, the rims either
1 exterior; 2 interior
(photo Mentesh Tepe rounded or flattened, but never thickened, and no
Excavations) carination is attested. The connection between the
neck and the shoulder is curved, the bottoms flat
or flat-round.
times including grit. It is clearly already linked Common ware includes large jars (Fig. 143),
rather with the later Chalcolithic material than with small jars and jugs (Fig. 144,4–12) and large basins
the Neolithic. Traces of combing to smooth the out- (Fig. 145). The necks of the first two groups (mostly
side surface appear during this period, although between 3.5 and 6 cm in height) are either vertical
they do not seem systematic nor frequent yet. or slightly everted. Many of the jars and small jars
Some of the major shapes that develop during peri- have an applied decoration at the junction with the
od III are already present here, but with specific dif- shoulder: either pellets arranged in groups (from 2
ferences. This is the case for ‘‘pans’’, a typical thick to 6) or a band, often decorated with finger impres-
disc-shape (35–40 cm wide, 2–3 cm thick) with a sions (Fig. 143,1–3). The same pellet decoration,
very low wall or thickened rim, which have here a sometimes associated with painted decoration with
rather rough greenish outer surface and only bitumen, is attested on the basins. Other large and
spaced thumb-prints along the rim for decora- small jars are without a neck and with a slightly in-
tion.195 A vessel such as this is attested at Alike- verted rim decorated with applied long pellets.201
mek.196 Similarly, ‘‘mangals’’ with perforations un- Tableware mostly includes open vases of
der the rim are attested, but they have semi-circular small dimensions like plates, bowls and beakers202
lugs decorated with a small notch in the middle ap- (Fig. 146), as well as small jugs or juglets with ver-
plied horizontally along the rim and do not show tical neck (Fig. 144,1–3). The bottoms of the bowls
comb-traces on the outside (Fig. 142,1–2). The
same shape of mangal is known at Kviriatskali in
northeastern Georgia.197
198 Although this is not initially a feature of decoration but of
Period III pottery constitutes over 80% of the
smoothing out the joints of the coils or slabs, some of the
ceramic finds. At this stage the preliminary study pots that we have excavated had been so carefully combed
does not allow the determination of specific changes on the outside that the desire for an aesthetic result is beyond
between the three major architectural phases of any doubt.
199
This does not exclude the eventual presence of chamotte or
this period, all dated within the second half of the of occasional minerals observed at a macroscopic level.
5th mill. BCE. 200
It is probably not intentional and could be due to the place-
ment of the small vessels inside larger ones in the kilns so
that they were not oxidized.
201
Applied pellets along the rim are known in the S̆omu-S̆ulaveri
195
It is not yet clear whether those that we found could have culture, see Bastert-Lamprichs, this volume.
even been used on a fire. 202
There is no real standardization of the shapes, even if there
196
I wish to thank T. Akhundov who showed me some of the is a tendency towards it, and it is not always possible to dis-
ceramic material from Alikemek at the Institute of Archaeolo- tinguish with certainty groups of plates from those of bowls
gy. or of beakers, but most of these small open vessels are rather
197
Varazashvili 1992. deep.
Fig. 143
and Transition to the
Late Chalcolithic.
1–5 Common Ware,
jars (drawings B. Lyon-
net, PAO J. Lhuillier)
Fig. 144
Late Chalcolithic.
1–3 Tableware, juglets;
4–12 Common Ware,
small jars and jugs
(drawings B. Lyonnet,
PAO J. Lhuillier)
and goblets are often flat-round with an omphalos.

Because the tableware is more carefully made, the
comb-traces are often less visible, especially since
it is mostly burnished or even polished on both
faces. Most of the vessels in this group are painted
with bitumen (inside for opened shapes, outside for
closed shapes). The thickness of the bitumen,
which had to be heated for its use, did not allow a
careful design. Decoration mostly consists of chev-
rons203 or vertical bands crossing at the bottom
and sometimes associated with dots (for opened
shapes), or in a horizontal band at the base of the
neck or large dots on the shoulder (for closed
shapes).204 The rim is generally always painted with
a band, in some cases very wide.205 A few motifs
of waves, semi-circles, or stylized human figures
with raised hands206 have also been noticed. Some
of the tableware combines applied pellets together
with the painted decoration. Finally, incised decora-
tion is also attested, either chevrons on the out-
side, or straight lines sometimes making a motif in-
side bowls (Fig. 146,4–5).207 One unique sherd
presents three very fine incisions together with ap-
plied round buttons (Fig. 147).208
Kitchen Ware includes closed vessels, pans
and mangals. Numerous large necked-pots with
curved body and flat bottoms have been discov-
ered. Their rims are generally decorated with inci-
sions or comb-impressions (Fig. 148,3.5).209 Similar
pots of much smaller dimensions are also attested
(Fig. 148,4.6). Large circular pans in the shape of a
thick disc, entirely polished and often covered with
bitumen on the top but roughly finished on the
bottom part, are also an important component of
203 The decoration of painted chevrons is frequent at Kültepe, Ali-

kemek (Munchaev 1982, pl. 44–47), Dalma (Hamlin 1975), and
at sites in northwestern Iran (Fazeli et al. 2005). the kitchen ware of this period. Their thickened Fig. 145
204 Dots are known in northern Mesopotamia (cp., for instance, rim is generally decorated with incisions or comb- the Middle Chalcolithic
Akkermans 1988, pl. 76,113.115.117.122) and have recently impressions (Fig. 148,1–2). Both the pots and the and Transition to the
also been discovered in a cave at Areni I, Armenia (Wilkinson
et al. 2012).
pans are mainly tempered with obsidian and/or grit. Late Chalcolithic:
Common Ware, basins
205 Bands, whether large or not, and often associated with wa- A great number of hole-mouth pots are mostly red (drawings B. Lyonnet,
ves, are well known at the end of the Ubaid period (cp., for or black-burnt and sooted and can also be included PAO J. Lhuillier)
instance, Matsutani 1991, pl. 86). with kitchen ware, although they have mainly a ve-
206
This type of decoration is already known in the S̆omu-S̆ulaveri
culture with applied figures (Chataigner 1995, 104 pl. 45), as getal temper and, therefore, were probably not
well as in the Tripol’e-Cucuteni culture (Chernykh 1982, pl. 41). used directly with fire (Fig. 148,8–12). Finally, man-
207
Such incisions inside bowls are well known in northern Meso- gals with perforations under the rim are extremely
potamia during the transitional period from the Ubaid to the
Late Chalcolithic, see, for instance, R. Matthews 2003, fig. 3.13,
frequent (Fig. 148,7). The mangals of period III have
11–13. only slightly raised and rounded lugs and are al-
208 ways combed on the outside (Fig. 149,1–2). Man-
The sherd can be compared with Gawra incised and impres-
sed ware in period XI (Tobler 1950, pl. 79,a) and periods XI– gals obviously were not used for cooking directly
IX (Tobler 1950, pl. 152,514) and also known in other places
in northern Mesopotamia, for instance at Brak (R. Matthews over a fire, since their wide bottom, even though
2003, fig. 3.10; 3.15,7) though at Mentesh Tepe the buttons often tempered with obsidian, chamotte or grit, is
are not stamped. extremely thin and fragile, but they are always
209
This is considered as the hallmark shape of the Sioni culture
(for references see footnote 18); it is still known during the
sooted on the outside and probably stood close to
Leilatepe culture (Akhundov 2007, fig. 16). the hearths.
Fig. 146
Late Chalcolithic: Table-
ware (drawings B. Lyon-
net, PAO J. Lhuillier)
Period IV pottery only represents about 10%

of the total number, but the sherds are often very
large, and almost complete pots can be restored.
This pottery is handmade, also using the large band
system. No evidence of textile impressions inside
the pots or between layers of slabs has been iden-
tified. The temper is not easily recognizable, but Fig. 147
may contain small chamotte, organic, mineral and the Middle Chalcolithic
even some vegetal inclusions. The exterior is entirely and Transition to the
polished. Two main groups can be distinguished: Late Chalcolithic: table-
ware with fine incisions
early and late. and applied buttons
The early group comes exclusively from the (photo Mentesh Tepe
collective burial under the kurgan,210 dated to the Excavations)
second half of the 4th and very beginning of the 3rd
mill. BCE (Figs. 150,1–3; 151,1–2). It has red-brown-
ish, but no real black colors, although we cannot
exclude that the heat changed the initial appear- base (Fig. 152,2.4). Pots are often decorated with
ance of the pots. The polishing on the outside is large grooves forming a zigzag on the shoulder,
rather faint. Almost all of the rims are simple, round sometimes ladder-like (Figs. 152,1; 154,2). Other
and slightly everted. The bottoms are flat, slightly pots have the same kind of zigzag decoration made
concave. Many of the pots are closed containers less carefully by incisions with a pointed tool
with one or two opposite handles usually from the (Fig. 154,1). Many of the small pots show a large
rim to the shoulder. Their general shape is curved ‘‘dimple’’ on the side (Fig. 152,3), the function of
with rare cases of prominent shoulder. One jug pre- which is not clear. Large opened bowls are not very
sents a rather high vertical neck and one small han- frequent and have a carination under the rim with
dle at the junction with the shoulder. Only three two opposite horizontal pierced lugs (Fig. 154,3).212
vases have open shapes, one with handles and the Besides this common ware and tableware, a kitchen
others without.211 ware is also frequently attested. The shapes are al-
The late group comes from the pits and from most uniform, with a slightly thickened rim, cari-
one of the individual burials, all dated between ca. nated shoulder and small flat bottom (Fig. 154,4).
2800 and 2400 BCE (Figs. 152–154). Most of the The carination is often underlined by finger impres-
pots are black-burnished outside, sometimes very sions (Fig. 154,5), or sometimes by a lug, and the
shiny or with regular and close strokes of burnish- body under the carination frequently covered with a
ing making a decoration. Exceptional cases are red, rough layer of added clay (‘‘wet ware’’).
red and black, or yellowish. The inside is dark gray
and rarely burnished, except for the open shapes, Conclusions
which are not frequent. Slips are attested, usually
black-gray, but sometimes becoming red inside and As already mentioned, Mentesh Tepe is an excep-
black outside (Fig. 153,1–2); however, we cannot tional settlement, the long duration of which helps
state that real red-black-burnished (RBBW) ware is to point out both the traditional features and the
attested. The section of the pots often shows a innovations. Among the ceramics, some shapes
typical bichrome effect. Most of the shapes are (hole-mouth jars) or decorations (applied pellets)
carinated and with a prominent shoulder. The rims as well as the techniques of fabrication (with large
are simple, sometimes thickened but with no typi- bands) or the use of vegetal temper213 appear al-
cal rail-rim attested. Handles similar to those of
the earlier group, and various shapes of lugs are 212
We will not go into detailed comparisons for this material ei-
frequent: Nakhichevan- (Fig. 153,1), pierced hori- ther, but general comparisons can be made with similar deco-
zontal (Fig. 154,3), and not pierced vertical lugs rations at Arslantepe VIB terminal and VIC (Palumbi 2008,
(Fig. 152,3–4), as well as one case of triangular lug fig. 6.27), and similar shapes at Karnut I, Armenia (Badalyan/
Avetisyan 2007, 138–147).
on a rim. Bottoms are small, flat of flat-concave. 213
The use of vegetal temper (without or not mineral temper) is
The shoulder is often underlined by a groove at its known since the Neolithic in the Southern Caucasus, both in
the S̆omu-S̆ulaveri culture (Kiguradze 1976; Narimanov 1987,
117 where it is said that it represents 15–20% of the total
210 Not all the pots have been restored yet and other shapes
amount of the sherds; Hansen et al. 2006), and in the Kamil-
may be found later. tepe area (Aliyev/Helwing 2009). It is also known in Northern
211
Without going into detailed comparisons for this material, we Mesopotamia (Amuq A, see Balossi-Restelli 2006, 114–115;
can just state that most of it is comparable to early Kura-Ara- Proto-Hassuna, see Le Mière 2000, 127; Halaf, see Nieuwen-
xes sites or levels as shown in Palumbi 2008 (Samshvilde, huyse 2000, 158), in the Levant and Jordan (Balossi-Restelli
Berikldeebi IV, Treli, Didube, Arslantepe VIA and VIB1, etc.). 2006, Tab. 1.1.) and in Eastern Anatolia (Mezraa Teleilat, Ba-
Fig. 148
Late Chalcolithic:
Kitchen Ware
(drawings B. Lyonnet,
PAO J. Lhuillier)
ready during our period I (first half of the 6th mill.

BCE) and are still present in the second half of the
5th mill. BCE.
The perpetuation of these features clearly
shows continuity in the development of the local
culture.
Other features, nevertheless, also point out
breaks and innovations. The first ones probably
date to the end of the S̆omu-S̆ulaveri culture, at the
end of the 6th mill. BCE, and could come from the
East (Mil and Muğan Steppe), as the introduction of
painted ceramics seems to indicate.214 Since our
documentation on this phase is still very scanty, we
can only consider this as a hypothesis. Another re-
markable break dates to the time of the change in
architecture from circular to rectangular structures.
Pottery of this period (III) is often painted, and its
relations with the East are very clear, like the dec-
oration in chevrons that has many parallels at Kül-
tepe (Nakhichevan), Alikemek or at Dalma and
other northwest Iranian sites.215 Yet, the influence
from northern Mesopotamia is also undeniable, like
incised geometric lines inside bowls,216 incised and
applied motifs close to those of Gawra XI,217 painted
bands, dots or waves.218
This period actually corresponds to a time of
profound changes in all the Ancient Near East, a
time when the Ubaid culture disappeared and was
replaced by what some scholars have called ‘‘North-
ern Early Uruk’’ and others ‘‘Late Chalcolithic 1’’. It
is during this period too that the ‘‘Sioni’’ assem-
blage (use of the comb, incised rims, mangals)219
spread over a large area, extending from the high-
lands of eastern Anatolia220 to the northern Cauca-
Fig. 149
sus.221 Although Mentesh Tepe shares many com- Mentesh Tepe. End of
mon features with sites associated with the Sioni the Middle Chalcolithic
culture as far as material culture is concerned, its and Transition to the
Late Chalcolithic:
exceptional rectangular architecture as well as the mangal; – 1 exterior;
abundance of painted ware lead us to consider that 2 interior (photo
this settlement belongs to a different sphere, at this Mentesh Tepe
Excavations)
point unique.
lossi-Restelli 2006, 262). It is therefore surprising to read that

chaff-faced ceramics could have its origins in Southern Cauca-
The innovations linked to northern Mesopota-
sus during the Chalcolithic period (Marro 2010). mian influence that have been underlined during
214
There are very few cases of painted ware in the S̆omu-S̆ulaveri Mentesh period III are only the prolegomenae of a
culture (Munchaev 1982, 107–115; Guliyev et al. 2009), while phenomenon that started shortly after the begin-
they are frequent at Kamiltepe (Aliyev/Helwing 2009).
215 See footnote 203. ning of the 4th mill. BCE, when relations between
216
See footnote 207. the two areas222 developed even more and reached
217
See footnote 208. the northern Caucasus, where the Majkop culture
218
For dot decoration see footnote 202; for bands see footnote started.223 The end of this period, around the middle
203.
219 Some of the most typical shapes of this assemblage like man-
gals and pans appear earlier (period II of Mentesh Tepe), but
the use of comb does not seem frequent then, although we 222
It is at that time that sites like Leilatepe, Boyuk Kesik, Alkhan
still lack data to be affirmative. Tepe or Berikldeebi were created. A new contemporaneous
220
Marro 2007. settlement has just been recently discovered near Gabala in
221 Nechitajlo 2007 (mangals); Korenevskij 1998 (mangals and in- Azerbaijan, at the foot of the Great Caucasus.
223
cised rims). Lyonnet 2000; Lyonnet 2007 b; Lyonnet 2007 c.
Fig. 150
Mentesh Tepe. Early
Bronze Age pottery
from the collective
burial under kurgan.
1–4 jars and juglets
(photo Mentesh
Tepe Excavations)
Fig. 151
Mentesh Tepe. Early
Bronze Age pottery
from the collective
burial under kurgan.
1–2 bowls (photo Men-
tesh Tepe Excavations)
Fig. 152
Mentesh Tepe. Early
Bronze Age pottery
from pits and individual
burials (photo Mentesh
Tepe Excavations)
of the 4th mill. BCE, is marked by the development the transformations seen both in northern Mesopo-
of the Kura-Araxes culture, the origins of which are tamia and in the Caucasus. Mentesh Tepe certainly
still debated.224 It may not be coincidental, then, if demonstrates that local metallurgy was develop-
this culture ultimately spread in the other direction, ing.226 Nevertheless, we need to know much more
that is, into northern Mesopotamia and the Levant. about the local cultures in both areas in order to
The reasons behind the profound changes better understand the circulation of goods, ideas
that happened during the Late Chalcolithic period and people that took place between these different
before the Late Uruk expansion are still unclear.225 neighboring areas.
The search for raw material cannot alone explain
224
Marro et al. 2011 for a very early date.
225 226
Lyonnet 2007c. Courcier, in this article.
Fig. 153
Mentesh Tepe. Early Bronze Age pottery from pits and individual burials (photo Mentesh Tepe Excavations)
Fig. 154
Mentesh Tepe. Early Bronze Age pottery from pits and individual burials (photo Mentesh Tepe Excavations)
The metallurgical evidence at Mentesh Tepe: levels and their date is therefore not secure: Kura-
preliminary results of archaeometallurgical Araxes or Chalcolithic.
analyses Awls were discovered in almost all the sectors
of the site,229 but the spatial distribution of the
Antoine Courcier other metallurgical artifacts is not so homogenous.
Nevertheless, we have been able to distinguish four
Four seasons of excavations at Mentesh Tepe227 areas with a concentration of metallurgical artifacts
have unearthed abundant metallurgical evidence. (wastes, metallurgical remains, metal objects) and
Most derives from the Chalcolithic levels and dates which cover contiguous sectors.
to the second half of the 5th mill. BCE, some is re-
lated to the Kura-Araxes culture (second half of the Spatial distribution
4th and first half of the 3rd mill. BCE). The number
of metallurgical finds is exceptional. It is the first A first concentration area with 10 artifacts has been
time indeed that so many items, including objects, identified in zone 1, which corresponds to a large
ores and waste, have been discovered in a settle- burial under kurgan dated to the early phase of the
ment of such an ancient date in the Caucasus. In Kura-Araxes period. Unfortunately, the chamber was
order to better understand and characterize the deeply dug into the Chalcolithic levels and totally
metallurgy at Mentesh Tepe, we have conducted damaged them, so that we have no secure date for
archaeometallurgical studies with the laboratory of this material.
the Deutsches Bergbaumuseum at Bochum, under A second concentration area with 20 metallur-
the direction of Andreas Hauptmann. gical artifacts has been localized in the Chalcolithic
The aim of this short article is to present briefly levels of three sectors (zones 10, 12 and 13), near
the discoveries and the results of this archaeome- walls or close to hearths or ovens.230
tallurgical research. A third group with 16 metallurgical artifacts
comes mainly from the Chalcolithic levels231 of
The metallurgical evidence three other sectors (zones 11, 14 and 16), close to
hearths/ovens and to a Chalcolithic jar.
Altogether, 92 artifacts dealing with metallurgy have The fourth and most important concentration
been found at Mentesh Tepe (Tab. 20). Their diver- area with 32 metallurgical artifacts is situated in the
sity underlines the importance of metallurgical ac- Chalcolithic levels of six sectors (zones 18, 20, 21–
tivities at this settlement. Most of the metallurgical 24). The artifacts were found near postholes,
‘‘chaı̂ne opératoire’’ has been unearthed: minerals, hearths/ovens and a silo in dwelling contexts or
host-rocks, ores, ceramic finds related to metal- craft areas. In zone 18, close to a second large bur-
lurgy, slag, remains of metal production (oxide frag- ial under a kurgan not yet excavated, two Kura-Ara-
ments, unidentified artifacts) and metal objects xes ornaments were also unearthed.
(awls, ornaments, unidentified fragments) (Tabs. 20; Finally, 20 other metallurgical artifacts have
21). been discovered individually in the other excavated
sectors (zones 2, 3, 5–9, 15, 17 and 19). Among
Relative dating them, seven come from secure Chalcolithic levels,
close to hearths/ovens.
Most of the artifacts come from the Chalcolithic lev- The 30 other artifacts come from the upper or
els, especially from levels dating to the second half disturbed levels, and their date is not secure.
of the 5th mill. BCE (Tab. 21). However, some also
come from disturbed levels or areas228 and may
date to the Early Bronze Age (Kura-Araxes Culture). Archaeometallurgical analyses
In three sectors (zones 8, 16 and 18), for instance, Research protocol
ornaments typical of this period have been discov-
ered (see below). However, in some cases we have Studies on this material started in 2009. They are
not been able to date them, because they were made by myself, under the direction of Prof. Dr. A.
found in deep Kura-Araxes pits dug into Chalcolithic Hauptmann and Dr. M. Prange, in the laboratory of
227 229 Zones 1, 2, 6, 7, 8, 10, 11, 13, 14, 15, 16, 18, 19, 20, 22 and 24.
For the excavations on this site, see Lyonnet et al., in this ar-
ticle. 230 Their exact function is still not determined, but, anyhow, they
228 are not exclusively associated with metallurgical activities.
The date of the metal artifacts found in the upper levels
231
which were excavated in the first two campaigns (2008, 2009) In these zones, nine objects were found in Kura-Araxes pits.
is not secure due to the destruction of the mound. On the Only one (a spiral ornament) is clearly related to the Kura-Ara-
contrary, metal artifacts coming from two last campaigns xes culture. The date of the others is not secure: Chalcolithic
(2010, 2011) are mainly dated to the Chalcolithic period. or Kura-Araxes.
Inventory no. Description Dating Inventory no. Description Dating

MET-01-2008 Slag Unsecure MET-27-2010 Awl Chalcolithic
MET-02-2008 Ore Unsecure MET-28-2010 Ore Chalcolithic
MET-03-2008 Awl Unsecure MET-29-2010 Unidentified object Chalcolithic or EBA
MET-04-2008 Curved awl Unsecure MET-30-2010 Unidentified object Chalcolithic
MET-05-2008 Semi-manufactured awl EBA MET-31-2010 Unidentified artifact Chalcolithic
MET-06-2008 Awl fragment Unsecure MET-32-2010 Curved awl Chalcolithic
MET-07-2008 Unidentified artifact Unsecure MET-33-2010 Awl fragment Chalcolithic
MET-08-2008 Unidentified artifact Unsecure MET-34-2010 Unidentified object Chalcolithic
MET-09-2008 Ingot mould Unknown MET-35-2010 Awl fragment Chalcolithic
MET-10-2008 Awl Unsecure MET-36-2010 Ore Chalcolithic
MET-11-2008 Unidentified artifact Unsecure MET-01-2011 Awl fragment Chalcolithic
MET-01-2009 Awl fragment Unsecure MET-02-2011 Curved awl Chalcolithic
MET-03-2009 Pendant EBA MET-03-2011 Ring fragment Chalcolithic
MET-04-2009 Awl Unknown MET-04-2011 Awl Chalcolithic
MET-05-2009 Small ring Unsecure MET-05-2011 Unidentified object Chalcolithic
MET-06-2009 Awl fragment Unsecure MET-06-2011 Mineral Chalcolithic
MET-07-2009 Awl fragment Chalcolithic or EBA MET-07-2011 Awl fragment Chalcolithic
MET-08-2009 Curved awl fragment Unsecure MET-08-2011 Host rock EBA
MET-09-2009 Awl Unknown MET-09-2011 Awl fragment Chalcolithic
MET-11-2009 Crucible Chalcolithic MET-10-2011 Awl Chalcolithic
MET-01-2010 Unidentified artifact Chalcolithic or EBA MET-11-2011 Awl fragment Chalcolithic
MET-03-2010 Unidentified object Chalcolithic MET-13-2011 Awl Chalcolithic
MET-04-2010 Awl fragment Chalcolithic MET-14-2011 Curved awl Chalcolithic
MET-05-2010 Unidentified artifact Chalcolithic MET-15-2011 Awl fragment Chalcolithic
MET-06-2010 Awl fragment Chalcolithic MET-16-2011 Ore Chalcolithic
MET-07-2010 Unidentified artifact Chalcolithic or EBA MET-17-2011 Awl fragment Chalcolithic
MET-09-2010 Unidentified artifact Chalcolithic or EBA MET-19-2011 Awl fragment Chalcolithic or EBA
MET-10-2010 Awl Chalcolithic MET-20-2011 Awl fragment Chalcolithic or EBA
MET-11-2010 Unidentified object Chalcolithic MET-21-2011 Unidentified object Chalcolithic or EBA
MET-12-2010 Unidentified object Chalcolithic MET-22-2011 Awl fragment EBA
MET-13-2010 Oxide fragment Chalcolithic MET-23-2011 Awl fragment Chalcolithic
MET-14-2010 Awl Chalcolithic or EBA MET-24-2011 Awl fragment Chalcolithic or EBA
MET-15-2010 Oxide fragment Chalcolithic MET-25-2011 Unidentified artifact Chalcolithic
MET-16-2010 Unidentified object Chalcolithic MET-26-2011 Awl fragment Chalcolithic
MET-17-2010 Oxide fragment Chalcolithic MET-27-2011 Awl fragment Chalcolithic
MET-18-2010 Awl fragment Chalcolithic MET-28-2011 Spiralled ornement EBA
MET-19-2010 Unidentified artifact Chalcolithic MET-29-2011 Curved awl Chalcolithic or EBA
MET-20-2010 Awl fragment Chalcolithic MET-30-2011 Spiralled ornement EBA
MET-21-2010 Mould Chalcolithic MET-31-2011 Awl fragment Chalcolithic
MET-22-2010 Awl Chalcolithic or EBA MET-32-2011 Awl fragment Chalcolithic
MET-23-2010 Curved awl Chalcolithic MET-33-2011 Spiralled ornament EBA
Tab. 20 MET-24-2010 Oxide fragment Chalcolithic MET-34-2011 Curved awl Chalcolithic
Mentesh Tepe.
List of the metallurgical MET-25-2010 Oxide fragment Chalcolithic MET-35-2011 Unidentified object Chalcolithic
artifacts (seasons MET-26-2010 Oxide fragment Chalcolithic MET-36-2011 Unidentified artifact Chalcolithic
2008–2011)
Type Dating Total
Chalcolithic Chalcolithic or EBA EBA Unsecure date
Mineral 1 1
Host-rock 1 1
Ore 3 1 4
Refractory ceramic 2 1 3
Slag 1 1
Unidentified artifact 5 3 3 11
Oxide fragment 6 6
Unidentified object 8 2 10
Awl (semi-manufactured) 1 1
Awl (fragment) 20 10 1 31
Awl (curved) 5 2 7
Awl (complete) 4 4 2 10
Ring 1 1 2
Spiraled ornament 3 3 Tab. 21

Mentesh Tepe. Relative
Pendant 1 1 dates of the metallurgi-
cal evidence, organized
Total 55 21 8 8 92 according to types and
numbers
the Deutsches Bergbau Museum in Bochum. Lead Host-rock, mineral and ores
isotopic analyses have been carried out by Dr. S.
Klein at the Institute of Geosciences in the Goethe An ore fragment was discovered during the first
University of Frankfurt. campaign in a level dated to the Kura-Araxes peri-
Different kinds of analyses have been od. During the last campaign, a fragment of host-
achieved: metallographies, chemical analyses (XRD, rock, characterized by a very thin copper impregna-
EDS, ICP-MS) and isotopic analyses.232 tion, was discovered in a level also dated to this
period. Neither of them has been analysed yet.
Results A rather important quantity of green-brown
mineral (45 g) was discovered in a Chalcolithic pit.233
Slag and copper oxide fragments
It is a mineral composed of muscovite-calcite-albite
In 2008, immediately under the surface of zone 1, a with quartz and probably heulantite. This mineral
small and light slag was collected. Analyses are could be used as a flux for smelting process, or as
programmed in the archaeometallurgical laboratory temper for the crucible or mould (see below).
of Baku, where this slag is kept. Three fragments of ores come from Chalco-
During the third campaign of excavation (2010), lithic contexts. XRD analyses made on them led to
six small piles of copper oxide fragments were found the identification of three types of ores: azurite,234
in contiguous sectors (zones 20, 21, 24) in Chalco- cuprite,235 and digenite-hematite.236 Contrary to
lithic levels. Unfortunately, the soil sampled with what we wrote previously, these ores are not semi-
them has ruined this evidence, and, therefore, we smelted.237
could not analyse them. Not only do all these ore fragments suggest the
Although we lack analyses of these finds, they existence of local metallurgical activities at Mentesh
nevertheless suggest that metal activities were Tepe, they also point out that during the Chalcolithic
going on at the site, and this is confirmed by other
metallurgical artifacts. 233 Inventory number: Met-06-2011; Analysis number: 4646-11.
234 Inventory number: Met-36-2010; Analysis number: 4632-10.
235
Inventory number: Met-16-2011, Analysis number: 4647-11.
236
Inventory number: Met-28-2010; Analysis number: 4628-10.
Sulfidic ore containing 8.1% sulfur, according to ICP-MS ana-
lyses.
232 237
For details, see Courcier et al., in prep. Courcier et al. 2012, 212 Tab. 2.
Fig. 155
Mentesh Tepe. Frag-
ments of ceramic finds
related to metallurgy.
1 ingot-mold,
Met-09-2008; 2 cru-
cible, Met-11-2009;
3 mold, Met-21-2010
period the metallurgists at the site were already able tom of an ordinary Chalcolithic pot with vegetal
to smelt oxide (cuprite) and hydrocarbonate (azurite) temper (Fig. 155,2). Different kinds of analysis have
copper ores, as well as probably also sulphidic cop- been made (thin section, XRD and EDS).241 They
per ores (digenite-hematite). allowed the identification of a vitrified part (1–
1.2 mm) containing many little prills of copper. This
Ceramic finds related to metallurgy potsherd, thus, seems to have been used as a cru-
cible. The copper prills are in arsenical copper (3–
Three fragments of ceramic finds related to metal- 5% As), with some amounts of lead (<1%)242 and
lurgy have been found at Mentesh Tepe (Fig. 155). few sulfur inclusions (0.33–3.09% sulfur), which
Two of them have been investigated in detail in the provide a small amount of evidence for the smelt-
laboratory.238 ing of a sulfidic ore. In the present state of re-
The first one, a fragment of ingot mold239 was search, we still have some difficulty in determining
discovered during the first campaign (Fig. 155,1). whether it is a smelting or a melting crucible, and
No traces of metallurgical use are visible at a ma- we are still working on this matter.
croscopic glance and comparisons of its shape date The third fragment of ceramic find243 was dis-
it to the Kura-Araxes period. This ingot mold has covered in a Chalcolithic level with an important
not been yet analysed. concentration of metallurgical artifacts (zone 21). Its
The second one240 was discovered among rectilinear edges form a right angle and suggest a
many other potsherds and represents the round bot- general rectangular shape, so that it could more
238 241
Courcier et al. 2012. I would like to extend my warmest thanks Courcier et al. in prep.
to Stephen Merkel, who helped me to study these ceramics. 242
In one case, we detected a small prill of argentiferous lead
239
Inventory number: Met-09-2008. (47.1% Ag, 1.17% Pb), but it is probably not representative.
240 243
Inventory number: Met-11-2009; Analysis number: 4007-12. Inventory number: Met-21-2010; Analysis number: 4008-12.
probably be a mold than a crucible (Fig. 155,3). It Finally, the unidentified artifacts discovered in
seems to have endured higher temperatures than the upper levels250 are in copper with low amounts
the first crucible, in view of cracks on its external of arsenic (0.107–0.382% As). Some traces of gold
edge. The different analyses made on this fragment (145 ppm), nickel (82 ppm) and antimony (105
(thin section, XRD and SEM-EDS)244 show that the ppm) were detected in two of these.251 Unfortu-
copper prills trapped into the slaggy encrustation nately, and contrary to our previous hypothesis,252
visible along its edge do not present any evidence we were not able to determine their nature (metal
of alloying, since no tin, arsenic, lead or antimony or ore), since the oxides products detected by XRD
was detected. Sulfur was not detected in any signifi- could be due either to oxide metal or to weathered
cant amount either, and no sulfide inclusions were ores.
found to indicate the smelting of sulfidic ore. How-
ever, if its shape is definitely that of a mould frag-
ment, we could not find any evidence that it had Unidentified objects
yet been used: the interior surface did not present
any trace of metal melting or of casting. Fluxing from Eight unidentified objects have been discovered
fuel ash may have played a part in the observed mainly in the Chalcolithic levels. Some were found
vitrification of its surface. In the present state of our on floor levels or in well identified structures: one
research, we therefore consider that this find had comes from a Chalcolithic silo,253 while another was
not been used as a crucible or as a mold, but that, found close to a Chalcolithic hearth/oven.254 Two
by one way or another, it happened to fall into a others were found in Kura-Araxes pits.255 Contrary to
hearth used for copper metallurgy. the unidentified artifacts, which are only character-
ized by oxides products, the cross section made on
Unidentified artifacts these objects revealed metal. Moreover, their metal-
lographies allow the identification of microstructures.
Besides the objects listed above, 11 heavily corroded According to ICP-MS, the majority of the Chal-
unidentified artifacts have been discovered.245 They colithic unidentified objects are in non-alloyed cop-
are characterized by fragments sticking together be- per. Only two256 are in arsenical copper (1.39–
cause of oxide products. Five were found in sectors 1.5% As), while four257 present high amounts of
of the Chalcolithic levels, where other metallurgical antimony and nickel traces (115–395 ppm Sb;
artifacts have also been found (zones 17, 18, 20), 100–1780 ppm Ni). This characteristic could be ex-
three come from Kura-Araxes intrusive pits into plained by the ores that were used.
Chalcolithic layers, whereas the last three were dis- The unidentified objects coming from Kura-
covered in the disturbed upper levels and therefore Araxes pits are in non-alloyed copper. Only one258
their date is not secure. presents high amounts of antimony and nickel
According to ICP-MS analyses, the Chalcolithic traces (105 ppm Sb; 69.99 ppm Ni). Their chemical
unidentified artifacts are all of copper.246 Only one247 compositions are very close to those dated to the
presents a high amount of arsenic traces (2600 ppm), Chalcolithic period.
while another contains a high amount of lead traces The metallographies realized on the Chalco-
(2700 ppm).248 lithic unidentified objects show the different steps
The unidentified artifacts coming from the of their production: casting, cold-working and an-
Kura-Araxes intrusive pits are also in copper but
their amount of arsenic is higher (0.44–0.63%
250 Inventory
As).249 numbers: Met-07-2008, Met-08-2008 and
Met-11-2008; Analysis numbers: 4184-09, 4185-09 and
4187-09.
251
Inventory numbers: Met-07-2008 and Met-08-2008.
252
Courcier et al. 2012, 212.
253
Inventory number: Met-35-2011. This artifact corresponds to a
very small fragment, found during the sifting of the content of
244
Courcier et al. in prep. a silo. I would like to sincerely thank A. Decaix for this dis-
245 covery.
Inventory numbers (those with * have been analyzed):
254
Met-07-2008 *, Met-08-2008 *, Met-11, 2008 *, Met-01-2010 Inventory number: Met-03-2010. It was found close to a Chal-
*, Met-05-2010, Met-07-2010 *, Met-09-2010, Met-19-2010, colithic rectangular oven (structure 50).
Met-31-2010 *, Met-25-2011*, Met-36-2011*. 255 Inventory numbers: Met-29-2010 and Met-21-2011; Analysis
246
Three Chalcolithic unidentified artifacts were analyzed: numbers: 4626-10 and 4667-11.
Met-31-2010, 25-2011, Met-36-2011; Analysis numbers: 256
Inventory numbers: Met-03-2010 and Met-05-2011; Analysis
4630-10, 4648-11 and 4649-11. numbers: 4615-10 and 4654-11.
247 257
Inventory number: Met-25-2011. Inventory numbers: Met-11-2010, Met-30-2010, Met-03-2010
248
Inventory number: Met-36-2011. and Met-05-2011; Analysis numbers: 4620-10, 4629-10,
249 Inventory numbers: Met-01-2010 and Met-07-2010; Analysis 4615-10 and 4654-11.
258
numbers: 4613-10 and 4618-10. Inventory number: Met-21-2011; Analysis number: 4667-11.
Fig. 156
Mentesh Tepe, awls. 1 semi-manufactured Met-05-2008; 2–31 fragments: 2 Met-01-2011; 3 Met-09-2011; 4 Met-24-2011; 5 Met-12-2011; 6 Met-10-2010;
7 Met-33-2010; 8 Met-27-2010; 9 Met-27-2011; 10 Met-19-2011; 11 Met-18-2011; 12 Met-17-2011; 13 Met-15-2011; 14 Met-20-2010; 15 Met-18-2010; 16 Met-04-2010;
17 Met-31-2011; 18 Met-06-2008; 19 Met-06-2010; 20 Met-11-2011; 21 Met-32-2011; 22 Met-23-2011; 23 Met-01-2009; 24 Met-22-2011; 25 Met-08-2010;
26 Met-22-2010; 27 Met-02-2010; 28 Met-04-2009; 29 Met-06-2009; 30 Met-07-2011; 31 Met-20-2011; 32–39 curved: 32 Met-26-2011; 33 Met-14-2011;
34 Met-02-2011; 35 Met-23-2010; 36 Met-04-2008; 37 Met-32-2010; 38 Met-34-2011; 39 Met-08-2009; 40–46 complete: 40 Met-29-2011; 41 Met-10-2008;
42 Met-14-2010; 43 Met-04-2011; 44 Met-13-2011; 45 Met-03-2008; 46 Met-10-2011 (photography and montage A. Courcier)
nealing. In some cases,259 an important final cold- search, we can now propose an identification for
working was carried out. some of them: a fragment of blade260 and two in-
According to the shape and the manufacturing gots.261
process that have been recognized during our re-
259 Inventory numbers: Met-03-2010 and Met-30-2010; Analysis 260

Inventory numbers: Met-03-2010.
261
numbers: 4615-10 and 4629-10. Inventory numbers: Met-11-2010 and Met-05-2011.
Objects ones are also characterized by high amounts of

nickel traces (100 to 3440 ppm Ni). Metallographic
Two types of objects have been excavated at Men- analyses of the Chalcolithic awls269 show the fol-
tesh Tepe: ornaments (rings, spiral ornaments) and lowing procedure: casting, cold-working, followed
awls. by annealing. Some270 also present a final intensive
cold-working.
Ornaments All the awls from Kura-Araxes pits are in unal-
loyed copper (Tab. 22). One presents high amounts
Six ornaments have been found: a pendant, two of nickel traces (205.319 ppm Ni).271 According to
rings and three spirals. However, only the ring frag- metallographies, the manufacturing process is the
ment262 comes from a Chalcolithic floor. All the same as that identified for the Chalcolithic awls.
others come from Kura-Araxes contexts and their The awl discovered in Kura-Araxes level is in unal-
shapes, moreover, are characteristic of this cul- loyed copper with nickel traces (120 ppm; Tab. 3).
ture.263 Finally, the Kura-Araxes semi-manufactured awl
The Chalcolithic ring fragment is in arsenical (mentioned in footnote 267) is in arsenical copper
copper (1.09% As), with high amounts of antimony (3.12% As) with traces of nickel, selenium and anti-
traces (225 ppm Sb). The fibrous structure shows a mony (134 ppm Ni; 535 ppm Se; 1790 ppm Sb).
heavily worked alloy. According to metallographic analysis, this semi-
Most of the Kura-Araxes ornaments are in manufactured awl was casted, soft hammered and
non-alloyed copper and only two264 are in arsenical annealed.
copper (1.77% and 6.326% As). The pendant pre-
sents not only a high percentage of arsenic (>6% Origins of the ores
As), but also high amounts of antimony traces
(4095 ppm). Two ornaments in unalloyed copper According to a request on our GIS data,272 32 ore
present respectively high amounts of antimony and deposits are available within a radius of 50 km
nickel traces (220 ppm Sb;265 394.74 ppm Ni266). from Mentesh Tepe and the number increases to 40
The manufacturing of the Kura-Araxes ornaments, as within a radius of 100 km. Moreover, ancient traces
shown by metallographies, gives evidence of cast- of mining were observed in some of these places273
ing, cold-working and annealing. during a survey carried out in 2007, although we
were not able to date them.
Awls The copper ores that have been discovered at
Mentesh Tepe demonstrate the interest of its inha-
Awls represent the majority of the metal finds (49) bitants for this raw material. Even though no metal-
(Fig. 156). Most of them (29) are dated to the Chal- lurgical furnace has yet been discovered, the work
colithic period (Tab. 21), and two come from a se- done on the ores (crushed azurite) and the clay
cure Kura-Araxes level, but the rest cannot be dated finds related to metallurgy suggest that they were
with certainty between the two periods, because 16 locally used.
were discovered in Kura-Araxes pits, while 8 were The origin of these ores can be identified
found in the disturbed upper level. We have distin- through chemical and isotopic analyses. The
guished four types of awls: semi-manufactured,267 amount of arsenic detected, not only in the Kura-
intentionally curved, fragmented and complete. Araxes objects, but also in those dated to the Chal-
The majority of the Chalcolithic awls are in colithic period, suggests ores containing arsenic.
non-alloyed copper, but 8 awls contain arsenic We cannot exclude the use of sulfide ore either,
(1.42% to 3.25% As;268 Tab. 22). Some of the last since we have detected some particles of CuS in
the crucible fragment and also because of the dis-
covery of a fragment of digenite (see above). Stran-
262 Inventory number: Met-03-2011; Analysis number: 4652-11.
263 Gambaschidze et al. 2001, 248.251.265; Glonti et al. 2008,
gely enough, the amount of sulfur remains very low
182–183.
in the objects (<0.001–0.545% sulfur). Most of the
264
Inventory numbers: Met-33-2011 and Met-03-2009; Analysis
numbers: 4678-11 and 4690-11.
265 269
Inventory number: Met-05-2009; Analysis number: 4692-11. Inventory numbers: Met-14-2011, Met-23-2010, Met-04-2011,
266 Met-13-2011, Met-18-2011; Analysis numbers: 4621-10,
Inventory number: Met-30-2011; Analysis number: 4675-11.
267
Inventory number: Met-05-2008. This semi-manufactured awl 4625-10, 4653-11, 4660-11, and 4664-11.
270 Inventory number: Met-04-2011 and Met-26-2011; Analysis
comes from a Kura-Araxes structure.
268
Inventory numbers (according to an increasing per cent in As): number: 4653-11 and 4671-11.
271
Met-27-2011, Met-32-2011, Met-15-2011, Met-27-2010, Inventory number: Met-07-2009; Analysis number: 4693-11.
272
Met-18-2011, Met-13-2011, Met-04-2011 and Met-26-2011; Cassard et al. 2010.
Analysis numbers: 4672-11, 4677-11, 4662-11, 4664-11, 273 Dagkesaman, Gosha, Siny-Yar, Sojudlu, Novogarilovka, Khar-
4660-11, 4653-11, 4671-11. Khar and Khanlar.
Tab. 22
Mentesh Tepe. ICP-MS
results of the awls,
organized according to
increasing percent of
arsenic for each period
Fig. 157
Location of Mentesh
Tepe, of the ore depo-
sits and of the geologi-
cal formation men-
tioned in the text.
1 Mentesh Tepe; 2 Ala-
verdi; 3 Dagkesaman;
4 Akadzor; 5 Fioletovo;
6 Frolova-Balka; 7 Gos-
ha; 8 Khanlar; 9 Siny-
Yar; 10 Khar-Khar;
11 Sojudlu; 12 Novoga-
rilovka; 13 ophiolitic
formation of Sevan-
Akara; 14 Sotk;
15 Kadjaran;
16 Tey and Lichk
metallurgical artifacts contain high amounts of on the ores from Mentesh with those achieved on
antimony and nickel traces (0.283–4095 ppm Sb; the Armenian deposits,276 on the one hand, and on
0.424–3440 ppm Ni). All of these characteristics the Georgian deposit at Madneuli,277 on the other.
make us consider that copper ores containing nick- According to isotopic correlations, we can propose
el and indeed even arsenic was used at Mentesh either the Kvemo-Kartli region in Georgia278 or the
Tepe. The presence of nickel in ores is mainly a Lori,279 Gegharkunik280 and Syunik281 regions in Ar-
characteristic of ophiolithic host rocks, which are at- menia (Fig. 157). However, the signature of some
tested in the Caucasus. Within this ophiolithic belt, other metallurgical finds from Mentesh Tepe sug-
the Sevan-Akara ophiolitic formation274 is close to gests another origin.282 Unfortunately, because of
Mentesh Tepe (Fig. 157) and overlaps with the the actual lack of isotopic studies on deposits within
southern part of the Kedabek and Dagkesan ore dis- Azerbaijan, we cannot yet give a definitive state-
tricts (Azerbaijan) and with the eastern part of the ment.
Gegharkunik one (Armenia), all also famous for their
numerous copper deposits. Furthermore, arsenic Concluding remarks
mineralization is attested in some of them.275 Thus,
it is possible that the ores used at Mentesh Tepe The excavations at Mentesh Tepe have given an im-
came from this western region of Sevan-Akara. portant amount of new data to better understand
Lead isotopic analyses also give some answers the metallurgical process during the 5th mill. BCE.
to this question and seem, similarly, to point at de- The evidence we have attests the existence of most
posits west and southwest of Mentesh Tepe. This
statement comes from the comparison of the results
276 Meliksetian et al. 2009.
277 Dschaparidze 2011.
274 This ophiolitic formation spreads from the upper reaches of
278
the Tutkhum River, near the village Qişlak (Qarabağ region), Madneuli deposit.
279
to Lake Sevan (Armenia); Adamia et al. 1981, 437–447; Gam- Alaverdi, Ankadzor, Fioletovo and Frolova-Balka deposits.
280
krelidze 1982, 99–101; Gasanov 1996. Sotk deposit.
275 Devebojnum, Sotk, Kedabek, Dagkesan: See THŁHÆ*Œ@- et al.; 281
Kadjaran, Tey and Lichk deposits.
282
8/ıC$+@- et al. 1987; `/Æ/-H/+* et al. 2005. Courcier et al. in prep.
of the steps of the ‘‘chaı̂ne opératoire’’, from the Contrary to the impressive assemblage of
ore to the objects. The raw material, probably col- bone tools known from the S̆omu-S̆ulaveri culture,
lected locally, corresponds not only to oxide and the collection discovered at Mentesh Tepe comprises
hydo-carbonate copper ores, but also to sulfidic only 37 items. This can probably be explained by
ores, as shown by the first investigations made on the importance that metal had already gained dur-
fragments of crucibles dating to the Chalcolithic ing the Chalcolithic.286 Thus, only three bone tools
period. According to chemical analyses, arsenical come from the Chalcolithic levels (two awls and
copper alloyed artifacts coexist with unalloyed cop- one shovel found in a pottery kiln), while six have
per objects already during this early period. How- been discovered in the Neolithic levels, although
ever, the amount of arsenic could also be explained excavation of these has only just started (essen-
by the use of copper-arsenic ores, which are fre- tially awls, one of which was discovered in an infant
quent in the deposits in the Caucasus. The metallo- burial).
graphic analyses demonstrate the practice of cast- The industry comprises mostly tools, and only
ing, cold-working and annealing. We hope that two pieces can be considered as waste from pro-
future excavations and new analyses will further duction, probably because all the rest has been put
augment our knowledge about these metallurgical together with the discarded faunal remains.287 The
activities at Mentesh Tepe. state of preservation of this industry is generally
rather poor.
The bone industry from Mentesh Tepe Most of the discoveries (26) concern pointed
objects, essentially awls (24) made on long bones of
Gaëlle Le Dosseur ovi-caprids, sometimes young individuals (Fig. 158,
1–2). Metapods (6) were often used, but tools on
The exploitation of bone materials in the Near East other bones (a shinbone and a radius) have also
has a long tradition. Equipment made from these been identified. For the tools with handle included,
materials had become part of daily life since at the distal part of the bone (5) was preferred to the
least the Natufian.283 At that time an extremely rich proximal one (1), as is generally the case from Neo-
repertoire of tools, weapons and jewelry was cre- lithic times onwards.
ated as well as efficient technical means to produce Most of these products were achieved through
them. During the subsequent Neolithic, all of these percussion. The length and regularity of some of
achievements became stabilized, although some them shows a good mastery of this technique. Most
changes are observed in the equipment composi- of the tools obtained with this technique were re-
tion: weapons and jewelry made of bone are rare, worked, either through abrasion (12), or by scraping
while tools are the main part of the production (2) or a combination of both (2), mainly on the active
which is now quite diversified.284 While the bone in- part. Some flakes were used without any further
dustry is part of the intimate sphere of the domestic transformation, which demonstrates a will to work
house, it is also implied in interregional exchanges, quickly and efficiently.
so that mutual influences are visible from the north- The collection also comprises nine cutting or
ern to the southern Levant. During the Chalcolithic, blunt tools: 3 shovels, 1 beveled tool, and 3 blunt
a huge change is observed in the choices of raw ma- tools.
terials: ivory is much more often exploited, to create – The shovels (Fig. 158,4) are made on the scapu-
beautiful objects that could be assigned to the cate- lae of large ruminants, probably bovids. Their ac-
gory of art productions like sculptures.285 The use tive part is wide. One was found in a Chalcolithic
of metal tools could have facilitated the emergence pottery kiln and is blackened because of its posi-
of this craftsmanship, as well as regional influences. tion as discard in the fire. The active part is lo-
Whereas the bone industry from the Near East cated on the proximal part of the bone, while
has become quite well known, only very little infor- the base is on the distal articulation. This base
mation has been reported on the bone industry had been perforated to insert a handle. It shows
produced in the neighboring Caucasus. The study important traces of wear (longitudinal striations
of Mentesh Tepe is, therefore, one contribution in at the active part, removal scars creating a con-
this direction, and further work on material from cavity, etc.). The shovels could have been used
the Neolithic period both at Mentesh and Aruchlo to work earth (to cultivate, to mix clay with temper
in the frame of the ANR/DFG project should provide for ceramics, to mix clay with straw for bricks,
further data in the future. etc.); we plan experimentations on this usage.
The scapulae were worked directly without major
283
Campana 1989; Stordeur 1988; Belfer Cohen 1988.
284 286
Le Dosseur 2008. Courcier, in this article.
285 287
Perrot 1968. This will be studied separately next season.
Fig. 158
Mentesh Tepe. Bone
implements (drawings
G. Le Dosseur)
transformation: only the spine had been knocked In conclusion, the small sample of Mentesh
by percussion and the base perforated.288 Tepe is dominated by tools used for transformation
– The beveled cutting tool has been made on the activities. Among them, awls are ubiquitous and dif-
shinbone of a large ruminant and obtained through ficult to date. On the contrary, shovels on scapulae
percussion. It is not well preserved. might have a better significance on a chrono-cultur-
– The blunt tools category is quite heterogenous. al level. These objects are not so common in the
The only common feature is the presence of red early prehistoric period. In the southern Levant,
particles on the active parts, probably resulting massive cutting tools with a handle at the level of
from work on a flexible material spread with ochre. the proximal articulation of the selected bone are
Finally, the sample contains the waste of a horn core known in Neolithic289 and Chalcolithic290 contexts.
showing several transversal grooves (Fig. 158,3). Their active part is not as large as our tool, and the
They seem to be the result of the sawing of the horn bone chosen is not the same (radius), but they are
in three to four parts, in which only the external part nevertheless comparable. They are on the contrary
was needed. rather abundant at S̆omu-S̆ulaveri culture sites, like
289
Stordeur 1999.
288 290
The technique for perforation has not yet been identified. Perrot, personal comm. (material from Beersheba).
Goy Tepe,291 where they are dated to the Neolithic. other regions further away. We will first present the
This similarity between the two periods contributes results concerning the general landscape and envir-
to the discussion on the heritage of Neolithic popu- onmental conditions, and then comparative studies
lations to Chalcolithic ones in our area. on the material culture.
The main raw material used is bone. Quite of-
ten people extracted this material from domesti- Holocene landscape and human modes
cated animals (ovi-caprines) at hand, but large ru- of occupation in the Kura Valley (Azerbaijan)
minants, wild or domesticated, were also used. The
selections, probably oriented by technical advan- Vincent Ollivier and Michel Fontugne
tages, are quite systematic: metapods of small
ruminants for awls, scapulae of big mammals for Introduction
shovels.
The main characteristics of the tools demon- Current geomorphological studies raised the ques-
strate a will to produce in a quick but efficient way: tion about the Caspian eustatic influence on the
quite often tools are made on plain bones only Kura Valley landscape mutations since the last gla-
lightly shaped, while rapid techniques are frequently cial maximum.293 The Caspian Basin is a vast endor-
used (percussion, abrasion). Part of the blanks ob- heic system (371 000 km2), particularly responsive to
tained by percussion are probably the result of a the flows of its contributors, to climatic oscillations
proper, well controlled débitage: the blanks are fairly (from regional to extra regional scales) and to solar
long and regular. We cannot exclude that others, insolation. For these reasons, its levels have changed
with more random shapes, were collected from kitch- dramatically both erratically and cyclically over var-
en wastes. ious time scales,294 causing vast modifications in
As has been observed on grinding materials292 both the volume and the area of the water body. In
and in several areas of the site, ochre is quite often numerous examples around the world, hydrosystems
attested on bone tools. It would be useful to analyse have reacted to eustatic variations along the geologi-
this red material found on stones and bones and to cal time scale by regressive incisions during low-
determine, through a joint study by specialists, its stands and important rapid valley infilling during
possible uses and the links between two industries. high sea levels. Considering this important para-
Similarly, the discovery of bone shovels that were meter for landscape mutation throughout the Qua-
possibly used to work the earth needs further cross ternary period, modes of human occupation must
studies and experimentations. have either reacted to or been strongly influenced
during the Holocene in the entire area under study,
which concerns the Mil Steppe in the Lower Qara-
Part IV: bagh and the Tovuz region (Fig. 159). The question
here is to identify and quantify the real impact of
Studies concerning the three areas eustatism in the various hydrosystems of this study
of the Middle Kura Valley area and its effect on the human mode of occupa-
tion during the Holocene.
Bertille Lyonnet and Barbara Helwing
Geomorphological research in the Tovuz Region
With this section of our work we enter into the stu-
dies dealing with the three geographic areas of the The Tovuz region is from the point of view of geo-
Middle Kura Valley. In this part of the project, our archaeology actually the area investigated in most
specialists study and compare the material culture, detail of the three study regions. The geomorpho-
or the landscape and geomorphology of the three logical research concerns the morphosedimentary
places in order to detect differences and similarities organization of the terrace levels of the Agstafa,
and the chronological development. This phase of Hesensu, Tovuz, Arenji, Zeyem and Şamkir çayi or
our research is just at its very beginning, and most riverlets, along which archaeological sites are lo-
of the time it was possible to work on only two cated.
areas and not yet on all the collections. Neverthe-
less, the project gives a hint at the potential of such Environmental components
studies, which will aid in a better determination of
the local cultures, of their local development, and The area of study, which is oriented northwest to
of their relations between themselves and/or with southeast on the right bank of the Kura River and
291 293
F. Guliyev, personal comm. Ollivier et al. 2011.
292 294
Hamon, in this article. Kazancı et al. 2004.
Fig. 159
South Caucasia: map
and location of the
study regions
bounded between the cities of Agstafa and Şamkir, Hydrosystem characteristics

is composed of mountain-, hill-, foothill- and plain-
landscapes. The climate of the mountainous part is The torrential units studied are part of the Kura hy-
considered as sub-Mediterranean and semi-humid drosystem. This hydrosystem has been directly con-
at lower altitudes (with beech, oak, and pine for- nected to the Caspian Sea since at least the Upper
ests) and thermo-moderate and humid in the higher Pleistocene. The Agstafa, Hesensu, Tovuz, Arenji,
parts (alpine meadows). On the other hand, the Zeyem and Şamkir çayi hydrological systems have a
plain and foothill ecosystems are associated with mature equilibrium profile and share quite the same
dry-hot semi-deserts and dry steppes with mild win- basin morphometric characteristics. Their watershed
ters. On the right bank of the Kura River, the main area encompasses ca. 2400 to 2650 km2, their length
lithology and outcrops are Jurassic, cretaceous, vol- measures between 133 to 120 km, and their mean
canic, metamorphic and sedimentary rocks. On the discharge varies between 0.91 m3/s (Tovuz çayi) and
left bank, the geology is constituted by Miocene 13.6 m3/s (Agstafa çayi). Globally, their annual water
transgressive deposits and Pliocene tectonized for- supply derives 45% from groundwater, 35% from
mations, rich in pebbles, silts and gypsum. Qua- the snowmelt and 20% from rain precipitation. The
ternary deposits predominate in both parts of the actual physiography of the Azerbaijan Kura Basin,
Kura Valley. With a minimum estimated thickness with an average declivity of 0.17% from the Geor-
of 80 meters, they belong mostly to alluvial fans gian border to the Caspian shore, presents a large
and terraces from the torrential rivers of the Lesser and relatively flat valley, in which the eustatic var-
Caucasus foothills and to the Kura River. Morphocli- iations can be expressed on a wide spatial scale.
matic and climato-eustatic variations have shaped The Tovuz region, located in the upstream part of
these formations since the end of the last climatic this geomorphological environment, has significant
cycle. tectonic activity with an average uplifting of 4 to
6 mm/year. This geodynamic contributes to the ma- ertial impact time on the upstream hydrosystem Fig. 160
Mentesh Tepe area.
turity of the longitudinal river profiles and to the morphogenical trend and the palaeo-shoreline loca- Zeyem çay, preliminarily
amplification of the eustatic signatures in the ter- lizations over time are based on the interpolation studied stratigraphy
race morphosedimentary expressions. Five terrace of radiochronological data and topographic altime-
levels were identified during the geomorphological try of the Azerbaijan territory.296 The second strat-
field work analysis, and their organization seems to egy was to focus on the archaeological sites discov-
follow relative level oscillations of the Caspian Sea ered during field work in the course of the previous
since the last glacial maximum. French Foreign Affair Ministry mission directed by
Bertille Lyonnet. Around 60 sites were visited, in-
Geomorphology and geoarchaeology: cluding visible mounds and scatters of sherds on
preliminary result the surface. Despite the diachronic aspect of the
sites observed, we have noticed that in this study
The geomorphological and geoarchaeological stud- area the major settlements from Neolithic to Bronze
ies were focused on two main strategies. The first Age period (notably: Göy Tepe, Mentesh Tepe and
one was to define precisely the landscape compo- Soyuq Bulaq) concerns primarily the higher terrace
nents and driving forces of the system. Each of the levels of the Kura River and its tributaries.
tributaries of the Kura River was analyzed in this These high terrace formations started to settle
regard. A total of 26 stratigraphical sections were during the Upper Pleistocene at 18 020 ) 50 BP
decrypted and 21 radiocarbon analyses are in pro- (19 788–19 291 cal. BCE), a period of major Cas-
cess. The estimation of palaeohydrodynamism in the pian Sea transgression (Fig. 160). This dynamism,
respective stratigraphy studied is based on macro- characterized by high energy deposits (pebbles and
facies sedimentological analysis.295 The eustatic in- sands), continued until 11 610 ) 35 BP (11 661–
295 Adapted from Hjulström 1935; Neill 1967; Neill 1968; Miall
296
1988; Wu et al. 1973 notably. SRTM NASA, cp. Farr et al. 2007.
Fig. 161 11 369 cal. BCE) during the late glacial period. A event (Fig. 161). During the high relative level of
Mentesh Tepe area. major regressive incision, which means a very im- the Holocene Caspian Sea, the flood from the Les-
Holocene flood area
and silt deposit (Zeyem portant deepening of the downstream to upstream ser Caucasus and its foothills covered the high ter-
çayi and Hesensu çayi) river valley, started between the late glacial era and races and caused the deposition of silt sediments.
the beginning of the Holocene (radiocarbon dating This dynamic of overflow over the Pleistocene/late
in progress). This event could be linked to the con- glacial alluvial fan, where the settlements were lo-
text of the Mangyshlak marine regression. This mor- cated, provided thin sediments that are favorable
phogenic trend was immediately followed by an infor agriculture. The development of a humid zone
tensive period of sedimentation, which continued with thalwegs, in which runoffs were more or less
until 3365 ) 30 BP (1742–1606 cal. BCE, Fig. 160). concentrated, at least until the Chalcolithic period,
During this period, the hydrography and sedimenta- contributed to a relatively effective in situ water
tion were potentially connected to the dynamics of supply. The geological setting also played an im-
the Caspian Holocene transgression, and settle- portant role in water availability in the area of ar-
ments were located on the periphery of river banks. chaeological sites. Lithology and faults control the
Gradually, the next phases of marine regression dis- proximity of aquifers and spring re-emergence. Yet,
connected the sites from the river proximity by the the eustatic movement can be considered here as
return to incision morphogenic trends. the main parameter that controlled the hydrosystem
Field work and investigations on the upper morphogenical adjustment over time.297 Subse-
terrace levels enabled us to identify a specific area quent to the Bronze Age, the thalwegs were signifi-
with a longitudinal transition of the outcrop facies. cantly deeper due to the lowstand of the Caspian
The soils in the upstream part are composed of Sea. In this context, the human occupation seems
silty deposits, while the downstream ones, in the to have decreased progressively, probably because
direction of the Kura, consist of pebbles and sands. the Neolithic and Chalcolithic settlement areas were
The Neolithic, Chalcolithic and Bronze Age occupa- 30 meters above the riverbeds within these new en-
tions are mainly located in this silty deposit area. vironmental conditions. Nowadays, the main parts
The stratigraphy of the right bank Kura tributaries of the higher terraces are relatively arid areas, colo-
clearly shows a fitting with erosion contact between nized by a steppe cover of Artemisia. The main ar-
silty and pebbly deposits. This is particularly visible gument for this aridity is the very deep position of
in the natural cutting of the Zeyem and Hesensu the groundwater due to morphostructural conditions
rivers (Fig. 161). A radiocarbon dating at 4495 )
35 BP (3352–3090 cal. BCE) in the silty levels close
to Mentesh Tepe confirms the chronology of the 297
Newson 2002.
(graben and cretaceous limestone syncline). In ad- around 4–6 mm/year in the proximal Lesser Cauca-
dition, the landscape morphosedimentary expression sus foothills. Downstream the uplifting values are ne-
and the sedimentological feature of the Quaternary gative with subsidence rates around ––4 mm/year,
formations outside the silty area, which are domi- decreasing in the estuarine area direction.
nated by sands and quasi open-work pebbles de-
posits, support the water infiltration, which is ob- Hydrosystem characteristics
servable at the sole infilling by seeps and spring re-
emergence. Work is in progress in this area and Few developed hydrosystems prevail in the Lower
needs to be further developed in order to complete Qarabağ plain. This is due to the specific climatic
the fluvio-eustatic model and its relation to human conditions (mainly subtropical and semi-arid) and
occupations. Nonetheless, the first environmental the geomorphological/lithological context predomi-
answers can be proposed. nated by low declivity and sandy-silty sedimentary
cover. The precipitation, with an annual average
rainfall under 400 mm/year, is absorbed by the high
Geomorphological research in the Mil Steppe area permeability of upstream soils, and the river drain-
of the Lower Qarabağ age is uncertain. Downstream the underground water
re-emerges and contributes to the Kura water table.
The second area of study is situated in the middle It is also important to note that more than 70% of
of the Mil Steppe in the Lower Qarabağ, in the Ağ- the river supply from the foothills is provided by
cabedi Rayon. As in the Tovuz region, archaeologi- this groundwater. The fluvial morphology of the Af-
cal sites dating to the Neolithic period are under şar, Gargar and Qarasu çayi is characterized by
excavation by a German-Azerbaijan team within the more winding riverbed with less cached thalwegs in
French-German ANR-DFG project ‘‘Ancient Kura’’ comparison with the Tovuz region hydrosystems.
and the Academy of Science of Azerbaijan. Four The mean discharge of the Qarasu and Gargar çayi,
Kura tributaries are under study in this open land- one of the biggest unsustainable torrential rivers in
scape: the Afşar, Gargar and Qarasu streams and – our study area, is around 4.2 m3/s with a reduced
largest of all – the Araxes River. The tributaries of watershed inferior to 2500 km2. Finally, one of our
the Kura River are precious testimonies of the in- key rivers is the Araxes River. With a length of
land impact of Holocene eustatic variations on an 1072 km (cp. the Kura River with 1514 km), a wa-
archaeologically rich, yet poorly known region of tershed of 102 000 km2 (218 906 km2 for the Kura)
Azerbaijan. Because they are closer to the Caspian and a mean discharge of 285 m3/s it is the major
hydrosystem than the Tovuz region, their study will Kura tributary with important geomorphological im-
provide us with clues for quantifying chronologically plications beyond Azerbaijan and Armenia.
and spatially the landscape changes and their long
term impact on ancient societies. In addition, the Geomorphology and geoarchaeology:
geomorphological analysis of the Araxes will supply preliminary results
valuable comparative data connected to the Arme-
nian upstream area, where the Neolithic archaeolo- Initially the palaeo-landscapes of the Ağcabedi ra-
gical sites of Aratashen and Aknashen are located. yon are difficult to determine without serious field
work. The soft morphology of the plain and hills en-
Environmental components countered there requires rather a micro-topographic
analysis than a determination of the large landforms.
The Lower Qarabagh Mil Steppe is an open land- However, gradually the Quaternary heritage and the
scape with xerophytic vegetation that has developed forms and superficial formations that make up the
in a plain punctuated only by small hills in direction actual landscape architecture become recognizable.
of the Lesser Caucasus foothills. The climate is mildly We used the same strategies as in the Tovuz region,
subtropical, warm and dry. The calcareous substra- that is, an analysis of upstream – downstream sys-
tum of the Jurassic and Cretaceous periods is cov- tematics of the Afşar, Gargar and Qarasu rivers, and
ered by a very thick Neogenic sedimentation (around a focus on the archaeological sites which are fre-
200 m thickness), in which marine and fluvial Qua- quently represented there by mounds or tepe con-
ternary sandy-clayey deposits are significant. Limnic taining abundant data about occupations during
gray/brown and halomorphic/salsodic soils are preva- the 6th to 4th mill. BCE. As in the Tovuz region, we
lent. Closer to the Kura River and to the Kura-Araxes observed five alluvial terrace levels along these
convergence are well developed swampy areas and Kura tributaries, with the same morphosedimentary
riparian vegetation. As in the Tovuz region, a signi- organisation (Fig. 162). A total of 19 radiocarbon
ficant tectonic activity exists. Close to the Ağcabedi samples was made on seven key stratigraphies in
area, the uplifting process is greater with rates determinant locations in order to understand each
Fig. 162
Kamiltepe region.
Morphosedimentary
organization
river’s evolution and their link with the major for- ventional morphogenetic events and characterizing
cing. The geomorphological landscape is structured a high base level for the hydrosystems, this could
by hills, which are in fact the inheritance from the be interpreted as a first indication of the eustatic
Upper Pleistocene and late glacial marine transgres- impact on the landscape and the human occupa-
sion. The micro plateau surfaces, separated by small tions.
slope failures (3 to 4 m), represent the Holocene
terrace complex (Fig. 162). Discussion and conclusion
According to the curve in relative sea level,298
during the Pleistocene-late glacial highstand, the Rises in the marine sea level can be very fast on a
shoreline was at 3 or 4 km from the Kamiltepe area global scale,299 but in the case of the Caspian Sea,
and ca. 50 meters a.s.l. Then, during the late gla- its endorheic character make them even more rapid
cial-Holocene transition, the sea level fell, and the and reactive to the climatic and tectonic settings.
coastline receded ca. 120 km from this site area. Based on this, the hydrosystem and human occupa-
Today, some Holocene soils are still saturated with tions of the Caspian Basin should be even more
salt in a few areas (for example, the downstream sensitive to regional eustatic changes. Alluvial val-
part of the Gargar River). The hydrosystem must leys provide the link between processes that con-
have followed the landscape mutations imposed by trol sediment flux to the continental margin and
variations in the sea level, but the lack of relief in- processes that control dispersal into the basin. The
ferred by the low piedmont declivity hinders the volume of sediment delivered to the margin reflects
visibility of the terrace fittings. However, we could hinterland drainage areas and large-scale relief.300
determine that the sites are located on the terraces The valley incision probably characterizes the low-
number 3 and 4 (Fig. 162). The first observations est of lowstands.301
show that the younger sites are above the flood Pleistocene and late glacial regressive or
plain, while the older ones, like MPS 4 (dated to transgressive variations are of large amplitudes
the first 150 years of the 6th mill. BCE), are posi- (þ/%150 m) and are likely to affect significantly the
tioned in the middle of it and present flood levels fluvial morphology and the morphogenic dynamic
in the top part of their stratigraphy. Opposite con-
299
Melt Water Pulse-1A: Deschamps et al. 2012.
298 Mamedov 1997; Rychagov 1997; Kroonenberg et al. 1997; 300
Blum 2003.
301
Mikhailov et al. 2003; Hoogendoorn et al. 2005. Posamentier 2001.
of the Kura River and its tributaries. The relative sus upstream watershed ‘‘flashflood’’ inferred to cli-
mobility of the Holocene curve of the Caspian Sea matic conditions on the other. In each case, ancient
does not exceed the %21 m/%20 m of altitude in societies had to occupy the most appropriate level
the transgressive phases, i.e. about 6 m higher than of fluvial terraces. More investigations are neces-
the current level. However, the modest Holocene sary in order to build a precise pattern, but the
variations, around 6 m in range, also seem signifi- multiplicity of the sites, of the identified palaeo-
cant enough to have caused the crossing of a geo- channels and of the on-going radiocarbon dating
morphic threshold for the Middle Kura Valley hydro- concerning this area will certainly lead to the dis-
systems. covery of some determinant indications.
This positive feedback of the Kura tributaries These preliminary results and interpretations
to the eustatic movements is expressed through inspire further research work, while giving some de-
valley infilling during the highstand by the multipli- tails about the main morphogenical driving forces
city of sediment inputs available from upstream. that possibly influenced the Holocene human occu-
The regressive incisions during lowstands back up pations in the Kura-Araxes Lowlands. The need for
far into the inland (700 km) and spread to moun- further cross analyses between archaeology and
tain torrential units. Favourable climatic conditions geomorphology is obvious.
are necessary to promote each kind of develop-
ment of morphogenic trends. In view of the impor- Archaeological landscape studies:
tant slopes encountered, the foothills hydrosystems The Mil-Qarabağ Plain and the
can be considered as sensitive units reacting strongly Kvemo Kartli Survey Projects: a preliminary
to medium/high amplitude, but also to the long-term account of the first two field seasons (2010–2011)
changes in the relative level of the Caspian Sea.
The on-going work must be reinforced within this Andrea Ricci
perspective, but important correlation between the
Kura hydrosystem evolution and the Holocene eu- Landscape archaeology plays a critical role in ad-
statism seems to be confirmed. dressing the core research questions of the Ancient
In the Mentesh Tepe area, we noticed that Kura Project. Intensive survey and geoarchaeologi-
modern villages and cultivated areas are located on cal investigations are conducted to understand
the silty deposits that were generally occupied dur- early settlement developments and human-environ-
ing the Neolithic period. Obviously, agricultural ac- mental dynamics in the studied regions. So far de-
tivities require the same approach to the past mor- tailed archaeological surface investigations have
phogenic evolution. Water for irrigation is artificially been conducted both in Azerbaijan and Georgia. In
diverted to terraces areas in the high dry steppe. the former, during the summer of 2010 and 2011,
During the mid-Holocene, irrigation was natural (mar- for a total of seven weeks, a small team carried out
ine transgression and high torrential levels) and ac- initial archaeological landscape explorations along
tive through the overflow floods. the Qaraçay and Qarasu river basins, in the south-
The Chalcolithic sites are seldom present on western part of the Aǧcab di Rayon, south of the
e
the surface, and the majority is buried beneath a Kura River (Fig. 1). The results of the first year of
sedimentary cover, as has been demonstrated by investigation of the so-called Mil Plain Survey (MPS)
the survey mentioned above and BTC pipeline in- have been recently discussed with an emphasis on
vestigations.302 Sometimes associated with Neoli- the Neolithic presence.304 In the second year, field
thic sites, they are also often located downstream work was conducted in collaboration with Vincent
from the piedmont, in the Kura floodplain. In the Ollivier. In September 2011, a two-week period of
lowering phase of the sea level, this area became preliminary research was completed in the region of
particularly exposed to gully erosion in the context of the Kvemo Kartli (KKS), along the Mas̆avera river
the 5500 cal. BP regression.303 This could explain basin, in southeastern Georgia (Fig. 1). The third re-
the low outcrop representation of the Chalcolithic gion to be investigated within the Ancient Kura Pro-
sites in the Tovuz region. ject is located around Mentesh Tepe and scheduled
In the Lower Qarabağ plain, two main geo- for 2012. Because of the brevity of these first two
morphologic phenomena can be distinguished in seasons, this account cannot be dealt with here in
the human modes of occupation context: the river detail or thoroughly; instead, by presenting some
groundwater upwelling following the marine trans- preliminary results of this pilot research, we intend
gressions, on the one hand, and the Lesser Cauca- to provide a basis for discussion and a foundation
for future work in the region.
302
Maynard 2011.
303 304
Hoogendoorn et al. 2005. Helwing et al. 2012.
First of all, field studies aimed at gaining an the Mas̆avera River as well as along both banks of
initial acquaintance with the general topographic, the Qarasu and Qaraçay rivers in Azerbaijan. In
natural and modern land-use characteristics of the Fig. 163, the polygons in orange color illustrate the
two regions in order to understand the ancient and areas that have been actually walked by surveyors,
modern landscape taphonomic processes. In parti- spaced every 15 to 20 m. Transects with 100 m long
cular, at the centre of this part of the research blocks of collection (less frequently 50 m long) were
stands an evaluation of the degree of landscape walked specifically within the 1 km radius around
preservation so as to identify areas of landscape of the site of Kamiltepe (see below: Intensive survey-
survival, where intensive survey could be conducted ing around Kamiltepe; Fig. 170), as well as on the
more successfully. In the field, attention was paid Aruchlo Plain (Fig. 168). Concentrations of artifacts,
to the documentation of the overall archaeological soil signature, morphology and remote-sensing sig-
landscape signature and to recording sites and ar- nal make the definition of a site possible; field
chaeological features for the entire course of the scatters of pottery and archaeological landscape
Holocene. This long-durée approach lags somewhat features are documented too.
behind the original and specific research questions The location of sites and landscape features
addressed by the Ancient Kura Project, but it is ne- are recorded with a field computer or with hand-
cessary to gain an understanding of both the early held GPS. The acquired field data and the subse-
settlement choices and population dynamics to then quent dating of features and sites convey into a
better comprehend the human-environment dynam- Geographical Information System (GIS), within the
ics that occurred in the individual regions. frame of which spatial analyses are conducted and
maps are produced.
Archaeological landscape investigation techniques
Remote sensing data is one of the basic tools for Taphonomic processes
our archaeological landscape investigations. Orthor- M i l P l a i n – A ǧ c a b d i R a y o n
e
ectified LANDSAT imageries together with Soviet
1 : 50 000 topographic maps serve as the main The complex modern ‘‘Upper Qarabağ canal’’ irriga-
mapping reference and offer broad regional cover- tion scheme, mainly composed of a series of open
age for environmental analyses.305 On this basis, all canals, was built and expanded from the 1950s on-
other remote sensing datasets are geo-referenced. wards in the southwestern area of the Aǧcab di e
In particular, CORONA imageries serve as the key Rayon307 to sustain farming activities. Major modern
remote sensing tool for the detection of sites and villages are located along this articulated channel
landscape features.306 The following sets of images system that divides the region into two main dis-
are currently at our disposal: tinct zones (Fig. 164):
– Azerbaijan: two sets of imagery from missions Landscape of destruction: is a largely irrigated
DS1048 and DS1110, respectively, taken on 22 eastern zone, where the provincial capital Aǧcab di,
e
September 1968 and 24 May 1970. sparse villages and farmsteads are located. Mostly
– Georgia: two sets of imagery from mission DS1115 flat and under intensive crop, cotton and lucerne
recorded on the 14 September 1971 and 10 Sep- cultivation, this area stretches eastwards towards
tember 1971. the Aqgöl und the Sari Su lakes. Over the last two
In addition, the 3arc Shuttle Radar Topographic Map decades, after the collapse of the Soviet Union, fol-
(SRTM) with 90 m/pixel resolution provides elevation lowing the governmental agrarian reforms, the col-
data. lective and state farms were reallocated to private
Remote sensing enables the first regional en- people. Consequently, the agricultural lands were
vironmental, topographic and taphonomic assess- greatly parcelized: small plots and orchards re-
ments. It also supports potential archaeological fea- placed the larger fields of the late 1960s (Fig. 165).
tures detection. Ground checking is subsequently Targeted visits in this area documented the occur-
conducted to verify the signal. In addition, a combi- rence of archaeological artifacts, mostly scatters of
nation of survey techniques is applied in the field. pottery sherds, as modern occupation and irrigation
General reconnaissance is first conducted on a vehi- have leveled and removed ancient features for the
cle to gain acquaintance with the ground and to most part. This zone is, therefore, labeled a land-
identify areas of potential interest. Low intensity scape of destruction.308 Here extensive archaeologi-
pedestrian survey has been completed so far along cal researches could be conducted restrictedly, only
305 307
Tucker et al. 2004. Gaude 2005.
306 308
Galiatsatos et al. 2009; Kennedy 1998; Ur 2003. Taylor 1972; Wilkinson 2003, 41 ff.
by investing a great amount of time and energy or a stable landscape of survival,309 where numerous Fig. 163
The areas of investiga-
sampling specific areas for more intensive targeted archaeological sites and features have been so far tion along the Qaraçay
survey without little, if any chance of recording an preserved (Fig. 166). However, large portions of and Qarasu rivers
untouched original archaeological signature. land have started to attract interest for undertaking
Landscape of survival: the zone west of the agricultural activities, and since at least 2010 they
major irrigation canal, where modern occupation is have been widely ploughed. In this area, a danger-
only represented by isolated single or double farms, ous potential for a rapid and extensive destruction
locally called firma, is characterized by a gradually of the archaeological remains exists.
raising terrain that, moving westwards towards the
Qarabağ Mountain ranges, becomes more incised Kvemo Kartli
by the natural river courses. This zone is more arid,
not well suited for agriculture due to the presence The area of investigation extends in the lower Ma-
on the surface of conspicuous secondary loess de- s̆avera depression. Two zones have been selected
posits and it is, therefore, only locally irrigated by for the 2011 field activities.
deep wells. Sparse large fields, for the majority still
under state ownership, are cultivated with crops,
but also extended fallow lands covered by low wild
vegetation characterize this marginal environment: 309
Taylor 1972; Wilkinson 2003, 41–43.
Fig. 164
Mil Plain Survey area. A
Landsat image of the
investigated area with
the two zones on the
opposite sides of the
major modern irrigation
channel
"
Fig. 165
Mil Plain Survey area. A landscape of destruction between the villages of Salmanbejjai and Uzdzirkend, south of city of Aǧcab di.
e
On the left a Corona image (1110-1057DA111; 24 May 1970; courtesy of the USGS); on the right, a Google Earth image (28 July
2009). Note the growth of the inhabited areas and a profound reorganization of the field system
Fig. 166
Mil Plain Survey area. A landscape of survival immediately southwest of Kamiltepe. On the left a Corona image (1110-1057DA111;
24 May 1970; courtesy of the USGS); on the right a Google Earth image (28 July 2009). Note the absence of inhabited areas and
the appearance of only a few tracks and cultivated fields over the past 30 years
Fig. 167
The Aruchlo Plain. On the left a Corona image (DS1115-2154DF086; 20 Sept 1971; courtesy of the USGS); on the right a Google
Earth image (28 July 2009). Note the growth and profound reorganization of the field system
Fig. 168 The Aruchlo Plain ger villages (i.e. Nakhiduri, Khidiskuri and Tashtiku-
Intensive survey
on the Aruchlo Plain:
lari), while farmsteads are sparsely located over the
dispersion of sherds Fertile alluvial deposits compose the rather flat and entire plain. Modern intensive human activities
(preliminary mapping) triangular in form Aruchlo Plain, which extends have erased most of the archaeological landscape
some 7 sq km immediately northwest of the Mas̆a- features; only the tell settlements have at least
vera and Khrami river confluence. Similarly to large partially survived, while sites such as KKS 3, situ-
portions of the lower Mas̆avera Valley, the area is ated approx. 750 m to the northwest of Aruchlo 1,
intensively exploited for vegetables and cereals. have been completely flattened by recent farming
The average 400 mm annual rainfall is implemented practices. Intensive transect-based collection docu-
with artificial canals for constant water supply. In mented field scatters over the entire plain, echoing
particular, a major water supply feature irrigates the the intensive modern transformations of the area
plain by diverting water from the Mas̆avera River, (Fig. 168).
some 6 km upstream (Fig. 167). Possibly in associa-
tion with the construction of this large canal sys- T h e S̆ u l a v e r i P l a i n
tem, the field system of the Aruchlo Plain was ac-
cordingly re-organized with a northwest-southeast Located some 7 km downstream from the Mas̆avera
orientation and small- to middle-sized plots. At the River, the part of the S̆ulaveri Plain east of the ma-
northern and southern margins of the plain lie lar- jor North-South modern road presents only very
few sparse single houses. There, fields of cereals and ing the next summer season, efforts will be devoted Fig. 169
Mil Plain Survey. The
onions are larger in size than those on the Aruchlo to refine their chronological attributions. 46 sites – all periods –
Plain. They are mostly not artificially irrigated, but a documented during the
high rate of fluvial dynamics might have obliterated Intensive surveying around Kamiltepe 2010 and 2011 MPS
seasons, mapped
ancient remains.310 over a combined
For a detailed recovery of the archaeological re- CORONA (CORONA
mains around Kamiltepe, transect-base intensive 1110-1057DA111;
Results of the survey investigations collections were conducted over an area of 1 km in 24 May 1970) and
SRTM DEM 2002 image
The Mil Steppe Survey (MPS) radius around the site (Fig. 170). Transects were set (courtesy of the USGS)
radiating from Kamiltepe or, more rarely, parallel to
So far the MPS has documented a total of 46 sites the Qarasu River. Surveyors walked along two par-
dated from the Neolithic to late Medieval times. allel lines subdivided into series of 100 m long
Forty of them are new discoveries (Fig. 169; Tab. 23). blocks, collecting archaeological artifacts, mainly
At the present state of research, the dating of these pottery and lithics, infrequently brick fragments.
occupations remains tentative and preliminary. Dur- When possible, transects were set where the fields
offered better conditions of visibility. However, the
ground was often covered by compact soil and
310 Von Suchodoletz et al. 2011; pers. comm. von Suchodoletz dense wild vegetation – predominantly low artemi-
2012. sia bushes – as the majority of plots remain uncul-
Site MPS number – name Site Morphology Lenght (m) Width (m) Height (m) Preliminary chronological attribution
MPS 1 – Kamiltepe Mound % % % Neolithic, Iron Age; see excavation report
MPS 2 Shallow mound % % %
MPS 3 Shallow mound % % % Iron Age, Antique
MPS 4 Flat site % % % Neolithic
MPS 5 Low mound 70 70 1.5 Neolithic
MPS 6 Low mound 90 50 1 Neolithic and Medieval
MPS 7 Multiple mounds % % % %
MPS 8 Low mound 115 100 1 Neolithic?, Medieval
MPS 9 – Örenqal Walled site 630 630 % Medieval
e
MPS 10 Low mound 90 60 0.7 Neolithic, Iron Age, Medieval

MPS 11 Flat site 30 35 0 Iron Age, Antique
MPS 12 % % % % Iron Age, Antique, Medieval
MPS 13 Flat site 30 30 0 Medieval
MPS 14 Low mound 100 100 1 Neolithic, Chalcolithic, Late Antique?,
Medieval
MPS 15 Hill top site 140 60 0.5 Iron Age, Medieval
MPS 16 Tell 230 130 3 Neolithic?, Chalcolithic, Bronze Age?,
Iron Age, Antique, Medieval
MPS 18 Multiple mounds 490 220 % Neolithic, Chalcolithic?, Late Antique,
Medieval
MPS 19 Tell 135 110 6 Neolithic, Chalcolithic?, Bronze Age,
Medieval?
MPS 21 Multiple mounds % % % Neolithic, Bronze Age, Medieval
MPS 22 – Gurtepesi Low mound 60 60 1 Medieval
MPS 23 – Bölle Tepe Low mound und 230 320 1.5 Neolithic, Iron Age, Medieval
lower flat part
MPS 24 % % % % Medieval
MPS 25 Low mound 40 40 1 Medieval
MPS 26 Low mound 40 30 0.5
MPS 27 Low mound 50 50 0.5
MPS 28 – Mehrabli Tepe % % % % Iron Age, Late Antique, Medieval
MPS 29 Low mound 50 40 0.5 Iron Age, Medieval
MPS 30 Low mound 120 70 1 Iron Age, Late Antique, Medieval
MPS 31 Low mound 35 35 0.5 Iron Age
MPS 33 Low mound % % 3 Medieval
MPS 34 Flat site 40 30 0 Late Antique, Medieval
MPS 35 Low mound 90 70 1 %
MPS 36 Shallow mound % % % Neolithic, Medieval
MPS 37 % % % % Neolithic, Medieval
MPS 38 % % % % Late Antique?, Medieval?
MPS 39 Low mound 200 90 1 Iron Age, Late Antique, Medieval
MPS 40 Low mound 250 130 0.5 Iron Age, Late Antique, Medieval
MPS 41 Flat site 50 50 0 Late Antique, Medieval
MPS 42 Shallow mound 350 150 1 Neolithic, Medieval
and flat site
MPS 43 Flat site 190 200 0 Neolithic, Medieval
Tab. 23
Gazetteer of surveyed MPS 45 % % % % Neolithic, Late Antique, Medieval
sites and their MPS 46 Low mound 40 40 2 Neolithic, Medieval
preliminary dating
tivated. Some areas, especially those located to the of 18 sites (Fig. 171). They are all situated along Fig. 170
Mil Plain Survey. Sites
east, were inaccessible due to the presence of crops the Qaraçay, whereby MPS 21 is found close to the and dispersion of diag-
or onion fields. These conditions made the detec- confluence of this river with the Qarasu riverlet. nostic and non-diag-
tion of archaeological finds generally difficult, and These Neolithic sites are often situated along the nostic pottery sherds
in the area around
in the future further transects might be collected, if margins of the modern river valley, on low natural Kamiltepe, mapped
fields present better ground visibility and the area ridges or on river terraces that slope down towards over a CORONA
of intensive collection could possibly be extended the Qaraçay plain. They are usually found every 1 1110-1057DA111 image
(24 May 1970; courtesy
beyond the 1 km radius. Nevertheless, this inten- to 1.5 km, as for example MPS 16, MPS 18, MPS of the USGS)
sive collection enabled the recording of areas of 19, MPS 21, but closer locations are documented
higher concentrations of artifacts and the discovery too, as in the case of MPS 36 and MPS 37
of three new sites: one cluster of artifacts, named (Fig. 172), or of the three adjacent sites MPS 42,
MPS 4, which turned out to be an archaeological MPS 43 and MPS 45. The former is composed of a
site, although no mound formation could be ob- shallow mound (Fig. 173) and a large extended flat
served; MPS 5, a small low mound of circular form, area. Most of the sites are small, ranging between
located ca. 850 m to the northwest of Kamiltepe and 0.5 and 1 ha in size, but always abundant is the
immediately north of a small stream; and MPS 13, a amount of ceramics present on the surface. This
flat Medieval occupation site situated along the Qa- Neolithic presence is often recorded on low mounds
rasu River, some 1.1 km to the south of Kamiltepe. with shallow deposits, and is often not covered by
later occupations.
The neolithic landscape
MPS 18
Along the Qaraçay River, the major natural water
course in the area around Kamiltepe, a dense Pot- This site is the largest prehistoric presence so far
tery Neolithic occupation is documented by a series recorded by the MPS. It is apparently composed of
Fig. 171
Mil Plain Survey. The Neolithic sites documented during the MPS 2010 and 2011 seasons, mapped over a combined CORONA (CORONA 1110-1057DA111;
24 May 1970) and SRTM DEM 2002 image (courtesy of the USGS)
Fig. 172
Neolithic sites MPS 36
and MPS 37 located at
a very close distance
to each other on
a terrace overlooking
the Qaraçay Valley
(photo A. Ricci)
a series of small low mounds that form a large and

complex multi-mounded site (Fig. 174). Pottery and
lithic fragments cover an area of over 8 ha on a low
natural ridge some 4.8 km southwest of Kamiltepe.
Sherds include fragments comparable to examples
found at Kamiltepe, but others do not find parallels
in the thus far known assemblage of the regional
Pottery Neolithic (Fig. 175). This suggests that the
occupation of this site lasted for a long period of
time. Possibly the entire area of MPS 18 was not
occupied simultaneously; rather, shifting, consecu-
tive and smaller occupations resulted in the actual
large dispersal of artifacts and small distinct mounds.
Magnetometric prospections were conducted on the
site,311 and also four sections exposed by modern
destructions were cleaned and documented. In par-
ticular, section 1801 retrieved the most interesting
findings (Fig. 176). This section is approximately
315 cm long and up to 140 cm high. Here, some role for the local 6th mill. BCE communities. Future Fig. 173
40–50 cm below the modern surface, clear anthro- excavations on MPS 18 might provide useful data shallow mound of site
pogenic deposits were observed, although in the for the constructing a fine chronology of the regio- MPS 42 in front of site
upper part painted and undecorated sherds were nal Pottery Neolithic. MPS 19 (photo A. Ricci)
also collected (Fig. 177), and one geometric painted
sherd shows clear analogies with Kamiltepe exam- MPS 36
ples and possible stylistical resemblances also with
the Hajji Firuz assemblage.312 In the southern (left Located approximately 6.5 km upstream from Ka-
in the picture) corner of the section a wall com- miltepe, this site lies on a high alluvial terrace that
posed of four unbaked mudbricks placed on top of overlooks the right bank of the Qarasu River
each other, reaching a total height of 51 cm, were (Fig. 172). On the surface abundant pottery sherds
visible. To the right of this wall a still completely were collected, the majority being Neolithic in date
preserved jar (Fig. 178) was standing in vertical po- (Fig. 179). It was impossible to determine the exact
sition on a floor level, which appeared darker in extent of the site, which has been partly but deeply
color than the immediately overlying stratum. The damaged by modern installations. These have also
vessel is irregular in shape and shows a chaff sur- exposed an approximately 120 cm and 550 cm long
face with a flat base and an uneven simple rim section. Here, some 35 cm below the modern sur-
sharply folded inwards towards the inside on one face, anthropogenic deposits composed of a series
side. Similar rims have been documented at Kamil- of burnt and ashy layers were visible (Fig. 180). No
tepe.313 To the right of this jar, grinding stones and built structure appears in this section, suggesting a
larger pottery fragments were present. Three un- possible, rather ephemeral occupation of at least
baked mudbricks are found some 50 cm to the right this part of the site.
of the vessel. Possibly, the two small mudbrick
walls delimited a small room on the floor of which Other neolithic sites
the jar was standing. In addition, in the northwest-
ern corner a group of at least four burned mud- Furthermore, two other sites (MPS 23 and MPS 46)
bricks was found. These mudbricks were no longer yielded an abundance of Neolithic material. They are
in situ, but in a shifted position. Due to the optimal located in the river basin to the North (the Xanasen-
conditions of preservation of this context and çay), where only the first, very preliminary general re-
thanks to the presence of abundant materials, MPS connaissance by car was conducted. Both sites are
18 emerges as an ideal candidate for conducting distinct mounds round in form, and they do not ex-
further detailed investigations. These might shed ceed 2 metres in height (Figs. 181; 182). More de-
light on the characteristics and duration of occupa- tailed investigations will be conducted in this area
tions at this site, which might have played a focal during the next field season (summer 2012).
This newly acquired data hints at a 6th mill.
BCE landscape dotted with small camps or ephem-
311 eral occupations with only a few more stable sites,
312
Voigt 1983, 94–98. which were possibly inhabited and in use for longer
313
D‘Anna, in this article. periods of time, such as Kamiltepe with its massive
Fig. 174
Mil Plain Survey. Topo-
graphic mapping of site
MPS 18 (map A. Kwast)
Fig. 175
Mil Plain Survey.
Neolithic sherds
collected on site
MPS 18 (photo A. Ricci)
Fig. 176
Mil Plain Survey. Site
MPS 18: section 1801
(photo A. Ricci)
Fig. 177 platform and possibly MPS 18 with built structures,

Mil Plain Survey. Site
MPS 18, Neolithic
but with a shifting on-site occupation. The lack of
sherds collected in the architecture as evidenced in the exposed section at
upper part of section MPS 36 suggests the presence of temporary camps
1801 (photo A. Ricci)
or short-living settlements. This landscape of highly
mobile Neolithic communities with an incipient re-
cognition of pivotal sites did not actually lead to a
later conspicuous tell formation phenomenon. So
far information on sites of later periods remains ex-
tremely scanty. While site MPS 16, a shallow tell
located on a flat ground on the right bank of the
Qarasu River alluvial plain, some 50 m away from
the actual course of the river, provides Chalcolithic
materials,314 possibly only at site MPS 19 and
Fig. 179
MPS 21 dubious Bronze Age materials have been Mil Plain Survey. Neolithic sherds collected on site MPS 36 (photo
A. Ricci)
collected. Future investigations on these shallow

anthropogenic mounds can add information on
the occupation of the Qaraçay River basin during
several periods of time. And prospect survey will
need to further include the banks along the Qara-
su River and the areas between and away from
the two streams in order to verify possible Neo-
lithic presence in these zones as well.
The Medieval landscape

A series of small Medieval sites have been docu-
mented along both the Qaraçay and Qarasu rivers.
Striking is the presence of seven small sites (MPS
25, MPS 26, MPS 27, MPS 29, MPS 30, MPS 31,
MPS 32, MPS 33) along the eastern bank of the
Qarasu. They are small occupations, ranging be-
tween 0.1 and 0.2 ha in size, but a shallow mound-
Fig. 178
Mil Plain Survey.
ing, differences in soil color and distinct pottery
Site MPS 18, The com- dispersals make them clearly identifiable in the
plete jar discovered landscape and distinct from one another. Other
during the cleaning
of section 1801 on site
possibly contemporaneous occupations – such as
MPS 18 (photo A. Ricci) MPS 7, MPS 8, MPS 13 or MPS 10 – are found
farther downstream. The nature of these small sites
314
Helwing, in this volume. still has to be clarified, and we are aware that at
least some of them might have been used as a

small cemetery. In any case, they hint to a sparse
occupation of farmsteads or rural settlements. Pos-
sibly they were connected to major centres located
farther East, as for example Örenqal , located some
e
10 km away in this direction. This site is a fort,
double-walled, square in form, ca. 630 " 630 m in
size and surrounded by a ditch. Örenqal was lar-
e
gely investigated during the 1960s by the A. A. Ies-
sen expedition.315 The site was once a major city
among a series of citadels located along a major
irrigation scheme, between Qalatepe and the Araxes
River. Two linear features, possible hollows, are
visible on the CORONAs (Fig. 183) and partly still in
the present-day landscape (Fig. 184). They both di-
vert from Örenqal and lead towards the Qaraçay/
e
Qarasu basin. The northern one turns southwards
around site MPS 7, and might be related to a further
linear feature that runs intermittently along the Fig. 180
eastern bank of the Qarasu, some 50 to 100 m East Mil Plain Survey. The section on site MPS 36 (photo A. Ricci)
to the series of small sites located in that area. For
the next seasons, a program is also planned to in-
vestigate this historical landscape and to address
forms of ancient and historical water managements
in the Mil Steppe.
The Kvemo Kartli Survey (KKS)
During the rather short first survey season con-

ducted in September 2011 in the Marneuli Region,
a total of 13 sites were documented. In the Aruchlo
Plain eight sites were mapped (Fig. 185). These in-
clude the already known Aruchlo and Aruchlo 2, but
also KKS 3, a site located some 800 m northwest of
Aruchlo, from which possibly prehistoric materials
were retrieved, as mentioned in earlier reports.316
Three mound sites (KKS 5, KKS 6, KKS 8) are situ-
ated along the Mas̆avera’s alluvial terrace at the
Fig. 181
southern margin of the Aruchlo Plain; a fourth small Mil Plain Survey. The shallow mound of site MPS 23 from a distance (photo A. Ricci)
tell (MPS 4) is located in a more inland position.
These three archaeological sites range between 0.2
to 0.5 ha in size, and are located approximately
550–750 m apart from each other. KKS 8 (Fig. 186)
might yield some early material, whereas the re-
maining sites can be dated to a later date, from the
Iron Age to the late Medieval period.
Three sites were discovered in the S̆ulaveri
Plain (Fig. 187). KKS 9 is thought to be a small kur-
gan, and it has been bulldozed in the central part.
Unfortunately, only a few pottery fragments were
collected; large squarish limestone slabs could be
observed in their original in situ context. Located in
a large corn field, KKS 11 shows typical Kura-Araxes
315
¨*ææ*B 1959. Fig. 182
316
8$B\/*- 1982, 106. Mil Plain Survey. Site MPS 46 (photo A. Ricci)
Fig. 183
Mil Plain Survey.
The linear features
between MPS 8 and
the Qarasu Valley on
a CORONA image
of 24 May 1970
(1110-1057DA111;
courtesy of the USGS)
Fig. 184
Mil Plain Survey. Traces
of a possible linear
hollow immediately
south to site MPS 7
(photo A. Ricci)
Fig. 185
The sites surveyed
on the Aruchlo Plain
in 2011
Fig. 186
Site KKS 8
(photo A. Ricci)
Fig. 187 material. Unfortunately, it is impossible to deter- Conclusions

The sites surveyed
on the S̆ulaveri Plain
mine the extent of this site, since modern agricul-
in 2011 tural activities had completely flattened it even be- The preliminary survey activities conducted in sum-
fore the late 1960s, as documented by the CORONA mer 2010 and 2011 have shown the potential and
imagery at our disposal. Finally KKS 12 is found in the necessity of conducting further intensive investi-
a more inland position, upstream along the S̆ulaveri gations in both research regions.
çay. The site is situated on a high natural hill, but In the Mil Steppe, profitable survey research
only a handful of material was collected due to low can be carried out especially in the westernmost
ground visibility. zone, where the archaeological landscape is well
Generally speaking, sherds, lithic fragments preserved. It offers a unique opportunity to investi-
and mudbricks were found abundantly on the surfa- gate patterns of settlement, human-environmental
ce where modern disturbances have reached the dynamics and address questions related to socio-
anthropogenic deposits of the KKS sites, whereas economic practices and decision-making processes
material was scarce when compact soil covered the in settlement choices. In this area, further investiga-
sites, a condition probably due to formation and tions should concentrate both on continuing sur-
taphonomic processes of the sites and by a low veys along the Qaraçay and Qarasu river basins as
artifact density. Furthermore, the presence of high well as in areas adjoining them. Such studies should
grown late summer vegetation hid large portion of further be integrated with geomorphological investi-
the ground. gation.
Fig. 188
Visualization of the
Capsian sea level at
5 ka BCE (sea level
data taken from Hoo-
gendorn et al. 2005)
In the Mas̆avera Depression, investigations and Aruchlo in Eastern Georgia, were only briefly
will be conducted to document the overall occupa- visited. The Mil Steppe is a plain situated between
tion of the valley. If possible, field activities will the confluence of the Kura and Araxes rivers, in
take place during other seasons of the year, as for southern Azerbaijan. The western boundary of the
example in early spring after the first ploughing of Mil Steppe is determined by the Qarabağ Moun-
the field, and before sowing, when less dense vege- tains, which mainly consist of Mesozoic sedimentary
tation covers the surface. This should enable a high- rocks of volcanic origin, like tuff conglomerates,
er ground visibility, i.e. better conditions for surface clayey slates and tuff sandstones.317 Several small
collections. rivers drain the mountain area and run in a north-
Moreover, the same landscape archaeology eastern direction as tributaries to the River Kura. At
approach will be implemented on the area around the beginning fo the 6th mill. BCE the Caspian Sea
Mentesh Tepe, to verify whether further prehistoric level was nearly 7 m higher than today. Consequently,
sites still exist in that region too. extensive areas of the present day floodplain of the
Kura River lay below sea level (Fig. 188).318 The lo-
Ancient and modern landscape along the Qaraçay cal climate in the forelands of the Qarabağ Moun-
and the Mas̆avera valleys – first results tains corresponds to a humid subtropical climate
(Cfa-climate according to the Köppen-Geiger classifi-
Wiebke Bebermeier, Frank Schlütz and Arian Goren cation) and shifts in the floodplain of the Kura River
to a semi-arid climate (BSk, Köppen-Geiger).319 At
Introduction Agdam weather station (252 m asl), 35 km north-
As a pilot study in the reconstruction of ancient

landscape around archaeological sites, geomorpho- 317
logical field work concentrated around the site of Caucasus map – Cyrillic 1:500 000.
318
Hoogendorn et al. 2005.
Kamiltepe in the Mil Steppe, while the two other 319 Kottek et al. 2006 Cfa: warm temperate, fully humid, hot sum-
research foci, the Tovuz area with Mentesh Tepe, mer; BSk: arid, steppe, cold arid.
west of Kamiltepe, the annual average temperature ductivity and total and inorganic carbon contents
is 12.3 # C, and annual rainfall averages 447 mm. were determined. The evaluation of the laboratory
The mean monthly temperatures do not fall below analyses and OSL procedures are still in progress
0 # C throughout the year.320 and will not be presented within this paper.
Since summer 2010 the landscape develop- The chronological frame of the profile MPSA1
ment in the hinterland of Kamiltepe has been inves- is based on AMS radiocarbon dates. Analyses were
tigated with a geoarchaeological approach in order carried out at the Curt Engelhorn Zentrum Archäo-
to establish a model of landscape evolution and metrie GmbH Radiocarbon Laboratory in Heidel-
palaeoenvironmental conditions in the Mil Steppe. berg, Germany. The software Oxcal (v. 4.1) and
Such knowledge may contribute to a better under- the IntCal09 calibration curve by Reimer et al.323
standing of the living conditions of the investigated were used for the calibration of the radiocarbon
archaeological cultures. It appears that the Neolithic dates.
occupation in the Mil Steppe was rather short-lived
and that it underwent a dramatic decline during the Sediment profile MPSA1
later half of the 6th mill. BCE, with little evidence for
permanent occupation during the 4th and 3rd mill. MPSA1 is located at an undercut bank of the River
BCE.321 The causes for such a rapid decline still re- Qaraçay, approximately 900 m southeast of Kamil-
main unclear.322 One goal of the landscape archae- tepe (Fig. 189). The profile reached a depth of 4.62 m
ological research within the project is to evaluate if and could be subdivided in nine major sediment
and how changes in the palaeoenvironmental con- sequences:
ditions (for example, trends towards drier or cooler – Unit I [0–69 cm depth]: yellowish fine sandy silt,
conditions) altered the living conditions of the inha- granular structure, well compacted, dry
bitants in such a way that changes in settlement in- – Unit II [69–95 cm depth]: pale brownish silt,
tensity or locational choices would have been a massive structure, well compacted, dry
consequence. – Unit III [95–140 cm depth]: pale brownish clayey
silt, top: massive structure, at the bottom a
Landscape Archaeology poorly developed prismatic structure
– Unit IV [140–143 cm depth]: yellowish brown
Wiebke Bebermeier and Arian Goren silt, good compacted, massive structure
– Unit V [143–196 cm depth]: pale brownish clay-
Methods ish silts, dry, poorly developed prismatic struc-
ture
The reconstruction of the landscape development is – Unit VI [196–292 cm depth]: alternating layers of
based on the analysis of sediment archives found yellowish silt and grayish brown clay
in alluvial fans, endorheic basins or alluvial flood- – Unit VII [292–342 cm depth]: alternating layers
plains. Within this project, sediment archives were of yellowish silt and grayish brown clay, layer
investigated in order to derive palaeoenvironmental boundaries are disturbed and contorted
proxy data. The sediments were acquired from ex- – Unit VIII [342–386 cm depth]: grayish brown clay
posed natural or artificial outcrops or extracted using – Unit IX [386–462 cm depth]: pale brownish fine
an Edelmann hand-auger. The sediment sequences sandy silt, granular structure, well compacted
of the outcrops were described macroscopically The chronological frame is given by seven radiocar-
with respect to parameters such as grain size, color, bon ages (Tab. 24).
organic content and compaction, and other fea- The sediment units consisting of yellowish to
tures, such as charcoal, carbonate concretions etc. brownish silt (units I, II, IV and IX) are interpreted
Sampling was done at 10–20 cm intervals, extract- as reworked loess, which was eroded in the upper
ing 2 cm thick bulk samples. Very thin sediment catchment of the Qaraçay River and transported
layers were sampled at 1 cm intervals. Furthermore, and deposited by the river. Also sediment units III
4 samples for optically stimulated Luminescence and V are composed of reworked loess, but the
dating (OSL) were taken. content of clay within the grain-size distribution
Mineralogical and geochemical analysis is and the poorly developed prismatic structure indi-
conducted at the Physical Geography Laboratory, cate the beginning of post-depositional soil forma-
Department of Earth Sciences, Free University Ber- tion processes and, thus, point to a stabilisation of
lin. Furthermore, bulk parameters like electrical con- the geomorphodynamic conditions.324 The alternat-
ing layers of silt and clay point to recurring shifts in
320
FAO Rome 2001.
321 323
Ricci, in this article. Reimer et al. 2009.
322 324
Aliyev/Helwing 2009; Helwing 2009. Muhs et al. 2004.
Fig. 189
Mil Plain. Lithostrati-
graphy and location
of profile MPSA1
the environmental conditions: while clay is depos- sists of clay and silt. The boundaries of the sedi-
ited under a stillwater environment, the silty layers ment layers are disturbed and contorted. These
represent the deposition of reworked loess and, characteristics may point to a former periglacial en-
hence, point to fluvial processes. Unit VII also con- vironment, in which permafrost and cryoturbation
14 14
misia, Liguliflorae and the Aster-type as well as some
Lab. Depth Material C age BCE C age midpoint
No. (cm) (uncalibrated) (2!) BCE
spores of the Ustilago-, Glomus- and Riccia-type.
Their low concentration diminished farther down,
12422 162 charcoal 16,960 ) 194 18,280 ) 675 and no pollen grains or spores were found below
12426 184 charcoal 8,855 ) 43 8,010 ) 210 250 cm.329
A preliminary interpretation of the stratigraphy
12421 206 charcoal 7,277 ) 33 6,150 ) 75
of the profile allows some conclusions to be drawn
12420 218 charcoal 8,248 ) 35 7,300 ) 155 about the landscape development in the floodplain
12425 245 charcoal 10,430 ) 50 10,360 ) 220 of the River Qaraçay. Sediment characteristics of
the lowest unit point to a phase of fluvial activity
Tab. 24 12419 252 charcoal 7,724 ) 38 6,550 ) 80
AMS radiocarbon with the deposition of reworked loess transported
ages for sediment 12423 372 charcoal 6,748 ) 32 5,670 ) 50 downstream from the upper catchment area. The
profile MPSA1 overlying clay unit points to rather stillwater condi-
tions, according for example to an oxbow lake of
the River Qaraçay, which was later regularly af-
occurred.325 Comparable to the clay layers in unit fected by flood events (unit VI). The soil formation
VI, the grayish to brownish clay in unit VIII points phases in unit III and V point to low flood activities,
to a deposition in a stillwater environment. which might be related to a meandering of the riv-
The AMS radiocarbon dates show several age erbed to a more distant part of the floodplain.
reversals. The oldest radiocarbon age dates sedi- Altogether profile MPSA1 reflects a very dy-
ment unit V, at a depth of 1.62 m to 18,280 ) 675 namic river system as a result of frequent changes
BCE. The youngest age is determined for a depth of in environmental conditions in its catchment area.
3.72 m, dating unit VIII to 5,670 ) 50 BCE. The high The River Qaraçay was most likely, until the con-
magnitude and frequency of the age reversals duction of hydraulic measures during the Soviet
raises the question as to how reliable these dates period, a braided to meandering river system with
are. The following example should illustrate this frequent dislocations of branches. The question as
problem. Layer VIII is dated to 5,670 ) 50 BCE and is to whether changes in the palaeoenvironmental
located directly below layer VII, which shows char- conditions triggered the settlement gaps will remain
acteristics of cryoturbation. Sedimentological and unanswered until a valid chronology of profile
palynological proxies obtained from sediment cores MPSA1 is established. However, the results from
of Lake Van in eastern Anatolia (Turkey) show a Lake Van suggest that during the Neolithic period a
maximum in the Holocene tree pollen value and a climate optimum prevailed.
high percentage of organic matter preservation for
the period between 6200–4000 BCE.326 The authors Sediment profiles MPSA2 and MPSA3
conclude that the Holocene climatic optimum with
warmer and wetter conditions is represented in this Three kurgans – the site of Üç Tepe – are located
data.327 The possibility of having more favourable approximately 11 km north of Kamiltepe. The north-
climatic conditions than today in eastern Turkey ernmost kurgan had been excavated by the Iessen
and permafrost conditions north of the Caucasus Expedition and yielded a Late Antique burial in an
Mountains simultaneously is inconceivable under a older, possibly Bronze Age kurgan.330 Parts of this
climatological perspective. kurgan were further destroyed and bulldozed by an
Analysis on the source plants of the charcoal excavator as part of infrastructure earthworks in
using the 213C values328 and an evaluation of dislo- 2010. The lowermost 2 m of the kurgan were ex-
cations of organic material by soil biota have to be posed, but not affected by the construction work.
carried out, before a valid age model for the profile The diameter of this kurgan totals 40 m. Two pro-
can be established. The upcoming OSL dates will files were described in order to document the stra-
contribute to this task. However, the numerous age tigraphy of this archaeological site and to obtain
reversals point to a fluvial environment leading to palaeoenvironmental proxy data from the buried
the entry of ‘‘old wood’’ from the slopes of the upper palaeosoil.
catchment area. Both profiles present very similar characteris-
The palynological analyses of MPSA1 revealed tics with regard to sedimentological parameters like
only a few pollen grains of Chenopodiaceae, Arte- grain-size, soil compaction and structure, but they
325
Bockheim/Tarnocai 1998.
326 329
Landmann 1996. For the methods and further results, see Schlütz, in this arti-
327
Landmann 1996. cle.
328 330
Schütt et al. 1997. ¨*ææ*B 1965.
were differently affected by fire. The lithostratigra- (dry) steppe used as grazing ground before the Fig. 190
Mil Plain. Lithostratigra-
phy of MPSA3 is presented in the Fig. 190: construction of the Tepe. As long as laboratory ana- phy of MPSA2 (a) and
– Unit I [0–181 cm depth]: very well compacted lysis and radiocarbon dating are still in progress, a MPSA 3 (b); (c) shows
yellowish silt, massive structure, rooted, carbo- further interpretation must be postponed. a cross profile through
the nearly almost
nate concretions up to 1 cm. destroyed kurgan
– Unit II [81–145 cm depth]: very well compacted Aruchlo in diameter
silt, which was affected by fire, few carbonate
concretions. Layers and lentils of burned material The Neolithic mound of Aruchlo is situated on the
are embedded within this unit; their degree of uppermost terrace of the River Mas̆avera, approxi-
burning is represented by different colors. mately 10 m above the present-day riverbed
– Unit III [145–210 cm depth]: poorly compacted silt (Figs. 191). A transect of three drillings was con-
with a prismatic structure, carbonate precipitation ducted in a North-South direction along an irriga-
along the prism columns, charcoal and snails. tion canal from the tell towards the River Mas̆avera.
The palynological samples from and below the The same methodological approach as in Azerbaijan
buried soil horizon (Figs. 190: 120–125 cm) were was applied to the extracted sediment profiles. A
dominated by Chenopodiaceae, but were also quite chronology of the drillings has not been established
rich in other pollen types like Artemisia, Poaceae, yet.
Potentilla-, Anthemis-, Aster-, Polygonum aviculare- The layer of the three profiles B1–B3 can be
type, Fabaceae, Malvaceae, Erodium and Ephedra parallelized and distinguished into three sedimen-
fragilis-type. Their relatively rich spore contents tary units. Profile B2 is representative for all three
included spores of coprophilous Ascomycetes profiles:
(Sporormiella-type, Delitschia), smut fungi (Uro- – Unit I [0–220 cm depth]: grayish loamy clay, inter-
cystis-type) and some spores of the Glomus-type. calations of thin layers of poorly rounded detrials
This combination of pollen and spores point to a smaller than 0.5 cm in depths of 55 cm and 83 cm.
Fig. 191 – Unit II [220–253 cm depth]: pale grayish to search an interpretation of this unit must be post-
Mas̆avera area.
Lithostratigraphy
brownish fine sandy silt poned for the future.332
of the drillings B1–B3 – Unit III [253–440 cm depth]: grayish loamy clay,
in the lowermost 10 cm poorly rounded detrials Geoarchaeology and pollen studies
smaller than 0.5 cm.
Units I and III are interpreted as fluvial deposits Frank Schütz
predominantly of the Mas̆avera River or one of its
former tributaries. The layers of poorly rounded det- Methods
rials in the drillings point to short transportation
distances.331 It is hypothesized that the riverbed of During the 2010 field work in the Mil Steppe and
a former small creek was used to build the irrigation around Aruchlo 150 samples were collected from
canal and that this creek deposited the detrials dur- archaeological trenches and 73 samples from soil
ing flood events. The sandy silt in unit II is prelimina- drillings and from the excavated sediments at the
rily interpreted as reworked loess, either correlated
to colluvial or fluvial processes. Without further re-
332
Late Pleistocene to Holocene fluvial dynamics of central South
Caucasia are being investigated at the University of Leipzig,
Institute of Geography as a contribution to the palaeoenviron-
mental history of the region. First results of the DFG-funded
331
Mills 1979. project are published in Von Suchodoletz et al. 2011.
River Qaraçay (MPSA1) and a Bronze Age kurgan semi-desert of the region, are represented by val-
(MPSA2, MPSA3). In addition, eleven surface sam- ues of up to 2%. Pocaceae pollen grains that can
ples (plant litter and soil surface) and 14 samples be denoted by their size and pore morphology to
of recent pollen (flowers) and spores (fungi) for the Cerealia-type did not exceed 4%. Pollen types
type slides were collected. Three short cores from indicative for open ground and/or trampling found
lake and fluvial sediments of in total 350 cm in are the Plantago lanceolata-, Polygonum aviculare-
length were received and sampled in the laboratory. type and Erodium. Spores of the Glomus-, Sporor-
About 134 samples have been processed in miella- and Podospora-type were generally com-
the HF (hydrofluoric acid) pollen laboratory of the mon. Species of Glomus are living in symbiosis with
Topoi Excellence Cluster Department of Earth the roots of grasses and herbal plants. Their occur-
Sciences, Institute of Geographical Science at the rence reflects that soil was collected with the sur-
Freie Universität Berlin, and 80 of them have been face samples. The Sporormiella- and Podospora-
qualitatively analysed by now. The type slides will type are often viewed as grazing indicators, since
be prepared in a different laboratory at the Univer- fungi producing spores of these types are common
sity of Göttingen. on animal dung. Nevertheless, other kinds of de-
The results are summarized here and prelimin- caying organic plant material are inhabited by those
ary conclusions are given. The identification and species as well. For instance, Asgari and Zare338 re-
nomenclature are mostly based for pollen grains on port species of the genus Preussia (Sporormiella-
type slides and Beug,333 and for spores on van type) from dead cereal leaves from Iran. Beside pol-
Geel et al.,334 van Geel and Aptroot,335 and Ellis and len and spores, even shells of the single-celled Tes-
Ellis.336 taceae were found. The living of this kind of amoe-
bae is restricted to the availability of water.
Surface samples (OP) The surface sample OP 109 taken at the site
of profile MPSA1 is dominated by Artemisia accom-
From the area around Kamiltepe four surface sam- panied by pollen of Poaceae, Liguliflorae and Car-
ples have been so far analyzed. Three of them were duae, well reflecting the recent Artemisia steppe
from harvested wheat fields used as pastures for with some grasses and Asteraceae including thistles.
sheep and goat at the time of sampling and rich in From the area of the natural reserve at the
weeds especially in wormwood (Artemisia spec.). Lake Ag Göl (Fig. 188) two surface samples were
Sums of 100–120 pollen grains of Artemisia and analyzed. Sample OP 127 taken in a dry bed of an
Chenopodiaceae (goosefoot family) were counted ephemeral river a few hundred meters away from
to determine the so-called Artemisia/Chenopodia- the lake was dominated by Chenopodiaceae, possi-
ceae (A/C) ratio, that is often used to distinguish bly due to the salty character of the site, while Arte-
between deserts with low A/C ratios and steppes misia and Poaceae were less common. With saws
with high A/C ratios. Samples OP 100 and OP 115 of Chironomids, shells of Arcella (Testaceae) and
taken between Kamiltepe and site MPS 4 reflect the remains of Foraminifera, this sample also con-
the man-made Artemisia semi-desert coming up tained zoological microfossils used in palaeoecolo-
after harvest in their pollen spectra by high A/C ra- gical studies. The OP 129 was taken from the sedi-
tios of 10 and 60, respectively. On the opposite, ment surface of the dried-up Ag Göl near a belt of
sample OP 119 from near a group of Chenopodia- Chenopodiaceae at the lakeshore from a phase of
ceae family plants showed an A/C ratio of only higher lake level. It was consequently dominated
0.14. Sample OP 95.5 from inside a thistle thicket by pollen of Chenopodiaceae, but also yielded pol-
at the east side of Kamiltepe revealed a moderate len of Artemisia, Poaceae, Fabaceae, the Plantago
A/C ratio of 2.7. In all samples the amount of tree lanceolata-type and a few of trees (Pinus, Carpinus
pollen (in diminishing order: Carpinus betulus, Os- betulus, Quercus). The later were possibly brought
trya-type, Pinus sylvestris-type, Quercus, Alnus, Ul- from abroad by long distance wind transport, as
mus/Zelkova, Myrtaceae) was more or less distinctly the vicinity of the lake is empty of forests.
below 5%, as is typical for the semi-deserts in the This study on surface samples proves the
Caucasus lowland.337 The bushes of Capparis spi- strong influence of human land use on pollen
nosa (caper bush), characteristic for the Artemisia spectra archived in sediments. For instance, the
values of the A/C ratios of nearby surface samples
span from rates pointing to Artemisia steppe to
those typical for Chenopodiaceae deserts, depen-
333
Beug 2004. ding on the actual land use practise. This demon-
334
Geel et al. 2003.
335
Geel/Aptroot 2006.
336
Ellis/Ellis 1988; Ellis/Ellis 1997.
337 338
Connor et al. 2004. Asgari/Zare 2010.
strates the high potential of palynological analyses level, some pollen grains of Artemisia and the An-
of suitable archives to reconstruct former human themis-type were found in half of the six analyzed
impact. samples. Spores of the Glomus-type were more fre-
However, the data point to possible difficul- quent than pollen grains. Charcoal particles became
ties in reconstructing former climate changes. For abundant in the uppermost 20 cm below the floor
comparison and to disentangle the influence of cli- level. The yellow floor layer itself yielded only char-
mate and humans on the vegetation in former times, coal particles. The material deposited on that floor
sediment archives (e.g. lakes) from remote archaeo- was rich in pollen of Erodium (stork’s bill), and also
logical settlements are needed. some grains of Liguliflorae, Chenopodiaceae and
Poaceae occurred. Spores were presented only by
Outcrops and drillings those of the Riccia-type. In spite of the HF treat-
ment during the pollen preparation, even some
The profile MPSA1 (Fig. 189) from the River Qara- dendritic phytoliths were found, possibly the resi-
çay was sampled about every 25 cm during the dues of a concentration of grass or straw material.
field campaign. As mentioned above, pollen (Che- Pieces of charred grass epidermis in one of the next
nopodiaceae, Artemisia, Liguliflorae, Aster-type) samples point towards the same direction. The fill-
and spores (Ustilago-, Glomus-, Riccia-type) were ing above the second floor was rich in pollen of
rare in the upper part of the profile and disap- Poaceae and the Cerealia-type accompanied by
peared totally below 250 cm. Chenopodiaceae and the Anthemis-type. The next
In the silty loess-like sediments of a 5 m pro- samples above yielded Chenopodiaceae, Erodium,
file (EB 03) drilled with an Edelmann hand-auger in Liguliflorae, Pinus and the Riccia-type. Again burnt
a slight depression between MPSA1 and Kamiltepe, fragments of grass epidermis appeared.
the presence of pollen and spores strongly dimin- In 2009 and 2010, five samples for palynolo-
ished from the top to the bottom and fully disap- gical analyses were taken by Neçef Museybli (Na-
peared below 200 cm. tional Academy of Sciences in Baku) from layers
For the palynological results from the profiles dating to the first half of the 6th mill. BCE at the site
MPSA2 and MPSA3 see above. of Hesensu, near Agstafa. The sample 320–340 cm
and an ash layer at 180 cm from square 6 as well as
Archaeological trenches the sample 280 cm from below a vessel in square 2
(Kamiltepe, Aruchlo and Hesensu) were nearly pollen and spore sterile. Two further
samples from square 2, an ash layer at 260 cm and
The pollen spectra of 16 samples from pit P12 in a sample at 130–150 cm, pictured quite a rich pol-
Kamiltepe spanning a vertical distance of about len content dominated by Artemisia followed by the
100 cm can be grouped in those rich in Aster-type Anthemis-type, Poaceae, Malvaceae, Chenopodia-
(samples from bottom to top 31, 32, 39, 42, 44), ceae and Caryophyllaceae. Fungi were represented
Artemisia (35, 36, 37, 37,5) and Chenopodiaceae very sporadically by spores of the Glomus- and the
(33, 45). Sample 43 revealed the highest diversity Sordaria-type.
containing in addition pollen of Liguliflorae, Ero-
dium, Poaceae, Cerealia- and Aconitum-type and Mudbricks
also spores of the Glomus-, Riccia- and Urocystis-
type as well as some unidentified spores and pol- The yellowish mudbricks and the mortar exposed
len forms. The spectrum of sample 34 is dominated during the excavation at Kamiltepe were free of any
by the Cerealia-type and Liguliflorae. Three of the pollen or spores, while charcoal particles occurred
samples (38, 39, 40) provided almost no pollen very occasionally. Concluded from the absence of
grains. Hence, the pollen spectra probably do not the Glomus-type, the extraction of mud for building
reflect a directed vegetation development over the base of Kamiltepe reached down below the root
time. As these samples represent refuse layers de- horizon, derived from the results of profile MPSA1
posited on the flank of the tepe, the spectra more and the drilling EB 03 (see above) into more than
likely reflect different sources and moments of 200–250 cm depth.
waste disposal, including successional stages of ve- The gray mudbricks and the mortar sampled
getation on the flank. at Aruchlo provided sharply contrasting results.
In a trench at the base of Aruchlo the transi- Their palynological spectra were rich in pollen,
tion from the natural soil to the floor and sediment spores and charcoal particles. The most common
fillings of the lowermost building of the tepe and a pollen taxa found were the Anthemis-type, Cheno-
superimposed building were sampled approximately podiaceae, Poaceae, Artemisia and the Liguliflorae,
every 5 to 10 cm over a vertical distance of 210 cm. but also pollen grains of the Cerealia-type and of
In the 70 cm of natural soil below the oldest floor Xanthium appeared. Spores were represented by the
Glomus- and the Riccia-type. Possibly most of the could become the nest cavities of solitary bees.
pollen and spores were incorporated during the They are tunneled frequently into the unconsoli-
processing of the mudbricks. In the case of the Ka- dated soft material of cultural layers in the walls,
miltepe mudbricks, such an immixture operation is even of very recent trenches, and are filled with
not evident. loads of pollen. The practise of drying samples for
soil analyses by opening the plastic bags in which
House floors they are stored may turn them into good archives
for airborne pollen, but avoid their use for any
The pollen content of samples from the house floor geoarchaeological purpose. To prevent the growing
in planum 323 at Kamiltepe was dominated by Che- and sporulation of fungi, paper bags can be used.
nopodiaceae (samples 1, 2) and Liguliflorae (sam- In that way samples can dry and are shielded
ple 3), with the other three samples (4, 5, 6) being against pollution.
nearly pollen sterile. Spores were represented by Clumps of pollen grains. Pollen grains of Che-
the Glomus-type only in sample 2. nopodiaceae stuck together were quite common in
Pollen grains of Salix from the house floor at a few sub-fossil as well as in some surface samples.
site MPS 4 might hint at the use of willow twigs. They could be the result of an interruption of the
Besides Salix, pollen of Chenopodiaceae and Poa- pollen-ripening process in the anthers due to cli-
ceae occurred in small amounts, while spores were matic or mechanical stress like grazing, for instance.
missing. Possibly, like in case of the bricks of Ka- Any processes in the soil that lead to clumps
miltepe, spore sterile mud was taken to build the should also be investigated. Mixed clumps to be in-
house floor. The few pollen grains may have fallen terpreted as bee loads did not occur in our material.
into the mud while the floor was been laid. Charcoal particles. So called ‘‘charcoal’’ parti-
cles or flitter represent burnt plant material not only
Skeletons of trees and shrubs, but of grasses and herbs as
well. Those particles were very common in several
Samples were taken from the abdominal region of of the cultural layers and bricks. After preparation
two skeletons and in addition from near one shoul- in the laboratory, some samples consisted for the
der on Kamiltepe. Both skeletons are probably of most part of charred particles only. Those particles
Medieval age. The three samples revealed pollen in were rare in the surface samples and the cores
a quite high amount and ample composition. Most drilled outside the archaeological settlements, indi-
common pollen grains were those of Chenopodia- cating that higher concentrations of them can indi-
ceae followed by Artemisia and Poaceae, pointing cate former human settlement activity quite well.
to nearby dry steppe and/or nutrient rich sites (Che-
nopodiaceae). Other pollen types point to grazing Palynological conclusions
(Liguliflorae, Euphorbia) and trampling (Polygonum and perspectives
aviculare-type) as well as to locations with open
water (Sparganium-type). In addition, pollen of the Compared with surface samples taken along cli-
Anthemis-, Aster-type, Euphorbia and Malvaceae matic and altitudinal transects in comparable land-
were found. Spores of the Sordaria-type, the Glo- scapes,339 higher Chenopodiaceae values can be
mus-type and other conidiospores were quite com- correlated to desert conditions. Nevertheless, the
mon, but not so in the sample from the shoulder. surface samples from near Kamiltepe indicate that
The findings of Juglans pollen may stand for the strong human influence leads to high Chenopodia-
cultivation of walnut trees since at least Medieval ceae values in the surface samples on horticultural
times, as this pollen type was not found in any ol- fields. Thus, higher amounts of Chenopodiaceae
der materials. Further sampling and preparation ex- pollen in archaeological samples are not necessarily
periments are needed to reach conclusions about an indicator for dry climatic conditions, but may
the absence of intestinal parasite eggs in the abdo- point to nutrient rich places due to human activity
minal samples. Possibly their detection needs more in and around a settlement.
exhaustive strategies in the field and the laboratory. The absence of Sporormiella-type spores in
the samples from Kamiltepe could be interpreted in
Some general remarks on the such a way that required organic material, i.e. dung
palynological investigations as well as straw, did not occur in an appropriate
way. Possibly the conditions were too dry, and the
Contamination. To prevent contamination by recent settlement was carefully cleaned.
pollen and spores during field work, great caution
is needed in choosing, extracting and handling the
samples. A very special source of contamination 339
Connor et al. 2004; Djamali et al. 2009.
The first sediment layers on the former house ken off-site in the Mil Steppe (i.e. Lake Ag Göl)
floors seem to be richer in palynomorphs than the and/or in the currently not accessible Nagorno-Qara-
sediments above and the floor layers themselves, bagh Mountains in the west would be necessary.
making them possibly good archives for recon- Both, surface samples and off-site profiles could
structing the former use of the rooms. Higher con- amend enormously the interpretations of the paly-
centrations of pollen of the Cerealia-type in combi- nological spectra in their geoarchaeological context.
nation with phytolith analyses could support the
identification of rooms used to store or process cer- Archaeozoological studies
eals. Spores of the Glomus-type in sediments from
inside former buildings may point to phases of up- Norbert Benecke
coming vegetation or the input of soil material.
Careful analyses of the flowering periods of The excavations on the prehistoric tell sites Aruchlo
the taxa represented by the pollen types in the (Georgia), Kamiltepe and Mentesh Tepe (Azerbaijan)
mudbricks and mortar could help to limit the sea- have produced large samples of animal remains.
son of processing these construction materials. They provide great insight into the subsistence
Spectra with pollen of Malvaceae are men- economy of these sites, and they have yielded
tioned very seldom in literature. Here, they occurred some data for reconstructing the ancient environ-
only in the profiles MPSA1 and 2 (Üç Tepe), Hesen- ment of the surroundings. The collections studied
su and with one of the skeletons. Due to the orna- so far consist of about 50 000 remains from Neo-
mental character of Malvaceae and the rarity of pol- lithic layers at Aruchlo and Kamiltepe and of nearly
len of the mallow family in palynological archives, 1100 remains from Neolithic and Early Bronze Age
some of the findings may originate from intentional deposits at Mentesh Tepe. The study is work in
human activity. progress and Mentesh Tepe material remains to be
Further work in the field should include the studied in greater detail in 2012.
sampling of surface samples from sites of different According to the archaeozoological evidence
land use in order to estimate the human influence in all three sites the settlement’s population relied
on the composition of pollen spectra under a given on animal breeding rather than hunting. The assem-
Fig. 192
Aruchlo and Kamiltepe. climatic background. In addition, knowledge about blages mainly consist of remains of domestic mam-
Relative frequencies of the overall landscape and climate development mals with sheep and goats being the most frequent
domestic mammals and over time is strongly needed to quote the amount species according to bone counts, followed by cat-
game (wild mammals),
based on fragment of local changes by former cultures documented in tle and pig. There are only small differences in the
counts (NISP) on-site archives. Therefore, a palynological core ta- percentages of the main food animals between the
Neolithic sites Aruchlo and Kamiltepe (Fig. 192).
Within the ovicaprids sheep are much more frequent
than goats at both sites. The sheep/goat ratio is
Aruchlo Kamiltepe 4 : 1 in Aruchlo and 5 : 1 in Kamiltepe. The Neolithic
Sheep/Goat 52.00 % 64.00 % assemblage from Mentesh Tepe exhibits a slightly
Cattle 32.00 % 24.00 % different species composition. Here, sheep/goat con-
Pig 14.00 % 10.00 % stitutes 71% and cattle 28% of the total amount of
Game 2.00 % 2.00 % domestic mammals, while pig is represented by
only 1%.
70% The prevailing species, sheep and goat, are
multi-purpose animals, i.e. they can be sources of
60% meat, milk, and wool or hair. Analysis of both mor-
Aruchlo tality pattern and sex ratio may shed light on the
50% Kamiltepe relative importance of meat production, dairying,
and wool/hair production in animal husbandry prac-
40% tices of the sites. So far, only the collection from
Aruchlo has yielded enough reliable data for recon-
30% structing the ancient herd composition. The results
of age determination on the mandibles of sheep/
20% goat are summarized in Fig. 193. Only a small num-
ber of animals were killed as juvenile individuals,
10% i.e. in the first year of life. About 60% of the ani-
mals were older than two years of age when slaugh-
0% tered. Sex determination on isolated horn cores,
Sheep/Goat Cattle Pig Game cranial fragments and pelvic bones indicates that
among both adult sheep and goats, female indivi- Aşağı Pınar 5-2 Aruchlo
duals clearly outnumber male individuals. Summa-
(N=1111) (N=109)
rizing the ageing evidence and the sex ratio, it be-
comes clear that sheep and goats were primarily 1-2 0.40 %
exploited for their meat and obviously also for their 3-4 2.80 % 2.80 %
milk. Age-structure and sex-ratio as seen in the ovi-
5-8 13.10 % 9.10 %
caprids from Aruchlo exhibit great similarities with
those for the sheep and goats from the Neolithic 9-11 9.40 % 3.70 %
site Aşaǧı Pınar in Thrace, where the exploitation of 12-17 10.00 % 13.80 %
milk was also attested by the analysis of organic 18-24 6.20 % 11.00 %
residue on pottery shards. >24 49.00 % 55.00 %
Morphologically, the sheep, goats, cattle and >>24 7.40 % 2.60 %
pigs from all three sites represent animals which >>>24 2.00 % 2.00 %
were in a well-advanced stage of domestication,
obviously being under human control for a long 60%
period. In general, they are significantly smaller in
body size than their wild relatives. First successful Aşağı Pınar 5-2 (N=1111)
50%
aDNA studies show a high variability in mitochondrial Aruchlo (N=109)
haplotypes in sheep and cattle from Aruchlo.340
This is in clear contrast to Western Anatolia or 40%
Southeast Europe where a strongly reduced haplo-
type variability was observed in these species indi- 30%
cating a quick spread of animals originating from a
small founder population in the areas of domesti-
20%
cation (‘‘bottleneck’’-effect). The high genetic varia-
bility seen in the Neolithic domestic animals from
South Caucasia might be indicative for a close proxi- 10%
mity to the areas of primary domestication. Further
research is urgently needed here since it seems that 0%
the Southern Caucasus played a more important 1-2 3-4 5-8 9-11 12-17 18-24 >24 >>24 >>>24
role in the process of animal domestication than Age in months
was previously thought.
Only a small number of the bones recovered
from the three sites belong to wild animals, indicat-
ing that species of this group contributed little to tepe, the wild mammal fauna is dominated by ga- Fig. 193
Aşaǧı Pınar and Aruch-
the diet of the former inhabitants and, obviously, zelle. Less frequent species are red deer, wild pig, lo. Age structure in
they were exploited only rarely as a resource for raw fox, badger and hare. In contrast to Aruchlo, fowl- sheep/goat on the
materials. The percentage of wild species among ing was of some importance at this site: about 600 basis of tooth develop-
ment in mandibles
the mammal remains varies between 2% in Aruchlo bird bones were recovered here. Species identifica-
and Kamiltepe and 6% in Mentesh Tepe. tion shows that nearly 90% of these bones come
There are significant differences in the com- from little bustard (Tetrax tetrax). Nowadays, this
position of wild species between the sites. The col- species is a migrant or winter visitor in the Mil
lection from Aruchlo is dominated by red deer, au- steppe forming large flocks in this period of the
rochs, roe deer and European bison. Less frequent year, thus making this bird easily available for hunt-
species are gazelle, hare, various carnivores (fox, ing. Most of the wild mammals and birds encoun-
badger, otter) and beaver. This points to the simul- tered at Kamiltepe point to a local environment
taneous exploitation of a variety of habitats, i.e. where open landscapes, mainly steppes, prevailed.
forests (red deer, bison), open woodland (aurochs) A few bone finds of carp fish and sturgeon docu-
and grassland (gazelle). Therefore, the surrounding ment fishing activities at the site. The collection of
of Aruchlo can be characterized as a mosaic of dif- wild animals from Mentesh Tepe available and
ferent landscapes. Fishing is documented at Aruchlo studied so far is too small for any conclusions on
by a few bone finds of carp fish (barbel). In Kamil- the environment of the surroundings.
340
Geörg 2011; Scheu 2011.
Soil samples treated Floated volume (l) Minimum size Maximum size
by flotation of sample (l) of sample (l)
Mil Plain sites 91 4,156 5 220

Tab. 25 Mentesh Tepe 96 1,880.35 1 267
Samples from
Mil Plain sites and TOTAL 187 6,036.35 % %
Mentesh Tepe 2011
First results of the archaeobotanical studies that constituted the base of subsistence, in order
at Kamiltepe and Mentesh Tepe to reconstruct agricultural practices in relation to lo-
cal environmental conditions and to understand
Alexia Decaix341 various aspects of the exploitation of wild plant re-
sources, notably wood. By the analysis of charcoal
Introduction resulting from the use of wood as fuel we also aim
at reconstructing the plant formations that would
Few recent archaeobotanical studies have been un- have grown around the sites in the past and their
dertaken so far in the Caucasus region where this possible evolution through time due to the impact
type of research is still in its early stages.342 This is of human and natural factors.
in stark contrast to the importance of the area for The analysis of archaeobotanical samples from
understanding phenomena of diffusion of early ag- the Mil Plain sites and Mentesh Tepe presented
ricultural economies and their adaptation to local here has just started, and the present paper has
conditions. thus to be considered as a preliminary report that
The global aim of archaeobotanical studies is is liable to be enriched and further detailed in a near
to understand interactions between human socie- future.
Fig. 194 ties and their environment. Among the questions of
Setting of the particular interest during the the first study of past
flotation device
plant economies at the Mil Plain sites and at Men- Methods of sampling and analysis
close to Kamiltepe
(photo A. Decaix) tesh Tepe is the identification of cultivated species Flotation
During the field work in summer 2011, soil samples

for archaeobotanical analysis were collected from
various features and layers in both sites. A total of
217 samples intended for archaeobotanical analysis
were collected in 2011. While some of these corre-
spond to hand-picked remains (mostly larger char-
coal fragments), 187 consist of sediment samples
of varying volumes (Tab. 25).
At the Mil Plain sites, charred plants remains
were separated from the surrounding soil matrix by
flotation carried out at some distance from the
excavation trenches where water was extracted
from a canal with with the help of an electric pump
(Fig. 194). Soil samples were kept in strong plastic
sugar bags and transported by car from the differ-
ent operations of excavation to the flotation area.
At Mentesh Tepe, soil samples were brought
daily in plastic bags to the excavation house where
flotation was carried out.
For both sites, the archaeological sediment
was slowly poured into a tank filled with water. A
pump was used to bring water into the tank and,
341 This archaeobotanical study will be further developed in co- despite water and electricity problems, the rate of
operation with Reinder Neef (DAI), who has already docu- flow was good and the system proved functional.
mented a larger part of the Aruchlo samples, and Margareta The use of a graduate bucket gave us the volumes
Tengberg (University Paris 1 Panthéon-Sorbonne), Supervisor
of this PhD. (in litres) of the sediment samples from the differ-
342
Hovsepyan/Willcox 2008; Badalyan et al. 2010. ent units. The heavy fraction (animal bones, pottery
Number of Floated Sorted Average Minimum Maximum Taxa Seeds and Charcoal
samples volume (l) volume (l) density density density identified fruits fragments
studied of remains analysed analysed
Mil Plain sites 45 1,655.5 2.69 12.15 0.05 51.39 26 20,112 370
Mentesh Tepe 54 706.1 2.32 23.63 1 1,389 30 16,687 382 Tab. 26

Studied samples
Total 99 2,365.6 5.01 % % % 37 36,799 752 from the Mil Plain sites
and Mentesh Tepe
sherds, pebbles . . .) was recovered on a grid (mesh Identifications of seeds and fruits were made with
size: 2 mm) put just under the surface of the water. the help of a binocular microscope (Nikon SMZ645).
Light charred remains that floated on the surface Charcoal pieces were studied with the help of a
were caught in a smaller sieve with a mesh of high-power reflected-light microscope (Olympus
either 0.3 mm or 0.5 mm (flour sieve). After extrac- BX51M) allowing magnifications up to "1000.
tion, the light fraction was put to dry in the shade
as found in the sieves.343 Preliminary results from Kamiltepe
and the Mil Plain sites
Conservation of remains
45 samples from Kamiltepe and MPS 4, 5, 16 and
Except for the mineralized nutlets belonging to the 18 have been sorted for seed remains and a preli-
borage family (Boraginaceae) and hackberry fruits, minary identification. The density of seed remains
all macrobotanical remains at the Mil Plain sites is variable from one sample to another, varying
and Mentesh Tepe are preserved by carbonisation from less than one item per litre (11 samples) to
and are generally found in areas where ashy sedi- more than 50 (1 sample) with a mean density of
ments have accumulated, in pits and ditches but 12.15 remains per litre (Tab. 26). Moreover, the pre-
also in room fills and on occupational floors. Few servation of remains seems to vary from one group
samples appeared directly associated with hearths of plants to another, with the remains of wild plant
and ovens during the excavation but it is neverthe- species generally better preserved than cereals,
less likely that many of the ashy deposits result even though this preliminary observation needs to
from the emptying out of such structures. Generally, be confirmed by a more systematic evaluation. In
it seems that the large majority of our samples general, seed remains are abundant in the samples
come from secondary deposits and thus probably while there are few and small-sized charcoal frag-
correspond to an admixture of plant remains result- ments. The analysis of these remains has allowed
ing from various activities (fuel burning, food pre- the identification of 26 taxa belonging to 16 botani-
paration, crop-processing . . .). In any case, no in cal families.
situ burning of building elements or of stored food-
stuffs due to accidental fire was observed at the Studies of seeds and fruits
sites in 2011.
The preliminary analysis of 20,112 seed and fruit
Study in the laboratory remains from the Mil Plain sites has allowed the
identification of 20 taxa belonging to 10 different
Once the samples had dried, they were conditioned botanical families. Cultivated plant species largely
in labeled bags and sent for analysis to the Labora- dominate the samples numerically even though wild
tory of Bioarchaeology (UMR 7209) at the Natural species are also attested in most samples as can
History Museum of Paris.344 Among the 187 flota- be seen in Tab. 27.
tion samples, 99 have been sorted for seed remains
and preliminary identifications (see below). Char- Crop plants
coal pieces from hand-picked samples (N ¼ 30)
were isolated and will be identified at a later stage. Among the cultivated plants, cereals and in particu-
lar barley (Hordeum vulgare) are by far the most
343 frequently encountered crops. Indeed, barley grains
In Kamiltepe, we were assisted during the flotation and sort-
ing of the heavy fraction by Ms. Golnaz Ahadi, student at the (caryopses) are present in almost all of the samples
University of Teheran. In Mentesh Tepe, students from the studied so far and sometimes in large numbers
University of Baku helped with this task. (>100). Rachis segments, though, are extremely
344
We would like to thank the French Embassy and especially
the Cultural Department, as well as the German Embassy, for
rare and are usually too poorly preserved to be
their help in this matter. identified. This may imply that we are dealing with
KAMILTEPE
KAMILTEPE AND MIL PLAIN SITES
TOTAL NA
TOTAL NA
TOTAL NA
TOTAL NA
TOTAL NA
TOTAL NA
TOTAL %
(MPS 1)
MPS 16
MPS 18
MPS 4
MPS 5
Cereals Hordeum vulgare, caryopsis 5,575 101 70 270 2 6,018 29.92
Hordeum vulgare, hulled caryopsis 4,324 12 76 14 % 4,426 22.01
Hordeum vulgare var. nudum, caryopsis 11 % % % % 11 0.05
Hordeum vulgare asymetrical, caryopsis 27 % % % % 27 0.13
Hordeum vulgare, rachis segment 17 2 % 1 % 20 0.10
Triticum aestivum/durum, caryopsis 266 15 36 20 % 337 1.68
cf. Triticum, caryopsis 22 2 % 1 % 25 0.12
Triticum, rachis segment 8 1 % % % 9 0.04
Cerealia, caryopsis 2,748 150 227 532 5 3,662 18.21
Cerealia, rachis segment 6 % % % % 6 0.03
Total cereals 13,004 283 409 838 7 14,541 72.30
Pulses Lens culinaris 7 % 6 % % 13 0.06
cf. Lens culinaris 1 % % % % 1 0.00
Cultivated Fabaceae 4 % 4 % % 8 0.04
Total pulses 12 % 10 % % 22 0.11
Oil/fibre plants Linum cf usitatissimum 17 1 % 2 % 20 0.10
Total oil/fibres plants 17 1 % 2 % 20 0.10
Wild Poacaee Aegilops, caryopsis 2 % % % % 2 0.01
Aegilops, rachis segment 5 % % % % 5 0.02
Panicoideae, caryopsis 1 % % % % 1 0.00
Poaceae, caryopsis 897 29 55 155 2 1,138 5.66
Total wild Poaceae 905 29 55 155 2 1,146 5.70
Wild Fabaceae cf. Astragalus 2 % % 15 1 18 0.09
cf. Medicago % % 1 % % 1 0.00
cf. Trigonella % % % 1 % 1 0.00
Total wild Fabaceae 2 % 1 16 1 20 0.10
Wild taxa Adonis 13 1 6 3 % 23 0.11
Boraginaceae 13 % 7 6 % 26 0.13
Buglossoides cf. arvensis/sibthorpianum % % 0 5 % 5 0.02
Capparis 489 % 6 35 % 530 2.64
Chenopodium cf. glaucum/rubrum 2,032 3 7 2 2 2,046 10.17
Cyperaceae/Polygonaceae 11 % % % % 11 0.05
Galium % % % 3 % 3 0.01
cf. Rumex 11 % % % % 11 0.05
Thymelaea 1 % % % % 1 0.00
Total wild taxa 257 4 26 54 2 2,656 13.21

Tab. 27
Kamiltepe and the Mil Undetermined 1,567 20 35 85 % 1,707 8.49
Plain sites. Preliminary
results from the ana- TOTAL 18,077 337 536 1,150 12 20,112 100
lysis of seeds and fruits
the storage of crops and that the post-harvesting

process took place outside the site or that the by-
products were used for other purposes, for exam-
ple, making mudbricks. When sufficiently well pre-
served, the barley grains present longitudinal scars
on the sides, indicating hulled barley, for which the
inner glumes remain attached to the grain even
after threshing (Fig. 195). A large majority of the
grains are symmetrical, indicating a predominance
(or perhaps even the exclusiveness) of the two-row
subspecies (subsp. distichum). No naked barley
(var. nudum) has so far been identified with cer-
tainty, even though a very low proportion of the
caryopses shows a more squat and rounded form
than what is usually characteristic for hulled barley.
The grains of naked wheat (Triticum aestivum/
durum type) appear in low numbers in a few sam-
ples and therefore, according to these first results,
wheat does not seem to have been a major crop
species cultivated at Kamiltepe.
Lentil (Lens culinaris) is attested through a
low number of seeds (N ¼ 14). represented in the region by several species, gener- Fig. 195
Hordeum vulgare,
To the cereal and pulse species we can add a ally growing on periodically inundated soils, often hulled caryopses
fair number of flax seeds (Linum cf usitatissimum) along rivers and canals. The tamarisks are usually (Kamiltepe)
(N ¼ 20) (Fig. 196). halophilous and regenerate easily even after re-
peated pruning.
Wild species Two species belonging to the Salicaceae fa-
mily were recognized: Populus (poplar) and Salix
12 wild plants taxa belonging to 9 different botani- (willow). Some of the charcoals belonging to this
cal families have been identified at Kamiltepe and family could not be identified precisely, as the
the Mil Plain sites: Boraginaceae (Buglossoides cf. structure of the wooden rays (discriminating in the
arvensis/sibthorpianum), Capparaceae (Capparis), case of the Salicaceae since rays are homogenous
Chenopodiaceae, Cyperaceae/Polygonaceae, Faba- in poplar and heterogeneous in willow) could not
ceae (Astragalus, Trigonella), Poaceae (Aegilops and always be determined in our samples.
some other not yet identified types), Ranunculaceae The last taxa identified as belonging to the ri-
(Adonis), Rubiaceae (Galium) and Thymelaeaceae parian forest is the reed (Phragmites).
(Thymelaea). Furthermore, the genus Acer was recognized
Among the above mentioned taxa, seeds from thanks to its diffuse porous and its numerous spiral
the goosefoot (Chenopodiaceae) and caper (Cap- thickenings. The maple tree is characteristic from
paris) family are particularly abundant (Fig. 197). the mixed deciduous forest in the Caucasus re-
gion.346
Charcoal study A genus from the Cornaceae family, the dog- Fig. 196
wood (Cornus), was also identified. This wood is Linum usitatissimum
The analysis of 370 charcoal fragments from Kamil- characterized by its diffuse porous wood and its (Kamiltepe)
tepe chosen randomly from flotation samples has al- scalariform perforations. Fruits from this tree are
lowed the identification of six wood taxa (Tab. 28). used today to prepare juices or for medicinal prop-
The ash tree (Fraxinus), characterized by its erties.347
ring-porous wood as well as its homogeneous and
multiseriate rays, was identified in several samples. Preliminary results from Mentesh Tepe
Several Fraxinus species are recorded for the Cau-
casus region growing in forests and along rivers.345 54 samples coming from different periods attested
The tamarisk (Tamarix) easily recognizable by at the site have been sorted for seed remains and
its large and heterogeneous rays with a storied struc- preliminarily identified. The density of seed remains
ture was identified in most samples. This genus is
346
Gabrielian/Fragman-Sapir 2008. Fig. 197
345 347
Akhalkatsi 2009. Gabrielian/Zohary 2004. Capparis (Kamiltepe)
KAMILTEPE MPS 1
Frequence (%)
BOTA 006
BOTA 061
BOTA 080
BOTA 081
BOTA 093
BOTA 094
BOTA 099
BOTA 111
TOTAL
Riparian forest Ash tree Fraxinus % 12 20 16 8 2 22 12 92 24.86
cf. Fraxinus % 11 6 4 15 2 7 6 51 13.78
Reed Phragmites % 3 2 % 2 1 1 % 9 2.43
Poplar/willow Populus/Salix % 4 15 % 3 2 2 5 31 8.38
cf. Populus/Salix % 1 % % 1 % % % 2 0.54
Tamarisk Tamarix 1 % % 21 3 17 % % 42 11.35
cf. Tamarix 2 1 % 1 2 9 10 2 27 7.30
Total riparian forest 3 32 43 42 34 33 42 25 254 68.65
Mixed deciduous Maple tree Acer % % % % 1 % % % 1 0.27

forest
cf. Acer % % % % 1 % % % 1 0.27
Buckthorn cf. Cornus % % 1 % % 2 % % 3 0.81
Total mixed deciduous forest % % 1 % 2 2 % % 5 1.35
Parenchyma tissue 1 % % % % % % % 1 0.27
Undetermined 16 18 6 8 14 15 8 25 110 30
Tab. 28
Kamiltepe. Preliminary Total others 17 18 6 8 14 15 8 25 111 30
results from the
analysis of charcoal TOTAL 17 18 8 8 86 19 8 25 370 100
fragments
is variable from one sample to another, varying Studies of seeds and fruits
from one item per litre to 1389, with a mean den-
sity of 23.63 remains per litre (Tab. 26). Moreover, The preliminary analysis of 16,687 seed and fruit
the preservation of remains is very good in some remains from Mentesh Tepe has allowed the identi-
samples where, for example, glumes enclosing bar- fication of 23 taxa belonging to 11 different botani-
ley grains are still preserved. cal families. Wild species seem to be the most fre-
Thanks to the study of almost 17,000 seeds quently encountered plant remains (see below).
and fruits and 382 charcoal fragments, 30 taxa be-
longing to 18 botanical families were identified. Crop plants
Fig. 198 Among the botanical remains, cereals constitute the

Triticum aestivum/ second most important group with barley and naked
durum (Mentesh Tepe)
wheat (Fig. 198). Among the pulses, lentil and grass
pea are attested (Fig. 199). Flax, also identified,
could have been used as a textile plant – a use
well known in the Near East since the Neolithic348 –
or for the extraction of oil.
The presence of charred and mineralized
seeds of grapevine also suggests practices of fruit
cultivation and perhaps the production of wine,
even though we do not yet have any firm evidence
for this, neither in the form of specific archaeologi-
cal structures nor as press remains, as is the case
on other sites in the Caucasus region (Fig. 200,1–
2).349
348
Zohary/Hopf 2003.
349
Areshian et al. 2012.
TOTAL NA
MENTESH TEPE
TOTAL %
Cereals Hordeum vulgare, hulled caryopsis 3 0.02
cf H. vulgare var. nudum, caryopsis 1 0.01
H. vulgare, caryopsis 950 5.69
H. vulgare, rachis segment 16 0.10
cf Hordeum sp., caryopsis 2 0.01
Triticum aestivum/durum, caryopsis 239 1.43
T. aestivum/durum, rachis segment 10 0.06
T. dicoccum type, caryopsis 55 0.33
T. dicoccum type, rachis segment 3 0.02
Little Triticum, caryopsis 8 0.05
cf Triticum sp., caryopsis 27 0.16
Cerealia, caryopsis 3,668 21.98
Cerealia, rachis segment 712 4.27
Total cereals 5,694 34.12
Pulses Lens culinaris 132 0.79 Fig. 199
Lathyrus 2 0.01 Lathyrus (Mentesh Tepe)
cf Lathyrus 2 0.01
Little ‘‘Pisum’’ 14 0.08
Cultivated Fabaceae 67 0.40
Total pulses 217 1.30
Oil/fiber Linum usitatissimum 83 0.50
plants
Total oil/fiber plants 83 0.50
Fruits Vitis 5 0.03
cf Vitis 1 0.01
Total fruits 6 0.04
Wild Aegilops type, caryopsis 20 0.12
Poaceae
Aegilops, rachis segment 31 0.19
Bromus type, caryopsis 93 0.56
Panicoideae 71 0.43
Poaceae 1,257 7.53
Total wild Poaceae 1,472 8.82
Wild Astragalus type 10 0.06
Fabaceae
Melilotus/Medicago type 1 0.01
Trigonella type 6 0.04
Fabaceae 2 0.01
Total wild Fabaceae 19 0.11
Wild taxa Adonis 22 0.13
Asteraceae type 2 0.01
Boraginaceae 5,728 34.33
Chenopodiaceae 534 3.20
Cyperaceae/Polygonaceae 19 0.11
Galium 92 0.55 Fig. 200
Thymelaea 1 0.01 Vitis. 1 charred seeds; 2 mineralized seeds (Mentesh Tepe)
Total wild taxa 6,398 38.34

Undetermined 2,798 16.77 3
TOTAL 16,687 100.00 Tab. 29
Mentesh Tepe. Preliminary results from the analysis of seeds and fruits
Frequence
Mentesh Tepe
FLOT 020
FLOT 024
FLOT 025
FLOT 029
FLOT 038
FLOT 043
FLOT 070
FLOT 089
FLOT 137
FLOT 145
TOTAL
(%)
Riparian forest Ash tree cf. Fraxinus 5 % 4 % % % % % % % 9 2.36
Poplar/willow Populus/Salix 9 % 12 1 23 18 % % 3 16 82 21.47
cf. Populus/Salix % % 3 % % 4 % % % 6 13 3.40
Total riparian forest 14 % 19 1 23 22 % % 3 22 104 27.23
Shibliak Buckthorn Rhamnus 1 1 % % % % % % % % 2 0.52
cf. Rhamnus % 1 % % % % % % % % 1 0.26
Elm cf. Ulmus 12 % 10 19 6 4 2 % % 22 75 19.63
Total shibliak 13 2 10 19 6 4 2 % % 22 78 20.42
Mixed deciduous forest Maple tree Acer 1 % % 3 4 13 37 % % % 58 15.18
cf. Acer 1 1 1 1 % 1 % % % % 5 1.31
Oak cf. Quercus % % 3 % % % % % % % 3 0.79
Common yew cf. Taxus baccata % % % % % 1 % % % % 1 0.26
Gymnosperm % % % % % % % % % 1 1 0.26
Total mixed deciduous forest 2 1 4 4 4 15 37 % % 1 68 17.80
Parenchyma tissue % % 2 % % % % % % % 2 0.52
Undetermined 21 17 15 23 16 9 5 11 8 5 130 34.03

Tab. 30
Mentesh Tepe. Total others 21 17 17 23 16 9 5 11 8 5 132 34.55
Preliminary results
from the analysis TOTAL 50 20 50 47 49 50 44 11 11 50 382 100.00
of charcoal fragments
Wild species to three different plant formations: riparian forest,

so-called shibliak (open shrublands) and mixed de-
15 wild plant taxa belonging to eight different bo- ciduous forests.
tanical families have been identified at Mentesh As in Kamiltepe and the Mil Plain sites, riparian
Tepe: Boraginaceae, Chenopodiaceae, Cyperaceae/ forest seems to have constituted the main source of
Polygonaceae, Fabaceae (Astragalus, Trigonella), the wood collected. Ash tree and poplar/willow be-
Poaceae (Aegilops, Panicoideae and some other long to this formation.
not yet identified types), Ranunculaceae (Adonis), The second source of wood is the shibliak
Rubiaceae (Galium) and Thymelaeaceae (Thyme- shrublands. These are open plant formations domi-
laea). nated by shrubs and today they occur at altitudes
As said above, wild taxa seem to be dominant between 500 and 1000–1600 m. Buckthorn (Rham-
in our samples. However, this is mainly due to sam- nus) and elm (Ulmus) are characteristic species
ple 79 that contained a concentration (5.244 items) growing in this type of formation.350 Buckthorn is
of Boraginaceae seeds (cf Anchusa). If we exclude characterized by the presence of parenchyma and
these from our general counting, the proportion of the disposition of pores in dendritic groups, both
wild taxa immediately drops from 38% to 10%, which easily visible in transverse section. Elm was identi-
seems to be more representative of our samples. fied thanks to the presence of parenchyma in tan-
The presence of wild seeds that in many cases gential bands and the disposition of latewood pores
may represent weeds, as well as that of rachis seg- in groups.
ments in most of the samples, indicates that har- The third vegetation type attested is the de-
vests were treated locally. ciduous mixed forest, with maple tree (Acer), oak
(Quercus) and yew (Taxus baccata). Oak porous
Charcoal study wood is easily recognizable because of the disposi-
The study of 382 charcoal fragments has allowed

the identification of several woody taxa belonging 350
Gabrielian and Fragman-Sapir 2008.
tion of its pores in dendritic groups. Yew grows now sites of Mentesh Tepe and Kamiltepe (Azerbaijan).
above 1400 m351 and is characterised by the pre- The aim of this article is to present the preliminary
sence of spiral thickenings and the absence of resin results on the material of each site, and then to
canals. compare them through a broader discussion includ-
Forest species are not very frequent in the sam- ing the regional context. A study of the macrolithic
ples from Mentesh Tepe, and this vegetation type, tools discovered at Aruchlo is also planned in the
thus, seems of less important for the collection of frame of our project.
wood.
Kamiltepe and the Mil Plain sites
Conclusion
From our study of ca. 200 macrolithic objects exca-
Altogether, 37 taxa belonging to 12 botanical fa- vated at three neolithic sites (Kamiltepe, MPS 4,
milies were common to both sites. MPS 5), we can underline the main characteristics
At both the Mil Plain sites and Mentesh Tepe of the industry of this period in the Mil Plain. It is
we have seen that the cultivated crops were almost mostly composed of grinding, hammering and pol-
the same with the predominance of barley, followed ishing tools. The toolkit is characterized by a high
by naked wheat as well as the presence of lentils. rate of fragmentation and dispersion. We present it
The collections of wild taxa also appear similar here by chronological order (Tab. 31).
in their composition and probably represent weeds At MPS 4 (beginning of the 6th mill. BCE),
cleaned out of the crops rather than an authenti- 30 objects have been discovered, among which
cally wild flora. The presence of flax in both sites 18 flakes and ten tools. They include: three frag-
may suggest the use of this plant for fibres, and ments of grinding tools, together with three percus-
thus the existence of textile crafts, or for the extrac- sion tools, among which is a hammer-stone made
tion of oil. of flint; two flaked cobbles made of metamorphic
At both sites, the principal source of wood is limestone; a scraper made out of a flake; two rub-
the riparian forest followed, to a lesser degree, by bing tools made of metamorphic rocks and lime-
wood from mixed deciduous forests. The presence stone cobbles with polished surfaces; and one large
of shrub formations of shibliak type has been no- flake of metamorphic rock used as a scraping tool
ticed only at Mentesh Tepe. on its distal and proximal edges.
Further studies will bring forward more de-
tailed information on the plant economies of these
Tools Kamiltepe MPS 4 MPS 5 TOTAL
sites and also allow comparisons between them
and those of other sites in the Southern Caucasus. Grinder 1 1 2
Undetermined 12 3 3 18
Macrolithic tools from Neolithic and grinding tools
Chalcolithic sites in the Southern Caucasus: Mortars 0
Mentesh Tepe and the Mil Plain sites
Pestles 1 1
Caroline Hamon Hammerstones 28 3 7 38
Hammer-anvil 3 3
Along with the development of agro-pastoral econo-
mies in the Southern Caucasus, an important evo- Hand polisher 1 1
lution of the techniques and material culture took Flat polisher 1 1 2
place. Macrolithic tools are therefore a key issue for
the reconstitution of food production and craft Grooved 1 1
polisher
activities within Neolithic and Chalcolithic settle-
ments.352 The techno-functional study of macrolithic Handstone 1 1
implements allows the positioning of food practices Smoother 7 2 9
and craft activities within the general cultural evolu-
Scraper 5 2 7
tion, the resource procurement strategies and the
environmental context. In 2011, within the ANR/DFG Undetermined 3 1 4
tools
program Ancient Kura, a typological, technological
and functional study was made on the macrolithic Choping tool 6 2 2 10
tools discovered at the Neolithic and Chalcolithic Flakes 71 20 3 94
fragments Tab. 31
351 Mil Steppe sites.
Gabrielian/Fragman-Sapir 2008. TOTAL 140 30 21 191 Categories of
352
Hamon 2007; Hamon 2008b; Hamon 2009. macrolithic tools
Fig. 201
Kamiltepe macrolithic
tools. 1 grinding tool;
2 pestle; 3 hammer-
stone; 4–7 limestone
and basalt cobbles
industry (photos
C. Hamon)
Among the 140 objects found in Kamiltepe

(mid 6th mill. BCE), only 60 are attributed to the un-
disturbed neolithic layers (Fig. 201).
Very few grinding tools have been found up
until now, and they are all fragmentary. Sandstone
seems more frequently used than basalt (Fig. 201,1).
They consist of one half-grinder with a double active
surface made on a coarsely shaped sandstone slab.
Another part of a grinding tool made on a fragment
of basalt was shaped by pecking; it presents a con-
cave active surface. One pestle was made on a frag-
ment of porous basalt; its face shows traces of pol-
ishing while its ends were used to crush and grind
(Fig. 201,2).
Among polishing tools, four cobbles, probably
used for ceramic processing, and one lower polisher
have been identified.
Cores, hammer-stones and flakes made from
metamorphic and limestone cobbles represent 2/3 of
the collection. These industries made on compact
basalt and metamorphic limestone amount to about
100 pieces (Fig. 201,3–7). Among them, a few ham-
mer-stones were re-using previous chopping-tools
and cores, and an important quantity of retouched or
unretouched flakes was obtained by lateral or distal
flaking. Several scrapers and splintered tools were
made on such flakes. Some hammer-anvils show sev-
eral circular surfaces of fine percussion. Three tools
with a trapezoidal shape present traces of hammer-
ing and flaking of their distal and larger edge. Fig. 202
Kamiltepe.
A production of limestone vessel was also re- Limestone vessel
cognized (Fig. 202). Among several fragments of with ochre residue
limestone vessel, two fragments present traces of (photos C. Hamon)
shaping: one show traces of polishing both on its
internal and external side, while its rim was flat- texts that have been discovered so far. Nonethe-
tened and regularized by scraping and polishing. less, the high fragmentation of the grinding tools,
Half a vessel is 10 cm in diameter and 4 cm deep; the diversity of the hammer-stone types and of the
its wall is rather thick (2.5 cm), its base is flat and raw materials that have been used rather evokes
its internal side present traces of pecking and pol- an opportunist and occasional toolkit. Nevertheless,
ishing. Its internal surface is entirely red-colored craft activities are attested by polishers, burnishers
and it may have been used as a mortar. Numerous and hand-stones involved in ceramic and lithic pro-
fragments of limestone presenting some technical ductions. The abundance of basalt and limestone
traces were probably used as a raw material stock. industries, especially hammer-stones, cores and
At MPS 5 (mid 6th mill. BCE), 21 macrolithic (un-)retouched scrapers on flakes, is the main ori-
tools have been unearthed. Four fragments of grind- ginality of the Mil Plain sites compared to other Neo-
ing tools are made of basalt and sandstone. Among lithic sites of the Southern Caucasus. This specificity
the seven hammer-stones, one is made of a re-used could be linked either to proper raw procurement
basalt core and two others present important traces strategies due to the lack of stones of good quality
of red color. Four chopping tools are made on me- near the sites, or to different cultural networks.
tamorphic limestone. Two burnishers and two other
polishing tools were also identified. Mentesh Tepe
From this short study on three sites of the Mil
Plain area, we can already infer a development and At Mentesh Tepe, 250 tools coming mostly from the
long duration of a local tradition. The low number Chalcolithic levels have been studied353. They corre-
of grinding implements is difficult to interpret, since
it could only be due to the actual state of the ex- 353 Some of the tools that have been studied also date to the
cavation process and depend on the kinds of con- Kura-Araxes period.
10 kilometer radius from the site in the terraces of

Tools Total no.
the surrounding river, and their size also fits with
Querns 15 that of the tools. Soft limestone is also available in
Grinders 27 the valley, 10 kilometers south from the site. Further
analysis will precise the origin of the different types
Grinding undet. 14 of basalts and possible exogenous sources.
Mortars 2 56 grinding tools, among which are 15 querns
Pestles 39
and 27 grinders, have been excavated. They are
made in three different qualities of basalts. At least
Pestles-handstones 1 two types of grinding tools have been identified, with
Grinding slabs 7 either short or overlapping grinders. They show a
high morphological diversity and a large range of
Multidirectional grinders 3
sizes (Fig. 203,1–2); their active surfaces vary in
Crushing tools 7 width, are flat to highly concave, and are or are not
Slabs 1 associated to distal edges. Their shaping is gener-
ally well-made, and the natural surface of the initial
Anvils 2
cobble is well-regularized by flaking and pecking,
Hammerstones 40 so as to obtain a regular convex back which helps
Hammers 2
the handling. Some grinders show two opposite ac-
tive surfaces due to a re-use. A rough-out of a grin-
Anvil-hammers 6 der and a grinder in rejuvenation were also found.
Lower polishers 3 Further studies should determine if this technical
variability has a functional or a chronological mean-
Burnishers 24
ing. Five querns and grinders have been used for
Handstones 4 the processing of cereals. To complete the grinding
Sharpeners 1 toolkit, eight grinding slabs show by their ovoid ac-
tive surface that they have been used in a multidir-
Chisels 1
ectional or circular low breadth motion (Fig. 203,3).
Scrapers 3 Some cobbles were probably used with these slabs.
Other tools 8 As for pestles, two of them served for the proces-
sing of plants and mineral coloring. These two main
Flakes 5 categories of tools appear complementary in their
Tab. 32 Fragments 10 function, as querns and grinders were mainly used
Mentesh Tepe. for cereal processing, and grinding slabs and cob-
Categories of TOTAL 225
macrolithic tools bles for other plant and mineral transformations.
Pounding and crushing activities also took a
good place in the processing techniques. Surpris-
spond to 20 techno-functional types, involved in ingly, only two mortars, one made of granite and
grinding, percussion and abrading activities. A first the other of limestone, have been identified. This
sample of 33 tools was selected for use-wear analy- contrasts with the numerous cylindrical and tronco-
sis. The methodology of functional analysis relies nical pestles made of sandstones and metamorphic
on a combined use-wear and residue analysis.354 In rocks (40 ex.; Fig. 203,4). Despite variable dimen-
the field, particular attention was paid to the pro- sions (from 5 to 25 cm long and 4 to 10 cm in diam-
tection of the active surfaces in order to prevent as eter), they show a low degree of morphological
much as possible pollutions or alterations both of variability. One or both ends present a high density
the use-wear and of the residues (ochre, phytoliths, of hammering impacts together with polished aspe-
etc). A first analysis was made at low magnification rities, and worn-out retouches on their periphery.
with a stereomicroscope Nikon SMZ800 comple- Theses traces were created by alternating throwing
mented by silicon imprints. percussion and circular grinding, which corresponds
The range of the stones chosen for the tools to the action of pounding. Several of these pestles
is particularly varied: basalts, sandstones, granites, show a more occasional use of their faces as ham-
quartz and limestones were used for more or less mer-stones, crushing tools or hand-stones. Two out
specialized functional activities (Tab. 32). Most of of five analyzed pestles were used for grinding
these various stones can be found within a 5 to minerals for dye, while two others have processed
vegetal abrasive matters. Together with these pes-
tles, seven crushing tools present traces of ham-
354
Hamon 2008 b. mering on their ends and others of strong percus-
Fig. 203
Mentesh Tepe
macrolithic tools.
1 Kura-Araxes quern
and grinder; 2 grinder;
3 pestle; 4 grinding
slab; 5 crushing tool;
6 hammerstone;
7 lower polisher;
8 burnisher;
9 handstone
(photos C. Hamon)
Fig. 204
Mentesh Tepe lime-
stone items. 1 door-
socket; 2–3 fragments
of vessel with traces of
scraping and polishing
(photos C. Hamon)
sion concentrated in a circular cavity on their faces the combination of microflaking and of polishing ac-
(Fig. 203,5). Their low morphological and use-wear tivity. It was used for the scraping of a soft mineral
variability indicates that these tools probably did matter, such as limestone for example.
not change very much through time, but they are The important range of raw materials avail-
reliable indicators of activities. able at Mentesh Tepe made possible the produc-
The percussion activities are represented by tion of a diversified macrolithic toolkit, involved in
standardized hammer-stones made of a green me- grinding, pounding, abrading and scraping opera-
tamorphic rock. They are circular or ovoid in shape, tions either for food or for craft activities. Most of
and all of them present impacts of percussion and the tools are made on local cobbles and show a
abrasion only on the sides and on the ends (Fig. low degree of shaping. Most of the querns that we
203,6). This specific type of hammer-stone was studied were used for cereal grinding, although
mostly used for stone working, such as lithic debit- mineral processing is also frequent. Different types
age or quern shaping. Once again, their relative of tools, such as grinders, pestles and polishers
standardization prevents to use them as chronologi- also show traces of red color. While hammer-stones
cal markers but makes them a good technical indi- attest of lithic debitage on the site, the numerous
cator. This contrasts with eight other hammers burnishers that have been found confirm the pro-
which have different shapes and functional zones duction of ceramic on the site.
and which were also used for percussion activities. Besides these macrolithic tools, three types of
Among the few polishing and abrading tools, limestone objects have been discovered on the
three lower polishers have been discovered. One site: door sockets, vessels and small objects. The
was used for the polishing of a hard mineral object, 13 door sockets, mostly found in situ, have been
such as an ornament or a stone tool (like an axe) shaped by flaking and scraping: the circular cavity
(Fig. 203,7). 24 cobbles of small size and made of located on their upper face shows concentric stria-
hard stones (sandstones, quartz, metamorphic rocks) tions and smoothing resulting from the contact with
present both shiny polished surfaces and transver- the lower part of the door (Fig. 204,1). Several
sal striations mostly on their sides and secondarily fragments of vessels were also found at different
on their faces (Fig. 203,8). They were used for the stages of shaping, including sawn blocks. On the
regularization of a soft mineral matter. These tools better preserved vessels, traces of scraping of the
can be interpreted as smoothers on half-humid clay bottom part and of polishing on the external sur-
or as burnishers on dry clay for ceramic processing. face are visible (Fig. 204,2–3). The inside was regu-
Only four hand-stones show an association of pol- larized by scraping with a 4-teeth tool following
ishing and striations on their faces and edges re- either a circular or a vertical motion.
sulting from the softening or processing of a soft
organic matter (animal or vegetal fibres; Fig. 203,9). Comparison and discussion
Finally, very few flaked cobbles and retouched
flakes of basalt used for scraping activities were Despite some technological differences, the same
discovered. Among them, one chisel has been iden- range of macrolithic tools is present on both Men-
tified: it shows a double flaked edge resulting from tesh Tepe and the Mil Plain sites. Grinding, percus-
Kamiltepe and other Mil Plain sites Mentesh Tepe
Grinding tools Very fragmented and poor morphological indicator Chronological/functional variability of grinding tools
types; cereal and ochre processing
Crushing tools Absence of mortars and pestles? Standardized and varied types of pestles; high rate
of involvement of crushing tools in the processing of
Occasional tools for crushing activities (food or craft) plant and ochre
Hammering tools Variability of hammerstones: linked to poor raw Standardized types of hammerstones
material access? Or function? (lithic debitage and shaping)
Polishing tools Burnishers for ceramic
Polishers for mineral matters
Handstones for suppling activities
Scrapers etc Half of the industry based on flaked cobbles and Occasional tools on basalt flakes
retouched flakes of basalt and limestone
Ochre processing Tab. 33

Comparison of the
Limestone vessels Mil Plain and Mentesh
Tepe macrolithic tools
sion and abrading tools together with limestone the processing of ochre has also been observed on
vessels are present on all of them (Tab. 33). pounders and grinders everywhere. Finally, lime-
As far as food processing is concerned, the stone vessel production using scraping and polish-
numerous and varied types of grinding tools discov- ing techniques is also well represented on both
ered at Mentesh Tepe contrast with the poor toolkit sites. However, Mentesh is richer in rough-outs and
from the Mil Plain sites. This is also true for the pieces at different technical stages, and gives more
pounding and crushing tools, which are numerous complete information on the production process of
and standardized at Mentesh, but rare at the Mil these stone vessels. The range of tools that has
Plain sites. Different hypothesis in relation with the been identified, together with the limestone vessels
function of the sites or their economy could be pro- and the processing of ochre, had already been de-
posed to explain these differences. There is no scribed as characteristic for the sites of the S̆ula-
doubt that Mentesh Tepe was a permanent settle- veri-S̆omu culture on the Georgian side.356
ment during its main period of occupation (second The most striking particularity of the Mil Plain
half of the 5th mill. BCE), as shown by the architec- collection definitely relies in the flake industry on
ture and by the discovery of food processing imple- limestone and basalt cobbles. Very few retouched
ments in different areas of the site. The lack of grind- flakes of this type were used at Mentesh and on
ing and pounding tools or their hastily made aspect the S̆ulaveri-S̆omu sites.
at the Mil Plain sites, on the contrary, could be in- Nonetheless, the analysis of a wider range of
terpreted as an indicator for a non-permanent oc- tools from future excavations is necessary for a fi-
cupation, although it could also be due to a more dif- ner determination and understanding of the techni-
ficult access to raw material. It is interesting to cal variability of the macrolithic collections in their
notice, for instance, that the Mentesh grinding tools chronological, functional and cultural frame.
are mostly made of basalt, following the tradition of
most similar tools in neolithic Southern Caucasus,355 Preliminary report on the lithic industries
whereas in the Mil Plain, sandstone is mostly used. of Kamiltepe and Mentesh Tepe
But we cannot totally exclude either the possible im-
pact of the different field work options and strate- L. Astruc, A. Samzun and B. Gratuze
gies of excavations between Mentesh and Mil Plain
to explain these obvious differences. The aim of this contribution is to present a first
Craft activities show more common features overview of the lithic industries from sites in the
between the sites. The burnishers and polishers in- Kamiltepe area and from Mentesh Tepe, as part of
volved in ceramic processing and lithic tool produc- the ANR/DFG project Ancient Kura. At all of these
tion have been found on all the sites and share more sites, excavations are still in progress, and our study
or less the same technological features. Similarly, is just beginning, therefore we cannot yet propose
355 356
Badalyan et al. 2007; Hamon 2009. Hamon 2009; Kiguradze 1986.
a detailed analysis of both assemblages. In the Mil last group is composed of ‘quartzite’ (36.2%), a gray
Plain sites, the collection in each operation is often flint with a wackestone texture, a chalcedonic flint, a
insufficient in number; at Mentesh Tepe, questions jasper-like flint and unspecified pieces (burnt flint
about the chronological attribution of the lithic and unidentified stones; Tab. 34; Fig. 205).
pieces are not yet solved due to the long occupa- On a macroscopic view, various kinds of obsi-
tion of the site and the presence of many pits or dian are attested. Most of the studied elements are
burials. Our goal in this article is, therefore, to illus- made of transparent obsidian, slightly coffee-colored
trate how these assemblages differ, to underline (or smoked), with frequent zonations (ashy, cloudy
their multi-composition, and to show what they can or dappled) or with filaments (gray, black or brown
add to our knowledge about the ancient commu- in color). It can also be associated to red and black
nities in the Middle Kura Valley in terms of the cir- opaque zones (1 specimen). A milky gray variety of
culation of raw material, technical practices, regio- obsidian, sometimes slightly bedded, is also present.
nal differences and/or cultural affinities with other The sources for the non-obsidian raw material
groups in the Caucasus. are still unknown at this state of research. A first
assessment of the obsidian deposits was made
Kamiltepe and the Mil Plain sites through LA-ICP-MS analysis at IRAMAT:358 13 sam-
ples have been analyzed. They are indicative of the
The Azerbaijan-German Expedition in the Mil Plain357 Armenian outcrops of Gutansar (N = 1), Gegham
is centred on the site of Kamiltepe and various (N = 4) and Syunik (N = 8). The latter group is com-
other small sites around in the Qarasu Valley. A posed of Syunik 3 (Mets Sevkar, Pork Sevkar,
first diagnosis of the lithic industries uncovered dur- N = 7) and Syunik 2 (Mets Satanakar, N = 1). The
ing the excavations was realized in October 2011. correlation between the macroscopic look of the
Altogether, the lithic collection in each operation is obsidian, its chemical composition and its techno-
rather small. We mainly focused on Operations 1 typological characterization will be explored later.
and 10 (mid-6th mill. BCE), which are at the basis of
a little more than 700 artifacts for this period and Techno-typology
give us a preliminary insight on the way the raw
material was selected then and on the techno-typol- The collection of operation 1 is composed of waste
ogy. In addition, we viewed the collection of Opera- (63%), mainly made of ‘quartzite’ 32%) and obsi-
tion 16 in order to identify various aspects of the dian (49%). Fragments of blocks (N ¼ 7) and cores
lithic collection of the 1st half of the 4th mill. BCE. (N ¼ 3) are only attested for the former. ‘Quartzite’
was exploited for the production of flakes (bladelet-
Raw material like blanks are scarce in the sample). On the con-
trary, obsidian was used for the production of both
During the mid-6th mill. BCE, the procurement of raw flakes and blades/bladelets. Few types could be
material at Kamiltepe is rather balanced between identified (only 10% of the total assemblage, or
obsidian (48.4%) and non-obsidian material. The 15% for obsidian and 5% for ‘quartzite’).
Block Waste Undet. Flake Blade(let) Core TOTAL %
‘Quartzite’ 7 137 0 90 4 3 241 36.24
Wackestone 0 3 2 1 0 0 6 0 .9
texture, grey Flint
Chalcedonic Flint 0 0 0 1 0 0 1 0.15
Jasper-like 0 8 0 2 0 0 10 1.50
red Flint
Flint (Burnt) 0 0 0 2 2 0.30
Unidentified 0 68 0 13 1 0 82 12.33
Non Obsidian 7 216 2 110 5 3 343 51.58

Tab. 34
Kamiltepe, Obsidian 0 206 42 56 18 0 322 48.42
op. 1. Collection
sorted according TOTAL 7 422 44 166 23 3 665 100
to raw material
358 Centre Ernest-Babelon, CNRS, Orléans, cp. Gratuze 1999,

357
Aliyev/Helwing 2009; Helwing/Aliyev, in this article. Figs. 2; 3.
Fig. 205
Mil Plain lithic industry
(Kamiltepe op. 1
[1, 3–7] and MPS 16
[2]; (drawings L. Astruc).
1 Used blade; 2 Sickle
element with longitudi-
nal gloss; 3 Pièce
esquillée; 4–5 Geo-
metrics; 6–7 Operation
1, sickle elements with
oblique gloss (1, 3 ob-
sidian; 2, 6, 7 flint)
‘Quartzite’ occurs as tabular blocks or pebbles for unidirectional blades/bladelets production (in-
from riverbeds or old terraces. It was brought to cluding pressure flaking) may eventually be further
the site as a raw material, as shown by one pebble documented with a larger collection. Very frequently,
and several fragments of blocks. It was at least if not constantly, obsidian pieces present traces of
partly processed in situ, as indicated by the pre- use (at least one or two sequences).
sence of three cores. One of the three cores is made The typology (Tab. 35) is clearly dominated
on a thick flake. The chaı̂ne opératoire is based on by ‘pièces esquillées’ found on ‘quartzite’, chalce-
a production of flakes or elongated flakes with a donic flint and obsidian. Sickle elements are made
minimum of shaping. Unipolar and crossed exploita- of the finest ‘quartzite’ or the wackestone fine flint.
tion are common. The natural morphology of the Blanks are respectively blade(let)s or flakes. The in-
tabular blocks is used, and flakes with cortical or serts were hafted obliquely as is shown by the dis-
natural surfaces are common (on the butts, lateral tribution of the macroscopic gloss found at the edge
or residual). Unipolar exploitation led to the produc- of the blanks, sometimes with traces of bitumen.
tion of elongated blanks up to 60 mm in length. No rejuvenation of the edge has been registered so
As far as obsidian is concerned, no cores or far. Finally, geometrics are attested as shown by
technical pieces such as shaping flakes, opening of the presence of one trapeze-shaped piece made by
platforms or crested blades have been identified so a tiny direct retouch on a flake and made of a fine
far. The only exception is a small irregular crested wackestone texture, and by that of a lunate made of
bladelet. The methods of débitage and techniques a very characteristic jasper-like red flint.
ticed for Operation 1. The unique obsidian element

‘Quartzite’ Pièce esquillée 1
studied by LA-ICP-MS comes from the main source
Lateral retouch 3 exploited there, i.e. Syunik 3. An exhausted core
Sickle (oblique hafting, 8 used for crossed production of flakes, and two deb-
3 with bitumen) ris extracted from the bulbar surface of large blanks
Wackestone Trapeze 1 show that obsidian was partly processed on the site.
grey Flint
Back 1
Mentesh Tepe
Sickle (oblique hafting) 1
Chalcedonic Flint Pièce esquillée 1 At Mentesh Tepe,359 four seasons of excavations

have led to the identification of four main chronolo-
Jasper-like red Flint Lunate 1
gical phases, from the Neolithic to the Early Bronze
Undet. Flint Sickle (undet.) 2 Age. Most of the material comes from the levels da-
Obsidian Pièce esquillée and fragments 40 ted to the second half of the 5th mill., but many in-
Burin and burin spalls trusive pits caused problematic mixings that are not
Tab. 35 Lateral retouch
Kamiltepe, op. 1. Tool
yet solved.
Lateral retouch/Burin
types sorted according Tool fragments The total amount of lithic material registered
to raw material so far consists of 5,582 elements,360 mostly made
of obsidian (96.5%). A first evaluation of the waste
shows that 98.9% is made of obsidian and 1.1% of
A chronological diversity of the
other raw materials. Obsidian cores or fragments of
assemblages in the Kamiltepe area
knapped blocks are dominant with 31 specimen vs.
The general trends observed in Operation 1 are only one flint core. The retouched pieces represent
confirmed by the small assemblage collected in Op- ca. 30.5% for obsidian and 39% for non-obsidian
eration 10 (N ¼ 43, with 28 obsidian blanks). In the raw materials. If we take into account the blanks
mid-6th mill. BCE, sickles were made of flint blanks only, 76.8% of the obsidian and 76.5% of the non-
hafted obliquely. ‘Pièces esquillées’ are well attested obsidian ones are typological tools: retouched flakes
on different kinds of obsidian blanks. A side-scraper and blade(let)s, backed bladelets, truncations, bur-
made of a fine sedimentary rock is also present, so ins and burin spalls, end-scrapers, side-scrapers,
that more varied raw materials may have been used denticulates, notches, ‘pièces esquillées’, drills and
than is visible in our limited sample. pointed tools (Tab. 36; Fig. 208).
The collection from Operation 16 (dated to
the 1st half of the 4th mill. BCE) is at the time being Obsidian sources
very limited (N ¼ 65). ‘Quartzite’, flint and obsidian
are attested. The flint sources seem to slightly differ The macroscopic look of the volcanic glasses is ex-
from that of the mid-6th mill. BCE, with a larger di- tremely varied. A high proportion consists of trans-
versity of rocks, either translucent or opaque, with parent obsidians either uncolored or slightly smoked
fine to slightly rough textures, with a chalcedonic (café), which may present different kinds of zona-
appearance or a mudstone texture, and with colors tions (‘bedded’, ‘speckled’, ‘filamentous’). The colors
varying from beige to brown. The relation between of these zones are mainly black and gray, and in
bladelets made of flint and their use as sickle ele- some cases, white or reddish. A sample of 42 pieces
ments is obvious as shown by five of the seven was analyzed by LA-ICP-MS (Figs. 206; 209): nine dif-
specimens collected that present a macroscopic ferent chemical groups have been identified, coming
gloss. These instruments are nevertheless different from Gegham (N ¼ 18), Tsakhunjats 1 (N ¼ 6) or 2
from the earlier ones because the distribution of (N ¼ 1), Chikiani (Paravani) (N ¼ 3), Sjunik 3 (Mets
the gloss is lengthwise. Several of these inserts Sevkar, Pork Sevkar) (N ¼ 2), Gutansar (N ¼ 1) and
bear traces of bitumen. In fact, the difference in the Arteni 3 (N ¼ 1). Furthermore, nine tools are che-
manufacture of sickles for both periods is visible in mically related to Sarıkamış North (6 from Hamamlı,
the choice of the raw material and in the way the 3 for Handere) and one is from Sarıkamış South
inserts were produced and inserted. Rejuvenation (Mescitti-Şehitemin). These last elements can come
of the edge is clear on some of the blade(let)s with either from Sarıkamış itself or from secondary
a tiny microdenticulation. sources, i.e. terraces from the Akhurian and the Ara-
Besides some obsidian flakes and waste, the xes rivers.
presence of an exhausted obsidian core used to pro-
duce flakes by crossed exploitation has to be no- 359
Lyonnet/Guliyev, in this article.
ticed in operation 16. The macroscopic appearance 360 Inventory of the 2008–2011 excavations, without the heavy
of the obsidian is in general the same as that no- residues from flotation.
A production in situ?
Debitage/Types of tools Obsidian Non Obsidian Raw
material
Whatever the contexts, obsidian flakes and debris
Cores 31 1
are numerous at Mentesh Tepe and several cores are
present. Among the samples that have been ana- Rejuvenation Flake 3 %
lyzed through LA-ICP-MS, we selected four cores or Flakes 1,796 32
their fragments to identify their origins. The results
Bladel(let)s 343 66
show a source at north Sarıkamış (Hamamlı) for one,
Chikiani for two others, and Gegham for the last Waste 1,572 18
one. The in situ productions, thus, were not limited Total Débitage 3,745 117
to one relatively nearby source.
Retouched Flakes 418 %
Two concentrations of obsidian remains corre-
sponding to two distinct knapping sequences have Retouched Blade(let)s 608 21
also been identified. One of them, a Chalcolithic Sickle elements 10 38
workshop (ST 287), contained 94 flakes or waste
Truncations 249 5
of obsidian weighing one kilogram altogether. The
raw material is transparent (slightly café in color), Burins 121 %
speckled with black and ashy gray. Some of the Burin spalls 48 1
flakes present brownish orange zones too. The che-
End-Scrapers on Flakes 11 %
mical analysis points to sources at Gegham. The
knapping sequence of this collection consists in the End-Scrapers on Blades 2 %
shaping of a core, probably a blade(let) core. This Side-Scrapers on Flakes 9 1
concentration of remains probably corresponds to Side and End-Scraper 1 %
the refuse collected at the time of the cleaning of the
floor of the workshop, while the rest of the knap- Denticulates 15 1
ping was done either here or at any other place, Drills 5 2
inside or outside of the village. The second con- Pointed tools 3 5
centration (total weight 1.570 kg) was found in a
Kura-Araxes structure. In that case, the obsidian is Notches on Flakes 15 %
made of a very characteristic transparent raw mate- Notches on Blades 19 %
rial, full of bubbles and small spherulites organized Pièces esquillées’ 40 %
in layers. The cortical part on both sides of the
block is fine, beige and presents longitudinal lines, Fragment of tools 71 1 Tab. 36
while the natural surface on the contrary is rough Total Tools 1,645 75 Mentesh Tepe.
Preliminary inventory
with a lot of bubbles. LA-ICP-MS analyses show that Total inventory 5,390 192 of the lithic collection
the block comes from Chikiani. The raw material (2008– 2011)
shape is known, because two fragments fit together
and consist of a cylindrical tabular block (150 mm of raw material. Flake cores and flakes showing nat-
long " 92. 4 mm wide " 70 mm thick). It was ural or cortical surfaces are rare. The products are
knapped directly without any preparation or shap- in general irregular in shape, but were used without
ing, producing slices that ended up in flakes with any further modification (so-called Non Formal Tools)
residual cortex. Butts are large, often roughly fa- and most of them consist of pièces esquillées.
cetted. The largest flake was made transversally on Unipolar bladelets and blades are also com-
the whole surface of the block (89 " 66.4 mm " mon. No blade or bladelet core has been discov-
20.3 mm). These remains are the results of the ered and technical pieces are rare: no tablets,
knapping sequence of a block for a crude flake pro- nearly no crested blade(let)s. So far, we have iden-
duction in situ. Eight flakes probably coming from tified only one shaping flake of very good quality
the same block have been discovered in another that can be directly related to the skilled débitage
Kura-Araxes context. of blades and bladelets. Actually, we are in front of
a well known behavior pattern: a strict selection of
Obsidian and non-obsidian technology blanks coming from the plein débitage and even
the lateral blade(let)s are scarce in the collection,
The chronological distribution of the collection still while the size of the blanks varies. Considering the
being uncertain, our comments here will be limited. techniques, a great homogeneity in the preparation
As far as obsidian is concerned, the flake produc- for the detachment is noticeable. Butts are in most
tion is well represented. Flakes are the results of a cases facetted, with a clearly limited point of com-
domestic unipolar or crossed exploitation of blocks paction. The preparation to detachment is short on
Fig. 206
Binary diagram
of Y/Zr–Nb/Zr ratios
for the artifacts
from Kamiltepe and
related outcrops
Fig. 207
Location of Kamiltepe
and Mentesh Tepe and
obsidian sources
(DAO, L. Astruc)
the surface of débitage, or not present. Pressure

flaking is attested. A further detailed analysis and
careful documentation of the blanks should allow
us to identify precisely the techniques of detach-
ment for each chronological phase. Several blanks
showing different kinds of preparation were made
of unusual obsidians.
Various other raw materials are present at
Mentesh Tepe: a fine to rougher chalcedonic flint
(translucent and yellowish, whitish when damaged
by post-depositional action), used for flake produc-
tion and present in small quantity; a fine jasper-like
flint, orange-red in color; a pink to red-brown flint
showing a wackestone semi-transparent texture con-
taining many fossils; a flint of a fine mudstone tex-
ture, opaque and brownish, becoming white when
damaged by post-depositional wear and which
could be argillite.361 This last material is the most
commonly found. Technical pieces such as shaping
flakes or under-crests are scarce, but no core was
found. There is a whole range of sizes for these
products, and their technological characters (meth-
ods and techniques) also can be different. Further
analysis should show if these differences are due to
the organization of craftsmanship or to chronologi-
cal variations. Unipolar blades and bladelets ware
made of jasper-like flint and wackestone flint. On
the contrary, the uncommon chalcedonic flint is ex-
clusively used for crude flake production.
Thus, the non-obsidian raw materials were not
equally used at Mentesh Tepe.
From a typological point of view, there is no
clear dichotomy between obsidian and non-obsidian
material, especially as far as ‘pièces esquillées’, lat-
eral retouch, truncation, pointed blade(let)s, micro-
denticulate, end-scrapers and denticulates are con-
cerned. Retouches made by pressure flaking are
known on both obsidian and flint.
The diversity in unipolar blade and bladelet
productions at Mentesh has already been under-
lined. On the opposite, tool types and rejuvenation
sequences show that the behavior on obsidian and
on flint was similar, except maybe for burins, only
made of obsidian and not of flint. Finally, the use
of pressure flaking is well attested both on flint
and obsidian.
Sickle elements Obsidian inserts are bladelets or blades. They can be Fig. 208
Mentesh Tepe lithic
rejuvenated with lateral retouches, sometimes using industries (drawings
At Mentesh Tepe, sickle elements are made of obsi- pressure flaking. Inserts made of flint are common M. Godon). 1–2 Sickle
dian or flint. Except for one case of a flake made of and more easily recognizable since a macroscopic element with longi-
tudinal gloss (flint);
chalcedonic flint that was hafted obliquely, they are gloss is present all along the edge. Like obsidian, 3 Blade (flint);
inserted longitudinally to the haft which was made they can be rejuvenated by pressure flaking or are 4 Bladelet (obsidian)
of another material and, thus, form composite tools. micro-denticulated. Remains of bitumen are gener-
ally present opposite to the working edge.
One obsidian bladelet was found still insert-
361
R. Arazova, personal communication. ed in an unknown filler. It was used for harvesting
Fig. 209
Binary diagram of
Y/Zr–Nb/Zr ratios for
the artifacts from
Mentesh Tepe and
related outcrops
and rejuvenated with direct retouches. The filler sar on the basis of LA-ICP-MS analysis on a sample
(maximum thickness 3.2 mm) still kept the print of of 13 artifacts). On the contrary, at Mentesh Tepe
the groove in the handle (longitudinal traces due 96.5% of the collection is made up of obsidian, the
to its carving) and of fingerprints. In this case, the origins of which are extremely varied (Gegham, Tsa-
haft was slightly curved and the groove has a khunjats 1 and 2, Syunik 3, Gutansar, Arteni, Chi-
V-shaped section (internal width 5 mm, external kiani, and Sarıkamış 1 and 2). Although the sample
width 9.5 mm). The total height of the fragment is analyzed by LA-ICP-MS is important in number
15.1 mm and the part of the bladelet coming out of (N = 42), we still do not have a clue concerning the
the filler is 5.8 mm. At least one other blade was importance played by each source.
inserted at the proximal part of the bladelet. The The choice of our samples was intended to
bladelet measures 54 mm in length and was rejuve- document the diversity of macroscopic aspects and
nated by a fine direct retouch up to 7 mm long in to answer contextual or techno-typological ques-
some parts. It had been inserted longitudinally and tions. In both sites this preliminary study already
presents characteristic traces of wear: longitudinal stresses the fact that each raw material – i.e. each
striation, slight smoothing of the edge and fine al- obsidian and non-obsidian component – has to be
ternative scaling. From this data, we can reconstruct analyzed separately in order to understand the be-
a composite sickle made of a slightly curved haft haviour of these ancient communities at the time of
with lithic elements inserted at right angles. their tool production. The shortages in both collec-
tions prevent us to further comment the technology
Conclusions and typology or to make comparative studies.
Nevertheless, among the easily recognizable activ-
The lithic collections of Kamiltepe and Mentesh ities performed at Kamiltepe and Mentesh we can
Tepe differ considerably. The inhabitants of Kamil- stress that of cereal harvesting.362 In the Southern
tepe acquired in similar proportion obsidians and Caucasus, sickles made of a haft in which sickle
different kinds of flint. The latter are considered as
mainly local, while the volcanic glasses originated 362 ˝/(ŁC/B@- 1966; ˝/(ŁC/B@- 1971; ˚@(@ÆŒ@-/ 1978; T(/H@-
from more distant sources (Syunik, Gegham, Gutan- -/ 1986.
elements are inserted obliquely, such as those a standard. This stands in contrast to the western
mainly found at Kamiltepe and in the S̆omu-S̆ula- sites under investigation, Mentesh Tepe and Aru-
veri culture,363 are considered as archaic. The chal- chlo. The Neolithic levels of these sites are contem-
cedonic flake bearing a highly developed oblique porary with the Mil Plain sites, but the evidence
gloss discovered at Mentesh remains an exception obtained so far shows long-term stable settlements
and is therefore either related to another activity or with a highly complex architectural record. The new
dating to the earliest Neolithic levels just recently excavations at Aruchlo have provided as yet un-
discovered. Later on, sickles are mostly made with known data concerning house architecture, building
inserts hafted longitudinally,364 similar to what has techniques, settlement structure, burial rituals and
mainly been found at Mentesh Tepe and in the Lei- also two ditches. Mentesh Tepe, halfway between
latepe culture.365 Aruchlo and Kamiltepe, is exceptional in its long-
Our study, even if very preliminary indeed, term stratigraphic record with four major periods
shows how the lithic industries can play a role for dating from the early 6th to the middle of the
the identification of the technical traditions and of 3rd mill. BCE and with almost no hiatus except dur-
the behaviors related to subsistence activities. They ing the first half of the 4th mill BCE. No excavated
are also a good marker of the mobility of these site in the Southern Caucasus ever presented such
communities or of the circulation of raw material. a promising possibility for studying the evolution
Further field work in both sites will be of primary of the local culture and of the environment on such
interest as the Neolithic and the beginning of agri- a long-time scale. The data it provides on the very
culture and herding is only known in few well-strati- poorly known 5th mill. BCE is especially remark-
fied sites for Azerbaijan. able.
The gap in the occupation of Mentesh Tepe
during the first half of the 4th mill. BCE is filled by
Part V: the presence of a site of this period in the Mil
Plain, where abundant chaff-faced pottery known
The Ancient Kura Project: general conclusions over a very wide area has been found. Furthermore,
imported pottery found on the same site also indi-
Bertille Lyonnet and Farhad Guliyev, cates that the Mil Steppe sustained relations with
Barbara Helwing and Tevekkül Aliyev, the Iranian Central Plateau at the same time. We
Svend Hansen and Guram Mirtskhulava hope that excavations there will provide data on
the transition to the Early Bronze Age, which thus
Already at this very preliminary state, it is evident far is only attested at Mentesh through a collective
that the collaborative efforts of the Ancient Kura burial under kurgan and further followed by prob-
Project have opened a large window onto a so far able seasonal occupations.
little known cultural landscape of the Southern Cau- Throughout the millennia covered by this re-
casus. In each area we have discovered Neolithic search, it appears that the three foci of our research
settlements starting around 6000 BCE, equipped project develop in a parallel way, but with strongly
with a full Neolithic subsistence set that comprised divergent tendencies. Fundamental differences in
domesticated animals and plants, and with complex architecture are one of the obvious visible features
architectural features, but they totally differ in many that distinguish the western from the eastern end
aspects. of the Middle Kura Valley: the round buildings and
In the Mil Plain, for the first time ever in this enclosures in Aruchlo that are so characteristic of
region has a settlement been attested that dates to the S̆omu Tepe – S̆ulaveri Gora group, and that ap-
the beginning of the 6th mill. BCE. The discovery of pear also in the Neolithic layers in Mentesh Tepe,
a complex ditch system resembling a causeway contrast with the rectangular architecture constructed
camp is another new feature. The monumental plat- at the Mil Plain sites. Even the concept of the build-
form discovered at the mid-sixth mill. site of Kamil- ing material differs, as in Aruchlo hand-shaped and
tepe is equally unique. Altogether, the Mil Steppe straw-tempered mudbricks were the standard, while
sites seem to have frequently shifted in location, so the bricks used in Kamiltepe were dug fresh from
that the formation of true settlement mounds usual- the ground and were used untempered and without
ly considered a hallmark of Neolithic sedentary life previous kneading.
must be regarded there as an exception rather than At Mentesh Tepe, the sudden change ob-
served in architecture, from round to rectangular
363
structures, during the transitional phase from the
Arazova 2008, sites of S̆omutepe, Toyretepe, Gargalartepe, to Middle to the Late Chalcolithic period shows that
which can now be added Göy Tepe (F. Guliyev, pers. comm.).
364
Arazova 2008, Fig. 4; Tab. 10. wide regional relations were already developing at
365
Akhundov 2008, Fig. 7,30-34, site of Boyuk Kesik. that time.
Indeed, the differences noted between the Preliminary results of the archaeobotany ana-
eastern and western Middle Kura Valley apply to lyses already provide evidence for the development
many other features of the material culture. During of individual traditions in the various regions with
the Neolithic period, Aruchlo and Mentesh Tepe used regard to mixed farming strategies and to variable
unpainted pottery, with occasional plastic decora- harvesting techniques, but further research on this
tion or a red slip. The Mil Plain ceramics, on the data is still needed in order to gain a better under-
contrary, are painted with geometric motifs in a standing of the possible local species in each area
style that resembles the material from the Iranian and their adaptation to environmental changes
Highlands. The Chalcolithic period has only been in- throughout the millennia.
vestigated so far at Mentesh Tepe, but the changes Environmental conditions also strongly shape
observed in the ceramic material during the second the archaeological record as it can be discerned
half of the 5th mill. BCE, combined with those men- from the landscape archaeology. The widely mean-
tioned in the architecture, point to influences from dering rivers in the Mil Plain left natural elevations
northern Mesopotamia, even though local features suitable for occupation, while the river terraces
are still clearly visible. farther up the river seem to have provided safe
Similarly, Aruchlo and Mentesh Tepe yielded settling ground. Furthermore, the environment is a
a rich bone industry, while the Mil Plain sites only key factor for modern land-use that in turn strongly
yielded metapodial points so far. Other differences affects the preservation of the archaeological land-
can be observed in the lithic record, both in the scapes: the Mil Plain area is still partially preserved
macrolithic tools and in the chipped stone indus- because of the absence of irrigation, whereas the
try, not only on the type of tools but also in the Aruchlo-Mentesh area is highly cultivated in the
choice of the raw material. Changing preferences piedmont zone, as well as largely affected by recent
are obvious in the choice of materials, with an constructions of pipelines.
overwhelming percentage of obsidian blades in As demonstrated by all of these combined
Aruchlo and Mentesh as opposed to the rather ad results, even if they are certainly very preliminary,
hoc flake industry on flint and obsidian attested the Ancient Kura project has a high potential to
in Kamiltepe, indicating different networks and ac- provide fundamental new insights into the devel-
cessibility of raw materials procurement. This may opment of prehistoric cultures in the Southern Cau-
furthermore be the case for the early usage of casus, and to open new avenues of collaborative
copper, newly attested for Neolithic Aruchlo, and research between areas and between specialists.
with a complete range of the chaı̂ne opératoire, The strategies of excavations differ in each area,
including the melting of copper found in Mentesh but the combination of the results and the com-
Tepe. parative analysis certainly brings in new data for a
Hence, it is evident that both ends of the Mid- much better understanding of the cultural process.
dle Kura River did not participate in the same net- Nevertheless, the archaeological landscape that
works. Obviously, the geographical setting played a now slowly begins to emerge before our eyes still
significant role in channeling traffic, and a precise holds an array of unsolved questions for the fu-
contextualisation of the archaeological sites through ture.
landscape reconstruction is necessary. Geomorpho-
logical research has demonstrated the crucial role
played by the oscillations of the Caspian Sea level Acknowledgements
and, to a lesser extent, by global climatic events.
The environmental conditions indicate a gradual The project ‘‘Ancient Kura’’ is a multidisciplinary in-
change from a moderate climate in the western ternational project supported through a grant for
area under study to a semi-arid climate in the Mil scientific collaboration between France and Ger-
Plain. These are reflected closely in both the bota- many from the Agence Nationale de la Recherche
nical and zoological record. The preferred hunting (ANR, France) and the Deutsche Forschungsgemein-
game in Aruchlo was red deer, as opposed to ga- schaft (DFG, Germany). This co-funding scheme has
zelle in the Mil Plain. The zoological study has also enabled us to collaborate in a truly new way, and
raised the question concerning the eventual local we would like to thank the responsibles at ANR
occurrence of the animal livestock available to the and DFG for their trust and support. Research is
Neolithic settlers in the Southern Caucasus (possi- conducted jointly by the German Archaeological In-
ble primary domestication) instead of its importa- stitute Berlin (DAI), the Centre National de la Re-
tion from other areas, and the growing importance cherche Scientifique Paris (CNRS), the Institute for
of ovi-caprids during the Chalcolithic period has to Archaeology and Ethnography of the Azerbaijan Na-
be studied in relation with possible environmental tional Academy of Sciences in Baku and the Centre
changes. for Archaeological Research ‘‘Lordkipanidse’’ at the
National Museum in Tblissi. In the Ancient Kura pro- Uhl, Michael Ullrich, Maruchi Yoshida, Orhan Zama-
ject, three previously independent research projects nov, Dimitri Zhvania and Jean Zivzivadze.
based on excavations in Aruchlo (German-Georgian None of this work would have been possible
excavations since 2005), Mentesh Tepe (French – without the generous and trustful collaboration be-
Azerbaijan excavations since 2007) and Kamiltepe tween our institutions in France and Germany, Geor-
(German – Azerbaijan excavations since 2009) since gia and Azerbaijan. We would like to warmly thank
2010 have joined their forces by combining their the responsible directors, Dr. Maya Ragimova at the
individual perspectives on natural resources and Institute for Archaeology and Ethnography of the
environmental factors behind the cultural develop- National Academy of Sciences Azerbaijan, and Prof.
ment in the Southern Caucasus since early sedent- Dr. David Lordkipanidze, Director General of at the
ism. Previous and additional financial support for Georgian National Museum and Dr. Zurab Makha-
these investigations was provided to the Mentesh radze, Director of the Archaeological Centre of the
Tepe excavations by CNRS (LIA AzAr2) and the Georgian National Museum, for their continuous
French Ministry of Foreign and European Affairs, to support and interest.
Aruchlo by DFG and DAI, and to Kamiltepe by DAI.
We gratefully acknowledge the support of all these
institutions.
The co-direction of the fieldwork lies with Te- Bibliography
vekkül Aliyev (Academy of Sciences Baku) and Bar-
bara Helwing (DAI) for the explorations in the Mil Akhundov 2007
Steppe, with Bertille Lyonnet (CNRS) and Farhad T. Akhundov, Sites des migrants venus du Proche-Orient
Quliyev (National Academy of Sciences Azerbaijan) en Transcaucasie. In: B. Lyonnet (ed.), Les cultures du
Caucase (VIe–IIIe millénaires avant notre ère). Leurs rela-
for the Mentesh Tepe excavations and with Svend
tions avec le Proche Orient (Paris 2007) 95–122.
Hansen (DAI) and Guram Mirtskhulava (Tbilisi Na- Adamia et al. 1981
tional Museum) for the work at Aruchlo. The field- S. A. Adamia/T. Chkhotua/M. Kekelia/M. Lordkipanidze/
work teams consist of archaeologists, specialists I. Shavishvili/G. Zakariadze, Tectonics of the Caucasus
and students, who all contributed greatly to the and adjoining regions: implications for the evolution of
work and we thank all of them: Joni Abuladze, Arzu the Tethys ocean. Journal of Structural Geology 3, 1981,
Abdullayeva, Golnaz Ahadi, Vusal Aliyev, Aygün No. 4, 437–447.
Aliyeva, Leı̈la Aliyeva, Laurence Astruc, Tarana Ba- Aitken 1958
bayeva, Ejdar Babazade, Katrin Bastert-Lamprichs, M. J. Aitken, Magnetic prospecting I. Archaeometry Ar-
Wiebke Bebermeier, Giorgi Bedianashvili, Florian chaeometry 1, 1958, No. 1, 24–29.
Aitken 1974
Becker, Norbert Benecke, Irma Berdzenishvili, Ulf
M. J. Aitken, Physics and archaeology (Oxford 1974).
Böttcher, Laurence Bouquet, Gaëlle Bruley-Chabot, Akhalkatsi 2009
Guram Chikovani, Antoine Courcier, Maria Bianca M. Akhalkatsi (ed.), Conservation and sustainable use
D’Anna, Alexia Decaix, Zaki Djaffarov, Jörg Fassbin- of crop wild relatives in Samtskhe-Javakheti (Tbilisi
der, Michel Fontugne, Svetlana Gasimowa, David 2009).
Gagoshidze, Ivan Gatsov, Felix Geitel, Nugsar Giko- Akkermans 1988
rashvili, Arian Goren, Jens Greif, Etibar Hadjiev, Nar- P. M. M. G. Akkermans, The Period IV Pottery. In: M. N.
giz Hadjizadeh, Caroline Hamon, Ilia Heit, Jens Hel- van Loon (ed.), Hammam et-Turkman I. Report on the
wing, Fuad Huseinov, Muzaffar Huseynov, Nedjat University of Amsterdam’s 1981–1984 Excavations in
Syria I. Publications de l’Institut historique-archéologi-
Ibrahimli, Valeh Ilakbarov, Zaur Ismailov, Vladimir
que neerlandais de Stamboul (Leiden 1988) 181–285.
Ioseliani, Bahman Jahani, Bakhtiar Jellilov, Tamar Akkermans/Schwartz 2003
Kentschoshvili, Julia Koch, Marina Kurdaze, Anne P. M. M. G. Akkermans/G. M. Schwartz, The archaeology
Kwast, Gaëlle Le Dosseur, Roland Linck, Ainhoa Lin- of Syria (Cambridge 2003).
cot, Zurab Makharadze, Yakub Mamedov, Mirdjafar Aliyev/Helwing 2009
Mirgadirov, Guram Mirtskhulava jr., Elina Muradova, T. Aliyev/B. Helwing, Kamiltepe in der Milebene. Archäo-
Djelal Musayev, Dimitri Narimanishvili, Petranka Ne- logische Untersuchungen 2009. Archäologische Mittei-
delcheva, Daniel Neumann, Georg Neumann, Anna lungen aus Iran und Turan 41, 2009, 23–45.
Okujava, Vincent Ollivier, Laure Pecqueur, Modwene Apakidze 1999
Poulmarc’h, Henriette Raddatz, Pascal Raymond, J. Apakidze, Lapislazuli-Funde des 3. und 2. Jahrtausends
v. Chr. in der Kaukasusregion – ein Beitrag zur Herkunft
Karine Raynaud, Andrea Ricci, Anaı̈ck Samzun, Badri
des Lapislazuli in Troia. Studia Troica 9, 1999, 511–
S̆arashenidze, Alexander Schäfer, Frank Schlütz, 525.
Christopher Schöbinger, Susan Schorr, Reza Sefid- Apakidze 2009
garan, Shahin Selimbeyov, Lisa Shillito, Hadn Tagi- J. Apakidze, Die Spätbronze- und Früheisenzeit in West-
zadeh, Helen Taylor, Levan Tchabashvili, Meda To- und Zentralkaukasien: chronologische Studien zur Kol-
deraş, Gia Turashvili, Tea Udesiani, Regina Anna chis-Kultur 1600–700 v. Chr (Rahden/Westf. 2009).
Arazova 2008 Belfer-Cohen 1988

R. Arazova, Sickles of early farmers of Azerbaijan. In: A. Belfer-Cohen, The Natufian settlement at Hayonim
L. Longo/N. Skakun/M. Saracino/M. Dalla Riva (eds.), Pre- Cave (Jerusalem 1988).
historic technology 40 years later: functional studies Belshe 1957
and the Russian legacy. Proceedings of the International J. C. Belshe, Recent magnetic investigations at Cam-
Congress, Verona (Italy) 2005. British Archaeological Re- bridge University. Advances in Physics 6, 1957, No. 22,
port International Series 1783 (Oxford 2008) 435–438. 192–193.
Areshian et al. 2012 Benech 2005
G. E. Areshian/B. Gasparyan/P. S. Avetisyan/R. Pinhasi/ C. Benech, Etude des plans d’urbanisme. Dossiers de
K. Wilkinson/A. Smith/R. Hovsepyan/D. Zardaryan, The l’archéologie 308, 2005, 12–19.
Chalcolithic of the Near East and South-Eastern Europe: Beug 2004
discoveries and new perspectives from the cave com- H.-J. Beug, Leitfaden der Pollenbestimmung für Mittel-
plex Areni-1, Armenia. Anitiquity 86, 2012, No. 331, europa und angrenzende Gebiete; mit 13 Tabellen
115–130. (München 2004).
Asgari/Zare 2010 Blum 2010
B. Asgari/R. Zare, Two new species of Preussia from Iran. M. D. Blum, Sediment supply to the shelf margin and
Nova Hedwigia 90, 2010, No. 3/4, 533–348. beyond: alluvial valley responses to sea-level change.
Arimura et al. 2012 In: H. R. Roberts/N. C. Rosen/R. F. Fillon/J. B. Anderson
M. Arimura/B. Gasparyan/C. Chataigner, Prehistoric sites (eds.), Shelf margin deltas and linked down slope pe-
in Northwest Armenia: Kmlo 2 and Tsaghkahovit. In: troleum systems. 23rd Annual GCSSEPM Foundation Bob
R. Matthews/J. Curtis (eds.), Proceedings of the 7th ICAANE, F. Perkins Research Conference, December 7–10, 2003,
London 2010 (Wiesbaden 2012) 135–150. Houston, Texas, USA (Red Hook 2010) 70–95.
Aspinall et al. 2008 Boardman/Jones 1990
A. Aspinall/C. F. Gaffney/A. Schmidt, Magnetometry for S. Boardman/G. Jones, Experiments on the effects of
archaeologists. Geophysical methods for archaeology 2 charring on cereal plant components. Journal of Archae-
(Lanham 2008). ological Science 17, 1990, No. 1, 1–11.
Badaljan et al. 1993 Bockheim/Tarnocai 1998
R. S. Badaljan/C. Edens/R. Gorney/P. L. Kohl/D. Stronach/ Bockheim J. G./Tarnocai C., Recognition of cryoturbation
A. V. Tonikijan/S. Hamayakjan/S. Mandrikjan/M. Zardar- for classifying permafrost-affected soils. Geoderma 81,
jan, Preliminary report on the 1992 excavations at Ho- 1998, No. 3–4, 281–293.
rom, Armenia. Iran 31, 1993, 1–24. Bull et al. 1999
Badalyan et al. 2007 I. D. Bull/I. A. Simpson/P. F. van Bergen/R. P. Evershed,
R. Badalyan/P. Lombard/P. Avetisyan/C. Chataigner/J. Cha- Muck ‘‘n’’ molecules: organic geochemical methods for
bot/E. Vila/R. Hovsepyan/G. Willcox/H. Pessin, New data detecting ancient manuring (1999).
on the Late Prehistory of the Southern Caucasus. The ex- Bull et al. 2005
cavations at Aratashen (Armenia): preliminary report. In: I. D. Bull/M. M. Elhmmali/P. Perret/W. Matthews/D. J. Ro-
B. Lyonnet (ed.), Les cultures du Caucase (VIe –IIIe millén- berts/R. P. Evershed, Biomarker evidence of faecal de-
aires avant notre èra. Leurs relations avec le Poche-Orient position in archaeological sediments at Çatalhöyük, Tur-
(Paris 2007) 37–61. key. In: I. Hodder (ed.), Inhabiting Çatalhöyük. Reports
Badalyan/Avetsiyan 2007 from the 1995–1999 seasons. Çatalhöyük Project 4.
R. S. Badalyan/P. S. Avetsiyan, Bronze and early Iron Monograph of the British Institute of Archaeology at An-
Age archaeological sites in Armenia 1. Mt. Aragats and kara 38 (Cambridge 2005) 415–420.
its surrounding region. British Archaeological Report In- Bullock et al. 1985
ternational Series 1697 (Oxford 2007). P. Bullock/N. Federoff/A. Jongerius/G. Stoops/T. Tursina,
Badalyan et al. 2010 Handbook for soil thin section description (Wolverhamp-
R. S. Badalyan/A. A. Harutyunyan/C. Chataigner/F. Le ton 1985).
Mort/J.-E. Brochier/A. Balasecu/V. Radu/R. Hovsepyan, Campana 1989
The settlement of Aknashen-Khatunarkh, a neolithic site D. V. Campana, Natufian and protoneolithic bone tools.
in the Ararat Plain (Armenia): Excavation results 2004– British Archaeological Report International Series 494
2009. TÜBA-AR. Turkish Academy of Sciences Journal of (Oxford 1989).
Archaeology 13, 2010, 185–218. Campbell et al. 2003
Balossi Restelli 2006 S. Campbell/E. Carter/S. Gauld, Elusive complexity: new
F. Balossi Restelli, The development of ‘‘cultural regions’’ data from late Halaf Domuztepe in South Central Tur-
in the Neolithic of the Near East. British Archaeological key. Paléorient 29, 2003, No. 2, 117–133.
Report International Series 1482 (Oxford 2006). Canti 1998
Becker 1999 M. G. Canti, The micromorphological identification of fae-
H. Becker, Duo and quadro-sensor configuration for high cal spherulites from archaeological and modern materi-
speed/high resolution magnetic prospecting with cae- als. Journal of Archaeological Science 25, 1998, No. 5,
sium magnetometer. In: J. W. E. Fassbinder/W. E. Irlinger 435–444.
(eds.), Archaeological prospection. Proceedings of the Cassard et al. 2010
3rd International Conference on Archaeological Prospec- D. Cassard/A. Courcier/M. Billa/C. Gateau/A. Lafitte/B.
tion, vol. 108. Arbeitshefte des Bayerischen Landesam- Lyonnet/F. Maldan/J.-F. Mallon, NavigaSIG#: Les cul-
tes für Denkmal pflege (München 1999). tures du Caucase et la métallurgie, 6e –3e millénaires
(Sud Russie, Républiques du Nord Caucase, Géorgie, Ar- David et al. 2008
ménie, Azerbaı̈djan, Iran du NO, Turquie orientale) (Un- A. David/N. Linford/P. Linford/L. Martin/A. Payne, Geo-
published report, 2010). physical survey in archaeological field evaluation (Swin-
Chataigner 1995 don 2008).
C. Chataigner, La Transcaucasie au Néolithique et au Deschamps et al. 2012
Chalcolithique. British Archaeological Report Internatio- P. Deschamps/N. Durand/E. Bard/B. Hamelin/G. Camoin/
nal Series 624 (Oxford 1995). A. L. Thomas/G. M. Henderson/J. Okuno/Y. Yokoyama,
Chelidze/Gogelia 2004 Ice-sheet collapse and sea-level rise at the Bolling warm-
L. Chelidze/D. Gogelia, Arukhlo I: An early-farming site. ing 14,600 years ago. Nature 483, 2012, No. 7391, 559–
Journal of Georgian Archaeology 1, 2004, 46–92. 564.
Chelidze/Chikovani 2008 Djamali et al. 2009
L. Chelidze/G. Chikovani, Pottery of the early farmers from M. Djamali/J.-L. de Beaulieu/P. Campagne/V. Andrieu-Po-
Eastern Georgia. In: A. G. Sagona/M. Abramishvili (eds.), nel/P. Ponel/S. A. G Leroy/H. Akhani, Modern pollen rain-
Archaeology in Southern Caucasus: perspectives from vegetation relationships along a forest-steppe transect
Georgia. Ancient Near Eastern Studies, supplement 19 in the Golestan National Park, NE Iran. Review of Pa-
(Leuven 2008) 27–36. laeobotany and Palynology 153, 2009, No. 3–4, 272–
Chernykh 1992 281.
E. N. Chernykh, Ancient metallurgy in the USSR (Cam- Le Dosseur 2008
bridge 1992). G. Le Dosseur, La place de l’industrie osseuse dans la
Chikovani et al. 2010 néolithisation au Levant Sud. Paléorient 34, 2008, No. 1,
G. Chikovani/Z. Shatberashvili/G. Gogochuri/G. Gamkre- 59–89.
lidze, A new site of the Eneolithic-Bronze Age from Te- Dschaparidze 2011
tritsqaro. In: G. Gamkrelidze (ed.), Rescue archaeology I. Dschaparidze, Kupfer und Bronzeartefakte aus der
in Georgia: the Baku-Tbilissi-Ceyhan and South Cauca- Bronzezeit Georgiens. Unpublished PhD thesis, Ruhr
sian pipelines (Tbilisi 2010) 95–109. University Bochum (Bochum 2011).
Clark 1996 Dunlop/Özdemir 1997
A. Clark, Seeing beneath the soil – prospecting methods D. J. Dunlop/Ö. Özdemir, Rock magnetism: fundamentals
in archaeology (London 1996). and frontiers (Cambridge 1997).
Connor et al. 2004 Ellis/Ellis 1988
S. E. Connor/I. Thomas/E. V. Kvavadze/G. J. Arabuli/G. S. M. B. Ellis/J. P. Ellis, Microfungi on miscellaneous sub-
Avakov/A. Sagona, A survey of modern pollen and vege- strates: an identification handbook (1988).
tation along an altitudinal transect in southern Georgia, Ellis/Ellis 1997
Caucasus region. Review of Paleobotany and Palynology M. B. Ellis/J. P. Ellis, Microfungi on land plants: an iden-
129, 2004, No. 4, 229–250. tification handbook (1997).
Cotkin et al. 1999 FAO Rome 2001
S. J. Cotkin/C. Carr/M. L. Cotkin/A. E. Dittert/D. T. Kremser, FAO Rome, FAOCLIM 2. World-wide agroclimatic data-
Analysis of slips and other inorganic surface materials on base. Food and Agriculture Organisation of the United
Woodland and Early Fort ancient ceramics, South-Central Nations (2001).
Ohio. American Antiquity 64, 1999, No. 2, 316–342. Farr et al. 2007
Courcier 2007 T. G. Farr/P. A. Rosen/E. Caro/R. Crippen/R. Duren/
A. Courcier, La métallurgie dans les pays du Caucase au S. Hensley/M. Kobrick/M. Paller/E. Rodriguez/L. Roth/
Chalcolithique et au début de l’âge u Bronze: bilan des D. Seal/S. Shaffer/J. Shimada/J. Umland/M. Werner/
études et perspectives nouvelles. In: B. Lyonnet (ed.), M. Oskin/D. Burbank/D. Alsdorf, The Shuttle Radar To-
Les cultures du Caucase (VIe –IIIe millénaires avant notre pography Mission. Reviews of Geophysics 45, 2007,
ère). Leurs relations avec le Proche Orient (Paris 2007) doi:10.1029/2005RG000183.
199–232. Faßbinder 1994
Courcier et al. 2012 J. W. E. Faßbinder, Die magnetischen Eigenschaften und
A. Courcier/B. Lyonnet/F. Guliyev, Metallurgy during the die Genese ferrimagnetischer Minerale in Böden im Hin-
Middle Chalcolithic Period in the Southern Caucasus: In- blick auf die magnetische Prospektion archäologischer
sight through recent discoveries at Mentesh-Tepe, Azer- Bodendenkmäler (Buch am Erlbach 1994).
baijan. In: P. Jett/B. McCarthy/J. G. Douglas (eds.), Scien- Faßbinder 2009
tific Research in Ancient Asian Metallurgy. Proceedings J. W. E. Faßbinder, Geophysikalische Prospektionsme-
of the 5th Forbes Symposium at the Freer Gallery of Art thoden – Chancen für das archäologische Erbe. In: E. Em-
(London 2012) 205–224. merling (ed.), Toccare – Non Toccare. ICOMOS – Hefte des
Courcier et al. in prep. Deutschen Nationalkomitees 47 (München 2009) 10–32.
A. Courcier/A. Hauptmann/B. Jalilov/S. Merkel/M. Prange/ Faßbinder et al. 1990
M. Ragimowa, Archaeometallurgical research on chalco- J. W. E. Faßbinder/H. Stanjek/H. Vali, Occurrence of mag-
lithic artifacts from Mentesh-Tepe (5th mill. BCE): preli- netic bacteria in soil. Nature 343, 1990, 161–163.
minary results., in prep. Faßbinder/Stanjek 1993
Courty et al. 1989 J. W. E. Faßbinder/H. Stanjek, Occurrence of magnetic
M. A. Courty/P. Goldberg/R. Macphail, Soils and micro- bacteria in archaeological soil. Archaeologia Polona 31,
morphology in archaeology (Cambridge 1989). 1993, 117–128.
Faßbinder/Becker 2003 Gambaschidze et al. 2001

J. Faßbinder/H. Becker, Die Magnetometerprospektion in I. Gambaschidze/A. Hauptmann/R. Slotta/Ü. Yalçın, Geor-
Aşağı Pınar. In: N. Karul/Z. Eres/M. Özdogan/H. Parzinger gien. Schätze aus dem Land des Goldenen Vlies. Kata-
(eds.), Aşağı Pınar: Einführung, Forschungsgeschichte, log der Ausstellung des Deutschen Bergbau-Museums
Stratigraphie und Architektur. Studien im Thrakien-Mar- Bochum in Verbindung mit dem Zentrum für Archäolo-
mara-Raum 1, Archäologie in Eurasien 15 (Mainz 2003) gische Forschungen der Georgischen Akademie der Wis-
39–41. senschaften Tbilissi vom 28. Oktober 2001 bis 19. Mai
Faßbinder/Gorka 2009 2002. Veröffentlichungen aus dem Deutschen Bergbau-
J. W. E. Faßbinder/T. H. Gorka, Beneath the soil – ar- museum (Bochum 2001).
chaeological prospecting with a caesium magnetometer. Gamkrelidze 1982
In: M. Reindel/G. A. Wagner (eds.), New technologies for I. P. Gamkrelidze, Nappe structures of the Caucasus. In:
archaeology: multidisciplinary investigations in Palpa M. Mahel’ (ed.), Alpine structural elements: Carpathian-
and Nasca, Peru (Berlin 2009) 49–69. Balkan-Caucasus-Pamir orogene zone (Bratislava 1982)
Fazeli et al. 2005 95–114.
H. Fazeli/E. H. Wong/D. T. Potts, The Qazvin Plain revis- Gasanov 1996
ited: A reappraisal of the chronology of northwestern T. A. Gasanov, Geodynamics of ophiolithes in the struc-
Central Plateau, Iran, in the 6th to 4th mill. BC. Ancient ture of the Lesser Caucasus and Iran (Baku 1996).
Near Eastern Studies 42, 2005, 3–82. Gaude 2005
Fröhlich et al. 2003 A. Gaude, Socio-economic situation and land use con-
N. Fröhlich/M. Posselt/G. A. Wagner, Excavations in a flicts in the Ag-Göl National Park region, Azerbaijan
‘‘blinde mode’’. Magnetometer survey, excavations and (Greifswald 2005).
magnetic susceptibility measurements of a multiperiod Geel et al. 2003
cite at Bad Homburg, Germany. Archaeologia Polona 41, B. van Geel/J. Buurman/O. Brinkkemper/J. Schelvis/A.
2003, 167–169. Aptroot/G. B. A. van Reenen/T. Hakbijl, Environmental
Furholt/Müller 2011 reconstruction of a Roman Period settlement site in Uit-
M. Furholt/J. Müller, The earliest monuments in Europe – geest (The Netherlands), with special reference to co-
architecture and social structures (5000–3000 cal BC). prophilous fungi. Journal of Archaeological Science 30,
In: M. Furholt/F. Lüth/J. Müller (eds.), Megaliths and 2003, 873–883.
identities: early monuments and neolithic societies from Geel/Aptroot 2006
the Atlantic to the Baltic. 3rd European Megalithic Stud- B. van Geel/A. Aptroot, Fossil ascomycetes in Quaternary
ies Group Meeting, 13th –15th of May 2010 at Kiel Uni- deposits. Nova Hedwigia 82, 2006, No. 3–4, 313–329.
versity. Institut für Ur- und Frühgeschichte der CAU Kiel Geörg 2011
and Römisch-Germanische Kommission Frankfurt a.M. C. Geörg, Paläopopulationsgenetik von Schwein und
Frühe Monumentalität und soziale Differenzierung 1 Schaf in Südosteuropa und Transkaukasien. Unpubl. PhD
(Bonn 2011) 15–32. dissertation, University of Mainz (Mainz 2011).
Gabrielian/Fragman-Sapir 2008 Glonti et al. 2008
E. Gabrielian/O. Fragman-Sapir, Flowers of the Transcau- L. Glonti/M. Ketshoveli/G. Palumbi, The cemetery at
casus and adjacent areas. Including Armenia, Eastern Kvatshelebi. In: A. G. Sagona/M. Abramishvili (eds.), Ar-
Turkey, Southern Georgia, Azerbaijan and Northern Iran chaeology in Southern Caucasus: perspectives from
(Ruggell 2008). Georgia. Ancient Near Eastern Studies, supplement 19
Gabrielian/Zohary 2004 (Leuven 2008) 153–184.
E. Gabrielian/D. Zohary, Wild relatives of food crops na- Gratuze 1999
tive to Armenia and Nakhichevan. Flora Mediterranea B. Gratuze, Obsidian characterization by laser ablation
14, 2004, 5–80. ICP-MS and its application to prehistoric trade in the
Gaffney et al. 2000 Mediterranean and the Near East: sources and distribu-
C. Gaffney/J. A. Gater/P. Linford/V. Gaffney/R. White, tion of obsidian within the Aegean and Anatolia. Journal
Large-scale systematic fluxgate gradiometry at the Ro- of Archaeological Science 26, 1999, 869–881.
man city of Wroxeter. Archaeological prospection 7, Greaves 2007
2000, 81–100. A. Greaves, Trans-Anatolia: Examining Turkey as a bridge
Galiatsatos et al. 2009 between East and West. Anatolian Studies 57. Transa-
N. Galiatsatos/T. J. Wilkinson/D. N. M. Donoghue/G. Philip, natolia. Proceedings of the conference held at the Brit-
The Fragile Crescent Project (FCP): Analysis of settle- ish Museum, 31 march to 1 april 2006, 2007, 1–16.
ment landscapes using satellite imagery. In: Proceed- Guliev et al. 2009
ings of CAA 2009: Making history interactive (Williams- F. Guliev/F. Guseynov/H. Almamedov, Excavations of a
burg 2009) 1–15. Neolithic settlement at Goytepe (Azerbaijan). In: Eura-
Gallis 1996 sien-Abteilung des Deutschen Archäologischen Instituts
K. Gallis, Ausgrabungen der neolithischen Siedlungen (ed.), Azerbaijan – Land between East and West. Trans-
und Friedhöfe von Platia Magoula Zarkou, Souphli- fer of knowledge and technology during the ‘‘First Glo-
Magula und Makrychori 2 in Thessalien. In: E. Alram- balization’’ of the VII.–IV. mill. BC. International Sympo-
Stern (ed.), Die Ägäische Frühzeit, 2. Serie, Forschungs- sium Baku, April 1–3, 2009 (Berlin 2009) 26–30.
bericht 1975–1993. Das Neolithikum in Griechenland Guliyev/Nishiaki 2012
mit Ausnahme von Kreta und Zypern (Wien 1996) 521– F. Guliyev/Y. Nishiaki, Ecavations at the Neolithic settle-
562. ment of Göytepe, the Middle Kura Valley, Azerbaijan,
2008–2009. In: R. Matthews/J. Curtis (eds.), Proceedings Henrickson/McDonald 1983

of the 7th ICAANE, London 2010 (Wiesbaden 2012) 71–84. E. F. Henrickson/M. McDonald, Ceramic form and func-
Hamlin 1975 tion: an ethnographic search and an archaeological ap-
C. Hamlin, Dalma Tepe. Iran 13, 1975, 111–128. plication. American Anthropologist 85, 1983, 630–643.
Hamon 2007 Hjulström 1935
C. Hamon, Modes de subsistance et activités dans le F. Hjulström, Studies of the morphological activity of riv-
Chalcolithique du Caucase Nord: étude fonctionelle des ers as illustrated by the River Fyris. Inaugural disserta-
outils en pierre de la culture de Maikop. In: B. Lyonnet tion (Uppsala 1935).
(ed.), Les cultures du Caucase (VIe –IIIe millénaires avant Hoogendoorn et al. 2005
notre ère). Leurs relations avec le Proche Orient (Paris R. M. Hoogendoorn/J. F. Boels/S. B. Kroonenberg/M. D.
2007) 189–198. Simmons/E. Aliyeva/A. D. Babazadeh/D. Huseynov, De-
Hamon 2008a velopment of the Kura delta, Azerbaijan. A record of Ho-
C. Hamon, From Neolithic to Chalcolithic in the Southern locene Caspian sea-level changes. Marine Geology 222,
Caucasus: economy and macrolithic implements from 2005, 359–380.
Shulaveri-Shomu sites of Kwemo-Kartli (Georgia). Pa- Hovsepyan/Willcox 2008
léorient 34, 2008, No. 2, 85–135. R. Hovsepyan/G. Willcox, The earliest finds of cultivated
Hamon 2008b plants in Armenia: evidence from charred remains and
C. Hamon, Functional analysis of stone grinding and pol- crop processing residues in pisé from the Neolithic set-
ishing tools from the earliest Neolithic of north-western tlements of Aratashen and Aknashen. Vegetation History
Europe, Journal of Archaeological Science 35, 1502–1520. and Archaeobotany 17, 2008, 63–71.
Hamon 2009 Jeunesse 2011
C. Hamon, Macrolithic tools of Georgian sites of the C. Jeunesse, Enceintes à fossé discontinu et enceintes à
Shulaveri-Shomu culture (VIIIth –Vth mill. B.C.), Paléorient pseudo-fossé dans le Néolithique d’Europe centrale et
34, 2008, No. 2, 85–135. occidentale. In: A. Denaire/C. Jeunesse/P. Lefranc (eds.),
Hansen 2011 Nécropoles et enceintes danubiennes du Ve millénaire
S. Hansen, Innovation Metall. Kupfer, Gold und Silber in dans le Nord-Est de la France et le Sud-Ouest de l’Alle-
Südosteuropa während des fünften und vierten Jahrtau- magne (Strasbourg 2011) 31–72.
sends v. Chr. Das Altertum 56, 2011, 275–314. Kazancı et al. 2004
Hansen et al. 2006 N. Kazancı/T. Gulbabazadeh/S. A. G. Leroy/O. Ileri, Sedi-
S. Hansen/G. Mirtskhulava/K. Bastert-Lamprichs/N. Bene- mentary and environmental characteristics of the Gilan-
cke/I. Gatsov/P. Nedelcheva, Aruchlo 2005–2006. Be- Mazanderan Plain, northern Iran: Influence of long- and
richt über die Ausgrabungen in einem neolithischen Sied- short-term Caspian water level fluctuations on geomor-
lungshügel. Archäologische Mitteilungen aus Iran und phology. Journal of Marine Systems 46, 2004, No. 1–4,
Turan 38, 2006, 1–34. 145–168.
Hansen et al. 2007 Kennedy 1998
S. Hansen/G. Mirtskhulava/K. Bastert-Lamprichs/J. Görs- D. Kennedy, Declassified satellite photographs and ar-
dorf/D. Neumann/M. Ullrich/I. Gatsov/P. Nedelcheva, chaeology in the Middle East: case studies from Turkey.
Aruchlo 2007. Bericht über die Ausgrabungen im neo- Antiquity 72, 1998, 553–561.
lithischen Siedlungshügel. Archäologische Mitteilungen Kiguradze 1986
aus Iran und Turan 39, 2007, 1–30. T. Kiguradze, Neolithische Siedlungen von Kvemo-Kartli,
Hansen et al. in press Georgien (München 1986).
S. Hansen/G. Mirtskhulava/K. Bastert-Lamprichs, Neo- Kiguradze 2000
lithic settlements of the 6th mill. BC in the southern Cau- T. Kiguradze, The Chalcolithic – Early Bronze Age tran-
casus. In: O. Niewenhuyse/A. Russel/R. Bernbeck/P. M. sition in the Eastern Caucasus. In: C. Marro/H. Haupt-
M. G. Akkermans (eds.), Interpreting Late Neolithic in mann (eds.), Chronologies des Pays du Caucase et de
Upper Mesopotamia (in press). L’Euphrate aux IVe –IIIe Millénaires. Varia Anatolica 11
Helwing 2009 (Paris 2000) 321–328.
B. Helwing, Azerbaijan in the Chalcolithic: A view from Kiguradze/Sagona 2003
the Southwest. In: Eurasien-Abteilung des Deutschen T. Kiguradze/A. Sagona, On the origins of the Kura-Ara-
Archäologischen Instituts (ed.), Azerbaijan – Land be- xes cultural complex. In: A. T. Smith/K. S. Rubinson
tween East and West. Transfer of knowledge and tech- (eds.), Archaeology in the borderlands: investigations in
nology during the ‘‘First Globalization’’ of the VII.–IV. Caucasia and beyond. The Cotsen Institute of Archaeol-
mill. BC. International Symposium Baku, April 1–3, 2009 ogy, University of California, Los Angeles Monograph 47
(Berlin 2009) 62–66. (Los Angeles 2003) 38–94.
Helwing et al. 2012 Kottek et al. 2006
B. Helwing/T. Aliyev/A. Ricci, Mounds and Settlements in M. Kottek/J. Grieser/C. Beck/B. Rudolf/F. Rubel, World
the Lower Qarabakh – Mil Plain, Azerbaijan. In: R. Hof- Map of the Köppen-Geiger climate classification updated.
mann/F.-K. Moetz/J. Müller (eds.), Tells: Social and Envi- Meteorologische Zeitschrift 15, 2006, No. 3, 259–263.
ronmental Space. Proceedings of the International Kroonenberg et al. 1997
Workshop ‘‘Socio-Environmental Dynamics over the Last S. B. Kroonenberg/G. V. Rusakov/A. A. Svitoch, The wan-
12,000 Years: The Creation of Landscapes II (14th –18th dering of the Volga delta: a response to rapid Caspian sea-
March 2011)’’ in Kiel. Universitätsforschungen zur prä- level change. Sedimentary Geology 107, 1997, nos. 3–4,
historischen Archäologie 207 (Bonn 2012) 67–77. 189–209.
Lancelotti 2010 (eds.), Proceedings of the 7th ICAANE, London 2010.

C. Lancelotti, Fuelling Harappan hearths: human-envi- (Wiesbaden 2012) 85–98.
ronment relations as revealed by fuel exploitation and Makharadze 2007
use. Unpubl. PhD dissertation (Cambridge 2010). Z. Makharadze, Nouvelles données sur le Chalcolithique
Landmann 1996 en Géorgie orientale. In: B. Lyonnet (ed.), Les cultures
G. Landmann, Dating Late Glacial abrupt climate changes du Caucase (VIe –IIIe millénaires avant notre ère). Leurs
in the 14,570 yr long continuous varve record of Lake relations avec le Proche Orient (Paris 2007) 123–132.
Van, Turkey. Paleogeography, Paleoclimatology, Paleoe- Mamedov 1997
cology 122, 1996, 107–118. A. Mamedov, The Late Pleistocene-Holocene history of
Le Borgne 1955 the Caspian Sea. Quaternary International 41–42, 1997,
E. Le Borgne, Susceptibilité magnétique anormale de sol 161–166.
superficiel. Annals of Geophysics 11, 1955, 399–419. Marro 2007
Le Borgne 1960 C. Marro, Upper-Mesopotamia and Transcaucasia in the
E. Le Borgne, Influence du feu sur les propriétés magné- Late Chalcolithic period (4000–3500 BC). In: B. Lyonnet
tiques du sol et sur celles du schiste et du granite. An- (ed.), Les cultures du Caucase (VIe –IIIe millénaires avant
nals of Geophysics 16, 1960, 159–195. notre ère). Leurs relations avec le Proche Orient (Paris
Lyonnet 2007a 2007) 77–94.
B. Lyonnet (ed.), Les cultures du Caucase (VIe –IIIe mil- Marro 2010
lénaires avant notre ère). Leurs relations avec le Proche C. Marro, Where did late Chalcolithic chaff-faced ware
Orient (Paris 2007). originate? Cultural dynamics in Anatolia and Transcauca-
Lyonnet 2007b sia at the dawn of urban civilization (ca 4500–3500 BC).
B. Lyonnet, Introduction. In: B. Lyonnet (ed.), Les cul- Paléorient 36, 2010, No. 2, 35–55.
tures du Caucase (VIe –IIIe millénaires avant notre ère). Marro et al. 2011
Leurs relations avec le Proche Orient (Paris 2007) 11–20. C. Marro/V. Bakhshaliyev/S. Ashurov, Excavations at Ov-
Lyonnet 2007c cular Tepesi (Nakhchivan, Azerbaijan). Second prelimin-
B. Lyonnet, La culture du Maikop, la Transcaucasie, ary report: the 2009–2010 seasons. Anatolica Antiqua 19,
l’Anatolie orientale et le Proche-Orient: relations et 2011, 53–100.
chronologie. In: B. Lyonnet (ed.), Les cultures du Cau- Matsutani 1991
case (VIe –IIIe millénaires avant notre ère). Leurs rela- T. Matsutani, Tell Kashkashok. The excavations at Tell
tions avec le Proche Orient (Paris 2007) 133–162. n. II Supplement to the Bulletin of the Institute of Orien-
Lyonnet 2009a tal Culture (Tokio 1991).
B. Lyonnet, Périphérie de la Mésopotamie à la période R. Matthews 2003 a
d’Uruk (4e millénaire): le cas des régions du Caucase. R. Matthews, Traces of early complexity. Late fifth to
In: J.-M. Durand/A. Jacquet (eds.), Centre et Périphérie: early fourth-millennia investigations: the early Northern
approches nouvelles des Orientalistes, Actes du collo- Uruk. In: R. Matthews (ed.), Excavations at Tell Brak 4:
que organisé par l’Institut du Proche-Orient Ancien du Exploring an Upper Mesopotamian regional centre,
Collège de France, la Société Asiatique et la CNRS (UMR 1994–1996. McDonald Institute Monographs (Oxford
7192) les 31 mai et 1er jiun 2006 (Paris, Collège de 2003) 25–52.
France). Cahiers de l’Institut du Proche-Orient Ancien du R. Matthews 2003 b
Collège de France (Paris 2009) 1–28. R. Matthews, Surface investigations. In: R. Matthews
Lyonnet 2009b (ed.), Excavations at Tell Brak. Vol. 4: Exploring an Upper
B. Lyonnet, Surveys and excavations in Western Azer- Mesopotamian regional centre, 1994–1996. McDonald
baijan: settlement changes and relations with surround- Institute Monographs (Oxford 2003) 7–24.
ing areas, from the Neolithic to the Bronze Age. In: Eu- R. Matthews et al. 2010
rasien-Abteilung des Deutschen Archäologischen Institut R. Matthews/Y. Mohammadifar/W. Matthews/A. Motarjem,
(ed.), Azerbaijan – Land between East and West. Trans- Investigating the Early Neolithic of western Iran: the
fer of knowledge and technology during the ‘‘First Glo- Central Zagros Archaeological Project (CZAP). Antiquity
balization’’ of the VII.–IV. mill. BC. International Sympo- 84, 2010, No. 323.
sium Baku, April 1–3, 2009 (Berlin 2009) 41–47. W. Matthews 2001
Lyonnet et al. 2008 W. Matthews, Micromorphological analysis of occupa-
B. Lyonnet/T. Akhundov/K. Almamedov/L. Bouquet/ tional sequences. In: R. J. Matthews/N. Postgate (eds.),
A. Courcier/B. Jellilov/F. Huseynov/S. Loute/Z. Makharadze/ Contextual analysis of the use of space at two Near
S. Reynard, Late Chalcolithic kurgans in Transcaucasia. Eastern Bronze Age sites [data-set] (2001 York).
The cemetery of Soyuq Bulaq (Azerbaijan). Archäologi- W. Matthews 2005
sche Mitteilungen aus Iran und Turan 40, 2008, 27–44. W. Matthews, Micromorphological and microstratigra-
Lyonnet/Guliyev 2010 phic traces of uses and concepts of space. In: I. Hodder
B. Lyonnet/F. Guliyev, Recent discoveries on the Neo- (ed.), Inhabiting Çatalhöyük. Reports from the 1995–
lithic and Chalcolithic of Western Azerbaijan. TÜBA-AR. 1999 seasons. Çatalhöyük Project 4. Monograph of the
Turkish Academy of Sciences Journal of Archaeology 13, British Institute of Archaeology at Ankara 38 (Cambridge
2010, 219–228. 2005) 355–398.
Lyonnet/Guliyev 2012 W. Matthews 2010
B. Lyonnet/F. Guliyev, Recent research on the Chalcolithic W. Matthews, Geoarchaeology and taphonomy of plant
Period in Western Azerbaijan. In: R. Matthews/J. Curtis remains and microarchaeological residues in early urban
environments in the Ancient Near East. Quaternary Inter- Muhs et al. 2004
national 214, 2010, No. 1–2, 98–113. D. Muhs/J. McGeehin/J. Beann/E. Fisher, Holocene loess
W. Matthews/Postgate 1994 deposition and soil formation as competing processes,
W. Matthews/J. N. Postgate, The imprint of living in an Matanuska Valley, southern Alaska. Quaternary Research
early Mesopotamian city: questions and answers. In: 61, 2004, 265–276.
R. Luff/P. Rowley-Conway (eds.), Whither enviromental Nebieridze 2010
archaeology? Papers from the Association of Environ- L. Nebieridze, The Tsopi chalcolithic culture. Studies of
mental Archaeology Conference, Selwyn College, Cam- the Society of Assyriologists, Bibliologists and Cauca-
bridge, 1991. Oxbow Monograph 38 (Oxford 1994) 171– siologists 6 (Tbilisi 2010).
212. Nechitajlo 2007
W. Matthews et al. 1996 A. Nechitajlo, La céramique de Ust’-Dzheguta, établisse-
W. Matthews/C. French/T. Lawrence/D. Cutler, Multiple ment du début de la culture de Maikop en Karachaevo-
surfaces: the micromorphology. In: I. Hodder (ed.), On Tcherkessie. In: B. Lyonnet (ed.), Les cultures du Cau-
the Surface. Çatalhöyük 1993–1995. Monograph of the case (VIe –IIIe millénaires avant notre ère). Leurs rela-
Çatalhöyük Project 1, McDonald Institute for Archaeolo- tions avec le Proche Orient. Éditions Recherche sur les
gical Research (Cambridge 1996) 301–342. Civilisations (Paris 2007) 163–178.
W. Matthews et al. 1997 Neill 1967
W. Matthews/C. A. I. French/T. Lawrence/D. F. Cutler/M. K. C. R. Neill, Mean-velocity criterion for scour of coarse
Jones, Microstratigraphic traces of site formation pro- uniform bed-material. 12th Proceedings of the Interna-
cesses and human activities. World Archaeology 29, tional Association of Hydrological Research, Fort Collins
1997, No. 2, 281–308. (Colorado 1967) 46–54.
Maynard 2011 Neill 1968
D. Maynard, The BTC Pipeline archaeological excava- C. R. Neill, A re-examination of the beginning of move-
tions in Azerbaijan. http://archaeologydataservice.ac.uk/ ment for coarse granular bed materials. Report of the
archives/view/btcpipeline_leap_2011/ Wallingford Hydraulics Research Station 68, 1968, 1–
Meliksetian et al. 2009 37.
K. Meliksetian/E. Pernicka/R. Badalyan, Composition and Neubauer et al. 1999
some considerations on the provenance of Armenian W. Neubauer/A. Eder-Hinterleitner/S. S. Seren/M. Doneus/
Early Bronze Age copper artifacts. In: Archaeometallurgy P. Milichar, Kombination archäologisch-geophysikali-
in Europe 2007: 2nd International Conference, Grado scher Prospektionsmethoden am Beispiel der römischen
and Aquileia, 17–21 June 2007 (Milano 2009) 41–58. Zivilstadt Carnuntum. Archaeologica Austrica 83, 1999,
Miall 1988 1–26.
A. D. Miall, Reservoir heterogenities in fluvial sand- Newson 2002
stones – Lessons learned from outcrop studies. Ameri- M. D. Newson, Geomorphological concepts and tools for
can Association of Petroleum Geologists Bulletin 72, sustainable river ecosystem management. Aquatic Con-
1988, No. 6, 682–697. servation: Marine and Freshwater Ecosystems 12, 2002,
Le Mière 2000 No. 4, 365–379.
M. Le Mière, L’occupation proto-Hassuna du Haut-Kha- Nieuwenhuyse 2000
bur occidental d’après la céramique. In: B. Lyonnet O. Nieuwenhuyse, Halaf settlement in the Khabur
(ed.), Prospection archéologique. Haut-Khabur occiden- headwaters. In: B. Lyonnet (ed.), Prospection archéolo-
tal (Syrie du N.E.) 1. Bibliothèque Archéologique et His- gique. Haut-Khabur occidental (Syrie du N. E.) 1. Bib-
torique 155 (Beyrouth 2000) 127–149. liothèque Archéologique et Historique (Beyrouth 2000)
Mikhailov et al. 2003 151–260.
V. N. Mikhailov/V. I. Kravtsova/D. V. Magritskii, Hydrolo- Nieuwenhuyse 2008
gical and morphological processes in the Kura River Del- O. P. Nieuwenhuyse, Feasting in the steppe – Late Neo-
ta. Water ressources – Vodnye Resursy 30, 2003, No. 5, lithic ceramic change and the rise of the Halaf. In: J. M.
495–508. Córdoba/S. Martı́nez Lillo/M. Molist/M. C. Pérez Dı́e/I. Ru-
Miller 2003 bio (eds.), Proceedings of the 5th International Congress
M. Miller, Technical aspects of ornament production in on the Archaeology of the Ancient Near East (Madrid
Sitagroi. In: E. S. Elster/C. Renfrew (eds.), Prehistoric 2008) 692–708.
Sitagroi: Excavations in Northeast Greece, 1968–1970. Nieuwenhuyse et al. 2010
Final Report 2. Monumenta Archaeologica 20 (Los An- O. P. Nieuwenhuyse/P. M. M. G. Akkermans/J. van der
geles 2003) 369–382. Plicht, Not so coarse, nor always plain – the earliest
Mills 1979 pottery of Syria. Antiquity 84, 2010, 71–85.
H. H. Mills, Downstream rounding of pebbles: a quanti- Ollivier et al. in press
tative review. Journal of Sedimentary Research 49, 1979, V. Ollivier/M. Fontugne/B. Lyonnet, Caspian Sea mobility
No. 1, 295–302. during the Late Quaternary: impact on the upstream hy-
Mosar et al. 2010 drosystem and the Neolithic – Chalcolithic settlements
J. Mosar/T. Kangarli/M. Bochud/U. Glasmacher/A. Rast/ of the Middle Kura Valley (Azerbaijan). Quarternary
M.-F. Brunet/M. Sosson, Cenozoic-recent tectonics and Science Review, in press.
uplift in the Greater Caucasus: a perspective from Azer- Özdoǧan 2007
baijan. Geological Society special publication 2010, M. Özdoǧan, Amidst Mesopotamia-centric and Euro-cen-
No. 340, 261–280. tric approaches: The changing role of the Anatolian Pe-
ninsula between the East and the West. Anatolian Stud- Schleifer et al. 2003
ies 57. Transanatolia. Proceedings of the conference N. Schleifer/J. W. E. Fassbinder/W. E. Irlinger/H. Stanjek,
held at the British Museum, 31 march to 1 april 2006, Investigation of an Eneolithic Chamer-group ditch sys-
2007, 17–24. tem near Riekofen (Bavaria) with archaeological, geo-
Palumbi 2008 physical and pedological methods. In: Soils and ar-
G. Palumbi, The Red and the Black. Social and cultural chaeology. Papers of the 1st International Conference
interactions between the Upper Euphrates and Southern on Soils and Archaeology, Szazhalombatta, Hungary,
Caucasus communities in the fourth and third mill BC. Stu- 30 May–3 June 2001, vol. 1163. British Archaeological
di di Preistoria Orientale 2 (Roma 2008). Report International Series (Archaeopress 2003) 59–
Pernicka/Anthony 2010 64.
E. Pernicka/D. Anthony, The invention of copper metal- Schmidt 2001
lurgy and the Copper age of Old Europe. In: D. Anthony/ A. Schmidt, Geophysical data in archaeology. A guide in
J. Y. Chi (eds.), The lost world of Old Europe. The Da- good practice (York 2001).
nube valley 5000–3500 BC (New York, Princeton, Ox- Schumm 1977
ford 2010) 163–177. S. A. Schumm, The fluvial system (New York 1977).
Perrot 1968 Schütt et al. 1997
J. Perrot, La préhistoire palestinienne. In: Supplément B. Schütt/J. Berking/M. Frechen/P. Frenzel/A. Schwalb/
au Dictionnaire de la Bible 8 (Paris 1968) 423–440. C. Wrozyna, Late Quaternary transition from lacustrine
Piperno 2006 to a fluvio-lacustrine environment in the north-western
D. R. Piperno, Phytoliths: a comprehensive guide for Nam Co, Tibetan Plateau, China. Quaternary Interna-
archaeologists and paleoecologists (Lanham, New York, tional 218, 1997, No. 1–2, 104–117.
Toronto, Oxford 2006). Scollar et al. 1990
Posamentier 2001 I. Scollar/A. Tabbagh/A. Hesse/I. Herzog, Archaeological
H. W. Posamentier, Lowstand alluvial bypass systems: prospecting and remote sensing. Topics in remote sen-
Incised vs. unincised. American Association of Petrol sing (Cambridge 1990).
Geologists Bulletin 85, 2001, No. 10, 1771. Selimkhanov 1978
Reimer P. J. et al. 2009 I. R. Selimkhanov, Ancient tin objects of the Caucasus
P. J. Reimer/M. G. L. Baillie/F. G. McCormac/R. W. Reimer/ and the results of their analysis. In: A. D. Franklin/J. S.
E. Bard/A. Bayliss/J. W. Beck/P. G. Blackwell/C. E. Buck/ Olin/T. A. Wertime (eds.), The search for ancient tin
T. J. Heaton/C. B. Ramsey/G. S. Burr/R. L. Edwards/M. Fried- (Washington 1978) 53–58.
rich/B. Kromer/P. M. Grootes/T. P. Guilderson/I. Hajdas/ Shahack-Gross et al. 2003
A. G. Hogg/K. A. Hughen/K. F. Kaiser/S. W. Manning/D. A. R. Shahack-Gross/F. Marshall/S. Weiner, Geo-ethnoarch-
Richards/J. R. Southon/S. Talamo/C. S. M. Turney/J. van aeology of pastoral sites: The identification of livestock
der Plicht/C. E. Weyhenmeyer, IntCal09 and Marine09 enclosures in abandoned Massai settlements. Journal of
radiocarbon age calibration curves, 0–50 000 years CAL Archaeological Science 30, 2003, No. 4, 439–459.
BP. Radiocarbon 51, 2009, No. 4, 1111–1150. Shahack-Gross et al. 2005
Reinhold 2007 R. Shahack-Gross/R.-M. Albert/A. Gilboa/O. Nagar-Hil-
S. Reinhold, Die Spätbronze- und frühe Eisenzeit im man/I. Sharon/S. Weiner, Geoarchaeology in an urban
Kaukasus: materielle Kultur, Chronologie und überregio- context: The uses of space in a Phoenician monumental
nalle Beziehungen. Universitätsforschungen zur prähis- building at Tel Dor (Israel). Journal of Archaeological
torischen Archäologie 144 (Bonn 2007). Science 32, 2005, No. 9, 1417–1431.
Ryan 2011 Shillito 2011
P. Ryan, Plants as material culture in the Near Eastern L. Shillito, Simultaneous thin section and phytolith ob-
Neolithic: Perspectives from the silica skeleton artifac- servations of finely stratified deposits from Neolithic
tual remains at Çatalhöyük. Journal of Anthropological Çatalhöyük, Turkey: implications for paleoeconomy and
Archaeology 30, 2011, No. 3, 292–305. Early Holocene paleoenvironment. Journal of Quaternary
Rychagov 1997 Science 26, 2011, No. 6, 576–588.
G. Rychagov, Holocene oscillations of the Caspian Sea, Shillito et al. 2011 a
and forecasts based on palaeogeographical reconstruc- L.-M. Shillito/W. Matthews/M. J. Almond/I. D. Bull, The
tions. Quaternary International 41–42, 1997, 167–172. microstratigraphy of middens: capturing daily routine in
Rye 1981 rubbish at Neolithic Catalhoyuk, Turkey. Antiquity 85,
O. S. Rye, Pottery technology: principles and reconstruc- 2011, No. 329, 1024–1038.
tion. Manuals on archeology 4 (Washington 1981). Shillito et al. 2011 b
Scheu 2011 L.-M. Shillito/I. D. Bull/W. Matthews/M. J. Almond/J. M.
A. Scheu, Palaeogenetische Studien zur Populations- Williams/R. P. Evershed, Biomolecular and micromorpho-
geschichte von Rind und Ziege mit einem Schwerpunkt logical analysis of suspected faecal deposits at Neolithic
auf dem Neolithikum in Südosteuropa. Unpubl. PhD dis- Çatalhöyük, Turkey. Journal of Archaeological Science
sertation (Mainz 2011). 38, 2011, No. 8, 1869–1877.
Schleifer 2004 Stöllner et al. 2010
Schleifer N., Ghost features: a proposal for appropriate T. Stöllner/I. Gambaschidze/A. Hauptmann/G. Mindiash-
management and a forum for discussion. Newsletter of vili/G. Gogocuri/G. Steffens, Goldbergbau in Südgeorgien.
the International Society of Archaeological Prospection Neue Forschungen zum frühbronzezeitlichen Bergbau in
1, 2004, 6–9. Georgien. In: S. Hansen/A. Hauptmann/I. Motzenbäcker/
E. Pernicka (eds.), Von Majkop bis Trialeti: Gewinnung Cave, Israel: Diagenetic Processes and Archaeological
und Verbreitung von Metallen und Obsidian in Kauka- Implications. Journal of Archaeological Science 29, 2002,
sien im 4.–2. Jt. V. Chr.: Beiträge des Internationalen No. 11, 1289–1308.
Symposiums in Berlin vom 1.–3. Juni 2006. Kolloquien Weiner 2010
zur Vor- und Frühgeschichte 13 (Bonn 2010) 139–160. S. Weiner, Microarchaeology (Cambridge, New York 2010).
Stordeur 1988 Wilkinson 2003
D. Stordeur, Outils et armes en os du gisement natou- T. J. Wilkinson, Archaeological landscapes of the Near
fien de Mallaha (Eynan) Israel. Mémoires et travaux du East (Tuscon, AR 2003).
Centre de Recherche Français de Jérusalem 6 (Paris 1988). Wilkinson et al. 2012
Stordeur 1999 K. N. Wilkinson/B. Gasparian/R. Pinhasi/P. Avetisyan/
D. Stordeur, Néolithisation et outillage osseux, la révo- R. Hovsepyan/D. Zardaryan/G. E. Areshian/G. Bar-Oz/
lution a-t-elle eu lieu? In: M. Julien/A. Averbouh/D. Ram- A. Smith, Areni-1 Cave, Armenia: A Chalcolithic-Early
seyer (eds.), Préhistoire d’os. Recueil d’études sur l’in- Bronze Age settlement and ritual site in the southern
dustrie osseuse préhistorique offert à Henriette Camps- Caucasus. Journal of Field Archaeology 37, 2012, No. 1,
Fabrer (Aix-en-Provence 1999) 261–272. 20–33.
von Suchodoletz et al. 2011 Wu et al. 1973
H. von Suchodoletz/D. Faust/D. Wolf, Investigations of T. H. Wu/S. K. Vyas/N. Y. Chang, Probabilistic adjustment
the fluvial dynamics in Eastern Georgia as a contribution of seepage. Journal of soil mechanics and foundation
to paleoenviromental research of the region. Annals of division. Proceedings of the American Society of Civil
Agrarian Science 9, 2011, No. 3, 13–17. Engineers 99, 1973, No. 4, 323–340.
Taylor 1972 Yang/Song 1979
C. C. Taylor, The study of settlement patterns in pre-Saxon C. T. Yang/C. C. S. Song, Dynamic adjustment of allu-
Britain. In: P. Ucko/R. Tringham/G. W. Dimbleby (eds.), vial chanels in Rhodes D. D. In: G. P. Williams (ed.),
Man, settlement and urbanism (London 1972) 109–113. Adjustments of the fluvial system (London 1979) 55–
Tekin 2011 67.
H. Tekin, Karavelyan kazılarından insan ve yapı betimli Yener et al. 2000
bir Halaf kabının tanıtımı. TÜBA-AR. Turkish Academy of K. A. Yener/C. Edens/T. Harrison/J. Vestraete/T. J. Wilkin-
Sciences Journal of Archaeology 14, 2011, 309–312. son, The Amuq valley regional project, 1995–1998. Amer-
Thompson/Oldfield 1986 ican Journal of Archaeology 104, 2000, No. 2, 163–
R. Thompson/F. Oldfield, Environmental magnetism 220.
(London, Boston 1986). Zohary/Hopf 2003
Tobler 1950 D. Zohary/M. Hopf, Domestication of plants in the Old
A. J. Tobler, Excavations at Tepe Gawra II, Levels IX–XX World. The origin and spread of cultivated plants in
(Philadelphia 1950). West Asia, Europe and the Nile Valley (New York 2003).
Tucker et al. 2004
C. J. Tucker/D. M. Grant/J. D. Dykstra, NASA’s global THŁHÆ*Œ@- et al. 1961
orthorectified Landsat data set. Photogrammetric Engi- N. T. THŁHÆ*Œ@-/8. T. ˆ/łª/Œ/T. 6$º%/B@-, ˆ*@º@ªŁ'
neering & Remote Sensing 70, 2004, No. 3, 313–322. TH*(Æ/Ø+I/B/, ($+Bß* ?@º*HBß* ŁæŒ@?/*Cß* (`/Œ$).
Ur 2003 TÆŁÆ$ºº/*- 1982
J. Ur, CORONA satellite photography and ancient road ˛. T. TÆŁÆ$ºº/*-, 3B*@ºŁ% Ł Æ(@BH/ B/ %*((Ł%@(ŁŁ
networks: a Northern Mesopotamian case study. Anti- ˝/ıŁ\*-/BæŒ@Ø T66— (`/Œ$ 1982).
quity, 2003, No. 295, 102–115. TºŁ*-/˝/(ŁC/B@- 2001
Vandiver 1987 ˝. TºŁ*-/¨. ˝/(ŁC/B@-, ˚$º:%$(/ æ*-*(B@ª@ TH*(-
P. B. Vandiver, Sequential slab construction: a conser- Æ/Ø+I/B/ - 9?@ı$ ?@H+B*ª@ 9B*@ºŁ%/ (`/Œ$ 2001).
vative southwest Asiatic ceramic tradition, ca. 7000– TCŁ(ı/B@- 1987
3000 BC. Paléorient 13, 1987, No. 2, 9–35. R. T. TCŁ(ı/B@-, O@ıæŒ@* ?@æ*º*BŁ*. O*º@-*Œ Ł *ª@
Vavilov/Dorofeev 1992 Œ$º:%$(/ - C*H@ºŁ%* Ł B*@ºŁ%* ª@(B@ª@ ˜/ª*æ%/B/
N. I. Vavilov/V. F. Dorofeev, Origin and geography of (8@æŒ-/ 1987).
cultivated plants (Cambridge, New York 1992). T(/H@-/ 1986
van der Veen 2007 —. T(/H@-/, ˚/C*BBß* @($+Ł' %($+/ (/BBŁı H*Cº*-
M. van der Veen, Formation processes of desiccated +*º:\*æŒ@-æŒ@%@-@+\*æŒŁı ?º*C‚B H/?/+B@ª@ TH*(-
and carbonized plant remains – the identification of Æ/Ø+I/B/ (?@ +/BBßC 9Œæ?*(ŁC*B%/º:B@-%(/ææ@º@ªŁ-
routine practice. Journal of Archaeological Science 34, \*æŒŁı Łææº*+@-/BŁØ) (`/Œ$ 1986).
2007, No. 6, 968–990. `/Æ/-H/+* et al. 2005
Vitelli 1989 ´. 8. `/Æ/-H/+*/T. T. ˝/+Ł(@-/ˆ. 6. `/ªŁ(@-/6. T.
K. Vitelli, Were pots first made for food? Doubts from `*Œ%/łŁ/8. 8. 8/C*+@-/´. ˆ. —/C/H/B@-/N. ˜. 8$-
Franchthi. World Archaeology 21, 1989, No. 1, 17–29. æ/*-/S. T. Tı$B+@-, 8ŁB*(/º:B@-æß(:*-ß* (*æ$(æß
Voigt 1983 TH*(Æ/Ø+I/B/ ($æº@-Ł' #@(CŁ(@-/BŁ', H/Œ@B@C*(-
M. M. Voigt, Hajji Firuz Tepe, Iran: the Neolithic settlement, B@æ%Ł (/HC*ø*BŁ', B/$\Bß* @æB@-ß ?(@ªB@HŁ(@-/BŁ')
University Museum Monograph 50 (Philadelphia 1983). (`/Œ$ 2005).
Weiner et al. 2002 ´/(/H/ł-ŁºŁ 1992
S. Weiner/P. Goldberg/O. Bar-Yosef, Three-dimensional ´. ´/(/H/ł-ŁºŁ, —/BB*H*Cº*+*º:\*æŒ/' Œ$º:%$(/
Distribution of Minerals in the Sediments of Hayonim ¨@(@-Tº/H/BæŒ@ª@ Æ/ææ*ØB/ (WÆŁºŁæŁ 1992).
¨*ææ*B 1959 ˝/(ŁC/B@- 1987

T. T. ¨*ææ*B, W($+ß /H*(Æ/Ø+I/BæŒ@Ø (˛(*B-˚/ºŁBæ- ¨. ˝/(ŁC/B@-, ˚$º:%$(/ +(*-B*Øł*ª@ H*Cº*+*º:-
Œ@Ø) /(ı*@º@ªŁ\*æŒ@Ø 9Œæ?*+Ł!ŁŁ (8@æŒ-/, ¸*BŁB- \*æŒ@-æŒ@%@-@+\*æŒ@ª@ B/æ*º*BŁ' TH*(Æ/Ø+I/B/
ª(/+ 1959). (9?@ı/ 9B*@ºŁ%/ VI–IV %ßæ. +@ B.9.) (`/Œ$ 1987).
¨*ææ*B 1965 6*C‚B@-/˚@(@ÆŒ@-/ 1983
A. A. ¨*ææ*B, ¨H Łæ%@(Ł\*æŒ@ª@ ?(@łº@ª@ 8Łº:æŒ@- 6. T. 6*C‚B@-/ˆ. S. ˚@(@ÆŒ@-/, W*ıB@º@ªŁ' +(*-B*Ø-
˚/(/Æ/ıæŒ@Ø æ%*?Ł. In: A. A. ¨*ææ*B/˚. ˚$łB/(*-/ łŁı ?(@ŁH-@+æ%-: C*H@ºŁ%-9B*@ºŁ% (¸*BŁBª(/+ 1983).
(eds.), W($+ß /H*(Æ/Ø+I/BæŒ@Ø /(ı*@º@ªŁ\*æŒ@Ø 9Œæ- R/Bæ*B et al. 2009
?*+Ł!ŁŁ II (1956–1960). 8/%*(Ł/ºß Ł Łææº*+@-/BŁ' 6. R/Bæ*B/ˆ. 8Ł(!ı$º/-//˚. `/æ%*(%-¸/C?(ŁH, T($ıº@,
?@ /(ı*@º@ªŁŁ 666— 125 (8@æŒ-/, ¸*BŁBª(/+ 1965) ?@æ*º*BŁ* 9?@ıŁ B*@ºŁ%/ - ˆ($HŁŁ. In: ˆ*(C/BæŒŁØ
10–35. T(ı*@º@ªŁ\*æŒŁØ ŁBæ%Ł%$% (ed.), TH*(Æ/Ø+I/B –
˚@(*B*-æŒŁØ 1998 æ%(/B/, æ-'Hß-/7ø/' -@æ%@Œ Ł H/?/+. ˛ÆC*B HB/BŁ-
6. ˚@(*B*-æŒŁØ, ˇ@æ*º*BŁ* ‘‘˙/C@Œ’’ $ ª@(@+/ ˚Łæº@- 'CŁ Ł %*ıB@º@ªŁ'CŁ - ?*(Ł@+ «?*(-@Ø ªº@Æ/ºŁH/!ŁŁ»
-@+æŒ/ (BŁIBŁØ æº@Ø). In: T. `*ºŁBæŒŁØ (ed.), 8/- VII–IV %ßæ. +@ B. 9 (`*(ºŁB 2009) 80–82.
%*(Ł/ºß ?@ ŁH$\*BŁ7 Łæ%@(ŁŒ@-Œ$º:%$(B@ª@ B/æº*+Ł' O*(Bßł 1982
6*-*(B@ª@ ˚/-Œ/H/ 1. T(ı*@º@ªŁ' (6%/-(@?@º: 1998) ¯. O*(Bßł, 3B*@ºŁ% ˇ(/-@Æ*(*IB@Ø VŒ(/ŁBß Ł
96–147. 8@º+/-ŁŁ. In: ´. 8/ææ@B/˝. 8*(?*(% (eds.), 3B*@ºŁ%
˚@(@ÆŒ@-/ 1978 666—. T(ı*@º@ªŁ' 666— (8@æŒ-/ 1982) 166–320.
ˆ. ˚@(@ÆŒ@-/, ˜(*-B*ØłŁ* I/%-*BBß* @($+Ł' Ł Łı OŁŒ@-/BŁ 1999
?(@ŁH-@+Ł%*º:B@æ%: (- æ-*%* 9Œæ?*(ŁC*B%/º:B@-%(/- ˆ. OŁŒ@-/BŁ, ˇ@æ*º*BŁ* (/BB*H*Cº*+*º:\*æŒ@Ø Œ$º:-
æ@º@ªŁ\*æŒ@ª@ Łææº*+@-/BŁ'). 6@-*%æŒ/' /(ı*@º@ªŁ' %$(ß ŁH OŁB%Ł. In: W($+ß $BŁ-*(æŁ%*%/ «¸*ªŁ' Ł Œ@C-
4, 1978, 36–52. ?/BŁ'», vol. 2. (WÆŁºŁæŁ 1999) 72–84.
˚$łB/(*-/ 1993
˚. ˚$łB/(*-/, 2IBßØ ˚/-Œ/H - IR –II %ßæ. +@ B.9.:
3%/?ß Œ$º:%$(B@ª@ Ł æ@!Ł/º:B@-9Œ@B@CŁ\*æŒ@ª@ (/H-
-Ł%Ł' (6. ˇ*%*(Æ$(ª 1993).
˚$łB/(*-//¸ŁæŁ!ßB/ 1979
˚. ˚$łB/(*-//ˆ. ¸ŁæŁ!ßB/, ˛ -@HBŁŒB@-*BŁŁ ?/ł*B-
B@ª@ H*Cº*+*ºŁ' - ˙/Œ/-Œ/H:*: (Œ ?@æ%/B@-Œ* ?(@Æ-
º*Cß). 6@-*%æŒ/' /(ı*@º@ªŁ' vl. 2, 1979 5–18
8/ıC$+@- et al. 1987
T. ¨. 8/ıC$+@-/T. 6. ˇŁ(Ł*-/6. ˝. `/ªŁ(@-//W. W.
¨æC/Łº-H/+*/¨. 8. 8/º$C'B/W. T. T+Łº@-, ˝@-ß*
CŁB*(/ºß C*+B@-?@(#Ł(@-ßı ($+ ˚/+/Æ*ŒæŒ@ª@
(/Ø@B/ 8/º@ª@ ˚/-Œ/H/. ˜@Œº/+ß TŒ/+*CŁŁ ˝/$Œ
TH*(Æ/Ø+I/BæŒ@Ø 666— (`/Œ$ 1987).
8$B\/*- 1982
P. 8$B\/*-, 3B*@ºŁ% ˚/-Œ/H/. In: B. 8/ææ@B/˝. 8*(-
?*(% (eds.), T(ı*@º@ªŁ' 666—, 3B*@ºŁ% 666— (8@æ-
Œ-/ 1982) 93–164.
˝/(ŁC/B@- 1966
¨. ˝/(ŁC/B@-, ˜(*-B*Øł/' H*Cº*+*º:\*æŒ/' Œ$º:-
%$(/ ˙/Œ/-Œ/H:'. In: ˜@Œº/+ß Ł æ@@Æø*BŁ' /(ı*@º@-
ª@- 666— B/ VII 8*I+$B/(@+B@C Œ@Bª(*ææ* +@Łæ%@-
(ŁŒ@- Ł ?(@%@Łæ%@(ŁŒ@- - ˇ(/ª* (8@æŒ-/ 1966) 121–
126.
˝/(ŁC/B@- 1971
¨. ˝/(ŁC/B@-, ˛ H*Cº*+*ºŁŁ 9?@ıŁ 9B*@ºŁ%/ - TH*(-
Æ/Ø+I/B*. 6@-*%æŒ/' /(ı*@º@ªŁ' 3, 1971, 3–14.
Svend Hansen Bertille Lyonnet

Barbara Helwing UMR 7192/CNRS
Katrin Bastert-Lamprichs Collège de France
Norbert Benecke 52 rue du Cardinal Lemoine
Daniel Neumann 75005 Paris
Michael Ullrich France
Andrea Ricci
Maria Bianca D’Anna Michel Fontugne
Deutsches Archäologisches Institut CEA Laboratoire des Sciences du Climat
Eurasien-Abteilung et de l’Environnement
Im Dol 2–6 Avenue de la Terrasse
14195 Berlin 91198 Gif-sur-Yvette cedex
Germany France
Guram Mirtskhulava Ilia Heit

Georgische Akademie der Wissenschaften Universität Mainz
Uznadze Str. 14 Institut für Vor- und Frühgeschichte
GEO-0102 Tbilisi Schillerstrasse 11
Georgia Schönborner Hof – Südflügel
D-55116 Mainz
T e v e k k ü l A l i y e v Germany
Farhad Guliyev
Azerbaijan National Academy of Sciences Vincent Ollivier
Institute of Archaeology and Ethnography ANR Ancient Kura
31 H. Javid Prospekti 7 Traverse des Dominicaines
AZ-1143 Baku 13100 Aix-en-Provence
Azerbaijan France
J ö r g F a s s b i n d e r Bernard Gratuze
Julia Koch CNRS, UMR 5060, IRAMAT
Roland Linck Centre Ernest Babelon
Florian Becker 3D rue de la Férollerie
Bayerisches Landesamt für Denkmalpflege 45071 Orléans cedex
Abt. Denkmalerfassung und Denkmalforschung France
Referat ZII Siedlungs- und Kulturlandschaftsdokumentation
Hofgraben 4 G a ë l l e L e D o s s e u r
D-80539 München ArScAn, UMR 7041
Germany Ethnologie préhistorique
21 allée de l’Université
Joachim Wahl 92023 Nanterre cedex.
Regierungspräsidium Stuttgart France
Landesamt für Denkmalpflege
Baden-Württemberg Alexia Decaix
Arbeitsstelle Konstanz, Osteologie Université Paris 1
Stromeyersdorfstraße 3 Arscan UMR 7041/Muséum National d’Histoire Naturelle
78467 Konstanz UMR 7209
Germany case postale 56, 55 rue Buffon
75005 Paris
Susan Schorr France
Ainhoa Lincot
Freie Universität Berlin Antoine Courcier
Malteserstr. 74–100 ANR Ancient Kura
D-12249 Berlin 890 route de Bailleul
Germany 59270 Berthen
France
Wiebke Bebermeier
F r a n k S c h l ü t z Laurence Bouquet
Arian Goren Inrap (Institut National de Recherches en
Freie Universität Berlin Archéologie Préventive)
Malteserstr. 74–100 56 Bd de Courcerin, Espaces multiservices-Lot 34
D-12249 Berlin 77183 Croissy-Beaubourg
France
Lisa-Marie Shillito Pascal Roymand

The University of York G a ë l l e B r u l e y - C h a b o t
Dept. of Archaeology Inrap (Institut National de Recherches
The King’s Manor Archéologiques Préventives)
York 34–36 avenue Paul Vaillant-Couturier
Y01 7EP United Kingdom 93120 La Courneuve
France
Felix Geitel
Freie Universität Berlin Anaick Samzun
Institut für Vorderasiatische Archäologie Inrap (Institut National de Recherches
Hüttenweg 7 Archéologiques Préventives)
14195 Berlin direction interregionale Centre-Ile de France
Germany 31 rue Delizy
93698 Pantin Cedex
France
Abstract —$#%"$
The multidisciplinary international research project Q*º:7 C*I+Łæ!Ł?ºŁB/(B@ª@ C*I+$B/(@+B@ª@

‘‘Ancient Kura’’ aims at the diachronic reconstruc- B/$\B@ª@ ?(@*Œ%/ «˜(*-B'' ˚$(/» '-º'*%æ' (*-
tion of human-landscape interaction in the southern Œ@Bæ%($Œ!Ł' B/ +Ł/ı(@BB@Ø @æB@-* /B%(@?@ª*B-
Caucasus from the begin of sedentism to the Early B@ª@ -ºŁ'BŁ' B/ º/B+ł/#% 2IB@ª@ ˚/-Œ/H/ -
Bronze Age. It combines research in three areas ?*(Ł@+ æ B/\/º/ @æ*+º@æ%Ł ?@ 9?@ı$ (/BB*Ø
around the archaeological sites of Aruchlo in Geor- Æ(@BHß. ˇ(@*Œ% -Œº7\/*% - æ*Æ' #$B+/C*B%/º:-
gia, Mentesh Tepe in western and Kamiltepe in Bß* Łææº*+@-/BŁ' ?@ 9Œ@º@ªŁŁ, º/B+ł/#%B@Ø
southern Azerbaijan with comparative research con- /(ı*@º@ªŁŁ, ÆŁ@/(ı*@º@ªŁŁ, / %/ŒI* Ł ŁH$\*BŁ*
cerning environment, archaeological landscapes, C/%*(Ł/º:B@Ø Œ$º:%$(ß B/ /(ı*@º@ªŁ\*æŒŁı ?/-
bio-archaeology and material culture. The article C'%BŁŒ/ı T($ıº@ - ˆ($HŁŁ, 8*B%*ł W*?* - H/?/+-
presents an interim report of the whole project, B@C Ł ˚/CŁº:%*?* - -@æ%@\B@C TH*(Æ/Ø+I/B*. ´
with chapters by authors on their individual, specia- æ%/%:* ?(*+æ%/-º*B ?(*+-/(Ł%*º:BßØ @%\‚% @
lized studies. One of the main foci is on the Neo- ?(@-*+‚BB@Ø æ@-C*æ%B@Ø (/Æ@%*, Œ/I+/' ªº/-/
lithic period, which is represented in all three re- B/?Łæ/B/ æ@@%-*%æ%-$7øŁCŁ æ?*!Ł/ºŁæ%/CŁ B/
search areas. The Neolithic shows two distinct @æB@-* ?(@-*+‚BBßı ŁCŁ Łææº*+@-/BŁØ. ´ !*B%(*
regional traditions, known as S̆ulaveri-S̆omutepe in ?(@*Œ%/ B/ı@+Ł%æ' 9?@ı/ B*@ºŁ%/, ?(*+æ%/-º*B-
the northwestern and as Mil Steppe Painted Ware B/' B/ Œ/I+@C ŁH $?@C'B$%ßı ?/C'%BŁŒ@- Ł æ@æ-
in the southern area, which reflect different cultural %/-º'7ø/' +-/ (/HºŁ\Bßı - (*ªŁ@B/º:B@C ?º/B*
affiliations in these neighboring regions. The Chal- #*B@C*B/: «N$º/-*(Ł-N@C$%*?*» B/ æ*-*(@-
colithic is investigated in great detail at Mentesh H/?/+* Ł «(/æ?ŁæB/' Œ*(/CŁŒ/ 8Łº:æŒ@Ø æ%*?Ł»
Tepe and has also been touched upon in the Kamil- B/ 7ª@--@æ%@Œ*. —/HBŁ!/ C*I+$ 9%ŁCŁ Œ$º:%$(-
tepe region. The Early Bronze Age is represented BßCŁ #*B@C*B/CŁ ?(@æº*IŁ-/*%æ' %/ŒI* Ł - Łı
through burials and pits at Mentesh Tepe. The dia- @(Ł*B%/!ŁŁ B/ æ@?(*+*º:Bß* %*((Ł%@(ŁŁ. 3?@ı/
chronic and supra-regional comparative approach ı/ºŒ@ºŁ%/ +*%/º:B@ Łææº*+@-/B/ B/ 8*B%*ł
adopted by the project provides a new perspective W*?* Ł %@º:Œ@ \/æ%Ł\B@ H/%(@B$%/ - (/Ø@B* ?/-
concerning specific adaptations within variable ha- C'%BŁŒ/ ˚'CŁº:%*?*. 3?@ı/ (/BB*Ø Æ(@BHß
bitats reflected in the faunal and botanical record, ?(*+æ%/-º*B/ ('+@C H/ı@(@B*BŁØ - 8*B%*ł
ranging from temperate to arid climatic zones. Indi- W*?*. 6(/-BŁ%*º:BßØ /B/ºŁH B/ +Ł/ı(@BB@Ø Ł
vidual sites show differing degrees of integration B/+-(*ªŁ@B/º:B@Ø @æB@-* @%Œ(ß-/*% B@-ß* ?*(-
into regional raw material procurement systems that æ?*Œ%Ł-ß +º' ŁH$\*BŁ' #@(C /+/?%/!ŁŁ Œ
crosscut and overlie the cultural affiliations. ŁHC*B*BŁ'C æ(*+ß @ÆŁ%/BŁ' @% H/æ$łºŁ-@ª@ +@
$C*(*BB@ª@ ŒºŁC/%/, \%@ B/łº@ @%(/I*BŁ* -
Æ@%/BŁ\*æŒ@C Ł H@@º@ªŁ\*æŒ@C C/%*(Ł/º*. ´ %@
I* æ/C@* -(*C', ŁH$\*BŁ* Œ/I+@ª@ ?/C'%BŁŒ/ -
@%+*º:B@æ%Ł +/‚% -@HC@IB@æ%: ?(@æº*+Ł%: (/H-
ºŁ\B$7 æ%*?*B: ŁB%*ª(/!ŁŁ B/æ*º*BŁ' - (*ªŁ@-
B/º:B$7 æŁæ%*C$ +@Æß\Ł Ł (/H(/Æ@%ŒŁ æß(:',
\%@, - æ-@7 @\*(*+:, @Æ$æº/-ºŁ-/*% Œ$º:%$(B$7
?(ŁB/+º*IB@æ%: @Æø*æ%-/.

Ancient Kura 2010 2011 The First Two Sea

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Ancient Kura 2010 2011 The First Two Sea

Hochgeladen von

Copyright:

Verfügbare Formate

Deutsches Archäologisches Institut !

DIETRICH REIMER VERLAG GmbH BERLIN

Herausgeber: Svend Hansen und Mayke Wagner

Die Redaktion behält sich vor, Manuskripte zu kürzen.

! 2012 by Deutsches Archäologisches Institut, Eurasien-Abteilung

L y o n n e t, B., G u l i y e v, F., H e l w i n g, B., A l i y e v, T., H a n s e n, S. und M i r t s k h u l a v a, G.,

L y o n n e t, B., G u l i y e v, F., H e l w i n g, B., A l i y e v, T., H a n s e n, S. and M i r t s k h u l a v a, G.,

A Ancient Kura 2010-2011:

ARCHÄOLOGISCHE MITTEILUNGEN AUS IRAN UND TURAN

with contributions by L. Astruc, K. Bastert-Lamprichs, W. Bebermeier, F. Becker, N. Benecke, L. Bouquet,

Schlagwörter: Südkaukasus, Neolithisierung, Landschaftsarchäologie, Chalkolithikum, Bioarchäologie

Introduction The special geographical situation of the South-

ally, by scraping the surface on a large scale,28 an

In summary, Kamiltepe appears as a more

historical age, but it covered two structures of Neo-

Semi-subterranean round building

A round structure (SU 519; 520) was uncovered

edres and cones, occasionally also as spheres. One

50 m can be calculated for the outer ditch. The

Description of the ditches – from the

side, and slopes more gently towards the inside of

tection of the structure, and this technique has

wide, which is filled with ash, pottery fragments and

Excavations at MPS 16 mill. BCE sites in Azerbaijan like Menteshtepe. In ad-

which however were deeply disturbed by plough-

Magnetometry of neolithic sites in the Mil Plain Magnetometer prospection

Here we present our processed high resolu-

Site MPS 19 is a hill site on the slope of the hilly

formation processes of complex deposits, or those

ples.71 More analyses will be undertaken to char- Results

The orientation of the phytoliths in different Post-depositional alterations

This is followed by a further lens of microchar- TSM 5 – Kamiltepe, operation 2,

Activities within buildings – MPS 4, TSM 6 – Kamiltepe, operation 2,

deposits will help to further characterise these and

The following report provides a brief and prelimin-

on this class of containers. They display triangular,

they could have been used as cooking trays.107 In-

to distinguishing between different contexts, people

acterized by thick walls and coarse fabrics. One of

The pottery of the site MPS 16

Pottery collected from MPS 16 mostly derives from

The bead workshop from site MPS 4

Kamiltepe: mineralogical studies It appears that there is no continuous profile

We have found indications of anthropogenic

, quartz decreases the shrinkage rate of the sherds

Besides feldspar minerals, other silicates like

and baryte (the last two less pronounced) occurred

The mineralogical composition was investigated

shifted over time. The Chrami likely flowed closer

The settlement Aruchlo I (Fig. 76) is located a few

possible to recognize coeval circular walls when the

The small circular wall/ring wall (AR05A108a),

Work of the former excavations was concentrated to

trenchments were superimposed by settlement

The cremation grave in Building Complex III Fig 98

127 The bone industry is currently being studied by D. Zhvania as

The few ceramic finds (Fig. 102) from these

During the three excavation seasons conducted in

Aruchlo Body sherds Diagnostic Total no.

2009 5,667 854 6,521

2010 4,484 761 5,245

Body Body Body Diagnostic Diagnostic Diagnostic Total no. %

Neolithic 3,729 2,964 3,483 652 606 582 12,016 71.94