Addressing the Desert Kites Phenomenon and Its Global Range Through a Multi-proxy Approach

Abstract

This paper argues that the wide geographical distribution of desert kites, which are huge archaeological structures of stone visible from satellite images, must be more broadly acknowledged as a momentous factor in the study of their variability and function. This is important so that researchers can more accurately understand and interpret their impact on biodiversity, landscapes and subsistence patterns. The first results and perspectives of the Globalkites research project are discussed and presented. Often considered as hunting traps, the kites could have also been used for animal husbandry. In a broader archaeological context, where kites seem to have been oper-ating from the Neolithic to recent historical times, we propose an interdisciplinary approach at the crossroads of anthropology (archaeology and ethnology), geomatics and Geographic Information Systems (GIS), geostatistics, mathematics and computer-ized data processing and geoarchaeological and bioarchaeological sciences (isotope studies, paleoclimatology, archaeozoology…). The principal aims of the project are to clearly articulate the variability of the structures and their relationship with the function and chronology of the kites. It is also crucial to discuss the wide distribution of these structures across the Middle East and Central Asia as a global phenomenon and the ideas that explain the dispersal and movements of people and/or traditions must be addressed.

Full text

Addressing the Desert Kites Phenomenon and Its Global
Range Through a Multi-proxy Approach
Rémy Crassard & Olivier Barge & Charles-Edmond Bichot &
Jacques Élie Brochier & Jwana Chahoud &
Marie-Laure Chambrade & Christine Chataigner &
Kamel Madi & Emmanuelle Régagnon & Hamida Seba &
Emmanuelle Vila
#
Springer Science+Business Media New York 2014
Abstract This paper argues that the wide geographical distribution of desert kites,
which are huge archaeological structures of stone visible from satellite images, must be
more broadly acknowledged as a momentous factor in the study of their variability and
function. This is important so that researchers can more accurately understand and
interpret their impact on biodiversity, landscapes and subsistence patterns. The first
results and perspectives of the Globalkites research project are discussed and presented.
Often considered as hunting traps, the kites could have also been used for animal
husbandry. In a broader archaeological context, where kites seem to have been oper-
ating from the Neolithic to recent historical times, we propose an interdisciplinary
approach at the crossroads of anthropology (archaeology and ethnology), geomatics
and Geographic Information Systems (GIS), geostatistics, mathematics and computer-
ized data processing and geoarchaeological and bioarchaeological sciences (isotope
studies, paleoclimatology, archaeozoology…). The principal aims of the project are to
clearly articulate the variability of the structures and their relationship with the function
and chronology of the kites. It is also crucial to discuss the wide distribution of these
structures across the Middle East and Central Asia as a global phenomenon and the
ideas that explain the dispersal and movements of people and/or traditions must be
addressed.
J Archaeol Method Theory
DOI 10.1007/s10816-014-9218-7
R. Crassard (*)
:
O. Barge
:
J. Chahoud
:
M.<L. Chambrade
:
C. Chataigner
:
E. Régagnon
:
E. Vila
CNRS, UMR 5133 ‘Archéorient’,Maisondel’Orient et de la Méditerranée, Lyon, France
e-mail: remy.crassard@mom.fr
C.<E. Bichot
École Centrale de Lyon, UMR 5205 ‘Liris’,Écully,France
J. É. Brochier
Aix-Marseille Université (AMU), CNRS, UMR 7269 ‘Lampea’, Aix-en-Provence, France
K. Madi
:
H. Seba
Université de Lyon, CNRS, Université Lyon 1 UMR 5205 ‘Liris’, Villeurbanne, France

Keywords Desert kites
.
Arid zones
.
GIS
.
Geoarchaeology
.
Hunting
.
Pastoralism
.
Interdisciplinary research
Introduction: The Desert Kites Phenomenon and the Globalkites Project
‘Desert kites’ (or ‘kites’) are stone constructions made of two long low walls called
antennae (also known as ‘guiding walls’ or ‘tails’) that converge into an enclosure (also
known as a ‘head’) flanked or not by several small cells (or compartments, or logettes,
also known as ‘hides’). The form of the enclosures varies, the antennae can reach a
length of several kilometers, while the size of the enclosure itself covers a surface of a
few hundred square meters to several hectares. Their distribution over the landscape
appears to be discontinuous and their density is very variable: 0.1 kite per 100 km
2
in
the Negev and Northern Sinai (Holzer et al. 2010) to 50 kites per 100 km
2
in some
areas of Syria (Échallier and Braemer 1995). ‘Desert kites’ were discovered in the Near
East and hence named by British aeroplane pilots who were flying over the Syrian–
Jordanian desert. Yet their distribution in the Arabian Peninsula has only become a
focus of studies during the last few years (Brunner 2008;Skorupka2010; Kennedy and
Bishop 2011). Kites even reach the confines of Uzbekistan (Betts and Yagodin 2000)
and the Caucasus (Barge and Brochier 2011; Barge et al. 2013;Gasparyanet al. 2013;
Brochier et al. 2014). Other types of trapping structures are known elsewhere: mainly
funnel-shaped, lined with wood or stone-lined constructions are known in North Africa,
Northern and Southern America and Scandinavia (e.g. Custred 1979;Ingold1980;
Riemer 2009; Benedict 2011; Storemyr 2011;Brink2013). This ‘kites phenomenon’
was most probably not either a continuous or a simultaneous process; thus, it therefore
covers a wide area. The range of our study is restricted to the Old World, and more
precisely, to the Middle East and Central Asia.
Dating kites in these particular regions is still difficult, as few remains were found in
what were usually poorly stratified structures or on the surface. The dating of Near
Eastern kites ranges from the 7th millennium BC (Helms and Betts 1987)torecent
historical periods, as documented by travelers’ testimonies (Burckhardt 1831). The
very rare radiometric or relative age measurements are the most convincing span from
the end of the Chalcolithic to the end of the Bronze Age (Échallier and Braemer 1995;
Holzer et al. 2010).
In order to consider kites as a wide range phenomenon, the Globalkites project was
developed by an interdisciplinary group of researchers. The project, whose aims and
methods will be developed here, covers a large geographical area, with a central
spectrum extending to about 600 km from north to south and 300 km from west to
east (Barge et al. 2013). This particular zone where kites are the most numerous
includes southern Syria, eastern Jordan and northern Saudi Arabia. Alongside this
Near Eastern core, similar structures were recently discovered in particularly distant
regions, such as in Armenia, western Uzbekistan and Kazakhstan, and in Arabia.
Observations from satellite imagery have revealed the existence of more than 4,300
structures (Fig. 1; see also the regularly updated and freely accessible interactive map
on the website www.globalkites.fr).
Consequently, in the last few years, the number of inventoried kites has increased
fivefold and the known distribution zone was greatly extended, suggesting some
Crassard et al.

exceptional potential for in-depth study of this phenomenon. More can therefore be
concluded regarding the use of kites across time and space. The main issues that the
project concemed itself with relates to the kites’ function and their dating: is it possible
to distinguish kites used as hunting traps from those meant as corrals for domesticated
animals or which are in the process of being used as such? Furthermore, what is the
time frame of the use of the kites? Other fundamental issues were also the focus of the
project such as the geographical spread and explanations for this extension: either
cultural convergence or real interregional technical diffusion? Finally, the adaptive
strategies by human groups must be addressed, as should issues of bio-economical
regulations in a fragile environment context such as potential overexploitation of wild
animal species and the ownership of an exclusive territory by humans.
This paper presents the main (old and new) problems linked to the kite phenomenon,
by means of the study of regional faunas and environments. Three fields of approach
Fig. 1 The inventory of kites at the scale of their distribution area; see also the Globalkites Interactive Map
freely accessible at www.globalkites.fr
Addressing the Desert Kites Phenomenon

will be presented: geomatics, geoarchaeology and computer sciences. All present and
future studies of the Globalkites project will subsequently be based on these analyses
and heuristic tools in order to focus on the kite phenomenon in its areas of wide
distribution.
The Function of Kites: A Device for Hunting and/or for Animal Husbandry?
Several theories about the function of kites have been posited since the 1920s. These
argued in favor either on their function as hunting traps or on their use for pastoral
activities (Rees 1929;Maitland1927;Fowden1999). Decades later, the issue is far
from being solved. Recent studies consider kites in the Levant in general as mass
killing traps for wild game, especially gazelle (Holzer et al. 2010; Bar-Oz et al. 2011;
Bar-Oz and Nadel 2013;Zederet al. 2013). By contrast, earlier studies have elaborated
on the progressive and multiple uses of kites. The function of these structures could
have shifted from trapping herds to taming wild animals and managing livestock
(Échallier and Braemer 1995). In all the cases, the concept, construction and the use
of kites suggests the collaboration of either group of people or of a community. These
communal activities are reflected by the large number of kites recorded, the wide
geographic range of kite dispersal, the size scale of kites, and by their possible function
for trapping a large number of animals.
In addition, a wide range of discovered rock art engravings depict ‘kite’ trapping
scenes with animal and human figures (Jordan: Harding 1953; Betts and Helms 1986;
Hershkovitz et al. 1987;MacDonald2005;Syria:Picalauseet al. 2004 and Uzbekistan:
see Betts and Yagodin 2000). It is nevertheless impossible to identify whether the
scenes reflect either the hunting of wild games or the management of livestock (Fig. 2).
The Hunting Model
Hunting as a possible function for kites reflects community involvement, a preference
for hunting traps and the occurrence of wild animals in the vicinity of kites. Travelers’
accounts dating to the sixteenth, nineteenth and twentieth centuries describe scenes of
gazelles hunting using kite structures in Syria and Jordan (Burckhardt 1831; Wright
1895; Sinclair and Fergusson 1902;Musil1928;Doughty1931;Simpson1994;Jabbur
and Conrad 1995; for more details, see Legge and Rowley-Conwy 1987 and Legge and
Rowley-Conwy 2000). Meanwhile, ethnographical studies (Fig. 3) refer to Bedouin
societies who practice communal hunting using trapping techniques similar to kites and
outline the different roles and statuses of hunters, as well as the complexity of mass
trapping and killing device. For instance, in the Sahara and in Arabia, nets are used
instead of stone-lined kites (Desombre 1946;Chapelle1957; Sergeant 1976; Bin Aqil
2004;Baroin2006). These traps made of perishable materials were also used outside
the desert zones of Africa (Dupré 1976). This type of hunting recalls a long tradition
can be seen in engravings and painted scenes in the Akakus and the Tasili-n-Ajjer (Le
Quellec and Civrac 2010).
Much wild game, in terms of ecology and ethology, is likely to be trapped by means of
a system of kites: bovids (sheep, goat, aurochs, gazelle, oryx
…), equids, camelids and
ostriches. These species occur in specific biotopes across the dispersal area of kites from
Crassard et al.

the deserts (Arabia, Jordan), steppes (Syria) and the semi-arid plateaus (Transcaucasia) to
Mediterranean ecosystems (the Levant), and mountainous zones (e.g. Transcaucasia).
Fig. 2 Rock art: a Khishâm-2, rock B37, Hemma, Syria (Picalause et al. 2004:2);b drawing of Cairn of Hani
rock arts, Jordan (Harding 1954); c tracing panel of rock drawing from Wadi Hashad, northeast Jordan
(Hoyland 2001:95);d painting scene of hunting using nets, Akakus and Tasili-n-Ajjer (Le Quellec and Civrac
2010: 255); e Khishâm-2, rock G28, Hemma, Syria (Picalause et al. 2004:2);f drawing of rock engravings
from Sinai (Bar-Oz and Nadel 2013); g tracing of hunting scene petroglyph, Har Micha, central Negev (Degen
2010:149);h Hippone mosaic, northeastern Algeria (Fowden 1999: 132); and i drawing of an engraving of
Sassanian royal hunt scene of deer in arteficial compound, Iran (Reed 1965)
Addressing the Desert Kites Phenomenon

Herbivorous ungulates are preys suitable for trapping by means of a system of kites.
They occur in groups and/or herds and some of them protect themselves by regrouping
Fig. 3 Ethnographical observation: a hunting traps according to Toubou tribes, Sahara (after Baroin 2006:39)
and b reconstruction of a gazelle hunt at Abu Hureyra, Syria (Legge and Rowley-Conwy 2000)
Crassard et al.

in defensive formations. For example, gazelles were once very common in steppe and
arid areas and accounted for the most favorable candidates for kite-type trap systems.
Observations on gazelle behaviour (Gazella gazella and Gazella dorcas)revealedthat
these herbivores tend to run parallel to low walls/fences and thus can be easily directed
by antennae of kites toward the trapping enclosure (Holzer et al. 2010). In the case of
goitered gazelle, (Gazella subguttorosa), corresponding data is lacking, but it is known
that under stress or in danger, these gazelles react instinctively by escaping in a closely
packed formation and run in a straight course, backward or forward, in front of the
pursuer (Kingswood and Blank 1996). Gazelles in general migrate seasonally accord-
ing to the availability of forage (Harrison 1968;Martin2000a). They are gregarious and
occur in small groups. But, in the case of goitered gazelle, they migrate in herds
numbering in the hundreds or thousands. Group structures change between the warm
and hot season. Three configurations can occur: female adults with juveniles, solitary
male adults and sub-adult males (Kingswood and Blank 1996).
Unfortunately, little is known about the exploitation of trapped animals within the
kites immediately after hunting. The only evidence was generated from the ethno-
graphical observation of Bedouin societies and from accounts of travelers who have
either seen or heard about kite use in Syria and Jordan.
Archaeozoological data clearly reveals that bone remains of wild ungulates such as
gazelles, aurochs, ibex, equids and cervids were the principal game hunted. They
occurred from the Paleolithic times in the Levant, northern Syria, Mesopotamia and
Transcaucasia. In Arabia, equids, antelopes and camelids were the main culled species.
During the Neolithic, along with the domestication of sheep, goats, cattle and pigs,
hunting practices still evolved and continued to focus on the same species found in the
vicinity, but tended to gradually decrease in importance with time (Davis 1982; Vila
1991; Bar-Oz 2004; Helmer et al. 2004;Gourichon2004;Gourichonet al. 2006;
Uerpmann 2008; Sapir-Hen et al. 2009;Martin1998;Martinet al. 2009; Tsahar et al.
2009;Bălăşescu et al. 2010; Chahoud and Vila 2011).
Hunted game was either trapped alive (Musil 1928; Jabbur and Conrad 1995:288;
Fowden 1999) or mass slaughtered, and the entire game carcasses or body parts were
carried away (Sergeant 1976;Baroin2006). However, remainders of meat processing
including discarded material, evidence of butchery, and meat conservation are not
apparent in the archaeozoological record. Scarce archaeological data reveal mass kill
spectrum or settlements specialized in carcass treatment at sites close to the kites.
Recent studies on the Khabur region have revealed one butchering site (Kuran, mid 4th
millennium BC) with remains of carcass disarticulation (bones of gazelle feet). This
Northern Syrian site is located near the Hemma kites group (Zeder et al. 2013).
The Husbandry Model
The pastoralism model suggests the use by transhumant societies of kites as enclosures
for gathering domesticated animals. Some authors have interpreted the scenes of
trapped animals in rock engravings as livestock management (Rees 1929; Échallier
and Braemer 1995; Jabbur and Conrad 1995).
In the Middle East, pastoralism and transhumance were the dominant economic
practices until the last century (Bar-Yosef and Belfer-Cohen 1989;MashkourandVila
2003; Ur and Hammer
2009). This region is a natural distribution zone for four
Addressing the Desert Kites Phenomenon

domestic ungulates highly exploited from the Neolithic period onwards: sheep, goat,
cattle and pig. The breeding pattern, based mainly on ovicaprine transhumance, implies
the annual migration of pastoral populations with their flocks. This indicates the need
for seasonal camps to protect the animals in isolated fragile environments, as well as
enclosures for concentrating livestock—the rounding-up of the latter for inspection or
other purposes (D’Hont 1994). Kite structures could reflect this use, hence the inter-
pretation of kites as structures related to livestock rearing.
Current research dates the domestication process of these ungulates back to the 8th
millennium BC (Helmer et al. 1998). The main source of animal protein, previously
based on wild game meat, was replaced during the Neolithic by livestock and dairy
products. Domesticated animals became the daily basis of diet and the centre of the
economy of Neolithic and Bronze Age societies, according to animal remains attested
on most archaeological sites (Grigson 1987, 2006; Vila 1998, 2006;Martin2000b;
Horwitz et al. 2001, 2002; Peters et al. 2005). In archaeozoological records, caprines
dominate faunal assemblages in the Near East, Arabia and Transcaucasia since the
Neolithic period (Vila 1998, 2006;Martin2000b; Horwitz et al. 2001, 2002;Grigson
2006;Bălăşescu et al. 2010; Vila and Helmer 2014). It is possible to state that the
economy was based on husbandry and not hunting activities based on
archaeozoological data from Bronze Age sites in Syria, taken from a concentration of
several hundred kites (e.g. Khirbet al-Umbashi (Southern Syria): Vila 2004; Al-Rawda
(Steppe, east of Hama): Vila and Al Besso 2014). If it is posited that kites were used for
hunting gazelles, it should expected that there is a significant frequency of wild game
bone in the assemblages. Consequently, if a direct relationship exists between kites and
settlement, the question arises of whether kites were already in existence, but not used
during the Early Bronze Age of this area or whether they were subsequently built.
Alternatively, one can assume kites were used by the inhabitants of the region for the
management of livestock.
Archaeozoological Analyses
In order to identify the function of kites, a complete analysis of faunal remains from
archaeological sites needs to be carried out in order to correlate kite and faunal
distribution. The study of the function of kites is considered in close connection with
an investigation on animal resources and human subsistence strategies. Hunting and/or
pastoralism models using kites imply a need for the gathering of large numbers of
animals, choices in consumption and an exploitation of animal products. Analysis of
the animal remains whether either slaughtered games or livestock, mortality profiles,
body part selection, provides valuable insights into both the human use of animals in
the past, as well as probably on the methods of acquiring these resources, whether
related to kites or not.
So far, no global analyses of archaeozoological records have taken into account the
variability and distribution of kites in relation to the availability of animal resources,
their procurement and their consumption. This should be done in order to examine the
correlation between kites as massive traps/enclosures and the demands of ancient
societies. Furthermore, there does not appear to be a direct relation between the
consumed wild game found on settlement sites and hunting and capture zones. By
integrating structure, geographic location and function of kites and correlating these
Crassard et al.

with animal occurrence (current and historical) and bone assemblages in the areas of
kites (Middle East, Arabia and the Caucasus), it would be possible to establish several
functional patterns. These patterns could be used to test the validity of each model in
order to suggest a possible connection with kites. If the ecology and behaviour of
animals were considered as vectors, it would be possible to examine the consistency of
hunting and/or pastoral uses in conjunction with the regional management of animal
resources.
Several analyses were carried out with this aim in mind: (a) the study of animal
remains from the excavation of kites; (b) the recording and mapping of current and
ancient distribution of species. This project focused on the ungulates; (c) the creation
and filling of a database of archaeozoological analyses which include regional and
chronological features of sites. The main issue is to homogenize the dataset from
different publications and current studies to be able to analyze and compare ancient
animal distribution, the animal-based economy and the evolution of wild game or
livestock dependency during the Holocene. (d) Analyses of the ecology and ethology of
animals likely to be trapped by kites. For instance: the classification of gazelles remains
complex despite recent phylogenetic analyses (Bärmann et al. 2013;Lerpet al. 2013).
Nevertheless, three main species of gazelles in the Middle East can be distinguished
(Harrison 1968): mountain gazelle (G. gazella), dorcas (G. dorcas) and the goitered or
Persian gazelle (G. subgutturosa). Mountain gazelles are endemic to coastal regions
and mountain ranges of the Levant and the Arabian Peninsula. The dorcas gazelle is a
subspecies in the southern Levant, Sahara and East Africa. The goitered gazelle is
widespread in the Near and Middle East and in an area spanning from the Arabian
Peninsula to south Mongolia and the Caucasus (Kingswood and Blank 1996). This
gazelle subspecies lives in desert and semiarid biotopes of all types. (e) Ethnographical
study must be conducted to e xplore different forms of pastora lism according to
available studies, to make observations relevant to current livestock management and
transhumance practices and by examining hunting techniques. These ethnographical
records can serve as a benchmark for the study of the function of kites.
Three patterns should be cross-examined according to period and region for a global
approach on the management and use of animal resources, using theses analyses:
archaeozoological data, animal occurrence and behaviour and ethnographical studies.
Cross-matching these patterns in this larger framework with other aspects of kites in the
Globalkites project could enable a better understanding of the function of kites and their
possible relevance to a cultural timeline. It would then be possible to examine the
viability of the three-tiered system relating trap-hunters and prey, on the one hand, and/
or enclosure-shepherd and flock and the kites themselves, on the other. In other words,
throughout these archaeozoological, ethological and ethnographic analyses, a feasibility
study is proposed of: animal occurrence, mobility and role in ancient society, activities
of the hunters/pastoralists who might have built and used the kites, their need for such
trapping system and the benefits of these structures.
Environmental Perspectives
The environmental contexts of the kites are periodically mentioned and described in the
literature and authors generally agree on the fact that the location of kites in the
Addressing the Desert Kites Phenomenon

landscape reflects choices, most often recurrent. These were guided by a profound
knowledge of the environment and of the biotic communities that constitute it, and by a
special awareness of faunal behaviour. This T raditional Ecological Knowledge (Berkes
2008: 7) is, according to B.D. Smith (2013: 10), a fundamental aspect of societies who
used structures such as large-scale traps. Several criteria have been suggested for the
establishment of kites based specifically on traditional ecological knowledge.
Review of the Criteria Relevant to the Choice of Location for Kites
The proposals that explain the choice of locations exclusively emphasize the links
between environmental characteristics and the behaviour of wild fauna. The criterion
most frequently mentioned for the choice of kite construction is the search for food, and
less importantly for water as these were guided by daily and seasonal cycles. Many
authors posit the view that kites were constructed and oriented at the regional scale in
relation to faunal migration routes (Helms and Betts 1987; Perevolotsky and Baharav
1991;Betts1998; Betts and Yagodin 2000;DeomandSala2009; Morandi Bonacossi
and Iamoni 2012; Kempe and Al-Malabeh 2013;Brochieret al. 2014). While at the
local level, their construction is connected to daily itineraries as well as to the
availability of pasture zones and water holes (Helms and Betts 1987; Perevolotsky
and Baharav 1991; Betts and Yagodin 2000; Holzer et al. 2010;Nadelet al. 2013).
Topography is the second criterion most commonly suggested (Helms and Betts 1987;
Betts and Yagodin 2000;Holzeret al. 2010; Morandi Bonacossi and Iamoni 2012;
Nadel et al. 2013; Quenet and Chambrade 2013;Brochieret al. 2014), i.e. essentially
the taking advantage of a change of slope hiding the enclosure until the animals either
approach or enter it and therefore either preventing or substantially curtailing any
likelihood of flight. It was concluded that the direction of either the sun or the
prevailing winds was less important in the choice of location for kites (Échallier and
Braemer 1995;Nadelet al. 2013).
These hypotheses are inspired or comforted by ethnographical observations carried
out on European or North American large-scale traps (e.g. Ingold 1980; Arkush 1986;
Gordon 1990). Nevertheless, the basis of some of these hypotheses was questioned
despite the recurrence of patterns for kite location (Échallier and Braemer 1995). It can
be argued that the study of environmental contexts and their relationship to kites has not
until now been the focus of neither a systematic and comprehensive approach, nor of a
comparative study at a scale that covers their area of distribution.
A New Systematic and Global Approach for the Environment of Kites
The study of the environment of kites is designed to determine the role of environ-
mental conditions in the construction of structures dedicated to subsistence activities.
The analysis of resources and environmental constraints enables researchers to under-
stand how and why these activities were carried out. The Globalkites project addresses
these environment studies by means of three spatial analyses. At the micro-regional
scale, the reconstruction of the kites in their natural context aims to understand choices
for the location of kites and explaining their use(s) and the way(s) they operated. At the
regional level, this means a study of the environmental zone which was inhabited by
the human groups who used the kites. A comparison of data at the macro-regional scale
Crassard et al.

will enable both the similarities and the discrepancies in the choices of where
kites w ere placed as well as the exploitation strategies of kites according to
environmental zones to be grasped. The environment is tackled through three
aspects that were researched: climate, natural resources and topography.
Furthermore, owing to the relationship between environment, animal presence
and kites’ location, environmental studies and archaeozoological analysis are
carried out in closely interrelated manner.
Climate
The Near East has a Medite rranean climatic regime, with rainy winter and dry
summers. This has been the case for approximately 17,000 years (Issar and Zohar
2007: 53). The climate is characterized by a substantial decrease of rainfall from the
West to the Southeast. In the current state of knowledge, kites are restricted to semi-arid
to desert bio-climatic zones, which means they are found where annual rainfall ranges
from 100 to 600 mm. Kites are generally absent from the most humid zones and in
certain cases from the driest regions. This specificity needs to be explored in detail, and
it is important to establish if the reasons for such distribution are preferentially human
and/or environmental. The succession of climatic oscillations during the Holocene had
a greater or weaker impact depending on periods and regions (Migowski et al. 2006;
Robinson et al. 2006; Verheyden et al. 2008;Develleet al. 2011). This climatic stress
possibly had an influence on the development and abandonment of kites, especially
when these were used over a quite short time span.
Climatic aridity and seasonality are causes for constraints on vegetation and fauna,
as well as on human settlement. These constraints can lead to a number of adaptations,
especially those affecting lifestyles and subsistence and the development of socioeco-
nomic and technical innovations (e.g. pastoral nomadism, hydraulic installations). Were
kites one of the forms of adaptation to these constraints and were they used year-round,
or during one or several seasons in particular? As widely observed, climatic seasonality
in part influences the subsistence activity cycle, particularly guided by the differential
distribution in space and time of the quantity and quality of potential resources (Monks
1981). The hypothesis of an establishment of kites in relation to faunal seasonal
migrations will be explored further, as will be the issue of seasons during which the
kites would have provided the best returns.
Natural resources
Climatic characteristics of arid zones are the reason for the predominantly low and open
natural vegetation, at times concentrated in certain micro-catchments and microenvi-
ronments more favorable to its development. They are also the cause of a temporary
access to surface water in large areas. Échallier and Braemer (1995:50, 60) are not
convinced of the potential link between the vegetation and the hydrographic network
and the kites themselves. However, vegetation distribution redirects faunal daily
itineraries and could have therefore influenced the implantation of kites. Furthermore,
access to drinking water remains an important element to be taken into account. The
authors mention the recurrent location of southern Syrian kites at the margins of zones
where water was found at certain times of the year.
Addressing the Desert Kites Phenomenon

Moreover, Burckhardt (1831: 220) observed gazelle hunts carried out with devices
akin to the kites in concept, noted that enclosures were located close to water holes and
sources, since these attracted gazelles in the summer. Water and natural vegetation are
therefore studied in a seasonal perspective, i.e. the differential distribution and avail-
ability during the year of these two resources. The presence of water resources would
have provided reasons important enough for the selection of specific locations for kites.
Some animals are ‘obligate drinkers’, i.e. they need regular access to a water resource.
This need becomes critical during the dry season. For other species, pasture grounds
suffice to ensure water requirements, some animals even practicing a mixed strategy
(Speth 2013: 179). The natural resources including vegetation and water and the faunal
distribution from the Holocene period, based on archaeozoological analyses of the
Globalkites project, are currently being mapped together.
Topography
An observation of relief and terrain, understood from the features that it comprises (e.g.
plateaus, mountain ranges), enables researchers to outline diversified catchments. Some
initial preliminary observations can be put forward based on spatially orientated
inventory of kites.
Those kites at the highest elevations are located in Armenia, since their altitudes are
approximately 1,500 m above sea level (asl). Apart from those in Armenia, elevations
above 1,500 m concern only restricted zones within the distribution area of the Near
Eastern kites: the Anti-Lebanon/Hermon ranges and the central Jabal al-‘Arab in Syria,
and some very localized sectors in southwestern Jordan and northwestern Arabia, as
well as in the southwest of the Arabian Peninsula. Apart from those kites in Armenia,
only 7 % of kites are located at an elevation above 1,000 m asl. These are in Syria on
the eastern slopes of the Jabal al-‘Arab, Southern Palmyrena ranges and north of the
Anti-Lebanon. The norm is a location of kites at between 500 and 1,000 m asl (70 %),
with two thirds of the kites found between 500 and 750 m. This upper limit in elevation
(1,500 m), therefore, does not seem to be a sufficiently relevant datum, and in any case
cannot be compared at the macro-regional level. Yet on a regional scale, elevation must
be confronted to wild fauna occurrence in higher mountains, as well as to vegetation.
One observes primarily, at the macro-regional scale, that kites are absent from plains,
endoreic depressions typical of arid areas and from the core of mountainous sectors.
They are, in contrast, preferentially implanted on the latter’s foothills and the piedmont,
as well as on plateaus (most often of basalt). Were kites effectively associated with
these environments in particular? For what reasons and why are they absent from
certain types of environments? An answer to these questions lies in the confrontation of
data related to the environment and to that relevant to subsistence activities.
At a finer scale, the systematic analysis of the positioning of kites and of their
various elements, in relation to local topography and components of the landscape,
greatly contributes to the study of the kites’ inner workings. It was often observed that
the existence of a change in slope hid the enclosure from the corridor formed by the
antennae. New recu rrent elements should be identified. This concerns for
instance the position of the cells in relation to the enclosure, sometimes also
hidden by a change i n slope (as observed in Jordan). Another element could
also for example be the existence of a corridor, such as a river or streambed,
Crassard et al.

which would have enabled the easy redirection of the animals towards the kite
(see Arkush 1986 on the hunting of pronghorns).
The analysis of environmental contexts consequently yields an overview on poten-
tial and the constraints linked to animal capture zones and environments. It also clarifies
the relative difficulties in the selection of geographic location of kites and issues of
human subsistence. Therefore Geographic Information System (GIS) is essential in
order to observe kites from a global perspective.
GIS as a Critical Tool for a Geographical Understanding of the Kites
Phenomenon
In terms of all the archaeological issues related to the kites, one the most important is
their geographical dimension. More precisely, the spatial component is intertwined with
most issues posited by these installations—distribution of animal species, layout and
arrangement of environmental settings, control and territorial layout, economic net-
works, diffusion phenomena, etc… The spatial component is related to the ecosystem
which is that it is to the anthropo-systemic sphere. The approach therefore
derives from general geography, both human and e nvironmental. T he question
of scale is also ever-present, from the macro-regional scale to that of the
architectural element.
The use of GIS becomes from then on unavoidable: the variety and quantity of data
and their spatial and multi-scalar character in themselves justify its use for the recording
and consultation of inventories. Representational functionalities ensure moreover being
able to face the (in principle recurrent) need for mapping documentation. The antici-
pated use of GIS also is due to the very functions of analysis, and these are suitable to
make the tool heuristic since it allows the cross-matching of data, distributional
analysis, analysis of distances and geo-statistics.
Inventories
The study of kites at the scale of their area of distribution presupposes an inventory
which is as comprehensive as possible. Indeed, the information source which shows the
greatest homogeneous character consists of high-resolution satellite imagery. The
images at the public’s disposal, provided by Google Earth and Microsoft Bing, are
complementary in that they give an almost complete and comprehensive high resolu-
tion coverage. The inventory partly derived from recent publications on kites (Kempe
and Al-Malabeh 2013; Kennedy 2012; Morandi et al. 2012), was completed by
meticulous and detailed research and today has reached a figure of more than 4,300
items (Barge et al. 2013). The precise geographical location of these kites is the GIS’s
first feature class. It includes a feature attribute table with some twenty fields which
allows information to be stored on size (length of antennae, enclosure surface…)as
well as and information on figures, numbers, counts or on the presence of architectural
elements (number of small cells, position of the latter, presence of a funnel-shaped
entrance…) This information is being entered digitally. If the detailed information
required on the totality of kites in the inventory is too long and too painstaking a
process, a meaningful sample could allow numerous analyses (see below). At this
Addressing the Desert Kites Phenomenon

general scale, environmental and zoological data are also in the process of being
incorporated. The set of these data are gathered in a spatial database, which moreover
includes a Digital Elevation Model (DEM).
It is at either the regional or micro-regional scale that one can best grasp the issues
relating on the one hand to kite implantations in a given environmental setting and
territorial structuration, and on the other to the functioning of kites. To this end,
‘windows’ were established while taking into account both the kite distribution and
data availability (from fieldwork or if this was lacking from bibliography). A spatial
regional database, which was interfaced with the general spatial database, corresponded
to each window. One of these windows concerns the region of the Aragats in Armenia,
where three fieldworks have already taken place. Jointly to the punctual implantation of
kites, one finds a feature class of archaeological sites. Kites and their constitutive
elements are described in a feature dataset, which includes a feature class called ‘walls’.
Within it, these ‘walls’ of kites were digitally recorded either from high-resolution
imagery or from a ground plan in the field (drawn with a Trimble Differential GPS).
In this case, the attribute fields enable the description of the construction technique
(simple or double facing), and the height and width of walls in homogeneous sections.
This feature dataset also includes specific feature classes meant to describe the small
cells (types, dimensions), subdivisions, as well as the eventual presence of other
elements (for instance tombs). It also comprises feature classes meant to record work
in the field (soundings, profiles and sections, field samples). Kites which were studied
were also the focus of aerial photography using a kite. Photogrammetry from this
pictures enables to obtain ortho-photographs and DEM at a very refined scale (Fig. 4).
This set of documents is stored in a spatial database, which coincides with the regional
window.
Fig. 4 Cell L14-01, example of ortho-photography (to the left) and of hilshade derived from a DEM (to the
right) produced by photogammetry and stored in the spatial database
Crassard et al.

Mapping
Mapping is a source of documentation which is of prime importance when
dealing with the issue of kites. The functionalities of automatic representation,
just like those of publishing, are subsequently applied. The scale of documents
produced can be very variable, from general distribution area to architectural
components of a kite. For greater scales, Mercator transverse projections are
used, with a UTM strip coinciding with the relevant zone. For small scales, a
conical Lambert consistent projection concentrated on the distribution area of
kites to be defined.
Simple distribution maps can therefore be drawn (Fig. 1). In particular, it is
quite relevant to separately map the different descriptors obtained from satellite
imagery. This enables researchers to pinpoint if elements of size or presence of
certain architectural attributes show a particular form i n terms of spatial distri-
bution. Figure 5 forinstanceshowsthatthelengthofantennaeisarather
important criterion in isolated kit es (Fig. 5a) or whether the orientati on of kites
conforms to that of slopes (Fig. 5c). This example taken from Armenia never-
theless possesses limited significance in that this region shows quite consistent
morphological unity. Undoubtedly, these maps will enlighten clearer differences
when they will be redrawn and published at a regional scale, and they will be
very useful in discerning regional peculiarities.
Maps at more refined scales are interesting in that they reveal the insertion of kites
into the environment and expose their architecture. It is at this scale that the functioning
of kites or sets of kites can be grasped. Figure 6 shows the alignment of kites and their
position in relation to the edge of the volcanic flow. Here the antennae point uphill and
encircle the rim of the plateau which is of low inclination and enclosures on the slope of
the flow’s edge and entrance where the slope changes. The plans of kites reveal their
architecture, the eventual specificities or superimpositions. Kite 67 (Fig. 7), for exam-
ple, is probably later than a more ancient kite since it is only a pointed cell survives (to
the north-east), while its enclosure wall was destroyed to the west by medieval
occupation. Finally, these documents also help in locating where fieldwork has been
carried out (Fig. 8).
Fig. 5 Examples of theme-based mapping: a length of antennae, b surfaces of enclosures in hectares, and c
direction of opening
Addressing the Desert Kites Phenomenon

Modes and Methods of Analysis
In many ways, it is difficult to select the most efficient types of analysis, since
possibilities offered by the GIS are so varied. The possibilities include: spatial cross-
matching of information, closest neighbor analysis, analyses of distances, distributions
and the combinations of descriptors (Rodier et al. 2011). The orientation of GIS
research relies heavily on information that both previously gathered and to be collected
in the future.
However, several approaches can already be imagined because they were tested in
Armenia. Statistical analysis and a geo-statistic examination of details (from satellite
imagery) make scholars expect interesting results. Multiple correspondence analyses
are aimed at revealing original combinations that are recurrent. What is projected is to
identify in more secure fashion the regional specificities that are perceived in a rather
more obscure and empirical manner, and subsequently seek to interpret these results.
These specificities can be rendered discrete and eventually explained eventually by
functional, chronological and environmental factors. Conducted in Armenia, this re-
search has enabled emphasizing the morphologically homogeneous character of this
sub-set. It has also permitted the allotment of a distinct function from that of kites to a
set of installations which, without possessing of all attributes of kites, do present some
of their salient features.
Cross-matching certain characteristics of kites with environmental parameters can
also yield interesting functional clues. In Armenia for instance, the link between the
direction of the opening of kites and the orientation of slopes was proven (Barge and
Brochier 2011). The direction of the opening of kites is in effect an interesting element
of data that some associate with the hunting of migrating animals (Morandi Bonacossi
and Iamoni 2012). However, it seems quite awkward to state this in the absence of
reliable data on migrations on the one hand, and given the unavailability of regional
comparisons for the direction of openings on the other. In Armenia, given the general
orientation of slopes towards the east and southwest, it is difficult to determine with
precision the part played by gradients and slopes. Consequently preferential orienta-
tions could not either be confirmed or disproven. At the level of the entire set of kites,
Fig. 6 The Aragats kites (Armenia) at the scale of a single slope
Crassard et al.

these analyses will undoubtedly provide relevant clues because topographical contexts
often stand in greater contrast and the directions of openings become more clearly
apparent.
At the same scale, but with less certainty because the necessary data remains
unavailable, it was planned to search for regularities in the relationship between
environmental data and the location of kites. One can in fact observe that kites are
located both in zones marked by aridity, in sectors void of any tree cover and places
with a marked topography. These cross-matches should entail confirmation of this fact
and yield more precise details on the matter. This type of cross-matching should also be
conducted on the archaeozoological data. In particular, localized data on fauna from
different periods at archaeological sites could be compared to kite distribution and
Fig. 7 Example of a kite plan: Kite 67 drawn with a Trimble DGPS
Addressing the Desert Kites Phenomenon

could provide answers on the species concerned by these huge trapping devices. In the
cases where the available data does allow this, the relationship between kites and
settlement patterns of a given territory (sedentary, nomadic) can on the one hand be
assessed, and settlement dynamics examined. Such analysis can be undertaken by
carrying out analyses of relative distances between archaeological sites (occupied at
different periods) and the kites themselves. Finally, but on a more experimental level,
the functions of complex distances that include directions of displacement (supposed
migrations) and the roughness of terrain (geographical constraints) should be called
upon to attempt to create a model of animal displacement, the aim being to test
functional hypotheses.
Geoarchaeology and Chronometry: Two Complementary Approaches
An unavoidable prerequisite of any archaeological discourse is the dating which is both
relative and absolute. This problem remains despite a century after initial discovery
particularly difficult to tackle. Of the thousands of kites known in the Near East, very
few have been excavated and even fewer subjected to chronological analysis. The most
numerous and the most reliable data come from kites in the Sinai and the Negev (the
analyses concern five of the seventeen known kites) where the constructions were used
between the end of the fourth and the third millennium BC (Kobusiewicz 1999; Holzer
et al. 2010;Nadelet al. 2010). But these dates only provide information only on the
chronology of kites that are very particular morphologically and outside the Syro-
Jordanian region, where they are both much more numerous and varied. Elsewhere,
chronological indications are rare and rather unclear. In any case, they do not allow one
to assess whether the kites were used over millennia from the very beginning of the
Neolithic either beforehand, or only during one or several brief periods. However it
appears that the end of the fourth and the third millennium were the main periods of use
(Échallier and Braemer 1995). The great morphological diversity observed in both
Jordan and Syria (Helms and Betts 1987; Échallier and Braemer 1995), the frequent
Fig. 8 Cell L14-01 (Aragats, Armenia): localization of fieldwork (topographical profile, soundings) and
photographs at ground level
Crassard et al.

modification in plan and of course the accounts of usage in the modern period argue
that kites were used over a long time frame.
Two complementary chronological approaches have been developed in the frame-
work of the Globalkites project: geoarchaeological analysis which rests essentially on
the analysis of anthropogenic dust and absolute chronological measurements.
Anthropogenic Dust and Relative Chronology
The kite cells, which are generally deep, are the places where sediment, both wind-
blown and/or from runoff, have most commonly accumulated since abandonment of
the constructions. Consequently, they preserve the traces of a long history of visits by
men, by their livestock or by wild animals. The method, developed during the study of
the Syrian steppe’s pastoral enclosures (Brochier 2014), was successfully transposed to
the case of the Aragats kites in Armenia. Without entering into the arcane intricacies of
a complex methodology, something already detailed by Brochier et al. (2014), the
Armenian example has brought to light the fact that since their abandonment the cells
were filled by detrital sediment mixed with biogenic remains originating from domes-
ticated and/or wild ruminants visiting the kite and its immediate surroundings. What
was in fact observed was the presence of carbonated faecal spherulites, siliceous
phytoliths (although part of them no doubt were the result of natural decay of local
vegetation) and above all numerous remains of silicified unicellular algae very classi-
cally associated with ruminants’ faeces (Brochier2002). Settlements, characterized
mainly by wood ash (calcitic pseudomorphs after calcium oxalate, or POCC;
Brochier and Thinon 2003) are, in all cases, extremely hard to identify. Following
pioneering works by Smol (1985), one has become aware of the balance between
diatom frustules and chrysophycean statospores during the course of time, as a function
of anthropogenic disturbance which was the reason for the catchment’s eutrophisation.
It is this very feature that allows us to follow anthropogenic disturbances in a given
geographical sector, to measure their impact and, by comparison with spectra obtained
in dated rural settlements, to order the different stratigraphic sequences observed. The
example of the Aragats (Fig. 9) demonstrates that the studied kites are considerably
earlier in age than the Middle Ages. This fact is confirmed by some rare cases of
superimposition which is a particularity that has been discussed elsewhere (Helms and
Betts 1987: 50; Échallier and Braemer 1995: 54; Zeder et al. 2013: 115). They
additionally enable one to proceed to a screening of samples for radiocarbon analysis.
Absolute Chronology
Apart from exceptional examples from the Negev and Sinai, the absolute dating of
these constructions is a particularly difficult task. In the framework of the Globalkites
project, two methods seem in our opinion to be applicable in many ca ses: the
measurement of residual radiocarbon and Optically Stimulated Luminescence (OSL).
In some geological contexts and specific environments such as the case of the Aragats
the uranium-thorium (U/Th) method applied to carbonated crusts also appears to be
able to provide results.
During excavation carried out in the cells of the kites, it was rare to encounter either
wood charcoal or carbonized seeds, and finding them deep down in stratigraphic layers
Addressing the Desert Kites Phenomenon

was an even scarcer occurrence. The procedure that routinely applied consists in
sampling, was hing on a fine mesh (0.5 mm) and sorting out under a binocula r
magnifying glass the residue from several kilograms of sediment sample extracted
from the soil, while bearing in mind the stratigraphy. Experience demonstrated that
carbonated material, both in Armenia and Kazakhstan, despite being not very abundant,
is often present in sufficient quantities to be analyzed through modern AMS techniques.
One should in particular emphasize the frequent occurrence in these fills of two types of
biogenic carbonates potentially yielding reliable results: millimeter-large fragments of
bird eggshell and limestone seeds of Boraginaceae of the Lithospermum genus. Indeed
the ability to gather the tens of milligrams necessary to produce a date is possible. The
presence of charcoal and of bone splinters is much more random. This type of research,
though necessary, nevertheless in all cases only provides researchers with a terminus
ante quem. The greatest difficulty resides in the determination of a terminus post quem.
The problem can only be solved on a case-by-case basis, depending on geological,
stratigraphic and architectural contexts, by calling upon radiocarbon analyses, OSL or
U/Th.
A Mathematical and IT Processing Approach to Kite Recognition
An important aspect of the project is the development of a pattern recognition tool that
will be based on the specific shapes and sizes of kites. This tool will allow a
comprehensive inventory of these structures worldwide. The systematic identification
Fig. 9 L59.01 cell. Examples of the evolution of relative proportions of diatom frustules to chrysophycaean
statospores; the horizontal grey stripe marks the position of a level yielding some bone splinters. Absolute
chronological benchmarks in italics; in normal fonts dates estimated by comparison with the spectra of known
periods. Note modern (and present day) eutrophisation of humid zones
Crassard et al.

of kites on satellite images from Google Earth enables to delimit this phenomenon from
a spatial viewpoint. It also allows a determination of other hunting or pastoral devel-
opments that are analogous to kites but must be culturally excluded.
The identification of objects in images is a fundamental scientific issue in many
fields and industrial applications. Indeed, the recognition of objects and shapes in
images is one of the most difficult problems of computer vision. In fact, the same
object can seem completely different depending on the standpoint, the scale, the
lighting, quality of images and many other factors such as partial occlusion and
background add-on. Applied to the analysis of satellite images of Google Earth-type,
we will process the images in order to minimize these drawbacks, especially lighting,
color changes and background add-on.
We can classify object recognition methods into four main approaches (van de
Sande et al. 2010): geometry and matching, sliding windows, deformable models,
and feature extraction and classification. The latter approach currently gives the best
results regardless of the type of objects. However, the other approaches may be
particularly effective according to the context. In particular, when the desired object
has a structured form, which is the case of kites, the geometry and matching approaches
are particularly interesting. We propose in this study to go beyond the traditional
boundaries of geometry-matching and feature extraction-classification approac hes
and combine the advantages of both approaches in a new model for object detection
based on a graph modeling of images. A graph is a modeling tool composed of a set of
vertices and a set of edges that connect the vertices.
Graphs are frequently used in image recognition. The main advantage of using
graphs for the representation of images is the integration of spatial information in the
representation. Indeed, conventional representations such as color histograms, texture
descriptors, etc. provide no information on how the regions of interest of the image are
arranged. The graph representation allows us to describe the structure of the image, i.e.,
how the parts are arranged relatively to each other. Moreover, according to the chosen
types of relationships, this graph representation (Fig. 10) can be invariant under certain
transformations such as rotations of the image or translations of parts of the image.
Graph theory is a powerful tool for studying complex combinatorial structures.
Among the major research topics in the field that develop methods for the design
and manipulation of complex structures, we can cite: graph embedding, graph decom-
position and partitioning and graph Kernels (Haussler 1999;Donget al. 2008;Bunke
and Riesen 2011;Lagraaet al. 2014). Graph kernels construct the similarity between
complex objects using the similarity of parts of these objects (Haussler 1999;Gärtner
et al. 2003). The basic idea is that it is easier to design a similarity function between
small parts than similarity function between objects made of these parts. To do so,
graph decomposition and partitioning are important issues. Graph embedding generally
aims to represent the graph structure by a vector, for example by computing its distance
to a set of graph prototypes (Emms et al. 2007).
The problem of kite recognition using graphs can be viewed as a set of steps as
follows; Step 1 is characterized by modeling a kite by a graph and constructing a set of
graph prototypes for kites. An initial model for kite recognition is established, based on
archaeological and geographical knowledge. This model will be gradually improved
with the feedback of archaeologists and geographers (experts on the results provided by
the model). During Step 2, the following tasks are realized for each processed image:
Addressing the Desert Kites Phenomenon

(a) extract the kite from the image using image processing tools, (b) represent the kite
with a graph, and (c) compare the obtained graph with the prototypes of Step 1 to verify
Fig. 10 Main steps of the kite recognition tool: a a kite as it appears on a satellite image, b the kite after image
processing (segment detection and noise suppression), and c a graph representation of a kite
Crassard et al.

that it is really a kite (and not a river for example). Figure 10 illustrates the results of
tasks 1 and 2 using the actual version of the tool under implementation.
To achieve these two steps, several challenges must be addressed. The first challenge
is to identify the most appropriate graph representation of a kite. This representation
must bring together the topological structure of the kite and its corresponding seman-
tics. Then, it proposes efficient algorithms for matching the obtained representations.
This turns into sorting, retrieving the algorithmic methods of graph matching, and then
adapt them to the research issues of kite identification. A second challenge is to enable
a kite identification that tolerates errors or imprecision in the query. Indeed, the kites
have several forms depending on their state of preservation. Thus, it is suitable to offer
approximate matching algorithms that express situations where kites are more or less
preserved. A third challenge in this project is related to the complexity of the graph
matching algorithms that have exponential execution time in the general case.
However, for some subclasses of graphs, algorithms with polynomial complexity have
been proposed. So it is important to identify subclasses of graphs corresponding to kites
in order to apply the most efficient algorithms, or propose new algorithms.
Other application areas are also seeking and assessing basic tools for searching and
comparison of structures represented by graphs. We can cite for example: biology and
biochemical networks, organic and molecular chemistry, the analysis of documents, the
semantic web and databases, etc. Thus developing an efficient and fast platform for
recognizing objects by graph matching is an interesting scientific advance in the field of
pattern recognition with applications in several domains.
Final Remarks and Conclusions
The kite is a landmark that reveals a way of occupying territory. It is an architectural
feature of social groups, which hence left a reflection of their territory and catchment.
However, these people have sometimes left very few traces enabling their identifica-
tion. If the extent and density of these installations is taken into account, the kite is a
massive phenomenon whose role was probably crucial in the development of societies
in these arid regions. Therefore the perspective is inverted in relation to traditional
archaeological research, where groups are most often identified by their domestic
settlements, with their subsistence modes still undefined. The kites, as a phenomenon,
are therefore a challenge for the archaeologist in that they require creative and new
approaches. The approaches used by of the Globalkites project are multiple and
complementary.
On the larger scale, digital tools (GIS and pattern recognition) could answer
questions which have been scarcely addressed until now. The GIS this is at the core
of the project, both due to its capacities in structuring inventories and to its ability to
produce cartographic documents. This centrality also means that the relevance of
results from analyses is still difficult to assess. The more related approaches
(geoarchaeology, environmental studies, and archaeology) will have produced solid
results (chronological data in particular), the more one will be able to mobilize the
range of possible analyses with the GIS and the more results are likely to be relevant.
Furthermore, the mathematical and IT processing approach for kite recognition will
propose a collection of models for kites that can be used for other archaeological
Addressing the Desert Kites Phenomenon

remains, provide answers to the questions surrounding kites and their environments,
and develop approaches for the identification of kites. It will also provide new
approaches for pattern recognition on images that can be adapted to different contexts
and application frameworks, propose scalable pattern recognition algorithms, and
finally validate these algorithms in a real application domain. Analyzing high-
resolution images to locate kites, carrying out a field survey to examine structures,
excavation and sampling, are all stages in the process. The recording and the analysis of
dusts might be what breaks new ground in the future when it comes to chronology. The
results of surveying and dating, combined with a description of the kites in GIS data are
to be analyzed concomitantly with the mapping of animal distribution,
archaeozoological recording and the environmental information.
Obviously, the kites phenomenon, because of its global range, was probably more
influential than previously thought. The human–animal relationships mediated by these
constructions have both ecological implications (for instance the environmental impact
of these activities), and socioeconomic ones (i.e. territorial demarcation and control). In
order to better understand the actually recognized importance and particularly wide-
spread distribution of this phenomenon, it will be useful to reassess data quality and to
refocus studies both within a precise context and in a global perspective.
Acknowledgements GLOBALKITES research (2013–2016) is financed by the Agence Nationale de la
Recherche, France (No. ANR-12-JSH3-0004-01, to RC). Kites shapes recognition program (2013-2016) is
financed by a Labex IMU grant from the University of Lyon, France (to OB, CEB, HS, EV). We wish to thank
the three reviewers for their very helpful comments, as well as Martin Makinson and Alicia Colson for English
editing. For field work in Armenia, we thank Arkadi Karakhanyan, Iren Kalantaryan and Pavel Avetisyan. For
field work in Kazakhstan, we thank Zhaken T aimagambetov, Renato Sala, Jean-Marc Deom and Constantin
Plakhov. For a first visit in Jordan, we wish to acknowledge the help from Wael Abu-Azizeh and Mohammad
Tarawneh.
References
Arkush, B. S. (1986). Aboriginal exploitation of pronghorn in the Great Basin. Journal of Ethnobiology, 6(2),
239–255.
Bălăşescu, A., Vila, E., Radu, V., Badalyan, R., & Chataigner, C. (2010). Production animale et économie de
subsistence au Néolithique dans la plaine de l'Ararat (Arménie). Annales d'Université "Valahia"
Târgovişte—Section d'Archéologie et d'Histoire , 12(1), 25–38.
Barge, O., & Brochier, J.É. (2011). Visible from space, understood during the fieldwork: the example of
“desert kites” in Armenia, Conference on cultural heritage and new technologies, November 2011,
Vienna. http://www.stadtarchaeologie.at/?page_id=3892. Accessed 14 April 2014.
Barge, O., Brochier, J.É., Chahoud, J., Chataigner, C., Chambrade, M.-L., Karakhanyan, A., Régagnon, E.,
Crassard, R. (2013). Towards a new approach to the ‘kites phenomenon’ in the Old World: the
GLOBALKITES Project. Antiquity, 87(338), Project Gallery. Available at http://antiquity.ac.uk/projgall/
barge338/. Accessed 14 April 2014
Bärmann, E. V., Börner, S., Erpenbeck, D., Rössner, G. E., Hebel, C., & Wörheide, G. (2013). The curious
case of Gazella arabica. Mammalian Biology, 78,220–225.
Baroin, C. (2006). La chasse et le statut des chasseurs au Sahara et en Arabie. In I. Sidéra, E. Vila, & P. Erikson
(Eds.), La chasse. Pratiques sociales et symboliques (pp. pp.87–pp.96). Paris: De Boccard.
Bar-Oz, G. (2004). Epipalaeolithic Subsistence Strategies in the Levant: A Zooarchaeological Perspective.
Boston: Brill Academic.
Bar-Oz, G., & Nadel, D. (2013). Worldwide large-scale trapping and hunting of ungulates in past societies.
Quaternary International, 297,1–7.
Crassard et al.

Bar-Oz, G., Zeder, M., & Hole, F. (2011). Role of mass-kill hunting strategies in the extirpation of Persian
gazelle (Gazella subgutturosa) in the northern Levant. PNAS, 108(18), 7345–7350.
Bar-Yosef, O., & Belfer-Cohen, A. (1989). The Origins of Sedentism and Farming Communities in the
Levant. Journal of World Prehistory, 3(4), 447–498.
Benedict, J. B. (2011). Tundra Game Drives: an Arctic-Alpine Comparison. Arctic, Antarctic, and Alpine
Research, 37(4), 425–434.
Berkes, F. (2008). Sacred Ecology (2nd ed.). New York: Routledge.
Betts, A. V. G. (Ed.). (1998). The Harra and the Hamad: excavations and surveys in eastern Jor dan. Vol. 1,
Sheffield Archaeological Monographs 9. Sheffield: Sheffield Academic Press.
Betts, A. V. G., & Helms, S. (1986). Rock art in eastern Jordan: kite carvings? Paléorient, 12(1), 67–72.
Betts, A. V. G., & Yagodin, V. (2000). Hunting traps on the Ustiurt plateau, Uzbekistan. In D. Christian & C.
Benjamin (Eds.), Realms of the Silk Roads: Ancient and Modern Silk Road Studies, 4 (pp. 29–45).
Turnhout: Brepols.
Bin Aqil, A.J. (2004). Kneis l wa’el fi Hadramut [Hunting Caribou in Hadramout]. Sana'a/Riyadh: King Fahd
National Library. [in Arabic]
Brink, J. W. (2013). The Barnett site: A stone drive lane communal pronghorn trap on the Alberta Plains,
Canada. Quaternary International, 297,24–35.
Brochier, J. É., Barge, O., Karakhanyan, A., Kalatarian, I., Chataigner, C., Chambrade, M.-L., & Magnin, F.
(2014). Kites on the margins: the Aragats kites in Armenia. Paléorient, 40(1), 25–53.
Brochier, J.É. (2002). Les sédiments anthropiques. Méthodes d’étude et perspectives. In J.C. Miskovsky (Ed),
Géologie de la Préhistoire: méthodes, techniques, applications (pp. 453-477). Paris: GÉOPRÉ éditions.
Brochier, J.É. (2014) Des cercles de pierre dans la steppe, usages et chronologie. In C. Castel, O. Barge, N.
Awad, & M. Al-Maqdissi (Eds.) Des villes neuves aux franges du désert de Syrie à la fin du 3
e
millénaire
: Travaux de la mission archéologique franco-syrienne d’Al-Rawda 2002-2008 ; Travaux de la mission
syrienne de Tell Sha’irat 2007-2008 et Tell Es Sour 2009, Travaux de la Maison de l’Orient. Lyon:
Maison de l’Orient (in press).
Brochier, J. É., & Thinon, M. (2003). Calcite crystals, starch grains aggregates or.POCC? Comment on ‘calcite
crystals inside archaeologial plant tissues’. Journal of Archaeological Science, 30(9), 1211–1214.
Brunner, U. (2008). Les pièges de chasse antiques au Yémen. Chroniques Yéménites, 15,29–34.
Bunke, H., & Riesen, K. (2011). Recent advances in graph-based pattern recognition with applications in
document analysis. Pattern Recognition, 44,1057–1067.
Burckhardt, J. L. (1831). Notes on the Bedouins and Wahabys, collected during his travels in the east.
Association for Promoting the Discovery of the Interior of Africa (Vol. 2). London: Colburn and Bently.
Chahoud, J., & Vila, E. (2011). The role of animals in ancient Sidon: an ove rview of ongo ing
zooarchaeological studies. Archaeology and History in the Lebanon, 34(35), 259–284.
Chapelle, J. (1957). Nomades noirs du Sahara. Paris: Plon (Re-edited: 1982, Paris: L’Harmattan).
Custred, G. (1979). Hunting technologies in Andean culture. Journal de la Société des Américanistes, 66,7–
19.
D’Hont, O. (1994). V ie quotidienne des ‘Agēdāt’. Techniques et occupation de l’espace sur le Moyen-
Euphrate. Damascus: IFPO
Davis, S. J. M. (1982). Climatic change and the advent of domestication the succession of ruminant
artiodactyls in the Late Pleistocene-Holocene in the Israel region. Paléorient, 8(2), 5–15.
Degen, D. E. (2010). A hunting scene from the Negev: the depiction of a desert kite and throwing weapon.
Israel Exploration Journal, 60(2), 146–165.
Deom, J., Sala, R. (2009). Ancient wild game management in arid regions: new evidences from the Northern
Ustyurt Plateau (Kazakhstan), In B.T. Zhanaev (Ed.) Nauchnye chteniya pamyati N.E. Masanova: sbornik
materialov nauchno-prakticheskoi konferentsii Almaty 25-26 Aprelya 2008 (pp. 135-141). Almaty: Daik-
Press. [Original in Russia n]. http://www.lgakz.org/Texts/LiveTexts/Ustyur t%20traps%202009.pdf.
Accessed 15 April 2014.
Desombre, S. (1946). La guémira : Mission Alger-Lac Tchad 1937. Paris: Lesourd.
Develle, A.-L., Gasse, F., Vidal, L., Williamson, D., Demory, F., Van Campo, E., Ghaleb, B., & Thouveny, N.
(2011). A 250ka sedimentary record from a small karstic lake in the Northern Levant (Yammoûneh,
Lebanon). Palaeogeography, Palaeoclimatolology, Palaeoecology, 305(1–4), 10–27.
Dong, Q., Yang, X., Zhao, J., & Tang, Y. Y. (2008). Embedding a family of disjoint 3D meshes into a crossed
cube.
Information Sciences, 178, 2396–2405.
Doughty, C.M. (1931). Arabia deserta. Textes choisis par Edward Garnett. London: Jonathan Cape (French
translation: 1990, Paris: Payot).
Dupré, G. (1976). La chasse au filet chez les Nzabi (République Populaire du Congo). Cahiers ORSTOM, 13
(4), 343–355.
Addressing the Desert Kites Phenomenon

Échallier, J.-C., & Braemer, F. (1995). Nature et fonction des “Desert Kites”: données et hypothèses nouvelles.
Paléorient, 21(1), 35–63.
Emms, D., Wilson, R. C., & Hancock, E. (2007). Graph embedding using quantum commute times. In F.
Escolano & M. Vento (Eds.), Graph-Based Representations in Pattern Recognition (pp. 371–382). Berlin:
Springer.
Fowden, G. (1999). ‘Desert kites’: ethnography, archaeology and art. In: Humphrey, J.H. (Ed.), The Roman
and Byzantine Near East: Some Recent Archaeological Research vol. 2. Journal of Roman Archaeology:
Supplementary Series, 31, 107-136
Gärtner, T., Flach, P., & Wrobel, S. (2003). On graph kernels: Hardness results and efficient alternatives. In B.
Schölkopf & M. K. Warmuth (Eds.), Learning Theory and Kernel Machines (pp. 129–143). Berlin:
Springer.
Gasparyan, B., Khechoyan, A., Bar-Oz, G., Malkinson, D., Nachmias, A., Nadel, D. (2013). The northernmost
kites in south-west Asia: the fringes of the Ararat Depression (Armenia) Project. Antiquity, 87(336),
Project Gallery. http://antiquity.ac.uk/projgall/bar-oz336/. Accessed 14 April 2014.
Gordon, B. C. (1990). World rangifer communal hunting. In L. Davis & B. Reeves (Eds.), Hunters of the
recent past (pp. 277–303). London: Routledge.
Gourichon, L. (2004). Faune et saisonnalité : l’organisation temporelle des activités de subsistance dans
l’Épipaléolithique et le Néolithique précéramique du Levant nord (Syrie). Unpublished PhD dissertation.
Lyon: Université Lyon 2. http://theses.univ-lyon2.fr/documents/lyon2/2004/gourichon_l#p=0&a=top.
Accessed 14 April 2014.
Gourichon, L., Helmer, D., & Peters, J. (2006). À la croisée des pratiques cynégétiques et de l’iconographie
des animaux sauvages. Haut et Moyen Euphrate X
e
et IX
e
millénaires av. In I. Sidéra, E. Vila, & P.
Erikson (Eds.), La chasse. Pratiques sociales et symboliques (pp. pp. 107–pp. 123). Paris: De Boccard.
Grigson, C. (1987). Shiqmim: pastoralism and other aspects of animal management in the Chalcolithic of the
northern Negev. In T. E. Levy & I. Shiqmim (Eds.), British Ar chaeological Report, International Series
356. Oxford: Archaeopress.
Grigson, C. (2006). Farming? Feasting? Herding? Lar ge mammals from the Chalcolithic of Gilat. In T. E.
Levy (Ed.), Archaeology, Anthropology and Cult. The Sanctuary at Gilat, Israel (pp. pp. 215–pp. 319).
London: Equinox.
Harding, G. L. (1953). The cairn of Hani. Annual of the Department of Antiquities of Jordan, 2,8–56.
Harrison, D. L. (1968). The Mammals of Arabia (Vol IIth ed.). London: Ernest Benn Ltd.
Haussler, D. (1999). Convolution kernels on discrete structures. UCSC-CRL-99-10. Technical report, UC
Santa Cruz. Santa Cruz, CA: University of California.
Helmer, D., Gourichon, L., & Stordeur, D. (2004). À l’aube de la domestication animale. Imaginaire et
symbolisme animal dans les premières sociétés néolithiques du Nord du Proche- Orient.
Anthropozoologica, 39(1), 143–163.
Helmer, D., Roitel, V., Saña Segui, M., Willcox, G. (1998). Interprétations environnementales des données
archéozoologiques et archéobotaniques en Syrie du Nord de 16000 BP à 7000 BP, et les débuts de la
domestication des plantes et des animaux. In M. Fortin, O. Aurenche (Eds.), Espace naturel,
espace habité en Syrie du Nord (10
e
-2
e
millénaires av. J-C.) (pp. 9-33). Canadian Society for
Mesopotamian Studies, Bulletin 33, Travaux de l a Maison de l'Orient 28, Lyon: Maison de
l'Orient, Toronto: CSMS.
Helms, S., & Betts, A. V. G. (1987). The Desert "Kites" of the Badiyat Esh-Sham and North Arabia.
Paléorient, 13(1), 41–67.
Hershkovitz, I., Ben-David, Y., Arensburg, B., Goren, A., & Pinchasov, A. (1987). Rock engravings in
southern Sinai. In G. Gvirtzman, A. Shmueli, Y. Gardos, I. Beit-Arieh, & M. Harel (Eds.), Sinai Part 2,
Human Geography / Gvirtzman Gdaliahu (pp. 605–616). Givatayim: Eretz Ha-Tzvi.
Holzer, A., Avner, U., Porat, N., & Horwitz, L. K. (2010). Desert kites in the Negev desert and northeast Sinai:
Their function, chronology and ecology. Journal of Arid Envir onments, 74(7), 806–817.
Horwitz, L. K., Tchernov, E., Mienis, H. K., Hakker-Orion, D., & Bar-Yosef Mayer, D. E. (2002). The
archaeozoology of three Early Bronze Age sites in Nahal Besor, northwestern Negev. In E. C. M. van den
Brink & E. Yanai (Eds.), In quest of ancient settlements and landscapes. Archaeological studies in honour
of Ram Gophna (pp. pp. 107–pp. 133). Tel Aviv: Ramot Publishing.
Horwitz, L. K., Tchernov, E., & Mienis, H. K. (2001). Archaeozoology and archaeomalacology of site 917 in
the ‘Uvda Valley. Atiqot, 42,121–127.
Hoyland, R. G. (2001). Arabia and the Arabs: from the Bronze Age to the coming of Islam.London:
Routledge.
Ingold, T. (1980). Hunters, Pastoralists and Ranchers. Reindeer economies and their transformations.
Cambridge, UK: Cambridge University Press.
Crassard et al.

Issar, A. S., & Zohar, M. (2007). Climate change: environment – Environment and History of the Near East
(2nd ed.). New York. Berlin Heidelberg: Springer-Verlag.
Jabbur, J.S., & Conrad, L.I. (1995). The bedouins and the desert: aspects of nomad life in the Arab East.New
York: Sunny Press, State University of New York Press. [Translated from Jabbur J.S. 1988 ‘ ﻜﺘاﺐابﺪو
واباﺪﻳﺔ ’].
Kempe, S., & Al-Malabeh, A. (2013). Desert kites in Jordan and Saudi Arabia: Structure, statistics and
function, a Google Earth study. Quaternary International, 297, 126–146.
Kennedy, D. (2012). Kites- new discoveries and new type. Arabian archaeology and epigraphy, 23, 145–155.
Kennedy, D., & Bishop, M. C. (2011). Google Earth and the archaeology of Saudi Arabia. A case study from
the Jeddah area. Journal of Archaeological Science, 38(6), 1284–1293.
Kingswood, S. C., & Blank, D. A. (1996). Gazella subgutturosa. Mammalian Species, 518,1–10.
Kobusiewicz, M. (1999). Excavations at Sinai-10, the Kite Site, Romythi locality. In F. Eddy & F. W endorf
(Eds.), An archaelogical investigation of the Central Sinai (pp. 173–180). Boulder: University Press of
Colorado.
Lagraa, S., Seba, H., Khennoufa, R., M’Baya, A., & Kheddouci, H. (2014). A distance measure for large
graphs based on prime graphs. Pattern Recognition, 47, 2993–3005.
Le Quellec, J.-L., & Civrac, M.-A. (2010). La chasse au filet sur les peintures rupestres du Sahara central
et dans l’Antiquité. Cahier de l’AARS (Association des Amis de l'Art Rupestre Saharien), 14, 255–
262.
Legge, A. J., & Rowley-Conwy, P. A. (1987). Gazelle killing in Stone Age Syria. Scientific American, 255(8),
88–95.
Legge, A. J., & Rowley-Conwy, P. A. (2000). The exploitation of animals. In A. M. T. Moore, G. C. Hillman,
&A.J.Legge(Eds.),Village on the Euphrates (pp. 423–471). Oxford: Oxford University Press.
Lerp, H., Wronski, T., Butynski, T. M., & Plath, M. (2013). Speciation of Arabian gazelles. In P. Michalak
(Ed.), Speciation: Natural Processes, Genetics and Biodiversity (pp. 59–82). New York: Nova Science.
MacDonald, M.C. (2005). Of Rock-art, ‘Desert Kites’ and Mesayid. In A.V. Sedov, I.M. Smulyanskaya
(Eds.), Arabi a Vita lis: Arabskij Vostok, I slam, drevnyaya, Aravi ya: Sbornik Naychnykh statej,
posvyashchennyj 60-letiyu V.V. Naumkina (pp. 332-345). Moscow: Rossijskaya Akademiya Nauk
Maitland, R. A. (1927). The ‘works of the old man’ in Arabia. Antiquity, 1,197–203.
Martin, L. (1998). The faunal remains from Dhuweila. In A. V. G. Betts (Ed.), Harra and the Hammad:
Archaeological Explorations in the Jordanian Basalt Desert (pp. 159–184). Sheffield: Sheffield
University Press.
Martin, L. (2000a). Gazelle (Gazella sp.) behavioural ecology: predicting animal behaviour for prehistoric
environments in southwest Asia. Journal of Zoology, 250,13–30.
Martin, L. (2000b). Mammal remains from the eastern Jordanian Neolithic, and the nature of caprine herding
in the steppe. Paléorient, 25(2), 87–104.
Martin, L., McCorriston, J., Crassard, R. (2009). Early Arabian pastoralism at Manayzah in Wadi Sana,
Hadramawt. Proceedings of the Seminar for Arabian Studies, 39, 271–282.
Mashkour, M., & Vila, E. (2003). Archaeometrical approaches for tracking ancient Bedouins; a pilot project in
northern Mesopotamia from 4th to 1st millenniaB.C.InT.Herzog,W.Holzwarth(Eds.),
Orientwissenschaftliche Hefte. Mitteilungen des SFB “Differenz und Integration” 4/1/ Nomaden und
Sesshafte - Fragen, Methoden, Ergebnisse (pp. 1-20). Halle (Saale): Orientwissenschatliche Zentrum der
Martin-Luther-Universität Halle-Wittenberg, Heft 9.
Migowski, C., Stein, M., Prasad, S., Negendank, J. F. W., & Agnon, A. (2006). Holocene climate variability
and cultural evolution in the Near East from the Dead Sea sedimentary record. Quaternary Research, 66,
42–431.
Monks, G. G. (1981). Seasonality studies. In M. B. Schiffer (Ed.), Advances in archaeological method and
theory (pp. 177–240). New York: Academic Press.
Morandi Bonacossi, D., & Iamoni, M. (2012). The Early History of the Western Palmyra Desert Region The
Change in the Settlement Patterns and the Adaptation of Subsistence Strategies to Encroaching Aridity: A
First Assessment of the Desert-kite and Tumulus Cultural Horizons. Syria, 89,31–58.
Musil, A. (1928). The manners and customs of the Rwala Bedouins. Oriental Explorations and Studies 6.New
York: American Geographical Society.
Nadel, D., Bar-Oz, G., Avner, U., Boaretto, E., & Malkinson, D. (2010). Walls, ramps and pits: the
construction of the Samar Desert kites, southern Negev, Israel. Antiquity, 84(326), 976–992.
Nadel, D., Bar-Oz, G., Avner, U., & Malkinson, D. (2013). Ramparts instead of walls: building techniques of
large game traps (kites) in the Negev Highland. Quaternary International, 297,147–154.
Perevolotsky, A., & Baharav, D. (1991). The distribution of desert kites in eastern Sinai and sub-regional
carrying capacity: an ecological perspective. Journal of Arid Envir onments, 20,239–249.
Addressing the Desert Kites Phenomenon

Peters, J., Driesch, A. von den, Helmer, D. (2005). The Upper Euphrates-Tigris Basin: Cradle of agro-
pastoralism? In J.-D. Vigne, J. Peters, D. Helmer (Eds.) The first steps of animal domestication,
Proceedings of the 9th International Council of Archaeozoology (ICAZ) Conference (Durham, 2002)
(pp. 96-123). Oxford: Oxbow Books.
Picalause, P., Cauwe, N., Lemaitre, S., Vander Linden, M., & Van Berg, P. L. (2004). Desert-kites of the
Hemma plateau (Hassake, Syria). Paléorient, 30(1), 89–99.
Quenet, P., & Chambrade, M.-L. (2013). Nouveaux kites en Jezireh syrienne. Studia Or ontica, XI, 61-76.
Reed, C. A. (1965). Imperial Sassanian hunting of pig and fallow deer, and problems of the survival of these
animals today in Iran. Postilla, 92,1–23.
Rees, L. W. B. (1929). The Trans-Jordan deserts. Antiquity, 3,389–407.
Riemer, H. (2009). Prehistoric trap hunting in the eastern Sahara deserts: A re-evaluation of the game trap
structures. In H. Riemer, F. Förster, N. Pöllath, & M. Herb (Eds.), Desert animals in the eastern Sahara:
Status, economic significance, and cultural reflection in antiquity (pp. 175–188). Köln: Heinrich-Barth-
Institut.
Robinson, S. A., Black, S., Sellwood, B. W., & Valdes, P. J. (2006). A review of palaeoclimates and
palaeoenvironments in the Levant and Eastern Mediterranean from 25,000 to 5000 years BP: setting
the environmental background for the evolution of human civilization. Quaternary Science Reviews, 25,
1517–1541.
Rodier, X., Barge, O., Saligny, L., Nuninger, L., & Bertoncello, F. (2011). Information spatiale et archéologie.
Paris: Errance.
Sapir-Hen, L., Bar-Oz, G., Khalaily, H., & Dayan, T. (2009). Gazelle exploitation in the early Neolithic site of
Motza, Israel: the last of the gazelle hunters in the southern Levant. Journal of Archaeological Science,
36(7), 1538–1546.
Sergeant, R. B. (1976). South Arabian hunt. London: Luzac.
Sinclair,W.F.,&Fergusson,D.(1902).The travels of Pedro Teixeira: with his “Kings of Harmuz” and
extracts from his “Kings of Persia”. London: Hakluyt Society.
Skorupka, M. (2010). Les “desert kites” yéménites, une relecture critique des données. Chroniques Yéménites,
16,5–14.
Smith, B. D. (2013). Modifying landscapes and mass kills: Human niche construction andcommunal ungulate
harvests. Quaternary International, 297,8–12.
Smol, J. P. (1985). The ratio diatom frustules to chrysophycean statospores: a useful palaeolimnological index.
Hydrobiologia, 123,199–204.
Speth, J. D. (2013). Middle Paleolithic large-mammal hunting in the southern Levant. In J. L. Clarck & J. D.
Speth (Eds.), Zooarchaeology and Modern Human origins: Human hunting behavior during the Later
Pleistocene (pp. 19–43). Dordrecht: Springer.
St Simpson, J. (1994). Gazelle-hunters and Salt-collectors: A Further Note on the Solubba. Bulletin of the
American School of Oriental Research, 293,79–81.
Storemyr, P. (2011). The ancient stone-built game traps at Gharb Aswan and beyond, Lower Nubia and Upper
Egypt. Sahara, 22
,15–28.
Tsahar, E., Izhaki, I., Lev-Yadun, S., & Bar-Oz, G. (2009). Distribution and extinction of ungulates during the
Holocene of the southern Levant. PLoS ONE, 4(4), e5316.
Uerpmann, M. (2008). Neolithic Faunal Remains from al-Buhais 18 (Sharjah, UAE). In H.-P. Uerpmann & S.
Jasim (Eds.), The Natural Environment of Jebel al-Buhais: Past and Present. The Archaeology of Jebel
al-Buhais, Sharjah, United Arab Emirates (2nd ed., pp. (pp. 97–132). Tübingen: Kerns Verlag.
Ur, J., & Hammer, E. (2009). Pastoral nomads of the 2nd and 3rd millennia A.D. on the Upper Tigris River,
Turkey: the Hirbemerdon Tepe Survey. Journal of Field Archaeology, 34(1), 37–56.
van de Sande, K. E. A., Gevers, T., & Snoek, C. G. M. (2010). Evaluating Color Descriptors for Object and
Scene Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 32(9), 1582–1596.
Verheyden, S., Nader, F. H., Cheng, H. J., Edwards, L. R., & Swennen, R. (2008). Paleoclimate reconstruction
in the Levant region from the geochemistry of a Holocene stalagmite from the Jeita cave, Lebanon.
Quaternary Research, 70, 368–381.
Vila, E. (1991). Note sur un dépôt de gazelles dans une fosse uruk d'El Kowm 2-Caracol (Syrie). Cahiers de
l'Euphrate, 8,55–60.
Vila, E. (1998). L’exploitation des animaux en Mésopotami e aux IV
e
et III
e
millénaires avant J.- C.
Monographie du CRA 21. Paris: CNRS Éditions.
Vila, E. (2004). Apport de l’étude de la faune à la connaissance des modes de vie. In F. Braemer, J.-C.
Échallier, & A. Taraqji (Eds.), Khirbet al Umbashi, Villages et campements de pasteurs dans le “désert
noir” (Syrie) à l’âge du Br onze (pp. 271–288). Beirut: IFPO.
Crassard et al.

Vila, E. (2006). Données archéozoologiques de la période Halaf à l’Âge du Fer. In B. Lion & C. Michel (Eds.),
De la domestication au tabou: le cas des suidés au Proche-Orient ancien (pp. 137–153). Paris: De
Boccard.
Vila, E., & Al Besso, M. (2014). L’homme et la steppe au Bronze ancien IV. Étude archéozoologique. In C.
Castel, O. Barge, N. Awad, & M. Al-Maqdissi (Eds.), Des villes neuves aux franges du désert de Syrie à
la fin du 3
e
millénaire : Travaux de la mission archéologique franco-syrienne d’Al-Rawda 2002-2008 ;
Travaux de la mission syrienne de Tell Sha’irat 2007-2008 et Tell Es Sour 2009, Travaux de la Maison de
l’Orient. Lyon: Maison de l’Orient (in press).
Vila E., & Helmer, D. (2014). The expansion of sheep herding and the development of wool production in the
ancient Near East: an archaeozoological and iconographical approach. In C. Bréniquet, C. Michel (Eds.),
Wool economy in the ancient Near East and the Aegean: from the beginnings of sheep husbandry to
institutional textile industry (pp. 22–40). Oxford: Oxbow Books.
Wright,W.(1895).An account of Palmyra and Zenobia with travels and adventures in Bashan and the desert.
London: T. Nelson Edition.
Zeder, M. A., Bar-Oz, G., Rufolo, S., & Hole, F. (2013). New perspectives on the use of kites in mass-kills of
Levantine gazelle: a view from northeastern Syria. Quaternary International, 297,110–125.
Addressing the Desert Kites Phenomenon