Content uploaded by Arlene M. Rosen
Author content
All content in this area was uploaded by Arlene M. Rosen on Jan 14, 2015
Content may be subject to copyright.
393
Diversity and Complexity in Early Farming
Communities of Southwest Asia: New
Insights into the Economic and
Environmental Basis of Neolithic
C¸ atalho¨yu¨k
Neil Roberts and Arlene Rosen
School of Geography, University of Plymouth, Drake
Circus, Plymouth PL4 8AA, United Kingdom (neil
.roberts@plymouth.ac.uk)/Institute of Archaeology,
University College London, 31–34 Gordon Square,
London WC1H OPY, United Kingdom. 24 X 08
CA⫹ Online-Only Material: Supplement A
Intensive but localized cultivation of cereal crops on alluvial
wetlands is thought to have provided the ecological basis for
the primary Neolithic settlement that spread across southwest
Asia and southeast Europe. New excavations at C¸ atalho¨yu¨k
provide an opportunity to test this horticultural model via
multiple data sources. Geoarchaeological studies show that
the main Neolithic occupation coincided with a period of
active river alluviation. Most of the area surrounding C¸ atal-
ho¨yu¨k was therefore under floodwaters each spring that would
have seriously damaged any autumn-sown cereal crops. In-
dependent evidence that the cereal crops consumed at C¸a-
talho¨yu¨k were grown under rain-fed conditions derives from
the paucity of silicified multicellular wheat phytoliths at the
site. Together, these and other data suggest that the bulk of
the cereal agriculture was not carried out in the immediate
vicinity of C¸ atalho¨yu¨k but was at least 13 km and 3 h away
in dryland soils. In turn, this implies that there may have
been seasonal fission and fusion of C¸ atalho¨yu¨k’s population,
with systematic exploitation of, and impact on, a range of
different ecological zones. This phase of nucleated settlement
ended when river flooding ceased, coincident with a multi-
decadal drought from 6300 to 6140 BC.
It has been widely accepted that Neolithic farmers in south-
west Asia employed a standardized strategy calculated to min-
imize economic risk and maximize returns in the form of
high crop yields. This strategy was centered on the spatially
intensive exploitation of moist alluvial soils for cereal and
pulse crops that were grown outside their natural dryland
habitats (Sherratt 1980; Roberts 1991). The success of this
economic package led to it being replicated across southeast
Europe and is believed to have underpinned the rapid spread
of Neolithic settlement from Anatolia through southeast Eu-
䉷 2009 by The Wenner-Gren Foundation for Anthropological Research.
All rights reserved. 0011-3204/2009/5003-0006$10.00. DOI: 10.1086/
598606
rope to the Great Hungarian Plain (van Andel and Runnels
1995; Colledge, Conolly, and Shennan 2004). In this model,
subsistence farming activities provided the primary criteria in
deciding village locations, and those Neolithic sites found in
nonalluvial settings are considered deviations from the norm.
As we obtain more and better archaeobotanical and ar-
chaeozoological evidence from Neolithic sites in this region,
there are increasing indications that Neolithic diet and sub-
sistence strategies included a wider range of foodstuffs than
those provided by core domesticates. For example, Neolithic
peoples of the Jordan Valley exploited wild grass seeds, tubers,
and a range of wild animals (Bar-Yosef et al. 1991; Colledge
2001). This broader range of primary resource exploitation
takes the focus off farming as a sole provider of subsistence
goods, relegating it to being one of the many sources of food,
albeit an important contributor. Diversity consequently re-
duces the risk associated with primary crop failure. If farming
was only one of several subsistence strategies, then, in turn,
it would not have been the sole factor influencing choice of
site locations. Rather than always privileging direct proximity
to prime agricultural land, other factors such as such as trade,
ease of transport, defense, or even cosmology may have dic-
tated primary site locations.
Exploring these alternative bases for Neolithic site econo-
mies requires us to employ multiple sources of high-quality
evidence. The multidisciplinary investigation of C¸ atalho¨yu¨k
since 1993 (Hodder 1996, 2007), along with the associated
KOPAL (Konya Basin paleoenvironments) project (Roberts
et al. 1999), provides an ideal “laboratory” for such an ex-
ploration. Neolithic C¸ atalho¨yu¨k (7400–6200 BC; all dates in
calendar years) is important both in terms of its record of
symbolic expression and cultural complexity and as one of
the best-studied of the first wave of agricultural settlements
to appear in southwest Asia and the Balkans.
In order to examine the nuances of the relationship between
Neolithic subsistence economy and site selection, it is im-
portant to find a direct link between a particular archaeo-
logical site, the remains of its subsistence resources, and its
immediate landscape setting. Geoarchaeological research pro-
vides us with this essential link by providing a reconstruction
of the landforms, soil resources, and hydrology within the
immediate vicinity of the site that can be tied directly to the
excavated on-site evidence for subsistence (Goldberg and
Macphail 2006). The landscape as it exists today is often very
different from the landscape and environmental setting as it
was in the past, and this can all too often be disregarded by
researchers investigating Neolithic lifeways. C¸ atalho¨yu¨k is no
exception, as some past discussions of economy at C¸ atalho¨yu¨k
have assumed that the environment and landscape setting was
essentially similar to that of today (Cohen 1970; Fairbairn
2005) and that subsistence farming in alluvial soils adjacent
to the site provided the economic basis for this large, populous
community (Todd 1976).
Here we combine a geoarchaeological reconstruction of
394 Current Anthropology Volume 50, Number 3, June 2009
Figure 1. Geomorphological map of the western Konya Plain, Turkey,
including the C¸ ars¸amba fan and the regional site survey area. Map shows
the extent of C¸ ars¸amba alluvium at the end of the Neolithic occupation
of C¸ atalho¨yu¨k, ∼6200 BC.
past landscape and environment at C¸ atalho¨yu¨k with on-site
evidence for farming strategies, particularly as indicated in
the phytolith record at the site. Joining these independent
data sets offers an alternative explanation for the site’s eco-
nomic, ecological, and social basis.
Geoarchaeology and Alluvial Stratigraphy
C¸ atalho¨yu¨k is located on the C¸ ars¸amba River alluvial fan in
the Konya Plain of central Anatolia (fig. 1). Alluvial landscapes
are inherently dynamic as a result of river alluviation and
incision and lateral channel migration (Brown 1997), and
systematic vibro-coring, off-site backhoe trenching, and
cleaning of irrigation ditch sections have shown significant
alluviation to have taken place around C¸ atalho¨yu¨k (Roberts
1982; Roberts, Boyer, and Parish 1996, Roberts, Boyer, and
Merrick 2007). This means that modern distributions of soil
types (Driessen and de Meester 1969) do not provide a reliable
guide to those that existed during the Neolithic and Chal-
colithic (cf. French 1970; Fairbairn 2005).
Two principal alluvial units can be distinguished that were
deposited above pale gray marl representing the bed of a large
lake that occupied the Konya Plain during the last glacial
period. The Upper Alluvium dates from the Bronze Age to
post-Byzantine times (Boyer, Roberts, and Baird 2006), and
the stratum at C¸ atalho¨yu¨k is 1.5–2.0 m thick. Beneath this
lies a Lower Alluvium stratum comprising very dark gray,
heavy clay that was laid down in a standing-water backswamp
environment during seasonal flooding. From cores taken
through the Neolithic (east) mound (Roberts Boyer, and Par-
ish 1996) and from a deep sounding excavated to “natural
deposits” in 1999 (Farid 2007), we know that this backswamp
clay began to form just before the foundation of Neolithic
C¸ atalho¨yu¨k. The C¸ ars¸amba River, having breached an east-
west-trending sand ridge, started to prograde across the marl
plain during the early Holocene.
Alluviation and flooding continued throughout the Neo-
lithic occupation of the site. This is well revealed in a 4-m-
deep KOPAL trench that was mechanically and then hand
excavated over a basal area of m ∼20 m from the
8m# 8
northern edge of the Neolithic mound (Roberts, Boyer, and
Merrick 2007). This trench uncovered a sequence beginning
at the base with Pleistocene lake marl—into which were cut
shallow circular pits—followed by a ∼50-cm-thick unit of
Neolithic cultural refuse mixed with flood sediments. The
Neolithic refuse layer was overlain by ∼80 cm of backswamp
clay, in turn capped by a weakly developed buried soil com-
395
Figure 2. Key stratigraphic sequences at and adjacent to C¸ atalho¨yu¨k. a,
Topographic cross-profile, showing section/core locations. b, Detailed
sedimentary profiles with
14
C-dated units (calibrated, multiple age de-
terminations; see CA⫹ online supplement table A1 for details). 1 p
Pleistocene lake marl, 2 p black organic marsh clay, 3 p lower alluvial
backswamp clay, 4 p Neolithic cultural fill, 5 p river sands and gravels,
6 p buried soil complex, 7 p upper alluvium (mid-late Holocene).
plex (fig. 2b). The latter, dated to 5800–6200 BC (CA⫹ online
supplement table A1), marks the end of alluvial deposition
and flooding and the establishment of a stable ground surface
around C¸ atalho¨yu¨k. Analysis and topographic survey of mul-
tiple cores and sections around C¸ atalho¨yu¨k has shown that
the thickness of the Lower Alluvium and the elevation of its
basal contact with the marl varies by over a vertical meter
across distances of hundreds of meters. In other words, the
surface on which alluviation had commenced was not hori-
zontal, like the modern ground surface, but undulating, pri-
marily as a result of wind deflation of the top of the marl as
the glacial-age lake shrank (Roberts et al. 1999; Kuzucuoglu,
Parish, and Karabiyikoglu 1998). Early Holocene alluviation
infilled the lowest-lying terrain, which had been occupied by
perennial marshes, so that topographic irregularities even-
tually became smoothed out.
Evidence indicates that a main channel of the C¸ars¸amba
River flowed next to C¸ atalho¨yu¨k. Core 95E, taken off the
396 Current Anthropology Volume 50, Number 3, June 2009
western edge of the east mound, contains coarse-grained
channel deposits interbedded with Neolithic cultural fill (fig.
2b), and a clear paleochannel of the river, 40–50 m wide and
2.5 m deep, was exposed in a cleaned irrigation ditch section
∼500 m south of the site (see Roberts, Boyer, and Parish 1996
for details). This PC1 paleochannel section and other se-
quences are notable for the absence of any identifiable levee-
type deposits. Even if present, levees cannot have formed a
very significant part of the Neolithic landscape or agricultural
focus around C¸ atalho¨yu¨k.
The Neolithic Site Environment
C¸ atalho¨yu¨k was thus founded next to a branch of the C¸ar-
s¸amba River, surrounded by a gently undulating landscape of
marshy flood basins and intervening natural marl hummocks.
It was one of several settlements, including Boncuklu (now
under excavation; Baird 2007), on the plain during the ace-
ramic Neolithic (8000–7000 BC). These first Neolithic people
encountered a complex mosaic of depositional environments
and vegetation associations. The presence of marsh localities
and shallow, seasonal-perennial lakes is also implied by some
of the bird bones at C¸ atalho¨yu¨k (Russell and McGowan 2005).
The vegetation across the plain would have included a com-
plex assortment of grassy steppe, sedges, and Phragmites
within the marshy areas, shrubs on the better-drained soils,
and riparian trees along the river margins (Asouti 2005; Rosen
and Roberts 2006). Just as the low-lying basins would have
been inundated longest by floodwaters, so topographic
“highs” would have been least flood prone. It is on marl
hummocks that C¸ atalho¨yu¨k and other Neolithic sites on the
C¸ ars¸amba fan were most likely situated. Progressive accu-
mulation of collapsed mud-brick building layers in these nu-
cleated settlements then raised them further above the zone
most at risk from flood inundation.
During the ceramic Neolithic period (7000–6200 BC), set-
tlement on the C¸ ars¸amba fan was concentrated at a single
large site—C¸ atalho¨yu¨k—and it is highly unlikely that any
other significant population centers existed within 30 km
(Baird 2005). The nature of the backswamp clay indicates that
seasonal flooding at this time was more severe and the land-
scape wetter than it was on the modern Konya Plain before
recent river regulation. This may be linked to the wetter cli-
mate that prevailed in central Anatolia during the early Ho-
locene, with precipitation around 25% higher than today
(Jones, Roberts, and Leng 2007). Notwithstanding any cli-
matic changes, the modern river hydrology offers a guide to
the flood regime that would have prevailed during the Ne-
olithic. Twentieth-century irrigation schemes involve inter-
basin water transfers from lakes Beysehir and Sugla and have
substantially altered the flow regime of the C¸ars¸amba River
in its lower reaches. A record of unregulated flood hydrology
is instead provided by flow data from the river upstream at
Bozkir (fig. 1). The Taurus Mountain region above Bozkir
receives high precipitation, much of it as snowfall, whose
melting gives rise to the main annual river flood peak in
March and April (fig. 3). After the spring flood peak, the
summer months (July-September) are marked by low flows,
while the autumn and winter are characterized by higher base
flows on which are superimposed periodic flood events fol-
lowing heavy rainfall. These are smaller in scale and much
shorter in duration than the snowmelt-generated late spring
flood (Rosen and Roberts 2006). It seems highly likely that
the same annual river cycle also operated during Neolithic
times, although the timing of these phases around C¸ atalho¨yu¨k
may have differed by a few weeks.
Human Ecology
During the main 2-month flood period, most of the landscape
surrounding C¸ atalho¨yu¨k would have been inundated and the
settlement mound made virtually an island. This would have
had highly significant consequences for human ecology, in-
cluding the seasonal cycle of food procurement activities. On
the one hand, seasonal wetlands offered rich and varied nat-
ural resources, including wildfowl, summer grazing for equids
and bovids, and club-rush tubers. On the other hand, some
domesticated animals and crops would have been less well
suited to year-round activity in the immediate vicinity of
C¸ atalho¨yu¨k. It is difficult to envisage how herds of sheep and
goat, for example, could have been provided with adequate
grazing and browsing during the spring flood period, and
neither animal adapts easily to having wet feet. We therefore
infer that sheep and goat, the most common domestic live-
stock at the site (Russell and Martin 2005), must have been
herded on drier ground away from C¸ atalho¨yu¨k during much
of the spring season, with perhaps limited penning of some
animals on-site for food, lambing, and dung production. Iso-
topic analysis of caprine bone from C¸ atalho¨yu¨k shows that a
wide array of different habitats were used for grazing/browsing
during the year from ∼7100 BC onward, consistent with
movement of animals across an extensive landscape range
(Pearson et al. 2007).
There are similar potential implications for crop cultiva-
tion, particularly of wheat and barley, both of which are abun-
dantly represented in the charred plant and phytolith remains
from Neolithic C¸ atalho¨yu¨k (Fairbairn et al. 2002; Fairbairn,
Near, and Martinoli 2005b; Rosen 2005). Autumn-sown crops
would not have been able to mature toward ripening in the
alluvial backswamp soils if they were standing in water for
several weeks during the spring flood. Growing these crops
on marl hummocks would have been hazardous depending
on the flood height. As figure 3 shows, flood discharge varied
greatly between years, so that a crop growing at the upper
limit of flood waters in one year would be completely drowned
the next. In any case, marl has much lower soil fertility (nu-
trient status, organic matter) than river alluvium (Driessen
and de Meester 1969). Marl soils are abundant across the
Konya Plain and yet have consistently been avoided for ag-
ricultural settlement from Neolithic until recent times. As
397
Figure 3. River flood regime annotated to show possible fission and fusion
of C¸ atalho¨yu¨k’s population, along with seasonal activities located close
to and more distant from the site. Lines show maximum, mean, and
minimum monthly water flows of the C¸ ars¸amba River at Bozkir for the
period 1964–1980.
backswamp basins gradually infilled, so marl hummocks grad-
ually became smaller in extent, and eventually they would
have been covered entirely by river alluvium and spring flood-
waters before the end of C¸ atalho¨yu¨k’s east mound occupation.
The only reliable means of growing cereal crops locally would
have been via late spring sowing. Such flood-recession farming
is well known in other parts of the world (e.g., Bryan 1929),
but it would have required the photoperiodic response of
cereals to have changed their natural timing from autumn to
spring germination soon after initial domestication (Fair-
bairn, Near, and Martinoli 2005b). There is as yet no direct
evidence for spring-sown cereals at this early time period, but
the possibility of such sowing practices has been suggested as
a special adaptation that allowed cereals to spread into Eu-
ropean temperate habitats from the Near East (Barker 1985,
146; Bogucki 1996). Bogaard (2004, 163–164) argues com-
pellingly against this suggestion. She cites instances where the
weedy flora was used to estimate spring sowing (cf. Gluza
1983; Bakels and Rouselle 1985) but proposes that rather than
indicators of sowing season, the weeds are simply reflecting
more intensive cultivation. Furthermore, spring-sown cereals
are typically planted in February–March rather than in
May–June as would have been required at C¸ atalho¨yu¨k.
An alternative explanation is that the bulk of the cereal
crops were not grown at C¸ atalho¨yu¨k but merely stored and
consumed there, having been grown farther afield. There is
increasing evidence from macrobotanical remains as well as
phytoliths (Rosen 2005) to support this latter argument,
namely that most cereal cultivation took place on well-drained
soils at slightly higher elevations away from the flood-prone
areas of the plain. Analysis of the seed flora by Fairbairn et
al. (2002) showed no evidence for the use of damper alluvial
soils for arable fields, with only dryland weed seeds having
significant statistical correlations with crop products (CA⫹
online supplement table A2). This is despite the fact that
wetland species are very abundant, with Scirpus maritimus
being the most abundant seed found on the site (A. Fairbairn,
personal communication). The nearest substantial extent of
nonalluvial dryland soils lies at least 13 km south of C¸ atal-
ho¨yu¨k on low-relief Neogene terraces (fig. 1). A smaller area
of dry sand ridge soils lies somewhat nearer to C¸ atalho¨yu¨k
(6 km) but their free-draining character and low nutrient
status would have made their potential crop yields unreliable
in below-average rainfall years, and it is difficult to see them
playing a key role in cereal cultivation.
Phytolith Evidence
Phytolith formation in cereals and other monocotyledons
such as reeds and sedges is influenced by soil and water con-
ditions during the growth period. Thus, the growth environ-
ment can influence the number and types of plant cells that
398 Current Anthropology Volume 50, Number 3, June 2009
Figure 4. Photograph of large multicelled phytoliths from club-
rush (Scirpus sp.) basket remains.
become silicified, allowing researchers to distinguish between
cereals that were cultivated in moist alluvial fields and those
that were grown through rain-fed cultivation in lighter, well-
drained soils. This is related to two main factors: (1) the level
of evapotranspiration, which controls the extent of passive
silica uptake in grasses and sedges, and (2) the amount of
dissolved silica that is available to the plants. Plants acquire
silica from the soil both actively, as a part of their physiological
programming, and passively. The actively acquired silica is
deposited within plant hairs and specialized individual silica
cells. The passively deposited silica can be an environmental
indicator since its uptake is significantly greater under con-
ditions of high-evapotranspiration growth in clay-rich soils
and a continuous supply of soil water for much of the growth
period (Parry and Smithson 1964; Hutton and Norrish 1974;
Hayward and Parry 1980). Under conditions of accelerated
silica uptake, plants form very large multicelled phytoliths
that can consist of hundreds of adjacent cells. These are es-
sentially fossilized sections of epidermal tissue. Experimental
data from the semiarid region of the northern Negev Desert
in southern Israel have shown that irrigated crops of emmer
(Triticum dicoccum) produce silica skeletons that commonly
exceed 100 adjacent cells per phytolith and sometimes pro-
duce forms with as many as 300 joined cells. Conversely,
wheat grown in dry farming conditions depending primarily
on rain-fed agriculture produce many fewer multicelled silica
skeletons (for more detailed information on the experimental
work, see Rosen and Weiner 1994; Rosen 1999).
Results of the geoarchaeological investigations outlined
above indicate the environmental conditions that existed
around the site of C¸ atalho¨yu¨k would have constituted the
ideal situation for the production of large multicelled phy-
toliths. This included very clay-rich sediments, the presence
of standing water for many weeks lasting into the warm spring
growth period, and relatively high evapotranspiration rates in
the later period of the growing season (see Rosen 1999). In-
deed, remains from the site show that plants that grew in the
marshy settings around the site, such as Scirpus and Phrag-
mites, are characterized by exceedingly large multicelled phy-
toliths (Rosen 2005). This is clearly demonstrated by the re-
mains of baskets constructed from sedges that still exist as
complete outlines of the original basket preserved as a silica
skeleton (fig. 4). Given these conditions we would expect to
see large numbers of silicified cells from wheat phytoliths as
well, if cereals had been cultivated in the moist wetland soils
around C¸ atalho¨yu¨k.
Wheat-husk phytoliths from a range of different contexts
in Building 1 at C¸ atalho¨yu¨k were examined, and the numbers
of their cells were counted. The results from counts of 250
multicelled phytoliths showed that very few of the phytoliths
were large and multicellular, and there were no phytoliths
with more than 70 adjoined cells. The number with greater
than 10 cells per phytolith was less than 15% of the assem-
blage. This is equivalent to samples from the experimental
fields in Israel, which received an allotment of 450 mm of
water throughout the growing period (CA⫹ online supple-
ment table A3). This suggests that the wheat obtained greater
amounts of water throughout the growth period than it would
receive today under dry farming conditions in the vicinity of
C¸ atalho¨yu¨k but not enough to account for the amount of
water available if they were grown in heavy, silica-rich wetland
soils. Based on independent evidence (Jones, Roberts, and
Leng 2007), annual rain/snowfall in the dry-farming zone
nearest to C¸ atalho¨yu¨k would have been around 350 mm in
Neolithic times, and this would have been more effective than
the equivalent amount in Israel because of the cooler climate
and lower evapotranspiration in Anatolia.
Implications for Neolithic Economic and
Social Organization
If the people of C¸ atalho¨yu¨k were indeed conducting most of
their cereal farming activities on higher ground and in better-
drained soils, then their principal fields must have been lo-
cated at least 13 km and more than 3 h walking distance away
from the site. It seems unlikely that farmers would maintain
a permanent residence at such a distance from their fields
since during periods of sowing, weeding, and harvesting they
would be required to work intensively for long days in their
plots. There would also be the problem of guarding the fields
from birds, grazers, and any other people traveling across the
landscape. Having fields too far for a daily “commute” may
seem suboptimal, but it is not hard to find contemporary
examples of traditional agrarian societies where this is the
case. In the Paramo zone of the northern Andes, for example,
native farmers live in temporary grass-made Choza huts near
their fields but after harvest time return to their permanent
homes in villages that lie at lower altitudes.
The farming mode proposed here has many implications
for population movements and seasonal activities of C¸ atal-
ho¨yu¨k’s Neolithic society. It suggests that some of the pop-
399
ulation splintered into task groups throughout the year and
left a base population back in the main settlement, which bird
bones and other evidence demonstrate was occupied year-
round (Russell and McGowan 2005; Fairbairn et al. 2005a).
This splintered segment could have consisted of people whose
duties included tending the fields, moving sheep and goats
to take advantage of changes in pasture productivity through-
out the year, as well as task groups that foraged and hunted
in the upland areas around the site. These wild resources
ended up back at C¸ atalho¨yu¨k and included wild hackberries,
wood timber, acorns, deer, and so on (Asouti and Hather
2001; Fairbairn, Near, and Martinoli 2005b; Russell and Mar-
tin 2005). Additionally, many nonbiotic resources such as
ground stone and salt were exchanged and transported on a
systematic basis over distances of hundreds of kilometers to
reach C¸ atalho¨yu¨k. In the case of obsidian—which accounts
for almost all of the chipped stone tools at the site—this
involved movement of material mainly from the Go¨llu¨ Dag
and Nenezi Dag sources in Cappadocia that lie 160 km away
(Cessford and Carter 2005). This implies a high degree of
seasonal mobility among at least some human groups. These
seasonal activities would have required only temporary shel-
ters to be constructed, which need not have left a substantial
trace in the archaeological record. The archaeological survey
on the C¸ ars¸amba fan undertaken by Baird (2005) focused on
the alluvial plain, where settlement mounds are abundant,
rather than on the surrounding drylands. The best-known
specialist site is the small rock shelter at Pınarbas¸i, a full day’s
walk southeast of C¸ atalho¨yu¨k, which probably went through
a phase of seasonal occupation at the end of the ceramic
Neolithic (Baird 2002). Other sizeable Neolithic sites that are
partly contemporary with C¸ atalho¨yu¨k (e.g., Can Hasan) and
from which food might have been obtained via exchange lie
at least a 3-day walk away (fig. 1).
We know from the abundant remains of wheat, barley,
lentils, and other cultivated crops found in domestic storage
bins and other contexts that C¸ atalho¨yu¨k had a strong agri-
cultural focus (Fairbairn, Near, and Martinoli 2005b; Atalay
and Hastorf 2005). By contrast, the location of the site does
not reflect the specific set of conditions that would have been
most favorable for low-risk crop production. We cannot es-
cape the supposition that the decision of Neolithic people to
locate C¸ atalho¨yu¨k within the wetland environment was for
considerations other than maximizing local agricultural pro-
ductivity. This might argue for a lesser focus of agricultural
production in the C¸ atalho¨yu¨k economy and a greater role for
other types of subsistence products, which may have provided
buffering mechanisms against crop failure. The abundant wild
plant resources, such as the bulbs of sedges (Scirpus maritimus,
etc.) within the marsh, wild crucifers, and hackberries, al-
monds, and acorns in the uplands could have provided the
kind of buffer that would allow survival in the face of an
unreliable natural environment (Fairbairn et al. 2007). By not
relying solely on agricultural products, the Neolithic villagers
would have been continuing a pattern of exploitation of ec-
otones that previous Epipaleolithic peoples relied on to reduce
subsistence risk before the emergence of the Neolithic peoples
in the basin. Other advantages to living within the wetland
zones might have included access to marl and clay sources
for construction and plastering materials, as well as the ease
of waterborne transportation, perhaps using reed watercraft
and rafts along the river. The choice of C¸ atalho¨yu¨kasalo-
cation for settlement may, of course, have also had a strong
noneconomic (e.g., symbolic) explanation.
It seems unlikely to be accidental that the period of regional
population concentration at C¸ atalho¨yu¨k coincided precisely
in time with the phase of most active river flooding. Plausibly
in response to increased flood risk, the east mound at C¸a-
talho¨yu¨k grew rapidly in size both laterally and vertically to
reach more than 20 m in height, thereby raising houses above
the highest spring floodwaters. Nucleation in turn created
protourban conditions of life, in which social cohesion and
economic well-being was partly sustained by seasonal fission
and fusion of a total population of 3,500–8,000 (Cessford
2005). It remains unknown how such a system functioned in
the apparent absence of community leadership, extrahouse-
hold food storage, or economic specialization, but increased
ritual and symbolic complexity seems to have played a key
role.
The suggestion that population nucleation was—at least in
part—a response to the environmental challenge of extreme
seasonal river flooding is given further support by the fact
that nucleated settlement came to an end when the “flooding
phase” ceased ∼6200 BC (fig. 5). During the subsequent Early
Chalcolithic period, occupation continued at C¸ atalho¨yu¨k but
moved to a new location on the other side of the river channel
to form the smaller west mound. This switch to a new, low-
lying site could have been possible only if flooding was no
longer a significant risk. C¸ atalho¨yu¨k West is one of 15 similar
small- to medium-sized Early Chalcolithic sites recorded on
or near the C¸ ars¸amba fan by the Baird (2005) survey. Al-
though overall regional population appears to have increased,
it became based in dispersed settlements associated with cereal
crops that could have been cultivated in fertile and no longer
flood-prone alluvial soils adjacent to villages.
While the change in flood regime at the end of the Neolithic
is potentially explicable in terms of a shift in the main river
channel away from C¸ atalho¨yu¨k to another part of the C¸ar-
s¸amba fan, it also coincided with a major perturbation to
global climate—the “8 ka event” (Alley and A
´
gu´stsdo´ttir
2005). This climatic event, triggered by the catastrophic col-
lapse of the remaining North American ice sheet over Hudson
Bay, led to the release of a pulse of meltwater into the Atlantic
Ocean that prompted drought conditions over much of Africa
and Asia (Gasse 2000). Dating of this event in Greenland ice
cores has shown that it lasted 160 years, from 6300 to 6140
BC (Thomas et al. 2007), the same time as the flooding phase
ended and dry conditions began at C¸ atalho¨yu¨k.
400 Current Anthropology Volume 50, Number 3, June 2009
Figure 5. Chronological chart, showing archaeological periods and re-
corded site numbers on or near the C¸ ars¸amba fan (from Baird 2005),
individual site occupations, and changing river flood regimes.
Conclusion
Here we have employed multiple lines of evidence from C¸a-
talho¨yu¨k, but especially from geoarchaeology and phytolith
analysis, to reconstruct an economic and ecological basis for
this key settlement site. Neolithic C¸ atalho¨yu¨k was located in
an extensive alluvial wetland with a strongly seasonal natural
regime and a spring flood that led to the settlement mound
being surrounded by water for up to 2 months of the year.
The seasonal wetlands provided a range of environmental
resources from club rush tubers and wildfowl to marl clay
for wall plaster, while the C¸ ars¸amba River—which then flowed
next to the site—provided a vital means of transport and
communication. However, unless planted in late spring as
flood-recession farming, the local alluvial soils cannot have
provided the locus for cereal crop cultivation. Phytolith and
other archaeobotanical data in fact suggest that most of these
crops were dry-farmed in nonalluvial soils located too far
from C¸ atalho¨yu¨k to have been visited on a daily basis. Dryland
cultivation would have been aided by the wetter and more
reliable rainfall regime that then prevailed.
These findings imply that not all Neolithic settlement in
the East Mediterranean region was reliant primarily on in-
tensive but localized cultivation of high-water-table alluvial
soils for cereal crops (Bogaard 2004). While a model involving
horticulture in alluvial soils may apply in many instances, in
the Konya basin at least, a very different and more complex
model of settlement and social organization operated, ap-
parently successfully and sustainably, for almost a millennium.
It involved a single nucleated settlement system, with fission
and fusion of population on a seasonal basis, which enabled
a diversity of natural resources to be utilized from a wide
landscape region. The ecological “footprint” of Neolithic C¸a-
talho¨yu¨k was therefore spatially extensive as well as locally
intensive.
Acknowledgements
Our thanks go to I. Hodder and to many members of the
C¸ atalho¨yu¨k team, along with D. Baird, P. Boyer, W. Eastwood,
M. Jones, G. Hillman, and S. Colledge for valuable discussions
and information and anonymous reviewers whose comments
have greatly improved this manuscript. Diagrams were drawn
by Brian Rogers. This paper is dedicated to the memory of
Andrew Sherratt, whose insights enlightened this and so many
other debates.
References Cited
Alley, R. B., and A. M. A
´
gu´stsdo´ttir. 2005. The 8k event: cause
and consequences of a major Holocene abrupt climate
change. Quaternary Science Reviews 24:1123–1149.
Asouti, E. 2005. Woodland vegetation and the exploitation of
fuel and timber at Neolithic C¸ atalho¨yu¨k: report on the
wood charcoal macro-remains. In Inhabiting C¸ atalho¨yu¨k:
reports from the 1995–99 seasons. C¸ atalho¨yu¨k research
project 4, pt. A. I. Hodder, ed. Pp. 213–258. Cambridge/
London: McDonald Institute for Archaeological Research/
British Institute of Archaeology at Ankara.
Asouti, E., and J. Hather. 2001. Charcoal analysis and the
reconstruction of ancient woodland vegetation in the
Konya Basin, south-central Anatolia, Turkey: results from
the Neolithic site of C¸ atalho¨yu¨k East. Vegetation History
and Archaeobotany 10:23–32.
Atalay, S., and C. A. Hastorf. 2005. Foodways at C¸ atalho¨yu¨k.
In C¸ atalho¨yu¨k perspectives: themes from the 1995–99 sea-
401
sons. C¸ atalho¨yu¨k research project 4. I. Hodder, ed. Pp.
109–123. Cambridge/London: McDonald Institute for Ar-
chaeological Research/British Institute of Archaeology at
Ankara.
Baird, D. 2002. Early Holocene settlement in central Anatolia:
problems and prospects as seen from the Konya Plain. In
The Neolithic of central Anatolia. F. Gerard and L. Thissen,
eds. Pp. 139–159. Istanbul: EGE Yayinlari.
———. 2005. The history of settlement and social landscapes
in the Early Holocene in the C¸ atalho¨yu¨k area. In C¸ atal-
ho¨yu¨k perspectives: themes from the 1995–99 seasons. C¸a-
talho¨yu¨k research project 6. I. Hodder, ed. Pp. 55–74. Cam-
bridge/London: McDonald Institute for Archaeological
Research/British Institute of Archaeology at Ankara.
———. 2007. The Boncuklu Project: the origins of sedentism,
cultivation and herding in central Anatolia. Anatolian Ar-
chaeology 13:14–18.
Bakels, C. C., and R. Rouselle. 1985. Restes botaniques et
agriculture de Ne´olitique ancien en Belgique at aux Pays-
Bas. Helinium 25:37–55.
Barker, G. A. 1985. Prehistoric farming in Europe. Cambridge:
Cambridge University Press.
Bar-Yosef, O., A. Gopher, E. Tchernov, and M. Kislev. 1991.
Netiv Hagdud: an early Neolithic village site in the Jordan
Valley. Journal of Field Archaeology 18:406–424.
Bogaard, A. 2004. Neolithic farming in Central Europe: an
archaeobotanical study of crop husbandry practices. Lon-
don: Routledge.
Bogucki, P. 1996. The spread of early farming in Europe.
American Scientist 84:242–253.
Boyer, P., N. Roberts, and D. Baird. 2006. Holocene environ-
ment and settlement on the C¸ ars¸amba alluvial fan, south
central Turkey: integrating geoarchaeology and archaeo-
logical field survey. Geoarchaeology 21:675–699.
Brown, A. G. 1997. Alluvial geoarchaeology: floodplain ar-
chaeology and environmental change. Cambridge: Cam-
bridge University Press.
Bryan, K. 1929. Floodwater farming. Geographical Review 19:
444–456.
Cessford, C. 2005. Estimating the Neolithic population of
C¸ atalho¨yu¨k. In Inhabiting C¸ atalho¨yu¨k: reports from the
1995–99 seasons. C¸ atalho¨yu¨k research project 4, pt. A. I.
Hodder, ed. Pp. 323–326. Cambridge/London: McDonald
Institute for Archaeological Research/British Institute of
Archaeology at Ankara.
Cessford, C., and T. Carter. 2005. Quantifying the consump-
tion of obsidian at Neolithic C¸ atalho¨yu¨k. Journal of Field
Archaeology 30:305–315.
Cohen, H. R. 1970. The palaeoecology of south central An-
atolia at the end of the Pleistocene and the beginning of
the Holocene. Anatolian Studies 20:119–137.
Colledge, S. 2001. Plant exploitation on epipalaeolithic and
early Neolithic sites in the Levant. Oxford: BAR Inter-
national.
Colledge, S., J. Conolly, and S. Shennan. 2004. Archaeobo-
tanical evidence for the spread of farming in the eastern
Mediterranean. Current Anthropology 45:S35–S58.
Driessen, P. M., and T. De Meester. 1969. Soils of the C¸umra
area, Turkey. Wageningen: Pudoc.
Fairbairn, A. 2005. A history of agricultural production at
Neolithic C¸ atalho¨yu¨k East, Turkey. World Archaeology 37:
197–210.
Fairbairn, A., E. Asouti, J. Near, and D. Martinoli. 2002.
Macro-botanical evidence for plant use at Neolithic C¸a-
talho¨yu¨k, south-central Anatolia, Turkey. Vegetation His-
tory and Archaeobotany 11:41–54.
Fairbairn, A., E. Asouti, N. Russell, and J. Swogger. 2005a.
Seasonality. In C¸ atalho¨yu¨k perspectives: themes from the
1995–99 seasons. C¸ atalho¨yu¨k research project 6. I. Hodder,
ed. Pp. 92–108. Cambridge/London: McDonald Institute
for Archaeological Research/British Institute of Archae-
ology at Ankara.
Fairbairn, A., D. Martinoli, A. Butler, and G. Hillman. 2007.
Wild plant seed storage at Neolithic C¸ atalho¨yu¨k East, Tur-
key. Vegetation History and Archaeobotany 16:467–479.
Fairbairn, A., J. Near, and D. Martinoli. 2005b. Macrobotan-
ical investigation of the north, south and KOPAL area ex-
cavations at C¸ atalho¨yu¨k East. In Inhabiting C¸ atalho¨yu¨k:
reports from the 1995–99 seasons. C¸ atalho¨yu¨k research
project 4, pt. A. I. Hodder, ed. Pp. 137–201. Cambridge/
London: McDonald Institute for Archaeological Research/
British Institute of Archaeology at Ankara.
Farid, S. 2007. Neolithic excavations in the south area, east
mound, C¸ atalho¨yu¨k 1995–99. In Excavating C¸ atalho¨yu¨k.
South, north and KOPAL area reports from the 1995–99
seasons. C¸ atalho¨yu¨k research project 3. I. Hodder, ed. Pp.
41–344. Cambridge/London: McDonald Institute for Ar-
chaeological Research/British Institute of Archaeology at
Ankara.
French, D. H. 1970. Notes on site distribution in the C¸umra
area. Anatolian Studies 20:139–148.
Gasse, F. 2000. Hydrological changes in the African tropics
since the Last Glacial Maximum. Quaternary Science Re-
views 19:189–211.
Gluza, I. 1983. Neolithic cereals and weeds from the locality
of the Lengyel culture at Nowa Huta-Mogila near Cracow.
Acta Palaeobotanica 23:123–184.
Goldberg, P., and R. I. Macphail, eds. 2006. Practical and
theoretical geoarchaeology. Oxford: Blackwell.
Hayward, D. M., and D. W. Parry. 1980. Scanning electron
microscopy of silica deposits in the culms, floral bracts and
awns of barley (Hordeum sativum Jess). Annals of Botany
46:541–548.
Hodder, I., ed. 1996. On the surface: C¸ atalho¨yu¨k 1993–95.
C¸ atalho¨yu¨k research project vol. 1. Cambridge/London:
McDonald Institute for Archaeological Research/British In-
stitute of Archaeology at Ankara.
———, ed. 2007. Excavating C¸ atalho¨yu¨k: south, north and
KOPAL area reports from the 1995–99 seasons. C¸ atalho¨yu¨k
research project vol. 3. Cambridge/London: McDonald In-
402 Current Anthropology Volume 50, Number 3, June 2009
stitute for Archaeological Research/British Institute of Ar-
chaeology at Ankara.
Hutton, J. T., and K. Norrish. 1974. Silicon content of wheat
husks in relation to water transpired. Australian Journal of
Agricultural Research 25:203–212.
Jones, M. D., N. Roberts, and M. J. Leng. 2007. Quantifying
climatic change through the LGIT based on lake isotope
palaeohydrology from central Turkey. Quaternary Research
67:463–473.
Kuzucuoglu, C., R. Parish, and M. Karabiyikoglu. 1998. The
dune systems of the Konya Plain (Turkey): their relation
to the environmental changes in central Anatolia during
the Late Pleistocene and Holocene. Geomorphology 23:
257–271.
Parry, D. W., and F. Smithson. 1964. Types of opaline silica
depositions in the leaves of British grasses. Annals of Botany
n.s. 28:169–185.
Pearson, J. A., H. Buitenhuis, R. E. M. Hedges, L. Martin, N.
Russell, and K. C. Twiss. 2007. New light on early caprine
herding strategies from isotope analysis: a case study from
Neolithic Anatolia. Journal of Archaeological Science 34:
2170–2179.
Roberts, N. 1982. A note on the geomorphological environ-
ment of C¸ atal Hu¨yu¨k, Turkey. Journal of Archaeological
Science 9:341–348.
———. 1991. Late Quaternary geomorphological change and
the origins of agriculture in south central Turkey. Geoar-
chaeology 6:1–26.
Roberts, N., S. Black, P. Boyer, W. J. Eastwood, H. Griffiths,
M. Leng, R. Parish, J. Reed, D. Twigg, and H. Yigitbasioglu.
1999. Chronology and stratigraphy of Late Quaternary sed-
iments in the Konya Basin, Turkey: results from the KOPAL
project. Quaternary Science Reviews 18:611–630.
Roberts, N., P. Boyer, and J. Merrick. 2007. The KOPAL on-
site and off-site excavations at C¸ atalho¨yu¨k 1996–2001. In
Excavating C¸ atalho¨yu¨k: south, north and KOPAL area re-
ports from the 1995–99 seasons. C¸ atalho¨yu¨k research pro-
ject 3. I. Hodder, ed. Pp. 553–572. Cambridge/London:
McDonald Institute for Archaeological Research/British In-
stitute of Archaeology at Ankara.
Roberts, N., P. Boyer, and R. Parish. 1996. Preliminary results
of geoarchaeological investigations at C¸ atalho¨yu¨k. In On
the surface: C¸ atalho¨yu¨k 1993–95. I. Hodder, ed. Pp. 19–40.
Cambridge/London: McDonald Institute for Archaeologi-
cal Research/British Institute of Archaeology at Ankara.
Rosen, A. M. 1999. Phytoliths as indicators of prehistoric
irrigation farming. In Prehistory of agriculture: new ex-
perimental and ethnographic approaches. P. C. Anderson,
ed. Pp. 193–198. Los Angeles: UCLA Institute of
Archaeology.
———. 2005. Phytolith indicators of plant and land use at
C¸ atalho¨yu¨k. In Inhabiting C¸ atalho¨yu¨k: reports from the
1995–99 seasons. C¸ atalho¨yu¨k research project 4, pt. A. I.
Hodder, ed. Pp. 203–212. Cambridge/London: McDonald
Institute for Archaeological Research/British Institute of
Archaeology at Ankara.
Rosen, A. M., and N. Roberts. 2006. The nature of C¸ atal-
ho¨yu¨k: people and their changing environments on the
Konya Plain. In C¸ atalho¨yu¨k perspectives: themes from the
1995–99 seasons. C¸ atalho¨yu¨k research project 6. I. Hodder,
ed. Pp. 39–53. Cambridge/London: McDonald Institute for
Archaeological Research/British Institute of Archaeology at
Ankara.
Rosen, A. M., and S. Weiner. 1994. Identifying ancient irri-
gation: a new method using opaline phytoliths from emmer
wheat. Journal of Archaeological Science 21:132–135.
Russell, N., and L. Martin. 2005. The C¸ atalho¨yu¨k mammal
remains. In Inhabiting C¸ atalho¨yu¨k: reports from the
1995–99 seasons. C¸ atalho¨yu¨k research project 4, pt. A. I.
Hodder, ed. Pp. 33–98. Cambridge/London: McDonald In-
stitute for Archaeological Research/British Institute of Ar-
chaeology at Ankara.
Russell, N., and K. J. McGowan. 2005. C¸ atalho¨yu¨k bird bones.
In Inhabiting C¸ atalho¨yu¨k: reports from the 1995–99 sea-
sons. C¸ atalho¨yu¨k research project 4, pt. A. I. Hodder, ed.
Pp. 99–110. Cambridge/London: McDonald Institute for
Archaeological Research/British Institute of Archaeology at
Ankara.
Sherratt, A. 1980. Water, soil and seasonality in early cereal
cultivation. World Archaeology 11:313–330.
Thomas, E. R., E. W. Wolff, R. Mulvaney, J. P. Steffensen, S.
J. Johnsen, C. Arrowsmith, J. W. C. White, B. Vaughn, and
T. Popp. 2007. The 8.2 ka event from Greenland ice cores.
Quaternary Science Reviews 26:70–81.
Todd, I. A. 1976. C¸ atalhu¨yu¨k in perspective. Menlo Park, CA:
Cummings.
Van Andel, T. H., and C. N. Runnels. 1995. The earliest farm-
ers in Europe. Antiquity 69:489–500.