Archaeobotanical Evidence for the Spread of Farming in the Eastern Mediterranean

Abstract

A major topic of debate in Old World prehistory is the relative importance of population movement versus cultural diffusion in explaining the spread of agriculture into and across Europe following its inception in southwestern Asia. An important set of data that has surprisingly been largely absent from this debate is the preserved crops and associated weeds of the earliest farmers. An analysis of archaeobotanical data from 40 aceramic Neolithic sites in southwestern Asia and southeastern Europe shows that there are vegetational signatures that characterize the different geographical regions occupied by the Early Neolithic farmers. On this basis it is argued that the compositional similarities of the crop package between the Levantine core, Cyprus, and Greece are indicative of both the routes of migration of early farming groups and the early agricultural practices of Europe's first farmers.

Full text

S35
Current Anthropology Volume 45, Supplement, August–October 2004
䉷2004 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved 0011-3204/2004/4504S4-0002$10.00
Archaeobotanical
Evidence for the
Spread of Farming in
the Eastern
Mediterranean
1
by Sue Colledge, James Conolly,
and Stephen Shennan
A major topic of debate in Old World prehistory is the relative
importance of population movement versus cultural diffusion in
explaining the spread of agriculture into and across Europe fol-
lowing its inception in southwestern Asia. An important set of
data that has surprisingly been largely absent from this debate is
the preserved crops and associated weeds of the earliest farmers.
An analysis of archaeobotanical data from 40 aceramic Neolithic
sites in southwestern Asia and southeastern Europe shows that
there are vegetational signatures that characterize the different
geographical regions occupied by the Early Neolithic farmers. On
this basis it is argued that the compositional similarities of the
crop package between the Levantine core, Cyprus, and Greece
are indicative of both the routes of migration of early farming
groups and the early agricultural practices of Europe’s first
farmers.
sue colledge is a postdoctoral fellow of the Institute of Ar-
chaeology, University College London. Born in 1955, she was ed-
ucated at the University of Birmingham (B.Sc., 1976) and the
University of Sheffield (Ph.D., 1994). Among her publications is
Plant Exploitation on Epipalaeolithic and Early Neolithic Sites
in the Levant (British Archaeological Reports International Series
986). james conolly is Lecturer in Archaeology at the Insti-
tute of Archaeology, University College London (31–34 Gordon
Square, London WC1H0PY, U.K. [j.conolly@ucl.ac.uk]). He was
born in 1968 and received his B.A. from the University of To-
ronto in 1990 and his Ph.D. from University College London in
1997. He is coauthor, with M. Lake, of Geographical Information
Systems (Cambridge: Cambridge University Press, in press).
s tephen shenna n is Professor of Theoretical Archaeology and
Deputy Director of the Institute of Archaeology, University Col-
lege London. Born in 1949, he was educated at Cambridge Uni-
versity (B.A., 1971; Ph.D., 1977). His most recent book is Genes,
Memes, and Human History: Darwinian Archaeology and Cul-
tural Evolution (London: Thames and Hudson, 2002). The pre-
sent paper was submitted 3iv03and accepted 22 i 04.
[Supplementary materials appear in the electronic edition of this
issue on the journal’s web page (http://www.journals.uchicago/
edu/CA/home.html).]
1. We thank Michael Charles, Andrew Garrard, David Harris, and
five anonymous reviewers who read and commented on drafts of
this paper. We are also grateful to the staff and students of the
Institute of Archaeology, University College London, who contrib-
Explaining the transition to agriculture is a long-standing
and central problem in European prehistoric archaeology
that traces its history to Gordon Childe’s (1929) diffu-
sionist model. In 1965 Grahame Clark mapped the ra-
diocarbon dates associated with the earliest Neolithic
sites and demonstrated a cline oriented roughly north-
west to southeast across Europe to the Levant, confirm-
ing Childe’s earlier proposals regarding the primacy of
agriculture in the eastern Mediterranean. Less than a
decade later Ammerman and Cavalli-Sforza (1973)in-
vestigated the processes underlying the expansion of
farming and developed a spatial model based in part on
Fisher’s (1937) wave of advance for advantageous genes,
proposing a demic expansion of farmers of ca. 1km per
year from an (assumed) origin in Jericho. Since then it
has been shown that there are regions that deviate sub-
stantially from this trend surface, notably along the Med-
iterranean coasts, where movement rates are faster than
predicted, and the Iberian peninsula, which experiences
much slower rates of Neolithization (Cavalli-Sforza
1996), but the 1km/year approximation has remained a
good generalization of the expansion of farming com-
munities across the continent. Debate continues, how-
ever, as to whether Neolithization was caused by a move-
ment of people or one of ideas. Recent sophisticated
quantitative analysis of the radiocarbon record has
shown that the relationship between the decline of in-
digenous Mesolithic populations and the first appearance
of farming communities was complex (Gkiasta et al.
2003). Population diffusion (e.g., through intermarriage
between hunter-gatherer and early farming groups) and
cultural diffusion independent of population movement
have been shown to give equally convincing explana-
tions of the spatial and temporal patterns of Neolithic
expansion (Richards 2003, Bentley et al. 2002, Nowak
2001, Price et al. 2001, Price 2000). The fragile consensus
is that a complex mixture of demic expansion and cul-
tural diffusion was responsible for the origin and spread
of the Neolithic into Europe from its first appearance in
southwestern Asia, although debate continues over
which of these processes was dominant in different
regions of the continent (see, in particular, Price 2000).
From the perspective of European archaeologists it is
the first Neolithic communities in Greece—from about
7000 BC
2
—that mark the beginning of the process of
European “Neolithization.” However, this date is nearly
three millennia after the emergence of agricultural com-
munities in the eastern Mediterranean, and Near Eastern
archaeologists have shown that an expansion of early
farming groups throughout the Levant and into Cyprus
and central Anatolia occurs well before the Neolithic
first appears in southeastern Europe (Cauvin 1989, Pel-
tenburg et al. 2001,O
¨zdog˘an 1997). With few exceptions
(e.g., Perle`s 2001), accounts of European Neolithization
pay little more than lip-service to the substantial re-
search that has been undertaken on the spread of early
uted to discussions after we presented this paper at a research sem-
inar in January 2003.
2. All dates BC are calibrated.

S36 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
Fig. 1. Eastern Mediterranean aceramic Neolithic sites referred to in the text. (Locations 25,35, and 37 refer to
two sites each.)
farming practices in the eastern Mediterranean, and, con-
versely, Near Eastern archaeologists rarely look beyond
Anatolia for further evidence to address the central ques-
tion of why farming spread and how farming practices
were adapted to European temperate environments. It is
in this regard that we draw attention to the fact that
although the fundamental elements of the European crop
package are derived from founder species that evolved
in southwestern Asia, archaeobotanical data have never
been referred to in any models of Neolithization in an
interregional comparative context.
As we demonstrate here, when examined in a spatial
and chronological framework the archaeobotanical rec-
ord provides insight into the transition to farming at both
local and regional levels. As well as providing informa-
tion on the use of domestic crops, the accompanying wild
species, and, more important, associated weeds, the data
can be used to construct a comprehensive picture of the
evolution and adaptation of agricultural systems over
space and time. In this paper we present the results of
comparative analyses of archaeobotanical data from 40
aceramic Neolithic sites from the eastern Mediterranean
(fig. 1). The results of the analysis clarify regional dif-
ferences in crop composition and refine the chronologies,
sources, and routes of dispersal of the earliest domestic
crops from southwestern Asia into southeastern Europe.
3
Integrating Archaeobotanical Data into
Models of the Spread of Farming
Despite concentrated research into the archaeobotanical
foundations of early agriculture in localized areas of the
Levant and southeastern Anatolia, there has never been
a systematic, pan-regional comparative analysis of ar-
chaeobotanical assemblages of key sites from south-
western Asia and Europe. To address this problem, we
have assembled a relational database containing records
of the plant taxa represented on pre- and early Neolithic
sites, linked to a database of
14
C dates for sites in Europe
4
and to additional dates for the southwestern Asian sites.
3. This paper presents initial findings arising from a larger project
(The Origin and Spread of Neolithic Plant Economies in the Near
East and Europe) funded by the Arts and Humanities Research
Board (U.K.), directed by Stephen Shennan and James Conolly of
the Institute of Archaeology, University College London, in col-
laboration with James Steels of the University of Southampton. Sue
Colledge is the project’s research fellow.
4. http://ads.ahds.ac.uk/catalogue/collections/blurbs/283.cfm.

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS37
Fig. 2. The data model. Primary fields are denoted by asterisks; arrows show one-to-many relationships.
At the time of writing, the database contains detailed
archaeobotanical information from 166 sites comprising
a total of 243 phases with ca. 1,000 associated radiocar-
bon dates and covering an area from southern Iran to
northwestern Europe (including Bulgaria, Romania,
Hungary, the former Yugoslavia, Italy, the Czech Re-
public, Austria, France, and Germany as well as coun-
tries in the eastern Mediterranean). The time span rep-
resented by these sites is ca. 15,500 years, between ca.
21,000 bp and ca. 5,500 bp. The data model is shown in
figure 2.
Each taxon (e.g., species) that has been identified by
the archaeobotanists responsible for the different sites
represents a separate record in the archaeobotanical data
table, and to date 6,121 entries have been made (our data
consist of records of remains that have been preserved
by charring, with very few mineralized specimens or
identifications made from impressions included). Num-
bers and ubiquity scores (i.e., number of samples in
which a taxon occurs as a percentage of the total number
of samples) relate to the representation of these taxa ac-
cording to major cultural phases rather than to individual
samples. The records include references to a total of 719
taxa that have been listed in the published reports (in-
cluding wild and domestic cereals and pulses, fruits, oil
plants, and many wild or weed species). Archaeobotan-
ical reports have been critically reviewed prior to enter-
ing information in the database; notes made by the au-
thors have been added to accompany the records,
providing details of the identification criteria they used,
for example, to distinguish between the wild progenitors
and domestic species. Archaeological literature relating
to the dating and phasing at the sites has also been re-
ferred to, and this has thrown light on inaccuracies in
published articles, most significantly with respect to the
contextual or chronological association of early finds of
domestic crops (e.g., Jericho [see below and Colledge
2004]).
Once complete, this publicly available database will
be an invaluable resource for the study of the origins of
and the transition to agriculture in southwestern Asia
and Europe. We demonstrate its relevance and value here
and show how multivariate analysis of archaeobotanical
data can contribute to our understanding of the spread
of the Neolithic crop package.
The Origins of Farming in Southwestern Asia
and Southeastern Europe: Spatial and
Temporal Dynamics
The first evidence of domestic crops
5
occurs in south-
western Asia in the first centuries of the tenth millen-
nium BC, probably at the beginning of the climatic ame-
lioration following the Younger Dryas stadial. The
earliest domestic cereals were emmer (Triticum dicoc-
cum), einkorn (Triticum monococcum), and hulled bar-
ley (Hordeum vulgare, hereafter referred to as Hordeum
sativum). Together with flax (Linum usitatissimum) and
four pulses—lentil (Lens culinaris), pea (Pisum sativum),
bitter vetch (Vicia ervilia), and chick pea (Cicer arie-
tinum)—they constitute the “founder crops” of Neo-
lithic agriculture (Zohary 1996). This assemblage of spe-
5. We refer here only to the “founder crops” which formed the basis
of Neolithic agriculture, not including, therefore, the early finds of
domestic rye from Epipalaeolithic contexts at Abu Hureyra (Hill-
man 2000:379–84).

S38 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
Fig. 3. Sites with references to domestic hulled barley, 10,000–6500 Cal BC. (For key see figure 1.)
cies (or crop package) was adopted either in its entirety
or in subsets and spread rapidly beyond the places in
which the first domestication events occurred (Garrard
1999) (figs. 3and 4).
Evidence for the earliest domestic cereal species on
aceramic Neolithic sites (culturally defined as Pre-Pot-
tery Neolithic A [PPNA] and Early Pre-Pottery Neolithic
B [EPPNB]; table 1) and, by extension, the timing of the
first domestication events are considered by some re-
searchers to be unreliable. These researchers dispute the
authenticity of the archaeobotanical finds for two rea-
sons: they question (1) whether a distinction between
wild and domestic grains/chaff can be made, given the
paucity of the remains and the poor quality of preser-
vation, and (2) whether domestic cereals can be assigned
to the earliest Neolithic phases, given the ambiguities
of radiocarbon dating and the questionable stratigraphic
integrity of the samples in which they were found (see,
e.g., Nesbitt 2002). However, if domestic cereal taxa from
PPNA and EPPNB contexts are omitted from the ar-
chaeobotanical record—the rationale being that all early
samples are intrusive or incorrectly dated—the question
remains why domestication events had not occurred in
the millennium between the time when cultivation of
wild species began in the Epipalaeolithic (Hillman et al.
2001; Hillman 2000:376–96) and the first unequivocal
signs of domestic crops linked with the more reliable
socio-demographic evidence of large-scale production by
farming communities in the Middle Pre-Pottery Neo-
lithic B (MPPNB) (Harris 1998:69–71). This “long-ges-
tation” argument is contrary to the views held by Hill-
man and Davies (1992:144), who, on the basis of
experimental evidence, have proposed that domestica-
tion could have taken place within 20–200 years after
the initial attempts were made at cultivating stands of
wild cereals. The semi-tough mutant forms in the wild
populations (i.e., the precursor domestic species) would
have experienced a positive selective advantage only un-
der certain conditions. The prerequisite conditions iden-
tified by Hillman and Davies are (1) harvesting of cereals
by sickle or uprooting, (2) harvesting while crops were
partially ripe or near ripe, (3) annual extension or shifts
of the area under cultivation, and (4) the taking of each
year’s seed stocks from the harvests of the previous sea-
son’s new plots (pp. 124–32). Given these preconditions,
they calculate, the rate of domestication would have
been relatively rapid.
This “short-gestation” model places the PPNA in a
pivotal role in the emergence of crop-based subsistence
(Harris 2002:69–70), and this is also in concordance with
the argument that climatic stability immediately follow-
ing the Younger Dryas stadial was an essential factor in

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS39
Fig. 4. Sites with references to domestic emmer/einkorn, 10,000–6500 Cal BC. (For key see figure 1.)
the relatively rapid shift to domestic cereal species (Rich-
erson, Boyd, and Bettinger 2001). We find these argu-
ments more persuasive, and our temporal and spatial
models therefore include the archaeobotanical evidence,
albeit tenuous, for domestic cereals from the earliest Ne-
olithic phases.
The geographical distributions of the wild progenitor
species of the three cereals overlap but differ in their
relative extents (fig. 5). The common focus of all three
species is the “Near Eastern arc” (i.e., the “fertile cres-
cent”). Zohary and Hopf describe the natural range of
wild emmer (Triticum dicoccoides) as extending
throughout Israel, Jordan (particularly the Jordan Valley
catchment, where it is most widespread), southwestern
Syria, Lebanon, southeastern Turkey, northern Iraq, and
western Iran (Zohary and Hopf 2000:44; Valkoun, Giles
Waines, and Konopka 1998). Wild einkorn (Triticum
boeoticum) is found as far west as the southern Balkans
and reaches Iran in the east; it has its distribution centre
in the Near Eastern arc and is prevalent in northern Sy-
ria, southern Turkey, and northern Iraq (Zohary and Hopf
2000:35; Valkoun, Giles Waines, and Konopka 1998). The
centre of the distribution of wild barley (Hordeum spon-
taneum) also lies in the Near Eastern arc. From Israel
and Jordan in the southwest its range extends north to-
wards southern Turkey and southeast into Iraqi Kurdi-
stan and southwestern Iran (Zohary and Hopf 2000:65;
Valkoun, Giles Waines, and Konopka 1998). It is in the
area covered by the western arm of the arc that the first
domestic cereals are thought to have originated. On the
basis of genetic evidence, it has also been postulated that
the domestication events which gave rise to the three
founder-crop cereals occurred only once or at most very
few times (Zohary 1996,1999).
The Earliest Evidence for Domestic Cereals in
the Western Arm of the Near Eastern Arc
The Levantine corridor. The earliest dated evidence for
the use of domestic cereals in southwestern Asia comes
from PPNA levels at the site of Iraq ed-Dubb (fig. 1, table
2) in the southern Levant in the first half of the tenth
millennium BC
6
(Colledge 2001:143–44; Kuijt n.d.). Van
Zeist and Bakker-Heeres have also recorded the presence
of domestic cereal species in PPNA phase Ia at Tell As-
wad in the central Levant in the late tenth millennium
BC (van Zeist and Bakker-Heeres 1982). Hopf (1983) iden-
tified domestic einkorn, emmer, and barley from charred
6. Calibrated date ranges were estimated using OxCal v3.8.

S40 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
table 1
Aceramic Neolithic Chronological Periods and Ap-
proximate Date Ranges
Chronological Period Start Cal BC End Cal BC
Pre-Pottery Neolithic A ca. 9800 ca. 8800
Early Pre-Pottery Neolithic B ca. 8800 ca. 8400
Middle Pre-Pottery Neolithic B ca. 8400 ca. 7500
Late Pre-Pottery Neolithic B ca. 7500 ca. 7000
Final Pre-Pottery Neolithic B/
Pre-Pottery Neolithic C
ca. 7000 ca. 6500
source: Kuijt (2000:3–13).
Fig. 5. Approximate distribution of the wild progeni-
tor species of wheat and barley in the Eastern Medi-
terranean. A, T. boeoticum; B, T. dicoccoides; C, H.
spontaneum (after Valkoun, Giles Waines, and Kon-
opka
1998
, Zohary and Hopf
2000
).
remains and impressions in plaster from PPNA levels at
Jericho. Close scrutiny of the exact provenance of the
samples in which the Jericho cereals were found has re-
vealed that, although they are within levels defined cul-
turally as PPNA (Bar-Yosef and Kra 1994:6; Bar-Yosef and
Gopher 1997:251), the radiocarbon evidence shows that
they are later than the Iraq ed-Dubb and Aswad finds
and chronologically equivalent to the EPPNB. Two of
the samples were taken from trench D1(stage VIII con-
texts of Kenyon’s stratigraphic system), associated with
four radiocarbon dates that fall in the middle of the ninth
millennium BC (Burleigh 1981:501–4;1983:760–65).
7
The other two samples in which Hopf identified do-
mestic cereals were both from trench Tr1(stage X con-
texts), approximately 2m above stage VIII, just below
the PPNA/B boundary sealing the tower and closely as-
sociated with a date for stage IX of 9200 Ⳳ70 bp (BM-
1789 [Burleigh 1983:760–65])(for a more detailed account
see Colledge 2004).
Northern Levant and southeastern Anatolia. The
well-documented and dated presence of PPNA (or PPNA-
like) settlements at C¸ ayo¨nu¨ and Hallan C¸ emi could be
interpreted as the terminal extension of a northern dis-
tribution of Levantine PPNA sites, including Mureybit,
Qermez Dere, and M’lefaat, all of which date roughly to
the late eleventh/early tenth millennium BC. There are,
however, no known sites dating to the PPNA/EPPNB
between the Damascus Basin and southeastern Anatolia
which have unequivocal evidence of the use of domestic
cereals. At the early Neolithic sites of Jerf al Ahmar
(Willcox and Fornite 1999, Willcox 1996), Tell Mureybit
(van Zeist and Bakker-Heeres 1984), and Dja’de (Willcox
1996) in the Euphrates Valley, it has been proposed that
there was cultivation of wild cereals, with no signs of
domestic species in this area until the MPPNB.
8
For ex-
ample, a date of 8700 Ⳳ75 bp is associated with domestic
cereals at Halula (Willcox 1996). The first evidence for
domesticated cereals outside of the Levantine corridor
7. Omitting earlier dates which are considered to be unreliable (see
Bar-Yosef and Kra 1994:6; Bar-Yosef and Gopher 1997:251).
8. Willcox (2004) has recently claimed on the basis of metric anal-
ysis that in the later levels at Jerf al Ahmar and at Dja’de (both
equivalent to the EPPNB) there are “plump-type” wheat and barley
grains similar in size to domestic cereal species. At both sites,
however, all wheat chaff and over 90% of barley chaff were recorded
as having wild-type abscission scars.
(sensu stricto) comes from C¸ ayo¨nu¨ in southeastern An-
atolia, where van Zeist and de Roller (1991–92) identified
domestic einkorn and emmer in the phases assigned to
the EPPNB, including the grill building, channelled
building, and associated basal pits (O
¨zdog˘an 1999,Bi-
c¸ akc¸i 1998). This brackets the first use of domesticates
to between 9320 Ⳳ95 bp (GrN-6243) and 8980 Ⳳ80 bp
(GrN-6244), which give a calibrated range of approxi-
mately 8700 BC–7800 BC, roughly equivalent in time to

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS41
table 2
Radiocarbon and Calendar Dates Associated with the
Earliest Domestic Cereals in the Levant
Site Lab Number Date bp 1sigma Cal BC 2sigma
Jericho BM-1321 9,230 80 8450 180
Jericho BM-1787 9,280 100 8525 275
Jericho BM-1226 9,320 220 8600 700
Jericho BM-252 9,320 150 8700 450
Tell Aswad GIF-2372 9,640 120 8950 350
Tell Aswad GIF-2633 9,730 120 9125 475
Iraq ed-Dubb AA-38145 9,941 72 9475 275
Iraq ed-Dubb AA-38140 9,952 64 9475 275
Iraq ed-Dubb OxA-2567 9,959 100 9600 400
sources: Burleigh (1981,1983), de Contenson (1973), Kuijt
(n.d.).
the first dated use of cereals at Jericho. Debate continues
over whether southeastern Anatolia was a separate “cen-
tre” of domestication for Triticum monococcum (see
Heun et al. 1997). Nesbitt (2002:127), after examination
of the Nevali C¸ ori glume wheat chaff, stated that it was
all of the domestic type, but, to date, full results have
not been published by the primary analyst.
Early Agricultural Migration and Colonization
Previous research by Peltenburg et al. (2001), Perle`s
(2001), O
¨zdog˘an (1997), van Andel and Runnels (1995),
and Cauvin (1989), among others, emphasizes migration
as a causal factor during the early phases of the spread
of farming. More specifically, it has been argued that
after the emergence of domestic cereals in the western
arm of the Near Eastern arc, early farming communities
colonized three adjacent regions from the late tenth/
early ninth millennium BC onward. In chronological or-
der, these were (1) Cyprus, (2) central Anatolia, and (3)
Crete and Greece. Each of these events can be conceived
of as an influx of farmers into favourable ecological
zones, replacing through assimilation or displacement
what were in all probability very small populations of
local hunter-gatherers. The following paragraphs present
the evidence for migration in more detail.
the colonization of cyprus
Prior to excavations in the 1990s it was considered un-
likely that there were any aceramic Neolithic settle-
ments on Cyprus earlier than those of Khirokitian date,
ca. 7000 BC. The ninth-millennium-BC dates for Kala-
vassos Tenta that Todd (1987) recorded were thought to
be anomalous, as were the circular stone-based houses,
which, it was argued, could not have predated those at
Khirokitia. The indefinable traces of pre-Khirokitian sed-
entary occupation only reinforced ideas that the original
dispersal episodes from the Levantine corridor (to coastal
Syria and across to Cyprus) occurred in the Late Pre-
Pottery Neolithic B (LPPNB) (e.g., Cauvin 1989). The sug-
gestion by Simmons et al. (1999) that the inhabitants of
the eleventh-millennium-BC rock shelter at Akrotiri-Ae-
tokremnos were transitory foragers who briefly exploited
the available resources and then left the island did noth-
ing to dispel these ideas. Recent excavations, however,
have identified six Neolithic sites that date to the first
half of the ninth millennium BC, showing that there was
much earlier permanent migration and colonization by
Neolithic farming communities (Peltenburg et al. 2001).
Domestic cereals and subsistence animals (domestic pig,
morphologically wild cattle, sheep, and goat) have been
documented, together with a chipped stone inventory
broadly analogous to the EPPNB in the Levant. The ab-
sence of indigenous foragers and equivalent animal or
plant species (i.e., wild progenitors of the domestic spe-
cies) in the Pleistocene/early-Holocene records means
that the crops and livestock were brought to the island
by the early settlers.
Peltenburg argues that at this time Cyprus was in-
cluded in a PPNB “interaction sphere,” a network of
contacts between different regions in the Levant that
facilitated trade and exchange of materials. Aspects of
subsistence, technology, settlement organization, and
ideology have been used to infer links with the northern
Levantine corridor in the EPPNB (Peltenburg et al. 2001:
38–39). He also suggests that the migratory populations
originated from coastal Syria (from sites which have yet
to be found and/or have since been flooded) and that their
ability to adjust their lifeways to the unpredictable sea-
level changes of this time would have prepared them well
for resettlement in new, unknown territories on the
island.
emergence of agriculture in central
anatolia
For the purposes of this discussion we need only high-
light that evidence for Late Pleistocene occupation in
central Anatolia is absent, although the coastal areas of
the Mediterranean and Sea of Marmara have provided
data on hunter-gatherer activities (O
¨zdog˘an 1997:17).
This lack of evidence can perhaps be explained in terms
of the erasure of traces of earlier habitation in areas like
the fertile plains around Konya and Karaman by signif-
icant alluvial sediments (Baird 2002, Roberts 1991). An
extensive survey of this area designed in part to address
this problem has, however, so far failed to recover defin-
itive evidence for Pleistocene settlement (Baird 2002).
There is correspondingly limited data about early Ho-
locene hunter-gatherers, but the ongoing excavations at
Pinarbas¸ıare likely to improve this situation (Baird 2002:
142–43; Watkins 1996). There must have been at least
short-term occupation by mobile groups, if not more per-
manent settlement, from around the late tenth millen-
nium BC, as small amounts of Anatolian obsidian are
found at a few Levantine PPNA sites (Binder 2002:81;
Cauvin et al. 1998). Recent evidence for an early-ninth-
millennium obsidian “workshop” at Kaletepe is perhaps
significant in this regard (Balkan-Atlıand Binder 2000).
The earliest farming community yet known in central

S42 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
Anatolia is As¸ıklıho¨yu¨ k, near Kaletepe and close to ob-
sidian sources. As¸ıklıho¨yu¨ k was occupied for several cen-
turies from approximately 8400 BC,
9
and the community
shares many basic characteristics with PPNB sites of the
Levant, among them rectangular housing with extensive
plastering and a technological variant of the so-called
naviform core-reduction method (Balkan-Atlı1994,
Binder 2002). The range of domestic crops present at
As¸ıklıho¨yu¨ k is similar to that found on the earliest farm-
ing settlements in the Levant (Asouti and Fairburn 2002).
The faunal assemblage, however, appears to be composed
entirely of wild species. Later Anatolian aceramic Ne-
olithic sites, such as Can Hasan III in southern central
Anatolia (mid-eighth millennium BC), are similar in na-
ture, with broadly analogous plant economies and de-
pendence on wild animals.
The paucity of pre-Neolithic central Anatolian sites is
in striking contrast with the situation in the Levantine
core, where Terminal Pleistocene hunter-gatherer settle-
ments are abundant. The evidence from central Anatolia
draws us naturally towards a model in which Late Pleis-
tocene and Early Holocene hunter-gatherers were largely
absent or present in only small numbers. Given the sud-
den appearance of large PPNB-like villages using do-
mestic crop species common to the Levant and no evi-
dence for settlements equivalent in form to the PPNA,
the most parsimonious explanation for the origins of
farming in central Anatolia is that there was an influx
of farmers during the first half of the ninth millennium
BC.
the colonization of crete and mainland
greece
Crete and Greece have traditionally been considered sep-
arately in terms of Neolithization, as there is no con-
vincing evidence for any settlement of Crete prior to the
earliest Neolithic of Knossos, whereas indigenous
hunter-gatherers had lived in very low densities for some
millennia on the Greek mainland (Runnels 1995). There
can be no doubt, therefore, that Crete was colonized by
migrant farmers (Broodbank and Strasser 1991), but the
relationship between hunter-gatherers and farming com-
munities on the mainland is controversial (Halstead
1996, Tringham 2000). Although the consensus is solidly
in favour of an exogenous origin for the latter, an im-
portant focus of research in recent years has been the
development of models that attempt to describe the or-
igins, impetus, and sequence of events that brought early
Neolithic settlers to Crete and Greece (Perle`s 2001, Hal-
stead 1996). Unfortunately, the relationship between the
latest hunter-gatherer and the earliest Neolithic settle-
ments on the mainland remains ambiguous, and the
chronic low density of sites dated to this period (i.e., ca.
7000 BC) prohibits detailed modelling of the transition
(Andreou, Fotiadis, and Kotsakis 1996:596–97).
9. Radiocarbon dates for Anatolian sites have been obtained from
the CANeW project (Central Anatolian Neolithic e-Workshop) at
http://chez.com/canew/cadata.htm.
Perle`s (2001) provides an excellent summary of the
character of the earliest Neolithic sites on Crete and
Greece in terms of their chronological framework and
potential origin. The evidence from the mainland sites
of Gediki, Sesklo, Argissa, and Franchthi Cave and Knos-
sos on Crete is strongly suggestive of an initial wave of
settlement from the seventh millennium BC, but Perle`s
stresses the difficulty in pinpointing an origin for the
migrants. Although they certainly arrived from the east
and brought with them a domestic crop and livestock
package and new forms of material culture, it is unclear
whether Crete and Greece were first settled by farming
communities whose origins were in Anatolia or Cypro-
Levantine groups followed a coastal Anatolian and “is-
land-hopping” route to the Aegean (Broodbank and Stras-
ser 1991, Demoule and Perle`s 1993). The hints of
aceramic Neolithic settlement in northwestern Anatolia
might indicate a more northerly route into southeastern
Europe (O
¨zdog˘an1997:17–19), but the lack of comparable
Early Neolithic data in northeastern Greece militates
against this model, at least for the first phases of colo-
nization. A more southerly coastal and island-hopping
route for Early Neolithic settlement, in which migrant
farmers arrived from coastal Anatolia in a “jump-dis-
persal” process (van Andel and Runnels 1995), is there-
fore more likely (cf. Davis 1992:702).
Temporal and Spatial Parameters of the
Earliest Domestic Cereals
If the migration hypothesis for the initial spread of farm-
ing in the Eastern Mediterranean is valid, then we would
expect to find corroborative patterns in the archaeobo-
tanical record. Our analysis of the taxonomic composi-
tion of the sites has demonstrated that archaeobotanical
records are of sufficiently high resolution to detect “veg-
etational signatures” on the basis of similar patterns of
taxon presence. These are evident in the form of clusters
of sites that show strong spatial patterning. We propose
that the data are therefore suitable for charting the spread
of species or groups of species between and across the
different regions. This is illustrated in the example we
describe here, in which we present the results of our
exploration of the compositional variation of aceramic
Neolithic phases of sites in Jordan, Syria, Israel, Turkey,
Cyprus, Greece, and Crete (tables 3and 4), covering a
period of from ca. 10,000 BC to ca. 6500 BC.
Presence/absence of taxa formed the basis of the anal-
ysis. We used correspondence analysis,
10
an established
multivariate statistical technique used by both ecolo-
gists and archaeologists and well suited to the explora-
tory analysis of binary data (Greenacre 1984; Shennan
1997:308–52), to investigate these data. Domestic crops,
wild cereals, and weed taxa, excluding taxa that occurred
in less than 5% of phases, were selected. [A detailed
explanation of the rationale for data selection and the
10. Using CANOCO 4.5, and Canodraw for Windows (ter Braak and
Smilauer 2002).

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS43
table 3
Sites and Phases Included in the Analysis,
10,000
–
8000
Cal BC
Site Phase Symbol
10,000–8000 Cal BC
Aswad Level Ia AsIa
Iraq ed-Dubb PPNA levels IeDII
Jerf al Ahmar Trench A15 (Bx, By,
Bz, Cx, Cy) and
Trench A90 (Bc,
Cx, Cy, Ex, Ey)
JA
Mureybit Level II MuII
Mureybit Level III MuIII
Netiv Hagdud Loci 1002–1007,
1013, and 1014
NH
8800–8400 Cal BC
Aswad Level Ib AsIb
C¸ ayo¨nu¨ Grill building,
basal pits, and
channelled building
CAIb
Dja’de PPNA/EPPNB
contexts
DJ
Kissonerga
Mylouthkia
Level IA KMIA
Mureybit Level IV MuIV
Nevali C¸ ori PPNB contexts NC
Wadi Jilat 7Trench A WJ7I
table 4
Sites and Phases Included in the Analysis,
8400
–
6500
Cal BC
Site Phase Symbol
8400–7500 Cal BC
’Ain Ghazal M/LPPNB contexts AG
As¸ıklıho¨yu¨ k Phase 2AsH
Aswad Level II AsII
Beidha Phases A and B BE
C¸ ayo¨nu¨ Cobble-paved building CAIc
C¸ afer Ho¨yu¨ k Phase I: levels III and
IV
CHI
C¸ afer Ho¨yu¨ k Phase II: levels V to
VIII
CHII
C¸ afer Ho¨yu¨ k Phase III: levels IX to
XIII
CHIII
Ghoraife´ Level I GhI
Halula MPPNB contexts HAI
Hacilar PPN levels HrI
Jericho Trench D1(stages
XVII, XIV, XXIII),
trench EI, II, V
(stages XII, XIII),
trench FI (stages
XXI, XXIV) and
trench TrI (stage
XIX)
JEII
Nahal Hemar Strata 3and 4Nhe
Wadi Jilat 7Trench B (squares
1–4,5–8)
WJ7II
7500–6500 Cal BC
Abu Hureyra Level 2AAH2A
Abu Hureyra Level 2BAH2B
Argissa Magoula AMI
Azraq 31 Square 1Az31
Basta Phases 3and 4BA
Bouqras Levels 1–10 BQ
C¸ ayo¨nu¨ Cell building CAId
C¸ ayo¨nu¨ Large room building CAIe
Cape Andreas
Kastros
Aceramic contexts CAK
Can Hasan III Trenches 48K and
49L
CHaIII
C¸ atalho¨yu¨ k Pre-level XIIA CtHPXII-A
C¸ atalho¨yu¨ k Pre-level XIIB CtHPXII-B
C¸ atalho¨yu¨ k Pre-level XIIC/D CtHPXII-C/D
Dhali-Agridhi Layer III DA
El Kowm I Phase A, Level IX EKIa
El Kowm II El Kowm 2Caracol EKII
Franchthi Cave Zone VI trenches FAS
and FAN
FCVI
Ghoraife´ Level II GhII
Gediki Aceramic contexts Gk
Halula M/LPPNB contexts HAII
Khirokitia Trench E KhE
Khirokitia Small sounding KhTh
Khirokitia Trench W KhW
Kissonerga
Mylouthkia
Level IB KMIB
Knossos Stratum X KNH
Parekklisha-
Shillourokambos
Phase 4PSIII
Ramad Level I RaI
Ramad Level II RaII
Ras Shamra Trench Vc RsVc
Sesklo “Aceramic” level SkI
Wadi Fidan A Areas 006 and 007 WFA
Wadi Fidan C Area 001 WFC
Wadi Jilat 13 Trenches B, C WJ13
full data set are provided in the electronic edition of this
issue on the journal’s web page.] The resultant data ma-
trix comprises 60 phases and 70 taxa.
Results. In the correspondence analysis plot of figure
6, the horizontal axis (axis 1) separates a majority of the
Jordanian sites, the Israeli, Greek, Cypriot, and Cretan
sites, and the Syrian sites in the Damascus Basin and on
the Mediterranean coast from the Turkish sites (with the
exception of one, CAIe) and the Syrian sites in the Eu-
phrates Valley and the central steppe region (with the
exception of one, EKI). There is also grouping of phases
along axis 2: the Damascus Basin/Mediterranean coastal
Syrian sites are distinct from the Jordanian/Greek/Cyp-
riot/Cretan sites (the Israeli sites are divided between
two groups). The analysis has therefore resulted in strong
geographic patterning of the aceramic Neolithic phases,
most noticeably forming a distinction between the Jor-
danian/Greek/Cypriot/Cretan sites and the Syrian/Turk-
ish sites.
In the bi-plot of figure 7, the distributions of taxa and
phases are superimposed. For clarity, only the domestic
crops (i.e., founder crops and secondary domesticates)
and wild progenitor species (if present in the data set)
are highlighted with distinguishing symbols and labelled
in full. Most noticeably, the wild cereals are separated
from all the domestic crop species (including cereals,
pulses, and flax) on axis 1. This strong separation be-
tween wild and domestic species therefore matches the
clear distinction between the two sets of sites on axis 1.
The wild cereals correspond with the distribution of the
Turkish and Syrian sites in the Euphrates Valley and
central steppe; the domestic crops are associated with
the Jordanian, Israeli, Greek, Cypriot, and Cretan sites

S44 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
Fig. 6. Correspondence analysis plot of aceramic Neolithic site phases for domestic cereals, pulses, flax, wild
cereals, and weed taxa. (For phase codes, see tables
3
and
4
.)
and the Syrian sites in the Damascus Basin and on the
Mediterranean coast. On axis 2the glume wheats and
hulled barley are clearly distinguished from the free-
threshing wheats and naked barley. Pea and lentil are
aligned with the former group and chick pea and flax
with the latter. As with the obvious correspondence be-
tween site groups and wild or domestic crop types on
axis 1, then, there is an equally clear correlation between
the hulled and naked cereals and the Jordanian/Greek/
Cypriot/Cretan sites versus the Damascus Basin/Medi-
terranean coastal Syrian sites on axis 2.
Before interpreting these patterns it is necessary to
take into account both the limitations of the method
and other factors that may be influencing the distribu-
tion of phases and sites.
Limitations. Zohary and Hopf (2000:247) warn against
the “unevenness” of the archaeobotanical record with
reference to their attempts to reconstruct agricultural
origins and diffusion. Similar caution should also be ap-
plied to this study. Over a period of ca. 40 years (from
the 1960s) many archaeologists and archaeobotanists
have been responsible for the excavation, sampling, and
analyses of the sites recorded in our database, and there
has no doubt been great variation in their methods and
approaches. It is inevitable, therefore, that the resultant
data, in part, reflect these disparities. For example, the

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS45
Fig. 7. Correspondence analysis bi-plot of aceramic Neolithic sites for domestic cereals, pulses, flax, wild cere-
als, and weed taxa. (For country key, see figure
6
.)
methods by which the plant remains were recovered
from the sites vary, and it may be that the taxonomic
comparisons we have made are between sites where sam-
ples were hand-picked versus those where there was re-
covery by flotation using graduated sieves. In the former
case only taxa represented by plant parts large enough
to be visible to the naked eye would have been recog-
nized and extracted, whereas with flotation recovery
even the smallest weed seed or fragment of chaff would
have been retained and subsequently identified. The de-
gree to which the archaeobotanical record is represen-
tative of the range of taxa preserved on different sites
may also be dependent on the relative thoroughness of
sampling—for example, there may be variations in the
number of samples taken, the size of the samples, and
the range of context types sampled. In some instances
there were no records of recovery, sampling, and pro-
cessing and therefore no means of establishing whether
the data were compromised because of inconsistent or
inefficient methodologies. The taphonomic processes

S46 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
table 5
Numbers of Weed Taxa in
Early Neolithic Phases by
Region
Region Mean
Southern Levant 5.92
Central Levant 14.13
Northern Levant 15.67
Cyprus 5.50
Southeast Turkey 7.63
Central Turkey 12.50
Crete and Greece 2.00
operating at the sites are doubtless varied, but analysis
at the level of presence/absence precludes further explo-
ration of the extent to which these have been a factor
in deposition and preservation of plant materials. A more
in-depth examination (see, e.g., Colledge 2001:97–108)
would be necessary to establish the possible influence
of taphonomy on the composition of samples from the
different sites and thus on the overall data set. Interest-
ingly, in a similar study, Lange found that there had been
very little measurable effect (in terms of the results of
correspondence analysis) on archaeobotanical samples
from Roman sites: “Although it is likely that tapho-
nomic processes do alter the original composition of the
assemblages, it has been found in this study that these
processes have not been responsible for the most im-
portant trends in the data” (1990:135).
It is beyond the scope of this paper to deal with all the
possible anomalies that may have arisen because of the
non-standard approaches to the identification and/or re-
cording of taxa by different archaeobotanists. An element
of bias is bound to have been introduced as a conse-
quence. However, the fact that correspondence analysis
resulted in coherent patterns of sites clustered in accor-
dance with their geographic location suggests to us that,
despite the potential problems just described, there is a
substantial degree of integrity in the primary data that
warrants interpretation at an appropriate level. In this
regard, it is significant that the results show composi-
tionally distinctive but internally homogeneous clusters.
In some instances, sites and phases that cluster together
were analyzed by a single archaeobotanist (e.g., van Zeist
studied the Damascus Basin sites and many of the Eu-
phrates Valley sites). Other clusters contain sites and
phases examined by several analysts. Moreover, some
archaeobotanists (e.g., van Zeist, Willcox) were respon-
sible for the identifications of remains at sites that fall
into two widely spaced clusters (e.g., Cyprus and the
northern Levant [Syrian sites in the Euphrates Valley and
central steppe]). We argue therefore that neither idiosyn-
cratic methods of analysis nor differential skill levels of
the archaeobotanists responsible for the collection of the
primary data are ultimately responsible for the observed
patterning in the correspondence analysis. Instead, we
propose that these patterns in part reflect regional var-
iations relating to the emergence and initial dispersal of
the founder crops.
Interpretation. Here we are concerned primarily with
the spatial dynamics of the origins and dispersal of the
earliest (i.e., founder) crop cereals.
The overlapping phases of the southern Levantine (i.e.,
all the Jordanian sites and one Israeli site), Cypriot, Cre-
tan, and Greek sites reflect a similarity in the cereal and
pulse crops represented but, most significant, in the prev-
alence of domestic glume wheats and hulled barley that
could be indicative of the earliest migratory routes taken
by farming communities after the establishment of the
founder-crop cereals. The northern Levantine, south-
eastern Turkish, and central Anatolian sites cluster to
the left of the plot, partly because of the presence of wild
cereals, possibly indicating their use at certain of the
sites both prior and subsequent to the emergence of the
domestic species. The grouping of the central Levantine
sites (Damascus Basin and coastal sites) in the lower
right quadrant reflects the frequent occurrence of free-
threshing wheats and naked barley, as well as chick pea
and flax, in the plant assemblages from this region.
The weed taxa are also of fundamental importance, for
these too affect the overall distribution of the phases and
sites manifest in the correspondence analysis plots. In
this respect calculations of the mean number of weed
taxa represented in the site clusters are informative (ta-
ble 5). For example, for sites associated with the domestic
glume wheats and hulled barley (i.e., the upper right
quadrant), the mean number of taxa is low (4.30) in com-
parison with that of the group of sites associated with
the wild cereals (i.e., the upper left and lower left quad-
rants), which is 12.88. Interestingly, the mean number
of taxa for the group of sites associated with the free-
threshing wheats and naked barley (i.e., the lower right
quadrant) is also high (12.75). If our plots do in fact rep-
resent the spatial dynamics of the migration of peoples
with their crops to new locations beyond the Levantine
core, then the low numbers of weed taxa found in as-
sociation with the earliest domestic cereals could be ac-
counted for by the fact that cleaned grain stocks were
being transported and sown in fields recently cleared of
the local wild flora. By contrast, those areas in which it
is proposed that the cereals evolved—and in which many
of the segetal species also probably originated (Zohary
1973:647)—show a greater diversity of weeds, presuma-
bly because early fields cultivated with relatively prim-
itive techniques would have been heavily infested.
The characteristics of the weed component of assem-
blages of charred remains included with crop species as-
sociated with early farming systems have already been
proven informative about the dynamics, both anthro-
pogenic and ecological, of the spread of agriculture (see,
e.g., Bogaard 2002; Colledge 1998,2002; Hillman 2000;
Hillman et al. 2001). The significance of the numbers of
taxa and taxon diversity has been demonstrated by more
comprehensive quantitative research on the develop-
ment of weed floras through time in central Europe (Wil-
lerding 1986,Ro¨ sch 1998). These studies have shown
that the numbers of weed species found in association
with crops in archaeobotanical assemblages were lowest

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS47
in the initial stages of colonization in the Early Neo-
lithic, as our study indicates, with a significant increase
by the Late Neolithic. Willerding lists a total of 44 weed
taxa in the Early Neolithic (an increase of only 36 from
the preceding Mesolithic period) and an addition of an-
other 76 by the Middle/Late Neolithic (1986:309). Ro¨ sch
distinguishes native species (apophytes) and immigrants,
brought in by farmers (anthropochores), and finds apo-
phytes dominant in the Early Neolithic but decreasing
through time as the numbers of anthropochores increase.
He proposes directional origins for the latter and suggests
that the relative mobility of farmers and/or the inten-
sification of farming practices could explain the differ-
ences in weed floras through time (1998:121–22). More
complex analytical work is required to account for the
processes that gave rise to the patterns of weed taxa.
However, even at this early stage in our research we find
it encouraging that our data might also be used to doc-
ument the spread of the weed floras associated with the
dispersal of the earliest domestic crops.
In more general terms, we propose that the vegeta-
tional signatures that characterize the suites of plants
found on the early settlements can be accounted for in
terms of both the different regional uses of particular
elements of the exported crop package and an initial re-
duction in taxon diversity. The reduction in taxon di-
versity relates to both the ecological conditions of the
fields in the newly colonized regions and the anthro-
pogenic effect of transportation of the crop package. The
specific anthropogenic factors related to the observed
correspondence in assemblage characteristics between
the southern Levant, Cyprus, Crete, and Greece deserve
further exploration, but at this stage we tentatively pro-
pose a combination of (1) a greater investment of time
in the cultivation of crops in order to maintain stocks
once geographically isolated from their source (for ex-
ample, better field maintenance and crop-processing
techniques that resulted in “cleaner” grain stocks and
correspondingly lower weed presence in the assem-
blages) and (2) an early and continued use of domestic
cereals in preference to wild cereals. This can be con-
trasted with the situation in the northern Levant and
southeastern and central Anatolia, where at certain sites
wild cereals were used throughout the aceramic Neo-
lithic and where field maintenance and crop processing
were less fastidious, resulting in greater infestation of
weeds. These are only hypotheses, but we forward them
as one set of possible reasons for the observed patterning
in our comparative analysis of archaeobotanical assem-
blages.
Summary and Conclusion
The regionally distinctive composition of archaeobotan-
ical assemblages from aceramic Neolithic sites in south-
western Asia and southeastern Europe has been deter-
mined through multivariate analysis. The structure of the
data lends support to a hypothesis of agricultural colo-
nization of Cyprus, central Anatolia, Crete, and Greece
from a Levantine core over a period of nearly three mil-
lennia. The subsequent foundation of Neolithic settle-
ments in southeastern Europe is the beginning of a long
and complex process of colonization of new landscapes,
together with the displacement, acculturation, and/or as-
similation of indigenous European Mesolithic hunter-
gatherers.
Our findings support a model in which cereal domes-
tication first emerges in the Levantine corridor during
the PPNA of the early tenth millennium BC and is then
restricted to a few sites in the southern and central Le-
vant for the next 400–500 years. Domestic cereals appear
at approximately the same time, about 8700 BC, in
southeastern Anatolia and on Cyprus in EPPNB con-
texts. Only in southeastern Anatolia has the possibility
of in situ domestication been considered. Cyprus is no-
table for representing the first definite evidence of a tar-
geted migration by farming communities, sometime in
the early ninth millennium BC. Two or three centuries
later, by 8500 BC, central Anatolia was first settled by
agricultural colonists, probably from a northern Levan-
tine source. Finally, at the end of the eighth millennium,
agricultural colonists arrived at approximately the same
time in southern Greece and Crete, and the archaeobo-
tanical evidence suggests Cyprus and/or the Levant as
the likely source of these migrants. The archaeobotanical
assemblages in each of these regions possess distinctive
vegetational signatures that can in part be explained by
differing regional patterns in the use of particular do-
mesticates. Another major source of the patterning is the
reduction in taxon diversity, including weed taxa, as crop
packages were transported from their areas of origin.
From a methodological perspective, this paper has
demonstrated that the quantification and multivariate
analysis of properly collated archaeobotanical data, even
recorded at the level of taxonomic presence/absence, is
of sufficiently high resolution to contribute to our un-
derstanding of the earliest phases of the transition to
farming. It is with some optimism, therefore, that we
continue our documentation and comparative analysis
of archaeobotanical assemblages from early Neolithic
sites in Europe.
Comments
peter bellwood
Department of Archaeology and Natural History,
Research School of Pacific and Asian Studies,
Australian National University, Canberra, ACT
0200
,
Australia (peter.bellwood@anu.edu.au). 13 v 04
My comments are confined to the underlying claim,
made in the abstract and introduction but clearly not a
central issue for the authors, that crop packages of the
type identified can be equated with the movement of
actual colonizing farmers rather than with cultural dif-
fusion amongst hunter-gatherers. I agree wholeheartedly

S48 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
with this claim and note that genetic evidence exists
that could be argued to support it. For instance, King and
Underhill (2002) point to very significant correlations
between the distributions of painted pottery and anthro-
pomorphic figurines in the Neolithic Levant, Anatolia,
and southeastern Europe and the distribution of Y-chro-
mosome haplogroup Eu9. This correlation is said to sup-
port a hypothesis of demic diffusion, at least of males,
out of southwestern Asia. While later than the PPNB,
these relationships could nevertheless be tracking an ear-
lier population expansion.
On a more general theoretical level, how can we
equate consistent and coherent economic packages of the
type described, clearly migratory out of the Levant in
terms of the movements of the crops and weeds them-
selves, with colonizing farming groups, in this case of
PPNA and EPPNB origin? I have recently tried to answer
this question in terms of the movements of agricultural
complexes and language families in many parts of the
world (Bellwood 2004), and related issues have been dis-
cussed in depth by many authors, with many differing
shades of opinion, in Bellwood and Renfrew (2003). Some
recent debates within archaeology have tended towards
the conclusion that “packages,” “cultures,” and other
normative entities have never existed, in step with a
general trend in social anthropology to deny the exis-
tence of discrete societies. World history from this view-
point reflects an eternal process of creolization. To me
this is a little like throwing out the baby with the bath
water. Widespread and internally homogeneous com-
plexes of material culture and/or crops have existed from
time to time, as this paper shows. They reflect relatively
punctuated episodes of expansion. Migration of a pop-
ulation into a new territory, carrying cultural items
honed to ensure its very survival, is one such form of
punctuated expansion that is easy to visualize with the
hindsight of recent history in many parts of the world.
This paper takes an important stance in favour of pop-
ulation movement as a significant factor in prehistory.
The authors do not discuss their reasons in any depth,
since the focus is clearly on the crop package itself rather
than on the mechanisms behind its spread. Perhaps they
can expand on this issue in their reply.
laurent bouby
CEPAM, CNRS UMR
6130
, Sophia-Antipolis,
250
Rue
Albert Einstein,
06560
Valbonne, France (bouby@
cepam.cnrs.fr). 14 v 04
Colledge and colleagues are to be commended for their
useful review of archaeobotanical data from aceramic
Neolithic sites of southwestern Asia and southeastern
Europe in close connection with
14
C dates. Of special
interest is the chronological and spatial analysis that al-
lows them to reintroduce archaeobotanical information
into the debate about the cultural and demographic pro-
cesses underlying the spread of agriculture. Multivariate
analysis is particularly useful for investigating the tem-
poral and spatial patterning of archaeological data, and
it is probably not yet sufficiently employed in archaeo-
botany. Colledge et al. cautiously point to some limi-
tations of the archaeobotanical record due to disparities
in sampling, recovery, and analytical methodologies and
taphonomic processes. Crop-processing activities should
also be mentioned, as they directly affect the relative
proportions of crop grains, chaff, and weed seeds. I am
also convinced that the coherent geographical patterning
of the correspondence analysis results indicates that the
above-mentioned limitations do not suffice to blur the
original variability between sites. I am wondering if a
more sophisticated quantification than the simple pres-
ence/absence record would not be useful for getting a
more accurate perception of these differences. For ex-
ample, it might permit discriminating between the really
cultivated dominant taxa and other domesticated plants
that may have been transported as weeds and are there-
fore recorded in small numbers. Of course, it is very
difficult to compare sites or taxa using rough counts of
plant remains, in particular because of some of the lim-
itations described by Colledge and colleagues, but pres-
ence/absence does not eliminate all the problems (the
most numerous taxa being the most easily and fre-
quently recorded), and it necessarily involves a reduction
of information (e.g., Kadane 1988, Jones 1991). The use
of relative quantification (such as percentage) or a
semiquantitative scale of abundance may be an alter-
native solution and produces information that can easily
be used for multivariate analysis (Bouby and Marinval
2004).
The results and discussion on weeds are very inter-
esting. It is shown in the paper that weeds are of some
importance in the discrimination of geographical groups
and that they are more diversified in the areas where
cereals would have been domesticated, including those
in which wild cereals were still in use throughout the
aceramic Neolithic. The hypothesis of the reduction of
weed taxon diversity with the transport of crop packages,
correlated with a transition from weak to greater in-
vestment in field maintenance and crop-processing ac-
tivities between areas of origin of domesticated plants
and newly colonized regions, is of great interest. How-
ever, the role in weed taxon diversity assigned to the use
of wild cereals is not very clear. Perhaps a little more
attention should be paid to the fact that not all of the
herbaceous wild plants are necessarily weeds involun-
tarily collected with crops. Some of them may have been
gathered on purpose, especially in areas where people
were accustomed to collecting wild cereals and pulses
(i.e., herbaceous wild plants).
As an archaeobotanist working in the western Medi-
terranean, I have been especially intrigued by the clear
separation between the hulled cereals, peas, and lentils
that are dominant in Jordan, Greece, Cyprus, and Crete
and the free-threshing wheats, naked barley, chick peas,
and flax associated with the Damascus basin and Med-
iterranean coastal Syrian sites. In the western Mediter-
ranean a similar opposition is also documented during
the early Neolithic. While hulled wheats are dominant
in Italy and in the earliest Neolithic sites of southern

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS49
France, clearly connected with the northwestern Italian
peninsula (the Ligurian group), in the subsequent Cardial
culture and, to a lesser extent, in Spain free-threshing
wheats and naked barley play a major part (Marinval
1992, Binder et al. 1993). At the present time no clear
hypothesis is proposed to explain this separation. Will
some kind of indirect relations someday be hypothesized
between coastal Syria and the northwestern Mediterra-
nean? However this may be, we will look forward to
Colledge et al.’s future work with special interest, hoping
that results for the whole Mediterranean and Europe will
soon be available. It would be of great interest to take
into account archaeozoological data. Correspondence
analysis applied to faunal records has proved useful in
pointing to geographical patterning of sites from Greece
to the Iberian Peninsula (Vigne and Helmer 1999, Vigne
2003). Two broad areas, not exactly corresponding to
those perceived by archaeobotany, are distinguished;
Greece and southern Italy, on the one hand, are separated
from central and northern Italy, southern France, and
eastern Spain, on the other. Hunting plays a major part
in the distinction: important in the western group, it is
insignificant in Greece and southern Italy. However, in
the western region some sites seem to be connected with
the central Mediterranean and could document coloni-
zation from that area. Processes of diffusion of animal
breeding seem to be more diversified in the western part
of the Mediterranean basin than in the central and east-
ern part, implying direct colonization over rather long
distances as well as more significant acculturation of
Mesolithic people (Vigne and Helmer 1999, Vigne 2003).
It is time to reintroduce agriculture, through archaeo-
botanical results, into the debate.
julie hansen
Department of Archaeology, Boston University,
675
Commonwealth Ave., Boston, MA
02215
, U.S.A.
(jmh@bu.edu). 21 v 04
This is a fascinating paper that piques one’s curiosity for
more information. The authors have, for the first time,
utilized agricultural remains to elucidate the question of
the spread of agriculture. What a novel idea! Earlier
works have proposed models using such diverse data as
language and trade in (now invisible) goods. What better
way to address the spread of agriculture than to look at
agricultural products? Those of us who have written
about the plant remains from pre- and post-Neolithic
sites in the Eastern Mediterranean have generally tab-
ulated what was found where and when and implicitly,
if not explicitly, acknowledged a demic diffusion along
the lines of a wave of advance. The utilization of cor-
respondence analysis to examine the actual distribution
in time and space of the crops, their precursors, and weed
assemblages provides a new dimension of analysis that
proves to be extremely fruitful.
As exciting as this prospect is, it is not without its
problems, many of which the authors themselves take
pains to explicate. The first of these is the effect of var-
iability in recovery methods on the presence or absence
of some species. In particular, their comment that the
“low numbers of weed taxa found in association with
the earliest domestic cereals could be accounted for by
the fact that cleaned grain stocks were being transported
and sown in fields recently cleared of the local wild
flora,” while probably correct, needs to be qualified by
the recognition that three of the four Greek sites that
they use for the analysis were not systematically water-
sieved and produced few plant remains overall. The in-
tensive flotation programs at many of the Near Eastern
sites have no counterparts in Greece or Crete with the
possible exception of Franchthi Cave. Here, however,
stratigraphic problems inside the cave (e.g., crosscutting)
and taphonomic issues outside on paralia (e.g., wet-dry
cycles breaking up and dispersing carbonized plant re-
mains) make the Early Neolithic material from this site
both scarce (nonexistent in paralia samples) and without
clear contexts.
A second problem that is not fully discussed is that,
although only presence/absence of taxa on the sites is
considered, the contexts from which those taxa were re-
covered have a significant effect on what is or is not
present. At Franchthi Cave no specific contexts can be
identified for the material from Zone VI, which appears
to be scatter from hearths and general debris on the cave
floor and fill. It is unlikely that this represents the full
range of botanical material that might have been present
on the site, especially since the primary living area may
have been outside the cave at this time. Samples from
Ghediki, Argissa, and Knossos come from fairly small
areas of the site and are equally unlikely to be represen-
tative of all the contexts. Thus, the absence of weed flora
from these sites may be a function of contexts sampled
as much as agricultural practices.
Another issue that needs clarification is the assign-
ment of botanical material to a specific phase on the
basis of calibrated radiocarbon dates rather than the as-
sociated material culture that defines the phase. I have
struggled with the Near East chronology for decades try-
ing to determine a clear and accurate breakdown of the
periods and phases. This is a daunting task, given the
very large number of dates (in general a good thing) and
the diverse nomenclature used for the various phases in
different areas and by different researchers, and it is one
best left to those more familiar with both the archaeo-
logical details of each site and the intricacies of radio-
carbon dating. At Jericho Colledge et al. have moved the
plant remains from the PPNA to the EPPNB on the basis
of the radiocarbon chronology, despite the fact that they
were found in a PPNA stratum. While I applaud their
attempts to grapple with the Near Eastern chronology, I
cannot accept that material that is in an otherwise
clearly PPNA context should be assigned to the EPPNB
because the radiocarbon dates are later than PPNA dates
at Aswad and Iraq ed-Dubb. We define cultural phases
on the basis of the material culture contained in the
strata. If there is a later date for these strata, then there
should be a longer chronology for the phase or a slightly
different chronology for Jericho than for Aswad or Iraq

S50 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
ed-Dubb. Perhaps there is a more detailed explanation
in the cited article by Colledge (2004), which I have not
yet seen.
The database being compiled and the further analyses
to come hold tremendous promise for our understanding
of the origins and spread of agriculture, as well as for
many other questions palaeoethnobotanists have been
trying to address, and I for one am extremely grateful for
these researchers’ efforts.
david r. harris
Institute of Archaeology, University College London,
31-34
Gordon Square, London WC
1
H
0
PY, U.K.
(david.harris@ucl.ac.uk). 20 v 04
Ever since Grahame Clark illustrated, with the radio-
carbon dates available to him in 1965, the Neolithic
spread of agriculture to and through Europe from the
Near East that Gordon Childe had earlier postulated,
there has been a need to test the hypothesis against ar-
chaeobotanical data. Archaeobotany was only a fledgling
subdiscipline in the 1960s, and when it did come to be
practiced more widely and systematically in Europe in
the 1970s and 1980s most specialists confined their at-
tention to local sites and areas within national bound-
aries. Even as recently as 1998, a resolution passed at a
conference in Venice on the Neolithic transition in Eu-
rope recommended that an AMS radiocarbon-dating pro-
gram of “well-provenienced and well-identified samples
of...plantremains of wheat, barley, and rye” from early
Neolithic sites in Europe and western Asia should be
undertaken (Ammerman and Biagi 2003:343). Now, al-
most 40 years after Clark’s paper appeared, Colledge,
Conolly, and Shennan have taken the long-awaited first
step in that direction by producing a critical analysis and
synthesis of the archaeobotanical data now available
from aceramic sites in the Levantine–east Mediterranean
region. Their very welcome paper marks a turning point
in the protracted debate about how agriculture spread
into and across Europe during the early Neolithic. It
shows that, despite the many limitations of the archaeo-
botanical record, comparative, multivariate analysis of
the published data on the plant assemblages and radio-
carbon dates for the principal pre- and early Neolithic
sites in the region can yield new, well-founded interpre-
tations of the probable routes of agricultural spread and
the nature of early Neolithic agricultural practices.
A particular strength of their analysis is that it is not
limited to the founder crops of the early Neolithic (ein-
korn and emmer wheat, hulled barley, lentils, peas, chick
peas, bitter vetch, and flax) but also includes many wild
and weedy taxa. The latter have tended to be overlooked
in interpretations of the spread of the so-called Neolithic
crop package, but their incorporation into the large re-
lational database that Colledge et al. are now creating
allows much closer examination of the role of weeds and
wild plants in the earliest agricultural systems of south-
western Asia and Europe. The different ecological re-
quirements of weedy taxa are a potentially powerful
means of differentiating between, for example, systems
of cultivation on lowland alluvial soils and on thinner,
less fertile upland soils and determining whether irri-
gation was practiced.
The value of including weedy taxa in their analysis is
also demonstrated by the authors’ working hypothesis
of two contrasted routes by which agriculture may have
reached Europe: the southern Levantine-Cypriot-Aegean
coastal route and the northern Levant–southeast-central-
Anatolian inland route. This distinction helps to focus
attention on what are likely to have been key areas and
environments in the dispersal of crops and weeds west-
ward from the Levant, and it should stimulate targeted
prospecting for early Neolithic sites, for example, along
the southern Turkish coast.
Colledge et al.’s discussion of the possible significance
of the contrasts between the domestic and especially the
weed floras of the southern Levant–Aegean and the
northern Levant–Anatolian routes prompts me to won-
der whether the lower weed diversity they report for the
southern route, beyond the Levantine core area, might
reflect both environmental and agrarian differences with
the northern one. Did the PPNB agricultural package
spread (probably by demic diffusion) west to the Aegean
mainly by means of the episodic selection of patches of
seasonally watered alluvial soils along river channels
that would naturally have supported ephemeral or at
least sparse vegetation, requiring less clearance and less
susceptible to weed infestation than upland areas with
denser and more diverse vegetation? Could the increase
in weed diversity later in the Neolithic on which Col-
ledge et al. comment reflect the extension of agriculture
from small patches of alluvial soil cultivated initially
into larger, ecologically more diverse areas upslope, per-
haps associated with the development of systems of
shifting cultivation?
These may be speculations that take us too far from
the available evidence, but they illustrate the great po-
tential value—of which this paper is the first fruit—of
the uniquely comprehensive and interactive archaeo-
botanical database now being created.
konstantinos kotsakis
Department of Archaeology, Aristotle University of
Thessaloniki, GR-
5400
6Thessaloniki, Greece
(kotsakis@hist.auth.gr). 17 v 04
The authors of this paper are to be congratulated for
bringing together the archaeobotanical evidence from the
Eastern Mediterranean in a comprehensive synthesis.
There is a lot of merit in an approach dealing with the
large-scale, despite the obvious fact that the finer details
are, by definition, given less attention. If nothing else,
large-scale syntheses offer a sense of historical process
and are necessary for integrating local histories and ar-
chaeologies into one apparently seamless narrative. It
follows that there are two domains which, in a broader
discussion, deserve our attention: on the level of the ev-
idence, the integrity and comparability of the facts, and

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS51
on the level of interpretation, the semantics of our an-
alytical concepts.
I will discuss the second domain first. Of course, prob-
lems of recognition and definition aside, “domesticated”
species are a biological entity, but as an element of what
we understand by the term “agriculture” domesticates
are predominantly a medium of human agency. The
problem is that, as such, they transcend the strict bio-
logical definition: they can be manipulated, preferred, or
neglected, and their presence/absence becomes inextri-
cably linked to the social domain, with all its referents.
The transition to agriculture thus becomes a historically
situated process in which “domesticates” as an analyt-
ical category eventually lose the essentialist content that
was borrowed from their biological meaning in the first
place. Obviously, this remark points towards the small-
scale and the local and builds a discussion around the
ontology of agriculture or—to use more philosophical
terms—around the distinction between the essence and
the existence of agriculture. Significant variability of the
“package” of domesticates throughout the Neolithic can
be traced in the archaeobotanical samples from Greece,
testifying to the active modification of the package in
particular localities and regions (Valamoti and Kotsakis
n.d.).
The second issue is the reliability of the data, and it
becomes more critical once we try to move away from
the integrative approach adopted in the article. I can only
comment, however, for the samples from Greece. With
the exception of Franchthi and in part Knossos, these
early Neolithic archaeobotanical samples were collected
under variable and often unspecified conditions. All the
excavations included in the database were concluded by
the early 1970s. The Thessalian samples in particular
were collected in the 1960s without any water-sieving,
relying on the hand-picking of visible charred plant re-
mains, and information on their archaeological context
is usually scant to nonexistent. For example, Gediki, a
rescue dig carried out in 1962, was limited to a trench
covering an area of no more than 2m
2
in an undefined
context. Even for the well-studied site of Franchthi, the
evidence for domesticated plants in the earliest Neo-
lithic levels is restricted to 27 specimens of emmer, 8of
lentil, and 5of barley. These observations might be
enough to cast serious doubt on the reliability of the few
samples available, but it can be argued that, unfortunate
as it may be, they represent all we have. In any case, it
should be kept in mind that the criteria for the identi-
fication of taxa used in the 1960s or even the 1970s were
by no means always comparable to those of today. Con-
sequently, the value of these samples as representative
of the earliest agriculture of Greece and the relative sig-
nificance of the various species supporting a general ar-
gument of colonization are at least questionable.
The discussion of weed taxa further highlights the sig-
nificance of sampling error. Because systematic sampling
with graded sieves was not applied in those early exca-
vations, the particularly low mean number of weed taxa
in Greece could well be the result of the failure to collect
weeds by hand-picking rather than the sowing of cleaned
grain stocks transported from the Levantine core in fields
cleared of the local wild vegetation.
However, in a way this final remark brings the dis-
cussion back to the ontology of agriculture and the role
of “domesticates.” As is very clearly presented in the
paper, even on this interregional scale the variability of
practice is remarkable: it can be observed in the degree
of preference for wild cereals, variable field maintenance
practices, and differences in crop-processing techniques.
Other areas of variability can reasonably be assumed to
exist on the basis of evidence from subsequent Neolithic
phases (Halstead 1994,1996). They include different
combinations of species, regionally and culturally spe-
cific, and plants that are not part of the normal package.
If these practices can vary so much, would it not be
equally helpful to explore more closely the existence of
diverse agricultures rather than derivation from a pos-
tulated original essence?
mehmet o¨ zdog˘an
Edebiyat Fakultesi, Prehistorya Anabilim Dali,
Istanbul University,
34459
Istanbul, Turkey
(mozdo@atlas.net.tr). 23 v 04
The origin and spread of a Neolithic way of life is prob-
ably one of the most discussed issues in prehistoric ar-
chaeology. Problems ranging from the essential ques-
tions of where, when, and how farming began to detailed
auxiliary topics have been rigorously discussed at both
the theoretical and the empirical level. For more than
half a century the expansion of the Neolithic way of life
has been at the center of all this discussion. The models
suggested have taken extreme views, with neither side
providing convincing evidence. In the first years of re-
search there was only a diffusionist model, suggesting
massive and organized migration of farmers colonizing
and thus bringing “civilization” to other lands. Later
there was a strong antidiffusionist trend, implying au-
tochthonous or parallel developments and the emergence
of farming in Europe independent of the Near East. Still
later other models suggested acculturation, culture con-
tact, a wave of advance, moving frontiers, etc. Theoret-
ical biases that obscure the hard evidence have tended
to hinder the development of acceptable solutions. In
this respect, Colledge et al.’s paper is most welcome in
focusing on a new facet of the problem. The research
design is simple but reasonable; rather than confine
themselves to the details of a site, a species, or a mic-
roregion, the researchers have adopted a supraregional
perspective. It may be argued that this approach is apt
to overlook or misinterpret certain facts because of its
geographic scope, but its advantages in drawing a mean-
ingful picture without becoming trapped in details are
evident.
The article includes a large number of ideas that can
be considered new or at least different from the conven-
tional ones. Among these is Perle` s’s maritime and/or
coastal expansion model. An unbiased reassessment of
the evidence strongly implies that there were multiple

S52 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
paths in the westward movement of the Neolithic way
of life. In identifying trajectories in the expansion of this
model, the coastal option must certainly be taken into
account, along with multiple routes along the plateau.
This idea of Childe’s was, needless to say, almost totally
ignored during the era of antidiffusionism.
There are, however, some points that need to be re-
considered if not revised. The aim here is to identify
“vegetational signatures” on the basis of “similar pat-
terns of taxon presence.” This would work, however,
only if there was an organized movement of people from
point A to point B carrying with them the original pack-
age of Neolithic components. In the present state of the
evidence, it is more or less clear that the westward en-
demic movement of “eastern” farmers was not an or-
ganized one but more like an infiltration from all parts
of the core to all parts of the new area. The original
components of the Neolithic package did not move to-
gether, and, accordingly, sites in the newly settled area
will reveal different random collections of elements
found somewhere in the core area. This would explain
what Colledge et al. point to as the “unevenness of the
evidence” on origins and diffusion.
A number of minor errors need to be corrected, among
them the comparison of the Central Anatolian assem-
blages with the PPNB sites of the Levant. The planning
and organization of settlements are among the most
striking differences between the Levant and Central An-
atolia, and naviform cores are absent in Central Anatolia
except at Kaletepe, where they were produced solely for
export to the Levant. Thus, despite the close interaction
between the Levant and the Central Plateau, there are
significant differences between them.
Overall, I share most of Colledge et al.’s views and
appreciate their approach, but it is evident that botanical
evidence alone is not enough to produce a general pic-
ture. One of the important biases of Neolithic archae-
ology is its overemphasis on subsistence patterns at the
expense of cultural traits.
edgar peltenburg
Department of Archaeology, University of Edinburgh,
Old High School, Edinburgh EH
11
LT, U.K.
(ejpeltenburg@aol.com). 20 v 04
A major issue in appraisals of the initial spread of farming
is the need for more sound empirical data that bridge the
regions of the Near East and Europe. In assembling a
radiocarbon-based, innovative and large-scale analysis of
critical archaeobotanical evidence, Colledge et al. over-
come significant weaknesses in current debates. Their
methodology and evaluation will play a significant role
in ongoing discussions about Neolithic dispersals.
One relevant complexity that emerges from their anal-
ysis concerns weed taxa. They propose, on the one hand,
that low mean numbers reflect transport of cleaned grain
stocks to secondary locales and, on the other, that cereal
domestication first emerged in the southern Levant,
which has low numbers. Reliance is placed on dates from
three sites, but those from Jericho are equivalent to rel-
evant dates from C¸ ayo¨nu¨ in the north, and Danielle Stor-
deur’s recent reinvestigations of Tell Aswad (personal
communication) raise serious questions about the exis-
tence of any Early PPNB there. This leaves the anoma-
lously early dates from Iraq ed-Dubb to sustain the south-
ern Levant primacy argument, so further evidence is
clearly required.
The primacy argument bears on the issue of routes
and mechanics of dispersal. If there must be some lin-
gering doubt about South Levantine primacy, then the
dates of domestic seeds from Mylouthkia in Cyprus are
roughly as early as the secure dates of domestic founder
crops on the mainland. (I am using Colledge et al.’s
dates rather than the lower OxCal v3.5calibrated dates
that provided a framework for discussion in Peltenburg
2003.) It is often presumed that this means that there
must be earlier, undiscovered agricultural sites on the
mainland but they probably lie outside the moderately
well-known Levantine corridor, closer to the island. In
other words, in situ domestication may be polycentric.
Alternatively, we may need to revisit Binford’s mar-
ginal-zone hypothesis, which sees human groups arti-
ficially producing the stands of grain that characterized
optimal zones where wild founder crops were abundant.
In that case, we may need to think of cultivators spread-
ing plants and animals which, by dint of close man-
agement, developed the morphological traits of domes-
tication. Although I still believe that farmers migrated
to Cyprus, these ambiguities in our empirical evidence
are not addressed in this article.
Focus on the archaeobotanical data to the exclusion
of other, associated evidence inevitably provides a lim-
ited perspective on general developments. For example,
the Levant/Cyprus–Crete/Greece route for farmers needs
to take account of disparities between the regions. These
disparities suggest that Cypriot migrants were unlikely
to be involved, since they eschewed pottery and failed
to make a success of cattle, two features that figure
prominently in the earliest Cretan Neolithic. The main-
land of the Levant also presents problems, since, al-
though Colledge et al. refer to “Mediterranean coastal
Syrian sites,” only Ras Shamra is cited, and it is too late
to provide evidence for the proposed process. These fac-
ets suggest a more complex picture in which we still
cannot rule out Anatolia as the main source for Aegean
domesticates.
Recent evidence from the Aegean points to a more
general concern with Colledge et al.’s stance. Through-
out there seems to be a premise that similar vegetational
signatures may be equated with the migration of farmers,
without consideration of other possibilities. Yet, Cyclops
Cave on Youra Island has yielded evidence for the do-
mestication of pig and caprid in Mesolithic times (Samp-
son and Katsarou n.d.). If sustained, it would be impor-
tant for indigenous arguments. Together with the
evidence from Cyprus, which was stocked with mor-
phologically wild animals (Horwitz, Tchernov, and
Hongo 2004), it suggests that foragers and cultivators
may have proactively and strategically widened their

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS53
subsistence base, and therefore we cannot exclude their
appropriation of cereal farming. Colledge et al. have pro-
vided a remarkably detailed insight into patterns of ag-
ricultural expansion. It now remains to look more closely
at individual regions and to assess indigenous appropri-
ations and migrations as long-term historically consti-
tuted processes with successes, reversals, and mixed
adaptive adjustments between farmers and hunter-gath-
erers—a modern distinction that may prove to have been
blurred in the Eastern Mediterranean during these early
times.
george willcox
Arche´ orient, Centre National de la Recherche
Scientifique, Jale` s, Berrias,
07460
France (g.willcox@
wanadoo.fr). 13 v 04
Colledge and colleagues make a significant contribution
to our understanding of early agriculture. Their article
presents data obtained from a new archaeobotanical
database created with the intention of providing a better
understanding of the emergence of agriculture in the
Near East and its subsequent diffusion into Europe. One
cannot but admire the quantity and the quality of the
meticulous work which went into the construction of
the database.
Applying multivariate analyses to information ob-
tained from the database, Colledge et al. lend support
to the claim that the origins of crop domestication are
to be found in the southern and central Levant core
area, from which crops spread to Cyprus and then to
Crete and Greece on the one hand and into southeastern
and then central Anatolia on the other. The interpre-
tation of a core area is partly based on the earliest do-
mestication reported at two sites (Aswad and Iraq ed-
Dubb), but there are no direct AMS dates on the
domesticated cereals themselves (I have recently ar-
ranged for the emmer grains from the original Aswad
excavations to be AMS-dated and am awaiting the re-
sults). The plant remains from the early layers at these
sites were few and of poor quality. Two other PPNA
sites in the south, Zahrat Adh-Dhra 2(Edwards et al.
2002) and Netiv Hagdug, which have excellent, well-
dated plant remains, show as in the north no signs of
domestication. Many archaeologists in recent years
have abandoned the concept of centres or core areas; at
the Fourth International Congress on the Archaeology
of the Ancient Near East in Berlin, March 29–April 3,
2004, a workshop entitled “Towards New Frameworks:
Supra-Regional Concepts in Near Eastern Neolithiza-
tion,” attended by a number of distinguished scholars,
produced a consensus in favour of a polycentric ap-
proach. I have argued that across the entire region local
cereals occurring near the sites were taken into culti-
vation during the PPNA and early PPNB, leading to
independent domestication events (Willcox 2002:136,
table 1). This is demonstrated by the fact that the ar-
chaeobotanical cereal assemblages correspond to the
differences in the natural distributions of the wild ce-
reals; for example, single-grained einkorn is found as a
dominant cereal only in the north, while emmer dom-
inates together with barley in the south. DNA data and
a much more detailed analysis of the distribution of
wild cereals (the traditional maps are misleading) sup-
port this hypothesis (Ishii, Mori, and Ogihara 2001:902,
fig. 3). Concerning the introduction of agriculture into
Cyprus, both emmer (southern Levant) and single-
grained einkorn (northern Levant) are found on early
PPNB Cypriot sites. Given the time scales involved and
the complexity of the region, the hypothesis of migra-
tion from a single core area is less plausible than a
model which involves multiple contacts from different
areas at different times. The generally poor preservation
of archaeobotanical material on Cypriot sites and the
fact that the evidence comes primarily from wells may
have contributed to the small number of weed taxa (ta-
ble 5) compared with sites on the dry continental
steppe, where the preservation is often excellent.
Colledge et al. suggest, and I agree, that the stable
climatic conditions following the Younger Dryas pro-
vided a favourable situation for the first farming com-
munities. But for the length of the gestation period for
domestication, they suggest that the data agree with
the model proposed by Hillman and Davies (1992), in
which plants, once taken into cultivation, were rapidly
domesticated. While this may be true in one area, one
should not assume that it was the case for the region
as whole (for further arguments for slow domestication
rates see Willcox 1999:487 and 2004:149).
Archaeobotanical data vary tremendously according to
preservation conditions, sampling methods, and archae-
ological context. Despite this, the “vegetational signa-
tures” described by Colledge et al. for the earliest periods
are real and result from the fact that the sites are located
in very different vegetation zones with different cereal
assemblages. With regard to the data, we are presented
only with the correspondence analysis plots and denied
the raw data. It would also be interesting to define the
geographical locations of the sites using natural ecolog-
ical boundaries or even the new cultural areas now
emerging as opposed to the modern political frontiers
used here, where, for example, Jordan and Israel occupy
the same as well as different vegetation zones while
northern Syria and southeast Anatolia are part of the
same natural and cultural region.
Colledge et al. should be congratulated for their con-
tribution to our knowledge of the origins and spread of
crop plants. However, a central question arises from their
work concerning the relationship of the spread of crops
to cultural diffusion on the one hand and to population
movement on the other. Future work combining ar-
chaeobotanical information from the database with
other archaeological data will further our knowledge
about this fascinating period in human history.

S54 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
Reply
sue colledge, james conolly, and
stephen shennan
London, U.K. 8vi04
We thank all the reviewers for their insightful com-
ments. We are encouraged that, on the whole, they too
see value in a pan-regional approach to archaeobotanical
data, and we look forward to the second phase of our
comparative research on southeastern and central Eu-
rope. We trust that the following response addresses their
concerns with the current paper.
At a general level, Kotsakis usefully highlights the dif-
ferences between the study of domestication from a bo-
tanical and zoological perspective (i.e., its “essence”) and
domestication as a historical and social construct (i.e.,
its “existence”). We have concentrated here on the spa-
tial-temporal dynamics of the former as a route to un-
derstanding the latter. By noting regional similarities in
the composition of archaeobotanical remains, especially
between the earliest Greek Neolithic sites and those
from Cyprus and the Levant, and by cautiously rejecting
environmental and taphonomic processes as the cause,
we have in fact highlighted how blanket “wave-of-ad-
vance” models fail to explain how agriculture arrived in
Europe. In this regard, O
¨zdog˘an takes issue with our use
of the heuristic device of “vegetational signatures” to
describe the regionally coherent patterning of sites on
the basis of their archaeobotanical remains. Although it
is true that the full suite of founder crops did not nec-
essarily move as a “package,” we suggest that the reasons
that regionally distinctive clusters of sites exist is in part
shared ancestry. We are certainly not rejecting central
Anatolia’s contribution to early Neolithic Greece (and,
by extension, southeastern Europe), only pointing out
that this should not be the default point of origin of
Europe’s first farmers as is often assumed (e.g., van Andel
and Runnels 1995). If anything, given that we show that
early agriculture in the eastern Mediterranean is region-
ally variable, O
¨zdog˘an, Kotsakis, and Peltenburg should
be reassured that what we are postulating is not the
spread of an agricultural “essence” from a single point
of origin but the diffusion and evolution of variable but
historically contingent practices. We wholeheartedly
agree with Kotsakis that domestication and domesticates
are a “medium of human agency” and that the docu-
mented variability in the use of certain species and ag-
ricultural practices requires further investigation. Har-
ris’s comments on the changing ecological niches
exploited by early farmers that may give rise to the pat-
terns are also useful in this context and merit further
investigation.
More pragmatically, Hansen and Kotsakis note the dif-
ferences in the methods of recovery and processing of
the samples and the possible correlation with taxon var-
iability. We have made reference to this in our text, but
we will elaborate. Recovery of charred plant remains was
by flotation for 56 of the 60 phases in our study. For 30
of these bucket flotation was used, and a flotation ma-
chine was employed for a further 17. For 9of the phases
there was no record of the method of flotation. We could
find no record of what recovery method had been used
for only two sites (Gediki and Knossos), and for an
equally small number the plant material was retrieved
using dry sieving (Jericho and Nahal Hemar). However,
these latter sites do not form outliers from those that
were subject to more rigorous recovery and in fact main-
tain their regional groupings. As far as we are able to
discern, therefore, the inclusion of a minority of sites
with poor recovery methods has not influenced the gen-
eral patterns we have identified.
It is worth noting that at several sites there are ar-
chaeobotanical reports by more than one researcher (e.g.,
Beidha [Helbaek 1966, Colledge 2001], Tell Mureybit
[van Zeist and Bakker-Heeres 1984, Willcox and Fornite
1999], Argissa Magoula [Hopf 1962; Kroll 1981,1983],
Sesklo [Renfrew 1966; Kroll 1981,1983], and others). In
these cases the different researchers have often used dif-
ferent methods of recovery. For example, at Beidha, Hel-
baek examined impressions in burnt plaster/daub, and
during later excavations at the site flotation of some of
the ashy layers and hearths revealed additional taxa not
previously identified. Similarly, at Argissa Magoula Hopf
(1962) listed taxa found in baked daub/clay, and ca. ten
years later Kroll (1981,1983) revisited the site and took
samples for flotation from the early layers. Our records
represent amalgamations of the data from all sets of
researchers.
We also acknowledge that differences in mesh sizes
used to retrieve the plant remains will influence the
overall composition of the archaeobotanical samples. As
far as we can establish, for 31 of the 60 phases there are
no records of sieve sizes, and on those for which records
exist 22 used sieves with meshes !1mm. Of relevance
to our plot is the fact that a majority of these 22 phases
are located in the quadrant with the lowest mean number
of weed taxa (i.e., equivalent to sites in the southern
Levant, Cyprus, Crete, and Greece). Four of the 7phases
with mesh sizes ≥1mm are also in this quadrant; the
other 3are to the left of the plot (upper and lower left
quadrants), correlating with sites that had the highest
mean number of weed taxa. It does not seem to be the
case that recovery methods were responsible for the low
number of weed taxa in the islands and Greece, as there
appears to be little correlation between number of weed
taxa recovered and the size of mesh used.
Hansen, Kotsakis, and Willcox also suggest that the
different contexts sampled and/or lack of contextual in-
formation about samples at the sites may undermine the
case we have made for patterning on the basis of taxo-
nomic composition. We have not overlooked this pos-
sibility, pointing out that the range of context types sam-
pled may well determine the degree to which the suite
of crops and wild taxa is representative of the full com-
plement originally present at the sites. For 17 of the 60
phases used in our analysis there is no information on
context, and a further 12 phases rely on samples taken

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS55
from sections and/or balks and three from unspecific
“ashy layers.” This leaves a total of 28 phases for which
adequate data are available and/or well-defined contexts
were sampled. Only in these cases would it have been
possible to make some assessment of variation of taxa
between the different context types. However, our
method of recording was not at the level of individual
samples, and we have used presence/absence rather than
absolute numbers in our analyses. This has enabled us
to sidestep instances in which, for example, we would
have had to compare the contents of burnt storage con-
tainers with general occupation deposits, which would
have increased the chances of skewing the data. Instead
we have combined the taxonomic data from the sites
(e.g., total numbers of taxa and ubiquity scores) according
to major cultural phases. The use of presence/absence
data precludes the likelihood that any patterning is the
result of contextually determined aberrantly high or low
proportions of taxa. It has also allowed us to include 8
phases for which records of numbers of taxa were not
given, only references to identifications (with or without
ubiquity). Bouby rightly points out that this method has
necessarily involved a reduction of information that oth-
erwise might have enabled us to determine, for example,
the relative significance of crops at the different sites.
We are fully aware that this is the case. As a conse-
quence, however, we have avoided the pitfalls of making
the unjustified assumption that numerical values some-
how reflect importance (Jones 1991:63). The very fact
that we have relied on presence/absence data alone has
enabled us to make comparisons unimpeded by quan-
titative differences derived from any context-related var-
iations (in turn a consequence of, for example, processing
activities) in numbers of taxa. We agree that an approach
based on relative/semi-quantitative assessment of the
data may permit a more in-depth study of the develop-
ment of plant-based subsistence/agriculture, and in sub-
sequent work we have already used ubiquity scores to
explore differences in crop and weed taxa between sites.
We agree with Bouby that wild or weed species are
likely to be transported onto sites for many different
reasons, including although not exclusively as contam-
inants of crop harvests. We did not refer explicitly to
crop processing as a cause of the possible “filtering out”
of wild or weed taxa but agree that this is undoubtedly
responsible in part for the presence or absence of certain
taxa from the sites in our plots. We have emphasized
that a more in-depth examination would be necessary to
establish the possible influence of taphonomy on the
composition of samples from the different sites and thus
on the overall data set.
Willcox questions the suitability of some of our data
because of the poor quality of the plant material; he men-
tions this in reference to the evidence from Levantine
PPNA sites and from Cypriot early PPN contexts. We
are aware of the problems with these samples and others
and acknowledge the limitations of our method of anal-
ysis as a result of not being able to make allowances for
variables such as preservation. However, records of the
relative state of preservation of taxa (e.g., distortion, frag-
mentation, etc.) are patchy, and when they occur the
descriptions are mostly subjective, making comparison
between sites on this basis impossible. In the same way,
therefore, that we have not made judgements about the
accuracy of identifications of taxa (i.e., based on the skill
levels of the researchers who made the identifications),
we have not felt able, in the absence of adequate infor-
mation about their overall condition, to include or ex-
clude certain taxa/samples from our analyses (the ex-
ceptions being cases in which the researchers comment
that plant items are definite contaminants). Willcox
quotes the Cypriot sites at which wells were sampled
and suggests that poor preservation/context type may be
the reason few weed taxa were recovered. The Cypriot
sites, including those where hearths, occupation depos-
its, floors, and burials were sampled, are all located in
the top right quadrant of our plot. If the sites where
samples were taken from wells (phases KMIA, KMIB,
PSIII) had formed discrete clusters or outliers in the plot
(and distinct from the other Cypriot sites), there would
perhaps have been some foundation for suspecting that
extraordinary circumstances were responsible for the
taxonomic differences.
Hansen and Peltenburg raise issues of phasing and
chronology. We recognize that Near Eastern chronologies
are complex, and therefore we have made every effort to
consult as many sources as possible to establish the cor-
rect dating sequences for the phases. This has included
reference to publications that cite potential errors in
radiocarbon dates, whether resulting from the dating
techniques or from incorrectly reported stratigraphic as-
sociation between dated samples and archaeological lev-
els. Hansen says, “I cannot accept that material that is
in an otherwise clearly PPNA context should be assigned
to the EPPNB because the dates are later than PPNA
dates at Aswad and Iraq ed-Dubb,” but we have not sug-
gested that the PPNA cultural affinity is incorrect at
Jericho. We have merely said that it is chronologically
equivalent to the EPPNB elsewhere in the Levant. The
presence of domestic cereals at Jericho in PPNA contexts
does not mean that they are equivalent in date to the
PPNA cereals at Aswad and Iraq ed-Dubb. Stordeur’s
reinvestigations at Tell Aswad, referred to by Peltenburg,
are as yet unpublished, and we therefore have not been
able to incorporate these data into our analysis. We un-
derstand that the PPNA contexts are under review at
Aswad, but until this information is published we have
used the established evidence (van Zeist and Bakker-
Heeres 1982) that shows cereal domesticates in mid-
tenth-millennium-bp contexts.
Finally, Bellwood invites us to comment on why we
have here favoured a model of population movement to
account for the initial spread of agriculture. Although our
work is part of a wider trend toward revisiting the role of
population dynamics in the spread of agriculture (e.g., Bell-
wood 2004, Bellwood and Renfrew 2003, Bentley et al.
2002, Shennan 2002), we rely on population movement
as a primary factor in the initial spread of agriculture only
because the evidence strongly favours demic diffusion for
this phase of the early Neolithization. We are, of course,

S56 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
not the only group promoting demic diffusion for the
spread of agriculture in southeastern Europe; in her com-
prehensive review of the origins of the Greek Neolithic
Perle`s (2001), for example, refers to demic diffusion as the
“inescapable hypothesis.” We also see population move-
ment and displacement of hunter-gatherers as the primary
mechanism underlying the initial spread of agriculture
into central Europe, with cultural diffusion occurring only
in parts of the west-central, southwest, and northwest of
the continent (Bentley et al. 2002, Zvelebil 2000). What
is interesting about this is that, as Bellwood shows (2004),
the rapid dispersal of early farmers in the eastern Medi-
terranean and southeastern Europe has many cross-cul-
tural parallels. The wider recognition of this may stim-
ulate further research into the underlying socio-demo-
graphic dynamics of early agricultural societies and the
development of models to explicate their expansionist
tendencies.
References Cited
ammerman, a. j., and p. biagi. Editors. 2003.The wid-
ening harvest: The Neolithic transition in Europe, looking
back, looking forward. Boston: Archaeological Institute of
America. [drh]
ammerman, a. j., and l. l. cavalli-sforza. 1973.
“A population model for the diffusion of early farming in Eu-
rope,” in The explanation of culture change: Models in prehis-
tory. Edited by C. Renfrew, pp. 343–57. London: Duckworth.
andreou, s., m. fotiadis, and k. kotsakis. 1996.
Review of Aegean prehistory 5: The neolithic and bronze age
of northern Greece. American Journal of Archaeology 10:
537–97.
asouti, e., and a. fairburn. 2002. “Subsistence econ-
omy in Central Anatolia during the neolithic: The archaeobo-
tanical evidence,” in The neolithic of Central Anatolia: Inter-
nal developments and external relations during the
9
th–
6
th
millennia BC. Edited by F. Ge´rard and L. Thissen, pp. 181–92.
Istanbul: EGE Yayınları.
b a i r d , d . 2002. “Early Holocene settlement in Central Anato-
lia: Problems and prospects as seen from the Konya Plain,” in
The neolithic of Central Anatolia: Internal developments and
external relations during the
9
th–
6
th millennia BC. Edited by
F. Ge´ rard and L. Thissen, pp. 139–59. Istanbul: EGE Yayınları.
balkan-atli, n. 1994.La Ne´ olithisation de l’Anatolie. Paris:
Institute Franc¸ais d’E
´tudes Anatoliennes/De Boccard.
balkan-atli, n., and d. binder. 2000. L’atelier ne´oli-
thique de Ko¨mu¨ rcu¨ -Kaletepe: Fouilles de 1999.Anatolia Anti-
qua 8:199–214.
bar-yosef, o., and a. gopher. 1997. “Chapter 10: Dis-
cussion,” in An early Neolithic village in the Jordan Valley.
Edited by O. Bar-Yosef and A. Gopher, pp. 247–66. (American
School of Prehistoric Research Bulletin 43.) Cambridge: Har-
vard University, Peabody Museum of Archaeology and
Ethnology.
bar-yosef, o., and r. s. kra. 1994. “Dating eastern
Mediterranean sequences: Introductory remarks,” in Late Qua-
ternary chronology and palaeoclimates of the Eastern Mediter-
ranean. Edited by O. Bar-Yosef and R. S. Kra, pp. 1–12. Tucson:
Radiocarbon/Cambridge: American School of Prehistoric
Research.
bellwood, p. 2004.First farmers: The origins of agricultural
societies. Oxford: Blackwell. [pb]
bellwood, p., and c. renfrew. Editors.2003.Examining
the farming/language dispersal hypothesis. Cambridge: Mc-
Donald Institute for Archaeological Research. [pb]
bentley, a. r., t. d. price, j. lu¨ ning, d. gronen-
born, j. wahl, and p. d. fullagar. 2002. Prehis-
toric migration in Europe: Strontium isotope analysis of Early
Neolithic skeletons. current anthropology 43:799–803.
bic¸ akc¸ i , e . 1998. “An essay on the chronology of the Pre-pot-
tery Neolithic settlements of the East-Taurus Region (Tur-
key),” in Light on top of the black hill: Studies presented to
Halat C¸ ambel. Edited by G. Arsebu¨ k, M. J. Mellink, and W.
Schirmer, pp. 137–50. Istanbul: EGE Yayınları.
binder, d. 2002. “Stones making sense: What obsidian could
tell about the origins of the Central Anatolian neolithic,” in
The neolithic of Central Anatolia: Internal developments and
external relations during the
9
th–
6
th millennia BC. Edited by
F. Ge´ rard and L. Thissen, pp. 79–90. Istanbul: EGE Yayınları.
binder, d., j. e. brochier, h. duday, d. helmer,
p. marinval, s. thiebault, and j. wattez. 1993.
L’abri Pendimoun a` Castellar (Alpes-Maritimes): Nouvelles
donne´es sur le complexe culturel de la ce´ ramique imprime´e
me´diterrane´ enne dans son contexte stratigraphique. Gallia Pre´-
histoire 35:177–251.[lb]
bogaard, a. 2002. Questioning the relevance of shifting cul-
tivation to Neolithic farming in the loess belt of Europe: Evi-
dence from the Hambach Forest experiment. Vegetation His-
tory and Archaeobotany 11:155–68.
broodbank, c., and t. strasser. 1991. Migrant farm-
ers and the Neolithic colonization of Crete. Antiquity 65:
233–45.
bouby, l., and p. marinval. 2004. Fruits and seeds from
Roman cremations in Limagne (Massif Central) and the spatial
variability of plant offerings in France. Journal of Archaeologi-
cal Science 31:77–86.[lb]
burleigh, r. 1981. “Appendix C: Radiocarbon dates,” in The
architecture and stratigraphy of the Tell. Edited by K. M. Ken-
yon and T. A. Holland, pp. 501–4. London: British School of
Archaeology in Jerusalem.
———. 1983. “Appendix D: Additional radiocarbon dates for Jeri-
cho,” in Excavations at Jericho, vol. 5, The pottery phases of
the Tell and other finds. Edited by K. M. Kenyon and T. A.
Holland, pp. 760–65. London: British School of Archaeology in
Jerusalem.
cauvin, j. 1989. “La ne´ olithisation au Levant et sa premie`re
diffusion,” in Ne´ olithisations. Edited by O. Aurenche and J.
Cauvin, pp. 3–36. British Archaeological Reports International
Series 516.
cauvin, m.-c., a. gourgaud, b. gratuze, n. ar-
naud, g. poupeau, j.-l. poidevin, and c. cha-
taigner. Editors. 1998.L’obsidienne au Proche et Moyen-
Orient: Du volcan a` l’outil. British Archaeological Reports
International Series 738.
cavalli-sforza, l. l. 1996. “The spread of agriculture and
nomadic pastoralism: Insights from genetics, linguistics, and
archaeology,” in The origins and spread of agriculture and pas-
toralism in Eurasia. Edited by D. R. Harris, pp. 51–69. London:
UCL Press.
childe, v. g. 1929.The Danube in prehistory. Oxford: Clar-
endon Press.
clark, j. g. d. 1965. Radiocarbon dating and the expansion
of farming culture from the Near East over Europe. Proceed-
ings of the Prehistoric Society 31:58–73.
colledge, s. m. 1998. “Identifying pre-domestication culti-
vation using multivariate analysis,” in The origins of agricul-
ture and crop domestication. Edited by A. B. Damania, J. Val-
koun, G. Willcox, and C. O. Qualset, pp. 121–31. Aleppo:
ICARDA.
———. 2001.Plant exploitation on Epipalaeolithic and Early Ne-
olithic sites in the Levant. British Archaeological Reports In-
ternational Series 986.
———. 2002. “Identifying pre-domestication cultivation in the
archaeobotanical record using multivariate analysis: Presenting
the case for quantification,” in The dawn of farming in the

colledge, conolly, and shennan Archaeobotany and the Spread of Farming FS57
Near East. Edited by R. T. J. Cappers and S. Bottema, pp.
141–52. Berlin: Ex Oriente.
———. 2004. “Reappraisal of the archaeobotanical evidence for
the emergence and dispersal of the ’founder crops,’” in Neo-
lithic Revolution: New perspectives on south-west Asia in the
light of recent discoveries in Cyprus. Edited by E. Peltenburg
and A. Wasse, pp. 49–60. Oxford: Oxbow.
contenson, h. de. 1973. Chronologie absolue de Tell As-
wad (Damasce´ ne, Syrie). Bulletin de la Socie´te´ Pre´historique
Franc¸aise 70:253–56.
davis, j. 1992. Review of Aegean prehistory 1: The islands of
the Aegean. American Journal of Archaeology 96:699–756.
demoule j-p., and p. perle` s . 1993. The Greek neolithic:
A new review. Journal of World Prehistory 7:355–416.
edwards, p., j. medows, m. metzger, and s.
ghattas. 2002. Zahrat Adh-Dhra 2: A new Pre-Pottery Neo-
lithic A site on the Dead Sea plain in Jordan. Bulletin of the
American School of Oriental Research 327:1–15.[gw]
fisher, r. a. 1937. The wave of advance of advantageous
genes. Annals of Eugenics (London) 7:355–69.
garrard, a. 1999. Charting the emergence of cereal and
pulse domestication in South-west Asia. Environmental Ar-
chaeology 4:67–86.
gkiasta, m., t. russell, s. shennan, and j.
steele. 2003. The Neolithic transition in Europe: The radio-
carbon record revisited. Antiquity 77:44–62.
greenacre, m. j. 1984.Theory and applications of corre-
spondence analysis. London: Academic Press.
halstead, p. 1994. “The North-South divide: Regional paths
to complexity in prehistoric Greece,” in Development and de-
cline in the Mediterranean Bronze Age. Edited by C. Mathers
and S. Stoddart, pp. 195–219. Sheffield: Sheffield University
Press. [kk]
———. 1996. “The development of agriculture and pastoralism
in Greece: When, how, who, and what?,” in The origins and
spread of agriculture and pastoralism in Eurasia. Edited by D.
R. Harris, pp. 296–309. London: UCL Press. [kk]
harris, d. r. 1998. “The spread of Neolithic agriculture from
the Levant to western Central Asia,” in The origins of agricul-
ture and crop domestication. Edited by A. B. Damania, J. Val-
koun, G. Willcox, and C. O. Qualset, pp. 65–82. Aleppo:
ICARDA.
———. 2002. “Development of the agro-pastoral economy in the
Fertile Crescent during the Pre-Pottery Neolithic period,” in
The dawn of farming in the Near East. Edited by R. T. J. Cap-
pers and S. Bottema, pp. 67–83. Berlin: Ex Oriente.
helbaek, h. 1966. “Pre-Pottery Neolithic farming at Beidha
(appendix A),” in Five seasons at the Pre-Pottery Neolithic vil-
lage of Beidha in Jordan. Edited by D. Kirkbride, pp. 61–66.
Palestine Exploration Quarterly 98.
heun, m., r. scha¨ fer-pregl, d. klawan, r. cas-
tagna, m. accerbi, b. borghi, and f. salamini.
1997. Site of einkorn domestication identified by DNA finger-
printing. Science 278:1312–14.
hillman, g. c. 2000. “The plant food economy of Abu Hu-
reyra 1and 2,” in Village on the Euphrates: From foraging to
farming at Abu Hureyra. Edited by A. M. T. Moore, G. C. Hill-
man, and A. J. Legge, pp. 327–422. Madison and New York:
Oxford University Press.
hillman, g. c., and m. s. davies. 1992. “Domestica-
tion rate in wild wheats and barley under primitive cultiva-
tion: Preliminary results and archaeological implications of
field measurements of selection coefficient,” in Pre´ histoire de
l’agriculture: Nouvelles approches experimentales et ethno-
graphiques. Edited by P. Anderson-Gerfaud, pp. 113–58. Mono-
graphie du CRA 6.
hillman, g. c., r. hedges, a. moore, s. col-
ledge, and p. petitt. 2001. New evidence of Late Gla-
cial cereal cultivation at Abu Hureyra on the Euphrates. The
Holocene 11:383–93.
hopf, m. 1962. “Bericht u¨ ber die Untersuchungen von Samen
und Holzkohlresten von der Argissa-Magula aus den pra¨ keram-
ischen bis mittelbronze-zeitlichen Schichten,” in Die
deutschen Ausgrabungen auf der Argissa-Magula in Thessa-
lien.Vol.1.Das pra¨ keramische Neolithikum sowie die Tier-
und Pflanzenreste. Edited by V. Milojcic, J. Boessneck, and M.
Hopf, pp. 101–10. Bonn: Habelt.
———. 1983. “Appendix B: Jericho plant remains,” in Excava-
tions at Jericho, vol. 5,The pottery phases of the Tell and
other finds. Edited by K. M. Kenyon and T. A. Holland, pp.
576–621. London: British School of Archaeology in Jerusalem.
horwitz, l. k., e. tchernov, and h. hongo. 2004.
“The domestic status of the early Neolithic fauna of Cyprus: A
view from the mainland,” in Neolithic Revolution: New per-
spectives on southwest Asia in light of recent discoveries on
Cyprus. Edited by E. Peltenburg and A. Wasse. Oxford: Oxbow.
[ep]
ishii, t., n. mori, and y. ogihara. 2001. Evaluation
of allelic diversity at chloroplast microsatellite loci among
common wheat and its ancestral species. Theoretical Applied
Genetics 103:896–904.[gw]
jones, g. 1991. “Numerical analysis in archaeobotany,” in
Progress in Old World palaeoethnobotany. Edited by W. van
Zeist, K. Wasylikowa, and K.-E. Behre, pp. 63–80. Rotterdam:
Balkema. [lb]
kadane, j. b. 1988. “Possible statistical contributions to pa-
leoethnobotany,” in Current paleoethnobotany: Analytical
methods and cultural interpretations of archaeological plant
remains. Edited by C. Hastorf and V. S. Popper, pp. 206–14.
Chicago: University of Chicago Press. [lb]
king, r., and underhill, p. 2002. Congruent distribu-
tion of Neolithic painted pottery and ceramic figurines with Y-
chromosome lineages. Antiquity 76:707–14.[pb]
kroll, h. 1981. Thessalische Kulturpflanzen. Zeitschrift fu¨r
Archa¨ ologie 15:97–103.
———. 1983.Kastanas: Ausgrabungen in einem Siedlungshu¨ gel
der Bronze- und Eisenzeit Makedoniens,
1975
–
1979
, Die Pflan-
zenfunde. Berlin: Speiss.
k u i j t , i . 2000. “Life in Neolithic farming communities: An in-
troduction,” in Life in Neolithic farming communities: Social
organization, identity, and differentiation. Edited by I. Kuijt,
pp. 3–13. London: Kluwer Academic/Plenum Publishers.
———. n.d. Architecture, stratigraphy, and chronology of the Pre-
Pottery Neolithic A and Late Natufian period occupations of
Iraq ed-Dubb, Jordan. Journal of Field Archaeology. In press.
lange, a. g. 1990.De Horden near Wijk bij Duurstede:
Plant remains from a native settlement at the Roman frontier:
A numerical approach. Nederlandse Oudheden, Kromme Rijn
Projekt 3. Amersfoort: Rijksdienst voor het Oudheidkundig
Bodemonderzoek.
marinval, p. 1992. “Approche carpologique de la ne´olithisa-
tion du sud de la France,” in Pre´ histoire de l’agriculture: Nou-
velles approches expe´ rimentales et ethnographiques, pp.
255–63. (Monographie du CRA 6.) Paris: CNRS. [lb]
nesbitt, m. 2002. “When and where did domesticated cereals
first occur in southwest Asia?” in The dawn of farming in the
Near East. Edited by R. T. J. Cappers and S. Bottema, pp.
113–32. Berlin: Ex Oriente.
nowak, m. 2001. The second phase of Neolithization in east-
central Europe. Antiquity 75:582–92.
o¨ zdog˘ a n , m . 1997. The beginning of neolithic economies in
southeastern Europe: An Anatolian perspective. Journal of Eu-
ropean Archaeology 5:1–33.
———. 1999.“C¸ ayo¨nu¨,”in Neolithic in Turkey: The cradle of
civilisation. Edited by M. O
¨zdog˘ an and N. Balkan, pp. 36–63.
Istanbul: Arkeoloji Ve Sanat Yayınları.
peltenburg, e. 2003. The colonisation and settlement of
Cyprus. Investigations at Kissonerga-Mylouthkia 1976–1996.
(Lemba Archaeological Project 3; Studies in Mediterranean Ar-
chaeology 70[4].) Sa¨vedalen: A
˚stro¨ms Fo¨ rlag. [ep]
peltenburg, e., s. colledge, p. croft, a. jack-
son, c. mc cartney, and m. a. murray. 2001. Ne-
olithic dispersals from the Levantine corridor: A Mediterranean
perspective. Levant 33:35–64.
perle` s , c . 2001.The Early Neolithic in Greece. Cambridge:
Cambridge University Press.

S58 Fcurrent anthropology Volume
45
, Supplement, August–October
2004
price, d., r. a. bentley, j. lu¨ ning, d. gronen-
born, and j. wahl. 2001. Prehistoric human migration
in the Linearbandkeramik of Central Europe. Antiquity 75:
593–603.
price, t. d. Editors. 2000.Europe’s first farmers. Cambridge:
Cambridge University Press.
renfrew, i. 1966. A report of recent finds of carbonised cereal
grains and seeds from prehistoric Thessaly. Thessaliki 5(E):
21–36.
richards, m. 2003. The Neolithic invasion of Europe. An-
nual Review of Anthropology 32:135–62.
richerson, p., r. boyd, and r. bettinger. 2001.
Was agriculture impossible during the Pleistocene but manda-
tory during the Holocene? A climate change hypothesis. Amer-
ican Antiquity 66:387–411.
roberts, n. 1991. Late Quaternary geomorphological change
and the origins of agriculture in south central Turkey. Geoar-
chaeology 6:1–26.
ro¨ sch, m. 1998. The history of crops and crop weeds in south-
western Germany from the Neolithic period to modern times,
as shown by archaeobotanical evidence. Vegetation History
and Archaeobotany 7:109–25.
runnels, c. 1995. Review of Aegean prehistory 4: The Stone
Age of Greece from the Palaeolithic to the advent of the Neo-
lithic. American Journal of Archaeology 99:699–728.
sampson, a., and s. katsarou. n.d. Cyprus, the Aegean,
and Near East during the PPN. Neo-lithics. In press. [ep]
shennan, s. 1997.2d edition. Quantifying archaeology. Edin-
burgh: Edinburgh University Press.
———. 2002.Genes, memes, and human history: Darwinian ar-
chaeology and cultural evolution. London: Thames and
Hudson.
simmons, a., et al. 1999.Faunal extinction in an island
society: Pygmy hippopotamus hunters of Cyprus. New York:
Plenum.
ter braak, c. j. f., and p. smilauer. 2002.CANOCO
reference manual and Canodraw for Windows user’s guide:
Software for canonical community ordination. Version 4.5. Ith-
aca: Microcomputer Power.
todd, i. a. 1987.Vasilikos Valley Project
6
: Excavations at
Kalavassos-Tenta
1
. (Studies in Mediterranean Archaeology
71[6].) Astrom: Go¨ teberg.
tringham, r. 2000. “Southeastern Europe in the transition to
agriculture in Europe: Bridge, buffer, or mosaic,” in Europe’s
first farmers. Edited by D. Price, pp. 19–56. Cambridge: Cam-
bridge University Press.
valkoun, j., j. giles waines, and j. konopka.
1998. “Current geographical distribution and habitat of wild
wheats and barley,” in The origins of agriculture and crop do-
mestication. Edited by A. B. Damania, J. Valkoun, G. Willcox,
and C. O. Qualset, pp. 293–99. Aleppo: ICARDA.
van andel, t., and c. runnels. 1995. The earliest
farmers in Europe. Antiquity 69:481–500.
van zeist, w., and j. a. h. bakker-heeres. 1982. Ar-
chaeobotanical studies in the Levant 1: Neolithic sites in the
Damascus Basin—Aswad, Ghoraife´ , and Ramad. Palaeohistoria
24:165–256.
———. 1984. Archaeobotanical studies in the Levant 3. Late-Pa-
laeolithic Mureybit. Palaeohistoria 26:171–99.
van zeist, w., and g. j. de roller. 1991–92. The
plant husbandry of Aceramic Cayo¨nu¨ , SE Turkey. Palaeohisto-
ria 33–34:65–96.
valamoti, s.-m., and k. kotsakis. 2004. Transitions to
agriculture in the Aegean: The archaebotanical evidence. Paper
presented to the Workshop “New Perspectives on the Origins
and Spread of Farming in Southwest Asia an Europe,” London.
[kk]
vigne, j. d. 2003. “L’exploitation des animaux a` Torre Sabea:
Nouvelles analyses sur les de´buts de l’e´levage en Me´diterrane´e
centrale et occidentale,” in Torre Sabea: Un e´ tablissement du
Ne´olithique ancient en Salento. Edited by J. Guilaine and G.
Cremonesi, pp. 325–59. Rome: Ecole Franc¸aise de Rome. [lb]
vigne, j. d., and d. helmer. 1999. “Nouvelles analyses
sur les de´ buts de l’e´ levage dans le Centre et l’Ouest me´ diterra-
ne´ens.” Le Ne´ olithique du Nord-Ouest Me´ diterrane´ en (Actes
du XXIVe`me Congre` s Pre´ historique de France, Carcassonne,
26
–
30
September
1994
), pp. 129–46. Paris: Socie´te´ Pre´ histo-
rique Franc¸aise. [lb]
watkins, t. 1996. “Excavations at Pınarbas¸ı: The early
stages,” in On the surface: C¸ atalho¨yu¨k
1993
–
1995
. Edited by I.
Hodder. British Institute of Archaeology at Ankara Monograph
22. Cambridge: McDonald Institute Monographs.
willcox, g. 1996. Evidence for plant exploitation and vegeta-
tion history from three Early Neolithic pre-pottery sites on the
Euphrates (Syria). Vegetation History and Archaeobotany 5:
143–52.
———. 1999. “Agrarian change and the beginnings of cultivation
in the Near East: Evidence from wild progenitors, experimental
cultivation, and archaeobotanical data,” The prehistory of food.
Edited by J. Hather and C. Gosden, pp. 479–500. London: Rout-
ledge. [gw]
———. 2002. “Geographical variation in major cereal compo-
nents and evidence for independent domestication events in
the Western Asia,” in The dawn of farming in the Near East.
Edited by R. T. J. Cappers and S. Bottema, pp. 133–40. Berlin:
Ex Oriente. [gw]
———. 2004. Measuring grain size and identifying Near Eastern
cereal domesticates. Journal of Archaeological Science 31:
145–50.
willcox, g., and s. fornite. 1999. Impressions of wild
cereal chaff in pis from 10th millennium uncal BP at Jerf el
Ahmar and Mureybit, northern Syria. Vegetation History and
Archaeobotany 8:21–24.
willerding, u. 1986.Zur Geschichte der Unkra¨ uter Mit-
teleuropas.Go¨ ttinger Schriften zur Vor- und Fru¨ hgeschicht 22.
zohary, d. 1996. “The mode of domestication of the founder
crops of southwest Asian agriculture,” in The origins and
spread of agriculture and pastoralism in Eurasia. Edited by D.
R. Harris, pp. 142–58. London: UCL Press.
———. 1999. Monophyletic vs. polyphyletic origin of the crops
on which agriculture was founded in the Near East. Genetic
Resources and Crop Evolution 46:133–42.
zohary, d., and m. hopf. 2000.3d edition. Domestica-
tion of plants in the Old World. Oxford: Oxford University
Press.
zohary, m. 1973.Geobotanical foundations of the Middle
East. Amsterdam: Gustav Fischer Verlag.
zvelebil, m. 2000. “The social context of the agricultural
transition in Europe,” in Archaeogenetics: DNA and the popu-
lation prehistory of Europe. Edited by C. Renfrew and K. Boyle,
pp. 57–79. Cambridge: McDonald Institute for Archaeological
Research.