The emergence of pottery in Africa during the tenth millennium cal BC: New evidence from Ounjougou (Mali)

Abstract

New excavations in ravines at Ounjougou in Mali have brought to light a lithic and ceramic assemblage that dates from before 9400 cal BC. The authors show that this first use of pottery coincides with a warm wet period in the Sahara. As in East Asia, where very early ceramics are also known, the pottery and small bifacial arrowheads were the components of a new subsistence strategy exploiting an ecology associated with abundant wild grasses. In Africa, however, the seeds were probably boiled (then as now) rather than made into bread.

Full text

Research
The emergence of pottery in Africa
during the tenth millennium cal BC:
new evidence from Ounjougou (Mali)
E. Huysecom1∗, M. Rasse2, L. Lespez3, K. Neumann4, A. Fahmy5,
A. Ballouche6, S. Ozainne1, M. Maggetti7, Ch. Tribolo8&S.Soriano
9
New excavations in ravines at Ounjougou in Mali have brought to light a lithic and ceramic
assemblage that dates from before 9400 cal BC. The authors show that this first use of pottery
coincides with a warm wet period in the Sahara. As in East Asia, where very early ceramics are
also known, the pottery and small bifacial arrowheads were the components of a new subsistence
strategy exploiting an ecology associated with abundant wild grasses. In Africa, however, the seeds
were probably boiled (then as now) rather than made into bread.
Keywords: Africa, Sahara, Sub-Saharan, Sahel, Early Holocene, bifacial arrowheads, ceramics
The emergence of pottery in Asia and Africa
Prehistoric populations in Japan, Siberia and China first began to produce ceramic wares
between 15 000 and 10 000 cal BC, more than 5000 years earlier than in the Near East
(Yasuda 2002: 119-42; Kuzmin 2006). The emergence of pottery in East Asia is linked with
the climatic amelioration at the Pleistocene–Holocene transition and coincides with the
appearance of lithic industries marked by distinctive small bifacial arrowheads (Habu 2004:
26-36). This technological complex is usually regarded as an expression of the intensified
1Department of Anthropology and Ecology, University of Geneva, Geneva, CH-1211, Switzerland (Email:
eric.huysecom@unige.ch; sylvain.ozainne@unige.ch)
2Lab. Ledra, CNRS-UMR IDEES 6228, University of Rouen, Mont St-Aignan, F-76821, France, and
lab. AnTET – Arscan UMR 7041 CNRS, University of Paris-X Nanterre, F-92023, France (Email:
michel.rasse@univ-rouen.fr)
3Lab. Geophen-LETG-UMR 6554 CNRS, University of Caen-Basse Normandie, Caen, F-14000, France (Email:
laurent.lespez@unicaen.fr)
4Institute of Archaeological Sciences, Goethe University, D-60323 Frankfurt, Germany (Email:
k.neumann@em.uni-frankfurt.de)
5Department of Botany, University of Helwan, Cairo, Egypt (Email: afahmy658@gmail.com)
6Lab. Environmental Studies on Anthropogenic Systems (LEESA/UA), University of Angers, Angers, F-49000,
France (Email: aziz.ballouche@univ-angers.fr)
7Department of Geosciences, University of Fribourg, Fribourg, CH-1700, Switzerland (Email:
marino.maggetti@unifr.ch)
8CRP2A, UMR 5060, University of Bordeaux, Pessac, F-33607, France (Email: ctribolo@u-bordeaux3.fr)
9Lab. AnTET – Arscan UMR 7041 CNRS, University of Paris-X Nanterre, F-92023, France (Email:
sylvain.soriano@mae.u-paris10.fr)
∗Author for correspondence
Received: 30 October 2008; Accepted: 30 January 2009; Revised: 20 February 2009
ANTIQUITY 83 (2009): 905–917
905

The emergence of pottery in Africa during the tenth millennium cal BC
exploitation of plant and animal resources, often including small-seeded grasses (Richerson
et al. 2001).
In Africa, the earliest pottery has been found in the large mountain massifs of the Central
Sahara, in the Eastern Sahara and the Nile Valley. About 30 14C and luminescence dates
have placed the emergence of ceramics in the Sahara and the Nile Valley between the
end of the tenth and the beginning of the ninth millennium cal BC (Close 1995: 24-7;
Roset 2000; Jesse 2003: 40-42; Haaland 2007: 171-5). This can be related to the sudden
onset of a warmer and wetter climate in the Early Holocene that enabled the re-settling
of the Sahara after the hyperarid phase of the last glacial maximum, the ‘Ogolien’ (Nelson
et al. 2002: 97-9). The origin of the earliest African pottery is controversial and has been
much discussed, with three hypothetical scenarios proposed. The first theory places the
emergence of ceramics in the Nile Valley, based principally on the early exploitation of
aquatic resources and wild cereals in this region (Haaland 1992: 47). The second suggests
an origin somewhere south of the Sahara (Close 1995: 23), but until recently the oldest
finds of sub-Saharan ceramics were only dated to the eighth millennium cal BC, both at
Lothagam in Kenya (Robbins 1974), and in the Ravin du Hibou at Ounjougou in Mali, for
Phase 2 of its Holocene occupation sequence (Huysecom et al. 2004: 584). A third assumes
that pottery was invented by relict populations who had survived in ecological refuge zones
of the Sahara during the hyperarid Late Pleistocene (Jesse 2003: 43). Within the framework
of the international research project ‘Palaeoenvironment and Human Population of West
Africa’ (Huysecom 2002), we have discovered ceramic sherds at the site of Ravin de la
Mouche at Ounjougou, associated with an original lithic industry and in stratified contexts
dated from before the end of the tenth millennium cal BC. This discovery throws new light
on the chronology of the emergence of ceramics in Africa and its environmental context.
The Early Holocene sequence at Ounjougou
The research programme at Ounjougou (14◦20N, 3◦30W) began in 1997 and since
2004 has developed in two parallel ravines, Ravin du Hibou and Ravin de la Mouche,where
several ceramic sherds were discovered in layers that could be attributed to the initial phases
of the Holocene (Figure 1). In our latest field season in September 2007, we established
the definitive chronostratigraphic sequence for these two ravines and clarified the position
of the pottery and the associated lithic assemblage. In general, the Holocene sedimentary
sequence here is primarily composed of channel infilling due to a high-energy flow of water,
strongly contrasting with the underlying Pleistocene silts and more recent Holocene silty
formations (Rasse et al. 2006). It is now possible to divide the Early Holocene into five
large chronostratigraphic units, identified from top to bottom as HA4 to HA0. The high
precision chronological ranges in the titles of the next sections are based on Bayesian analysis
results of 14 C and OSL dates (see Technical Appendix).
The HA4 formation (6700-8100 cal BC)
The most recent formation, HA4, of fine-grained particle size and particularly well-
developed in the Ravin du Hibou, has yielded artefacts from cultural Phase 2 of the
906

Research
E. Huysecom et al.
Figure 1. A) Location of the site of Ounjougou; B) general stratigraphic sequence for the Holocene of Ounjougou; C)
stratigraphic sequence of the Early Holocene, with the position of the potsherds and 14C and OSL samples; D) OxCal plot of
the series of 14C and OSL dates for the Early Holocene sequence (OxCal v. 4.0.5: Bronk Ramsey 2008. IntCal04 atmospheric
curve). Figures in light grey represent the prior distributions for each sample, those in dark grey the posterior modelled
distributions (hpd). Indicated on the left for each sample, are: laboratory ID, BP age or raw OSL age (in parentheses), and
the individual agreement index (A) (in parentheses). The model agreement index (Amodel) for the data series is 99.6% (see
Technical Appendix).
Holocene occupation at Ounjougou, dated to the eighth millennium cal BC by five
14C dates on charcoal and two OSL dates (between 8080 +
−55 BP and 8700 +
−75 BP,
Figure 1, Table 1). The material culture is characterised by a microlithic quartz industry,
with geometric segments, associated with ceramics and grinding tools (mortar and pestles),
in a Sudanian savannah context combined with Syzygium gallery-forests (Huysecom et al.
2004; Eichhorn & Neumann in press).
907

The emergence of pottery in Africa during the tenth millennium cal BC
Table 1. Radiocarbon and OSL dates, with details of the OxCal plot results. The columns on the right of the table give the 1σand 2σranges of
Bayesian hpd (highest posterior density), the mean (μ), the agreement indice for each date (A) and the agreement index for the whole model
(Amodel). See also Technical Appendix.
Calibrations Bayesian hpd (Amodel =99.6)
Unit/transitions Sample Material 14CBP +
−OSL Age +
−1σ2σ1σ2σμA
End HA4 7032 6654 7096 6173 6729
HA4
01/5 4Quartz grains 9800 1000 7067 6829 7124 6544 6892 93.5
Ly-9335 Wood charcoal 8085 50 7174 6867 7291 6826 7121 6868 7137 6836 7026 116
Ly-9334 Wood charcoal 8080 55 7173 6848 7295 6820 7134 7046 7183 6923 7083 132
Ly-9339 Wood charcoal 8115 50 7172 7052 7312 6861 7182 7070 7306 7056 7160 89.2
Ly-6804 Wood charcoal 8210 60 7317 7086 7450 7066 7322 7143 7451 7077 7243 101
00/5/4 Quartz grains 11000 1200 7684 7234 7846 7070 7466 63
ETH-20214 Wood charcoal 8700 75 7816 7598 8166 7582 7817 7599 8166 7582 7765 99
HA3/HA4 8439 7926 8521 7683 8132
HA3
ETH-27144 Wood charcoal 9150 70 8452 8286 8551 8256 8535 8293 8601 8263 8415 92.6
ETH-27143 Wood charcoal 9365 70 8741 8556 8815 8350 8730 8559 8800 8359 8626 104
ETH-27142 Wood charcoal 9500 75 9120 8659 9152 8629 8812 8645 8891 8572 8739 102
04/21/3Quartz grains 10700 900 8800 8351 8896 7996 8501 134
ETH-28745 Wood charcoal 9515 70 9121 8744 9155 8638 8916 8766 9054 8715 8859 109
ETH-31278 Wood charcoal 9610 70 9184 8841 9233 8790 8939 8799 9040 8756 8886 97.6
ETH-23540 Wood charcoal 9590 70 9148 8836 9224 8773 9036 8861 9131 8821 8966 111
HA2/HA3 9116 8935 9191 8862 9027
HA2 ETH-31279 Wood charcoal 9510 70 9121 8734 9151 8638 9178 9024 9231 8943 9094 84
ETH-28746 Wood charcoal 9785 70 9315 9201 9441 8925 9296 9157 9366 8927 9208 87.1
HA1/HA2 9477 9152 9959 8932 9376
908

Research
E. Huysecom et al.
The HA3 formation (8100-9000 cal BC)
In Ravin de la Mouche, below HA4, unit HA3 comprises a succession of several coarse
sand lenses and grey sandy silt layers with organic remains. This indicates a meandering
river, with a coarse load, flowing in a floodplain with permanent ponds. Six 14Cdates
and one OSL date allow us to place this formation of HA3 between 8100 and 9000 cal
BC (between 9150 +
−70 BP and 9610 +
−70 BP, Figure 1, Table 1). Charcoal, pollen and
phytoliths in this layer indicate the existence of open grassland with a few Sahelo-Sudanian
tree species and a dense gallery-forest with the riverine tree Syzygium. Uapaca, Celtis, Palms
and Marantaceae in the undergrowth were also constituents of the denser woody vegetation
along the watercourses. It points to the existence of a permanent water source, which is
essential for palm growth (Str¨
omberg 2004), and for Uapaca (Arbonnier 2000). There is
some evidence of fire, but only in the gallery-forest (Neumann et al. 2009).
With its coarser and finer laminae, HA3 represents the earliest Holocene rhythmic
sediments resulting from alternating wet and dry season alluvial sedimentation. The
archaeological finds of HA3 are characterised by some isolated quartz flakes and two small
rounded ceramic fragments (diameter 15mm, thickness 7-8mm) which might represent a
re-deposition of older material.
The HA2 formation (9000-9400 cal BC)
Under HA3, unit HA2 is composed of silty sandy alluvial sediments with grey Pleistocene
silty blocks, particularly at the base of the unit (HA2/a-b), which indicates an important
incision of the meandering river and river bank erosion into the Pleistocene deposits (Lespez
et al. 2008). These alluvial sediments experienced a slight pedogenesis, particularly visible
in the upper part of the unit (HA2/c-d). Two 14C dates place the HA2 sequence, which
is archaeologically sterile, between 9000 and 9400 cal BC (9510 +
−70 BP and 9785 +
−
70 BP, Figure 1, Table 1). HA2 is a fossil soil with a massive structure and few channels
with clay coating, and phytoliths originating from vegetation developed in situ on the
ancient land surface. The spectrum of Poaceae short cell phytoliths points to a grassland
similar to the modern northern Sahel, with dominating annuals (Barboni et al. 2007;
Neumann et al. 2009). Among the Poaceae, short cell phytoliths identifiable to subfamily
level – panicoid morphotypes – (Piperno 2006) are well represented. The grass subfamily
Panicoideae comprises numerous annual species with edible grains which are commonly
collected in the Sahel today (Harlan 1989) and were also used by prehistoric populations
in the Sahara during the Early and Middle Holocene (Fahmy & Barakat 1999; Wasylikowa
2001). The phytolith samples also contain an allochtonous component originating from
episodic flooding of the fossil soil, with morphotypes from Marantaceae and palms indicating
a dense gallery-forest with a shade-tolerant undergrowth.
The HA1 formation (before 9400 cal BC)
Below HA3 and HA2, unit HA1 is 2-5m thick and consists of several sequences composed
of red to ochre cross-bedded coarse sand and gravel with grey Pleistocene silt blocks,
suggesting high-energy flows running through a braided river (Lespez et al. 2008). The
909

The emergence of pottery in Africa during the tenth millennium cal BC
phytolith samples in HA1 are very poor, due to the low amount of silt, and are therefore
not interpretable in terms of the vegetation during this period. Only a few micro-charcoals
were found in HA1, and gave a Pleistocene age, showing the importance of re-deposited
sediments. OSL dating failed in the coarse and badly (if at all) bleached sediments. However,
the 14C and OSL dates of units HA2, HA3 and HA4 constitute a reliable terminus ante
quem of 9400 cal BC for unit HA1 and the re-deposition of the archaeological material
(Figure 1, Table 1). The archaeological remains of HA1 include the oldest ceramic sherds and
a rich lithic industry. The artefacts, although out of archaeological context and reworked,
were discovered in a well-characterised sedimentary sequence, accurately positioned in the
stratigraphic sequence between the Pleistocene deposits and the first dated early Holocene
sediments (HA2). They represent a former occupation on the river banks, the sites having
been eroded by fluvial activity and their material re-deposited in HA1. The good preservation
of two ceramic sherds and the slight wearing on the ridges of the lithic material show clearly
that they had not been transported over long distances.
The HA1 assemblage
The lithic assemblage of unit HA1 numbers 479 objects, primarily knapped from
quartz cobbles. It is characterised by small bifacial fusiform or oval foliate points
Figure 2. Bifacial arrowheads from unit HA1, directly
associated with the sherds.
(Figure 2), obtained by bifacial shaping,
in some cases by pressure flaking, and
by the absence of geometric microliths.
Among the other retouched tools are drill
bits, borer, burins and end-scrapers. This
toolkit, although in secondary position,
can be securely attributed to the Early
Holocene, since these types are significantly
absent in the MSA industries of the region,
particularly drill bits, borers and bifacial
points. The latter, based on their size,
morphology and shaping technique, are
quite different from those recovered in the recent phases of the MSA at Ounjougou (Robert
et al. 2003). Moreover, retouched tools are rare or absent in MSA industries at Ounjougou,
although here they represent 6.2 per cent of the entire assemblage. The other retouched
tools within HA1 unit (sidescrapers, retouched flakes, scaled pieces and denticulates) could
also be observed in MSA industries so they could not be securely attributed to Early
Holocene.
Three ceramic sherds from the base of the stratigraphic unit HA1A (Figure 3) are
associated with this industry (their dimensions are respectively 100, 35 and 15mm). Their
thickness ranges from 4.5 to 7mm. Only one form could be reconstituted as a hemispherical
bowl with a simple rim and a diameter of 0.21m. One sherd shows a number of impressions
which could not be precisely identified. Microscopic analysis of two samples shows a silicate
matrix, free of carbonates, with 20-30 per cent of the volume being non-plastic inclusions.
These are mainly well-rounded quartz monocrystals with a thin recrystallization border, very
910

Research
E. Huysecom et al.
Figure 3. Ceramic sherds from stratigraphic unit HA1, older than 9400 cal BC, including a bowl fragment (A) and a
decorated sherd (B).
similar to those observed in local sandstones and clays. Therefore, a local to regional origin
of the analysed samples can be inferred. Mineralogical analyses by X-ray diffraction of the
clays from the closest outcrops confirm the presence of kaolinite, which is lacking in the
studied material. This points to firing temperatures higher than 550◦C, because kaolinite
is not stable above this temperature in oxidizing firing conditions. Evidently, the studied
samples are not fragments of a heat-hardened clay, but of a fired clay, i.e. of a true ceramic
object.
The HA0 formation (before 9400 cal BC)
AtthebaseofRavin de la Mouche, the earliest sedimentary sequence (HA0) of the
Pleistocene–Holocene transition is composed of reworked Pleistocene silts. This unit was
directly cut into a channel developed within the yellow Pleistocene silts of formation U4,
dated by OSL between 45 and 40ka (Rasse et al. 2004), and reflects a brutal hydrologic
episode with significant reworking of the banks of the Yam´
e. It is archaeologically sterile.
Discussion: emergence of pottery south of the Sahara
At Ounjougou, new stratigraphic and chronological data for the beginning of the Holocene
support a terminus ante quem of 9400 cal BC for an archaeological assemblage characterised
by the presence of ceramics and lithics in small bifacial forms. From an archaeological
viewpoint, if we consider all of the 14C dates for African sites with ceramics and contemporary
with the HA1, HA2 and HA3 formations at Ravin de la Mouche (Figures 4 and 5, Table 2),
it can be observed that few of them have been dated earlier than 9000 cal BC. They are
concentrated in two different regions: in the large mountain massifs of the Central Sahara
(Adrar Bous 10 and Tagalagal; Roset 2000) and in the Eastern Sahara and the Nile Valley
(Bir Kiseiba E-79-8, Sarurab 2 and Wadi el Akhdar; Connor 1984; Khabir 1987; Sch¨
on
1996).
911

The emergence of pottery in Africa during the tenth millennium cal BC
Figure 4. Summary of 14C dates from African sites with ceramics contemporaneous with the HA1, HA2 and HA3 form-
ations at Ravin de la Mouche at Ounjougou, in chronological order. The Ravin de la Mouche 14 C dates are here represented
as simple calibrations.
After a review of the evidence, we have decided to exclude a series of dates lacking a
clear stratigraphic context from the discussion: the earliest 14C date of Uadi Ti-n-Torha in
Libya 9080 +
−70 BP (R-1036, Barich 1974: 149), Tamaya Mellet in Niger 9350+
−170 BP
912

Research
E. Huysecom et al.
Figure 5. Map of African archaeological sites with ceramics contemporaneous with the HA1, HA2 and HA3 formations at
Ounjougou: 1) Ounjougou/Ravin de la Mouche; 2) Tagalagal; 3) Adrar Bous 10; 4) Wadi el Akhdar; 5) Bir Kiseiba E-79-8;
6) Sarurab 2 (map data: SRTM and FAO).
(Gif-1728, Paris et al. 1993: 385), Bir Kiseiba E-80-4 in Egypt 9220+
−120 BP (SMU-925,
Close 1984: 347) and finally the Site Launey AK-AF 094-18 in Algeria 9210 +
−115 BP
(UW-97, Maˆ
ıtre 1971: 57; Maˆ
ıtre 1974: 101). The discoveries of Temet in Niger are not
included in this discussion either. The excavation of the lacustrine deposits of Temet yielded
a date indicating the contemporaneity of this site with the HA2 formation at Ounjougou, at
the junction of the tenth and ninth millennium cal BC (9550 +
−100 BP; Roset 1983, 1996).
This is also one of the few sites that contain bifacial arrowheads comparable to those found
in formation HA1 of Ravin de la Mouche, indicating a clear relationship between the two
areas. However, Temet contains only whole or broken stone bowls made out of fibrolite. The
use of pottery here is only suggested by a fragment of a short, toothed object, on a plaquette
of chloritic schist. This object was interpreted as a potter’s comb after the observation of
impressed motifs on surface-find sherds (Roset 1983: Figure 15). The latter sherds cannot,
however, be reliably correlated with the occupation of the site during the Early Holocene. In
addition, this object may also be a fragment of a disc decorated with incisions, without any
necessary connection to ceramic production. Until proof of the contrary, the populations
of Temet appear to have opted for the use of carved and polished stone and not fired clay
for the fabrication of some of their containers.
Only one comparable site can be seen as potentially contemporary with the appearance of
ceramics at Ounjougou, before the HA1/HA2 transition: Bir Kiseiba, in the southern part
of the Egyptian Sahara, this site having also yielded grinding equipment. The site E-79-8
at Bir Kiseiba yielded three sherds discovered during the excavation of sandy sediments
913

The emergence of pottery in Africa during the tenth millennium cal BC
Table 2. Summary of 14C dates and calibrations from African sites with ceramics contemporaneous
with the HA1, HA2 and HA3 formations at Ounjougou. RavindelaMouchedates appear here as
simple calibrations in chronological order.
Country Site N˚ Lab 14CBP +
−Cal BC 1σCal BC 2σ
Egypt Bir Kiseiba E-79-8 SMU-757 8920 130 8274 7846 8323 7610
Niger Adrar Bous 10 N 9030 190 8535 7940 8700 7612
Egypt Bir Kiseiba E-79-8 SMU-861 9060 80 8428 8207 8536 7969
Soudan Wadi el Akhdar UtC-6536 9080 50 8310 8246 8437 8226
Niger Adrar Bous 10 9100 150 8563 8004 8713 7823
Niger Tagalagal 9100 120 8541 8221 8633 7955
Niger Adrar Bous 10 9130 65 8437 8277 8541 8245
Niger Tagalagal 9150 90 8525 8279 8610 8235
Mali Mouche HA3 ETH-27144 9150 70 8450 8285 8550 8255
Egypt Bir Kiseiba E-79-8 SMU-914 9180 140 8597 8275 8790 7971
Niger Tagalagal 9330 130 8753 8348 9120 8284
Soudan Sarurab 2 HAR-3476 9339 110 8755 8354 9116 8293
Egypt Bir Kiseiba E-79-8 SMU-927 9350 120 8780 8352 9120 8295
Mali Mouche HA3 ETH-27143 9365 70 8739 8555 8814 8348
Soudan Sarurab 2 HAR-3475 9370 110 8796 8461 9120 8308
Niger Tagalagal 9370 130 8817 8350 9132 8298
Egypt Bir Kiseiba E-79-8 SMU-758 9440 230 9147 8481 9393 8224
Mali Mouche HA3 ETH-27142 9500 75 9119 8658 9150 8628
Mali Mouche HA2 ETH-31279 9510 70 9119 8733 9150 8636
Mali Mouche HA3 ETH-28745 9515 70 9120 8743 9154 8637
Mali Mouche HA3 ETH-23540 9590 70 9147 8835 9222 8772
Egypt Bir Kiseiba E-79-8 SMU-928 9610 150 9220 8813 9336 8561
Mali Mouche HA3 ETH-31278 9610 70 9183 8839 9231 8789
Mali Mouche HA2 ETH-28746 9785 70 9313 9200 9440 8923
Egypt Bir Kiseiba E-79-8 SMU-858 9820 380 10015 8759 10660 8326
(Connor 1984), found just below ground surface, as well as at depths of 0.10 and 0.60m.
In the publication, the excavator indicates for the deepest sherd: ‘it is possible that the sherd
might have been moved to this depth by traffic over the surface of the site’ (Connor 1984: 240).
Three other sherds were nearby surface finds. The seven 14C dates obtained on charcoal,
unfortunately, have large error margins and as a result a broad range for the calibration,
which ranges from the end of the eleventh millennium to the beginning of the eighth
millennium cal BC (between 9820 +
−380 BP and 8920 +
−130 BP). Without stratigraphic
context to clearly correlate the three sherds and the dates, it is not possible to go further in
the interpretation of this site in terms of dating the emergence of ceramics.
From a palaeoenvironmental viewpoint, geomorphological and sedimentological analyses
in Ravin de la Mouche indicate a powerful hydrologic regime for this period that remodelled
the landscape on the valley floor. This allows us to identify a relationship between the
emergence of the ceramic and lithic assemblage with one of the humid phases of the
Pleistocene–Holocene transition recently recognised in West Africa (DeMenocal et al.
2000; Duplessy et al. 2005; L´
ezine et al. 2005). This corresponds most probably to the
abrupt resumption of the African monsoon after the Younger Dryas, between 10 050 and
914

Research
E. Huysecom et al.
9350 cal BC, the early Holocene monsoon front reaching 14◦N around 9500 cal BC in
West Africa (Garcin et al. 2007). The palaeoenvironmental data from Ounjougou and other
terrestrial sites in the Sahel (Waller et al. 2007; Neumann et al. in press) show that the onset
of the monsoon had an immediate effect on the landscape. A vast tropical grassland spread
across the former desert areas, and panicoid grasses with edible grains became available in
abundance. As in the Near East (Haaland 1995, 2007; Hillman 1996) and in East Asia,
the massive presence of wild cereals triggered the development of new resource exploitation
behaviour, linked with technological innovations for collection, storage and processing. Heat
treatment of the wild cereals before consumption increases the digestibility of the starch-rich
grains by amylase in the human body (Stahl 1989). While the baking of bread became the
predominant form of processing in the Near East, we hypothesize that the small grains of
the tropical African Panicoideae were boiled in a container, as practised today in the Sahel.
As with East Asia, African ceramics were part of a new technological complex, together with
the production of small bifacial arrowheads for hunting in the open tropical savannas.
Conclusion
Thus, with a solid stratigraphic and chronological context at Ounjougou, there is no doubt
that ceramics appeared in sub-Saharan West Africa at least as early as in the Nile Valley,
some time before 9400 cal BC. This innovation must be coupled with the re-establishment
of the tropical grassland during the Early Holocene. Starting in the middle of the tenth
millennium cal BC, the new technological complex may have rapidly diffused northwards,
together with the advancing monsoon front, the greening of the Sahara and the massive
expansion of edible Panicoid grasses.
Acknowledgments
We would like to thank the Swiss National Foundation for Scientific Research (FNRS), the Swiss-Liechtenstein
Foundation for Foreign Archaeological Research (SLSA), the State and Academic Society of Geneva, the Faculty
of Science of the University of Geneva, the French Minister of Foreign Affairs and the European Community.
The palaeoenvironmental studies conducted by the researchers of Caen, Angers, Rouen and Frankfurt have
been supported by the INSU-CNRS (program CAPHASS-ECLIPSE II), the French National Research Agency
(ANR) and the German Research Foundation (DFG) (program APPD). Our research in Mali benefits from
the support of the Cultural Mission of Bandiagara, the Institute of Human Sciences in Bamako, the University
of Mali, the Swiss Consulate in Bamako, as well as the inhabitants of the villages of Dimmbal and Gologou.
Thanks to Rebecca Miller and Richard Byer for translation and language editing and thanks to Philippe Lahos
for discussing the Bayesian analysis.
Technical Appendix
After calibration of the 14C data (IntCal04), OSL and 14 C data sets have been treated together with the
Bayesian statistical method. This method consists in using secure stratigraphic information to recalculate the
probability density functions (pdf) for each sample (Bronk Ramsey 2000). In short, it allows the pinpointing
of the most likely part of each pdf (called ‘posterior’ data, often referred to as ‘highest posterior density’ or hpd)
when considering the stratigraphic relationships. The direct and interesting consequence is a reduction of the
uncertainty for each individual value and thus an increase of the chronological resolution (Figure 1).
All calculations have been done with the Oxcal 4.0 software (Bronk Ramsey 2008). OSL data have been
integrated to the 14C ones following the recommendations of Rhodes et al. (2003). The A index gives an
indication of the consistency of the data within the data set: when lower than 60%, the data is considered as an
915

The emergence of pottery in Africa during the tenth millennium cal BC
outsider and can be discarded. In complement to the posterior for each data, one can calculate the most likely
interval for the transitions between the phases. The general chronological ranges used to describe the HA4-HA0
units are based on the intervals calculated for those transitions (68.2% confidence interval and μ). In particular,
this allows concluding that the transition between HA1 and HA2 occurred between 9477 and 9152 cal BC with
68.2% confidence interval (between 9959 and 8932 with 95.4% confidence interval; μ=9376. See Table 1).
References
ARBONNIER, M. 2000. Arbres, arbustes et lianes des
zones s`
eches d’Afrique de l’Ouest. Paris:
CIRAD-MNHN-UICN.
BARBONI,D.,L.BREMOND &R.BONNEFILLE. 2007.
Comparative study of modern phytolith
assemblages from inter-tropical Africa.
Palaeogeography, Palaeoclimatology, Palaeoecology
246: 454-70.
BARICH, B.E. 1974. La serie stratigrafica dell’Uadi
Ti-N-Torha (Acacus, Libia). Per una interpretazione
delle facies a ceramica saharo-sudanesi. Origini 8:
7-157.
BRONK RAMSEY, C. 2000. Comment on the use of
Bayesian statistics for 14C dates of chronologically
ordered samples: a critical analysis. Radiocarbon
42(2): 199-202.
– 2008. Oxcal 4.0 manual http://c14.arch.ox.ac.uk/
oxcalhelp/hlp contents.html. Accessed on 20 July
2008.
CLOSE, A.E. 1984. Report on site E-80-4, in A.E. Close
(ed.) Cattle-keepers of the Eastern Sahara. The
Neolithic of Bir Kiseiba: 325-49. Dallas (TX):
Department of Anthropology, Southern Methodist
University.
– 1995. Few and far between. Early ceramics in North
Africa, in W.K. Barnett & J.W. Hoopes (ed.) The
emergence of pottery. Technolog y and innovation in
ancient societies: 23-37. Washington (D.C.) &
London: Smithsonian Institution Press.
CONNOR, D.R. 1984. Report on site E-79-8, in A.E.
Close (ed.) Cattle-keepers of the Eastern Sahara.
The Neolithic of Bir Kiseiba: 217-50. Dallas (TX):
Department of Anthropology, Southern Methodist
University.
DEMENOCAL,P.,J.ORTIZ ,T.GUILDERSON,J.ADKINS,
M. SARNTHEIN,L.BAKER &M.YARUSINSKY. 2000.
Abrupt onset and termination of the African humid
period: rapid climate responses to gradual
insolation forcing. Quaternary Science Reviews 19:
347-61.
DUPLESSY,J.-C.,E.CORTIJO &N.KALLEL. 2005.
Marine records of Holocene climatic variations.
C. R. Geoscience 337: 87-95.
EICHHORN,B&K.NEUMANN. In press. Holocene
vegetation change and land use at Ounjougou
(Mali), in D.Q. Fuller & M.A. Murray (ed.) Flora,
past cultures and archaeobotany in Africa. Wal n u t
Creek (CA): Left Coast Press.
FAHMY,A.&H.BARAKAT. 1999. Wild grasses as
‘Neolithic’ food resources in the Eastern Sahara: a
review of the evidence from Egypt, in M. Van der
Veen (ed.) The exploitation of plant resources in
ancient Africa: 33-46. New York: Kluwer.
GARCIN, Y., A. VINCENS,D.WILLIAMSON,G.BUCHET
&J.GUIOT. 2007. Abrupt resumption of the
African Monsoon at the Younger Dryas-Holocene
climatic transition. Quaternary Science Reviews 26:
690-704.
HAALAND, R. 1992. Fish, pots and grain: Early and
Mid-Holocene adaptation in the Central Sudan.
African Archaeological Review 10: 43-64.
– 1995. Sedentism, cultivation, and plant
domestication in the Holocene Middle Nile region.
Journal of Field Archaeology 22: 157-73.
– 2007. Porridge and pot, bread and oven: food ways
and symbolism in Africa and the Near East from
the Neolithic to the present. Cambridge
Archaeological Journal 17(2): 167-83.
HABU, J. 2004. Ancient Jomon of Japan. Cambridge:
Cambridge University Press.
HARLAN, J.R. 1989. Wild-grass seed harvesting in the
Sahara and Sub-Sahara of Africa, in D.R. Harris &
G.C. Hillman (ed.) Foraging and farming: 79-98.
London: Unwin Hyman.
HILLMAN, G.C. 1996. Late Pleistocene changes in wild
plant-foods available to hunter-gatherers of the
northern Fertile Crescent: possible preludes to
cereal cultivation, in D.R. Harris (ed.) The origins
and spread of agriculture and pastoralism in Eurasia:
159-203. London: UCL Press.
HUYSECOM, E. 2002. Palaeoenvironment and human
population in West Africa: an international research
project in Mali. Antiquity 76: 335-6.
HUYSECOM,E.,S.OZAINNE,F.RAELI,A.BALLOUCHE,
M. RASSE &S.STOK ES. 2004. Ounjougou (Mali): a
history of Holocene settlement at the southern edge
of the Sahara. Antiquity 78: 579-93.
JESSE, F. 2003. Early ceramics in the Sahara and the
Nile Valley, in L. Krzyzaniak, K. Kroeper & M.
Kobusiewicz (ed.) Cultural markers in the later
prehistory of northeastern Africa and recent research
(Studies in African Archaeology 8): 35-50. Poznan:
Poznan Archaeological Museum.
KHABIR, A.M. 1987. New radiocarbon dates for Sarurab
2 and the age of the Early Khartoum tradition.
Current Anthropology 28: 377-80.
916

Research
E. Huysecom et al.
KUZMIN, Y.V. 2006. Chronology of the earliest pottery
in East Asia: progress and pitfalls. Antiquity 80:
362-71.
LESPEZ,L.,M.RASSE,Y.LEDREZEN,C.TRIBOLO,E.
HUYSECOM &A.BALLOUCHE. 2008. L’´
evolution
hydrog´
eomorphologique de la vall´
ee du Yam´
e (Pays
Dogon, Mali): signal climatique et hydrosyst`
eme
continental en Afrique de l’Ouest entre 50 et 4 ka.
G´
eomorphologie: relief, processus, environnement 3:
169-85.
L´
EZINE, A.-M., J.-C. DUPLESSY &J.-P.CAZET. 2005.
West African monsoon variability during the last
deglaciation and the Holocene: evidence from fresh
water algae, pollen and isotope data core KW31,
Gulf of Guinea. Palaeogeography, Palaeoclimatology,
Palaeoecology 219: 225-37.
MAˆ
ITRE, J.-P. 1971. Contribution `
alaPr
´
ehistoire de
l’Ahaggar. I. T´
efedest centrale. Alger: CRAPE.
– 1974. Nouvelles perspectives sur la Pr´
ehistoire r´
ecente
de l’Ahaggar. Libyca 22: 93-143.
NELSON,K.,M.C.GATTO,F.JESSE &M.N.ZEDENO.
2002. Holocene settlement of the Egyptian Sahara. 2:
The pottery of Nabta Playa.NewYork:Kluwer
Academic & Plenum.
NEUMANN, K., A. FAHMY,L.LESPEZ,A.BALLOUCHE &
E. HUYSECOM. 2009. The Early Holocene
palaeoenvironment of Ounjougou (Mali):
phytoliths in a multiproxy context. Palaeogeography,
Palaeoclimatology, Palaeoecology. 276(1-4): 87-106
PARIS, F., A. PERSON &J.-F.SALI `
EGE. 1993.
Peuplements et environnements holoc`
enes du
bassin de l’Azawagh oriental (Niger), in J. Devisse
(ed.) Vall ´
ees du Niger: 378-92. Paris: R´
eunion des
Mus´
ees Nationaux.
PIPERNO, D.R. 2006. Phytoliths.A comprehensive guide
for archaeologists and palaeoecologists. Lanham
(MD): Altamira Press.
RASSE, M., A. BALLOUCHE,E.HUYSECOM ,CH.
TRIBOLO,S.OZAINNE,Y.LEDREZEN,S.STOK ES &
K. NEUMANN. 2006. Evolution g´
eomorphologique,
enregistrements s´
edimentaires et dynamiques
pal´
eoenvironnementales holoc`
enes `
a Ounjougou
(Plateau dogon, Mali, Afrique de l’Ouest).
Quaternaire 17/1: 61-74.
RASSE,M.,S.SORIANO,CH.TRIBOLO,S.STOKE S &E.
HUYSECOM. 2004. La s´
equence pl´
eistoc`
ene
sup´
erieur d’Ounjougou (Pays dogon, Afrique de
l’Ouest): ´
evolution g´
eomorphologique,
enregistrements s´
edimentaires et changements
culturels. Quaternaire 15/4: 329-41.
RHODES,E.J,C.BRONK RAMSEY,Z.OUTRAM,C.BAT T,
L. WILLIS,S.DOCKRILL &J.BOND. 2003. Bayesian
methods applied to the interpretation of multiple
OSL dates: high precision sediment ages from Old
Scatness Broch excavations, Shetland Isles.
Quaternary Science Reviews 22: 1231-44.
RICHERSON, P.J., R. BOYD &R.L.BETTINGER. 2001.
Was agriculture impossible during the Pleistocene
but mandatory during the Holocene? A climate
change hypothesis. American Antiquity 66(3):
387-411.
ROBBINS, L.H. 1974. Archaeology in the Turkana
District, Kenya. Science 176: 359-66.
ROBERT, A., S. SORIANO,M.RASSE,S.STO KES &E.
HUYSECOM. 2003. First chrono-cultural reference
framework for the West African Palaeolithic: new
data from Ounjougou (Dogon Country, Mali).
Journal of African Archaeology 1(2): 151-69.
ROSET, J.-P. 1983. Nouvelles donn´
ees sur le probl`
eme
de la n´
eolithisation du Sahara: A¨
ır et T´
en´
er´
e, au
Niger. Cahiers de l’ORSTOM, s´
er. G´
eologie 13(2):
119-42.
– 1996. La c´
eramique des d´
ebuts de l’Holoc`
ene au
Niger nord-oriental: nouvelles datations, bilan des
recherches, in G. Aumassip, J.D. Clark & F. Mori
(ed.) The prehistory of Africa (UISPP 13, section
15): 175-82. Forl`
ı: ABACO.
– 2000. C´
eramique et n´
eolithisation en Afrique
saharienne, in J. Guilaine (ed.) Premiers paysans du
monde: naissances des agricultures: 263-90. Paris:
Errance – Collection des Hesp´
erides.
SCH ¨
ON, W. 1996. Ausgrabungen im Wadi el Akhdar, Gilf
Kebir (SW- ¨
Agypten) (Africa Praehistorica 8).
Cologne: Heinrich-Barth-Institut.
STAH L, A.B. 1989. Plant-food processing: implications
for dietary quality, in D.R. Harris & G.C. Hillman
(ed.) Foraging and farming: 171-94. London:
Unwin Hyman.
STR ¨
OMBERG, C.A.E. 2004. Using phytolith assemblages
to reconstruct the origin and spread of
grass-dominated habitats in the Great Plains during
the late Eocene to early Miocene. Palaeogeography,
Palaeoclimatology, Palaeoecology 207(3-4):
239-75.
WALLER, M.P., A.A. STREET-PERROTT &H.WANG.
2007. Holocene vegetation history of the Sahel:
pollen, sedimentological and geochemical data from
Jikariya Lake, north-eastern Nigeria. Journal of
Biogeography 34: 1575-90.
WASYLIKOWA, K. 2001. Site E-75-6: vegetation and
subsistence of the Early Neolithic at Nabta Playa,
Egypt, reconstructed from charred plant remains, in
F. Wendorf & R. Schild (ed.) Holocene settlement of
the Egyptian Sahara. Volume 1: The archaeology of
Nabta Playa: 544-608. New York: Kluwer.
YASUDA, Y. (ed.) 2002. The origins of pottery and
agriculture. New Delhi: Lustre Press & Roli
Books.
917