journal of anthropological archaeology 17, 143–165 (1998)
article no. AA980321
Livestock Husbandry, Pastoralisms, and Territoriality:
The West African Record
Augustin F. C. Holl
Department of Anthropology, University of California, San Diego, 9500 Gilman Drive,
La Jolla, California 92093-0532
Received October 1, 1997; revision received November 1, 1997; accepted January 10, 1998
Evidence for livestock found in West African Holocene sites are considered in relation to patterns
of settlement distribution in the context of changing climatic circumstances. The development of
livestock husbandry appears to coincide with the advent of a drier Middle Holocene climatic,
followed by a relatively fast expansion all over West Africa. Pastoral–nomadic strategies developed
in the northern Saharan part are hardly distinguishable from Late Stone Age settlement–subsistence traditions, while genuine agro-pastoral systems emerged in the south, along the Sahara–Sahel
margins. In all the cases, stone tumuli burials, rock art stations, and intermittent or permanent
settlements were used as de facto territorial markers, effective signposts of West African Collector,
herder, and peasant territoriality. © 1998 Academic Press
ety could only function while the outside
world not only existed but also allowed for
those reaction from it—reactions which
were economic, social, political, cultural, in
a word, a multifaceted response—which ensured that the nomads remains nomads’’
(Khazanov 1984: 3). Khazanov views clearly
refer to specialized trade and market-oriented pastoral nomadism, more or less integrated within the fabric of Near-eastern
and Central Asian centralized state societies
from the ca. fourth millenium B.C. onward.
It is doubtful that this scenario is applicable
without serious alteration to all cases of development of pastoral nomadism (Holl
1990). The archaeology of pastoral nomadism is strongly constrained not only by the
ambiguity of cultural remains, but also by
the low preservability and visibility of sites
(Cribb 1991). In this regard, the analysis of
livestock husbandry systems developed by
West African prehistoric societies presented
in this paper will consider the documented
faunal repertoires with due emphasis on
chronology, environment, attested domestic species, regional variability, settlement
INTRODUCTION
In prehistoric contexts, livestock husbandry was probably inextricably imbedded in the fabric of social systems. As socioeconomic subsystem, it ranges from the
rearing of few animals within the domestic
sphere of activity to specialized long-range
trade and market-oriented pastoralisms.
The social, economic, and political impact of
animal husbandry thus increases from one
end of the spectrum, that of small household-based systems to the other, that of
more specialized pastoralist ones. Pastoralism involving diverse animal species and
varying patterns of socioeconomic relationships, is a multifaceted phenomenon with
many possible definitions (Khazanov 1984;
Cribb 1991; Smith 1992a,b; Bar-Yosef and
Khazanov 1992; Frankfort 1990). According
to Khazanov (1984), its nomadic variants developed in specific circumstances. He states
that ‘‘. . . nomads could never exist on
their own without the outside world and its
non-nomadic societies, with their different
economic systems. Indeed, a nomadic soci143
0278-4165/98 $25.00
Copyright © 1998 by Academic Press
All rights of reproduction in any form reserved.
144
AUGUSTIN F. C. HOLL
FIG. 1. Regions and sites with Late Stone Age sites and settlements with livestock remains.
types and distribution, as well as processes
of territorial demarcation if any. The study
area under consideration here is limited in
the north by the Ténéré in Niger in the East
and the mouth of Wadi Dra in Western
Sahara in the West, its southern limit corresponding to the northern sahel margins on
the latitude of the Niger bend (Fig. 1).
The chronological and cultural background from which varying animal husbandry strategies developed can fruitfully
be considered is that of Early Holocene
West African Late Stone Age societies
(Fig. 1). Late Stone Age sites with and
without pottery and bone tools dated to
ca. 9500 to 6000 B.P. have been recorded in
different areas in West Africa. (Fig. 1) In
Niger, few sites were tested at Tagalagal
in the Aı̈r; Temet, Fachi (Roset 1974, 1983),
and Adrar Bous (Clark et al. 1973) in the
Ténéré, and Tamaya Mellet in the Azawagh valley. In Mali further west, Early
Holocene settlements were found in the
Erg-in-Sakane, Erg Jmeya, and Hassi el
Abiod areas in the north and desertic part
of the country (Petit-Maire and Riser 1983;
Petit-Maire et al. 1983) and the Sahelian
southwest in the Vallée du Serpent (McDonald and Allsworth 1994) and the
Mema region (Togola 1996). An extensive
Late Stone Age complex with microlthic
components has been recorded all over
the Sudano-Sahelian belt, from Nigeria to
Senegambia (Andah 1979; Philipson 1982;
Ballouche et al. 1993; Connah 1981;
Breunig et al. 1996).
In general, with minor variation from
one area to the other, the Early Holocene
(ca. 11,000 – 6000 B.P.) climate was humid
and wet, with abundant rainfall (Williams
and Faure 1980). West African vegetation
belts shifted northward, with most of the
present-day desert turned into a more or
less rich grassland (Schulz 1991), with ex-
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
tensive fresh to low brackish water lakes.
With the onset of the Mid-Holocene arid
climatic phase (ca. 5000 – 4000 B.P.), vegetation belts shrunk and/or shifted southward. It is basically in this environmental
context that livestock husbandry appears
to have expanded at a relatively fast speed
in West Africa. This formulation does not
have to be read as environmental determinism; the timing (ca. 5000 – 4000 B.P.) at
archaeological time scale is inescapable,
but the expansion of livestock in some
areas and not in others suggests the operation of other parameters.
Archaeological sites with livestock remains dated from ca. 6000 B.P. onward
have been recorded in the Ténéré at Adrar Bous, Arlit in the Aı̈r (Smith 1992a,b),
in numerous localities of the In-Gall Tegidda-n-Tesemt basin (Grébénart 1985,
Paris 1984, 1992), the Tilemsi (Smith 1980;
Gaussen and Gaussen 1988), the Taoudenni
basin in Northern Mali (Petit-Maire and
Riser 1983), Western Sahara (Petit-Maire
1979a), and finally, the Dhar Tichitt region
in Southeastern Mauritania (Munson 1971;
Holl 1985a,b, 1986). Several other sites,
with evidence of livestock but dated to a
later ca. 4000 –3000 B.P. time range, have
been tested during the past decades.
Some of them, like the Kintampo-culture
settlements Mumute and Tereso, are located as far south as the forest–savanna
ecotone in Ghana (Flight 1976; Andanquah 1993; Stahl 1993), while others are
found in the southwestern part of the
Chad basin (Connah 1981; Breunig 1995;
Breunig et al. 1996). Single-site sequences
present important well-contextualized archaeological data, but their integration in
a study geared toward an analysis of prehistoric pastoral territories is not yet feasible. This paper will thus be based on a
sample of the best documented regional
sequences, with the hope to gain some
insight on prehistoric herders’ social territories.
145
REGIONAL SEQUENCES
The Taoudenni Basin, Malian Sahara
The Taoudenni basin is located between 22°–30° northern latitude and 0°
309–5° 309 western longitude. It is an extended desertic sand plain delimited in
the north and northeast by series of sandstone cliffs (Fig. 1). Within a research
program focused on palaeoclimatic investigations, archaeological surveys and excavations have been carried out in three
distinct study areas: the Erg In Sakane,
Erg Jmeya, and Hassi el Abiod (PetitMaire et al. 1983; Petit-Maire and Riser
1983; Petit-Maire 1991). The Basin appears
to have been settled on a relatively intensive scale during the Early Holocene by
pottery-using hunter– collector folks in a
context of humid climatic conditions. Subsistence systems were based on an intensive exploitation of wild grain and game,
as well as aquatic resources.
Approximately 100 sites of varying extent and morphological characteristics
have been recorded, mapped, and some
of them tested. Settlement size varies
from few square meters to 1 ha. Detailed
publications are not yet available for all
the tested archaeological sites. Site MN
36, situated in the Hassi el Abiod area is
the notable exception (Decobert and
Petit-Maire 1985). It is extended over
1600 m2. The recorded faunal material
comprises abundant remains of fish, tortoises, snakes, crocodiles, and hippopotami. Bones of large antelopes as well as
giraffes, rhinoceros, buffalo, wild boar,
and lions have also been collected. The
settlement was a seasonal camp, visited
intermittently between ca. 8500 to 4500
B.P. Five cattle bones belonging to the
same animal have been collected at Site
AR 7 in the Erg In Sakane area and dated
to 6970 6 120 B.P. (Petit-Maire and Riser
1983: 19). Their association with the rest
of the archaeological record is, however,
146
AUGUSTIN F. C. HOLL
highly uncertain. According to Guérin
and Faure (1983: 245), ‘‘these 5 teeth differ from the rest of the material collected
at AR 7 by their very white patina, and
their surface aspects, they are perhaps of
a recent age.’’ In addition, as far as domestic animals are concerned, a second
cattle phalanx was found at Site MK 2
and dated to 4750 6 120 B.P. Formerly
assigned to a buffalo, that phalanx appeared to be too small in size to Gauthier (1991: 173–174) who considers it to
belong to Bos taurus and not to the illdefined Bos ibericus.
If considered in relation to changing
Holocene climatic patterns, two main sequences of occupation with slightly distinct subsistence systems but comparable
site-location strategies are attested in the
Taoudenni basin. An earlier Late Stone
Age hunter– collector occupation sequence dated from ca. 8500 to 5000 B.P.
during the Early Holocene humid phase,
followed by a later collector– herder sequence, singled out by the presence of
cattle during the mid-Holocene dry phase.
As can be judged from the frequency of
recorded faunal remains, hunting of wild
game and the exploitation of aquatic resources and wild grain were predominant
subsistence pursuits during both settlement sequences. Sites were preferentially
located near freshwater lakes and lagoons;
pottery, grindstones, and grinders are
abundant in almost all the recorded localities. There is no remarkable change in
stone tool types and knapping techniques.
The adoption of cattle husbandry thus
clearly seems to coincide with the advent
of drier climatic conditions. It can be suggested that livestock husbandry based on
cattle herding may have been an alternative settlement–subsistence strategy devised by the Taoudenni Basin Mid-Holocene folks to face decreasing productivity
of their home range. Site-location strategies were based on a ‘‘traditional Late
Stone Age’’ model of mobility, with dis-
persed small sites at one end of the settlement spectrum, and at the other, relatively larger sites located along more or
less permanent water places.
The Western Sahara along the Atlantic
Coast
Along the Atlantic coast, an extensive
but narrow stretch of land, 2000 km long
north–south and 50-km-wide east–west in
average, was surveyed by N. Petit-Maire
and her research team (Petit-Maire 1979a,b,
1980). The investigated area is limited in
the north by wadi Draa and Cape Timiris
in the south. Twenty-one archaeological
sites, ranging in time from ca. 8000 to 2000
B.P. have been recorded and mapped.
They consist of several shallow surface
sites with more or less loose scatters of
cultural evidence, concentrated or dispersed burials, shell-middens, and relatively bulkier settlement evidence found
at Tintan and Chami. Subsistence systems
devised by Western Sahara mobile prehistoric folk consisted of hunting, fishing,
wild plant gathering and the exploitation
of sea mollusks, a subsistence package to
which livestock husbandry was added
later. In the context of a drier Mid-Holocene climate, fresh water may have been a
critical resource.
The prehistoric sequence recorded in
the Western Sahara study area is divided
into two chronological periods. The earliest dated from ca. 8000 to 4500 B.P. is
characterized by an extremely low density
of settlement consisting of low shell-middens found along the coast, with very little
evidence of material culture (Petit-Maire
1979a). The second chronological period is
dated from ca. 4000 to 2000 B.P. It is characterized by a higher density of settlements, ranging from shell-middens and
cemeteries to bulkier sites. Archaeological
data are much more diversified; stone
tools, pottery, and grinding equipment as
well as faunal remains of wild and domes-
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
tic mammals have been recorded. Two
sites, Tintan on the coast and Chami farther inland, have been submitted to more
intensive archaeological investigation.
The faunal evidence from Tintan, the
coastal site, consists exclusively of sea
mollusks. Five species have been identified; these are Anadara senilis, Conus papilionaceus, Murex hoplites, Cymbium neptuni,
and Semifusus morio (Elouard 1979: 228 –
229). Tintan was probably settled intermittently, on a seasonal basis; its position in
the subsistence–settlement spectrum appears clearly as a special-purpose site.
Chami is located approximately 100 km
farther inland, near a freshwater spring.
Fireplaces, grinding tools, pottery, and
faunal remains are much more abundant
on the settlement, extended over 1 ha.
Animal bones are distributed into remains
of large wild mammals, Phacochoerus aethiopicus (wild boar), Oryx algazel (oryx),
Loxodonta africana (elephant), Ceratotherium simum (rhinoceros), Lepus capensis
(hare), and Gazella sp., and livestock with
Bos sp. (cf. Taurus) and Ovis sp. Frequency
distribution of species recorded in the
Chami faunal sample is unknown. It is
however specified that cattle is one among
the three predominant species (PetitMaire 1979a: 234). The presence of a freshwater spring and a diversified faunal
material, as well as bulkier settlement evidence, suggest that Chami may have
been located in an optimal herding area.
It was probably part of the Late Stone
Age herder– collector home range, settled
preferentially during dry seasons.
Evidence of material culture shows
strong similarities between Tintan and
Chami archaeological records. Both settlements are contemporaneous and situated
100 km from each other; the former on the
coast and the latter in the hinterland. The
recorded difference in ecological settings
thus suggests that both sites may have
been part of a broad Late Stone Age collector– herder home range with comple-
147
mentary components. The coast may have
been visited by dispersed small groups of
herders during rainy seasons when fresh
water was easily available, with a focus on
the exploitation of marine resources at
several dispersed localities, Tintan being
one of them. During dry seasons, humans,
livestock, and wild mammals may have
converged toward permanent freshwater
places situated farther in the hinterland.
Repeated occupation of the same place
has resulted in the formation of bulkier
and extensive archaeological sites, as is
the case for Chami. As has been shown for
the Taoudenni basin, evidence for livestock, in this case cattle and sheep, also
appeared in the western Sahara archaeological sequence during a climatic phase
characterized by a sustained trend toward
increased aridity. Livestock husbandry
thus appears as a new alternative integrated within an earlier mobile Late Stone
Age settlement–subsistence tradition.
The Tilemsi Basin, South Central Mali
The Tilemsi is an extensive ‘‘fossil’’
river drainage basin, 700 km in length
from the southern part of the Tanezrouft
in the north to the Niger river in the south
and 400 km wide (Fig. 1). It is limited in
the northeast by the Adrar-n-Ifogha plateau (Gaussen and Gaussen 1988). Several
Late Stone Age sites ranging from low
mounds (Smith 1975, 1980, 1992a,b), surface sites with more or less dense scatters
of archaeological evidence (Gaussen and
Gaussen 1988; Raimbault 1991), cemeteries with stone tumuli, carnelian bead
workshops, and rock-art stations (Lhote
and Tomasson 1967) have been recorded.
Surveys and few archaeological test excavations have been carried out. The
Nécropole de la Frontière, a cemetery with
20 stone tumuli, measuring 260 m in
length and 150 m in width, has been recorded in the north, on the edge of the
Tanezrouft. One of the tumuli has been
148
AUGUSTIN F. C. HOLL
excavated and dated to 4750 6 80 B.P. The
deceased was buried with five pieces of
pottery, eight ground axes, and two bone
tools, a spatula and a point. Several surface sites have been found in the extensive
Asselar depression, situated on the western part of the Tilemsi basin. A well
known well of brackish water, attracting
present-day populations of pastoralists
for their annual ‘‘cure salée,’’ is situated at
the center of the depression. An archaeological site, 3 to 4 ha in surface area was
discovered in the proximity of the well. A
sample of 5170 stone tools collected from
the surface comprises 710 arrowheads, 600
end scrapers, 1399 geometric microliths,
810 borers, 598 blades, bladelets, backed
and retouched flakes, and finally, 118
ground-stone pieces. The collected faunal
material consists of freshwater mollusks
(Melania sp.), a large quantity of catfish,
and few pieces wild boar, gazelle, antelope, and rhinoceros bone. No livestock
remains have been reported. The situation
is different at Aguendemen, a site located
at 20 km of Asselar, where cattle bones
have been found in a context with quite
similar material culture (Gaussen and
Gaussen 1988: 83).
Two mounds have been tested at
Karkarichinkat A and B in the south
(Smith 1975, 1980, 1992a,b). Both mounds,
inhabited from ca. 4000 to 3300 B.P., are
located in the bed of the Tilemsi river.
Pottery is abundant and stone tool repertoires consist of grinding equipment, microliths, end scrapers, burins, borers, bifacially retouched arrowheads, and finally
ground axes. Numerous terracota figurines representing cattle have also been
collected. Cattle bones are predominant
among faunal remains (41.4% of mammalian bones). In addition, goats bones
(21.8%) were found as well as those from
few wild mammals, gazelles, wild boar,
and jackals (Smith 1975, 1980). Aquatic resources are represented by abundant remains of fish, with nile perch predomi-
nant, and mollusks. Plant macroremains
of wild fruits, Celtis integrifolia, Grewia sp.,
and Acacia nilotica, and imprints of Pennisetum sp. (McIntosh and McIntosh 1983)
are attested. According to Smith (1975),
both mounds may have been settled according to seasonal variation in resources.
The high frequency of cattle figurines,
supported by the predominance of cattle
and sheep/goat bones, provides strong evidence for the development of a pastoral–
nomadic way of life and tradition.
Peculiar and intriguing archaeological
sites were discovered in the southeast,
north, and northeast of Gao. They consist
of high concentrations of carnelian beads
workshops clustered around the localities
of In Begouen and Telataye (Gaussen and
Gaussen 1988). The recorded material culture comprises pottery, carnelian at every
stage of bead manufacture process, thousands of stone borers, ground axes, and
other knapped flint tools. None of the
workshop has been tested and the chronology of this specific component of the
Tilemsi archaeological sequence is still
obscure. Rock-art stations are concentrated in the northern part of the basin,
along streambeds in the Adrar-n-Ifogha
(Lhote and Tomasson 1967) and the Timetrine massif. Cemeteries with varying
numbers of stone tumuli are located in the
foothills on the western edge of the Adrar
and the Timetrine massif. Thirty settlements have been mapped during a recent
survey conducted in the latter area (Raimbault 1991: 138 –140).
Very few settlements have been tested in
the Tilemsi area. The chronology of the recorded archaeological sequence is still in
need of further clarification. Judged from
the overall similarities in material culture,
the Tilemsi archaeological record has been
partitioned into four regional variants,
termed Facies A (Asselar), B (In Begouen), K
(Karkarichinkat), and T (Telataye) (Gaussen
and Gaussen 1988). If read from the background of sharply contrasted seasons which
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
may have characterized the Mid-Holocene
climate, the Tilemsi basin was probably settled by Late Stone Age pastoral–nomadic
groups. The landscape is dotted with intermittent habitation sites, special-purpose
sites, burial grounds, and rock-art stations,
which can be considered as different but
complementary signposts of pastoral nomad social territories. Their home range
may have comprised highlands, plateaus,
and massifs such as the Adrar-n-Ifogha and
the Timetrine massif in the north and northeast and the Niger river and its floodplain in
the south, between them being the Tilemsi
river valley oriented north–south, the extensive Asselar depression with brackish water
sources, lagoons, and ponds in the west,
and the rolling dunes regions of In Begouin
and Telataye in the southeast. Intensive carnelian bead manufacture is a later development, ranging in time from the beginning of
the first millenium A.D. to 1600 A.D. (Holl
1994, 1995a). Pastoral–nomadic groups were
then involved in long-distance trade relationships with developing markets and
urban centers from the Niger river and surrounding lands. Long-range pastoral nomadism with camels, involved in long-distance trans-Saharan trade, may have been
developed during this later period.
The Dhar Tichitt, Southeastern Mauritania
The Dhar Tichitt region is located in
southwestern Sahara in Mauretania, by
18° 209/18° 279 N and 9° 059/9° 309 W (Fig.
1). The study area, slightly more than 600
km2 in extent, measures 44 km in length
and 14 km in width. The physical landscape consists of a series of sandstone
cliffs, oriented NW–SE, situated at elevation varying from 40 to 60 m above sandy
flats and interdunal depressions with remains of former freshwater lakes (Munson 1971; Holl 1985a,b, 1986, 1989, 1993;
Amblard and Pernès 1989). Several surveys and archaeological excavations have
been carried since the beginning of this
149
century. The archaeological record comprises massive settlement evidence with
dry-stone masonry distributed among
well defined villages of varying size, stone
tumuli, rock engravings and paintings,
and, finally, more or less extensive scatters of surface material.
With 46 recorded sites, the Dhar Tichitt
prehistoric sequence spans approximately
2000 years, from ca. 4000 to 2000 B.P., with
settlement divided into two major categories. Villages with dry-stone features, organized into walled compounds and ranging in size from 500 m2 to 18 ha, are
located above the sandstone cliff; surface
sites with scatters of cultural material are
found in the sandy flats and interdunal
depressions, in general near former freshwater lakes. It is highly probable that
many small surface scatters were wiped
out by hydrologic agencies or hidden by
drifting sand or even escaped observation.
As suggested by palaeoclimatic research
(Hugot 1977; Holl 1986: 26 –29; Person et al.
1995), the Late Holocene climate was characterized by two sharply contrasted seasons: a longer dry season and a shorter
rainy season with stormy rains. However,
the climate appears to have been slightly
wetter from ca. 3000 to 2500 B.P.
Dhar Tichitt Late Holocene people
were semisedentary agro-pastoralists with
broad-spectrum subsistence systems. They
include hunting, fishing, wild fruits and
grain collecting, agriculture of bulrush
millet (Pennisetum sp.), and, finally, livestock herding of sheep/goats and cattle.
Fishing is attested by remains of catfish
and Nile perch bones. Three samples of
faunal remains have been collected. The
first came from controlled excavation of a
compound dated to ca. 2500 B.P. (Gif 6083)
and systematic surface sampling of the
12-ha settlement of Akhreijit, the second
from a surface site (Site 46) located in the
interdunal depression near a former
fresh-water lake, and the third through
judgemental sampling implemented dur-
150
AUGUSTIN F. C. HOLL
ing a survey conducted by the Dhar Tichitt Prehistoric Mission in 1980 (Holl
1985a, 1986).
Seventeen taxa have been identified in
the Akhreijit village samples, Nile perch
(3 vertebras), 14 of wild mammals and 2 of
livestock. Wild mammals are distributed
into 80 gazelles (39.02%), 39 bovids (19.02%),
10 oryx (4.8%), 9 addax (4.3%), 5 eland
(2.43%), 3 (1.46%) each of Tragelaphus sp.
and Gazella dama, 2 each (0.97%) of Hippotragus equinus, Mellivora capensis, and Kobus sp., and finally, 1 (0.48%) each of Equus
sp., Panthera leo, Cricetomys sp., and
Genetta genetta. Domestic animals are represented by 34 (16.58%) cattle and 12
(5.85%) sheep/goat bones. Site 46 sample
comprises 16 identified taxa; one fragment
of Crocodylus niloticus carapace, 145 burned
ostrich eggshell fragments, and five large
Nile perch vertebras measuring 26 to 32
mm in length and 25 to 32 mm in diameter
have been collected. Wild mammals are
distributed into 21 (35%) Bovidae sp., 12
(20%) Gazella dorcas, 6 (10%) H. equinus, 5
(8.33%) Oryx algazel, 3 (5%) Addax nasomaculatus, 2 (3.33%) Tragelaphus sp., and 1
(1.66%) each for Equus sp., Taurotragus derbianus, G. dama, and finally, Hippopotamus
amphibius. Three domestic species are attested: 4 (6.66%) for cattle, 1 (1.66%) for
sheep/goat, and 2 (3.33%) for camel. The
camel bones singled out by their white
patina, are clearly intrusive. The third
sample is patently characterized by its
loose chronological resolution. Sixteen
taxa have been determined; domestic
mammals are represented by 27 (32.14%)
cattle and 9 (10.71%) sheep/goat bones.
Eleven wild mammals are distributed into
13 (15.47%) Bovidae sp., 6 (7.14%) H. amphibius, 5 (5.95%) each for O. algazel and G.
dorcas, 4 (4.76%) each for T. derbianus, G.
dama, and H. equinus, 2 (2.38%) A. nasomaculatus, and, finally, 1 (1.19%) each for
Equus sp., Acinonyx jubatus, M. capensis and
Ceratotherium simum (white rhinoceros). In
addition, two jawbones and two dorsal
spines of catfish as well as a tens of connected vertebras of Python mollure were
collected.
The proportion of domestic animals
varies from 22.43% (Akhreijit village) to
8.32% (Site 46) to 42.85% (third sample).
Cattle bones are always predominant with
proportions varying from 32.14% (third
sample) to 16.5% (Akhreijit village) to
6.66% (Site 46). This predominance of cattle is independently supported by higher
frequencies of representations in rock engravings (Amblard and Vernet 1984; Beyries and Boeda 1981). Some of the scenery
depicts small cattle herds in line followed
by a shepherd. In addition, large cattle
enclosures have been found on the periphery of three walled settlements, at
Akhreijit, Chebka, and El Khimiya. Among
all the studied samples, 8.28% of the determined cattle bones belong to young animals; kill-of pattern does not necessary
reflect herd management strategies. However, read in conjunction with frequent
depictions of cows with overemphasized
udders, milking adult animals may have
been one of the predominant targets of
cattle-herding, putting aside the important but unsubstantiated social value of
livestock herds.
The Dhar Tichitt settlement system
comprised permanent sedentary villages,
characterized by heavy investment in the
construction of habitation features, situated above the sandstone cliff on the
shore of intermittent streams, and intermittent dry season camps, without built
facilities and set in the sandy lowland and
interdunal depressions. The home range
was a narrow one, not exceeding 20 km if
judged by the regional distribution of settlement recorded in the study area. Seasonal moves to dry season camping areas
may have concerned a minute fraction of
the population of each village only, de
facto special-purpose task-groups. Repeated occupation of the same places has
generated extensive surface scatters of
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
cultural material, as is the case for the site
of Goungou, tested by Munson (1971). The
long-term dynamics of the Dhar Tichitt
settlement systems is beyond the scope of
this paper; suffice it to specify that
through time, settlements coalesced into
four subregional clusters organized around
large 4 to 18 ha central villages, Akhreijit,
El Khimiya, Chebka, and finally, Tijot
(Holl 1993). It can nonetheless be considered that the area was inhabited by seminomadic peasant– herder communities,
practicing short-range seasonal moves
from permanent villages to intermittent
dry-season camps and vice-versa.
SELECTED SPECIES, PASTORALISMS,
AND TERRITORIALITY
There is a differential distribution of domestic species, depending on areas and
judged from the available faunal record.
Cattle and sheep/goat bones are found in
virtually all the reviewed regional sequences, with the notable exception of the
Taoudenni basin. This feature recorded in
the Erg In Sakane, Hassi el Abiod, and Erg
Jmeya does not seem to result from a bias
in field observation. If this is the case, the
herding system devised by the Taoudenni
basin Mid-Holocene herder–collectors appears to have been based exclusively on
cattle husbandry. The record from the Asselar depression in the Tilemsi basin is
poorly investigated; cattle bones were recorded, but the absence of sheep/goat
bones does not seem to be significant. In
all the remaining cases, in Western Sahara, Adrar Bous, Aı̈r, Central Tilemsi,
Dhar Tichitt, and the Eghazer basin which
will be presented later as a case study,
livestock remains comprise cattle and
sheep/goat bones in varying proportions,
sometime complemented by the presence
of domestic shepherd dogs (Paris 1984).
Where available, frequency distribution of
cattle and sheep/goat bones always present
a predominance of the former over the
151
latter. As suggested by rock engravings
and the few available morphometric data
(Holl 1986: 93–94; Smith 1992b: 42), West
African prehistoric cattle belong to the
small-size, short horn, humpless breed
(Smith 1986, 1992a,b; Grigson 1991; Clutton-Brock 1993).
Two types of livestock herding systems
can be reconstructed if the available faunal remains are considered in relation to
settlement features, regional distribution
of sites, and inferred climatic patterns.
The first type comprises long-range pastoral-nomadic herder–collectors with more
or less shallow settlement records distributed over territories extended over more
than 100 km. Such systems are found in
the Taoudenni, Tilemsi, Western Sahara,
and Eghazer basin. The social landscape is
dotted with stone-tumuli cemeteries, rockart stations, and seasonal camps, de facto
territorial markers. The second type consists of a single case, that of the Dhar
Tichitt sequence characterized as seminomadic agro-pastoralists. In this case, the
practice of agriculture associated with
genuine village life is associated with livestock husbandry. The territorial range is
narrow, less than 20 km, with probably
specialized groups of herders moving
from one water place to another during
dry seasons. Rock-art stations and stonetumuli burials are dispersed throughout
the landscape without any remarkable
concentrations.
Territoriality in prehistoric pastoral context is a difficult issue. Societies and traditions waxed and waned; and the fluidity
of pastoralist social systems precludes any
attempt at the reconstruction of long-term
bounded spatial units. More probably,
pastoralist territoriality involved changing
flexible networks of places comprising
grazing lands, groves of fruit trees and
wild grain, water-places, settlement areas,
burial places, and symbolically charged
localities (Casimir 1992). ‘‘Territorial functioning refers to an interlocked system of
152
AUGUSTIN F. C. HOLL
sentiments, cognitions, and behaviors that
are highly place specific, socially and culturally determined and maintaining, and
that represent a class of person–place
transactions concerned with issues of setting management, maintenance, legibility,
and expression’’ (Taylor (1988: 6) in Casimir 1992: 19). In some cases, the Saharan
archaeological record provides glimpses
on prehistoric pastoralists territoriality.
This is the case for the extensive and
widespread distribution of megalithic
burials, dispersed or clustered in cemeteries with numerous stone tumuli (Camps
and Camps 1964; Lihoreau 1993; Savary
1966; Paris 1984, 1990), the regional distribution of rock paintings in the Tassili-nAjjer for example (Lhote 1976a; Holl 1989,
1995b), and rock engravings along the
Wadi Djerat (Lhote 1976b). Three examples will be briefly reviewed here.
In their study of Djebel Mazela cemetery at Bou Nouara in eastern Algeria,
the largest megalithic cemetery in
northern Africa extended over 400 ha
and comprised 3000 to 4000 burials.
Camps and Camps (1964: 87) have
shown that the megalithic builders were
peasants and sheep herders. The clay
material used to manufacture the vessels found in some of the tested burials
is distributed into three groups of distinct origins. The cemetery is interpreted
as ‘‘a high place used to materialize the
unity of widely scattered tribal segments’’ (Camps and Camps 1964: 87).
Another relevant example is provided
by the cemetery of Iwelen studied by Paris
(1990). It is located on the southern slope
of Wadi Greboun in the Aı̈r mountain
range, on the shore of an intermittent
river, and comprises 64 burials distributed
into five monument types. All the burials
have been excavated and range in date
from ca. 5000 B.P. (5020 6 250 B.P.) to ca.
900 B.P. (1160 6 90 B.P.). Copper artifacts
found in some tombs attest to a strong
connection with the Eghazer basin. A hab-
itation site associated with series of rock
paintings and engravings were also found
at the same locality. The long-term use of
large cemeteries was certainly connected
to some concept of territoriality.
A third case, that of the Fadnoun on the
northern side of the Tassili-n-Ajjer, presents an interesting pattern of distribution
of megalithic burials (Savary 1966). The
surveyed area, situated at 25° 289/26° 229
latitude N and 7°549/8° 549 longitude E,
with elevation above sea level (asl) ranging from 600 m asl in the north to 1600 m
in the south, measures 4250 km2. The plateau landscape is tormented with deeply
incised south–north river valleys. These
valleys (Wadi Djerat, Imirhou, Ad, Tadjeradjeri, Ouret, Tiflamine, etc.) with remnant water sources and sometimes wide
flat bottoms, were probably used as
transhumance corridors by prehistoric
pastoralists who painted the artworks
found in the Tassili-n-Ajjer, in places generally situated at more than 1400 m asl. As
is the case for Wadi Djerat (Lhote 1976b),
thousands of rock engravings are found
along the wadi valleys. According to Savary (1966) analysis based on low-altitude
air photographs, there are no cemeteries;
megalithic burials are dispersed almost
evenly throughout the landscape, with a
preferential concentration in the southern
half of the surveyed area, where 306
clearly visible monuments of a total of 428
have been recorded. The average density
is approximately 1 monument per 10 km2,
varying from 1 per 8.54 km2 in the south to
1 per 13.5 km2 in the north. The Fadnoun
case suggests that prehistoric pastoralist
groups roaming in the area between the
Outer and the Inner Tassili may have
been constituted in average of relatively
small size units, without specific large
burial grounds.
As far as monumental burials are concerned, two different strategies are represented in the cases reviewed here. At
Iwelen and Djebel Mazela there are large
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
long-term cemeteries, probable gathering
‘‘high-places’’ for usually dispersed tribesmen. In the Fadnoun, tribe segments,
probably at the level of clans, adopted a
strategy consisting of shifting burial
places for the deceased without any noticeable preferential pattern of location
and clustering. Rock engravings, rock
paintings, megalithic burials, and occupation sites thus appear as meaningful archaeological signposts of prehistoric pastoral nomad territories.
PASTORAL-NOMADS TERRITORIES:
THE EGHAZER BASIN
Situated between 16°/18° 109 latitude N
and 6°/9° longitude, the In Gall–Tegiddan-Tesemt region, here termed the Eghazer
basin, is located between the Aı̈r mountain range in the East, the Tigidit cliff in
the South and the Azawagh valley in the
West. It is watered by numerous seasonal
streams, among which the Eghazer is the
most important (Poncet 1983). The area
surveyed by the Programme Archéologique
d’urgence (Poncet 1983; Bernus et al. 1984;
Paris 1984; Grébénart 1985) extends over a
surface of about 51,000 km2. This study
area can be divided into four morphological units: in the central position, there is
an extensive clayey depression watered
by the Eghazer river. The Aı̈r mountain
range, located in the East with numerous
rock-engraving stations (Lhote et al. 1987),
is bordered on its western side by foothills
with elevation varying from 400 to 600 m
asl. In the south, the Tigidit cliff stretches
200 km east–west.
The Aı̈r mountain range has been
poorly surveyed; it has, however, been observed that there are numerous valleys,
more or less deep with sand and silty sand
at their bottom. ‘‘The vegetation is characterized by narrow strips of grass and
trees at valleys bottoms, where southern
species (Acacia albida etc.), willingly visited by pastoralists can be found’’ (Poncet
153
1983: 26). The foothills zone comprises the
Tadarast plateau with series of deeply cut
valleys for water courses originating from
the Aı̈r mountain. During the rainy seasons, both areas are dotted with important
ponds, some of them deep enough to last
for entire dry seasons. As such, they are
optimal dry season camping areas for nomadic pastoralist groups (Poncet 1983: 23).
The Eghazer basin is composed of two
subunits: the clayey zone along the
Eghazer river drainage stricto sensu, surrounded on its northern and western periphery by sandy formations and areas of
drifting sand, along the Sekkiret valley. In
the south, the sandstone cliff of Tigidit,
with boulders and sometime tormented
landscape with isolated hilltops, sets the
southern limit of the Eghazer basin.
Past climatic history of the region is as
yet poorly understood. If we rely on research carried out on sediment cores from
Segguedine in the Ténéré (Pomel et al.
1991), preliminary palaeoclimatic work
conducted by Durand and Paris (1986) at
Chin Tafidet, and the general reconstitution of Saharan Holocene palaeoclimatic
sequences (Williams and Faure 1980), it
appears that the Late Holocene (ca. 5000
to 2000 B.P.) climate was characterized by
a neat trend toward aridity, with noticeable local variations: the climatic pattern
may have been that of contrasted seasons
of unequal duration, with a longer dry
season and a shorter wet season with
stormy moonsonal rains. Under these circumstances, the capacity of sediments to
retain water may have played a crucial
role in the timing of plant growth and
regional distribution of water and pasture.
The Eghazer Basin: Resources Potentials
For pastoral–nomadic societies, the
optimal socio-economic strategy may
have involved the development of sitelocation strategies comprising seasonal
moves from one part of the landscape to
154
AUGUSTIN F. C. HOLL
another. Water resources are divided
into two categories: fresh and saline. In
fact there is a gradation from exclusively
fresh to highly brackish springs and
ponds. The former are generally found
in the Aı̈r and the foothills zone and the
latter at major saline sources at such
places as Gélélé and Azelik wan Birni in
the central part of the Eghazer basin.
The nature of the sedimentary matrix
also determines two major zones: an extensive zone with a sand matrix in the
northern and western periphery of the
basin, surrounding a central clayey one.
In rainy seasons, saline water ponds, considered
important for the health of livestock, are found
almost everywhere in the clayey area; a dense
and strong annual vegetation grows almost everywhere after rains, thus constituting rich grazing lands. The flooded areas of the Eghazer wan
Agadez and its tributaries provide optimal conditions for the growth of annual species such as
Sorghum aethiopicum, which constitute first order
grazing lands. (Poncet 1983: 24)
High quality pastures from the clayey
areas, which seem to be confined to
lands situated below the 400 m asl contour line, do not last very long; the
growth season is longer in areas with
sand matrix where the dispersion and
lower quality of grazing lands are compensated by longer duration, which in
some contexts encompasses the whole
dry seasons. It is therefore not surprising that areas of salines sources are famous for their influence on the health of
livestock and are thus visited every year
by pastoral–nomadic folks (Bernus 1981)
during their annual cure salée. According
to Schulz (1991), in the clayey depression of the Eghazer basin, in years with
moderate rainfall, the average output of
a harvest of wild grain of Sorghum aethiopicum and Panicum laetum may have
amounted to 250 kg/ha, with an effective
quantity of grain varying from 130 to 150
kg/ha. Considering the productive capacities of such an environmental con-
text, it can be infered that the Eghazer
river drainage basin may have been a
crucial area for the reproduction and
maintenance of actual Late Holocene
pastoral–nomadic societies, from ca. 4500 to
2000 B.P.
Beyond sedimentary matrixes with their
distribution into complementary sandy and
clayey zones, the presence of saline sources
and the growth patterns of vegetation,
which were, without any doubt, crucial for
pastoral–nomadic societies, copper ore has
also been exploited in the study area. The
exact dating of the beginning of this technological innovation is still a matter of scientific debate. According to Grébénart
(1985), a Copper I sequence started around
ca. 2000 B.C., while for Killick et al. (1988), it
is not older than ca. 1000 B.C. Whatever the
case, ‘‘copper mineralization and even native copper are found in contact zones between the Agadez sandstone and the clayey
formation of the Eghazer, or more specifically, in the faulted contact zones’’ (Poncet
1983: 16). If considered as a scarce and strategic resource, attempts to control a copper
zone may have been implemented by actual
social systems. ‘‘A direct correlation between the presence of copper and some of
these tumulus will therefore not be surprising’’ (Bernus and Gouletquer 1976: 26).
If considered from the perspective of a
whole package of resources, comprising
soils, pastures, grain, salt, water, and copper, we will expect site location strategies
to be geared toward timely and optimal
exploitation of the Eghazer basin resources potentials. Such an exploitation
may have included flexible strategies with
seasonal moves from one resource area to
another, with attempts to single out key
zones with territorial markers such as
stone tumuli and burial grounds. Following these flexible strategies, some areas
and localities may have been settled for
short periods of stay, while other may
have been reoccupied more or less regularly, during the same season each year.
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
Archaeological data available today are
probably a palimpsest of these different
settlement strategies and their relatively
loose resolution does not allow a detailed
reconstitution of the dynamics of the past
flexible socioeconomic landscape. The
taxonomic typochronological approach,
with its emphasis on Saharan and Sahelian Neolithic, Copper I and Copper II,
and Early and Late Iron Age (Grébénart
1985), quickly turns into a handicap. With
most of the sites consisting of relatively
shallow surface scatters, several patterns
of the recorded archaeological evidence
cross-cut the boundaries of the inferred
chronological units based on the Stone–
Copper–Iron model. Our capacity to study
meaningful socioeconomic change is severely hampered. Copper metallurgy is
contemporaneous with the so-called Saharan and Sahelian Neolithic; vessels attributed to the Saharan neolithic are
found in Sahelian neolithic contexts as
well as in Copper I and II sites; Copper
technology occurred in a Late Stone Age
context with similar and long lasting stone
knapping and potting traditions. Pottery
from sites with evidence of iron metallurgy is different but copper production is
still present. The Eghazer basin record
provides an exciting spectrum of archaeological data, ranging in date from ca. 4500
to 2000 B.P., and related to technological
innovation and use of metal items, livestock herding strategies, spatial distribution and location of habitation sites,
megalithic burials, and cemeteries.
Distribution of Settlements and Nature
of Sites
Surveys have resulted in the recording
of several hundreds archaeological sites.
With the exception of the Aı̈r mountain
range, which has not been surveyed in
detail, evidence for human settlements
has been found in three other morphographic units: the foothills, the Tigidit
155
cliff, and the extensive depression of the
Eghazer drainage basin. The chronological resolution of our data base is far from
satisfactory; few sites have been tested
and dated, and many others are dated
from their associated cultural remains
found on the surface. Four categories of
settlements will be considered in this discussion: Late Stone Age sites, localities
with evidence of copper production, settlements with evidence of iron metallurgy,
and cemeteries. Only few of them have
been tested in varying degrees; it is the
case for 13 Late Stone Age localities, 10
with evidence of copper metallurgy, 9
with remains of iron production, and 30
with megalithic burials. Most habitation
sites are a palimpsest of multiple episodes
of occupation, congruent with the dynamics of pastoral–nomadic societies; the time
range under consideration spans over
2000 years, from ca. 4500 to 2500 B.P.
Megalithic burials and cemeteries are
lumped in the category of ‘‘preislamic’’
monuments, their chronological sequence
ranging from ca. 4000 to 1000 B.P.
Among the 148 recorded Late Stone Age
settlements, 64 are located in the Eghazer
depression, 10 in the foothills zone, and 74
along the Tigidit cliff (Fig. 2; Table 1). The
highest concentrations are found in the
northern southwest sandy zone, while 12
sites are located within the area of annual
cure salée. Two of the tested Late Stone
Age sites are located in the Eghazer depression: Chin Tafidet in the western
sandy zone and Anyokan in the central
clayey area. Tuluk 211, a stone axe workshop, is situated in the foothills zone,
while 8 others are found along the Tigidit
cliff. The surface extent of the tested settlements varies from a maximum of 28.20
ha at Shin Wasararan on the cliff plateau,
to a minimum of 0.12 ha at Afunfun 176,
179 and Tamat. No preserved dwelling
features have been recorded (Grébénart
1985; Paris 1984).
At Chin Tafidet, dated to ca. 4500 –3400
156
AUGUSTIN F. C. HOLL
FIG. 2. Long-term and cumulative Late Holocene pastoral–nomadic territories of the Eghazer basin.
B.P., 18 complete skeletons of cattle, 3 of
sheep/goats, and 2 of shepherd dogs have
been recorded (Paris 1984, 1992). The
presence of cut marks on cervical vertebrae suggest that these animals, 4 to 8
years old, were sacrificed and buried in
association with humans. Of the 75 recorded human burials, 12 have been excavated. The dead were buried without
grave goods in clusters of tombs compris-
TABLE 1
Distribution of Archaeological Sites According to Morphographic Units
Depression of
the Eghazer
Foothills of
the Aı̈r
Tigidit cliff
Morphographic units
Nb
%
Nb
%
Nb
%
Total
Néolithic sites
Sites with copper
Sites with iron
Undetermined
Megalithic cemeteries
Total
64
15
2
7
48
136
43,24
57,69
10,52
53,84
46,60
44,01
10
3
—
2
32
47
6,75
11,53
—
15,38
31,06
15,21
74
8
17
4
33
136
50
30,76
89,47
30,76
32,03
44,01
148
26
19
13
103
309
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
ing humans and cattle. Four such clusters
have been identified (Paris 1992). It is
worth noting the fact that shepherd dogs
and sheep/goats burials are not associated
with human burials. A similar feature of
patterned association between human
and cattle burials has been recorded at
Ikawaten in the north, but the collected
material is not yet published in detail.
From these mortuary data, it can be inferred that small groups of pastoral nomads used to bury their dead in one of
their major dry season camping areas, a
burial program emphasizing strong relationships between humans and cattle.
In the southeastern part of the study
area, Afunfun 161 and 177 are dated to ca.
3000 B.P. Respectively, 4 burials of 16 and
5 of 25 have been excavated. In the former
site, two of the excavated burials have
grave goods, consisting of pottery, while
one of the deceased has been buried with
an infant sheep. In the latter, from both
surface inspection and excavation, ceramic wares are attested in 18 burials with
one of the deceased buried with an infant
sheep. In general the population of grave
goods is more diversified, comprising ceramic wares, polished stone axes, stone
arm-rings, and landsnail shells. The material part of Afunfun burial program suggests an emphasis on sheep/goat husbandry.
Chin Tafidet, Ikawaten, and Afunfun
161 and 177 are extensive settlements.
They appear to have been focal points of
at least two variants of pastoral–nomadic
systems actually present in the Eghazer
basin: the western variant, comprising two
subvariants (Chin Tafidet and Ikawaten),
with an emphasis on cattle husbandry,
and the eastern variant, with sheep and
goats as predominant livestock. It is
highly probable that these pastoral–nomadic groups may have moved in the central part of the basin during the rainy seasons for the annual cure salée and the
harvest of wild grain of S. aethiopicum and
157
Panicum laetum. According to Schulz (1991),
in years with an average amount of rainfall, the productivity of natural fields of
wild grain was important enough for people to live without agriculture.
Compared to the number of Late Stone
Age settlements, that of sites with evidence of copper metallurgy suggests a
drastic decrease in the density of occupation (Fig. 2; Table 1). This observation is
certainly an artifact of the adopted typochronological framework. The recorded
material evidence and radiocarbon dates,
however, suggest that copper metallurgy
is but one facet of the actual socioeconomic systems. As far as spatial distribution of sites is concerned, 15 of the recorded sites are located in the central part
of the Eghazer depression, within the area
of the annual cure salée, with the highest
concentration in the copper ore zone
along the Guélélé–Azelik–Sekkiret anticlinal, while 3 are found in the foothills, and
8 along the Tigidit cliff. It can be inferred
that during their annual sojourn in the
depression, the procurement and processing of copper ore was imbedded in a
larger activity package, comprising livestock herding and harvesting of wild
grain. Settlements with evidence of copper metallurgy are dated from ca. 4000 to
1300 B.P. Among the tested sites, four are
located in the Eghazer depression, two in
the foothills zone, and four along the Tigidit cliff. Their surface extent varies from
20 to 0.03 ha (Tables 1 and 2). Evidence of
copper production is attested by series of
furnaces, slag heaps, and worn-out and
lost copper artifacts. The major copper
smelting sites have been found at Afunfun
162 and 175, Eres-n-Enadan, and Aghtauzu in the southeast and Azelik, Sekkiret valley, and Ikawaten in the northwest, while minor centers of production
have been recorded in the foothills zone at
Tuluk and Tyeral.
Settlements with evidence of iron metallurgy appear to be confined to the south
158
AUGUSTIN F. C. HOLL
TABLE 2
Major Characteristics of the Tested Settlements
Locality
Surface
(Ha)
Situation
Nature
Date B.P.
Peculiarities
?
3000
?
?
?
3400
3400
2900
3200
3400
3200/2800
?
?
Humans and animals
Humans and animals
—
—
—
Hundreds of tumulus
Humans and animals
—
—
—
—
—
Stone axes workshop
2800/2500
4100/3100
?
2900
2500/2000
—
2200/1300
3300/2900
2350
2400
Clusters of furnaces
Clusters of furnaces
—
—
Reworked
—
Clusters of furnaces
Clusters of furnaces
—
Tumulus
—
—
2450
2200/2000
—
—
—
—
2100
1 Tumulus
—
—
6 Pits
Late Stone Age settlements
1-Afunfun 161 burials
2-Afunfun 176 burials
3-Afunfun 177
4-Afunfun 179
5-Anyokan
6-Asaquru
7-Chin Tafidet burials
8-Efey Washaran
9-Mio
10-Orub
11-Shin Wasararan
12-Tamat
13-Tuluk 211
Cliff
Cliff
Cliff
Cliff
Depression
Depression
Depression
Cliff
Cliff
Cliff
Cliff
Cliff
Foothills
2,00
0,12
1,50
0,12
12,5
110
2,36
0,19
1,13
2,16
28,2
0,12
1,75
Camp/village
Village/camp
Nécropole
Village/camp
Camp
Cemetery
Village/camp
Village/camp
Village/camp
Village/camp
Village/camp
Camp
Camp
Settlement with evidence of copper
1-Afunfun 162
2-Afunfun 175
3-Afunfun 216
4-Agtauzu 178
5-Azelik 210
6-Eres n Enadan
7-Ikawaten
8-Sekiret
9-Tuluk
10-Tyeral
Cliff
Cliff
Cliff
Cliff
Depression
Depression
Depression
Depression
Foothills
Foothills
12,00
20,00
?
12,56
8,00
17,66
5,00
?
0,03
1,75
Village?
Camp
Camp
Village?
Camp
Camp
Camp
Camp
Camp
Camp
Settlements with evidence of iron
1-Chin Oraghen 105
2-Efey Washaran
3-Ekne wan Ataram
4-In Taylalen II 15
5-Jibo 136
6-Mio 169
7-Shin Ajeyn
8-Tamat 157
9-Teguef n’Agar
Cliff
Cliff
Cliff
Cliff
Cliff
Cliff
Cliff
Cliff
Cliff
1,75
0,50
3,00
1,50
0,50
1,13
1,75
1,00
0,78
of the study area, with 17 of a total of 19
sites located along the Tigidit cliff (Fig. 2;
Table 1). It is highly improbable that the
central part of the Eghazer drainage was
not visited during this period. This
skewed distribution has therefore to be
linked with the natural distribution of iron
ore. The 9 tested localities with early evi-
Village
Village/camp
Village
Village
Camp
Village
Village
Camp
Village
—
1 Pit
—
10 Pits
dence of iron metallurgy, dated from ca.
2600 to 2100 B.P. and are located along the
Tigidit cliff, most of them on the plateau
(Fig. 2; Table 1). Their surface extent varies from 3 to 0.6 ha; settlements are now
genuine small villages of sedentary or
semisedentary communities. Among the
tested sites, remains of iron smelting fur-
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
159
TABLE 3
Characteristics of the Tested Megalithic Cemeteries
Site
Location
Surface
(Ha)
No. of
burials
No. of tested
burials
Diversity
Date B.P.
1-Afunfun 8
2-Agadez
3-Asaquru
4-Azelik
5-Imosaden
6-Shi Mumenin
7-Shin Wasadan
8-Tegaza
9-Tezzigart
10-Tin Tegeis
Cliff
Foothills
Depression
Depression
Foothills
Cliff
Depression
Depression
Foothills
Depression
?
0.35
.10
?
?
?
22.50
7
?
10
?
77
.100
20
31
8
177
25
?
50
1
6
7
1
2
2
4
5
1
1
?
1
Maximum
1
?
?
5
5
7
3
?
—
3350
—
—
730
2450
—
—
—
naces have been found at Ekne wan Ataram, In Taylalen II 15, and Teguef
n’Agar. In general, the total density and
diversity of archaeological remains is
higher. Series of pits without any connection with iron production have been recorded at In Taylalen II 15, Shin Ajeyn,
and Teguef n’Agar (Grébénart 1985: 263–
330); the diameters of the excavated pits
vary from 1.40 to 1.00 m, with depths varying from 0.20 to 0.70 m. Due to the loose
sedimentary matrix in which these pits
were dug, it is highly probable that they
were originally deeper and have collapsed
after abandonment. Considered in this
perspective, these pits may have been
used as storage features distributed
among different domestic units of actual
settlers, even if positive plant macroremains have not been recorded. The settlements of this period may correspond to a
relative decline of the previous pastoral–
nomadic socioeconomic systems, with a
shift toward a bulkier and long-lasting
mixed-farming sedentary communities.
Preislamic burial grounds are almost
evenly distributed in the study area; 48
are located in the depression of the
Eghazer, 32 in the foothills zone, and 33
along the Tigidit cliff (Table 1). Of a total
of 103, 39 are found in the central area of
the annual cure salée. It is as if megalithic
burial grounds were used as territorial
markers to emphasize claims for exclusive control of strategic resources from
the Eghazer drainage basin. Very few
(30) megalithic burials have been tested;
one of them from the large cemetery of
Asaquru, is, however, dated as early as
3400 B.P. (Table 3).
In general, if we combine all categories
of sites, 136 are located in the Eghazer
depression and along the Tigidit cliff,
while 47 have been recorded in the foothills zone (Table 1; Fig. 2). The most remarkable feature as far as the regional
distribution of settlement is concerned, is
the contrast and reversal of the density of
habitation relative to burial sites in the
Eghazer depression. Burial sites with
megalithic tombs are highly visible, long
lasting, and even permanent; as such they
were probably used as signposts geared to
signal the appropriation of territories.
Symbols of power were probably crucial
within this new social juncture; they may
have generated a peculiar kind of social
demands which, in turn, may have enhanced and sustained the invention or
adoption of metal production.
Due to the strict contemporaneity between ‘‘Saharan’’ and ‘‘Sahelian Neolithic’’ on the one hand and both taxonomic units and localities with evidence of
copper metallurgy (i.e. Copper I and II) on
the other, added to the impossibility to
160
AUGUSTIN F. C. HOLL
differentiate archaeological artifacts belonging to each dwelling episode in the
recorded surface sites certainly visited
several times for centuries and even millenia, the Late Holocene occupation sequence of the Eghazer basin can be divided into two major periods. Period I (ca.
4500 to 2500 B.P.) was characterized by the
presence of pastoral–nomadic societies divided into two major variants: one focused
on cattle husbandry and the other on
sheep and goats. Social differentiation
and territorial strategies singled out by
stone-tumuli cemeteries have generated
and sustained the development of craft
specialization geared to produce highly
valued goods in copper. Period II follows
and lasted from ca. 2600 B.P. to 2000 –1500
B.P. It is characterized by the adoption of
iron metallurgy, with a shift of settlement
toward the south, with bulkier probably
more permanent dwelling units with storage features. This period coincided with
the decline of ancient pastoral–nomadic
lifestyle, ending with the arrival of new
settlers (Early Tuareg) from the north possessing camels.
Distribution Patterns of Metal Artifacts
Metals artifacts, in copper and iron,
have been recorded in burials as well as
habitation sites. Assemblages collected
from habitation and metal producing sites
are more diversified and heterogenous.
Stone tools were still in use in daily life
activities. Metal tools useful for productive tasks are almost absent from the studied assemblages; this absence is ambiguous as such tools may have been highly
curated and when worn out, recycled.
Whatever the case, however, the recorded
sets of metal implements probably result
from palimpsests of multiple episodes of
occupation, through lost and discard of
worn out artifacts. Assemblages from
burials comprise a narrower spectrum of
artifacts, essentially ‘‘jewels’’ and arms,
the former preferentially in copper and
the latter in iron (Table 4). These features
of metal artifacts distribution suggest a
patterned hierarchy of metal goods resulting from quite neat social demands. In
this perspective, jewels in copper and iron
weapons seem to have been parts of social
tactics and strategies of distinction.
Cemeteries and Social Territories
Burial evidence can be divided into
three major categories: isolated tombs
within habitation sites as is the case at
Efey Washaran 151 and 183, Jibo 136, and
Shin Wasasaran; clusters of tombs as attested at Chin Tafidet, Afunfun 161, and
Afunfun 177 (Paris 1984); and hundreds of
stone tumuli, some of them isolated, but
more often situated in small burial
grounds or within large cemeteries extending over several hectares. In this part
of our discussion, we will focus on burial
grounds with stone tumuli.
Several hundred burials in megalithic
monuments have been recorded; they are
distributed in varying concentrations into
116 localities. Excavations have been carried out at 10 localities, with the number
of tested burials varying from 8 at Shi
Mumenin to 1 at Afunfun 8, Azelik, Tezzigart, and Tin Tegeis. Five of the tested
sites, Asaquru, Azelik, Shin Wasadan,
Tegaza, and Tin Tegeis, are located in the
Eghazer depression; 3, Agadez, Imosaden,
and Tezzigart, are in the foothills zone;
and 2, Afunfun 8 and Shi Mumenin, along
the Tigidit cliff (Table 3). Uncovered archaeological data are not always reported
with the needed accuracy. From the available information, it appears that surface
extent of cemeteries varies from 22.5 to
0.35 ha, and the number of burials varies
from 4 to 177 (Poncet 1983; Paris 1984).
Types of megalithic monuments are extremely diversified, within as well as between sites. Stone circles, quadrangular
constructions, long rectilinear tumuli, and
161
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
TABLE 4
General Distribution of Metal Artifacts and Slag
Copper artifacts
Locality
Sites with copper
Afunfun 162
Afunfun 175
Aghtauzu 178
Azelik
Eres n Enadan
Ikawaten 185
Tuluk
Tyeral
Sites with iron
Chin Oraghen 105
Efey Washaran 151 (tomb)
Efey Washaran 183 (tomb)
Ekne wan Ataram
In Taylalen II 15
Jibo 136
(tomb)
Mio 169
Tamat 157
Teguef n’Agar
Shin Ajeyn
Shin Wasararan (tombs)
Megalithic burials
Asaquru
Mon. D
Afunfun 8
Tegaza
Mon. 4
Mon. 5
Tezzigart Mon. 2
Mon. 3
Iron artifacts
Total
Weapons
Jewels
Others
Weapons
Jewels
Others
Copper
Iron
3
2
2
2
2
2
2
2
4
2
2
2
1
2
2
2
15
3
1
6
4
1
2
2
5
2
1
1
2
2
2
2
8
4
1
2
2
2
2
2
26
6
1
8
2
2
2
2
23
3
1
6
5
1
2
2
29
10
1
9
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
3
3
10
3
2
2
2
2
2
2
2
1
2
1
2
2
2
2
2
2
2
2
2
2
2
2
3
2
2
2
2
1
2
2
4
1
1
1
2
2
2
2
1
2
2
3
2
7
2
2
2
2
1
2
2
2
2
2
4
2
11
2
2
1
2
4
3
10
3
2
2
2
2
2
2
2
2
2
4
2
2
1
2
1
9
2
22
2
1
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
6
2
1
2
1
2
2
2
2
1
1
1
2
2
2
2
2
2
2
2
2
2
1
2
6
2
1
2
2
1
2
2
2
1
Note. Mon., monument.
crescent-shaped, tronconic, disc-shaped
monuments have been recorded. Stone
circles located in the foothills zone are
considered to be early Berber burials, and
quadrangular constructions from the
western part of the Tigidit cliff dated to
the first millenium A.D. (Paris 1984) are
not taken into account in this work.
The recurrent use of the same area to
bury the deceased of a community is in
itself an important social fact. In this regard, the number of burials, their bulk,
the total extent of the cemetery, and the
location of tombs relative to each other are
highly meaningful. Considered from this
research perspective, even the location of
burial grounds relative to each other all
over the landscape can be linked to the
dynamics of social systems and conceptualized as social strategies geared to the
long-term buildup of a social landscape.
The process may have been discontinuous
in some areas and continuous in others,
according to the vagaries of climatic hazards, social interaction, and a combination
of both.
Recorded cemeteries can be ranked according to their size and number of mega-
162
AUGUSTIN F. C. HOLL
TABLE 5
Hierarchy and Regional Spatial Patterning of Megalithic Cemeteries
Rank
I (.100)
Southwestern
group
Eastern group
Tuluk
(TTA 16)
12
(TTA 25, 26, 28, 32,
33, 35, 40, 41, 42,
47, 51, 52)
—
Shin Wasadan
(TTS 48)
8
(TTA 4, 7, 9, 15,
TTS 9, 10, 11,
31)
3
(TTS 2, 80, 82)
7
20
5
17
III (5-19)
1
(IG 27)
Asawas
(TTA 44)
12
(AG 32, 33, 34, 39, 62,
73, 75, 106, 108,
109, 117)
2
(AG 35, 53)
IV (1-4)
Total
10
16
8
22
II (20-100)
Anyokan
(AG 31)
4
(AG 30, IG 21,
23, 32)
Northern
group
Southeastern group
lithic burials. Five of them, Anyokan, Asawas, Tuluk, Shin Wasadan, and Asaquru,
are large cemeteries with more than 100
monuments each. One of the tested burial
from Asaquru is dated to ca. 3350 B.P., and
another one, Monument D, contained one
iron spearhead and one copper and two
iron jewels. At Shin Wasadan, one tested
burial is dated to ca. 2450 B.P. If considered as nodal points of the political landscape, these large cemeteries located on
the border zone of the cure salée area seem
to control almost equivalent territorial
units, with the Eghazer river as a boundary line (Fig. 2).
Cemeteries with 20 to 100 monuments
are unevenly distributed among inferred
territorial units, their frequency in each
varying from 4 to 12 (Table 5). They were
probably burial grounds for lower segments of actual pastoral–nomadic societies. Smaller cemeteries, with 5 to 19
and 1 to 4 monuments may have been
used for the inhumation of smaller clan
members, extended families, or households units. In order to build a large
megalithic monument, mobilization of a
labor force is crucial; the status of a deceased is therefore indirectly indicated
by the material characteristics of his or
her burial. The deceased were also potential ‘‘consumers’’ of prestige items
Western group
Asaquru (IG 2)
6
(TTS 3, 38, 49,
53, 74, 88)
4
(TTS 75, 83, 92,
93)
5
16
produced by crafts specialists. If considered from the perspective of the dynamics of pastoral–nomadic societies, it can
be said that at the climax of its Late
Holocene occupation, the Eghazer basin
may have been part of the territorial
range of five tribal units, each comprising varying number of segments (clans,
extended families, and domestic units).
Depending on actual circumstances, fission/dispersal or fusion/aggregation, conflicts and competition, as well as cooperation and solidarity, may have generated
concentration and/or dispersal of prehistoric pastoral–nomadic folk. In such a context, metal producers had provided their
kinsmen with emblems of prestige and
power.
CONCLUDING REMARKS
Livestock husbandry was adopted following a general ecological crisis in West Africa. The direction of expansion is clearly a
north/northeast–south/southwest one. Cattle and sheep/goats seem to have been part
of the same package, with variation from
area to another. Agro-pastoral mixed farming societies develop in the south along the
shifting Sahara–Sahel boundaries, in the
Dhar Tichitt in Mauretania, the southwestern part of the Chadian plain in Nigeria,
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
and possibly, in the central part of the
Tilemsi valley in Mali. Further north, in
Western Sahara, Taoudenni basin, upper
Tilemsi, Eghazer basin, as well as the Aı̈r
and the Ténéré, Late Stone Age prehistoric
herder–collectors adopted a properly pastoral–nomadic lifestyle. In general, the landscape was dotted with isolated or clustered
burials, rock-art stations, and intermittent
habitation sites, de facto territorial markers,
in ever changing social territories.
REFERENCES CITED
Amblard, S., and R. Vernet
1984 Des gravures rupestres intégrées à une
structure d’habitat: L’exemple d’Akhreijit.
Journal des Africanistes 54: 67–78.
Amblard, S., and J. Pernès
1989 The identification of cultivated pearl millet
(Pennisetum) amongst plant impression on
pottery from Oued Chebbi (Dhar Oualata,
Mauritania). The African Archaeological Review 7: 117–126.
Andah, B.
1979 The Later Stone Age and neolithic of Upper
Volta viewed in a West African context.
West African Journal of Archaeology 9: 87–110.
Andanquah, J.
1993 The Kintampo complex: A case study of
early sedentism and food-production in
sub-sahelian West Africa. In The archaeology
of Africa: Foods, metals and towns, edited by T.
Shaw, P. Sinclair, B. Andah, and A. Okpoko,
pp. 255–260. Routledge, London.
Ballouche, A., K. Kuppers, K. Neumann, and H. P.
Wotzka
1993 Aspects de l’occupation humaine et de
l’histoire de la végétation au cours de
l’Holocène dans la région de la chaı̂ne de
Gobnangou, SE Burkina Faso. Berichte des
Sonderforschungsbereichs 268(I): 13–31.
Bar Yosef, O., and A. Khazanov (eds.)
1992 Pastoralism in the Levant: Archaeological materials in anthropological perspectives. Prehistory Press, Madison, WI.
Bernus, E.
1981 Touaregs nigériens: Unité culturelle et diversité
régionale d’un peuple pasteur. ORSTOM,
Paris.
163
Bernus, E. S. Bernus, P. Cressier, P. L. Gouletquer,
and Y. Poncet
1984 La Région d’In Gall - Tegidda-n-Tesemt (Niger)
I: Introduction, Méthodologie - Environnement. Etudes Nigériennes, Niamey.
Bernus, S., and P. L. Gouletquer
1976 Du cuivre au sel: Recherches ethno-archéologiques sur la région d’Azélik (campagnes 1973–1975). Journal des Africanistes
46(1–2): 7– 68.
Beyries, S., and E. Boeda
1981 Relevés des gravures rupestres sur le Dhar Tichitt. Unpubl. DEA thesis, University Paris
X, Nanterre.
Breunig, P.
1995 Gajiganna und Konduga: Zur Frühen Besiedlung des Tschadbeckens in Nigeria. Beiträge zur Allgemeinen und Vergleichenden Archäologie 15: 3– 48.
Breunig, P., K. Neumann, and W. Van Neer
1996 New research on the Holocene settlement
and environment of the Chad basin in Nigeria. African Archaeological Review 13: 111–
145.
Camps, G., and H. Camps
1964 La Nécropole Mégalithique de Djebel Mazela.
Arts et Métiers Graphiques, Paris.
Casimir, M. J.
1992 The dimensions of territoriality: An Introduction. In Mobility and territoriality, edited
by M. J. Casimir and A. Rao, pp. 1–26. Berg.
New York/Oxford.
Clark, J. D., M. A. J. Williams, and A. B. Smith
1973 The geomeorphology and archaeology of
Adrar Bous, central Sahara: A preliminary
report. Quaternaria 17: 245–297.
Clutton-Brock, J.
1993 The spread of domestic animals in Africa. In
The archaeology of Africa: Foods, metals and
towns, edited by T. Shaw, P. Sinclair, B. Andah, and A. Okpoko, pp. 61–79. Routledge,
London.
Connah, G.
1981 Three thousand years in Africa. Cambridge
Univ. Press, Cambridge.
Cribb, R.
1991 Nomads in archaeology. Cambridge Univ.
Press, Cambridge.
Decobert, M., and N. Petit-Maire
1985 An early neolithic midden and necropolis in
the Malian Sahara. Nyame Akuma 26: 26 –32.
Durand, A., and F. Paris
1986 Peuplements et climats Holocènes de
l’Azawagh (Niger nord-occidental): Premiers rèsultats. In Changements globaux en
164
AUGUSTIN F. C. HOLL
Afrique durant le Quaternaire, pp. 127–130.
ORSTOM, Paris.
Elouard, P.
1979 Mollusques testacés. In Le Sahara Atlantique
à l’Holocène: Peuplement et écologie, edited by
N. Petit-Maire, pp. 228 –232. Mémoires du
CRAPE, Alger.
Flight, C.
1976 The Kintampo culture and its place in the
economic prehistory of West Africa. In Origins of African plants domestication, edited by
J. Harlan, A. B. M. Stemler, and A. de Wet,
pp. 211–221. Mouton, the Hague/Paris.
Frankfort, H.-P. (ed.)
1990 Nomades et Sédentaires en Asie Centrale: Apports de l’Archéologie et de l’Ethnologie. Editions du CNRS, Paris.
Gaussen, J., and M. Gaussen
1988 Le Tilemsi Préhistorique et ses Abords: Sahara et
Sahel Malien. Editions du CNRS, Paris.
Gauthier, A.
1991 Mammifères fossiles holocènes du Sahara
malien: Interprètation paléoécologique. In
Paléoenvironnements due Sahara: Lacs Holocènes à Taoudenni (Mali), edited by N.
Petit-Maire, pp. 173–176. Editions du CNRS,
Marseille.
Grébénart, D.
1985 La Région d’In Gall - Tegidda-n-Tesemt (Niger)II: Le Néolithique Final et les Débuts de la
Métallurgie. Etudes Nigeriennes, Niamey.
Grigson, C.
1991 An African origin for African cattle?—some
archaeological evidence. The African Archaeological Review 9: 119 –144.
Guérin, C., and M. Faure
1983 Mammifères. In Sahara ou Sahel: Quaternaire
récent du bassin de Taoudenni (Mali), edited
by N. Petit-Maire and J. Riser, pp. 239 –272.
Marseille.
Holl, A.
1985a Subsistence patterns of the Dhar Tichitt
Neolithic, Mauretania. The African Archaeological Review 3: 151–162.
1985b Background to the Ghana empire: Archaeological investigations on the transition to
statehood in the Dhar Tichitt region. Journal
of Anthropological Archaeology 4: 73–115.
1986 Economie et Société Néolithique du Dhar Tichitt
(Mauritanie). Editions Recherches sur les Civilisations, Paris.
1989 Social issues in Saharan prehistory. Journal
of Anthropological Archaeology 8: 313–354.
1990 Les formes du pastoralisme au Sahara néolithique. In Nomades et Sédentaires en Asie
Centrale: Apports de l’Archéologie et de
l’Ethnologie, edited by H.-P. Frankfort, pp.
141–155. Editions du CNRS, Paris.
1993 Late neolithic cultural landscape in southeastern Mauritania: An essay in spatiometrics. In Spatial boundaries and social dynamics,
edited by A. Holl and T. E. Levy, pp. 95–133.
International Monographs in Prehistory,
Ann Arbor.
1994 The cemetery of Houlouf in northern Cameroon (AD 1500 –1600): Fragments of a past
social system. The African Archaeological Review 12: 133–170.
1995a Réseaux d’échanges préhistoriques dans la
plaine tchadienne. Sahara 7: 17–28.
1995b Pathways to elderhood: Research on past
pastoral iconography, the paintings from
Tikadiouine (Tassili-n-Ajjer). Origini 18: 69 –
113.
Hugot, G.
1977 Un Secteur du Quaternaire Lacustre Mauritanien: Tichitt. Mémoires de l’IMRS, Nouakchott.
Khazanov, A.
1984 Nomads and the outside world. Cambridge
Univ. Press, Cambridge.
Killick, D., N. J. Van der Merwe, R. B. Gordon, and D.
Grébénart
1988 Reassessment of the evidence for early metallurgy in Niger, West Africa, Journal of Archaeological Science 15: 367–394.
Lhote, H.
1976a Vers d’autres Tassili. Arthaud, Paris.
1976b Les Gravures Rupestres de l’Oued Djerat, Tassili-n-Ajjer. Mémoires de CRAPE, Alger.
Lhote, H., P. Colombel, and H. Bouchard
1987 Les Gravures du Pourtour Occidental et du centre de l’Aı̈r. Recherche sur les Civilisations,
Paris.
Lhote, H., and R. Tomasson
1967 Gravures rupestres de la haute vallée du
Tilemsi (Adrar des Iforas, République du
Mali). In Sixieme Congrès Panafricain de Préhistoire, pp. 235–241. Dakar.
Lihoreau, M.
1993 Djorf Torba: Nécropole Saharienne Antéislamique. Karthala, Paris.
MacDonald, K. C., and P. Allsworth-Jones
1994 A Reconsideration of the West African macrolithic conundrum: New factory sites and
an associated settlement in the vallée du
Serpent, Mali. The African Archaeological Review 12: 73–104.
McIntosh, S. K., and R. J. McIntosh
1983 Current directions in West African Prehistory. Annual Review of Anthropology 12:215–58.
LIVESTOCK HUSBANDRY, PASTORALISMS, AND TERRITORIALITY
Munson, P. J.
1971 Tichitt tradition: A late prehistoric occupation in
the southwestern Sahara. Ph.D. thesis, University of Illinois at Urbana–Champaign.
Paris, F.
1984 La Région d’In Gall - Tegidda-n-Tesemt (Niger)III: Les Sépultures du Néolithique Final à
l’Islam. Etudes Nigeriennes, Niamey.
1990 Les sépultures monumentales d’Iwelen (Niger). Journals des Africanistes 60: 47–74.
1992 Chin Tafidet, village néolithique. Journal des
Africanistes 62: 35–53.
Person, A., N. E. Saoudi, and S. Amblard
1995 Nouvelles recherches, objectifs et premiers
résultats sur les paléoenvironnements Holocéne des sites archéologiques de la région des
Dhars Tichitt et Oulata (Mauritanie Sud-Orientale). Journal des Africanistes 65: 9 –29.
Petit-Maire, N.
1979a Le Sahara Atlantique à l’Holocène: Ecologie et
Peuplement. Mémoire de CRAPE, Alger.
1979b Prehistoric palaeoecology in the Sahara Atlantic coast in the last 10 000 years: A synthesis. Journal of Arid Environment 2: 85– 88.
1980 Holocene biogeographical variations along
the northwestern African coast (28°–19° N):
Paleoclimatic implications. Palaeoecology of
Africa 12: 365–377.
1991 Paléoenvironnements du Sahara: Lacs Holocènes à Taoudenni (Mali). Editions du
CNRS, Marseille.
Petit-Maire, N., and J. Riser (eds.)
1983 Sahara ou Sahel? Le Quaternaire récent du Bassin de Taoudenni (Mali). Editions du CNRS,
Marseille.
Petit-Maire, N., J. C. Celles, D. Commelin, and M.
Raimbault
1983 The Sahara in Northern Mali: Man and his
environment between 10,000 and 3,500
years b.p. (preliminary results). The African
Archaeological Review 1: 105–125.
Phillipson, D.
1982 The Later Stone Age in sub-Saharan Africa.
In The Cambridge history of Africa I, edited by
J. D. Clark, pp. 410 – 477. Cambridge Univ.
Press, Cambridge.
Pomel, S., E. Schulz, and R. Baumhauer
1991 Changement dans le région de bordure des savanes à l’Holocène: Seguedine (Kaouar, Nord-Est
Niger). Interactions de l’homme et du climat.
Paper presented at the Cinquième Colloque
Mega-Tchad: L’Homme et le Milieu végétal,
18 –20 Septembre 1991, Paris.
Poncet, Y.
1983 La Région d’In Gall - Tegidda-n-Tesemt (Niger): Atlas. Etudes Nigeriennes, Niamey.
165
Raimbault, M.
1991 Mission du CNRS au Sahara malien. In Archeologie du Mali, edited by M. Raimbault
and K. Sanogo. Karthala, Paris.
Roset, J. P.
1974 Un gisement néolithique ancien près de Fachi (Erg du Ténéré). Cahiers ORSTOM Sciences Humaines 11: 105–110.
1983 Nouvelles données sur le problème de la
néolithisation du Sahara méridional: Aı̈r et
Ténéré, au Niger. Cahiers ORSTOM Sciences
Humaines 13: 119 –142.
Savary, J. P.
1966 Monuments en pierres sèches du Fadnoun (Tassili n Ajjer). Arts et Métiers Graphiques,
Paris.
Schulz, E.
1991 A neolithic revolution in the Sahara? Background to the ressources of human alimentation
during the Holocene of the Central and Southern
Sahara. Paper presented at the Cinquième
Colloque Mega-Tchad: L’Homme et le Milieu végétal, 18 –20 Septembre, Paris.
Smith, A. B.
1975 A note on the flora and fauna from the
post-palaeolithic sites at Karkarichinkat
Nord and Sud. West African Journal of Archaeology 5: 201–204.
1980 Domesticated cattle in the Sahara and their
introduction to West Africa. In The Sahara
and the Nile, edited by M. A. J. Williams and
H. Faure, pp. 489 –501. Balkema, Rotterdam.
1986 Cattle domestication in north Africa. The
African Archaeological Review 4:197–203.
1992a Pastoralism in Africa: origins and development
ecology. Hurst, London.
1992b Origins and spread of pastoralism in Africa.
Annual Review of Anthropology 21: 125–141.
Stahl, A. B.
1993 Intensification in the west African Late
Stone Age: A view from central Ghana. In
The Archaeology of Africa: Foods, Metals and
Towns, edited by T. Shaw, P. Sinclair, B.
Andah, and A. Okpoko, pp. 261–273. Routledge, London.
Togola, T.
1996 Iron Age occupation in the Méma region,
Mali. African Archaeological Review 13: 91–
110.
Williams, M. A. J., and H. Faure (eds.)
1980 The Sahara and the Nile. Balkema, Rotterdam.