Shuwa-Arab economics:
livestock management and animal consumption
Augustin F.C. Holl
Museum of Anthropology, Department of Anthropology, Center of AfroAmerican and African Studies, University of Michigan, Ann Arbor, USA
Abst r a c t / Zu s a m me n fa s s u ng
Shuwa-Arab communities reached the Chad basin in the late 14th century AD and from then on spread to their
present day locations in the Chad republic, Cameroon, Niger, and Nigeria. The ethnoarchaeological research
conducted on their settlements in the Houlouf region in northern Cameroon allowed for an interesting assessment
of their patterns of subsistence, the importance of livestock in their daily life, as well as patterns of animal use
and meat consumption. The taphonomic trajectories of faunal remains and their implications for the understanding of Shuwa-Arab economics are evaluated at three settlement types: permanent villages, semi-permanent villages, and dry season camps, sampled during three separate seasons of fieldwork.
Shuwa-Araber Gemeinschaften erreichten das Tschadbecken im späten 14. Jahrhundert AD. Von dort aus breiteten sie sich zu ihren heutigen Ansiedlungen in Tschad, Kamerun, Niger und Nigeria aus. Die an ihren Fundplätzen in der Houloufregion Nordkameruns durchgeführten ethnoarchäologischen Untersuchungen ermöglichen
eine interessante Einschätzung hinsichtlich Subsistenzmuster, Bedeutung von Vieh im Alltag sowie Nutzung von
Tieren und Fleischverzehr. Die taphonomischen Prozesse denen Tierreste unterworfen sind, sowie deren Konsequenzen für das Verständnis der Wirtschaft der Shuwa-Araber werden anhand von drei Siedlungstypen evaluiert,
nämlich Dörfern mit dauerhafter Bewohnung, Dörfern mit semi-permanenter Bewohnung und Trockenzeitlagern,
deren Beprobung während dreier Geländeaufenthalte stattfand.
Keywords
Ethnoarchaeology, Shuwa-Arab, patterns of subsistence, site layout, livestock management, meat
consumption, animal bones, pastoral economics
Ethnoarchäologie, Shuwa-Araber, Subsistenzmuster, Fundplatzgrundriss, Viehwirtschaft, Fleischverzehr, Tierknochen, Weidewirtschaft
I nt roduct ion
Present-day Shuwa-Arab communities have adjusted to
the vagaries of the shifting balance of power between
the multiple communities of the Chad basin and Central
Sudan. Many are now fully sedentary, relying more on
agriculture than livestock husbandry. Others remain
fully semi-nomadic, residing one part of the year in
their main rainy season villages, while spending the dry
season in camps that can be located as far as 150 kilometers away. Relationships to neighboring communities, the inhabitants of which are predominantly sedentary and employ agricultural and fishing subsistence
strategies, are very complex (Holl 2003). They range
from partnership to rivalry, articulated on unequal power relationships. Even if they are sedentary and rely
more on agricultural products, the management of livestock and the handling of animal resources are what set
the Shuwa-Arab apart in the Chadian plain ethnic mosaic (Podlewski 1966, 1971, Lebeuf 1969, 1987, Zeltner
1970, 1979, 1980, 1988, Conte & Hagenburger-Sacripanti 1977, Hagenburger-Sacripanti 1977a, b, Frantz
1981, Tijani 1986, Holl & Levy 1993). How are their
socio-economic systems organized? How do they manage their livestock resources? And what patterns of consumption of animal resources did they develop? These
are the key issues addressed in this paper.
The investigation of settlement structures and patterns of
subsistence in contemporary societies is one of the success stories of ethnoarchaeological research (Yellen 1977,
Binford 1978, Hodder 1982, Atherton 1983, Agorsah
1985, 1990, David & Kramer 2001, Holl 2003). This kind
of research analyzes the patterns emerging from residual
regional settlements. Despite the dominant optimism, the
retrodiction of ethnoarchaeological findings to past archaeological situations is severely handicapped by significant differences in scale. If ethnoarchaeology has any
impact on archaeological research beside the obvious
situation of documented historical and cultural continu-
38
Augustin F.C. Holl
Fig. 1: Map of the study area
ity, it is through its role as a “hypotheses generator”. Ethnoarchaeology does not solve historical problems. It provides a series of hypotheses that can help make sense of
the archaeological record at hand (Holl 2003). This iconoclastic position is not shared by most Africanist ethnoarchaeologists despite its obvious accuracy, as shown in the
analysis to follow.
or Syria, used Egypt only as a corridor to the west or
the south; others remained for a shorter or longer time.
Being notoriously anarchic, turbulent and always in opposition to any outside control, they represented an element of danger and unrest, too often causing disruption
of economic life by their raids on villages or by holding
up caravans” (Hrbek 1977: 69).
Short Historical Background
The Mameluk dynasty of Egypt launched a number of
heavy military campaigns against Arab mobile pastoralists and forced many out of the Nile valley. These groups
moved west, and spread slowly to reach the Chad basin
in the 14th century AD. They adapted to a broad range
of social and political circumstances, as part of the elite
in some cases and as pariahs in others (Zeltner 1970,
1979, Garcin 1985, Bianquis 1990, Holl & Levy 1993,
Holl 2003). From the 10th to the 14th century, the territories situated between the Nile and northern Darfur
were settled by Arabs (MacMichael 1922). Their changing adaptation to their new ecological and cultural settings generated a great deal of regional variations outlined by Hrbek (1977). Those settled along the
northeastern part of the territorial range continued to
rear camels and sheep. They consist of different clans of
various tribes but nonetheless formed large groups
known as the Kababish (from Kabsh i.e. “ram”). Those
who migrated further south, to southern Darfur, southern Kordufan, and westwards to Wadday and the Lake
Chad region, adopted cattle husbandry and became
known as Baqqara (from Baqara i.e. “cow”). “Although
they preserved their nomadic way of life and their tribal
system intact, they intermarried with the local peoples,
Shuwa-Arab communities started to settle in the eastern confines of the Chad basin at the end of the 14th
century AD. Their arrival in that part of Central Sudan
was the result of long and protracted migratory histories that took their ancestors from their Arabian homeland to the Nile delta in the 7th century (Hasan 1973,
Fisher 1975, 1977, Hiskett 1976, 1994, Hrbek 1977,
Garcin 1985, Bianquis 1990). The ancestors of presentday Shuwa-Arab were among the first Arab tribes to
settle in Egypt after its conquest by Moslem troops.
Some of these groups decided to stick to their traditional way of life, that of mobile pastoralists. Suspicious and resentful of centralized bureaucratic organizations, they gradually trickled southwards along the
Nile River and, step by step, interacted with and later
disturbed a number of Nubian Christian kingdoms. The
then city-centric and dominant view of these groups is
well captured in the following excerpt:
“On the fringes of the cultivated lands and beyond
them roamed Arab nomad tribes not yet assimilated to
a settled life. Some of these tribes coming from Arabia
Shuwa-Arab economics: livestock management and animal consumption
39
Fig. 2: Cattle herd grazing in the arbustive savanna.
Fig. 3: Prime grazing from the yaere, the hinterland depression.
inaugurating thus another process, that of the gradual
Africanization of the Sudanese Arabs” (MacMichael
1922, Cunnison 1966, Hasan 1973, Hrbek 1977: 79, Kropacek 1985). In the Chad basin, Shuwa-Arab communities were spread out between the Darfur and Lake Chad
(Denham et al. 1828, Barth 1965, Lebeuf 1969, 1987,
Lange 1987, Nachtigal 1987a, b), in Wadday, Bagirmi,
Bornu and many others local chiefdoms.
T he d at a ba se
The material analyzed in this article was collected in the
Houlouf region, officially known as the El-Birke Can-
ton, in three field seasons, in 1987, 1989, and 1991. The
study area is located along the southern margins of the
Sahel in northern Cameroon and measures 500 square
kilometers, 25 kilometers W-E and 20 N-S, at 11° 55’/12°
05’ latitude north and 14° 50’/ 15° 05’ longitude west.
The western, northern, and southern boundaries of the
study area are arbitrary. The eastern one corresponds to
the sinuous course of the Logon River (Fig. 1). The land,
generally flooded during the rainy season, is flat with
elevation ranging from 296 to 290 m above sea level.
The vegetation consists of an arbustive savanna, with
thorny trees and shrubs (Fig. 2) and extensive grasslands, prime grazing land for herders (Fig. 3). A number
of settlements, ranging across the local sites’ spectrum,
40
Augustin F.C. Holl
Fig. 4: Map of Djidat II, a permanent settlement.
were sampled in different parts of the study area. The
studied settlement sample thus consists of four permanent villages: Abuzrega, Djidat I, Djidat II (Figs. 1 and
4) and Marafin, all located on land generally above the
average flood level; 11 dry season camps arranged in
two clusters (Fig. 5) (Amachita and Agedipse) located in
the hinterland depression (Yaere); and finally, 12 semipermanent villages, distributed all over the study area
(Fig. 6). Some of the members of the research crew of
Shuwa-Arab ethnic background were very familiar with
all the aspects of life in these different kinds of settlements. Permanent and semi-permanent were constant
features of the cultural landscape. Dry season camps on
the other hand were never set by the same group on the
same spot from one year to the next. Even the pattern of
association that resulted in dry-season camps clusters
was in a constant state of flux.
Pat t e r n s of subsist e nce
In all the cases, Shuwa-Arab subsistence systems include grain agriculture and livestock husbandry. But
the balance in terms of labor investment and social values varies according to local situations and historical
circumstances. A third element, fishing, can be added
to that fundamental pair, but its position within the systems is slightly ambiguous as most of the catches are
sold in town markets as sun-dried fish. Fish is nonetheless an important component of the Shuwa-Arab diet.
Fishing
In the study area, fishing is a strictly seasonal activity
confined to the flood period (September–November)
when the hinterland depression – Yaere – is trans-
Shuwa-Arab economics: livestock management and animal consumption
41
Fig. 5: Map of Amachita V (Yaere V), a dry-season camp.
formed into an extensive water body. The productivity
of fishing expeditions that settled in camps on available
higher grounds is however low at the peak of the floodperiod because of the high dispersion of fish populations. It is higher at the onset of the flood and during the
receding phase (Blache & Miton 1962, Blache 1964,
Laure 1974). Different tools are part of the fishermen
gear; they include dug-outs, nets, basket-traps, batteries of fish hooks, as well as a range of cutting tools.
Fishing parties generally consist of young adult males.
They settled in fishing camps for several weeks in a row
during the optimal season and process their catches on
a daily basis. Two complementary techniques are used,
sun-drying on the one hand for small size fish, and
smoking for the larger specimens on the other hand.
Sun-drying on a woven reed mat does not require special installations. It is as such archaeologically invisible. Smoking is performed in especially built struc-
42
Augustin F.C. Holl
Fig. 6: Map of Ngada I, a semi-permanent settlement.
tures. The examples recorded near the Amachita camp
cluster, situated on a low archaeological mound (Holl
2002), are elongated and horse-shoe shaped. They consist of a low mud-wall 0.20 to 0.30 m in height. Wood
twigs are placed on top of the substructure, and the fish
to be smoked is laid out flat upon this constructed surface. The hollow space within the structure and underneath the platform is filled with slow burning fuel
suited for producing thick smoke, generally a combination of almost dry and moist wood, leaves and grass.
Depending on the size of the catch, quantities of both
sun-dried and smoked fish are transported to the markets of the nearby cities of Kusseri and Ndjamena, or
the entire product is stored in the camp in especially
dug clay lined pits for the whole two-month fishing season for sale at a later date.
Agriculture
All the communities from the study area are involved
one way or the other in agricultural production. However, the levels of involvement vary from one set of
groups to another, within and across ethnic affiliations.
As far as the Shuwa-Arabs are concerned, this involvement is minimal for some of the groups inhabiting
semi-permanent villages. It is much more important for
those living in permanent settlements. The documented agricultural activities are divided into two main
components. Grain agriculture devoted to the cultiva-
tion of sorghum and maize for the production of the
staple food on the one hand, and on the other, intensive
gardening geared toward the production of green vegetables (lettuces, cucumbers, and tomatoes) sold in the
markets of nearby towns.
Intensive gardening is carried out during the first half
of the dry season. It is a labor-intensive process involving the frequent watering of carefully designed
green vegetables plots. The production of green vegetables generally starts with the receding of the flood
and lasts as long as the water-table is high enough to
allow optimal access to water. The gardens are located
along the northern edge of the Yaere, approximately
one kilometer south of the settlement of Madaf. It is
predominantly a young adult male activity from which
they raise some cash. Plots are owned by individuals,
close kin or friends, and labor-cooperation seems to
be minimal. Finally, beside its cash value, the contribution of intensive gardening to the villagers’ diets is
negligible.
The bulk of agricultural production revolves around
the cultivation of sorghum and maize. Red sorghum is
preferentially cultivated by the inhabitants of semipermanent settlements. They provide some explanations for their choice: 1) red sorghum growth and maturation cycle is shorter; consequently, it can be sown
after the beginning of the receding flood and harvested
Shuwa-Arab economics: livestock management and animal consumption
just in time for the move to the dry season camping
area. 2) It does not attract birds; the amount of loss to
natural competitors is minimal. 3) It is tastier and much
more nutritious. White sorghum is preferentially grown
by the inhabitants of permanent villages. Its out-put per
acreage is higher but it is the favorite target of birds.
Different bird-scaring devices are developed to protect
the harvest. They include classic scarecrows, noisemaking devices, as well as the active involvement of
children stationed on elevated platforms with their
slingshots made of two parallel rubber bands tied to a
Y-shape handle. The proportion of maize in the local
agricultural output is still minute; Bilbede appears to
be an interesting case of shift from sorghum to maize
production if judged from the large amount of corncobs found in the site.
There are no differences in tools and field preparation
techniques used in the cultivation of either red and white
sorghum or maize. In most cases, settlements are surrounded by one or more rings of fenced cultivated plots.
This is not the case for Danguerchem, Gobrem, and
Ngumati. All these sites were abandoned at the time of
fieldwork; the absence of evidence for cultivated plots is
certainly due to poor preservation. Ndom has six fenced
fields for six households, ranging in size from 65 x 30 m
for the largest to 30 x 25 m for the smallest. The situation
is radically different for permanent settlements. All are
surrounded by several rings of cultivated fields. Mishiskwa is located on a shallow mound; the rings of fenced
fields are not attached to the rest of the village features.
They are also comparatively narrow, stretched on less
than 500 m outwards. Abuzrega, Djidat I, Djidat II, and
Marafin have more complex catchment areas stretched
on a radius of two to three kilometers. In fact, the whole
central-western part of the study area has a dense pattern
of dendritic field systems visible on low altitude aerial
photographs. The shift toward heavier reliance on agricultural production, already suggested with the relatively high frequency of storage platforms, is clearly visible
in the landscape. The remaining settlements present
more diverse situations as far as the nature, distribution,
and size of cultivated fields are concerned. The frequency of cultivated plots per site varies from 10 (Bilbede) to
2 (Bawara). There are three settlements, Bawara, Ngada
II, and Ngada I, with two to four cultivated plots. In the
first site with two fields for seven household units, the
recorded plots measure 48 x 37 m for the larger specimen, and 17 x 14 m for the smaller. At Ngada II, the
second with three fields for eight household units, two of
the cultivated plots measuring 40 x 30 m for the larger
and 15 x 12 m for the smaller are devoted to sorghum,
and the last one 25 x 17 m to maize. And finally, at Ngada I, the third site with four plots for nine household
units, two of the largest plots, measuring 22-23 x 25 m,
are devoted to the cultivation of gourds for calabashes
used as canteens and other containers and a kind of
43
herbal tea. The smaller plots measuring 20 x 18 m and
15 x 7 m are sorghum fields.
Bilbede, Gallis, Ngada III, and Ngada IV have six to ten
cultivated plots. Gallis has six fields for nine household
units, ranging in size from 58 x 40 m for the largest
specimen to 28 x 25 m for the smallest one. Both Ngada
III and Ngada IV have nine fields each for fourteen and
twelve household units, respectively. In the former, the
largest cultivated plot measures 44 x 34 m and the smallest 13 x 10 m. In the latter, they measure 37 x 25 m and
20 x 13 m, respectively. Finally, Bilbede has the highest
number of cultivated plots, ten for nine household units,
as well as a relatively large threshing area. They vary in
size from 60 x 40 m for the largest to 16 x 16 m for the
smallest. If measured by the number of documented cultivated plots, the extent to which the inhabitants of these
last settlements are engaged in agriculture appears to be
extremely variable. Because the number of cultivated
fields is systematically lower than the number of documented household units, some family groups may have
pooled their plots or, quite unlikely, relied on exchange
for their supply of agricultural products.
Pat t e r n s of spa ce allo cat ion
The multiple needs of livestock in terms of fodder, water, and protection are embedded in Shuwa-Arab settlement layout. Evidence for livestock husbandry is found
at all the tested settlements. Their nature, characteristics, and frequency vary from village to village. However, they generally include cattle corrals, sheep/goat
pens, livestock houses, tethering as well as milking
poles, and milking stations. The number and distribution of these features, in dry season camps, semi-permanent, and permanent villages, are a good indication
of the important role livestock play in the lives of site
inhabitants. The differential allocation of space to human habitation stricto-sensu versus livestock provides
interesting insights (Table 1). Livestock space takes the
largest proportion of dry season camps, varying from
684 (Amachita V) to 2,344.35 m2 (Agedipse III). The
human space on the other hand ranges from 162.19
(Agedipse II) to 656.70 (Agedipse III) m2. There are
three main situations if human space is factored to that
of livestock: 1) lower values [0 to 1] mark a strong predominance of livestock space in the site’s layout, the
lower the figure, the larger the difference. 2) The value
<1> points to a well balanced allocation with equal
share of the site’s space devoted to human habitation
and livestock management. 3) Higher values [> 1] mark
the predominance of human habitation stricto-sensu in
the site’s layout. Livestock is an essential element in the
spatial layout of all Shuwa-Arab settlements. However,
things appear to be tailored differentially as each of the
three settlements categories stands out.
44
The value recorded from dry season camps varies from
0.09 (Amachita IV) to 0.38 (Amachita V) indicating
quite clearly that these localities can safely be also
called livestock camps (Table 1). The figures for semipermanent settlements that range from 0.12 (Bawara) to
1.04 (Gallis) cover two trends. The first, indicated by
values ranging from 0.12 to 0.31, documented at Bawara,
Gobrem, Mishiskwa, and Ndom (Table 1), is characterized by a layout of features typical for an overall communal management of livestock herds. The other, with
values varying from 0.43 to 1.04, found at Bilbede, Danguerchem, Gallis, Ngada I, II, III, IV, and Ngumati,
points to a shift toward a less communal and more individualized household management strategy. This is visibly manifested by an increased reliance on fences. Agriculture and other economic activities such as raising
poultry play more important roles in these settlements.
Permanent settlements present a totally different but not
surprising situation (Table 1). A larger proportion of the
built space is allocated to human habitation but the
share of livestock is very important, ranging from
1,010.90 (Djidat I) to 4,051.11 (Djidat II) m2. The tabulated values vary from 1.13 (Djidat II) to 2.09 (Djidat I).
This trend does not point to a decreasing importance of
livestock in the settlement economies. In this case, livestock management is exclusively a household affair.
Each extended family has its own cattle enclosures,
sheep/goat pens, and livestock houses. A closer look at
the distribution of features built for livestock provides
additional insight into intra- and inter-site variations.
These features essentially belong to three categories,
relatively large size corrals for cattle, smaller pens for
sheep/goats, and houses for both cattle and sheep/goat
(Table 2). Dry season camps and semi-permanent settlements exhibit strong similarities in the frequency
distribution of cattle corrals and sheep/goat pens. The
number of corrals varies from 1 to 4 and that of sheep/
goat pens from 1 to 10. There is, however, a significant
difference in the number of livestock houses which
ranges from 1 to 5 in the dry season camps and from 2
to 24 in the semi-permanent villages. The term livestock house is somewhat ambiguous since in the dry
season camp context most of these houses are used for
horses. The relatively large number of livestock houses
in semi-permanent settlements – rainy season villages –
is partly correlated to the local tse-tse fly infestation that
generally peaks during the rainy season and the ensuing
annual flood. During that time of the year, cattle herds
are kept and fed indoors throughout the day and taken
out for watering at nightfall. The unusually small number of livestock houses (2) at Danguerchem is to some
extent due to the poor preservation of the settlement
record. It is very likely that there were many more livestock houses than detected during the mapping of the
site (Table 2). As far as permanent settlements are concerned, the frequency distribution of livestock features
alone is misleading. The number of cattle corrals varies
Augustin F.C. Holl
from 2 to 5 but they are generally smaller in size and
each belongs to a specific household. The number of
sheep/goat pens, which ranges from 4 to 10, is surprisingly low if compared to the settlements’ population
size. Finally, the number of livestock houses mirrors
that of semi-permanent settlements for the same reasons. Chicken coops and straw storage platforms complement the range of features used for the management
of livestock. A few chicken coops are found in dry season camps (Table 2), but they are particularly concentrated at two sites, Bilbede and Ngada IV. The implication of this pattern will be discussed later as it represents
a trend toward specialization and intensification.
In summary and from the perspective of settlements’ constructed features, it is not an exaggeration to suggest that
livestock needs are the crucial element placing constraints
on the size and shape of Shuwa-Arab settlements.
A s s e s si ng l ive st o ck u se
A more straightforward archaeological approach to the
analysis of subsistence patterns is adopted in this portion of the paper. Faunal samples have always played an
important if not ambiguous role in the identification of
past socio-economic systems (Cribb 1991, Sadr 1991,
Smith 1992).
“Traditionally, sites with a predominance of domestic
animals in the faunal remains have been attributed to pastoralists. In some instances, researchers have tried to go
further, to infer milk, meat, or wool production from the
age and sex profile of the domestic animal remains. …
But faunal samples, so easily distorted by taphonomic
processes, are rarely representative enough to allow such
studies. … Furthermore, most interpretations of animal
mortality curves are based on a tentative model. The validity of which has been challenged” (Sadr 1991: 13).
Sadr’s critique of faunal remains analyses is important
and hits the nail on its head. The vagaries of taphonomy
and the hazards of preservation and recovery make an
accurate assessment of livestock management systems
extremely difficult despite the optimistic promises of
age/sex curves. The state of preservation of the archaeological record at hand and the degree of resolution of
the determination of the collected remains can nonetheless lead to interesting interpretations of past management of animal resources. As sound as it is, Sadr’s
critique is unfortunately not backed by any useful suggestion on how to extract good information from archaeological faunal samples, and one is tempted to ask,
what to do? As will be shown in this analysis, even an
ethnoarchaeological study of single occupation sites
like the Yaere dry season camps does not provide an
easy and ready to generalize answer.
Shuwa-Arab economics: livestock management and animal consumption
Site
Size
(ha)
Number
of features
Livestock features
E
H
Total
45
Livestock
space
(m2)
Human
pace
(m2)
Human/
Livestock
Dry-season camps
Amachita I
Amachita II
Amachita III
Amachita IV
Amachita V
Amachita VI
Amachita VII
0.38
0.25
0.28
0.40
0.38
0.38
0.38
23
29
27
19
25
15
21
5
6
7
2
5
4
4
2
1
3
1
1
7
7
10
2
5
5
5
1,615.39
1,183.80
1,155.36
2,000.00
684.00
930.23
952.29
268.50
326.00
227.00
198.04
265.00
187.77
251.71
0.16
0.29
0.19
0.09
0.38
0.20
0.26
Agedipse I
Agedipse II
Agedipse III
Agedipse IV
0.86
0.38
1.76
0.44
38
15
44
23
6
4
14
4
5
2
1
-
11
6
15
4
2.132.63
934.60
2,344.35
859.27
556.29
162.19
656.70
187.94
0.26
0.17
0.28
0.21
Semi-permanent settlements
Bawara
0.78
Bilbede
0.67
Danguerchem
0.54
Gallis
3.20
Gobrem
1.20
Mishiskwa
1.05
Ndom
0.63
Ngada I
3.36
Ngada II
0.60
Ngada III
2.00
Ngada IV
2.03
Ngumati
0.59
27
74
21
92
35
55
36
63
59
67
107
30
5
6
3
7
4
4
5
2
2
1
2
5
8
11
2
13
10
24
10
16
12
12
17
4
13
17
5
20
14
28
15
18
14
13
19
9
2,071.68
1,626.91
1,276.75
1,099.60
2,826.13
2,885.94
2,840.28
2,202.86
1,611.33
2,438.56
1,433.58
892.38
265.49
706.40
862.84
1,150.63
878.00
905.73
381.99
964.03
1,167.40
1,300.40
1,077.61
634.64
0.12
0.43
0.67
1.04
0.31
0.31
0.13
0.43
0.72
0.53
0.75
0.71
Permanent settlements
Abuzrega
4.15
Djidat I
1.76
Djidat II
4.52
Marafine
2.01
105
92
190
74
6
11
15
9
13
18
35
21
18
29
50
28
1,415.11
1,010.90
4,051.11
1,375.46
2,411.47
2,118.61
4,595.56
2,091.88
1.70
2.09
1.13
1.52
Table 1: Patterns of space allocation
Sa mpl i ng a nd on- sit e f ield
met ho dolog y
Animal bones were not collected from all four permanent villages because of their overall, extremely low
density and the high intensity of trampling. The faunal
samples considered in this discussion were systematically collected from the semi-permanent sites and dry
season camps within the confines of each of the investigated settlements. The sampling space was delimited
by the outer perimeter of houses. Visible animal bones
were handpicked by a crew of fifteen workers, walking
in parallel lines with each covering a two meter wide
transect. It was thus possible to collect the maximum
of the accessible faunal material. The sampling techniques used in this case favour large bones (Fig. 7). Not
surprisingly, and despite the well-known high consumption of fish among Shuwa-Arab from the study
area, fish bones are literally absent from the collected
46
Settlement
Augustin F.C. Holl
Cattle
Corrals
Sheep/Goat
pens
Livestock
houses
Others
Total
Dry-season camps
Amachita I
Amachita II
Amachita III
Amachita IV
Amachita V
Amachita VI
Amachita VII
1
1
1
2
2
1
1
4
5
6
2
3
3
1
3
2
1
1
1
-
5
7
10
2
7
5
5
Agedipse I
Agedipse II
Agedipse III
Agedipse IV
1
1
4
1
5
3
10
3
5
2
1
-
1
1
11
7
15
5
Semi-permanent settlements
Bawara
Bilbede
Danguerchem
Gallis
Gobrem
Mishiskwa
Ndom
Ngada I
Ngada II
Ngada III
Ngada IV
Ngumati
1
1
3
2
4
2
1
2
1
1
2
3
4
5
5
3
2
4
4
1
2
8
2
8
11
2
13
10
24
10
16
12
12
17
4
29
45
-
13
45
5
20
17
28
15
22
14
15
72
9
Permanent settlements
Abuzrega
Djidat I
Djidat II
Marafine
2
2
5
3
4
9
10
6
13
18
35
21
-
19
29
50
30
Table 2: Distribution of livestock features
samples. All bones scattered over the sites’ surface or
in their immediate periphery are not necessarily refuse
from human meat consumption. Some result from the
scattering out of skeletal elements of dead animals.
Other are dispersed and/or accumulated by dogs, jackals, and other scavengers. Two dense bones scatters
were found and mapped. One was discovered at few
hundred meters east of Gobrem, on the right bank of
the Abani, a seasonal stream, and the other less than
one hundred meters south of Ngada IV.
The bones collected from the studied sites had to be identified and recorded directly in the field, a task carried out
by Dr. Anne Bridault and myself (Holl et al. 1991). For
understandable reasons, these bones were not intended to
be brought back to the university laboratory. The material collected from each site was arranged directly on the
ground into anatomic elements, species, and size. This
tactic allowed for an immediate visual assessment of the
overall distribution and frequency of anatomic elements,
as well as the range of species represented (Fig. 7).
Shuwa-Arab economics: livestock management and animal consumption
47
Fig. 7: Sample of animal bones collected from Ngada I
T he b one s c at t e r s
Ngada IV bone scatter
The bones’ scatter from Ngada IV consists of the carcass of a calf that died two years prior to the 1989 fieldseason. The cause of death was not made explicit to the
field crew. The bulk of the calf remains, also comprising most of the bones, are contained within a dry and
parched skin. With the exception of three goat scapulas, all the remaining 128 mapped bones resulted from
the dismemberment and dispersal of the calf’s skeleton
by dogs. Gnaw marks were found on a left patella.
Limbs bones, skull bones, cervical as well as portions
of lumbar vertebras were found widely scattered around
the ribcage which was still preserved in the dried skin.
The set of processes leading to the formation of the
bone scatter described is clear. Initially, a dead calf was
abandoned at some distance from the village, although
it is not known if the young animal died in the village
with the carcass being transported later to an acceptable distance from the settlement, or if this happened at
the location where it was found. This was followed by
scavenging domestic dogs, rarely fed in the African
countryside, which took the opportunity to feed on the
available meat. This functioned to scatter most of the
limb bones away from the axial skeleton, and at the
same time bringing in few sheep-goat bones. The bone
scatter is comprised exclusively of domestic animals
remains; it does not, however, provide any crucial information on the settlement-subsistence status of the
village inhabitants. Alone, it does not help discriminate
between a dry season camp, a semi-permanent village,
and a permanent settlement.
Gobrem bone scatter
The Gobrem bone scatters are in fact located on the
“ghost” of a completely eroded twin settlement that was
situated on the right bank of the Abani seasonal stream.
It is spread over 395.20 m2 and divided into four subcomponents termed bone scatter (BS) I to IV. All the 237
recorded bones belong to cattle, with the density varying
from 0.75/ m2 to 0.38/ m2. BS-I is found at the northwestern end of the grid sample. It contains 76 bones, 28 from
the appendicular and 48 from the axial skeleton. Jaws
and cranium bones are rare with two cases of left and
right mandibles, the latter with a fully erupted M3. Two
individuals are represented in the studied sample, a mature and a juvenile animal. All anatomical elements,
from head to hooves are represented, suggesting that
both animals were probably discarded, with the carcasses later being fed upon by domestic dogs and scavengers
like golden jackals. 15 bones out of a total of 76 present
evidence of gnawing by carnivores. Interestingly
enough, all the bones with gnaw marks belong to the
axial skeleton: cervical, thoracic, and lumbar vertebras,
suggesting late access to the carcass. This is generally
the case when an animal dies during the flood season in
a river bed, with the carcass accessible later after the
receding of water. BS-II is located at few meters southeast of BS-I. 94 cattle bones were recorded, 49 from the
48
axial skeleton, and 45 from the limbs. Gnaw marks are
documented on 30 out of 94 bones. Most of these marks
are found on vertebras, mandibles and cranial bones.
However, 5 limbs bones, scapula, pelvis, tibia, and tarsal bones were found with tooth marks. BS-III is located thirteen meters southeast of BS-II. The recording
grid measures 10 m in length and 8 m in width. 31 cattle bones have been mapped. 22 belong to the axial and
9 to the appendicular skeleton. The sample represents
a single individual with gnaw marks found on five
bones, the proximal end of the left and right femurs,
both the left and right side of the pelvis, and lateral portions of the sacrum. The pattern of tooth marks suggests the scavenger fed on the upper part of the rear
limbs. Finally, the remaining sub-component BS-IV is
spread over 91.20 m2 east of the three previous ones. It
measures 15.20 m in length and 6 m in width and includes 36 cattle bones, 20 from the axial skeleton and
12 from the limbs. The scatter seems to be comprised
of the skeleton of two individuals. In order to clarify
that issue, the sampled area was divided into four equal
units, quadrant I to IV (Q-I to Q-IV). Ten bones were
recorded from Q-I; they include the sacrum, which is
gnawed on its edge, two left scapulas both with gnaw
marks, one proximal and one distal femur, a left humerus, one specimen each of cervical, thoracic, and
lumbar vertebras, and finally, a right pelvic fragment.
The subset consists of elements of the upper limbs and
backbone; with gnaw marks found on seven out of ten
bones. Ten bones have also been recorded in Q-II; there
are four vertebras, one cervical and three thoracic; two
proximal ends of ribs, a left proximal femur, a right
proximal humerus, and two tarsals. Five of the bones
have gnaw marks. Q-III has seven bones, this time with
a complete cranium of a young mature ox with well
preserved horn-cores. Interestingly, the specimen has
extensive gnaw marks, and the dried feces of a jackal
were found on the cranium. The rest of the sample includes a left ulna, a right proximal humerus, a left distal femur, a cervical vertebra, a rib fragment, and a
tarsal bone. Gnaw marks are rare on the rest of the
sample, suggesting the scavenger focused on the cranium. Finally, Q-IV, containing nine bones, mostly includes elements of the axial skeleton. There are four
vertebras, three cervical and one lumbar, two mandibles, left and right, one left proximal femur, one scapula blade, and a tarsal bone. Gnaw marks are found on
two bones, the scapula blade and the lumbar vertebra.
The Gobrem bone scatters appear to have resulted from
the dismemberment of dead animals by scavengers.
In both Ngada IV and Gobrem cases, there is no human
contribution to the scatter of domestic animals bones
described above. The strong predominance of cattle
can easily lead any archaeologist to rightly infer the
existence of a livestock husbandry. But the material at
hand does not allow further clarification.
Augustin F.C. Holl
Fau n a l r e m a i n s f rom s e m ip e r m a ne nt set tle me nt s
Faunal remains were collected in semi-permanent settlements and dry season camps. The relatively higher
density of occupation and the ensuing higher degree of
trampling precluded any significant bone sampling
from the permanent villages. The total amount of bones
collected varies considerably from one semi-permanent settlement to the other, ranging from 7 (Danguerchem) to 795 (Gallis) (Table 3). The proportion of
undetermined splinters is high in general, varying
from 4 (Danguerchem) to 275 (Ngada III). The recorded faunal remains are distributed into two major categories, domestic and wild animals (Table 3). Three wild
animals species, gazelle, elephant, and hare, are represented in small proportions at Bawara, Gallis and
Ndom (Table 3). Despite the well known importance of
fish in Shuwa-Arab subsistence and economy (Holl
2003), they are practically absent in the material record
with the exception of a single fish bone found at Ngada
I. Domestic animals are partitioned into two sub-categories. One includes all animals used for direct subsistence purposes including cattle, sheep/goat, poultry,
and other bovidae. The other is made of non-food animals like horse, donkey, and dogs, the bones of which
were probably spread all over the sites by dogs and
other scavengers. Horses are prestigious riding animals
owned by few individuals. In this sub-category, the
relatively large predominance of donkey bones is an
accurate reflection of the importance this animal species has in Shuwa-Arab life. Donkeys are used to carry
people and a broad range of loads in a multiplicity of
situations. In all the recorded cases, cattle bones are
predominant, conveying an accurate picture the Shuwa-Arab animal species hierarchy (Table 3). Sheep and
goats are always in second position in all the tested
settlements. A number of settlements present particularly interesting developments. The bone sample from
Bilbede, with a relatively important proportion of
chicken skeletal remains, confirms the impression
gained from the observed high frequency of chicken
coops (Table 3, Fig. 8). Chicken bones amount to 18
remains representing 3% of the sample. Cattle is largely predominant with 174 pieces (29.2%), followed by
Bovidae sp. with 84 (14.1%), and sheep/goat with 55
(9.2%). Horse and donkey are represented by 1 (0.2%)
and 4 (0.6%) bones, and unspecified equids by 14
(2.3%). The Ngada site cluster presents the broadest
range of variation in terms of represented taxa (Table
3, Fig. 9). The recorded faunal spectra range from six
(Ngada III and Ngada IV) to eight taxa (Ngada I and
Ngada II). Despite the persistent predominance of cattle with proportions varying from 40.6% to 27.2%, each
of the settlements has a specific combination of species.
In addition to the 150 (27.2%) cattle remains, the Ngada
I sample includes 76 (13.2%) Bovidae sp., 61 (11%)
Shuwa-Arab economics: livestock management and animal consumption
Fig. 8: The faunal spectrum from Bilbede. Key: 1: Bos taurus; 2: Ovis/Capra sp.; 3: Bovidae sp.;
4: Equus caballus; 5: Equus asinus; 6: Equus sp.; 7: Canis familiaris; 8: Gallus gallus; 9: Gazella sp.;
10: Loxodonta africana; 11: Lepus capensis; 12: Fish; 13: Undetermined.
49
50
Augustin F.C. Holl
Fig. 9: The faunal spectrum from Ngada I, II, III, and IV. Key: 1: Bos taurus; 2: Ovis/Capra sp.;
3: Bovidae sp.; 4: Equus caballus; 5: Equus asinus; 6: Equus sp.; 7: Canis familiaris; 8: Gallus
gallus; 9: Gazella sp.; 10: Loxodonta africana; 11: Lepus capensis; 12: Fish; 13: Undetermined.
sheep/goat, 15 (2.7 %) Equus sp., 3 (0.5%) donkey,
2 (0.3%) chicken, and finally, 1 (0.2%) each for horse
and fish (Table 3). Ngada II has the largest proportion
of chicken bones with 20 (5.6%) remains as well as
those from domestic dogs. The rest of the sample consists of cattle (98, 27.4%), Bovidae sp, (54, 15.1%),
sheep/goat (29, 8.1%), donkey (9, 2.5%), and finally,
horse and Equus sp. (7, 1.9% each). In addition to the
standard domestic animals, which is strongly dominated by cattle with 210 (34.6%) skeletal remains, the
Ngada III sample includes only a few chicken bones
(5, 0.8%). Finally, Ngada IV has a handful of dog remains, with 3 (0.6%) bones. The rest of the sample with
cattle largely predominant (185, 40.6%) is divided into
sheep/goat, Bovidae sp., donkey, and Equus sp., with
the proportion varying from 10.1% to 0.6%.
Fau n a l r e m a i n s f rom t he
d r y sea son ca mps
The dry season camps sampled in this study were inhabited for two to three months – March – May/June in 1990. An important precipitation deficit in the 1990
rainy season allowed for an excellent preservation of
these camps. They were not flooded that year as is generally the case. At first glance, the faunal remains collected from the sampled dry season camps provide a
desperately monotonous picture of the represented faunal spectra. They are nonetheless the most accurate
indication of pastoralism. However, a closer look at the
evidence reveals interesting taphonomic differences
and patterns of meat consumption during that part of
the year. In general, the samples are small with frequency varying from a maximum of 65 at Amachita VI
1
-
-
Wild animals
Gazella sp.
Loxodonta africana
Lepus capensis
Fish
246
596
-
Table 3: Faunal remains from semi-permanent villages
27
87
18
-
-
Gallus gallus
Canis familiaris
Undetermined
TOTAL
1
4
14
9
4
Equus caballus
Equus asinus
Equus sp.
-
174
55
84
Bilbede
24
7
15
Bawara
Domestic animals
Bos taurus
Ovis/Capra sp.
Bovidae sp.
Site
Taxa
4
7
-
-
-
-
3
-
Danguerchem
259
795
-
3
3
3
-
7
15
307
88
110
Gallis
14
231
-
-
-
4
-
176
24
13
Gobrem
54
206
-
1
1
-
-
12
6
65
47
20
Ndom
243
552
1
-
2
-
1
3
15
150
61
76
Ngada I
129
357
-
-
20
4
7
9
7
98
29
54
Ngada II
275
606
-
-
5
-
15
14
210
43
44
Ngada III
171
445
-
-
3
3
3
185
46
44
Ngada IV
8
54
-
-
-
4
1
5
24
4
8
Ngumati
Shuwa-Arab economics: livestock management and animal consumption
51
52
Augustin F.C. Holl
Taxa
Site
Bos taurus
Ovis/Capra sp.
Limicolaria sp.
Undetermined
Total
Amachita I
Amachita II
Amachita III*
Amachita IV
Amachita V
Amachita VI
Amachita VII
3
27
24
37
10
62
15
55
7
11
14
3
2
2
2
6
2
3
2
20
8
-
58
56
46
57
15
65
19
Agedipse camps cluster
Agedipse I
Agedipse II
Agedipse III
Agedipse IV
19
60
43
7
4
9
13
2
1
-
23
69
56
10
Amachita camps cluster
Table 4: Faunal remains from the dry-season camps
to a minimum of 10 at Agedipse IV (Table 4). Two species, cattle and sheep/goat are represented, with a few
specimens of land snail shells. The Limicolaria sp.
shells recorded are not food refuse. They result from
the natural death of land snails that fed on the camp’s
refuse.
The short term occupation and the small size of all the
collected samples offer the opportunity to address interesting issues revolving around the management of
animal products. The represented body parts and relative age of the animals can be used to tackle this issue.
A m a ch it a c a mps clu st e r
In the Amachita camp cluster, the amount of collected
faunal remains varies from 65 (Amachita VI) to 15
(Amachita V). Land snail shells are more frequent
ranging from two to six specimens. A single donkey
bone was found at Amachita III.
Amachita I is a four household camp inhabited by people from Djidat. The faunal material includes 3 cattle
and 55 sheep/goat bones (Table 4). All cattle bones belong to an adult animal represented by two mesial humerus fragments and a scapula. It is clearly part of a
share of meat chopped from the animal’s upper front
limb and brought to the dry season camp. The rest of
the sample consists of goat bones. They are partitioned
into 3 cranium fragments with horn-cores, 6 jawbone
fragments, 11 ribs, and a set of bones representing the
lower limbs. They include two distal ends of tibiae, one
belonging to a younger animal, three tarsal bones, and
an ulna fragment. And finally, 29 chopped bones. With
the exception of one tibia fragment, all the goat bones
belong to adult animals. The striking absence of vertebras and upper limb bones indicates that these animals
were not butchered in-situ during the occupation of the
camps, it rather suggests that sun-dried goat meat was
brought to the camp.
Amachita II, with six household units, was settled by
the families from Magourde, a village located 8-10 kilometers NE of the Yaere. The animal bone sample
consists of 56 remains, 27 from cattle, 7 from goat, and
2 land snail shells, as well as 20 undetermined pieces
(Table 4). The goat bones recorded include one cranium, four jaws, a left pelvic fragment, and a distal tibia
with the astragal still attached. A large proportion of
the cattle bones recorded, 20 out of 27, belong to calves.
They are comprised of 4 crania, three of them hornless, 4 right and 3 left mandibles, 2 pelvic fragments,
1 right scapula, 1 right and 2 left humeri, 1 right radioulna, 1 right tibia, and 1 left metacarpal. The presence
of a sacrum fragment with an attached left pelvis half
suggests that at least one of the calf carcasses was not
butchered. The bones of this individual were scattered
after death and abandonment. Adult cattle are represented by two radio-ulnae (right and left), two metacarpals (right and left), and finally, a complete right hindlimb with the femur, tibia, calcaneum, and astragal, in
Shuwa-Arab economics: livestock management and animal consumption
anatomic connection. This set of bones also suggests
the adult animal skeleton was scattered after death and
abandonment. Paradoxically, the cattle bones from
Amachita II do not seem to stem from meat consumption, and are more likely from the post-abandonment
scattering of dead animal bones.
Amachita III is an eight household unit camp settled by
people from Chouaram. The faunal sample includes eight
undetermined fragments, 2 Limicolaria sp. shells, a donkey tertiary phalanx, as well as 24 cattle and 11 sheep/goat
bones (Table 4). Seven of the cattle bones are from calves.
They are distributed into: 1 cranium, 1 cranial fragment,
1 jawbone, 1 jawbone fragment, 1 radius, 1 right tibia, and
1 metacarpal. Adult cattle bones are comprised of: 1 right
horn-core, 2 molar teeth, 1 cervical vertebra, and 2 pelvic
fragments, 1 proximal femur, 1 right radio-ulna, 2 distal
radio-ulna, 1 right tibia, 4 carpals, and finally, 1 astragal
and 1 tarsal. The uncovered goat remains belong to an
adult animal that appears to have been butchered. They
consist of 1 cranium and 1 cranial fragment, 1 molar, 1 rib
fragment, 1 sacrum fragment, 2 humeri (left and right),
2 radio-ulnae (left and right), 1 mandible, and 1 proximal
metacarpal. The faunal assemblage thus seems to be comprised of the remains of three animals, one calf, one adult
ox, and one goat.
Amachita IV, a nine household unit camp, was settled
by families from Bam-Kala. The collected faunal material consists of 57 pieces made up of 6 Limicolaria sp.
shells, 14 sheep/goat bones and 37 cattle bones (Table
4). The goat bones belong to a single adult animal and
consist of 4 ribs, 1 complete and 2 mandible fragments,
1 cranium, and 6 fragments. Cattle bones belong predominantly to calves with 29 pieces out of a total of 37.
Adult cattle are represented by lower limbs bones, 1
proximal ulna, and 7 phalanxes, 5 primary, 1 secondary, and 1 tertiary. Calves are represented by three
complete skulls, 1 occipital bone, 1 complete vertebral
column from the atlas to the sacrum, 1 ear bone, 2 mandibles (left and right), 4 jaw fragments, and 1 molar for
the axial skeleton. The limb bones include 1 complete
and 5 mesial portions of humeri, 1 left femur, 1 radioulna, 1 left tibia, 2 complete and 2 proximal ends of
metacarpals, and 3 metatarsals. The presence of a complete vertebral column strongly suggests that the bones
of at least one of the calves were scattered after it died.
The adult cattle bones as well as those from the goat are
more likely to represent food refuse.
Amachita V is a seven-household unit camp. The
amount of animal bone remains collected from the site
is surprisingly small, with a total of 15 pieces divided
into 2 land snail shells, 3 sheep/goat and 10 cattle bones
(Table 4). Sheep/goat bones are limited to 2 ribs and one
femur from an adult animal. Two individuals, one adult
and a calf, are represented among the cattle remains.
53
One skull fragment, 1 molar, and 1 second phalanx belong to a calf, with the adult’s bones consisting of 3
fragments and 1 distal end from a humerus, and 2 proximal and 1 distal ends of radius. The eclectic representation of anatomic elements seems to suggest that the
Amachita V faunal remains probably resulted from
food consumption.
Amachita VI was settled by eight families from the
semi-permanent village of Alaya I. The faunal sample
collected from the camp is comprised of 65 pieces, 3
land snail shells and 62 cattle bones belonging to calves
and adult animals (Table 4). Calves are represented by
2 skulls, 4 occipital fragments, 1 horn, 2 jaw fragments,
1 atlas, 1 sacrum, 1 complete pelvis, 2 right scapulas, 1
rib fragment, 1 right humerus, 2 femurs (right and left),
1 proximal femur, 1 radio-ulna, 1 left tibia, 1 metacarpal, 1 metatarsal, and 1 calcaneum. Adult cattle are represented by 3 mandibles (2 right and 1 left), 1 jaw fragment, 5 lumbar vertebras, 1 pelvis, 1 right femur, 2
distal humeri (left and right), 1 right radio-ulna, 2 tibiae
(1 complete right and 1 distal left), 2 right metacarpal,
1 metatarsal, 2 astragal, and 2 tarsals. The bone sample
seems to include the remains of at least four animals,
two calves and two adult individuals. The absence of
any anatomic connection between the recorded pieces,
and the breakage of some marrow-rich long bones suggest the animals were butchered at the site.
Amachita VII, the remaining site from the camp cluster
was settled by seven families from Mahanna, a village
located some 20 km northeast of the Yaere, the hinterland depression. The bone sample is small in size,
amounting to 19 pieces and consists of two land snail
shells, two sheep/goat and 15 cattle bones (Table 4).
Sheep/goat is represented by 2 lumbar vertebras of a
mature animal. The collected cattle bones are those of
an adult individual. They are comprised of 1 left mandible, 1 sacrum fragment, 2 scapula proximal ends, 8
radio-ulna fragments, 1 right distal radius, 1 right proximal ulna, and 1 proximal tibia. The sample seems to
have resulted from the consumption of small shares of
meat that may have been brought from the main village.
T he Age d ips e c a mp clu st e r
Four faunal samples were collected from the Agedipse
camp cluster. They range in size from a minimum of 10
pieces (Agedipse IV) to a maximum of 69 (Agedipse II).
The frequency of land snails, with a single specimen
found at Agedipse IV, is negligible when compared to
the Amachita cluster (Table 4). Cattle bones are predominant in all cases, with the frequency varying from
7 (Agedipse IV) to 60 (Agedipse II). The proportion of
sheep/goat bones is surprisingly low and varies from 2
recorded at Agedipse IV to 13 found at Agedipse III.
54
Agedipse I was settled by eight household units from
the village of Kulkule. The collected faunal remains
amount to 23 pieces divided into 19 cattle and 4 sheep/
goat bones (Table 4). The sheep/goat bones belong to an
adult individual. They are represented by 1 mandibular
fragment, 1 right radio-ulna, 1 right scapula, and 1 left
humerus. The cattle bones recorded belong to three
separate individuals, a calf, a young animal, and an
adult. The calf is represented by 2 mandibles (left and
right), 1 complete right humerus, and 1 mesial humerus.
The young animal is represented by 1 cranium, 1 mandibular fragment, 1 atlas fragment, 2 scapulas (right and
left), 2 humeri (left and right), 1 right femur, 1 right
tibia, 2 radio/ulna (left and right), and 1 left metatarsal.
The adult individual is represented by 2 left calcaneum
fragments and 1 second phalanx. The axial skeleton is
poorly represented with only five elements recovered.
The skewed distribution of anatomical parts, with
limbs bones largely predominant, may have resulted
from the transfer of meat shares brought from outside
to the site. Alternatively, it can be suggested that calves
or very young animals in a poor health during the peak
of the dry season may have been butchered and shared
with the inhabitants of nearby settlements.
Agedipse II was settled by five family groups originating from Magourde, a mixed farming community. The
faunal material collected amounts to 69 pieces and consists of 60 cattle and 9 sheep/goat bones (Table 4).
Sheep/goat are represented by 2 skulls, 2 horns, 2 mandibles (left and right), 2 vertebras, and 1 astragal, all
from adult animals. The cattle bones belong to at least
four individuals, one calf and three adults. The calf is
represented by 1 skull, 1 horn, 1 proximal tibia, 1 metacarpal, and 1 metatarsal. Adult animals are represented
by 3 skulls, 2 mandibles, 5 mandibular fragments, 17
vertebras, 1 sacrum, 1 complete and 4 scapula fragments, 5 pelvis fragments, 5 rib fragments, 2 humerus
(left and right), 2 femurs (left and right), 1 right radioulna, 5 metacarpals, 4 metatarsals, and 3 phalanxes.
An interpretation of the Agedipse II faunal sample is
particularly delicate. It is the smallest camp of the
whole cluster with the largest sample of animal bones.
All the collected bones were found without anatomic
connection; a feature that generally indicates scattered
dead animal carcasses. The material at hand thus seems
to have resulted from deliberate food consumption and
refuse disposal.
Agedipse III, the largest camp of the cluster was settled
by sixteen households from the village of Hinduk. The
faunal sample is comprised of 56 animal bones, 13 from
sheep/goat and 43 from cattle (Table 4). The sheep/goat
bones belong to two adult animals, as suggested by the
two recorded skulls, one of a ewe and the other of a male
goat. The remaining pieces include, 2 jaw fragments, 1
left pelvic fragment, 2 metacarpals (right and left), and
Augustin F.C. Holl
6 phalanxes. The recorded cattle bones are those of at
least three individuals, one young and two adults. The
young animal is represented by 1 scapula and 1 complete metacarpal. The adult bones consist of 2 scapulas,
one complete and the other fragmented, 4 femurs (2 left
and right), 4 humeri (2 left and 2 right), 4 tibiae (2 left
and 2 right), 4 radio-ulnas (2 right and 2 left), 2 metacarpals, 5 astragals, 5 calcanei, 5 tarsals, and 6 carpals.
There is an interesting difference between the patterning of sheep/goat and that of cattle bones. Elements
from the axial skeleton are absent from the sample of
the latter and more or less well represented in the assemblage of the former. The ewe and the male goat appear to have been butchered in the dry season camp and
probably shared. Cattle meat appears to have been
brought to the camps as limbs pieces, as if especially
processed to be carried away.
Agedipse IV, the remaining camp of the cluster, was
settled by five family groups originating from Tumam.
The sample of faunal material is the smallest of all the
documented cases with only 10 pieces, 1 land snail
shell, 2 sheep/goat and 7 cattle bones (Table 4). All the
collected bones are part of the axial skeleton, 2 jaw
fragments from the sheep/goat represented, and 1 skull,
1 skull fragment, 1 jaw fragment, and 3 vertebral fragments from the calf. The interpretation of the faunal
assemblage is particularly difficult. No gnaw marks
were observed on the studied bones, but the selective
representation of head bones suggests the activity of
scavengers after the abandonment of the camp.
The faunal remains from the dry-season camps result
from two distinct processes, human consumption of
livestock meat and the dispersal of skeletal elements
after the natural death of the animals. Meat consumption appears to have operated along two overlapping
tactics; in one, the consumed meat came from animals
butchered at the site. In the other, shares of meat, either
sun dried, smoked, or not, were brought from the main
village to the dry-season camps.
Pat t e r n s of i nt e n si f icat ion
The concept of intensification is used here in a strictly
relational perspective. It aims to single out any significant difference in output well beyond local subsistence
needs. Production intensification seems to be driven by
the combination of a few independent variables: 1) the
constant pressure exerted on all adult males aged eighteen years and above to raise cash to pay for the annual
per capita tax; 2) the proximity of large markets from
the nearby towns of Kusseri in Cameroon and
N’Djamena in Chad Republic; and 3) for the younger
segment of the population, the appeal of and desire to
access certain sides of mass-consumption.
Shuwa-Arab economics: livestock management and animal consumption
55
Fig. 10: View of a goat-milking station from Djidat I.
There are few indications of production intensification
in two components of the production-subsistence continuum. One involves poultry and the other goats. The
first one is documented in the settlements layout by the
presence of sets of chicken coops. They are found in a
number of permanent and semi-permanent settlements.
In permanent villages, the frequency of chicken coops
is confined within a narrow range of 5-9. In fact, localities such as Abuzrega and Djidat II that are devoid
of milking installations have the highest number of
chicken coops, 8 and 9 respectively. In the semi-permanent settlements, chicken coops are recorded in four
villages out of seven with frequency varying from 1 to
27. Two of the villages, Bilbede and Ngada IV, with 22
and 27 chicken coops, respectively, stand out as poultry
production centers.
The second component of the production-subsistence
continuum subject to intensification revolves around
the production of goat milk. It is indicated in site layout
by the presence of milking stations. Dairy products are
an essential component of the herders’ diets and each
family owning livestock, cattle, or sheep/goat, or both,
collect a certain amount of milk for household consumption during the adequate period of the year. This
household milking is carried out without special installations. The documented milking-stations, exclusively
for goats, are special purpose features consisting of a
series of equidistant wooden poles set in pens, round
huts, courtyards, and even along walls (Fig. 10). They
are used to streamline the milking process. As far as
the social division of labor is concerned, milking operations, transportation, as well as market transactions
are all carried out by females. Milking-stations have
been recorded at two permanent villages, Djidat I and
Marafin, occurring 10 and 5 times, respectively. The
former village is clearly specialized in animal products, with seven chicken coops and ten milking-stations
for eleven households. The latter has five milking-stations and five chicken coops for ten households.
Cr af t s a nd special i zed occupat ion s
A small array of crafts and occupational specializations are represented in the material record from the
tested settlements. For example, the Imam (Marabout),
the Moslem scholar in charge of the Friday Mosque and
the “boarding” Koranic school of Djidat II as well as
the operator of the engine-pulled mill also located at
Djidat II. He is not a resident of that settlement and has
to commute from Houlouf every day. Blacksmiths installations or products have been found at two settlements, at Djidat II and Amachita V. The former consists of a blacksmith workshop equipped with a furnace,
bellows, blow-pipes, as well as a supply of scrap metal.
In the latter, two freshly made pieces of equipment, an
iron axe and a socketed adze, were lost in the dry season camp. A shoe-polisher’s box was found in house 35
at Ngada IV suggesting that the house owner may have
56
Augustin F.C. Holl
been involved in a part-time work in the nearby towns,
Kusseri in Cameroon or Ndjamena, the capital of the
Chad Republic, on the opposite river bank. And finally,
a reed-mat weaving workshop was recorded in house
26 at Ngada IV. There is a high demand for such mats
in villages and cities during the dry seasons, when they
are used to build open shelters in courtyards.
Conclu sion
Subsistence activities and a certain but narrow range of
more specialized crafts are carried out in many of the
settlements investigated in this project. In terms of diversity, intensity, and frequency according to settlement categories, permanent villages easily rank first,
followed by semi-permanent villages, and finally, dry
season camps. With a single exception represented by
iron artifacts lost by a blacksmith at Amachita VI, livestock husbandry is the key activity performed in all dry
season camps. The analysis of animal bones brings to
light interesting results. Sophisticated quantitative
analyses of faunal remains can be applied to a very
narrow range of settlements, namely short term single
occupation sites as is the case for the tested dry season
camps that were settled between March and May/June
1990. In this case only, it is possible to address issues
concerning the number of animals represented in the
sample (MNI), the body parts brought to the site, as
well as patterns of meat consumption, with an accurate
control of the evidence. In all other cases, from the
semi-permanent as well as permanent villages, the palimpsestic nature of the faunal assemblages precludes
any reasonable assessment of such variables as the
number of butchered animals. At best, the proportional
analysis of different parameters from faunal remains
based on the crudest yardstick, the number of rests
(NR) can be performed. Finally, the distributional pattern of species in each of the investigated faunal assemblages alone is not discriminating enough to allow
for a subtle differentiation between distinct degrees of
reliance on livestock products.
Ack nowle dge me nt s
The research reported in these pages was funded by
National Geographic Society Grant # 6378-98 and
3715-87. I wish to thank Mahamat Bahr Maaruf, Sultan
of Logone-Birni for his understanding, kindness and
support in the field. I am grateful to the people of Houlouf for their hospitality and to the Shuwa-Arabs from
the sampled sites to have allowed this research to take
place.
Bibl iog r aphy
Agorsah E.K., 1985.
Archaeological implications of traditional house construction
among the Nchumuru of northern Ghana. Current Anthropology
26: 103-115.
Agorsah E.K., 1990.
Ethnoarchaeology: the search for a self corrective approach to
the study of past human behaviour. African Archaeological Review 8: 189-208.
Atherton J., 1983.
Ethnoarchaeology in Africa. The African Archaeological Review 1: 75-104.
Barth H., 1965.
Travels and Discoveries in North and Central Africa, Vol. II.
London: Frank Cass.
Bianquis T., 1990.
L’Egypte depuis la conquête Arabe jusqu’à la fi n de l’empire
Fatimide (1171). In: M. el Fasi and I. Hrbek (eds.). Histoire Générale de l’Afrique, Vol. III: 189-220. Paris: UNESCO and Nouvelles Editions Africaines.
Binford L.R., 1978.
Nunamiut Ethnoarchaeology. New York: Academic Press.
Blache J., 1964.
Les Poissons du bassin du Tchad et du bassin adjacent du Mayo
Kebbi: Étude Systématique et Biologique. Paris: Editions ORSTOM.
Blache J. & F. Miton, 1962.
Première Contribution a la Connaissance de la Pêche dans le
Bassin hydrographique Logone-Chari lac Tchad. Paris: Editions ORSTOM.
Conte E. & F. Hagenburger-Sacripanti, 1977.
Habitation et vie quotidienne chez les Arabes de la rive sud du
lac Tchad. Cahiers ORSTOM - Sciences Humaines XIV(3):
289-323.
Cribb R., 1991.
Nomads in Archaeology. Cambridge: Cambridge University Press.
Cunnison I., 1966.
Bagara Arabs. Oxford: Clarendon Press.
David N & C. Kramer, 2001.
Ethnoarchaeology in Action. Cambridge: Cambridge University
Press.
Denham F.R.S., H. Clapperton & Dr. Oudney, 1828.
Narrative of Travels and Discoveries in Northern and Central
Africa in the Years 1822, 1823, and 1824. Vol 2. London: John
Murray.
Fisher H.J., 1975.
The Central Sahara and Sudan. In: R. Gray (ed.). The Cambridge
History of Africa, Vol. 4: 58-141. Cambridge: Cambridge University Press.
Fisher H.J., 1977.
The Eastern Magrib and the Central Sudan. In: R. Oliver (ed.).
The Cambridge History of Africa, Vol. 3: 232-330. Cambridge:
Cambridge University Press.
Shuwa-Arab economics: livestock management and animal consumption
Frantz C., 1981.
Fulbe continuity and change under five flags atop West Africa:
Territoriality, ethnicity, stratification and national integration.
In: J.G. Galaty and P.C. Salzman (eds.). Change and Development
in Nomadic and Pastoral Societies: 89-114. Leiden: E.J. Brill.
Garcin J.C., 1985.
L’Egypte et le monde musulman (du XIIè au début du XVIè siècle). In: D.T. Niane (ed.). Histoire Générale de l’Afrique, Vol. IV:
405-432. Paris: UNESCO and Nouvelles Editions Africaines.
Hagenburger-Sacripanti F., 1977a.
Les Arabes dit “Suwa “ du Nord-Cameroun. Cahiers ORSTOM
- Sciences Humaines XIV(3): 223-249.
Hagenburger-Sacripanti F., 1977b.
Eléments de magie et de sorcellerie chez les Arabes d’Afrique
Centrale. Cahiers ORSTOM - Sciences Humaines XIV(3): 251288.
Hasan Y.F., 1973.
The Arabs and the Sudan. Khartoum: Khartoum University
Press.
Hiskett M., 1976.
The Nineteenth century Jihads in West Africa. In: J.E. Flint (ed.).
The Cambridge History of Africa, Vol 5: 125-169. Cambridge:
Cambridge University Press.
Hiskett M., 1994.
The Course of Islam in Africa. Edinburgh: Edinburgh University Press.
Hodder I., 1982.
Symbols in Action. Cambridge: Cambridge University Press.
57
Laure J., 1974.
Valeur nutritionnelle des Produits de la Pèche conserves artisanalement au Cameroun et au Tchad. Paris: Editions ORSTOM.
Lebeuf A.M.D., 1969.
Les Principautés Kotoko: essai sur le caractère sacré de
l’autorité. Paris: CNRS.
Lebeuf A.M.D., 1987.
Le Royaume du Bagirmi. In: C. Tardits (ed.). Princes et Serviteurs du Royaume. 171-225. Paris: Société d’ethnographie.
MacMichael H.A., 1922.
A History of the Arabs in the Sudan. Cambridge: Cambridge University Press.
Nachtigal G., 1987a.
Sahara and Sudan (Vol. III): The Chad Basin and Bagirmi.
Translated and annotated by A.G.B. Fisher and H.J. Fisher. London: C. Hurst.
Nachtigal G., 1987b.
Voyages et Explorations (1869-1875): Tome I: Au Tibesti, du
Bornu an Bagirmi; Tome II: Au Ouadaï. Paris: Centre d’études
sur l’histoire du Sahara.
Podlewski A.-M., 1966.
La dynamique des principales populations du Nord-Cameroun I
(entre Bénoué et lac Tchad). Cahiers ORSTOM - Sciences Humaines III(special issue).
Podlewski A.M., 1971.
La dynamique des principales populations du Nord-Cameroun
II, “Piémont et plateau de l’Adamaoua”. Cahiers ORSTOM - Sciences Humaines VIII, (special issue).
Holl A.F.C., 2002.
The Land of Houlouf: Genesis of a Chadic Chiefdom (BC 1900 –
AD 1800). Ann Arbor: University of Michigan Museum of Anthropology.
Sadr K., 1991.
The Development of Nomadism in Ancient Northeast Africa.
Philadelphia: University of Pennsylvania Press.
Holl A.F.C., 2003.
Ethnoarchaeology of Shuwa-Arab Settlements. Lanham: Lexington Books.
Smith A.B., 1992.
Pastoralism in Africa: Origins and Development Ecology. London: Hurst and Co.
Holl A.F.C., T.E. Levy, C. Lechevalier & A. Bridault, 1991.
Of Men, Mounds and Cattle: Archaeology and Ethnoarchaeology in the Houlouf Region (Northern Cameroon). West African
Journal of Archaeology 21: 6-36.
Tijani K., 1986.
The Shuwa-Arab. In: M. Adamu and A.H.M. Kirk-Greene (eds.).
Pastoralists of the West African Savanna. 62-73. Manchester:
Manchester University Press.
Holl A.F.C. & T.E. Levy, 1993.
From the Nile Valley to the Chad Basin: Ethnoarchaeology of
Shuwa-Arab settlements. Biblical Archaeologist 56: 166-179.
Yellen J., 1977.
Archaeological Approaches to the Present. New York: Academic Press.
Hrbek I., 1977.
Egypt, Nubia and the Eastern Deserts. In: R. Oliver (ed.). The
Cambridge History of Africa, Vol. 3: 10-98. Cambridge: Cambridge University Press.
Zeltner J.C., 1970.
Histoire des Arabes sur les rives du Lac Tchad. Annales de
l’Université d’Abidjan 2.
Zeltner J.C., 1979.
Les Arabes dans la région du Tchad: Problèmes d’origine et de
chronologie. Sarh: Centre d’études Linguistiques.
Kropacek L., 1985.
La Nubie de la fi n du XIIè siècle à la conquête par les Funj au
début du XVIè siècle. In: D. T. Niane (ed.). Histoire Générale de
l’Afrique, Vol. IV: 433-459. Paris: UNESCO and Nouvelles Editions Africaines.
Zeltner J.C., 1980.
Pages d’histoire du Kanem: Pays Tchadien. Paris: L’Harmattan.
Lange D., 1987.
A Sudanic Chronicle: The Bornu Expeditions of Idris Alauma
(1564-1576). Wiesbaden: Frantz Steiner Verlag.
Zeltner J.C., 1988.
Les Pays du Tchad dans la tourmente 1880-1903. Paris:
L’Harmattan.