Journal of Archaeological Science 31 (2004) 1395–1411
http://www.elsevier.com/locate/jas
Reconstruction of spatial organization in abandoned Maasai
settlements: implications for site structure in the Pastoral
Neolithic of East Africa
Ruth Shahack-Grossa,), Fiona Marshallb, Kathleen Ryanc, Steve Weinera
a
Department of Structural Biology, Weizmann Institute of Science, Rehovot 76100, Israel
Anthropology Department, Washington University, One Brookings Drive, St. Louis, MO 63130, USA
c
MASCA, University of Pennsylvania Museum of Archaeology and Anthropology, 33rd and Spruce Streets,
Philadelphia, PA 19104-6324, USA
b
Received 13 November 2003; received in revised form 22 February 2004
Abstract
The analysis of spatial organization in archaeological sites is important for the interpretation of economic and social issues. In
East Africa, the appearance of mobile herders, adoption of pastoralism by some hunter–gatherers, and spread of competing pastoral
groups, create a complex archaeological record and interpretive problems associated with the beginnings of food production. Spatial
analyses could contribute to their resolution, but are difficult because most sites lack macroscopic features. We present a geoethnoarchaeological study of abandoned pastoral Maasai settlements that allows us to evaluate the ‘‘archaeological visibility’’ of
ephemeral features such as hearths, trash pits, gates, houses and fences. Micromorphology, mineralogy and phytolith analyses show
that features containing ash have the highest visibility. Livestock enclosures, a feature studied by us previously, can also be identified
based on this suite of techniques. Large livestock gates have poor visibility but may be recognized. Small gates, fences and house
floors could not be detected using the methods applied here. Identifying livestock enclosures, trash pits and cooking hearths based
on this approach, and houses based on post-hole positions, will contribute to a better understanding of the spread of food
production in Africa. Our findings will also contribute to studies of pastoralists in other regions of the world.
Ó 2004 Elsevier Ltd. All rights reserved.
Keywords: Maasai pastoralists; Ethnoarchaeology; Mineralogy; Micromorphology; Phytoliths; Geoarchaeology; Site structure
1. Introduction
The spatial organization of human settlements often
reflects sociopolitical, economic, and belief systems, as well
as social relations among families [17,22,29,32–34,44,60].
Interpretation of past societies from spatial data has been
best developed for agricultural societies living in permanent village or urban settings with considerable investment
in architecture [34,36,41]. Recognition and interpretation
of ephemeral spatial features characteristic of highly
mobile hunter–gatherer or pastoral societies has posed
a major challenge for archaeologists. This is largely
) Corresponding author. Tel.: C972-8-9342267; fax: C972-89344136.
E-mail address: [email protected] (R. Shahack-Gross).
0305-4403/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jas.2004.03.003
because mobility constrains the use of building material
and the accumulation of material culture, and sites do not
often preserve well [5,15,22,23,49,50,63].
Studies of one of the fundamental societal changes
in the Holocene in Africadthe beginnings of food
productiondhave not benefited from detailed spatial
data because, unlike other regions of the world, settled
villages are rare among late hunter–gatherer and early
food producing societies. In Africa animals rather than
plants were the earliest domesticates (cattle, donkeys),
and mobile herders the earliest food producers. This
resulted in a different trajectory of intensification and
societal change than in other regions, and a distinctive
material record ( for details see [38]). In keeping with this
pattern, in East Africa, mobile hunter–gatherer and
pastoral groups predominated during the initial period
1396
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
of adoption of food production (termed the Pastoral
Neolithic, ca. 5000–1300 BP). A complex mosaic of
mobile hunter–gatherer and pastoral societies remains in
the region to this day.
The lack of sites with recognizable architecture has
constrained interpretation of the spread of food production in East Africa. A basic understanding of socioeconomic differentiation and change in East Africa
during the adoption of food production has been
obtained from many lines of evidence, especially lithics,
ceramics, and fauna. Still poorly understood, however,
are the nature of interactions between hunter–gatherers
and early pastoralists, variations in the organization of
competing pastoral societies, and household level analyses of gender-based or specialized activities and social
relations [24,46,56]. These issues are fundamental to
understanding the process of spread of food production
in the region and many of them could be addressed with
additional spatial data on late Holocene settlements. To
understand social relations and economic activities
within residential settlements for example, archaeologists need to be able to differentiate domestic household space, public space, and areas of livestock holding
within domestic settlements. In order to achieve this,
information on the size and layout of houses, spatial relations among houses, house-enclosure relations, and
the size of settlements in relation to the number of
domestic units is essential [25,40].
In recent years, micromorphological and geochemical
approaches have contributed to the detection of activity
areas in archaeological sites, allowing identification of
ephemeral or poorly preserved hearths, trash accumulations and livestock enclosures (e.g., [9,13,14,16,26,37,
39,43,55]). These studies have been carried out however,
primarily in European and Eastern Mediterranean
environmental contexts and are thus particularly relevant to soil types, plants, and climates, characteristic of
these regions. In order to carry out geoarchaeological
studies in Africa, these factors (i.e., soils, plants, climate)
need to be characterized locally. Archaeologists have not
however, employed geoarchaeological approaches to
spatial studies widely in Africa. To date, spatial research
in East Africa has focused on location of settlements on
the landscape, and examining activity areas within sites
based on artifact type and density [8,51]. East African
sites of the Pastoral Neolithic period commonly lack
macroscopic features. Hearths are rare. Post-holes have
been recognized from a few sites such as Narosura [48].
The only hut floor so far excavated is from the Elmenteitan site of Ngamuriak [52]. Therefore micromorphological and geochemical research to enhance the
detection of ephemeral features in East African sites is
necessary in order to understand the broad layout of
hunter–gatherer and early pastoral sites.
In this paper, we discuss geoarchaeological approaches
to reconstruction of spatial organization of East African
pastoral sites, based on geo-ethnoarchaeological research
on abandoned Maasai settlements. Fieldwork was conducted in southern Kenya, near Rombo town, among
the Kisongo Maasai (Fig. 1a). We present a study of
the microstructure, mineralogy, and phytolith contents of
house floors, cooking hearths, gates, fences, and trash
pits, sampled from four abandoned pastoral Maasai
settlements. Using this method, we characterize and
evaluate the potential for ‘‘archaeological visibility’’ of
these features.
In the following section, we present background ethnographic and ethnoarchaeological data on East African
Maasai pastoralists, focusing on the structure of settlements in relation to the social activities that take place in
and around them.
1.1. Maasai: ethnographic background
The Maasai are Nilotic speaking pastoralists. Their
society is relatively egalitarian with a clan based social
organization and an age set system. It is however, sometimes termed a gerentocracy with more power going to
elders and to men [12,20,21,42,57–59]. Adult men are in
charge of management of herds, family labor, and care
of livestock. Until recently, warrior age sets were responsible for defense of livestock and settlements. Herding is
primarily carried out by boys and young men. Women
are responsible for the household. They milk livestock,
open gates in the morning to let the family herds out for
grazing, collect firewood and water, build and maintain
houses, clean, cook, sew, and care for children.
Most features of Maasai settlements are determined
by their focus on livestock [62]. Domestic settlements,
enkang in Maa or boma in Swahili, are the primary
settlements used throughout the year by Maasai families. Several other more temporary settlement types exist
(e.g., seasonal cattle camps, warrior encampments and
meat-feasting sites). Several polygamous extended families (3–12 households, 10–50 people) live together in
domestic settlements in order to share labor for herding
and to protect the herds [4,28,30]. Although animals are
herded together, each household is economically independent. The size of domestic settlements varies greatly,
depending upon how many families are living there and
the size of their herds. Cattle holdings may range between 20 and 350 head per family [4,6,30]. Mbae reports
a range of settlement size between 20 and 80 m. Family
settlements last between a few and 20 years, with an
average of 4 years [62].
The spatial organization of domestic settlements
follows set rules which reflect social organization, including polygamy, kinship, and seniority [40] (see also
[29] for the Luo). In a typical settlement, houses are
arranged in a roughly circular pattern around a central
cattle enclosure. The entire settlement is ringed with a
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
1397
thorn fence in order to protect livestock from predators.
The central enclosure is surrounded by smaller livestock
enclosures for calves, sheep and goats [3,40] (see for
example Fig. 2a). Kinship and seniority is indicated by
the position of houses relative to gates. Each adult male
has his own gate in the perimeter fence, and the house of
his first wife is built to the right of this. Houses of
subsequent wives are placed on either side of the gate
[40,42,58]. Houses, averaging 6!3 by 1.5 m are occupied by a woman, her children, and young livestock;
they are made of a mixture of ash, cattle dung, and mud
over a wooden frame [3,40]. Within houses spatial
activities are centered around the cooking hearth where
food preparation, distribution, and consumption occur.
Women sweep their hearths daily and dispose off ashes
in secondary refuse. Separate daytime activity areas for
men and women are common. The men’s area is usually
under a shade tree within 15 m of the settlement [40].
This is the place where men discuss herd management
and politics, rest, play games, and manufacture or repair
tools. The only features in this area are a hearth and
possibly a stake windbreak [40]. The women’s area is
close to the household gate, where they may go to
prepare food and make ornaments. Children also play in
this area.
2. Materials and methods
2.1. Fieldwork
Fieldwork was initiated in the 1990s by one of us
(K.R.) who focused on collection of ethnographic data
from Maasai settlements in southern Kenya [53,54].
Geoarchaeologically oriented fieldwork was conducted
by R.S.-G. in May 1999 and January 2001. One inhabited
and four abandoned bomas were sketch mapped using
compass and tape, and sediments were sampled inside and
outside the abandoned bomas. Ole Koringo, a long time
collaborator of K.R. located abandoned bomas and
features within them, and estimated dates of boma abandonment, based on his life history in the area. In order to
study the taphonomy of features we asked Ole Koringo to
locate bomas of varying ages. We were able to sample four
abandoned bomas: abandoned approximately one year
before sampling (termed AB1), 20 years before sampling
(termed AB20), 30 years before sampling (termed AB30),
and one abandoned 40 years before sampling (termed
AB40) (Fig. 1b). Bomas in the study area were typically
occupied for some 10–15 years.
Fig. 1. (a) Map of Kenya showing the Rombo area near Mt.
Kilimanjaro, where fieldwork was conducted. (b) Topographic map of
the study area showing the location of the inhabited boma (IB) and the
four abandoned bomas (AB1 to AB40). Elevation contours are in feet.
1398
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
Fig. 2. (a) Sketch map of the inhabited boma in 1999. Note the central
cattle enclosure surrounded by smaller enclosures for calves and
caprines. The houses are built around the enclosures on the inner
perimeter of the boma. (b) Sketch maps of the four abandoned bomas.
Note that not all caprine enclosures and concentrations of hearthstones were located in AB20, AB30, and AB40. Sampling pits in all
sites are plotted, and are denoted as follows: control regional
sediments (c), cattle and caprine enclosures (m), gates (g), houses or
only hearths within houses (h), and trash pits (t). Each gate was
sampled by three pits, one at the gate and the other two several meters
to the left/right of the gate. Each house was sampled by three pits, one
inside the house at the living area, one inside the house at the hearth,
and one outside the house.
Sampling locations within bomas included houses
(inside and outside), cooking hearths, gates, fences
and trash pits. The central cattle enclosure and one or
more caprine enclosures had been previously sampled
(reported in [55]) (Fig. 2b). Information regarding the
formation and placement of these features was collected
through conversations with Ole Koringo and his son,
Ole Parmitoro. Several samples of regional sediment
were collected outside each boma. These samples served
as controls to which all analyses of feature sediments
were compared (Fig. 2b).
Each feature was sampled through excavation of
1!1 m test pit. These were excavated to a depth of
about 20 cm below the contact between the feature and
regional sediments. Ole Koringo and R.S.-G. located
gates visually in the inhabited boma and in AB1, and
based on Ole Koringo’s memory for the older abandoned bomas. R.S.-G. sampled fences only in the
inhabited boma and in AB1, where she could clearly
observe their woody remnants. Ole Koringo located
houses in the older abandoned bomas by finding their
hearthstones. Women normally use three local basalt
boulders as hearthstones on which they balance metal
cooking pots. Hearthstones are rarely removed from
abandoned bomas. Hearths were sampled directly between such hearthstones, and living floors were sampled
about 0.5 m away from the hearthstones. Sediments outside the houses, but still within boma perimeters, were
sampled 3–4 m away from hearthstones. Ole Parmitoro
located one trash pit 3–4 m away from the main cattle
gate of AB1. Test pits for the collection of regional sediment samples were dug to approximately 30 cm below
surface. Thickness of stratigraphic units (distinguished
primarily by color) within test pits was measured. Loose
sediment samples (at least 5 g each) from different layers
were collected in plastic bags. Care was taken to sample
the central portion of such layers in order to avoid contamination by over and underlying layers. In cases
where no stratigraphy was observed, samples were taken
in 5 or 10 cm intervals from top to bottom of the test
pits. All loose samples were used for both mineralogical
and phytolith analyses. Block sediment samples were
collected from all feature types and regional sediments,
and were tightly wrapped with paper and masking tape.
Plant samples were taken from extant fences and
included the trees Acacia mellifera (Eiti in Maa, used
for internal fencing and as fuel) and Acacia tortilis
(Entepesi in Maa, used for external fencing). These were
ashed (550 (C/4 h) and analyzed for phytoliths and
mineralogy.
2.2. Laboratory techniques
Micromorphological, mineralogical and phytolith
analyses were performed using methods and instrumentation described in Shahack-Gross et al. [55]. These
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
1399
include a polarizing light microscope (Nikon Labophot2pol), Fourier Transform Infrared (FTIR) spectroscopy
(MIDAC Corp., Costa Mesa, CA, USA), X-Ray powder
Diffraction (XRD) (Rigaku Corp., Japan), Energy
Dispersive X-ray spectrometry (EDS) (Jeol 6400 Scanning Electron Microscope with an EDS Link, Oxford
Instruments), and quantitative phytolith analyses based
on methods developed by Albert et al. [1]. Phytolith
morphologies were quantified from three hearths and
three control regional sediment samples. These were
compared to samples from livestock enclosures previously reported by Shahack-Gross et al. [55]. Phytolith
morphologies follow Mulholland and Rapp [45] and
Albert and Weiner [2]. Micromorphological descriptions
follow Bullock et al. [10] and Courty et al. [13].
the walls and roof, followed by the collapse of the roof
(Fig. 4b). Walls come down last. This was observed in
AB1 in May 1999 and in January 2001. The cooking
hearthstones are the only visible objects left after houses
have been abandoned and degraded.
3. Results
3.1.5. Hearths
Wood is the only source of fuel for the cooking
hearth. Species that do not produce a lot of smoke are
selected, commonly A. mellifera. During the early stages
of lighting a fire, some grasses are also used. Dung is
never burned as fuel. Dung may be burned inside
a house, however, if a pest (i.e., a snake or scorpion) has
entered, because the smoke is very strong. Metal pots
are used for cooking and balanced on three local basalt
boulder hearthstones. Profiles of recently used hearths
from AB1 show a white-gray layer of wood ash, about
10 cm thick (Fig. 3d). Usually, the ash layers overlay
a red-burned soil layer of 1–2 cm thickness. Below, the
sediment did not seem to be heat affected. Charcoal
particles were often observed. In AB20, AB30, and
AB40, profiles between hearthstones always revealed
a grayish ash layer, and sometimes a red-burned layer.
Pieces of charcoal were always present. The ash layers
were usually thin, averaging about 5 cm (Fig. 3e).
3.1. Field observations
3.1.1. Regional sediments (i.e., control samples)
Analyses of all features were compared to those of
the regional sediments. The regional sediments in the
study area are brown or red clays. No stratigraphy
was usually observed at the depths dug in this study
(Fig. 3a). The observed gross structure of regional sediments is blocky and coloration is uniform along the
profile.
3.1.2. Settlements
Sketches of the inhabited boma and four abandoned
bomas are shown in Fig. 2. The structure and features in
the most recently abandoned boma, AB1, were fairly
well preserved. The settlement’s perimeter was clearly
visible, as were the perimeters of enclosures and houses
(Fig. 4a). Less than 20 years after abandonment, older
bomas were completely degraded at the time of sampling. No aboveground features, except for occasional
hearthstones, remained (Fig. 4c).
3.1.3. Houses
Built by women from a wooden frame and covered by
plaster, prepared from a mixture of dung and ash,
traditional Maasai houses are oval. The plaster mixture is about 80% by volume cattle dung and 20% by
volume ash. The wooden frame is made from dense
and compact wood, and known by the Maasai to be less
subject than others to the depredations of termites. This
includes wood from the following trees: A. mellifera,
Eseki (a sacred tree, not identified), Grewia bicolor,
Emapitet (not identified) and A. tortilis. Inside the
houses there are storage and sleeping areas, and a central
cooking hearth.
The process of disintegration of houses starts with
the disarticulation of the dung and ash plaster from
3.1.4. Living floors
Stratigraphic units were not visible in profiles opened
inside houses (Fig. 3b). The only exception is an upper
finely laminated thin layer (1 cm) in the AB1 location.
This layer appeared in all test pits opened in and around
the AB1 location. As a result, we attributed it to a
flooding event that we were told triggered the abandonment of this settlement in December 1998. Profiles
outside houses, in general, also did not have macroscopically visible stratigraphy (Fig. 3c).
3.1.6. Trash pits
A trash pit (about 1!1 m) that is still in use was
observed near the inhabited boma. Trash was of household origin, mostly composed of ash from the cooking
hearth. Bones from meals were also thrown into the
trash after they had been crushed by humans (and later,
dogs) for the extraction of marrow. Other items thrown
into the trash pit include broken calabashes and old
clothing. According to Lekatoo Ole Parmitoro, trash
pits are left open and covered by grass only before the
trash is about to be burned. The trash is burned every
month or two, or earlier if it is rotten or smelly. Every
woman living in the boma has her own trash pit that she
uses as long as she lives in the boma.
One trash pit was sampled in the AB1 location. The
profile of this feature is relatively deep (ca. 40 cm) and
complex (Fig. 3f). The profile was covered by some 6 cm
of local sediment that was deposited on top of the pit
during the 1998 flood. The uppermost anthropogenic
1400
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
Fig. 3. Photographs of sediment profiles in open test pits. (a) Regional sediment in the vicinity of AB1. (b) Sediment inside a house in the AB20
location. (c) Sediment outside the same house as in (b). Note the similarity between the sediments shown in (a), (b), and (c). (d) A hearth profile in the
AB1 location. Note an upper light layer of ash, overlying a red-burned regional sediment. (e) A hearth profile in the AB30 location. Note a very thin
light layer of ash overlying a red-burned regional sediment. ( f) Sediments in the trash pit at the AB1 location. Note larger size and complexity of this
profile relative to those of the cooking hearths. The upper dark layer is probably a humus horizon that formed after the inundation of this locality in
1998. The upper light layer is light orange colored, overlying a black colored layer that contains light and dark orange colored ash lenses. The black
layer is in direct contact with the regional sediment below. Arrows indicate the contacts between burned features and underlying regional sediments.
Scale bar: 10 cm.
layer was a thin (1 cm) black layer that overlaid a 13 cm
thick pale orange layer that contained gray ashy
nodules. This pale orange layer overlaid a black layer
(19 cm thick) that contained a dark orange lens with
several gray ashy nodules. The black layer overlaid the
regional sediment. No bones were observed in the open
profile.
3.1.7. Gates
Maasai settlements in the study area have two types
of gates, wide ones used by both people and livestock,
and narrow ones used primarily by people. Test pits
opened in both types of gates and several meters away
from gates did not show well-defined stratigraphy, and
in fact resemble profiles of control regional sediments.
Fig. 4. Features observed at abandoned bomas. (a) AB1 in May 1999. Note the elevated dung heaps (1, 2) of caprine enclosures, intact house (4), and
fences (5). (b) AB1 in January 2001, view similar to that in (a). Note the caprine dung heaps (1, 2), the flat dung floor of the cattle central enclosure
(3), house (4), same as in (a), but the roof has collapsed, and the absence of enclosure fences, now demarcated by grass growth (5). Field of view in (a)
and (b) is approximately 20 m. (c) View of AB30. Note the absence of above ground features except for three basalt hearthstones (indicated by
arrow).
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
3.1.8. Fences
Fences are made from thorny tree branches. The
branches of A. mellifera have relatively small thorns and
are thus used for internal fencing (i.e., small stock
enclosures and around houses), while the branches of
A. tortilis have large thorns, thus it is used for the defensive external fence surrounding the whole settlement.
In January 2001, all the fences in AB1 had decayed, i.e.,
in less than 2 years after abandonment (Fig. 4b). Old
wooden fences, exposed for many years to the forces
of nature, are not re-used because they are no longer
strong. People prefer to cut fresh trees for building a
new boma. Because the decay of the wood is above
ground, no test pits were opened at these locations.
In summary, field observations in opened test pits
show that living floor and gate profiles are similar to
those of regional sediments. This indicates that such
features, as well as fenced areas, will be difficult to
recognize visually in archaeological sites in East Africa.
On the other hand, features containing ash (i.e., hearths
and trash accumulations) may be recognized visually
based on color differences. Note, however, that ash
accumulations, as observed in the taphonomic sequence
of cooking hearths in this study, tend to thin with time.
This observation indicates that ash accumulations may
also be difficult to recognize visually in archaeological
sites in East Africa.
3.2. Micromorphology
In this section we describe the micromorphology of
regional sediments and ash that was obtained experimentally from Acacia trees used by people living in
the study area (Table 1). These descriptions serve for
comparison with the micromorphology of boma features. The micromorphology of house floors and gate
areas is similar to that of regional sediments and is thus
not elaborated on in this section.
3.2.1. Regional sediments
Five embedded blocks were prepared from regional
sediments (Fig. 5a). All five are similar, composed of
a clay-rich groundmass that contains randomly distributed grains of basalt and basalt-derived minerals (i.e.,
feldspars and pyroxenes), and occasional pedogenic
carbonate nodules. The concentration of the grains and
nodules is, on average, 10–20%. The clay-rich groundmass shows no preferred orientation (i.e., an undifferentiated birefringence, or, b-fabric), probably indicating
little pedogenic activity. The microstructure is complex,
including poorly developed angular blocky, granular,
and crumbly structures. Evidence for bioturbation
includes earthworm granules (sensu [11]), possible
earthworm casts, and an overall granular appearance.
1401
3.2.2. Experimental wood ashing
The ashes of the wood and bark of A. mellifera and
A. tortilis were embedded on microscope slides. They are
composed of pseudomorphs of calcite after Ca-oxalate
crystals. Some rhombohedral crystals are arranged in
linear arrays (Fig. 5b).
3.2.3. Hearths
Block samples collected from hearths show characteristic features of wood ash. The ash is composed
of rhombic calcite crystals, some present in the form
of linear arrays similar to those observed in ashed
A. mellifera and A. tortilis (Fig. 5c). Pieces of charcoal
are always present, and in recent ashes from AB1 some
woody tissues that are only partially burned still contain
Ca-oxalate druses and crystals embedded in them. The
ashy layer also contains lumps of red-burned regional
sediment. The burned regional sediments do not differ
micromorphologically from unburned regional sediments.
Blocks of hearths from AB20 and AB40 show that
the calcitic wood ash and charcoal fragments are mixed
with local regional sediments, but the layer of ash can be
observed clearly (Fig. 5d). One block of ash from AB1
was taken from a house with a collapsed roof. Here,
a thin layer (about 0.5 cm) of vegetal debris from the
roof accumulated on top of the ash. The ash itself was
heavily stained with organic compounds and appears to
be in the process of dissolution. On top of the vegetal
layer, a thin layer (0.2 cm) of regional sediment had
already accumulated by 2001 (Fig. 5e, g).
3.2.4. Trash pit
One block obtained from the trash pit found outside
the perimeter of AB1 shows a mixture of ash and sediment. Certain areas appear to be completely charred,
some contain vegetal matter that is either not burned or
partially burned. Many areas are organically stained.
The microstructure is relatively massive, but more porous in the charred areas. Vertical, water-draining channels occur in the thick charred layer. The upper ashy
layer (orange colored) contains phosphatic nodules and
a few bone fragments that appear to be burned (Fig. 5f).
The intermediate ashy layer (gray colored) seems to be
composed of well-preserved wood ash, based on the occurrence of linear arrays of calcite pseudomorphs after
Ca-oxalate crystals. It also contains a few bone fragments that appear to be burned. The lower layer is composed primarily of charred vegetal matter.
In summary, only ash-containing features could be
differentiated from regional sediments based on micromorphological criteria. However, calcitic ash seems to be
dissolved by organic solutions, and probably also rainfall,
in a matter of a few years. This contributes to the thinning
1402
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
Table 1
Micromorphological descriptions of sampled sediments (refer to text for terminology)
Context (Sample nos.)
Microstructure
Related distributiona
Fine fraction
Notes
Mostly granular and
crumb with compound
packing voids. Also poorly
developed angular blocky
structure. Other voids are
vughs, chambers and
channels.
Open to closed porphyric.
Light to dark brown clay
with undifferentiated
b-fabric.
Coarse fraction includes
20 mm to 4 mm grains of
basalt, feldspars and
pyroxene, poorly sorted and
angular. Occasional
pedogenic carbonate
nodules (up to 200 mm),
earthworm granules and
possible earthworm casts
(up to 300 mm).
Crack and laminated
microstructure at top of
samples 31 and 181 due to
flooding event, together with
crystallitic b-fabric
(carbonate probably
originates from within
boma sediments).
b
Regional sediments
Representative
(KEN-19, 31, 133,
181, 225)
Living floor profiles
AB1 (KEN-150, 168)
AB20 (KEN-121, 122)
Hearths
AB1 (KEN-143, 160,
161)
AB20 (KEN-44)
Same as regional sediments. Same as regional sediments. Same as regional sediments.
Same as regional sediments. Same as regional sediments. Same as regional sediments.
Upper (1 cm): Vegetal
Monic.
matter, Spongy structure.
Open porphyric.
Middle (w10 cm): Ash.
Angular to sub-angular
blocky structures. In certain
areas vesicular structure due
to dissolution. Mostly
planar voids and vughs, few
channels and chambers.
Closed porphyric.
Lower: Burned and
unburned regional sediment.
Granular and crumb
structures, sometimes
slightly compacted.
Compound packing voids
and vughs.
Upper (3 cm): Regional
Closed porphyric.
sediment, granular structure.
Middle (7 cm):
(a) Three cm of regional
sediment mixed with
enclosure sediment
(probably cattle), granular
structure.
Porphyric.
None.
White, gray and
yellowish calcite.
Crystallitic b-fabric.
Upper parts are mixtures
of clay from regional
sediments and calcite from
ash, with a crystallitic
b-fabric. Lower parts are
similar to regional
sediments.
Same as regional
sediments.
Light to dark brown clay
in granules of regional
sediments, and light
brown clay mixed with
opaline phytoliths in
granules of organicpoor cattle enclosure
sediments. Crystallitic
b-fabric.
Probably from collapse of
house roof and/or walls.
Organic staining especially
in upper parts. Calcite
crystals are rhombohedral,
up to 15 mm. Coarse fraction
contains granules of burned
regional sediment (0.1–
5 mm) and angular pieces
of charcoal (0.01–15 mm).
Possible phosphatic nodules.
Few small areas contain
spherulites.
Upper parts may also
include charcoal fragments.
Accumulated, postabandonment regional
sediment.
This layer may be interpreted
as originating from the
collapse of the house roof
and/or walls, covering the
hearth. Contains a few
phosphatic nodules.
Enclosure sediments show
faint ‘‘undulating’’ structure
and contain relatively many
carbonatic minerals. These
may be attributed to
degraded Ca-oxalates that
were originally from the ash
that was mixed with the dung
to plaster the house roof and
walls.
1403
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
Table 1 (continued )
Context (Sample nos.)
Trash pit
AB1 (KEN-193)
Microstructure
Related distributiona
Fine fraction
Notes
(b) Four cm of ash and
charcoal mixed with
regional sediment,
vesicular microstructure.
Porphyric.
Gray to brown calcite
mixed with clay.
Crystallitic b-fabric.
This layer represents the
hearth. The ash is still
calcitic and the vesicular
structure implies that it is in
a process of dissolution.
A few phosphatic nodules
occur.
Lower: Regional
sediment, sub-angular
blocky structure.
Closed porphyric.
Same as regional
sediments.
Upper (1 cm): Black layer
containing a mixture of
organic matter, regional
sediment, calcite and
spherulites. Granular
microstructure.
Middle a (5 cm): Dark
orange layer containing
vegetal matter, charcoal,
regional sediment, calcitic
ash, possible enclosure
sediments and phosphatic
nodules. Granular to
massive microstructures,
compound packing voids.
Middle b (4 cm): Pale
orange layer containing
material as in the layer
above it, but with
relatively more ash.
Massive to granular
structure, compound
packing voids.
Porphyric.
Yellow, brown and black
organic matter and clay.
Crystallitic b-fabric.
Porphyric.
Gray, yellow, orange and
brown, calcite, organic
matter and clay.
Undifferentiated and
crystallitic b-fabrics.
Porphyric.
Gray to brown calcite
and clay, crystallitic
b-fabric.
Chitonic.
Brown to black organic
matter, undifferentiated
and crystallitic b-fabrics.
Lower: Black layer
containing charred and
partially charred vegetal
matter with small amounts
of calcitic ash. Spongy
structure, simple packing
voids.
This layer probably
represents the lower part of
the flood-deposited material,
mostly originating from
anthropogenic sediments
from AB1.
This layer seems to represent
burned material with its ash
partially dissolved, and
formation of phosphatic
nodules.
This layer seems to represent
better preserved ash, some of
it still contains linear arrays
of calcite pseudomorphs
after Ca-oxalate rhombs.
This layer contains a few
burned and unburned bone
fragments (0.05–0.9 mm).
The lower part of the layer
contains vertical, water
draining, channels.
This layer seems to represent
a burning event with low
oxygen availability, thus
forming charcoal. The
internal organization of this
layer is disturbed due to
water infiltration from the
layer above it. There are
a few calcium oxalate druses.
a
The relationship between the coarse and fine units. Porphyric: large units occur in a dense groundmass of smaller units; monic: presence of units
of one size group; chitonic: coarse units surrounded by smaller units [10].
b
Previously reported in Shahack-Gross et al. [55].
of ash profiles, probably together with compaction of
these sediments. Therefore, small ashy features such as
cooking hearths may be difficult to recognize in archaeological sites in East Africa (and probably elsewhere),
while ashy trash accumulations are more likely to preserve and be recognized visually in archaeological sites.
3.3. Mineralogical characterization of bulk samples
Table 2 shows the minerals present in regional sediments, experimentally ashed trees and boma features.
Regional sediments in the study area are largely composed of clay minerals ( for a more detailed characterization see [55]). The FTIR spectra of sediments
collected from house floors resemble those of regional
sediments (Fig. 6a). The FTIR spectra of ashes collected
from hearths at the AB1 location are composed of
calcite (Fig. 6b). Fresh ash of A. mellifera and A. tortilis,
experimentally burned in the laboratory, is composed
mainly of calcite, some ammonium sulfate and traces of
apatite (spectrum not shown here), from either the wood
or the bark of the trees. In this context, the absence of
1404
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
Fig. 5. Photomicrographs of thin sections (a, c–f), of a grain mount (b), and a photograph of a scanned thin section (g). Scale bar in (a) also applies to
(c–f). (a) Regional sediment sampled at the AB20 location. Note an overall granular/crumb structure of clay-rich groundmass and presence of grains
of pyroxene (1), feldspar (2), and iron oxides (3). (b) Linear arrays of rhombohedral calcite crystals that were formed as pseudomorphs after Caoxalate crystals, derived from experimentally ashed bark of Acacia mellifera. (c) Ash from a hearth at the AB1 location. The gray groundmass is
calcitic, the black particles are charcoal, and the granular feature at the upper right is burned regional sediment. (d) Ash from a hearth at the AB20
location, showing extensive mixing with regional sediment. (e) Ash from the AB1 location showing signs of dissolution. (f ) Sediment from the trash
pit at the AB1 location, from the orange colored upper layer. Note an overall granular appearance, burned bone (4) and Ca-phosphate nodule (5).
Most of the calcitic ash had been dissolved. (g) The upper part of a thin section of a hearth in the AB1 location, where the roof of the house had
fallen. Note pristine calcitic ash in the lower part, a layer of dissolving ash above it (higher magnification of this layer is Fig. 5e), and a top layer of
vegetal matter covered by regional sediment.
traces of apatite in the hearth ash is puzzling. The
absence of ammonium sulfate in hearth ash is explained
by its high solubility in water. Upon dissolution of fresh
hearth ash in 1 N HCl, its insoluble remains have an
FTIR spectrum characterized by a broad absorption
centered around 1050 cm1 (Fig. 6c). Such a spectrum
was also obtained from the acid insoluble fraction (AIF)
of the experimentally ashed tree barks, but almost
nothing remained of the experimentally ashed wood. An
orange-colored layer from the trash pit in the AB1
location has an FTIR spectrum similar to that of the
AIF of hearth ash (Fig. 6d, cf. Fig. 6c). The material
present in the AIF of the hearth ash (Fig. 6c) is probably
burned regional sediment, with its clay component
undergoing a crystallographic transformation due to
pyrolisation. This interpretation is supported by the
orange color of the sediment from the trash pit (Fig. 6d).
The more soluble minerals in the sediment of the orange
layer of the trash pit must have dissolved, as the only
material identified in this layer by FTIR spectroscopy is
the burned regional sediment.
In summary, just as in the field and micromorphological observations, the only really indicative mineralogical signals in bulk samples are from ash-containing
features.
3.4. Quantitative phytolith analyses
Phytoliths in the study area can preserve for at least
thousands of years [55]. Phytolith concentrations from
regional sediments and boma features are shown in Fig. 7.
Concentrations are presented as millions of phytoliths per
1 g of total sediment. Concentrations in regional sediments, sediments from living floors, hearths, gates, and
around gates are all similar. A significant difference in
phytolith concentrations exists only for the trash pit
sediments, with average concentrations higher than 10
million phytoliths in 1 g of sediment. For comparison, the
phytolith concentrations from enclosure sediments range
between 1 and 55 million phytoliths in 1 g of sediment,
averaging more than 10 million phytoliths in 1 g of
sediment [55].
1405
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
Table 2
Minerals present in regional, enclosure and other anthropogenic sediments (the average weight % of the acid insoluble fraction (AIF) is shown for
each category; the table also includes the results from the experimental ashing of two wood species)
Category
Sub-category
a
Regional sediments
Cattle enclosure
sedimentsa
Caprine enclosure
sedimentsa
Organic-rich
Average weight % of AIF
Clay (kaolinite).
Vermiculite, plagioclase,
olivine, iron oxides,
sodium nitrate, possible
carbonates, opal,
unidentified peaks at 1010,
754 and 746–748 cm1.
Monohydrocalcite.
89.3G5.87 (n ¼ 15)
Clay, organic matter.
Clay, opal.
Organic-rich
Clay, monohydrocalcite,
ammonium sulfate,
organic matter.
Regional sediments
overlain by enclosure
sedimentsa
Living floors
Experimental wood
ashing (Acacia mellifera
and Acacia tortilis)
Minor components
Organic-poor
Organic-poor
Hearths
Major components
Clay, monohydrocalcite,
opal, Mg-rich calcite.
Clay.
Clay.
Recent ash (AB1)
Old ash (AB20,
AB30, AB40)
Wood
Calcite.
Clay, calcite.
Bark
Calcite.
Calcite.
Trash pit
Burned regional sediment
Gates
Clay.
a
Monohydrocalcite,
organic matter.
Mg-rich calcite,
unidentified peaks at 1155,
754 and 615 cm1.
Organic matter, possible
Ca-oxalate, unidentified
peaks at 879, 688 and
574 cm1.
Carbonate, opal, sodium
nitrate, unidentified peaks
at 688 and 754 cm1.
Sodium nitrate, unidentified
peak at 688 cm1.
Possible clay.
Unidentified peak at
688 cm1.
Ammonium sulfate in
A. tortilis only.
Possible clay, ammonium
sulfate in A. tortilis only.
Opal, possible phosphate
minerals, sodium nitrate.
Unidentified peaks at 688
and 749 cm1.
67.1G21.55 (n ¼ 3)
After ashing: 46.1G0.07
(n ¼ 2)
82.8G9.81 (n ¼ 16)
55.6G16.92 (n ¼ 5)
After ashing: 32.6G11.95
(n ¼ 4)
71.3G15.34 (n ¼ 11)
87.0G6.36 (n ¼ 14)
91.4G3.92 (n ¼ 12)
16.0G1.60 (n ¼ 5)
80.9G4.42 (n ¼ 4)
0.7G0.91 (n ¼ 2)
3.4G3.56 (n ¼ 2)
72.5G3.41 (n ¼ 5)
88.0G6.69 (n ¼ 18)
Previously reported in Shahack-Gross et al. [55].
Exceptionally high phytolith concentrations are found
in one sample of regional sediment. This sample is from
AB30 where the boma perimeter was not observed, thus
this sample may actually represent a livestock enclosure
sediment or a livestock gate. Sediments from presumed
house floors in AB20 and AB30 have phytolith concentrations similar to those of regional sediments, except for
two samples taken from the upper 5 cm of these profiles.
In addition, two sediments sampled outside houses have
relatively high phytolith concentrations that may be due
to the presence of dung around the houses. Two hearths
sampled in AB20 also had phytolith concentrations
significantly higher than in regional sediments. In addition, two gate samples had relatively higher concentrations of phytoliths than regional sediments. These are
from the wide gate of AB1 that served for the passage of
livestock.
Phytolith concentrations in acacia tree bark, as
represented in the ashing experiment are very low. Thus
identifying the presence of wooden fences based on phytoliths is not feasible. Overall, it appears that phytolith
concentrations in the range of 0 to 4 millions in 1 g of
sediment characterize regional sediments and most of the
boma features. Higher phytolith concentrations represent
anthropogenic addition of material rich in phytoliths.
Small amounts of additional grassy material result in
concentrations in the range of 4–8 million phytoliths in
1 g of sediment, while addition of large amounts as in the
trash pit and livestock enclosures result in much higher
concentrations of phytoliths.
Analyses of phytoliths with characteristic morphologies were performed on samples of hearth ash, cattle
enclosure sediments, caprine (i.e., sheep and goats) enclosure sediments and regional sediments. The phytolith
1406
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
Fig. 6. FTIR spectra of: (a) sediment from a house floor sampled in
AB1. The peaks at 1035, 913, 542 and 472 cm1 are from clay. The
peak at 1384 cm1 is from sodium nitrate, a salt usually forming in
sediments in dry areas. The peak at 688 cm1 is probably from silicate
minerals that are found in sediments in the study area. This spectrum is
essentially similar to those of regional sediments in the study area (cf.,
Fig. 9a of [55]). (b) An ash layer sampled in a hearth in AB1. The
absorption around 1450 cm1 together with the peaks at 875 and
713 cm1 are from calcite. (c) The acid insoluble fraction of the sediment in (b) showing broad absorptions around 1049, 473 and 567 cm1
indicative of burned regional sediment. (d) A dark orange-colored
sediment from the trash pit outside AB1. Note the similarity of this
material to that in (c), indicating that this sediment is composed mostly
of burned regional sediment. The peak at 604 cm1 together with the
absorption around 564 may indicate the presence of Ca-phosphate.
assemblages are quite similar to one another, with no
consistent differences that can be used as indicators of
specific anthropogenic features (Fig. 8).
4. Discussion
The results of this study are relevant to understanding site taphonomy, to feature recognition and spatial
analyses of structures of mobile hunter–gatherer and
pastoral settlements, and to design of excavation strategies. The study shows that a range of features from
open-air Maasai settlements, hearths, trash pits and livestock gates, can be differentiated from sediments outside
settlements (regional sediments), using a suite of techniques that characterize the micromorphology, mineralogy
and phytoliths of the anthropogenic sediments. Table 3
summarizes the indicators that can be used to identify
each feature ( for more details on livestock enclosures see
[55]). Hearths and trash pits can be identified micromorphologically, mineralogically, and based on their
phytolith concentrations. Livestock gates may be identified based on their phytolith contents alone. Techniques
employed in this study did not however, allow identification of living floors and fences. These findings are
especially significant for archaeologists seeking to identify spatial patterning in mobile settlements because they
are counterintuitive. Houses and fences are by far the
broadest and most visible features of present day pastoral
settlements, but could not be identified. Smaller scale
features such as hearths, trash pits and livestock gates
could be identified. Fortunately these identifiable features
are spatially significant. In recent years, ethnoarchaeological studies have shown that hearths and trash piles are
key to the recognition of broad spatial patterning and
more relevant to the social organization of settlements
than previously realized (e.g., [15,18,35,40,47]). Gates are
also key to understanding the spatial layout and social
structure (number of households, kinship and seniority)
of African settlements [29,40]. There is also much
variation among contemporary pastoral ethnic groups
in the spatial layout of such features, and spatial analyses
may contribute to differentiating among Neolithic
pastoral groups. In addition, such spatial data will allow
the hypothesis that hunter–gatherers adopting food
production in East Africa lived in settlements that varied
spatially from pastoral norms to be tested.
Below we discuss factors underlying the identification
of the most prominent features that could be identified
in Maasai settlements, namely hearths and trash pits.
Both features are composed of wood ash, characterized
by the mineral calcite. The source of the calcite is the
combustion of Ca-oxalate monohydrate (whewellite,
CaC2O4H2O) crystals that are commonly found in
wood [19,31,61]. Our experimental wood ashing shows
that well-preserved ash is characterized by the presence
of linear arrays of calcite pseudomorphs after Caoxalate crystals. This accords with Gourlay and Grime’s
[27] histological study of African acacia species, where
they show that rhombic Ca-oxalate crystals are arranged
inside the woody tissue in long chains. This study also
shows that combustion of A. mellifera is similar to that
of coal, which may explain the Maasai preference for
this tree species as fuel. Gourlay and Grime [27] point
out as well that the presence of Ca-oxalates in acacia
trees makes them less palatable to termites, which
supports the Maasai explanation for their preference of
certain trees for building materials. This suggests that
there are reasons for choices of fuel and building
materials to have been similar in the region in the past,
and for the likelihood of detection of present features to
be relevant to the past.
Dissolution of calcitic ash in one hearth in the AB1
location was also observed. In this case the hearth was
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
1407
Fig. 7. Phytolith concentrations in sediments from various boma features and from the experimentally ashed trees. Note the low amount of phytoliths
in ashed tree bark from the study area. Also, note that in all features except for the trash pit, the range and values of phytolith concentrations are
similar to those of regional sediments. Only in the trash pit, where grasses are added deliberately, are phytolith concentrations significantly higher
than in regional sediments.
covered by roof-fall (i.e., organic matter). The decomposition of organic matter involves the formation of
organic acids which may be the agents dissolving the
calcitic ash. It is important to note that even if the ash
dissolves, its insoluble component of burned regional
sediment remains. The latter can be distinguished from
unburned regional sediment by infrared spectroscopy,
and may serve as an indicator of hearth or trash
Fig. 8. Percentages of characteristic phytolith morphological groups from regional sediments (dotted), hearths (white), cattle enclosure sediments
(oblique dashes), and caprine enclosure sediments (black). sc refers to short cells, and ppp refers to parallelpiped morphology. Phytolith
morphologies are roughly grouped into taxonomic categories of grasses, dicotyledons and wood/bark. For detailed descriptions of phytolith
morphologies refer to Albert and Weiner [2]. Note overall similarity in percentages of phytolith morphologies between all four groups of sediments.
1408
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
Table 3
Summary of geoarchaeological indicators that can be used to identify
features in pastoral sites that lack permanent architecture
Feature
Micromorphology
Mineralogy
Phytolith
concentration
Enclosures
Hearths
Trash pits
Gates
House floors
Fences
O
O
O
–
–
–
O
O
O
–
–
–
O
–
O
O
–
–
accumulations. Taken together, our evidence suggests
that hearths and trash pits may be identified mineralogically either by the presence of calcite in cases of good
preservation, or presence of burned regional sediment in
cases of poor preservation. This contrasts with the
findings of previous studies of the insoluble remains of
wood ash in the Mediterranean region which suggested
identification based on the presence of wood phytoliths
and siliceous aggregates (e.g., [1]). Wood phytoliths were
not diagnostic and siliceous aggregates not found in our
study. This highlights the importance of studying the
visibility of features in different regions of the world.
Following this discussion of site taphonomy and
feature recognition, the important question is how likely
are these indicators to survive into the archaeological
record? These degraded features of Maasai pastoral sites
(hearths, trash pits and livestock gates) can be identified
based on a suite of indicators that persist up to 40 years
after site abandonment. It is highly likely that they will
persist in archaeological sites. Conditions that are especially conducive to this are the following. Thick
features such as enclosure and trash pit sediments preserve well micromorphologically at depths where they
are protected from fragmentation by root action and
bioturbation. Indicative minerals should similarly preserve for long periods of time as long as the conditions
in the sediment correspond to their stability fields. Opal
( phytoliths), for example, is stable in the pH range
between 5 and 9 [31], and the mineral dahllite is stable
above pH 7 [7]. Considering the alkaline nature of East
African soils, opal phytoliths and authigenic dahllite are
likely to persist for long periods of time. Therefore, we
do not think that phytolith concentrations will change
much over long periods of time, especially in the deeper
parts of features, away from mechanical mixing through
root action and bioturbation. We conclude that archaeologists can identify livestock enclosures, trash accumulations and hearths in East African Pastoral Neolithic
sites using the combination of the lines of evidence
presented here.
Having discussed identifiable features, it is also important to examine the reasons for the lack of visibility
of houses and fences in our study. No compacted features
in living floors of houses were detected microscopically in
this study (see also [26]). This may relate to loading
associated with the human frame, as compaction was
observed microscopically in regional sediments that
underlie livestock enclosures [55]. This indicates that the
soil type found in the study area may form compacted
surfaces only under very heavy loads (i.e., cattle vs.
human; w200 kg vs. w60–70 kg). Thus, in order to locate
living floors, or house perimeters, in East African sites,
the traditional post-hole feature is more useful than
compacted features. Wooden fences were not identified in
this study mainly because the two Acacia species used in
our study area do not produce distinctive phytoliths. In
other regions of the world, however, trees and bushes
used for construction of fences and/or houses might contain distinctive phytoliths and could be used as indicators
for such features.
Our findings are also significant for future excavation
strategies. This study shows that the size and placement
of identifiable features may hamper their detection
during excavation. Cooking hearths have a small area
and are thin. We advocate therefore that slight color
changes in excavated profiles/areas, and especially in
thin (1–2 cm) layers or lenses should be recorded and
possibly analyzed, as these may represent poorly preserved hearths. In addition trash pits in the study area
are located outside the site’s perimeter, and as a result
excavations of pastoral and hunter–gatherer sites should
extend beyond the site’s perimeter (this is also necessary
for obtaining regional sediments as controls). Finally the
current practice of focusing on small excavated areas in
large sites minimizes the chance of encountering more
than one feature ( for example, a whole excavation may
be placed in one large livestock enclosure). As a result
we suggest that future excavations focus on horizontal
rather than vertical aspects of sites. Once a general
understanding of site-structure has been achieved,
certain areas may be dedicated for vertical excavation.
In summary, in order to start looking for patterns in
internal arrangement of Holocene sites in East Africa,
it is necessary to conduct excavations on a scale that
will locate as many features as possible and allow
redundancy in spatial patterning to emerge.
5. Conclusion
This geo-ethnoarchaeological study of abandoned
pastoral Maasai settlements provides the first evaluation
of the ‘‘archaeological visibility’’ of ephemeral features
such as hearths, trash pits, gates, houses and fences in
African contexts. Micromorphology, mineralogy and
phytolith analyses show that features containing ash
have the highest visibility and allow clear identification
of trash pits and cooking hearths on mobile pastoral
settlements. Large livestock gates may also be recognized. Interestingly the most visible features of pastoral
settlements today, fences, house floors and small gates
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
could not be detected using the methods applied here.
These findings regarding identification of hearths, trash
pits and livestock gates complement and significantly
extend previous geoarchaeological identification of
livestock enclosures in Africa, and studies of the remains
of wood ash in other regions [1,55]. They also have
important implications for the design of future research
on African pastoral sites.
Taken together micromophological, mineralogical
and phytoliths analyses, combined with identification of
hearths based on stone features, house floors based on
post-holes, and lithic, ceramic and faunal analyses, will
enable archaeologists to reconstruct relationships among
features within mobile sites. These data may be used to
address issues of socio-economic identity of sites, the
nature of interactions between hunter–gatherers and early
food producers, variation in the organization of competing pastoral societies, and household level analyses of
gender-based or specialized activities during the period of
advent of food production into East Africa. This approach will also be useful in addressing the interpretation
of pastoral, and other ephemeral sites associated with
mobile people in other regions of the world.
Acknowledgements
We thank Dr. Karega-Munene, Head and Senior
Research Scientist, Archaeology Division, National
Museums of Kenya for his collegial support and friendship over many years of collaborative research. We
thank all members of our team, S. M. Kahinju, P. N.
Kunoni, C. Ngange, K. Seberini and J. Ndungu, W. R.
Fitts, Parmitoro Ole Koringo, Ole Kaaka, Lekatoo Ole
Parmitoro, and Ole Seela for their hard work and
unfailing good humor. We thank the staff of the
Archaeology and Osteology Divisions of the National
Museums of Kenya for their courtesy and encouragement. Very special thanks go to the families of Ole Kaaka
and Ole Koringo, our main Maasai consultants for many
years, for their hospitality, patience, and friendship.
Permission to conduct the study was granted by the
President’s Office of the Government of Kenya, in collaboration with the National Museums of Kenya. R.S.-G.
thanks R. M. Albert for discussions about issues related
to phytolith data, and F. Berna for initiation of heating
experiments of sediments. The study was supported by
an NSF Doctoral Dissertation Improvement Grant,
a Wenner-Gren Foundation grant and the Kimmel
Center for Archaeological Science, Weizmann Institute,
to R. S.-G., an Israel Science Foundation grant to S. W,
and grants to K. R. from the NSF, the Social science
Research Council, the Boise Fund, Sigma Xi, and the
University of Pennsylvania Museum of Archaeology and
Anthropology. We are indebted to Diane GiffordGonzalez and Bernard Mbae for ‘‘showing the way’’.
1409
References
[1] R.M. Albert, A. Tsatskin, A. Ronen, O. Lavi, L. Estroff, S. LevYadun, S. Weiner, Mode of occupation of Tabun Cave, Mt
Carmel, Israel during the later Mousterian: a study of the
sediments and phytoliths, Journal of Archaeological Science 26
(1999) 1249–1260.
[2] R.M. Albert, S. Weiner, Study of phytoliths in prehistoric ash
layers using a quantitative approach, in: J.D. Meunier, F. Colin
(Eds.), Phytoliths, Applications in Earth Science and Human
History, Balkema, Rotterdam, 2000, pp. 251–266.
[3] K.B. Andersen, African Traditional Architecture, Oxford University Press, Nairobi, 1978.
[4] K. Arhem, Pastoral Man in the Garden of Eden, Uppsala
Research Reports in Cultural Anthropology, 1985.
[5] E.B. Banning, I. Kohler-Rollefson, Ethnographic lessons for the
pastoral past: camp locations and material remains near Beidha,
Southern Jordan, in: O. Bar-Yosef, A. Khazanov (Eds.),
Pastoralism in the Levant: Archaeological Materials in Anthropological Perspectives, Monographs in World Archaeology No.
10, Prehistory Press, Chicago, 1992, pp. 181–204.
[6] S. Bekure, B.E. Grandin, Introduction, in: S. Bekure, P.N. de
Leeuw, B.E. Grandin, P.J.H. Neate (Eds.), Maasai Herding: an
analysis of the livestock production system of Maasai pastoralists
in eastern Kajiado District, Kenya, ILCA Systems Study No. 4,
Addis Ababa, Ethiopia, 1991, pp. 1–5.
[7] F. Berna, A. Matthews, S. Weiner, Solubilities of bone mineral
from archaeological sites: the recrystallization window, Journal of
Archaeological Science, in press. doi: 10.1016/j.jas.2003.12.003.
[8] J.R.F. Bower, C.M. Nelson, A.F. Waibel, S. Wandibba, The
University of Massachusetts’ Later Stone age/Pastoral ‘Neolithic’
comparative study in central Kenya: an overview, Azania 12
(1977) 119–146.
[9] J.E. Brochier, P. Villa, M. Giacomara, Shepherds and sediments:
geo ethnoarchaeology of pastoral sites, Journal of Anthropological Archaeology 11 (1992) 47–102.
[10] P. Bullock, N. Fedoroff, A. Jongerius, G. Stoops, T. Tursina,
Handbook for Soil Thin Section Description, Waine Research
Publications, Wolverhampton, England, 1985.
[11] M. Canti, Origin of calcium carbonate granules found in buried
soils and Quaternary deposits, Boreas 27 (1998) 275–288.
[12] T. Chieni, P. Spencer, The world of Teleila: reflections of a Maasai
woman in Matapato, in: T. Spear, R. Waller (Eds.), Being
Maasai, James Curry Ltd, London, 1993, pp. 157–173.
[13] M.A. Courty, P. Goldberg, R.I. Macphail, Soils and Micromorphology in Archaeology, Cambridge University Press, Cambridge, 1989.
[14] M.A. Courty, R.I. Macphail, J. Wattez, Soil micromorphological
indicators of pastoralism, with special reference to Arene
Candide, Finale Ligure, Italy, Rivista di Studi Liguri, A. LVII
(1991) 127–150.
[15] R.L.D. Cribb, Nomads in Archaeology, Cambridge University
Press, Cambridge, 1991.
[16] R.P. Evershed, P.H. Bethell, P.J. Reynolds, N.J. Walsh, 5b
stigmastanol and related 5b-stanols as biomarkers of manuring:
analysis of modern experimental material and assessment of
the archaeological potential, Journal of Archaeological Science
24 (1984) 485–495.
[17] J.W. Fisher, H.C. Strickland, Ethnoarchaeology among the Efe
Pygmies, Zaire: spatial organization of campsites, American
Journal of Physical Anthropology 78 (1989) 473–484.
[18] J.W. Fisher, H.C. Strickland, Dwellings and fireplaces: keys to Efe
Pygmy campsite structure, in: C.S. Gamble, W.A. Boismier (Eds.),
Ethnoarchaeological Approaches to Mobile Campsites, International Monographs in Prehistory, Ethnoarchaeological Series 1,
Ann Arbor, Michigan, 1991, pp. 215–236.
1410
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
[19] V.R. Franceschi, H.T. Horner, Calcium oxalate crystals in plants,
The Botanical Review 46 (1980) 361–427.
[20] J.G. Galaty, Land and livestock among Kenyan Maasai, Journal
of African Studies 16 (1981) 68–88.
[21] J.G. Galaty, Being ‘‘Maasai’’; being ‘‘people-of-cattle’’: ethnic
shifters in East Africa, American Ethnologist 9 (1982) 1–20.
[22] C.S. Gamble, W.A. Boismier, Ethnoarchaeological Approaches to
Mobile Campsites, International Monographs in Prehistory,
Ethnoarchaeological Series 1, Ann Arbor, Michigan, 1991.
[23] D.P. Gifford, Ethnoarchaeological observations of natural processes affecting cultural materials, in: R.A. Gould (Ed.), Explanation in Ethnoarchaeology, University of New Mexico Press,
Albugquerque, 1978, pp. 77–101.
[24] D.P. Gifford-Gonzalez, Gender and early pastoralists in East
Africa, in: S. Kent (Ed.), Gender in African Prehistory, Altamira
Press, Walnut Creek, California, 1998, pp. 115–137.
[25] D.P. Gifford-Gonzalez, Early pastoralists in East Africa: ecological and social dimensions, Journal of Anthropological Archaeology 17 (1998) 166–200.
[26] P. Goldberg, I. Whitbread, Micromorphological study of a
Bedouin tent floor, in: P. Goldberg, D.T. Nash, M.D. Petraglia
(Eds.), Formation Processes in Archaeological Context, Prehistory Press, Madison, Wisconsin, 1993, pp. 165–188.
[27] I.D. Gourlay, G.W. Grime, Calcium oxalate crystals in African
acacia species and their analysis by scanning proton microprobe
(SPM), IAWA Journal 15 (1994) 137–148.
[28] B.E. Grandin, P.N. de Leeuw, M. de Souza, Labour and livestock
management, in: S. Bekure, P.N. de Leeuw, B.E. Grandin, P.J.H.
Neate (Eds.), Maasai Herding: an analysis of the livestock production system of Maasai pastoralists in eastern Kajiado District,
Kenya, ILCA Systems Study No. 4, Addis Ababa, Ethiopia, 1991,
pp. 71–82.
[29] I. Herbich, M. Dietler, Space, time, and symbolic structure
in the Luo homestead: an ethnoarchaeological study of
‘‘settlement biography’’ in Africa, in: J. Pavuk (Ed.), Actes du
XIIe Congres International des Sciences Prehistoriques, Bratislava, Czechoslovakia, September 1991, vol. 1, Archaeological
Institute of the Slovak Academy of Sciences, Nitra, 1993,
pp. 26–32.
[30] K. Homewood, W.A. Rodgers, Maasailand Ecology: Pastoralist
Development and Wildlife Conservation in Ngorongoro, Tanzania,
Series in Ecology and Resource Management, Cambridge University Press, Cambridge, 1991.
[31] P. Karkanas, O. Bar-Yosef, P. Goldberg, S. Weiner, Diagenesis in
prehistoric caves: the use of minerals that form in situ to assess the
completeness of the archaeological record, Journal of Archaeological Science 27 (2000) 915–929.
[32] S. Kent, Analyzing Activity Areas: An Ethnoarchaeological Study
of Use of Space, University of New Mexico Press, Albuquerque,
1984.
[33] S. Kent, Method and Theory for Activity Area Research,
Columbia University Press, New York, 1987.
[34] S. Kent, A cross-cultural study of segmentation, architecture, and
the use of space, in: S. Kent (Ed.), Domestic Architecture and the
Use of Space: An Interdisciplinary Cross-cultural Study, Cambridge University Press, Cambridge, 1990, pp. 127–152.
[35] S. Kent, Studying variability in the archaeological record: an
ethnoarchaeological model for distinguishing mobility patterns,
American Antiquity 57 (1992) 635–660.
[36] I. Kuijt, Life in Neolithic Farming Communities: Social
Organization, Identity, and Differentiation, Kluwer Academic/
Plenum Publishers, New York, 2002.
[37] R.I. Macphail, G.M. Cruise, M.J. Allen, J. Linderholm, P.
Reynolds, Archaeological soil and pollen analysis of experimental
floor deposits; with special reference to Butser Ancient Farm,
Hampshire, UK, Journal of Archaeological Science 31 (2004)
175–191.
[38] F. Marshall, E. Hildebrand, Cattle before crops: the origins and
spread of food production in Africa, Journal of World Prehistory
16 (2002) 99–143.
[39] W. Matthews, C.A.I. French, T. Lawrence, D.F. Cutler, M.K.
Jones, Microstratigraphic traces of site formation processes and
human activities, World Archaeology 29 (1997) 281–308.
[40] N.B. Mbae, The ethnoarchaeology of Maasai settlements and refuse disposal patterns in the Lemek area, in: P. Robertshaw (Ed.),
Early Pastoralists of South-Western Kenya, British Institute in
Eastern Africa Memoir No. 11, British Institute in Eastern Africa,
Nairobi, 1990, pp. 279–292.
[41] J. Mellaart, Catal Huyuk: A Neolithic Town in Anatolia,
MacGraw-Hill, New York, 1967.
[42] M. Merker, Die Masai, D. Reimer, Berlin, 1910.
[43] W.D. Middleton, T.D. Price, Identification of activity areas by
multi element characterization of sediments from modern and
archaeological house floors using inductively coupled plasmaatomic emission spectroscopy, Journal of Archaeological Science
23 (1996) 673–687.
[44] H. Moore, Space, Text, and Gender: An Anthropological Study
of the Marakwet of Kenya, Cambridge University Press, Cambridge, 1986.
[45] S.C. Mulholland, G. Rapp Jr., A morphological classification of
grass silica bodies, in: G. Rapp Jr., S.C. Mulholland (Eds.),
Phytolith Systematics, Plenum Press, New York, 1992, pp. 65–89.
[46] K.K. Mutundu, Ethnohistoric Archaeology of the Mukogodo in
North-central Kenya, Cambridge Monographs in African Archaeology 47 (B.A.R. International Series 775), Cambridge, 1999.
[47] J.F. O’connell, Alyawara site structure and its archaeological
implications, American Antiquity 52 (1987) 74–108.
[48] K. Odner, Excavations at Narosura: a Stone Bowl site in the
southern Kenyan highlands, Azania 7 (1972) 25–92.
[49] L.H. Robbins, Turkana material culture viewed from an
archaeological perspective, World Archaeology 5 (1973) 209–214.
[50] P.T. Robertshaw, The archaeology of an abandoned pastoralist
camp-site, South African Journal of Science 74 (1978) 29–31.
[51] P. Robertshaw, Early Pastoralists of South Western Kenya,
Memoirs of the British Institute in Eastern Africa no. 11, British
Institute in Eastern Africa, Nairobi, 1990.
[52] P. Robertshaw, F. Marshall, Ngamuriak, in: P. Robertshaw (Ed.),
Early Pastoralists of South Western Kenya, Memoirs of the
British Institute in Eastern Africa no. 11, British Institute in
Eastern Africa, Nairobi, 1990, pp. 54–72.
[53] K. Ryan, A layman’s pharmacopoeia: everyday use of medicinal
plants in Kenya’s Maasailand, The Explorers Journal 77 (1999)
8–11.
[54] K. Ryan, K. Munene, S.M. Kahinju, P.K. Kunoni, Ethnographic
perspectives on cattle management in semi-arid environments:
a case study from Maasailand, in: R.M. Blench, K.C. MacDonald
(Eds.), The Origins and Development of African Livestock:
Archaeology, Genetics, Linguistics and Ethnography, University
College London Press, London, 2000, pp. 462–477.
[55] R. Shahack-Gross, F. Marshall, S. Weiner, Geo-ethnoarchaeology of pastoral sites: the identification of livestock enclosures in
abandoned Maasai settlements, Journal of Archaeological Science
30 (2003) 439–459.
[56] A.B. Smith, K. Sadr, J. Gribble, R. Yates, Excavations in the
south western Cape, South Africa, and the archaeological identity
of prehistoric hunter–gatherers within the last 2000 years, South
African Archaeological Bulletin 46 (1991) 71–91.
[57] T. Spear, R. Waller, Being Maasai: Ethnicity and Identity in East
Africa, James Curry Ltd, London, 1993.
[58] P. Spencer, The Maasai of Matapato: A Study of Rituals of
Rebellion, Indiana University Press, Bloomington, Indiana, 1988.
[59] P. Spencer, Becoming Maasai, being in time, in: T. Spear,
W. Waller (Eds.), Being Maasai, James Curry Ltd, London, 1993,
pp. 140–156.
R. Shahack-Gross et al. / Journal of Archaeological Science 31 (2004) 1395–1411
[60] P.J.Watson, Archaeological Ethnography in Western Iran, Viking
Fund Publications in Anthropology #57, The University of
Arizona Press, Tucson, Arizona, 1979.
[61] J. Wattez, M.A. Courty, Morphology of ash of some plant
materials, in: N. Fedoroff, L.M. Bresson, M.A. Courty (Eds.),
Soil Micromorphology, AFES, Plaisir, 1987, pp. 677–683.
1411
[62] D. Western, T. Dunne, Environmental aspects of settlement site
decisions among pastoral Maasai, Human Ecology 7 (1979)
75–98.
[63] J.E. Yellen, Archaeological Approaches to the Present:
Models for Reconstructing the Past, Academic Press, New
York, 1977.