MANUSCRIT ACCEPTAT

MANUSCRIT ACCEPTAT
An ethnoarchaeological study of livestock dung
fuels from cooking installations in northern
Tunisia
Marta Portillo ; M. Carme Belarte ; Joan Ramon ; Nabil Kallala
; Joan Sanmartí Rosa M. Albert
Revista

Quaternary International. Volume 431, Part A, 28 February 2017, Pages
131-144

DOI

https://doi.org/10.1016/j.quaint.2015.12.040

Disponible
02/02/2016
en línia

Data de publicació

28/02/2017

Per citar aquest document:
Marta Portillo, M. Carme Belarte, Joan Ramon, Nabil Kallala, Joan Sanmartí, Rosa M. Albert,
An ethnoarchaeological study of livestock dung fuels from cooking installations in
northern Tunisia, In Quaternary International, Volume 431, Part A, 2017, Pages 131-144,
ISSN 1040-6182, https://doi.org/10.1016/j.quaint.2015.12.040.
Aquest arxiu PDF conté el manuscrit acceptat per a la seva publicació.

1

MANUSCRIT ACCEPTAT

Abstract
Livestock dung is a valuable material in many rural communities worldwide. In our
research area, the site of Althiburos and its surroundings, now el Médéïna, in
northwestern Tunisia, dung is the main source of fuel for domestic purposes, primarily the
processing and cooking of foods. Ovicaprine dung is daily used in traditional mud tannur
type ovens, namely tabouna. The archaeological record shows that mud constructed
cooking installations were common during the first millennium BC. Previous studies of
phytoliths and dung spherulites at Numidian Althiburos suggested the use of vegetal and
fecal matter for fuel purposes. We present here the results of the continuation study based
on the comparison between archaeological results (a selection of cooking installations, six
hearths and two mud ovens) and those obtained from the ethnographic study of dung fuel
materials from the site area. The present study builds up on ethnographic observations
and informal interviews (dung collection, management, storage, waste disposal and
cooking and baking activities), temperature measurements within the burning fuel, as well
as modern material sampling (fresh dung, burned pellets, dung ashes and fuel trash paths)
which was followed by integrated studies of phytoliths and calcitic microfossil analyses
(dung spherulites and wood ash pseudomorphs) for comparative purposes. The results
obtained provided direct evidence regarding the type of fuel sources: dung, wood and a
mixing of dung and vegetal matter (wood and agricultural by-products). Dung was used as
source of fuel material across time (from the Early Numidian occupation phase, 10th–9th
century BC, to the last centuries BC) and space (in different excavation areas and type of
installations). Such integrated studies demonstrate the value of combining different
microarchaeological techniques and the use of ethnoarchaeological material from site
areas.

Keywords
Numidians ; Fuel ; Cooking ; Phytoliths ; Dung spherulites ; Ash pseudomorphs

2

MANUSCRIT ACCEPTAT

1. Introduction
It is well known that livestock dung is a valuable source of fuel in many rural
communities across the world (Miller, 1984; Anderson and Ertug-Yaras, 1998;
Reddy, 1998; Sillar, 2000). Among traditional societies dung is an important
secondary product that can be used in varied range of ways (i.e. fertilizer either in
its organic form or after being burned, building material, container making, etc.), in
addition to fuel for domestic uses. The use of dung as fuel has been commonly
related to situations of deforestation, in areas where the woody vegetation is
sparse or not available (Bottema, 1984; Miller, 1984; Anderson and Ertug-Yaras,
1998). However, ethnographic studies on fuel selection show how it may relate not
only to availability and economic factors (i.e. cost, efficiency and redundancy, as
this material may be needed for other main purposes such as manure), but also by
other cultural choices or social values (i.e. taboo on using specific materials)
(Shahack-Gross et al., 2004; Picornell et al., 2011). Ethnographic research in the
Moroccan Rif, an area in western Maghreb where Mediterranean woodlands are
still abundant, illustrates that despite wood has traditionally been the main source
of fuel for domestic purposes, the use of dung cakes (primarily mixed cow and goat
dung) is linked to the firing of pottery (Zapata et al., 2003). In this later example
the use of dung cakes responds to technical reasons related to their burning
properties (slower and more regular than wood), and shows a limited
consumption, as only a partial amount of the household dung production is used
for firing and most of the production is devoted to other main purposes, such as
manure and tempering.
In our research area, the site of Althiburos and its surroundings, located in a small
fluvial valley on the northern edge of the Ksour massif, in northwestern Tunisia
(Fig. 1), livestock dung is the main source of fuel for domestic use, primarily the
processing and cooking of foods, in addition to ceramic production. The area,
where firewood is sparse, is particularly interesting because the present-day rural
communities use ovicaprine dung as the main fuel source, including daily cooking
and baking in mud cylindrical tannur type ovens, locally called tabouna (Portillo
and Albert, 2011).

3

MANUSCRIT ACCEPTAT

1.
2.

Download high-res image (455KB)
Download full-size image

Fig. 1. Map showing the general location of Althiburos, in northern Tunisia.

A variety of archaeological domestic cooking installations from all time periods
have been studied in the Mediterranean region. Many studies have focused mainly
in prehistoric and protohistoric ovens and hearths, based on macroscopic
descriptions and analogies to ethnographic parallels. A number of recent studies
have explored such fire installations using microarchaeological techniques
(Weiner, 2010 and references therein), thus emphasizing on the identification of
the mineralogical signatures of heating associated to the installations, and taking
into account formation and degradation processes that are critical for interpreting
the archaeological record (Albert et al., 2000, 2003; Albert and Cabanes, 2007;
Berna et al., 2007; Gur-Arieh et al., 2012, 2014). Possible depositional route-ways
and taphonomic histories for such firing contexts may relate to fuels used, foods
accidentally burnt during preparation and cooking, as well as materials
accidentally or deliberately discarded into the fire (i.e. destroying infested seeds)
(Van der Veen, 2007; Matthews, 2010). In addition, secondary depositional
pathways and re-use for storage or trash may introduce burnt and un-burnt
materials that are unrelated to the original function of the installations (i.e. fill
deposits, inclusion of building materials).
Mud constructed cooking installations are also common in Numidian
archaeological contexts belonging to the first millennium BC (Belarte, 2011;
Ramon and Maraoui, 2011). In previous studies at Numidian Althiburos we have
addressed plant and dung exploitation, including food processing activities,
through the combined use of opal phytoliths– microscopic bodies composed of
pure silica present in the tissues of many vegetal species (Piperno, 1988, 2006;
4

MANUSCRIT ACCEPTAT

Pearsall, 1989) and dung spherulites– calcitic particles that form in animal guts
and can be found in dung (Brochier et al., 1992; Canti, 1997, 1998, 1999). Livestock
dung assemblages have been examined through such microfossils in varied
contexts belonging to different occupational phases of the site (covering most of
the 1st millennium BC), including room floors and midden deposits, but also in
combustion structures such as hearths and ovens (Portillo and Albert, 2011, 2016;
Portillo et al., 2012). The aim of this previous research was to evaluate the
potential use of plant and dung microfossils, in integration with other
bioarchaeological evidence from faunal and charred plant macro-remains, for
delineating domestic activities being carried out at the site, including food
processing and grazing and foddering of herds. The presence of phytolith-rich
layers in association with large amounts of fecal spherulites in certain areas of the
site suggested that plants were deposited onsite as livestock dung or dungproducts. Thus, such microfossil associations in combustion structures suggested
the use of plant and dung material for fuel purposes. Additionally, we carried out a
pilot ethnoarchaeological study in order to better understand the manner in which
both vegetal and dung microfossils were embedded in the archeological
assemblages and, more widely, their role in site formation processes (Portillo et al.,
2012). For this purpose, the research included the study of selected modern dung
materials obtained from the surroundings of the site, such as fresh dung pellets,
penning soils and dung sub-products (i.e. building material and burned dung
pellets from a tabouna oven). Sampling of modern materials, as well as laboratory
extraction procedures, was conducted in a manner similar to that of archaeological
contexts for comparative purposes.
The research reported upon here represents the continuation of these earlier
studies focusing on fuel exploitation and food processing in cooking installations
using similar integrated microarchaeological and ethnoarchaeological approaches.
This study examines phytolith and calcitic microfossil assemblages from a
selection of archaeological installations, primarily hearths and mud ovens, in
addition to modern fuel materials from the site area. The current study includes
the analyses of calcitic ash pseudomorphs (calcite pseudomorphs after calcium
oxalate crystals' heating to at least 450 °C, primarily originating from wood and
dicotyledonous leaves) (Wattez and Courty, 1987; Brochier and Thinon, 2003;
Canti, 2003). For this purpose, a pilot reference collection from selected modern
woody specimens is also presented here. Thus, following the recently developed
methods of Gur-Arieh et al. (2013, 2014), based on the calculation of the ratio of
ash pseudomorphs to dung spherulites, we intend to discriminate between wood
versus dung-dominated ashes.
The present work addresses the taphonomy of all three microfossil lines of
evidence (phytoliths, spherulites and ash pseudomorphs) based on the comparison
between archaeological results and those obtained from the ethnographic study of
dung fuel materials from the site area. Fieldwork included ethnographic
observations and informal interviews (dung collection, management, storage,
waste disposal and cooking and baking activities), in addition to time
measurements from the initial fire lighting and the end of cooking and
temperature measurements within the burning tabouna fuels. The later issue,
which has not been previously addressed in our research, is critical for evaluating
5

MANUSCRIT ACCEPTAT

taphonomic questions related with microfossil preservation (i.e. melting of
phytoliths, the impact on ash pseudomorphs and dung spherulites total
abundances). Our ethnographic work treated the modern installations and
materials as if they were archeological. The comparison of these new reference
modern materials to the results of the archaeological material from this study, as
well as from previous studies of such archaeological installations (hearths and
ovens), will provide the opportunity of characterizing the type of fuel matter and
food processing behaviors. Additionally, it may allow us to assess whether there is
continuity or change of the tradition in the use of fuel sources across time (from
the Early Numidian occupation phase, 10th–9th century BC, to the last centuries
BC), as well as in different site areas and types of firing installations (Fig. 2).

1.
2.

Download high-res image (747KB)
Download full-size image

Fig. 2. General plan of the site showing the locations of the excavation areas (zone 2, northern edge
of the capitolium).

6

MANUSCRIT ACCEPTAT
1.1. Research area

The ethnoarchaeological study was conducted at the site area, now el Médéïna
(meaning the “small town”), located in the High Tell region (Governorate of El Kef),
less than 50 km from the Algerian border (Fig. 1). The High Tell region is
characterized by a succession of alluvial plains, valleys and plateaux, with an
average altitude of 700 m.a.s.l. At present the landscape is open due to
deforestation. The area receives ca. 400–600 mm of annual rain on average, with
high annual variability (El Kef rainfall station), and is situated within the
continental Mediterranean climate belt, with wet winters and dry summers
(average temperatures around 7.3 °C in winter and 26.5 °C in summer) (Kallala
and Sanmartí, 2011). The economic production of the area is based on cereal
agriculture, although small permanent rivers in the valley also allow horticultural
crops, in addition to livestock farming of ovicaprines and cattle. Rain-fed arable
crops include primarily free-threshing wheat and barley, and yields are subject to
high year-to-year variation (Latiri et al., 2010).
Our research was carried out in two small villages, El Souidat and Gouasdya (about
10 km from Dahmani), both inhabited by families of the Ouarten tribe that
maintain a certain traditional way of life in many aspects. Ouarten families subsist
on farming, mostly sheep and goats, and cows to a lesser extent, in addition to
cereal agriculture. Their diet is based on cereals prepared in a wide range of ways,
and commonly accompanied by legumes and vegetables, whereas meat and eggs
are consumed to a lesser extent. Traditional households commonly comprise a
main house for a nuclear family and separate houses for married sons and their
families. Households include central courtyards where varied activities are carried
out daily, as well as small vegetable back-gardens, cooking and storage
installations, and livestock pens (Fig. 3). Electricity was only introduced recently,
and they do not benefit from running water.

7

MANUSCRIT ACCEPTAT

1.
2.

Download high-res image (1MB)
Download full-size image

Fig. 3. Livestock dung management, storage and fuel use at El Souidat, June 2010. a) Household
surrounded by open-air dung drying spaces. Arrows point at the location of main penning areas.
The inset (b) shows a detail of swept areas. c) Livestock dung storage structures, locally called
Kamur, close to drying spaces. d–e) Tabouna ovens usually located out-doors close to penning
areas; note the placement of a tabouna in use within a pen in the same household (e).

1.2. The site

Althiburos is a multi-period urban settlement site, mainly known as a Roman city.
It is located on a relatively large plateau (about 28 ha) surrounded by limestone
hills and defined by the courses of the Wadi el Médéïna and the Sidi Baraket
streams (Fig. 1). Recent excavations close to the capitolium have revealed a
complete occupation sequence dating from the Early Numidian phase (10th–8th
century cal BC) to the Middle Ages. Ongoing research focuses on the formation and
development of Numidan states, about which little is known archaeologically
(Kallala and Sanmartí, 2011; Sanmartí et al., 2012). Ancient Numidians followed
economic strategies based on mixed farming practices which integrated cereal
agriculture, wheat (Triticum aestivum/durum) and barley (Hordeum vulgare), and
to a lesser extent emmer wheat (T. dicoccum) and common millet (Panicum
miliaceum), legumes including lentils (Lens culinaris) and peas (Pisum sativum),
and fruits such as vine (Vitis vinifera) and figs (Ficus carica), in addition to
husbandry, sheep (Ovis aries), goats (Capra hircus), cattle (Bos taurus), and pig (Sus
domesticus) (Kallala and Sanmartí, 2011; López-Reyes and Cantero, 2016;
Valenzuela, 2016).

8

MANUSCRIT ACCEPTAT

Two main types of cooking installations have been reported in the field as hearths
and ovens. Such broader categories show many variants in size and shape.
Interestingly, some of these features were cylindrically-shaped and described as
mud ovens similar to modern tabounas (Kallala et al., 2008; Ramon and Maraoui,
2011).

2. Materials and methods
2.1. Ethnoarchaeological fieldwork and modern materials

Fieldwork was carried out during the summers of 2008, 2010 and 2014 (periods of
around 7–10 days per year) and included ethnographic observations, informal
interviews and sampling materials from rural communities from site vicinity, in
which traditional non-mechanized farming practices on the verge of extinction are
still common. This study builds on previous work on livestock dung materials,
including burned sheep/goat pellets from a tabouna oven from Gouasdya (Portillo
et al., 2012). The present work enlarges the study of dung exploitation,
management and storage, as well as waste disposal and cooking and baking
activities with the sampling of new fuel materials from El Souidat. The samples
examined comprised ovicaprine fresh dung from a pen in the same household,
non-burned mixed fuel, burned dung pellets within a tabouna oven, fuel ashes
from its bottom part, and fuel trash residues stored in a metal bin (Table 2).
Additionally, we conducted time and temperature measurements in the same
tabouna installation (El Souidat, June 2014, Fig. 4). Temperature along the cooking
activity, in this case bread (Khobz) baking, was recorded every few minutes using a
portable digital thermometer (Dostman Serie P615) equipped with two detectors:
one was placed within the burning dung fuel at the bottom; the second as close as
possible to the oven's wall, in order to record the baking temperature (Figs. 4a and
5). Note that a few branches of woody shrubs growing in the same location were
used for lighting the fire.
Table 1. Location and field descriptions of sediment samples from ovens (FR),
hearths (FY) and in situ vase (VP), including control samples (C) from their vicinity
(Early Numidian sub-phases EN2: 9th century BC and EN3: 8th–early 7th century
BC; Late Numidian sub-phases LN1: 4th century–146 BC and LN2: 146–27 BC).
Feature Sector Phase

UnitSample

Description

2641

2692
FR2640

4b

Ashy black upper fill with abundant charcoal, varied seeds,
faunal remains and pottery
Soft ashy oven fill with abundant charcoal, seeds, bones and
ceramic fragments

2712

Soft ashy sediment with abundant charcoal fragments,
bottom oven fill

LN2

2670-C

Yellowish compacted sediment with lime nodules and

9

MANUSCRIT ACCEPTAT

Feature Sector Phase

UnitSample

Description
pottery, SL 6070, control FR2640

270224
270219-C

290416

Compacted burnt clay material and ceramic fragments
Compacted burnt clay material, about 5 cm thick
Compact gray–green sediments with ashes, varied pottery
and animal bones, control FY290411
Ashy black ceramic vase fill with abundant charcoal
fragments and seeds

2072-C

7d

Ashy black thick layer with ceramic fragments, control
FY290401 and FY290416

2075
VP2069

Ashy black sediment with abundant charcoal and pottery

290409-C

FY290411 3–4

Burnt stony and clay material with abundant charcoal, faunal
remains and pottery

290411

FY290416 3–4

280208

290412-C

FY290401 3–4

Light brown layer with lime nodules and clay material,
control FY270355

290401

FY280208 3–4

Ashy black sediment composed of charcoal

270353-C

FY270355 8a

Compact gray–green sediments with ashes, varied pottery
and animal bones, SL2708, control FY2710

270355

2d

Ashy black sediment with abundant charcoal fragments

2708-C

FY2710

Soft black pit fill with abundant charcoal, ash and seeds,
control FR270223

2710

FR270223 7a

Oven fill containing tabouna pieces and charred seeds

Soft black pit fill containing an in situ ceramic vase,
composed of abundant charcoal, ash and seeds, control
VP2069

LN1

EN3

LN1

EN3

EN2

EN2

EN2

LN1

Table 2. Provenance, description of samples and main phytolith, ash pseudomorph
and dung spherulite results obtained from modern fuel samples.

10

MANUSCRIT ACCEPTAT
Ash
Spheruli
Phytoli
AI
Phytolit Multicel pseudomo tes 1 g
Sample
ths 1 g Grass
Ratio ash
F
hs
led
rphs 1 g of of ashed
descripti
of AIF phytoli
pseudomorphs/sph
(%
weather phytolit
ashed
material
on
(millio ths (%)
erulites
)
ing (%) hs (%) material (million
n)
(million)
)
Sheep/go
at fresh
dung
41.
135.9
from
1
livestock
penning

91.3

5

32.5

0.5

529

0.001

Nonburned
mixed
ovicaprin
e
dung/wo 32 100.6
ody
(Marrubi
um)
tabouna
fuel

90.3

4.3

22.3

0.15

338

0.005

Sheep/go
at fuel
from a
tabouna 39.
62.7
oven,
1
close to
upper
wall

84.6

8.7

14.1

0.49

314

0.002

Sheep/go
at fuel
from a
37.
tabouna
119.3
7
oven,
bottom
part

80.8

12.9

29.5

0.95

164

0.006

Trash
burned
fuel
residues 42 138.1
stored in
a bin,
close to
tabouna

84.8

9.1

16.7

0.59

153

0.004

11

MANUSCRIT ACCEPTAT
Ash
Spheruli
Phytoli
AI
Phytolit Multicel pseudomo tes 1 g
Sample
ths 1 g Grass
Ratio ash
F
hs
led
rphs 1 g of of ashed
descripti
of AIF phytoli
pseudomorphs/sph
(%
weather phytolit
ashed
material
on
(millio ths (%)
erulites
)
ing (%) hs (%) material (million
n)
(million)
)
oven

1.
2.

Download high-res image (864KB)
Download full-size image

Fig. 4. a) Tabouna baking measurements, El Souidat, June 2014. Note the placement of the
thermometer detectors (arrows): one within the burning fuel at the bottom ventilation hole, and
the second as close as possible to the inner oven's wall in order to record the baking temperatures.
b) A metal bin close to the oven containing a mixing of non-burned fuel (primarily ovicaprine dung,
left) and the same type of burned-fuel material (right) re-used in the same installation. c) Initial
stages of firing. d) Bread cakes (Khobz) being baked. Also note the placement of the thermometer
detector (arrow).

12

MANUSCRIT ACCEPTAT

1.
2.

Download high-res image (298KB)
Download full-size image

Fig. 5. Histogram showing the temperatures measured during the Tabouna baking using two
detectors, within the burning fuel at the bottom and close to the oven's inner wall (temperatures in
C° and time expressed in hh:mm:ss). Note that the baking started 30 min after the lighting and only
took 5 min for baking 12 Khobz cakes.

2.2. Archaeological materials

Eighteen archaeological samples were selected for phytolith and calcitic
microfossil studies (Table 1). The contexts investigated here belong to the
excavation area located in the northern edge of the capitolium (zone 2, Fig. 2) and
to different occupation phases, from the Early Numidian (sub-phases EN2: 9th
century BC, and EN 3: 8th–early 7th century BC) to the last centuries BC (Late
Numidian phases, LN1: 4th–mid 2nd century BC, and LN2: mid 2nd–1st century
BC), thus offering the opportunity to learn more about fuel use and food processing
activities through time. These contexts were described in the field as combustion
fillings and fire installations corresponding to six well-defined hearths and two
ovens (Belarte, 2011; Ramon and Maraoui, 2011). Eight control sediment samples
from their vicinity were also analyzed for comparative purposes. Samples were
taken during the 2007, 2008, 2009 and 2014 excavation seasons.
The hearths examined differed in diameter and shapes (round/ovoid). They were
built from alluvial pebble layers, or had a foundation surface of compacted clay
material. All six hearths were covered by gray ashy sediments that included wood
charcoal and charred seeds (Belarte, 2011; Ramon and Maraoui, 2011) (Table 1,
Fig. 6a–b). Like the ethnographic installations, ovens were cylindrically-shaped.
Oven FR270223 (Late Numidian sub-phase, LN1) was described as a mud oven
similar to modern tabounas (Ramon and Maraoui, 2011) (Table 1, Fig. 6c–d). Gray
and/or white ash was found, with abundant macroscopic remains from charcoal,
13

MANUSCRIT ACCEPTAT

seeds and pottery fragments in their upper fillings, as well as concentrations of
charred botanical remains at their bottom. A similar macroscopic composition was
also attested in the filling of an in situ ceramic vessel found in an indoor domestic
space that was interpreted by the excavators as a storage area (VP2069, sub-phase
LN1) (Fig. 7). The latter is included here for the sake of comparison; it may relate
to dumping episodes from nearby installations, and/or perhaps the storage of fuel
remains, as observed in traditional households in the site area today.

1.
2.

Download high-res image (1MB)
Download full-size image

Fig. 6. a–b) Hearth FY290411, EN2; c–d) Mud oven FR270223, LN1. Extracted from Ramon and
Maraoui, 2011: a–b) Figs. 4.58 and 4.59, pp. 224, a–b; c–d) Figs. 4.136 and 4.138, pp. 243.

14

MANUSCRIT ACCEPTAT
1.
2.

Download high-res image (467KB)
Download full-size image

Fig. 7. In situ vase VP2069, LN1.

2.3. Phytolith analyses

Phytolith extraction and quantitative analyses from archaeological and modern
samples followed the methods of Albert et al. (1999). A weighed aliquot of 1 g of
dried sediment was treated with an equivolume solution of 3 N HCl and 3 N HNO3.
Organic matter was oxidized with 33% H2O2. Phytoliths were concentrated using
2.4 g/ml sodium polytungstate solution [Na6(H2W12O40)·H2O]. Microscope slides
were mounted with 1 mg of dried sample using Entellan New (Merck). A minimum
of 200 phytoliths with diagnostic morphologies were counted at 400×. Phytolith
quantification was based on the acid insoluble fraction (AIF), which is the fraction
that remains after the acid and peroxide treatment. AIF allows comparisons
between samples independently of the diagenesis suffered by the sediment.
Phytoliths that were unidentifiable because of dissolution were counted and
recorded as weathered morphotypes. Morphological identification was based on
modern plant reference collections from the Mediterranean area (Albert and
Weiner, 2001; Tsartsidou et al., 2007; Albert et al., 2008, 2011; Portillo et al., 2014)
and standard literature (Twiss et al., 1969; Brown, 1984; Piperno, 1988, 2006;
Mulholland and Rapp, 1992; Rosen, 1992; Twiss, 1992). The terms used follow the
International Code for Phytolith Nomenclature (Madella et al., 2005). Slides were
examined using an Olympus BX41 optical microscope and digital images were
taken with an Olympus Color View Ilu camera.
2.4. Calcitic microfossils: dung spherulite and ash pseudomorph analyses

The methods used are similar to those developed by Canti (1999). An accurately
weighed amount of between 0.5 and 1 mg of dried sediment was placed on a
microscope slide using a pipette. Slides were mounted using Entellan, as described
above for phytoliths. Note that if the sediment is rich in clays or aggregates, they
can be dispersed by sonication (Gur-Arieh et al., 2013). This was not necessary in
the present study. Slides were examined at 400× magnification under the optical
microscope with crossed polarized light (XPL) for spherulites, whereas
pseudomorphs where examined in plane polarized light (PPL). Both microfossil
numbers found in a known number of randomly chosen fields were recorded and
then related to the initial sample weight. The initial weight in modern samples
should be around 0.5 mg in order to avoid microfossil overloading.
The data of the ash pseudomorph and spherulite quantification were used to
calculate ratio values, which may allow the distinction between wood and dungdominated ashes (Gur-Arieh et al., 2013, 2014). Additionally, ash pseudomorphs
were compared to a pilot modern plant reference collection (Table 3, Fig. 8a–b–c).
The selected species includes Aleppo pine (Pinus halepensis) and olive tree (Olea
europaea), which were common in the Numidian layers' macro-botanical record
(Kallala and Sanmartí, 2011; Cantero, 2016; López-Reyes and Cantero, 2016).
Samples were processed by burning small dried branches in an oven furnace (at
550 °C for 4 h), after washing in deionized water and sonication for 10 min to
15

MANUSCRIT ACCEPTAT

remove external contamination. The estimated pseudomorph numbers is based in
abundances per weight of ashed plant material.
Table 3. Provenance, description of samples and quantities of ash pseudomorphs
obtained from modern plant species.
Species

Family

Ash pseudomorphs 1 g of ashed
material (million)

Provenance

Juniperus
oxycedrus

Cupressaceae

Oued
Mellègue

166

Marrubium sp.

Lamiaceae

Souidat

171

Olea europaea

Oleaceae

Oued
Mellègue

226

Pinus halepensis Pinaceae

Oued
Mellègue

502

Pistacia lentiscus Anacardiaceae

Oued
Mellègue

569

1.
2.

Download high-res image (812KB)
Download full-size image

Fig. 8. Photomicrographs of phytoliths and calcitic microfossils identified in modern samples. The
photographs have been taken at 400×. (a–b–c) ash pseudomorphs obtained from small branches,
Oued Mellègue, NO Le Kef, June 2014: a) Aleppo pine (Pinus halepensis), b) mastic (Pistacia
lentiscus), c) olive tree (Olea europaea); (d–e–f) microfossils obtained from modern fuel samples, El
Souidat, June 2014: d) dung spherulites, e) multicellular structure of dendritic long cells with
papillae from the husks of Hordeum sp., f) melted multicelled phytoliths from grasses.
16

MANUSCRIT ACCEPTAT
Archaeological samples were compared to the modern fuel materials listed in Table 2. Modern
dried dung pellets were also ashed in laboratory-controlled conditions (at 500 °C for 4 h using a
muffle furnace). After ashing, all three microfossils (phytoliths and calcitic spherulites and
pseudomorphs) from modern dung samples were treated and examined following the abovedescribed methodology.

3. Results
The main results are described separately below.
3.1. Ethnographic observations

Traditional means of cooking and baking among Ouarten families are mud
cylindrical tabouna ovens. These daily activities are exclusively the domain of
women. The tabouna is usually built with soil mixed with organic materials
(primarily chaff, agricultural by-products and dung). Most households have their
own out-doors ovens for daily bread production. In addition to bread cakes
(Khobz) preparation, these installations are used for cereal processing, including
de-husking hulled barley by roasting. The main source of fuel is air-dried
ovicaprine dung pellets.
Livestock penning supply a regular and predictable production of dung, which can
be exploited by these communities in many ways, including fuel, in addition to
manure and building material. In summer, herds graze on cereal stubble in fallow
agricultural fields, and on wild vegetation in stream margins. Dung is left to dry in
open-door spaces for a few days, usually close to penning areas, and once dried it
is stored in specific stone-made installations, locally called kamur, primarily for
winter fuel purposes (Fig. 3a-b-c). These materials are used for firing local pottery,
which is female work as well. The sub-products of different domestic structures
and various firing episodes can be re-used, and then discarded in trashing areas or
used as manure in household gardens. Fresh dung pellets from cattle, but also from
donkeys– depending on the wealth of the family, or on exceptional periods when
such animals are needed (i.e. threshing crops) are used for plastering walls and
floors. Threshing floors are also tempered during the summer with a mixture of
fresh dung, chaff and water. Similar patterns concerning livestock dung
exploitation have been reported in ethnographic studies conducted in western
Maghreb (i.e. drying and storage, firing pottery, plastering walls and floors,
container making, fertilizing) (Zapata et al., 2003; Peña-Chocarro et al., 2009,
2015).
3.2. Burning experiments and modern fuel materials

Fig. 5 shows temperature and time measurements obtained from the tabouna
bread (Khobz) baking. The woody material used for lighting the fire produces a fast
increase of temperature (up to 800 °C in a few minutes). Then, it stays relatively
steady between 500 and 600 °C within the burning dung fuel placed in the bottom,
and decreases to around 400 °C during the cooking time. The baking starts 30 min
after the lighting, and it takes only 5 min to prepare 12 bread cakes. The bread is
baked by placing Khobz cakes on the oven's inner walls (Fig. 4d). Temperatures
measured within the oven's wall decreased much faster, and were around 100–
17

MANUSCRIT ACCEPTAT

150 °C during the cooking process. Thus, temperatures were measured also after
the baking was completed. Fuel within the oven's bottom was about 200 °C 2 h
later. This data is consistent with previous experiments conducted in Uzbekistan
by Gur-Arieh et al. (2013), on mud traditional installations using different fuel
materials (wood, mixed cow/sheep-goat dung cakes and mixed wood/dung).
Similar observations have been made on tannur baking measurements within the
burning fuel in the Upper Khabur (northeastern Syria), where the main fuel source
is woody material from cotton production, while ovicaprine dung is only residually
used (Portillo et al., 2014).
Samples obtained from the same tabouna installation and fuel materials, including
ovicaprine fresh dung, non-burned mixed wood material for the initial fire lighting,
burned pellets, dung ashes and trash paths, were also examined (Table 2).
Concentrations of all three types of micro-remains (phytoliths, spherulites and ash
pseudomorphs) vary significantly between burned and non-burned fuel materials.
Phytoliths were abundantly identified in all samples (over 100 million phytoliths
in 1 g of AIF in most of the samples, Table 2). Most of these microfossils belonged
to a livestock early summer grass-rich diet (agricultural by-products, primarily
from barley, Fig. 8e). In general, phytoliths were well preserved and noted in
anatomical connection (multi-celled phytoliths constitute between 14 and 32%),
but with a higher dissolution degree in burned samples (phytoliths weathering
reaching around 13% in ashes form the oven bottom). Phytoliths showing
evidence for partial melting, resulting in deformations due to high temperatures
(although the morphology of most original cells may be preserved and therefore
identified, Fig. 8f) were also observed. These were common in samples from the
oven's bottom and from the stored burned fuel residues located in a metal bin for
re-use purposes (Fig. 4b). As expected, strong differences were noted between the
calcitic microfossil abundances. Accordingly to previous studies showing that
spherulite concentrations are particularly high in fresh dung ovicaprine pellets
(over 500 million spherulites per gram of ashed dung, Table 2), much smaller
amounts were found in ashy fuel remains (about 300 million in the upper part, and
around 100 million spherulites 1 g of sediment in the bottom and stored remains
for re-use). It should be noted that dung spherulites are well preserved at
temperatures lesser than ∼650–700 °C (Matthews, 2010; Shahack-Gross, 2011). In
contrast, all samples contained low amounts of pseudomorphs (related both to the
animal's diet– barley chaffs, and the dung-dominate fuel type). Pseudomorphs
included druses and prisms, as well as rhombs to a lesser extent. The calculation of
the pseudomorphs/spherulites ratios showed distinctive low values (around 0,
Table 2), which in conjunction with large phytolith concentrations are
characteristic of rich-dung ashes, according to quantitative models proposed by
Gur-Arieh et al. (2013, 2014).
Additionally, modern samples from small branches of selected woody species were
also analyzed for comparative purposes. Table 3 details all the analyzed woody
species according to their different families and provenance, with indication of the
number of ash pseudomorphs per gram of ashed organic material. Ash
pseudomorph abundances were relatively high in most of the samples (over one
hundred million pseudomorphs per 1 g of ashed material). Pseudomorphs
included prismatic crystals and occasionally druses; and, significantly, no raphides
18

MANUSCRIT ACCEPTAT

or styloids, which are commonly produced by monocotyledonous plants. The
results of this pilot study are also in agreement with published data (Wattez and
Courty, 1987; Brochier, 1996; Brochier and Thinon, 2003; Canti, 2003).
Interestingly, fresh ashes of Aleppo pine (Pinus halepensis) showed abundant
characteristic large and regular-shaped rhomboids (ranging 15–50 μm long and 2–
5 μm width, Fig. 8a). The presence of such characteristic elongated prisms in Pinus
has been previously reported (Brochier and Thinon, 2003, and references therein).
These morphologies were also commonly observed in our archaeological samples.
It should be noted that Aleppo pine is the most common woody species in the
Numidian occupation of the site (Cantero, 2016). These rhomboids were larger
than those noted in Juniperus oxycedrus and Pistacia lentiscus (ranging 10–25 μm
long), but also in Olea europaea (around 10–35 μm long and 3 μm width, Fig. 8b–c).
These later genera are also present in the archaeological charred assemblages.
3.3. Archaeological results

Table 4 shows the location and field descriptions of the samples analyzed, together
with the quantitative phytolith and calcitic microfossil results (ovens– FR, hearths–
FY, in situ vessel–VP and corresponding control samples–C). The mineralogical
results, which are expressed as the acid insoluble fraction, showed a similar
mineralogical distribution (AIF average around 26%, Table 4), indicating that
carbonates, phosphates and other non-siliceous materials were major components
of the sediments examined. Phytoliths, ash pseudomorphs and dung spherulites
were noted in different amounts in all the samples, with two exceptions, control
samples 2670-C and 290412-C (from the vicinity of two installations, FR and FY
respectively), where calcitic microfossils were scarce or even absent.
Table 4. Location and main phytolith, ash pseudomorph and dung spherulite
results obtained from ovens (FR), hearths (FY) and in situ vase (VP) sediments,
including control samples (C) from their vicinity.
AI
Phytolit Multic
Spherul
Grass
Ash
Unit- F Phytoli
hs
ell
ites 1 g
Ratio ash
Featur
phytoli
pseudomo
samp ( ths 1 g
weathe phytoli
of
pseudomorphs/sp
e
ths
rphs 1 g of
le % of AIF
ring
ths
sedime
herulites
(%)
sediment
)
(%)
(%)
nt
2641

23. 28.900.
82.2
1 000

12.2

14.2

520.000

2.080.00
0.25
0

2692

28. 55.600.
81.9
9 000

10.5

17.9

340.000

332.000 1.02

2712

32. 38.300.
78.8
3 000

14.1

18

2.290.000

306.000 7.48

2670- 14. 470.00
73.2
C
8 0

16.1

5.1

0

0

FR264
0

0

19

MANUSCRIT ACCEPTAT
AI
Phytolit Multic
Spherul
Grass
Ash
Unit- F Phytoli
hs
ell
ites 1 g
Ratio ash
Featur
phytoli
pseudomo
samp ( ths 1 g
weathe phytoli
of
pseudomorphs/sp
e
ths
rphs 1 g of
le % of AIF
ring
ths
sedime
herulites
(%)
sediment
)
(%)
(%)
nt
2702 26. 27.100.
85.7
24
4 000

6.1

4.7

191.000

58.000

3.29

2702 29. 50.500.
83.8
19-C 5 000

8.3

23

40.000

13.000

3.08

8

2.3

130.000

69.000

1.88

2708- 29. 22.000.
72.6
C
7 000

17.9

8.4

322.000

610.000 0.53

2703 30. 53.500.
89.6
55
7 000

6.1

16.7

272.000

148.000 1.84

2703
6.300.0
24
88.6
53-C
00

4.9

7.2

18.000

54.000

FY280 2802 20. 5.700.0
82.5
208
08
9 00

9.8

15.4

111.000

230.000 0.48

2904 33. 1.100.0
80
01
4 00

9.9

2.1

193.000

331.000 0.58

2904 21. 560.00
85.6
12-C 7 0

7

9.5

0

36.000

0

16.3

0

20.000

99.000

0.20

2904 30. 781.00
65.8
11
9 0

19

0

52.000

350.000 0.15

2904
17.800.
29
81.5
09-C
000

10.3

4.3

78.000

410.000 0.19

8

8.9

85.000

2.900.00
0.03
0

17.5

8.2

2.700.000

496.000 3.08

FR270
223

2710
FY271
0

FY270
355

FY290
401

22. 25.000.
85.3
3 000

FY290 2904 28. 397.00
71
416
16
2 0

FY290
411

2075
VP206
9

36. 71.100.
83.7
1 000

2072- 19. 27.300.
74.1
C
6 000

0.33

Phytoliths were abundant in most of the samples (from 1 to 71 million phytoliths
per gram of AIF, Table 4). The only exceptions are again control samples 2670-C
20

MANUSCRIT ACCEPTAT

and 290412-C, in addition to hearths FY290411 and FY290416, which yielded a
lesser amount (400–700.000 phytoliths/1 g of AIF). These latter hearth samples,
belonging to the Early Numidian phase, were composed of compacted burnt clay
material and also showed a higher dissolution index of phytoliths (16–19%).
Partially melted phytoliths, which were common in our modern tabouna fuel
ashes, were also observed in the bottom filling oven sample 2712 (oven FR2640).
The morphological results indicated that grasses dominated in most of the samples
with around 70–80% of all the counted morphotypes (Table 4), whereas wood,
bark and dicotyledonous leaf phytoliths constitute around 5–10% of the
assemblages. Multicellular structures– multi-celled or interconnected phytoliths,
from both floral parts of cereals, primarily wheat and barley (Fig. 9a), and the
leaves and the stems of grasses (Fig. 9b–c), were also noted in most of the samples
in different amounts. Again, it is worth noting here that multi-celled phytoliths
were not observed in those samples with the highest rate of dissolution
(FY290411 and FY290416, Table 4). In contrast, samples corresponding to fillings
yielded multi-celled phytoliths in higher amounts (around 10–20%).

1.
2.

Download high-res image (644KB)
Download full-size image

Fig. 9. Photomicrographs of phytoliths and calcitic microfossils identified in archeological samples
(at 400×). a) multicellular phytoliths of dendritic long cells with rondels from the husks of Hordeum
sp., b) multicellular phytoliths with hairs form grass leaves/stems, c) multicellular phytoliths with
hairs and stomata cells from grass stems, d) weathered phytolith, e) ash pseudomorph resembling
to Pinus morphologies, f) dung spherulite.

Concentrations of ash pseudomorphs range between 18.000 and 2.7 million per 1 g
of sediment (Table 4). Again, control samples 2670-C and 290412-C were the only
exceptions. Samples 2712 and 2072, from oven (FR2640) and vessel fillings
(VP2069) respectively, were by far the richest (over 2.2 million/1 g of sediment).
Pseudomorphs included mostly prisms and rhombs. These latter were
morphologically similar to those observed in our Pinus halepensis reference
21

MANUSCRIT ACCEPTAT

samples obtained from burned branches (Fig. 9e). In turn, dung spherulites were
noted also in all samples, again with the exception of control sample 2670-C.
Concentrations of dung spherulites constitute between 13.000 and 2.9 million per
1 g of sediment (Table 4, Fig. 9f). The richest samples corresponded to oven
(FR2640, sample 2641) and vase fillings (VP2069, sample 2075).
The data obtained from ash pseudomorph and spherulite absolute numbers were
then used to calculate ratio values, following the quantitative models developed by
Gur-Arieh et al. (2013, 2014). The results indicated distinctive low values (around
0) in most of the samples within installations (n = 6), which can be interpreted as
dominated by dung ash. Most of our control samples from the vicinity of the same
installations showed similar ratio values around 0 (n = 4). These values are
consistent with modern dung fuel standards obtained from ethnoarchaeological
research in central Asia (Gur-Arieh et al., 2013), as well as our reference tabouna
fuel materials from the site area, as previously argued. In contrast, only one of the
samples showed values higher than 5 and can be confidently interpreted as
dominated by wood ash (oven FR2640, sample 2712). Based in further dissolution
experiments conducted by Gur-Arieh et al. (2014), the so-called “gray area” may
indicate either well-preserved ashes composed of dung mixed with wood, or
partially dissolved dung ashes. This is the case for the rest of our samples within
installations with ratio values around 1–3 (n = 4), and some of the control samples
as well (n = 2). Our results reveal that most of the sediments examined within
installations are composed by dung ash, in addition to mixtures of well-preserved
dung and wood or partially dissolved dung-dominated ash.

4. Discussion
Two general issues have been addressed in this study: i) establishing the nature of
fuel sources, preservation and formation processes; ii) tracing fuel use and
domestic activities (i.e. food processing, storage, dumping) at Numidian Althiburos
through time. These overall trends are discussed below.
In this study, using microfossil plant evidence from phytoliths and wood ash
pseudomorphs, in addition to dung direct indicators from spherulites, we were
able to identify fuel sources. The presence of spherulites within all installations
examined confirms that the use of dung as fuel was practiced in all the studied
occupation phases, covering most of the first millennium BC (at least from the 9th
to the 1st centuries BC). However, the range of results obtained from the absolute
quantities of calcitic micro-remains, particularly ash pseudomorphs to dung
spherulites, shows significant differences between both main types of studied
installations, ovens (FR) and hearths (FY).
Our ethnographic observations revealed that the choice of dung for fueling in
traditional mud ovens (both for cooking and firing pottery) is a common practice
followed by Ouarten families living at the site area today. The modern materials
examined and the baking temperatures recorded from such cooking installations
provided a reference assemblage for dung-dominated fuels. One of the main traits
noted is that both plant and fecal microfossil abundances are significantly lower in
archaeological assemblages compared to modern dung fuel materials. This general
22

MANUSCRIT ACCEPTAT

pattern is consistent with data obtained from ethnoarchaeological studies that
have followed similar quantitative approaches (Tsartsidou et al., 2008; Portillo
et al., 2012; Gur-Arieh et al., 2013; Portillo et al., 2014). This is especially true for
calcitic micro-remains, both dung spherulites and wood ash pseudomorphs, which
are dramatically lower in archaeological ashes compared to modern ashes (GurArieh et al., 2014).
In previous studies we have addressed particularly the case of one of these
cooking/baking installations described as a mud oven similar to modern tabounas,
dating to the Late Numidian sub-phase LN1 (FR270223, 4th–2nd century BC,
Fig. 6c–d). The oven filling contained tabouna pieces, ashy sediments and abundant
macro-botanical remains. In the same domestic context, findings from a small pit
located near the oven yielded abundant concentrations of free-threshing wheat
(Triticum aestivum/durum), and hulled barley (Hordeum vulgare) seeds to a lesser
extent (Kallala et al., 2008; Ramon and Maraoui, 2011; López-Reyes and Cantero,
2016). The morphometric study of distinctive dendriform phytoliths from both
oven and pit fillings, allowed the identification of the processed cereal as bread
wheat (T. aestivum) (Portillo and Albert, 2011). Consistently with this previous
study, the microfossil data reported here indicated that oven fillings are also
composed by mixtures of dung and wood or partially dissolved dung-dominated
ash. This is consistent with findings from charred macro-botanical weeds and
legumes, which are commonly related to livestock feeding (i.e. Medicago sp.,
Melilotus sp.) and woody specimens such as Ficus carica and Vitis vinifera in the
same oven fillings. The leaves of these later species are also used as leafy hay by
modern rural communities in the western Maghreb, and the slow burning
temperatures of Ficus wood are appreciated for firing pottery, while dried vine
branches are also used in bread ovens as fuel (Zapata et al., 2003). Interestingly,
flotation samples also provided direct evidence from a well-preserved charred
ovicaprine dung pellet (López-Reyes and Cantero, 2016). Similar to the modern
mud constructed cooking installations from the site area, it was fueled with dung
(likely ovicaprine) and perhaps also mixed dung-wood. The use of this oven can be
related to baking and cereal processing (i.e. dehusking hulled cereals).
A recent microarchaeological study on cooking installations from Late Bronze and
Iron Age Levantine contexts has proved that the use of dung as fuel was practiced
in both periods, although wood-dominated fuel material was the most common
source (Gur-Arieh et al., 2014). This is also the case of the second oven installation
examined in this study (FR2640), which was excavated in 2014. Oven FR2640 is
dated to the Late Numidian (LN2, mid 2nd–1st century BC). Interestingly, the stone
wall in direct contact with the installation showed evidence of burning on its
surface, thus indicating its exposure to flames. The upper filling included abundant
faunal and vegetal charred remains, especially cereal seeds (sampled for further
zooarchaeological and archaeobotanical analyses), varied pottery fragments,
building material and large concentrations of dung micro-remains. On the
contrary, the bottom fill did not provide evidence of pottery or bones, but it was
composed of abundant charred wood pieces. Our results showed high
concentrations of pseudomorphs, indicative of wood-dominated ash. Interestingly,
their morphologies compared favorably to Aleppo pine (Pinus halepensis)
pseudomorphs. The results from the macro-botanical dataset may confirm the use
23

MANUSCRIT ACCEPTAT

of pinewood within the installation. In addition, evidence from partially melted
phytoliths (indicative of high temperatures, above 700 °C, for a long time
exposure) was also noted within the oven bottom filling. Their presence may
provide some insights into burning temperatures within the installation. Partial
melting of phytoliths was also noted in our modern tabouna fuel residues, both
within the bottom of the installation, as well as within the burned fuel stored in a
metal bin for re-use purposes placed close to the oven (Fig. 4b). As previously
argued, maximal temperatures reached around 800 °C in the initial fast firing,
related to the use of woody material, and between 700 and 600 °C during the
following 30 min (Figs. 5 and 8f). In the above mentioned study on Levantine
cooking installations, Fourier Transform Infrared spectroscopy (FTIR) indicated
that within the so-called “Canaanite ovens” maximal temperatures ranged between
500 and 900 °C, and that they were operated in a similar way to most tannurs in
the Levant, that is, internally fueled (Gur-Arieh et al., 2014). Previous ethnographic
research in the Upper Khabur (northern Syria) showed that modern tannurs are
operated similarly (from their interior) and fueled with wood-dominated materials
(primarily cotton branches), whereas ovicaprine dung sporadically used within the
installations bottom. The recorded baking temperatures were above 600 °C
(Portillo et al., 2014).
Another point to emerge from this study is the presence of dung-dominated ashes
in most of the hearths examined belonging to different site areas and occupational
phases, regardless of their morphological characteristics. Most of the hearth
assemblages yielded plant charred remains. In contrast, animal bones were scarce
or completely absent in these domestic installations, thus indicating that bones
were not used as fuel, and not even discarded into the fire (Portillo et al., 2012).
Because all of the installations examined here included direct microfossil evidence
from dung spherulites, we conclude that the use of fecal material for fuelling
hearths was a common practice in the Numidian settlement through time. The
results reported here show that only two of the studied installations indicate
either mixed wood-dung fuel and/or partially dissolved dung-dominated ashes
(FY2710 and FY270355, NA3 sub-phase and NR1 sub-phase, respectively, Table 1).
Here again, the noted calcitic rhomboid microfossils compared favorably to Aleppo
pine (Pinus halepensis) pseudomorphs common in our modern plant reference
material. It should be noted that the charred-wood dataset obtained from
combustion installations and related filling deposits revealed that P. halepensis was
the most common source of fuel plant material in the Numidian settlement through
all its occupational phases (Cantero, 2016). Findings from other woody species,
including juniper (Juniperus sp.) and olive tree (Olea europaea), were also noted in
hearth assemblages from the Late Numidian phase (LN1), although to a lesser
extent. According to our ethnographic observations, olive wood is commonly used
for building materials and crafts (i.e. for the central piece of rotary querns). A
similar use is reported in the Moroccan Rif, although dead wood may be used for
fuel as well (Zapata et al., 2003). In order to explore wood fuel use, further
research will address burning experiments and sampling in modern traditional
ovens used for pine nut exploitation following a similar methodological approach.
Although it is no longer common in the site area due to deforestation, P. halepensis
occupies more than 56% of the total forested area of Tunisia (around 297.000 ha,
Sghaier and Ammari, 2012), mostly in its northwest and central parts. Aleppo pine
24

MANUSCRIT ACCEPTAT

plays a major ecologic, economic and social role in this country. It is exploited for
wood and seeds, and its so-called zgougou nuts are widely consumed (Morales
et al., 2015).
As previously argued, a key question is whether the assemblages examined
represent originally in situ remaining burnt fuel of the last use/s of the
installations, or rather post-abandonment dumping events. This is especially
important in such type of features whose closed structure may allow secondary
uses as bins. This may be the case at least of the upper oven fillings, characterized
by typical fill-like materials related to household domestic waste (i.e. animal bones,
charcoal, seeds, building materials and ceramics), in addition to dung material, as
described above. It is clear that micromorphological analyses would provide clues
to a more detailed interpretation, thus defining indicators that may be useful in
assessing the integrity of the fill sediments associated to the cooking installations,
including preservation. They may also provide an insight into whether or not ashes
were formed in situ. As pointed out by micromorphological evidence, both
installation type (hearths vs ovens or walled features) and location (in-door vs outdoor), in addition to burial conditions (fast vs slow) may determine microfossil
preservation. We therefore assume that walled (ovens or walled features, such as
ceramic vessels and bins) and in-door installations may be less exposed to
dissolution, and micro-remains may be better preserved, as noted in recent studies
(i.e. Gur-Arieh et al., 2014; Kadowaki et al., 2015). The last case exposed here
addresses some insights regarding preservation questions, in addition to storage
and dumping fuel behaviors.
We also examined the filling sediments of an in situ large ceramic vessel (VP2069)
from an indoor space that is described as a storage area by the excavators, since
another large pottery container was also found within it (Fig. 7). It is dated to the
Late Numidian (LN1, 4th to mid-2nd century BC). VP2069 was especially rich in
charred remains, particularly seeds. The vase itself was deposited in a pit, whose
fillings were also examined for comparative purposes. Both were reported in the
field as ashy black fill sediments with abundant charcoal fragments and wellpreserved seeds, primarily from crops. The macro-botanical dataset from both
vase fillings is dominated by Hordeum vulgare and Triticum aestivum/durum, in
addition to weeds and legumes (i.e. Lens culinaris, Vicia faba, Medicago sativa,
Melilotus sp.) and fruit seeds that include Ficus carica and Vitis vinifera (LópezReyes and Cantero, 2016). Well-preserved charred cereal stems were also noted, in
addition to clusters of complete charred ovicaprine dung pellets. These fillings
were also rich in microfossil evidence from dung spherulites. Additionally, our
phytolith results indicate concentrations of multi-celled or interconnected
phytoliths, from both floral parts of cereals, primarily wheat and barley, and from
leaves and the stems as well (Fig. 9a–b–c), with a low dissolution rate. Overall,
these findings indicate that the vase filling corresponds to well-preserved dungdominated ashes with abundant charred plant material, mostly related to
agricultural crops and their sub-products. These assemblages provide us with a
good example of the excellent preservation conditions inside close walled-indoors.
In contrast, the associated pit fillings showed a higher phytolith dissolution rate, in
addition to evidence of partial melting, which is indicative of exposure to high
temperatures. Additionally, dung spherulite abundances decreased dramatically
25

MANUSCRIT ACCEPTAT

and charred ovicaprine dung pellets were not present either. Both micro and
macro-botanical remains were similar to the vase fillings, in addition to
concentrations of wood ash pseudomorphs, which are indicative of dung fuel and
mixed wood-dung material and/or partially dissolved dung-dominated ashes; the
latter may be the case outside the vase. In general, the vase fillings examined
compare favorably to the above-described oven tabouna assemblages from the
same Late Numidian phase. We conclude that the context examined may relate to
dumping episodes from nearby firing installations (likely including domestic
ovens) and/or perhaps the storage of fuel remains for burned fuel use purposes
(i.e. manure, fuel re-use), as observed in traditional households in the site area and
in other areas of the Maghreb today (Zapata et al., 2003).

5. Conclusions
These studies have shown how phytoliths and calcitic microfossil associations
(dung spherulites in addition to wood ash pseudomorphs) provide direct evidence
from the nature of fuel matter in a selection of cooking installations from Numidian
Althiburos, primarily ovens and hearths. The results indicated different types of
fuel sources: livestock dung (at least from ovicaprines), wood (mainly Pinus
halepensis), and a mixing of dung and vegetal matter (wood and agricultural byproducts). Dung was used as source of fuel material across time (at least from the
9th to the last centuries BC), and in different site areas and types of cooking
installations, both hearths and ovens.
Additionally, this study examines ways in which ethnoarchaeological approaches
contribute to addressing taphonomic issues, which are fundamental for
interpreting archaeological contexts. Our ethnographic study shows that the
temperatures recorded within tabouna oven dung-fuels reached around 800 °C in
the initial firing, related to the use of woody material in the lighting, and decreased
to 600–700 °C during most part of the cooking activity. The modern materials
examined from the same oven installation provided a reference assemblage for
dung-dominated fuels (i.e. partial melting of phytoliths which were also noted in
archeological oven samples and indicative of high combustion temperatures in
such walled installations).
We have also addressed a comparative case of vase fillings containing abundant
well-preserved mixed wood-dung material from an in-door building space. This
latter context may relate to dumping and/or perhaps the storage of fuel remains
for re-use or fertilizing, common practices also observed in modern households.
We therefore conclude that installation type (hearths/ovens or walled features,
such as ceramic vessels and bins) may determine both micro and macro-remains
preservation. Although depending on burial conditions and other postdepositional processes, it is clear that potentially walled and indoor installations
may be less exposed to dissolution and microfossils may be better preserved.
The results reported upon here demonstrate the value of integrating
microarchaeological and ethnoarchaeological techniques to traditional
macroscopic archaeological research. We assume that, despite present-day
conditions are not completely analogous to those of past times,
26

MANUSCRIT ACCEPTAT

ethnoarchaeological approaches provides us with a reference framework for
better understanding farming practices in this still poorly investigated region in
northern Africa.

Acknowledgements
The project has been supported by the National Heritage Institute of Tunisia (INP),
the University of Barcelona, the Catalan Agency for Administration of University
and Research Grants (AGAUR, 2006-EXCAVA00011 and 2009SGR 1418), the
Spanish Ministry of Innovation and Science (MICINN, HUM2006-03432/HIST and
HAR2009-13045), the Spanish Ministry of Culture (MECUL, CUL/3348/2009 and
CUL/2165/2009), the Spanish Agency for International Development Cooperation
(AECID, A/017890/08), and the Spanish Ministry of Competitiveness and Economy
(MINECO, HAR2012-39189-C02-0 and HAR2013-42054-P). The first author
research has been funded by postdoctoral fellows from the Spanish Ministry
Science and Education (MEC) at the Institute of Archaeology– University College
London, and the Beatriu de Pinós (AGAUR) and the Juan de la Cierva (MICINN) at
the University of Barcelona. She is currently part of the Prehistory Consolidated
Research Team at the University of the Basque Country, UPV/EHU (IT-622-13).
Modern reference and archaeological samples were processed at the phytolith
laboratory at the University of Barcelona.
We are grateful to many members of the Althiburos excavations team for their
helpful discussions, specifically Daniel López-Reyes and Francisco José Cantero
(UB), Silvia Valenzuela (Sheffield University), Mounir Torchani (INP, Tunis
University) and Halib Soltani (Tunis University). Thanks are extended to Francisco
Hernández-Camacho (UB) and Breanne Clifton (Connecticut University), who
helped in the phytolith laboratory at the UB. Ethnographical record and modern
material sampling was made possible thanks to the kind help of the site workers
and their families. Thanks are due to Gouasdya and El Souidat inhabitants, with
special thanks to Habib and Fadhel Souidi families, for providing us with valuable
information on the site area and local husbandry practices, as well as for their kind
hospitality.

References






Albert and Cabanes, 2007
R.M. Albert, D. CabanesFire in prehistory: an experimental approach to
combustion processes and phytolith remains
Israel Journal of Earth Sciences, 56 (2007), pp. 175-189
Albert and Weiner, 2001
R.M. Albert, S. WeinerStudy of phytoliths in prehistoric ash layers using a
quantitative approach
J.D. Meunier, F. Colin (Eds.), Phytoliths, Applications in Earth Sciences and Human
History, A.A. Balkema Publishers, Lisse (2001), pp. 251-266
Albert et al., 1999
R.M. Albert, O. Lavi, L. Estroff, S. Weiner, A. Tsatskin, A. Ronen, S. Lev-YadunMode
of occupation of Tabun Cave, Mt Carmel, Israel during the Mousterian period:
a study of the sediments and phytoliths
Journal of Archaeological Science, 26 (1999), pp. 1249-1260
27

MANUSCRIT ACCEPTAT





















ArticlePDF (401KB)
Albert et al., 2000
R.M. Albert, O. Bar-Yosef, L. Meignen, S. WeinerPhytoliths in the Middle
Paleolithic deposits of Kebara cave, Mt. Carmel, Israel: Study of the plant
materials used for fuel and other purposes
Journal of Archaeological Science, 27 (2000), pp. 931-947
ArticlePDF (870KB)
Albert et al., 2003
R.M. Albert, O. Bar-Yosef, P. Goldberg, L. Meignen, S. WeinerPhytolith and
mineralogical studies of hearths from the Middle Paleolithic levels of
Hayonim cave (Galilee, Israel)
Journal of Archaeological Science, 30 (2003), pp. 461-480
ArticlePDF (872KB)
Albert et al., 2008
R.M. Albert, R. Shahack-Gross, D. Cabanes, A. Gilboa, S. Lev-Yadun, M. Portillo, I.
Sharon, E. Boaretto, S. WeinerPhytolith-rich layers from the late Bronze and
Iron Ages at Tel Dor (Israel): mode of formation and archaeological
significance
Journal of Archaeological Science, 35 (2008), pp. 57-75
ArticlePDF (2MB)
Albert et al., 2011
R. Albert, X. Esteve, M. Portillo, A. Rodríguez-Cintas, D. Cabanes, I. Esteban, F.
HernándezPhytolith CoRe, Phytolith Reference Collection
(2011)
Retrieved Jan 21, 2015, from
http://phytcore.org/phytolith/index
Anderson and Ertug-Yaras, 1998
S. Anderson, F. Ertug-YarasFuel fodder and faeces: an ethnographic and
botanical study of dung fuel use in central Anatolia
Environmental Archaeology, 1 (1998), pp. 99-110
Belarte, 2011
M.C. BelarteLes sondages dans la zone 1
N. Kallala, J. Sanmartí (Eds.), Althiburos I. La fouille dans l'aire du capitole et dans
la nécropole méridionale. Serie Documenta, Institut Català d'Arqueologia Clàssica,
Tarragona (2011), pp. 45-151
Berna et al., 2007
F. Berna, A. Behar, R. Shahack-Gross, J. Berg, E. Boaretto, A. Gilboa, I. Sharon, S.
Shalev, S. Shilshtein, N. Yahalom-Mack, J.R. Zorn, S. WeinerSediments exposed to
high temperatures: reconstructing pyrotechnological practices in Late
Bronze and Iron Age strata at Tel Dor (Israel)
Journal of Archaeological Science, 34 (2007), pp. 358-373
ArticlePDF (2MB)
Bottema, 1984
S. BottemaThe composition of modern charred seed assemblages
W. van Zeist, W. Casparie (Eds.), Plants and Ancient Man, A.A. Balkema, Rotterdam
(1984), pp. 207-211
Brochier, 1996
J.E. BrochierFeuilles ou fumiers? Observations sur le rôle des poussières
sphérolitiques dans l'interprétatation des dépôts archaéologiques holocènes
Anthropozoologica, 24 (1996), pp. 19-30
Brochier and Thinon, 2003
J.E. Brochier, M. ThinonCalcite crystals, starch grains aggregates or…POCC?
Comment on ‘calcite crystals inside archaeological plant tissues’
Journal of Archaeological Science, 30 (2003), pp. 1211-1214
28

MANUSCRIT ACCEPTAT





















ArticlePDF (253KB)
Brochier et al., 1992
J.E. Brochier, P. Villa, M. Giacomarra, A. TagliacozzoShepherds and sediments:
geo-ethnoarchaeology of pastoral sites
Journal of Anthropological Archaeology, 11 (1992), pp. 47-102
ArticlePDF (17MB)
Brown, 1984
D.A. BrownProspects and limits of a phytolith key for grasses in the central
United States
Journal of Archaeological Science, 11 (1984), pp. 345-368
ArticlePDF (1MB)
Cantero, 2016
F.J. CanteroRessources forestières à Althiburos à partir de l'étude des
charbons de bois
N. Kallala, J. Sanmartí, M.C. Belarte (Eds.), Althiburos II. L'aire du capitole et la
nécropole méridionale: études, Institut Català d'Arqueologia Clàssica, Tarragona
(2016)
(in press)
Canti, 1997
M.G. CantiAn investigation of microscopic calcareous spherulites from
herbivore dungs
Journal of Archaeological Science, 24 (1997), pp. 219-231
ArticlePDF (2MB)
Canti, 1998
M.G. CantiThe micromorphological identification of faecal spherulites from
archaeological and modern materials
Journal of Archaeological Science, 25 (1998), pp. 435-444
ArticlePDF (247KB)
Canti, 1999
M.G. CantiThe production and preservation of faecal spherulites: animals,
environment and taphonomy
Journal of Archaeological Science, 26 (1999), pp. 251-258
ArticlePDF (240KB)
Canti, 2003
M.G. CantiAspects of chemical and microscopic characteristics of plant ashes
found in archaeological soils
Catena, 54 (2003), pp. 339-361
ArticlePDF (3MB)
Gur-Arieh et al., 2012
S. Gur-Arieh, E. Boaretto, A. Maeir, R. Shahack-GrossFormation processes in
philistine hearths from tell es-Safi/Gath (Israel): an experimental approach
Journal of Field Archaeology, 37 (2012), pp. 121-131
Gur-Arieh et al., 2013
S. Gur-Arieh, E. Mintz, E. Boaretto, R. Shahack-GrossAn ethnoarchaeological
study of cooking installations in rural Uzbekistan: development of a new
method for identification of fuel sources
Journal of Archaeological Science, 40 (2013), pp. 4331-4347
ArticlePDF (4MB)
Gur-Arieh et al., 2014
S. Gur-Arieh, R. Shahack-Gross, A.M. Maeir, G. Lehmann, L.A. Hitchcock, E.
BoarettoThe taphonomy and preservation of wood and dung ashes found in
archaeological cooking installations: case studies from Iron Age Israel
Journal of Archaeological Science, 46 (2014), pp. 50-67
ArticlePDF (4MB)
29

MANUSCRIT ACCEPTAT




















Kadowaki et al., 2015
S. Kadowaki, L. Maher, M. Portillo, R.M. Albert, Ch Akashi, F. Guliyev, Y.
NishiakiGeoarchaeological and palaeobotanical evidence for prehistoric
cereal storage in the southern Caucasus: the Neolithic settlement of Göytepe
(mid. 8th millennium BP)
Journal of Archaeological Science, 53 (2015), pp. 408-425
ArticlePDF (8MB)
Kallala and Sanmartí, 2011
N. Kallala, J. SanmartíSynthèse des résultats: deux mille ans d'histoire
d'Althiburos
N. Kallala, J. Sanmartí (Eds.), Althiburos I. La fouille dans l'aire du capitole et dans
la nécropole méridionale. Serie Documenta, Institut Català d'Arqueologia Clàssica,
Tarragona (2011), pp. 31-43
Kallala et al., 2008
N. Kallala, J. Sanmartí, M.C. Belarte, J. Ramon, R. Álvarez, M.B. Moussa, S. Bechrifiia,
X. Bermúdez, J. Campillo, N. Chebbi, T. Fadrique, R. Jornet, D. López, Z.B.H.N. Loum,
B. Maraoui, J. Noguera, J.M. Puche, V. Revilla, N. Tarradell, M. Torchani, S.
ValenzuleaRecherches sur l'occupation d'Althiburos (région de Kef, Tunisie)
et de ses environs à l'époque numide
Pyrenae, 39 (1) (2008), pp. 67-113
Latiri et al., 2010
K. Latiri, J.P. Lhomme, M. Annabi, T.L. SetterWheat production in Tunisia:
progress, inter-annual variability and relation to rainfall
European Journal of Agronomy, 33 (2010), pp. 33-42
ArticlePDF (1MB)
López-Reyes and Cantero, 2016
D. López-Reyes, F.J. CanteroAgriculture et alimentation à Althiburos à partir de
l'étude des graines et des fruits
N. Kallala, J. (Dir.) Sanmartí, M.C. Belarte (Eds.), Althiburos II. L'aire du capitole et
la nécropole méridionale: études, Institut Català d'Arqueologia Clàssica, Tarragona
(2016)
(in press)
Madella et al., 2005
M. Madella, A. Alexandre, T.B. Ball, ICPN Working GroupInternational code for
phytolith nomenclature 1.0
Annals of Botany, 96 (2005), pp. 253-260
Matthews, 2010
W. MatthewsGeoarchaeology and taphonomy of plant remains and
microarchaeological residues in early urban environments in the Ancient
Near East
Quaternary International, 214 (2010), pp. 98-113
ArticlePDF (3MB)
Miller, 1984
N.F. MillerThe use of dung as a fuel: an ethnographic example and an
archaeological application
Paléorient, 10 (2) (1984), pp. 71-79
Morales et al., 2015
J. Morales, S. Mulazzani, L. Belhouchet, A. Zazzo, L. Berrio, W. Eddargach, A. Cervi,
H. Hamdi, M. Saidi, A. Coppa, L. Peña-ChocarroFirst preliminary evidence for
basketry and nut consumption in the Capsian culture (ca. 10,000–7500 BP):
archaeobotanical data from new excavations at El Mekta, Tunisia
Journal of Anthropological Archaeology, 37 (2015), pp. 128-139
ArticlePDF (3MB)
Mulholland and Rapp, 1992
30

MANUSCRIT ACCEPTAT



















S.C. Mulholland, G. Rapp Jr.A morphological classification of grass silica-bodies
G. Rapp Jr., S.C. Mulholland (Eds.), Phytolith Systematics: Emerging Issues,
Advances in Archaeological and Museum Science, Plenum Press, New York (1992),
pp. 65-89
Pearsall, 1989
D.M. PearsallPaleoethnobotany: a Handbook of Procedures
Academic Press, San Diego (1989)
Peña-Chocarro et al., 2009
L. Peña-Chocarro, L. Zapata, J.E. González Urquijo, J.J. IbáñezEinkorn (Triticum
monococcum L.) cultivation in mountain communities of the western Rif
(Morocco): an ethnoarchaeological project
A.S. Fairnbairn, E. Weiss (Eds.), From Foragers to Farmers. Gordon Hillman
Festschrift, Oxbow, Oxford (2009), pp. 103-111
Peña-Chocarro et al., 2015
L. Peña-Chocarro, G. Pérez Jordà, J. Morales Mateos, L. ZapataStorage in
traditional farming communities of the western Mediterranean:
ethnographic, historical and archaeological data
Environmental Archaeology, 20 (2015), pp. 379-389
Picornell et al., 2011
Ll Picornell, E. Asouti, E. AlluéThe ethnoarchaeology of firewood management
in the Fang villages of Equatorial Guinea, central Africa: Implications for the
interpretation of wood fuel remains from archaeological sites
Journal of Anthropological Archaeology, 30 (2011), pp. 375-384
Piperno, 1988
D.R. PipernoPhytolith Analysis: an Archaeological and Geological Perspective
Academic Press, San Diego (1988)
Piperno, 2006
D.R. PipernoPhytoliths: a Comprehensive Guide for Archaeologists and
Paleoecologists
AltaMira Press, Lanham (2006)
Portillo and Albert, 2011
M. Portillo, R.M. AlbertHusbandry practices and livestock dung at the
Numidian site of Althiburos (el Médéina, Kef Governorate, northern Tunisia):
the phytolith and spherulite evidence
Journal of Archaeological Science, 38 (2011), pp. 3224-3233
ArticlePDF (1MB)
Portillo and Albert, 2016
M. Portillo, R.M. AlbertLes activités domestiques de la période numide à
travers de l'étude des microrestes végétaux et fécaux : phytolithes et
sphérolithes
N. Kallala, J. (Dir.) Sanmartí, M.C. Belarte (Eds.), Althiburos II. L'aire du capitole et
la nécropole méridionale: études, Institut Català d'Arqueologia Clàssica, Tarragona
(2016)
(in press)
Portillo et al., 2012
M. Portillo, S. Valenzuela, R.M. AlbertDomestic patterns in the Numidian site of
Althiburos (northern Tunisia): the results from a combined study of animal
bones, dung and plant remains
Quaternary International, 275 (2012), pp. 84-96
ArticlePDF (2MB)
Portillo et al., 2014
M. Portillo, S. Kadowaki, Y. Nishiaki, R.M. AlbertEarly Neolithic household
behavior at Tell Seker al-Aheimar (Upper Khabur, Syria): a comparison to
ethnoarchaeological study of phytoliths and dung spherulites
31

MANUSCRIT ACCEPTAT





















Journal of Archaeological Science, 42 (2014), pp. 107-118
ArticlePDF (1MB)
Ramon and Maraoui, 2011
J.R. Ramon, B. MaraouiLes sondages dans la zone 2
N. Kallala, J. Sanmartí (Eds.), Althiburos I. La fouille dans l'aire du capitole et dans
la nécropole méridionale. Serie Documenta, Institut Català d'Arqueologia Clàssica,
Tarragona (2011), pp. 153-391
Reddy, 1998
S.N. ReddyFueling the hearths in India: the role of dung in
paleoethnobotanical interpretation
Paléorient, 24 (2) (1998), pp. 61-70
Rosen, 1992
A.M. RosenPreliminary identification of silica skeletons from Near Eastern
archaeological sites: an anatomical approach
G. Rapp Jr., S.C. Mulholland (Eds.), Phytolith Systematics: Emerging Issues,
Advances in Archaeological and Museum Science, Plenum Press, New York (1992),
pp. 129-147
Sanmartí et al., 2012
J. Sanmartí, N. Kallala, M.C. Belarte, J. Ramon, B. Maraoui, R. Jornet, S.
MiniaouiFilling gaps in the protohistory of the Eastern maghreb: the
althiburos archaeological project (El Kef, Tunisia)
Journal of African Archaeology, 10 (2012), pp. 21-44
Sghaier and Ammari, 2012
T. Sghaier, Y. AmmariCroissance et production du pin d'Alep (Pinus halepensis
Mill.) en Tunisie
Ecologia Mediterranea, 38 (1) (2012), pp. 39-57
Shahack-Gross, 2011
R. Shahack-GrossHerbivorous livestock dung: formation, taphonomy, methods
for identification, and archaeological significance
Journal of Archaeological Science, 38 (2011), pp. 205-218
ArticlePDF (2MB)
Shahack-Gross et al., 2004
R. Shahack-Gross, F. Marshall, K. Ryan, S. WeinerReconstruction of spatial
organization in abandoned Maasai settlements: implications for site
structure in the pastoral Neolithic of East Africa
Journal of Archaeological Science, 31 (2004), pp. 1395-1411
ArticlePDF (986KB)
Sillar, 2000
B. SillarDung by preference: the choice of fuel as an example of how Andean
pottery production is embedded within wider technical, social and economic
practices
Archaeometry, 42 (1) (2000), pp. 43-60
Tsartsidou et al., 2007
G. Tsartsidou, S. Lev-Yadun, R.M. Albert, A. Miller-Rosen, N. Efstratiou, S.
WeinerThe phytolith archaeological record: strengths and weaknesses based
on a quantitative modern reference collection from Greece
Journal of Archaeological Science, 34 (2007), pp. 1262-1275
ArticlePDF (606KB)
Tsartsidou et al., 2008
G. Tsartsidou, S. Lev-Yadun, N. Efstratiou, S. WeinerEthnoarchaeological study of
phytolith assemblages from an agro-pastoral village in Northern Greece
(Sarakini): development and application of a Phytolith Difference Index
Journal of Archaeological Science, 35 (2008), pp. 600-613
ArticlePDF (386KB)
32

MANUSCRIT ACCEPTAT












Twiss, 1992
P.C. TwissPredicted world distribution of C3 and C4 grass phytoliths
G. Rapp Jr., S.C. Mulholland (Eds.), Phytolith Systematics: Emerging Issues,
Advances in Archaeological and Museum Science, Plenum Press, New York (1992),
pp. 113-128
Twiss et al., 1969
P.C. Twiss, E. Suess, R.M. SmithMorphological classification of grass phytoliths
Soil Science Society of America Proceedings, 33 (1969), pp. 109-115
Valenzuela, 2016
S. ValenzuelaAlimentation et élevage à Althiburos à partir des restes
fauniques
N. Kallala, J. Sanmartí, M.C. Belarte (Eds.), Althiburos II. L'aire du capitole et la
nécropole méridionale: études, Institut Català d'Arqueologia Clàssica, Tarragona
(2016)
(in press)
Van der Veen, 2007
M. Van der VeenFormation processes of desiccated and carbonized plant
remains – the identification of routine practice
Journal of Archaeological Science, 34 (2007), pp. 968-990
ArticlePDF (1MB)
Wattez and Courty, 1987
J. Wattez, M.A. CourtyMorphology of ash of some plant remains
N. Fédoroff, L.M. Bresson, M.A. Courty (Eds.), Micromorphologie des sols – Soil
Micromorphology. Association Francaise pour l'Étude du Sol, Plaisir (1987), pp.
677-682
Weiner, 2010
S. WeinerMicroarchaeology. Beyond the Visible Archaeological Record
Cambridge University Press, New York (2010)
Zapata et al., 2003
L. Zapata, L. Peña-Chocarro, J.J. Ibáñez Estévez, J.E. González
UrquijoEthnoarchaeology in the Moroccan Jebala (Western Rif): wood and
dung as fuel
Africa Praehistorica, 15 (2003), pp. 163-175

33