RESEARCH ARTICLE

Spatial analyses of archaeobotanical record

reveal site uses and activities at Early to

Middle Holocene Takarkori (Libya, Central

Sahara)

Savino di Lernia1,2, Fabrizio BuldriniID
3, Assunta FlorenzanoID

4*, Anna Maria Mercuri4,

Varinia Nardi5, Rocco RotunnoID
1*

1 Dipartimento di Scienze dell’Antichità, Sapienza University of Rome, Rome, Italy, 2 GAES, University of

Witwatersrand, Johannesburg, South Africa, 3 Dipartimento di Scienze Biologiche, Geologiche e Ambientali,

Università di Bologna, Bologna, Italy, 4 Laboratorio di Palinologia e Paleobotanica, Dipartimento di Scienze

della Vita, Università di Modena e Reggio Emilia, Modena, Italy, 5 The Archaeological Mission in the Sahara,

Sapienza University of Rome, Rome, Italy

* assunta.florenzano@unimore.it (AF); rotunno.rocco@gmail.com (RR)

Abstract

This study investigates botanical remains from the Takarkori site in the Tadrart Acacus

region (SW Libya) to reconstruct socio-economic and cultural characteristics of human

groups during the Holocene. By analyzing micro- and macrofossils of plant origin, we aim to

understand the availability and management of environmental resources and how plant taxa

were used by humans. The exceptional preservation of archaeobotanical material across all

occupation levels, facilitated by the region’s geomorphological and environmental condi-

tions, provides a unique opportunity to study pre-Pastoral and Pastoral Neolithic activities

within a comprehensive diachronic framework. Our research extends previous investiga-

tions by examining the spatial distribution of archaeobotanical remains in association with

site furniture and material correlates, offering insights into the functional use of space within

the site. Also, the features of plant assemblages and their distribution patterns indicate the

planning in the use of plant resources and the diverse uses beyond subsistence, including

ritual and cultural practices. The findings contribute to a deeper understanding of Holocene

environmental and cultural dynamics, highlighting the importance of archaeobotanical data

in archaeological research.

Introduction

The complex relationship between humans and plants has been widely addressed in archaeo-

logical research. Plant remains as macrofossils and pollen provide crucial insights into the sub-

sistence strategies and land use patterns of ancient communities [e.g. 1–4] Archaeobotany,

also including the analyses of phytoliths, starch grains, lipids and other molecules, has emerged

as a vital field for identifying activity areas in prehistoric habitation sites [e.g. 5–8]. By

PLOS ONE

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 1 / 40

a1111111111

a1111111111

a1111111111

a1111111111

a1111111111

OPEN ACCESS

Citation: di Lernia S, Buldrini F, Florenzano A,

Mercuri AM, Nardi V, Rotunno R (2024) Spatial

analyses of archaeobotanical record reveal site

uses and activities at Early to Middle Holocene

Takarkori (Libya, Central Sahara). PLoS ONE

19(10): e0310739. https://doi.org/10.1371/journal.

pone.0310739

Editor: Raven Garvey, University of Michigan,

UNITED STATES OF AMERICA

Received: November 21, 2023

Accepted: September 4, 2024

Published: October 23, 2024

Copyright: © 2024 di Lernia et al. This is an open

access article distributed under the terms of the

Creative Commons Attribution License, which

permits unrestricted use, distribution, and

reproduction in any medium, provided the original

author and source are credited.

Data Availability Statement: All relevant data are

within the manuscript and its Supporting

Information files.

Funding: Funds for research came from Sapienza

University of Rome (Grandi Scavi di Ateneo) and

the Minister of Foreign Affairs (DGPCC), entrusted

to S.D.L. Funds for APC were provided by A.M.M.

The funders had no role in study design, data

collection and analysis, decision to publish, or

preparation of the manuscript.

https://orcid.org/0000-0002-0376-6760
https://orcid.org/0000-0003-4759-6406
https://orcid.org/0000-0001-9235-8082
https://doi.org/10.1371/journal.pone.0310739
http://crossmark.crossref.org/dialog/?doi=10.1371/journal.pone.0310739&domain=pdf&date_stamp=2024-10-23
http://crossmark.crossref.org/dialog/?doi=10.1371/journal.pone.0310739&domain=pdf&date_stamp=2024-10-23
http://crossmark.crossref.org/dialog/?doi=10.1371/journal.pone.0310739&domain=pdf&date_stamp=2024-10-23
http://crossmark.crossref.org/dialog/?doi=10.1371/journal.pone.0310739&domain=pdf&date_stamp=2024-10-23
http://crossmark.crossref.org/dialog/?doi=10.1371/journal.pone.0310739&domain=pdf&date_stamp=2024-10-23
http://crossmark.crossref.org/dialog/?doi=10.1371/journal.pone.0310739&domain=pdf&date_stamp=2024-10-23
https://doi.org/10.1371/journal.pone.0310739
https://doi.org/10.1371/journal.pone.0310739
http://creativecommons.org/licenses/by/4.0/


analyzing charred seeds, wood, pollen and other plant remains, researchers infer past dietary

preferences, and the use of plants for medicinal, ritual, and construction purposes [9]. The

examination of remains from cultivated crops, along with those from wild plants, and the

reconstruction of agricultural practices, offer essential insights into the extent of agricultural

activities, innovations in farming, and the level of sedentism within ancient communities.

Moreover, the study of pollen and phytoliths reveal changes in vegetation, land use, and envi-

ronmental conditions, contributing to our understanding of past ecosystem dynamics and

their interplay with human activities [10–13]. Spatial analysis techniques provide a powerful

toolkit for investigating the distribution, density, and arrangement of archaeological contexts

and features, including plant remains [e.g. 14]. Geographic Information Systems (GIS) and

geospatial analysis methods offer efficient means to map, analyze, and interpret spatial data

related to habitation sites [15–17]. These techniques enable researchers to identify settlement

patterns, recognize activity areas, and uncover patterns of land use and resource exploitation.

By integrating spatial analysis with archaeological and archaeobotanical data, researchers

can discern the functional zoning of prehistoric settlements, identify areas of intensive activi-

ties such as food processing, craft production, and communal spaces, and explore the socio-

economic organization within the archaeological sites [18, 19]. Furthermore, the combination

of archaeobotany and spatial analysis facilitates the identification of subtle site boundaries,

transitional zones, and the detection of hidden or ephemeral features that might be overlooked

by traditional excavation methods alone [20]. The main objective of this work is to apply these

interdisciplinary approaches to the Takarkori site, one of the most important archives of Holo-

cene Saharan prehistoric societies [21], aiming to extract comprehensive environmental and

cultural information from the spatial distribution of botanical artefacts. This research will con-

tribute to a deeper understanding of Holocene Saharan communities, how they organized

their activities spatially within the site, and their adaptive strategies during changing and vary-

ing environmental and climate settings.

The context

Located in south-western Libya, the Takarkori rock shelter opens on the western slopes of the

Tadrart Acacus, a massif in the central Sahara. The shelter looks to the west and extends for

about 80 m in a north-south direction on a wide terrace, bordered to the east by a rock cliff

about 30 m high [21]. The region is currently characterized by hyper-arid climate with low

rainfall and vegetation limited to desert savannah with specialized species, such as acacia

(Vachellia tortilis (Forssk.) Galasso & Banfi subsp. raddiana (Savi) Kyal. & Boatwr., formerly

named Acacia tortilis Forssk. subsp. raddiana (Savi) Brenan) and tamarisk (Tamarix aphylla
(L.) H.Karst.) usually close to synanthropic areas [22].

From a geomorphological point of view [23], this area can be defined as a vast geosyncline

characterized by large rock formations (Tadrart Acacus massif and Messak plateau), large

dune fields (Erg Titersine, Uan Kasa, and Edeyen of Murzuq), valleys and fossil river networks

locally termed ‘wadi’ (Wadi Takarkori, Wadi Tanezzuft, and Wadi Barjuj). The Tadrart Acacus

massif extends approximately 150 km north-south (26–24˚ North), near the current border

between Libya and Algeria (Fig 1). From an orographic point of view, this formation is similar

to the vast mountainous complex of the Algerian Tassili-n-Ajjer, from which it is separated by

the Wadi Tanezzuft to the west, while the vast expanses of the Erg Titersine and Uan Kasa

encircle its northwestern and eastern limits respectively [23, 24]. The Wadi Takarkori sepa-

rates the Tadrart Acacus to the north from the Algerian Tadrart to the south.

The rock shelter (Fig 2A) was investigated as part of the larger Wadi Takarkori Project [25,

26], a territorial research program in the southern Tadrart Acacus. The investigations were

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 2 / 40

Competing interests: The authors have declared

that no competing interests exist.

https://doi.org/10.1371/journal.pone.0310739


Fig 1. Study area. A) Map of main localities cited in the text; dashed rectangle indicates (B) the study area in

southwestern Libya (Background images from ESRI and OPM).

https://doi.org/10.1371/journal.pone.0310739.g001

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 3 / 40

https://doi.org/10.1371/journal.pone.0310739.g001
https://doi.org/10.1371/journal.pone.0310739


carried out through four excavation campaigns, from 2003 to 2006, while subsequent activities

on the site were interrupted due to geopolitical events linked to the ‘Arab Spring’ [27]. The

archaeological sequence of the Takarkori shelter encompasses over 5000 years, from the

Fig 2. The context and sampling. A) a view of the rock shelter from south-west; B) site topography and excavation sectors; C) 5 l volumetric sampling: the

meshes used for dry sieving in the field (decreasing width: 10, 2 and 0.5 mm); D) sampling strategy with relative and absolute frequency of archaeobotanical

remains retrieved from sampling evidencing absence of documentation/recording gaps (photos ©The Archaeological Mission in the Sahara).

https://doi.org/10.1371/journal.pone.0310739.g002

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 4 / 40

https://doi.org/10.1371/journal.pone.0310739.g002
https://doi.org/10.1371/journal.pone.0310739


occupation of hunter-gatherer-fishers locally called Late Acacus, to the Early, Middle and Late

phases of the Pastoral Neolithic. The stratigraphic excavation, secured by a significant number

of 14C dates [28], led to the definition of additional sub-phases (Table 1). Early Holocene use

(Late Acacus (LA) phase) was by hunter-gatherer-fishers, followed by a prolonged Pastoral

Neolithic occupation (Early Pastoral (EP), Middle Pastoral (MP), Late Pastoral (LP) phases)

marked by cattle and ovicaprine herders in the Middle and Late Holocene. Stone structures,

fireplaces, grinding stones, and potsherds indicate semi-residential utilization by LA dwellers

[29]. Palaeoenvironmental and palaeovegetational reconstructions indicate grasslands and

sparse wooded savannahs dominated, along with permanent freshwater habitats characterized

by pondweeds and reeds [25], and a progressively more unstable climatic context towards its

ending, culminating with an arid phase centered around the so-called 8.2 ka “event” [30]. The

EP phase displays fireplaces, stone structures, and human graves. The environment was

warmer, where wet habitats diminished while dry shrubs and herbs, mainly belonging to the

Asteraceae family and Chenopodioideae sub-family, expanded. The MP phase highlights a cat-

tle-focused economy with evidence of dairying and saw increasing aridity and the spread of

xerophytes like acacias. Lastly, organic remains, cuvette hearths, and ovicaprine dung point to

specialized LP goat herders adapted to extremely dry conditions and desert savannah.

Data collection and methods

The excavations

The stratigraphic excavations were organized into four different sectors (Fig 2B), using a 1x1

m grid set up on the entire area of the terrace (2200 m2). The total excavated area is 143 m2.

The Main Sector (MS therein) with 117 m2 is the largest area of investigation: here the excava-

tion was interrupted for conservation purposes. The bedrock was reached instead in the

Northern Sector (NS therein), an 8 m2 trench, in which the oldest occupation levels, related to

LA1, were identified immediately lying on the bedrock, ~1.7 m below the present surface. The

Southern Sector (SS therein) covers an area of 20 m2 including a conical tumulus (T1), that did

not yield any human remains. The Western Sector (WS), a 9 m2 trench, is the only excavated

Table 1. Main socio-cultural phases and eco-climatic features.

Cultural Phase Sub-phase uncal BP cal BP Socio-cultural features Eco-climatic features

Late Acacus (LA) LA1 8900–

8500

10170–

9400

Hunter-gatherer-fishers Wet savannah with permanent water reservoirs, grasslands and sparse tree

cover

LA2 8500–

7900

9500–8600

LA3 7900–

7400

9000–8000

Early Pastoral (EP) EP1 7400–

6900

8300–7600 Early herders Increasing aridity with seasonality. Grassland cover of shrub-steppe and

savannah

EP2 6900–

6400

7800–7300

Middle Pastoral

(MP)

MP1 6100–

5500

7100–6200 Mature Neolithic

pastoralism

Strong seasonality with wet and dry spells, grasslands, rare tree cover and

shrub-steppe

MP2 5500–

5000

6400–5600

Late Pastoral (LP) LP1 5000–

4000

5900–4300 Specialized nomadic

herders

Hyperarid desert savannah with Vachellia and Tamarix associations

Main socio-cultural phases and eco-climatic features identified in the Takarkori area (modified, after [25, 28]). The calibrated dates express the maximum chronological

range and overlaps are statistically possible. For the calibration: OxCal online version 4.4.4 [31].

https://doi.org/10.1371/journal.pone.0310739.t001

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 5 / 40

https://doi.org/10.1371/journal.pone.0310739.t001
https://doi.org/10.1371/journal.pone.0310739


area beyond the rain dripline, outside of the vault cover. Due to their location outside the drip

line of the shelter and being affected by various natural and anthropogenic formation pro-

cesses, the deposits in SS and WS do not provide a clear stratigraphy allowing a solid archaeo-

botanical sampling [21, 25]. Therefore, this paper reports on the distribution pattern of the

archaeobotanical record from the MS and NS sectors.

Categories of Archaeological Contexts (ACs)

The classification system adopted here, describing the unique characteristics of archaeological

deposits in arid environments, considers the physical and spatial characteristics of the context,

the nature of the sediment, as well as the most frequent organic and inorganic components

(Table 2). It includes four main categories of Archaeological Contexts (ACs) proposed by [21]:

• Matrices: “the sedimentary by-products of human occupation, consisting of organic matter

(ovicaprid excrement, plant macromaterials, faunal specimens, etc.) mixed with sands”;

• Fixtures: “a large category, highly diversified, and with different functional meanings (. . .),

the Fixtures are the clear outcome of deliberate human intervention and are characterized by

an unambiguous functional attribution”;

• Burials: the rock shelter yielded the remains of 15 individuals of women and children [26].

The burials belong to different cultures, mainly of Pastoral age;

• Physiogenetic contexts: these layers mainly consist of surface aeolian sand, collapsed rocks,

altered bedrock.

Table 2. Main features of the Archaeological Contexts (ACs) at Takarkori.

Category of AC Type of AC Physical state Sediment

spatial configuration spatial consistency hardness texture

Matrixes M1—coarse sand diffuse low very low poorly sorted

M2—humified organic sand diffuse low low well sorted

M3—cemented organic sand clear high very high sorted

M4—organic sand diffuse low low sorted

M5—stone accumulation diffuse high low well sorted

M6—plant accumulation clear high low poorly sorted

M7—dung diffuse high high sorted

M8—floor diffuse high high sorted

Fixtures F1—pit (filling) clear high low poorly sorted

F2—post hole (filling) clear high low poorly sorted

F3—kettle clear high very high poorly sorted

F4—hearth clear high low sorted

F5—ash dump clear high high well sorted

F6—informal stone structure clear low - -

F7—formal stone structure clear high - -

Burials B—burial clear high - -

Physiogenetic Ph1—aeolian ingression diffuse low very low poorly sorted

Ph2—vault/wall collapse; altered bedrock clear low - -

Spatial configuration refers to the visibility/readability of ACs’ limits; spatial consistency defines the uniformity/homogeneity of sediment (modified, after [21]).

https://doi.org/10.1371/journal.pone.0310739.t002

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 6 / 40

https://doi.org/10.1371/journal.pone.0310739.t002
https://doi.org/10.1371/journal.pone.0310739


Sampling for plant macroremains

The plant materials were collected using three different types of sampling, named as follows:

• Sieve: sieving of all the excavated sediment from each stratigraphic unit using a 4-mm-diam-

eter mesh.

• Volumetric: sieving standard sediment volumes (5 liters) taken from each stratigraphic unit,

with a series of three overlapping meshes of decreasing diameter (10, 2 and 0.5 mm)

(Fig 2C).

• Spot: direct collection of particularly significant materials which include different ecofacts.

Moreover, they also include artifacts and relevant structural elements such as plant fiber

weavings or wooden poles. These ‘spot’ samples were systematically identified with a unique

code (ID) and recorded by Electronic Total Station (ETS).

For the purpose of this study the plant remains from Sieve and Volumetric sampling are

primarily included in statistical elaborations, while Spot sampling of accumulations of Seeds

[4], and handpicked Basketry/Cordage remains [32] are considered only as additional ele-

ments for the analysis of the spatial layout. Seeds coming from Sieve and Volumetric sampling

are considered when useful for the general interpretation and represent additional and unpub-

lished material.

All excavation squares were included in the sample collection (Fig 2D), confirming the reli-

ability of the sampling strategy.

Plant macroremain extraction and analysis

A systematic layer-by-layer sampling was conducted at the site, with a total number of at least

2000 samples collected in the field. Among them, 1112 samples were analyzed in detail for

their rich content in plant macroremains. Ecofacts and artifacts were manually extracted from

each sample, weighed with a digital scale, and measured in length with a digital calipers Mitu-

toyo Digimatic CD– 15D. The state of preservation was evaluated through a 5-step qualitative

scale (1: bad, 2: poor, 3: sufficient, 4: good, 5: excellent; from hardly identifiable to very well

preserved–mummified–record). Plant macroremains were observed with a stereomicroscope

Leica Wild M10 (magnification of 25x to 80x). Carpological and xylo-anthracological atlases,

keys and reference collections were used to identify remains with diagnostic characters [1, 19,

33–40]. Charcoal and wood ecofacts were identified through the analysis of wood anatomy,

observed on cross and longitudinal sections, or along fresh, hand-made fractures.

Intra-site spatial analysis

Intrasite spatial analyses aim at investigating patterns and relationships within and between

small units of observations across time and space [8, 17, 41]. We analyze here the spatial distri-

bution of archaeobotanical remains within several occupation horizons employing spatial sta-

tistical methods within the range of Point Pattern Analyses (PPA) [8, 42, 43], via native

algorithms both in R, and additional plugins in the QGIS environment (ver. 3.16.16).

For each occupation sub-phase and material class category selected (see further), we per-

formed a chi-squared test of complete spatial randomness (CSR) using quadrat counts test

(QCT) methods to assess the spatial pattering of each observation [14, 44].

To understand the directionality of the pattern evidenced by QCT (i.e., if the deviation

from the CSR is driven either by a clustering or dispersion pattern), we performed a distance-

based Point Pattern Analysis (PPA). For each sub-phase and each material category we per-

formed a Monte Carlo test for the point pattern. It generates randomly distributed points

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 7 / 40

https://doi.org/10.1371/journal.pone.0310739


within the study region and uses the Ripley’s K function for each set of generated simulated

points (100 iterations) and compares them to the K-function for the original set of points to

determine whether the point pattern deviates at a given distance from complete spatial ran-

domness (CSR) as null hypothesis.

The “spatstat” library [45] and “maptools” library in R (R Core team 2022) were used via

the QGis plugin interface as R dependencies, which allows the performance of statistical analy-

ses of spatial point patterns (SPPs) represented in 2D providing data on the intensity and type

of the pattern and the relationships among and within different types of point [45, 46].

To assess the actual location of the identified spatial pattern within the site, the spatial anal-

ysis of the botanical record was explored out using the “Kernel Density Estimation” (KDE)

plugin. It is a spatial interpolation technique that provides an estimate of the density of points

in a hypothetical continuous surface, producing clusters or areas with a high density of events.

This surface is obtained by cumulating the density functions for individual events, considering

for each of them a bandwidth (radius r of predefined value). In this study, we chose a radius of

r = 1 m. This selection considers the sampling strategy and the source of the materials used.

These plant remains mainly result from sifted sediment from each stratigraphic unit and are

associated with the centroid of the corresponding square. The reference values are represented

by a chromatic gradient that expresses density curves defined by the k function, which gives a

weight to the points that fall within the search radius [47, 48]. In this way, quality concentra-

tion maps (heatmaps) are produced, which allow us to visually ascertain the distributional

trend of the materials and their interaction both with other plant categories that make up the

record and with the main archaeological structures.

Given the palimpsest nature of the site and its pluri-stratigraphic layout, we consider here

materials assigned to each sub-phase as pertaining to a coarsely contemporaneous occupation

phase. Though inherent issues related to time-average phenomena, such approach has been

widely adopted in intra-site analysis for archaeological interpretation and supported by influ-

ential research [49–53].

Where spatial analysis revealed potentially significant distributions, the degree of associa-

tion between the specific categories of material was statistically measured by calculating Pear-

son’s correlation coefficient r. The analysis was carried out with Past© software (v. 4.03).

Results: The archaeobotanical record

The Takarkori archaeobotanical record here considered consists of a total of 20,540 specimens

of different nature (Table 3: 99% from ‘Sieve’ and “Volumetric” sampling, and 1% from ‘Spot’

samples).

Ecofacts

“Ecofacts” include a total of 20,419 elements (excluding samples from Seed Accumulations

[4]) that, based on their different botanical nature, were divided in the following 11 categories

(Fig 3):

• Charcoals: loose, scattered fragments of charred wood (size� 0.5 cm). In most cases, they

were fragments of Acacia/Vachellia (probably V. tortilis (Forssk.) Galasso & Banfi subsp.

raddiana (Savi) Kyal. & Boatwr.), sometimes also Tamarix sp. (probably T. aphylla (L.) H.

Karst.), more rarely Nerium oleander L., Cupressus sp. (probably C. dupreziana A.Camus) or

Ficus sp.

• Barks: fragments of rhytidome, or rind, of stems or branches of woody plants. They are

mostly not identifiable, but they are probably attributable to Acacia/Vachellia or Tamarix,

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 8 / 40

https://doi.org/10.1371/journal.pone.0310739


being the most frequent woody species in this area; only one fragment resembled Phoenix
dactylifera L.

• Grass stems: parts of caulis of Poaceae, clearly recognizable by their cylindrical hollow section

and the structure articulated in nodes and internodes; some stems of Cyperaceae were recog-

nized by the triangular section; in a few cases, the cylindrical stem, approximately 0.5–1 cm

diameter, with epidermis longitudinally striated and a dense spongy pith, was attributable to

species of Typhaceae family.

• Leaves: fragments of mummified leaves (rarely entire leaves); some of them were desiccated

leaves of Poaceae, recognizable for their typical narrow, elongated shape with apex acute and

parallel veins.

• Wood Sticks: splinters, sticks or small pieces of wood whose shape and section, usually flat-

tened and irregular, suggests they are the product of a break (not necessarily intentional).

• Roots: fragments of root systems, usually easily recognizable by their twisted and irregular

morphology and the presence of lateral rootlets. They were parts of the root system of Poa-

ceae, recognizable by their twisted shape and small diameter (1.5–2 mm maximum in the

thickest roots).

• Twigs: fragments of branching of woody plants, easily distinguishable by their circular or

sub-circular cross-section; rarely identifiable at the species level for their too small diameter,

in a few cases they were little branches of Acacia/Vachellia.

Table 3. Plant ecofacts and artifacts, per category and type of sampling.

Class Category Sieve/Volumetric Spot* Number of specimens

Ecofacts Charcoals 7155 2 7157

Barks 5328 3 5331

Grass stems 1297 2 1299

Leaves 7 - 7

Wooden Sticks 1832 22 1854

Roots 61 - 61

Twigs 2021 - 2021

Thorns 24 - 24

Inflorescences 10 1 11

Fruits 2631 1 2632

Seeds 22 30** 22

Total 20388 31 20419
Artifacts Cordage - 38 38

Basketry 4 49 53

Wooden tool - 25 25

Wooden pole - 5 5

Total 4 117 121
Grand total 20540

*spot samples of handpicked Basketry and Cordage remains [31]and accumulation of Seeds [4] are discussed elsewhere.

**number of specimens in Spot of Seed accumulations are not counted here, but available in [4].

Eleven categories of ecofacts and 4 categories of artifacts were identified (Table 3).

https://doi.org/10.1371/journal.pone.0310739.t003

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 9 / 40

https://doi.org/10.1371/journal.pone.0310739.t003
https://doi.org/10.1371/journal.pone.0310739


Fig 3. Some examples of ecofacts and chronology. A) Twig (MP2); B) Bundle of grass partially burned on one side (LA3);

C) Stems and sticks (LA2); D) Intact Balanites aegyptiaca (L.) Delile fruit (LP1) and endocarp (LA3); E) Charcoals from a

fireplace (MP2); F) Leaves still in place (MP2); G) (LA3) and H) (EP1) Wood remains with bark (photos ©The

Archaeological Mission in the Sahara).

https://doi.org/10.1371/journal.pone.0310739.g003

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 10 / 40

https://doi.org/10.1371/journal.pone.0310739.g003
https://doi.org/10.1371/journal.pone.0310739


• Thorns: sharp elements with enlarged base and pointy apex, made of sclerified tissues,

belonging to epidermal tissues of woody plants; some of them may have been lost from the

barks of trunks or branches of the spiny acacia trees.

• Inflorescences: groups of flowers or spikelets of Poaceae (Paniceae, Andropogoneae), borne

by the same stem, which is often divided into several branches; in few cases, the inflores-

cences were identified as Pennisetum sp.

• Fruits: fruits of angiosperms, nearly always fleshy (drupes). Fruits of Balanites aegyptiaca (L.)

Delile were common, largely being well preserved endocarps, sometimes still maintaining

the fruit flesh; fragmented pods of Fabaceae (possibly Cassia, in a few cases Acacia/Vachel-
lia), some drupes resembling those of Celtis and a few syconia from Ficus were also observed.

A single case of gymnosperm, included in this category for simplicity, is the strobilus of

Cupressus dupreziana A.Camus. Very few fruits of Phoenix dactylifera L. with preserved peri-

carp and esocarp were also found.

• Seeds: seeds from Sieve and Volumetric sampling belonging to different angiosperm species;

among them, Acacia/Vachellia, and Citrullus (C. colocynthis (L.) Schrad.), and also fragments

of P. dactylifera.

Artifacts

This class includes four categories, totaling 121 elements, often in an exceptional state of pres-

ervation (Fig 4). Their low numerical incidence does not allow for a statistically relevant spatial

examination, but they have high functional/interpretive value:

• Basketry: includes plant fiber artifacts, usually stems of herbaceous plants, twisted or inter-

woven into more or less long sections, thus defining a recognizable weave (Fig 4A and 4B).

• Cordage: includes elements composed of filaments of heterogeneous materials (herb stems,

bark and animal fibers), woven or twisted together (Fig 4C and 4D).

• Wooden pole: wooden structural element with a supporting function, well recognizable by

the diameter of the rounded section and the development in the longitudinal direction

(Fig 4E).

• Wooden tool: includes a range of worked wooden tools, clearly recognizable as such, of vari-

ous shapes and sizes (e.g., hooks, points, and spatulas) (Fig 4F).

Results: The spatial analysis

Stratigraphic distribution and spatial patterning of the archaeobotanical

record

Charcoals is the most widespread category (35% out of total), with close relationship with fire

installations and fire related activities, a complex topic that will be subject of a forthcoming

study.

The remaining most frequent classes of archaeobotanical remains are Barks (26% of total),

Fruits (13% of total), Sticks (9%) and Twigs (10% of total) which together make up ca. 58% of

the record (Table 4). Their ubiquity and frequency, combined with their inherent interpreta-

tive potential, make them the four categories under scrutiny here.

The high rate of frequentation of the site, though with chronological trends and differences,

is directly related to the abundance of plant remains. As expected, much of plant remains

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 11 / 40

https://doi.org/10.1371/journal.pone.0310739


Fig 4. Some examples of artifacts. A) and B) basketry remains (LA3); C) (MP1) and D) LA3) cordage; E) wooden pole (LP1); F) a wooden point (LA3)

(photos ©The Archaeological Mission in the Sahara).

https://doi.org/10.1371/journal.pone.0310739.g004

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 12 / 40

https://doi.org/10.1371/journal.pone.0310739.g004
https://doi.org/10.1371/journal.pone.0310739


comes from Matrices, i.e., sands with abundant organic materials, regardless of their socio-cul-

tural context, followed by Fixtures while the other ACs host less records (Table 5).

Matrices represent the most common type of Archaeological Context found in the site and

may function as good sedimentary traps where organic remains are retained after the deposi-

tional event. Interesting is also the association of ecofacts with Fixtures, which by representing

“structures évidentes” [54], may be related to activities directly linked to their use or functional

domain [21, 55], either as de facto, primary or secondary refuse [56, 57]. Functional

Table 4. Number of plant remains, by cultural phase.

Class Category Chronological sub-phases Total

LA1 LA2 LA3 EP1 EP2 MP1 MP2 LP1

Ecofacts Charcoals 190 988 1921 1578 664 171 1325 320 7157

Barks 323 1565 1131 525 675 117 886 109 5331

Culms of grasses 50 292 319 231 155 21 211 20 1299

Leaves 0 1 3 0 0 0 3 0 7

Fruits 19 178 464 414 482 69 888 118 2632

Inflorescences 1 5 1 2 0 0 2 0 11

Wood Sticks 79 587 524 223 172 35 211 23 1854

Roots 1 3 9 7 11 1 27 2 61

Twigs 117 503 481 305 210 43 314 48 2021

Seeds 0 0 1 3 8 2 8 0 22

Thorns 1 18 0 2 0 0 2 1 24

Artifacts Cordage 1 6 10 5 5 10 1 0 38

Basketry 4 14 25 5 0 4 1 0 53

Wooden artifacts 1 3 10 4 1 0 6 0 25

Wooden pole 1 1 0 1 1 0 0 1 5

TOTAL 788 4164 4899 3305 2384 473 3885 642 20540

https://doi.org/10.1371/journal.pone.0310739.t004

Table 5. Percentage of plant remains, by Archaeological Contexts (ACs).

Class Category* Archaeological Context

Matrices Fixtures Burials Physiogenetic

Ecofacts Charcoals 71.1 28.8 0 0.1

Barks 81.5 17.7 0 0.7

Culms of grasses 82.2 17.3 0 0.5

Leaves 100.0 0 0 0

Fruits 82.3 17.7 0 0

Inflorescences 54.6 45.4 0 0

Sticks 78.6 21.0 0.1 0.3

Roots 93.4 6.6 0 0

Twigs 83.5 15.4 0 1.1

Seeds 81.8 18.2 0 0

Thorns 91.7 8.3 0 0

Artifacts Cordage 64.5 35.5 0 0

Basketry 74.3 22.3 2.8 0

Wooden tool 88.0 12.0 0 0

Wooden pole 66.7 33.3 0 0

*in italics the four categories considered for geostatistical and spatial analysis in this study.

https://doi.org/10.1371/journal.pone.0310739.t005

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 13 / 40

https://doi.org/10.1371/journal.pone.0310739.t004
https://doi.org/10.1371/journal.pone.0310739.t005
https://doi.org/10.1371/journal.pone.0310739


interpretation of these distribution depends on contextual analysis and variables, for which a

systematic description is needed and accomplished through the interpolation of the spatial

analysis results and an ad hoc interpretative framework [15, 16, 42, 58, 59].

Chrono-cultural spatial patterns of selected archaeobotanical remains

The bulk of archaeobotanical remains (Barks, Fruits, Sticks and Twigs) predominantly origi-

nate from the Late Acacus phase (Fig 5). This is attributed to the nature of occupation, which

has been interpreted as semi-residential and characterized by prolonged periods of occupation

[21]. A steady decrease in total raw frequency of plant remains is the trend from the Early Pas-

toral onwards among all categories except for Fruits, which show an inverse trend. Indeed, a

second peak in total frequency is recorded in the Middle Pastoral 2 sub-phase, driven mostly

by the great number of Fruits which duplicate in comparison to all the preceding sub-phases.

The material remains of the Late Acacus 1 sub-phase (LA1) were brough to light only in the

Northern Sector. Among the categories under study (S1 Table), Barks and Twigs are the most

represented, followed by Sticks and Fruits. Most of the material comes from matrix layers and

to a lesser extent from some fixtures, 15% of the Sticks from hearths (F4) and 25% of the Barks

from formal structures (F7).

The archaeological deposit of the Late Acacus 2 (LA2) is present in all the excavated sectors,

as mirrored by the large quantity of plant remains (> 2800 items). Barks and Sticks were the

most represented, largely coming from matrix-type layers, especially organic sands (M4) and

floors (M8).

Late Acacus 3 (LA3) yielded also a high amount of material (2600). Barks is the most com-

mon, while Sticks, Twigs and Fruits show similar amounts. Most of the ecofacts (an average of

66%) come from organic sands (M4), followed by ash accumulations (F5) (S1 Table).

Less than two thousand ecofacts comes from the Early Pastoral 1 (EP1) archaeological con-

texts, mainly organic sands (M2 and M4) and to a lesser extent from some fixtures, such as for-

mal stone structures (F7) (S1 Table).

About 1400 remains come from the Early Pastoral 2 (EP2), with frequencies fairly similar to

the previous EP1, both quantitatively and stratigraphically: however, Fruits are more common.

The M4 organic sand represents the major depositional context (Table 5).

266 plant remains come from the Middle Pastoral 1 (MP1) sub-phase, an exiguity largely

due to post-depositional phenomena. Most of them come from organic sands (M4), followed

by some structural elements, particularly informal structures (F6) (S1 Table).

The Middle Pastoral 2 (MP2) accounts for about 2200 plant remains, with a particular high

frequency of Fruits and Barks. The stratigraphic context of MP2 is made almost exclusively of

organic sand layers, which together with fire-installation, i.e., hearths, are the main features of

the occupation layers (S1 Table).

Slightly less than 300 remains come from the Late Pastoral 1 (LP1), with Barks and Fruits

more numerous. The stratigraphic context is mostly represented by organic sands (M4),

although structural elements are also common (S1 Table).

The deviation from Complete Spatial Randomness CSR observed in the Quadrat Counts

Test QCT analysis for the spatial distribution of botanical remains could potentially be attrib-

uted to various biases introduced during the excavation process. However, the consistent spa-

tial behavior differing from CSR and the presence of different directionality, as observed

through both QCT and SPP (Spatial Point Pattern), suggest that there may be causal factors at

play, including primary and second-order effects.

Specifically, for the Late Acacus horizons, a strong deviation from CSR is observed for

Barks, Sticks, and Twigs (S2 Table). This indicates a non-random distribution and clustering

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 14 / 40

https://doi.org/10.1371/journal.pone.0310739


Fig 5. Percentage and trends of selected ecofacts. A) Selected ecofacts (Barks, Fruits, Sticks and Twigs) by sub-phases

and B) divided by main ACs.

https://doi.org/10.1371/journal.pone.0310739.g005

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 15 / 40

https://doi.org/10.1371/journal.pone.0310739.g005
https://doi.org/10.1371/journal.pone.0310739


of these materials within the sub-phases. On the other hand, in the Middle Pastoral layers,

there is a greater susceptibility to deviations from CSR for Fruits (S1 Fig; S2 Table).

These findings suggest that there may be underlying factors influencing the distribution of

materials in the archaeological context being studied and the nature and causes of it are further

addressed by a contextual approach.

Organization and use of the space

Activities and functions: An interpretative framework

The plant materials here observed could have been used in different ways and differently

entered the archaeological record, hence indicating different types of activities and use of

spaces. Nevertheless, based on the archaeobotanical and ethnobotanical research carried out as

part of archaeological research at Holocene sites in the region and other chronologically com-

parable contexts [60–66], some broad functional categories can be hypothesized (Table 6). Pos-

sible plant uses can be identified in what concerns the realm of dwelling structure and

habitation and living space, from their construction to maintenance and dismission. Wood,

together with leaves, is used as the main structural element in regional vernacular architecture,

as documented by ethnographic and ethnobotanical research in North Africa (e.g., among the

ethnic groups Tebu in the Southern Sahara and Tuareg in the Central Sahara; [67–70]).

Barks and Wood sticks can be associated as structural elements with functions of support or

reinforcement. Bark roofing, paneling and floor installments are recorded quite diffusely in

vernacular architecture both archaeologically [71–73], and in ethnobotanical accounts. More-

over, barks can be processed to extract tannins, saponins, lipids and antimalarian compounds

[74, 75].

Table 6. Selected categories of plant remains and their possible uses, based on ethnobotanical and ethnographic

research.

Ecofact categories Possible/Attested use

Barks building material

bracing elements

floor installation

tannin extraction

phytomedicine

paneling

roofing

fuel

Fruits/seeds/flowers food

fodder

medicinal use

oil extraction

Wooden Sticks building material

bracing elements

fuel

Twigs building material

bracing elements

roofing huts

bedding

food preparation (e.g., stirring etc.)

food enhancement (e.g., tamarix salt)

https://doi.org/10.1371/journal.pone.0310739.t006

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 16 / 40

https://doi.org/10.1371/journal.pone.0310739.t006
https://doi.org/10.1371/journal.pone.0310739


Cordage, when associated to Barks, Twigs, or Wood sticks, can be indicative of its use as

binding device for erecting hut walls, panels, windbreaks or enclosures [32, 76].

Twigs or branches can be associated with uses in the domestic sphere, such as bedding, pro-

cessing waste and fuel, or roof coverings for huts [3, 62, 65]. This use as building material is

common among the Tuareg kel Tadrart: wood branches of tamarisk species are used for the

walls and internal divisions, whereas acacia branches are employed for the hut cover [77].

Also, plant bedding is widely documented in ethnographic accounts whereas more difficult

and painstaking is its archaeological preservation and identification [78].

Leaf arrangements, matting, and similar devices have been identified in hunter-gatherer

archaeological sites through micromorphological approaches, though occurrences of unam-

biguous macro-remains are rare [62, 78–80].

In the Sahara, present-day desert populations such as the Tuareg Kel Tadrart still use these

categories of ecofacts [77], and adopt the ‘multipurpose approach’ on exploiting the same

plant species for food, medicine and other uses [65]. Pollen accumulations as well as the tech-

nological analysis of basketry remains also testify to the skilled transport of inflorescences and

grass stems at Takarkori [25, 32]. An exceptionally impressive find is the documented use of a

bundle of grass stems, partially burnt, in connection with a fireplace (Fig 3B).

Fruits are among the most common archaeobotanical remains that are prevalently inter-

preted as food evidence, as accumulation, storage or votive offer depending on the specific

context. As seen above, the record of Takarkori includes many drupes of Balanites aegyptiaca.

This tree species, typical of the arid zones of the Sahara and the Sahel, is used today by the Tua-

reg Kel Tadrart for food, as well as fodder for camels and goats: in the dry season, animals can

also consume its leaves and branches [66]. In many parts of Africa, the Middle East and the

Indian subcontinent, this is a known oil plant and different plant organs, from fruits to roots,

are used for the treatment of some ailments, whose efficacy is corroborated by phytochemical

research [81]. Fruits, seeds and charcoals of the desert date have been recorded in the Tadrart

Acacus [79], and are also reported from Egyptian predynastic settlements and cemeteries [e.g.

82, 83 and references therein]. Remarkable is also the presence of seeds of Citrullus colocynthis,
a viny plant spread even today on desert sands. The plant has known purgative effects, and it is

a vermifuge. The dried pulp of the peeled fruit without seeds is used also for other medicinal

purposes. The powder is irritating to mucous membranes and is usually administered mixed

with other drugs on account of its griping action [84].

Within these frames of references and analogies (Table 6), and while still considering cau-

tionary tales, an interpretative proposition regarding the arrangement of space and activities

within the site is here proposed.

Space and activities during the Late Acacus

Late Acacus 1 (10,200–9400 years cal BP). The LA1 horizon, relating to the first occupa-

tion of the site by hunter-gatherer-fisher groups, was intercepted exclusively in the Northern

Sector (Fig 6). This area is characterized by the presence of important structural elements such

as a formal stone structure composed of large slabs, likely associated with two hearths on the

east towards the shelter wall. The interpretation of the structure as a hut is supported by the

high number of Barks some of large size, which together with the presence of an actual wooden

post are interpreted as the remains of one or more support posts LA1_CL1, (Table 7).

Dwelling structures realized with the use of light wooden structures in various arrange-

ments have been hypothesized at some archaeological sites, suggesting the employment of

poles made of branches or tree roots (often of Acacia/Vachellia, or Tamarix sp.) [85]. The pres-

ence and distribution of Twigs around the structure may be related to the remains of some

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 17 / 40

https://doi.org/10.1371/journal.pone.0310739


kind of installation with covering function, hypothesizing a composite framework made of

trees and shrubs. The clustering of this material category only in this part of the quadrants

may moreover corroborate such reconstruction, just in perpendicular location where the

vault/roof would have been located.

All Fruits are strictly outside of the structure, in the area between the hearths and the shelter

wall, and many of them have a fragmentary state of conservation. Albeit with low numbers,

they could define a processing area of food resources, LA1_CL2. The negative correlations

between Fruits and the other ecofacts (Barks, Sticks and Twigs; S3 Table) would confirm that

the fruits were concentrated in a specific area of food processing.

Late Acacus 2 (9500–8600 years cal BP). The prolonged, semi-residential occupation, as

well as advanced forms of resource management of LA2 horizon are reflected in the frequency

and distribution of the archaeobotanical record under study (Fig 7). The analyses allowed the

identification of an area of high density of remains in the Northern Sector (LA2_CL1) in asso-

ciation with a structure interpreted as a hut/dwelling structure. The frequency of Barks and

Sticks suggest their use as building material (the perishable walls?), or as roofing supported by

the remaining pole (Fig 8). Their distribution follows the almost circular outline of the stone

Fig 6. LA1 Kernel Density Estimation. LA1 Kernel Density Estimation of selected ecofact categories (Barks, Fruits, Sticks and Twigs), with plotting of relevant

artifacts discussed in the text.

https://doi.org/10.1371/journal.pone.0310739.g006

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 18 / 40

https://doi.org/10.1371/journal.pone.0310739.g006
https://doi.org/10.1371/journal.pone.0310739


arrangement, indicating its spatial and probable structural association. Pearson’s r coefficient

confirms the Barks-Sticks positive association with a particularly high value (S3 Table), sup-

porting an intended, joint use of these ecofacts, presumably as part of the same wood

materials.

In this same area (LA2_CL2), the presence of Fruits and various artifacts, especially bas-

ketry and cordage, may define an area for processing activities. Millet (Panicum sp., including

P. laetum Kunth) and sorghum (Sorghum sp., namely S. bicolor (L.) Moench), for instance,

were systematically harvested, cultivated and stored inside baskets [4, 32], or in shallow con-

cavities dug into the sandstone bedrock, the so-called kettles [86]. The ethnobotanical and

technological study on basketry and cordage clarified the type of use of these artifacts as

Table 7. Spatial clusters by chronological sub-phases.

Chronological

sub-phase

Cluster ID Cluster location and description Main ACs and ID Ecofacts+ artifacts Spatio-functional

interpretation

Late Acacus 1 LA1_CL1 NS. High number of Barks (81 fragments, ca. 25% out

of a total of 323)

F7 –formal stone

structure (398)

Barks, Twigs,
Wooden post

Hut with plant roofing

sustained by wooden posts

LA1_CL2 NS. Fragmented Fruits strictly outside of the structure,

between the hearths and the shelter wall

F4—hearth (393, 354) Fruits Processing area of food

resources

Late Acacus 2 LA2_CL1 NS. High frequency of Barks and Sticks. Their

distribution follows the circular outline of the stone

arrangement

F7—formal stone

structure (374)

Barks (329 items, ca.

20% out of a total of

1565)

Sticks (82 items, ca.

14% out of a total of

587).

Wooden pole

Hut with perishable wall

and roofing

LA2_ CL2 NS. High presence of Fruits and artifacts, especially

basketry and cordage

F4-hearth (378)

M4-organic sand (376,

372, 379, 383)

M8-floor (384, 387)

Fruits (37 items, ca.

20% out of a total of

178)

Basketry

Food processing area;

Outside activity area with

rug or mat lined floors

LA2_CL3 Located in the MS, it associates with formal stone

structures

F7—formal stone

structure (174);

F1—pit (filling) (220,

228), F3 (229, 240, 242)

Fruits Storage and processing

area

LA2_CL4 Located in the west-central area of MS, perimetral to F7 F7—formal stone

structure (180)

Barks and Sticks Wooden fence with basal

stone inlay

Late Acacus 3 LA3_CL1 MS. Several artifacts found in a defined zone between

the shelter wall and the other clusters

F5-ash dump (163)

F6-informal stone

structure (151)

M8-floor (107)

Basketry (13)

Cordage (12)

Wooden tools (9)

Multifunctional area

LA3_CL2 MS. Located south of the main structure F7-131. Barks
are more frequent

F7—formal stone

structure (131)

Barks
Sticks
Twigs

Roofing

Walling

Hut

LA3_CL3 Center of MS F4 –hearth (71, 124,

197, 152)

F5-ash dump (249)

F6—informal stone

structure (261, 151)

M4-organic sand

(103,69, 132)

M8-floor (107)

Barks
Sticks
Twigs

Roofing

Walling

Structure latent

LA3_CL4 Located in the MS southern part. Characterized by fire-

related structures and ash accumulations with many

Fruits

F1- pit (filling) (99)

F4-hearth (97, 134)

F5-ash dump (145,

138, 121, 168, 287, 267,

163)

F6- informal stone

structure (153, 154)

Fruits 139 remains, ca

30% out of a total of

464

Discard/for secondary

refuse management

(Continued)

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 19 / 40

https://doi.org/10.1371/journal.pone.0310739


remains of containers made of woven plant fibers, used for the storage of wild cereals [32].

However, given their spread presence across the site, the interpretation of these elements as

parts of woven mats, probably connected to dwelling areas and implicated in the preparation

Table 7. (Continued)

Chronological

sub-phase

Cluster ID Cluster location and description Main ACs and ID Ecofacts+ artifacts Spatio-functional

interpretation

Early Pastoral 1 EP1_CL1 MS shelter wall F7—formal stone

structure (68–67)

F4 –hearth (88–92)

F1 - (91)

Barks
Basketry

Cordage

Wooden tool

Informal dwelling area

EP1_CL2 MS southwestern part.

Associated with structural/formal evidence, and a

wooden post.

Barks, Twigs, and Sticks show similar accumulation

densities.

Fruits form a well-defined cluster

F1 –pit (filling) (85–

86)

F4—hearth (81)

F5—ash dump (80)

Barks
Twigs
Sticks
Wooden post

Fruits (76 ecofacts

-14% of the total).

Processing area

Food/Resource storage

area (?)

EP1_CL3 Southern part of MS.

Overlapping clusters of Barks, Sticks, and Twigs.
Fruits show a richer and denser cluster

F1-pit (269)

M2 –humified organic

sand (38)

Barks
Sticks
Twigs
Fruits (102 items

-25% of the total).

Processing

Storage

Maintenance

EP1_CL4 Near the shelter wall in MS in proximity of burial H6 B- burial (H6)

(proximity)

Barks
Twigs
Sticks

Burial coating

Inhumation wrapping

Early Pastoral 2 EP2_CL1 MS northeastern part. F4—hearth (73) Barks
Wooden post

Sticks

Wooden structure (hut/

windbreak)

EP2_CL2 Location in the MS inner corner. Cluster with high

density of fruits in low state of preservation associated

with concentrations of Sticks and Twigs.

F1—pit (filling) (62,

64, 87, 244)

F4 –hearth (61, 63)

F5 –ash dump (59, 57,

60–270).

Fruits (118 items; 8%

of the total)

Sticks
Twigs

Secondary refuse

management

Middle Pastoral 1 MP1_CL1 NS central-eastern area. F6—informal stone

structure (360–361)

Barks, Fruits
Basketry, Sticks
Twigs

Temporary shelter

MP1_CL2 MS southeastern part. Clusters of varying intensities of

different remains.

F6—informal stone

structure (246)

Fruits (51%)

Barks
Sticks
Twigs

Secondary refuse

management

Middle Pastoral 2 MP2_CL1 MS southern area F4 –hearth (233, 231,

209, 198, 204, 206, 194)

F5 –ash dump (207,

243, 227),

Barks
Twigs
Wooden tools

Basketry

Small animal enclosure

MP2_CL2 MS southern area F6—informal stone

structure (33)

Fruits
Sticks

Animal management

/Fodder

Late Pastoral 1 LP1_CL1 MS northeastern part.

Gradually decreasing towards south. Hotspot in

association with fixtures.

F7 –formal stone

structure (21),

F1 –pit (filling) (29, 23,

19, 16),

F2—post hole (filling)

(18)

F4 –(24, 26, 8, 5, 4)

Barks
Sticks
Twigs

Multifunctional area.

Food processing

outside dwelling

LP1_CL2 MS central area. towards the shelter wall. M7- dung (6)

F2- post hole (filling)

(18)

Barks
Sticks
Twigs
Fruits
Wooden pole

Fenced area for animal

management// fodder

https://doi.org/10.1371/journal.pone.0310739.t007

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 20 / 40

https://doi.org/10.1371/journal.pone.0310739.t007
https://doi.org/10.1371/journal.pone.0310739


Fig 7. LA2 Kernel Density Estimation. Kernel Density Estimation of selected ecofacts (barks, fruits, sticks and twigs) in LA2, with

plotting of relevant artifacts discussed in the text.

https://doi.org/10.1371/journal.pone.0310739.g007

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 21 / 40

https://doi.org/10.1371/journal.pone.0310739.g007
https://doi.org/10.1371/journal.pone.0310739


of bedding or floor systematization/levelling through matting and rug-like furnishings can-

not be ruled out [62, 78]. In the Main Sector, a space with possible storage and processing

purposes was evidenced by well-preserved fruits and can be identified in the central area

close the shelter wall, LA2_CL3, where also formal stone structures, pits, and kettles are

found.

The central-southern zone of the Main Sector is bordered to the west and south by the pres-

ence of a rather extensive fire-related area likely connected to the main domestic-dwelling

unit, characterized by high maintenance given the dearth of remains. In the central-western

area, the formal stone structure (F7–180) was interpreted as an area for Barbary sheep (Ammo-
tragus lervia) corralling, based on the abundance of fodder and animal coprolites [87]. This

interpretation is now further supported by the abundance of Barks and Sticks around the

perimeter of the structure, and LA2_CL4 might represent ephemeral remains of the perishable

parts of a wooden fence reinforced with stones at the base. Examples of such structures can be

found in some of the ethnographic contexts investigated by [76], as part of their study of the

continuity between prehistoric and historic animal management practices. The spatial visual

correspondence between Barks and Twigs is supported by high r-value (S3 Table), indicating a

positive correlation between these materials. A good correlation was also found for the associa-

tion Barks-Twigs and Sticks (S3 Table).

These agglomerations represent different functional zones within the archaeological site,

each with specific features and potential purposes related to resource management, dwelling,

processing activities, and animal management.

Fig 8. Structural and archaeobotanical features from LA2 horizon. A) close up of the wooden pole (bottom) and twigs associated with the B) hut-like

structure F7-374; and C) a fireplace with still in place preserved twigs and sticks.

https://doi.org/10.1371/journal.pone.0310739.g008

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 22 / 40

https://doi.org/10.1371/journal.pone.0310739.g008
https://doi.org/10.1371/journal.pone.0310739


Late Acacus 3 (9000–8000 years cal BP). The LA3 occupation levels, characterized by

several Fixtures, especially hearths and ash accumulation, are among the richest of plant

remains with more than 80% from the central-southern and western parts of the Main Sector,

leaving the northern and north-central space almost free. Several artifacts, LA3_CL1, found

between the shelter wall and the clusters of ecofacts (Fig 9), are part of a large multifunctional

area with external activities linked to a dwelling structure, a hut, 131, where LA3_CL2 is

localized.

The KDE highlights a clustering in the center of the Main Sector, LA3_CL3, where the spa-

tial outline may hint to the presence of some kind of “structures latentes”, sensu Leroi-Gour-

han. It is tempting to recognize in those characters the residues of a more formal structure

resembling in fact hut-like features, disrupted and whose building materials could have been

reused later. This same spatial layout is clearer for the actual hut structure 131, where we how-

ever observe different conservation dynamics and formation process.

Huts roofed by tamarisk branches and lined inside by mats have been hypothesized for

other contemporary sites like the structure 1/90 at site E-75-6 in the Nabta Playa basin [88]

and the already aforementioned Sudanese examples [85]. Typha- roofing for a hut like struc-

ture has been advanced for the remains unearthed at Uan Tabu [65, 89].

The stones arranged in a circular fashion of structure 131 may have functioned as structural

reinforcement and bracing for the lightweight structure of wooden material of which the few

remains in question here are left. The almost clean internal surface, void of macro-remains but

for few small evidences, should be considered the outcome of cleaning activities as specified by

various intra-site analyses [55, 90]. Bigger elements are usually thrown out whereas smaller

ones are projected towards the liminal area, walls edges etc., or embedded in the earthen floor

further compacted through successive trampling or “under the rug/mat” effects [78, 91–93]

and testified also by the widely planar micromorphological structure of soils sampled in the

site itself [25].

LA3_CL4, another area of interest in the southern part of the Main Sector, is characterized

by numerous fire-related structures, especially ash accumulations which could be assigned to a

discard maintenance area, for secondary refuse management (sensu [94]) as hinted by the fact

that about 1/3 of the fruits in this area came from levels of ash accumulations and are either

fragmentary or partially preserved.

These agglomerations represent various functional and structural aspects of the archaeolog-

ical site, including dwelling areas, potential structures, and refuse management areas.

Space and activities during the Early Pastoral

Early Pastoral 1 (8300–7600 years cal BP). The earliest horizon of the Early Pastoral

phase is characterized by important cultural changes [95]. The EP1 occupation shows stone

structures and enclosure, as well as the presence of burials in close interaction with the habita-

tion spaces [26]. The distribution pattern of plant remains in EP1 markedly differs from the

previous sub-phases, with relatively very poor remains (Fig 10). Although post-depositional

processes may in part explain this configuration–in particular the humification of organic mat-

ter–the presence of well-defined clusters associated with fixtures suggests specific uses of the

shelter. In detail, only a low quantity of Barks and Twigs comes from the Northern Sector,

whereas some major clusters are visible in the Main Sector.

In this context, the analysis of plant remains and artifacts define three main areas of inter-

est, respectively, in the northeastern, southwestern, and southern parts of the Main Sector. The

first cluster, EP1_CL1, is close to the shelter’s wall where several fixtures are located: a concen-

tration of Barks in association with formal structures, likely remains of the perishable part of a

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 23 / 40

https://doi.org/10.1371/journal.pone.0310739


Fig 9. LA3 Kernel Density Estimation. Kernel Density Estimation of selected ecofacts (barks, fruits, sticks and twigs) in LA3,

with plotting of relevant artifacts discussed in the text and indication of clusters specified in Table 7.

https://doi.org/10.1371/journal.pone.0310739.g009

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 24 / 40

https://doi.org/10.1371/journal.pone.0310739.g009
https://doi.org/10.1371/journal.pone.0310739


Fig 10. EP1 Kernel Density Estimation. Kernel Density Estimation of selected ecofacts (barks, fruits, sticks and twigs) in EP1, with plotting of

relevant artifacts discussed in the text and indication of clusters specified in Table 7.

https://doi.org/10.1371/journal.pone.0310739.g010

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 25 / 40

https://doi.org/10.1371/journal.pone.0310739.g010
https://doi.org/10.1371/journal.pone.0310739


dwelling structure, whereas the other plant categories are rather negligible or null in amount.

Fruits cluster in the northeastern corner of this area. The presence of a few artifacts and struc-

tural elements such as two hearths and one pit could define an area related to the domestic

sphere.

The second area of interest is in the southwestern part of the Main Sector, EP1_CL2, in asso-

ciation with the presence of some structural/formal evidence: two pits, one hearth, an ash accu-

mulation, and a wooden post where Barks, Twigs and Sticks show similar accumulation

densities. Here Fruits—only Balanites aegyptiaca—form a well-defined cluster. These desert

dates show different state of preservation: some fruits still fleshy and other with only stones,

some of them broken. This instance suggests food processing activities, probably to extract oil.

Their association with pits could be related to the presence of some containers perhaps of per-

ishable nature. At Takarkori, numerous grinding tools have been uncovered, yet no direct evi-

dence link them to oil extraction [21, 96]. Combined molecular and isotopic techniques in

pottery organic residue analysis revealed the presence of diagnostic plant lipids, including leaf

waxes and seed oils, suggesting the processing of grasses, seeds and aquatic plants at the site [5].

The third area of interest is in the southern part of the Main Sector, EP1_CL3: Barks, Sticks,

and Twigs form largely overlapping clusters, while Fruits again show a richer and denser

cluster.

A marked cluster, shared by all the categories here considered, is in the vicinity of Burial

H6, EP1_CL4.

Early Pastoral 2 (7800–7300 years cal BP). The EP2 cultural horizon shows a degree of

continuity with the previous sub-phase. Two main areas are defined, located in the Main Sec-

tor (Fig 11).

The first, EP2_CL1, in the northeastern part, features a cluster of Barks associated with a

wooden post; together with a second cluster of Barks and a minor cluster of Sticks, a couple of

meters south: these remains could represent the ephemeral evidence of wooden structure,

either a hut or a windbreak. The presence of a hearth and minor concentrations of fruits sup-

port the interpretation of this area as a kind of dwelling context.

A second area, EP2_CL2, is of interest in the southern portion, between the shelter wall

and fixtures such as pits ash accumulations, and hearths. The position in the inner corner of

the shelter, together with the presence of fruits in low state of preservation partly associated

with an ash dump as well as concentrations of sticks and twigs, could define a place for second-

ary refuse management.

Space and activities during the Middle Pastoral

Middle Pastoral 1 (7100–6200 years cal BP). The MP1 sub-phase is characterized by a

more marked climate seasonality [97]. Compared with the Early Pastoral, we record major cul-

tural discontinuities, expressed in particular by the exploitation of cattle secondary products

and the full maturity of herding management [98]. These changes are reflected in the different

setting of the shelter, characterized by the presence of small hearths, informal structures (F6),

and burials (B). We record in this subphase an abrupt decrease in the number of plant remains

(<2% out of the overall total) (S1 Table). Without again dismissing post-depositional erosional

phenomena, this decrease can be also explained by the change in the shelter use, characterized

now by a discontinuous and more seasonal occupation [25]. Plant remains mainly cluster in

the central part of the Northern Sector and in the southeastern corner of the Main Sector. The

central-eastern area, instead, close to the shelter wall, houses the burial area of two individuals

where basketry remains could be cautionary interpreted as related to the layout and/or prepa-

ration of the burial [26].

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 26 / 40

https://doi.org/10.1371/journal.pone.0310739


Fig 11. EP2 Kernel Density Estimation. Kernel Density Estimation of selected ecofacts (barks, fruits, sticks and twigs) in EP2, with plotting of

relevant artifacts discussed in the text and indication of clusters specified in Table 7.

https://doi.org/10.1371/journal.pone.0310739.g011

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 27 / 40

https://doi.org/10.1371/journal.pone.0310739.g011
https://doi.org/10.1371/journal.pone.0310739


In the Northern Sector, the area occupied by some stone structures overlaps with the main

clusters of plant remains, MP1_CL1 helping to define a temporary shelter for short-time visi-

tors (Fig 12). In the southeastern portion of the Main Sector, it is possible to identify another

area of aggregation, MP1_CL2, characterized by the presence of different remains associated

with an informal structure: likely an area used for secondary refuse management (Fig 12). The

correlation between Sticks and Fruits is slightly higher than that between Sticks and Barks

(S3 Table).

Middle Pastoral 2 (6400–5600 years cal BP). The shelter in this sub-phase is mainly used

in the southern portion of the Main Sector, where numerous hearths and ash accumulations

are preserved (Fig 13). Fruits and Barks are here particularly abundant, followed by Twigs and

Sticks. Barks and Twigs show a very similar distributional pattern, with higher density toward

the outer edge of the area in question, while Fruits tend to be more concentrated along the

southern wall of the shelter.

In fact, plant remains are concentrated in the southern area of the Main Sector, MP2_CL1,

characterized by the presence of numerous hearths and ash accumulations, and a few artefacts

on perishable materials (six wooden tools and some basketry remains). Differently from Fruits

and Sticks, Barks and Twigs show a density peak immediately south of F6–33, an informal

stone structure, also testified by the significant positive correlation between them, much higher

than Fruits and Barks (S3 Table) MP2_CL2.

In a herding framework, such a distribution in the innermost part of the shelter, between

the eastern and southern walls, together with the presence of small hearths interpreted as “cor-

ral fires” [99], suggests the presence of an animal management area: the plant remains, such as

Barks, Sticks, and Twigs could represent what is left of an enclosure made of perishable mate-

rial (as also testified by the presence of cordage), whereas the high density of Fruits could be

interpreted as forage [66].

Space and activities during the Late Pastoral

Late Pastoral 1 (5900–4300 years cal BP). The Middle Pastoral experiences gradual cli-

matic deterioration, which intensifies significantly in the early stages of the Late Pastoral [100].

In Late Pastoral 1, we observe a substantial reorganization of the settlement system and a shift

towards an exclusive reliance on small livestock exploitation for subsistence, mainly goat. This

phase is predominantly found in the Main Sector, where plant remains are concentrated

towards the shelter’s innermost region, displaying a reasonably varied spatial distribution.

However, a common feature is the presence of isolated clusters and a preference for areas adja-

cent to the shelter’s wall (Fig 14). These isolated clusters correspond to a low correlation index,

except for Barks and Twigs (S3 Table).

Within this area, we identify numerous fixtures, including a formal stone structure, four

pits, a post hole, and five hearths. This delineates a more organized zone between the shelter

wall and the central space. Barks, Sticks, and Twigs tend to cluster more densely in the north-

eastern part of the Main Sector, gradually diminishing towards the south, LP1_CL1. These

materials are often associated with structural elements such as pits, hearths, and ash accumula-

tions, suggesting a multifunctional area encompassing activities like fire-related tasks, storage,

and processing.

A study on caprine dung [101] highlights a primary stabling area in the central portion of

the Main Sector, near the shelter wall where LP1_CL2 is located. This may imply the presence

of fencing structures, supported by evidence like the existence of a post hole on its outer

boundary and a wooden pole. Fruits are also prevalent near the shelter wall, likely resulting

from digested fodder, as they are found within dung levels.

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 28 / 40

https://doi.org/10.1371/journal.pone.0310739


Fig 12. MP1 Kernel Density Estimation. Kernel Density Estimation of selected ecofacts (barks, fruits, sticks and twigs) in MP1, with

plotting of relevant artifacts discussed in the text and indication of clusters specified in Table 7.

https://doi.org/10.1371/journal.pone.0310739.g012

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 29 / 40

https://doi.org/10.1371/journal.pone.0310739.g012
https://doi.org/10.1371/journal.pone.0310739


Fig 13. MP2 Kernel Density Estimation. Kernel Density Estimation of selected ecofacts (barks, fruits, sticks and twigs) in MP2, with

plotting of relevant artifacts discussed in the text and indication of clusters specified in Table 7.

https://doi.org/10.1371/journal.pone.0310739.g013

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 30 / 40

https://doi.org/10.1371/journal.pone.0310739.g013
https://doi.org/10.1371/journal.pone.0310739


Fig 14. LP1 Kernel Density Estimation. Kernel Density Estimation of selected ecofacts (barks, fruits, sticks and twigs) in LP1, with plotting of

relevant artifacts discussed in the text and indication of clusters specified in Table 7.

https://doi.org/10.1371/journal.pone.0310739.g014

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 31 / 40

https://doi.org/10.1371/journal.pone.0310739.g014
https://doi.org/10.1371/journal.pone.0310739


In the central-northern zone of the shelter, we observe weaker concentrations of Fruits and

Barks. This corresponds to the formal structure F7–21, interpreted as a small installation

linked to animal management activities [101].

General discussion: Plant and site use across the Holocene

The archaeobotanical record at Takarkori shows the biodiversity of wild plants that were dis-

tributed in the area in the early to late Holocene. Foragers and herders possessed a deep eco-

logical and environmental knowledge of the environment, as mirrored in the quantity and the

careful selection of the different available plant species. The use of plants for the preparation,

construction, and maintenance of the living floors and dwelling arrangements are mostly evi-

dent during the Late Acacus, and clearly indicates a semi-permanent type of occupation by for-

agers. This implies a significant investment in time and effort for the realization of facilities.

Spatial configurations reflect a commitment to long-term inhabitation. The intricate rela-

tionships and features, characterized by postholes, wooden poles and archaeobotanical

remains, along with its consistent use, signifies a substantial investment in establishing a per-

manent settlement, commencing with the initial habitation in the LA1 and further growing

along the LA2 and LA3 subphases. Investing in dwellings has consistently shown a strong link

to reduced residential mobility and a deliberate intention for longer stay [102–107]. Addition-

ally, the size and construction investment in huts are directly linked to the expectation of pro-

longed habitation [108]. The absence of macroremains from the immediate vicinity of

structure 131 (LA3) can moreover be interpreted as a result of a specific settlement strategy,

contrasting with the evidence of the hut structure 374 in the previous LA2 sub-phase. The lat-

ter was likely more temporary and allowed to be dismissed without further “caring”, whereas

for the larger, more durable, and stable structure 131, important and “structural” items like

poles and side mats might have been deliberately removed as part of an “anticipated mobility”

strategy following Kent’s framework [108].

At Takarkori, the site organization layout furthermore points towards a high investment

into resource management, as testified by the storage facilities made of basketry and of the dif-

ferential distribution of ecofacts. Along with the presence of an enclosure for corralling wild

animals, it clearly indicates a planned use of various resources. This programming is hence

also visible in the systematic and reiterated use of similar spatial layouts in the site, which finds

close correlates also in other sites of the region, like Ti-n-Torha East in the northern Tadrart

Acacus, where the presence of hut remains are recorded [109, 110], in the remains retrieved

from Uan Tabu [65], Uan Afuda and Fozzigiaren [111]. If hence the ways domestic architec-

ture is realized may reflect a “habitus” (sensu Bourdieu), the similarity across a sub-region may

be indicative of at least a someway unitarian cultural entity. Evidence of residential architec-

ture and increased sedentism was furthermore advanced for the circular outlined structures

and floor with leafy remains in the sheltered site of Ti-n-Hanakaten in the Tassili-n-Ajjer in

Algeria [64, 112, 113], which further sustain close relationships and cultural affinity between

the two areas of Central Sahara [114, 115]. The significant level of reiteration observed in the

spatial arrangement at Takarkori over time can be attributed to two possible factors, besides

the spectrum of plant species available in the region. Firstly, the presence of material remnants

may evoke reminiscences of past occupations, leading to the preservation of previous patterns

as part of an active social tradition [49, 53, 116–119]. Alternatively, these material remains

may provide favorable conditions that enhance the attractiveness of specific locations within

the site for certain activities [94, 102, 120]. Areas designated for waste disposal, for instance,

may continue to serve that purpose, while activity areas may be reused for similar or related

tasks, facilitating the reuse of debris as recycled materials [15, 56, 121].

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 32 / 40

https://doi.org/10.1371/journal.pone.0310739


Moreover, the spatial redundancy in the site’s layout may also be influenced by strictly

functional factors related to the topography and physical characteristics of the location. These

factors contribute to meeting the fundamental requirements of dwelling structures, aiming to

optimize cost and energy efficiency [117, 122], particularly effective in sheltered locations.

Functionality and maximization in the exploitation of space in terms of accessibility and

arrangements are factors that certainly also affected the later occupations of the Pastoral Neo-

lithic. The reconstruction of possible site layout in the Early Pastoral evidence how the con-

comitant use of the shelter as burial ground in some way influenced the organization of the

site. The plant remains associated with the burial H6 might be linked to the layout of the grave

installments, as liner or coating in association with the stone arrangements, a feature of this

period [26]. The transition to a productive subsistence economy must have led, at least at its

beginnings, to a degree of changes in mobility strategies and site organization, as reflected in

the decrease of clear and well-defined dwelling installments. The identification of some latent

structure as evidenced by plant remains, testifies on the other hand that the site had multiple

functions, and was certainly shared by both humans and their livestock.

The presence and the increase of socio-economic importance of animals might be con-

nected to the parallel increase in Balanites aegyptiaca, which as testified from different ethno-

graphic sources is a valuable and still used fodder, especially during resting times and when

animals are fenced for some period. This can be the case of the agglomerations of desert date

fruits in the Middle Pastoral periods, where several elements testify to the co-presence of ani-

mals and humans at the site during the time of occupation. These became shorter but repeated

and inserted into a scheduled occupation of mountain areas part of a systematic transhumant

system of seminomadic herders [100]. Nomadic and highly specialized goat herders character-

ize the demise of Takarkori’s occupation with short-term and episodic use of use-specific

installments, fodder remains and thick dung layers [55, 101].

The results here presented shed further insights into the complex and multifaceted relation-

ships between plant materials and the dwellers of the Takarkori rock shelter. The need to pres-

ent a diachronic picture of this interconnection hampered the investigations of multiple lines

of interpretation and nuances that are evident throughout the millennial occupation of the

site. The presence of multiple phases of occupation, which exceeds dozens of generations,

implies the equal multiplicity of behaviors that can be reflected within the material and spatial

vestiges recognized in the putative living space. Trends and patterns are in fact generalizations

that offer us a reconstructed image of reality that can, however, be approximated if interro-

gated and analyzed with the right degree of formalization and analyticity.

Final remarks

The archaeobotanical evidence from Takarkori indicates a significant biodiversity of wild

plants, reflecting the extensive ecological knowledge of foragers and herders in the early to late

Holocene. The complexity and diversity of plant remain assemblages reflect a long history of

occupation, spanning multiple generations, and highlight the diverse and evolving nature of

human-environment interactions at the site evidencing:

• The careful selection and use of plants for different purposes, such as construction materials

for living floors and dwellings, suggest a deep knowledge of local ecology and vegetation and

a significant investment in semi-permanent structures during the Late Acacus period.

• The consistent presence of certain plant species across different occupation phases reflects a

systematic approach to resource management, including the use of specific plants for food,

fodder, and other utilitarian purposes.

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 33 / 40

https://doi.org/10.1371/journal.pone.0310739


• The spatial distribution of plant remains, including their association with specific structures

and activity areas, indicates deliberate planning in the use of plant resources, such as the use

of Balanites aegyptiaca for animal fodder during the Middle Pastoral period.

• The evidence of plant materials in burial contexts and the presence of leafy remains in shel-

tered areas highlight the diverse uses of plants beyond subsistence, including ritual and cul-

tural practices.

Supporting information

S1 Fig. Monte Carlo test of spatial randomness for the point pattern of selected ecofacts.

Monte Carlo test of spatial randomness for the point pattern (Ripley’s K function) of selected

ecofact categories, according to sub-phases (from the oldest LA1, top left, to the youngest LP1,

bottom right). Observed K-function is on y axes (black line) compared to the expected ran-

domly distributed points (red dotted line) in the expected range of 95% confidence envelopes

(gray area) for the hypothesis of complete spatial randomness, obtained from 100 independent

randomizations.

(TIF)

S1 Table. Percentage of selected plant remains. The percentage of selected plant remains, by

ACs (in brackets the absolute frequency) and chronological sub-phases.

(DOCX)

S2 Table. QCT results for the archaeobotanical selected samples by sub-phase.

(DOCX)

S3 Table. Pearson’s correlation coefficient r between selected ecofacts by chronological

sub-phase.

(DOCX)

S1 File.

(XLSX)

Acknowledgments

This research is part of the activities of the Archaeological Mission in the Sahara, Sapienza

University of Rome, directed by S.d.L. We thank the Department of Antiquities in Tripoli and

Ghat.

Author Contributions

Conceptualization: Savino di Lernia, Anna Maria Mercuri, Rocco Rotunno.

Data curation: Fabrizio Buldrini, Varinia Nardi, Rocco Rotunno.

Formal analysis: Fabrizio Buldrini, Varinia Nardi, Rocco Rotunno.

Funding acquisition: Savino di Lernia.

Investigation: Savino di Lernia, Fabrizio Buldrini, Assunta Florenzano, Anna Maria Mercuri,

Varinia Nardi, Rocco Rotunno.

Methodology: Savino di Lernia, Fabrizio Buldrini, Anna Maria Mercuri, Rocco Rotunno.

Project administration: Savino di Lernia, Anna Maria Mercuri.

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 34 / 40

http://www.plosone.org/article/fetchSingleRepresentation.action?uri=info:doi/10.1371/journal.pone.0310739.s001
http://www.plosone.org/article/fetchSingleRepresentation.action?uri=info:doi/10.1371/journal.pone.0310739.s002
http://www.plosone.org/article/fetchSingleRepresentation.action?uri=info:doi/10.1371/journal.pone.0310739.s003
http://www.plosone.org/article/fetchSingleRepresentation.action?uri=info:doi/10.1371/journal.pone.0310739.s004
http://www.plosone.org/article/fetchSingleRepresentation.action?uri=info:doi/10.1371/journal.pone.0310739.s005
https://doi.org/10.1371/journal.pone.0310739


Resources: Savino di Lernia, Assunta Florenzano, Anna Maria Mercuri.

Supervision: Savino di Lernia, Anna Maria Mercuri.

Visualization: Savino di Lernia, Assunta Florenzano, Varinia Nardi, Rocco Rotunno.

Writing – original draft: Savino di Lernia, Fabrizio Buldrini, Assunta Florenzano, Varinia

Nardi, Rocco Rotunno.

Writing – review & editing: Savino di Lernia, Assunta Florenzano, Anna Maria Mercuri,

Rocco Rotunno.

References
1. Cappers RT, Bekker RM, Jans JE. Digital seed atlas of the Netherlands. Barkhuis; 2006.

2. Miller NF. Archaeobotany: The use of plant remains in archaeology. Routeledge; 2021.

3. Mercuri AM, D’Andrea AC, Fornaciari R, Höhn A. Plants and people in the African past: Progress in

African archaeobotany. Springer; 2018.

4. Mercuri AM, Fornaciari R, Gallinaro M, Vanin S, di Lernia S. Plant behaviour from human imprints and

the cultivation of wild cereals in Holocene Sahara. Nature Plants. 2018; 4: 71–81. https://doi.org/10.

1038/s41477-017-0098-1

5. Dunne J, Mercuri AM, Evershed RP, Bruni S, di Lernia S. Earliest direct evidence of plant processing

in prehistoric Saharan pottery. Nature Plants. 2016; 3: 16194. https://doi.org/10.1038/nplants.2016.

194

6. Dunne J, Höhn A, Neumann K, Franke G, Breunig P, Champion L, et al. Making the invisible visible:

tracing the origins of plants in West African cuisine through archaeobotanical and organic residue anal-

ysis. Archaeol Anthropol Sci. 2022; 14: 30. https://doi.org/10.1007/s12520-021-01476-0

7. Lucarini G, Radini A. First direct evidence of wild plant grinding process from the Holocene Sahara:

Use-wear and plant micro-residue analysis on ground stone tools from the Farafra Oasis, Egypt. Qua-

ternary International. 2020; 555: 66–84. https://doi.org/10.1016/j.quaint.2019.07.028

8. Gillings M, Hacıgüzeller P, Lock G. Archaeological Spatial Analysis: A Methodological Guide. Rout-

ledge; 2020.

9. Mercuri AM. Plant Use. Oxford Research Encyclopedia of Anthropology. 2022. https://doi.org/10.

1093/acrefore/9780190854584.013.50

10. Pearsall DM. Paleoethnobotany: A Handbook of Procedures. 3rd ed. New York: Routledge; 2016.

https://doi.org/10.4324/9781315423098

11. Jolly D, Prentice IC, Bonnefille R, Ballouche A, Bengo M, Brenac P, et al. Biome reconstruction from

pollen and plant macrofossil data for Africa and the Arabian peninsula at 0 and 6000 years. Journal of

Biogeography. 1998; 25: 1007–1027.

12. Mercuri AM. Genesis and evolution of the cultural landscape in central Mediterranean: the ‘where,

when and how’ through the palynological approach. Landscape Ecol. 2014; 29: 1799–1810. https://

doi.org/10.1007/s10980-014-0093-0

13. Miller NF, Johnson M. Plant remains and ancient landscapes: A multiscale perspective. Annual

Review of Anthropology. 2023; 52: 271–289.

14. Conolly J, Lake M. Geographical Information Systems in Archaeology. Cambridge: Cambridge Uni-

versity Press; 2006. https://doi.org/10.1017/CBO9780511807459

15. Achino KF, Barceló JA. Spatial Prediction: Reconstructing the “Spatiality” of Social Activities at the

Intra-Site Scale. J Archaeol Method Theory. 2019; 26: 112–134. https://doi.org/10.1007/s10816-018-

9367-1

16. Carrer F. Interpreting Intra-site Spatial Patterns in Seasonal Contexts: an Ethnoarchaeological Case

Study from the Western Alps. J Archaeol Method Theory. 2017; 24: 303–327. https://doi.org/10.1007/

s10816-015-9268-5 PMID: 29266121

17. Verhagen P. Spatial Analysis in Archaeology: Moving into New Territories. In: Siart C, Forbriger M,

Bubenzer O, editors. Digital Geoarchaeology: New Techniques for Interdisciplinary Human-Environ-

mental Research. Cham: Springer International Publishing; 2018. pp. 11–25. https://doi.org/10.1007/

978-3-319-25316-9_2

18. Champion L, Fuller DQ, Ozainne S, Huysecom É, Mayor A. Agricultural diversification in West Africa:

an archaeobotanical study of the site of Sadia (Dogon Country, Mali). Archaeol Anthropol Sci. 2021;

13: 60. https://doi.org/10.1007/s12520-021-01293-5 PMID: 33758626

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 35 / 40

https://doi.org/10.1038/s41477-017-0098-1
https://doi.org/10.1038/s41477-017-0098-1
https://doi.org/10.1038/nplants.2016.194
https://doi.org/10.1038/nplants.2016.194
https://doi.org/10.1007/s12520-021-01476-0
https://doi.org/10.1016/j.quaint.2019.07.028
https://doi.org/10.1093/acrefore/9780190854584.013.50
https://doi.org/10.1093/acrefore/9780190854584.013.50
https://doi.org/10.4324/9781315423098
https://doi.org/10.1007/s10980-014-0093-0
https://doi.org/10.1007/s10980-014-0093-0
https://doi.org/10.1017/CBO9780511807459
https://doi.org/10.1007/s10816-018-9367-1
https://doi.org/10.1007/s10816-018-9367-1
https://doi.org/10.1007/s10816-015-9268-5
https://doi.org/10.1007/s10816-015-9268-5
http://www.ncbi.nlm.nih.gov/pubmed/29266121
https://doi.org/10.1007/978-3-319-25316-9%5F2
https://doi.org/10.1007/978-3-319-25316-9%5F2
https://doi.org/10.1007/s12520-021-01293-5
http://www.ncbi.nlm.nih.gov/pubmed/33758626
https://doi.org/10.1371/journal.pone.0310739


19. Van der Veen M. The Exploitation of Plant Resources in Ancient Africa. New York: Kluwer Academic/

Plenum Publishers; 1999.

20. Clark AE, Gingerich JAM, editors. Intrasite Spatial Analysis of Mobile and Semisedentary Peoples:

Analytical Approaches to Reconstructing Occupation History. Salt Lake City: University of Utah

Press; 2022.

21. Biagetti S, di Lernia S. Holocene deposits of Saharan rock shelters: The case of Takarkori and other

sites from the Tadrart Acacus Mountains (Southwest Libya). African Archaeological Review. 2013; 30:

305–338. https://doi.org/10.1007/s10437-013-9138-z

22. Mercuri AM. Human influence, plant landscape evolution and climate inferences from the archaeobo-

tanical records of the Wadi Teshuinat area (Libyan Sahara). Journal of Arid Environments. 2008; 72:

1950–1967.

23. Zerboni A, Perego A, Cremaschi M. Geomorphological Map of the Tadrart Acacus Massif and the Erg

Uan Kasa (Libyan Central Sahara). Journal of Maps. 2015; 11: 772–787. https://doi.org/10.1080/

17445647.2014.955891

24. Cremaschi M. Late Quaternary geological evidence for environmental changes in south-western Fez-

zan (Libyan Sahara). In: Cremaschi M, di Lernia S, editors. Wadi Teshuinat—Palaeoenvironment and

Prehistory in South-Western Fezzan (Libyan Sahara). CNR-All’Insegna del Giglio; 1998. pp. 13–48.

25. Cremaschi M, Zerboni A, Mercuri AM, Olmi L, Biagetti S, di Lernia S. Takarkori rock shelter (SW

Libya): an archive of Holocene climate and environmental changes in the central Sahara. Quaternary

Science Reviews. 2014; 101: 36–60. https://doi.org/10.1016/j.quascirev.2014.07.004

26. di Lernia S, Tafuri MA. Persistent deathplaces and mobile landmarks: The Holocene mortuary and iso-

topic record from Wadi Takarkori (SW Libya). Journal of Anthropological Archaeology. 2013; 32: 1–15.

https://doi.org/10.1016/j.jaa.2012.07.002

27. di Lernia S, Gallinaro M. Libya Before and After the Conflict: What Future for Its Cultural Heritage? In:

Castillo A, editor. Archaeological Dimension of World Heritage: From Prevention to Social Implica-

tions. New York, NY: Springer; 2014. pp. 73–87. https://doi.org/10.1007/978-1-4939-0283-5_6

28. Cherkinsky A, di Lernia S. Bayesian Approach to 14 C Dates for Estimation of Long-Term Archaeolog-

ical Sequences in Arid Environments: The Holocene Site of Takarkori Rockshelter, Southwest Libya.

Radiocarbon. 2013; 55: 771–782.

29. Lernia S di. Saharan Hunter-Gatherers: Specialization and Diversification in Holocene Southwestern

Libya. Taylor & Francis; 2022.

30. Li H, Renssen H, Roche DM, Miller PA. Modelling the vegetation response to the 8.2 ka bp cooling

event in Europe and Northern Africa. Journal of Quaternary Science. 2019; 34: 650–661. https://doi.

org/10.1002/jqs.3157

31. Bronk Ramsey C. Bayesian analysis of radiocarbon dates. Radiocarbon. 2009; 51: 337–360.

32. di Lernia S, N’siala IM, Mercuri AM. Saharan prehistoric basketry. Archaeological and archaeobotani-

cal analysis of the early-middle Holocene assemblage from Takarkori (Acacus Mts., SW Libya). Jour-

nal of Archaeological Science. 2012; 39: 1837–1853.

33. Cappers RT, Neef R, Bekker RM. Digital atlas of economic plants. Barkhuis; 2009.

34. Fahn, Werker, Baas. Wood anatomy and identification of trees and shrubs from Israel and adjacent

regions. Wood anatomy and identification of trees and shrubs from Israel and adjacent regions. Brill;

1986. Available: https://brill.com/display/title/2821

35. Neumann K, Schoch W, D&#233, Tienne P, Schweingruber FH. Woods of the Sahara and the Sahel.

Woods of the Sahara and the Sahel. 2001 [cited 22 Sep 2023]. Available: https://www.cabdirect.org/

cabdirect/abstract/20026791720

36. Ozenda P. Flore et végétation du Sahara septentrional et central. 3rd ed. Paris: Editions du CNRS;

2004. Available: https://www.soumbala.com/pays/pays-d-afrique-de-l-ouest-1/mali/flore-du-sahara-

septentrional-et-central-3e-edition.html

37. Wasylikowa K. Holocene flora of the Tadrart Acacus area, SW Libya, based on plants macrofossils

from Uan Muhuggiag and Ti-n-Torha Two Caves archaeological sites. Origini. 1992; 16: 125–159.

38. Wasylikowa K. Flora of the 8000 years old archaeological site E-75-6 at Nabta Playa, Western Desert,

Southern Egypt. Acta Palaeobotanica. 1997; 37: 99–205.

39. Van Der Veen M. Consumption, Trade and Innovation:: exploring the botanical remains from the

Roman and Islamic ports at Quseir al-Qadim, Egypt. Africa Magna Verlag; 2011.

40. Zohary D, Hopf M, Weiss E. Domestication of plants in the Old World, Fourth Edition. Oxford: Oxford

University Press; 2012.

41. Hietala H, Larson PA. Intrasite Spatial Analysis in Archaeology. CUP Archive; 1984.

PLOS ONE Spatial analyses of archaeobotanical record for assessing site uses and activities at Takarkori (Libya)

PLOS ONE | https://doi.org/10.1371/journal.pone.0310739 October 23, 2024 36 / 40

https://doi.org/10.1007/s10437-013-9138-z
https://doi.org/10.1080/17445647.2014.955891
https://doi.org/10.1080/17445647.2014.955891
https://doi.org/10.1016/j.quascirev.2014.07.004
https://doi.org/10.1016/j.jaa.2012.07.002
https://doi.org/10.1007/978-1-4939-0283-5%5F6
https://doi.org/10.1002/jqs.3157
https://doi.org/10.1002/jqs.3157
https://brill.com/display/title/2821
https://www.cabdirect.org/cabdirect/abstract/20026791720
https://www.cabdirect.org/cabdirect/abstract/20026791720
https://www.soumbala.com/pays/pays-d-afrique-de-l-ouest-1/mali/flore-du-sahara-septentrional-et-central-3e-edition.html
https://www.soumbala.com/pays/pays-d-afrique-de-l-ouest-1/mali/flore-du-sahara-septentrional-et-central-3e-edition.html
https://doi.org/10.1371/journal.pone.0310739


42. Carr C. The Nature of Organization of Intrasite Archaeological Records and Spatial Analytic

Approaches to Their Investigation. Advances in Archaeological Method and Theory. 1984; 7: 103–

222.

43. Orton C. Point pattern analysis revisited. Archeologia e Calcolatori. 2004; 15: 299–315.

44. Bivand RS, Pebesma E, Gómez-Rubio V. Applied Spatial Data Analysis with R. New York, NY:

Springer; 2013. https://doi.org/10.1007/978-1-4614-7618-4

45. Baddeley A, Turner R. Spatstat: an R package for analyzing spatial point patterns. Journal of statistical

software. 2005; 12: 1–42.

46. Baddeley A, Rubak E, Turner R. Spatial point patterns: methodology and applications with R. CRC

press; 2015.

47. Baxter MJ. Spatial k-means clustering in archaeology–variations on a theme. Academia (Accessed

February 17, 2017) https://www academia edu/18142974/Spatial_k-means_clustering_in_archaeol-

ogy_–_variations_on_a_theme. 2015.

48. Baxter MJ, Beardah CC, Wright RV. Some archaeological applications of kernel density estimates.

Journal of Archaeological Science. 1997; 24: 347–354.

49. Bailey G. Time perspectives, palimpsests and the archaeology of time. Journal of Anthropological

Archaeology. 2007; 26: 198–223. https://doi.org/10.1016/j.jaa.2006.08.002

50. Bargalló A, Gabucio MJ, Rivals F. Puzzling out a palimpsest: testing an interdisciplinary study in level

O of Abric Romanı́. Quaternary International. 2016; 417: 51–65.

51. Moclán A, Huguet R, Márquez B, Álvarez-Fernández A, Laplana C, Arsuaga JL, et al. Identifying activ-

ity areas in a neanderthal hunting camp (the Navalmaı́llo Rock Shelter, Spain) via spatial analysis.

Archaeol Anthropol Sci. 2023; 15: 44. https://doi.org/10.1007/s12520-023-01746-z

52. Malinsky-Buller A, Hovers E,