Occasional Papers in Archaeology 60

Occasional Papers in Archaeology 60
Occasional Papers in Archaeology 60
Rita Peyroteo Stjerna
on
Death in the Mesolithic
or the
Mortuary Practices of the Last Hunter-Gatherers of the
South-Western Iberian Peninsula, 7th–6th Millennium BCE
Dissertation presented at Uppsala University to be publicly examined in Geijersalen, Centre
for the Humanities, English Park Campus, Uppsala, Friday, 26 February 2016 at 13:00 for
the degree of Doctor of Philosophy. The examination will be conducted in English. Faculty
examiner: Clive Bonsall (University of Edinburgh).
Abstract
Peyroteo Stjerna, R. 2016. On Death in the Mesolithic. Or the Mortuary Practices of the
Last Hunter-Gatherers of the South-Western Iberian Peninsula, 7th–6th Millennium BCE.
Occasional papers in archaeology 60. 511 pp. Uppsala: Department of Archaeology and
Ancient History, Uppsala University. ISBN 978-91-506-2525-7.
The history of death is entangled with the history of changing social values, meaning that a shift
in attitudes to death will be consistent with changes in a society’s world view.
Late Mesolithic shell middens in the Tagus and Sado valleys, Portugal, constitute some of the
largest and earliest burial grounds known, arranged and maintained by people with a hunting,
fishing, and foraging lifestyle, c 6000–5000 cal BCE. These sites have been interpreted in the
light of economic and environmental processes as territorial claims to establish control over
limited resources. This approach does not explain the significance of the frequent disposal of
the dead in neighbouring burial grounds, and how these places were meaningful and socially
recognized. The aim of this dissertation is to answer these questions through the detailed analysis
of museum collections of human burials from these sites, excavated between the late nineteenth
century and the 1960s.
I examine the burial activity of the last hunter-gatherers of the south-western Iberian Peninsula
from an archaeological perspective, and explain the burial phenomenon through the lens of
historical and humanist approaches to death and hunter-gatherers, on the basis of theoretical
concepts of social memory, place, mortuary ritual practice, and historical processes. Human
burials are investigated in terms of time and practice based on the application of three methods:
radiocarbon dating and Bayesian analysis to define the chronological framework of the burial
activity at each site and valley; stable isotope analysis of carbon and nitrogen aimed at defining
the burial populations by the identification of dietary choices; and archaeothanatology to
reconstruct and define central practices in the treatment of the dead.
This dissertation provides new perspectives on the role and relevance of the shell middens
in the Tagus and Sado valleys. Hunter-gatherers frequenting these sites were bound by shared
social practices, which included the formation and maintenance of burial grounds, as a primary
means of history making. Death rituals played a central role in the life of these hunter-gatherers
in developing a sense of community, as well as maintaining social ties in both life and death.
Keywords: death, Late Mesolithic, hunter-gatherers, social memory, place, burial practices,
mortuary ritual, historical process, south-western Iberian Peninsula, archaeothanatology,
radiocarbon dating and Bayesian analysis, stable isotopes (carbon and nitrogen), shell
middens, museum collections
Rita Peyroteo Stjerna, Department of Archaeology and Ancient History, Box 626, Uppsala
University, SE-75126 Uppsala, Sweden.
© Rita Peyroteo Stjerna 2016
ISSN 1100-6358
ISBN 978-91-506-2525-7
urn:nbn:se:uu:diva-271551 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-271551)
Appendix (online-only material): http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-271551
Proofreading by Jill Kinahan
Cover design by author and Tore Stjerna. Cover drawing (Skeleton 11, Vale de Romeiras
1959, Sado valley) by Dario de Sousa courtesy of the National Museum of Archaeology,
Lisbon, Portugal
Printed in Sweden by Kph Trycksaksbolaget, Uppsala 2016
To my mom,
who worked hard to raise me, and whose
love for History has always been an
inspiration.
Contents
Acknowledgements ......................................................................................................... xvii
Abbreviations .................................................................................................................. xxiii
List of Figures .................................................................................................................. xxv
List of Tables .................................................................................................................. xxxi
Introduction to the dissertation ...................................................................................... 35
Death in the Mesolithic sites of the Tagus and Sado valleys, Portugal ............................. 35
Aims of the dissertation ........................................................................................................... 36
Research questions and theoretical framework ............................................................... 36
Methodological choices ...................................................................................................... 36
The outcome ........................................................................................................................ 37
Outline of the dissertation....................................................................................................... 38
PART ONE: DEATH ........................................................................................................... 41
Introduction to part one .......................................................................................................... 43
Chapter 1 Death, Hunter-Gatherers, and the Historical Process................................ 44
The human engagement with death ....................................................................................... 44
Origins of human burial ..................................................................................................... 45
Material nature of death ..................................................................................................... 47
Archaeological thought and death.......................................................................................... 48
Chronology and typology ................................................................................................... 49
The cultural paradigm, normative and diffusionist......................................................... 49
The “New Archaeology” and the social approach to death .......................................... 50
Postprocessual reaction: rituals in action ......................................................................... 52
Historical and humanist approaches to death and hunter-gatherers ............................ 53
Deathways of hunter-gatherers and historical processes.......................................... 55
Memory in the landscape: rituals in practice .............................................................. 55
Death in deep time: why does it matter? ............................................................................... 56
Death paradigms ................................................................................................................. 56
Death studies and archaeology: an archaeological manifesto........................................ 57
PART TWO: BODIES OF KNOWLEDGE .......................................................................... 59
Introduction to part two .......................................................................................................... 61
Chapter 2 Archaeological Material: Human Remains in the Shell Middens of the
Tagus and Sado valleys, Portugal .................................................................................... 62
General description of the shell midden sites in the Tagus and Sado valleys, Portugal.. 62
Material Selection...................................................................................................................... 64
Sites selected ........................................................................................................................ 67
A note on the museum collections .............................................................................. 67
Notation of the human remains ............................................................................. 68
A note on the historical documentation ..................................................................... 69
Tagus valley ............................................................................................................... 69
Sado valley ................................................................................................................. 69
Sites excluded....................................................................................................................... 71
Sites with human remains ............................................................................................. 71
Tagus valley, Fonte da Moça: Fonte da Moça 1 and 2 ........................................ 72
Tagus valley, Muge: Fonte do Padre Pedro .......................................................... 72
Tagus valley, Muge: Flor da Beira .......................................................................... 72
Tagus valley, Muge: Cabeço da Amoreira ............................................................. 73
Tagus valley, Magos: Cova da Onça ...................................................................... 73
Tagus valley, Magos: Cabeço dos Morros ............................................................. 74
Sado valley: Poças de S. Bento ............................................................................... 74
Sites without known human remains .......................................................................... 76
Total funeral .............................................................................................................................. 77
Preservation and research bias .......................................................................................... 77
Organic preservation ..................................................................................................... 77
Settlement network bias ................................................................................................ 78
Visibility of the archaeological record ......................................................................... 78
Representativeness of the sample under study................................................................ 78
Hunter-gatherer death rate vs burial population............................................................. 79
A note about longevity among hunter-gatherers ....................................................... 81
Understanding the context of data production: brief research history ............................. 81
Portuguese shell middens and the birth of prehistoric archaeology in Portugal ........ 82
The antiquity of mankind ............................................................................................. 82
From the antiquity of mankind to the antiquity of nations: human remains and
discourses of identity .......................................................................................................... 83
Historical background ................................................................................................... 83
Prehistoric remains ........................................................................................................ 84
The 1950s–60s: towards a systematic archaeology .................................................... 86
The 1980s–90s: old collections, physical anthropology, and selected fieldwork... 88
Maturity in the twenty-first century: Sado valley out of the archives ..................... 89
Chapter 3 Methods ............................................................................................................. 91
Methodological choices in this study ..................................................................................... 91
Radiocarbon dating method and Bayesian analysis.............................................................. 93
Introduction ......................................................................................................................... 93
Chronological resolution in prehistory ....................................................................... 93
Enquiry-dependent chronologies: the radiocarbon dating programme in this
study ................................................................................................................................ 93
This section..................................................................................................................... 95
Structure .................................................................................................................... 95
Terminology and conventions ................................................................................ 95
Radiocarbon dating method .............................................................................................. 97
Basic principles............................................................................................................... 97
Interpretation of 14C measurements ................................................................................. 99
Carbon fluctuation through time: calibration curves .............................................. 100
The place of the sampled organism in the environment ........................................ 100
Reservoir offsets (ΔR) ........................................................................................... 100
Reservoir offsets in this study ......................................................................... 101
Dietary sources and the calibration of samples of human bone ...................... 102
Sources of protein component in human bone collagen: calculation
estimate............................................................................................................... 103
Method marine 1 ......................................................................................... 103
Method marine 2 ......................................................................................... 104
Statistical treatment and modelling............................................................................ 106
Why modelling? ...................................................................................................... 106
The Bayesian approach .......................................................................................... 108
OxCal ............................................................................................................................ 110
Reliability and stability of the models ....................................................................... 112
Statistical criteria ..................................................................................................... 113
Sensitivity analysis: prior beliefs and standardised likelihoods ......................... 113
Calibration and modelling in this study .................................................................... 114
Limits of current standardised likelihoods .......................................................... 115
Estimated aquatic protein intake .................................................................... 115
Endpoint values ................................................................................................ 115
ΔR ....................................................................................................................... 116
Sensitivity analysis of regional reservoir offsets ...................................... 116
Selection and screening of samples in this study .......................................................... 118
Nature and context ...................................................................................................... 118
Contamination.............................................................................................................. 118
In the laboratory: pre-treatment and measurement of samples in this study ...... 118
Reliability of the isotopic measurements .................................................................. 119
Quality parametres ................................................................................................. 119
δ13C values: AMS or IRSM ................................................................................... 120
Screening of samples in this study .................................................................. 121
Stable isotope analysis of carbon and nitrogen .................................................................. 124
Introduction ....................................................................................................................... 124
Burial analysis and past human diets ......................................................................... 124
This section................................................................................................................... 126
Terminology and conventions .............................................................................. 126
Principles of dietary reconstruction based on the isotopes of carbon and
nitrogen............................................................................................................................... 127
Knowledge of the past environment......................................................................... 127
Variation of δ13C .................................................................................................... 127
Variation of δ15N .................................................................................................... 128
Freshwater reservoirs ............................................................................................. 128
Assumptions in this study........................................................................................... 129
Archaeothanatology: chaîne opératoire of mortuary practices .............................................. 130
Introduction ....................................................................................................................... 130
Enquiry-dependent taphonomies: the archaeothanatological approach .............. 130
This section................................................................................................................... 131
Structure .................................................................................................................. 131
Terminology and conventions .............................................................................. 132
Anatomical terminology ................................................................................... 132
Directional terms ......................................................................................... 132
Motions of the body.................................................................................... 133
Other terms .................................................................................................. 133
Mortuary terminology....................................................................................... 133
Archaeothanatology and archaeology............................................................................. 134
Brief history of the method and impact in archaeology ......................................... 134
Archaeothanatology: main principles ........................................................................ 136
Aims of the method ............................................................................................... 136
Taphonomy of mortuary practices....................................................................... 137
Basic principles ....................................................................................................... 137
First basic principle: observation of the joints .............................................. 138
Second basic principle: observation of the internal and external
environments of the cadaver ........................................................................... 140
The nature of the deposits .................................................................................... 142
Primary deposit ................................................................................................. 142
Secondary deposit ............................................................................................. 143
The space of decomposition of the cadaver ....................................................... 144
Empty and filled spaces of decomposition ................................................... 144
Wall-effect (effet de parois) ............................................................................. 145
The initial position of the cadaver in the feature ............................................... 145
Cranium and mandible ..................................................................................... 146
Observation of the mandible ..................................................................... 146
Hyoid .................................................................................................................. 146
Vertebral column .............................................................................................. 147
Observation of the cervical vertebrae (C1–C7) ....................................... 147
Observation of the lumbar vertebrae (LV) .............................................. 147
Thoracic cage: sternum and ribs ..................................................................... 148
Bones of the shoulder girdle and upper limbs .............................................. 149
Bones of the shoulder girdle ...................................................................... 149
Bones of the hands ...................................................................................... 149
Bones of the pelvic girdle and lower limbs ................................................... 150
Bones of the pelvic girdle ........................................................................... 150
Bones of the lower limbs ............................................................................ 150
Patellae ..................................................................................................... 150
Bones of the feet .................................................................................... 151
Initial position of the cadaver: additional observations about constraint
effects ................................................................................................................. 151
Sediment and hypercontracted skeletons ................................................. 151
Wrappings..................................................................................................... 151
The grave features .................................................................................................. 151
Deposits containing more than one individual .................................................. 152
Funerary contexts: one single structure ......................................................... 153
Funerary contexts: more than one structure ................................................. 154
Post-depositional manipulations of the cadavers ............................................... 155
Archaeothanatology in this study .................................................................................... 155
Source material: the challenges .................................................................................. 155
Graphic documentation: drawings ....................................................................... 156
Graphic documentation: field photographs........................................................ 157
Human remains ...................................................................................................... 157
Archaeothanatology in the lab: the strategies........................................................... 158
Chapter 4 Results .............................................................................................................. 161
Introduction ............................................................................................................................ 161
Tagus valley: Moita do Sebastião .......................................................................................... 163
Site background: a short history of research ................................................................. 163
Chronology of the burial activity .................................................................................... 165
Previous 14C measurements ........................................................................................ 165
Objectives of the dating programme ........................................................................ 166
Sampling strategy ......................................................................................................... 166
Results ........................................................................................................................... 168
Calibration and chronological models ...................................................................... 171
Internal consistency and reliability of the model ............................................... 171
Stable isotopes: carbon and nitrogen.............................................................................. 175
Previous measurements: δ13C and δ15N ................................................................... 175
Sampling strategy ......................................................................................................... 176
Results ........................................................................................................................... 176
Archaeothanatology .......................................................................................................... 179
Source material and analysis strategy......................................................................... 179
Results of the archaeothanatological analysis .......................................................... 179
The nature of the deposits .................................................................................... 179
The space of decomposition of the cadaver ....................................................... 183
The initial position of the cadaver in the feature ............................................... 183
The grave features .................................................................................................. 189
Deposits containing more than one individual .................................................. 196
Post-depositional manipulations of the cadavers ............................................... 196
Other observations................................................................................................. 197
The mortuary programme at Moita do Sebastião .............................................. 197
Tagus valley: Cabeço da Arruda ........................................................................................... 199
Site background: a short history of research ................................................................. 199
Chronology of the burial activity .................................................................................... 201
Previous 14C measurements ........................................................................................ 201
Objectives of the dating programme ........................................................................ 202
Sampling strategy ......................................................................................................... 202
Results ........................................................................................................................... 202
Calibration and chronological models ...................................................................... 207
Internal consistency and reliability of the model ............................................... 207
Stable isotopes: carbon and nitrogen.............................................................................. 211
Previous measurements: δ13C and δ15N ................................................................... 211
Sampling strategy ......................................................................................................... 212
Results ........................................................................................................................... 212
Archaeothanatology .......................................................................................................... 215
Source material and analysis strategy......................................................................... 215
Results of the archaeothanatological analysis .......................................................... 219
The nature of the deposits .................................................................................... 219
The space of decomposition of the cadaver ....................................................... 219
The initial position of the cadaver in the feature ............................................... 222
The grave features .................................................................................................. 222
Deposits containing more than one individual .................................................. 223
Post-depositional manipulations of the cadavers ............................................... 223
The mortuary programme at Cabeço da Arruda ................................................ 223
Sado valley: Arapouco............................................................................................................ 227
Site background: a short history of research ................................................................. 227
Observations about one mislabelled individual ....................................................... 230
Chronology of the burial activity .................................................................................... 232
Previous 14C measurements ........................................................................................ 232
Objectives of the dating programme ........................................................................ 233
Sampling strategy ......................................................................................................... 233
Results ........................................................................................................................... 233
Calibration and chronological models ...................................................................... 233
Stable isotopes: carbon and nitrogen.............................................................................. 235
Previous measurements: δ13C and δ15N ................................................................... 235
Sampling strategy ......................................................................................................... 236
Results ........................................................................................................................... 236
Archaeothanatology .......................................................................................................... 238
Source material and analysis strategy......................................................................... 238
Results of the archaeothanatological analysis .......................................................... 238
The nature of the deposits .................................................................................... 238
The space of decomposition of the cadaver ....................................................... 241
The initial position of the cadaver in the feature ............................................... 244
The grave features .................................................................................................. 247
Deposits containing more than one individual .................................................. 248
Post-depositional manipulations of the cadavers ............................................... 251
The mortuary programme at Arapouco .............................................................. 251
ARA1961, Sk5 ................................................................................................... 251
ARA1961, Sk6 ................................................................................................... 256
ARA1961, Sks 11 and 12 ................................................................................. 259
ARA1961, Sk13 ................................................................................................. 263
Sado valley: Cabeço das Amoreiras ...................................................................................... 267
Site background: a short history of research ................................................................. 267
Chronology of the burial activity .................................................................................... 271
Previous 14C measurements ........................................................................................ 271
Objectives of the dating programme ........................................................................ 271
Sampling strategy ......................................................................................................... 271
Results ........................................................................................................................... 272
Calibration and chronological models ...................................................................... 272
Internal consistency and reliability of the model ............................................... 273
Stable isotopes: carbon and nitrogen.............................................................................. 276
Previous measurements: δ13C and δ15N ................................................................... 276
Sampling strategy ......................................................................................................... 276
Results ........................................................................................................................... 276
Archaeothanatology .......................................................................................................... 279
Source material and analysis strategy......................................................................... 279
Results of the archaeothanatological analysis .......................................................... 280
The nature of the deposits .................................................................................... 280
The space of decomposition of the cadaver ....................................................... 280
The initial position of the cadaver in the feature ............................................... 280
The grave features .................................................................................................. 285
Deposits containing more than one individual .................................................. 287
Post-depositional manipulations of the cadavers ............................................... 287
The mortuary programme at Cabeço das Amoreiras ........................................ 290
Sado valley: Vale de Romeiras .............................................................................................. 292
Site background: a short history of research ................................................................. 292
Chronology of the burial practice ................................................................................... 296
Previous 14C measurements ........................................................................................ 296
Objectives of the dating programme ........................................................................ 296
Sampling strategy ......................................................................................................... 298
Results ........................................................................................................................... 298
Calibration and chronological models ...................................................................... 298
Internal consistency and reliability of the model ............................................... 299
Stable isotopes: carbon and nitrogen.............................................................................. 302
Previous measurements: δ13C and δ15N ................................................................... 302
Sampling strategy ......................................................................................................... 302
Results ........................................................................................................................... 302
Archaeothanatology .......................................................................................................... 305
Source material and analysis strategy......................................................................... 305
Results of the archaeothanatological analysis .......................................................... 305
The nature of the deposits .................................................................................... 305
The space of decomposition of the cadaver ....................................................... 309
The initial position of the cadaver in the feature ............................................... 314
The grave features .................................................................................................. 317
Deposits containing more than one individual .................................................. 320
Post-depositional manipulations of the cadavers ............................................... 322
Other observations................................................................................................. 327
Animal bones ..................................................................................................... 327
Other burials ...................................................................................................... 327
The mortuary programme at Vale de Romeiras ................................................. 329
VR1959, Sk11 .................................................................................................... 329
VR1959, Sk7 ...................................................................................................... 331
VR1959, Sk23 .................................................................................................... 332
VR1959, Sk19 .................................................................................................... 334
Sado valley: Cabeço do Pez ................................................................................................... 338
Site background: a short history of research ................................................................. 338
Chronology of the burial activity .................................................................................... 342
Previous 14C measurements ........................................................................................ 342
Objectives of the dating programme ........................................................................ 342
Sampling strategy ......................................................................................................... 343
Results ........................................................................................................................... 343
Calibration and chronological models ...................................................................... 344
Internal consistency and reliability of the model ............................................... 344
Stable isotopes: carbon and nitrogen.............................................................................. 348
Previous measurements: δ13C and δ15N ................................................................... 348
Sampling strategy ......................................................................................................... 349
Results ........................................................................................................................... 349
Archaeothanatology .......................................................................................................... 352
Source material and analysis strategy......................................................................... 352
Results of the archaeothanatological analysis .......................................................... 354
The nature of the deposits .................................................................................... 354
The space of decomposition of the cadaver ....................................................... 354
The initial position of the cadaver in the feature ............................................... 355
The grave features .................................................................................................. 357
Deposits containing more than one individual .................................................. 359
Post-depositional manipulations of the cadavers ............................................... 360
Other observations................................................................................................. 360
The mortuary programme at Cabeço do Pez ..................................................... 361
Sado valley: Várzea da Mó..................................................................................................... 362
Site background: a short history of research ................................................................. 362
Chronology of the burial activity .................................................................................... 364
Previous 14C measurements ........................................................................................ 364
Objectives of the dating programme ........................................................................ 364
Sampling strategy ......................................................................................................... 364
Results ........................................................................................................................... 364
Calibration and chronological models ...................................................................... 364
Stable isotopes: carbon and nitrogen.............................................................................. 365
Previous measurements: δ13C and δ15N ................................................................... 365
Sampling strategy ......................................................................................................... 365
Results ........................................................................................................................... 365
Archaeothanatology .......................................................................................................... 366
The nature of the deposit, the space of decomposition, and the initial
position of the cadaver .......................................................................................... 366
Post-deposition manipulations of the cadaver ................................................... 367
Chapter 5 Analysis and Synthesis of the Results .......................................................... 369
Introduction ............................................................................................................................ 369
Chronology of the burial activity .......................................................................................... 370
Moita do Sebastião, Tagus valley .................................................................................... 370
Chronological relationship of the burial areas ......................................................... 371
Burial activity: duration, start/end, frequency ......................................................... 372
Alternative model ................................................................................................... 375
Cabeço da Arruda, Tagus valley ...................................................................................... 377
Chronological relationships of the burial areas........................................................ 377
Burial activity: duration, start/end, frequency ......................................................... 377
Alternative model ................................................................................................... 382
Chronology of burial activity: implications for the local setting, Tagus valley ......... 383
Arapouco, Sado valley ...................................................................................................... 388
Antiquity of the site ..................................................................................................... 388
Burial activity: duration, start/end, frequency ......................................................... 388
Cabeço das Amoreiras, Sado valley ................................................................................ 388
Antiquity of the site ..................................................................................................... 388
Burial ground during the Neolithic?.......................................................................... 388
Burial activity: duration, start/end, frequency ......................................................... 389
Vale de Romeiras, Sado valley ......................................................................................... 390
Burial activity: duration, start/end, frequency ......................................................... 390
Coeval group of burials? ............................................................................................. 391
Unexpected modern 14C date of a child ................................................................... 391
Cabeço do Pez, Sado valley ............................................................................................. 392
Chronological relationship of the burial areas and isolated burials....................... 392
Burial activity: duration, start/end, frequency ......................................................... 394
Várzea da Mó, Sado valley ............................................................................................... 396
Chronology of burial activity: implications for the local setting, Sado valley ........... 397
Tagus vs Sado .................................................................................................................... 399
Isotopic signatures: defining the people in each burial place ........................................... 402
Moita do Sebastião, Tagus valley .................................................................................... 402
Cabeço da Arruda, Tagus valley ...................................................................................... 407
Cabeço da Amoreira, Tagus valley .................................................................................. 412
Isotopic signatures and implications for the local setting, Tagus valley .................... 415
Arapouco, Sado valley ...................................................................................................... 423
Cabeço das Amoreiras, Sado valley ................................................................................ 424
Vale de Romeiras, Sado valley ......................................................................................... 426
Cabeço do Pez, Sado valley ............................................................................................. 428
Várzea da Mó, Sado valley ............................................................................................... 432
Poças de S. Bento, Sado valley ........................................................................................ 432
Isotopic signatures and implications for the local setting, Sado valley ...................... 434
Tagus vs Sado .................................................................................................................... 439
The treatment of the dead: synthesis of the archaeothanatological analysis .................. 443
PART THREE: DEATH IN PLACE ................................................................................... 449
Introduction to part three ..................................................................................................... 451
Chapter 6 The Places for the Dead as Places of Shared Narratives: Discussion .... 453
Introduction ............................................................................................................................ 453
Burial as the chosen way of treating the cadaver: boundaries of the phenomenon ...... 455
Treatment of the dead ...................................................................................................... 455
Chronological and spatial boundaries............................................................................. 456
The living and the dead .................................................................................................... 459
Seasonal model of settlement pattern ....................................................................... 459
Dietary patterns ............................................................................................................ 460
Tagus valley ............................................................................................................. 460
Sado valley ............................................................................................................... 461
Lithic analyses .............................................................................................................. 462
Other data ..................................................................................................................... 463
Tagus valley ............................................................................................................. 464
Sado valley ............................................................................................................... 465
Side by side: geographic segmentation of shared landscapes................................. 467
Sado valley ............................................................................................................... 467
Tagus valley ............................................................................................................. 468
Shared landscapes ................................................................................................... 468
Death in place ......................................................................................................................... 469
Burial ground as the chosen space for the dead............................................................ 469
Cemeteries and shell middens in the Tagus and Sado valleys ..................................... 472
The dead in a daily landscape: residential or non-residential sites? ....................... 472
Areas reserved for the dead ........................................................................................ 474
Grave markers ........................................................................................................ 474
Site formation and burial activity ............................................................................... 475
Burials and shells .................................................................................................... 476
Shell symbolism ........................................................................................................... 477
Places for the dead? ..................................................................................................... 478
Mobility: permanent places for the dead, but also for the living? ......................... 479
From death to burial ......................................................................................................... 480
Transportation of the dead ......................................................................................... 482
Topographical relationships of death ............................................................................. 485
Places of mortuary narratives ................................................................................................ 485
Conclusion: on the history of death and the last hunter-gatherers of the southwestern Iberian Peninsula...................................................................................................... 487
Appendix .......................................................................................................................... 489
References ........................................................................................................................ 491
Acknowledgements
Hunter-gatherers entered my life when I was little. Long before I was born, my
dad travelled through Africa and met some of the last hunter-gatherers of this
continent. With him, he brought photographs of these encounters, offerings and
objects of exchange, and vivid memories which he shared with me and my sister as
we were growing up. I have been curious about the strange lifeways of huntergatherers since then, but only many years later did I find that pursuing
Archaeology was my way forward. My dad passed away early last summer, while I
was working in a dig in Norway and writing the sections of this dissertation about
the treatment of the dead among the last hunter-gatherers of Portugal, where I was
born. As strange and awkward as everything felt this summer, I now realize how
everything is interconnected and how through death we make sense of life.
As I write these lines I think of people and institutions to whom I wish to
express my gratitude for helping me throughout the long and sinuous path that is a
doctoral research programme.
My supervisors have been the backbone of this PhD, and I am very grateful for
their support and encouragement from the very beginning when I first approached
them, one by one, presenting the first sketches of the project Death in the Mesolithic
back in 2009–10. I was fortunate to work with three excellent archaeologists who
never gave up on me and inspired me to try always to do better. Mariana Diniz has
been my mentor since I was an undergraduate student in archaeology at the
University of Lisbon. With her, I took my first steps into the Mesolithic and Early
Neolithic. She was the first to suggest I should pursue a career in archaeological
research, which at the time I thought it was something completely out of my
league. Fortunately, I listened to her advice and ended up working hard for it, and
loving every single moment of this amazing journey. Liv Nilsson Stutz came into
my life through her groundbreaking PhD thesis. I was inspired by her work and
was ecstatic when I received a copy of her book with a dedication. Since then,
Embodied rituals and ritualized bodies (2003), has been on my reference shelf, and
which through continuous handling required the laminating of its cover. I am
extremely grateful to Liv for believing in this project from the beginning and for
everything she generously taught me throughout these years. The main supervisor
of this project is Kjel Knutsson, without whom this would have been very different and certainly a much more difficult endeavour. He has been the engine moving
this dissertation forward from day one. I am forever grateful to Kjel for taking me
on board and guiding me through the process of writing a PhD in archaeology.
Thank you.
Throughout this PhD I met some wonderful researchers who warmly
welcomed and supported me in big and small things, and some of them became
dear friends. The first person that comes to mind is Ana Cristina “Catia” Araújo. I
met her in a test pit in the shell midden of Poças de S. Bento in the Sado valley, in
the summer of 2010. Since then we became good friends and I consider her the
unofficial advisor of this dissertation. I thank her for the countless discussions on
Mesolithic hunter-gatherers while we were digging, but especially when we were
not. Catia has been the helpdesk of this project and I thank her for her immense
generosity and time spent reading and commenting on several drafts of this
dissertation.
José Paulo Ruas photographed the human burials preserved in paraffin blocks
at the National Museum of Archaeology, Lisbon, as well as all site plans. I am
thankful for his professionalism, but mostly for his friendship, support, and
delicious late dinners with Catia.
Two persons worked very close with me for the completion of this book, and
not only did they do their job beautifully, but they were also strong and supportive
when it was hard for me to move on. Susanna Berglund patiently listened to detailed descriptions of burial positions and to the effects of decomposition on a
body underground. She made the five illustrations that show the patterns of my
burial analyses, and I watched with amazement how the archaeological skeletal
remains of the past came to the present through her skilful and talented strokes. I
am thankful to Adam Zaars for posing in the required burial positions in order to
get the anatomy right. Jill Kinahan worked with me intensively during the last stage
of this project. Jill revised the text of this dissertation and her professionalism and
skill never was compromised despite the (crazy) tight schedule. I am forever
grateful for her dedication, effort, and continued support.
With Mari Tõrv I shared the ups and downs of being a PhD student. Like me,
she wrote her dissertation by examining the human remains excavated years ago,
analysing the mortuary practices of the last hunter-gatherers of Estonia, her home
country. I thank her for being such a great companion throughout this journey, the
best colleague, and a good friend, whether we met in Denmark, Sweden, Germany,
Serbia, or Turkey.
I would like to thank Amy Gray Jones, Emily Hellewell, and Sara Gummesson
for their enthusiasm for Mesolithic bones. I also wish to thank the Iberian Mesolithic
(and Early Neolithic) front of young and talented women archaeologists for the
companionship and support: Miriam Cubas, Inés Lópes-Dóriga, Helena Reis, and
Diana Nukushina.
Kim von Hackwitz co-organized the Ancient Death Ways (ADW) workshops
with me and is possibly one of the most professional persons I know. I am grateful
for our (heated) discussions, lunches, and the occasional wine. Thank you.
I wish to express my gratitude to the participants of the ADW sessions who
made these meetings so enjoyable and productive. I am especially thankful to Chris
Scarre and Åsa Larsson for their continued support and discussions. Another most
enjoyable meeting took place at the Meso Burials conference in Halle, Germany,
and I thank Judith M. Grünberg for kindly inviting me and for believing in my
work.
Mesolithic sites like Motala in Sweden are unique and I am grateful to Fredrik
Molin for the opportunity to put my hands and rubber boots in what is one of the
richest mud in the history of Mesolithic archaeology. My dear friend Emily
Sunding was the person behind this visit and she made sure I felt most welcome
into the world of Swedish archaeology.
With Joakim Goldhahn I worked on my first excavation in Sweden. I learned a
lot from him, and I am grateful for the amazing afterwork rock art experience. He
introduced me to the Swedish Bronze Age, but most importantly, to Liv Nilsson
Stutz’s work! In retrospect, I think it was thanks to Joakim that I came to do my
PhD in Sweden. Life is funny.
The Department of Archaeology and Ancient History in Uppsala was my
academic home during the course of my PhD and I am honoured to have had the
opportunity of carrying out my doctoral studies at this prestigious University. I
want to thank everybody in the Department. I especially want to thank Karl-Johan
Lindholm and Paul Sinclair, for continuous support and encouragement since the
beginning of this project, as well as Anneli Ekblom, C.-G. Ojala, Frands
Herschend, Lars Karlsson, and Paul Wallin. My PhD colleagues have also been
most supportive and I would like to thank Anna Shoemaker, Erika Lindgren
Liljenstolpe, Fredrik Tobin, Jhonny Thérus, Joakim Kjellberg, Ezekiel Mtetwa,
Patrik Klingborg, Michael Neiss, Rikke Wulff Krabbenhöft, and Therese Ekholm.
Especial thanks to Magdalena Fraser and Michel Guinard with whom I discussed
several points of my thesis.
At the neighbouring department of archaeology in Stockholm I would like to
thank Kerstin Lidén for kindly inviting me to present my research at the
Archaeological Research Laboratory. The feedback and input I received after this
seminar proved to be most useful and I am especially grateful to Gunilla Eriksson
for her comments and suggestions.
In 2014 I had the opportunity to take a short course in radiocarbon dating and
Bayesian chronological analysis at the University of Oxford. After this experience I
just fell in love with the method. I am indebted to Christopher Bronk Ramsey
along with Thomas Higham, and Michael Dee for guidance and advice on how to
deal with the particularities of my data set.
Göran Possnert at the Tandem Laboratory in Uppsala taught me how to select
and collect samples for radiocarbon dating, and inspired me to assemble my very
own toolkit. I thank him for his time and generosity. Likewise, I wish to thank Karl
Håkansson for taking the time to answer all my questions regarding reliability of
results and quality parameter checks.
I am grateful for the opportunity to participate in several courses and workshops organized by the Nordic Blade Technology Network. I learned a lot from
these meetings and I would like to warmly thank everyone involved. I especially
want to thank Berit Valentin Eriksen, Helena Knutsson, Jarmo Kankaanpää,
Mikkel Sørensen, and Tuija Rankama, for sharing their knowledge about the
pioneer settlement of Fennoscandia. I also would like to thank Astrid Nyland,
Inger Marie Berg-Hansen, Noora Taipale, and Sigrid Staurset for their companionship and discussions. Daniel Groß provided important German stationary which
greatly improved the organization of my papers!
This dissertation stands on the shoulders of researchers that before me worked
with the human remains excavated in the Mesolithic shell middens of the Tagus
and Sado valleys. I am especially grateful to Mary Jackes who gave me her full
support from day one, providing unpublished data, and continuous inspiration
from her profound knowledge and enthusiasm for these collections. Likewise, I
wish to thank Chris Meiklejohn, Mirjana Roksandic, and David Lubell, for being
there for me. Cláudia Umbelino generously provided her database of the Sado
collections. Most importantly, she shared her research experience and gave me
important advice on how to avoid several pitfalls. I am indebted, and extremely
thankful to the time she spent going back to her notes helping me out with my
(odd) questions.
Some of the last writing of this dissertation was carried out while doing fieldwork in Norway in the prehistoric sites at the E18 Tvedestrand-Arendal. Emotionally, I was not in a good shape, but I was lucky to be working with great people. I
want to warmly thank everybody in the project, especially Lars Sundström for his
support, discussions, and generosity, as well as Kim Darmark, Joachim Åkerstrøm,
and Annette Marie Strandli.
Stanley H. Ambrose, Cleia Detry, Lars Larsson, John Meadows, António
Monge Soares, Peter Rowley-Conwy, and Rick Schulting kindly provided relevant
bibliography, advice and/or personal communications when I needed it the most.
Grégor Marchand, Éva David, Elisabeth Bellon, and J. P. Cunha Ribeiro helped
me out when I was trying to locate J. Roche’s archives in France and I am grateful
for their efforts.
I wish to express my appreciation to Pedro Alvim for providing several images
used in this dissertation. I thank him for all the discussions and encouragement,
and I deeply regret that Pedro did not have the chance to finish his PhD.
Sit Tibi Terra Levis.
Work on historical collections would not be possible without the dedication of
museums and their curators. I am forever indebted to the National Museum of
Archaeology (MNA) in Lisbon for supporting this project from the very beginning.
This PhD was made possible in part due to the support of Luis Raposo, and
António Carvalho, past and current directors of MNA. I would like to thank
Ana Isabel Santos, Luísa Guerreiro, Helena Figueiredo, Maria do Carmo Vale,
Ana Melo, as well as Luis and Paulo for their precious assistance. I am thankful to
the director Miguel Ramalho of the Geological Museum (MG) in Lisbon, and to
José António Anacleto. I thank Maria José Cunha at the Museum of Natural
History, University of Porto (MHNUP), for her commitment and continued
support. I also wish to thank Marta Lourenço and Hugo Cardoso at the National
Museum of Natural History and Science, University of Lisbon (MUNHAC) for
their assistance.
***
This research was funded by The Portuguese National Science Foundation (FCT,
Ministério da Ciência, Tecnologia e Ensino Superior). I was awarded with an individual
doctoral grant (SFRH/BD/72758/2010) which gave me the lifetime opportunity
to dedicate my professional life (2010–15) to archaeological research exclusively.
Financial support for radiocarbon and isotopic analyses, and scientific illustrations was generously provided by the Societas Archaeologica Upsaliensis (SAU,
Uppsala), Birgit och Gad Rausings Stiftelse för Humanistisk Forskning, Valsgärdefonden/Mårten Stenbergers fond (Uppsala University), and Helge Ax:son Johnsons Stiftelse.
***
On a more personal level, I wish to thank my sister Ana Romeiro for her
unconditional friendship and for kindly providing shelter for me and for my bone
samples, always when needed.
Watain has been in my life for many years. This dissertation was planned,
thought, and partially written while in the studio or working in gigs with them.
Some of my best ideas came while being in this environment, and I can never
forget the creative blast while Watain were recording the song Total Funeral for the
Lawless Darkness album. Thank you for being a part of me.
My last words are for Tore, my husband and best friend. You walked with me
throughout this extraordinary journey, and I love you for that. It was hard, but I
think we made it.
Svartsjö, 11th January 2016
Rita Peyroteo Stjerna
Abbreviations
APMH: Personal Archive of Manuel Heleno, Arquivo Pessoal de Manuel Heleno
Excav.: excavation
Ind.: individual
m.a.s.l: metres above sea level
MG: Geological Museum, Museu Geológico (Lisbon, Portugal)
MHNUP: Museum of Natural History, University of Porto, Museu de História
Natural da Universidade do Porto (Porto, Portugal)
MNA: National Museum of Archaeology, Museu Nacional de Arqueologia (Lisboa,
Portugal)
MNI: minimum number of individuals
MUNHAC: National Museum of Natural History and Science, Museu Nacional de
História Natural e da Ciência da Universidade de Lisboa (Lisbon, Portugal)
n/a: not available
n/d: not determinable
n/p: not published
Sk/Sks: skeleton/skeletons
Archaeological sites
ARA: Arapouco, Sado valley
CAM: Cabeço da Amoreira, Tagus valley
CAMS: Cabeço das Amoreiras, Sado valley
CAR: Cabeço da Arruda, Tagus valley
CP: Cabeço do Pez, Sado valley
MS: Moita do Sebastião, Tagus valley
VM: Várzea da Mó, Sado valley
VR: Vale de Romeiras, Sado valley
Chemical elements mentioned in the text
Ba: barium
C: carbon
Ca: calcium
Mg: magnesium
N: nitrogen
Sr: strontium
V: vanadium
Zn: zinc
xxiii
List of Figures
Figure 1.1. The archaeological study of death through its material remains. .......................... 48
Figure 2.1. Iberian Peninsula and the location of the shell midden sites of the Tagus and
Sado valleys, Portugal. ........................................................................................................ 62
Figure 2.2. Tagus valley: shell midden sites with human remains. ............................................ 66
Figure 2.3. Sado valley: shell midden sites with human remains. .............................................. 66
Figure 2.4. Excavation report of Cabeço do Pez 1959, page 1. ................................................ 70
Figure 2.5. Poças de S. Bento, Sado valley: site plan. Area with the human remains
excavated in 1960. ............................................................................................................... 76
Figure 2.6. Excavations at the Sado middens were among the earliest sites where the
Wheeler method was applied in Portugal. ....................................................................... 87
Figure 2.7. Daily life in a Muge shell midden, Tagus valley, Portugal. ..................................... 90
Figure 3.1. Calculation of the percentage of marine resources in a diet, estimated from
the δ13C or δ15N value of bone collagen. ....................................................................... 103
Figure 3.2. Mathematical equation: linear interpolation between two known points. ......... 104
Figure 3.3. Graphic representation of the calculation of percentage of marine foods in
a diet. ................................................................................................................................... 105
Figure 3.4. Application of linear interpolation equation for the calculation of the
percentage of marine foods in a diet. ............................................................................. 105
Figure 3.5. Example of a probability density function. ............................................................ 107
Figure 3.6. Bayes’s theorem applied to archaeology. ................................................................ 108
Figure 3.7. Example of informative prior beliefs (or, the archaeology) incorporated in a
Bayesian chronological model. ........................................................................................ 109
Figure 3.8. Example of a posterior density estimate which occupies part of the
calibrated probability distribution. .................................................................................. 112
Figure 3.9. Sensitivity analysis of reservoir offsets. ................................................................... 117
Figure 3.10. Archaeological contexts with human remains. Identification and
interpretation of cultural phenomena with the methods of archaeothanatology. .... 134
Figure 3.11. Classification of joints in archaeothanatology. .................................................... 139
Figure 3.12. Place of disposal of the cadaver. ............................................................................ 141
Figure 3.13. The two basic principles of archaeothanatology and the reconstruction of
aspects of the funerary practices. .................................................................................... 142
Figure 3.14. Terminology: a cemetery accommodates more than one funerary structure. . 154
Figure 4.1. Moita do Sebastião, Tagus valley: excavation areas with human skeletons. ...... 164
Figure 4.2. Moita do Sebastião, Tagus valley, 1952–54: excavated area with human
remains and individuals selected for 14C measurements. ............................................. 168
Figure 4.3. Chronological model for the burial activity at Moita do Sebastião, calibrated
using marine 1 diet values. ............................................................................................... 172
Figure 4.4. Start and End of burial activity at Moita do Sebastião derived from the
chronological model, calibrated with values for marine 1 diet. .................................. 173
Figure 4.5. Duration of burial activity at Moita do Sebastião derived from the
chronological model, calibrated with values for marine 1 diet. .................................. 173
xxv
Figure 4.6. Moita do Sebastião, Tagus valley: correlation between the δ13C (‰) and
δ15N (‰) values of 20 samples of human bone collagen from 18 individuals. ........ 176
Figure 4.7. MS1952–54, Sk9 was the only cadaver lying on the back with the lower limbs
in complete extension. ...................................................................................................... 190
Figure 4.8. The upper body of MS1952–54, Sk30 presents strong medial tendency. .......... 191
Figure 4.9. MS1952–54, Sk5. The opening of the lower portion of the rib cage indicates that
the grave pit was wide enough to allow the expansion of the skeletal elements. ..... 192
Figure 4.10. MS1952–54, Sk12. The position of the upper right limb in abduction
suggests that the grave pit was relatively wide at this level. ......................................... 192
Figure 4.11. MS1952–54, Sk15. ................................................................................................... 193
Figure 4.12. Cabeço da Arruda, Tagus valley 1964–65: site plan with the location of the
skeletons. ............................................................................................................................ 203
Figure 4.13. Probability density function for the calibration of two bone collagen
samples (Beta–127451 and AA–101343) from CAR1937, Sk6................................... 207
Figure 4.14. Chronological model for the burial activity at Cabeço da Arruda, calibrated
using marine 1 diet values. For each of the dates two distributions have been
plotted, one in outline which is the result produced by the scientific evidence alone,
and a solid one which is based on the chronological model used. Calibrated ranges
of 95% and 68% probability are given by the lower and higher square brackets
respectively. The large square brackets (left) and the OxCal keywords define the
overall model. Model-definition command files in appendix, Tagus valley, Cabeço
da Arruda, Model 1, Marine 1. ............................................................................................ 208
Figure 4.15. Start and End of burial activity at Cabeço da Arruda derived from the
chronological model, calibrated with values for marine 1 diet. .................................. 209
Figure 4.16. Duration of burial activity at Cabeço da Arruda derived from the
chronological model, calibrated with values for marine 1 diet. .................................. 209
Figure 4.17. Cabeço da Arruda, Tagus valley: correlation between the δ13C (‰) and
δ15N (‰) values of 16 samples of human bone collagen from 14 individuals. ........ 212
Figure 4.18. Field photograph of the human burials excavated in 1964 at Cabeço da
Arruda, Tagus valley. ........................................................................................................ 216
Figure 4.19. Field sketch by O. Veiga Ferreira showing the relative spatial position of the
burials excavated in 1964 at Cabeço da Arruda, Tagus valley. .................................... 216
Figure 4.20. Nineteenth century excavations at Cabeço da Arruda, Tagus valley. ............... 217
Figure 4.21. The only site plan available for Arapouco, Sado valley. ..................................... 229
Figure 4.22. Re-identification of individual from Arapouco, Sado valley. ............................. 231
Figure 4.23. This paraffin block was mislabelled Arapouco, Skeleton 5. ................................... 232
Figure 4.24. Probability density function estimated for the death of ARA1962, Sk2A ....... 235
Figure 4.25. Arapouco, Sado valley: correlation between the δ13C (‰) and δ15N (‰)
values of 12 samples of human bone collagen. ............................................................. 237
Figure 4.26. ARA1961, Sk5 illustrates several common principles governing burial
practice at Arapouco, Sado valley. .................................................................................. 253
Figure 4.27. ARA1961, Sk5 presents typical patterns of decomposition of the body in a
sediment-filled environment. ........................................................................................... 254
Figure 4.28. The CV of ARA1961, Sk6 present a similar movement pattern as observed
in ARA1961, Sk5. .............................................................................................................. 255
Figure 4.29. ARA1961, Sk6. The striking displacement upwards of the upper right limb
contrasts with the minor movements observed in this feature. .................................. 257
Figure 4.30. ARA1961, Sk7 presents a displacement upwards of the right shoulder girdle
similar to the movement pattern observed in ARA1961, Sk6. ................................... 259
xxvi
Figure 4.31. ARA1961, Sks 11 (adult female) and 12 (child, c 3 yrs) were buried
simultaneously. .................................................................................................................. 260
Figure 4.32. Reconstruction hypothesis of the initial position of individuals 11 and 12
in the grave feature, Arapouco, Sado valley................................................................... 262
Figure 4.33. ARA1961, Sk13 shares the basic common characteristics observed in all
burials at Arapouco, but was placed in an unusual position. ...................................... 264
Figure 4.34. This illustration presents several key elements observed in the burials at
Arapouco. ........................................................................................................................... 266
Figure 4.35. Cabeço das Amoreiras, Sado valley: site plan and profiles of the main
excavation area (Talhão T) where all human remains were found (I to VIII)............ 268
Figure 4.36. Cabeço das Amoreiras, Sado valley: site plan of the excavated areas. .............. 269
Figure 4.37. Chronological model for the burial activity at Cabeço das Amoreiras,
calibrated using marine 1 diet values. ............................................................................ 273
Figure 4.38. Start of burial activity at Cabeço das Amoreiras derived from the
chronological model 1, calibrated with values for marine 1 diet. ............................... 274
Figure 4.39. End of burial activity at Cabeço das Amoreiras derived from the
chronological model 1, calibrated with values for marine 1 diet. ............................... 274
Figure 4.40. Duration of burial activity at Cabeço das Amoreiras derived from the
chronological model, calibrated with values for marine 1 diet. .................................. 275
Figure 4.41. Cabeço das Amoreiras, Sado valley: correlation between the δ13C (‰) and
δ15N (‰) values of 10 samples of human bone collagen from 5 individuals. .......... 277
Figure 4.42. CAMS1958, Sk7. The transverse compression exerted on the skeleton
indicates significant physical limits on the cadaver....................................................... 286
Figure 4.43. Cabeço das Amoreiras, Sado valley, 1958: field photograph of individual 3
(complete skeleton) and individual 2 (cranium). ........................................................... 289
Figure 4.44. Cabeço das Amoreiras, Sado valley, 1958: field photograph of individual 2
(cranium). ........................................................................................................................... 290
Figure 4.45. Vale de Romeiras, Sado valley: site plan and profiles of the excavation area. . 293
Figure 4.46. Chronological model for the burial activity at Vale de Romeiras, calibrated
using marine 1 diet values. ............................................................................................... 300
Figure 4.47. Start of burial activity at Vale de Romeiras derived from the chronological
model 1, calibrated with values for marine 1 diet. ........................................................ 300
Figure 4.48. End of burial activity at Vale de Romeiras derived from the chronological
model 1, calibrated with values for marine 1 diet. ........................................................ 301
Figure 4.49. Duration of burial activity at Vale de Romeiras derived from the
chronological model, calibrated with values for marine 1 diet values. ....................... 301
Figure 4.50. Vale de Romeiras, Sado valley: correlation between the δ13C (‰) and δ15N
(‰) values of 9 samples of human bone collagen from 7 individuals. ...................... 303
Figure 4.51. VR1959, Sk5. Although poorly preserved this case is suggestive of a mixed
space of decomposition.................................................................................................... 309
Figure 4.52. VR1959, Sk6. The striking displacement upwards of the upper left limb
contrasts with the minor movements observed in the feature.................................... 311
Figure 4.53. Profile A–A', Vale de Romeiras, Sado valley. ...................................................... 320
Figure 4.54. Profile B–B', Vale de Romeiras, Sado valley. ....................................................... 321
Figure 4.55. Profile E-E', Vale de Romeiras, Sado valley, shows the close proximity and
partial overlapping of the grave features VR1959, Sks 9, 12. ...................................... 322
Figure 4.56. VR1959, Sk12 contains the human remains of an adult male. .......................... 323
Figure 4.57. Detail of the site plan of VR1959, Sks 9, 12, showing disarticulated long
bones on the left side of Sk9. .......................................................................................... 324
xxvii
Figure 4.58. VR1959, Sk9 and disarticulated long bones (one femur, and one tibia?)
on his left side. ................................................................................................................... 324
Figure 4.59. VR1959, Sk13. This feature is poorly preserved possibly due to its proximity
to the modern surface of the site (c 10 cm deep). ........................................................ 325
Figure 4.60. VR1959, Sk1. The cranium and mandible are not in the feature. ..................... 326
Figure 4.61. VR1959, Sk8. Grave feature with the remains of a 3–4 yrs ± 12 months
child..................................................................................................................................... 328
Figure 4.62. VR1959, Sk11 is representative of a typical grave feature at Vale de
Romeiras, Sado valley. ...................................................................................................... 330
Figure 4.63. VR1959, Sk7 is a typical example of a hypercontracted skeleton commonly
observed in the grave features at Vale de Romeiras, Sado valley. .............................. 332
Figure 4.64. VR1959, Sk23 is another example of a hypercontracted skeleton although
placed in an atypical square shaped burial position. ..................................................... 333
Figure 4.65. This illustration presents the key elements observed in two burials at Vale
de Romeiras (12, 19) and one at Cabeço das Amoreiras (7). ...................................... 335
Figure 4.66. This illustration presents several typical elements observed in the burials at
Vale de Romeiras. ............................................................................................................. 337
Figure 4.67. Cabeço do Pez, Sado valley: site plan of the areas excavated in 1956. ............. 340
Figure 4.68. CP1959, Sk1. ............................................................................................................ 341
Figure 4.69. Chronological model for the burial activity at Cabeço do Pez, calibrated
using marine 1 diet values. ............................................................................................... 345
Figure 4.70. Start and End of burial activity at Cabeço do Pez derived from the
chronological model, calibrated with values for marine 1 diet. .................................. 345
Figure 4.71. Start of burial activity at Cabeço do Pez derived from the chronological
model, calibrated with values for marine 1 diet. ........................................................... 347
Figure 4.72. End of burial activity at Cabeço do Pez derived from the chronological
model, calibrated with values for marine 1 diet. ........................................................... 347
Figure 4.73. Duration of burial activity at Cabeço do Pez derived from the chronological
model, calibrated with values for marine 1 diet. ........................................................... 348
Figure 4.74. Cabeço do Pez, Sado valley: correlation between the δ13C (‰) and δ15N
(‰) values of 27 samples of human bone collagen from 21 individuals. ................. 352
Figure 4.75. CP1956, Sk9 in two blocks of paraffin. The distal end of the right humerus
of CP1956, Sk10 is preserved in the block with skeleton 9......................................... 359
Figure 4.76. Burial area B at Cabeço do Pez, Sado valley. ....................................................... 360
Figure 4.77. Burial area A at Cabeço do Pez, Sado valley. ....................................................... 361
Figure 4.78. Várzea da Mó, Sado valley: site plan and profile of the area with human
remains. .............................................................................................................................. 363
Figure 4.79. VM1959, Sk1 preserved in paraffin block at MNA. ........................................... 363
Figure 4.80. Probability density function estimated for the death of VM1959, Sk1 ............ 365
Figure 4.81. Várzea da Mó, Sado valley: correlation between the δ13C (‰) and δ15N
(‰) values of one sample of human bone collagen from the only burial known
at the site. ........................................................................................................................... 366
Figure 5.1. Moita do Sebastião, Tagus valley, 1952–54: site plan and 14C dated human
remains. .............................................................................................................................. 371
Figure 5.2. Chronological model for the burial activity at Moita do Sebastião, as
presented in chapter 4. ..................................................................................................... 372
Figure 5.3. MS1952–54, Sks 5, 9, 12, 15: 14C data. .................................................................... 373
Figure 5.4. Hypothetical chronological model 2 for the burial activity at Moita do
Sebastião, Tagus valley. .................................................................................................... 376
xxviii
Figure 5.5. Probability density function for the calibration of sample AA–101343
(Jackes et al. 2014) from bone collagen of CAR1937, Sk6. ......................................... 378
Figure 5.6. Probability density function for the calibration of sample Beta–127451
(Cunha and Cardoso 2002–03; Cunha et al. 2003) from bone collagen of
CAR1937, Sk6. .................................................................................................................. 378
Figure 5.7. Chronological model for the burial activity at Cabeço da Arruda, calibrated
using marine 1 diet values. ............................................................................................... 379
Figure 5.8. Position of the burials excavated in 1937 and 1964–65 relatively to the
bottom of the shell midden, Cabeço da Arruda, Tagus valley. ................................... 383
Figure 5.9. Chronological models for the burial activity at Moita do Sebastião and
Cabeço da Arruda, calibrated using marine 1 diet values. ........................................... 384
Figure 5.10. Chronological model for the burial activity at Cabeço da Amoreira,
calibrated using marine 1 diet values. ............................................................................. 385
Figure 5.11. Chronological model for the burial activity at Moita do Sebastião, Cabeço
da Arruda and Cabeço da Amoreira in the Tagus valley. ............................................ 387
Figure 5.12. Cabeço das Amoreiras, Sado valley: site plan and 14C dates. ............................. 389
Figure 5.13. Vale de Romeiras, Sado valley: site plan and 14C dates. ...................................... 391
Figure 5.14. Cabeço do Pez, Sado valley: site plan and 14C dates. .......................................... 393
Figure 5.15. Chronological model for the burial activity at Cabeço do Pez without the
later date Ua–46930 (CP1956, Sk2), calibrated using marine 1 diet values. .............. 394
Figure 5.16. Duration of burial activity at Cabeço do Pez derived from the chronological
model run without Ua–46930, CP1956, Sk2, calibrated with values for marine
1 diet. .................................................................................................................................. 395
Figure 5.17. Start and End of burial activity at Cabeço do Pez derived from the
chronological model run without Ua–46930, CP1956, Sk2, calibrated with values
for marine 1 diet. ............................................................................................................... 395
Figure 5.18. Chronological model for the burial activity at Arapouco, Cabeço das Amoreiras,
Vale de Romeiras, Cabeço do Pez and Várzea da Mó at the Sado valley. ................. 398
Figure 5.19. Chronological model for the burial activity at the Tagus (Muge) and Sado
valleys. ................................................................................................................................. 401
Figure 5.20. Moita do Sebastião, Tagus valley: variation of δ13C (‰) over time (cal BCE,
95% probability) of 21 samples of human bone collagen from 18 individuals. ....... 406
Figure 5.21. Moita do Sebastião, Tagus valley: variation of δ15N (‰) over time (cal BCE,
95% probability) of 20 samples of human bone collagen from 18 individuals. ....... 406
Figure 5.22. Cabeço da Arruda, Tagus valley: variation of δ13C (‰) over time (cal BCE, 95%
probability) of 16 samples of human bone collagen from 14 individuals. ................ 411
Figure 5.23. Cabeço da Arruda, Tagus valley: variation of δ15N (‰) over time (cal BCE,
95% probability) of 16 samples of human bone collagen from 14 individuals. ....... 411
Figure 5.24. Cabeço da Amoreira, Tagus valley: variation of δ13C (‰) over time (cal BCE,
95% probability) of 5 samples of human bone collagen from 5 individuals............. 414
Figure 5.25. Cabeço da Amoreira, Tagus valley: variation of δ15N (‰) over time (cal BCE,
95% probability) of 4 samples of human bone collagen from 4 individuals............. 414
Figure 5.26. Moita do Sebastião, Cabeço da Arruda and Cabeço das Amoreiras, Tagus
valley: bone collagen δ13C over time (cal BCE, 95% probability). ............................. 416
Figure 5.27. Moita do Sebastião, Cabeço da Arruda and Cabeço das Amoreiras, Tagus
valley: bone collagen δ15N over time (cal BCE, 95% probability).............................. 416
Figure 5.28. Maps 1 and 2 represent temporal moments of burial activity at the Tagus
valley and display the 14C dated human burials at each site......................................... 418
Figure 5.29. Maps 3 and 4 represent temporal moments of burial activity at the Tagus
valley and display the 14C dated human burials at each site......................................... 419
xxix
Figure 5.30. Map 5 represents temporal moments of burial activity at the Tagus valley
and display the 14C dated human burials at each site. .................................................. 421
Figure 5.31. Map 6 represents temporal moments of burial activity at the Tagus valley
and display the 14C dated human burials at each site. .................................................. 422
Figure 5.32. Cabeço das Amoreiras, Sado valley: bone collagen δ13C over time
(cal BCE, 95% probability). ............................................................................................. 425
Figure 5.33. Cabeço das Amoreiras, Sado valley: bone collagen δ15N over time
(cal BCE, 95% probability). ............................................................................................. 425
Figure 5.34. Vale de Romeiras, Sado valley: bone collagen δ13C over time
(cal BCE, 95% probability). ............................................................................................. 427
Figure 5.35. Vale de Romeiras, Sado valley: bone collagen δ15N over time (cal BCE,
95% probability). ............................................................................................................... 427
Figure 5.36. Cabeço do Pez, Sado valley: variation of δ13C (‰) over time (cal BCE,
95% probability) of 11 samples of human bone collagen from 5 individuals. ......... 430
Figure 5.37. Cabeço do Pez, Sado valley: variation of δ15N (‰) over time (cal BCE,
95% probability) of 10 samples of human bone collagen from 5 individuals. ......... 430
Figure 5.38. Poças de S. Bento, Sado valley: correlation between the δ13C (‰) and δ15N
(‰) values of 9 samples of human bone collagen from 8 individuals. ...................... 434
Figure 5.39. Sado valley: correlation between the δ13C (‰) and δ15N (‰) values of 66
samples of human bone collagen from 53 individuals. ................................................ 435
Figure 5.40. Sado valley: bone collagen δ13C over time (cal BCE, 95% probability). .......... 437
Figure 5.41. Sado valley: bone collagen δ15N over time (cal BCE, 95% probability)........... 437
Figure 5.42. Sites with human remains at the Sado valley: minimum number of individuals
(MNI), mean δ13C (‰) and δ15N (‰) values, and respective standard deviations
derived from human bone collagen. ............................................................................... 438
Figure 5.43. Individuals buried in the Tagus and Western Sado valleys derived their isotopes
from a mixed diet of terrestrial and marine-estuarine resources (M) while the people
buried in the East Sado obtained their protein from terrestrial food sources (T). ... 440
Figure 5.44. Tagus and Sado valleys: correlation between the δ13C (‰) and δ15N (‰)
values from human bone collagen samples. .................................................................. 442
Figure 5.45. This illustration presents several typical elements observed in the burials
at the Tagus and Sado valleys. ......................................................................................... 447
Figure 6.1. Chronological data for the burial activity in the Tagus and Sado valleys
plotted with two major events: palaeoclimatic change 8200 years ago (c 6250 BCE)
and the onset of the Neolithic in Portugal. ................................................................... 458
Figure 6.2. Burial grounds in the Sado valley with indication of sources of protein
(mixed/terrestrial) and frequency of caries (F.C.) of the burial populations. ........... 466
Figure 6.3. Moita do Sebastião, Tagus valley: reinterpretation of the structures identified
by J. Roche (1972) at Moita do Sebastião. ..................................................................... 473
Figure 6.4. Burials at Arapouco in the Sado valley, lying in close proximity without
disturbance. ........................................................................................................................ 474
Figure 6.5. Tagus valley, Moita do Sebastião, 1952–54, individual 21.................................... 476
Figure 6.6. Profile F–F’, Vale de Romeiras, Sado valley. ......................................................... 477
Figure 6.7. From death to the burial of the individual: post-mortem changes of the
cadaver. ............................................................................................................................... 481
Figure 6.8. 60 and 30 km radii from a central point in the regions of the Tagus and Sado
shell middens. .................................................................................................................... 484
xxx
List of Tables
Table 2.1. Tagus valley: 13 shell midden sites identified, nine of which with recorded human remains. .. 65
Table 2.2. Sado valley: 11 shell midden sites identified, six of which with known human remains. ....... 65
Table 2.3. Museums curating archaeological material excavated at the shell middens of the Tagus and
Sado valleys. ............................................................................................................................ 68
Table 2.4. Death rate among hunter-gatherers. ................................................................................... 80
Table 3.1. Sado valley: human remains with radiocarbon dates published before 2013. ........................ 94
Table 3.2. Tagus valley: human remains with radiocarbon dates published before 2013. ...................... 94
Table 3.3. Reference record of reservoir ages for the Portuguese coast and estuaries. ............................. 102
Table 3.4. Measurements of problematic calibration from samples on human bone collagen. ................ 123
Table 3.5. Arguments supporting the nature of a deposit in primary position. .................................... 143
Table 3.6. Arguments supporting the space of decomposition in a sediment-filled environment. ............ 145
Table 3.7. Funerary contexts in one single structure. ......................................................................... 153
Table 4.1. Estimated number of individuals excavated each year at Moita do Sebastião, Tagus valley. 163
Table 4.2. The various MNI estimates proposed for Moita do Sebastião, Tagus valley. ..................... 164
Table 4.3. Moita do Sebastião, Tagus valley: samples of human bone (14) collected and submitted
for 14C measurement in the course of this study. ....................................................................... 167
Table 4.4. Human bone collagen samples (19) from Moita do Sebastião, Tagus valley: 14C
measurements and estimation of non-terrestrial carbon intake (% marine). ............................... 169
Table 4.5. Human bone collagen samples (18) from Moita do Sebastião, Tagus valley: calibrated
date ranges according to two different approaches for the calculation of the marine percentage
(M1, M2). ............................................................................................................................ 174
Table 4.6. Posterior density estimates for the dates of archaeological events (Start/End) and the
duration of burial activity at Moita do Sebastião. .................................................................... 175
Table 4.7. Moita do Sebastião, Tagus valley: carbon (δ13C) and nitrogen (δ15N) measurements on
human bone collagen samples. ................................................................................................. 177
Table 4.8. Moita do Sebastião, Tagus valley: source material used for the archaeothanatological
analysis. ................................................................................................................................ 180
Table 4.9. Summary of the nature of the deposits with human remains at Moita do Sebastião,
Tagus valley. .......................................................................................................................... 181
Table 4.10. Summary of the space of decomposition of the human remains in the grave features at
Moita do Sebastião, Tagus valley. .......................................................................................... 185
Table 4.11. Summary of the reconstruction of the initial position of the cadaver in the grave features
at Moita do Sebastião, Tagus valley. ...................................................................................... 187
Table 4.12. Summary of the key observations used in the reconstruction of the grave features at Moita
do Sebastião, Tagus valley. ..................................................................................................... 194
Table 4.13. Estimated number of individuals excavated each year at Cabeço da Arruda, Tagus
valley. .................................................................................................................................... 200
Table 4.14. Cabeço da Arruda, Tagus valley: samples of human bone (6) collected and submitted
for 14C measurement in the course of this study. ....................................................................... 204
Table 4.15. Human bone collagen samples (15 samples, 14 individuals) from Cabeço da Arruda,
Tagus valley. .......................................................................................................................... 205
xxxi
Table 4.16. Human bone collagen samples (14+1) from Cabeço da Arruda, Tagus valley: calibrated
date ranges according to two different approaches for the calculation of the marine percentage
(M1, M2). ............................................................................................................................ 210
Table 4.17. Posterior density estimates for the dates of archaeological events (Start/End) and the
duration of burial activity at Cabeço da Arruda. ..................................................................... 211
Table 4.18. Cabeço da Arruda, Tagus valley: carbon (δ13C) and nitrogen (δ15N) measurements on
human bone collagen samples. ................................................................................................. 213
Table 4.19. Cabeço da Arruda, Tagus valley: source material assessed for archaeothanatological
analysis. ................................................................................................................................ 218
Table 4.20. Summary of the nature of the deposits with human remains at Cabeço da Arruda,
Tagus valley. .......................................................................................................................... 220
Table 4.21. Summary of the space of decomposition of the human remains in the grave features at
Cabeço da Arruda, Tagus valley. ........................................................................................... 221
Table 4.22. Summary of the reconstruction of the initial position of the cadaver in the grave features
at Cabeço da Arruda, Tagus valley. ....................................................................................... 224
Table 4.23. Summary of the key observations used in the reconstruction of the grave features at Cabeço
da Arruda, Tagus valley. Observed pressures on the body are presented in a number sequence.
The arguments supporting these observations are indicated by the same number in the respective
column. .................................................................................................................................. 226
Table 4.24. Arapouco, Sado valley: minimum number of individuals (MNI) estimated for the
material curated at MNA. .................................................................................................... 229
Table 4.25. Arapouco, Sado valley: sex and age estimates of 22 individuals. ..................................... 230
Table 4.26. Samples of human bone (9) from Arapouco collected and submitted for 14C measurement
in the course of this study. ....................................................................................................... 234
Table 4.27. Human bone collagen sample (1) from Arapouco, Sado valley. ....................................... 234
Table 4.28. Human bone collagen sample from Arapouco, Sado valley: calibrated date ranges
according to two different approaches for the calculation of the marine percentage. ....................... 235
Table 4.29. Arapouco, Sado valley: carbon (δ13C) and nitrogen (δ15N) measurements on human
bone collagen samples (12) collected at MNA. ........................................................................ 237
Table 4.30. Arapouco, Sado valley: source material used for the archaeothanatological analysis. ......... 239
Table 4.31. Summary of the nature of the deposits with human remains at Arapouco, Sado valley. .... 240
Table 4.32. Summary of the space of decomposition of the human remains in the grave features at
Arapouco, Sado valley. .......................................................................................................... 242
Table 4.33. Summary of the reconstruction of the initial position of the cadaver in the grave features
at Arapouco, Sado valley. ...................................................................................................... 245
Table 4.34. Summary of the key observations used in the reconstruction of the grave features at
Arapouco, Sado valley. .......................................................................................................... 249
Table 4.35. Cabeço das Amoreiras, Sado valley: sex and age estimates. ............................................ 270
Table 4.36. Cabeço das Amoreiras, Sado valley: stratigraphy and depth of the human burials in
relation to top layer. ............................................................................................................... 270
Table 4.37. Samples of human bone (4) from Cabeço das Amoreiras collected and submitted for 14C
measurement in the course of this study.................................................................................... 272
Table 4.38. Human bone collagen samples (3) from Cabeço das Amoreiras, Sado valley. .................. 272
Table 4.39. Human bone collagen samples (3) from Cabeço das Amoreiras, Sado valley: calibrated
date ranges according to two different approaches for the calculation of the marine percentage
(M1, M2). ............................................................................................................................ 275
Table 4.40. Cabeço das Amoreiras, Sado valley: carbon (δ13C) and nitrogen (δ15N) measurements
on human bone collagen samples collected at MNA. ................................................................ 278
Table 4.41. Cabeço das Amoreiras, Sado valley: source material used for archaeothanatological
analysis. ................................................................................................................................ 279
xxxii
Table 4.42. Summary of the nature of the deposits with human remains at Cabeço das Amoreiras,
Sado valley. ........................................................................................................................... 282
Table 4.43. Summary of the space of decomposition of the human remains in the grave features at
Cabeço das Amoreiras, Sado valley. ....................................................................................... 283
Table 4.44. Summary of the reconstruction of the initial position of the cadaver in the grave feature at
Cabeço das Amoreiras, Sado valley. ....................................................................................... 284
Table 4.45. Summary of the key observations used in the reconstruction of the grave features at Cabeço
das Amoreiras, Sado valley. ................................................................................................... 288
Table 4.46. Vale de Romeiras, Sado valley: minimum number of individual (MNI) estimated for
the material currently curated at MNA. ................................................................................. 294
Table 4.47. Vale de Romeiras, Sado valley: sex and age estimates of 20 individuals. ........................ 294
Table 4.48. Vale de Romeiras, Sado valley: stratigraphy and depth of the human burials. ................. 295
Table 4.49. Samples of human bone (13) from Vale de Romeiras collected and submitted for 14C
measurement in the course of this study.................................................................................... 297
Table 4.50. Human bone collagen (3) samples from Vale de Romeiras, Sado valley. ......................... 298
Table 4.51. Human bone collagen (2) samples from Vale de Romeiras, Sado valley: calibrated date
ranges according to two different approaches for the calculation of the marine percentage
(M1, M2). ............................................................................................................................ 299
Table 4.52. Vale de Romeiras, Sado valley: carbon (δ13C) and nitrogen (δ15N) measurements on
human bone collagen samples collected at MNA. .................................................................... 304
Table 4.53. Vale de Romeiras, Sado valley: all individuals identified and excavated (n=23) and
the respective source material used for the archaeothanatological analysis.................................... 306
Table 4.54. Summary of the nature of the deposits with human remains at Vale de Romeiras,
Sado valley. ........................................................................................................................... 307
Table 4.55. Summary of the space of decomposition of the human remains in the grave features at
Vale de Romeiras, Sado valley. .............................................................................................. 312
Table 4.56. Summary of the reconstruction of the initial position of the cadaver in the grave features
at Vale de Romeiras, Sado valley. .......................................................................................... 315
Table 4.57. Summary of the key observations used in the reconstruction of the grave features at Vale
de Romeiras, Sado valley. ....................................................................................................... 318
Table 4.58. Cabeço do Pez, Sado valley: minimum number of individual (MNI) estimated for the
material curated at MNA. .................................................................................................... 338
Table 4.59. Cabeço do Pez, Sado valley: sex and age estimates of 20 individuals............................... 339
Table 4.60. Samples of human bone (5) from Cabeço do Pez collected and submitted for 14C
measurement in the course of this study.................................................................................... 343
Table 4.61. Human bone collagen samples (5) from Cabeço do Pez, Sado valley: 14C measurements
and estimation of non-terrestrial carbon intake (% marine). ..................................................... 344
Table 4.62. Human bone collagen samples (5) from Cabeço do Pez, Sado valley: calibrated date
ranges according to two different approaches for the calculation of the marine percentage. ............ 346
Table 4.63. Posterior density estimates for the dates of archaeological events (Start/End) and the
duration of burial activity at Cabeço do Pez. ........................................................................... 346
Table 4.64. Cabeço do Pez, Sado valley: carbon (δ13C) and nitrogen (δ15N) measurements on human
bone collagen samples. ............................................................................................................ 350
Table 4.65. Cabeço do Pez, Sado valley: list of individuals considered for archaeothanatological
analysis, based on the preservation of the material. ................................................................... 353
Table 4.66. Summary of the nature of the deposits with human remains at Cabeço do Pez, Sado
valley. .................................................................................................................................... 354
Table 4.67. Summary of the space of decomposition of the human remains in the grave features at
Cabeço do Pez, Sado valley. ................................................................................................... 355
xxxiii
Table 4.68. Summary of the reconstruction of the initial position of the cadaver in the grave features
at Cabeço do Pez, Sado valley. ............................................................................................... 356
Table 4.69. Summary of the key observations used in the reconstruction of the grave features at
Cabeço do Pez, Sado valley. ................................................................................................... 358
Table 4.70. Sample of human bone (1) from Várzea da Mó collected and submitted for 14C
measurement in the course of this study.................................................................................... 364
Table 4.71. Human bone collagen sample (1) from Várzea da Mó, Sado valley: 14C measurement
and estimation of non-terrestrial carbon intake (% marine). ..................................................... 364
Table 4.72. Human bone collagen sample from Várzea da Mó, Sado valley: calibrated date ranges
according to two different approaches for the calculation of the marine percentage. ....................... 365
Table 4.73. Várzea da Mó, Sado valley: source material used for archaeothanatological analysis. ...... 366
Table 5.1. Representativeness of the archaeological burial population at Moita do Sebastião in
relation to the original group. .................................................................................................. 375
Table 5.2. Cabeço da Arruda, Tagus valley: stratigraphy and burial activity at the site. .................... 380
Table 5.3. Representativeness of the archaeological burial population at Cabeço da Arruda in
relation to the original group. .................................................................................................. 381
Table 5.4. Posterior density estimates for the dates of archaeological events (Start/End) and the
duration of burial activity at Cabeço do Pez. ........................................................................... 395
Table 5.5. Representativeness of the archaeological burial population at Cabeço do Pez in relation
to the original group. .............................................................................................................. 396
Table 5.6. Moita do Sebastião, Tagus valley: chronology and stable isotope values (δ13C and δ15N)
of human bone collagen. .......................................................................................................... 404
Table 5.7. Cabeço da Arruda, Tagus valley: chronology and stable isotope values (δ13C and δ15N)
of human bone collagen. .......................................................................................................... 409
Table 5.8. Cabeço da Amoreira, Tagus valley: stable isotope values (δ13C and δ15N) of human bone
collagen without associated 14C date. ....................................................................................... 412
Table 5.9. Cabeço da Amoreira, Tagus valley: chronology and stable isotope values (δ13C and δ15N)
of human bone collagen. .......................................................................................................... 413
Table 5.10. Cabeço das Amoreiras, Sado valley: chronology and stable isotope values (δ13C and δ15N)
of human bone collagen. .......................................................................................................... 424
Table 5.11. Vale de Romeiras, Sado valley: chronology and stable isotope values (δ13C and δ15N)
of human bone collagen. .......................................................................................................... 426
Table 5.12. Cabeço do Pez, Sado valley: stable isotopes values (δ13C and δ15N) without associated
14
C date................................................................................................................................. 428
Table 5.13. Cabeço do Pez, Sado valley: chronology and stable isotope values (δ13C and δ15N) of
human bone collagen............................................................................................................... 429
Table 5.14. Poças de S. Bento, Sado valley: stable isotopes values (δ13C and δ15N) without
associated 14C date. ................................................................................................................ 433
Table 5.15. Average δ13C values for the human bone collagen samples from individuals buried in
the Tagus and Sado valleys..................................................................................................... 441
Table 5.16. Average δ15N values for the human bone collagen samples from individuals buried in
the Tagus and Sado valleys..................................................................................................... 441
Table 6.1. Factors that can influence the development of rigor mortis.................................................. 481
Table 6.2. Estimated walking speed for a small group carrying a human corpse ................................. 483
xxxiv
Introduction to the dissertation
Death is a physical and a metaphysical phenomenon with a material representation.
Because human attitudes towards death are never random, the intrinsic materiality
of death is especially interesting for archaeology. Death’s materiality offers a window to past attitudes to the biological death of the individual and to the cultural
aspects of the human engagement with death.
Archaeological evidence shows that the origins of mortuary burial practice go
back to at least the Middle Palaeolithic, but until the Late Mesolithic the burial of
the cadaver seems to be an exceptional occurrence. Likewise, the earliest places of
frequent burial practice go back to the Middle and Upper Palaeolithic, becoming
more common in postglacial scenarios in Europe and in the Near East.
Death in the Mesolithic sites of the Tagus and Sado valleys,
Portugal
The Mesolithic shell middens of the Tagus and Sado valleys are remarkable because of the large number of human bodies buried in these sites, and constitute the
archaeological material investigated in this dissertation. These sites aggregate some
of the largest and earliest burial grounds known, arranged and maintained by
populations with an exclusive hunting, fishing, and foraging lifestyle. Here, in the
south-western Atlantic coast of Europe, the rise of sea levels during the Atlantic
climatic optimum (c 6400–4350 cal BCE) resulted in the formation of large estuaries. The typical scattered settlement pattern known for the Pre-boreal, Boreal, and
very early Holocene (c 9530–6400 cal BCE) became clustered by these new ecosystems in interior estuarine regions (Araújo 2003). Today, far from sight and
influence of marine waters, these Late Mesolithic shell middens can be very large
archaeological sites, many of them with well preserved human remains. This new
form of settlement was accompanied by a different relation to death. At the Tagus
and Sado valleys, burial practice was a common way of disposal of the dead. New
born babies, children, young and old adults, men and women, were carefully buried
in these places. From the 13 middens identified in the Tagus valley, nine have
human burials with a total of at least 263 individuals (Bicho et al. 2013; Cunha and
Cardoso 2001, 2002–03; Jackes and Meiklejohn 2008; Meiklejohn et al. 2009a; Paço
1938; Roksandic 2006; Santos et al. 1990). Similarly, in the Sado valley, there are 11
known middens, six of which with human burials to a total of c 113 individuals
35
Introduction
(Cunha and Umbelino 1995–97; Diniz et al. 2014). This particular engagement
with death contrasts with earlier and contemporaneous sites in the region; burial as
a mortuary practice was known in the Iberian Peninsula, however, it was not
common practice (Peyroteo Stjerna, forthcoming).
Shell midden sites with burial grounds have been interpreted in the light of economic and environmental processes as territorial claims to establish control over
limited resources (Arnaud 1989, 621). In highly productive estuarine environments
such as these, this is a compelling argument. However, subsistence alone does not
explain the significance of the frequent disposal of the dead in neighbouring burial
grounds, and how these sites were meaningful and socially recognized. Furthermore, the classic functionalist approach tends to reduce hunter-gatherers to highly
adaptive humans with a functional responsive behaviour, and ignore that they were
also agents of history with an historical consciousness and an intrinsic symbolic
expression. Human burials in archaeological contexts are an opportunity and a
challenge for researchers to investigate the non-utilitarian spheres that structured the
hunter-gatherer lifestyle.
Aims of the dissertation
Research questions and theoretical framework
The aim of this dissertation is to understand the relevance of the formation and
maintenance of burial grounds, and the significance of mortuary practices at the
sites of the Tagus and Sado valleys. In this study, I examine the burial activity of
the last hunter-gatherers of the south-western Iberian Peninsula from an explicitly
archaeological perspective, and intend to explain the burial phenomenon through
the lens of historical and humanist approaches to death and hunter-gatherers, on
the basis of theoretical concepts of social memory, place, mortuary ritual practice, and
historical processes. Theoretical framework formulated to address the research
questions stems from the need to bridge archaeological evidence obtained from a
set of archaeological methods strongly rooted in the natural sciences, with the
actions of people in their historical context.
Methodological choices
The human burials excavated from the nineteenth century to the 1960s at the shell
middens of the Tagus and the Sado valleys are at the centre of this research and
constitute the archaeological material of this study. Understanding the significance
of the burial activity at these sites requires we define how consistent and deliberate
these practices were, who was buried in these places, as well as when and how. The
examination of the burial activity through the analysis of the human remains in
these archaeological contexts relies on the application of three methodologies:
36
Introduction
radiocarbon dating and Bayesian analysis, stable isotope analysis of carbon and nitrogen, and
archaeothanatology. Following the aims of this dissertation, these methods are used to
investigate human burials in terms of time and practice, which can be outlined as
follows:
• to examine the chronological framework of the burial activity at each
site and each valley;
• to characterize the dietary patterns of the burial populations in an
attempt to define more closely who was buried in these sites;
• to reconstruct and define the central practices and gestures involved in
the treatment of the dead.
The outcome
Evidence obtained from archaeothanatological and isotopic analysis within the
chronological framework of the burial activity of each site, demonstrates that the
mortuary programme at the Tagus and Sado valleys was governed by a core set of
common practices that defined the proper burial. Despite the common mortuary
programme, the people buried in each valley were part of regional bands with
contrasting dietary patterns, suggesting that different groups were buried in each
valley and in each burial ground. More significantly, these results challenge former
interpretations which explained the Late Mesolithic settlement pattern in these
regions in the framework of seasonal and logistic models.
This dissertation provides new perspectives on the role and relevance of the
shell middens of the Tagus and Sado valleys. People frequenting these sites did not
merely adapt to a new environment. Hunter-gatherer communities were structured
within specific social scenarios in a historically situated world view. People were
bound by shared social and historical practices, which included the formation and
maintenance of burial grounds, as a primary means of history making. In this
study, I wish to suggest that mortuary ritual practice had a central role in the
formation and maintenance of these places, because this social practice is an act of
cultural production and a powerful memory aid. Furthermore, I propose that the
frequent burial of the dead in these burial grounds was an act of historical
consciousness motivated by the encounter and interaction with other bands frequenting the same environments. With this approach, I wish to stress the active
roles of memory and historical production, performed by mortuary ritual practice,
which were central in the formation and reconstruction of these places. This aspect
of social memory, the memory transmitted through bodily practices, has been
widely neglected, particularly in hunter-gatherer studies. In this view, and in the
scope of hunter-gatherer studies, the human burials are material representations of
the historical practices of the last hunter-gatherers of the south-western Iberian
Peninsula, and were recorded in these shell middens as depositional narratives of
an historical process.
The topographical placement of the dead within the landscape of the living was
a central element in the social dynamics of the last hunter-gatherers of the Tagus
37
Introduction
and Sado valleys. The formation of permanent burial grounds has been associated
with an increasingly sedentary lifestyle. However, the more or less permanent use
of these sites remains an open question, and it may have varied over time and
between regions. While the funerary function of the places is undeniable, the
residential nature of the sites remains unclear. As discussed throughout this
dissertation, dietary patterns of protein consumption identified in the different
burial grounds reflect a strong connection of the burial population to the ecology
of each site. However, while this may indicate that these sites were used during
long periods, it does not exclude the possibility of mobility, exploration, and use of
other territories.
Among other possible features and interpretations, these shell middens were
mutually a place for the dead, and a place of ritualized codes. Death rituals played a
central role in the life of these hunter-gatherers in developing a sense of community within the landscape, as well as maintaining social ties in both life and death.
Outline of the dissertation
The dissertation is structured in three parts. In part one, Death, I introduce and
explore the theoretical framework of this research by addressing central issues of
the human engagement with death framed as acts of cultural practice and historical
consciousness. In this part, I outline a short history of death in deep time with a
focus on burial as a mortuary practice. This is followed by a brief examination on
the core relationship between the discipline of archaeology and the spaces
dedicated to the dead. I conclude this part with considerations on the role of
archaeology in death studies and reflect upon the significance of studying death.
In part two, Bodies of Knowledge, I present the source material and archaeological
evidence of this dissertation, from chapters 2 to 5. Chapter 2, Archaeological Material,
presents a general description of the archaeological sites, the material selection, and
an examination of the historical contexts of data production. Chapter 3, Methods,
outlines the three methodological approaches applied in this study which were
explicitly targeted towards the study of mortuary practices: radiocarbon dating and
Bayesian analysis, stable isotopic analysis of carbon and nitrogen, and archaeothanatology. In
this chapter, I present the central principles of each method and their application
to this particular data set. Chapter 4, Results, presents the data obtained from the
radiocarbon and Bayesian analysis, stable isotopic analysis of carbon and nitrogen,
and archaeothanatological analysis. This chapter is organized by archaeological site.
In the last section of Bodies of Knowledge, Chapter 5, Analysis and Synthesis of the
Results, I present a synthesis of the results and examine the intersections of the
empirical evidence from each archaeological site in the context of each valley and
inter-valley.
In the last part of this dissertation, part three, Death in Place, I discuss the archaeological evidence from the perspective of mortuary behaviour on the basis of
the concepts of social memory, place, mortuary ritual practice, and historical
38
Introduction
processes. Chapter 6, The Places of the Dead as Places of Shared Narratives: Discussion,
evaluates the boundaries of the mortuary phenomenon under study, in terms of
practice, space, and people. I discuss the significance of the observed mortuary
ritual practices in the formation and maintenance of the shell middens at the Tagus
and Sado valleys. The volume is completed with the last section of this chapter,
Conclusion: on the history of death and the last hunter-gatherers of the south-western Iberian
Peninsula, where I address the central arguments in this dissertation and emphasize
the relevance of further research in the field of archaeology of death so that we can
better understand the social and historical dynamics of the last hunter-gatherers of
the south-western Iberian Peninsula.
The appendix provides supplementary data and it is available online-only. Tables
in the appendix are indicated by the letter A followed by a sequential number (e.g.
Table A.1).
39
PART ONE: DEATH
Döden: Därför spelar du schack med Döden?
Riddaren: Han är en svår och skicklig taktiker men än så länge har jag inte givit upp
en enda pjäs.
(Bergman 1957)
Introduction to part one
Death is a complicated field of research because it affects all of us. This is also why
the field is so intriguing and captivating. Although each individual encounters the
experience of death in different ways, the way we deal with the death of others and
with our own personal awareness of death, has a major influence on how we, as
researchers, think about death.
The opening part of this dissertation is dedicated to the subject of death as a
central academic field of research, and is written from an archaeological point of
view. In the first section, The human engagement with death, I outline a brief history of
death in deep time. The burial practices of the last hunter-gatherers of the southwestern Iberian Peninsula are the object of this dissertation and this version of the
history of death is biased towards these elements: burial and (Mesolithic) huntergatherers. This section is followed by Archaeological thought and death, where I discuss
the importance of the places of the dead in the development of the discipline of
archaeology. I then turn to the theoretical framework of this dissertation to address
central issues of the human engagement with death framed as acts of cultural
practice and historical consciousness – Deathways of hunter-gatherers and historical
processes. Having death as a point of departure, I explore the role of mortuary ritual
practices as powerful memory aids in the formation, maintenance, and reconstruction of places for the dead, by stitching together the spheres of death, memory,
and place, with the thread of the historical process – Memory in the landscape: rituals in
practice. In the last section, Death in deep time: why does it matter?, I argue for the
potential of death studies to investigate changes of world view in a given society,
and why knowing about how past societies dealt with death, grief, bereavement
and commemoration, actually matters, highlighting why archaeology is in a
privileged position to contribute to an interdisciplinary dialogue in the field of
death studies.
43
Chapter 1
Death, Hunter-Gatherers, and the Historical
Process
The human engagement with death
The human experience of death is unique in the animal kingdom and it reflects in a
unique way the complexity of the human species. The paradox of the human
engagement with death lies in its universality on the one hand, and its heterogeneity on the other. The universal concept of death, as the inevitable fate of all living
beings, has paradoxically, a heterogeneity of meanings that only an emotional and
creative complexity can explain. The invention of death (see Taylor 2003) rests on this
antagonism, and it is one of the most particular aspects of the human condition.
Awareness of death is possibly the most powerful challenge to human significance
and destiny, and a cornerstone in the development of humanity (Davies 2005). For
social and intelligent beings death is an emotional experience, both in terms of
bereavement and consciousness. Despite symbolic and religious beliefs, grief was
certainly powerful in the close-knit bands of past hunter-gatherers (Davies 2005;
Tarlow 1999). Social experience is centred on emotions and materials (Gamble
2013; Gamble et al. 2011), and ritual responses to death act as bodily and material
ways to cope with the crisis of death. Ritual responses to death have an integrating
effect. Mortuary ritual practices make the link between death and life, while allowing a perspective upon life and the future (Davies 2005). Mortuary rituals act on
the level of individual relationships but more importantly, they act on the cosmological level, between humans and the world around them (Nilsson Stutz 2014).
For this reason, D. Davies (2005, 90) has argued in his book A brief history of death
that “humanity became stronger because of death-awareness”.
The debate about the origins and human concept of death, as well as its meaning in the context of human evolution, is ongoing and multivocal. The development of language, carrying with it the communication and memory of the inevitability of death, along with cognitive developments associated with symbolic
thought and the earliest ritual practices, are probably the origin of the development
of a necessity/capacity for intentional and reflective funerary ritual practice. The
core of this debate lies in the timing rather than the factors: when, in human evolution, did humans develop this consciousness of death? With modern humans only
(c 35,000 years ago, in Europe), or is it an earlier development? For the immense
time span between the Middle and Upper Palaeolithic, only c 150 possible funerary
44
Death, Hunter-Gatherers, and the Historical Process
deposits have been identified (Pettitt 2011; Taylor 2003, 31), and in many of these
contexts it is difficult to confirm the intentionality of the practices as funerary behaviour (see Zilhão and Trinkaus 2002, 519). Another set of questions in the scope
of the debate about the origins of funerary behaviour relates to the significance
and impact, in human evolution, of clear manifestations of intentional funerary
ritual practices. These questions are important not only from the perspective of the
development of ritual and symbolic behaviour associated with a consciousness of
death, but also in the cognitive context of the formation of the self, society and
cosmologies (Parker Pearson 2005, 146). In any field of research, questions related
to theme of origins are always complex and controversial, and the debate on the
origins of funerary behaviour is no exception (see Pettitt 2011).
Origins of human burial
In this brief overview I focus on the origins of human burial and on the formation
of the first cemeteries in Europe. Following the aims of this dissertation, this review is focused on the act of human burial as a form of funerary practice, and for
this reason some preliminary observations should be made.
The first observation concerns the different contexts where human remains can
be found. The presence of human remains in a site is not necessarily correlated
with funerary practice. Human remains can be found in a variety of cultural contexts. Human remains can be worn as ornaments, as well as amulets and relics; can
be involved in an array of ancestral and religious rituals, as well as in demonstrations of power; can be found in scenarios of violence and conflict, as well as in
contexts of accidental death; the list goes on. The funerary context is just one of
the many contexts where human remains can be present, and the intentionality of
the actions should be demonstrated (Duday 2009).
A second preliminary observation is that burial is just one of many possible cultural options in the complex diversity of funerary practices. Also, the act of burial
is frequently just one part of the various episodes involved in the funerary process.
Burial is, however, a practice that may be well preserved in the archaeological record (see Lyman 1994), hence the traditional interpretative focus on this practice,
often ignoring other variables. As noted above, regarding the context of occurrence of human remains, a buried corpse is not necessarily correlated to a funerary
practice either. Also, it should be emphasized that the development of a consciousness of death is not synonymous with marking its occurrence. Following
these preliminary observations, a cautious and evidence-based approach is essential
when interpreting the presence of human remains in archaeological contexts.
According to the available evidence, the earliest intentional burials known date
from c 120,000 to 90,000 years ago, with evidence from Anatomically Modern
Humans (AMH) and possibly Neanderthals (Pettitt 2011). These early burials are
interpreted as the earliest evidence of deliberate creation of a space in which to
deposit and cover a corpse (Pettitt 2011, 57). The remains of ten archaic AMH at
the Mugharet-es-Skhūl cave, Mount Carmel, Israel, are the earliest burials known,
45
Chapter 1
dated to between 130,000 and 100,000 years ago. Archaeological evidence indicates
that at least four individuals were placed in artificially excavated grave cuttings,
attesting to the intentionality of the burial act (Belfer-Cohen and Hovers 1992;
Pettitt 2011, 59–62). Djebel Qafzeh cave, Israel, contains the remains of 13 archaic
AMH dated between 100,000 and 90,000 years ago, presenting what seems to be
evidence for organization of the space with a dedicated area for the dead (BarYosef and Vandermeersch 1993; Pettitt 2011, 63, 68–72). The reasons why these
individuals were buried at Skhūl and Qafzeh are unknown and the nature of the
ritual practices behind such depositions remain a question of debate. While some
authors argue in favour of a degree of symbolic practices involved in the burials
(Hovers et al. 2003; Pettitt 2011), other researchers question the ritual dimension
of these early burials based on their simplicity and variability (Shea 2001). Evidence
from possibly the earliest Neanderthal burials known comes from Tabun cave, Israel. The dating of the Tabun sequence is hotly debated (Bar-Yosef 2000; Millard
and Pike 1999), but it seems probable that at least one of the three individuals
found in the cave was buried c 120,000 years ago. While there is strong evidence
for Neanderthal burials c 30,000 years ago, burial practices were rare and the archaeological data seems to indicate a more varied funerary activity among
Neanderthals, even when compared with contemporary AMH populations (Pettitt
2011, 79).
The earliest burials of AMH in Europe emerge in the Mid Upper Palaeolithic,
c 29,000–21,000 BP, and present clear evidence of mortuary ritual practices (Pettitt
2011, 168). The earliest burial known in the Iberian Peninsula is dated from this
period, c 25,000–24,000 BP (Pettitt et al. 2002), and was found at Lagar Velho in
the Lapedo valley, Portugal. This is an individual burial of a 4–5 year old child,
buried in a shallow pit in a rock shelter, and was accompanied by a complex set of
grave goods (Duarte et al. 1999; Zilhão and Trinkaus 2002). Despite the relatively
complex burials known for this period throughout Europe, the preferred relationship with the dead is mobile. The bodies of the dead were fragmented and moved
across the landscape. This mobility was possible through complex interaction with
the cadaver and processing of skeletal elements into personal ornaments, amulets
and relics. The archaeological evidence for the symbolic fragmentation of the human body is particularly strong for the European Mid and Late Upper Palaeolithic,
also represented in parietal and portable art (Pettitt 2011), remaining a common
practice throughout the Mesolithic (Gray Jones 2011).
The deliberate burial of dead (i.e., excavation of a grave; placement of the cadaver in the feature; covering of the cadaver) is a practice only known for two
hominids: Neanderthals and AMH. However, throughout human history, burial as
a funerary practice was always rare and certainly not the funerary norm (Pettitt
2011). Overall, burial evidence clusters into certain periods in particular regions,
suggesting that burial practices were not transmitted continuously throughout the
Middle and Upper Palaeolithic. Burial practice becomes more frequent in the later
stages of the Upper Palaeolithic, during the Late Pleistocene, between 15,000 and
12,000 BP, in several regions of Europe and the Near East. During this time,
46
Death, Hunter-Gatherers, and the Historical Process
several places were clearly reserved for the formal burial of the dead suggesting
incipient forms of cemeteries. These practices, first identified in the Late Pleistocene, during the final stages of the Upper Palaeolithic and the Early Mesolithic,
continue into the Holocene (see Meiklejohn et al. 2009b). Human burial in dedicated spaces for the dead then becomes more frequent, suggesting that burial
practice was charged with a new meaning, possibly in the context of new cosmologies and world views (Nilsson Stutz 2014; Pettitt 2011, 249). Burial was no longer
an odd phenomenon dedicated largely to single individuals placed in secluded
caves and rock shelters.
The formation of cemeteries is a phenomenon in the landscape (Pettitt 2011,
249) and presents a new relationship with death. The earliest examples of this practice can be traced in the Late Pleistocene (c 15,000–11,000 BP), but largely in
caves. A new topographical relationship with death, with the formation of
dedicated spaces for the dead in open air sites, is clearly an emergent phenomenon
in the context of postglacial societies, with well known examples in the European
Mesolithic. Although as we shall see later in this dissertation, burial practice in
dedicated sites remains exceptional in the Iberian Peninsula, until the Late Mesolithic cemeteries at the Tagus and Sado valleys, c 8000 years ago (see Peyroteo
Stjerna, forthcoming). The adoption of one or another funerary practice varies
regionally and according to specific social dynamics.
The variability and discontinuity in mortuary practices through time and space
suggests that burial practice is not the end point of a development of a funerary
practice, nor the end point of an internal social development of hunter-gatherers.
Modern western society retains the idea that, “the ways of physically removing the
dead from the living have changed very little during the course of human history”
(Davies 2005, 49), based on the misleading idea that earth-burial and cremation
have always been the two basic forms of funerary practice. However, burial practice is one of many possible cultural options for the treatment of the dead, but
many of these other practices do not leave traces in the archaeological record.
Also, burial practice is not necessarily more complex than any other cultural option
for the treatment of the dead, to which can be added that absence of burial is not
equivalent to absence of mortuary practice either. From this perspective, mortuary
practices cannot be approached in a progressive evolutionary way: burial is not a
step in the directional development of hunter-gatherers.
Material nature of death
One of the main difficulties when analysing archaeological data is presented by the
diversity of cultural responses, which is a central characteristic of human
behaviour. Attitudes towards death constitute a good example of this great diversity. Is there a common denominator in cultural responses to death? One of the
fundamental aspects of being human is the material form in which ideas are expressed and understood. The diversity of cultural responses to death has in common the materiality of death (Fahlander and Oestigaard 2008).
47
Chapter 1
Death is a material event in both biological and cultural dimensions (Nilsson
Stutz 2003a, 2008; Robb 2013). The death of the individual is a biological phenomenon with a corporeal representation: the deceased is a material entity. The
human engagement with death, as an emotional and metaphysical phenomenon
has also a material representation: by the manipulation of space, of objects, and the
human body itself. Herein lies the advantage of the discipline of archaeology in the
field of death studies: the material nature of death (fig. 1.1).
Figure 1.1. The archaeological study of death through its material remains.
Archaeological thought and death
In popular culture, archaeology and the spaces of death are practically intertwined.
We are all familiar with two of the most popular fictional archaeologists: Dr. Henry
Walton Indiana Jones Jr. and Lara Croft from the video game Tomb Raider. Both
characters are known for their adventures in ancient tombs, chasing mysterious
artefacts of ancient funerary practices, while fighting the troops of doom. Despite
the fact that neither of these fictional characters pursues archaeology in the modern sense, there is a hint of truth behind this representation of the discipline.
In fact, the development of archaeology as a scientific subject is closely related to
mortuary spaces. This is because, the spaces reserved for the dead are frequently
preserved and maintained over generations, as places of commemoration and
memory. Often, these can also assume monumental proportions and are frequently
easy to identify because of their visibility in the landscape; typically by contrast
48
Death, Hunter-Gatherers, and the Historical Process
with settlement sites built with perishable materials which are easy to renovate but
decay quickly, becoming more subtle in the landscape, and in the archaeological
record.
In this section, I present a brief overview of the development of archaeological
thought in terms of the importance of the places of the dead in the development
of the discipline. In the last part of this section Historical and humanist approaches to
death and hunter-gatherers, I outline the theoretical framework of this dissertation
formulated to address the particular questions of this research.
Chronology and typology
It was in the late eighteenth century that the first systematic excavations took
place, with the recovery of a great quantity of artefacts that constituted the assemblages of the first museums (Chapman and Randborg 1981, 2–3). It is in this context that in the mid nineteenth century, C. J. Thomsen and his collaborator J.
Worsaae developed the method of relative chronology (Trigger 2007, 124) beginning with a chronological approach to the analysis of funerary contexts, named The
Worsaae Law. According to this principle, the objects that accompanied the funerary deposition would be in use in most cases, during the same time period. This
principle was at the root of all chronological associations in funerary context
(Rowe 1962, 129), and constitutes the basis of the first typological series, being the
first contribution of mortuary archaeology to the study of past human societies.
The cultural paradigm, normative and diffusionist
In the first half of the twentieth century, in continuity with the chronological and
typological approaches of the previous century, the great interest in funerary contexts stemmed from the assumption of its potential as cultural indicator (Childe
1945). The comparative analysis of the great quantity of artefacts, recovered mostly
from funerary contexts, motivated the recognition and definition of cultural areas
in space and time, with the development of typological series always becoming
more refined. The normative concept of culture, introduced by the new theoretical
framework developed by G. Childe (1929) guided research during this period.
Changes in material culture were interpreted from a diffusionist perspective, replacing previous unilinear and evolutionist theories. In the context of the normative and diffusionist paradigm, the appearance, diffusion and decline of cultural
groups could be identified through behaviour patterns (Childe 1956, 9–10).
Likewise, this approach explained funerary rituals within the same normative and
diffusionist framework (Chapman and Radsborg 1981, 4). Even though the definition of culture involves a great variety of aspects, as recognized by G. Childe
(1956, 122), many of these so-called cultural groups, such as the European
bell beaker group, were defined from data recovered and known from funerary contexts only. Despite substantial criticism of these assumptions, the cultural and
normative interpretation of the funerary ritual as an ethnic marker remained
49
Chapter 1
popular in the archaeological narrative throughout the twentieth century (Chapman
and Randsborg 1981, 4).
The “New Archaeology” and the social approach to death
Archaeological data retrieved from funerary contexts during the first half of the
twentieth century was used as evidence to support arguments of the cultural paradigm, but the data was never interpreted according to the perspective of the
funerary practices per se. The rare interpretative proposals based their observations
on ethnographic data known from anthropological studies, which was applied to
archaeology in a more or less speculative way. In 1969, P. Ucko called attention to
this problem with the article Ethnography and archaeological interpretation of funerary
remains, also known in the literature as Ucko’s cautionary tale (Chapman and Randsborg 1981; Nilsson Stutz 2003a, 109; Parker Pearson 1999, 21; Tarlow 1999, 11).
This cautionary text, written from the critical analysis of ethnographic cases,
resulted from reflection on the relationship between the society of the living and
the material remains of their funerary practices. Ucko criticized diffusion theories
and demonstrated that, even within societies with strict and formal funerary
practices, there will be variation in the ritual treatment of the dead, as well as in the
different individual and collective interpretations. One society or culture has not
one, but various ways to treat their dead. He argued that this variability, which the
archaeological method should be able to discuss, may in fact indicate differences in
social status rather than being the result of external influence by ways of cultural
diffusion (Ucko 1969, 270). Ucko emphasized, although without being specific,
that social status may have been of various types, and called attention to the fact
that absence of grave goods was not necessarily indicative of low status or poverty
of the individual or social group. Likewise, the similarities found in disposal of the
dead may not have directly correlated to contact between people from different
groups (Ucko 1969, 275). Ucko did not reject ethnographic parallels but their incautious use, and defended the use of knowledge obtained through ethnography by
archaeologists in order to expand their interpretative horizons when reading the
archaeological material (Ucko 1969, 262, 264). This text is considered emblematic
of the positivist fervour in the archaeology of funerary practices during the 1970s,
but the optimism and cautions expressed by Ucko, gradually faded (Nilsson Stutz
2003a, 110). As criticized by Chapman and Randsborg (1981, 8), many researchers
have ignored this text. However, its most important premises, especially those
aligned with the historical-culturalist critique and with the demand to integrate a
social approach to death studies, found continuity through the work of Saxe (1970)
and Binford (1971).
The archaeological spaces of the dead were once again at the centre of archaeological debate and constituted the central argument in Binford’s critique of historical-culturalist perspectives (Chapman and Randsborg 1981, 6). From the 1960s
onwards, with the expansion of archaeology as a discipline, some of the earlier
pessimism was mitigated, about the potential of archaeology to access certain
50
Death, Hunter-Gatherers, and the Historical Process
spheres of human action, besides the traditional economic and technological
questions. The new inquiry proposed for the archaeological analysis of funerary
contexts had a significant role in the emergence of the New Archaeology, and
determined the bases for its theoretical development and methodological application. The body of work developed by L. Binford (1962, 1965, 1968) was in this
context fundamental. Binford denied any limitation on accessing the past and
introduced the conviction that questions of social organization, already explored by
anthropology in the early twentieth century by Durkheim and his followers, could
potentially be addressed through archaeological analysis. Binford’s argument was
based on a systemic approach to the analysis of the internal dynamics of cultural
systems, through which the relations between material culture and other elements
of the cultural system could be identified (Chapman and Randsborg 1981, 6). The
analysis of funerary contexts played a central role in the demonstration of this
argument, culminating with the publication of the highly cited thesis of A. Saxe
(1970), and a compilation of seminal texts edited by J. Brown (1971) with
contributions from Binford, Saxe, and Brown. These works had the merit of finally
introducing the social dimension of funerary practices to the context of the
archaeological research.
Contrary to the irregularities proper of the material recovered in settlement
areas, the funerary spaces were considered representative of an intentional
behaviour, enclosing a special interest for the approach of the New Archaeology.
In this perspective, the analysis of the material remains in the funerary context is
considered particularly informative in terms of organization of the society. Following this premise, Binford (1971) considered that the social structure of a given
society would be mirrored in the funerary practices of the group. In this view, the
social organization would determine the patterns and variability of the funerary
practices. For the New Archaeology, the treatment of the dead was correlated with
the nature of the social system of the group, rather than to cultural specificities,
such as the identification of certain funerary rituals with particular ethnic groups,
as claimed by the historical-culturalist archaeologists. The critique of Binford
(1971) was aimed at shifting the focus from ethnic differences, and to concentrate
on a new scale of analysis focused on the differences within social groups. In this
approach, funerary rituals should be interpreted as evidence of social variability,
status differentiation, and social complexity (Binford 1971, 4). This new framework
was based on a central theoretical element borrowed from social anthropology:
theory of roles and the concept of social persona which Binford (1971) adapted and
applied to archaeological questions, through a new methodology based on quantitative analysis and studies of compared ethnography (Trigger 2007, 399, 405).
Binford recognized five dimensions in which the social persona may be represented
through funerary practices: age, sex, hierarchy or social ranking, social affiliation
and circumstances of death (Binford 1971, 17). In this model, human groups are
divided according to their subsistence practices – hunter-gatherers, farmers, and
shepherds – in an increasing state of social complexity, which would be reflected in
the treatment of the dead. In this view, the more complex the society, the more
51
Chapter 1
dimensions would be represented in funerary contexts, suggesting that social
complexity is proportional to the complexity of the group’s funerary practices.
Funerary practices were then analysed through the lens of social complexity and
explanations of social diffusion were rejected. In this perspective, the variability of
funerary ritual was a direct consequence of the social structure of the population
under study (Trigger 2007, 405), and the treatment of the dead was assumed to be
related to the nature of the social system rather than to its cultural specificities. In
this logic, the more complex a society is the more complex its funerary practices
will be.
Despite the relevance of funerary contexts in the development of archaeology
as a discipline, it was only in the 1970s (Brown 1971) that the discipline developed
a theoretical framework able to deal with and interpret mortuary data (Chapman
2003, 305; Chapman and Randborg 1981, 2). It was under the umbrella of the New
Archaeology that the archaeological approach to material remains of past funerary
practices was coined as The Archaeology of Death (Chapman et al. 1981).
Unfortunately, the investigation of the spaces for the dead was the conduit for the
analysis of social structure of the living, rather than the study of human engagement with death.
Postprocessual reaction: rituals in action
Theoretical discussion initiated by the New Archaeology, explicitly advanced by
Binford’s (1977) call for theory building was a turning point in the history of archaeological thought (Trigger 2007). As in any discipline, this paradigm shift generated
much controversy and heated debates. A few years later, the postprocessual
reaction to the premises of the New Archaeology strongly criticized the positivist,
functionalist and evolutionary approach to human societies; emphasizing symbolic,
structural, and critical approaches to the study of archaeological data (Hodder
1982; Trigger 2007). The postprocessual reaction introduced archaeology to a
broad range of theoretical positions, particularly those in the historical and social
sciences (Hodder 2012) that the New Archaeology had ignored. Its impact in the
field of death studies derived particularly from the recognition that ritual practices
are a fundamental aspect of human societies (Hodder 1980). This was a major shift
of focus, because the ritual perspective of mortuary practices had been naively
neglected in the processual approach to the archaeology of death (Nilsson Stutz
2003a, 119). Early postprocessual works argued that variation in mortuary
treatment of the dead could not be interpreted as a reflection of everyday social life
mirroring certain dimensions of social identity, and emphasized the importance of
multiple lines of evidence in order to interpret mortuary ritual practices in their
social context (Pader 1982, quoted in Nilsson Stutz 2003a, 122). Despite the clear
shift in perspective, ritual was often approached as a form of ideology and
symbolic power (Shanks and Tilley 1982; Trigger 2007, 451), and status and power
remained central aspects in the postprocessual mortuary narratives (Parker Pearson
1982, 1984). Unfortunately the idiosyncrasies of the human engagement with death
52
Death, Hunter-Gatherers, and the Historical Process
remained largely ignored (Tarlow 1999, 12), despite the creative influx, broader
range of theoretical positions, and the open spectrum of the questions promoted
by postprocessual archaeologists. Nevertheless, postprocessual archaeologies
brought into studies of death theoretical propositions which remain highly
influential in today’s agenda, which in multiple ways informed the reading of the
archaeological data in this dissertation.
***
Throughout the history of archaeology, the material remains of death have always
been a primary source in the development of theoretical and interpretative
approaches. However, for the most part, the specificities of death-related contexts,
and of the human engagement with death as an event or a process, have been
largely overlooked (Robb 2013; Tarlow 1999). The description and analysis of
grave goods, and the spatial analysis of mortuary contexts remained central in
archaeological studies. Other studies, focusing on the human remains recovered in
mortuary contexts continued, providing relevant information about the life of the
individual, by assessment of age at death, sex, height and robustness, as well as
identification of diseases, pathologies or body modifications, along with recent
improvements on the characterization of past diets, and genomic data of ancient
populations. These are invaluable lines of evidence for archaeological research,
inside and outside the spectrum of death studies. However, in the scope of
funerary archaeology, the aim is to investigate funerary practices in all their dimensions, as well as to document, identify and contextualize cultural responses to
death, from the material remains of past populations. This emphasis is important,
because these are in fact independent questions, although related, but frequently
confused with the questions of physical and biological anthropology.
In recent years, the subject of death has gained new importance, certainly
encouraged by a broader spectrum of perspectives motivated by the postprocessual
reaction and a new maturity of the discipline in the twenty-first century, which
would hardly have happened without the critical approach to method and theory
promoted by the New Archaeology. Archaeology of death now problematizes
central aspects of the human engagement with death, having death at the centre of
the research and mortuary ritual practices as the point of departure (see Nilsson
Stutz and Tarlow 2013), exploring dimensions such as cognition and symbolic
thought, emotion and experience, the body, ritual and social relations; through the
lens of human attitudes towards death in the archaeological record (see e.g., Brandt
et al. 2015; Croucher 2012; Fahlander and Oestigaard 2008; Murphy 2008; Nilsson
Stutz 2003a; Pettitt 2011; Sofaer 2006; Tarlow 1999).
Historical and humanist approaches to death and hunter-gatherers
Theoretical framework formulated to address the research questions of this dissertation stems from the need to bridge archaeological evidence obtained from a set
of archaeological methods strongly rooted in the natural sciences, such as biology
53
Chapter 1
and chemistry, with the actions of people in their historical context. In this study,
I examine the burial activity of the last hunter-gatherers of the south-western
Iberian Peninsula from an explicitly archaeological perspective, and attempt to
explain the burial phenomenon through the lens of historical and humanist
approaches to death and hunter-gatherers, on the basis of theoretical concepts of
social memory, place, mortuary ritual practice, and historical processes.
Humans are historical creatures. People shape history through actions and
representations, and in this sense, “history is the process of cultural construction
through practice” (Pauketat 2001, 87). History is shaped by the relations between
individuals and between individuals and groups (Robb and Pauketat 2013, 21). The
way people acted, the way they represented themselves and embodied their
traditions, has shaped history (Pauketat 2001, 86). Practices are historical processes
which often involve a material representation (Pauketat 2001; Sassaman and Holly
2011). Monuments and cemeteries are among the most obvious material manifestations of historical practices. Other archaeological favourites, such as patterns of
settlement and exchange also express historical processes as they may reveal
movement of people and objects, migration, interaction, alliances, and so forth
(Sassaman and Holly 2011, 8).
Hunter-gatherers are makers of history and enjoy historical consciousness (Sassaman and Holly 2011), like any other human group. Hunter-gatherers live in a
world they understand, with a past and a future, and act culturally within this
historically situated world view (Ingold 2000; Sassaman and Holly 2011). The formation of burial grounds and shell middens by the last hunter-gatherers at the
Tagus and Sado valleys in Portugal, are acts of historical production. This
approach provides insights into the long-term history of the cultural production of
these places. It allows us to examine the significance of mortuary practices during
the formation, maintenance and development of these shell middens. To understand the mortuary practices of the last hunter-gatherers of the south-western
Iberian Peninsula through the archaeological material from the shell middens in
the Tagus and Sado valleys, I work in the framework of social memory, concepts
of place, mortuary practices, and historical agency. These theoretical principles
allow me to discuss the central questions of this dissertation, enquiring about the
relevance of the formation and maintenance of burial grounds, and the significance
of mortuary practices and formal places for the dead in the lives of these huntergatherers.
History was recorded in the shell middens at the Tagus and Sado valleys, not as
a literary enterprise, but through non-literate acts of history-making (Randall 2011;
Sassaman and Holly 2011). In this perspective, depositional narratives are viewed
as the primary means of history-making (Randall 2011), and burial practice as an
act of cultural production. From this point of view, depositional narratives
constitute the acts that give meaning to these places, rather than the shell mound
in its monumental and unitary form (see Kidder 2011; Pauketat and Alt 2003).
When burial grounds in shell middens are approached as historical processes we
can focus on the history of people in their context and move from functionalist
54
Death, Hunter-Gatherers, and the Historical Process
interpretations that regard these deposits as simply refuse heaps, of humans and
food debris.
Deathways of hunter-gatherers and historical processes
Deathways of hunter-gatherers are varied and may present a more or less complex
set of related activities. As discussed in an earlier section of this chapter, the social
interaction between the living and the dead has a long history. Mortuary practices
did not evolve in a linear fashion and present significant regional cultural variation
over time (Pettitt 2011). At the shell middens of the Tagus and Sado valleys, burial
was a common method of disposal of the dead. This particular way of interacting
with the dead is expressed through depositional narratives the material remains of
which are the object of this dissertation. These burials represent social practices
and the history of these hunter-gatherers unfolds from the way they engaged with
death. The history of death itself correlates with the history of human relationships
(Davies 2005, 28). The encounter with “others” often leads to changes in cultural
production, as profound as the creation of new cosmologies and new social relationships, in order to cope with diversity or resistance to change (Sassaman and
Holly 2011, 8). Cultural diversity among hunter-gatherers is a consequence of interaction rather than isolation (Sassaman 2011, 189). Following this, could the new
way of dealing with the dead that was recorded at the Tagus and Sado valleys,
characterized by the burial of the dead in dedicated burial grounds, be motivated
by encountering other bands?
Memory in the landscape: rituals in practice
Places with burial grounds reveal a commitment of the living to the dead, expressed by continued depositional events reproduced over a more or less long
period of time. These practices required bodily performances which were central in
processes of remembering. Social memory was created and renewed at these places
through mortuary ritual practices. The stress on memory emphasizes the role of
ritual practice in the construction of the landscape. This approach acknowledges
the archaeological record of burial practices as material traces of memory construction. In this perspective, mortuary rituals are viewed as central elements in the
construction of places. Following this, I would like to argue that it is through mortuary ritual practices that these sites were remembered and kept meaningful in the
landscape.
Memory, understood as shared memory of a social order, requires certain
mechanisms to be maintained and transmitted within the social group over time
and across generations. Memory can be inscribed in monuments and texts. In archaeological contexts, monuments are evident traces of memory and commonly
emphasized, but incorporated memory leaves archaeological traces as well (Preucel
and Meskell 2004, 216; van Dyke 2010, 279). Memory can be materialized in representations, objects, ritual behaviours, and places (van Dyke and Alcock 2003). Of
particular interest to this dissertation is the ritual behaviour which can be revealed
55
Chapter 1
through the observation of the remains of mortuary practices in the archaeological
context (Nilsson Stutz 2003a).
As Connerton (1989) demonstrates, embodied or incorporated memory, manifested through performative rituals and physical behaviour, is much more effective
in the production and transmission of social memory. In his work How Societies
Remember, Connerton (1989) shows through various case-studies that memory is
transmitted in non-textual and non-cognitive ways through ritual practice. This is
shown to be a powerful system to enhance memory because ritual practice is a
formalized, sequential, and repetitive performative language that implies continuity
with the past (Connerton 1989, 45). Thus, Connerton (1989) argues, incorporating
practices through embodied behaviour provides a particularly effective system of
mnemonics. Nevertheless, inscribed memory is commonly considered the privileged form for the transmission of the memories of a given society. In archaeology,
this aspect of social memory, the memory transmitted through bodily practices has
been widely neglected, particularly in hunter-gatherer studies, as authors such as
Sassaman and Holly (2011) have pointed out. However, it has been shown that
systems of inscription, even the more complex, do not correlate to the capacity of
a society to remember (Connerton 1989, 102). Thus, following Connerton’s (1989)
premises, mortuary ritual practices are determinant in the process of memorialization of these places.
Death in deep time: why does it matter?
Death paradigms
The history of death is entangled with the history of changing social values, as
shown by anthropological and sociological research, as well as by philosophical and
theological approaches to death studies. In this view, the shift in the meaning of
death, i.e., death’s paradigm shift will be consistent with the changes of world view
of a given society (Davies 2005).
The human engagement with death is complex and varied and can be expressed
in multiple ways. Regardless of beliefs and symbolic associations to death, dead
bodies require a physical action from the living, which largely relates to a world
view (Davies 2005; Nilsson Stutz 2014). Mortuary ritual practices are defined
within a cosmology in the way people see the world and express their emotions
and materiality (Nilsson Stutz 2014; Tarlow 1999). The history of death and the
history of human relationships are closely related, as well as the place of death and
human identity (Davies 2005, 38). Changes of world view present significant consequences in the cultural understanding of death and have a strong impact upon
mortuary ritual practices (Davies 2005). This materiality of change, in world views
and attitudes to death, is part of the archaeological record and their interpretation
56
Death, Hunter-Gatherers, and the Historical Process
provides relevant assets for the understanding of past human relations with death
and with each other.
Death studies and archaeology: an archaeological manifesto
Death studies and archaeology are fully interconnected. Death studies deal with the
complex nature of death within its biological, emotional, metaphysical, and
material dimensions. This complexity requires multidimensional approaches and
challenges researchers from diverse fields. Archaeology can provide a solid bridge
to this interdisciplinary dialogue. Archaeology has access to human behaviour over
long periods of time and in vast geographic areas. Archaeology brings temporal
depth to the great variety of human agency with an exceptional perspective on
what makes us human. This database of knowledge – on what makes us human –
has no parallel, and it is of great interest to us as archaeologists, to work with other
disciplines that deal with death, ritual and society. It is only within an interdisciplinary framework that we can reach a meaningful and holistic approach within the
humanities.
The richest information about past mortuary practices is of archaeological nature (Nilsson Stutz 2014, 714), and it is our responsibility as archaeologists dealing
with ancient death ways to provide a robust and theoretically informed explanation
of death in deep time. This knowledge about past human engagement with death,
grief, bereavement and commemoration, matters in terms of human evolution, in
terms of understanding our place in the world, in terms of understanding other
world views, and ultimately, it matters in terms of understanding what makes us
human.
57
PART TWO: BODIES OF KNOWLEDGE
Jöns: Varför håller du på med såna kludderier?
Målaren: Jag tänkte man skulle påminna folk om det faktum att de ska dö.
Jöns: Det blir de inte gladare av.
Målaren: Varför i helvete ska man alltid göra folk glada? Man kan ju skrämma dem
lite ibland.
Jöns: Då blundar de och tittar inte på din tavla.
Målaren: Du kan tro de tittar. En dödskalle är mycket intressantare än ett naket fruntimmer.
(Bergman 1957)
Introduction to part two
The biological remains of a human body found in archaeological context are invaluable elements of research, and no other find brings us so emotionally close to
the past. In archaeological sites, when preservation allows, the most common type
of human remains are hard tissues. Skeletal elements, such as bone and teeth, are
the most durable tissues of the body, and it is only under very exceptional conditions that soft tissues are preserved. Nevertheless, the bones of the individual are
extremely informative in terms of the life and death of the person within their
social and cultural context.
The second part of this dissertation, Bodies of Knowledge, is structured around the
empirical evidence which forms the basis of the central arguments in this dissertation. In Chapter 2, Archaeological Material, I present the historical source material of
this study in terms of its possibilities and limitations. I start with a general description of the shell middens of the Tagus and Sado valleys and discuss the selection of
material, outlining the primary biases of the assemblages. The chapter concludes
with a brief history of research from the perspective of the context of data production which I believe is fundamental for a full understanding of any archaeological
material in historical collections. In Chapter 3, Methods, I outline the methodological
approaches to the study of the mortuary practices at the sites. The chapter is divided into three main sections, each dedicated to a method: radiocarbon dating and
Bayesian analysis, stable isotope analysis of carbon and nitrogen, and archaeothanatology:
chaîne opératoire of mortuary practices. The following chapter, Chapter 4, Results, is
organized by archaeological site where I present the results obtained in the course
of this study. In the last chapter of Bodies of Knowledge, Chapter 5, Analysis and
Synthesis of the Results, I examine the results in terms of the chronology of burial
activity, in terms of defining the people in each burial ground through their
isotopic signatures, and in terms of the treatment of the dead through archaeothanatology, emphasizing the common elements identified in the treatment of the
dead. The results are examined at the scale of the site and valley, and thereafter
compared across the Tagus and Sado valleys.
61
Chapter 2
Archaeological Material: Human Remains in the
Shell Middens of the Tagus and Sado valleys,
Portugal
General description of the shell midden sites in the Tagus
and Sado valleys, Portugal
The archaeological material under analysis is the human remains excavated from
the nineteenth century to the 1960s at the Late Mesolithic shell midden sites of the
Tagus and Sado valleys in Portugal (fig. 2.1).
Figure 2.1. Iberian Peninsula and the location of the shell midden sites of the Tagus and
Sado valleys, Portugal.
These sites are characterized by the anthropogenic accumulation of marine animals, terrestrial fauna, and artefacts made of stone and bone. This type of site is
known as a shell midden after its most visible feature, the sea shells. Here, in the
south-western Atlantic coast of Europe, the relatively rapid rates of sea level rise
during the Atlantic climatic optimum (c 6400–4350 cal BCE) resulted in the forma62
Archaeological Material
tion of large estuaries in the Tagus and Sado Rivers with marine waters moving
towards the inland areas of these river valleys (Arnaud 1989; van der Schriek et al.
2007; Walker et al. 2009). The scattered settlement pattern known for the Early
Mesolithic, during the Pre-boreal, Boreal, and very early Holocene (c 9530–6400
cal BCE), becomes clustered in the innermost areas of large estuaries, by the new
ecosystems formed during the Holocene Atlantic chronozone (Araújo 2003, 2015).
Today, far from the sight and influence of marine waters, these Late Mesolithic
shell middens can be very large archaeological sites, many of them with well preserved human remains. These sites are also rich in faunal remains (Detry 2003,
2007; Lentacker 1986, 1994; Marques-Gabriel 2015; Rowley-Conwy 2004, 2015)
providing evidence for the consumption in situ of a varied range of food resources, both marine and terrestrial (Umbelino 2006; Umbelino et al. 2007). According to these studies there is no evidence for domesticated plants or animals,
except for the domestic dog (Detry and Cardoso 2010; Diniz et al. 2012). Stone
tool production is also well documented for all stages of the chaîne opératoire (Araújo
1995–97; Diniz and Nukushina 2014; Marchand 2001; Nukushina 2012). The relatively few ceramic fragments found in some of these contexts are attributed to later
occupations, to the Early and Late Neolithic (Diniz 2010) and Chalcolithic (Santos
et al. 1972).
The shell middens at the Tagus valley are located in the Ribatejo region, c 80
kilometres north-east from the river mouth in Lisbon. The sites are located on the
banks of three tributaries to the Tagus River: Fonte da Moça, Muge and Magos.
The shell middens of the Sado valley are located in the Alentejo region, c 120
kilometres south of Muge and c 80 kilometres east from the river mouth in
Setúbal. The sites are located near or on the banks of the Sado River. A common
geographic trait of all these sites is their location by the palaeo-estuaries of large
rivers. In the Tagus valley all sites are located at the low terraces by the shores of
the palaeo-estuary, subject to occasional floods. This location close to the aquatic
resources seems to be the privileged position for the development of the site. At
Sado, however, some larger sites are situated further inland and all are located at
relatively higher altitudes. Here, a location with good visibility through the valley
seems to be more important than proximity to aquatic resources (Diniz and Arias
2012, 147).
There is a fundamental difference in site formation in these two valleys which is
reflected in their stratigraphic complexity and ultimately in the spatial position of
the human remains in the sites. In the Tagus valley, these artificial mounds are
formed by the superposition of cultural layers and can reach c 4 to 5 metres in
classic vertical stratigraphy (Roche 1965, 1967). In contrast, the middens in the
Sado valley are relatively shallow and visually difficult to identify. Here, the site
formation is mainly on a horizontal logic of occupation with a current stratigraphic
depth of c 60 to 80 cm at most (Arnaud 1989; Diniz and Arias 2012; Larsson
1996). However, according to the site plans of the Sado valley sites, the human
remains recovered in 1959 from Cabeço das Amoreiras were found at a depth of
c 120–150 cm and in 1960, at Poças de S. Bento, between c 130 and 170 cm deep.
63
Chapter 2
The stratigraphic difference between the sites at the Tagus and Sado valleys is possibly related to the relative location of the sites to aquatic sources but also to ecological constraints under study (Diniz and Arias 2012).
Another common characteristic is the placement of the human remains in clusters within the sites, as observed from the available field records from the old excavations. Although the records are sparser for the Tagus valley, this seems to be
also the case at Cabeço da Arruda, Moita do Sebastião, and Cabeço da Amoreira.
Unfortunately there are no records for the other sites with human remains at the
Tagus valley. At the Tagus valley most of the burials known were recovered from
the bottom layers of the shell middens. However, there are some cases of burials
through the stratigraphy and in the upper layers, confirmed by recent fieldwork in
both Cabeço da Arruda (Roksandic 2006) and Cabeço da Amoreira (Bicho et al.
2013; Roksandic 2006). Similarly, in the Sado valley, most burials were found in the
sandy bottom layers. In general, the burials seem to be spatially organized in tight
areas within a larger area of the site. The relation between these burial clusters and
the remaining areas of the sites remains an open question.
These sites are remarkable because of the large number of human bodies buried
within them. A review of the archaeological data on mortuary practices during the
earliest stages of the Holocene, i.e., Early Mesolithic, in the Iberian Peninsula, prior
to the Late Mesolithic burial grounds in the Tagus and Sado valleys (Peyroteo
Stjerna, forthcoming), shows strong evidence for activities involving manipulation
of human remains, but the practice of funerary burial was rare. This review indicates that there is an increase of burial practice from the Early to the Late Mesolithic in the Iberian Peninsula, as the evidence for open air sites increases as well.
However this growth in burial practice in open air sites during the Mesolithic in
Iberia is only valid when we add the data from the great concentration of burials
from the open air shell middens at the Tagus and Sado valleys. The mortuary practices in the Tagus and Sado valleys are not only in contrast with previous Early
Mesolithic sites but with contemporaneous sites in the Iberian Peninsula, stressing
the regional and social character of mortuary ritual. Interestingly, this regional relationship with death is consistent with the particular settlement known in this region, also unique in the Iberian Peninsula –the formation of large shell midden
sites clustered by the banks of the palaeo-estuaries of the Tagus and Sado Rivers.
This relationship will be discussed further in chapter 6.
Material Selection
From the 13 middens identified in the Tagus valley, nine have human remains
yielding a total of at least 263 individuals (Bicho et al. 2013; Cunha and Cardoso
2001, 2002–03; Jackes and Meiklejohn 2008; Meiklejohn et al. 2009a; Paço 1938;
Roksandic 2006; Santos et al. 1990) (table 2.1, fig. 2.2), although the historical records seem to account for c 340 individuals (chapter 4). At the Sado valley, there
64
Archaeological Material
are 11 known middens, six of which with human burials yielding a total of c 113
individuals (Cunha and Umbelino 1995–97; Diniz et al. 2014) (table 2.2, fig. 2.3).
Table 2.1. Tagus valley: 13 shell midden sites identified, nine of which with recorded human remains.
Tributary Site
MNI
Reference
Fonte da Fonte da Moça 1
Moça
Fonte da Moça 2
fragments of bones
and teeth
fragments of bones
Santos et al. 1990
Muge
Magos
Santos et al. 1990
Fonte do Padre Pedro 2–3
Meiklejohn et al. 2009a
Flor da Beira
4–5
Meiklejohn et al. 2009a
Cabeço da Arruda
110
Moita do Sebastião
85
Cunha and Cardoso 2002–03;
Roksandic 2006
Jackes and Meiklejohn 2008
Cabeço da Amoreira
29
Cova da Onça
32 or 36
Cabeço dos Ossos
–
Magos de Cima
–
Magos de Baixo
–
Cabeço da Barragem
–
Cabeço dos Morros
1, one cranium
Bicho et al. 2013; Cunha and Cardoso 2001;
Roksandic 2006
Cunha and Cardoso 2002–03;
Meiklejohn et al. 2009a
Paço 1938, 74
Table 2.2. Sado valley: 11 shell midden sites identified, six of which with known human remains.
Site
MNI
Reference
Arapouco
32
Cunha and Umbelino 1995–97
Cabeço do Rebolador
0
Cabeço das Amoreiras 6
Report, MNA, APMH, 2/3/7/3, p. 29–32
Cunha and Umbelino 1995–97
Poças de S. Bento
16
Cunha and Umbelino 1995–97; Diniz et al. 2014
Vale de Romeiras
26
Cunha and Umbelino 1995–97
Cabeço do Pez
32–36
Cunha and Umbelino 1995–97
Várzea da Mó
1
Cunha and Umbelino 1995–97
Barrada das Vieiras
–
Fonte da Mina
0
Santos 1967, 1968
Barranco da Moura
0
Santos 1967, 1968
Barrada do Grilo
0
Correspondence, MNA, APMH, 5/1/654/46
65
Chapter 2
Figure 2.2. Tagus valley: shell midden sites with human remains. Site, MNI: Fonte da Moça
1 and 2, fragments of bones and teeth; Fonte do Padre Pedro, 2–3; Flor da Beira, 4–5;
Cabeço da Arruda, 110; Moita do Sebastião, 85; Cabeço da Amoreira, 29; Cabeço dos
Morros, 1; Cova da Onça, 32 or 36 (see table 2.1 for references).
Figure 2.3. Sado valley: shell midden sites with human remains. Site, MNI: Arapouco, 32;
Poças de S. Bento, 16; Cabeço das Amoreiras, 6; Vale de Romeiras, 26; Cabeço do Pez, 32–
36; Várzea da Mó, 1 (see table 2.2 for references).
66
Archaeological Material
Sites selected
This dissertation relies largely on source material from historical collections and
documentation in museum archives as well as on previously published literature.
The source material and documentation available varies greatly from site to site,
depending on the approach of the different excavation teams but also on the curation of the material. The selection of the sites for this study was based on the
following criteria:
• maintenance of the human remains in natural blocks of concretion or
artificial preservation in blocks of paraffin wax;
• access to and quality of the field records regarding graphic documentation of the human remains such as photographs, drawings, and site
plans;
• access to written documentation, such as field notes, diaries, letters, and
reports.
Following these criteria, the material selection comprises two sites in the Tagus
valley: Moita do Sebastião and Cabeço da Arruda, and five sites in the Sado valley:
Arapouco, Cabeço das Amoreiras, Vale de Romeiras, Cabeço do Pez, and Várzea
da Mó. The archaeological evidence retrieved from these seven sites is presented in
chapter 4 and constitutes the core of the argument in this dissertation. The sites of
Cabeço da Amoreira in the Tagus valley and Poças de S. Bento in the Sado valley
are not treated in detail, but for a more robust analysis of the local setting they
have been included in the discussion of the results (chapters 5 and 6).
A note on the museum collections
The current location of the collections results from the long and complex history
of research which became more complicated due to the lack of a central administration for archaeological work in Portugal, only established in the early 1980s
(Fabião 1999, 2011). The archaeological material excavated at the shell middens of
the Tagus and Sado valleys is housed in three main museums (table 2.3).
The scattered provenance of the assemblages results from the institutional affiliations of the excavation leaders, who normally chose to deposit their newly
excavated archaeological material in the affiliated museum. Two other institutions
hold a small part of the assemblages. These are the National Museum of Natural
History and Science, University of Lisbon (Museu Nacional de História Natural e da
Ciência, MUHNAC) which keeps the cranial remains of a few individuals excavated
in the nineteenth century, possibly from Cabeço da Arruda; and the Research
Centre for Anthropology and Health, University of Coimbra, which holds the
human remains excavated in the Tagus valley from the 2000s to the present.
67
Chapter 2
Table 2.3. Museums curating archaeological material excavated at the shell middens of the Tagus and
Sado valleys.
Museum
Excavation year/team
Site
Geological Museum, Lisbon
Museu Geológico, MG
19th c./Geological Commission
Tagus valley, all
1960s/Dir. J. Roche
Cabeço da Amoreira
Cabeço da Arruda
Museum of Natural History,
University of Porto
Museu de História Natural da Universidade do Porto, MHNUP
1930s/Dir. M. Corrêa
Cabeço da Amoreira
Cabeço da Arruda
1950s/Dir. J. Roche
Moita do Sebastião
National Museum of Archaeology,
Lisbon
Museu Nacional de Arqueologia, MNA
1950s–present
Sado valley, all
There are no comprehensive inventories published or available in the museums.
The only exception is Moita do Sebastião, an inventory of human remains for
which was published in detail (Ferembach 1974; Roche 1972). Inventories of the
human remains recovered at the Tagus valley have been drawn up by independent
researchers at least four times: in 1969 by C. Meiklejohn, in the 1980s by C. Meiklejohn and M. Jackes, in 1995 by E. Cunha and C. Umbelino (later with F.
Cardoso), and more recently by M. Roksandic (see Jackes and Meiklejohn 2004;
Meiklejohn et al. 2009a). The material from the Sado valley was studied by E.
Cunha and C. Umbelino (Cunha and Umbelino 1995–97, 2001).
Notation of the human remains
Typically, the skeletons were numbered sequentially by each excavation team, resulting in duplicate numbers in several site assemblages. In this dissertation, all
skeleton numbers are accompanied by the year of excavation and abbreviation of
the name of the site (e.g., CAR1937, Sk7, meaning Cabeço da Arruda, 1937,
Skeleton 7). In the beginning of this dissertation, I provide a list of abbreviations,
including those adopted for the names of the sites.
In recent excavations the nomenclature is more explicit, and in these cases I follow the excavator’s terminology. This is the case for the material excavated in the
early 2000s at Cabeço da Arruda and Cabeço da Amoreira, where the first part of
the nomenclature refers to the name of the site (abbreviated), followed by the year
of excavation and sequential number of the individual (e.g., CA–00–01, meaning
Cabeço da Arruda, 2000, Skeleton 1) (Roksandic 2006). Similarly, recent material
excavated at Cabeço da Amoreira is first identified by the type of context (burial),
followed by the year of excavation and sequential number of the skeleton (e.g.,
Burial 2011.2) (Bicho et al. 2013).
68
Archaeological Material
A note on the historical documentation
Tagus valley
Field notes of the nineteenth century excavations in the Tagus valley are extremely
rare. The only documentation known was identified in the 1980s in the archives of
the Geological Museum in Lisbon, but was possibly recently moved to a central
archive (see Jackes and Alvim 2006). These short notes were published and commented upon by M. Jackes and P. Alvim (2006). The few known field photographs
taken in the nineteenth century have also been published (Cartailhac 1886; Ribeiro
1884).
The written documentation for the fieldwork done in the 1930s and the years
between the 1950s and 1960s, and produced by archaeologists established in institutions in Portugal is published, at least for the most part, and partially facsimiled
(Abrunhosa 2012; Cardoso and Rolão 1999–2000; Jackes et al. 2014). During the
course of this study I tried to locate the personal archives of Jean Roche, a French
archaeologist who carried out extensive work at the sites of the Tagus valley between 1950s and 1960s. J. Roche (1913–2008) (Debénath 2008) was an abbot.
After his death, his belongings remained at his monastery but his archives were
allegedly destroyed due to lack of storage facilities of the religious group, and apparently due to lack of interest from French institutions allegedly contacted by the
religious institution (Grégor Marchand, pers. comm.). Fortunately, J. Roche, unlike
most pioneer archaeologists, published regularly, offering detailed analyses and
extensive descriptions of stratigraphy, spatial organization, and archaeological material (Roche 1956, 1959, 1964–65, 1965, 1967, 1972, 1974, 1989).
Sado valley
Systematic fieldwork at the Sado valley begins with Manuel Heleno (1894–1970)
who was the director of the National Museum of Archaeology (MNA) from 1929
to 1964 (Raposo 2013). It was under his leadership that the first and largest excavations took place at the shell middens of the Sado valley, between 1955 and 1964.
Recently, the museum acquired from his family, the Personal Archive of Manuel
Heleno (Arquivo Pessoal de Manuel Heleno, APMH). This is an invaluable research
source and consists of graphic and written documentation, extensively used in the
course of this study. In this dissertation, references to documentation from the
Archive of Manuel Heleno are identified with the museum’s reference code (MNA,
APMH). In the case of written documentation, the original date and/or location of
the document may be added for further reference (e.g., MNA, APMH,
5/1/664/89, 2/2, Água Derramada, 19 Aug. 1964).
The graphic documentation in APMH comprises hundreds of photographs (c 12 cm
x 8 cm) taken during fieldwork at numerous sites excavated during his directorship.
Unfortunately several of these photographs were already in very poor condition
when acquired by the museum.
The written documentation consists mainly of excavation reports and letters received by M. Heleno, from Jaime Pereira Roldão, who was the main excavator and
69
Chapter 2
responsible for all fieldwork. The original documents are archived at MNA and are
available in digital format (jpeg file) at the museum. The excavation reports
(relatórios) are the field notes handwritten by J. Roldão in a diary format. The entries
are direct and simple and while J. Roldão is not very detailed in his descriptions, he
writes in a systematic manner. Typically, he starts by identifying the area under
excavation, including basic stratigraphic information, and normally, at the end of
each entry he lists the material recovered during that particular day. In between
these two almost obligatory fields – area under excavation and list of material – he
writes down some of his observations and short descriptions. Unfortunately he did
not make any sketches to illustrate his notes, just plain written text (fig. 2.4).
Figure 2.4. Excavation report of Cabeço do Pez 1959, page 1. Handwritten by J. Roldão.
(MNA, APMH, 2/3/7/2, p.1)
The archive of correspondence consists of handwritten letters sent from J. Roldão to
M. Heleno (MNA, APMH, 5/1/654, Box 12, Sado valley, shell middens). Most
letters were sent from Torrão and a few were sent from neighbouring localities,
such as Vale do Guizo. The letters are short, and like the excavation reports, not
very rich in detail.
This documentation is particularly useful to reconstruct the chronology of the
excavations and to understand how much time and effort was dedicated to each
site. It also informs us on excavation methods and documentation, and most
70
Archaeological Material
importantly, it helps us understand which archaeological materials were selected to
recover and register. The excavation of human remains was the main focus of the
fieldwork, and it is clear that if no human remains were found, the excavation work
would halt and move to another site. Yet there is an acute interest and attention to
the recovery of microliths, even the very small.
The letters mention in more or less detail work done at: Cabeço do Pez (1959),
Cabeço do Rebolador (also known as Vale do Guizo) (1959), Vale de Romeiras
(1959, 1960), Várzea da Mó (1959), Barrada do Grilo (1960), and Poças de S.
Bento (1960, 1964). There are only a few comments about Arapouco (1960), and
the fieldwork at Cabeço das Amoreiras is not mentioned in the available
correspondence. The correspondence is incomplete and the letters reporting the
first excavations at Cabeço do Pez in 1956, the fieldwork at Cabeço das Amoreiras
in 1959, and the excavations at Arapouco in 1961 and 1962 are missing from the
archive. Written data directly related to the recovery of human remains is limited.
Most of it seems to be lost (the reports only mention the excavation of one
skeleton at Cabeço do Pez identified on 1st June 1959), and the few descriptions in
the letters are very brief.
Other sources of graphic documentation at MNA consist of field sketches,
drawings and site plans. The archives of drawings and site plans are in the department of inventory at MNA. The site plans are in Archive A and are identified with
the letter A along with serial number (e.g., A118). Archive B contains all detailed
drawings such as those of human skeletons or animal bones, which are identified
with the letter B along with serial number. A separate folder holds the original field
sketches which have no archive number. The material dated to 1956 was drawn by
A. Paiva, while all documentation after 1958 was drawn by Dario de Sousa. During
the course of this dissertation I digitalized (tiff format) all photographs, sketches
and drawings with human remains at MNA. All site plans in the archives of the
museum were photographed and processed digitally (tiff and jpeg formats) by José
Paulo Ruas in 2013.
Sites excluded
Sites with human remains
Several sites with human remains were not analysed in detail in the scope of this
dissertation. This was in part due to time constraints, and the sites with best
documentation were prioritized. However, most of these sites have little or no
documentation, and the human remains are in general poorly preserved not allowing the type of analysis proposed in this study. These sites are: Fonte da Moça 1
and 2, Fonte do Padre Pedro, Flor da Beira, Cabeço da Amoreira, Cova da Onça,
and Cabeço dos Morros in the Tagus valley; and Poças de S. Bento in the Sado
valley.
71
Chapter 2
Tagus valley, Fonte da Moça: Fonte da Moça 1 and 2
The sites of Fonte da Moça 1 and 2 are located on the banks of the tributary with
the same name, north of the Muge area (fig. 2.2). Site 1 was excavated in 1981–83.
The human remains recovered from FM1 are fragments of crania and mandibles,
unidentified bones and various loose teeth. The skeletal elements were found in
layer 4 just above the basal terrace sands (layer 5) where a few perforated shells
were recovered (Santos et al. 1990, 34). Site 2 is located c 2.3 kilometres from site 1
and is partially destroyed. Fieldwork done in 1986, on the basal sand layer, identified several fragments of human bone including cranial fragments (Santos et al.
1990, 37). The minimum number of individuals identified on both sites is
unknown.
Tagus valley, Muge: Fonte do Padre Pedro
The shell midden at Fonte do Padre Pedro is located on the right bank of the
Muge tributary (fig. 2.2), and was already destroyed in 1880 (Roche and Ferreira
1967, 20). Documentation about this site is extremely limited. The human remains
recovered by C. Ribeiro consist of one femur found on the surface layer along with
fragments of shells and faunal remains (Cardoso and Rolão 1999–2000, 121;
Ribeiro 1884, 281).
The site was surveyed by M. Corrêa in June 1930, but the two test pits revealed
just a few fragments of shells, which were followed by a sterile layer. The only
artefact mentioned in O. V. Ferreira’s field notes is “one beautiful flint point”
(Cardoso and Rolão 1999–2000, 144). M. Corrêa visited the site in August 1930
reporting that it was completely destroyed by the vineyard (Cardoso and Rolão
1999–2000, 144). The last note about Fonte do Padre Pedro is from August 1931,
when apparently a skeleton was found while planting the vineyard (Gonçalves
1986).
The human remains are curated at MG in mixed and disarticulated state. Recent
anthropological analysis by M. Roksandic indicates a MNI of 2–3 (one child, and
one or two adults) (Meiklejohn et al. 2009a, 13). It is unclear if this estimate
includes the latest skeleton found in 1931.
Tagus valley, Muge: Flor da Beira
Flor da Beira is located on the right bank of the Muge tributary between Fonte do
Padre Pedro and Cabeço da Arruda (Paço 1938) (fig. 2.2). The site was discovered
in 1935 by H. Cabaço (Paço 1938) and reported to be destroyed (Paço 1940).
The human remains are curated at MG in mixed and disarticulated state. Recent
anthropological analysis by M. Roksandic indicates a MNI of 4–5 (Meiklejohn et al.
2009a, 13). There is no documentation for this site and the first published
reference is probably by A. Paço (1938) (Umbelino 2006, 89).
72
Archaeological Material
Tagus valley, Muge: Cabeço da Amoreira
Cabeço da Amoreira is a shell midden site located on the left bank of Muge
tributary on the opposite side to Cabeço da Arruda, and just c 700 metres east of
Moita do Sebastião (fig. 2.2). The site was excavated in the nineteenth century by
the Geological Commission, in the 1930s by M. Corrêa, in the 1960s by J. Roche,
in 2001–03 by J. M. Rolão, and since 2008 by N. Bicho.
Graphic documentation is limited but published, at least partially (Abrunhosa
2012; Cardoso and Rolão 1999–2000). Unfortunately the human remains are in
general poorly preserved, stored in disarticulated state, mixed and mislabelled in
some cases. The MNI currently curated at MHNUP is 21 (Cunha and Cardoso
2001, 326), but according to the excavation reports this number could be between
28 and 29. Recent excavations have recovered 7–8 individuals (Bicho et al. 2013;
Roksandic 2006). Field notes (Cardoso and Rolão 1999–2000, 121, 160), site plans
and early literature (Newell et al. 1979; Roche 1964–65) indicate that 11–12
individuals were excavated in the 1930s, however, the anthropological analysis
indicates 15 individuals (10 adults and 5 non-adults) in the assemblage (Cunha and
Cardoso 2001, 326; 2002–03, 175). Of these, only three are almost complete
skeletons; seven are incomplete and two individuals are represented by one bone
each. Sex determination was possible for six skeletons: three males and three
females (Cunha and Cardoso 2001; 2002–03). Field documentation from J. Roche’s
excavations in the 1960s indicates the excavation of 17 individuals (Cardoso and
Rolão 1999–2000).
The MNI adopted in this study (29) follows the estimates by E. Cunha and F.
Cardoso (2001) (MNI 21), in addition to the recently excavated material (MNI
7–8) (Bicho et al. 2013; Roksandic 2006) (table 2.1).
Tagus valley, Magos: Cova da Onça
Cova da Onça, also known as Arneiro do Roquete, was the first shell midden to be
identified in Portugal, in April 1863 (Ribeiro 1884). The site is located on the right
bank of Magos tributary (fig. 2.2) near another midden named Cabeço dos Ossos
(meaning, hill of the bones). The two sites are located in a larger area named
Quinta da Sardinha and are separated from each other just by a small depression
(Paço 1938, 67). Both Cova da Onça and Cabeço dos Ossos have been destroyed.
Apparently, at the time of the discovery of Cova da Onça, the site was not
excavated because C. Ribeiro did not get the permission from the landowner.
Later, A. Paço (1938, 1940) mentions that the Museum of the Geological Commission (nowadays MG) was storing some material under the name of Cova da Onça,
but unfortunately he does not mention of which kind.
There is no historical documentation known for this site. It is only known that
the site was relocated by H. Cabaço in 1935 (Paço 1938), and that the skeletal
remains of more than 30 individuals are curated at MG. The human remains are
stored in disarticulated state and all skeletons are mixed in a few containers (Cunha
and Cardoso 2002–03; Meiklejohn et al. 2009a). The material is marked with a
73
Chapter 2
handwritten label from 1915. It is unknown when the skeletons from Cova da
Onça were excavated and it is difficult to explain the year written on this label
(1915) because there are no excavations reported between 1885 and the 1930s
(Paço 1938; Roche 1965). There is however a brief reference to the recovery of
unspecified anthropological material (Corrêa 1940; Paço 1938). A. Paço mentions
only the recovery of some lithic and anthropological material, as well as some
shells excavated from test pits at Cabeço dos Ossos, Cova da Onça and Cabeço
dos Morros (Paço 1938, 69).
The MNI is 32 (Cunha and Cardoso 2002–03) or 36 (8 non-adults; 28 adults,
Meiklejohn et al. 2009a). One radiocarbon date on human bone confirms the
Mesolithic chronology of the human remains (Cunha and Cardoso 2002–03) (table
A7).
Cova da Onça was surveyed in 1984 by A. M. Flores and V. Gonçalves, to
evaluate the preservation of the shell midden. During these few weeks, the team
recovered abundant lithic material, two perforated shell beads (Trivia monacha), two
fragments of ceramics on the top layer, and identified a hearth, and an unspecified
negative feature. There is no reference to the excavation of human remains
(DGPC 2015).
Tagus valley, Magos: Cabeço dos Morros
Cabeço dos Morros is located on the left bank of Magos tributary (fig. 2.2). The
site was described as a large shell midden possibly similar to Cabeço da Arruda, but
mostly destroyed (Paço 1938, 67).
In the 1860s, C. Ribeiro recovered a few fragments of human remains (Cardoso
and Rolão 2000, 87). In the 1930s, one cranium (no mandible) was found by H.
Cabaço. This cranium was sent to M. Corrêa (Institute of Anthropology,
University of Porto) who recognized similarities with the crania found in Muge
which M. Corrêa identified as the Homo afer taganus race (Paço 1938, 74) (see below,
Understanding the context of data production: brief research history). The site was recently
excavated by J. M. Rolão between 1997 and 2001 (Detry 2008, 53), and there is no
reference to the excavation of human remains.
Sado valley: Poças de S. Bento
Poças de S. Bento is located at the Sado valley (fig. 2.3), and unlike most sites, it
stands on a plateau c 3 kilometres from the river valley at 85 m.a.s.l. (Arnaud 1989;
Araújo 1995–97; Diniz and Arias 2012).
Archaeological material and documentation available for Poças de S. Bento is
curated at MNA. The site was first excavated by the museum staff from September
1960 (MNA, APMH, 5/1/654/54, 6 Sep. 1960) until the end of November. The
second and last field season organized by the museum, started in August 1964
(MNA, APMH, 5/1/664/89, 2/2, Água Derramada, 19 Aug. 1964) until at least
the end of September of the same year (MNA, APMH, 5/1/654/94, 2/2, 27 Sep.
1964).
74
Archaeological Material
The site was excavated in 1986–88 by J. M. Arnaud and L. Larsson (Arnaud
2000; Arnaud and Larsson 1994; Larsson 1996, 2010). The trial excavation in 1986
uncovered one skull in the basal sand. According to the excavators, these remains
were possibly associated with skeleton 11, excavated in the 1960s (Larsson 1996,
2010). The skull was dated (Ua–425, 5390 ± 110 BP, Larsson 2010; Lars Larsson,
pers. comm.), and despite its stratigraphic position it represents a relatively recent
date, possibly contemporaneous with a few polished stone tools excavated at the
site (Arnaud 1990). The isotopic values of this measurement are not available and
the calibration of this sample is uncertain. If the individual had a full terrestrial diet,
the calibrated date range for his/her death would be 4448–3984 cal BCE (95%
confidence). The ranges become slightly more recent when calibrated with the
marine curve. For example, with an assumed marine diet of 30%, the calibrated
date range is 4365–3819 cal BCE (95% confidence) (table A14).
J. M. Arnaud and L. Larsson identified the remains of one individual (individual
13) eccentric to the main burial area identified in 1960 (Larsson 1996, 131). Unlike
the skeletons found in the early excavations in the bottom layer, this individual was
found on the top layer, just above the layer with shells. The cranium and mandible
of the skeleton were not identified and the bones of the trunk were poorly
preserved (Larsson 1996, 130). This material was not radiocarbon dated (Lars Larsson pers. comm.).
Graphic documentation at MNA consists of two site plans with all burials in the
excavation context (MNA, 1960, D. Sousa, A118, A119) (fig. 2.5), several individual drawings of the human remains (individuals 1 to 12) and one drawing of a
group of faunal remains. The quality of the 68 field photographs available for
Poças de S. Bento (MNA, APMH, 2/11/57) is poor and most images are unclear.
The human skeletons are in disarticulated state and stored in containers. The
material is poorly preserved and incomplete due to fragmentation. The assemblage
was studied by a team of anthropologists in the 1990s (Cunha and Umbelino
1995–97, 2001; Umbelino and Cunha 2012) who estimated a minimum number of
individuals of 15 (11 adults; 4 non-adults). The poor preservation of the remains
did not allow sex determination or more precise age estimates (Cunha and
Umbelino 1995–97, 168).
Recent excavations directed by M. Diniz and P. Arias (2010–ongoing) identified
the burial of a dog (Canis familiaris) on the opposite side of the burial area with the
human remains (Diniz et al. 2012), as well as the burial of a female adult near the
area where all human burials were found in the 1960s and 1980s (Diniz et al. 2014).
75
Chapter 2
Figure 2.5. Poças de S. Bento, Sado valley: site plan. Area with the human remains excavated
in 1960. (Adapted from a photograph by J. P. Ruas. MNA, 1960, D. Sousa, A119)
Sites without known human remains
The human remains excavated in the shell middens at the Tagus and Sado valleys
constitute the core material of this research. The sites without known human
remains were not investigated in the scope of this dissertation.
In the Tagus valley these sites are located on the banks of the Magos tributary:
Cabeço dos Ossos (also known as Monte dos Ossos), Magos de Cima, Magos de
Baixo, and Cabeço da Barragem. Although speculative, the lack of known human
remains in these sites could be explained by the poor preservation of the middens
when identified, due to floods and farming.
In the Sado valley, the sites without known human remains are: Cabeço do
Rebolador (or Vale do Guizo), Fonte da Mina, Barranco da Moura, Barrada das
Vieiras (also known as Portancho), and Barrada do Grilo. Cabeço do Rebolador
was excavated in the summer of 1959 (MNA, APMH, 2/3/7/3, pp. 29–32) and no
human remains were found. Fonte da Mina and Barranco da Moura were partially
excavated, but the small test pits had few results and no human remains were
found (Santos 1967, 1968). Barrada das Vieiras (or Portancho) was the first shell
76
Archaeological Material
midden identified at the Sado valley but apparently it was never excavated, or at
least never reported to be. Barrada do Grilo was largely excavated by the museum
staff but no human remains were found (MNA, APMH/5/1/654/46, 2/2, Torrão,
7 Aug. 1960). In the inventory of MNA, the site of Barrada do Grilo is wrongly
identified as the same site of Várzea da Mó, when these are in fact two sites
(MNA, APMH/5/1/654/26, Torrão, 2 Aug. 1959; Santos et al. 1972). For this
reason, Barrada do Grilo was mistakenly published as a site with one burial
(Marchand 2001) when in fact this burial was excavated at Várzea da Mó.
Total funeral
The archaeological remains of the mortuary practices of the last hunter-gatherers
of the Tagus and Sado valleys most likely represent just a portion of the total
funeral. Practices involving the treatment of the body after death, such as washing
or clothing of the cadaver, the ritual words, the stories told and the songs intoned,
the dancing, and any other bodily gestures performed before, during and after the
placement of the cadaver in the grave, are normally not accessible to archaeology
because they do not leave material traces. Our reconstruction of the mortuary
practices is thus limited to what remains in the archaeological record and to what
the archaeologist can read from it.
Another related limitation is expressed by the archaeological notions of incompleteness and the total record (see Lucas 2012, 18–73). The archaeologist is well
aware of the fragmentary nature of the archaeological record, which can be understood in terms of its representativeness of past human actions, preservation of the
material remains and rate of recovery during archaeological excavation (Lucas
2012, 19). In order to overcome these challenges, it is important to consider
carefully the relationship between the archaeological assemblage and the human
groups under study. How representative are the burials in relation to the original
group? Are we dealing with a full collection as excavated? What is the relationship
of the excavated sample to the site as a whole? Are the excavated areas representative of the cemetery as a whole or of only discrete areas of the cemetery (Mays
2010)? How does our sample signify the universe of funerary practices carried out
at a certain place and time in the past?
Preservation and research bias
Organic preservation
The first aspect to consider concerns the preservation of organic material in the
Mesolithic sites known in the Iberian Peninsula. A review of the archaeological
data available for the mortuary practices during the earliest stages of the Holocene
in the Iberian Peninsula (Peyroteo Stjerna, forthcoming) shows that human
remains are preserved in c 37 sites, which is less than 20% of the c 200 Early and
77
Chapter 2
Late Mesolithic sites known in Iberia. However, 117 of these sites have faunal
assemblages (Gallego Lletjós 2013, 452), indicating bone preservation in at least
60% of the sites. Moreover, some of these sites have been excavated quite intensively, and if these sites had had human remains, these would have been preserved
and found, at least in many of them.
Settlement network bias
Another aspect to consider when evaluating the representativity of the data is
related to inevitable research bias. It is true that the concentration of sites with
human remains corresponds to the areas where archaeological research is also
more intense (Gallego Lletjós 2013). Also, and despite the geoclimatic variability in
Iberia, it has been shown that when a specialized systematic survey is done, new
sites are identified (Arias et al. 2009). Yet new finds have been very limited and
have not changed the current Mesolithic paradigm for settlement patterns, despite
recent survey programmes and large archaeological rescue projects in the context
of local construction and land development (Araújo 2015).
Visibility of the archaeological record
The visibility of these sites is also an important issue to consider. The apparent
lack of open air sites without shell layers and potentially with human remains could
be a result of research bias, as these sites are obviously more difficult to identify
than caves or shell middens. I have shown (Peyroteo Stjerna, forthcoming) that the
shell midden/shell layer component is not a determinant variable for the presence
of human remains during the Mesolithic. There is evidence for the disposal and
preservation of bone in both situations, with and without shell layers. Interestingly,
the sites with shell midden layers tend to contain more inhumations than the sites
without. The relationship between shells and burials is discussed further in chapter
6, and is likely not just the result of good bone preservation in such environments,
but of cultural choices.
Representativeness of the sample under study
The sites selected for this study are Moita do Sebastião and Cabeço da Arruda in
the Tagus valley and Arapouco, Cabeço das Amoreiras, Vale de Romeiras, Cabeço
do Pez and Várzea da Mó in the Sado valley.
Moita do Sebastião and Cabeço da Arruda are the largest shell midden sites
known in Portugal, with the highest number of human burials. The sites have been
extensively excavated and the current assemblage is likely representative of the
population buried in these sites. The analysis may be biased if we consider these
two sites as representative of the mortuary practices at the scale of the entire Tagus
valley. For this reason, Cabeço da Amoreira in the Muge area in the Tagus valley
was included in the discussion, although not treated in detail because of the limitations of the source material.
78
Archaeological Material
In the case of the data selection for the Sado valley, I included all sites with human remains with the exception of Poças de S. Bento due to poor preservation of
the skeletal material and unclear documentation. Overall, the assemblage is smaller
than for the Tagus valley, however, the nature of the documentation for the Sado
material allows a more detailed analysis. With the possible exception of Vale de
Romeiras, none of these sites were completely excavated, and ongoing fieldwork
has shown that a few more human burials may be preserved in these sites (Diniz et
al. 2014). However, I believe we can be confident about the representativeness of
the current assemblage due to the efficient recovery methods employed by the
museum staff in the 1950–60s (Arnaud 1989, 615, 2000, 24–27). The strategy in
these earlier excavations, when more than 100 burials were recovered, consisted of
opening long narrow trenches along the recognized limits of the site. Then, if
human remains were found these trenches were expanded. In all sites the human
remains were found in one trench, later expanded as the main excavation area.
One exception is the large site Cabeço do Pez, where the site plan from 1956
(MNA) shows two areas with human remains.
Of great utility to this study, are the skeletons preserved in natural or artificial
blocks. In some cases, the human remains were recovered with the sediment
concretion enveloping the bones, maintaining the skeleton in the position it was
when excavated. In other cases, the skeletons were covered with paraffin wax and
transported in blocks to the museums. This later procedure contaminated the
bones limiting reliable chemical analysis of the human remains. However, the
position of the skeleton in the archaeological deposit was maintained.
The great limitation of this analysis lies in the quality of the field documentation. The field notes when available are short and the description of the human
remains is in general brief and superficial. The graphic documentation is absent for
many sites and of variable quality. Each site has particular constraints and many
possibilities, and these are described in detail in chapter 4.
Hunter-gatherer death rate vs burial population
Representativeness of the archaeological burials in relation to the original population can be estimated by the calculation of death rate of an assumed group size,
within a chronological framework.
Ethnographic evidence indicates that nomadic foragers live most of their lives
in small groups, often fewer than 25 persons, although they might periodically
gather in larger groups (Kelly 2013, 172). Recent reanalysis of Binford’s (2001)
dataset of 339 hunter-gatherer groups (Hamilton et al. 2007) suggests the following
numbers for group units of nomadic foragers:
• Families: 4–5 people
• Residential group: 14–17 people
• Social aggregations (e.g., at winter camps): 50–60 people
• Periodic aggregations: 150–180 people
• Ethnic population: 730–950 people
79
Chapter 2
Death rate patterns observed in several groups of historical hunter-gatherers
(Gurven and Kaplan 2007) (table 2.4) along with knowledge of typical huntergatherer group sizes can be useful to test different scenarios.
Table 2.4. Death rate among hunter-gatherers. Data from ethnographic studies, compiled by M. Gurven
and H. Kaplan (2007).
No. yrs
studied
11
No. of
individuals
500
No. of
deaths
94
Deaths/year
estimate (no.)
8.5
Deaths/year
estimate (%)
1.7
14
971
353
25.2
2.6
Agta
Tropical to humid
24
sub-tropical, Philippines
176
117
4.9
2.8
Hadza
Topical savannah,
Tanzania
10
706
125
12.5
1.8
Hiwi1
Tropical savannah,
Venezuela
2
375
107
53.5
14.3
Group
Environment
!Kung
Kalahari desert,
Botswana/Namibia
Ache
Tropical forest,
Paraguay
To test one possible scenario at the Tagus and Sado valley sites one may speculate
on the number of deaths per year (table 2.4) in one residential group of between
15–17 people. If such a group had low death rate, such as the value estimated for
the !Kung from the Kalahari desert (1.7 percent per year), this group would experience 0.3 deaths per year, or three deaths every ten years (30 deaths in 100 years). In
the case of higher death rate, such as the value estimated for the Agta from the
forests of the Philippines (2.8 percent per year), a residential group of 15–17 people would experience 0.5 deaths per year, or 5 deaths every ten years (50 deaths in
100 years). Osteological analyses of the human remains from the Tagus and Sado
valleys do not show evidence for inter-personal violence (Cunha et al. 2004; Jackes
2004), thus the high death rate estimated for violent groups such as the Hiwi, can
be excluded from the current estimate.
In an archaeological population, calculating the death rate of a residential group
of hunter-gatherers is not sufficient to estimate the representativeness of the burial
population in relation to the living group. Chronological resolution is fundamental,
and without this control one can only speculate how representative these burial
grounds are of the people using them, and of their mortuary practices. M. Jackes
and colleagues (1997, 652) have suggested that the group size of the huntergatherers frequenting the sites at the Tagus valley was small, and calculated that
only approximately three women of child-bearing age in each generation lived in
the area of Muge. This suggestion and other aspects related to whether these burial
grounds provide a full picture of how the remains of the dead were handled is
discussed further after the analysis of the radiocarbon data in chapters 5 and 6.
1
The Hiwi are known for well documented inter-group violence (Gurven and Kaplan 2007).
80
Archaeological Material
A note about longevity among hunter-gatherers
M. Gurven and H. Kaplan’s (2007) study on the death rate of hunter-gatherers
suggests that modal lifespan is 68–78 years (Gurven and Kaplan 2007, 349), in
contradiction to previous claims suggesting that among early humans, few individuals passed forty years, and it was exceptional to pass fifty (Vallois 1961, 222).
The authors conclude that extensive longevity appears to be a novel feature of
Homo sapiens although juvenile mortality is high in hunter-gatherer populations
(Gurven and Kaplan 2007).
Following this, in this dissertation I have assumed that a human generation
represents 25 years. For a maximal human lifespan I have considered 75 years,
though the average was presumably shorter (see Whittle et al. 2007, 131–2).
Understanding the context of data production: brief research
history
The Mesolithic shell middens located in Portugal have been the subject of research
for more than 150 years and the history of their research overlaps with major developments in archaeological thought in Portugal and elsewhere. The aim of this
section is to summarize and contextualize the most relevant phases of shell midden
research in Portugal within its historical background. Understanding the context of
production of the archaeological data is fundamental when studying historical
collections, as it allows better evaluation of the possibilities and limitations of the
archaeological material at our disposal.
This brief review follows a chronological order and links each research phase to
its historical background, emphasizing the scientific and/or political perspectives
of the time which, for the purposes of this dissertation, I consider relevant to highlight. Also, for the scope of this work, I treat the context of data production in the
historical collections (nineteenth century–1960s) in greater detail. This section
starts with the discovery of the Portuguese shell middens and the birth of prehistoric archaeology in Portugal in the context of contemporaneous scientific debate on the antiquity of mankind. Then, I discuss the research work on the shell middens in the
light of the use of archaeology in political discourse, in a section titled From the
antiquity of mankind to the antiquity of nations: human remains and discourses of identity.
After defining these two early phases of research, I focus on the first modern
approaches to archaeological research during the 1950s and 1960s in the field and
in the laboratory, and I present a summary of the research done over the last 30
years.
81
Chapter 2
Portuguese shell middens and the birth of prehistoric archaeology in
Portugal
The first shell middens identified in Portugal are located in the Tagus valley by the
Muge and Magos tributary rivers. These archaeological sites were discovered by
geologists commissioned to survey the country for geological and natural resources
(Fabião 2011, 105–7). Between 1863 and 1885, the Geological Commission (Comissão
Geológica do Reino) identifies five and excavates four shell midden sites, all of them
with human remains (Cova da Onça, Fonte do Padre Pedro, Cabeço da Arruda,
Moita do Sebastião, and Cabeço das Amoreiras). During this pioneer phase, more
than 250 human skeletons were excavated at Fonte do Padre Pedro, Cabeço da
Arruda, Moita do Sebastião, and Cabeço da Amoreira (Paula e Oliveira 1884,
1888–92; Pereira da Costa 1865, Ribeiro 1884). The Portuguese middens were
immediately compared to the well known køkkenmøddinger (or kitchen middens) in
Denmark discovered in 1831 (Anderson 2007). Interestingly, in the earliest publications about the Portuguese shell middens, the sites are named kjökkenmöddings or
kioekkenmoeddings, after the Danish term, even though the texts were written in
Portuguese and/or French by Portuguese researchers (Paula e Oliveira 1884,
1888–92; Pereira da Costa 1865; Ribeiro 1884).
The antiquity of mankind
The goals of this national geological inventory were naturally very distinct from an
archaeological survey. However, the theme of the antiquity of humankind was not
marginal to the members of this commission (Fabião 2011, 107). Interest in human
antiquity is well framed in a European trend at the time (Renfrew 2008; Schnapp
1997; Trigger 2007), and reinforced by the clear antiquity of certain finds recovered
during fieldwork.
The works of C. Darwin On the Origin of Species (first published in 1859 and
translated into French in 1862) and The Descent of Man (first published in 1871 and
translated into French in 1872) were known among Portuguese scholars and writers (Fabião 2011; Pereira 2001). However, The Geological Evidence of the Antiquity of
Man by C. Lyell (first published in 1863 and translated into French in 1864) had a
greater influence at the time, at least among the scholars involved in geological and
archaeological fieldwork. This is well illustrated by Pereira da Costa, the author of
the first study published about the Portuguese shell middens and credited as the
first work in prehistoric archaeology in Portugal: Da existência do homem em epochas
remotas no valle do Tejo. Primeiro opusculo. Noticia sobre os esqueletos humanos descobertos no
Cabeço da Arruda, which translates to On the antiquity of man in the Tagus valley. Part
one. Note about the skeletons found at Cabeço da Arruda – bilingual edition in Portuguese
and in French (Pereira da Costa 1865). This pioneering study shows a clear knowledge of C. Lyell’s work and an evident concern with the antiquity of mankind in
the Portuguese territory, which subsequent studies followed.
Following the developments in the natural sciences, during the nineteenth
century there was an increase in the use of the term: archaeology – parallel to its
82
Archaeological Material
emergence as an independent discipline (Schnapp 1997, 275–6). At this time many
societies were formed in Europe focused on themes of anthropology, ethnology
and prehistory. In Portugal, the Association of the Portuguese Archaeologists was
founded in 1863 (Associação dos Arqueólogos Portugueses), and the Ethnological
Museum in 1897 (Museu Ethnologico), nowadays the National Museum of Archaeology (Museu Nacional de Arqueologia, MNA). The first course in anthropology and
archaeology started in 1885 at the University of Coimbra: Antropologia, Palaeontologia
Humana e Arqueologia Pré-historica – Anthropology, Human Palaeontology and Prehistoric Archaeology (Fabião 1999).
The birth of prehistoric archaeology is closely related to the developments in
biology, geology, and other sciences within the Positivist paradigm at the time
(Diniz and Gonçalves 1993–94; Trigger 2007). In Portugal, this happened in a clear
geological framework but with contributors informed by other natural sciences:
Pereira da Costa, for example, was a Professor in mineralogy and geology with a
degree in medicine as well (Fabião 2011, 106). This is an important point because
these prehistoric shell middens were intriguing not only because of their geological
and anthropogenic formation, but also because they enclosed the biological remains of past humans, thus providing the impetus for the development of
prehistoric archaeology in Portugal.
From the antiquity of mankind to the antiquity of nations: human
remains and discourses of identity
Historical background
Portugal has been an autonomous territory with stable borders since the thirteenth
century (Ventura 2007). Perhaps for this reason, “there have been no remarkable
cases in Portugal of nationalistic appropriation of archaeological interpretations or
uses of archaeological monuments as national symbols” (Fabião 1996, 90). Nevertheless, some historical moments received special attention in the formation of a
national identity: the foundation of the country and its defence against other peninsular Christian kingdoms, in particular against the Muslim power; the several
conflicts with Castile (Spain); and the illustrious Age of Discoveries (Fabião 1996, 90).
In the early nineteenth century, A. Herculano (1810–77), the first Portuguese
historian in a modern sense, defined the birth of the nation as an act of political
will. Based on this argument, Herculano rejected any relationship between the
Portuguese people and any other prior populations in the Portuguese territory. In
this view, and breaking with the previous humanist paradigm, Herculano urged
special protection of “the truly Portuguese monuments rather than for the relics of
remoter ages”, that is, protection of the medieval roots rather than those of Roman
heritage or earlier (Fabião 1996, 93). This thesis of the medieval origins of the
Portuguese state was adopted by other influential and popular historians. Ever
since, all Portuguese archaeological work with nationalistic purposes has
83
Chapter 2
attempted, without great success, to integrate the ancient heritage into a concept of
historical identity (Fabião 1996).
Prehistoric remains
During the second half of the nineteenth century, the pioneer studies of the newly
born discipline of prehistoric archaeology were concerned with the discovery of
the great antiquity of humankind, greatly inspired by the revolutionary works of
C. Darwin and C. Lyell.
The nineteenth century was rich in important changes of scientific paradigms
but also in political transformations of national character in Europe. In this context, and throughout the twentieth century, archaeology was used and abused by
national states in order to support various claims of territorial and cultural identity
(Trigger 2007). Following this, in the early twentieth century, there was a significant emphasis on the idea of a primeval Portugal. This idea, supposedly based on
archaeological evidence, proposed that the pre-figuration of the Portuguese nation
was to be found in prehistoric times, hence in contradiction to the previous mentioned thesis of a medieval origin for the Portuguese people. Two main personalities, both influential archaeologists but also great adversaries (Fabião 1999, 2011),
were involved in this shift of identity discourse: M. Corrêa and M. Heleno. Interestingly and not coincidentally, both were involved in excavations at the prehistoric
shell midden sites in the Tagus valley (M. Corrêa) and in the Sado valley
(M. Heleno).
M. Corrêa (1888–1960) was the pre-eminent researcher of the Muge middens
during the 1930s, the second great moment of research into these shell middens.
M. Corrêa had a degree in medicine and was the founder of the Institute of
Anthropology of University of Porto (Cardoso 1999). He proposed that the human remains excavated in the Muge shell middens were of African origin (Corrêa
1919, 1933). This theory was strongly debated and opposed by other anthropologists, such as H. V. Vallois (Vallois 1930, 1940) (see Ferembach 1974, 12–5), but
M. Corrêa was confident in his assumptions and even referred to the Muge specimens as Homo afer taganus. Later however, M. Corrêa revised his theories and accepted other interpretations, such as the Mediterranean origin of the Muge populations (Corrêa 1956). Yet this idea of an African ancestry went well with the national propaganda at the time. The multiracial and pluricontinental Portugal were important ideological concepts in the nationalistic dictatorship with political and
institutional control over Portugal and several territories in Africa and Asia. For
the first time, although for a brief period, the Muge finds with claimed African
ancestry, were placed in the national history.
In the same vein, but in a purely nationalistic rhetoric, M. Heleno, director of
the Museum of Ethnology Leite de Vasconcelos in Lisbon (1929–64) (nowadays
the National Museum of Archaeology, MNA), had put forward the idea that the
Portuguese nation and people had shared moral and unity determined since the
New Stone Age. He emphasized that this unity descended from the Muge Men, who
prevailed throughout the Bronze to the Iron Age, developing through the Roman
84
Archaeological Material
period to medieval times when Portugal was finally constituted a state, although
rejecting M. Corrêa’s thesis of a possible African ancestry in the Early Stone Age
(Fabião 2011, 171–72). Furthermore, and based on no firm archaeological
evidence, Heleno compared the spread of the megalithic culture to a prehistoric
imperialism which source of development was of Portuguese genesis (Fabião
1996). Because of the animosity between these two personalities, M. Corrêa and M.
Heleno (see Fabião 1999, 2011), the National Museum of Archaeology was never
involved in the excavations at the Tagus (Muge and Magos) shell middens. It is in
this context that the first shell midden sites were identified in the Sado valley, c 100
kilometres south from Muge (Barradas 1936). However it was only some decades
later, between 1955 and 1964 that M. Heleno and the National Museum, focused
on systematic fieldwork in the region, excavating and identifying several shell middens in the Sado valley (see below). Unfortunately, contrary to M. Corrêa, M.
Heleno never published anything about the vast archaeological material excavated
during these years in the Sado valley, and the question of what Heleno thought
about these human remains and the origins of the Portuguese people remains a
mystery.
Overall, these ideas had a certain impact and some of the misconceptions remained in the archaeological discourse for many years. However, never again were
prehistoric remains part of the narrative on the genesis of the nation, a topic that
was not in the processual agenda of Portuguese archaeology after the 1970s.
From the late nineteenth century throughout the first half of the twentieth
century, European archaeology studied its human osteological collections from
racial perspectives, aiming at finding national identities based on narratives of
origins, selected relations and networks. Although based on different motivations,
these narratives had strong nationalistic accents. As discussed, in Portugal, the
Tagus valley collections were implicated, although modestly, in arguments related
to identity, ancestry, and origins. These observations were especially interesting to a
colonial power such as Portugal, interested in supporting its ancestral connections
with Africa. However, this African perspective ceased to exist, and in fact during
the second half of the twentieth century, the southern coast of the Mediterranean
disappears from European archaeological maps (see Diniz 2007).
In a country such as Portugal, with a very long history, located in a geographic
space where no reasons besides political ones could justify its independence, it
seems quite obvious that nationalistic discourses favoured historical periods rather
than prehistory, classical antiquity or the early Middle Ages. Even so, Portuguese
archaeologists with nationalistic aims have tried, with little success to relocate the
roots of the nation deeper in history, chiefly as an attempt to achieve social
relevance in the eyes of both political power and public opinion. But, ‘in a country
with so many national “glorious events” to celebrate, who needs poor and controversial archaeological remains from remote ages?’ (Fabião 1996, 105).
85
Chapter 2
The 1950s–60s: towards a systematic archaeology
In the second half of the twentieth century, archaeology became more systematic
discipline with its own field methods and recording techniques (Trigger 2007). In
Portugal, archaeology remained a fragmented discipline, marked by individual
personalities, and with few exceptions, it remained an amateur discipline (Fabião
2011, 177). Nevertheless, in the case of shell midden research, the data production
during this period was outstanding, pioneering in the application of methods and
techniques, such as the first radiocarbon dating in Portugal carried out in the Muge
middens, or the use of the Wheeler method at the Sado valley (fig. 2.6).
The excavations done at the Muge sites remained under the control of M.
Corrêa (Institute of Anthropology of University of Porto). Corrêa did not carry out
the fieldwork himself but invited the French archaeologist Jean Roche to direct the
excavations (Roche and Ferreira 1967). Interestingly, and following the research
tradition at these sites, J. Roche was assisted by Octávio da Veiga Ferreira, member
of the Geological Services of Portugal (Serviços Geológicos de Portugal) – the institution
born from the earlier Geological Commission behind the discovery of the first
middens in the nineteenth century. J. Roche was meticulous and published his
finds regularly. The same can be said about O. V. Ferreira who assisted J. Roche
and directed the excavations in his absence. It was during these decades that the
first radiocarbon measurements were done on archaeological material from
Portuguese sites, obtained from samples of charcoal collected by J. Roche at Moita
do Sebastião, and later from Cabeço da Arruda and Cabeço da Amoreira. This
sampling for dating was done at the request of M. Corrêa who wanted solid data to
establish the absolute chronology of the Tagus shell middens (Roche and Ferreira
1967, 32). Also, the first and only monograph ever published on any of the
Portuguese shell middens was authored by J. Roche (1972) on the site of Moita do
Sebastião.
It is also during these decades, between 1955 and 1964, that M. Heleno, director
of the National Museum of Archaeology focused his attention on the shell midden
sites discovered in the 1930s in the Sado valley. Most middens were identified
during this period and more than 100 skeletons were recovered. These were the
first and the largest excavations ever carried out at these sites but unfortunately the
archaeological material remained unexplored for several decades and nothing was
ever published. Curiously, during this time the Sado middens were never
mentioned in the literature, except in a brief note by Heleno (1956). This is well
illustrated by the published work by J. Roche, O. V. Ferreira, and D. Ferembach,
who were active and prolific researchers of the Muge middens, but never
mentioned the middens in the Sado valley, which could have been used for
comparison with their finds. This omission in the literature was most probably
because nobody knew anything about the archaeological material excavated at the
Sado sites, except M. Heleno and the staff of the museum. The archaeology of the
Sado valley remained unknown until the 1980s, just blooming in the twenty-first
century.
86
Archaeological Material
Figure 2.6. Excavations at the Sado middens were among the earliest sites where the
Wheeler method was applied in Portugal. Above: sketch with detail of excavation area at
Cabeço do Pez, 1958. Below: field photograph at Arapouco, 1961. (MNA, 1958, D. Sousa;
MNA, APMH, 1961, F644)
87
Chapter 2
The 1980s–90s: old collections, physical anthropology, and selected
fieldwork
Since the identification and excavation of the shell midden sites in the Tagus and
Sado valleys in the late nineteenth century and in the mid twentieth century respectively, the archaeological research was focused on excavation of artefacts and
human remains. The researchers in charge of fieldwork flooded their institutions
and museums with excavated material: hundreds of human skeletons, thousands of
lithic artefacts, and several kilograms of shells and sediment. With a few
remarkable exceptions, most of this material remained unstudied for more than
one hundred years, stored in the archives of the Geological Museum in Lisbon
(MG), the Institute of Anthropology, University of Porto (now MHNUP), and the
National Museum of Archaeology in Lisbon (MNA). Fortunately there was a shift
in focus in the early 1980s, and research was targeted towards detailed studies of
the historical collections along with selected fieldwork.
In 1981–82, J. M. Arnaud developed an interdisciplinary project focused on the
last hunter-gatherers of the Sado valley (Arnaud 1986, 1989), based on the analysis
of the collections at MNA and new fieldwork consisting of survey, relocation of
sites, excavation and test pits (Arnaud 2000; Arnaud and Larsson 1994). Arnaud’s
(1985, 1987, 1989, 1990, 1994, 2000, 2002) greatest contribution to the history of
research of the Sado shell middens lies in his extraordinary capacity for archaeological synthesis, and his work remains the central reference for the study of the
Sado collections, in particular the pioneer paper outlining the model for the
settlement pattern in the region, presented at the European Mesolithic Conference
in Edinburgh in 1985 (Arnaud 1989).
It is also in the early 1980s that a group of Canadian researchers, M. Jackes, D.
Lubell, and C. Meiklejohn addressed the historical collections of the Muge middens and presented pioneer studies strongly based on detailed osteological analyses
(Jackes 1988; Jackes and Lubell 1995, 1999; Jackes et al. 1997a–b; Lubell and
Jackes 1985, 1988; Lubell et al. 1989, 1994; Meiklejohn et al. 1986). At this time the
questions of race and origins, so in fashion in the 1930s to 1940s, had become
completely irrelevant, and the team worked mainly on questions related to the
Mesolithic-Neolithic transition. This research group presented the first
radiocarbon dates on human bone collagen, as well as the first isotopic analysis, on
bones from Portuguese sites (Lubell and Jackes 1985, 1988; Meiklejohn et al.
1986). Throughout the years of their research, Jackes and colleagues have produced a significant body of work dedicated to the Muge sites, not only at the level
of the osteological collections, but also on the laborious task of reconstructing
spatial organization and stratigraphic history of the sites, based on detailed study of
the human remains and historical documentation (Jackes, forthcoming; Jackes and
Alvim 2006; Jackes et al. 2014, 2015, forthcoming; Roksandic and Jackes 2014).
Finally, in the 1990s a team of anthropologists examined the Sado osteological
collection as a whole, and published it for the first time (Cunha and Umbelino
1995–97, 2001; Cunha et al. 2002, 2003). B. Kaufmann from the University of
88
Archaeological Material
Basel studied the material in the 1980s but his results were never published
(Umbelino 2006, 139). From E. Cunha and C. Umbelino’s research we now have
access to site estimates of MNI, demographic profiles (age at death, sex), morphological information (height, robustness), as well as elements of health and
pathology. Along with this pioneering anthropological study this team provided the
first radiocarbon dates on human bone collagen from the Sado valley (Cunha and
Umbelino 2001).
The last decades of the twentieth century are marked by other ground-breaking
studies on zooarchaeology of the middens (Lentacker 1986, 1994; Rowley-Conwy
2004), as well as by the first comprehensive studies on Mesolithic lithic technology,
this time focused on the Sado collections (Araújo 1995–97; Marchand 2001).
Maturity in the twenty-first century: Sado valley out of the archives
Archaeological research on the shell middens of the Tagus and Sado valleys gained
new impetus in the twenty-first century, particularly at the level of the Sado valley
shell middens. New interdisciplinary fieldwork started and is ongoing at both
valleys (Arias et al. 2015; Bicho et al. 2011, 2013; Diniz and Arias 2012; Rolão et al.
2006), along with ever more sophisticated studies of the museum collections of
fauna (Detry 2003, 2007; Marques-Gabriel 2015), lithic technology (Diniz and
Nukushina 2014; Nukushina 2012), and history of research (Abrunhosa 2012).
The human remains continued to be a central element of research. C. Umbelino
expanded her pioneering study on Mesolithic and Neolithic osteological collections
with an in-depth and remarkable study on palaeodiets based on trace elements and
isotopic analysis (Umbelino 2006; Umbelino et al. 2007). Other in-depth isotopic
analyses of carbon and nitrogen followed, this time exclusively dedicated to the
Sado collections (Fontanals-Coll et al. 2014; Guiry et al. 2015). Initial DNA
analyses have been done on samples from Muge (Bamforth et al. 2003) and Sado
(Chandler et al. 2005), and were compared with findings from Neolithic populations. While the Muge assemblage suggested genetic continuity between the
Mesolithic and Neolithic in Portugal, the study based on the Sado assemblages
seems to indicate discontinuity between the Mesolithic and Neolithic groups.
These are interesting results but without further evidence should be interpreted
with caution, not only because these analyses were done during the early stages of
the development of the method (Matisso-Smith and Horsburgh 2012), but also
because the sample sizes were small, requiring further testing to confirm or reject
these hypotheses.
The research at the shell middens of the Tagus and Sado valleys is strong, and
the following years promise to be exciting in terms of upcoming results from ongoing excavations. Hopefully, and perhaps more importantly, researchers will remain interested in the historical collections in the archives of these, and other
museums, which hold abundant and some of the richest archaeological material we
will ever have.
89
Chapter 2
***
Despite the unusual number of human burials, these sites have been studied almost
exclusively from the perspective of settlements for the living. The archaeological
remains of the dead were first used to determine or reject an African ancestry for
the Portuguese nation. Afterwards, within the framework of processual archaeology, these sites were in the centre of the debate about ecological and economic
issues related to the lifestyle and strategies of the last hunter-gatherers of Europe.
Models of seasonal occupation have been proposed (Arnaud 1989; Marchand
2001; Rowley-Conwy 2015); detailed studies on palaeodiets (Umbelino 2006) and
palaeo-populations focused on the Mesolithic-Neolithic transition have been done
(Jackes and Lubell 1995; Jackes and Meiklejohn 2008), and just very recently brief
syntheses on mortuary practices were published (Jackes and Lubell 2012; Jackes et
al. 2014; Roksandic and Jackes 2014; Roksandic 2006; Umbelino and Cunha 2012).
Although the dead have been the main reason why these sites have been excavated
and internationally recognized, their place in archaeological interpretation has been
largely ignored. The following painting by the French archaeologist H. Breuil
(1949) (fig. 2.7) illustrates this perspective well, making no allusion to mortuary
practices (Peyroteo Stjerna 2015). Interestingly, by this time, more than 250 burials
had been recovered from the Muge shell middens. The painting is accompanied by
a page long text with only a very brief reference to burial practices:
They must have lived in straw or reed huts quite near their heaps of shell-fish, in
which they often buried their dead; the bodies were generally doubled up.
(Breuil 1949, 93)
Figure 2.7. Daily life in a Muge shell midden, Tagus valley, Portugal. (Reproduced from
Breuil 1949, 94–5)
90
Chapter 3
Methods
Methodological choices in this study
Understanding the significance of the mortuary activity at the shell middens of the
Tagus and Sado valleys requires we define how consistent and deliberate these
practices were, who was buried in these places, as well as when and how. The human
remains recovered from these archaeological sites are at the centre of this research
and I relied on three methodological approaches: radiocarbon dating and Bayesian
analysis, stable isotope analysis of carbon and nitrogen, and archaeothanatology.
To answer the question when, in relation to the chronology of the burial activity,
I developed a radiocarbon dating programme, using the radiocarbon measurements on human bone collagen from 15 individuals previously dated (16 measurements) and 30 new dates obtained during the scope of this study. The radiocarbon data, previous and new, was carefully scrutinized for reliability of measurements. The statistical treatment and modelling of the data follows the Bayesian
approach to the interpretation of radiocarbon dates. Each site was analysed
individually and a model for the chronology of the burial activity in each site has
been proposed.
To answer the question who, in order to define the people in each burial place,
I considered previously published isotopic signatures of carbon and nitrogen, and
plotted these with 30 new signatures obtained in the course of this dissertation.
Despite the current limitations to the application of the isotopic method to this
particular data set, discussed further in this chapter, these data can be particularly
informative in terms of mortuary practices, which can now be analysed within a
greater chronological resolution compared to previous studies.
To answer the question how, in relation to the treatment of the dead, I analysed
82 individuals with the method of archaeothanatology. This number represents a
natural population with all age groups and both sexes represented. In several cases,
due to the fragmented nature of the material and unclear documentation, the
analysis was limited to a few elements. This is however a dynamic method, which
with the combination of various observations allows a close reconstruction of the
chaîne opératoire from the human remains recovered in the archaeological context to
the original funerary feature.
In this chapter, I present the central principles of each method in relation to the
investigation of the mortuary activity based on the analysis of the human remains
91
Chapter 3
in archaeological context. The first section is dedicated to the radiocarbon dating
method and Bayesian analysis, slightly overlapping in terms of central principles
with the second section which is focused on stable isotope analysis of carbon and
nitrogen. The third and last section of this chapter is centred on the method of
archaeothanatology where I also introduce and discuss relevant terminology.
92
Radiocarbon dating method and Bayesian analysis
Introduction
Chronological resolution in prehistory
Time in prehistoric sites is frequently defined by a couple of absolute dates and a
significant amount of typology. Typological relative dating of lithic and ceramic
assemblages may provide a satisfactory broad range chronological framework, and
in many cases absolute dating methods are frequently used simply to support the
typological argument with a suitable calendar date. This approach to the prehistoric
time stems from an enquiry concerned with a generic chronological framework;
however, the large narrative over a long time period is not the only possible approach to prehistoric sites, as it in fact may be not only misleading but wrong
(Bayliss et al. 2007; Whittle et al. 2011). Also, as good practice, the archaeologist
should always aim for better and more precise chronologies, because time is our
trade.
Chronology is the cornerstone of any archaeological interpretation, and the
consistent use of radiocarbon data combined with accurate stratigraphic data in
prehistoric contexts allows us to refine our chronologies. Radiocarbon data can be
used explicitly because it offers the best probabilistic estimates “of real dates when
things happened by the agency of particular people in specific places” (Bayliss et al.
2007, 1). This is important because the construction of refined chronologies brings
into focus the social context and actions of people (Whittle et al. 2011). A calendrical time scale can provide solid evidence for when an activity occurred, and for how
long, allowing more meaningful comparisons, beyond the stratigraphic and typological series (Bayliss et al. 2007).
Our enquiry into chronology will influence the scale of our archaeological
narrative. In this dissertation, first, I focus on the short term scale: on the human
burial as an event. Second, I focus on the short term event in relation to the long
term scale: the duration of the burial activity in each site. With this chronological
enquiry my aim is to model formally the specific parameters of a particular problem: the use of each site as a burial ground.
Enquiry-dependent chronologies: the radiocarbon dating programme in this
study
A radiocarbon dating programme was developed in the course of this study to
determine the chronology of the burial activity in each site and in each valley. The
chronological models proposed are thus dependent on this interpretative and
research framework.
93
Chapter 3
The shell midden sites of the Tagus and Sado valleys are attributed to the Late
Mesolithic which in Portugal ranges from c 6350–5250 cal BCE (Araújo 2015)
slightly overlapping with the earliest phase of the Mesolithic-Neolithic transition
c 5500–5300 (Carvalho 2009, 2010). This chronological framework was established
on typological analysis (Araújo 1995–97; Breuil and Zbyszewski 1947; Marchand
2001; Nukushina 2012; Roche 1967, 1972) and on several radiocarbon dates
(Arnaud 1989, 2000; Bicho et al. 2011, 2013; Cunha and Cardoso 2001, 2002–03;
Cunha and Umbelino 2001; Cunha et al. 2003; Delibrias and Roche 1965; Detry
and Cardoso 2010; Jackes et al. 2013, 2014; Lubell and Jackes 1985, 1988; Larsson
1996; Martins et al. 2008; Meiklejohn et al. 1986, 2009; Roksandic 2006; Roche
1957; Soares and Cabral 1984; Umbelino 2006).
This study examines the use of the shell midden sites in the Tagus and Sado valleys as burial grounds. One of the main problems at the beginning of this research
was the lack of direct dates on the human remains excavated from these sites, as
well as the unclear association between the radiocarbon dates on non-human material and the burial activity. This was particularly problematic at the sites of the Sado
valley where only four individuals from four sites had direct dates (table 3.1), all
four with problematic calibrations (see below, Reliability of the isotopic measurements).
The situation was better in the case of the Tagus valley where 20 individuals from
four sites were dated, six of which from recent excavations (table 3.2).
Table 3.1. Sado valley: human remains with radiocarbon dates published before 2013.
Site
Arapouco
Poças de S. Bento
Cabeço das Amoreiras
Cabeço do Pez
Individual,
year of excavation
2A, 1962
Skull 1986
5, 1958
4, 1956
Table 3.2. Tagus valley: human remains with radiocarbon dates published before 2013.
Site
Moita do Sebastião
Cabeço da Amoreira
Cabeço da Arruda
Cova da Onça
Individual,
year of excavation
CT, 19th c.
7, 1937
A, 19th c.
1 individual
(without no.)
22,
19th
c.
CAM-00-01
D,
24, 19th c.
CAM-01-01
N, 19th c.
29, 19th c.
Burial 2011.1
III, 19th c.
41, 19th c.
Burial 2011.2
6, 19th c.
16, 1952–54
19th
c.
42, 19th c.
CA–00–01
CA–00–02
94
Methods: Radiocarbon dating and Bayesian analysis
The radiocarbon dating programme was centred on the seven sites under study:
two sites in the Tagus valley, Moita do Sebastião and Cabeço da Arruda, and five
sites in the Sado valley, Arapouco, Cabeço das Amoreiras, Vale de Romeiras,
Cabeço do Pez, and Várzea da Mó. The objectives of the dating programme were
site-specific and are described in detail for each site in chapter 4. The general aim
was to establish a chronology for the use of the sites as burial grounds, which can
be outlined by specific objectives as follows:
• to define the duration of the human burial activity, in each valley in
general and in each site in particular;
• to estimate the frequency of the burial activity, in each valley in general
and in each site in particular;
• to estimate when the burial activity started and when it ended, in each
valley in general and in each site in particular;
• to identify synchronic and diachronic similarities and differences in
burial patterns.
The dating programme was planned in two phases. The first phase consisted of the
selection of samples organized around site-specific and general questions. The
second phase of sampling selection was centred on particular questions generated
by the results of the first phase.
This section
Structure
The section Radiocarbon dating method and Bayesian analysis concentrates on methodological issues related to the main principles of the radiocarbon method and the
interpretation of radiocarbon measurements. This methodological review was
written from the perspective of the analysis of unburnt prehistoric human bone in
Holocene archaeological contexts in the temperate zone of the northern
hemisphere, concerning south-western Europe in particular, and for this reason,
issues relevant to other archaeological contexts may not have been addressed.
Methodological choices for the selection and screening of samples in this study are
presented in the last part of this section.
The results of the radiocarbon dating programme, the analysis and interpretation of the absolute dating evidence, and the formal chronological models are
presented by archaeological site in chapter 4. The results and the chronological
models are analysed and contextualized in chapter 5.
Terminology and conventions
Radiocarbon determinations and calibrated dates are reported following the revised
conventions published in Radiocarbon (Millard 2014). Accordingly, the following
elements are considered and included along with the presentation of the data:
• The laboratory code for each determination.
95
Chapter 3
•
•
•
•
•
•
•
•
•
•
Identification of the sampled material. Biological materials should be
identified to genus and species when possible.
The laboratory pre-treatment method applied and the quality control
measurements.
It should be clear whether the δ13C values were obtained by accelerator
mass spectrometry (AMS) during the process of 14C measurement or
independently by isotope-ratio mass spectrometry (IRMS). This is important because the AMS values should not be used for dietary reconstructions or reservoir corrections (Millard 2014, 557; Taylor and
Bar-Yosef 2014, 117).
The laboratory measurement is reported as a conventional radiocarbon
age in radiocarbon years BP (14C yrs BP).
The calibration curve(s) used and respective version number(s).
The reservoir offset used, when applicable.
The software used for calibration and respective version number.
The calibrated date is given as a range (or ranges).
The associated probability, such as 68% or 95%, is given for each range.
The terms “1-sigma” (1σ) and “2-sigma” (2σ) should not be used to
describe calibrated dates (Millard 2014, 557).
The calendar timescale used is clearly identified (e.g., cal BP, cal BCE).
These conventions are the recommended minimum requirements for good
practice (Millard 2014). The advantages of following these standards are that it
allows the replication of results and ensures that the published data can be used by
other researchers as methods and techniques improve.
In this dissertation, calibrated/calendar ages are primarily reported as
“cal BCE” (Before Common Era). This option was based on the history of
research of these sites, which have been studied from a Mesolithic-Neolithic perspective, and where dates are typically reported as cal BCE. Additionally, the radiocarbon tables may present a second calendar timescale, “cal BP” (Before Present,
where Present is 1950 CE), for easier comparison with other sites and research
traditions where cal BP is conventionally used.
All radiocarbon measurements presented in this study are calibrated with OxCal
4.2 (Bronk Ramsey 2009) using atmospheric curve IntCal13 (Reimer et al. 2013).
Samples where carbon is derived from atmospheric and aquatic reservoirs are calibrated using the atmospheric curve IntCal13, the aquatic curve Marine13 (Reimer
et al. 2013) and the regional reservoir offset. In each sample, the percentage of
non-atmospheric carbon derived from aquatic reservoirs is estimated using two
methods clearly identified as marine 1 and marine 2. The calibrated ranges quoted in
italics derive from mathematical modelling and are posterior density estimates. The
presentation of the results for each site adopts in general the structure presented by
A. Whittle and colleagues (2011).
96
Methods: Radiocarbon dating and Bayesian analysis
Radiocarbon dating method
The principles of radiocarbon dating were established after the Second World War
(Libby et al. 1949) following the development of atomic physics and rapidly tested
and applied to archaeological samples of known historical date (Arnold and Libby
1949). This major scientific innovation transformed archaeology (Trigger 2007,
382) and was the most important breakthrough in the study of prehistory since the
scientific demonstration of the antiquity of man in the mid 1800s (Renfrew 2008,
37). The first set of samples (Arnold and Libby 1951), submitted from sites around
the world, included classic prehistoric sites. A charcoal sample came from the
Palaeolithic cave Lascaux in France and a sample from a wooden platform from
the Mesolithic open air site Star Carr in England. Later, in 1960, W. Libby
(1908–80) was awarded the Nobel Prize in Chemistry for his method of age
determination:
Professor Willard Libby has been selected to be the prize-winner for his method of
age determination of materials of biological origin by use of carbon-14 as a
measurer of time. His method has obtained widespread use and has become indispensable in archaeology, geology, geophysics and other sciences. Fortunately, it is so
simple – which is probably not always the case with chemical research distinguished
with the Nobel Prize – that everyone should be able to understand the conditions
and principles for its execution. (Westgren 1960)
Radiocarbon dating is the most useful, reliable and widely used scientific method
for determining the age of a carbon-containing substance formed over the last fifty
thousand years (Bronk Ramsey 2008; Bronk Ramsey et al. 2006; Reimer et al.
2013).
The method works from the observation that the chemical element, radiocarbon, is absorbed and fixed into organisms during their lifespans, and decays at a
constant rate over time after their biological death (Arnold and Libby 1949; Libby
et al. 1949). The decay rate can be counted and an age can be determined for when
the organism died, that is, when the organism stopped absorbing radiocarbon from
the environment. For the estimation, interpretation and critique of radiocarbon
measurements it is important to understand the basic principles of the method (see
Bronk Ramsey 2008; Taylor and Bar-Yosef 2014). As discussed further, the
archaeological material under study in this dissertation offers a set of complexities
which are difficult to analyse and interpret if the radiocarbon dating methodology
is not well understood.
Basic principles
Radiocarbon (14C) is the radioactive isotope of the element carbon, while 12C and
13C are its stable isotopes (non-radioactive) (Pollard et al. 2007). Carbon and its
isotopes enter the biosphere mainly through the process of photosynthesis in
plants, algae, and cyanobacteria, and enter the food chain directly via herbivore
animals and indirectly in the case of carnivores and omnivores. In reality, the 14C
97
Chapter 3
method dates the photosynthesis event when 14C enters the food chain, however,
“terrestrial food chains are also sufficiently short that in most instances the carbon
in any animal’s diet will have been fixed from the atmosphere within the past few
years” (Bronk Ramsey 2008, 253). Thus carbon is incorporated in living organisms
from the food they eat. Animals will use incorporated carbon in the production of
structural proteins such as keratin (present in hair, horns, nails, claws, hooves,
feathers, and beaks), collagen (present in bone and tooth dentine), and also in the
formation of exoskeletons (e.g., reptile shells, shells in snails, bivalves and other
molluscans). This is useful for archaeological research because, once formed, these
structural proteins retain the isotopic record (14C, 13C, 12C, among other elements)
of that moment in time, and many of these structures survive relatively well in
archaeological contexts.
From the point of view of 14C dating, the next stage in the cycle of incorporated
carbon that is important to keep in mind is that, over time unstable atoms such as
14C go through a process of radioactive decay, also known as nuclear decay or
radioactivity, and will fall off at a regular and known exponential rate independently of all environmental conditions (Bronk Ramsey 2008; Pollard et al. 2007).
When the organism dies the uptake of 14C ceases and the steady concentration of
14C in the organism begins to decline. The age of the dead organism can thus be
calculated by measuring the amount of 14C left in the sample, with reference to the
known rate of 14C decay. The rate of decay used in the laboratories for the calculation of 14C ages is the so-called Libby’s half-life of 5568 ± 30 years (Anderson and
Libby 1951) and it is the conventional value in use since 1965 until the present
(Bronk Ramsey 2008, 254; Stuiver and Suess 1966, 536).
In the context of the application of the 14C dating method to archaeological
samples, one must acknowledge that the production of 14C is not constant through
time (Anderson and Libby 1951). However this fluctuation has been examined and
continuously studied for different timescales following the reference work of M.
Stuiver and H. Suess (1966) (see Reimer et al. 2013). Furthermore, research has
shown that the fluctuation of 14C in the atmosphere is different between the
northern and southern hemispheres, a variation also known as the hemispheric offset
(Hogg et al. 2002, 2013; McCormac et al. 1998, 2004). Yet the fluctuation within
each hemisphere does not seem to vary locally or regionally, at least significantly
(Bronk Ramsey 2008, 251–252).
Radiocarbon produced in the atmosphere is in dynamic equilibrium with that
on the surface of the hydrosphere (Bronk Ramsey 2008, 253). The concentration
of 14C in waterbodies is generally lower than that in the contemporary atmosphere
due to a slower mixing rate between the chemical elements in the atmosphere and
the water surface (Bronk Ramsey 2008; Keaveney and Reimer 2012; Stuvier and
Polach 1977; Stuvier et al. 1986). Because of this relative depletion in 14C, the organisms grown in aquatic reservoirs, either in marine, brackish or freshwater
habitats, will variably present relatively older 14C ages. This offset is a phenomenon
known as the reservoir effect (ΔR) and is discussed further below (Stuvier and Polach
1977; Stuvier et al. 1986).
98
Methods: Radiocarbon dating and Bayesian analysis
Unlike the concentration of 14C in the atmosphere, the concentration of 14C in
aquatic ecosystems has great geographical variation (see CHRONO 2015). Various
factors, such as the upwelling phenomenon in marine environments (Stuvier and
Braziunas 1993; Monge Soares 1993; Monge Soares and Dias 2006) or the local
multi-sources of carbon in freshwater systems (Keaveney and Reimer 2012;
Lanting and van der Plicht 1998) may contribute greatly to these variations. When
applying the radiocarbon dating method to archaeological samples, it is fundamental to identify the sources of 14C for each archaeological sample, because in the case
of animal samples, their foods may derive from a greater or lesser variety of
sources of different primary reservoirs.
In summary, in the context of the application of the 14C dating method to
archaeological samples, the key points to keep in mind about the carbon cycle and
14C in the environment are:
• The production of 14C is not constant through time. This fluctuation
has been examined and continuously studied for different timescales.
• Food sources from different primary reservoirs, terrestrial or from the
various aquatic systems, will determine the 14C incorporated in each
archaeological sample.
• Once formed in animal tissue, structural proteins retain the isotopic
record (14C, 13C, 12C, among other elements) of that moment in time,
and many of these structures survive relatively well in the archaeological
record.
• Once the organism dies, the uptake of 14C ceases and the steady
concentration of 14C begins to decline (i.e., decay) in a steady exponential rate, independently from all environmental conditions.
The 14C age obtained from the laboratory is a measurement of the isotope ratios
and 13C/12C converted to a radiocarbon “date”, based on the assumption
of constant 14C concentration equal to that of the atmosphere in 1950 and a decay
rate of 14C based on Libby’s half-life reference (Bronk Ramsey 2008, 260).
Following this, it is important to acknowledge that the laboratory measurement
reflects the isotopic composition of the sample and not its “true” age. The only
way to convert it to a calendar age in absolute terms is through the process of
calibration (Bronk Ramsey 2008, 261–262; 2009, 341).
14C/12C
Interpretation of 14C measurements
When used for chronological purposes, 14C measurements must be interpreted on
the basis of the current knowledge of the past environment. In this process, two
main aspects have to be taken into consideration: the carbon fluctuation through
time and the place of the organism (sample) in the environment (Bronk Ramsey
2008, 260). This conversion from isotopic measurement to calendar age is known
as calibration and provides a statistical estimate for the “true” age of the sample. For
this reason, once the calibration process is resolved, the 14C dates should go
99
Chapter 3
through statistical treatment in order to construct the best possible chronological
model based on the currently available evidence.
Carbon fluctuation through time: calibration curves
It is known that 14C ages do not directly express calendar years. As observed since
the early stages of the method, the concentration and production of carbon in the
atmosphere fluctuates through time (Anderson and Libby 1951; De Vries 1958;
Stuiver and Suess 1966).
IntCal13 (Reimer et al. 2013) is the current internationally-agreed calibration
curve for the mid-latitude northern hemisphere atmospheric reservoirs and can be
applied from 0 to 50,000 cal BP. The calibration curve Marine13 (Reimer et al.
2013) is used for samples derived from any marine reservoir, from the northern or
southern hemisphere, but must be applied with the available data for regional
oceanic variations (see below, Reservoir offsets). These curves replace IntCal09 and
Marine09. With some minor changes, the calibration curve IntCal13 is similar to
IntCal09 from 0–12,000 cal BP, but it is slightly different between 12,000–13,900
cal BP, and it presents greater updates for the older portion of the curve between
14,000–50,000 cal BP (Reimer et al. 2013). The data available from 0 to 12,000
cal BP, and to some extent between 12,000–13,900 cal BP, is fairly robust, and
later updates have not changed this portion of the calibration curve substantially
(Bronk Ramsey et al. 2006; Reimer et al. 2004, 2009, 2013). Thus, the European
Mesolithic chronologies can be calibrated with a high degree of confidence and
resolution, and it is unlikely that the continuous revisions of the curve will affect
these significantly.
The International Calibration (IntCal) curves are produced by the IntCal Working Group (IWG) and are recommended and ratified by the International Radiocarbon Conference. These provide the most comprehensive, up-to-date consensual
knowledge of the past variation in 14C (Bronk Ramsey et al. 2006; Reimer et al.
2009). Alternative curves are available and may be used, however, it is imperative
to explicitly state which curve or data set has been used and to present the original
uncalibrated 14C data. In this study, the calibration curves used are IntCal13 and
Marine13 (Reimer et al. 2013).
The place of the sampled organism in the environment
Reservoir offsets (ΔR)
As discussed, when applying the 14C dating method to archaeological samples it is
fundamental to know the sources of 14C for each archaeological sample. Radiocarbon data must be always interpreted in the environment context and we must identify the reservoir source(s) of the carbon present in our sample (Bronk Ramsey
2008).
A marine calibration curve must be applied in the case of organisms grown in
aquatic environments or samples which indirectly, via the food chain, incorporate
carbon fixed in non-terrestrial reservoirs. Such samples are depleted in 14C and will
100
Methods: Radiocarbon dating and Bayesian analysis
present an offset in 14C age in relation to coeval material containing atmospheric
carbon exclusively. Thus, if these samples are calibrated with the terrestrial curve
only the dates of these specimens will appear older than their real calendar date.
The Marine13 curve is the best current estimate to correct these 14C ages, but
because of the large variability of marine reservoir corrections in some regions it
must be applied together with the local marine reservoir offset (ΔR).
A number of factors related to time and location contribute to the high
variability of the ΔR parameter (Lanting and van der Plicht 1998). The ΔR varies
with the size of the carbon reservoirs, and shallow lakes for example may present
identical values to the atmosphere (Stuiver et al. 1986). The ΔR can also change
over time depending on local factors such as latitude, sedimentary basins,
upwelling regions and other oceanographic factors (Reimer et al. 2009, 1114).
Reservoir offsets in this study
The Western Atlantic coast of the Iberian Peninsula is influenced by a dynamic
upwelling which strongly affects the fluctuations in the ocean reservoir offset on
the western coast of Portugal during the Holocene (Monge Soares and Dias 2006).
This is a marine phenomenon that occurs in the open ocean and along coastlines,
when deep water rises and replaces the surface water. The upwelling activity varies
through time and research has shown that coastal values of ΔR in these regions are
not constant as previously suggested (Stuiver et al. 1986) and may exhibit significant variability (Monge Soares 1993; Monge Soares and Dias 2006; Monge Soares
and Martins 2009; Stuiver and Braziunas 1993).
It is estimated that throughout the Holocene, the ΔR values on the Portuguese
coast varied between 940 ± 50 to –160 ± 40 14C years (Monge Soares and Dias
2006). Wind-driven coastal upwelling is influenced by climate conditions and it has
been suggested that positive high ΔR may be correlated to a strong upwelling as
opposed to weak or even nonexistent upwelling denoted by low or negative ΔR
values (Monge Soares and Dias 2006; Monge Soares and Martins 2009). Along a
dynamic coast such as the Portuguese coast, the determination of a mean ΔR is
meaningless and the archaeologist must use the determined ΔR value that is closer
to the 14C age to be calibrated (Monge Soares and Martins 2009, 2881), which for
this study corresponds to the time frame c 7000 BP. Palaeo-estuarine environments such as the Tagus and Sado are also affected by the depletion of 14C relative
to the concentration in the atmosphere, and the appropriate reservoir correction
must be applied (table 3.3). As good practice, all adopted reservoir corrections
must be clearly stated and referenced (Millard 2014; Stuiver and Polach 1977).
101
Chapter 3
Table 3.3. Reference record of reservoir ages for the Portuguese coast and estuaries. *Originally published
as a positive value due to a typographical error (Monge Soares pers. comm.).
Region in Portugal
Time frame
ΔR
weighted mean in 14C yrs
Reference
West coast
Modern
250 ± 25
Monge Soares and Dias 2006
West Coast of Estremadura,
Alentejo and Algarve regions
c 7000 BP
95 ± 15
Monge Soares pers. comm.
Tagus valley (Muge, Magos)
c 7000 BP
140 ± 40
Martins et al. 2008
Sado valley
c 7000 BP
–100 ± 155*
Martins et al. 2008
Dietary sources and the calibration of samples of human bone
Collagen is the main organic component in bone and one of the most common
molecular forms of carbon used in 14C dating. In archaeological contexts, the
collagen in bones may be preserved for thousands of years, “particularly in cool
and stable environments, and is extremely resistant to post-mortem diagenetic
alteration of stable isotope ratios” (Ambrose 1993, 72). Bone collagen is formed
primarily by the protein consumed by the organism. Consequently, the collagen
component in the bone of a specimen will reflect the total carbon and nitrogen
ingested by the organism (Ambrose and Norr 1993), so when we are dating a
skeleton, what is being measured is the radiocarbon age of this dietary protein
(Bayliss et al. 2004, 568). To interpret the 14C measurements of a sample of bone
collagen two main aspects have to be taken into consideration: the bone collagen
turnover and, most importantly, the sources of carbon which formed the sampled
collagen.
The bone collagen turnover is a slow process. The rate of this turnover is not
well-known, but it is estimated that complete replacement may take 10–30 years
(Ambrose 1993) depending on growth and ageing of the individual. Experimental
work has consistently shown that bone collagen reflects the diet of a human adult
over a much longer period of time than 10 years (Hedges et al. 2007).
As discussed, the sources of carbon which formed the bone collagen in the
sample can be varied. This can be a particularly complex task when analysing
samples of human bone. Humans, as other omnivores, may have a very diverse
diet and obtain protein foods grown in different environments. Because terrestrial
and aquatic reservoirs contain different concentrations of carbon at a given moment in time, it is essential to know where the individual derives its carbon, i.e., the
protein component of its diet. In this process, the reservoirs of carbon must be
identified and the proportions of consumed carbon, if derived from more than one
reservoir source, must be estimated (Bronk Ramsey 2008, 260). The measurements
of the stable isotope ratios of carbon (13C) and nitrogen (15N) on bone collagen can
be used to some extent to identify and estimate the sources of protein consumed
by the dated individuals (Ambrose 1993; Richards and Hedges 1999). This is a
fundamental step for the correct interpretation of 14C measurements because “any
102
Methods: Radiocarbon dating and Bayesian analysis
uncertainty in our knowledge of the contribution of non-terrestrial food sources
compromises our ability to get good chronological resolution” (Bronk Ramsey
2008, 260). Even if there is no evidence of non-terrestrial intake, it cannot be ruled
out without chemical evidence (Bronk Ramsey 2008, 261).
There is significant variation in the isotopic composition of food sources in different habitats and climates (Ambrose 1993, 85) and the diet of the individual cannot be inferred solely on the basis of its measured isotopes. One must be careful
with these estimates, because, as S. Ambrose (1993, 100) has pointed out
“translating the delta (δ) values into precise percentages of marine versus terrestrial
resource use would be unwarranted without further information on the isotopic
composition of the local food web”.
Here, for the purposes of calibration of 14C measurements on samples of human bone, I present the methodology used in this study to estimate the proportions of atmospheric and non-atmospheric sources of carbon present in each
sample.
Sources of protein component in human bone collagen: calculation estimate
In this study, for the purpose of correction of radiocarbon measurements, the
marine (i.e., non-atmospheric) fraction of each sample is calculated by the
application of two methods, indicated as method marine 1 and method marine 2 or in a
simplified form, marine 1 and marine 2 (M1, M2). All estimates were calculated by
using the carbon stable isotope values of each sample (δ13C). Both methods
assume a margin of error of ± 10%.
Method marine 1
In this method, the percentages of marine versus terrestrial resources in a diet are
estimated from the δ13C or δ15N value of bone collagen (Ambrose 1993, 83). This
method presented by Ambrose (1993) draws from pioneering work developed in
the 1980s on the reconstruction of palaeodiets from stable isotopes in human
skeletons (Schwarz et al. 1985; White and Schwarz 1989). It determines a value for
the marine fraction needed for the correction of 14C ages, and because it requires a
detailed knowledge of the local ecosystem, it may provide relatively accurate
information about the diet of the individual (fig. 3.1).
% marine (±10) =
measured value
estimated average value
estimated value
δ13C or δ15N
–
–
δ13C or δ15N
δ13C or δ15N
sample
diet-tissue fractionation factor
terrestrial resources
estimated average value
estimated average value
–
δ13C or δ15N
δ13C or δ15N
marine resources
terrestrial resources
× 100
Figure 3.1. Calculation of the percentage of marine resources in a diet, estimated from the
δ13C or δ15N value of bone collagen. (Adapted from Ambrose 1993; Schwarz et al. 1985)
103
Chapter 3
In this equation, the estimated average δ13C value of terrestrial dietary resources is related to
the type of plants, C3 and/or C4, available for consumption. In the case of C3 plant
environments the average δ13C value is –26.5‰ (Ambrose 1993). C4 plants fix
carbon with average values of –12.5‰ (Smith 1972). Likewise, this equation
considers an estimated average δ13C value of marine resources. This average value will
depend on the adopted endpoint values, which in this study are –12‰ to –20‰,
with an average value of –16‰ (see section, Stable isotope analysis of carbon and nitrogen). Finally, this equation also takes into account the isotope enrichment in
collagen relative to that of the diet, which can be quite variable, but where terrestrial resources are of C3 plants, such as in Mesolithic Europe, the estimated value is
+5‰ (Ambrose 1993, 84) – diet-tissue fractionation factor. Accordingly, an individual
with a measured δ13C value of –16.9‰ will present an average of 43.8 ± 10 % of
marine protein:
% marine (±10) =
–16.9‰ – –26.5‰ – 5‰
–16.0‰ – –26.5‰
× 100 = 43.8
Method marine 2
In this method, the percentages of marine versus terrestrial resources are estimated
from the measured δ13C value of bone collagen by the application of a simple
linear interpolation equation (fig. 3.2) between the adopted endpoint values
(Arneborg et al. 1999; Richards and Hedges 1999) (figs. 3.3, 3.4). Like the previous
method, it determines an adequate value for the marine fraction needed for the
correction of 14C ages, however, it is less detailed in the assessment of the
individual’s place in the environment thus potentially less accurate regarding the
diet of the individual. It is a practical method when the aim is the calibration of 14C
ages, particularly when details of the local environment are not available.
y-y0
y1-x0
=
⇔ y = y0 + y1-y0
x-x0
x1-x0
x-x0
x1-x0
Figure 3.2. Mathematical equation: linear interpolation between two known points. The
equation is solved for y which is the unknown value at x (Meijering 2002).
104
Methods: Radiocarbon dating and Bayesian analysis
100
90
80
% marine (y)
70
60
50
40
30
20
10
0
terrestrial endpoint
marine endpoint
δ13C (x)
Figure 3.3. Graphic representation of the calculation of percentage of marine foods in a diet.
Percentage is estimated from the δ13C value of bone collagen applying linear interpolation
between two known points. The terrestrial endpoint (x0) corresponds to 0% marine diet
(y0), and the marine endpoint (x1) corresponds to 100% marine diet (y1).
% marine (±10) =
estimated value
measured value
δ13C
δ13C
–
sample
terrestrial endpoint
estimated value
estimated value
δ13C
δ13C
–
marine endpoint
terrestrial endpoint
× 100
Figure 3.4. Application of linear interpolation equation for the calculation of the percentage
of marine foods in a diet. Percentage is estimated from the δ13C value of bone collagen.
Accordingly, an individual with a measured δ13C value of –16.9‰ will present an
average of 38.8 ± 10 % of marine protein:
% marine (±10) =
–16.9 – –20.0
–12.0 – –20.0
× 100 = 38.8
The linear method using δ13C values (method marine 2) has been criticized because
while the carbon values help to identify a marine diet, it is unlikely to identify a
freshwater diet, the impact of which on the real age of the sample may be significant (Cook et al. 2001). This critique can be extended to method marine 1, and
when using any of these methods the significance of the δ15N values of each
105
Chapter 3
sample must be evaluated. A more reliable approach for environments where the
consumption of freshwater fish may be high – based on archaeological data,
proximity to a rich riverine biome and elevated δ15N/low δ13C – would be to apply
the linear interpolation using δ15N values (Cook et al. 2001, 458). In this dissertation, while every δ15N value was carefully evaluated, the estimate of aquatic food
intake was based on δ13C values, because of the uncertainties in the knowledge
about δ15N trophic level variation in the regions under study, and the lack of clear
indications of heavy reliance on freshwater fish (estuarine environments with
influence of marine waters).
In summary, all 14C measurements need to go through the process of calibration
to convert 14C ages into 14C dates (i.e., calendar dates):
• In every case, an up-to-date atmospheric calibration curve must be used,
such as IntCal13.
• In the case of organisms such as humans, that may derive their carbon
from terrestrial and/or non-terrestrial reservoirs, an estimate of the terrestrial versus aquatic protein component of the individual’s diet must
be calculated, in accordance with the original living environment.
• If it is shown that the protein intake is of mixed origin, terrestrial and
aquatic, an up-to-date marine calibration curve must be used, such as
Marine13, and applied according to the calculated marine percentage.
• As discussed, the concentration of carbon in the atmosphere varies
through time and between hemispheres. However, the concentration of
carbon in aquatic habitats varies not only though time but regionally as
well, depending on various local factors. Thus, the marine curve must
be applied together with the local or regional reservoir offset (ΔR) of
the time period under study.
Statistical treatment and modelling
Why modelling?
As discussed, the concentration of 14C in the different primary reservoirs varies
over time and for this reason, all 14C measurements must go through the process
of calibration. The outcome of the calibration process is always given by a
probability density function for the age of the sample (fig. 3.5) which should be
handled with appropriate statistical methods in order to overcome the uncertainties
introduced by the calibration process and improve the calendar precision of each
14C date (Bronk Ramsey 2008).
106
Methods: Radiocarbon dating and Bayesian analysis
Figure 3.5. Example of a probability density function. Calculated by OxCal 4.2 (Bronk Ramsey 2009) using IntCal13 curve (Reimer et al. 2013) for the calibration of a fictitious 14C age
of 7000 ± 50 BP.
The visual inspection of probability distributions is a common method for the
interpretation of 14C dates. It consists of the informal interpretation of graphs or
tables of calibrated 14C dates by taking their widest limits for the start and end of
the dated activity (Bayliss et al. 2007). However, calibrated 14C dates are probabilistic and the visual method does not take into account the normal statistical scatter
on the 14C dates. As emphasized by A. Bayliss and colleagues (2007, 5), in any set
of dates of an archaeological phase part of the “probability distributions of the
calibrated radiocarbon dates will lie outside – earlier or later than – the actual
calendar span of that phase. If this scatter is not taken into account, then it will
appear that the archaeological activity started earlier, finished later and continued
for longer than was actually the case”. The visual approach can be significantly
misleading because it overestimates the spread of the events, particularly when
looking at a series of calibrated 14C dates (Bayliss et al. 2007, 9; Bayliss 2009, 131;
Bronk Ramsey 2009, 339).
Bayesian analysis is a statistical method increasingly applied to the analysis and
interpretation of calibrated 14C dates and its application has been demonstrated to
be very effective in archaeological research (Bronk Ramsey 2008, 265). Bayesian
statistics allows us to combine the calibrated 14C ages with the appropriate archaeological (prior) information with greater precision and to find the most probable
solution (Bayliss et al. 2007). This approach, known as Bayesian chronological modelling
or Bayesian statistical modelling, accounts for the inevitable statistical scatter of the 14C
dates around the actual values, and establishes a statistical distribution of the group
of 14C dates giving the best range of possible fits and likelihoods with probabilities
107
Chapter 3
(Bayliss 2009, 131; Bronk Ramsey 2008, 264; 2009). For this reason, Bayesian statistical modelling is a significant step towards a more exact absolute chronology
because it can improve the chronological resolution of a given set of 14C dates.
The Bayesian approach
Bayesian statistics were introduced to archaeology by a series of papers published
in the early 1990s (Buck et al. 1991, 1992, 1994a–b) and have been successfully
implemented for the treatment of 14C data in archaeological contexts (Bayliss 2009;
Bayliss and Bronk Ramsey 2004; Bayliss et al. 2007, 2011; Bronk Ramsey 2009).
The strength of any statistical method for the interpretation of archaeological
problems comes from a combined approach of statistical methods with the explicit
use of archaeological data. Without this explicit correlation models may bias data in
ways that were not intended (Bronk Ramsey 2000; Steiner and Rom 2000).
The Bayesian approach to the interpretation of calibrated 14C dates is a statistical process and it requires clear and unequivocal choices, based on contextual
archaeological data, about how to model every single 14C date in the study (Bayliss
et al. 2011, 38) (fig. 3.6).
Figure 3.6. Bayes’s theorem applied to archaeology. (After Bayliss 2009, 127, fig. 3).
The figure presents the original Bayes’s theorem adapted to archaeology and its
application to the interpretation of calibrated 14C dates (Bayliss 2009; Bayliss and
Bronk Ramsey 2004; Bayliss et al. 2007). In archaeological terms, and in the
context of chronological modelling of archaeological data, the standardised likelihoods
express the scientific dating evidence, such as the 14C dates (Bayliss et al. 2007;
Bronk Ramsey 2009). This includes the 14C measurements of the samples and any
calibration data required, such as the calibration curve and reservoir offsets (Bayliss
et al. 2007).
The dates (or, the standardised likelihoods) must be contextualized in previous
archaeological knowledge, here referred to as prior beliefs (Bayliss et al. 2007; Bronk
Ramsey 2009). The prior beliefs (or, the archaeology) can be expressed by relative
dating evidence, such as the stratigraphic sequence of the dated samples. The
stratigraphic data are based on archaeological evidence, known as informative prior
beliefs, and affect the output of the Bayesian chronological model strongly (Bayliss
108
Methods: Radiocarbon dating and Bayesian analysis
et al. 2007, 5, 14) (fig. 3.7). The archaeological information (or, the informative
prior beliefs) must be correct or the resulting chronological estimates suggested by
the model will be wrong. The validity of any chronological modelling will always
depend on the archaeological interpretation of the excavated evidence. In any
model, the informative prior beliefs must be explicitly defined; their reliability must
be critically evaluated and alternative readings of excavation records may be
discussed as well (Bayliss et al. 2007, 14, 22).
Figure 3.7. Example of informative prior beliefs (or, the archaeology) incorporated in a
Bayesian chronological model. The stratigraphic relationships between samples are shown
with the earliest at the bottom. Solid lines indicate relationships determined by excavation;
dashed lines indicate relationships inferred on the basis of the 14C results. (After Meadows et
al. 2007)
The mathematical distribution for the age of the samples is also a prior belief
included in the model and is known as uninformative prior belief (Bayliss et al. 2007, 5).
In contrast to the informative prior beliefs, the mathematical assumptions about
the distributions of dated events (or, the uninformative prior beliefs) “have to be
grossly wrong before they are problematic for the estimation of accurate
chronologies” (Bayliss et al. 2007, 22).
109
Chapter 3
The posterior beliefs (or, an answer) are the outcome of the Bayesian chronological
modelling. These are quantitative estimates of the dates of events based on both
archaeological and statistical data. These dates are expressed as probability
distributions known as posterior density estimates. The outcome of the chronological
modelling may change as new data become available or as the data are modelled
from different perspectives, which must always be strongly rooted in archaeological evidence. Because the posterior beliefs are interpretative estimates, and are not
absolute, the posterior density estimates should be given in italics (Bayliss et al. 2007).
OxCal
The calibration of 14C measurements is normally done with specific developed
software. There is a range of user friendly freeware available, such as BCal (Buck et
al. 1999), CALIB (Stuiver and Reimer 1993), CalPal (Weninger and Jöris 2008), and
OxCal (Bronk Ramsey 2009), among others.
OxCal is one of the most widely used calibration programs. It is versatile, continuously updated and is available in both online and downloadable versions.
OxCal can be used as a simple calibration program or to build rigorous
chronological models of all kinds, from simple to very complex (Bronk Ramsey
1995, 425–6; 1998, 461). The Bayesian statistics for the analysis of 14C dates (Buck
et al. 1992) have been incorporated in OxCal since its first release (Bronk Ramsey
1995). OxCal (Bronk Ramsey 1995, 2009) is the software used for the calibration,
statistical treatment and modelling of all 14C measurements used in this dissertation. In this section, I highlight some of the most relevant functions for the data
set under analysis. For a complete set of analysis operations, models and other
options consult OxCal 4.2 Manual (Bronk Ramsey 2015).
In OxCal, one of the most useful and simplest ways to deal with a set of dates is
to group them in a phase model. This is also the most common way of grouping
dates in OxCal and should not be confused with the archaeological sense of the
term phase. In OxCal, a phase means a group of elements for which there are no
fixed relations (Bronk Ramsey 1995, 426) which imposes no internal constraints
(Bronk Ramsey 1998, 463). The phase model is useful when the relative
chronology of the dated elements is unknown. In OxCal, a phase model can be
expressed in code view (Bronk Ramsey 2015) as the following:
Sequence()
{
Boundary(“Start”);
Phase()
{
R_Date(“A”,8050,25);
R_Date(“B”,7980,30);
R_Date(“C”,8020,25);
R_Date(“D”,8000,25);
};
Boundary(“End”);
};
110
Methods: Radiocarbon dating and Bayesian analysis
The sequence model should be used if the relative chronology of the group of dated
elements is known. This is the second most useful grouping in OxCal and differs
from the phase model because it constrains the group of dated elements within it,
to be in a given chronological order (Bronk Ramsey 1995, 426, 1998, 463), which
must be based on explicit archaeological evidence. In the sequence model it is
assumed that A is older than B which is older than C which is older than D; and it
can be expressed in code view (Bronk Ramsey 2015) as the following:
Sequence()
{
Boundary(“Start”);
{
R_Date(“A”,8050,25);
R_Date(“B”,7980,30);
R_Date(“C”,8020,25);
R_Date(“D”,8000,25);
};
Boundary(“End”);
}
In both phase and sequence models, the 14C dates under analysis are framed by
two undated events: a start event and an end event. In OxCal these events are
termed boundaries (Bronk Ramsey 2000, 199). Boundaries are date estimates which
have not been dated directly by 14C measurements and are calculated by OxCal
from the dated events of the series. Boundaries do not depend on any particular
14C date, but on the whole assemblage of considered dates. Boundary events
should be always included when modelling a discrete group of events. This is
important because, in practice, archaeological events are not truly independent but
related in some way over a limited time span, rather than in an infinite timescale.
The introduction of boundaries in a model is a way to make the events relate
statistically to each other. C. Bronk Ramsey (2001, 357) has made this point very
clear when stating that “models without boundaries will tend to bias to spans
which are too long and so, although for small numbers of events it usually makes
little difference, they should really be included in almost any model”. Often,
however, the use of boundaries in the analysis will have little effect in the outcome
of the model, because the dates are often well resolved and not very high in
number (Bronk Ramsey 2000, 201).
The OxCal program is flexible and each of these phase or sequence groups can
admit many other possibilities, containing other sequences or phases (Bronk
Ramsey 1995, 426, 1998, 463), and more or less complex Bayesian chronological
models can be explicitly defined this way (Bronk Ramsey 2009, 2015).
Once the model is defined in OxCal, the program calculates the probability
distributions of the individual calibrated 14C results (Stuiver and Reimer 1993),
after which it relates the distributions with the prior information (i.e., our prior
archaeological information). This process produces posterior density estimates for
111
Chapter 3
the calendar age of each sample, which typically occupies part of the calibrated
probability distribution (Bayliss et al. 2007, 6) (fig. 3.8).
Figure 3.8. Example of a posterior density estimate which occupies part of the calibrated
probability distribution. Calculated by OxCal 4.2 (Bronk Ramsey 2009) using IntCal13
curve (Reimer et al. 2013) for the calibration of a fictitious 14C age of 7000 ± 50 BP in a
fictitious phase model.
OxCal allows the analysis of phases and sequences, and of more complex relations
such as the definition of time boundaries and limits, such as the probability distribution for the dates of the earliest (terminus post quem) and last events (terminus ante
quem) (Bronk Ramsey 1995, 426). The program can also generate a probability
distribution for the time difference between any two events (Bronk Ramsey 1995,
426, 2015). With OxCal it is also possible to combine a number of 14C dates
available for one same entity into one probability distribution (Bronk Ramsey
1995, 426). The graphical output is another advantage of OxCal, as it shows the
details imposed in the model allowing the reproducibility of the results by others
(Bronk Ramsey 1995, 430).
Reliability and stability of the models
While the complexity of a society cannot be represented by any model, these are
tremendously useful to make sense of reality. For this reason, and for the power of
interpretation they may have, we must make sure that we are not imposing a model
which is inconsistent with the dating evidence. Every model must be checked for
reliability and stability.
112
Methods: Radiocarbon dating and Bayesian analysis
Statistical criteria
Two main statistical criteria produced by OxCal can be used for reliability and
consistency testing both on the model as a whole and on individual elements: the
index of agreement (A) and convergence (Bronk Ramsey 1995). These criteria allow the
user to test and select the most appropriate and reliable chronological model
(Bayliss et al. 2007, 6; Bronk Ramsey 1995, 428). These agreement indices of
OxCal will highlight cases where the prior model is inconsistent with the data
(Bronk Ramsey 2000, 201).
The index of agreement (A) is related to the standardised likelihoods (the dates)
and it expresses how well any posterior distribution (the modelled dates) agrees
with the prior distribution (Bronk Ramsey 1995, 429). Always, for each dated element, an agreement index A is calculated, which has a value of c 100% but can be
sometimes higher or fall as low as 60% (Bronk Ramsey 1995, 427). The index of
agreement is high (≥ 60%) when the posterior distribution is situated in a highprobability region of the prior distribution. If the index of agreement is low
(< 60%) it means that the posterior distribution falls in a low-probability region
and the result is regarded problematic (Bronk Ramsey 1995, 429). This may indicate the presence of a statistical outlier, although a very low index of agreement
may suggest that a sample is residual or intrusive.
The overall index of agreement (Aoverall) is calculated for the model as a whole
from the individual agreement indices (Bronk Ramsey 1995, 429) and provides a
measure of the consistency between the prior information and the 14C results. The
overall index of agreement is typically c 100% with a threshold value of 60%, and
models which produce values lower than this should be subject to critical
re-examination (Bayliss et al. 2007, 6; Bronk Ramsey 1995, 429).
The index of convergence is the second criterion that must be checked. The
convergence index indicates how quickly the sampler algorithm (Markov chain
Monte Carlo, MCMC) is able to produce a representative and stable solution to the
model (see Bronk Ramsey 1995, 429). In practice, a model is unstable when the
convergence is poor (< 95%) and the results should not be used (Bronk Ramsey
1995, 429). In such case the sampling procedure can be examined in more detail to
determine whether particular parts of the model are unstable.
Sensitivity analysis: prior beliefs and standardised likelihoods
Always when the data allows, it is advantageous to suggest and provide alternative
chronological models. These comparisons of alternative models are known as
sensitivity analyses because they measure the sensitivity of our estimates to different
models (Bayliss et al. 2007, 7). Sensitivity analyses are useful because they allow us
to argue for the reliability of the preferred model (Bayliss 2009, 134; Bayliss and
Whittle 2007) and identify which elements of the model are vulnerable to incorrect
information (Bayliss et al. 2007, 22).
113
Chapter 3
In a sensitivity analysis, alternative models may explore different readings of the
archaeological data (prior beliefs) and/or examine the robustness of the 14C
measurements and any calibration data required (standardised likelihoods).
Archaeological evidence (prior beliefs) can be used to analyse and compare a
range of statistically plausible models by incorporating different archaeological
interpretations (different priors), to determine whether they provide similar
estimates or not; if the results of the different models are similar it demonstrates
that the conclusions are insensitive to the priors (Bayliss et al. 2007; Bronk Ramsey
2000, 201).
The standardised likelihoods (i.e., 14C measurements, calibration data) may also
be assessed to test the validity of a model. In this case we may want to examine the
effect of varying the calibration data (e.g., reservoir offsets) on the outcome of the
chronological model (Bayliss et al. 2007, 17–8).
Calibration and modelling in this study
In this study, all chronological models are defined by a Bayesian approach to the
analysis and interpretation of 14C measurements in archaeological contexts
(fig. 3.6).
The standardised likelihoods (the dates) consist of 46 14C dates on human bone: 30
from new measurements and 16 previously published (45 individuals). The
atmospheric calibration curve used is IntCal13 (Reimer et al. 2013). The proportion of aquatic protein intake was estimated from the measured δ13C value of each
sample by the application of two methods explicitly indicated in each case as
marine 1 (Ambrose 1993) and marine 2 (Arneborg et al. 1999; Richards and
Hedges 1999). The adopted endpoint values for marine (100%) and terrestrial
(100%) diets are –12‰ to –20‰ respectively. Samples with mixed diet were calibrated using Marine13 curve (Reimer et al. 2013) and regional ΔR accordingly; ΔR
= 140 ± 40 14C yrs BP for the Tagus valley and ΔR = –100 ± 155 14C yrs BP for
the Sado valley (table 3.3).
The prior beliefs (the archaeology) are based on a critical review of the archaeological data in archive documentation and published literature. Stratigraphic relations of the dated material were retrieved from field notes, graphic documentation
such as drawings and site plans, and published data. The archaeological priors from
the historical excavations are limited and the modelling is therefore constrained to
the available evidence. Because of this limitation, the Bayesian model is defined by
simply incorporating the information that the dates derive from a coherent
archaeological activity. Despite the limitations, the application of the method allows the estimate of posterior densities of the radiocarbon measurements, which are
more precise than the date estimates (see Bayliss 2009, 127, 131).
The models are outlined in OxCal 4.2 (Bronk Ramsey 2009) which provides the
posterior beliefs for each model and event.
For this study, the most suitable kind of model is the site-based model. These
are robust models, potentially providing good chronological resolution enabling
regional comparisons as well (Bronk Ramsey 2013, 34). In this study, all models are
114
Methods: Radiocarbon dating and Bayesian analysis
defined as phase models with two boundaries for the start and end of activity. As
discussed, this is a useful approach when we do not have explicit archaeological
data indicating a clear order of a group of elements (Bronk Ramsey 1995). The
phase model simply assumes that the start boundary event occurs before the dated
events, and that these occur before the end boundary event. With a phase grouping
of elements such as this, it is assumed that the dated events belong to the burial
activity in the site, but no particular constraints or fixed relations are assumed or
imposed. The reliability of this approach comes from the fact that each sample
corresponds to a burial event which is part of the burial activity at the site.
Following C. Bronk Ramsey (1998, 462, 2015) “failure to do this will mean that the
events are assumed to be entirely independent apart from the constraint applied.
To use no model at all is in fact to assume that all of the events are truly
independent; this is in effect a model in itself, and in many cases a very unreasonable one”. Each model is presented in detail in chapter 4.
Limits of current standardised likelihoods
Calibration and modelling in this study relies on a set of informed assumptions.
These are the estimated aquatic protein intake values, the adopted endpoint values
for marine (100%) and terrestrial (100%) diets, and the regional ΔR.
Estimated aquatic protein intake
The marine (i.e., non-atmospheric) fraction of each sample was calculated by the
application of two methods (see above, Sources of protein component in human bone
collagen: calculation estimate). The values are slightly different in some cases but show
agreement on the overall impression of the main source of protein in the diet of
each individual. To determine if these differences would have an impact in the
outcome of the chronological interpretation, the models were run independently
with the values calculated with methods marine 1 and 2. In every case, the results
were similar, demonstrating the robustness of the estimates.
Endpoint values
The knowledge of the regional endpoint values for marine (100%) and terrestrial
(100%) diet requires an in-depth study of the local ecosystem of food chains at the
time the shell middens were in use. In the absence of such studies, the adopted
endpoint values are –12‰ to –20‰ (see section, Stable isotope analysis of carbon and
nitrogen). These values may change slightly with further comprehensive studies of
the regional ecosystems. Nevertheless, this should have little impact on the
chronological interpretation of the sites. While R. Hedges (2004) pointed out that
1‰ change in the terrestrial endpoint corresponds to c 10% change in the estimate
of marine protein, experimental work has shown that the exact choice of endpoints
should not appreciably change the calibrated ages (Arneborg et al. 1999, 159).
115
Chapter 3
ΔR
In this study, the human remains were calibrated according to the ΔR of the region
where they were buried. This approach assumes that the place of burial and the
location(s) where the aquatic foods were collected were in the same region, within
the same ΔR. However, this may not be the case. Aquatic foods, such as fish,
could have been captured in other locations, such as by the coast, carrying a different reservoir offset (table 3.3). Thus, some individuals may have consumed foods
from various aquatic offsets, marine (ΔRCoast= 95 ± 15) and estuarine
(ΔRTagus= 140 ± 40; ΔRSado= –100 ± 155). However, the ΔR values are not significantly different and few individuals derive their isotopes from the intake of 50% or
more of marine-estuarine foods (see chapter 4).
Sensitivity analysis of regional reservoir offsets
To examine the impact of the different regional ΔR values in the outcome of the
chronological models I tested the calibration of a fictitious 14C date of 7000 ± 50
BP with an assumed marine/estuarine diet of 50 ± 10%. The calibrated date ranges
are identical (fig. 3.9). When calibrated with the coastal ΔR the range is 5990–5761
cal BCE (95% confidence), which is identical when calibrated with any of the
estuarine ΔR values, 5990–5762 cal BCE (95% confidence).
The sensitivity analysis shows that varying the relevant reservoir offsets (marine,
estuarine) does not affect the outcome of the calibrated dates and posterior density
estimates. The analysis shows that we can be confident about the calibration of
human bone collagen from the Tagus and Sado valleys, even when the aquatic
reservoirs may be unclear.
116
Methods: Radiocarbon dating and Bayesian analysis
Figure 3.9. Sensitivity analysis of reservoir offsets. A fictitious 14C age of 7000 ± 50 BP was
calibrated with three reservoir offsets (ΔRCoast= 95 ± 15; ΔRTagus= 140 ± 40; ΔRSado= –100
± 155) but presents identical calibrated date ranges. Sensitivity analysis of the reservoir
offsets shows no significant effect on the outcome of the calibrated date ranges. Calculated
by OxCal 4.2 (Bronk Ramsey 2009) using IntCal13 curve (Reimer et al. 2013).
117
Chapter 3
Selection and screening of samples in this study
Nature and context
Articulated bones in archaeological contexts are among the most reliable category
of samples for 14C dating (Bayliss et al. 2011, 38). In this dissertation, the articulated human skeleton was the targeted sample which dates the event that is of interest in this research: the practice of burying human remains. In this dissertation,
only samples of Homo sapiens bone in recorded archaeological context were selected
and considered for the construction of chronological models of burial activity. The
selection of samples of this nature ensures a direct relationship between the sample
and the human activity of interest, otherwise difficult to establish, particularly
when studying archaeological material from historical excavations. The selection
follows the single-entity sampling approach (Ashmore 1999) and each 14C date
comes from a bone sample from a single individual. This approach ensures that
each 14C date estimates the time of death of the individual and dates the burial
event within the period of burial activity at the site (Bayliss et al. 2011).
Contamination
Contamination of the sample is another important factor to take into consideration
when choosing a sample for 14C dating because contamination of a sample may
result in an erroneous date (Bayliss et al. 2011, 38; Brock and Dee 2013, 45). This
is a significant issue in the archaeological material under study in the scope of this
dissertation. Several human burials excavated from the nineteenth century to the
1960s in the shell middens of the Tagus and Sado valleys were covered with
paraffin wax in order to maintain their original position or have been treated with a
binding product to avoid the collapse of fragile bone elements. These substances
cover not only the surface of the bones but often penetrate through the bone.
Despite the rigorous selection to avoid contamination, 14 out of 52 samples
selected in the scope of the radiocarbon dating programme were rejected after pretreatment while 7 out of 52 samples did not yield enough collagen for reliable
measurements (see below, Reliability of the isotopic measurements; chapter 4).
In the laboratory: pre-treatment and measurement of samples in this study
All samples selected in the course of this dissertation’s radiocarbon dating
programme were processed by the Tandem Laboratory at the Department of
Physics and Astronomy, Uppsala University, in 2013 and 2014, and are identified
as Ua–number. 14C measurements were done using accelerator mass spectrometry
(AMS) and the stable isotopes 13C and 15N were measured by EA-CF-IRMS
(Elemental Analyzer Continuous Flow IRMS). The samples were pre-treated
following the HCl method as reported by the laboratory:
1. The surface is mechanically cleaned (scraping, in some cases sand blasting).
2. The sample is ultrasonically cleaned in boiled distilled water, pH=3.
118
Methods: Radiocarbon dating and Bayesian analysis
3. Grinding in mortar.
4. 0.8M HCl is added, stirring at 10°C for 30 minutes (apatite removed).
Soluble fraction is named fraction A.
5. Distilled water kept at pH=3 is added to the insoluble fraction, which is
stirred for 6–8 hours at 90°C. Insoluble part is named fraction C and
soluble part is named fraction D. Fraction D should give the most relevant age, since it contains most of the organic parts (the “collagen”) of
the original bone. However, information on the influence of contaminants could be obtained from the other fractions. In critical cases they
should preferably be dated as well. The quality of the bone (and the reliability of the age) could be judged by the chemical yields in the different
stages of preparation. The fraction to be 14C–dated is combusted to
CO2 and then converted to graphite using a Fe-catalyst reaction. The
age of fraction D has been measured in the present investigation.
6. Fourteen samples were pre-treated with boiling in acetone, ultrasonification first with ether and then with ethanol, and finally filtrated; after
that, processed as usual (5.).
Reliability of the isotopic measurements
Quality parameters
Collagen can be extremely resistant to post-mortem alterations (Ambrose 1993,
72), thus reflecting the isotopic signal that would have been measured in vivo. However, the preservation of collagen in human bone is dependent on factors such as
temperature, extremes of pH, the presence of organic acids, and so forth (Pollard
et al. 2007, 21), and can vary greatly within archaeological sites.
Degradation (diagenetic alteration) and contamination (presence of exogenous
substances) of the bone can significantly compromise the validity of the 14C
measurements and other chemical analysis. A set of parameters can be considered
to evaluate the sample quality and reliability of the measurements: chemical indicators, elemental data and stable isotopic ratios of nitrogen and carbon (van Klinken
1999). However, a single quality parameter may be insufficient on its own to clearly
indicate contamination or degradation of the sample and should be used in
combination with other quality parameter checks (van Klinken 1999). The screening of sample quality is normally done in the 14C laboratory. Nonetheless, the
archaeologist should have a basic knowledge of the acceptance ranges for each of
the indicators, for each type of sample, in order to critically assess the data received
from the laboratory and evaluate the validity of the results.
Chemical indicators: one of the chemical indicators for bone collagen extracts is
the collagen yield (expressed as weight percentage, wt %). Modern fresh bone has
c 22 wt % collagen but it drops steadily after burial (van Klinken 1999).
Well-preserved prehistoric bone should have more than 1 wt % collagen (Ambrose
1993, 75; van Klinken 1999, 689). Samples with values ranging between 0.5 to
119
Chapter 3
2 wt % should be checked with other parameter to confirm the integrity of the
sample (van Klinken 1999).
Elemental data: The degree of preservation of collagen is indicated by the atomic
carbon and nitrogen ratios (C:N) measured for each sample (DeNiro 1985; Ambrose
1993). Experimental work has shown that collagen C:N ratios between 2.9 and 3.6
indicate minor post-mortem alterations (DeNiro 1985), and this indicator can be
used as one of the parameters to assess the quality of the chemical measurements
of a sample (van Klinken 1999). C:N ratios between the 3.4 and 3.6 range may
indicate some contamination, by lipids, carbonates, humic acids or other carbonrich substances (Ambrose 1993). Van Klinken (1999) recommended that the C:N
parameter should be used with other indicators because in some cases it may be
insufficient on its own, and proposed a slightly different range between 2.7 and 3.3
(van Klinken et al. 2000).
Stable isotopic ratios of nitrogen and carbon: δ13C and δ15N values can be used for
quality checks. Out of range values are of concern and should be evaluated in
combination with other quality parameters before accepting or discarding the sample. Values between –22 to and over –24‰ for δ13C may indicate degradation
and/or contamination of the sample. δ15N values for quality checks are more
limited and only meaningful when combined with the local trophic level values
(van Klinken 1999; van Klinken et al. 2000).
Following this, contaminated or low-collagen samples are expected to have:
• variable, but mostly higher C:N ratios;
• variable, but mostly more negative δ13C;
• mostly, more positive δ15N values, in the case of low-collagen samples
(Ambrose 1990; DeNiro and Hastorf 1985; van Klinken 1999).
The usefulness of some of these parameters will vary with the laboratory procedures which the archaeologist should be aware of. In the case of the samples
measured in the scope of this study (Ua–number) the parameter collagen yield was
not provided by the laboratory because this indicator is not used as a measure of
bone quality. In this case, the atomic carbon and nitrogen ratios (C:N), and δ13C
and δ15N values can be used for quality checks.
δ13C values: AMS or IRSM
It is important to be clear whether the δ13C values were obtained by accelerator
mass spectrometry (AMS) during the process of 14C measurement or independently by isotope-ratio mass spectrometry (IRMS). In the case of previously published measurements, if the method is not explicit, it is prudent to confirm with the
correspondent author or with the laboratory.
The δ13C values associated with the dating process are normally offset from the
values obtained by an IRMS instrument (Schulting and Richards 2001, 320; Taylor
and Bar-Yosef 2014, 117). The offset is variable and although the AMS- and
IRMS-based values may be highly correlated (Schulting and Richards 2001, 320)
with an average offset lower than 2‰, it can be as high as 10‰ (Taylor and Bar120
Methods: Radiocarbon dating and Bayesian analysis
Yosef 2014, 117). For this reason, only IRMS-based δ13C values can be used for
dietary reconstructions or other isotopic-based environmental analysis such as
reservoir corrections (Millard 2014, 557; Taylor and Bar-Yosef 2014, 117).
Screening of samples in this study
Each measurement was screened based on the acceptance ranges of the atomic
ratio of carbon and nitrogen (C:N) in combination with the stable isotopic ratios of
nitrogen and carbon (δ13C, δ15N). Samples with C:N ratios outside the 2.9–3.6
range have been discarded as well as samples within the 3.4–3.6 range combined
with an out of range δ13C and/or δ15N value (table A1).
Each δ13C value was checked and in every case it is indicated whether the values
were obtained by AMS or IRMS. Most δ13C values measured from human bone
collagen were obtained by IRMS, however, in some cases, the δ13C values reported
and published alongside the 14C ages were obtained on an AMS instrument (table
3.4), and this can be outlined as follows:
• All δ13C values obtained for the radiocarbon programme in the course
of this study (Ua–number) were measured by EA-CF-IRMS (Elemental
Analyzer Continuous Flow IRMS).
• All δ13C values published by Jackes et al. 2014; Lubell and Jackes 1985,
1988; Meiklejohn et al. 1986; Roksandic 2006; , are IRMS-based values
obtained independently from the dating process (Mary Jackes, pers.
comm.; Lubell et al. 1994).
• All δ13C values obtained for dietary studies (Fontanals-Coll et al. 2014;
Guiry et al. 2015; Umbelino 2006) are IRMS-based values.
• All δ13C values published by Cunha and Cardoso 2002–03; Cunha and
Umbelino 2001; Cunha et al. 2003, are AMS-based values associated
with the dating process (Cláudia Umbelino, pers. comm.) which cannot
be used for dietary studies or reservoir corrections. When possible,
these 14C measurements were corrected with recent IRMS-based δ13C
values obtained for the same individual (table 3.4). The calibration of
14C ages for which there are no IRMS-based δ13C values is problematic
and may be significantly misleading, particularly in the case of complex
estuarine environments such as the sites under study. Thus, all 14C
measurements from individuals without IRMS-based δ13C values are
excluded until these values are obtained.
• The excluded measurements and corresponding AMS-based δ13C values
are discussed in chapter 5 in comparison with the isotopic series
available for each site.
Following the selection criteria for retaining or rejecting determinations, the 14C
data set of this study comprises 46 14C dates on human bone: 30 from new measurements and 16 previously published (45 individuals). This data set comprises all
reliable 14C dates on human bone collagen for the sites of Moita do Sebastião and
121
Chapter 3
Cabeço da Arruda at the Tagus valley, and Arapouco, Cabeço das Amoreiras, Vale
de Romeiras, Cabeço do Pez and Várzea da Mó at the Sado valley.
122
1958
MNA
5
Cabeço
das
Amoreiras
Cabeço
do Pez
Cabeço
do Pez
Sado
Sado
Sado
4
1956
MNA
1956
MNA
1962
MNA
Arapouco 2A
Sado
4
1937
MHNUP
Identification of the
human remains
Ind. Excav.
Museum
16 1952–54
MHNUP
Tagus Cabeço da 6
Arruda
Tagus Moita do
Sebastião
Valley Site
Sac–1558
Beta–125109
Beta–125110
Sac–1560
Beta–127451
Beta–127499
Lab no.
Age BP
6740±110
6760±40
7230±40
7200±130
7550±100
7120±40
14C
–19.3
–22.6
–20.8
–16.9
–19.0
δ13C (‰)
VPDB
–16.8
Measurements 14C and δ13C by AMS
Cunha and Umbelino 2001
Cunha and Umbelino 2001
Cunha and Umbelino 2001
Cunha and Umbelino 2001;
Umbelino 2006
Cunha and Cardoso 2002–03;
Cunha et al. 2003
Cunha and Cardoso 2002–03;
Cunha et al. 2003
Reference
Excluded until IRMS values are available.
Excluded until IRMS values are available.
Calibrated with recent IRMS–based δ13C value
(–19.0) obtained by Fontanals-Coll et al. 2014.
Calibrated with recent IRMS–based δ13C value
(–17.2) obtained by Guiry et al. 2015.
Combined (OxCal, R_Combine) with new AMS
date (Jackes and Lubell 2014) and respective
IRMS–based δ13C value.
14C
Excluded until IRMS values are available.
Obs.
Table 3.4. Measurements of problematic calibration from samples on human bone collagen. δ13C values reported and published alongside the 14C ages were obtained on an
AMS instrument.
Stable isotope analysis of carbon and nitrogen
Introduction
Burial analysis and past human diets
Human food choices are culturally defined in an environmental framework, with
dietary patterns being highly dependent on ecological constraints, particularly in
the case of past populations. In this sense, ethnographic data is revealing by showing a systematic relation between hunter-gatherer diets and their environment
(Kelly 2013). Subsistence studies have been a central subject in the anthropology of
hunter-gatherers, however the ethnographic methods used to estimate the foods
consumed are often vague and inconsistent, and in many cases stereotyped (Kelly
2013, 40–44). Direct analysis of human remains is a reliable approach to overcome
these biases. Stable isotope analysis of human remains in archaeological context is
a scientific method widely used for assessing past dietary patterns (see Makarewicz
and Sealy 2015). The method of dietary analysis works from the observation that
the isotopic composition of human tissue, such as bone collagen or tooth enamel,
largely reflects that of dietary protein (Ambrose and Norr 1993; DeNiro 1985).
Additionally, the method recognizes that classes of dietary resources (i.e., plants,
omnivores and carnivores) have characteristically different isotopic ratios
(Ambrose 1993; DeNiro 1985), and present a systematic difference between the
isotopic composition of the consumer and the ingested food resource (i.e.,
fractionation factor) (Ambrose 1993).
The comparative analysis of dietary patterns can reveal social aspects of the
lifeways of hunter-gatherers. This is highly relevant for archaeological research
because it offers a tool to discuss past relations to food at various levels, such as
gender and age, group identity, and differential access to food resources. In the
scope of this dissertation, centred on the mortuary practices of the last huntergatherers of the Tagus and Sado valleys, the analysis of dietary patterns is aimed at
identifying and comparing intra- and inter-site variation at the isotopic scale of
each individual in each burial place. In this framework, the method of dietary
analysis is applied to identify broad patterns in terms of the main sources of protein intake which can be outlined as follows:
• to identify trends in terms of sources of protein in the diet, in each
valley in general and in each site in particular;
• to identify synchronic and diachronic dietary patterns;
• to discuss homogeneity and/or heterogeneity of the isotopic signatures
of each burial ground and each valley.
124
Methods: Stable isotope analysis of carbon and nitrogen
In this dissertation, I aim to discuss some aspects on which isotopic data can be
particularly informative in terms of mortuary practices. Individual isotopic signatures may clarify whether the burial grounds are places of funerary activity for one
or various dietary groups; bearing in mind that a group may comprise several
sub-dietary groups defined by gender, age, or status. The following hypotheses are
possible:
1. One dietary group uses these sites to bury their dead.
2. More than one dietary group uses these sites to bury their dead.
a. Each group uses one main burial place (i.e., one group per site).
b. Each group uses various burial places (i.e., more than one group at
each site).
Also, in the light of the funerary activity and based on the isotopic data, I will discuss briefly whether the isotopic signatures reflect the ecology of the burial places,
or rather the ecology from elsewhere, taking into account the current state of
research and insufficient environmental data. How local are the deceased buried in
these sites?
The first study applying isotopic evidence from the human remains recovered in
the Portuguese shell middens was centred on the Mesolithic-Neolithic transition
(Lubell et al. 1994). D. Lubell and colleagues (1994) compared the isotopic profiles
from the Muge Mesolithic with those obtained from Neolithic sites in Portugal.
This pioneering study indicated a decrease of marine foods in the diet as well as a
decrease in diversity of food choices through time, suggesting a significant shift in
dietary patterns from the Mesolithic to the Neolithic, characterized by the intensification of terrestrial sources and increased homogeneity of food choices towards
the Neolithic (Lubell et al. 1994). A more in-depth study on Mesolithic and
Neolithic palaeodiets from Portuguese sites, based on the analysis of trace elements and stable isotopes, confirmed the pattern of Mesolithic heterogeneity based
on a mixed diet obtained from marine and terrestrial sources (Umbelino 2006;
Umbelino et al. 2007). Latest studies centred in the Sado valley assemblages once
again confirmed the regional heterogeneity already known for the Mesolithic. With
a larger data set and greater resolution, these studies were able to show significant
inter-site variability, indicating that the heterogeneity found in Mesolithic diets may
derive from the differences found between different groups (Fontanals-Coll et al.
2014; Guiry et al. 2015). This suggestion, already hypothesised by J. Martins and
colleagues (2008, 83), contradicts the seasonal model for the settlement pattern
proposed for the Sado valley (Arnaud 1989), which suggested that most of the
population living in the valley moved seasonally between large basecamps while
making short incursions to smaller temporary camps for specialized activities.
While the seasonal scenario may admit heterogeneity at the individual level, which
could be explained by differential access to food sources depending on gender, age
or status; it does not account for significant inter-site variability, because in such
model the sites were presumably seasonally used by the same people.
125
Chapter 3
New stable isotope data of carbon and nitrogen were obtained in the course of
this dissertation in order to further test these scenarios along with higher chronological resolution. The selection of human bone collagen samples was fully
dependent on the goals of the radiocarbon dating programme which are described
in detail for each archaeological site in chapter 4.
This section
The section Stable isotope analysis of carbon and nitrogen presents a brief overview of the
main principles of dietary analysis and interpretation of isotopic measurements of
carbon and nitrogen, making explicit the set of informed assumptions used in this
dissertation in the interpretation of human bone collagen measurements of δ13C
and δ15N.
Like the previous section Radiocarbon dating method and Bayesian analysis, this
methodological review is written from the perspective of the analysis of unburnt
prehistoric human bone in Holocene archaeological contexts in the temperate zone
of the northern hemisphere, concerning south-western Europe in particular, and
for this reason, issues relevant to other archaeological contexts may not have been
addressed. This is the case, for example, for contexts in environments of C4 pathway plants (e.g., maize, millet) which were not available in Mesolithic Europe.
Other stable isotopes of elements such as hydrogen, oxygen, sulphur, or strontium,
are also of interest for palaeodietary and palaeoenvironmental reconstruction, but
are not discussed in the scope of this study.
The results of the isotope measurements of carbon and nitrogen are presented
by archaeological site in chapter 4 and plotted with previously published data. In
chapter 5 the isotopic data is analysed and contextualized from the perspective of
funerary practices.
Terminology and conventions
Carbon–12 (12C), carbon–13 (13C), nitrogen–14 (14N), and nitrogen–15 (15N) are
the natural isotopic variants of carbon and nitrogen, and can be used as tracers in
biological systems because they do not decay over time (Pollard et al. 2007). The
variation of stable isotopes in the environment and food chains is usually very
small. For this reason, the measurement of their abundances involves measurement
of the ratio of the heavier to the lighter isotope, with references to the ratio of a
standard reference material: Vienna Peedee Belemnite (VPDB) for 13C, and atmospheric N2 (AIR) for 15N ratios (see e.g., Mays 2010, 266; Tykot 2006, 132). Thus,
the measured isotope ratios of a given sample are expressed as delta (δ) units in
parts per thousand (permil: ‰) (Ambrose 1993, 65; Mays 2010, 266). The δ13C
values are always negative numbers, because they represent less 13C than in the
reference standard while the δ15N values are always positive numbers because most
food resources have more 15N than the standard (Ambrose 1993, 65). In this dissertation, the isotopic measurements are presented rounded to one decimal place.
In this study, the analysis of intra- and inter-site variability of isotopic signatures
is based on the measure of standard deviation (SD) which quantifies the dispersion
126
Methods: Stable isotope analysis of carbon and nitrogen
of a set of data values (VanPool and Leonard 2011). Variation is considered small
when SD is equal or lower than 0.5‰, moderately high when it is equal or higher
than 1.0‰, and high when equal or higher than 1.5‰ (chapters 4 and 5).
Principles of dietary reconstruction based on the isotopes of carbon
and nitrogen
The method is based on the principle that carbon and nitrogen ratios (13C/12C and
15N/14N) will reflect those of the food resources, adjusted for the fractionation
factor related to the relationship between trophic level and isotopic value of foods
and consumers (Ambrose 1993; Pollard et al. 2007, 182). Since the range of stable
isotope values in a variety of food resources is known, the measurements of δ13C
and δ15N in human bone collagen will reflect which foods were consumed by the
individual (Schulting and Richards 2002, 153). Overall, the δ13C values measured
from human bone collagen samples will give an indication of the proportion of
protein derived from terrestrial and/or marine food sources, while the δ15N values
varying according to the trophic level will indicate the position in the food chain of
the protein foods consumed (Ambrose 1993; Ambrose and Norr 1993).
Knowledge of the past environment
For an accurate assessment and interpretation of the isotopic values obtained from
the human samples it is important to establish the stable isotope ecology of the
region (see Eriksson et al. 2008), because environmental factors can influence isotopic variations (Hedges et al. 2004; van Klinken et al. 1994, 2000). This can be
achieved through the robust sampling and analysis of contemporary assemblages
of archaeological fauna species in order to characterize the extent of the isotopic
variation in the region and in the foods consumed (Eriksson et al. 2008; Hedges
and Reynard 2007; Makarewicz and Sealy 2015). This type of analysis is yet to be
done in the Tagus and Sado valleys and without specific isotopic faunal data from
these regions we have to rely on studies done for other regions in Europe.
Variation of δ13C
The endpoint value normally assumed for archaeological human bone collagen
δ13C values from the European Holocene are about –21/20‰ for individuals deriving their protein from terrestrial C3 pathway plants, or from animals subsisting
on those plants. The δ13C in human bone collagen is elevated with the consumption of foods grown in aquatic environments. The endpoint δ13C value of about
–12 ± 1‰ indicates that the protein portion of the diet was obtained from
marine/aquatic food sources (Chisholm et al. 1982; Richards and Hedges 1999;
Schoeninger et al. 1983; Schulting and Richards 2001, 2002).
Observations made on large data sets of δ13C measurements from across
Europe show a geographic trend suggesting that isotopic data may vary with
latitude presenting a strong correlation to a complex of climatic factors (van
127
Chapter 3
Klinken et al. 1994, 2000). The data suggests that throughout the Holocene the
terrestrial endpoint δ13C values are on average consistently elevated by 1–2‰ in
southern compared with northern Europe (Pollard et al. 2007, 178, 180; van
Klinken et al. 1994, 2000). Following this, and without specific stable isotope
values for Mesolithic terrestrial fauna from the Tagus and Sado valleys, the
terrestrial endpoint δ13C value assumed in this dissertation is –20‰.
Variation of δ15N
It is well attested that δ15N values increase at successive trophic levels in the food
chain. However, recent studies have shown that there are several difficulties in
assessing the trophic level in human palaeodiet (Hedges and Reynard 2007; van
Klinken et al. 2000). The 15N enrichment value in humans is unclear, and although
it is normally assumed to be about 2 to 4‰ for mammals (DeNiro and Epstein
1981; Schoeninger and DeNiro 1984) higher values up to 6‰ have been suggested.
This discrepancy results from the assumption that the δ15N values of plants
consumed by humans and fauna are identical (Styring et al. 2015), but also from
the fact that nitrogen values are specific to regions and ecosystems, which once
again emphasizes the importance of defining the past environment under study
(Ambrose 1993).
Experimental work has shown that breastfeeding infants will present higher
δ15N values than the mother due to bioenrichment from breast milk consumption
(Fogel et al. 1989). In this case, the child will present δ15N values 1 to 3‰ greater
than the adult females in the sample, depending on whether they were
breastfeeding exclusively (Katzenberg et al. 1996).
Overall, the δ15N values estimated for archaeological human bone collagen
from the European Holocene range from 4 to 10‰ for individuals deriving their
protein from terrestrial sources, depending on the trophic level of the foods
consumed. Because the food chains in aquatic systems are typically longer than
terrestrial ones, the values are enriched, and range from 10 to 22‰ for protein
derived mainly from marine/aquatic sources (Richards and Hedges 1999).
Freshwater reservoirs
The consumption of freshwater fish is an additional complication for the reconstruction of palaeodiets (Hedges and Reynard 2007; Keaveney and Reimer 2012).
G. Cook and colleagues (2001) have demonstrated that foods from freshwater
sources may induce a reservoir age in human bone collagen, in a similar way to
marine foods. While the consumption of marine protein can be attested through
elevated δ13C values, freshwater sources of protein are more difficult to detect
(Hedges and Reynard 2007). The δ15N values derived from both marine and
freshwater foods are generally higher than those of terrestrial origin (Schoeninger
and DeNiro 1984), however, these high values are dependent on the productivity
of the aquatic system.
The consumption of freshwater fish may be indicated by the enriched δ15N
values along with low δ13C values (Bonsall et al. 2002; Cook et al. 2001). In these
128
Methods: Stable isotope analysis of carbon and nitrogen
cases, additional archaeological lines of evidence can support the isotopic interpretation, such as fish bones recovered from the site and the proximity to a riverine
biome (Hedges and Reynard 2007). Freshwater systems are complex and without
knowing the ecology of the region all interpretations must be taken with caution
(Keaveney and Reimer 2012).
Assumptions in this study
Despite the sources of variation and uncertainties we can still benefit from stable
isotope analysis of human remains, even if the past environment is not perfectly
known. The present analysis and interpretation of human bone collagen measurements of δ13C and δ15N relies on several assumptions and published values for
European Early Holocene. These can be listed as follows:
• The human bone collagen of an adult individual reflects the diet
averaged over the 10 years prior to death (Hedges et al. 2007).
• The human bone collagen from a diet based on terrestrial C3 pathway
plants, and/or from animals subsisting on those plants, has a δ13C value
of –20‰. This is the assumed terrestrial endpoint for a diet derived
from protein sources of terrestrial origin (Richards and Hedges 1999;
Schulting and Richards 2001, 2002).
• The human bone collagen from a diet based on marine/aquatic
resources has a δ13C value of –12‰. This is the assumed marine endpoint for a diet derived from protein sources of marine/aquatic origin
(Richards and Hedges 1999; Schulting and Richards 2001, 2002).
• The human bone collagen δ15N values range from 4 to 10‰ for a diet
based on terrestrial sources of protein and 10 to 22‰ for protein
sources of marine/aquatic origin (Richards and Hedges 1999).
Based on these premises it is possible to compare objectively the isotopic data sets
available for these sites and gain insights into broad dietary patterns in terms of the
main sources of protein intake, which may be useful to clarify elements of intraand inter-site variation at the isotopic scale of each individual in each burial
ground.
Recent methodological developments are introducing a Bayesian approach to
quantitative diet reconstruction estimates. These are powerful tools which can
handle the different sources of uncertainty associated with past human diets,
including the basis for necessary corrections to collagen 14C dates that have a
reservoir effect (Fernandes 2015; Fernandes et al. 2014). One example of this type
of modelling was introduced by FRUITS (Food Reconstruction Using Isotopic
Transferred Signals) (Fernandes et al. 2014) presenting the way forward in terms of
dietary reconstruction of the past and associated reservoirs, which will be of great
interest to apply to the assemblages of the Tagus and Sado valleys in future
research.
129
Archaeothanatology: chaîne opératoire of mortuary practices
Introduction
Enquiry-dependent taphonomies: the archaeothanatological approach
The method of disposal of a cadaver has different taphonomic consequences. The
complete skeletons of vertebrates can be preserved if quickly buried in areas of
rapid sedimentation, or if they become buried in a slower manner in places
sheltered from transportation forces (Lyman 1994, 137). Some animals, and humans in particular, are important agents behind the covering with sediments (i.e.,
burial) of faunal or human remains, contributing to the preservation of relatively
complete and articulated skeletonized bodies.
In archaeology, the objectives of a taphonomic analysis are varied but the ultimate goal is identical: to identify and separate the taphonomic processes, and lead
to conclusions regarding the human behaviour behind the archaeological context
(Lyman 1994, 5). Taphonomic methods depend largely on the nature of the
assemblage, but most importantly, on the nature of the enquiry. Archaeothanatology is a taphonomic approach to the study of human remains in archaeological
context. The method is dependent on an enquiry, firmly rooted in the physicality
of the body, documenting and explaining differences between the living appearance of the human skeleton and the human remains found in the archaeological
context. This type of analysis provides the tools to distinguish natural processes
from those that are the result of human action. Often, the evidence is insufficient
to unequivocally determine the nature of the processes, however, this method is
one of the most useful and reliable approaches for the study of the human remains
when the objective is to identify and describe the chaîne opératoire of intentional
gestures behind the deposit of the human cadaver in a grave feature (see Duday
2009; Nilsson Stutz 2003a).
This dissertation examines the archaeological features with human remains
excavated in the shell midden sites in the Tagus and Sado valleys. Despite the
unusual number of human burials, there was never a systematic study centred on
the mortuary contexts of these valleys. Recent studies have applied the archaeothanatological approach to part of the assemblages of Moita do Sebastião and
Cabeço da Amoreira in the Tagus valley (Jackes et al. 2014; Roksandic and Jackes
2014), while the material from the Sado sites remained unexplored.
The archaeological material (documentation and human remains) selected for
this dissertation was carefully evaluated in order to determine which contexts
contained high enough resolution for an archaeothanatological analysis (see below,
Source material: the challenges). The quality of the source material varies largely from
130
Methods: Archaeothanatology
site to site and their potential and limitations are described in detail for each site in
chapter 4.
In this study, the objectives of the archaeothanatological analysis were to reconstruct central aspects of the mortuary practices in each context in particular, which
can be outlined as follows:
• to identify the nature of the deposit (primary, secondary);
• to describe the space of decomposition of the cadaver (filled, empty,
mixed);
• to reconstruct the initial position of the cadaver in the feature;
• to reconstruct the grave features, such as size and shape;
• to detect the initial presence of perishable materials deposited along
with the cadaver that in different ways affected the body during the
process of decomposition, such as structures placed behind the bodies
(e.g., headrest, cushions, platforms), or wrappings of the body at the
time of the burial;
• to clearly define the deposits containing more than one individual;
• to identify post-depositional manipulations of the cadaver.
The general aim is to reconstruct the mortuary programme in each valley in general
and in each site in particular by defining typical practices as well as the atypical.
This section
Structure
The section Archaeothanatology: chaîne opératoire of mortuary practices concentrates on
methodological issues related to the main principles of the method of archaeothanatology. This methodological review is written from the perspective of the
analysis of articulated and semi-articulated skeletonized human remains, and for
this reason, issues relevant to other archaeological contexts may not have been
addressed. For a complete overview of the method consult The archaeology of the
dead: lectures in archaeothanatology (Duday 2009), as well as reference texts presented
by H. Duday (2006), H. Duday and M. Guillon (2006), and L. Nilsson Stutz (2003).
First, I present a brief history of the method and its impact in archaeology.
Then, I outline the main principles of archaeothanatology, describing the principles
and central aspects of the analysis. In this section I focus on the reconstruction of
the initial position of the cadaver and present in detail some basic observations at
the level of each bone element. Lastly, I discuss the application of archaeothanatology in this study by examining the possibilities and limitations of the method in
the analysis of the historical source material used in this dissertation. Before the
methodological review, I present the terminological choices and conventions used
throughout this study.
The results of the archaeothanatological analysis are presented by archaeological
site in chapter 4. The presentation of the results for each site adopts in general the
131
Chapter 3
structure presented by L. Nilsson Stutz (2003a). The results are analysed and
discussed in chapters 5 and 6.
Terminology and conventions
The lack of unequivocal terminology is a frequent problem in funerary archaeology
(see Boulestin and Duday 2006; Knüsel 2014). Often, the nomenclature used in
different studies varies, making it difficult to interpret and compare sites and
contexts. Frequently, while one term may be used to describe different practices,
one phenomenon may be described by using several different terms. To further
complicate the problem, different research traditions use terminology in different
ways, and in some cases the translations may not reflect the original meaning of the
term, thus it may be good practice to use the original expression along with the
translated term.
Overall, the terminological problem is well known in archaeology, not only in
the field of funerary archaeology. The use of clear and explicit terminology is
fundamental, and its adoption must be well understood. The description of human
remains in archaeological context must be carried out by using specific and welldefined terminology, and by centring the descriptions around the cardinal points of
the body applying anatomical nomenclature. The advantages of following these
procedures are that they allow the replication of results and ensure that the
published data can be used unambiguously by other researchers. Below, I provide
the references used throughout this study for the anatomical terminology in
general, and for the mortuary terminology in particular.
Anatomical terminology
Archaeothanatology uses the international anatomical nomenclature for the human
skeleton (Duday 2009, 16), and in this dissertation I follow the standard reference
as presented by T. D. White and P. A. Folkens (2005). Anatomical terms of bones
are used for the description of the human remains in the feature while anatomical
terms for the parts of the human body are preferred in the interpretation of the
contexts.
The human remains are described in relation to the Anatomical Position of the
human body, also referred to as Standard Erect Position (Bass 2005; White and
Folkens 2005). In this standard, the body is standing and looking forward, with all
limbs in extension, and the palms of the hands are facing forwards with the
thumbs pointing away from the body (White and Folkens 2005, 67, 70).
Directional terms
The description of the human remains in the archaeological feature takes the body
as the point of reference (Duday 2009, 16). Terms of location are always referred
from the perspective of the body and never from that of the observer. For example, when describing a location on the right it refers to the right side of the body.
132
Methods: Archaeothanatology
In archaeothanatology, it is essential to describe the lateralization of the bones
as they are found in the archaeological feature. This is done by indicating the side
of appearance of the bone elements in a plan view and in reference to anatomical
terms of direction, such as anterior, posterior, lateral, medial, proximal, distal, or a
composite view, such as anterior with a lateral component (Courtaud 1996; Duday
et al. 1990; Nilsson Stutz 2003a, 205).
Motions of the body
The motions of the body occur as a function of the movement of the joints, thus
all terms related to motion are described as a function of the joints, following the
standard anatomical terms of motion, such as flexion, extension, abduction,
adduction and rotation (Schwartz 2007; White and Folkens 2005). Additionally,
following D. H. Ubelaker (1989), I use the terms semi-flexed (angle between 90 to
180 degrees), flexed (angle inferior to 90 degrees), and contracted (angle close to
zero degrees), also referred to as hyperflexed. For example, when describing the
upper limb the appropriate expression of motion is flexed elbow rather than flexed
upper limb.
Other terms
For terminological clarity the position of the body in the feature is expressed as
lying on the back, lying on the front, lying on the lateral right/left side.
In this dissertation the expressions body, corpse, cadaver, and deceased are used
interchangeably to refer to a dead human body.
Burial refer to the covering of human remains with sediment soon after disposal
(Lyman 1999, 406), but also to the context where the human remains have been
placed and covered (i.e., burial feature), which may also be referred to as grave feature.
See further discussion below in The nature of the deposits.
Mortuary terminology
The clear use of an explicit terminology is essential for the accurate description and
interpretation of archaeological contexts. This can be particularly problematic in
the field of mortuary studies because terminology can also express human
intention (see Boulestin and Duday 2006; Knüsel 2014). Thus, inconsistencies in
the use of appropriate terminology may be an obstacle to a clear understanding of
the mortuary context.
In this dissertation, I follow the guidelines for the mortuary archaeology in
Portugal (Duarte 2003), adapted to the terminology and concepts used by archaeothanatology (Duday 2009; Duday et al. 1990; Nilsson Stutz 2003a) which are
further explained along with the principles of archaeothanatology.
133
Chapter 3
Archaeothanatology and archaeology
Brief history of the method and impact in archaeology
The methodological principles of archaeothanatology emerged in France in the late
1970s and early 1980s in the context of the exponential growth of rescue archaeology accompanied by an increase of excavation of cemeteries in urban areas. The
method, initially termed anthropologie de terrain (Duday et al. 1990), was developed to
overcome field difficulties when excavating and documenting human remains in
archaeological contexts. This was done by applying basic knowledge of biology –
regarding human anatomy and the decomposition process – to archaeological
questions, in order to correctly identify and argue for the cultural phenomena behind the archaeological context (fig. 3.10).
Figure 3.10. Archaeological contexts with human remains. Identification and interpretation
of cultural phenomena with the methods of archaeothanatology.
Simultaneously, in the UK and particularly in North America, other approaches
were developed on the basis of forensic research, but were of great interest and
applicability to archaeological contexts, notably by the syntheses and publication of
data related to the taphonomic behaviour of the human cadaver (Haglund and
Sorg 1997).
The lack of specific archaeological methodology capable of responding to the
particular questions related to funerary contexts motivated a group of researchers
134
Methods: Archaeothanatology
to formalize a set of tools and practical procedures for the fieldwork done in
France (Duday and Guillon 2006, 118). The methodology evolved from a practical
need, and it is today integrated in the discipline of archaeothanatology (Duday
2009). Its relevance comes from the urgency to resolve various taphonomic
problems in archaeological deposits with human remains, while providing an
understanding of the gestures associated with the funerary practice.
During the first decades of its development, the approach was known as
l’anthropologie du terrain, a term still in use today by some authors. However, the term
carries some conceptual problems. First, because there are different approaches to
the discipline of anthropology: in the countries of Romance languages in Europe,
such as France, Spain or Portugal, anthropology is the discipline that studies the
human being in its biological dimension, while in the Anglo-Saxon countries as
well as in Northern Europe, the discipline has a greater focus on the cultural
dimension of human behaviour (see Roberts 2009, 2–6). Second, because the term
terrain related to the fieldwork and directly associated with the context of excavation in archaeology, does not have the same meaning in anthropology. In this
sense, B. Boulestin and H. Duday (2006) suggest the use of the term
archaeothanatology. This expression is better framed for the goals and area of expertise of the method, since it is the discipline of thanatology that deals with the biological and social aspects of death. This new terminology was established by H.
Duday (2009), one of the main proponents of the method, in the first methodological synthesis of the discipline The archaeology of the dead: lectures in
archaeothanatology, emphasizing that the expression l’anthropologie du terrain must be
abandoned in order to prevent terminological confusion with other disciplines
(Duday 2009, 3).
Despite the relatively short expansion of archaeothanatology outside the francophone world, the method reached its maturity. The approach has been defined
through more than three decades of research and it follows a strategy of continuous accumulation of data and archaeological references, with numerous contexts
being described, supplying the tools of analysis that allow the comparison of
archaeological observations with the reconstruction of intentional practices. To the
researcher following this method, the collection and publication of references is
taken as a duty always when excavating and studying human remains in archaeological contexts (Nilsson Stutz 2003a, 150). In recent years, some methodological
texts were published in English (Duday 2006, 2009; Duday and Guillon 2006),
allowing the expansion of the method to a wider audience, and exposing it to
important methodological debates.
In Portugal, archaeothanatology has a recent history and few professionals
apply it in the context of archaeological excavations. However, one of the first
references to the method in Portugal, about the methodological potential of the
French approach anthropologie de terrain was published in the late 1980s, Antropologia
de campo e arqueologia funerária (Moura 1989). This short article seems to have gone
almost completely unnoticed and it is rarely cited, but it remains one of the most
detailed and complete reference on the terminology and methodological principles
135
Chapter 3
of anthropologie de terrain, published in Portuguese. Almost ten years later, a new
article was published on theme, with a very similar title Ler os ossos: antropologia de
campo e arqueologia funerária (Silva 1997), however, less detailed in terms of method,
but presenting a case study in Portugal of the excavation of a collective megalithic
tomb at Alcalar, the tholos of Monte Canelas I. In the past ten years, important
steps were taken in order to implement the methods of archaeothanatology to
fieldwork, in particular to the contexts of rescue archaeology (Neves et al., n.d.).
Since 2006, there has been a series of practical courses, organized by a private
company of archaeology, Dryas/Styx in collaboration with the University of
Coimbra. These sessions have a strong interdisciplinary approach and involve the
disciplines of archaeology, physical anthropology, and medicine, with the goal of
bridging the gap between archaeologists and physical anthropologists, by teaching
the methodological principles of archaeothanatology (Ferreira et al. 2008). The
method has been successfully applied to various prehistoric contexts, such as the
recent interventions at the Late Neolithic funerary complex of Monte Canelas in
Alcalar, south of Portugal (Neves and Silva 2010), as well as to the recently
excavated burials at the shell middens of Cabeço da Arruda and Cabeço da
Amoreira in the Tagus valley (Roksandic 2006), and Poças de S. Bento in the Sado
valley (Diniz et al. 2014).
Archaeothanatology: main principles
Aims of the method
Traditionally, research developed in the context of the archaeology of death is
focused on architectonic aspects of the graves and on detailed analysis of grave
goods (see chapter 1), though the death of the individual is the reason for all
funerary actions, and the human body the central element of funerary practices.
The aim of archaeothanatology is to reconstruct past ritual actions in response to
death, by the identification and reconstruction of funerary treatments, having as
point of departure the material remains of the dead. In contrast to traditional approaches, in archaeothanatology human remains are the central question, with
which all methodological processes are concerned (Duday 2006, 2009).
In this approach, the knowledge of what happens to the human body after the
biological death of the individual is essential for the argumentation and
taphonomic interpretation of the various aspects represented in the funerary
context. In archaeology, it is universally accepted that the knowledge of site formation processes is essential for a proper archaeological excavation and its interpretation. Following this, it should be implicit why understanding the taphonomic processes related to the decomposition of the human body are fundamental for the
excavation and interpretation of funerary contexts. This detailed knowledge about
the physical aspects of death is the basis of the taphonomic theory developed by
archaeothanatology, contributing in this way with unparallel insights about past
mortuary practices.
136
Methods: Archaeothanatology
Taphonomy of mortuary practices
Taphonomic history begins with death (Lyman 1994) and the identification of its
processes affecting archaeological human remains is one of the pillars of archaeothanatology. Taphonomy is an independent discipline with its roots in disciplines
such as palaeontology, geology, chemistry and biology, which studies the natural
and non-natural processes involved in the formation of geological and archaeological contexts. By identifying patterns with origins in different agents, the
taphonomic analysis allows the identification of post-depositional phenomena
affecting the archaeological assemblages during the period these have been buried
to their recovery during archaeological excavations.
The goal of archaeothanatology is to identify taphonomic patterns caused by
human agents in order to reconstruct the context created by the actions that
constitute the mortuary practices. Departing from taphonomy, as theoretical foundation, archaeothanatology has been responsible for the development of a taphonomy of mortuary practices, particularly focused on the handling of the human
body, equipping the archaeologist with an important set of tools and establishing a
new confidence and new possibilities for a modern archaeology of death (Nilsson
1998; Nilsson Stutz 2003a, 2008).
Archaeothanatology is thus an interdisciplinary archaeological method, which
applies knowledge of biology about human anatomy and decomposition processes
of the human body, to the interpretation of human remains in archaeological
context, supporting every observation on taphonomic theory, with the objective of
identifying the chaîne opératoire of intentional gestures involved in the funerary
practices of past populations.
Basic principles
In archaeothanatology, the source of information is the human skeleton, more or
less complete, which must be documented rigorously. All elements must be taken
into consideration and their relations carefully evaluated (Nilsson Stutz 2003a,
156). The basic principle of the method is the analysis of the spatial distribution
and orientation of the bones. During excavation, it is important to record in detail
all skeletal elements in order to identify their spatial distribution within the excavated feature, along with a rigorous documentation of the orientation of the bone
and its face of appearance (Courtaud 1996; Duday et al. 1990; Nilsson Stutz 2003a,
205). Archaeothanatology is based on the identification of the post-depositional
dynamics of the movements of the bones, and it is important to understand that
there will always be a difference, more or less pronounced, between the initial position of the cadaver and its situation that the archaeologist will find in the context
of excavation.
In the framework of the study of mortuary practices, it is fundamental to take a
critical and dynamic stand point, and in every situation evaluate the intentionality
of the actions represented in the archaeological context. The presence of human
bones in the archaeological site is not evidence of intentionality of the deposit.
137
Chapter 3
Likewise, a deposit with human remains, even if intentional, is not by itself evidence of funerary ritual. In the study of funerary practices it is necessary to prove
first the intentionality of the deposit, which then may be followed by the identification of the evidence of the funerary practice (Duday 2009).
The most frequent kind of human remains found in archaeological context consist of hard tissue, that is, skeletal elements, such as bone and teeth, which are the
most durable tissues of the body. The skeletonization of the body consists of the
removal of all soft tissue from the bone and is one of the last chronological stages
of decomposition (Lyman 1994; Pinheiro 2006). The decomposition of the body
starts immediately after death. It is a complex process, subject to many variables
and intervenients influencing the transformation, faster or slower, of the corpse
into a skeleton (Pinheiro 2006). It is also an active process, and one of significant
taphonomic consequences, the impact of which in the archaeological context is
commonly neglected. The way the body decomposes may have important outcomes in the way the human remains are found in the archaeological feature. The
process of decomposition is predictable, and when understood, may provide clear
understanding of the non-natural and intentional processes involved in the mortuary practice. The first and second basic principles of archaeothanatology, observation
of the joints and observation of the internal and external environments of the cadaver respectively, are based on the understanding of these processes.
First basic principle: observation of the joints
During decomposition, as the soft tissues dissolve, the skeletal elements begin to
collapse, losing their anatomical integrity. Experimental fieldwork (Hill 1979a–b;
Hill and Behrensmeyer 1984, 1985) has shown that the different types of joints
have different resistance levels during the process of decomposition, suggesting
that the anatomical characteristics of the joint will influence the chronological order of its disarticulation. Thus, the chronological order of joint disarticulation in a
skeleton will largely depend on the anatomy of the joint and its associated soft
tissue (Lyman 1994, 160).
A joint is a region of contact of two or more skeletal elements and is defined by
its range of motion and according to the type of tissue holding it together
(Schwartz 2007). The basic relative chronology for the decomposition and break
down of the joints can be used to establish the relative chronology of the various
events of the taphonomic history of the archaeological deposit with human
remains (Nilsson Stutz 2003a, 152). These observations are of extreme importance
for the application of archaeothanatology, and constitute a tool of great utility to
identify central aspects related to the observed practice.
In archaeothanatology, the joints are classified into labile joints if they break
down early in the process of decomposition, and persistent joints, if they resist
decomposition for longer time (Duday et al. 1990) (fig. 3.11).
138
Methods: Archaeothanatology
Figure 3.11. Classification of joints in archaeothanatology. Joints are classified as persistent
or labile according to their relative resistance during the process of decomposition.
139
Chapter 3
Labile joints can remain intact for a variable length of time, between some weeks
to a few months. These are the temporomandibular (TM) joints, the joints of the
cervical vertebrae (CV) (in particularly C2–C3 and C3–C4), the sternocostal joints,
the scapulothoracic joints, the joints of the hands, the hip joints, and the distal
joints of the feet (Duday 2009). The persistent joints can retain their function for
several months or years, retaining the anatomical integrity of the skeletal elements
involved. These are the atlanto-occipital joint, the joints of the lumbar vertebrae
(LV), the lumbosacral joint, the sacroiliac joints, the knee joints, the ankle joints,
and the tarsal joints (Duday 2009). The disarticulation of the bones held by these
joints can happen due to the effects of gravity on the initial position of the cadaver
in the feature (Duday 2006, 35). This is the case of the lumbar portion of the
vertical column or the sacroiliac joints. Despite being some of the most persistent
structures of the body, decaying very late in the process, these elements are
commonly found disarticulated due to the phenomenon of gravity (Duday 2006,
35). For this reason, the sacroiliac joints have been referred to as relatively labile
which in this case could be explained by the natural oblique position of the
os coxae in relation to the sacrum, rather than to the relative resistance of the joint
(Nilsson Stutz 2003a).
Second basic principle: observation of the internal and external environments of the cadaver
The process of decomposition depends on the internal (endogenous) and external
(exogenous) environment of the cadaver. While decomposition in the endogenous
environment is predictable, the processes in the exogenous surroundings can occur
at different paces and levels of activity within the same cadaver, depending on the
different micro-environments created around the corpse (Pinheiro 2006, 87). The
dynamics between these two spaces will influence the distribution pattern of the
bone elements in the archaeological context.
To reconstruct the taphonomic history of the deposit it is necessary to consider:
first, the initial volume of the cadaver and the decomposition of soft tissue in its internal environment, and second, the external environment of the cadaver, the space of
decomposition. In both internal and external environments, the decomposition of
organic matter will result in empty spaces, which in the case of a buried cadaver,
will be filled by sediment. The observation of these dynamics constitutes one of
the basic principles of the archaeothanatological method (Nilsson Stutz 2003a,
151). This analysis is important because all empty spaces created during the process
of decomposition in the space of disposal can potentially contribute to the
instability of the bones and influence their spatial distribution pattern in the
archaeological feature (fig. 3.12).
The dissolution of soft tissue will create empty spaces inside the initial volume
of the cadaver. These new empty spaces will not only destabilize the initial position
of the bone elements but will constitute pockets where the bones can collapse and
gain new stability. The disarticulated bones will find a new balance when their
movement is blocked by the filling of sediment or by the presence of some other
140
Methods: Archaeothanatology
element. The new equilibrium will be achieved also in relation to the external
environment. This equilibrium is disrupted during archaeological excavation, when
the sediment is removed while lifting the bones (Duday and Guillon 2006, 130;
Duday 2009, 17).
***
The processes described here are complex to observe and need a dynamic and
critical approach. However, their effects are predictable and distinguishable from
other post-depositional taphonomic processes, such as disturbances from the
activity of a rodent, or geological movements of the soil. The observations, based
on the two basic principles of archaeothanatology, provide the argumentative
support for the identification and reconstruction of cultural phenomena associated
with the mortuary practices, such as the nature of the deposits, the space of
decomposition of the cadaver, the initial position of the cadaver in the feature, the
grave features, deposits containing more than one individual and post-depositional
manipulations of the cadavers (fig. 3.13).
Figure 3.12. Place of disposal of the cadaver. The external and internal environments will
influence the distribution pattern of the bone elements during decomposition of soft tissue
and other organic material in the feature.
141
Chapter 3
Figure 3.13. The two basic principles of archaeothanatology and the reconstruction of aspects of the funerary practices.
The nature of the deposits
The variability of the cultural responses to death carries an assortment of ways of
dealing and manipulating the human remains. Archaeothanatology distinguishes
between two types of archaeological contexts where human remains can be found:
the primary deposit and the secondary deposit, also termed primary or secondary burial
(Duday 2009, 14; Nilsson Stutz 2003a, 206).
Primary deposit
The archaeological context with human remains is a primary deposit when the archaeothanatological analysis indicates the direct deposition of a fresh cadaver in
the feature while it still retains its anatomical integrity (Duday and Guillon 2006,
125; Duday 2009). The primary deposit is where the breakdown of the organic
matter takes place, being through decomposition, in the case of burial, or through
fire, in the case of cremation, for example.
The positive identification of a primary deposit is easily argued when the bones
articulated by labile joints retain their anatomical position. However, preservation conditions do not always allow such clear identification, and in the case of the rare
cremations in primary deposits it becomes more complicated (see Duday 2009,
152–53). In fact, the anatomical preservation of all labile joints is rare (Duday and
Guillon 2006, 131) but the observation of some of the labile joints in anatomical
position demonstrates the integrity of the cadaver at the time of disposal, indicating the primary nature of the deposit. It is important to emphasize that the nonpreservation of the labile joints is not sufficient to exclude the interpretation of a
deposit in primary position (Duday and Guillon 2006, 128). The disorder of the
142
Methods: Archaeothanatology
skeletal elements is a weak argument supporting the intentional manipulation of
the bones and it is necessary to take into account other factors that may have
disturbed the deposit.
The general topography of the body can demonstrate the primary nature of the
deposit if the overall anatomical order of the skeletal elements has been maintained, even if few or none of the other diagnostic criteria (i.e., labile joints) has
been preserved (Duday 2009, 27).
The primary nature of the deposit may be attested by the observation of fluids of
decomposition altering the colour of the sediment involved, as well as by the identification of remains of organisms associated with the process of decomposition
(Duday and Guillon 2006, 131).
The primary nature of a deposit can be demonstrated if one of or all these
criteria are identified in the archaeological feature (table 3.5).
Table 3.5. Arguments supporting the nature of a deposit in primary position.
Primary deposit
Arguments
Maintenance of labile articulations
Maintenance of general anatomical integrity of the body
Observation of products of decomposition
Secondary deposit
The secondary deposit, more expressively termed as funeral in multiple episodes (funérailees
en plusieurs temps, or funérailles decallées) emphasizing the dynamics involved in this
type of practice (Nilsson Stutz 2003a, 206), contains the manipulated remains of
the deceased and is the result of the treatment of the remains in multiple episodes.
This practice is characterized by the occurrence of one or more distinct and
successive phases, between the moment of death and the final deposit. In the
archaeological context we may find the last stage of manipulation, or rather an
intermediate phase of a complex process (see Gray Jones 2011).
This process results in the disarticulation of the skeleton, total or partial, and
the bones articulated by labile joints are more dispersed or even missing. It is
almost common place in archaeology to consider that the absence of the bones of
the hands and/or feet indicates the secondary nature of the deposit. This is
however a weak argument because the absence of these bones may be explained in
many cases by taphonomic agents and/or poor excavation methods (Duday 2009,
89).
The secondary deposit is easily identified when the context consists of a selection of skeletonized or cremated bones placed in a container much smaller than
the human body. In most cases the clear identification of secondary deposits is
difficult and requires a multiple set of observations which cannot be explained by
143
Chapter 3
excavation bias or taphonomic agents such as bones in disarticulation, or absence
of small bones.
Other observations, such as the retention of specific skeletal elements, such as
crania or long bones (Andrew and Bello 2006, 17), or a particular spatial organization of the bones in the archaeological feature, may constitute strong arguments
supporting the identification of a secondary deposit.
The space of decomposition of the cadaver
Empty and filled spaces of decomposition
The decomposition process starts soon after death, with the process taking place in
an empty space or in a filled space (Duday 2009, 32; Nilsson Stutz 2003a, 252)
(fig. 3.12). The relevance of the identification of the environment (i.e., space) of
decomposition is twofold: first, it represents a relevant aspect of the mortuary
practice, and second, it presents significant taphonomic consequences in the spatial
distribution and orientation of the human remains in the archaeological feature.
The space of decomposition is empty if the cadaver is in an open air space, or if
the body is placed in a hard container, such as a coffin or a crypt chamber. The
space of decomposition is filled if the cadaver is covered with sediment soon after
death, either intentionally, or by accidental or natural causes. Within these two
possibilities, empty or filled space, other empty spaces can be created by the
decomposition of organic elements external to the cadaver. These are secondary
empty spaces and may be detected by their eventual effect on the spatial distribution
and orientation of the skeletal elements in the feature. The identification of these
secondary spaces is of great interest to the study of mortuary practices, because it
allows the identification of perishable materials deposited along with the cadaver,
which do not leave any trace, besides the taphonomic impact on the feature. Even
if the cadaver is initially located in an empty environment, it will eventually be
covered with sediment, unless the remains are placed in a durable and hermetic
container such as a stone sarcophagus. This taphonomic process occurs at different speeds during the process of decomposition, depending on different factors.
The chronology of the filling of the different spaces can be progressive and
immediate (colmatage progressif) or delayed (colmatage différé) (Duday 2009, 52; Nilsson
Stutz 2003a, 254). The filling of the internal volume of the cadaver is dependent on
the characteristics of the space of decomposition and the properties of the
sediment. If the cadaver is buried in direct contact with the soil, in a filled space
sufficiently porous and fluid, while the soft tissues dissolve, their original space is
immediately and progressively substituted by sediment without the formation of an
empty space. In the case of immediate filling, the movement of the bones will be
minor even in the regions of contact of labile joints, since the penetration of
sediment will allow the maintenance of the original equilibrium of the skeletal
elements (Duday 2009, 38) (table 3.6). The term hourglass effect is used by archaeothanatology to describe situations when the infilling is particularly fluid. This effect
occurs in the presence of very fine sediment, such as small grained sand or
144
Methods: Archaeothanatology
powdery ash, and is a diagnostic criterion for the placement of the cadaver in
direct contact with the soil (Duday 2006, 41).
Table 3.6. Arguments supporting the space of decomposition in a sediment-filled environment.
Filled space of decomposition
Arguments
Maintenance of bones in original unbalanced position
No movement outside the initial volume of the cadaver
When the cadaver is not in immediate contact with sediment, separated either by a
more resistant kind of wrapping creating an empty space around the corpse, or by
the deposition of the corpse in an empty space such as a hard container, the filling
will be delayed and may not be fluid and progressive. In this case, the movements
of the bones will be more pronounced, depending on the forces of gravity towards
the initial position of the cadaver (Duday 2009, 35). Basically, if the body decomposes in an open space, at least some bones will have the tendency to fall outside
the initial volume of the body during decomposition. The most relevant diagnostic
criteria are given by the position of the os coxae and the lower limbs, in which the
documentation of the patellae is one of the key diagnostic elements (Duday 2009,
36) (see below, Patellae).
Wall-effect (effet de parois)
The balanced position of skeletal elements is determined not only by the filling
pattern of sediment, but also by the relation between the bones and the limits of
the feature, and between other elements in the space of decomposition. Archaeothanatology recognizes the presence of non-sedimentary supports, responsible by
the maintenance of the bones in equilibrium. This phenomenon is known as walleffect (effet de parois) (Duday 2006, 41; 2009, 40 Nilsson Stutz 2003a, 256). The walleffect consists of any obstacle that constrains the movement of the bones while
providing support and preventing their collapse. This effect may be produced by
the physical limits of the feature or by the presence of elements with slower
decomposition rate placed close to the body. Elements covering or involving the
body, or even clothing can restrict the movement of the bones.
The initial position of the cadaver in the feature
If the methods of excavation and documentation are sufficiently detailed and
rigorous, it is generally possible to reconstruct the initial position of the cadaver in
the feature. This analysis must take into account the natural movements of the
skeleton and the various phenomena occurring during decomposition. The postdepositional movements follow simple and logical rules, dictated by the relative
chronology of decomposition of the joints and by a key taphonomic principle, the
law of universal gravitation (Duday and Guillon 2006, 128). This principle affects
145
Chapter 3
all movements but its impact depends largely on the initial position of the cadaver
in the feature and on the topography of the grave feature (Duday 2006, 34).
Here, I present several key observations at the level of single bone elements that
have been described in the literature, which are relevant to the current study. It is
important to emphasize that I chose this format to facilitate the presentation.
Archaeothanatology is a dynamic analysis and the skeletal elements, influencing
one another, must be taken as a whole. The interpretation should not be carried
out from the observation of one single element but of a composite observation of
various elements in a dynamic process.
Cranium and mandible
Observations at the level of cranium and mandible may indicate the initial position
of the cadaver in the feature, and the space of decomposition to a certain extent.
The identification of the original position of the head may be an important
aspect for the study of mortuary practices. It is important to understand if the
movement of rotation of the head is natural, that is, the result of taphonomic
processes, or if it is intentional, and potentially related to a mortuary practice. The
key observations are at the level of the temporomandibular (TM) joint and
mandible, and at the level of the cervical vertebrae (C1–C7), the latter described in
a separate section below.
Observation of the mandible
The TM joint is one of the first joints to decay during the process of decomposition (Hill 1979a–b).
In the archaeological feature, if the TM joint is detached, it indicates a
taphonomic rotation. If the body is laid on its back, after dissociation at the level
of the TM joint, the mandible tends to fall forwards while the cranium rolls in the
reverse way.
If the TM is not detached and the cranium and mandible are well articulated, it
indicates either a taphonomic or an intentional rotation. The relative position of
the mandible is an important element in the analysis, but not determinant, because
the maintenance of this element in anatomical articulation is not sufficient to argue
for the initial position of the head (Duday 2009, 19).
The analysis of the rotation of the head must be done with caution and based
on a multiple set of observations (see below, Observation of the cervical vertebrae).
Hyoid
The hyoid bone is particularly poorly represented in archaeological assemblages
(Mays 2010, 31). Observations of the hyoid when possible, offer relevant information about the space of decomposition of the cadaver. The hyoid bone tends to
collapse during decomposition. However, if its anatomical position is maintained,
it suggests an hourglass effect, indicating that the sediment could penetrate imme-
146
Methods: Archaeothanatology
diately and progressively fill the empty spaces created as soft tissue decayed (Duday
2009, 56, 57).
Vertebral column
The most relevant observations are at the level of the cervical vertebrae, and lumbar vertebrae to a certain extent. The rotation observed at the level of the thoracic
vertebrae may offer significant information about the initial position of the upper
body.
Observation of the cervical vertebrae (C1–C7)
Observations at the level of the cervical vertebrae (CV) may indicate the nature of
the deposit, the space of decomposition, and the initial position of the cadaver in
the feature.
The atlanto-occipital joint is a persistent joint and articulates the occipital bone
in the cranium with the first CV (C1, atlas). The remaining joints are labile: the
atlanto-axial joint (C1, atlas–C2, axis), and the intervertebral joints, in particular at
the level of C2–C3 and C3–C4 (Duday 2006, 35). Thus, the maintenance of the
CV in articulation is a strong indicator of the primary position of the remains in a
filled space of decomposition.
The rotation of the head usually involves all CV. In archaeothanatology it is particularly important to document the position of the superior portion of the cervical
vertebrae (Duday 2009, 18; Duday and Guillon 2003, 132). If these are in continuity and in anatomical connection while presenting a degree of rotation in conformity with the rotation of the head, the intentionality of the rotation is confirmed.
In some cases, observations of the formation of an arc with the CV well articulated (Nilsson Stutz 2003a, 272) indicate that the neck was initially flexed forwards
and downwards bringing the head towards the body in the feature. This movement
can occur if the head is placed in an elevated position, such as in a sitting or semisitting position, but even a moderate downslope may have an impact on this
movement. In cases like this, the maintenance of the CV in articulation can be
explained by the progressive and immediate infilling of sediment preventing the
collapse of the bone elements.
Observation of the lumbar vertebrae (LV)
The intervertebral joints at the level of the lumbar vertebrae (L1–L5) are persistent
joints, as well as the lumbosacral joint (L5–sacrum). Despite the late decomposition of these joints, they are commonly found disarticulated due to the effects of
gravity in relation to the initial position of the cadaver in the feature (Duday 2006,
35).
Observations at the level of the lumbosacral joint may reveal important aspects
of the initial position of the cadaver. In some cases an angle can be observed between the L5 and the sacrum indicating the movement forwards of the sacrum
followed by the lumbar vertebrae. Some case studies have shown that the migra147
Chapter 3
tion forwards of the sacrum appears to be associated with the descent and movement downwards of the vertebral column, which could be explained by the decomposition of a cadaver in a sitting or semi-sitting position where the upper body
is in a more or less vertical position (Nilsson Stutz 2003a, 269).
Thoracic cage: sternum and ribs
Often, very little is preserved from the thoracic cage (TC). This skeletal region is
particularly vulnerable to decay in the soil because of the relatively low density of
the ribs and due to the intense activity in this area during the decomposition
process promoting the dissolution of the hard tissues. Observations at the level of
the TC may reveal important aspects about the feature, such as the space of decomposition, the initial position of the cadaver, and other aspects related to the
grave feature.
One of the most frequent disarticulations observed during decomposition are
the collapse of the TC. The dissolution of the organs in the thoracic and abdominal cavities during decomposition leaves significant empty spaces resulting in the
collapse of the bones. Commonly, the initial volume of the TC is not maintained,
and can be outlined as follows:
• If the body is laid on its back, the ribs of both right and left hemithorax
will fall symmetrically towards the bottom of the feature. This displacement is known in archaeothanatology as the flattening of the rib cage
(Duday 2006, 34, 2009, 16).
• The hemithoraces may fall vertically if the body lies on a downslope or
in a sitting position. In a filled space, the ribs collapse downwards or
neatly vertically (Nilsson Stutz 2003a, 271).
• If the body is laid on one side, the ribs of the side on which the body
lies will remain in the same position, while the ribs of the opposed side
will collapse inside the thoracic void (Duday 2006, 34).
• Notably, the collapse of the TC may affect the movement of other
bones, such as the bones of the hands, or objects, or clothing elements
placed in front of the thorax or abdomen.
• In a few cases, the initial volume of the TC can be maintained (Duday
2006, 44). If the anatomical position of the ribs is maintained it indicates
one of the following cases:
o The sediment could penetrate immediately, and progressively fill the
empty spaces created as soft tissue decayed (Duday 2009, 56). This is a
rare occurrence because the TC is rich in soft tissue and a very active
region of the body during the process of decomposition; thus disarticulations at this level are very common, even in filled environments. The
maintenance of the thoracic volume is in this case explained by an
hourglass effect.
o Or, it may suggest the decomposition of the body in a narrow central
pit. This particular form of pit is elevated on the lateral sides where the
148
Methods: Archaeothanatology
upper limbs lie in a raised position relative to the thoracic cage. Thus,
the usual displacement of the thoracic cage is prevented by the bilateral
wall-effect supporting and keeping the ribs in equilibrium (Duday 2006,
44).
Bones of the shoulder girdle and upper limbs
Bones of the shoulder girdle
Observations at the level of the bones of the shoulder girdle (clavicles and scapulae) may offer relevant information about the grave feature. The key observations
involve three bone elements: the clavicle and scapula, and the humerus at the level
of the upper limb.
During the decomposition process the clavicles may become vertical or adopt a
steeply oblique position, with the acromial ends directed upwards and the sternal
ends rotated downwards. In archaeothanatology, this movement is termed the
verticalization of the clavicles (Duday 2006, 43; 2009, 45). This movement presents
strong evidence for constraint effects, indicated by a bilateral pressure at the level
of the shoulder girdle. In some cases, the constraint effect may influence only one
clavicle, indicating a particular pressure at that side of the feature. The transverse
compression can happen when the cadaver lies on its back in a narrow space, such
as a soft container (e.g., tight wrapping in a shroud) or a hard container (e.g., narrow coffin, anthropomorphic pit) (Duday 2006, 43; Duday and Guillon 2006, 144).
A clear case may be indicated by the observation of the projection upwards and
forwards of the shoulder girdle and the humeri. Typically, the clavicles are verticalized, the glenoid fossae are directed upwards and the axillary borders of the scapulae are vertical. Likewise, the humeri are projected upwards and forwards, rotated
inwards and located very close to the TC.
A different pattern can be observed when the scapula is completely exposed
presenting the anterior view. This phenomenon may indicate a combination of
movements, such as the rotation of the scapula upwards (even if moderate) while
the ribs slid towards the medial axis of the body exposing the anterior side of the
scapula (Nilsson Stutz 2003b, 3). This movement pattern suggests a transfer of
body weight towards the opposing side, which could indicate a sloping floor, even
if moderate.
Bones of the hands
Observations at the level of the bones of the hands may indicate aspects related to
the nature of the deposit, the space of decomposition and the initial position of the
cadaver in the feature.
The joints of the hands are labile and the articulated state of the bones of the
hands is a strong argument for the primary nature of the deposit, as well as for the
filled space of decomposition. When found in situ, at least some bone elements,
may aid in the reconstruction of the position of disturbed upper limbs (Nilsson
Stutz 2003b, 5).
149
Chapter 3
Bones of the pelvic girdle and lower limbs
Bones of the pelvic girdle
Observations at the level of the pelvic girdle may indicate the space of decomposition, the initial position of the cadaver and elements of the grave feature.
The joints at the level of the pelvic girdle, lumbosacral and sacroiliac joints are
persistent and some of the last to decompose. However, a common movement
during decomposition is the collapse of the bones of the pelvic girdle (Duday
2006, 34–35). Following the law of gravity, if the body is laid on its back, the
os coxae tend to collapse towards the rear. If the body is laid on one side, the
os coxae from the opposed side will collapse inside the pelvic cavity. These
movements are common even in filled environments because of the instability
caused by the decomposition of masses of soft tissues in the pelvic cavity and buttocks (Duday 2006, 35). If the body is laid on its back in a wide space, the collapse
of the bones of the pelvic girdle may be more pronounced, with a progressive
disarticulation at the level of the pubic symphysis until the os coxae lie towards the
bottom of the feature (Duday 2006, 40).
The collapse of the pelvic girdle can be prevented, even in empty spaces, by
bilateral pressure from a narrow container of soft or hard material (Duday 2006,
44, 2009, 41). The collapse of these bones can also be prevented, at least to some
extent, by the rapid penetration of fluid sediment, such as by an hourglass effect.
Observations at the level of the sacrum may indicate particular aspects of the
original position of the body. The rotation forwards and downwards of this bone
and following disarticulation at the lumbosacral joint, with the formation of an
angle between the L5 and the sacrum, may indicate a sloping floor and/or a body
in a sitting position (see above, Observations of the lumbar vertebrae).
Bones of the lower limbs
Patellae
Observations at the level of the patellae may offer relevant information about the
space of decomposition.
The position of the patellae is a diagnostic criterion for the space of decomposition and it is important to record their exact positions in the feature (Duday 2009,
35). During decomposition, as the os coxae become disarticulated, the femora tend
to rotate outwards causing the patellae to fall and collapse in an empty space
outside the knee joint (Duday 2009, 36). The movement of the patellae can be
more moderate and slid in either direction, dragged by decomposition fluids
constrained by the environment of the feature and by the initial position of the
cadaver (Duday 2009, 34). The position of the patellae in anatomical position at
the time of the excavation is a strong argument of a filled space of decomposition
(Nilsson Stutz 2003a, 265).
150
Methods: Archaeothanatology
Bones of the feet
Observations at the level of the bones of the feet may indicate aspects related to
the nature of the deposit and the space of decomposition.
The bones of the feet are articulated by persistent joints at the level of the ankle
and tarsal joints, and by labile joints on the distal portion of the feet. The maintenance in anatomical position of the distal portion of the feet is a diagnostic
criterion for the primary nature of the deposit and indicative of decomposition in a
filled environment (Duday 2009, 57).
Initial position of the cadaver: additional observations about constraint effects
Sediment and hypercontracted skeletons
The gradual penetration of sediment during the process of decomposition can
accentuate the closure of the angles between the bones of the limbs enhancing the
flexion of the joints (Duday 2006, 43, 47; 2009, 53–54). This constraint effect may
occur if the body lies in direct contact with the sediment and affects the joints in
flexion which become hypercontracted by the penetration of fine and fluid sediment.
Explanations such as ties and/or tight wrappings must be based on arguments
other than just the hyperflexion of the joints (Duday 2006, 43; 2009, 54).
Wrappings
Narrow grave features exert an obvious constraint effect on the cadaver. In some
cases, the observed constraining effect cannot be explained by the narrowness of
the feature, and may require alternative explanations such as the wrapping of the
body (see Nilsson Stutz 2003a, 299). Wrapping of the body is difficult to confirm if
direct elements such as the wrapping itself or items holding the wrapping are not
preserved. Several indicators affecting the distribution of the bones must be
addressed to confirm this hypothesis. The decomposition of a wrapped or tied
body will present a general wall-effect, which is typically particularly visible at the
level of the upper body, and can be outlined as follows (Nilsson Stutz 2006):
• the TC is compressed and presents a “narrow” appearance;
• the shoulder girdles are projected forwards and upwards and the
clavicles are verticalized;
• the upper limbs are rotated inwards and possibly upwards, even if only
slightly, and are placed in very close proximity to the TC.
While these indicators could be the effect of a narrow grave, the possibility of
wrapping should be carefully considered, in particular if the limits of the grave are
wider than the area occupied by the cadaver.
The grave features
This category consists of elements related to the physical appearance of the deposit
where the human remains were found. It includes aspects related to the size
(perimeter and depth), shape (e.g., oval, round, rectangle), and topography of the
151
Chapter 3
deposit, such as characteristics of the walls and the bottom of the feature. Aspects
related to the surrounding sediment and filling of the deposit may also be included
in this category.
One of the aims of the analysis of the grave features is the identification and
description of the structure(s) that contained the human remains at the time of the
deposit, i.e., the container. In archaeothanatology, the term container refers to any of
the possible ways to contain a cadaver, such as a simple pit dug out of the ground,
a shroud (see above, Wrappings), a wood roof, or a box such as a coffin or
sarcophagus, or even a chamber. Thus, one deposit may accommodate several
containers involving the cadaver. Archaeological observations, such remains of
wood, nails or other architectonic elements, as well as by the identification of the
limits of the grave pit, may allow the direct identification of the container(s). The
archaeothanatological analysis enhances the identification of such structures, even
without the direct observation of the remains, and also when the pit limits are
difficult to recognize in the field.
The characteristics of the container, where the remains are deposited, will influence the taphonomic history of the deposit, as well as the position of the skeletal
elements recovered in the archaeological context. In archaeothanatology, key
observations used in the reconstruction of the grave features are indirect indicators
showing the effect of the container on the organization of the skeleton (Duday
2006). This requires a multiple set of observations, in particular those related to
constraint effects, such as wall-effects or fluid penetration of sediment causing
hypercontraction. Likewise, indicators of collapse of bone elements are also valuable for the reconstruction of the grave features. Original empty spaces in the
grave feature can promote the scattering of the bone elements. Other elements,
such as structures placed behind the bodies (e.g., headrest, cushions, and wooden
platforms) can also be responsible for disorder of the bones, and potentially can be
identified through archaeothanatology.
Deposits containing more than one individual
Often, archaeological features with human remains contain more than one individual. Archaeothanatology is particularly useful in this kind of analysis because it
allows reconstruction of the original context and recognition of whether the deposits are synchronous (simultaneous or successive/consecutive) or asynchronous
(Duday 2009, 72).
The synchronous deposit of more than one individual is easily argued when the archaeothanatological analysis indicates that the corpses were in anatomical integrity
when they came in contact with each other (Duday 2009, 76). It is normally
difficult to determine if the deposits were made at the same time (simultaneous) or
within a short period of time (successive/consecutive), because even the labile
joints may remain intact for some days, or even weeks. The argument for a synchronous deposit (either simultaneous or successive/consecutive) becomes more
difficult to support if the bodies are not intertwined, making it hard to determine
the time gap between the depositions. Other observations, such as the common
152
Methods: Archaeothanatology
arrangement of the bodies (e.g., facing each other) may support the main argument, but are not conclusive.
The asynchronous deposit of more than one individual in a feature is the common
practice in collective burials (see definition below). The practice of reduction can be
included in this category (Duday 2009, 72), and entails the removal of the remains
of a previous deposit to give place to the deposit of a new individual (see below,
Post-depositional manipulations of the cadavers). In all these cases, the deposit of a new
individual is an independent episode from the previous deposition(s) in the feature.
Funerary contexts: one single structure
The synchronous or asynchronous deposit of human remains may take place in one
single structure. Deposits in one single structure may indicate different mortuary
practices and their description and definition should be explicit to avoid ambiguity
in the interpretation. Funerary contexts in one single structure are distinguished by
the number of individuals in the structure, one or more than one individual, and
most significantly, by the chronology of disposal of the human remains, synchronous or asynchronous. Accordingly, funerary deposits in one single structure can
be: individual, multiple, or collective deposits (table 3.7).
Table 3.7. Funerary contexts in one single structure.
Individual deposit
Multiple deposit
Collective deposit
one structure
one structure
one structure
one individual
two or more individuals
synchronous deposit (successive/consecutive)
two or more individuals
asynchronous deposit
e.g., one individual in a pit
e.g., two inds. buried synchronously in the same pit
e.g., family tomb
One single funerary structure defined for the mortuary remains of one individual is
an individual deposit. This structure can be as simple as a pit dug in the soil, or as
complex as a pyramid containing the remains of one individual. A container with
the cremated remains of one individual also belongs in this category.
The distinction between multiple and collective deposits is based on the chronology of disposal. These terms are often used interchangeably. However, these
expressions reflect specific cultural options that should be made clear by the use of
appropriate terminology. A multiple deposit contains the synchronous disposal
(successive or consecutive) of two or more individuals in one structure. In archaeothanatology, a multiple deposit such as a double burial, can be demonstrated
by showing that the cadavers “retained their anatomical integrity when they came
in contact with each other” (Duday 2009, 75–6). If the labile joints of one individual were disturbed by the placement of a new cadaver in the feature, it indicates
that the deposits were not synchronous. Multiple deposits can accommodate various individuals, deposited in a short time frame between each disposal, as in contexts of mortality crisis related to catastrophes, massacres or plagues, as long as it
153
Chapter 3
can be demonstrated that the cadavers were in anatomical integrity at the time of
the deposit, thus indicating a relatively short period between the disposals (Duday
2009, 98). Mortuary practices in a collective deposit follow a very different cadence. A
collective funerary structure is used over a more or less long period of time where
two or more individuals are deposited during separate events (Duday 2009, 105).
Passage graves, dolmens, and family tombs are good examples of collective
funerary deposits. These are single funerary structures successively re-opened for
the deposit of new individuals over a period of time.
Funerary contexts: more than one structure
After considering the different types of disposal of human remains in one single
structure, according to the number of individuals and time frame, we can focus on
contexts with more than one funerary structure, i.e., cemeteries. Essentially, a cemetery
accommodates a number of funerary structures regardless of the number of
individuals and tempo of disposal. In its simplest definition, a cemetery can be
defined as a funerary complex that accommodates more than one funerary structure, organized in a more or less complex manner (Duday 2009, 13). In modern
cemeteries it is not uncommon to find the three types of funerary deposits
described above: individual, multiple, and collective (fig. 3.14), but this seems to be
less common in earlier periods. As a last observation on terminology, I have
chosen to avoid the term necropolis and opted for the term cemetery instead, but
more often I use the term burial ground. Necropolis is a common term used in
archaeology and it has been used to cover a variety of categories where human
remains are found in archaeological contexts, including cemeteries. Besides its
vague definition, the term has an urban connotation derived from the Hellenistic
Greek language, the city of the dead, which may not be the most suitable for deeper
chronologies.
Figure 3.14. Terminology: a cemetery accommodates more than one funerary structure. A
cemetery may contain one or all three categories of funerary structures.
154
Methods: Archaeothanatology
Post-depositional manipulations of the cadavers
Post-depositional manipulations of the cadavers refer to the handling of human
remains after deposit in the primary feature (Nilsson Stutz 2003a, 309). This deliberate action is part of the funerary process in multiple episodes and can be identified by the retrieval of selected bone elements from a primary deposit or by the
identification of human remains in secondary position (see above, The nature of the
deposits). In the analysis of post-depositional manipulations it is important to keep
in mind the idea of intentionality of the practice (Duday 2009, 14, 89) which
should not be interpreted in the same way as the taphonomic or accidental disturbance of the remains. Likewise, the practice of reduction should not be considered in
this category because this gesture, while literally being a post-depositional manipulation, has a practical intention behind it, aimed at giving the necessary space to a
new deposit (Duday 2009, 72). Nevertheless, it may be relevant to discuss the
presence or absence of reductions when discussing mortuary practices as it may
provide relevant information about the spatial organization of the mortuary areas
and the space allocated for the dead.
Archaeothanatology in this study
Archaeothanatology emerged in the context of field archaeology and the application of the method requires detailed documentation of the human remains during
the process of excavation. The method lays great emphasis on the field situation
and it has been argued that assessment of the material may not be possible if key
observations are not documented in situ (Duday 2006, 30). Depending on the
quality of the field documentation the method of archaeothanatology may be
applied in post-excavation stages even when the archaeothanatological approach
was not carried out in the field. A thorough excavation and documentation of the
archaeological context with of the human remains, particularly when including key
observations about the spatial distribution of the bones and their tridimensional
registration, and high resolution graphic documentation showing diagnostic
elements of analysis, may allow the application of the method with few restrictions
(see e.g., Nilsson Stutz 2003a). Furthermore, recent studies (see e.g., Tõrv 2015),
including the present dissertation, consistently show that archaeothanatology has
great potential for the analysis of museum collections from old excavations that
otherwise would remain unexplored, although with acknowledged restrictions
(Tõrv and Peyroteo Stjerna 2014).
Source material: the challenges
The archaeological material selected for this study comes from historical excavations containing some of the most exceptional material from the European
Mesolithic. Early excavations impose important limitations to the analysis of the
material, not only because of the methods of excavation and registration but also
due to curation problems (see chapter 2).
155
Chapter 3
In this dissertation, the archaeothanatological analysis relies on graphic documentation (drawings and field photographs), and on the human remains preserved
in natural or artificial blocks. The written documentation (field notes, diaries,
reports, letters, published articles, and one monograph) was used as supplementary
documentation, except in the case of the analysis of the Moita do Sebastião where
the main source of information is the written documentation published in a monograph. The completeness and detail of each analysis is dependent on the character
and quality of the available source material, presented in detail for each site in
chapter 4.
Graphic documentation: drawings
The field drawings available for the Sado valley are not only more abundant but of
greater quality than the material available for the Tagus sites. This is because the
National Museum of Archaeology (MNA), which was in charge of the excavations
in the Sado valley, kept a professional artist on the staff, who very skilfully drew
most of these shell middens. The material dated to 1956 was sketched by A. Paiva,
and all documentation after 1958 by D. Sousa.
The few drawings available for the Tagus sites consist of very simple sketches
of the human remains, site plans, and site profiles. Most of this documentation is
published (Abrunhosa 2012; Cardoso and Rolão 1999–2000; Jackes and Alvim
2006; Roche 1965, 1967, 1972).
The human skeletons excavated in the Sado valley were individually drawn at a
scale of 1:20. The skeletons were completely exposed and cleaned for the task of
documentation, and if there were any other materials close to the skeleton, such as
lithics or shell beads, these were removed before the drawing of the bones in situ
(Dario de Sousa, pers. comm.). The drawings are in plan view and the measurements are accurate. In a few cases, the relative depth of some skeletal elements is
documented. The skeletons are also documented in the site plans and in the site
profiles, although in much less detail. The limitations of these drawings can be
outlined as follows:
• in some cases the sketched view is somewhat distorted, presenting a mix
of plan and profile views;
• there is a common lack of bone landmarks, possibly due to limited
knowledge of human anatomy, making it difficult to identify the side of
appearance of the bone elements;
• despite the realism and attention to detail presented in the drawings, it
seems that in some cases the artist opted to represent the bone elements
in articulation, when in fact they were not (e.g., articulations at the level
of the pelvic girdle).
Other drawings available for the Sado sites include site plans and profiles with and
without the human skeletons, and detailed drawings of concentrations of faunal
remains.
156
Methods: Archaeothanatology
Graphic documentation: field photographs
Most of the few field photographs available for the Tagus sites are published
(Abrunhosa 2012; Cardoso and Rolão 1999–2000; Cartailhac 1886; Ribeiro 1884;
Jackes et al. 2014; Roche 1972). The material available for the Sado valley is in the
archives of MNA and is in general poorly preserved. Overall, the field photographs
are not abundant and are in general of poor quality, presenting several limitations:
• most photographs are not in plan view, resulting in distortion of depth
and of the relative position of the skeletal elements;
• the photographs have no scale;
• the photographs present a general view of one or several individuals and
the lack of close-up views restricts access to diagnostic skeletal elements; also in most cases it is unclear if the indistinctness of diagnostic
elements is due to preservation, disturbance and/or exposure;
• bones that are underneath others, or covered by sediment, do not appear;
• the common size for the Sado valley photographs is c 12 x 8 cm presenting one whole skeleton, or several individuals;
• all photographs are in black and white making it difficult to distinguish
colouration of the sediments.
These limitations introduce a significant uncertainty to the analysis. Nevertheless,
the drawings and photographs when available are valuable documents and provide
information that would not be available otherwise. The strategy of analysis is to
combine and intersect the documentation, without relying on one single source.
Human remains
One particularity of these collections is the maintenance of the human remains in
natural blocks of carbonate matrix or artificial preservation in blocks of paraffin
wax. The material from the Tagus sites is normally heavily covered with a hard
matrix of sediment and shells and in several cases, the human remains were kept
within this cement, at least partially. Also, throughout the twentieth century, one of
the most common procedures in recovering the skeletons was to cover them with
paraffin wax and transport them in blocks to the museum. This procedure restricts
a series of bioanthropological observations (Cunha and Umbelino 1995–97, 167),
as well as chemical analysis, but it is very conducive to archaeothanatology, since it
provides access to the remains as they were when excavated decades ago. Unfortunately, in the 1980s, an anthropologist with a different research agenda removed
the wax from most of the blocks curated at MNA, and as a consequence most
material is removed from its context.
The material maintained in blocks is extraordinary and should be preserved.
However, this apparently perfect source material can be deceptive. The human
remains preserved in blocks are not completely excavated and several bones and
diagnostic elements, remain within the sediment, or underneath other bones, and
157
Chapter 3
are also covered by wax. Another problem when working with material preserved
in blocks, is that while the wax or sediment hides certain features, the block itself
emphasizes other elements, and this distortion if not critically addressed can lead
to misinterpretations. Here again, I would like to emphasize that when working
with material from historical excavations, we should not rely on one source – even
if that source is the block with the human remains as they were when originally
excavated.
Archaeothanatology in the lab: the strategies
Archaeothanatological analysis from historical sources is challenging but possible
(Tõrv and Peyroteo Stjerna 2014). The method has no cultural or chronological
boundaries, and can be applied both in the field and in the laboratory.
One strategy is to cross-reference the different sources critically and to combine
multiple observations in order to demonstrate the validity of the interpretation.
When applying archaeothanatology to post-excavation documentation, the analysis
should not rely on one kind of data, even if apparently very complete.
A second strategy is always to refer to comparative material and similar patterns
published in the literature. This is a valuable way to compare and test observations
with reference material fully excavated according to the principles of the method.
Unfortunately, this is not always possible either because of the lack of case studies
available, or due to the singularity of the patterns under investigation. Following
this, in order to carry out more precise analyses, more comparative material is
needed, ideally based on common protocols and terminology. Comparative data
must be explicit, clear, and published. This is a fundamental strategic link between
work in the field and future possibilities in the laboratory. It is based on a growing
body of knowledge that Archaeothanatology in the lab can proceed as a strong and
reliable method for the study of mortuary practices from the almost unlimited
source material that has been stored for decades in our museums (Tõrv and
Peyroteo Stjerna 2014).
A third strategy is to understand the context of data production, allowing the
informed and necessary critical assessment of the documentation. Access to
archives and original documentation is the closest we can get to the fieldwork
situation. Knowledge about the general context of the excavation may be as important as the context of the human remains (see Tõrv 2015).
Archaeothanatology offers a step-by-step, observation-based argument to confirm or disprove hypotheses. When carrying out this type of analysis, all apparently
small pieces of information can be revealing. Presumed good sources should not
overshadow other sources that at first seem to be less informative. Archaeothanatology in the lab must be done through multiple filters while cross-referencing with
published and well described reference material. This way, it is possible to achieve
sufficient resolution for the archaeothanatological analysis of historical collections.
Interpretations of the material may remain open, possibly to be resolved with new
data or further developments of the method. In fact, this is not a limitation of the
method, but an opportunity and a research challenge (Tõrv and Peyroteo Stjerna
158
Methods: Archaeothanatology
2014). Despite the limitations, archaeothanatology is certainly one of strongest and
most reliable methods to assess, study, and retrieve new valuable data from the
extensive archaeological assemblages at our disposal in museums all over the
world.
159
Chapter 4
Results
Introduction
This chapter is organized by archaeological site starting with Moita do Sebastião
and Cabeço da Arruda in the Tagus valley, followed by Arapouco, Cabeço das
Amoreiras, Vale de Romeiras, Cabeço do Pez, and Várzea da Mó, in the Sado
valley. The sites are introduced by the Site background: a short history of research containing a synthesis of basic elements about the site. This section is followed by the
presentation of the results of this dissertation, which is organized in three parts,
according to the three methodologies used in this study: radiocarbon dating and
Bayesian analysis, stable isotope analysis of carbon and nitrogen, and archaeothanatology. The
results are analysed, contextualized and synthesized in chapter 5.
In the section Chronology of the burial activity, I present the results of the 14C dating
programme, calibration and statistical treatment of all 14C dates on human bone
collagen available for each site. In this section, I present:
• Previous 14C measurements
• Objectives of the dating programme (macro-goals related to the general
aims of this dissertation and micro-goals which are site-specific)
• Sampling strategy
• Results
• Calibration and chronological models
The chronological data is followed by the presentation of the results obtained for
the stable isotopic analysis of carbon and nitrogen of human bone collagen, which can be
outlined as follows:
• Previous measurements: δ13C and δ15N
• Sampling strategy
• Results
In the third section, I present the results of the archaeothanatological analysis. For each
central aspect analysed I give the whole set of results summarized in a table.
Several burial contexts are described in detail (e.g., burials at Arapouco, and at Vale
de Romeiras in the Sado valley) as illustrative examples depending on the quality of
the source material of each site. The presentation of the results is organized
according to the following categories:
161
Chapter 4
•
•
•
•
•
•
•
162
The nature of the deposits
The space of decomposition of the cadaver
The initial position of the cadaver in the feature
The grave features
Deposits containing more than one individual
Post-depositional manipulations of the cadavers
The mortuary programme
Tagus valley: Moita do Sebastião
Site background: a short history of research
Moita do Sebastião (also known as Fonte da Burra) is a shell midden site located
on a low terrace on the left bank of the tributary Muge, c 4 kilometres from the
confluence with the Tagus River. The site has excellent access and visibility
towards the river and towards Cabeço da Arruda on the opposite margin. The
neighbouring site of Cabeço da Amoreira is just c 700 metres east (fig. 2.2).
The site was first excavated in the nineteenth century by the Geological Commission: in 1880 (Ribeiro 1884), and in 1884–85 (Paula e Oliveira 1888–92).
Dozens of human skeletons were excavated (table 4.1), and most of them, if not
all, were found in the basal terrace sands (Jackes and Alvim 2006). This assemblage
is curated at the Geological Museum in Lisbon (MG).
The next and last seasons of excavations were directed by J. Roche and O. V.
Ferreira, between 1952 and 1954, after the destruction of the top layers, which
were estimated to be at c 2.5 metres from the bottom (Roche and Ferreira 1967).
The 1950s excavation area was presumably contiguous to the areas excavated in
the nineteenth century (Alvim 2009–10) (fig. 4.1). In these three years, 34 graves
(sépultures, Roche 1972) were registered, and some of them contained the remains
of more than one individual (Ferembach 1974; Roche 1972). The 1952–54 collection is curated at the Museum of Natural History, University of Porto (MHNUP).
Most human remains from Moita do Sebastião are preserved in disarticulated
state. A few individuals are maintained in blocks of carbonate matrix, at least
partially, including two individuals excavated in the nineteenth century.
Table 4.1. Estimated number of individuals excavated each year at Moita do Sebastião, Tagus valley.
Year of excavation
Number of individuals
Reference
1880
MS + CAR, MNI 120
Ribeiro 1884
1884
13
Paula e Oliveira 1888–92
1885
39
Paula e Oliveira 1888–92
1952–53
MNI 27 (sépultures 1–27)
Cardoso and Rolão 1999–2000; Roche 1972
1954
MNI 7 (sépultures 28–34)
Roche 1972
Total recovery, MNI = 86 + unknown number for 1880
163
Chapter 4
Figure 4.1. Moita do Sebastião, Tagus valley: excavation areas with human skeletons. The
human remains excavated in 1952–54 are identified in grey circles, A, B, and C. The areas
excavated in the nineteenth century are identified by circle (1880) and ellipse (1884–85), but
the exact location of the human burials is unknown. (Reproduced from Alvim 2009–10, 44,
fig. 13).
The MNI calculated for the assemblages at both MG and MHNUP is 85 (Jackes
and Meiklejohn 2008), which is lower than what was probably recovered in the
field (table 4.1). Ferembach (1974) reported the analysis of 136 individuals: 96
individuals curated in Lisbon (MG) and 40 in Porto MHNUP (table 4.2). The
number of individuals recovered at Moita do Sebastião was certainly higher than
85, and as previous authors have pointed out it is likely that some osteological
material may not have survived in the current collections (Jackes and Alvim 2006).
Table 4.2. The various MNI estimates proposed for Moita do Sebastião, Tagus valley.
Estimated MNI
Collections
Reference
136
MG and MHNUP
Ferembach 1974
79
MG and MHNUP
Jackes et al. 1997
85
MG and MHNUP
Jackes and Meiklejohn 2008
42
MHNUP
Cunha and Cardoso 2002–03
164
Results: Tagus valley, Moita do Sebastião
Most original written documentation available for Moita do Sebastião is from the
excavations of 1952–54. These belong to a private collection but are published
(Cardoso and Rolão 1999–2000), at least partially. This documentation consists of
field notes in a diary format handwritten by O. V. Ferreira. The field notes that J.
Roche certainly wrote are presumably destroyed (see chapter 3, A note on the
historical documentation). The graphic documentation is also published (Cardoso and
Rolão 1999–2000; Jackes et al. 2014; Roche 1972) and consists of site plans, a few
field photographs, and simple sketches of the human remains. A few short notes
from the nineteenth century have been identified and published (Jackes and Alvim
2006).
Among the middens of the Tagus and Sado valleys, Moita do Sebastião is the
only site with a dedicated monograph, published in two volumes. The first volume
is centred on the archaeological data (Roche 1972), and the second volume
presents the anthropological analysis of all human remains excavated at the site
(Ferembach 1974). These volumes contain detailed descriptions of the grave
features, and a site plan with the location of the human remains.
Chronology of the burial activity
Previous 14C measurements
In 1954, J. Roche collected and submitted a sample of charcoal from Moita do
Sebastião for 14C dating (Roche 1957). The charcoal was collected under the shell
layers in the bottom level of the midden. This was the first 14C measurement
(Sa–16, 7350 ± 350 BP) obtained from any site in Portugal, and was analysed in
1957 at the Laboratoire d’Electronique Physique du Centre d’Etudes Nucléaires de Saclay in
France (Roche 1957, 1965, 160). Some years later, a second sample on charcoal
(H–2119/1546), from the same excavations was submitted for 14C dating
(Delibrias and Roche 1965) (table A3).
The first 14C measurements on human bone from Moita do Sebastião (TO–131,
TO–132, TO–133, TO–134, and TO–135) were carried out in the 1980s (Lubell
and Jackes 1985; Lubell et al. 1994; Meiklejohn et al. 1986), collected from material
excavated in the nineteenth century, and selected according to the different enveloping sediments. The results were consistent, with the exception of TO–135 (individual CT), which revealed a more recent date in comparison with the other samples (Lubell and Jackes 1985). The measurements were done by AMS at the
Isotrace Laboratory of the University of Toronto, Canada, and independently
measured for stable isotopes. The individual results for the C:N ratio were not
published but were reported to be within the accepted quality range of 2.9–3.6
(Lubell et al. 1994, 204).
The human remains recovered during J. Roche’s excavations in 1952–54
remained undated until individual 16 was submitted for dating in the early 2000s
(Beta–127499, Cunha and Cardoso 2002–03; Cunha et al. 2003). Unfortunately, the
only δ13C value available for this skeleton was obtained by AMS (Cláudia
165
Chapter 4
Umbelino, pers. comm.) which cannot be used for dietary reconstructions or reservoir corrections (Millard 2014, 557; Taylor and Bar-Yosef 2014, 117). For this
reason, this sample is excluded from the 14C data set until IRMS-based δ13C values
are available (see chapter 3, Reliability of the isotopic measurements).
Objectives of the dating programme
Most 14C measurements available from the historical collections of the Tagus
middens are from samples from Moita do Sebastião. Before this radiocarbon
dating programme, six human skeletons were directly dated: five from the nineteenth century material and one from 1952–54, the latter , as discussed, with a
problematic calibration.
The general aims of the dating programme for Moita do Sebastião followed the
macro-goals of the radiocarbon programme of this dissertation outlined in chapter
3, and were:
• to determine the duration of use of this site for burial activity;
• to determine when the burial activity started and when it ended;
• to estimate the frequency of the burial activity in the site.
The site-specific aims of the dating programme for Moita do Sebastião were:
• to date the three main burial areas identified in 1952–54;
• to establish their chronological relationship to each other and to the
nineteenth century collections.
Sampling strategy
Fourteen samples of human bone were selected for 14C analysis (table 4.3). The
sampling strategy was focused on the material excavated in 1952–54 which,
contrary to the nineteenth century collection, is plotted (fig. 4.2) and well
described. These human remains were excavated from one layer at the bottom of
the shell midden and were found clustered in three main areas (A, B, and C)
(Roche 1972).
Sample pre-selection was based on the analysis of J. Roche’s site plan, and was
targeted towards burials in the three main clusters, as well as individuals found in
relatively isolated positions. The sampling was constrained for the preservation of
the collection. One sample only was obtained from area B (MS1952–54, Sk18), and
just two samples from area C (MS1952–54, Sks 22, 25), one of which was outside
the main cluster (Sk22). Nevertheless, it was possible to get a good spatial coverage
of the whole burial area, in particular of the largest concentration of burials located
in area A, where only one skeleton was previously dated (Beta–127499, MS1952–
54, Sk16).
166
Lab no.
Ua–46263
Ua–47977
Ua–46264
Ua–47978
Ua–47979
Ua–46265
Ua–47980
Ua–46266
Ua–46267
Ua–46268
Ua–47981
Ua–46269
Ua–46270
Ua–46271
Ind.
1
5
9
10*
14
15*
17
18
22
25
30
31
33
34
Adult, old
Adult
Adult
Non-adult,
c 2 yrs
Non-adult,
c 1 yr
Adult
Adult
Adult,
25–35 yrs
Adult
Adult,
35–50 yrs
Adult
Adult
Adult
Adult
Age
Bone
♂
♀
♀
Tibia–R
Tibia–R
Tibia–R
Humerus–R
Tibia
n/d
♂
Femur–R
Tibia–R
Femur–R
Tibia–R
Tibia–L
Tibia–L
Tibia–L
Long bone–R
Tibia–L
n/d
♂
♂
♂
♂
♀
♂
♂
n/d
Sex
Area A,
outside cluster
Area A
Isolated
Area A
Area C,
outside cluster
Area C
Area B
Area A
Area A
Area A
Area A
Area A
Area A
Area A
Site plan
–
–
Previous stable isotope analysis of same bone (Umbelino 2006).
–
–
Ind. 19 was the first choice but was not possible to sample because it is on
display. Ind. 20 was not sampled because it is well preserved in a block.
–
–
Previous stable isotope analysis of femur–R (Umbelino 2006).
Doubtful quality.
Previous stable isotope analysis of the same bone (Umbelino 2006).
–
The feature with ind. 1 had the remains of two adults: one male and one female
(Ferembach 1974; Roche 1972).
Ind. 3 was the first choice because of its central position in the burial area, but
was not possible to sample.
–
Obs.
Table 4.3. Moita do Sebastião, Tagus valley: samples of human bone (14) collected and submitted for 14C measurement in the course of this study. All samples are from
contexts excavated in 1952–54 in the bottom layer. This material is curated at MHNUP. Age and sex estimation by *C. Umbelino (2006), otherwise by D. Ferembach
(1974).
Chapter 4
Figure 4.2. Moita do Sebastião, Tagus valley, 1952–54: excavated area with human remains
and individuals selected for 14C measurements. Burial features were numbered from 1 to 34
(Roche 1972, 116, fig. 29). Samples selected for 14C measurements are indicated by rounded
rectangles. Individual 14 was selected for dating but the results were of doubtful quality.
Results
Nineteen 14C measurements on human bone collagen samples from Moita do
Sebastião are available now, one of which (Beta–127499, MS1952–54, Sk16), as
discussed, has a problematic calibration and was excluded from the present study.
Five samples belong to human remains excavated in the nineteenth century and 14
samples, including Beta–127499, belong to human remains excavated in 1952–54
(table 4.4). One of the 14 samples (Ua–47979, individual 14) submitted as part of
this study revealed poor quality during laboratory analysis and was rejected (table
A1).
168
1952–54
MHNUP
1952–54
MHNUP
1952–54
MHNUP
1952–54
MHNUP
1952–54
MHNUP
1952–54
MHNUP
30
31
1
9
18
22
Ua–46267
Ua–46266
Ua–46264
Ua–46263
Ua–46269
Ua–47981
Ua–46265
1952–54
MHNUP
15
7120±43
7095±45
7621±50
7483±48
7141±40
7058±44
6986±40
–16.9
–16.9
–17.0
–17.0
–17.4
–17.4
–17.9
–18.2
Ua–47978
1952–54
MHNUP
10
6753±46
Measurements
14C by AMS; δ13C and δ15N by IRMS
14C Age BP
Lab no.
δ13C (‰)
VPDB
Ua–46270
6743±44
–18.4
Identification of the
human remains
Ind.
Excav.
Museum
33
1952–54
MHNUP
12.8
10.8
10.8
10.5
10.6
9.8
10.2
9.0
δ15N (‰)
AIR
10.5
3.2
3.1
3.2
3.2
3.2
3.2
3.2
43.8
43.8
42.9
42.9
39.0
39.0
34.3
31.4
29.5
3.2
3.4
M1
C:N
38.8
38.8
37.5
37.5
32.5
32.5
26.3
22.5
20.0
M2
% marine (±10)
This study
This study
This study
This study
This study
This study
This study
This study
This study
Reference
continued…
Table 4.4. Human bone collagen samples (19) from Moita do Sebastião, Tagus valley: 14C measurements and estimation of non-terrestrial carbon intake (% marine).
Samples are sorted so that the lowest marine percentage is on the top of the table. Marine percentage was estimated from the measured δ13C value of each sample by the application of method marine 1 and method marine 2, based on adopted endpoint values of –12‰ and –20‰ for marine (100%) and terrestrial (100%) diets respectively.
TO–132
TO–134
Ua–46268
TO–131
TO–135
1952–54
MHNUP
19th c.
MG
19th c.
MG
1952–54
MHNUP
1952–54
MHNUP
19th c.
MG
19th c.
MG
1952–54
MHNUP
17
24
41
34
25
22
CT
16
Beta-127499
Ua–46271
Ua–47980
TO–133
19th c.
MG
29
Ua–47977
1952–54
MHNUP
5
Table 4.4. continued
7120±40
6810±70
7240±70
7243±45
7236±41
7160±80
7180±70
7105±42
7200±70
7138±42
AMS:
–16.8
–15.3
–16.1
–16.2
–16.3
–16.7
–16.8
–16.8
–16.9
–16.9
n/a
13.4
12.2
14.0
11.9
11.2
11.9
9.5
10.4
9.3
n/a
n/a
n/a
3.2
3.2
n/a
n/a
3.1
n/a
3.1
–
59.0
51.4
50.5
49.5
45.7
44.8
44.8
43.8
43.8
–
58.8
48.8
47.5
46.3
41.3
40.0
40.0
38.8
38.8
Cunha and Cardoso 2002–03;
Cunha et al. 2003
Lubell and Jackes 1985; Lubell et al. 1994;
Meiklejohn et al. 1986
Lubell and Jackes 1985; Lubell et al. 1994;
Meiklejohn et al. 1986
This study
This study
Lubell and Jackes 1985; Lubell et al. 1994;
Meiklejohn et al. 1986
Lubell and Jackes 1985; Lubell et al. 1994;
Meiklejohn et al. 1986
This study
Lubell and Jackes 1985; Lubell et al. 1994;
Meiklejohn et al. 1986
This study
Results: Tagus valley, Moita do Sebastião
Calibration and chronological models
The chronological model was built on 18 14C dates, 13 of which are new dates
obtained in the scope of this study, and five were previously published.
This is a phase model which groups all dated burial events in a phase of burial
activity. As recommended (Bronk Ramsey 1998, 2000, 2001) and discussed in
chapter 3, this grouping is constrained by a start of burial activity event, and an end of
burial activity event (i.e., boundaries). This calibration approach imposes no internal
constraints to the relative order of the dated events, which is particularly useful
when these are from one stratigraphic set. All it assumes is that the start event
occurs before the dated events, and that these all occur before the end event. For
the command files defined for this model see appendix.
Internal consistency and reliability of the model
Agreement index (A) for each dated event is good (fig. 4.3), except for the sample
Ua–46264 (individual 9) which is in poor agreement (Marine 1: A=38.3%; Marine
2: A=38.2%) but in good convergence (Marine 1: C=98.4%; Marine 2: C=97.9%).
This sample, Ua–46264 (individual 9), is the earliest dated burial in the site and is
not an outlier to the model, rather, its greater age may suggest an earlier event relatively older than the group of events in the considered phase. The most recent
burial event dated in the site (TO–135, individual CT) shows a satisfactory agreement index (Marine 1: A=80%; Marine 2: A=75.1%) but which is lower than the
remaining events in the group (A>100%). This sample is not in poor agreement,
but its relatively lower index and recent age may suggest a more recent event,
relatively later than the main grouping.
The model shows satisfactory overall agreement (Marine 1: Aoverall=81.9%;
Marine 2: Aoverall=81.7%), indicating that the 14C dates are in accordance with the
prior information incorporated in the model, which assumes one phase of burial
activity and does not impose a relative order to the dated events.
The results obtained when running the model with the different marine values
(M1, M2) are statistically insignificant (table 4.5). For this reason, all further observations and graphic presentation are based on the results obtained with the model
run with the values of M1. The results obtained with M2 are presented in the
tables.
The chronological model for Moita do Sebastião, calibrated with M1 values, estimates the start of the burial activity to have been in 6365–6051 cal BCE
(95% probability; start of burial activity) and its end to have been in 5603–5351 cal BCE
(95% probability; end of burial activity) (fig. 4.4; table 4.6). The burial activity is
estimated to have continued for between 469 and 816 years (95% probability: site use for
burial activity), or 523 and 696 years (68% probability) (fig. 4.5; table 4.6).
171
Chapter 4
Figure 4.3. Chronological model for the burial activity at Moita do Sebastião, calibrated
using marine 1 diet values. For each of the dates two distributions have been plotted, one
in outline which is the result produced by the scientific evidence alone, and a solid one
which is based on the chronological model used. Calibrated ranges of 95% and 68% probability are given by the lower and higher square brackets respectively. The large square
brackets (left) and the OxCal keywords define the overall model. Model-definition command files in appendix, Tagus valley, Moita do Sebastião, Model 1, Marine 1.
172
Results: Tagus valley, Moita do Sebastião
Figure 4.4. Start and End of burial activity at Moita do Sebastião derived from the chronological model, calibrated with values for marine 1 diet. Posterior density estimated for the
start of the burial activity is 6365–6051 cal BCE (95% probability) or 6246–6088 cal BCE
(68% probability). Posterior density estimated for the end of the burial activity is 5603–5351
cal BCE (95% probability) or 5557–5437 cal BCE (68% probability). Calibrated ranges of 95%
and 68% probability are given by the lower and higher square brackets respectively.
Figure 4.5. Duration of burial activity at Moita do Sebastião derived from the chronological
model, calibrated with values for marine 1 diet. Posterior density estimated for the duration
of the burial activity is between 469 and 816 years (95% probability), or 523 and 696 years (68%
probability). Calibrated ranges of 95% and 68% are given by the lower and higher square
brackets respectively.
173
Chapter 4
Table 4.5. Human bone collagen samples (18) from Moita do Sebastião, Tagus valley: calibrated date
ranges according to two different approaches for the calculation of the marine percentage (M1, M2). Results
are presented in both (a) cal BCE and (b) cal BP in 95.4% confidence range. The calibrated date ranges
are probability distributions derived from scientific dating alone. The posterior density estimates (in italic)
are derived from Bayesian modelling. Samples are sorted so that the most recent dates are at the top of the
column.
(a) Cal BCE
Age BP
Unmodelled
cal BCE, M1
Calibrated date range
(95% confidence)
Modelled
Modelled
cal BCE, M1
cal BCE, M2
Posterior density estimate
(95% probability)
5640–5390
5640–5474
5667–5487
Ind.
Lab no.
14C
33
Ua–46270
6743±44
CT
TO–135
6810±70
5616–5305
5642–5421
5656–5429
10
Ua–47978
6753±46
5644–5391
5644–5473
5670–5484
15
Ua–46265
6986±40
5846–5576
5843–5611
5876–5635
30
Ua–47981
7058±44
5890–5631
5887–5632
5964–5648
18
Ua–46266
7095±45
5902–5632
5896–5636
5968–5657
17
Ua–47980
7105±42
5902–5636
5898–5639
5969–5661
22
Ua–46267
7120±43
5968–5652
5965–5655
5977–5685
5
Ua–47977
7138±42
5973–5671
5972–5672
5979–5714
31
Ua–46269
7141±40
5980–5715
5979–5718
5986–5734
34
Ua–46271
7236±41
5994–5731
5991–5734
6003–5748
41
TO–134
7160±80
5991–5646
5992–5652
6007–5668
24
TO–132
7180±70
6001–5676
5997–5681
6017–5707
25
Ua–46268
7243±45
6001–5733
5998–5736
6005–5741
29
TO–133
7200±70
6021–5701
6017–5707
6031–5721
22
TO–131
7240±70
6025–5701
6021–5705
6026–5712
1
9
Ua–46263
Ua–46264
7483±48
7621±50
6336–5996
6426–6111
6218–5998
6292–6023
6235–6010
6322–6046
Unmodelled
cal BP, M1
Calibrated date range
(95% confidence)
Modelled
Modelled
cal BP, M1
cal BP, M2
Posterior density estimate
(95% probability)
7589–7339
7589–7423
7616–7436
(b) Cal BP
Ind.
Lab no.
14C
Age BP
33
Ua–46270
6743±44
CT
TO–135
6810±70
7565–7254
7591–7370
7605–7378
10
Ua–47978
6753±46
7593–7340
7593–7422
7619–7433
15
Ua–46265
6986±40
7795–7525
7792–7560
7825–7584
30
Ua–47981
7058±44
7839–7580
7836–7581
7913–7597
18
Ua–46266
7095±45
7851–7581
7845–7585
7917–7606
17
Ua–47980
7105±42
7851–7585
7847–7588
7918–7610
22
Ua–46267
7120±43
7917–7601
7914–7604
7926–7634
continued...
174
Results: Tagus valley, Moita do Sebastião
Table 4.5b. continued
5
Ua–47977
7138±42
7922–7620
7921–7621
7928–7663
31
Ua–46269
7141±40
7929–7664
7928–7667
7935–7683
34
Ua–46271
7236±41
7943–7680
7940–7683
7952–7697
41
TO–134
7160±80
7940–7595
7941–7601
7956–7617
24
TO–132
7180±70
7950–7625
7946–7630
7966–7656
25
Ua–46268
7243±45
7950–7682
7947–7685
7954–7690
29
TO–133
7200±70
7970–7650
7966–7656
7980–7670
22
TO–131
7240±70
7974–7650
7970–7654
7975–7661
1
9
Ua–46263
Ua–46264
7483±48
7621±50
8285–7945
8375–8060
8167–7947
8241–7972
8184–7959
8271–7995
Table 4.6. Posterior density estimates for the dates of archaeological events (Start/End) and the duration
of burial activity at Moita do Sebastião. Data derived from model described in fig. 4.3
Modelled
cal BCE, M1
Modelled
cal BCE, M2
Modelled
cal BCE, M1
Modelled
cal BCE, M2
Distribution
Post. density estimate (68% probability)
Post. density estimate (95% probability)
Start of burial activity
6246–6088
6274–6101
6365–6051
6390–6068
End of burial activity
5557–5437
5580–5459
5603–5351
5613–5366
Span of burial activity,
duration
Overall model
agreement (Aoverall)
523–696 yrs
515–706 yrs
469–816 yrs
469–828 yrs
81.9%
81.7%
87.9%
81.7%
Stable isotopes: carbon and nitrogen
Previous measurements: δ13C and δ15N
The first measurements of stable isotopes (δ13C, δ15N) on human bone from Moita
do Sebastião were carried out in the 1980s by H. Schwarcz at the McMaster
University in Canada (TO–131, TO–132, TO–133, TO–134, and TO–135). The
precision of analysis for δ13C is ± 0.1‰ and for δ15N is ± 0.2‰ (Lubell et al. 1994,
204). These five samples were also independently measured for 14C dating (see
above).
In the early 2000s, three samples from the 1950s excavations (Sks 10, 15, 31)
were analysed in the scope of a palaeodietary study (Umbelino 2006). Like the first
measurements, these analyses were carried out at the McMaster University and the
precision of the results are ± 0.1‰ for δ13C and ± 0.2‰ for δ15N (Umbelino
2006, 208–9). The C:N ratios were not provided but the laboratory confirmed the
quality of the collagen and of the measurements (Umbelino 2006, 315). The δ13C
value published for MS1952–54, Sk16 (Beta–127499) along with its 14C date
(Cunha and Cardoso 2002–03; Cunha et al. 2003) was obtained via AMS and not
by IRSM (Cláudia Umbelino, pers. comm.). This sample was not included in
175
Chapter 4
Umbelino’s palaeodietary study (2006), and as previously discussed it should not be
used for dietary reconstructions and/or reservoir corrections.
All previous measurements were included in this study (n=8), as long as they
fell within the accepted quality ranges as discussed.
Sampling strategy
In this study, the selection of samples of human remains for δ13C and δ15N measurements was fully dependent on the goals of the radiocarbon dating programme
(see above).
Results
Samples Ua–number (n=13) were measured by IRSM and the precision of the
measurements for both δ13C and δ15N is ± 0.1‰.
MS1952–54, Sks 10, 15, and 31 have multiple measurements (this study;
Umbelino 2006) and present different values even when the same bone was sampled (Sks 10, 31) (table 4.7).
At Moita do Sebastião, the human collagen δ13C values (n=21/no. individuals=18) range from –18.4 to –15.3‰ with a mean value of –16.9 ± 0.7‰. The
human collagen δ15N values (n=20/no. individuals=18) range from +9.0 to
+14.0‰ with a mean value of +11.1 ± 1.3‰ (fig. 4.6; table 4.7). Measurement of
δ15N was not possible for the sample of MS1952–54, Sk15 (Umbelino 2006).
16
15
14
13
δ15N (‰)
12
11
10
9
8
7
6
-22 -21 - 20
-19 -18 -17 -16 -15
δ13C (‰)
-14 -13 -12
Figure 4.6. Moita do Sebastião, Tagus valley: correlation between the δ13C (‰) and δ15N
(‰) values of 20 samples of human bone collagen from 18 individuals. Individuals 10
(triangle) and 31 (circle) have two sets of values.
176
Tibia–R
♀
♂
♂
♂
♀
♀
Adult,
35–50 yrs
Adult,
25–35 yrs
Adult,
25–35 yrs
Adult
Adult
Adult
Adult
10
1952–54
15
1952–54
15
1952–54
30
1952–54
31
1952–54
31
1952–54
1
1952–54
n/d
Tibia–L
♀
Adult,
35–50 yrs
10
1952–54
Tibia–L
Tibia–R
Tibia–R
Humerus–R
Femur–R
Tibia–L
Tibia–R
♀
Adult
33
1952–54
Bone
Sex
Age
Ind.
Identification of the human remains
Ua–46263
n/a
Ua–46269
Ua–47981
n/a
Ua–46265
n/a
Ua–47978
7483±48
analysis for isotopes
7141±40
7058±44
analysis for isotopes
6986±40
analysis for isotopes
6753±46
–17.0
–16.7
–17.4
–17.4
–16.2
–17.9
–16.6
–18.2
10.5
11.2
10.6
9.8
n/a
10.2
11.5
9.0
Measurements
14C by AMS; δ13C and δ15N by IRMS
Age BP
Lab no.
δ13C (‰) δ15N (‰)
VPDB AIR
Ua–46270
6743±44
–18.4
10.5
3.2
n/a
3.2
3.2
n/a
3.2
n/a
3.4
3.2
C:N
43
46
39
39
50
34
47
31
30
38
41
33
33
48
27
43
23
20
% marine
(±10)
M1 M2
This study
Umbelino 2006
This study
This study
Umbelino 2006
This study
Umbelino 2006
This study
This study
Reference
continued…
Table 4.7. Moita do Sebastião, Tagus valley: carbon (δ13C) and nitrogen (δ15N) measurements on human bone collagen samples. Samples from the nineteenth century assemblages were collected at MG and samples from the 1952–54 material were collected at MHNUP. Age and sex estimation by *C. Umbelino (2006), Mary Jackes (pers.
comm.), otherwise by D. Ferembach (1974). Samples are sorted so that the lowest δ13C values are on the top of the table. Individuals with multiple measurements are sorted
together.
♂
♂
♀
♂
♂
n/d
♂
n/d
♀
♂
Adult
Adult
Non–adult,
c 2 yrs
Adult
Adult
Adult
Adult
Adult,
old
Non-adult,
c 1 yr.
Adult
Adult
5
1952–54
18
1952–54
22
1952–54
29
19th c.
17
1952–54
24
19th c.
41
19th c.
34
1952–54
25
1952–54
22
19th c.
CT
19th c.
n/d
♂
Adult
9
1952–54
Table 4.7. continued
n/p
n/p
Tibia
Tibia–R
n/p
n/p
Femur–R
n/p
Femur–R
Tibia–R
Long bone–R
Tibia–L
TO–135
TO–131
Ua–46268
Ua–46271
TO–134
TO–132
Ua–47980
TO–133
Ua–46267
Ua–46266
Ua–47977
Ua–46264
6810±70
7240±70
7243±45
7236±41
7160±80
7180±70
7105±42
7200±70
7120±43
7095±45
7138±42
7621±50
–15.3
–16.1
–16.2
–16.3
–16.7
–16.8
–16.8
–16.9
–16.9
–16.9
–16.9
–17.0
13.4
12.2
14.0
11.9
11.2
11.9
9.5
10.4
12.8
10.8
9.3
10.8
n/a
n/a
3.2
3.2
n/a
n/a
3.1
n/a
3.2
3.1
3.1
3.2
59
51
51
50
46
45
45
44
44
44
44
43
59
49
48
46
41
40
40
39
39
39
39
38
Lubell and Jackes 1985; Lubell et al.
1994; Meiklejohn et al. 1986
Lubell and Jackes 1985; Lubell et al.
1994; Meiklejohn et al. 1986
This study
This study
Lubell and Jackes 1985; Lubell et al.
1994; Meiklejohn et al. 1986
Lubell and Jackes 1985; Lubell et al.
1994; Meiklejohn et al. 1986
This study
Lubell and Jackes 1985; Lubell et al.
1994; Meiklejohn et al. 1986
This study
This study
This study
This study
Results: Tagus valley, Moita do Sebastião
Archaeothanatology
Source material and analysis strategy
Source material for Moita do Sebastião is restricted to the excavations by J. Roche
in 1952–54. It comprises field photographs of variable quality (Cardoso and Rolão
1999–2000; Jackes et al. 2014; Roche 1972), simple sketches, written documentation from field notes (Cardoso and Rolão 1999–2000) and from the monograph Le
Gisement Mésolithique de Moita do Sebastião I (Roche 1972) and II (Ferembach 1974).
Few skeletons are preserved, or partially preserved, in blocks of concretion, including two individuals excavated in the nineteenth century (table 4.8). Unfortunately,
most skeletal material is in disarticulated state, often in containers with more than
one individual.
Overall, the archaeothanatological analysis is limited, and in several cases it relies on indirect observations only, such as the descriptions of the excavators. The
analysis was carried out on a total of 18 individuals, two excavated in the nineteenth century and 16 individuals excavated in 1952–54 (c 40% MNI 1952–54). In
several cases, the analysis is restricted to a few portions of the skeleton where
diagnostic criteria can be observed. In some cases it was possible to obtain data
supporting general observations regarding the treatment of the dead, even when
detailed analysis was not possible. For the remaining individuals excavated in
1952–54 (18 burial features), and for most of the assemblage from the nineteenth
century, there is not enough source material available, at least at the time of this
study.
Results of the archaeothanatological analysis
The nature of the deposits
Analysis of the nature of the deposits at Moita do Sebastião suggests a pattern of
primary burials (table 4.9), indicating that the cadavers were placed in the burial
features before the degradation of the labile joints. In four cases the labile articulations are poorly preserved and/or unclear in the documentation (MS1952–54, Sks
20, 21, 27, 32), however, the overall position of the bones strongly indicates that
the bodies were disposed while still retaining their general anatomical integrity,
supporting the interpretation of burials in primary position.
J. Roche (1972) noted that some grave features contained the remains of more
than one individual. The documentation does not allow detailed analysis of these
cases, and the possibility of placement of skeletal elements in secondary position
cannot be excluded. Deposits presumably containing more than one individual will
be discussed further below.
179
yes
n/a
n/a
n/a
n/a
27, 1952–54
30, 1952–54
31, 1952–54
32, 1952–54
33, 1952–54
n/a
15, 1952–54
yes
n/a
14, 1952–54
yes
yes
12, 1952–54
21, 1952–54
n/a
10, 1952–54
20, 1952–54
n/a
9, 1952–54
n/a
yes
5, 1952–54
yes
yes
3, 1952–54
19, 1952–54
yes
43, 19th c.
17, 1952–54
yes
Block
yes
yes
yes
yes
n/a
n/a
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
n/a
Graphic documentation w/human
remains
Drawings Site plan
Field
photographs
n/a
n/a
n/a
Source material
10, 19th c.
Individual
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
n/a
n/a
Written
documentation
Poor preservation and limited documentation. Limited analysis.
Poor preservation and limited documentation. Limited analysis.
Poor preservation and limited documentation. Limited analysis.
Poor preservation and limited documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
The block contains the remains of probably two individuals, possibly disturbed. The bones are heavily
covered with concretion and the analysis is extremely limited.
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
–
Upper body is poorly preserved. Archaeothanatology is limited and partially possible on the lower limbs.
–
–
–
Lower body is not preserved in the block. MS19thc., Sk10 described by Ferembach (1974) seems to be a
different skeleton as the one currently identified.
Archaeothanatology is limited and partially possible at the level of the lower limbs.
Observations
Table 4.8. Moita do Sebastião, Tagus valley: source material used for the archaeothanatological analysis. Material from the nineteenth century is curated at MG and the
assemblage from 1952–54 is located at MHNUP.
Primary
Primary
Primary
Primary
Primary
Primary
♂?
♂
♂
♂
♂
♀
Adult
Adult
Adult
Adult
Adult, 25–35 yrs ♂
♂
Adult
Adult, 35–50 yrs ♀
♀
Adult
Adult
Adult
43, 19th c.
3, 1952–54
5, 1952–54
9, 1952–54
10, 1952–54*
12, 1952–54
14, 1952–54
15, 1952–54*
17, 1952–54
19, 1952–54
Primary
Primary
Primary
Primary
Primary
n/d
Adult
10, 19 c.
Nature of the deposit
Sex
Age
th
Individual
Maintenance of labile articulations: bones of left hand, metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of left foot.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: bones of right hand, metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of right foot.
Maintenance of the general anatomical integrity of the body.
continued...
Maintenance of labile articulations: metatarsals and phalanges of left foot; other labile articulations are unclear in the
documentation. Maintenance of the general anatomical integrity of the body.
Labile articulations are unclear in the documentation, except for the bones of the hands which are partially disarticulated.
The overall position of the bones indicates that the general anatomical integrity of the body has been maintained.
Maintenance of labile articulations: bones of right and left hand, metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the lower body.
Labile articulations are poorly preserved. The bones of the left hand are covered with concretion and although their
articulation is not clear, these are concentrated in front of pelvic girdle.
The skeleton is well articulated and the general anatomical integrity of the body has been maintained.
Arguments
Table 4.9. Summary of the nature of the deposits with human remains at Moita do Sebastião, Tagus valley. Age and sex estimation by *C. Umbelino (2006) otherwise by
D. Ferembach (1974).
♀
n/d
n/d
♂
♀
♀
♀
Adult
Adult
Non-adult
Adult
Adult
Adult
Adult
20, 1952–54
21, 1952–54
27, 1952–54
30, 1952–54
31, 1952–54
32, 1952–54
33, 1952–54
Table 4.9. continued
Primary
Probably primary
Primary
Primary
Probably primary
Probably primary
Probably primary
Maintenance of labile articulations: bones of the left hand, metatarsals and phalanges left foot.
The skeleton is well articulated and the general anatomical integrity of the body has been maintained.
Labile articulations are poorly preserved and unclear in the documentation.
The overall position of the bones indicates that the general anatomical integrity of the body has been maintained.
Maintenance of labile articulations: bones of right hand, some metatarsals and phalanges of right foot.
The skeleton is well articulated and the general anatomical integrity of the body has been maintained.
Maintenance of labile articulations: bones of right hand, metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: TM joint is preserved but the documentation is unclear for other joints.
Maintenance of the general anatomical integrity of the body.
The skeleton is heavily covered with concretion and the labile articulations are not exposed.
The lower limbs are fragmented but seem well articulated in anatomical integrity.
The labile articulations are poorly preserved and unclear in the documentation.
The skeleton is well articulated and the general anatomical integrity of the body has been maintained.
Results: Tagus valley, Moita do Sebastião
The space of decomposition of the cadaver
Analysis of the space of decomposition of the cadavers buried in Moita do
Sebastião indicates a general pattern of filled spaces (table 4.10). The cadavers were
covered with sediment immediately after their placement in the features. The
decomposition of the bodies took place in this sediment-filled environment. In
several cases the analysis is unclear due to the poor preservation of the remains. In
every case, the movement of the bones is limited suggesting that the space of
decomposition was filled and that the sediment could penetrate immediately.
The initial position of the cadaver in the feature
The dominant position at Moita do Sebastião consists of a cadaver lying on the
back (n=15 to 17). In one case only (MS1952–54, Sk10), it is possible that the
cadaver was lying on the right side (see Jackes et al. 2014), but the data is not
conclusive (table 4.11).
The head of the cadaver was often raised in relation to the trunk as illustrated
by the rotation of the crania forwards and downwards towards the body in the
feature. In a few cases, the excavators noted a volume of sand behind the crania
(e.g., MS1952–54, Sks 3, 5, 9) (Roche 1972). In at least one case (MS1952–54,
Sk12), the head was not elevated and was placed at the same level as the trunk.
The position of the upper limbs varied considerably, but they were normally
placed close to the body, symmetrically (n=6 to 8) or asymmetrically (n=6), in
extension, semi-flexion (mostly), flexion or hyperflexion. In most cases, at least
one of the hands was placed on the abdominal or pelvic region. MS1952–54, Sk15
presents the exception to this pattern and his arms were placed in abduction, however, the forearms were lying towards the body and the hands were placed on the
pelvic region. Two burials (MS1952–54, Sks 9, 12) present a particular disposal of
one of the forearms. In these cases, one of the elbows is in hyperflexion (9–left,
12–right), the forearm having been placed in front of the respective arm with the
hand lying on the respective shoulder.
Typically, the lower limbs were in flexion at the hip and knee joints, and the feet
were positioned towards the buttocks in a constrained position. According to J.
Roche (1972), the feet were usually close to the pelvic region, and they were often
crossed or placed one against the other (MS1952–54, Sks 5, 12, 14, 15, 16, 22, 30).
Hyperflexion is noted at the level of the knees only, probably in just two to four
cases (MS19th c., Sk43; MS1952–54, Sks 19, 21, 31–right). MS1952–54, Sk9 was the
only individual buried with the lower limbs in full extension (fig. 4.7). MS1952–54,
Sk15 presents a similar position but in this case the limbs were placed in slight
semi-flexion. Often, the rotation of the lower limbs is unclear in the documentation, but it seems that in just a few cases the limbs were rotated to the right
(MS1952–54, Sks 17, 19–tendency, 31) or to the left (MS1952–54, Sks
12–tendency, 14). Normally, the knees would be directed forwards or towards one
side while the feet were brought up towards the buttocks. In a few cases, the lower
183
Chapter 4
limbs were probably placed towards the trunk. This is documented for MS1952–
54, Sk10, but the documentation is unclear in most cases.
184
Sex
n/d
♂?
♂
♂
♂
♀
♀
♂
♂
Age
Adult
Adult
Adult
Adult
Adult
Adult, 35–50 yrs
Adult
Adult
Adult, 25–35 yrs
Individual
10, 19th c.
43, 19th c.
3, 1952–54
5, 1952–54
9, 1952–54
10, 1952–54
12, 1952–54
14, 1952–54
15, 1952–54
Filled
Probably filled
Filled
Probably filled
Filled
Filled
Filled
Probably filled
Filled
Space of decomposition
Maintenance of bones in original unbalanced position:
the bones of the feet are rotated and placed in front of each other.
No movement outside initial volume of the cadaver.
Unclear documentation but no movement outside initial volume of the cadaver.
continued...
Maintenance of bones in original unbalanced position: moderate collapse of the bones of os coxae; the bones of the
feet are well articulated while rotated and placed in front of each other (left in front of right).
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position:
moderate collapse of the bones of the hands in front of TC; the patellae are suspended on the distal end of the
femora.
Diagnostic criteria at the level of the upper body are not preserved.
Maintenance of bones in original unbalanced position: phalanges of right foot.
Diagnostic criteria are only partially exposed.
Maintenance of bones in original unbalanced position:
moderate collapse of bones of left hand which placed in front of abdominal region; metatarsals of foot.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: perfect articulation of the bones of right hand rotated and
flexed at the wrist; moderate collapse of the bones of the left hand lying in the lower abdominal region.
The left patella lies in articulation at the distal end of the femur.
No movement outside initial volume of the cadaver.
Analysis not possible at the level of the upper body.
Maintenance of bones in original unbalanced position: phalanges of the feet.
Poor preservation of diagnostic criteria.
Maintenance of bones in original unbalanced position: bones of left hand in front of pelvic region.
No movement outside initial volume of the cadaver.
Arguments
Table 4.10. Summary of the space of decomposition of the human remains in the grave features at Moita do Sebastião, Tagus valley.
♂
♀
♀
n/d
n/d
♂
♀
♀
♀
Adult
Adult
Adult
Adult
Non-adult
Adult
Adult
Adult
Adult
17, 1952–54
19, 1952–54
20, 1952–54
21, 1952–54
27, 1952–54
30, 1952–54
31, 1952–54
32, 1952–54
33, 1952–54
Table 4.10. continued
Filled
Probably filled
Filled
Filled
Probably filled
Unknown
Filled
Filled
Filled
Maintenance of bones in original unbalanced position: bones of the left hand are rotated and lie in front of LV; bones
of the feet are rotated and placed in front of each other (left in front of right).
No movement outside initial volume of the cadaver.
Poor preservation and unclear documentation.
No movement outside initial volume of the cadaver.
Poor preservation and unclear documentation.
Maintenance of bones in original unbalanced position:
bones of right hand are rotated and placed in front of left forearm; a portion of the right foot seems to be well articulated while rotated outwards. According to J. Roche (1972, 127) the CV can be followed as an arc forwards.
No movement outside initial volume of the cadaver.
Poor preservation and unclear documentation.
Maintenance of bones in original unbalanced position: moderate collapse of TC; bones of right hand.
No movement outside initial volume of the cadaver.
Poor preservation and no graphic documentation.
No movement outside initial volume of the cadaver.
The feature is disturbed and contains the remains of possibly two individual. The bones are only partially exposed and
covered with a hard matrix of concretion.
Poor preservation and unclear documentation.
Maintenance of bones in original unbalanced position: bones of right hand are rotated at the wrist.
No movement outside initial volume of the cadaver.
Poor preservation and unclear documentation.
Maintenance of bones in original unbalanced position:
bones of left hand are rotated and lie in front of left shoulder girdle; bones of left foot.
No movement outside initial volume of the cadaver.
Poor preservation and unclear documentation.
Maintenance of bones in original unbalanced position:
bones of the hands lying in front of each other on the lower abdominal region.
No movement outside initial volume of the cadaver.
n/d
♂?
♂
♂
♂
♀
♀
♂
♂
Adult
Adult
Adult
Adult
Adult
Adult, 35–50 yrs
Adult
Adult
Adult, 25–35 yrs
10, 19th c.
43, 19th c.
3, 1952–54
5, 1952–54
9, 1952–54
10, 1952–54
12, 1952–54
14, 1952–54
15, 1952–54
Sex
Age
Individual
Raised
Rotation downwards (neck flexed forwards)–
tendency R
n/a
Rotation–L
n/a
Raised
Slight rotation downwards (neck flexed forwards)
Raised
Slight rotation downwards (neck flexed forwards)
Raised
Rotation downwards (neck flexed forwards)–
tendency R
n/a
Head
n/a
Initial position of the cadaver in the feature
On the back
On the back
On the back
On the back,
or on the right?
On the back
On the back
On the back
Probably on the back
Trunk
On the back
In semi-flexion
In extension–R
In extension?–L
In hyperflexion–R
In extension–L
n/a
In flexion–R
In hyperflexion–L
In semi-flexion
In flexion–R
In semi-flexion–L
n/a
Upper limbs
In extension–R
In semi-flexion–L
continued...
Hip and knee joints: semi-flexion
Feet against one another
Hip and knee joints: flexion
Rotation–L
Feet crossed and towards buttocks
Hip and knee joints: flexion
Rotation–tendency L
Feet towards buttocks
Hip and knee joints: flexion
In adduction towards trunk
Feet towards buttocks?
Hip and knee joints: extension
Hip and knee joints: flexion
Feet towards buttocks
Hip and knee joints: flexion
Rotation–R
Feet towards buttocks
Hip joints: prob. flexion
Knee joints: prob. hyperflexion
Feet towards buttocks
Lower limbs
Hip joint–R: flexion
Knee joint–R: n/a
Hip and knee joints–L: n/a
Table 4.11. Summary of the reconstruction of the initial position of the cadaver in the grave features at Moita do Sebastião, Tagus valley. The initial position of the limbs is
indicated as –R or –L and refers to the right or left limb, when their positions are different. Rotation –R or –L indicates the rotation of the skeletal element(s) towards the
right or the left side.
♀
♀
n/d
n/d
♂
♀
♀
Adult
Adult
Adult
Non-adult
Adult
Adult
Adult
Adult
19, 1952–54
20, 1952–54
21, 1952–54
27, 1952–54
30, 1952–54
31, 1952–54
32, 1952–54
33, 1952–54
♀
♂
Adult
17, 1952–54
Table 4.11. continued
On the back
On the back
Probably on the back
On the back
On the back
On the back
Raised
Rotation downwards (neck flexed forwards)–
tendency L
Raised
Rotation downwards (neck flexed forwards)
On the back
On the back
Raised
On the back
Rotated–R and downwards (neck flexed forwards)
Raised
Rotation downwards (neck flexed forwards)
n/a
Rotation downwards (neck flexed forwards)
Raised
Rotation downwards (neck flexed forwards)
Raised
Rotation downwards (neck flexed forwards)
Raised
Rotation downwards (neck flexed forwards)
In extension or
semi-flexion–R
In semi-flexion–L
In semi-flexion
In semi-flexion (c
90˚)–R
In semi-flexion–L
In extension
n/a
n/a
In flexion–R
In semi-flexion–L
In semi-flexion–R
In hyperflexion–L
In semi-flexion
Hip joints: semi-flexion
Knee joints: flexion
Rotation–R
Feet towards buttocks
n/a
Hip joint–R: semi-flexion
Knee joint–R: hyperflexion
Hip joint–L: semi-flexion (c 90˚)
Knee joint–L: flexion
Rotation–R
Feet towards buttocks
n/a
Feet next to one another towards
buttocks
n/a
Hip joints: prob. flexion
Knee joints: prob. hyperflexion
Prob. feet towards buttocks
Hip joints: flexion
Knee joints: flexion?
Feet crossed
Hip joint–R: semi-flexion
Knee joint–R: n/a
Hip joint–L: flexion
Knee joint–L: hyperflexion
Left foot towards buttocks–
tendency R
Hip and knee joints: flexion
Rotation–R
Feet towards buttocks
Results: Tagus valley, Moita do Sebastião
The grave features
J. Roche (1972) described the grave features as natural depressions, or as pits excavated in the sand, possibly not very deep. The limits of the graves were difficult to
determine during excavation, but according to the archaeologists, the floor of the
graves was prepared to accommodate the cadaver (MS1952–54, Sks 3, 5, 7, 14, 30).
Archaeothanatological analysis suggests that the lateral pressures and walleffects observed on the skeleton in the feature indicate that the shape and size of
the pit was just sufficient to contain the body in the planned burial position (table
4.12). In most cases the design of the grave had a significant impact on the final
arrangement of the body in the feature, visible by the constrained positions of
most skeletons, particularly at the level of the lower limbs. Often, the lower limbs
were supported by the wall of the grave, keeping the feet towards the buttocks.
While the pit imposed physical limits on the cadaver, by maintaining the lower
limbs clumped together, it also provided lateral support preventing the collapse of
the skeletal elements. At the level of the upper body, this is demonstrated by bilateral pressure exerted on the body, observed in most cases.
It is possible that in some cases there might have been an extra support offered
by some kind of container such as a light wrapping. MS1952–54, Sk9 (fig. 4.7)
presents one of these cases, where the strong bilateral pressures could suggest the
existence of a soft wrapping at the time of the burial. This adult male (Ferembach
1974) was lying on the back, with the head slightly elevated in relation to the trunk.
The upper limbs were flexed or hyperflexed at the elbows, the right hand was
placed on the chest and the left hand lay on the left shoulder. The most striking
element of this burial is the placement of the lower limbs in complete extension,
which is unique at Moita do Sebastião.
Another striking feature of this burial is the compression towards the medial
axis of the body. The analysis suggests an overall bilateral pressure exerted on the
body. Some diagnostic elements are unclear in the documentation, but the available
arguments are strong. The wall-effect and consequent pressure on the upper body
is well illustrated by the alignment and inwards rotation of the right humerus, the
verticalization of the left clavicle, the alignment of the bones of the upper left limb,
and the medial compression and verticalization of the ribs. These pressures seem
to suggest the possibility of the existence of a soft wrapping surrounding the
cadaver at the time of the burial. At the lower body, the lateral support of the
skeletal elements seems to suggest another possible scenario. The os coxae were
maintained in anatomical position, even after the slight outwards rotation of the
femora, suggesting significant lateral support at this level. While the femora rotated, the patellae remained in perfect anatomical position, indicating direct contact
with sediment, preventing the collapse of these unstable elements, which seems to
invalidate the hypothesis of soft wrapping, at least at this level. The two latter
arguments suggest that the overall compression of the skeleton could be explained
by the narrow characteristics of the pit. In this scenario, the walls of the grave
would have exerted a significant supporting effect on the skeletal elements. This
189
Chapter 4
hypothesis could be excluded if some element in the feature suggested that the pit
was wider than the area occupied by the body, but that was not the case. Overall,
the hypothesis of a narrow pit seems to explain most of the observations in the
feature; however, the wrapping hypothesis is possible, at least around the upper
body, and cannot be completely excluded.
Figure 4.7. MS1952–54, Sk9 was the only cadaver lying on the back with the lower limbs in
complete extension. Bottom figure, shows MS1952–54, Sk9 (superior right) lying on a sloping
floor, slightly more elevated at the level of the head. (Photographs after Cardoso and Rolão
1999–2000, 200, fig. 38; Roche 1972, Plate III–1)
190
Results: Tagus valley, Moita do Sebastião
Two other cases (MS19th c., Sk10; MS1952–54, Sk30) present strong arguments
supporting the hypothesis of extra support provided by a soft wrapping of the
cadaver. In these burials, the skeletal elements of the upper body present strong
medial tendency, which is particularly well illustrated by the compression of the TC
and verticalization of the ribs. The bilateral pressures are further demonstrated by
the verticalization of the clavicles, the projection forwards and upwards of the
shoulder girdles, and the strong alignment of the upper limbs placed close to the
upper body. The skeletal elements of the lower body of MS19th c., Sk10 are not
preserved in articulation and there is no graphic documentation for this feature. In
the case of MS1952–54, Sk30 (fig. 4.8), the os coxae have collapsed outwards and
do not present any particular support at this level. The lower limbs are not preserved, but the position of the bones of the feet, placed towards the buttocks,
indicate that the limbs were flexed at the hips and knees. It is plausible, but not
conclusive, that the upper bodies of these cadavers could have been contained in a
soft wrapping at the time of the burial. The placement of the bodies in a narrow
pit remains valid, and possibly the most likely hypothesis.
Figure 4.8. The upper body of MS1952–54, Sk30 presents strong medial tendency. This is
illustrated by the compression of the TC and verticalization of the ribs. Bilateral pressures
are further demonstrated by the verticalization of the clavicles, the projection forwards and
upwards of the shoulder girdles, and the strong alignment of the upper limbs placed close
to the upper body. (Photograph after Jackes et al. 2014, 134, fig. 10.2)
In some cases, the grave pits were somewhat wider allowing a relatively expanded
position of certain skeletal elements, particularly at the level of the upper body
(MS1952–54, Sks 5, 12, 15, 31). This is visible by the position of the humeri in
abduction and/or by the expansion and partial opening of the rib cage (figs. 4.9,
4.10). Apart from MS1952–54, Sk15 (fig. 4.11), whose lower limbs were in slight
semi-flexion at the hips and knees, all other cadavers were constrained by a walleffect at the lower end of the grave.
191
Chapter 4
Figure 4.9. MS1952–54, Sk5. The opening of the lower portion of the rib cage indicates that
the grave pit was wide enough to allow the expansion of the skeletal elements. (Photograph
after Cardoso and Rolão 1999–2000, 201, fig. 39)
Figure 4.10. MS1952–54, Sk12. The position of the upper right limb in abduction suggests
that the grave pit was relatively wide at this level. This is in contrast with the wall-effect at
the lower end of the grave keeping the lower limbs aligned, and the feet in forced position
directed towards the buttocks. (Photograph after Cardoso and Rolão 1999–2000, 201, fig.
39)
192
Results: Tagus valley, Moita do Sebastião
Figure 4.11. MS1952–54, Sk15. The position of the arms in abduction and the lower limbs
slightly semi-flexed at the hips and knees, suggests that the pit was wide enough to contain
the body without exerting any particular pressures on the skeletal elements in the feature.
(Photograph after Cardoso and Rolão 1999–2000, 200, fig. 15)
The floor of the graves was often irregular and sloping, however, the documentation is not always clear. The moderate downslope is suggested by the rotation forwards and downwards of the head, towards the body in the feature, registered in
several cases. At least in one case (MS1952–54, Sk31) this rotation was very pronounced and the neck was in hyperflexion, as indicated by the flexion of the cervical vertebrae which could be followed in an arc forwards (Roche 1972, 127). However, as noted by J. Roche, the elevation of the head was sometimes aided by the
placement of a volume of sand functioning as a headrest. Further elements supporting the sloping nature of the floors are indicated by the displacement observed
at the level of persistent joints of the vertebral column (MS1952–54, Sks 3, 5, 9),
suggesting the effect of an irregular surface and the rearrangement of the bones
during the process of decomposition.
193
n/d
♂?
♀
♀
♀
♀
♂
Adult
Adult
Adult
Adult
Adult
Adult, 35–50 yrs
Adult
Adult
10, 19th c.
43, 19th c.
3, 1952–54
5, 1952–54
9, 1952–54
10, 1952–54
12, 1952–54
14, 1952–54
♂
Sex
Age
Individual
1. Prob. wall-effect, bilateral, upper body;
2. Wall-effect, left lower end.
Design of the grave feature.
1. Sloping floor;
2. Wall-effect, left side;
3. Wall-effect, left lower end.
Design of the grave feature.
n/a
1. Overall strong bilateral compression and constraint
effects.
Design of the grave feature and/or soft wrapping.
1. Wall-effect, right side;
2. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, bilateral;
2. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, lower end.
Prob. design of the grave feature.
1. Strong bilateral compression and constraint effects
on upper body.
Lower body: n/a.
Design of the grave feature and/or soft wrapping.
Pressures
continued...
1. Suggested by alignment of the upper limbs affecting shoulders and arms;
2. Feet are forced towards buttocks.
1. Suggested by exposure of right scapula indicating movement of the body towards left side;
2. Alignment of upper left limb, involving shoulder and arm;
3. Lower limbs are constrained and clumped together. Feet are forced towards
buttocks.
n/a
1. Alignment of right arm. Alignment of upper left limb and projection forwards
and upwards of shoulder and arm. Medial compression of TC. Maintenance of the
os coxae in anatomical position. Alignment of lower limbs.
1. Alignment of upper right limb and projection forwards and upwards of shoulder
and arm. Pattern accentuated by the sloping floor causing a transfer of body weight
from the right side to the medial axis;
2. Alignment and clumping of lower limbs.
1. Alignment of upper limbs and projection forwards and upwards of shoulders
and arms. Pattern enhanced by sloping floor, slightly raised at the superior end as
well as on lateral sides, causing a transfer of body weight towards medial axis;
2. Alignment and clumping of lower limbs.
1. Suggested by constrained position of feet lying close to buttocks.
1. Alignment of upper limbs positioned close to TC. Projection forwards and
upwards of shoulders and arms. Medial compression of TC.
Arguments
Table 4.12. Summary of the key observations used in the reconstruction of the grave features at Moita do Sebastião, Tagus valley. Observed pressures on the body are presented in a number sequence. The arguments supporting these observations are indicated by the same number in the respective column.
♂
♀
♀
n/d
♂
♀
♀
Adult
Adult
Adult
Adult
Non-adult
Adult
Adult
Adult
Adult
17, 1952–54
19, 1952–54
20, 1952–54
21, 1952–54
27, 1952–54
30, 1952–54
31, 1952–54
32, 1952–54
33, 1952–54
♀
n/d
♂
Adult, 25–35 yrs
15, 1952–54
Table 4.12. continued
1. Weigh of sediment from above;
2. Wall-effect, bilateral, upper body;
3. Wall-effect, lower end.
Design of the grave feature.
1. Prob. wall-effect, bilateral, upper body.
Design of the grave feature.
1. Weigh of sediment from above;
2. Wall-effect, lower end.
Design of the grave feature.
1. Weigh of sediment from above;
2. Strong bilateral compression and constraint effects
on upper body.
3. Prob. wall-effect, lower end.
Design of the grave feature and/or soft wrapping.
n/a
1. Overall constraint effect.
Design of the grave feature?
1. Wall-effect, bilateral, upper body;
2. Prob. wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, bilateral, upper body;
2. Wall-effect, lower left end.
Design of the grave feature.
1. Prob. bilateral compression;
2. Prob. wall-effect, lower end.
Prob. design of the grave feature.
Unclear documentation
No particular pressures.
1. Crushed cranium;
2. Alignment and compression of shoulders;
3. Suggested by constrained position of lower limbs clumped together, visible by
the placement of feet towards buttocks.
1. Suggested by alignment of shoulders and arms.
1. Crushing of frontal of cranium and movement inwards towards the calvaria
while the back of the cranium is perfectly preserved;
2. Alignment of upper limbs positioned close to TC. Projection forwards and
upwards of shoulders and arms. Medial compression of TC;
3. Suggested by constrained position of lower limbs clumped together, visible by
the placement of feet towards buttocks.
1. Crushed cranium; hyperflexion of neck;
2. Possible wall-effect at the lower end of the feature keeping the lower limbs
clumped together, visible by the feet placed towards the buttocks.
n/a
1. Suggested by constrained position of lower limbs clumped together.
1. Alignment and compression of shoulders and upper limbs;
2. Suggested by constrained position of lower limbs clumped together.
1. Alignment and compression of shoulders and upper limbs;
2. Left foot towards buttocks.
1. Suggested by alignment of the upper limbs affecting shoulders and arms;
2. Suggested by constrained position of lower limbs clumped together, visible by
the placement of feet towards buttocks.
–
Chapter 4
Deposits containing more than one individual
J. Roche (1972) describes 34 grave features excavated between 1952 and 1954,
some of which containing the remains of more than one individual. That is the
case of the graves 1, 7, 13, and 18. MS19th c., Sk20 was also identified as containing
the remains of two individuals (Ferembach 1974). In every case, the remains were
found disturbed and fragmented and the temporal relation between the individuals
is unclear (Roche 1972). The site plan from the 1952–54 shows only individual
burials (n=34). Also, in the descriptions by both J. Roche and O. V. Ferreira
(Cardoso and Rolão 1999–2000; Roche 1972) it is clear that in each of these features, one individual was always identified as the main skeleton in the grave, suggesting that the remains of the second or third individual were older and deposited
separately over a period of time. It seems likely that new burials may have disturbed previous deposits, and the older elements were displaced to accommodate
the new cadaver. Field notes from June 1880 describe the skeletons placed in very
close proximity, often overlapping, at least partially (Jackes and Alvim 2006). This
practice, known as reduction, was presumably common at Moita do Sebastião, indicating the continued preference for these areas of the site for the burial of the
dead. Based on the current documentation, there is no evidence to support the
interpretation of simultaneous deposits at the Moita do Sebastião, but this
hypothesis cannot be rejected either.
Post-depositional manipulations of the cadavers
At Moita do Sebastião, the bodies are apparently complete and lie generally undisturbed in the grave feature. As discussed above, some deposits were possibly
disturbed to give place to a new burial.
It is unknown if the practice of re-opening the grave for the removal or rearrangement of selected elements occurred at Moita do Sebastião. Some cases seem
to indicate post-depositional manipulations of the cadavers but the data is not
conclusive. MS1952–54, Sk14 was found without the cranium while a fragment of
the mandible remained in situ (Roche 1972, 122). The remains of this individual
were disturbed and poorly preserved, and it is unclear if the missing cranium could
be explained by preservation, disturbance or deliberate action. Skeletons of children were normally found incomplete (Roche 1972, 125–6). This could be explained by the intrinsically low resistance to decay of the bones of infant and juvenile skeletons (Mays 2010, 28). Two cases, however, are suggestive of postmanipulation of selected skeletal elements. One case is MS1952–54, Sk23, which
has been described as an isolated cranium of a child deposited in a pit (Roche
1972, 125). The second case (MS1952–54, Sk24), was found next to the latter deposit, and contained most skeletal elements of the upper body of a child, including
the cranium, while the lower limbs were missing. Overall, the poor preservation of
the material and unclear documentation does not allow reliable observations at this
level, but the hypothesis of post-manipulations of the cadavers is plausible and
cannot be excluded.
196
Results: Tagus valley, Moita do Sebastião
Other observations
In the 1952–54 excavations, J. Roche (1972) documented the presence of ochre
and other pigments, objects of adornment, charcoal and ashes, as well as the concentration of selected shells placed in the grave features.
Ochre and other pigments were found in small quantities and in small fragments in at least five graves (MS1952–54, Sks 3, 5, 6, 11, 14) (Roche 1972, 132).
Objects of adornment were identified in several graves, most notably in
MS1952–54, Sks 1, 5, 6, 8, 11, 25 (Roche 1959, 408). These consisted mostly of
pierced shells, such as Theodoxus fluviatilis (also known as Neritina fluviatilis), and
were sometimes stained with ochre.
The floor of some graves was covered with selected shells (Roche 1972, 132).
A layer of unopened grooved carpet shells (Ruditapes decussatus) was placed in the
grave before the placement of MS1952–54, Sk3. MS1952–54, Sk12 had a great
quantity of non-pierced land snails (Helix pisana) all around the deposit (Roche
1972, 121). MS1952–54, Sk33 was buried with a great quantity of peppery furrow
shells (Scrobicularia plana). The most striking floor preparation was found in a child
burial (MS1952–54, Sk25), with the body being laid on its back on a layer of
Theodoxus fluviatilis shells.
Fragments of charcoal and ashes are commonly found mixed in the grave sediment. J. Roche (1972) noted that in some cases, the relation between fire activity
and the graves is more evident. According to the archaeologist, two cases
(MS1952–54, Sks 1, 3) indicated that a fire was lit close to the cadavers. A third
case (MS1952–54, Sks 34) suggested an intense fire near the cadaver, the heat of
which was sufficient to have an impact on the cranium, at least superficially (Roche
1972, 132).
The mortuary programme at Moita do Sebastião
Archaeothanatological analysis of the human remains excavated at Moita do
Sebastião is limited. Most human remains are preserved in disarticulated state, and
the graphic documentation is limited to a few individuals. Nevertheless, the analysis indicates a general similarity of burial practices observed at the site and the minor detected variations do not affect the overall pattern of the treatment of the
dead.
In the sample analysed, the cadavers were placed in individual grave features
while retaining their anatomical integrity (individual primary deposits). The feature
was then immediately covered with sediment, providing a filled space of decomposition.
The bodies were typically lying on the back in pits somehow adjusted to the size
of the cadaver in the planned burial position. The upper limbs were normally
placed close to the body and the hands were placed in front of the upper body in
variable positions. The lower limbs were normally in flexion at the hips and knees,
but rarely placed in front of the trunk. Often, the lower limbs were in contraction,
and the knees were directed forwards, or alternatively directed towards one side. In
197
Chapter 4
these cases, the feet were placed in a forced position towards the buttocks. These
constrained positions, particularly at the level of the lower limbs, were maintained
by the physical limits of the grave which provided support and kept the limbs
clumped together.
Some cases provide clear indications of the preparation of the graves; notably
by the arrangement of headrests made of sand and the inclusion of shells before
the placement of the cadaver in the grave.
Few cases present deviant elements, but there are exceptional instances of features with, for example, possible wrappings, unique initial positions in the grave,
and particular preparation of the feature. Despite the striking and unusual elements, all these cases follow the common core characteristics of burial practices
observed in the sample analysed from Moita do Sebastião.
198
Tagus valley: Cabeço da Arruda
Site background: a short history of research
Cabeço da Arruda is a shell midden site located on a small headland on the edge of
the flood plain, on the right bank of the tributary Muge, about 5 kilometres from
the confluence with the Tagus River. This position makes the site vulnerable to
occasional floods, particularly during the winter (Roche 1967, 79, 80, 1974, 25).
The site has excellent visibility towards the river and towards Moita do Sebastião
and Cabeço da Amoreira on the opposite margin (fig. 2.2). Cabeço da Arruda is
possibly the largest shell midden known reaching c 5 metres in depth in the central
area (Roche 1967, 79).
The site was identified and test excavated in 1863 by C. Ribeiro of the Geological Commission, the same year the first middens were identified. The first systematic excavations were carried out in 1865 by F. A. Pereira da Costa, and continued
in 1880 by C. Ribeiro who aimed at presenting the site at the 9th International
Congress of Prehistoric Anthropology and Archaeology in Lisbon (Ribeiro 1884).
In 1884–85, F. de Paula e Oliveira expanded the excavations. His work was presented in a posthumous publication in 1892, becoming the first volume on a prehistoric site in Portugal (Paula e Oliveira 1888–92).
In the 1930s, M. Corrêa and collaborators from the Institute of Anthropology
Porto University, started a new period of excavations at Cabeço da Arruda
(Cardoso and Rolão 1999–2000). The first season in the summer of 1933 was relatively short possibly due to the excavation of few human remains (MNI 1)
(Cardoso and Rolão 1999–2000, 172). The second season, in 1937 was longer (August to November), and presumably with better results (MNI 10).
In 1964–65 the site was excavated by J. Roche and O. V. Ferreira, under the
Geological Services of Portugal, and the fieldwork was focused on the analysis of
the complex stratigraphy of the site (Roche 1967; 1974, 27).
The last field season was in 2000, directed by J. Rolão, aiming at assessing the
state of preservation of the site while reinforcing and protecting the profiles
(Roksandic 2006).
Written sources indicate that c 80 individuals were excavated in the nineteenth
century, while the excavations of 1933, 1937 and 1964–65 recovered c 11 and 13
individuals, respectively (table 4.13). The currently accepted MNI is 110 based on
the assemblages at both MG and MHNUP (Cunha and Cardoso 2002–03) and the
two burials excavated in 2000 (Roksandic 2006).
199
Chapter 4
Table 4.13. Estimated number of individuals excavated each year at Cabeço da Arruda, Tagus valley.
Year of excavation
Number of individuals
Reference
1863
45
Pereira da Costa 1865
1880
MS + CAR, MNI 120
Ribeiro 1880
1884
13
Paula e Oliveira 1888–92
1885
?
–
1933
1
Abrunhosa 2012; Cardoso and Rolão 1999–2000
1937
10
Abrunhosa 2012; Cardoso and Rolão 1999–2000
1964
9 (Sks 1–9)
Cardoso and Rolão 1999–2000; Roche 1974
1965
4 (Sks 10–13)
Cardoso and Rolão 1999–2000; Roche 1974
2000
2 (CA–00–01, CA–00–02)
Roksandic 2006
Total recovery, MNI = 84 + unknown number for 1880
Like Moita do Sebastião, the current MNI is probably lower than the number excavated at the site. The number of individuals excavated at Cabeço da Arruda
could presumably go up to 150, admitting that some of the osteological material
may not have survived in the current collections (Jackes and Meiklejohn 2008,
238), or even up to 170, according to original nineteenth century labels at MG.
Most human remains excavated in the nineteenth century by the Geological
Commission are curated at MG. A few individuals, represented by their crania, are
maintained at the National Museum of Natural History and Science (MUNHAC),
Lisbon. The collections from the 1930s are curated at the Museum of Natural History, University of Porto (MHNUP). The MG also holds the collections from
1964–65. Most human remains are preserved in disarticulated state, and just a few
bone elements are still in articulation due to an enveloping heavy carbonate concretion.
Original written documentation known for Cabeço da Arruda is from the excavations in 1933, 1937 and 1964–65. Both archives are published (Abrunhosa 2012;
Cardoso and Rolão 1999–2000), at least partially. The 1930s material consists of
various short loose notes, as well as field diaries, handwritten by Santos Júnior and
M. Corrêa (Abrunhosa 2012; Cardoso and Rolão 1999–2000). The 1964–65 documentation consists of a field diary handwritten mostly by O. V. Ferreira, but also
by M. Andreata (Cardoso and Rolão 1999–2000). J. Roche’s field notes have not
survived (see chapter 3, A note on the historical documentation). The graphic documentation is extremely limited and consists of site plans, profiles, a few field photographs and simple sketches of the human remains (Abrunhosa 2012; Cardoso and
Rolão 1999–2000; Cartailhac 1886; Ribeiro 1884; Roche 1967).
200
Results: Tagus valley, Cabeço da Arruda
Chronology of the burial activity
Previous 14C measurements
Five 14C measurements are available on non-human samples (table A4). Two of
these measurements are from material from the nineteenth century (Beta–152956
and Beta–271927), one of which (Beta–152956) on bone of a domestic dog (Canis
familiaris) (Detry and Cardoso 2010). The 1960s excavations were dated by J. Roche
from two samples of charcoal (Sa–196 and Sa–197) from the upper and bottom
layers respectively (Delibrias and Roche 1965; Roche 1965, 161). One sample on
charcoal (TO–10215) from the 2000s excavations was also dated (Roksandic 2006).
The first measurements on human bone from Cabeço da Arruda (TO–354,
TO–355, TO–356, TO–359, and TO–360) were carried out in the 1980s. The
samples were collected from the nineteenth century material and selected randomly
(Lubell and Jackes 1988; Lubell et al. 1994). The measurements were done by AMS
at the Isotrace Laboratory of the University of Toronto, Canada, and independently measured for stable isotopes. The individual results for the C:N ratio were
not published but were reported to be within the accepted quality range of 2.9–3.6
(Lubell et al. 1994, 204).
The excavations from 1937 were dated by one 14C measurement on individual 6
(Beta–127451, Cunha and Cardoso 2002–03; Cunha et al. 2003). This measurement
has been questioned for being too old (Jackes and Meiklejohn 2004; Jackes and
Lubell 2012), and CAR1937, Sk6 was recently re-dated (AA-101343, Jackes et al.
2014). M. Jackes and colleagues (2014) suggest that the original date (Beta–127451)
is anomalous, although not statistically different from the new measurement
(AA–101343). One of the problems with the first 14C date (Beta–127451) was that
the δ13C value published along with the sample was obtained by AMS (Cláudia
Umbelino, pers. comm.), which cannot be used for dietary reconstructions or reservoir corrections (Millard 2014, 557; Taylor and Bar-Yosef 2014, 117). Now, with
the new measurements (AA–101343), which include an IRMS–based δ13C value,
this problem can be overcome, and the Beta–127451 date can be corrected with
the new δ13C value. Following this, instead of rejecting the first 14C date (Beta–
127451), and because the dates are not statistically different, I opted to combine
the two 14C measurements available for CAR1937, Sk6 (Beta–127451 and AA–
101343). This was done by using the OxCal function R_Combine and the recent
δ13C value (AA–101343) obtained by IRMS.
The two individuals (CA–00–01 and CA–00–02) excavated in 2000 are also
dated (Roksandic 2006). The C:N ratios were not originally published but were
reported recently (Jackes and Lubell 2015). These two samples were also independently measured for stable isotopes.
The human remains recovered during J. Roche’s excavations in 1964–65 were
not previously dated.
201
Chapter 4
Objectives of the dating programme
Cabeço da Arruda is a large site with a high number of human burials found in
different levels of the shell midden.
The general aims of the dating programme for Cabeço da Arruda followed the
macro-goals of the radiocarbon programme of this dissertation, outlined in chapter
3, and were
• to determine the duration of use of this site for burial activity;
• to determine when the burial activity started and when it ended;
• to estimate the frequency of the burial activity in the site.
The site-specific aims of the dating programme for Cabeço da Arruda were:
• to establish the chronology of the burial areas identified in 1964 (basal
sand) and in 1965 (upper layers) (fig. 4.12), which were undated;
• to establish the chronology of the burial area excavated in 1937. At the
time of this programme, only one skeleton was dated from this collection (Beta–127451, CAR1937, Sk6), which provided the oldest date
known at the Tagus valley;
• to define the temporal relation between the burial areas identified in
1964–65, their relationship to each other and to the nineteenth century
and 1937 collections.
Sampling strategy
Six samples of human bone were selected for 14C analysis (table 4.14). The sampling strategy was focused on the material excavated in 1937 and in 1964–65. Sampling of the 1937 collection was based on the stratigraphic position of each individual. Five samples were selected from different depths in the midden, including
two from the lower levels. The sampling of the 1964–65 collection was severely
constrained by the current state of preservation of the collection. The human remains are mixed and partially sorted by bone type. It was possible to identify one
individual only, which was selected for sampling (CAR1964, Sk4). Unfortunately, it
was not possible to identify the remains from the upper layers excavated in 1965.
Results
Fifteen 14C measurements on human bone collagen samples from 14 individuals
from Cabeço da Arruda are available now. Five samples belong to human remains
excavated in the nineteenth century. Seven samples belong to human remains excavated in 1937, two of which were sampled from the same skeleton (individual 6).
One sample is from the material excavated in 1964 and two samples are from the
burials excavated in 2000 (table 4.15).
202
Results: Tagus valley, Cabeço da Arruda
Figure 4.12. Cabeço da Arruda, Tagus valley 1964–65: site plan with the location of the
skeletons. The human remains excavated in 1964 were found in the lower levels (layers 5–
7), while the material excavated in 1965 was recovered from the upper levels (layer 2) at c 4
metres from the bottom (Roche 1974). Arrow: indicates the southern edge where CA–00–01
and CA–00–02 were excavated in 2000 (Roksandic 2006, 44). (Site plan after Cardoso and
Rolão 1999–2000, 227, fig. 57).
203
Lab no.
Ua–46272
Ua–47975
Ua–46273
Ua–46274
Ua–46275
Ua–47976
Individual
1, 1937**
2, 1937*
3, 1937*
9, 1937
10, 1937**
4, 1964
Adult
n/d
n/d
Adult
Adult
♂
Adult
♂
♀
Adult
Adult, 20–35 yrs
Sex
♀
Age
Tibia–R
81.19.103.4
Femur–L
Tibia
Tibia–R
Radius–R
Tibia–R
Bone
Basal sand, layer 5–7
0.25 m from basal sand
0.30 m from basal sand
1.20 m from basal sand
0.85/0.95 m from basal sand
1.40 m from basal sand
Stratigraphy
–
Previous analysis of the same bone for stable
isotope analysis (Umbelino 2006).
–
–
–
Previous analysis of the same bone for stable
isotope analysis (Umbelino 2006).
Obs.
Table 4.14. Cabeço da Arruda, Tagus valley: samples of human bone (6) collected and submitted for 14C measurement in the course of this study. Samples from the 1937
assemblage were collected at MHNUP. The sample from the 1964 material (individual 4) was collected at MG. Age and sex estimation by *A. Ataíde (1940) and **C.
Umbelino 2006.
2000
19th c.
MG
1937
MHNUP
1937
MHNUP
1937
MHNUP
19th c.
MG
CA–00–02
III
9
10
3
42
TO–359
Ua–46273
Ua–46275
Ua–46274
TO–360
TO–10216
TO–10217
2000
CA–00–01
6960±70
7198±40
7263±46
7200±41
6990±110
7040±60
6620±60
–17.2
–17.3
–17.4
–17.4
–17.7
–17.9
–18.1
–18.9
TO–355
19th c.
MG
D
6780±80
Measurements
14C by AMS; δ13C and δ15N by IRMS
14C Age BP
Lab no.
δ13C (‰)
VPDB
TO–354
6970±60
–19.0
Identification of the
human remains
Ind.
Excav.
Museum
A
19th c.
MG
11.8
10.1
11.4
10.6
11.2
10.6
10.5
10.3
δ15N (‰)
AIR
12.2
n/a
3.1
3.2
3.1
n/a
3.2
3.4
41
40
39
39
36
35
32
25
24
n/a
n/a
M1
C:N
35
34
33
33
29
27
24
14
13
M2
% marine (±10)
continued...
Lubell and Jackes 1988; Lubell et al. 1994
This study
This study
This study
Lubell and Jackes 1988; Lubell et al. 1994
Jackes and Lubell 2015; Roksandic 2006
Jackes and Lubell 2015; Roksandic 2006
Lubell and Jackes 1988; Lubell et al. 1994
Lubell and Jackes 1988; Lubell et al. 1994
Reference
Table 4.15. Human bone collagen samples (15 samples, 14 individuals) from Cabeço da Arruda, Tagus valley. 14C measurements and estimation of non-terrestrial carbon
intake (% marine). Samples are sorted so that the lowest marine percentage is on the top of the table. Marine percentage was estimated from the measured δ13C value of each
sample by the application of method marine 1 and method marine 2, based on adopted endpoint values of –12‰ and –20‰ for marine (100%) and terrestrial (100%)
diets respectively.
1964–65
MG
1937
MHNUP
1937
MHNUP
1937
MHNUP
1937
MHNUP
19th c.
MG
4
2
1
6
6
N
Table 4.15. continued
TO–356
Beta–127451
AA–101343
Ua–46272
Ua–47975
Ua–47976
6360±80
7550±100
7351±70
7166±41
7116±44
7261±45
12.5
n/a
AMS:
–19.0
–15.3
10.9
11.2
9.5
9.7
–16.6
–16.6
–16.7
–17.1
n/a
n/a
3.3
3.1
3.1
3.1
59
–
47
47
46
42
59
–
43
43
41
36
Lubell and Jackes 1988; Lubell et al. 1994
Cunha and Cardoso 2002–03; Cunha et al. 2003
Jackes et al. 2014
This study
This study
This study
Results: Tagus valley, Cabeço da Arruda
Calibration and chronological models
The chronological model was built on 15 14C dates, six of which are new dates
obtained in the scope of this study, and nine were previously published.
The two 14C measurements available for CAR1937, Sk6 (Beta–127451 and
AA–101343) were combined using the OxCal function R_Combine and calibrated
using the only δ13C value (AA–101343) obtained by IRMS. The combined result
shows good agreement (A=90.8%) (fig. 4.13).
This is a phase model which groups all dated burial events in a phase of burial
activity (Bronk Ramsey 1998, 2000, 2001). This grouping is constrained by a start of
burial activity event, and an end of burial activity event (i.e., boundaries) which imposes
no internal constraints to the relative order of the dated events. All it assumes is
that the start event occurs before the dated events, and that these all occur before
the end event. For the command files defined for this model see appendix.
Figure 4.13. Probability density function for the calibration of two bone collagen samples
(Beta–127451 and AA–101343) from CAR1937, Sk6. Calculated by OxCal 4.2 (Bronk
Ramsey 2009) using function R_Combine and IRMS–based δ13C value obtained for AA–
101343. The posterior density estimated for the death of CAR1937, Sk6 is 6133–5852 cal
BCE (95% probability), 6065–5969 cal BCE (54% probability), or 5050–5914 cal BCE (14%
probability).
Internal consistency and reliability of the model
Agreement index (A) for each dated event is high (A≥91%), except for the sample
TO–356 (CAR19th c., SkN) which is in poor agreement (Marine 1: A=49%, Marine
2: A=51.5%) but in good convergence (Marine 1, 2: C=99.2%) (fig. 4.14). This
sample, TO–356 (CAR19th c., SkN), is the most recent burial known in the site and
is not an outlier to the model, rather, its younger age may suggest a later event,
relatively more recent than the group of events in the considered phase.
The model shows satisfactory overall agreement (Marine 1: Aoverall=85.8%;
Marine 2: Aoverall=87.7%), indicating that the 14C dates are in accordance with the
207
Chapter 4
prior information incorporated in the model, which assumes one phase of burial
activity and does not impose any relative order to the dated events.
The results obtained when running the model with the different marine values
(M1, M2) are statistically insignificant (table 4.16). For this reason, all further observations and graphic presentation are based on the results obtained with the
model run with M1. The results obtained with M2 are presented in the tables.
Figure 4.14. Chronological model for the burial activity at Cabeço da Arruda, calibrated
using marine 1 diet values. For each of the dates two distributions have been plotted, one
in outline which is the result produced by the scientific evidence alone, and a solid one
which is based on the chronological model used. Calibrated ranges of 95% and 68% probability are given by the lower and higher square brackets respectively. The large square
brackets (left) and the OxCal keywords define the overall model. Model-definition command files in appendix, Tagus valley, Cabeço da Arruda, Model 1, Marine 1.
208
Results: Tagus valley, Cabeço da Arruda
The chronological model for Cabeço da Arruda, calibrated with M1 values, estimates the start of the burial activity to have been in 6270–5918 cal BCE (95% probability; start of burial activity), and its end to have been in 5424–4843 cal BCE
(95% probability; end of burial activity) (fig. 4.15; table 4.17). The burial activity is estimated to have continued for between 557 and 1093 years (95% probability: site use for
burial activity), or for 682 and 918 years (68% probability) (fig. 4.16; table 4.17).
Figure 4.15. Start and End of burial activity at Cabeço da Arruda derived from the chronological model, calibrated with values for marine 1 diet. Posterior density estimated for the
start of the burial activity is 6270–5918 cal BCE (95% probability) or 6127–5976 cal BCE
(68% probability). Posterior density estimated for the end of the burial activity is 5424–4843
cal BCE (95% probability) or 5289–5046 cal BCE (68% probability). Calibrated ranges of 95%
and 68% probability are given by the lower and higher square brackets respectively.
Figure 4.16. Duration of burial activity at Cabeço da Arruda derived from the chronological
model, calibrated with values for marine 1 diet. Posterior density estimated for the duration
of the burial activity is between 557 and 1093 years (95% probability) or 682 and 918 years (68%
probability). Calibrated ranges of 95% and 68% are given by the lower and higher square
brackets respectively.
209
Chapter 4
Table 4.16. Human bone collagen samples (14+1) from Cabeço da Arruda, Tagus valley: calibrated date
ranges according to two different approaches for the calculation of the marine percentage (M1, M2). Results
are presented in both (a) cal BCE and (b) cal BP in 95.4% confidence range. The calibrated date ranges
are probability distributions derived from scientific dating alone. The posterior density estimates (in italic)
are derived from Bayesian modelling. Samples are sorted so that the most recent dates are at the top of the
column.
(a) Cal BCE
Age BP
Unmodelled
cal BCE, M1
Calibrated date range
(95% confidence)
Modelled
Modelled
cal BCE, M1 cal BCE, M2
Posterior density estimate
(95% probability)
Ind.
Lab no.
14C
N
TO–356
6360±80
5217–4738
5464–4963
5466–4951
CA–00–01
TO–10217
6620±60
5604–5280
5602–5287
5613–5317
D
TO–355
6780±80
5735–5391
5734–5428
5754–5475
42
TO–359
6960±70
5831–5501
5827–5495
5866–5528
A
TO–354
6970±60
5907–5617
5902–5617
5973–5652
2
Ua–47975
7116±44
5913–5637
5905–5640
5970–5666
CA–00–02
TO–10216
7040±60
5965–5622
5914–5621
5976–5653
III
TO–360
6990±110
5975–5509
5972–5511
5982–5553
1
Ua–46272
7166±41
5977–5696
5975–5696
5980–5719
3
Ua–46273
7198±40
6000–5741
5993–5745
6017–5769
9
Ua–46274
7200±41
6005–5746
5998–5748
6021–5771
4
Ua–47976
7261±45
6033–5759
6021–5766
6061–5816
10
6
Ua–46275
R_Combine:
AA–101343
Beta–127451
7263±46
7418±58
6065–5794
6223–5912
6049–5798
6133–5852
6072–5836
6172–5892
Unmodelled
cal BP, M1
Calibrated date range
(95% confidence)
Modelled
Modelled
cal BP, M1
cal BP, M2
Posterior density estimate
(95% probability)
(b) Cal BP
Ind.
Lab no.
14C
Age BP
N
TO–356
6360±80
7166–6687
7413–6912
7415–6900
CA–00–01
TO–10217
6620±60
7553–7229
7551–7236
7562–7266
D
TO–355
6780±80
7684–7340
7683–7377
7703–7424
42
TO–359
6960±70
7780–7450
7776–7444
7815–7477
A
TO–354
6970±60
7856–7566
7851–7566
7922–7601
2
Ua–47975
7116±44
7862–7586
7854–7589
7919–7615
CA–00–02
TO–10216
7040±60
7914–7571
7863–7570
7925–7602
III
TO–360
6990±110
7924–7458
7921–7460
7931–7502
1
Ua–46272
7166±41
7926–7645
7924–7645
7929–7668
3
Ua–46273
7198±40
7949–7690
7942–7694
7966–7718
9
Ua–46274
7200±41
7954–7695
7947–7697
7970–7720
continued...
210
Results: Tagus valley, Cabeço da Arruda
Table 4.16b. continued
4
Ua–47976
7261±45
7982–7708
7970–7715
8010–7765
10
6
Ua–46275
R_Combine:
AA–101343
Beta–127451
7263±46
7418±58
8014–7743
8172–7861
7998–7747
8082–7801
8021–7785
8121–7841
Table 4.17. Posterior density estimates for the dates of archaeological events (Start/End) and the duration
of burial activity at Cabeço da Arruda. Data derived from model described in fig. 4.14
Modelled
cal BCE, M1
Modelled
cal BCE, M2
Modelled
cal BCE, M1
Modelled
cal BCE, M2
Distribution
Post. density estimate (68% probability)
Post. density estimate (95% probability)
Start of burial activity
6127–5976
6154–6003
6270–5918
6311–5952
End of burial activity
5289–5046
5276–5026
5424–4843
5444–4823
Span of burial activity,
duration
Overall model
agreement (Aoverall)
682–918 yrs
713–956 yrs
557–1093 yrs
582–1126 yrs
85.8%
87.7%
85.8%
87.7%
Stable isotopes: carbon and nitrogen
Previous measurements: δ13C and δ15N
The first measurements of stable isotopes (δ13C, δ15N) on human bone from
Cabeço da Arruda were carried out in the 1980s by H. Schwarcz at the McMaster
University in Canada (TO–354, TO–355, TO–356, TO–359, and TO–360). The
precision of analysis for δ13C is ± 0.1‰ and for δ15N is ± 0.2‰ (Lubell et al. 1994,
204). These five samples were also independently measured for 14C dating (see
above).
In the early 2000s, two samples from the 1937 excavations (individuals 1 and
10) were analysed in the scope of a palaeodietary study (Umbelino 2006). Like the
first measurements, these analyses were done at the McMaster University and the
precision of the results are ± 0.1‰ for δ13C and ± 0.2‰ for δ15N (Umbelino
2006, 208–9). The C:N ratios were not provided but the laboratory confirmed the
quality of the collagen and of the measurements (Umbelino 2006, 315). The δ13C
value published for CAR1937, Sk6 (Beta–127451) along with its 14C date (Cunha
and Cardoso 2002–03; Cunha et al. 2003) was obtained via AMS and not by IRSM
(Cláudia Umbelino, pers. comm.). This sample was not included in Umbelino’s
palaeodietary study (2006), and as previously discussed it should not be used for
dietary reconstructions and/or reservoir corrections. This individual was recently
re-dated (AA–101343, Jackes et al. 2014) and new stable isotope measurements
were carried out by IRSM with a precision of analysis of ± 0.1‰ for δ13C and
± 0.15‰ for δ15N.
211
Chapter 4
The latest human remains excavated at Cabeço da Arruda (CA–00–01 and
CA–00–02) were analysed for stable isotopes (TO–10216 and TO–10217)
(Roksandic 2006), independently from the 14C measurements. The C:N ratios were
not originally published but were recently reported (Jackes and Lubell 2015).
All previous measurements were included in this study (n=10), as long as they
fell within the accepted quality ranges as discussed.
Sampling strategy
In this study, the selection of samples of human remains for δ13C and δ15N measurements was fully dependent on the goals of the radiocarbon dating programme
(see above).
Results
Samples Ua–number (n=6) were measured by IRMS and the precision of the
measurements for both δ13C and δ15N is ± 0.1‰.
CAR1937, Sks 1, 10 have multiple measurements from the same bone element.
In the case of individual 10, the results are similar, but different in the case of individual 1 (table 4.18).
At Cabeço da Arruda, the human collagen δ13C values (n=16/no. individuals=14) range from –19.0‰ to –15.3‰ with a mean value of –17.3 ± 1.0‰. The
human collagen δ15N values (n=16/no. individuals=14) range from +9.5‰ to
+12.‰ with a mean value of 11.0 ± 0.9‰ (fig. 4.17, table 4.18).
16
15
14
13
δ15N (‰)
12
11
10
9
8
7
6
-22 -21 - 20 -19 -18 -17 -16 -15 -14 -13 -12
δ13C (‰)
Figure 4.17. Cabeço da Arruda, Tagus valley: correlation between the δ13C (‰) and δ15N
(‰) values of 16 samples of human bone collagen from 14 individuals. Individuals 1 (triangle) and 10 (circle) have two sets of values.
212
♂
n/d
n/d
♂?
n/d
♂
Adult
Non-adult
Adult
Non-adult,
10–11 yrs
Adult
Adult
Adult
Adult
D
19th c.
CA–00–01,
2000
CA–00–02,
2000
III
19th c.
9
1937
10
1937
10
1937
3
1937
♂
♂
Adult
A
19th c.
Sex
Age
Ind.
Identification of the human remains
Tibia–R
Femur–L
Femur–L
Tibia
n/p
Tibia
Rib
n/p
n/p
Bone
Ua–46273
n/p
Ua–46275
Ua–46274
TO–360
TO–10216
TO–10217
TO–355
7198±40
analysis for isotopes
7263±46
7200±41
6990±110
7040±60
6620±60
6780±80
–17.3
–17.2
–17.4
–17.4
–17.7
–17.9
–18.1
–18.9
10.1
11.2
11.4
10.6
11.2
10.6
10.5
10.3
Measurements
14C by AMS; δ13C and δ15N by IRMS
Age BP
Lab no.
δ13C (‰) δ15N (‰)
VPDB AIR
TO–354
6970±60
–19.0
12.2
3.1
n/p
3.2
3.1
n/p
3.2
3.4
n/p
n/p
C:N
40
41
39
39
36
35
32
25
24
M1
34
35
33
33
29
27
24
14
13
M2
% marine (±10)
This study
continued…
Umbelino 2006
This study
This study
Lubell and Jackes 1988;
Lubell et al. 1994
Jackes and Lubell 2015;
Roksandic 2006
Jackes and Lubell 2015;
Roksandic 2006
Lubell and Jackes 1988;
Lubell et al. 1994
Lubell and Jackes 1988;
Lubell et al. 1994
Reference
Table 4.18. Cabeço da Arruda, Tagus valley: carbon (δ13C) and nitrogen (δ15N) measurements on human bone collagen samples. Samples from the nineteenth century and
1964–65 assemblages were collected at MG. Samples from the 1937 material were collected at MHNUP. Samples are sorted so that the lowest δ13C values are on the top of
the table. Individuals with multiple measurements are sorted together.
♀
n/d
n/d
♀
♀
♀?
♂
Adult,
20–24 yrs
Adult
Adult
Adult
Adult
Adult
Adult,
old
42
19th c.
4
1964–65
2
1937
1
1937
1
1937
6
1937
N
19th c.
Table 4.18. continued
n/p
Fibula–R
Tibia–R
Tibia–R
Radius–R
Tibia–R
n/p
TO–356
AA–101343
n/p
Ua–46272
Ua–47975
Ua–47976
TO–359
6360±80
7351± 70
analysis for isotopes
7166±41
7116±44
7261±45
6960±70
–15.3
–16.6
–15.7
–16.6
–16.7
–17.1
–17.2
12.5
10.9
12.0
11.2
9.5
9.7
11.8
n/p
3.3
n/p
3.1
3.1
3.1
n/p
59
47
55
47
46
42
41
59
43
54
43
41
36
35
Lubell and Jackes 1988;
Lubell et al. 1994
Jackes et al. 2014
Umbelino 2006
This study
This study
This study
Lubell and Jackes 1988;
Lubell et al. 1994
Results: Tagus valley, Cabeço da Arruda
Archaeothanatology
Source material and analysis strategy
Source material for Cabeço da Arruda is restricted to the excavations by M. Corrêa
in 1937 and by J. Roche in 1964–65, however, with great limitations (table 4.19).
The documentation is relatively more abundant for the assemblage from the
1930s (MNI 11). It comprises a few field photographs of low quality (Abrunhosa
2012), simple sketches, and written documentation from field notes (Cardoso and
Rolão 1999–2000, 172–79) and one article (Ataíde 1940). The skeletal material is
mostly in disarticulated state, and just a few elements remain in small blocks of
concretion.
The material from the 1960s (MNI 13) is poorly documented. The burials excavated in 1964 (n=9) were described briefly by O. Veiga Ferreira and M. Andreata
(Cardoso and Rolão 1999–2000, 224–32), but the documentation for 1965 (n=4) is
extremely limited. O. Veiga Ferreira was sick during this campaign (Cardoso and
Rolão 1999–2000, 236–37) and he did not describe the burials, except for individual 13, which he mentions briefly. J. Roche (1974) made short descriptions of each
burial and commented on their relative chronology. He published a short synthesis,
comparing Cabeço da Arruda and Moita do Sebastião, regarding burial positions,
orientation of the bodies, and grave goods (Roche 1974). J. Roche’s notes are not
available (see chapter 2, A note on the historical documentation). The graphic documentation available is limited to a few photographs showing several individuals
(Cardoso and Rolão 1999–2000; Roche 1974) (fig. 4.18), one sketch indicating the
relative spatial position of the burials (1 to 9) (fig. 4.19), and one site plan with the
human burials (fig. 4.12). The human remains are preserved in disarticulated state,
and sorted by bone type. Archaeothanatological analysis is extremely limited and it
relies strongly on a few descriptions by the excavators.
Most human burials were excavated in the nineteenth century (MNI 80). There
are no field notes available for these excavations and the burials were not described. A few field photographs are published but it is impossible to observe the
skeletal elements in any detail (Cartailhac 1886, 56; Ribeiro 1884) (fig. 4.20). Some
of the material is preserved in small blocks of concretion but are of limited use for
archaeothanatology.
Overall, archaeothanatological analysis is extremely constrained by the absence
of appropriate graphic documentation and by the lack of human remains preserved
in blocks. The analysis was carried out partially on a total of 15 individuals, 10
excavated in 1937 (possibly 100% MNI, 1937) and 5 excavated in 1964 (c 39%
MNI, 1964–65). In most cases, the analysis is restricted to one line of evidence,
such as one description, or one photograph. Despite the difficulties it was possible
to obtain sufficient data supporting general observations regarding the treatment
of the dead, even when detailed analysis was constrained.
215
Chapter 4
Figure 4.18. Field photograph of the human burials excavated in 1964 at Cabeço da Arruda,
Tagus valley. (Reproduced from Cardoso and Rolão 1999–2000, 235, fig. 61)
Figure 4.19. Field sketch by O. Veiga Ferreira showing the relative spatial position of the
burials excavated in 1964 at Cabeço da Arruda, Tagus valley. (Reproduced from Cardoso
and Rolão 1999–2000, 235, fig. 61)
216
Results: Tagus valley, Cabeço da Arruda
Figure 4.20. Nineteenth century excavations at Cabeço da Arruda, Tagus valley. (Reproduced from Ribeiro 1884, Plates I, II)
217
1, 1937
2, 1937
3, 1937
4, 1937
5, 1937
6, 1937
7, 1937
8, 1937
9, 1937
10, 1937
1, 1964
2, 1964
3, 1964
4, 1964
5, 1964
6, 1964
7, 1964
8, 1964
9, 1964
10, 1965
11, 1965
12, 1965
13, 1965
Individual
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
Block
yes
yes
yes
n/a
yes
yes
yes
yes
yes
n/a
n/a
w/ other burials
n/a
n/a
w/ other burials
w/ other burials
w/ other burials
w/ other burials
w/ other burials
n/a
n/a
n/a
n/a
Field photographs
sketch
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
sketch w/ 1–9
sketch w/ 1–9
sketch w/ 1–9
sketch w/ 1–9
sketch w/ 1–9
sketch w/ 1–9
sketch w/ 1–9
sketch w/ 1–9
sketch w/ 1–9
n/a
n/a
n/a
n/a
Drawings
Graphic documentation w/human remains
Source material
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
Site plan
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes, ltd
yes
yes, ltd
yes
yes
yes
yes
yes
yes
yes, ltd
yes, ltd
yes, ltd
yes, ltd
Written
documentation
Limited documentation. Archaeothanatology is limited.
Limited documentation. Archaeothanatology is extremely limited.
Poor preservation and unclear documentation. Archaeothanatology is extremely limited.
Limited documentation. Archaeothanatology is extremely limited.
Limited documentation. Archaeothanatology is limited.
Limited documentation. Archaeothanatology is limited.
Limited documentation. Archaeothanatology is extremely limited.
Limited documentation. Archaeothanatology is extremely limited.
Limited documentation. Archaeothanatology is extremely limited.
Unclear documentation. Archaeothanatology is extremely limited.
Limited documentation and very fragmented material. Archaeothanatology is not possible.
Limited documentation. Archaeothanatology is extremely limited.
Limited documentation and very fragmented material. Archaeothanatology is not possible.
Limited documentation. Archaeothanatology is extremely limited.
Limited documentation. Archaeothanatology is extremely limited.
Limited documentation. Disturbed and fragmented material. Archaeothanatology is not possible.
Limited documentation. Archaeothanatology is limited.
Limited documentation. Archaeothanatology is extremely limited.
Limited documentation. Disturbed and fragmented material. Archaeothanatology is not possible.
Limited documentation and very fragmented material. Archaeothanatology is not possible.
Limited documentation and very fragmented material. Archaeothanatology is not possible.
Limited documentation. Disturbed and fragmented material. Archaeothanatology is not possible.
Limited documentation and very fragmented material. Archaeothanatology is not possible.
Observations
Table 4.19. Cabeço da Arruda, Tagus valley: source material assessed for archaeothanatological analysis. Material from 1937 is curated at MHNUP and the assemblage
from 1964–65 is located at MG.
Results: Tagus valley, Cabeço da Arruda
Results of the archaeothanatological analysis
The nature of the deposits
Analysis of the nature of the deposits at Cabeço da Arruda suggests a pattern of
primary burials (table 4.20) indicating that the cadavers were placed in the burial
features before the degradation of the labile joints. In most cases the labile articulations are unclear in the documentation, however, the overall position of the bones
strongly indicates that the bodies were disposed while retaining their general anatomical integrity, supporting the interpretation of burials in primary position. In
three cases (CAR1937, Sks 7, 8, 9) the analysis was not possible, and the nature of
these deposits is unknown.
The space of decomposition of the cadaver
Analysis of the space of decomposition of the cadavers buried in Cabeço da Arruda suggests a pattern of filled spaces (table 4.21). The cadavers were covered
with sediment immediately after their placement in the features. The decomposition of the bodies took place in this sediment-filled environment. In several cases
(CAR1937, Sks 2, 3, 5, 6, 10; CAR1964, Sks 5, 8), the interpretation was based on
the observation that the bones of the lower limbs were rotated and suspended, but
in articulation. These are not the typical diagnostic criteria for the identification of
the space of decomposition, but the maintenance of these skeletal elements in
unbalanced position, indicates support during decomposition. In every case, the
movement of the bones is limited suggesting that the space of decomposition was
filled and that the sediment could penetrate immediately.
The low resolution of the source material does not allow the identification of
original empty spaces in the grave feature and/or the formation of secondary
empty spaces outside the volume of the cadaver within the overall filled environment. The hypothesis of mixed spaces of decomposition cannot be excluded.
219
♀
♀
♂
n/d
♀?
n/d
n/d
♂
n/d
n/d
n/d
n/d
n/d
Adult
Adult
Adult
Non-adult
Non-adult
Adult
Non-adult
Non-adult
Adult
Adult, 20–35 yrs
Adult
Adult
Non-adult
Adult
Adult, young
1, 1937**
2, 1937*
3, 1937*
4, 1937
5, 1937
6, 1937
7, 1937
8, 1937
9, 1937
10, 1937**
2, 1964
4, 1964,
5, 1964
7, 1964
8, 1964
n/d
n/d
Sex
Age
Individual
Probably primary
Probably primary
Probably primary
Probably primary
Probably primary
Probably primary
Unknown
Unknown
Unknown
Probably primary
Probably primary
Probably primary
Probably primary
Probably primary
Primary
Nature of the deposit
Unclear documentation.
Maintenance of the general anatomical integrity of the body (according to field description).
Limited documentation and poor preservation.
Maintenance of labile articulations: unknown, but presence of some phalanges.
Maintenance of the general anatomical integrity of the body.
Limited documentation.
Maintenance of labile articulations: bones of the right hand.
Maintenance of the general anatomical integrity of the body.
Limited documentation.
Maintenance of the general anatomical integrity of the body.
Limited documentation.
Maintenance of labile articulations: bones of hands.
Maintenance of the general anatomical integrity of the body.
Limited documentation.
Maintenance of the general anatomical integrity of the body.
Unclear documentation
Unclear documentation.
Maintenance of labile articulations: some bones of left hand; some bones of feet.
Maintenance of the general anatomical integrity of the body.
Unclear documentation.
Maintenance of the general anatomical integrity of the body.
Poor preservation and unclear documentation.
Maintenance of the general anatomical integrity of the body.
Limited documentation.
Maintenance of labile articulations: CV, metatarsals and phalanges of left foot.
Unclear documentation.
Maintenance of the general anatomical integrity of the body.
Unclear documentation.
Maintenance of the general anatomical integrity of the body.
Unclear documentation.
Arguments
Table 4.20. Summary of the nature of the deposits with human remains at Cabeço da Arruda, Tagus valley. Age and sex estimation by *A. Ataíde (1940) and **C.
Umbelino (2006).
♀
♀
♂
n/d
♀?
n/d
n/d
♂
n/d
n/d
n/d
n/d
n/d
Adult
Adult
Adult
Non-adult
Non-adult
Adult
Non-adult
Non-adult
Adult
Adult, 20–35 yrs
Adult
Adult
Non-adult
Adult
Adult, young
1, 1937
2, 1937
3, 1937
4, 1937
5, 1937
6, 1937
7, 1937
8, 1937
9, 1937
10, 1937
2, 1964
4, 1964
5, 1964
7, 1964
8, 1964
n/d
n/d
Sex
Age
Individual
Probably filled
Filled
Probably filled
Unknown
Unknown
Probably filled
Unknown
Unknown
Unknown
Probably filled
Probably filled
Probably filled
Probably filled
Probably filled
Filled
Space of decomposition
Limited documentation.
Maintenance of bones in original unbalanced position: bones of lower limbs are rotated and suspended.
Limited documentation.
Maintenance of labile articulations: bones of hands in front of the os coxae.
Limited documentation.
Maintenance of bones in original unbalanced position: bones of lower limbs are rotated and suspended.
Limited documentation.
Maintenance of bones in original unbalanced position: bones of lower limbs are rotated and suspended
(according to field description).
Limited documentation.
–
–
Maintenance of bones in original unbalanced position: lower limbs are rotated in front of the upper body.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: bones of the lower limbs are rotated and suspended.
No movement outside initial volume of the cadaver
–
Maintenance of bones in original unbalanced position: some bones of the left hand are rotated and flexed at
the wrist and lie on the lower abdominal region; several bones of the feet.
Maintenance of bones in original unbalanced position: bones of lower limbs are rotated.
No movement outside initial volume of the cadaver
Maintenance of bones in original unbalanced position: bones of lower limbs are rotated and suspended.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: bones of feet are rotated.
Arguments
Table 4.21. Summary of the space of decomposition of the human remains in the grave features at Cabeço da Arruda, Tagus valley.
Chapter 4
The initial position of the cadaver in the feature
The most common position at Cabeço da Arruda consists of a cadaver lying on the
back (n=12) with the lower limbs in flexion or hyperflexion (table 4.22). In three
cases (CAR1937, Sks 7, 8, 9), it was not possible to determine if the bodies were
lying on the back or on the side.
The position of the upper limbs was varied, but they were always placed close
to the body, in extension, semi-flexion, flexion or hyperflexion.
At Cabeço da Arruda, the bodies were placed in the feature with the lower
limbs in semi-flexion, flexion or hyperflexion at the level of the hip and knee
joints. In most cases the feet were forced towards the buttocks in a constrained
position. In a few cases (i.e., CAR1937, Sks 1, 2) that was not the case. Often, the
rotation of the lower limbs was unclear in the documentation. In some cases, the
knees were directed forwards (CAR1964, Sk5), or towards the right or left side
(CAR1934, Sks 1, 2, 10). In seven cases the hip and knee joints were in hyperflexion (CAR1937, Sks 4, 5, 7, 8, 9; CAR1964, Sks 2, 5), and the limbs were in adduction and rotated towards the trunk. Other burials (CAR1964, Sks 9, 13), without
sufficient documentation for archaeothanatological analysis, have been described
in a similar contracted position (Roche 1974). In a few cases (CAR1937, Sks 7, 8,
9), clumping of the body was enhanced during decomposition by the progressive
closure of the angles between the bone segments and the skeletons became hypercontracted.
The grave features
The size and shape of the deposits with human remains were probably difficult to
determine during excavation. J. Roche (1974) described the grave features as natural depressions, or as shallow pits excavated in the basal sand layers. CAR1965, Sks
10, 11, 12, 13 were excavated in the upper layers (layer 2), c 4 metres from the
bottom of the shell midden, but these features were not described, except for short
descriptions of the human remains (Roche 1974).
The archaeothanatological analysis suggests that the lateral pressures and walleffects observed on the skeletons in the features indicate that the shape and size of
the pit was just sufficient to contain the body (table 4.23). In several cases, the
design of the grave had a significant impact on the final arrangement of the body
in the feature, visible by the constrained position of several skeletons, particularly
at the level of the lower limbs. Often, the lower limbs were supported by the wall
of the grave, keeping the feet towards the buttocks. While the pit imposed a physical limit to the cadaver, by maintaining the lower limbs clumped together, it also
provided lateral support preventing the collapse of the skeletal elements. In some
cases (CAR1937, Sks 1, 2, 3, 6), the grave pits were somewhat wider allowing a
relatively expanded position, at least at the level of the lower limbs.
Analysis of the upper body is limited and it is not possible to provide further
descriptions at this level. Likewise, the low resolution of the source material does
222
Results: Tagus valley, Cabeço da Arruda
not allow the analysis of the presence or absence of extra support offered by some
kind of container, such as a light wrapping of the cadaver at the time of burial.
The floor of the graves was possibly uneven and sloping, but the documentation is in most cases unclear. The moderate downslope is suggested by the rotation
forwards and downwards of the head, towards the body in the feature.
Deposits containing more than one individual
At Cabeço da Arruda, all deposits apparently contain the remains of one individual.
Post-depositional manipulations of the cadavers
J. Roche (1974) indicates that a few skeletons have been disturbed by more recent
burials. That is the case in CAR1964, Sks 6, 9, which were disturbed by the burial
of CAR1964, Sk7. Also, CAR1964, Sk3 was buried above the lower limbs of
CAR1964, Sk2. All these cases indicate the removal and disturbance of previous
burials to give place to a new cadaver. Despite this practice of reduction, the bodies at Cabeço da Arruda are apparently complete and lie generally undisturbed in
the grave feature.
The low resolution of the source material does not allow further observations at
this level. It is unknown if the practice of re-opening of the grave for the removal
or rearrangement of selected elements occurred at Cabeço da Arruda.
The mortuary programme at Cabeço da Arruda
Archaeothanatological analysis of the human remains excavated at Cabeço da
Arruda is extremely constrained by the general lack of documentation. Despite the
limitations, it is possible to observe some patterns in the mortuary practices at the
site.
In the sample analysed, the cadavers were typically placed in individual grave
features while retaining their anatomical integrity (individual primary deposits). The
feature was then immediately covered with sediment, providing a filled space of
decomposition.
Typically, the cadavers were lying on the back with the upper limbs positioned
close to the body in various arrangements. Often, the lower limbs were placed in
contraction, with the feet forced towards the buttocks.
In most cases, the pressures exerted on the bodies in the features can be explained as the result of the design of the grave. However, the data is not conclusive
and the possibility of extra support provided by soft wrapping of the body at the
time of burial cannot be completely excluded.
Overall, the mortuary programme at Cabeço da Arruda presents a consistent set
of practices and all cases observed follow a common way of burying the dead.
223
♀
♂
n/d
♀?
Adult
Adult
Non-adult
Non-adult
Adult
2, 1937
3, 1937
4, 1937
5, 1937
6, 1937
n/d
♀
Sex
Adult
Age
1, 1937
Individual
Prob. rotation–L
Slightly raised
Rotated–L and downwards
(neck flexed forwards)
Rotation downwards
(neck flexed forwards)
Rotated–R and downwards
(neck flexed forwards)
n/d
Rotation–L
Head
Initial position of the cadaver in the feature
On the back
On the back
On the back
On the back
On the back
On the back
Trunk
In extension–R
n/a–L
n/a
In hyperflexion
In extension–R
In semi-flexion?–L
In flexion–R
In extension–L
In extension–R
In semi-flexion–L
Upper limbs
continued…
Hip and knee joints–R: semi-flexion or flexion
Hip joint–L: flexion
Knee joint–L: hyperflexion
Left foot towards buttocks
In hyperflexion
In adduction in front of trunk
Feet towards buttocks
In hyperflexion
In adduction in front of trunk
Feet towards buttocks
Hip joint–R: semi-flexion (c 90)
Knee joint–R: n/a
Hip joint–L: semi-flexion
Knee join– L: prob. hyperflexion
Left foot towards buttocks
Hip joints: semi-flexion
Knee joints: semi-flexion
Rotation–L
Hip joints: semi-flexion
Knee joints: semi-flexion
Rotation–R
Lower limbs
Table 4.22. Summary of the reconstruction of the initial position of the cadaver in the grave features at Cabeço da Arruda, Tagus valley. The initial position of the limbs is
indicated as –R or –L and refers to the right or left limb, when their positions are different. Rotation –R or –L indicates the rotation of the skeletal element(s) towards the
right or the left side.
n/d
n/d
♂
n/d
n/d
n/d
n/d
n/d
Non-adult
Adult
Adult, 20–35 yrs
Adult
Adult
Non-adult
Adult
Adult, young
8, 1937
9, 1937
10, 1937
2, 1964
4, 1964
5, 1964
7, 1964
8, 1964
n/d
Non-adult
7, 1937
Table 4.22. continued
n/d
Rotation downwards
(neck flexed forwards)
n/d
n/d
Rotation downwards
(neck flexed forwards)
Cranium: tendency–R
Maxilla and mandible: rotation–L
n/d
n/d
n/d
On the back
On the back
On the back
On the back
On the back
On the back
n/d
n/d
n/d
In extension or semiflexion
In semi-flexion
n/a
In semi-flexion–R
In extension or semiflexion–L
In extension or semiflexion
In flexion–R
n/a–L
In hyperflexion
In hyperflexion
In hyperflexion
Hip joints: semi-flexion (c 90°)
Knee joints: hyperflexion
Feet towards buttocks
Hip joints: semi-flexion (c 90°)
Knee joints: hyperflexion
Feet towards buttocks
In hyperflexion
Feet towards buttocks
Hip joints: semi-flexion or flexion
Knee joints: hyperflexion
Feet towards buttocks
In hyperflexion
Rotation–tendency L
Feet towards buttocks
Prob. in semi-flexion or flexion
Rotation–tendency L
n/a – L
In adduction?
In hyperflexion
Feet towards buttocks
Clumping
In hyperflexion
Feet towards buttocks
Clumping
In hyperflexion
Feet towards buttocks
Clumping
♀
♀
♂
n/d
♀
n/d
n/d
♂
n/d
n/d
n/d
n/d
n/d
Adult
Adult
Adult
Non-adult
Non-adult
Adult
Non-adult
Non-adult
Adult
Adult, 20–35 yrs
Adult
Adult
Adult
Adult
Adult, young
1, 1937
2, 1937
3, 1937
4, 1937
5, 1937
6, 1937
7, 1937
8, 1937
9, 1937
10, 1937
2, 1964
4, 1964
5, 1964
7, 1964
8, 1964
n/d
n/d
Sex
Age
Individual
1. Overall strong bilateral compression and constraint effects.
Limited documentation.
1. Overall strong bilateral compression and constraint effects.
Limited documentation.
1. Overall strong bilateral compression and constraint effects.
Limited documentation.
1. Prob. wall-effect, upper right side.
Limited documentation.
1. Prob. wall-effect, lower end.
Limited documentation.
1. Prob. wall-effect, lower end.
Limited documentation.
1. Prob. wall-effect, lower end.
Limited documentation.
1. Prob. wall-effect, lower end.
Limited documentation.
1. Prob. wall-effect, lower end.
Limited documentation.
1. Overall strong bilateral compression and constraint effects.
Unclear documentation.
1. Overall strong bilateral compression and constraint effects.
Unclear documentation.
1. Sloping floor;
2. Prob. wall-effect, left side and lower end.
Unclear documentation.
n/a
n/a
1. Prob. wall-effect, upper right side.
Unclear documentation.
Pressures
1. Alignment of lower limbs and constraint effect forcing feet towards
buttocks.
1. Alignment of lower limbs and constraint effect forcing feet towards
buttocks.
1. Suggested by constrained position of lower limbs. Feet towards buttocks.
1. Suggested by constrained position of lower limbs. Feet towards buttocks.
1. Suggested by constrained position of lower limbs. Feet towards buttocks.
1. Suggested by projection forwards and upwards of right shoulder and arm.
1. Hypercontracted skeleton.
1. Hypercontracted skeleton.
1. Hypercontracted skeleton.
1. Transfer of body weight from the right to the left side. Suggested by LV
and the os coxae which present a tendency towards the left side;
2. Alignment of lower left limb in constrained position. Left foot towards
buttocks.
1. Suggested by clumping of the skeleton.
1. Suggested by clumping of the skeleton.
n/a
n/a
1. Alignment of upper right limb and projection forwards and upwards of
right shoulder and arm. Suggested by oblique position of left clavicle and
alignment of upper right limb.
Arguments
Table 4.23. Summary of the key observations used in the reconstruction of the grave features at Cabeço da Arruda, Tagus valley. Observed pressures on the body are presented in a number sequence. The arguments supporting these observations are indicated by the same number in the respective column.
Sado valley: Arapouco
Site background: a short history of research
Arapouco (also known as Ara Pouca) is the western-most shell midden site in the
Sado valley (fig. 2.3). The site is located on the left margin of the Sado River at
47 m.a.s.l. (Diniz et al. 2012, 13), and has excellent visibility over the river valley.
Archaeological material and documentation available for Arapouco are curated
at the National Museum of Archaeology in Lisbon (MNA). The year of its identification is unknown. The earliest reference to Arapouco comes from one letter
(MNA, APMH, 5/1/654/50, Torrão, 25 Aug. 1960) written by J. Roldão to
M. Heleno, where Roldão mentions his intentions of approaching the owner of the
land (Santa Casa da Misericórdia) to request a fieldwork permit. At the time of this
letter, J. Roldão was working in the region, at the sites of Barrada do Grilo and
Vale de Romeiras, while planning the fieldwork for Poças de S. Bento.
Graphic documentation in the archives of MNA indicates that Arapouco was
excavated by the museum staff in 1961 and 1962. Unfortunately, all written documentation for the years of 1961, 1962, and 1963 relative to the shell middens is
missing from the museum archives. One site plan of Arapouco (MNA, 1961,
D. Sousa, A120a) shows the areas excavated in the field season of 1961 (fig. 4.21).
This is the only site plan available, thus, there is no information about site profiles
or precise location of the human burials. According to Dario de Sousa who did the
site drawings, all human burials were in the basal sandy layers of the site (Dario de
Sousa, pers. comm.), which is apparently a common feature of the burials in the
Sado middens. There are no field drawings with the human remains excavated at
Arapouco. Nevertheless, the museum’s artist has confirmed that he did all field
drawings for all the shell middens excavated by the museum after 1958 (Dario de
Sousa, pers. comm.); however, these are missing from the museum. Other researchers have commented about the missing field documentation for Arapouco
(Umbelino 2006, 141).
Archives of photography of Arapouco consist of 154 photographs (plus 34
photographs in poor preservation) from 1958, 1961 and 1962 (MNA, APMH,
2/11/8). Photographs from 1958 were probably taken during a preliminary survey,
before the first excavation in 1961. The 1961 archive consists of photographs from
15 individuals. Photographs from 1962 do not show human remains.
Human remains excavated in 1961 are numbered from 1 to 15. The sequence of
human remains excavated in 1962 starts again with the number 1, but the series is
distinguished by an additional A, such as 1A, 2A, 3A, and so forth. The assumption about the correlation between the year of excavation and the numbers of the
skeletons is based on two observations: (1) the field photographs have handwritten
notes on the reverse, and all photographs with human remains are dated to 1961
227
Chapter 4
and the skeletons are identified from 1 to 15; (2) in the inventory list at MNA
(fichas de desmontagem) all skeleton numbers from 1962 are identified with the additional A, such as 1A 1962, 2A 1962, 3A 1962, and so forth.
Four human skeletons are preserved in blocks of paraffin wax (ARA1961, Sks
5, 7; ARA1962, Sks 1A, 3A). The remaining skeletons are in disarticulated state
sorted in containers.
The human remains were studied by a team of anthropologists and the MNI
curated at the museum is 32 (Cunha and Umbelino 1995–97, 2001; Umbelino and
Cunha 2012). This number suggests that if 15 individuals were excavated in 1961,
as suggested by the field photographs, 17 individuals may have been recovered in
1962. The burial population consists of 6 non-adults, and 26 young (20–30 yrs) and
mature adults (>35 yrs) (table 4.24). The adult group comprises 14 males and 10
females. In two cases the sex of the individual was not possible to determine. The
group of non-adults (Umbelino 2006, 141) consists of two adolescents (c 15 yrs ±
36 months), and four children, three of which are under 6 years old. The youngest
child buried at Arapouco is 1.5 to 2.5 years old, followed by a 3 year ± 12 months,
and a 3 to 7 year old child. The oldest of these children is c 9 yrs ± 24 months.
228
Results: Sado valley, Arapouco
Figure 4.21. The only site plan available for Arapouco, Sado valley. Dots: area of the shell
midden. Long narrow rectangles: test trenches excavated in 1961 (Sonda 1 to 5). Talhão T: main
excavation area where 15 human burials were recovered in 1961. (Adapted from a photograph by J. P. Ruas. MNA, 1961, D. Sousa, A120a)
Table 4.24. Arapouco, Sado valley: minimum number of individuals (MNI) estimated for the material
curated at MNA.
MNI
Adults
♀
10
32
Non-adults
Reference
♂
14
26
6
Cunha and Umbelino 1995–97
229
Chapter 4
The following table (table 4.25), courtesy of one of the anthropologists who studied this material in detail (Umbelino 1999), is the reference used in this study regarding sex and age estimates for Arapouco.
Table 4.25. Arapouco, Sado valley: sex and age estimates of 22 individuals. (Data from Cunha and
Umbelino 1995–97, 2001; Umbelino 1999)
Individual
Container
Age
Sex
1
6229
Adult
♂
3
6223
Adult, mature
♀
4
6221
Adult, mature
♀
6
6227
Adult, mature
♀
8
6224
Adult, mature
♂
9
6236
Adult
♀
13
6226
15 yrs±36 m.
n/d
14
6231
9 yrs±24 m.
n/d
11
6228
Adult, young
♀
12
6228
3 yrs±1 m.
n/d
2A
6958
Adult, mature
♀
4A
6163
Adult, young
♀
5A
6222
Adult, mature
♂
6A
6189
Adult, mature
♂
7A
6209/6202
Adult
♂
9A
6121
Adult, mature
♂
11A
24046/6217
Adult
♂
13A1
6216
Adult, young
♀
13A2
6198
Adult, young
♂
14A
6207
Adult, mature
♂
15A
6120
Adult, mature
♂
16A
6215
Adult, mature
n/d
Observations about one mislabelled individual
In the course of this dissertation I identified one mislabelled individual in the
Arapouco assemblage. This was individual 5 (adult, possibly male), whose remains
are preserved in two blocks of paraffin wax (fig. 4.22). Until October 2013, these
two blocks were identified as two different individuals. The block with the upper
skeleton was identified as Vale de Romeiras and did not have an individual number,
while the block with the lower limbs was identified as Arapouco, skeleton 5, container
6225 A and B. Another block, containing mostly the upper skeleton of an adult,
possibly female, was identified with the label Arapouco, skeleton 5, container 6225–A
(fig. 4.23). The re-identification of skeleton 5 from Arapouco was carried out by
the comparison of the field photographs with the human remains preserved in
230
Results: Sado valley, Arapouco
blocks and by the clear mismatch between the blocks containing the upper and
lower skeletons which were originally identified as Arapouco, skeleton 5. The provenience of the individual mislabelled as Arapouco, skeleton 5 is unknown. The
observation of the sediment seems to exclude it from Arapouco, and place it
possibly at Cabeço do Pez. This tentative identification is based on similarities of
the soil found in this block and in the block with the remains of CP1956, Sk2. The
soil in both blocks is powdery brown sediment containing very small fragments of
shells and a few complete shells of common cockle.
M. Fontanals-Coll and colleagues (2014) sampled a tibia from the individual
mislabelled as Arapouco, Skeleton 5, which they identified as a female adult.
Because of this mislabelling, the isotopic data for this individual was wrongly
published by the team as Arapouco, Skeleton 5. The skeleton is likely to be from
another site, presumably Cabeço do Pez. This problem is discussed further in
chapter 5, Isotopic signatures: defining the people in each burial place.
Figure 4.22. Re-identification of individual from Arapouco, Sado valley. The field photograph on the right (c) is identified as Arapouco, Skeleton 5 and it matches the human remains in the two paraffin blocks (a, b). In 2013 the block with the upper skeleton (a) was
wrongly labelled as Vale de Romeiras without a skeleton number, while the portion with the
lower limbs (b) was identified as Arapouco, skeleton 5, container 6225 A and B. (Left, photographs by author at MNA. Right, MNA, APMH, F398)
231
Chapter 4
Figure 4.23. This paraffin block was mislabelled Arapouco, Skeleton 5. The provenience of this
burial is unknown. This skeleton in preserved in one paraffin block at MNA. (Photograph
by J. P. Ruas)
Chronology of the burial activity
Previous 14C measurements
The first 14C measurements of samples from the Sado valley shell middens were
submitted in the late 1980s (Arnaud 1989). One sample of shells (Q–2492) reported to be recovered from the middle layers during the 1961–62 excavations was
the first 14C date available for Arapouco (table A8).
The first measurements on human bone from Arapouco (Sac–1560) were
carried out in the late 1990s (Cunha and Umbelino 2001; Umbelino 2006). The
sample was collected from one individual (2A) excavated in 1962. The δ13C value
232
Results: Sado valley, Arapouco
published along with this 14C measurement (Sac–1560) was obtained by AMS
(Cláudia Umbelino, pers. comm.) and cannot be used for dietary reconstructions or
reservoir corrections (Millard 2014, 557; Taylor and Bar-Yosef 2014, 117). Fortunately, recent IRMS measurements on this individual (Guiry et al. 2015) allow the
calibration of this 14C date.
Objectives of the dating programme
Contextual information about the burials of Arapouco is poor. The basic aims of
the dating programme in relation to the macro-goals outlined in chapter 3 were:
• to estimate the duration of the use of this site for burial activity;
• to estimate when the burial activity started and when it ended.
The site-specific aims of the dating programme for Arapouco were:
• to confirm the apparent antiquity of the site indicated by a non-human
sample (Q–2492, 6374–5706 cal BCE, 95% confidence) (table A8).
Sampling strategy
Nine samples of human bone were selected for 14C analysis (table 4.26). The sampling strategy was targeted towards the burials with photographs and/or preserved
in blocks.
Results
None of the nine samples (Ua–number) provided reliable results (table A1). Three
samples (Ua–46937, Ua–47968, and Ua–47969) did not yield enough collagen for
reliable 14C measurements. In three samples (Ua–47970, Ua–47971, and Ua–
47972) the C:N ratio is higher than the acceptance range (2.9–3.6). In two samples
(Ua–46938 and Ua–46940) the C:N ratio is within the 3.4–3.6 range suggesting
contamination which is further confirmed by the high δ15N values (34.5 ‰ and
21.5 ‰ respectively) which are out of range. One of the nine samples (Ua–46939)
was later recognized to be collected from a individual mislabelled as skeleton 5; as
discussed above its provenience is unclear and it is therefore excluded from the
analysis (table A2).
Calibration and chronological models
Only one 14C measurement on human bone collagen from Arapouco is available
(table 4.27). This sample (Sac–1560) was calibrated with the recently published
δ13C values obtained by IRMS (table 4.28, fig. 4.24). This one date does not allow
the construction of a chronological model for the burial activity at Arapouco.
233
Chapter 4
Table 4.26. Samples of human bone (9) from Arapouco collected and submitted for 14C measurement in
the course of this study. With one exception (Ua–46939), all samples are from the 1961 assemblages and
were collected at MNA. Sample Ua–46939 was collected from a skeleton mislabelled as Arapouco, skeleton 5 and its provenience remains unclear. *Samples pre-treated with boiling in acetone, ultrasonification
first with ether and then with ethanol, and finally filtrated; after that processed as usual (see chapter 3).
Ind. Lab no.
Age
Sex
Bone
Site plan
Obs.
1
Ua–46937
Adult
♂
Femur–R
999.49.20
Trench T
Not measurable
3
Ua–46938
Adult,
35–50 yrs
♀
Femur–L
999.50.32
Trench T
Doubtful quality
5
Ua–47968*
Adult
♂
Fibula–L
Trench T
Paraffin block
Not measurable
6
Ua–47969*
Adult,
>40 yrs
♀
Radius–R
999.66.5
Trench T
Not measurable
7
Ua–47970*
Adult,
young
n/d Radius–R
Trench T
Paraffin block
Doubtful quality
8
Ua–47971*
Adult,
35–50 yrs
♂
Humerus–L Trench T
999.51.43
Doubtful quality
11
Ua–46940
Adult,
20–35 yrs
♀
Long bone
Trench T
Multiple–double burial
Doubtful quality
13
Ua–47972*
Non-adult,
15 yrs±36 m.
n/d Radius
999.61.24
Trench T
Doubtful quality
?
Ua–46939
Adult
♀
unknown
Paraffin block
Mislabelled as ARA1961, Sk5
Provenience is unknown
Femur
Table 4.27. Human bone collagen sample (1) from Arapouco, Sado valley. 14C measurement and estimation of non-terrestrial carbon intake (% marine). Marine percentage was estimated from the measured
IRMS-based δ13C value by the application of method marine 1 and method marine 2, based on adopted
endpoint values of –12‰ and –20‰ for marine (100%) and terrestrial (100%) diets respectively.
Identification of the Measurements
14C by AMS; δ13C and δ15N by IRMS
human remains
Ind.
2A
234
Excav.
Museum
1962
MNA
Lab no.
δ13C
(‰)
VPDB
δ15N
(‰)
AIR
C:N
M1
M2
Sac–1560 7200±130* –17.2#
12.1#
3.4#
41
35
14
C Age BP
% marine
(±10)
Reference
*Cunha and Umbelino 2001;
Guiry et al. 2015
#
Results: Sado valley, Arapouco
Table 4.28. Human bone collagen sample from Arapouco, Sado valley: calibrated date ranges according to
two different approaches for the calculation of the marine percentage. Results are presented in both cal BCE
and cal BP in 95.4% confidence range.
Individual
Lab no.
14C
Age (BP)
2A, 1962
Sac–1560
7200±130
Calibrated date range
(95% confidence)
cal BP
Calibrated date range
(95% confidence)
cal BCE
8181–7662, M1
8193–7673, M2
6232–5713, M1
6244–5724, M2
Figure 4.24. Probability density function estimated for the death of ARA1962, Sk2A is
6232–5713 cal BCE (95% confidence, marine 1) or 6082–5820 cal BCE (67% confidence,
marine 1). Calibrated ranges of 95% and 68% probability are given by the lower and higher
square brackets respectively.
Stable isotopes: carbon and nitrogen
Previous measurements: δ13C and δ15N
In the early 2000s, one sample from the 1961 excavations (individual 9) was analysed at the McMaster University in Canada in the scope of a palaeodietary study,
however, the sample did not yield enough collagen for reliable measurements
(Umbelino 2006; Umbelino and Cunha 2012).
Recently, two teams sampled this assemblage for stable isotope analysis of carbon and nitrogen (Fontanals-Coll et al. 2014; Guiry et al. 2015). M. Fontanals-Coll
and colleagues (2014) analysed four samples at the Biological Anthropology
Department at the Universitat Autònoma de Barcelona. Unfortunately, the only
235
Chapter 4
sample that produced results was wrongly labelled as skeleton 5 at the time of the
sampling (see above, Site background: a short history of research). This sample was obtained from the tibia of an adult female while the real skeleton 5 from Arapouco is
an adult male preserved in a paraffined block and his long bones were not sampled. As discussed, the real provenience of this adult female is unknown and for
this reason these measurements (table A2) must be excluded from the Arapouco
series. Despite the poor collagen preservation of this assemblage, E. Guiry and
colleagues (2015) obtained 12 reliable measurements (table 4.29) from a total of 19
samples. This material was measured on a Thermo-Finnigan Delta Plus XL isotope
ratio mass spectrometer coupled via continuous flow to a Carlo Erba elemental
analyzer at the Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany. The instrumental error for δ13C and
δ15N measurements was ± 0.1‰ and ± 0.2‰, respectively (Guiry et al. 2015).
The δ13C value published for individual 2A (Sac–1560) along with its 14C date
(Cunha and Umbelino 2001; Umbelino 2006) was obtained by AMS and not by
IRMS (Cláudia Umbelino, pers. comm.). This sample was not included in
Umbelino’s palaeodietary study (2006) and as previously discussed it should not be
used for dietary reconstructions and/or reservoir corrections.
Sampling strategy
In this study, the selection of samples of human remains for δ13C and δ15N measurements was fully dependent on the goals of the radiocarbon dating programme
(see above).
Results
None of the Ua–number samples (n=8) provided reliable results (tables A2, A9).
Overall, the collagen preservation at Arapouco is poor. Despite the multiple samples collected (n=32) which were analysed by four teams, only less than one third
of the samples (n=12) presented measurements within the accepted quality ranges.
At Arapouco, the human collagen δ13C values (n=12/no. individuals=12) range
from –17.9‰ to –16.1‰ with a mean value of –17.0 ± 0.5‰. The human collagen
δ15N values (n=12/no. individuals=12) range from +11.0‰ to +13.0‰ with a
mean value of +12.2 ± 0.4‰ (table 4.29, fig. 4.25).
236
Results: Sado valley, Arapouco
Table 4.29. Arapouco, Sado valley: carbon (δ13C) and nitrogen (δ15N) measurements on human bone
collagen samples (12) collected at MNA. Samples are sorted so that the lowest δ13C values are on the top
of the table.
Identification of the human remains
Individual
Age
4A, 1962
Adult, ♀
young
Adult, n/d
mature
Adult ♂
16A, 1962
11A, 1962
2A, 1962
7A, 1962
Sex
Adult, ♀
mature
Adult ♂
Adult,
mature
Adult,
9A, 1962
mature
14A, 1962 Adult,
mature
Adult,
8, 1961
mature
13A2, 1962 Adult,
young
Adult,
6, 1961
mature
Adult,
6A, 1962
mature
3, 1961
Bone
♀
♂
♂
♂
♂
♀
♂
Long bone
999.64.68
Long bone
999.57.24
Long bone
999.55.96
Long bone
Cont. 6958
Long bone
999.69.106
Long bone
999.50.44
Long bone
999.54.41
Long bone
999.65.99
Long bone
999.51.35
Rib
999.53.77
Rib
Cont. 6227
Long bone
999.62.53
Measurements
δ13C and δ15N by IRMS
Lab no.
δ13C
δ15N
(‰)
(‰)
VPDB AIR
7749.3
–17.9
11.6
Reference
C:N
% marine
(±10)
M1 M2
3.5
34
26
Guiry et al. 2015
7741.3
–17.6
12.1
3.3
37
30
Guiry et al. 2015
7746.3
–17.3
12.0
3.3
40
34
Guiry et al. 2015
7751.3
–17.2
12.1
3.4
41
35
Guiry et al. 2015
7745.3
–17.2
11.7
3.3
41
35
Guiry et al. 2015
6885.3
–16.8
12.7
3.3
45
40
Guiry et al. 2015
7739.3
–16.8
12.7
3.3
45
40
Guiry et al. 2015
7750.3
–16.8
12.0
3.3
45
40
Guiry et al. 2015
7748.3
–16.8
11.9
3.4
45
40
Guiry et al. 2015
7742.3
–16.5
13.0
3.3
48
44
Guiry et al. 2015
6884.3
–16.5
12.6
3.4
48
44
Guiry et al. 2015
7747.3
–16.1
12.5
3.3
51
49
Guiry et al. 2015
16
15
14
13
δ15N (‰)
12
11
10
9
8
7
6
-22 -21 -20 -19 -18 -17 -16 -15 -14 -13 -12
δ13C (‰)
Figure 4.25. Arapouco, Sado valley: correlation between the δ13C (‰) and δ15N (‰) values
of 12 samples of human bone collagen.
237
Chapter 4
Archaeothanatology
Source material and analysis strategy
Archaeothanatological analysis was carried out on a total of 16 individuals (table
4.30), 50% of the MNI (32, Cunha and Umbelino 1995–97) for Arapouco. The
source material comprises field photographs available for 15 individuals, and paraffin blocks of four individuals. For the remaining individuals excavated at this site
(n=15) there is no source material available, at least at the time of this study.
The first phase of analysis took place at MNA, where the four individuals
preserved in paraffin blocks (ARA1961, Sks 5, 7; ARA1962, Sks 1A, 3A) were
analysed and photographed. The second phase of analysis consisted of the reassessment of my observations done in the museum, crosschecked with the photographed material and the archival field photographs where available (ARA1961,
Sks 5, 7). The analyses based on field photographs exclusively (ARA1961, Sks 1, 2,
3, 4, 6, 8, 9, 11, 12, 13, 14, 15; ARA1962, Sks 1A, 3A) followed the same analytical
protocol (see chapter 3, Archaeothanatology).
Results of the archaeothanatological analysis
The nature of the deposits
Analysis of the nature of the deposits at Arapouco indicates an unequivocal pattern
of primary burials (table 4.31). The cadavers were placed in the burial features
before the degradation of the labile joints, retaining their anatomical integrity. The
two cases that raise questions about this pattern come from the analysis of two of
the child skeletons analysed (ARA1961, Sks 10, 12). These questions are in part
due to the unclear documentation available for these individuals and to the normal
degree of fragmentation of juvenile skeletons which are not fully formed, resulting
in a greater degree of fragmentation and scattering of the bones in comparison
with an adult skeleton. In the case of ARA1961, Sk10 the documentation is also
unclear and further analysis is not possible. For ARA1961, Sk12 the graphic
documentation is somewhat better but of very low resolution. This 3 year old child
(ARA1961, Sk12) is buried with an adult female (ARA1961, Sk11) and while
ARA1961, Sk11 is clearly a primary deposit, the analysis of the child skeleton is not
conclusive. The right temporomandibular joint is preserved and supports the interpretation of a primary deposit, as this is one of the first articulations to break
down during the decomposition process. However, other labile articulations if
preserved are not observable in the documentation. The general topography of the
body is quite chaotic and the observations at this level are unclear. This scattering
could be the result of post-depositional disturbance of the feature; however, the
skeleton of the adult in the same feature remains in articulation invalidating this
hypothesis (see below, Deposits containing more than one individual).
238
n/a
n/a
n/a
n/a
n/a
n/a
yes
yes
10, 1961
11, 1961
12, 1961
13, 1961
14, 1961
15, 1961
1A, 1962
3A, 1962
yes
n/a
n/a
6, 1961
7, 1961
n/a
yes
yes
8, 1961
yes
yes
n/a
4, 1961
5, 1961
9, 1961
yes
n/a
3, 1961
yes
n/a
n/a
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
n/a
n/a
Field photographs
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
Drawings
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
Site plan
Graphic documentation w/human remains
2, 1961
Block
Source material
1, 1961
Individual
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
Written documentation
–
Fragmented material. Archaeothanatology is limited.
–
–
Fragmented material and unclear documentation.
Archaeothanatology is limited.
–
Very fragmented material and unclear documentation.
Archaeothanatology is not possible.
–
–
–
–
–
–
–
–
–
–
Observations
Table 4.30. Arapouco, Sado valley: source material used for the archaeothanatological analysis. This material is curated at MNA.
Primary
Primary
♀
♀
♂
Adult,
35–50 yrs
Adult,
35–50 yrs
Adult
Adult,
>40 yrs
Adult,
young
Adult,
35–50 yrs
Adult
3, 1961
6, 1961
10, 1961
Probably primary
n/d
n/d
n/d
Adult
Adult
1A, 1962
3A, 1962
15, 1961
Primary
Probably primary
Primary
Unclear, primary?
Probably primary
n/d
n/d
n/d
14, 1961
13, 1961
12, 1961
11, 1961
Primary
Primary
♀
♀
Primary
♂
Unknown
Primary
n/d
n/d
Primary
♀
Probably primary
Non-adult,
1.5–2.5 yrs
Adult,
20–35 yrs
Non-adult,
3 yrs±1 m.
Non-adult,
15 yrs±36 m.
Non-adult,
9 yrs±24 m.
Adult
9, 1961
8, 1961
7, 1961
5, 1961
4, 1961
Primary
n/d
Adult
2, 1961
Nature of the deposit
Adult
1, 1961
Primary
Sex
♂
Age
Individual
Maintenance of labile articulations: metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
The labile articulations are poorly preserved and unclear in the documentation.
However, the skeleton is well articulated and the general anatomical integrity of the body has been maintained.
The labile articulations are poorly preserved and unclear in the documentation.
However, the overall position of the bones indicates that the general anatomical integrity of the body has been maintained.
The preservation of the skeleton is poor.
However, the overall position of the bones indicates that the general anatomical integrity of the body has been maintained.
Maintenance of labile articulations: most metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: bones of right hand, most metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: TM joint is preserved but the documentation is unclear for other joints.
Maintenance of labile articulations: phalanges of left hand, metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Most labile articulations are unclear in the documentation except for the bones of the hands which are slightly disarticulated.
However, the skeleton is well articulated and the general anatomical integrity of the body has been maintained.
Maintenance of labile articulations: metatarsals and phalanges of left foot; other joints are not clear in the documentation.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of right foot; other joints are not clear in the documentation.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: bones of right hand, metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of left foot.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: bones of left hand, metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Poor preservation of diagnostic criteria and unclear documentation
Arguments
Table 4.31. Summary of the nature of the deposits with human remains at Arapouco, Sado valley.
Results: Sado valley, Arapouco
The space of decomposition of the cadaver
Analysis of the space of decomposition of the cadavers buried in Arapouco indicates a general pattern of filled spaces (table 4.32). The cadavers were covered with
sediment immediately after their placement in the features. The decomposition of
the bodies took place in this sediment-filled environment. Some cases at Arapouco
present very strong indicators not only of a filled space of decomposition but also
for very rapid and fluid penetration of sediment. This is demonstrated by the limited or moderate collapse of the thoracic cage (ARA1961, Sks 3, 4, 5, 6, 13). In
several cases the bones of the pelvic girdle are maintained in anatomical position
and the collapse of the bones is prevented by the immediate penetration of sediment (ARA1961, Sks 2, 4, 5, 7, 9, 13, ARA1962, Sk3A). Observations at the level
of the pubic symphysis are not always possible due to the nature of the documentation, but at least in one case (ARA1961, Sk8) the collapse of the os coxae occurs
laterally, which is a common movement even in filled spaces of decomposition.
Few cases suggest mixed spaces of decomposition (ARA1961, Sks 6, 12) with
original empty spaces in the grave feature and/or by the formation of secondary
empty spaces outside the volume of the cadaver within an overall filled environment. One of these cases is presented by ARA1961, Sk6. The analysis of this feature, discussed further below, suggests the existence of perishable material, functioning as a cushion, behind the upper body creating a secondary empty space in
this area while it decayed. In the second case, ARA1961, Sk12, the documentation
is more ambiguous. The analysis suggests the possibility of original empty spaces in
the feature and/or of the creation of secondary empty spaces by the decomposition of a perishable element in the feature, allowing the scattering of the bones of
this child. This perishable element could have been a container, such as a pouch or
a basket accommodating the child. The analysis is not conclusive and it should be
emphasized that children’s bones have a tendency to spread out more, both because they are smaller and because the articulations are less firm. Thus, some of
these scattering patterns could be possible even in a filled space. The burial of this
child (ARA1961, Sk12) will be discussed further throughout this text.
241
Sex
♂
n/d
♀
♀
♂
♀
n/d
♂
Age
Adult
Adult
Adult, 35–50 yrs
Adult, 35–50 yrs
Adult
Adult, >40 yrs
Adult, young
Adult, 35–50 yrs
Individual
1, 1961
2, 1961
3, 1961
4, 1961
5, 1961
6, 1961
7, 1961
8, 1961
Filled
Filled
Filled. Mixed?
Filled
Filled
Filled
Filled
Filled
Space of decomposition
continued...
Maintenance of bones in original unbalanced position: the right patella is suspended on the distal end of the right
femur; the bones of left foot are rotated. Also, the cranium is maintained in a semi-suspended position.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: moderate collapse of the TC and of os coxae. Also, the
cranium is maintained in a semi-suspended position.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: the CV can be followed as an arc forward; limited collapse
of the TC; the patellae are suspended on the distal end of the femora; the bones of left foot are slightly rotated
outwards. Also, the cranium is maintained in a semi-suspended position.
Movement at the level of the upper right limb suggests empty space behind this area.
Maintenance of bones in original unbalanced position: the CV can be followed as an arc forwards; the bones of the
right hand are slightly rotated inwards; limited collapse of the TC; limited collapse of os coxae; the right patella is
suspended on the distal end of the right femur. Also, the cranium is maintained in a semi-suspended position.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: limited collapse of TC; no collapse of left os coxae.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: limited collapse of TC.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: left os coxae.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: the bones of left hand are rotated and flexed at the wrist
and the bones of left foot are rotated outwards. Also, the cranium is maintained in a semi-suspended position.
No movement outside initial volume of the cadaver.
Arguments
Table 4.32. Summary of the space of decomposition of the human remains in the grave features at Arapouco, Sado valley.
n/d
n/d
Adult, 20–35 yrs
Non-adult, 3 yrs±1 m.
Non-adult, 15 yrs±36
m.
Non-adult, 9 yrs±24 m.
Adult
Adult
Adult
11, 1961
12, 1961
13, 1961
14, 1961
15, 196
1A, 1962
3A, 1962
n/d
n/d
n/d
n/d
♀
n/d
Non-adult, 1.5–2.5 yrs
10, 1961
♀
Adult
9, 1961
Table 4.32. continued
Filled
Filled
Filled
Filled
Filled
Unclear, mixed?
Filled
Unknown
Filled
Maintenance of bones in original unbalanced position: limited collapse of the os coxae.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: right patella is in a semi-suspended position.
No movement outside initial volume of the cadaver.
No movement outside initial volume of the cadaver.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: limited collapse of TC; limited collapse of the bones of os
coxae; the patellae are suspended on the distal end of the femora; the bones of right foot are rotated.
No movement outside initial volume of the cadaver.
Overall, the bones are found disarticulated, scattered and fragmented.
Movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: moderate collapse of TC; the right patella is suspended on
the distal end of the right femur; the bones of right foot are slightly rotated.
No movement outside initial volume of the cadaver.
Poor preservation and unclear documentation.
Maintenance of bones in original unbalanced position: the bones of left hand which are slightly rotated; no collapse
of os coxae.
No movement outside initial volume of the cadaver.
Chapter 4
The initial position of the cadaver in the feature
The most common position at Arapouco consisted of a cadaver lying on the back
(n=10) with the limbs in flexion (table 4.33). The second most common position
was the body lying on the lateral side (n=5) with the limbs in flexion and arranged
close to the body. It was not possible to reconstruct the initial position of two of
the children analysed at this site (ARA1961, Sks 10, 12).
In the case of the individuals placed on the back, the head of the cadaver was
often rotated forwards and downwards (n=8) towards the body in the feature, with
the neck in flexion or hyperflexion (ARA1961, Sks 1, 2, 5, 6, 7, 8, 13, 14). In only
two features (ARA1961, Sk15; ARA1962, 3A) was this not the case. In the case of
the individuals placed on the lateral side the head was rotated slightly upwards and
backwards (ARA1961, Sks 3, 4). The analysis was not possible in five cases
(ARA1961, Sks 9, 10, 11, 12; ARA1962, Sk1A).
The position of the upper limbs was variable but they were always placed close
to the body. In seven cases (ARA1961, Sks 1, 2, 6, 8, 11, 15; ARA1961, Sk1A) the
limbs were arranged symmetrically with both elbows in semi-flexion (n=2), flexion
(n=1), or hyperflexion (n=3). In three cases (ARA1961, Sks 5, 7; ARA1961, Sk3A)
they were placed in asymmetrical positions, with one of the elbows in semiextension and the other in extension (n=2) or flexion (n=1). For the other seven
individuals it was not possible to identify the symmetry of the upper limbs due to
the fragmentation of the bones and/or due to unclear documentation. The upper
limbs were frequently flexed at the level of the elbow joints, from slight or semiflexion to strong or hyperflexion. The extended position of at least one of the
limbs was noted in only four cases (ARA1961, Sks 5, 7, 9, 13).
At Arapouco, the bodies were placed in the feature with the lower limbs always
in flexion at the level of the hip and knee joints. The rotation of the limbs presents
the variation within this pattern. In the case of the bodies placed on the back
(n=10), three individuals presented a rotation towards the right side (ARA1961,
Sks 2, 7, 8) while three were rotated towards the left (ARA1961, Sks 1, 15;
ARA1962, Sk3A). In three cases the lower limbs were not rotated and the knees
were directed forwards (ARA1961, Sks 5, 6, 9). In only two cases were the lower
limbs in abduction (ARA1961, Sk6, 11). In the case of the bodies placed on the
lateral side the lower limbs were always in flexion or hyperflexion positioned close
to the upper body. The feet of the cadavers were commonly (n=9) positioned
towards the buttocks (ARA1961, Sks 2, 3, 4, 5, 7, 9, 11, 14; ARA1962 Sk3A)
indicating a constrained position of the lower limbs.
Often, the bodies lay in the grave feature in a clump (n=5), on the back or on
the lateral side, with the lower limbs in hyperflexion in front of the upper body
(ARA1961, Sks 2, 3, 4, 7, 14). This constrained position is particularly striking in
the case of skeletons 3 and 4 which became hypercontracted.
244
♂
n/d
♀
♀
♂
♀
n/d
Adult
Adult
Adult, 35–50 yrs
Adult, 35–50 yrs
Adult
Adult, >40 yrs
Adult, young
1, 1961
2, 1961
3, 1961
4, 1961
5, 1961
6, 1961
7, 1961
Sex
Age
Individual
Rotation–R and downwards
(neck flexed forwards)
Rotation–L and downwards
(neck flexed forwards)
Rotation–R and downwards
(neck flexed forwards)
Rotated slightly upwards and backwards
(neck in extension)
Rotated slightly upwards and backwards
(neck in extension)
Rotation–R and downwards
(neck flexed forwards)
Rotation–L and downwards
(neck slightly flexed forwards)
Head
Initial position of the cadaver in the feature
On the back
On the back
On the back
On the left
On the right
On the back
On the back
Trunk
In extension–R
In semi–flexion (90°)–L
In semi–flexion
In extension–R
In semi–flexion (90°)–L
In hyperflexion–R
n/a–L
n/a
In hyperflexion
In hyperflexion
Upper limbs
continued...
Hip and knee joints: flexion
Rotation–R
Feet towards buttocks
Clumping
Hip and knee joints: flexion
Abduction
Knees directed forwards
Hip joints: flexion
Knee joints: hyperflexion
Knees directed forwards
Feet towards buttocks
Hip and knee joints: hyperflexion
Rotation–L
Feet towards buttocks
Clumping
Hip and knee joints: hyperflexion
Rotation–R
Feet towards buttocks
Clumping
Hip joints: flexion
Knee joints: hyperflexion
Rotation–R
Feet towards buttocks
Clumping
Hip and knee joints: flexion
Rotation–L
Lower limbs
Table 4.33. Summary of the reconstruction of the initial position of the cadaver in the grave features at Arapouco, Sado valley. The initial position of the limbs is indicated
as –R or –L and refers to the right or left limb, when their positions are different. Rotation –R or –L indicates the rotation of the skeletal element(s) towards the right or the
left side.
♀
n/d
♀
n/d
n/d
n/d
n/d
Adult
Non-adult, 1.5–2.5 yrs
Adult, 20–35 yrs
Non-adult, 3 yrs ± 1 m.
Non-adult, 15 yrs ± 36 m.
Non-adult,9 yrs ± 24 m.
Adult
Adult
Adult
9, 1961
10, 1961
11, 1961
12, 1961
13, 1961
14, 1961
15, 1961
1A, 1962
3A, 1962
n/d
n/d
♂
Adult, 35–50 yrs
8, 1961
Table 4.33. continued
Rotation–L
n/d
Rotation–L
Rotation downwards (neck flexed forwards)
Rotation downwards (neck flexed forwards)
n/d
n/a
(head is not in the feature)
n/d
n/d
Rotation–R and downwards
(neck flexed forwards)
On the back
On the right
On the back
On the left
On the left
n/a
On the back
n/a
On the back
On the back
In flexion–R
In semi–flexion (90°)–L
In hyperflexion
In semi–flexion
One limb in flexion
In extension–R
n/a–L
n/a
In semi–flexion
n/a
In extension or semi–
extension–R
In extension–L
In flexion
Hip and knee joints: flexion
Rotation–R
Feet towards buttocks
Probably in flexion
Rotation–prob. R
Hip joints: flexion
Knee joints: hyperflexion
Rotation–L
Hip and knee joints: hyperflexion
Rotation–L
Feet towards buttocks
Clumping
Hip and knee joints: flexion
Rotation–L
n/a
Hip and knee joints: flexion
Abduction
Feet towards buttocks
n/a
Hip joints: flexion
Knee joints: hyperflexion
Knees directed forwards
Feet towards buttocks
Hip joints: flexion
Knee joint–R: hyperflexion
Knee joint–L: flexion
Rotation–R
Results: Sado valley, Arapouco
The grave features
The size and shape of the deposits with human remains at Arapouco were not
described by the excavators. The archaeothanatological analysis indicates that the
bodies were placed in pits dug into the ground. The lateral pressures and walleffects observed on the skeleton in the feature (table 4.34) indicate that in most
cases the shape and size of the pit was just sufficient to contain the body. However, in some cases the grave pits were somewhat wider allowing a relatively expanded position of the body particularly at the level of the lower body (ARA1961,
Sks 6, 8, 11). The analysis further indicates that in some cases the design of the
grave had a significant impact on the final arrangement of the cadaver in the
feature, visible by the clumping of some bodies (ARA1961, Sks 2, 7, 14), and by
the extremely constrained positions of some individuals (ARA1961, Sks 3, 4) indicating an increased hypercontraction of the skeleton during decomposition. In
these latter cases, the bodies were laid on the lateral side in individual small narrow
pits with the limbs in hyperflexion in front of the thoracic cage. The analysis shows
not only a strong bilateral pressure towards the body, but also evidence for important constraint effects. The wall-effect is visible in the alignment of the cranium
and the back on one side, and of the lower limbs on the other side. This pattern
was enhanced during decomposition by the progressive closure of the angles
between the limb segments causing the hypercontraction of the skeleton. These
wall-effects probably correspond to the physical limits of the pit, just wide enough
to contain the individual.
Frequently, the bottom of the feature was irregular and sloping from the upper
to the lower end as well as towards one of the sides of the grave. This is demonstrated by the flexion and the collapse forwards and downwards of the head, registered in eight to nine cases, three of which are very pronounced (ARA1961, Sks 5,
6, 13). This is also revealed by the flexion of the cervical vertebrae which can be
followed in an arc forwards (ARA1961, Sks 5, 6), and by the position of the crania
in anterior view lying in front of the chest (ARA1961, Sks 5, 6, 7, 8, 13) or towards
one of the shoulders (ARA1961, Sks 2, 14). This flexion pattern of the CV will be
discussed further below. Other observations suggesting a sloping floor are the
rotation forwards and downwards of the ribs (ARA1961, Sks 5, 6, 7, 11, 14;
ARA1962, Sks 1A, 3A), and the forward rotation of the sacrum (ARA1961, Sks 5,
8). The irregularity of the surface of the graves is also suggested by the general
distribution of the bones indicating a transfer of body weight with following patterns of movement towards one side of the feature (ARA1961, Sks 1, 2, 5, 7, 11,
14, 3A). Thus, the cadavers were often placed in a half-sitting position propped up
by the sloping characteristics of the floor of the grave.
Often, the narrow shape of the pit imposed a significant physical limit to the
cadaver offering lateral support and preventing the collapse of the bones. However, the analysis suggests that in a few cases there might have been extra support
offered by some kind of soft container such as light wrapping. This is possibly the
case for ARA1961, Sk11, where the pressures observed seem to indicate the
247
Chapter 4
existence of a soft wrapping at the time of the burial involving the upper body, as
suggested by the bilateral compression of the upper limbs. This lateral pressure
constraining and supporting the upper body is in contrast with the relatively wide
space available to accommodate the position of the lower limbs which lie in abduction. Furthermore, this soft wrapping at the level of the upper body (11) possibly
also offered support to the body of the 3 year old child (12) placed in front of the
left hemithorax of this adult, which will be discussed further below. The data available for ARA1961, Sk6 is more ambiguous and alternative explanations are possible. The alignment of the upper limbs contrasts with the position of the lower
limbs which are in abduction. In this case, I suggest that it is possible that some
kind of binding, such as that of a rope could have provided the lateral support
observed, while allowing the penetration of fluid sediment in the feature as well as
the flexibility of movement observed by the disarticulation of the upper limb upwards at the level of the shoulder. This is also the only case (ARA1961, Sk6) where
the analysis suggests the possibility of the presence of perishable structures behind
the body, as discussed further in this text.
Deposits containing more than one individual
Most deposits with human remains at Arapouco contain the remains of one individual.
One grave feature at Arapouco contains the remains of two individuals
(ARA1961, Sks 11, 12). This is a multiple double deposit of one female adult (11)
and one 3 year old child (12). The adult was laid in the pit and the child was placed
in front of her left hemithorax. This seems to be a synchronous deposit, first because individual 11 was still in anatomical integrity when the second individual (12)
was placed in the feature, and second because it seems as if the adult (11) was arranged to hold the child (12).
Another two individuals (ARA1961, Sks 9, 10), also a female adult and a child
(1.5 to 2.5 yrs), have been described briefly as a double burial (Cunha and Umbelino 1995–97, 2001; Umbelino and Cunha 2012). Based on the graphic documentation, I argue that it is not evident that these are simultaneous deposits. The bodies
are in very close proximity (particularly the heads) and do not disturb each other,
but a thick layer of sediment separates both bodies. The body of the child (10) lies
on a layer of sediment which covers the left shoulder girdle and left arm of the
adult (9). The bodies are not intertwined and do not come in contact with each
other and there is no evidence of a common arrangement of the bodies. ARA1961,
Sk9 was deposited first, covered with sediment, and only then was individual
ARA1961, Sk10 buried. The interval between these deposits is not possible to
determine; it could have been some minutes or several years. Based on the current
documentation, there is no evidence to support the interpretation of a simultaneous deposit, but this hypothesis cannot be completely rejected either.
248
Sex
♂
n/d
♀
♀
♂
♀
n/d
Age
Adult
Adult
Adult, 35–50 yrs
Adult, 35–50 yrs
Adult
Adult, >40 yrs
Adult, young
Individual
1, 1961
2, 1961
3, 1961
4, 1961
5, 1961
6, 1961
7, 1961
1. Weight of sediment from above;
2. Wall-effect, upper right side;
3. Prob. back pressure, upper right side;
4. Prob. wall-effect, upper left side.
Design of the grave feature and prob. light wrapping
and organic pocket behind upper body.
1. Wall-effect, right side;
2. Wall-effect, left side.
Design of the grave feature.
1. Wall-effect, right side;
2. Wall-effect, left side;
3. Overall constraint effect.
Design of the grave feature.
1. Wall-effect, right side;
2. Wall-effect, left side;
3. Overall constraint effect.
Design of the grave feature.
1. Weight of sediment from above;
2. Wall-effect, right side;
3. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, right side;
2. Wall-effect, upper left side;
3. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, upper right side;
2. Wall-effect, lower left side.
Design of the grave feature.
Pressures
continued...
1. Affecting knees and legs in particular, forcing feet towards buttocks;
2. Alignment of upper left limb, involving left shoulder and arm. Pattern enhanced by the sloping floor causing a transfer of body weight towards the left
side.
1. Extreme flexion of neck;
2. Alignment of upper right limb and projection forwards and upwards of right
shoulder and arm. Pattern enhanced by the sloping floor causing a transfer of
body weight towards the right side;
3. Alignment of lower limbs and constraint effect towards legs forcing feet
towards buttocks.
1. Extreme flexion of neck;
2. Alignment of upper right limb, involving arm and forearm, not affecting right
shoulder;
3. Affecting right arm in particular;
4. Suggested by the alignment of left arm.
Hypercontracted skeleton
1. Alignment of upper right limb and projection forwards and upwards of right
shoulder and arm;
2. Alignment of lower left limb, involving leg and foot. Pattern enhanced by the
sloping floor causing a transfer of body weight towards the left side.
1. Alignment of right limbs. Pattern enhanced by the sloping floor causing a
transfer of body weight towards the right side;
2. Alignment of left arm;
3. Alignment of lower limbs and constraint effect towards legs forcing feet
towards buttocks.
Hypercontracted skeleton
Arguments
Table 4.34. Summary of the key observations used in the reconstruction of the grave features at Arapouco, Sado valley. Observed pressures on the body are presented in a
number sequence. The arguments supporting these observations are indicated by the same number in the respective column.
♀
n/d
n/d
n/d
n/d
Adult
Non-adult, 1.5–2.5 yrs
Adult, 20–35 yrs
Non-adult, 3 yrs±1 m.
Non-adult, 15 yrs±36 m.
Non-adult, 9 yrs±24 m.
Adult
9, 1961
10, 1961
11, 1961
12, 1961
13, 1961
14, 1961
15, 1961
1A, 1962 Adult
3A, 1962 Adult
n/d
n/d
♀
n/d
♂
Adult, 35–50 yrs
8, 1961
Table 4.34. continued
1. Bilateral compression;
2. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, lower end;
2. Weight of sediment from above.
Design of the grave feature.
1. Wall-effect and bilateral pressure;
2. Wall-effect, lower left side.
Design of the grave feature.
1. Wall-effect, left side;
Design of the grave feature.
1. Wall-effect, upper end;
2. Wall-effect, right side;
3. Wall-effect, left side.
Design of the grave feature.
n/d
1. Bilateral compression, upper body;
2. Pressure from body weight towards right side.
Prob. soft wrapping involving upper body.
n/d.
1. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, lower end.
Design of the grave feature.
1. Towards shoulder girdles involving the arms as well. Pattern enhanced by the
sloping characteristics of the floor of the grave, slightly more elevated on the
lateral sides than on the centre of the feature.
2. Alignment of lower limbs and constraint effect towards lower limbs forcing
feet towards buttocks.
1. Alignment of right leg;
2. Visible by the even levelling and flattening of the bones in a horizontal plane.
1. Verticalization of clavicles. This lateral pressure affected shoulders but not the
upper limbs.
2. Alignment of lower left limb position in constrained position.
1. Alignment of lower limbs. Pattern enhanced by the sloping floor causing a
transfer of body weight towards the left side.
1. Extreme rotation forwards and downwards of head and shoulders;
2. Alignment of buttocks, hips and left foot;
3. Alignment of left shoulder and knees.
1. Alignment of the arms, which are rotated inwards and positioned close to TC.
Alignment of upper right limb, involving shoulder and arm. Suggested by the
alignment of left arm, not conclusive.
2. Movement of abdomen and pelvis towards right side of the feature, affecting
lumbar vertebrae, pelvic girdle and possibly lower limbs at the level of the hips.
This transfer of body weight and following pressure happened as the vertebral
column settled down on an irregular and sloping floor provoking a shift of axis
towards the right side.
–
–
1. Alignment of lower limbs and constraint effect towards legs forcing feet
towards buttocks.
1. Alignment and flexion of left foot.
Results: Sado valley, Arapouco
Post-depositional manipulations of the cadavers
At Arapouco, the skeletons are apparently complete and lie undisturbed in the
grave feature.
The only suggestion for the post-depositional manipulation of the cadavers
comes from the double grave of ARA1961, Sks 11 and 12. The skull of individual
11 (cranium and mandible) is missing from the feature. One possible hypothesis is
that the individual was originally buried complete and the skull was later removed.
The removal of the skull could be explained by (1) post-depositional surface disturbance or (2) by a deliberate act involving the re-opening of the grave and the
removal of the skull. In the case of a sloping grave floor such as that of this burial,
the head would be the most elevated body part in the feature; hence the closest
body part to the surface and easily removed without disturbing other bone elements. Another possible hypothesis is that the individual was buried without the
head. In this case, the head would have been removed soon after death, before an
advanced stage of decomposition, because the body was buried retaining its anatomical integrity. The viability of this hypothesis depends on whether the atlas (C1)
and axis (C2) were present or also missing from the feature, which could not be
confirmed at the time of this study. However, these scenarios are working
hypotheses and are not possible to test with the current documentation.
The mortuary programme at Arapouco
The burial practices observed at Arapouco present a strong similarity and the
minor detected variations do not affect the overall pattern of the treatment of the
dead.
While the tables presented above may be consulted for the diagnostic criteria
and arguments of analysis, I have chosen to describe ARA1961, Sk5 as a case that
clearly illustrates the common principles governing burial practice at Arapouco. In
this section I describe also ARA1961, Sks 6, 11, 12 and 13 as cases that present
unusual features within the assemblage of Arapouco with several striking elements
from an archaeothanatological perspective.
ARA1961, Sk5
ARA1961, Sk5 consisting of the remains of an adult, possibly male, lying on its
back in the grave feature, is particularly well-preserved in a paraffin block
(fig. 4.26).
This case combines several common elements observed in the human burials at
Arapouco. The first common element is the primary nature of the deposit. In the
case of this skeleton, the maintenance of the bones of the right hand in anatomical
connection, as well as the maintenance of the articulations of the metatarsals and
phalanges of the feet in perfect anatomical position are strong arguments supporting this interpretation. The general topography of the body is also indicative of a
burial in primary position. The few disarticulations observed at the level of the
251
Chapter 4
cervical vertebrae or at the bones of the left hand, are indicative of minor movements into empty spaces created during the process of decomposition.
The second common element is the decomposition of the body in an environment filled with sediment. This is illustrated by several groups of bones which were
maintained in place even though their original position was quite unbalanced. This
is the case for the connection maintained between the cervical vertebrae (CV)
which can be followed as an arc forwards and downwards (fig. 4.27). A similar
movement can be observed in the feature ARA1961, Sk6 between the C3–C7,
while the atlas (C1) and axis (C2) became detached (fig. 4.28). This forward flexion
of the neck probably also occurs in other features where there is rotation of the
head forwards and downwards but the documentation is unclear and in many cases
the CV are not visible or preserved. In the case of ARA1961, Sk5, only some CV
are exposed. These are flexed forwards with a tendency towards the right side and
form an arc behind the cranium. The cervical joints that are visible are not articulated but are in contact. The atlas (C1) is not visible and the axis (C2) is only
partially visible on its posterior side below the cranium. Thus, the persistent joint
occipital–atlas (C1) is not exposed. C3 presents its posterior view while C4 presents its superior and posterior view. C5 is only partially visible and its lateralization is not clear. The segment C6–C7 is not exposed. The position of the CV
shows that one portion, C2–C3, has tipped with the cranium and C1 possibly
moved below the cranium, while the other portion, C4–C5, has stayed behind the
collapse. The close contact between the CV has been maintained by the immediate
penetration of fine sediment throughout the process of decomposition, otherwise
the bones would not display this gradual support. Also, the bones of the right hand
were rotated slightly inwards but retained a perfect articulation. Another strong
argument is the position of the right patella suspended on the distal end of the
right femur. These are all indicators of decomposition in a filled space where these
bones could be supported, even if in unstable positions.
Other evident movements in this feature can be explained by the dynamics of
the decomposition process. The cranium has collapsed slightly forwards, but was
maintained in a semi-suspended, elevated position within the feature. Likewise, the
dispersal of the bones of the left hand took place in the volume of the initial cadaver and cannot be attributed to empty spaces outside the cadaver. ARA1961,
Sk5, like most burials at Arapouco, presents very strong indicators not only of a
filled space of decomposition but for a very rapid and fluid penetration of sediment. The collapse of the thoracic cage even in filled volumes is in this case moderate and the position of the ribs is kept in order. This is also the case for the collapse of the os coxae, which in most of the burials is very limited. In the particular
case of ARA1961, Sk5 the elevation of the bones of the left forearm by a pocket of
sediment is an interesting example of the formation of a filled volume by the interplay of an active area during decomposition where very fluid sediment filled up all
tiny empty spaces created during a dynamic stage of decay. Overall, the limited
movement of the bones clearly indicates that the space of decomposition was filled
and that the sediment penetrated immediately.
252
Results: Sado valley, Arapouco
Figure 4.26. ARA1961, Sk5 illustrates several common principles governing burial practice
at Arapouco, Sado valley. This skeleton in preserved in two paraffin blocks at MNA. (Photograph by J. P. Ruas)
253
Chapter 4
Figure 4.27. ARA1961, Sk5 presents typical patterns of decomposition of the body in a
sediment-filled environment. Detail of the CV which can be followed as an arc forwards
and downwards indicating the forward flexion of the neck and the gradual support of the
skeletal elements by the rapid penetration of fluid sediment. (Photographs by author at
MNA)
254
Results: Sado valley, Arapouco
Figure 4.28. The CV of ARA1961, Sk6 present a similar movement pattern as observed in
ARA1961, Sk5. The CV are flexed forwards and downwards, indicating the forward flexion
of the neck and the gradual support of the skeletal elements by the rapid penetration of
sediment. (MNA, APMH, F367)
The initial position of the cadaver ARA1961, Sk5 presents various common
elements observed in the human burials at Arapouco. The cadaver was laid on its
back in a small pit with a sloping floor, possibly slightly more elevated at the superior left side of the feature. The head was directed forwards and downwards towards the chest, slightly rotated to the right with the neck flexed forwards. This
pattern was accentuated during decomposition, first because the body was lying on
a slope and second, because the cranium has a tendency to roll after the joints of
the neck break down. Also, the head is the most elevated body part in the feature
hence the closest body part to the surface of the grave, thus the pressure of the
sediment on the back of the head would increase the flexion of the neck, which is
a common observation for the burials at Arapouco. The right upper limb of
ARA1961, Sk5 was placed in complete extension while the left upper limb was
semi-flexed at the elbow with the left hand placed in front of the abdomen. The
lower limbs were flexed at the level of the hips and hyperflexed at the knees with a
tendency towards the left side.
At Arapouco, the main pressures observed on the bodies in the grave features
consist of the weight of the sediment from above enhancing the flexion of the
neck and the rotation forwards and downwards of the head, and the wall-effects at
the lateral sides and lower end of the graves with consequent pressures providing
support as well as constraining the movement of the limbs. This is well illustrated
by ARA1961, Sk5 where there are three main pressures. One is the movement
255
Chapter 4
pattern and consequent pressure from the weight of the sediment from above
visible by the extreme flexion of the neck. The second is the lateral pressure and a
wall-effect on the right side of the feature, visible in the alignment of the upper
right limb, involving the right shoulder girdle and arm. This lateral pressure was
accentuated by the sloping grave floor, slightly more elevated on the superior left
side, causing a transfer of body weight towards the right side. Hence, there is
movement pattern and consequent pressure from body weight from the upper left
side against the right wall of the grave. The third main pressure observed is a walleffect at the lower end of the feature visible in the alignment of the lower limbs
and the constraint effect towards the legs forcing the feet towards the buttocks.
The pressures observed are the result of the design of the grave feature.
ARA1961, Sk6
Likewise, the grave feature ARA1961, Sk6 presents all the general elements of the
burials known at Arapouco. However, from an archaeothanatological perspective,
this case offers an unusual example of localised movement of the bones at the level
of the upper right limb which is in contrast with the overall pattern of limited
movement of the bones in the grave features, suggesting in this case the formation
of a secondary empty space (fig. 4.29).
As with the typical burials at Arapouco, ARA1961, Sk6 was laid in a pit with a
sloping floor, on its back, with the head flexed forwards and downwards, and
slightly rotated to the left. The upper right limb was semi-flexed at the elbow and
placed laterally to the upper body with the right hand lying on the right hip. The
upper left limb was semi-flexed at the elbow with the forearm lying in front of the
lower abdomen. The lower limbs were flexed at the level of the hips and knees, in
slight abduction, while the feet were lying medially in close proximity.
The primary nature of this deposition is demonstrated by the anatomical preservation of the labile joints of the left foot. Also, the few disarticulations observed,
at the level of the cervical vertebrae (C1–C2, C2–C3), at the bones of the right
hand, or at the bones of the left forearm, are the result of movements into empty
spaces created during the process of decomposition. Overall, the skeleton is well
articulated and the general topography of the body confirms its primary position.
The extreme disarticulation of the upper right limb at the level of the shoulder
girdle indicates an important movement of the bones but it is not indicative of a
secondary deposition.
The secondary empty space can be observed at the level of the upper right limb.
The limb was pushed upwards and forwards becoming disarticulated at the level of
the shoulder while retaining its articulations at the elbow. This radical movement
upwards, in an overall filled environment such as this, could be explained by the
existence of an organic element behind the body in this area creating a secondary
empty space after its decomposition to which the limb could migrate.
256
Results: Sado valley, Arapouco
Figure 4.29. ARA1961, Sk6. The striking displacement upwards of the upper right limb
contrasts with the minor movements observed in this feature. The disarticulation of the
limb occurs at the level of the shoulder joint and did not affect the articulations at the
elbow or at the wrist which remain in articulation. While the bones of the limb are displaced upwards the scapula migrated towards the bottom of the feature. (MNA, APMH,
F368, F367)
257
Chapter 4
The analysis indicates four main pressures on the body in the feature. One is the
movement pattern following pressure from the weight of the sediment from
above, visible by the extreme flexion of the neck. The second is the lateral pressure
and a wall-effect on the right side of the feature, visible in the alignment of the
upper right limb, involving the arm and forearm, but not affecting the right shoulder girdle. At this level, there is a pressure from the back affecting the arm in particular. The fourth pressure is a wall-effect at the left side of the feature suggested
by the alignment of the left arm; however, this observation is not conclusive
because the visibility is reduced at this level.
The striking displacement upwards of the upper right limb is in contrast to the
minor movements observed in this feature. The disarticulation of the limb occurs
at the level of the shoulder joint and did not affect the articulations at the elbow or
at the wrist which remain in articulation. While the bones of the limb are displaced
upwards the scapula moved towards the bottom of the feature. Oddly, neither the
movement upwards nor the wall-effect seems to have affected the shoulder girdle.
The bones of the right shoulder girdle do not show the effects of a lateral pressure.
The right clavicle is slightly displaced downwards but this is consistent with the
general tendency of the bones to move downwards, supporting the interpretation
of a sloping grave floor.
These movements could be explained by a combination of elements in the feature. One is the presence of a perishable element behind the body in this area, such
as cushioning or padding, creating an empty space during its decay into which
these bones could fall. Additionally, the existence of a soft wrapping may be suggested that could support the limb, at the same time allowing the penetration of
sediment. One hypothesis is that the upper body was bound by some kind of stout
cord, such as rope. Support from a rope would have provided the lateral or bilateral pressure visible in the alignment of the upper right limb, affecting the arm and
forearm and possibly the left arm. If the rope had been placed at the level of the
arm it would allow the exclusion of the right shoulder girdle from the set of
movements affecting the limb. If the body had been wrapped in a more encompassing material, such as a shroud, this would also affect the shoulder girdle while
binding by a rope/cord would permit greater flexibility of movement. Furthermore, binding by a rope would have allowed the immediate penetration of sediment as attested in this feature. Thus, as the body collapsed forwards and downwards towards the sloping bottom, the right limb moved upwards in the reverse
direction, becoming disarticulated at the level of the shoulder joint while maintaining its alignment by support from the rope. Additionally, while the perishable
structure behind the body decayed, the scapula descended, but only slightly, as the
fine sediment penetrated providing gradual support and preventing further collapse.
These elements – cushioning/padding and wrapping – have possibly affected
the left upper limb as well. Unfortunately the documentation is unclear and further
observations are not possible. These supports are suggested by the observations at
258
Results: Sado valley, Arapouco
the level of the upper body and there are no indicators of wrapping or padding at
the level of the lower limbs.
Other cases, such as ARA1961, Sks 7 and 11 (figs. 4.30, 4.31), present a similar
disarticulation at the level of the right shoulder girdle where the upper right limb is
projected forwards and upwards. However, in these cases the movement upwards
and forwards affects not only the limb but the shoulder girdle as well. Also, in
neither of these cases is the disarticulation as accentuated as in the case of Sk6.
Figure 4.30. ARA1961, Sk7 presents a displacement upwards of the right shoulder girdle
similar to the movement pattern observed in ARA1961, Sk6. (Left: MNA, APMH, F370.
Right: photograph by J. P. Ruas)
ARA1961, Sks 11 and 12
At Arapouco, the cadavers are typically placed individually in the grave feature.
The deposit containing the remains of an adult female (ARA1961, Sk11) and a 3
year old child (ARA1961, Sk12) is the only clear case of a double burial at the site
(fig. 4.31). Furthermore, from an archaeothanatological perspective this burial
presents several striking elements. One is the apparent bilateral compression of the
upper body which is in contrast with the wide space available to accommodate the
lower body. The other singular feature of this burial is the possible placement of
the child in a container of sorts arranged in front of the adult. As discussed, another unusual element of this burial is the missing skull of the adult as well as the
position of the lower limbs rotated outwards in abduction. Despite these deviant
259
Chapter 4
elements, the feature presents several of the common elements observed in the
burials at Arapouco.
Individual 11 is clearly a primary deposit. The skeleton is well articulated and
the decomposition of the body took place in this feature. This is well demonstrated
by the preservation of many labile articulations such as those of the bones of the
right hand and of most bones of the feet. The few disarticulations observed at the
level of the right elbow joint and some of the bones of the right hand are indicative
of movements into empty spaces created during the process of decomposition, and
are not indicative of a secondary deposition. The general topography of the body
also indicates a burial in primary position. Individual 12 is possibly a primary deposit, however, as previously discussed, this interpretation is not conclusive both
because the documentation is unclear and because the skeleton is very fragmented.
Figure 4.31. ARA1961, Sks 11 (adult female) and 12 (child, c 3 yrs) were buried simultaneously. The remains of the child were found concentrated in front of the left hemithorax of
the adult, but also disarticulated and scattered in the feature. (MNA, APHM, F414)
260
Results: Sado valley, Arapouco
Individual 11 is well articulated and the decomposition of the body took place in a
sediment-filled environment. Individual 12 presents a more ambiguous situation.
As discussed, although the right labile temporomandibular joint is in articulation,
the other few identifiable bones are disarticulated, scattered, and fragmented. One
exception is a tibia and a fibula, found articulated, but in isolation at the lower end
of the feature. As mentioned, the skeletal remains of children are more prone to
scattering due to the smaller size and fragility of the joints. Also, its position in
front of the adult’s upper body would have influenced the movement of its bones,
because the thoracic region and abdominal cavity are rich in soft tissue and very
active areas during decomposition. However, the wide scattering of individual 12
requires further explanation and may suggest a mixed space of decomposition for
the child. The different spaces of decomposition could be explained if the adult
(11) had been placed directly in the pit while the child (12) had been lain in some
kind of container and placed in front of the adult, after which, the feature would
have filled with sediment. This container could have been of soft or hard material
providing enough space for the movement of the bones while the articulations
broke down.
As with most burials at Arapouco, individual 11 was laid in a pit with a sloping
floor, possibly slightly more elevated on the upper right side of the feature and
lower at the region of the pelvis. The body was positioned on its back. As discussed, the skull of individual 11 (cranium and the mandible) is not in the feature
and the original position of the head is unknown.
The upper limbs were slightly flexed at the level of the elbows and the hands
were placed on the pelvic region. However, the position of the lower limbs is
unusual. They present the typical flexion at the hips and knees, but are rotated in
abduction, suggesting an atypical wide pit, at least at this level, while the feet are
crossed and placed medially.
There are two main pressures on the adult in the feature. One is a bilateral pressure visible through the alignment of the arms, which are rotated inwards lying
very close to the thoracic cage. On the right side of the feature this wall-effect is
visible in the alignment of the upper right limb, involving the shoulder girdle and
arm. On the left side this wall-effect is suggested by the alignment of the left arm,
although this observation is not conclusive because the visibility is reduced at this
level. The second is the pressure from the body weight towards the lower and right
side, visible by the movement of the abdomen and the pelvis to the right side of
the feature, affecting the lumbar vertebrae, pelvic girdle and possibly the lower
limbs as well, at the level of the hips. This transfer of body weight and consequent
pressure occurs as the vertebral column settles down on an irregular and sloping
grave floor, provoking a shift of axis towards the right side. The pressures
observed suggest the existence of a soft wrapping at the time of the burial
surrounding the upper body, as indicated by the bilateral compression of the upper
limbs. This constraining element is in contrast with the wide space available to
accommodate the lower limbs.
261
Chapter 4
Figure 4.32. Reconstruction hypothesis of the initial position of individuals 11 and 12 in the
grave feature, Arapouco, Sado valley. The analysis suggests that the child (12) was arranged
in some kind of soft wrapping such as a pouch or even a semi-hard container, such as a
basket, placed in front of the adult female (11) and her arms were wrapped around it.
(Drawing by Susanna Berglund)
262
Results: Sado valley, Arapouco
The initial position of individual 12 in the feature is unknown. Its remains were
found concentrated in front of the left hemithorax of individual 11. A disarticulated femur and some unidentifiable fragments are found laterally to the left femur
of individual 11. An articulated tibia and fibula are found at the lower right end of
the feature together with foot bones, however, it is unclear if these belong to this
child or are in fact remains from another burial.
It seems likely that the bilateral pressure observed at the level of the upper body of
the adult female could be related to the arrangement of the 3 year old child in the
feature. If this was the case, it could be suggested that the child was arranged in
some kind of soft wrapping such as a pouch or even a semi-hard container, such as
a basket, placed in front of the adult and with her arms wrapped around it
(fig. 4.32).
ARA1961, Sk13
At Arapouco, the choice of the initial position of the cadaver in the grave feature is
consistent. One burial, ARA1961, Sk13, presents a few unusual elements, such as
the rotation of the head towards the knees and the placement of at least one of the
upper limbs between the thighs. Nevertheless, the burial shares the basic common
elements observed in all the burials at Arapouco (fig. 4.33).
The burial is a primary deposit as demonstrated by the maintenance of the
bones of the feet in perfect articulation and by the general topography of the body.
Also, like most burials at Arapouco, the decomposition of the body took place in
filled space. This is well illustrated by the limited collapse of the bones of the pelvic girdle. The normal displacement of these bones was somewhat prevented by a
rapid and fluid penetration of sediment. Other strong arguments supporting the
decomposition in a sediment-filled environment are the position of both patellae in
anatomical connection at the knee joints, as well as the bones of the right foot
which were rotated but retained a perfect articulation.
The body was laid in a small pit, possibly slightly more elevated on the left side
of the feature. The individual was laid on its left side, with the head directed forwards and downwards towards the lower limbs, probably with the neck in hyperflexion similar to ARA1961, Sks 5, 6. The CV are not visible in the documentation,
but the relative position of the cranium suggests that the cranium collapsed forwards towards the lower limbs. This pattern was possibly enhanced during decomposition, firstly because the body was in a very constrained position, particularly at
the upper end of the feature, and secondly, because the cranium has a tendency to
roll after the joints of the neck break. Also, the pressure of the sediment on the
back of the head would have an impact as well. The upper right limb was placed in
complete extension with the forearm lying between the thighs. The upper left limb
is not visible in the documentation. The buttocks and hips seem to lie on the lowest part of the feature. The lower limbs were placed in front of the upper body,
flexed in adduction at the level of the hips and knees. The feet lying at the lower
263
Chapter 4
end of the feature, were crossed and aligned with the buttocks, with the left foot
placed over the right foot.
Three main pressures bore down on the body in the feature. One is the walleffect from the upper end of the pit maintaining the extreme rotation of the head
and shoulders forwards and downwards. The second is a wall-effect on the right
and lower end of the feature visible in the alignment of the buttocks, hips and left
foot. The third pressure was a wall-effect on the left side suggested by the alignment of the left shoulder and knees. The pressures observed are likely to be the
result of the design of the grave feature.
Figure 4.33. ARA1961, Sk13 shares the basic common characteristics observed in all burials
at Arapouco, but was placed in an unusual position. The body lay on the left side with the
head rotated towards the knees, and at least one of the upper limbs was placed between the
thighs. (MNA, APMH, F376)
264
Results: Sado valley, Arapouco
***
The mortuary programme at Arapouco was consistent and homogeneous, as suggested by the archaeothanatological analysis.
The cadavers were typically placed in individual grave features while retaining
their anatomical integrity (individual primary deposits). The feature was then
immediately covered with sediment, providing a filled space of decomposition. In
several cases the analysis indicates a rapid and fluid penetration of sediment.
The cadavers were commonly laid on the back, or alternatively on the lateral
side, always with the limbs in flexion usually arranged close to the body. Typically,
the bodies were placed in flexed positions becoming contracted in the small grave
features. The floor of the graves tended to be uneven and sloping, and a halfsitting position of the body was common. In these cases, the head of the cadaver
was often rotated forwards and downwards towards the body (fig. 4.34).
In most cases, the observed pressures exerted on bodies in the features can be
explained as the result of the design of the grave. Normally, these wall-effects and
consequent pressures correspond to the physical limits of the pit, just wide enough
to contain the individual.
Few cases present deviant elements, but there were exceptional instances of features with, for example, more than one individual, possible perishable elements
inserted behind the bodies, and possible wrappings and/or containers. Despite the
striking and unusual elements, all these cases followed the common core characteristics of burial practices observed at Arapouco.
265
Chapter 4
Figure 4.34. This illustration presents several key elements observed in the burials at
Arapouco. The cadavers were commonly laid on the back, or alternatively on the lateral
side, always with the limbs in flexion usually arranged close to the body. The floor of the
graves tended to be uneven and sloping, and a half-sitting position of the body was common. In these cases, the head of the cadaver was often rotated forwards and downwards
towards the body. In most cases, the observed pressures exerted on bodies in the features
can be explained as the result of the design of the grave. Normally, these wall-effects and
consequent pressures correspond to the physical limits of the pit, just wide enough to contain the individual. (Drawing by Susanna Berglund)
266
Sado valley: Cabeço das Amoreiras
Site background: a short history of research
Cabeço das Amoreiras (also known as S. Romão) is a shell midden site located on
the eastern part of the Sado valley, and is situated on the left margin of the river at
52 m.a.s.l. (Diniz et al. 2012) (fig. 2.3).
Archaeological material and documentation available for Cabeço das Amoreiras
are curated at the National Museum of Archaeology in Lisbon (MNA). The first
field season started on the 26th of August of 1958 as indicated by the excavation
diary written by J. Roldão (MNA, APMH, 2/3/7/2, p. 4–7)2. This document
reports the first five days of excavation at Cabeço das Amoreiras. During these five
days no human remains were found but the graphic documentation indicates that
all human burials (n=8) were recovered in 1958. The burials were excavated sometime after these first five days but those pages are missing from the archive. The
site was excavated again in 1985–86 (Arnaud 1986) but no human remains were
found.
Graphic documentation in the archives of MNA consists of two site plans with
the human burials in context (MNA, 1958, D. Sousa, A115, A116) (figs. 4.35,
4.36), twelve drawings of osteological material of human remains and fauna (MNA,
1958, D. Sousa, B45–56), and field sketches (human remains, numbered I–VIII;
fauna, numbered a–c). All drawings are from 1958 by Dario de Sousa, MNA, and
include a few notes on stratigraphy and orientation of the features.
Archives of photography of Cabeço das Amoreiras consist of 49 photographs
from 1958 (MNA, APMH, 2/11/4) of variable quality. The archive contains photographs of the eight individuals excavated at the site (MNA, APMH).
One human skeleton is preserved in a block of paraffin wax (CAMS1958, Sk4),
and the remaining skeletons are in disarticulated state sorted in containers. The
human remains were studied by a team of anthropologists and the MNI curated at
the museum is six (Cunha and Umbelino 1995–97, 2001; Umbelino and Cunha
2012). Graphic documentation indicates the excavation of eight individuals, however, one of them is represented solely by its cranium (CAMS1958, Sk2). The
burial population studied by E. Cunha and C. Umbelino (1995–97, 2001; Umbelino and Cunha 2012) consists of five adult males, one non-adult, and one indeterminable. This analysis suggested that four of the adult males were of mature age
(>35 yrs) (table 4.35).
This report is in a bundle with other Sado site reports under the archive title Report of Cabeço do
Pez 1958 (Relatório do Cabeço do Pez 1958) with no reference to Cabeço das Amoreiras.
2
267
Chapter 4
Figure 4.35. Cabeço das Amoreiras, Sado valley: site plan and profiles of the main excavation area (Talhão T) where all human remains were found (I to VIII). (Adapted from a
photograph by J. P. Ruas. MNA, 1958, D. Sousa, A115)
268
Results: Sado valley, Cabeço das Amoreiras
Figure 4.36. Cabeço das Amoreiras, Sado valley: site plan of the excavated areas. Bold dots:
area of the shell midden. Test trenches are indicated from A to I. Talhão T: main excavation
area where all human remains were found. (Adapted from a photograph by J. P. Ruas.
MNA, 1958, D. Sousa, A116)
A note on the back of the photograph of CAMS1958, Sk6 is in disagreement with
the anthropological analysis. The excavator describes the human remains as a small
group of bones of a child (pequeno conjunto de ossos de criança) while the osteological
analysis indicates that these are the remains of a mature adult. This inconsistency
could be the result of mislabelling of the skeletons. The graphic documentation is
unclear, but the misinterpretation in the field is possibly due to the fragmentary
state of the material.
The stratigraphic data is provided by the graphic documentation (table 4.36).
The top layer is described as brown soil (terra castanha), the second layer as black
soil (terra preta), the third layer is generally thicker and is described as grey soil and
269
Chapter 4
shells (terra cinzenta e conchas), and the fourth layer is described as chestnut brown
sand. The human remains were excavated in the chestnut brown sand, and some of
them were presumably lying directly on the soft bedrock (piçarreira, chão natural).
Table 4.35. Cabeço das Amoreiras, Sado valley: sex and age estimates.
Ind.
Container Age
Sex
Notes
Reference
1
–
n/d
n/d
Graphic documentation
shows concentration of
bone fragments of unidentifiable material.
This study
2
–
Adult?
n/d
Graphic documentation
shows one cranium.
This study
3
6253(?)
Adult, mature ♂
–
Cunha and Umbelino 1995–97,
Umbelino 1999
4
999.68.1 Adult
Preserved in paraffin
block.
This study
5
6239
Adult, mature ♂
–
Cunha and Umbelino 1995–97;
Umbelino 1999
6
6237
Adult, mature ♂
–
Cunha and Umbelino 1995–97;
Umbelino 1999
7
6238
Adult, mature ♂
–
Cunha and Umbelino 1995–97;
Umbelino 1999
8
6233
Non-adult,
7.5–16 yrs
–
Cunha and Umbelino 1995–97;
Umbelino 1999
♂?
n/d
Table 4.36. Cabeço das Amoreiras, Sado valley: stratigraphy and depth of the human burials in relation
to top layer. (Data from field drawings, MNA, 1958, D. Sousa)
Ind.
Layer
Depth (cm)
1
4, chestnut brown sand
n/a
6
4, chestnut brown sand
n/a
7
4, chestnut brown sand
120, cranium; 140, post-cranial skeleton
8
4, chestnut brown sand
125
4
4, chestnut brown sand
130
2
4, chestnut brown sand
140
3
4, chestnut brown sand
140
5
4, chestnut brown sand
150
270
Results: Sado valley, Cabeço das Amoreiras
Chronology of the burial activity
Previous 14C measurements
The first 14C measurements of material from Cabeço das Amoreiras were carried
out on material excavated in 1985–86: one sample on shells (Q–AM85B2b) and
one on charcoal (Q–AM85B2a) (Arnaud 2000), both collected from under a compact layer of shells containing a few fragments of cardium pottery (Arnaud 2000,
29; see Diniz 2010) (table A10).
The first measurements on human bone from Cabeço das Amoreiras (Beta–
125110) were carried out in the late 1990s (Cunha and Umbelino 2001), collected
from one individual (5) excavated in 1958.
Objectives of the dating programme
One human skeleton only (CAMS1958, Sk5) was directly dated before this radiocarbon dating programme. This date (Beta–125110) was the earliest 14C date available for the site, many centuries older than the other two dates available on nonhuman samples (Q–AM85B2a and Q–AM85B2b), but it was coeval with the only
human burial dated at Arapouco (Sac–1560, ARA1962, Sk2A).
The burial area of this site was small and had relatively few burials. It was expected that the human burials would be coeval in time.
The general aims of the dating programme for Cabeço das Amoreiras followed
the macro-goals of the radiocarbon programme of this dissertation, outlined in
chapter 3, and were:
• to estimate the duration of the use of this site for burial activity;
• to estimate when the burial activity started and when it ended;
• to estimate the frequency of the burial activity in the site.
The site-specific aims of the dating programme for Cabeço das Amoreiras were:
• to confirm the apparent antiquity of the burial activity at this site indicated by a previous measurement on human bone (Beta–125110) contrasting with more recent dates on charcoal and shells;
• to clarify if this site was used as burial ground during both the Late
Mesolithic and the Early Neolithic time frame.
Sampling strategy
Four samples of human bone were selected for 14C analysis (table 4.37). It was
previously reported that CAMS1958, Sk6 did not yield enough collagen for isotopic analysis (Fontanals-Coll et al. 2014), and it was excluded from pre-selection.
The sampling strategy was targeted towards a good spatial and stratigraphic coverage and it was possible to obtain samples from all pre-selected individuals.
271
Chapter 4
Results
Two of the four samples submitted as part of this study revealed poor quality during laboratory analysis (Ua–46935 and Ua–46936) and were rejected (table A1).
Three 14C measurements on human bone collagen samples from Cabeço das
Amoreiras are available now (table 4.38). The three remaining undated individuals
(individuals 3, 6 and 8) revealed poor quality for 14C and isotopic measurements.
The remains of individuals 1 and 2 were not identified at MNA.
Table 4.37. Samples of human bone (4) from Cabeço das Amoreiras collected and submitted for 14C
measurement in the course of this study. All samples are from the 1958 assemblage and were collected at
MNA. *Samples pre-treated with boiling in acetone, ultrasonification first with ether and then with ethanol, and finally filtrated; after that processed as usual (see chapter 3).
Ind. Lab no.
Age
Sex
Bone
Stratigraphy
3
Ua–46935
Adult,
mature
♂
Humerus–L Layer 4,
999.43.8
140 cm
Previous stable isotope analysis
(Fontanals-Coll et al. 2014;
Umbelino 2006)
Ua–46935: doubtful quality
4
Ua–47973*
Adult
♂?
Fibula–L
Layer 4,
130 cm
Paraffin block
7
Ua–47974*
Adult,
mature
♂
Fibula
999.46.4
Layer 4,
Previous stable isotope analysis
120–140 cm (Fontanals-Coll et al. 2014)
8
Ua–46936
Non-adult,
7.5–16 yrs
n/d Tibia
999.47.1
Layer 4,
125 cm
Observations
Not measurable
Table 4.38. Human bone collagen samples (3) from Cabeço das Amoreiras, Sado valley. 14C measurements and estimation of non-terrestrial carbon intake (% marine). Samples are sorted so that the lowest
marine percentage is on the top of the table. Marine percentage was estimated from the measured δ13C value
of each sample by the application of method marine 1 and method marine 2, based on adopted endpoint
values of –12‰ and –20‰ for marine (100%) and terrestrial (100%) diets respectively.
Identification of the
human remains
Ind. Excav.
Museum
4
1958
MNA
7
1958
MNA
5
1958
MNA
% marine Reference
Measurements
14C by AMS; δ13C and δ15N by IRMS
(±10)
14C Age BP δ13C (‰)
Lab no.
δ15N (‰) C:N M1 M2
AIR
VPDB
Ua–47973 6484±39
–20.0
7.6
3.3 14
0
this study
Ua–47974 6645±42
–20.0
8.2
3.3
Beta–
125110
–19.0#
9.5#
2.9# 24
7230±40*
14
0
this study
13
*Cunha and Umbelino 2001;
#Fontanals-Coll et al. 2014
Calibration and chronological models
The chronological model was built on three 14C dates, two of which are new dates
obtained in the scope of this study. The sample Beta–125110 was calibrated with
the recently published IRMS–based δ13C values because the originally δ13C value,
published along with the 14C measurements, was an AMS-based value which
should not be used for calibration (see chapter 3, Reliability of the isotopic
measurements).
272
Results: Sado valley, Cabeço das Amoreiras
This is a phase model which groups all dated burial events in a phase of burial
activity (Bronk Ramsey 1998, 2000, 2001). This grouping is constrained by a start of
burial activity event, and an end of burial activity event (i.e., boundaries) imposing no
internal constraints to the relative order of the dated events. All it assumes is that
the start event occurs before the dated events, and that these all occur before the
end event. For the command files defined for this model see appendix.
Internal consistency and reliability of the model
Agreement index (A) for each dated event is high (A ≥102.6%) (fig. 4.37). The
model shows high overall agreement (Marine 1: Aoverall=111.3%, Marine 2: Aoverall=
111.6%), which indicates that the 14C dates are in accordance with the prior information incorporated in the model. The index of convergence (C) is good (≥96%).
The results obtained when running the model with the different marine values
(M1, M2) are statistically insignificant. For this reason, all further observations and
graphic presentation are based on the results obtained with the model run with the
values of M1. The results obtained with M2 are presented in the tables (table 4.39).
Figure 4.37. Chronological model for the burial activity at Cabeço das Amoreiras, calibrated
using marine 1 diet values. For each of the dates two distributions have been plotted, one
in outline which is the result produced by the scientific evidence alone, and a solid one
which is based on the chronological model used. Calibrated ranges of 95% and 68% probability are given by the lower and higher square brackets respectively. The large square
brackets (left) and the OxCal keywords define the overall model. Model-definition command files in appendix, Sado valley, Cabeço das Amoreiras, Model 1, Marine 1.
Because of the small number of measurements and the scattering of burials over
time, the model estimates a wide time range for the start and end dates of the burial activity probability at the site at 95%, accommodating a wide range of possibilities (figs. 4.38, 4.39). In this case, the variance cannot be determined precisely and
the start and end dates will be very uncertain. These estimates can be misleading
without further archaeological data allowing related early and/or later burial activi273
Chapter 4
ties at the site. The likelihood is greater at 68% probability, and in this case, the
posterior density estimated for the start of the burial activity is 6433–5998 cal BCE
(68% probability), and 5460–5028 cal BCE (68% probability) for the end of the burial
activity. These estimates are consistent with the earliest burial known at the site
represented by CAMS1958, Sk5 (6156–5914 cal BCE, 95% probability), and with the
latest burial known, indicated by CAMS1958, Sk4 (5483–5329 cal BCE, 95% probability). The burial activity is estimated to have continued for between 487 and 744
years (95% probability: site use for burial activity) or for 548 and 667 years (68% probability)
(fig. 4.40).
Figure 4.38. Start of burial activity at Cabeço das Amoreiras derived from the chronological
model 1, calibrated with values for marine 1 diet. Posterior density estimated for the start
of the burial activity is 7754–5928 cal BCE (95.4% probability), 7159–5928 cal BCE (92.5%
probability), or 6433–5998 cal BCE (68.2% probability). Calibrated ranges of 95% and 68%
probability are given by the lower and higher square brackets respectively.
Figure 4.39. End of burial activity at Cabeço das Amoreiras derived from the chronological
model 1, calibrated with values for marine 1 diet. Posterior density estimated for the end of
the burial activity is 5488–3714 cal BCE (95.4% probability), 5488–4212 (93.4% probability), or
5460–5040 (67.9% probability). Calibrated ranges of 95% and 68% probability are given by
the lower and higher square brackets respectively.
274
Results: Sado valley, Cabeço das Amoreiras
Figure 4.40. Duration of burial activity at Cabeço das Amoreiras derived from the chronological model, calibrated with values for marine 1 diet. Posterior density estimated for the
duration of the burial activity is between 487 and 744 years (95% probability), or 548 and 667
years (68% probability). Calibrated ranges of 95% and 68% probability are given by the lower
and higher square brackets respectively.
Table 4.39. Human bone collagen samples (3) from Cabeço das Amoreiras, Sado valley: calibrated date
ranges according to two different approaches for the calculation of the marine percentage (M1, M2). Results
are presented in both (a) cal BCE and (b) cal BP in 95.4% confidence range. The calibrated date ranges
are probability distributions derived from scientific dating alone. The posterior density estimates (in italic)
are derived from Bayesian modelling. Samples are sorted so that the most recent dates are at the top of the
column.
(a) Cal BCE
Age BP
Unmodelled
cal BCE, M1
Calibrated date range
(95% confidence)
Modelled
Modelled
cal BCE, M1 cal BCE, M2
Posterior density estimate
(95% probability)
Ind.
Lab no.
14C
4
Ua–47973
6484±39
5486–5316
5483–5329
5484–5330
7
Ua–47974
6645±42
5626–5478
5620–5481
5620–5481
5
Beta–125110
7230±40
6206–5919
6156–5914
6203–5983
Unmodelled
cal BP, M1
Calibrated date range
(95% confidence)
Modelled
Modelled
cal BP, M1
cal BP, M2
Posterior density estimate
(95% probability)
(b) Cal BP
Ind.
Lab no.
14C
Age BP
4
Ua–47973
6484±39
7435–7265
7432–7278
7433–7279
7
Ua–47974
6645±42
7575–7427
7569–7430
7569–7430
5
Beta–125110
7230±40
8155–7868
8105–7863
8152–7932
275
Chapter 4
Stable isotopes: carbon and nitrogen
Previous measurements: δ13C and δ15N
In the early 2000s, one sample (CAMS1958, Sk3) was analysed in the scope of a
palaeodietary study (Umbelino 2006). These analyses were done at the McMaster
University in Canada and the precision of the results are ± 0.1‰ for δ13C and
± 0.2‰ for δ15N (Umbelino 2006, 208–209). The C:N ratios were not provided
but the laboratory confirmed the quality of the collagen and of the measurements
(Umbelino 2006, 315).
Recently, two teams sampled this collection for stable isotope analysis of carbon
and nitrogen (Fontanals-Coll et al. 2014; Guiry et al. 2015). M. Fontanals-Coll and
colleagues (2014) analysed four samples at Biological Anthropology Department at
the Universitat Autònoma de Barcelona, and duplicate measurements were taken
in the laboratories of the Institute of Environmental Science and Technology at
the same university. The analytical error is below 0.2‰ for both δ13C and δ15N.
One sample (CAMS1958, Sk6) presented values outside the accepted quality range
and was rejected (Fontanals-Coll et al. 2014). E. Guiry and colleagues (2015)
sampled four individuals and the measurements were made on a Thermo-Finnigan
Delta Plus XL isotope ratio mass spectrometer coupled via continuous flow to a
Carlo Erba elemental analyzer at the Department of Human Evolution, Max
Planck Institute for Evolutionary Anthropology, Leipzig, Germany. The instrumental error for δ13C and δ15N measurements was ±0.1‰ and ±0.2‰, respectively (Guiry et al. 2015).
The δ13C value published for CAMS1958, Sk5 (Beta–125110) along with its 14C
date (Cunha and Umbelino 2001) was obtained via AMS and not by IRSM (Cláudia
Umbelino, pers. comm.). This sample was not included in Umbelino’s palaeodietary study (2006), and as previously discussed it should not be used for dietary
reconstructions and/or reservoir corrections.
All previous measurements were included in this study (n=8) as long as they fell
within the accepted quality ranges as discussed.
Sampling strategy
In this study, the selection of samples of human remains for δ13C and δ15N measurements was fully dependent on the goals of the radiocarbon dating programme
(see above).
Results
The samples Ua–number (n=2) were measured by IRSM and the precision of the
measurements for both δ13C and δ15N is ± 0.1‰.
At Cabeço das Amoreiras, the human collagen δ13C values (n=10/no. individuals=5) range from –20.0‰ to –16.0‰ with a mean value of –18.8 ± 1.1‰. The
human collagen δ15N values (n=10/no. individuals=5) range from +7.6‰ to
+13.4‰ with a mean value of +9.6 ± 1.5‰. Individual 6 presents a relatively high
276
Results: Sado valley, Cabeço das Amoreiras
set of values. When these are excluded the values range from –20.0‰ to –18.5‰
(–19.1 ± 0.6‰) for carbon and from +7.6‰ to +10.0‰ (+9.2 ± 0.8‰) for nitrogen.
Different bones from individuals 3, 5 and 7 have been sampled and analysed at
different laboratories and present different values (fig. 4.41, table 4.40).
16
15
14
13
δ15N (‰)
12
11
10
9
8
7
6
-22
-21
-20
-19
-18
-17
δ13C
-16
-15
-14
-13
-12
(‰)
Figure 4.41. Cabeço das Amoreiras, Sado valley: correlation between the δ13C (‰) and δ15N
(‰) values of 10 samples of human bone collagen from 5 individuals. Individuals 3 (triangle), 5 (circle) and 7 (square) have multiple set of values.
277
♂
♂
♂
♂?
♂
♂
♂
♂
♂
♂
Adult,
mature
Adult,
mature
Adult,
mature
Adult
Adult,
mature
Adult,
mature
Adult,
mature
Adult,
mature
Adult,
mature
Adult,
mature
3, 1958
3, 1958
3, 1958
4, 1958
7, 1958
7, 1958
7, 1958
5, 1958
5, 1958
6, 1958
Sex
Age
Ind.
Identification of the human remains
Long bone
999.45.48
Long bone
999.44.50
Long bone
Long bone
Long bone
999.46.79
Fibula
999.46.4
Fibula–L
Long bone
Long bone
999.43.69
Long bone
Bone
analysis for isotopes
analysis for isotopes
6645±42
6484±39
analysis for isotopes
analysis for isotopes
6887.6
6889.3
analysis for isotopes
analysis for isotopes
Beta–125110* 7230±40*
n/p
6886.3
Ua–47974
Ua–47973
n/p
6888.3
–16.0
–18.5
–19.0#
–18.9
–19.2
–20.0
–20.0
–18.5
–18.9
Measurements
14C by AMS; δ13C and δ15N by IRMS
14C Age BP
Lab no.
δ13C (‰)
VPDB
n/p
analysis for isotopes –19.3
13.4
10.0
9.5#
9.7
9.4
8.2
7.6
9.5
9.2
δ15N (‰)
AIR
9.3
3.2
3.3
2.9#
2.9
3.3
2.8
2.9
n/a
3.3
2.9
C:N
52
29
24
25
22
14
14
29
25
21
M1
50
19
13
14
10
0
0
19
14
9
M2
Guiry et al. 2015
Guiry et al. 2015
*Cunha and Umbelino 2001;
et al. 2014
#Fontanals-Coll
Fontanals-Coll et al. 2014
Guiry et al. 2015
This study
This study
Umbelino 2006
Guiry et al. 2015
Fontanals-Coll et al. 2014
% marine (±10) Reference
Table 4.40. Cabeço das Amoreiras, Sado valley: carbon (δ13C) and nitrogen (δ15N) measurements on human bone collagen samples collected at MNA. Samples are sorted
so that the lowest δ13C values are on the top of the table. Multiple samples of the same individual are sorted together.
Results: Sado valley, Cabeço das Amoreiras
Archaeothanatology
Source material and analysis strategy
Archaeothanatological analysis was carried out on a total of 5 individuals, 63% of
the number of individuals (n=8) excavated at Cabeço das Amoreiras (table 4.41).
Analysis was not possible on the remains of CAMS1958, Sks 1 (sex, age: n/d) and
6 (child) due to poor preservation of the material. CAMS1958, Sk2 is represented
by its cranium only, and this feature will be discussed throughout this section. The
source material comprises field photographs, drawings, a site plan of human remains, and the paraffin blocks of one individual.
The first phase of analysis took place at the MNA, where the individual preserved in paraffin blocks (CAMS1958, Sk4) was analysed and photographed. The
second phase of analysis consisted of the reassessment of my observations done in
the museum, crosschecked with the photographed material and the archival
graphic documentation. The analyses based on field photographs and individual
drawings exclusively (CAMS1958, Sks 3, 5, 7, 8) followed the same analytical
protocol (see chapter 3, Archaeothanatology).
Table 4.41. Cabeço das Amoreiras, Sado valley: source material used for archaeothanatological analysis.
This material is curated at the MNA.
Ind.
Source material
Block
Observations
Graphic documentation w/human Written
remains
documentation
Field
Drawings Site plan
photographs
n/a
yes
yes
n/a
Few small bone fragments of unidentifiable material.
Archaeothanatology is not possible.
1, 1958
n/a
2, 1958
n/a
yes
yes
yes
n/a
Isolated human cranium and limited
documentation.
Archaeothanatology is not possible.
3, 1958
n/a
yes
yes
yes
n/a
–
4, 1958
yes
yes
yes
yes
n/a
–
5, 1958
n/a
yes
yes
yes
n/a
–
6, 1958
n/a
yes
yes
yes
n/a
Very fragmented material.
Archaeothanatology is not possible.
7, 1958
n/a
yes
yes
yes
n/a
–
8, 1958
n/a
yes
yes
yes
n/a
–
279
Chapter 4
Results of the archaeothanatological analysis
The nature of the deposits
Analysis of the nature of the deposits at Cabeço das Amoreiras indicates an
unequivocal pattern of primary burials (table 4.42). The cadavers were placed in the
burial features before the degradation of the labile joints, retaining their anatomical
integrity. In two cases (CAMS1958, Sks 5, 8) the documentation is unclear and the
diagnostic criteria are poorly preserved. However, in both cases, the overall position of the skeletal elements indicates that the bodies were disposed while still
retaining their general anatomical integrity, supporting the interpretation of burials
in primary position. One possible exception to this pattern is presented by a cranium (CAMS1958, Sk2) lying disarticulated on the left side of CAMS1958, Sk3,
which, as discussed further, may suggest a burial in a secondary position (see below, Post-depositional manipulations of the cadavers).
The space of decomposition of the cadaver
Analysis of the space of decomposition of the cadavers buried in Cabeço das
Amoreiras indicates a general pattern of filled spaces (table 4.43). The cadavers
were covered with sediment immediately after their placement in the features. The
decomposition of the bodies took place in this sediment-filled environment. In
two cases (CAMS1958, Sks 5, 8) the analysis is unclear due to the poor preservation of the remains. In every case, the movement of the bones is limited suggesting
that the space of decomposition was filled and that the sediment could penetrate
immediately.
The initial position of the cadaver in the feature
At Cabeço das Amoreiras all cadavers (n=5) were lying on the back in the grave
feature (table 4.44).
The position of the upper limbs was variable but they were always placed close
to the body. In two cases the limbs were arranged symmetrically (CAMS1958, Sks
4, 7) either in semi-flexion at the level of the elbows, with the hands towards the
pelvic region (Sk4) or in extension along the body with the hands lying partially
under the buttocks (Sk7). In two cases the limbs were placed in asymmetrical positions (CAMS1958, Sks 5, 8), with the right limb in slight semi-flexion at the elbow
and the hand placed on the right hip, while the left limb was in semi-flexion (c 90°)
at the elbow, with the forearm laid in front of the abdomen and the hand on the
right side of the trunk. The symmetry of the upper limbs was not possible to identify in one case (CAMS1958, Sk3) due to the fragmentation of the bones and unclear documentation. In this case, the right upper limb was in flexion at the elbow
and the hand was placed towards the left shoulder lying in front of the face.
At Cabeço das Amoreiras, the bodies were placed in the feature with the lower
limbs always in flexion or hyperflexion at the level of the knee joints. The degree
of flexion at the hip joints is variable and in one case these are almost in extension
(CAMS1958, Sk7). The lower limbs were rotated to the right in two cases
280
Results: Sado valley, Cabeço das Amoreiras
(CAMS1958, Sks 4, 8) and to the left in one case (CAMS1958, Sk3). In two cases
the lower limbs had a slight tendency towards the right (CAMS1958, Sk5) or the
left side (CAMS1958, Sk7).
Clumping of the body with the lower limbs brought up towards the upper body
was moderate (CAMS1958, Sks 3, 4), with the limbs being placed on one of the
sides. Clumping at the level of the lower body was common and in every case
(n=5) the feet of the cadavers were positioned towards the buttocks indicating a
constrained position of the lower limbs.
281
♂
♂?
♂
♂
Adult, mature
Adult
Adult, mature
Adult, mature
Non-adult, 7.5-16 yrs
3, 1958
4, 1958
5, 1958
7, 1958
8, 1958
n/d
Sex
Age
Individual
Probably primary
Primary
Probably primary
Primary
Primary
Nature of the deposit
The labile articulations are poorly preserved and unclear in the documentation.
However, the overall position of the bones indicates that the general anatomical integrity of the
body has been maintained.
Maintenance of labile articulations: metatarsals and phalanges of right foot.
Maintenance of the general anatomical integrity of the body.
The labile articulations are poorly preserved and unclear in the documentation. However, the
overall position of the bones indicates that the general anatomical integrity of the body has been
maintained.
Maintenance of labile articulations: metatarsals and phalanges of left foot.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: some metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Arguments
Table 4.42. Summary of the nature of the deposits with human remains at Cabeço das Amoreiras, Sado valley.
♂?
♂
♂
Adult
Adult, mature
Adult , mature
Non-adult, 7.5–16 yrs
4, 1958
5, 1958
7, 1958
8, 1958
.
♂
Adult, mature
3, 1958
n/d
Sex
Age
Individual
Probably filled
Filled
Probably filled
Filled
Filled
Poor preservation of diagnostic criteria.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: the patellae are suspended on the distal
ends of the femora; the bones of the right foot are rotated.
No movement outside initial volume of the cadaver.
Poor preservation of diagnostic criteria.
Maintenance of left os coxae in original unbalanced position.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: the bones of the left foot are in contraction and rotated inwards; the left patella is suspended on the distal end of the right femur. Also,
the cranium is maintained in a semi-suspended position.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: the bones of the feet which are rotated
outwards; the right patella is suspended on the distal end of the right femur. Also, the cranium is
maintained in a semi-suspended position.
No movement outside initial volume of the cadaver.
Space of decomposition Arguments
Table 4.43. Summary of the space of decomposition of the human remains in the grave features at Cabeço das Amoreiras, Sado valley.
Sex
♂
♂?
♂
♂
Individual Age
Adult, mature
Adult
Adult, mature
Adult, mature
Non-adult, 7.5–16 yrs n/d
3, 1958
4, 1958
5, 1958
7, 1958
8, 1958
Trunk
Raised
Raised
Raised
On the back
On the back
On the back
Raised
On the back
Rotation–R and downwards
(neck flexed forwards)
Raised
On the back
Rotation–L and downwards
(neck flexed forwards)
Head
Initial position of the cadaver in the feature
In semi–flexion–R
In semi–flexion (c 90°)–L
In extension
In semi–flexion–R
In semi–flexion (c 90°)–L
In semi–flexion
In flexion–R
n/a–L
Upper limbs
Hip joints: semi–flexion
Knee joints: flexion
Rotation–R
Feet towards buttocks
Hip joints: slight semi–flexion
Knee joints: hyperflexion
Rotation–tendency L
Feet towards buttocks
Knees directed downwards
Hip and knee joints: hyperflexion – R
Hip and knee joints: flexion – L
Rotation–R
Feet towards buttocks
Clumping
Hip joints: semi–flexion
Knee joints: flexion
Rotation–tendency R
Knees directed forwards
Hip and knee joints: hyperflexion
Rotation–L
Feet towards buttocks
Clumping
Lower limbs
Table 4.44. Summary of the reconstruction of the initial position of the cadaver in the grave feature at Cabeço das Amoreiras, Sado valley. The initial position of the limbs is
indicated as –R or –L and refers to the right or left limb, when their positions are different. Rotation –R or –L indicates the rotation of the skeletal element(s) towards the
right or the left side.
Results: Sado valley, Cabeço das Amoreiras
The grave features
The size and shape of the deposits with human remains at Cabeço das Amoreiras
were not described by the excavators. The archaeothanatological analysis suggests
that the bodies were placed in pits dug into the ground. The lateral pressures and
wall-effects observed on the skeletons in the feature (table 4.45) indicate that the
shape and size of the pits was just sufficient to contain the body. The analysis further indicates that in some cases the design of the grave had a significant impact on
the final arrangement of the body in the feature, visible by the constrained positions of some individuals (e.g., CAMS1958, Sks 3, 4). However, in one case the pit
was somewhat wider allowing a relative expanded position of the body particularly
at the level of the lower body (CAMS1958, Sk8).
The floor of the graves tended to be irregular and sloping from the upper to the
lower end. This is visible in some of the drawings in profile and confirmed by the
raised position of the head in relation to the trunk (CAMS1958, Sks 5, 7, 8) or by
the flexion and collapse forwards and downwards of the head (CAMS1958, Sks 3,
4). The fragmented state of the material does not allow detailed observations at the
level of ribs and sacrum the rotations of which during the process of decomposition could clarify aspects of the original position of the body in the feature in relation to the characteristics of the floor of the grave, such as in the case of bodies
lying in a half-sitting position. Also, the possibility cannot be excluded that the
elevation of the head could have been aided by some kind of support, other than
the floor of the pit, but the documentation is unclear.
Often, the narrow shape of the feature imposed significant physical limits to the
cadaver, offering lateral support and preventing the collapse of the bones. However, the analysis suggests that in some cases there might have been an extra support offered by some kind of soft container such as a light wrapping. CAMS1958,
Sk7 is one of these cases, where the strong pressures observed could suggest the
existence of a soft wrapping at the time of the burial. As in all cases observed at
Cabeço das Amoreiras, CAMS1958, Sk7 was lying on the back with the head
slightly elevated in relation to the trunk. However, this case presents the only example of an individual placed with the upper limbs in complete extension adjacent
to the upper body with the hands lying very close to the hips. The lower limbs are
in hyperflexion at the knees which are directed downwards in the opposite direction of the feet which were placed towards the buttocks (fig. 4.42).
The most striking feature of this burial is the overall compression through the
medial axis of the body, which is observed throughout the skeleton. The analysis
suggests two main pressures exerted on the body in the feature. One is a bilateral
pressure visible in the alignment of the upper limbs which are rotated inwards.
This is illustrated by the right scapula which presents the anterior view with a lateral component, and is rotated forwards and upwards, while the right clavicle became verticalized and disarticulated from the scapula indicating strong projection
forwards of the shoulder.
285
Chapter 4
Figure 4.42. CAMS1958, Sk7. The transverse compression exerted on the skeleton indicates
significant physical limits on the cadaver. This case could suggest the existence of extra
support provided by a soft wrapping at the time of the burial, but the data is not conclusive. (MNA, APMH, F807)
The bone preservation in the left side is poor and the diagnostic elements at the
level of the shoulder girdle are not preserved. However, the compression and verticalization of the ribs, and the alignment of the arm and forearm placed very close
to the thoracic cage, are strong indicators of a lateral pressure at this level. The
second main pressure observed in this feature is an overall wall-effect at the lower
end of the grave visible by the contraction and clumping of the lower limbs which
were maintained on an extremely constrained position. The pressures observed
may suggest a tight wrapping of the body. However, the overall compression of
the skeleton could also be explained by the narrow characteristics of the pit where
the walls of the grave could have had the same supporting effect on the skeletal
elements, which seems to be a common element at this burial ground. Nothing in
this feature suggests that the pit was wider than the area occupied by the body, and
the hypothesis of an unwrapped body placed in a very narrow pit remains as valid
as the wrapping possibility.
286
Results: Sado valley, Cabeço das Amoreiras
Likewise, in the feature containing the remains of CAMS1958, Sk5 there is a
strong bilateral compression of the upper body indicated by the oblique position of
the right clavicle and the verticalization of the left clavicle, which are clear indications of strong projection forwards of the shoulders. However, in this case, other
diagnostic elements such as the scapulae and the thoracic cage are not preserved,
and further observations are not possible with the current documentation.
Deposits containing more than one individual
At Cabeço das Amoreiras all deposits with human remains contained the remains
of one individual.
Post-depositional manipulations of the cadavers
At Cabeço das Amoreiras, the skeletons are apparently complete and lie undisturbed in the grave feature. There is evidence suggesting the post-depositional
manipulation of the cadavers.
This is documented by the feature with the disarticulated cranium of
CAMS1958, Sk2 lying on the left side of the complete skeleton of CAMS1958, Sk3
(fig. 4.43). This deposit consists of a cranium poorly preserved, apparently without
the mandible or any postcranial bones. It is unclear whether this cranium was
originally placed in the feature in articulation with its postcranial skeleton, which
would later be completely removed, or if the cranium was deposited already in
disarticulated state. The field photograph of the cranium (fig. 4.44) shows fragments of what could be a portion of the pelvic girdle, yet it is unclear if these
fragments are skeletal elements. Unfortunately, the remains of CAMS1958, Sk2 are
not inventoried in the material curated at the MNA. Nevertheless, the description
on the back of this photo seems to support the hypothesis of a feature with an
isolated cranium: “cranium II (fragments) without skeleton”. Also, the drawing of
the remains in the feature presents the cranial fragments only. In any case, this
feature suggests the handling of the human remains after death. It is not clear,
however, if the cranium was placed in the feature already in disarticulated state and
clean of most soft tissue, or if this is the original place of decomposition of
CAMS1958, Sk2, the postcranial skeleton of which has been removed sometime
after the decomposition of the soft tissues.
The relationship between CAMS1958, Sk2 and Sk3 is also ambiguous. The relative position of the body of individual 3 towards cranium 2 could suggest a common and intentional arrangement. The remains of both individuals lie at the same
depth, in the same layer (4: chestnut brown sand) and in very close proximity, but
do not disturb or come in contact with each other. Interestingly, individual 3 presents a general rotation towards the left side where the cranial fragments of individual 2 are deposited. The interval between these deposits is not possible to determine and despite the close proximity and suggestive arrangement of the remains
the evidence supporting the interpretation of a related and/or simultaneous deposit is not conclusive.
287
Sex
♂
♂?
♂
♂
Age
Adult, mature
Adult
Adult, mature
Adult , mature
Non-adult, 7.5–16 yrs n/d
Individual
3, 1958
4, 1958
5, 1958
7, 1958
8, 1958
1. Wall-effect, upper left side;
2. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, upper right side;
2. Wall-effect, upper left side;
3. Wall-effect, lower end.
Prob. soft wrapping.
1. Wall-effect, upper right side;
2. Wall-effect, upper left side;
3. Wall-effect, lower end.
Prob. design of the grave feature.
1. Wall-effect, upper right side;
2. Wall-effect, upper left side;
3. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, upper right side;
2. Wall-effect, left and lower side.
Design of the grave feature.
Pressures
1. Suggested by the alignment of left arm;
2. Alignment and constraint effect towards lower limbs forcing feet
towards buttocks.
1. Alignment of upper right limb, involving shoulder, arm and forearm;
2. Alignment of left arm and forearm;
3. Alignment and constraint effect towards lower limbs forcing feet
towards buttocks.
1. Alignment of upper right limb, involving shoulder, arm and forearm;
2. Alignment of upper left limb, involving the shoulder and arm;
3. Alignment of and constraint effect towards lower limbs forcing feet
towards buttocks.
1. Alignment of the upper right limb, involving right shoulder and arm;
2. Suggested by projection forwards and upwards of left shoulder;
3. Affecting lower limbs and forcing feet towards buttocks. Also, visible
by contraction of left foot.
1. Projection forwards and upwards of right shoulder;
2. Alignment of and constraint effect towards lower limbs forcing feet
towards buttocks.
Arguments
Table 4.45. Summary of the key observations used in the reconstruction of the grave features at Cabeço das Amoreiras, Sado valley. Observed pressures on the body are
presented in a number sequence. The arguments supporting these observations are indicated by the same number in the respective column.
Results: Sado valley, Cabeço das Amoreiras
Figure 4.43. Cabeço das Amoreiras, Sado valley, 1958: field photograph of individual 3
(complete skeleton) and individual 2 (cranium). (MNA, APMH, F790)
289
Chapter 4
Figure 4.44. Cabeço das Amoreiras, Sado valley, 1958: field photograph of individual 2
(cranium). Back of photograph: Cranium II (fragments) without skeleton (Craneo II [fragmentos]
sem esqueleto). (MNA, APMH, F792)
The mortuary programme at Cabeço das Amoreiras
Cabeço das Amoreiras has a relatively low number of individuals (n=8), one of
which is represented by an isolated cranium (CAMS1958, Sk2). Nevertheless, it is
possible to observe some patterns in the mortuary practices at the site.
The individuals buried at Cabeço das Amoreiras were predominantly adult
males except in the case of two individuals belonging to juvenile skeletons
(CAMS1958, Sks 6, 8) the sex of which is not determinable. Also, sex and age of
individuals 1 and 2 are not possible to determine due to fragmentation of the
material.
290
Results: Sado valley, Cabeço das Amoreiras
At Cabeço das Amoreiras the cadavers were typically placed in individual grave
features while retaining their anatomical integrity (individual primary deposits). The
feature was then immediately covered with sediment, providing a filled space of
decomposition. As discussed, one case (CAMS1958, Sk2) suggests the postdepositional manipulation of the cadaver, which could either indicate the
re-opening and removal of the postcranial skeleton from an individual primary
deposit or, more likely, it could indicate the secondary deposition of selected
elements, in this case, the cranium.
In the case of the individual primary deposits, the cadavers were always lying on
the back with the upper limbs positioned close to the body in various arrangements. The lower limbs were normally rotated to one side and were always flexed
or hyperflexed at the knees. The degree of flexion at the hip joints was more variable. In some cases the hip joints were flexed or hyperflexed and the limbs were
brought up towards the upper body, but in other cases these joints were almost in
extension. Interestingly, in any of these variants, the feet of the cadavers were always placed towards the buttocks in a forced position, and clumping at the level of
the lower body was a common characteristic of burial practices at Cabeço das
Amoreiras.
The floor of the graves tended to be uneven and sloping from the upper to the
lower end of the feature. The heads of the cadavers were often rotated forwards
and downwards towards the body, and in some cases the head was in a more elevated position than the trunk.
In most cases the pressures exerted on bodies in features can be explained as
the result of the design of the grave. Typically, the wall-effects and consequent
pressures corresponded to the physical limits of the pit, just wide enough to contain the individual. However, some observations may suggest the tight wrapping of
the body at the time of the burial, at least in one case (CAMS1958, Sk7), but the
data is not conclusive. In contrast to the typical narrow graves, in one case
(CAMS1958, Sk8) the pit was somewhat wider allowing a relatively expanded
position of the body, particularly at the level of the lower body.
Overall, the mortuary programme at Cabeço das Amoreiras presents a consistent set of practices and all cases observed follow a common way of burying the
dead.
291
Sado valley: Vale de Romeiras
Site background: a short history of research
Vale de Romeiras is a shell midden site located on the eastern side of the Sado
valley (fig. 2.3). The site is located on the right margin of the Sado River at
55 m.a.s.l. (Diniz et al. 2012), and has excellent visibility over the river valley. From
here we can see the site of Cabeço das Amoreiras on the opposite margin some
kilometres away, and the site of Cabeço do Pez on the neighbouring hilltop.
Archaeological material and documentation available for Vale de Romeiras are
curated at the National Museum of Archaeology in Lisbon (MNA). The year of
identification of the site is unclear, but according to the literature it can be assumed
to be sometime between 1950 and 1955 (Heleno 1956, 229). The documentation
indicates that Vale de Romeiras was excavated in two seasons in 1959 and 1960 by
the staff of MNA. The excavation diary (MNA, APMH, 2/3/6/1) shows that the
first field season started on the 18th of August 1959, but unfortunately, like other
documents, this report is incomplete. During the first five days at the site, the excavators identified three skulls and some bone elements in articulation, and at least
two burials were completely excavated. Yet the site plans indicate that c 23 individuals were recovered in 1959. Most of the burials were excavated after these first
five days but those pages are missing from the archives. The site was excavated
briefly in 1960, but since no human remains were found, the work was dismissed
and the field team moved to Poças de S. Bento (MNA, APMH, 51/1/654/51 and
53, Aug. and Sep. 1960).
Graphic documentation consists of one site plan with all burials in the excavation context (fig. 4.45), several individual drawings of the human burials and fauna,
and photographs. Unfortunately, the quality of the 58 photographs available for
Vale de Romeiras (MNA, APMH, 2/11/83) is poor and most images are extremely
blurry and indistinguishable. Drawings of the human remains are, however, the
best graphic documentation available for any of the shell middens in Portugal. This
documentation is available in the format of field sketches with 33 drawings of
human remains and fauna, a small booklet of 10 pages with all burials in 1:20 scale
and eight detailed ink drawings of individuals 4, 7, 8, 10, 11, 16, and 18, as well as
faunal concentrations (MNA, 1959, B37–B44). All drawings are from 1959 by
Dario de Sousa, MNA, and include a few notes on stratigraphy and orientation of
the features.
292
Results: Sado valley, Vale de Romeiras
Figure 4.45. Vale de Romeiras, Sado valley: site plan and profiles of the excavation area.
(Adapted from a photograph by J. P. Ruas. MNA, 1959, D. Sousa, A120)
293
Chapter 4
Human skeletons are in disarticulated state and preserved in containers. The osteological material was studied by a team of anthropologists in the 1990s (Cunha and
Umbelino 1995–97, 2001; Umbelino and Cunha 2012). The MNI identified in the
museum is 26 while just 23 are represented in the site plan. The burial population
studied by E. Cunha and C. Umbelino consists of 20 adults (five females, five
males, ten indeterminate), six of which are young adults under 30 years old, and six
are non-adults in the first infancy (Cunha et al. 2002, 191, 192) (table 4.46).
Table 4.46. Vale de Romeiras, Sado valley: minimum number of individual (MNI) estimated for the
material currently curated at MNA.
MNI
Adults
♀
♂
5
26
5
20
Non-adults
Reference
Indet.
10
6
Cunha and Umbelino 1995–97
The following table (table 4.47), courtesy of one of the anthropologists who
studied this material in detail (Umbelino 1999), is the reference used in this study
regarding sex and age estimates for Vale de Romeiras.
Table 4.47. Vale de Romeiras, Sado valley: sex and age estimates of 20 individuals. (Data from Cunha
and Umbelino 1995–97, 2001; Umbelino 1999)
Individual
Container
Age
Sex
3
4
5
7
8
9
10
11
12
13
14
15
16
17
18
19
20
22
1 or 21?
23?
6186
6115
6194
6156
6123
6203
6155
6115
6122
6118
6161
6189
6157
6185
6162
6195
6119
6116
6190
6197
Adult
Adult
Adult
Adult
Non-adult, 3–4 yrs±12 m.
Adult
Adult, young
Adult, young
Adult
Adult
Adult
Adult
Non-adult, 12 yrs±36 m.
Non-adult, 3 yrs±12 m.
Non-adult, 2–6 yrs
Adult, young
Adult
Adult
Adult, young
Adult
n/d
♂
n/d
♀
n/d
♂
♂
♀
n/d
n/d
n/d
n/d
n/d
n/d
n/d
♀
n/d
♂
♀
n/d
294
Results: Sado valley, Vale de Romeiras
Stratigraphic data is provided by the graphic documentation (table 4.48). Most
burials (n=14) were found lying on the soft bedrock (layer 6, piçarreira) but at various depths ranging from 10 to 84 cm. One short entry in the excavation diary
describes the first two burials excavated at the site, possibly individuals 1 and 2,
being recovered at 20–30 cm from the top. In the same entry the excavator predicts that the maximum depth of the shell midden is probably not more than 50
cm. This is because individual 1 was found in the soft bedrock layer, at only 20 cm
deep. However, most burials were recovered at lower depths, suggesting a significant erosion of the midden in the south-west area, possibly due to a downslope
position, where also individual 1 was found. Likewise, individuals 5, 13 and 14
buried in the same area, were found on the bedrock layer at depths ranging from
10 to 18 cm.
Table 4.48. Vale de Romeiras, Sado valley: stratigraphy and depth of the human burials. Samples are
sorted so that the upper layer is on top of the table. (Data from field drawings, MNA, 1959, D. Sousa)
Individual
Layer
Depth (cm)
2
2, grey soil with (few) shells
15
8
2, grey soil with shells
35
15
4, yellowish clay-like soil, beginning
48
17
4, yellowish clay-like soil, beginning
48
16
4, yellowish clay-like soil, beginning
49
18
4, yellowish clay-like soil
54
21
5, chestnut brown sand
45
19
5, chestnut brown sand
60
4
5/6, chestnut brown sand/soft bedrock
50
13
6, soft bedrock
10
14
6, soft bedrock
10
5
6, soft bedrock
18
1
6, soft bedrock
20
6
6, soft bedrock
30
7
6, soft bedrock
32
3
6, soft bedrock
37
9
6, soft bedrock
43, cranium; 62, pelvic girdle
11
6, soft bedrock
60
12
6, soft bedrock
60
10
6, soft bedrock
70
22
6, soft bedrock
75
20
6, soft bedrock
84
23
6, soft bedrock
84
295
Chapter 4
Three burials (Sks 4, 19, 21) were found in layer 5, described as chestnut brown
sand (terra acastanhada, areenta), at depths ranging from 45 to 60 cm. Interestingly
both burials 19 and 21 are eccentric to the burial ground, lying on the northernmost area of the site. Individual 4 is in the interface of layer 5/6 and well positioned with other burials. Three of the four burials found in yellowish soil (layer 4,
terra amarelada, barrenta) are children (Sks 16, 17, 18) buried next to each other, c 50
cm deep, on the west part of the burial area. No burials were found in layer 3,
described as a black soil (terra preta), suggesting a possible hiatus in the burial activity. Layer 2 is the layer containing shells (grey soil with shells, terra cinzenta e conchas)
where just two individuals were found (Sks 2, 8). The top layer (layer 1, brownish
soil, terra acastanhada, vegetal) covered the layer with shells and had no burials.
Chronology of the burial practice
Previous 14C measurements
Three 14C measurements on the material excavated in 1959 were carried out in the
late 1980s: one sample on unidentified fauna (ICEN–144) and two samples on
shells (ICEN–146 and ICEN–150) (Arnaud 2000) (table A11).
Previous attempts to date the human remains (individuals 8 and 12) were unsuccessful due to insufficient quantity of bone collagen (Cunha et al. 2002, 189;
Umbelino 2006, 137) (table 4.49).
Objectives of the dating programme
The dates available on non-human material (ICEN–144, ICEN–146 and ICEN–
150) indicate an early use of the site, which is coeval with the non-human sample
available for Arapouco (Q–2492) but several centuries older than both the human
(Beta–125109) and non-human (Q–2496 and Q–2497) measurements available for
the neighbouring site Cabeço do Pez.
The burial area of this site is relatively small but with high concentration of human burials. The general aims of the dating programme for Vale de Romeiras followed the macro-goals of the radiocarbon programme of this dissertation, outlined
in chapter 3, and were:
• to define when the burial activity started and when it ended;
• to estimate the duration of the use of this site for burial activity;
• to estimate the frequency of the burial activity in the site.
The site-specific aims of the dating programme for Vale de Romeiras were:
• to test the suggestion of a coeval group of burials;
• to test the suggestion of this site being used as burial ground for the
neighbouring site Cabeço do Pez.
296
♂
♀
Adult
Adult
Ua–46967
Ua–47982*
Ua–46968
Ua–47983*
Ua–46969
Ua–47984*
Ua–47985*
Ua–46970
Ua–46971
Ua–47986*
Ua–46972
Ua–47987*
Ua–47988*
4
7
8
9
11
12
13
14
16
18
19
20
22
Tibia–R
999.25.7
Bone
Adult
Adult
Adult
Adult, 20–35 yrs
Non-adult, 2–6 yrs
Non-adult, 12 yrs±36 m.
n/d Tibia-L
n/d Tibia
999.14.15
n/d Femur
999.19.2
n/d Tibia–L
999.15.12
♀
Tibia–R
999.29.9
n/d Femur–R
999.16.1(?)
♂
Fibula
999.6.6
Adult
Tibia–R
999.27.19
Non-adult, 3–4 yrs±12 m. n/d Tibia
999.18.18
Adult
♂
Radius–R
999.26.10
Adult, 20–35 yrs
♀
Humerus–L
999.24.53
Adult
n/d Long bone
Sex
Age
Ind. Lab no.
Observations
Layer 6, bedrock, 75 cm
Layer 6, bedrock, 84 cm
Layer 5, brownish soil, sand like, 60 cm
Layer 4, yellowish soil, 54 cm
Layer 4, yellowish soil, 49 cm
Layer 6, bedrock, 10 cm
Not measurable
Doubtful quality
Doubtful quality
Doubtful quality
Doubtful quality
Doubtful quality
Doubtful quality
Doubtful quality
Layer 6, bedrock, 60 cm
Layer 6, bedrock, 10 cm
Not measurable
Previous 14C dating without results
(Cunha et al. 2002, 189)
Not measurable
Not measurable
Layer 6, bedrock, 60 cm
Layer 6, bedrock, 43-62 cm
Layer 2, grey soil with shells, 35 cm
Layer 6, bedrock, 32 cm
Layer 5/6, brownish sand/bedrock, 50 cm Previous stable isotope analysis by Umbelino 2006
Ua–46967: doubtful quality
Stratigraphy
Table 4.49. Samples of human bone (13) from Vale de Romeiras collected and submitted for 14C measurement in the course of this study. All samples are from the 1959
assemblage and were collected at MNA. *Samples pre-treated with boiling in acetone, ultrasonification first with ether and then with ethanol, and finally filtrated; after that
processed as usual (see chapter 3).
Chapter 4
Sampling strategy
Thirteen samples of human bone were selected for 14C analysis (table 4.49). The
sampling strategy was targeted towards a good spatial and stratigraphic coverage.
In some cases it was not possible to obtain samples from the pre-selected individuals due to preservation/curation of the material.
Results
Ten of the thirteen samples submitted as part of this study revealed poor quality
during laboratory analysis and were rejected (table A1). Three 14C measurements
on human bone collagen samples from Vale de Romeiras are available now (table
4.50). One of these measurements (Ua–46968, VR1959, Sk8) has a modern age and
for this reason it was excluded from the chronological modelling, but its context is
discussed further in chapter 5.
Table 4.50. Human bone collagen (3) samples from Vale de Romeiras, Sado valley. 14C measurements
and estimation of non-terrestrial carbon intake (% marine). Samples are sorted by lowest to highest %
marine. Marine percentage was estimated from the measured δ13C value of each sample by the application of
method marine 1 and method marine 2, based on adopted endpoint values of –12‰ and –20‰ for marine
(100%) and terrestrial (100%) diets respectively.
Identification of the
human remains
Measurements
14C by AMS; δ13C and δ15N by IRMS
Ind. Excav.
Museum
Lab no.
14C Age
BP
δ13C (‰)
VPDB
δ15N (‰)
AIR
C:N M1
M2
9
1959
MNA
Ua–47983
6625±51
–20.5
7.7
3.4
10
0
This study
19
1959
MNA
Ua–46972
7640±55
–20.2
8.7
3.2
12
0
This study
8
1959
MNA
Ua–46968
380±30
–18.3
14.4
3.4
30
21
This study
% marine
(±10)
Reference
Calibration and chronological models
The chronological model is limited to two 14C dates, both obtained in the scope of
this study (table 4.51, fig. 4.46). Individual 8 (Ua–46968) was not included in the
model because its date is modern and is out of the scope of this study (table A2).
This is a phase model which groups all dated burial events in a phase of burial
activity (Bronk Ramsey 1998, 2000, 2001). This grouping is constrained by a start of
burial activity event, and an end of burial activity event (i.e., boundaries) imposing no
internal constraints to the relative order of the dated events. All it assumes is that
the start event occurs before the dated events, and that these all occur before the
end event. For the command files defined for this model see appendix.
298
Results: Sado valley, Vale de Romeiras
Internal consistency and reliability of the model
Agreement index (A) for each dated event is high (A≥97%). The model shows
high overall agreement (Marine 1: Aoverall=104%, Marine 2: Aoverall=101.2%), which
indicates that the 14C dates are in accordance with the prior information incorporated in the model. The index of convergence (C) is good (≥97%).
The results obtained when running the model with the different marine values
(M1, M2) are statistically insignificant. For this reason, all further observations and
graphic presentation are based on the results obtained with the model run with the
values of M1. The results obtained with M2 are presented in the tables (table 4.51).
This model was built with a small number of measurements. For this reason, and
because of the scattering of the dates, the model estimates a wide time range for
the start and end dates of the burial activity at 95% probability, and accommodates
a large range of possibilities, although with low likelihood. As in the model for
Cabeço das Amoreiras (see above), the variance cannot be determined precisely
and the start and end dates will be very uncertain. Without further archaeological
data allowing related early and/or later burial activities at the site the estimates for
start and end dates should be considered with caution (figs. 4.47, 4.48). The burial
activity is estimated to have continued for between 779 and 1074 years (95%
probability: site use for burial activity) or for 842 and 978 years (68% probability) (fig. 4.49).
Table 4.51. Human bone collagen (2) samples from Vale de Romeiras, Sado valley: calibrated date
ranges according to two different approaches for the calculation of the marine percentage (M1, M2). Results
are presented in both (a) cal BCE and (b) cal BP in 95.4% confidence range. The calibrated date ranges
are probability distributions derived from scientific dating alone. The posterior density estimates (in italic)
are derived from Bayesian modelling. Samples are sorted so that the most recent dates are at the top of the
column.
(a) Cal BCE
Ind.
Lab no.
14C
9
Ua–47983
6625±51
Unmodelled
cal BCE, M1
Calibrated date range
(95% confidence)
5633–5471
19
Ua–46972
7640±55
6598–6371
6593–6370
14C
Modelled
Modelled
cal BP, M1
cal BP, M2
Posterior density estimate
(95% probability)
7574–7424
7580–7435
8542–8319
Age BP
Modelled
Modelled
cal BCE, M1
cal BCE, M2
Posterior density estimate
(95% probability)
5625–5475
5631–5486
6596–6415
(b) Cal BP
Ind.
Lab no.
9
Ua–47983
6625±51
Unmodelled
cal BP, M1
Calibrated date range
(95% confidence)
7582–7420
19
Ua–46972
7640±55
8547–8320
Age BP
8545–8364
299
Chapter 4
Figure 4.46. Chronological model for the burial activity at Vale de Romeiras, calibrated
using marine 1 diet values. For each of the dates two distributions have been plotted, one
in outline which is the result produced by the scientific evidence alone, and a solid one
which is based on the chronological model used. Calibrated ranges of 95% and 68% probability are given by the lower and higher square brackets respectively. The large square
brackets (left) and the OxCal keywords define the overall model. Model-definition command files in appendix, Sado valley, Vale de Romeiras, Model 1, Marine 1.
Figure 4.47. Start of burial activity at Vale de Romeiras derived from the chronological
model 1, calibrated with values for marine 1 diet. Posterior density estimated for the start
of the burial activity is 8459–6409 cal BCE (95.4% probability), 7736–6409 cal BCE (78.2%
probability), 8456–6425 cal BCE (68.2% probability), or 7283–6425 cal BCE (64.9%). Calibrated
ranges of 95% and 68% probability are given by the lower and higher square brackets respectively.
300
Results: Sado valley, Vale de Romeiras
Figure 4.48. End of burial activity at Vale de Romeiras derived from the chronological
model 1, calibrated with values for marine 1 diet. Posterior density estimated for the end of
the burial activity is 5601–3594 cal BCE (95.4% probability), 5601–4249 cal BCE (79.5%
probability), 5597–3597 cal BCE (68.2% probability), or 5597–4749 cal BCE (64.9% probability).
Calibrated ranges of 95% and 68% probability are given by the lower and higher square
brackets respectively.
Figure 4.49. Duration of burial activity at Vale de Romeiras derived from the chronological
model, calibrated with values for marine 1 diet values. Posterior density estimated for the
duration of the burial activity is between 779 and 1074 years (95% probability), or 842 and 978
years (68% probability). Calibrated ranges of 95% and 68% probability are given by the lower
and higher square brackets respectively.
301
Chapter 4
Stable isotopes: carbon and nitrogen
Previous measurements: δ13C and δ15N
In the early 2000s, one sample from the 1959 excavations (VR1959, Sk4) was
analysed in the scope of a palaeodietary study (Umbelino 2006). These analyses
were done at the McMaster University in Canada and the precision of the results
are ± 0.1‰ for δ13C and ± 0.2‰ for δ15N (Umbelino 2006, 208–209). The C:N
ratios were not provided but the laboratory confirmed the quality of the collagen
and of the measurements (Umbelino 2006, 315).
Recently, one team sampled this collection for stable isotope analysis of carbon
and nitrogen (Guiry et al. 2015), and obtained seven reliable measurements (table
4.52) from a total of 17 samples. This material was measured on a ThermoFinnigan Delta Plus XL isotope ratio mass spectrometer coupled via continuous
flow to a Carlo Erba elemental analyzer at the Department of Human Evolution,
Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany. The instrumental error for δ13C and δ15N measurements was ±0.1‰ and ±0.2‰, respectively (Guiry et al. 2015).
All previous measurements were included in this study (n=8) as long as they fell
within the accepted quality ranges as discussed.
Sampling strategy
In this study, the selection of samples of human remains for δ13C and δ15N measurements was fully dependent on the goals of the radiocarbon dating programme
(see above).
Results
Samples Ua–number (n=2) were measured by IRSM and the precision of the
measurements for both δ13C and δ15N is ± 0.1‰.
VR1959, Sks 4, 9, and 19 have multiple samples on different bones (this study;
Guiry et al. 2015; Umbelino 2006) and present different values (fig. 4.50, table
4.52).
At Vale de Romeiras, the human collagen δ13C values (n=10/no. individuals=7)
range from –20.5‰ to –18.4‰ with a mean value of –19.2 ± 0.7‰. The human
collagen δ15N values (n=9/no. individuals=7) range from +7.7‰ to +10.3‰ with
a mean value of +9.5 ± 0.9‰. The measurement of δ15N was not possible in the
sample of individual 4 (without lab no.) (Umbelino 2006).
One of the two sets of values obtained for individual 9 (Ua–47983) seems to be
an outlier in the series (fig. 4.50). When excluded, the values range from –20.2‰ to
–18.4‰ (–19.1 ± 0.6‰) for carbon and from +8.7‰ to +10.3‰ (+9.7 ± 0.6‰)
for nitrogen.
When considering the 14C dated individuals (n=2) the values range from
–20.5‰ to –20.2‰ (–20.4 ± 0.2‰) for carbon and from +7.7‰ to +8.7‰ (+8.2
± 0.7‰) for nitrogen.
302
Results: Sado valley, Vale de Romeiras
16
15
14
13
δ15N (‰)
12
11
10
9
8
7
6
-22
-21
-20
-19
-18
-17
-16
-15
-14
-13
-12
δ13C (‰)
Figure 4.50. Vale de Romeiras, Sado valley: correlation between the δ13C (‰) and δ15N (‰)
values of 9 samples of human bone collagen from 7 individuals. Individuals 9 (square) and
19 (triangle) have two sets of values.
303
♂
♂
♀
n/d
♀
♂
♂
♀
Adult
Adult
Adult, 20–35 yrs
Adult, 20–35 yrs
Adult
Adult, young
Adult
Adult
Adult, young
Adult
9, 1959
9, 1959
19, 1959
19, 1959
5, 1959
1 (21?), 19593
4, 1959
4, 1959
11, 1959
23?, 1959
Long bone
999.24.61
Long bone
999.22.8
Radius–R
999.26.10
Long bone
999.26.68
Tibia–R
999.29.9
Rib
999.29.53
Long bone
999.21.21
Long bone
999.10.12
Long bone
999.25.65
Long bone
Bone
7731.3
7730.3
n/p
6904.3
6901.3
7735.3
7728.2
Ua–46972
7726.3
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
7640±55
analysis for isotopes
–18.6
–18.7
–18.4
–19.0
–19.1
–19.3
–19.8
–20.2
–18.7
Measurements
14C by AMS; δ13C and δ15N by IRMS
14C Age BP
Lab no.
δ13C (‰)
VPDB
Ua–47983
6625±51
–20.5
10.1
10.1
–
10.0
10.0
10.3
9.1
8.7
9.6
δ15N (‰)
AIR
7.7
3.3
3.3
n/a
3.3
3.3
3.5
3.5
2.8
3.3
2.9
C:N
28
27
30
24
23
21
16
12
27
10
M1
18
16
20
13
11
9
3
0
16
0
M2
% marine (±10)
Guiry et al. 2105
Guiry et al. 2105
Umbelino 2006
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
This study
Guiry et al. 2015
This study
Reference
3 E. Guiry and collegues (2015) were in doubt about the identification of this adult female: 1 or 21. According to my analysis, this is likely to be individual 1, because
the skeletal elements of individual 21 are very fragmented and sex identification would not be possible.
n/d
♀
Sex
Age
Ind.
Identification of the human remains
Table 4.52. Vale de Romeiras, Sado valley: carbon (δ13C) and nitrogen (δ15N) measurements on human bone collagen samples collected at MNA. Samples are sorted so
that the lowest δ13C values are on top of the table. Multiple samples of the same individual are sorted together.
Results: Sado valley, Vale de Romeiras
Archaeothanatology
Source material and analysis strategy
Archaeothanatological analysis was carried out on a total of 19 individuals, c 83%
of the MNI (23 in site plan, 26 estimated by Cunha and Umbelino 1995–97) for
Vale de Romeiras. One of these 19 individuals (VR1959, Sk8) revealed a recent 14C
date and its analysis will be considered separately, thus the sample size for the
general observations is 18. The completeness and detail of each analysis is dependent on the character and quality of the available source material which in several
cases is very limited. At Vale de Romeiras, the preservation of the human remains
is relatively poor, in particular at the level of the upper body. The source material
comprises deteriorated field photographs, available for 14 individuals, good individual drawings and a site plan of the burials (table 4.53).
Results of the archaeothanatological analysis
The nature of the deposits
Analysis of the nature of the deposits at Vale de Romeiras suggests a pattern of
primary burials (table 4.54) indicating that the cadavers were placed in the burial
features before the degradation of the labile joints. In several cases the diagnostic
criteria are poorly preserved, however, the overall position of the skeletal elements
strongly indicates that the bodies were disposed while still retaining their general
anatomical integrity, supporting the interpretation of burials in primary position.
One possible exception to this pattern is presented by two disarticulated long
bones lying on the left side of VR1959, Sk9, which, as discussed further, may suggest a deposit in a secondary position (see below, Post-depositional manipulations of the
cadavers).
305
n/a
n/a
22, 1959
23, 1959
n/a
n/a
yes
yes
yes
yes
yes
yes
n/a
n/a
n/a
n/a
n/a
n/a
n/a
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
yes
yes
yes
yes
yes
9, 1959
10, 1959
11, 1959
12, 1959
13, 1959
14, 1959
15, 1959
16, 1959
17, 1959
18, 1959
19, 1959
20, 1959
21, 1959
yes
yes
yes
yes
yes
n/a
n/a
n/a
n/a
n/a
4, 1959
5, 1959
6, 1959
7, 1959
8, 1959
1, 1959
2, 1959
3, 1959
Source material
Block
Graphic documentation w/human remains
Drawings Site plan
Field
photographs
n/a
yes
yes
yes
n/a
yes
yes
yes
n/a
yes
yes
yes
Individual
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
Written documentation
Poor preservation and unclear documentation. Limited analysis.
Very fragmented material. Archaeothanatology is not possible.
Very fragmented material. Includes fragments of cranium, long bones and a few disarticulated
hand bones. Archaeothanatology is not possible.
–
Poor preservation and unclear documentation. Archaeothanatology is extremely limited.
–
–
Modern date with a calibrated date range of 1445–1632 cal CE (95% confidence).
Archaeothanatology is presented separately in last section.
–
–
–
–
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
Very fragmented material. Archaeothanatology is not possible.
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
Very fragmented material. Includes small fragments of cranium and long bones.
Archaeothanatology is not possible.
Poor preservation and unclear documentation. Limited analysis.
Poor preservation and unclear documentation. Limited analysis.
Observations
Table 4.53. Vale de Romeiras, Sado valley: all individuals identified and excavated (n=23) and the respective source material used for the archaeothanatological analysis.
This material is curated at the MNA.
Sex
♀
♂
n/d
n/d
♀
♂
♂
♀
n/d
n/d
n/d
Age
Adult, 20–35 yrs
Adult
Adult
Adult
Adult
Adult
Adult, 20–35 yrs
Adult, 20–35 yrs
Adult
Adult
Adult
Individual
1, 1959
4, 1959
5, 1959
6, 1959
7, 1959
9, 1959
10, 1959
11, 1959
12, 1959
13, 1959
14, 1959
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Probably primary
Primary
Primary
Nature of the deposit
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Maintenance of labile articulations: bones of right hand.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of right foot.
Maintenance of the general anatomical integrity of the body.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Maintenance of labile articulations: bones of right hand; metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Maintenance of labile articulations: metatarsals and phalanges of right foot.
Maintenance of the general anatomical integrity of the body.
Arguments
Table 4.54. Summary of the nature of the deposits with human remains at Vale de Romeiras, Sado valley.
continued...
n/d
n/d
n/d
♀
n/d
♂
n/d
Non-adult, 12 yrs ± 36 m.
Non-adult, 3 yrs ± 12 m.
Non-adult, 2–6 yrs
Adult, 20–35 yrs
Adult
Adult
Adult
16, 1959
17, 1959
18, 1959
19, 1959
20, 1959
22, 1959
23, 1959
Table 4.54. continued
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of labile articulations: moderate collapse of bones of right hand.
Maintenance of the general anatomical integrity of the body
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body suggests a burial in primary position.
Results: Sado valley, Vale de Romeiras
The space of decomposition of the cadaver
Analysis of the space of decomposition of the cadavers buried in Vale de Romeiras
indicates a general pattern of filled spaces (table 4.55). The cadavers were covered
with sediment immediately after their placement in the feature. The decomposition
of the bodies took place in this sediment-filled environment. In several cases the
analysis is unclear due to the poor preservation of the remains. In every case, the
movement of the bones is limited suggesting that the space of decomposition was
filled and that the sediment could penetrate immediately.
Few cases suggest mixed spaces of decomposition (VR1959, Sks 5, 6) with
original empty spaces in the grave feature and/or by the formation of secondary
empty spaces outside the volume of the cadaver within an overall filled environment. The analysis of VR1959, Sk5 suggests a mixed space of decomposition by
the formation of a secondary empty space outside the volume of the cadaver
within an overall filled environment but there is no clear indication of original
empty spaces in the feature. Unfortunately, the analysis is constrained by the poor
preservation of the material however it is possible to advance some observations
(fig. 4.51).
Figure 4.51. VR1959, Sk5. Although poorly preserved this case is suggestive of a mixed
space of decomposition. The collapse forwards and downwards of the skull (cranium and
mandible), right shoulder girdle, and bones of the right arm and forearm, contrasts with the
position of the pelvic girdle which stayed behind, probably remaining close to the original
burial position. These movements suggest the formation of a secondary space in front of
the upper body where the bones could migrate during the process of decomposition while
being supported by the penetration of sediment in a filled environment. (MNA, 1959, D.
Sousa)
309
Chapter 4
The mixed space of decomposition is suggested by the significant collapse forwards and downwards of the skull (cranium and mandible), right shoulder girdle,
and bones of the right arm and forearm. This movement contrasts with the position of the pelvic girdle which seems to have stayed behind, remaining probably
close to its original burial position. This contrasting movement could be explained
if the upper body was rotated forwards and suspended over what would become a
void, by the liberation of space during decomposition of perishable material placed
in front of the body.
While this secondary empty space was formed the skeletal elements of the
upper body migrated forwards and downwards regaining equilibrium in these new
empty spaces. Also, the skull was likely the most elevated element in the feature
and the closest to the surface of the grave. In this case, the pressure of the sediment would have had an impact on its collapse along with other skeletal elements
of the upper body. While these movements suggest the formation of a pocket of
empty space in front of the upper part of the body, other elements suggest that
decomposition took place in a filled environment. Most diagnostic criteria are not
preserved, but the mandible follows the movement of the cranium suggesting that
the labile temporomandibular joint was maintained. Also, the skeletal elements
involved in this migration forwards and downwards, remained in order, suggesting
continuous support by the penetration of sediment preventing further displacements.
The analysis of VR1959, Sk6 is also limited due to unclear documentation
(fig. 4.52). Nevertheless, some patterns in this feature suggest localized movements
in a mixed space of decomposition. Overall, the skeleton is well articulated and the
movement of the bones is limited. The exception is noted at the level of the upper
left limb which is disarticulated upwards suggesting the existence of perishable
material, functioning as a cushion, behind this area creating a secondary empty
space while it decayed to where the skeletal elements could migrate, in a similar
pattern as observed in the feature ARA1961, Sk6.
310
Results: Sado valley, Vale de Romeiras
Figure 4.52. VR1959, Sk6. The striking displacement upwards of the upper left limb contrasts with the minor movements observed in the feature. The documentation is limited but
the disarticulation at the level of the shoulder joint and consequent movement presents
similarities to the pattern observed at ARA1961, Sk6. (Top: MNA, APMH, F1416. Bottom:
MNA, 1959, D. Sousa)
311
♀
♂
n/d
n/d
♀
♂
♂
♀
Adult, 20–35 yrs
Adult
Adult
Adult
Adult
Adult
Adult, 20–35 yrs
Adult, 20–35 yrs
Adult
1, 1959
4, 1959
5, 1959
6, 1959
7, 1959
9, 1959
10, 1959
11, 1959
12, 1959
n/d
Sex
Age
Individual
Filled
Filled
Filled
Filled
Filled
Filled. Mixed?
Mixed?
Filled
Filled
Space of decomposition
Poor preservation but no movement outside initial volume of the cadaver.
continued...
Maintenance of bones in original unbalanced position: the right patella has slid slightly downwards, but is maintained in a suspended position between the femur and tibia. No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: the bones of right hand are rotated and flexed at the wrist.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: moderate collapse of pelvic girdle; the bones of the feet
are slightly rotated inwards. Also, the cranium is maintained in a semi-suspended position.
No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: moderate collapse of the right hemithorax as well as the
right hip bone; the right patella is suspended on the distal end of the femur. Also, the cranium is maintained in a
semi-suspended position. No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: the patellae are suspended on the distal ends of the
femora. Movement at the level of the upper left limb suggests empty space behind this area.
Poor preservation and unclear documentation.
However, significant collapses suggest pocket of empty space in front of upper body.
Maintenance of bones in original unbalanced position: the bones of right hand are slightly rotated inwards; the
patellae are suspended on the distal ends of the femora; the bones of the feet. Also, the cranium is maintained in a
semi-suspended position. No movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: the bones of right foot are rotated inwards.
No movement outside initial volume of the cadaver.
Arguments
Table 4.55. Summary of the space of decomposition of the human remains in the grave features at Vale de Romeiras, Sado valley.
n/d
♀
n/d
Adult
Non-adult, 12 yrs±36 m.
Non-adult, 3 yrs±12 m.
Non-adult, 2–6 yrs
Adult, 20–35 yrs
Adult
Adult
Adult
14, 1959
16, 1959
17, 1959
18, 1959
19, 1959
20, 1959
22, 1959
23, 1959
n/d
♂
n/d
n/d
n/d
n/d
Adult
13, 1959
Table 4.55. continued
Filled
Filled
Probably filled
Filled
Filled
Filled
Filled
Filled
Probably filled
Poor preservation but no movement outside initial volume of the cadaver. Also, the cranium is collapsed forwards but is maintained in a semi-suspended position.
Poor preservation but no movement outside initial volume of the cadaver. Also, the cranium is collapsed forwards but maintained in a semi-suspended position.
Poor preservation and unclear documentation but no movement outside initial volume of the cadaver.
Maintenance of bones in original unbalanced position: the bones of the right hand are rotated inwards but collapse is moderate. Also, the cranium is maintained in a semi-suspended position.
Poor preservation but no movement outside initial volume of the cadaver.
Poor preservation but no movement outside initial volume of the cadaver.
Poor preservation but no movement outside initial volume of the cadaver. Also, the cranium is maintained in a
semi-suspended position.
Poor preservation but no movement outside initial volume of the cadaver. Also, the cranium is maintained in a
semi-suspended position.
Poor preservation but no movement outside initial volume of the cadaver. Also, the cranium is maintained in a
semi-suspended position.
Poor preservation and unclear documentation but no movement outside initial volume of the cadaver.
Chapter 4
The initial position of the cadaver in the feature
The most common position at Vale de Romeiras consists of a cadaver lying on the
back (n=15 to 16) with the lower limbs in flexion or hyperflexion (table 4.56). In
two cases the body was lying on one side, always on the left, with the limbs in flexion and arranged close to the body. Some of the bodies lying on the back were
generally rotated to the left side (VR1959, Sks 4, 9, 10).
The position of the upper limbs varied considerably but they were always placed
close to the body, symmetrically or asymmetrically, in extension, semi-flexion,
flexion or hyperflexion at the elbows.
At Vale de Romeiras, the bodies were placed in the feature with the lower limbs
in flexion or hyperflexion at the level of the hip and knee joints (n=14). In nine
cases the hip and knee joints were in hyperflexion (VR1959, Sks 1, 6, 7, 11, 13, 16,
17, 22, 23), and in flexion in three to four cases (VR1959, Sks 2, 10, probably 14,
18). In one case (VR1959, Sk20) the hip joints were flexed while the knees were in
hyperflexion. In three cases (VR1959, Sks 9, 12, 19) the hip joints were in semiflexion, always with the knees in hyperflexion, directed forwards (Sk9), downwards
(Sk19) or forwards with a tendency to the left side (Sk12). In one case (VR1959,
Sk5) it was not possible to identify the position of the lower limbs. The rotation of
the lower limbs was variable. In six cases (VR1959, Sks 1, 7, 13, 16, 17, 23) the
lower limbs were in adduction in front of the trunk. The rotation of the limbs to
the left side was also common (VR1959, Sks 2, 6, 10, 20) even if just as a tendency
(VR1959, Sks 9, 12). VR1959, Sk22 combines these two rotations and the lower
limbs were placed in front of the trunk with a tendency towards the left side. The
rotation to the right was less common (VR1959, Sks 11, 18, 19). In two cases it
was not possible to observe the rotation of the lower limbs (VR1959, Sks 5, 14).
Clumping of the body was common, and in most cases the analysis suggests
that the limbs were somewhat nested on the upper body and the feet were forced
towards the buttocks. The pattern was enhanced during decomposition by the
progressive closure of the angles between the segments and several skeletons became hypercontracted (VR1959, Sks 1, 6, 7, 13, 16, 17, 22, 23). The nesting of the
lower limbs could explain the general poor preservation of the bones at the level of
the upper body, since the placement of the thighs in this region would add a mass
of soft tissue increasing the decomposition activity in the area while contributing
to a lower preservation of the bones.
314
Sex
♀
♂
n/d
n/d
♀
♂
♂
♀
Age
Adult, 20–35 yrs
Adult
Adult
Adult
Adult
Adult
Adult, 20–35 yrs
Adult, 20–35 yrs
1, 1959
4, 1959
5, 1959
6, 1959
7, 1959
9, 1959
10, 1959
11, 1959
Individual
Raised
Rotation–L and downwards
(neck flexed forwards)
Raised
Rotation downwards
(neck flexed forwards)
Raised
Rotation–L and downwards
(neck flexed forwards)
Raised
Rotated–L and downwards
(neck flexed forwards)
On the left
Prob. rotated downwards
(neck flexed forwards)
Rotation downwards
(neck flexed forwards)–tendency L
Initial position of the cadaver in the feature
Head
n/a
On the back
On the back
On the back
On the back
On the left
Prob. on the back
On the back
Trunk
On the back
In flexion
In flexion
In flexion–R
In hyperflexion–L
In semi–flexion (c 90°)–R
In semi–flexion–L
In flexion–R
n/a–L
In semi–flexion–R
Prob. in extension–L
In semi–flexion–R
In flexion–L
Upper limbs
In flexion
continued…
Hip and knee joints: flexion
Rotation–R
Feet towards buttocks. Clumping
Hip and knee joints: flexion
Rotation–L
Feet towards buttocks. Clumping
Hip joints: semi–flexion
Knee joints: hyperflexion
Rotation–tendency L
Feet towards buttocks. Clumping
Hip and knee joints: hyperflexion
Rotation–tendency L
In adduction in front of trunk
Feet towards buttocks. Clumping
Hip and knee joints: hyperflexion
Rotation–L
Feet towards buttocks. Clumping
n/a
Hip and knee joints: flexion
Rotation–L
Feet towards buttocks
Lower limbs
Hip and knee joints: hyperflexion
In adduction in front of trunk
Feet towards buttocks. Clumping
Table 4.56. Summary of the reconstruction of the initial position of the cadaver in the grave features at Vale de Romeiras, Sado valley. The initial position of the limbs is
indicated as –R or –L and refers to the right or left limb, when their positions are different. Rotation –R or –L indicates the rotation of the skeletal element(s) towards the
right or the left side.
n/d
n/d
n/d
n/d
♀
n/d
♂
n/d
Adult
Adult
Non-adult, 12 yrs±36 m. n/d
n/d
Adult
Non-adult, 3 yrs±12 m.
Non-adult, 2–6 yrs
Adult, 20–35 yrs
Adult
Adult
Adult
12, 1959
13, 1959
14, 1959
16, 1959
17, 1959
18, 1959
19, 1959
20, 1959
22, 1959
23, 1959
Table 4.56. continued
Rotation–L and downwards
(neck flexed forwards)
Rotation downwards
(neck flexed forwards)
Rotated to the left
Raised
Rotation downwards
(neck flexed forwards)–prob. tendency L
Rotation–R and slightly downwards
Rotation downwards
(neck flexed forwards)– tendency R
Prob. raised
Rotation–L and downwards
(neck flexed forwards)
On the left
n/a
Rotation–L
On the back
On the back
On the back
On the back
On the back
On the back
On the left
On the back
On the back
On the back
In semi–flexion
In semi–flexion–R
n/a–L
In semi–flexion
In extension
In extension?–R
In semi–flexion–L
n/a
In flexion–R
n/a–L
In semi–flexion (c 90°)–R
In extension–L
In extension?
Prob. extension–R
Extension–L
Hip and knee joints: hyperflexion
In adduction in front of trunk. Clumping
Hip joints: flexion
Knee joints: hyperflexion
In adduction in front of trunk
Rotation–L
Feet towards buttocks. Clumping
Hip joints: flexion
Knee joints: hyperflexion
Rotation–L
Hip joints: semi–flexed
Knee joints: hyperflexed
(knees directed downwards)
Feet towards buttocks
Hip and knee joints: flexion
Rotation–R
Feet towards buttocks
Hip and knee joints: hyperflexion
In adduction in front of trunk
Feet towards buttocks. Clumping
Hip and knee joints: hyperflexion
In adduction in front of trunk
Feet towards buttocks. Clumping
Hip and knee joints: prob. flexion
Hip joint–R: semi–flexed
Knee joint–R: hyperflexed
(knee directed forwards or downwards?)
Foot towards buttocks
n/a–L
Rotation–tendency L
Hip and knee joints: hyperflexion
In adduction in front of trunk
Feet towards buttocks. Clumping
Results: Sado valley, Vale de Romeiras
The grave features
The size and shape of the deposits with human remains at Vale de Romeiras were
not described by the excavators. Archaeothanatological analysis suggests that the
bodies were placed in pits dug into the ground. The lateral pressures and walleffects observed on the skeleton in the feature indicate that the shape and size of
the pit was just sufficient to contain the body (table 4.57). The construction of
these simple features seems to be dependent on the planned burial position.
Furthermore, the analysis indicates that in most cases the design of the grave had a
significant impact on the final arrangement of the body in the feature, visible by
the constrained positions of most skeletons. At Vale de Romeiras, the compressing
effect observed on most skeletons can be explained by the narrow shape of the
graves. While the pit imposed a significant physical limit to the cadaver, it also
provided lateral support preventing the collapse of the skeletal elements. It is possible that in some cases there might have been an extra support offered by a tight
wrapping at the time of the burial. However, the analysis did not reveal any clear
evidence for wrapping of the body that could not be explained by the impact of
the limits of the grave and the rapid penetration of fine sediment. Thus, the
analysis suggests that the forced positions observed at Vale de Romeiras could be
sustained by the limits of the grave features without an extra support such as tight
wrappings. Although the use of wrappings is not explicit in the analysis, it is clear
that this kind of additional support would facilitate the neat placement of the cadaver in very small and narrow features, such as those observed at Vale de
Romeiras.
The bottom of the graves tended to be irregular and sloping from the upper to
the lower end of the feature. This is visible in some of the drawings in profile and
confirmed by the raised position of the head in relation to the trunk (VR1959, Sks
7, 9, 10, 11, prob. 14, 19) and/or by the common flexion and collapse forwards
and downwards of the head (VR1959, Sks 4, prob. 5, 9, 11, 14, 17, 19, 22, 23). The
fragmented state of the material does not allow detailed observations at the level of
ribs and sacrum the rotations of which during the process of decomposition could
clarify aspects of the original position of the body in the feature, in relation to the
characteristics of the floor of the grave, such as in the case of bodies placed in a
half-sitting position. Also, it cannot be excluded that the elevation of the head
could be aided by some kind of support, other than the floor of the pit, but the
documentation is unclear.
317
Sex
♀
♂
n/d
n/d
♀
♂
♂
♀
Age
Adult, 20–35 yrs
Adult
Adult
Adult
Adult
Adult
Adult, 20–35 yrs
Adult, 20–35 yrs
Individual
1, 1959
4, 1959
5, 1959
6, 1959
7, 1959
9, 1959
10, 1959
11, 1959
Arguments
1. Wall-effect, upper left side;
2. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, upper right side;
2. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, left side;
2. Overall strong lateral pressure and compressing effect.
Design of the grave feature.
1. Wall-effect, right side;
2. Wall-effect, left side and distal end;
3. Overall strong lateral pressure and compressing effect.
Design of the grave feature.
1. Wall-effect, upper right side;
2. Wall-effect, lower end.
Design of the grave feature.
continued...
1. Alignment of upper right limb and projection forwards and upwards of
shoulder and arm;
2. Alignment and clumping of lower limbs including feet;
3. Hypercontracted skeleton.
1. Alignment of upper right limb and projection forwards of shoulder and arm;
2. Alignment and clumping of lower limbs including feet. This wall-effect corresponds to the physical limits of the pit and was accentuated by the slope characteristics of the bottom, lower at the level of the pelvic girdle, enhancing the
nesting of the lower body.
1. Alignment of upper right limb and projection forwards of shoulder and arm;
2. Alignment and clumping of lower limbs including feet. This wall-effect corresponds to the physical limits of the pit and was accentuated by the slope characteristics of the bottom, lower at the level of the pelvic girdle, enhancing the
nesting of the lower body.
1. Alignment of upper left limb and projection forwards of shoulder and arm,
accentuated by the transfer of body weight from the right towards the left side;
2. Alignment and clumping of lower limbs including feet. This wall-effect corresponds to the physical limits of the pit and was accentuated by the slope characteristics of the bottom, lower at the level of the pelvic girdle, enhancing the
nesting of the lower body.
1. Alignment and clumping of lower limbs including feet;
2. Hypercontracted skeleton.
1. Wall-effect, upper right side;
1. Alignment of upper right limb, involving the arm;
2. Overall strong lateral pressure and compressing effect. 2. Contraction and clumping of lower limbs which were maintained on an
Design of the grave feature.
extreme constrained position;
3. Hypercontracted skeleton.
1. Wall-effect, upper left side;
1. Alignment of the upper left limb, involving shoulder and arm, accentuated by
2. Wall-effect, lower end.
the sloping floor causing a transfer of body weight towards the left side;
Design of the grave feature.
2. Alignment of lower limbs and constraint effect towards lower limbs forcing
feet towards buttocks.
n/a
n/a
Pressures
Table 4.57. Summary of the key observations used in the reconstruction of the grave features at Vale de Romeiras, Sado valley. Observed pressures on the body are presented
in a number sequence. The arguments supporting these observations are indicated by the same number in the respective column.
n/d
n/d
n/d
♀
n/d
♂
n/d
Adult
Adult
Non-adult, 12
yrs±36 m.
Non-adult, 3
yrs±12 m.
Non-adult,2–6 yrs
Adult, 20–35 yrs
Adult
Adult
Adult
13, 1959
14, 1959
16, 1959
17, 1959
8, 1959
19, 1959
20, 1959
22, 1959
23, 1959
n/d
n/d
n/d
Adult
12, 1959
Table 4.57. continued
1. Alignment and clumping of lower right limb.
1. Alignment of the upper left limb, involving at least the arm;
2. Projection forwards of left shoulder, possibly accentuated by slope characteristics of the bottom, slightly more elevated on the superior left side;
3. Suggested by clumping of lower limbs.
1. Suggested by clumping of lower limbs.
–
1. Suggested by the alignment of upper right limb, involving arm and possibly
shoulder;
2. Suggested by the nesting of the lower limbs.
1. Overall strong lateral pressure and compressing effect. 1. Constraining the movement of the bones maintaining these clumped together;
Design of the grave feature.
2. Hypercontracted skeleton.
1. Wall-effect, left side and lower end;
1. Alignment and clumping of lower limbs;
2. Overall strong lateral pressure and compressing effect. 2. Hypercontracted skeleton.
Design of the grave feature.
1. Prob. wall-effect, left side and distal end.
Poor preservation and unclear documentation.
The observations do not indicate any main pressures
towards the body in the feature.
Poor preservation.
1. Probably wall-effect, upper right side;
2. Probably wall-effect, lower end.
Design of the grave feature.
1. Overall strong lateral pressure and compressing effect. 1. Constraining the movement of the bones maintaining these clumped together;
Design of the grave feature.
2. Hypercontracted skeleton.
1. Wall-effect, left side.
1. Affecting bones of lower limbs in particular, visible by alignment and
2. Overall strong lateral pressure and compressing effect. clumping of legs;
Design of the grave feature.
2. Hypercontracted skeleton.
1. Wall-effect, right side;
2. Wall-effect, left side;
3. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, left side;
1. Alignment of upper left limb, involving at least the arm;
2. Wall-effect, lower end
2. Alignment and clumping of lower limbs;
3. Overall strong lateral pressure and compressing effect. 3. Hypercontracted skeleton.
Design of the grave feature.
1. Wall-effect, lower end.
Poor preservation.
Chapter 4
Deposits containing more than one individual
At Vale de Romeiras the deposits with human remains contain the remains of one
individual. In the site plan (1959), the deposits with individuals 2 and 3, 9 and 12,
18 and 23 seem to be somehow related, either by being buried simultaneously
(18 and 23) or by being disturbed by a later inhumation (2 and 3; 9 after 12). However, the analysis does not support any of these hypotheses.
In the site plan (fig.4.45), the features containing the remains of VR1959, Sks 2,
3 are in very close proximity and seem to overlap. In both features, the human
remains are poorly preserved and an archaeothanatological analysis is not possible
(table 4.53). Stratigraphically, the grave feature containing individual 3 (depth: 37
cm, layer 6: soft bedrock) seems to antedate the deposit with individual 2 which
was found in an upper layer (depth: 15 cm, layer 2: grey soil with few shells). One
hypothesis is that the disposal of VR1959, Sk2 had disturbed the grave of VR1959,
Sk3, in a deeper layer. However, this does not seem to be the case when looking at
the site profile, because grave 2 does not intersect with grave 3 (fig. 4.53).
Figure 4.53. Profile A–A', Vale de Romeiras, Sado valley. The close proximity and possible
overlapping of the remains of VR1959, Sks 2, 3, as suggested by the observation of the site
plan, is not confirmed in the site profile. The relation of these two burials remains ambiguous. Roman numerals: human remains. Letters: animal bones. (Adapted from a photograph by
J. P. Ruas. MNA, 1959, D. Sousa, A120-detail)
Another hypothesis could be suggested by the relative completeness of VR1959,
Sk3 in comparison to the sparse remains in the feature with VR1959, Sk2. This
could indicate an older date for the latter, suggested by its disturbance at the time
of the burial of VR1959, Sk3. However, the fragmentary state of VR1959, Sk2
could also be explained by its proximity to the top soil. Unfortunately, despite the
suggestive reference in the site plan, the documentation is unclear and it is not
possible to test any of these hypotheses.
This is also the case for the features VR1959, Sks 18, 23. In the site plan they
seem to be placed in close proximity, suggesting a common arrangement. However, in the site profile, it is clear that is not the case. Stratigraphically, VR1959,
320
Results: Sado valley, Vale de Romeiras
Sk23 was buried first. It was found at a depth of 84 cm at the time of the excavation (layer 6: soft bedrock) while VR1959, Sk18 was found at a depth of 54 cm
(layer 4: yellowish soil), suggesting a later deposit (fig. 4.54).
Figure 4.54. Profile B–B', Vale de Romeiras, Sado valley. The close proximity and possible
overlapping of the remains of VR1959, Sks 18, 23 as suggested by the observation of the
site plan is not confirmed in the site profile. Roman numerals: human remains. Letters: animal
bones. (Adapted from a photograph by J. P. Ruas. MNA, 1959, D. Sousa, A120-detail)
VR1959, Sks 9, 12 are shown in the site plan as partially overlapping, which in this
case, is also confirmed in the site profile (fig. 4.55). The graves were found in the
same layer (6: soft bedrock), each containing the remains of an adult male overlapping at the level of the lower body. The analysis suggests that these features do not
constitute a double burial, but separate events in time, possibly involving postdepositional manipulation of skeletal elements, which will be discussed further
below. VR1959, Sk12 was the first to be deposited. At the time of the excavation
this feature was found at a depth of 60 cm (layer 6). The remains are poorly preserved but the skeleton seems complete, articulated and undisturbed, except for
the bones of the left lower limb which are all missing. VR1959, Sk9 was found
lying on a sloping floor in layer 6 with the cranium at a depth of 43 cm and the
pelvic girdle 20 cm lower, at a depth of 62 cm. According to the site plan, the
lower body of VR1959, Sk9 lies in front of the lower body of VR1959, Sk12, at a
depth of c 60 cm. Despite the overlapping, the disposal of VR1959, Sk9 does not
seem to disturb VR1959, Sk12, and both individuals seem to maintain their anatomical integrity when they come in contact with each other, supporting the hypothesis of synchronous multiple deposit. However, the absence of the bones of
the lower left limb of VR1959, Sk12 is a strong argument supporting the hypothesis of separate burial events, because the removal of these elements could only
have happened after decomposition of most soft tissue and before the disposal of
the second individual VR1959, Sk9.
321
Chapter 4
Figure 4.55. Profile E-E', Vale de Romeiras, Sado valley, shows the close proximity and
partial overlapping of the grave features VR1959, Sks 9, 12. Roman numerals: human remains. Letters: animal bones. (Adapted from a photograph by J. P. Ruas. MNA, 1959, D.
Sousa, A120-detail)
Post-depositional manipulations of the cadavers
At Vale de Romeiras, the skeletons are apparently complete and lie undisturbed in
the grave feature. In a few cases there is also evidence suggesting the postdepositional manipulation of the cadavers.
One case is documented by VR1959, Sk12 containing the remains of an adult
male lying on the back with the upper limbs in extension placed laterally to the
trunk, and the lower right limb hyperflexed at the level of the knee, folded backwards with the foot lying behind the buttocks (fig. 4.56). As discussed, despite the
fragmentary state of this material, the skeleton seems to be complete, except for
the bones of the lower left limb which are all absent. The general topography of
the body indicates a burial in primary position, decomposed in a filled environment
as suggested by the limited movement of the bones in the feature. The missing
bones, the left femur, tibia, fibula, and possibly the patella and bones of the left
foot, have been removed probably after the decomposition of the body in the
grave. It seems unlikely that the absence of the whole lower limb could be
explained by taphonomic processes such as preservation and/or animal activity.
Despite the poor preservation of the bones in the feature, most elements are present and recognizable in the documentation, even if in a very fragmentary state,
thus, it is unlikely that the strongest and most dense bones in the body, the femur
and tibia, would decompose completely before all other elements. Also, it seems
unlikely that an animal would completely remove these large bones without disturbing other elements in the feature. Accordingly, this case seems to suggest the
re-opening of the grave and the human removal of the bones of the lower left
limb, probably after decomposition of the soft tissues was advanced or completed.
As discussed, VR1959, Sk12 lies partially under VR1959, Sk9 (fig. 4.57). Interestingly, this latter feature (VR1959, Sk9) presents another case of postdepositional manipulation, which could be possibly be related to the removal of
the lower left limb of the individual lying underneath (VR1959, Sk12). VR1959,
Sk9 was lying on the back with the lower limbs in hyperflexion and rotated upwards (fig. 4.58). The overall arrangement of the bones indicates a general rotation
towards the left side of the feature. The graphic documentation shows two long
bones lying on the left side of VR1959, Sk9, in which direction the individual is
rotated, suggesting a common arrangement. One of these bones is a human femur.
322
Results: Sado valley, Vale de Romeiras
The other element is unclear but could be a tibia. These bones were found in the
upper layer (2, greyish sand) and very close to VR1959, Sk9. The presence of these
bones lying on the side of VR1959, Sk9 coincide with the missing long bones in
VR1959, Sk12. Although it is not possible to be completely sure if these disarticulated elements belong to individual 12, it is suggestive that these have been removed and later (impossible to determine precisely when), individual 9 was buried,
partially on top of 12, and the removed bones then were placed on the side of the
new burial. This is a probable scenario; however, the correlation of the events is
not possible to confirm.
Figure 4.56. VR1959, Sk12 contains the human remains of an adult male. Despite the poor
preservation of the bone material, the skeleton seems complete, articulated and undisturbed, except for the bones of the lower left limb which are all missing, possibly due to
the post-depositional manipulation of the cadaver. (MNA, 1959, D. Sousa)
323
Chapter 4
Figure 4.57. Detail of the site plan of VR1959, Sks 9, 12, showing disarticulated long bones
on the left side of Sk9. (MNA, 1959, D. Sousa, A120-detail)
Figure 4.58. VR1959, Sk9 and disarticulated long bones (one femur, and one tibia?) on his
left side. (MNA, 1959, D. Sousa)
324
Results: Sado valley, Vale de Romeiras
Other cases are less clear and difficult to describe due to the unclear nature of the
documentation. Fragments of unidentifiable disarticulated long bones are found
close to the features VR1959, Sks 4, 10, apparently in the same layer, but association with the burial deposits is not possible to define. Nothing suggests a common
arrangement of the burial feature or a deliberate post-depositional manipulation of
the foreign bones. These disarticulated elements could belong to a disturbed burial,
such as VR1959, Sks 2, 3. Likewise, VR1959, Sk13 contains a few bone elements
found at the level of the lower limbs, possibly foreign to the feature (fig. 4.59).
Figure 4.59. VR1959, Sk13. This feature is poorly preserved possibly due to its proximity to
the modern surface of the site (c 10 cm deep). The cranium is not preserved but a few
fragments of the maxilla are visible in the drawing. A concentration of bone elements are
found at the level of the lower limbs which seem to be foreign to the feature, possibly
indicating a previous burial in the same location rather than the post-depositional manipulation of human bones. (MNA, 1959, D. Sousa)
The deposit is poorly preserved and the analysis is limited. Although none of the
diagnostic criteria are preserved, there is no clear indication of movements outside
of the initial volume of the body, suggesting a primary deposit in a filled space of
decomposition. In this scenario, this concentration of foreign skeletal elements
could be explained by the presence of a previous human burial in the same location, and the possible post-depositional manipulation of bone elements would be
in this case difficult to assert. Furthermore, the cranium of VR1959, Sk13 is not
325
Chapter 4
present, possibly due to preservation. A few fragments are visible from what seems
to be the remains of the maxilla invalidating the hypothesis of intentional removal
of the cranium. The poor preservation of this feature could be explained by surface
disturbance, as this deposit was excavated at a depth of 10 cm from the modern
surface (fig. 4.55).
One other feature in the same area as VR1959, Sk13 seems to suggest the manipulation of cranial elements, however, this hypothesis should be viewed with
caution due to the fragmentary nature of the documentation, and could likewise be
the result of taphonomic causes. This is the case of VR1959, Sk1 whose skull
(cranium and mandible) was not found in the grave feature (fig. 4.60).
Figure 4.60. VR1959, Sk1. The cranium and mandible are not in the feature. (MNA, 1959,
D. Sousa)
The postcranial skeleton is apparently complete, articulated and undisturbed. This
individual was recovered close to the current top layer of the site, c 20 cm deep,
which could explain the disturbance of the feature. The absence of the cranium
and mandible could also be explained by a deliberate act involving the re-opening
of the grave and removal of these elements, sometime after the break down of the
labile cervical joints. It is also possible that the individual was buried without the
head, which could be demonstrated by traces of trauma at the level of the cervical
vertebrae. Unfortunately, all these scenarios remain as working hypotheses, and are
not possible to test with the current documentation.
326
Results: Sado valley, Vale de Romeiras
As discussed throughout this section, at Vale de Romeiras, several burials are
poorly preserved and the possible disturbance of the features, taphonomic or deliberate, is difficult to determine. However, in most cases, rather than the result of
post-depositional manipulations of the cadavers, the disturbance of the graves can
be explained by the original and/or current proximity to the surface of the site,
rendering the skeletons more vulnerable to taphonomic agents. That seems to be
the case of VR1959, Sk2 (depth: 15 cm) which is not just poorly preserved, but
possibly disturbed, and several skeletal elements are missing. The scenario of reopening of the grave, for the deliberate removal of selected bones is not possible
to demonstrate. The reduction of the feature to accommodate new deposits is
another possible explanation. Nevertheless, this does not seem to be the case, because the burials found in the same area (VR1959, Sks 4, 3, 7) are well aligned in a
deeper level at a depth of c 35 cm (Sks 3, 7) and 50 cm (Sk4), and as discussed
above, VR1959, Sk2 seems to be a later interment, thus not disturbed by any of
these deposits, which excludes the possibility of a case of reduction of VR1959,
Sk2. Likewise, VR1959, Sk5 (fig. 4.51), found 18 cm deep, lies in close proximity to
other burials (VR1959, Sks 1, 14) but it is not disturbed by them. Again, its fragmentation and partial preservation can be explained by its close proximity to the
surface of the site as well as by the probable greater decomposition activity, increased by the inclusion of perishable material placed in front of the body, as discussed above. In any case, the possibility of post-depositional manipulation of the
cadaver is not demonstrated.
Other observations
Animal bones
At Vale de Romeiras, several concentrations of animal bones have been found.
These concentrations consist of apparently disarticulated bones, and the nature of
these deposits is unknown. It is possible that some of these elements were deposited or discarded in articulated state and intentionally buried, but the material is
fragmented and the documentation is unclear. Some of the concentrations of animal bones lie in close proximity to the features with human remains.
Only the case of VR1959, Sk1 and the animal group h are recorded in the same
layer (4) and depth (48 cm) (fig. 4.54). It is unclear if these are simultaneous deposits and there is no evidence for a common arrangement of the features. Unfortunately, the documentation is very limited at this level and it is impossible to determine a relationship between the features with human remains and the deposits
with animal bones.
Other burials
VR1959, Sk8 is 14C dated to the early modern period in Europe with a calibrated
date range of 1445–1632 cal CE (95% confidence), and for this reason it was not
included in the general analysis.
327
Chapter 4
The deposit contains the remains of a child of 3 to 4 yrs ± 12 months in primary position, as suggested by the maintenance of the general anatomical integrity
of the body (fig. 4.61). The material is poorly preserved but there is no evidence
for movement outside the initial volume of the cadaver, suggesting a filled space of
decomposition. The child was lying on the left side with the upper limbs in extension placed in front of the body and the lower limbs in flexion at the level of the
knees and rotated to the left. The body was placed in a small pit but wide enough
to contain the body in a natural and unforced position, which has not been observed in any other burial at Vale de Romeiras. Although the material is fragmented the analysis did not indicate any particular pressures on the body in the
feature. An unidentified flat bone lies at the distal end of the left forearm of the
child. The bone does not seem to belong to the skeleton of this individual and it is
also unclear if the bone is human. Unfortunately the limited documentation does
not allow a further observation.
Figure 4.61. VR1959, Sk8. Grave feature with the remains of a 3–4 yrs ± 12 months child
buried in the centre of the burial area in the second layer (grey soil with shells) at a depth of
35 cm from the modern surface. 14C measurements on a tibia revealed a surprisingly modern date with a calibrated date range of 1445–1632 cal CE (95% confidence). (MNA, 1959,
D. Sousa)
328
Results: Sado valley, Vale de Romeiras
The mortuary programme at Vale de Romeiras
At Vale de Romeiras, bone preservation is relatively poor and in several cases the
diagnostic criteria are not preserved and/or are unclear in the documentation.
Nevertheless, the archaeothanatological analysis indicates a strong similarity of
burial practices observed at the site and the minor detected variations do not affect
the overall pattern of the treatment of the dead. The bodies were normally laid in
small pits in contracted positions, with the limbs nested on the upper body, and
with the feet forced towards the buttocks. In several cases this pattern was enhanced during decomposition and the skeletons become hypercontracted.
While the tables presented above may be consulted for the diagnostic criteria
and arguments of analysis, I have chosen to describe VR1959, Sks 11, 7 which I
consider representative of the practices at Vale de Romeiras. VR1959, Sks 19, 23
are also further described, as these burials present variations to the common practice and reveal striking elements from an archaeothanatological perspective. The
selection of the burials for description was constrained to those graves whose remains are relatively well preserved and hold sufficient documentation for consistent argumentation.
VR1959, Sk11
VR1959, Sk11 consisting of the remains of an adult female is particularly well
documented by graphic documentation (fig. 4.62). This case combines several
common elements observed in the burials at Vale de Romeiras. The cadaver was
laid on its back in a small pit with a sloping floor (fig. 4.54), possibly slightly more
elevated at the superior right side of the feature. The head was rotated to the left
and was the most elevated body part in the feature; hence the closest body part to
the surface of the grave. The upper limbs were flexed at the elbows and the hands
were placed in front of the abdomen, with the right hand lying in front and below
the left hand. The lower limbs were flexed at the level of the hips and hyperflexed
at the knees, rotated upwards towards the upper body, with a tendency towards the
right side. The feet were positioned medially in front of each other while placed
towards the buttocks.
There were two main pressures on the body in the feature. One was a lateral
pressure visible in the alignment of the upper left limb, affecting the left clavicle
and humerus. While the left scapula is not visible in the documentation, there is a
clear verticalization of the left clavicle, and projection forwards and upwards of the
left humerus which is rotated inwards lying very close to the thoracic cage. This
alignment is explained by a wall-effect on the left side of the feature, accentuated
by the sloping floor of the grave, slightly more elevated on the superior right side,
causing a transfer of body weight towards the left side. While the cadaver slowly
slid to the left, the right shoulder girdle stayed behind and the right scapula became
exposed between the humerus and thoracic cage.
329
Chapter 4
Figure 4.62. VR1959, Sk11 is representative of a typical grave feature at Vale de Romeiras,
Sado valley. Normally, the cadavers were laid on their backs in small pits with sloping
floors. The upper limbs were placed close to the upper body and the lower limbs were
clumped together, typically flexed at the level of the hips and hyperflexed at the knees,
rotated upwards towards the upper body. The feet were positioned medially and placed
towards the buttocks (MNA, 1959, D. Sousa, B42)
330
Results: Sado valley, Vale de Romeiras
The second main pressure observed is not only a strong lateral pressure exerted on
the body, but also the evidence for significant constraint effects. The lateral pressure and wall-effect at the lower end of the feature is visible in the alignment and
clumping of the lower limbs including both feet. This wall-effect corresponded to
the physical limits of the pit, and its impact was increased by the sloping bottom,
lower at the level of the pelvic girdle, enhancing the nesting of the lower body. The
pressures observed are the result of the design of the grave feature. This contraction, affecting the lower limbs in particular, is a typical trait observed at Vale de
Romeiras. In several cases, this pattern is enhanced during decomposition and the
skeleton becomes hypercontracted.
VR1959, Sk7
VR1959, Sk7 is illustrative of a hypercontracted skeleton. This burial is relatively
well preserved and the documentation, although limited, allows a more in depth
observation of these typical features at Vale de Romeiras (fig. 4.63).
In this case as well, the cadaver was laid on its back in a small pit with a moderate sloping floor. The head is poorly preserved but the analysis suggests that the
head was the most elevated body part in the feature and the closest body part to
the surface of the grave (fig. 4.53).
The right upper limb was placed in semi-flexion at the level of the elbow in c 90
degrees with the forearm laid in front of the trunk. The left upper limb was placed
in semi-flexion with the left hand lying on the pelvic region towards the right thigh.
The lower limbs were hyperflexed at the level of the hips and knees, rotated
upwards, and placed in adduction in front of the trunk, with a slight tendency towards the left side of the feature.
Two main pressures bore down on the body. One was the wall-effect on the
right side of the feature, visible in the alignment and projection forwards of the
right upper limb, involving the right shoulder girdle and the arm. The second main
pressure, is similar to what has been observed in the previously described feature
(VR1959, Sk11), indicating strong lateral pressure on the body, and significant
constraint effects. The lateral pressure and a wall-effect at the lower end and left
side of the feature are visible in the alignment and clumping of the lower limbs
including both feet. This wall-effect corresponds to the physical limits of the pit,
just big enough to contain the individual. This pattern was enhanced during decomposition by the penetration of fluid sediment which progressively closed the
angles between the skeletal segments, causing a contracted position to become a
hypercontracted skeleton. The pressures observed are the result of the grave design, although as discussed above, the possibility of wrapping cannot be excluded,
but the data is not conclusive.
331
Chapter 4
Figure 4.63. VR1959, Sk7 is a typical example of a hypercontracted skeleton commonly
observed in the grave features at Vale de Romeiras, Sado valley. (MNA, 1959, D. Sousa)
VR1959, Sk23
VR1959, Sk23 is another example of a hypercontracted skeleton, but in this case
the burial position is noted as unusual at Vale de Romeiras. Unfortunately, the
feature contains poorly preserved remains and the analysis is limited. However, it is
possible to advance some observations.
The striking element of this feature is presented by the unusual position of the
lower limbs. Like the typical burials at Vale de Romeiras, VR1959, Sk23 was placed
332
Results: Sado valley, Vale de Romeiras
on the back with the upper limbs in hyperflexion rotated towards the upper body.
In this case, however, the limbs were lying apart, at each lateral side, leaving the
trunk completely exposed. The grave feature containing the cadaver seems to have
been smaller than usual. According to the notes in the field drawing, the maximum
length occupied by the skeleton at the time of excavation was c 49 cm and the
maximum width c 32 cm. These atypical features constitute a burial feature that
resembles a square grave, which could be called a square shaped burial (fig. 4.64).
Figure 4.64. VR1959, Sk23 is another example of a hypercontracted skeleton although
placed in an atypical square shaped burial position. (MNA, 1959, D. Sousa)
The pit was possibly slightly more elevated at the superior end of the feature as
noted by the rotation of the head to the right and downwards, towards the body in
the feature. This movement could be explained if the cranium was suspended over
what would become a void by the liberation of space during decomposition in the
thoracic cage. A moderate slope elevation would have had an impact on the
333
Chapter 4
collapse forwards and the pressure of the sediment would have accentuated this
movement.
The thoracic skeleton is only partially preserved. The few ribs observable are
displaced and rotated downwards but their lateralization is not possible to determine. A few bone fragments are found lying in the region of the thoracic cage but
the documentation does not allow further observation.
The bones of the upper limbs are only partially exposed and are covered by the
bones of the lower limbs, particularly at the level of the shoulder girdles and arms.
The arms were lying laterally to the trunk and the forearms were crossed in front
of the abdomen with the right hand placed towards the left hip and the left hand
towards the right hip. The left forearm was placed in front of the right forearm.
The bones of the hands are not visible in the documentation possibly due to poor
preservation. Like the bones of the hands, the bones of the feet are not visible in
the documentation possibly due to poor preservation.
The analysis suggests an overall lateral pressure exerted on the body, and a general wall-effect from all sides of the pit. These pressures are visible in the extremely
constrained and hyperflexed position of the skeleton, which is particularly striking
at the level of the lower limbs. This overall wall-effect corresponds to the physical
limits of the pit and the pressures observed are the result of the almost square
design of the grave.
VR1959, Sk19
VR1959, Sk19, presents the exception to the nesting of the lower limbs on the
upper body. This grave is somewhat eccentric to the main burial area, and is also
the oldest burial known at the site. The feature contains the remains of a poorly
preserved skeleton of an adult female, and presents all general elements of the
burials known at Vale de Romeiras. The cadaver was placed on its back on a sloping floor, with the head rotated forwards and downwards towards the upper body.
The upper limbs were placed close to the body lying in almost complete extension
laterally to the trunk. The lower limbs were placed clumped together with the feet
towards the buttocks as normally observed at this site. However, in this case, the
lower limbs are not rotated towards the upper body, but folded backwards, in such
a way that the knees were directed downwards and the legs would lie under the
thighs while the feet were placed under the buttocks. Interestingly, this unusual
position of the lower limbs is somewhat similar to one case observed at Cabeço
das Amoreiras (CAMS1958, Sk7), and to a certain extent is also similar to VR1959,
Sk12 previously discussed for its missing lower limb (fig. 4.65).
334
Results: Sado valley, Vale de Romeiras
Figure 4.65. This illustration presents the key elements observed in two burials at Vale de
Romeiras (12, 19) and one at Cabeço das Amoreiras (7). The distinctive trait of these burials is presented by the unusual position of the lower limbs, which were in extension at the
hips and hyperflexion at the knees; and while the knees were rotated downwards the feet
were directed upwards and placed towards the buttocks. These burials share most general
characteristics found in other features, such as the placement of the body on its back in a
small pit with a sloping floor. (Drawing by Susanna Berglund)
335
Chapter 4
***
The mortuary programme at Vale de Romeiras is consistent and homogeneous, as
suggested by the archaeothanatological analysis.
The cadavers were typically placed in individual grave features while retaining
their anatomical integrity (individual primary deposits). The feature was then
immediately covered with sediment providing a filled space of decomposition. In
many cases the analysis is limited due to the poor preservation of the archaeological material, but, in most cases the remains provide satisfactory indicators to
support these conclusions.
The cadaver was normally laid on the back, and in just a few cases the body was
placed on the lateral left side. The arrangement of the upper limbs was variable but
they were always placed close to the body. The lower limbs were consistently
placed in forced positions, flexed or hyperflexed at the hips and knees, rotated
towards the trunk or towards the right or left side. The clumping of the body was
common. The limbs were nested on the upper body and the feet always placed
towards the buttocks (fig. 4.66). The floor of the graves tended to be uneven and
sloping. In these cases, the head of the cadaver was often rotated forwards and
downwards towards the body in the feature. The physical limits of the pit had a
significant impact on the final arrangement of the body in the feature and the walleffects and consequent pressures, visible by the contracted and hypercontracted
position of most skeletons can be explained as the result of the design of the grave
feature.
Few cases present deviant elements, but these are exceptional instances of
features with, for example, possible perishable elements inserted behind or in front
of the body, rare post-depositional manipulations of the cadaver, or the atypical
initial position of the cadaver in the grave. Despite the striking and unusual
elements, all these cases follow the common core characteristics of burial practices
observed at Vale de Romeiras.
336
Results: Sado valley, Vale de Romeiras
Figure 4.66. This illustration presents several typical elements observed in the burials at Vale
de Romeiras. The cadavers were normally laid on the back with the limbs nested on the
upper body. The lower limbs were consistently placed in forced positions, flexed or hyperflexed at the hips and knees, rotated towards the trunk or towards the right or left side. The
floor of the graves tended to be uneven and sloping. The head of the cadaver was often
rotated forwards and downwards towards the body. The physical limits of the pit had a
significant impact on the final arrangement of the body in the feature. The wall-effects and
consequent pressures are visible in the contracted and hypercontracted position of most
skeletons and can be explained as the result of the design of the graves. (Drawing by
Susanna Berglund)
337
Sado valley: Cabeço do Pez
Site background: a short history of research
Cabeço do Pez (also known as Quinta de Baixo) is a shell midden site located on
the eastern part of the Sado valley (fig. 2.3). It is situated on the right margin of the
river Sado at 57 m.a.s.l. (Diniz et al. 2012), on a hilltop next to Vale de Romeiras,
with good visibility over the river. Cabeço do Pez is one of the largest sites known
at the Sado valley and it has the highest number of human burials.
Archaeological material and documentation available for Cabeço do Pez are
curated at the National Museum of Archaeology in Lisbon (MNA). The site was
identified in the 1930s (Barradas 1936). According to the documentation it was
excavated only 20 years later in 1956 and 1958–59 by the staff of MNA. The site
plans indicate the excavation of 27 individuals in 1956 and two individuals in 1959.
However, the anthropological analysis indicates the excavation of a minimum
number of individuals between 32 and 36 (Cunha and Umbelino 1995–97) (table
4.58). The burial population consists of 26 adults (seven females, six males, ten
indeterminate), six of which are young adults under 30 years old (Cunha et al.
2002, 191). Among the non-adults (n=6) two are aged between 1 to 2 years old,
and one is under 1 year old (Cunha et al. 2002, 191). Small scale excavations took
place in 1983 (Arnaud 1986) and 2010 (Arias et al. 2015) but no human remains
were found.
Table 4.58. Cabeço do Pez, Sado valley: minimum number of individual (MNI) estimated for the material curated at MNA.
MNI
Adults
♀
7
32–36
♂
6
26
Non-adults
Reference
Indet.
13
6
Cunha and Umbelino 1995–97
The following table (table 4.59), courtesy of one of the anthropologists who studied this material in detail (Umbelino 1999), is the reference material used in this
study regarding sex and age estimates for Cabeço do Pez.
Skeletal preservation at Cabeço do Pez is generally poor. Eight individuals are
preserved or partially preserved in blocks of paraffin wax (CP1956, Sks 2, 9, 10, 11,
17, 21, 27; CP1959, Sk2) while the remaining skeletons are in disarticulated state
sorted in containers.
Graphic documentation is very limited and previous authors have commented
on the lack of field drawings available for this site (Umbelino 2006, 144). The individuals excavated in 1956 are documented in one site plan (MNA, 1956, A. Paiva)
which shows two main areas with human burials (trenches A, B), and the position
of the skeletons represented in a simple manner (fig. 4.67).
338
Results: Sado valley, Cabeço do Pez
Table 4.59. Cabeço do Pez, Sado valley: sex and age estimates of 20 individuals. (Data from Cunha and
Umbelino 1995–97, 2001; Umbelino 1999)
Individual
Container
Age
Sex
1
3
4
5
6
8
13
14
16
19
20
22
24
25
26
27
15(9?)
6258
6256
6257
6249
6247
6243
6250
6241A
6247-16
6255
6247A
6247A
6240
6240
6247A
6234
6242
n/d
♀
♀
♀
n/d
♀
♂
♀?
♂
n/d
n/d
♀?
n/d
♂
n/d
♂
♀ (adult)
A
–
Adult
Adult
Adult, young
Adult, mature
Adult
Adult, young
15–23 yrs
Adult
Adult
Adult
Adult
Adult
Adult
Adult, mature
Adult
Adult
3 individuals:
Adult
Non-adults, 6–9 yrs; 10–11 yrs±30 m.
Adult
♀?
The two individuals excavated in 1959 are documented in individual sketches
(MNA, 1959, D. Sousa) (fig. 4.68). The final drawings are missing from the archives of MNA, but their existence is known by a publication in 1974 (Santos et al.
1974). Other drawings consist of one plan with the topography of the site (MNA,
1956, A. Paiva, A10), a plan with a detail of a pit structure in stone of unknown
chronology (MNA, 1956, A. Paiva, A11), and several sketches from 1958–59
showing excavation trenches and grids, profiles, as well as concentrations of faunal
remains (MNA, D. Sousa). Like the individual skeleton drawings, the final versions
of the site plans from 1958–59 are not in the archives of MNA and their existence
is known only from the same article (Santos et al. 1974).
Archives of photography of Cabeço do Pez consist of 90 photographs from
1957 and 1958 (MNA, APMH, 2/11/61). However, the human remains were excavated in 1956 and 1959 and the archive does not contain photographs from
these field seasons. Several photographs, poorly preserved, were taken in 1950,
labelled under the name of Quinta de Baixo (MNA, APMH, 2/11/60), which is
the name by which Cabeço do Pez used to be known. These early photographs do
not indicate excavation work at the time, and were possibly taken during the first
surveys of the area.
339
Chapter 4
Figure 4.67. Cabeço do Pez, Sado valley: site plan of the areas excavated in 1956. (Adapted
from a photograph by J. P. Ruas. MNA, 1956, A. Paiva)
340
Results: Sado valley, Cabeço do Pez
Figure 4.68. CP1959, Sk1. This sketch is in the archives of MNA but the final version of the
drawing is missing. Its existence is known from a publication in 1974 (Santos et al. 1974).
(MNA, 1959, D. Sousa)
Written documentation consists of diary reports from 1958 and 1959. Unfortunately, several pages are missing from the documents and the reports are incomplete. In the documentation from 1959 (Cabeço do Pez, 23 May to 25 July 1959,
MNA, APMH, 2/3/7/3, pp. 1–28) J. Roldão describes briefly the excavation of
one skeleton in layer 2. It was identified on the 1st of June 1959 but excavated almost two months later (20–24 July 1959). The skeleton was found in the limits of
the trench in one of the baulks of the grid system. The site plan for this year shows
a second area with the human remains of a second individual, but these are not
described in the available documents. The field sketch shows that the skeletal material was poorly preserved and fragmentary. In his notes, J. Roldão highlights the
identification of several areas related to fire activity, which he identifies as ashes or
hearths (cinzeiros and lares). One of these is described as very large, 200 cm x
120 cm, with a depth of 10 to 15 cm (diary entry, 23 June 1959). He also describes
341
Chapter 4
recovering a great quantity of tiny sea snails (Trivia monacha?) almost every day,
most of them concentrated in particular areas. The exact location of the fire activities and the pockets of small sea snails are difficult to reconstruct, and a correlation
with the human remains is not possible to determine.
Stratigraphic documentation of the individuals excavated in 1956 is limited. According to the site profiles, most skeletons were recovered from the sand layer on
top of the bedrock. Not all skeletons are represented in the site profiles and their
stratigraphic location remains unclear. The basal sand layer, where apparently most
skeletons were found, is covered by a relatively thick layer (c 25 cm) described as
dark brown soil with shells, hearths and ashes (terra castanha escura; conchas, lares e
cinzas). This dark layer with fire activity is covered by the top soil. The two
individuals recovered in 1959 were excavated from this dark layer with shells
described as layer 2. CP1959, Sk1 was recovered at a depth of c 80 cm while
CP1959, Sk2, apparently in the same layer, was recovered closer to the surface at
c 18–35 cm deep.
Chronology of the burial activity
Previous 14C measurements
Four 14C measurements are available on non-human samples (table A12). Unfortunately, not all are usable due to the nature of the samples and to the incomplete
published data. This is the case for a measurement made on a conglomerate of
shell and charcoal (Q–2498, Arnaud 1989, 619) which makes it impossible to securely calibrate. Also, a measurement on animal bone (Q–2499, Arnaud 1989),
which unknown δ13C is not possible to securely calibrate with the terrestrial curve
only.
The first two measurements on human bone from Cabeço do Pez (Sac–1558
and Beta–125109) were carried out in the late 1990s (Cunha and Umbelino 2001)
both collected from CP1956, Sk4. However, both δ13C values published along with
the 14C dates were obtained by AMS (Cláudia Umbelino, pers. comm.) which cannot be used for dietary reconstructions or reservoir corrections (Millard 2014, 557;
Taylor and Bar-Yosef 2014, 117), and these 14C measurements have to be excluded
until IRMS-based δ13C values are available.
Objectives of the dating programme
Cabeço do Pez is a large site with at least two main areas of burial activity (trenches
A, B) (fig. 4.67). Before this radiocarbon dating programme, one human skeleton
(CP1956, Sk4) was directly dated but with problematic calibration. Previous measurements on non-human samples (Q–2496 and Q–2497) suggested a relatively
more recent use than most shell middens in the valley. Cabeço das Amoreiras also
presented slightly more recent dates on non-human samples (Q–AM85B2a and
Q–AM85B2b).
342
Results: Sado valley, Cabeço do Pez
The general aims of the dating programme for Cabeço do Pez followed the
macro-goals of the radiocarbon programme of this dissertation, outlined in chapter
3, and were:
• to define when the burial activity started and when it ended;
• to estimate the duration of the use of this site for burial activity;
• to estimate the frequency of the burial activity in the site.
The site-specific aims of the dating programme for Cabeço do Pez were:
• to define the temporal relation between the two main burial areas (areas
A, B);
• to define the temporal relation between the apparently isolated burials
(CP1956, Sks1, 2) and the two main burial areas.
Sampling strategy
Five samples of human bone were selected for 14C analysis (table 4.60). The
sampling strategy was targeted towards a good spatial coverage of the material
excavated in 1956.
Table 4.60. Samples of human bone (5) from Cabeço do Pez collected and submitted for 14C measurement
in the course of this study. All samples are from the 1956 assemblages and were collected at MNA.
Ind. Lab no.
Age
Sex
Bone
Site plan
Observations
2
Ua–46930 Non-adult,
1.5–2 yrs
n/d Long bone
T. B
(isolated)
Previous stable isotope analysis of a
fibula (Fontanals-Coll et al. 2014)
5
Ua–46931 Adult,
35–50 yrs
♀
T. A
–
9
Ua–46932 Adult
n/d Fibula
T. B
Previous stable isotope analysis of a
rib (Fontanals-Coll et al. 2014)
11
Ua–46933 Adult
♂?
Long bone
T. B
–
27
Ua–46934 Adult
♂
Humerus–R T. A
999.89.2
Tibia
999.77.92
Previous stable isotope analysis of a
long bone (Umbelino 2006) and a
humerus (Fontanals-Coll et al. 2014)
Results
Seven 14C measurements on human bone collagen samples from Cabeço do Pez
are available now (table 4.61), two of which (Beta–125109 and Sac–1558, CP1956,
Sk4) have a problematic calibration and were excluded from the present study (see
chapter 3, Reliability of isotopic measurements).
343
Chapter 4
Table 4.61. Human bone collagen samples (5) from Cabeço do Pez, Sado valley: 14C measurements and
estimation of non-terrestrial carbon intake (% marine). Samples are sorted so that the lowest marine percentage is on top of the table. Marine percentage was estimated from the measured δ13C value of each sample
by the application of method marine 1 and method marine 2, based on adopted endpoint values of –12‰
and –20‰ for marine (100%) and terrestrial (100%) diets respectively.
% marine
(±10)
Identification of the Measurements
14C by AMS; δ13C and δ15N by IRMS
human remains
Ind. Excav.
Museum
Lab no.
11
1956
MNA
9
Reference
δ13C (‰)
VPDB
δ15N (‰)
AIR
C:N
M1
M2
Ua–46933 6788±46
–20.7
6.7
3.3
8
0
This study
1956
MNA
Ua–46932 6780±48
–19.8
10.4
3.2
16
2
This study
27
1956
MNA
Ua–46934 6734±51
–19.7
13.5
3.3
17
4
This study
2
1956
MNA
Ua–46930 5579±41
–19.1
11.5
3.6
23
11
This study
5
1956
MNA
Ua–46931 6791±43
–18.3
13
3.5
30
21
This study
14C
Age BP
Calibration and chronological models
The chronological model was built on five 14C dates, all obtained in the scope of
this study (fig. 4.69, table 4.62).
This is a phase model which groups all dated burial events in a phase of burial
activity (Bronk Ramsey 1998, 2000, 2001). This grouping is constrained by a start of
burial activity event, and an end of burial activity event (i.e., boundaries) imposing no
internal constraints to the relative order of the dated events. All it assumes is that
the start event occurs before the dated events, and that these all occur before the
end event. For the command files defined for this model see appendix.
Internal consistency and reliability of the model
Agreement index (A) for each dated event is high (A≥60%). The model shows
high overall agreement (Marine 1: Aoverall=107.4%, Marine 2: Aoverall=100.8%),
which indicates that the 14C dates are in accordance with the prior information
incorporated in the model. The index of convergence (C) is good (≥96%).
The results obtained when running the model with the different marine values
(M1, M2) are statistically insignificant. For this reason, all further observations and
graphic presentation are based on the results obtained with the model run with the
values of M1. The results obtained with M2 are presented in the tables (table 4.62).
The chronological model for Cabeço do Pez, calibrated with M1 values, estimates the start of the burial activity to have been in 6909–5626 cal BCE (95.4%
probability; start of burial activity) (figs. 4.70, 4.71, table 4.63), and its end to have been
in 4471–3170 cal BCE (95.4% probability; end of burial activity) (figs. 4.70, 4.72, table
4.63). The burial activity is estimated to have continued for between 1192 and 1412
344
Results: Sado valley, Cabeço do Pez
years (95% probability: site use for burial activity), or for 1246 and 1359 years (68%
probability) (table 4.63, fig. 4.73). However, the posterior density estimated for the
duration of burial activity at Cabeço do Pez, calculated without CP1956, Sk2,
which date is coeval with Middle Neolithic groups in Portugal (Ua–46930), is
between 0 and 123 years (95.4% probability), or 0 and 58 years (68.2% probability). The
latter chronological model is discussed in chapter 5.
Figure 4.69. Chronological model for the burial activity at Cabeço do Pez, calibrated using
marine 1 diet values. For each of the dates two distributions have been plotted, one in
outline which is the result produced by the scientific evidence alone, and a solid one which
is based on the chronological model used. Calibrated ranges of 95% and 68% probability
are given by the lower and higher square brackets respectively. The large square brackets
(left) and the OxCal keywords define the overall model. Model-definition command files in
appendix, Sado valley, Cabeço do Pez, Model 1, Marine 1.
Figure 4.70. Start and End of burial activity at Cabeço do Pez derived from the chronological model, calibrated with values for marine 1 diet. Posterior density estimated for the start
of the burial activity is 6909–5626 cal BCE (95% probability) or 6028–5658 cal BCE (68%
probability). Posterior density estimated for the end of the burial activity is 4471–3170 cal
BCE (95% probability) or 4430–4038 cal BCE (68% probability). Calibrated ranges of 95% and
68% probability are given by the lower and higher square brackets respectively.
345
Chapter 4
Table 4.62. Human bone collagen samples (5) from Cabeço do Pez, Sado valley: calibrated date ranges
according to two different approaches for the calculation of the marine percentage. Results are presented in
both (a) cal BCE and (b) cal BP in 95.4% confidence range. The calibrated date ranges are probability
distributions derived from scientific dating alone. The posterior density estimates (in italic) are derived from
Bayesian modelling. Samples are sorted so that the most recent dates are at the top of the column.
(a) Cal BCE
Ind.
Lab no.
14C
2
Ua–46930
5579±41
Unmodelled
cal BCE, M1
Calibrated date range
(95% confidence)
4466–4246
27
Ua–46934
6734±51
5713–5496
5705–5510
5716–5532
5
Ua–46931
6791±43
5727–5509
5711–5536
5721–5558
11
Ua–46933
6788±46
5746–5563
5733–5566
5725–5108
9
Ua–46932
6780±48
5733–5543
5726–5552
5733–5566
Modelled
Modelled
cal BP, M1
cal BP, M2
Posterior density estimate
(95% probability)
6410–6215
6436–6278
Age BP
Modelled
Modelled
cal BCE, M1
cal BCE, M2
Posterior density estimate
(95% probability)
4461–4266
4487–4329
(b) Cal BP
Ind.
Lab no.
14C
2
Ua–46930
5579±41
Unmodelled
cal BP, M1
Calibrated date range
(95% confidence)
6415–6195
27
Ua–46934
6734±51
7662–7445
7654–7459
7665–7481
5
Ua–46931
6791±43
7676–7458
7660–7485
7670–7507
11
Ua–46933
6788±46
7695–7512
7682–7515
7674–7057
9
Ua–46932
6780±48
7682–7492
7675–7501
7682–7515
Age BP
Table 4.63. Posterior density estimates for the dates of archaeological events (Start/End) and the duration
of burial activity at Cabeço do Pez. Data derived from model described in fig. 4.69
Modelled
Modelled
cal BCE, M1 cal BCE, M2
Posterior density estimate
(68% probability)
Modelled
Modelled
cal BCE, M1 cal BCE, M2
Posterior density estimate
(95% probability)
Start of burial activity
6028–5658
6020–5656
6909–5626
6843–5624
End of burial activity
4430–4038
4438–4058
4471–3170
4486–3240
Distribution
Span of burial activity,
duration
Overall model
agreement (Aoverall)
346
1246–1359 yrs 1229–1334 yrs
1192–1412 yrs 1178–1383 yrs
107.4%
107.4%
103.1%
103.1%
Results: Sado valley, Cabeço do Pez
Figure 4.71. Start of burial activity at Cabeço do Pez derived from the chronological model,
calibrated with values for marine 1 diet. Posterior density estimated for the start of the
burial activity is 6824–5626 cal BCE (95.1% probability) or 6017–5658 cal BCE (67.7% probability). Calibrated ranges of 95% and 68% probability are given by the lower and higher
square brackets respectively.
Figure 4.72. End of burial activity at Cabeço do Pez derived from the chronological model,
calibrated with values for marine 1 diet. Posterior density estimated for the end of the
burial activity is 4471–3170 cal BCE (95.4% probability) or 4430–4038 cal BCE (68.2% probability). Calibrated ranges of 95% and 68% probability are given by the lower and higher
square brackets respectively.
347
Chapter 4
Figure 4.73. Duration of burial activity at Cabeço do Pez derived from the chronological
model, calibrated with values for marine 1 diet. Posterior density estimated for the duration
of the burial activity is between 1192 and 1412 years (95% probability), or 1246 and 1359 years
(68% probability). Calibrated ranges of 95% and 68% probability are given by the lower and
higher square brackets respectively.
Stable isotopes: carbon and nitrogen
Previous measurements: δ13C and δ15N
In the early 2000s, one sample (CP1956, Sk27) was analysed in the scope of a palaeodietary study (Umbelino 2006). These analyses were done at the McMaster
University in Canada and the precision of the results are ± 0.1‰ for δ13C and
± 0.2‰ for δ15N (Umbelino 2006, 208–209). The C:N ratios were not provided
but the laboratory confirmed the quality of the collagen and of the measurements
(Umbelino 2006, 315).
The δ13C values published for CP1956, Sk4 (Sac–1558 and Beta–125109) along
with its 14C dates (Cunha and Umbelino 2001) were obtained via AMS and not by
IRSM (Cláudia Umbelino, pers. comm.). These samples were not included in
Umbelino’s palaeodietary study (2006), and as previously discussed these should
not be used for dietary reconstructions and/or reservoir corrections.
Recently, two teams sampled this assemblage for stable isotope analysis of carbon and nitrogen (Fontanals-Coll et al. 2014; Guiry et al. 2015). M. Fontanals-Coll
and colleagues (2014) analysed nine samples at Biological Anthropology Department at the Universitat Autònoma de Barcelona, and duplicate measurements were
taken in the laboratories of the Institute of Environmental Science and Technology
at the same university. The analytical error is below 0.2‰ for both δ13C and δ15N.
Two of these samples (individuals 10, 3A) did not yield enough collagen for
analysis (Fontanals-Coll et al. 2014).
348
Results: Sado valley, Cabeço do Pez
E. Guiry and colleagues (2015) obtained 14 reliable measurements (table 4.64)
from a total of 16 samples. This material was measured on a Thermo-Finnigan
Delta Plus XL isotope ratio mass spectrometer coupled via continuous flow to a
Carlo Erba elemental analyzer at the Department of Human Evolution, Max
Planck Institute for Evolutionary Anthropology, Leipzig, Germany. The instrumental error for δ13C and δ15N measurements was ±0.1‰ and ±0.2‰, respectively (Guiry et al. 2015).
All previous measurements were included in this study (n=22) (table 4.64;
fig. 4.74) as long as they fell within the accepted quality ranges as discussed.
Sampling strategy
In this study, the selection of samples of human remains for δ13C and δ15N measurements was fully dependent on the goals of the radiocarbon dating programme
(see above).
Results
The samples Ua–number (n=5) were measured by IRSM and the precision of the
measurements for both δ13C and δ15N is ± 0.1‰.
Individuals 2, 5, 9 and 27 have multiple samples on different bones (this study;
Fontanals-Coll et al. 2014; Guiry et al. 2015; Umbelino 2006) and present different
values (table 4.64, fig. 4.74).
At Cabeço do Pez, the human collagen δ13C values (n=27/no. individuals=21)
range from –20.7‰ to –17.2‰ with a mean value of –19.3 ± 0.7‰. The human
collage δ15N values (n=27/no. individuals=21) range from +6.7‰ to +13.5‰
with a mean value of +9.8 ± 1.5‰ (table 4.64, fig. 4.74).
When considering the 14C dated individuals (n=5) and the Ua– values in the
case of individuals with more than one set of measurements, the values range from
–20.7‰ to –18.3‰ (–19.5 ± 0.9‰) for carbon and from +6.7‰ to +13.5‰
(+11.0 ± 2.7‰) for nitrogen.
When considering the 14C dated individuals (n=5) and all sets of isotopic results
the values range from –20.7‰ to –17.2‰ (–19.1 ± 1.0‰) for carbon and from
+6.7‰ to +13.5‰ (+10.8 ± 2.0‰) for nitrogen. When excluding CP1956, Sk2,
which 14C date is coeval with Middle Neolithic groups in Portugal (see chapter 5),
the ranges remain very similar (–19.2 ± 1.1‰; +10.6 ± 2.1‰).
349
n/d
n/d
n/d
n/d
♂
♂
♂
♂
♂
♀?
n/d
♀
Adult
Adult
Adult
Adult
Adult
Adult
Adult
Adult
15–23 yrs
Adult
Adult
Adult, young
A, 1956–59
9, 1956
9, 1956
B, 1956–59
27, 1956
27, 1956
27, 1956
27, 1956
13, 1956
14, 1956
6, 1956
8, 1956
♀?
♂?
Adult
11, 1956
15 (9)?, 1956 Adult
Sex
Age
Ind.
Identification of the human remains
Long bone
999.87.16
Rib
999.40.47
Rib
999.95.6
Long bone
999.78.12
Vertebra
999.79.35
Long bone
999.86.12
Long bone
Humerus–R
999.89.2
Humerus
Cranium
Fibula
Rib
Cranium
Long bone
Bone
6891.3
7761.3
6894.3
7762.3
7768.3
6892.6
n/p
n/p
Ua–46934
n/p
Ua–46932
n/p
n/p
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
6734±51
analysis for isotopes
6780±48
analysis for isotopes
analysis for isotopes
–19.5
–19.5
–19.6
–19.7
–19.7
–18.6
–18.7
–19.5
–19.7
–19.8
–19.8
–20.0
–20.0
Measurements
14C by AMS; δ13C and δ15N by IRMS
14C Age BP
Lab no.
δ13C (‰)
VPDB
Ua–46933 6788±46
–20.7
8.7
8.8
9.2
8.5
8.9
10.6
9.8
10.0
13.5
8.2
10.4
9.0
9.3
3.2
3.3
3.3
3.3
3.3
3.3
–
2.9
3.3
2.9
3.2
2.9
2.9
δ15N (‰) C:N
AIR
6.7
3.3
19
19
18
17
17
28
27
19
17
16
16
15
15
8
M1
6
6
5
4
4
18
16
6
4
3
3
1
0
0
M2
% marine (±10)
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Umbelino 2006
continued...
Fontanals-Coll et al. 2014
This study
Fontanals-Coll et al. 2014
This study
Fontanals-Coll et al. 2014
Fontanals-Coll et al. 2014
This study
Reference
Table 4.64. Cabeço do Pez, Sado valley: carbon (δ13C) and nitrogen (δ15N) measurements on human bone collagen samples. All samples were collected at MNA. Samples
are sorted so that the lowest δ13C values are on the top of the table. Multiple samples of the same individual are sorted together.
Fibula
Rib
999.88.12
Long bone
999.74.21
Tibia
999.77.92
Rib
999.77.95
♀
♂
n/d
♀
Adult
Adult
Non-adult, 1.5–2 yrs n/d
n/d
Adult, mature
Non-adult, 1.5–2 yrs n/d
♂
Adult
Adult
Adult
Adult, 35–50 yrs
Adult, 35–50 yrs
21, 1956
25, 1956
20, 1956
19, 1956
2, 1956
2, 1956
16, 1956
1, 1956
5, 1956
5, 1956
♀
n/d
n/d
Adult
24, 1956
Rib
999.85.60
Rib
999.89.18
Rib
999.83.21
Long bone
Rib
999.84.17
Humerus
n/d
Non-adult, 5–7 yrs
Rib
999.92.17
Long bone
17, 1956
n/d
Adult
26, 1956
Table 4.64. continued
7766.2
Ua–46931
6893.3
7767.3
n/p
Ua–46930
7763.3
7758.2
7764.3
n/p
7759.2
n/p
7760.3
analysis for isotopes
6791±43
analysis for isotopes
analysis for isotopes
analysis for isotopes
5579±41
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
analysis for isotopes
–17.2
–18.3
–18.5
–18.9
–18.4
–19.1
–19.2
–19.2
–19.3
–19.3
–19.4
–19.4
–19.5
12.4
13.0
10.2
9.6
11.9
11.5
8.8
9.7
8.6
9.2
8.9
9.3
8.7
3.5
3.5
3.2
3.3
3.2
3.6
3.3
3.4
3.4
2.9
3.5
2.9
3.3
41
31
29
25
30
23
22
22
21
21
20
20
19
35
21
19
14
20
11
10
10
9
9
8
8
6
Guiry et al. 2015
This study
Guiry et al. 2015
Guiry et al. 2015
Fontanals-Coll et al. 2014
This study
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Fontanals-Coll et al. 2014
Guiry et al. 2015
Fontanals-Coll et al. 2014
Guiry et al. 2015
Chapter 4
16
15
14
13
δ15N (‰)
12
11
10
9
8
7
6
-22
-21
-20
-19
-18
-17
-16
-15
-14
-13
-12
δ13C (‰)
Figure 4.74. Cabeço do Pez, Sado valley: correlation between the δ13C (‰) and δ15N (‰)
values of 27 samples of human bone collagen from 21 individuals. Individuals 2 (circle), 5
(dash), 9 (square) and 27 (triangle) have multiple sets of values.
Archaeothanatology
Source material and analysis strategy
The archival documentation available for the burials at Cabeço do Pez is limited
and although some skeletons are maintained in blocks, most of them are poorly
preserved, limiting the possibilities of the archaeothanatological analysis. This is
unfortunate because this is a large site with minimum number of individuals of 32
(Cunha and Umbelino 1995–97) clustered in two main areas (A and B).
Archaeothanatological analysis was carried out on a total of seven individuals,
six from the material excavated in 1956 (26% of 27 individuals in site plan) and
one from the two individuals excavated in 1959. The completeness and detail of
each analysis is largely compromised by the character and quality of the available
source material (table 4.65). The source material comprises paraffin blocks of eight
individuals, in various states of preservation, drawings for the individuals excavated
in 1959, and two site plans of the burials. One of these blocks contains the remains
of CP1959, Sk2, but the material was extremely fragmentary at the time of the
352
Results: Sado valley, Cabeço do Pez
excavation and archaeothanatological analysis is not possible; although this burial
will be commented upon briefly throughout this text. For the remaining individuals
excavated at this site (n=20) there is no source material available, at least at the
time of this study.
Table 4.65. Cabeço do Pez, Sado valley: list of individuals considered for archaeothanatological analysis,
based on the preservation of the material. This material is curated at MNA.
Ind.
Source material
Observations
2
1956
yes
Graphic documentation w/
Written
human remains
doc.
Field
Drawings Site plan
photographs
n/a
n/a
yes
n/a
9
1956
yes
n/a
n/a
yes
n/a
Burial area B
10
1956
yes
n/a
n/a
yes
n/a
Burial area B. The upper body is poorly
preserved. Archaeothanatology is limited
and only possible on the lower limbs.
11
1956
yes
n/a
n/a
yes
n/a
Burial area B. Poor preservation.
Archaeothanatology is limited.
17
1956
yes
n/a
n/a
yes
n/a
Burial area B
21
1956
yes
n/a
n/a
yes
n/a
Burial area A. Poor preservation.
Archaeothanatology is limited.
27
1956
yes
n/a
n/a
yes
n/a
Burial area A
1
1959
n/a
n/a
yes
yes
n/a
–
2
1959
yes
n/a
yes
yes
n/a
Very fragmented material.
Archaeothanatology is not possible.
Block
Isolated burial. Coeval with Middle Neolithic groups (4466–4246 cal BCE, 95%
confidence). Extremely poor preservation.
Archaeothanatology is not possible.
The first phase of analysis took place at the MNA, where the eight individuals
preserved in paraffin blocks were analysed and photographed (CP1956, Sks 2, 9,
10, 11, 17, 21, 27; CP1959, Sk2). The second phase of analysis consisted of the
reassessment of my observations done in the museum, crosschecked with the
photographed material.
353
Chapter 4
Results of the archaeothanatological analysis
The nature of the deposits
Analysis of the nature of the deposits at Cabeço do Pez indicates a pattern of
primary burials (table 4.66). The cadavers were placed in the burial features before
the degradation of the labile joints, retaining their anatomical integrity. In some
cases the documentation is unclear and the diagnostic criteria are poorly preserved,
however, the overall position of the skeletal elements indicates that the bodies
were disposed while still retaining their general anatomical integrity. This is also the
case for the poorly preserved CP1959, Sk2 (non-adult), where the maintenance of
the general anatomical integrity of the few preserved elements suggests a burial in
primary position.
Table 4.66. Summary of the nature of the deposits with human remains at Cabeço do Pez, Sado valley.
Ind.
Age
Sex
Nature of the deposit
Arguments
9
1956
Adult
n/d
Primary
10
1956
Adult
n/d
Primary
Maintenance of labile articulations: some bones of right hand;
metatarsals and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
Poor preservation of the upper body.
Maintenance of labile articulations: metatarsals and phalanges
of left foot.
The bones of the lower body are well articulated suggesting the
maintenance of the general anatomical integrity of the body.
11
1956
Adult
♂?
Primary
Poor preservation and heavily covered with paraffin wax.
Maintenance of labile articulations: TM joint, CV, metatarsals
and phalanges of feet.
Maintenance of the general anatomical integrity of the body.
17
1956
Non-adult,
5–7 yrs
n/d
Probably primary
Poor preservation of diagnostic criteria.
Maintenance of the general anatomical integrity of the body.
21
1956
Adult
♀
Primary
Poor preservation and heavily covered with paraffin wax.
Maintenance of labile articulations: metatarsals and phalanges
of feet.
Maintenance of the general anatomical integrity of the body.
27
1956
Adult
♂
Primary
Poor preservation.
Maintenance of labile articulations: cervical vertebrae; a few
metatarsals and phalanges of a foot.
Maintenance of the general anatomical integrity of the body.
1
1959
Adult
n/d
Primary
Maintenance of labile articulations: metatarsals and phalanges
of left foot.
Maintenance of the general anatomical integrity of the body.
The space of decomposition of the cadaver
Analysis of the space of decomposition of the cadavers buried in Cabeço do Pez
indicates a general pattern of filled spaces (table 4.67). The cadavers were covered
with sediment immediately after their placement in the feature. The decomposition
of the bodies took place in this sediment-filled environment. In several cases the
analysis is unclear due to the poor preservation of the remains. In every case, the
movement of the bones is limited suggesting that the space of decomposition was
filled and that the sediment could penetrate immediately.
354
Results: Sado valley, Cabeço do Pez
Table 4.67. Summary of the space of decomposition of the human remains in the grave features at Cabeço
do Pez, Sado valley.
Ind.
Age
Sex
9
1956
Adult
n/d
Space of
decomposition
Filled
10
1956
Adult
n/d
Filled
The upper body is poorly preserved. The bones of the lower body
are well articulated. Maintenance of bones in original unbalanced
position: bones of left foot.
11
1956
Adult
♂?
Filled
Maintenance of bones in original unbalanced position: bones of feet
are rotated and placed in front of each other. No movement outside
initial volume of the cadaver
17
1956
Non-adult,
5–7 yrs
n/d
Filled
Poor preservation of diagnostic criteria but no movement outside
initial volume of the cadaver.
21
1956
Adult
♀
Filled
Maintenance of bones in original unbalanced position: moderate
collapse of left os coxae, bones of feet are rotated inwards and
placed towards one another (right behind left). No movement
outside initial volume of the cadaver.
27
1956
Adult
♂
Filled
Poor preservation of diagnostic criteria. Maintenance of bones in
original unbalanced position: moderate collapse of right os coxae
(left is not preserved). No movement outside initial volume of the
cadaver.
1
1959
Adult
n/d
Filled.
Maintenance of bones in original unbalanced position: left patella is
suspended on the distal end of femur; bones of left foot are rotated
outwards. Also, the cranium is in a semi-suspended position.
Arguments
Maintenance of bones in original unbalanced position: some bones
of right hand are rotated and flexed at the wrist, as well as several
bones of feet. No movement outside initial volume of the cadaver.
The initial position of the cadaver in the feature
At Cabeço do Pez all bodies analysed were lying on the back (n=7) (table 4.68) and
the remaining burials, sketched in the site plan, confirm this pattern.
The position of the upper limbs was consistent, typically placed close to the
body in semi-flexion at the level of the elbow joints. One exception is the burial of
a child (CP1956, Sk17) where the upper left limb was lying in semi-flexion but in
slight abduction. In the few cases where both limbs are well preserved, the analysis
indicates that the positions were symmetrical (n=3), semi-flexed at the elbows and
the hands placed in front of the body. This observation is not available for the
remaining four burials analysed.
At Cabeço do Pez, the bodies were placed in the feature with the lower limbs in
flexion or hyperflexion at the level of the hip joints and always in hyperflexion at
the level of the knees. The lower limbs were rotated to the right (CP1956, Sks 11,
17, 27–tendency) or with a tendency towards the left side (CP1956, Sks 10, 21;
CP1959, Sk1). The lower limbs are not observable in one case (CP1956, Sk11).
Clumping of the body was common (n=5), with the limbs somewhat nested on
the upper body and the feet forced towards the buttocks (CP1956, Sks 9, 10, 21,
27; CP1959, Sk1). In the other two cases this is not observable due to fragmentation of the remains (CP1956, Sks 11, 17), although this was very likely the case of
individual 17.
355
n/d
n/d
♂?
♀
♂
n/d
Adult
Adult
Adult
Non-adult, 5–7 yrs
Adult
Adult
Adult
9, 1956
10, 1956
11, 1956
17, 1956
21, 1956
27, 1956
1, 1959
n/d
Sex
Age
Individual
Trunk
Rotated–L and downwards
(neck flexed forwards)
Rotated–R
Prob. rotated–L
Rotated–L and downwards
(neck flexed forwards)
Rotated–R
Prob. rotated–L
On the back
On the back
On the back
On the back
On the back
On the back
Prob. rotated–L and downwards On the back
(neck flexed forwards)
Head
Initial position of the cadaver in the feature
In semi–flexion?–R
In semi–flexion–L
In semi–flexion–R
n/a–L
n/a–R
In semi–flexion–L
In semi–flexion–R
In semi–flexion–L
In semi–flexion
n/a
In semi–flexion
Upper limbs
Hip and knee joints: hyperflexion
Rotation–L
Feet towards buttocks. Clumping
Hip and knee joints: hyperflexion
Rotation–tendency R
In adduction in front of trunk. Clumping
Hip joints: flexion
Knee joints: hyperflexion
Rotation–tendency to L
Feet towards buttocks. Clumping
Hip joint–R: flexion
Hip joint–L: hyperflexion
Knee joints: n/a
Rotation–L
n/a
Hip joints: flexion
Knee joints: hyperflexion
Rotation–tendency L
Feet towards buttocks. Clumping
Hip joints: flexion
Knee joints: hyperflexion
Rotation–R
Feet towards buttocks. Clumping
Lower limbs
Table 4.68. Summary of the reconstruction of the initial position of the cadaver in the grave features at Cabeço do Pez, Sado valley. The initial position of the limbs is indicated as –R or –L and refers to the right or left limb, when their positions are different. Rotation –R or –L indicates the rotation of the skeletal element(s) towards the right
or the left side.
Results: Sado valley, Cabeço do Pez
The grave features
The size and shape of the deposits with human remains at Cabeço do Pez were not
described by the excavators. The archaeothanatological analysis suggests that the
bodies were placed in pits dug into the ground. The lateral pressures and walleffects observed on the skeleton in the feature (table 4.69) indicate that the shape
and size of the pit was just sufficient to contain the body. Overall, the pressures
exerted on the upper body are difficult to observe due to the poor preservation of
the diagnostic criteria. Nevertheless, the analysis suggests that in most cases the
design of the grave had a significant impact on the final arrangement of the cadaver in the feature, visible by the clumping of the bodies (CP1956, Sks 9, 10, 21;
CP1959, Sk1) becoming hypercontracted in one case (CP1956, Sk27). The narrow
shape of the graves imposed a significant limit to the cadaver providing lateral
support and preventing the collapse of the bones. These wall-effects probably
correspond to the physical limits of the pit, just wide enough to contain the
cadaver.
The documentation is limited but some cases suggest that the bottom of the
graves were irregular and sloping from the upper to the lower end of the feature.
This is indicated by the observation of the flexion and collapse forwards and
downwards of the head (CP1956, Sks 9, 17; CP1959, Sk1), suggesting that, at least
in these cases, the head was possibly slightly more elevated in the grave feature in
relation to the trunk. However, the fragmented state of the material does not allow
detailed observations at the level of ribs and sacrum rotations of which during the
process of decomposition could clarify aspects of the original position of the body
in the feature, in relation to the characteristics of the floor of the grave, such as in
the case of bodies placed in a half-sitting position.
357
n/d
n/d
♂
n/d
Adult
Adult
Adult
Non-adult, 5–7 yrs
Adult
Adult
Adult
9, 1956
10, 1956
11, 1956
17, 1956
21, 1956
27, 1956
1, 1959
♀
n/d
♂?
Sex
Age
Individual
1. Weight of sediment from above;
2. Overall strong lateral pressure and compressing effect.
Design of the grave feature.
1. Wall-effect, upper right side;
2. Overall constraint effect.
Design of the grave feature.
1. Wall-effect, lower end.
Design of the grave feature.
1. Wall-effect, upper right side.
Design of the grave feature.
1. Extreme flexion of the neck suggested by significant rotation forwards and
downward of the cranium lying in front of TC;
2. Suggested by alignment of arms close to TC, possible medial tendency of
TC, and alignment of lower limbs.
1. Alignment of upper right limb, involving right girdle
(arm is not preserved in the block);
2. Hypercontracted skeleton.
1. Alignment and clumping of lower limbs.
1. Alignment of upper right limb, involving right shoulder girdle and arm.
1. Alignment of upper right limb, involving right shoulder and arm;
2. Alignment of upper left limb, involving left shoulder girdle and arm.
1. Alignment and clumping of lower limbs.
Upper body is poorly preserved.
1. Wall-effect, lower left end.
Design of the grave feature.
1. Wall-effect, upper right side;
2. Wall-effect, upper left side.
Design of the grave feature.
1. Alignment of upper right limb and projection forwards and upwards of
right shoulder and arm. Pattern enhanced by the sloping floor causing a
transfer of body weight towards the right side;
2. Alignment of upper left limb and projection forwards and upwards of left
shoulder and arm;
3. Alignment and clumping of lower limbs.
Arguments
1. Wall-effect, upper right side;
2. Wall-effect, upper left side;
3. Wall-effect, lower end.
Design of the grave feature.
Pressures
Table 4.69. Summary of the key observations used in the reconstruction of the grave features at Cabeço do Pez, Sado valley. Observed pressures on the body are presented in
a number sequence. The arguments supporting these observations are indicated by the same number in the respective column.
Results: Sado valley, Cabeço do Pez
Deposits containing more than one individual
At Cabeço do Pez most deposits with human remains contain the remains of one
individual but there is some evidence supporting the hypothesis of double burials
at the site.
In the site plan, the deposits with the individuals 4 and 5, 9 and 10, 12 and 13
and 25 and 27 seem to be somehow related, either by being buried simultaneously
or by being disturbed by a later inhumation. Unfortunately, in most cases the
documentation is limited and a detailed analysis is not possible.
In one case only (CP1956, Sks 9, 10) the documentation is robust enough to
support the scenario of a multiple burial of two adults. The individuals were lying
in grave positioned in very close proximity. Individual 10 was possibly the first to
be deposited and shortly after individual 9 was placed on its left side. This is demonstrated by the position of the right humerus of individual 10 which lies behind
the left elbow of individual 9 (figs. 4.75, 4.76), suggesting that individual 10 was
still in anatomical integrity when the second individual (9) was placed in the
feature. The upper body of CP1956, Sk10 is poorly preserved, but its lower limbs
are well articulated and not disturbed, attesting to the primary position of the body
in the feature and supporting the double burial hypothesis. The bodies are partially
intertwined and come in contact with each other without disturbance of any joints,
suggesting a common arrangement of the bodies and a synchronous deposit.
Figure 4.75. CP1956, Sk9 in two blocks of paraffin. The distal end of the right humerus of
CP1956, Sk10 is preserved in the block with skeleton 9. Its position behind the left elbow
of CP1956, Sk9 suggests a common arrangement of the bodies and a synchronous deposit.
Preserved in several paraffin blocks at MNA. (Photographs by J. P. Ruas)
359
Chapter 4
A similar case (CP1956, Sks 25, 27) is suggested by observations from the site plan
where the bodies also seem to intertwine also at the level of the elbow joints,
suggesting a common arrangement (fig. 4.77). Individual 27 is partially preserved in
a paraffin block but the material is very fragmented at the level of the upper body
and there is no evidence suggesting either disturbance or common arrangement
with individual 25. Thus, in this case, the multiple burial hypothesis is not conclusive since it cannot be tested against other lines of evidence.
In the site plan, two other cases are represented in very close proximity
(CP1956, Sks 4 and 5; Sks 12 and 13) (figs. 4.76, 4.77), suggesting a common and
synchronous arrangement. Unfortunately, the current documentation is limited to
the site plan and the hypothesis of multiple double deposits cannot be tested.
Post-depositional manipulations of the cadavers
At Cabeço do Pez, the skeletons are apparently complete and lie undisturbed in the
grave feature. However, the unclear nature of the documentation does not allow a
reliable assessment at this level.
Other observations
The analysis of the paraffin blocks showed that some grave features may contain
other elements, possibly mixed in the grave filling. This is the case for CP1956, Sk9
where one small fragment of a non-human long bone was found lying in front of
the left os coxae. Another case (CP1956, Sk11) presented a fragment of a long
bone of a child, possibly a humerus, lying close to the thoracic vertebrae of the
adult. In either case it is not possible to determine the nature of these inclusions.
Figure 4.76. Burial area B at Cabeço do Pez, Sado valley. The drawing suggests the hypothesis of double burial of Sks 9 and 10, which is confirmed by the skeletal remains preserved
in block. CP1956 Sks 12, 13 lie in close proximity but the data is not conclusive regarding
the simultaneous deposit of these individuals. (MNA, 1956, A. Paiva)
360
Results: Sado valley, Cabeço do Pez
Figure 4.77. Burial area A at Cabeço do Pez, Sado valley. The drawing suggests the hypotheses of multiple double burials of Sks 25 and 27 (superior left side), and possibly 4 and 5 (inferior
left corner), but the data is not sufficient to support these hypotheses. (MNA, 1956, A. Paiva)
The mortuary programme at Cabeço do Pez
The archaeothanatological analysis at Cabeço do Pez is limited due to the scarce
and unclear nature of the documentation. Also, the skeletons maintained in paraffin blocks are poorly preserved, particularly at the level of the upper body. Nevertheless, it is possible to observe some patterns in the mortuary practices at the site.
At Cabeço do Pez, the cadavers were typically placed in individual grave
features while retaining their anatomical integrity (individual primary deposits). The
feature was then immediately covered with sediment providing a filled space of
decomposition. Some cases suggest the synchronous burial of two individuals in a
common grave feature, but as discussed, in one case only the data is sufficient to
support the scenario of a double burial.
The bodies were normally laid on the back with the upper limbs placed close to
the body, which were commonly semi-flexed at the elbows. The lower limbs were
brought up towards the upper body and rotated to one side, and were always
hyperflexed at the knees. The feet of the cadavers were always placed towards the
buttocks in a forced position. The clumping of the cadavers at the level of the
lower body is a common feature of these burials.
The pressures exerted on bodies in features can be explained as the result of the
narrow design of the grave where the wall-effects and consequent pressures correspond to the physical limits of the pit, just wide enough to contain the individual.
However, the documentation is limited and extra support of cadavers, provided by
soft wrappings, cannot be excluded.
The data is limited but some cases suggest that the floor of the graves tended to
be uneven and sloping from the upper to the lower end of the feature.
Overall, and despite the limitation of this data set, the mortuary programme at
Cabeço do Pez suggests a consistent set of mortuary practices and all cases
observed follow a common way of burying the dead, both in area A and area B.
361
Sado valley: Várzea da Mó
Site background: a short history of research
Várzea da Mó is a shell midden site located on the eastern side of the Sado valley
(fig. 2.3). It is situated on the right margin of the Sado River, but the precise location of the site is unknown, and it was not yet relocated during the scope of
current fieldwork in the region (Diniz et al. 2012).
Archaeological material and documentation available for Várzea da Mó are
curated at the National Museum of Archaeology in Lisbon (MNA). The year of
identification of the site is unclear but according to a short article by M. Heleno
(1956, 229) the site was already known before the identification of Vale de Romeiras, Cabeço das Amoreiras, and Arapouco. The excavation diary (MNA, APMH,
2/3/17/1) shows that the first field season started on the 3rd of August 1959, but
unfortunately, like other documents, this report is incomplete and the last entry
does not correspond to the last day of excavation. In these notes there are no
references to the identification of human remains.
Graphic documentation in the archives of MNA consists of one site plan with
one human skeleton plotted in the excavation area and profile (MNA, 1959, D.
Sousa, A117), a few poorly preserved photographs (MNA, APMH, 2/11/84), and
individual sketches of the only human skeleton excavated at the site (figs. 4.78,
4.79).
According to the documentation, this individual was excavated at a depth of c
65 cm. The notes in the field sketch and in the final version of the site profile are
contradictory regarding the layer where the human remains were found. In the
sketch the layer is described as grey soil with few shells, while in the site profile the
skeleton is plotted in the bottom layer of white sand, under the grey soil.
The skeleton belongs to an adult individual, probably middle aged (Cunha et al.
2002, 193). It is preserved in a block of paraffin but the bone elements are very
fragmented.
362
Results: Sado valley, Várzea da Mó
Figure 4.78. Várzea da Mó, Sado valley: site plan and profile of the area with human remains. (Adapted from a photograph by J. P. Ruas. MNA, 1959, D. Sousa, A117)
Figure 4.79. VM1959, Sk1 preserved in paraffin block at MNA. (Photograph by J. P. Ruas.
Drawing by D. Sousa, MNA)
363
Chapter 4
Chronology of the burial activity
Previous 14C measurements
One sample of shells (ICEN–273), reported to be collected from the middle layers
of the shell midden excavated in 1959, was dated in the 1980s (Arnaud 2000) (table
A13).
Objectives of the dating programme
The aim of the dating programme for Várzea da Mó was to date the only human
burial known at the site and to contextualize it in the chronological framework of
the burial activity in the valley.
Sampling strategy
One sample of human bone from VM1959, Sk1 was collected for 14C analysis (table 4.70).
Table 4.70. Sample of human bone (1) from Várzea da Mó collected and submitted for 14C measurement
in the course of this study. The sample is from the 1959 assemblages and was collected at MNA.
Ind.
Lab no.
Age
Sex
Bone
Stratigraphy
Observations
1
Ua–46310
adult
n/d
Radius
Bottom, c 65 cm
Paraffin block
Results
The only human burial known for this site is now dated (table 4.71).
Table 4.71. Human bone collagen sample (1) from Várzea da Mó, Sado valley: 14C measurement and
estimation of non-terrestrial carbon intake (% marine). Marine percentage was estimated from the measured
δ13C value of the sample by the application of method marine 1 and method marine 2, based on adopted
endpoint values of –12‰ and –20‰ for marine (100%) and terrestrial (100%) diets respectively.
Identification of the
human remains
Ind. Excav.
Museum
1
Measurements
14C by AMS; δ13C and δ15N by IRMS
14C Age BP
Lab no.
δ13C (‰)
VPDB
1959, MNA Ua–46310 6305±44
–20.5
δ15N (‰)
AIR
8.1
% marine
(±10)
C:N M1 M2
Reference
3.6
This study
10
0
Calibration and chronological models
The 14C measurement on human bone collagen from Várzea da Mó dates the
burial activity at the site to have been in 5345–5071 cal BCE (95% confidence, M1)
or 5462–5207 cal BCE (95% confidence, M2) (fig. 4.80, table 4.72).
364
Results: Sado valley, Várzea da Mó
Figure 4.80. Probability density function estimated for the death of VM1959, Sk1 is 5345–
5071 cal BCE (95.4% confidence, marine 1), 5345–5194 cal BCE (79.9% confidence, marine 1), or 5309–5214 cal BCE (68.2% confidence, marine 1). Calibrated ranges of 95% and
68% probability are given by the lower and higher square brackets respectively.
Table 4.72. Human bone collagen sample from Várzea da Mó, Sado valley: calibrated date ranges according to two different approaches for the calculation of the marine percentage. Results are presented in both
cal BCE and cal BP in 95.4% confidence range.
Ind.
Lab no.
14C
Age BP
1, 1959
Ua–46310
6305±44
Calibrated date range
(95% confidence)
Calibrated date range
(95% confidence)
cal BP
cal BCE
7294–7020, M1
7411–7156, M2
5345–5071, M1
5462–5207, M2
Stable isotopes: carbon and nitrogen
Previous measurements: δ13C and δ15N
The first attempt to measure stable isotopes on the only individual buried at
Várzea da Mó did not yield enough collagen (Fontanals-Coll et al. 2014).
Sampling strategy
One sample was collected from the VR1959, Sk1 in the scope of the radiocarbon
dating programme (chapter 3).
Results
Sample Ua–46310 was measured by IRMS and the precision of the measurements
for both δ13C and δ15N is ± 0.1‰. At Várzea da Mó, the human collagen δ13C
value (n=1) is –20.5‰ and the δ15N value (n=1) is +8.1‰ (fig. 4.81).
365
Chapter 4
16
15
14
13
δ15N (‰)
12
11
10
9
8
7
6
-22 -21 -20 -19 -18 -17 -16 -15 -14 -13 -12
δ13C (‰)
Figure 4.81. Várzea da Mó, Sado valley: correlation between the δ13C (‰) and δ15N (‰)
values of one sample of human bone collagen from the only burial known at the site.
Archaeothanatology
The individual buried at Várzea da Mó is preserved in a paraffin block but the
skeletal elements are poorly preserved and very fragmented. Archival documentation consists of field photographs, sketches and one site plan with the burial (table
4.73).
Table 4.73. Várzea da Mó, Sado valley: source material used for archaeothanatological analysis. This
material is curated at MNA.
Individual Source material
Observations
Block Graphic documentation w/human remains Written
documentation
Field photographs Drawings Site plan
1, 1959
yes
yes
yes
yes
n/a
Fragmented material.
Archaeothanatology is very
limited.
The nature of the deposit, the space of decomposition, and the initial position of the cadaver
The primary nature of VR1959, Sk1 is suggested by the maintenance of the general
anatomical integrity of the body. The labile articulations are unclear in the documentation, however, the concentration of several hand bones on the pelvic region
indicates moderate collapse of these labile joints.
366
Results: Sado valley, Várzea da Mó
Decomposition of the body was possibly in a sediment-filled environment as
suggested by the moderate collapse of the bones of the hands. Also, the skeletal
elements remained in order, and there is no movement outside the initial volume
of the cadaver.
The cadaver was lying on the back with the upper limbs along the body and the
hands placed in front of the lower abdomen. The lower right limb was semi-flexed
(c 90°) at the hip, suggesting that the knee was directed forwards. Other skeletal
elements of the lower limbs were presumably not in the feature at the time of the
excavation. This could be explained by disturbance of the feature, but the documentation does not allow further analysis.
The cranium was poorly preserved and several fragments are concentrated on
the upper right side of the TC, in the paraffin block. However, the field sketch
indicates that these fragments were found in anatomical position and this placement in front of the TC was possibly after excavation, at the time the block was
being prepared for transportation to the museum.
The verticalization of the clavicles and ribs suggests significant compression
towards the medial axis of the body. Most diagnostic elements are unclear in the
documentation, however, the rotation of these elements is a strong indicator of a
bilateral pressure exerted on the body. This bilateral wall-effect could be explained
by the narrow shape of the pit, imposing a significant limit to the cadaver while
offering lateral support and preventing the collapse of the bones. In this scenario,
the size and shape of the deposit was just sufficient to contain the body in a
relatively constrained position.
Post-deposition manipulations of the cadaver
Nothing noted.
367
Chapter 5
Analysis and Synthesis of the Results
Introduction
The methodological choices of this dissertation were explicitly targeted towards the
human remains excavated in the shell middens of the Tagus and Sado valleys. This
strategy was aimed to understand the significance of the mortuary activity in these
sites. The data obtained in the course of this study offers significant evidence to
start defining who was buried in these places, as well as when and how. In this chapter, Analysis and Synthesis of the Results, I examine the results presented in chapter 4
in the light of the general aims and questions of this dissertation, in terms of time,
people, and funerary practices.
I start with an examination of the Chronology of the burial activity in each site and in
each valley. The analysis is presented by following the objectives of the dating
programme, the general aims of which were to define the chronology of the burial
activity in terms of duration, start/end, and frequency (see chapter 3, Enquirydependent chronologies: the radiocarbon dating programme in this study), along with sitespecific questions (see chapter 4, Site/Objectives of the dating programme). In each case
I evaluate the representativeness of the results, and compare with previously published proposals when available, as well as with the current chronological framework proposed for the Late Mesolithic in Portugal, c 6350–5250 cal BCE (Araújo
2015). This is an interpretative chronology for the burial activity based on the
chronological models presented in chapter 4, and new data may refine and/or
redefine some of these scenarios. The models are based on 14C dates on buried
human bone exclusively. Other dates are available, but in most cases the contexts
are unclear and the relationship with the burials is unknown. It is not possible at
the moment to discuss the chronology of other contexts in most middens because
there is a lack of a reliable data set, along with a clear definition of what these other
uses are (i.e., ritual, non-funerary, domestic?). The discussion in this dissertation is
on the use of these sites for burial practice. Ongoing (Bicho et al. 2013; Diniz and
Arias 2012) and future research will possibly clarify these and other questions that
have persisted for decades. The analysis at the level of each site is followed by a
discussion on the implications of the chronological data for the local setting, at the
scale of each valley, Tagus vs Sado.
The second section of this chapter is centred on the Isotopic signatures as a means
to define the people in each burial place. Here I integrate the data obtained in this study
369
Chapter 5
with previously published isotopic signatures (Fontanals-Coll et al. 2014; Guiry et
al. 2015; Lubell et al. 1994; Umbelino 2006), and examine the isotopic data from
the perspective of funerary practices. The aim is to identify isotopic signature patterns that may reflect homogeneity or heterogeneity of the burial groups at the
scale of the site, valley and inter-valley. In this section, I introduce the temporal
dimension as presented in the section Chronology of the burial activity, and investigate
the impact of time in the observed patterns and outlier signatures. Even at its most
elementary level of analysis, the isotopic data (δ13C and δ15N) were proved to be a
powerful data set for the interpretation of these sites, which future studies will be
able to refine.
In the third and last section of this chapter, I present a synthesis of the results
of the archaeothanatological analysis presented in chapter 4. The treatment of the dead:
synthesis of the archaeothanatological analysis is presented in terms of central aspects
about the nature of the deposits, the space of decomposition of the cadaver, the
initial position of the cadaver in the feature, the grave features, the deposits containing more than one individual, and the post-depositional manipulation of the
cadavers. The mortuary programme described for each site in the previous chapter
is examined here at the scale of the valley and compared across the Tagus and
Sado valleys.
Chronology of the burial activity
Moita do Sebastião, Tagus valley
The chronological model proposed for the burial activity at Moita do Sebastião is
presented in chapter 4 and is based on 18 14C measurements on human remains
which can be considered representative of the assemblage:
1. Five measurements are from the nineteenth century excavations (MNI
52) selected on the basis of representative enveloping sediment (Lubell
and Jackes 1985). Despite the differences in sediment the 14C dates are
consistent, with one exception (TO–135, individual CT) which presents
a more recent date than the other four samples.
2. Thirteen measurements are from the 1952–54 excavations (MNI 34)
and represent a good spatial coverage of the burial area. With one
exception (Ua–46264, individual 9) which presents an earlier date, the
14C results for these samples are highly consistent.
3. The results obtained for both groups of samples, nineteenth century
and 1952–54, are highly coherent and can be analysed as one series.
This is consistent with the stratigraphic data available for the site which
indicates that despite the c 3 metres depth of the midden, most skeletons were excavated from the basal sand layer (Jackes and Alvim 2006).
370
Analysis and Synthesis of the Results
Chronological relationship of the burial areas
The new dates obtained for the burials excavated in 1952–54 are coeval with previous 14C measurements of the burials recovered in the areas excavated in the nineteenth century. Two of the new dates (Ua–47978, MS1952–54, Sk10 and Ua–
46270, MS1952–54, Sk33) are consistent with TO–135 (MS19th c., SkCT), which
was rightly interpreted as an outlier (Jackes and Lubell 2012). Now, this burial is
well established in the chronology of the site consistent with a later burial activity.
The burials in the excavation areas of both the nineteenth century and 1952–54
were in use at the same time and do not correspond to different burial phases.
However, some of the older burials are located in the 1952–54 excavation in a
central area of the shell midden (area A) (fig. 4.1). Also, the new 14C dates indicate
that the different burial areas (A, B and C) identified in 1952–54 are coeval in time.
The 14C data demonstrates that the different spatial areas are synchronic (fig. 5.1).
However, the chronological sequence suggests that the burial activity clusters in
three chief moments in time, possibly discontinuously, with relatively short hiatuses between them without significant burial activity (fig. 5.2).
Figure 5.1. Moita do Sebastião, Tagus valley, 1952–54: site plan and 14C dated human remains. Dash stroke: earlier burials (1, 9). Full stroke: main burial activity (5, 15, 17, 18, 22, 25,
30, 31, 34). Polygon: later burials (10, 33). Unreliable measurements (14). (Site plan reproduced from Roche 1972, 116, fig. 29)
371
Chapter 5
Figure 5.2. Chronological model for the burial activity at Moita do Sebastião, as presented in
chapter 4. Shade: highlights the moment of most frequent burial activity at the site c 5950–
5650 cal BCE.
Burial activity: duration, start/end, frequency
The new 14C dates suggest an earlier use of the site than previously considered.
This is demonstrated by the measurements obtained for MS1952–54, Sks 9, 1 (Ua–
46264, Ua–46263, 1952–54, area A) presenting the earliest dates on human bone
collagen known at the Tagus valley. The posterior density estimated for the time of
372
Analysis and Synthesis of the Results
their burial is 6292–6023 cal BCE (95% probability) for individual 9 and 6218–5998
cal BCE (95% probability) for individual 1. Interestingly, MS1952–54, Sks 9 (male
adult, Ferembach 1974), was the only individual buried with the lower limbs in full
extension, which is unique at both the Tagus and Sado valleys. According to J.
Roche (1972, 117), the feature containing MS1952–54, Sk1 (sépulture I) included the
remains of two adults probably disturbed at the time of the burial of MS1952–54,
Sk5 (fig. 5.3). The 14C measurements confirm that the burial of individual 5 is more
recent (Ua–47977, 5972–5672 cal BCE, 95% probability).
Figure 5.3. MS1952–54, Sks 5, 9, 12, 15: 14C data. (Adapted from a photograph in Cardoso
and Rolão 1999–2000, 200, fig. 38 and site plan in Roche 1972, 116, fig. 29)
The second main moment of burial activity at Moita do Sebastião is where most
14C dates cluster (n=13) indicating that the burial activity was continuous and more
frequent during this time span. The new dates support the suggestion of Jackes and
Lubell (2012) of a main group of burials clustering at c 5850 cal BCE. Burials from
all excavated areas fall into this time span, suggesting that when the burial practice
was more frequent, the extent of the burial area was as large as the total burial area
known at Moita do Sebastião. The earliest burials from this second time cluster are
represented by a child buried in area C (MS19th c., Sk22, TO–131) with a posterior
density estimated for the time of burial of 6021–5705 cal BCE (95% probability), and
MS19th c., Sk29 (TO–133 with a posterior density estimated for the time of burial
of 6017–5707 cal BCE (95% probability). The 14C dates for this time span are consis373
Chapter 5
tent with the dates of most individuals, and all present identical ranges. One of the
latest dates known for this main burial activity is found in area A, MS1952–54,
Sk15 (Ua–46265), with a posterior density estimated for the time of burial at 5843–
5611 cal BCE (95% probability).
The third and last period of concentrated burial activity at Moita do Sebastião is
indicated by the 14C dates of three individuals: MS1952–54, Sks 10, 33 in area A
(Ua–47978, Ua–46270) and MS19th c., SkCT (TO–135) which precise location is
unknown. These three dates are coeval with a posterior density estimated for the
time of their burial around 5650–5450 cal BCE. The new 14C dates securely
demonstrate the later burial activity at the site as already suggested by a previous
date (TO–135, MS19th c, SkCT).
Moita do Sebastião is a large site with high concentration of human burials
(MNI 85). The chronological model suggested assumes one general phase of burial
activity, with the burials taking place more or less continuously during a period of
time, as defined by the start/end of burial activity. In this scenario the model suggests
that the burial activity was clustered in three main periods, in a slightly discontinuous fashion, with short hiatuses between them. The burial activity was more
frequent between c 5950–5650 cal BCE, with an earlier moment at c 6200–6000
cal BCE and a later cluster at c 5650–5450 cal BCE.
In this model, the burial activity is estimated to have continued for between 469
and 816 years (95% probability), or for 523 and 696 years (68% probability). This
estimate suggests a frequency of at least one burial every 10 years, or two burials
every two generations, based on the minimum number of individuals recovered at
the site (n=85) and assuming the widest span of burial activity of 816 years. If the
chronological model considered more than one well defined phase, such as in the
hypothetical model 2 (see below, Alternative model), the frequency of burial practice
within each phase would have been higher.
If Moita do Sebastião was used as the burial ground of one residential group of
hunter-gatherers (n=15 to 17), with a death rate of 0.3 or 0.5 per year (see chapter
2, Hunter-gatherer death rate vs burial population), the burial ground would aggregate in
816 years, c 245 or 408 human burials, respectively. The current MNI (n=85,
Jackes and Meiklejohn 2008) accounts for 35% or 21% of this hypothetical burial
population. D. Ferembach (1974) calculated a MNI of 136 individuals. In this case,
the human remains recovered would account for 56% or 33% of the total burial
population. If we consider a smaller span of burial activity, of c 600 years, the current MNIs would represent a larger part of the population, but probably not more
than half of the group (table 5.1).
374
Analysis and Synthesis of the Results
Table 5.1. Representativeness of the archaeological burial population at Moita do Sebastião in relation to
the original group. Estimates assume a residential group unit of nomadic foragers of 15 to 17 people (after
Binford 2001; Hamilton et al. 2007). Death rate estimates after M. Gurven and H. Kaplan (2007).
Death rate/year
Burial activity span
Total burial population
MNI 85
MNI 136
0.3
816 yrs
245
35%
56%
0.5
816 yrs
408
21%
33%
0.3
600 yrs
180
47%
76%
0.5
600 yrs
300
28%
45%
The proposed model can account for the previous dates on charcoal of unidentified species (H–2119/1546 and Sa–16, table A3), although one of the dates (Sa–
16) has a large margin of error (±350) suggesting a slightly earlier use of the site.
The new 14C dates introduce an earlier moment of burial activity at Moita do
Sebastião. The current posterior density estimated for the start of the burial activity
at Moita do Sebastião is 6365–6051 cal BCE (95% probability) which may be slightly
earlier than the previous estimate at c 6150 cal BCE (Jackes and Lubell 2012). The
new dates support the previous suggestion of M. Jackes and D. Lubell (2012) of a
main burial activity between c 5950–5650 cal BCE, which is where most dates
cluster. Furthermore, the new data set confirms the later burial activity at the site,
already suggested by one previous date (TO–135), estimating the end of the burial
activity at 5603–5351 cal BCE (95% confidence).
The burial activity at Moita do Sebastião starts while the estuarine environment
is abruptly established in the region (c 6100 cal BCE, van der Schriek et al. 2007)
but ends while this apparently favourable ecosystem is still predominant. Also, the
proposed model suggests that the burial activity at Moita do Sebastião is coeval
with the currently proposed dates for the beginning of the Late Mesolithic in Portugal (Araújo 2015), at c 6350 cal BCE. The burial activity at Moita do Sebastião
ends a few centuries before the end of the Mesolithic at c 5250 cal BCE.
Alternative model
As a working hypothesis, a second chronological model could be proposed, by
explicitly defining three main periods of burial activity as three distinct phases, with
clear boundaries defined for each phase (fig. 5.4). In this model, chronological
relations are imposed and the order of the three phases is assumed; clear hiatuses
between phases, and duration of phases are defined. However, the archaeological
data (i.e., prior beliefs) do not clearly indicate burial activity in three phases. The
resolution of the stratigraphic data is low and we know only that the burials were
excavated from the basal sand layer, at the bottom of the shell midden. This is not
to suggest that the burial activity could not have happened in three distinct phases,
but to caution against the statistical manipulation of the data, which, without a
strong archaeological basis, can significantly mislead representation of the data.
Thus, the current archaeological documentation (the prior beliefs) does not
securely allow the construction of a robust second model.
375
Chapter 5
Figure 5.4. Hypothetical chronological model 2 for the burial activity at Moita do Sebastião,
Tagus valley. Posterior densities estimated for the start and end of the burial activity are
7001–6093 cal BCE (95% probability) and 5597–4800 cal BCE (95% probability), respectively.
Phase 1 of burial activity is estimated to have started 6515–6072 cal BCE (95% probability)
and ended 6251–5850 cal BCE (95% probability). Phase 2 of burial activity is estimated to
have started 5917–5752 cal BCE (95% probability) and ended 5831–5686 cal BCE (95% probability). Phase 3 of burial activity is estimated to have started 5712–5482 cal BCE (95% probability) and ended 5609–5333 cal BCE (95% probability). Observe that the current archaeological data does not acknowledge the construction of a model that explicitly assumes three
distinct phases of burial activity. Model-definition command files in appendix, Tagus valley,
Moita do Sebastião, Hyp. Model 2, Marine 1.
376
Analysis and Synthesis of the Results
Cabeço da Arruda, Tagus valley
The chronological model proposed for the burial activity at Cabeço da Arruda is
presented in chapter 4 and is based on 15 14C measurements on the human remains of 14 individuals. This is a large site with a large number of human burials
(MNI 110) in a complex stratigraphy. Nevertheless, the current set of dates can be
considered representative of the assemblage:
1. The samples from the nineteenth century assemblages (n=5), for which
precise location in the midden is unknown, were selected randomly
(Lubell et al. 1994). The 14C measurements are consistent, with one
exception (TO–356, individual N) which presents a more recent date
than the other four samples.
2. The material now dated from the 1937 and 1964 excavations was selected according to the position of the burials in the shell midden, and
despite the lack of dates from the upper layers, this set represents a
comprehensive stratigraphic coverage.
3. Overall, the 14C measurements obtained for the three groups of samples, nineteenth century, 1930s and 1960s are consistent and can be
analysed as one series.
Chronological relationships of the burial areas
The chronology of the material excavated in the nineteenth century, in 1937 and
1964 overlaps in time. However, the material excavated in the nineteenth century
generally presents more recent dates. The 14C data indicates that the different
spatial areas are synchronic. Also, the chronological sequence suggests that the
burial activity took place continuously.
Burial activity: duration, start/end, frequency
The earliest burial activity at Cabeço da Arruda is represented by the two 14C dates
available for CAR1937, Sk6. This individual was excavated at c 0.80 metres above
the basal sand, and c 3.80 m from the top of the midden (Cardoso and Rolão
1999–2000, 178). Despite doubts expressed about the early date of this burial
(Jackes and Meiklejohn 2004; Jackes and Lubell 2012) it remains the earliest burial
known at Cabeço da Arruda, even after being recently re-dated (Jackes et al. 2014).
The posterior density estimated for the time of burial is 6133–5852 cal BCE (95%
probability), obtained by combining the two dates available (Beta–127451 and
AA–101343). The combined calibrated date range is 6223–5912 cal BCE (95%
confidence, fig. 4.13) which is slightly older but consistent with the single calibration of the new measurement (AA–101343) which presents a calibrated date range
of 6205–5811 cal BCE (95% confidence) (fig. 5.5). However, the combined
calibrated date range is not as old as the original date (Beta–127451), the calibrated
date range for which is 6413–5996 cal BCE (95% confidence) (fig. 5.6), mostly due
to its large margin of error (±100). Nevertheless, the 14C data available for this
377
Chapter 5
individual indicates an earlier burial activity, which according to M. Jackes and D.
Lubell (2012) was expected to be no earlier than 5750–5650 cal BCE.
Figure 5.5. Probability density function for the calibration of sample AA–101343 (Jackes et
al. 2014) from bone collagen of CAR1937, Sk6. The calibrated measurements for this sample indicate that the probability density estimated for the death of individual 6 is 6205–5811
cal BCE (95% confidence) or 6066–5899 cal BCE (68% confidence).
Figure 5.6. Probability density function for the calibration of sample Beta–127451 (Cunha
and Cardoso 2002–03; Cunha et al. 2003) from bone collagen of CAR1937, Sk6. The calibrated measurements for this sample indicate that the probability density estimated for the
death of individual 6 is 6413–5996 cal BCE (95% confidence) or 6288–6063 cal BCE (61%
confidence).
378
Analysis and Synthesis of the Results
This earlier burial (CAR1937, Sk6) is followed by the main period of burial at
Cabeço da Arruda, where most 14C dates cluster (n=9), indicating that burial activity was more frequent between c 6000–5600 cal BCE (fig. 5.7). Burials excavated
from all areas fall into this time span, suggesting that when the burial practice was
more frequent, the extent of the burial area was as large as the total burial area
known at Cabeço da Arruda. Burial practice is frequent after the burial of
CAR1937, Sks 10, 9, 3 (Ua–46275, Ua–46274, Ua–46273) and CAR1964, Sk4
(Ua–47976) followed by CAR1937, Sk1 (Ua–46272). CAR1937, Sk2 (Ua–47975),
CAR19th c., Sks III, A (TO–360, TO–354), and CA–00–02 (1960s area,
TO–10216) are coeval within this main activity. The later burials of this main activity are indicated by the 14C dates of two individuals: CAR19th c., Sk 42 and D
(TO–359, TO–355) (table 5.2).
Figure 5.7. Chronological model for the burial activity at Cabeço da Arruda, calibrated using
marine 1 diet values. Shade: highlights the period of most frequent burial activity at the site
c 6000–5600 cal BCE
379
Chapter 5
Table 5.2. Cabeço da Arruda, Tagus valley: stratigraphy and burial activity at the site.
Post. density estimate
cal BCE, 95% prob.
Individual/Lab no.
Stratigraphy
Reference
Burial activity
N, 19th c./TO–356
Upper layers
(based on
sediment matrix)
Jackes and Lubell 2012 5464–4963
Latest
CA-00-01/TO–10217
Upper layers
Roksandic 2006
5602–5287
Late
42, 19th c./TO–359
Basal sand
Jackes et al. 2013
5827–5495
Main–Late
D, 19th c./TO–355
Basal sand
Jackes et al. 2013
5734–5428
Main–Late
III, 19th c./TO–360
Basal sand
Jackes et al. 2013
5972–5511
Main
A, 19th c./TO–354
Basal sand
Jackes et al. 2013
5902–5617
Main
1, 1937/Ua–46272
c 1.40 m
from basal sand
Cardoso and Rolão
1999–2000, 178
5975–5696
Main
3, 1937/Ua–46273
c 1.20 m
from basal sand
Cardoso and Rolão
1999–2000, 178
5993–5745
Main
2, 1937/Ua–47975
c 0.85 m from basal
sand
Cardoso and Rolão
1999–2000, 178
5905–5640
Main
9, 1937/Ua–46274
c 0.30 m
from basal sand
Cardoso and Rolão
1999–2000, 179
5998–5748
Main
10, 1937/Ua–46275
c 0.25 m from basal
sand, under skeleton 9
Cardoso and Rolão
1999–2000, 179
6049–5798
Main
CA-00-02/TO–10216
Basal sand (1960s area) Roksandic 2006
5914–5621
Main
4, 1964/Ua–47976
Basal sand
Roche 1974
6021–5766
Main
6, 1937/R_Combine:
AA–101343
Beta–127451
c 0.80 m
from basal sand
Cardoso and Rolão
1999–2000, 178
6133–5852
Early
The only date securely attributed to the upper layers of the shell midden corresponds to the later occurrence of burial activity at Cabeço da Arruda (TO–10217,
individual CA–00–01, 5602–5287 cal BCE, 95% probability) which is in good agreement with the remaining 14C measurements from Cabeço da Arruda.
Individual N (nineteenth century) is the latest burial known at the site. It presents a much more recent date (TO–356, 5466–4951 cal BCE, 95% probability) and
is in poor agreement with the remaining dates available, suggesting an episode of
later burial activity at the site. Its exact location and depth are unknown, although
the enveloping sediment seems to correspond to the characteristic matrix of the
upper layers (Jackes and Lubell 2012). More dates from the upper layers could
eventually clarify this relatively later burial activity, but unfortunately the only
380
Analysis and Synthesis of the Results
material securely identified from these upper layers, excavated in 1965 was mixed
with material recovered in 1964 from the bottom of the midden.
Like Moita do Sebastião, Cabeço da Arruda is a large site with a high concentration of human burials (MNI 110). The chronological model suggested assumes one
general phase of burial activity. The burials take place continuously during a period
of time, as defined by the start/end of burial activity, and are more frequent between
c 6000–5600 cal BCE, with earlier and later burial episodes.
In this model, the burial activity is estimated to have continued for between 557
and 1093 years (95% probability), or for 682 and 918 years (68% probability). This
estimate suggests a frequency of at least one burial every 10 years, or two burials
every two generations, based on the minimum number of individuals recovered at
the site (n=110) and assuming the widest span of burial activity of 1093 years.
If Cabeço da Arruda was used as the burial ground of one residential group of
hunter-gatherers (n=15 to 17), with a death rate of 0.3 or 0.5 per year (see chapter
2, Hunter-gatherer death rate vs burial population), the burial ground would aggregate in
1093 years, c 328 or 547 human burials, respectively. The current MNI (110)
accounts for 34% or 20% of this hypothetical burial population. If we accept, as
discussed, that the MNI could be higher than 170 individuals, the human remains
recovered would account for c 52% or 31% of the total burial population. If we
consider a smaller span of burial activity, of c 800 years, the current MNIs would
represent a larger part of the population, but probably not more than half of the
group (table 5.3).
Table 5.3. Representativeness of the archaeological burial population at Cabeço da Arruda in relation to
the original group. Estimates assume a residential group unit of nomadic foragers of 15 to 17 people (after
Binford 2001; Hamilton et al. 2007). Death rate estimates after M. Gurven and H. Kaplan (2007).
Death rate/year
Burial activity span
Total burial population
MNI 110
MNI 170
0.3
1093 yrs
328
34%
52%
0.5
1093 yrs
547
20%
31%
0.3
800 yrs
240
46%
71%
0.5
800 yrs
400
28%
43%
The proposed model can account for three (Sa–197, Beta–152956, and Beta–
271927) of the five previous dates on non-human material (table A4). Two of these
samples are on unidentified species of charcoal. One was collected by J. Roche in
the 1960s from the lower layers (Sa–197) and it has a relatively recent date (5977–
4721 cal BCE, 95% confidence). The second date (Beta–271927) was recently
measured from charcoal inside the crushed skull of a child excavated in the
nineteenth century (Jackes et al. 2013). This charcoal presents a very similar date to
the measurement obtained for the articulated dog (Canis familiaris) found at Cabeço
da Arruda (Beta–152956), the calibrated date range for which is 6023–5851
cal BCE (95% confidence). A third date on unidentified charcoal (TO–10215,
6426–6100 cal BCE, 95% confidence) is slightly older than the range proposed in
381
Chapter 5
the model, which could be the result of old wood effect (Jackes et al. 2014). A
more recent date, also on unidentified charcoal (Sa–196) collected by J. Roche
from the upper layers in the 1960s indicates a more recent activity at the site, with
a calibrated date range of 4706–3196 cal BCE (95% confidence), but its association
with the burials is unknown.
The new 14C dates strengthen the chronology of the main period of burial activity between c 6000–5600 cal BCE. The new dates also support the previous estimates for the start and end of the burial activity at the site. The time span for the
use of the site for burial practices is not altered but refined. The posterior density
estimated for the start of the burial activity at Cabeço da Arruda is 6270–5918 cal BCE
(95% probability), during the first few centuries of the Late Mesolithic in Portugal,
when the estuarine environment is well established in the region (c 6100 cal BCE,
van der Schriek et al. 2007). The posterior density estimated for the end of the burial
activity is 5424–4843 cal BCE (95% probability) suggesting that the burial activity at
Cabeço da Arruda ends as saltmarshes gradually contracted (c 5555 cal BCE,
van der Schriek et al. 2007). This latest period of burial activity at the site, represented by the date of one later burial (CAR19th c., SkN, TO–356), overlaps with
the end of the Late Mesolithic (c 5250 cal BCE) and the Early Neolithic in
Portugal (c 5500–4500 cal BCE, Carvalho 2010).
Alternative model
As a working hypothesis, a second chronological model could be proposed by
explicitly defining the main phases of burial activity in relation to the stratigraphic
sequence. The stratigraphic data of the relative position of the burials could be
included as the prior beliefs (or, the archaeology in a Bayesian Model) for the construction of a chronological model. In this model, individual CA–00–02
(TO–10216) and CAR1964, Sk4 (Ua–47976), in the bottom of the shell midden,
would inaugurate the burial activity at the site. CA–00–01 (TO–10217) c 2 metres
from the bottom would be included in a last phase of activity (fig. 5.8). The burials
found at different depths could be grouped in one or two middle phases, according
to their stratigraphic position. However, the calibrated date ranges and the posterior density estimates (model 1, fig. 4.14) show that depth and antiquity of the
burial are not entirely correlated. This is the case for CA–00–02 (TO–10216)
found in the base of the midden. This burial in the basal sand, presents a relatively
more recent 14C date than CAR1937, Sk6 (Beta–127451, AA–101343) which was
found 0.80 metres from the base, as well as more recent than CAR1937, Sk3
(UA–46273) at 1.20 metres from bottom. Other dated burials, however, seem to
follow a vertical chronology, with individual CA–00–01 (TO–10217) at 2 metres
from the bottom presenting one of the most recent dates known at the site. However, without any further information, such as the description of the stratigraphic
layers, the depth of the burial cannot be assumed to be indicative of relative
chronology.
382
Analysis and Synthesis of the Results
Figure 5.8. Position of the burials excavated in 1937 and 1964–65 relatively to the bottom of
the shell midden, Cabeço da Arruda, Tagus valley. The exact depth of the burials excavated
in the nineteenth century is unknown (dashed rectangles) and their stratigraphy is assumed
from the characteristics of the enveloping sediment. (Data from Cardoso and Rolão 1999–
2000; Jackes and Lubell 2012; Jackes et al. 2013; Roche 1974; Roksandic 2006)
Chronology of burial activity: implications for the local setting, Tagus
valley
The chronology of the burial activity proposed for the shell midden sites of the
Tagus valley is based on the data available for the sites under study, which are also
the largest sites in the region: Moita do Sebastião and Cabeço da Arruda. For a
more robust analysis of the local setting, I introduce in this section a supplementary data set of the published 14C dates on human bone collagen from the
neighbouring site Cabeço da Amoreira. The data available for the other sites with
human remains at the Tagus valley is extremely limited and does not allow further
observations at this level.
The new 14C dates suggest that the first human burials in the Muge region were
placed at Moita do Sebastião between 6365–6051 cal BCE (posterior density estimate for
the start of burial activity, 95% confidence) as indicated by the burial of MS1952–54, Sks
9, 1. These are followed closely by the burial of CAR1937, Sk6 at Cabeço da
Arruda on the other margin of the river, indicating the start of the activity at the
site at 6270–5918 cal BCE (posterior density estimate for the start of burial activity, 95%
confidence).
This earlier period of burial activity is followed by the most frequent period of
burial practice in the Muge region. The burial activity concentrates in time and
space at Moita do Sebastião and Cabeço da Arruda between c 6000–5600 cal BCE.
The burial activity is frequent and synchronous in both sites (fig. 5.9).
383
Chapter 5
Figure 5.9. Chronological models for the burial activity at Moita do Sebastião and Cabeço da
Arruda, calibrated using marine 1 diet values. Shade: highlights the period of most frequent
burial activity c 6000–5600 cal BCE.
384
Analysis and Synthesis of the Results
The data is less robust for Cabeço da Amoreira (tables A5, A6) but here, the burial
activity is less frequent (fig. 5.10) and apparently less continuous when compared
with Moita do Sebastião and Cabeço da Arruda. The oldest human remains known
at Cabeço da Amoreira were excavated in 2000 (CAM–00–01, TO-11819–R) on
the basal sand layers of the site (Jackes and Lubell 2012; Meiklejohn et al. 2009a).
The posterior density estimated for the time of this burial is 6048–5702 cal BCE
(95% probability) which is coeval with the beginning of the frequent burial activity at
both Moita do Sebastião and Cabeço da Arruda. However, another measurement
from the same context of human remains presents a much later date (CAM–00–01,
TO–10218) (Roksandic 2006) with a posterior density estimate of 5508–5224 cal
BCE (95% probability). These disparate dates suggest mixing in the deposit, possibly
taphonomic. A single burial from Cabeço da Amoreira (burial 2011.1) is coeval
with the main period of burial activity at Moita do Sebastião and Cabeço da
Arruda, although in its later stages. This burial was excavated from the upper layers
of the midden (Wk–32143, Bicho et al. 2013) and presents a posterior density
estimate of 5878–5633 cal BCE (95% confidence).
Figure 5.10. Chronological model for the burial activity at Cabeço da Amoreira, calibrated
using marine 1 diet values. For each of the dates two distributions have been plotted, one
in outline which is the result produced by the scientific evidence alone, and a solid one
which is based on the chronological model used. Calibrated ranges of 95% and 68% probability are given by the lower and higher square brackets respectively. The large square
brackets (left) and the OxCal keywords define the overall model. Model-definition command files in appendix, Tagus valley, Cabeço das Amoreira, Model 1, Marine 1.
385
Chapter 5
The burial activity continues between c 5600–5400 in the three sites but it is less
frequent. The activity apparently ends at Moita do Sebastião, continuing episodically at Cabeço da Arruda and Amoreira. The last burials at Arruda and Amoreira
are coeval with the gradual contraction of estuarine habitats (starting c 5500 until
3800 cal BCE, van der Schriek et al. 2007). After the regional establishment of an
open landscape in the region around 5000 cal BCE (van der Schriek et al. 2007)
these sites are no longer used for burial practices. It is important to remember that
the upper layers of Moita do Sebastião were destroyed just before it was excavated
by J. Roche and O.V. Ferreira between 1952–54. Nevertheless, this material
represents a strong chronological framework. The nineteenth century excavations
at Moita do Sebastião were done before the destruction of the site and these dates
are coeval and consistent with the material excavated by J. Roche after the destruction of the top layers. This is further consistent with the suggestion that most nineteenth century material was excavated from the basal layers (Jackes and Alvim
2006).
In summary, in the Muge region, the first human burials are placed at Moita do
Sebastião, and are immediately followed by a first burial at Cabeço da Arruda on
the opposite margin of the river, in the transition between the 7th and 6th millennium cal BCE. During the first half of the 6th millennium cal BCE, between
c 6000–5600 cal BCE, the burial activity is frequent at these two sites, with occasional burials at Cabeço da Amoreira, near Moita do Sebastião. The burial activity
ceases at Moita while the estuarine environment is still predominant in the region.
It continues at Cabeço da Arruda and Amoreira but in a less frequent fashion. The
last burials known were placed in these two sites by the end of the 6th millennium,
coinciding with a decline in the tidal influence, gradually shifting from estuarine to
a fluvial landscape (van der Schriek et al. 2007, 381). Overall, the burial activity is
estimated to have continued in the Muge region for between 773 and 1113 years
(95% probability), or for 846 and 1005 years (68% probability) (fig. 5.11).
This updated chronology for the burial activity at the Muge sites in the Tagus
valley does not contradict the current chronological framework established for the
Late Mesolithic in Portugal (c 6350–5250 cal BCE, Araújo 2015), but strengthens
it.
386
Analysis and Synthesis of the Results
Figure 5.11. Chronological model for the burial activity at Moita do Sebastião, Cabeço da
Arruda and Cabeço da Amoreira in the Tagus valley. For each of the dates two distributions have been plotted, one in outline which is the result produced by the scientific evidence alone, and a solid one which is based on the chronological model used. Calibrated
ranges of 95% and 68% probability are given by the lower and higher square brackets respectively. The large square brackets (left) and the OxCal keywords define the overall
model. Model-definition command files in appendix, Tagus valley, Moita do Sebastião, Cabeço
da Arruda and Cabeço da Amoreira, Model 1, Marine 1.
387
Chapter 5
Arapouco, Sado valley
Antiquity of the site
There is one reliable 14C date only on human bone collagen from Arapouco
(ARA1962, Sk2A, Sac–1560, Cunha and Umbelino 2001). This 14C date has a
probability density estimate of 6232–5713 cal BCE (95% confidence) which is
consistent with a previous measurement on shells from the middle layers excavated
in 1961–62 with a calibrated date range of 6374–5706 cal BCE (95% confidence)
(table A8), and confirms the early chronology of the site. The association of these
shells with the human burials is unknown.
Burial activity: duration, start/end, frequency
Current 14C data does not allow the estimation of duration, start/end, and frequency of the burial activity at Arapouco. This is a large site with a MNI of 32, and
despite the poor collagen preservation at this site further 14C measurements should
be tried in the future.
Cabeço das Amoreiras, Sado valley
The chronological model proposed for the burial activity at Cabeço das Amoreiras
is presented in chapter 4 and is based on three 14C measurements on human remains. Despite the low number of dates this model can be considered representative of this assemblage, because the eight individuals recovered from this site (one
of which is represented by an isolated cranium) were found in one stratigraphic
layer (4, brownish sand) and lie in very close proximity.
Antiquity of the site
The new 14C dates do not confirm the antiquity of the burial activity at Cabeço das
Amoreiras suggested by the burial previously dated (Beta–125110, individual 5).
Burial ground during the Neolithic?
New measurements reveal a later use of the site for burial purposes, and suggest
that the burial episodes were spread over time. Despite the apparent scattering, all
dates introduced in the model show good agreement and the model did not consider any outliers. The measurements on human bone collagen strongly indicate
that the burial activity is a Mesolithic phenomenon.
The proposed model can account for the previous dates on non-human material (table A10). The measurement on shells (Q–AM85B2b, layer 2b, structure B)
with a calibrated date range of 5476–5221 cal BCE (95% confidence) is coeval with
the latest burial date available (Ua–47973, CAMS1958, Sk4). The measurement on
unidentified charcoal (Q–AM85B2a, layer 2a, structure B) is relatively recent
(5199–4707 cal BCE, 95% confidence). This charcoal was collected in the same
structure as the dated shells, and without further data this sample should be
388
Analysis and Synthesis of the Results
considered with caution. This later date on charcoal fits well with the few
fragments of Neolithic ceramic found at this site (Diniz 2010), but this does not
match the burial activity at the site.
Burial activity: duration, start/end, frequency
CAMS1958, Sk5 (Beta–125110) is the earliest burial known at the site with a posterior density estimate of 6156–5914 cal BCE (95% probability) (fig. 5.12). The latest
use of the site for burial practice is represented by CAMS1958, Sk4 (Ua–47973)
with a posterior density estimate of 5483-5329 cal BCE (95% probability), dated
several centuries after the first burial at the site. The start and end of the burial
activity at Cabeço das Amoreiras, represented by the dates of these two burials, is
well framed in the currently proposed dates for the Late Mesolithic in Portugal
(c 6350–5250 cal BCE; Araújo 2015).
The six to eight burials known at Cabeço das Amoreiras were excavated in a
small area of c 25 m2. The human remains were undisturbed and it was expected
that these burials would be coeval in time. The new 14C results are surprising,
showing a long duration of burial activity between 487 and 744 years (95%
probability). At Cabeço das Amoreiras, burials were not frequent but the practice
continued over a long period of time. It is also interesting that the burials were
found undisturbed in a small area which was apparently used for one episodic
burial every 100 years, or over four generations.
Figure 5.12. Cabeço das Amoreiras, Sado valley: site plan and 14C dates.
389
Chapter 5
Vale de Romeiras, Sado valley
The chronological model proposed for the burial activity at Vale de Romeiras is
presented in chapter 4 and is based on only two 14C measurements on human remains. Unfortunately, several samples submitted for analyses did not yield enough
collagen or presented results that did not assure the quality of the measurements.
This is a very small number of measurements compared with the MNI (26) buried
in this site. Nevertheless, these are the first successful 14C dates obtained on human
bone collagen from Vale de Romeiras, and although new measurements should be
tried in the future, the preliminary model presented here provides valuable indications about the burial activity at the site.
Burial activity: duration, start/end, frequency
The earliest period of burial activity known at Vale de Romeiras is represented by
the burial of the VR1959, Sk19 (Ua–46972) placed on top of the basal sand,
isolated and apparently outside the area of the shell midden (fig. 5.13). The
posterior density estimate for this burial is 6593–6370 cal BCE (95% probability) and
indicates an early use of the site for burial practices. The second date (Ua–47983,
VR1959, Sk9) is several centuries later, with a posterior density estimate of 5625–
5475 cal BCE (95% probability). This burial is in the shell midden area and it was
excavated also in the basal layer. These two 14C measurements indicate a long duration of burial activity of 779 to 1074 years (95% probability) but it is unknown if the
burial activity was more frequent during some periods of time or scattered over a
long period. Despite the apparent scattering, the two dates introduced in the
model show good agreement and the model did not indicate any outliers.
If the burial activity was scattered over time, this estimate suggests a frequency
of not more than two burials every two generations, based on the MNI (26) recovered at the site and assuming the widest span of burial activity of 1074 years.
However, the spatial relations of the burials and the homogeneity of the burial
modes seem to indicate otherwise. The burial frequency was likely higher during
certain period(s) of time, but further dates are required to support this hypothesis.
The start of the burial activity at Vale de Romeiras is represented by the burial
of VR1959, Sk19 (6593–6370 cal BCE, 95% probability) which is slightly earlier than
the currently proposed dates for the Late Mesolithic in Portugal (c 6350–5250
cal BCE). The end of the burial activity is represented by the burial of VR1959,
Sk9 (5625–5475 cal BCE, 95% probability) which is well positioned in the
chronological framework of the Late Mesolithic.
Previous 14C measurements on non-human samples from shells recovered from
the middle layers (Arnaud 2000) suggested an early use of the site, presenting
calibrated date ranges of 6416–6081 cal BCE (ICEN–150) and 6370–6076 cal BCE
(ICEN–146) (95% confidence) (table A11). Unfortunately, the relation of these
samples to the human burials is unknown.
390
Analysis and Synthesis of the Results
Figure 5.13. Vale de Romeiras, Sado valley: site plan and 14C dates.
Coeval group of burials?
It has been suggested that the burials at Vale de Romeiras were possibly coeval,
based on the spatial arrangement of the burials, organized in a semicircle. Also,
because of the assumed small size of the site, it was suggested that this place was
used as burial ground for the neighbouring site Cabeço do Pez (Arnaud 1989).
The 14C measurements indicate that the burials at Vale de Romeiras do not represent one coeval group, contra earlier suggestions (Arnaud 1989, 621) of a
synchronic burial ground. However, these dates do not invalidate the earlier
suggestion, which should be reformulated, to focus on possible synchronic burial
subgroups within the burial ground. The chronological relationship with Cabeço
do Pez is discussed below in Chronology of burial activity: implications for the local setting,
Sado valley.
Unexpected modern 14C date of a child
A third successful 14C measurement on human bone from this site (Ua–46968,
VR1959, Sk8) revealed a surprising calibrated date range of 1445–1632 cal CE
(95% confidence) (table A2) falling into the early modern period in Europe. For
this reason this measurement was not included in the model with the two Mesolithic dates. This individual is a child of 3–4 yrs ± 12 months and was interred in
the centre of the burial area in the second layer (grey soil with shells) (fig. 5.13).
A bone from this child was previously measured for 14C dating but did not yield
results (Cunha et al. 2002, 189). The recent sample (Ua–46968) of a tibia passed all
391
Chapter 5
quality checks and the quality of the measurements was further confirmed by the
laboratory.
Despite the very unusual practice, it is probably not so strange that a local child
of this time could be buried in a place like this. These inland sites have large concentrations of shells, and human bones are found occasionally on the surface.
Local populations would have known about these places, possibly recognized them
as ancient burial grounds, and incorporated them in local legends and mysticism.
In Portugal, in the fifteenth and sixteenth centuries, and until the mid-nineteenth
century, the dead were buried in religious buildings, and very few burials took place
outside these precincts. In rural areas, such as Vale de Romeiras, the majority of
the burials took place inside churches, and the open air burial of the poor who died
in charitable hospitals for example, was viewed as a great indignity (Queiróz and
Rugg 2003). The burial of this child in this most unusual place suggests a deviant
practice of possibly a non-Christian minority which at the time were highly
persecuted groups (Medina 2004). Nevertheless, it would be interesting to obtain a
new measurement on this individual in order to confirm this most unusual episode.
Cabeço do Pez, Sado valley
The chronological model proposed for the burial activity at Cabeço do Pez is
presented in chapter 4 and is based on five 14C measurements on human remains.
Two samples are from area T.A, two are from area T.B, and one is from an
isolated burial (fig. 5.14). This is a relatively low number of samples when compared with the MNI (n=32–36) buried in the site. Nevertheless, the four samples
represent a good spatial coverage of both areas T.A and T.B (1956) where most
individuals are closely placed. Also, the results obtained for the 14C measurements
are remarkably consistent, except for the sample collected from the individual
buried outside these two main clusters (Ua–46930, CP1956, Sk2), which presents a
much later date.
Chronological relationship of the burial areas and isolated burials
The four 14C measures obtained from the two main areas (T.A and T.B, 1956) are
coeval and present a calibrated date range of c 5700–5500 cal BCE (95% confidence), indicating a concentration in time and space of the burial activity at Cabeço
do Pez.
One sample was collected from one of the three burials (CP1956, Sk2,
Ua–46930) isolated outside the two main areas. The burial of this 1.5–2 year old
child indicates a more recent burial activity at the site with a posterior density estimate of 4461–4266 cal BCE (95% probability). The chronology of this burial is a
strong argument supporting the already suggested interpretation (Arnaud 2000, 32)
of Cabeço do Pez as a shell midden developing during the Late Mesolithic with
Neolithic episodes yet to be clearly defined.
392
Analysis and Synthesis of the Results
Figure 5.14. Cabeço do Pez, Sado valley: site plan and 14C dates.
393
Chapter 5
Burial activity: duration, start/end, frequency
When the 14C dates are plotted without the later date of Neolithic chronology
(Ua–46930, CP1956, Sk2), the model indicates that the Mesolithic duration of the
burial activity is remarkably short. In this case, the posterior density estimated for
the duration of the burial activity is between 0 and 123 years (95% probability), or
0 and 58 years (68% probability) (figs. 5.15–17; table 5.4). If the burial of CP1956, Sk2
(Ua–46930) is included in the model, the posterior density estimated for the duration of the burial activity is between 1192–1412 years (95% probability), or 1246 and
1359 years (68% probability) (as presented in chapter 4).
The posterior density estimated for the start of the burial activity at Cabeço do
Pez is coeval with the currently proposed dates for the Late Mesolithic in Portugal
(c 6350–5250 cal BCE).
At Cabeço do Pez, during the Mesolithic, the burial activity concentrates in a
very short period of time. The areas in use during the Mesolithic are located in two
main clusters, T.A and T.B, where at least 24 burials were excavated. The few isolated burials could be of Neolithic chronology, as suggested by CP1956, Sk2, but
this hypothesis must be confirmed with further data. At the moment, and assuming that the areas T.A and T.B were used during the Mesolithic for an estimated
maximum time span of 123 years, it may be assumed that there was at least one
burial every 5 years, indicating a high frequency of burial practice at the site.
Figure 5.15. Chronological model for the burial activity at Cabeço do Pez without the later
date Ua–46930 (CP1956, Sk2), calibrated using marine 1 diet values. For each of the dates
two distributions have been plotted, one in outline which is the result produced by the
scientific evidence alone, and a solid one which is based on the chronological model used.
Calibrated ranges of 95% and 68% probability are given by the lower and higher square
brackets respectively. The large square brackets (left) and the OxCal keywords define the
overall model.
394
Analysis and Synthesis of the Results
Figure 5.16. Duration of burial activity at Cabeço do Pez derived from the chronological
model run without Ua–46930, CP1956, Sk2, calibrated with values for marine 1 diet. Posterior density estimated for the duration of the burial activity is between 0 and 123 years (95%
probability), or 0 and 58 years (68% probability). Calibrated ranges of 95% and 68% probability
are given by the lower and higher square brackets respectively.
Figure 5.17. Start and End of burial activity at Cabeço do Pez derived from the chronological model run without Ua–46930, CP1956, Sk2, calibrated with values for marine 1 diet.
Posterior density estimated for the start of the burial activity is 5826–5578 cal BCE (95%
probability) or 5706–5631 cal BCE (68% probability). Posterior density estimated for the end
of the burial activity is 5678–5446 cal BCE (95% probability) or 5649–5565 cal BCE (68%
probability). Calibrated ranges of 95% and 68% probability are given by the lower and higher
square brackets respectively.
Table 5.4. Posterior density estimates for the dates of archaeological events (Start/End) and the duration
of burial activity at Cabeço do Pez. Data derived from model described in fig. 5.15.
Start of burial activity
Modelled (cal BCE)
Marine 1
Posterior density estimate
(68% probability)
5706–5631
Modelled (cal BCE)
Marine 1
Posterior density estimate
(95% probability)
5826–5578
End of burial activity
5649–5565
5678–5446
Span of burial activity, duration
Overall model agreement
(Aoverall)
0–58 yrs
0–123 yrs
122%
122%
Distribution
395
Chapter 5
The 14C measurements on human bone are consistent with the previous dates on
shell samples excavated from the middle layers in 1983 (Arnaud 1989) with a
calibrated date range of 5750–5494 cal BCE (Q–2497) and 5518–5299 cal BCE
(Q–2496) (95% confidence) (table A12). The relation of these shell samples with
the burials is unknown.
At Cabeço do Pez, the burial activity during the Mesolithic is short but frequent
between 5826–5578 and 5678–5446 cal BCE (95% probability), with a frequency of one
burial every five years. The data indicates that the site was also used during the
Neolithic for burial purposes, as indicated by the date of the isolated burial
CP1956, Sk2. The later activity at the site needs further research and the dating of
CP1956, Sks 1, 3 also outside the two main clusters may eventually clarify this later
episodic and apparently spread out burial activity. Nevertheless, the current model
seems to be representative of the burial activity at Cabeço do Pez with one main
period during the Late Mesolithic and a more episodic activity during the
Neolithic.
If Cabeço do Pez was used as the burial ground of one residential group of
hunter-gatherers (n=15 to 17), with a death rate of 0.3 or 0.5 per year (see chapter
2, Hunter-gatherer death rate vs burial population), the burial ground would aggregate in
123 years, c 37 or 62 human burials, respectively. The current lowest MNI (32)
accounts for 87% or 52% of this hypothetical burial population (table 5.5). This
estimate allows that some individuals may be buried in the site much later
(maximum MNI 36).
Table 5.5. Representativeness of the archaeological burial population at Cabeço do Pez in relation to the
original group. Estimates assume a residential group unit of nomadic foragers of 15 to 17 people (after
Binford 2001; Hamilton et al. 2007). Death rate estimates after M. Gurven and H. Kaplan (2007).
Death rate/year
Burial activity span
Total burial population
MNI 32
0.3
123 yrs
37
87%
0.5
123 yrs
62
52%
Várzea da Mó, Sado valley
One human burial was excavated at this site (VM1959, Sk1). The burial of this
individual has a probability density estimate of 5345–5071 cal BCE (95% confidence), which is coeval with the later stages of the currently proposed dates for the
Late Mesolithic in Portugal (c 6350–5250 cal BCE).
A 14C date on a sample of shells (ICEN–273, table A13) recovered from the
middle layers excavated in 1959, has a calibrated date range of 6067–5892 cal BCE
(95% confidence) and indicates an earlier use of the site. The relation of this
deposit of shells and the human burial is unknown.
396
Analysis and Synthesis of the Results
Chronology of burial activity: implications for the local setting, Sado
valley
The chronology of the burial activity proposed for the shell midden sites of the
Sado valley is based on the data available for: Arapouco, Cabeço das Amoreiras,
Vale de Romeiras, Cabeço do Pez, and Várzea da Mó (fig. 5.18). The 14C data
available for Poças de S. Bento (MNI 15) consists of one measurement on human
bone of uncertain calibration (Ua–425, PSB1986, Skull) and several non-human
samples (table A14). The chronological interpretation of the burial activity at this
site is limited, but the data is included here for a more robust analysis at the level
of the local setting.
The new 14C dates suggest that the first human burials in the Sado valley were
placed at Vale de Romeiras, as indicated by the burial of VR1959, Sk19 (Ua–47983,
6593–6370 cal BCE, 95% probability). This is an isolated early date (fig. 5.18) followed by burials at Arapouco (ARA1962, Sk2A, Sac–1560, 6232–5713 cal BCE,
95% confidence) and at Cabeço das Amoreiras (CAMS1958, Sk5, Beta-125110,
6156–5914 cal BCE, 95% probability).
These earlier burials are followed by the most active period of burial practice in
the Sado valley. Between c 5750–5500 cal BCE the burial activity is frequent at
Cabeço do Pez, and is coeval with the burials of VR1959, Sk9 (Ua–47983) in the
neighbouring site of Vale de Romeiras, and with the burial of CAMS1959, Sk7
(Ua–47974) at Cabeço das Amoreiras.
The only burial known at Várzea da Mó (VM1959, Sk1, Ua–46310) is also the
last burial of Mesolithic chronology known at the Sado valley (5345–5071 cal BCE,
95% confidence). This burial is later than any known Mesolithic burial activity in
the valley, but future data from Vale de Romeiras and Poças de S. Bento may
eventually change this scenario.
Cabeço do Pez is also used for burial practices during the Neolithic. This also
seems to be the case at Poças de S. Bento (Larsson 2010). At Cabeço do Pez, this
is suggested by the burial of a child, outside the main burial areas, presenting a
posterior density estimate of 4461–4266 cal BCE (95% probability), coeval with the
early stages of the Middle Neolithic in Portugal (c 4500–3500 cal BCE, Neves and
Diniz 2014). At Poças de S. Bento, burial practices coeval with the Neolithic are
indicated by the 14C of a skull excavated in 1986 in the basal sand (Larsson 2010).
The calibration of this measurement is unreliable due to the lack of stable isotopic
measurements. However, its calibrated date range should be around 4400–3800
cal BCE (95% confidence), depending on the higher or lower offset (see chapter 2,
Poças de S. Bento).
397
Chapter 5
Figure 5.18. Chronological model for the burial activity at Arapouco, Cabeço das Amoreiras,
Vale de Romeiras, Cabeço do Pez and Várzea da Mó at the Sado valley. Shade: highlights
the most frequent burial activity c 5750–5500 cal BCE. For each of the dates two distributions have been plotted, one in outline which is the result produced by the scientific evidence alone, and a solid one which is based on the chronological model used. Calibrated
ranges of 95% and 68% probability are given by the lower and higher square brackets respectively. The large square brackets (left) and the OxCal keywords define the overall
model. Model-definition command files in appendix, Sado valley, human remains, except VR8
(Ua–46968) and CP2 (Ua–46930), Model 1, Marine 1.
In summary, in the Sado valley, the first human burial is placed at Vale de
Romeiras in the very early stages of the Late Mesolithic. Only a few centuries later,
new burials take place at Arapouco and Cabeço das Amoreiras, also isolated
chronologically. The most frequent period of burial activity takes place between
c 5750–5500 cal BCE with burials mainly at Cabeço do Pez, but also at Vale de
Romeiras and Cabeço das Amoreiras. After c 5500 the burial activity is less frequent, and one burial is known at Cabeço das Amoreiras, followed by the last
burial of Mesolithic chronology at Várzea da Mó.
398
Analysis and Synthesis of the Results
There is a lack of dates for Arapouco which could arguably strengthen the
chronology of the earlier phases of burial activity in the valley. Also, further 14C
dates at Vale de Romeiras could clarify the chronology of this site, suggested to be
the Cabeço do Pez burial ground (Arnaud 1989, 621).
This updated chronology for the burial activity at these four sites does not contradict the current chronological framework established for the Late Mesolithic in
Portugal (c 6350–5250 cal BCE; Araújo 2015), rather it strengthens it. Nevertheless, Cabeço do Pez presents the interesting case of two moments of burial activity:
one during the Mesolithic and one coeval with the early stages of the Middle
Neolithic in Portugal. Further dates on the isolated burials found in this site could
potentially strengthen the case for the Neolithic use of the site for burial practice.
This is possibly also the case at Poças de S. Bento, but the 14C data on human
remains is insufficient for further analysis of the burial activity at the site.
Tagus vs Sado
At both Tagus and Sado valleys, the burial activity is more frequent between
c 6000 and 5500 cal BCE. This is well attested by the burial frequency during this
time period at Moita do Sebastião (c 6000–5600 cal BCE), Cabeço da Arruda
(c 5900–5700 cal BCE), and Cabeço do Pez (c 5600 cal BCE). Based on the spatial
distribution of the burials and on the homogeneity of the burial modes, I suggest
that this could also be the case at Arapouco and Vale de Romeiras, although at
present there are few 14C measurements to confirm this hypothesis.
In both valleys, we find early episodes of human burial activity, during the very
first phases of the Late Mesolithic in Portugal (c 6350 cal BCE). The earliest
human burial known in any of these shell middens (VR1959, Sk19, 6593–6370
cal BCE) is found at Vale de Romeiras in the Sado valley. Early burials are found
also at Moita do Sebastião in the Tagus valley, although with slightly later posterior
density estimate (MS1952–54, Sk9, 6218–5998 cal BCE; MS1952–54, Sk1, 6218–
5998 cal BCE). These first episodes of burial activity are followed by the earliest
burials known at Arapouco (Sado, ARA1962, Sk2A, 6232–5713 cal BCE), Cabeço
das Amoreiras (Sado, CAMS1958, Sk5, 6156–5914 cal BCE) and Cabeço da Arruda
(Tagus, CAR1937, Sk6, 6133–5852 cal BCE).
Overall, the sites are used for burial practices more or less synchronously. The
one exception is Várzea da Mó at the Sado valley, with one burial known, coeval
only with the later burial activity at Cabeço da Amoreira and Cabeço da Arruda in
the Tagus valley.
After c 5500 cal BCE, the frequency of burial practice decreases dramatically, in
both the Tagus and Sado valleys. At this time period, the burial activity seems to be
more active at Cabeço da Amoreira and Cabeço da Arruda in the Tagus valley,
with two coeval burials known at the Cabeço das Amoreiras and Várzea da Mó in
the Sado valley.
In both valleys, certain sites such as at Moita do Sebastião and Cabeço da
Arruda in the Tagus valley and Cabeço do Pez in the Sado valley, show frequent
399
Chapter 5
and continuous burial activity over a relatively short period of time. Other sites,
such as Cabeço da Amoreira in the Tagus valley or Cabeço das Amoreiras at Sado,
present a more scattered activity pattern along with lower frequency of burials,
with a more or less continuous burial activity over a long period of time. Arapouco
and Vale de Romeiras in the Sado valley present a high spatial concentration of
human burials, however, the current chronological data do not allow further
observations at this level.
Burial activity is a Late Mesolithic phenomenon at both the Tagus and Sado
valleys. Two exceptions are presented by the burial of a child at Cabeço do Pez
(CP1956, Sk2, 4461–4266 cal BCE), and one adult (?) at Poças de S. Bento
(PSB1986, Skull, c 4400–3800 cal BCE) which are coeval with Middle Neolithic
groups in Portugal. Another interesting case of the later use of these sites for
human burial is given by the early modern burial of a child at Vale de Romeiras
(VR1959, Sk8, 1445–1632 cal CE).
In terms of chronological patterns, the burial activity at the Tagus and Sado
valleys is coherent, although much more frequent at Moita do Sebastião and
Cabeço da Arruda in the Tagus valley, possibly related to differential demography.
This updated chronology for the burial activity at these sites does not contradict
the current chronological framework established for the Late Mesolithic in
Portugal, rather it strengthens it.
400
Analysis and Synthesis of the Results
Figure 5.19. Chronological model for the burial activity at the Tagus (Muge) and Sado valleys. For each of the dates two distributions have been plotted, one in outline which is the
result produced by the scientific evidence alone, and a solid one which is based on the
chronological model used. Calibrated ranges of 95% and 68% probability are given by the
lower and higher square brackets respectively. The large square brackets (left) and the
OxCal keywords define the overall model. Model-definition command files in appendix,
Tagus and Sado valleys, human remains, Model 1, Marine 1.
401
Chapter 5
Isotopic signatures: defining the people in each burial place
Moita do Sebastião, Tagus valley
At Moita do Sebastião, the human collagen samples (n=21/no. individuals=18)
display stable isotope values ranging between –18.4‰ and –15.3‰ (–16.9 ± 0.7‰,
mean ± SD) for carbon, and +9.0 to +14.0‰ (+11.1 ± 1.3‰) for nitrogen
(n=20/no. individuals=18) (table 5.6). These values are indicative of the consumption of a mixture of marine-estuarine and terrestrial food resources, although with
significant intra-individual variation among some individuals. Some individuals
were sampled more than once and present slightly different stable isotope values.
However, if the multiple measurements (Umbelino 2006) are excluded, the ranges
remain the same and the means are not statistically different: –17.0 ± 0.8‰ for
δ13C and +11.0 ± 1.4‰ for δ15N. For consistency in the case of multiple sets of
values, the discussion is based on the individual values obtained in the course of
this study (Ua–number).
The highest δ15N values derive from three individuals: MS19th c., SkCT; 1952–
54, Sks 22, 25. Two are children under two years old (Sks 22, 25) and their elevated
δ15N values (+12.2, +14.0) may be explained as a result of bioenrichment from
breast milk consumption (Fogel et al. 1989). However, individual CT is an adult
male (Mary Jackes, pers. comm.) and his high δ15N value (+13.4) is clear outlier to
the Moita do Sebastião series. The mean δ15N value calculated without these three
individuals is +10.6 ± 1.0‰ presenting slightly smaller standard deviation,
however, still moderately high.
The δ13C and δ15N values fluctuate over time and do not follow any particular
tendency (figs. 5.20, 5.21). The earliest burials at Moita do Sebastião (MS1952–54,
Sks 9, 1) present equivalent stable isotope values (table 5.6), also consistent with
the mean values of the series. The δ13C values derived from the individuals buried
during the main period of burial activity (n=13) are consistent with the mean
values of the series and present small intra-individual variation (–16.9 ± 0.5‰).
The δ15N values present larger standard deviation and indicate moderate intraindividual variation (+11.1 ± 1.4‰). As discussed, the highest δ15N values are
derived from the two children in the group (MS1952–54, Sks 22, 25) possibly due
to a breastfeeding effect. However, even when these two individuals are excluded
from the group the standard deviation is still moderately high (+10.7 ± 1.0). The
latest period of burial activity at the site is represented by three individuals
(MS19th c., SkCT; 1952–54, Sks 10, 33). Two of these individuals (Sks 10, 33)
present slightly lower δ13C values indicating a group with higher consumption of
terrestrial foods. In contrast, the values derived from individual CT are obtained
from marine-estuarine sources. This distinct isotopic profile (MS19th c., SkCT)
presents the highest set of values for δ13C and δ15N among the adults.
Overall, the intra-individual variation at Moita do Sebastião is low to moderate
suggesting that the people buried in this place had a fairly similar pattern of protein
402
Analysis and Synthesis of the Results
consumption. Towards the end of the burial activity the isotopic signatures are
more heterogeneous and the individuals buried during this later stage present not
only the lowest but also the highest δ13C values of the series. The new δ13C value
obtained for MS1952–54, Sk10 (–18.2, Ua–47978) is slightly lower than the
previous value available for this same individual (–16.6, Umbelino 2006). While the
latter value is not in contrast with most of the series, the new value is. This difference could be explained by contamination of the sample or other difficulties
during the measuring process. The C:N ratio for the new measurement (Ua–47978,
C:N=3.4, δ13C= –18.2) is acceptable but it is at the limit of the quality parameter
(2.9–3.4/3.6) (see chapter 3, Reliability of the isotopic measurements). Unfortunately, the
C:N parameter is not available for the earlier measurements of this skeleton
(Umbelino 2006), and for this reason the new measurements are those used in the
discussion of the results.
One 14C measurement (Beta–127499, Cunha and Cardoso 2002–03; Cunha et al.
2003) was excluded from this study because the only δ13C value available was obtained on an AMS instrument during the radiocarbon dating process making the
calibration of this 14C date highly uncertain. Nevertheless, the original AMS
measurement of –16.8‰ for carbon (Beta–127499), is in line with the current
isotopic series for Moita do Sebastião suggesting possible low offset from an
isotope-ratio mass spectrometry (IRMS) measurement, which must be confirmed
with future measurements.
403
♂
♀
♂
♂
♂
Adult
Adult, 35–50 yrs
Adult, 25–35 yrs
Adult
Adult
Adult
Non-adult
CT, 19th c.
10, 1952–54
15, 1952–54
30, 1952–54
18, 1952–54
17, 1952–54
22, 1952–54
n/d
♂
♀
Adult
33, 1952–54
Sex
Age
Individual
Identification of the human remains
Ua–46267
Ua–47980
Ua–46266
Ua–47981
Ua–46265
Ua–47978
TO–135
Ua–46270
Lab no.
Age BP
7120±43
7105±42
7095±45
7058±44
6986±40
6753±46
6810±70
6743±44
14C
Measurements (IRMS)
5965–5655
5898–5639
5896–5636
5887–5632
5843–5611
5644–5473
5642–5421
Post. density estimate
(95% probability)
5640–5474
–16.9,
this study
–16.8,
this study
–16.9,
this study
–17.4,
this study
–17.9,
this study
–16.2,
Umbelino 2006
–18.2,
this study
–16.6,
Umbelino 2006
–15.3,
Lubell et al. 1994
–18.4,
this study
δ13C (‰)
12.8,
this study
9.5,
this study
10.8,
this study
9.8,
this study
10.2,
this study
9.0,
this study
11.5,
Umbelino 2006
13.4,
Lubell et al. 1994
10.5,
this study
δ15N (‰)
continued...
Main
Main
Main
Main
Main
Late
Late
Late
Burial activity
Table 5.6. Moita do Sebastião, Tagus valley: chronology and stable isotope values (δ13C and δ15N) of human bone collagen. Samples are sorted so that the most recent dates
are at the top of the column.
♀
n/d
♂
n/d
♀
♀
n/d
Adult
Adult, old
Adult
Adult
Non-adult
Adult
Adult
Adult
Adult
31, 1952–54
34, 1952–54
41, 19th c.
24, 19th c.
25, 1952–54
29, 19th c.
22, 19th c.
1, 1952–54
9, 1952–54
♂
♂
♂
Adult
5, 1952–54
Table 5.6. continued
Ua–46264
Ua–46263
TO–131
TO–133
Ua–46268
TO–132
TO–134
Ua–46271
Ua–46269
Ua–47977
7621±50
7483±48
7240±70
7200±70
7243±45
7180±70
7160±80
7236±41
7141±40
7138±42
6292–6023
6218–5998
6021–5705
6017–5707
5998–5736
5997–5681
5992–5652
5991–5734
5979–5718
5972–5672
–17.0,
this study
–17.0,
this study
–16.1,
Lubell et al. 1994
–16.9,
Lubell et al. 1994
–16.2,
this study
–16.8,
Lubell et al. 1994
–16.7,
Lubell et al. 1994
–16.3,
this study
–17.4,
this study
–16.7,
Umbelino 2006
–16.9,
this study
10.8,
this study
10.5,
this study
12.2,
Lubell et al. 1994
10.4,
Lubell et al. 1994
14.0,
this study
11.9,
Lubell et al. 1994
11.2,
Lubell et al. 1994
11.9,
this study
10.6,
this study
11.2,
Umbelino 2006
9.3,
this study
Early
Early
Main
Main
Main
Main
Main
Main
Main
Main
Chapter 5
-12
-13
Bone collagen δ13C (‰)
-14
-15
-16
-17
-18
-19
-20
-21
6400 6300 6200 6100 6000 5900 5800 5700 5600 5500 5400 5300
cal BCE 95%
Figure 5.20. Moita do Sebastião, Tagus valley: variation of δ13C (‰) over time (cal BCE,
95% probability) of 21 samples of human bone collagen from 18 individuals. Individuals 10
(circle), 15 (triangle) and 31 (open square) have two sets of δ13C values.
16
15
14
Bone collagen δ15N (‰)
13
12
11
10
9
8
7
6
6400 6300 6200 6100 6000 5900 5800 5700 5600 5500 5400 5300
cal BCE 95%
Figure 5.21. Moita do Sebastião, Tagus valley: variation of δ15N (‰) over time (cal BCE,
95% probability) of 20 samples of human bone collagen from 18 individuals. Individuals 10
(circle) and 31 (open square) have two sets of δ15N values. The non-adults are indicated by grey
square.
406
Analysis and Synthesis of the Results
Cabeço da Arruda, Tagus valley
At Cabeço da Arruda, the human collagen samples (n=16/no. individuals=14)
display stable isotope values ranging between –19.0‰ and –15.3‰ (–17.3 ± 1.0‰)
for carbon, and +9.5 to +12.0‰ (+11.0 ± 0.9‰) for nitrogen (n=16/no. individuals=14) (table 5.7). These values are indicative of the consumption of a mixture of marine-estuarine and terrestrial food resources, although with chronological
and intra-individual variation. Some individuals were sampled more than once and
present slightly different stable isotope values. However, if the multiple measurements (Umbelino 2006) are excluded, the ranges remain the same and the means
are not statistically different: –17.4 ± 1.0‰ for δ13C and +10.9 ± 0.9‰ for δ15N.
For consistency in the case of multiple sets of values, the discussion is based on
the individual values obtained in the scope of this study (Ua–number).
The δ13C variation over time is not linear and does not follow any particular
tendency (fig. 5.22). The δ13C values are more homogeneous during the early/main
burial activity and more heterogeneous within the main/later burials. In comparison, the δ15N variation is more homogeneous over time (fig. 5.23). The earliest
burial at Cabeço da Arruda (CAR1937, Sk6) presents a slightly higher marineestuarine diet component than the mean values of the series. Stable isotope values
derived from the individuals (n=11) buried during the main period of burial activity are consistent with the mean values of the series and present moderate intraindividual variation. Isotopic signatures are –17.6 ± 0.8‰ and +10.8 ± 0.9‰ for
carbon and nitrogen, respectively. The individuals with the oldest dates in this
group present coherent δ13C values and slight variation regarding the δ15N values.
Following burial episodes, samples dating to between c 5800–5700 cal BCE present the greatest intra-individual variation, suggesting a heterogeneous diet among
these individuals. One adult male (CAR19th c., SkA) presents the lowest δ13C value
(–19.0‰) paired with a relatively high δ15N value (+12.2‰). This set of values is
unusual at Cabeço da Arruda and suggests a significant consumption of freshwater
foods. Alternatively, these values could be explained as the result of contamination
of the sample, showing an anomalous negative value for carbon paired with a more
positive value for nitrogen (see chapter 3, Reliability of the isotopic measurements).
However, the problem of contamination is difficult to confirm because the values
of carbon and nitrogen are in the normal ranges, although outlier to the series.
Also, the C:N ratio was reported to be within the accepted quality range (Lubell et
al. 1994, 204), although the precise values were not published. Thus, without further data, this set of values is accepted, and the hypothesis of a diet rich in
freshwater foods is put forward as a possible explanation for this particular isotopic profile. The last episode of burial activity at the site is represented by one old
adult male (CAR19th c., SkN). The stable isotopes derived from this individual are
the highest at Cabeço da Arruda and are clearly indicative of a high intake of
marine-estuarine foods.
Intra-individual variation at Cabeço da Arruda is moderately high, particularly
after c 5800 cal BCE. While most individuals present δ13C values between
407
Chapter 5
c –16.5‰ and –18.0‰, a smaller group presents highly variable values outside this
range. This suggests that some of the individuals buried at Cabeço da Arruda had
fairly different patterns of protein consumption, contributing to the isotopic
heterogeneity represented at the site.
408
Adult, old
N, 19th c.
♂?
♀
Adult
1, 1937
♀
Non-adult, 10–11 yrs
Adult
2, 1937
♂
III, 19th c.
Adult
A, 19th c.
♀
n/p
Adult, 20–24 yrs
42, 19th c.
♂
n/d
♂
Sex
CA–00–02, 2000 n/p
Adult
D, 19th c.
CA–00–01, 2000 Non-adult
Age
Individual
Identification of the human remains
Ua–46272
TO–360
TO–10216
Ua–47975
TO–354
TO–359
TO–355
TO–10217
TO–356
Lab no.
7166±41
6990±110
7040±60
7116±44
6970±60
6960±70
6780±80
6620±60
5975–5696
5972–5511
5914–5621
5905–5640
5902–5617
5827–5495
5734–5428
5602–5287
Age BP Post. density estimate
(95% probability)
6360±80
5464–4963
14C
Measurements (IRMS)
–16.6,
this study
–15.7,
Umbelino 2006
–17.7,
Lubell et al. 1994
–17.9,
Lubell et al. 1994
–16.7,
this study
–19.0,
Lubell et al. 1994
–17.2,
Lubell et al. 1994
–18.9,
Lubell et al. 1994
–18.1,
Lubell et al. 1994
–15.3,
Lubell et al. 1994
δ13C (‰)
11.2,
this study
12.0,
Umbelino 2006
11.2,
Lubell et al. 1994
10.6,
Lubell et al. 1994
9.5,
this study
12.2,
Lubell et al. 1994
11.8,
Lubell et al. 1994
10.3,
Lubell et al. 1994
10.5,
Lubell et al. 1994
12.5,
Lubell et al. 1994
δ15N (‰)
continued…
Main
Main
Main
Main
Main
Main–Late
Main–Late
Late
Latest
Burial activity
Table 5.7. Cabeço da Arruda, Tagus valley: chronology and stable isotope values (δ13C and δ15N) of human bone collagen. Samples are sorted so that the most recent dates
are at the top of the column.
n/d
n/d
Adult
Adult
Adult
Adult
9, 1937
4, 1964–65
10, 1937
6, 1937
♀?
♂
♂
Adult
3, 1937
Table 5.7. continued
R_Combine:
AA–101343
Beta–127451
Ua–46275
Ua–47976
Ua–46274
Ua–46273
7418±58
7263±46
7261±45
7200±41
7198±40
6133–5852
6049–5798
6021–5766
5998–5748
5993–5745
–16.6,
Jackes et al. 2014
–17.4,
this study
–17.2,
Umbelino 2006
–17.1,
this study
–17.4,
this study
–17.3,
this study
10.9,
Jackes et al. 2014
11.4,
this study
11.2,
Umbelino 2006
9.7,
this study
10.6,
this study
10.1,
this study
Early
Main
Main
Main
Main
Analysis and Synthesis of the Results
-12
-13
Bone collagen δ13C (‰)
-14
-15
-16
-17
-18
-19
-20
-21
6200
6000
5800
5600
5400
cal BCE 95%
5200
5000
4800
Figure 5.22. Cabeço da Arruda, Tagus valley: variation of δ13C (‰) over time (cal BCE, 95%
probability) of 16 samples of human bone collagen from 14 individuals. Individuals 10
(triangle) and 1 (circle) have two sets of δ13C values.
16
15
14
Bone collagen δ15N (‰)
13
12
11
10
9
8
7
6
6200
6000
5800
5600
5400
cal BCE 95%
5200
5000
4800
Figure 5.23. Cabeço da Arruda, Tagus valley: variation of δ15N (‰) over time (cal BCE,
95% probability) of 16 samples of human bone collagen from 14 individuals. Individuals 10
(triangle) and 1 (circle) have two sets of δ15N values.
411
Chapter 5
Cabeço da Amoreira, Tagus valley
Material from Cabeço da Amoreira was not sampled or analysed in the scope of
this study. However, the published data are important resources for comparison
with the data from Moita do Sebastião and Cabeço da Arruda, allowing a more
robust analysis of the local setting.
At Cabeço da Amoreira, the human collagen samples of 14C dated individuals
display stable isotope values ranging between –20.1‰ and –15.8‰ (–17.0 ± 1.8‰)
for carbon (n=5), and +8.2 to +13.9‰ (+11.8 ± 2.5‰) for nitrogen (n=4) (tables
5.8, 5.9, figs. 5.24, 5.25). If all isotopic data is included, dated and non-dated, the
mean value is –16.4 ± 1.5‰ for carbon (n=9), which is very similar to the mean
value obtained for the 14C dated individuals only. In the case of δ15N (n=8), the
standard deviation for the whole data set is slightly lower +12.0 ± 1.7‰, although
still high. This apparently high intra-individual variation depends on the values
derived from one individual, who presents the lowest isotopic values in the series.
The isotopic signature for this individual (CAM–01–01, bone 139) indicates a full
terrestrial diet in contrast to the remaining measurements of the series which
present a balanced consumption of marine-estuarine and terrestrial foods. If this
set of values is excluded the stable isotope values for the series range between
–16.9‰ and –15.8‰ for carbon and +11.9 and +13.9 for nitrogen. The mean
values for the dated series are very similar when all values are taken into account,
dated and non-dated: –16.3 ± 0.5‰ (n=4) or –16.0 ± 0.6‰ (n=8) for carbon and
+13.0 ± 0.8‰ (n=3) or +12.6 ± 0.7‰ (n=7) for nitrogen.
The intra-individual variation at Cabeço da Amoreira is relatively low suggesting
that the people buried in this place had a fairly similar pattern of protein
consumption. One individual (CAM–01–01, bone 139) is an outlier and presents a
radically different isotopic signature suggesting a very different pattern of protein
consumption.
Table 5.8. Cabeço da Amoreira, Tagus valley: stable isotope values (δ13C and δ15N) of human bone
collagen without associated 14C date. Samples are sorted so that the lowest δ13C values are on the top of the
column.
Identification of the human remains
Measurements (IRMS)
Individual
Age
Sex
Lab no.
δ13C (‰)
δ15N (‰)
C:N
Reference
7, 1930–33
Adult
n/d
n/a
–16.5
11.9
n/a
Umbelino 2006
4, 1930–33
Adult
n/d
n/a
–15.7
12.7
n/a
Umbelino 2006
8, 1930–33
Adult, mature n/d
n/a
–15.6
12.0
n/a
Umbelino 2006
6, 1930–33
Adult, mature n/d
n/a
–14.8
12.5
n/a
Umbelino 2006
412
n/d
♀
Adult
Adult, young
Non-adult
Non-adult
CAM–01–01. 2001
bone 139, main ind.
Burial 2011.2
Burial 2011.1
CAM–00–01, 2000
n/d
n/d
n/d
?
CAM–01–01. 2001
Sex
Age
Individual
Identification of the human remains
7300±80
7132±41
Wk–32143
TO–11819–R
6910±40
6550±70
Wk–32142
TO–10225
6329±40
Wk–26796
Age BP
14C
Lab no.
Measurements (IRMS)
6048–5702
5878–5633
5666–5466
5613–5296
Post. density estimate
(95% probability)
5292–4921
–16.4,
Meiklejohn et al. 2009a
–16.0,
Bicho et al. 2013
–15.8,
Bicho et al. 2013
–20.1,
Roksandic 2006
–16.9,
Bicho et al. 2011
δ13C (‰)
–
13.9,
Bicho et al. 2013
12.9,
Bicho et al. 2013
8.2 ,
Roksandic 2006
12.3,
Bicho et al. 2011
δ15N (‰)
Table 5.9. Cabeço da Amoreira, Tagus valley: chronology and stable isotope values (δ13C and δ15N) of human bone collagen. CAM–00–01, TO–11819–R replaces the
previous measurements TO–10218 (Meiklejohn et al. 2009a). Samples are sorted so that the most recent dates are at the top of the column.
Chapter 5
-12
-13
Bone collagen δ13C (‰)
-14
-15
-16
-17
-18
-19
-20
-21
6000
5800
5600
5400
cal BCE 95%
5200
5000
4800
Figure 5.24. Cabeço da Amoreira, Tagus valley: variation of δ13C (‰) over time (cal BCE,
95% probability) of 5 samples of human bone collagen from 5 individuals.
16
15
14
Bone collagen δ15N (‰)
13
12
11
10
9
8
7
6
6000
5800
5600
5400
cal BCE 95%
5200
5000
4800
Figure 5.25. Cabeço da Amoreira, Tagus valley: variation of δ15N (‰) over time (cal BCE,
95% probability) of 4 samples of human bone collagen from 4 individuals.
414
Analysis and Synthesis of the Results
Isotopic signatures and implications for the local setting, Tagus valley
Overall, the isotopic signatures for the individuals buried in the shell middens of
the Tagus valley show low to moderate intra-individual variability. The profiles
reflect the consumption of a mixture of marine-estuarine and terrestrial resources,
suggesting the consumption of foods that could have been collected by the shores
of the estuary near the shell midden sites, as well as the intake of terrestrial foods
that could have been collected in the immediate vicinity of the sites and/or further
inland. In fact, the isotopic signatures are derived mostly from terrestrial foods,
plants and animals. Few individuals (c n=8) derive their isotopes from the intake of
50% or more of marine-estuarine food sources. These results are in agreement
with previous trace element studies suggesting an overall inter-site homogeneity,
characterized by a mixed and diverse diet, based on marine-estuarine and terrestrial
resources including a range of faunal and floral foods (Umbelino 2006).
Samples from individuals dating to the main period of burial activity at the
Tagus valley are homogenous, and most individuals present similar isotopic
signatures. The values become more heterogeneous towards the end of the activity
(figs. 5.26, 5.27). Interestingly, only two individuals stand out as clear outliers to
the general pattern:
• one adult male buried at Cabeço da Arruda (CAR19th c., SkA, 5902–
5617 cal BCE) has a signature usually associated with considerable intake
of freshwater fish supplemented with terrestrial sources of protein
(δ13C= –19.0‰, δ15N= +12.2‰);
• one adult buried at Cabeço da Amoreira (CAM–01–01, bone 139,
5613–5296 cal BCE) has a typical terrestrial signature (δ13C= –20.1‰,
δ15N= +8.2‰).
415
Chapter 5
-12
MS
CAR
CAM
-13
Bone collagen δ13C (‰)
-14
-15
-16
-17
-18
-19
-20
-21
6400
6200
6000
5800
5600
5400
cal BCE 95%
5200
5000
4800
Figure 5.26. Moita do Sebastião, Cabeço da Arruda and Cabeço das Amoreiras, Tagus valley: bone collagen δ13C over time (cal BCE, 95% probability). At the Tagus valley, the main
period of burial activity is homogenous in terms of dietary sources, and most individuals
present similar isotopic signatures. Towards the end of the activity the values become more
heterogeneous. Dashed outline: the two outliers of the series (CAR19th c., SkA; CAM–01–01,
139). Individuals MS1952–54, Sks 10 (open square), 15 (grey square with outline), 31 (grey square),
and CAR1937, Sks 1 (open triangle), 10 (grey triangle with outline) have two sets of δ13C values.
16
MS
CAR
CAM
15
Bone collagen δ15N (‰)
14
13
12
11
10
9
8
7
6
6400
6200
6000
5800
5600
5400
cal BCE 95%
5200
5000
4800
Figure 5.27. Moita do Sebastião, Cabeço da Arruda and Cabeço das Amoreiras, Tagus valley: bone collagen δ15N over time (cal BCE, 95% probability). Dashed outline: the two outliers of the series (CAR19th c., SkA; CAM–01–01, 139). Individuals MS1952–54, Sks 10
(open square), 31 (grey square), and CAR1937, Sks 1 (open triangle), 10 (grey triangle with outline)
have two sets of δ15N values.
416
Analysis and Synthesis of the Results
When the isotopic data is analysed over time the homogeneity of the profiles
becomes more evident. During the earliest period of burial activity (fig. 5.28,
map 1), the individuals buried at Moita do Sebastião (n=2) and Cabeço da Arruda
(n=1) present equivalent isotopic signatures (–16.9 ± 0.2‰ for carbon and +10.7
± 0.2‰ for nitrogen), indicating no inter-site variability at the isotopic level.
At the start of the main burial activity at the Tagus valley (fig. 5.28, map 2) there
is a moderate variation in the trophic levels of the foods consumed, as suggested
by the nitrogen values (+9.7‰ to +12.2‰, +10.9 ± 1.1‰). In every case the intraindividual variation is lower than +2.0‰. At Moita do Sebastião, MS19th c., Sk22
consumed marine-estuarine foods on average one trophic level higher than
MS19th c., Sk29. The δ15N enrichment is just +1.8‰, which is lower than the
estimated trophic level enrichment of nitrogen for mammals (2 to 4‰, see chapter
3, Stable isotope analysis of carbon and nitrogen). Similarly, at Cabeço da Arruda the
isotopic signatures suggest that CAR1937, Sk10 consumed marine-estuarine foods
on average one trophic level higher than CAR1937, Sk4, however the δ15N
enrichment is just +1.7‰. At the start of the main burial activity at the Tagus
valley the intra-individual carbon variability is low, and carbon values range from
–17.4‰ and –16.1‰ (–16.8 ± 0.5‰) at the three sites. Overall, the inter-site
variability is low to moderate and only two individuals present an average
consumption of foods from different trophic levels, as indicated by the nitrogen
enrichment (+2.5‰) between MS19th c., Sk22 (adult female, –16.1‰/+12.2‰)
and Cabeço da Arruda 4 (adult, 17.1‰/+9.7‰).
During the main period of burial activity the isotopic signatures are fairly
consistent and the intra-individual variability is low to moderate. For the benefit of
comparison I divided this main period into two maps (fig. 5.29, maps 3 and 4),
although the division is artificial because most of these individuals are probably
contemporaneous and could be represented on one map. In map 3, the isotopic
values derived from Moita do Sebastião (n=7) range from –17.4‰ and –16.2‰
(–16.7 ± 0.4‰) for carbon, and +9.3‰ to +14.0‰ (+11.7 ± 1.5‰) for nitrogen.
As discussed, the large standard deviation observed for the nitrogen values is
derived from two children possibly benefiting from breastfeeding, and without
their values the range is +9.3‰ to +11.9‰ (+11.0 ± 1.1‰). Nevertheless, there is
a moderate variability between three adult males. One adult indicates the average
consumption of protein from lower trophic levels (MS1952–54, Sk5; δ15N=
+9.3‰). The other two adults (MS19th c., Sk24; 1952–54, Sk34; δ15N= +11.9‰)
show an average intake of protein one trophic level higher (+2.6‰) than the latter
individual. At Cabeço da Arruda (n=7), the isotopic values range from –19.0‰
and –16.6‰ (–17.5 ± 0.8‰) for carbon and +9.5‰ and +12.2‰ for nitrogen.
One adult male (CAR19th c., SkA) presents an odd isotopic profile (δ13C= –19.0‰,
δ15N= +12.2‰) in relation to this series. If his values are excluded, the isotopic
values display a shorter range from –17.9‰ and –16.6‰ (–17.3 ± 0.5‰) for
carbon and +9.5‰ to +11.2‰ (+10.5 ± 0.7‰) for nitrogen.
417
Chapter 5
Figure 5.28. Maps 1 and 2 represent temporal moments of burial activity at the Tagus valley
and display the 14C dated human burials at each site. The burials are identified by skeleton
number, excavation year, 14C date (cal BCE, 95% probability), δ13C, and δ15N.
418
Analysis and Synthesis of the Results
Figure 5.29. Maps 3 and 4 represent temporal moments of burial activity at the Tagus valley
and display the 14C dated human burials at each site. The burials are identified by skeleton
number, excavation year, 14C date (cal BCE, 95% probability), δ13C, and δ15N.
419
Chapter 5
As observed at Moita do Sebastião, at Cabeço da Arruda there is also a moderate
variability between three individuals, one indicating the consumption of protein
from lower trophic levels (CAR1937, Sk2; δ15N= +9.5‰) while two individuals
(CAR19th c., SkIII; 1937, Sk1; δ15N= +11.2‰), present an average intake of
protein of a higher trophic level, although the increase is only +1.7‰ (trophic level
increase, 2 to 4‰). If the values of the two breastfeeding children at Moita do
Sebastião (MS1952–54, Sks 22, 25) and the outlier at Cabeço da Arruda
(CAR19th c., SkA) are excluded, the inter-site variability is low (–17.1 ± 0.5‰ for
carbon and +10.7 ± 0.9‰ for nitrogen). The variability observed for the nitrogen
values remains moderate due to the low values derived from one individual at each
site (MS1952–54, Sk5; δ15N= +9.3‰ and CAR1937, Sk2; δ15N= +9.5‰),
confirming once again the low intra-individual variability between the burial places.
Overall, the individuals buried at Moita do Sebastião and Cabeço da Arruda
present similar isotopic signatures and the deceased could belong either to the
same group or to different groups deriving their isotopes from similar sources.
As discussed, one individual presents an odd isotopic profile (CAR19th c., SkA,
adult male, δ13C= –19.0‰, δ15N= +12.2‰), possibly derived from an important
consumption of freshwater foods, suggesting a different environment where this
individual would have spent most of his time over the last years of his life.
Likewise, the burials represented in map 4 present low δ13C inter-site variation
ranging from –17.9‰ and –16.0‰ (–17.0 ± 0.6‰). The variation of δ15N is higher
(9.5‰ to 13.9‰; +11.0 ± 1.6‰) mainly due to the values derived from a child
buried at Cabeço da Amoreira (burial 2011.1). If these values are excluded the
range is smaller from +9.5‰ to +11.8‰ (+10.4 ± 0.9‰), but still moderate due
to the difference between MS1952–54, Sk17 (δ13C= –16.8‰, δ15N= +9.5‰) and
CAR19th c., Sk 42 (δ13C= –17.2‰, δ15N= +11.8‰) derived from foods of one
trophic level higher (+2.3‰).
The overall homogeneity observed so far changes during the later stages of
burial activity (fig. 5.30, map 5), as indicated by the larger isotopic variability. At
Moita do Sebastião (n=3) the isotopic values range from –18.4‰ and –15.3‰
(–17.3 ± 1.7‰) for carbon and +9.0‰ to +13.4‰ (11.0 ± 2.2‰) for nitrogen.
This heterogeneity derives from two distinct signature groups buried in the site.
One is represented by two individuals with a range from –18.2‰ to –18.4‰
(–18.3 ± 0.1‰) for carbon and +9.0‰ to +10.5‰ (+9.8 ± 1.1‰) for nitrogen.
A second group is represented by one individual with higher values both for carbon (–15.3‰) and for nitrogen (+13.4‰). At Cabeço da Arruda (n=2) the variability is low (δ13C= –18.5 ± 0.6‰, δ15N= +10.4 ± 0.1‰) and similar to one of
the groups buried at Moita do Sebastião. At Cabeço da Amoreira (n=2) the interindividual variability is much higher (–18.0 ± 3.0‰, 10.6 ± 3.3‰). Here, one individual presents a clearly marine-estuarine diet (δ13C= –15.8‰, δ15N= +12.9‰)
while the other individual derives its isotopes from terrestrial sources (δ13C=
–20.1‰, δ15N= +8.2‰). Each individual presents similar isotopic signatures to
the two groups identified at Moita do Sebastião.
420
Analysis and Synthesis of the Results
Figure 5.30. Map 5 represents temporal moments of burial activity at the Tagus valley and
display the 14C dated human burials at each site. The burials are identified by skeleton
number, excavation year, 14C date (cal BCE, 95% probability), δ13C, and δ15N.
During this later period of burial activity there are two distinct isotopic signatures
representing different dietary groups buried in these sites. One group (n=4) was
buried at Moita do Sebastião and Cabeço da Arruda and derived their isotopes
from a mixture of atmospheric and aquatic reservoirs but with a high intake of
terrestrial foods (mixed, high terrestrial). The other main group (n=2), buried at Moita
do Sebastião and Cabeço da Amoreira also depended on mixed reservoirs but the
intake of marine foods was higher (mixed, high aquatic), and in contrast with the
other group. A possible third group is represented by one individual buried at
Cabeço da Amoreira, which could be considered an outlier to the series due to its
low carbon values clearly indicative of a full terrestrial diet. Despite their
differences, the two main groups were coherent with the ecology of the sites. The
latter individual was clearly an outlier, possibly from inland origin. The two dietary
groups (mixed, high terrestrial and mixed, high aquatic) were found at Moita do
Sebastião in distinct areas within the burial ground. The mixed, high terrestrial group
(MS1952–54, Sks 10, 33) was buried in the area excavated in the 1950s while the
mixed, high aquatic individual (MS19th c., SkCT) was buried in the nineteenth century
area. Likewise, at Cabeço da Amoreira the individuals with distinct isotopic
signatures were excavated from two different areas. The outlier CAM–01–01 (139)
was excavated at the J. Rolão area (c 20–60 cm below present day surface, Roksandic 2006), while the burial 2011.2 with the mixed, high aquatic diet was found in
421
Chapter 5
the current excavation area on the top of the midden. During this later phase at
Cabeço da Arruda, the dominant signature corresponds to the mixed, high terrestrial
diet also found at the 1950s area at Moita do Sebastião.
The very last burials known in the valley present coherent isotopic signatures of
a mixed diet with high intake of aquatic foods (n=2, δ13C= –16.1 ± 1.1‰,
δ15N= +12.4 ± 0.1‰) (fig. 5.31, map 6). Interestingly, the individual with the
highest intake of aquatic foods was buried at Cabeço da Arruda, and presents a
signature which is in contrast with the mixed, high terrestrial profile of previous burial
episodes (fig. 5.30, map 5).
Figure 5.31. Map 6 represents temporal moments of burial activity at the Tagus valley and
display the 14C dated human burials at each site. The burials are identified by skeleton
number, excavation year, 14C date (cal BCE, 95% probability), δ13C, and δ15N
Overall, the isotopic signatures derived from Moita do Sebastião, Cabeço da
Arruda, and Cabeço da Amoreira are indicative of the consumption of a mixture of
marine-estuarine and terrestrial resources. These results are consistent with the
ecology of the area where the sites are located, and suggest that a significant
portion of the protein consumed could be obtained locally. As discussed, two
burials are clear outliers to this pattern and may suggest that these individuals were
originally from elsewhere. The overall homogeneity of the stable isotope profiles
indicates that the people buried in these places derived their protein from similar
sources, but it is difficult to establish if these individuals belong to one or various
groups, based on the current isotopic data. The people buried in these places were
422
Analysis and Synthesis of the Results
very likely consuming resources collected nearby the sites, as indicated by the intake of marine-estuarine sources. However, in most cases, more than half of the
protein consumed is derived from terrestrial food sources which could have been
obtained from elsewhere, nearby or at some distance.
There is a slight shift during the latest period of burial activity regarding the
isotopic signatures. Instead of one homogeneous dietary group, the people buried
in these sites during the later phases, present two distinct isotopic profiles,
representing two dietary groups. One group is possibly more local and is buried at
Moita do Sebastião and Cabeço da Amoreira (mixed, high marine-estuarine, c 50% of
protein from nearby the shell middens). The other group presents a predominantly
inland pattern of food consumption (mixed, high terrestrial, deriving much less from
the immediate area of the shell middens), and is represented by individuals buried
at Cabeço da Arruda and also at Moita do Sebastião.
Only two burials are clear outliers. One was at Cabeço da Arruda during the
main period of burial activity and the other at Cabeço da Amoreira during the later
phases. These are individuals whose isotopic signatures do not match the main
patterns. This could be explained by personal food preferences or social
impediments. Another possible explanation is that these individuals derived their
protein from other environments and were possibly “external” to the main
group(s) identified, but were nevertheless, provided with a permanent place in
these burial grounds (see discussion in chapter 6).
Arapouco, Sado valley
At Arapouco, the human collagen samples (n=12/no. individuals=12) display stable isotope values ranging between –17.9‰ to –16.1‰ (–17.0 ± 0.5‰) for carbon,
and +11.0‰ to +13.0‰ (+12.2 ± 0.4‰) for nitrogen, which are indicative of the
consumption of a mixture of marine-estuarine and terrestrial food resources. These
results are in agreement with previous trace element analysis (n=8). Trace elements
indicated high intake of marine protein (rich in zinc), as indicated by the high
Zn/Ca ratio, balanced by significant consumption of vegetables, as indicated by
the Mg/Ca index and Ba ratio (ORBa) (Umbelino 2006; Umbelino and Cunha
2012).
The lack of reliable 14C data available for Arapouco does not allow comparisons
over time. However the intra-individual variation is small indicating that the people
buried at Arapouco shared a similar pattern of protein consumption.
As discussed, one skeleton wrongly labelled as skeleton 5 from Arapouco was
sampled for isotopic analysis by M. Fontanalls-Coll and colleagues (2014). The
provenience of this individual is unknown, but I suggest, based on the visual
analysis of the sediment, that this skeleton could be from Cabeço do Pez. The
values for the mislabelled individual (δ13C= –20.0‰; δ15N= +9.0‰, FontanallsColl et al. 2014) indicate the consumption of a full terrestrial diet which is in
contrast with the mixed diet observed for the burial population of Arapouco.
423
Chapter 5
Cabeço das Amoreiras, Sado valley
Stable isotope values derived from most individuals (n=9/no. individuals=4)
buried at Cabeço das Amoreiras indicate consumption of foods primarily from
terrestrial sources. This is supported by the analysis of trace elements (n=2) which
suggested high consumption of vegetables complemented with a medium intake of
terrestrial animals as indicated by the ratio of Ba (ORBa) validated by the Mg/Ca
and V indexes (Umbelino 2006; Umbelino and Cunha 2012). One mature adult
male (CAMS1958, Sk6) is exceptional to this pattern and presents the highest set
of isotopic values (δ13C= –16.0‰, δ15N= +13.4‰), indicating an important intake
of marine-estuarine protein in contrast with the remaining individuals buried at
Cabeço das Amoreiras. When CAMS1958, Sk6 is excluded from the series
(n=9/no. individuals=4), the intra-individual variation is smaller and the mean
values are –19.1 ± 0.6‰, instead of –18.8 ± 1.1‰ for carbon and +9.2 ± 0.8‰,
instead of +9.6 ± 1.5‰ for nitrogen.
Some individuals were sampled more than once and present slightly different
stable isotope values (table 5.10), but are statistically coherent. For consistency, the
Ua–number values were preferred in the analysis over time. Over time (n=3), the
δ13C variation is generally small (–19.7 ± 0.6‰) but moderate when considering
the δ15N values (+8.4 ± 1.0‰) where there is a small decrease over time (figs. 5.32,
5.33).
The individuals buried at Cabeço das Amoreiras present a consistent isotopic
signature, particularly when considering the long time span of burial activity.
As discussed, one individual (CAMS1958, Sk6) presents the exception to this
pattern and indicates a very different pattern of protein consumption. Unfortunately it remains undated.
Table 5.10. Cabeço das Amoreiras, Sado valley: chronology and stable isotope values (δ13C and δ15N) of
human bone collagen. Samples are sorted so that the most recent dates are at the top of the column.
Identification of the
human remains
Ind.
Sex
Lab no.
14
4, 1958 Adult
♂?
Ua–47973
7, 1958 Adult
♂
5, 1958 Adult
♂
424
Age
Measurements (IRMS)
C Age BP
δ13C (‰)
δ15N (‰)
6484±39
Post. density
estimate
(95% prob.)
5483–5329
–20.0,
this study
7.6,
this study
Ua–47974
6645±42
5620–5481
–20.0,
this study
–18.9,
Fontanals-Coll et al.
2014
–19.2,
Guiry et al. 2015
8.2,
this study
9.7,
Fontanals-Coll et al.
2014
9.4
Guiry et al. 2015
Beta–125110
7230±40
6156–5914
–19.0,
Fontanals-Coll et al.
2014
–18.5,
Guiry et al. 2015
9.5,
Fontanals-Coll et al.
2014
10.0,
Guiry et al. 2015
Analysis and Synthesis of the Results
-12
-13
Bone collagen δ13C (‰)
-14
-15
-16
-17
-18
-19
-20
-21
6200 6100 6000 5900 5800 5700 5600 5500 5400 5300 5200
cal BCE 95%
Figure 5.32. Cabeço das Amoreiras, Sado valley: bone collagen δ13C over time (cal BCE,
95% probability). Individuals 5 (open square), and 7 (grey square) have multiple δ13C values.
16
15
Bone collagen δ15N (‰)
14
13
12
11
10
9
8
7
6
6200 6100 6000 5900 5800 5700 5600 5500 5400 5300 5200
cal BCE 95%
Figure 5.33. Cabeço das Amoreiras, Sado valley: bone collagen δ15N over time (cal BCE,
95% probability). Individuals 5 (open square), and 7 (grey square) have multiple δ15N values.
425
Chapter 5
Vale de Romeiras, Sado valley
The isotopic signature for Vale de Romeiras (n=10/no. individuals=7) is clearly
terrestrial. These results are supported by previous trace element analysis (n=10)
suggesting the high consumption of vegetables supplemented by a relatively low
intake of terrestrial animals as indicated by the ratio of Ba (ORBa) validated by the
Mg/Ca and V indexes (Umbelino 2006; Umbelino and Cunha 2012). The mean
value for all isotopic measurements is –19.2 ± 0.7‰ (n=10) for carbon and
+9.5 ± 0.9‰ (n=9) for nitrogen.
Some individuals were sampled more than once and present slightly different
stable isotope values (table 5.11), but are generally statistically consistent. In one
case (VR1959, Sk9) the values are less consistent (δ13C= –20.5‰, δ15N= +7.7‰
vs δ13C= –18.7‰, δ15N= +9.6‰), but remain within the range of a terrestrial
isotopic signature. For consistency, the Ua–number values were preferred in the
analysis over time. Unfortunately, the sample size for the analysis over time is
limited (n=2). The δ13C variation is insignificant (–20.4 ± 0.2‰) and slightly higher
when considering the δ15N values (+8.2 ± 0.7‰) where there is a small decrease
over time (figs. 5.34, 5.35). Nevertheless, the individuals buried at Vale de
Romeiras present a consistent isotopic signature, particularly when considering the
long time span of the burial activity. However, this pattern should be tested with a
larger sample.
Table 5.11. Vale de Romeiras, Sado valley: chronology and stable isotope values (δ13C and δ15N) of
human bone collagen. Samples are sorted so that the most recent dates are at the top of the column.
Identification of the human remains
Measurements (IRMS)
Ind.
Age
Sex
Lab no.
14
9, 1959
Adult
♂
Ua–47983
19, 1959
♀
Adult,
20–35 yrs
Ua–46972
426
C Age BP
δ13C (‰)
δ15N (‰)
6625±51
Post. density
estimate
(95% prob.)
5625–5475
–20.5,
this study
–18.7,
Guiry et al. 2015
7.7,
this study
9.6,
Guiry et al. 2015
7640±55
6593–6370
–20.2,
this study
–19.8,
Guiry et al. 2015
8.7,
this study
9.1,
Guiry et al. 2015
Analysis and Synthesis of the Results
-12
-13
Bone collagen δ13C (‰)
-14
-15
-16
-17
-18
-19
-20
-21
6600
6400
6200
6000
cal BCE 95%
5800
5600
5400
Figure 5.34. Vale de Romeiras, Sado valley: bone collagen δ13C over time (cal BCE, 95%
probability). Individuals 9 (grey square), and 19 (black square) have two sets of δ13C values.
16
15
14
Bone collagen δ15N (‰)
13
12
11
10
9
8
7
6
6600
6400
6200
6000
cal BCE 95%
5800
5600
5400
Figure 5.35. Vale de Romeiras, Sado valley: bone collagen δ15N over time (cal BCE, 95%
probability). Individuals 9 (grey square), and 19 (black square) have two sets of δ15N values.
427
Chapter 5
Cabeço do Pez, Sado valley
At Cabeço do Pez, the human collagen samples (n=13/no. individuals=9) display
stable isotope values ranging between –20.7‰ and –17.2‰ (–19.3 ± 0.7‰) for
carbon, and +6.7‰ to +13.5‰ (+9.8 ± 1.5‰) for nitrogen (tables 5.12, 5.13).
When considering the 14C dated individuals (n=5) and the Ua–number values, in
the case of individuals with more than one set of measurements, the range of
values is maintained, but the standard deviation for the nitrogen values increases
(+11.0 ± 2.7‰).
Overall, the δ13C variation is small to moderate, while the intra-individual
variation of δ15N values is high (figs. 5.36, 5.37). The highest δ15N values derive
from three individuals, two adults (CP1956, Sks 5, 27), and one isolated child
(CP1956, Sk2). Individuals 2 and 5 have two sets of measurements (this study;
Fontanals-Coll et al. 2014; Guiry et al. 2015), and individual 27 was sampled four
times (this study; Fontanals-Coll et al. 2014; Guiry et al. 2015; Umbelino 2006).
Table 5.12. Cabeço do Pez, Sado valley: stable isotopes values (δ13C and δ15N) without associated 14C
date. Samples are sorted so that the lowest δ13C values are at the top of the column.
Identification of the human remains
Measurements (IRMS)
Reference
Individual
Age
Sex
Lab no.
δ C (‰)
δ N (‰)
A, 1959
Adult
n/d
n/a
–20.0
9.3
13
15
Fontanals-Coll et al. 2014
B, 1959
Adult
n/d
n/a
–19.8
8.2
Fontanals-Coll et al. 2014
13, 1956
15–23 yrs
♂
7768.3
–19.7
8.9
Guiry et al. 2015
14, 1956
Adult
♀?
7762.3
–19.7
8.5
Guiry et al. 2015
6, 1956
Adult
n/d
6894.3
–19.6
9.2
Guiry et al. 2015
8, 1956
Adult, young
♀
7761.3
–19.5
8.8
Guiry et al. 2015
15 (9)?, 1956
Adult
♀?
6891.3
–19.5
8.7
Guiry et al. 2015
26, 1956
Adult
n/d
7760.3
–19.5
8.7
Guiry et al. 2015
17, 1956
Non–adult, 5–7 yrs
n/d
n/a
–19.4
9.3
Fontanals-Coll et al. 2014
24, 1956
Adult
n/d
7759.2
–19.4
8.9
Guiry et al. 2015
21, 1956
Adult
♀
n/a
–19.3
9.2
Fontanals-Coll et al. 2014
25, 1956
Adult, mature
♂
7764.3
–19.3
8.6
Guiry et al. 2015
20, 1956
Adult
n/d
7758.2
–19.2
9.7
Guiry et al. 2015
19, 1956
Adult
n/d
7763.3
–19.2
8.8
Guiry et al. 2015
16, 1956
Adult
♂
7767.3
–18.9
9.6
Guiry et al. 2015
1, 1956
Adult
n/d
6893.3
–18.5
10.2
Guiry et al. 2015
428
Adult
Adult
Adult
27, 1956
9, 1956
11, 1956
♂?
n/d
♂
Ua–46933 6788±46
Ua–46932 6780±48
Ua–46934 6734±51
Ua–46931 6791±43
♀
Adult, 35–50 yrs
5, 1956
Age BP
Ua–46930 5579±41
14C
Non-adult, 1.5–2 yrs n/d
Lab no.
2, 1956
Sex
Age
Measurements (IRMS)
Individual
Identification of the human remains
5733–5566
5726–5552
5705–5510
5711–5536
Post. density estimate
(95% probability)
4461–4266
–20.7,
this study
–19.8,
this study
–20.0,
Fontanals-Coll et al. 2014
–19.7,
this study
–19.5,
Fontanals-Coll et al. 2014
–18.7,
Umbelino 2006
–18.6,
Guiry et al. 2015
–18.3,
this study
–17.2,
Guiry et al. 2015
–19.1,
this study
–18.4,
Fontanals-Coll et al. 2014
δ13C (‰)
6.7,
this study
10.4,
this study
9.0,
Fontanals-Coll et al. 2014
13.5,
this study
10.0,
Fontanals-Coll et al. 2014
9.8,
Umbelino 2006
10.6,
Guiry et al. 2015
13.0,
this study
12.4,
Guiry et al. 2015
11.5,
this study
11.9,
Fontanals-Coll et al. 2014
δ15N (‰)
Main
Main
Main
Main
Late
Burial activity
Table 5.13. Cabeço do Pez, Sado valley: chronology and stable isotope values (δ13C and δ15N) of human bone collagen. Samples are sorted so that the most recent dates are
at the top of the column.
Chapter 5
-12
-13
Bone collagen δ13C (‰)
-14
-15
-16
-17
-18
-19
-20
-21
5800
5600
5400
5200 5000 4800
cal BCE 95%
4600
4400
4200
Figure 5.36. Cabeço do Pez, Sado valley: variation of δ13C (‰) over time (cal BCE, 95%
probability) of 11 samples of human bone collagen from 5 individuals. Individuals 2 (grey
square), 5 (open circle), 9 (open square), and 27 (grey square with outline) have multiple δ13C values.
16
15
Bone collagen δ15N (‰)
14
13
12
11
10
9
8
7
6
5800
5600
5400
5200 5000 4800
cal BCE 95%
4600
4400
4200
Figure 5.37. Cabeço do Pez, Sado valley: variation of δ15N (‰) over time (cal BCE, 95%
probability) of 10 samples of human bone collagen from 5 individuals. Individuals 2 (grey
square), 5 (open circle), 9 (open square), and 27 (grey square with outline) have multiple δ15N values.
The possible anomalous δ15N value of +13.5‰ obtained for individual 27 (this study) is
not displayed.
430
Analysis and Synthesis of the Results
In the case of CP1956, Sk5 (5711–5536 cal BCE, 95% probability), the two different
sets present moderate deviations but both indicate a relatively significant
consumption of marine protein, c 30 to 40 ± 10%, placing this individual in an
outlier position. On the other hand, the four sets available for CP1956, Sk27
(5705–5510 cal BCE, 95% probability) are consistent in terms of δ13C values but the
high δ15N value obtained in this study (+13.5‰) seems to be anomalous when
considering the consistent values from the other three measurements with a mean
value of +10.1 ± 0.4‰. Thus, when this possibly anomalous value is excluded, the
δ15N values obtained for CP1956, Sk27 are no longer eccentric to the series. The
two sets of values for CP1956, Sk2 are more consistent and its slightly deviant
position in the series can be explained by the recent date obtained for this
1.5–2 year old child (4461–4266 cal BCE, 95% probability) coeval with Middle
Neolithic groups in Portugal. Also, its relatively high nitrogen value could suggest a
breastfeeding effect. If this was the case, the δ15N value for the mother of this
child could be estimated to be between +8.5‰ and +10.5‰ by lowering 1 to 3‰
(Katzenberg et al. 1996) from the bioenriched δ15N value for the child. In this case,
the δ13C values derived from this child reflect the diet of his/her mother and indicate the consumption of terrestrial foods, which is in clear continuity with the
isotopic signatures observed during the Late Mesolithic at Cabeço do Pez.
One adult (CP1956, Sk11) presents the lowest set of values in the series
(δ13C= –20.7‰; δ15N= +6.7‰) indicative of a terrestrial diet based primarily on
low trophic level food sources, such as vegetables. Although its 14C date (5733–
5566 cal BCE, 95% probability) is consistent with the main burial activity in the site,
its isotopic signature places this individual in an outlier position.
As discussed, three individuals buried at Cabeço do Pez present different
patterns of protein consumption (CP1956, Sks 2, 5, 11). Two are adults buried
during the main activity of the site. One of these adults presents a mixed diet of
terrestrial resources complemented by protein of marine-estuarine origin (CP1956,
Sk5). The other adult (CP1956, Sk11), indicates a full terrestrial diet, but based
exclusively on plants. The third individual is a child (CP1956, Sk2) buried at
Cabeço do Pez several centuries after the main burial activity at the site, and its
outlier position derives possibly from a breastfeeding effect. When these three
outliers (CP1956, Sks 2, 5, 11) are excluded, as well as the possibly anomalously
high δ15N value obtained for CP1956, Sk27 (+13.5‰), the intra-individual
variation is smaller and the isotope values range between –20.0‰ to –18.5‰
(–19.4 ± 0.4‰) for carbon, and +8.2‰ to 10.6‰ (+9.3 ± 0.7‰) for nitrogen,
which are indicative of a typical terrestrial food diet.
The strong terrestrial component observed in the isotopic signatures derived
from most individuals buried at Cabeço do Pez is supported by previous trace
element studies (n=11). Trace element analysis indicates high consumption of
vegetables, such as tubers, legumes and fruits, supplemented with medium
consumption of animal food sources (Umbelino 2006; Umbelino and Cunha
2012). As discussed, two individuals, buried during the main period of burial
activity, are clearly eccentric to this pattern. While one individual follows the terres431
Chapter 5
trial diet but largely from vegetarian sources, the other individual presents a mixed
diet of terrestrial sources supplemented with a medium consumption of marineestuarine sources of protein. Both signatures contrast with the whole series of
Cabeço do Pez.
Two 14C measurements on individual 4 (Beta–125109 and Sac–1558, Cunha and
Umbelino 2001) were excluded from this study because the two δ13C values
available were obtained on an AMS instrument during the radiocarbon dating
process making the calibration of this 14C date highly uncertain. Nevertheless, one
of these original AMS measurements is –19.3‰ for carbon (Sac–1558) and is in
line with the current isotopic series for Cabeço do Pez, suggesting a low offset
from an isotope-ratio mass spectrometry (IRMS) measurement, which must be
confirmed with future measurements. The second measurement (–22.6‰,
Beta–125109) presents a greater offset and remains unreliable for calibration and
dietary analysis.
Várzea da Mó, Sado valley
The only human burial known at Várzea da Mó presents an isotopic signature
derived from terrestrial sources as indicated by the values –20.5‰ for carbon and
+8.1‰ for nitrogen.
Poças de S. Bento, Sado valley
Poças de S. Bento was not sampled or analysed in the scope of this study.
However, the published data are important resources for comparison with the
studied material from the Sado valley, allowing a more robust analysis of the local
setting.
At Poças de S. Bento, the human collagen samples (n=9/no. individuals=8)
display stable isotope values ranging between –17.9‰ and –16.9‰ (–17.4 ± 0.3‰)
for carbon, and +11.7 to +12.8‰ (+12.3 ± 0.4‰) for nitrogen (table 5.14), which
are indicative of the consumption of a mixture of marine-estuarine and terrestrial
food resources. These results are in agreement with previous trace element analysis
(n=6) showing the highest Zn/Ca ratio in the Sado series explained by significant
intake of marine protein, such as fish and crustaceans (Umbelino 2006; Umbelino
and Cunha 2012).
The lack of reliable 14C data available for Poças de S. Bento does not allow
comparisons over time. However, the intra-individual variation is small indicating
that the people buried at Poças de S. Bento shared a similar pattern of protein
consumption (fig. 5.38).
432
♂
n/d
n/d
n/d
n/d
n/d
n/d
Adult
Adult
Adult
Adult
Adult
Non-adult, 12–15 yrs
Adult
Adult
Adult
5, 1960
3, 1960
7, 1960
7, 1960
11, 1960
1, 1960
13, 1960
12, 1960
8, 1960
n/d
n/d
Sex
Age
Individual
Identification of the human remains
Long bone
999.37.30
Long bone
999.41.1
Long bone
999.42.5
Long bone
999.30.13
Long bone
999.40.5
Long bone
999.35.3
n/a
Long bone
999.31.20
Long bone
999.33.3
Bone
7754.3
6898.3
6895.3
6899.3
7752.2
7755.2
n/a
6896.6
6897.3
Lab no.
–16.9
–17.0
–17.0
–17.3
–17.4
–17.5
–17.6
–17.7
–17.9
δ13C (‰)
Measurements (IRMS)
12.5
12.6
12.8
12.2
12.7
12.1
11.7
11.8
12.1
δ15N (‰)
3.3
3.3
3.3
3.4
3.5
3.5
n/a
3.4
3.3
C:N
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Guiry et al. 2015
Umbelino 2006
Guiry et al. 2015
Guiry et al. 2015
Reference
Table 5.14. Poças de S. Bento, Sado valley: stable isotopes values (δ13C and δ15N) without associated 14C date. Samples are sorted so that the lowest δ13C values are at the
top of the column.
Chapter 5
16
15
14
13
δ15N (‰)
12
11
10
9
8
7
6
-22
-21
-20
-19
-18
-17
δ13C
-16
-15
-14
-13
-12
(‰)
Figure 5.38. Poças de S. Bento, Sado valley: correlation between the δ13C (‰) and δ15N (‰)
values of 9 samples of human bone collagen from 8 individuals. Individual 7 (circle) has two
sets of values.
Isotopic signatures and implications for the local setting, Sado valley
The isotopic signatures of carbon and nitrogen derived from the sites of the Sado
valley present two dietary groups. One group derived its stable isotopes from a
mixture of terrestrial foods complemented by c 40 ± 10% of protein from marineestuarine sources, and was buried in the western sites of the Sado valley, at
Arapouco and Poças de S. Bento. The other main dietary group derived its
isotopes from a full terrestrial diet, and was buried in the eastern side of the valley
in the sites of Cabeço das Amoreiras, Vale de Romeiras, Cabeço do Pez, and
Várzea da Mó (fig. 5.39).
The intra-individual variation within each site is generally small suggesting that
the people interred in each burial ground shared a fairly homogeneous pattern of
protein consumption. Few individuals are noted as real outliers, and are found only
at the eastern sites:
• one undated adult male at Cabeço das Amoreiras (CAMS1958, Sk6)
presents a mixed diet with significant intake of protein from marine-
434
Analysis and Synthesis of the Results
•
•
estuarine sources (δ13C= –16.0‰, δ15N= +13.4‰), which is in contrast
with the remaining individuals in this burial ground;
one adult female at Cabeço do Pez (CP1956, Sk5, 5711–5536 cal BCE)
presents a relatively significant intake of marine-estuarine resources
supplemented by the consumption of terrestrial foods (δ13C= –18.3‰,
δ15N= +13.0‰, this study; δ13C= –17.2‰; δ15N= +12.4‰, Guiry et al.
2015);
also at Cabeço do Pez, another adult deviates from the general pattern
(CP1956, Sk11, 5733–5566 cal BCE) due to its possibly exclusive
vegetarian diet, as suggested by the very low isotopic values
(δ13C= –20.7‰; δ15N= +6.7‰).
16
Arapouco
Poças de S. Bento
Cabeço das Amoreiras
Vale de Romeiras
Cabeço do Pez
Várzea da Mó
15
14
13
δ15N (‰)
12
11
10
9
8
7
6
-22
-21
-20
-19
-18
-17
-16
δ13C (‰)
-15
-14
-13
-12
Figure 5.39. Sado valley: correlation between the δ13C (‰) and δ15N (‰) values of 66 samples of human bone collagen from 53 individuals. The inter-site variation suggests that the
burials grounds were used distinctively by at least two populations. One main group was
defined in the west, characterized by an isotopic profile indicative of a mixed diet
(Arapouco and Poças de S. Bento). The second main group was defined in the eastern sites
where individuals with typical terrestrial diet buried their dead (Cabeço das Amoreiras, Vale
de Romeiras, Cabeço do Pez, and Várzea da Mó). The outliers of the series are indicated
with a modified notation: CAMS1959, Sk6 (open diamond), CP1956, Sks 5 (open triangle), and
11 (grey triangle). CP1956, Sk2 (dashed outline) is a child, possibly breastfeeding, and coeval
with the early stages of the Middle Neolithic in Portugal.
435
Chapter 5
At the Sado valley, while the intra-site variation is small, suggesting a homogeneous
dietary population buried in each site, the inter-site variation indicates that the
burial grounds were used distinctively by at least two populations. One main group
was defined in the west, characterized by an isotopic profile indicative of a mixed
diet, and used Arapouco and Poças de S. Bento as their burial grounds. The second
main group was defined in the eastern sites, where individuals with a typically
terrestrial diet buried their dead, and used Cabeço das Amoreiras, Vale de
Romeiras, Cabeço do Pez, and Várzea da Mó as their burial grounds. The individual isotopic profiles overlap within the two main dietary groups and it is
speculative to distinguish eventual subgroups, which could represent different
communities using specifically determined burial ground(s) in particular, as was
briefly suggested before (Fontanals-Coll et al. 2014, 544).
Unfortunately, the 14C data available is limited for the western sites (figs. 5.40,
5.41). Nevertheless, it seems that the groups in the west and those in the east used
the sites simultaneously, as demonstrated by the earliest burials known in the
valley, from both the western (Arapouco) as well as the eastern sites (Cabeço das
Amoreiras) (fig. 5.18).
I suggested that the skeleton wrongly labelled as skeleton 5 from Arapouco
could be a burial from Cabeço do Pez. The assignment of this burial to Cabeço do
Pez is based on the visual analysis of the sediment within the individuals preserved
in paraffin blocks at MNA (Arapouco, Cabeço das Amoreiras, and Cabeço do
Pez). Thus, it is unknown if the sediment could also match that of the individuals
buried at the neighbouring site Vale de Romeiras. The isotopic values of the mislabelled individual (δ13C= –20.0‰, δ15N= +9.0‰; Fontanalls-Coll et al. 2014) indicate the consumption of a full terrestrial diet which is in contrast to the mixed diet
observed for the burial population of Arapouco, but matches the signature profiles
of both Cabeço do Pez and Vale de Romeiras. I sampled this skeleton (adult
female, Ua–46939, table A2) for 14C dating, and the calibrated date range for the
time of her burial is 5787–5634 cal BCE (95% confidence) which is consistent with
the main burial activity at Cabeço do Pez.
The distinct isotopic profiles noted between the western and eastern groups are
coherent with the zooarchaeological evidence. While fish remains of marineestuarine species are relatively abundant at Arapouco and Poças de S. Bento in the
west (Arnaud 1989; Marques-Gabriel 2015), mammal bones are more frequent at
Cabeço das Amoreiras, Vale de Romeiras, and Cabeço do Pez (Arnaud 1989;
Rowley-Conwy 2015). Fish remains are also found at Cabeço das Amoreiras, but in
much lesser quantity than in the western sites, and without species predominantly
marine, such as those of shark families such as the Triakidae and the Lamnidae
found at Arapouco and Poças de S. Bento (Marques-Gabriel 2015). Unlike
Arapouco, mammal remains are relatively abundant at Poças de S. Bento, in
particular rabbits (Rowley-Conwy 2015).
436
Analysis and Synthesis of the Results
-12
CAMS
VR
CP
VM
-13
Bone collagen δ13C (‰)
-14
-15
-16
-17
-18
-19
-20
-21
6600 6400 6200 6000 5800 5600 5400 5200 5000 4800 4600 4400 4200
cal BCE 95%
Figure 5.40. Sado valley: bone collagen δ13C over time (cal BCE, 95% probability).
CAMS1958 Sks 5 (open diamond), 7 (grey diamond), VR1959 Sks 9 (grey square), 19 (black square),
CP1956 Sks 2 (dashed outline), 5 (open triangle), 9 (grey triangle), 27 (grey triangle with outline) have
multiple δ13C values.
16
CAMS
VR
CP
VM
15
Bone collagen δ15N (‰)
14
13
12
11
10
9
8
7
6
6600 6400 6200 6000 5800 5600 5400 5200 5000 4800 4600 4400 4200
cal BCE 95%
δ15N
Figure 5.41. Sado valley: bone collagen
over time (cal BCE, 95% probability).
CAMS1958 Sks 5 (open diamond), 7 (grey diamond), VR1959 Sks 9 (grey square), 19 (black square),
CP1956 Sks 2 (dashed outline), 5 (open triangle), 9 (grey triangle), 27 (grey triangle with outline) have
multiple δ15N values.
437
Chapter 5
The distinct food choices noted between the western and eastern groups seem to
follow a geoecological pattern (figs. 5.42, 5.43). People buried in the western sites
would have easier access to marine resources, not only because the estuary was
probably richer in marine foods where Arapouco is located, but also because of the
accessibility to the coast by boat. The location of the other western site, Poças de
S. Bento, in a relatively inland area c 3 kilometres from the river, remains an
apparently odd choice. In this case, the populations buried here would get an important part of their food outside the immediate area of this site. In the eastern
side of the valley, the terrestrial signatures observed could be explained if the local
environment was less rich in marine foods in this portion of the valley. Also, it can
be suggested that the individuals buried at the eastern sites derived most of their
carbon and nitrogen from other ecosystems, either in the valley, which was
probably rich in terrestrial resources and/or elsewhere. Interestingly, despite the
important component of shells of aquatic molluscs, in these eastern sites as well, its
consumption by the burial population seems to be low and occasional.
Figure 5.42. Sites with human remains at the Sado valley: minimum number of individuals
(MNI), mean δ13C (‰) and δ15N (‰) values, and respective standard deviations derived
from human bone collagen. The burial populations of the western-most sites display isotopic signatures indicative of the consumption of a mixture of terrestrial and marineestuarine resources. In contrast, the people buried at the eastern sites generally present
isotopic values consistent with diets derived entirely from terrestrial sources.
Two individuals with a mixed diet of marine-estuarine and terrestrial foods were
buried at the eastern sites where the burial populations present a full terrestrial diet.
One of these individuals was buried at Cabeço das Amoreiras (CAMS1958, Sk6)
438
Analysis and Synthesis of the Results
but unfortunately remains to be dated. This mature adult male was buried in the
centre of the burial area and could be one of the earliest burials in the site. If this
were the case, this individual would be contemporaneous with the population
buried in the western site of Arapouco, presumably from the same dietary group.
The other dietary outlier is found at Cabeço do Pez (CP1956, Sk5). This mature
adult female was buried during the period of main activity at the site and despite
being contemporaneous with most burials she presents a dietary pattern of
terrestrial sources of protein, complemented by the intake of marine-estuarine
foods, which is in contrast with the full terrestrial profile of the site. This burial
(5711–5536 cal BCE, 95% probability) is more recent than the only burial known at
Arapouco, but it is contemporaneous with the use of the Poças de S. Bento, in the
west (table A14), with an overall similar dietary profile (δ13C = –17.4 ± 0.3; δ15N =
+12.3 ± 0.4). As a working hypothesis it could be suggested that CP1956, Sk5 was
originally from the western dietary group, possibly from the site of Poças de S.
Bento.
Tagus vs Sado
Current isotopic data suggests that the protein sources of choice vary regionally
and possibly culturally. The isotopic signatures are consistent with the
zooarchaeological evidence, and suggest that the dietary patterns were influenced
and reflected the geography and ecology of the sites where the individuals were
buried. Overall, while the isotopic signatures from the Tagus sites present a
balanced intake of marine-estuarine and terrestrial foods, the individuals buried in
the Sado sites can be grouped into two isotopic populations (figs. 5.43, tables 5.15,
5.16). The isotopic population characterized by a mixed diet of terrestrial and
marine-estuarine resources was typically buried at the western sites. The population
deriving their isotopes mainly from terrestrial sources was buried at the eastern
sites of the valley.
This isotopic contrast between the individuals buried in the Tagus valley and
the western sites of the Sado valley, with those buried in the eastern portion of the
Sado valley (fig. 5.44), suggests that these burial grounds were used by at least two
major regional and/or cultural groups of hunter-gatherers. One is defined by the
consumption of a mixture of marine-estuarine and terrestrial foods in the west, and
the other in the east deriving their protein from terrestrial sources. However, while
the dietary profiles in each site of the Sado valley remained highly homogeneous
over time, the sites of the Tagus valley became more heterogeneous during the
later episodes of the burial activity. At the later stages of burial activity at the Tagus
valley, two dietary groups were buried in each site, both with a mixed diet, but
while one group had a high marine intake the other group relied on a greater
quantity of terrestrial sources.
439
Chapter 5
Figure 5.43. Individuals buried in the Tagus and Western Sado valleys derived their isotopes
from a mixed diet of terrestrial and marine-estuarine resources (M) while the people buried
in the East Sado obtained their protein from terrestrial food sources (T).
440
Analysis and Synthesis of the Results
Table 5.15. Average δ13C values for the human bone collagen samples from individuals buried in the
Tagus and Sado valleys.
Valley
Site
No. samples
Average δ13C
Standard deviation
Tagus
Moita do Sebastião
21
–16.9
0.7
Tagus
Cabeço da Amoreira
9
–16.4
1.5
Tagus
Cabeço da Arruda
16
–17.3
1.0
Sado
Arapouco
12
–17.0
0.5
Sado
Poças de S. Bento
9
–17.4
0.3
Sado
Cabeço das Amoreiras
10
–18.8
1.1
Sado
Vale de Romeiras
10
–19.2
0.7
Sado
Cabeço do Pez
27
–19.3
0.7
Sado
Várzea da Mó
1
–20.5
–
Table 5.16. Average δ15N values for the human bone collagen samples from individuals buried in the
Tagus and Sado valleys.
Valley
Site
No. samples
Average δ15N
Standard deviation
Tagus
Moita do Sebastião
20
11.1
1.3
Tagus
Cabeço da Amoreira
8
12.0
1.7
Tagus
Cabeço da Arruda
16
11.0
0.9
Sado
Arapouco
12
12.2
0.4
Sado
Poças de S. Bento
9
12.3
0.4
Sado
Cabeço das Amoreiras
10
9.6
1.5
Sado
Vale de Romeiras
9
9.5
0.9
Sado
Cabeço do Pez
27
9.8
1.5
Sado
Várzea da Mó
1
8.1
–
In both valleys just few individuals stand out as outliers. In two cases, the
individuals present isotopic signatures not found in any other of the sites analysed.
This is the case for the first clear isotopic outlier found at Cabeço da Arruda in the
Tagus valley (CAR19th c., SkA, 5902–5617 cal BCE) which presents a freshwater
signature not comparable with any other individual. Likewise, one individual
buried at Cabeço do Pez (CP1956, Sk11, 5733–5566 cal BCE) has no isotopic
parallel, and although it presents a clear terrestrial diet in agreement with the site’s
profile, this is the only signature that could be derived from an entirely vegetarian
diet.
Most commonly, the outliers present isotopic signatures that are coherent with
the dominant profile observed in other sites. This is the case for the second outlier
found at the Tagus valley at Cabeço das Amoreiras (CAM–01–01, bone 139, 5613–
5296 cal BCE), which presents a clear terrestrial signature, comparable with the
individuals buried at both Vale de Romeiras (9, 5625–5475 cal BCE) and Cabeço
441
Chapter 5
das Amoreiras (7, 5620–5481 cal BCE) in the Sado valley. However, without
further analyses this correlation is not possible to establish as individuals with a
similar terrestrial signature could be originally from any other region. Likewise, the
only outlier known at Cabeço da Amoreira in the East Sado, presents a mixed
isotopic signature particularly coherent with the profiles of the West Sado as
indicated by the high nitrogen value. Similarly, one adult female buried at Cabeço
do Pez (CP1956, Sk5, 5711–5536 cal BCE) with the only mixed signature known in
this burial ground could be originally from any of the western populations.
16
15
14
13
δ15N (‰)
12
11
10
9
Moita do Sebastião, Tagus
Cabeço da Arruda, Tagus
Cabeço da Amoreira, Tagus
Arapouco, Sado (West)
Poças de S. Bento, Sado (West)
Cabeço das Amoreiras, Sado (East)
Vale de Romeiras, Sado (East)
Cabeço do Pez, Sado (East)
Várzea da Mó, Sado (East)
8
7
6
-22
-21
-20
-19
-18
-17
δ13C
-16
-15
-14
-13
-12
(‰)
Figure 5.44. Tagus and Sado valleys: correlation between the δ13C (‰) and δ15N (‰) values
from human bone collagen samples.
In summary:
• The individuals buried in the sites of the Tagus valley derive their
isotopes from mixed sources of terrestrial and marine-estuarine origin.
• At the Sado valley, the individuals buried in the western sites derive
their isotopes from mixed sources of terrestrial and marine-estuarine
442
Analysis and Synthesis of the Results
•
•
•
•
origin while those buried in the eastern sites obtained their protein from
terrestrial food sources.
The distinct isotopic profiles found in the western and eastern sites of
the Sado valley suggest that different groups were using distinct sites to
bury their dead.
The few isotopic outliers in each site suggest that these distinct groups
were rarely mixed.
The isotopic profile of each site at the Sado valley remains homogenous
over time while at the Tagus valley there is increased intra-site
heterogeneity during the later episodes of burial activity in the region.
The profiles remain characterized by the consumption of mixed
sources, but one group is more dependent on marine foods (mixed,
high marine-estuarine) while the other obtains their protein mainly from
terrestrial sources (mixed, high terrestrial).
The analysis does not indicate any particular pattern associated with age
or sex of the individual, which is possibly in part due to the poor preservation of the material which increases the difficulties of accurate
estimates.
The treatment of the dead: synthesis of the
archaeothanatological analysis
Archaeothanatological analysis was carried out on a total of 82 individuals: 33 were
excavated at the Tagus valley, and 49 at the Sado valley. For the material from the
Tagus valley, the analysis was particularly limited and constrained by the nature and
availability of the source material. In several cases the observations were restricted
to a few parts of the skeleton. Often, the analysis relied on indirect observations
only, such as the descriptions of the excavators. Despite the difficulties, it was
possible to obtain sufficient data supporting general observations regarding the
treatment of the dead, even when detailed analysis was constrained. The sample for
the Sado valley is better documented. The data set represents slightly less than half
of the burials known in the region, but the consistency of the results indicates that
the sample is representative of the assemblage. Observations supporting the
reconstruction of the treatment of the dead at the Tagus and Sado valleys are
presented in chapter 4.
Overall, the mortuary practices identified in the burial grounds of both valleys
are comparable, with strong similarities. The analysis suggests some intra- and
inter-site variations but these are minor and do not affect the overall pattern of the
treatment of the dead.
The cadavers were typically interred in individual grave features soon after
death, and retained their anatomical integrity (individual primary deposits). In
several cases the diagnostic criteria are unclear. However, in most cases the general
443
Chapter 5
topography of the skeletal elements in the features strongly indicates that the
corpses were placed in the graves before degradation of the labile joints,
supporting the interpretation of burials in primary position.
The secondary nature of the deposits was rare, and presumably absent in most
sites. These subtle practices, when detected, involved the manipulation of a few
selected bone elements. In only two cases (CAMS1958, Sk2; VR1959, Sk9), both
found in the Sado valley, is the data sufficient to support the hypothesis of secondary deposition of skeletal elements related to post-depositional manipulation of
the cadavers, although neither of the cases is conclusive. At the Tagus valley, one
case (MS1954–54, Sk23) described by the excavators as the isolated cranium of a
child deposited in a pit, is suggestive of a secondary deposition. Unfortunately, the
low resolution of the source material for the Tagus sites does not allow further
observations at this level. It is possible that contexts with disarticulated human
remains, containing for example just a few elements of the skeleton, were not
documented in the same detail as the complete burials. This could have been the
case for the nineteenth century excavations, but it is unlikely for the fieldwork
done in the twentieth century. The historical documentation shows a thorough
recovery system always when human remains were found, which is also detected in
the painstaking recovery of microliths, even the very small lithic debris.
In rare cases, the analysis confirms the multiple nature of primary deposits
containing the remains of two individuals. These striking episodes of double
burials are well attested at Arapouco (ARA1961, Sks 11, 12) by the synchronous
deposit of a adult female and a three year old child, and at Cabeço do Pez
(CP1956, Sks 9, 10) by the deposit of two adults.
In every burial ground, the bodies were normally laid on the back, with the
upper limbs arranged close to the body, and with the lower limbs in flexion at
various degrees, with the feet often placed towards the buttocks. In one case only
(MS1952–54, Sk9) were the lower limbs placed in complete extension. The knees
of the cadavers lying in the features were directed forwards, upwards or towards
the right or left side. In rare cases the knees were directed downwards
(CAMS1958, Sk7; VR1959, Sks 12, 19). The practice of laying the cadaver on the
lateral side was unusual in most sites. At Arapouco only, was this a common
alternative. Placement on the lateral side was further noted in two burials only at
Vale de Romeiras, and a possible one at Moita do Sebastião. This practice was
presumably absent elsewhere.
The size and shape of the grave features were rarely described by the
excavators. Archaeothanatological analysis suggests that the cadavers were placed
in pits dug into the ground, just wide enough to contain the body. The construction of these simple features was possibly dependent on the planned burial
position. However, it is clear that in most cases the design of the grave had significant impact on the position of the cadaver. Clumping of the body was common,
particularly at the level of the lower limbs which were commonly placed in
contraction. At some sites (Moita do Sebastião, Cabeço da Arruda, Arapouco, and
Cabeço das Amoreiras), the grave features were somewhat wider allowing a
444
Analysis and Synthesis of the Results
relatively expanded position of the cadaver, at the level of the upper or the lower
body. This was particularly common at Moita do Sebastião. In contrast, at Vale de
Romeiras the graves were notably narrow, and the cadavers were frequently
compressed, with their skeletons becoming hypercontracted in the feature. This
phenomenon is noted in two cases at Arapouco, and in one case at Cabeço do Pez,
but was apparently absent at Cabeço das Amoreiras. Hypercontraction of the body
is also noted at Cabeço da Arruda in at least three cases, but at Moita do Sebastião
this is occasionally noticed at the level of the lower limbs only, as the result of the
clumping of these elements.
Normally, the pressures observed on the bodies in the features can be explained
as the result of the physical limits of the pit. A few cases at Moita do Sebastião
(19th c., Sk10; 1952–54, Sks 9, 30), Arapouco (1961, Sks 6, 11, 12), and Cabeço das
Amoreiras (1958, Sk7), suggest that there might have been extra support offered by
some kind of soft container, such as a light wrapping of the cadaver at the time of
the burial. The low resolution of the source material available for Cabeço da
Arruda does not allow further observations at this level. At the other sites, the
wrapping hypothesis is unlikely.
The floor of the graves tended to be uneven and sloping from the upper to the
lower end of the feature, and often towards one side. Typically, the head was in a
more elevated position than the trunk and pelvic region. Unfortunately, the
fragmented state of most of the material does not allow detailed observations of
skeletal elements that could clarify aspects of the original position of the body in
the feature, in relation to the characteristics of the bottom, such as in the case of
bodies placed in a half-sitting position.
After the placement of the cadaver in the grave, the feature was covered with
sediment providing a filled space of decomposition. In several cases, the analysis
indicates rapid penetration of fluid sediment provided by a sand rich environment.
Arapouco at the Sado valley presents several good examples of rapid penetration
of sediment and the hourglass effect was often detected. In several cases the
documentation is unclear and/or the diagnostic criteria are not preserved. In every
case however, movement of the bones is limited suggesting that decomposition of
the cadaver occurred underground in a sediment-filled environment. A few cases
suggest mixed spaces of decomposition (ARA1961, Sks 6, 12; VR1959, Sks 5, 6) by
the formation of secondary empty spaces outside the volume of the cadaver within
an overall filled environment. The analysis of these features suggests the existence
of perishable material inserted in the grave along with the cadaver. Some of these
elements may have functioned as a cushion placed behind the upper body
(ARA1961, Sk6; VR1959, Sk6), while other elements were possibly placed in front
of the upper body (VR1959, Sk5). One exceptional case (ARA1961, Sk12) suggests
the existence of an element functioning as a container, such as a pouch or a basket
accommodating the body. As discussed, neither of these interpretations are conclusive but they strongly suggest the existence of other elements in the graves,
placed along with the cadaver.
445
Chapter 5
Typically, the cadavers lie underground and undisturbed in their graves. Some
skeletons were found disturbed in the feature, possibly due to the disposal of a
new cadaver. The practice of reduction of the skeletal elements of an individual to
accommodate the burial of a new cadaver was presumably common at Moita do
Sebastião. This practice was also noted at Cabeço da Arruda, and possibly at Vale
de Romeiras. These gestures, rather than suggesting disrespect towards the principle of integrity of the body, strongly indicate the continued preference of certain
areas for the burial of the dead.
In rare cases the data suggest the post-depositional manipulation of selected
bone elements in association with secondary deposits of skeletal elements. One
further case that may suggest post-depositional manipulation of the cadavers is
ARA1961, Sk11. The skull (cranium and mandible) of this adult female was
missing from the feature. As discussed in chapter 4, it is unclear if this was a case
of disturbance or deliberate removal of the skeletal element. However, it is
interesting to entertain the hypothesis, although not possible to confirm it, that the
isolated cranium found at Cabeço das Amoreiras (CAMS1958, Sk2) could belong
to this female buried at Arapouco.
The main principles governing the mortuary practices at the Tagus and Sado
valleys are remarkably consistent and homogeneous, as suggested by the archaeothanatological analysis.
Few cases present deviant elements, but exceptional instances were noted in
almost every site. Despite the common core characteristics of burial practices, the
data indicate some particulars that may distinguish practices carried out in the
different burial grounds. Site specificities are apparently not consistent with any
specific trend over time, but further chronological data may refine this observation.
These are noted particularly at the level of the characteristics of the grave features,
as well as in the initial position of the cadavers, and can be briefly outlined as
follows:
• At Arapouco, the bodies were often placed on the lateral side with the
limbs in flexion arranged close to the body. While this position was not
the most common at the site, the lateral placement of the body was rare,
and largely absent in other burial grounds.
• The individuals buried at Cabeço das Amoreiras (MNI 8) were
predominantly mature adult males (n=4). The remaining individuals
were two juveniles, one represented by its cranium, and another adult
(male or female), the sex of which could not be determined. At other
sites, the burial populations were representative of natural populations.
• While at all sites most cadavers were placed in constrained grave pits,
this aspect was particularly striking at Vale de Romeiras. At Vale de
Romeiras, the bodies were normally laid in very small pits in contracted
positions, with the limbs nested on the upper body and the feet forced
towards the buttocks.
• In slight contrast, at Moita do Sebastião, while the lower limbs were
often placed in a constrained position with the feet forced towards the
446
Analysis and Synthesis of the Results
•
buttocks, the lower limbs were possibly never rotated towards the upper
body.
At Cabeço da Arruda and Cabeço do Pez the analysis did not reveal any
particular characteristic of the initial position of the bodies or the grave
features, most likely due to the low resolution of the source material.
These are slight variations on a common theme. Further analyses of this material
may provide additional elements about subtle and particular practices in the
different burial grounds.
Overall, the mortuary programme at the Tagus and Sado valleys was consistent
and based on strong common principles indicating a shared mindset on the
treatment of the dead (fig. 5.45).
Figure 5.45. This illustration presents several typical elements observed in the burials at the
Tagus and Sado valleys. The cadavers were normally laid on the back with the upper limbs
placed close to the body and the lower limbs in contraction with the feet forced towards
the buttocks. The floor of the graves tended to be sloping. The head of the cadaver was
often rotated forwards and downwards towards the body. The physical limits of the feature
had a significant impact on the final arrangement of the body in the feature. The walleffects and consequent pressures are visible in the contracted position of most skeletons
and can be explained as the result of the design of the grave.
(Drawing by Susanna Berglund)
447
PART THREE: DEATH IN PLACE
Riddaren: Då är levandet en orimlig fasa. Ingen människa kan leva med Döden för
ögonen och vetskapen om alltings intighet.
Döden: De flesta människor reflekterar varken över Döden eller intigheten.
Riddaren: En dag står de ju ändå på livets yttersta näs.
Döden: Ja den dagen…
(Bergman 1957)
Introduction to part three
The human remains analysed in this dissertation provide concrete archaeological
evidence for an interpretation of the significance of burial activity at the sites of the
Tagus and Sado valleys, as outlined in the beginning of this study. Mortuary
evidence offers a strong basis for inferences about broad historical processes and
explanations about the human engagement with death in particular, with a focus
on the burial phenomenon. The narrative of this dissertation is built on the
archaeological data obtained in the course of this study under three main lines of
enquiry:
• the reconstruction of the mortuary gestures,
• the chronological framework of the burial activity,
• and the definition of the people buried in these places by the
identification of their dietary choices.
To provide additional evidence supporting the central arguments in this dissertation, I call upon further empirical data, such as studies of lithic technology, faunal
evidence, palaeoecological analyses, osteological analyses, and palaeodemography.
Empirical evidence is evaluated through the lens of theory and concepts that
relate most closely to the aspects of mortuary behaviour of interest in this
dissertation:
• the manipulation of human remains,
• the places for the dead,
• and the (re)construction of social memories in mortuary ritual context.
The arguments supporting the analysis of the mortuary practices in the context of
site formation processes are justified in a historical framework, on the basis of the
concepts of place, social memory, and mortuary ritual practice, as presented in chapter 2.
The explicit use of this framework allows us to link the actions of the people using
these sites to the materiality of their mortuary practices recovered in the archaeological context. The use of these concepts is twofold. First, they frame the human
impact on the formation and maintenance of these burial grounds. Second, and
additionally, they allow inferences about the significance of mortuary practices and
attitudes towards death on these people.
In the last part of this dissertation, I return to the central themes of this
research: death, memory, and place. I evaluate the empirical evidence by tracing the
boundaries of the mortuary phenomenon under study, in terms of practice, time,
space, and people. Then, I discuss this particular engagement with death, in terms
of the significance of the formal practice of human burial in dedicated spaces, and
its relation to the formation and maintenance of these shell midden sites. Based on
the empirical evidence available, I suggest an interpretation of the sites at the
Tagus and Sado valleys by stitching together the spheres of death, memory and
451
place with the thread of historical agency in hunter-gatherer communities. In the
final section of this thesis, Conclusion: on the history of death and the last hunter-gatherers of
the south-western Iberian Peninsula, I recapitulate the central arguments in this dissertation and emphasize the need for broad examination of the history of human relationships with death in the chrono-cultural periods that frame the Late Mesolithic
in the south-western Iberian Peninsula.
452
Chapter 6
The Places for the Dead as Places of Shared
Narratives: Discussion
Introduction
Archaeological evidence indicates a strong core set of common practices associated
with the treatment of the cadaver during the Late Mesolithic at the Tagus and Sado
valleys. This is illustrated by the attitudes towards the body after death, which is
particular to these regions, by contrast with contemporaneous sites in the Iberian
Peninsula (Peyroteo Stjerna, forthcoming). In this dissertation, I suggest that the
burial activity at the Tagus and Sado valleys is a significant historical practice with a
central role in the formation and maintenance of these sites, rather than a
peripheral activity resulting from the practical need to dispose a cadaver. The
reasons supporting the structural significance of the burial activity at these sites can
be outlined as follows:
• The topographical placement of the dead within the landscape of the living is a
central aspect in the social dynamics of the last hunter-gatherers of the
Tagus and Sado valleys.
• Hunter-gatherer communities were structured within specific social
scenarios in a historically situated world view. At the Tagus and Sado
valleys, people were bound by shared social and historical practices, including the
formation and maintenance of burial grounds.
• The formation and maintenance of burial grounds is a primary means of
history making. Burial practices are non-literary depositional narratives
of an historical process. The burial of the dead in these burial grounds was an
act of historical consciousness, and was motivated by the encounter and
interaction with other bands frequenting the same environments.
• Social practices such as death rituals are acts of cultural production with
a powerful memory aid. Mortuary ritual practices had a central role in the
formation and maintenance of these sites as meaningful social environments.
• Death rituals in general and burial practice in particular played a central
role in the life of these hunter-gatherers, and in developing a sense of
community within the landscape as well as maintaining the social ties in
both life and death.
453
Chapter 6
These reasons support the argument for the structural significance of the burial
activity at the Tagus and Sado valleys. They are articulated from the evidence
obtained in the course of this study: radiocarbon and isotopic evidence, and
archaeothanatological data. Studies of lithic technology (Araújo 1995–97; Diniz
and Nukushina 2014; Marchand 2001, 2005; Nukushina 2012; Roche 1967, 1972),
faunal evidence (Detry 2003, 2007; Lentacker 1986, 1994; Marques-Gabriel 2015;
Rowley-Conwy 2015), palaeoecological analyses (Mateus and Queiróz 1993; van
der Schriek et al. 2007), palaeodemographic data, and osteological analysis of
human remains (Cunha and Umbelino 1995–97; Jackes, forthcoming; Jackes and
Meiklejohn 2008), provide further empirical evidence supporting my arguments.
The shell middens of the Tagus and Sado valleys have been interpreted as
hunter-gatherer territorial claims to establish control over economic resources
(Arnaud 1989, 621). This is a compelling explanation; however, this approach does
not address the question of the significance of mortuary ritual practices in the
formation and maintenance of these places (Peyroteo Stjerna 2015). A broad and
widely used argument to explain mortuary contexts in relation to the people under
study, was presented by Saxe (1970) and Binford (1971), and links mortuary
treatment to social structure (see chapter 1). In this view, the treatment of the dead
reflects the dimensions of the social persona, such as age, sex, social rank and
affiliation (Binford 1971). In this framework, death rituals are interpreted as
indicative of the complexity of the social structure, and understood as claims of
corporate groups to resources and territories (Saxe 1970). The merit of these
explanations of the burial phenomenon lies in their simplicity, applicability to a
diversity of empirical evidence, and general independence of chronological and
spatial boundaries. However, these explanations tend to emphasize certain
elements, such as the analysis of funerary architecture and grave goods,
interpreting the elaboration of mortuary practices as representative of the
complexity of a given society, which correlation is speculative in most cases.
In the following sections, I evaluate the empirical data at the level of the burial
practice, Burial as the chosen way of treating the cadaver: boundaries of the phenomenon, and at
the level of the spaces dedicated to these burials, Death in place. These two scales,
the burial and the burial ground, are discussed further within an historical framework, based on the concepts of place, social memory, and mortuary ritual practice.
In the following section, Places of mortuary narratives, I aim to explicitly connect these
theoretical concepts to the materiality of the mortuary practices of the huntergatherers under study. Following this, I suggest an explanation for the significance
of the burial practices at the Tagus and Sado valleys. This interpretative scenario is
based on the arguments listed above supported by multiple lines of archaeological
evidence.
454
Discussion
Burial as the chosen way of treating the cadaver: boundaries
of the phenomenon
Treatment of the dead
During the Late Mesolithic at the sites of the Tagus and Sado valleys, the burial of
the cadaver was a typical way of dealing with the body of the deceased. The
cadavers were placed in individual grave features, soon after death, while retaining
their anatomical integrity. After laying the body in the grave, the feature was
immediately covered with sediment. The metamorphosis of the cadaver through
the process of decomposition occurred largely underground. The physical
transformation of the individual was hidden from the sphere of the living, and the
bodies remained undisturbed in the subterranean world. Transformation was
concealed, but the dead were afforded a common and permanent location within
the landscape of the living, in an accessible open air place. The placement of the
body in the grave pit was careful and meticulous. The last image of the deceased
was lifelike and faithful to what the individual looked like when he/she was alive in
a position of repose. For the hunter-gatherers of the Tagus and Sado valleys, burial
of the cadaver in a dedicated place for the dead was a common practice. While we
cannot know their intentions, or why this was the chosen way of dealing with the
dead, some aspects can be considered that distinguish the practice of primary
burial in dedicated common burial grounds from other mortuary practices, which
seem to be particularly significant to these sites:
• relevance of the integrity of the body;
• transformation of the cadaver was hidden from the living and happened
underground;
• lifelike position and last image of the deceased before concealment in
the grave feature;
• the dead were afforded a common and permanent place in the
landscape of the living, in an open air and accessible place.
Unusual practices, such as the placement of more than one individual in the grave
feature, or the occasional post-depositional manipulation of skeletal elements, did
not affect the principles governing burial practice at these sites. Exceptional
instances highlight central aspects about the proper way of dealing with the
cadavers. The double burial identified at Arapouco (ARA1961, Sks 11, 12) is a
striking example of the significance of the lifelike position of the deceased in the
grave. The young woman and child in this feature were not just disposed of
simultaneously in the grave. The adult was placed holding the child on her body
and the pair were clearly arranged in a realistically lifelike position (see chapter 4).
This unusual case demonstrates the considerable effort from the living to stage the
dead in a familiar way. Likewise, the few post-depositional manipulations of the
identified cadavers, presumably happen after decomposition of most soft tissue.
455
Chapter 6
These are exceptional instances and demonstrate the significance of the principle
of hidden metamorphosis of the cadaver. On occasion, manipulation of skeletal
elements seems to have been related to processes of reduction. The episodes of
reduction could be used to argue against the principle of the relevance of the
integrity of the body. However, this type of handling was unusual in most sites,
and could be explained by ill marked graves and consequent unintentional manipulation of the remains. Here, the practice of reduction suggests the relevance of
these particular places for the burial of the dead, without intentional disrespect
towards the principle of the integrity of the body.
Chronological and spatial boundaries
The way of dealing with the cadaver in the Tagus and Sado valleys contrasts with
earlier and contemporary practices known from other regions in the Iberian
Peninsula (Peyroteo Stjerna, forthcoming). In the Early Mesolithic, the presence of
human remains was more common in cave or rock shelter contexts, a pattern that
decreases in occurrence through time during the Mesolithic. Yet this practice was
not abandoned when open air sites became the most common option for the
disposal of human remains. At the same time, during the Late Mesolithic in Iberia,
there was an increment of sites with human remains accompanied by the intensification of burial practice. Nevertheless, the statistical increase of burial practice in
open air sites in the Peninsula is due to the material from the Tagus and Sado
valleys; without these sites the data for both Early and Late Mesolithic does not
differ (Peyroteo Stjerna, forthcoming). Burial as a mortuary practice was known in
the Iberian Peninsula but it was never a common practice until the Late Mesolithic
in these two river valleys.
This is a Late Mesolithic phenomenon which in the Iberian Peninsula was clustered in the Tagus and Sado valleys. Overall, these sites were used for burial activity more or less synchronously. In both the Tagus and Sado valleys, burial activity
was more frequent between c 6000 and 5500 cal BCE. In both valleys, there were
early episodes of human burial activity, during the very first phases of the Late
Mesolithic in Portugal, around 6350 cal BCE. After c 5500 cal BCE, the frequency
of burial practice decreases dramatically, in both valleys, with a few last episodes
not later than c 5000 cal BCE contemporaneous with several Early Neolithic sites
in Portugal. Two apparent outliers are presented by the burial of a child at Cabeço
do Pez, and one individual at Poças de S. Bento, in the Sado valley, around 4300
cal BCE, both coeval with Middle Neolithic sites in Portugal. The burials of
Neolithic chronology require further research to understand their significance in
these sites. At present, other than the calendrical date, nothing in the graves
indicates that these individuals are Neolithic and not hunter-gatherers.
Death rituals observed at the sites of the Tagus and Sado valleys are materialized in the practice of human burial in dedicated burial grounds. This singular engagement with death is contemporaneous with the reconfiguration of the settlement pattern and other cultural changes that define and distinguish the Early and
456
Discussion
the Late Mesolithic in Portugal (Araújo 2015). During both periods, archaeological
sites are located in places corresponding to the innermost areas of palaeoestuaries.
However, while in the Early Mesolithic the sites were scattered throughout the
territory by small rivers, in the Late Mesolithic the settlement pattern was clustered
in the large estuaries of the Atlantic period, such as in the Tagus and Sado, which
were newly occupied during this later phase. The concentration of people in very
specific regions has no parallel in the Early Mesolithic, when highly mobile groups
probably of small single-family units, moved regularly between the littoral and the
interior regions in logistic and short-term residential camps (Araújo 2012, 2015).
As pointed out by A. C. Araújo (2015) the convergence of people to the confined
and ecologically rich new estuarine environments of the Atlantic have certainly
triggered many of the Late Mesolithic behaviours, such as the use of long-term
residential sites, the new lithic technology, and a new attitude towards death.
Current explanations for the reconfiguration of the settlement pattern from the
Early to the Late Mesolithic (Bicho et al. 2010; Zilhão 2003) are based on palaeoclimatic evidence which indicates significant changes in the local ecosystems
around 8200 years ago, c 6250 BCE (Mateus and Queiróz 1993; van der Schriek
et al. 2007). The earliest 14C dates at the Tagus and Sado valleys correspond to the
start of burial activity in some of the sites. These are the first moments of
occupation of these sites, c 6350 cal BCE (fig. 6.1), and are coeval with
environmental changes in Portugal and elsewhere. In fact, at the Tagus valley
where the palaeogeography and palaeoenvironment are well known, the burial
activity is more frequent between c 6000–5600 cal BCE, coinciding with the rapid
establishment of inner estuarine tidal mudflat and saltmarsh environments c 6100
cal BCE (van der Schriek et al. 2007). While the chronological match between
environmental changes and the new settlement pattern is not a coincidence, the
particular social mechanisms allowing this territorial and social reconfiguration
remain unclear (Araújo 2015).
Likewise, the processes leading to the decrease of burial activity around 5500
cal BCE, at both the Tagus and Sado valleys, are not well established, and in this
case, the environmental factors seem to play a less important role. The decrease of
burial activity in the two valleys is coeval with the first phase of the Early Neolithic
in Portugal, between c 5500–5300 cal BCE (Carvalho 2010). It coincides with
environmental changes well documented in the Tagus valley, defined by the
contraction of estuarine habitats, between c 5500–3800 cal BCE, and the establishment of an open landscape around c 5000 cal BCE (van der Schriek et al.
2007). In this case, the socio-cultural explanation for the end of the phenomenon,
defined by the expansion of the Neolithic life style, seems to be more compelling,
because despite the decline of the environmental conditions, the estuarine
environment in the Tagus valley was still productive and the same mix of habitats
remained until c 3800 cal BCE (van der Schriek et al. 2007). Palaeoenvironmental
reconstruction of the Sado valley is under study (Costa et al. 2015; Freitas et al.
2013), and the spatial and temporal extension of the marine influence on the valley
remains unclear. However, preliminary data seem to indicate that the less frequent
457
Chapter 6
use of the Sado sites is not correlated to a palaeoenvironmental event (Diniz and
Arias 2012).
Figure 6.1. Chronological data for the burial activity in the Tagus and Sado valleys plotted
with two major events: palaeoclimatic change 8200 years ago (c 6250 BCE) and the onset
of the Neolithic in Portugal.
458
Discussion
The living and the dead
Clustering of the settlement pattern observed from the Early to the Late Mesolithic
was the outcome of the convergence of small and highly mobile groups to an
apparently more confined territory, broadly concentrated in the large estuaries of
the Tagus and Sado valleys. This shift brought together small groups probably
organized in single families (Araújo 2015) which then lived side by side. The
organization of these groups in the new territories is poorly understood, and it is
unclear whether these groups merged, at least partially, or if they remained
autonomous. Traditionally, the hunter-gatherers of the Tagus and Sado valleys
have been studied as two independent groups. This divide exists primarily from the
segmented research history of these sites (see chapter 2). Research teams focused
on one or the other valley, and centred their efforts on particular sites, to the
detriment of long-range aspects of the archaeology of the last hunter-gatherers. In
the following sections, I discuss the evidence supporting the unity of the
populations frequenting the sites of the Tagus and Sado valleys, and the
contrasting data suggesting that autonomous groups lived side by side and buried
their dead in neighbouring burial grounds.
Seasonal model of settlement pattern
Early interpretations of the settlement pattern at the Sado valley (Arnaud 1989)
were based on the concept of logistic movement (Binford 1980) applied to multiple
lines of evidence, such as faunal remains, material culture, and site location. In this
model, most of the population living in the valley moved seasonally between large
base camps, occupying one site during the spring/summer (e.g., Poças de S.
Bento), and moving to another, just a few kilometres away (fig. 2.3), for the
autumn/winter (e.g., Cabeço do Pez), while making short incursions to smaller
temporary camps for specialized activities. This explanation for the settlement
pattern in the Sado valley was a preliminary model based on limited data which
needed further evidence for its validation, refinement or rejection (Arnaud 1989),
but so far, no other model was so well systematized. The sites of the Tagus valley
were never interpreted in such explicit manner and have been broadly referred to
as long-term residential sites of semi-sedentary hunter-gatherer groups; based on
the generally large size of the sites, the high number of human burials in each site,
and the several structures identified only at Moita do Sebastião which have been
interpreted as postholes and storage pits, thus suggesting a permanent domestic
use of the site (Roche 1972).
Faunal analyses are somewhat contradictory. While some authors emphasize the
seasonal use of the sites within each valley (Lentacker 1986; Marques-Gabriel 2015,
312; Rowley-Conwy 2015), it has also been noted that the evidence may indicate a
year-round use of the sites. The non-seasonal use of the sites was pointed out by
A. Lentacker (1986, 22) in her study of the Tagus assemblages. Lentacker identified
seasonal patterns of some species, but emphasized that the remains of mammals
such as wild rabbit, red deer, and wild boar, possibly representing all age459
Chapter 6
categories, could indicate visits to the sites throughout the year, instead of stays
during particular seasons. This pattern was supported further by the study of shell
remains at the sites of the Sado valley carried out by M. Deith (Arnaud 1990, 440)
which indicated that shellfish were collected in the autumn/winter at all sites, by
contrast with the suggestion of sites specialized in spring/summer vs autumn/winter activities. Conversely, and like the seasonal model proposed by J.
Arnaud (1989) for the Sado valley, a broad zooarchaeological model advanced for
the Portuguese Late Mesolithic sites suggested short seasonal movement between
summer and winter base camps, and radial incursions out from each to logistic
camps (Rowley-Conwy 2015). In this model, based on the differential representation of skeletal parts, P. Rowley-Conwy (2015) suggested that Cabeço da Arruda
was possibly a winter hunting camp across the Muge River from the base camps of
Moita do Sebastião and Cabeço da Amoreira, where hunting parties camped and
prepared deer carcasses for transport back across the river to the base camps.
According to this author, the data seemed also to accommodate the alternative
hypothesis that the Muge sites may not have been occupied in the summer. At the
Sado valley, the faunal analysis suggested that the eastern sites which are located
relatively more inland, such as Vale de Romeiras and Cabeço do Pez, were used in
the winter, while in the summer, the groups moved west using Arapouco as a
logistic fishing camp, and Poças de S. Bento as a rabbit hunting site (RowleyConwy 2015).
Seasonal models imply that the burial activity followed a cyclical movement, and
the deceased were buried wherever the group was stationed at the time. In this
scenario, the populations having access to a variety of foods throughout the year,
from marine-estuarine to terrestrial origin, should present individual isotopic
signatures indicative of a more or less balanced intake of protein from seasonal
reservoirs. More significantly, the itinerancy between the sites, with the burial of
the deceased occurring according to the location of the group would result in low
inter-site variability at the level of the dietary patterns of the burial population,
which does not seem to be the case at least at the Sado valley, as discussed below.
Dietary patterns
Tagus valley
At the Tagus valley, isotopic data of carbon and nitrogen shows low to moderate
inter-site and intra-individual variability suggesting that the people buried in the
valley derived their protein from similar sources. The analysis over time indicates a
general homogeneity of the isotopic signatures suggesting a balanced consumption
of marine-estuarine and terrestrial foods. This pattern changes during the later
episodes of burial activity, when two dietary groups (mixed, high terrestrial and mixed,
high marine-estuarine) were buried at Moita do Sebastião and Cabeço da Amoreira,
while apparently at Cabeço da Arruda only one of these groups (mixed, high
terrestrial) was buried. The pattern is confirmed with the last burial known at
Cabeço da Arruda presenting a contrasting profile, similar to those of mixed, high
460
Discussion
marine-estuarine signatures found at Moita do Sebastião and Cabeço da Amoreira in
preceding burial episodes. The heterogeneity noted towards the last phases of
burial activity is identified in all sites and each burial ground accommodates individuals from both dietary groups (mixed, high terrestrial and mixed, high marineestuarine), further supporting the similar patterns in the use of these sites for burial
practice.
The dietary patterns observed in the Tagus valley suggest similar sources of
protein consumption varying only in proportion during the later phases. In either
case, the foods could be obtained locally in the estuarine environment and/or from
inland territories. Current dietary data is not conclusive in terms of possible
differences between the groups buried in each site. Further analyses on other
isotopic elements, such as strontium or barium may identify different geographic
patterns (Burton et al. 2003; Price et al. 2002) within this population, which could
potentially clarify similarities and differences that are masked at present.
Sado valley
Dietary patterns observed at the Sado valley are somewhat surprising and do not
conform to current seasonal models for the settlement pattern in the region.
Unlike the sites at the Tagus valley, the inter-site variability at the Sado valley is
high, indicating that the people buried in this valley derived their protein from
different sources. In contrast, the intra-site variability is low, suggesting that the
people in each burial ground shared similar patterns of protein consumption.
The burial populations in the western sites of the Sado valley, at Arapouco and
Poças de S. Bento, present a dietary pattern similar to those buried in the Tagus
valley, indicating the balanced consumption of marine-estuarine and terrestrial
foods. However, this is not the case in the eastern portion of the valley, where the
burial populations at Cabeço das Amoreiras, Vale de Romeiras, Cabeço do Pez,
and Várzea da Mó derived their protein from a full terrestrial diet. Interestingly,
this dual dietary pattern is coherent with the zooarchaeological evidence, which
indicates a significant consumption of marine-estuarine sources in the western sites
while the remains of terrestrial foods dominate the inventories of the eastern sites
(Arnaud 1989). The dualism found in the faunal remains is one of the pillars of the
seasonal models, proposing that most of the population would live in the eastern
hunting sites during the autumn/winter, moving to the west in the spring/summer
when the marine-estuarine environment was more productive. The isotopic data of
the human remains suggests otherwise. Dietary patterns of the burial populations
indicate a division of the valley, which is not only geographic and ecological but
also according to the population. The burial grounds of the west (Arapouco and
Poças de S. Bento) were used by populations with a similar mixed dietary pattern
while the burial grounds of the east (Cabeço das Amoreiras, Vale de Romeiras,
Cabeço do Pez, and Várzea da Mó) were used by populations with a full terrestrial
diet. Dietary patterns and zooarchaeological evidence are in agreement and suggest
a strong territorial pattern of people using the western sites exclusively while
another group was concentrated in the eastern portion of the valley.
461
Chapter 6
At the Sado valley, dietary patterns remain consistent over time strengthening
the population segmentation of the territory, both in terms of food catchment areas
and burial grounds. Current isotopic data is not clear in terms of possible differences between subgroups in the west and in the east, and further analyses should
be pursued. At the Sado valley, the isotopic evidence of carbon and nitrogen
shows that distinct groups were using these sites not only simultaneously but with
clear territorial borders, one defined in the west and the other defined in the east.
Despite the dietary differences, the association of aquatic shells with the burial
grounds remains a common trait, in the east and in the west (see below Cemeteries
and shell middens in the Tagus and Sado valleys).
Lithic analyses
Occasionally, the archaeological material from the two valleys was investigated
from a broad perspective (Arnaud 1987; Carvalho 2009). When the valleys are
compared some striking aspects emerge. One remarkable difference is related to
the choices of raw material. People at the Tagus valley worked with regional flint
(Araújo 1993; Cardoso 2004; Carvalho 2003, 2008; Zilhão 2000), while people at
the Sado valley used local raw material (Pimentel et al. 2015) with generally poor
knapping properties (Araújo 1995–97; Arnaud 1989, 2000; Diniz and Nukushina
2014; Marchand 2001; Nukushina 2012). The Tagus region is rich in sources of
quality flint for knapping; however, this raw material was not transported to the
Sado valley, c 100 kilometres away. The choice of raw material indicates no movement of stone resources from one valley to the other, supporting the idea of selfcontained valleys with separate groups clustering towards one or the other region.
Another striking difference between the valleys lies in the procurement strategies
of lithic raw material. While at the Tagus valley, flint with good knapping properties was obtained at a regional level, not at the immediacies of the sites, at the Sado
valley the choice of materials was micro-local, with the use of pebbles of siliceous
slates, chert, jasper, and quartz found in close proximity to the sites, although with
inferior knapping quality (Pimentel et al. 2015).
The use of rock from the immediate vicinity is consistent with the local lithic
production well documented at the Sado valley sites. All lithic studies are
unanimous that all stages of the chaîne opératoire are present in the site’s
assemblages, suggesting that lithic tools were produced, used, and discarded at the
site (Araújo 1995–97; Diniz and Nukushina 2014; Marchand 2001; Nukushina
2012). The presence of all stages of the chaîne opératoire of lithic production seems
also to be the case at the Tagus sites (Marchand 2001).
While the local tool production does not contradict the seasonal models, the
proposed specialized character of the sites seems more difficult to uphold, based
on current lithic evidence. In the seasonal scenario, the lithic tools should present a
certain degree of specialization and heterogeneity, at least between the logistic sites,
which does not seem to be the case. In the Sado valley, recent lithic studies (Diniz
and Nukushina 2014; Pimentel et al. 2015) challenged the specialized character of
the sites as proposed in the seasonal models. This is the case of Arapouco, which
462
Discussion
has been interpreted as a specialized fishing camp (Arnaud 1989, 2000; RowleyConwy 2015) due to its downstream location, and relative abundance of marineestuarine fish remains in relation to other sites in the valley (Marques-Gabriel 2015;
Rowley-Conwy 2015). Lithic analysis presents an unspecialized toolkit dominated
by multitasking tools, such as retouched flakes (69%) among other “common
fond” tools (end-scrapers, denticulated flakes, and others), which seems inconsistent with the specialized interpretation of the site (Diniz and Nukushina 2014).
The analysis suggests that this pattern of tool production cannot be explained by
constraints of the raw material, and specialized tools could be produced using the
available rock. The raw material used at Arapouco for the production of these
flakes is predominantly siliceous rocks, of equal or better quality than the raw
material used at Cabeço das Amoreiras, which was mainly used for the production
of specialized tools such as geometric microliths (Nukushina 2012; Pimentel et al.
2015). So far, the specialization of the lithic industry has not been verified in other
sites either (Nukushina 2012). Lithic studies demonstrate and confirm the use of
locally available rock but most significantly, they indicate that the raw material used
in each site is different, suggesting a preference for rock available in the immediate
vicinity (Pimentel et al. 2015).
The frequency of retouched tools, such as geometric microliths, is a common
trait at the Late Mesolithic sites (Araújo 2003; Arnaud 1987; Carvalho 2009). Their
typological frequency has been used to determine the relative chronology of the
sites (Roche 1972), or to suggest particular functions for each site (Arnaud 1989;
Marchand 2001). The most common microlith is the trapeze. These are dominant
in the assemblages of Moita do Sebastião and Cabeço da Arruda in the Tagus
valley (Roche 1967, 1972), and Arapouco, Poças de S. Bento, and possibly Vale de
Romeiras in the Sado valley (Araújo 1995–97; Diniz and Nukushina 2014;
Nukushina 2012). Triangles are predominant only at Cabeço da Amoreira in the
Tagus valley. These are residual at Moita do Sebastião and Cabeço da Arruda
(Roche 1972), and at all the Sado sites (Nukushina 2012, 78). Segments are
abundant in the assemblages of the eastern portion of the Sado valley, at Cabeço
das Amoreiras, Cabeço do Pez, and Várzea da Mó (Marchand 2001; Nukushina
2012), but are rare at the Tagus valley, found in the assemblages of Cabeço da
Amoreira only. While there is some chronological patterning, with the trapezes
being more common in the sites with earlier dates, and the triangles and segments
in the sites with the most recent dates, this variability does not clearly support arguments towards different functions of the sites, which in fact seem typologically
and technologically homogeneous. At present, the typological and technological
data are insufficient and future comparative analyses of the assemblages are
necessary to clarify and understand homogeneity and heterogeneity in the lithic
material of both valleys.
Other data
Other lines of evidence provide additional elements for the analysis of the burial
populations in the sites at the Tagus and Sado valleys. In this section I summarize
463
Chapter 6
and evaluate central aspects of various osteological analyses and palaeodemographic data from both valleys, along with preliminary observations about objects
of adornment recovered at the Tagus sites. These data highlight aspects of
heterogeneity between the sites, further supporting the hypothesis of autonomous
groups living side by side and burying their dead in neighbouring burial grounds.
Tagus valley
Osteological analysis of the skeletal material from Moita do Sebastião and Cabeço
da Arruda presents several elements suggesting heterogeneity between the
populations of each site, which can be outlined as follows:
• Individuals buried at Cabeço da Arruda present a higher fertility rate
than those at Moita do Sebastião, as suggested by the reconstruction of
demographic profiles analysed in terms of population growth (or
decline) and fertility (Jackes and Meiklejohn 2008). According to these
authors, this difference suggests an expanding population at Cabeço da
Arruda while the values from Moita do Sebastião are aligned with a
typical hunter-gatherer population with a slow growth rate.
• The burial population at Cabeço da Arruda contains a higher number of
older adults than at Moita do Sebastião, suggesting a greater life expectancy at Cabeço da Arruda (Jackes, forthcoming).
• Dental wear is stronger in the individuals buried at Moita do Sebastião
than in those at Cabeço da Arruda (Jackes, forthcoming).
• Frequency of dental caries is higher at Moita do Sebastião than at
Cabeço da Arruda. In contrast with the material from Cabeço da
Arruda, at Moita do Sebastião caries are common at late adolescence,
and older adults present very high levels of caries (Jackes 2009, forthcoming).
• Bone cortex is on average thinner in the osteological series of Cabeço
da Arruda than in the series of Moita do Sebastião (Jackes, forthcoming;
Jackes and Meiklejohn 2008, 218). These authors suggested that this
could be explained by lower activity levels of the population buried at
Cabeço da Arruda and/or by the consumption of high oxalic acid and
salt rich plants, which could be found in the estuary, suggesting high
reliability on foods found in the immediacies of the site (Jackes, forthcoming; Jackes and Meiklejohn 2008, 218).
According to M. Jackes and C. Meiklejohn (2008), the different patterns observed
between Moita do Sebastião and Cabeço da Arruda seem to suggest higher levels
of sedentism of the latter population. In this scenario, differential sedentism could
explain the higher fertility levels noted at Cabeço da Arruda, as well as the indicators of reduced physical activity and the eventual higher reliance on local plants
(Jackes, forthcoming; Jackes and Meiklejohn 2008). This scenario remains hypothetical, and further research should clarify the suggested model of divergent sedentism levels between the populations of Moita and Arruda. While the isotopic
464
Discussion
signatures indicate the consumption of protein from similar sources, the differential dental wear and caries observed in the assemblages of Moita do Sebastião and
Cabeço da Arruda suggests different food habits and/or genetic differences
between the burial groups.
Objects of adornment analysed by J. Roche (1959) indicate differential use of
these elements, further supporting the idea of heterogeneity between the burial
populations at Moita do Sebastião and Cabeço da Arruda, as well as at Cabeço da
Amoreira. Pierced shells are abundant at Moita do Sebastião and were found in
significant quantities in several graves (MS1952–54, Sks 1, 5, 6, 8, 11, 25). Pierced
shells were also found at Cabeço da Amoreira, but were very rare at Cabeço da
Arruda where only three elements were identified. Stone pendants made from raw
materials that could be found locally (Roche 1959, 409) were also found but are
rare, with four examples known at Moita do Sebastião, two from Cabeço da
Amoreira, and possibly one from Cabeço da Arruda. Four teeth found at Cabeço
da Amoreira were possibly used as pendants. This type of pendant is unknown at
both Moita do Sebastião and Cabeço da Arruda. J. Roche’s (1959) observations
suggest a distinction in use of objects of adornment between the individuals buried
in the different sites, and while pierced shells are common at Moita do Sebastião,
these elements were not used by the individuals buried at Cabeço da Arruda.
Unfortunately, comparative observations of the objects of adornment are not
available for the sites at the Sado valley.
Sado valley
Palaeodemographic analysis of the skeletal material from the Sado valley indicates a
common demographic pattern between the populations buried in each site
(Umbelino 2006, 146):
• The proportion of adults (c 80%) and non-adults (20%) buried in each
site is almost identical (Cunha et al. 2003), with the exception of Várzea
da Mó where just one burial of an adult was found.
• Sexual distribution is difficult to determine due to preservation of the
material. However, most sites present a balanced number of male and
female burials (Arapouco, Vale de Romeiras, and Cabeço do Pez), with
the exception of Cabeço das Amoreiras where all adults seem to be
male. Sexual determination was not possible at Poças de S. Bento
(Cunha and Umbelino 1995–97).
• Age at death is equally difficult to establish but it seems that all sites accommodate non-adults, adults, and individuals over 50 years old (Cunha
and Umbelino 1995–97; Cunha et al. 2003).
Frequency of dental caries in the Sado valley series indicates some variation and
heterogeneity between the sites, even attending to differential preservation of the
material (Cunha et al. 2003) (fig. 6.2). In a sample of 48 individuals, a total of 18
present caries (38%). Arapouco (7/15) and Vale de Romeiras (6/12) present a
similar pattern of c 50% of the individuals with caries, despite the marked
465
Chapter 6
differences in the sources of protein consumption. In contrast, at Cabeço do Pez,
located next to Vale de Romeiras, only 2 out of 17 individuals had caries (12%).
Cabeço das Amoreiras, also with a terrestrial isotopic profile, presents the highest
percentage of individuals with caries (3/4, 75%).
Figure 6.2. Burial grounds in the Sado valley with indication of sources of protein
(mixed/terrestrial) and frequency of caries (F.C.) of the burial populations.
The combined evaluation of the patterns of dietary sources of protein and
frequency of caries for each burial population at the Sado valley, offers some
indicators about possible subgroups coexisting within the eastern group defined by
the terrestrial protein pattern. Unfortunately, due to poor preservation of the
material from Poças de S. Bento, dental analysis has not been possible, and without
further data, it remains unclear if the burial populations of the western sites,
Arapouco and Poças de S. Bento, correspond to one or more groups. In the
eastern group, characterized by the consumption of terrestrial protein, two subgroups emerge: one with high frequency of caries buried at Cabeço das Amoreiras
and Vale de Romeiras, and the other with low levels of caries buried at Cabeço do
Pez. Despite the common pattern of protein consumption, these subgroups
present different patterns of frequency of caries suggesting different food habits
and/or genetic differences between the burial groups.
The common demographic pattern and inter-site variability of dietary patterns,
seems to support the scenario of different burial grounds being used by particular
groups where natural populations were buried according to their group of origin.
466
Discussion
Side by side: geographic segmentation of shared landscapes
Several lines of evidence indicate that the burial grounds at the Tagus and Sado
valleys were used side by side by different units of hunter-gatherers, suggesting the
geodemographic segmentation of the valleys. This term, borrowed from geographical
marketing research (see See and Openshaw 2001) is used in this dissertation to
express the phenomenon of coexistence of different groups of a population