McDONALD INSTITUTE MONOGRAPHS

Temple landscapes Fragility, change and resilience of Holocene environments in the Maltese Islands

By Charles French, Chris O. Hunt, Reuben Grima, Rowan McLaughlin, Simon Stoddart & Caroline Malone

Volume 1 of Fragility and Sustainability – Studies on Early , the ERC-funded FRAGSUS Project Temple landscapes

McDONALD INSTITUTE MONOGRAPHS

Temple landscapes Fragility, change and resilience of Holocene environments in the Maltese Islands

By Charles French, Chris O. Hunt, Reuben Grima, Rowan McLaughlin, Simon Stoddart & Caroline Malone

With contributions by Gianmarco Alberti, Jeremy Bennett, Maarten Blaauw, Petros Chatzimpaloglou, Lisa Coyle McClung, Alan J. Cresswell, Nathaniel Cutajar, Michelle Farrell, Katrin Fenech, Rory P. Flood, Timothy C. Kinnaird, Steve McCarron, Rowan McLaughlin, John Meneely, Anthony Pace, Sean D.F. Pyne-O’Donnell, Paula J. Reimer, Alastair Ruffell, George A. Said-Zammit, David C.W. Sanderson, Patrick J. Schembri, Sean Taylor, David Trump†, Jonathan Turner, Nicholas C. Vella & Nathan Wright

Illustrations by Gianmarco Alberti, Jeremy Bennett, Sara Boyle, Petros Chatzimpaloglou, Lisa Coyle McClung, Rory P. Flood, Charles French, Chris O. Hunt, Michelle Farrell, Katrin Fenech, Rowan McLaughlin, John Meneely, Anthony Pace, David Redhouse, Alastair Ruffell, George A. Said-Zammit & Simon Stoddart

Volume 1 of Fragility and Sustainability – Studies on Early Malta, the ERC-funded FRAGSUS Project This project has received funding from the European Research Council (ERC) under the European Union’s Seventh Framework Programme (FP7-2007-2013) (Grant agreement No. 323727).

Published by: McDonald Institute for Archaeological Research University of Cambridge Downing Street Cambridge, UK CB2 3ER (0)(1223) 339327 [email protected] www.mcdonald.cam.ac.uk

McDonald Institute for Archaeological Research, 2020

© 2020 McDonald Institute for Archaeological Research. Temple landscapes is made available under a Creative Commons Attribution-NonCommercial- NoDerivatives 4.0 (International) Licence: https://creativecommons.org/licenses/by-nc-nd/4.0/

ISBN: 978-1-902937-99-1

Cover design by Dora Kemp and Ben Plumridge. Typesetting and layout by Ben Plumridge.

On the cover: View towards Nadur lighthouse and Għajnsielem church with the Channel to Malta beyond, from In-Nuffara (Caroline Malone).

Edited for the Institute by James Barrett (Series Editor). Contents

Contributors xi Figures xiii Tables xvi Preface and dedication xix Acknowledgements xxi Foreword xxiii

Introduction Caroline Malone, Simon Stoddart, Chris O. Hunt, Charles French, 1 Rowan McLaughlin & Reuben Grima 0.1. Introduction 1 0.2. Background to FRAGSUS as an archaeological project 3 0.3. Environmental research in Malta and the Mediterranean 5 0.4. The development of the FRAGSUS Project and its questions 6 0.5. Archaeological concerns in Maltese prehistory and the FRAGSUS Project 8 0.6. The research programme: the sites and their selection 9 0.7. Investigating the palaeoenvironmental context 10 0.8. Archaeological investigations 11

Part I The interaction between the natural and cultural landscape – insights into the 17 fifth–second millennia bc

Chapter 1 The geology, soils and present-day environment of Gozo and Malta 19 Petros Chatzimpaloglou, Patrick J. Schembri, Charles French, Alastair Ruffell & Simon Stoddart 1.1. Previous work 19 1.2. Geography 19 1.3. Geology 21 1.4. Stratigraphy of the Maltese Islands 23 1.4.1. Lower Coralline Limestone Formation 23 1.4.2. Globigerina Limestone Formation 23 1.4.3. Chert outcrops 25 1.4.4. Blue Clay Formation 26 1.4.5. Greensand Formation 28 1.4.6. Upper Coralline Limestone Formation 28 1.4.7. Quaternary deposits 29 1.5. Structural and tectonic geology of the Maltese Islands 29 1.6. Geomorphology 29 1.7. Soils and landscape 31 1.8. Climate and vegetation 32

Chapter 2 Chronology and stratigraphy of the valley systems 35 Chris O. Hunt, Michelle Farrell, Katrin Fenech, Charles French, Rowan McLaughlin, Maarten Blaauw, Jeremy Bennett, Rory P. Flood, Sean D. F. Pyne-O’Donnell, Paula J. Reimer, Alastair Ruffell, Alan J. Cresswell, Timothy C. Kinnaird, David Sanderson, Sean Taylor, Caroline Malone, Simon Stoddart & Nicholas C. Vella 2.1. Methods for dating environmental and climate change in the Maltese Islands 35 Rowan McLaughlin, Maarten Blaauw, Rory P. Flood, Charles French, Chris O. Hunt, Michelle Farrell, Katrin Fenech, Sean D.F. Pyne-O’Donnell, Alan J. Cresswell, David C.W. Sanderson, Timothy C. Kinnaird, Paula J. Reimer & Nicholas C. Vella 2.1.1. Data sources for chronology building 35 2.1.2. Pottery finds 41

v 2.2. Basin infill ground penetrating radar surveys 41 Alastair Ruffell, Chris O. Hunt, Jeremy Bennett, Rory P. Flood, Simon Stoddart & Caroline Malone 2.2.1. Rationale 41 2.2.2. Geophysics for basin fill identification 41 2.2.3. Valley locations 43 2.3. The sediment cores 43 Chris O. Hunt, Michelle Farrell, Rory P. Flood, Katrin Fenech, Rowan McLaughlin, Nicholas C. Vella, Sean Taylor & Charles French 2.3.1. Aims and methods 43 2.3.2. The core descriptions 49 2.3.3. Magnetic susceptibility and XRF analyses of the cores 59 2.4. Age-depth models 64 Maarten Blauuw & Rowan McLaughlin 2.4.1. Accumulation rates 64 2.5. A local marine reservoir offset for Malta 65 Paula J. Reimer 2.6. Major soil erosion phases 65 Rory P. Flood, Rowan McLaughlin & Michelle Farrell 2.6.1. Introduction 65 2.6.2. Methods 66 2.6.3. Results 67 2.6.4. Discussion 68 2.6.5. Conclusions 71

Chapter 3 The Holocene vegetation history of the Maltese Islands 73 Michelle Farrell, Chris O. Hunt & Lisa Coyle McClung 3.1. Introduction 73 Chris O. Hunt 3.2. Palynological methods 74 Lisa Coyle-McClung, Michelle Farrell & Chris O. Hunt 3.3. Taxonomy and ecological classification 75 Chris O. Hunt 3.4. Taphonomy 75 Chris O. Hunt & Michelle Farrell 3.5. The pollen results 87 Michelle Farrell, Lisa Coyle-McClung & Chris O. Hunt 3.5.1. The Salina cores 87 3.5.2. Wied Żembaq 87 3.5.3. Xemxija 87 3.5.4. In-Nuffara 87 3.5.5. Santa Verna 95 3.5.6. Ġgantija 105 3.6. Synthesis 107 3.6.1. Pre-agricultural landscapes (pre-5900 cal. bc) 107 3.6.2. First agricultural colonization (5900–5400 cal. bc) 108 3.6.3. Early Neolithic (5400–3900 cal. bc) 109 3.6.4. The later Neolithic Temple period (3900–2350 cal. bc) 110 3.6.5. The late Neolithic–Early Bronze Age transition (2350–2000 cal. bc) 111 3.6.6. The Bronze Age (2000–1000 cal. bc) 112 3.6.7. Late Bronze Age, Punic and Classical periods (c. 1000 cal. bc to ad 1000) 112 3.6.8. Medieval to modern (post-ad 1000) 113 3.7. Conclusions 113

vi Chapter 4 Molluscan remains from the valley cores 115 Katrin Fenech, Chris O. Hunt, Nicholas C. Vella & Patrick J. Schembri 4.1. Introduction 115 4.2. Material 117 4.3. Methods 117 4.4. Radiocarbon dates and Bayesian age-depth models 117 4.5. Results 117 4.5.1. Marsaxlokk (MX1) 127 4.5.2. Wied Żembaq (WŻ) 127 4.5.3. Mġarr ix-Xini (MĠX) 128 4.5.4. Marsa 2 128 4.5.5. Salina Deep Core 133 4.5.6. Xemxija 1 and 2 152 4.6. Interpretative discussion 153 4.6.1. Erosion – evidence of major events from the cores 153 4.7. Environmental reconstruction based on non-marine molluscs 155 4.7.1. Early Holocene (c. 8000–6000 cal. bc) 155 4.7.2. Mid-Holocene (c. 6000–3900 cal. bc) 155 4.7.3. Temple Period (c. 3900–2400 cal. bc) 155 4.7.4. Early to later Bronze Age (2400–c. 750 cal. bc) 155 4.7.5. Latest Bronze Age/early Phoenician period to Late Roman/Byzantine 156 period (c. 750 cal. bc–cal. ad 650) 4.8. Concluding remarks 156 4.9. Notes on selected species 157 4.9.1. Extinct species 157 4.9.2. Species with no previous fossil record 158 4.9.3. Other indicator species 158

Chapter 5 The geoarchaeology of past landscape sequences on Gozo and Malta 161 Charles French & Sean Taylor 5.1. Introduction 161 5.2. Methodology and sample locations 164 5.3. Results 165 5.3.1. Santa Verna and its environs 165 5.3.2. Ġgantija temple and its environs 174 5.3.3. Skorba and its immediate environs 183 5.3.4. Taċ-Ċawla settlement site 188 5.3.5. Xagħra town 190 5.3.6. Ta’ Marżiena 192 5.3.7. In-Nuffara 192 5.3.8. The Ramla valley 193 5.3.9. The Marsalforn valley 195 5.3.10. Micromorphological analyses of possible soil materials in the Xemxija 1, 196 Wied Żembaq 1, Marsaxlokk and Salina Deep (SDC) cores 5.4. The Holocene landscapes of Gozo and Malta 213 5.5. A model of landscape development 217 5.6. Conclusions 221

Chapter 6 Cultural landscapes in the changing environments from 6000 to 2000 bc 223 Reuben Grima, Simon Stoddart, Chris O. Hunt, Charles French, Rowan McLaughlin & Caroline Malone 6.1. Introduction 223 6.2. A short history of survey of a fragmented island landscape 223 6.3. Fragmented landscapes 225 vii 6.4. The Neolithic appropriation of the landscape 227 6.5. A world in flux (5800–4800 cal. bc) 227 6.6. The fifth millennium bc hiatus (4980/4690 to 4150/3640 cal. bc) 228 6.7. Reappropriating the landscape: the ‘Temple Culture’ 230 6.8. Transition and decline 236 6.9. Conclusion 237

Part II The interaction between the natural and cultural landscape – insights from 239 the second millennium bc to the present: continuing the story

Chapter 7 Cultural landscapes from 2000 bc onwards 241 Simon Stoddart, Anthony Pace, Nathaniel Cutajar, Nicholas C. Vella, Rowan McLaughlin, Caroline Malone, John Meneely & David Trump† 7.1. An historiographical introduction to the Neolithic–Bronze Age transition 241 into the Middle Bronze Age 7.2. Bronze Age settlements in the landscape 243 7.3. The Bronze Age Phoenician transition and the Phoenician/Punic landscape 246 7.4. Entering the Roman world 250 7.5. Arab 250 7.6. Medieval 251 7.7. The Knights and the entry into the modern period 251

Chapter 8 The intensification of the agricultural landscape of the MalteseArchipelago 253 Jeremy Bennett 8.1. Introduction 253 8.2. The Annales School and the Anthropocene 254 8.3. The Maltese Archipelago and the longue durée of the Anthropocene 255 8.4. Intensification 257 8.5. Population 258 8.5.1. Sub-carrying capacity periods 258 8.5.2. Post-carrying capacity periods 260 8.6. The agrarian archipelago 262 8.6.1. The agricultural substrate 262 8.6.2. The development of agricultural technology 262 8.7. Discussion: balancing fragility and sustainability 264

Chapter 9 Locating potential pastoral foraging routes in Malta through the use of a 267 Geographic Information System Gianmarco Alberti, Reuben Grima & Nicholas C. Vella 9.1. Introduction 267 9.2. Methods 267 9.2.1. Data sources 267 9.2.2. Foraging routes and least-cost paths calculation 268 9.3. Results 271 9.3.1. Garrigue to garrigue least-cost paths 271 9.3.2. Stables to garrigues least-cost paths 273 9.4. Discussion 276 9.4. Conclusions 283

Chapter 10 Settlement evolution in Malta from the Late Middle Ages to the early twentieth 285 century and its impact on domestic space George A. Said-Zammit 10.1. The Medieval Period (ad 870–1530) 285 10.1.1. Medieval houses 288 viii 10.1.2. Giren and hovels 289 10.1.3. Cave-dwellings 292 10.1.4. Architectural development 292 10.2. The Knights’ Period (ad 1530–1798) 293 10.2.1. The phase ad 1530–1565 293 10.2.2. The phase ad 1565–1798 293 10.2.3. Early modern houses 294 10.2.4. Lower class dwellings 297 10.2.5. Cave-dwellings and hovels 298 10.2.6. The houses: a reflection of social and economic change 298 10.3. The British Period (ad 1800–1900) 298 10.3.1. The houses of the British Period 299 10.3.2. The effect of the Victorian Age 300 10.3.3. Urban lower class dwellings 301 10.3.4. Peasant houses, cave-dwellings and hovels 301 10.4. Conclusions 302

Chapter 11 Conclusions 303 Charles French, Chris O. Hunt, Michelle Farrell, Katrin Fenech, Rowan McLaughlin, Reuben Grima, Nicholas C. Vella, Patrick J. Schembri, Simon Stoddart & Caroline Malone 11.1. The palynological record 303 Chris O. Hunt & Michelle Farrell 11.1.1. Climate 303 11.1.2. Farming and anthropogenic impacts on vegetation 307 11.2. The molluscan record 308 Katrin Fenech, Chris O. Hunt, Nicholas C. Vella & Patrick J. Schembri 11.3. The soil/sediment record 310 Charles French 11.4. Discontinuities in Maltese prehistory and the influence of climate 313 Chris O. Hunt 11.5. Environmental metastability and the longue durée 314 Chris O. Hunt 11.6. Implications for the human story of the Maltese Islands 316 Charles French, Chris O. Hunt, Caroline Malone, Katrin Fenech, Michelle Farrell, Rowan McLaughlin, Reuben Grima, Patrick J. Schembri & Simon Stoddart

References 325

Appendix 1 How ground penetrating radar (GPR) works 351 Alastair Ruffell

Appendix 2 Luminescence analysis and dating of sediments from archaeological sites 353 and valley fill sequences Alan J. Cresswell, David C.W. Sanderson, Timothy C. Kinnaird & Charles French A2.1. Summary 353 A2.2. Introduction 354 A2.3. Methods 355 A2.3.1. Sampling and field screening measurements 355 A2.3.2. Laboratory calibrated screening measurements 355 A2.4. Quartz OSL SAR measurements 356 A2.4.1. Sample preparation 356 A2.4.2. Measurements and determinations 356 ix A2.5. Results 357 A2.5.1. Sampling and preliminary luminescence stratigraphies 357 A2.5.2. Gozo 357 A2.5.3. Skorba 363 A2.5.4. Tal-Istabal, Qormi 363 A2.6. Laboratory calibrated screening measurements 363 A2.6.1. Dose rates 367 A2.6.2. Quartz single aliquot equivalent dose determinations 367 A2.6.3. Age determinations 371 A2.7. Discussion 372 A2.7.1. Ġgantija Temple (SUTL2914 and 2915) 372 A2.7.2. Ramla and Marsalforn Valleys (SUTL2917–2923) 373 A2.7.3. Skorba Neolithic site (SUTL2925–2927)s 373 A2.7.4. Tal-Istabal, Qormi (SUTL2930) 376 A2.7. Conclusions 376

Appendix 2 – Supplements A–D 379

Appendix 3 Deep core borehole logs 401 Chris O. Hunt, Katrin Fenech, Michelle Farrell & Rowan McLaughlin

Appendix 4 Granulometry of the deep cores 421 (online edition only) Katrin Fenech

Appendix 5 The molluscan counts for the deep cores 441 (online edition only) Katrin Fenech

Appendix 6 The borehole and test excavation profile log descriptions 535 Charles French & Sean Taylor

Appendix 7 The detailed soil micromorphological descriptions from the buried soils and 549 Ramla and Marsalforn valleys Charles French A7.1. Santa Verna 549 A7.2. Ġgantija Test Pit 1 551 A7.3. Ġgantija WC Trench 1 552 A7.4. Ġgantija olive grove and environs 553 A7.5. Skorba 553 A7.6. Xagħra town 554 A7.7. Taċ-Ċawla 555 A7.8. In-Nuffara 555 A7.9. Marsalforn Valley Profile 626 556 A7.10. Ramla Valley Profile 627 556 A7.11. Dwerja 556

Appendix 8 The micromorphological descriptions for the Malta deep cores of Xemxija 1, 557 Wied Żembaq 1, Marsaxlokk and the base of the Salina Deep Core (21B) Charles French & Sean Taylor

Appendix 9 The charcoal data 563 Nathan Wright

Index 565

x Contributors

Dr Gianmarco Alberti Dr Rory P. Flood Department of Criminology, Faculty for Social School of Natural and Built Environment, Queen’s Wellbeing, University of Malta, Msida, Malta University, University Road, Belfast, Northern Email: [email protected] Ireland Email: [email protected] Jeremy Bennett Department of Archaeology, University of Prof. Charles French Cambridge, Cambridge, UK Department of Archaeology, University of Email: [email protected] Cambridge, Cambridge, UK Email: [email protected] Dr Maarten Blaauw School of Natural and Built Environment, Queen’s Dr Reuben Grima University, University Road, Belfast, Northern Department of Conservation and Built Heritage, Ireland University of Malta, Msida, Malta Email: [email protected] Email: [email protected]

Dr Petros Chatzimpaloglou Dr Evan A. Hill Department of Archaeology, University of School of Natural and Built Environment, Queen’s Cambridge, Cambridge, UK University, University Road, Belfast, Northern Email: [email protected] Ireland Email: [email protected] Dr Lisa Coyle McClung School of Natural and Built Environment, Queen’s Prof. Chris O. Hunt University, University Road, Belfast, Northern Faculty of Science, Liverpool John Moores Ireland University, Liverpool, UK Email: [email protected] Email: [email protected]

Dr Alan J. Cresswell Dr Timothy C. Kinnaird SUERC, University of Glasgow, East Kilbride, School of Earth and Environmental Sciences, University of Glasgow, Glasgow, Scotland University of St Andrews, St. Andrews, Scotland Email: [email protected] Email: [email protected]

Nathaniel Cutajar Prof. Caroline Malone Deputy Superintendent of Cultural Heritage, School of Natural and Built Environment, Queen’s Heritage Malta, Valletta, Malta University, University Road, Belfast, BT7 1NN, Email: [email protected] Northern Ireland Email: [email protected] Dr Michelle Farrell Centre for Agroecology, Water and Resilience, Dr Steve McCarron School of Energy, Construction and Environment, Department of Geography, National University of Coventry University, Coventry, UK Ireland, Maynooth, Ireland Email: [email protected] Email: [email protected]

Dr Katrin Fenech Dr Rowan McLaughlin Department of Classics & Archaeology, University School of Natural and Built Environment, Queen’s of Malta, Msida, Malta University, University Road, Belfast, Northern Email: [email protected] Ireland Email: [email protected]

xi John Meneely Prof. Patrick J. Schembri School of Natural and Built Environment, Queen’s Department of Biology, University of Malta, University, University Road, Belfast, Northern Msida, Malta Ireland Email: [email protected] Email: [email protected] Dr Simon Stoddart Dr Anthony Pace Department of Archaeology, University of UNESCO Cultural Heritage, Valletta, Malta Cambridge, Cambridge, UK Email: [email protected] Email: [email protected]

Dr Sean D.F. Pyne-O’Donnell Dr Sean Taylor Earth Observatory of Singapore, Nanyang Department of Archaeology, University of Technological University, Singapore Cambridge, Cambridge, UK Email: [email protected] Email: [email protected]

Prof. Paula J. Reimer Dr David Trump† School of Natural and Built Environment, Queen’s University, University Road, Belfast, Northern Dr Jonathan Turner Ireland Department of Geography, National University Email: [email protected] of Ireland, University College, Dublin, Ireland Email: [email protected] Dr Alastair Ruffell School of Natural and Built Environment, Queen’s Prof. Nicholas C. Vella University, University Road, Belfast, Northern Department of Classics and Archaeology, Faculty Ireland of Arts, University of Malta, Msida, Malta Email: [email protected] Email: [email protected]

George A. Said-Zammit Dr Nathan Wright Department of Examinations, Ministry for School of Social Science, The University of Education and Employment, Government of Malta, Queensland, Brisbane, Australia Malta Email: [email protected] Email: [email protected]

Prof. David C.W. Sanderson SUERC, University of Glasgow, East Kilbride, University of Glasgow, Glasgow, Scotland Email: [email protected]

xii Figures

0.1 Location map of the Maltese Islands in the southern Mediterranean Sea. 2 0.2 Location of the main Neolithic archaeological and deep coring sites investigated on Malta and Gozo.  11 0.3 Some views of previous excavations on Malta and Gozo.  12–13 0.4 Some views of recent excavations.  14 1.1 The location of the Maltese Islands in the southern Mediterranean Sea with respect to Sicily and North Africa.  20 1.2 Stratigraphic column of the geological formations reported for the Maltese Islands.  22 1.3 Geological map of the Maltese Islands.  22 1.4 Typical coastal outcrops of Lower Coralline Limestone, forming sheer cliffs.  23 1.5 Characteristic geomorphological features developed on the Lower Coralline Limestone in western Gozo (Dwerja Point).  24 1.6 The Middle Globigerina Limestone at the Xwejni coastline. 24 1.7 An overview of the area investigated in western Malta. 25 1.8 The end of the major fault system of Malta (Victorian Lines) at Fomm Ir-Riħ.  26 1.9 An overview of the western part of Gozo where the chert outcrops are located. 27 1.10 Chert outcrops: a) and c) bedded chert, and b) and d) nodular chert. 27 1.11 Four characteristic exposures of the Blue Clay formation on Gozo and Malta.  28 1.12 Map of the fault systems, arranged often as northwest–southeast oriented graben, and strike-slip structures.  30 2.1 Summary of new radiocarbon dating of Neolithic and Bronze Age sites on Gozo and Malta.  36 2.2 Summed radiocarbon ages for the main sediment cores.  36 2.3 The location of the Birżebbuġa Għar Dalam and Borġ in-Nadur basins and their GNSS-located GPR lines.  42 2.4 The core locations in Malta and Gozo.  44 2.5 Radiocarbon activity in settlement cores.  48 2.6 The Xemxija 2 core by depth. 51 2.7 The Wied Żembaq 1 and 2 cores by depth. 52 2.8 The Mġarr ix-Xini core by depth.  54 2.9 The Marsaxlokk 1 core and part of 2 by depth.  55 2.10 The resistivity and magnetic susceptibility graphs for Xemxija 1 core.  60 2.11 The resistivity and magnetic susceptibility graphs for Xemxija 2 core.  60 2.12 The multi-element data plots for Xemxija 1 core.  61 2.13 The multi-element data plots for Wied Żembaq 1 core.  62 2.14 The multi-element data plots for Marsaxlokk 1 core.  63 2.15 RUSLE models of soil erosion for the Maltese Islands in September and March.  69 2.16 R and C factors and their product.  70 3.1 Valley catchments and core locations in the Mistra area of Malta.  79 3.2 The modern pollen spectra. 81 3.3 Pollen zonation for the Salina Deep Core.  82–3 3.4 Pollen zonation for the Salina 4 core.  88–9 3.5 Pollen zonation for the Wied Żembaq 1 core.  92–3 3.6 Pollen zonation for the Xemxija 1 core.  96–7 3.7 Pollen zonation for the pit fills at In-Nuffara.  101 3.8 Pollen and palynofacies from the buried soils below the temple at Santa Verna.  102 3.9 Pollen and palynofacies from Test Pit 1 on the southwestern edge of the Ġgantija platform.  104 3.10 Photomicrographs (x800) of key components of the palynofacies at Santa Verna and Ġgantija.  106 4.1 Marsaxlokk 1 molluscan histogram.  120 4.2 Wied Żembaq 1 molluscan histogram.  122 4.3 Mġarr ix-Xini molluscan histogram.  129 4.4 Marsa 2 molluscan histogram.  134 4.5 Salina Deep Core molluscan histogram.  138 4.6 Marine molluscan histogram for the Salina Deep Core.  139 xiii 4.7 Xemxija 1 molluscan histogram.  144 4.8 Base of Xemxija 2 molluscan histogram.  145 5.1 Location map of the test excavation/sample sites and geoarchaeological survey areas on Gozo and Malta.  164 5.2 Plan of Santa Verna temple and the locations of the test trenches.  166 5.3 Santa Verna excavation trench profiles all with sample locations marked.  167 5.4 The red-brown buried soil profiles in Trench E, the Ashby and Trump Sondages within the Santa Verna temple site.  170 5.5 Santa Verna soil photomicrographs.  172–3 5.6 Plan of Ġgantija temple and locations of Test Pit 1 and the WC Trench excavations, with as-dug views of the WC Trench and TP1.  175 5.7 Section profiles of Ġgantija Test Pit 1 on the southwest side of Ġgantija temple and the east-west section of the Ġgantija WC Trench on the southeast side. 176 5.8 Ġgantija TP 1 photomicrographs.  178 5.9 Ġgantija WC Trench 1 photomicrographs.  180 5.10 Section profiles of Trench A at Skorba showing the locations of the micromorphological and OSL samples.  183 5.11 Skorba Trench A, section 1, photomicrographs.  185 5.12 Skorba Trench A, section 2, photomicrographs.  186 5.13 Taċ-Ċawla soil photomicrographs.  189 5.14 A typical terra rossa soil sequence in Xagħra town at construction site 2.  191 5.15 Xagħra soil photomicrographs.  191 5.16 In-Nuffara photomicrographs.  193 5.17 The Marsalforn (Pr 626) and Ramla (Pr 627) valley fill sequences, with the micromorphology samples and OSL profiling/dating loci marked.  194 5.18 Ramla and Marsalforn valley profiles soil photomicrographs.  195 5.19 Photomicrographs of the Blue Clay and Greensand geological substrates from the Ramla valley.  199 5.20 Xemxija 1 deep valley core photomicrographs.  202 5.21 Wied Żembaq 1 deep valley core photomicrographs.  206 5.22 Marsaxlokk and Salina Deep Core photomicrographs.  210 5.23 Scrub woodland on an abandoned terrace system and garrigue plateau land on the north coast of Gozo.  213 5.24 Terracing within land parcels (defined by modern sinuous lanes) on the Blue Clay slopes of the Ramla valley with Xagħra in the background.  216 6.1 The location of the Cambridge Gozo Project survey areas.  224 6.2 Fieldwalking survey data from around A. Ta Kuljat, B. Santa Verna, and C. Għajnsielem on Gozo from the Cambridge Gozo survey and the FRAGSUS Project.  227 6.3 The first cycle of Neolithic occupation as recorded by the Cambridge Gozo survey using kernel density analysis for the Għar Dalam, Red Skorba and Grey Skorba phases.  229 6.4 The first half of the second cycle of Neolithic occupation as recorded by the Cambridge Gozo survey using kernel density analysis implemented for the Żebbuġ and Mġarr phases. 232 6.5 The second half of the second cycle of Neolithic occupation as recorded by the Cambridge Gozo survey using kernel density analysis for the Ġgantija and Tarxien phases. 233 7.1 Kernel density analysis of the Tarxien Cemetery, Borġ in-Nadur and Baħrija periods for the areas covered by the Cambridge Gozo survey. 244 7.2a The evidence for Bronze Age settlement in the Mdina area on Malta. 245 7.2b The evidence for Bronze Age settlement in the Rabat (Gozo) area. 245 7.3 Distribution of Early Bronze Age dolmen on the Maltese Islands.  246 7.4 Distribution of presses discovered in the Mġarr ix-Xini valley during the survey. 248 7.5 The cultural heritage record of the Punic tower in Żurrieq through the centuries. 249 7.6 The changing patterns of social resilience, connectivity and population over the course of the centuries in the Maltese Islands.  252 8.1 An oblique aerial image of the northern slopes of the Magħtab land-fill site, depicting landscaping efforts including ‘artificial’ terracing.  256 8.2 RUSLE estimates of areas of low and moderate erosion for Gozo and Malta.  259 9.1 a) Sheep being led to their fold in Pwales down a track; b) Sheep grazing along a track on the Bajda Ridge in Xemxija, Malta.  269 xiv 9.2 Least-cost paths (LCPs), connecting garrigue areas, representing potential foraging routes across the Maltese landscape.  271 9.3 Density of LCPs connecting garrigue areas to random points within the garrigue areas themselves. 272 9.4 Location of ‘public spaces’, with size proportional to the distance to the nearest garrigue-to-garrigue LCP.  273 9.5 LCPs connecting farmhouses hosting animal pens to randomly generated points within garrigue areas in northwestern (A) and northeastern (B) Malta. 274 9.6 As for Figure 9.5, but representing west-central and east-central Malta. 274 9.7 As for Figure 9.5, but representing southern and southwestern Malta. 275 9.8 Location of ‘public spaces’, with size proportional to the distance to the nearest outbound journey.  276 9.9 a) Public space at Tal-Wei, between the modern town of Mosta and Naxxar; b) Tal-Wei public space as represented in 1940s survey sheets. 277 9.10 Approximate location of the (mostly disappeared) raħal toponyms.  279 9.11 Isochrones around farmhouse 4 representing the space that can be covered at 1-hour intervals considering animal walking speed.  280 9.12 Isochrones around farmhouse 2 representing the space that can be covered at 1-hour intervals considering animal walking speed (grazing while walking).  281 9.13 a) Isochrones around farmhouse 5 representing the space that can be covered at 1-hour intervals; b) Isochrones around farmhouse 6; c) Isochrones around farmhouse 7. 282 10.1 The likely distribution of built-up and cave-dwellings in the second half of the fourteenth century. 286 10.2 The lower frequency of settlement distribution by c. ad 1420.  286 10.3 The distribution of settlements just before ad 1530.  288 10.4 The late medieval Falson Palace in Mdina.  289 10.5 A girna integral with and surrounded by stone dry walling. 290 10.6 A hovel dwelling with a flight of rock-cut steps.  291 10.7 The hierarchical organisation of settlements continued, with the addition of Valletta, Floriana and the new towns around Birgu. 295 10.8 An example of a seventeenth century townhouse with open and closed timber balconies. 296 10.9 An example of a two-storey razzett belonging to a wealthier peasant family.  297 10.10 The distribution of built-up settlements in about ad 1900.  299 10.11 An example of a Neo-Classical house.  301 11.1 Summary of tree and shrub pollen frequencies at 10 sample sites. 304 11.2 Summary of cereal pollen frequencies at 14 sample sites.  305 11.3 Schematic profiles of possible trajectories of soil development in the major geological zones of Malta and Gozo. 311 11.4 The main elements of a new cultural-environmental story of the Maltese Islands throughout the last 10,000 years.  317 A2.1 Marsalforn valley, Gozo.  360 A2.2 Marsalforn valley, Gozo.  361 A2.3 Ramla valley, Gozo.  361 A2.4 Ġgantija Test Pit 1, Gozo.  361 A2.5 Skorba Neolithic site; trench A, East section; trench A, South section.  362 A2.6 Skorba, Trench A, South section.  362 A2.7 Tal-Istabal, Qormi, Malta.  364 A2.8 Tal-Istabal, Qormi, Malta.  364 A2.9 Photograph, showing locations of profile sample and OSL tubes, and luminescence-depth profile, for the sediment stratigraphy sampled in profile 1. 365 A2.10 Photograph, and luminescence-depth profile, for the sediment stratigraphy sampled in profile 3.  365 A2.11 Photograph, and luminescence-depth profile, for the sediment stratigraphy sampled in profile 2.  366 A2.12 Photograph, and luminescence-depth profile, for the sediment stratigraphy sampled in profiles 4 and 6.  366 A2.13 Photograph, and luminescence-depth profile, for the sediment stratigraphy sampled in profile 5.  367 A2.14 Apparent dose and sensitivity for laboratory OSL and IRSL profile measurements for SUTL2916 (P1). 370 A2.15 Apparent dose and sensitivity for laboratory OSL and IRSL profile measurements for SUTL2920 (P2). 370 A2.16 Apparent dose and sensitivity for laboratory OSL and IRSL profile measurements for SUTL2913 (P3). 370 A2.17 Apparent dose and sensitivity for laboratory OSL and IRSL profile measurements for SUTL2924 (P4). 370 xv A2.18 Apparent dose and sensitivity for laboratory OSL and IRSL profile measurements for SUTL2929 (P5). 371 A2.19 Apparent dose and sensitivity for laboratory OSL and IRSL profile measurements for SUTL2928 (P6). 371 A2.20 Apparent dose and sensitivity for laboratory OSL and IRSL profile measurements for SUTL2931 (P7). 371 A2.21 Probability Distribution Functions for the stored dose on samples SUTL2914 and 2915.  374 A2.22 Probability Distribution Functions for the stored dose on samples SUTL2917–2919.  374 A2.23 Probability Distribution Functions for the stored dose on samples SUTL2921–2923.  375 A2.24 Probability Distribution Functions for the stored dose on samples SUTL2925–2927.  375 A2.25 Probability Distribution Function for the stored dose on sample SUTL2930.  376 SB.1 Dose response curves for SUTL2914. 385 SB.2 Dose response curves for SUTL2915. 385 SB.3 Dose response curves for SUTL2917. 386 SB.4 Dose response curves for SUTL2918. 386 SB.5 Dose response curves for SUTL2919. 387 SB.6 Dose response curves for SUTL2921. 387 SB.7 Dose response curves for SUTL2922. 388 SB.8 Dose response curves for SUTL2923. 388 SB.9 Dose response curves for SUTL2925. 389 SB.10 Dose response curves for SUTL2926. 389 SB.11 Dose response curves for SUTL2927. 390 SB.12 Dose response curves for SUTL2930. 390 SC.1 Abanico plot for SUTL2914.  391 SC.2 Abanico plot for SUTL2915.  391 SC.3 Abanico plot for SUTL2917.  392 SC.4 Abanico plot for SUTL2918.  392 SC.5 Abanico plot for SUTL2919.  392 SC.6 Abanico plot for SUTL2921.  393 SC.7 Abanico plot for SUTL2922.  393 SC.8 Abanico plot for SUTL2923.  393 SC.9 Abanico plot for SUTL2925.  394 SC.10 Abanico plot for SUTL2926.  394 SC.11 Abanico plot for SUTL2927.  394 SC.12 Abanico plot for SUTL2930.  395 SD.1 Apparent ages for profile 1, with OSL ages. 397 SD.2 Apparent ages for profile 2, with OSL ages. 397 SD.3 Apparent ages for profile 3, with OSL ages. 398 SD.4 Apparent ages for profiles 4 and 6, with OSL ages.  398 SD.5 Apparent ages for profile 5, with OSL ages. 399 SD.6 Apparent ages for profile 7.  399

Tables

1.1 Description of the geological formations found on the Maltese Islands.  21 2.1 The cultural sequence of the Maltese Islands (with all dates calibrated). 37 2.2 Quartz OSL sediment ages from the Marsalforn (2917–2919) and Ramla (2921–2923) valleys, the Skorba temple/buried soil (2925–2927) and Tal-Istabal, Qormi, soil (2930).  40 2.3 Dating results for positions in the sediment cores. 45 2.4 Summary stratigraphic descriptions of the sequences in the deep core profiles. 57 2.5 Mean sediment accumulation rates per area versus time for the deep cores. 64 2.6 Radiocarbon measurements and ΔR values from early twentieth century marine shells from Malta. 65 2.7 Calibrated AMS 14C dates of charred plant remains from Santa Verna palaeosol, Gozo.  68 2.8 Physical properties of the catchments.  68 2.9 Normalized Diffuse Vegetation Index (NDVI) for the catchments in 2014–15 and average rainfall data for the weather station at Balzan for the period 1985 to 2012.  69 3.1 Semi-natural plant communities in the Maltese Islands.  76 xvi 3.2 Attribution of pollen taxa to plant communities in the Maltese Islands and more widely in the Central Mediterranean.  77 3.3 Characteristics of the taphonomic samples from on-shore and off-shore Mistra Valley, Malta.  80 3.4 The pollen zonation of the Salina Deep Core with modelled age-depths.  84 3.5 The pollen zonation of the Salina 4 core with modelled age-depths.  90 3.6 The pollen zonation of the Wied Żembaq 1 core with modelled age-depths.  94 3.7 The pollen zonation of the Xemxija 1 core with modelled age-depths.  98 3.8 The pollen zonation of the fill of a Bronze Age silo at In-Nuffara, Gozo.  103 3.9 Summary of the pollen analyses of the buried soil below the Santa Verna temple structure.  103 3.10 Summary of the pollen analyses from the buried soil in Ġgantija Test Pit 1.  105 3.11 Activity on Temple sites and high cereal pollen in adjacent cores.  105 4.1 List of freshwater molluscs and land snails found in the cores, habitat requirement, palaeontological record and current status and conservation in the Maltese Islands. 118 4.2 Molluscan zones for the Marsaxlokk 1 core (MX1).  121 4.3 Molluscan zones for the Wied Żembaq 1 core (WŻ1).  123 4.4 Molluscan zones for the Wied Żembaq 2 core (WŻ2).  125 4.5 Integration of molluscan zones from the Wied Żembaq 1 and 2 cores.  128 4.6 Molluscan zones for the Mġarr ix-Xini 1 core (MĠX1).  130 4.7 Molluscan zones for the Marsa 2 core (MC2).  135 4.8 The non-marine molluscan zones for the Salina Deep Core (SDC).  140 4.9 Molluscan zones for the Salina Deep Core (SDC).  142 4.10 Molluscan zones for the Xemxija 1 core (XEM1).  146 4.11 Molluscan zones for the Xemxija 2 core (XEM2).  148 4.12 Correlation and integration of molluscan data from Xemxija 1 (XEM1) and Xemxija 2 (XEM2).  151 5.1 Micromorphology and small bulk sample sites and numbers.  162 5.2 Summary of available dating for the sites investigated in Gozo and Malta.  163 5.3 pH, magnetic susceptibility, loss-on-ignition, calcium carbonate and % sand/silt/clay particle size analysis results for the Ġgantija, Santa Verna and the Xagħra town profiles, Gozo.  168 5.4 Selected multi-element results for Ġgantija, Santa Verna and Xagħra town buried soils, and the Marsalforn and Ramla valley sequences, Gozo.  169 5.5 Summary of the main soil micromorphological observations for the Santa Verna, Ġgantija and the Xagħra town profiles, Gozo.  181 5.6 pH, magnetic susceptibility and selected multi-element results for the palaeosols in section 1, Trench A, Skorba.  184 5.7 Loss-on-ignition organic/carbon/calcium carbonate frequencies and particle size analysis results for the palaeosols in section 1, Trench A, Skorba.  184 5.8 Summary of the main soil micromorphological observations of the buried soils in sections 1 and 2, Trench A, Skorba.  188 5.9 Summary of the main soil micromorphological observations of the possible buried soils at Taċ-Ċawla.  189 5.10 Field descriptions and micromorphological observations for the quarry and construction site profiles in Xagħra town. 190 5.11 Sample contexts and micromorphological observations for two silo fills at In-Nuffara.  192 5.12 Summary of the main soil micromorphological observations from the Ramla and Marsalforn valley fill profiles.  196 5.13 Main characteristics of the Upper and Lower Coralline Limestone, Globigerina Limestone, Blue Clay and Greensand. 197 5.14 Summary micromorphological descriptions and suggested interpretations for the Xemxija 1 core.  200 5.15 Summary micromorphological descriptions and suggested interpretations for the Wied Żembaq 1 core.  207 5.16 Summary micromorphological descriptions and suggested interpretations for the Marsaxlokk 1 core.  209 5.17 Summary micromorphological descriptions and suggested interpretations for the base zone of the base of the Salina Deep Core.  211 8.1 Carrying capacity estimates for the Neolithic/Temple Period of the Maltese Archipelago.  258 8.2 Summary of population changes in the Maltese Archipelago.  261 11.1 Summary of the environmental and vegetation changes in the Maltese Islands over the longue durée. 306 xvii 11.2 Summary of events revealed by the molluscan data in the deep cores. 309 11.3 Major phases of soil, vegetation and landscape development and change during the Holocene.  312 11.4 Occurrence of gypsum in FRAGSUS cores and contemporary events.  314 A2.1 Sample descriptions, contexts and archaeological significance of the profiling samples used for initial screening and laboratory characterization.  358 A2.2 Sample descriptions, contexts and archaeological significance of sediment samples SUTL2914–2930.  360 A2.3 Activity and equivalent concentrations of K, U and Th determined by HRGS.  368 A2.4 Infinite matrix dose rates determined by HRGS and TSBC.  368 A2.5 Effective beta and gamma dose rates following water correction.  369 A2.6 SAR quality parameters.  369 A2.7 Comments on equivalent dose distributions of SUTL2914 to SUTL2930.  372 A2.8 Quartz OSL sediment ages.  372 A2.9 Locations, dates and archaeological significance of sediment samples SUTL2914–2930.  373 SA.1 Field profiling data, as obtained using portable OSL equipment, for the sediment stratigraphies examined on Gozo and Malta.  379 SA.2 OSL screening measurements on paired aliquots of 90–250 μm 40% HF-etched ‘quartz’. 380 SA.3 OSL screening measurements on three aliquots of 90–250 μm 40% HF-etched ‘quartz’ for SUTL2924. 382 SA.4 IRSL screening measurements on paired aliquots of 90–250 μm 15% HF-etched ‘polymineral’.  382 SA.5 IRSL screening measurements on three aliquots of 90–250 μm 15% HF-etched ‘polymineral’ for SUTL2924.  383 A3.1 Stratigraphy and interpretation of the Salina Deep Core.  401 A3.2 Stratigraphy and interpretation of the Salina 4 core.  405 A3.3 Stratigraphy and interpretation of the Salina 2 core.  407 A3.4 Stratigraphy and interpretation of the Xemxija 1 core.  408 A3.5 Stratigraphy and interpretation of the Xemxija 2 core.  411 A3.6 Stratigraphy and interpretation of the Wied Żembaq 1 core.  413 A3.7 Stratigraphy and interpretation of the Wied Żembaq 2 core.  413 A3.8 Stratigraphy and interpretation of the Mgarr ix-Xini core.  414 A3.9 Stratigraphy and interpretation of the Marsaxlokk core.  416 A3.10 Stratigraphy and interpretation of the Marsa 2 core.  417 A3.11 Stratigraphy and interpretation of the Mellieħa Bay core.  418 A3.12 Key to the scheme for the description of Quaternary sediments. 419 A4.1 Marsa 2.  421 (online edition only) A4.2 Mgarr ix-Xini.  424 (online edition only) A4.3 Salina Deep Core.  427 (online edition only) A4.4 Wied Żembaq 2.  429 (online edition only) A4.5 Wied Żembaq 1.  430 (online edition only) A4.6 Xemxija 1.  432 (online edition only) A4.7 Xemxija 2.  435 (online edition only) A4.8 Marsaxlokk 1.  438 (online edition only) A5.1 Marsa 2.  442 (online edition only) A5.2 Mgarr ix-Xini.  456 (online edition only) A5.3 Salina Deep Core non-marine.  466 (online edition only) A5.4 Salina Deep Core marine.  478 (online edition only) A5.5 Wied Żembaq 2.  490 (online edition only) A5.6 Wied Żembaq 1.  496 (online edition only) A5.7 Xemxija 1.  502 (online edition only) A5.8 Xemxija 2.  516 (online edition only) A5.9 Marsaxlokk 1.  528 (online edition only) A8.1 Xemxija 1 core micromorphology sample descriptions.  557 A8.2 Wied Żembaq 1 core micromorphology sample descriptions.  559 A8.3 Marsaxlokk core micromorphology sample descriptions.  560 A8.4 Salina Deep Core micromorphology sample descriptions.  561 A9.1 The charcoal data from the Skorba, Kordin, In-Nuffara and Salina Deep Core.  563 xviii Preface and dedication

Caroline Malone

The FRAGSUS Project emerged as the direct result As this and the two other volumes in this series of an invitation to undertake new archaeological report, the original Cambridge Gozo Project was the fieldwork in Malta in 1985. Anthony Bonanno of the germ of a rich and fruitful academic collaboration University of Malta organized a conference on ‘The that has had international impact, and has influenced Mother Goddess of the Mediterranean’ in which successive generations of young archaeologists in Colin Renfrew was a participant. The discussions that Malta and beyond. resulted prompted an invitation that made its way to As the Principal Investigator of the FRAGSUS David Trump (Tutor in Continuing Education, Cam- Project, on behalf of the very extensive FRAGSUS team bridge University), Caroline Malone (then Curator of I want to dedicate this the first volume of the series the Avebury Keiller Museum) and Simon Stoddart to the enlightened scholars who set up this now 35 (then a post-graduate researcher in Cambridge). We year-long collaboration of prehistoric inquiry with our eagerly took up the invitation to devise a new collab- heartfelt thanks for their role in our studies. orative, scientifically based programme of research on prehistoric Malta. We dedicate this volume to: What resulted was the original Cambridge Gozo Project (1987–94) and the excavations of the Xagħra Joseph Attard Tabone Brochtorff Circle and the Għajnsielem Road Neo- Professor Anthony Bonanno lithic house. Both those sites had been found by local Professor Lord Colin Renfrew antiquarian, Joseph Attard-Tabone, a long-established figure in the island for his work on conservation and and offer our profound thanks for their continuing site identification. role in promoting the prehistory of Malta.

xix

Acknowledgements

This volume records research undertaken with fund- They were helped by Vincent Van Walt, who provided ing from the European Research Council under the technical assistance. Al Ruffell and John Meneely did European Union’s Seventh Framework Programme geophysical evaluation and GRP location of the cores. (FP/2007-2013)/ERC Grant Agreement n. 323727 During fieldwork, Tim Kinnaird and Charles French (FRAGSUS Project: Fragility and sustainability in were assisted by Sean Taylor, Jeremy Bennett and small island environments: adaptation, cultural change Simon Stoddart. We are grateful to the Superintend- and collapse in prehistory – http://www.qub.ac.uk/ ence of Cultural Heritage, Malta and Heritage Malta sites/FRAGSUS/). All the authors of this volume are for permission to undertake the analyses and much indebted to the ERC for its financial support, and to practical assistance. the Principal Investigator of the FRAGSUS Project, For Chapter 5, we would like to thank all at Her- Prof. Caroline Malone (Queen’s University, Belfast, itage Malta, the Ġgantija visitor’s centre and the Uni- UK), for her central role in devising the project and versity of Malta for their friendly and useful assistance seeing this research through to publication. throughout. In particular, we would like to thank For Chapter 2, we extend warm thanks to the George Azzopardi, Daphne Caruana, Josef Caruana, staff of the 14CHRONO centre at QUB, especially Nathaniel Cutajar, Chris Gemmell, Reuben Grima, Stephen Hoper, Jim McDonald, Michelle Thompson Joanne Mallia, Christian Mifsud, Anthony Pace, Ella and Ron Reimer, all of whom took a keen interest in Samut-Tagliaferro, Mevrick Spiteri, Katya Stroud, the FRAGSUS Project. The success of the FRAGSUS Sharon Sultana and Nick Vella. We also thank Tonko Project in general and the radiocarbon dating exercise Rajkovača of the McBurney Laboratory, Department has depended on their work. We thank the Physical of Archaeology, University of Cambridge, for mak- Geography Laboratory staff at the School of Geogra- ing the thin section slides, the Physical Geography phy, University College Dublin, for the use of their Laboratory, Department of Geography, University of ITRAX XRF core scanner. In particular, we would like Cambridge, and the ALS Global laboratory in Seville, to thank Dr Steve McCarron, Department of Geogra- Spain, for processing the multi-element analyses. phy, National University of Ireland, Maynooth and Dr For Chapter 6, Reuben Grima wrote the first Jonathan Turner, Department of Geography, National draft of this contribution, receiving comments and University of Ireland, University College, Dublin. additions from the other authors. We thank Prof. Patrick Schembri for sourcing and For Chapter 7, Simon Stoddart wrote the first collecting the Acanthocardia samples from the Natural draft of this contribution, receiving comments and Museum of Natural History. Sean Pyne O’Donnell additions from the other authors. thanks Dr Chris Hayward at the Tephrochronology For Chapter 9, we thank Sharlo Camilleri for pro- Analytical Unit (TAU), University of Edinburgh, for viding us with a copy of the GIS data produced by the help and advice during microprobe work. Dr Maxine MALSIS (MALtese Soil Information System) project. Anastasi, Department of Classics and Archaeology, We are grateful to Prof. Saviour Formosa and Prof. University of Malta, helped identify the pottery from Timmy Gambin, both of the University of Malta, who the settlement cores. Dr Frank Carroll helped show facilitated the donation of LiDAR data, together with us the way forward; but sadly is no longer with us. computer facilities, as part of the European project Chris Hunt, Rory Flood, Michell Farrell, Sean Pyne ERDF156 Developing National Environmental Monitor- O’Donnell and Mevrick Spiteri were the coring team. ing Infrastructure and Capacity, from the former Malta xxi Acknowledgements

Environment and Planning Authority. A number of the grant. The research team also wants to record our individuals were happy to share their recollections of indebtedness to the administrators of the grant within shepherding practices in Malta and Gozo over the last our own institutions, since this work required detailed sixty or seventy years; others facilitated the encoun- and dedicated attention. In particular we thank Rory ters. We are grateful to all of them: Charles Gauci, Jordan in the Research Support Office, Stephen Hoper Grezzju Meilaq, Joseph Micallef, Louis Muscat, Ċettina and Jim McDonald – CHRONO lab, and Martin and Anġlu Vella, Ernest Vella and Renata Zerafa. Stroud (Queen’s University Belfast), Laura Cousens Simon Stoddart would like to thank Prof. Martin (Cambridge University), Glen Farrugia and Cora Jones and Rachel Ballantyne for their advice in con- Magri (University of Malta), the Curatorial, Finance structing Figure 11.4. The editors would like to thank and Designs & Exhibitions Departments in Heritage Emma Hannah for compiling the index. Malta and Stephen Borg at the Superintendence of Firstly, the FRAGSUS Project is the result of Cultural Heritage. Finally, we thank Fr. Joe Inguanez a very generous research grant from the European (Emeritus Head of Department, Department of Soci- Research Council (Advanced Grant no’ 323727), with- ology, University of Malta) for offering us theleitmotif out which this and its two partner volumes and the of this volume while a visiting scholar in Magdalene research undertaken could not have taken place. We College, Cambridge: ‘Mingħajr art u ħamrija, m’hemmx heartily thank the ERC for its award and the many sinjorija’ translating as ‘without land and soil, there is administrators in Brussels who monitored our use of no wealth’.

xxii Foreword

Anthony Pace

Sustainability, as applied in archaeological research designed by the participating institutional partners and heritage management, provides a useful perspec- and scholars, was to explore untapped field resources tive for understanding the past as well as the modern and archived archaeological material from a number conditions of archaeological sites themselves. As often of sites and their landscape to answer questions that happens in archaeological thought, the idea of sus- could be approached with new techniques and meth- tainability was borrowed from other areas of concern, ods. The results of the FRAGSUS Project will serve to particularly from the modern construct of develop- advance our knowledge of certain areas of Maltese ment and its bearing on the environment and resource prehistory and to better contextualize the archipela- exploitation. The term sustainability entered common go’s importance as a model for understanding island usage as a result of the unstoppable surge in resource archaeology in the central Mediterranean. The work exploitation, economic development, demographic that has been invested in FRAGSUS lays the founda- growth and the human impacts on the environment tion for future research. that has gripped the World since 1500. Irrespective of Malta and Gozo are among the Mediterranean scale and technology, most human activity of an eco- islands whose prehistoric archaeology has been nomic nature has not spared resources from impacts, intensely studied over a number of decades. This transformations or loss irrespective of historical and factor is important, yet more needs to be done in the geographic contexts. Theories of sustainability may field of Maltese archaeology and its valorization. provide new narratives on the archaeology of Malta Research is not the preserve of academic specialists. and Gozo, but they are equally important and of It serves to enhance not only what we know about central relevance to contemporary issues of cultural the Maltese islands, but more importantly, why the heritage conservation and care. Though the archae- archipelago’s cultural landscape and its contents ological resources of the Maltese islands can throw deserve care and protection especially at a time of light on the past, one has to recognize that such extensive construction development. Strict rules and resources are limited, finite and non-renewable. The guidelines established by the Superintendence of sense of urgency with which these resources have to Cultural Heritage have meant that during the last two be identified, listed, studied, archived and valued is decades more archaeological sites and deposits have akin to that same urgency with which objects of value been protected in situ or rescue-excavated through a and all fragile forms of natural and cultural resources statutory watching regime. This supervision has been require constant stewardship and protection. The idea applied successfully in a wide range of sites located in of sustainability therefore, follows a common thread urban areas, rural locations and the landscape, as well across millennia. as at the World Heritage Sites of Valletta, Ġgantija, It is all the more reason why cultural resource Ħaġar Qim and Mnajdra and Tarxien. This activity management requires particular attention through has been instrumental in understanding ancient and research, valorization and protection. The FRAGSUS historical land use, and the making of the Maltese Project (Fragility and sustainability in small island historic centres and landscape. environments: adaptation, cultural change and col- Though the cumulative effect of archaeological lapse in prehistory) was intended to further explore research is being felt more strongly, new areas of and enhance existing knowledge on the prehistory interest still need to be addressed. Most pressing are of Malta and Gozo. The objective of the project as those areas of landscape studies which often become xxiii Foreword

peripheral to the attention that is garnered by prom- FRAGSUS Project, will bear valuable results that will inent megalithic monuments. FRAGSUS has once only advance Malta’s interests especially in today’s again confirmed that there is a great deal of value world of instant e-knowledge that was not available in studying field systems, terraces and geological on such a global scale a mere two decades ago. settings which, after all, were the material media in FRAGSUS also underlines the relevance of which modern Malta and Gozo ultimately developed. studying the achievements and predicaments of past There is, therefore, an interplay in the use of the term societies to understand certain, though not all, aspects sustainability, an interplay between what we can learn of present environmental challenges. The twentieth from the way ancient communities tested and used the century saw unprecedented environmental changes very same island landscape which we occupy today, as a result of modern political-economic constructs. and the manner in which this landscape is treated in Admittedly, twentieth century developments cannot contested economic realities. If we are to seek factors be equated with those of antiquity in terms of demog- of sustainability in the past, we must first protect its raphy, technology, food production and consumption relics and study them using the best available meth- or the use of natural resources including the uptake ods in our times. On the other hand, the study of the of land. However, there are certain aspects, such as past using the materiality of ancient peoples requires climate change, changing sea levels, significant envi- strong research agendas and thoughtful stewardship. ronmental degradation, soil erosion, the exploitation The FRAGSUS Project has shown us how even small and abandonment of land resources, the building and fragile deposits, nursed through protective legislation maintenance of field terraces, the rate and scale of and guardianship, can yield significant information human demographic growth, movement of peoples, which the methods of pioneering scholars of Maltese access to scarce resources, which to a certain extent archaeology would not have enabled access to. As reflect impacts that seem to recur in time, irrespec- already outlined by the Superintendence of Cultural tively of scale and historic context. Heritage, a national research agenda for cultural herit- age and the humanities is a desideratum. Such a frame- Anthony Pace work, reflected in the institutional partnership of the Superintendent of Cultural Heritage (2003–18).

xxiv Chapter 3

The Holocene vegetation history of the Maltese Islands

Michelle Farrell, Chris O. Hunt & Lisa Coyle McClung

3.1. Introduction with a dry Early Holocene becoming more humid Chris O. Hunt after c. 4000 cal. bc, while the northwest and southeast show an opposite trend with a wetter Early Holocene The history of climate and environmental change in the and progressive desiccation after c. 4000 cal. bc (Hunt lands around the Mediterranean Sea is dramatic, and et al. 2007). Within this very broad pattern there are our still emerging understanding has changed radically considerable regional differences (e.g. Finné et al. over the last 60 years and is still changing. Pioneering 2011) and in the central Mediterranean, changes in pollen work by Bonatti (1966) first provided evidence seasonality are superimposed on these trends (Peyron that relative to the Holocene (the last 11,500 years) the et al. 2011, 2017). Late Pleistocene was a time of drought and cold in the Furthermore, the impact of a series of short, region. But for many years alternative viewpoints held intense aridification phases (some of which seem to currency, especially the work of Vita-Finzi (1969) who relate to events in the North Atlantic (Vellinga & Wood held, on the basis of widespread Late Pleistocene grav- 2002; Wiersma & Renssen 2006) seem to have had dra- els, that this had been a period of high precipitation. matically different degrees of severity in different parts This view was finally laid to rest only in the 1980s and of the basin (e.g. Asioli et al. 1999; Peyron et al. 2011; 1990s by further pollen work on lacustrine deposits Jaouadi et al. 2016). In some Mediterranean countries, (e.g. Bertoldi 1980; Bottema & Woldring 1984; Alessio these extreme events had significant landscape impact et al. 1986; Follieri et al. 1988; Bottemaet al. 1990), and (e.g. di Rita & Magri 2009; Zielhofer & Faust 2008; Ziel- analysis of the sedimentology and biotic components hofer et al. 2010). Some of these events, at c. 6200, 4500, of Late Pleistocene gravels (e.g. Barker & Hunt 1995). and 2300 bc, disrupted the environmental foundations Vita-Finzi (1969) did, however, pioneer the rec- of human societies, leading to changes in agricultural ognition of the scale and impact of climatic variability systems, social organization and migration. The most within the Holocene in countries bordering the Med- dramatic of these was the event at c. 6200 bc which iterranean at a time when most researchers thought seems to have triggered dispersal across the Mediter- of the period as extremely stable climatically. Recog- ranean of people carrying Neolithic technology from nition of this climatic variability and its impacts was Anatolia (e.g. Zilhāo 2001; Weninger et al. 2006; Berger made more complex because of very strong patterns & Guilaine 2009; Zeder 2008), although later events of human impacts in some Mediterranean countries, also had severe consequences for people in marginal which were difficult to disentangle unequivocally from situations and coincide with major technological and the climatic signal (e.g. Hunt 1998; Grove & Rackham societal change (e.g. Carroll et al. 2012; Soto-Berelov 2003). Only with the rise of isotope-based palaeocli- et al. 2015). mate studies and high-resolution dating did it become The scale and nature of human interference possible to separate the climatic and anthropogenic with – and shaping of – landscape and the interrela- signals (e.g. Sadori et al. 2008). tionships between human activities and climate have More recent work has started to show that within also emerged. Patterns of vegetation modification the Mediterranean Basin the overall trend and timing associated with the spread of farming and later waves of Holocene climate change differs from region to of extensification and intensification have been widely region (Peyron et al. 2011). In broad terms, the north- reported in the northern Mediterranean countries, east and southwest of the basin seem to be in phase, where agriculturalists cleared landscapes of trees (e.g. 73 Chapter 3

Drescher-Schneider et al. 2007; Tinner et al. 2009), but the early Holocene, giving rise to deep inlets. During are more subtle in the southern and eastern countries the mid- to late Holocene these inlets were infilled which were less heavily vegetated in the early Holocene by sediments eroded from the Maltese landscape, and where pastoral activity and vegetation manage- producing coastal plains. ment were more significant (Roberts 2002; Asouti et Schembri and Lanfranco (1993) inferred an early al. 2015; Jaouadi et al. 2016). Episodes during which Holocene landscape covered by oak-pine sclerophyll agricultural activity expanded and/or intensified are woodland on biogeographical grounds. The early often marked by major valley alluviation phases in the Holocene vegetation prior to human colonization was northern Mediterranean countries (Hunt et al. 1992; otherwise completely unknown, although evidence Barker & Hunt 1995; Hunt & Gilbertson 1995; Hunt for burning and clearance of pine-juniper woodland 1998). The interaction between people and environment appears to be present before 3262–2889 cal. bc at emerges particularly from landscape-scale integrated Marsa (Carroll et al. 2012). Post-colonization prehis- archaeological-palaeoenvironmental surveys and by toric environments seem to be characterized by open, combining pollen patterns with settlement data (e.g. often steppic landscapes, some lentisk scrub and Malone & Stoddart 1994; Barker 1996; Barker et al. considerable arable activity and grazing (Evans 1971; 1996; Stoddart et al. 2019). Hunt 2001, 2015; Schembri et al. 2009; Carroll et al. In the Maltese Islands, Pleistocene temperate 2012; Marriner et al. 2012; Djamali et al. 2013; Gambin stages were characterized by fairly dense shrubby or et al. 2016). There is a suggestion of an interruption to small arboreal vegetation, slope stability and marked cereal cultivation around 2300 cal. bc, probably linked palaeosol formation, with calcareous tufa deposits with general regional aridity at this time (Carroll et forming around springs. Blown sands accumulated al. 2012). The coastal deposits seem to be marked by rapidly at the end of temperate stages as sea level fell, discontinuous sedimentation and there are issues with exposing the sandy sea bed to wind erosion. Pleistocene dating as the result of recycling and most probably cold-stage environments were mostly rather arid, with intrusion of material during the coring process at some minimal, very open vegetation. Consequently, high sites (Carroll et al. 2012; Djamali et al. 2013; Gambin sediment mobility led to deposition of wind-blown et al. 2016). Strong taphonomic imprints seem to be loessic silts, thick – often muddy – colluvial slope present in pollen assemblages from archaeological sites deposits resulting from wash and mud-flow, and (Hunt 2001, 2015), further complicating the process of gravelly alluvium of ephemeral rivers (Hunt 1997). reconstructing Holocene vegetation. Charcoal analysis The presence of interglacial palaeosols and the absence at Skorba provides evidence for trees including carob, of interglacial beach sediments in Pleistocene deposits hawthorn and ash in the early Neolithic and olive, around the coasts of the Maltese archipelago inclined perhaps cultivated, in the later Neolithic (Trump 1966). Hunt (1997) to the view that sea level during the cur- rent (Holocene) interglacial was relatively higher than 3.2. Palynological methods during the Last Interglacial, when sea level globally Lisa Coyle-McClung, Michelle Farrell & Chris O. Hunt was c. 6 m above present day sea levels. Carroll et al. (2012) also noted that some Roman portside locations Cores were sub-sampled for palynology with slices at Marsa were very close to the modern sea level. Both 1 cm thick taken for analysis and bagged in self-seal these observations negate the suggestions of Furlani polythene bags. These were stored at 4° C in cold et al. (2013) of long-term tectonic stability and instead stores at Queen’s University Belfast and Liverpool suggest that generally the Maltese Islands are sinking John Moores University until analysed. Sub-samples very slowly relative to modern sea levels, most prob- of 2.5–5 cm3 volume were taken from each sample bag. ably as the result of tectonic factors. Samples were prepared for pollen analysis fol- At the beginning of the Holocene, it is likely that lowing standard methods (Moore et al. 1991), including the Maltese landscape was rather different from today. treatment with 10 per cent hydrochloric acid to remove Sea levels were much lower – probably about 80 m carbonates, hot 10 per cent potassium hydroxide to dis- below present-day sea level (Lambeck et al. 2011), so the aggregate the sediment before sieving through 120 μm major islands (Malta, Gozo, and Cominotto) mesh to remove larger mineral and organic material, would have been part of one larger landmass. The and 5 per cent sodium pyrophosphate and sieving legacy of Pleistocene slope and aeolian sedimentation through 6 μm mesh to remove clay-sized particles. was most probably a mantle of silty loessic sediment, Heavy liquid separation using sodium polytungstate at which has now been mostly stripped away by erosion. a specific gravity of 1.95 (Zabenskie & Gajewski 2007) Deep valleys scoured out during the Last Glaciation was employed to separate the remaining mineral and would have been flooded by the rising sea level in organic fractions. The organic residues were stained 74 The Holocene vegetation history of the Maltese Islands

using aqueous safranin and mounted on microscope dinoflagellate cysts and a few microfossils of uncertain slides in silicone oil. affinity. Identification follows the catalogues of Bas Slides were counted at a magnification of ×400, van Geel (i.e. van Geel 1978; van Geel et al. 1981, 1989, with ×1000 magnification and oil immersion used 2003) and McCarthy et al. (2011). for critical identifications. Pollen and spores were A number of semi-natural habitat-specific plant identified using the keys of Mooreet al. (1991), Reille groupings may be recognized in the Maltese Islands (1992) and Beug (2004) and the reference collections at (Table 3.1). Our palynological analyses suggest a more Queen’s University Belfast and Liverpool John Moores diverse flora existed in the past, but it is likely that University. Wherever possible, counts of at least 300 similar associations existed, although not present in pollen grains per sample were made (Benton & Harper quite the same locations, or covering the same spatial 2009, 608). Non-pollen palynomorphs (NPPs) were extent as at present. also identified and recorded during pollen counting. Interpretation of pollen diagrams requires attri- The sum used to calculate percentages consisted bution of taxa to ecological groupings. Scrutiny of the of all terrestrial pollen and spores (TLPS). Percentages literature and field experience amongst theFRAGSUS of taxa not included within the main sum (i.e. aquatics environmental team suggests that taxa encountered and NPPs) were calculated using the main sum plus during pollen counting can be attributed to one or the sum for the taxon group. more of the semi-natural plant communities shown in Pollen diagrams for the FRAGSUS sequences were Table 3.1. The attribution of palynologically recognized plotted using Tilia (Grimm 1987) and zoned according taxa to plant communities appropriate for the Maltese to significant changes in the proportions of major taxa. Islands is shown in Table 3.2. Composite diagrams were compiled for selected taxa from the FRAGSUS pollen counts, those of Carroll 3.4. Taphonomy et al. (2012) and Djamali et al. (2013), and extracted Chris O. Hunt & Michelle Farrell from Gambin et al. (2016) by measuring their pollen diagrams. The data were placed into 20-year bins and It has long been recognized that the relationship calculated and graphed in Tilia (Grimm 1987). All between pollen-producing vegetation and pollen pollen diagrams were put into Adobe Illustrator for arriving at depositional settings is far from straight- final preparation. forward, and that further changes to assemblages occur after deposition. Taphonomy (Efremov 1940; 3.3. Taxonomy and ecological classification Behrensmeyer & Kidwell 1985) is the study of this Chris O. Hunt relationship between life and death assemblages – in the case of pollen and spores it includes the processes The Maltese vegetation is in many ways comparable of production by the parent plants, dispersal into the in physiognomy and taxonomy with that of heavily environment, transportation by animal vectors, wind human-impacted landscapes elsewhere in the low- and water, deposition and burial by sedimentary lands of the Mediterranean Basin, although there are processes, post-depositional alteration by decay and differences because of the high rate of endemism in diagenetic processes, extraction from sediments and the region and in the Maltese Islands in particular investigation by palynologists (for a recent review (Schembri 1993). Taxonomic attribution in this study see Hunt & Fiacconi 2018). Most pollen taphonomic follows Reille (1992, 1995, 1998), Moore et al. (1991), work has dealt with peat bogs and lakes – fairly simple Beug (2004) and type collections at Queen’s Univer- environments with well-defined taphonomic pathways sity Belfast and Liverpool John Moores University. and generally good pollen preservation, but in the The taxonomic level reached is not uniform: in some case of the FRAGSUS Project, sediments laid down cases specific identification is possible, but with some in these environments were not available for study. taxonomic groupings, in particular the Poaceae and Therefore, sediments deposited in coastal shallow Asteraceae, sub-family or family is the best that could marine and marginal marine settings were sampled be achieved. This is unfortunate, as members of these in the FRAGSUS boreholes, as previous work (Carroll large taxonomic groupings have more detailed envi- et al. 2012) shows that pollen and other palynomorphs ronmental requirements than can be allocated at are preserved in these waterlogged deposits. We also higher taxonomic levels, but quite normal in current sampled soils and redeposited soils associated with palynological research. In addition to the pollen and archaeological sites where these seemed potentially spores, a small number of NPPs were recorded in viable for palynology. Such sites are widely acknowl- the pollen diagrams. These include fungal spores edged to be highly challenging to the palynologist and micorrhyzae, some types of green algal spores, since preservation is frequently poor and subject to 75 Chapter 3

Table 3.1. Semi-natural plant communities in the Maltese Islands (partly after Schembri 1993). Association Characteristic taxa Comments Palynology Sclerophyll Pine (Pinus halepensis), holm oaks Very small stands of probably Woodland trees, most notably pines, are woodland (Quercus ilex, Quercus spp.), with secondary woodland, strongly mainly wind-pollinated. They produce oleaster (Olea europaea) and carob impacted by human activity and substantial quantities of pollen and their (Ceratonia siliqua), and understorey of mostly situated on steep rocky height enables them to disperse their lentisk (Pistacia lentiscus), buckthorn ground. Today there are also more- pollen widely. Pine pollen disperses (Rhamnus lycioides), hawthorn or-less recent plantations, often over immense distances and is highly (Crataegus spp.) and a ground flora characterized by non-native pines resistant to decay including polypody fern (Polypodium spp.), and ivy (Hedera spp.) Maquis Oleaster (Olea europaea), lentisk Scrub, often characterized by Many maquis species are insect- (Pistacia lentiscus), carob (Ceratonia thorny, glaucous or aromatic pollinated. They produce little pollen siliqua), buckthorn (Rhamnus lycioides), shrubs. There are fairly extensive which typically is not widely dispersed yellow germander (Teucrium flavum), areas of maquis in rocky areas on white hedgenettle (Prasium majus) the sides of widien, at the foot of inland cliffs and on abandoned farmland. Carob seems to have been a late prehistoric introduction. Today Maltese maquis is often dominated by introduced eucalypts Garrigue Conehead thyme (Thymus capitatus), Low scrub, often characterized by Many garrigue plants are insect- Maltese yellow kidney-vetch (Anthyllis cushion-shaped low bushes and pollinated and produce little pollen hermanniae), tree germander (Teucrium clumps of plants. It is extensive which is usually not widely dispersed fruticans), erica (Erica multiflora), in the Maltese Islands, some Maltese spurge (Euphorbia melitensis), natural, but some resulting from squill (Scilla spp.) and other geophytes degradation of woodland or maquis Steppe Esparto (Lygeum spartum), other Steppe habitats are often dominated Some steppe plants, such as grasses, are grasses (e.g. Hyparrhenia pubescens, by grasses with an admixture of wind-pollinated and produce abundant Andropogon distachyos, Brachypodium other herbaceous plants and bulbs. pollen. Grass pollen is difficult to identify retusum, Phalaris truncata, Stipa In degraded steppes, the proportion beyond the family level. Pollen of thistles capensis, Aegilops geniculata), thistles of spiky plants such as thistles, and the lettuce group, although insect- (Carlina involucrata, Notobasis syriaca, unpalatable plants like wormwood, pollinated, are particularly durable so Galactites tomentosa), plantains bulbs including asphodels, often present in larger proportions in (Plantago spp.), sandworts and rosette plants such as plantains, pollen assemblages than their parent allies (Caryophyllaceae), asphodels ephemerals such as sandworts and plants are in the vegetation. Again, pollen (Asphodelus aestivuus), and other lilies the lettuce group (Lactuceae) tends of these taxa cannot often be identified (e.g. Urginea maritima) to rise relative to the grasses beyond family or sub-family level Coastal Grasses (Elymus farctus, Sporobolus Coastal habitats include saltmarshes, Coastal grasses and chenopods are not arenarius, Ammophila arenaria), seakale dunes and shallow saline rocky separable palynologically from other (Salsola), glasswort (Salicornia) and soils adjacent to coasts. Dunes are members of their families, but very high other Chenopods (Chenopodiaceae), very rare in the Maltese Islands. percentages of chenopod pollen are sea-lavender (Limonium-type), They are typically characterized by often associated with saltmarshes. Sea thrift (Armeria-type), rock samphire specialized grasses. Saltmarshes are lavender and thrift are indistinguishable (Crithmum) marked by sea-lavender and some palynologically. Apart from the grasses, chenopods. The saline coastal soils most coastal taxa are insect pollinated are characterized by rock samphire and are low pollen producers. and sea-lavender Rupestral Caper bush (Capparis spinosa), Maltese Rock faces and boulder screes Most rupestral taxa are insect-pollinated rock centaury (Palaeocyanus crassifolius, have a highly specialized flora and are low pollen producers. Their the National Plant of Malta) and and provide important refuges for pollen is often not widely dispersed Maltese cliff orache (Atriplex lanfrancoi) endemic plants Freshwater Reeds (Phragmites australis), mints Wetlands are very rare in the Reeds are not distinguishable (Mentha), buttercups (Ranunculus), Maltese Islands, and perennial palynologically from other members knotgrass (Persicaria), poplar (Populus wetlands are vanishingly rare. of the grass family. Most of the other alba), willows (Salix pedicellata, Salix There are few semi-permanently herbaceous water plants are insect- alba), elm (Ulmus canescens), bay wet valley-floors, characterized pollinated. The waterside woodland (Laurus nobilis) by dense stands of reeds with trees are wind-pollinated and fairly other wetland taxa such as mints, prolific pollen producers, except for the buttercups and knotgrass. There are insect-pollinated bay. Poplar pollen is also a few patches of relict riverine not durable and rarely preserves, while woodland characterized by poplars, bay pollen is completely non-durable willows, elms and bay

76 The Holocene vegetation history of the Maltese Islands

Table 3.1 (cont.). Association Characteristic taxa Comments Palynology Disturbed Ironwort (Sideritis), nettles Urtica( ), Disturbed habitats result from These species are insect-pollinated and figworts (Scrophularia), borage (Borago human and geomorphic activity (for generally low pollen producers officinalis), mustards (Brassicaceae), instance landslides). The disturbed honeyworts (Cerinthe), bindweeds ground species are mostly annuals (Convulvus arvensis-type), spurges which mature quickly and disperse (Euphorbia), bedstraws (Galium), seeds prolifically purslane (Portulaca oleracea), corn salad (Valerianella) Cultivated Corn cockle (Agrostemma githago), Agricultural weeds such as corn Most, apart from the cereals, are insect- cow wheat (Melampyrum), love-in-a- cockle, cow wheat, love-in-a-mist pollinated. All are low pollen producers, mist (Nigella), knotweed (Polygonum are aliens which arrived with early but cereal pollen is shed in quantity aviculare), carob (Ceratonia siliqua), agricultural plants. Olives only during threshing cereals, barley (Hordeum), wheat seem to have been domesticated (Triticum), eucalypts (Eucalyptus), quite late in prehistory vines (Vitis)

Table 3.2. Attribution of pollen taxa to plant communities in the Maltese Islands and more widely in the Central Mediterranean. axaT where the indicated ecology is followed by an asterisk are suggested not to be native to the Maltese Islands during the Holocene because of a combination of (1) absence of macrofossil evidence (2) extremely low or very sporadic pollen counts (3) incompatible ecology. Catholic = not ecologically determinate. Pollen taxon Ecology Pollen taxon Ecology Abies high montane forest* Pistacia maquis Betula montane forest* Rhamnus-type maquis Carpinus montane forest* Rosaceae maquis Corylus-type montane forest* Rosmarinus-type maquis Fagus montane forest* Salvia officinalis-type maquis Ostrya montane forest* Scabiosa garrigue, maquis, disturbed sites Quercus (deciduous type) montane forest Centaurium garrigue Tilia montane forest* Scilla-type garrigue Hedera woods Spergularia media-type garrigue, rocky places, Helleborus woods sand Helleborus viridis-type woods Tuberaria garrigue, rocky places Ilex aquifolium woods Linum garrigue, rocky places Osmunda regalis woods Rubiaceae garrigue, rocky places Pinus woods Adiantum capillus-veneris rocky places Polypodium woods Campanula rocky places Pteridium aquilinum woods Centranthus ruber-type rocky places Quercus woods Gladiolus rocky places Quercus ilex-type woods Linaria-type rocky places Daphne maquis Sedum-type rocky places Ephedra maquis Theligonum rocky places Erica maquis Scrophularia disturbed, rocky Ericaceae maquis Borago officinalis disturbed, rocky, weed of Fumana maquis cultivation Juniperus maquis Brassicaceae disturbed, rocky, weed of cultivation Ligustrum-type maquis Cerinthe disturbed, rocky, weed of Olea europaea maquis cultivation Origanum maquis Convolvulus arvensis-type disturbed, rocky, weed of Phillyrea maquis cultivation

77 Chapter 3

Table 3.2 (cont.). Pollen taxon Ecology Pollen taxon Ecology Euphorbia disturbed, rocky, weed of Caryophyllaceae dry catholic cultivation Chenopodiaceae dry catholic Galium disturbed, rocky, weed of Cirsium-type dry catholic cultivation Cistus dry catholic Portulaca oleracea disturbed, rocky, weed of cultivation Dipsacaceae dry catholic Valerianella disturbed, rocky, weed of Helianthemum dry catholic cultivation Lactuceae dry catholic Sideritis disturbed, weed of cultivation Lathyrus dry catholic Urtica disturbed, weed of cultivation Polygala dry catholic Agrostemma githago weed of cultivation Potentilla-type dry catholic Melampyrum weed of cultivation Senecio-type dry catholic Nigella weed of cultivation Acanthus catholic Polygonum aviculare-type weed of cultivation Apiaceae catholic Centaurea cyanus weed of cultivation Asteroideae catholic Ceratonia siliqua cultivated Cyperaceae catholic Cereal-type cultivated Fabaceae catholic Eucalyptus cultivated Gentiana catholic Vitis cultivated Lamiaceae catholic Centaurea steppe, garrigue, rocky places Lotus-type catholic Sanguisorba minor-type steppe, garrigue, rocky places Poaceae catholic Acacia steppe Pteropsida (monolete) catholic Ambrosia-type steppe Pteropsida (trilete) catholic Arenaria-type steppe Ranunculus catholic Artemisia steppe Rubus catholic Asphodelus steppe Rumex catholic Botrychium steppe Tamarix catholic Carduus-type steppe Trifolium-type catholic Carlina/Onopordum-type steppe Valeriana officinalis-type catholic Hippophae steppe Alnus waterside* Limonium steppe Equisetum waterside Lygeum spartum steppe Filipendula waterside Malva sylvestris-type steppe Mentha-type waterside Matricaria-type steppe Littorella-type waterside Ophioglossum steppe Persicaria maculosa-type waterside Plantago lanceolata-type steppe Ranunculus acris-type waterside Plantago major/media-type steppe Sphagnum waterside Rumex obtusifolius steppe Typha emergent aquatic Scabiosa columbaria-type steppe Isoetes aquatic Succisa-type steppe Potamogeton aquatic Allium-type dry catholic Sparganium-type aquatic major taphonomic alteration (Edwards et al. 2015). It Surface sediment samples were taken in the was therefore necessary to understand the taphonomic Mistra Valley on Malta, and from the shallow marine pathways associated with these depositional environ- inlet into which it drains, to explore the influence of ments in the Maltese Islands. pedogenic processes and marine sedimentation on 78 The Holocene vegetation history of the Maltese Islands

N c

MIS C3

MIS C2

a MIS B6

MIS B8 MIS B10 MIS B5 MIS B9

MIS B1 MIS B4

MIS C4 0 200 m

0 10 km

b

MIS C8

MIS C5

MIS C7

0 2 km

Figure 3.1. Valley catchments and core locations in the Mistra area of Malta: a) location of the Mistra catchment; b) locations of onshore surface sediment samples from the Mistra catchment; c) locations of the remaining onshore surface sediment samples from the Mistra catchment and offshore surface sediment samples from Mistra Bay (M. Farrell). pollen assemblages within a restricted area which is seasonal cyclic wetting and drying and are likely ana- likely to have had relatively uniform pollen rain (Fig. logues for samples from archaeological sites and buried 3.1; Table 3.3). The Mistra Valley was selected because, soils. These environments are not suitable for long- compared with many other localities in the Maltese term survival of pollen unless rapid burial occurs. In Islands, traditional agricultural land-uses are still situations of rapid burial, pollen may survive because present and the intensity of anthropogenic alteration of bacteria and fungi preferentially scavenge other types semi-natural vegetation is comparatively low. Stands of organic material which are less resistant to attack of planted pines are present in the valley, mostly on than the sporopollenin from which plants make the higher ground, and although oak woodland is not resistant exine of pollen and spores (Moore et al. 1991). present within the valley it is present on the slopes Here well represented taxa are widely acknowledged of the adjacent Pawles Valley. It therefore offers a to be resistant to microbial attack and oxidation, princi- reasonable, but not perfect analogue for pre-modern pally Pinus, Lactuceae and various Asteraceae such as land-uses and the type of depositional environments Senecio-type (Havinga 1964, 1967). Nevertheless, some sampled in this project. indications of land use and vegetation are also present, Strong taphonomic biases are most probably for instance the cereal pollen and agricultural weeds seen in all samples (Table 3.3; Fig. 3.2). Comparing like Chenopodiacae in C3, derived from an arable field, these biases is instructive for the interpretation of the maquis plants such as Olea europaea, Phillyrea, and FRAGSUS core and archaeological site assemblages. Ceratonia siliqua in samples C5 and C7 where scrub is The onshore samples were obtained from microbially present, and the extremely high Pinus associated with active, strongly oxidative environments affected by pine scrub woodland in C7. Additionally, Florenzano 79 Chapter 3

Table 3.3. Characteristics of the taphonomic samples from on-shore and off-shore Mistra alley,V Malta. Sample code Ecology of Location Major palynological Key accessory characteristics characteristics MIS C2 Limestone pavement with low Pinus 28%, Lactuceae 41%, Cirsium-type 3%, Asphodelus 2%, Brassicaceae garrigue Senecio-type 11%, Poaceae 4% 2%, Sedum-type, Linum, Carlina/Onopordum- Dominant species: Poaceae, Asphodelus type, Rumex, Rosaceae, Apiaceae, Acanthus, sp., Psoralea bituminosa, Urginea Asteroideae, Euphorbia, Mentha-type present maritima Also present: Sedum sp., Cirsium sp., Galactites tomentosa, Fumana sp., Lotus sp., Foeniculum vulgare MIS C3 Border beween small arable fields Pinus 15%, Chenopodiaceae Caryophyllaceae 2%, Senecio-type 2%, close to the sea 12%, Lactuceae 42%, Poaceae Asteroideae 2%, Sedum-type, Linum, Cirsium- Dominant species: cultivars (wheat, 4%, Brassicaceae 11%, Cereal- type, Centaurea, Rumex, Rosaceae, Apiaceae, beans, potatoes), Poaceae, Borago type 7% Tamarix present officinalis, Brassicaceae Also present: Calendula arvensis, Senecio bicolor, Adonis microcarpa, Ficus carica MIS C4 Coastal maquis Pinus 57%, Ericaceae 5%, Poaceae 2%, Ceratonia siliqua 2%, Quercus, Dominant species: Erica sp., Ceratonia Asphodelus 11%, Lactuceae Acacia, Carlina/Onopordum-type, Cirsium- siliqua, Pistacia lentiscus, Prasium majus, 14% type, Centaurea, Caryophyllaceae, Poaceae, Foeniculum vulgare, Rosa sp., Senecio-type, Rumex, Rosaceae, Asteroideae, Asteraceae Poaceae, Brassicaceae, Tamarix present Also present: Hedera helix, Asphodelus sp., Teucrium sp., Cirsium sp., Galium sp., Fumana sp., Psoralea bituminosa MIS C5 Abandoned agricultural terrace with Pinus 40%, Lactuceae 35%, Cirsium-type 2%, Brassicaceae 2%, Linum, regenerating scrub Senecio-type 10%, Poaceae 5% Carlina/Onopordum-type, Centaurea, Dominant species: Poaceae, Ceratonia Asphodelus, Chenopodiaceae, Helianthemum, siliqua, Eucalyptus sp., Foeniculum Rosaceae, Apiaceae, Asteroideae, Euphorbia, vulgare, Asparagus aphyllus, Lotus sp. Ceratonia siliqua, Cereal-type, Tamarix present Also present: Cirsium sp., Galactites tomentosa, Arundo donax MIS C7 Semi-open pine scrub/maquis Pinus 97% Olea europaea, Phillyrea, Poaceae, Eucalyptus Dominant species: Pinus halepensis, present Olea europaea Also present: Acacia, Asparagus aphyllus, Poaceae MIS C8 Limestone pavement with low Pinus 22%, Lactuceae 40%, Carlina/Onopordum-type 2%, Sanguisorba garrigue Senecio-type 12%, Poaceae 3% minor-type 2%, Chenopodiaceae 2%, Dominant species: Thymus capitatus, Euphorbia 2%, Ericaceae, Tuberaria, Scilla- Poaceae, Carlina involucrata, Urginea type, Plantago lanceolata-type, Cirsium-type, maritima Artemisia, Lygeum spartum, Asphodelus, Also present: Aster squamatus, Helianthemum, Allium-type, Rumex, Fabaceae, Atractylis gummifera, Euphorbia Eucalyptus, Tamarix present melitensis, Asphodelus aestivus, Allium sp. Asparagus aphyllus MIS B1 Shallow marine. Nearshore, shallow, Pinus 15–30%, Quercus 2–5%, Ostrya 0–9%, Ericaceae sandy bottom Chenopodiaceae 3–11%, 0–7%, Sedum-type 0–9%, Plantago lanceolata- MIS B4 Shallow marine. Deeper water, Lactuceae 3–20%, Rosaceae type 0–6%, Carlina/Onopordum-type 0–4%, eelgrass meadow 1–7%, Poaceae 5–14%, Artemisia 0–2%, Caryophyllaceae 0–2%, Brassicaceae 4–12% Helianthemum 0–3%, Senecio-type 0–11%, MIS B5 Shallow marine. Shallow, sandy Matricatria-type 0–4%, Rumex 0–6%, bottom Asteroideae 0–5%, Cyperaceae 0–3%, MIS B6 Shallow marine. Shallow, sandy Eucalyptus 0–4%, Ceratonia siliqua 0–4%, bottom on edge of eelgrass meadow Cereal-type 0–5%, Alnus 0–2%, Tamarix 0–4%, Mentha-type 0–1%, Littorella-type 0–2% MIS B8 Shallow marine. Deeper water, eelgrass meadow MIS B9 Shallow marine. Deeper water, sandy bottom MIS B10 Shallow marine. Deeper water, eelgrass meadow

80 The Holocene vegetation history of the Maltese Islands

Sea

Land

s

e

r

o

p

s

&

n

e

l

l

o

p

d

e n

p a

l

y

200 400 600 l

t

-

a

t a

l

e

l

o

p

e

T

r

y

t

o

-

t

t a i

h

L

t 5 n

e

M

s

a 5 u

e

u

n

l

p

q

i

y

A

l values represents less '+' thanNote: 1% t

i

- 5 s

s

i

a

s

Disturbed Cult. Waterside i

n n

e

o

Cereal-type

s

v t 10

r

u

a

t a r

p

e s

y

C u l

l

a 5

u

c

v

l u

o

E

. a v

i 5 t

n

b

e

r o

d

o C

n

h i 5

) p

e u

t

E e

l 5

o

n o

20 m

(

a

Brassicaceae e

a

c

d

i i

t

s r

p

U

o 5 r

e

t P 5

Catholic

.

f

f

i

d 20

n

u

Poaceae

e

a

e

d

Cyperaceae i

5

o

r

s

e

t

u

s

.

h f

t f

A 10 i

n

d a

n c

u

A

e 5

a

e

c Apiaceae

e

a 5

p

b

y

a t

-

F

s 5

u

t

o

s

L

u 5

b

x

i

. u

r

f

R

f

a

i 5

d

m

n

a

u

T

s

x 5

e u

l

m u

c

u

e

n

R 10

p u

y

n

t - a

R

m 5

u

i

l

l

e

A

p 5

y

t

-

o i

Dry Catholic c

20 e

n

e

S

.

f

f

i

d

n

u

e

a

e e

c

p

u

y 20 40 t

t

- c

a

m

L

u

i

m

s

r

u

i

e

C

m a

e 5 e

c

h

t

a

l

n Dipsacaceae l

a

y 5 i

l

h

e p

H

o 5 y

r

a

C

5

.

f

f

i

d 20

n

u

Chenopodiaceae

s

m

u

l

u

t e

r

d

a

o

e

Steppe p

h 20

s

p

p

y

s

t m

-

A

u

a

i e

r

g

.

a

f

y

f c

L i

i

e

r

d 5

t a

p

i

n

a

y

s

i

t

u

-

M

a

m 5 m

e

e

e

t r e u

r

p u

p

d

A r

y

a y

t t

t

o 5

-

-

n s p

a

e

t u o

C

u a

n

l a

i

d 5

O

o

r

d /

e

a

e

e a

c

C

p n n

m

i 5 l y

/

a

r r

t l

-

o a

r

o j

C o

g a

n 5

a i

m t

n m

o

a

g

a l

a P

b 10

t r

n

o

s a

i l

u P

g 5

n

a

a

S i

c 5

a

c

A m

u 5

n

e o

Rubiaceae

g p

i

l

y 5

t e

-

h

a

i T

r

e 5

a

p

n y

i

t

L -

5

m

e

u

p .

d

f

y

f

e

t i

-

S 10 d

a

l n

l

i

u

Maquis Garrigue Rocky c

S

m 5 a

u i

r

n

i

a

L

r

e 5 a

b a e

e u

a

r

T

p y

l 5

o l

r i

u h

e P

.

f 5 a

f

i e

l

a

d

i

O

c n

a 5

u

t

s

e

i

e

a

P

p

e 5

y c

t a

-

s s

i

l o

i

R 10 g

a

r

f

a

e r

Ericaceae d

p 10

e y

t

h -

s

p

u

E

l

y 5

s r

u

o

n C i

p 5

Woodland r a

C

5 a

y

r

t

s

s

O 10

u

c

r

e

u Q 10 The modern pollen spectra (M. Farrell).

%

s 20 40 60 80 100

u

n

i

P

e

l

p

m

a S MIS B1 MIS B4 MIS B5 MIS B6 MIS B8 MIS B9 MIS C2 MIS C3 MIS C4 MIS C5 MIS C7 MIS C8 MIS B10 Figure 3.2.

81

Chapter 3

Zonation

e

p

y SDC-14 SDC-13 SDC-12 SDC-11 SDC-10 SDC-09 SDC-08 SDC-07 SDC-06 SDC-05 SDC-04 SDC-03 SDC-02 SDC-01

t

-

e

m p

y

u

i

t

l -

l

o

A i 10

c

e

n

e

S 10

.

f

f

i

d

n

u

Dry Catholic Dry e a

e

e c

u

p 20 40 t

y

c

t

-

a

L

m

u i

s

r

i

e

C 10 p

y

t

-

a

l

Dipsacaceae 10 a

g

y

l

s

o

m

u

P 10 r u

y

m

h

e t

h a

t

L 10

n

a

i

l

e Note: exaggeration represent x10 dot-shaded curves

H 10

s

u

t

s i

C 10 Caryophyllaceae 10

20 m

u

i

m h

Chenopodiaceae .

f c u

f

y s

i

r s

d t

o

l o

n

g B

u 10

o

s

i

u

h m l

p e

u

t

d

Steppe O

r 10

o

a

h

p

s p

s

e

A

m 10

. p

u

f

y

f e

i t

g - a

d i

y

s

n m L

i 10

u

u

m

d

e r

t

o r

p

A 10

o

n

O

/

Centaurea

a 10

n

i

l

r

a

C 10

a

t

e a

l

p

o

y

e

t

- c

r

n

o

a

l n

i

o

m

g

a a

t

e

b 20 40

r n

p

o a

y

l e

s

t

i P

p

- r u

y

t e

g

-

b

n s

u

a u

r l

S 10

s o

i

u

d

h

a

t l

n

G 10

a

r

m t

u n

e n

o

C 10

g

a i

l l

e u

h n

T

a 10

e p

p

m y

t

a

-

C 10

m

u

d

e

.

f

S 10 f i

d

Garrigue Garrigue/Rocky Rocky n u

Rubiaceae undiff. Rubiaceae 10

m

u

a

i n

i r

L

a 10

r

e e

p

b

y a

u

t

s

-

T 10

o

i a

i

m

e

b

d

u p

a

i

e

y

r c

t

S u m

- 10

e

a

a r t

i

r a

n

a e g

l

l

C

u 10

u

v

g

r

e

m

p

u

S

r 10 t

s

u

g

i

L 10

a a

e

e

r

a

y

Maquis e l 20 p

l

i

p o

r h

y

t

u P

-

e

s

u a

e n

l i

r

O 10 a

m

s

o

R 10

.

f

f

i

d

a

i

n

c

u 20 40

a

t

e

s

i e a

P p e

y c

t

a

-

s

e s

i

l o

a

i

R

e 10 g

c

a

a

r

c f

i

r

a

r

E 10

d

e

s h

u

p

r

E 10

e

p

m

Woodland s i

i

l u

n

i

a

u d

J

g 10 o

e

p

r

y

l a

o

d

P n 10

u

s

m

u s

r

o

O 10

b

e

l

l

e

s

H

u 10

c

r

e

u

Q 10

s 20

Non-native u

n

i

P

a

i

s l

i

u

T 10

n

i

p

r

a

C 10

e a

y p

r

y

t

t

s -

s

O 10

u l

y

r

o

C 10

a

l

u

t

e

B 10

s

e

i

b A 10 %

Pollen zonation for the Salina Deep Core (L. Coyle-McClung, C.O. Hunt & M. Farrell). Pollen zonation for the Salina Deep Core (L. Coyle-McClung,

Depth (cm) Depth P

B

.

l

a 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600 2700 2800 2900

c

s

r

a

e

Y C

B

/

D 5000 6000 7000 8000 9000

A

s

r

a

e Y 3000 4000 5000 6000 7000 Figure 3.3.

82

The Holocene vegetation history of the Maltese Islands

n

o i

t

a

n

o Z

SDC-14 SDC-13 SDC-12 SDC-11 SDC-10 SDC-09 SDC-08 SDC-07 SDC-06 SDC-05 SDC-04 SDC-03 SDC-02 SDC-01

s

e

r

o

p

s

d

n &

a

n

L

e

l

d l

e

o

t

p

a

v

d i

t n

l

a

l

l u C

a

/

t 100 200 300 400

o d

T e

b

r

u

t

s i

D

c

i

l

o

h

t

c

i

l a C

o

h t

a C

y

y )

k r

e

c

D

p

o

y Note:curves represent x10 dot-shaded exaggeration e R

t

/

- p

e

a

p

r )

u

e

o ) e

t

g

i p e

S p

r

s

y p

r

o t

y

a

-

t

d

s

-

G

o n

a

i

P

a 20 40 60 80 100 ( r

t

i

a )

Maquis

8

s

e

l m

6

i

p

a

3

)

)

- y

V Woodland

e t m ( e

V

-

u p p

a 3 i d

l

y

y

i

6

n

t

t

H 10 r -

- h

2

) -

a A

p

(

r

e

V

o

o

p

2

z

d

i p

y

6

H s

t

h 10

2

- r

o

-

Arnium

a

r

(

r a V

e

c

o t

1

d

e

p p

6

i

H

h s 10

r

2 t

o -

T

i

o (

r ) V

r

a e 9 o d

v

6 p p

H

i 10

y

Non-pollen palynomorphs 1 a

- t

i

D

- r (

V

a

a

r 3 d

s

l

4 o

)

H 10 a

1

h

e

-

V

p ( p )

V

o

y

e

0

d

t

c

p

- r

4

H

y 10 e

1

t

-

C -

(

V

m

2

d

u

1 i

H 10

1

- Sordaria m

(

V

o

t

A d

e

5

H a 10

5

- h

V C

(

d

A

H 10

7

-

V

d

H 10

Sporormiella 10 s

M

A

V 10

a

e

a

z

i

h

c

s

o

d 20 40 60 e

u

s

n

P

o a

t

h

e

p

g

y e

o

T e 10 p

m

y p

t

a

y

m

- t

t

u

-

o

a

n

P

e s 10

g

p o

a l

y

u t h

-

c

p e

Triticum

a

a

S

/ p 10

h

a

y

t m

t

n

n

-

a e

i

r v m

e M

a

u 10 A

c . e

i

f

s d

c

r r

e

o a

i

P

l 10 H

a .

f

e

c

r

a

i

e C

l 10

a

e

r

.

s

f

i

e C

f 10 t

i

i

r

d

Disturbed/Cultivated Land Cultivars Waterside Aquatics e

n

d

i

u

e

S 10 a

m p

i

s

y

r

u

i

i

t

l

l a

-

l

a s

a

i

u

n

s

G

i 10 h

n c

i p e

f

f o

v

r

r

o

c

a

S

o 10

s

g

u

l

a

r u

o

v

l

B 10 o

v

a

n i

o

b

r

C

e 10

a o

a

e

h

e

c

p c

a

. r

a

t

u E c

e 10 l

i

e

s

o

d

s

a n

a

i

r

c

)

B

a 10 l e

t

u

t e

l r

i

o

r

t

P

( 10

a

a

c

i

d

t i

r

s

U

p 10

o

r

e

t

P 10 Pteropsida (monolete) indet. (monolete) Pteropsida 10 Catholic

20 40

Poaceae

.

f Cyperaceae

f 10

i

d

n

u

.

e

f

f a

i

e

d

d

n i

u o

r

e 20 40

e

a t

s e

c

A

a i

.

f

f

m i

a

a

d n

L 10 n a

i

u t

n

e

e

a e

G e

p 10

c

y

t a

i -

p

m

A

u 10

i

l

o

f

i

.

r

f

f

T 10 i

d

n

u

Tamarix

x 10

s e

u

l m

u

u

c

R 10

n

u

n

) a

R 10 m

c % (

h

t

p (cont.). e D

1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600 2700 2800 2900

Years cal. BP cal. Years C

B

/

D 5000 6000 7000 8000 9000

A

s

r

a

e Y 3000 4000 5000 6000 7000 Figure 3.3

83 Chapter 3

Table 3.4. The pollen zonation of the Salina Deep Core with modelled age-depths. Zone Defining Depth Date ranges Characteristics Interpretation taxa (cm) cal. bc (bp) for zone bases (2σ) SDC-01 None 3000– Little to no pollen present Palaeosols and gravels 2710 virtually lacking organic matter. A few isolated grains of Pistacia, Poaceae, Lactuceae SDC-02 Pistacia- 2710– 7042–6553 bc Pinus increasing (8–10%), low Quercus (4%). Pistacia Lentisk-dominated Poaceae- 2660 (8992–8503 bp) rising (5–20%), Ericaceae low (0.5–1%), Phillyrea scrub, dry grassy Pinus present. Poaceae high but declining (20–40%). Some steppe, some pine- Asteroideae (8–15%), Lactuceae, Ophioglossum, oak woodland, Botrychium (2–8%), Chenopodiaceae (4–6%), Asphodelus, some wetland, soil Cyperaceae (1–3%), Theligonum, Plantago lanceolata- disturbance declining type. Persicaria maculosa-type present. Pseudoschizaea decreasing (10–20%). Vesicular arbuscular mycorrhizae (VAMs) and Tripterospora-type also present SDC-03 Poaceae- 2660– 6858–6419 bc Pinus fluctuating (5–11%), Quercus low (1–4%), Pistacia Dry grassy steppe, Pistacia 2510 (8808–8369 bp) declines (4–10%), Phillyrea and Ericaceae present. with some scrub and a Poaceae increase (30–48%), Asteroideae (14–20%) little woodland, some and Chenopodiaceae rise (3–12%), some Lactuceae, wetland, episodically Ophioglossum, Botrychium (2–8%), and Asphodelus disturbed soils (1–5%). Theligonum, Sedum-type, Helianthemum, Cyperaceae present. Pseudoschizaea fluctuates (2–16%), Tripterospora-type and VAMs present. A layer at 25.75 m contains no organic material (highlighted in Fig. 3.3) SDC-04 Pistacia- 2510– 6489–6135 bc Pistacia high and rising (15–40%), Pinus low (4–6%), Steppe partly being Poaceae- 2430 (8439–8085 bp) Quercus rising (2–7%). Phillyrea and Ericaceae present. replaced by lentisk Asteroideae Poaceae (15–20%), Asteroideae high and declining scrub and with oak (10–18%). Theligonum, Ophioglossum, Botrychium, woodland expanding Chenopodiaceae low and declining, Helianthemum, slightly, some wetland, Lactuceae and Cyperaceae low (1–3%). Aquatics present, soils less disturbed than including Potamogeton. Pseudoschizaea and VAMs present previously SDC-05 Poaceae- 2430– 6350–6037 bc Low Pinus (6–12%), Quercus rising (1–6%). Pistacia Steppe with a little Pistacia- 2265 (8300–7987 bp) declines (6–2%), low Phillyrea (0.5–2%) and Ericaceae, woodland, scrub and Asteroideae Theligonum, Tuberaria, Helianthemum present. wetland, with some Asteroideae high (15–38%). Poaceae high and rising eroding soils (15–36%), Lactuceae (6–16%), Cyperaceae (6–8%), Chenopodiaceae (0–12%), Sedum (0–2%), Ophioglossum (3–8%), Botrychium (1–8%) and Asphodelus (1–2%) all increase. Persicaria maculosa-type present. Pseudoschizaea rises significantly (5–30%), other NPPs present include VAMs, Podospora-type, Tripterospora-type and Valsaria varispora-type SDC-06 Cerealia- 2265– 6067–5821 bc Pinus peaks (6–16%). Quercus (2–6%) and Pistacia Steppe with a little Poaceae- 1965 (8017–7771 bp) (0.5–5%) both low, Phillyrea rises (1–8%). Ephedra woodland, scrub and Pinus- fragilis-type, Ericaceae, Tuberaria, Sedum-type, wetland. Initiation Phillyrea Theligonum and Asphodelus are present. Ophioglossum of cereal cultivation (1–2%), Botrychium (2–6%) and Chenopodiaceae and livestock grazing (0.5–3%) have low peaks in the middle of the zone. with some indicators Poaceae (28–44%) and Asteroideae (15–22%) are of disturbed ground high, and Lactuceae (4–9%) fairly high. The Cerealia- and significant fungal types Hordeum and Avena/Triticum are first recorded indicators of disturbed and Plantago lanceolata-type peaks early in the zone ground and animal and rises again at the end. Convolvulus arvensis-type, dung Urtica, Sideritis, Borago officinalis and Brassicaceae are present. Typha and Persicaria maculosa-type are present. Pseudoschizaea initially increases from 15–44% before declining to 22%. Diporotheca rhizophila-type and Tripterospora-type expand slightly, and Sordaria-type appears for the first time

84 The Holocene vegetation history of the Maltese Islands

Table 3.4 (cont.). Zone Defining Depth Date ranges Characteristics Interpretation taxa (cm) cal. bc (bp) for zone bases (2σ) SDC-07 Plantago 1965– 5625–5419 bc Pinus declines slightly (2–10%), Quercus and Pistacia Steppe with slightly lanceolata 1755 (7575–7369 bp) low and rising slightly (2–5%), Phillyrea falling after a declining minor -Cerealia- low peak (1–7%), Ericaceae, Sedum-type, Theligonum woodland, with scrub Poaceae and Tuberaria rise slightly (0.5–3%). Plantago lanceolata- increasing slightly. type increases (5–16%), Ophioglossum, Botrychium, Strong indications Chenopodiaceae (1–5%), Lactuceae (6–12%) and of grazing and the Asteroideae (22–34%) rise slightly while Poaceae spread of short-grazed decrease (16–30%). Brassicaceae, Convolvulus arvensis- grassland, some type, Urtica, Hordeum-type and Avena/Triticum-type indications of grazing are present. Persicaria maculosa-type and Typha are pressure. Continued present. Pseudoschizaea high and fluctuating (15–45%). cereal cultivation, and VAMs rise slightly (0.5–2%). Diporotheca rhizophila- strong soil erosion type, Tripterospora-type and Podospora-type remain low. Sordaria-type and Valsaria varispora-type are occasionally present SDC-08 Plantago 1755– 5404–5145 bc Pinus rising (6–14%). Quercus (1–4%), Pistacia (0.5–2%), Steppe with minor lanceolata- 1650 (7354–7095 bp) Phillyrea (0.5–4%), Ericaceae, Tuberaria, Theligonum are scrub and woodland. Poaceae- all low. Plantago lanceolata-type rises strongly (15–42%). Further expansion of Cerealia- Ophioglossum, Botrychium, Chenopodiaceae and short-grazed grassland Asteroideae Cyperaceae remain low (1–4%). Lactuceae peak and fall with grazing animals (5–20%). Asteroideae (12–22%) and Poaceae (14–30%) indicated by dung remain high. Convolvulus arvensis-type, Brassicaceae, fungi. Some cultivation Borago officinalis and Urtica increase slightly. Avena/ of wheat and barley Triticum-type and Hordeum-type are present. Persicaria and marked phases of maculosa-type and Typha are present. Pseudoschizaea soil erosion highly variable (5–45%), VAMs rise slightly (1–3%), Diporotheca rhizophila-type, Podospora-type and Tripterospora-type increase slightly. Sporormiella-type makes its only appearance in the profile and Cercophora- type, Valsaria varispora-type, Sordaria-type and Chaetomium-type are occasionally present SDC-09 Plantago 1650– 5202–4759 bc Pinus (5–15%) and Quercus (1–5%) peak in the middle Steppe with woodland lanceolata- 1510 (7152–6709 bp) of the zone. Pistacia (1–9%) and Phillyrea (<1%–6) and a little scrub. Poaceae- decline sharply near the base of the zone. Ericaceae, There was a slight Pinus- Theligonum, Tuberaria and Spergularia media-type are recovery of woodland, Cerealia- present. Plantago lanceolata-type (7–16%) declines at the but scrub seems to Asteraceae beginning of the zone then rises again. Ophioglossum, have been cleared or Botrychium, Chenopodiaceae (1–3%) and Cyperaceae grazed down. Grazing (3–6%) remain low. Asteroideae and Poaceae remain pressure seems to have significant (20–30%). Convolvulus arvensis-type, remained strong, but Brassicaceae, Galium, Sideritis and Urtica are present cultivation may have but sporadic, and cereals are represented by Hordeum- waned locally type. Persicaria maculosa-type and Typha are present. Pseudoschizaea (6–52%) highly variable but peaks in the middle of the zone. VAMs, Diporotheca rhizophila-type and Tripterospora-type present consistently. Podospora- type, Sordaria-type, Chaetomium-type and Arnium-type occasionally present SDC-10 Plantago 1510– 4566–4177 bc Pinus and Quercus low (2–6%). Pistacia peaks (6–18%). Degrading steppe with lanceolata- 1435 (6516–6127 bp) Phillyrea and Ericaceae are low (0–2%). Tuberaria, encroaching scrub Pistacia- Theligonum, Sedum-type, Asphodelus, Ophioglossum, and a little woodland. Lactuceae Botrychium, Helianthemum, Polygala-type and Woodland is minor but Cyperaceae are all low. Plantago lanceolata-type lentisk scrub expands. declines then rises (3–12%), Lactuceae peak (12–20%). Markers of degraded, Asteroideae (12–20%) and Poaceae (16–22%) decline perhaps rather over- slightly. Convolvulus arvensis-type and Brassicaceae are grazed steppe increase present. Persicaria maculosa-type present. Psuedoschizaea but fungal spores often declines (8–13%). VAMs, Diporotheca rhizophila-type associated with animal and Tripterospora-type are present dung become less frequent so it is possible that the degradation

85 Chapter 3

Table 3.4 (cont.). Zone Defining Depth Date ranges Characteristics Interpretation taxa (cm) cal. bc (bp) for zone bases (2σ) of the steppe might be climatic, or livestock were moved elsewhere. Cultivation appears to have ceased locally SDC-11 Poaceae- 1435– 4237–3685 bc Pinus rises (8–10%), Quercus falls slightly (2–4%), Degrading steppe Plantago 1320 (6187–5635 bp) Pistacia falls (0.5–2%). Phillyrea fluctuates (1–23%). and scrub with a little lanceolata- Tuberaria, Theligonum, Asphodelus, Ophioglossum, woodland. Woodland Phillyrea- Botrychium, Chenopodiaceae low (0–2%). Plantago is minor. Lentisk Pistacia- lanceolata-type peaks (5–18%). Lactuceae (7–14%), scrub is replaced by Lactuceae- Asteroideae (18–31%), Poaceae (12–24%) all rise then Phillyrea (false privet). Asteroideae decline. Brassicaceae, Convolvulus arvensis-type and Short-grazed grassland Persicaria maculosa-type all present. Pseudoschizaea rises, and disturbed falls and rises again (12–24%) and VAMs, Diporotheca ground indicators are rhizophila-type, Tripterospora-type and Podospora-type present and markers are present for soil erosion are more prevalent than previously SDC-12 Poaceae- 1320– 3823–2979 bc Pistacia rises, then declines (2–27%). Quercus is low and Degrading steppe Pistacia- 1195 (5773–4929 bp) declines gently (1.5–4%). Pinus is low (1–5%). Ericaceae, and scrub with a little Plantago Asphodelus, Plantago lanceolata-type (1–7%), Tuberaria, woodland. Woodland is lanceolata- Chenopodiaceae, Cyperaceae and Convolvulus minor, lentisk replaces Asteroideae arvensis-type (0.5–2%) are all low. Lactuceae (15–37%), Phillyrea but then wanes. Asteroideae (14–27%) and Poaceae (12–26%) all Short-grazed grassland fluctuate. Pseudoschizaea fluctuates (4–16%). VAMs may be replaced by (2–6%) have a low peak and Diporotheca rhizophila- taller vegetation, but type and Tripterospora-type are present. Valsaria there are short phases varispora-type is recorded at the beginning of the zone. of degradation and soil Sordaria-type and Arnium-type appear at the end erosion SDC-13 Poaceae- 1195– 3047–2355 bc Pinus, Quercus, Pistacia, Phillyrea, Ericaceae, Tuberaria, Steppe with very little Pistacia- 1165 (4997–4305 bp) Theligonum, Olea europaea are all very low (0.5–2%). woodland or scrub. A Lactuceae- Plantago lanceolata-type peaks, declines then rises very open landscape Asteroideae- again (0.5–16%). Botrychium, Ophioglossum (0.5–6%), characterized by grassy Olea Asteroideae (12–27%) and Poaceae (15–33%) peak then herbaceous vegetation decline. Lactuceae decline (8–9%). Cyperaceae (2–6%), with signs of grazing Convolvulus arvensis-type and Persicaria maculosa-type are present. Pseudoschizaea declines (7–19%), while VAMs, Diporotheca rhizophila-type, Tripterospora-type and Podospora-type all rise slightly

SDC-14 Olea- 1165– 2717–2233 bc Pinus (4–6%), Quercus (1–5%), Pistacia (6–15%), Phillyrea Degraded steppe and Ligustrum- 1125 (4667–4183 bp) (1–5%), Olea europaea, Ericaceae, Tuberaria, Theligonum, with a little scrub Poaceae- Asphodelus (1–2%), Chenopodiaceae (5–12%), Lactuceae and woodland. The Plantago (15–29%), Asteroideae (12–27%) all rise. Ligustrum rise of scrub taxa and lanceolata- vulgare-type appears. Plantago lanceolata-type (1–5%), decline in Plantago Pistacia- Cyperaceae (1–6%) and Poaceae (10–14%) decline. may suggest declining Phillyrea- Convolvulus arvensis-type and Persicaria maculosa-type grazing pressure, but Lactuceae are present. Pseudoschizaea (5–10%), VAMs (1–2%), the rise in composites Diporotheca rhizophila-type, Podospora-type and and chenopods and Tripterospora-type are present decline in grasses might suggest a more arid landscape and/or are a sign of grazing pressure Top of sampled interval has an interpolated date of 2464– 1979 cal. bc (4414–3929 cal. bp)

86 The Holocene vegetation history of the Maltese Islands

et al. (2015) have argued that over-representation a high resolution record spanning the period c. 7030– of Cichorieae, members of the Lactuceae tribe, may 6600 cal. bc (7923 bp; UBA-34973) to 2400–2050 cal. bc actually be an indicator of open habitats and grazing (3793 bp; UBA-34970), although the coring technique activity in Mediterranean landscapes, rather than the caused a series of short sampling gaps. The sampled result of differential pollen preservation. sequence was laid down during a period of rapidly The off-shore samples provide an analogue rising sea level (Lambeck et al. 2011), so distances for assemblages from shallow marine layers in the from the shoreline (and thus parent vegetation) to the FRAGSUS cores. The assemblages are comparatively core site will have increased during this interval. This uniform (Table 3.3), characterized by lower percentages may have impacted slightly on pollen taphonomy, of degradation-resistant taxa (e.g. Pinus, Lactuceae and although no strong trends in taphonomic indicators various Asteraceae), than are present in the onshore soil seem to be visible in the record. Of particular note in samples. These taxa are, however, still over-represented this core are the events in zone SDC-06 commencing at relative to their occurrence in the catchment vegeta- 6080–5920 cal. bc (7145 bp; UBA-35586) which appear tion. Ostrya and Corylus-type, which are not present to show the impacts of initial agricultural activity on in the Maltese Islands, reflect long distance dispersal the Maltese Islands. by wind from Sicilian or more distant vegetation to The palynological results from the Salina 4 core the north. Shoreline taxa such as Chenopodiaceae are shown in Figure 3.4 and Table 3.5. The sampled and Brassicaceae are well-represented. Some anemo- interval in this core overlaps with the later part of the philous taxa, such as Pinus, Quercus and Poaceae, are Salina Deep Core, with dates ranging from before over-represented relative to their occurrence in the c. 5630–5500 cal. bc (6644 bp; UBA-30090) to 1260–1050 vegetation, while several entomophilous taxa present cal. bc (2953 bp; UBA-30083) (Table 2.3). onshore, principally maquis and garrigue species such as Pistacia, Olea europaea, Tuberaria and geophytes such 3.5.2. Wied Żembaq as Asphodelus, are either not present or are poorly The pollen zonation for the Wied Żembaq 1 core is represented, although the entomophilous Asteraceae shown in Figure 3.5 and Table 3.6. The sampled interval and Lactuceae are over-represented. runs from before c. 3630–3370 cal. bc (4707 bp; UBA- Although the on-shore and off-shore samples 28262) to after 910–800 cal. bc at 2.15 m down-profile reflect the extant vegetation imperfectly as the result (2707 bp; UBA-29042). The Wied Żembaq 2 core, which of these taphonomic biases, the overall aspect of that was not anlaysed for pollen, exhibited a date range vegetation can be seen. The taphonomic biases are from c. 3260–2890 cal. bc (4366 bp; UBA-34969) to after predictable and can therefore be allowed for in the 1880–1630 cal. bc at 2.95 m down-profile (3428 bp; interpretations of the palaeo-ecological sequences that UBA29043). follow in the next sections. 3.5.3. Xemxija 3.5. The pollen results The pollen zonation for the Xemxija 1 core is shown Michelle Farrell, Lisa Coyle-McClung & Chris O. Hunt in Figure 3.6 and Table 3.7. The sampled interval here runs from c. 8780–8460 cal. bc (9353 bp; UBA-25001) Some 25 deep cores were taken from a wide selection to some time in the twentieth century when natural of valley and near-shore environments on Malta and deposition was terminated by the emplacement of Gozo (see Chapter 2) (Fig. 2.4), but only the Salina Deep quarried material (made ground) to produce a dry area Core, Salina 4, Wied Żembaq 1 and Xemxija 1 on Malta for scout camping. Age control at the base of this core were selected and fully analysed for their palynological is, however, problematical and assemblages from the records, each with an associated AMS radiocarbon part of the core below 6.6 m show signs of significant chronology (Table 2.3) and age-depth models (Tables taphonomic bias. It is also rather probable that there 3.4–3.7). In addition, buried soils were sampled and are major erosive discontinuities in the core, especially analysed from the Santa Verna and Ġgantija temple below pollen zone XEM-05, dated to c. 4330–4050 cal. bc sites on Gozo, and the primary fill of a Bronze Age silo (5357 bp; UBA-29041). Sedimentation above this level at In-Nuffara on Gozo. seems to have been more continuous, but there are significant taphonomic biases evident in pollen zones 3.5.1. The Salina cores XEM-06 and XEM-10 to XEM-13. The results of the palynological analysis of the Salina Deep Core are shown in Figure 3.3. A total of fourteen 3.5.4. In-Nuffara zones, labelled SDC-01 to SDC-14, were identified and The results of pollen analysis of the primary fill of are summarized in Table 3.4. This core appears to have a Bronze Age silo at In-Nuffara, Gozo, are shown 87

Chapter 3

Zonation

e

p

y SA-08 SA-07 SA-06 SA-05 SA-04 SA-03 SA-02 SA-01

t -

m

u

i

l e

l

p

A 10

y

t

-

o

i

c 20 e

n

e

S

e

a

e

e c

p

u 20 40

t

.

y

f

c t

f

- i

a

L d

m

n

u

i

u

s

r e

i a

l

p

C 10 a

y

t g

-

y

l

Dry Catholic Dry a

l

o

l m

i

P t 10 u

n

m

e

t e

o

h

t

P 10

n

a

i

l

e

H 10

s

Note: exaggeration represent x10 dot-shaded curves u

t

s i C 10

20 40

Caryophyllaceae

.

f

f

i

d

n

u

s 20 40 60 80 100

u

l

e

e

p d Chenopodiaceae

y

o

t

h -

e

p

a

i

s

p

s

A

y 10

o

t

r

-

b

a

i

r m

e

a

A 10

c p

i

r

y

t

t

-

a

a

i

e

s

e

m i

Steppe M

10 p

p

u

y

m

y

d t

t

e

r -

-

t

o r

a s

i

p A u

r 10

o u a

n

d b

r

O

m a

/

u C

a

l 10

n

o

i

l c

r

a a

s

C 10

e o

i

e p

b

y

p

a

t

y

-

c

t

e

-

S

a 10

p

s

a i

y

t

c

t

a - c

l

a u

o

i

S

e 10

d

c e

n

m a

l

/

r

o

o

j g

a

a

t 20 m

n

a o e

l

g p

P

a

y

t

t

-

n

s

i a

l

r

t

P 10 s

e

v

l

y

s

Acacia 10

a

v

l

m a

u

i

M 10

n

o

m i

L 10

Garrigue Rocky m u

n

o

Rubiaceae 10

e .

g

f i

l p f

i

e y

t d

h

-

n

T 10

e u m

p u

y

d

t

e

. -

f r

S 10

f

i o

d n

i

n

m u

Centaurea 10

a

m

b

r u

o

n

i

s

i

L 10 u

g

Maquis n

a

a a

e

e

S 10 r

a

y

l p

l

i

o

r h

u

P 10

e

a

e l

O 10

a

i

c 20

a

t

s

i

e

P

n

h

p

e

e a

a

p D

e 10

y

c

t

a

-

s

e s

i

l o

a

i

R

e 10 g

c

a

a

r

c f

i

r

a

r

E 10

d

m

e

u h

i

p

d

E 10 e o

p

p

y

y

. l

t

f

- f o

i

s

P

a 10

d

r u

n

e

o

u d

u

e

s d

i

H

u 10 c

c

e r e

e

d p

u

y

s

t

Q u

- 10

c x

r

e

l e

i

u

s

Q

u 10

c

r

e

Non-native Woodland u

Q 10

s

u

n i

P 10 e

a

y p

s

r

y

t

u

t

l s

-

u

s

O

t 10

u

e l

y b

r

s o

u

C 10

n

i

p

r

a

C 10

Alnus 10

s

u

g

a F 10 %

Pollen zonation for the Salina 4 core (M. Farrell).

Depth (cm) Depth P

B

. 300 350 400 450 500 550 600 650 700 750 800 850 900 950

l

a 1000 1050 1100

c

s

r

a

e

Y C

B

/

D 4000 5000 6000 7000

A

s

r

a

e Y 2000 3000 4000 5000 Figure 3.4.

88

The Holocene vegetation history of the Maltese Islands

n

o

i

t

a

n

o Z SA-08 SA-07 SA-06 SA-05 SA-04 SA-03 SA-02 SA-01

200 400 600 800

d

Total land pollen & spores & pollen land Total e

b

r

u

t

s

i

D

Catholic

Dry Catholic Dry

) curves dot-shaded representx10 Note: exaggeration

m

e

a

r p

o

p

f

(

e

A)

t

.

a

S

p t

s

e )

a

) e

a

h

e 20 40 60 80 100 p

n Maquis

c

y i

p

r

t o

y

i

-

e

t

n -

a

g

i

i

o

a

Woodland

h

r

b

)

C

c

( o

o

e

s

l

p t

p 6

i

G s

l 10

y

1

o

e

t

0 r

-

-

D

e

a

(

t

r

M

p

6

o )

i

T 10

r

0 h e

T

0 p

p (

-

y o

t 9

c M

-

r

6

T

a 10 e

1

i

-

r C

(

V a

d 2 d

r

1

H

o 10

1

S -

(

V

A d

5

H 10 a

5

l

- l

e V i

d

m

H r 10

o

r

o a

e

p

S

a 10

z

e i

h p

c

y

s

t

-

VAMs

o 10

d m

u

u

i e

s

n

n

P

a 10

o

g t

r

e

a

g

p e

o

S e p 10

m p

y

t

y

a

-

t

t

-

a

o

l

l a

P

e 10 s

e

p

r

o

l

y

o

e

t

t u

- t p

c

i

a

y

a

L 10

t

h

-

t

m

s

i n

r a

e

i

c

r

a M

a 10

s c

i

u

s

l

r

u

e

e

c

p

P 10

n

y

t u

-

n

m a

u

R 10

e

e

d

a

r p

u

o y

q

t

i H

-

l 10

i

s

s

u

s

i

a

n

t i o

i

a

n g

V

y 10

o a

c

t

h

a t a

i

r

e

e g

r

C u 10 a

a

t m

n

m e

e

C t 10

s

o

r

g

e

A

a 10

p

l

l a y

l

t e

l

DisturbedCultivars weeds Cultivar Waterside- Aquatics palynomorphs Non-pollen

e g

s

i

i

n

s N

a 10

i n

r

e

e

v l

r

a

e

a

V

h 10

s t

u n

l

i

r

u

e

v

l

C 10 o

v

a

n i

o

b

r

C 10

o

h

p

u E 10

20

Brassicaceae

a

c

i

t

r e

a

U 10

e

c

a

o

P 10

.

f

f

Catholic i

d

n 20

u

e

Cyperaceae a

e

d i

o

r 20

e

t s

s

u

A

h

t

n

a

e

c

a

A e 10

c

a

i

p

e

A p 10

y

t -

m

Fabaceae

u 10

i

l

e

o

f

p

i

r

y

.

t

f T 10

- f

i s

d

u

t

n

o

u

L 10

s

u

b

.

u

x

f

i

f R

r 10 i

a

d

n m

u

a

T x 10

e

m

) u

R 10 m

c %

(

h

t

p

e

D P

(cont.).

B

. 300 350 400 450 500 550 600 650 700 750 800 850 900 950 l

a 1000 1050 1100

c

s

r

a

e

Y C

B

/

D 4000 5000 6000 7000

A

s

r

a

e Y 2000 3000 4000 5000 Figure 3.4

89 Chapter 3

Table 3.5. The pollen zonation of the Salina 4 core with modelled age-depths. Zone Defining taxa Depth Date ranges Characteristics Interpretation (cm) cal. bc (bp) for zone bases (2σ) SA-01 Chenopodiaceae- 1055– 5608–5238 bc High Chenopodiaceae (30%), Cyperaceae An extremely open, rather Cyperaceae 1040 (7558–7188 bp) (29%). Some Lactuceae (12%), Senecio-type degraded landscape (4%), Asteroideae (4%), Poaceae (5%), characterized by ruderals Brassicaceae (4%). A littlePinus, Quercus, and with coastal vegetation Rosaceae, Matricaria-type, Artemisia, prominent, perhaps with Cirsium-type, Asphodelus and Rumex some woodland at a distance. (0.5–2%). Mentha-type and Potamogeton Freshwater wetlands with present. Pseudoschizaea and VAMs present sedges and slow-moving waters higher in the estuary SA-02 Lactuceae- 1040– 5566–5130 bc High Plantago lanceolata-type (10–24%), An extremely open landscape. Plantago-Poaceae 940 (7516–7080 bp) Lactuceae (23–36%). Some Senecio-type Perhaps some woodland and (4–9%), Chenopodiaceae (7–16%), Rumex scrub at a distance. Decline (2–4%), Asteroideae (2–4%), Cyperaceae of ruderals suggests the (3–6%), Poaceae (6–9%), Brassicaceae vegetation is less degraded in (1–4%). Low Pinus, Quercus, Rosaceae, aspect than SA-01 and there is Artemisia, Asphodelus, Cirsium-type, Lotus- evidence for cereal cultivation type, Fabaceae (0.5–2%). Olea europaea, and animal grazing. Declining Centaurea, Theligonum, Malva sylvestris- soil fungi suggest maybe type, Matricaria-type, Tamarix, Euphorbia, reduced soil erosion Hordeum-type, Mentha-type, Littorella-type sometimes present. Sporormiella-type, Sordaria-type, Pseudoschizaea present, Cercophora-type, Delitschia-type, Tripterospora-type, VAMs occasionally present SA-03 Lactuceae- 940–780 5063–4670 bc High Lactuceae (18–32%), Pistacia An open, rather degraded Pistacia- (7013–6620 bp) (2–14%), Chenopodiaceae (10–26%). Some landscape with prominent Chenopodiaceae Plantago lanceolata-type (4–9%), Senecio- ruderals, encroaching type (8–12%), Rumex (1–9%), Asteroideae garrigue and scrub, woodland (3–6%), Cyperaceae (2–8%), Poaceae at a distance. Some evidence (3–9%), Brassicaceae (2–6%). A little Pinus, for cereal cultivation and Quercus, Daphne, Ericaceae, Olea europaea, animal grazing. Some Rosaceae, Sedum-type, Theligonum, Malva evidence for soil erosion from sylvestris-type, Carlina/Onopordum-type, the soil fungi Artemisia, Asphodelus, Helianthemum, Cirsium-type, Tamarix, Fabaceae, Hordeum-type, Mentha-type, Littorella- type, Delitschia-type, Sordaria-type, Pseudoschizaea are present, Coniochaeta Type A, Sporormiella-type, Tripterospora- type and VAMs are occasionally present SA-04 Brassicaceae- 780–690 4431–4026 bc High Brassicaceae (6–24%), Lactuceae An open, fairly degraded Lactuceae (6381–5976 bp) (12–20%), Chenopodiaceae (10–30%). landscape with prominent Some Plantago lanceolata-type (7–8%), ruderals and with a little Caryophyllaceae (8–10%), Senecio-type scrub and woodland. There is (4–12%), Rumex (2–6%), Asteroideae evidence for cereal cultivation (1–5%), Poaceae (1–5%). A littlePinus , and animal grazing and for Quercus, Pistacia, Olea europaea, Rosaceae, some soil erosion Sedum-type, Theligonum, Matricaria-type, Carlina/Onopordum-type, Asphodelus, Helianthemum, Cirsium-type, Fabaceae, Euphorbia, Hordeum-type, Ranunculus acris-type, Convolvulus arvensis- type. Delitschia-type, Sordaria-type, Pseudoschizaea and VAMs are present, Sporormiella-type is occasionally present SA-05 Lactuceae- 690–590 4077–3716 bc High Lactuceae (21–22%), An open, rather degraded Caryophyllaceae- (6027–5666 bp) Caryophyllaceae (10–28%), Plantago landscape with prominent Plantago lanceolata-type (10%). Some ruderals, rising ephemerals, Chenopodiaceae (2–8%), Senecio-type and some scrub and

90 The Holocene vegetation history of the Maltese Islands

Table 3.5 (cont.). Zone Defining taxa Depth Date ranges Characteristics Interpretation (cm) cal. bc (bp) for zone bases (2σ) (2–10%), Rumex (2–6%), Asteroideae woodland at a distance. There (1–6%), Cyperaceae (1–12%), Poaceae is some evidence for cereal (1–5%), Brassicaceae (2–11%). A little cultivation, grazing and soil Pinus, Quercus, Rosaceae, Centaurea, erosion Sedum-type, Theligonum, Rubiaceae, Plantago media/major-type, Artemisia, Asphodelus, Cirsium-type, Tamarix, Fabaceae, Euphorbia, Ranunculus acris-type, Agrostemma githago. Vitis and Hordeum-type are occasionally present. Delitschia-type, Sordaria-type, Pseudoschizaea and VAMs are present, Sporormiella-type and Tripterospora-type are sometimes present SA-06 Caryophyllaceae- 590–420 3386–2560 bc High Caryophyllaceae (28–45%) and An open, very degraded, Lactuceae (5336–4510 bp) Lactuceae (8–16%). Some Pistacia probably rather dry landscape (2–4%), Plantago lanceolata-type (1–20%), with some scrub and distant Chenopodiaceae (8–10%), Senecio-type woodland. Ruderals and (2–5%), Rumex (1–5%), Asteroideae ephemerals prominent. Some (1–7%), Cyperaceae (2–8%), Poaceae evidence for cereal cultivation (2–9%), Brassicaceae (2–8%). A little Pinus, and a little animal grazing. Quercus, Olea europaea, Rosaceae, Carlina/ Soil erosion was present but Onopordum-type, Asphodelus, Cirsium- perhaps a little less prevalent type, Tamarix, Agrostemma githago, than in earlier zones Hordeum-type, Ericaceae, Sedum-type, Rubiaceae, Matricaria-type, Polygala- type, Fabaceae, Apiaceae, Urtica are occasionally present. Sordaria-type and Pseudoschizaea are present SA-07 Lactuceae- 420–315 2220–1885 bc High Lactuceae (15–25%), Asteroideae An open, rather degraded Asteroideae (4170–3805 bp) (6–14%). Some Plantago lanceolata- landscape with ruderals type (2–9%), Chenopodiaceae (3–6%), and ephemerals prominent. Caryophyllaceae (1–8%), Senecio-type Some scrub and garrigue (1–4%), Rumex (1–9%), Cyperaceae and a little distant woodland. (0–11%), Poaceae (1–7%), Brassicaceae Arable agriculture and animal (1–3%). Some Ericaceae, Rosaceae, grazing were present and Theligonum, Artemisia, Asphodelus, there were pulses of soil Cistus, Cirsium-type, Fabaceae. Pinus, erosion at the bottom and top Quercus, Daphne, Linum, Acacia, Plantago of the zone media/major-type, Agrostemma githago, Centaurea cyanus-type, Hordeum-type, Vitis, Ranunculus acris-type, Littorella-type occasionally present. Delitschia-type, Sporormiella-type, Sordaria-type, Pseudoschizaea and VAMs sometimes present SA-08 Chenopodiaceae 315–300 1326–1064 bc Chenopodiaceae dominant (97%). Probably a strong taphonomic (3276–3014 bp) Caryophyllaceae, Asteroideae, imprint. Little can be deduced Brassicaceae are also present in small regarding the landscape, amounts but the Chenopodiaceae include saltmarsh, coastal and exposed open ground taxa, including arable weeds Top of sample interval has an interpolated date of 1259– 1006 cal. bc (3209–2956 cal. bp)

91

Chapter 3

Zonation

e

p

y WZ-07 WZ-06 WZ-05 WZ-04 WZ-03 WZ-02 WZ-01

t -

m

u

i

l l

A 10

e

p

y

t

-

o

i

c 20 40 e

n

e S

Dry Catholic Dry

e

a

e

c

u 20 40 60 80 t

c

e

a Note: exaggeration represent x10 dot-shaded curves

p

L

.

y

f

t

f

- i

d

m

n

u

i 20

u

s

r

e

i a

l

p C

a

y

t g

-

y

l

a

l

o

l m

i

P t u 10

n

m

e

t e

o

h

t

P 10

n

a

i

l

e

H 10

s

u

t

s i

C 10 Caryophyllaceae 10

20 m

u

i

h

Chenopodiaceae

c

y

r

t m

.

f o

u

f

s i

B 10

s

d

o

l n

g u

o

s

i 20

u

h m l

p

e

u

t

d

O r

o a

h

p

e s p

s

p

A

y

m 10

t

u -

e

a

i

g

Steppe r

y

a

L 10 c

i

e

r

t p

a y

t

-

M 10

m a

i

u

s

i

d

r

m

o 20

e

p

t

r

o

A

n

O

/

a e

n

p

i 20

l

y

r

t

a

-

a C

.

s

f

i

f e

i

c

p

c d

y

u n

t

- S

u 10

a

t

o

a

g

l

a

a

o

i

t

e 20

d n

c

e

a

n l

m a P

l

/

r

o

e o

j

g

p

a

a

y

t

t 20 m -

n

s

i a o

l

r

g t

P

s

a

t e

v

n

l

a y

l

e

s

P 10

p

.

a

f

y

v f

t

l i

-

a d

a

i

n

M

r 10

u

a

n

e

r

A 10

Centaurea 10

m

u

n

o

Rubiaceae 10

e e

Steppe/Garrigue/Rocky g

i

p p l

y y

e

t t

h

. - -

f r T

a 10 f

i i o

r

d n

i a

n

n

i

m

u

L 10 a

m

b

r u

o

n

i

s

i

L 10

e u

g p

y

n

t

a -

S 10

m

u

a d e

i

e r

p

S

a

y 10

r

t

- e

b

m

u

u

T

r 10 t

s

u

g

a a

i

e

e

L 10 r

a

y

l p

l

i

o

r h

u

P 10

e

Maquis

a

e m

l

u

O

n 10

a

g

i

r e

a p O

i 10

y

c

t

- a

t

s

s

i u

P n 10

m

a

h

e

R 10 n

h

p

e

a

a

D

e 10

c

a

s e

e

o a

p R

e 10

y

c

t

a

-

c s

i i

l r

a i

E

n 10 g

a

a

r

f

m

u a

m r

F 10

u d

e n

i

s l h

i

u

p

u r

E 10 q e

a p

i

n

m

u

u

J

i 10

d

i m

r

u

i e

t

d m

P 10

o u

i

p l

y

o

l f

i o

u

P 10

q

a

x

s

Non-native Woodland e l

I

u 10

c

r

e

u

Q 10

s 20

u

n

i

P

a

y

r

t

s O 10 %

Pollen zonation for the Wied Żembaq 1 core (M. Farrell). Pollen zonation for the Wied

P

Depth (cm) Depth B

.

l

a 100 150 200 250 300 350 400 450 500 550

c

s

r

a

e

Y C

B

/

D 2000 3000 4000 5000

A

s

r

a

e 0 Y 1000 2000 3000 Figure 3.5.

92

The Holocene vegetation history of the Maltese Islands Zonation

WZ-07 WZ-06 WZ-05 WZ-04 WZ-03 WZ-02 WZ-01

s

e

r

o

p

s

&

n

e

l

l

o

p

d

n

a

l

l

a

t

o 100 200 300 400 500 T

c

i

l

o

Cultivars

h

t

c a

i

l C

o

h

t

a

y

C

k

c

y

r o

R

D

/

e

u

g i

r

) ) r Note: dot-shaded represent curves exaggeration x10

a e e

p p G

/

y y

t t e

Steppe - -

p

a

m r

p

u o

e

i

t

p

n

) s S

r

e

o

A

r

p 20 40 60 80 100 (

e

y

t Maquis

1 t

-

p

6

i

a

r

2

i

-

r T Woodland

(

a V

9

d

d

r

6

H o 10

1

S -

(

V

A d

5

H 10 a

5

l

- l

e V i

d

m

H r 10

o

r

o

p

S 10 s

M

A

V 10

a

e

a

z

i

h

c

s e

o p

d

y

t 20 40 u

-

e

a s

l

l

P e

e

r p

e

y

o

p t

t

- t

y

i

t a

L

- 10

h

a

t

s

n

o e

l

e u

M 10 p c

y Cultivars Waterside Aquatic NPPs a

t

-

m

s

i

r a

i c

r

a

a

s c

i 20 u

s

l

r

u e

e

c p

P

n

y

t u

-

n

m a

e

u R

p 10

e

y

t

d

- r

e o

r

H

a 10 m l

u

u

r s

c

i

i Disturbed weeds Cult. y

t

i v

p

V

a 10

m

a

m l

u e

n

M 10 o

g

y a

l i

o b

r

P 10

o

h

.

p t

. u e

t E

d 10

e

n

i d

) n

i

e

)

t 20

e e

l t

i

r e

l

t

Brassicaceae

(

o

a

a

n c

i

d o

t

i

r

s

m

U

(

p 10

o a

r

d

e i

t

s

P

p 10

o

r

e

t P 10 20

Catholic Poaceae

Cyperaceae 10

.

f

f

i

d

n

u

e

a

e

d i

o

r 20 40

e

s t

s u

A h

t

n

a

c

A 10

e

Apiaceae p 10

y

t -

m

Fabaceae

u 10

i

l

e

o

f p

i

r

y

t

T 10 -

s

u

t

x o

i

L r 10

a

m

a

T 10 .

f

f

i

d

n

u

x

e 20

m %

) u

R

m

c

(

h

t

p

e

P

D B

. (cont.).

l

a 100 150 200 250 300 350 400 450 500 550

c

s

r

a

e

Y C

B

/

D 2000 3000 4000 5000

A

s

r

a

e 0 Y 1000 2000 3000 Figure 3.5

93 Chapter 3

Table 3.6. The pollen zonation of the Wied Żembaq 1 core with modelled age-depths. Zone Defining taxa Depth Date ranges Characteristics Interpretation (cm) cal. bc (bp) for zone bases (2σ) WZ-01 Lactuceae- 550–520 3587–3352 bc High Lactuceae (9–30%), Poaceae A predominantly open landscape Poaceae-Senecio- (5537–5302 bp) (7–25%), Senecio-type (5–37%), mosaic with grassland indicators, type-Rumex Rumex (10–20%) and Plantago ruderals and ephemerals, some taxa lanceolata-type (1–11%). Some typical of scrub and garrigue. There Pinus, Quercus, Olea europaea, are indicators of cereal cultivation and Ephedra fragilis-type, Rosaceae, grazing animals Phillyrea, Sedum-type, Theligonum, Asphodelus, Chenopodiaceae, Helianthemum, Tamarix, Lotus- type, Trifolium-type, Apiaceae, Cyperaceae, Brassicaceae, Euphorbia, Hordeum-type, Mentha- type and Littorella-type (>2%). Vitis is occasionally present. Sordaria-type, Arnium-type, Sporormiella-type, VAMs and Pseudoschizaea are sometimes present WZ-02 Asteroideae- 520–450 3330–3047 bc Some very low counts, particularly A very open degraded landscape with Lactuceae- (5280–4997 bp) in the top half of this zone. Very some intermittent cereal cultivation Senecio-type- high Asteroideae (6–49%), high and animal grazing and very strong Poaceae Lactuceae (6–54%), Senecio-type soil erosion. It is likely that these (2–38%), Poaceae (4–22%). Some assemblages bear a strong taphonomic Pinus (0–18%), Theligonum (0–5%), imprint consistent with erosion Asphodelus (0–8%), Brassicaceae and redeposition of material from (0–12%), Persicaria maculosa-type soil profiles since many common (0–23%). Pseudoschizaea are very grains in this zone are known to be high (2–39%). Sordaria-type, highly resistant to degradation and Tripterospora-type and VAMs are preferentially preserved in soils. It may occasionally present therefore reflect a cluster of significant erosion and sedimentation events WZ-03 Lactuceae- 450–420 2922–2542 bc Very high Lactuceae (45–51%) with A very open landscape with a little Senecio-type- (4872–4492 bp) some Senecio-type (8–11%), Poaceae scrub and woodland at a distance. It Poaceae (8–12%), Cirsium-type (1–12%) and is likely, given the high percentage Asteroideae (5–6%). A littleOstrya, of degradation-resistant grains, that Quercus, Olea europaea, Ephedra soil erosion continued but this could fragilis-type, Theligonum, Centaurea, also reflect a vegetation dominated by Plantago lanceolata-type, Matricaria- ruderals and ephemerals. There are type, Asphodelus, Caryophyllaceae, indications of a little cereal cultivation and Cyperaceae. Ilex aquifolium, but the lack of dung fungi in this zone Sanguisorba minor-type, makes it difficult to be sure that animals Arenaria-type, Chenopodiaceae, were grazing Brassicaceae, Polygonum aviculare- type and Hordeum-type are sometimes present WZ-04 Lactuceae- 420–390 2728–2322 bc High Lactuceae (10–28%), Plantago An open landscape with a little scrub Plantago- (4678–4272 bp) spp. (7–18%), Asteroideae (9–12%), and distant woodland. Soil erosion Asteroideae- Poaceae (6–9%), Cirsium-type probably continued but at a reduced Poaceae (7–10%), Artemisia (4–10%). Some rate and parts of the landscape may Linaria-type (1–7%), Carlina/ have started to stabilize. Taxa such as P. Onopordum-type (1–5%), Senecio- lanceolata and Artemisia are consistent type (2–4%). A littleOstrya, Pinus, with animal grazing; cereal agriculture Quercus, Theligonum, Centaurea, and vine cultivation were also present Matricaria-type, Asphodelus, Caryophyllaceae, Tamarix, Cyperaceae, Brassicaceae. Pistacia, Olea europaea, Ericaceae, Tuberaria, Sanguisorba minor-type, Linum, Arenaria-type, Helianthemum, Vitis, Hordeum-type are occasionally present

94 The Holocene vegetation history of the Maltese Islands

Table 3.6 (cont.). Zone Defining taxa Depth Date ranges Characteristics Interpretation (cm) cal. bc (bp) for zone bases (2σ) WZ-05 Lactuceae- 390–260 2529–2104 bc Lactuceae (25–51%) very high. A very open steppic landscape with Poaceae-Carlina/ (4479–4054 bp) High Poaceae (15–20%), Carlina/ abundant ruderals and ephemerals. Onopordum- Onopordum-type (6–20%), Some distant woodland but little scrub. type-Plantago- Plantago spp. (5–16%), Cirsium- Abundant Plantago suggests a great deal Cirsium-type type (4–10%). Some Asteroideae of animal grazing. Cereal cultivation (1–6%). Ostrya, Pinus, Quercus, was present at times. Areas of wetland Olea europaea, Rosaceae, were present. High Lactuceae and other Sanguisorba minor-type, Asphodelus, degradation-resistant grains suggest Caryophyllaceae, Helianthemum, continuing soil erosion Senecio-type, Rumex, Cyperaceae, Brassicaceae, Ranunculus acris-type and Persicaria maculosa-type are often present WZ-06 Lactuceae- 260–220 1501–1120 bc Very low count size in uppermost An extremely open, eroding landscape Chenopodiaceae- (3451–3070 bp) sample of this zone. Very high with abundant ruderals and ephemerals. Carlina/ Lactuceae (33–47%) and high Plantago most probably reflects animal Onopordum-type Chenopodiaceae (16–17%). Carlina/ grazing. The high Chenopodiaceae most Onopordum-type (9–11%), Plantago probably represents nearby saltmarsh lanceolata-type (2–18%), Senecio- vegetation in this coastal locality type (7–9%), Artemisia (1–8%) are present in all samples WZ-07 Lactuceae- 220–120 1021–819 bc Some low counts in this zone, An extremely degraded steppic Pinus- (2971–2769 bp) particularly towards the top. Very landscape with woodland at a distance Asteroideae high Lactuceae (54–81%), high and marked by strong soil erosion. Pinus (2–17%) and Asteroideae The vegetation was characterized (2–16%). Plantago major/media- by ruderals and ephemerals, with type, Carlina/Onopordum-type, nearby saltmarsh. It is possible that Chenopodiaceae are present in the relatively high Pinus reflects low most samples local pollen productivity enhancing the regional pollen signal, rather than an expansion of woodland Top of sampled interval has an interpolated date of cal. ad 47–763 (1903–1187 cal. bp) in Figure 3.7 and Table 3.8. Chronological control is do not show strong evidence that corrosion resistant provided by radiocarbon dates on cereal grains recov- taxa are present in percentages typical of microbially ered from the basal fills, ceramic typology and pollen active, strongly oxidizing environments, as seen in biostratigraphy. The pollen analyses show significant modern soils (Fig. 3.2; Table 3.3). This, along with the taphonomic biases but generally suggest extremely very high percentages of spores of liverwort and other open environments. lower plants is perhaps consistent with a perennially humid, local sheltered location and deposition in very 3.5.5. Santa Verna damp sediments. The Peridinium cysts are suggestive Pollen analyses of the buried soil preserved below the of unshaded, mildly eutrophic freshwater. The later floor sequence of the temple structure at Santa Verna, assemblage from the buried lower organic (Ah) horizon Gozo, exposed by the archaeological investigation, (context SV2/1) is too small for reliable interpretations are summarized in Tables 3.9 and 3.11 and Figure to be made. 3.8. Pollen counts are very low and the assemblages The combination of very abundant thermally are very small and probably taphonomically influ- mature (charred) material, plant cell walls and cuticle enced, but the assemblages from the B horizon of the and fungal hyphae is typical of middens, while the high buried soil (contexts SV2/2, 2/3, 5/1, 5/2; see Chapter inertinite in the upper part of Santa Verna 2 (Fig. 3.8) is 5), which most likely dates from the earlier Neolithic, very typical of seasonally wet biologically active soils. 95

Chapter 3

Zonation

e

p

y

t XEM-13 XEM-12 XEM-11 XEM-10 XEM-09 XEM-08 XEM-07 XEM-06 XEM-05 XEM-04 XEM-03 XEM-02 XEM-01 -

m

u

i

l l

A 10

e

p

y

t

-

o i

c

e 20 40

n

e S

Dry Catholic Dry

.

f

f

i

d

n

u

x10 exaggeration represent Note: curves dot-shaded

e

a

.

f e

f

c i

e

u d

p 20 40 60 80 100 t

y

n

c

t

- u

a

L

e m

a

u

i

e

s

c r

i

m a

C c u 10

a

m s

e

p

i h

t

D 10

n

a

i

l

e

H

e 10

s a

u

e

t

c

s

i

a

l

C

l 10

y

h

p

o 20

y r

a C

20 40

Chenopodiaceae

m

u

i

h

c

y 20 40

r

t m

u o

s

B

s

o

l

g

o

s

i 20

u

h m l

p

e u

t

d O r

o

a

h

p

s p

e s

p

m

A 10

y u

t

- e

e

a g

i

y

p s

L

y 10 o

t

r

Steppe - b

a

i

r m

a

A 10 c

.

i

f

r

f t

i

a a

i d

s

n i

M 10

u

m

a

e

t

e

r r

A u 10

a

e

t

p

n

e

y

e

p

t

- C

y 10

t

m

-

s

u

u

d

r u

e 20 o

d

r p

p

y a

o

t

C -

n

a

t O

/

a

l

a

a

o

n i

i

e 20

l d

c r

e

a n

a m C

l

/

r

o

o

j

g

a

a

e

t

p 20 m

n

y

o

a

t

l

-

g r

P

a

o

t

e

n

n i

a

a

h l

m

p P

s 10

a

o u

i

b

l p

r

o p

o

i f

s i

i

s

H 10

u

u

t g

b

n

e

o a

p S

x 10

y e

t

-

m s

u

u

l

R 10

o

i

m

d

u

a l n

o

G 10

e

g

i

p l

e

y

t

h

-

T

a 10

i e

r

p

a

y

t n

Garrigue/Rocky Rocky i -

L 10

m

u

d

e

.

f

S 10 f

i

d

n

u

Rubiaceae undiff. Rubiaceae 10

m

u

a e

i n

i r

p

L

a 10 y

r

t

- e

b

m

u

u

r

T 10 t

s

u

g

i

L 10

e

a a

p

e

e

y r

a

t

y

- l

p 20

l s

i i

o

l

r

h

Maquis a

u

P

n e

i

c

a

i

f e

f l

o

O 10

a

i

v

l

a

S 10

a

i

c

.

f 20 a

f t

i

s

i

d

e

P

n n

u h

p

e

a

a

e

D 10

c

a

e

s

p

o

y

t

R 10

-

s

i

l i

g

a 20 r

f

a

r Ericaceae

d

e

s h

u p

r

E 10 e

p

i

n

m

u

u

J

i 10

d

m i s

i r

e l u

e i

a

p t

d

y g

P 10 o

t

e

- r p

s

i

y

l a

d

i

d

o

r

i

P

n 10

v

u

Woodland

s

m

u s

r

o

O 10

b

e

l

l

e

s

u

H 10

c

r

e

u Q 10

Non-native

s 20 40 60 80

u

n

i

P

e a

y p

r

y t

t

s -

s

O 10

u

l

y

r

o

C 10

s

u

n l

A 10

s

u

g

a

F 10

s

e

i

b Pollen zonation for the Xemxija 1 core (M. Farrell & L. Coyle-McClung).

A 10

%

Depth (cm) Depth P

B

.

l 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950

a 1000

c

s r

a

e 0

Y C

B /

D

1000 2000 3000 4000 5000 6000 7000 8000 9000

A

s r

a

e 0 Y 1000 2000 3000 4000 5000 6000 7000 -2000 -1000 Figure 3.6.

96

The Holocene vegetation history of the Maltese Islands

n

o i

t

a

n

o Z

XEM-13 XEM-12 XEM-11 XEM-10 XEM-09 XEM-08 XEM-07 XEM-06 XEM-05 XEM-04 XEM-03 XEM-02 XEM-01

s

e

r

o

p

s

&

n

e

l

l

o

p

d

n

a

l

500 1000 1500 l

a

t

o

T

c

i

l Waterside

o

h

t

c

i

l a C

o

h

t

a

C

y r

D

)

e x10 represent curves exaggeration dot-shaded Note:

p

y

t

-

a

r

) o

e p Steppe

s p

o

y

t ) d

.

-

o

a p )

i P

Maquis

s 20 40 60 80 100 r e

(

a

p a

8

s

y y l

6 r

t

- a

3 Woodland

a

-

V

a

b

r (

V

e

o

)

3 d

.

D

p

6 (

H

p

s

10

2 s

o

4 -

r

1 a ) V

e

c

2 t

e

d

-

e

p

p

H

i

V 10 h r

y

t d

t

T

-

o

(

H r

a

.

9 o r

f

6 o p ) c

i 10

1 h

e

- D

p

p (

V

o

y

t 3 d c

-

r

4

H 10 e a

1

i

- C

r

(

V

a

d

2 d

r

1

Non-pollen palynomorphs H 10 o

1

S -

(

V

A d

5

H 10

5

-

V

d H 10

20

VAMs

a

e

a

Aquatics z i

h

c

s

o

d 20 40 60 80 u

e

s

P s

e

t

e

o

s

e

I 10

p a

y h

t

- p

y

m

T 10

u

i

n

n

a

o

g

t

r 20

e

a

g

p

e o

S

p

m

y

t

a

- t

a o

l

l

P 10

e

r

o m

t

u

t

i

n

L 10

g

a

h m

p u

e t

S

e

e 10

e

p

p p s

y i

y

y

t

t

t u -

- -

q

a a

m

E 10

s

h

u

t o

e a l

l

n

d u

e u r

c

e

d o

M

a 10

p n

H

y e m

.

t

f

Disturbed Cultivar Waterside p -

i

a

c

i l

s

i

r i

e F s

a 10

p

n c

i

y e

s

t

v r -

r

e

a

a

i

P

l 10

s

a

u

e l

r

u

.

e

v

f

l

C

f 10

i o

d

v

a n

n

i

u o

b

r

C 10 e

o

a

h

e

p c

u a

E

c 10 i

s

s

a

r

.

B t 10

e a

c

d

i

t

n

r

i

.

) U

t 10

e

e

t

d

e

l

n i

i

r

t )

(

e

t

a

e

d l

i

o s

n p

o

o 20 40 r

m e

(

t

P a

d i

s

p

o 20 r

e

t P

20 40 60 Poaceae

Catholic

e

a

e

c

a

r

e 20 40 60 80

p

.

f y C

f

i

d

n

u

e

a

e

d

i

. o

f r

f 20 40 i e

t

a d

s

n

n

A

a u

i

r

e

e

l a

a

e

c

V 10

a i

m

a

a

n

L 10

a

i

t

n

. e

f

f

G i 10

d

n

.

u

f

f

i e

a d

e n

c

u

20 a

i

e

p a

e A

c

a

e

b p

a

y

t

F 10 -

s

u

t

o

L 10

s

u

b

. u

x

f i

R

f r 10

i

a

d

n m

u

a

T x 10

s e

u

l m

u u

c

R 10

n

u

n

a R 10 %

(cont.).

Depth (cm) Depth

P

B

. 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 l

a 1000

c

s r

a

e 0 Y 1000 2000 3000 4000 5000 6000 7000 8000 9000

0 Years AD/BC Years 1000 2000 3000 4000 5000 6000 7000 -2000 -1000 Figure 3.6

97 Chapter 3

Table 3.7. The pollen zonation of the Xemxija 1 core with modelled age-depths. Zone Defining taxa Depth Date ranges Characteristics Interpretation (cm) cal. bc/ad (bp) for zone bases (2σ) XEM-01 Botrychium- 1000– 7675–6861 bc Extremely low pollen counts. High The very limited pollen counts and Poaceae 950 (9625–8811 bp) Botrychium (11–32%) and Poaceae the high percentages of spores and (6–34%) with sporadic but strong other degradation-resistant grains occurrences of Pteropsida (trilete) constrain confident interpretation (0–38%), Senecio-type (0–23%), but assemblages are consistent Cyperaceae (0–20%), Carduus-type with an open, grassy, virtually (0–19%), Ophioglossum (0–16%), treeless steppe with ephemerals Ericaceae (0–15%), Lactuceae and ruderals prominent and with (0–6%) and the occasional presence some scrub. The high percentages of of Pinus, Persicaria maculosa-type Pseudoschizaea may reflect eroding and Sphagnum. Pseudoschizaea soils is common (12–54%) and Tripterospora-type is present in one sample XEM-02 Senecio-type- 950–790 7066–6444 bc Low pollen counts. High Senecio- The low pollen counts and Botrychium- (9016–8394 bp) type (10–48%), Pteropsida considerable proportion of Pteropsida (trilete) (2–39%), Botrychium degradation-resistant grains (Trilete)-Poaceae (1–30%), Poaceae (6–11%). Some constrain interpretation. Pinus, Quercus, Olea europaea, Nevertheless, assemblages are Ericaceae, Pistacia, Daphne, Carlina/ consistent with an open grassy Onopordum-type, Asphodelus, steppe with some ruderals and Ophioglossum, Lactuceae, ephemerals, some scrub and distant Persicaria maculosa-type. Very woodland. Locally, a wetland high Pseudoschizaea (75–87%), a with standing water is suggested few VAMs, some Sordaria-type, by the green alga Debarya and the Tripterospora-type, Podospora-type wetland taxa, while the very high and Debarya sp. Pseudoschizaea is consistent with considerable soil erosion XEM-03 Lactuceae- 790–715 5722–5255 bc High Lactuceae (22–35%) and Rising pollen counts may point to Cyperaceae- (7672–7205 bp) Cyperaceae (22–27%), some declining sedimentation rates, but Poaceae Poaceae (5–10%), Carlina/ the high proportion of degradation- Onopordum-type (1.5–6%), Apiaceae resistant grains is consistent with (0–22%), Senecio-type (0–12%), continued soil erosion and complex Carduus-type (0–6%). Pinus, taphonomic pathways. The rise Pistacia, Ophioglossum and Typha in ephemerals and decline in are often present. Pseudoschizaea grassland indicators suggests strong decline rapidly; VAMs and degradation of steppe habitats. Some Tripterospora-type are present scrub and wetlands were present XEM-04 Lactuceae- 715–660 4923–4439 bc Very high Lactuceae (66–94%). The very high Lactuceae most Poaceae- (6873–6389 bp) Some Pinus, Ericaceae, likely reflect a taphonomic bias, Cyperaceae Chenopodiaceae, Cyperaceae, most probably reflecting eroding Poaceae. Persicaria maculosa-type soils, but may also reflect a high and Typha are occasionally present incidence of ephemerals and ruderals, given that Pinus and other degradation-resistant taxa are not heavily represented. There are hints from the rest of the assemblage of distant woodland, some scrub and wetlands with standing water, but the dominant signal is steppic XEM-05 Lactuceae- 660–570 4251–3989 bc High Lactuceae (15–40%), The assemblages show a less strong Cyperaceae- (6201–5939 bp) Cyperaceae (6–25%), Poaceae taphonomic imprint than those lower Poaceae- Senecio- (4–23%), Senecio-type (4–22%), in the sequence. Assemblages are type- Pistacia Pistacia (4–19%). Some Pinus consistent with a complex landscape (2–10%), Quercus (1–4%), Ericaceae with a mosaic of steppe, garrigue, (0–6%), Carlina/Onopordum-type maquis and local wetlands with (2–6%), Plantago lanceolata-type perennial pools, and a little distant (0–14%), Chenopodiaceae (1–3%), woodland. Soil erosion continued. Brassicaceae (0–9%), Hordeum-type There are clear signals for arable (0–9%), Sparganium-type (0–18%). agriculture and animal grazing

98 The Holocene vegetation history of the Maltese Islands

Table 3.7 (cont.). Zone Defining taxa Depth Date ranges Characteristics Interpretation (cm) cal. bc/ad (bp) for zone bases (2σ) Olea europaea, Rosaceae, Phillyrea, Tuberaria, Theligonum, Gladiolus-type, Carduus-type, Centaurea, Artemisia, Asphodelus, Ophioglossum, Botrychium, Caryophyllaceae, Cirsium-type, Ranunculus, Rumex, Apiaceae, Tamarix, Asteroideae and Potamogeton are often present in small amounts. Pseudoschizaea and VAMs are present. Sordaria-type, Cercophora-type and Debarya sp. are often present XEM-06 Poaceae- 570–520 3521–3102 bc Count sizes very variable. High These samples show a variable Cyperaceae- (5471–5052 bp) Poaceae (3–51%), Cyperaceae taphonomic imprint and count Lactuceae- (11–31%), Lactuceae (0–42%), sizes are very variable. There are Senecio-type Senecio-type (0–27%). Some Pinus indications of a mosaic of steppe, (4–5%), Quercus (0–5%), Pistacia garrigue, maquis, local wetland (0–8%), Ericaceae (0–6%), Phillyrea with standing water and distant (0–4%), Plantago lanceolata-type woodland. There are signs of (0–5%), Carlina/Onopordum-type animal grazing (0–9%), Carduus-type (0–5%), Centaurea (0–2%), Ophioglossum (0–2%), Chenopodiaceae (0–13%), Cistus (0–2%), Ranunculus (0–6%), Apiaceae (0–3%), Persicaria maculosa- type (0–5%). Pseudoschizaea, VAMs, Sordaria-type, Cercophora-type and Debarya sp. are sometimes present XEM-07 Chenopodiaceae- 520–500 3048–2686 bc Extremely low pollen counts. These assemblages have very low Cyperaceae- (4998–4636 bp) High Chenopodiaceae (26–34%), pollen counts, but do not show Caryophyllaceae Cyperaceae (12–75%) and a strong taphonomic imprint. Caryophyllaceae (5–17%). Some They most probably derived from Persicaria maculosa-type and Poaceae a very local pollen catchment. (0–35%) The indications are of a wetland environment with standing water and perhaps nearby saltmarsh XEM-08 Poaceae- 500–470 2824–2503 bc Some low pollen counts. Very high Rather small pollen counts Asteroideae- (4774–4453 bp) Poaceae (34–54%). High Asteroideae constrain interpretation, but Ranunculus (5–31%) and Ranunculus (4–8%). there appears to be fairly strong Some Ericaceae (0–6%), Rubiaceae ecological coherence, suggesting (0–4%), Caryophyllaceae (2–5%), that taphonomic alteration Chenopodiaceae (0–8%), Rumex was limited. There is a strong (1–5%), Pteropsida (monolete) successional pattern, suggesting (1–10%), Equisetum, Littorella-type, marshland taxa giving way to Potamogeton and Isoetes aquatics in the local environment. It is likely that some of the Poaceae pollen reflects reeds rather than grasses. Away from the wetland was steppe, maquis and perhaps a little distant forest XEM-09 Lactuceae- 470–370 2401–2092 bc Some low pollen counts in this zone. The pollen catchment for this Senecio-type- (4351–4042 bp) High Lactuceae (11–64%) and Senecio- zone appears larger than for Pinus- Poaceae- type (8–38%). Some Cyperaceae the previous zones. An open, Cyperaceae (1–22%), Pinus (1–10%), Carlina/ strongly eroding landscape with Onopordum-type (0–11%), Poaceae rather degraded steppe, scrub (1–8%). Ericaceae, Pistacia, Daphne, and a little distant woodland is Tuberaria, Carduus-type, Asphodelus, indicated, with the site likely Ophioglossum, Chenopodiaceae, a wetland containing shallow Caryophyllaceae and Persicaria pools. There are signs of some maculosa-type are generally present grazing animals and occasional and Hordeum-type is occasionally cereal cultivation

99 Chapter 3

Table 3.7 (cont.). Zone Defining taxa Depth Date ranges Characteristics Interpretation (cm) cal. bc/ad (bp) for zone bases (2σ) present. Pseudoschizaea and VAMs are common, Debarya sp. is often present and Sordaria-type, Tripterospora-type and Valsaria-type are sometimes present XEM-10 Senecio-type- 370–290 736–321 bc Some very low pollen counts. Counts are very low and there Lactuceae- (2686–2271 bp) High Senecio-type (20–34%), appears to be a reasonably Poaceae- Lactuceae (10–36%), Poaceae strong taphonomic imprint in Cyperaceae- (5–23%), Cyperaceae (5–17%), Pinus this zone, with high percentages Pinus- (7–14%). Ericaceae, Pistacia, Plantago of degradation-resistant grains. lanceolata-type, Carlina/Onopordum- Nevertheless, a vegetation mosaic type, Asphodelus, Ophioglossum, with degraded steppe, some scrub Botrychium, Convolvulus arvensis- and distant woodland is suggested. type and Persicaria maculosa-type The wetland suggested for previous are sometimes present. zones appears to have been drier, Pseudoschizaea, VAMs, Tripterospora- without permanent pools. There type and Sordaria-type are were disturbed soils and some soil occasionally present erosion took place XEM-11 Pinus-Lactuceae- 290–215 ad 141–570 Some low pollen counts in the The extremely high percentages Poaceae (1809–1380 bp) lower half of this zone. Very high of Pinus pollen are consistent with Pinus (47–86%). Some Lactuceae, nearby pine woodland: in this Late Poaceae, Cyperaceae, Carlina/ Holocene Maltese context, this most Onopordum-type. Persicaria likely reflects a pine plantation maculosa-type is often present. rather than natural vegetation. Other VAMs, some Pseudoschizaea and vegetation was likely a mosaic of occasional Podospora-type are degraded steppe with a little scrub present with eroding soils. It is likely that the depositional site was seasonally dry XEM-12 Lactuceae- 215–150 ad 1092–1509 Very high Lactuceae (5–71%) A very variable taphonomic signal Cyperaceae (858–441 bp) and Cyperaceae (5–78%). Some is present in this zone, with a major Chenopodiaceae (2–12%), Poaceae influx of degradation-resistant (3–9%), Pistacia (1–9%), Senecio- Lactuceae grains in the upper part. type (2–8%), Pinus (2–6%). A little The assemblages are otherwise Ericaceae, Daphne, Phillyrea, Carlina/ consistent with a mosaic landscape Onopordum-type, Carduus-type, with grassy and degraded steppe, Ambrosia-type, Ophioglossum, scrub and distant woodland. There Botrychium, Caryophyllaceae, is evidence for eroding soils and for Persicaria maculosa-type, Filipendula a local wetland environment with and Typha in some samples. Some near-permanent pools supporting VAMs, Pseudoschizaea, and a little emergent aquatics by the end of the Debarya sp. in one sample zone XEM-13 Lactuceae-Pinus- 150–130 ad 1845–1963 Abundant Lactuceae (16–42%), Counts are extremely low and the Cyperaceae (105–13 bp) Pinus (15–26%), Cyperaceae (11– proportion of degradation-resistant 16%). Some Carlina/Onopordum-type grains is high, suggesting a marked (3–16%) and Carduus-type (2–4%). taphonomic imprint. The high Pinus Tuberaria, Plantago lanceolata- is likely consistent with the presence type, Ophioglossum, Botrychium, of pine woodland, most probably Chenopodiaceae, Senecio-type and a plantation, though further Poaceae are sometimes present. away or less sizeable than that Pseudoschizaea and Debarya sp. are suggested for XEM-11. The other occasionally present taxa are suggestive of a degraded steppeland, some grassland and some scrub or garrigue. Soils were eroding and shallow pools were present on the deposition site Natural The uppermost 1.3 m of this core sedimentation appears to be made ground of appears to have twentieth century age. ceased during the twentieth century 100

The Holocene vegetation history of the Maltese Islands Zonation IN-04 IN-03 IN-02 IN-01

100 200 300 400

Total Land Pollen & Spores & Pollen Land Total

)

e

p

y

t -

m 20

Note differing x-axis scales for some taxa

u

i

n

Damaged grains Damaged )

r

e

A

(

p

y

1

t

6

- 20

2 a

i -

r

V

a

d

d

r H

o

S

(

a

5 20

5

-

V

d H

20

a

e

a

VAMs

z

i

e

h

p c

y

s

t

o

-

e d a

s

p u

o y

e

l 5 10 t

s

-

u

P

e a c

p

h a

t

y

t

n m

-

e

s 5 a

i

i r

M r

c

a

a

c

i

s

s

u r e

l 5 e

p

u

P

y

c

t

n -

u

m

n

u 5 a

e

R

d

r

o

H

. 5 f

c

o

g

a s

s

i

h t

u 5 i t

t i

V

p

g

y

l a

a

m

c 5

m

u E

e

t

s

o

Disturbed weeds Cult. Cultivars Waterside palynomorphs Non-pollen r

5 g

A

a

l

l

e

g a

i 5 i

b

N

r

o

h

p 5

u E

5 10

Brassicaceae

a

c

i t

r 5

U

.

f

e f

i

e a

d 5 e p

n c

y

Poaceae

t

a u

-

r

s

e i e

l

a p

a

Catholic e

y 5

n

i d

C i

c

i

o

f

r f

e

o

t

s 5 a

n A

a

i

r

e

l

.

f 5 a

e f

i a V

d e

n c

a u

i

p e 5

a A

e e

c p

a

y

t b

- 5 a

m F

u

i

l

o

f

i 5 r

T

s

u

t 5

o

x L

i

.

f r

f

a i

d

m

n 5

a

u T

x

e

m 5

u

R e

p

y

t

-

o i

c

e

n 5 10 e S Dry Catholic Dry

20 40 60 80

e

p

y

t

-

Lactuceae undiff. Lactuceae

m

u i

s

r

i 5

a

C l

m

a

u

g

y

m

l

e

e 5 o

a

h

P

t e

c n

a

a

l

i

l l

e

y

e 5

a

h

H

e

p c

o a

i

y

r d

a 5

o

C

p

o

Steppe n

s

e

u

l

h 5

e

C

d

o

h

.

f p

f 5

s

a i

i

A

s d

i

e n

m p

u

e

y

t

a

t 5 r -

e

r

A

m

u

e u

a

t p

d

r

y

n

t

o 5 e

-

p C

a

t o

a n

l

O

o

/

e

a s

c

i

n r

n

i

l e

a

r

l

n

a 5 10

e

o

C

v

g

-

a s

t

u

l n

l

i

a 5

Rocky places Rocky l p

P

a

c

m

u

t

n

a

m i

u 5 10 d

n A

o

g

i e l

p

e 5 y

h

t

T -

a

i

r

e

Maquis a p

n

y 5

i

t

L

-

a

e

m

a

u

p

d 5

o e

r

S

u

e

a

e

l 5

O

a

i

e

c

p

a

y

t

t

s 5

- i

s

i P

l

i

g

a

r

f 5

.

f

Rosaceae undiff. Rosaceae a

f r

i

d

d

e

n

h

e

u 5

p

s p E

y

u

t

c -

r x

e

e

l

u

i 5

s

Q Pollen zonation for the pit fills at In-Nuffara (M. Farrell).

Non-native Woodland u

c

r

e

u 5

Q

s

u

n 5 10 i

P

a

y

r

t

s 5 O %

0 Depth (cm) Depth 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Figure 3.7.

101 Chapter 3 0 0 0 0 0 0 4 8

0

0 s

0

e s

0 i

0 e c

i 0 a

c 2

f

a 4

o f

n

o

y n

l

y

e r l

u

a P

e r t

a P

u a

t

a m

0 m

y

l

l

s 0

0

y l a

l 0

m 1

a

s

i 1 m r

n e

m r

a

h 0 e

T

s h g r 8 0

o T

m 8

s c

i i

t

n

d

a 0 a

e

u 6

v 0

g r

i

r r e

6 o q A

e

u d

c

t i r e e

t d -

l

v 0

i a

u a

i

t r

4 0 u

e

o S m a

4

d - q A l

y i m l

l

0 t

o S y n a

2

l

0

t e l a

t

i

n 2

a l P

n a

r m i

t r l P e

r m e

h

e n

T I

h

e

t

T i

n 0

i

t r 2

e 0

n 8 lgal I A

lgal

e

r 0

u A

t 6

a 0

e r

m 4 u

s

t

t

y

l

a s

l 0

y

a

m 4 s c

0 t y

l s m r

l s l

m 2 y

e

e

a

a t u l

c

i

h

g s 0 a n

s T

y i

m r

m

n e c 2

g d e

t

u

i

i r

i r h

u s n t

n

y T

i

n u F

c

d e P

e

i

F r i r

c t

n

n

o

e P e

e e

c

C a a

n

h h

o

p p

y y

t C

s h h

l n l

e r

a a

a s

0 0

o

g

g t

p e

r 2 2

n n

s

P l o n

u u

l

p

F

F a

a

s

g l

n P l a

u g

n

F

e

u l

c

F

i

e

c t

i l

0

l u

c

M A V

i

c

n 1

o t

e

u l

&

h l

c

t

s

l M A V

l & o a P

a

s

C l

n

l w

e a l

l l

l

w e

l c

l

o

0 P t o

e e n

g

2 c

n

a a

d t

l

e e i

n l

e

l l

d P

a

s

i o

r s

m p

l P

a

e u

r s t

i t r

n

e

i

e

e e

a t d

a t z

o

n

e p

a I n

c a

W u

l n

l

H

a a r

W e

. e

n

f

i

p

c

g

y

m

a

l

C m u

u n

. i

e S

t g

l d

a

u o h

p s

o

o C r

c p S

e y

m

c

L

. u

n

o

r t

l

g h

u a

h

s s

t r

A C

o r o r p S

e e

c c

o a

h

t

h P

a

s i

l n A

a o r

e

e r

c

e

a

C

o

h P

g e

a

a i l t p

a n p

a

e e r

e

l P

e e a C

S t p

i r

p y

t

e

p

s

i

e l

l

e e

m r A p p

S t y y e B

t

t

s

s

i

l n e

l

e p

y

d

t

e B

a i B

i

m

s

q i u

i

a m

s r e

i M

C

e

e

t r A

e

a

a

a

e

e

e

c

c

c

a

a

l

a l

y e o P

o P h a

p e

c o

u

t y r

c a

a

a C

i

L

e

s

i p

y

m t

m

a

e

u

c

m

d s

e

s

u

t r A n a

i r E o

l

e d

r a

g c l

o

o

u p d i

t

o h

c o

n p

a

o a

L d O

O i

s /

s

d a

W p

n

o i

o r

l

o r e a

C t

W P

s s

i r i r

e e

t P t P

m

m

u

u

i

i

d

d

m

i r

i r u

e e s

s

t P t P

o

l

g 7

a

7 e

o

a

i 1

y

1 r t e

h 0

5 s c

0 p

2 a

O

O 2 5 s

t

2

t s

n

n a e

u r a

y r

u p a t

u with the presence of a taxon shown by dot (C.O. Hunt). Pollen and palynofacies from the buried soils below temple at Santa Verna,

s H

H C O r n r n

O

O

e

s e s

u

)

u C

) V

n C

V n

m

m

i d P

i P

c d (

c

( e a

e h a t

h

s t

t

s t

y p p n

l y e

n l e

a D 5 0 5 0 5 0 5 D 5 a 5 0 5 0 5 0 n 3 3 2 2 1 1 0 9 n 2 2 1 1 0 0 S a A 1 1 1 1 1 1 1 1 1 1 1 1 1 S a A Figure 3.8.

102 The Holocene vegetation history of the Maltese Islands

Table 3.8. The pollen zonation of the fill of a Bronze Age silo at In-Nuffara, Gozo. Zone Contexts Defining Depths Period Characteristics Interpretation taxa (cm) IN-01 46, 41 Lactuceae- 138–125 Borġ Dominated by Lactuceae (62–70%). Also Assemblages highly biased Senecio in-Nadur present are Pinus (2–5%), Tamarix (1–3%), taphonomically and showing Olea (1–2%), Rosaceae (0.5–4%), Rumex signs of the recycling of (0.5–1.5%), Euphorbia (0.5–1%), Plantago material derived from soil lanceolata-type (1–2%), Carlina/Onopordum- profiles through erosion. type (3–6%), Cirsium-type (1%), Centaurea Landscape predominantly (0.5%–1), Senecio-type (6–8%), Poaceae open steppe, with some (1–3%), Asteroideae (2–4%), Asphodelus scrub. Considerable evidence (1%), Brassicaceae (1–2%). VAMs (2–5%) for grazing and consequent and Pseudoschizaea (2%) present in small vegetation degradation in the amounts wider landscape IN-02 41–40 Lactuceae- 125–72 Punic Dominated by Lactuceae (70–80%. Also Assemblages highly biased, Carlina/ present are Pinus (3–7%), Plantago lanceolata- with erosional recycling of Onopordum- type (0.5–1%), Carlina/Onopordum-type material from soil profiles. type-Senecio (5–8%), Senecio-type (4–9%), Asteroideae Landscape mostly very open (1–4%). VAMs present (4–5%) steppe, probably heavily degraded by grazing IN-03 39–37 Lactuceae- 72–42 Roman– Dominated by Lactuceae (74–79%). Also Assemblages highly biased Carlina/ Knights present are Pinus (2–6%), Tamarix (1–2%), and with recycling of Onopordum- Caryophyllaceae (1–3%), Plantago lanceolata- material from soil profiles. type type (0.5–1%), Carlina/Onopordum-type Landscape rather open (6–8%), Senecio-type (2–6%), Asteroideae grassy grazed steppe with (0.5–4%), Poaceae (0.5–2%), Brassicaceae some scrub. Minor cereal 1–2%). VAMs present in all samples cultivation. Fungal spores (2–16%). First appearance of Hordeumtype provide evidence for grazing and spores of Arnium-type and adjacent to the deposition Sordaria-type site IN-04 36–34 Lactuceae- 42–0 nineteenth– Dominated by Lactuceae (71–76%). Also Highly biased assemblages Brassicaceae- twentieth present are Pinus (4–5%), Olea europaea with evidence for recycling Pinus centuries (1–2%), Tamarix (1–2.5%), Chenopodiaceae of material from soil profiles. ad (0.5–1%), Plantago lanceolata-type (0.5%), Landscape rather open Carlina/Onopordum-type (2–4%), Senecio- increasingly degraded steppe type (1–5%), Asteroideae (2%), Brassicaceae with some scrub. Possible (4–7%). VAMs (1–38%), Arnium-type increase in grazing pressure (2–19%), Sordaria-type (2–18%) present. First at top of zone appearance of Eucalyptus

Table 3.9. Summary of the pollen analyses of the buried soil below the Santa Verna temple structure. Contexts Defining taxa Depths Period Characteristics Interpretation (cm) SV2/2, Phaceros- 110–140 Earlier Woodland taxa 10% (Pinus, Very small, highly taphonomically 2/3, 5/1, Anthroceros- Neolithic Cupressaceae, Ostrya, Pteris, Pteropsida, influenced assemblages. An open 5/2 Lactuceae pre-5500 Ophioglossum, Pteridium aquilinum), steppic landscape with some cal. bc (pre- maquis taxa 1% (Erica), steppe taxa 15% woodland patches, grazed land and 7500 cal. bp) (Carlina/Onopordum-type, Lactuceae, cereal cultivation. The abundant Artemisia, Caryophyllaceae, Poaceae, waterside plants and Pseudoschizaea Cirsium-type, Bellis-type, Bidens-type, are consistent with damp, sheltered Armeria, Plantago), cultivated taxa 9% soils close to water. The planktonic (cereals), waterside taxa 61% (Phaceros, Peridinium cysts suggest the Anthroceros, Sphagnum, Lycopodium cf. presence of mildly eutrophic fresh annotinum, Selaginella, Huperzia selago, water Cyperaceae, 24% Peridinium cysts, 8% Pseudoschizaea SV2/1 Pinus- 100 Early Presence of Pinus, Lactuceae, Phaceros Extremely small assemblage, Lactuceae- Neolithic and Sphagnum in which all recorded taxa are Phaceros- 5500–3800 known to be degradation-resistant. Sphagnum cal. bc Suggestive perhaps of the presence (7500–5800 of woodland, steppe and damp cal. bp) habitats

103 Chapter 3 0 0 0

8

s

e

i 0

c 0

a

f

e r 0

o

u

t 4 n

a y s

l

m

m

s

a P

y

i l

l

n

a a

g r m r 0

o e

0

h c

i 1 T t

a

u

d 0

e q A

v 8

i r

e

d -

l 0 i

6

o S

l 0

a

4

m i

n A

0 t

n 2

e a l

l P u r

e

h

p S 5 r e 2 u t 0 a 2 m 5

1

y

l

l e

t 0

a i

n 1

i

t r

r m

e 5

e n I h T 5 7 0 7 5 6 0 6 5 5 0 5 5 4 0 4 t e 5 d 3 n I 0 s e 3 d m 5 s i s ' i 2

n

W

e r 0

a

. u

m t 2

t

l u

r g a

u s 5

o m

n

n 1

C

0

e y

l

l c e

8 l l l i

l t a 0 o

o 0

s p

a 1

e 4

P

m r

d .

p u t

9

e

s

e y

e t s 1

t q

h 5

s e d u

1

T

e A y a c

n

w I

c

e p

c h

t

p y

o

e t

n

y m

l c

e

u s T o

b i

y

M e n

y

a r

i

h

r t a

d

c

e

A o a

a

o B

t c

t

i

s l

p a

t

S

i

a

e s

a r d

e

y

r

y

o

c

a S

g e

p

o C

m

o r

i

n u

i

e

n s

p S

i

h

e u

d

r

C

p e r

i r

y

y

o

t

h

p

t

a

e P

l

e s

s

l a

i

t

a o e L

i

r e o

z e c

c r A l d

a d i

a c

i

S e

g

s

v n

s

u 5

a r i

r

F

x e

B e

e - d

m l p 0

i u

y 3 t

R e e

s

p p i

S o

5 l

y

l

p t m

e 2

e

p e u

s

t B

y

n 0 t

m

S

e

e

e 2

d

m h

i

a

t

u h 5

B

n

d

p

a

r 1

i

y

a l

o

i

h

e

s

p

s 0 l

i

o

H a e

r 1

m n

g

o

e

n

p O

t

/

r

u s e 5

a l

p F A

n

a y

i

t

l

g

r

n

s

a

u

u 5

C F u

d l r

m

a

u a i

e 0

C

p

n 3

m ) y

a i t

r

e

e n 5

a a

i

d r G 2

A

e

i r

e

a

u 0

e e

e

m

h

r

c

p

a 2

c

a y

A e

i r l t

l

c 5

y

T ( a a

h 1

s o

s p

o

P

g i

o 0

g b

b e

y

e 1 r a

a

u

a c

e

a

s

c

c r C i t r

5

u

a a

M A V

t n S l

l

e

i

c a

s / r p a

s

u

a

d L

a P l

e r

t

o C

o

t n

o m

c e

u

e i p S

C

W s s d

a

i 5

&

n

u r

I

d

i

n

e

n

t

s

u

e

l

r

p P

4 l

o

P

r 5

.

o

1 P

e f

t

f

a

a

r 0 P

e

e 0

2 d

a

e 3

e l

e

c

H c

7 i

t 5

a

1 u

s 2

e

r m c 0 o

a

e 0 & 2 R

o

c

s 2

t f

a

l a

t l

n e s

l

a 5

u s

w a O Pollen and palynofacies from Test Pit 1 on the southwestern edge of the Ġgantija platform, with presence a taxon shown by Pit 1 on the southwestern Pollen and palynofacies from Test

e 1 l

l r

H l

p

p e

0

u

c

O t 1

C

)

s n

a )

C u

j a

m

5 m i n

c i

l

d P t (

c (

P

e

h n h

t

s t

p

y p a

l e e

0 0 D 0 0 0 0 0 0 a 0 0 0 0 g D 8 9 8 9 n 3 2 1 0 3 2 1 0 1 1 1 1 1 1 1 1 G A Figure 3.9. dot (C.O. Hunt).

104 The Holocene vegetation history of the Maltese Islands

Table 3.10. Summary of the pollen analyses from the buried soil in Ġgantija Test Pit 1. Defining Depths Period Characteristics Interpretation taxa (cm) Lactuceae- 110–130 Tarxien Woodland/maquis taxa 34% (Pinus, Quercus, Moderately taphonomically biased Hedera-Pinus Cupressaceae, Hedera, Pteropsida), steppe taxa assemblages, but suggestive of 43% (Lactuceae, Poaceae, Carduus-type, Carlina/ some woodland (the rather high Onopordum-type, Artemisia, Helianthemum, Bidens- Hedera could have been growing on type, Bellis-type, Rumex), cereals 5%, agricultural stonework), open steppe, arable land, weeds 9% (Brassicaceae, Sideritis-type, Lathyrus, damp habitats. Algal assemblage Chenopodiaceae), waterside taxa 5% (Tamarix, Mentha- suggestive of rather eutrophic type, Anthoceros). Algae 37% (mostly Type 119) standing water Lactuceae 85–110 Tarxien Woodland/maquis taxa 7% (Pinus, Olea europaea, Strongly taphonomically biased Rosaceae aff.Prunus , Pteropsida, Pteridium aquilinum), assemblages dominated by Lactuceae. steppe taxa 85% (Centaurea scabiosa-type, Lactuceae, Suggestive of a little woodland/scrub, Caryophyllaceae, Poaceae, Armeria-type, Geranium, very open, dry, somewhat degraded Carduus-type, Carlina/Onopordum-type, Artemisia, steppe, minimal damp habitats and Bidens-type), cereals 1%, agricultural weeds 1%, somewhat eutrophic standing water waterside taxa 1%. Algae 64% (mostly Type 119 and Peridinium sp.)

Table 3.11. Activity on Temple sites and high cereal pollen in adjacent cores. Pollen data from FRAGSUS Project, Carroll et al. (2012), Djamali et al. (2013) and Gambin et al. (2016) and periods of human activity (C. Malone, pers. comm.). Grey shading indicates periods not sampled for palynology in coring or during archaeological excavation. Archaeologically attested or inferred periods of activity on megalithic sites are indicated with an ‘a’. Anomalously high cereal pollen values are indicated with an ‘h’. et al . Tas-Sil ġ temple Tas-Sil ġ middens Bor ġ in-Nadur temple Wied Ż embaq core Tarxien/Ħal Saflieni temple Marsa core (Carroll 2012) Buġibba temple Salina Bay (Carroll et al . 2012) Tal Qadi temple Burmarrad BM1 and BM2 cores (Djamali et al . 2013; Gambin et al . 2016) Xemxija temple Xemxija core Ġgantija temple Ġgantija soils Santa Verna temple Santa Verna soils

Later h a h h Bronze Age Tarxien a Cemetery Thermi a Ware Tarxien a h a ? h a h Saflieni a a a ? a a Ġgantija a a h a h ? h a a h a a Mgarr ? a a h ? h ? h a a Żebbuġ ? a h ? h ? a Earlier h a a h Neolithic

The algal microfossils suggest that the sites were in temple are summarized in Tables 3.10 and 3.11 and receipt of slightly eutrophic standing or slow-moving Figure 3.9. Pollen counts are very low, but peaks in water or sediments derived therefrom. Representative pollen occurrence and the coincident peak in vesicular material is illustrated in Figure 3.10. arbuscular miccorhyzae (VAM) suggests episodes of soil stability around 115 and 85 cm. The combination 3.5.6. Ġgantija of very abundant thermally mature (charred) material, Pollen analyses of the buried soil as exposed in Test plant cell walls and cuticle and fungal hyphae is typical Pit 1, located immediately to the southwest of Ġgantija of middens, while the high inertinite in the lower part 105 Chapter 3

a b c d

e f g h

i

j

k

l

m

Figure 3.10. Photomicrographs (×800) of key components of the palynofacies at Santa Verna and Ġgantija: a) Spherule, Ġgantija 80–85 cm; b) Capillaria egg, Ġgantija 95–100 cm; c) fungal spore, Santa Verna 2, 105–115 cm; d) fungal zoospore, Ġgantija 95–100 cm; e) thermally mature (charred) material, Santa Verna 2, 105–115 cm; f) Concentricystes sp., Santa Verna 2, 105–115 cm; g) Type 119, Ġgantija 80–85 cm; h) Spirogyra sp. aplanospore, Ġgantija 80–85 cm; i) plant cuticle, Santa Verna 2, 105–115 cm; j) fungal hypha, Santa Verna 2, 105–115 cm; k) Saeptodinium sp. cyst showing large compound archaeopyle, Ġgantija 95–100 cm; l) Peridinium sp. cyst showing incipient compound archaeopyle, Santa Verna 55, 120–140 cm; m) vesicular arbuscular miccorhyza, Santa Verna 2, 105–115 cm (C.O. Hunt). 106 The Holocene vegetation history of the Maltese Islands

of the section is very typical of seasonally wet biologi- sedimentary discontinuities associated with erosion cally active soils. The nematode Capillaria gives rise to episodes. The latter point may indicate that the mod- a hepatitis-like zoonotic disease causing liver failure eled ages could be unreliable, but the progressively in humans, but it can also affect a variety of animal consistent spread of dates over the last c. 8000 years hosts. The algal microfossils suggest that the site was suggests otherwise (see Chapter 2). in receipt of fairly eutrophic slow moving or standing Some time after about 8000 cal. bc, shallow marine water or sediments derived therefrom. Representative sedimentation became established at the Salina Deep material is illustrated in Figure 3.10. Core site (Table 2.4). Thereafter, continuing sedimen- Two phases of sedimentation are evident, both tation was off-set by rising sea level, and it may be dating to the Tarxien phase. The earlier assemblages suggested that sediment accumulation was essentially (130–110 cm) suggest some soil erosion and a land- continuous. The reducing environment offered by scape with woodland, open steppe and damp habitats shallow marine sediments led to preservation of suf- close to water. A rather eutrophic water body nearby ficient pollen to recognize the biostratigraphy shown is suggested by the algal microfossil assemblage. The in Figure 3.3 and Table 3.4. As with the modern marine later assemblages (110–85 cm) show rather strong soil samples from the Mistra catchment discussed above, erosion and a drier landscape with a little woodland, the assemblages show taphonomic biases but these open steppe and less evident damp habitats. The water can be accounted for in the interpretation (Table 3.4 body had by this time become slightly less eutrophic. and below). These are the oldest reliable data available from which to reconstruct Maltese Holocene vegetation 3.6. Synthesis history. Previously the earliest Holocene palynological data from Malta were from the BM1 sequence from 3.6.1. Pre-agricultural landscapes (pre-5900 cal. bc) Burmarrad, with the oldest sediments there dated to The basal sections of the deeper cores are characterized c. 5350 cal. bc (7300 cal. bp) (Marriner et al. 2012; Djamali by extremely sparse pollen assemblages, which show et al. 2013; Gambin et al. 2016). evidence of strong taphonomic bias. The Salina Deep It is well documented that attempting to dis- Core has a basal section which predates 7042–6553 entangle natural climatically forced events from cal. bc (8992–8503 cal. bp), consisting of poorly sorted anthropogenic impacts in the palynological record is gravels interbedded with incipient palaeosol mate- extremely difficult (e.g. Caló et al. 2012; Carroll et al. rial, some of which contain pedogenic carbonates. 2012; Djamali et al. 2013; Gambin et al. 2016; Chapman These deposits relate to terrestrial fluvial and collu- 2018). Prior to the arrival of farming people in the vial environments, most probably present during the Maltese Islands, any changes in vegetation would have Late Pleistocene and perhaps the Early Holocene, by most likely resulted from environmental processes. The comparison with similar deposits described by Hunt basal pollen zones in the Salina Deep Core (SDC-02 to (1997). The very sparse pollen survival in these deposits 05) suggest the cyclic spread of Pistacia (lentisk) scrub may be consistent with dry terrestrial environments into steppe landscapes in Salina Deep Core zone SDC- supporting steppe and scrub vegetation, as might be 02 at 7042–6553 cal. bc (8992–8503 cal. bp) and zone expected during interstadial episodes (Hunt 1997), SDC-04 at 6489–6135 cal. bc (8439–8085 cal. bp). The but extremely low numbers of pollen grains and poor repeated spread of Pistacia was most probably driven pollen preservation caution against reliance on these by periods of increasing effective humidity within a assemblages. generally very dry period. Pistacia is generally accepted The basal 3.4 m of the Xemxija 1 core is also rather to be under-represented in modern day pollen assem- problematic as the sediments are heavily pedogenically blages (Wright et al. 1967; Collins et al. 2012; Djamali altered. The pollen assemblages contained within are et al. 2013), and the percentages found in these zones extremely small and appear to have been significantly indicate dense lentisk-dominated vegetation. Djamali altered by strongly oxidizing and biologically active et al. (2013) review the edaphic conditions required for terrestrial environments. Indeed, the micromorpho- Pistacia species – an expansion in this taxon may be logical analysis of this core suggests that at least the indicative of high moisture availability either through basal c. 1.2 m of sediments are probably eroded soils humidity or increased precipitation (Tinner et al. from the immediate catchment, which are similar to the 2009; Carroll et al. 2012; Djamali et al. 2013; Gambin soils observed beneath several of the Neolithic temple et al. 2016). The very short duration of these episodes sites and would have undergone some pedogenesis seems to have precluded significant expansion of other once re-deposited in the base of the valley bottom (see scrub or woodland taxa, but the substantial decline in Chapter 5). The abrupt changes in pollen assemblages Pseudoschizaea and vesicular arbuscular mycorrhizae throughout this core may also suggest that there are (VAMs) in the Salina Deep Core zones SDC-02 and 107 Chapter 3

SDC-04 also point to the dense vegetation canopy chronologically with the widely reported 8.2 ka bp reducing soil erosion. Equivalent chronologically to the event which led to cooling in the North Atlantic and second Pistacia peak, a humid episode among general aridification in the Mediterranean coastlands (e.g. aridity is reported in Sicily and in southeastern Spain Barber et al. 1999; Weninger et al. 2006; Bini et al. 2018; at c. 6450 cal. bc (8400 cal. bp) (e.g. Reed et al. 2001; Chapman 2018), although the second part of this zone Tinner et al. 2009). shows a small increase in Quercus (oak) pollen, sug- Intervening zones in the Salina Deep Core SDC-03 gesting perhaps the end of climate severity. at 6858–6419 cal. bc (8808–8369 cal. bp) and SDC-05 at 6350–6037 cal. bc (8300–7987 cal. bp) show an increase 3.6.2. First agricultural colonization (5900–5400 cal. bc) in herbs including grasses, Chenopodiaceae, Aspho- A key point in the Salina Deep Core is at 22.65 m, delus and sporadic occurrences of Plantago lanceolata, where pollen zone SDC-06, commencing at 6067–5821 Urtica, Brassicaceae and Convolvulus arvensis-type. cal. bc (8017–7771 cal. bp), is marked by assemblages While some species of Asphodelus can be favoured by containing pollen of cereals. These include both Avena/ nutrient-deficient soils resulting from over-grazing, it is Triticum type (oats/wheat) and Hordeum-type (barley) also known to colonize rapidly newly exposed ground and a ruderal flora includingConvolvulus, Borago, Scro- surfaces following fires (Pantis & Margaris 1988; Abel- phularia and Sideritis, all of which are often associated Schaad & López-Sáez 2013). Fungal ascospores are with arable farming (Fig. 3.3). It is very likely that arable also present, with Podospora-type (Dietre et al. 2012) agriculture was established in the vicinity of the pres- and Tripterospora-type considered to be coprophil- ent-day Salina salt pans, since cereal pollen is produced ous, growing on the dung of large herbivores, while in low quantities and is not well-dispersed (Edwards Diporotheca rhizophila is associated with nitrogen-rich & McIntosh 1988). Pollen of Plantago lanceolata-type environments (van Geel et al. 2003). Studies of mod- (ribwort plantain) and Urtica (nettle) may indicate ern palynological assemblages have shown that these grazing. The presence of a range of fungal ascospores, ascospores are not dispersed over long distances including the coprophilous taxa Podospora-type, Tripter- (Blackford & Innes 2006) and therefore they can be ospora-type, Sordaria-type and Arnium imitans-type and interpreted as indicative of grazing animals close the nitrophilous Diporotheca rhizophila-type is strongly to the deposition site, whether these were naturally suggestive of the presence of grazing animals locally. present or introduced by people. It has been argued The assemblages also show very high incidences of that navigation in the central Mediterranean Sea may Pseudoschizaea, which is of likely fungal origin (Mila- have occurred as early as c. 8000 cal. bc (9950 cal. bp) nesi et al. 2006) and associated with damp calcareous (Broodbank 2013) and that temporary exploration and soils. Large numbers of this palynomorph deposited occupation of Malta might have occurred prior to the in a marine environment are likely to reflect enhanced establishment of permanent settlements (Gambinet al. soil erosion. All these indicators suggest that the base 2016). No archaeological evidence has been identified of this zone may reflect the initial colonization of this from this period, however, and whilst the present day area by people using cereals and domestic animals, mammalian fauna of Malta is relatively impoverished, and thus also reflect the beginning of the Neolithic comprising only around 21 species (Schembri 1993), it in the Maltese Islands, suggesting that it occurred at is possible that a dwarfed form of red deer was present 6067–5821 cal. bc (8017–7921 cal. bp). Such an early in the earlier Holocene (Hunt & Schembri 1999). The date was not expected on archaeological grounds, presence of coprophilous fungal spores could therefore but is broadly consistent with other first Neolithic represent either native herbivores or (less likely) wild dates in the central and western Mediterranean (e.g. species introduced by early pioneer settlers. Regression de Vareilles et al. 2020; Zeder 2008; Zilhao 2001), and of scrub and woodland was accompanied by rising strengthens the hypothesis for a seaborne Neolithic Pseudoschizaea and VAMs, indicating increased erosion diaspora from the eastern Mediterranean in response in the catchment area (Argant et al. 2006; Ejarque et to instability caused by the 8.2 ka bp climate event (i.e. al. 2011; Estiarte et al. 2008; Gambin et al. 2016; van Weninger et al. 2006; Zeder 2008). Geel et al. 1989). This is likely to have been related to The vegetation suggested by assemblages in the decline in vegetation cover, as the exposed soil Salina Deep Core zone SDC-06 otherwise comprises surface would have been more vulnerable to removal a very grassy steppe with some maquis and garrigue via surface run-off. and very minor woodland, suggesting a relatively The expansion of steppe habitats in Salina Deep dry seasonal Mediterranean climate. The rise of the Core zones SDC-03 and SDC-05 may be consistent drought-resistant but frost-sensitive Phillyrea (false with periods of general precipitation decline in Malta. privet) might suggest strengthening summer droughts The second of these events (in zone SDC-05) coincides and relatively warm winters at this time. 108 The Holocene vegetation history of the Maltese Islands

3.6.3. Early Neolithic (5400–3900 cal. bc) it has been suggested that an increase in this taxon Plantago lanceolata-type (ribwort plantain, very often seen in the Burmarrad sequence at c. 4650 cal. bc (6600 associated with grazed pasture) rises strongly in the cal. bp) may reflect the establishment and development Salina Deep Core zone SDC-07 at 5625–5419 cal. bc of terraces in the landscape (Djamali et al. 2013). Cli- (7575–7369 cal. bp) shortly after the date of the first matically, there appears to have been very little change known archaeological evidence for human activity through these events, which thus probably reflect cul- in the Maltese Islands (Fig. 3.4). Coincident with this, tural processes and variously successful experiments pollen of Lactuceae rises. High percentages of Lactu- with agriculture. Broadly contemporary early cereal ceae are generally accepted to be indicative of poor cultivation in a landscape characterized by steppe with pollen preservation as the pollen is relatively resistant some woodland is also evident on Gozo at this time, to decay (Havinga 1967; Mercuri et al. 2006). The pol- in the soils underlying the Santa Verna Temple (Table len assemblages from the Salina Deep Core, however, 3.9). The palynological evidence for early cereal cul- also contain many other taxa and the condition of the tivation in Malta is supported by the recovery during pollen was generally good, so this does not necessarily FRAGSUS excavations at Santa Verna and Skorba of appear to be a preservation issue. It is possible that this archaeobotanical remains of barley, wheat and lentils rise reflects the liberation of the resistant Lactuceae dating from c. 5400–4900 cal. bc (7350–6850 cal. bp). This pollen grains from the soil, caused by over-grazing period of intensive and extensive arable and pastoral of vegetation exposing the soils to erosion. Other taxa agriculture occurs during the Għar Dalam and Skorba resistant to corrosion, oxidation and microbial attack cultural phases. – including Pinus and fern spores – do not, however, From about 7000 years ago (5050 cal. bc) there may rise at this time. Alternatively, Florenzano et al. (2015) have been significant changes in effective moisture. It have argued that an abundance of Lactuceae pollen seems highly probable that, at Burmarrad and in other in Mediterranean contexts is related to intensively localities where land-use was not intensive in Malta, as grazed pasture, and given the abundance of other in Sicily and southern Spain (Tinner et al. 2009; Reed et pastoral indicators, this is a plausible explanation for al. 2001), Pistacia and other vegetation was responding some of the high percentages seen in the Salina Deep to an episode of relatively high effective moisture, Core and at other sites. Cereal pollen is low in this with widespread major increases in precipitation at period, suggesting either low intensity cultivation or this time. The presence of Pistacia pollen does seem, cultivation relatively distant from the core site. however, to have been extremely localized, for instance In Salina Deep Core pollen zone SDC-08 at 5404– in the BM1 core it was rising strongly by 5050 cal. bc 5145 cal. bc (7354–7095 cal. bp) P. lanceolata-type rises (7000 cal. bp), while in the nearby BM2 core it was still further and the nitrophilous Urtica (nettles), Rumex a relatively minor component of assemblages at that (docks), ruderals including Brassicaceae and Convol- time (Fig. 11.1) (Djamali et al. 2013; Gambin et al. 2016). vulus arvensis-type and rising coprophilous fungal Although high tree pollen is present, a differ- spores (with Sporormiella-type now present in addition ent pattern is visible in the Marsa 1 core, where the to those taxa discussed earlier) are consistent with basal zone, with an interpolated age of c. 5050 to 4850 major expansions of pastoral agriculture close to this cal. bc (7000 to 6800 cal. bp) is characterized by pollen site. Relatively high P. lanceolata-type and Lactuceae suggestive of scrub dominated by Pinus (pine) and are visible in the basal zone of the Salina 4 core (Fig. Juniperus or Tetraclinis (juniper or sandarac gum), with 3.4), but are not present in the Burmarrad BM1 and very little Pistacia. There is also pollen indicative of a BM2 cores (Djamali 2014; Djamali et al. 2013; Gambin little steppe and cereal agriculture and there is much et al. 2016), or at Xemxija (Fig. 3.6) (where, however, microscopic charred wood, suggestive of significant pollen preservation is poor in sediments of this age). and repeated fire (Carroll et al. 2012). Thereafter the The signal is probably localized, suggesting that areas landscape at Marsa appears to have become signif- of intensive grazing were restricted in size at this time. icantly more open and shortly after c. 4550 cal. bc Cereal pollen shows a short-lived peak in the two (6500 cal. bp) cereal and P. lanceolata-type pollen had FRAGSUS Salina cores and at Burmarrad, synchronous become more common, suggesting an expansion of with the rise in P. lanceolata-type, suggesting that arable arable and pastoral agriculture. This most probably farming became more extensive, for a while (Fig. 11.2). points toward patches of pine-juniper/sandarac gum A slight increase in Theligonum pollen accompa- scrub woodland being cleared. These findings may nies other evidence for increased agricultural activity, point to development of a different flora on the dry and the taxon is consistently present throughout the Globigerina Limestone plateau in southern Malta, in remainder of the Salina Deep Core sequence (Fig. 3.3). contrast to the lentisk-dominated vegetation associated Theligonum is found in dry, rocky environments and with the better water-retaining properties and hill-side 109 Chapter 3

springs from the perched Upper Coralline Limestone reflect social and/or cultural rather than climatic factors. aquifer associated with the Blue Clay lowlands and Archaeological excavations by the FRAGSUS Project at muddy colluvial cover of hillsides in northern Malta Skorba and Santa Verna have identified a hiatus in the (Hunt 1997). cultural record between c. 4800–3800 cal. bc (6750–5750 Areas of intense arable agriculture and grazing cal. bp), which corresponds broadly with the reduced seem to have shifted during the interval between agricultural activity in the palynological record (see c. 5050 and 4550 cal. bc (7000–6500 cal. bp), whether in Volume 2). The continuation of cereal agriculture at response to climatic or to socio-cultural factors. In both some sites suggests, however, that there might have the Salina Deep and Salina 4 cores, P. lanceolata-type been some sort of re-organization of the locations and cereals reach minima and peaks of Pistacia sug- preferred by the population and/or changes in social gest patches of lentisk scrub around 4550 cal. bc (6500 structures and economic activities at this time, rather cal. bp) (Figs. 3.3 & 3.4). The patchiness of vegetation at than a substantial de-population of the islands. This Salina is indicated by the contrasting record from the view is supported by the very small number of exca- Salina Bay core of Carroll et al. (2012), which shows vated Neolithic sites. an opposite trend, with a maximum of Plantago and very little Pistacia or other woodland or maquis taxa 3.6.4. The later Neolithic Temple period at the same time. In the Burmarrad BM1 pollen dia- (3900–2350 cal. bc) gram, P. lanceolata-type peaks at c. 4900 cal. bc (6850 After 6000 years ago (4050 cal. bc), Maltese landscapes cal. bp), then declines to a minimum at c. 4600 cal. bp generally remained extremely open, with cereal cultiva- (6550 cal. bp), while Pistacia rises strongly from c. 4950 tion increasing and grazing evident at most core sites. cal. bc (6900 cal. bp), peaking at over 60 per cent at Spores of soil fungi suggest continued soil erosion, c. 4750 cal. bc (6700 cal. bp) before declining to c. 45 per while rising values of corrosion-resistant pollen such cent at c. 4550 cal. bc (6500 cal. bp). There are peaks of as Lactuceae and various Asteraceae may suggest that Pistacia and Phillyrea around 3850 cal. bc (5800 cal. bp) soil erosion was accelerating, perhaps because of agri- at Xemxija (Fig. 3.6). cultural intensification, or expansion of arable activity Vitis (vine) appears at c. 4750 cal. bc (6700 cal. bp) into more geomorphologically marginal areas. At the at Burmarrad (Djamali et al. 2013), perhaps reflecting same time, the percentages and diversity of pollen of the introduction of grapes. It is also possible, because agricultural weed taxa both start to increase, suggest- vines are extremely low producers of pollen, that these ing that farmers may have been losing productivity as were native plants under-represented in the pollen weeds proliferated and soils degraded and continued rain (Pagnoux et al. 2015). The Burmarrad BM2 pollen to erode. All of this is consistent with the island being diagram (Gambin et al. 2016) has an interval with no relatively intensively used for agriculture during the pollen preservation at this time. later Neolithic (see Chapter 6 and Volume 2). There is evidence for a general decline in agri- There are, however, some distinct elements of cultural activity around 4550 cal. bc (6500 cal. bp) in spatial and chronological patterning. The Salina Deep many of the Maltese records. Arboreal and scrub Core shows three episodes when Pistacia (lentisk) pollen expands at several sites, and cereal pollen scrub regenerated, between c. 4450 cal. bc (6400 cal. bp) percentages generally decline apart from at Marsa 1 and 3850 cal. bc (5800 cal. bp), between c. 3450 cal. bc where they rise sharply (Carroll et al. 2012) (Figs. 11.1 (5400 cal. bp) and c. 2700 cal. bc (4650 cal. bp) and & 11.2). The rise in arboreal and scrub pollen might c. 2400 cal. bc (4350 cal. bp) and c. 2300 cal. bc (4250 suggest that climatic wetness was rather high at this cal. bp) (Fig. 3.3). The second peak in Pistacia is also time. Although this may in part reflect a relaxation of recorded at the Burmarrad BM2 core, where it forms grazing pressure associated with agricultural decline, over 60 per cent of the pollen assemblages deposited percentages of tree and shrub pollen are higher at during the Ġgantija and early Saflieni phases (Fig. 4550 cal. bc (6500 cal. bp) than during humid episodes 11.1) (Djamali et al. 2013; Gambin et al. 2016). The first before the 8.2 ka bp event when grazing livestock were and second peaks seem to be visible at Xemxija, but most likely absent. This suggestion of possible higher records further away seem to be out of phase. This effective humidity contrasts with many areas around therefore suggests the presence of very localized stands the Mediterranean (e.g. Magny et al. 2011; Sadori et al. of lentisk scrub in an area close to the Burmarrad core 2016; Jaouadi et al. 2016; Bini et al. 2018), which show sites which was not used heavily for arable agriculture evidence of climatic disruption and rainfall decline or grazing. It has been hypothesized that Pistacia spp. around 4550 cal. bc (6500 cal. bp). It is thus possible to were managed for food, fuel and fodder in the Levant infer that in Malta, the widespread, although not uni- during the pre-pottery Neolithic (Asouti et al. 2015) versal decline in agriculture at this time may perhaps and it is conceivable that something similar occurred 110 The Holocene vegetation history of the Maltese Islands

on Malta during the later Neolithic. That management (Figs. 3.4 & 11.1; Table 2.1). Both vines and carob have was occurring might be inferred from the Burmarrad extremely poor dispersal so the very low percentages BM1 and BM2 diagrams (Djamali et al. 2013; Gambin recovered may indicate that these were actually fairly et al. 2016), because the expansion of Pistacia is not abundant in the vegetation when first recorded, and followed by substantive successional expansion of absence in the pollen diagrams prior to this cannot other maquis or woodland taxa for about 2000 years, be taken as conclusive evidence that they were not until after c. 3050 cal. bc (5000 cal. bp) when Pistacia present in the Maltese Islands. Indeed, carob charcoal declines and short-lived peaks of Olea and Quercus are was recovered from the early Neolithic site at Skorba observed. Substantiating this suggestion further would (Metcalfe 1966), suggesting that it was present in Malta require intensive investigations of macro-fossil wood from well before the first appearance of the pollen. and charcoal on Maltese archaeological sites. Carob was thought to be an eastern Mediterranean During the later Neolithic period, from the species domesticated during the Chalcolithic (Ramón- Żebbuġ phase (3900–3600 cal. bc) onward, cereal per- Laca & Mabberley 2004), although recent genetic work centages are remarkably high at some sites (Tables 2.1 suggests populations of wild carob spreading from & 3.11). In the Salina Bay core of Carroll et al. (2012), several glacial refugia such as in Morocco, Iberia, from c. 3850–2350 cal. bc (5800 to 4300 cal. bp), there are Sicily, Greece and southern Turkey during the early at times very high values (over 10 per cent) for cereal Holocene (Viruel et al. 2016). Traditionally, carob pods pollen. This especially occurred during the Ġgantija have been used as animal fodder and famine food in phase (3450–3200 cal. bc), when there are cereal pollen Mediterranean countries (Forbes 1998) and could have percentages of the same order at Wied Żembaq and been used in Maltese prehistory, wild or domesticated. one very high value at Xemxija, and again during the On the other hand, grape pips first appear at Tas- Tarxien phase (2850–2350 cal. bc), when there are very Silġ in the Early Bronze Age (Fiorentino et al. 2012), high cereal pollen percentages at Tas-Silġ (Fig. 11.2). significantly after the first appearance of the pollen. There are also some very high values in the segment Nevertheless, the appearance of pollen grains of these of the Marsa core assigned to the Żebbuġ to Ġgantija species may perhaps indicate rising numbers of the phases (3900–3200 cal. bc), but unfortunately these plants and possible widening of the subsistence base sediments are followed by a depositional hiatus. The during this time. percentages at all these sites are greater than would be expected in a landscape devoted to cereal growing 3.6.5. The late Neolithic–Early Bronze Age transition and perhaps reflect threshing on cereal-processing sites (2350–2000 cal. bc) rather than cultivation per se. Even in the middle of a During the Tarxien phase (2850–2350 cal. bc), there is field of cereals, it is rare for their pollen to form more possible evidence for aridification at Ġgantija (Table than 8 per cent of the assemblage (Faegri & Iversen 3.10), Tas-Silġ (Hunt 2015) and Wied Żembaq (Fig. 3.5), 1975), but threshing releases into the environment in the form of woodland contraction, steppe vegetation abundant cereal pollen, which would otherwise remain apparently taking on a drier aspect and diminished trapped within the husks, so sites where threshing clear evidence for areas of damp vegetation. Cereal occurred are often marked by very high cereal pollen pollen also declines markedly at this time at many values (e.g. O’Brien et al. 2005). This perhaps raises sites. Woodland decline occurs during the Tarxien the possibility of cereal processing, or the handling Period at Xemxija, Burmarrad (Gambin et al. 2016), of processed cereals, adjacent to temple sites such as Salina 4, Salina Bay (Carroll et al. 2012) and in the Tas-Silġ, Borġ in-Nadur (close to the Wied Żembaq Salina Deep Core, but trends in other indicators are core site), the Tarxien Ħal Saflieni complex (near the less clear. Nevertheless, it can be argued that effective Marsa core sites), Xemxija (adjacent to the Xemxija precipitation may have declined during the Tarxien core site), and Buġibba (close to the Salina core sites). Period (cf. Gambin et al. 2016), with the corollary that Only the Burmarrad cores, which are not far from the socio-economic systems may have come under increas- Tal-Qadi temple, do not conform to this pattern, but ing stress through declining agricultural productivity this area was not cultivated and instead appears to and perhaps the drying of some karst water sources. have been covered with dense lentisk scrub. Dispersal In the Salina 4 core (Fig. 3.4), Carroll’s (2012) of cereal pollen within this environment would likely Salina Bay core, the Wied Żembaq core (Fig. 3.5) and have been minimal. in the Tas-Silġ middens, there is a short episode around The occurrence of pollen of vines in the later 2350 to 2050 cal. bc (4300 to 4000 cal. bp) during which Neolithic Żebbuġ phase at Salina 4, and in the Ġgantija cereal pollen disappears or becomes very rare and and Tarxien phases at Wied Żembaq, and of carob in pollen of trees and shrubs increases. The only contrary the Tarxien phase at Salina 4 is possibly of significance trend is at Burmarrad BM2 (Gambin et al. 2016) where 111 Chapter 3

cereal pollen rises at this time. The behaviour of P. agricultural weeds, pasture land taxa and coprophilous lanceolata-type (which may be taken to reflect grazed fungal spores for widespread cereal cultivation and land) through this episode varies: at Salina 4 it declines pastoral agriculture. This coincides with evidence markedly and at Wied Żembaq it declines more gently, for extensive arable agriculture at this time in the whereas at Burmarrad BM2 (Gambin et al. 2016) and macrofossil record at Tas-Silġ (Fiorentino et al. 2012). Salina Bay it rises sharply (Carroll et al. 2012). This epi- The number of sites of Later Bronze Age date sampled sode is known elsewhere around the Mediterranean as during this and previous research is rather small and the culmination of a time of climatic stress (e.g. Magny there are marked taphonomic biases to assemblages et al. 2009, 2011; Jaouadi et al. 2016; Ruan et al. 2016), at Wied Żembaq, Tas-Silġ (Hunt 2015) and Xemxija. but there seems to be no strong climatic evidence from Nevertheless, the relatively uniform percentages for the Maltese pollen records at this point. tree and shrub pollen are suggestive of there being When considering all the available records, the no major climatic change at this time, while relatively decline of cereal pollen at Salina 4 and Salina Bay high cereal pollen percentages can be used to infer that (Carroll et al. 2012), Tas-Silġ and Wied Żembaq and the cereals were grown and quite probably processed at rising trend of tree and shrub pollen at Salina (Deep Tas-Silġ. Core & Salina 4), Xemxija, and Wied Żembaq close to 2350 cal. bc (4300 cal. bp) may reflect abandonment of 3.6.7. Late Bronze Age, Punic and Classical periods agricultural land and regeneration of woody vegetation (c. 1000 cal. bc to ad 1000) rather than a climatic signal per se. It is worth noting By the start of the Classical Period, pollen assemblages that in Malta during the early Holocene, only times from Marsa (Carroll et al. 2012) are marked by relatively of climatic wetness seem to have supported abundant high incidences of Caryophyllaceae, Chenopodiaceae, woody vegetation, so alternatively the rise of woody spiny Asteraceae, Arabis, Spergula-type, and rather low vegetation at this time may conceivably reflect a rise Poaceae and tree/shrub pollen. This is likely to reflect a in effective moisture. highly degraded landscape (Carroll et al. 2012). There On the other hand, archaeological evidence sug- is ample evidence for eroding soils (e.g. at Xemxija), gests that there was most probably some form of with relatively high counts for corrosion-resistant taxa, societal disruption close to 2350 cal. bc (4300 cal. bp), such as Pinus, Lactuceae and various Asteraceae, and reflected in the apparent abandonment of temple often very strong representation of soil organisms such sites and short-lived appearance of the Thermi ware, as VAMs and Pseudoschizaea. followed later after c. 2000 cal. bc by the Tarxien Cem- Tree and shrub pollen persisted relatively etery culture (e.g. Malone et al. 2009b; Fiorentino et al. unchanged from the Late Bronze Age and throughout 2012; Malone & Stoddart 2013) (see Chapters 6 & 7). the Punic period at Xemxija and Marsa (Carroll et al. The pollen evidence suggests that as part of this dis- 2012). In contrast, arboreal pollen values are higher at ruption, cereal cultivation and pastoral activity were Wied Żembaq and at Tas-Silġ, mostly because of high abandoned in coastal localities such as Salina, Xemxija percentages of Pinus, which could be taphonomic or and Wied Żembaq. The decline in pollen of woody reflecting overall lower pollen productivity locally plants and rise in cereal pollen at Burmarrad counters rather than reflecting changes to the vegetation. this trend and would be consistent with relocation of Cereal pollen remains high and rises in the lat- agriculture, and perhaps human populations, toward est Punic period, peaking in the Roman Period at the interior of the island (see Chapter 7). This may be Marsa (Carroll et al. 2012). Cereal pollen also appears consistent with an element of population continuity at In-Nuffara, but declines slightly at Tas-Silġ and at the end of the Temple Period and into the Bronze disappears at Wied Żembaq (Hunt 2015) at the same Age. The redistribution of agricultural activity and time. The long record from Xemxija shows very strong perhaps population away from the coast seems a evidence for possible discontinuities and often has consistent response to the well-known rise of maritime very low pollen counts, but nonetheless reflects a raiders in the Early Bronze Age (Grima 2007; Wiener degraded steppe landscape adjacent to a drying former 2013) since at this point coastal localities would have marshy area. become very vulnerable. It is not possible to separate the pollen of culti- vated and wild or feral (descended from cultivated 3.6.6. The Bronze Age (2000–1000 cal. bc) forms) olives, but since Olea pollen is recorded prior to After the decline in cereal pollen around the end of the Neolithic at Xemxija, it is almost certainly a native the Temple Period (2350 cal. bc), which may reflect species and therefore likely derives from oleaster (wild partial population decline, by c. 2050 cal. bc (4000 olive) for most of the Holocene. Its behaviour prior cal. bp) there is evidence from the pollen of cereals, to the Punic Period is in-phase with the rise and fall 112 The Holocene vegetation history of the Maltese Islands

of tree and shrub pollen and out of phase with the poor chronological resolution. Pollen indicative of pollen of cereals, supporting its likely derivation from degraded steppe vegetation is particularly evident oleaster. Only in the later Punic and Roman periods, in the basal half of the core from Santa Marija Bay on from c. 400 cal. bc (2450 cal. bp) onwards at Burmarrad Comino with very low percentages of Poaceae and high and Marsa, is Olea pollen in phase with cereal pollen, percentages of Centaurea and Gladiolus, taxa which are perhaps suggesting that it derives from the cultivation today associated with degraded steppe environments of domesticated olive trees. This corresponds with finds on the islands (Carroll et al. 2012). of olive stones at Marsa and Tas-Silġ (Fiorentino et al. Around ad 1600–1650 (assuming constant sedi- 2012). Recent investigations have shown that the pollen mentation rates), there is a sharp change in the Santa productivity and dispersal of modern Olea europaea Marija Bay core, with a decline in Gladiolus and rise in are very limited, with pollen percentages in surface Poaceae and Plantago (Carroll et al. 2012). At In-Nuffara, samples declining by 60–90 per cent within 100 m of there is a minor increase in Poaceae which is undated, the cultivation site (Florenzano et al. 2017). Values of but predates the appearance of the exotic Eucalyptus olive pollen as high as 8 per cent at Burmarrad (Gam- and thus is probably earlier than the British Period (or bin et al. 2016) and 5 per cent at Marsa (Carroll et al. pre-nineteenth century). At Għajn il-Kbira on Gozo, 2012) during the Roman period may therefore reflect molluscan assemblages relating to the nineteenth extensive olive groves (Anastasi & Vella 2018; Vella et century or earlier are marked by a rise in the grass- al. 2017). Elsewhere in the Mediterranean, such as at land species Truncatellina callicratis (Hunt & Schembri Kouremenos in Crete, there is certainly evidence of 2018). These changes were most likely caused by the intensifying olive tree presence from the later Neolithic policy of the Knights, who encouraged the creation at about 3600 cal. bc onwards (Langgut et al. 2019), of terrace systems (see Chapters 7 & 8) and improve- coincident with the opening-up of the landscape and ment or manufacture of soils and the development of intensification of grazing (Stoddartet al. 2019), although sheep runs (Blouet 1997), but may also reflect a phase the early stages of olive tree management may not of higher effective humidity during the Little Ice Age necessarily imply domestication and cultivation of the (Sadori et al. 2016). tree (Canellas-Bolta et al. 2018, 73, & references therein). Finally, the most recent parts of the In-Nuffara A recent review of the evidence for olive growing in deposit, and the cores from Xemxija, Santa Marija Bay, Puglia, southeast Italy, shows that wild olives have Salina Bay and Marsa (Carroll et al. 2012) all show a been cultivated there since the early Neolithic, and strong rise in Pinus, consistent with the widespread selective cultivation culminated in the first appearance planting of pines as ornamentals during the British of the domestic type during the Middle Bronze Age or Period. The Australian exotic Eucalyptus, was also mid-second millennium bc (Caracuta 2020). planted during British times and was present at In-Nuf- At Xemxija the very high peak of Pinus pollen fara and in the Mistra Valley (Hunt & Vella 2004/2005). during the period c. 150 cal. bc to cal. ad 1150 (1800 Maize and cotton were also present in the field-fill in to 800 cal. bp) is notable, and can probably be best the Mistra Valley (Hunt & Vella 2004/2005). The wider interpreted as evidence for a Roman period pine landscape generally remained extremely degraded. plantation that persisted through the Arab period and into Norman times. Supporting evidence for this 3.7. Conclusions interpretation may be provided by the observation of al-Idrisi, made around ad 1150, incorporated in a later The work done by the FRAGSUS Project has signifi- text by al-Himyari (Brincat 1995) that good timber for cantly changed our understanding of the sequence and ship repair had been available on Malta. causes of Holocene vegetation change in the Maltese Islands, as well as some aspects of the relationship 3.6.8. Medieval to modern (post-ad 1000) between the environment and the human popula- All assemblages of Medieval to early Modern age show tion. For the first time, analysis of multiple sediment a very strong taphonomic imprint, being dominated sequences and archaeological sites has allowed us by corrosion-resistant taxa such as Lactuceae. All other to differentiate between climatic and anthropogenic evidence points towards very degraded environments, signals at times during the Holocene. These entwined mostly degraded steppe, although most records have issues will be returned to in the concluding Chapter 11.

113

Temple landscapes

The ERC-funded FRAGSUS Project (Fragility and sustainability in small island environments: adaptation, cultural change and collapse in prehistory, 2013–18), led by Caroline Malone (Queens University Belfast) has explored issues of environmental fragility and Neolithic social resilience and sustainability during the Holocene period in the Maltese Islands. This, the first volume of three, presents the palaeo-environmental story of early Maltese landscapes. The project employed a programme of high-resolution chronological and stratigraphic investigations of the valley systems on Malta and Gozo. Buried deposits extracted through coring and geoarchaeological study yielded rich and chronologically controlled data that allow an important new understanding of environmental change in the islands. The study combined AMS radiocarbon and OSL chronologies with detailed palynological, molluscan and geoarchaeological analyses. These enable environmental reconstruction of prehistoric landscapes and the changing resources exploited by the islanders between the seventh and second millennia bc. The interdisciplinary studies combined with excavated economic and environmental materials from archaeological sites allows Temple landscapes to examine the dramatic and damaging impacts made by the first farming communities on the islands’ soil and resources. The project reveals the remarkable resilience of the soil-vegetational system of the island landscapes, as well as the adaptations made by Neolithic communities to harness their productivity, in the face of climatic change and inexorable soil erosion. Neolithic people evidently understood how to maintain soil fertility and cope with the inherently unstable changing landscapes of Malta. In contrast, second millennium bc Bronze Age societies failed to adapt effectively to the long-term aridifying trend so clearly highlighted in the soil and vegetation record. This failure led to severe and irreversible erosion and very different and short-lived socio-economic systems across the Maltese islands.

Editors: Charles French is Professor of Geoarchaeology in the Department of Archaeology, University of Cambridge. Chris O. Hunt is a Professor in the School of Biological and Environmental Sciences, Liverpool John Moores University, Liverpool. Reuben Grima is a Senior Lecturer in the Department of Conservation and Built Heritage, University of Malta. Rowan McLaughlin is Senior Researcher in the Department of Scientific Research at the British Museum and honorary research scholar at Queen’s University Belfast. Caroline Malone is a Professor in the School of Natural and Built Environment, Queen’s University Belfast. Simon Stoddart is Reader in Prehistory in the Department of Archaeology, University of Cambridge.

Published by the McDonald Institute for Archaeological Research, University of Cambridge, Downing Street, Cambridge, CB2 3ER, UK.

Cover design by Dora Kemp and Ben Plumridge.

ISBN: 978-1-902937-99-1 ISBNISBN 978-1-902937-99-1 978-1-902937-99-1

9 781902 937991