excavations at edited by john - university of cambridge · per il periodo storico, la ceramica...
TRANSCRIPT
Bova Marina
Archaeological
Project:
Survey and
Excavations
at
Umbro
Preliminary
Report,
1999
Season
edited by John
Robb with contributions by Marina
Ciaraldi, Lin Foxhall, Doortje
Van Hove, and David Yoon
Department of Archaeology
University of Southampton
Southampton SO17 1BJ
United Kingdom
tel. 00-44-23-80592247
fax 00-44-23-80593032
email [email protected]
Bova Marina Archaeological Project 1999
2
ACKNOWLEDGEMENTS
I would like to thank Dottoressa Elena Lattanzi, Soprintendente, and Dottoressa Emilia Andronico,
Ispettrice, of the Soprintendenza Archeologica della Calabria, for their help and guidance in this research. As in
past years, we are grateful to Sebastiano Stranges and to Luigi Saccà for their help, advice, and friendship; to
Brian McConnell and Laura Maniscalco for archaeological advice; to our landlords, Antonino and Silvana
Scordo, and to our cooks, Mariella Catalano and Annunziata Caracciolo. Mary Anne Tafuri translated the
project summary and Fiona Coward helped assemble the final documentation. Finally, I would like to thank the
field staff (David Yoon, Lin Foxhall, Paula Lazrus, Keri Brown and Starr Farr) and post-excavation staff and
analysts (Umberto Albarella, Marina Ciaraldi, Sonia Collins, Kathryn Knowles, Doortje Van Hove, Jayne Watts,
and David Williams) for their expertise, and all the crew members listed below for their hard work and
enthusiasm.
We gratefully acknowledge funding from the British Academy (Excavation Grant A-AN4798/APN7493
Supplementary post-excavation funding), the Arts and Humanities Research Board (Research Grant AHRB/RG-
AN4798/APN8592), the Department of Archaeology, University of Southampton, and the School of
Archaeology and Ancient History, University of Leicester.
BOVA MARINA ARCHAEOLOGICAL PROJECT: CREW, 1999 SEASON
Director John Robb
Survey Co-Director David Yoon
Survey Co-Director Lin Foxhall
Field Supervisor Keri Brown
Field Supervisor Paula Kay Lazrus
Lab Manager Starr Farr
Artist (Southampton) Kathryn Knowles
Computing (Southampton) Doortje Van Hove
Faunal analysis (Birmingham) Umberto Albarella
Paleobotany (Birmingham) Marina Ciaraldi
Cook Mariella Catalano
Cook Annunziata Caracciolo
Crew Members Siân Anthony
Francesca Binyon
Rebecca Crowson-Towers
Glenn Dunaway
Lauren Dumford
Mark Ellis
Anne Forbes
Helen Forbes
Janet Forbes
Lucy Heaver
Alex Hopson
Jenny House
Charlotte John
Claire Rees
Kathryn Simms
Barney Skinner
Jayne Watts
Matthew Wortley
Child Johanna Farr
Infant Nicholas Robb
Bova Marina Archaeological Project 1999
3
RIASSUNTO ITALIANO
Nel 1999, il Progetto Archeologico Bova Marina ha intrapreso la sua terza campagna di scavo. Le
finalità principali della ricerca includevano l’attività di ricognizione di varie zone intorno al comune di Bova
Marina e lo scavo del sito preistorico di Umbro. Tali attività hanno avuto luogo dal 29 agosto al 26 settembre
1999, con un gruppo di 25 persone. I responsabili dello scavo erano John Robb (scavo preistorico e
amministrazione generale), David Yoon (ricognizione di superficie, archeologia Romana) e Lin Foxhall
(ricognizione di superficie, archeologia Greca).
Ricognizione di superficie
Finalità e metodi. Lo scopo principale dell’attività di ricognizione era rivolto alla comprensione dello
sviluppo, avvenuto nel corso della preistoria ed in epoca storica, dei modelli insediamentali legati al territorio di
Bova Marina. Si è continuato quanto intrapreso nel 1997 e 1998. Il metodo utilizzato consiste nel systematic
transect walking, mediante il quale gruppi di 4-5 persone, poste ad intervalli di 10 m., camminano su territori
delimitati registrandone le caratteristiche storiche e geografiche. Nel 1999, il nostro gruppo di ricognitori ha
coperto una superficie di 102 ettari, portanto l’area totale finora esplorata sistematicamente a 274 ettari. Le zone
ricognite comprendono due grossi appezzamenti contigui (intorno ad Umbro e nella zona costiera di San
Pasquale), e numerose porzioni di territorio sparse intorno al comune. I lavori del 1997 e 1998 si erano
concentrati prevalentemente nei territori di Umbro e San Pasquale. Nel 1999, la direzione della ricerca era
rivolta ad estendere la porzione di territorio ispezionato ad ambienti nuovi come colline e pianori interni.
Tutti i ritrovamenti sono stati catalogati e datati per quanto possibile. E’ da tenere presente tuttavia che
la sequenza cronologia della ceramica presente in questa regione è piuttosto variegata. Alcuni siti preistorici,
chiaramente Neolitici, possono essere facilmente individuati grazie alle decorazioni presenti sui frammenti
ceramici. Tuttavia altri periodi, quali l’Eneolitico e l’età del Bronzo, sono caratterizzati da ceramica inornata la
cui identificazione si basa, per la maggior parte dei casi, sulla forma del vaso, rendendo quindi molto difficile la
lettura di frammenti di piccole dimensioni. Per il periodo storico, la ceramica Greca e Romana è facilmente
identificabile, tuttavia, pochi dati si posseggono sulla ceramica medievale di questa regione, forse a causa dello
scarso numero di siti medievali individuati.
Risultati. Fino ad oggi sono stati identificati 37 siti che variano da una piccola concentrazione di
frammenti preistorici non databili a grandi villagi di epoca Greca e Romana con complessa articolazione interna.
I siti sono elencati e descritti al paragrafo 2.1.4. E’ da tenere presente che tale elenco aggiorna quello
precedenemente fornito in quanto alcuni siti scoperti nel passato sono stati successivamente ridatati.
Nel considerare i siti in ordine cronologico, fino ad oggi, non si ha traccia di attività Paleolitica o
Mesolitica. I siti di epoca Neolitica sono manifestati in una numerosa varietà che comprende terrazzi costieri,
basse colline prospicenti valli fluviali e pianori interni. Non si sono evidenziati insediamenti riferibili
all’Eneolitico o all’età del Bronzo (benché tracce di frequentazione di età del Bronzo sono state messe in
evidenza durante gli scavi ad Umbro, cfr. più avanti); tale aspetto può essere riferibile in parte a problemi di
datazione dei pezzi raccolti durante l’attività di ricognizione (vedi sopra). Le testimonianze di epoca Greca
includono un grande villaggio (Mazza, cfr. più avanti), e numerosi piccoli cascinali interni (siti 18 e 33). Molti
sono situati su alti terrazzi. In epoca Romana si hanno per la prima volta tracce di siti su fondo valle, che
suggeriscono il passaggio a forme di agricoltura intensiva. I resti ceramici di epoca Greca e Romana mostrano
due apici: uno che va dal periodo Tardo Arcaico a quello Classico (VI-V secolo a.C.) e l’altro in epoca
Imperiale (III-V sec d.C.). Gli insediamenti più tardi sembrano occupare gli stessi territori, ma in minore
consistenza, fino all’VIII sec. circa. L’epoca Medievale rimane oscura tuttavia nel XIX sec. si registrano di
nuovo numerosi insediamenti a carattere rurale.
Oltre all’attività di ricognizione sono state portate a termine numerose indagini secondarie rivolte allo
studio dell’uso del territorio. Esse comprendono il riconoscimento, in ambito geologico, di quelle materie prime
che possono essere state utili in epoca preistorica (da continuare nel 2000), e il GIS (Geographical Information
System) computer modelling volto alla ricostruzione delle antiche dinamiche di sfruttamento del territorio in
termini di esigenze relative ai diversi modelli economici preistorici.
Raccolta di superficie intensiva dal sito di Mazza. Si è intrapresa una raccolta di superficie intensiva
sul sito Greco di Mazza. Questo ampio villaggio ha finora restituito le uniche ceramice Greche di imporazione
del periodo arcaico/classico, un possibile frammento Attico ed un possibile frammento Laconico. Le finalità
della raccolta intensiva erano di determinare i limiti geografici, l’intervallo cronologico e la distribuzione
spaziale interna del sito. Come prima cosa è stata stabilita una griglia di punti di riferimento su tutto il sito e
sono stati collocati punti di raccolta ogni 30 metri all’interno della stessa. Si è poi proceduto alla raccolta di tutti
i manufatti presenti in un’area di 10 metri quadrati per ogni punto di raccolta. I risultati potranno essere
Bova Marina Archaeological Project 1999
4
analizzati quantitativamente. L’analisi dei dati provenienti dal sito di Mazza è ancora in corso, tuttavia alcune
considerazioni sono già possibili. Il sito sembra possedere una componente Romana preponderante che si
aggiunge alle evidenze di epoca Greca. Un considerevole numero di tegole di epoca Romana, insieme a tegole di
epoca Greca, è stato ritrovato nella parte alta del sito. Tegole di epoca Greca sono state ritrovate in piccole
quantità sul resto della zona superiore del sito e sul fianco sud-est della collina. La ceramica fine Greca
comprende una considerevole quantità di tazze di fattura locale e forme a cratere. Gli affioramenti rocciosi
presenti nella parte centrale del sito somigliano a quelli utilizzati nelle fondazioni di Locri Epizephyrii, e
possono essere aver avuto lo stesso scopo. L’insediamento si estende con minore densità su tutto il pianoro di
Mazza. Il limite meridionale del sito potrebbe essere stato, in antico, una zona industriale, come testimonia l’alta
concentrazione di resti di fusione e lavorazione dei metalli. Resti di concotto e scorie di fornace sono stati
raccolti e sono attualmente in corso di analisi presso l’Univesità di Leicester.
Il sito di Mazza è stato abitato lungo un ampio arco cronologico durante il periodo classico. I materiali
raccolti hanno indicato la presenza sia di abitazioni che di attività industrali nonché di un probabile grande
edificio in prossimità della sommità del sito. Si è accertata inoltre la presenza di ceramica preistorica. Studi
futuri sul sito si concentreranno sulla relazione tra lo stesso e le città di Reggio e Locri Epizephyrii e sulle
problematiche relative all’occupazione indigena della zona. L’individuazione, nell’area studiata, di una serie di
piccoli insediamenti rurali Greci e la considerevole distanza da qualsiasi città Greca rappresentano un
interessante aspetto da analizzare nella comprensione della colonizzazione Greca e dell’occupazione di territori
rurali a scopo coloniale. In genere, gli studi sulla colonizzazione Greca si sono concentrati su informazioni
provenienti da siti urbani come Locri e dalle zone circostanti come Metaponto. Le datazioni antiche di alcuni siti
provenienti dalla ricognizione di Bova sono sorprendenti, esse potrebbero indicare che l’insediamento Greco si
sia diffuso ampiamente sul territorio coloniale nell’arco di poche generazioni, o che le popolazioni indigene si
siano, almeno in parte, ellenizzate piuttosto rapidamente. La funzione di questi siti, e la loro relazione con le
città Greche rimane piuttosto incerta, non si hanno inoltre informazioni circa una occupazione autoctona che sia
essa contemporanea o leggermente più antica. Lo scavo di un insediamento Greco di piccole dimensioni come il
sito 18 (Umbro) potrebbe fornire dati da mettere in relazione con siti urbani come Locri o ‘cascinali’ presenti su
territorio Greco, allo scopo di risolvere alcune questioni.
Note su siti soggetti a minaccia. Benché numerosi degli insediamenti presenti nel comune di Bova
Marina siano stati danneggiati dall’erosione o da processi geomorfologici, alcuni dei siti osservati sono risultati
soggetti ad azioni distruttive provocate dall’uomo quali, ad esempio, l’attività edilizia. Tutti i siti minacciati sono
a conoscenza della Soprintendenza; questi includono principalmente la villa Romana o cascinale di Panaghia,
San Pasquale, in corso di danneggiamento da attività di costruzione e attualmente vincolato, e la villa Romana o
insediamento urbano di Torrente Siderone (Amigdala), sotto la SS 106 presso la città moderna di Bova Marina,
attualmente in corso di scavo da parte della Soprintendenza.
Scavi ad Umbro
Umbro appare come un complesso sito a più fasi, posto a circa 4 km nell’entroterra, in prossimità
dell’antico confine tra i comuni di Bova Marina e Bova Superiore. Scoperto nel 1990 da S. Stranges e L. Saccà,
il sito è stato soggetto a ripetute attività di raccolta di superficie e, nel 1998, ad una nostra prima campagna di
scavo. Lo scavo del 1998 stabilì che l’insediamento consisteva di una piccola area, alla base di un dirupo
intensamente occupato durante il Neolitico, e di numerose concentrazioni di materiale, alla sommità dello stesso,
probabilmente di età del Bronzo. L’insediamento Neolitico principale ha restituito una datazione al
radiocarbonio che va dalla metà del VI millennio BC alla metà del V millennio BC (Stentinello). E’ inoltre
attestata l’occupazione durante la facies di Diana e, probabilmente in maniera sporadica, durante l’età del Rame.
Nel 1999, i lavori si sono concentrati in tre trincee: la Trincea 1, la Trincea 6 e la Trincea 7.
Trincea 1: l’area Neolitica. Tale trincea fu iniziata sottoforma di sondaggio stratigrafico nel 1998 ed
esteso allo scopo di comprendere la zona centrale del sito di epoca Neolitica posto alla base del dirupo. Essa
comprendeva una trincea di un metro per sei metri, con una piccola estensione laterale al centro. Gli scavi sono
proceduti in aree di un metro quadrato con tagli arbitrari di 10 cm.; il terreno rimosso è stato setacciato in griglie
di 5 mm. Gli scavi sono proceduti fino ad arrivare alla base rocciosa nella parte meridionale della trincea; nel
margine settentrionale della stessa lo scavo si è interrotto ad 1,5 m. dal piano di calpestio; il materiale antropico
presente a questa profondità era considerevolmente diminuito.
La Trincea 1 ha restituito numerosi manufatti Neolitici, ma nessuna evidenza di strutture abitative quali
capanne. E’ evidente che il sito consisteva in un’area ristretta (con uno spazio abitativo di forse 20 metri
quadrati), occupata da piccoli gruppi di persone lungo il corso di diversi millenni. Tutti i manufatti erano
considerevolmente frammentati. Sono stati identificati in tutto cinque strati. Lo Strato I consisteva di 5-10 cm. di
terreno di superficie inquinato. Lo Strato II era formato da uno spesso livello di sedimento bruno chiaro con
Bova Marina Archaeological Project 1999
5
abbondante presenza di materiale roccioso di crollo, probabilmente disturbato sian in epoca recente che in
antico. Gli Strati I e II contenevano manufatti di tutti i periodi. Lo Strato III, di sedimento chiaro giallognolo,
conteneva materiale roccioso di crollo di diverse dimensioni e una serie di lenti sabbiose apparentemente
derivate dal disgregamento dei frammenti rocciosi. Il materiale ceramico presente è riferibile alle facies di
Stentinello e Diana, e sembra essere riferibile a fasi di Diana con presenza di intrusioni di epoche precedenti; ciò
non sorprende visto che l’area Neolitica è posta sul fondo di un bacino roccioso. Lo Strato IV consiste di un
terreno più denso e compatto di colore bruno. Esso contiene prevalentemente ceramica di tipo Stentinello ed è
databile al V millennio BC. In ultimo, lo Strato V è alla base dell’area scavata e contiene terreno argilloso di
colore bruno scruro con presenza di ceramica Stentinello databile al VI millennio BC. I manufatti Neolitici
raccolti comprendono ceramica tipica degli stili di Stentinello e Diana, alcuni esempi di ceramia Neolitica di
altro tipo, quali ceramica a pittura rossa, ed alcuni probabili frammenti di epoca Eneolitica (prevalentemente
dagli Strati II e III). L’industria litica era ricca di ossidiana ed includeva numerose piccole schegge e lamelle.
Sono state ritrovate un’ascia in anfibolite ed una piccola riproduzione di ascia in fillite, si è attestata inoltre la
presenza di numerosi ciottoli in pietra metamorfica di importazione, alcune macine e alcuni frammenti di
concotto e di ocra rossa. I resti faunistici non sono ancora stati studiati, tuttavia è già evidente la preponderanza
degli ovini e dei caprini, con presenza di maiale; i bovini sono scarsamente rappesentati e l’ittiofauna è assente.
Sono stati conservati campioni di flottazione provenienti da ogni Strato. L’analisi dei semi e dei resti vegetali
carbonizzati mostrano che la coltura di alcune specie era praticata; i resti vegetali includono la veccia, il
Triticum aestivum s.l., il Triticum dicoccum, e l’orzo.
La Trincea 6: l’area di età del Bronzo. Nel 1999 si iniziata la pulitura del profilo della rupe dal quale la
ceramica dilavava dal bordo accanto alla strada che attraversa il sito. Dal fianco del dirupo è emerso un vaso
intero, si è così deciso di aprire una trincea per effettuare ulteriori indagini. Nel complesso si è messa in
evidenza la deposizione di un gruppo di tre vasi comprendenti due tazzine attingitoio con ansa soprelevata ed
una grande olla dotata di tre anse orizzontali poste vicino all’orlo. I vasi erano affiancati e posti su uno stesso
piano di calpestio distinguibile da quella di riempimento sopra per la grandezza e l’orientamento delle pietre. Si
è attestata la presenza di piccoli frammenti di concotto e carbone. Subito sotto il gruppo di vasi è stato ritrovato
un inusuale frammento di corno fittile. Sulla superficie del battuto si è attestata la presenza di un crollo con
pietre di medie e grandi dimensioni (fino a 30 cm. ca.). Il crollo sembrava di maggiore volume ed intensità ad est
del gruppo di vasi. Non è chiaramente comprensibile se ci si trovi di fronte ad un fenomeno erosivo naturale o al
crollo di qualche strutture abitativa. In termini di datazione, i vasi sono chiaramente riferibile al Bronzo Antico
ed il corno fittile è tipico dello stile Rodì-Tindari-Vallelunga. In stretto accordo con quanto suggerito dallo stile
ceramico, la datazione al radiocarbonio, fornita da un frammento di carbone proveniente dal piano di posa del
gruppo dei vasi, è riferibile al 1720-1580 BC (cal.)(intervallo 2-sigma).
Il gruppo ceramico della Trincea 6 potrebbe rappresentare quanto rimane del crollo di una capanna,
benchè sono scarsi i resti di frequentazione domestica all’interno della trincea e segni di strutture tipiche dell’età
del Bronzo quali capanne con soglie in pietra. Una seconda possibilità potrebbe essere rituale: i vasi potrebbero
rappresentare una qualche deposizione rituale idiosincratica. Una terza possibilità è rappresentata da una
eventuale deposizione funeraria. Confronti effettuati con contesti della Sicilia orientale hanno messo in evidenza
la presenza ricorrente, durante il Bronzo Antico, di sepolture situate lungo margini scoscesi di pietra calcarea o
di altro tipo, così come appare essere Umbro. Si tratta in genere di piccole camere artificiali raramente più
grandi dei di 2 metri di diametro e 1,5 metri di altezza, spesso raggiungibili attraverso uno stretto passaggio. In
diversi casi, inoltre, gli archeologi hanno messo in luce aree livellate o piattaforme che contenevano gruppi di
vasi rappresentanti il corredo funerario volontariamente posto all’esterno della tomba piuttosto che all’interno di
essa (es. Santa Febronia, Maniscalco, 1996).
Trincea 7. La Trincea 7 consiste di un piccolo (un metro per un metro) pozzetto scavato
immediatamento al di sotto del margine del dirupo, lungo il limite meridionale del bacino che racchiude il sito
Neolitico. Si è sperato di recuperare in questa area materiale di epoca preistorica. Tuttavia, pochi resti sono stati
ritrovati, e ad una profondità di 1,5 metri lo scavo della trincea è risultato impraticabile.
Sondaggi di Scavo a San Pasquale
Nel 1999 è stato sondato il sito Neolitico di San Pasqule, situato su di un basso terrazzo lungo il
margine orientale della valle di S. Pasquale. Si è sperato di verificare la presenza di un “villaggio” Neolitico, per
il quale la zona offriva le caratteritiche tipiche. Due pozzetti di un metro per un metro sono stati scavati in una
zona situata lungo il margine occidentale del sito, dove in superficie emergevano frammenti di ceramica
preistorica. I risultati si sono rivelati piuttosto scoraggianti. Entrambi i sondaggi hanno evidenziato che non
esistevano depositi archeologici al di sotto dello strato superficiale; il terreno si è rapidamente trasformato in uno
strato di sabbia sterile. Appare evidente che il materiale ceramico di epoca preistorica proviente dal “sito” deriva
Bova Marina Archaeological Project 1999
6
dal dilavamento della collina posta a nord-est di esso, in parte coltivata ed in parte coperta da una fitta macchia
mediterranea. Indagini future potrebbero concentrarsi sull’individuazione del sito originario.
Indagini Future
La campagna di scavo del 1999 ha aperto la strada a nuove prospettive di ricerca.
Ricognizione. Nelle campagne future, speriamo di poter continuare ed incrementare l’attività di
ricognizione, che stà fornendo nuove informazioni sull’utilizzazione del territorio. Allo stesso tempo speriamo di
indagare due problemi metodologici.
Ricognizioni di altura a Bova Superiore. Sia le indagini storiche che la nostra ricostruzione informatica
al GIS hanno dimostrato l’importanza, in diversi periodi, degli insediamenti d’altura in zone interne. Un
sondaggio basato solo sull’indagine di zone costiere sarebbe inevitabilmente incompleto e non potrebbe fornire
l’intera consistenza e l’evoluzione delle dinamiche insediamentali del territorio di Bova Marina; questo potrebbe
spiegare inoltre la scarsa presenza di siti di età del Ferro o di epoca Medievale. Nel continuare i sondaggi a Bova
Marina, speriamo inoltre di poterci estendere al comune di Bova Superiore e di completare il quadro delle
conoscenze sulle strategie insediamentali antiche, investigando un completa serie di diversi ambienti.
Sondaggi di scavo in siti di superficie non datati. Sono inoltre stati ritrovati una serie di siti non datati
che potrebbero rappresentare i nosti “periodi mancanti”. L’unico modo per indagare gli stessi consiste nello
scavo programmato di sondaggi di un metro quadrato, in grado di fornire materiale datante meno danneggiato,
campioni per datazioni assolute al radiocarbonio, ed informazioni sullo stato di conservazione del sito.
Nell’immediato futuro ci auguriamo di poter sondare due aree entro un raggio di 200-300 metri dallo scavo di
Umbro: Limaca, dove una concentrazione di materiale potrebbe rappresentare una successiva occupazione di età
del Bronzo (si è ottenuto un permesso nel 1999 che è però arrivato troppo tardi perché si potesse praticamente
procedere allo scavo); e una zona a circa 200 metri a sud-ovest di Umbro dove due concentrazioni di frammenti
di epoca preistorica sono visibili in superficie. Quest’ultima potrebbe rappresentare un villaggio Neolitico
all’aperto; si considera critica una eventuale verifica che possa permettere di comprendere le dinamiche
insediamentali del territorio di Umbro.
Lo scavo ad Umbro: Ci auguriamo di poter continuare gli scavi sia del sito Neolitico che di quello di
età del Bronzo dell’area di Umbro. Gli scavi saranno limitati nelle finalità e si concentreranno sulla
individuazione della base della stratigrafia, sulla verifica dell’esistenza di eventuali frequentazioni pre-
Neolitiche e sul recupero di un maggior numero di campioni per analisi, per esempio, di tipo paleobotanico.
Nell’area di età del Bronzo Antico (Trincea 6), speriamo di poter verificare la natura del gruppo ceramico,
ovvero se formi o meno parte di una sepoltura.
Altre attività di scavo: Come già accennato, l’area di Umbro è ricca di piccole concentrazioni di
materiale archeologico. In alcune di esse ci auguriamo di poter effettuare dei sondaggi esplorativi (vedi sopra).
Un ulteriore sito da indagare è rappresentato dal piccolo insediamento Greco posto a circa 200 meri a sud di
Umbro. La presenza di piccoli insediamenti interni di epoca greca, come il cascinale di Umbro, si sono rivelati
una scoperta sorprendente e potrebbero risultare di notevole importanza nella comprensione del processo di
colonizzazione di questa zona.
Note sull’organizzazione della documentazione
La documentazione relativa alla campagna del 1999 risulta divisa in due parti. La prima parte (questo
volume) consiste in un rapporto descrittivo illustrato. La seconda parte è formata dalla compilazione di dati di
appendice per il progetto e per gli archivi della Soprintendenza. Sono inclusi cataloghi degli oggetti suddivisi
per provenienza e categoria di oggetto, disegni di tutti i frammenti diagnostici, il giornale di scavo e schede
standard per siti, unità stratigrafiche e frammenti ceramici.
Bova Marina Archaeological Project 1999
7
CONTENTS
Riassunto Italiano 3
1. Introduction 8
2. Field Survey and Landscape Studies 9
2.1. Field Survey (David Yoon) 9
2.2. Intensive collections at Mazza (Lin Foxhall) 19
2.3. Geological reconnaissance 20
2.4. GIS analysis of land use (Doortje Van Hove) 21
3. The Excavations at Umbro 24
3.1. Introduction: previous work, goals and methods for this season 24
3.2. Description of 1999 Trenches 25
3.3. Other areas explored 30
4. Umbro Excavations: Description of Finds 32
4.1. Ceramics 32
4.2. Lithics 34
4.3. Other Artifacts: daub, worked shell and bone, ground and polished stone, and ochre 35
4.4. Faunal Remains and Human Remains 37
4.5. Paleobotanical Remains (Marina Ciaraldi) 38
5. Test Excavations at San Pasquale 41
6. Conclusions and Future Research Directions 43
6.1. Umbro: dating, stratigraphy, and site function 43
6.2. The Landscape of Bova Marina from Neolithic through Recent Times 43
6.3. Future Work 43
Bibliography 45
Figures and Tables 47
NOTE ON THE ORGANIZATION OF THIS REPORT
Description of the 1999 field season is presented in two parts. This report, which comprises the first part,
presents a self-contained, synthetic description of the survey, excavations and finds. The second part presents
detailed records of various kinds for archive purposes: catalogs of finds inventoried within each artifact bag and
listed by kind of artifact, ceramic inventories and drawings of all diagnostic sherds, the excavation diary, and so
on.
Bova Marina Archaeological Project 1999
8
1. INTRODUCTION
1999 was the third year of the Bova
Marina Archaeological Project. In 1997, a small
crew of 5 Southampton students led by John Robb
carried out field survey in the comune of Bova
Marina. In 1998, a larger crew of 10 Southampton
students and about 5 staff, directed by John Robb
and David Yoon, surveyed further areas, and
excavated numerous exploratory trenches at the
Neolithic site of Umbro.
These campaigns are described in the
1997 and 1998 preliminary reports (Robb 1997;
1998), which also detail the circumstances in which
fieldwork was carried out and the geography,
geology, and archaeology of the region. To avoid
repetition, these topics will not be extensively
described here.
In 1999, we had the largest field season to
date, with 18 (10 from Southampton and 8 from
Leicester) crew members and 10 staff or
contributing specialists, directed by John Robb,
David Yoon and Lin Foxhall. We worked in Bova
Marina from August 29 through September 25. The
season had three major goals:
(1) to carry out extensive excavations in
the Neolithic area of Umbro, with particular
attention to economic and contextual information
about the Neolithic habitation there;
(2) to carry out test excavations in the
Neolithic site of San Pasquale, in order to assess its
potential for large-scale excavations;
(3) to continue the field survey, extending
it to new areas, particularly inland; the overall
survey goal is to understand the evolution of
settlement in the area of Bova Marina throughout
both prehistoric and historic times.
In addition, we aimed to carry out a number of
complementary minor projects such as GIS
environmental modelling, geological raw materials
survey, and intensive gridded collection at the
Greek site of Mazza with the goal of clarifying its
internal organization and evolution.
The 1999 field season ran smoothly, with
no major hitches, and these goals were
accomplished. In some cases the results were
disappointing, as in the San Pasquale test
excavations, but in other cases surprises emerged,
as in the Bronze Age ritual deposition discovered at
Umbro. The purpose of this report is to describe the
methods and results of this field season in detail,
and to interpret their significance for both the
archaeology of the region and the development of
future fieldwork.
Bova Marina Archaeological Project 1999
9
2. FIELD SURVEY AND LANDSCAPE STUDIES
2.1. Field Survey (David Yoon)
2.1.1. Background
The overall goal of the BMAP field survey
is to understand how people used different areas of
the Calabrian landscape for various purposes such
as habitation, farming, herding, foraging,
specialized production, and so on, in all periods
from the Paleolithic through modern times. As
detailed in previous reports, the local landscape is
both very rugged, with very little level ground, and
highly varied, with very diverse environments
located close to each other. The contrast between
narrow coastal strips and valley bottoms and
mountainous interior is especially marked.
Historically, there appear to have been oscillations
between coast-oriented settlement such as in
Roman times and inland settlement as in medieval
times. As also detailed previously, a number of
landscape features make archaeological field survey
challenging and its results sometimes problematic.
These include heavy erosion on slopes and deep
alluviation on valley bottoms, intensive and
destructive historical land use, the rapid
proliferation of fenced-in non-surveyable areas, and
the fact that pottery fragments from many periods
are not particularly diagnostic.
In 1997, a strategy of visiting previously
known sites and systematically walking a variety of
small areas provided an efficient introduction to the
local archaeological sequence and also provided
useful experience for adapting survey methods to
local conditions. In 1998, the same survey strategy
was continued, together with an effort to assemble
larger contiguous tracts of systematically surveyed
territory, especially around Umbro and San
Pasquale.
The work in 1997 and 1998 raised a
number of questions which we hoped to address in
1999. Some periods (Paleolithic, Copper Age, Iron
Age, Medieval) were absent or nearly absent from
our collections. Does this represent an actual
scarcity of occupation in those periods or did it
instead reflect poor recognition of artifacts by us or
poor choice of survey areas? The recognition of
several small Greek sites of early date raised many
questions about the process of Greek colonization
in the region, and how pre-existing native
communities interacted with the Greek presence.
The differences observed so far in Greek and
Roman site locations (Greek sites on hills, Roman
sites in valleys), despite similarities of technology
and social organization, should be associated with
important differences in land use and economic
organization.
In the 1999 survey, co-directed by David
Yoon and Lin Foxhall, we therefore sought in 1999
to improve the quality of our evidence as it relates
to these questions, by:
• doing more survey of hilltops and high-
elevation areas
• continuing to build up larger blocks of
completely surveyed territory
• obtaining more detailed information about the
locations and internal organization of Greek
and Roman sites
Survey of hilltops and high-elevation areas was
intended to improve the representation of these in
our sample, and particularly to see whether the
missing time periods would be located in these
places. Larger contiguous blocks of surveyed
territory enable better understanding of the relative
placement of contemporary sites, which is
important for interpreting the organization of local
communities of any period. Because Greek and
Roman village sites are relatively large and
complex, they are difficult to interpret through
ordinary fieldwalking methods. The controlled
surface collections at Deri (Site 9, San Pasquale) in
1997 provided some information on the internal
structure of one large Roman site. We decided in
1999 to do similar controlled surface collections at
Mazza, the largest known Greek site in the study
area, to get comparable information for that period.
2.1.2. Procedures
The methods used in 1999 were essentially
the same as in previous years. We worked in crews
consisting either of one of the co-directors and
three to four students or (more rarely) both co-
directors and six to eight students. All survey was
carried out within defined “areas,” zones .1-2 ha in
size whose boundaries, location, and geology and
land use were recorded systematically, All areas
without previously identified sites were walked in
parallel transects with the crew members spaced 10
meters apart (in some locations, where it was
difficult to maintain this interval due to steep slopes
or thick vegetation, we were able only to
approximate it). Previously identified sites were
collected more intensively, using either a systematic
grid of 10 m2 collection areas or nonsystematic
collection of any artifacts observed on the site. In
1997 and 1998 some areas with previously reported
Bova Marina Archaeological Project 1999
10
sites were surveyed less systematically, but we
attempted for the sake of comparability to use
systematic methods for all areas in 1999.
The 1999 survey was conducted from 29
August to 17 September. This included 2 days of
fieldwalking with two crews, 5 days of fieldwalking
with one crew, 0.5 days of fieldwalking with a
double-size crew, and 3.5 days of intensive surface
collection with a large crew at Mazza. In all this
represents 73 person-days of work, of which 45.5
were used for fieldwalking and 27.5 for intensive
collection at Mazza.
The fieldwalking survey in 1999 covered a
total of 102.0 ha, of which 5.8 ha had been
surveyed in previous years and 96.2 ha were new.
In all, excluding repeat visits to the same area, we
have surveyed a total of 274.5 ha in the three years
of the project so far. We defined 53 new survey
areas in 1999, numbered 102 through 154. Some of
these (Areas 102, 103, 130, 131, 133, and 134) may
overlap to some degree with areas from previous
years. All new areas were surveyed using
systematic transect walking except Areas 150
through 154, which were areas at Mazza surveyed
using a systematic grid of collection areas instead
(see below).
The survey work of the past three years
has been concentrated in several locations. The two
largest clusters are around Umbro (Figure 1, Figure
2) and in the lower part of the San Pasquale valley.
In each of these places numerous survey areas make
up almost a square kilometer of contiguous
coverage. Most of this was done in 1997 and 1998,
but a few areas were added at Umbro in 1999,
especially to the northeast. A third cluster of survey
areas is located in the middle part of the San
Pasquale valley, but in this case it consists of
several disconnected fragments, due to difficulties
of access. A few of these areas were done in 1997
and 1998, but most were done in 1999. Several
smaller clusters have been selected to represent
particular types of location: Mazza (Figure 3) and
Capo Crisafi for coastal hills (1997, plus the
controlled collections in 1999), M. Rotonda and M.
Vunemo for high inland hills (both newly done in
1999), and M. Silipone to compare the valley of the
Torrente Sideroni with the San Pasquale valley
(mostly done in 1998, plus two small areas in
1999). Several small isolated patches occur as well,
mostly to investigate known sites or to survey small
patches of accessible land near the modern town of
Bova Marina; a few of these have been done each
year.
2.1.3. Ceramics and chronology
One of the biggest problems for survey in
the past was the lack of a well-defined local
ceramic sequence which would be used to date sites
found on survey. During 1999, considerable
progress was made at identification of the pottery
from the survey. We were able to work several days
in the Museo Archeologico Nazionale in Reggio di
Calabria before the field season, restudying some of
our 1997 and 1998 collections and also comparing
them to excavated materials from the Roman site at
San Pasquale (Deri). We are grateful to Dssa.
Emilia Andronico and the staff of the museum for
their assistance with this work.
Based on this research, a new provisional
ceramic classification was set up before the field
season (and modified slightly during the field
season), which was used for classification of the
1999 survey finds. This system will require
additional modification as our knowledge of the
regional pottery improves, but it has improved our
ability to recognize chronological information in
our surface collections, by enabling us to associate
the better-known wares with changes in the local
productions.
Of the remaining problems in this ceramic
chronology, the most important is the prehistoric
period. Prehistoric pottery can usually be
distinguished from that of later periods, but that is
often the limit of our resolution so far. The standard
chronological types for prehistoric pottery in
southern Italy, based on whole vessels from burials,
make use of both form and decoration. For some
periods, such as the Neolithic, surface decoration is
often useful, but after the Neolithic, most vessels
were probably undecorated and vessel form is the
most distinctive criterion. Unfortunately, the small
eroded fragments in our surface collections have
little evidence for either, but especially for vessel
form. Traces of impressed decoration sometimes
survive, making the Neolithic period more visible
than other portions of prehistory, while post-
Neolithic periods are far less easy to identify. It is
likely that some very broad divisions should be
possible on the basis of fabric and surface
treatment, although these are not likely to be as
precise as the existing categories. Of our prehistoric
sites, some have been dated by chance survival of a
diagnostic element or two, some have a suggested
date based on a subjective assessment of similarity
to the excavated assemblage from Umbro, and
some remain undated.
The chronology of the Greek period is
founded on the black-gloss finewares which,
although mostly of regional manufacture, reflect
stylistic trends common throughout the Greek
world. These are associated with reddish brown
Bova Marina Archaeological Project 1999
11
sandy coarseware, beige plainwares, large and
coarse storage jars, and sandy beige transport
amphoras of Greek-derived form. Roof tile appears
for the first time in the Greek period as well,
predominantly in a sandy, light-colored fabric and
taking the form of flat tile (tegula) with wide
flanges and cover tile (imbrex) with a flat top and
angular sides.
The assemblages of the
Hellenistic/Republican and early Imperial periods
remain ill-defined, because very few imports of
well-dated types have been found to confirm dates
in this range. The assemblages of local
commonwares seem to show a gradual changing
mixture, however, of the sandy brown cooking ware
and fine plainware of the Greek period and the
gritty light-colored wares of the later Roman
period. An orange variant of the Greek cooking
ware is likely to belong to this period.
The late Roman period is well defined by
the presence of imports, including African Red Slip
ware, plainwares, and amphoras from North Africa,
as well as small quantities of Late Roman C ware
and Late Roman amphoras from the eastern
Mediterranean. The types and quantities found in
Bova Marina suggest a peak of imports between the
late 2nd or early 3rd and late 5th centuries AD. The
associated local wares include light-colored
coarseware and amphoras tempered with well-
sorted grit (around 1 mm in size) and light-colored
fine plainware. Roman roof tile, compared to
Greek, tends to be in grittier fabrics and to have
slightly different forms, with thinner flanges on the
tegulae and a curved shape to the imbrices.
Probably at the later end of the sequence of Roman
coarsewares is a brown fabric tempered with
abundant grit, mica, and sometimes chunks of
schist. The precision of dates within the Roman
period is limited so far, however, because the sites
appear in most cases to have been occupied for
quite long periods, several centuries in the case of
Sites 9 and 22.
There is a large, ill-defined gap between
the end of the Roman period and the recent
assemblage associated with farmhouses of the 19th
and 20th centuries. It is not clear how much of this
results from an actual scarcity of settlement and
how much from poor recognition of the pottery of
those periods. Judging from the few available
reports, from the 9th to 11th centuries, ceramic
assemblages in Calabria seem to consist of coarse
cooking wares, semi-fine or finely sandy wares with
red painted decoration, and Italian-Byzantine
amphoras. All of these are continuations from Late
Antiquity, although the details may vary. In
addition, occasional lead-glazed pottery begins by
the 9th century, either with a thick green glaze or
with a sparse glaze overall but large patches and
streaks of thick glaze. In the 11th century the first
tin-glazed wares with green and brown decoration
appear. Very little has been published on late
medieval pottery in Calabria, but based on other
parts of southern Italy one would expect tin-glazed
protomaiolica decorated in various colors, sgraffito
wares, lead-glazed cooking wares, and unglazed
semi-fine wares with red painted decoration
(narrower, finer painting than on the early medieval
version). Until a better sense of the local
productions can be obtained, it may be possible for
early medieval pottery to be mistaken for late
ancient, and for late medieval pottery to be
mistaken for early modern.
There is a clearly modern assemblage,
often associated with abandoned farmhouses,
including large beige water jars, red casseroles with
a very thin green glaze, beige plates, bowls, and
jars with a thin greenish or brownish glaze,
earthenwares with decorated opaque glazes, and
hard, well-fired, usually dark red curved roof tiles.
This assemblage probably dates to the late 19th and
20th centuries. Some of our collections resemble
this assemblage in some respects but also differ
substantially. These collections may be associated
with visible ruins, as at Site 10 (Torre Crisafi), but
often are not. Characteristics include red casseroles
decorated with white slip under a yellow glaze, jars
in a hard, finely granular reddish fabric, and curved
roof tile in a soft brown fabric containing large
chunks of gneiss and schist. In some cases, such as
Torre Crisafi, there is good reason to assign this
material to the early modern period, but in most
cases there is no evidence for an absolute date, and
some of it may extend back to the late Middle
Ages. Comparison with better-dated assemblages
from elsewhere will therefore be important for
establishing the chronological ranges of these types.
2.1.4. Results to date: sites found
We assigned eight new site numbers in
1999, bringing the total to 37 (Figure 1). We
recognize, however, that not all of these site
numbers represent comparable entities. They range
from a few scattered artifacts in a field all the way
to a large and dense concentration of finds such as
Mazza (Site 12). The issue is further complicated
by the different levels of material culture typical of
different periods: a thin scatter of artifacts which
would define a full-fledged prehistoric site might be
less than the general off-site “background” scatter
of artifacts in areas of intensive Roman land use.
As a first approximation toward a more
useful classification, we have divided them into two
categories: "sites", meaning locations where the
Bova Marina Archaeological Project 1999
12
artifact concentration is sufficiently obvious that
one can define its extent, and "scatters", meaning
areas or locations where artifacts are more
abundant than usual, but not abundant enough to
form a definite concentration.
The list here reviews all 37 locations to
which we have given site numbers so far. Note that,
for the sites defined in 1997 and 1998, it
incorporates updated information where available
from revisiting the location or restudying the
collections. Some sites have been redated or
reinterpreted, and hence this information
supersedes earlier descriptions.
Site 1 (Canturatta A, Area 2). Site, ca. 1
ha? Neolithic. This site, reported to us by S.
Stranges and L. Saccà , is located on the eastern
edge of the S. Pasquale valley, on the slopes just
below a steep ridge. It is in a field with some vines
and almond trees, and evidence for more extensive
vine cultivation in the past. The 1997 collections
from this area produced one Impressed Ware or
Stentinello sherd, and Stranges reports having
found several Stentinello sherds there. Test
excavations in 1999, described in detail elsewhere
in this report, failed to demonstrate the existence of
significant subsurface archaeological deposits.
Site 2 (Canturatta B, Areas 2, 50). Site, ca.
2 ha? Roman. This site is located on the mild slopes
on the eastern edge of the lower S. Pasquale valley,
just below a steep ridge. Abundant fragments of
Roman commonwares and tile were found,
especially in the northeastern corner of Area 50.
The collections need to be reexamined, but no
clearly datable diagnostics have been noted so far.
Site 3 (Pisciotta A, Area 5). Site, ca. 0.5
ha. Bronze Age. This site consists of a scatter of
impasto sherds found on a steep rocky hillside at
the eastern edge of the S. Pasquale valley. The
artifacts appear to be associated with a particular
level near the base of the hillside, are highly
localized, and include fairly large pieces of vessels.
This implies that the site is located on the slope
itself rather than consisting of slopewash from
above. The pottery includes pieces with carination,
high raised strap handles, or horizontal lug handles,
suggesting a Bronze Age date.
Site 4 (Pisciotta B, Area 6). Scatter.
Roman? This area is located near the southern end
of the top of the ridge overlooking the eastern side
of the S. Pasquale valley. A scatter of ancient tile
and pottery fragments, possibly Roman, was found
in this area, in contrast to adjoining areas, but the
quantity of material found is small enough that it
may be background scatter associated with a site
elsewhere.
Site 5 (Pisciotta C, Area 7). Site, 0.4 ha.
Bronze Age? This site, reported to us by S.
Stranges and L. Saccà, is located on a rocky hill
along the ridge to the east of the S. Pasquale valley.
A scatter of prehistoric pottery was found on the
peak of the hill and just below it on the western
side. Most of this pottery was nondiagnostic, but
one rim would be compatible with a Bronze Age or
Iron Age date, and previous collections described
by Saccà would support a Bronze Age date. If so,
like Site 3, it may be worth further investigation to
see how similar it is to recently discovered BA
deposits at the base of a rocky bank at Umbro (see
below).
Site 6 (Deri A, Areas 9, 10, 83). Site, ca. 4
ha. Roman. This site is just above the east bank of
the Fiumara di S. Pasquale near the sea. A rescue
excavation in advance of partially completed
highway construction several years ago found a
Late Roman synagogue as well as other structures
and some burials (Costamagna 1991). The principal
data from our survey consist of a series of 15
irregularly placed 10 m2 collection units in the field
just north of the intended highway bridge (Area 10)
as well as collections from transect-walking in
adjoining areas. In addition, a few diagnostic pieces
were selected from the disturbed area around the
construction. There is at least a thin scatter of
artifactual material throughout the entire field, but a
dense concentration only in the southwestern half
of Area 10. The pottery appears to cover a time
range at least from the 1st century BC to the 5th
century AD. Dated sherds include two rims of a
form resembling Campana A Lamboglia 36 with a
grayish brown slip (2nd to 1st century BC), a
fragment of a thin-walled cup decorated with
barbotine dots (late 1st century BC to early 2nd
century AD), Eastern Sigillata B Hayes 60 (80-150
AD), African Red Slip (ARS) Hayes 8A (75/90-
180/200 AD), ARS cookware Hayes 23B (150-220
AD), African amphora Keay 27B (300-450 AD),
ARS Hayes 50B (350-400+ AD), ARS Hayes 67
(350-450 AD), and ARS Hayes 84 small (440-500
AD). Additional finds include Campanian (black
sand) amphora, Late Roman Amphora 2, and a
marble mosaic tessera.
Site 7 (Deri B, Areas 9, 84). Site, ca. 1 ha.
Prehistoric. This site, which overlaps the northern
part of the Roman site at Deri, is a sparse scatter of
prehistoric artifacts in a cultivated field in the
bottom of the S. Pasquale valley. The finds include
a small amount of prehistoric impasto, and one
fragment of obsidian. There was also a piece of
worked flint nearby in Area 10. The only diagnostic
artifact was a horizontal lug handle. This site is
noteworthy for being a prehistoric site in a valley
bottom location, suggesting that landscape
alteration in the valleys has not been so total as to
Bova Marina Archaeological Project 1999
13
preclude all possibility of finding sites. It may be
the one erroneously called “Torre Varata” in Tinè
(1992).
Site 8 (Pisciotta D, Area 13). Scatter.
Historic. This is a sparse scatter of pottery found in
a small area about 30 meters in diameter on a
recently reforested slope. The artifacts are not
prehistoric, but have yet to be reexamined to
determine their date.
Site 9 (Umbro A, Area 16; Figure 2)).
Site, < 0.1 ha. Neolithic, Copper Age?, Bronze
Age, Roman. This site has been the principal focus
of the excavations in 1998 and 1999 and is
discussed in detail in elsewhere in this report. To
summarize briefly, this site, previously explored by
S. Stranges and briefly examined by Tinè (1992),
is located on a plateau near the border between
Bova Marina and Bova Superiore, where the
bedrock of calcareous sandstone projects out to
form steep cliffs. At the foot of one of the cliffs was
found a dense scatter of Neolithic pottery and
obsidian; this area was excavated in both 1998 and
1999. Two areas on top of this same cliff where
undated prehistoric pottery had been found on the
surface were also excavated, one in 1998 yielding a
poorly dated assemblage, possibly post-Neolithic,
and the other in 1999 yielding an assemblage
including parts of some Early Bronze Age whole
vessels. Abundant small pieces of human bone were
found partway up the cliff; although associated with
Neolithic pottery and obsidian, these were
subsequently dated by radiocarbon to the late
Roman period. The presence of Roman burials at
this location is somewhat enigmatic, because the
nearest site found as yet which may possibly have
Roman occupation is Site 33, about half a kilometer
away. However, Area 130, located about 250
meters to the southwest, yielded one fragment of
Italian sigillata, so there is some other evidence for
Roman activity in the area.
Site 10 (Torre Crisafi, Area 26). Site, ca.
0.5 ha. Modern. This site occupies the peak of the
promontory above Capo S. Giovanni. The central
part of the peak is now occupied by the shrine of
the Madonna del Mare. However, around the edges
of the peak and around the ruined coastal
watchtower, abundant pottery and tile fragments
were collected. Although this site was originally
considered both Classical and medieval in date,
part of the collection was reexamined in 1999 and
the site has been redated. It appears to be modern,
but probably earlier than the assemblage associated
with recent abandoned farms in Bova Marina. It
might be useful to compare this assemblage with
late medieval pottery as well. It is worth noting that
the tower at this point dates to the 16th
-17th
centuries.
Site 11 (Cimitero, Areas 28, 29). Site?
Historic. A large amount of pottery was found on
the slope west of the modern Cimitero S. Pietro,
overlooking the modern town of Bova Marina. A
small amount was also found at the top of the hill.
The artifacts are not prehistoric, but have yet to be
reexamined to determine their date.
Site 12 (Mazza, Areas 30-37, 102, 103,
150-153; Figure 3)). Site, ca. 8 ha. This site,
reported to us by S. Stranges and L. Saccà and the
location several years ago of a small test excavation
by L. Costamagna for the Soprintendenza
Archeologica della Calabria, was partially surveyed
in 1997. At that time the presence of a substantial
Greek site, presumably a village, was confirmed.
Restudy of the collections in 1999 showed the
presence of prehistoric and Late Roman pottery in
addition to a wide range of Greek pottery, from
early Archaic to Hellenistic. Greek pottery included
the base of an early Archaic cup with bichrome
decoration, numerous cup and krater pieces of the
6th to 5th centuries BC, Italiote transport amphoras
of the 5th to 4th centuries BC, and a possible
Corinthian “frying pan” of the Hellenistic period.
Late Roman material identified in the 1997
collections includes ARS Hayes 91C/D (6th to 7th
century AD), ARS Hayes 105 (late 6th to 7th
century AD), and Late Roman C Hayes 3C/D (late
5th century AD). In 1999, we did a small amount of
new fieldwalking (Areas 102 and 103) and
conducted intensive surface collections on a regular
grid over all the parts of the site to which we could
obtain access. These intensive collections are
reported in detail below. It should also be noted
that the site probably continues to the east, and
possibly to the north as well, but that access to that
area was blocked by fences in both 1997 and 1999.
Site 13 (Pisciotta E, Area 47). Site, ca. 0.5
ha. Roman. This site was reported to us as a
possible Roman kiln site by S. Stranges and L.
Saccà . It is located in a small ravine cut into a
slope on the eastern edge of the S. Pasquale valley.
On the slope north of the ravine, abundant tile and
commonware sherds were found. A drystone wall
foundation, possibly of a small rectangular
building, was observed in the eroding hillside. Two
rounded depressions in the bottom of the ravine
may be the features suggested to have been kilns.
Full reexamination of the collection remains to be
done, but a brief inspection showed a small but
diverse assemblage of Roman (and possibly also
Greek or medieval) commonwares, including a
fragment of African amphora. Surface collections
contain no evidence of pottery production such as
misfired wasters or a predominance of one or two
fabrics. Thus, it may be best interpreted as a typical
small Roman site, probably a farm. The round
Bova Marina Archaeological Project 1999
14
features of unknown date and function may or may
not be associated.
Site 14 (Panaghia A, Areas 51, 55). Site,
ca. 0.5 ha? Roman. This site, reported to us by S.
Stranges and L. Saccà , is located on the western
side of the middle part of the San Pasquale valley.
Collections from Area 51, done in 1997, include a
variety of Roman commonwares and tile, probably
predominantly early imperial, and a fragment of a
fineware with a bright orange slip (possibly either
ARS chiara A or Eastern Sigillata B). When the
site was revisited in 1998, a trench had been dug in
the site, apparently for construction, which had cut
through a Roman structure with tile roof, brick
walls, and opus signinum (cocciopesto) floor before
being halted. In the disturbed ground around this
trench we collected more Roman artifacts,
including the base of an ARS bowl (chiara D
fabric) with feather-rouletting on the interior, dating
to the 5th or 6th century AD. We made a series of
controlled collections in the field extending north
of this trench (Area 55). For the most part these
collections produced little or no Roman material,
except for a low-density scatter about 100 meters
away from the structure. Thus, the site is apparently
fairly small, probably a single farm. The area to the
east is inaccessible, being fenced-off orange groves,
but it may be worthwhile to investigate how far the
artifact scatter extends to the southwest.
Site 15 (Agrillei, Area 52). Site, < 1 ha.
Prehistoric. This site is located on a small
promontory at the southern end of the Agrillei ridge
near the border between Bova Marina and Palizzi
Marina, overlooking the sea. The presence of a
prehistoric site here was reported to us by L. Saccà.
Collections in 1997 and 1998 produced a
significant quantity of prehistoric impasto sherds,
mostly from the southern face of the hill, just below
the crest. None of the fragments appear to be useful
chronological indicators. The presence of a flat area
at the summit and buried terrace walls eroding out
of the slope suggests modern earth-moving, so the
finds may not be in their original context. Several
probably Greek or Roman sherds indicate the
possibility of a Greek or Roman component as well,
but the quantity is not enough to rule out their being
background scatter.
Site 16 (Umbro B, Area 58). Site, ca. 0.4
ha. Prehistoric. This site is on a hill about 150
meters north of Site 9 at Umbro. The main
concentration of artifacts is on a small saddle of
land along the top of the hill and along the upper
northern face of the hill. Collections in 1998
produced a large quantity of prehistoric impasto;
the assemblage differs somewhat from the Neolithic
pottery at Site 9, particularly in the presence of a
slip on many fragments, but no obvious
chronological diagnostics were found. It is likely
that the assemblage is post-Neolithic, but
excavation may be needed to determine a more
specific date.
Site 17 (Limaca, Area 66). Scatter.
Prehistoric. On the northern slope of a rounded hill
east of Umbro we collected one impasto sherd and
one piece of worked flint. The quantity is too small
to demonstrate the existence of any concentration
that could be called a site, but ground visibility was
poor, so it may be that additional items were
missed.
Site 18 (Umbro C, Areas 24, 68, 130).
Site, 0.2 ha. Greek. The main concentration of finds
is in Area 24, which is the sloping top of a small
rocky outcrop. A less dense scatter of artifacts
occurs also in the adjoining areas. The finds include
black-gloss pottery, probably of the late 6th to early
5th century BC, as well as Greek commonwares
and tile. The small but fairly dense sherd scatter
suggests something like a small single farm site. A
few Roman sherds have been found in nearby areas
as well; it is not clear whether these relate to reuse
of this location or to some other, as yet unidentified
site.
Site 19 (Penitenzeria, Area 72). Scatter.
Prehistoric?, Greek. This area is a large, gently
sloping field on the large plateau extending west
and south from the rock outcrops at Umbro. A
small scatter of artifacts was collected from the
southern edge of Area 72. This included four
probable fragments of prehistoric impasto, all
nondiagnostic, as well as a fragment of Greek
black-gloss and some Greek or Roman
commonwares. This location should be revisited to
assess whether there is a significant artifact
concentration, and if so to determine its extent and
obtain more evidence of its date.
Site 20 (Buccisa A, Area 76). Scatter.
Greek/Roman? This area is an olive grove on the
south face of M. Buccisa. At the southern edge, a
dense concentration of Greek or Roman tile was
found. No definitely ancient pottery was found in
association with this tile. Thus, it is uncertain
whether this tile scatter is the result of an ancient
structure in this location or of reuse for building
modern terrace walls. Even in the latter case,
however, it is likely that the site the tile came from
should be nearby.
Site 21 (Buccisa B, Area 79). Scatter.
Prehistoric. In a small, level area at the foot of the
southwest end of M. Buccisa, a small scatter of
pottery was found in a cultivated field. Most was
modern, but three fragments of prehistoric impasto,
all nondiagnostic, were also present. This is unusual
for such a small area, and although this small
Bova Marina Archaeological Project 1999
15
quantity could be background scatter, it is also
possible that there is a prehistoric site in this
location or nearby.
Site 22 (Sideroni/Amigdala, Areas 88, 89).
Site, > 1.5 ha. Prehistoric?, Roman, Medieval. The
site is located on the eastern bank of the Torrente
Sideroni, amidst the built-up area of modern Bova
Marina. Much of Area 88 is directly under a
highway overpass. Collections were made from
vacant lots in Area 89 and the northern part of Area
88 and from disturbed areas under the highway
overpass and near recent construction in Area 88.
These disturbed areas produced very abundant
finds in good condition; the vacant lots, despite
poor visibility due to grassy ground cover,
produced moderately abundant finds as well. The
total size of the site is impossible to estimate,
because only a few places are not currently built
over. No clear evidence was found regarding
whether the site should be considered as a village
or as a villa. The site was under excavation by the
Soprintendenza Archeologica della Calabria in
1999; it is hoped that their investigation will enable
better interpretation of the nature of the site's
occupation. The earliest well-dated sherds were a
fragment of a thin-walled cup, perhaps of the 1st
century AD, and the rim of a bowl in a gray fabric
covered with a dark gray slip, possibly a Sardinian
or Sicilian imitation of Campanian ware of the 2nd
or 1st century BC. The principal concentration of
finds ranges from the 2nd to the 5th century AD.
ARS forms include Hayes 23, Hayes 32/58, Hayes
52B, Hayes 196, Hayes 59, Hayes 61, Lamboglia
52B, and a small fragment of a decorated lamp.
Several pieces of African amphora were found,
including an Africana II spike and a Keay 62a rim.
Identified eastern Mediterranean imports include
Late Roman Amphoras 2 and 3. Some red-painted
commonware sherds of late ancient or early
medieval date were found. Early medieval
occupation was demonstrated by the presence of a
lamp of the so-called lucerna a ciabatta type, dated
approximately to the 8th century AD (Garcea and
Williams 1987; Ceci 1992). Four possibly
prehistoric impasto sherds were found in the
northern part of the site (Area 89), but they were
nondiagnostic.
Site 23 (Carusena, Area 91). Scatter. Two
prehistoric sherds were found in this small area in
1998. Given the size and location of the area this
was unexpected, so a site number was given to this
scatter. In 1999 part of the ridge just above Area 91
was walked (Area 148), yielding no prehistoric
artifacts. Although part could not be done at the
time, it is likely that there is no concentrated
prehistoric site in the vicinity, and that the minor
scatter in Area 91 is as much as will be found.
Although one of the sherds found in 1998 was a
rim, it did not appear to be chronologically
diagnostic.
Site 24 (Zaccaria, Area 96). Scatter.
Roman. This area is on the middle part of the slope
of M. Silipone, overlooking the valley of the
Torrente Sideroni. This part of the slope yielded
several fragments of ancient pottery and tile,
probably Roman. There was no clear concentration
anywhere, but the quantity is much greater than in
the surrounding areas. The quantity is not enough to
rule out being background scatter related to an as
yet unlocated site, but because the area is located
on a soft clay slope, it is also possible that a small
site here has been largely obscured by
geomorphological processes.
Site 25 (Sant'Aniceto, Area 98). Site, ca.
0.6 ha. Prehistoric (Final Bronze Age or Iron
Age?), Roman/Medieval?, Modern. This site is
located at the top of a very steep hill overlooking
the valley of the Torrente Vena. The presence of a
medieval church on this hill was reported to us by
S. Stranges and L. Saccà. A difficult and time-
consuming climb is required to get to the top of the
hill. The most obvious feature at the top is a small
ruined church built of mortared stone masonry,
approximately 9-10 meters square with an apse on
the eastern side. Some other traces of wall
foundations are visible several meters to the north.
The predominant finds include red straw-tempered
roof tile, red earthenware with a thick green glaze,
light-colored plainware with reddish sherd temper,
and light-colored to orange-brown gritty
coarseware. The last of these resembles Roman
coarsewares, but no datable finewares or imports
were found. The others are unusual in our
collections and appear most likely to be early
modern, or possibly late medieval, although no
recognizably medieval types were found. On the
present evidence the church is most likely to be
modern, although an earlier date cannot be ruled
out. In the eroding slope just west of the church a
somewhat different pottery assemblage was found,
consisting mostly of impasto, probably of
protohistoric date. Three decorated pieces of coarse
pottery were found, one with a wide band of dark
brown paint and two with finger-impressed applied
cordons; these could be of either protohistoric (Iron
Age?) or medieval date.
Site 26 (Climarda, Area 99). Site,
unknown extent. Greek?, Roman. This area consists
of the fields at the foot of the south end of the hill
on which Site 25 is located, as well as the banks of
the Torrente Vena running through these fields.
Most of the finds in this area were modern,
probably related to an abandoned farmhouse at the
foot of the hill. In the eroding river banks, however,
and archaeological stratum could be seen about 1 to
Bova Marina Archaeological Project 1999
16
1.5 meters below the present ground surface. From
the banks and the stream bed, and also in small
quantities from the fields, we recovered numerous
fragments of ancient pottery and tile, including
black-gloss (Greek or Republican), ARS form
Hayes 96 (early 6th century AD), and the spike of a
Late Roman African amphora. It is not clear from
the small collection made in 1998 whether the
apparent hiatus between the Greek/Republican
occupation and the Late Roman is real or not. Due
to the depth of the overburden it is difficult to
determine the extent of the site from the surface
evidence, but the depth of the deposit also suggests
that this may be a rare site with relatively
undisturbed archaeological contexts. Thus, it may
merit further archaeological testing, at least to the
extent of cleaning the bank profiles.
Site 27 (Vadicamo, Area 100). Site, < 0.01
ha. Roman/Medieval. Area 100 is a long, badly
eroded ridge of fine clay partially reforested with
eucalyptus trees. At the northwest end of the ridge,
a small gravely patch that looked like a possible
relict deflation surface had a concentration of ten
sherds of late Roman or early Medieval
commonware and one fragment of Roman tile. The
sherds all appear to be of the same vessel, a jar of a
form known from other late Roman and early
Medieval sites. This unusual concentration in such
a small area, and especially the combination of tile
with the remains of what may have been a whole
vessel, suggests that this may have been a late
Roman or early Medieval burial. If so, however, it
has entirely eroded away, because the artifacts were
all found resting on the apparent deflation surface.
Site 28 (Papagallo, Area 101). Site, < 1
ha? Prehistoric, Greek/Roman. Area 101 is located
in an eroding badlands formation on the east slope
of a clay ridge near the border between Bova
Marina and Condofuri Marina. From the relatively
small area of this slope that we investigated, we
collected several fragments of prehistoric impasto
and Greek or Roman commonware. The prehistoric
pottery appeared to be later than the Neolithic
assemblage at Umbro, but did not include any clear
chronological diagnostics. The Greek or Roman
pottery needs to be reexamined to determine its
date. This area should be revisited to determine the
extent of the site and to see whether contexts can be
identified in the eroding banks from which the
artifacts are eroding.
Site 29 (Pisciotta F, Areas 8, 80, 81, 82,
87). Scatter. Roman. The fields on the mildly
sloping eastern side of the S. Pasquale valley, just
to the east of the Torrente Turdari, contain a higher
than usual frequency of Roman pottery. There does
not appear to be a clear concentration anywhere
within these areas, however, so it may all be a
diffuse scatter associated with Site 2 or 13.
Site 30 (Umbro D, Area 130). Site, ca. 0.6
ha. Neolithic?, Bronze Age? This site, reported to
us by S. Stranges, is located on the edge of an area
at Umbro partially surveyed in 1998. It is a long,
narrow scatter of artifacts along the southern edge
of a plateau area at Umbro, where the slope begins
to become broken and steeper. The finds are mostly
nondiagnostic fragments of prehistoric pottery,
along with one fragment of worked flint. The
majority of the pottery resembles the Neolithic
pottery from the excavations at Umbro in fabric,
and one fragment appears to have impressed
decoration, but some are somewhat finer and may
be later. One of these may be a fragment of a horn-
shaped handle or some similar projecting element.
Another rim, though of coarse fabric, appears to
have the attachment for a strap handle projecting
above the rim.
Site 31 (Umbro E, Area 131). Site, ca.
0.25 ha. Neolithic? This site is located on the
central plateau at Umbro, in an area not surveyed
previously. It is at the southern edge of the central
plateau, at the top of some low cliffs overlooking
the field containing Site 30. A scatter of prehistoric
pottery (plus one piece of obsidian) was found
along the top of the cliff. No useful diagnostic
fragments were found, but the fabrics generally
resemble the Neolithic pottery from the excavations
at Umbro. That, together with the obsidian, makes a
Neolithic date most likely for this site.
Site 32 (Marcasita, Areas 136, 137). Site,
ca. 0.8 ha. Neolithic?, Roman. This site is at the
end of a low hill projecting into the eastern side of
the S. Pasquale valley, overlooking the floodplain.
It is located directly across the river bottom from
Site 14 (Panaghia). It comprises a fairly dense and
well-defined scatter of brick, tile, and pottery in
two clusters on the northeast and southwest sides of
the ridge. Some of the brick and tile are modern,
related to an abandoned farmhouse just to the west,
but some are clearly Roman. Roman imports
include one fragment of Campanian (black sand)
amphora and several nondiagnostic fragments of
African Red Slip ware, African plainware, and
African amphora. The Roman assemblage does not
contain anything that can be dated very closely but
would support a date range from the 1st century AD
(possibly somewhat earlier) to the 4th century AD
or later. The size and density of the Roman
component conform to what would be expected of a
single farm site. The site also yielded a fair amount
of prehistoric pottery, in fabrics resembling the
Neolithic impasto and figulina from the
excavations at Umbro. One fragment may have
traces of impressed decoration. Thus, the
Bova Marina Archaeological Project 1999
17
prehistoric assemblage may be of Neolithic date.
The prehistoric pottery is less eroded than usual for
surface assemblages in Bova Marina, which may
mean that plowing has recently cut into or is
currently cutting into the archaeological levels.
Site 33 (Cromidi, Area 117). Site, < 1 ha?
Greek, Roman? A small cluster of ancient pottery
and tile, including one fragment of Greek black-
gloss. It is located about 500 meters northeast of
the Neolithic site at Umbro, in the valley by the
foot of the cliffs of M. Buccisa, near some springs.
The surrounding parts of Area 117 also yielded two
fragments of African Red Slip ware, as well as a
fair amount of Greek and Roman tile. Thus, it is
likely that there was a small occupation, possibly a
farm or a seasonal fieldhouse, used in both Greek
and Roman periods.
Site 34 (Vunemo, Area 123). Site, < 0.1
ha? Roman? A small, fairly level area of vines and
fruit trees was the only part of Area 123 to yield
artifacts in any quantity. A concentration of
artifacts in this one location included Greek or
Roman coarse pottery and tile. No obvious
chronological diagnostics were found, but the
overall character of the assemblage may suggest an
early Roman date.
Site 35 (Cecilia, Area 104). Scatter.
Roman. This consists of a small quantity of artifacts
found during fieldwalking on the north face of a hill
along the border between Bova Marina and Palizzi
Marina. A large piece of Roman tegula was found
on the slope below the crest, and a few scraps of
pottery, some of which may be Roman, were found
near the top. A stone weight for an olive press,
possibly Late Roman or later in date, was found
built into the wall of a recent fieldhouse at the base
of the hill to the north.
Site 36 (Panaghia B, Area 143). Scatter.
Roman. This area, on the terraces of the western
side of the S. Pasquale valley, produced a diffuse
scatter of Roman pottery and tile, none of it clearly
diagnostic. There is no clear concentration here or
in the slopes above this area, so although it could
be a small, low-density site, it is at least as likely to
be dense background scatter, possibly related to the
Roman site at Panaghia (Site 14).
Site 37 (Panaghia C, Area 145). Scatter.
Greek, Roman? This area, on a slope overlooking
the western side of the S. Pasquale valley, produced
a diffuse scatter of pottery and tile, mostly modern
(related to an abandoned farmhouse at the upper
end of the area) but also including a few pieces of
Greek or Roman coarseware and tile. No
concentration of ancient artifacts was observed, so
this is likely to be background scatter related to a
site located elsewhere, possibly the Roman site at
Panaghia (Site 14).
2.1.4. Discussion of survey results
It may be useful at this point to give a
brief review of what we have learned so far from
our survey.
We have not yet found any evidence of
Paleolithic activity. Given sporadic Paleolithic
finds elsewhere in the area (at Torre Mozza and
Gunì in Palizzi Marina, for instance; S. Stranges
and L. Saccà, pers. comm.) it is unlikely that Bova
Marina was simply uninhabited before the
Neolithic. Paleolithic settlements are likely to be
archaeologically unobtrusive, and coastal sites are
probably now under water due to rising sea levels at
the beginning of the Holocene. However, the lack
of Paleolithic sites may also be the result of
geomorphological processes having eliminated
older ground surfaces in the study area. It is also
possible that we have failed to recognize them
adequately: our survey has recovered very few
chipped stone artifacts so far, possibly due to our
reliance on inexperienced crews. As the number of
later prehistoric sites grows, our understanding of
site location improves.
Neolithic sites seem to occur in a wide
variety of settings, including coastal valley terrace
(Site 1), low hill overlooking a river valley (Site
32), and inland plateau (Sites 9 and 31). Later
prehistoric periods are more difficult to distinguish
(see discussion of ceramic chronology above), but a
similar diversity is likely.
In the Classical periods, Greek settlement
includes a large village (Site 12) as well as small
farmsteads (Sites 18 and 33). As with the
prehistoric sites, however, most are located on high
ground, although some exceptions occur (Site 33).
In the Roman period for the first time numerous
sites are known from valley locations (Sites 2, 6,
14, 22, and 26), suggesting a shift toward more
intensive agricultural use of valley bottoms. The
ceramic evidence for the Greek and Roman periods
shows two peaks: one in late Archaic to Classical
times (6th to 5th centuries BC) and one in late
imperial times (3rd to 5th centuries AD). Late
ancient settlement seems to continue in diminishing
abundance but in the same locations as Roman
settlement, to the 8th century or so (Site 22). After
that, medieval settlement remains obscure, but by
the 19th century abundant rural settlement again
existed throughout Bova Marina, like in Greek and
Roman times.
In relation to the goals we had set for the
1999 season, we made reasonable progress given
Bova Marina Archaeological Project 1999
18
the constraints of such a short field season. The
detailed evidence from Mazza is not only likely to
be important for understanding Greek settlement
patterns in the region, it also revealed significant
prehistoric and Roman occupations, adding to our
knowledge of those periods. The size of this site
compared to others, however, and the anomalously
high artifact density at the highest part of the hill,
raises many questions about the internal structure of
the site. Geophysical prospection, for example by
resistivity/conductivity or magnetometer survey,
may help explain the differences in Greek and
Roman settlement patterns, by allowing us to
compare sites in terms of number and arrangement
of recognizable structures.
The recognition of a number of small,
rural Greek sites in the study area, a considerable
distance from any Greek town, is a valuable
addition to the understanding of Greek colonization
derived from the study of urban sites such as Locri
and their immediate surroundings as at Metaponto.
The early date of some of these sites is somewhat
surprising; it indicates that Greek settlement spread
very widely through the colonial areas within a few
generations. What remains unclear is the function
of these sites, and their relationship on the one hand
to the towns and on the other to the native
population. Excavation at a small, early Greek site
such as Site 18 may provide evidence that can be
compared to that from urban sites such as Locri in
order to address such questions.
The additional evidence from higher,
interior locations (M. Rotonda, M. Vunemo) is a
valuable complement to our evidence from Umbro.
In particular, it shows two things. First, Umbro is
unusual among the higher locations in Bova
Marina, in the amount of prehistoric settlement to
be found there. Although both M. Rotonda and M.
Vunemo produced some archaeological evidence,
nothing like the cluster of prehistoric sites at
Umbro was found. Second, even in highland areas
that appear fairly barren at present, Greek and
Roman settlement and activity were present to some
degree. This suggests a problem with the present
definition of our survey. In early modern times,
Bova Superiore and Bova Marina formed a single
unit, and it is likely that agricultural practices of the
time made use of land at all elevations within this
extended territory. There is no reason to suppose
that the prehistoric, Greek, and Roman periods
were different; interior areas in Bova Superiore
were probably used by the same people using parts
of Bova Marina, for cultivation of grain or other
crops in the lower parts, for haying or pasture at
higher altitudes, and even higher places could have
been used for other purposes such as hunting,
wood-cutting, charcoal-burning, or mineral
extraction. Moreover, since we know that the
principal focus of medieval settlement in the region
was the interior town of Bova, the apparent absence
of medieval settlement from our survey may be a
result of working too near the coast. For these
reasons, a survey concentrated entirely on the lower
areas within the borders of modern Bova Marina is
incomplete, and runs the danger of missing
important parts of past settlement systems. This
danger is greatest for the warmer periods of the
past, when arable cultivation could have extended
higher than in recent times; such periods may
include important parts of the Bronze Age and the
Middle Ages.
The problems with the chronological
sequence remain worrisome. It appears likely that a
combination of continued survey and statistical
seriation of our assemblages will improve the
resolution for the Greek and Roman periods,
already the best-known part of the sequence.
Expanding the survey to higher areas near Bova
Superiore or Amendolea Vecchia may improve our
knowledge of the medieval part of the sequence.
The problems with the prehistoric sequence will
require a different approach, however. Apart from
chance finds of freshly disturbed prehistoric sites,
an unlikely occurrence, survey is not likely to
produce assemblages in good enough condition to
allow systematic comparison of the attributes used
in traditional classifications with those more readily
visible in our surface collections, such as paste,
surface treatment, and firing. Surface collections
also lack secure archaeological contexts, so one
does not know whether they represent an umixed
assemblage and one cannot associate them with
absolute dates. What may be required, instead, is a
program of small test excavations at selected
prehistoric sites, to obtain samples of less eroded
artifacts from better-defined contexts, possibly in
association with materials suitable for radiocarbon
dating. Such excavations, carefully targeted at sites
likely to have intact deposits of enigmatic date,
would not only enable the excavated sites to be
dated, but would improve the precision of the
dating of our survey in general.
Finally, we were asked by the
Soprintendenza Archeologica della Calabria to
watch for archaeological sites at imminent risk of
destruction due to human activity. While numerous
sites are subject to destructive geomorphological
processes such as erosion and human practices such
as plowing, we observed only two sites actually
threatened by active construction. One of these is
Panaghia, an inland Roman villa or farmstead site
where digging foundations for new houses had
destroyed part of the site already. Construction has
now been stopped at Panaghia following legal
action by the Soprintendenza. The other site is
Siderone/ Amigdala, located beneath the SS106
Bova Marina Archaeological Project 1999
19
overpass on the eastern side of the modern town of
Bova Marina. This Roman villa or village site has
been largely covered by modern building, and parts
still preserved in gardens are currently being built
upon. Again, the Soprintendenza has taken legal
action to halt building and has also begun rescue
excavations.
2.2. Intensive Collections at Mazza (Lin
Foxhall)
Controlled surface collections were
carried out over four days on site 12, Mazza, to
determine the geographical limits, the
chronological range and patterning, and the internal
spatial organisation of this large and complex
historic-period site.
A baseline was established along the fence
near the summit of the site. Using a compass, tapes
and ranging rods, a grid oriented along cardinal
points was plotted over the accessible areas of the
site with stations marked with chaining pins every
30 m. Each station was assigned a number and a
letter designating its position within the grid, and
each was then used as the centre point for a circle
10m2 in area. Circles were demarcated using two
attached to a piece of string 1.78 m in length. One
nail was held at the station point while the other
was used to inscribe the circumference of a circle
with a 1.78 m radius. All artefacts were collected
from within each of these areas. Part of the site,
near the (probably) early modern-modern structures
previously identified as a ‘temple’, is fenced off so
it was not possible to include this area in our
intensive systematic survey. However, the intensive
survey area did cover the section of the site with
modern (19th century?) farmhouse buildings, just
south of the summit. A total of 105 small
collections were made (Figure 3, Figure 4).
Spatial data is summarized in a set of
distribution maps of selected artifact categories.
Each map depicts the number of fragments of that
category collected at each location; collection units
with no items of that category are left blank. These
numbers are based on preliminary sorting in 1999
and may be subject to revision. A few problems
with the data which are relevant to reading these
maps should be noted. First, one bag, from
collection H8, became separated from its contents
while the finds were being washed. These artifacts
presumably became mixed with those of one or
more other bags. Second, an assortment of pottery
and tile became mixed with the contents of a
different bag, which originally contained samples
of metal-smelting debris. This group of pottery and
tile fragments is shown in the circle in the lower
part of each map. The quantity and types resemble
those of the units around H8, but it is not known
whether these are in fact some or all of the missing
finds. Finally, several units (L6 to L12) were
incorrectly placed too far to the east. These were
redone once the error was noted, but the initial
mistaken collections are included on the maps
between K and L.
Preliminary analysis of the material
suggests that temporal and spatial patterns are
discernable in the controlled surface collections.
Very large amounts of ancient roof tile were found
at the highest part of the site, smaller quantities on
the rest of the upper areas and the SE slopes, and
much less around the edges. Natural rock outcrops
in the central portion of the site S and SW of the
summit (Area 151), look similar to those used for
house foundations at Locri Epizephyrii, and might
perhaps have served the same purpose. The
northern slopes of the site showed low level artefact
scatter almost to the gully. On the southern and
southeastern slopes artifact scatter continued as far
as we could survey to the edges of the steep slopes,
though density levels decreased dramatically
toward the edges. The southern end of the site
appears to have a concentration of smelting/metal
working remains. Burnt daub and furnace slag were
collected and a large piece of probably furnace slag
too large to collect was photographed. Slag samples
were taken to Leicester University for analysis.
Ceramic data has not yet been fully
analysed, but preliminary results suggest a
preponderance among the Greek fine-wares of cups
and krater-like shapes. Pottery is most abundant on
the eastern, southern, and southwestern parts of the
plateau, as well as on the southeastern terrace
below; an area of at least 8 hectares in all. The
artifact scatter is variable in density, but the only
major anomaly is the dense concentration of tile at
the summit. Greek pottery, as represented by black-
gloss wares, occurs throughout this area. Roman
pottery, as represented by African imports, occurs
mostly in the eastern and southeastern parts of the
site. Prehistoric pottery is most abundant in the
southern and southwestern parts of the site, which
is also the area where the metal-smelting evidence
occurred.
To conclude, intensive collection at Mazza
has showed that there is both chronological
patterning within the site, with Greek remains
everywhere and an unsuspected Roman component
concentrated near the highest part of the site. It has
also shown that there is functional differentiation
within the site, with possible residential and
industrial areas identifiable. Further analysis of this
data should clarify these patterns.
Bova Marina Archaeological Project 1999
20
2.3. Geological Reconnaissance
Prehistoric people made use of a range of
geological raw materials; archaeologically
demonstrable examples include clay, flint, obsidian,
fossil shark teeth, hard stones for polished axes, and
stone for querns, mills and other ground stone tools.
The sources of some of these are known. Obsidian
used at Umbro generally came from Lipari (R.
Tykot, pers. comm.) and some grey flint came from
outcrops in the Saracena area of Condofuri Marina.
The polished stone axe found at Umbro in 1999
(see below) was made from amphibolite from the
Sila (P. Barrier, pers. comm.), and the miniature
axe replica was made from a softer phyllite schist
available locally. The source of other materials
remains unknown. Petrographic analysis has shown
that pots probably made at Umbro were not made
from clay from the two sources nearest to the site
(see below), and we have little idea where the red,
brown, and yellow flint used at Umbro came from.
Hard stone for making axes has been the
object of a number of sourcing studies. Throughout
Southern Italy and Sicily, prehistoric people used
granite, diorites, serpentine and other greenstones,
amphibolite, and other stones for axes, adzes, and
similar tools. Often greenstones transported from
long distances were used for smaller or finer tools,
particularly amulets, while amphibolite and diorite
were used for less fine axes and more resistant and
commoner stones such as granite were used large
pebble tools (Leighton 1999; O’Hare 1990). These
studies make clear both that hard stones were
transported over considerable distances, with some
coming ultimately from the Alps, and that materials
were often carefully matched with a tool’s intended
form and use. It is also probable that the trade in
stone and axes not only carried on to supply useful
tools but was socially and symbolically important
as well.
With these considerations in mind, we
decided to investigate the availability of hard stones
in the Bova Marina region. If appropriate raw
materials were available, could they have been an
economic resource for local axe production or for
trade? If raw materials were not available locally,
were they available inland in the highlands of
Aspromonte? If so, what were the implications for
trade? If raw materials were available and were not
used, why would this have been so?
The geology of Aspromonte is highly
varied, but the potential sources of polishable hard
stone seem limited. The territory of Bova Marina
itself, like much of the coastal strip, consists of
harder and softer sedimentary stones (limestones,
sandstones, clays, shales) and of metamorphic
rocks including predominantly phyllites, schists and
gneisses. The latter are often too soft or too riven
by cleavage planes to be used for tools. Harder and
finer grained metamorphic and igneous stones
occur at higher altitudes 10-20 km inland; these
include granites, diorites, and possibly other stones
(Servizio Geologico dell’Italia 1970; 1976).
Given this, hard stones could have been
available locally in two ways. First, fragments
could have washed down from the highlands in
river valleys. Secondly, a band of relatively soft
sandstone conglomerate stretches across the
territory at about 3-6 km inland. As road cuts in the
Fiumara di Amendolea valley, Monte Bucissa, and
alnog the Bova Superiore road show, this
conglomerate contains strata dense with cobbles of
varied stones which must have originally come
from older strata higher inland. A cursory
inspection on Monte Bucissa showed that most of
these were granites and gneisses, but they also
contain some unidentified close-grained stones
which may have been suitable for tool-making. The
only crystalline material was a coarse-grained and
fault-ridden quartzite. These conglomerate strata
would have provided another potential source of
stone for tools.
The logical strategy for assessing raw
material availability, therefore, was through
surveying river valleys where both recently
transported stone and cobbles eroding from
conglomerates would be available. In addition to
recording raw materials present, we hoped to
collect data on how much stone of each type was
available, to assess the “richness” of potential
source zones quantitatively. As a pilot experiment,
we recorded stone resources at tweve locations
along the Fiumara di Amendolea. These were
spaced approximately a kilometer apart, extending
inland from the coast to above the ancient castle of
Amendolea. At each point, the crew laid out a
circle enclosing ten square meters, and collected all
cobbles larger than 15 cm long by 10 cm high by 5
cm wide. They then matched these to a type
collection they carried with them and tallied how
many cobbles of each type of stone were present.
When a new type of stone was found, a fragment
was added to the collection. Although cumbersome,
this procedure allowed survey to be carried out
without a trained geologist present on the crew and
without collecting and transporting hundreds of
kilos of cobbles. The type collection was retained
for later identification of samples by a trained
petrologist.
Results are pending at this time. However,
it is clear from this exercise that plentiful cobbles
of an appropriate size and shape are available in the
major river valleys, and at least some of them are
Bova Marina Archaeological Project 1999
21
hard, potentially usable stones. It is also clear that
different points in river valleys varied in their raw
material “richness;” this may have had implications
for landscape use and social rights of access.
2.4. GIS Analysis of Prehistoric Land Use
(Doortje Van Hove)
Much archaeological research focuses
upon human settlements, ceremonial sites, and the
material that is found within them rather than on the
spaces in between. This is because ‘off-site
archaeology’ (Foley 1981) provides poorly studied
material to work with and immense methodological
difficulties. However, humans live in a world
composed of many different kinds of places, so
restricting our view of prehistoric people to discrete
sites will inevitably and unnecessarily narrow our
understanding of how they lived. In emphasising
the fuller landscape, this research highlights human
usage of land, one of the key means by which
humans categorise and lay claim to their
surroundings.
This research is aimed at understanding
how people used the landscape in prehistoric
Southern Italy through the use of computer GIS
(Geographical Information Systems). The overall
goal is to create a simulated environment
representing one possible situation in which
Southern Italians could have lived throughout the
different periods of human occupation of the area.
Particular attention is paid to land use and resource
exploitation within the possible economies of each
period, and the possible cultural categorisation and
social use of off-site areas.
To investigate the scope of the research, a
pilot study was done which focussed on a very
limited area during the Neolithic period. This
preliminary research has drawn together elements
of archaeology, human psychology, geography and
simulation modelling to provide novel insights into
human-landscape relations. The research simulated
environments and human land uses around three
Neolithic sites: Castello Bova Superiore, Umbro
and San Pasquale (see Figure 5). The quantities of
archaeological and environmental data and the
spatial dimensions of the area examined were very
limited. This prototype allowed developments to
be made in the techniques and interpretative
frameworks, before applying the methodology on a
larger scale.
GIS modelling, based upon relevant data
from IGM (Istituto Geografico Militare)
topographic and geological maps, and limited field
survey site distributions from the Bova Marina
Archaeological Project, was used. Environmental
relations were modelled using current and historical
Mediterranean geographical and ecological
information, and comparative studies from
Mediterranean prehistory. In the broader area
around Umbro, incorporation of different kinds of
environments allowed us to model the travel
distances and sizes of specific activity-zones of the
Neolithic cultures in the region.
As basis for the analysis, topographic and
geological maps were digitised. The base map used
was the IGM 1:25000 map quadrant for Bova
Marina. Eight categories of environment were
defined on the basis of the elevation, slope, river
and geology maps. For this purpose, parameters
were established on the basis of literature,
assumptions and observations of the terrain in order
to represent and evaluate the present and past
landscape. General map algebra was used, which
involves calculations with different categories of
various maps to generate a new map, which can
give novel insights into a newly generated
environment. The eight different environments are
coastal plain, river channel, river valley, clay
terrace, sedimentary plateau, sedimentary hill,
metamorphic plateau and metamorphic hill.
To evaluate the potential usefulness of
each zone on the map and its possible utility for
humans, a combination of yield values and
accessibility had to be attributed to each
environmental niche. Thus, the environment
reconstruction map was reclassified to hold
estimated resource yield values. These valued were
estimated based on the usefulness of each
environmental zone to a specific kind of Neolithic
economy. Four kinds of possible economy in the
Neolithic have been selected for the analysis:
foraging, mixed foraging-farming, subsistence
horticulture and intensive pastoralism.
In assessing how people will use their
environment, not only the qualitative yield but also
the accessibility of the area plays a significant role.
For each site, a cost-surface model was built to
assess the possible access to areas around the
different sites. Therefore, a friction surface was
created based upon a valid and plausible
assumption of the cost traversing a sloping
landscape. The area is quite hilly, and therefore it
was considered that a given distance will be
traversed with different time and effort costs on
level or on sloping ground. In addition, walking
uphill or downhill will affect time-intervals as well.
It becomes increasingly more difficult and time
consuming as the slope increases. To understand
the values of the cost-surface model, a calibration
was made, based on a plausible assumption about
the landscape. Five classes of ‘distance’ or ‘cost’
were defined to calculate what it takes to get from
Bova Marina Archaeological Project 1999
22
each site to the different environmental zones
defined earlier.
Combining yield and accessibility values,
three kinds of dynamic simulations were run on a
fixed population-size (50 people) for purposes of
simplicity. Each simulation can be considered as a
sequence of events. The first simulation shows how
far people need to travel before reaching enough
land for their particular economies. The assessment
of the size of land needed per population and
economy was based on literature, especially the
work of Susan Gregg (Gregg 1988). The value of
specific land use and its land size desired will
depend on the economy used. Farmland will be
more important in the horticulturalist and mixed
farming-foraging economies than in the pastoralist
and forager ones. For foraging a larger territory is
required, although it does not have the same impact
on the land. For herding, a slightly larger terrain
will also be necessary as animal husbandry is more
mobile than farming. This will have an effect on
the time and distance needed to cover to achieve
the suitable land size for the amount of people
concerned.
A second kind of simulation is
vegetational growth. One starts from a set of empty
maps, which all grow a particular kind of
vegetation during a fixed time period. After the
vegetation has iterated for a particular amount of
time, the simulation stops and assumptions can be
made about when a vegetation type reaches its
climax state as each type grows at its own rate. It
was assumed that wet plants will grow faster than
plants on heavy soil, depending on the soil texture,
temperature and precipitation. Beach vegetation is
conditioned to withstand a lot of extreme climatic
and edaphic conditions and will grow at an average
rate. Varied plants were given a rate of growth in
between the other types of vegetation.
For the purpose of simplicity, both
previous simulations were combined. First, the
natural vegetation grows for a fixed amount of
years until it is mature enough to let people use it
for their specific purposes. Humans then use the
land for farming, herding or foraging for a certain
period, after which these are abandoned and the
natural vegetation can grow back slowly. Because
of the different impacts that foraging, farming and
herding can have on the environment, different
parameters were used in the regeneration of the
vegetation. This is because land that has been used
for herding and agriculture is a lot more degraded
than when it was used for foraging and it will take
longer for natural vegetation to grow back. Hunter-
gatherers tend not to exhaust their environments in
order to reserve the possibility of returning to them
after several seasons. The result is a total of 36
maps, defining land use per site, economy and
specific type of impact on the environment. Some
examples are displayed as Figure 5.
When simulating these land use maps,
some basic assumptions were used, which applied
to the yield and accessibility. For people to use a
particular landscape zone, the yield has to be
sufficiently high in order to be selected. In
addition, it is assumed that foraging people will
travel further than those herding and farming.
Therefore the cost assumptions are different per
specific land use. This will influence the land use
maps and is interesting in comparative studies. It
was also moderately based upon literature about
site catchment analysis (Higgs and Vita-Finzi 1972:
27-36; Jarman, Bailey and Jarman 1982: 26-46).
The analysis thus resulted in maps
illustrating anthropogenic environmental impact
and potential spatial dimensions of the cultural
world that humans lived in. These results have two
basic implications. One is economic and ecological:
human landscape use and environmental impact can
be examined in a systematic way, in contrast to
isolated single site-based studies. The second is
cultural: humans use the landscape, including ‘off-
site’ areas for many other non-economic uses.
Many of these uses are conditioned by the existent
human use of those areas.
In this regard, it is interesting to observe
that sites are preferably ecotonally located,
enabling the exploitation of two or even three
resource zones: cultivation on a plateau, some
pasture for grazing and a possibility for hunting and
gathering in wooded alluvial margins.
The primary result of the GIS study to date
is to underline the importance of non-farming land
uses for site distribution. Instead of the commonly
held view that competition for farmland was of
greatest importance in farming societies, the
simulation showed that even in these, human
interaction was probably structured by non-farming
uses such as the needs of foraging and pasture. The
sizes of the different specific land uses such as for
foraging, mixed foraging-farming, farming and
herding show a different picture. In the region
under consideration, farmland is stretched out in
small patches around the sites, while foraging and
herding zones are continuous and large. Neolithic
groups were probably more mobile, using bigger
territories and moving frequently within them, than
is usually assumed (e.g. (Higgs and Vita-Finzi
1972). This means that the distribution of sites,
even in mountainous areas like Calabria, should
not be limited to available farmland. Observing the
contacts between the areas of human impact
associated with different sites enables inference of
possible elements of South Italian social and
Bova Marina Archaeological Project 1999
23
cultural behaviour. These might include exchange
and competition.
GIS study of prehistoric land use in
Calabria will be continued as part of an ongoing
PhD thesis. Future developments include
comparisons with actual land use at each site, as
indicated by archaeological data, which could
further support studies of site potential and the
development of an environmental and economical
site-type model for Southern Italy. Other future
possible adaptations of the analysis and approach
will seek to highlight a more pluralistic view on
Neolithic land use. These would implement a
higher level of detail and environmental data such
as water resources, soil types and economic
resource needs (biomass, nutritional components of
animals and plants). An expanded site database
would also be introduced. In addition, different
parameters considering human resource needs can
be implemented and compared to see whether they
make a significant difference in land use around
each of the analysed sites. Finally, ongoing research
will address current problems representing the
temporal dimension in modelling. In part, it will
examine population shifts across the landscape in
response to resource exhaustion and the differences
of human land use during the whole period of
occupation of this area.
Bova Marina Archaeological Project 1999
24
3. THE EXCAVATIONS AT UMBRO
3.1. Introduction: previous work, goals and
methods for this season
Umbro is one of several names for a
calcareous sandstone and limestone plateau about a
kilometer in diameter located along the unpaved
Bova Marina – Bova Superiore road just south of
the border between the two comunes (Figure 7).
The plateau consists of sloping fields between 360
and 400 meters above sea level, surrounded on all
sides by sporadic, interrupted cliffs 10-20 meters
high. On some sides, the cliffs were probably
penetrated by shallow caves in the past; in other
zones they afford vertical sheltering walls. To the
east and south underlying, impermeable clay beds
are exposed, and there may have been springs at the
foot of the cliffs here in the past.
The prehistoric site at Umbro is located
along the east margin of the plateau (Figure 6,
Figure 7). It has been the subject of about a decade
of investigation before the work described here (see
Robb 1997; 1998; Stranges 1992; Stranges and
Saccà 1994; Tinè 1992). Stranges, Saccà and co-
workers have made periodic informal surface
collections since their original discovery of the
Neolithic site in the early 1990s. Tinè conducted a
walkover of the site in 1992, and dug a shallow,
.5m square test pit, unfortunately in a sterile area of
the site. The site was re-surveyed by the Bova
Marina Archaeological Project in 1997 and 1998.
Finally, the only excavations at the site were
conducted in 1998 by the Bova Marina
Archaeological Project. These excavations put a
one by two meter sondage in the main Neolithic
area (Trench 1), excavated a one by one meter
sondage along the south slopes of the cliffs (Trench
3) and on the valley floor just below Trench 1
(Trench 5), and an area excavation atop the cliffs
which revealed post-Neolithic, probably Bronze
Age deposits.
These researches established several basic
facts about the prehistoric use of the site:
• The general area of Umbro was occupied in the
Neolithic and in at least one phase of the
Bronze Age, possibly the Middle Bronze Age
(Tinè 1992).
• Occupation was patchy; for instance, the
Neolithic area, though rich in archaeological
materials, was restricted to a small area below
the main cliffs at the site.
• Human bone at the Neolithic site was actually
of Late Roman date, though no other Roman
remains to speak of were found here. The
origins of this deposit remain a puzzle.
• The main Neolithic area was dated
radiometrically to the mid-6th
and mid-5th
millennia (both Stentinello periods). It was
probably also occupied later in the Late
Neolithic (Diana) and possibly in the Copper
Age as well.
• The depositional sequence at the main
Neolithic area is complex and difficult to
interpret, particularly in the upper levels; in
contrast, other areas such as the Bronze Age
area in Trench 4 typically have shallow and
simple stratigraphies.
We approached the 1999 excavations with
a number of research questions in mind:
(1) one basic goal was to clarify and
elaborate the dating of Neolithic sequence in
Trench 1.
(2) the Trench 1 Neolithic sondage was
undertaken primarily to explore the cultural
sequence and the potential for further investigation.
In 1999 we hoped to excavate a much broader area
to yield substantial data on Neolithic economy,
settlement and culture.
(3) investigations in other parts of the site
were intended to clarify the presence or absence of
any further Neolithic deposits, for instance on the
clifftops above Trench 1 and the fields atop the
plateau. They were also intended to provide further
dating material for the post-Neolithic but poorly
dated area located in Trench 4.
The basic methods remained unchanged
from those used in 1998 (see Robb 1998 for
description). To summarize, all excavations were
located within an overall site grid whose datum was
off the edge of the site to the east. For each trench a
local trench datum was established from which
depths could be measured and converted to
absolute depths. For Trenches 1 and 7, all soil was
removed in 10 cm arbitrary levels; records were
kept of how these related to the sometimes clear,
sometimes fuzzy natural soil units. Trench 6 was
excavated in arbitrary 10 cm levels except for the
removal of large rock fall along the east side of the
trench, where this was impractical. All deposits
removed were sifted in 5 mm screens. Finds were
bagged separately for each level and 1x1 meter
square. When features such as pits or postholes
were encountered, they were excavated and finds
bagged separately as separate contexts. Carbon
Bova Marina Archaeological Project 1999
25
samples were photographed in situ and removed
with metal tools only to foil envelopes. Flotation
samples were also removed from selected levels as
detailed below. Following excavation, all trenches
were mapped, profiled and photographed.
3.2. Description of 1999 Trenches
3.2.1. Trench 1 (continued excavation)
Trench 1 was located just below the cliffs
in the densest area of Neolithic artifacts (Figure
10). It was dug into sediments more than 2 meters
deep, which had accumulated on top of large
boulders which had fallen from the rock face before
the Neolithic period, creating a small talus cone
about 3 meters high at the base of the cliffs.
Excavations in Trench 5, at the base of these
boulders, in 1998 had established that these
boulders lie directly upon sterile clay and thus
predate any human occupation of the site.
In 1999 we opened a long narrow trench
extending six meters north-south by one meter east-
west. After excavating the top 10-20 centimeters in
four of these six squares, we changed strategy and
excavated one of them (-8n/-36e) and the adjacent
square to the west (-8n/-37e) as a deep sondage.
Thus at the beginning of the 1999 season Trench 1
essentially consisted of a one by two meter
sondage. The goal of the 1999 season was to finish
excavating all six meters of the original north-south
trench down to bedrock or sterile soil (Figure 8).
By the end of the season, this goal was
largely accomplished through work with trowels,
handpicks and pick. The only major difficulty was
the removal of excavated soil, since there was little
unexcavated area atop the talus for a backdirt heap,
while shipping soil down to the base of the slope
with buckets and ropes was very hard work and
unsafe. In the end, we devised a system by which
soil was poured down a plastic pipe about 20 cm in
diameter and three meters long which dumped it
directly in a screen hung from a tripod. Finds were
then passed back up to the trench to be bagged
inside small plastic containers. This system may
have resulted in slight micro-damage in transit to a
few delicate finds, but allowed excavation to
continue rapidly without undue crew fatigue. The
problem of ultimately carrying the soil back up the
three meter drop to the trench to fill the trench back
in remains unresolved.
In the northernmost two squares of the
trench (-6n/-36e and –7n/-36e), we stopped at the
end of the season in sediments which, though still
including sporadic Neolithic material, were
increasingly sterile. In –8n/-36e, most of the
sediments had been removed in the previous year’s
work, but we excavated a further 10-20 cm. At the
bottom, about half the square consisted of bedrock
and open, air-filled fissures were gaping between
buried boulders along the east side of the pit.
Deposits in the next three squares to the south (-
9n/-36e, -10n) were relatively shallower, as the
underlying bedrock sloped upward. Excavation in
these squares finished when bedrock was reached.
We also excavated another 30-40 cm in the
westward extension square (-8n/-37e) which
reached a depth of about 280 cm below the trench
datum.
Unexcavated deposits remain at the base
of the northern half of the square, at the bottom of
the westward extension, and below a large boulder
just northeast of the trench. It is clear that the
underlying boulders slope sharply downwards as
they approached the cliff, and the ultimate depth of
deposits close to the cliff face remains unknown.
The stratigraphy was described in terms of
five strata, as in the 1998 sondage (Figure 9; Table
1). These were more or less found throughout the
new, larger excavated area, with some ambiguity at
times.
Stratum I. This stratum consisted of a
layer of topsoil 5-10 cm thick, with an irregular
conformation and extensively disturbed by erosion,
soil washing in, and vegetation. A loam containing
sand, silt and clay, it was a light brownish color
(Munsell color 10YR 4/2 dry). It contained cultural
finds from all periods within the Neolithic.
Stratum II. This stratum consisted of the
same basic sediment as Stratum I, but was less
disturbed with less vegetation. It was slightly
sandier in places, probably due to the
decomposition of sandstone slabs, and consisted of
a mixed sandy loam (Munsell color 10YR 5/3 dry).
It contained some rock fall, particularly small rocks
20-40 mm in diameter, and much penetration by
small roots. Stratum II varied greatly in thickness,
from about 20 cm up to about 60 cm. While its top
surface was irregular, its bottom surface was far
closer to horizontal. It contained cultural finds from
all periods within the Neolithic as well as almost all
the possible Copper Age sherds.
Stratum III. This stratum consisted of
light, yellowish, gritty sediments similar to those in
Stratum II (sandy loam, Munsell color 10YR 5/4
dry), and in places the distinction between the two
strata was not particularly clear. It contained more
rock fall of all sizes. Rocks generally lay roughly
horizontal, suggesting gradual rather than
catastrophic accumulation. The differences between
Statum II and Stratum III may be due in part to less
root penetration in Stratum III, which might leave
larger, undecomposed rocks while Stratum II was
Bova Marina Archaeological Project 1999
26
slightly sandier. The north end of the trench
contained considerable rock fall in this stratum with
accompanying sandy lenses. Stratum III was a
relatively horizontal level between 30 and 50 cm
thick. It contained both Stentinello wares and
abundant Diana wares (see below for further
discussion).
Stratum IV. This stratum was clearly
distinguishable from Stratum III in section. It was a
denser, browner, and more compact sandy clay with
much less rock fall in it (Munsell color 10YR 4/4,
dry). The rocks in this layer lay more or less
horizontal. In the northern end of the trench,
Stratum IV formed a homogeneous layer between
30 and 40 cm thick. In the southern half of the
trench, Stratum IV was less thick and overlay rising
bedrock. It contained Stentinello pottery.
Stratum V. This stratum made up the
bottom 20-30 cm of excavated deposits in the
northern end of the square. Although bedrock
underlay this sediment directly in –8n/-36e, its
depth at the northern end of the trench is unknown.
It consisted of a dark brown (Munsell color 10YR
4/4, dry) dense, compact, clayey sediment with
little rock fall in it. It contained small flecks of
charcoal and disintegrated animal bone, resembling
midden deposits. It contained Stentinello pottery.
This definition of the stratigraphy is based
upon the six-meter north-south strip of squares in –
36e. In the westward extension square (-8n/-37e) all
strata appear to slope downward sharply as one
approaches the cliff face. This clearly has
implications for both dating the site and
interpreting the living area available at any one
time. Clarifying the depth of the stratigraphy close
to the cliff wall is a major goal of future excavation
seasons.
Nowhere in the excavated area did we
encounter clear evidence of structures. Several
faint, possible pits or post-holes all turned
unconvincing upon full excavation, and may easily
have been due to ancient disturbances by rodents or
roots. Concentrations of small, dispersed, lightly
fired fragments of daub were noted in certain
levels, notably in Stratum III. Here the fragments
consisted of flattish pieces ca. 5 cm thick, smoothed
on both sides and with occasional impressions of
sticks up to 3 cm in diameter. These may be
evidence of a house, wall, or shed; alternatively,
they may be from a non-habitation structure such as
a wall or daub-lined pit. When found, they were
scattered in the soil at a consistent depth but
without contextual position or orientation.
Likewise, few features were found, and
none which were unequivocally interpretible.
Several large sand lenses were encountered. One, in
the far southwestern corner of the pit, contained
almost pure brownish sand ca. 25 cm deep in a
neatly delimited area about 25 cm in area. The
source of such sand is not clear as it does not occur
on the site, nor in most geological contexts. Hence
it may have been transported to the site for a
specific purpose. In contrast, several sand lenses in
–9n/-36e, -8n/-36e, and –6n/-36e were clearly
derived from decomposing bedrock. They consisted
of yellowish sand in broad, flat, generally
horizontal lenses, diffusely bounded. The sand was
very similar to that demonstrably freed from
dissolving tabular sandstone fragments in the upper
levels of the north end of the pit, and they
contained virtually no cultural material.
The only feature showing the cultural use
of fire consisted of a small area about 15 cm in
diameter and about 10 cm deep, running into the
west wall of the trench deep in the westward
extension (-8n/-37e). This feature contained
abundant charcoal and ash, but no really associated
artifacts. It is very similar to one encountered in the
1998 excavations of Trench 5 (Robb 1998). Its
form suggests it is not natural burning and it may
represent a small fire area used for cooking or for a
special purpose.
The form of the site has evolved
substantially during its use. The earliest levels
excavated so far consist of Early Neolithic deposits
on top of sloping boulders in the north half of the
trench; in this period, the site would have afforded
at most about six square meters of approximately
level area. Only after enough sediments
accumulated to cover bedrock to the south and
make it more or less level was this area also
occupied. Even so, level area for occupation
consisted of at most ca. 15 square meters. This
implies that the site was never used by very large
groups, particularly during the earlier Neolithic,
and that certain functions would have required the
use of off-site areas. For instance, the herding of
animals could have been carried out conveniently in
the basin enclosed by rocky arms of the cliffs just
below the site.
Radiocarbon dating and stratigraphic
interpretation. Radiocarbon dates for Trench 1
from the 1998 excavations placed Stratum V in the
6th
millennium BC and Stratum IV in the 5th
millennium BC, with a date of ca. 3000 BC in
Stratum III probably representing an intrusive or
sporadic Copper Age occupation (Table 6).
Following the 1999 season, we obtained two further
dates for Trench 1. One was for the transition
between Strata II and III, and dated it to 4945-4615
BC (Umbro-13, Table 6). The other was for the
center of Stratum III, and dated it to 5660-5485 BC
(Umbro-15, Table 6).
Bova Marina Archaeological Project 1999
27
These dates are incompatible to some
extent with the chronostratigraphic interpretation
established in 1998, and there are four possible
interpretations.
(1) One is a “high” chronology in which
Strata III-V all date to the 6th
millennium BC,
Stratum II/III date to the 5th
millennium BC. To
argue this, one needs to argue that Diana wares
were in use for most of the 5th
millennium BC at the
same time as Stentinello wares, and to discount one
date, Umbro-4, as intrusive into an earlier stratum.
(2) A second is a “low” chronology in
which Stratum V belongs to the 6th
millennium,
Stratum IV belongs to the 5th
millennium, and
Stratum III is a Diana ware level undated but later
than Stentinello times. To argue this, one needs to
discount two dates (Umbro-13 and Umbro-15) as
based upon earlier material remixed into later
strata.
(3) A third is a “middle” chronology in
which Stratum V belongs to the 6th
millennium,
Stratum IV is undated but dates to between the 6th
and 5th
millennia, and Stratum III dates to the 5th
millennium. To argue this, one needs to argue that
Stentinello and Diana wares were used together
during the 5th
millennium, and to discount one date,
Umbro-15, as based upon earlier material remixed
into a later level.
(4) The final possibility is that the
stratigraphy is completely mixed and no sense can
be made of it.
Of these, the last possibility can be readily
discounted. Sediment profiles show a distinct
difference at least between Strata II/III and Strata
IV/V, a difference which wholesale mixing would
obliterate. Moreover, Strata IV/V contain no Diana
wares, while Stratum II/III contain both Diana
wares. Again, wholesale mixing of levels would
obliterate this difference. The first possibility can
probably also be discounted on archaeological
grounds. Stratum III contains abundant Diana
pottery, and while we might imagine an early Diana
date in the earlier 5th
millennium BC (cf. Leighton
1999), dating a Diana stratum to the earlier 6th
millennium would place it far earlier than any other
known Diana site (Skeates 1994) and must be an
error.
This leaves the “low” chronology and the
“middle” chronology as possible interpretations.
Both are plausible archaeologically. Given the
shape of the site, with Trench 1 intensively
occupied and lying below steep slopes to the south,
it is clear that the stratigraphy was not sealed in
each period but was accumulative. In other words,
early levels contain only early ceramics, charcoal,
and other materials, but later levels contain both
later materials and earlier material washed in or
mixed in. This provides grounds for understanding
Stentinello pottery mixed in the upper part of the
stratigraphy and early dates from upper levels, and
is consistent with both the “low” or “medium”
interpretation.
Here it is worth noting some conclusions
which both views point towards, and some
unresolved issues.
• In both scenarios, Stratum V is a Stentinello
level dating to the 6th
millennium BC, and
Stratum IV is a Stentinello level dating to
sometime between the 6th
and 5th
millennium.
• In both scenarios, Stratum III is a Diana level.
• In both scenarios, Stratum II may be a Copper
Age level with substantial mixing of earlier
material, as most of the few possible Copper
Age sherds are found in it along with
Stentinello and Diana wares.
• One key unresolved issue is the chronology of
Stratum IV, which may date to anywhere
between the 6th
and 5th
millennium BC. More
radiocarbon dates from Stratum IV may help
resolve this.
• A second key unresolved issue is the dating of
the Diana level, Stratum III, and the related
issue of whether Diana and Stentinello wares
were in use simultaneously or whether their
association in Stratum III represents mixing of
earlier and later material. The chronological
issue may be resolvable through further
radiocarbon dates for Stratum III. The ceramic
question may be impossible to resolve with
archaeological evidence from Umbro.
Lest we despair, it should be pointed out that these
problems are normal rather than exceptional.
Virtually every site explored from this period has
similar difficulties. For instance, the regional
cornerstone, the Lipari chronology based upon the
Lipari Acropolis, Contrada Diana, and Castellaro
Vecchio (Bernabò Brea and Cavalier 1956; 1960;
1980), achieved a clear periodization by
discounting numerous sherds in “mixed” levels.
This method assumed in advance of the data that
the ceramics seriate into short, clearly-bounded
periods and that any finds to the contrary are
merely contingent errors of site preservation which
should be ignored. If one allows other possibilities
such as the contemporary use of several styles for
long periods, quite different interpretations are
possible.
Bova Marina Archaeological Project 1999
28
3.2.2. Trench 6
Trench 6 proved one of the most
interesting surprises of the 1999 field season. We
began investigating this area simply to understand
what archaeological deposits existed on the west
side of the site. At this point, prehistoric sherds
were visible eroding out of a bank about a meter
high, about three meters east of the Bova-Marina-
Bova Superior road.
The first effort was to cut a vertical profile
along the face of this bank; we expected to find
archaeological material tumbled out of context at
the base of the rocky peak here, but wanted to
check this. A profile about three meters long was
cleaned and mapped, running generally north-south
(Figure 11). As this work was completed, it was
clear that most of the sherds were localized at one
point where large fragments of a single vessel were
visible in the profile. We then decided to cut a
trench back from the profile to explore further. A
two meter by two meter trench was laid out,
anchored on the site grid. Because the profile cut
across the grid at an angle, the actual trench
excavated was a trapezoid including one complete
one meter square and parts of two others. Finally,
an extension of half a meter thick was opened up on
the eastern and southern margins of the trench; this
was excavated only to the upper zones of the rock
fall in Stratum II (Figure 12, Figure 15). Trench 6
was excavated by grid squares in 10-cm levels and
all soil was screened in 5 mm mesh.
Five strata were defined in Trench 6
(Figure 13).
Stratum I. This consisted of topsoil 5-10
cm thick, a clayey loam (Munsell color 10YR 3/2
dry) with much root penetration and very little
archaeological material.
Stratum II. The basic sediment was the
same as in Stratum I though slightly darker in color
(clayey loam, Munsell color 10YR 5/2 dry). It
contained far less vegetation but much rock fall.
Rocks included both large tumbled stones and
many smaller irregular rocks, tightly packed
together. Overall, the stratum was between thirty
and fifty centimeters thick. Archaeological material
was infrequent. This stratum contained two
fragments of a ground stone quern (see below),
possibly in secondary context.
Stratum III. This level included a
prehistoric floor or surface. The basic soil
continued to be a sandy, clayey loam (Munsell
10YR 4/2 dry). There was no sign of plaster or
paving, but the soil was compacter than above;
there were fewer large rocks but a dense
concentration of small stones; many small stones
were lying flat; and most of the few sherds found
scattered through the level were lying flat. This
level was about fifteen centimeters thick. It was
excavated away only in the westernmost part of the
Trench, and left in situ for further excavation in the
eastern areas. The radiocarbon samples for Trench
6 came from this stratum.
Stratum IV. This stratum was visible only
at the western edge of the pit where erosion and
excavation had exposed it. It appears to be a
whitish, sterile soil of variable thickness (a greyish-
white concreted sand and clay, Munsell color 10YR
5/2). It appears to contain a few pieces of
archaeological material but may be the top of sterile
sediments derived from decomposing bedrock.
Stratum V. Bedrock.
Pot deposition and other features. As we
excavated Trench 6, the vessel appearing in the
profile turned out to be part of a group of three pots
(Figure 14, Figure 16). The pot visible in the
profile was a basin between thirty and forty
centimeters in maximum diameter, apparently
undecorated on the surface but with three
horizontal handles under the rim. Its overall form is
yet to be understood as it was extensively crushed.
The other two pots consisted of small attingitoi or
cup-dippers with tall upraised handles. One was
relatively thin-walled, the other cruder and thicker.
The three pots were placed close together, with
their walls almost touching. Their bases rested on
the same plane at 70 cm below trench datum, a
place corresponding with the floor level defined in
Stratum III. As these pots were excavated, they
were mapped and photographed, and their fill was
collected completely as a series of flotation samples
(see below). Under the two attingitoi, we found a
fragment of a fourth pot, which consisted only of
one horn of a horned bowl.
Large and small rock fall covered the
entire area including the pot deposition group, but
grew thicker to the east. Larger stones were more
common in the southeastern corner of the square.
No order was visible in the rock fall. No other
possible features or structures were visible, though
the bedrock at the southwestern corner of the trench
may have been artificially shaped to some extent.
Dating. Both the vessel styles and absolute
datations agree in assigning the deposition to the
Early Bronze Age. While analysis of the large bowl
must await its reconstruction, the form of the
attingitoi is very similar to ones from Mursia,
Pantelleria, varied Rodi-Tindari-Vallelunga sites
such as Ciavolaro in Sicily, and S. Domenico
Ricadi in Calabria. The ceramic horn is distinctive
of the Rodì-Tindari-Vallelunga style (Tusa 1993;
Bova Marina Archaeological Project 1999
29
Castellana 1996). A radiocarbon sample from
charcoal on the surface below the vessel group gave
a date of 3390 +/- 60 BP (1780-1520 BC
calibrated).
Because of the limited excavations carried
out at this location to date, it is not entirely clear
what kind of deposition this group of pots
represents. They may represent the remnant of a
house context. The daub fragments associated with
them may argue in favor of this, but the finds from
the trench are far sparser than one would expect
from a domestic context, including virtually no
animal bone or lithics, and there is yet no sign of
any structure such as the typically substantial stone-
footed Bronze Age huts known from eastern Sicily
and Lipari. A second possibility is ritual: the pots
may represent an idiosyncratic ritual deposition of
some kind, though no comparable sites have been
published. At this point, burial seems an equally
plausible possibility. In eastern Sicily, EBA burials
are commonly found in cemeteries of small rock-
cut tombs, located along elevated ridges or cliffs of
limestone or other calcareous stone. Umbro is in
just such a location, and the vessel group lies 2-3
meters from an outcropping, low bedrock ridge,
now buried. Furthermore, in several cases, outside
rock-cut tombs archaeologists excavated leveled
areas or platforms which contained groups of
pottery vessels, interpreted as grave offerings
deposited outside tombs rather than inside (see for
instance Santa Febronia, Maniscalco 1996). Pottery
vessels used as grave goods in the Bronze Age
often consisted of groups including a bowl or basin
and several cups or dippers, interpreted as a
complete pottery service used for ritual functions
(Maniscalco 1999). This is precisely the
composition of the Umbro pot group. It is possible
that there are rock-cut tombs in the rocky bank just
behind the pottery group, and that the pots
represent external grave goods for such tombs, with
the rock fall representing remnants of a wall closing
the tombs (B. McConnell and L. Maniscalco, pers.
comm.). Needless to say, we hope to excavate
further in this area, to ascertain the actual nature of
the site.
3.2.3. Trench 7
Trench 7 continues the exploration of
outlying parts of the site already begun with Area 2,
Trench 3, Trench 4 and Trench 5.
The trench consisted of an exploratory one
meter square pit (Figure 17, Figure 19). It was
located on a small (ca. 4 meters in either direction)
level area at the very crest of the southern slope
below the cliffs, just below the small field
described below and next to the “south slope
profile” (see Robb 1998). The “cliffs” above it
were only about two meters tall, consisting of very
irregularly sloping bedrock outcrops. It was located
about six meters southeast of Trench 3. This level
patch, one of very few on the hillside, had been
used in the previous year for a backdirt heap for
Trench 3, and some soil remaining after Trench 3
was backfilled was removed without screening. By
these means, the terrace was cleared to a uniform
surface, except for a low, flat boulder in its center.
Beyond a few historic sherds clearly in secondary
context, no artifacts were evident either in this soil
removal or in the surface thus cleared. Once the
working ground was cleared, before laying out the
trench, we cleaned away some lose, irregular soil
adhering to the cliff wall directly above where we
wished to place the trench; this was to prevent this
soil from falling into the trench later. About half a
cubic meter of soil was removed in this way,
leaving a vertical surface. Again, only a few
historic sherds out of context were found.
Trench 7 was located directly against the
rocky wall rather than in the center of the level,
open area. Trench 3 had shown that such level
pockets adhering to steep slopes could contain
considerable depth of soil, but that they were often
stratigraphically complex. Moreover, there were no
indications on the surface of the level area of
underground deposits here. A primary
consideration, however, was that of caves. The
shelving, angled bedrock at this point creates the
possibility of small pocket caves in the cliff wall.
With increasing evidence that the cliff-tops
immediately above were not settled, and that the
hillslope deposits such as in Trench 3 and Area 2
are not in original context, we wondered whether
prehistoric hillside deposits might originally have
been in small caves or shelters, now buried or
collapsed, high along the rocky walls of the basin.
Trench 7 was excavated using the same
methods as elsewhere: excavation with trowel,
handpick, and pick proceeded in 10 cm arbitrary
levels, and all soil was screened in 5 mm mesh. A
trench datum was set up using the cleft in the large
flat boulder north of the trench to lodge a datum
spike; this was the only secure point available from
which depths could be practically measured.
Excavation continued to a total depth of 150 cm
below trench datum, or ca. 110 cm below the
cleared ground surface.
Two strata were defined in the pit, both of
extremely irregular dimensions (Figure 18). The
upper one was a very loose loam containing sand,
silt and clay. It contained stones of all dimensions
jumbled without consistent orientation, and was
much disturbed by rootlets. The maximum depth
was about 120 cm below trench datum, or close to
Bova Marina Archaeological Project 1999
30
the bottom of the pit. This sediment clearly
represents soil filtering in among the rocks of the
cliffs, a process visible everywhere on the cliff
slopes. It did not appear internally stratified in any
sense, and it is not clear when it was deposited or
formed. It contained a few sherds from all dates
including prehistoric, medieval and modern. The
second, lower stratum was very similar in its basic
substance, but was slightly less loose in texture,
slightly browner. Again, it contained both
prehistoric and historic sherds. It thus seems to be
an ancient version of soil and rocky infill of
irregularities in the cliffs.
Trench 7 finds included only ceramics.
The density of finds was very low and three levels
produced no artifacts at all. Prehistoric and historic
sherds were mixed down to the bottom of the
trench, suggesting that all the fill excavated was
mixed and of recent date.
We stopped excavating Trench 7 because
the field season ended, and because it was not clear
how to continue excavating safely in a small trench
more than a meter deep in loose deposits partially
below a rocky overhang. There was no indication
how deep soil deposits ultimately run here on the
cliff-side; if we assume that the cliffs were
originally vertical all around the sides of the basin
enclosing the Neolithic site, and that the current
slopes of approximately forty-five degrees have
been formed by millennia of rock fall and soil infill,
there could potentially be twenty meters of deposits
here.
With this in mind, we need to return to the
question of how to check for possible cave deposits
in the rocky walls of the basin. Figuring out how to
investigate this safely and practically remains a
challenge for the future.
3.3. Other Areas Explored
With the discovery of distinct
archaeological areas on the west side of the site
above the cliffs, and with continued questioning
about the nature of the upper stratigraphy of Trench
1 below the cliffs, we decided it was important to
re-check whether there were any archaeological
deposits in the intervening parts of the site – the
highest peak of the rocky cliffs. In particular, we
were concerned to know whether there was any
possibility that the artifacts in Trench 1 and
elsewhere below the cliffs had fallen there from
cliff-top deposits, and whether the newly-found
Bronze Age deposits in Trench 6 were in situ, as
they seemed to be, or could have eroded from
deposits higher on the peak.
We had, of course, thoroughly examined
the entire ground surface of these peaks in previous
years, and we did so again. Perhaps half the surface
consists of irregular, shelving bedrock outcrops;
between these are pockets of soil of varying depths,
and the whole area is sparsely covered in scrubby
vegetation. Ground visibility was thus patchy but
reasonable in places. All previous examinations had
failed to find any archaeological material of any
date.
This season, we decided to check for
possible archaeological deposits below the surface
(for instance, surviving in bedrock crevices). Since
there were no known archaeological deposits in this
area, and since the soil covering the area consisted
of thin, irregular pockets, we excavated it
informally by cleaning the bedrock off with a
shovel. A team of L. Saccà and a student carried
out the work. Most of the soil removed was
screened, but not all was, as it rapidly became
apparent that the sediments were entirely sterile and
that the team could monitor it adequately in the
process of removing it. This strategy carried the
risk of damaging real archaeological deposits
through loosely controlled excavation, but we felt
that any deposits encountered could be recognized
rapidly enough to stop, grid, and excavate in a
controlled way with only limited damage;
meanwhile, we needed rapid information without
committing ourselves for a full-scale excavation in
an area of patchy, sterile soil and bedrock. All
cleaned areas were backfilled when finished. In the
event, this strategy was justified.
We cleaned the bedrock in five areas,
three on the high area of the cliff-top, one along the
cliff edge above Trench 3, and one to the southeast
above Trench 7. In all of them, the soil consisted of
a loose, dark loam with much organic material.
Atop the cliffs, the northernmost area
consisted of the largest pocket of soil close to the
actual cliff edge which could be safely worked at. It
lay directly above Trench 1, about three meters in
from the actual lip of the cliff. The soil filled a
shallow trough about two meters east-west and
perhaps .75 meters north-south. The deepest part of
the excavation was about 40 cm deep. The other
two areas cleared atop the cliffs were located
southwest of here, on the very highest point of the
hill. These consisted of two parallel troughs about
four meters long and less than a meter wide, each
filling in a v-shaped recess created by shelving
bedrock (Figure 20). The maximum soil depth in
both was 30 cm. None of these trenches contained
any artifacts.
Along the cliff edge above Trench 3, we
cleared a thin covering of loose topsoil from
Bova Marina Archaeological Project 1999
31
bedrock in a series of small pockets. None
contained any cultural material.
Southeast of the rocky peak, there is a
small level field which lies above the southern
slope of the basin between the cliffs. The profile
just below this field was cleaned and recorded in
1998 (“South slope profile”, Robb 1998), and
Trench 7 was excavated just below it in 1999. If
archaeological material was washing into the basin
from somewhere high on the southern slopes, this
seemed one candidate for a source area. We did not
clean this area down to bedrock, as the soil was of
unknown depth and we were reluctant to intervene
here substantially without setting up a formal
excavation area. However, we cleared the ground
vegetation and removed the top 10 cm of soil using
the same techniques as above. The only finds
included a number of sherds from a single modern
vessel. This confirms the generally sterile nature of
this part of the site inferred from the south slope
profile and from Trench 7.
These exercises were very useful in
understanding the site. It seems virtually certain
that the Trench 1 deposits and the south slope
deposits in the area of Trench 3 were probably
originally deposited below the cliffs; it is difficult
to imagine substantial enough deposits above the
cliffs which would not leave at least some residual
archaeological material up above. Likewise, the
Bronze Age deposits in Trench 6 seem localized
and in their original location. The presence of an
elevated, archaeologically empty area between the
Neolithic and Bronze Age areas confirms the
separation of these deposits.
While on the topic of other areas
investigated, it is worth illlustrating some of the
1998 trenches briefly; these trenches were
described in the previous preliminary report (Robb
1998) but no photographic docmentation was then
available. Figure 21 shows Area 2, a zone on the
southern slope of the basin where we cleaned a
profile within a cleft between tumbled boulders but
found only redeposited material in secondary
position. Figure 22 shows Trench 3, a one meter by
one meter test pit further up the same slope, and
Figure 24 depicts Trench 5, a three square meter
trench located below Trench 1 in the very bottom
of the basin. The boulders looming over the trench
are those underlying Trench 1. Finally, Figure 23
depicts Trench 4, a broad, shallow four by five
meter pit located on a rocky outcrop above the
cliffs. It yielded no structures or architecture, but a
small assemblage of post-Neolithic pottery which is
probably Bronze Age in date.
Bova Marina Archaeological Project 1999
32
4. UMBRO EXCAVATIONS: DESCRIPTION OF FINDS
In this section, finds from the Umbro
excavations are discussed. Note that this discussion
includes both new finds from the 1999 excavations
and previously described finds from the 1998
excavations; the reason is that our understanding of
earlier finds has often changed as research has
progressed.
4.1. Ceramics
4.1.1. Neolithic and Copper Age wares: Trench 1.
The typical wares found in the Neolithic
area of Umbro have already been described (Robb
1998), and pending statistical analysis, little
additional comment can be made. They will only be
briefly presented here. Pottery styles found in the
area below the cliffs at Umbro include Stentinello
coarse wares and finewares, Diana wares, and
possible Copper Age sherds.
Stentinello coarse wares (Figure 25) are
often thick-walled and relatively crudely made,
executed in dark brown, light brown, red and
orange fabrics. Some vessels in this category may
have been bowls or cooking pots, while others
appear to have been large storage vessels. They
were decorated with a range of impressed designs,
including “stab and drag” punctuation, shell and
rocker impressions, and arrays of short parallel
lines, often arranged vertically. They are similar not
only to coarse wares at other Stentinello sites such
as Acconia and Capo Alfiere, but to coarser
Impressed Ware vessels in general.
Stentinello finewares (Figure 25) are
usually thinner-walled, with dark, glossy burnished
surfaces apparently intended to highlight the
impressions filled with yellow, red or white pastes.
Decoration was by impression, with geometrical
arrays of small motifs forming elaborate design
schemes. The most common motifs were “v”
impressions, small straight lines and fine grids of
diagonal lines. These were recombined in bands
below the rim of a vessel, with vertically “hanging”
strips of decoration or further banding below.
Decoration also occurred elsewhere on vessels, as
shown by a small handle with an elaborate strip of
impressions down its back. Many of these vessels
appear to have been small open bowls, probably for
serving food and drink.
Diana wares (Figure 26) include both
coarse and fine fabrics, though the coarse fabrics
can be difficult to distinguish from other Neolithic
wares. The Umbro Diana finewares include many
shallow carinated bowls with reddish, brownish or
buff fabrics, highly burnished, occasionally slipped
with a red slip, and usually undecorated save for a
thin line below the rim and for the famous trumpet
handles attached horizontally below the rim.
In addition to Diana and Stentinello wares,
other Neolithic wares included a few red-painted
sherds on a buff fabric, including some from
relatively deep in Stratum IV.
A few sherds unusual for Neolithic wares
(though not entirely unparalleled) and with generic
Copper Age parallels were found in the 1998
excavations (Figure 27). These included several
body sherds with large circular bosses, and several
large horizontal handles lacking the typical Diana
flaring. More such sherds were found in 1999, and
the greater extent of excavation allowed us to
define their stratigraphic level as Stratum II. These
included several large, thick body sherds with a
double band of finger-pinched cordoned
decoration, perhaps from the same vessel. Another
sherd was incised with a series of evenly-spaced
vertical lines ending in spirals at the top. An
especially unusual fragment was decorated with a
broad, shallow groove interrupted by a pair of
protuberances. Finally, several sherds bore incised
designs consisting of two parallel zig-zag lines, the
space between which was filled with rough
hatching (cf. Cocchi Genick 1996 for parallels from
the Southern Italian Copper Age).
4.1.2. Stampini and other ceramic-production
related implements
The Umbro excavations have yielded six
stampini, or small ceramic punches, to date (Figure
29). They consist of small fired clay cylinders with
simple geometrical motifs in raised clay at the end
or on the side. These small tools were used to
impress unfired vessels to create the stereotypical
Stentinello pottery designs. At Umbro, they have
been found in all strata of Trench 1. Examples
include:
• a complete stampino with a v-motif at one end;
the edges of the v are finely ticked (Bag 1021;
length 70 mm).
• a fragmentary handle of a stampino (Bag
1035).
Bova Marina Archaeological Project 1999
33
• a fragmentary stampino, approximately 48 mm
long and 10 mm in diameter and slightly
broken; the end was flared but the exact motif
was eroded and indistinct (Bag 1152).
• a fragmentary stampino, whose end had broken
off. The surviving end appears to have been
triangular, suggesting that the motif on the end
was originally a “v” (Bag 1170).
• a fragmentary stampino ending in an eroded
but distinct “v” impression (Bag 1195).
• a fragmentary stampino with a cylindrical
body. Unusually, this fragment bore a raised
motif for impressing clay on the side of the
body rather than the end. The motif consisted
of a thin ridge of clay forming a straight line
with tick marks along its edge (Bag 1225).
These numerous tools for making pots suggest that
ceramic production was a common activity at
Umbro in the Stentinello period. This impression is
reinforced by the presence of other implements
which could have been used for forming or
decorating pots, including a shark’s tooth, a
smoothed bone, several small marine shells with
worked edges. Finally, one large, reddish sherd had
two edges worn round rounded and smoothed by
protracted use in smoothing or polishing some
material (Figure 29). All of these could have been
used in forming, decorating or burnishing pots prior
to firing them.
4.1.3. Thin section analysis of prehistoric ceramics
Fifteen sherds were transported to the
University of Southampton for thin section analysis
(we are grateful to the Soprintendenza
Archeological della Calabria for permission to
carry out this work). These sherds represented a
variety of prehistoric wares in use at the site,
including Stentinello coarse wares (sample 4) and
decorated (2, 3, 9, 11) and undecorated (7, 8, 15)
fine wares, Diana vessels (10, 13, 14), Early
Bronze Age vessels, possible Copper Age vessels
(6, 12), and red-painted wares (1) which were
considered possible imports. A piece of daub was
also examined (5). Finally, samples of clay from
two clay sources were examined. One clay sample
was from the field directly east of the site at a
distance of 10 meters east of the site datum point.
The other was taken from a pure clay bank in the
Bova-Bova Superiore road cut just south of the
Umbro plateau where the road begins to rise
steeply, about 500 m south of the Neolithic site.
The thin-section analysis was carried out
by Sonya Collins and Jayne Watts under the
supervision of Dr. David Williams; we are grateful
to these researchers for their work, discussed here
and reproduced in the attached documentation. All
sherds were impregnated with a consolidating glue
and then thin-sectioned. Sherd 1 proved too soft
and crumbly to thin-section, but was
macroscopically described and illustrated. The clay
samples were worked into lumps, fired,
impregnated with consolidating glue and thin-
sectioned.
Thin section analyses, while preliminary,
suggest a number of trends.
First, virtually all the sherds had the same
composition. Common mineral inclusions in the
pastes include mica, quartz, and feldspar, with
possible metamorphic rock sources including gneiss
and micaschist. All of these are consistent with the
kinds of minerals to be found locally resulting from
the decomposition of metamorphic rocks common
in Aspromonte at altitudes of 500 meters and
above. Hence, these minerals are available, for
instance, in the form of sand in river valleys, and by
crushing up parent rock from nearby outcrops. The
only exceptions were the daub sample (5) and a rim
(14). The daub contained similar mineral inclusions
but included more of them and larger grains.
Sample 14, a small, fine greyish Diana rim with the
characteristic tubular handle, contained an unusual
base clay with abundant foraminifera fossils and
contained mica, quartz, and black iron grains.
Given the high mica content, it seems likely that it,
too, came from some area of Aspromonte, but it
may have been made in a different area, from
different raw clay, or using different temper.
The fact that all the other sherds had a
common composition suggests that they were all
locally made, even when they are made in very
different styles such as red-painted wares,
originally thought to be a potential import from
Northern Calabria. This emphasis on local
production is a common finding in prehistoric
Italian thin-section studies (for instance, Skeates
1992).
The second surprising finding emerged
from comparison of the pottery with clay sources
near the site. Both clay sources near the site
contained high levels of organic elements in the
form of microfossils. They were clearly
distinguishable from the clays used in all the sherds
except for sample 14 (discussed above) and to a
lesser extent sample 15. This suggests that,
although there is strong evidence that people were
making pottery at Umbro (see below), and although
large quantities of apparently suitable clay are
available within a few meters of the site, people
preferred to transport clay from somewhere else to
make pots. This is especially interesting as regards
the daub sample, as one would expect a bulky
material such as daub to have been made from the
Bova Marina Archaeological Project 1999
34
closest available clay to save work. It may be that
the potters of Umbro knew the qualities of the local
clays intimately and preferred to choose the most
suitable for a particular task even if they were not
the closest to hand.
These investigations require further
samples and further contextualization in the local
geology and should be considered preliminary
indications only.
4.1.4. Bronze Age wares: Trench 6.
Ceramics from Trench 6 were completely
different from the Neolithic wares found elsewhere
on the site. The general paste was soft, coarse and
light, ranging from orange to brown in color and
usually with a black oxidation stripe in the center of
the fabric. There was no surface decoration evident
beyond an occasional groove or ticking along a rim.
The repertory of vessel forms was also distinct,
with everted rims, thick, beaded lips, ring bases,
and tall raised strap handles. These wares closely
resembled pottery found in Trench 4, and the two
probably date to the same period, the Early Bronze
Age.
In addition to these wares, a few sherds of
a very different style were found, of a black fabric
with a very glossy highly burnished surface. These
sherds represented at least two vessels. Two sherds
were decorated with broad, shallow grooves. These
wares may be similar to Apennine pottery, although
the grooves are not excised. Alternatively, they may
be the wares which Tinè (1992) considered similar
to Maltese Borg-in-Nadur wares.
The pottery deposition from Trench 6
contained three vessels and a fragment of a fourth.
These vessels are currently under restoration by
conservators of the Museo Nazionale di Reggio
Calabria.
• Vessel 1 (Figure 28) was a small, thin-walled
cup-dipper (attingitoio). The fabric was a soft
buff/orange inside and outside, with fine grit
temper and a thin oxidation stripe. There was
no surface decoration. The lip was everted, and
a tall handle was raised directly from the lip.
Both this cup and Vessel 2 resemble in form,
though not necessarily in fabric or surface
treatment, contemporary vessels from Mursia
(Pantelleria), Capo Graziano, Rodi-Tindari-
Vallelunga sites, and S. Domenico Ricadi
(Tropea)(Bernabò Brea 1957; Castellana 1996;
Tusa 1993).
• Vessel 2 (Figure 28) was a small, thick-walled
cup-dipper (attingitoio). The fabric was dark
brown externally and orange-brown internally,
with a black oxidation stripe and small grit
inclusions. The surface showed occasional
signs of having been burnished, but is now
eroded. There was a shallow groove the
outside surface under the rim.
• Vessel 3 was a large impasto bowl, executed in
a soft orange-brown fabric. There was no
observable surface decoration, though the
surface was eroded and in places covered with
a tenacious concretion. The precise form of the
vessel is not yet known, as it is still under
reconstruction, but it measured 20-25 cm in
diameter at the mouth, 10-15 cm in diameter at
the base, and was 15-20 cm tall. Below the rim
there were three horizontal strap handles.
• Below these was found a fragment of a fourth
vessel, a ceramic horn. This was made of a
dark brown, burnished fabric. The concave
side was flattish, while the convex side of the
horn was rounded. This cross-section makes it
resemble more closely the horns from the
Rodi-Tindari-Vallelunga style horned cups
(see Castellana 1996) than the later Ausonian
wares.
4.2. Lithics
Lithics in 1999 were found only in Trench
1. One flake was found in Trench 6, but it was a
crude quartzite flake and may be natural rather than
humanly manufactured. The large number of lithics
found in Trench 1 brings the total lithic assemblage
there to 401 items (Table 3). Of these, 17 (4.2%)
were made from flint; 384 (95.8%) were made from
obsidian. This overwhelming use of obsidian is
typical of Stentinello sites in southern Calabria,
eastern Sicily and Lipari (see Ammerman 1987),
though obsidian use falls to about 50% in the
further margins of Stentinello culture area (e.g.
Stentinello, Orsi 1890; Capo Alfiere, Morter 1992).
To date, only a small sample of about
twenty pieces has been examined by a lithics
analyst (G. Marshall, see Robb 1998); the finds
from the 1999 season do not appear to change the
conclusions of this analysis. A full analysis will
take place during the 2001 study season. Marshall’s
analysis showed that obsidian probably entered
Umbro in the form of small blade cores, which
were then used to produce blades until they were
too small, at which point they were were crushed
using bipolar percussion to create small flakes for
expedient use.
Obsidian provenience analysis in progress
at the University of South Florida has shown that
all samples examined from Umbro came from
Lipari (R. Tykot, pers. comm.). Flint came from a
variety of sources, as at least five distinct colors of
Bova Marina Archaeological Project 1999
35
raw material are known (grey, brown, yellow, red,
and pink). While grey flint of poor quality is found
at Saracena in the hinterlands of Condofuri Marina,
the sources of the other varieties are unknown. As
elsewhere in Calabria, Sicily and Lipari, obsidian
was used primarily for bladelets and unmodified
flakes; flint was used disproportionately for formal
tools and modified flakes, probably for tasks which
required its somewhat tougher cutting edge.
4.3. Other Artifacts
4.3.1. Daub
Daub, or intonaco, is a common Neolithic
technology. To date, 1172 daub fragments have
been recovered at Umbro (Table 4). Of these, 3
came from Trench 3, 5 from Trench 4, 4 from
Trench 5, and 46 from Trench 6. These serve to
document the use of daub at least in the Trench 6
area. The vast majority of fragments (1142) come
from Trench 1, the Neolithic area of the site.
Almost all of the daub fragments
recovered consist of tiny fragments less than a
centimeter or so in maximum dimension. Only
about 80 pieces are larger. The largest fragment
found is about 10 cm. No daub fragments were
found in original position; all were scattered
through the soil, though there were slightly greater
concentrations between 150-180 cm below datum
(in the lower part of Stratum II and in Stratum III)
in the central part of the trench.
All fragments found were burnt to some
degree, usually only lightly but intensely in a few
cases. This does not necessarily imply that burning
was part of the use of daub here; rather, burning
probably helped preserve daub and daub fragments
which were not burned at some point simply
disintegrated and were not preserved. However,
some burning clearly happened before the daub was
completely fragmented and quite possibly during its
use-life; this is suggested by a number of flat
fragments consistently burned on one side only (see
below).
A handful of fragments display
impressions of sticks or reeds; the largest
impression is of a stick about four centimeters in
diameter, but impressions between one and two
centimeters are more common.
Between twenty and thirty fragments
appear to come from a single artifact or structure.
These fragments are flat Running cross-wise
through these fragments in places are impressions
of three reeds or sticks about one centimeter in
diameter, laid edge to edge. The daub thus formed a
flattish sheet about five centimeters thick, enclosing
on a framework of three parallel reeds with about a
centimeter to spare on the inside and outside. One
side was smoothed, and occasionally betrays
possible plank impressions or smoothing marks.
The other side is roughened. The rough side was
consistently burned to a blackish color; the
smoothed side was burned far less intensely, and at
most shows some reddening.
Daub is stereotypically interpreted as
having been used for house construction. However,
sites of Neolithic houses built in daub often yield
very large amounts of it (Shaffer 1985; Tozzi and
Tasca 1989). There is no signs of such large
quantities of daub at Umbro, nor is it clear that
there would have been room for a house even 3-4
meters square in the constricted area below the
cliffs at Trench 1. Rather, the daub may have been
used in some other kind of artifact or facility. The
group of flat daub fragments may belong to some
feature such as a partition, daub-lined storage pit, a
large clay-daubed storage basket, or even a clay-
lined oven or hearth. It is possible that refitting
these daub fragments may shed some light on the
original structure they formed part of.
4.3.2. Shell, Worked Bone and other utilized
Fauna
Several kinds of organic materials were
utilized to make tools. As noted previously (Robb
1998), a small fossil shark’s tooth from Trench 3
may have been kept as a curiosity or used as a tool
for some purpose such as incising unfired pottery.
Similarly, two fragments of boar’s tusk from the
upper strata of Trench 1 may have been used for
tools or ornaments rather than simply representing
culinary debris; one may be artifically smoothed at
the end.
One small marine shell, from a clam-like
bivalve, was found in Stratum II of Trench 1. Its
edges were artificially smoothed and rounded all
around, and a small hole at its apex may have been
naturally or artificially created. This example
closely resembles a worked shell found in 1998 in
the same stratum, as well as an unworked shell from
the same stratum and a burnt shell fragment from
Stratum IV. As these show, small marine shells
were probably commonly used as tools for
smoothing or scraping soft materials.
One worked bone implement was found
(Figure 29). This was a long bone, probably a
radius or metapodial from a sheep-sized animal,
which had been split lengthwise and smoothed all
around. The non-articular end was sharpened to a
point, now broken, and the edges of the fragment
were rounded. The result was a small awl of a type
common in Neolithic sites. Its overall dimensions
were 37 mm long by 10 mm wide by 8 mm thick.
Bova Marina Archaeological Project 1999
36
4.3.3. Ground and Polished Stone
Excavations to date have found between
ten and twenty artifacts of ground and polished
stone. The exact number is difficult to determine,
since some pieces are clearly modified by humans,
while others pieces consist of stone not found
naturally on the site but do not display clear traces
of working or shaping. It should also be noted that
many of these are difficult for excavators to identify
and a certain number of ground stone artifacts,
particularly fragments or ones not obviously
modified, may have been discarded accidentally
during excavation.
Non-flaked stone artifacts fall into several
categories:
• polished stone axes and other tools, and
possible fragments from working or damaging
them
• fragments of large grinding stones, of
limestone or sandstone, with clearly worn
surfaces
• fragments of large metamorphic or igneous
pebbles, sometimes showing worked facets or
surfaces
• small pebbles of unusual stone without any
clear signs of human modification
Polished stone axes and other tools. To
date, two polished stone axes from Umbro have
been found. The first was a broken fragment from
the pointed (butt) end of a small, greenish-black
axe or adze tool which was found in Stratum II of
Trench I in 1998 (see Robb 1998 for further detail).
Excavations in 1999 uncovered a polished
stone axe in Stratum II of Trench I (-10n/-
36e)(Figure 30). Because of its position close to the
surface (120-130 cm below datum, 10-20 cm
below surface, in loose, sloping sediments), it is
almost sure not to be in original depositional
context. The axe is made of blackish amphibolite
originally from the Sila (P. Barrier, pers. comm.). It
measures 85 mm long by 47 mm high by 36 mm
wide, and weighs 254 grams. The axe has an
unusually round, stocky or barrel-shaped form
which is unlike most published examples of
prehistoric axes from this region. Most of the body
is highly polished and glossy. However, the small
(butt) end is blunted with post-polishing damage of
some kind, and the large (blade) end is in the
process of being reworked. The blade end has been
broken off flat to form a cylindrical end; a diagonal
spall has been flaked off this flat surface and
partially smoothed, possibly as a beginning of re-
fashioning a point to it; and in a band around the
end on the sides of the axe the polished surface has
been removed by pecking with a hard stone. The
most likely interpretation seems that this piece was
originally a much larger axe which was in the
process of being remodelled into a smaller axe
when it was discarded. The thickness, too great for
a small axe, and the highly polished surface remain
from the larger axe, which has been trimmed to a
new, shorter length and was just starting to be
thinned and re-pointed.
Both of these axes are reasonably similar
to other axes known from the area, for instance the
three from loc. Cavalli near Bova Superiore in the
Bruno Casile collection (examined in 1997 by kind
permission of Sig. B. Casile). Unfortunately, such
axes could have been used in any period from the
Early Neolithic through the Bronze Age, and the
relatively high stratigraphic positions of both
Umbro examples does little to resolve their dating
(see above). Polished stone axe manufacture may
also be evident in the form of occasional thin spalls
of hard stones alien to the site, whether worked or
not.
A more unusual polished stone artifact was
the replica stone axe (Figure 30). Found close to
the surface in Trench 1 (-10n/-36e, surface to 120
cm), it cannot be considered in a secure
stratigraphic context. It consists of a small, shaped
and polished axe made of a locally available,
streaky white, black and brown phyllite with mica
(P. Barrier, pers. comm.). The surface is uniformly
smooth except where small areas are broken away.
The blade end is sharpened, but displays no use
damage. The axe measures 4.5 mm thick, 10 mm
high, and at least 17 mm long (part is missing). The
stone is soft and has strong cleavage planes, which
would rule out the possibility of using it as a chisel
or other small hand tool. Rather, it appears to be
simply a replica axe, perhaps for use as a toy or
amulet. Tiny replica axes or adzes are not unknown
in the Southern Italian Neolithic, though most
reported examples seem to be greenstone.
Large grinding stones. Several stone
fragments represent large querns or grinding stones
made from relatively soft sedimentary rock. One
flat chunk of sandstone (135 x 112 x 48 mm, from
Stratum II in the north end of Trench 1) had been
shaped into an oval form; one flat face of it had
been ground very smooth, while the other face and
the side had been left rough. This fragment is
probably a large piece of a two-handed rubber or
mano. A second flat piece of sandstone with one
very well-worn side (41 x 35 x 31, from the same
general context) is probably part of a similar tool.
Large querns are known from Umbro from
only two fragments. This is a piece of a limestone
slab 120 mm by 110 mm and 41 mm thick. One
Bova Marina Archaeological Project 1999
37
edge of the fragment is raised with a lip ca. 1 cm
high, so that the piece forms part of a very shallow
basin. The top surface is smoothed. This piece
probably formed part of a large flat quern or metate
of a type common in Italian Neolithic villages. A
second flat limestone fragment without a lip but of
similar dimensions (90 x 85 x 27) and smoothness
probably comes from the same or a similar quern.
Note that these two fragments were found
in Trench 6, in an Early Bronze Age context. Two
granite pebbles were also found in this context, one
of which had a smooth facet and the other of which
had possible ochre stains. It is possible that these
are genuine EBA querns and grinding stones
deposited in either a ritual or a domestic context;
alternatively, they may be Neolithic tools re-used as
building stone during the Bronze Age.
Raw materials for these tools could have
come from the site of Umbro itself, for limestone,
or from nearby sandstone outcrops.
Large pebbles. A third class of ground
stone tools consisted of pebbles of granite and other
non-local rock. These tools were identifiable in
some cases by their raw material, and in other cases
by evident modification. A total of eight were
collected, two from Trench 6 and six from Trench
1.
• granite pebble with two grinding facets (86 x
46 x 31 mm, broken; Bag 1066)
• granite pebble with one side ground smooth
(88 x 88 x 44 mm, broken; Bag 1121)
• granite pebble with several smooth grinding
facets and a slight pink/red stain, perhaps from
ochre (60 x 41 x 45 mm, broken; Bag 1134)
• pebble of unidentified metamorphic stone with
several oblique grinding facets (76 x 58 x 37
mm, broken; Bag 1135)
• granite pebble with one side probably
artificially smoothed (57 x 40 x 37 mm,
broken; Bag 1149)
• pebble of unidentified stone with one side
probably artificially smoothed (69 x 58 x 39
mm, broken; Bag 1149)
• granite pebble with smoothed surfaces (79 x 58
x 35 mm; Bag 1160, from Trench 6)
• granite pebble without visibly smoothed facets
but with possible ochre stain (68 x 56 x 45
mm; Bag 1150, from Trench 6)
It is likely that other examples were discarded
inadvertently by excavators.
The raw material for these pebble tools
most probably came from the sandstone/
conglomerate outcrops 200 meters north of Umbro.
Their function was clearly grinding, as
demonstrated by the flat facets worn on many of
their surfaces, and they were evidently curated and
reused for some time, to judge from the multiple
grinding facets many display. They may not have
been the normal tools for grinding grain, and at
least some may have been used to crush and grind
ochre pigments, as several display reddish stains. It
is not clear why most of them were broken; perhaps
they were also used as hammerstones.
Small pebbles. Finally, a number of small
pebbles were found which were imported into the
site and probably used as expedient tools. These
include six examples from Trench 1 (Bag 1010;
Bag 1136, 2 examples; Bag 1161; Bag 1211; Bag
1216). Only one shows a possible smoothed facet
(Bag 1211); the rest are identified solely by their
non-local raw material. Other examples have
probably been discarded by excavators. These
pebbles are usually found unbroken and their
maximum dimension is usually less than 50 mm.
They may have been used for some purpose such as
smoothing and burnishing unfired vessels.
4.3.4. Ochre
To date, eight small fragments of ochre
have been found. All come from Trench 1, where
they are found throughout the stratigraphy in all
levels. The largest fragment is 21 mm in diameter,
but most are smaller than 5 mm. They display no
form or wear facets, but have clearly been imported
to the site by humans for use as a pigment. This
interpretation is corroborated by ochre stains on
grinding stones (see above) and the use of red
pigment to color impressions in vessels.
It is worth noting that several levels in
Trench 1 displayed slight concentrations of small,
chalklike nodules. These were not collected, as they
were soft, disintegrating, and at times difficult to
recognize. Such white lumps may represent a
natural soil concretion, but they may also represent
the remnants of materials used to make a white
paste for encrusting impressed pottery.
4.4. Faunal Remains and Human Remains
Faunal Remains. The 1999 excavations
recovered 892 fragments of animal bone; of these,
5 came from Trench 6, 3 from Trench 7, and all the
rest from Trench 1. Counting material from the
1998 excavations, this brings the total faunal
remains from Trench 1 to 1340 pieces (Table 2).
This body of material will be studied by a trained
faunal analyst following the 2000 field season. The
following preliminary comments should be
Bova Marina Archaeological Project 1999
38
understood as based on rapid impressions in the
field by people untrained in animal bone
identification.
Taphonomically, the animal bone from
Umbro is remarkable for its poor preservation.
Most pieces are less than one centimeter in
maximum dimension and are probably not
identifiable. The most common identifiable pieces
will almost certainly be tooth fragments. This
marked degradation is not typical of extensive
Neolithic village sites, in my experience, and serves
to underscore the particular nature of Umbro. It
probably derives from a very long-term use of a
very restricted habitation area. It may also have
something to do with the particular use of animal
resources or food preparation or consumption at the
site.
Between one third and one quarter of all
fragments are burnt, usually to a shade between
black and white. Burning is found on bones from
all strata of Trench 1 down to the earliest levels.
This indicates exposure to extreme heat, as opposed
to a brownish color which can indicate exposure to
lesser temperatures. Intense burning on a par with
cremation seems hardly compatible with thermal
alteration in the process of cooking. Instead, it
suggests that the sediments around Umbro were
intensely heated periodically or repeatedly. Such
intense burning no doubt contributed to the high
fragmentation of the faunal remains.
As an informal impression, the commonest
kind of animal at the site seems to have been a
small herbivore, most likely sheep and/or goat,
though at this stage deer cannot be excluded.
Several bones from very young individuals
(lambs?) were found; these may prove informative
on the season of occupation of the site. Pig has
been identified as definitely present from several
pieces submitted to faunal analysts (U. Albarella,
pers. comm.; D. Serjeantson, pers. comm.), and
teeth resembling pig are moderately common. It is
unclear whether the pig would have been
domesticated, wild boar, or feral. Cattle are
probably represented by several fragments, but
seem very uncommon. One fragment of cattle bone
appears to be a nearly complete metapodial with
numerous cut marks.
Other animals present include bird of
some kind, snake and rodents. It is likely that some
snake and rodent bones are intrusive modern
specimens, especially as at least snakes now inhabit
the site. Fish bones appear to be completely absent.
Human Remains. In contrast to the 1998
excavations which yielded a large assemblage of
disarticulated bone from Trench 3 (which later
proved to be probably Roman in date), the 1999
excavations yielded only three human skeletal
fragments. All came from Stratum II in Trench 1.
One consisted of a proximal pedal phalange
fragment. The other two consisted of small cranial
vault fragments whose precise location on the skull
could not be determined. Like the four fragments of
human bones and teeth found in Trench 1 in 1998,
these seem to be isolated fragments probably
transported in from deposits higher up the southern
slopes of the site and without any particular
significance for the function of this part of the site.
Supporting this interpretation is the fact that all
seven fragments were found in Stratum II, a
relatively high stratum probably representing mixed
deposits.
4.5. Paleobotanical Remains (Marina Ciaraldi)
In 1999, we began a flotation program.
Flotation samples were collected from all the
features in Trench 6, for which the entire fill was
bagged. This yielded four samples (one each from
the three pots and one from the surrounding soil
matrix). Samples were also taken from four
representative levels in Trench 1, for which 20 liter
samples were collected. For consistency, all Trench
1 samples were taken from the same grid square (-
7n/-36e); one sample was taken from each of strata
II-V. Finally, one sample was taken from a feature
in Stratum IV of Trench 1 (–8n/-36e). Flotation
samples were first sifted in 5 mm mesh, like all
excavated soil, and the finds were recorded and
reunited with other finds from the bag. The samples
were then floated by hand using basins and hand
strainers. The coarse fraction was collected using 1
mm nylon mesh, and the fine fraction was collected
using a .5 mm nylon mesh. These fractions were
then dried in fine mesh bags. The residue and the
flotation (fine) fraction were both sorted and
identified for seed and charcoal analysis by Dr.
Marina Ciaraldi (Birmingham University Field
Archaeological Unit). In the case of the flotation
fraction, a standard low power stereomicroscope
was used. Identifications were made using a
reference collection of modern material and various
identification atlases.
None of the four samples from the Trench
6 (Early Bronze Age) pots and feature fill
contained organic remains. This is unsurprising as
most of these samples were less than a liter in
volume, and the largest was less than five liters.
The remaining five samples all yielded
botanical remains (Table 5). Sample 9 represents
Stratum II, possibly mixed later Neolithic deposits;
Sample 7 represents Stratum III (Diana); and
Samples 5, 8, and 6 represent Stentinello deposits
Bova Marina Archaeological Project 1999
39
from Stratum IV, Stratum V and a feature
respectively.
Many modern rootlets and some modern
seeds were observed in sample 9 and, to a minor
degree, in sample 7. The remaining three samples
presented a very similar sandy matrix. All the
samples contained some charcoal (although in very
small quantity) and all samples but N.6 contains
some small bones. The general preservation of the
charred plant remains was generally poor. The
cereal grains had clearly undergone carbonization
under intense heat and they presented a damaged
surface and numerous holes. The same applied to
some of the chaff elements. The identification to
species level was therefore difficult and in many
cases impossible. Most of the cereal grains
recovered were fragmented and distorted and they
were assigned to the wide category of cereals.
In some cases it was possible to attribute
some of the grains to barley (Hordeum vulgare L.).
However, even in these cases it was impossible to
observe details such as the symmetry of the
embryo, the presence of hull on the surface of the
grain or eventual signs of germination. It was
therefore impossible to establish whether these
grains belonged to the six-row or to the two-row
variety. In two cases some barley grains were
identified simply as Hordeum sp. because they
were rather small and they could have been either
tail grains (underdeveloped grains) of cultivated
barley or a species of barley.
Only two wheat grains were found and
only one was identified as bread wheat (Triticum
aestivum s.l.). The diagnostic features used to
identify this grain were its roundness and its curved
bottom profile. Most of the charred chaff found was
damaged too and although the presence of emmer
(Triticum dicoccum Schank.) was confirmed by the
identification of some of the better preserved glume
bases and forks, it was impossible to establish with
certainty the presence of einkorn (Triticum
monococcum L.).
None of the weed seeds found is particular
informative in the interpretation of the assemblage
mainly because they could not be identified to
species level. One possible olive pit was recovered
from a good Early Neolithic context and identified
in the field (identification by L. Foxhall); however,
this did not form part of the material available for
identification by a trained paleobotanist, and the
identification must be considered tentative. If it is
indeed an olive pit, it is likely to represent wild
olive and may have been used for animal fodder (L.
Foxhall, pers. comm.).
The paucity of the plant remains recovered
does not allow to make any general statment on the
economy of the site. However, considering the
general dearth of data from the Neolithic of the
area, the plant material from Umbro provides an
important contribution to the more general
reconstruction of the Neolithic economy of
Calabria.
Overall the plant assemblage from Umbro
represents a typical assemblage from a Southern
Italian Neolithic site, with the simultaneous
presence of three different cereals: barley, emmer
(and einkorn?) and a free-threshing wheat. This
group of cereals represents the typical Neolithic
“package” of Southern Europe (Zohary and Hopf
1993). The absence of pulses from the plant
assemblage may be the consequence of the
generally poor preservation of the plant material.
The agricultural component of the economy may
also be attested by the common finds of grinding
stone fragments (see above).
There are only two Neolithic sites of the
Stentinello facies from Calabria whose plant
assemblage has been studied: Capo Alfiere,
Crotone (CZ) (Costantini 1988) on the Ionic coast
and Acconia area C, Curingia (CZ) (Ammerman
1987) on the Tyrrenic coast. The information on the
plant assemblage from Acconia is very limited and
it only attests the presence of barley (Costantini and
Stancanelli 1994). For the plant assemblage from
Capo Alfiere we only have a preliminarily report
but there is evidence of a higher number of species.
The plant assemblage from Capo Alfiere reflects
closely the findings from Umbro although there is
no reference to the presence of chaff at Capo
Alfiere (Costantini and Stancanelli 1994).
The presence of grains, chaff and a few
weeds (possibly associated with the cereal crops) is
an important indication that part of the crop
processing took place on site. The crop processing
may have happened at different levels. It could
have taken place as a large-scale crop cleaning
directly on site or it may just have taken place as
daily cleaning of cereals to be eaten. A higher
density of charred plant remains was recovered
from the feature fill (Sample 6, T1 8n.36e –230-
240). Its composition was very similar to the rest of
the plant assemblage. This could perhaps be
interpreted as evidence of the fact that crop
cleaning took place on a daily basis and that the
plant assemblage recovered from the feature fill
reflects this activity.
It is evident from the data that there is no
clear differentiation between the assemblages from
the different strata and occupational periods. If in
some cases this can be due to a certain degree of re-
mixing of the layers (especially for Stratum II and
III), on the other hand this might be interpreted as a
sign of continuation between the Stentinello and
Bova Marina Archaeological Project 1999
40
Diana occupation levels. This seems in agreement
with the hypothesis that these two kinds of ware
might have been in use at the same time over a
certain number of centuries (Whitehouse 1986).
However the data are too scarce to be used with
confidence in confirming this statement.
Bova Marina Archaeological Project 1999
41
5. TEST EXCAVATIONS AT SAN PASQUALE
San Pasquale is a Neolithic site located on
a natural clay terrace at the eastern margin of the
San Pasquale valley. It was found by S. Stranges
and L. Saccà and first surveyed by the Bova Marina
Archaeological Project in 1997 (see survey site
catalog, above, in which it is referred to as
Canturatta A, Area 2). Both Stranges and Saccà and
our survey have found pottery to include mostly
small, highly eroded fragments with a few
diagnostic impressed or Stentinello pieces; there
are also some fragments which may be later
prehistoric and some apparently Roman fragments
which may be associated with Site 2.
The natural setting of the site would have
been quite attractive for Neolithic settlement. It is
located about 100 m from the present day coastline
and about 30 m above sea level, overlooking the
broad, level valley of the Torrente di S. Pasquale.
The terrace itself consists of pure clayey and sandy
deposits, as its eroding margins show. A deep
ravine divides it into an eastern and a western part.
Like most such terraces, it has now been built up
with a cascade of small agricultural terraces, but it
seems likely that this reflects the underlying shape
of the hill substantially. In prehistoric times, the
torrente may have had marshy areas at its mouth,
and the valley would have probably been wooded.
Plenty of level ground would have been available
for building and farming, as well as marine
resources and a variety of montane resources from
the Agrillei ridge just behind the site. The ridge
would have also sheltered the site from northern
and eastern winds. In addition, there may well have
been springs available at points around the base of
the Agrillei ridge; a large clump of reeds near the
eastern end of the terrace suggests some ground
moisture, and a spring is still extant at a similar
stratigraphic level below the ridge on the western
side of the valley.
Such a location – on a low terrace
commanding ecotonal resources – is often
considered virtually stereotypical of Neolithic
villages, based on areas such as the Materano and
the Tavoliere (Jarman and Webley 1975; Jones
1987; Tinè 1983). Hence, if we suppose that the
Calabrian Neolithic was based around large villages
as in other areas, San Pasquale seems an obvious
area to look for one. Thus, we were especially
interested in test-excavating San Pasquale to
investigate sites complementary to Umbro, which is
clearly not a large, open village.
At present, the area of the site is used for
several kinds of agricultural production. On the
western part, the terrace is in derelict scrub, and
had recently been burnt over at the time of
excavation; hence surface visibility was excellent.
Across the ravine on the broader eastern part, there
was an active and fenced vineyard and, further east,
a grove of olive and almond trees with some
scrubby vegetation between them.
As a preliminary to test excavations, we
examined current ground surfaces as well as
interviewing past surveyors (especially D. Yoon
and L. Saccà) as to the location of artifact
concentrations. It is clear that some cultural
materials are found throughout the terrace, but that
there are at least two distinct concentrations. One is
located in the center of the western part of the
terrace. Here we found, for instance, scattered
prehistoric and historic sherds, all small eroded
fragments. A second is located along the southern
margin of the center of the eastern plateau, between
a large outcropping boulder (collected as “Location
B”) and the nearby olive trees (collected as
“Location A”). Finds here seemed to be
predominantly Roman or modern, though at least
one flake of obsidian suggests prehistoric
occupation.
We decided to excavate test pits on the
western part of the terrace in the center of the
concentration of sherds there (Figure 31). Two one
by one meter test pits were excavated, Trench 1 and
Trench 2 (Figure 32, Figure 33). The two pits were
aligned to a hypothetical north-south grid and
spaced 16 meters apart. They were excavated in ten
centimeter arbitrary levels. Work was carried out
with a crew of 8 on September 9th
.
The results were disappointing, but at least
they were clear. In both pits, the ground surface
was littered with small, tabular gravel, consisting
largely of schist fragments evidently originating
from the slopes of Agrillei just 40-60 meters to the
northeast. When this was removed, the soil rapidly
turned to a brownish sandy clay which contained
virtually no archaeological materials. As the
trenches grew deeper, the soil grew yellower and
both stones and bulbs and roots from surface
vegetation became fewer and fewer. Both trenches
were stopped between 40 and 50 cm deep, when it
was clear that the soil below the surface was
essentially sterile and was rapidly turning identical
to the 20+ meter thick deposit of clay visible in the
eroding banks of the terrace. Both trenches were
backfilled.
Bova Marina Archaeological Project 1999
42
While these excavations were taking
place, three crew cleaned profiles in eroding places
at four locations around the edges of the terrace. In
all four places, with profiles up to three meters
deep, we encountered stratigraphy identical to that
in the test trenches: a litter of stone fragments on
the surface, followed by a thin transition to sterile
geological deposits.
The main, and apparently indisputable,
conclusion is that the archaeological deposits in this
particular place are entirely superficial. A
suggestion was made by L. Saccà that farmers
earlier in the century, to improve the quality of land
for vines, dumped gravel into the field; few remains
were visible on the surface because the real site lay
buried a meter or more deep. However, this concept
does not account for a number of facts: some
artifacts are indeed found on the surface, the scatter
of stones is entirely superficial rather than a thick
stratum well mixed with the soil; and nowhere in
exposed sections is there any evidence of a buried
site outcropping at any depth.
Instead, the most likely interpretation is
that both the sheet of thin schist fragments and the
scatter of artifacts identified as the site have washed
in from the ridge to the northeast. This is the
geological source of the schist, which has nothing
to do with the soft sedimentary deposits of the
terrace. It would account for the apparent mixing of
the fragments, and would also explain their highly
fragmented and eroded state, which is so severe that
virtually none of them retains its original surface.
The concentration of both stones and sherds on the
surface of the field where we excavated can be
explained through repeated cycles of alluviation
and deflation.
We thus concluded that there is a genuine
Neolithic site somewhere in the vicinity of the area
excavated, probably at no great distance. Although
some of the relevant areas have been fieldwalked
and do indeed yield low-level concentrations of
both prehistoric and historic material, it is unclear
whether these represent the real site or are also
slopewash. The task for the future is to walk and re-
walk all areas in the environs, including some steep
and thickly brushy areas, to check where the site
may have been.
Bova Marina Archaeological Project 1999
43
6. CONCLUSIONS AND FUTURE RESEARCH DIRECTIONS
6.1. Umbro: Dating, Stratigraphy, and Site
Function
The dating and stratigraphy of Umbro are
gradually becoming clearer, thanks to numerous
small test excavations and radiocarbon dates. The
site was first occupied in the early 6th
millennium
BC by people using Stentinello wares. The area
occupied then consisted of a small zone in the
shelter of low, east-facing cliffs. Stentinello
occupation appears to have continued, continuously
or intermittently, at least through the middle of the
5th
millennium BC and likely for longer, as Late
Neolithic Diana wares are common. There also
appears to have been some Copper Age occupation,
although this is poorly defined. Following a lapse
of perhaps a thousand years, the site was reutilized
in the Early Bronze Age by people using ceramics
akin to the Eastern Sicilian Rodi-Tindari-
Vallelunga wares. Nothing is known of later Bronze
Age or Iron Age use of the site, but a small Greek
farmstead was occupied from Archaic times
onward.
The nature of occupation in each period is
less clear and more interesting. In the Neolithic,
Umbro was a very small site, which could have
been occupied by perhaps five to ten people at
most. Influenced by areas such as Puglia, the
Materano, the Adriatic coastal strip and eastern
Sicily, archaeologists have usually understood
Neolithic settlement as centered around villages
housing many families. If this settlement pattern
was found in Calabria as well, then Umbro must be
understood as a special function site of some kind.
It would have been occupied by a fraction of
society performing some combination of
specialized tasks such as hunting, herding, making
pottery, or conducting rituals. Such a group may
have been a nuclear family, or it may have been
recruited along age, gender, ritual or other lines.
However, we must also consider the possibility that
Neolithic settlement in Calabria did not follow the
pattern established elsewhere; it could have been
decentralized, without fixed, aggregated villages
(cf. Whittle 1996). Evidence from Trench 1 gives
some idea of the nature of activities carried on at
the site. Pots were made, probably in the dry
season; cereals were prepared for consumption;
domestic animals were eaten; exotic materials were
fashioned into axes and cutting tools.
The function of Umbro in the Early
Bronze Age is currently equally ambiguous; as
known from limited preliminary excavations, it may
have been a burial site, a ritual site or a domestic
site.
The function of Umbro is inextricably
linked to use of the surrounding landscape. For
both the Neolithic and the Early Bronze Age,
understanding the function of the site depends on
understanding other sites in the neighborhood –
whether Neolithic villages existed nearby, and
whether there were domestic or funerary Bronze
Age sites nearby.
6.2. The Landscape of Bova Marina from
Neolithic through Recent Times
General trends in settlement as known
from the field survey have been discussed above
and are only briefly summarized here. Prehistoric
settlement is theoretically poorly understood for the
Neolithic and not known for later periods.
However, survey data shows early settlement by
Greeks, or acculturation of native populations to
Greek customs. Surprisingly, hellenization occurred
early even at small, inland farmsteads as well as in
cities such as Rhegion. Probably the densest
settlement in antiquity took place in the later
Roman period, between the second and fifth
centuries A.D., when intensive occupation of
coastal plains and river valleys probably reflects
villa-based intensive agriculture. Little is known of
the following early medieval period until the
foundation of inland hilltop towns such as Bova
Superiore. Population rose again in historic times,
with dense peasant land use bespoken by numerous
now-abandoned farmhouses. Perhaps the final turn
in recent settlement history is a general shift in the
last century back to the coastlands with the gradual
abandonment of traditional mountain villages.
6.3 Future Work
Future work includes several logical
developments of this season’s researches.
Our field survey plans include three
primary objectives:
(1) to continue survey in the comune of
Bova Marina, exploring especially inland
territories.
(2) to expand survey inland to the comune
of Bova Superiore, in order to understand
settlement in the entire range of environmental
zones locally available and used in historical times.
Bova Marina Archaeological Project 1999
44
Areas to target specifically include the hills
between Umbro and Bova Superiore, the immediate
surroundings of Bova Superiore, and the Campi di
Bova, a small highland plain traditionally used as
pastures for people based at Bova.
(3) to begin a program of small test
excavations in undated prehistoric sites, to check
the presence of periods such as the Copper and
Bronze Ages which are very difficult to identify
from survey materials and to verify the presence of
Neolithic sites in the surroundings of Umbro. The
targets for 2000 include a probable Bronze Age site
at Limaca, about 200 meters north of Umbro, and
two undated ceramic scatters about 300 meters
southwest of Umbro.
The excavation goals are straightforward.
We hope to continue excavations in the Neolithic
area, though on a limited scale. The specific goals
will be to further define the chronology of the site,
to ascertain the beginning of occupation at the base
of the sequence near the cliff, and to recover further
scientific samples (for example, for flotation). We
hope to conduct substantial excavations in the Early
Bronze Age area, primarily to understand the nature
of the site and, if it turns out a burial site, to learn
about EBA funerary ritual. Finally, we hope to
conduct test excavations in the Greek site about
200 meters south of Umbro. As noted earlier,
excavation of a small, early inland Greek farmstead
has the potential to yield much new information
about social processes of acculturation in the early
colonial period.
Bova Marina Archaeological Project 1999
45
BIBLIOGRAPHY
Ammerman, A.
1987 Ricenti contributi sul neolitico della
Calabria. Atti, Riunione Scientifica
dell'I.I.P.P. 26:333-349.
Bernabò Brea, L.
1957 Sicily before the Greeks. Thames and
Hudson, London.
Bernabò Brea, L., and M. Cavalier
1956 Civiltà preistoriche delle Isole Eolie e
del territorio del Milazzo. Bullettino di
Paletnologia Italiana 66:7-98.
1960 Meligunìs Lipára. Volume I: La
stazione preistorica della contrada Diana e
la necropoli preistorica di Lipari.
Pubblicazioni del Museo Eoliano di Lipari.
S. F. Flaccovio, Palermo.
1980 Meligunìs Lipára, Volume IV:
l'acropoli di Lipari nella preistoria.
Publications of the Museo Eolio. Flaccovio,
Palermo.
Castellana, G.
1996 La stipe votiva del Ciavolaro nel
quadro del Bronzo Antico Siciliano.
Assessorato regionale beni culturali
ambieltali e pubblica istruzione, Agrigento.
Ceci, M.
1992 Note sulla circolazione delle lucerne
a Roma nell'VIII secolo: i contesti della
Crypta Balbi. Archeologia Medievale
19:749-766.
Cocchi Genick, D.
1996 Manuale di Preistoria. III. Età del
Rame. Octavo, Firenze.
Costamagna, L.
1991 La sinagoga di Bova Marina nel
quadro degli insediamenti tardoantichi della
costa jonica meridionale della Calabria.
Mèlanges de l'école française di Rome 2.
Costantini, L.
1988 Cereali e legumi del sito neolitico di
Capo Alfiere, Unpublished manuscript.
Costantini, L., and M. Stancanelli
1994 La preistoria agricola dell' Italia
centro-meridionale: il contributo delle
indagini archeobotaniche. Origini 18:149-
243.
Foley, R.
1981 A model of regional archaeological
structure. Proceedings of the Prehistoric
Society 47:1-17.
Garcea, F., and D. Williams
1987 Appunti sulla produzione e
circolazione delle lucerne nel napoletano tra
VII e VIII secolo. Archeologia Medievale
14:537-545.
Gregg, S.
1988 Foragers and farmers: population
interaction and agricultural expansion in
prehistoric Europe. University of Chicago
Press, Chicago.
Higgs, E., and C. Vita-Finzi
1972 Prehistoric economies: a territorial
approach. In Papers in economic
prehistory, edited by E. Higgs, pp. 27-36.
Cambridge University Press, Cambridge.
Jarman, M., and D. Webley
1975 Settlement and land use in
Capitanata, Italy. In Palaeoeconomy, edited
by E. Higgs, pp. 177-231. Cambridge
University Press, Cambridge.
Jarman, M. G., G. Bailey, and H. Jarman
1982 Early European agriculture.
Cambridge University Press, Cambridge.
Jones, G.B.D.
1987 Apulia. Society of Antiquaries,
London.
Leighton, R.
1999 Sicily before history. Duckworth,
London.
Maniscalco, L.
1996 Early Bronze Age funerary ritual and
architecture: monumnetal tombs at Santa
Febronia. In Early Societies in Sicily, edited
by R. Leighton, pp. 81-87. Accordia
Research Centre, London.
1999 The Sicilian Bronze Age pottery
service. In Social Dynamics of the
Prehistoric Central Mediterranean, edited
by R. H. Tykot, J. Morter, and J. E. Robb,
pp. 185-194. Accordia Research Centre,
London.
Bova Marina Archaeological Project 1999
46
Morter, J.
1992 Capo Alfiere and the Middle
Neolithic period in eastern Calabria,
Southern Italy. Unpublished Ph.D.
Dissertation, University of Texas, Austin.
O'Hare, G.
1990 A preliminary study of polished stone
artefacts in prehistoric southern Italy.
Proceedings of the Prehistoric Society
56:123-152.
Orsi, P.
1890 Stazione neolitica di Stentinello.
Bullettino di Paletnologia Italiana 16:177-
200.
Robb, J.
1997 Bova Marina Field Survey:
Preliminary Report, 1997 Season.
Department of Archaeology, University of
Southampton, Southampton.
Robb, J. E.
1998 Bova Marina Archaeological
Project: Survey and Excavations at Umbro.
Department of Archaeology, University of
Southampton, Southampton.
Servizio Geologico dell'Italia
1970 Carta Geologica dell'Italia. Cassa
per il Mezzogiorno. 1:25000.
1976 Carta Geologica dell'Italia. IGM.
1:500000.
Shaffer, G.
1985 Architectural resources and their
effect on certain neolithic settlements in
Southern Italy. In Papers in Italian
Archaeology IV: the Cambridge
conference, edited by C. Malone, and S.
Stoddart, pp. 101-117. BAR International
Series, 245. British Archaeological Reports,
Oxford.
Skeates, R.
1992 Thin-section analysis of Italian
neolithic pottery. In Papers of the Fourth
Conference of Italian Archaeology. Volume
3: New developments in Italian
archaeology, edited by E. Herring, R.
Whitehouse, and J. Wilkins, pp. 29-34.
Accordia Research Center, London.
1994 A radiocarbon date-list for prehistoric
Italy (c. 46,400 BP - 2450 BP/400 cal. BC).
In Radiocarbon dating and Italian
prehistory, edited by R. Skeates, and R.
Whitehouse, pp. 147-288. Accordia
Specialist Studies on Italy, 3. Accordia
Research Center, London.
Stranges, S.
1992 Importante ritrovamento
stentinelliano a Bova Marina. Calabria
Sconosciuta 15:51-52.
Stranges, S., and L. Saccà
1994 Nuove acquisizioni sulla preistoria
nella Jonica Reggina. Calabria Sconosciuta
17:19-22.
Tinè, S.
1983 Passo di Corvo e la civiltà neolitica
del Tavoliere. Sagep, Genova.
1992 Bova Survey 1992. Istituto Italiano di
Archeologia Sperimentale, Genova.
Tozzi, C., and G. Tasca
1989 Ripa Tetta. Atti Convegno Nazionale
sulla Preistoria, Protostoria e Storia della
Daunia 11:39-54.
Tusa, S.
1993 La Sicilia nella preistoria. 2nd ed.
Sellerio, Palermo.
Whitehouse, R.
1986 Siticulosa Apulia revisited. Antiquity
60:36-44.
Whittle, A.
1996 Neolithic Europe: the creation of new
worlds. Cambridge University Press,
Cambridge.
Zohary, D., and M. Hopf
1993 Domestication of plants in the Old
World. Clarendon, Oxford.
Bova Marina Archaeological Project 1999
47
FIGURES AND TABLES
Figures
1. Map of areas surveyed and sites found, 1997-1999.
2. Map of areas surveyed in the Umbro area.
3. Map of areas surveyed in the Mazza area.
4. Map of grid collection units at Mazza.
5. GIS reconstruction of prehistoric land use.
6. Umbro: site map.
7. Umbro: general view of cliffs with Neolithic site below, Bronze Age site above (from north).
8. Trench 1 plan.
9. Trench 1 section.
10. Trench 1 general view at end of excavations.
11. Trench 6 profile of road cut before excavation of trench.
12. Trench 6 plan of excavations.
13. Trench 6 profile of excavations.
14. Trench 6 plan of pot group.
15. Trench 6 general view of Bronze Age surface and rock fall.
16. Trench 6 Bronze Age pots in situ.
17. Trench 7 plan.
18. Trench 7 profile.
19. Trench 7 general view.
20. Clearing bedrock near peak of site.
21. Area 2 general view (1998 excavations).
22. Trench 3 general view (1998 excavations).
23. Trench 4 general view (1998 excavations).
24. Trench 5 general view (1998 excavations).
25. Umbro: Stentinello pottery.
26. Umbro: Diana pottery.
27. Umbro: Other prehistoric pottery.
28. Umbro: Early Bronze Age vessels from Trench 6 deposition.
29. Umbro: miscellaneous tools: (a) stampini, (b) worked bone, (c) worked shell.
30. Umbro: ground stone (a) axe, (b) miniature replica axe.
31. San Pasquale general view of location of test pits.
32. San Pasquale Trench 1 test pit.
33. San Pasquale Trench 2 test pit.
Bova Marina Archaeological Project 1999
48
Tables
Table 1. Trench 1 finds and taphonomic characteristics by depth.
Table 2. Summary of faunal assemblage from Trench 1.
Table 3. Summary of lithic assemblage from Trench 1.
Table 4. Summary of daub frequencies in Trench 1.
Table 5. Paleobotanical samples.
Table 6. Radiocarbon dates.
Bova Marina Archaeological Project 1999
49
1. Map of areas surveyed and sites found, 1997-1999.
Bova Marina Archaeological Project 1999
50
2. Map of areas surveyed in the Umbro area.
Bova Marina Archaeological Project 1999
51
3. Map of areas surveyed in the Mazza area.
Bova Marina Archaeological Project 1999
52
4. Map of grid collection units at Mazza.
Bova Marina Archaeological Project 1999
53
5. GIS reconstruction of prehistoric land use.
Bova Marina Archaeological Project 1999
54
6. Umbro: site map.
Bova Marina Archaeological Project 1999
55
7. Umbro: general view of cliffs with Neolithic site below, Bronze Age site above (from north).
Bova Marina Archaeological Project 1999
56
8. Trench 1 plan.
Bova Marina Archaeological Project 1999
57
9. Trench 1 section.
Bova Marina Archaeological Project 1999
58
10. Trench 1 general view at end of excavations.
Bova Marina Archaeological Project 1999
59
11. Trench 6 profile of road cut before excavation of trench.
Bova Marina Archaeological Project 1999
60
12. Trench 6 plan of excavations.
Bova Marina Archaeological Project 1999
61
13. Trench 6 profile of excavations.
Bova Marina Archaeological Project 1999
62
14. Trench 6 plan of pot group.
Bova Marina Archaeological Project 1999
63
15. Trench 6 general view of Bronze Age surface and rock fall.
Bova Marina Archaeological Project 1999
64
16. Trench 6 Bronze Age pots in situ.
Bova Marina Archaeological Project 1999
65
17. Trench 7 plan.
Bova Marina Archaeological Project 1999
66
18. Trench 7 profile.
Bova Marina Archaeological Project 1999
67
19. Trench 7 general view.
Bova Marina Archaeological Project 1999
68
20. Clearing bedrock near peak of site.
Bova Marina Archaeological Project 1999
69
21. Area 2 general view (1998 excavations).
Bova Marina Archaeological Project 1999
70
22. Trench 3 general view (1998 excavations).
Bova Marina Archaeological Project 1999
71
23. Trench 4 general view (1998 excavations).
Bova Marina Archaeological Project 1999
72
24. Trench 5 general view (1998 excavations).
Bova Marina Archaeological Project 1999
73
25. Umbro: Stentinello pottery.
Bova Marina Archaeological Project 1999
74
26. Umbro: Diana pottery.
Bova Marina Archaeological Project 1999
75
27. Umbro: Other prehistoric pottery.
Bova Marina Archaeological Project 1999
76
28. Umbro: Early Bronze Age vessels from Trench 6 deposition.
Bova Marina Archaeological Project 1999
77
29. Umbro: miscellaneous tools: (a) stampini, (b) worked bone, (c) worked sherd.
Bova Marina Archaeological Project 1999
78
30. Umbro: ground stone (a) axe, (b) miniature replica axe.
Bova Marina Archaeological Project 1999
79
31. San Pasquale general view of location of test pits.
Bova Marina Archaeological Project 1999
80
32. San Pasquale Trench 1 test pit.
Bova Marina Archaeological Project 1999
81
33. San Pasquale Trench 2 test pit.
Bova Marina Archaeological Project 1999
82
Table 1. Trench 1 finds and taphonomic characteristics by depth.
Stratum Depth Ceramics Obsidian Flint Fauna Daub Ground
stone
% Fauna
burnt
sherd
diameter
(mean)
obsidian
length
(mean)
I 0-130 528 59 2 186 54 1 37 29.7 15.7
II 130-160 1172 110 6 329 269 10 29 33.9 19.9
III 160-190 710 106 7 338 361 8 19 32.6 17.2
IV 190-230 388 71 2 343 330 0 17 32.6 19.0
V* 230-250 99 28 0 135 80 1 20 26.1 17.7
Bova Marina Archaeological Project 1999
83
Table 2. Summary of faunal assemblage from Trench 1.
stratum depth unburned
fragments
burned
fragments
1 0-100 44 13
1 0-120 8 6
1 0-130 2 1
1 0-147 3 1
1 100-110 22 40
1 110-120 16 7
1 120-130 18
1 135-150 4 1
2 110-120 21 22
2 120-130 27 9
2 130-140 47 14
2 140-150 50 26
2 150-160 77 18
2 160-170 13 5
3 160-170 62 28
3 170-180 82 18
3 180-190 74 13
3 190-200 49 3
3 200-210 7 2
4 200-210 69 17
4 210-220 76 11
4 220-230 83 12
4 230-240 25 13
4 240-250 30 7
5 240-250 67 3
5 250-260 24 19
5 260-270 16 2
5 270-280 1 3
Bova Marina Archaeological Project 1999
84
Table 3. Summary of lithic assemblage from Trench 1.
stratum raw
mat.
blade blade
fragment
chunk core flake tool
1 f 1 1
1 o 4 5 5 1 42 2
2 f 1 4 1
2 o 3 12 12 4 72 7
3 f 1 1 4 1
3 o 3 20 12 1 63 7
4 f 1 1
4 o 3 10 7 1 48 2
5 o 5 2 20 1
Bova Marina Archaeological Project 1999
85
Table 4. Summary of daub frequencies from Trench 1.
stratum depth Total
fragments/level
Average
fragments/
square
1 0-100 8 2.66
1 0-120 5 5
1 0-147 3 3
1 100-110 13 6.5
1 110-120 9 9
1 120-130 7 7
1 135-150 9 9
2 110-120 9 2.25
2 120-130 9 2.25
2 130-140 36 7.2
2 140-150 58 5.8
2 150-160 128 11.64
2 160-170 29 29
3 160-170 107 11.89
3 170-180 91 10.11
3 180-190 46 5.11
3 190-200 76 12.67
3 200-210 41 41
4 200-210 129 16.13
4 210-220 116 16.57
4 220-230 40 10
4 230-240 39 9.75
4 240-250 6 3
5 240-250 46 15.33
5 250-260 34 17
Bova Marina Archaeological Project 1999
86
Table 5. Paleobotanical samples.
Sample N. 9 7 5 8 6
Feature/Context T1 7n-36e
(160-170)
str.II
T1 7n.36e
str.III
T1 7n.36e
(230-240)
Str IV
T1 7n.36e
(250-260)
str.V
T1 8n.37e
(230-240)
Sample vol. 20 20 20 20 2.5
Seed/Litre
flot/res f f f f f
Type of context stratum stratum stratum stratum feature fill
Period/phase Neol Neol Neol Neol Neol
common
name
Taxon body
part
pres.
cereals Cerealia gr carb 4 6 10 6 6
cereals Cerealia c.n. carb 1
barley Hordeum cf.
vulgare
gr carb 6 1 2 4 3
(wild?)
barley
Hordeum sp. gr carb 1 1
heinkorn/
emmer
Triticum
monococcum/
dicoccum
gl.b. carb 3 2 3 1 1
heinkorn/
emmer
Triticum
monococcum/
dicoccum
f carb 1 1 1
emmer Triticum
dicoccum
gl.b. carb 1 1 1 2
emmer Triticum
dicoccum
f carb 1
bread
wheat
Triticum aestivum
s.l.
gr carb 1
wheat Triticum sp. gr carb 1
wheat Triticum sp. r.in. carb 1 3 2
Atriplex sp. s carb 1
vetch Vicia/Lathyrus s carb 1
Trifolium/Medica
go/ Melilotus
s carb 1
grape Vitis vinifera s mo 1
grasses Gramineae (Poa
type)
s carb 1 1
grasses Gramineae s carb 1
modern
seeds
Charcoal carb 5fr 3fr 10fr 15fr 15fr
eggs/coprolites carb x 3
cyst 2 1 2
small bones x x x x
Bova Marina Archaeological Project 1999
87
Table 6. Radiocarbon dates.
Sample Material Trench Level ID Uncal.
Date
Calibrated
date*
Notes
Umbro-1 Charcoal 3 210-220
cm
Beta-
122935
1770 +/-
60 BP
AD 120 -
415 (AD
250)
Late
Roman
Umbro-2 Charcoal 3 220-230
cm
-- -- -- sample
not
analyzed
Umbro-3 Charcoal 1 (-8n/-
37e)
181 cm,
sublocus 1
(Stratum
3)
Beta-
122937
4330 +/-
50 BP
3045 -
2880 BC
(2910
BC)
Copper
Age
Umbro-4 Charcoal 1 (-8n/-
37e)
210-220
cm (Base
of Stratum
3)
Beta-
122938
5790 +/-
50 BP
4780 -
4515 BC
(4680
BC)
Middle
Neolithic
Umbro-5 Charcoal 1 (-8n/-
36e)
250-260
cm
(Stratum
5)
Beta-
122939
6750 +/-
50 BP
5685 -
5565 BC
(5600
BC)
Early
Neolithic
Umbro-6 Animal
bone
4 Stratum I Beta-
122940
modern modern AMS
date; no
valid
result
Umbro-7 Human
bone
3 180-190
cm
Beta-
125061
1600 +/-
40 BP
AD 395-
560 (AD
440)
AMS
date; Late
Roman
Umbro-13 Charcoal 1 (-7n/-
36e)
160-170
cm (Strata
II/III)
Beta-
135146
5930 +/-
70 BP
4945-
4615 BC
(4785
BC)
AMS date
Umbro-15 Charcoal 1 (-7n/-
36e)
180-190
cm (top of
Stratum
III)
Beta-
135147
6620 +/-
60 BP
5660-
5485 BC
(5610
BC)
AMS date
Umbro-19 Charcoal 6 (-60n/-
6e)
60-70 (III:
same level
as base of
pot scatter
Beta-
135150
3390 +/-
60 BP
1865-
1520 BC
(1680
BC)
AMS date
* calibrated date is 2 sigma, 95% probability range with intersection of radiocarbon age with calibration curve in
parentheses
**Umbro-8, 9, 10, 11, 12, 14, 16, 17, 18, and 20 are charcoal samples archived and not dated.