Journal of Mediterranean Archaeology 9. I (1996) 39-82

Obsidian Procurement and Distribution in the Central andWestern Mediterranean

Robert H. Tykot

Department of Anthropology, University of South Florida, Tampa, Florida 33620-8100, USA

Abstract

Obsidian has long been recognized as an indicator of long-distance, maritime-based exchangenetworks in the Neolithic central and western Mediterranean. Earlier studies have identified andchemically characterized the major island sources, but few subsequent efforts have been directed atdetermining the provenance of significant numbers of artefacts from secure archaeologicalcontexts. This paper presents new interpretations of obsidian procurement and distribution basedon the chemical and visual sourcing of more than 2700 artefacts from island and mainland sitesin France and Italy, and discusses the spatially and temporally dynamic economic and social roleof obsidian. Finally, it is suggested that long-distance prestige exchange of obsidian and othermaterials was an important way of maintaining ethnic or km connections in increasingl)'sedentary N eoUthie societies.

Introduction

Archaeologists commonly use the conceptsof 'trade' and 'exchange' to explain thepresence of non-local raw materials orartefacts on archaeological sites. With theintroduction of instrumental methods ofchemical analysis and their successfulapplication to archaeological provenancestudies of materials like obsidian (Cannand Renfre\v 1964), it became possible todefine an object not only as an 'import',but also to identify the specific source ofits raw material components. This estab­lished both beginning and end points in thelithic chaine operatoire, and opened thepossibility of examining the intermediatebehaviors in greater detail.

Interpreting the social context of variousmodes of exchange in prehistoric societies

has been the subject of considerableresearch for some decades (e.g. Clark1952), and the possibility of equating cer­tain modes of trade with particular levelsof social complexity (Service 1962) becamea central interest of processual archaeologyand its goal of achieving higher levels ofinference from static archaeological data(Renfrew 1972; Sahlins 1972; Wilmsen1972; Adams 1974). It seemed apparentthat exchange could only be fully under­stood within both social and economiccontexts relevant to the society in question,constructs which nevertheless remain diffi­cult to establish without historical or eth­nographic data (Polanyi 1957a, 1957bjDalton 1975, 1977; Akalu and Stjernquist1988).

Many studies in the 1970s and 1980smade use of generalizing, heuristic mathe-

40 Ty/wt

matical models to interpret artefact distri­bution patterns. Among the best knownare the gravity model (Bradley 1971;Chappell 1986), used to describe inter­action zones in which different sources'compete' for market share, and fall-offcurves (Hodder 1974, 1978; Hodder andOrton 1976; Renfrew 1975, 1977; Warren1981) of artefact frequency versus increas­ing distance from their source. The shapeof the fall-off curve is determined byparticular exchange nlechanisms, and theslope by factors such as demand, transport­ation costs, and the availability of alterna­tive materials.

There are a number of inherent problemsin these models. Ammerman (1979;Ammerman et al. 1978; Ammerman andAndrefsky 1982), in particular, has notedthat these models assume exchange was notsporadic, \vas not disrupted, and that theydo not take into consideration temporalvariations in behavior, population growth,changes in settlement size, and the hetero­geneity of 'dropping rates' (Ammermanand Feldman 1974, 1978). There are alsovariables such as differential participationin exchange networks, seasonal activities,and the likelihood that several exchangemechanisms were in effect at the sametime, that need to be taken into considera­tion. Objects may have moved alone (bytrade or gift exchange); along with indi­viduals (traders, craftspeople, brides); orwith groups of people (migration, coloni­zation, war, foraging). Local variation inresources, transportation, population den­sity, and social organization would haveproduced regionally distinct situationswhich may be amplified by differences inarchaeological fieldwork methods, sam­pling strategy and sample size (Clark 1979;Hodder 1980; Knapp 1985). Lastly, theproblem of equifinality, that differentexchange mechanisms may have resulted in

the same distribution of artefacts, has alsobeen recognized. Despite these considera­tions, few have made an explicit interpre­tive connection between Polanyi's oft-citedmodes of exchange (reciprocity, redistri­bution, and market exchange), the multiplemechanisms characteristic of each, andtheir manifestations in the archaeologicalrecord (Sheridan 1982).

These problems, along with the unevenquality of most archaeological data, ledBietti Sestieri (1985: 115) to suggest that thequantification of existing data was a higherpriority than the elaboration of theoreticalmodels:

...one should do no less than askwhether, in reality, the more fruitfulapplication of mathematical methods toarchaeology is in the elaboration of dataon a quantitative and statistical basis,independently of the degree of formali­zation of the model.

This criticism is particularly rdevant toobsidian in the central and westernMediterranean, where the sources them­selves were until recently not well charac~

terized, and where few provenance studiesexamined significant numbers of artefacts- many from insecure or inappropriatearchaeological COntexts which precludedthe application of mathematical models.Perhaps because of limited capabilities inproducing quantifiable exchange data,many recent Mediterranean lithic studieshave emphasized procurement and pro­duction (e.g. Ammerman and Andrefsky1982; Torrence 1981a, 1981b, 1982, 1983,1984, 1986), and USe (e.g. Hurcombe 1992a,1992b; Vaughan 1985, 1990), rather thandistribution of artefacts (e.g. Pollmann1993).

The quantification and interpretation ofobsidian provenance data is the focus ofthe work that follows (see Tykot 1995a).

~......

~;::l~

tJ'"...&­~5;::l

;;';;..'"'

o0­V>

§:.~;::l

\:l...cr;

~3~

n'"'~....2­~;::l~

~1::

~$:'"'~r:.:

~l:l;::l\

.. ,1 :;:'-~i -~-. J

\/ ~/..-'<! (1 II' \ """-, /

j (~\! ....... /

_~ symbols) and archaeological sites (small circles) with obsidian in the central and westernlediterranean, (Nearly all of the mainland sites are Neolithic.)

·Y;YJ '''\F~ I ",~ ~);, '" ""c<-=:; \~\.j/~'/-- ,~~ 'J .. Y-O:tJ

~~ \ I ( }! (~r""!/ }--., '" \. " 1<\- ", / ' (-......., '- rJ \'~ /"'--., ~ , I' ' v" ' " ,,' ,,', ~ ; \~'! \ .< j'''./ '. \-<0;; I" (~l' ",~-.\

hI' fi "'~/ \/ "~\JJ :-,~ •

M), ; "I ~1;(j.J{:1' )((If - i- r \}+-

\

.//./: -::. ,,", ;f- )~~\, ~1' \ 1foJ

Jr'1

I

, 'lP', <" v' " ,,,/ '~ ,,"" " , "" ~ ,'."' ' ' '.:>/i, ,}i--;~'['> 'h~ j;il:-._;/)), .l;;)r~\~~;'.":?· -\~-.,.r;r=-< C·/\,F,' t.-).;;, ((, "" ~ '- \(: I"'_~ '1 '- ~ 1 [~1'. .1' f' ,r--~ ...--~.J"'" ~/v" ' ,J, ' - ' ~

y'-/;J ).. J I /( ,'P • ~.. ~ \ .'1.· .! "~...;..or-~ (r ')T:",- ~~.S'~.r.~'-~l.'; '/?~-'tf:·~ • ..,,' ~ '----, / ('7. ;l '<;J;,..1jl/ ' ~02, ( ~ ~_ /~ < .~ ,~F

J~~;r'.,+\'2:::;.{ f'JIF,/ ,,~i :i'~" -:""£;;'" \ '( 7Ti-::'t'f~C::;(/0,' ~ ,>/ .J '<1:3 ., •!.rl\'" ~', \'\!I)' _'I0~"(,"~ ~-~d,J'\! ,J I ,.'\. ,,,4 .~. q, ,~J' I. 'j--'"

IL/'-~-/\ (( JI( / .. ~-..~ ,. ;~..... _' ~-)~~_\ ) ,,'_ ~ I), \ \y'/ !t1"-'- ",-/ r\! 'J/' r \.. I l., [ /~ II.:! , \'(/ ;'jl~~~h//"-\..<\ r'." "\ •• \.., ...... I"

_ /" ('v-

l J~(I. ;;:." t... \ \~ ',--k~ '? . Monte Arcil ; Palmarola- ~ "'0.' j~, \~~

l 7:1 ! ·1 ~u .. ,'I,---7--- .. H ~,-"..... -~v ,J L . \ . . 'II!'\/ ' lpari . ~ '0 -

! \ . ." \t' '--,

/K'~ '-~,,__.\. /:;~-~~~- ' ,~. '}".l'~ *"_____\ r'~)' < /7~ __~~-" \~,\ ~..?----"'",. __ .-./.'t,' -'"'q'~' / ~~1 ~ ~J"_ It' ,,;,-yJor'. ,; t::;

,.", j : .. "",C.i/ ,c' " P \'0,--/':'-"~ / f ",,<i?"J ~ .'- .. \.', ' ' -- ' antelleria'.______._ /' "r:" · ~~", ", --'J I", -'L/---'&\ j> '~. ] • .ce

../ I I -- : - '/

"---\,"\ ([~1c} ( J P 'r~~:j~",.:(/, ,,) -- ? / \ ",.;: '--

,./ __ .2/-- , v,,!' -" -- '" ,-.,~j

,/ 1'-' ' .,, " '. .\ I -, .' -' .~--,

______ \, I) . I "",m I I·-/ I \, .I' "lb'd sl(oson,ou ' \ \" r\ J~l (

SI ian source' (I n:o Islands not shown >J - ~_.r-- "\., ,m- .

42 Tykot

The application of new analyticalmethods permits the construction of statis­tically significant geographic and chrono­logical patterns of obsidian source exploi­tation and new interpretive schemes re­establish exchange as critically importantfor understanding all stages of the chaineoperatoire. As the most visible indicator ofNeolithic interactions, obsidian use is alsorelevant to discussions of the earliestsettlement of the Mediterranean islands,the transition from hunting and gatheringto an agricultural way of life, long-distanceexchange networks, craft specialization,and the development of social differen­tiation and other precursors of morecomplex Copper and Bronze Age societies.

Mediterranean Obsidian Sources

In the Mediterranean region, the existenceof a limited number of obsidian sourcessuitable for the production of stone toolshas been known for quite a long time.Virtually all obsidian artefacts found atarchaeological sites in the central and west­ern Mediterranean come from four Italianisland sources: Lipari, Palmarola, PanteII­eria, and Sardinia (Figure 1). The Carpath­ian sources in southeast Slovakia andnortheast Hungary (Williams-Thorpe,Warren and Nandris 1984, 1987; Biro et al.1986; Bigazzi et al. 1990) are responsible fora few artefacts found in northern Italy,presumably others in Dalmatia (Martinelli1990), and some as far east as GreekMacedonia (Kilikoglou et al. 1996). Obsid­ian from MeIos has been confirmed at buta single site west of the Balkan peninsula(Bigazzi et al. 1986) and a few pieces ofAnatolian obsidian have been identified inEastern Europe and Greece (Renfrew andAspinall 1990). The converse also appearstrue: as yet, not a single piece of centralMediterranean obsidian has been docu-

mented east of the southern tip of theItalian peninsula, despite the claims forSardinian obsidian in Bosnia made a num­ber of years ago in a conference paper(Rasson et al. 1977; cf. Williams-Thorpe1995: 231).

Lipari

Lipari, located some 30 km north of thenortheastern part of Sicily, is the largest ofthe Aeolian Islands at ca. 38 km2.Transparent, black obsidian of excellentquality may be found today in severallocalities on Lipari, but the Forgia Vecchiaand Rocche Rosse (including PuntaCastagna) flows, and the Fossa cone on thenearby island of Vulcano, are from historiceruptions and were thus unavailable forprehistoric exploitation (Buchner 1949;Pichler 1967, 1980; Keller 1967, 1970;Bigazzi and Bonadonna 1973; Cortese et al.1986; Gillot and Cornette 1986; Lefevreand Gillot 1994). While older obsidianflows exist at Acquacalda, ValloneGabellotto (including Papesca beach andthe Pomiciazzo-Lami flow), and Montedella Guardia (including Praia di Vinci),fission-track dates on 66 artefacts of Lipariobsidian from archaeological sites in Italy(Arias-Radi et al. 1972; Bigazzi and Radi1981; Bigazzi et al. 1982; Arias et al. 1984,1986) indicate that Gabellotto was the pri­mary source used in antiquity. Chemicalanalyses of samples from both recent andprehistoric flows conclusively demonstratethat there are negligible differences in theirelemental composition (Belluomini andTaddeucci 1971; Francaviglia 1986; Tykot1995a, 1995b).

Palmarola

Palmarola is the westernmost of thePontine Islands, located west of Naples inthe Gulf of Gaeta, about 3S km from the

Obsidian Procurement and Distribution in the Central and Western Mediterranean 43

mainland. Obsidian flows are found to thesouth of Monte Tramontana in a domalcrust that transects the island, and alongthe east coast of Palmarola down to itssoutheastern tip at Punta Vardella where itis found in fist-sized, black, opaque blocks(Buchner 1949; Barberi et al. 1967; Herold1986). Although analyses have not turnedup any chemical differences in obsidianfrom the different source localities onPalmarola (Herold 1986), making it impos­sible to determine whether Punta Vardellawas the only source utilized, the small sizeof the island (less than 3 km l ) renders thisdistinction of minimal archaeological sig­nificance.

Pantelleria

Pantelleria, a small pear-shaped island ofabout 8 by 13 km, lies in the Strait ofSicily, about 90 km east of Cap Bon,Tunisia. Pantelleria is the type locality forperalkaline rocks, in particular its Na- andFe-rich greenish obsidian known as Pantell­erite (Foerstner 1881; MacDonald andBailey 1973). Pantellerian obsidian is thusreadily differentiable from other westernMediterranean sources on a visual basis(e.g. Cann and Renfrew 1964). Francaviglia(1988) has isolated five chemical sourcegroups on Pantelleria: three vertically-dif­ferentiated sources exposed at Balata deiTurchi; Gelkhamar; and lago di Venere.Chemical analyses of artefacts fromPantelleria, Malta, Sicily and the mainlanddemonstrate that the upper (i.e. morerecent) Balata dei Turchi flows were mostcommonly used, but that pitchy (and morerecent still) Gelkhamar obsidian was alsoemployed - even in Sicily (Francaviglia1988; Francaviglia and Piperno 1987;Tykot 1995a, 1995b). The large variation infission-track dates on archaeologicalartefacts (Arias-Radi et al. 1972; Arias et al.

1984) also suggests the use of multiplesource flows, but these dates cannot becorrelated as yet with Francaviglia's chemi­cal source groups. like Palmarola, PantelI­eria was not settled during the Neolithic,and neither differential access to thesources nor organized extraction of obsid­ian is likely to have existed.

Sardinia

Sardinia, unlike all other Mediterraneanislands with obsidian sources, is a largelandmass with an area of 24,000 km2 and ahistory of occupation dating back to theUpper Paleolith ic (for a detailed review, seeTykot 1995a, 1997b). Obsidian beds in theMonte Arci volcanic complex were firstdescribed by della Marmora (1839-40) andlater by Washington (1913); a comprehen­sive survey of the Monte Arci zone wasundertaken by Puxeddu (1958) as part ofhis thesis in archaeology at the Universityof Cagliari. This seminal contributionincludes a detailed description of obsidiansources and archaeological sites in theMonte Arci region. In a zone of about 200km2 which today includes 19 towns orhamlets, Puxeddu found 246 locations withobsidian, including four which he classifiedas sources. The later realization that at leastthree chemical groups (SA, SB, SC) wererepresented among analysed archaeologicalmaterial raised questions about whichsources were being utilized, since only onegeological source (Conca Cannas) had beenanalysed (Cann and Renfrew 1964; Hallamet al. 1976), and both translucent andopaque obsidian had long been recognizedin archaeological assemblages. Followingdetailed geological surveys of the entireMonte Arci complex (Beccaluva, Deriu et

al. 1974; Beccaluva, Maccioni et al. 1974;Assorgia et al. 1976), several attempts weremade to chemically characterize the multi-

44 Tykot

?km

5~ Basalt

6~ Submarine Basic Lavas

7kME,j Alluvial Sediments

41]1 Trachyte & Trachyrhyolite

3~ Andesite & Dacite

1 Em Perlite & Obsidian

2 [}] Rhyolite & Rhyodacite

··i'.':·'~'-~:~~ ;}-t;~i!&~i~\L~~i~~i.-' -', ," "·:'o~~·::-'~'·':~"":;"·:~'~'",;~"r~):;i:;':;;

lN~~r~II'l!ii..~~"~flI1~rw~~~t.~\~l~'~"'i;'<;~:"''-:«1(,""i''''''''''""",~ =._""", """~'i""" ""'""""'

f{!~Xtll[li~$:"~ 1d•·••:.1••••••~.~111·~-';~:~~~~~;;~:~[.1~~'l~I!!~t;i!~~: ••. >:~~~ _••~;.'N,""';'"'''''''''''''''''' ·.·.' ..d.C,'"=.... 'v.~.. _ ~

~11~lf~::"':,E-!T;"~~"~~~l_t\\i(f41t1:t(t"''''"~~ ' "'"-i;?ii ,." '. . " .', '.' M"","'?J~'." .. ""--".,,;.»'" , ";l!Hi:!:!_",~..•••••••:••: i, •••• :, •.• ~11}{}~~(~~f~!t'~::·;.'·':"'~i;';:"'~;~:' ~/~ . '.' .'. ;';;'.' ... '~d~"a.:. ' ...;'11 '_"<:;";\""1::;,.'"'''':'~?:i,."'':''' '", .. :z '. '..... , '»-..-7~h~.l/h .. _ ";~~"__':"__"-"'~~_'~ "':.~<>\~;:H:::··,::: .> ':. ·>-;"·i~~·.n;"': ..~. /=.:=:::.:'~~e)J';I(~~~::.lI\1;"'",,,,,,.... ..... ,..". "''''. .'. ." %%a Sc,~""""""""""" ,~l:;''":'''",,;,;;%~:;~~?"m~~~~¥t~t~'·('~qr"aMl'~_·""W ." ','. '.' 0.,.~~ ~_•• '.'""",.••.. "",'. ""', ~"'. 0,. '~/.= ' _ '~"''''..,'".,.",. .' ",,'," ,'. . M."•.,~ ")'.. "'_. _ _~ c. >, c",_,,_

~f ;~i£~'e; ",,· ;"'%\;:C~-:-<~tr;:i!ff~i.(.1~r;:?'~"·::::::;SB·2;:;::;::.:::;:::.;;~ .::. '~jrl'::d\,~ J~~IJI~~;};~?<f;'~':;_;:~;:_;~:;~;:)t~:~~~~t~iJ;qi~~):~:~~:Uim~11111 !111~il!rll~'~1~11 ~I~lll!jl~jll';~~ii~~j~{~'{hl~W)1t:}""""'.""~~" '" 'I mC011~ ~"":-""""""""".',i·.,~.:.; •..~.. @Y""'-·'-'I'\IIII;.,\1 l .~, I'~ _ "-:';"'I:-~:'-';'("_~<'..-'"·····.. ·,?!rljrullc . // . . \ '=" III ·c.-.,." .•.. _' .....,,,'..... ".-"'". "" '. '''" ~" ...'.. ''',."'".,,.,X~;;:~~:;~.(.:~•..•~_~_cr~._r/1i<"::'::.: ::." ')DI"~ -.;;Pi!., ~..-';\;';,'ii.;;;;;~ii'i,!.~:(:~Ji~4'_••..":'" '" .,'""""'~" ';@,jj'!t~~llf.'I'~Aii?),\,,~~~::;~\i;~;,/, ~ ~: :-5 :}:-~.. :;:::.. ~t'.~:~f:~~~~:~>:;;~~:{:~~,~»:~(::';<-,': '/"/0 ~>. = :.:.; ~.' ~% ";--':"':~;;";:i""_'::,:,.",~.:;!;'~;:~~t:~ ~O~~/I ~~~A':::::~~@~":'I~1J~1. ,~~:~;,y,~:~o~:';i~;':~~;~'Y{:~;--;",."",,~ f.(.,w",~, "':""'>:'. //- "":"..',"","",i; I',,' ,,"-:'s"" '"'''''<''0'/' -'",''; ',' . '''''-''i~,,-,.,,,,,,..,, ,'.,"" '"..'...'''<'',h:"""'" '.'."""'''. ,:: . . '''.' , " ",.",."'''''''':''", '.,:~'~;~:~:i~\->~[~';:;~l' ·15·"~~~i~~. ~_~ ~~'''''~.'''\':~'''''';'~';:';"c{iii~'£'i~~;;m\\i ~~;:\. .m ~ \"''\.''i1{\i;!i~;;fii''::i'';;;!""'(~:;"t(;",;·;:,,,<,. ' :::;~rwa":t~ 'f ~1;,!:';f;,.!;"":,:"",,,,,;',"j:ii~}:~~~~~J~~~t~k;r:l~'))};':''%1A:~~~lltl?oI@lililljl~ll~!f%~~~~~?t~:s;~;~\;":i};~~7; =M,o" Sol'"",""rO:;i;~:~~~ir;2'~;:~j:_P)~s 'il'~' '~'iI1l§!!y'<"C'.''''''; ~:'""":1;",,,, 8t0IJfffi!!f1il%:iW#1;] W%r;; ,,' t.%~~if':i!~[i'b:,';i;;f!fii!R;f1."',:,"',.,"',.,'''••• I'~,..,;..» ..,"-',;""',:.."."., ..: 1

i~i~.tli'f~I,r,t~i~~."'.:"''''.'''' ..'''-.-:".,".,~,,, ..;,:,.,''.:,..,' ':"":""":..i[*~~rJ§:\~;_'!ii;~~iti~~~'i"'i! ~t'''.,"m , '.' ...

Obsidian Procurement and Distribution in the Central and Western Mediterranean 45

pIe obsidian outcrops. Although all weresuccessful in this regard, the only availableinformation from the first study comesfrom a very brief conference paper (Mac­key and Warren 1983); an independenteffort by Francaviglia (1986), also not fullypublished, provides no details about theobsidian deposits themselves; and theunpublished dissertation by Herold (1986)makes no attempt to match chemically­defined geological source groups with arch­aeological materials. More recently, myown survey of the Monte Arci zone (Figure2) located the SC source in situ for the firsttime, and geological material from the fivesources (SA, SB1, SB2, SCI, SC2)represented among archaeological artefacts

has been fully described and chemicallycharacterized (Tykot 1992, 1995a, 1997a).

Most importantly, it was demonstratedthat all archaeologically-significant dis­tinctions among Mediterranean obsidiansources (Lipari, Palmarola, Pantelleria, SA,SBI, SB2, SC, Melos, Giali) can be madeon the basis of major/minor element com­position, permitting quantitative yet inex­pensive and minimally-destructive analysisusing the electron microprobe (with wave­length dispersive X-ray spectrometers) ofhundreds of archaeological artefacts fromNeolithic sites in the central Mediterranean(Table 1). Furthermore, it was discoveredthat the frequency of each obsidian sourcerepresented in a lithic assemblage could be

Table 1

SA SBI SB2 SC

SiOr 74.72 ± 0.25 73.87 ± 0.35 75.08 ± 0.27 72.72 ± 0.35(74.02-75.63) (73.22-74.59) (74.33-75.72) (71.74.73.77)

AlrO l 13.40±0.15 13.63 ± 0.10 12.97 ± 0.14 13.92 ± 0.19(12.86-13.83) (13.46-13.84) (12.60-13.30) (13.18-14.83)

TiOr 0.09 ± 0.01 0.17 ± 0.03 0.13 ± 0.02 0.27 ± 0.03(0.05.0. I 2) (0.13-0.25) (0.10.0.20) (0.17-0.37)

FerOl 1.25 ± 0.09 1.33 ± 0.20 1.1 5 ± 0.09 1.52 ± 0.19(1.01-1.59) (0.81-1.68) (0.92-1.44) (1.01-2.06)

MgO 0.08 ± 0.01 0.12 ± 0.04 0.11 iO.02 0.20 ± 0.06(0.05-0.13) (0.06-0.21) (0.06-0.15) 0.06-0.38

CaO 0.59 ± 0.03 0.74 i 0.03 0.57 ± 0.02 0.87 ± 0.08(0.52-0.71 ) (0.69-0.83) (0.53-0.67) (0.67-1.19)

Na20 3.45 i 0.11 3.38 ± 0.10 3.36 i 0.06 3.31 ± 0.12(3.05-3.82) (3.22-3.58) (3.11-3.48) (2.81-3.84)

KlO 5.24±0.14 5.55 ± 0.20 5.46 ± 0.11 5.89 i 0.18(4.93-5.92) (5.12-6.10) (5.23-5.86) (4.84-6.43)

Pp5 0.08 ± 0.01 0.10 ± 0.02 0.08 i 0.01 0.14iO.01(0.02-0.10) (0.07-0.13) (0.06-0.12) (0.09-0.18)

MnO 0.06 ± 0.01 0.04 i 0.01 0.04 i 0.01 0.03 ± 0.01(0.04-0.08) (0.03.0.06) (O.oz-O.Q7) (0.01-0.07)

BaO 0.02 ± 0.02 0.05 i 0.02 0.02 i 0.02 0.11 ± 0.02(0.00-0.08) (0.01.0.10) (0.00-0.07) (0.03-0.18)

n 207 26 130 341

Compositional analyses of obsidian from Monte Arci, Sardinia. (Means ± one standarddeviation are given for n geological and archaeological samples analysed, with the rangein parentheses below. All data are given as percentages, with totals normalized to99.00%.)

46 Ty/wt

reasonably estimated by non-destructivevisual examination of the whole assem­blage; the reduced accuracy of the visuallydetermined source frequencies may be lesssignificant than the sampling error com­monly associated with chemical analysis ofselected numbers of artefacts (Tykot1995a).

Distribution Patterns

Obsidian artefacts have been identified atover 1000 archaeological sites in the centraland western Mediterranean, including vir­tually every site on Sardinia and Corsica,and many if not most Neolithic sites inSicily, the Italian peninsula, and Mediter­ranean France (Figure 1). Few sites withobsidian finds are known to the north ofthe Po River, on the eastern side of theAdriatic, or in North Africa. Obsidiantools are particularly abundant on or nearthe source islands, accounting for up to100% of lithic assemblages, while theplethora of distant sites with obsidianmasks the actual paucity of artefacts in thatmaterial found at most sites. In stark con­trast to the eastern Mediterranean (seePerles 1987, 1990), obsidian use is strictlyassociated with pottery-using agropastoral­ists, beginning in the Early Neolithicperiod. In the following discussion, I com­bine the results of my own research withprevious analyses in order to describe theexploitation and distribution of Sardinianobsidian in particular, as well as the moregeneral chronological and geographicpatterns apparent in central and westernMediterranean obsidian use (Figure 3).

Sample Selection

The imprecise nature of the archaeologicalrecord presents many problems for theinterpretation of obsidian distribution

patterns. Excavations, like publications, areof uneven quality, and those sites exca­vated may be unrepresentative of settle­ment patterns in their location, chronologyor site type (e.g. emphasis on cave sites vs.open-air settlements). For the Neolithicperiod, we are currently limited to placingsite-contexts within a timespan of a fewcenturies at best (Tykot 1994), so thatdescribing 'contemporary' obsidian use atmultiple sites is likely to conflate manygenerations of human activity andhomogenize short-term patterns that mayactually be heterogeneous (either inten­tionally or randomly). Lithic assemblagesfrom older collections underrepresentsmall pieces of debitage, while increasingthe probability that each artefact collectedrepresents a different reduction event.Furthermore, obsidian artefacts in museumcollections may not be a random selectionof the population of tools used at a site.Finally, some sites have small numbers ofartefacts analysed making it impossible toconclude much about their distributionpattern with any statistical confidence,although some attempt is made here topool the results for multiple sites, and insome cases by combining preliminarily thedistributional evidence from Neolithic,Copper and Bronze Age sites before spe­cificall y examining each age or cultureperiod in detail.

Sardinia

Within Sardinia, it appears that a distinc­tion can be made between an Oristano areasupply zone, where lithic assemblages arecomprised entirely of obsidian which mostlikely was acquired directly from itssource, and the rest of the island, whereobsidian probably was obtained indirectlythrough exchange (for terminology, seeRenfrew 1969, 1977) and may even be less

Obsidian Procurement and Distribution in the Central and Western Mediterranean 47

Isolino di Varese•

Pa~so di Corvo·

·Po~gioOhvastro

• Pianaccia diuvero

Gaione •

• San PoloAreneCandide

•Combes

Les Terres Blanches•

~~Ca~erine J1Tlbaldi

~cabre

• J ~o;:::tl; t~col~Tigglanese

l I CalanchiBasI

Liscla Pilastx:,Cala Villamarina

M d'A dd' I' Li Murionte cco 1 _

Grotta SaCorona Grona FiliestruSa Ucca de su Tmtlmolu \

Palas de Casteddu MoliaMes'e Arrius Simaxis

Cuccuru s'Arriu PuisterisCantoniera Frumini San Gemiliano

Serra de Castius S PietroBuon Cammino S~ Coddu

Barbusl \ T ..su Carroppu erramamlTracasi Grotta San BartolomeoSanta Gilla Grolta S. Elia

c>~Skorba

I 250km

Figure 3 1\'eolithic sites with 10 or more obsidian analyses.

48 Tykot

frequent than flint, quartzite or other lithicmaterials. Surface collections of manythousands of obsidian tools from sites inthe Oristano-Campidano area and incIud.ing the Monte Arci zone itself illustrate theready availability and exploitation of thismaterial in prehistoric Sardinia (Contu1990-91', Atzeni 1992). Puxeddu (1958) infact identified 11 collection centers and 74reduction sites on Monte Arci, based onthe lithic forms (cores, trim, flakes, blades)found at each. Unfortunately, there is noassociated material to date these procure,ment and production activities, althoughthe quantity and type of artefacts thatPuxeddu collected suggests that they areprimarily Neolithic.

Obsidian continued to be the mostimportant lithic raw material used inSardinia during the Copper, Bronze andIron Ages (Contu 1990-91), but its ubiq,uitous presence from earlier use precludesknowing whether or not fresh obsidian wasobtained from geological sources or wasrecycled from earlier site occupations.Hydration dating of obsidian from Sar­dinian sites has frequently identified reusedor redeposited artefacts (Le. those with agesmuch older than their context) (e.g.Michels et al. 1984; Stevenson 1989; Dysonet al. 1990; Stevenson and Ellis n.d.).

Chemical and visual analyses of obsidianfrom 61 sites in Sardinia indicate that allfour Monte Arci sources were utilized atNeolithic sites in the Oristano region, butthat types SA and SC obsidian were mostcommonly exploited (Tykot 1995a). Theavailability of considerable quantities oftype SC obsidian on the west-central coastof Sardinia suggests that the 'value' (flakingquality plus less tangible factors) of typeSC obsidian at least equaled that of typeSA, since the procurement cost (time andeffort) for type SC was probably somewhatgreater than for type SA. Type SA obsidian

has always been considered to be of finerquality than type SC, but the roughly equaluse of type SC in the Oristano region sug,gests that at least the local craftspeople didnot make a big distinction on that basis.

In Sardinia, preliminary examination ofobsidian relative use frequencies by prov­ince of the multiple Monte Arci sourcesreveals significant geographic differences(Figure 4). In southern Sardinia (Cagliariprovince), SB obsidian is hardly used at all« 5%), whereas in northern Sardini.a(Sassari province) it accounts for 33% ofthe obsidian tested; type SA is twice as fre·quent in southern Sardinia as in the north(36% vs 19%). Since the frequency of typeSC is fairly constant, this distribution canbe explained best by the relative proximityof the SB sources to northern sites, and ofthe SA source to southern sites.Surprisingly, type SB is less common at theOristano province sites, for which the SBsources are the closest.

Looking at individual sites in Sardiniawith 10 or more chemically analysedartefacts, greater heterogeneity in obsidianuse patterns appears than in the provincialcomparison above. In southern Sardinia,type SB is noticeable only at Tracasi;although much more common overallamong tested samples from northernSardinia, it is in fact not represented at allat three of the six sites there with 10 ormore analyses. The chronology of the sitesappears to be at least part of the explana~

ti.on. Malia (Tanda 1980, 1984) and LisdaPilastru (G. Pitzalis, unpublished) are LateNeolithic (Ozieri) sites, and Monted'Accoddi (Tine and Traverso 1992) isFinal NeolithiclChalcolithic, whereas mostof the artefacts from Grotta Filiestru(Trump 1983) are Early and MiddleNeolithic, as are those from Monte Maiore(Foschi Nieddu 1982, 1987) - both siteswhere type SB represents more than 45%

ObsIdian Procurement and Distribution in the Central and Western Mediterranean 49

%

100

80

60

149 50 202

.SCDSBCJSA

20

aCagliari OristanolNuoro Sassari

Province

% 584 322 947100

.SCDSB2

80 ~SNSB2GSA

60

40

20

0Cagliari OristanolNuoro Sassari

Province

Figure 4 Obsidian source frequency in Sardinia, by province. Top: combined results fromchemical analyses (Cann and Renfrew 1964; Hallam et al. t 976; Mackey and Warren1983; Francaviglia 1986; Michels et al. 1984; Tykot 199Sa); more than two--thirds of theartefacts represented are Neolithic. Bottom: results of visual assessment (Tykot 1995a);more than 90% of the artefacts represented come from Neolithic sites. Numbers ofartefacts represented are at the top of each bar.

50 Tylwt

of the obsidian assemblage. Sa 'Ucca de suTintirriolu (Loria and Trump 1978), withnearly 20% type SB obsidian, is LateNeolithic, but is right next door to Filiestruin the Bnou Ighinu Valley, so that the reuseof locally available type SB cannot be ruledout. Visual determinations also revealed notype SB obsidian at Molia among 63artefacts examined (in addition to the 20chemically analysed), and only 14 (our of189 = 7%) type SB artefacts at Li MLlti(also Late Neolithic; Puglisi 1941-42).11%of 114 Middle Neolithic obsidian artefactsfrom Cala Villamarina (Lilliu 1959) Werevisually identified as type SB.

At Grotta Filiestru, the only site inSardinia with all four Neolithic periods instratigraphic succession, 86 randomlyselected obsidian artefacts were chemicallyanalysed, and an additional 581 werevisually provenanced (Figure 5). Together,nearly all of the obsidian found at the site~ constituting a steady 20~30% of thelithic assemblage (Trump 1983) ~ wasexamined. Both sets of data indicate thatthe use of type 5B obsidian steadilydecreased over time, to be replaced pri­marily by type SC obsidian. This situationis paralleled at Monte Maiore~Thiesi,where 58% of 26 Early Neolithic (Filiestruphase) and only 36% of 11 MiddleNeolithic samples analysed are of type 5B2obsidian.

Corsica and the Tuscan Archipelago

The problem of sample size is best illus­trated by the earlier work of Hallam et al.(1976), which indicated that type 5B obsid­ian was particularly well represented inCorsica, especially at Curacchiaghju-Leviewhere 8 of the 9 analysed artefacts werefrom that source. In contrast, chemicalanalyses now of several hundred artefactssuggests that type SB accounts for less than

20% of the obsidian in Corsica, and evenless in the southern part of the island(Figure 6). I wrote several years ago (Tykot1992: 65) that the SB obsidian sources werereadily accessible to sea-borne travelers tothe Gulf of Oristano and that this couldaccount for such high percentages inCorsica when its use in Sardinia was mini~

mal. In fact, SB turned out to be muchmore significant in northern Sardinia thanpreviously thought (33% instead of 6%)and the question is now reversed: why doestype 5B constitute only 12% of obsidianassemblages from southern Corsica? Inaddition, why is type 5B twice as commonin northern Corsica than in the south?

If we look at the data for individual sites,chronology again appears to provide someof the answer. Curacchiaghju (Lanfranchi1987), Pietracorbara (Magdeleine 1991) and5trette (Magdeleine and Ottaviani 1986) areEarly Neolithic or have Early Neolithiccomponents (Lanfranchi 1993), making thelow frequency of SB obsidian at Basi (levels7-6) (BaiIloud 1969a, 1969b) the exceptionamong Early Neolithic sites. Furthermore,type SB obsidian appears more frequentlyin Early Neolithic (65%) than in laterNeolithic (26%) levels at both Pietra­corbara and Strette. Among the other sites,only the Chalcolithic dolmen of Cardiccia(P. Nebbia and ].-c. Ottaviani, unpub­lished) has more than a single piece of typeSB obsidian. As noted above, tombcontexts are more likely to reflect single (orat least short-term) production events, andthe size of the artefacts certainly does notexclude the possibility that all of the SBobsidian may have come from a singlecore. As in northern Sardinia, then, typeSB obsidian is of major significance in theEarly Neolithic but declines in frequencyin the Middle and Late Neolithic. Thequantity of obsidian imported to northern

Obsidian Procurement and Distribution in the Central and Western Mediterranean 51

%100

80

60

40

27 20 29 14.SCDSB2CSBIDSA

60

40

Cardial Filiestru Bonu Ighinl

Culture PhaseOzieri

Figure 5 Obsidian source frequency for four Neolithic cultural phases at Grona Filiestru.Cardial and Filiestru are Early Neolithic; Bonu Ighinu is Middle Neolithic; and Ozieri isLate Neolithic. Top: samples analysed by electron microprobe (Tykot 1995a), plus 4Ozieri artefacts analysed by NAA (Mackey and Warren 1983). Bottom: samples attrib­uted by visual assessment. Numbers of artefacts represented are at the top of each bar.

52 Tykot

Province

oUlJ~Ulill.illillJill~Corse-du-Sud Haute-Corse

.SCOSBFl.•...·::. SALJ

Corsica also appeurs to increase over time:it accounts for just 6% of the lithic assem­blage at Early Neolithic Strette (Magdeleineand Ottaviani 1986), while at nearby LateNeolithiciChalcolithic Monte Grosso, 85 %of the stone tools are in obsidian(Magdeleine 1973).

Two hundred and fourteen obsidianartefacts from Basi have been analysed, 197of them from specific stratigraphiccontexts (Figure 7), the most analyses fromany site in the western Mediterranean.Although a chi-square test of sourcefrequency for each of nine stratigraphiclevels indicates some relationship betweenthe two variables, no chronological patternis evident despite the millennium-longhiatus in occupation between the EarlyNeolithic (Cardial) and Late Neolithic(Basien) levels. Furthermore, even thoughsome of the excavated contexts actuallycontained mixed material (Lewthwaite1983: 153; Camps 1988: 78), it is stillsurprising that type SB2 obsidian is so rare.Both SA and SC obsidian are wellrepresented in every level (each a minimumof 25%), while both SBI and SB2 are neververy significant (maximum of 10% total).The two SB obsidian types are neverthelesspresent in five of the eight levels with morethan two analyses. In the early Cardialcontexts at Basi only a few percent of thestone tools were of obsidian, most of thelithics being in imported (Sardinian) flint;in the Busien period, however, obsidianaccounted for 56-78% of the lithicassemblage, the rest made of quartz andother local rocks (Bailloud 1969a, 1969b).Similarly, at Longone-Bonifacio onlyimported flint tools are present in the firstCardial phase of the Early Neolithic(Lanfranchi 1993).

It has been taken for granted thatSardinian obsidian made its way to themainland via Corsica and the Tuscan

83345

Figure 6 Obsidian source frequency inCorsica, by province. All but oneof the analysed sites are Neolithic.All attributions are based onchemical analyses (Hallam er al.1976; Crisci et al. 1994; Tykot1995a). Numbers of artefactsrepresented are at the top of eachbar.

20

40

80

60

0/0

100

6.1 5e 5d

Stratigraphic Level

30

6.3

29

6.2

11 29 26 30

5c

20

5b

20

5a

.SCDSB2~ SBtr::l....', ','.... SALJj

o0­V>

is.:.is,::s

~Q

":;:....~

~

'"~ls;JV>':;&­~0",::s

;;-;:r"'"Q,::s

[.,,::s;:::...

~~'"....,::s

$:'";:::...~.

~.,,::s

,sidian source frequency at Basi. Levels 7 and 6 are Early Neolithic; Level 5 is later Neolithic (Basien). All analysesIresented by microprobe (Tykot 1995a). Numbers of artefacts represented are at the top of each bar. Vi

VJ

54 Tykot

archipelago, minimizing the open-waterdistances traveled (e.g. Phillips 1975;Hallam et aL 1976: 99; Bagolini 1980: 42;Williams-Thorpe, Warren and Courtin1984: 141). In fact, obsidian has beenfound on the islands of Elba, Capraia,Pianosa and Giglio. All 14 artefacts that Ianalysed from the Cardial Impressed Waresite of La Scola on Isola Pianosa (Ducci andPerazzi 1991) are from Sardinia and includefive of the SB2 type, emphasizing again theimportance of this source in the EarlyNeolithic. Two artefacts from Elba are alsoSardinian (Hallam et al. 1976), while bothLipari and Monte Arci obsidian have beendocumented on Capraia, albeit without anyreal context (Arias et al. 1984; Bigazzi et al.1986). None of the extensive EarlyNeolithic finds from Le Secche on Isola delGiglio have been analysed, although theexcavator attributes them to Lipari(Brandaglia 1985: 59~60). Lipari obsidianwas not extensively distributed so far northin the Early Neolithic, however, and thepresence of cores and reduction debris sug­gests to me that Sardinian and/orPalmarolan obsidian might account for asizeable portion of the Le Secche assem­blage.

Southern France, Northern and Central Italy

In Europe and the Mediterranean regions,obsidian appears to have been distributedas part of a few almost mutually exclusivesystems: Anatolia and the Levant; theAegean and Greece; central Europe; andthe central/western Mediterranean. Theonly sure examples of Melian obsidianwest of the Balkans are three pieces fromGrotta del Leone-Agnano from an inse­cure context (Bigazzi et aL 1992b, 1986;Bigazzi and Radi 1981); their significanceshould not be over-estimated. Only twopieces of Carpathian obsidian have been

identified in Italy (one each at Grotta dellaTartaruga-Trieste and Sammardenchia diPozzuolo-Udine), although visual analysissuggests that more may be present atFornace Cappuccini~Faenza (Randle et al.1993; Polglase 1989; Antoniazzi et al. 1990:55). Five analysed, but unattributed, sam­ples from Le Crestair~Mornas andBeauvallon-Valence in France may alsocome from Carpathian sources (Crisci et aI.1994), but this needs to be confirmed byquantitative analyses. Finally, two uniden­tified pieces from Bellori-Verona andMisano Adriatico-Forli have extremelylow Na and very high Fe (Williams~Thorpeet al. 1979), compositions inconsistent withobsidian.

A few pieces of obsidian have been foundat sites north and west of Toulouse alongthe Garonne and its tributaries (Villeneuve-sur-Lot, Condom, St. Michel~du~Touch,Capdenac-le-Haut), on both sides of thesouthern Alps (Pierre Chatel-Belley inFrance, and Isolino di Varese in Italy), andeven in Spain near Barcelona (Guilaine andVaquer 1994; Pollmann 1993; Courtin1989). Several of these sites are about 200km from the Mediterranean coast, but onlythe obsidian from Isolino di Varese (ca.150 km north of Genoa) has been analysed(Hallam et aI. 1976; Williams-Thorpe et al.1979). The number of sites now withanalysed obsidian in the western Mediter­ranean is actually quite considerable,including 37 in France and over 100 in Italyand Sicily, and the list of Neolithic siteswith 10 or more analyses is growing (Figure3). Nevertheless, only five excavated sitesoutside Sardinia and Corsica have morethan 20 chemical analyses.

In southern France, obsidian from EarlyNeolithic sites is extremely rare, and theone piece analysed from Peiro Signado~

Portiragnes comes from Lipari (Crisci et al.1994). Obsidian is found more regularly -

Obsidian Procurement and Distnbution in the Central and Western Mediterranean 55

but still in very small quantities - atFrench Middle Neolithic (Chasseen) sitesin the form of blades, flakes and cores. AtGiribaldi-Nice, the 58 obsidian artefactsfound comprise only 0.5% of the lithicassemblage; the number of pieces found atthis site is surpassed only by the 70 fromLa Cabre/Le Grenouiller-Agay (Binder andCourtin 1994). Provenance determinationsof 143 obsidian artefacts from French sitesshow them to be predominantly ofSardinian origin. Only 24 analysed arte­facts are from Lipari, and most of themappear in earlier rather than later Neolithiccontexts, providing a contrast to theincreasing quantities of Sardinian obsidian.At the Chasseen site of La Ca bre/LeGrenouiller-Agay, 37 of 41 analysedobsidian artefacts are Sardinian, as are all22 pieces analysed from the Late Chasseensite of Sainte Catherin~Trets (Hallam et al.1976; Crisci et al. 1994). Since Monte Arciis considerably closer to southern Francethan IS Lipari, the predominance ofSardinian obsidian (by whatever mechan­ism and route it took) is not surprising;what is shocking is that 110 of the 117artefacts of Sardinian obsidian, from 27different sites in southern France, arespecifically of type SA. This regionalemphasis on a single Monte Arci source isunparalleled and deserves explanation.

Sardinian obsidian could have reachedsouthern France by several routes: directlyfrom the Monte Arci supply zone; viaCorsica; or via Tuscany and Liguria. Thefirst and second choices would supposemuch greater confidence and capability inopen-water crossings than most scholarsare willing to credit to Neolithic sailors,even if the ships hugged the western coastof Corsica before crossing 150 km of opensea to the Cote d'Azur. Interaction betweensouthern France and northern Italy isdocumented by the distribution of jadeitite

and eclogite axes from western Alpinesources in Liguria and Piedmont (Ricq-deBouard et al. 1990; Ricq-de Bouard 1993;Ricq-de Bouard and Fedele 1993). Theobsidian at southern French sites couldtherefore also be derived from the north­ern Italian regions, transported overinland, riverine or coastal routes along withanimal or plant products too (Phillips1982: 27-30, 43). The distribution patternof obsidian closely resembles that of theeclogites, which Ricq-de Bouard andFedele (1993) argue were not distributed bycoastal routes. Honey-colored flint, fromsources in the Rhone valley, and glauco­phane schists, from alluvial sources in thelower Durance valley, were widely dis­tributed on a more regional level. Hallam etal. (1976: 103), using an exponential fall­off curve model, argue that the density ofobsidian finds in northern Italy and south­ern France supports a down-th~line recip­rocal exchange mechanism. They concludethat Proven~e was the last in a networkseries from Sardinia to Corsica to Tuscanyto Liguria, and perhaps over the Alps intoProvence from the Po Valley (Hallam et al.1976: 99; see also Williams-Thorpe,Warren, and Courtin 1984: 141). The smallquantities reaching Provence indicated thatdirect trade with Sardinia, with or withoutmiddlemen, was very unlikely. Bloedow(1987), however, is particularly critical ofthe evidence for 'trade' in the NeolithicMediterranean, since there is no proof thatgoods were exchanged. For the westernMediterranean, he specifically argues thatparties from individual settlements mayhave travelled to the sources, if not exclu­sively, at least in many cases (Bloedow1987: 114).

Since obsidian from at-least two if notthree Monte Arci sources are commonlyfound at sites in northern Italy (see below),Corsica, and even in the Oristano area, the

56 Tykot

concentration on type SA obsidian insouthern France can only be explained by(1) conscious selection of that type (1 a)during acquisition from their neighbors''supply' (in a down~the-line sense) and/or(1 b) by procurers in the supply zone; or by(2) deriving the obsidian finds at so manysites ultimately from just a few 'shipments'(by any of the above routes) to the areawhich happened to be all of type SA obsid­ian. While type SA obsidian has beenfound at many sites in northern Italy, thereis no evidence there of its selection overthe other Sardinian varieties, either at thesite or regional level, and since the indige­nous Sardinians did not select type SAover the other varieties either, it seemsunlikely that Sardinian sailors would havecarried only type SA with them on theirvoyages. Selection, therefore, of type SAobsidian by 'merchants' from southernFrance (whether in northern Italy, Corsica,or even visiting Neolithic settlements in theCabras-Oristano area) would have been apeculiarly French phenomenon. Specificrock types were intentionally selected forground stone implements with widely dif­ferent frequencies in neighboring areas ofFrance, indicating that the availability andquality of the raw materials were not theonly factors influencing selection (Ricq-deBouard and Fedele 1993; 16), circum­stances which could equally well apply toobsidian (translucent vs opaque, with mostof the available translucent obsidian oftype SA, especially in the Middle-LateNeolithic). The second hypothesis seemshighly unlikely (too many sites with obsid­ian, at least a 1500-year time span), and theother, also unlikely, alternative appears tobe direct procurement from Monte Arci bysailors from southern France who onlyknew about the SA source and who passedaround directions to Conca Cannas forgenerations.

No obsidian from Palmarola has beenidentified yet in France, but two excep­tional pieces of obsidian from Pantelleriaare known from a Final Neolithic dolmentomb at San Sebastien (Williams-Thorpe,Warren and Courtin 1984), nearly 850 lin­ear km from their source. The absence ofPalmarola obsidian among the manyartefacts that have now been analysed fromsouthern France, considering its wide­spread (but modest) use in northern Italy,lends additional support to the hypothesisof French selection of translucent obsidianand/or frequent procurement via Sardinia/Corsica rather than Tuscany/Liguria.

In Northern Italy, obsidian is stilI only aminor part of lithic assemblages, with flintfrom southern Alpine sources the primaryraw material used during the Neolithic(Ferrari and Pessina 1994; Barfield 1987,1981). Of the obsidian present, all threeSardinian obsidian sources (SA, SB, sqare almost equally well represented. Incontrast to more-distant southern France,Lipari and Palmarola are also significantcontributors. In fact, the usual case seemsto be for at least two if not three islandsources to be represented at each site innorthern and western Italy where morethan a few artefacts have been analysed. Atthe Italian Middle Neolithic (YBQ) site ofGaione-Parma, obsidian from Sardinia(sq, Palmarola and Lipari is present, andthe strong tendency towards blades beingof Lipari obsidian and cores and trim ofSardinian obsidian suggests source~based

differences existed in the forms producedand circulated (Ammerman et al. 1990;Polglase 1990). The underlying motivationfor circulation of different forms of obsid­ian (cores = utilitarian, reduced on-site;blades = prestige, little-used?) emphasizesthe complexity of Neolithic exchange sys­tems and our difficulty reconstructingthem.

Obsidian Procurement and Distribution in the Central and Western Mediterranean 57

More definitive evidence comes fromArene Candide-Savona, where 53 obsidianartefacts from several Neolithic strati­graphic contexts have been analysed(Ammerman and Polglase 1993, 1996). Inthe Early Neolithic, obsidian comprisesabout 5% of the lithic assemblage at AreneCandide, and comes equally from Sardiniaand Palmarola. All three Sardinian varietiesare represented, with type SB by far themost important. In the Middle Neolithic,obsidian from Lipari replaces much of theSardinian SA and SB contribution (the fre­quency of opaque [Palmarola + SC] obsi­dian remains constant), and by the LateNeolithic, nearly all obsidian artefacts arefinished blades of Lipari obsidian.Ammerman and Polglase (1993, 1996) linkthis shift to the high quality of Lipari obs­idian and its circulation as a prestige itemin the later Neolithic. One must note,however, that they analysed only eight LateNeolithic artefacts from Arene Candide,and Sardinian obsidian is actually betterrepresented than Lipari among LateNeolithic artefacts from other sites innorthern Italy (combined analyses ofW illiams~Thorpe et al. 1979). Further­more, visual examination of 99 obsidianartefacts from Gaione (Polglase 1990)suggests equal use of Lipari and othersources during the Middle Neolithic, whilevisual examination of 334 obsidian arte­facts from Fornace Cappuccini-Faenza(Polglase 1989) may indicate that non­Lipari obsidian was more common therefrom the Early Neolithic through theChaIcolithic. Finally, my visual proven­ancing of 213 artefacts from PoggioOIivastro-Viterbo (Bulgarelli et al. 1993)revealed only 4 of Lipari obsidian in LateNeolithic levels. As usual, more analysesfrom more stratigraphically-controlledsites are warranted.

Some obsidian from Monte Circeo-

Latina, in southern Latium, was indicatedto be of type SA (Hallam et al. 1976: fig. 4),presumably making it the most southerlyoccurrence of Sardinian obsidian on themainland, but no analysis was reported anda later distribution map also produced byscholars at Bradford University shows onlyPalmarola obsidian from that site(Crummett and Warren 1985: fig. Al.2).Obsidian from Palmarola has been identi­fied at a number of sites in central peninsu­lar Italy, as well as in the Foggia area of theTavoliere, and as far south as the Gulf ofTaranto at Grotta Sant'Angelo-Cosenza(Hallam et al. 1976). Obsidian found at S.Domino in the Tremiti Islands was alsoreported to be from Palmarola (CornaggiaCastiglioni et at. 1962, 1963), but this attri­bution should not be taken for granted(e.g. Crummett and Warren 1985; Bigazzi etat. 1992a) since the analysis may be unreli­able. Of no small significance is the singlepiece of Pantellerian obsidian identified inan Early Neolithic context at VillaBadessa-Pescara, on the Adriatic coast ofcentral Italy (Bigazzi et al. 1992b).

Sardinian obsidian seems not to havemade it to the eastern shores of Italy,although it accounts for all but one of 25analysed samples from the Lake Regionnorth of the Po River. In the Trieste area atthe head of the Adriatic, most of the obsi­dian is from Lipari, and a few pieces arefrom Palmarola and Carpathian sources.Obsidian has been found at several coastalsites in Dalmatia, accounting for up to 2­3% of Danilo lithic assemblages (Martinelli1990). No analyses of this material havebeen published, although it had beenreported that some artefacts from Bosniawere of Sardinian origin (Rasson et al.1977), a conclusion unsupported by thedistributional evidence and perhaps due totrace element similarities between Sardin­ian and Anatolian obsidian (Williams-

58 Tykot

Thorpe 1995: 231). A couple of pieces ofAnatolian obsidian have been reported inGreece and eastern Europe (Renfrew andAspinall 1990), but Carpathian sources aremore likely to account for most of theDalmatian finds. Obsidian is found moresporadically in Albania, where it may beMelian in origin (K.M. Petruso, pers.comm. 1994).

Southern Italy, Sicily, Malta and NorthAfrica

In southern Italy, the Lipari sourceaccounts for virtually all of the obsidiananalysed to date. In Calabria, obsidian isparticularly plentiful (90% of lithicassemblages at sites on the west coast),since alternative lithic raw materials werenot available locally (Ammerman 1985a).Even though sites may be more than 100km from the Lipari source, cores were dis­carded at an earlier stage of reduction thanin northern Italy where they would be evenmore precious (Ammerman 1979, 1985b;Ammerman and Polglase 1993). StrongEarly-Middle Neolithic (Impressed Wareand Stentinello) ceramic parallels betweenCalabria and Sicily and the Aeolian Islandslink these regions in a single cultural unit,separate from Puglia and other areas ofsouthern Italy (Ammerman et al. 1978:191). Obsidian from Passo di Corvo­Foggia has been cited as Melian in origin(Pedes 1992: 145-146), but in fact the ESRanalyses performed by Mello (1983) couldnot distinguish between Melos Monte,Arci, and Palmarola - the last being themost likely possibility. To date, not a sin­gle piece of central Mediterranean obsidianhas been identified east of Italy's heel.

All the obsidian found in Malta comesfrom Lipari and Pantelleria, not Melos(Cann and Renfrew 1964; Hallam et al.1976; contra Cornaggia Castiglione et al.

1963). Visual identification of 300 artefactsfrom Skorba indicate that more than 85 %come from Lipari (Cann and Renfrew1964), indicating that Malta's socioeco­nomic affiliation was primarily to thenorth, certainly a much shorter open­water distance than to Pantelleria andNorth Africa. In Sicily, obsidian is largelypresumed to be from Lipari, but few siteshave been analysed. Nearly 40% of 152obsidian artefacts from Grotta dell'Uzzohowever, come from Pantelleria (Franca~viglia and Piperno 1987), a surprisingamount if obsidian were the only resourceobtained after such a large open-watercrossing. Pantellerian obsidian has alsobeen identified in Bronze Age contexts atMonte Cofano-Trapani (Francaviglia andPiperno 1987), and on the island of Usticato the north (Tykot 1995b). Along with thesporadic finds at Villa Badessa and SanSebastien, it appears that Pantellerianobsidian was more widely distributed tothe north than on the North Africanmainland. The distribution pattern forPantellerian obsidian is clearly significantfor our interpretation of the westernspread of domesticates from the easternMediterranean during the Early Neolithic.A coastal North African route (Lewthwaite1986a, 1986b, 1989) could have ultimatelysupplied domesticated animals and plantsto Sicily and the Italian peninsula, with anopen-water crossing of about 150 km be­tween Cap Bon and southwest Sicily. Ifregular crossings did take place, however,one would expect that Pantelleria (laO kmfrom Capo Granitola-Sicily) would havebeen visited often as well, resulting in thedistribution of a significant quantity ofobsidian (southward, if Sicilians travelledto North Africa; northward, if NorthAfricans went to Sicily).

Pantelleria, only 90 km from theTunisian coast, is presumed to have been

Obsidian Procurement and Dzstnbution in the Central and Western Mediterranean 59

the source of most obsidian found inNorth Africa. Obsidian is being identifiedat more and more sites, although rarely insignificant quantity (Camps 1964, 1974;and the multiple volumes in the seriesAtlas Prehistorique de la Tunisie,Collection de l'Ecole Franc;:aise de Rome81: 1985,1987,1989,1992). Not a singleanalysis of obsidian found at NorthAfrican sites has been published, however,even though several samples were analysedin 1976 (unpublished data; Williams~

Thorpe 1995: 229). These few analyseshave been cited by Crummett and Warren(1985: 108), whose map (fig. A1.2) indicatesthat Lipari is the source of obsidian froman inland site (Tebessa?), while Pantelleriais the source of material from a site in theBizerte region. The proximity of Pantelleriato the Tunisian coast makes it the mostlikely source of obsidian, since Lipari isconsiderably further away (ca. 400 km),and there is no evidence that material fromTibesti in northwestern Chad ever reachedthe Mediterranean coast nearly 1800 km tothe north. Some obsidian artefacts report~

ed from Tunisian and Algerian sites havebeen assigned a provenance on a visualbasis (Camps 1964, 1974), but areinsufficient in number to establish the rela­tive contributions of Lipari and Pantelleriato North African lithic assemblages. Iexamined 34 pieces of obsidian collectedduring surface surveys and excavation onthe small island of Zembra, and found allto be green in transmitted light and thusfrom Pantel1eria (Tykot and Vigne n.d.).

Discussion

Procurement

During the Early Neolithic (Cardial andFiliestru periods), there appear to havebeen few settlements in the Gulf of

Oristano area, permitting unhinderedaccess to Monte Arci and its obsidianresources. In extant traditional societies,lithic raw materials are often acquired inthe course of other activities such as sea­sonal tnmshumance, in other words as anern bedded strategy (Binford 1979). Shep­herds, farmers, hunters, or fishermenpassing through the Campidano plain tothe west of Monte Arci would have foundit particularly easy to locate type SB2 obs­idian in secondary geological contexts onthe western flanks of Monte Arci and in

situ deposits right nearby. I suspect that themore concentrated Conca Cannas obsidianflow was less visible to the casual observer,particularly non-local prospectors comingeven from adjacent regions. More distantvisitors arriving by boat in the Gulf ofOristano would have been closer to theSB2 source, and even if they entered thecoastal lagoons south of Arborea, wouldstill have been 13 km from the ConcaCannas source. While they might havebeen able to get somewhat closer if theMannu or Mogoro rivers were navigable,obsidian otherwise was transported byhuman hands, since the only animals(cattle) capable of carrying a load ofobsidian nodules must have beenextremely rare in the Early Neolithic.Residents of areas east of Monte Arcicould have come across type SC obsidianin secondary deposits without having toascend the steep ridge to Punta Pizzighinu.

In the Middle Neolithic (Bonu Ighinu),the settlement of the Cabras and Oristanoarea meant that residents of neighboringand/or distant communities could obtainobsidian without going to Monte Arcithemselves. Agricultural intensificationmay have led to fewer seasonal rounds,long distance hunting trips, etc., ultimatelyresulting in less embedded procurement ofobsidian. In any case, ethnographic data

f.L

60 Tykot

suggest that direct procurement from rawmaterial sources in another community'sterritory is rare, in part because therepeated reciprocal exchanges that typicallycharacterize privileged expeditionary rela­tionships between communities obviatesthe need for such trips (Feblot-Augustinsand Perles 1992). Nevertheless, both agri­culturalists in California and hunter-gath­erers in Australia were able to quarry intheir neighbors' territory freely or inexchange for small gifts (Bryan 1950;Gould et al. 1971). In such cases, a sequen­tial production strategy is often employedto produce utilitarian items from the rawmaterials back at their own village; luxurygoods are more likely to be completed atthe quarry or in its immediate area by localresidents (Ericson 1981).

The Oristano province locals, certainlyknowledgeable about all the available obs­idian sources, apparently preferred typesSA and SC over SB2, but passed on alltypes (including SB 1) to Corsica and themainland, mostly in unfinished form(preform cores or raw nodules). Lilliu(1989) writes that oro nero sardo wastransported to the mainland by local mer­chants from the Gulf of Oristano, and evensuggests that the manufacture of ground­stone objects may have been centered inthe Monte Arci zone because of stimula­tion from obsidian workshops there (Lilliu1986). It is uncertain, however, to whatextent specialists were involved in the pro­curement, production or transport of obs­idian from Monte Arci. In the EarlyNeolithic, the island's low populationdensity and the incipient level of newly­introduced domestic food productionwould seem to make full-time lithic spe­cialists unlikely for socioeconomic reasons,in contrast to the Middle Neolithic whenincreases in settlement numbers and moreintensive agropastoralist activities are

accompanied by innovative exploitationand production of ground stone vases andfigurines in addition to widespread obsi­dian distribution (see Tykot 1997b for adetailed review of the Neolithic periods).Excavation and analysis of lithic quarriesand workshops would provide directinformation on raw material selection,extraction and reduction technology, andknapping behavior, as well as chronologicalchange in production, exchange and tech­nology (Ericson 1984). No workshop sitesin Sardinia have been excavated, nordetailed studies done to determine lithicreduction skill and efficiency. There issome evidence, however, that knappingwas better controlled and forms morestandardized in the Middle Neolithic thanin the Early Neolithic levels at GrottaFiliestru (Hurcombe and Phillips n.d.),while use-functions of individual toolswere also more specific in the later period(Hurcombe 1992a, 1992b). This is consis­tent with a general impression of progres­sive typological standardization andmaximization of core material by the LateNeolithic (Garibaldi 1993).

Labor specialization must have existed inthe central Mediterranean in the EarlyNeolithic, judging from the technologicalexpertise represented in the extensive flintmine complex of Defensola-Vieste in theGargano peninsula, which may have beenworked year round (di Lernia et al. 1990­91). Likewise, Leighton (1992) argues thatraw materials for ground stone tools weretaken from local sources in southern Italyand worked at nearby settlements, ulti­mately being exchanged to adjacent com­munities as finished objects. The involve­ment of particular communities with pro­curement and manufacturing - villagecraft specialization - would naturally havebeen favored when raw material resourceswere only locally available.

Obsidian Procurement and Distribution zn the Central and Western Mediterranean 61

Finally, it has been suggested that accessto the Monte Arci sources may havebecome restricted by the Bronze Age,based on the presence of only one type ofobsidian at some sites (Michels et al. 1984).As noted above, the likelihood of most sitecollections representing multiple genera­tions of human activity, and the strongpossibility of reuse of locally available obs­idian in the Bronze Age, makes it extremelydifficult to test this hypothesis. Obsidianartefacts from tomb contexts may well befrom a single source, but are also morelikely the result of reduction from a singlecore, as are artefacts from residential siteswhere core trim and debitage are alsofound. Differentiating between direct pro­curement from a single source, and acqui­sition of obsidian of a single type fromintermediaries is virtually impossible.

Distribution

A now substantial number of analysescontinues to support the notion of mini­mal overlap in obsidian exchange betweenNeolithic Europe, the ,vestern Mediter­ranean, the Aegean, and the Near East. Thelimits of obsidian distribution probablyalso applied to the movement of othermaterials and ideas, and may reflectcultural or ethnic boundaries. It shouldalso be understood that these distributionpatterns represent cumulative actions Overlong periods of time; it may have takengenerations for some obsidian to travelfrom its geological source to its findspot(Williams-Thorpe 1995).

Provenance determinations of more thana few artefacts each from an increasingnumber of Neolithic sites in the central andwestern Mediterranean have contributedgreatly to our understanding of obsidiandistribution from each island source,although the number of analysed speci-

mens from the mainland, especially penin­sular Italy, remains inadequate (Figure 3).Obsidian from Monte Arci is found inEarly Neolithic contexts throughoutSardinia, as well as in Corsica, the Tuscanarchipelago, and northern Italy, althoughthere is some evidence that not much if anyobsidian reached southern Corsica in thevery first phase of the Early Neolithic(Lanfranchi 1993) and few excavationselsewhere have produced fine enough dis­tinctions within the Early Neolithic toaddress this possibility. Obsidian fromLipari, Palmarola, and Pantelleria was alsodistributed in the Early Neolithic, withlong~distance circulation of obsidian fromall four islands peaking during the Middle­Late Neolithic in southern France(Chasseen) and northern Italy (VBQ andLagozza). Courtin (1967: 105, fig. 5)implied long ago that Sardinian obsidianreached southern France directly and inde­pendently of Tuscany and Liguria, whileCamps (1976a, 1976b) specifically indicatedthat Monte Arci obsidian probably passedalong the west Corsican coast and thenacross the open sea to Provence, near St.Tropez. The presence of Pantellerian obsi­dian in Early Neolithic levels at Grottadell'Uzzo in Sicily is direct evidence ofopen-water crossings of at least 100 km, toa tiny island destination, and attests to thenavigational skills of early seafarers. Theopen-water distance between Corsica andsouthern France is of the same order ofmagnitude (150 km), indicating that suchtrips could well have taken place on aregular basis during the Neolithic. Phillips(1992, 1986, 1982) more reasonably sug­gests that obsidian probably reachedsouthern France by several routes,including directly from Sardinia, but alsoby cabotage along the Ligurian coast andoverland from the Po Valley.

Regional differences in form and fre-

62 Tykot

quency of the multiple obsidian sourcespresent at individual sites are apparenteven when the obvious factors of relativesource distance, quality and abundance aretaken into account (Figure 8). High qualitytranslucent obsidian from Lipari predomi­nates throughout southern Italy, Sicily, andMalta, overshadowing the more limitedquantities of opaque Palmarolan andPantellerian obsidian circulating in thesame areas and ultimately reaching thesame or even more distant distributionalendpoints in North Africa, Languedoc, andTrieste. Three major varieties of MonteArci obsidian are exclusively found inSardinia and Corsica, predominate insouthern France, and 'compete' with Lipariand Palmarola obsidian in northern Italy.Hurcombe and Phillips (n.d.) haveobserved that type SA obsidian may havereached Gtotta Filiestru in northernSardinia in finished form, whereas abun­dant evidence for the reduction of type SCexists at the cave site. The large quantity ofblade cores and debitage in southernCorsica indicates tools of Sardinian obsi­dian were produced there too, and sixworkshop sites have been identified insouthern France (Phillips 1982, 1992;Binder 1987) even though the amount offlaked material was small. In northernItaly, Sardinian obsidian also appears tohave arrived as cores rather than finishedblades, in contrast to obsidian from Lipari(Ammerman et aL 1990; Ammerman andPolglase 1993). At Arene Candide, bladesof Sardinian or Palmarolan obsidian arealso noticeably larger than those frommore distant Lipari (Ammerman andPolglase 1996).

The relative use frequency of the differ­ent Sardinian sources is particularly signifi­cant for our understanding of procurementand distribution activities, and raises ques­tions about the differential functional

quality of the various obsidian sources aswell as their aesthetic characteristics. TypeSA - and some SB2 - obsidian is quitetranslucent and glassy and thus similar inappearance to Lipari obsidian, while typeSC is opaque, less glassy, and resemblesPalmarola obsidian (and Pantelleria tooexcept for thin pieces in which the greencolor is more readily apparent). Glassy obs­idian is considered too brittle for certaintasks, and generally dulls quickly; in thosecases flint would actually be the preferablestone tool material. Knapping experimentsand use-wear analysis of obsidian assem­blages will be necessary to determinewhether types SA and SC obsidian werebest suited - and employed - for differ­ent tool forms and use-functions.

The distributional evidence indicates thattypes SA, SB2 and SC obsidian were allused at Neolithic sites in central and north­ern Sardinia, with type SB2 the least fre­quent of the three, and rarely found insouthern Sardinia. In the Early Neolithic,however, it seems that type SB2 obsidianwas much more commonly used and evenpredominates in some assemblages, notonly in Sardinia but also in northernCorsica, the Tuscan archipelago, andnorthern Italy (Figure 9). This could resultfrom its ready accessibility in a sparselypopulated area of the island. By the LateNeolithic, types SA and SC are the onlysources regularly found at individual sites,with type SC obsidian actually accountingfor a greater percentage of most assem­blages. For the most part, the relative fre­quencies of the multiple Monte Arci obsi­dian sources represented in northernSardinia are quite close to those in Corsicaand northern Italy. Southern France is thenotable exception, where more than 95%of the Sardinian obsidian for the entireNeolithic is type SA. The most parsimon­ious explanation for the presence of almost

143 189 63 128 159 25

0:;":;&­~

§;:;-;r'

'"()'";:l~.";:lP.-

~'"On

;:,-<;:l

5::'"P.-

'"~'".";:l

o~P.:­:=;";:l

~g~'"::l'"~l'>;:lP.-

tJ

.Pa&I PIo Li.MAesc~SB

LillI SA

MaltaSicilyS. ItalyC. ItalyN. Italy:;'rance

;idian source frequency in the Neolithic central and western Mediterranean, by region. All analyses by chemical orsical methods; attribution to unspecified Sardinian sources (MA) is based on fission~trackdating. Numbers of artefacts'esented are at the top of each bar. 0\

I.N

64 Tykot

exclusively type SA obsidian in southernFrance, for a period spanning more than amillennium, is particular selection of thattranslucent variety for reasons unimpor­tant or not applicable to their neighbors atArene Candide and elsewhere in northernItaly where a diversity of sources is repre­sented. If Sardinian (and/or Corsican) boatstravelled to southern France or northernItaly, they likely would have carried obsi­dian from multiple Monte Arci sourcesand left it all behind before returninghome, probably laden ,',lith mainlandproducts. Since mostly type SA is found insouthern France, and the current in theLigurian Sea runs counterclockwise (i.e.northern Italy would probably have beenvisited before southern France bySardinian/Corsican sailors), it is morelikely that French merchants specificallyselected type SA obsidian over the othervarieties available, whether that was atports of call in Liguria, Corsica, or evenSardinia. Lipari obsidian, when available sofar from its source, was also apparentlyacceptable, but Palmarolan obsidian ­even though it is well attested at AreneCandide - was not. Differences in obsi­d ian selection criteria between southernFrance and northern Italy may also bertidted to the availability of alternativelithic materials, or evolved cultural prefer­ences or ethnic ties dating from the begin­ning of the Early Neolithic; in southeasternFrance, neolithisation involved the accul­turation of Cardial characteristics byCastelnovian (Late Mesolithic) peoples,whereas in Liguria there seem to have beenno Late Mesolithic antecedents (Binder1987, 1989; Vaquer 1990; Biagi et al. 1989).

Obsidian continued to be used inSardinia and Corsica after the Neolithic,judging from its regular presence atCopper, Bronze and even Iron Age sites(Contu 1990-91) and from obsidian

hydration dating studies (Michels et al.1984; Dyson et al. 1990; Stevenson andEllis n.d.). There is also growing evidenceof its continued use (at diminished levels)at contemporary mainland sites, andalthough this may represent recycling oflocally available material, there is otherevidence that contacts between Sardiniaand Tuscany ,vere maintained after theNeolithic (Vigliardi 1980).

Exchange

In reconstructing prehistoric exchangesystems, Renfrew (1993a) has emphasizedthe consideration of the full range ofhuman interactions, and the possibilitythat these interactions may have been char­acterized by 'communication' rather than'trade' if the acquisition of goods played asecondary or minor role. What has becomestrikingly clear is that multiple, distinctsystems of production and exchange oper­ated simultaneously In NeolithicMediterranean societies, pertaining to dif­ferent categories of material goods (Pedes1992; Skeates 1993). Cardial ceramics, forexample, were also being exchanged in theEarly Neolithic, although the distancesinvolved were on the order of 50-70 kmrather than the hundreds traveled by obsi­dian (Barnett 1990a, 1990b). We also haveevidence for Neolithic trade of flint,greenstone, and various kinds of fineceramics up and down the Italian penin­sula, along the Po Valley, and across thehead of the Adriatic to Dalmatia. It shouldbe no surprise that topography played animportant role in the routes these materialsfollowed; cabotage along either theTyrrhenian or Adriatic shoreline mademore sense than transport across moun­tainous inland areas - although thisoccurred in the Alps in the case of flint andgreenstone, and across the Appenines in

Obsidian Procurement and Distribution in the Central and Western Mediterranean 65

the case of Ripoli trichrome pottery.Finally, we know that Near Eastern domes­ticates were introduced to the centralMediterranean during the Neolithic (Tykot1995a, 1997b; Donahue 1992; Lewthwaite1986a; Costantini 1989), and along withother potential trade goods like salt,basketry and textiles which are notparticularly visible in the archaeologicalrecord, would have formed the basis forthe Neolithic economic system. We musttherefore consider that the distribution ofobsidian not only is linked to the circula­tion of other material goods of similarutilitarian, social, and/or symbolic signifi­cance, but that the particular exchangesystem it belonged to probably dependedupon dynamic factors such as local customand the availability of the material whichwere likely to change over time.

In the central and western Mediter­ranean, obsidian was not used during theUpper Paleolithic or Mesolithic, despitethe likelihood that its sources were knownto the pre-Neolithic residents of Corsica,Sardinia, and Sicily, and its presence fallsoff sharply after the Late Neolithic inmainland Italy (Lagozza) and southernFrance (late Chasseen). This situationparallels the use of amber in Britain, whereit was a high prestige item in the EarlyBronze Age Wessex culture, but is almostarchaeologically invisible in the precedingNeolithic and the succeeding MiddleBronze Age (Beck and Shennan 1991). Inthe eastern Mediterranean, pre-Neolithictrade has been considered an incentive forthe adoption of agriculture and theproduction of surplus wealth to acquirenon-local goods (Runnels and van Andel1988; d. also Tangri 1989; Runnels 1989).But in the central and western Mediter­ranean obsidian distribution coincides withthe expansion of village farming, and ismore likely the consequence of this new

sedentism. The dramatic shift in thesubsistence economy is accompanied bythe development of craft specialization ­evident not only in the skilled knapping ofobsidian and flint tools, but also in ceramicproduction and other activities - but craftspecialization is certainly not a prerequisitefor obsidian exploitation, particularly nearits source. In its 'supply zone' obsidian waslikely to have been primarily utilitarian infunction, whereas at greater distances itsdiminished availability added a prestigecomponent to its use. Jadeitite and eclogiteaxes found in northern Italy and southernFrance appear well used and less perfectlyfinished than more finely-made, lightly­used axes found in southern Italy whichprobably served a non-utilitarian or orna­mental purpose (Leighton 1992). While itwould be tidy to link the southward distri­bution of 'jadeite' from the western Alpsto the northward dispersion of Lipari obs­idian, this cannot be assumed since theymay have belonged to different exchangesystem categories depending on a site'slocation relative to the two opposingsources, and because interaction betweengroups may not have been symmetrical(Renfrew 1986; Champion 1989). Likewise,the circulation of Sardinian obsidian, flint,nephrite, steatite, and salt cannot be con­sidered within a single exchange category(Garibaldi 1993; Lilliu 1986).

The relative quantity of obsidian foundat sites of differing distances from itssource has been used to define fall-offcurves which may be characteristic of cer­tain exchange mechanisms of sociopoliticalsystems. The quantification of obsidianfrequency, by number or mass of tools,debitage, and cores relative to otherchipped stone tools or other measures ofsite size, is still theoretically desirable(Ericson and Baugh 1994), although often

27 72 10 10 14 26

~.Q\

'-J'-<'

~

III PI.SCDSBLillI SA

liestru Basi Strette Pietracorbara Pianosa Arene Candide

'bsidian source frequency at several Early Neolithic sites. All determinations by chemical analysis (Tykot 1995a;mmerman and Polglase 1993, 1996 for Arene Candide). Numbers of artefacts represented are at the top of each bar.

Obsidian Procurement and Distribution in the Central and Western Mediterranean 67

difficult to apply given the considerablevariation in reported excavation data in thelast century (see Guidi 1987). Such sys­tematic data are necessary for diachronicand spatial analysis of obsidian use, and anadmirable effort has been made recentlyfor the central and western Mediterraneanby Pollmann (1993) who confirms theempirical hypothesis that obsidian use wasgreatest during the Italian Middle-LateNeolithic (VBQ, Ripoli, Serra d'Alto,Lagozza, Diana; Chasseen in southernFrance), with presumed transportationroutes and geographic barriers significantlyinfluencing the quantities found. Theidentification of centers of redistributionshould be made cautiously, however, sincethe quantity of obsidian found at Pescale innorthern Italy (9S0 pieces) is no longerunusual, or surprising. Hundreds of obsi­dian artefacts have been found at FornaceCappuccini-Faenza (Polglase 1989;Antoniazzi et al. 1990; Montanari et al.1994), Podere Uliveto and La puzzolente­Coltano (Cocchi Genick and Sammartino1983; Sammartino 1986), and extrapolationfrom the surface finds at Gaione suggeststhat thousands may be present there in justthe plough-zone levels (Ammerman et al.1990).

While privileged access to non-localgoods may have enhanced the prestige oflocal elites - by both the exotic nature ofthe material and any accompanying exoticor secret knowledge (Renfrew 1993a;Helms 1988) - there is little evidence atpresent to indicate that obsidian fulfilledsuch a role in prehistoric Italy. Accordingto the structuralist, prestige-goods econ­omy model, elite sociopolitical status ischaracterized by control of commoditieswhich are scarce, require specialist pro­duction, and/or are associated with morepowerful social systems (Baugh andEricson 1992: 10). During the Early and

Middle Neolithic, however, social differ­entiation within culture groups was mini­mal, as was any political hierarchy betweengroups, and it is only in the Late Neolithic(Lagozza, Late VBQ, Diana) - after thezenith of obsidian distribution - that agri­cultural intensification and an increase inthe variety and quantity of material goodsin circulation signify the growth of rankingand increased emphasis on prestige goods(Phillips et al. 1977; Barker 1981; Shennan1982; Phillips 1993).

Conclusion

The typology, relative quantities andarchaeological contexts in which obsidianis found in the central Mediterraneansuggest that it was primarily utilitarian infunction, but that its presence was theresult of social and multi-level economicinteractions including prestige exchange,both of which probably had particularlocal characteristics. The basic exchangesystems that existed may be understood asintertwined local networks rather than as awhole world system (Renfrew 1993a: 7),and in combination with some direct long­distance contacts - mostly via coastalmaritime routes - resulted in the distri­bution of mainly raw materials and utilitar­ian products, although their 'prestige' valuewould have been enhanced at great dis­tances from their source. In the Neolithic,domesticated sheep, goat and cattle, andtheir secondary products, were among themost likely commodities exchanged, espe­cially so for the islands of Sardinia andCorsica where they must have been inten­tionally introduced and were a necessarydietary supplement to the limited indige­nous fauna (see Lewthwaite 1981). Again,specifically linking extra-insular exchangeof Sardinian obsidian with incomingdomesticates is tidy, but impossible to

0-.00

ITALY

-l:habcrt (Sainte-Eulalic) I. Isolino di VaTese 42. Grotta Ilella (MOnlecastrilli) 82. Torre Testa (Brindisi) '"";<:--

,s (Ventenac-Cabardcs) 2. Grotta I'ollera (Savona) 43. Ripoli (Corropoli) 83. Torre S. Su.•sanna g'cassanne) 3. Arene Candide (Finale Liguria) 44. Poggio Olivaslro (Vulei) (Masscria Guidone)rc-de-Roquelongue 4. Pianocc ia di Suvero 45. Valle OttaTa (Cittaducalc) 84. Masseria S. Gaetaoo (Guagnano)do (Portiragnes) 5. Monte Covolo (Brescia) 46. Catignano (La Slepara) 85. Laghi Alimini (Olranlo)Fonlaine (Congenies) 6. Rocca di Manerba 47. Villa Badessa (Rosciano) 86. Campi Latini (Galatone)

>aint-Loup (Tresques) 7. Riparo Valtenesi (Manerba) 48. Ponte Peschio (Genzano) 87. Gratia della Trinita (Ruffano): (Orange) 8. Bellori (Grezzana-Verona) 49. Polidoro (Roma) 88. Grotta Grande di Latronico (Potenza)~s Roches (Piolene) 9. Grolta G. Perrin (Sengia Bassa 50. Setteville (Tivoli) 89. Grotla del Romito (Papasidero)Combes (Piolenc) di San Cassiano) 51. Via Pontina (Roma) 90. Grotta Sanl'Angelo (Cassano Ionia)(Momas) 10. Sammardenchia di Pozzuolo 52. Paterno (Avezzaoo) 91. Acconia(Chateauneuf-du-Rhone) II. Grotla della Tartaruga (Trieste) 53. Santo Stefano (Ortucchio) 92. Bevilacqua (Acconia)(Valence) 12. Vlaka Jama (Trieste) 54. Fossacesia (Chieli) 93. Preslarona (Canolo)lonnaire 13. Riparo di Monrupino (Trieste) 55. Batteria (Monle Circeo-Roma) 94. Isola di Filicudi-en-Diois) 14. Grotta degli Zingari (Sgonico) 56. Campo Mezzomonte (Roma) 95. Isola di LipariBlanches (Menglon) 15. San QuiTino (Trieste) 57. Isola di Ponza 96. Castellaro Vecchio (Lipari)['OlivieT (lstTes) 16. Grotta Lonza (Monrupino) 58. Isola di Capri 97. Lipari Castellor (Grans) 17. Grotta dell'Ansa (San Pelagio) 59. Grotta delle Feld (Capri) 98. Isola di Ustieaeres (Vemegues·Cazao) 18. Gaione (Parma) 60. Pompeii 99. Monte Cofano (Trapani).ambesc) 19. Rau.a di Campegine 61. La Panettaria (Lucera) 100. Grotla dell'Uzzo(Trapani)=(Vauvenargues) 20. San Polo d'Enza 62. Lucera 101. Santa Panegia (Siracusa)e(Mime!) 21. Chiozza (Scandiano) 63. Casone (San Severo) 102. Arenella (Siracusa)erine (Trets) 22. Villa Agazzolti (Formigine) 64. Monle Aquilone (Manfredonia) 103. MUTsia (Pantelleria)Baume (Gemenos) 23. Cava Nuova (Fiorano) 65. Grotta Scaloria (Manfredonia) 104. Isola di Pantelleriana (La Londe-Ics-Maures) 24. Pescale (Prignano) 66. Passa di Corvo (Foggia) 105. Cala Pisana (Lampedusa)~glise SUpCrieure 25. Spilamberto (Modena) 67. Masseria Leonessa (Melli)·sur-Verdon) 26. Fomaee Cappuecini (taenza) 68. Bad MALTA,asse) 27. Misano Adriatico (Riccione) 68 Pulo di Allamura (Bari) 1. Skorba~ (Vi lIecraze) 28. Grolla del Leone (Agnano) 68. S. Candida (Bari)(Tourtout) 29. Piazza della Signoria (Firenze) 69. Gravina di Picciano (Matera) TUNISIA

(Ramaluelle) 30. PodeTe Uliveto (Livomo) 70. Serra d'Allo (Malera) 1. La Galiteen (Plan-de-la-Tour et 31. Isola di Capraia 7L Fonte di Vita (Matera) 2. Rernel (Bi7~rte)

ime) 32. Isola d'EI ba 72. GroHa Funeraria (Matera) 3. Environs de Bi7.ertelaint-Tropez) 33. La Scola (Isola Pianosa) 73. Murgecchia (Matera) 4. Djebel cd Dib (BechateuT)ontrouge 34. Grolta del Fontino (Vallerotana) 74. Murgia Timone (MateTa) 5. Ile de Zembrala~l, Agay) 35. Pienza 75. Grotla dei Pipistrelli (Matera) 6. KOTbares (Sainl-Rapha~I, Agay) 36. Cava BaTbieri (Pienza) 76. Pizzica Panlanello (Metaponto) 7. Sebkhet Halk el Mermzel (Hergla)~ Grenouiller 37. GTotla del Bealo Benincasa (Pienza) 77. Torre Sabea (Taranto),a~l, Agay) 38. GTolla dell'Orso (Sarteano) 78. Torre Canne (Fasano) ALGERIACarros) 39. Santa Maria in Selva (Treia) 79. Torre Bianca (Fasano) 1. La Marsa (Skikda)ice) 40. Argentano (GTosseto) 80. Fontanelle (Oslun;) 2. ATn Khiar (Annaba)lice) 41. Ischia di CastTO (Vilerbo) 81. Gratia Morelli (Ostuni) 3. Tebessa

\=-'";1 -"'J

~

/

~~f<-?)/~~~~] (/ ;rYl

7'/ ()

(:3; '> " /~Ci ",c", 'J

,. '"

l 250km

sites in Sardinia & Corsica not shown

ov~,

e:­§:por;

~

~'";;~~~

\:Jc;;''=<&­:::g;;-~"Q;3

K.,;3~

~'";:;l",-<;3

3:'"P...

;;Z=l.,;::,

'".,;::,

sidian distribution in the central and western Mediterranean: source determinations by chemical, physical, or visuallysis. (Nearly all of the sites shown are Neolithic, but many are represented by analyses of obsidian from unstratified~texts. In Italy, sites 7, 17, 34, 60, 94, 98, and 103 are post-Neolithic.)

0­'0

70 Tykot

demonstrate archaeologically. Neverthe­less, a similar proposal has been posited forthe exchange of cattle and obsidian at CatalHuyuk in Turkey (Sherratt 1982).

Simultaneously, ethnicity and the main­tenance of kin connections are likely tohave been significant factors In the creationof preferential social exchange partners,and the importance of alliances and cere­monial behavior in the functioning ofexchange networks cannot be underesti­mated. Fine ceramics, shell and otherexotic materials are commonly found inritual or ceremonial contexts, and these arejust the Neolithic commodities that arearchaeologically visible (Skeates 1993;Malone 1985). Such occasional ceremonialevents may also have provided the maincontext for the social exchange of obsidian,ground stone axes and other lessspectacular materials. Group identity andthe maintenance of kin relations wouldhave become important social issues withthe changes in mobility due to agriculturalsedentism in the Early Neolithic, and wide­spread exchange networks would haveintegrated dispersed communities throughthe common behaviors associated with theexchanged items, especially decoratedceramics and the eating and drinking habitsassociated with them (Chapman 1988).

Chronological change in prestige-typematerial goods in the central and westernMediterranean may be described in tetmsof three phases:

(i) In the Early Neolithic, ceramic pro­duction was an innovative technology, sothat highly decorated Cardial wares wereespecially prized, and were produced inmultiple workshops spread across much ofthe region. The great stylistic similaritymay have served to maintain social affili­ations among groups, while local modifi­cations may have operated as boundarymaintenance symbols and defined particu-

lar social groups.Oi) In the Middle and Late Neolithic, we

see the addition to the northern mainlandrepertoire of obsidian, a material availableonly in restricted, island localities, andonly sparsely available on the mainland inthe Early Neolithic. The generally smallquantities found at large numbers of indi­vidual sites surely implies its social ratherthan economic importance, and the trans­parency of obsidian from certain sourcesmight have been its most sought after fea­ture (Barker 1981). The best way to modellong-distance obsidian exchange in thisperiod may be as multiple small-scalemovements like those demonstrated byBarnett (1990a, 1990b, 1995) for fineceramics.

(iii) The third stage begins with theappearance of metal artefacts at laterNeolithic and Chalcolithic sites, both onthe islands and on the mainland. It is at thistime that elaborate chambered tombsappear in much of the central Mediter­ranean region (Whitehouse 1981; Jous­saume 1985; Guilaine 1992), but obsidian isnoticeably absent in them except for thoseon Sardinia and Corsica. Also, metal didnot replace stone as the material mostcommonly used for tools in the Chalco­lithic or even the Early Bronze Age formost of Italy; the absence of obsidian inlithic assemblages must mean then that itsprestige value declined, not that exchangewith the source islands was abandoned.In modeling distribution and exchangesystems, it is necessary to go beyond thesimple dispersion diagrams showing thegeographic extent of a single material's dis­tribution (Figure 10), and consider thechronology, quantity and quality of thematerial circulated, the economic andsocial context(s) in which obsidian andother products were acquired orexchanged, and the particularistic factors of

Obsidian Procurement and Distribution In the Central and Western Mediterranean 71

transportation methods and routes, differ­ential use-function and value pertaining toindividual prehistoric communities.

In conclusion, while we must be cautiousin our extraction of economic and socialinformation from lithic distribution pat­terns, this remains a worthwhile endeavor.Characterization alone is not enough; thederivation of structured models that inte­grate social and utilitarian function fullywithin specific exchange systems isnecessary to interpret 'trade' in ancientsocieties. The extent and significance ofobsidian distribution from the centralMediterranean sources are still beingrefined, but it is clear that comprehensivesourcing of obsidian assemblages providesa different yet clearer picture of sourceexploitation, production and exchangethan does selective analyses of smallnumbers of artefacts. Analyses of largenumbers of artefacts permits chronologicalcontrol at least at the site level, andtherefore insight into dynamic changes inobsidian source selection. The integrationof source data, form, function, andreduction sequence for whole assemblagesof obsidian artefacts from well-datedarchaeological contexts will ultimatelyprovide a more complete understanding ofprehistoric socio-economic systems andhuman behavior in the central and westernMediterranean.

Acknowledgments

This research was supported by grantsfrom Sigma Xi, The Scientific ResearchSociety; the American School of Prehis­toric Research; and the Department ofAnthropology and the Graduate School ofArts and Sciences Student Council,Harvard University. Archaeological sam­ples from Sardinia and Italy were gener­ously provided by the Soprintendenza

Archeologica per Ie Provincie di Cagliari eOristano (Dott. Vincenzo Santoni, soprin­tendente); the Soprintendenza Archeolog­ica per Ie Provincie di Sassari e Nuoro(Dott.ssa Fulvia Lo Schiavo, soprin­tendente); the University of Cagliari (Prof.Enrico Atzeni); the SoprintendenzaSpeciale al Museo Nazionale Preistorico edEtnografico 'L. Pigorini' (Dott. GiovanniScichilone, soprintendente); the Soprinten­denza Archeologica per la T oscana (Dott.Francesco Nicosia, soprintendente); andBrown University (Prof. R. RossHolloway). Dr. David Trump gave permis­sion to analyse the material from hisexcavations at Grotta Filiestru and Sa 'Uccade Su Tintirriolu in northern Sardinia, asdid Prof. Miriam S. Balmuth for the obsid­ian from Nuraghe Ortu Ct)midu in south­ern Sardinia. I thank Dott.sse LuisannaU sai, Donatella Salvi, Francesca Galli,Vanna Canalis, Paola Perazzi, and GraziaBulgarelli, and Dott. Rico Pellegrini fortheir help in accessing the various collec­tions. Dr. Joseph Cesari (Direction desAntiquites Prehistoriques de Corse) pro­vided most of the Corsican material,including many samples from recent exca­vations by Dr. Jacques Magdeleine.Additional obsidian from Corsican sitescame from the University of Cagliari (Prof.Enrico Atzeni). Archaeological samplesfrom Tunisia were supplied by Dr. Jean­Denis Vigne (CNRS, Paris). Finally, I thankthe anonymous reviewers for their usefulcomments and suggestions, which 1 hopehave resulted in improved clarity of thispresentation.

About the author

Robert H. Tykot, Assistant Professor ofAnthropology at the University of SouthFlorida, received his B.S. and M.A. inClassical Archaeology from Tufts

72 Tykot

University, and an M.A. and Ph.D. inAnthropology from Harvard Universitywhere he managed the ArchaeometryLaboratories for six years. He has been theassistant director of excavations at a BronzeAge site in Sardinia (see L.]. Gallin andR.H. Tykot, Metallurgy at Nuraghe SantaBarbara (Bauladu), Sardinia, Journal of FieldArchaeology 20 [1993] 335-45), and IS

currently investigating the Mesolithic­Neolithic transition in the central andwestern Mediterranean. His research inter­ests also include the sources of Greek andRoman marble sculpture and the spread ofagriculture in the New \Xlorld.

References

Adams, R.M.

1974 Anthropological perspectives on ancienttrade. Current Anthropology 15: 239-58.

Akalu, A., and P. StJernquist

191\1\ To what extent are ethnographic analogiesuseful for the reconstruction of prehistoricexchange? [n B. Hardh, L. Larsson, D.Olausson, and R. Petre (eds), Trade andExchange in Prehistory. Studies in Honour ofBerta Stiernquist. Acta ArchaeologicaLundensia 16: 5-13. Lund: LundsUniversitets Historiska Museum.

Ammerman, A.].

1979 A study of obsidian exchange networks inCalabria. World Archaeology 11: 95-110.

1985a Anthropology and rhe study of neolithicexchange systems in Calabria. Dialoghi diArcheologia 3: 25-33.

1985b The Acconia Survey: Neolithic Settlement andthe Obsidian Trade. Institute of ArchaeologyOccasional Publication 10. London:University of London, Institute ofArchaeology.

Ammerman, A.]., and W. Andrefsky, ]r.

1982 Reduction sequences and the exchange ofobsidian in Neolithic Calabria. [n] .E.Ericson and T.K. Earle (eds), Contexts forPrehistoric Exchange, 149-72. New York:Academic Press.

Ammerman, A.J., A. Cesana, e. Polg1ase, and M.Terrani

1990 Neutron activation analysis of obsidianfrom two neolithic sites in Italy. Journal ofArchaeological Science 17: 209-220.

Ammerman, A.]., and M.W. Feldman

1974 On the 'making' of an assemblage of stonetools. American Antiquity 39: 610-16.

1978 Replicated collection of site surfaces.American Antiquity 43: 734-40.

Ammerman, A.]., C. Matessl, and L.L. Cavalli­Sforza

1978 Some new approaches to the study of theobsidian trade in the Mediterranean andadjacent areas. In 1. Hodder (ed.), SpatialOrganisation of Culture, 179-96. London:Duckworth.

Ammerman, A.]., and e. Polglase

1993 The exchange of obsidian at Neolithic sitesin Italy. [n F. Healy and e. Scarre (eds),Trade and Exchange in European Prehistory.Oxbow Monograph 33: 101-107. Oxford:Oxbow Books.

1996 Analyses and descriptions of the obsidiancollections from Arene Candide.Quaternaria Nova. In press.

Antoniazzi, A., G. Bermond Montanari, G.Giusberti, M. Massi Pasi, D. Mengoli, G. Morico,and L. Prati

1990 Lo scavo preistorico a Fornace Cappucciniin Faenza. In Archeologia a Faenza, Ricerchee Scavi dal Nealitico al Rmascimento, 23-59.Bologna: Nuova A[b Editoriale.

Arias, C., C. Bernardes, G. Bigazzi, F.r. Bonadonna,M.F. Cesar, ].e. Hadler Neto, C.M.G. Lattes, ].X.Oliveira, A.M. Osorio Araya, and G. Radi

1986 ldentlfica.;iio cia proveniencia de manu­fatmados de obsidiana atraves da dat<;aocom 0 metodo do tra<;o de fissao. Cincia eCulrura 38: 285-308.

Arias, C., G. Bigazzi, F.P. Bonadonna, M. Cipolloni,J.C. Hadler, C.M.G. Lattes, and G. RaJi

1984 Fission track dating in archeology. A usefulapplication. In P.L. Parrini (ed.), ScientificMethodologies AIJplied to Works af Arts, 151­59. Milan: Montedison Progetto Cultura.

Arias-Radi, G., G. Bigazzi, and F.P. Bonadonna

1972 Le tracce di fissione. Un metodo per 10studio delle vie di commerciodell'ossidiana. Origini. Preistoria e Protostoria

Obsidian Procurement and Distribution in the Central and Western Mediterranean 73

delle Civilta Antiche 6: 155-69.

Assorgia, A., L. Beccaluva, G. di Paola, L. Maccioni,G. Macdotta, M. Puxeddu, R. Santacroce, and G.Venturelli

1976 II complesso vulcanico di Monte Arci(Sardegna Centro--Occidentale). Notaillustrativa alia carta geopetrografica 1:50.000. Bollettilla della Societa GeologicaItaliana 95: 371-40l.

Atzeni, E.

1992 Reperti neolltiCI dall'Oristanese. In SardiniaAntiqua: Studl in Onore dl Piero Meloni inOccasione del Suo Seuantesimo Complea.nno,35-62. Cagllari: Edizioni della Torre.

Bagolini, B.

1980 Introduzione al Neolitico dell'ItaliaSettentrionale. Nel Quadro dell'Evaltizionedelle Prime Culture A[!ricole Europe.Pordenone: Grafiche Editoriali ArtistichePordenonesi.

Bailloud, G.

1969a Fouilles de Basi (Serra-dl-Ferro - Corse):Campagne 1968. Corse Historique 9: 49-64.

1969b Fouille d'un habitat neolithique et torn~en :l.Basi (Serra-di-Ferro - Corse): Premiersresultats. Bulletin de ia Societe PrehistoireFran~a1se 66: 367-84.

Barberi, F., S. Borsi, G. Ferrara, and F. Innocenti

1967 Contribute alia conoscenza vulcanologica emagmatologica delle isole dell'arcipelagoPontino. Memone della Societa GeologicaIwliana 6: 581-606.

Barfield, L.H.

1981 Patterns of north Italian trade 5000-2000b.c. In G. Barker and R. Hodges (eds),Archaeology and Italian Society: Prehistoric,Roman and Mcdlcml Studies, 27-51. BritishArchaeological Reports International Series102. Oxford: British ArchaeologicalReports.

1987 Recent work on sources of Italian flint. InG. de G. Sieveking and M.H. Newcomer(eds), The Human Uses of Flint and Chert:Proceedings of the Fourth International FlintSymposium, Held at Brighton Polytechnic,lO­15 Apn11983, 231-39. Cambridge:Cambridge UniversitY Press.

Barker, G.

1981 Landsc{!.pe and Soclet"j. Prehiswric CentralItaly. London: Academic Press.

Barnett, W.K,

1990a Small-scale transport of early Neolithicpottery in the west Mediterranean.Antiquity 64: 859-65.

1990b Production and distribution of earlypottery in the west Mediterranean. In W.Kingery (ed.), The Changing Roles ofCeramics in Society: 26,000 B.P. to the Present,1.37-57. Westerville, Ohio: The AmericanCeramic Society.

1995 Putting the pot before the horse: earliestceramics and the Neolithic transition in thewestern Mediterranean. [n W.K. Barnettand J.W. Hoopes (eds), The Emergence ofPottery: Technology and Innovation in AncientSocieties, 79--88. Washington, DC:Smithsonian Institution Press.

Baugh, T.G., and J.E. Ericson

1992 Trade and exchange in a historicalperspective. [n J.E. Ericson and T. Baugh(cds), The American Southwest andMesoamerica: Systems of Prehistoric Exchange,3-20. New York: Plenum.

Beccaluva, L., M. Deriu, L. Maccioni, G. Macciotta,and G. Venturelli

1974 II massiccio vulcanico di M.te Arci(Sardegna centro-occidentale): Notapreliminare. Rendiconti della Soc leta Italianadi Mineralogia e Perrologla 30: 1069-80.

Beccall1va, L., L. Maccioni, G. Macciotta, and G.Venturelli

1974 Dati geo]oglci e pctrochimici suI massicciovl1lcanico di Monte Arci (Sardegna).Memorie della Socleta Geologica Italiana 13:387-405.

Beck, C., and S. Shennan

1991 Amber in Prehistoric Britain. OxbowMonograph 8. Oxford: Oxbow Books.

BelluominI, G., and A. Taddeucci

1971 Studi suIle ossidiane itaHane, III: Elementiminori. Perlodico dl Mineralogia 40: 11- 39.

Biagi, P., R. Maggi and R. Nisbet

1989 Liguria: 9,000 5,000 BC. In C. Bonsall (ed.),The Mesolithic In Europe: Papers Presented atthe Third International SympOSium, Edinburgh1985,533-40. Edinburgh: John Donald.

Bietti Sestieri, A.M.

1985 Organizzazione sociale e modeIlimatematici: l'archeo1ogia di Colin Renfrew.Dialoghi di Archeolagia 3: 109-IIS.

74 Tylwt

Bigazzi, G., and F. Bonadonna

1973 Fission track dating of the obsidian ofLipari Island (Italy). Nature 242: 322-23.

Bigazzi, G., F.P. Bonadonna, and G. Radi

1982 Fission track dating of obsidians andprehistory. Paper read at the FifthInternational Conference onGeochronology, Cosmochronology,Isotope Geology, June 27-July 2,1982,Nikko National Park, Japan.

Bigani, G., S. Meloni, M. Oddone, and G. Radi

1986 Provenance studIes of obsidian artefacts~

trace elements and data reduction. Journalof Radwanalytical and Nuclear Chemistry,Articles 98: 353-63.

1992a Nuovi dati sulla diffusione dell'ossidiananegli insediamenn preistorici Italiani. In E.Herring, R. Whitehouse, and). Wilkins(eds), Papers of the Fourth Conference ofItalian Archaeology 3: New Developments inItalian Archaeology, Part 1,9-18. London:Accordia Research Centre.

1992b Study on the diffusion of Italian obsidian inthe neolithic settlements. Atti del VI/lConvegno Nazionale sulla Attivit{z di Ricercanii Settori della Radwchimlca e delia ChimicaNucleare, delle Radiazioni e deiRadIoelement!, Torino, 16-19 Giugno 1992,243-47. Torino: C.N.R. and Universitadegli Studl di Torino.

Bigazzi, G., and G. Radi

1981 Datazione con Ie tracce di fissione perl'identificazione dena proveniema deimanufatti di ossidiana. Rivista di ScienzePreistoriche 36: 223-50.

Binder, D.

1987 Le Neolithlque Ancien PTOven~al: Typologie etTechnologie des Outil/ages Uthiques. GalliaPrehistoire, Supplement 24. Paris: C.N.R.S.

1989 Aspects de la neolithisation dans les airespadane, proven.;ale et Iigure. In O.Aurenche and). Cauvin (eds),Neolithisations, 199-225. BritishArchaeological Reports International Series516. Oxford: BritIsh ArchaeologicalReports.

Binder, D., and). Courtin

1994 Un point sllr la circulation de l'obsidiennedans Ie domaine Proven.;al. GalliaPrehistoire 36: 310-22.

Binford, L.R.

1979 Organization and formation processes:looking at curated technologies. Journal ofAnthropological Research 35: 255-73.

Biro, K.T., l. Pozsgai, and A. Vlader

1986 Electron beam microanalyses of obsidiansamples from geological and archaeologicalsites. Acta Archaeologica AcademiaeScientiarum Hungaricae 38: 257-78.

Bloedow, E.F.

1987 Aspects of ancient trade in theMediterranean: obsidian. Studi Micenei edEgeo-Anatolici 26: 59-124.

Bradley, R.

1971 Trade competition and artefactdistribution. World Archaeology 2: 347-51.

Brandaglia, M.

1985 11 Neolirico a ceramica impressa deII'lsoladel Giglio: L'industria litica. Studi perl'ecologic.! del QuateTllano 7: 53-76.

Bryan, K.

1950 Flint Quames - the Sources of Tools and, atthe Same Time, the Factories of the AmericanIndian. Peabody Museum Papers 17.Cambridge: Peabody Museum, HarvardUniversity.

Buchner, G.

1949 Ricerche sui giacimenti e sulle industrieossidianiche in Italia. Rivista di ScietlzePreistoriche 4: 162-86.

Bulgarelli, G.M., L. d'Erme, E. Pellegrini, and P.Petitti

1993 L'insediamento preistorico di PoggioOlivastro (Canino~VT):Considerazioni eprospettive. Bullettino di PaletnologiaIta/iana 84: 435-80.

Camps, G.

1964 Notes de protohisroire Nord-Africaine.Libyca 12: 293-297.

1974 Les Civilisations Prehistoriques de l'Afrique duNord et du Sahara. Paris: C.N.R.S.

1976a La navigation en France au Neolithique et a.I'age de Bronze. Archeologia Corsa I: 113­30.

1976b La question des navigations prehistoriquesdans Ie bassin occidental de laMediterranee. Comptes Rendl-ls, CongresPrehistonque de France 20: 53~65.

1988 PTehlstoire d'une lie: Les Ongines de [a Corse.

Obsidian Procurement and Distribution in the Central and Western MedIterranean 75

Paris: Errance.

Cann, ).R., and C. Renfrew

1964 The characterization of obsidian and itsapplication to the Mediterranean region.Proceedings of the PrehistorIC Society 3D: 111­33.

Champion, T.C.

1989 Introduction. In T.C. Champion (ed.),Centre and Periphery: Comparative Studies inArchaeology, 1-21. London: Unwin Hyman.

Chapman, ].C.

1988 Ceramic production and socialdifferentiation: the Dalmatian neolithic andthe western Mediterranean. Journal ofMediterranean Archaeology 1: 3-25.

Chappell, S.

1986 Alternative sources in regional exchangesystems: a gravity model approach.Proceedmgs of the Prehistoric Society 52: 131­42.

Clark, ).GD.

1952 Prehistoric Europe: The Economic Basis.London: Methuen.

Clark, J,R.

1979 Measuring the flow of goods witharchaeological data. Ecotwmic Geography 55:1-17.

Cocchi Genick, D., and F. Sammartino

1983 L'ossidiana utilizzara nelle industriepreistoriche dellivornese. Quaderni delMuseo di Storia Naturale di Livorno 4: 151­61.

Contu, E.

1990-91 L'ossidiana e 1a selce della Sardegna e laloro diffusione. Ongini 15: 241-53.

Cornaggia Castiglioni, 0., F. Fussi, and M. D'Agnolo

1962 Indagini sulla provenienza dell'ossidiana inuso nelle industrie preistoriche italiane. Auidella Societa Italzana dl SClenze Naturali e delMuseo Civico di Stona Naturale in Milano101: 12-19.

1963 Indagini sulla provenienza dell'ossidianautilizzata nelle industrie preistoriche delMediterraneo occidentale. Aui della SocletaItaliana di SCienze Naturali e del MuseoCWICO dl Storia Naturale in Milano 102: 310­22.

Cortese, M., G. Frazzetta, and L. La Volpe

1986 Volcanic history of Lipari (Aeolian Islands,Italy) during the last 10,000 years. Journal ofVolcanology and Geothermal Research 27:117-33.

Costantini, L.

1989 Plant exploitation at Grotta dell'Uzzo,Sicily: new evidence for the transition fromMesolithic to Neolithic subsistence insouthern Europe. In D.R. Harris and G.C.Hillman (eds), Foraging and Farming. TheEvolution of Plant Exploitation, 197-206. OneWorld Archaeology 13. London: UnwinHyman.

Courtin, ).

1967 Le probleme de l'obsidienne dans IeNeolithique du Midi de la France. Rivista diStudi Liguri 33: 93-109.

1989 Recherches sur l'obsidienne dans Ieneolithique du Midi de la France. In M.Ricq-de Bouard, Rapport Scientifique ATPArcheologie Metropolitaine. Origine etDiffusion des Differentes Matieres Premieresdans La France Mediterraneene et ses abords auNeolithique, 26-28. Valbonne: C.N.R.S.

Crisci, G.M., M. Ricq-de Bouard, U. Lamaframe,and A.M. de Francesco

1994 Nouvelle methode d'analyse et provenancede l'ensemble des obsidiennes neolithiquesdu Midi de la France. Gallia Prehistoire 36:299-309.

Crummett, ).G., and S.E. Warren

1985 Appendix I. Chemical analysis ofCalabrianobsidian. In A.). Ammerman, The AcconiaSurvey: Neolithic Settlement and the ObsidianTrade, 107-114. Institute of ArchaeologyOccasional Publication 10. London:University of London, Institute ofArchaeology.

Dalton, G.

1975 Polanyi's analysis of long-distance tradeand his wider paradigm. In). Sabloff andC.C. Lamberg-Karlovsky (eds), AncientCivilization and Trade, 63-132.Albuquerque: University of New MexicoPress.

1977 Aboriginal economies in stateless societies.In T.K. Earle and ).E. Ericson (eds),Exchange Systems in Prehistory, 191-212. NewYork: Academic Press.

76 TyJ<ot

della Marmora, A. Ferrero

1839-40 Voyage en Sardaigne. Ou DescriptionStatistique, Physlque et Politlque de cette lie,avec des Recherches sur ses ProductionsNarurelles et ses AntiqUites. Z volumes. Paris:A. Bertrand; Turin:]. Bocca.

di Lernia, S., G. Fiorentino, and A. Galiberti

1990-91 'Gar~ano prehistoric flmt mines project':the state of research in the neolithic mine ofDefensola - Vieste (Italy). OngiTll 15: 175­98.

Donahue, R.E.1992 Desperately seeking Ceres; a critical

examination of current models for thetransition to agriculture in MediterraneanEurope. In A.B. Gebauer and T.D. Price(eds), Transitions to Agriculture III Prehistory,73-80. Monographs in World Archaeology4. Madison, WI: Prehistory Press.

Ducci, S., and P. Perazzi

1991 Notiziario, Parte 1: Isola di Pianosa - LaScola. Studi e Matenali. SClenzadell'Antichita in Toscana 6: 308-309.

Dyson, S.L., L. Gallin, M. Klimkiewicz, R.].Rowland, Jr., and C.M. Stevenson

1990 Notes on some obsidian hydration dates inSardinia. Quaderni della SoprintendenzaArcheologica per Ie Provl1lce di Cagliari eOmtano 7: 25·42.

Ericson, J .E.

1981 Exchange and Production Systems inCalifornia Prehistory. British ArchaeologicalReports International Series 110. Oxford:British Archaeological Reports.

1984 Toward the analysis of lithic productionsystems. In J.E. Ericson and B.A. Purdy(eds), Prehistonc Quarries and LithicProduction, 1-9. Cambridge: CambridgeUniversity Press.

Ericson, J.E., and T.G. Baugh

1994 Systematics of the study of prehistoricregional exchange in North America. InT.G. Baugh and J.E. Ericson (eds),Prehistoric Exchange Systems III NorthAmerica, 3- 15. New York: Plenum Press.

Feblot-Augustins, ]., and C. Perles

1992 Perspectives ethno-archeologiques sur lesechanges longue distance. In F. Audollze,A. Gallay, and V. Roux (eds),Ethnoarcheologie, Justification, Problemes et

Limites, 195-209. ]uan-Ies-Pins: APDCA.

Ferrari, A., and A. Pessina

1994 Le rocce lltilizzate per la fabbricazione dimanufatti in pietra scheggiata in Friuli.Primi dati. Attl della XXIX RlUllioneScientifica, 'Preistoria e Protostoria del Fnuii­Venezia Qiulia e dell'Istria, ' 28-30 settembre1990,129-37. Firenze: Istituto Italiano diPreistoria e Protostoria.

Foerstner, H.

1881 Nota prdiminare sulla geologia dell'isola diPantelleria secondo gli stlldi fatti negli anni1874 e 1881. Bollettino del R. ComitatoQeologleo d'Italia 12: 532-56.

Foschi Nieddll, A.

1982 11 neolitico antico della Grotta Sa Korona diMonte Majore (Thiesi, Sassari): notapreliminare. In R. Montjardin (ed.), LeNeolithique Ancien Med/terraneen: Actes duC(Jlloque International de Prehistoire,Montpellier 1981. Archeologie enLanguedoc No. Special 1982: 339-47. Sete:federation Archeologique de I'Herault.

1987 La Grotta di Sa Korona di Monte Majore(Thiesi, Sassari). Primi risultati dello scavo1980. Am della XXVI Riu7llone Scientifica 'IINeolitico in ltalia,' Firenze, 7-10 novembre1985, Volume II: 859-70. Firenze: IstitlltoItaliano di Preistoria e Protostoria.

Francaviglia, V.M.

1986 Characterization of Mediterranean obsidiansources by classical petrochemical methods.Preistoria Alpina 20: 311-32.

1988 Ancient obsidian sources on Pantelleria(Italy). Journal of Archaeological Science 15:109-122.

Francaviglia, V" and M. Piperno

1987 La repartition et la provenance deI'obsidienne archeologique de la Grottadell'Uzzo et de Monte Cofano (Sicile).Revue d'Archeometrie 11: 31-39.

Garibaldi, P.

1993 L'ossidiana di Monte Arci. In Sardegna:Civilta di un'lsola Mediterranea, 30. Bologna:Nuova Alfa Editoriale.

Gillot, P.-Y., and Y. Cornette

1986 The Cassignol technique for potassium­argon dating, precision and accuracy:examples from the Late Pleistocene torecent volcanics from southern Italy.

Obszdwn Procurement and Distribution in the Central and Western Mediterranean 77

Chemica! Cjeology (Isotope Cjeoscience Section)59: 205-222.

Gould, R.A., D.A. Koster, and A.H.L Sontz

1971 The lithic assemblage of the WesternDesert aborigines of Australia. AmericanAntiquity 36: 149--69.

Guidi, A.

1987 The development of prehistoricarchaeology in Italy: a short review. ActaArchaeologica 58: 237-47.

Guilaine, J.

1992 The Megalithic in Sardinia, southernFrance and Catalonia. In R.H. Tykot andT.K. Andrews (eds), Sardinia in theMediterranean: A Footprint in the Sea. StudiesIn Sardinian Archaeology Presented to MiriamS. Balmuth, 128-36. Monographs inMediterranean Archaeology 3. Sheffield:Sheffield Academic Press.

Guilaine, J., and J. Vaquer

1994 Les obsidiennes al'ouest du Rhone. Gal!iaPrehistoire 36: 323~27.

Hallam, B.R., S.E. Warren, and C. Renfrew

1976 Obsidian in the western Mediterranean:characterisation by neutron activationanalysis and optical emission spectroscopy.Proceedings of the Prehistoric Society 42: 85­110.

Helms, M.W.

1988 Ulysses' Sail: An Ethnographic Odyssey ofPower, Knowledge, and GeographicalDistance. Princeton: Princeton UniversityPress.

Herold, G.

1986 M ineralogische, Chemische undPhysikalische Untersuchungen an denObsidianen Sardiniens und Palmarolas.Grundlagen zur RekonstruktionPrehistorischer Handelswege imMittelmeerraum. Unpublished PhDdissertation, Universitat (TH) FridericianaKarlsruhe.

Hodder, I.

1974 Regression analysis of some trade andmarketing patterns. World Archaeology 6:172-89.

1978 Some effects of distance on patterns ofhuman interaction. In I. Hodder (ed.), TheSpatial Organisation of Culture, 155-78.London: Duckworth.

1980 Trade and exchange: definitions,identification and function. [n R.E. Fry(ed.), Models and Methods in RegionalExchange, 151-56. SAA Paper 1.Washington, DC: Society for AmericanArchaeology.

Hodder, 1., and C. Orton

1976 Spatial Analysis in Archaeology. Cambridge:Cambridge University Press.

Hurcombe, L.

1992a New contributions to the study of thefunction of Sardinian obsidian artefacts. InR.H. Tykot and T.K. Andrews (eds),Sardinia in the Mediterranean: A Footprint inthe Sea. Studies in Sardinian ArchaeologyPresented to Minam S. Balmuth, 83-97.Monographs in MediterraneanArchaeology 3. Sheffield: SheffieldAcademic Press.

1992b Use Wear Analysis and Obsidian: Theory,Experiments and Results. SheffieldArchaeological Monographs 4. Sheffield:John Collis.

Hurcombe, L.H., and P. Phillips

n.d. Chronology and artefact function at theGrotta Filiestru. In M.S. Balmuth and R.H.Tykot (eds), Sardinian Stratlgraphy andMediterranean Chronology. Proceedl1lgs of theInternational Conference, Tufts University,Medford, Massachusetts, March 1995.

]oussaumc, R.

1985 Les Dolmens pour les Morts: Les megalithismesatravers le monde. Paris: Hachette.

Keller, ].

1967 Alter und Abfolge der vulkanischenEreignisse auf den Neolischen Inseln(Sizilien). Benchte NaturforschendeGesdlschaft Freiburg 57: 33 -67.

1970 Datierung der Obsidiane und Bimstuffe vonLipari. Neues Jahrbuch fur Geologie undPaldontologie, MonatshefteI: 90-[01.

Kilikoglou, V., Y. Bassiakos, K. Souvatzis, and A.P.Grimanis

1996 Carpathian obsidian in Macedonia, Greece.[n G. Summers and F. Summers (cds),Archaeometry '94. Proceedings of the 29thInternational Symposium on Archaeometry,May 9-14, Ankara, Turkey. Ankara: TheScientific and Technical Research Councilof Turkey. In press.

78 Tykot

Knapp, A.B.

1985 Production and exchange in the Aegean andEastern Mediterranean: an overview. InA.B. Knapp and T. Stech (eds), PrehistoricProductwn and Exchange: The Aegean andEastern Mediterranean. UCLA Institute ofArchaeology Monograph 25: I-II. LosAngeles: Institute of Archaeology, UCLA.

Lanfranchi, F. de

1987 Le neolithique de Curacchiaghiu. PositionchronoIogique et culture materielle. Sonimportance dans I'ensemble corso~sarde. InJ. Guilaine, J. Courtin, J.-L. Roudil and J.­L. Vernet (eds), Premieres CommunautePaysannes en Mediterranee Occidentale. Actesdu Colloque International du C.N.R.S.,Montpellier, 26-29 aVril 1983, 433~ 42.Paris: C.N.R.S.

1993 Le Neolithique ancien mediterraneen de laCorse. CorSIca Antica (juillet 1993): 2~9.

Lefevre, ].~c., and P.~Y. Gillot

1994 Datation potassium-argon de rochesvolcaniques du Pleistocene superieur et deI'Holocene: exemple de l'italie du sud;application a l'archeologie. Bulletin de laSociete Prehistorique Fran"aise 91: 145~48.

Leighton, R.

1992 Stone axes and exchange in south Italianprehistory: new evidence from oldcollections. Accordia Research Papers 3: 11­40.

Lewthwaite, ].

1981 Plain tails from the hills: transhumance inMediterranean Archaeology. In A.Sheridan and G. Bailey (eds), EconomicArchaeology, 57-66. British ArchaeologicalReports International Series 96. Oxford:British Archaeological Reports.

1983 The Neolithic of Corsica. [n C. Scarre (ed.),Ancient France: Neolithic Societies and theirLandscapes 6000-2000 be, 146-83.Edinburgh: Edinburgh University Press.

1986a The transition to food production: aMediterranean perspective. In M. Zvelebil(ed.), Hunters in Transition: MesolithicSOCleties of Temperate Eurasia and TheirTransitwn w Farmmg, 53-66. Cambridge:Cambridge University Press.

1986b From Menton to the Mondego in threesteps: application of the availability modelto the transition to food production in

Occitania, Mediterranean Spain andsouthern Portugal. ArqueoLogia 13: 95~119.

1989 Isolating the residuals: the Mesolithic basisof man-animal relationships on theMediterranean islands. In C. Bonsall (ed.),The Mesolithic In E"urope: Papers Presented atthe Third International Symposium, Edinburgh1985, 54I~55. Edinburgh: John Donald.

Lilliu, G.

1959 Ricerche sull'arcipelago della Maddalena:l'arcipelago nella preistoria e nell'antichitaclassica. Memorie della Societa GeograficaItaliana 25.

1986 Le miniere dalla preistoria all'eta tardo­romana. In F. Mannoni (ed.), Le Miniere e IMinawri della Sardegna, 7-18. Cagliari:ConSiglio Regionale della Sardegna.

1989 La Sardegna preistorica e Ie sue relazioniesterne. Studi Sardi 27: 11-36.

Loria, D., and D.H. Trump

1978 Le scoperte a 'Sa 'Ucca de su Tintirriolu' eil neolitico sardo. Monument! Antichi deiLincei 49: 117-253.

MacDonald, R., and D.K. Bailey

1973 Chapter N. Chemistry of Igneous Rocks,Part I: The chemistry of the peralkalineover-saturated obsidians. In M. Fleischer(technical ed,), Data of Geochemistry, 6thedition. U.S. Geological SurveyProfessional Paper 440-N- I. Washington,D.C.: U.S. Government Printing Office.

Mackey, M., and S.E. Warren

1983 The identification of obsidian sources inthe Monte Arci region of Sardinia. In A.Aspinall and S.E. Warren (eds), Proceedingsof the 22nd Symposium on Archaeometry,University of Bradford, Bradford, U.K. March30th - April3rd 1982,420-31. Bradford:University of Bradford.

Magdeleine, J.

1973 Le gisement prehistorique du MonteGrosso a Biguglia (resultats des premierssondages). Bulletin de la Societe des SciencesHiswriques et Naturelles de La Corse 608-609:9-26.

1991 Une deuxieme sepulture pre-neolithique deCorse. Bulletin de la Societe PrehlstoriqueFran(aise 88: 80.

Obsidian Procurement and Distribution In the Central and Western Mediterranean 79

Magdeleine, J., and ].-c. Ottaviani

1986 L'abri prehistorique de Strette. Bulletin desSCiences Historiques et Naturelles de la Corse650: 81-90.

Malone, C.

1985 Pots, prestige and ritual in Neolithicsouthern Italy. In C. Malone and S.Stoddart (eds), Papers in Italian Archaeology4: The CambrJdge Conference, 2: Prehistory,118-5 I. British Archaeological ReportsInternational Series 244. Oxford: BritishArchaeological Reports.

Martinelli, M.C.

1990 Industrie litiche di alcuni siti neolitici dellaDalmazia e loro raffronti concontemporanee industrie dell'Italia sud­Adriatica e delle Isole Eolie. Rassegna diArcheologia 9: 125-52.

Mello, E.

1983 Indagini scientifiche per I'individllazionedella provenienza dei manllfatti diossidiana. In S. Tine, Passo dl Corvo e lacivilta neotitica del Tavoliere, 122-24.Genova: Sagep.

Michels, J., E. Atzeni, I.S.T. Tsong, and G.A. Smith

1984 Obsidian hydration dating in Sardinia. InM.S. Balmuth and R.J. Rowland, Jr. (eds),Studies in Sardinian Archaeology, 83-113.Ann Arbor: University of Michigan.

Montanari, G. Bermond, 1\1. Massi Pasi, and D.Mengoli

1994 L'insediamento neolitico di FornaceCappuccini di Faenza (Ravenna). Campagnadi scavo 1990. Preistoria Alpma 27: 173~95.

Pedes, C.

1987 Les Industries Lithiques Taillees de Franchthi(Argolide, Grece) I: Presentation Generale et!ndustrJes Paleolithiques. Bloomington:Indiana University Press.

1990 Les Industries Lithiques Taillees de Franchthi(Argolide, Grece) I1:Les Industries duMesolithique et du Neo!ithique Initial.Bloomington: Indiana University Press.

1992 Systems of exchange and organization ofproduction in neolithic Greece. Journal ofMediterranean Archaeology 5: 115-64.

Phillips, P.

1975 Early Farmers of West Mediterranean Europe.London: Hutchinson.

1982 The Middle Neolithic in Southern France.

Chasseen Farmmg and Culture Process. BritishArchaeological Reports International Series142. Oxford: British ArchaeologicalReports.

1986 Sardinian obsidian and neolithic exchangein the west Mediterranean. In M.S.Balmuth (ed.), Studies in SardinianArchaeology 11,203-209. Ann Arbor:University of Michigan Press.

1992 Obsidian distribution and the Europeanneolithic. In R.H. Tykot and T.K. Andrews(eds), Sardinia in the.Mediterranean: AFootpTlnt in (he Sea. Studies in SardinianArchaeology Presented to Miriam S. Balmuth,71-82. Monographs in MediterraneanArchaeology 3. Sheffield: SheffieldAcademic Press.

Phillips, P., A. Aspinall, and S. Feather

1977 Stages of 'Neolithisation' in SouthernFrance. Proceedings of the Prehistoric Society43: 303-316.

Pichler, H.

1967 Neue Erkentnisse uber Art und Genese desVulkanismus der Neolischen lnseln(Sizilien). Geologische Rundschau 57: 102­126.

1980 The island of Lipari. Rendiconti della Societaltaliana della M ineralogia e Petrologia 36:415-40.

Polanyi, K.

1957a Aristotle discovers the economy. In K.Polanyi, C.M. Arensberg, and H.W.Pearson (eds), Trade and Market in the EarlyEmpires, 64-94. New Yark: Academic Press.

1957b The economy as instituted process. In K.Polanyi, C.M. Arensberg, and H.W.Pearson (eds), Trade and Market in the EarlyEmpires, 243-70. New York: AcademicPress.

Polglase, C.

1989 Competing sources, resource availabilityand utilization at the end of long-distanceobsidian exchange routes. Paper read at the54th Annual Meeting of the Society forAmerican Archaeology, Atlanta, Georgia.

1990 Appendice I. In M. Bernabo Brea, A.Ghiretti, C. Polglase and V. Visconti, I sitineolitici [ungo il tarrente Cinghio (Parma).Preistoria Alpina 24: 140-55.

Pollmann, H.-O.

1993 Obsidjan im nordwestmediterranen Raum.

80 Tykot

Seine Verbreitung und Nutzung imNeolithikum und Eneolithlkum. BritishArchaeological Reports International Series585. Oxford: Tempus Reparatum.

Puglisi, S.M.

1941-42 V illaggi sotto roccia e sepolcri megaliticinella Gallura. Rullettino di Paleonto[ogiaIraliana 5-6: 123-41.

Puxeddu, C.

1958 Giacimenti di ossidiana del Monte Arci inSardegna e sua irradiazione. Studi Sardi 14­I5(Parte I): 10-66.

Randle, K., L.H. Barfield, and B. Bagolini

1993 Recent Italian obsidian analyses. Journal ofArchaeological Science 20: 503-509.

Rasson, ].A., J.E. Ericson, and T.D. D'Altroy

1977 Trace elements In obsidian from neolithicBosnia: a scarce resource and marginality.Paper presented at the Symposium onArchaeometry and ArchaeologicalProspection, University of Pennsylvaniaand the University Museum, Philadelphia.

Renfrew, C.

1969 Trade and culture process in Europeanprehistory. Current Anthropology 10: 151­60.

1972 The Emergence of Civilisation. London:Methuen.

1975 Trade as action at a distance. In). Sabloffand C.C. Lamberg-Karlovsky (eds), AncientCivilization and Trade, 3-60. Albuquerque:University of New Mexico Press.

1977 Alternative models for exchange and spatialdistribution. In T.K. Earle and J.E. Ericson(eds), Exchange Systems in Prehistory, 71-90.New York: Academic Press.

1986 Introduction: Peer polity interaction andsocio-political change. In C. Renfrew and] .F. Cherry (eds), Peer Polity Interaction andSocio-,political Change, 1-18. Cambridge:Cambridge University Press.

1993 Trade beyond the material. In C. Scarre andF. Healy (eds), Trade and Exchange inPrehistoric Europe, 5-16. OxbowMonograph 33. Oxford: Oxbow Books.

Renfrew, C., and A. Aspinall

1990 Aegean obsidian and Franchthi Cave. [n C.Pedes, Les Industries Lithiques Taillees deFranchthi (Argolide, Greee) II: Les Industriesdu Mesolithique et du Neolithique Initwl,

257-70. Bloomington: Indiana UniversityPress.

Ricq-de Bouard, M.

1993 Trade in neolith ic jadeite axes from theAlps: new data. In C. Scarre and F. Healy(eds), Trade and Exchange in PrehistoricEurope, 61-67. Oxbow Monograph 33.Oxford: Oxbow Books.

Ricq-de Bouard, M., R. Compagnoni, J. Desmons,and F.G. Fedele

1990 Les roches alpines dans I'outillage pobneolithique de la France mediterraneenne.Gallia Prehistoire 32: 125-49.

Ricq-de Bouard, M., and F.G. Fedele

1993 Neolithic rock resources across the westernAlps: circulation data and models.Geoarchaeology 8: 1-22.

Runnels, C.

1989 Trade models in the study of agriculturalorigins and dispersals. Journal ofMediterranean Archaeology 2: 149-56.

Runnels, c., and T.H. van Andel

1988 Trade and the origins of agriculture in theeastern Mediterranean. Journal ofMediterranean Archaeology 1: 83-109.

Sahlins, M.

1972 Stone Age Economics. Chicago: Aldine.

Sammartino, F.

1986 Le industrie litiche del neolitico e dell'eddei metalli a Coltano. In R. Mazzanti, R.Grifoni Cremonesi, M. Pasquinucci, andA.M. Pult Quaglia, Terre e Paduli. RepertjDocumenti Imaggmi per la Storia di Colrano,61. Pontedera: Bandecchi and Vivaldi.

Service, E.

1962 Primitive Social Organisation. New York:Random House.

Shennan, S.].

1982 The emergence of hierarchical structure. InC. Renfrew and S.]. Shennan (eds), Ranking,Resource and Exchange: Aspects of theArchaeology of Early European Society, 9-12.Cambridge: Cambridge University Press.

Sheridan, A.

1982 Archaeological study of exchange:theoretical problems and a new approach.Archaeological Review from Cambridge I: 4­12.

Obsidian Procurement and Distribution in the Central and Western Mediterranean 81

Shetratt, A.

1982 MobIle resources: settlement and exchangein early agricultural Europe. [n C. Renfrewand S.]. Shennan (eds), Ranking, Resourceand Exchange: Aspects of [he Archaeology ofEarly European Society, 13-26. Cambridge:Cambridge University Press.

Skeates, R.

1993 Neolithic exchange in central and southernItaly. [n C. Scarre and F. Healy (eds), Tradeand Exchange in PrehlstoTic Europe:Proceedings of a Conference held at [helhllversityofBristol, April 1992,109114.Oxbow Monograph 33. Oxford: OxbowBooks.

Stevenson, C.M.

1989 The hydration dating of obsidian artefactsfrom Nuraghe Santa Barbara, Sardinia,Italy. Unpublished report on file with theauthor.

Stevenson, C.M., and ].G. Ellis

n.d. Dating Sardinian archaeological obsidian.[n M.S. Balmuth and R.H. Tykot (eds),Sardinian Stratigraphy and MediterraneanChronology. Proceedings of the InternationalConference, Tufts Uruversity, Medford,Mas.lachusetts, March 1995.

Tanda, G.

1980 Alcune considerazioni suI sito archeologicodi Molia-lllorai. Quaderlll Boloranesi 6: 63­77.

1984 Illoria. Loc. Molia. [n E. Anati (ed.), 1Sardi:Dal Paleolitico aWEul Romana, 296- 298.Cagliari: ]aca.

Tangri, D.

1989 On trade and assimilation in Europeanagricultural origins. Journal of MediterraneanArchaeology 2: 139-48.

Tine, S., and A. Traverso

1992 Monte d'Accoddi, 10 Anni di Nuovi ScaVI:Relazione Preliminare. Genova: IstitutoItaliano Archeologia Sperimentale.

Torrence, R.

1981a Obsidian in the Aegean: Towards aMethodology for the Study of PrehistoricExchange. PhD dissertation, University ofNew Mexico.

1981b Mining and working of obsidian on Melos.Der Anschnitt 33: 86-103.

1982 Obsidian quarries and their use. [n C.

Renfrew and ].M. Wagstaff (eds), An Is!andPolity: The Archaeo!ogy of Exploitation onMelos, 193-22 1. Cambridge: CambridgeUniversity Ptess.

1983 Time budgeting and hunter-gatherertechnology. [n G. Bailey (ed.), Hunter­Qatherer Economy in Prehistory, 11-22.Cambridge: Cambridge University Press.

1984 Monopoly or direct access? Industtialorganization at the Melos obsidian quarries.In ].E. Ericson and B.A. Purdy (eds),Prehistoric Quarries and Lithic Production,49-63. Cambridge: Cambridge UniversityPress.

1986 Production and Exchange of Stone Tools:PrehIstoric Obsidian III the Aegean.Cambridge: Cambridge University Press.

Trump, D.H.

1983 La Qrotra FUiestru a Bonu Ighinu, Mara (SS).Quademi della SoprintendenzaArcheologica per Ie Provincie di Sassari eNuoro 13. 5assari: Dessi.

Tykot, R.H.

1992 The sources and distribution of Sardinianobsidian. In R.H. Tykot and T.K. Andrews(cds), Sardlllla in the Mediterranean: AFootprint in the Sea. Studies in SardinianArchaeology Presented to MIriam S. Balmuth,57-70. Monographs in MediterraneanArchaeology 3. Sheffield: SheffieldAcademic Press.

1994 Radiocarbon dating and absolutechronology in Sardinia and Corsica. In R.Skeates and R. Whitehouse (eds),Radiocarbon Dating and Italian Prehistory,115-45. Accordia Specialist Studies on Italy3. London: Accordia Research Centre.

1995a Prehistoric Trade in the WesternMediterl'anean: The Sources andDistribution of Sardinian Obsidian. PhD.thesis, Harvard University. Ann Arbor:University Microfilms.

1995b Provenance of obsidian artefacts fromUstica. In R.R. Holloway and 5.S. Lukesh,UsUca I: The Results of the Excavations of theRegione Siciliana Soprinrendenza ai BeniCulturati ed Ambientali Provincia di Palermoin Collaboration with Brown University in1990 and 1991,51-53. ArchaeologiaTransatlantica 13. Louvain-la~Neuve:Universite de Louvain.

1997a Characterization of the Monte Arci

82 Tykot

(Sardinia) obsidian sources. Journal ofArchaeolog~calScience 24. [n press.

1997b Islands in the stream: stone age culturaldynamics in Sardinia and Corsica. In R.H.Tykot,). Morter and J.E. Robb (eds), SocialDynamics of the Prehistoric CentralMediterranean.

Tykot, R.H., and J.. D. Vigne

n.d. The provenance of archaeological obsidianfrom Zembra Island, Tunisia. In Pr. Chelbi(ed.), ile de Zemhra (ms 1993; Tunis).

Vaquer, J.1990 Le Neolithique en Languedoc Occidental.

Paris: C.N.R.S.

Vaughan, P.

1985 Use Wear Analysis of Flaked Stone Tools.Tucson: University of Arizona Press.

1990 Use-wear analysis of Mesolithic chipped­stone artefacts from Franchthi Cave. In C.Pedes, Les Industries Lithiques Taillres deFranehthi (Argolide, Greee) II: Les Industriesdu Mesolithique et du Neolithique Initial,239-53. Bloomington: Indiana UniversityPress.

Vigliardi, A.

1980 Rapporti tra Sardegna e Toscananell'Eneolitico finale-primo bromo: Lagrotta del Fontino nel grossetano. Aw dellaXXII Riunione Seientifica, 'Sardegna Centro­Settentriona!e, ' Ottobre 1978, 247~88.Fireme: Istituto Italiano di Preistoria eProtostoria.

Warren, S.E.

1981 Linear exchange mechanisms and obsidiantrade. Revue d'Areheometrie 5: 167-75.

Washington, H.

1913 Some lavas of Monte Arci, Sardinia.

American Journal of Science 36: 577-90.

Whitehouse, R.

1981 Megaliths of the central Mediterranean. In).D. Evans, B. Cunliffe, and C. Renfrew(eds), Antiquity and Man: Essays in Honour ofQlyn Damel, 106-127. London: Thames andHudson.

Williams~Thorpe,O.

1995 Obsidian in the Mediterranean and theNear East: a provenancing success story.Arehaeometry 37: 217-48.

Williams-Thorpe, 0., S.E. Warren, and L.H.Barfield

1979 The sources and distribution ofarchaeological obsidian in northern Italy.PmstoTia Alpina IS: 73-92.

Williams-Thorpe, 0., S.E. Warren, and]. Courtin

1984 The distribution and sources ofarchaeological obsidian from southernFrance. Journal of Archaeological Selenee 11:135-46.

Williams-Thorpe, 0., S.E. Warren, and ).G.Nandris

] 984 The distribution and provenance ofarchaeological obsidian in Central andEastern Europe. Journal of Archaeolog~calScience 11: 183~212.

1987 Characterization of obsidian sources andartefacts from central and eastern Europe,using instrumental neutron activationanalysis. In J. Korek (ed.), Proceedings of theInternational Conference on Lithic RawMaterial Used in Prehistory, Flint Mining andLithic Raw Material Characterization,Budapest, and Siimeg, 1986. 2: 271-79.Budapest.

Wilmsen, E.N.

1972 Social Exchange and Interaction. Museum ofAnthropology Paper 46. Ann Arbor:University of Michigan.