Palaeolithic mollusc exploitation at Riparo Mochi (Balzi Rossi, Italy): food and ornaments from the Aurignacian through Epigravettian.

STINER, MARY C.

Introduction

The early Upper Palaeolithic of Europe is distinguished by the sudden appearance of art, ornaments, and a quickened pace of technological diversification (Bietti 1997; Gamble 1986; Hahn 1972; Harrold 1989; Kozlowski 1990; Kuhn & Stiner 1998b; Leroi-Gourhan 1961; Mellars 1989; Otte 1990; Palma di Cesnola 1993; White 1982). This study explores an important branch of that process through the analysis of the marine molluscs used as ornaments and food during the Upper and Epi-Palaeolithic periods at Riparo (Abri) Mochi, on the northwestern Italian coast. The stratigraphic sequence spans 36,000 to about 9000 years before present (TABLE 1). The ornaments from this shelter differ from those of inland European sites in having been made almost exclusively from marine shells. Carved ivory and soft stone pendants and modified mammal teeth generally dominate Upper and Epi-Palaeolithic ornament assemblages of the continental interior (Hahn 1972; White 1989). Shell ornaments were emphasized in coastal areas and where marine fossil outcrops were available (e. g. Taborin 1993a).

TABLE 1. Age ranges for the Epi- through Upper Palaeolithic faunal assemblages from Riparo Mochi, Italy.

Palaeolithic associational
culture and layer direct dating dating

Late Epigravettian (layer A) -- 9000-12,000
Early Epigravettian (layer C) -- 17,000-19,000
Gravettian (layer D) -- 24,000-28,000
Middle Aurignacian (layer F) -- 27,000-32,000
Early Aurignacian (layer G) 32,000-36,000 --

Associational dating is based on nearby sites that contain similar industries dated by radiocarbon technique. (Sources: Bietti 1990; Hedges et al. 1994: 347; Palma di Cesnola 1993.)

The long Upper-Epi-Palaeolithic sequence of Riparo Mochi presents a rare opportunity to examine the influence of natural patterns of availability on humans' choices of food and raw material for ornament-making at one rich coastal source. The steep, rocky topography in this stretch of the Riviera lent considerable stability to the local marine environment, despite dramatic changes in sea level over the last 36,000 years (FIGURE 1). The shelter was never far from the Mediterranean shore, due to the high topographic relief of the Balzi Rossi coast. These biogeographic and geologic constants eliminate the most common causes of variation in archaeological comparisons of shell ornaments and shellfish exploitation -- biogeography and climate-driven reconfiguration of shorelines -- making it possible to test hypotheses about human-caused variation.

[Figure 1 ILLUSTRATION OMITTED]

The possibility that Palaeolithic ornaments, including those made from shells, signalled aspects of individual or group identity is frequently raised by archaeologists (Hahn 1972; Softer 1989; Taborin 1993b; White 1982; 1989). While Palaeolithic ornaments probably held considerable social significance, it is not clear just what we are seeing of early emblematic expression in piece-by-piece analyses of shells. Outside of decorated burials, disarray is the common condition of ornament assemblages, material fallen from use or forgotten. Upper Palaeolithic people's use of marine shells for ornament-making nonetheless can be examined for collector preferences. In cases where non-fossil species were emphasized, as at the Balzi Rossi sites, culturally imposed biases can be exposed by reference to natural standards of marine community structure and local species availability.

This presentation begins by introducing the full spectrum of marine mollusc species represented at Riparo Mochi. Next is a summary of assemblage formation histories, based on damage patterns and spatial associations of the shells to other archaeological materials across fine stratigraphic cuts. Four major categories of shell debris are recognized as a result -- food, ornaments, accidental inclusions with marine sponges, and co-resident species (land snails). The ornaments are examined for taxonomic and formal diversity among cultural phases, and for what shell and non-shell ornaments may reveal about Palaeolithic manufacturing effort, standardization and symbolling behaviour. Because the Early Aurignacian generally marks the florescence of ornament and art traditions in Europe (Hahn 1972; Leroi-Gourhan 1961; White 1989), possible evidence for an early `experimental' -- and thus more variable -- phase in ornament production is considered. For reasons that will become apparent, I also compare the rates at which various categories of shell were deposited relative to ungulate remains and lithic artefacts. The concluding discussion relates the results for Riparo Mochi to those for nearby and more distant sites of the same age.

Archaeological background

Riparo Mochi is one of a series of contiguous rock-shelters and caves in the base of the limestone Balzi Rossi (Baousse-Rousse, or Red Cliffs), between the towns of Menton and Ventimiglia on the Mediterranean Riviera, in the Italian province of Liguria (FIGURE 1). Known as the Grimaldi caves, this complex of seaside shelters has yielded rich Palaeolithic cultural remains and faunas dating to the Middle and Late Pleistocene through late Holocene geologic periods (Barral & Simone 1969; Leonardi 1935; Villeneuve et al. 1906-19). The tool industries and portable artwork of the Grimaldi sites are extraordinary for their abundance and styles (e.g. Bisson & Bolduc 1994; Kuhn & Stiner 1998a; White & Bisson 1998).

Riparo Mochi is a shallow cavity tucked between Grotta del Caviglione and Grotta di Florestano in the Balzi Rossi (Blanc 1953; Cardini & Biddittu 1967: 361-5; Laplace 1977; Lumley-Woodyear 1969; Renault-Miskovsky 1972). Work began there with modest testing in 1938, followed in 1941, 1942, 1949 and 1959 by large-scale excavations led by L. Cardini and A.C. Blanc. New research on the Mochi collections by the author(1) and others began in 1991 (Alhaique et al. 1997; Kuhn & Stiner 1992; 1998a; Stiner et al. 1999; in press).

Riparo Mochi stratigraphy

A.G. Segre identified nine thick stratigraphic units or `layers' in Riparo Mochi. Most of Layer A at the top of the series was removed in 1938 and appears to be early Holocene in age (c. 10-9000 BP). Layer B is nearly sterile, whereas Layer C contains an Epigravettian or Late Gravettian industry (compare Bietti 1997; Palma di Cesnola 1993). Layer D is a thick deposit containing abundant Gravettian artefacts and is considered discrete from the overlying stratum on the basis of sedimentary changes. Layer E is another near-sterile stratum, underlain by two strata containing early Upper Palaeolithic artefacts: Layer F contains a `Middle' Aurignacian industry with classic tool elements; Layer G contains an early or `proto-'Aurignacian industry, distinguished by abundant bladelets (Kuhn & Stiner 1998a; Laplace 1977; Palina di Cesnola 1993: 117-18). The Upper Palaeolithic layers are separated from 4 m of Middle Palaeolithic deposits (Layer I) by the nearly sterile Layer H. No human burials were found in Riparo Mochi, although at least 10 are known from later Upper and Epi-Palaeolithic deposits in other Balzi Rossi sites (Riviere 1887; Boule & Vallois 1957).

The quality of faunal preservation in Riparo Mochi is exceptional, due to favourable sediment chemistry and protection from the elements by the cliffs above. The possibility of stratigraphic mixing in the Riparo Mochi series is minimal, as most of the artefact-rich layers are bracketed by sterile/semi-sterile deposits and notable changes in sediment color and texture. The bedding planes in the Upper and Epi-Palaeolithic layers are nearly horizontal and were excavated in 10-cm cuts. One important advantage of the Mochi collections stems from Cardini's meticulous excavation style and his penchant for complete recovery of artefacts, faunal remains, and wood charcoal. All of the sediments were dry-sieved through fine mesh.

Mollusc classification and marine ecology

Because numerous quasi-independent nomenclatures have been applied to Mediterranean molluscs over the years, TABLE 2 presents both the taxonomic spectrum of and common synonyms for the marine molluscs of Riparo Mochi. All of the species continue to exist in the Lusitanian marine province (see also Leonardi 1935: 27), which includes the Mediterranean Sea, the Black Sea and the Atlantic Ocean off the North African and southern European coasts (Sabelli 1980: 45-7). TABLE 2 also provides information on adult shell size ranges, feeding habits, and substrate preferences by taxon, characteristics which are important to the analyses below. My identifications follow Tornaritis (1987) primarily, with synonyms (=) from Barnes (1994), Riedl (1970) and Sabelli (1980), among others.

TABLE 2. Molluscan taxa in the Upper and Epi-Palaeolithic cultural layers of Riparo Mochi, Italy.

 name
family genus species source

SCAPHOPODA (Class)

Dentaliidae Dentalium sp. --

GASTROPODA (Class)

ARCHAEOGASTROPODA
(Order)

Haliotidae Haliotis lamellosa Lamarck

Fissurellidae Fissurella sp. --

Patellidae Patella caerulea(*) L.
 Patella ferruginea(*) Gmelin
 Patella lusitanica Gmelin
 =rustica(*)

Trochidae Calliostoma sp. --
 Gibbula varia L.
 Gibbula adansoni Payr.
 Gibbula richardi Payr.
 Clanculus jussieui Payr.
 Clanculus cruciatus L.
 Clanculus corallinus Gmelin
 Jujubinus=Cantharidis Monter.
 corallinus
 Jujubinus=C. depictus Deshayes
 Jujubinus=C. exasperatus Pennant
 Jujubinus=C. striatus L.
 Monodonta=Gibbula articulata Lamarck
 Monodonta=G. mutabilis Philippi
 Monodonta turbinata(*) Born

Turbinidae Astrea=Turbo rugosa L.
 Homalopoma=Leptothyra L.
 =T. sanguineum

MESOGASTROPODA
(Order)

Littorinidae Littorina neritoides L.
 Littorina obtusata(1) L.
 Littorina rudis L.

Turritellidae Turritella communis Lamarck

Vermetidae Vermetus sp. --

Architectonicidae Architectonica=Solarium sp. --

Cerithiidae Cerithium rupestre Risso
 Cerithium tuberculatum Brug.
 Cerithium vulgatum Brug.

Cerithiopsidae Cerithiopsis tubercularis Montagu

Epitoniidae Epitonium=Scala commune Lamarck

Aporrhaidae Aporrhais L.
 =Chenopus pes-pelecani

Strombidae Strombus bubonius Lamarck

Eratoidae Trivia europaea=adriatica Montagu

Cypraeidae Cypraea=Talparia L.
 =Luria lurida

Naticidae Naticarius dillwyni Payr.
 Naticarius=Natica hebraea Martyn
 Naticarius=N. Payr.
 =Polynices guillemini
 Naticarius=N. Risso
 =Neverita josephina
 Naticarius=N. millepunctata Lamarck

Cassididae Phalium=Cassis undulatum Born

Cymatiidae Charonia nodifera Lamarck

NEOGASTROPODA
(Order)

Muricidae Ocinebrina=Tritonalia Payr.
 edwardsi
 Purpura lapillus L.

Pyrenidae= Columbella rustica L.
 Columbellidae Pyrene=Mitrella scripta L.

Buccinidae Buccinum undatum(1) L.
 Pisania maculosa=striata Lamarck

Nassariidae Cyclope=Nassa neritea L.
 Nassarius=Nassa L.
 =Arcularia gibbosula(1)
 Nassarius=N.=Hinia Muller
 incrassatus
 Nassarius=N.=H. costulata Renieri
 Sphaeronassa=Nassa mutabilis L.

Fasciolariidae Fusus sp. --

Mitridae Mitra sp. --

Conidae Conus mediterraneus Brug.
 =ventricosus

BIVALVIA or
PELECYPODA (Class)

FILIBRANCHIA
(Order)

Arcidae Striarca=Arca lactea L.

Glycymeridae Glycymeris --
 =Pectunculus sp.(*)

Mytilidae Mytilus galloprovincialis(*) Lamarck
 Lithophaga=Lithodomus L.
 lithophaga

Pectinidae Chlamys glabra L.
 Chlamys opercularis L.
 Chlamys varia L.
 Chlamys sp. --
 Pecten jacobaeus(*) L.
 Pecten maximus(*) L.

Ostreidae Ostrea edulus(*) L.

EULAMELIBRANCHIA
(Order)

Cardiidae Acanthocardia L.
 =Cardium tuberculatum(*)
 Cerastoderma=Cardium L.
 edule(1)(*)

Veneridae Callista=Meretrix L.
 =Pitaria chione(*)
 Chamelea=Venus L.
 =Chione gallina(*)

 common
family genus species name

SCAPHOPODA (Class)

Dentaliidae Dentalium sp. tusk shell

GASTROPODA (Class)

ARCHAEOGASTROPODA
(Order)

Haliotidae Haliotis lamellosa abalone,
 ormer
Fissurellidae Fissurella sp. keyhole
 limpet
Patellidae Patella caerulea(*) blue limpet
 Patella ferruginea(*) iron limpet
 Patella lusitanica lusitanian
 =rustica(*) limpet

Trochidae Calliostoma sp. painted top
 Gibbula varia top shell
 Gibbula adansoni top shell
 Gibbula richardi top shell
 Clanculus jussieui top shell
 Clanculus cruciatus top shell
 Clanculus corallinus top shell
 Jujubinus=Cantharidis top shell
 corallinus
 Jujubinus=C. depictus top shell
 Jujubinus=C. exasperatus top shell
 Jujubinus=C. striatus top shell
 Monodonta=Gibbula articulata top shell
 Monodonta=G. mutabilis top shell
 Monodonta turbinata(*) checkered top

Turbinidae Astrea=Turbo rugosa star shell
 Homalopoma=Leptothyra red turban
 =T. sanguineum

MESOGASTROPODA
(Order)

Littorinidae Littorina neritoides periwinkle
 Littorina obtusata(1) periwinkle
 Littorina rudis periwinkle

Turritellidae Turritella communis turrit shell

Vermetidae Vermetus sp. worm shell

Architectonicidae Architectonica=Solarium sp. sundial

Cerithiidae Cerithium rupestre horn shell
 Cerithium tuberculatum horn shell
 Cerithium vulgatum horn shell

Cerithiopsidae Cerithiopsis tubercularis small horn

Epitoniidae Epitonium=Scala commune wentletrap

Aporrhaidae Aporrhais pelecan foot
 =Chenopus pes-pelecani

Strombidae Strombus bubonius conch

Eratoidae Trivia europaea=adriatica false cowry

Cypraeidae Cypraea=Talparia cowry
 =Luria lurida

Naticidae Naticarius dillwyni moon snail
 Naticarius=Natica hebraea moon snail
 Naticarius=N. moon snail
 =Polynices guillemini
 Naticarius=N. moon snail
 =Neverita josephina
 Naticarius=N. millepunctata moon snail

Cassididae Phalium=Cassis undulatum helmet shell

Cymatiidae Charonia nodifera conch

NEOGASTROPODA
(Order)

Muricidae Ocinebrina=Tritonalia triton
 edwardsi
 Purpura lapillus violet snail

Pyrenidae= Columbella rustica dove shell
 Columbellidae Pyrene=Mitrella scripta dove shell

Buccinidae Buccinum undatum(1) common whelk
 Pisania maculosa=striata spotted
 pisania
Nassariidae Cyclope=Nassa neritea mud snail
 Nassarius=Nassa mud snail
 =Arcularia gibbosula(1)
 Nassarius=N.=Hinia mud snail
 incrassatus
 Nassarius=N.=H. costulata mud snail
 Sphaeronassa=Nassa mutabilis dog/basket
 whelk

Fasciolariidae Fusus sp. spindle shell

Mitridae Mitra sp. miter shell

Conidae Conus mediterraneus cone shell
 =ventricosus

BIVALVIA or
PELECYPODA (Class)

FILIBRANCHIA
(Order)

Arcidae Striarca=Arca lactea ark shell

Glycymeridae Glycymeris bittersweet
 =Pectunculus sp.(*)

Mytilidae Mytilus galloprovincialis(*) mussel
 Lithophaga=Lithodomus rock borer
 lithophaga

Pectinidae Chlamys glabra fan scallop
 Chlamys opercularis fan scallop
 Chlamys varia fan scallop
 Chlamys sp. fan scallop
 Pecten jacobaeus(*) scallop
 Pecten maximus(*) giant scallop

Ostreidae Ostrea edulus(*) flat oyster

EULAMELIBRANCHIA
(Order)

Cardiidae Acanthocardia cockle shell
 =Cardium tuberculatum(*)
 Cerastoderma=Cardium edible cockle
 edule(1)(*)

Veneridae Callista=Meretrix brown venus
 =Pitaria chione(*)
 Chamelea=Venus striped venus
 =Chione gallina(*)

family genus species diet

SCAPHOPODA (Class)

Dentaliidae Dentalium sp. C

GASTROPODA (Class)

ARCHAEOGASTROPODA
(Order)

Haliotidae Haliotis lamellosa HA

Fissurellidae Fissurella sp. HA

Patellidae Patella caerulea(*) HA
 Patella ferruginea(*) HA
 Patella lusitanica HA
 =rustica(*)

Trochidae Calliostoma sp. HD
 Gibbula varia HD
 Gibbula adansoni HD
 Gibbula richardi HD
 Clanculus jussieui HD
 Clanculus cruciatus HD
 Clanculus corallinus HD
 Jujubinus=Cantharidis HD
 corallinus
 Jujubinus=C. depictus HD
 Jujubinus=C. exasperatus HD
 Jujubinus=C. striatus HD
 Monodonta=Gibbula articulata HD
 Monodonta=G. mutabilis HD
 Monodonta turbinata(*) HD

Turbinidae Astrea=Turbo rugosa HA
 Homalopoma=Leptothyra HA
 =T. sanguineum

MESOGASTROPODA
(Order)

Littorinidae Littorina neritoides H
 Littorina obtusata(1) H
 Littorina rudis H

Turritellidae Turritella communis HD

Vermetidae Vermetus sp. --

Architectonicidae Architectonica=Solarium sp. --

Cerithiidae Cerithium rupestre HD
 Cerithium tuberculatum HD
 Cerithium vulgatum HD

Cerithiopsidae Cerithiopsis tubercularis HA

Epitoniidae Epitonium=Scala commune C-cor

Aporrhaidae Aporrhais H
 =Chenopus pes-pelecani

Strombidae Strombus bubonius HA

Eratoidae Trivia europaea=adriatica C-cor

Cypraeidae Cypraea=Talparia C-cor
 =Luria lurida

Naticidae Naticarius dillwyni C-biv
 Naticarius=Natica hebraea C
 Naticarius=N. C
 =Polynices guillemini
 Naticarius=N. C
 =Neverita josephina
 Naticarius=N. millepunctata C

Cassididae Phalium=Cassis undulatum C-urch

Cymatiidae Charonia nodifera C-var

NEOGASTROPODA
(Order)

Muricidae Ocinebrina=Tritonalia C
 edwardsi
 Purpura lapillus C

Pyrenidae= Columbella rustica O
 Columbellidae Pyrene=Mitrella scripta O

Buccinidae Buccinum undatum(1) C,SC
 Pisania maculosa=striata C,SC

Nassariidae Cyclope=Nassa neritea C,SC
 Nassarius=Nassa C,SC
 =Arcularia gibbosula(1)
 Nassarius=N.=Hinia C,SC
 incrassatus
 Nassarius=N.=H. costulata C,SC
 Sphaeronassa=Nassa mutabilis O

Fasciolariidae Fusus sp. C

Mitridae Mitra sp. SC,C

Conidae Conus mediterraneus C
 =ventricosus

BIVALVIA or
PELECYPODA (Class)

FILIBRANCHIA
(Order)

Arcidae Striarca=Arca lactea F

Glycymeridae Glycymeris F
 =Pectunculus sp.(*)

Mytilidae Mytilus galloprovincialis(*) F
 Lithophaga=Lithodomus F
 lithophaga

Pectinidae Chlamys glabra F
 Chlamys opercularis F
 Chlamys varia F
 Chlamys sp. F
 Pecten jacobaeus(*) F
 Pecten maximus(*) F

Ostreidae Ostrea edulus(*) F

EULAMELIBRANCHIA
(Order)

Cardiidae Acanthocardia F
 =Cardium tuberculatum(*)
 Cerastoderma=Cardium F
 edule(1)(*)

Veneridae Callista=Meretrix F
 =Pitaria chione(*)
 Chamelea=Venus F
 =Chione gallina(*)

family genus species substrate

SCAPHOPODA (Class)

Dentaliidae Dentalium sp. m,s

GASTROPODA (Class)

ARCHAEOGASTROPODA
(Order)

Haliotidae Haliotis lamellosa r

Fissurellidae Fissurella sp. r

Patellidae Patella caerulea(*) r
 Patella ferruginea(*) r
 Patella lusitanica r
 =rustica(*)

Trochidae Calliostoma sp. r
 Gibbula varia r,s,w
 Gibbula adansoni r,s,w
 Gibbula richardi r,s,w
 Clanculus jussieui r,g
 Clanculus cruciatus r,g
 Clanculus corallinus r,g
 Jujubinus=Cantharidis w
 corallinus
 Jujubinus=C. depictus w
 Jujubinus=C. exasperatus w
 Jujubinus=C. striatus w
 Monodonta=Gibbula articulata r
 Monodonta=G. mutabilis r
 Monodonta turbinata(*) r

Turbinidae Astrea=Turbo rugosa r,c
 Homalopoma=Leptothyra r,w
 =T. sanguineum

MESOGASTROPODA
(Order)

Littorinidae Littorina neritoides r
 Littorina obtusata(1) r
 Littorina rudis r

Turritellidae Turritella communis g,m,s

Vermetidae Vermetus sp. c,r

Architectonicidae Architectonica=Solarium sp. w,s

Cerithiidae Cerithium rupestre m,w,s
 Cerithium tuberculatum m,w,s,sp
 Cerithium vulgatum m,w,s

Cerithiopsidae Cerithiopsis tubercularis sp

Epitoniidae Epitonium=Scala commune r,c

Aporrhaidae Aporrhais m
 =Chenopus pes-pelecani

Strombidae Strombus bubonius s,c

Eratoidae Trivia europaea=adriatica r,c,w

Cypraeidae Cypraea=Talparia r
 =Luria lurida

Naticidae Naticarius dillwyni s
 Naticarius=Natica hebraea s
 Naticarius=N. s
 =Polynices guillemini
 Naticarius=N. s,g,m
 =Neverita josephina
 Naticarius=N. millepunctata s

Cassididae Phalium=Cassis undulatum s

Cymatiidae Charonia nodifera r

NEOGASTROPODA
(Order)

Muricidae Ocinebrina=Tritonalia r
 edwardsi
 Purpura lapillus r

Pyrenidae= Columbella rustica s,r,w
 Columbellidae Pyrene=Mitrella scripta r,s,c

Buccinidae Buccinum undatum(1) s
 Pisania maculosa=striata r

Nassariidae Cyclope=Nassa neritea s,m
 Nassarius=Nassa s
 =Arcularia gibbosula(1)
 Nassarius=N.=Hinia s,r,m
 incrassatus
 Nassarius=N.=H. costulata r,s
 Sphaeronassa=Nassa mutabilis s,m

Fasciolariidae Fusus sp. s,c,r

Mitridae Mitra sp. s,w,r

Conidae Conus mediterraneus r,w
 =ventricosus

BIVALVIA or
PELECYPODA (Class)

FILIBRANCHIA
(Order)

Arcidae Striarca=Arca lactea r

Glycymeridae Glycymeris s,m,g
 =Pectunculus sp.(*)

Mytilidae Mytilus galloprovincialis(*) r
 Lithophaga=Lithodomus r
 lithophaga

Pectinidae Chlamys glabra s
 Chlamys opercularis s
 Chlamys varia s
 Chlamys sp. s
 Pecten jacobaeus(*) r/s
 Pecten maximus(*) r/s

Ostreidae Ostrea edulus(*) r,s,m

EULAMELIBRANCHIA
(Order)

Cardiidae Acanthocardia s,m,g
 =Cardium tuberculatum(*)
 Cerastoderma=Cardium s,m,g
 edule(1)(*)

Veneridae Callista=Meretrix s
 =Pitaria chione(*)
 Chamelea=Venus s
 =Chione gallina(*)

 adult
 size
family genus species (mm)

SCAPHOPODA (Class)

Dentaliidae Dentalium sp. 35-50

GASTROPODA (Class)

ARCHAEOGASTROPODA
(Order)

Haliotidae Haliotis lamellosa 60-75

Fissurellidae Fissurella sp. 10-13

Patellidae Patella caerulea(*) 30-45
 Patella ferruginea(*) 60-80
 Patella lusitanica 30-45
 =rustica(*)

Trochidae Calliostoma sp. 20-40
 Gibbula varia 10-15
 Gibbula adansoni 8-15
 Gibbula richardi 8-15
 Clanculus jussieui 10-12
 Clanculus cruciatus 10-13
 Clanculus corallinus 9-15
 Jujubinus=Cantharidis 5-6
 corallinus
 Jujubinus=C. depictus 5-10
 Jujubinus=C. exasperatus 7-10
 Jujubinus=C. striatus 8-10
 Monodonta=Gibbula articulata 10-25
 Monodonta=G. mutabilis 10-15
 Monodonta turbinata(*) 20-35

Turbinidae Astrea=Turbo rugosa 30-55
 Homalopoma=Leptothyra 3-7
 =T. sanguineum

MESOGASTROPODA
(Order)

Littorinidae Littorina neritoides 5-7
 Littorina obtusata(1) 5-7
 Littorina rudis 5-7

Turritellidae Turritella communis 20-45

Vermetidae Vermetus sp. --

Architectonicidae Architectonica=Solarium sp. 15-20

Cerithiidae Cerithium rupestre 20-35
 Cerithium tuberculatum 20-65
 Cerithium vulgatum 20-65

Cerithiopsidae Cerithiopsis tubercularis 8-12

Epitoniidae Epitonium=Scala commune 20-30

Aporrhaidae Aporrhais 25-50
 =Chenopus pes-pelecani

Strombidae Strombus bubonius large

Eratoidae Trivia europaea=adriatica 10-12

Cypraeidae Cypraea=Talparia 50-60
 =Luria lurida

Naticidae Naticarius dillwyni 15-20
 Naticarius=Natica hebraea 30-45
 Naticarius=N. 15-20
 =Polynices guillemini
 Naticarius=N. 25-40
 =Neverita josephina
 Naticarius=N. millepunctata 30-45

Cassididae Phalium=Cassis undulatum 60-70

Cymatiidae Charonia nodifera v. large

NEOGASTROPODA
(Order)

Muricidae Ocinebrina=Tritonalia 7-18
 edwardsi
 Purpura lapillus --

Pyrenidae= Columbella rustica 15-20
 Columbellidae Pyrene=Mitrella scripta 15-18

Buccinidae Buccinum undatum(1) 70-80
 Pisania maculosa=striata 15-32

Nassariidae Cyclope=Nassa neritea 8-17
 Nassarius=Nassa 16-19
 =Arcularia gibbosula(1)
 Nassarius=N.=Hinia 10-15
 incrassatus
 Nassarius=N.=H. costulata 8-15
 Sphaeronassa=Nassa mutabilis 18-28

Fasciolariidae Fusus sp. --

Mitridae Mitra sp. 20-55

Conidae Conus mediterraneus 60-65
 =ventricosus

BIVALVIA or
PELECYPODA (Class)

FILIBRANCHIA
(Order)

Arcidae Striarca=Arca lactea 10-32

Glycymeridae Glycymeris 35-65
 =Pectunculus sp.(*)

Mytilidae Mytilus galloprovincialis(*) 20-100
 Lithophaga=Lithodomus 12-25
 lithophaga

Pectinidae Chlamys glabra 20-50
 Chlamys opercularis 20-25
 Chlamys varia 25-60
 Chlamys sp. 10-50
 Pecten jacobaeus(*) 85-130
 Pecten maximus(*) 100-150

Ostreidae Ostrea edulus(*) 100-160

EULAMELIBRANCHIA
(Order)

Cardiidae Acanthocardia 25-90
 =Cardium tuberculatum(*)
 Cerastoderma=Cardium 30-50
 edule(1)(*)

Veneridae Callista=Meretrix 70-90
 =Pitaria chione(*)
 Chamelea=Venus 35-40
 =Chione gallina(*)

(*) used mainly as a food source, all others were used primarily or exclusively as ornaments.

(1) tolerates or prefers brackish water.

Molluscan diet codes= (H) herbivore, (O) omnivore, (D) detritivore, (HA) herbivore that feeds mostly on algae, (HD) herbivore and detritivore, (SC) scavenger, (F) filter feeder, (C) carnivore feeds on sea urchins (-urch), coral (-cor), or bivalves (-biv).

Substrate codes= (r) rock and other firm surfaces, (m) mud, (f) floating matter and bubbles, (s) sand, (v) varied, (w) weeds, (c) corals, (g) gravel or coarse sand, (sp) sponges, (r/s) adheres to hard surfaces as juvenile but free swimming as adult. Nomenclature= (L.) Linnaeus; (Blainv.) Blainville; (Monter.) Monterosato; (Payr.) Payraudeau; (Brug.) Bruguiere.

Most of the marine molluscs found in Riparo Mochi associate with rocky (hard) surfaces of the littoral zone, fewer to soft fans of gravel or sand. Virtually none of the species in TABLE 2 favors brackish water, although some tolerate it (e.g. Cerastoderma edule, Barnes 1994; and Cyclope neritea, Sauriau 1991). The marine environment of the Balzi Rossi today is subject to persistent deep convection and upwelling (Longhurst 1998:136), conditions which favour high species diversity in undisturbed coastal settings. Local gyres sweep nutrient-rich outflow from large rivers and benthic water over the steep rock surfaces of the Balzi Rossi. While the Ligurian coast is much altered by recent development, and local species diversity reduced, the rich assortment of marine mollusc species ([is greater than] 55) in the Upper and Epi-Palaeolithic layers of Riparo Mochi testifies to both high diversity and stability there between 36,000 and 9000 years ago.

Taphonomic histories of the shells

The shells from Riparo Mochi can be divided into four distinct groups based on damage patterns. These groups are defined by the relative frequencies of

1 beach polish induced by wave action;

2 shell completeness based on the ratio, MNI/ NISP;

3 human-caused perforation (or sectioning in the case of tusk shells) and perforation-related breaks emanating from the perforation point; and

4 burning by fire.

There is little variation among the five Epi- and Upper Palaeolithic assemblages in the kinds of damage present, and thus TABLE 3 and FIGURE 2 summarize the taphonomic histories for all assemblages combined. The frequencies of beach polish and burning damage are expressed as percentages of NISP (number of identified specimens), because the degree of shell completeness varies among the shell classes affected by such damage. MNI (minimum number of individual animals) is the counting unit for human-caused perforations, because fragmentation is minimal for the shells that display this damage.

[Figure 2 ILLUSTRATION OMITTED]

TABLE 3. Summary of damage frequencies for various shell categories, all Epi- and Upper Palaeolithic layers combined.

 small
 marine tusk
 gastropods shells limpets
variable (1) (2) (3)

NISP and MNI counts
total NISP 2310 74 1498
total MNI 2197 63 846

percentage values based on NISP or MNI
beach polish (% of NISP) 11 4 <1
completeness (MNI/NISP) 95 85 56
perforation (% of MNI) 48 48 <1
burned (% of NISP) 5 4 3

 large common rare
 turbans bivalves bivalves
variable (4) (5) (6)

NISP and MNI counts
total NISP 257 9114 317
total MNI 129 861 109

percentage values based on NISP or MNI
beach polish (% of NISP) <1 0 20
completeness (MNI/NISP) 50 9 34
perforation (% of MNI) 0 <1 3
burned (% of NISP) 7 18 14

 small terrestrial
 horns gastropods
variable (7) (8)

NISP and MNI counts
total NISP 1832 6680
total MNI 1832 6675

percentage values based on NISP or MNI
beach polish (% of NISP) 24 0
completeness (MNI/NISP) 100 100
perforation (% of MNI) (*) (*)
burned (% of NISP) 35 <1

Perforation count includes tusk shells sectioned to make tube beads.

(*) Some holes are present, but these are due only to extreme beach polish in the case of Cerithiopsis, and from small (avian or mammalian) predators in the case of land snails.

() shell category code in FIGURE 2.

Shell Group I: ornaments

Small marine gastropods, tusk shells (Dentalium), and certain small colourful bivalves were used exclusively as ornaments or talismans. About half of these shells (48%) were perforated for suspension, and 4-5% were burned. Unperforated shells may have been cached and forgotten, or used in ways that did not require suspension (Taborin 1993a). While some of the shells in this damage complex display wave-induced polish (11% of gastropods, 4% of tusks), indicating that many if not all of the shells were collected from beaches, almost all of the shells are whole or nearly whole (95% of gastropods, 85% of tusks). Clearly, beach polish was of little concern to Palaeolithic collectors, but wholeness was essential. Breaks emanating from perforations occur at a rate of 1-9%, depending on the taxon, and many appear to have resulted from use rather than production errors. Perforation was normally achieved by impact- or pressure-punching with a small pointed tool. This was facilitated in some cases by pre-boring or pecking the shell's outer surface; none were drilled. Most of the perforations on gastropod shells are located near the aperture (FIGURE 3) and have rough outlines, but a few large specimens instead were pierced by abrasive `sawing'. Holes drilled by naticid or muricid predatory snails are easily distinguished from perforations made by humans: the former are perfectly round, bevelled and randomly placed, whereas human-made holes tend to have irregular outlines, or are elliptical, and their anatomical placement is quite consistent. Tusk shells were snapped transversely to make tube beads. A few limpet shells were perforated ([is less than] 1% of MNI) through the apex, and they may have been strung in pairs -- margin to margin -- to create small talisman containers, like those from the Late Epigravettian burials in Arene Candide in Liguria (Cardini 1980). Fan scallops (Chlamys) and ark shells (Striarca) usually were perforated at the umbo; some of these holes are human-made, but others appear to be natural and merely capitalized upon by humans. Cord-wear and red ocher staining are present but uncommon.

[Figure 3 ILLUSTRATION OMITTED]

Shell Group II: food

Limpets (Patella spp.), large turbans (Monodonta turbinata) and large bivalves were consumed almost exclusively as food. The bivalves are mostly mussels (Mytilus galloprovincialis) along with small quantities of other genera. All of these species are relatively large-bodied and inhabit intertidal to shallow subtidal waters. Limpets may be found in loose clusters on shoreline rocks. Mussels, the most abundant bivalve in Riparo Mochi by far, are distinguished by their tendency to live in dense colonies on shoreline rocks (e.g. Little & Kitching 1996). Much less common in the archaeological collections are sand dwelling and free-swimming bivalves (cf. TABLE 2). Little if any beach polish occurs on limpets or common bivalves ([is less than] 1%), but it is present on up to 3% of the rare bivalve taxa (TABLE 3). Shells in Group II were almost never perforated by humans, and fragmentation is extensive: only 9% of the common bivalve specimens are complete, 34% of rare bivalve taxa. The fracture edges are consistently sharp (FIGURE 4), contra-indicating accumulation and abrasion by wave action. The incidence of burning (usually mild) varies between 3% and 7% of NISP for most marine shells in Riparo Mochi, including limpets. However, burning is much more common on edible bivalves (all species, 14-18%), suggesting that fire was used to open them. The taphonomic histories of the rare bivalve taxa are not as straightforward as those of turbans, limpets and mussels. While most of the rare bivalves display damage typical of food processing, the incidence of beach polish is also high (20%), and thus it is possible that some were also used as ornaments or containers. The great majority of shells in Group II nonetheless arrived in the shelter while the animals were alive. Nearly all of the post-mortem damage occurred as humans processed the shells, often with the help of fire, consumed their contents and trampled the debris.

[Figure 4 ILLUSTRATION OMITTED]

Shell Group III: accidentals (Cerithiopsis tubercularis)

This enigmatic damage group contains only one species of `small horn', distinguished by the high co-incidence of beach polish (24% on average) and burning damage (35%). Small in size (8-12 mm), the horn shells were never deliberately perforated by humans, although extensive polishing by waves left holes in the surfaces of many (FIGURE 5). Apart from wave-caused abrasion, virtually all of the small horns are whole. These tiny shells (0.8-1.2 cm) are particularly abundant in the Gravettian Layer D. Cerithiopsis tubercularis, incorrectly identified as Bittium reticulatum by the author (in Kuhn & Stiner 1998a), displays a strong attraction to marine sponges (Halichondria panicea and Hymeniacidon perleve) during life, where fine red algae is also available as graze (Hayward et al. 1995: 525). The close proximity of small horn shells and sponges in life allows vacated shells to become lodged in the open structures of the sponges. The wave-worn condition of the horns from Riparo Mochi suggests that their transport to the shelter by humans was an accidental by-product of sponge collection. The prevalence of burning damage on the horn shells implies a consistent spatial connection between sponge use and hearth-centred activities in Riparo Mochi. No intact sponges are preserved in the Mochi collections, although a few crumbs were found. Use of marine sponges is also known from the Middle Palaeolithic site of Grotta dei Moscerini near Gaeta, Italy (Stiner 1994: 183ff).

[Figure 5 ILLUSTRATION OMITTED]

Shell Group IV: shelter co-residents

Numerous species of land snail (terrestrial gastropods) either inhabited Riparo Mochi and died there or were carried into the shelter by carnivorous birds or small mammals such as hedgehogs, mice or foxes. Some of the shells display tiny perforations but otherwise are undamaged (FIGURE 6; see also Stiner 1994: 172). There is no evidence linking the land snails to human activities; the fragile, whole shells occur in great quantities, yet burning damage -- so widespread in other shell categories -- is exceedingly rare on land snail shells ([much less than] 1%). Land snails are potentially edible and are known to have been exploited in North Africa during the Epi-Palaeolithic (e.g. Close & Wendorf 1990), but individuals of worthy size are rare in Riparo Mochi, and the human inhabitants evidently did not eat them.

[Figure 6 ILLUSTRATION OMITTED]

Spatial associations among shells, stone artefacts and ungulate bones

The four shell groups -- ornamental, food, accidental, and land snail -- display rather different quantitative relations to non-shell materials in the stratigraphic sequence of Riparo Mochi. Here, I examine the distributions of material across 63 fine cuts of the east trench, excavated in 1959. The cuts traverse the Gravettian (D) through Early Aurignacian (G) layers, younger layers A through C having been removed during earlier field campaigns. The total weights (g) of lithic artefacts and ungulate bones by cut provide two baselines against which shell frequencies (MNI) are compared in a Pearson's (r) correlation matrix (TABLE 4). Ungulate remains obviously testify to terrestrial foraging of some kind, and edible shellfish must reflect marine foraging.

TABLE 4. Peason correlation matrix comparing mollusc shell frequencies (MNI) to the weights (g) of lithic artefacts and ungulate remains from 63 fine cuts in the Gravettian through Early Aurignacian layers, 1959 excavations (east trench only), at Riparo Mochi. Bivalves and limpets are evaluated separately, because they display different frequencies of burning damage; turbans are too few to consider.

 mammal lithic ornamental
 remains artefacts shell
 (g) (g) MNI

weight ungulate remains (g) --
weight lithic artefacts (g) (*)0.93 --
ornamental shell MNI (*)0.74 (*)0.75 --
limpet (food) MNI -0.07 0.03 0.17
bivalve (food) MNI 0.06 0.15 0.27
land snail MNI -0.05 -0.10 0.01
small horn MNI (*)0.64 (*)0.53 (*)0.81

 (food) (food) land small
 limpet bivalve snail horn
 MNI MNI MNI MNI

weight ungulate remains (g)
weight lithic artefacts (g)
ornamental shell MNI
limpet (food) MNI --
bivalve (food) MNI 0.27 --
land snail MNI -0.04 0.17 --
small horn MNI -0.07 -0.03 -0.02 --

(*) r value >0.41 is statistically significant at the .001 level of probability (N cuts = 63, df = 61).

Strong, positive correlation coefficients exist among the frequencies of ornamental marine shells, sponge-dwelling Cerithiopsis shells, lithic artefacts and ungulate remains across the 63 cuts. The correlations are not affected by the subtraction of zero cells from the sample. Oddly, there is no correlation between marine limpets and marine bivalves, or either of these to other anthropogenic materials in the stratigraphic series. The limpet and bivalve shells represent food debris, yet their deposition was independent of that of lithic artefacts and the bones of large game animals. The spatial independence of terrestrial game and lithic debris on the one hand, and edible shellfish on the other, may reflect fine-scale temporal variation in foraging activities. There is no correlation between the quantities of land snails and any other class of debris (TABLE 4). It is, however, the lack of human-caused damage that refutes a behavioural connection between land snails and human activities in Riparo Mochi.

Ornaments: interassemblage variation and its causes

The ornamental shells from Riparo Mochi include many species of marine gastropod (FIGURE 7), along with a few scaphopods and bivalves (FIGURE 8). Living mollusc diversity and biomass by trophic level provide two standards with which to gauge Palaeolithic humans' investment in the search for raw material, and the extent to which the selection of shells for ornament-making was bound by cultural rules. Taxonomic representation and adult shell size distributions in the archaeological assemblages (TABLE 2) are compared to natural patterns of availability below.

[Figures 7-8 ILLUSTRATION OMITTED]

Human responses to natural availability

The Palaeolithic occupants of Riparo Mochi exploited many more marine gastropod species than bivalve species, the former mainly as ornaments and the latter as food (TABLE 5). FIGURE 9 compares the proportions of marine gastropod and bivalve species in the archaeological assemblages to a standard taken from Sabelli (1980: 23ff) for living marine communities worldwide. While species composition of one marine biographic province will differ from the next, strong parallels exist at higher taxonomic levels (Hayward et al. 1995; Sabelli 1980) due to the currents that connect populations in the world's oceans and seas (Gaines & Lafferty 1995; Longhurst 1998: 135-8). Moreover, the rocky coasts of the northern Mediterranean are relatively productive and tend to support many marine species, warranting the application of a global standard.

[Figure 9 ILLUSTRATION OMITTED]

TABLE 5. Relative abundances (percentages) of ornamental, food and small horn marine molluscs in the Epi- and Upper Palaeolithic layers of Riparo Mochi.

 with small horn shells included

 %
 total % % small
period & layer MNI ornaments food horns

Late Epigravettian (A) 1857 51 43 6
Early Epigravettian (C) 368 49 44 7
Gravettian (D) 2446 27 6 67
Middle Aurignacian (F) 526 37 60 2
Early Aurignacian (G) 1125 52 45 3

 without small horn shells

 total % %
period & layer MNI ornaments food

Late Epigravettian (A) 1741 54 46
Early Epigravettian (C) 341 52 47
Gravettian (D) 814 82 18
Middle Aurignacian (F) 513 38 62
Early Aurignacian (G) 1091 53 46

Most ornamental shells are marine gastropods; most food shellfish are bivalves; small horn shells associate with sponge collecting and use by humans.

None of the five shell ornament assemblages from Riparo Mochi differs significantly from Sabelli's global ratio of gastropod to bivalve species. The high proportion of gastropod species among the Palaeolithic ornaments is explained simply by humans' emphasis on gastropods as sources of raw material. This aspect of shell use was constant from the Upper through the Epi-Palaeolithic at Riparo Mochi.

The lack of cultural bias in the ratio of gastropod to bivalve species is a useful benchmark for the next comparison, in FIGURE 10, of the relative biomass (MNI) of carnivorous, omnivorous and herbivorous gastropods among the shell ornaments from Riparo Mochi. Commensurate with high taxonomic diversity, the class Gastropoda embodies myriad adaptations and dietary specializations (see TABLE 2). Here the standard for comparison is a census of beach-cast shells (MNI = 347) conducted by the author in 1997 on the Hatay coast of south-central Turkey (FIGURE 1). The census serves as a general model of the opportunities for collecting vacated shells, since the Palaeolithic people at Riparo Mochi obtained many or most of their raw material in this way. The Hatay coast is geographically removed from the Italian Riviera, but it resembles the geological, hydrological and presumed ecological characteristics of the Riviera prior to recent development. Rockdwelling taxa predominate at both localities, and locally steep shorelines consist of small sand and/or gravel beaches tucked among broad limestone cliffs and spits in both areas. Coastal currents (Longhurst 1998:136) push nutrient-laden river outflow along craggy limestone coasts.

[Figure 10 ILLUSTRATION OMITTED]

Carnivorous taxa should always be less abundant in animal communities than their herbivorous and omnivorous prey. The Hatay reference sample is true to this expectation (FIGURE 10), as carnivorous gastropod MNI is low (13%), and herbivore MNI (38%) and omnivore MNI (49%) are high. The archaeological assemblages from Riparo Mochi deviate strongly from the natural pattern of shell availability by trophic category. Humans favoured the shells of certain relatively rare carnivorous gastropods. The shells of herbivores were collected in the proportion predicted by the trophic organization of mollusc communities, but omnivorous species are strangely under-represented. The bias favouring carnivores is due to humans' strong preference for the shells of Cyclope neritea, a carnivorous scavenger (Morton 1960). This marine snail also displays some tendency toward omnivory, and thus its natural abundance could be higher than that of pure carnivores, but its natural abundance is still well short of that for dedicated omnivores and herbivores. Nassarius gibbosula, a close relative of Cyclope and similar in habit (Tornaritis 1987: 95-7), is relatively rare (1% of MNI) in the beach-cast assemblage from the Hatay, and requires considerable time to find. Because carnivores are most abundant in the archaeological samples, and must have been relatively uncommon in beach-cast situations then as now, it is likely that humans were responding to a perception of rarity.

Another human bias is evidenced by the size distribution of the ornamental shells. FIGURE 11 compares the range of adult shell sizes naturally attained by the genera represented in the Mochi assemblages (TABLE 2) to shell MNI by size category. The sharp peak between 4 and 8 mm for the archaeological shells is nearly identical to that observed by White (1989: 382) for nonshell ornaments laboriously carved from ivory and stone from the Perigord region of France. However, the archaeological size distribution is highly skewed relative to the natural median for all shells -- humans obviously favoured small shells 5-16 mm in length. Most beach-cast shells are not from mature animals, and thus are smaller on average, but the median shell size for the Hatay gastropods (26 mm) is well in excess of the archaeological median. The size distribution of the shell ornaments from Riparo Mochi is not explained solely by humans' preference for carnivorous types. The smallest, most abundant shells (TABLE 6) are the carnivorous scavenger, Cyclope neritea (FIGURE 3 & 7-Cn) and the herbivore, Homalopoma sanguineum (FIGURE 7-Hs). Cyclope has a cream to red-brown, ovoid shell (FIGURE 3), and Homalopoma a bright pink to red, spherical shell. In addition to preferring certain rare shells, it is clear that human collectors responded selectively to shell shape, size, and colour.

[Figure 11 ILLUSTRATION OMITTED]

TABLE 6. Relative abundances (MNI) of common and uncommon genera in the ornamental shell assemblages from the Epi- and Upper Palaeolithic layers of Riparo Mochi.

 Layer

shell types A C D F G

Cyclope & Homalopoma 51% 57% 73% 64% 45%
Cerithium & Nassarius 13% 9% 14% 3% 11%
rare/uncommon genera 36% 34% 13% 33% 43%

total assemblage MNI 668 131 662 174 589
N-genera 30 20 25 20 32

Inter-assemblage variation in the shell ornaments

Human preferences for certain ornamental shells are clear, but the favourite genera changed remarkably little between 36,000 and 9000 years ago. While human preferences are boldly apparent with respect to certain rare, small shells, FIGURE 12 reveals little if any variation in these preferences across the five Epi- and Upper Palaeolithic phases at Riparo Mochi (see also Leonardi 1935:27 on shells from Barma Grande). Cyclope and Homalopoma shells dominate in every culture phase (TABLE 6), and the second most common genera are usually Cerithium and Nassarius. However, there are differences in the proportions of all remaining genera combined among the five assemblages. It is the Gravettian assemblage from Layer D that stands apart from the others, due the scarcity of genera other than the four listed above. The Gravettian layer probably dates to the earlier shoulder of the Last Glacial Maximum, when sea level was declining, and its vertebrate fauna is the most terrestrially oriented of the series (see below). Declining sea level could have altered details of the local habitat and species availability, but, if this was the case, the changes were subtle.

[Figure 12 ILLUSTRATION OMITTED]

Non-shell ornaments

Ornaments made from materials other than shell are rare ([is less than] 1%) in Riparo Mochi. Of these, pendants made from red-deer canines (`pearl teeth' of Cervus elaphus) seem to cluster in the Gravettian and Epigravettian (FIGURE 13). Also present is one amorphous limestone pebble, with a natural hole, that may have been used as an ornament (FIGURE 13e). The Early Aurignacian layer instead yielded one perforated carnivore incisor, from fox or wild cat (FIGURE 13j), and four small basket-shaped beads, one carved from compact bone or possibly ivory (FIGURE 13k) and three faceted specimens carved from steatite or chlorite (FIGURE 13l-n) (Kuhn & Stiner 1998a). The most consistent features of the carved Aurignacian ornaments are their size, form and high manufacturing investment; considerably less shaping was required for the red-deer pearl teeth of the Gravettian and Epigravettian. While the carved beads mirror the natural forms of red-deer canines and Cyclope neritea shells (FIGURE 14), it is difficult to say which served as inspiration for the others.

[Figures 13-14 ILLUSTRATION OMITTED]

Interestingly, the carved ornaments from the Early Aurignacian of Riparo Mochi are indistinguishable from those found in coeval sites of the Perigord region (Collie 1928: 106-12; White 1989) in size, form and raw material. These ornaments also resemble in size and shape the most common ornamental mollusc at Riparo Mochi, Cyclope neritea (FIGURE 14). So rare and locally peculiar are the carved stone and bone pendants from the Early Aurignacian of Riparo Mochi that it seems likely that they were obtained from the continental interior by trade.

Search and manufacturing costs and the influence of local ecology

Shell-working probably took much less time than did searching local beaches for appropriate raw material. The techniques for making the few non-shell ornaments found in Riparo Mochi were much more elaborate, involving labour-intensive carving and shaping by abrasion (see White 1989; 1993). The extraordinary emphasis on shell for ornament-making throughout the Upper and Epi-Palaeolithic at Riparo Mochi is best explained by local availability and ease of manufacture. Certain small species were preferred, but the search image was otherwise eclectic. The high frequency of Cyclope neritea among the ornaments is typical of the Balzi Rossi sites in general, and also of the Late Epigravettian burials in the Ligurian site of Arene Candide (Cardini 1980). Shells from the Early Aurignacian of Riparo Fumane, which lies 350 m above modern sea level and 80 km inland on the eastern side of the Italian Peninsula (Bartolomei et al. 1994; Broglio 1995) and whose bone and stone tool industries closely resemble those from Mochi G (Kuhn and Stiner 1998a), are instead dominated by Homalopoma sanguineum (44%, N = 484; from Fiocchi 1996-7). The second most common ornamental shell at Fumane is Cyclope, however, and the taxonomic profile is quite similar to the Mochi assemblages overall. Interassemblage differences of this minor order could be explained by humans' tendency to reorganize species abundance patterns at the scale of customized shell strands or sets.

The French Palaeolithic cases nearest to the Balzi Rossi are most similar to the Mochi ornament assemblages in taxonomic content (Taborin 1993a). Farther away, species composition diverges much more. Levantine Palaeolithic cases are dominated by Dentalium (Bar-Yosef 1991; Mienis 1977; Reese 1989) and those in southernmost Turkey by Dentalium or Nassarius gibbosula and Columbella rustica (Kuhn et al. 1999). In other words, distance gradients have much to do with levels of taxonomic similarity among Palaeolithic shell-ornament assemblages. Cyclope was widely used in southern France and northwestern Italy, but it was not widely exchanged within Aurignacian or later Palaeolithic populations inhabiting the continental interior or other coastlines. The abundance of C. neritea in the Ligurian sites reflects a peculiarity of Riviera marine ecology, independent of changes in Palaeolithic stone and bone tool industries over time.

Ornament variation vs foraging variation

While humans' selection of shells for ornament-making changed very little between the Early Aurignacian and Epigravettian periods at Riparo Mochi, important economic changes are indicated by shifts in the relative contribution of edible molluscs to total game (Stiner et al. in press; Stiner n.d.). FIGURE 15 compares the proportions of small and large game in the five Palaeolithic layers, and FIGURE 16 the relative contribution of shellfish to the small game fraction of each layer. The abrupt increase in aquatic resources at Riparo Mochi in the Late Epigravettian coincides with the beginning of the Holocene, when sea level transgressed nearly to the foot of the shelter.

[Figures 15-16 ILLUSTRATION OMITTED]

Discussion

Few sites provide a better opportunity to investigate the causes of temporal and geographic variation in Palaeolithic shell ornaments than does Riparo Mochi. Whilst other aspects of material culture at this shelter changed considerably between 36,000 and 9000 years ago, humans' preferences for certain ornamental shells were as invariant as the environment from which these shells were obtained. The Balzi Rossi locality was a rich and relatively stable source of shell raw material during the Late Pleistocene. Palaeolithic people responded to this situation by collecting many species for ornament-making, although they favoured small, rare shells of the genera Cyclope, Homalopoma, Cerithium and Nassarius. Ornament use was not confined to grave goods. Shell-bead accumulation is closely tied to the deposition of terrestrial resources and stone artefacts in camp litter, but not to the products of marine foraging. Non-shell ornaments are rare, incidental additions to the Upper and Epi-Palaeolithic ornament repertoires. The rare and intricately carved basket-shaped beads of the Early Aurignacian (layer G) closely resemble coeval specimens from the French Perigord sites, and may have been obtained by Mochi residents via trade. If so, the distance-exchange gradient was very steep. In addition to molluscs, the Palaeolithic occupants of Riparo Mochi made use of marine sponges, whose frequencies based on the presence of Cerithiopsis shells also correlate with the prevalence of terrestrial game, stone tools, and ornaments, but not with edible shellfish.

There is nothing tentative or half-baked about the Early Aurignacian ornaments from Riparo Mochi. Ornament-making appears suddenly and abundantly in this Middle-Upper Palaeolithic sequence (Kuhn & Stiner 1998a), as it does elsewhere in Europe (Hahn 1972; Leroi-Gourhan 1961; Mellars 1989; Palma di Cesnola 1993; White 1982). The presence of small quantities of ornamental material in the Chatelperronian or Uluzzian layers of some European sites in no way undermines the general abruptness of this cultural transition (Broglio 1995; D'Errico et al. 1998; White 1989).

More surprising is the lack of variation in shell ornaments following the Early Aurignacian at Riparo Mochi. The Gravettian assemblage shows a slightly higher degree of taxonomic homogeneity, possible evidence for standardization, but the ornament assemblages otherwise are quite similar across the five Upper and Epi-Palaeolithic phases. The Aurignacian-Gravettian transition in the southern Europe has been interpreted from other materials to represent a significant break in cultural traditions (Bietti 1997; Palma di Cesnola 1993), perhaps associated with somewhat distinct human populations. While this assessment may be correct, no clear stylistic transition is apparent from the shell ornaments. Instead, the relatively stable local ecology and well-nourished waters of the Balzi Rossi coastline explain both the richness and uniformity in humans' choice of shells for ornaments from the beginning of the Upper Palaeolithic to its end. It seems that people of this rocky coast were rich with raw material from the sea, just as some contemporaneous peoples of the northern interior were rich with high-quality flint. In this context, Palaeolithic humans seldom turned to labour-intensive production of nonshell ornaments.

Shells would seem to be an excellent medium for signalling individual or group identities. However, the contents of the shell-ornament assemblages from Riparo Mochi are faithful principally to local raw material availability, and only secondarily to human perceptions of scarcity and appearance. Of course raw material is only one aspect of how decorative traditions are produced. Its availability nonetheless contributes heavily to the appearance of Palaeolithic ornament assemblages.

The results of this study underscore two aspects of emblematic behaviour during the Upper and Epi-Palaeolithic, one bound directly by human intention and another dictated by aspects of local ecology which humans simply adopted. Deliberate were humans' arrangements of shells on themselves and on the bodies of the deceased with the help of cordage, mastics and tiny containers. The makers were also selective with respect to shell size, shape and ecological rarity. Individual shells no doubt served as minimal elements in larger, more arresting combinations (e.g. beaded caps and bracelets in Grimaldi and other burials), and significant meaning may have been confined mainly this level of bead-working traditions. This point unfortunately leads us back to precious few burials for information on ornament assembly and the visual syntax of ornament use. That people consistently favoured a very narrow range of shell sizes, and that the size range is nearly identical to that for meticulously carved ornaments from the continental interior, is both intriguing and difficult to explain in ecological or practical terms. It suggests some kind of shared aesthetic, yet one that lasted more than 20,000 years. Such a widespread aesthetic may originate with certain objects of constant, natural form that humans find attractive and which other materials could be made to imitate. Ornaments of similar appearance but different raw material were somewhat interchangeable design elements, a minor concern relative to the larger compositions of which they were a part.

It is possible, even probable, that the focus on shells for ornament-making in coastal areas, and on certain locally available species, also served to identify human groups with regions, beyond any intentional signalling of social roles or life history states. The kinds of ornaments used by Palaeolithic people varied least over time where the source of raw material was most constant, and thus the impressions to be made by shell ornaments were born directly of local biogeography. The steep exchange gradient for ornaments -- most shells staying near the coast and most carved ornaments and mammal teeth remaining in the continental interior -- may mean that these kinds of visual expression were directed internally to the group and its most immediate neighbours. From here, research on shell ornaments is no longer an exercise in zoological classification, but rather a matter of how local ecology, labour and craft lend distinctiveness to human styles of ornamentation. Many of the primary hypotheses about shell ornaments nonetheless can be tested by reference to natural standards of molluscan ecology and community structure.

Acknowledgements. This research would not have been possible without the help of A. Segre, E. Segre-Naldini, F. Parenti, R. Sebastiani (Istituto Italiano di Paleontologia Umana), A. Bietti, F. Alhaique, A. Recchi (Universita di Roma), P. Cassoli and A. Tagliacozzo (Museo Pigorini). I am also grateful to S. Kuhn for advice on the quantitative analyses and arguments presented, to the anonymous ANTIQUITY reviewers for their valuable comments, and to D. Bar-Yosef and H. Mienis on matters of mollusc classification; all errors herein are mine. I dedicate this work to the memory of L. Cardini, whom I have come to know only from his notes and innovative field techniques. Thanks also to P. Mellars (Cambridge University) for supporting AMS radiocarbon dating of the earliest Upper Palaeolithic of Riparo Mochi. While the drawings herein are mine, some are redrafted from illustrations in Blanc (1953) and Leonardi (1935).

(1) A complete account of the Mochi faunas and archaeological background is in preparation (Stiner n.d.). In the meantime, requests for the mollusc data appendixes on which this study is based should be made in writing to the author.

References

ALHAIQUE, E, A. BIETTI, A. DEL LUCCHESE, S. GRIMALDI & A. RECCHI. 1997. I nuovi scavi al Riparo Mochi, il contributo dell'Istituto Italiano di Paleontologia Umana alla conoscenza dei Balzi Rossi, Atti del Convegno `Balzi rossi 1846-1996:150 anni di preistoria tra collezionisma e indagine scientifica' (Ventimiglia 1-2 febbraio 1997), in Rivista di Studi Liguri.

BARNES, R.S.K. 1994. The brackish-water fauna of northwestern Europe. Cambridge: Cambridge University Press.

BARRAL, L. & S. SIMONE. 1969. Sur la presence a la Grotte du Prince (Grimaldi, Ligurie Italienne) de breches a ossements rissiennes et de formations attributables a la mer du Mindel-Riss, Comptes Rendus de l'Academie de Sciences, Paris D 268: 637-40.

BARTOLOMEI, G., A. BROGLIO, P.F. CASSOLI, L. CASTELLETTI, L. CATTAMO, M. CREMASCHI, G. GIACOBINI, G. MALERBA, A. MASPERO, M. PRESANI, A. SARTORELLI & A. TAGLIACOZZO. 1994. La Grotte de Fumane: un site aurignacien au pied des Alpes, Preistoria Alpina 28(1992): 131-79.

BAR-YOSEF, D.E. 1991. Changes in the selection of marine shells from the Natufian to the Neolithic, in O. Bar-Yosef & F.R. Valla (ed.), The Natufian Culture of the Levant: 629-36. Ann Arbor (MI): International Monographs in Prehistory. Archaeological Series 1.

BIETTI, A. 1997. The transition to anatomically modern humans: the case of peninsular Italy, in G.A. Clark & C. Willermet (ed.) Conceptual issues in modern human origins research: 132-50. New York (NY): Aldine de Gruyter.

BISSON, M. & P. BOLDUC. 1994. Previously undiscovered figurines from the Grimaldi caves, Current Anthropology 35: 458-68.

BLANC, A.C. 1953. Il Riparo Mochi ai Balzi Rossi di Grimaldi: le industrie, Paleontografic Italica 50 (Bozze di Stampa, apparently never published).

BOULE, M. & H.V. VALLOIS. 1957. Fossil men. New York (NY): Dryden Press.

BROGLIO, A. 1995. Discontinuita tra musteriano e protoaurignaziano mediterraneo nella Grotta di Fumane (Monte Lessini, Prealpi Venete), VELEIA, Revista de Prehistoria, Historia Antigua, Arqueologia y Filologia Clasicas 12: 49-65.

CARDINI, L. 1980. Le necropoli Mesolitica delle Arene Candide (Liguria), Memorie dell'Istituto Italiano di Paleontologia Umana (Nuove Serie) 3: 9-31.

CARDINI, L. & I. BIDDITTU. 1967. L'attivita scientifica dell'Istituto Italiano di Paleontologia Umana, dalla sua fondazione. I. Liguria, Quaternaria 9: 353-83.

CLOSE, A. & F. WENDORF. 1990. North Africa at 18,000 BP, in C. Gamble & O. Softer (ed.), The world at 18,000 BP: Low latitudes (volume 2): 41-57. London: Unwin Hyman.

COLLIE, G. L. 1928. The Aurignacians and their culture. Beloit (WI): Logan Museum, Beloit College. Logan Museum Bulletin 1(1).

D'ERRICO, F., J. ZILHAO, M. JULIEN, D. BAFFIER & J. PELEGRIN. 1998. Neanderthal acculuturation in western Europe? A critical review of the eidence and its interpretations, Current Anthropology 39 (supplement): 1-44.

FIOCCHI, C. 1996-7. Le conchiglie marine provenienti dalla Grotta di Fumane (Monti Lessini-Verona), Atti dell'Istituto Veneto di Scienze, Lettere ed Arti 155: 1-22.

GAINES, S.D. & K.D. LAFFERTY. 1995. Modeling the dynamics of marine species: the importance of incorporating larval dispersal, in L. McEdward (ed.), Ecology of marine invertebrate larvae: 389-412. Boca Raton (FL): CRC Press.

GAMBLE, C. 1986. The Palaeolithic settlement of Europe. Cambridge: Cambridge University Press.

HAHN, J. 1972. Aurignacian signs, pendants, and art objects in Central and Eastern Europe, World Archaeology 3: 252-66.

HARROLD, F. 1989. Mousterian, Chatelperronian, and early Aurignacian in Western Europe: Continuity or discontinuity? in Mellars & Stringer (ed.): 677-713.

HAYWARD, P.J., G.D. WIGHAM & N. YONOW. 1995. Molluscs (Phylum Mollusca), in P.J. Hayward & J.S. Ryland (ed.), Handbook of the marine fauna of northwest Europe: 484-628. Oxford: Oxford University Press.

HEDGES, R.E.M., R. HOUSLEY, C. BRONK RAMSEY & G.J. VAN KLINKEN. 1994. Radiocarbon dates from the Osford AMS System: Archaeometry Datelist 18, Archaeometry 36: 337-74.

KNECHT, H., A. PIKE-TAY & R. WHITE (ed.). 1993. Before Lascaux: The complex record of the early Upper Paleolithic. Boca Raton (FL): CRC Press.

KOZLOWSKI, J. 1990. A multi-aspectual approach to the origins of the Upper Paleolithic in Europe, in Mellars (ed.): 419-37.

KUHN, S.L. & M.C. STINER. 1992. New research on Riparo Mochi, Balzi Rossi (Liguria): preliminary results, Quaternaria Nova II: 79-90.

1998a. The earliest Aurignacian of Riparo Mochi (Liguria, Italy), Current Anthropology 39 (supplement): S175-89.

1998b. Middle Paleolithic `creativity': reflections on an oxymoron? in S. Mithen (ed.), Creativity and human evolution and prehistory:. 143-64. London & New York (NY): Routledge.

KUHN, S.L., M.C. STINER & E. GULEC. 1999. Initial Upper Palaeolithic in south-central Turkey and its regional context: a preliminary report, Antiquity 73: 505-17.

LAPLACE, G. 1977. Il Riparo Mochi ai Balzi Rossi di Grimaldi (Fouilles 1938-1949): Les industries leptolithiques, Rivista di Scienze Preistoriche 32: 3-131.

LEONARDI, P. 1935. I Balzi Rossi: le faune 1. I molluschi pleistocenici della Burma Grande. Florence: Istituto Italiano di Paleontologia Umana. Anno XIII.

LEROI-GOURHAN, A. 1961. Les fouilles d'Arcy-sur-Cure, Gallia Prehistoire 4: 3-16.

LITTLE, C. & J.A. KITCHING. 1996. The biology of rocky shores. Oxford: Oxford University Press.

LONGHURST, A. 1998. Ecological geography of the sea. San Diego (CA): Academic Press.

LUMLEY-WOODYEAR, H. DE. 1969. Le Paleolithique Inferieur et Moyen du Midi Mediterraneen dans son cadre geologique I, Ligurie-Provence. Paris: CNRS. Ve Supplement a Gallia Prehistoire.

MELLARS, P. 1989. Major issues in the origin of modern humans, Current Anthropology 30: 349-85.

(Ed.) 1990. The emergence of modern humans: An archaeological perspective. Ithaca (NY): Cornell University Press.

1996. The Neandertal legacy. Princeton (NJ): Princeton University Press.

MELLARS, P. & C. STRINGER (ed.). 1989. The human revolution. Princeton (NJ): Princeton University Press.

MIENIS, H.K. 1977. Marine molluscs from the Epipaleolithic Natufian and Harifian of the Har Harif, Central Negev, Israel, in A.E. Marks (ed.), Prehistory and paleoenvironments in the Central Negev, Israel 2: 493-506. Dallas (TX): Southern Methodist University Press.

MORTON, J.E. 1960. The habits of Cylcope neritea, a style-bearing stenoglossan gastropod, Proceedings of the Malacological Society 34: 96-105.

OTTE, M. 1990. From Middle to Upper Paleolithic: The nature of the transition, in Mellars (ed.): 438-56.

PALMA DI CESNOLA, A. 1993. Il Paleolitico superiore in Italia. Florence: Garlatti & Razzai Editori.

REESE, D.S. 1989. Appendix D, The Natufian shells from Beidha, in B.F. Byrd (ed.), The Natufian encampment at Beidha: Late Pleistocene adaptation in the southern Levant: 102-4. Arhus: Jutland Archaeological Society. Publications 23(1).

RENAULT-MISKOVSKY, J. 1972. Contribution a la paleoclimatologie du Midi Medtierraneen pendant la derniere glaciation et le post-glaciale, d'apres l'etude palynologique du remplissage des grottes et abris sous roche, Bulletin Musee d'Anthropologie et Prehistoire de Monaco 18: 145-210.

RIEDL, P. 1970. Fauna und Flora der Adria. Hamburg & Berlin: Verlag Paul Parey.

RIVIERE, E. 1887. De l'antiquite de l'homme dans les Alpes' Maritimes. Paris.

SABELLI, B. 1980. Simon & Schuster's guide to shells (H.S. Feinberg, ed.). New York (NY): Simon & Schuster.

SAURIAU, P.-G. 1991. Spread of Cyclope neritea (Mollusca: Gastropoda) along the northeastern Atlantic coasts in relation to oyster culture and to climatic fluctuations, Marine Biology 109: 299-309.

SOFFER, O. 1989. The Middle to Upper Palaeolithic transition on the Russian Plain, in Mellars & Stringer (ed.): 714-42.

STINER, M.C. 1994. Honor among thieves: a zooarchaeological study of Neandertal ecology. Princeton (NJ): Princeton Univeristy Press.

N.d. The Upper Paleolithic faunas of Riparo Mochi. Monograph in preparation.

STINER, M.C., N. D. MUNRO & T.A. SUROVELL. In press. The tortoise and the hare: small game use, the Broad Spectrum Revolution and Paleolithic demography, Current Anthropology 41(1).

STINER, M.C., N.D. MUNRO, T.A. SUROVELL, E. TCHERNOV & O. BAR-YOSEF. 1999. Paleolithic population growth pulses evidenced by small game use, Science 283: 190-94.

TABORIN, Y. 1993a. La parure en coquillage au Paleolithique. Paris: CNRS. XXIXe supplement a Gallia Prehistoire.

1993b. Shells of the French Aurignacian and Perigordian, in Knecht et al. (ed.): 211-29.

TORNARITIS, G. 1987. Mediterranean sea shells (Cyprus). Nicosia: Proodos.

VILLENEUVE, L. DE, M. BOULE, R. VERNEAU & E. CARTAILHAC. 1906-19. Les grottes de Grimaldi (Baousse-Rousse): Geologie et paleontologie. Monte Carlo: Imprimerie de Monaco.

WHITE, R. 1982. Rethinking the Middle/Upper Paleolithic transition, Current Anthropology 23: 169-92.

1989. Production complexity and standardization of early Aurignacian bead and pendent manufacture: evolutionary implications, in Mellars & Stringer (ed.): 366-90.

1993. Technological and social dimensions of `Aurignacianage' body ornaments across Europe, in Knecht et al. (ed.): 277-300.

WHITE, R. & M. BISSON. 1998. Imagerie feminine du Paleolithique: l'apport des nouvelles statuettes de Grimaldi, Gallia Prehistoire 40: 95-132.

MARY C. STINER, Department of Anthropology, Building 30, University of Arizona, Tucson AZ 85721, USA. mstiner@u.arizona.edu

Received 8 April 1999 accepted 9 July 1999, revised 11 September 1999.