10. ARCH. VOL.

17 (3ª) 11/9/08 20:00 Página 125

Archaeofauna 17 (2008): 125-138

Shell as a Raw Material: Mechanical Properties and Working

Techniques in the Tropical Indo-West Pacific

KATHERINE SZABÓ
Anthropology Program, College of Liberal Arts and Social Sciences
University of Guam, Mangilao 96923, Guam U.S.A.

(Received 21 October 2006; Revised 7 January 2008; Accepted 17 June 2008)

ABSTRACT: Considering the temporal depth and geographical breadth of shell-working with-
in the archaeological record, the lack of literature addressing identification and analytical pro-
tocols is somewhat surprising. As a starting point, this paper investigates the microstructural and
physical properties of three major taxa selected for artefact production in the tropical Indo-
Pacific region. Against this baseline information, the technological choices of ancient Pacific
shell-workers are assessed, leading to the conclusion that from at least the terminal Pleistocene,
shell-workers altered their techniques in accordance with the various structural properties of dif-
ferent types of shell.

KEYWORDS: SHELL ARTEFACTS, SHELL-WORKING, MICROSTRUCTURE, SHELL

FRACTURE, Turbo, Trochus, Conus

RESUMEN: Dada la amplitud temporal y dispersión geográfica que manifiesta el trabajo en

conchas por lo que atañe al registro arqueológico, resulta llamativa la ausencia de trabajos que
traten sobre la identificación de estos restos y sus protocolos analíticos. A modo de punto de par-
tida, este trabajo investiga las propiedades microestructurales y físicas de tres grandes taxones
seleccionados para la producción de utensilios en la región tropical indopacífica. Sobre esta
base se evalúan las opciones tecnológicas de los antiguos artesanos de la concha en el Pacífico
y se concluye que cuando menos desde el Pleistoceno terminal estos artesanos acoplaron su tec-
nología de acuerdo con las propiedades estructurales de los diferentes tipos de concha.

PALABRAS CLAVE: UTILLAJE EN CONCHA, TRABAJO EN CONCHA,

MICROESTRUCTURA, FRACTURACIÓN DE LA CONCHA, Turbo, Trochus, Conus
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 126

126 KATHERINE SZABÓ

INTRODUCTION assess technological decisions made by Late Pleis-

tocene - Mid Holocene shell workers in Island
In order to understand the working of a partic- Southeast Asia and the western Pacific in relation
ular raw material, there must be some understand- to these. Three major taxa consistently selected as
ing of the material itself. It is only within this con- raw materials are introduced together with a dis-
text that technological decisions can be fairly cussion of their microstructural properties. In addi-
understood, and an appreciation of culturally dis- tion to microstructure, standard and idiosyncratic
tinctive technological choices can be gained. features of shell growth and repair will also be
While stone in particular has received a great deal considered, as these further influence fracture
of attention as a raw material, there is scant mechanics. Drawn from a study of shell-working
archaeological literature discussing the structural practices evidenced at seventeen sites across
and mechanical properties of molluscan shell. Island Southeast Asia and the western Pacific (see
Indeed, many studies of shell artefact production Figure 1 and Table 1), the relationships between
have explicitly taken their lead from archaeologi- various shell taxa utilised as raw materials and dif-
cal understandings of lithic properties and tech- ferent working techniques are investigated.
nologies (e.g. Cleghorn, 1977; Smith, 1991).
Unfortunately, this has often resulted in an obfus-
cation of our understanding of shell-working tech- SHELLS UTILISED AS RAW MATERIALS
nologies and it is questionable how far conclusions FOR ARTIFACT PRODUCTION
such as the ‘crude and uninspired’ nature of shell-
working (Cleghorn, 1977: 241 for mid-Holocene
The tropical Indo-Pacific is the world’s most
Philippines) represent fair judgments.
speciose molluscan biogeographic region with
What follows is an attempt to investigate the over 10,000 species of mollusc. Despite this diver-
physical properties of shell as a raw material, and sity, however, only a small number of species are

FIGURE 1
Map showing region showing location of sites under discussion. The boundary between Near and Remote Oceania represents the lim-
its of Pleistocene colonisation, with Remote Oceania first being colonised c. 3300 B.P.
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 127

SHELL AS RAW MATERIAL: MECHANICAL PROPERTIES AND WORKING TECHNIQUES IN THE TROPICAL 127

Archaeofauna 17 (2008): 125-138

10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 128

128 KATHERINE SZABÓ

TABLE 1
* Full data on radiocarbon dates, along with calibrations, are presented in Szabó 2005.
** No sign of working on recovered fragments, however the presence of the raw material is incongruous with the nature of the midden.
In the case of Ille Cave, the midden is composed of fresh- and brackish-water species. A Pamwak Rockshelter, the midden is dominat-
ed by estuarine/mangrove species. In both cases, raw materials derive from coral reef habitats.
Island Southeast Asian and western Pacific sites from which data are drawn for this study including geographical location, chronologi-
cal positioning, shells used as raw materials and formal shell artefact types. Data taken from Best(n.d.), Szabó(n.d.), Fox (1970), Vermeij
(1978), Solheim et al. (1979), Fredericksen et al. (1993), Summerhayes (2000), Spriggs (2001), Wickler (2001), Szabó (2007), Szabó
et al. (2007).

consistently selected for artefact production with important raw materials over an extended time-
only minor spatial and temporal variation (Szabó, span, though the frequent selection of fossil or
2005) (see Table 1). Space prohibits a discussion sub-fossil specimens for working complicates any
of all of these raw materials, so the focus here will discussion of fracture properties and their relation-
be on three of the taxa with the widest spatio-tem- ships to working. For further information on giant
poral distribution: Trochus niloticus, Turbo mar- clam structure and associated working techniques,
moratus and Conus spp. (Figure 2). The various readers are referred to Moir (1990), Smith (1991)
members of the Tridacnidae (giant clams) are and Szabó (2005).
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 129

SHELL AS RAW MATERIAL: MECHANICAL PROPERTIES AND WORKING TECHNIQUES IN THE TROPICAL 129

FIGURE 2
The three major species under consideration; (a) Trochus niloticus (Gastropoda: Trochidae); (b) Turbo marmoratus (Gastropoda:
Turbinidae); (c) Conus leopardus (Gastropoda: Conidae).

MICROSTRUCTURE AND FRACTURE eration. These are nacreous, prismatic and cross-
lamellar microstructures.
Molluscan shell is composed of multiple layers,
the structure and composition of which vary from
taxon to taxon. The outermost layer is generally a TROCHIDAE AND TURBINIDAE
proteinaceous coating (periostracum), however in
some taxa, such as the cypraeids (cowries), this is Trochid and turbinid shells have a nacreous
absent (Watabe, 1988: 69). The calcareous layers (‘mother-of-pearl’) interior, and simple prismatic
below the periostracum are either calcitic or arag- exterior layer. Nacreous microstructures are
onitic forms of calcium carbonate (CaCO3), or a always constructed of aragonite, while prismatic
combination of both, and may occur in a variety in forms can be generated from either calcite or arag-
microstructures and combinations. Different onite. In the case of tropical trochids and turbinids,
microstructures can, and frequently do, occur the prismatic layer is aragonitic (Watabe, 1988:
within the same shell. 74). Prismatic structures have a higher organic
Whether aragonite or calcite, CaCO3 crystals content than either nacreous or crossed-lamellar
are laid down in a variety of defined microstruc- microstructures, and this organic component is
tural types. These types include nacreous, foliate, very strong, with prismatic structures being greater
prismatic, cross-lamellar, spherulitic and homoge- in tensile strength than crossed-lamellar
neous/granular structures (Wilbur & Saleuddin, microstructures (Currey, 1988: 191-192). The
1983: 257). I will discuss only three in further elongate prismatic crystals are aligned perpendic-
detail, as these are the microstructural arrange- ular to the periostracum, and this, combined with
ments of direct relevance to the taxa under consid- the high organic content, means that cracks gene-
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 130

130 KATHERINE SZABÓ

rally travel directly through the prismatic layer These features of microstructure and fracture
with little lateral dissipation (Currey, 1988: 191). mean that fractures are unpredictable within the
In gastropod nacreous microstructures, nacre is Trochidae and Turbinidae on a number of levels.
laid down as rounded and flattened crystals dis- Firstly, the prismatic and nacreous layers respond
tributed across the growing edge of the shell. As differently to force, and secondly, fractures in the
they are deposited, stacks form. These stacks nacreous structure itself are unpredictable. This
expand both horizontally and vertically with unpredictability of fracture would be accentuated
growth, and eventually coalesce to form the solid when force is considerable and suddenly applied
sheets (Wilbur & Saleuddin, 1983: 260-261). (as in direct percussion). This ‘impact force’ is sub-
Organic material is found within and between stantially less predictable and able to be controlled
stacks (Watabe, 1988: 77). than, say, compressive force (e.g. cutting/sawing,
where compression is coupled with friction) where
Nacre is one of the toughest molluscan structures
force is not as great or imparted solely at the brief
and has the ability to stop cracks better than pris-
moment of contact. Given that the stress that
matic forms (Currey, 1988: 186). Unlike prismatic
induces fracture is a function of force per area, the
structures, where fractures travel easily through the
blunter the percussive object the less successful
surface running adjacent to the elongate crystals,
and predictable the fracture is likely to be. Thus, if
the overlapping sheets combined with conical
force can be concentrated on a smaller area, either
stacks of differing diameters in nacreous structures
through the use of a more angular percussive
make the path of the fracture much more complex.
instrument or secondary percussion, the risk of
As with prismatic structures, cracks travel through
undesired breakage is lessened. In terms of work-
the organics rather than the crystals, but the non-lin-
ing trochid and turbinid shell, then, greater preci-
earity of such organics in nacreous structures dissi-
sion and control would be achieved through cut-
pate and spread the force (Figure 3a). What results
ting, sawing, grinding or indirect percussion.
is a very rough fracture at the micron level (Currey,
1988: 186). Nacre is tough when force is applied
perpendicular to sheet direction, however it is cer-
tainly not tough if force is applied in the same direc- CONIDAE
tion as sheet orientation. In the case of the latter, the
nacre will peel apart easily into aragonitic plates Cones exhibit a nearly wholly crossed-lamellar
(Figure 3b). Separation of sheets also occurs with structure and are aragonitic. The structure is made
the loss of organics through diagenetic processes – up of first, second and third-order lamella. The
a taphonomic headache frequently encountered by third-order lamellae are the smallest structural unit
archaeomalacologists. and are tiny elongate crystals oriented in the same

a b

FIGURE 3
3a: Schematic diagram of the path of a fracture through nacre running from the outer to inner surface of the shell. After Currey (1988:
188). 3b: Schematic diagram of the path of a fracture through nacre running parallel to the outer and inner surfaces of the shell. After
Currey (1988: 188).
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 131

SHELL AS RAW MATERIAL: MECHANICAL PROPERTIES AND WORKING TECHNIQUES IN THE TROPICAL 131

direction. These are stacked to form second-order (Watabe, 1988: 82). This means that compressive
lamellae, which are in turn stacked to form first- or tensile force at the margin of the bivalve or lip
order lamellae (Watabe, 1988: 82). What results of the gastropod will result in a fracture that runs
are long rod-like structures, with the crystals lying parallel to the growing edge and does not tend to
horizontally across the long axis of the first-order journey into the body of the shell. In Conus, this
lamellae (Figure 4). Between lamella, the direction property of breakage may be magnified by the
changes markedly, usually by about 70° to 90° presence of sculptural growth lines on the external
(Currey & Kohn, 1976: 1615). In Conus, there are surface of the shell, which also run parallel to the
three differently oriented cross-lamellar layers lip (Vermeij, 1993: 50). This type of fracture is
sandwiched together. The inner and outer layers thought to be related to predator defense. It is also
are angled transversely to the long axis of the worthy of note that Conus spp. reabsorb about
shell, while the inner layer is oriented in a parallel 25% of the inner whorls of the shell as they grow,
direction to the same axis (Currey & Kohn, 1976: which serves both to provide a larger internal cav-
1616-1617). The organic content in crossed-lamel- ity for the animal and free CaCO3 to thicken the
lar formations is rather low compared to prismatic body whorl (Vermeij, 1993: 33-34).
and nacreous microstructures, with thin mem-
branes surrounding lower-order lamellae (Watabe,
1988: 91; Dauphin & Denis, 2000: 376).
DIAGENETIC PROCESSES AND FRACTURE
As with nacre, crossed-lamellar structures offer PATTERNS
no clean line of fracture through the shell. While
cracks can travel with relative ease for a short dis-
With the death of the animal, the organic com-
tance, the direction of lamellae will suddenly
ponents of the shell begin to break down swiftly.
change. Thus, energy has to travel along the tortu- For nacreous shell, the breakdown is particularly
ous path of alternating lamellae, or break through fast due to the higher organic content. Experimen-
the stacks (Figure 5). The three differently orient- tal work with the gastropod Calliostoma ligatum
ed layers of crossed-lamellar formations men- (Gastropoda: Trochidae) has shown that after only
tioned above further complicate the nature of three days the shell loses about half its original
cracking. strength in compression (Vermeij, 1993: 50-51).
The angle at which the lamellae are laid down The loss of strength is less drastic for shells with a
in relation to the growing edge is important to note smaller organic fraction (Zuschin et al., 2003: 57).
with regards to fracture patterns. In bivalves, the With regards to tropical depositional contexts, Ver-
first order lamellae are oriented parallel to the meij’s (1993: 51) observation that organic decay is
shell margin in a concentric arrangement, and in faster in warmer temperatures has especial signifi-
gastropods they are arranged parallel to the lip cance (also Zuschin et al., 2003: 57). In the con-

FIGURE 4
Schematic rendering of a generalised cross-lamellar microstructure showing first-, second- and third-order lamellae. Adapted from Su
et al. (2004: fig. 1d).
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 132

132 KATHERINE SZABÓ

FIGURE 5
Scanning Electron Microscope micrograph of fracture in Conus. After Currey (1988).

text of shell as a raw material for artifact produc- fragments (Figure 6a). The finer work of creating pre-
tion, then, nacre will be more prone to splitting forms, for fishhooks, rings and perhaps further
between sheets (shearing), and will not preserve as unidentified formal artifacts, was carried out by deli-
well as shells with a low organic fraction and non- cately chipping back the shell with a sharp point using
organic-reliant matrix (e.g. the crossed-lamellar either an indirect percussion or pressure flaking tech-
Conidae and Tridacnidae). If a shell shows evi- nique (Figure 6b). Resulting preforms and manufac-
dence of having been collected as a raw material turing waste display a regular series of small indenta-
post-mortem, such diagenetic processes have a tions along the worked margins (Figure 6b and c).
bearing on the working process itself. Where evidence of further manufacturing stages exist,
the technique applied is invariably grinding (on a flat
grindstone), with freehand manual abrasion being
PATTERN AND VARIATION IN SHELL- applied to less-accessible parts of the artifact, includ-
WORKING TECHNIQUES ing the inner margins of rings and fishhooks (Figure
6b, c and d) (see also Szabó, 2007). Based on current
The Trochidae evidence, these working techniques, and the order in
which they are performed, are standard throughout
the Holocene across Island Southeast Asia and Near
Worked Trochus niloticus fragments and artifacts
Oceania.
are consistently present at low levels in palaeolithic,
neolithic, and Metal Age1 archaeological deposits. The application and order of these techniques
From worked pieces recovered, it is apparent that demonstrates an intimate knowledge of the nature
direct freehand percussion was only applied as a tech- of the raw material. As detailed above, Trochus
nique for the initial stages of reduction – effectively, niloticus shell is composed of two sandwiched
to break up the shell into desired and manageable microstructures (prismatic and nacreous) that have
different fracture properties and thus it is difficult
1
Island Southeast Asia, lacking a separate ‘bronze’ or ‘iron to shape Trochus niloticus through the application
age’, has a ‘Metal Age’ visible from c. 2,200 b.p. of direct freehand percussion alone. The tightly
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 133

SHELL AS RAW MATERIAL: MECHANICAL PROPERTIES AND WORKING TECHNIQUES IN THE TROPICAL 133

FIGURE 6
Working of trochid shells (Trochus niloticus and Tectus pyramis): 4a: Tectus pyramis shell reduced through the application of direct free-
hand percussion. Kamgot, New Ireland province, Papua New Guinea (Test-pit 1, Unit II/III, spit 9). 4b: Trochus niloticus shell reduced
through the application of direct freehand percussion. Kamgot, New Ireland province, Papua New Guinea (Test-pit 2, Unit II, spit 8).
4c: Trochus niloticus ring preform; the inner surface has been chipped into shape with a sharp point and the upper surface has been
ground. Batu Puti, Palawan Island, Philippines (Square 4, 60 cm below surface). 4d: Broken Trochus niloticus fishhook preform; note
grinding on the faces and chipping around the inner margins. Kamgot, New Ireland province, Papua New Guinea (Test-pit 2, Unit II,
spit 4). 4e: Fully ground and abraded broken Trochus niloticus fishhook. Site 13A, Grande Terre, New Caledonia (Zone II, G3: 32 cm).

focused application of force, in the form of chip- of undesired fracture. Cutting or sawing would
ping with a sharp point, allowed Trochus-workers have allowed similar – if not greater – levels of
to effectively reduce the raw material with less risk control, however evidence for the application of
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 134

134 KATHERINE SZABÓ

this technique is currently lacking2. It is not, how- Asia. The majority of associated formal artifacts
ever, lacking from the sequence of techniques utilizes the sturdy spire and/or shoulder sections of
applied to Turbo marmoratus (see below). the shell, and includes rings in varying diameters
It is notable that there is no evidence for the use and widths, beads, and ground and perforated
of Trochus niloticus specimens collected post- spires.
mortem. In light of the high organic content of Reduction sequences for all ring and spire arti-
both prismatic and nacreous structures, it is possi- facts involve the initial removal of most, or all, of
ble that this is an intentional choice to avoid unpre- the body whorl of the shell. This was achieved
dictable fracture and workability associated with a through the application of one of two techniques:
loss of shell structural integrity. direct percussion or sawing. When direct percus-
sion was used, the first blow was directed at the
centre of the body close to the lip. Given the ten-
The Turbinidae dency of Conus shells to split parallel to the lip as
part of their predator defense mechanisms (see
Working of the shell of Turbo marmoratus has a above), this first blow frequently results in a long
long history in the Island Southeast Asian and Near sliver of shell with one long edge being the sharp-
Oceanic regions (see Table 1), with little spatio- ened lip surface (Figure 8a). These fragments are
temporal variation in the selection and ordering of frequently referred to as ‘knives’ in post-excava-
reduction techniques. Direct percussion was only tion sorts. Successive blows to the body were then
used in the initial stages of reduction; firstly to applied concentrically around the shell surface
remove the thickened, irregular keel at the shoulder until the anterior portion of the shell came away
of the body whorl, and then to detach large sections (Figure 8b). Only the body whorl required such
of body whorl (Figure 7a-c). A combination of cut- treatment, as this whorl is considerably thicker
ting and/or sawing, and chipping with a sharp point than the internal architecture, which fractures eas-
was used to generate blanks and preforms (Figure ily (see above). The cross-lamellar microstructure
7d-e). Where curved outlines were desired, the means that, while fracture surfaces may be rough,
edge of the cutting implement was positioned per- fractures do not travel for any considerable dis-
pendicular to the shell surface, while straight edges tance into the shell. What seems to influence frac-
were achieved through sawing with the cutting ture more is the growth lines which run down the
edge parallel. Where formal artifact types in Turbo shell in an anterior-posterior direction.
marmoratus have been identified, they are invari- The ‘broad’ Conus bracelets frequently recov-
ably one-piece fishhooks. Finishing of the artifact ered from deposits associated with the Lapita cul-
was achieved through manual freehand abrasion. tural complex (Figure 8c), generally utilize 2-5 cm
As with Trochus niloticus, there is currently no evi- of the posterior body whorl in their construction.
dence for the use of Turbo marmoratus shell col- The risk of longitudinal splitting of the body whorl
lected post-mortem. associated with direct percussion appears to have
prompted a labor-intensive solution in their manu-
facture. Manufacturing waste from a series of
The Conidae Lapita sites indicates that the anterior and posteri-
or of the shell were separated by means of sawing,
The working of Conus spp. shells is strongly concentrically, around the body whorl (Figure 8d).
associated with the mid-late Holocene. Worked The application of compressive force and abra-
shell assemblages associated with the Lapita cul- sion, rather than impact force, made fracture con-
tural complex of the western Pacific are invariably siderably more predictable. It is likely that this
dominated by Conus-working, and conids assume greater investment of time accorded to broad
a far greater importance as a raw material with the Conus rings had a direct link to value, with these
materialization of the neolithic of Island Southeast artifacts being the only ornaments in the Lapita
shell artifact repertoire that were consistently –and
often repeatedly – curated (Szabó & Summer-
2
Smith (1991, 2001) has claimed cutting as part of the hayes, 2002; Szabó, 2005) (Figure 8c). The final
reduction sequence for Trochus niloticus fishhook manufacture,
stages of shaping and finishing, for all artifacts,
however this interpretation has not been reduplicated by recent
work (Szabó, 2005) that interpreted these relatively smooth include a mixture of grinding, chipping with a
edges as fracture along natural growth lines. sharp point, and freehand abrasion (Figure 8e-f).
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 135

SHELL AS RAW MATERIAL: MECHANICAL PROPERTIES AND WORKING TECHNIQUES IN THE TROPICAL 135

FIGURE 7
Working of Turbo marmoratus: 5a: Whole Turbo marmoratus shell with raised keel at shoulder knocked away by direct percussion.
Golo Cave, Gebe Island, eastern Indonesia (LM6: 210-220 cm). 5b: Detached segments of the shoulder keel of Turbo marmoratus. Kilu
Cave, Buka Island, Papua New Guinea (Test-pit 2, Layer I:2). 5c: Reduced Turbo marmoratus shell; reduced through direct percussion
and chipping with a sharp point. Golo Cave, Gebe Island, eastern Indonesia (LM6:195-200 cm). 5d: Chipped nacreous layer of the
shoulder section of a Turbo marmoratus body; ventral view. Kilu Cave, Buka Island, Papua New Guinea (Test-pit 2, Layer I:17). 5e:
Cut and chipped Turbo marmoratus body piece; possible rotating fishhook preform; dorsal view. Palandraku Cave, Buka Island, Papua
New Guinea (Test-pit 1, Layer V/VI:13).

In contrast to Turbo marmoratus and Trochus CONCLUSION

niloticus, there is consistent evidence for the selec-
tion of post-mortem specimens as raw materials, as The preceding examples of artifact manufac-
indicated by taphonomic signatures such as beach- ture in Trochus niloticus, Turbo marmoratus and
rolling and the action of marine bio-eroders and Conus spp. all demonstrate that technological
epibionts (Figure 8g). The relatively low organic choices are attuned to the specific structural and
content of cross-lamellar structures means that the physical properties of each taxon. This is not to
use of post-mortem shell offers no great disadvan- say, however, that raw materials are in any way
tage in itself. If, however, the shell has been struc- prescribing the way in which they are to be
turally damaged by organisms such as sponges worked. While different raw materials are more or
which bore into the shell, structural integrity will less suited to particular working techniques, there
be compromised in idiosyncratic ways. are always a number of ways to achieve a desired
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 136

136 KATHERINE SZABÓ

FIGURE 8
Working of Conus spp.: 6a: Lip of Conus litteratus or Conus leopardus detached through direct percussion. Kamgot, New Ireland
province, Papua New Guinea (Test-pit 2, Unit II, spit 4). 6b: Body whorl of a Conus litteratus detached through direct percussion.
Naigani, Fiji (3ext, level A4). 6c: Broken and curated Conus sp. broad ring fragment. St Maurice-Vatcha, New Caledonia (surface). 6d:
Body whorl of a Conus sp. shell detached through sawing. Nenumbo, south-east Solomon Islands (YZ-23, base of Grey layer 2). 6e:
Conus litteratus or Conus leopardus ring preform; note zone of repaired predator damage on exterior. Naigani, Fiji (9, level B3). 6f:
Ground and abraded Conus sp. small ring. Naigani, Fiji (4, Level A3 feature 4).6g: Water-rolled, ground and perforated Conus sp. spire.
Leta Leta, Palawan Island, Philippines (no provenance details).
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 137

SHELL AS RAW MATERIAL: MECHANICAL PROPERTIES AND WORKING TECHNIQUES IN THE TROPICAL 137

end. In this respect, it is noteworthy that cutting Stephen Wickler and Jo Lyn Gunness. Useful com-
and sawing are techniques frequently applied to ments on earlier drafts were provided by Philip
Turbo marmoratus but not Trochus niloticus, Piper, Marcela Lucero Juez, Irvy Quitmyer and an
despite the structural parallels. An understanding anonymous reviewer.
of the properties of raw materials is clearly impor-
tant, but technological choices are not reducible to
these variables alone.
REFERENCES
If technological practices and choices are
understood as having strong cultural elements,
BELLWOOD, P. 1997: Prehistory of the Indo-Malaysian
then the mapping of technological choices
Archipelago. University of Hawaii Press, Honolulu.
assumes a cultural – and culture-historical – sig-
nificance (Dobres, 2000; Szabó, 2005). While this BELLWOOD, P.; NITIHAMINOTO, G.; IRWIN, G.; GUNADI;
is not the venue for an in-depth assessment of the WALUYO, A. & TANUDIRJO, D. 1998: Bird’s Head
relevance of the above conclusions for issues in Approaches: Irian Jaya Studies - A Programme for
Interdisciplinary Research. In: Bartstra, G. (ed.):
Asia-Pacific culture history, it is worthy of note
35,000 years of Prehistory in the Northern Moluccas:
that the selection of, and technological approaches
233-275. A.A. Balkema, Rotterdam.
to, raw materials discussed in this paper [together
with others discussed within Szabó (2005)] shows BEST, S. (n.d.): Excavations at Site VL21/5 Naigani
great spatio-temporal constancy over the Island Island, Fiji: A preliminary report (1981). Manuscript
on file. Department of Anthropology, University of
Southeast Asian and Pacific regions. This lends lit-
Auckland.
tle support to neolithic population replacement
theories (e.g. Bellwood, 1997). However, neither CLEGHORN, P.L. 1977: A Note on Flaked Shell Imple-
does it entirely support arguments for an indige- ments: An Experimental Study. Asian Perspectives
20(2): 241-245.
nous genesis for the Near Oceanic neolithic (e.g.
Smith & Allen, 1999), as links in shell-working CURREY, J.D. 1988: Shell Form and Strength. In: Trueman,
practices between the western Pacific Islands and E.R. & Clarke, M.R. (eds.): Form and Function. The
Island Southeast Asia are consistently present Mollusca 11: 183-210. Academic Press, San Diego.
throughout, at least, the Holocene. Formal shell CURREY, J.D. & KOHN, A.J. 1976: Fracture in the
artefact types and methods of manufacture speak crossed-lamellar structure of Conus shells. Journal of
more of regional continuity and a shared history Materials Science 11: 1615-1623.
across the Island Southeast Asia/Melanesia bound- DAUPHIN, Y. & DENIS, A. 2000: Structure and composi-
ary. What is clear, however, is that in order to tion of the aragonitic crossed lamellar layers in six
effectively understand the form and production species of Bivalvia and Gastropoda. Comparative
techniques of shell artefacts, the limitations and Biochemistry and Physiology Part A 126: 367-77.
possibilities inherent in the raw materials must be DOBRES, M.A. 2000: Technology and Social Agency:
considered. Outlining a practice framework for archaeology.
Blackwell Publishers, Oxford.
FOX, R.B. 1970: The Tabon Caves. National Museum of
the Philippines, Manila.
ACKNOWLEDGEMENTS
FREDERICKSEN, C.; SPRIGGS, M. & AMBROSE, W.R. 1993:
Pamwak Rockshelter: A Pleistocene Site on Manus
This paper is drawn from research conducted as Island, Papua New Guinea. In: Smith, M.A.; Spriggs,
a part of my doctoral thesis, and thanks are due to M. & Fankhauser, B. (eds.): Sahul in Review: Pleis-
my supervisor Atholl Anderson and the Depart- tocene Archaeology in Australia, New Guinea and
ment of Archaeology and Natural History, Aus- Island Melanesia: 144-152. Department of Prehisto-
tralian National University for providing support ry, Research School of Pacific Studies, Australian
in this endeavour. Many people/institutions National University, Canberra.
allowed access to archaeological material drawn MOIR, B.G. 1990: Comparative Studies of ‘Fresh’ and
upon in this study, and thanks are due to Wilfredo ‘Aged’ Tridacna gigas Shell: Preliminary Investigations
Ronquillo, Eusebio Dizon, Angel Bautista, Hazel of a Reported Technique for Pretreatment of Tool Mate-
Ramirez, Corazon Alvina, Peter Bellwood, Glenn rial. Journal of Archaeological Science 17: 329-345.
Summerhayes, Roger Green, Wilhelm Solheim II, SMITH, A. 1991: Mollusc of the Ancient Mariner - Shell
Stuart Bedford, Christophe Sand, Simon Best, artefacts: typology, technology and Pacific prehisto-
Archaeofauna 17 (2008): 125-138
10. ARCH. VOL. 17 (3ª) 11/9/08 20:00 Página 138

138 KATHERINE SZABÓ

ry. Unpublished BA (hons) thesis. La Trobe Univer- SZABÓ, K. n.d.: Shell artefacts from the Lapita site of
sity, Bundoora. Bourewa, Viti Levu, Fiji. Unpublished report on file.
SMITH, A. 2001: A continuous sequence? Trochus shell University of the South Pacific, Suva.
artefacts in Lapita and post-Lapita assemblages in the SZABÓ, K. & SUMMERHAYES, G. 2002: Worked Shell Arte-
Arawe Islands, West New Britain Province, Papua facts - New Data from Early Lapita. In: Bedford, S.;
New Guinea. In: Clarke, G.R.; Anderson A.J. & Sand, C. & Burley, D. (eds.): Fifty Years in the Field -
Vunidilo, T. (eds.): The Archaeology of Lapita Dis- Essays in Honour and Celebration of Richard Shutler
persal in Oceania. Pandanus Books, Canberra. Jr’s Archaeological Career: 91-100. New Zealand
SMITH, A. & ALLEN, J. 1999: Pleistocene shell technolo- Archaeological Association Monograph 25, Auckland.
gies: Evidence from Island Melanesia. In: Hall, J. & SZABÓ, K.; SWETE KELLY, M.C. & PEÑALOSA, A. 2004:
McNiven, I. (eds.): Australian Coastal Archaeology: Preliminary Results from Excavations in the Eastern
291-297. Archaeology and Natural History, Aus- Mouth of Ille Cave, Northern Palawan. In: Paz, V.
tralian National University, Canberra. (ed.): Southeast Asian Archaeology: Papers in Hon-
SOLHEIM, W.G. II; LEGASPI, A.M. & NERI, J.S. 1979: our of Wilhelm G. Solheim II: 209-224. University of
Archaeological Survey in Southeastern Mindanao. the Philippines Press, Manila.
National Museum of the Philippines and the Univer- SZABÓ, K.; BRUMM, A. & BELLWOOD, P. 2007: Shell
sity of Hawaii, Manila. Artefact Production at 32,000 – 28,000 BP in Island
SPRIGGS, M. 2001: How AMS Dating Changed My Life. Southeast Asia: Thinking Across Media? Current
In: Jones, M. & Sheppard, P. (eds.): Australasian Con- Anthropology 48(5): 701-723.
nections and New Directions: 365-374. Department of VERMEIJ, G.J. 1978: Biogeography and Adaptation: Patterns
Anthropology, University of Auckland, Auckland. of Marine Life. Harvard University Press, Cambridge.
SU, X.-W.; ZHANG, D.-M. & HEUER, A.H. 2004: Tissue VERMEIJ, G.J. 1993: A Natural History of Shells. Prince-
Regeneration in the Shell of the Queen Conch Strom-
ton University Press, Princeton.
bus gigas. Chemistry of Materials 16(4): 581-93.
WATABE, N. 1988: Shell Structure. In: Trueman, E.R. &
SUMMERHAYES, G.R. 2000: Recent Archaeological Inves-
Clarke, M.R. (eds.): Form and Function. The Mol-
tigations in the Bismarck Archipelago, Anir - New
lusca 11: 69-104. Academic Press, San Diego.
Ireland Province, Papua New Guinea. Bulletin of the
Indo-Pacific Prehistory Association 19: 167-174. WICKLER, S. 2001: The Prehistory of Buka: A Stepping
Stone Island in the Northern Solomons. Department
SZABÓ, K. 2005: Technique and Practice: Shell-Working
of Archaeology and Natural History and the Centre
in Island Southeast Asia and the Western Pacific.
for Archaeological Research, Canberra.
Unpublished PhD thesis. Department of Archaeology
and Natural History, Australian National University, WILBUR, K.M. & SALEUDDIN, A.S.M. 1983: Shell For-
Canberra. mation. In: Saleuddin, A.S.M. & Wilbur, K.M. (eds.):
SZABÓ, K. 2007: Shell fishhooks of the Lapita cultural Physiology (Part 1). The Mollusca 4: 235-287. Aca-
complex: An assessment of types, manufacturing demic Press, New York.
protocols, distributions and relationships. In: Ander- ZUSCHIN, M.; STACHOWITSCH, M. & STANTON, R. J. Jr
son, A.; Leach, B.F. & Green, K. (eds.): Vastly Inge- 2003: Patterns and processes of shell fragmentation
nious: The Archaeology of Pacific Material Culture: in modern and ancient environments. Earth-Science
227-241. University of Otago Press, Otago. Reviews 63: 33-82.

Archaeofauna 17 (2008): 125-138