Limpet shells as unmodified tools in Pleistocene Southeast Asia: an experimental approach to assessing fracture and modification Katherine Szab  o * , Brent Koppel Centre for Archaeological Science, University of Wollongong, Wollongong NSW 2522, Australia article info Article history: Received 13 September 2014 Received in revised form 12 November 2014 Accepted 13 November 2014 Available online 29 November 2014 Keywords: Limpet shells Scutellastra flexuosa Shell tools Pleistocene Southeast Asia abstract Pleistocene tools manufactured in shell are rarely identified. This may in part be due to the complexity of shell as a raw material and associated challenges in recognising and interpreting shell modification. A series of unusually-shaped Scutellastra flexuosa limpets from c. 30,000 year old deposits in Golo Cave, eastern Indonesia were identified as putatively modified during midden analysis. A pilot programme of investigations into the microstructure and natural fracture patterns of this species, coupled with a series of use-wear experiments, demonstrates that some S. flexuosa shells were used as scrapers. The shells were used in unmodified form and were ‘repurposed’ after having been gathered for subsistence pur- poses. Taken together with other forms of early shell-working already reported for Golo Cave, the identification of these new unmodified shell tools expands the corpus of shell tool use at the site and presents a picture of diversity and complexity not seen in the associated lithic assemblage. © 2014 Elsevier Ltd. All rights reserved. 1. Introduction Molluscan shell used as a raw material for tool production is uncommonly noted for Pleistocene archaeological assemblages in different global locations. Active modification of molluscs for use e typically as scrapers e has been noted for a number of coastal Mousterian sites in Italy and Greece (Douka and Spinapolice, 2012; Stiner, 1994). Choi and Driwantoro (2007) argued for the use of shell knives by Javanese Homo erectus, however fundamental problems with the argument (outlined in Szab o, 2013) means that these tools will be discounted here. For Homo sapiens, the delib- erate modification of the large, calcareous opercula of Turbo mar- moratus was reported at c. 32,000 e 28,000 BP for eastern Indonesia (Szab o et al., 2007), and preliminary observations of other contemporaneous types of shell working and modification at the same site were also made. It is one of these other putative forms of shell tool e the use of unmodified Scutellastra flexuosa limpets as scrapers e that we investigate here. The behaviour of shell as a raw material for artefact production is poorly understood. Different molluscan families can have pro- foundly different microstructures, and thus structural properties (Currey and Taylor, 1974; Watabe, 1988). Taphonomic processes, including recrystallization of the calcium carbonate building blocks and the degradation of the protein fraction which composes the matrix in which crystals are enmeshed, further complicate the re- sponses of shells to force. Given these various properties, diagnostic patterns of fracture in lithic materials are a poor analogue for un- derstanding the behaviour of shell, and new methods through which to recognise and interpret shell-working and modification are needed. Using a programme of experimental use-wear replication, as well as investigations into the structural nature and taphonomic proclivities of S. flexuosa limpets, we pilot a technique to help discern between (1) natural breakage and taphonomic modifica- tion, and (2) cultural patterns of surface and edge modification through use of the shells as unmodified tools. If demonstrated, ~30,000 year old use of shell as a raw material would include not only the focussed gathering and structured reduction of T. mar- moratus opercula (Szab o et al., 2007), but the repurposing of shell midden refuse for use as expedient tools. 2. Background to the site and sample Golo Cave is limestone solution cave located on Gebe Island, northern Maluku, eastern Indonesia (see Fig. 1). Excavations of a 5 Â 1 m trench and two separate 1 Â 1 m squares to bedrock in 1994 and 1996 by Bellwood and colleagues uncovered stratified archaeological deposits spanning from 32 kya just above bedrock to * Corresponding author. Tel.: þ61 2 4221 5846. E-mail address: kat@uow.edu.au (K. Szab o). Contents lists available at ScienceDirect Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas http://dx.doi.org/10.1016/j.jas.2014.11.022 0305-4403/© 2014 Elsevier Ltd. All rights reserved. Journal of Archaeological Science 54 (2015) 64e76