Journal of Archaeological Science (2000) 27, 135–150 doi:10.1006/jasc.1999.0446, available online at http://www.idealibrary.com on Bone Diagenesis and Radiocarbon Dating of Fish Bones at the Shag River Mouth Site, New Zealand Fiona Petchey and Thomas Higham Radiocarbon Dating Laboratory, School of Science and Technology, University of Waikato, Private Bag 3105, Hamilton, New Zealand (Received 9 October 1998, revised manuscript accepted 15 March 1999) The Shag River Mouth site in southern New Zealand has yielded a series of radiocarbon determinations from multiple sample types, which suggest that the site was occupied briefly in the 14th century . Radiocarbon determinations of bone disclose a greater variation which has been attributed to problems of sample degradation, contamination, inadequate pretreatment or the use of sub-fossil bones in prehistory. We examined the dating of bone at the site using barracouta (Thyrsites atun) fish bone. In order to determine whether this material could be dated accurately, we applied a series of analytical tests to ascertain the preservation state and the extent of post-depositional contamination. These data were used to predict which samples would be most useful for 14 C analysis. The resultant ages, with the exception of two problematic samples, are consistent with those on other acceptable marine and estuarine shell carbonate samples. Based on the available results, we suggest that barracouta fish bone from this site is well preserved and reliable for dating. We re-examined previous determinations of moa and rat bone from different parts of the site to evaluate their accuracy in the light of this new data.  2000 Academic Press Keywords: FISH BONE, RADIOCARBON DATING, BONE PRETREATMENT, SHAG RIVER MOUTH, NEW ZEALAND PREHISTORY. Introduction S hag River Mouth is an archaeological site at the mouth of the Shag or Waihemo River near Palmerston in North Otago, New Zealand (Figure 1). The site covers about 2 ha of a large sand dune and spit, which extends some 600 m in a north- easterly direction from an outcrop of Tertiary rocks (McFadgen, 1996). It consists of large middens con- taining a wide spectrum of marine and terrestrial flora and fauna, dwelling remains and artefacts typical of the Archaic phase of New Zealand culture. Anderson & Smith (1996a: 281–2) have suggested that the site was occupied during a transitional period towards the end of this Archaic phase. At its margins, the site encroaches onto the edge of an estuary and adjacent Salicornia dominated mud flats (Anderson, Allingham & Smith, 1996)(Figure 2). In 1988, after a series of smaller test excavations, Anderson and Allingham directed a large research programme at the site (Anderson, Allingham & Smith, 1996). The most extensive excavation comprised one large 10 10 m square on the highest part of the dune (SM/C: Dune) and smaller test excavations along it and down onto the lower parts of the site (Figure 2). Almost 50 radiocarbon determinations were obtained (Anderson, Smith & Higham, 1996). The charcoal, marine and estuarine shell, and moa eggshell determinations suggested an occupation in the 14th century . The close statistical agreement between these results and the lack of significant hiatuses implied that prehistoric occupation was brief (Anderson, Smith & Higham, 1996). Bone determinations were excluded from chronological analysis due to variability in the results, which Anderson, Smith & Higham (1996) attributed either to a sub-fossil origin for the dated bone, inadequate pretreatment and/or differences in preservation state. More recently, Anderson (1996) obtained rat gelatin determinations from the site. All were at odds with previous results from identical stratigraphical locations and indicated that there may be significant problems with rat bone as a reliable dating medium. Beavan & Sparks (1997) suggested, on unstated grounds, that the rat bone samples from Shag River Mouth were degraded, but Anderson (1998) argued that they were not, citing Fourier transform infra-red analysis (FTIR) in evidence. This has also raised questions about previous rat bone radiocarbon determinations from natural sites, on the basis of which Holdaway (1996) inferred human contact with New Zealand by 2000 years . In this paper, we examine whether barracouta (Thyrsites atun) fish bone from Shag River Mouth is reliable for dating. The ability to date fish bone accurately offers a number of advantages to archaeolo- gists, not least that it is ubiquitous in archaeological 135 0305–4403/00/020135+16 $35.00/0  2000 Academic Press