Radiocarbon analysis of a novel bone sample type: snapper and barracouta bone from New Zealand archaeological sites

Abstract

This dissertation investigates the potential of fish bone, specifically barracouta (Thyrsites atun) and snapper (Pagrus auratus), for routine radiocarbon analysis. Of particular interest to this study is the perceived reliability of bone ¹⁴C determinations in the New Zealand archaeological literature. Issues of bone preservation and contamination, dietary fractionation, species differences, and the reliability of different pretreatments are of key importance. Informed, critical assessment of bone determinations in New Zealand is, however, currently limited for a number of reasons. First, there have been no, or few comprehensive tests of bone pretreatment, species reliability or the influence of contamination. Second, confusion has resulted regarding the effectiveness of the varied radiocarbon pretreatments available, due in part to the complexity of some methods. This has been further hampered by the incorrect reporting of fractions isolated for ¹⁴C measurement. Third, inadequate sample selection procedures have resulted in burnt bone, sub-fossil bone and severely degraded bone, each with intrinsically different chemistries, being submitted for radiocarbon assay. Fourth, insecure provenance of the sample, or associated samples to be dated, mean that few comparisons of bone reliability could have, or can be made. Fifth, publication of results and procedures have been limited. Finally, there has been limited research into the radiocarbon measurement of bone in New Zealand due to a preconception about the reliability of bone determinations. These uncertainties with bone radiocarbon measurement are addressed in this dissertation.

The sources of ∆¹⁴C to fish are also critical. Bone collagen and its derivatives (e.g. gelatin, tripeptides, amino acids) are the most common bone fractions isolated for radiocarbon measurement. For fish, the ¹⁴C in collagen is derived either directly, or indirectly via diet, from dissolved inorganic carbon in sea water. A number of uncertainties have previously been identified with the measurement of ¹⁴C in marine animals, including the effects of hardwater, depleted carbon from depth, or terrestrial organic carbon. These factors are investigated, and it is determined that their impact on fish bone ¹⁴C determinations depend on the immediate environment and the type of fish selected for dating. Both barracouta and snapper occupy predominantly the well-mixed surface waters around New Zealand. Studies of marine shell and snapper otolith carbon from around New Zealand, suggest that hardwater and terrestrial organic carbon are of limited influence. This requires further testing and may depend on the specific dietary preferences of each species. Of particular concern, however, is the upwelling of depleted carbon, especially at the convergence of inshore and offshore waters or at the boundary of two water masses (i.e. the Subtropical Convergence). In addition, inbuilt age is often cited as reasons for anomalous ¹⁴C determinations. This also appears to be of nominal influence, due in part to the precision of the radiocarbon technique, but also because of the relatively rapid replacement of collagen.

A range of different bone pretreatment and assessment methods are discussed. From this review it is concluded that gelatinisation following a NaOH wash, should remove better than 8% contamination (i.e. a maximum error of 42 years in a well-preserved sample of 900 BP material if contaminated by modem carbon). This pretreatment method, in combination with techniques for assessing contamination and collagen preservation, will significantly improve the likelihood of a reliable radiocarbon determination. The use of an appropriate assessment methodology also provides a wealth of information about the sample and site taphonomy.

Snapper and barracouta samples from 7 archaeological sites are analysed. Fourier Transform Infrared (FTIR) spectra, stable isotope values, N% determinations and yield data, obtained prior to and during pretreatment, are presented along with ¹⁴C information. Material from Pleasant River, Long Beach, Shag River Mouth, Twilight Beach, Houhora and Rotokura is identified as well-preserved. Bone removed from Tata Beach is of suspect preservation state due to low levels of remaining collagen, and contamination identified in the FTIR spectra of the acid-insoluble fraction.

Radiocarbon results from Tata Beach are, however, variable, and largely inconclusive at the level of precision used. The remaining 6 sites produced reproducible fish gelatin determinations that are statistically indistinguishable from associated marine shell determinations, and in chronological agreement with charcoal samples after calibration using the marine calibration curve of Stuiver and Braziunas (1993), corrected according to the New Zealand reservoir value (∆R) (Higham and Hogg 1995). Determinations on purified tripeptides do not agree with these results. These are, however, not supported by archaeological evidence or associated ¹⁴C determinations on reliable charcoal and shell samples.

An analysis of 46 New Zealand archaeological sites with bone determinations (human, dog, fish, seal, and moa) obtained over a 40 year period is undertaken and a discard protocol applied according to the results of this dissertation. Eleven sites with bone determinations remain after application of the discard protocol. Not all are statistically indistinguishable from associated charcoal and marine shell pairs. Using geographic and climatic information, as well as intra site data, it is apparent that sites with problematic bone radiocarbon estimates are located in high rainfall areas, which have resulted in bones of poorer preservation. This, in combination with inadequate pretreatment (typically an acid wash, or acid/alkali/acid pretreatment) has resulted in some erroneous determinations.

In the light of these results it is suggested that a bone selection protocol, using a range of chemical methods, needs to be implemented in order to identify problematic samples prior to ¹⁴C analysis. Further, gelatinisation should be a minimum pretreatment. Bones <1000 years of age should be viewed with caution for radiocarbon dating when less than 40% “extractable collagen” remains, and where contamination is identified in the FTIR spectra of the acid-insoluble fraction. The adoption of complex and expensive biochemical purification techniques is not recommended.