White says that there remains a “fundamental archaeological problem” in the Neolithic and Bronze Age chronologies published in our paper. She says that dates from Ban Chiang and Ban Non Wat do not support a conclusion that there is similarity between results, not only from the same archaeological cultural periods but across sites. We refer the reader to our Figure 8 which shows the opposite and summarizes all of the key boundary information marking the dates of the transition to the Bronze Age. These data are based on Bayesian models incorporating the 160+ dates from the 5 sites that we have worked on. The boundaries overlap one another between 1200-1000 BC (at 68.2% probability). There is also good agreement in our results for the initial Neolithic settlement of Non Nok Tha, Ban Non Wat and Ban Chiang whether the determinations come from charcoal, shell or human bone. In addition, the transition from the late Bronze Age into the Iron Age is consistent between sites.

White ignores this profusion of data and focuses her attention instead on just one date from Ban Non Wat (OxA-11722). She compares this result with two of her own dates; CAMS-14264 and AA-15578, to demonstrate an apparently divergent picture from the one we have produced. CAMS-14264 is a date on phytolith carbon, which we have already shown to be unreliable and experimental at best. OxA-11722 is a clear outlier within the Ban Non Wat sequence as we show here in our supplementary figure (http://figshare.com/s/9d3...). It is 100% likely to be an outlying result. The sample comes from a fragment of charcoal in the upper part of the fill of Burial 28, a massive pottery vessel that contained the seated skeleton of a man. This person had been interred with a bivalve shell in his hands. We dated this shell at 3170±27 BP, a result matched by the other results for the initial Neolithic at this site as shown in the figure. OxA-11722 is either the result of inbuilt age of charcoal or is residual material from an earlier phase. This shows how necessary it is to obtain multiple determinations from a site sequence rather than rely on 2-3 results. In the supplementary figure shell and charcoal determinations are shown and one can see the good levels of agreement between them (charcoal dates are in blue outline and shell carbonate determinations are in black outline), with the exception of OxA-11722 and -13467.

This does not constitute a ‘fundamental problem’, as White suggests, instead it reflects the biased selection of one determination that supports her view and the concomitant rejection of the much larger dataset that disagrees with it. The only fundamental problem that we can see here is bad science on the part of White.

T. Higham & K. Douka
Oxford Radiocarbon Accelerator Unit