Extra paper
RADIOCARBON AND DENDROCHRONOLOGY 
Peter Ian KUNIHOLM & Maryanne NEWTON
Aegean Dendrochronology Project, B-48 Goldwin Smith Hall, Cornell University, Ithaca NY 14853-3201, USA.
PETER@dendro.mail.cornell.edu / MARYANNE@dendro.mail.cornell.edu
Radiocarbon calibration and wiggle-matching
The process of converting radiocarbon years to absolute (calendar) years depends on calibration of the radiocarbon time scale. That the amount of radiocarbon in the atmosphere has not been constant over time was recognised by Hans Suess and Walter Libby in the 1950s, but the earliest calibration curves based on absolutely dated tree-rings were not available until the 1970s. Since then, concerted international efforts to deal with issues of interlaboratory offsets, time-dependent regional differences, species differences, etc., have led to updates and revisions. The currently recommended curve for terrestrial samples within the range of tree-ring calibration (back into the 10th millennium cal BC) is INTCAL98, the curve employed in this exercise.
The result of tree-ring correction of the radiocarbon time scale is a non-monotonic - 'wiggly' - calibration curve. A single radiocarbon date does not, therefore, have a single corresponding calendar date. Additional uncertainty is added when the 14C error associated with the counting procedures of the radiocarbon lab translates into a much larger BC error range, depending on the nature of the calibration curve at the given date. In the Çatalhöyük 'wiggle-match' example highlighted here, this is made explicit by the error margins on both the Y- (14C) and X- (BC) axes. Such uncertainties are implicit in all the dates plotted on the calibration curve here (from the CANeW database). How does the investigator choose the mostly likely BC date given the broad ranges of possibilities for any single radiocarbon date?
One answer is dendrochronological 'wiggle-matching', a procedure in which knowledge of the exact number of years between 14C samples (in these examples, all 10-year samples) permits fixing of the X-axis for a given set of dates. This set of dates (vs. a single date) can only have one best fit (or at minimum a narrow range) against the wiggles of the calibration curve (see the examples from Çatalhöyük, Hacilar, Canhasan I, and Kösk Höyük). Another procedure known as archaeological wiggle-matching can approximate control of the calendar time scale if assumptions are made about lengths of occupation, etc. (for one example, see Newton and Kuniholm 1999, in which the length of occupation of some units at Çatalhöyük was approximated by testing James Mellaart's assumption that layers of wall plaster may represent annual applications).
In much of Europe north of the Alps and in the American southwest, absolute dendrochronological dates have been available for decades. But in the eastern Mediterranean, until a continuous tree-ring chronology from living trees back into prehistory becomes available, we hope that radiocarbon dating, archaeological wiggle-matching, and dendrochronological wiggle-matching can help to refine the broad span of dates that fill the radiocarbon calibration curve pictured here (Fig. 1). Our aim, ultimately, is to illuminate the temporal landscape of early settled life in terms of absolute (calendrical) age, one connected to the human life-spans of the people we study.
Toward the absolute dating of human settlement in Central Anatolia from the 9th through 5th millennia cal BC
Since 1989, when David French and the British Archaeological Research Institute in Ankara made charcoal samples from the excavations at Canhasan available, the Aegean Dendrochronology Project at Cornell University has been conducting dendrochronological analyses of charcoal samples from Neolithic and Chalcolithic sites in Central Anatolia. Our work is constrained by the availability of wood and charcoal samples from long-lived tree species (common in Anatolia are Juniperus, Pinus, Cedrus and Quercus), and so we report here on only a handful of the many sites that have been documented and investigated (compare the graphs above). The results from our analyses can be combined with a range of radiocarbon results from both other material (non-wood) from the same sites, and material from other sites in this time period (see the 14C charts compiled by Laurens Thissen and the CANeW team, Appendix I, this volume) to span some five millennia, providing a firm basis for understanding the temporal landscape - and great time depth - that framed the development of settled life on the Anatolian plateau. Also, since the climatological zone within which we work extends from the western Appennines in Italy to the southern Caucasus in eastern Georgia, we can look forward to expanding to the nexus of dendrochronologically investigated sites beyond Central Anatolia, and thus add a controlled temporal component to important questions about external relations.