Past Variations of Natural Radiocarbon as Recorded in U.K. Wood

Past variations in the natural radiocarbon content of the atmosphere over recent millenia have previously been documented through analyses of precisely dated organic materials, principally tree rings of bristlecone pine wood. Evidence for the causal factors of the secular variations has been obtained by correlation of the C profile with geophysical parameters such as the Earth's magnetic field, solar activity and climatic data. While the long-term trend in atmospheric 14C concentration and its origins are now firmly established, the existence of short-term fluctuations and the factors which could produce them are still in dispute. Bristlecone pine calibration of the radiocarbon timescale has produced a means of correcting archaeological data for past fluctuations, but this data record has subsequently been criticised, the main objections being that the 14C concentration of that wood may be atypical of world-wide C concentrations, as a result of its extreme growth conditions. Alternative dendrochronologies are at present being developed and subsequent analyses will prove or disprove the validity of the existing data. This thesis presents the preliminary results of a longterm project aimed at producing an absolute dendrochronology/ radiocarbon calibration system for the United Kingdom. Samples for this study come from the vast areas of submerged forests around the coast of the U. K. , and from archaeological sites. The principal features of the study - one of the first of this kind to make exclusive use of the liquid scintillation method - are, a) a very high sampling frequency (10 samples/century), b) abundant sample materials in perfect preservation and c) an emphasis on intercalibration and error assessment. The results of radiocarbon analyses of 5 "floating" chronologies, three oak and two pine, are presented. These clearly show that short-term fluctuations in atmospheric radiocarbon do exist, the variations being of the order of 2 - 3% in 40 - 50 years. The laboratory standardisation, by replicate analysis of an aged wood sample, indicates that the total error associated with a single analysis is largely a result of counting statistics, and thus these observed fluctuations are considered to reflect real changes in atmospheric 14C levels, rather than random experimental noise. Variations of this magnitude and time period have not previously been observed and it is believed that their detection in this study is a result of the enhanced sampling frequency employed. Statistical comparison of U. K, and American data has not shown reported systematic discrepancies between "normal" organic material and bristlecone pine wood but does suggest that, while longer-term trends in atmospheric levels are equivalent, the detailed structure of the C profile may be different. In addition to the major research programme described above, two short studies involving direct application of radiocarbon dating to archaeological problems were performed. The results of these have direct relevance to the limitations inherent in practical radiocarbon dating