ABSTRACT. Measurement of 14C in samples of wood of precisely known age have shown that the cosmogenic 14C content of the CO2 in the atmosphere has not remained constant during the period of time covered by radiocarbon dating. As the terrestrial atmosphere mixes with a time constant of less than 3 years, these variations must be essentially independent of geographic location. The 14C in atmospheric CO2 must be a quantity that, at a given time, pertains to the terrestrial atmosphere as a whole. Not only is its knowledge necessary for deriving accurate radiocarbon dates, but it is also valuable in connection with many geophysical and astrophysical problems. Unfortunately, progress in our knowledge of 14C variations in the terrestrial atmosphere has been delayed by hidden experimental errors in results obtained by many laboratories. By rigorous statistical analysis of the La Jolla results, it is now possible to show that the 14C variations are not simple random fluctuations but show distinct regularities. Similar patterns of variations have been found in the growth rate of trees during the last 5000 years. Measurements of radiogenic 10Be currently being done by European workers promise to conclusively elucidate the 14C findings.
THE DE VRIES VARIATIONSWillard F Libby, at the time when he and his co-workers published the first results of radiocarbon measurements around 1950(Libby, 1952, believed that the exchange of the carbon dioxide in the atmosphere with the bicarbonate of the oceans occurred quickly compared to all other changes in the ocean-atmosphere system. He, therefore, assumed that ocean and atmosphere can be considered as one and the same large carbon reservoir, reacting, as Libby put it, to changes in the 14C production rate "only extremely sluggishly." Libby considered variations such as those observed by H de Vries (1958) in the 14C content of wood from the 15th and 17th centuries AD to be within the experimental error of the measurements. However, I had been able to show (Suess, 1953) that the exchange rate of carbon dioxide between atmosphere and ocean was not sufficiently fast to exclude the possibility of variations of a few per cent, as indicated by the measurements of de Vries. In order to explain the results obtained by de Vries, Libby had also suggested the consideration of possible errors in the dendrochronologic dating that de Vries (pers commun, 1954) had used for his measurements. However, several other investigators used different tree-ring sequences to determine their sample ages. They at least qualitatively confirmed de Vries' results (Broecker, Olson & Bird, 1959;Willis, Tauber & Munnich, 1960;Ralph, 1959). As the cause of the 14C variations, these investigators considered changes in the 14C production rate through modulation of the cosmic ray flux by solar activity (see also Suess, 1958Suess, , 1968. This matter has been discussed from a more quantitative viewpoint by Stuiver (1961) and by Stuiver and Quay (1980 (Suess, 1970). This curve reached back to 5400 BC, as far back in tim...