Results of extensive and systematic radiocarbon dating of tufa samples and associated materials (wood, plants, moss, and peat) are presented in this paper. Samples were collected in Plitvice Lakes National Park, Central Croatia, Yugoslavia, where intensive precipitation of tufa has taken place over several geologic epochs in a typical Karst area. Problems associated with radiocarbon dating of tufa are discussed in a previous paper (Srdoč et al, 1980). The authors concluded that dating of tufa is feasible and reliable, at least in the specific case of Karst regions, where retention of groundwater is short. Tufa is precipitated as a result of decomposition of groundwater bicarbonates due to loss of CO2, usually in waterfalls and cascades, and it is practically contemporaneous with living organic material. However, radiocarbon activity of modern tufa is lower than that of organic material due to dissolution of inactive carbonates (limestone) in groundwater. The initial activity of groundwater depends on the geology and vegetation of the area and it is widely discussed in the literature (Geyh, 1973). The initial activity, Ao, of tufa could be easily measured on recent samples, had not bomb-test contamination upset the equilibrium. Namely, while the atmospheric CO2 and living matter responded relatively quickly to the injection of radiocarbon in the atmosphere, groundwater and, consequently, tufa showed a much slower and damped response. Therefore, we have based our estimation of the initial activity of tufa on measurements of samples of known age collected before the bomb-test era. An average value of the initial activity equal to 85% of modern has been obtained which is in accordance with data for groundwater activity in similar areas. Consequently, all radiocarbon ages of tufa samples in this paper are based on the Libby half-life of 14C (5568 years) and the initial radiocarbon activity equal to 85% of modern. The modern standard is 0.95 of the activity of NBS oxalic acid. Errors quoted are based on counting statistics only and do not include uncertainties in 14C half-life and in initial activity. The latter may introduce much larger errors, up to ± 1000 years, depending on the sampling site, as explained in the following sections. Whereas radiocarbon dating of Holocene tufa beds give reliable ages within the above specified errors, radiocarbon dating of old tufa beds is still questionable. Intrusion of minute quantities of modern calcareous material renders the ages of old (∼30,000 yr) tufa useless. Research on measuring the age of old tufa deposits by other methods is in progress.