ABSTRACT. Twenty-two radiocarbon activity measurements were made by accelerator mass spectrometry (AMS) on 2 Holocene stalagmites from Belgium (Han-stmlb) and from southwest France (Vil-stmlb). Sixteen thermal ionization mass spectrometric (TIMS) U/Th measurements were performed parallel to AMS analyses. The past dead carbon proportion (dcp) due to limestone dissolution and old soil organic matter (SOM) degradation is calculated with U/Tb ages, measured calcite 14C activity and atmospheric 14C activity from the dendrochronological calibration curves. Results show that the dcp is different for the 2 stalagmites: between 10,800 and 4780 yr from present dcp=17.5% (a=2.4; n=10) for Han-stmlb and dcp=9.4% (=1.6; n=6) between 3070 and 520 yr for Vil-stmlb. Despite a broad stability of the dcp during the time ranges covered by each sample, a slight dcp increase of about 5.0% is observed in the Han-stmlb sample between 8500 and 5200 yr.This change is synchronous with a calcite S13C increase, which could be due to variation in limestone dissolution processes possibly linked with a vegetation change. The dcp and S13C of the 2 studied samples are compared with 5 other modern stalagmites from Europe. Results show that several factors intervene, among them: the vegetation type, and the soil saturation leading to variable dissolution process systems (open/closed). The good correlation (R2=0.98) between the U/Th ages and the calibrated 14C ages corrected with a constant dcp validates the 14C method. However, the dcp error leads to large 14C age errors (i.e. 250-500 yr for the period studied), which is an obstacle for both a high-resolution chronology and the improvement of the 14C calibration curves, at least for the Holocene.