The trace-element geochemistry of speleothems is becoming increasingly used for reconstructing palaeoclimate, with a particularemphasis on elements whose concentrations vary according to hydrological conditions at the cave site (e.g. Mg, Sr, Baand U). An important step in interpreting trace-element abundances is understanding the underlying processes of their incorporation.This includes quantifying the fractionation between the solution and speleothem carbonate via partition coefficients(where the partitioning (D) of element X (DX) is the molar ratio [X/Ca] in the calcite divided by the molar ratio [X/Ca] in theparent water) and evaluating the degree of spatial variability across time-constant speleothem layers. Previous studies of howthese elements are incorporated into speleothems have focused primarily on stalagmites and their source waters in naturalcave settings, or have used synthetic solutions under cave-analogue laboratory conditions to produce similar dripstones. However,dripstones are not the only speleothem types capable of yielding useful palaeoclimate information. In this study, weinvestigate the incorporation of Mg, Sr, Ba and U into a subaqueous calcite speleothem (CD3) growing in a natural cave poolin Italy. Pool-water measurements extending back 15 years reveal a remarkably stable geochemical environment owing to thedeep cave setting, enabling the calculation of precise solution [X/Ca]. We determine the trace element variability of ‘modern’subaqueous calcite from a drill core taken through CD3 to derive DMg, DSr, DBa and DU then compare these with publishedcave, cave-analogue and seawater-analogue studies. The DMg for CD3 is anomalously high (0.042 ± 0.002) compared to previousestimates at similar temperatures (~8 °C). The DSr (0.100 ± 0.007) is similar to previously reported values, but datafrom this study as well as those from Tremaine and Froelich (2013) and Day and Henderson (2013) suggest that [Na/Sr] mightplay an important role in Sr incorporation through the potential for Na to outcompete Sr for calcite non-lattice sites. DBa inCD3 (0.086 ± 0.008) is similar to values derived by Day and Henderson (2013) under cave-analogue conditions, whilst DU (0.013 ± 0.002) is almost an order of magnitude lower, possibly due to the unusually slow speleothem growth rates(<1 lm a-1), which could expose the crystal surfaces to leaching of uranyl carbonate. Finally, laser-ablation ICP-MS analysisof the upper 7 lm of CD3, regarded as ‘modern’ for the purposes of this study, reveals considerable heterogeneity, particularlyfor Sr, Ba and U, which is potentially indicative of compositional zoning. This reinforces the need to conduct 2D mappingand/or multiple laser passes to capture the range of time-equivalent elemental variations prior to palaeoclimate interpretation.