A 4-year study in a central Texas cave quantifies multiple mechanisms that control dripwater composition and how these mechanisms vary at different drip sites. We monitored cave-air compositions, in situ calcite growth, dripwater composition and drip rate every 4-6weeks. Three groups of drip sites are delineated (Groups 1-3) based on geochemical variations in dripwater composition. Quantitative modeling of mineral-solution reactions within the host carbonate rock and cave environments is used to identify mechanisms that can account for variations in dripwater compositions. The covariation of Mg/Ca (and Sr/Ca) and Sr isotopes is key in delineating whether Mg/Ca and Sr/Ca variations are dictated by water-rock interaction (i.e., calcite or dolomite recrystallization) or prior calcite precipitation (PCP). Group 1 dripwater compositions reflects a narrow range of the extent of water-rock interaction followed by varying amounts of prior calcite precipitation (PCP). Group 2 dripwater compositions are controlled by varying amounts of water-rock interaction with little to no PCP influence. Group 3 dripwater compositions are dictated by variable extents of both water-rock interaction and PCP. Group 1 drip sites show seasonal variations in dripwater Mg/Ca and Sr/Ca, whereas the other drip sites do not. In contrast to the findings of most previous dripwater Mg/Ca-Sr/Ca studies, these seasonal variations (at Group 1 drip sites) are independent of changes in water flux (i.e., rainfall and/or drip rate), and instead significantly correlate with changes in cave-air CO2 concentrations. These results are consistent with lower cave-air CO2, related to cool season ventilation of the cave atmosphere, enhancing calcite precipitation and leading to dripwater geochemical evolution via PCP. Group 1 dripwater Mg/Ca and Sr/Ca seasonality and evidence for PCP as a mechanism that can account for that seasonality, have two implications for many other regions where seasonal ventilation of caves is likely: (1) speleothem trace-element records may provide seasonal signals, and (2) such records may be biased toward recording climate conditions during the season when calcite is depositing. Additionally, we use our results to construct a forward model that illustrates the types of speleothem Mg/Ca and Sr/Ca variations that would result from varying controls on dripwater compositions. The model provides a basis for interpreting paleo-dripwater controls from high frequency Mg/Ca and Sr/Ca variations for speleothems from caves at which long term monitoring studies are not feasible. ?? 2011 Elsevier Ltd.