Vadose cements in the Late Pleistocene Miami Limestone contain regions with two-phase aqueous fluid inclusions that have consistent vapor to liquid (V-L) ratios. When heated, these seemingly primary inclusions homogenize to a liquid phase in a range between 75??C and 130??C (mean = 100??C) and have final melting temperatures between -0.3?? and 0.0??C. The original distribution of Th was broadened during measurements because of fluid inclusion reequilibration. The narrow range of Th in these fluid inclusions suggest unusually consistent V-L ratios. They occur with small, obscure, single phase liquid-filled inclusions, which infer a low temperature origin (less than 60??C), and contradict the higher temperature origin implied by the two phase inclusions. The diagenetic environment producing these seemingly primary fluid inclusions can be inferred from the origin of the host calcite enclosing them. The ??18O composition of these cements (-4 to-5.5%., PDB) and the fresh water in the fluid inclusions are consistent with precipitation from low-temperature meteoric water. The carbon-isotope composition of the vadose cements that contain only rare two-phase fluid inclusions are comparable to the host rock matrix (??13C between 0 and +4%., PDB). Cements that contain common two-phase fluid-inclusions have a distinctly lighter carbon isotopic composition of -3 to -5%.. The carbon isotope composition of cements that contain common two-phase inclusions are about 6%. lighter than those of other vadose cements; models of early meteoric diagenesis indicate that this is the result of precipitation from water that has been influenced by soil gas CO2. Our hypothesis is that the primary fluid inclusions, those with consistent V-L ratios and the single-phase liquid inclusions, form at near-surface temperature (25??C) and pressure when consistent proportions of soil gas and meteoric water percolating through the vadose zone are trapped within elongate vacuoles. This study corroborates that Th measurements on two phase inclusions in vadose cements can be misleading evidence of thermal diagenesis, even if the measurements are well grouped. ?? 1988.