In this study, we investigate the foundations of the hydrated volcanic glass paleoelevation-paleoenvironment proxy by testing (1) the sensitivity of volcanic glass δD values to laboratory procedures, (2) the reliability of young glass reconstructions of observed meteoric water values, and (3) the paleoenvironmental reconstructions based on volcanic glass in comparison to those of other proxies. Volcanic glass δD values are insensitive to variations in laboratory pretreatments, displaying no significant shift in δD values with variations in duration of drying, rinsing, the temperature at which glass is dried, or with the grain size that is analyzed. However, clay adhering to the glass particles may skew δD results by ∼10‰, and HF acid pretreatment can cause significant isotopic shifts. Parent-water δD values calculated from young (≤0.3 Ma) glass δD values are in close accord with modern surface-water δD values across ∼3000 m of elevation in the central Andes. Combined glass δD values, pedogenic carbonate δ18O values, and clumped isotope paleothermometry from the 16.3 Ma Los Cristales tuff provide a key test of resistance to resetting of δD values in ancient glass hydration layers. These proxies indicate hydrated volcanic glass preserves its stable isotopic information over geologic time and that the climatic patterns of the high and dry Puna Plateau were already in place by 16.3 Ma.