Glassy microspherules collected from the plume of Etna volcano show a surprising compositional variability with SiO2 from 48 to 90 wt%, despite a very homogeneous basaltic source magma (SiO2 = 48%). This same variability is also observed in individual microspherules from their inner to outer parts. The most plausible hypothesis concerning these compositional variabilities is the interaction of the microspherules with the volcanic gases inside the crater prior to their atmospheric dispersal. In order to study this compositional variability we experimentally exposed basaltic glass to volcanic gases at Etna. The resulting glass composition documents a fundamental alteration mechanism, involving an apparent leaching process that can rapidly form nearly pure silica. Because microspherules erupted directly into the air from lava fountains remain unaltered, the duration of microspherules within a gas plume can explain the chemical variability of glass alteration. A major implication for the origin of glassy atmospheric particulates is that chemical compositions cannot be used as a reliable indicator of their source (e.g., meteoric, volcanic, anthropogenic).