The summer 2001 flank eruption of Mount Etna ended with repeated basaltic ash explosions. The explosions were similar to small-scale Vulcanian blasts and probably originated when a plug of magma at the top of the conduit became overpressurized. We use three independent methods to estimate the overpressure in the plug: (1) To determine the pressure differential required to fragment the plug, we use a shock tube apparatus to heat and pressurize ballistic blocks at temperature T up to 900¿C and pressure P up to 25 MPa and rapidly decompress them to ambient conditions; (2) we apply a model to determine the overpressure generated by microlite growth at depth because pyroclasts from the explosions have 35 vol % more microlites than preceding Strombolian product; and (3) we relate the range of ballistic blocks to their exit velocity and, with models for Vulcanian eruptions, the exit velocity to pressure in the conduit. Experiments and ejecta range bracket the overpressure driving the explosions between ~3 and ~5 MPa: below 2.5 MPa the blocks would not reach the observed range, but above 5 MPa the existing blocks with 20% porosity would not survive the explosions. Crystallization of 35 vol % microlites at a depth of ~100 m increases the pressure in the plug to ~6 MPa above atmospheric, enough to drive the explosions, with a possible contribution from deeper volatiles. Combining the different methods, we also constrain the exsolved water at fragmentation at ~0.05 wt % and provide inferences on how the explosions came to an end.