Wilson and Keil (1991) argued that a few hundred ppm of expanding volatiles present in early partial (basaltic) melts on the aubrite parent body (or on other differentiated asteroids < 100 km in radius), upon ascent of the magma to the surface of the body, would cause disruption into a spray of droplets moving with velocities in excess of the escape velocities of small asteroidal-sized bodies and thus escape and be lost into space 4.55 Ga ago. Hence, no such basaltic rocks exist as individual meteorites nor as clasts in brecciated aubrites. We report measurements of volatiles in enstatite chondrites, considered to be reasonable analogs of the precursor rocks of the aubrites (Qingzhen: EH3; Indarch, Abee: EH4; St. Mark's: EH5; Hvittis, Pillistfer: EL6, sinoite-bearing; Khairpur: EL6, no sinoite) and address in detail the nature and origin of volatile phases driving this pyroclastic volcanism. We find that 1. (1) Volatiles measured by high-temperature mass spectrometric degassing and released at <780°C are contaminants from terrestrial weathering, handling and cutting; those released at >925°C (principally CO, 1580–2830 ppm; N2, 110–430 ppm; Cl, 120–450 ppm; S, see discussion for abundance) are indigenous. The precise host phases and host sites of the indigeneous volatiles are uncertain; 2. (2) N2 release appears to be associated with the melting of the Fe,Ni-sulfide cotectic at 950°C. The amounts of gas released are not unlike those previously measured for iron meteorites and sulfides in irons, but exceed by several orders of magnitude the equilibrium N2 solubility in metallic Fe,Ni for solar nebula pressures. This suggests that either partial pressures during N2 incorporation were much higher than solar values or N2 is trapped in lattice defects and microcracks, possibly in sulfides; 3. (3) CO release over a narrow temperature interval of 1025–1255°C is violent and results in pronounced vesiculation of the residue. CO may be trapped in microvesicles and cracks in silicates or, less likely, may form by a reduction reaction during heating between carbon (graphite) and the minor ferrosilite in the enstatite; 4. (4) Cl release beginning at 950 (EH's) and 1025°C (EL's) appears not to be associated with lawrencite, which should sublime at T as low as 325°C; in EH's, some may be released from djerfisherite; 5. (5) Gas release and textural studies suggest that metamorphic reheating of EH's and EL's was to less than about 950°C. EL's did not form during metamorphic reheating of EH's by draining of cotectic Fe,Ni-sulfide liquid; 6. (6) Sinoite present in some EL6's did not form by condensation from the solar nebula, but during parent body metamorphism at T < 950°C via the reaction of SiO2, gaseous N2 and Si (in solid solution in the metallic Fe,Ni); 7. (7) As little as 1% by volume partial melting in an asteroid of any size will generate excess pressures in the melt much greater (10's of MPa) than the tensile strength of silicate materials and will initiate formation of cracks. This process will be enhanced by expansion of any exsolved or trapped gas phase in the asteroid and will lead to the formation of pathways (cracks growing into dikes) for melt migration; 8. (8) The pressure gradients implied when these pathways intersect the asteroid surface, coupled with the observed amounts of indigenous volatiles released from EH's and EL's, are easily sufficient to drive the pyroclastic volcanic eruptions envisaged by Wilson and Keil (1991).