Abstract— We conducted impact experiments into SiO2-based aerogel of uniform density (0.02 g cm−3) with spherical corundum projectiles. The highly refractory nature and mechanical strength of corundum minimizes projectile deformation and continuous mass loss by ablation that might have affected earlier experiments with soda-lime glass (SLG) impactors into aerogel targets. We find that corundum is a vastly superior penetrator producing tracks a factor of 2.5 longer, yet similar in diameter to those made by SLG. At velocities <4 km s−1 a cylindrical “cavity” forms, largely by melting of aerogel. The diameter and length of this cavity increase with velocity and impactor size, and its volume dominates total track volume. A continuously tapering, exceptionally long and slender “stylus” emerges from this cavity and makes up some 80–90% of the total track length; this stylus is characterized by solid-state deformations. Tracks formed below 4 km s−1 lack the molten cavity and consist only of a stylus. Projectile residues recovered from a track's terminus substantially resemble the initial impactors at V > 4 km s−1, yet they display two distinct surfaces at higher velocities, such as a blunt, forward face and a well-preserved, hemispherical trailing side; a pronounced, circumferential ridge of compressed and molten aerogel separates these two surfaces. Stringers and patches of melt flow towards the impactor's rear where they accumulate in a characteristic melt tip. SEM-EDS analyses indicate the presence of Al in these melts at velocities as low as 5.2 km s−1, indicating that the melting point of corundum (2054 °C) was exceeded. The thermal model of aerogel impact by Anderson and Cherne (2008) suggests actual aerogel temperatures <5000 K at comparable conditions. We therefore propose that projectile melting occurs predominantly at those surfaces that are in contact with this very hot aerogel, at the expense of viscous heating and associated ablation. Exposure to superheated aerogel may be viewed as extreme form of “flash heating.” This seems consistent with observations from the Stardust mission to comet Wild 2, such as relatively pristine interiors of rather large, terminal particles, yet total melting of most fine-grained dust components.