Abstract– Whether a target is penetrated or not during hypervelocity impact depends strongly on typical impactor dimensions (Dp) relative to the absolute target thickness (T). We have therefore conducted impact experiments in aluminum1100 and TeflonFEP targets that systematically varied Dp/T (=D*), ranging from genuine cratering events in thick targets (Dp << T) to the nondisruptive passage of the impactor through very thin films (Dp >> T). The objectives were to (1) delineate the transition from cratering to penetration events, (2) characterize the diameter of the penetration hole (Dh) as a function of D*, and (3) determine the threshold target thickness that yields Dh = Dp. We employed spherical soda-lime glass (SLG) projectiles of Dp = 50–3175 μm at impact velocities (V) from 1 to 7 km s−1, and varied target thicknesses from microns to centimeters. The transition from cratering to penetration processes in thick targets forms a continuum in all morphologic aspects. The entrance side of the target resembles that of a standard crater even when the back of the target suffers substantial, physical perforations via spallation and plastic deformation. We thus suggest that the cratering-to-penetration transition does not occur when the target becomes physically perforated (i.e., at the “ballistic limit”), but when the shock pulse duration in the projectile (tp) is identical to that in the target (tt), i.e., tp = tt. This condition is readily calculated from equation-of-state data. As a consequence, in reconstructing impactor dimensions from observations of space-exposed substrates, we recommend that crater size (Dc) be used for the case of tp < tt, and that penetration hole diameter (Dh) be used when tp > tt. The morphologic evolution of the penetration hole and its size also forms a continuum that strongly depends on both the scaled parameter D* and on V, but it is independent of the absolute scale. The condition of Dh = Dp is approached at D* > 50. The dependence of Dh on T and V, however, is very systematic. This has led to new and detailed calibration curves, permitting the reconstruction of Dp from the measurement of either crater diameter or penetration-hole size in Al1100 and TeflonFEP targets of arbitrary thickness. We also placed witness plates behind penetrated targets to intercept the down-range debris plume, which is generally a mixture of both target and impactor fragments and melts. These witness plates also reveal that the debris plume systematically and diagnostically depends on D*. Thick targets shed spall debris only, and target thickness must be less than crater depth (Tc) to allow projectile material on the witness plate. Concentric plume patterns, accented by characteristic “hole saw” rings, characterize penetrated Al-targets at D* = 1–10, but they give way to distinctly radial geometries at D* = 10–20. Most of the target debris occupies the periphery of the plume, while the projectile fragments or melts reside in its central parts. The periphery of the plume is also typically more fine-grained than its center. At D* > 50, the exit plume is dominated by solid projectile fragments that progressively coagulate and overlap with each other, giving rise to compound craters. The latter have irregular crater interiors on account of the heterogeneous mass distribution of a collisionally produced, aggregate impactor. Similarly, complex craters are observed on LDEF and Stardust and they are produced by aggregate cosmic-dust particles containing large, dense components within a relatively low-density, fine-grained matrix. The witness-plate observations can also be used to address the enigmatic clustering of impact sites observed on Stardust’s aerogel and aluminum surfaces. We suggest that this clustering is difficult to produce by the collision of particles from comet Wild 2 with the Stardust spacecraft, and that it is more likely due to particle disaggregation in the comet’s coma.