We present experimental data and a model for the low-velocity (subsonic, 0--1000 m/s) penetration of brittle materials by both solid and hollow (i.e., coring) penetrators. The experiments show that penetration is proportional to momentum/frontal area of the penetrator. Because of the buildup of a cap in front of blunt penetrators, the presence or absence of a streamlined or sharp front end usually has a negligible effect for impact into targets with strength. The model accurately predicts the dependence of penetration depth on the various parameters of the target-penetrator system, as well as the qualitative condition of the target material ingested by a corer. In particular, penetration depth is approximately inversely proportional to the static bearing strength of the target. The bulk density of the target material has only a small effect on penetration, whereas friction can be significant, especially at higher impact velocities, for consolidated materials. This trend is reversed for impacts into unconsolidated materials. The present results suggest that the depth of penetration is a good measure of the strength, but not the density, of a consolidated target. Both experiments and model results show that, if passage through the mouth of a coring penetrator requires initially porous target material to be compressed to <26% porosity, the sample collected by the corer will be highly fragmented. If the final porosity remains above 26%, then most materials, except cohesionless materials, such as dry sand, will be collected as a compressed slug of material. ¿ American Geophysical Union 1996