Impact by clusters of projectiles rather than a single projectile can result from several processes: atmospheric breakup, tidal breakup, and ejecta from a large primary impact. Experiments have been performed in order to establish the characteristics of such events over a wide range of impact velocities (15 m/s to 6 km/s). At very low impact velocities (15--200 ms/s), clustered impacts were produced by launching a grouped projectiles of aluminum shot, steel shot, iron filings, and sand. At moderate to high velocities (0.8--6 km/s), pyrex spheres were shattered above the target during passage through aluminum foil or paper, thereby forming a well-defined cluster of fragments of overall radius rc. The ratio of the overall radius rc of the cluster to the radius rs of a solid impactor of the same mass provides a measure of the cluster dispersion. Sand and compacted pumice targets were used in order to compare qualitatively the additional effect of slight differences in target strength. The experiments reveal marked contrasts between impacts by clusters of projectiles and impacts by a single solid body. ''Tight'' clusters (rc/rs20) displace a factor of 10 less mass. This reduction in cratering efficiency is largely expressed as a shallow crater with an aspect ratio (daimeter/depth) as large as 30 for open clusters. The size and velocity of the cluster (as well as the density and strength contrast between the impactor and target) can dramatically affect crater morphology. Open clusters impacting compacted pumice produce a flat, hummocky floor with an incipient multiring pattern, whereas a tight cluster impacting the same target produces a central floor mound. Oblique impacts by symmetrical clusters form a characteristic array of V-shaped ridges pointing uprange.

The apex angle of the ridges depends on the cluster dispersion and impact angle. Inventories of postimpact projectile material reveal that the projectile is largely retained on the surface and spread downrange from the impact direction. The amount of projectile material retained on the surfaces increases with decreasing impact angle (from the horizontal) and with increasing strength of the target relative to the projectile. Clustered impact craters and lunar secondary craters bear striking similarities over a broad range of morphologic features. On this basis and on the basis of reasonable models of ejecta curtain structure, we suggest that these experiments provide new clues for understanding ejecta emplacementaround large lunar impact craters. When the ejecta curtain around large impacts is viewed as a thick wall of debris and clustered impactors are viewed as a unit section of such a curtain, then experimental results indicate that the continuous ejecta facies even for lunar craters larger than 100 km could contain as much as 90% primary material. Such a conclusion contrasts with many existing models that derive mixing ratios implicitly based on single, noninteracting impact events. Beyond the continuous ejecta facies, impacting clusters of ejecta provide a physical basis for understanding the wide variety of secondary morphologies and the large range in spectral signatures of primary material in crater rays.