We made petrographic and major-element chemical analyses of 50 Impact-Melt Splash (IMS) samples from the Apollo 16 landing site to assess their origin and the nature of their subregolith source. These melts possess volumes that are significantly larger than those agglutinates but smaller than those represented by fragments of (generally holocrystalline) melt sheets. Most of these melts occur as splashes that drape various host rocks. The majority appear to have been deposited onto their hosts inside growing crater-cavities or during ballistic flight and subsequently ''piggy-backed'' to their collection site(s). Macroscopically, IMS's exhibit a glassy appearance but display varied crystallinity under the microscope. Schlieren-rich glasses dominate the holohyaline areas. Most IMS's contain generally feldspathic monomineralic and lithic clasts, while purely regolithic materials (i.e., agglutinates) are absent. Crystalline portions are predominately spherulitic, but definitive criteria to distinguish between crystallization from a melt or solid state devitrification are lacking. More than half of the IMS's possess a distinct crystallization front that we attribute to devitrification. Chemically, most IMS's are fairly homogeneous on the scale of a thin section; variation among samples is much greater (e.g., Al2O3) varied from ~24 to 32 wt%; FeO from ~2 to 6 wt%). Such a wide range of melt compositions suggests that more than one impact generated these 50 IMS's. The chemistry of most IMS's is unlike the local regolith and appears to represent varied mixtures of VHA impact-melt breccias and anorthosite (i.e., are dimict-like). The host rocks are predominantly dimict breccias (70%) or one of their component lithologies (VHA impact-melt breccias and anorthosite).

Because of the short surface-residence times for these samples (determined from microcrater populations on all samples, and absolute exposure ages of 10 specimens), many IMS's might be associated with South Ray and Baby Ray Craters. In the absence of absolute chronological ages for melt formation, we cannot distinguish between a ''young'' scenario, in which the melts were formed during excavation of South Ray and/or Baby Ray Craters, and an ''old'' scenario, in which previously formed melts were exposed as part of the discontinuous ejecta of either or both of these craters. We prefer the former, in part because of the tight compositional cluster of most IMS's. In either case, we conclude that most of these samples or their progenitors resided at depth within the South Ray and/or Baby Ray targets. Furthermore, the Cayley Formation in the southern part of the Apollo 16 landing site is a polymict deposit predominantly composed of VHA impact-melt breccias and anorthosites. Both lithologies must be intimately mixed as they are in dimict breccias, and large, coherent units of individual lithologies appear to be limited to <50 m in dimensions.