The Passive Baseline: Why Lunar Impact Cratering is Not Charge-and-Release

A random matrix theory (RMT) program has shown that single-source, charge-and-release systems exhibit Wigner-Dyson level repulsion, while superpositions of independent sources collapse to Poisson. A rigorous program needs a negative control: a system with no internal charge-and-release memory, which should not repel. Lunar impact craters are the natural candidate, since impacts are an externally driven, memoryless bombardment and the lunar surface has no reservoir to deplete. We test the spatial distribution of craters from the Robbins (2019) global database (1.3 million craters; we use the 6973 craters larger than 20 km and primary-dominated mare subsets). We first correct a subtle but decisive error. The classic Poisson value r-bar = 0.386 is the one-dimensional temporal baseline, whereas for two-dimensional nearest-neighbor spacings, complete spatial randomness (CSR) already gives r-bar about 0.57, because the nearest-neighbor operation excludes a disk around each point. Comparing crater spacings to 0.386 would falsely make any random 2D field look repulsive. The Rayleigh nearest-neighbor law of a 2D Poisson process furnishes the analytic reason the baseline must exceed 0.386 (its density vanishes at zero rather than peaking there), though the exact value has no simple closed form and is fixed by Monte Carlo. Crucially, this 2D correction does not invalidate the 0.386 baseline of the program's earlier studies, which are all strictly one-dimensional (beds along a depth axis, vein intercepts along a borehole, orbits along the semi-major-axis line); the 0.57 null applies only to unconstrained 2D point fields. This corrected geometric null is a reusable result for any future spatial RMT analysis. Calibrated against the correct CSR baseline, global large craters give r-bar = 0.695, modestly above CSR. Diagnostic tests reject both saturation (no diameter-cut trend) and hard-core exclusion (52 percent of nearest neighbors physically overlap). Resolving the field by terrain reveals two clean anchors. Old, isolated edge-maria (Mare Smythii, Mare Australe) are statistically indistinguishable from CSR (Clark-Evans R about 0.94, KS p above 0.1): impacts there are spatially random, exactly as a memoryless process should be. Young nearside maria (Oceanus Procellarum, Mare Nubium) are significantly clustered (Clark-Evans R about 0.79, z about -8), driven by secondary-crater chains and ejecta, not by any internal memory. Lunar craters thus occupy the opposite end of the spatial spectrum from the Earth and orbital systems: where charge-and-release carves repulsion, memoryless bombardment yields randomness or, where secondaries dominate, clustering. The RMT diagnostic cleanly separates all three states (clustering, randomness, and repulsion), and the negative control confirms the program's core claim from the other side: no charge-and-release mechanism, no level repulsion. This is the ninth system in the program and its essential negative control.