What are the real constraints on the existence and magnitude of the late heavy bombardment?

Geochronological evidence for and against a late heavy bombardment (LHB) spike in impact rates on the Moon and meteorite parent bodies is re-examined. In particular, we find that the sampling of impact melts on the Moon is strongly biased against older examples, possibly due to preferential surface deposition of such melts and/or blanketing and burial by basin ejecta (arguments that the bias might be due to pulverization of old impact melts at depth are incorrect). The apparent absence of melts older than 4 Ga thus provides much weaker evidence for a pre-nectarian lull in bombardment (which would define a post-nectarian cataclysm) than has been argued, although their absence compared with other rocks from depth may remain a weak constraint. Differences in rock-age histograms between meteorite parent bodies and the Moon may imply that different sampling biases exist for bodies in the asteroid belt; otherwise, the more straightforward interpretation is that the bombardment that affected the asteroids was more long-lasting than the lunar LHB. Since the strongest constraint on the dynamics of source populations for the LHB remains the very rapid decline in rate of basin formation from 3.90 to 3.85 Ga, we must re-establish that the associations of dated samples with particular basins are robust. Further studies of processes that can bias sampling are warranted, in particular more quantitative modeling of regolith and megaregolith evolution. In summary, we find that constraints either for or against the lunar cataclysm as a spike in the bombardment rate that commenced shortly before the formation of Nectaris are very weak. There may or may not have been a lull in bombardment before 3.9 Ga. Only the rapid decline in bombardment rate after Imbrium is fairly secure and can be adopted as a constraint to be matched by various dynamical scenarios for the impact history of the Moon. (c) 2007 Elsevier Inc. All rights reserved.