Sediments containing components produced by accretionary events should contain sufficient geochemical evidence to constrain several of the variables involved in modeling the environmental consequences of such events. We consider the geochemical record expected for 3 modes of accretion: accretion from an interstellar cloud, non-impacting accretion of weak materials subjected to tidal and atmospheric disruption, and the impact of dense asteroidal or cometary materials. To constrain an accretionary event, it is important to determine: 1) the siderophile abundance patterns; 2) the duration of the event; 3) the geographic extent of the anomaly; 4) the physical nature of the extraterrestrial materials and their siderophile concentrations; and 5) the source of the terrestrial component. In attempting to constrain the Cretaceous-Tertiary and the Antarctic Basin late-Pliocene events we find that the necessary evidence is only partially available. It seems clear that the Cretaceous-Tertiary event generated fallout on a worldwide basis, and that its duration was ⩽1 ka, but other features are not yet well defined. The siderophile pattern is generally chondritic, but variations from site to site (and even sample to sample) indicate differing geochemical fractionations during deposition and make it impossible to associate the projectile with a specific group of meteorites. Future studies of unusually well-preserved basal layers at Caravaca and DSDP 465A may improve the precision with which the siderophile patterns are determined. Magnesium concentrations and isotopic studies argue against a mantle ejecta component in the boundary layer, and thus against the oceanic impact of a dense projectile.