and deeper portions of epicontinental seas during accelerated outgassing associated with large-scale middle Cretaceous plate reorganization and development of the Pacific superplume (Larson, 1991a,b). These trace elements could have been concentrated as oxides and carbonates in basinal Cretaceous sediments in the presence of at least moderate amounts of benthic oxygen, especially during high sea-level and offshore sediment starvation; trace elements could also have been sequestered in solution in low-oxygen to anoxic water masses (OMZs). Expansion and benthic touchdown of the oceanic oxygen minimum zone during the Cenomanian-Turonian OAE would have remobilized sequestered trace elements from the seafloor and caused progressive advection of potentially toxic chemicals through the water column, with profound effects on the global marine biota. If the base of the food chain was affected by these advection events, it would have caused a far-reaching set of negative ecological feedback loops within the trophic web. The precise correlation of the first five C-T extinction steps, and two of the succeeding steps, with trace element enrichment horizons in epicontinental and continental shelf deposits of four continents, strongly supports the hypothesis of advection and trace element poisoning as a partial cause of C-T mass extinction. The relatively shallow water trace element enrichment layers found to be associated with C-T extinction events in high-resolution stratigraphic analysis, probably represent depositional sites of advecting trace elements from deeper ocean sources, situated at and above the oceanic redoxcline between the OMZ and the mixing zone.
A third hypothesis for C-T mass extinction, which draws on events described in the two previous hypotheses, focuses on the extraordinary rates and magnitudes of ocean-climate changes associated with the 1.46-m.y.-long C-T boundary interval, and their effects on a global biota narrowly adapted for the most part to the equable greenhouse environments that had been developing throughout the early and middle Cretaceous. Data from Pueblo, Colorado (Figures 3.3 and 3.4) are characteristic of many global boundary sections and show a series of exceptionally rapid, large-scale shifts in organic carbon values (representing rapid shifts in benthic oxygen levels); in d13C values (representing changes in carbon cycling) within the global positive d13C excursion; in d18O values (possibly representing rapid salinity and/or temperature changes); and in trace element values (probably representing one or more oceanic advection sequences). Virtually all late Cenomanian extinction events, and some lesser ones in the early Turonian, are correlative with one or more of these rapid, large-scale geochemical fluctuations. Whereas these geochemical signals can be strongly modified by diagenetic processes, the fact that similar changes occur in virtually all well-studied global C-T boundary sections suggests that they represent a primary ocean-climate signal. Major changes in ocean chemistry and temperature around the C-T boundary could well have been the primary cause of extinction events as they progressively exceeded the narrow adaptive ranges of many stenotopic Cenomanian-Cretaceous lineages. Of special interest in this theory is the general correlation of many apparently rapid environmental fluctuations with bedding rhythms representing 41,000- and 100,000-yr Milankovitch climate cycles (Barron et al., 1985; Fischer et al., 1985; Kauffman, 1988a; Glancy et al., 1993; Figures 3.3 and 3.5 herein). On the one hand this may suggest diagenetic modification of a primary signal in carbonate-rich versus carbonate-poorer facies of the C-T boundary interval, but to the degree that it represents a primary signal, it suggests that the Milankovitch climate cyclicity may have acted as an independent catalyst that drove an environmentally perched, greenhouse ocean-climate system to even greater levels of change, at rates dictated by the climate cycles themselves.
Finally, the possibility of extraterrestrial influences on the C-T mass extinction cannot be ruled out. The precise correlation of the first five late Cenomanian extinction events, or steps, with Ir enrichment of two to four or five times background levels leaves open the possibility of extraterrestrial sources for the iridium. Orth et al. (1989, 1990, 1993) have been cautious in suggesting extraterrestrial origins, pointing out instead an apparent similarity of the overall C-T trace element suite to those originating from deep mantle outgassing, and the fact that the C-T boundary interval was also a time of major plate rearrangement and superplume development (Larson, 1991a,b). On the other hand, four temporally clustered late Albian to late Cenomanian terrestrial impact craters are known with age error bars that overlap the C-T boundary (Grieve, 1982), suggesting an impact storm, or shower (sensu Hut et al., 1987); this is a conservative estimate of terrestrial impacts, when considering the amount of Cenomanian surface that has been subducted or is now covered by younger sediments/strata, vegetation, ice, and especially water. Because impacting is predictably spatially random (Grieve, 1982), and the late Cenomanian world was 80 to 82% covered by water near eustatic highstand, it is likely that at least four of five potential impactors would have fallen in the world seas and oceans during this impact shower. This statistically projects at least 20 impacts during the Cenomanian and early Turonian interval, the majority of which would be aquatic. Oceanic impacting would cause repeated, short-term stirring events. This hypothesis further predicts an extended duration for the Cenomanian impact shower, with high probability that it would overlap the C-T mass extinction interval. Recent discoveries of possible microtektites (Colombia) and multilamellate shocked quartz grains (Colorado) at two